Addition of Zinc, Manganese, and Iron to Growth Media Triggers Antibiotic Production in Bacterial Isolates From the Lower Atmosphere
With the growing number of antibiotic-resistant pathogens, there is a need for new antibiotics. Bacteria within the order Actinomycetales produce the majority of known antibiotic compounds but harbor cryptic secondary metabolic pathways that likely produce thousands more antibiotics awaiting discovery. This has recently renewed interests in bioprospecting for novel Actinomycetales in underexplored environments, such as the lower atmosphere, and activating cryptic secondary metabolic pathways in previously characterized members of this order. Many antibiotic-producing Actinomycetales have complex metabolic interactions with trace metals (such as iron, zinc, and manganese); adding these metals to bacterial growth media may trigger antibiotic production, which could lead to the discovery of novel drugs that may otherwise be overlooked during screening in the absence of such metals. To test this hypothesis, a collection of bacterial isolates, composed primarily of taxa within the order Actinomycetales, was obtained from the lower atmosphere. In support of our hypothesis, we found that several isolates produced antibiotics in the presence of trace metals, but not in their absence. Further investigation is required to identify the compounds being produced, but our results suggest that adding trace metals to growth media during high-throughput screening efforts may be a fruitful approach for identifying antibiotic-producing bacteria in large culture collections.
ccording to the World Health Organization (2015), pathogens are becoming more antibiotic-resistant than ever before, which is a problem caused and exacerbated by the overuse and misuse of existing antibiotics. As a result, there is a desperate need for novel antibiotics, but the approval rate of clinical antibiotics continues to decline (Donadio, Maffioli, Monciardini, Sosio, & Jabes, 2010). The order Actinomycetales within the phylum Actinobacteria, includes the genus Streptomyces, which produces two-thirds of known antibiotics (Barka et al., 2016; Watve, Tickoo, Jog, & Bhole, 2001). This genus is predicted to produce 150,000 to almost 300,000 antimicrobial compounds still awaiting discovery (Watve et al., 2001). Therefore, predictions about the next source of novel antibiotics often point to Actinomycetales (Bérdy, 2012; Goodfellow & Fiedler, 2010).
Goodfellow & Fiedler (2010) stated that by using selective techniques, such as sampling from understudied and extreme environments, novel Actinobacteria may be discovered. One such environment, is the lower atmosphere (Weber and Werth, 2015), which is defined by as the first 20km above ground level (Womack, Bohannan, & Green 2010). A multitude of both culture-dependent and culture-independent studies demonstrate that Actinobacteria are an omnipresent component of the aerial environment (Bowers et al., 2011; Fahlgren, Hagström, Nilsson, & Zweifel, 2010; Polymenakou, 2012; Shaffer & Lighthart, 1997; Weber & Werth, 2015). The lower atmosphere has several distinct advantages in the search for novel Actinomycetales. The lower atmosphere is a highly variable environment (Fahlgren et al., 2010) with dramatically oscillating temperatures (-56°C to 15°C), low relative humidity and high levels of ultraviolet radiation (Womack, Bohannan, & Green, 2010). These conditions may select for Actinomycetales over faster-growing bacterial taxa, such as many Proteobacteria (Weber & Werth, 2015). Exploring the lower atmosphere, given its potential to harbor antibiotic-producing bacteria, with selective cultivation methods may lead to the discovery of novel species and antibiotics. While not as commonly studied for their antibiotic-producing capabilities, Bacillus is another genus of bacteria that contains antibiotic-producing members and is commonly found in the lower atmosphere (Athukorala, Dilantha Fernando, & Rashid, 2009; Fahlgren et al., 2010; Shaffer & Lighthart, 1997).
Another approach to discover novel antibiotic compounds is to place a single organism under a wide array of culture conditions in an attempt to activate cryptic metabolic pathways that may produce uncharacterized metabolites with antibiotic properties known as the ‘One-Strain-Many-Compounds’ approach (Bode, Bethe, Höfs, & Zeeck, 2002). This method focuses on elucidating the up to 50 different pathways that have been found in a single Streptomyces strain, instead of screening massive numbers of isolates (Barka et al., 2016) in one set of conditions. The disadvantage of the latter approach is that it frequently results in the disposal of isolates, which do not produce antibiotic compounds under the one set of conditions tested (Bode et al., 2002). Culturing an isolate under altered conditions has been shown to encourage the production of antibiotics. Modifications of growth media that have been found to elicit the expression of secondary metabolite pathways include metals and amino acid additions (notably tryptophan; Palazzotto et al., 2015), as well as varied phosphate concentrations, carbon sources, nitrogen sources, and pH, to name a few (Bode et al., 2002; Iwai & Ōmura, 1982; Marwick, Wright, & Burgess, 1999; Weinberg, 1990).
Along these lines, amending growth media with trace metals is possibly a fruitful approach for activating cryptic secondary metabolic pathways, as trace metals are often required for the expression of secondary metabolite pathways (Haferburg et al., 2009; Huck, Porter, & Bushell, 1991; Iwai & Ōmura, 1982; Weinberg, 1990). Ochi and Hosaka (2013) indicated that rare earth metals and trace metals, notably manganese, zinc and iron may trigger the production of various secondary metabolic pathways. Actinobacteria have complex metabolic interactions with metals (Abbas & Edwards, 1989; Abbas & Edwards, 1990; Haferberg & Kothe, 2007; Paul & Banerjee 1983; Pettit, 2011). For instance, Yamasaki, Furuya, & Matsuyama (1998) showed that the impact of a metal ion on an isolate’s antibiotic production varies depending on the anionic species to which it is associated. Additionally, a single trace metal may inhibit antibiotic production in some cases, but encourage production in others, making trace metals potentially useful in the discovery of novel antibiotics (Abbas & Edwards, 1989; Abbas & Edwards, 1990; Bundale et al., 2015; Huck et al., 1991; Iwai & Ōmura, 1982; Weinberg, 1990).
In this study, the impact of amending media with three different metals (iron, manganese, and zinc) on the ability of bacterial isolates to produce antibiotics was examined. Isolates were obtained from the lower atmosphere and their ability inhibit the growth of Gram-positive and Gram-negative bacteria was assessed.
Materials and Methods
Air Sampling and Selection of Isolates
Air samples (180L) were collected onto 90mm petri plates containing Tryptic Soy Agar (TSA; Honeywell Specialty Chemicals, Seelze, Germany) amended with 50mg/mL cycloheximide (TSA-C), using a SAS SUPER 180 (Bioscience International, Rockville, MD) on the roof of the Gale Life Science Building on the Idaho State University campus (Pocatello, ID, USA). Plates were incubated at room temperature until Actinomycetales-like morphologies developed (approximately 5 days). Colonies with unique morphologies, several of which were Bacillus-like, were selected for isolation and purification.
After several streaks for purification, seven isolates of Actinomycetales-like and Bacillus-like morphologies and twelve isolates identified using the 16S rRNA gene from air samples collected between 2013 and 2015 were plated onto 60mm x 15mm petri dishes containing Luria Broth (LB; Thermo Fisher Scientific, Waltham, MA), King’s B (KB; Coldspring Harbor, 2009), tryptic soy agar (TSA), TSA with cycloheximide (TSA-C), TSA with 10mM (NH4)5Fe(C6H4O7)2 (TSA-Fe), TSA with 10mM MnCl2 (TSA-Mn), and TSA with 10mM ZnSO4•7H2O (TSA-Zn), KB with 10mM (NH4)5Fe(C6H4O7)2 (KB-Fe), KB with 10mM MnCl2 (KB-Mn), and KB with 10mM ZnSO4•7H2O (KB-Zn). The isolates were grown for one week at room temperature and then overlaid with molten LB agar amended with 1μL/mL of TSA broth containing either S. aureus or E. coli with average optical densities (measured at 600nm) of 1.885 and 1.648, respectively. Plates were incubated and monitored for zones of clearance around the isolates for seven days, indicating inhibition of S. aureus or E. coli growth.
Determination of Airborne Isolates Taxonomy
The twelve isolates utilized in the second set of inhibition assays were taxonomically identified by sequencing their 16S rRNA genes. DNA was extracted from isolates using the UltraClean Microbial DNA Isolation Kit (Mo Bio, Carlsbad, CA) following the manufacturer’s protocol, except only 25μL of the kit’s solution MD5 rather than 50μL was used in the final DNA elution step. The 16S rRNA gene was amplified using a Mastercycler Pro S (Eppendorf, Hamburg, Germany) in 20μL reaction volumes containing: 1x buffer solution, 200μM dNTPs, 3% DMSO, 0.5μM 27F primer, 0.5μM 1492R primer, 0.03UμL-1 Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltman, MA) using universal 27F 5’-AGAGTTTGATCMTGGCTCAG-3’ and 1492R 5’-TACGGYTACCTTGTTACGACTT-3’ primers using the following PCR conditions: initial denaturation at 98°C for 30 seconds; thirty cycles of 98°C for 30 seconds, 64°C for 45 seconds, 72°C for 30 seconds; final extension at 72°C for five minutes. Amplicons were purified using the MinElute PCR Purification Kit (QIAGEN, Hilden, Germany) and delivered to the Idaho State University Molecular Research Core Facility (Pocatello, ID, USA) for bidirectional sequencing. Contigs were formed from the bidirectional sequence reads using Sequencher version 5.1 (Gene Codes Corporation, Ann Arbor, MI). Sequences were taxonomically identified using the Basic Local Alignment Search Tool (BLAST).
The impact of four different complex media types with and without three different trace metal amendments on the ability of isolates to inhibit the growth of E. coli and S. aureus was tested. Antibiotic production was observed more frequently in media types containing trace metal amendments than in media lacking such amendments (Table 1). Three of the seven isolates inhibited S. aureus growth on all of the complex medias, while six of the seven isolates inhibited the growth of S. aureus on at least one of the trace metal-amended mediums. Many isolates failed to grow on TSA during the E. coli inhibition assay. Therefore, the impact of trace metals on the production of antibiotics that were effective against Gram-negative bacteria is unclear.
The inhibitory activity of isolate 31RC1 is of particular interest. This isolate only inhibited the growth of E. coli on TSA media. However, when iron and manganese were added, it became effective only to S. aureus. This indicates that the addition of metals may result in the production of completely different inhibitory compounds.
Eleven of the twelve isolates utilized in the second experiment were genera within Actinomycetales (Streptomyces, Nocardia, Rhodococcus) representing seven different species, and one isolate was a Staphylococcus spp. (Table 2). Inhibition assays were performed on KB media, and KB media amended with 10mM of iron, zinc, or manganese (Table 2).
None of the isolates inhibited E. coli on any media types. When the isolates were assayed against S. aureus, isolates that were identified as the same strain had distinct antibiotic-producing behaviors when trace metals were added. This was the most notable for the three isolates identified as Streptomyces pratensis, which had different inhibition capabilities, but all were able to inhibit S. aureus with the addition of zinc to the media. In some instances, metals improved antibiotic production capabilities especially in isolate 4oA2, where the addition of zinc alone resulted in inhibition. For every isolate that inhibited S. aureus on KB, inhibition was also observed on at least one medium with a metal amendment; while isolate QLW 21 produced no antibiotic on the zinc-amended media, it still produced antibiotics on manganese-amended and iron-amended media. Furthermore, replicate assay plates using the same isolate showed inconsistencies in inhibition such as in the case of isolate QAS 5, which inhibited S. aureus on manganese-amended media on only one of the four replicate assay plates.
With the need for novel antibiotics, there is renewed interest in screening large collections of bacteria to identify antibiotic-producing cultures. There is also renewed interest in identifying alternative sets of culturing conditions to trigger the production of uncharacterized secondary metabolites. By restricting screening assays to one culture condition, bacterial isolates of interest may be prematurely removed from further study. Using an array of conditions for initial antibiotic screens will likely facilitate the discovery of new antibiotics. Our data supports the hypothesis that adding trace metals to bacterial growth media may be an effective way to activate cryptic, secondary metabolite-producing pathways (Haferberg & Kothe, 2013).
Few metal-facilitated, antibiotic producing pathways have been studied in enough detail to determine the exact role that the metal ions play. It was hypothesized that metal ions are essential cofactors in antibiotic pathways (Iwai & Ōmura, 1982). The well-studied antibiotic production of actinorhodin in S. coelicolor A3(2) found that the addition of zinc to media may have allowed antibiotic producing, zinc-dependent enzymes to function when most intracellular zinc would otherwise be unavailable near the end of the exponential phase, thus predicting that the zinc affects translation directly and indirectly through feedback (Hesketh, Kock, Mootien, & Bibb, 2009).
Our results add to the growing collection of data that indicates that strain-to-strain variation in antibiotic production capacity is large, and culture identification at the species-level is not a reliable predictor of an organism’s ability to produce compounds with antibiotic properties. Schmidt et al. (2009) observed that bacterial isolates identified as the same species, respond to nickel using different mechanisms. This is perhaps a result of genetic mutations affecting antibiotic production akin to those examined by Higo, Hara, Horinouchi and Ohnishi (2012) or plasmid acquisition or loss. For instance, metabolic pathways encoding the production of molecules such as methylenomycin A have been found on transmittable plasmids (Wright & Hopwood, 1976).
Inconsistencies in the antibiotic production between inhibition assay replicates of the same isolate may be a result of the developmental stage reached by the isolate at the time of the assay. The stage of development triggers antibiotic production; generally occurring at the end of the exponential phase (Abbas & Edwards, 1990; Hesketh et al., 2009; Weinberg, 1990), but the mechanism is not entirely understood (Bibb, 2005). Nonetheless, trace metals, even if not the only determining factor, certainly appear to encourage antibiotic production and their addition to bacterial growth media should be considered when screening large collections of isolates or even single organisms for their ability to produce novel antibiotics.
The lower atmosphere is an accessible source of antibiotic-producing bacteria that should be further studied. Trace metals can trigger the production of antibiotics and should be included while screening isolates for antibiotic activity. Using replicate assay plates of isolates should also be considered during screening, as isolates inconsistently produced antibiotics. Although further study of this phenomenon is needed, it may be the result of the stage of development reached by the test colonies. Lastly, even if a bacterial strain is determined to be common and “well-studied” based on the identity of its 16S rRNA gene sequence, incorporating trace metal amendments into the ‘one-strain-many-compounds’ approach may result in the discovery of novel antibiotics.
We thank Abdullah Aljadi, Rachel Clinger, Naomi Veloso and Sora Matsunaga for their work on this study via their participation in the AMOEBA program at Idaho State University (National Science Foundation (NSF DUE 1140286; J.P. Hill, PI), and Jason T. Werth for technical support. Funding for this project was provided by a Dimensions of Biodiversity grant from the NSF (NSF DEB 1241069; C.F. Weber, PI).
Abbas, A. & Edwards, C. (1989). Effects of metals on a range Streptomyces species. Appl. Environ. Microbiol. 55(8): 2030-2035.
Abbas, A. & Edwards, C. (1990). Effects of metals on Streptomyces coelicolor growth and Actinorhodin production. Appl. Environ. Microbiol. 56(3): 675-680.
Athukorala, S. N. P., Dilantha Fernando, W. G., & Rashid, K. Y. (2009). Identification of antifungal antibiotics of Bacillus species isolated from different microhabitats using polymerase chain reaction and MALDI-TOF mass spectrometry. Can. J. Microbiol. 55: 1021-1032. doi:10.1139/W09-067
Barka E. A., Vatsa, P., Sanchez, L., Gaveau-Vaillant, N., Jacquard, C., Klenk, H-P., Clément, C., Ouhdouch, Y., & van Wezel, G. P. (2016). Taxonomy, physiology, and natural products of Actinobacteria. Micobiol. Mol. Biol. Rev. 80(1): 1-43. 25 November 2015 doi:10.1128/MMBR.00019-15.
Bérdy, J. (2012). Thoughts and facts about antibiotics: Where we are now and where we are heading. J. Antibiot. 65: 385-395. doi:10.1038/ja.2012.27
Bibb, M. J., (2005). Regulation of secondary metabolism in Streptomycetes. Curr. Opin. Microbiol. 8: 208-215. doi: 10.1016/j.mib.2005.02.016
Bode, H. B., Bethe, B., Höfs, R., & Zeeck, A. (2002). Big effects from small changes: Possible ways to explore nature’s chemical diversity. Chem. Bio. Chem. 3: 619-627. doi:10.1002/1439-7633(20020703)3:7<619::AID-CBIC619>3.0.CO;2-9
Bowers, R. M., Sullivan, A. P., Costello, E. K., Collett, J. L. Jr., Knight, R., & Fierer, N. (2011). Sources of bacteria in outdoor air across cities in the Midwestern United States. Appl., Environ. Microbiol. 77(18): 6350-6356. doi:10.1128/AEM.05498-11
Bundale, S., Begde, D., Nahikkar, N., Kadam, T., & Upadhyay, A. (2015). Optimization of culture conditions for production of bioactive metabolites by Streptomyces spp. Isolated form soil. Adv. Microbiol. 5: 441-451. doi:10.4236/aim.2015.56045
Donadio, S., Maffioli, S., Monciardini, P., Sosio, M., & Jabes, D. (2010). Antibiotic discovery in the twenty-first century: current trends and future perspectives. J. Antibiot. 63:423-430. doi:10.1038/ja.2010.62
Fahlgren C., Hagström, A., Nilsson, D., & Zweifel U. L. (2010). Annual variations in the diversity, viability, and origin of airborne bacteria. Appl. Environ. Micobiol. 76(9): 3015-3025. doi:10.1128/AEM.02092-09.
Goodfellow, M., & Fiedler, H-P. (2010). A guide to successful bioprospecting: informed by actinobacterial systematics. Antonie van Leeuwenhoek. 98(2): 119-142. doi:10.1007/s10482-010-9460-2
Haferburg, G., & Kothe, E. (2007). Microbes and metals: interactions in the environment. J. Basic Microbiol. 47: 453-467. doi:10.1002/jobm.200700275
Haferburg, G., Groth, I., Möllmann, U., Kothe, E., & Sattler, I. (2009). Arousing sleeping genes: shifts in secondary metabolism of metal tolerant Actinobacteria under conditions of heavy metal stress. Biometals 22: 225-234. doi:10.1007/s10534-008-9157-4
Haferburg, G., & Kothe, E. (2013). Activation of silent genes in actinobacteria by exploiting metal stress. In Amoroso, M. J., Benimeli, C. S., & Cuozzo, S. A. (Eds.), actinobacteria: application in bioremediation and production of industrial enzymes (56-73). Boca Raton, FL: CRC Press.
Hesketh, A., Kock, H., Mootien, S., & Bibb, M. (2009). The role of absC, a novel regulatory gene for secondary metabolism, in zinc-dependent antibiotic production in Streptomyces coelicolor A3(2). 74(6): 1427-1444. doi:10.1111/j.1365-2958.2009.06941.x
Higo, A., Hara, H., Horinouchi, S., & Ohnishi Y. (2012). Genome-wide distribution of AdpA, a global regulator for secondary metabolism and morphological differentiation in Streptomyces, revealed the extent and complexity of the AdpA regulatory network. DNA Res. 19: 259-273. doi:10.1093/dnares/dss010
Huck, T. A., Porter, N., & Bushell, M. E. (1991). Positive selection of antibiotic producing soil isolates. J. Gen. Microbiol. 137: 2321-2329.
Iwai, Y., & Ōmura, S. (1982). Culture conditions for screening of new antibiotics. J. Antibiot. 35(2): 123-141.
King’s B medium. (2009). Cold Spring Harbor Protocols. doi:10.1101/pdb.rec11326
Schmidt, A., Haferburg G., Schmidt, A., Lischke, U., Merten, D., Ghergel, F., Büchel, G., & Kothe, E. (2009). Heavy metal resistance to the extreme: Streptomyces strains from a former uranium mining area. Chem. Erde. Geochem. 60(2): 35-44. doi:10.1016/j.chemer.2007.11.002
Shaffer, B. T., & Lighthart B. (1997). Survey of culturable airborne bacteria at four diverse locations in Oregon: urban, rural, forest, and coastal. 2: 167-177. doi: 10.1007/s002489900046
Marwick, J. D., Wright, P. C., & Burgess, J. G. (1999). Bioprocess intensification for production of novel marine bacterial antibiotics through bioreactor operation and design. Mar. Biotechnol. 1: 495-507. doi:10.1007/PL00011806
Ochi, K., & Hosaka, T. (2013). New strategies for drug discovery: activation of silent or weakly expressed microbial gene clusters. Appl. Microbiol. Biotechnol. 97(1): 87-98. doi:10.1007/s00253-012-4551-9
Palazzotto, E., Renzone, G., Fontana, P., Botta, L., Scaloni, A., Puglia, A. M., & Gallo, G. (2015). Tryptophan promotes morphological and physiological differentiation in Streptomyces coelicolor. Appl. Microbiol. Biotechnol. 99: 10177-10189. doi:10.1007/s00253-015-7012-4
Paul A. K. & Banerjee A. K. (1983). Determination of optimum conditions for antibiotic production by Streptomyces galbus. Folia Microbiol. 28: 397-405. doi:10.1007/BF02879489
Pettit, R. K. (2011). Small-molecule elicitation of microbial secondary metabolites. Microbiol. Biotechnol. 4(4): 471-478. doi:10.1111/j.1751-7915.2010.00196.x
Polymenakou, P. (2012). Atmosphere: a source of pathogenic or beneficial microbes? Atmosphere 3: 87-102. doi:10.3390/atmos3010087
Watve, M. G., Tickoo, R., Jog, M. M., & Bhole, B. D. (2001). How many antibiotics are produced by the genus Streptomyces? Arch. Microbiol. 176: 386-390. doi:10.1007/s002030100345
Weber, C. F., & Werth, J. T. (2015). Is the lower atmosphere a readily accessible reservoir of culturable, antimicrobial compound producing Actinomycetales? Front. Microbiol. 6: 802. doi: 10.3389/fmicb.2015.00802
Weinberg, E. D. (1990). Roles of trace metals in transcriptional control of microbial secondary metabolism. Biol. Metals 2: 191-196. doi:10.1007/BF01141358
World Health Organization [WHO] (2015). Antibiotic Resistance. Retrieved April 25, 2016, from http://www.who.int/mediacentre/factsheets/antibiotic-resistance/en/
Womack, A. M., Bohannan, B. J. M., & Green J. L. (2010). Biodiversity and biogeography of the atmosphere. Phil. Trans. R. Soc. B 365, 3645-3653. doi:10.1098/rstb.2010.0283
Wright, L. F., & Hopwood, D. A. (1976). Identification of the antibiotic determined by the SCP1 plasmid of Streptomyces coelicolor A3(2). J. Gen. Microbiol. 95: 96-106.
Yamasaki, S., Furuya, N., & Matsuyama, N. (1998). Effects of specific ions in agar on antibiotic production by Burkholderia glumae. J. Fac. Agr. 43(3.4): 309-314.
rClone: A Synthetic Biology Tool That Enables the Research of Bacterial Translation
Understanding bacterial translation is of interest to many high school and college biology students, but hands-on research of this process has traditionally been inaccessible because of the expertise required for molecular cloning. There is also a need for more and better translational control elements that can be used for genetic circuits with applications in medicine, biotechnology, bioremediation, and biomaterials. We addressed these needs by inventing, constructing, and testing two new tools called rClone Red and rClone Blue that use Golden Gate Assembly for easy cloning of ribosome binding sites (RBSs), which control the initiation of translation. We validated rClone Red for student research by building libraries of mutated RBSs, selecting members of the library, determining their sequences, and measuring their strengths using a red fluorescence protein reporter gene (RFP). We compiled consensus sequences and tested them in rClone Blue with a blue chromoprotein reporter gene. Because our results showed higher RBS strengths for the blue chromoprotein gene than for the RFP gene, we developed and supported a hypothesis that an anti-RBS sequence was present in the RFP gene that sequestered the RBS with intramolecular mRNA base pairing. We demonstrated that rClone Red and rClone Blue are student-friendly research tools for mutational analysis of bacterial RBSs. In addition, we increased the number of RBS and bicistronic RBS elements for use in genetic circuits. We submitted 65 RBS and 62 bicistronic RBS sequences to the Registry of Standard Biological Parts to make them available to the worldwide synthetic biology research community.
Gene expression, the process by which the inherited information of genes is used to direct the function of cells, is regulated in all cells because not all genes are needed all the time or under all circumstances (Hijum, Medema, & Kuipers, 2009). Gene expression begins with transcription, the process by which the DNA base sequence of a gene is converted into RNA sequence information. For genes that encode proteins, the messenger RNA (mRNA) product of transcription is used during translation to encode the sequence of amino acids in a protein. The sequence of bases in mRNA is translated by the ribosome, which is composed of a large (50S) and a small (30S) subunit. Translation is initiated when the 16S ribosomal RNA (rRNA) of the small ribosomal subunit base pairs to a conserved sequence in the mRNA, called the ribosome binding site (RBS; Figure 1).
After the small ribosomal subunit binds to the RBS, the large ribosomal subunit attaches to the small subunit to begin translation of the mRNA into a chain of amino acids. The mRNA bases are read as triplet codons that interact by base pairing with anticodons in transfer RNA (tRNA) molecules, which carry amino acids to the growing protein chain (Malys & McCarthy, 2010). As shown in Figure 1, RNA-RNA base pairing typically involves the Watson-Crick base pairs of G with C, and A with U, but G can also base pair with U. The conventional understanding is that the strength of a given RBS is determined by the strength of its base pairing interactions with the 16S rRNA (Shine & Dalgarno, 1974). In natural bacterial genomes, there is a wide variety of RBS sequences and RBS translational strengths that have resulted from natural selection for global patterns of gene expression. The relationship between RNA base pairing and the strength of an RBS also explains how synthetic RBSs can be produced with widely varying strengths.
In addition to intermolecular base pairing, intramolecular base pairing affects the strengths of RBSs. The ability of RNA to engage in intramolecular base pairing is well established (Busan & Weeks, 2013). RBS elements can be disabled by RNA folding, as is the case in riboswitches (Breaker, 2012). The RNA in riboswitches adopts an OFF state when the RBS is bound by a complementary anti-RBS sequence within the mRNA. The addition of a small molecule ligand that binds to the folded RNA changes the RNA shape so that the RBS is available for interaction with the 16S rRNA during the ON state.
Understanding the function of RBSs informs the new discipline of synthetic biology, which uses engineering principles and molecular cloning methods for the construction of parts, devices, and systems, with applications in areas such as medicine, energy, and the environment (Khalil & Collins, 2010). Synthetic biologists have studied the function of RBSs as interchangeable parts that retain their strengths in the context of any gene expression device, but the interchangeability of simple RBSs has come under question (Mutalik et al., 2013). An RBS that functions efficiently upstream of one gene will not necessarily function efficiently upstream of a different gene. The C dog bicistronic RBS was intended to solve this inconsistency (Mutalik et al., 2013). As shown in Figure 1, the C dog RBS uses two RBSs to preserve the strength of the second RBS upstream of any gene of interest. The first RBS initiates translation of mRNA into a short leader polypeptide. The coding sequence for the leader polypeptide extends past the second RBS and ends at the start codon of the gene of interest. Translation of the mRNA into the leader polypeptide is hypothesized to disrupt gene-specific mRNA folding that could sequester the second RBS and reduce its strength of translation initiation.
We still have more to learn about the translation initiation in natural and synthetic systems. However, mutational analysis of RBSs has historically been challenging for high school and undergraduate researchers, because it requires substantial expertise in molecular cloning methods. For example, the introduction of mutations into bacterial translational control elements usually involves purification of restriction-digested DNA from agarose gels, a method that is difficult to master and time-consuming. In addition, an often complicated strategy involving compatible restriction enzyme sites must be tailored for the assembly of DNA parts during each molecular cloning project. We addressed these two problems with the design, construction, and testing of a new molecular tool called rClone. rClone uses Golden Gate Assembly (GGA), which eliminates the need for gel purification of DNA and enables the standardization of assembly for molecular cloning (Weber et al., 2011). We leveraged the simplicity and reliability of GGA for rClone, which enables the cloning of RBSs, as we did previously for pClone, which enables the cloning of promoters (Campbell et al., 2014). After RBSs and promoters are cloned, rClone and pClone help one measure their function easily using convenient reporter genes. These two plasmids make the mutational analysis of bacterial gene expression faster, cheaper, and more accessible to high school and college student researchers. We demonstrated the power of rClone for RBS mutational analysis by producing libraries of thousands of mutant simple RBSs and C dog RBSs, and investigating the libraries to learn about the sequence requirements for the control of bacterial translation. We used this mutational analysis to address the hypothesis that mutations in both simple RBSs and bicistronic C dog RBSs would affect the efficiency of translation. Our results culminated in consensus sequences that represent new and testable hypotheses about the sequence requirements for RBSs, as well as a collection of 127 new RBSs that can be used by the synthetic biology community.
Materials and Methods
For the construction of rClone Red, we used PCR and GGA to remove a RBS from tClone Red, which we built previously for the study of transcriptional terminators. We built rClone Blue by using PCR and GGA to replace the RFP reporter gene in rClone Red with a reporter gene that encodes a blue chromoprotein. We submitted rClone Red and rClone Blue to the Registry of Standard Biological Parts as part numbers BBa_J119384 and BBa_J119389, respectively (MIT Working Group, 2005; http://parts.igem.org/Part:Bba_J119384; http://parts.igem.org/Part:Bba_J119389). We used rClone Red and rClone Blue to construct four different mutant RBS libraries. We explored the libraries by picking colonies, determining the DNA sequences of the mutant RBSs they carry, and measuring the strengths of the RBSs with the fluorescence produced by the RFP reporter gene. Additional details for the experimental procedures by which we constructed and used rClone Red and rClone Blue can be found in Supplemental Information.
Cloning RBSs with rClone Red and rClone Blue
rClone plasmids use the type IIs restriction enzyme BsaI and GGA to clone RBSs (Figure 2A). rClone Red contains a “backward facing” Green Fluorescent Protein (GFP) expression cassette between a “forward facing” promoter and a “forward facing” RFP reporter gene. The GFP cassette is flanked by BsaI binding sites. The GFP and the BsaI sites are removed during GGA when BsaI cuts out the GFP cassette, leaving behind sticky ends. Annealed oligonucleotides that encode the RBS to be cloned also have sticky ends complementary to those produced by BsaI digestion of rClone Red. DNA ligase present in the GGA reaction attaches the oligonucleotides that have annealed to rClone Red. Two categories of clones result from the transformation of the GGA reaction into E. coli bacteria (Figure 2B). Transformation of the original rClone Red plasmid results in colonies that express GFP because the GFP cassette is still within the RFP cassette. Successful cloning of a new RBS sequence produces colonies that are not green. The strength of the RBS determines the amount of RFP produced in non-green colonies.
Cloning RBSs with rClone Blue works the same way as cloning RBSs with rClone Red (Figure 2A). rClone Blue contains a “backward facing” GFP expression cassette between a “forward” promoter and the AmilCP blue reporter gene. The GFP cassette is flanked by BsaI binding sites and is removed during GGA when BsaI cuts out the GFP cassette, leaving behind sticky ends. An RBS, or a library of RBSs, with complementary sticky ends, can be cloned using GGA. As with rClone Red, green and not green colonies can occur (Figure 2B, left side). Unsuccessful GGA results in green colonies because the GFP cassette was not removed. Colonies that do not express GFP are the products of successful GGA.
Library Design and Exploration
We used rClone Red and rClone Blue to study simple RBSs and C dog bicistronic RBSs. We developed two mutation strategies for both types of RBSs. One strategy produced 65,536 possible RBS sequences by varying all eight bases of the RBS and replacing them with N’s, where N refers to A, T, C or G. The other strategy preserved the middle two highly conserved bases, resulting in 4,096 different RBS sequences. We constructed libraries using both strategies for the simple RBS (Figure 3A) and the C dog bicistronic RBS (Figure 3B). The strengths of the RBSs determines the amount of blue chromoprotein produced in the non-green bacterial colonies. The strength of a given RBS determines the phenotype of colonies in a library (Figure 4). Colonies that resulted from failed assemblies are green whereas colonies from a successfully cloned RBS are not green. The intensity of the red fluorescence of a colony is determined by the strength of the RBS it contains. A strong RBS results in a visibly red colony whereas a weak RBS results in a colony that is not red.
We explored our simple RBS N6 library by picking 33 clones (Figure 5A). The RBS strengths are expressed as percentages compared to the strongest RBS we found in our libraries, which was C dog mutant C10. The strongest simple RBS we found in the library was RM, with a relative strength of 31.1%. Of the 33 clones we examined, 19 of them had a strength of less than 10%. We also picked 32 mutant clones from the simple RBS N8 library (Figure 5B). The strongest simple RBS among the clones we picked from the N8 library was B24 with a relative strength of 51.3%. Twenty-six of the 32 clones we picked from the simple RBS N8 library had a strength less than 10%.
We picked 26 clones from the C dog N6 library (Figure 6A). The strongest one we found in the N6 C dog library was D18, which had a relative strength of 84.9%. Only 3 of the 26 clones had a strength less than 10%. We picked 36 clones from the C dog N8 library (Figure 6B). The strongest C dog RBS we picked from the N8 C dog library was C10, which was the strongest of all the examined RBS clones and it was used as the relative standard for all comparisons. Ten of the 36 clones had a strength less than 10%.
Building and Testing RBS Consensus Sequences
A consensus sequence expresses the frequency at which each base occurs in each position of a DNA or RNA sequence for a library of sequences (Schneider & Stephens, 1990). We used the information from Figures 5 and 6 to construct a weighted consensus (see details in Supplemental Information) sequence for the RBS N6, RBS N8, C dog N6, and C dog N8 libraries, and the results are displayed in Figure 7. Figure 7 also shows a consensus sequence from 149 RBSs in the E. coli genome (Schneider et al., 1990). To validate our consensus sequences, we selected the base with the highest score and the second strongest base if its score was within half a standard deviation of the highest. For example, simple RBS N6 had a consensus formula of CRCGAGGT, where R stands for purine (A or G). Therefore, testing the simple RBS N6 consensus required two sequences: CGCGAGGT (RBS1) and CACGAGGT (RBS2). We cloned each validating consensus sequence into rClone Red and measured their relative RBS strengths (Figure 8). The strengths of the validating consensus clones for simple RBSs are lower than some members of the RBS N6 library from which the consensus was generated. For the simple RBS N8 library, one of the validating clones was weaker than some members of the simple RBS N8 library, but the other validating clone was stronger than all but one of the simple N8 sequences we had tested. For C dog N6, the validating sequence had the second highest score among the tested clones. Of the two C dog N8 validating clones, one was stronger than all of the tested clones and the other was the fifth strongest in its source library.
We also cloned all of the consensus validating sequences into rClone Red in their reciprocal simple RBS and C dog contexts. We gave consensus testing sequences two-letter abbreviations based on their context and origin, using R for simple RBS and C for C dog RBS. For example, we cloned the consensus validating sequence RBS1, derived from the simple RBS N6 library consensus, into a C dog RBS and called it CR1. Likewise, we tested the validating sequence called Cdog1 as a standalone simple RBS called RC1 (Figure 8). For three of the simple RBS sequences tested in the C dog context, the strength increased by an average of 2.4-fold. All of the C dog RBS sequences decreased by an average of 28-fold when they were tested as simple RBSs.
To investigate the interchangeability of simple RBSs and C dog RBSs upstream of a different reporter gene, we cloned all of the consensus validating sequences into rClone Blue (Figure 9). The amount of blue chromoprotein produced was proportional to RBS strengths. We categorized each clone as strong, medium or weak based on how blue the colonies appeared in white light. We used RFP strengths to categorize the same validating sequences when tested in rClone Red. rClone Blue and rClone Red shared three of the five strong validating sequences (CC1, CC2 and CC3). Only two of the five medium validating sequences (RR1 and RR4) and one weak validating sequence (RR3) were shared in rClone Red and rClone Blue.
Base Pairing of the 16S rRNA with RBSs
Our results provide support for the hypothesis that base pairing between the 16S rRNA of the ribosome and the RBS affects the strength of a given RBS. The strongest RBSs from all four libraries have an average of 6.75 out of 8 bases pairs with the 16S rRNA. The weakest RBSs from all four libraries have an average of only 2.25 base pairs with the 16S rRNA. Our results also showed that the strengths of simple and C dog RBSs were lower in rClone Red than they were in rClone Blue. We hypothesized that the RFP mRNA has a translation blocking anti-RBS that is absent in the blue chromoprotein mRNA. An anti-RBS within RFP mRNA would explain why many of the validating sequences worked better in rClone Blue than in rClone Red. To investigate the validity of our anti-RBS hypothesis, we explored the possible secondary structures formed in both rClone Red and rClone Blue using a web-based program called mFold (Zuker, 2003). The program searches all the possible intramolecular base pairing interactions for an RNA sequence. It presents the most stable structures and calculates their free energies as ΔG in kcal/mol. Free energies are a measure of the stability of structures and more negative free energies indicate more stable ones. We used mFold to predict the secondary structure and calculate the ΔG for the RC1 RBS in the context of both rClone Red and rClone Blue mRNAs (Figure 10). The RC1 RBS had a strength of only 4.0 in rClone Red but was among the strongest RBSs in rClone Blue (see Figure 9). The mFold results show that in the rClone Red mRNA, the RC1 RBS is base paired to a part of the RFP coding sequence immediately downstream of its start codon, forming a stem and loop structure. The RC1 RBS does not form a stable a stem and loop structure with the blue chromoprotein mRNA. The mFold results support our hypothesis that the RFP gene has an anti-RBS which sequesters simple RBSs and inhibits the initiation of translation.
We looked for base pairing between the anti-RBS in the RFP coding sequence and each of the consensus validating RBS sequences. Validating sequences that base pair with the 16S rRNA also base pair with the anti-RBS, which explains the preponderance of low strength RBSs in the rClone Red libraries. Every RBS sequence that had the potential to be a high strength RBS also was more likely to interact with the RFP anti-RBS. For example, seven of the eight bases of RC2 can pair with the 16S rRNA and with the anti-RBS. RC2 could be a strong RBS because it base pairs well with the 16S rRNA, but it also binds to the anti-RBS in the RFP mRNA.
Testing the Anti-RBS Hypothesis
To test whether the anti-RBS was the cause of the reduced RFP production of rClone Red RBS constructs, we mutated five bases in the three codons predicted to base pair with RC1 without changing the amino acids they encode (Figure 11). Two of the three codons specify serine which can be encoded by six codons. Having six codons to choose from provided more mutational options to reduce base pairing between the RFP mRNA and RC1 RBS. The result of introducing three synonymous codons is called RFP version 2. We used mFold to predict the structure and stability (ΔG) of the encoded mRNA of RFP version 2 and compared it to that of RFP version 1. As shown in Figure 11, the RBS is not sequestered in RFP version 2, and the RFP version 2 stem and loop structure is less stable than that of RFP version 1. We also compared the RFP produced with RC1 in the original rClone Red to the RFP produced with RC1 in rClone Red version 2 encoding RFP version 2. Consistent with the mFold analysis, rClone Red version 2 produced ten times more RFP than rClone Red version 1 when both used the same simple RBS. With a relative RFP production of 36%, RC1 would be moved from the weak category to the medium category in Figure 9.
Codon Usage Bias in the Anti-RBS
Our investigation uncovered several new RBS consensus sequences that differ from those published over forty years ago (Shine & Dalgarno, 1974). Because of the modular nature of both rClone plasmids, it was straightforward for us to clone the consensus verifying sequences for simple RBS and C dog RBS in rClone Red and rClone Blue. Because we compared the verifying sequences directly in both plasmids, we uncovered a previously undocumented anti-RBS in the RFP mRNA. When we produced rClone Red version 2, we confirmed that RC1 was stronger than it had been in rClone Red version 1. It is interesting that RC1 placed upstream of RFP version 2 had medium strength (36% in Figure 11) whereas RC1 in rClone Blue was among the strongest of the constructs we tested (Figure 9). A possible explanation for these observations might be connected to the fact that, for a given amino acid, E. coli uses some codons more often than others. This is referred to as codon usage bias. Researchers who want to produce a foreign protein in E. coli often conduct a process called codon optimization to pick the most frequently used codons to gain translation efficiency (Shin, Bischof, Lauer, & Desrosiers, 2015). Perhaps part of the reduction in the efficiency of translation for RFP mRNA was because the UCC serine codon found in rClone Blue is more optimal (17% usage in E. coli) than the UCA (12% usage) and AGU (13% usage) serine codons in RFP version 2 (Maloy, Stewart, & Taylor, 1996). As with any engineering solution to a problem, compromises must be made such as reducing the stem and loop structure while maximizing codon optimization. Nevertheless, we successfully redesigned and tested rClone Red to be more responsive to strong RBS sequences.
In the context of synthetic biology, understanding gene expression allows us to engineer bacterial cells to produce pharmaceuticals, attack cancer cells, neutralize environmental pollutants, and synthesize biofuels (Khalil & Collins, 2010). The Registry of Standard Biological Parts contains over 20,000 DNA parts and devices for use in synthetic biology (MIT Working Group, 2005). We submitted our collection of 65 simple RBSs and 62 bicistronic RBSs as four Registry parts (Bba_J119390 to BbaJ119393). Our contribution has increased the number of RBSs in the Registry from 56 to 183, all of which can be used for the design and construction of genetic circuits that enable bacteria to produce pharmaceuticals, biofuels, and chemical commodities. Various RBS strengths can be used to fine-tune the expression of genes encoding enzymes in a biosynthetic pathway. We also submitted information to the Registry about rClone Red (Bba_J119384) and rClone Blue (Bba_J119389) to make them available to the global synthetic biology research community, including high school and college student researchers.
Future Prospects for Mutational Analysis of Translation Initiation
There are many opportunities for mutational analysis of bacterial translation. For example, students could search for anti-RBS sequences in thousands of bacterial genomes to discover new examples of translational regulation (Li, Oh, & Weissman, 2012). For synthetic biology applications, learning how to disable anti-RBS elements would be important for improving the interchangeability of RBSs and achieving higher levels of gene expression. Another interesting topic for exploration would be the fitness cost of gene expression (Pope, McHugh, & Gillespie, 2010). There is a cost to bacterial cells of producing orthogonal proteins such as RFP and blue chromoprotein. Fitness cost could be measured by comparing growth rates of clones that produce RFP or blue chromoprotein with various RBS strengths. Fitness costs of promoters, RBSs, alleles, and protein degradation tags could be used to determine optimal orthogonal gene expression.
Suite of Synthetic Biology Cloning Tools
rClone Red and rClone Blue are members of a suite of synthetic biology tools that enables high school and undergraduate research students to perform original research on bacterial gene regulation and expression. Each tool leverages the ease with which GGA can be used to clone user-defined regulatory elements and measure their effects on expression of a reporter gene. pClone Red (Bba_J119137) and pClone Blue (Bba_J119313) enable students to conduct promoter experiments of their own design (Campbell & Eckdahl, 2015; Campbell et al., 2014; Eckdahl & Eckdahl, 2016; Eckdahl & Campbell, 2015). tClone Red (Bba_J119367) facilitates investigation of transcriptional terminators that use alternative RNA folding states to control gene expression. We have designed and constructed actClone (Bba_J100204) to study DNA binding sites for activator regulatory proteins, and repClone (Bba_J100205) to study repressor protein binding sites. Each of these tools use the same GGA strategy as rClone Red and rClone Blue to enable students to explore the sequence requirements of bacterial DNA regulatory elements. Our suite of synthetic biology tools is appropriate for high school and undergraduate researchers to conduct original research in bacterial gene regulation and expression.
We would like to thank Dr. Jay Meyers of Saint Joseph Central High School for his support and encouragement. Support from National Science Foundation (http://www.nsf.gov/) RUI grants MCB-1329350 to Missouri Western State University and MCB-1120578 to Davidson College is gratefully acknowledged, as well as the James G. Martin Genomics Program at Davidson College.
Breaker, R. R. (2012). Riboswitches and the RNA world. Cold Spring Harbor Perspectives in Biology, 4(2). doi: 10.1101/cshperspect.a003566
Busan, S., & Weeks, K. M. (2013). Role of Context in RNA Structure: Flanking Sequences Reconfigure CAG Motif Folding in Huntingtin Exon 1 Transcripts. Biochemistry, 52(46), 8219–8225. doi: 10.1021/bi401129r
Campbell, A. M., Eckdahl, T., Cronk, B., Andresen, C., Frederick, P., Huckuntod, S., Yuan, J. (2014). pClone: Synthetic Biology Tool Makes Promoter Research Accessible to Beginning Biology Students. CBE Life Sciences Education, 13(2), 285–296. doi: 10.1187/cbe.13-09-0189
Campbell, A. M., & Eckdahl, T. T. (2015). CourseSource pClone Intro Biology Lesson. Retrieved from http://www.coursesource.org/
Eckdahl, A. J., & Eckdahl, T. T. (2016). Mutational Analysis of Transcriptional Initiation in Bacteria. Journal of Young Investigators, 31(3), 1–8.
Eckdahl, T. T., & Campbell, A. M. (2015). CourseSource pClone Genetics Lesson. Retrieved from http://www.coursesource.org/
Khalil, A. S., & Collins, J. J. (2010). Synthetic biology: applications come of age. Nat Rev Genet, 11(5), 367–79. doi: 10.1038/nrg2775
van Hijum, A. S., Medema, M. H., Kuipers, O. P. (2009). Mechanisms and evolution of control logic in prokaryotic transcriptional regulation. Microbiol Mol Biol Rev, 73(3), 481-509.
Li, G.-W., Oh, E., & Weissman, J. S. (2012). The anti-Shine-Dalgarno sequence drives translational pausing and codon choice in bacteria. Nature, 484(7395), 538–541. doi: 10.1038/nature10965
Maloy, S. R., Stewart, V. J., & Taylor, R. K. (1996). Genetic Analysis of Pathogenic Bacteria: A Laboratory Manual. Plainview, N.Y: Cold Spring Harbor Laboratory Press.
Malys, N., & McCarthy, J. E. G. (2010). Translation initiation: variations in the mechanism can be anticipated. Cellular and Molecular Life Sciences, 68(6), 991–1003. doi: 10.1007/s00018-010-0588-z
MIT Working Group. (2005). Registry of Standard Biological Parts. Retrieved September 2, 2015, from http://partsregistry.org/Main_Page
Mutalik, V. K., Guimaraes, J. C., Cambray, G., Lam, C., Christoffersen, M. J., Mai, Q.-A., … Endy, D. (2013). Precise and reliable gene expression via standard transcription and translation initiation elements. Nature Methods, 10(4), 354–360. doi: 10.1038/nmeth.2404
Pope, C. F., McHugh, T. D., & Gillespie, S. H. (2010). Methods to determine fitness in bacteria. Methods in Molecular Biology (Clifton, N.J.), 642, 113–121. doi: 10.1007/978-1-60327-279-7_9
Schneider, T. D., & Stephens, R. M. (1990). Sequence Logos: A New Way to Display Consensus Sequences. Nucleic Acids Res, 18, 6097–6100.
Shin, Y. C., Bischof, G. F., Lauer, W. A., & Desrosiers, R. C. (2015). Importance of codon usage for the temporal regulation of viral gene expression. Proceedings of the National Academy of Sciences, 1121(45), 14030–14035.
Shine, J., & Dalgarno, L. (1974). The 3′-Terminal Sequence of Escherichia coli 16S Ribosomal RNA: Complementarity to Nonsense Triplets and Ribosome Binding Sites. Proceedings of the National Academy of Sciences, 71(4), 1342–1346.
Weber, E., Gruetzner, R., Werner, S., Engler, C., & Marillonnet, S. (2011). Assembly of designer TAL effectors by Golden Gate cloning. PLoS One, 6(5), e19722. doi: 10.1371/journal.pone.0019722
Zuker, M. (2003). Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Research, 31(13), 3406–3415. doi: 10.1093/nar/gkg595
Response of Acacia tortilis to Elephant Browsing in Tarangire National Park, Tanzania: Possible Above-Ground Compensation?
Large herbivore browsing leads to above-ground compensatory growth for some species of Acacia trees, but strength and variation of the relationship are poorly understood. Acacia tortilis is a keystone species in East African savannas and experiences a wide range of browsing pressure. In this study, terminal bud scale scars were used to measure yearly shoot elongation in A. tortilis experiencing various levels of elephant browsing at three mesic sites in northern Tarangire National Park, Tanzania. For all four years of growth, twig elongation remained very similar across elephant browsing pressure from minimal to heavy. Although never significant, twig elongation in 2013, 2011 and 2010 was somewhat higher in trees experiencing moderate elephant browsing pressure, suggesting a possible tendency toward compensatory growth. However, growth in 2012 had a slight negative correlation with browsing pressure. Further research can test whether A. tortilis compensates for moderate elephant browsing by elongating twigs more and if that growth slows slightly at higher browsing pressure. Overall A. tortilis tolerates elephant browsing extraordinarily well, but because the elephant population in Tarangire National Park is large and growing, managers should continue to monitor this keystone for decreased growth and tree mortality.
Woody plants provide energy and nutrients for many mammals in African savannas. Throughout Africa, Acacia spp. trees provide the primary food source for numerous browsers, and also serve as important habitat for birds (Dharani, Kinyamario, Wagacha & Rodrigues, 2009). Many species, including black rhinoceros, giraffe, grey duiker, dik-dik, grysbok, klipspringer, gerenuk, dibatag, bushbuck, and kudu, feed exclusively on woody browse (Owen-Smith, 1982). Because Acacia trees fix nitrogen, this also leads to higher forage quality in grasses underneath them compared to areas not under their canopies (Ludwig, De Kroon & Prins, 2008; Mopipi, Trollope & Scogings, 2009). Understanding the conditions that allow continued Acacia growth despite browsing will allow conservationists to monitor and prevent mortality of this keystone savanna species.
There are many variables in woody savanna which affect the productivity of trees, including Acacia spp. Disease decreases growth, rainfall increases growth, and fire has variable effects on growth based on the severity of the fire and the characteristics of various Acacia species (Dharani et al., 2009; Fornara, 2008; Mopipi et al., 2009; Otieno, Kinyamario & Omenda, 2001; Scogings, Johansson, Hjalten & Kruger, 2012). However, effects of herbivory and the dynamic interactions between browsers and woody plants have been heavily disputed. Several researchers found that intense large herbivore browsing results in compensatory above-ground plant growth and can lead to alternative stable states (Dublin, Sinclair & McGlade, 1990; Jachmann & Bell, 1985; Smallie & O’Connor, 2000). However, others have found that large herbivore populations, such as elephants (Loxodonta africana) or giraffes (Giraffa camelopardalis), have a negative effect on woody vegetation growth (Guldemond & Van Aarde, 2008; Chira & Kinyamario, 2009; Pellew 1984).
Past findings on the effects of browsing on sub-Saharan African Acacia species vary by mammal, tree species and response measurement. Browsing by non-elephant large mammals resulted in compensatory growth, leading to higher stem diameter growth of A. xanthophloea (Dharani et al., 2009). Similarly, simulated browsing of A. xanthophloea, A. tortilis, A. hockii (Pellew, 1984) and A. karroo (Stuart-Hill & Tainton, 1989) resulted in increased shoot growth and competitive ability when defoliation rates were between 25-50%. Du Toit, Bryant & Frisby (1990) and Chira & Kinyamario (2009) respectively found that heavy browsing by elephants led to an increase in shoot growth and higher nitrogen concentration in A. nigrescens foliage and coppice growth in A. brevispica. Conversely, Gandiwa, Magwati, Zisadza, Chinuwo, & Tafangenyasha (2011) found that South African A. tortilis with moderate to high levels of elephant browsing had smaller mean tree density, height, and basal area, compared to trees with low level of browsing. While recent patterns in precipitation, defoliating insects, disease, and soil fertility can also affect above-ground growth, current evidence suggests that compensatory growth responses occur only under certain levels of browsing pressure before tree mortality and/or resource depletion occurs.
Although compensatory growth in response to browsing has been observed in many Acacia species, the mechanism(s) and evolutionary trade-offs have been heavily disputed. Du Toit et al. (1990) found lower levels of tannins – naturally occurring plant polyphenols that discourage herbivore browsing by lowering metabolic efficiency (Chung, Wong, Wei, Huang & Lin, 1998) – and higher nitrogen and phosphorus content in heavily browsed A. nigrescens. They concluded that there is a tradeoff between shoot growth and defensive allo-chemicals for the species. The subsequent increase in palatability would lead to increased browsing, and create a positive feedback loop leading to the evolutionary development of Acacia compensatory growth as a necessary mechanism for the tree survival. Similarly, Fornara & du Toit (2007) found an inverse relationship between Acacia tolerance (regrowth abilities) and reproduction (seeds produced) in response to browsing, suggesting that above-ground growth came as a trade off with reproduction. Dharani et al. (2009) also observed increased browsing tolerance of A. xanthophloea after browsing. However, the study found that compensatory growth was limited to plant stems, while browsing still had a negative effect on tree height and canopy growth. They concluded that trees reallocated resources from leaves to stems and roots. Fornara (2008) similarly found that large herbivore pruning of A. nigrescens contributed to enhanced root size and subsequent re-sprouting abilities while negatively impacting maximum tree branch height, suggesting nutrient shunting from foliage to stems and roots.
It has also been hypothesized that the mechanism of shoot growth stimulation after browsing is not a reallocation of nutrients within the individual tree, but increased energy flow from the recycling of nutrients bound in leaf growth (Mopipi et al., 2009). However, Tanentzap & Coomes (2012) argued that herbivore grazing and browsing lead to decreased terrestrial carbon stock in general, and that the previous theory applies only in environments with sufficient resources to compensate for lost photosynthetic ability of consumed foliage. Other proposed mechanisms relating to resource allocation include: reduced intershoot competition due to woody browse loss, reduced leaf shading enabling enhanced photosynthesis capability, and reduced transpiration allowing more soil water retention (Fornara & du Toit, 2007; McNaughton, 1979). Numerous studies have resulted in various, sometimes contradictory, evidence and theories for compensatory plant growth mechanisms. Since no general conclusion has been reached, compensatory growth is likely species and location-dependent, influenced by browsing type and pressure, nutrient availability, inter- or intra-species competition, and perhaps other factors.
The mesic Acacia woodland in northern Tarangire National Park is composed largely of A. tortilis, one of the most prevalent and widespread Acacia species throughout Africa (Otieno et al. 2001), but has become less dense due to elephant utilization (Tanzania National Parks, 2012). Elephant herbivory can drive vegetation change in savanna ecosystems, and dense elephant populations can lead to over browsing and significant plant mortality (Van Aarde & Guldemond, 2008). Elephant browsing was found to be a more significant factor than rainfall behind changes in savanna vegetation (Hayward & Zawadska, 2010). With an average annual growth rate of about 7% from 1993-2006, there is no doubt that the Tarangire elephant population has intensified browsing pressure in the area, but it is unclear how the composition and woody plant-browser dynamics of the park savanna have changed (Foley, 2010). If this population growth continues, increased browsing pressure could threaten the survival and resilience of the park’s Acacia population.
Understanding the risk of A. tortilis decline and growth response to elephant browsing will help reduce tree mortality and prevent the loss of its ecosystem services. In this study, we inferred above-ground compensatory growth if greater growth occurred at moderate levels of branch removal by elephants. We hypothesized that A. tortilis trees under moderate levels of elephant browsing will have higher levels of annual shoot elongation because moderate levels of browsing pressure will stimulate more compensatory growth.
Tarangire National Park is located between S03°40’ and S03°35 and between E35°45’ and E37°0’ in the Arusha province of northern Tanzania and covers 2,850km2. As previously described by Ludwig et al. (2008), the park experiences a short rainy season from October to November, and a long rainy season from March to May, with the majority of the 650mm/year rainfall occurring between March and April. The park has several major habitat types including Acacia tortilis parkland, tall Acacia woodland dominated by A. xanthophloea and A. sieberiana, drainage line woodland, and deciduous savanna dominated by Combretum spp. and Commiphora spp. (Ludwig et al., 2008). This study focuses on three mesic sites located in the Lemiyon, Matete and Gursi areas of the northern part of the Park in order to focus on A. tortilis specifically.
At each of the three sites, we selected five to six trees with little (<30%), moderate (30-50%), and severe (>50%) elephant damage (47 trees total) from trees with multiple accessible unbrowsed twigs to measure. Tree measurements were taken for A. tortilis which fit strict criteria (<200m from roads, diameter at breast height (DBH) >10cm, and height >5m).
The percentage of canopy removed by elephant browsing was estimated qualitatively to the nearest 10%. Similar manual methods of estimating and quantifying browsing damage on tree canopy were described in Okula & Sis (1986), Riginos & Young (2007), Dharani et al. (2009) and Moncrieff (2011). It was not possible to judge the age of tree damage so that some years of measured growth could have occurred before any or some of the recorded damage occurred.
Elongation was measured using terminal bud scale scars on 3-4 twigs for each of the four previous years. We assigned twig measurements to the year in which growth began. To focus on growth response at the tree level, we averaged annual twig measurements within each tree.
Precipitation data to examine differences among years are unfortunately not available from this site, and reproduction occurred after the study time. All data was collected in the park in a short four-week period due to resource limitation.
Minitab ver. 17 (Minitab, 2014) was used to test for differences in growth among sites and years using a repeated measures ANOVA with year (fixed factor) crossed with site (random factor) and with trees nested within sites. Before analysis, a square root transformation was used to meet assumptions of normality and equality of variances. For each year linear and quadratic regressions from JMP ver. 11 (JMP®, 2007) were used to model relationships between annual shoot elongation and elephant damage. The best model fit was defined as that which had a larger R2 value. Due to the similarity of growth sites we combined them in further analyses and considered all trees independent.
Consistency of Growth Across Sites
Twig elongation ranged between 110-240mm across the three sites and four years, with the highest average growth recorded for site A in 2012 and the lowest for site B in 2013 and site A in 2010 (Figure 1).
The repeated measures ANOVA showed no differences in twig elongation between growth sites and years. It also measured the effects of an individual tree, and the interaction between site and year (Table 1). While growth year and the mesic site did not affect twig elongation individually (p-values > .05), the interaction between the two variables was significant (p = .01), as was the variation among trees within sites (p = .04). The repeated measures ANOVA explained 46% of the variation in annual growth measurements.
Effects of Browsing on Shoot Elongation
The relationships between elephant browsing damage and annual A. tortilis twig elongation had model p-values > .05 for all years (Figure 2). For 2013, 2011 and 2010 quadratic regression models were a better fit to the observed data, with higher R2 values (Figure 2). The general trend for these years showed that growth at moderate browsing levels (30-60% damage) was about 20mm greater than at low and high levels. For 2012, a linear regression line provided a better fit, showing a very slight (<5mm) decrease in shoot elongation as browsing pressure increases (Figure 2). Model R2 were low for all relationships, indicating substantial unexplained variation, but the quadratic models for 2013, 2011 & 2010 were higher than the linear model of 2012 growth (Figure 2).
The effects of herbivory on compensatory growth in woody plants have been heavily disputed. (Chira & Kinyamario, 2009; Dublin et al., 1990; Jachmann & Bell, 1985; Pellew, 1984; Smallie & O’Connor, 2000). It is clear that elephant herbivory can drive vegetation change in savanna ecosystems, and dense elephant populations can sometimes lead to over-browsing and significant plant mortality (Van Aarde & Guldemond, 2008).
Soil moisture and available nutrients are major confounding factors influencing Acacia growth (Mopipi et al., 2009). To reduce these variations, we measured trees only on mesic sites, i.e. sites with moderate and consistent soil moisture levels and nutrient availability in the context of the larger study area. Trees less than 200m from roads were excluded to limit the impact of dust from nearby vehicle traffic. We excluded immature trees (DBH < 10cm) since differences in plant developmental stages have been shown to influence Acacia browsing response (Bergstrom, 1992). Trees shorter than 5m were also excluded because they are more likely to be affected by the browsing of smaller mammals, which reduces Acacia growth due to the differences in browsing behavior of smaller browsers compared to elephants (Augustine, McNaughton & Samuel, 2004). Although Dharani et al. (2008) showed that rainfall plays a significant role in Acacia growth, levels of precipitation were not a major confounding factor in this study since sites were within about 20km of each other. Ben-Shahar (1996) also showed that fire and burn damage can significantly influence Acacia growth, but there was no recent history of fire in these areas (Tarangire, 2012), nor did we observe evidence of recent fires.
The three mesic sites used in this study were chosen to control for proximity to roads, tree maturity, and small mammal browsing; factors known to affect browsing response. The repeated measures ANOVA showed no significant difference between twig elongation across the three mesic sites (p = .52), which indicates that location does not explain the observed variations (Table 1) and growth across sites can be regarded as the approximately the same.
Annual variation in precipitation has been shown to be a significant driver of overall Acacia productivity (Mopipi et al., 2009; Otieno et al., 2001). However, in this study, tree measurements were taken for each of the last four years individually to control for variations in annual environmental conditions, such as drought, that might affect tree growth. There was no significant difference between twig elongation across years (p = .17) (Table 1). Thus, growth year did not explain the observed variations in twig elongation, and trees grew similar amounts across the four years. This is in line with Hayward & Zawadska (2010), who found that precipitation variables had less of an impact on above ground growth than elephant browsing.
Recent studies have found evidence that intense large herbivore browsing can sometimes result in compensatory above-ground plant growth (Dublin et al., 1990; Jachmann & Bell, 1985; Smallie & O’Connor, 2000), while at other times having a negative effect on woody vegetation growth (Chira & Kinyamario, 2009; Gulemond & Aarde, 2008; Pellew, 1984). In this study A. tortilis shoot elongation varied little across the very broad continuum of browsing pressure (Figure 2), and browsing pressure explained little of variations in growth (R2 = .07). Thus we conclude that, in terms of aboveground shoot elongation, the Tarangire National Park population tolerates elephant browsing remarkably well. This may help explain the dominance of A. tortilis in the park and other East African savanna ecosystems.
Our data show equivocal results on whether moderate levels of elephant browsing stimulate compensatory growth. Insignificant trends in 2013, 2011 and 2010 (Figure 2) suggest support for the hypothesis and the conclusions of Du Toit et al. (1990), Chira & Kinyamario (2009), and Stuart-Hill & Tainton (1989) that compensatory growth occurs. The slight, insignificant decrease in 2012 shoot elongation as browsing increases (Figure 2) suggests support for the conclusions of Dharani et al. (2009) and Fornara (2008) that compensatory growth does not occur and that all levels of elephant browsing have a negative effect on above-ground growth. However, in contrast to these studies cited above, our results for four years in northern Tarangire National Park indicate a striking robustness to elephant browsing in A. tortilis, since even trees with heavy browsing damage had approximately the same shoot growth as trees with light or no browsing damage.
It is important to note that a large proportion of above ground growth variation among trees within sites remain unexplained by elephant browsing damage (Table 1). Substantial genetic differences, small-scale variation in soil nutrients and other soil properties, differences in above- and below-ground competition, time of elephant damage, and other environmental factors likely account for differences in shoot growth among replicate trees. This large variation will tend to mask patterns in growth with browse damage. Also, since important influences on growth such as fire, water, and nutrient availability will vary at other sites, resistance to browsing at other sites could show different patterns.
The variation in Acacia growth response to herbivory studied over the last 20 years, as well as within this paper, means that a unified understanding of Acacia compensatory growth remains elusive. Future studies might attempt to measure multiple responses, e.g., shoot elongation, the radial increment of trunks, reproduction, concentration of secondary compounds, and below-ground growth in order to better understand this phenomenon. Since responses likely vary with water availability, studies across a water availability gradient and across years of varying precipitation might prove fruitful. Additionally, genetic variation within A. tortilis, seasonal timing of elephant browsing, previous elephant browsing, effects of other herbivores, and time since the last fire may also influence responses to elephant browsing. Future studies should also consider tree mortality and seedling establishment since growth measurements do not take mortality into account.
Based on shoot elongation data from four years, A. tortilis in northern Tarangire National Park appears to be tolerating browsing of the large elephant population well, with some insignificant suggestion of modest compensatory growth at moderate browsing levels. In one particular year, a slight decline in shoot growth in trees with heavier browsing is observed. Such an anomalous reverse in growth trends in this year could be evidence of herbivory tipping point, rather than the linear relationship suggested in other years, after which browsing is harming not stimulating above ground growth for A. tortilis. Since this tree species comprises such a large proportion of tree cover in the Park (Tanzania National Parks, 2012) and because the recent trend in elephant numbers is upward (Foley, 2010), we suggest like Gandiwa et al. (2011) in Zimbabwe, that Park managers should monitor growth responses and mortality of A. tortilis to ensure the vitality of this keystone species.
Thanks to my fellow students, Heather Hagerling, Nikki Hernandez, Mark Parlier, and Maxine Quinney, who suffered many an Acacia wound during the blistering late mornings of data collection. And thanks to our wonderful driver Walter Humphreys for navigating us daily throughout Tarangire National Park.
Augustine, J., McNaughton, D., & Samuel, J. (2004). Regulation of shrub dynamics by native browsing ungulates on east African rangeland. Journal of Applied Ecology, 41(1), 45-58.
Ben-Shahar, R. (1996). Woodland dynamics under the influence of elephant and fire in northern Botswana. Vegetation, 123(2), 153–163.
Bergström, R. (1992). Browse characteristics and impact of browsing on trees and shrubs in African savannas. Journal of Vegetation Science, 3(3), 315-324.
Chira, R. M., & Kinyamario, J. I. (2009). Growth response of woody species to elephant foraging in Mwea National Reserve, Kenya. African Journal of Ecology, 47(4), 598-605.
Chung, K. Y., Wong, T. Y., Wei, C. I., Huang, Y. W. & Lin, Y. (1998). Tannins and human health: a review. Critical Review of Food Science Nutrition, 38(6), 421-64.
Dharani, N., Kinyamario, J. I., Wagacha, P. W. & Rodrigues, A. J. (2009). Browsing impact of large herbivores on Acacia xanthophloea in Lake Nakuru National Park, Kenya. African Journal of Ecology, 47(2), 184-191.
Dublin, H. T., Sinclair, A. R. E., & McGlade, J. (1990). Elephant and fire as causes of multiple stable states in the Serengeti-Mara woodlands. Journal of Animal Ecology, 59, 1147–1164.
Du Toit, J. T., Bryant, J. P, & Frisby, K. (1990). Regrowth and palatability of Acacia shoots following pruning by African savanna browsers. Ecology, 71(1), 149-153.
Foley, C. A. H., & Faust, L. J. (2010). Rapid population growth in an elephant Loxodonta africana population recovering from poaching in Tarangire National Park, Tanzania. Oryx, 44(2), 205-212.
Fornara, D.A., & du Toit, J.T. (2007). Browsing lawns responses of Acacia nigrescens to ungulate browsing in an African savanna. Ecology, 88(1), 200.
Fornara, D.A., & du Toit, J.T. (2008). Responses of woody saplings exposed to chronic mammalian herbivory in an African savanna. Ecoscience, 15(1), 129-135.
Gandiwa, E., Magwati, T., Zisadza, P., Chinuwo, T., & Tafangenyasha, C. (2011). The impact of African elephants on Acacia tortilis woodland in northern Gonarezhou National Park, Zimbabwe. Journal of Arid Environments, 75(9), 809-814.
Guldemond, R., & Van Aarde, R. (2008). A meta-analysis of the impact of African elephants on savanna vegetation. Wild Journal of Wildlife Management, 72(4), 892-899.
Hayward, M. W., & Zawadzka, B. (2010). Increasing elephant Loxodonta africana density is a more important driver of change in vegetation condition than rainfall. Acta Theriologica, 55, 289-299.
Jachmann, H., & Bell, R. H. V. (1985). Utilization by elephant of the Brachystegia woodland of the Kasungu National Park, Malawi. African Journal of Ecology, 23, 245–258.
JMP®, Version 11. SAS Institute Inc. 1989-2007. [Computer software] Cary, NC.
Ludwig, F., De Kroon, H. & Prins, H. H. T.(2008). Impacts of savanna trees on forage quality for a large African herbivore. Oecologia, 155(3), 487-496.
McNaughton, S.J. (1979). Three grassland-herbivore dynamics. In: Serengeti, Dynamics of and Ecosystem. The University of Chicago Press, Chicago pp. 46-80.
Minitab. 17 Statistical Software. 2010. [Computer software]. State College, PA: Minitab, Inc. www.minitab.com.
Moncrieff, G. R., Chamaille-Jammes, S., Higgins, S. I., O’Hara, R. B., & Bond, W. J. (2011). Tree allometries reflect a lifetime of herbivory in an African savanna. Ecology, 92(12), 2310-2315.
Mopipi, K., Trollope, S. W. W & Scogings, P. F. (2009). Effects of moisture, nitrogen, grass competition and simulated browsing on the survival and growth of Acacia karroo seedlings. African Journal of Ecology, 47(4), 680-687.
Otieno, D. O., Kinyamario, J. I., & Omenda, T. O. (2001). Growth features of Acacia tortilis and Acacia xanthophloea seedlings and their response to cyclic soil drought stress. African Journal of Ecology, 39(2), 126-132.
Owen-Smith, N. (1982). Factors influencing the consumption of plant products by large herbivores. In: B. J. Huntley & B. H. Walker (Eds.), Ecology of tropical savannas (359-404). New York: Springer Berlin Heidelberg.
Pellew, R. A. (1984). The feeding ecology of a selective browser, the giraffe (Giraffa camelopardalis tippelskirchi). Journal of Zoology, 202, 57-81.
Riginos, C., & Young, T. P. (2007). Positive and negative effects of grass, cattle, and wold herbivores on Acacia saplings in an East African savanna. Oecologia, 153(4), 985-995.
Scogings, P.F., Johansson, T., Hjalten, J., & Kruger, J. (2012). Responses of woody vegetation to exclusion of large herbivores in semi-arid savannas. Australia Ecology, 37(1), 56-66.
Smallie, J. J., & O’connor, T. G. (2000). Elephant utilization of Colophospermum mopane: possible benefits of hedging. African Journal of Ecology, 38, 352–359.
Stuart-Hill, G. C., & Tainton, N. M. (1989). The competitive interaction between Acacia karroo and the herbaceous layer and how this is influenced by defoliation. Journal of Ecology, 26, 285-298.
Tanentzap, A. J., & Coomes D. A. (2012). Carbon storage in terrestrial ecosystems: do browsing and grazing herbivores matter. Biological Reviews of the Cambridge Philosophical Society, 87(1), 72-94.
Tanzania National Parks. (2012). Tarangire National Park. Retrieved September 10, 2014. http://www.tanzaniaparks.com.
Neurogenesis Unchanged by MTHFR Deficiency in Three-Week-Old Mice
The primary pathway for removing homocysteine, a potentially neurotoxic molecule, from circulation is via 5-methyltetrahydrofolate. This molecule is converted from folate via an enzyme known as methylenetetrahydrofolate reductase (MTHFR). Polymorphisms in the MTHFR gene have been linked to various pathologies (e.g. neurological disease) and animal models have been developed to study the in vivo effects of the deficiency. These models have revealed increased levels of apoptosis in the cerebellum and hippocampus and potential modulation of neurogenesis, which may contribute to the pathologies viewed. The aim of this study was to evaluate neurogenesis during late development. Brain tissue from 3-week-old male mice with different MTHFR genotypic was collected and stained for phosphohistone H3 (PH3) and 4′,6-diamidino-2-phenylindole (DAPI). Cell counts were performed in order to quantify PH3 positive cells in the hippocampus (dentate gyrus), cerebellum and cortex. There were no significant differences between genotype groups in all areas analysis. This result suggests that MTHFR status does not affect neurogenesis levels at this particular stage of development. Future research should consider the potential for confounding methylation pathways, such as the betaine-mediated mechanism, and investigating the effect on earlier developmental periods.
Folate metabolism is a key mechanism in the brain that allows the downstream alteration of a variety of proteins and plays a role in the synthesis of nucleotides (Kamen, 1997). These folate-mediated effects are necessary for the production of new neural cells and thus are essential to the overall health of the brain during development, adulthood and aging (McGarel, Pentieva, Strain & McNulty, 2015). One mechanism through which folates affect protein function is through the initial methylation of homocysteine to methionine. Homocysteine is a cytotoxic molecule when in high levels that produces a variety of negative effects, including endoplasmic reticulum stress, excitatory amino acid receptor over-activation, kinase hyperactivity and DNA damage (Ho, Ortiz, Rogers & Shea, 2002). These effects have been associated with many clinical pathologies in humans, being indicated as a contributing factor to cognitive impairment (Almeida et al., 2005), neural tube defects (Felkner, Suarez, Canfield, Brender & Sun, 2009), brain atrophy (den Heijer et al., 2003), stroke (Hankey & Eikelboom, 2001) and cardiovascular disease (Frosst et al., 1995; Wierzbicki, 2007). Thus, the ability for folates to methylate homocysteine to its nontoxic derivative methionine plays a large role in protecting against neurotoxicity. However, the links between these illnesses and homocysteine are still not fully understood, with many downstream biochemical pathways still needing to be discovered. Research into increased homocysteine levels and altered folate metabolism confirms a variety of cytotoxic effects in Caenorhabditis elegans (Ortbauer et al., 2016), Drosophila melanogaster (Blatch, Stabler & Harrison, 2015) and Sacchaomyces cerevisiae (Kumar et al., 2011). Thus, the need to utilize mammal models of increased levels of homocysteine is necessary to produce potential theories of illness.
One such model looks at the knockout of a particular enzyme in the homocysteine cycle, known as methylenetetrahydrofolate reductase (MTHFR). Folate itself cannot directly methylate homocysteine, thus it must first be converted from the form it is ingested to its primary circulating form, 5-methyltetrahydrofolate (5-methyl-THF). The key enzyme to this process is the aforementioned MTHFR, which catalyzes the production of 5-methyl-THF from a less abundant form 5,10-methyl-THF, which then methylates homocysteine. Thus, this enzyme is essential to both the metabolism of folate and homocysteine. This cycle is highlighted in Figure 1.
The occurrence of MTHFR deficiency is not uncommon in humans, with two common mutations producing reduced or lack of function. One of these deficiency-causing mutations is homozygous in approximately 18 percent of humans (Zittan et al., 2007). As many as 34 mutations in this gene, however, have been identified in individuals with homocystinuria, a genetic condition resulting in elevated levels of homocysteine that is associated with neurological and vascular problems (Leclerc, Sibani & Rozen, 2000). Studies into MTHFR in animal models have confirmed its role in preventing homocysteine-mediated neurotoxicity and associated pathologies (Chen et al., 2001; Chen, Schwahn, Wu, He & Rozen, 2005; Jadavji et al., 2012). Some of these pathologies include changes to motor control, mood and cognitive function, all of which have some connection to the altered morphology of the hippocampus and cerebellum specifically (Jadavji et al., 2012). The combination of the modelled results and the clinical implications of this gene thus offer significant intrigue into the role of folates in protecting against neurotoxicity.
Many of the associated pathologies, including cognitive impairment, brain atrophy and neural tube defects would suggest that the production of new and fully-functioning neurons is altered in response to high levels of homocysteine (Boot et al., 2003; Black, 2008). Neurogenesis, or the production of new neurons, occurs most prominently prenatally across species (Clancy, Darlington & Finlay, 2001, and is tightly linked to overall brain and cognitive function (Siwak-Tapp et al., 2007). Neurogenesis is also inherently linked with apoptosis, or cell death, as during development, neural pathways are trimmed via apoptosis. Additionally, adult neural death must be replaced by new cells in order to maintain the function of that brain region. Adult neurogenesis almost exclusively occurs in the dentate gyrus of the hippocampus and the olfactory bulb (Ming & Song, 2011), although there is research suggesting adult cell proliferation in the cerebellum (Ponti, Peretto & Bonfanti, 2008) and the cortex (Gould, Reeves, Graziano & Gross, 1999). As a result of this, it is of great interest to study the effect of increased homocysteine levels on the rates of neurogenesis in these regions to determine if blunted neurogenesis occurs as a result of increased homocysteine levels and if this mediates any of the observed pathologies.
Previous research has already begun to look in some of these regions with regards to neurogenesis and homocysteine levels. Research conducted by Jadavji et al. in 2012, used adult MTHFR knockout mouse models to test the association between neurogenesis and hyperhomocysteinemia-mediated pathologies. The results showed that homozygous knockouts for MTHFR had severely decreased cognitive functioning in object recognition tests, while apoptosis levels in the dentate gyrus were also elevated in the same genotype (Jadavji et al., 2012). This would suggest that without cell replacement in the dentate, cognitive function is greatly reduced, as a result of elevated homocysteine.
Similarly, research into other areas of the brain associated with neurogenesis appears to produce comparable results. In 2005, Chen, Schwahn, Wu, He & Rozen analyzed cell death levels in the cerebellum of MTHFR knockout mice, and found that homozygous knockouts also showed increased apoptosis of cerebellar neurons in both the intragranular and extragranular layers during the first two weeks of postnatal development (Chen, Schwahn, Wu, He & Rozen, 2005). As previously mentioned, pruning of neural pathways occurs during this time frame and alterations to cell death would result in changes to cerebellar patterning and could potentially contribute to neuropathology.
Thus, in order to determine if the pathologies viewed as a result of elevated homocysteine levels are partially mediated by alterations to neurogenesis levels, this study aimed to analyze the levels of neurogenesis in three of the regions associated with adult cell proliferation; the dentate gyrus, the cerebellum and the cortex. Analysis of neurogenesis in these regions in MTHFR knockout mice was used to determine if and propose how neurogenesis reductions may mediate homocysteine-related neuropathology.
Materials and Methods
All experiments were approved by Montreal Children’s Hospital Animal Care Committee in accordance to the Canadian Council on Animal Care guidelines. The mice used were from the C57B1/6 genetic background. The breeding of each status of MTHFR has been described previously (Jadavji et al., 2012). Mice were fed standard mouse chow (Envigo) and water ad libitum. They were sacrificed at 3 weeks from male mice.
Slides of sagittal-oriented brain tissue from 3-week-old animals were deparaffinized in fresh xylene twice for 5 minutes each. They were then rehydrated in absolute alcohol for 5 minutes, followed by reduced concentration alcohol gradually at 95, 80 and 70% alcohol for 3 minutes each. The slides were then rinsed using PBS twice for 5 minutes each before Antigen Unmasking Solution was used for 20 minutes at 95°C to retrieve antigens on the slides. 10 μg/mL Proteinase K in Tris/EDTA solution was used to perform cell permeabilization for 10 minutes at room temperature before 5% goat serum in PBS blocked non-specific binding sites in the samples for 30 minutes. This allows for the targeted antibody-antigen interaction to occur. These slices were then stained using 1:100 concentration rabbit anti-phospho histone H3 mitosis marker (PH3; Cell Signalling) as the primary antibody to identify cells that were proliferating overnight at 4°C. Alexa Fluor (Cell Signalling) antibody at a dilution of 1:200 was then used for 40 min at room temperature. 4′,6-diamidino-2-phenylindole (DAPI) was used as a general stain for the visualization of all cells and section were coverslipped with Vectashield Hardset Mounting Media in order to preserve the staining and were stored at 4°C.
The stained slides were imaged using Infinity Analyze software, where an image under fluorescence for PH3 and DAPI were captured for each of the regions of interest; the dentate gyrus of the hippocampus, the cerebellum and the cortex. The images for each fluor respectively, were combined using Image J software (NIH) and the final combined images were manually quantified for co-localization of the two immunofluorescent tags. As the cerebellum and cortex were larger and unable to be imaged all in one photo, three separate lobes were imaged for each cerebellar sample, and three sections from each cortex (anterior, medial and posterior) and the counts were averaged from every sample for each specimen into one mean per mouse.
Data was analyzed using SPSS and GraphPad software. One-way analysis of variance (ANOVA) was conducted on the three levels of MTHFR in each of the regions of interest, with p < .05 set as the significance threshold.
Previous research into MTHFR knockout mice would suggest that neurogenesis would likely be reduced in the mice with lost function of the enzyme. Previous research by Chen et al. (2001) confirmed the knockout status of MTHFR of this strain of mice, and can be referred to for Western blot results. As the goal of the research was to determine if MTHFR status had an impact on neurogenesis in the 3 week old mice, the co-localization of DAPI and PH3 staining was compared to determine if there was any difference between genotype groups.
First, in the dentate gyrus (Figure 2), the three statuses of MTHFR produced mean co-localized cell counts of 4.10 (n = 4, SD = 1.49) for the wild-type, 4.33 (n = 3, SD = 4.04) for the heterozygotes and 4.50 (n = 4, SD = 1.78) for the full knockouts. There was no difference between genotype groups (Figure 2(C), F(2,8) = 0.03, p = 0.974).
Next, the same measures were taken for the cerebellum (Figure 3). This procedure produced a mean co-localized cell count for the wild-type mice of 9.30, (n = 5,SD = 4.51), 8.55 (n = 6, SD = 4.49) for the MTHFR heterozygotes and 5.64 (n = 5, SD= 1.52) for the complete knockouts. The analysis of variance showed no difference between groups (Figure 3(C), F(2,13) = 1.29, p = 0.309).
Finally, the levels of co-localized cells were recorded for the cortex of the mice, as imaged in Figure 4. The mean co-localized cell count for the wild-type mice was 8.06 (n = 4, SD = 1.93), 6.19 (n = 4, SD= 3.71) for the heterozygous knockouts and 7.42 (n = 5, SD= 4.62) for the complete knockouts. There was no difference between groups (Figure 4(C), F(2,10) = 0.26, p = .773).
This research produced no statistically significant changes in the co-localization counts of PH3 and DAPI in any of the three main regions of interest, the dentate gyrus, cerebellum and cortex in response to MTHFR status. This would suggest that the relative level of MTHFR does not have a significant impact on neurogenesis at this point of development in male mice, and is not an ideal candidate for explaining altered neurogenesis and neuropathology seen in response to increased homocysteine. Despite this unexpected result, there are numerous reasons why this study may have failed to reject the null hypotheses.
The first reason may simply have to do with a small sample size. The levels of MTHFR had a range of n-values of 3 to 6, thus even after procedures to reduce variance by calculating a mean for each mouse before the genotype mean, there was still a high level or variance for most statuses. This high variance greatly contributes to the failure to reject the null hypothesis of the one-way ANOVAs, especially for the cerebellum which had the highest F-ratio. If the sample size was increased for each MTHFR level, it is possible that the resulting reduced variance could lead to a statistically significant effect. Procedural issues were not the only potential source of confound producing this result. As mentioned in the introduction, mice are generally into their “adolescent” period by three weeks, and thus have completed the majority of their developmental neurogenesis (Clancy, Darlington & Finlay, 2001). As no significant change in neurogenesis was viewed at this stage of development, it could perhaps be beneficial to look at younger mice that are still going through developmental neurogenesis. It is possible that earlier neurogenesis may be more sensitive to MTHFR status, suggested by the many studies showing developmental neurological defects, especially in the neural tube (Felkner, Suarez, Canfield, Brender & Sun, 2009). Future research should consider looking at mice earlier in neurological development for alterations to neurogenesis as a result of MTHFR status.
In addition to changing the age of the mice used, another possible reason for viewing this result could be an alternative method of methylating homocysteine. Betaine is well established as a supplementary methyl donor for the homocysteine to methionine conversion known to reduce homocysteine concentrations (Olthof & Verhoef, 2005). Derived from choline, betaine is regularly produced in the body through the catalytic activity of two enzymes, choline dehydrogenase (CHDH) and betaine aldehyde dehydrogenase (BADH). Betaine is then converted into its methyl-donating form via betaine homocysteine methyltransferase (BHMT). Thus, these results viewed may show that at this key stage of development there is an alternative method of reducing homocysteine and neuroprotection in response to MTHFR deletion. This method could involve the upregulation of any of the CHDH, BADH or BHMT enzymes, resulting in increased betaine-mediated methylation of homocysteine, preventing the neurotoxic effects. This theory is not completely hypothetical, as previous research mentioned by Chen, Schwahn, Wu, He & Rozen, in 2005, showed that betaine augmentation reduced the effect of MTHFR deletion in cerebellar development, the region closest in this study to having statistically significant effects (Chen, Schwahn, Wu, He & Rozen, 2005). Although the supplementation of that study came from an external source, it could be valuable to assess the relative transcript levels of betaine-related enzymes in future research of folate metabolism and the homocysteine cycle.
It may also be of value in future research to sample levels of methyl marks on methylated acceptors in these regions of interest. As methionine can be metabolized to S-adenosylmethionine, protein methylation would be expected to be viewed in mice with effective folate metabolism. This could potentially provide a valuable measure of whether homocysteine was being converted to methionine at regular levels or not, perhaps confirming confounding effects from betaine.
MTHFR has continuously been indicated in numerous pathologies and has been modelled effectively as a knockout paradigm. Age of the mice and confounding enzymatic pathways should both be considered in future research into the models. Despite no statistically significant results here, this study still contributes essential information into the mechanisms of neurogenesis and folate metabolism.
Almeida, O. P., Flicker, L., Lautenschlager, N. T., Leedman, P., Vasikaran, S. & van Bockxmeer, F. M. (2005). Contribution of the MTHFR gene to the causal pathway for depression, anxiety and cognitive impairment later in life. Neurobiology of Aging, 26(2), 251-257. doi:10.1016/j.neurobiolaging.2004.03.007.
Black, M. M. (2008). Effects of vitamin B12 and folate deficiency on brain development in children. Food and Nutrition Bulletin, 29(2): S126-S131.
Blatch, S. A., Stabler, S. P. & Harrison, J. F. (2015). The effects of folate intake on DNA and single-carbon pathway metabolism in the fruit fly Drosophila melanogaster compared to mammals. Comparitive Biochemistry and Physiology: Part B Biochemistry and Molecular Biology, 189, 34-39. doi: 10.1016/j.cbpb.2015.07.007.
Boot, M. J., Steegers-Theunissen, R. P., Poelmann, R. E., Van Iperen, L., Lindemans, J. & Gittenberger-de Groot, A. C. (2003). Folic acid and homocysteine affect neural crest and neuroepithelial cell outgrowth and differentiation in vitro. Developental Dynamics, 227(2): 301-308. doi: 10.1002/dvdy.10303
Chen, Z., Karaplis, A. C., Ackerman, S. L., Pogribny, I. P., Melnyk, S., Lussier-Cacan, S., Chen, M. F., Pai, A., John, S. W., Smith, R. S., Bottiglieri, T., Bagley, P., Selhub, J., Rudnicki, M. A., James, S. J. & Rozen, R. (2001). Mice deficient in methylenetetrahydrofolate reductase exhibit hyperhomocysteinemia and decreased methylation capacity, with neuropathology and aortic lipid deposition. Human Molecular Genetics, 10(5), 433-443.
Chen, Z., Schwahn, B. C., Wu, Q., He, X. & Rozen, R. (2005). Postnatal cerebellar defects in mice deficient in methylenetetrahydrofolate reductase. International Journal of Developmental Neuroscience, 23, 465-474. doi:10.1016/j.ijdevneu.2005.05.007.
Clancy, B., Darlington, B. & Finlay, B. L. (2001). Translating developmental time across mammalian species. Neuroscience, 105, 7-17.
den Heijer, T., Vermeer, S. E., Clarke, R., Oudkerk, M., Koudstaal, P. J., Hofman, A. & Breteler, M. M. (2003). Homocysteine and brain atrophy on MRI of non-demented elderly. Brain, 126, 170-175.
Felkner, M., Suarez, L., Canfield, M. A., Brender, J. D. & Sun, Q. (2009). Maternal serum homocysteine and risk for neural tube defects in a Texas-Mexico border population. Birth Defects Research, 85(6), 574-581. doi: 10.1002/bdra.20545.
Frosst, P., Blom, H. J., Milos, R., Goyette, P., Sheppard, C. A., Matthews, R. G., Boers, G. J., den Heijer, M., Kluijtmans, L. A., van den Heuvel, L. P. & Rozen, R. (1995). A candidate genetic risk factor for vascular disease: A common mutation in methylenetetrahydrofolate reductase, Nature Genetics, 10, 111-113.
Gould, E., Reeves, A. J., Graziano, M., S. & Gross, C. G. (1999). Neurogenesis in the neocortex of adult primates. Science, 15(286), 548-552. doi: 10.1126/science.286.5439.548.
Hankey, G. J. & Eikelboom, J. W. (2001). Homocysteine and stroke. Current Opinion in Neurology, 14, 95-102.
Ho, P. I., Ortiz, D., Rogers, E. & Shea, T. B. (2002). Multiple aspects of homocysteine neurotoxicity: Glutamate excitotoxicity, kinase hyperactivation and DNA damage. Journal of Neuroscience Research, 70(5), 694-702. doi: 10.1002/jnr.10416.
Jadavji, N. M., Deng, L., Leclerc, D., Malysheva, O., Bedell, B. J., Caudill, M. A. & Rozen, R. (2012). Severe methylenetetrahydrofolate reductase deficiency in mice results in behavioral anomalies with morphological and biochemical changes in hippocampus. Molecular Genetics and Metabolism, 106, 149-159. doi:10.1016/j.ymgme.2012.03.020.
Kamen, B. (1997). Folate and antifolate pharmacology. Seminars in Oncology, 24(5 Suppl 18), 30-39. PMID: 9420019.
Kumar, A., John, L., Maity, S., Manchanda, M., Sharma, A., Saini, N., Chakraborty, K. & Sengupta, S. (2011). Converging evidence of mitochondrial dysfunction in a yeast model of homocysteine metabolism imbalance. The Jounral of Biological Chemistry, 286(24), 21779-21795. doi: 10.1074/jbc.M111.228072.
Leclerc D, Sibani S, Rozen R. Molecular Biology of Methylenetetrahydrofolate Reductase (MTHFR) and Overview of Mutations/Polymorphisms. Austin (TX): Landes Bioscience; 2000.
McGarel, C., Pentieva, K., Strain, J. J. & McNulty, H. (2015). Emerging roles for folate and related B-vitamins in brain health across the lifecycle. The Proceedings of the Nutrition Society, 74(1), 46-55.
Olthof, M. R. & Verhoef, P. (2005). Effects of betaine intake on plasma homocysteine concentrations and consequences for health. Current Drug Metabolism, 6, 15-22.
Ortbauer, M., Ripper, D., Fuhrmann, T., Lassi, M., Auernigg-Haselmaier, S., Sitegler, C. & Konig, J. (2016). Folate deficiency and voer-supplemntation causes impaired folate metabolism: Regulation and adaption mechanisms in Caenorhabditis elegans. Molecular Nutrition & Food Research, 60(4), 949-956. doi: 10.1002/mnfr.201500819.
Ponti, G., Peretto, P. & Bonfanti, L. (2008). Genesis of neuronal and glial progenitors in the cerebellar cortex of peripuberal and adult rabbits. PLoS One, 3(6), e2366. doi:10.1371/journal.pone.0002366.
Siwak-Tapp, C. T., Head, E., Muggenburg, B. A., Milgram, N. W. & Cotman, C. W. (2007). Neurogenesis decreases with age in the canine hippocampus and correlates with cognitive function. Neurobiology of Learning and Memory, 88(2), 249-259. doi:10.1016/j.nlm.2007.05.001.
Wierzbicki, A. S. (2007). Homocysteine and cardiovascular disease: A review of the evidence. Diabetes and Vascular Disease Research, 4, 143-150.
Zittan, E., Preis, M., Asmir, I., Cassel, A., Lindenfeld, N., Alroy, S., Halon, D. A., Lewis, B. S., Shiran, A., Schliamser, J. E. & Flugelman, M. Y. (2007). High frequency of vitamin B12 deficiency in asymptomatic individuals homozygous to MTHFR C677T mutation associated with endothelial dysfunction and homocysteinemia. American Journal of Physiology: Heart and Circulatory Physiology, 293(1), 860-865. doi: 10.1152/ajpheart.01189.2006.
Access Full PDF: Neurogenesis Unchanged by MTHFR Deficiency in Three-Week-Old Mice
Structural Dynamics of Amyloid-β Aggregation in Alzheimer’s Disease: Computational and Experimental Approaches
The nucleation of amyloid-β (Aβ) oligomers, and the fibril formation that follows represents an important pathologic mechanism for Alzheimer’s disease (AD). This has motivated the search for therapeutics that specifically target Aβ, which holds promise to be a cure for AD. However, conventional biophysical approaches like X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy fall short of capturing this highly dynamic process. The aggregation of amyloid fibrils has been unravelled by a mix of novel approaches. For example, computational methods like molecular dynamics (MD) simulations provide atomistic predictions of structural dynamics of protein fibrils. On the other hand, various spectroscopy and spectrometry approaches allow unprecedented resolution and details to be experimentally observed and complement predictions offered by computational investigations. This review surveys these approaches in the context of studying Aβ peptide aggregation. We shall discuss how these methods help us understand the mechanistic aspect of the aggregative process. We emphasise that computational and experimental approaches must go together to obtain a comprehensive view on the structural dynamics of critical protein players in health and diseases.
Aβ aggregates as causative agent for Alzheimer’s disease
AD is the most common form of dementia, featuring irreversible memory loss and deterioration in linguistic, cognitive, and learning capacities. Insoluble Aβ plaques are a hallmark of AD, with elevated levels of the 42-residue Aβ42 in the brain compared to a more abundant Aβ40 in healthy individuals (Näslund et al., 1994). Researchers have long followed the path of “amyloid hypothesis” in finding AD’s pathologic mechanism. This hypothesis holds that aggregation and deposition of Aβ peptides are the earliest initiating factors of AD, leading to the formation of paired helical filaments of tau aggregates and eventually neuronal cell death (Karran, Mercken & De Strooper, 2011).
Both soluble Aβ oligomers and insoluble Aβ fibrils are neurotoxic. The former is a better correlate of cognitive dysfunction; Aβ oligomers block long-term potentiation in hippocampal neurons (Tomic, Pensalfini, Head & Glabe, 2009). In Aβ fibrils, β-strands of individual Aβ peptides aggregate to form cross-β structures with inter-strand hydrogen bonds (Figure 1A). This provides surfaces onto which Aβ peptides are favourably “docked-and-locked” (Karran et al., 2011). Therefore, it is intuitive for therapeutics to attempt preventing Aβ aggregates, in both oligomer and fibril states.
To design these therapeutic interventions, we need to thoroughly understand the nucleating step, stability, and reversibility of these aggregates. While techniques like NMR contributed greatly to studying the structure of Aβ fibrils, Aβ oligomer structure is less clear due to its heterogeneity, instability, and variability across experimental conditions (Cerasoli, Ryadnov & Austen, 2015) (Figure 1B).
The dynamic process of how Aβ is transiently misfolded, aggregates into oligomers, and subsequently forms protofibrils and fibrils is still elusive, due to the limitations of conventional biophysical techniques. In this review, we compare newer spectrometry and spectroscopy techniques with conventional biophysical approaches used in studying Aβ aggregation. In addition, computational strategies like MD simulation have generated insights into the dynamic properties of Aβ aggregates. We discuss their principles and usages and outline how a combination of experimental and computational methods contributes profound insights into Aβ aggregation. Both approaches should go hand in hand in order to understand, not only Aβ but critical peptides and proteins in human diseases.
Conventional Biophysical Techniques Only Give Snapshots of Aggregation
X-ray crystallography, NMR, and cryo-electron microscopy are established experimental methods to study protein structures and interaction at atomic resolution by capturing a particular instant of conformation. While they suffice to infer spatial arrangement and interactions, the study of time-dependent processes such as Aβ aggregation, which goes on for days, takes more than a few structural snapshots.
In the case of Aβ aggregation, solving a definitive structure is already a challenge. In its fibril state, Aβ becomes insoluble, and it is impossible to perform solution NMR or to grow crystals for typical X-ray diffraction analysis. It can at most be subjected to X-ray fibre diffraction at moderate resolution or solid-state NMR (ssNMR). Nevertheless, several sizes of Aβ oligomers can be soluble in sodium dodecyl sulphate (SDS), and such a sample had been subjected to solution NMR (Yu et al., 2009). These forms exhibit strong toxicity and neuropathogenicity (Ahmed et al., 2010). However, such solubilisation could shift the stabilization environment, possibly by adding detergent that does not properly reflect physiological situations. Moreover, the immense scale of Aβ aggregation complicates such inference (Robustelli, Stafford & Palmer, 2012).
Electron microscopy (EM) is one of the earliest approaches to studying amyloid fibrils (Shirahama & Cohen, 1967). Although EM has yielded insights into the basic structural features and might also capture several states of the aggregates, it is not meant to provide time-dependent data for analysis. In Aβ aggregation where the process is of utmost interest, EM, like NMR, is inadequate in understanding the dynamic details of Aβ fibril formation.
As shown in Figure 1B, amyloid fibrils can take on different morphology depending on the source. For example, in vitro-prepared fibrils can occur in varying layers. Although only one of the dozen conformations is displayed here, the remaining frames are merely a transient oscillation of flexible regions that highlight no key interaction events. Conventional methods thus leave us ignorant of processes that happened before and during aggregation, fragmentation, and secondary nucleation. As we are gradually convinced that Aβ aggregates can exist in any form, solving structures of each polymorph becomes less relevant and the spotlight shines on understanding and predicting how they are formed with molecular and biophysical details.
Molecular Dynamics Simulation can Computationally Model Aβ Aggregation
To overcome limitations of conventional biophysical techniques, computational MD simulations, provide a “time-lapse” dimension to the Aβ structure. Here we put the protein(s) of interest into a system surrounded with solvents and compute its/their movement by using Newton’s laws and sets of pre-defined parameters known as “force-fields”. These force-fields dictate the energy functions that model potential energies within bonded and non-bonded entities (Jernigan & Bahar, 1996). An all-atom conventional MD simulation of Aβ shows a trajectory from which inference can be drawn with regards to its stability and/or plasticity. Trajectories of wild-type and mutant protein can also be compared to identify critical residues in structure and function (Figure 2A). In the context of Aβ, a special variant of MD simulation known as steered MD (SMD) is particularly relevant. This is illustrated in Figure 2B, where the amyloid aggregate is taken as a control from which the fibril is pulled away. An “umbrella sampling” method samples conformations from varying distances of the pull to probe interaction across time (Lemkul & Bevan, 2010). This provides a detailed atomistic account of the reverse of aggregation, which is valuable in deciphering aggregation itself. Such framework can be enhanced to answer more specific biological questions. For example, robust free energy calculations performed on the SMD profile can quantify binding and aggregation of peptides (Perez, Morrone, Simmerling & Dill, 2016) (Figure 2B).
An early paper manipulated simulations to characterise the conformational heterogeneity of a short stretch Aβ16-22 mediated by electrostatic and hydrophobic interactions (Klimov & Thirumalai, 2003). Lemkul and Bevan (2010) revealed a critical role for the salt bridge between Asp23 and Lys28 of Aβ42. They observed that this interaction was maintained throughout trajectories of wild-type amyloid fibres, and free energy calculations confirmed their role in stabilizing amyloid aggregates. While they accounted for intrinsic molecular details behind plasticity, Aβ aggregation was not directly addressed.
Kahler, Sticht and Horn (2013) have contributed more in this area. Their simulations revealed that Aβ fibrils in larger aggregates twisted at larger angles, and large aggregates subsequently broke. Smaller fragments then emerged as seeds to propagate aggregation. Recent research revealed that water plays a critical role in driving fibril assembly in Aβ40-seeded growth (Schwierz, Frost, Geissler, & Zacharias, 2016). These studies offered many insights, but their reliability has always been a matter of debate. As we shall see later, intrinsic changes of the surroundings seem to play a role in Aβ aggregation, especially concerning water in Aβ fibrils. It remains hard to factor in delicate contexts into the simulation, and this is where experimental approaches come in to validate such bioinformatics insights.
Spectroscopic and Spectrometric Techniques can Manipulate Real-Time Aβ Aggregation
A variety of spectroscopic/spectrometric techniques have emerged for investigating Aβ and other amyloid aggregation, offering high-resolution details of transient intermediates. Firstly, infrared spectroscopy relies on the excited vibrations of atoms in a molecule upon infrared radiation absorption (Figure 3A). While it is highly sensitive for antiparallel β-sheets, which are abundant in amyloids (Bruker Optics Inc, 2013), isotope labelling is needed to comprehend its complex spectrum. Usually, one or two carbonyl residues are labelled at a time by 13C and 18O (Arnaud, 2009), and the shifts of their peaks due to motions of neighbouring charges are also monitored. Another useful vibrational spectroscopy method is Raman spectroscopy (Figure 3A), which excites molecules to a higher virtual energy state and measures inelastic scattering of light (Kurouski, Van Duyne & Lednev, 2015). These spectroscopies have provided valuable information on the role of water and the disruption of lipid membranes. 2D infrared spectroscopy identified water neighbouring the amide groups of the hydrophobic Leu17 and Leu34 (Kim, Liu, Axelsen & Hochstrasser, 2009), which warranted more kinetic experiments to study its role. Raman spectroscopy showed that Aβ40 peptide within anionic lipid bilayers changes from disordered or helical conformations to β-sheets over time (Kurouski et al., 2015). However, since they can only collect averaged signals from the entire population of Aβ monomers, it cannot characterize individual conformations in the heterogeneous population at equilibrium as in mass spectrometry.
Figure 3B shows a particularly relevant mass spectrometry for Aβ aggregation dynamics based on hydrogen-deuterium exchange (HDX-MS), in which backbone amide hydrogens, which are exposed to solvent comprising “heavy water” (deuterium oxide) as a result of protein unfolding or hydrogen-bond-breaking, are replaced by deuterium (Eyles & Kaltashov, 2004). Each protium-deuterium exchange will increase the protein mass by a single unit and be detected. Peptide digestion and liquid chromatography further characterize solvent-accessible protein sites. When applied to Aβ structural dynamics, HDX-MS readily differentiated between monomers, protofibrils, and fibrils (Kheterpal et al., 2003). This study revealed that 40% of the backbone amide hydrogens were located within the core of protofibrils and that this increased to 60% in mature fibrils (Kheterpal et al., 2003), indicating dynamic changes of the positioning of residues throughout fibril formation. The same technique also demonstrated how the core of Aβ42 (residues 20-35) “seeded” aggregation, while the hydrophobic C-terminus played a minor but indispensable role (Zhang et al., 2013).
Single molecule force spectroscopy (SMSF) offers a profound resolution to study the mechanical stability of Aβ secondary structures. It consists of the molecule-of-interest attached to a surface with a probe on each end. The probe exerting large force is a sharp tip at the free end of a cantilever-beam for atomic force microscopy (AFM-SMSF; Eghiaian & Rico, 2014; Figure 3C). Apart from measuring forces applied to the molecule, changes in reflection angle of a laser beam off the surface of the cantilever can give information on topography (Eghiaian & Rico, 2014). Of note, AFM-SMSF, coupled with other methods, was instrumental in showing nucleation as a critical step in fibril formation (Harper, Lieber & Lansbury, 1997). This technique showed that Aβ1-40 β-sheets attached to one another in a reversible zipping motion (Kellermayer et al., 2005). Currently, high-speed AFM is emerging and has already contributed to elucidating the formation of false fibril branching points (Milhiet et al., 2010).
Researchers are constantly refining the above techniques and combining them for optimization. For instance, HDX was coupled with Raman spectroscopy for determining the Aβ peptide psi dihedral angles, the nearby environment of aromatic amino acids, and overall core structure of the fibril (Kurouski et al., 2015). Nevertheless, the sophisticated set-up, accessibility, and requirement of training make these experimental techniques better candidates for validation than discovery.
Simulations and Experiments go Hand in Hand in Understanding Aβ Aggregation
It is evident now that simulations and experimental approaches are complementary to each other. One could benefit from the atomistic resolution that MD simulations propose for Aβ fibrillation and aggregation. The experimental evidence was solid for the biological relevance of these models. The idea of validations across different platforms is typically employed in refining experimentally solved structures (Raval, Piana, Eastwood, Dror & Shaw, 2012; Schröder, 2015). It has been extended by, for example, coupling modelling with NMR or small-angle X-ray scattering (SAXS) to understanding fluctuations of protein structures (Bernadó et al., 2010; Yang et al., 2007). In the Aβ field, the integration of in vitro and in silico approaches has offered meaningful insights into Aβ aggregation.
One contribution of combining the two approaches concerns the role of water. Water has been considered critical since early on as some drew parallelism of fibril formation with crystallization: In the latter process, escape of water molecules stands as a necessity (Thirumalai, Reddy & Straub, 2012), which resembles fibril formation where the landscape of solvation must change to accommodate aggregation. In Aβ fibrils, this happens early on (Tarus, Straub & Thirumalai, 2006) and is probably related to the establishment of hydrophobic interactions that organise aggregation. Many experimental and simulation studies have captured extensive conformational plasticity mediated by different contexts of water. Simulations exhibited discrepancies with respect to the amount of water within the pre-fibril nuclei (Krone et al., 2008; Tarus, Straub, & Thirumalai, 2006). The simulations of water channels, e.g., around the Asp23-Lys28 salt bridge (Lemkul & Bevan, 2010; Tarus et al., 2006), were validated by 2D infrared spectroscopy as previously discussed (Kim et al., 2009). This provides an example of integrating the two approaches: simulations described interesting interactions between different parts of Aβ with water, while experimental techniques addressed the discrepancies among in silico models.
Integrating in silico and in vitro approaches also contributed to understanding the influence of membrane lipid composition in Aβ42 embedment at the membrane, which modulates Aβ aggregation. This has first been primed by “lipidomics” strategies, characterising the compositions and prevalence of membrane lipid species across the human brain (Svennerholm, Boström, Jungbjer & Olsson, 1994). Contemporary large-scale shotgun MS-based lipid measurements (Han, 2010) added to this, by demonstrating that cholesterol redistribution is dependent upon aging and AD. Computational approaches contributed to the mechanistic explanation of these phenomena, e.g. in a study by Liguori, Nerenberg, Paul, and Head-Gordon (2013). They showed, with free energy calculations, that the abundant extracellular cholesterol favours extrusion of N-terminals of Aβ fibrils, probing it to oligomerise and aggregate. On the other hand, the helicity at the C-terminal of Aβ peptides contributed to retain themselves at the plasma membrane. Amongst other intrinsic and delicate components of the local environment which “omics” experiments have primed, MD simulations give fine resolution into their role for mediating Aβ association, folding, and oligomerization.
It is also interesting to look at the interplay of computational and experimental approaches in a broader, time-lapse view of Aβ research. Long before simulations were available, conventional structural biology had established that cross-β structures were common in aggregating amyloid fibrils. ssNMR provided direct evidence that the same peptide sequence could form a vast variety of polymorphs (Figure 1A). Although oligomers existed before protofibrils and fibrils, the rearrangement of weaker oligomer intermediates might be directed to either fibril formation or growth into various stable forms of oligomers, depending on the physical and chemical environment (Figure 4i-iv) as concluded in a ssNMR study (Tay, Huang, Rosenberry & Paravastu, 2013).
Once the oligomers are committed to forming protofibrils and fibrils, they start to elongate extensively. In MD-simulated seeded growth (Kahler et al., 2013), single layers of protofibrils attempted to elongate until stability broke and they dissociated into short protofibrils. Later, pairing occured and resulted in protofibril pairs and possibly triplets with varying molecular structures caused by subtle differences in growth conditions, as shown by EM and ssNMR (Petkova et al., 2005). The increase in β-sheet structures gave the fibril greater stability as revealed by Fourier transform infrared spectroscopy (Kodali, Williams, Chemuru & Wetzel, 2010). Additionally, this allowed further elongation into large fibrils (Figure 4v-x).
Subjected to the underlying molecular structures, these fibrils were polymorphic with unique physical properties. Nevertheless, all of these structures (including the oligomers) are in dynamic equilibrium with constant fragmentation and secondary nucleation which enable growing of new fibrils along the old ones, according to a series of kinetic, radio-labelling and cell viability experiments (Cohen et al., 2013). Altogether, this elaborate pathway of Aβ aggregation is the product of combining computational and experimental approaches in discovering biophysical basis at each step. Either one of these is indispensable in advancing mechanistic understanding of aggregation.
Conclusion and Future Perspectives
We presented an overview of the mechanism of Aβ aggregation and highlighted that computational and experimental approaches go hand in hand behind this. The synergy has contributed to identifying the distinct steps of aggregation as well as their biophysical underpinnings. This has allowed numerous attempts in identifying the so-called “wonder drugs”, which effectively antagonise aggregation and bring the least side effects (Karran et al., 2011). An understanding of the aggregative structural polymorphs can lead to possible strategies to target Aβ aggregation, e.g. by utilising the agents that disrupt salt bridges to destabilise the fibrils. For the theoretical biophysicists, the success of detailing the aggregative process also bears significant implications. It is this well-positioned biophysical problem that puts MD simulation to its best use, probing details that in vitro approaches often lack. The rise of simulation as a well-recognised approach to study biomolecular dynamics, as well as the excitement that the distributed computing project Folding@home (http://folding.stanford.edu/home) has brought to both the academic field and the media, bring protein folding and misfolding research to new heights. Further refinement, e.g., using patient-derived fibrils, would resolve ambiguities in this mechanism, considering the model in Figure 1B(iii) being the only, to our knowledge, published patient-derived Aβ fibril structure (Paravastu, Qahwash, Leapman, Meredith & Tycko, 2009). This would inspire further computational simulation and experimental validation to provide details on the basis of preferring one polymorph over another and their relevance to neuropathology. More broadly, combining different lines of investigations help greatly in understanding mechanistic aspects of many critical proteins in other diseases awaiting research efforts.
Ahmed, M., Davis, J., Aucoin, D., Sato, T., Ahuja, S., Aimoto, S., Elliott, J. I., Van Nostrand, W. E., & Smith, S. O. (2010). Structural conversion of neurotoxic amyloid-β(1-42) oligomers to fibrils. Nature Structural & Molecular Biology, 17(5), 561-567.
Arnaud, C. H. (2009). Tools for Amyloids. Chemical and Engineering News, 87.
Bernadó, P., Modig, K., Grela, P., Svergun, D. I., Tchorzewski, M., Pons, M., & Akke, M. (2010). Structure and Dynamics of Ribosomal Protein L12: An Ensemble Model Based on SAXS and NMR Relaxation. Biophysical Journal, 98(10), 2374-2382.
Bruker Optics Inc. (2013). Study of protein conformation with FTIR. In Bruker Optics Inc (Ed.), (pp. 1-3). USA.
Cerasoli, E., Ryadnov, M. G., & Austen, B. M. (2015). The elusive nature and diagnostics of misfolded Aβ oligomers. Frontiers in Chemistry, 3, 17.
Cohen, S. I. A., Linse, S., Luheshi, L. M., Hellstrand, E., White, D. A., Rajah, L., Otzen, D. E., Vendruscolo, M., Dobson, C. M., & Knowles, T. P. J. (2013). Proliferation of amyloid-β42 aggregates occurs through a secondary nucleation mechanism. Proceedings of the National Academy of Sciences, 110(24), 9758-9763.
Eghiaian, F., & Rico, F. (2014). High-Speed Atomic Force Microscopy: Imaging and Force Spectroscopy. FEBS Letters, 588(19), 3631-3638.
Eyles, S. J., & Kaltashov, I. A. (2004). Methods to Study Protein Dynamics and Folding by Mass Spectrometry. Methods, 34, 88-99.
Han, X. (2010). Multi-dimensional mass spectrometry-based shotgun lipidomics and the altered lipids at the mild cognitive impairment stage of Alzheimer’s disease. Biochimica et Biophysica Acta (BBA) – Molecular and Cell Biology of Lipids, 1801(8), 774-783.
Harper, J. D., Lieber, C. M., & Lansbury, P. T. J. (1997). Atomic Force Microscopic Imaging of Seeded Fibril Formation and Fibril Branching by the Alzheimer’s Disease Amyloid-β Protein. Chemistry and Biology, 4(12), 951-959.
Jernigan, R. L., & Bahar, I. (1996). Structure-derived potentials and protein simulations. Current Opinion in Structural Biology, 6(2), 195-209.
Kahler, A., Sticht, H., & Horn, A. H. C. (2013). Conformational Stability of Fibrillar Amyloid-β Oligomers via Protofilament Pair Formation – A Systematic Computational Study. PloS One, 8(7), 1-12.
Karran, E., Mercken, M., & De Strooper, B. (2011). The amyloid cascade hypothesis for Alzheimer’s disease: an appraisal for the development of therapeutics. Nature Reviews Drug Discovery, 10(9), 698-712.
Kellermayer, M. S. Z., Grama, L., Karsai, A., Nagy, A., Kahn, A., Datki, Z. L., & Penke, B. (2005). Reversible Mechanical Unzipping of Amyloid β-Fibrils. Journal of Biological Chemistry, 280(9), 8464-8470.
Kheterpal, I., Lashuel, H. A., Hartley, D. M., Walz, T., Lansbury, P. T. J., & Wetzel, R. (2003). Aβ Protofibrils Possess a Stable Core Structure Resistant to Hydrogen Exchange. Biochemistry, 42(48), 14092-14098.
Kim, Y. S., Liu, L., Axelsen, P. H., & Hochstrasser, R. M. (2009). 2D IR provides evidence for mobile water molecules in β-amyloid fibrils. Proceedings of the National Academy of Sciences, 106(42), 17751-17756.
Klimov, D. K., & Thirumalai, D. (2003). Dissecting the Assembly of Aβ16-22 Amyloid Peptides into Antiparallel β Sheets. Structure, 11(3), 295-307.
Kodali, R., Williams, A. D., Chemuru, S., & Wetzel, R. (2010). Aβ(1-40) forms five distinct amyloid structures whose beta-sheet contents and fibril stabilities are correlated. Journal of Molecular Biology, 401(3), 503-517.
Krone, M. G., Hua, L., Soto, P., Zhou, R., Berne, B. J., & Shea, J. E. (2008). Role of Water in Mediating the Assembly of Alzheimer Amyloid-β Aβ16−22 Protofilaments. Journal of the American Chemical Society, 130(33), 11066-11072.
Kurouski, D., Van Duyne, R. P., & Lednev, I. K. (2015). Exploring the Structure and Formation Mechanism of Amyloid Fibrils by Raman Spectroscopy: a Review. Analyst, 140(15), 4967-4980.
Lemkul, J. A., & Bevan, D. R. (2010). Assessing the Stability of Alzheimer’s Amyloid Protofibrils Using Molecular Dynamics. Journal of Physical Chemistry. B, 114(4), 1652-1660.
Liguori, N., Nerenberg, Paul S., & Head-Gordon, T. (2013). Embedding Aβ42 in Heterogeneous Membranes Depends on Cholesterol Asymmetries. Biophysical Journal, 105(4), 899-910.
Milhiet, P. E., Yamamoto, D., Berthoumieu, O., Dosset, P., Le Grimellec, C., Verdier, J. M., Marchal, S., & Ando, T. (2010). Deciphering the Structure, Growth and Assembly of Amyloid-like Fibrils Using High-speed Atomic Force Microscopy. PloS One, 5(10), 13240.
Näslund, J., Schierhorn, A., Hellman, U., Lannfelt, L., Roses, A. D., Tjernberg, L. O., Silberring, J., Gandy, S. E., Winblad, B., & Greengard, P. (1994). Relative abundance of Alzheimer Aβ amyloid peptide variants in Alzheimer disease and normal aging. Proceedings of the National Academy of Sciences, 91(18), 8378-8382.
Paravastu, A. K., Qahwash, I., Leapman, R. D., Meredith, S. C., & Tycko, R. (2009). Seeded growth of β-amyloid fibrils from Alzheimer’s brain-derived fibrils produces a distinct fibril structure. Proceedings of the National Academy of Sciences, 106(18), 7443-7448.
Perez, A., Morrone, J. A., Simmerling, C., & Dill, K. A. (2016). Advances in free-energy-based simulations of protein folding and ligand binding. Current Opinion in Structural Biology, 36, 25-31.
Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M., Mattson, M. P., & Tycko, R. (2005). Self-Propagating, Molecular-Level Polymorphism in Alzheimer’s β-Amyloid Fibrils. Science, 307, 262-265.
Raval, A., Piana, S., Eastwood, M. P., Dror, R. O., & Shaw, D. E. (2012). Refinement of protein structure homology models via long, all-atom molecular dynamics simulations. Proteins: Structure, Function and Bioinformatics, 80(8), 2071-2079.
Robustelli, P., Stafford, K. A., & Palmer, A. G. (2012). Interpreting Protein Structural Dynamics from NMR Chemical Shifts. Journal of the American Chemical Society, 134(14), 6365-6374.
Schröder, G. F. (2015). Hybrid methods for macromolecular structure determination: experiment with expectations. Current Opinion in Structural Biology, 31, 20-27.
Schwierz, N., Frost, C. V., Geissler, P. L., & Zacharias, M. (2016). Dynamics of Seeded Aβ40-Fibril Growth from Atomistic Molecular Dynamics Simulations: Kinetic Trapping and Reduced Water Mobility in the Locking Step. Journal of the American Chemical Society, 138(2), 527-539.
Shirahama, T., & Cohen, A. S. (1967). High-resolution electron microscopic analysis of the amyloid fibril. Journal of Cell Biology, 33(3), 679-708.
Svennerholm, L., Boström, K., Jungbjer, B., & Olsson, L. (1994). Membrane Lipids of Adult Human Brain: Lipid Composition of Frontal and Temporal Lobe in Subjects of Age 20 to 100 Years. Journal of Neurochemistry, 63(5), 1802-1811.
Tarus, B., Straub, J. E., & Thirumalai, D. (2006). Dynamics of Asp23−Lys28 Salt-Bridge Formation in Aβ10-35 Monomers. Journal of the American Chemical Society, 128(50), 16159-16168.
Tay, W. M., Huang, D., Rosenberry, T. L., & Paravastu, A. K. (2013). The Alzheimer’s amyloid-β(1-42) peptide forms off-pathway oligomers and fibrils that are distinguished structurally by intermolecular organization. Journal of Molecular Biology, 425(14), 2494-2508.
Thirumalai, D., Reddy, G., & Straub, J. E. (2012). Role of Water in Protein Aggregation and Amyloid Polymorphism. Accounts of Chemical Research, 45(1), 83-92.
Tomic, J. L., Pensalfini, A., Head, E., & Glabe, C. G. (2009). Soluble fibrillar oligomer levels are elevated in Alzheimer’s disease brain and correlate with cognitive dysfunction. Neurobiology of Disease, 35(3), 352-358.
Yang, L. W., Eyal, E., Chennubhotla, C., Jee, J., Gronenborn, A. M., & Bahar, I. (2007). Insights into Equilibrium Dynamics of Proteins from Comparison of NMR and X-Ray Data with Computational Predictions. Structure, 15(6), 741-749.
Yu, L., Edalji, R., Harlan, J. E., Holzmann, T. F., Lopez, A. P., Labkovsky, B., Hillen, H., Barghorn, S., Ebert, U., Richardson, P. L., Miesbauer, L., Solomon, L., Bartley, D., Walter, K., Johnson, R. W., Hajduk, P. J., & Olejniczak, E. T. (2009). Structural Characterization of a Soluble Amyloid β-Peptide Oligomer. Biochemistry, 48, 1870-1877.
Zhang, Y., Rempel, D. L., Zhang, J., Sharma, A. K., Mirica, L. M., & Gross, M. L. (2013). Pulsed Hydrogen-Deuterium Exchange Mass Spectrometry Probes Conformational Changes in Aβ Peptide Aggregation. Proceedings of the National Academy of Sciences, 110(36), 14604-14609.
Pulley Optimization for a Walking-Engine-Actuated Active Ankle-Foot Orthosis
Active orthotic devices for joint articulation have a vast number of applications that could benefit many people. Examples include joint articulation for people suffering from disabilities, increased load carrying capacity and walking distance for humans, and gait training. The main goal of this research is to help people with disabilities regain natural walking ability by replicating normal walking gait through the use of an active ankle-foot orthosis (AAFO). This research investigates the optimization of a pulley system for the primary actuator of an AAFO utilizing a high-efficiency pneumatic “Walking Engine.” In order to accurately replicate a healthy human gait, the AAFO device had to accurately reproduce the moment applied to the ankle during the gait cycle. The AAFO’s internal-combustion (IC) engine was characterized using a dual-combustion (limited-pressure) gas-power-cycle model. With the dual-combustion model, both a theoretical pressure-volume diagram and the thermodynamic engine efficiency were calculated. Using the calculated pressure output of the IC engine, the pulley system was optimized to best match the ankle moment of a healthy human gait, which was obtained from David Winter’s Biomechanics and Motor Control of Human Movement. The optimized pulley geometry is very complicated and additional research is necessary to utilize its design. The results of this research provide insight for the future development of untethered, lightweight, efficient AAFO devices.
With increasing advancements in robotics technology, robotic systems, such as the Walking Engine, are continuously being implemented in new and exciting ways. As a result, humans experience ever increasing interactions with robotic systems on both a professional and personal level. On the personal level, these robotic systems can be used to assist with physical impairment, amplify normal physical capabilities, and even act as an extension of human abilities for the 35.2 million people suffering from physical functioning disabilities (Dollar & Herr, 2007). Some well-publicized examples include the systems being developed by Boston Dynamics for assisting military personnel in carrying equipment (Big Dog, Little Dog) and exoskeletons that provide walking assistance for those normally bound to wheel chairs for mobility (ReWalk, EKSO) (Dollar & Herr, 2007). The main goal of this research is to design an active ankle-foot orthosis (AAFO) that helps people with disabilities regain natural walking ability by replicating the normal walking gait of a human. This paper investigates the combination of robotic system technology and orthotics in the form of an AAFO to accomplish this goal.
Active Ankle-Foot Orthosis
Orthotics is a specialty within the medical field concerned with the design, manufacture, and application of an orthosis. An orthosis is an externally applied device used to modify the structural and functional characteristics of the neuromuscular and skeletal system (Lusardi, Jorge, & Nielson, 2013). Orthotic devices can be designed to control, guide, limit, or immobilize an extremity, joint, or body segment. Under the International Standard terminology, orthoses are classified by an acronym describing the anatomical joints that they contain (Lusardi et al., 2013). For example, an orthosis applied to the ankle and foot, such as the one investigated in this research, is an ankle-foot orthosis (AFO). AFOs are commonly used in the treatment of disorders that affect muscle function such as stroke, spinal cord injury, muscular dystrophy, cerebral palsy, polio, multiple sclerosis, and peripheral neuropathy by: providing artificial joint articulation, increasing load-carrying capabilities; and, assisting in gait-training procedures (i.e. people re-learning to walk from an injury or disability) (Lusardi et al., 2013).
Currently, ankle-foot orthoses fall into the two main categories of passive and active (powered) orthoses. Figure 1 shows a comparison of a passive and active AFO system. Passive AFOs rely on an external energy source (the user) and either utilize mechanical components, such as springs, to assist with movement, or are rigid and simply immobilize the joint (Lusardi et al., 2013). Passive devices are typically lightweight with simple designs, resulting in low cost and high robustness. However, despite their benefits, many problems exist in current passive AFO designs. While most patients see improvement with their ability to walk with the aid of these devices, the gait is labored and very unnatural (Lenhart & Sumarriva, 2008). AAFO devices seek to combat this limitation by replicating human gait with an integrated power source. AAFOs are typically powered using a pneumatic or electric power source (Lusardi et al., 2013). This integrated power source allows for more precise movement and complex functionality of the ankle joint. This complexity adds extra weight and size to the apparatus (Figure 1). One major shortcoming of current AAFO designs is this tradeoff between size and functionality. An ideal AAFO design must be compact and lightweight, in addition to having the capability for the complex functions needed to accurately replicate human gate. The issue of complexity versus weight can be addressed by using an internal-combustion (IC) engine, such as a “Walking Engine”, to power the device.
High Efficiency Pneumatic Walking Engine
It is understood that hydrocarbon fuels have a much higher specific energy than electric batteries or compressed air (Mitchell, Gallant, Thompson, & Shao-Horn, 2011). While liquid petroleum gases, such as butane, propane, or methane, have specific energies of around 45-50MJ/kg, electric batteries have specific energies of only around 9-10MJ/kg. Since a compact and lightweight design is required, an AAFO utilizing a small IC engine as the primary means of ankle articulation was chosen for this research. Propane, a common liquid petroleum gas, was selected as the IC engine’s fuel source due to its associated ease of implementation and high energy density.
To further meet the efficient and lightweight AAFO design requirements, an IC “Walking Engine” was chosen over a typical IC engine due to its increased thermodynamic and system efficiency. In a typical IC engine, a portion of the energy obtained from combustion is used to compress the engine’s pistons with the use of a flywheel. Consequently, this energy does not contribute to the forward motion of the vehicle. On the contrary, with a “Walking Engine” the user of the AAFO can be viewed as the flywheel. The kinetic energy from the forward motion of the user also supplies the energy to compress the pistons, providing much higher system efficiency. Thermodynamically speaking, a typical IC engine converts merely 25% of the applied fuel energy into usable energy, suffering losses from exhaust (40%), coolant (30%), and friction (5%) (Heywood, 1988). Conversely, a “Walking Engine” converts a substantial 55% of the applied fuel energy into usable energy. This is accomplished with the implementation of a passive cooling system and reusing recovered exhaust energy in a pneumatic system.
A design for an AAFO has been developed that utilizes a forced-induction, “Walking Engine” actuation system, as shown in Figure 2. The design employs two opposing (antagonistic), asymmetric piston actuators to provide the required power for locomotion. The primary actuator is driven by a forced-induction, propane-powered, IC “Walking Engine”. Exhaust energy from the combustion engine is captured to charge a pneumatic system that will drive the second actuator.
The two grey cylinders shown in Figure 2 are the two actuating pistons. On the left is the primary actuator, which is the IC engine and provides the power required for the plantarflexion phase of the gait cycle. The green object in the schematic represents the fuel canister, while the blue tube represents the surrounding air. The fuel and the air will be forced into the combustion chamber with a hydraulic bellows pump. The bellows pump is shown under the heel of the foot bed. For ignition, a spark plug device will be activated, creating a spark. The exhaust from the engine is shown on the diagram as the red area connecting the primary cylinder to the secondary cylinder. The energy from the engine exhaust is captured and used to pneumatically power the secondary actuator. The secondary actuator provides the power for the dorsiflexion phase of the gait cycle. This design represents the optimal compromise between size, weight, and power.
Materials and Methods
The main goal of this research is to help people with disabilities regain natural walking ability by replicating the normal walking gait of a human through the use of an AAFO device. We investigated the optimization of a pulley system for the primary actuator of an AAFO utilizing a high-efficiency pneumatic “Walking Engine”. To accurately replicate a healthy human gait, the AAFO device had to properly reproduce the moment applied to the ankle during that gait. Thus, the IC engine used on the AAFO had to be properly characterized. Once the IC engine’s performance was accurately modeled, a pulley system was optimized to best match the ankle moment of a healthy human gait.
Walking Engine Characterization
The IC engine used as the “Walking Engine” in the AAFO device was a pre-manufactured, or off-the-shelf (OTS), Bosch pneumatic actuator. The OTS actuator, having an inner chamber diameter of 25mm and a maximum piston extension of 50mm, was modified for use as an IC engine. Figure 3 shows a computer-aided design (CAD) model of the modified OTS Bosch actuator.
To convert the Bosch actuator into an operational IC engine, several modifications were necessary. First, a hole was drilled into the bottom of the actuator to allow for the insertion of the spark plug ignition source. A block extension was then added to the bottom of the actuator to ensure that the actuator could withstand the stresses generated from the 10:1 compression ratio combustion. A metal piston extension was also added to the bottom of the actuator’s piston to protect the rubber-sealing device from the flames generated during combustion. Finally, for experimental combustion testing, a needle valve was attached to the bottom actuator port to allow the fuel/air mixture in, hold the pressure during compression and combustion, and allow the exhaust gas to escape the engine chamber.
The first step in characterizing the IC engine used on the AAFO was to calculate the gas power cycle for the engine. A gas power cycle, or thermodynamic engine cycle, consists of a linked sequence of thermodynamic processes that involve the transfer of heat and work into and out of the engine system that eventually returns the system to its initial state (Heywood, 1988). Each thermodynamic engine cycle for the AAFO engine corresponds to one step of the leg equipped with the device during the gait cycle (i.e., one cycle per step by the leg that wears the device). A pressure-volume diagram is often used to represent the thermodynamic cycle of an engine. A dual-combustion (limited-pressure) pressure-volume diagram was chosen to model the gas power cycle of the IC engine used to power the AAFO.
A dual-combustion engine cycle is a combination of both the Otto and the Diesel engine cycle (Heywood, 1988). Energy from combustion is added as heat in two different stages, q1’ and q1”(Figure 4). Heat addition occurs first at a constant volume, similar to the Otto cycle, and then at a constant pressure, resembling the Diesel cycle. A dual-combustion cycle was chosen to represent the AAFO’s IC engine because this cycle is a closer approximation to the real life behavior of IC engines. In typical applications, the combustion process of an IC engine does not occur perfectly at a constant volume or a constant pressure, but rather in two separate stages as indicated by the dual cycle (Heywood, 1988).
When calculating the gas power cycle of the “Walking Engine” some important assumptions about the engine process were made in order to simplify the calculations. The first major assumption, which is assumed in most engine cycles, is that the working fluid (the air/fuel mixture) of the engine remains in a gaseous state throughout the cycle. It is also assumed that the working fluid can be modeled solely as air and can always be treated as an ideal gas. Furthermore, this research assumes an ideal engine cycle in which internal irreversibilities and other complexities, such as the combustion process and the exhaust of the products of combustion, are removed (Heywood, 1988).
In addition to these initial assumptions, the IC engine cycle was approximated with the following assumptions.
⋅ All of the engine processes listed in Figure 4 are internally reversible.
⋅ A dual-heat-addition process from an external source replaces the combustion process.
⋅ The lower heating value (LHV; 2044kJ/mol) of propane was used with an 80% burn efficiency during combustion calculations (Linstrom & Mallard, 2001).
⋅ A perfect stoichiometric air/fuel mixture was also used for the combustion calculations.
Thermodynamic Engine Efficiency
The second step in characterizing the IC engine used to power the AAFO was to calculate the thermodynamic efficiency of the engine. The thermodynamic efficiency of an engine is a dimensionless relationship between the total energy supplied to the engine—through the combustion of fuel—and the amount of energy available to perform useful work (Heywood, 1988). In other words, thermodynamic efficiency indicates how well an energy conversion process is accomplished. The thermodynamic efficiency of a dual-combustion engine cycle is represented by Equation 1:
where, as Figure 4 shows, q1’ is heat addition at a constant volume, q1” is heat addition at a constant pressure, and q2 is heat rejection during exhaust. By using the fact that an energy exchange can be modeled by heat capacity, q1’, q1”, and q2 can be rewritten as the following:
By substituting in these relations, Equation 1 can be rewritten in the following, more convenient form:
Thus, the thermodynamic efficiency of the IC engine used to power the AAFO can be calculated using Equation 3 and the temperatures found from the gas power cycle. It is interesting to note that because the dual-combustion engine cycle is a combination of the Otto and Diesel cycles, the efficiency of the dual-combustion cycle will always fall somewhere between the Otto and Diesel.
Pulley Optimization Process
In order to accurately replicate a healthy human gait, the AAFO device had to properly reproduce the moment applied to the ankle during that gait. This was done with a pulley system connected to the AAFO’s actuating pistons. Figure 5 shows a conceptual model of the pulley system used to power the AAFO.
By finding the pressures in the actuator chambers and multiplying them by the cross-sectional area of the actuators’ pistons, AC, the force output of the engine system can be determined. Dividing the optimal ankle moment by the engine’s force output results in the optimal pulley geometry. Equation 4 represents this procedure.
|Ankle Moment = (Pressure ※ Ac) ※ Pulley Radius||(Eq. 4)|
Since our AAFO design utilizes a dual piston configuration, a dual pulley configuration is also necessary. The primary actuator (IC engine), which is responsible for the plantarflexion phase of the gait cycle, requires different pulley radii than the secondary actuator (pneumatic exhaust recovery system), which is responsible for the dorsiflexion phase of the gait cycle, to match the ankle moment. A dual pulley configuration has been designed to allow for the optimization of both the inner pulley, the primary actuator, and the outer pulley for the secondary actuator (Figure 5). However, this research focuses only on the optimization of the primary actuator pulley.
There is no question that AAFOs have the potential to improve the quality of life for many people, but current AAFO designs lack a balance between functionality and compactness. The results of this research will be used to help further the development of the “Walking Engine” AAFO by addressing these shortcomings.
Characterized Walking Engine
As stated earlier, before the primary actuator pulley could be optimized, the AAFO’s IC engine had to be characterized. Figure 6 shows the calculated ideal dual-combustion pressure-volume diagram that was used to model the IC engine. Each step by the leg equipped with the AAFO corresponds to one thermodynamic engine cycle that can be broken into two main phases: the swing phase and the stance phase. The engine cycle begins at the start of the stance phase where the heel of the foot comes into contact with the ground (HCR), which is indicated by point 1 in Figure 6. At this time the fuel/air mixture is injected into the engine chamber (points 1-3) via a bellows pump. This fuel is compressed, ignited, and expanded to the actuators maximum extension, from points 3 to 7, to provide the locomotion power for the plantarflexion portion of gait. The remainder of the gait cycle, the swing phase, begins when the toe of the foot leaves contact with the ground (TOR) and is comprised of points 7, 8, and 1. The swing phase corresponds to the exhaust phase of the engine, which is used to power the dorsiflexion portion of gait with the help of the secondary pneumatic actuator.
Table 1 shows the volume of the engine chamber, pressure and temperature in the engine chamber, and the number of moles of working fluid in the engine at each of the indicated points in Figure 6. Using these values, the engine’s work output was calculated by finding the area inside the P-V curve. The work output of the AAFO engine was calculated to be 29.80J, which is more than enough energy to actuate the ankle joint.
By substituting the temperatures of the key points in the engine cycle into Equation 3, the theoretical thermodynamic engine efficiency of the modified OTS actuator was calculated: a thermodynamic engine efficiency of 0.61. The P-V diagram calculated in Figure 6 will be validated with experimental combustion testing data and altered to match actual engine performance if necessary.
Optimized Primary Actuator Pulley Configuration
Once the AAFO engine was properly characterized, the primary actuator pulley system could then be optimized. The first step in optimizing the AAFO pulley was to optimize the ankle moment data. The ankle moment data that was replicated comes from David Winter’s Biomechanics and Motor Control of Human Movement. As Figure 7 shows, eliminating the sudden jumps in the data and removing the negative moment values, which would have been impossible to replicate, resulted in an optimal ankle moment curve. This smoothed moment data allowed for a smooth and precise optimal pulley configuration to be calculated.
The optimal primary actuator pulley radius was calculated by finding the pressure in the engine as a function of time, from the pressure-volume diagram, and rearranging Equation 4 to solve for the pulley radius. It is important to note that while the pulley radius may change it is always assumed that the edge of the pulley and the AAFO engine are in line with one another. Figure 8 shows the optimal pulley radius as a function of time alongside several constant diameter pulleys. This parametric polar plot of the pulley configuration starts at the green dot and moves along the geometry line with time ending at the red dot.
The results of this research show that the modified OTS actuator, which has a theoretical work output of 29.80J with a thermodynamic engine efficiency of 0.61, is able to provide more than enough energy to activate the ankle joint, which makes it an acceptable power source for the AAFO. Using a theoretical P-V diagram (Figure 6), which will be validated with combustion testing, the optimal pulley radius for the plantarflexion phase of the gait cycle was determined. While the optimal pulley radius for the primary actuator was calculated (Figure 8), it has a very complicated geometry.
This complicated pulley geometry stems from three main issues. The first issue is that during the gait cycle the ankle only rotates through a small 27° window. Because the ankle only rotates 27°, only 27° of the pulley perimeter—indicated by the red pie slice in Figure 8—is being used. This results in the squished pulley geometry that is shown. The second major issue is that during the gait cycle the ankle oscillates between negative and positive angle values, resulting in the pulley geometry doubling back on itself. This happens because each time the ankle rotates through a previous degree there is a different pressure in the AAFO engine and a different ankle moment that needs to be matched, resulting in different pulley radii at the same ankle angle. The final complication with the pulley geometry is the extreme pulley radii (2 to 15cm) that are experienced due to the relationship between the engine pressure and ankle moment. Because the ankle moment is such a small value during that beginning of the gait cycle, the radius necessary to match the moment is small, but when the moment reaches its larger values the radii necessary to reproduce this with the engine pressure are very large. It should also be noted that this pulley geometry is specific to the ideal P-V performance of the engine. The optimal pulley geometry will be subject to change when actual engine performance data becomes available through experimental combustion testing.
This paper introduced a unique AAFO design that utilizes a forced-induction, “Walking Engine” system to provide the power for locomotion. The implication of this research is that a modified OTS actuator used as an IC engine is a feasible power source for an AAFO. Providing 29.80J of work output and weighing less than 10oz, this design provides the balance between size and functionality that is not available with current designs. While the optimal pulley radius for the primary actuator was calculated, it has a very complicated geometry. Further research is necessary to determine a practical pulley system, such as a gear or chain torque converter, that allows for the successful implementation of the optimal pulley configuration. This research has laid necessary groundwork for future experiments to further the development of the “Walking Engine” AAFO. This work, characterizing the IC engine and optimizing the pulley radius, will be applied to future research used to construct a fully functional “Walking Engine” AAFO.
This project was funded by the National Science Foundation (NSF) and the Milwaukee School of Engineering (MSOE) Fluid Power Institute (FPI). A special thanks to Douglas Cook, Vito Gervasi, the Research Experience for Undergraduates (REU) staff, Milwaukee School of Engineering, The Center for Compact and Efficient Fluid Power, the Fluid Power Institute, and all other parties involved, because only with their help was this research possible.
This material is based upon work supported by the National Science Foundation under Grant No. EEC-0540834. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.
Blackwell, D. L., Lucas, J. W., & Clarke, T. C. (2014). Summary health statistics for US adults: national health interview survey, 2012. Vital and health statistics. Series 10, Data from the National Health Survey, (260), 1-161.
Dollar, A. M., & Herr, H. (2007). Active orthoses for the lower-limbs: challenges and state of the art. In 2007 IEEE 10th International Conference on Rehabilitation Robotics (pp. 968-977). IEEE. doi:10.1109/ICORR.2007.4428541
Heywood, J. (1988). Internal combustion engine fundamentals. New York: McGraw-Hill.
Lenhart, R. L., & Sumarriva, N. (2008). Design of Improved Ankle-Foot Orthosis. University of Tennessee Honors Thesis Projects.
Linstrom, P. J., & Mallard, W. G. (2001). The NIST Chemistry WebBook: A chemical data resource on the internet. Journal of Chemical & Engineering Data, 46(5), 1059-1063. doi:10.1021/je000236i
Lusardi, M. M., Jorge, M., & Nielsen, C. C. (2013). Orthotics and prosthetics in rehabilitation. Elsevier Health Sciences. doi:10.1016/b978-1-4377-1936-9.09990-2
Mitchell, R. R., Gallant, B. M., Thompson, C. V., & Shao-Horn, Y. (2011). All-carbon-nanofiber electrodes for high-energy rechargeable Li–O 2 batteries. Energy & Environmental Science, 4(8), 2952-2958. doi:10.1039/c1ee01496j
Winter, D. A. (2009). Biomechanics and motor control of human movement. John Wiley & Sons. doi:10.1002/9780470549148
Opportunity for Pharmaceutical Intervention in Lung Cancer: Selective Inhibition of JAK1/2 to Eliminate EMT-Derived Mesenchymal Cells
The Epithelial Mesenchymal Transition (EMT) has been implicated as a driving force behind the metastasis of epithelial derived cancers; it stimulates the acquisition of a migratory, drug resistant mesenchymal phenotype. Current proposals for targeting EMT-facilitated metastasis are ineffective, as a majority focus on the inhibition of EMT-initiating signals. Instead, this study used a novel approach aimed at directly inhibiting the mesenchymal phenotype by targeting mesenchymal survival pathways post mapping at specifically up-regulated points in the mesenchymal state. In vitro EMT models of three lung adenocarcinoma cell lines (A549, HCC827, HCC4006) were each treated with three concentrations (50nM, 0.5µM, and 5µM) of AZD1480, BAY87-2243, MK-2206, and GDC0994 at 48 hours, which targeted the JAK/STAT, PI3K-AKT, and MAPK pathways, respectively. MTT assays were used to quantify cell death and determine cell viability. Across all assays, the JAK 1/2 inhibitor AZD1480 resulted in the greatest elimination of EMT-xderived mesenchymal cells. 5µM AZD1480 significantly reduced cell viability in mesenchymal populations treated with AZD1480, supporting JAK1/2 as a potential therapeutic target in lung cancer. Future investigations include testing AZD1480’s effectiveness in decreasing cell viability among alternative epithelial-derived cell lines in vitro and its effectiveness as a suppressor of metastasis in vivo.
Lung cancer is currently the most prevalent form of cancer in the United States, causing over 158,000 deaths each year (American Lung Association, 2016). More than 90% of these deaths and other cancer-associated mortality can be attributed to metastasis (Ray & Jablons, 2009). An immense clinical need exists for novel treatments either targeting or preventing metastasis, especially in early stage cancer patients.
In order for distant metastasis to occur, primary tumor cells must disseminate through the blood vessels and invade a distant organ site (Brabletz, 2012). A biological phenomenon known as the Epithelial-Mesenchymal Transition (EMT) is a key facilitator of this process. EMT involves a series of changes which allow a cancer cell to transition from a stationary epithelial phenotype to a migratory, drug resistant mesenchymal phenotype. This process includes the disassembly of epithelial cell-junctions and a loss of epithelial polarity in exchange for mesenchymal characteristics, such as a fibroblast-like morphology and increased invasiveness (Thomson et al., 2010). A full EMT, defined as a complete transition from epithelial cell surface markers, such as E-cadherin and Occludin, to mesenchymal markers, such as Vimentin and N-cadherin, often requires ten days or more (Cao, Xu, Liu, Wan, & Lai, 2015; Rai & Ahmed, 2014).
After EMT has transpired, the newly formed mesenchymal cells enter the bloodstream through the process of intravasation (invasion of cancer cells through the basement membrane) and disseminate throughout the body, eventually invading a distant organ site, through the process of extravasation (cancer cells exit the capillaries and invade an organ; Fig 1). At this distant organ site, the cancer cells undergo a Mesenchymal-Epithelial Transition (MET), the reverse process of EMT, in order to colonize and form a secondary epithelial tumor. By reverting to their original epithelial phenotype, the cells regain the ability to proliferate rapidly and form cell-cell junctions, two characteristics that are necessary for successful colonization. These characteristics were lost during the original EMT, when the cells acquired the ability to metastasize (Kalluri & Weinberg, 2009). After MET has occurred and a second epithelial tumor is established, the process of metastasis is complete. Thus, EMT is considered an important process during the early stages of metastasis, while MET is considered an important process during the later stages of metastasis (Brabletz, 2012).
EMT induction primarily occurs through the activation of phosphorylation cascades by various cytokines and growth factors present within the tumor microenvironment, including Transforming Growth Factor Beta (TGFß), Hepatocyte Growth Factor (HGF), Epidermal Growth Factor (EGF), and Platelet Derived Growth Factor (PDGF), among others. As the cascades are activated, the expression of E-cadherin, an important cell-adhesion molecule which maintains the epithelial state, is repressed, and EMT ensues (Cao, Xu, Liu, Wan, & Lai, 2015). This repression is facilitated by the transcriptional activity of four EMT-initiating transcription factors: Snail, Slug, Zeb, and Twist. Snail, Slug, and Zeb repress E-cadherin through direct binding to the promoter region of CDH1, the gene which codes for E-cadherin. Twist, on the other hand, functions in a different mechanism and represses E-cadherin expression through association with other proteins, most notably SET8, which binds to the promoter region of CDH1 and promotes the transcription of N-cadherin through activation of CDH2, the gene which codes for N-cadherin (Lamouille, Xu & Derynck, 2014; Yang et al., 2011).
Targeting EMT-Mediated Metastasis
Four strategies have been proposed to target EMT-mediated metastasis (Fig 2). Current research has predominantly focused on targeting EMT-induction through the inhibition of EMT-initiating signals, such as TGFß and EGF, by targeting their respective receptors (Buonato & Lazzara, 2013; Halder, Beauchamp & Datta, 2005; Wendt & Schiemann, 2009). In a clinical setting, individual targeting of pathways which initiate EMT and MET is implausible due to the significant toxicity of such a treatment. Yet, there is currently no effective method to eliminate EMT-derived mesenchymal cells. They are resistant to both chemotherapy and radiation treatments, which allows them to drive metastasis and cause tumor reoccurrence even after standard treatment has been administered (Bosco, Kenworthy, Zorio & Sang, 2015; Creighton et al., 2009; Gupta et al., 2009; Thomson, 2005).
Objective: Targeting Mesenchymal Survival Pathways
Because of the important role EMT-derived mesenchymal cells play in driving the metastatic process and their resistance to standard treatment, the overarching goal was to develop an effective strategy to eliminate EMT-derived mesenchymal cells. A novel approach was taken by identifying survival pathways within the mesenchymal state and testing the effects of several small molecule inhibitors on cell viability of differential epithelial and mesenchymal populations in vitro.
Small molecule inhibitors were selected based on their potency towards their intended target. Inhibitors that demonstrated a lower IC50 (half maximal inhibitory concentration), which indicates the amount of drug necessary to inhibit the intended biological component by half, were considered more potent and were chosen for use in this study. These inhibitors include GDC-0994, MK-2206, AZD1480, BAY 87-2243. GDC-0994 and AZD1480 inhibit the Extracellular Signal–Regulated Kinase (ERK1/2) and the Janus-Kinase 1/2 (JAK1/2), respectively, in an ATP-competitive manner (Ioannidis et al., 2011; Robarge et al., 2014). MK-2206 acts as an allosteric inhibitor of Protein Kinase B (AKT). It is equally potent against both AKT1 and AKT2 but approximately five-fold less potent against AKT3 (Hirai et al., 2010). Meanwhile, BAY 87-224 targets Hypoxia Inducible Factor 1 alpha (HIF1A) through inhibition of mitochondrial complex I (Ellinghaus et al., 2013).
Materials and Methods
Cytoscape computes survival pathways of the mesenchymal context by visualizing, modeling, and analyzing genetic interaction networks (Cline et al., 2007). Using information from the CytoKegg plugin in Cytoscape and the Kegg Pathway Database, survival pathways were mapped in Cell Designer. These pathways were color-coded based on RNA sequence data previously generated by Dr. John Haley of the Stony Brook Proteomics Center. Nodes which were up-regulated in a two-fold manner, at minimum, were selected as targets for drug inhibition.
A549, HCC4006, and HCC827 adenocarcinoma cell lines were purchased from the ATCC; two plates of cells were generated per cell line (+/- TGFß). 5ng/ml TGFß was added to each TGFß+ plate in order to induce EMT. In preparation for future assays, a target concentration of 5×103 cells/100µl was achieved for each plate. Cells were maintained in RPMI Medium 1640 (1x) in a 37°C/5% CO2 environment; Glutamine, Pyruvate, Penicillin Streptomycin (Pen Strep), HEPES, and Fetal Bovine Serum (FBS) were added to the RPMI media. Cell counting was performed every two days using a Hausser Scientific hemocytometer.
Verification of EMT with Light Microscopy
After a ten-day period, images of all cell lines (+/- TGFß) were taken at 60x using a Nikon Eclipse TS100 light microscope.
Verification of EMT with Immunofluorescence
A549 cells (+/- TGFß) were plated on Lab-Tek® II Chamber Slides. Preparation of slides for immunofluorescence was performed following the protocol included in the Life Technologies Image-iT® Fixation/Permeabilization Kit. Mouse Anti E-cadherin (Santa Cruz Biotechnology Catalog #21791) and Mouse Anti-Vimentin (BD Biosciences Catalog #550513) were used as primary antibodies. Alexa Fluor 488 goat anti-mouse (Molecular Probes by Life Technologies Ref. A11001), Alexa Fluor 555 goat anti-mouse (Molecular Probes by Life Technologies Ref. A21422), and Vimentin (D21H3) XP® Rabbit mAb (Alexa Fluor® 555 Conjugate; Cell Signaling Technology #9855) were used as secondary antibodies. All cells were stained with DAPI (Molecular Probes by Life Technologies Ref. D3571). After antibody incubation, ProLong Diamond Anti-fade Mountant (Molecular Probes by Life Technologies Ref. P36965) was applied to prolong fluorescence signal after initial exposure to light. Fluorescence images were captured at 200x and 630x with a Zeiss Fluorescence Microscope and quantified with Image J (Schneider, Rasband, & Eliceiri, 2012).
Verification of EMT with Western Blot
Cell lysates were created from all cell lines (+/- TGFß) using ThermoScientific Pierce RIPA Buffer with protease inhibitors. 3×106 cells were lysed at the conclusion of the ten-day period in which EMT occurred. To prepare loading samples, 25µl of NuPAGE® LDS Sample Buffer (4x) was added to each lysate. ThermoScientific NuPAGE® Bis-Tris Precast Gels were used. Along with the loading samples, 25µl of Invitrogen Novel Sharp Pre-stained Protein Standards was loaded into one well, acting as a protein ladder. After gel electrophoresis, samples were transferred to nitrocellulose membrane using a LKB Electrophoretic Transfer Kit and NuPAGE® Transfer Buffer (2x). The membrane was blocked with 5% milk with TBST. Mouse Anti E-cadherin (Santa Cruz Biotechnology Catalog #21791) and Mouse Anti-Vimentin (BD Biosciences Catalog #550513) were used as primary antibodies. Mouse Anti E-cadherin was diluted 1:500, while Mouse Anti-Vimentin was diluted 1:100. Peroxidase Conjugated Goat Anti-Mouse (ThermoScientific Catalog #32430) was used as the secondary antibody; it was diluted 1:100. After multiple washes with TBST, Enhanced Chemiluminescence (ECL) was performed, with Super Signal Pico (ThermoScientific) acting as the ECL (Mruk & Cheng, 2011). Quantification was performed with Image J.
Preparation of Drugs
Small molecule inhibitors (GDC-0994, MK-2206 2HCl, AZD1480, BAY 87-2243, Bl2536) were purchased from Selleck Chemicals (Houston, TX). Experimental target concentrations for GDC-0994 (ERK1/2 inhibitor), MK-2206 2HCl (Pan-AKT inhibitor), AZD1480 (JAK1/2 Inhibitor), and BAY 87-224 (HIF1A Inhibitor) were 50nM, 0.5µM, and 5µM. The target concentration for Bl2536 (PLK inhibitor) was 1µM.
All cell lines (+/- TGFß) were plated in Costar White 96 Well Plates; 5×103 cells were plated in each well. Experimental groups were set up on each plate; all cell types were treated with the three distinct target concentrations of the experimental drugs. After addition of drugs, plates were left in a humidified incubator at 37°C and 5% CO2 for a 48-hour period. MTT assays were then performed to quantify cell death. Absorbance was measured using a Molecular Devices SpectraMax M2. Absorbance was measured in Optical Density (O.D.) units.
In each cell line, every experimental condition was performed in triplicate. A549 and HCC827 trials were performed twice. For each data set, average deviation was represented by positive y-error bars, which were graphed using Graph Pad Prism 6. Statistical Analysis was performed using a Student’s T-Test in Microsoft Excel. Significance was set at p<0.05.
Identification of Survival Pathways and Drug Targets
Using Cytoscape and Cell Designer, survival pathways within the mesenchymal state were identified and illustrated (Fig 3). These pathways were focused on due to the presence of multiple proteins which were specifically up-regulated during the transition from epithelial to mesenchymal state. Based on this map, targets for drug inhibition were selected. Because of their specific up-regulation, JAK1/2, AKT1/2/3, and ERK1/2 were chosen as targets. Additionally, HIF1A was selected due to its role in downstream Vascular Endothelial Growth Factor (VEGF) signaling, an important transcriptional regulator of angiogenesis (Karar & Maity, 2011).
Light Microscopy Analysis
The first test used to verify TGFß induces EMT was a qualitative analysis of cell characteristics using light microscopy. TGFß- cells remain colonized and display a significant proliferative capacity (Fig 4A, 4B, 4C). These characteristics are in line with characteristics expected of epithelial cells. Thus, based on the light microscopy analysis, we concluded that TGFß- cells remained in the epithelial state. Meanwhile, TGFß+ cells were no longer colonized and displayed an individualized, spindle-like morphology (Fig 4D, 4E, 4F). These characteristics are distinctive of cells in the mesenchymal state.
Analysis of Protein Expression with Immunofluorescence
In order to validate the results of the light microscopy, protein expression of TGFß- and TGFß+ cells were analyzed using immunofluorescence. TGFß- cells expressed over five times as much E-cadherin than Vimentin when quantified using fluorescent microscopy at both 200x and 630x magnification (Fig 5A, 5B, 5C). On the other hand, TGFß+ cells expressed over six times as much Vimentin as E-cadherin when quantified using fluorescent microscopy at both 200x and 630x magnification (Fig 6A, 6B, 6C).
Because the TGFß- cells demonstrate a high E-cadherin expression and low Vimentin expression, we concluded that the cells were in the epithelial state. Conversely, the TGFß+ cells demonstrate a high Vimentin expression with low E-cadherin expression and were thus in the mesenchymal state.
Analysis of Protein Expression with Western Blot
Finally, EMT was verified by Western blot. Unlike the immunofluorescence analysis in which the expression of two distinct protein markers was assessed, only expression of E-cadherin was examined in this test. Because cancer cells experience reduced E-cadherin expression as they undergo EMT and transition from an epithelial to mesenchymal state, this analysis tested for a decrease in E-cadherin expression between the TGFß- and TGFß+ cells of the same cell line. Because lanes 1 and 3 demonstrate a more expression of E-cadherin than lanes 2 and 4 (35% vs. 20% and 17% vs. 5%, respectively), we concluded that EMT occurred (Fig 7). Thus, the TGFß- cells expressing a higher level of E-cadherin were in the epithelial state, while the TGFß+ cells expressing a lower level of E-cadherin were in the mesenchymal state.
Cell Viability Assays
Cell viability assays were performed on three lung adenocarcinoma cell lines after treatment: A549, HCC827, and HCC4006. In each cell line, the effect of BAY 87-2243 (HIF1A inhibitor), MK-2206 (AKT inhibitor), AZD1480 (JAK1/2 inhibitor), and GDC-0944 (ERK1/2 inhibitor) at 50nM, 0.5µM, and 5µM, respectively, on cell viability were tested on differential epithelial (TGFß-) and mesenchymal (TGFß+) populations. DMSO and BI2536, a PLK inhibitor that is a known suppressor of tumor growth, served as control groups (Steegmaier et al., 2007). Drug effectiveness was determined by comparing cell viability in mesenchymal populations treated with the drug and cell viability in the DMSO control. The drug which demonstrated a consistent decrease in viability of the drug-treated mesenchymal population compared to the DMSO control was considered the most effective drug in eliminating EMT-derived mesenchymal cells.
In the A549 assay, AZD1480 effectively decreased cell viability among the mesenchymal populations, with greatest effect achieved at 5µM (Fig 8C). Other inhibitors had limited effectiveness in decreasing cell viability (Fig 8A, 8B, 8D). Although the results weren’t significant for most of the HCC827 and HCC4006 assays, a similar trend was observed, with AZD1480 continuing to serve as the most effective inhibitor (Fig 9C, 10C).
To date, a very limited amount of research has been conducted that focuses on direct elimination of EMT-derived mesenchymal cells. This study used the novel approach of targeting specific survival pathways within the mesenchymal state as a strategy to eliminate EMT-derived mesenchymal cells. Results demonstrate that AZD1480, a potent ATP-competitive inhibitor of JAK1/2, induced the greatest decrease in cell viability within mesenchymal populations among inhibitors tested. This suggests JAK1/2 is a key modulator of the mesenchymal state and a plausible target of inhibition within EMT-derived mesenchymal cells. JAK1/2 plays an important role in regulating the activation of EMT-initiating transcription factors.
Because of the regulation by JAK1/2, the cell is maintained in a distinctly mesenchymal state as long as JAK1/2 is constitutively expressed, making JAK1/2 a promising drug target. Targeting the JAK-STAT pathway is feasible because of its relative simplicity compared to other pathways (Harrison, 2012). Additionally, JAK1/2 is a preferable target for inhibition because the phosphorylation of STAT3 is one of its primary functions, thus making it less likely that its inhibition will have significant effects on non-malignant cells (Levy & Lee, 2012).
As an ATP-competitive inhibitor, AZD1480 inhibits the ATP-binding pocket of the kinase domain, thus preventing the phosphorylation of STAT3 by JAK1/2 (Hedvat et al., 2009). Consequently, STAT3’s constitutive activation of EMT-transcription factors is prevented. This is the hypothesized mechanism by which AZD1480 eliminates EMT-derived mesenchymal cells (Fig 11).
In the future, continued study of AZD1480’s effectiveness as an anti-metastatic drug is warranted. This includes testing AZD1480’s effect on JAK protein expression at 48, 72, and 96 hours post treatment to ensure there is no protein rebound at a particular time point. Further studies include testing the effectiveness of AZD1480 in eliminating mesenchymal cells derived from other epithelial cell lines (e.g., breast, renal, colon) in vitro in order to potentially broaden the applications. Additionally, studies can be conducted in immune-deficient mice to test the in vivo efficacy of AZD1480.
Besides metastasis, EMT-derived mesenchymal cells have been implicated in numerous other malignancies. For example, because these cells are innately resistant to traditional therapeutics, they remain after the initial rounds of therapy and cause cancer reoccurrence in the future (Guffanti F, 2014). Furthermore, mesenchymal cells have been shown to not only be resistant to therapy themselves, but also to induce chemo-resistance within the surrounding cells. Thus, this strategy acts not only as an anti-metastatic treatment method, but also as a preventative measure for cancer reoccurrence and as a method to combat chemo-resistance. With this advancement, it may become possible for a patient to adhere to a low toxicity, yet effective treatment method. Such a combination has eluded science thus far.
The author would like to thank Dr. John Haley and Dr. Serena McCalla for their guidance and support in making this research possible.
Association, A. L. (2016). Lung Cancer Fact Sheet. Retrieved March 1, 2016, from http://www.lung.org/lung-disease/lung-cancer/resources/facts-figures/lung-cancer-fact-sheet.html
Bosco, D. B., Kenworthy, R., Zorio, D. A. R., & Sang, Q.-X. A. (2015). Human Mesenchymal Stem Cells Are Resistant to Paclitaxel by Adopting a Non-Proliferative Fibroblastic State. PLOS ONE, 10(6), e0128511. http://doi.org/10.1371/journal.pone.0128511
Brabletz, T. (2012). EMT and MET in Metastasis: Where Are the Cancer Stem Cells? Cancer Cell, 22(6), 699–701. http://doi.org/10.1016/j.ccr.2012.11.009
Buonato, J. M., & Lazzara, M. J. (2014). ERK1/2 Blockade Prevents Epithelial-Mesenchymal Transition in Lung Cancer Cells and Promotes Their Sensitivity to EGFR Inhibition. Cancer Research, 74(1), 309–319. http://doi.org/10.1158/0008-5472.CAN-12-4721
Cao, H., Xu, E., Liu, H., Wan, L., & Lai, M. (2015). Epithelial-mesenchymal transition in colorectal cancer metastasis: A system review. Pathology Research and Practice, 211(8), 557–569. http://doi.org/10.1016/j.prp.2015.05.010
Cheng, G. Z., Zhang, W., Sun, M., Wang, Q., Coppola, D., Mansour, M., … Wang, L.-H. (2008). Twist Is Transcriptionally Induced by Activation of STAT3 and Mediates STAT3 Oncogenic Function. Journal of Biological Chemistry, 283(21), 14665–14673. http://doi.org/10.1074/jbc.M707429200
Cline, M. S., Smoot, M., Cerami, E., Kuchinsky, A., Landys, N., Workman, C., … Bader, G. D. (2007). Integration of biological networks and gene expression data using Cytoscape. Nature Protocols, 2(10), 2366–2382. http://doi.org/10.1038/nprot.2007.324
Creighton, C. J., Li, X., Landis, M., Dixon, J. M., Neumeister, V. M., Sjolund, A., … Chang, J. C. (2009). Residual breast cancers after conventional therapy display mesenchymal as well as tumor-initiating features. Proceedings of the National Academy of Sciences, 106(33), 13820–13825. http://doi.org/10.1073/pnas.0905718106
Derenzini, E., Lemoine, M., Buglio, D., Katayama, H., Ji, Y., Davis, R. E., … Younes, A. (2011). The JAK inhibitor AZD1480 regulates proliferation and immunity in Hodgkin lymphoma. Blood Cancer Journal, 1(12), e46. http://doi.org/10.1038/bcj.2011.46
Ellinghaus, P., Heisler, I., Unterschemmann, K., Haerter, M., Beck, H., Greschat, S., … Ziegelbauer, K. (2013). BAY 87-2243, a highly potent and selective inhibitor of hypoxia-induced gene activation has antitumor activities by inhibition of mitochondrial complex I. Cancer Medicine, n/a–n/a. http://doi.org/10.1002/cam4.112
Guffanti F, D. G. (2014). Ovarian Cancer Recurrence: Role of Ovarian Stem Cells and Epithelial-to-Mesenchymal Transition. Journal of Cancer Science & Therapy, 06(08). http://doi.org/10.4172/1948-5956.1000284
Gupta, P. B., Onder, T. T., Jiang, G., Tao, K., Kuperwasser, C., Weinberg, R. A., & Lander, E. S. (2009). Identification of Selective Inhibitors of Cancer Stem Cells by High-Throughput Screening. Cell, 138(4), 645–659. http://doi.org/10.1016/j.cell.2009.06.034
Halder, S. K., Beauchamp, R. D., & Datta, P. K. (2005). A Specific Inhibitor of TGF-β Receptor Kinase, SB-431542, as a Potent Antitumor Agent for Human Cancers. Neoplasia, 7(5), 509–521. http://doi.org/10.1593/neo.04640
Harrison, D. A. (2012). The JAK/STAT Pathway. Cold Spring Harbor Perspectives in Biology, 4(3), a011205–a011205. http://doi.org/10.1101/cshperspect.a011205
Hedvat, M., Huszar, D., Herrmann, A., Gozgit, J. M., Schroeder, A., Sheehy, A., … Zinda, M. (2009). The JAK2 Inhibitor AZD1480 Potently Blocks Stat3 Signaling and Oncogenesis in Solid Tumors. Cancer Cell, 16(6), 487–497. http://doi.org/10.1016/j.ccr.2009.10.015
Hirai, H., Sootome, H., Nakatsuru, Y., Miyama, K., Taguchi, S., Tsujioka, K., … Kotani, H. (2010). MK-2206, an Allosteric Akt Inhibitor, Enhances Antitumor Efficacy by Standard Chemotherapeutic Agents or Molecular Targeted Drugs In vitro and In vivo. Molecular Cancer Therapeutics, 9(7), 1956–1967. http://doi.org/10.1158/1535-7163.MCT-09-1012
Ioannidis, S., Lamb, M. L., Wang, T., Almeida, L., Block, M. H., Davies, A. M., … Zinda, M. (2011). Discovery of 5-Chloro- N 2 -[(1 S )-1-(5-fluoropyrimidin-2-yl)ethyl]- N 4 -(5-methyl-1 H -pyrazol-3-yl)pyrimidine-2,4-diamine (AZD1480) as a Novel Inhibitor of the Jak/Stat Pathway. Journal of Medicinal Chemistry, 54(1), 262–276. http://doi.org/10.1021/jm1011319
Kalluri, R., & Weinberg, R. A. (2009). The basics of epithelial-mesenchymal transition. Journal of Clinical Investigation, 119(6), 1420–1428. http://doi.org/10.1172/JCI39104
Karar, J., & Maity, A. (2011). PI3K/AKT/mTOR Pathway in Angiogenesis. Frontiers in Molecular Neuroscience, 4. http://doi.org/10.3389/fnmol.2011.00051
Lamouille, S., Xu, J., & Derynck, R. (2014). Molecular mechanisms of epithelial–mesenchymal transition. Nature Reviews Molecular Cell Biology, 15(3), 178–196. http://doi.org/10.1038/nrm3758
Levy, D. E., & Lee, C. (2002). What does Stat3 do? Journal of Clinical Investigation, 109(9), 1143–1148. http://doi.org/10.1172/JCI15650
Mruk, D. D., & Cheng, C. Y. (2011). Enhanced chemiluminescence (ECL) for routine immunoblotting. Spermatogenesis, 1(2), 121–122. http://doi.org/10.4161/spmg.1.2.16606
Provost, J. (2007). MTT Proliferation Protocol. Retrieved from http://web.mnstate.edu/provost/mtt proliferation assay protocol.pdf
Rai, H., & Ahmed, J. (2014). N-cadherin: a marker of epithelial to mesenchymal transition in tumor progression. The Internet Journal of Oncology, 10(1), 1–12.
Ray, M. R., & Jablons, D. M. (2009). Hallmarks of metastasis. Lung Cancer Metastasis: Novel Biological Mechanisms and Impact on Clinical Practice, 29–46. http://doi.org/10.1007/978-1-4419-0772-1_2
Robarge, K., Schwarz, J., Blake, J., Burkard, M., Chan, J., Chen, H., … Moffat, J. G. (2014). Abstract DDT02-03: Discovery of GDC-0994, a potent and selective ERK1/2 inhibitor in early clinical development. Cancer Research, 74(19 Supplement), DDT02–03–DDT02–03. http://doi.org/10.1158/1538-7445.AM2014-DDT02-03
Saitoh, M., Endo, K., Furuya, S., Minami, M., Fukasawa, A., Imamura, T., & Miyazawa, K. (2016). STAT3 integrates cooperative Ras and TGF-β signals that induce Snail expression. Oncogene, 35(8), 1049–1057. http://doi.org/10.1038/onc.2015.161
Steegmaier, M., Hoffmann, M., Baum, A., Lénárt, P., Petronczki, M., Krššák, M., … Rettig, W. J. (2007). BI 2536, a Potent and Selective Inhibitor of Polo-like Kinase 1, Inhibits Tumor Growth In Vivo. Current Biology, 17(4), 316–322. http://doi.org/10.1016/j.cub.2006.12.037
Thomson, S. (2005). Epithelial to Mesenchymal Transition Is a Determinant of Sensitivity of Non-Small-Cell Lung Carcinoma Cell Lines and Xenografts to Epidermal Growth Factor Receptor Inhibition. Cancer Research, 65(20), 9455–9462. http://doi.org/10.1158/0008-5472.CAN-05-1058
Thomson, S., Petti, F., Sujka-Kwok, I., Mercado, P., Bean, J., Monaghan, M., … Haley, J. D. (2010). A systems view of epithelial-mesenchymal transition signaling states. Clinical and Experimental Metastasis, 28(2), 137–155. http://doi.org/10.1007/s10585-010-9367-3
Uddin, N., Kim, R.-K., Yoo, K.-C., Kim, Y.-H., Cui, Y.-H., Kim, I.-G., … Lee, S.-J. (2015). Persistent activation of STAT3 by PIM2-driven positive feedback loop for epithelial-mesenchymal transition in breast cancer. Cancer Science, 106(6), 718–725. http://doi.org/10.1111/cas.12668
Wendt, M. K., Balanis, N., Carlin, C. R., & Schiemann, W. P. (2014). STAT3 and epithelial–mesenchymal transitions in carcinomas. JAK-STAT, 3(2), e28975. http://doi.org/10.4161/jkst.28975
Wendt, M. K., & Schiemann, W. P. (2009). Therapeutic targeting of the focal adhesion complex prevents oncogenic TGF-β signaling and metastasis. Breast Cancer Research, 11(5), R68. http://doi.org/10.1186/bcr2360
Yang, F., Sun, L., Li, Q., Han, X., Lei, L., Zhang, H., & Shang, Y. (2012). SET8 promotes epithelial-mesenchymal transition and confers TWIST dual transcriptional activities. The EMBO Journal, 31(1), 110–123. http://doi.org/10.1038/emboj.2011.364
Effects of High Fructose/Glucose on Nlrp3/Il1β Inflammatory Pathway
The Nod-Like Receptor Protein-3/Interleukin-1β (NLRP3/IL-1β) inflammatory pathway activation is associated with autoimmune diseases including gout, Muckle-Wells syndrome, familial Mediterranean fever and multiple sclerosis. Type 2 diabetes and insulin resistance have been associated with chronic inflammation; however, the mechanism by which the NLRP3 inflammatory pathway participates in this condition remains unclear. Past research shows that fructose could induce production of reactive oxygen species (ROS), which are involved in NLRP3 complex assembly and secretion of IL-1β. In this study, the activation of NLRP3 inflammatory pathway was investigated in mouse macrophage cells (J774A1) treated with fructose and glucose (25, 50 and 100mM) for 24 hours. The cell lysates were analyzed using western blot analysis for inflammatory protein components NLRP3, IL-1β, and caspase 1, as well as antioxidant enzymes including peroxiredoxins (PRXSO3, PRX1), superoxide dismutase (SOD) and catalase. Mitochondria-derived ROS and mitochondrial permeability were assessed by MitoSOX red staining and JC-1 assay respectively. Results suggest that the intracellular levels of pro-IL-1β, antioxidant proteins (PRXSO3, SOD) and ROS were more elevated in cells treated with fructose and glucose at 25, 50 and 100mM when compared to untreated cells. Although pro-IL-1β accumulation was observed, no extracellular IL-1β could be detected using enzyme-linked immunosorbent assay (ELISA).
Western culture has adopted a diet rich in energy-loaded carbohydrates. This increased consumption of high-energy foods has been accompanied by reliance on mechanical technology to do work, reducing necessary physical activity (Popkin, 2001). The ratio of energy consumed to energy spent is imbalanced in favor of consumption, which results in storage of fat cells as adipose tissue and uncontrolled deposition of fats could lead to an individual carrying an excess amount of weight, referred to as being overweight or obese. This condition can be defined using the body mass index (BMI) of an individual (Finucane et al, 2011). Higher BMIs correspond to excess weight and obesity. Studies (Finucane et al, 2011) show that the mean BMI worldwide has increased over the years and so has the rate of obesity. In 2008, over 500 million people worldwide were considered obese and about 1.46 billion were overweight (Finucane et al, 2011).
In obese individuals, enlarged fat cells secrete fatty acids and cytokine factors such as tumor necrosis factor-α (TNF-α), that are capable of causing a wide range of downstream effects such as having higher risks for a variety of diseases including coronary artery disease, hypertension, metabolic syndrome, gall bladder disease, cancer, osteoarthritis and type 2 diabetes (Rodriguez-Hernandez, Simental-Mendia, Rodriguez-Ramirez, & Reyes-Romero, 2013). Among the obesity related diseases, type 2 diabetes has recently been classified an autoimmune disease involving inflammation through NLRP3 activation (Bray, 2004; Gunter & Leitzmann, 2006; Hajer, Haeften, & Visseren 2008). Studies have investigated the relationship between type 2 diabetes, insulin resistance and IL-1β expression (Gao et al, 2014; Larsen et al, 2007). Larsen et al.’s experimental results (2007) showed that blockade of IL-1β expression in patients with type 2 diabetes improved β-cell function and promoted glycemic control while Goa et al.’s results (2014) showed that IL-1β presence in human adipocytes significantly reduced the gene expression of insulin signaling molecules and its absence improved insulin sensitivity. The secretion of IL-1β is regulated by the Nod-Like Receptor protein 3 (NLRP3) inflammasome. IL-1β secretion is carried out in two steps. The first signal, also known as the priming step, consists of activation of Nod-Like Receptor protein 3 (NLRP3) coupled with accumulation of pro- IL-1β – the inactivated precursor protein for IL-1β. Upon accumulation of the precursor, a second signal is needed to recruit the NLRP3 inflammasome complex, consisting of (NLRP3), adaptor protein apoptosis speck-like Protein (ASC) and activated caspase 1, consequently responsible for cleavage of pro-IL-1β to secretion as Il-1β (Figure 1). When NLRP3/IL-1β pathway is activated, ROS production is also observed (Jo, Kim, Shin, & Sasakawa et al, 2016). An article (Tschopp & Schroder, 2010) suggested that mitochondrial ROS is not only correlated with NLRP3 activation, but is involved in assembling the NLRP3 inflammasome complex.
Mitochondria are considered the main source of ROS in normal and altered metabolism, and are the site of aerobic carbohydrate metabolism (Lenaz, 2001). After consumption, carbohydrates are broken down to sugars, which are metabolized to yield acetyl-CoA, the primary substrate of the tricarboxylic acid cycle (TCA) taking place in the mitochondria. During the TCA cycle, Acetyl-CoA goes though series of oxidation steps carried out by co-enzymes NAD+ and FAD, which are reduced to NADH and FADH2 respectively. These energy rich hydrogen atoms are supplied to the electron transport chain (ETC), where they are re-oxidized simultaneously with the reduction of oxygen to water by complexes on the mitochondrial membrane, yielding ATP (adenosine triphosphate), the primary form of energy in cells. This process results in 1-2% of consumed oxygen being partially reduced to generate superoxide anion (O2●-) (Thannickal & Fanburgh, 2000). Antioxidants usually act collectively to suppress ROS and other free radicals by reducing them to a less reactive species as in the multistep reduction of superoxide to water. Superoxide dismutase (SOD) reduces O2●- to hydrogen peroxide (H2O2), which is further reduced by catalase (CAT) or Glutathione peroxidase (GPX) to water (Devasagayam et al., 2004). Peroxiredoxins (PRX) are able to directly reduce H2O2 to water, while GPX works to reduce H2O2 by using Glutathione (GSH), a thiol-based antioxidant, as a substrate. In an adverse physiological state referred to as oxidative stress, the body is overwhelmed with production of ROS resulting in an imbalance of ROS/antioxidant ratio (Nordberg & Arner 2001). Oxidative stress has been associated with several pathological conditions such as cancer, diabetes (Ceriello & Motz, 2004) and chronic inflammation as well as other autoimmune diseases (Reuter et al, 2010). Among the inflammatory pathways that are associated with chronic inflammation and autoimmune diseases is the NLRP3 Inflammasome/IL-1β Inflammatory pathway (Jo et al, 2016).
ROS direct involvement with NLRP3 inflammasome activation has not been completely elucidated. The purpose of this study was to investigate whether or not ROS production is proportional to the sugar concentration in macrophage cell culture and if this higher input of glucose and fructose would contribute to an inflammatory response through the NLRP3 inflammasome. Our findings will give insight on the impact of a sugar-rich diet on oxidative and inflammatory pathways, as well as provide a perspective on macrophage cellular processes involved with obesity-associated inflammation for other scientists and researchers to expand on.
Materials and Methods
Chemicals and Reagents
3, 3’, 5, 5’-tetramethylbenzidine (TMB) reagent, Antimycin A (AA) and β-actin antibody were obtained from Sigma. Mitochondrial superoxide indicator (MitoSOX red, was purchased from Molecular Probes® (InvitrogenTM). NLRP3 antibody, caspase 1 antibody and IL-1β antibody was obtained from Santa Cruz. PRX1 and PRXSO3 antibodies were obtained from Abcam. The secondary IgG peroxidase-linked mouse antibody, and rabbit antibodies were purchased from GE Healthcare. All reagents were of analytical grade.
Cell Culture and Experimental Conditions
Mouse monocyte (macrophage) cell line J774A1 was obtained from American type culture collection (ATCC). Cells were seeded at 1 x 105 cells/mL using RPMI 1640 medium supplemented with 10% FBS, penicillin, streptomycin and L-glutamine, and incubated at 37°C in a 5% CO2-supplemented atmosphere for at least 24 hours before treatments. The cells were treated with glucose or fructose at concentrations 25mM, 50mM and 100mM for 24 hours. Treatments were performed in triplicates.
Cell Harvesting and Lysates Preparation
After respective treatments, tissue culture plates were placed on ice and the attached cells were rinsed once with cold phosphate buffered saline (PBS) and lysed using total lysis buffer (50mM Tris, pH 7.4, 150mM Sodium Chloride (NaCl), 2mM ethylenediaminetetraacetic acid (EDTA), 0.2% TritonTM X-100, 0.3% IGEPAL®, and protease inhibitor cocktail). Lysates were removed from the plate, transferred to microcentrifuge tubes, and immediately frozen in liquid nitrogen to prevent protein degradation and enhance cell lysis.
Protein Concentration Determination: Bradford assay
The J774A1 lysates were thawed at room temperature. 10µL of bovine serum albumin (BSA) standards or cell lysates samples were transferred into separate clean tubes and 1000µL of the 1:4 diluted dye (Bradford) was added to the respective tubes. The tubes were incubated for five minutes at room temperature before their content was transferred into a cuvette. Absorbance was measured at 595nm and the protein content in the samples was calculated using a standard curve with a series of BSA standards to 1500 – 100mg/ml. The amount of protein in each condition was determined and adjusted to 0.5mg/ml by adding PBS buffer.
Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) and Immunoblot
Leamli SDS buffer was added to the lysate and 30µL of the mixture per condition was loaded on to polyacrylamide gels. The proteins were separated on 10% SDS-PAGE and transferred from the gel to a nitrocellulose membrane. The membrane was blocked for one hour at RT with 3% milk in PBS containing 0.05% Tween®20 (PBST) and then incubated with respective primary antibodies overnight at 4oC on a rotating platform. The membranes were washed three times with PBST for five minutes each and incubated with respective horseradish peroxidase-conjugated secondary antibodies. Western blots were developed utilizing, 3, 3’, 5, 5’ TMB liquid substrate system for membranes according manufacturer’s instructions.
Detection of Mitochondria-Derived ROS
Mitochondria-derived ROS, was detected using the mitochondrial superoxide indicator, MitoSOX red, a cationic dihydroethidium modified to target the mitochondria. According to the manufacturer, MitoSOX red is a cell-permeable dye that reacts with ROS to form ethidium, which upon binding to nucleic acids gives a bright red fluorescence. Briefly, cells were seeded onto eight well CultureSlides (BD Falcon™) and treated as described above. At the end of the respective treatments, the cells were rinsed with PBS and loaded with MitoSOX red (2.5μM) for ten minutes. The medium containing fluorescent probes was removed, and the cells were rinsed with PBS, and observed in an Olympus BX53 Fluorescence Microscope coupled to an Olympus DP73 digital camera.
Detection of Mitochondrial Membrane Potential (∆ψm)
Cells (5 x 105 cells/mL) were seeded in 12-well plates and 24 hours later were treated as indicated. After washing with PBS, cells were incubated in fresh medium containing 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazol-carbocyanine iodide (JC-1) for 15min. The dye was then removed; and cells were washed with PBS and fresh medium was added. Then live cells were immediately observed in an Olympus BX53 Fluorescence Microscope coupled to Olympus DP73 digital camera. Healthy cells, mainly JC-1 aggregates were observed at 540/570nm excitation/emission and the apoptotic or unhealthy cells with mainly JC-1 monomers at 485/535nm excitation/emission.
As ROS are a by-product of fructose and glucose metabolism, it was anticipated that an increase in substrate would cause an increase in ROS production. This increase in ROS could promote an environment suitable for NLRP3 inflammasome activation. In the vicinity of ROS, MitoSOX red produces a bright red fluorescence, which varies in intensity with amount of ROS present. Antimycin A was used as the positive control as it is known to be an inhibitor of complex III in the ETC and leads to robust ROS production and mitochondrial damage. J774A1 cells appeared to produce more red fluorescence when treated with fructose and glucose in contrast to the control, untreated cells (Figure 2a). In addition, fructose treatment elicits a more intense fluorescence than glucose at concentrations 25 and 50mM, as demonstrated by fluorescence intensity analysis (Figure 2b).
Mitochondrial Membrane Potential (∆ψm)
When the mitochondrion is constantly hyperactive, it has the tendency to go into a dysfunctional state evidenced by leaks in the mitochondrial membrane. These leaks could cause it to become depolarized (Tsujimoto & Shimizu, 2007). The JC-1 dye kit can be used to detect the overall health of the mitochondria. In hyperpolarized mitochondria the dye emits a red fluorescence while in depolarized mitochondria it emits green. Hence, the red fluorescence indicates healthier mitochondria while green indicates a disturbed mitochondrial function. In cells treated with glucose (25mM) and fructose (25mM, 50mM), there is a higher red intensity compared with control, indicating that the mitochondrion is polarized. However, as treatment concentration is increased, there is a decrease in red and increase in green fluorescence in glucose treatment, indicating mitochondrial depolarization occurs. It appears that the decrease in red and increase in green was more intense in fructose treated cells than glucose treated cells at 100mM concentration (Figure 3a & b).
Western Blot analysis for Antioxidant and Inflammatory Proteins
Western blot analysis was used to detect the presence of antioxidant and inflammatory proteins. Comparing the intensity and thickness of the bands can give insight on the amounts of proteins present across experimental conditions. The Bradford assay was used to measure the total amounts of protein in each sample and ensure the total protein loaded on to gels were constant throughout the treatment conditions. Increased expression of antioxidant protein is a marker of oxidative stress. As a response to counteract ROS, antioxidant enzymes are produced through stress response mechanisms to prevent oxidative damage (Nguyen et al, 2009). Compared with the control experiment, PRX1 levels did not seem to vary throughout the conditions, however, as implied by the prominent bands observed, the amounts of its oxidized form (PRXSO3) and catalase appear to be increased, in particularly with fructose treatment. SOD1 expression showed to be slightly increased in both fructose and glucose treatments (Figure 4).
The presence and amount of precursor proteins of NLRP3 inflammatory pathway can be used to assess NLRP3 inflammasome activation and determine the extent of inflammatory response in macrophage cells. The precursor protein, Pro-IL-1β, appeared to be increased only in fructose and glucose-treated cells. In addition, it seems to be expressed in higher amounts in fructose treatments compared to glucose treatments. The cleaved products of caspase 1, seen as the multiple lower bands, are also observed majorly in fructose condition, indicating some level of caspase 1 activation. NLRP3 protein, however, seems to remain constant throughout the conditions (Figure 5).
Discussion and Conclusions
The results appear to be consistent with the presumption that sugars can induce oxidative stress through an over active mitochondria and promote an inflammatory response in J774A1 macrophage cells. The MitoSOX red assay possibly indicates an increased ROS production in treated cells. In lower concentrations (25mM, 50mM), fructose seems to invoke more ROS production than glucose, however in the highest concentration, fructose and glucose treatments produced comparable amounts of ROS, perhaps because the antioxidant systems were increased to compensate the potentially higher amounts of ROS.
The mitochondrial permeability assay suggests that the lower concentrations of fructose (25mM, 50mM) and glucose (25mM) treatments actually cause the cells to become relatively healthy compared to the control. However, as the treatment concentrations increased, the membrane was depolarized, indicating an unhealthy cell. One of the key events during apoptosis and necrosis, or cell death, is an increase in mitochondrial membrane permeability by opening of permeability transition pore proteins, which are ion channels. This leak in the membrane leads to a decrease in membrane potential of the mitochondria (Tsujimoto & Shimizu, 2007). Another study (Russell et al., 2002) also found a dose-dependent increase in ROS production and mitochondria depolarization in neurons treated with glucose. This study (Russell et al., 2002) observed that the mitochondrial membrane goes through hyperpolarization before it is depolarized suggesting that depolarization might be time dependent. Future studies could investigate this by measuring membrane potential at varied time points.
As ROS levels appeared to increase with treatment, the intracellular antioxidant protein amounts were investigated. It appeared that the amounts of some antioxidant proteins (Catalase and PRXSO3) were increased in treated cells, indicating an upregulation of antioxidant protein expression and implying an increased level of oxidative stress. We found that fructose had a more observable effect on this response. This could be due to the differences in fructose and glucose metabolism. According to Cha et al (2008) glucose metabolism is more regulated than that of fructose. In the glycolytic pathway, fructose is able to bypass the rate-limiting step catalyzed by phosphofructokinase, an important control enzyme for the rate of glucose metabolism, allowing fructose to be metabolized at a faster rate. In fructose fed mice, it was found that fructose metabolic pathway, and not glucose, was essential to recruitment of macrophages and production of pro-inflammatory mediators including TNF-α in the visceral adipose tissue (Marek et al, 2015).
Although the results indicate the precursor to IL-1β – pro-IL-1β, was accumulated intracellularly, extracellular IL-1β remained undetected. We speculate that this might be due to the duration of incubation. A time course experiment in our research group (Sunasee et al, 2015) showed that pro-IL-1β accumulation is also time dependent. In the experiment using cellulose nanocrystals (CNCs) cationic derivatives, seven hours of treatment was the peak time for pro-IL-1β detection in J774A1 cells. Treatments at other time points did not seem to show as much pro-IL-1β accumulation. In the future, we would vary the lengths of treatments and investigate if this is a factor affecting pro- IL-1β accumulation and IL-1β secretion in fructose and glucose treatments. We also would like to investigate the effect of adding antioxidants in the current treatments and possibly, including lipids in treatments to create a more type 2-diabetic/metabolic-syndrome environment.
Future studies should focus on quantifying the cell counts and analyzing the MitoSox and JC-1 fluorescence imaging data statistically. Overall, our results corroborate with the existing information regarding the effects of excess sugar on ROS production by the mitochondria, and the potential link with activation of NLRP3 inflammasome pathway. By understanding the behavior of this pathway, researchers will be able to better understand autoimmune diseases and possibly create new ideas on mitochondria-targeted therapies.
I am extremely grateful to my mentor, Dr. Karina Ckless, whose continuous guidance and encouragement made it possible to execute this study. In addition, I would like to thank the SUNY Plattsburgh Redcay Honors center for their approval and sponsorship of the project.
Bray, G. A. (2004). Medical Consequences of Obesity. The Journal of Clinical Endocrinology & Metabolism, 89(6), 2583–2589. http://doi.org/10.1210/jc.2004-0535
Ceriello, A., & Motz, E. (2004). Is Oxidative Stress the Pathogenic Mechanism Underlying Insulin Resistance, Diabetes, and Cardiovascular Disease? The Common Soil Hypothesis Revisited. Arteriosclerosis, Thrombosis, and Vascular Biology, 24(5), 816–823. http://doi.org/10.1161/01.ATV.0000122852.22604.78
Cha, S. H., Wolfgang, M., Tokutake, Y., Chohnan, S., & Lane, M. D. (2008). Differential effects of central fructose and glucose on hypothalamic malonyl–CoA and food intake. Proceedings of the National Academy of Sciences of the United States of America, 105(44), 16871–16875. http://doi.org/10.1073/pnas.0809255105
Devasagayam, T. P. A., Tilak, J. C., Boloor, K. K., Sane, K. S., Ghaskadbi, S. S., & Lele, R. D. (2004). Free radicals and antioxidants in human health: Current status and future prospects. Journal of Association of Physicians of India, 52(OCT), 794–804.
Finucane, M. M., Stevens, G. A., Cowan, M. J., Danaei, G., Lin, J. K., Paciorek, C. J., Ezzati, M. (2011). National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. The Lancet, 377(9765), 557–67.
Gao, D., Madi, M., Ding, C., Fok, M., Steele, T., Ford, C., … Bing, C. (2014). Interleukin-1β mediates macrophage-induced impairment of insulin signaling in human primary adipocytes. American Journal of Physiology – Endocrinology and Metabolism, 307(3), E289–E304. http://doi.org/10.1152/ajpendo.00430.2013
Gunter, M. J., & Leitzmann, M. F. (2006). Obesity and colorectal cancer: epidemiology, mechanisms and candidate genes. The Journal of Nutritional Biochemistry, 17(3), 145–156. http://doi.org/10.1016/j.jnutbio.2005.06.011
Hajer, G. R., Haeften, T. W. van, & Visseren, F. L. J. (2008). Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. European Heart Journal, 29(24), 2959–2971. http://doi.org/10.1093/eurheartj/ehn387
Jo, E.-K., Kim, J. K., Shin, D.-M., & Sasakawa, C. (2016). Molecular mechanisms regulating NLRP3 inflammasome activation. Cellular & Molecular Immunology, 13(2), 148–159. http://doi.org/10.1038/cmi.2015.95
Larsen, C. M., Faulenbach, M., Vaag, A., Vølund, A., Ehses, J. A., Seifert, B., Donath, M. Y. (2007). Interleukin-1-Receptor Antagonist in Type 2 Diabetes Mellitus. The New England Journal of Medicine, 356(15), 1517–26. http://doi.org/http://dx.doi.org/10.1056/NEJMoa065213
Lenaz, G. (2001). The Mitochondrial Production of Reactive Oxygen Species: Mechanisms and Implications in Human Pathology. IUBMB Life, 52(3-5), 159–164. http://doi.org/10.1080/15216540152845957
Liu, H., & Heaney, A. P. (2011). Refined fructose and cancer. Expert Opinion on Therapeutic Targets, 15(9), 1049–1059. http://doi.org/10.1517/14728222.2011.588208
Marek, G., Pannu, V., Shanmugham, P., Pancione, B., Mascia, D., Crosson, S., … Sautin, Y. Y. (2015). Adiponectin Resistance and Proinflammatory Changes in the Visceral Adipose Tissue Induced by Fructose Consumption via Ketohexokinase-Dependent Pathway. Diabetes, 64(2), 508–518. http://doi.org/10.2337/db14-0411
Nguyen, T., Nioi, P., & Pickett, C. B. (2009). The Nrf2-Antioxidant Response Element Signaling Pathway and Its Activation by Oxidative Stress. Journal of Biological Chemistry, 284(20), 13291–13295. http://doi.org/10.1074/jbc.R900010200
Nordberg, J., & Arnér, E. S. J. (2001). Reactive oxygen species, antioxidants, and the mammalian thioredoxin system1. Free Radical Biology and Medicine, 31(11), 1287–1312. http://doi.org/10.1016/S0891-5849(01)00724-9
Popkin, B. M. (2001). The Nutrition Transition and Obesity in the Developing World. The Journal of Nutrition, 131(3), 871S–873S.
Reuter, S., Gupta, S. C., Chaturvedi, M. M., & Aggarwal, B. B. (2010). Oxidative stress, inflammation, and cancer: How are they linked? Free Radical Biology and Medicine, 49(11), 1603–1616. http://doi.org/10.1016/j.freeradbiomed.2010.09.006
Rodriguez-Hernandez, H., Simental-Mendia, L. E., Ndez, H., Rodriguez-Ramirez, G., & Reyes-Romero, M. A. (2013). Obesity and Inflammation: Epidemiology, Risk Factors, and Markers of Inflammation. International Journal of Endocrinology, 2013, e678159. http://doi.org/10.1155/2013/678159
Russell, J. W., Golovoy, D., Vincent, A. M., Mahendru, P., Olzmann, J. A., Mentzer, A., & Feldman, E. L. (2002). High glucose-induced oxidative stress and mitochondrial dysfunction in neurons. The FASEB Journal, 16(13), 1738–1748. http://doi.org/10.1096/fj.01-1027com
Sunasee, R., Araoye, E., Pyram, D., Hemraz, U. D., Boluk, Y., & Ckless, K. (2015). Cellulose nanocrystal cationic derivative induces NLRP3 inflammasome-dependent IL-1β secretion associated with mitochondrial ROS production. Biochemistry and Biophysics Reports, 4, 1–9. http://doi.org/10.1016/j.bbrep.2015.08.008
Thannickal, V. J., & Fanburg, B. L. (2000). Reactive oxygen species in cell signaling. American Journal of Physiology – Lung Cellular and Molecular Physiology, 279(6), L1005–L1028.
Tschopp, J., & Schroder, K. (2010). NLRP3 inflammasome activation: The convergence of multiple signalling pathways on ROS production? Nature Reviews. Immunology, 10(3), 210–215. http://doi.org/10.1038/nri2725
Tsujimoto, Y., & Shimizu, S. (2007). Role of the mitochondrial membrane permeability transition in cell death. Apoptosis: An International Journal On Programmed Cell Death, 12(5), 835–840.
Access Full PDF: High Fructose/Glucose on Nlrp3/Il1β Inflammatory Pathway
Total Electron Content (TEC) Variations and Correlation with Seismic Activity over Japan
Earthquakes are extremely dangerous physical phenomena. The ability to properly forecast them would go a long way in reducing the damage they cause. One earthquake forecasting method being researched uses the ionospheric Total Electron Content (TEC). Our investigation used TEC data from 2011 during certain days near and on the date of the earthquake off the coast of Tōhoku, Japan. We took advantage of the large amounts of GPS records obtained by the GPS Earth Observation Network (GEONET) of Japan which contained the TEC data needed. This data was used to visualize the TEC over the course of the day of the Tōhoku earthquake. The video produced abnormalities consistent with the predicted effect an earthquake has on the ionosphere. These abnormalities were shown to not be caused by solar and geomagnetic activity. These results suggest that detectable ionospheric activity precedes earthquakes. Ionospheric disturbances are also known to be caused by other confounding factors such as solar and geomagnetic activity. Careful analysis is included in this paper to exclude this class of disturbances from those that are seen to occur due to seismic and pre-earthquake activities. The hope is that the potential correlation between seismic and pre-earthquake activities may be used as an earthquake precursor towards the development of an earthquake forecasting method.
Recently, research has focused on the relationship between the state of the ionosphere and seismic activity (Heki & Ping, 2005; Oyama, Kakinami, Liu, Abdu, & Cheng, 2011). One ionospheric parameter that has been investigated is the Total Electron Content (TEC). TEC is defined as the number of electrons along a path between a receiver (rx) on the surface of the earth and a GPS satellite (st) in orbit. The TEC can be computed with the line integral from a GPS satellite and a receiver as described in (1).
r = is range or radial position (meters)
Θ = latitude (degrees)
Φ = longitude (degrees)
t = time (seconds)
With regards to the correlations between TEC and pre-earthquake and seismic activities, the TEC is an important parameter of study because it has the potential for showing the changes in the ionosphere due to these activities. It is because seismic and pre-earthquake activities create stress in rocks in the earth’s crust. These stresses are known to positively charge the rocks on the earth’s crust. As the positive charges accumulate at the rocks outer surfaces, they create a difference in potential which in turn creates a flow of charges that can travel fast and far from their point of origin. As the charges travel upward under the influence of the electric field lines between the surface of earth and the bottom of the ionosphere, they reach the bottom of the ionosphere, disturbing the equilibrium of the electrons in the ionosphere (Freund, Takeuchi & Lau, 2006). These disturbances can be seen in the TEC which makes TEC a potential candidate as an earthquake precursor. If TEC disturbances could be used as an earthquake precursor, tracking those disturbances could be used as part of an earthquake forecasting system which would improve earthquake warning systems, in turn saving countless lives.
This study uses TEC data from Japan and current knowledge of the Tōhoku Japan earthquake to determine whether pre-earthquake and seismic activities correlate with TEC changes around the time of the earthquake.
TEC disturbances can be observed using GPS signals. It is because the ionosphere creates a phase delay in the electromagnetic signals, sent from a receiver on earth to a GPS satellite in orbit. The phase delays change based on several variables: the frequency of the emitted signals, the path from the receiver to the satellite and the associated electron density along the path. The phase delay can be used to estimate the distance (called pseudo-range) from the GPS satellite and the receiver. More specifically, since the ionosphere is a dispersive medium (i.e., varies with frequency), using two signals with known GPS frequencies, we can measure the phase delay between the returned signals to estimate the TEC as shown in (2).
Prs = The differential pseudo-range measurements
f1,f2 = The GPS measurement frequencies: 1,575.42 and 1,227.60 MHz
TEC = The Total Electron Content (m/s2)
DCB = The biases in the measurements
From these measurements, we can test if seismic activity causes detectable disturbances in the ionosphere’s TEC.
The visualization of TEC data over Japan required GPS readings from the GPS Earth Observation Network (GEONET) of Japan (http://datahouse1.gsi.go.jp/terras/terras_english.html). From these readings, TEC values were obtained using the relationship in (2), resulting in multiple data points for each receiver over time. The GPS data was contained in receiver independent exchange format (RINEX) observation data files. These files contain phase delay measurements for GPS signals, as well as the time at which these measurements were made. These files were processed to extract TEC data over time, which was done using a FORTRAN program. The FORTRAN program was originally developed by Professor Kosuke Heki of Hokkaido University in Japan. The TEC data over time was then input into MATLAB (The Mathworks Inc., Natick, MA, 2013) for analysis and visualization. We averaged the data points at each site and time, resulting in one data point for each site every 30 seconds (the sampling period of the GPS receivers). This data was interpolated using a linear triangulation method, resulting in one grid of interpolated data over Japan every 30 seconds. We turned these grids into contour plots, each plot creating a frame for the video file produced.
Since disturbances in the ionosphere are not necessarily caused by seismic activity, we considered other possible sources of disturbances in our analysis. We obtained measurements of 10.7 cm solar flux (F10.7), the sunspot number (SSN), and earth’s geomagnetic storm activity (Kp index) over the month of March in 2011. This data was provided by the National Oceanic and Atmospheric Administration’s Space Physics Interactive Data Resource (SPIDR) (http://spidr.ionosonde.net/spidr/). This data was compared to other research (Hasbi et al., 2009) (Ouzounov et al., 2011) to determine whether any ionospheric irregularities could have been caused by solar or geomagnetic sources.
The video of TEC levels produced for February 11th and March 11th show the behavior of the ionosphere throughout the day. From the video on March 11th, after an initial enhancement (Figure 1A), a sudden depletion of roughly 30 TECU in 15 minutes was observed from 1:35 PM JST to 1:50 PM JST (Figure 1B). Large fluctuations were also observed, with a significant example from 4:25 PM JST (Figure 1C) to 5:00 PM JST (Figure 1D) showing levels decreasing by 50 TECU in 45 minutes. For comparison, plots from February 11th (one month before the earthquake) were taken at the same times (Figure 2). Data from the same times of day show no enhancements, depletions, or fluctuations.
Sunspot number data for March 2011 shows a sunspot count of roughly 100 on March 8, coming down to 60 on March 11th (Figure 3A). Geomagnetic storm activity data for March 2011 shows a Kp of roughly 5.3 on March 11th, with relatively low numbers before March 11th (Figure 3B). Solar flux (F10.7) data for March 2011 shows heavy activity on March 7th with measurements of more than 900 watt per square meter-hertz, with relatively calm activity for the rest of the month (Figure 3C).
It has been proposed that TEC anomalies can be caused by seismic activity. Our study looked at GPS data during the 2011 Tōhoku Earthquake to investigate the correlation between TEC and seismic activity. The investigation resulted in evidence supporting detectable TEC disturbances caused by pre-seismic activity.
TEC measurements taken over the course of the earthquake showed significant TEC anomalies during seismic activity compared to a time period with no seismic activity. The TEC on the day of the earthquake shows an initial enhancement, followed by a rapid depletion and large fluctuations. Comparatively, TEC levels on a day with no seismic activity behave smoothly.
While these results provide strong evidence indicative of seismic activity causing TEC disturbances, it is necessary to make sure no other source caused these enhancements, depletions, and fluctuations. While the values are relatively high for the sunspot number (SSN) and magnetic storm activity (Kp) (Figure 3A-B) compared to expected levels, prior work has shown these values alone are not significant to cause a disturbance of this magnitude in the ionosphere (Hasbi et al., 2009). The solar flux peaked for the time frame analyzed on March 7th, four days before the earthquake (Figure 3C). Based on previous studies, the solar flux around the day of the earthquake is likely not the cause of the TEC disturbances on the day of the earthquake due to the timing of the maximum flux (Ouzounov et al., 2011).
This study had limitations with certain systemic sources of error in the measurements of TEC, as reflected in the DCB term shown in (2). The data analyzed did not take into account the angle through which the GPS signals passed through the ionosphere. This leads to the GPS signals passing through a larger section of the ionosphere when the satellite is not directly above the receiver. Another source of error that should be considered is certain biases introduced from the GPS satellites and receivers, as well as the troposphere. These biases are introduced due to many factors, such as satellite ephemeris inaccuracies, clock errors, as well as atmospheric conditions and measuring equipment temperature. These errors were out of the scope of this study, and should be looked into further.
It is highly probable that the earthquake caused the enhancements, depletions, and fluctuations in the TEC. However, more detailed work is needed to help determine the exact cause and nature of TEC events surrounding seismic activity. Future investigations should focus on tracking and monitoring TEC for longer periods of time surrounding an earthquake to better characterize the beginning and length of disturbances. In addition, further research should be performed to determine how solar flux, magnetic storms, or other external events affect TEC disturbances.
The pattern of TEC enhancements, depletions, and fluctuations expected by previous works (Heki & Ping, 2005) were noticed in TEC measurements near the time of the earthquake. These TEC behaviors are quite different from the average expected behaviors resulting from a smooth ionosphere. These TEC behaviors are also seen not to result from or be correlated with other outside phenomena such as solar flux and geomagnetic storms. Our results support the hypothesis that seismic activity increases disturbances in TEC. Future research should focus on developing methods to isolate the effects of seismic activities from confounding factors and to remove errors from TEC measurements.
Freund, F. T., Takeuchi, A., & Lau, B. W. (2006). Electric currents streaming out of stressed igneous rocks – A step towards understanding pre-earthquake low frequency EM emissions. Physics and Chemistry of the Earth, 31(4-9), 389-396. doi:10.1016/j.pce.2006.02.027
Garner, T., Gaussiran II, T., B.W., T., R.B., H., Calfas, R., & Gallagher, H. (2008). Total electron content measurements in ionospheric physics. Advances in Space Research, 42, 720-726. doi:10.1016/j.asr.2008.02.025
Hasbi, A. M., Momani, M. A., Mohd Ali, M. A., Misran, N., Shiokawa, K., Otsuka, Y., & Yumoto, K. (2009). Ionospheric and geomagnetic disturbances during the 2005 Sumatran earthquakes. Journal of Atmospheric and Solar-Terrestrial Physics, 71(17-18), 1992-2005. doi:10.1016/j.jastp.2009.09.004
Heki, K., & Ping, J. (2005). Directivity and apparent velocity of the coseismic ionospheric disturbances observed with a dense GPS array. Earth and Planetary Science Letters, 236, 845-855. doi:10.1016/j.epsl.2005.06.010
Ouzounov, D., Pulinets, S., Romanov, A., Romanov, A., Tsybulya, K., Davidenko, D., . . . Taylor, P. (2011). Atmosphere-Ionosphere Response to the M9 Tohoku Earthquake Revealed by Joined Satellite and Ground Observations. Preliminary results. Earthquake Science, 24(6), 557-564. doi:10.1007/s11589-011-0817-z
Oyama, K., Kakinami, Y., Liu, J. Y., Abdu, M. A., & Cheng, C. Z. (2011). Latitudinal distribution of anomalous ion density as a precursor of a large earthquake. Journal of Geophysical Research, 116(A4). doi:10.1029/2010JA015948
Chemical Reduction and Deposition of Nanostructured Pt–Au Alloy
Nanostructured metal alloys made up of Pt and another metal are more efficient in catalysing reactions than pure Pt nanoparticles. However, few studies have investigated low heat, solvent-free chemical deposition techniques of nanostructured metal alloys. This paper investigates the deposition of Pt–Au nanostructured metal alloy on fluorine-doped tin oxide glass via the low heat, solvent-free polyol reduction and the effect of Pt:Au mass loading ratio on the catalytic performance. The deposition process involves drop-casting the metal precursors, H2PtCl6 and HAuCl4, on the glass substrates and reducing the precursors with vaporised ethylene glycol at 170°C for 15 minutes. The scanning electron microscope revealed that the structure of Pt–Au alloy changes from three-dimensional globular nanostructures to two-dimensional triangular and hexagonal nanoplates and to three-dimensional nanocrystals as the Au concentration increases. The X-ray photoelectron spectroscopy confirmed that the precursors on glass substrates were reduced to metallic Pt and Au. The electrocatalysis of CH3OH, with the Pt–Au glass substrates as work electrodes, showed that Pt–Au alloys have better catalytic performances than those of pure Pt and the catalytic rate peaks at a certain Pt:Au mass loading ratio.
Platinum (Pt) nanoparticles act as catalysts in proton exchange membrane (PEM) fuel cells powering machinery (Bing, Liu, Zhang, Ghosh, & Zhang, 2010; Ouyang & Cho, 2011). Using H2 or liquid fuels like CH3OH, PEM fuel cells, made up of acid-soaked PEM placed in between the anode and cathode catalyst, oxidize the fuel at the cathode and reduce the oxygen entering the cell. This creates a potential difference, V, that drives an electric current. Said electric current can be used to power a variety of applications. The fuel cell could use a Pt plate or Pt coated substrate as either the anode or cathode catalysts. It has been reported in studies that by combining Pt with other metals to form nanostructured metal alloys (NMA), the adsorption of carbonaceous poisoning species like CO is suppressed (Ren et al., 2010). Such poisoning species tend to permanently bind themselves to the catalyst, leaving less sites for the oxidation and reduction of chemical species responsible for driving the electric current. Less adsorption of poisoning species in NMA catalysts can lead to enhanced catalytic performance.
NMA can be chemically deposited on substrates in the same way as pure metal nanoparticles do. There are several deposition methods, of which the chemical reduction method is the most popular method (Herricks, Chen & Xia, 2004; Ouyang & Cho, 2011; Skrabalak, Wiley, Kim, Formo, & Xia, 2008). The chemical reduction method uses a reducing agent, such as NaBH4 and LiBEt3H, to reduce metal precursors to their pure metallic form (Gonsalvesa, Rangarajan & Wang, 2000). By controlling the experimental conditions, like the pH and precursor concentration, the shape, size and composition of NMA are well-controlled, making this method the most popular (Ouyang & Cho, 2011).
Most studies employ complex chemical reduction methods to produce NMA with high catalytic capabilities. They include heating at high temperatures over 300°C (Ganesan, Freemantle, & Obare, 2007; Jana, Dutta, Bera, & Koner, 2008) or using complicated postnanoparticle immobilisation processes like layer-by-layer deposition (Ouyang & Cho, 2011). A convenient chemical reduction method by Ouyang & Cho (2011) is the low heat, solvent-free polyol reduction. The reducing agent used is ethylene glycol (EG), which is vaporized under low heat (below 200°C) so that the vapour will reduce metal precursors. EG is then oxidized to aldehydes and carboxylic acids. Gaseous products from this reaction will escape into the air, leaving the NMA end-product free of any liquid organic compounds. The formed NMA will have well-defined shapes and good adhesion to glass substrates. Thus, no additional steps of mixing metal precursors with surfactants and additives, which control the shapes, are required.
Here we investigate the deposition of Pt–Au NMA of varying Pt:Au mass loading ratios on fluorine-doped tin oxide (FTO) glass substrates using the low heat, solvent-free polyol reduction by Ouyang & Cho (2011). We evaluate the hypothesis that the low heat, solvent-free polyol reduction method is able to produce the Pt–Au NMA that has better catalytic ability than that of pure Pt. We also investigate the role that the Pt:Au ratio plays in determining the catalytic capability of Pt–Au NMA.
Materials and Methods
An FTO glass sheet was cut into 1cm×1cm pieces for the SEM imaging and 1cm×6cm pieces for CV. They were then sterilized by subjecting them to sonication in an ultrasonic bath using distilled water and isopropyl alcohol. The chemicals used were EG, chloroplatinic acid hexahydrate (H2PtCl6.6H2O) and gold (III) chloride trihydrate (HAuCl4.3H2O) obtained from Sigma-Aldrich. The respective acid salts were then each dissolved in 90% ethanol to form 0.01M solution.
Deposition of Pt–Au NMA
NMA were deposited by reducing the metal precursors, H2PtCl6 and HAuCl4, with EG vapor. NMA of varying Pt:Au mass loading ratios were to be investigated. Hence, three 40ml solutions with different metal precursor mixing ratios were prepared. The mixing ratios were such that once they are reduced, they produce the required Pt:Au ratio (1:1, 1:2 and 1:3) on the glass substrates. A 40ml solution consisting of only H2PtCl6 without HAuCl4 was also prepared.
In a typical experiment, 10ml of the prepared solution was drop-casted onto the glass substrate. After drying the metal precursor layer at room temperature for 15 minutes, the substrate was placed into a large petri dish containing a small petri dish filled with EG. The large petri dish was then placed on a hot plate at 170°C for 15 minutes. The large petri dish was covered with a glass lid so as to create a closed environment, which allowed the vaporised EG to be contained within the dish and to effectively reduce the metal precursors. After 15 minutes, the glass cover was removed and the heating continued for another ten minutes in order to dry the substrate.
Characterization of Pt–Au NMA
To study the morphologies of NMA, SEM images of NMA were taken with a Zeiss Supra 40 field emission scanning electron microscope, with the extra high tension voltage fixed at 5kV under the secondary electron imaging mode. The XPS spectra of NMA, which are used to confirm if the metal precursors are successfully reduced to their metallic forms, were acquired using an Axis Ultra DLD X-ray photoelectron spectrometer equipped with an Al Ka1 X-ray source of 1486.6eV.
Electrochemical Catalysis of Pt–Au NMA
The CV was used to investigate the oxidation of CH3OH with the NMA-doped glass as the catalyst. The NMA glass substrate, a saturated calomel electrode and a Pt plate were used as the work, reference and counter electrode respectively. The electrolyte consisted of 0.5M H2SO4 and 0.5M CH3OH. The potential scan rate was 20mVs-1.
Pt–Au NMA Characterization Using SEM
Differing morphologies are observed under the SEM, for the three NMA of different Pt:Au mass loading ratios (Figure 1). For the 1:1–NMA, three-dimensional (3D) globular nanostructures were observed (Figure 1a). For the 1:2–NMA, 2D triangular and hexagonal nanoplates covered with mesh-like structures were observed (Figure 1b). For the 1:3–NMA, 3D nanocrystals were observed (Figure 1c).
Pt–Au NMA Characterization Using XPS
The three XPS spectra of the three NMA display four sets of peaks located around 71.0, 74.2, 83.7 and 87.5eV (Figure 2). The corresponding binding energies of the peaks will be compared with those of Pt and Au in existing literature to determine if the NMA are made up of metallic Pt and Au.
Electrochemical Catalytic Performance of Pt–Au NMA
From the oxidation of CH3OH on the Pt–Au NMA catalyst, the varying oxidation current shown in the cyclic voltammogram can determine the NMA’s catalytic performance. The 15th cyclic voltammograms of the NMA-FTO glass of different Pt:Au ratios, as well as the control FTO glass doped with only Pt, display two distinct current peaks (Figure 3). The right peak, If, which was produced by the forward sweep and the left peak, Ib, which was produced by the backward sweep when the potentials are ~ 0.7V and ~ 0.5V respectively. The If and Ib values for the Pt and various Pt–Au NMA catalysts are listed in Table 1.
Pt–Au NMA Characterization Using SEM
The differing morphologies exhibited by different NMA can be partially explained by existing literature and the specific surface energies of Pt and Au. Existing studies have shown that Au nanostructures are usually 2D (Cho, Mei, & Ouyang, 2012), while Pt nanostructures are 3D (Cho & Ouyang, 2011; Shen et al., 2008) when grown on substrates. By referring to their lowest surface energy crystallographic plane (111), given that Pt and Au have face-centered cubic structures, the specific surface energies of Pt and Au are 2.299 J m–2 and 1.283 J m–2 respectively (Vitos, Ruban, Skriver, & Kollar, 1998). The high surface energy of Pt causes Pt nanoparticles to cluster together, in order to minimize the surface area of Pt, for maximum stabilization, leading to the formation of 3D (rather than 2D) nanostructures. In contrast, the formation of 2D Au nanoplates (which have wide surface areas) is energetically feasible due to the low surface energy of Au. Thus, the highest concentration of Pt nanoparticles in the 1:1–NMA, compared to other NMA, can explain the observed overall 3D globular nanostructures dominated by Pt in the 1:1–NMA. Similarly, the more Au nanoparticles in the 1:2–NMA compared to the 1:1–NMA led to overall visible growths of 2D nanoplates dominated by Au in the 1:2–NMA but not in the 1:1–NMA.
However, no existing studies provide insight on the formation of 3D nanocrystals in the 1:3–NMA, which has more Au nanoparticles than that of 1:2–NMA. The transformation from 2D nanoplates to 3D nanocrystals may suggest that with increasing Au concentration, it is not energetic feasible for 2D nanoplates to exist. The higher total surface energy of Pt–Au system results in Pt and Au nanoparticles to cluster together to form crystalline structures for stabilization. Nevertheless, further research is necessary to explicate the transition from 2D nanoplates to 3D nanocrystals above the threshold Au concentration for Pt–Au NMA. Since the morphology controls the chemical reaction sites governing catalytic activity, understanding the morphological change can perhaps explain the different catalytic performances of Pt–Au NMA of varying ratios when used in fuel cells (Bing et al., 2010).
Pt–Au NMA Characterization Using XPS
As seen in the results section, for all three NMA, the binding energies associated with four sets of peaks are close to those reported in existing literature (Cho et al., 2012; Cho & Ouyang, 2011; Ye et al., 2011). These studies report that for metallic Pt and Au, the Pt4f bands lie around 71 and 74eV, and the Au4f bands lie around 84 and 88eV. Thus, from the spectra, the binding energies of the first two peaks (71.0 and 74.2eV) belong to Pt4f bands and the last two peaks (83.7 and 87.5eV) belong to Au4f bands. This indicates that the metal precursors were successfully reduced to their metallic forms in all NMA.
Electrochemical Catalytic Performance of Pt–Au NMA
From the voltammograms, the If peak was attributed to the three-step adsorption of CH3OH on Pt sites of NMA (Bagotzky, Vassilliey, & Khazova, 1977; Manoharan & Goodenough, 1992):
The adsorbed H species will subsequently be lost into the solution as H+ ions, while liberating electrons that contribute to the If peak. In addition, OH species will be adsorbed throughout the three-step oxidation process, leading to the oxidation of adsorbed carbonaceous species in Equation 1 (CH2OH, CHOH and COH) to form CH2O, HCOOH and CO2 (Bagotzky, Vassilliey, & Khazova, 1977; Manoharan & Goodenough, 1992). Should the adsorbed carbonaceous species fail to be oxidised by OH species, at higher voltages, adsorbed COH species will be oxidised to CO and CO2, thereby liberating electrons that contribute to the If peak as well according to Equation (2a) (Manoharan & Goodenough, 1992):
The catalyst’s performance can be determined by calculating the ratio of If/Ib. A lower ratio indicates a poorer catalytic performance. This is because a lower ratio corresponds to a higher Ib peak, which implies that more CO molecules are adsorbed to the catalyst and greater CO oxidation rate. The tabulated ratios in Table 1 show that the FTO glass doped with only Pt has the lowest ratio. Also, modifying the Pt:Au ratio would result in different tolerance performances, which the Pt:Au ratio of 1:2 yields the highest If/Ib ratio and best catalytic performance.
Comparing between Pt-doped catalyst with only Pt and Pt–Au NMA catalysts, Au plays a key role in enhancing the catalytic performance of Pt as described by Choi et al. (2006). It is because Au is responsible for increasing the oxidation rate of HCOOH by inducing a major oxidation reaction pathway, which reduces the formation of CO. Thus, lower amount of CO accumulates in Pt–Au catalysts, which reduce catalytic poisoning and lead to lower Ib peak in NMA compared to Pt-doped catalyst.
While the incorporation of small amount of Au leads to better catalytic performance, based on the current study that investigates three different Pt:Au ratios (1:1, 1:2 and 1:3), higher Au concentration above a threshold Pt:Au ratio of 1:2 can instead reduce the catalytic performance. According to Wang et al. (2010) and Ye et al. (2011), since the Fermi energies of Pt and Au are different, Au modifies the electronic structure of Pt–Au NMA, resulting in electrons tending to flow from Pt to Au. At high Au concentration, electrons are less available in Pt for adsorption of OH species that oxidize CO in Equations (2b) and (2c). In addition, as Pt (but not Au) is primarily responsible for the chemisorption activities, excessively high Au concentration in Pt–Au NMA implies that less Pt sites are available for adsorption of OH species. It is possible that less adsorbed OH species lead to more incomplete oxidation of CH3OH, more accumulation of CO and higher Ib peak. Hence, these explanations justify the peak catalytic performance at the Pt:Au ratio of 1:2.
In conclusion, this investigation had successfully deposited Pt–Au NMA on the FTO glass via the low heat, solvent-free polyol reduction. From SEM imaging, different Pt:Au loading ratios resulted in different morphologies, as attributed to differing total surface energies of Pt–Au systems. The XPS spectra confirmed that the precursors were reduced to metallic Pt and Au. Studying the electrochemical catalysis of methanol confirmed the hypothesis that the NMA-doped glass had a better catalytic performance than the Pt-doped glass. Also, another hypothesis that was verified is varying the Pt:Au ratio affects the catalytic performance, which its performance peaks at the Pt:Au ratio of 1:2. The investigated method provides a convenient alternative for producing high-performing catalysts compared to the high heat, solvent-based methods that are currently employed in industries. Future research can explain the changing morphologies of Pt–Au NMA with respect to varying Pt:Au ratio, which can assist other research seeking formulate new methods of controlling the morphology of Pt–Au NMA during the NMA production in order to develop optimally-performing NMA. Also, Pt–Au NMA of other Pt:Au ratios, apart from the three currently studied ratios (1:1, 1:2 and 1:3), can be characterized by other research to understand the catalytic performance change by increasing the Pt:Au ratio. This has important implications on situations where there is limited Pt or Au and there is a need to produce Pt–Au NMA of maximal catalytic capability with limited resources.
This project is conducted under the 2013 Undergraduate Research Opportunities Programme supported by the Faculty of Engineering at National University of Singapore. The author is grateful to his advisors, Dr. Karimbintharikkal G. Nishanthand Prof. Ouyang Jianyong for providing the necessary guidance for this project, as well as the Laboratory Office of Department of Materials Science and Engineering for providing the materials and equipment for this project.
Bagotzky, V. S., Vassiliev, Y. B., & Khazova, O. A. (1977). Generalized scheme of chemisorption, electrooxidation and electroreduction of simple organic compounds on platinum group metals. Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, 81(2), 229–238. doi: 10.1016/0368-1874(77)80299-2.
Bing, Y., Liu, H., Zhang, L., Ghosh, D., & Zhang, J. (2010). Nanostructured Pt-alloy electrocatalysts for PEM fuel cell oxygen reduction reaction. Chemical Society Reviews, 39(6), 2184–2202. doi: 10.1039/b912552c.
Cho, S. J., Mei, X., & Ouyang, J. (2012). Direct deposition of gold nanoplates and porous platinum on substrates through solvent-free chemical reduction of metal precursors with ethylene glycol vapor. Physical Chemistry Chemical Physics, 14(45), 15793–15801. doi: 10.1039/c2cp42164j.
Cho, S. J., & Ouyang, J. (2011). Attachment of platinum nanoparticles to substrates by coating and polyol reduction of a platinum precursor. The Journal of Physical Chemistry C, 115(17), 8519–8526. doi: 10.1021/jp2001699.
Choi, J. H., Park, K. W., Park, I. S., Kim, K., Lee, J. S., & Sung, Y. E. (2006). A Pt–Au nanoparticle electrocatalyst for methanol electro-oxidation in direct methanol fuel cells. Journal of the Electrochemical Society, 153(10), A1812–A1817. doi: 10.1149/1.2224055.
Ganesan, M., Freemantle, R. G., & Obare, S. O. (2007). Monodisperse thioether-stabilized palladium nanoparticles: synthesis, characterization, and reactivity. Chemistry of Materials, 19(14), 3464–3471. doi: 10.1021/cm062655q.
Gonsalvesa, K. E., Rangarajana, S. P., & Wang, J. (2000). Chemical synthesis of nanostructured metals, metal alloys, and semiconductors. Handbook of Nanostructured Materials and Nanotechnology, 1, 1–56. doi: 10.1016/b978-012513760-7/50004-6.
Herricks, T., Chen, J., & Xia, Y. (2004). Polyol synthesis of platinum nanoparticles: Control of morphology with sodium nitrate. Nano Letters, 4(12), 2367–2371. doi: 10.1021/nl048570a.
Jana, S., Dutta, B., Bera, R., & Koner, S. (2008). Immobilization of palladium in mesoporous silica matrix: preparation, characterization, and its catalytic efficacy in carbon–carbon coupling reactions. Inorganic Chemistry, 47(12), 5512–5520. doi: 10.1021/ic8004294.
Liu, Z., Ling, X. Y., Su, X., & Lee, J. Y. (2004). Carbon-supported Pt and PtRu nanoparticles as catalysts for a direct methanol fuel cell. The Journal of Physical Chemistry B, 108(24), 8234–8240. doi: 10.1021/jp049422b.
Mancharan, R., & Goodenough, J. B. (1992). Methanol oxidation in acid on ordered NiTi. Journal of Materials Chemistry, 2(8), 875–887. doi: 10.1039/jm9920200875.
Ren, H., Humbert, M. P., Menning, C. A., Chen, J. G., Shu, Y., Singh, U. G., & Cheng, W. C. (2010). Inhibition of coking and CO poisoning of Pt catalysts by the formation of Au/Pt bimetallic surfaces. Applied Catalysis A: General, 375(2), 303–309. doi: j.apcata.2010.01.018.
Shen, Q., Jiang, L., Zhang, H., Min, Q., Hou, W., & Zhu, J. J. (2008). Three-dimensional dendritic Pt nanostructures: sonoelectrochemical synthesis and electrochemical applications. The Journal of Physical Chemistry C, 112(42), 16385–16392. doi: 10.1021/jp8060043.
Skrabalak, S. E., Wiley, B. J., Kim, M., Formo, E. V., & Xia, Y. (2008). On the polyol synthesis of silver nanostructures: glycolaldehyde as a reducing agent. Nano Letters, 8(7), 2077–2081. doi: 10.1021/nl800910d.
Vitos, L., Ruban, A. V., Skriver, H. L., & Kollar, J. (1998). The surface energy of metals. Surface Science, 411(1), 186–202. doi: 10.1016/s0039-6028(98)00363-x.
Wang, C., Tian, W., Ding, Y., Ma, Y. Q., Wang, Z. L., Markovic, N. M., … & Sun, S. (2010). Rational synthesis of heterostructured nanoparticles with morphology control. Journal of the American Chemical Society, 132(18), 6524–6529. doi: 10.1021/ja101305x.
Xu, J. B., Zhao, T. S., Yang, W. W., & Shen, S. Y. (2010). Effect of surface composition of Pt–Au alloy cathode catalyst on the performance of direct methanol fuel cells. International Journal of Hydrogen Energy, 35(16), 8699–8706. doi: 10.1016/j.ijhydene.2010.05.008.
Ye, W., Kou, H., Liu, Q., Yan, J., Zhou, F., & Wang, C. (2012). Electrochemical deposition of Au–Pt alloy particles with cauliflower-like microstructures for electrocatalytic methanol oxidation. International Journal of Hydrogen Energy, 37(5), 4088–4097. doi: 10.1016/j.ijhydene.2011.11.132.
Access Full PDF: Chemical Reduction and Deposition of Nanostructured Pt-Au Alloy