ANALYSIS OF THE AMMONIUM ASSIMILATION PATHWAYS OF THE HUMAN COLONIC BACTERIUM, BACTEROIDES THETAIOTAOMICRON
BY
MICHAEL IAKIVIAK
DISSERTATION
Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Animal Sciences in the Graduate College of the University of Illinois at Urbana-Champaign, 2017
Urbana, Illinois
Doctoral Committee
Professor Roderick I. Mackie, Chair Professor Isaac K.O. Cann Assistant Professor Jason Ridlon Assistant Professor Patrick Degnan
Abstract
In ruminants, efficient rumen function and proper host metabolism is dependent
on the nitrogen supply in the feed. Assimilated ammonium accounts for up to 70% of the
microbial protein production, which satisfies up to 85% of the host protein requirements.
Similar numbers for the human colon have not been determined. However, colonic
bacteria are responsible for the production of ammonium, derived from host-secreted
urea and endogenous and dietary proteins, that provides the preferred nitrogen source
for microbial growth. Bacteroides thetaiotaomicron, a model organism for human gut
Bacteroidetes, encodes genes for the capture of ammonium through the two primary
pathways, the glutamate dehydrogenase (GDH) pathway and the glutamine
synthetase/glutamate synthase (GS/GOGAT) pathway.
To gain insight into the genomic features underlying ammonium uptake and
assimilation in this bacterium, comparative transcriptomic analysis using RNA-Seq was
employed on cultures growing under excess or limiting ammonium concentrations. A
single genomic locus, encoding for the GS/GOGAT pathway, was identified with highly
increased transcription when the organism grows under limiting ammonium
concentration. The relative contribution of each gene to ammonium assimilation was
assessed through construction of genomic deletion strains for each of the three GS, one
GOGAT, and two GDH genes. The deletion of two genes, the NADPH-dependent glutamate dehydrogenase (gdhA) and the glutamine synthetase type 3 (glnN2)
significantly impeded growth of the organisms under both nitrogen conditions. Taken
together, the results demonstrate the importance of the GDH pathway for constitutive
ii ammonium assimilation, and GlnN2 for glutamine biosynthesis. However, when the organism grows under nitrogen limitation, the GS/GOGAT pathway, including glnN1, is highly induced. To extrapolate the significance of the findings, a comparative bioinformatic analysis, using all of the available sequenced Bacteroides genomes, revealed high conservation of the critical genomic loci in gut species. Understanding of nitrogen metabolism in gut microbes is essential for a complete depiction of their ecological implications on the host´s metabolism in health and disease.
iii
Table of Contents
List of Figures ...... vii
List of Tables ...... xii
Chapter 1. Literature review ...... 1
1.1 Introduction ...... 1
1.2 Gut microbiota ...... 6
1.3 Bacterial nitrogen metabolism ...... 8
1.4 Enzymatic pathways of ammonium assimilation ...... 11
1.5 Ammonium assimilation in gut Bacteroidetes ...... 22
1.6 Proposition for research ...... 29
Chapter 2. Comparative genomic analysis for phylogenetic conservation of ammonium assimilation genes within Bacteroides and Prevotella ...... 31
2.1 Introduction ...... 31
2.2 Materials and Methods ...... 35
2.3 Results and Discussion ...... 36
Chapter 3. Nitrogen utilization patterns of B. thetaiotaomicron and genome wide transcriptional response to extracellular ammonium ...... 48
3.1 Introduction ...... 48
iv 3.2 Materials and Methods ...... 52
3.3 Results ...... 56
3.4 Discussion ...... 72
Chapter 4. Probing the genetic machinery of ammonium assimilation through phenotypic and competitive fitness analyses ...... 80
4.1 Introduction ...... 80
4.2 Materials and Methods ...... 84
4.3 Results ...... 91
4.4 Discussion ...... 103
Chapter 5. Discussion and Future Perspectives ...... 108
5.1 Introduction ...... 108
5.2 Major finding...... 110
5.3. Implication to the Enteric Paradigm ...... 115
5.4 Future Prospective ...... 128
5.5 Conclusion ...... 131
References ...... 134
v Appendix A. Media Composition ...... 167
Appendix B. Two New Xylanases with Different Substrate Specificities from the Human
Gut Bacterium Bacteroides intestinalis DSM 17393 ...... 169
Appendix C. Functional and modular analyses of diverse endoglucanases from
Ruminococcus albus 8, a specialist plant cell wall degrading bacterium ...... 206
vi List of Figures
Figure 1. Schematic diagram of microbial nitrogen cycling in the rumen and monogastrics animals ...... 4
Figure 2. Enteric paradigm for ammonium assimilation based on E. coli ...... 13
Figure 3. Catalytic mechanisms of ammonium assimilation ...... 15
Figure 4. Diagram of four genomic loci containing ammonium assimilation genes
...... 27
Figure 5. Phylogenetic conservation of loci encoding high affinity genes in
Bacteroides species ...... 39
Figure 6. Phylogenetic conservation of the locus encoding low affinity genes in
Bacteroides sp...... 40
Figure 7. Conservation of genes encoded by the high affinity loci in Bacteroides
fragilis strains ...... 42
Figure 8. Conservation of genes encoded by the low affinity locus in Bacteroides
fragilis strains ...... 43
Figure 9. Conservation of ammonium assimilatory genes in Prevotella species .. 45
Figure 10. Growth of B. thetaiotaomicron VPI-5482 on non-limiting concentrations
of various nitrogen sources ...... 57
Figure 11. Ammonium concentration of B. thetaiotaomicron VPI-5482 growing on
non-limiting concentrations of various nitrogen sources ...... 59
Figure 12. Growth of B. thetaiotaomicron VPI-5482 on limiting concentrations of
individual and mixed nitrogen sources ...... 60
vii Figure 13. Growth of B. thetaiotaomicron VPI-5482 on limiting concentrations of
mixed nitrogen sources ...... 61
Figure 14. Ammonium concentration of B. thetaiotaomicron VPI-5482 on limiting
concentrations of individual and mixed nitrogen sources ...... 62
Figure 15. Growth parameters of B. thetaiotaomicron VPI-5482 during growth in
continuous culture ...... 64
Figure 16. Production of short chain fatty acids by B. thetaiotaomicron VPI-5482 grown for 20 days in continuous culture ...... 65
Figure 17. Global transcriptional regulatory changes of B. thetaiotaomicron strain
VPI-5482 grown in continuous culture with limiting vs. excess ammonium ...... 68
Figure 18. RNA-Seq coverage map of the B. thetaiotaomicron VPI-5482 ammonium
assimilation loci with representative samples from limiting and non-limiting conditions ...... 75
Figure 19. Schematic diagram showing deletion of the ammonium transporter
gene (amtB) in B. thetaiotaomicron ...... 92
Figure 20. Growth of B. thetaiotaomicron mutant strains on limiting concentration
of peptides ...... 94
Figure 21. Growth of B. thetaiotaomicron deletion strains on varying
concentrations of glucose and ammonium ...... 95
Figure 22. Growth of B. thetaiotaomicron complementation strains on varying
concentrations of glucose and ammonium ...... 97
Figure 23. Growth of B. thetaiotaomicron barcoded strains on varying
concentrations of glucose and ammonium ...... 99
viii Figure 24. Competition between selected B. thetaiotaomicron barcoded deletion strains versus the barcoded wild-type strain when cultured on excess and limiting ammonium ...... 101
Figure 25. Competition between selected B. thetaiotaomicron barcoded deletion strains versus the B. thetaiotaomicron ΔglnN1 ΩBC04 strain when cultured on
excess and limiting ammonium ...... 102
Figure 26. Schematic diagram of ammonium assimilation in Bacteroides
thetaiotaomicron VPI-5482 under ammonium limitation and ammonium excess.
...... 111
Figure 27. Genomic context for the two Bacteroides intestinalis GH8 genes ...... 172
Figure 28. Optimal parameters for B. intestinalis Xyn8A, Rex8A, and Xyl3A activity
...... 179
Figure 29. Amino acid sequence alignment for Rex8A (BACINT_00927) and Xyn8A
(BACINT_04210) ...... 183
Figure 30. The two Bacteroides intestinalis GH8 genes encode xylan-degrading
enzymes with distinct properties ...... 185
Figure 31. Hydrolysis of XOS by Rex8A (BACINT_00927) ...... 189
Figure 32. Hydrolysis of XOS by Xyn8A (BACINT_04210) ...... 190
Figure 33. Mutational analysis reveals residues important for catalysis in Xyn8A
(BACINT_04210) and Rex8A (BACINT_00927) ...... 192
Figure 34. BACINT_00926 encodes a GH3 β-xylosidase, Xyl3A ...... 194
Figure 35. Expansion and diversity of Rex8A homologs in the gut microbiome . 196
ix Figure 36. Thin-layer chromatography screening of R. albus 8 endoglucanases.
...... 215
Figure 37. Domain architectures and purification of 7 glycoside hydrolase family 5 enzymes from Ruminococcus albus 8 ...... 224
Figure 38. Comparison of cellulose hydrolysis by the endoglucanases of R. albus
8 ...... 227
Figure 39. Time course analysis of hydrolysis of phosphoric acid swollen cellulose by Ra0259, Ra0325 and Ra2535 ...... 229
Figure 40. Enzyme kinetic analysis for individual endoglucanases ...... 230
Figure 41. Phylogenetic analysis of GH5 catalytic domains of R. albus 8 endoglucanases ...... 231
Figure 42. Binding of soluble and insoluble polysaccharides by the R. albus 8 endoglucanases ...... 233
Figure 43. Truncational analyses for identification of CBMs in Ra1831 and Ra2535
...... 235
Figure 44. Substrate binding characteristics of the CBM65 from endoglucanase
Ra2535 to insoluble substrate ...... 236
Figure 45. Amino acid sequence alignments of the analyzed CBMs in Ra2535 and
Ra1831 with homologous sequences in the NCBI ...... 238
Figure 46. Substrate binding characteristics of the CBM65 from endoglucanase
Ra2535 to soluble substrates ...... 239
Figure 47. Identification of a novel CBM in Ra1831 ...... 242
x Figure 48. Site directed mutagenesis of Ra1831 TM2 and binding of the mutants to soluble polysaccharides ...... 244
xi List of Tables
Table 1. Genes used as bioinformatic queries ...... 38
Table 2. Growth parameters of B. thetaiotaomicron VPI-5482 on non-limiting concentrations of various nitrogen sources ...... 53
Table 3. Qualitative assessment of read mapping to the closed genome of B.
thetaiotaomicron strain VPI-5482 from Illumina HiSeq4000 sequencing results... 67
Table 4. Genes of B. thetaiotaomicron with significant transcriptional changes
during continuous growth in limiting vs. excess ammonium ...... 70
Table 5. Relative expression and transcript abundance of the ammonium
assimilation genes of B. thetaiotaomicron strain VPI-5482 grown in continuous culture with limiting vs. excess ammonium ...... 71
Table 6. List of primers used for construction of deletion plasmids ...... 86
Table 7. List of primers used for construction of complementation plasmids ...... 88
Table 8. List of assorted primers used for mutant construction, sequencing and
qPCR analysis ...... 90
Table 9. Brain heart infusion supplemented (BHIS) medium ...... 167
Table 10. Composition of chemically defined medium (CDM) used for the culture
of B. thetaiotaomicron ...... 168
Table 11. Primers used in this study ...... 175
Table 12. Analysis of CD spectra using DICHROWEB ...... 186
Table 13. R. albus genes previously cloned, expressed, and demonstrated to exhibit cellulose degrading activity ...... 210
xii Table 14. Primers used in cloning R. albus 8 functional endoglucanases, their
predicted signal peptides, and characteristic of recombinant proteins ...... 212
Table 15. Primers used in cloning truncational mutants of Ra1831 and Ra2535 and characteristic of recombinant proteins ...... 218
Table 16. Primers used in site directed mutagenesis of ra1831 TM2 and characteristic of recombinant proteins ...... 219
Table 17. Binding parameters of Ra2535 truncational mutants for cello- and xylo- oligosaccharides determined by ITC ...... 240
Table 18. Screening of R. albus 8 putative family 5 endoglucanases for enzymatic
activity on different glucans ...... 247
xiii
Chapter 1. Literature review
1.1 Introduction
The mammalian gut microbiota impact host physiology, eliciting nutritional,
immunological, and developmental effects to the benefit and/or detriment of the host.
The relationship between gut microbiota, the entirety of the microbial population, and
the host has been of pronounced interest to animal production and human health over several decades. In any gut system, microbial interactions with the host can be mediated through direct epithelial contact or metabolite driven (e.g. through transport of fermentation products). Host animals have evolved in parallel with the symbiotic microorganisms that participate in metabolic cross-feeding were short chain fatty acids
made by microbes are used as energy sources, stimulating growth and development.
However, unfavorable interactions may occur through opportunistic pathogens or metabolic dysfunction detrimental to the host.
In production animals, commensal microorganisms are involved in growth efficiency and feed conversion as they consume undigested feedstock, making
fermentation end-products available to the host. Optimization of feed components
indigestible to the host has been primarily through microbial fermentation into
metabolizable short chain fatty acids (SCFAs). This phenomenon is most pronounced in
ruminants, are dependent on microbial foregut fermentation to provide the energy,
protein, and vitamin requirements of the host (Wolin, Miller, & Stewart, 1997). The beef
and dairy cattle industry is dependent on the bacterial degradation of the feed, and
1 through optimizing feed digestibility and feed conversion have improved rumen bacterial fermentation. Similar implications exist for other production animals, such as swine and poultry, as all animals have gut or colonic microbiota capable of host interaction and participate in symbiotic interactions with their host.
The gut microbiota of humans is similarly linked to various aspects of host nutrition and health. Most critical are the microbes to negatively and positively impact the health of the host (Ohland & Jobin, 2015). Microbes metabolize components of the diet, including dietary fiber and proteins, which bypass digestion in the stomach and small intestine and become substrates for fermentation in the colon. As the major end products of fermentation, SCFAs contribute 5-10% of the host energy requirements and act as an energy source for colonocytes, immune system modulators, and virulence regulators in pathogens (Den Besten et al., 2013; Sun & O'Riordan, 2013; Vogt, Pena-
Diaz, & Finlay, 2015). Healthy brain function, hematopoiesis and host immune system maturation are linked with the resident bacteria (Collins, Surette, & Bercik, 2012;
Sommer & Backhed, 2013; Trompette et al., 2014). Antithetically, gut microbes are also associated with a myriad of pathologies, including increased bowel inflammation, allergic airway disease, anxiety, autism, atherosclerosis, and colorectal cancer (Bennet,
Ohman, & Simren, 2015; De Palma et al., 2015; Hsiao et al., 2013; Karlsson et al.,
2012; E. Kim, Coelho, & Blachier, 2013; Tomasello et al., 2014; Trompette et al., 2014).
Nitrogen metabolism in the ruminant animal is primed by the microbiota in the rumen. The bacterial digestion of poor quality protein from forage and non-protein nitrogen from ammonium and urea is responsible for the growth of bacteria and accumulation and synthesis of bacterial protein. The bacterial protein is subsequently
2 used as the protein source of the host. In contrast, the monogastrics animal first digests and metabolizes any readily digestible protein before the microbial community encounters the remaining poor quality or host indigestible protein. In both scenarios, the microbiota predominantly encounters ammonium, urea, and host indigestible nitrogen sources as sources of bacterial protein production (Figure 1).
From a human health perspective, the nitrogen cycle in the colon is linked to
several disease states. Diets high in protein are associated with chronic kidney disease
and increased incidence of colorectal cancer (E. Kim et al., 2013; Marckmann, Osther,
Pedersen, & Jespersen, 2015). Production of hazardous protein fermentation products
from gut bacteria, such as indole, phenol, and p-cresol, are linked to cell damage and colon cancer (Bone, Tamm, & Hill, 1976; Chung, Fulk, & Slein, 1975). The gut microbiota is responsible for ammonium production from host derived urea and deamination of proteins (Moreau, Ducluzeau, & Raibaud, 1976). Individuals with hepatic encephalopathy or defects in the urea cycle can develop hyperammonemia, an accumulation of systemic ammonium, which is toxic to the central nervous system and the epithelial membrane (Andriamihaja et al., 2010; Batshaw, Tuchman, Summar, &
Seminara, 2014; Lin & Visek, 1991; Raabe, 1990; Shawcross & Jalan, 2005). Previous studies in rats and humans show a fiber-dependent response in ammonium production and utilization within the colon from dietary intervention (Birkett, Muir, Phillips, Jones, &
O'Dea, 1996; Kalmokoff et al., 2013). Later, in attempt to ameliorate hyperammonemia,
a low urease-producing defined microbial consortium (altered Schaedler flora) can
stably colonize a microbiota-depleted murine model and decrease the level of fecal
ammonium (Shen et al., 2015).
3
Figure 1. Schematic diagram of microbial nitrogen cycling in the rumen and monogastrics animals.
4 Ammonium is present at concentrations between 2-50 mM in the average human colon, capable of damaging colonocytes, affecting metabolism and promoting cancerous growth (Visek, 1978; Wolpert, Phillips, & Summerskill, 1970; O. Wrong,
Metcalfe-Gibson, Morrison, Ng, & Howard, 1965). Microorganisms in the colon incorporate ammonium directly into bacterial amino acids for proteins synthesis and growth. A diverse set of bacterial isolates from the pig hindgut revealed that a majority of the isolates preferentially assimilated ammonium for cell nitrogen (Takahashi, Benno,
& Mitsuoka, 1980). Most of the bacteria in the cecum of rabbits primarily incorporate ammonium directly into alpha-ketoglutarate to form glutamate through a NADPH dependent glutamate dehydrogenase (Forsythe & Parker, 1985). Importantly, it is unclear whether deleterious effects result from an increase in ammonium production or a dysbiotic population of microbes with decreased ammonium assimilation or irregular nitrogen cycling.
The capacity of the ruminant animal for ammonium incorporation into microbial protein as nitrogen source for the host was demonstrated by Arturi Virtanen in 1966
(Virtanen, 1966). A group of dairy cows was fed a protein-free diet for over a year with ammonium and urea supplementation as the only nitrogen sources. Animals amino acid requirements were sustained throughout a lactation cycle solely from microbial biomass derived from ammonium assimilation. Although a similar capacity for ammonium acquisition exists in all gut systems, it is most evident in the foregut fermenters, like the ruminant, whereas hindgut fermenters, like humans, do not rely on the microbes for nutrient supplementation to the same extent.
5 Importance and impact. The regulation of carbohydrate metabolism is currently intensively studied using in vitro and in vivo models, while nitrogen metabolism is frequently overlooked. The complexity of microbial metabolism is well defined in recent reviews that highlight α-ketoglutarate as a central regulatory metabolite for carbon and nitrogen metabolism (Huergo & Dixon, 2015; Ninfa & Jiang, 2005; Van Heeswijk,
Westerhoff, & Boogerd, 2013). The interconnection of these two essential nutrients remains poorly understood in human gut microbes. However, Bacteroides species are extensively studied and provide excellent model organisms with highly tractable genetic systems to expose this association. Future studies in gut microbiota require a comprehensive mechanistic description of regulatory models for both carbon and nitrogen metabolism. This knowledge and understanding is essential in order to selectively modulate the microbial community for therapeutic effects.
1.2 Gut microbiota
The impact ruminal microbes have on host health has been acknowledged for almost a century, whereas the importance of colonic microbes to human health has only been recognized in the past few decades. This disparity is evident in the knowledge of microbial ecology in the two analogous systems. As such, human microbial ecology and metabolism studies benefit from the foundation established in the ruminal microbiology field.
Foregut-fermenting ruminants and hindgut-fermenting monogastric animals possess similar microbial communities (i.e. the microbiota) in their respective fermentative compartments. This results from similar selective pressures (host immunology) and analogous functionality (nutritional inputs). Both environments are
6 dominated by the Bacteroidetes and Firmicutes phyla, with lower proportions of
Proteobacteria, Actinobacteria, and Verrucomicrobia. The dominant phyla are
comprised of organisms that have evolved to efficiently metabolize the carbohydrate
and nitrogenous compounds abundant in plant materials and host secretions.
The human gut microbiota and its impact on host health is an active and intense
research area. The Human Microbiome Project funded by the National Institutes of
Health (NIH) has analyzed stool samples from 242 healthy individuals and identified a
predominance of Bacteroidetes, with the dominant genus being Bacteroides
(Consortium, 2012). A separate project, Metagenomics of the Human Intestinal Tract
(MetaHIT) funded by the European Union, analyzed metagenomic data from 124
healthy individuals and the most frequently identified genera mostly belong to
Bacteroides and other Bacteroidetes (Qin et al., 2010). Clone library analysis of 16S sequences showed a similar trend with the Bacteroidetes as the most common taxa recovered (Eckburg et al., 2005).
As one of the dominant members of the gut community, bacteria within the
Bacteroidetes phylum are extensively studied to understand their physiology, especially in relation to host health. These organisms are strongly linked to beneficial host effects, including a lean phenotype, improved hematopoiesis, and immune system maturation
(Ley et al., 2005; Sommer & Backhed, 2013; Trompette et al., 2014). Alternatively,
Bacteroides fragilis and other species can become opportunistic pathogens and cause systemic infections (Gibson III, Onderdonk, Kasper, & Tzianabos, 1998; H. M. Wexler,
2007). Major interest has been focused on understanding fiber degradation in commensal Bacteroidetes in recent years. Organization and function of genetic loci for
7 the complete deconstruction of target polysaccharides have been described in several
Bacteroidales species (Dodd, Mackie, & Cann, 2011; Martens, Chiang, & Gordon, 2008;
Martens et al., 2011; McNulty et al., 2013; Rogers et al., 2013; Rogowski et al., 2015).
Although the ability of Bacteroides species to degrade the plant cell wall and host glycans as a carbon source is of major interest, of equal importance are the sources and mechanisms by which they obtain and metabolize colonic nitrogen.
1.3 Bacterial nitrogen metabolism
In rumen systems, ingested protein is readily degraded by resident microbes into peptides, free amino acids, and ammonium (G. Broderick & Craig, 1989; Chalupa,
1975; Newbold, Wallace, & McKain, 1990). Approximately 80% of available peptides are metabolized into microbial protein. The peptides can be degraded by some microbes and the free amino acids are rapidly deaminated and fermented (Allison,
1970; G. A. Broderick & Balthrop Jr., 1979; Chalupa, 1976; Wallace & Cotta, 1988).
Deamination of peptides and urea hydrolysis can occur more rapidly than ammonium assimilation, potentially accumulating ammonium in the blood and causing systemic complications (Bartley et al., 1976; Carroll, Barton, Anderson, & Smith, 1988; Ferguson,
Blanchard, Galligan, Hoshall, & Chalupa, 1988).
Approximately 16-43% of cultivable ruminal bacteria possess proteolytic activity
(Prins, van Rheenen, & van't Klooster, 1983). Microbial protein degradation is impacted by several factors including ruminal pH (Bach, Calsamiglia, & Stern, 2005; Kopecny &
Wallace, 1982), biochemical and structural properties of the protein (Chen, Strobel,
Russell, & Sniffen, 1987; Romagnolo, Polan, & Barbeau, 1994), and nutritional context of the protein being consumed (Kohn & Allen, 1995; Tománková & J., 1995). Although
8 amino acids can be incorporated by microbes, mixed ruminal communities preferentially consume peptides over amino acids (Ling & Armstead, 1995; Wallace, 1996). Total free amino acid pools in the rumen are extremely low due to rapid deamination by resident microbes (Wallace, 1996; Wallace, Onodera, & Cotta, 1997). The expelled ammonium contributes to the total ammonium pool which is responsible for 50-70% of the bacterial nitrogen synthesized (Mackie & White, 1990; Mathison & Milligan, 1971).
Furthermore, protein metabolism rates of colonic bacteria were found to be similar to that of the rumen (Richardson, McKain, & Wallace, 2013). Proteins that bypass intestinal digestion and endogenous host secretions are catabolized and ultimately converted into ammonium and ammonium assimilation provides the bacteria with the necessary nitrogen for protein biosynthesis. Amino acid deamination is suggested to be the primary source of ammonium in the lower gastrointestinal tract (O.
M. Wrong, Vince, & Waterlow, 1985). Another major source of colonic ammonium is derived from systemic urea secreted into the gut. Bacterial ureases hydrolyze approximately 15-30% of the host urea, approximately 7 grams per day (Jones,
Smallwood, Craigie, & Rosenoer, 1969; Walser & Bodenlos, 1959). Ureases are inversely regulated by the peptide or ammonium concentration, so that decreasing ammonium levels cause bacteria to increase their urease production to release ammonium, resulting in concentrations of ammonium that are maintained within a narrow range (Wozny, Bryant, Holdeman, & Moore, 1977). Total ammonium levels in the rumen range from 1mM to 40mM, and presumably never become limiting (<0.7mM) irrespective of diet (Russell & Strobel, 1987). Similarly, colonic levels also exist in the range of 2-50mM and since ammonium is in constant supply, this molecule is an
9 essential source of nitrogen for gastrointestinal microbes, shown to stimulate microbial growth as well as being essential for many bacteria (Bryant & Robinson, 1962).
The ammonium molecule presents important physiochemical challenges and opportunities for assimilation into biomass. The protonated and positively charged ammonium can exist in a deprotonated and gaseous form, ammonia. The gaseous ammonia is able to diffuse through the membrane and become protonated ammonium.
The pKa of ammonium is 8.95 at 35°C and at physiological pH (6.5-7.5) only 1% of the total ammonium/ammonia exist in the gaseous ammonia (Martinelle & Haggstrom,
1997). Depending on the total ammonium/ammonia concentration, gaseous ammonia diffusion is potentially responsible for ammonia transport across the cytoplasmic membrane. The ammonium transport protein, AmtB from Saccharomyces cerevisiae and E. coli, is not expressed unless the ammonium concentration is very low (Soupene,
He, Yan, & Kustu, 1998; Soupene, Lee, & Kustu, 2002; van Heeswijk et al., 1996;
Winkler, 2006). Further evidence provided by organisms completely lacking ammonium transport facilitators, Escherichia coli, Bacillus subtilis, Corynebacterium glutamicum,
Salmonella enterica and Saccharomyces cerevisiae are still able to grow optimally with excess concentrations of ammonium/ammonia (Detsch & Stulke, 2003; Meier-Wagner
et al., 2001; Soupene et al., 1998). In contrast, mixed ruminal microbes have an increased concentration gradient with 160 mg/L higher concentration in the cytoplasm compared to extracellular concentration, indicating active transport of ammonium
(Russell & Strobel, 1987). When the concentration of ammonium becomes too low, facilitated transport is required to move ammonium across the cytoplasmic membrane.
Ammonium analogs, like methyl ammonium and ethyl ammonium, have been used to
10 study the transport of ammonium across the bacterial membrane. Although useful, the results should be interpreted with caution for two primary reasons: transporters are very selective of their ligands especially for substrates as small as ammonium, and the transport of the three different molecules are not equal and can bias results (Kleiner,
1982; Kleiner & Castorph, 1982; R. Stevenson & Silver, 1977).
Once ammonium enters the cytoplasm, glutamate and glutamine are the key metabolic intermediates central to intracellular nitrogen cycling. Glutamate is the most abundant metabolite in the cell, 96 mM in E. coli, and directly links nitrogen metabolism with carbon metabolism via α-ketoglutarate, onto which ammonium is appended during ammonium assimilation (Bennett et al., 2009). Bacterial cells primarily incorporate ammonium into glutamate and glutamine, irrespective of the nitrogen conditions in the environment, which are then used as nitrogenous building blocks to a myriad of N- containing metabolites, including amino acids, purines, pyrimidines, and other metabolites.
1.4 Enzymatic pathways of ammonium assimilation
Current knowledge of enteric ammonium assimilation and regulation is largely based on research on E. coli, Klebsiella, Salmonella, and Bacillus, which does not necessarily reflect dominant gut microbes from Bacteroidetes and Firmicutes (Reitzer,
2003; Van Heeswijk et al., 2013). Research on Bacillus subtilis provides evidence for the divergence from the enteric paradigm in regards to the regulation and mechanisms of the ammonium assimilation genes (Gunka & Commichau, 2012).
11 The enteric paradigm is structured around two competing pathways that are inversely regulated depending on the ammonium concentration and nitrogen status of the cell (Figure 2). These two pathways are described as the high affinity pathway and
the low affinity pathway. The high affinity pathway is employed under low concentrations
of ammonium and contains three functional enzymes including an ATP-dependent glutamine synthetase (GS), glutamate synthase (GOGAT), and an ammonium transporter (AmtB). In contrast, the low affinity pathway is utilized under excess (non- limiting) concentrations of ammonium. This pathway consists of a NAD(P)H-oxidizing glutamate dehydrogenase (GDH). In B. subtilis, the GS/GOGAT system is solely responsible for the assimilation of ammonium, while the GDH enzymes run the reverse reaction for catabolism of glutamate.
In addition to the differences in energy expenditure and catalytic mechanisms, the regulation of these pathways have been characterized in detail. In E. coli, an elegant balance of transcriptional regulation and post-translational modification orchestrates the total contributions of both pathways. The non-competitive binding of metabolic intermediates further modulates the differential regulation and function of the catalytic enzymes and regulatory proteins.
1.4.1 The Low Affinity Pathway. Under non-limiting concentrations of
ammonium, ammonium exists in a state of equilibrium between the protonated
ammonium and deprotonated ammonia, with the pKa of ammonium being 8.95.
Deprotonated ammonia is able to pass through the cytoplasmic membrane and
becomes protonated ammonium in the cytoplasm. An ammonium transport deficient
strain of Klebsiella pneumonia constantly lost ammonia through outward diffusion when
12
Figure 2. Enteric paradigm for ammonium assimilation based on E. coli. Enzymatic pathways include low affinity and high affinity pathways with the functional proteins glutamate dehydrogenase (GDH, gdhA), glutamine synthetase (GS, glnA), glutamate synthase (GOGAT, gltB), and the ammonium transporter (AmtB). The regulatory network includes UTase, ATase, and GlnB that modify the activity of functional proteins, and transcriptional regulators, NRI, NRII, and Nac, which regulate transcription of functional genes. Proteins encapsulated in red are not identified in Bacteroidetes by sequence homology.
13 grown on nitrogen sources other than ammonium (Castorph & Kleiner, 1984). Once in the cytoplasm, a single enzyme, GDH, catalyzes the reductive amination of α- ketoglutarate into glutamate using NADPH, directly linking this enzyme to the carbon cycle (Figure 3A).
The proteobacterial glutamate dehydrogenase, encoded by gdhA, is a hexameric protein with 45-50 kDA monomers. This enzyme has a relatively low affinity for ammonium compared to glutamine synthetase, with Km values ranging from 1 to 36 mM,
and affinities for the other substrate, α-ketoglutarate and NADPH, at 0.2-6 mM and 12-
83 µM, respectively (Sharkey & Engel, 2008). A histidine can be covalently modified via
phosphorylation to inhibit the enzyme potentially responding to ATP concentrations in
the cell (Coulton & Kapoor, 1973a, 1973b). The resulting glutamate is subsequently
used for transamination reactions or to produce glutamine via glutamine synthetase
which particulates other transamination reactions to build the intracellular pool of amino
acids.
A separate category of GDH enzymes catalyzes the opposite reaction using
catabolic NADH-dependent instead of anabolic NADPH-dependent enzymes. The
catabolism of cytoplasmic glutamate consumes an NAD+ and releases α-ketoglutarate
and NADH to shuttle amino acids into carbon metabolism. For instance, B. lichiniformis
and B. megaterium utilized GDH for the purpose of ammonium assimilation in N-excess
(Bernlohr, Schreier, & Donohue, 1984; Meers, Tempest, & Brown, 1970; Phibbs &
Bernlohr, 1971), but Bacillus subtilis uses its GDH, RocG, for catabolism and runs this reaction in reverse, liberating ammonium from exogenously acquired glutamate and
14
Figure 3. Catalytic mechanisms of ammonium assimilation. Enzymatic mechanisms of (A) glutamate dehydrogenase (GDH, EC 1.4.1.4), (B) glutamine synthetase (GS, EC 6.3.1.2), and (C) glutamate synthetase (GOGAT, EC 1.4.1.13). Nitrogen atoms directly involved in the reactions are in blue.
15 releasing α-ketoglutarate into the carbon metabolism scheme (Belitsky & Sonenshein,
1998; Kane, Wakim, & Fischer, 1981).
1.4.2 The High Affinity Pathway. When extracellular ammonium becomes limiting, total ammonium/ammonia concentrations below 1mM, or pH levels decrease total gaseous ammonium concentration, the diffusion of ammonia into the cytoplasm is insufficient for growth. To compensate, the bacterial cell possesses an ammonium transporter, AmtB, that facilitates the translocation of ammonium across the cytoplasmic membrane. Almost universally, there is a small regulatory protein encoded next to the transporter, known as GlnK or P-II, capable of plugging the channel of AmtB, responsible for rapid regulation of ammonium transport (described in Section 1.4.3).
The AmtB ammonium transporter has an extremely high affinity for ammonium (1
– 150 µM, (Kleiner, 1985)) and is specifically induced when ammonium concentrations become limiting (<<1mM, (Jayakumar, Epstein, & Barnes, 1985)). A signal peptide is cleaved after membrane insertion resulting in an 11 transmembrane α-helices monomer that ultimately forms a homo trimer with three pores for transport (Khademi et al., 2004).
Some debate exists whether protonated ammonium or deprotonated ammonia is transported across the membrane, with evidence supporting both claims (Andrade &
Einsle, 2007). A more recent publication describes the transport of electrogenic ammonium through archaeal transporters using solid-supported membrane (SSM)- based electrophysiology (Wacker, Garcia-Celma, Lewe, & Andrade, 2014). In B. subtilis, the ammonium transporter (NrgA) and associated P-II like protein (NrgB) are essential for the organism to grow under acidic conditions, when gaseous ammonia concentrations are exceptionally low while protonated ammonium is still in excess
16 (Detsch & Stulke, 2003). Therefore, pH is an important factor for ammonium transport.
As pH decreases, the concentration of gaseous ammonium decreases and the flux of gaseous ammonium across the cell membrane is hampered and an ammonium transporter is employed.
Once inside the cell, ammonium concentrations fall significantly below the Km of
GDH and this enzyme, having a low affinity for ammonium, is not able to produce
enough glutamate to satisfy the demands of the cell. A high affinity protein, glutamine
synthetase is responsible for the rapid assimilation of ammonium under these
conditions, coupled to the consumption of ATP (Figure 3B). The affinity constants for
this enzyme are significantly higher than GDH with Km values for ammonium, ATP, and glutamate at 0.1 mM, 0.4 mM, and 4 mM, ensuring the forward reaction continues under limiting ammonium concentrations. In contrast to the proteobacterial enteric paradigm,
B. subtilis continues to use the GS/GOGAT system under non-limiting ammonium
concentrations (Pan & Coote, 1979). Other organisms, like Mycobacterium tuberculosis,
also require GS for growth and this enzyme shows potential as a drug target for this
organism (Mowbray, Kathiravan, Pandey, & Odell, 2014).
Large variations exist within the glutamine synthetase superfamily than can be
categorized into three families, GS Type 1 (GS1, glnA), GS Type 2 (GS2, glnII) and GS
Type 3 (GS3, glnN). The glutamine synthetase Type 1 is comprised of 52 kDa subunits
arranged into a dimer of hexameric rings held together by the apolar C-terminus binding
to a hydrophobic pocket formed by two subunits in the adjacent ring (Almassy, Janson,
Hamlin, Xuong, & Eisenberg, 1986; Colombo & Villafranca, 1986; Miranda-Rios,
Sanchez-Pescador, Urdea, & Covarrubias, 1987; Valentine, Shapiro, & Stadtman, 1968;
17 Yamashita, Almassy, Janson, Cascio, & Eisenberg, 1989). Crystallographic studies reveal an ordered mechanism as ATP binds first followed by glutamate. The phosphoryl transfer occurs yielding a γ-glutamyl phosphate before ammonium binds and deprotonates forming ammonia which attacks the carbonyl (Liaw & Eisenberg, 1994;
Liaw, Kuo, & Eisenberg, 1995). The GS Type 1 is commonly found in diverse organisms
(bacteria, archaea, plants, and mammals) and environments (gut, marine, root nodules, and soil) but are frequently found in organisms inhabiting the gastrointestinal tract
(Mathis, Gamas, Meyer, & Cullimore, 2000; J.M. Van Rooyen, Abratt, Belrhali, & Sewell,
2011; Wyatt et al., 2006).
Another family of GS enzymes encountered in the gut include the Type 3 glutamine synthetases. Originally discovered in Bacteroides fragilis, this enzyme
possesses many features of GS1 enzymes but has significant structural alterations (Hill,
Parker, Goodman, Jones, & Woods, 1989). For instance, the primary amino acid
sequence is significantly longer, from 450 amino acids in GS1 to 750 amino acids in
GS3 enzymes. Secondly, the quaternary structure, although a dimer of hexamers like
GS1, is inverted with the N-terminal regions contributing to ring dimerization (J. M. Van
Rooyen et al., 2011). The covalent modification sites present on GS1 are absent in this family, without any known regulatory features. Unlike the classical GS1 mode of regulation (discussed in section 1.4.3), GS3 likely possess novel regulatory features yet to be discovered.
Finally, GS type 2 are not commonly found in commensal gut microbes but are present in eukaryotes and plant symbionts, including bacteria and fungi (Darrow &
Knotts, 1977; Edmands, Noridge, & Benson, 1987; Joyner & Baldwin, 1966; Kumada,
18 Takano, Nagaoka, & Thompson, 1990). Conservation between bacterial and plant GS2 is high in amino acid sequence, reflecting a possible horizontal gene transfer (Shatters
& Kahn, 1989). This family is structurally different from families one and three, including
a smaller subunit of 36 kDa arranged in an octamer (Joyner & Baldwin, 1966).
After ammonium has been assimilated via glutamine synthetase, the glutamine
can then enter the amino acid pool or be used by glutamate synthase, also known as
glutamine oxoglutarate aminotransferase (GOGAT), encoded by gltBD, along with α-
ketoglutarate and NADPH to produce 2 molecules of glutamate (Figure 3C). Glutamate
synthase is an oxidoreductase, despite being termed an aminotransferase, composed
of two subunits, GltD and GltB. The smaller subunit, GltD, is 52 kDa in size and supplies
the reducing equivalents from NADPH to the active site of GltB. Within the large subunit
of GltB (135 kDa), the glutamine enters the glutaminase site of GltB which catalyzes the
formation of an enzyme-γ-glutamyl thioester through a cysteine residue. Subsequently,
the enzyme releases glutamate, and the ammonia is transferred through an
intramolecular tunnel to the synthase site. At the synthase site, ammonia attacks the α-
ketoglutarate and produces a 2-iminoglutarate which is then reduced to produce a
second glutamate (Suzuki & Knaff, 2005).
The complicated machinery of the high affinity pathway is more energy
demanding than the low affinity pathway. As such, a strong regulatory network is
required to minimize the energy expenditure of the cell to optimize growth. Significant
efforts have uncovered the intricate network of regulation that proteobacterial have
evolved to regulate the flux of metabolites through both pathways. The model has been
extended to include all gut organisms and is termed the enteric paradigm.
19 1.4.3 Mechanisms of Regulation. An elegant model of regulation has been proposed through studies of model organisms (Van Heeswijk et al., 2013). Global gene expression is also modulated primarily via NRI/NRII, a two-component system, and Nac which is itself transcriptionally regulated by NRI/NRII. Rapid repression or activation of enzymes occurs through the enzymatic modification of GS and GDH. The modulation of activity occurs through the regulatory proteins GlnB (also known as P-II), ATase
(adenylyltransferase/adenylyl-removing enzyme), and UTase
(uridylyltransferase/uridylyl-removing enzyme). In B. subtilis, the transcriptional regulators include TnrA and GlnR, which modulate transcription through protein-protein interaction with the functional enzymes (Fisher, 1999). Through very carefully fine-tuned enzymatic and regulatory pathways, organisms incorporate extracellular ammonium into intracellular α-ketoglutarate and glutamate to produce glutamate and glutamine, respectively.
In E. coli, the functional proteins are transcriptionally regulated by the aforementioned two-component system NRI/NRII, as well as several other regulators including Nac, CRP-cAMP, IHF, Lrp, and ArgR. The sensing protein, NRII, binds to ammonium and undergoes autophosphorylation, subsequently transferring the phosphate to the response regulator NRI. The phosphorylated NRI goes on to increase transcription of glnA, glnK, amtB, nac and other genes (Magasanik, 1989). Interestingly,
Nac represses gdhA without a co-effector molecule or covalent modification. The bacterium responds to amino acid deficit through Lrp by increasing transcription of gltBD, and to energy (ATP) deficit through Crp-cAMP inhibiting gltBD expression and
modulating a basal level of expression of glnA (Van Heeswijk et al., 2013).
20 In the model Firmicute, B. subtilis, three transcriptional regulators have been identified, GlnR, TnrA, and GltC (Fisher, 1999; Fisher & Wray Jr., 2002; Schumacher,
Chinnam, Cuthbert, Tonthat, & Whitfill, 2015; Wray Jr., Zalieckas, & Fisher, 2001). The
transcriptional activity of GlnR is mediated by the binding of glutamine synthetase,
stabilizing DNA interaction when bound, and is affected by pH. In contrast, TnrA is
inactive when bound to glutamine synthetase. Additionally, TnrA can be titrated away
from DNA by association with the membrane by interactions with GlnK and AmtB.
Finally, GltC is responsible for the activation of transcription of glutamate synthase
under increased glutamate demand during higher growth rates (Gunka & Commichau,
2012).
The P-II proteins are central to regulation of protein activity as they incorporate
signal from the intracellular metabolite pool and modulate enzymatic activity as well as
transcription (Figure 2). A P-II (GlnB) and a P-II like protein (GlnK) are encoded by glnB
and glnK, and glnK is commonly found adjacent to amtB (Arcondeguy, Jack, & Merrick,
2001; Blauwkamp & Ninfa, 2002; Detsch & Stulke, 2003; Forchhammer, 2008; van
Heeswijk et al., 1996). P-II proteins are homotrimers and possess binding sites for α-
ketoglutarate and ATP, as well as uridylylation sites by which UTase acts as an efficient
glutamine sensor. In addition, P-II proteins can undergo adenylylation in mycobacteria,
phosphorylation in cyanobacteria, or remain unmodified (Forchhammer, 2008; Gunka &
Commichau, 2012; Williams, Bennett, Barton, Jenkins, & Robertson, 2013). Several proteins directly interact with P-II including AmtB, ATase, NRII, and UTase in proteobacteria, as well as TnrA in B. subtilis. The transport of ammonium across AmtB is regulated by direct insertion of a P-II loop into the transport channel of the trimeric
21 AmtB, preventing ammonium transport. This interaction is inhibited by UTase uridylylation of P-II at the loop (Reitzer, 2003).
Glutamine synthetase type 1 (GS-1) activity is regulated via covalent modification
by the ATase in Proteobacteria. The ATase adenylylates a subunit of the homododecameric GS-1 and inactivates that subunit. Since a range of adenylylation states can exist (between 0-12), GS-1 can exist in a range of activities. ATase is also capable of activating GS-1 by the deadenylylation activity present within the same polypeptide. The regulation of adenylylation/deadenylylation is mediated by P-II interaction with the ATase. The regulatory activity of P-II towards ATase is dependent on its uridylylation state via the UTase ability to uridylylate or deuridylylate P-II. The
UTase uridylylation/deuridylylation activity is affected by glutamine and other small molecules (Figure 2). Finally, transcription of ammonium assimilatory genes is also affected by P-II through its interaction with NRII. Interaction between P-II and NRII are affected by the metabolites ATP and α-ketoglutarate, resulting in decreased autophosphorylation under energy and nitrogen abundance.
Although helpful in understanding how an organism modulates enzymatic activity through transcriptional and posttranslational means, this paradigm fails to explain gut nitrogen utilization. Extension of the enteric paradigm to the Bacteroidetes cannot be direct, as they lack homologs to the ammonium assimilation regulators, namely the two- component system (NRI/NRII), covalent modifiers (ATase/UTase), and the transcriptional regulators. In addition, conflicting reports exist concerning the dominant enzymatic activity that Bacteroidetes exhibit under varying nitrogen availability.
1.5 Ammonium assimilation in gut Bacteroidetes
22 For the purpose of this review, and in an attempt to determine a Genus and possibly Order-wide mechanisms for gut Bacteroidales to assimilate ammonium,
Bacteroides and Prevotella research into N-utilization will be treated as analogous in the context of the gut microorganisms. It should be noted that there is significantly more variation in nitrogen utilization patterns within Prevotella also evidenced by the genomic variations described in Chapter 2 (Bryant & Robinson, 1962; Nili & Brooker, 1995).
Research on the nitrogen utilization of Bacteroides and Prevotella species has been conducted since their isolation with original contributions from the laboratory of Dr.
Marvin P. Bryant. Nutritional characterization of Bacteroides and Prevotella has
repeatedly described the stimulatory effects or essentiality of ammonium as a main
nitrogen source (Bryant & Robinson, 1962; Bryant, Small, Bouma, & Chu, 1958; Nili &
Brooker, 1995; Pittman & Bryant, 1964; Varel & Bryant, 1974). The ability of Prevotella to scavenge ammonium is extremely high with an ammonium saturation constant of less than 50 µM (Schaefer, Davis, & Bryant, 1980). Results of these investigations have led to most of the medium formulations used today for all, including human, gut microorganisms.
Many Bacteroides/Prevotella species are capable of utilizing multiple nitrogen sources. Peptides are commonly supplemented to achieve maximal growth rates
(Griswold & Mackie, 1997). Bacteroidales also have amino acid uptake systems, which can feed into amino acid pools or even supplement growth for some Prevotella species
(R. M. Stevenson, 1979). Prevotella melaninogenica is also capable of fermenting amino acids and peptides (Wahren & Gibbons, 1970). Carbon skeletons from peptide degradation can be used for maintenance energy of Prevotella bryantii B14 (Russell,
23 1983). Nevertheless, proteolytic strains of Prevotella preferentially utilized ammonium
(Abou Akkada & Blackburn, 1963). When Prevotella are grown with 15N-labeled ammonium and peptides (1g/L) in the medium, 83% of total N was derived from ammonium and resulted in the formation of 15N-labeled amino acids within the cell
especially alanine and aspartate (Atasoglu, Valdes, Walker, Newbold, & Wallace, 1998;
Takahashi et al., 1980). Ammonium inhibits proteolytic activities (Sales-Duval, Lucas, &
Blanchart, 2002; Sales, Lucas, & Blanchart, 2000). Ammonium clearly serves as a
unifying feature within Bacteroidales and it is important to characterize these
assimilatory mechanisms.
The first work on ammonium assimilatory mechanisms of a Bacteroides was
published in 1984 describing the impact of the low affinity pathway (Yamamoto, Abe,
Saito, & Ishimoto, 1984). Yamamoto et al. describe the presences of glutamate
dehydrogenase activity capable of NADH and NADPH oxidation. The critical finding
here showed the use of GDH under excess and even more so under limiting ammonium
due to the increased GDH activity under N-limited batch and continuous cultures
(Yamamoto et al., 1984). These finding have also been demonstrated in later studies
with ruminal Prevotella species, P. ruminicola, P. brevis, and P. bryantii (Z. Wen &
Morrison, 1996, 1997).
Several studies showed the upregulation of GDH activity under low ammonium
concentration (Abrahams & Abratt, 1998; Abrahams, Iles, & Abratt, 2001; Baggio &
Morrison, 1996; Yamamoto et al., 1984; Yamamoto, Saito, & Ishimoto, 1987). Cultures
of B. fragilis growing under ammonium limiting conditions were spiked with a glutamine
synthetase inhibitor, methionine sulfoximine (Yamamoto et al., 1984) . The resulting
24 culture continued to grow and assimilate ammonium. These results highlight the primary deviation from the enteric paradigm set by E. coli, the increased intracellular activity of
glutamate dehydrogenase under ammonium limiting conditions.
Deeper investigation toward enzyme regulation reveals the organism synthesized
more GDH under low ammonium concentrations and rapidly regulated the enzyme
when pulsed with excess ammonium. The authors postulated the presence of post- translational modifications as the cause of reversible activation/inactivation mechanisms
(Yamamoto, Saito, et al., 1987). This trend is also true for ruminal isolates of Prevotella
including P. brevis GA233, P. bryantii B14, and P. ruminicola 23 (Z. Wen & Morrison,
1996, 1997). Baggio and Morrison identified the presence of a protein capable of
modulating activity of GDH downstream of the gene (Baggio & Morrison, 1996).
Analysis of the GDH enzymes shows that one GDH is specific for NADH while
the other, more dominant enzyme, is able to use NADH or NADPH. Kinetic analysis of
NADH and NADPH dependent ammonium assimilation revealed expectedly high Km
values for ammonium, 3.8 and 0.8 mM respectively (Yamamoto et al., 1984). Similar values were obtained when the enzyme was purified from cell extracts (Yamamoto,
Abe, & Ishimoto, 1987).
With the prevalence of genomic sequencing, the presence of two GDH enzymes has been found to be relatively ubiquitous in Bacteroides and in some Prevotella spp., encoded by gdhA (NAD(P)H-utilizing) and gdhB (NADH-specific). Analyzing expression patterns of both genes from B. fragilis Bf1 show the NADH-specific gdhB is expressed in response to growth on peptides, reflecting that NAD(P)H-dependent glutamate dehydrogenase is primarily responsible for the assimilation of ammonium under limiting
25 and non-limiting ammonium concentrations, while the NADH-specific glutamate
dehydrogenase produces α-ketoglutarate when glutamate is in excess (Abrahams &
Abratt, 1998; Abrahams et al., 2001).
To add further complexity, the gut Bacteroidales species typically possess multiple glutamine synthetase genes; B. thetaiotaomicron possesses three GS and two
GDH enzymes (Figure 4). A Bacteroides glutamine synthetase was first cloned in 1986 conferring ammonium utilization to a null mutant of E. coli (Southern, Parker, & Woods,
1986). Hybridization by the B. fragilis “glnA” to E.coli glnA showed the unique nature of the gene possessing no DNA homology to GS type 1. Further studies sequenced the gene revealing a novel family of GS significantly longer than the previously known GS1 and GS2 enzymes (Hill et al., 1989). Analyses of promoter transcription and enzymatic activity repeatedly demonstrated the transcription of glnN, from B. fragilis and P. bryantii
B14, is responsive to nitrogen concentration specifically expressing the gene under low
ammonium concentrations (Abratt, Zappe, & Woods, 1993; Kirk, Woodward, Ellis, &
Ricke, 2000; Z. T. Wen, Peng, & Morrison, 2003). Mutant strains of Prevotella bryantii
lacking GDH activity were still able to assimilate ammonium even though the growth
rate was slower. Attempts at identifying possible regulatory proteins for the B. fragilis
GS revealed only an elongation factor related protein, suggested to enhance the
translation of the gene (Abratt, Mbewe, & Woods, 1998).
Structural studies were performed on the GlnN of B. fragilis and revealed novel
features and architectures (J. van Rooyen, Belrhali, Abratt, & Sewell, 2011; J. M. van
Rooyen, Abratt, Belrhali, & Sewell, 2010; J. M. Van Rooyen et al., 2011; J.M. van
Rooyen, Abratt, & Sewell, 2006). The GlnN of Bacteroides is significantly larger than
26
Figure 4. Diagram of four genomic loci containing ammonium assimilation genes. Genes coding for proteins involved in ammonium assimilation are colored in yellow (high affinity pathway) and blue (low affinity pathway). Genes predicted to encode other nitrogen and carbon cycle genes are shaded in grey. Genes coding for hypothetical proteins and genes not associated with carbon or nitrogen metabolism are in white. Abbreviations: gdhA, glutamate dehydrogenase; glnA, glutamine synthetase Type 1; glnN, glutamine synthetase Type 3; gltB and gltD, glutamate synthase or GOGAT; amtB, ammonium transporter; glnK, P-II like protein.
27 the GS type1 and GS type 2 of bacteria, archaea, and eukaryotes, by 200 to 300 amino acids. Although GS3 folds and orients itself into a dimer of haxameric rings, like GS1 enzymes, they are inverted with the outside faces in GS1 forming the ring interfaces in GS3. The GS3 also lack known modification moieties for regulation. The energy demanding nature of glutamine synthetases likely requires a rapid inactivation mechanism through either covalent modification or allosteric inhibitors, however these features have not yet been identified.
Enzymatic characterization of the three glutamine synthetases of P. ruminicola
23 describes kinetic properties which are in agreement with known glutamine
synthetases. The enzymatic affinity for ammonium is high, with Kms of 0.48 and 0.43
mM for the two GS type 3 enzymes (J.N. Kim, Cann, & Mackie, 2012). In opposition to
previous research, the most recent report of ammonium assimilation in Prevotella
describes the transcriptomic response of P. ruminicola 23 to be entirely opposed to the
enteric paradigm. The GSIII-2 is more transcriptionally abundant when the organism is
grown in non-limiting as compared to limiting concentration of ammonium (J. N. Kim et
al., 2017).
In summary, three major discoveries from Bacteroides research have revealed a widely different nature of ammonium assimilation and regulation from the known enteric model. This potential paradigm shift includes (1) increased glutamate dehydrogenase
activity under nitrogen limitation, (2) increased glutamine synthetase transcription under
ammonium excess, and (3) an entirely new family of glutamine synthetases. Features of
regulation are critical to ascertain before a “true” enteric model can be applied to
microbial ammonium assimilation within the human or animal gastrointestinal tract. In
28 addition, the multiplicity of functional genes needs to be addressed to identify the roles for each putative glutamine synthetase and glutamate dehydrogenase encoded within gut Bacteroidales. Although investigations into ammonium assimilation within
Bacteroidetes are limited, the breadth of the information primarily describes deviations from the classical enteric paradigm. Without more foundation-building research into mechanisms employed by predominant gut microorganisms, the enteric paradigm of ammonium assimilation cannot accurately reflect nitrogen acquisition in mammalian systems.
1.6 Proposition for research
In our lab, work on various gut bacteria including Bacteroides intestinalis DSM
17393 (Hong et al., 2014), Prevotella ruminicola 23 (Dodd et al., 2009; Kabel et al.,
2011; J. N. Kim et al., 2012), Prevotella bryantii B14 (Dodd, Kiyonari, Mackie, & Cann,
2010; Dodd, Moon, Swaminathan, Mackie, & Cann, 2010), and Ruminococcus albus 8
has been recently carried out (Iakiviak, Mackie, & Cann, 2011; J.N. Kim, Henriksen,
Cann, & Mackie, 2014; Moon, Iakiviak, Bauer, Mackie, & Cann, 2011). In particular, N-
metabolism in gut bacteria has become an important research question driven by a lack
of understanding of gene function and regulation in gut bacterial nitrogen metabolism.
Recent work on P. ruminicola 23 (J. N. Kim et al., 2012) and R. albus 8 (J.N. Kim et al.,
2014) led to the enzyme characterization of ammonium assimilatory proteins and
demonstration of a P. ruminicola glutamine synthetase III (GSIII-2) up regulation in
response to non-limiting ammonium concentrations. This is contrary to the enteric
paradigm, where GS is up regulated during growth under ammonium limiting conditions.
These results have stimulated the search for definitive answers on the regulatory and
29 biochemical mechanisms of ammonium assimilation in gut microbes. To further characterize nitrogen metabolism at the genetic level, we propose to utilize B. thetaiotaomicron, a predominant member of the human colonic bacterial community that possesses the advantage of being amenable to genetic manipulations (Koropatkin,
Martens, Gordon, & Smith, 2008).
30 Chapter 2. Comparative genomic analysis for phylogenetic conservation of
ammonium assimilation genes within Bacteroides and Prevotella
2.1 Introduction
In general, ammonium assimilation occurs via two competing pathways specifically induced under different environmental conditions. Within gut systems, the
enteric paradigm describes the preferential use of a low affinity system when
ammonium is in excess and a high affinity system when ammonium is limiting. The low
affinity pathway consists of a single enzyme, glutamate dehydrogenase (GDH), which
appends ammonium onto an α-ketoglutarate (α-KG) through reductive amination to
yield glutamate. The glutamate can then be used to produce amino acids and other nitrogenous compounds within the cell. Previous reports identified the GDH to be predominantly expressed under ammonium excess in E. coli and K. pneumoniae
(Camarena, Poggio, Garcia, & Osorio, 1998; Rosario & Bender, 2005). In Bacteroides and Prevotella species, the GDH is constitutively employed but activity increases when ammonium concentration decreases (Abrahams & Abratt, 1998; Abrahams et al., 2001;
Baggio & Morrison, 1996; Z. Wen & Morrison, 1996, 1997; Yamamoto et al., 1984;
Yamamoto, Saito, et al., 1987). Alternatively, glutamate dehydrogenase can be used
catabolically to deaminate glutamate, releasing ammonium and an α-KG for carbon
catabolism. The catabolic GDH utilizes NADH, while the anabolic GDH oxidizes NADPH
(Van Heeswijk et al., 2013).
The high affinity strategy requires the concerted action of two different enzymes in the assimilatory pathway. First, ammonium is appended onto glutamate via glutamine
31 synthetase (GS), producing glutamine with energy from ATP and driving the reaction forward under low concentrations of ammonium. Second, the NADPH-consuming reductive transamination of glutamine and an α-KG into two molecules of glutamate is catalyzed by glutamate synthase (GOGAT). Three divergent families of glutamine synthetases exist with two families commonly identified in gut bacteria, the GS type 1
(GS1) and GS type 3 (GS3). The GS type 1 is a classical proteobacterial enzyme, extensively characterized in E. coli, Salmonella, and Klebsiella, specifically regarding their catalytic mechanisms and regulation (Van Heeswijk et al., 2013). The GS type 3 enzymes were later identified in Bacteroides fragilis and shown to possess novel structural features and significant sequence divergence from the GS1s (Southern et al.,
1986). To date, little is known about the regulatory nature of GS3 genes or their protein products.
Additionally, the high affinity system requires the facilitated transportation of ammonium when diffusion of gaseous ammonium is insufficient. The ammonium transporter, AmtB, allows for the internalization of exogenous ammonium. A small protein, termed P-II and encoded by glnK, is responsible for the physical regulation of
AmtB to avoid ammonium flooding into the cytoplasm. In the enteric paradigm, intercellular signals mediate the physical blockage of AmtB by P-II when the nitrogen status is adequate.
Classically, identification of enzymes involved in ammonium assimilation was performed using ‘whole cell’ techniques. Enzymatic activity was screened on cell lysates and function was therefore confirmed. Glutamate dehydrogenase activity in Bacteroides fragilis was demonstrated to dominate in non-limiting concentrations of ammonium and
32 increased in activity when ammonium became limiting (Abrahams & Abratt, 1998;
Abrahams et al., 2001; Yamamoto et al., 1984; Yamamoto, Saito, et al., 1987). With the
advent of genetic and molecular biological techniques, genes involved in ammonium
assimilation were identified, cloned, and expressed. The glutamate dehydrogenase and
glutamine synthetase of Bacteroides spp. were cloned and complemented an E. coli
strain lacking the ammonium assimilation pathways (Baggio & Morrison, 1996; Southern et al., 1986). Finally, with genome sequences available, the genomic context provided further information on the mechanism of ammonium assimilation.
As the genomic era expanded the knowledge of gene content in related organisms, the ability to generalize information gleaned from studying type strains became possible. Using Bacteroides thetaiotaomicron as a model gut microbe has proven useful in the study of the ubiquitous polysaccharide utilization loci (PULs) present within most gut Bacteroides and Prevotella species (Anderson & Salyers, 1989;
Dodd, Moon, et al., 2010). Similar approaches towards the analysis of ammonium
assimilatory genes are essential to understand the mechanism by which gut
Bacteroidales assimilate nitrogen.
Several ammonium assimilation enzymes have been identified in the Bacteroides
and Prevotella genera, including glutamate dehydrogenases with dual specificity for
NADH and NADPH, as well as an NADH-specific GDH. The dual-cofactor enzyme is
theorized to be the primary ammonium assimilatory enzyme under high and low
ammonium concentrations, while the NADH-specific enzyme is increased during peptide utilization with a catabolic role (Abrahams & Abratt, 1998; Abrahams et al.,
2001; Yamamoto, Saito, et al., 1987). The multiplicity of GS enzymes has been a point
33 of confusion for the assimilatory mechanism, as Prevotella and Bacteroides species
commonly carry a GS1 and two GS3 genes in their genomes. Enzymatic
characterization of the three GS enzymes does not reveal any obvious functional differences, but transcriptional analysis reveals that a single GS3 is responsive to ammonium concentration (J. N. Kim et al., 2012; J. N. Kim et al., 2017). The genomic context of Prevotella ruminicola nitrogen assimilation genes reveals a co-localization of the ammonium responsive catalytic GS3 with the ammonium transporter and the PII regulatory gene in the vicinity of the gltBD, encoding the glutamate synthase complex. A similar organization is identified within Bacteroides thetaiotaomicron.
Gut Bacteroides constitute an ideal model group to lay the foundations of our understanding on the fundamental principles that determine how our core gut bacteria colonize and persist in their host. First, they live and grow exclusively in the gastrointestinal tracts of animals, suggesting strong adaptation to the gut environment.
Second, as commensal and mutualistic bacteria, they can establish stable, long-term associations with their human hosts. Third, they are major constituents of the microbiota across human population, and constitute a foundational part of the microbial food webs
(A. G. Wexler & Goodman, 2017). We used the gut bacterium Bacteroides thetaiotaomicron, the first member of the Bacteroides to have its genome sequenced, and with a genetic system available, to model the gut Bacteroidales ammonium assimilation pathway. To determine the suitability of the model, it is essential to identify the conservation of ammonium assimilation genes in the most abundant gut
Bacteroidetes. To predict whether these genes and their genomic orientation are conserved, comparative genomics was used to search all sequenced genomes from
34 Bacteroides and Prevotella spp. To establish the catalog of genes involved in this pathway, we performed a complete analysis of genomic loci associated with ammonium assimilation genes from Bacteroidales. Within the Bacteroides genus, most species
maintain the conservation of three glutamine synthetases, two GS3s and one GS1, with
some exceptions. Conversely, the Prevotella genus shows little conservation, except
within ruminal fibrolytic species, possessing a similar multiplicity of genes. These
observations revealed Bacteroides thetaiotaomicron as a suitable model organism for
identifying the ammonium assimilation mechanism; as most gut Bacteroidales retain
similar genomic features.
2.2 Materials and Methods
Using previously identified genes implicated in ammonium assimilation, the
genome of B. thetaiotaomicron VPI-5482 was uploaded onto mgRAST (Overbeek et al.,
2014). Next, the genome sequence was used as the database and a BLAST search
was performed using the known genes as query sequences. A list of loci was compiled
for prediction of function in ammonium assimilation (Figure 4). The genes predicted to
be involved in ammonium assimilation of Bacteroides thetaiotaomicron were used as
bioinformatic queries to identify homologous genetic loci in other genomes of
Bacteroides and Prevotella.
Using Geneious v.9.0.2 (http://www.geneious.com, (Kearse et al., 2012)), all of
the available Bacteroidetes genomes were downloaded, including 197 Bacteroides and
122 Prevotella genomes. The Bacteroides and Prevotella genomes were aligned in
batches to B. thetaiotaomicron VPI-5482 or P. ruminicola 23 respectively, using Mauve
v.2.3.1 (http://darlinglab.org/mauve). The P. ruminicola genome was used as the
35 reference genome for Prevotella species, as it encodes homologs for all of the ammonium assimilation genes of B. thetaiotaomicron. A total of 16 alignments covered the 197 Bacteroides genomes, and 10 alignments represented 122 Prevotella genomes.
Using Mauve allowed for manual analysis of genomic variations throughout related organisms. Alignments were manually screened to identify the genomic loci for ammonium assimilation in all other genomes of Bacteroides and Prevotella species by
searching for homology in DNA sequence and bioinformatic annotations. Visualizing
and heat map generation was performed using the heatmap.2 command from the gplots
package in R (Team, 2004).
2.3 Results and Discussion
An initial hypothesis centered around the complexity of Prevotella ruminicola as a
gut relative to the Bacteroides. Ruminal Prevotella spp. harbor a multiplicity of genes,
and we expectedly found a similar genotype in B. thetaiotaomicron. Through
bioinformatic searches using MG-RAST, we identified the presence of a putative high
affinity pathway, coding for three glutamine synthetases (glnN1, glnN2, and glnA), a
glutamate synthase (gltB and gltD), an ammonium transporter (amtB) and a PII protein
(glnK) (Figure 4). We also identified the low affinity pathway coding for two glutamate dehydrogenases (the putatively NAD(P)H-specific gdhA and the putatively NADH- specific gdhB). Of the glutamine synthetases, one encodes a glutamine synthetase type
1 (GS1, glnA), which is homologous to the classical E. coli glnA. In addition, there are two genes encoding glutamine synthetase type 3 enzymes originally identified in
Bacteroides fragilis (Southern et al., 1986). This enzyme family is commonly found in gut organisms including Bacteroidetes (Bacteroides and Prevotella) and Firmicutes
36 (Ruminococcus) (J.N. Kim et al., 2014; Z. T. Wen et al., 2003). The multiplicity of the
glutamate synthases reflects the importance of this function to gut bacteria. There may
be multiple modes of regulation controlling the employment of each GS under specific conditions. Alternatively, the predicted genes coding for glutamine synthetases may be misannotated and perform different function within the bacterium.
Most notably, the high affinity pathway is localized on the genome in one large
locus encoding glnN1, amtB, glnK, and nearby, gltBD, as well as several genes for
amino acid biosynthesis, including lysine (dapL and dapF) and asparagine (asnB),
termed the High Affinity Assimilatory Locus or (HAAL). This co-localization can imply the
co-expression of these genes in response to a nitrogen requirement. Bacteroides
commonly co-localize and co-express genes in response to a polysaccharide carbon
source. These genes are organized in polysaccharide utilization loci (Schwalm &
Groisman, 2017). Our observations on the genomic organization of conserved
ammonium assimilation genes in Bacteroides spp. leads us to hypothesize that these
organisms may employ a similar regulatory mechanism for ammonium assimilation.
To understand the conservation and evolutionary importance of the ammonium
assimilatory genes, comparative genomics was used to thoroughly search sequenced
genomes of Bacteroides and Prevotella. In order to remove phylogenetic bias, all
sequenced genomes were aligned to the B. thetaiotaomicron genome and gene
conservation was screened by manually scanning the loci that are homologous to those
of B. thetaiotaomicron (Table 1). The Bacteroides alignment reveals a strong
conservation for preserving all of the genes predicted to assimilate ammonium (Figure 5
and Figure 6). In general, the genes in the large locus encoding the high affinity
37 Table 1. Genes used as bioinformatic queries.
Locus Tag Gene name Predicted function Gene Length High affinity locus (HAAL) 1 Chromosomal region 667,987 … 684,917 BT_0543 glnN1 Glutamine synthetase type 3 2,190 bp BT_0544 amtB Ammonium transporter 1,458 bp BT_0545 glnK PII regulatory protein 357 bp BT_0546 Hypothetical 726 bp BT_0547 dapL LL-DAP aminotransferase (glutamate-dependent) 1,233 bp BT_0548 dapF Diaminopimelate epimerase 804 bp BT_0549 Hypothetical 297 bp BT_0550 Glycerophosphodiester phosphodiesterase 738 bp BT_0551 asnB Asparagine synthetase (glutamine-dependent) 1,680 bp BT_0552 gltD Glutamate synthase SS 1,341 bp BT_0553 gltB Glutamate synthase LS 4,611 bp High affinity locus 2 Chromosomal region 973,619 … 975,121 BT_0785 glnA Glutamine synthetase type 1 1,503 bp High affinity locus 3 Chromosomal region 5,714,701 … 5,716,890 BT_4339 glnN2 Glutamine synthetase type 3 2,190 bp Low affinity locus Chromosomal region 2,472,243 … 2,484,264 BT_1969 Malic enzyme 2,292 bp BT_1970 gdhA Glutamate dehydrogenase (NADPH-dependent) 1,335 bp BT_1971 Catalase 1,467 bp BT_1972 PEP synthase 2,973 bp BT_1973 gdhB Glutamate dehydrogenase (NADH-dependent) 1,338 bp BT_1974 Aminopeptidase 1,164 bp
38
Figure 5. Phylogenetic conservation of loci encoding high affinity genes in Bacteroides species. Synteny of the high affinity system of Bacteroides thetaiotaomicron VPI-5482 was determined against sequenced Bacteroides genomes excluding B. fragilis. Genome alignments were performed in Mauve v. 2.3.1 and genes were manually categorized based on gene homology and locus synteny in the High Affinity Locus 1 (HAAL). Phylogenetic relationship of Bacteroides species depicted on the left is based on 16S sequences retrieved from www.arb-silva.de. Blocks of light blue signify syntenic conserved regions and blocks of dark blue signify conserved but non- syntenic genes. Blocks in dark grey describe genes absent from the genome sequence.
39
Figure 6. Phylogenetic conservation of the locus encoding low affinity genes in Bacteroides sp. Synteny of the low affinity system of B. thetaiotaomicron VPI-5482 was determined against sequenced Bacteroides genomes excluding B. fragilis. Blocks of light blue signify syntenic conserved regions and blocks of dark blue signify conserved but non-syntenic genes. Blocks in dark grey describe genes absent from the genome sequence.
40 pathway are conserved within Bacteroides, but not in strict synteny. The contiguous regions, shown in light blue, are relative to the glnN1, and a majority of the genomes encode the glnN1, amtB, and glnK genes syntenically, whereas the gltBD genes are found either in a syntenic orientation or in trans to the glnN1. As shown by the transcriptional data presented in Chapter 3, the co-localization of these genes is not essential, as they are transcribed in different operons.
The genomes presented in Figure 5 and Figure 6 are arranged in phylogenetic
order based on a 16S phylogenetic tree, using a cluster of two Parabacteroides
sequences as the out-group. Viewing the conservation based on phylogeny, there is a
clear distinction between members of the Bacteroides that are distantly related. Their
genomes display a greater divergence of gene conservation. For example, Bacteroides
pyogenes, B. coporsuis and B. propionicifaciens are phylogenetically distant to B.
thetaiotaomicron and there is an absence of genes in the high affinity locus and the
glutamine synthetase type 1. Other distantly related bacteria, including B.
paurosaccharolyticus, B. plebius, B. coprocola, B. salanitronis, B. coprophilus, B.
barnesiae, and B. reticulotermitis, also lack the glnA. While glnN2 is more conserved
and only absent in B. paurosaccharolyticus, which does contain a glnN1 homolog.
Therefore, the importance of individual gln genes based on conservation through
comparative genomics remains unclear.
When phylogenetic conservation of the glutamate dehydrogenases is analyzed,
strong conservation is observed. For instance, the putatively NADH-dependent gdhB is
strictly conserved, highlighting the importance of this enzyme. Although gdhA is nearly
completely conserved, it is absent in B. coprosuis and B. propionicifaciencs. As the gdh
41
Figure 7. Conservation of genes encoded by the high affinity loci in Bacteroides fragilis strains. Genes of the high affinity system of Bacteroides thetaiotaomicron VPI- 5482 were used a query against sequenced Bacteroides fragilis genomes. Genome alignments were performed in Mauve v. 2.3.1 and genes were manually categorized based on conservation. Blocks of light blue signify conserved genes and blocks in dark grey describe genes absent from the genome sequence.
42
Figure 8. Conservation of genes encoded by the low affinity locus in Bacteroides fragilis strains. Genes of the low affinity system of B. thetaiotaomicron VPI-5482 were used a query against sequenced B. fragilis genomes. Genome alignments were performed in Mauve v. 2.3.1 and genes were manually categorized based on conservation. Blocks of light blue signify conserved genes and blocks in dark grey describe genes absent from the genome sequence.
43 system is much more highly conserved compared to the high affinity system, we hypothesize the GDH to be the primary pathway for ammonium assimilation, and will be discussed further in later chapters. Other research has described the GDH pathway of
Bacteroides to be crucial in ammonium assimilation under all ammonium conditions
(Abrahams & Abratt, 1998; Abrahams et al., 2001; Baggio & Morrison, 1996; Z. Wen &
Morrison, 1996, 1997; Yamamoto, Abe, et al., 1987; Yamamoto et al., 1984; Yamamoto,
Saito, et al., 1987).
The B. fragilis genomes were also analyzed for their genomic conservation of
ammonium assimilatory genes. Results from this species differ drastically from most
other Bacteroides genomes (Figure 7 and Figure 8). The conservation of the high
affinity pathway was much lower, with only nine out of 80 genomes retaining the
catalytic GS/GOGAT system, with three genomes lacking the amtB/glnK genes. In
contrast to glnN1, the conservation of glnN2 and glnA was universal in B. fragilis. If the
gltBD genes are an essential component to the high affinity pathway, the glutamine
synthetase genes with parallel conservation should accompany the GOGAT complex in
forming the high affinity system. The glnN1 is co-conserved with gltBD in B. fragilis,
therefore strengthening the hypothesis that Locus 1 is the critical gene system
employed for ammonium assimilation under nitrogen limitation. In opposition to the poor
conservation of the putative high affinity locus, the genes encoding the glutamate
dehydrogenase enzyme are universally conserved (Figure 8). As B. fragilis is commonly
identified and isolated as an opportunistic pathogen, the primary source of nitrogen for
this organism may not be ammonium, and therefore conservation of the high affinity
system is not critical (H. M. Wexler, 2007).
44
Figure 9. Conservation of ammonium assimilatory genes in Prevotella species. Genes of the high and low affinity system of B. thetaiotaomicron VPI-5482 were used a query against sequenced Prevotella genomes. Genome alignments were performed in Mauve v. 2.3.1 and genes were manually categorized based on conservation. Blocks of light blue signify conserved genes and blocks in dark grey describe genes absent from the genome sequence. The blocks of P. copri DSM 18205 colored red describe a gene duplication.
45 The genomic analysis within Prevotella species was performed using the P. ruminicola genome as the reference genome for alignments, because of the parallel conservation of genes to B. thetaiotaomicron. Unlike the Bacteroides alignments, which display a moderate level of gene conservation, there was very little conservation of genes involved in ammonium assimilation (Figure 9). Only the eight genomes, including
P. ruminicola 23 and Ga6B6, P. copri, P. albensis, P. paludivivens, P. bryantii, P.
multisaccharivorax, Prevotella sp. P6B1, and Prevotella sp. 109, retained the High
Affinity Locus 1. Almost all Prevotella spp. investigated lack the multiplicity exhibited by
ruminal Prevotella and Bacteroides species. In addition, there is no universal
conservation of any ammonium assimilation gene, although the NADH-dependent GDH
encoding gene was the most conserved. This effect is likely due to the source of
isolation. Human Prevotella spp. have been mostly isolated from the oral cavity and
many are not strictly gut organisms, where other N-sources such as proteins and
peptides are readily available. The genomes likely reflect the lower importance of ammonium assimilation in non-gut Prevotella species.
Through the bioinformatic approach, we reveal the conservation of potentially essential nitrogen utilization genes and their organization among gut Bacteroidales.
Although Bacteroides species, excluding B. fragilis, tend to maintain both the high affinity and low affinity pathways, the conservation in Prevotella is low. Our comparative genomic analyses show that dominant commensal gastrointestinal microbes preferentially retain both ammonium assimilation pathways. Additionally, the conservation of the “low affinity” system is stronger in commensal Bacteroides species and opportunistic B. fragilis strains, implicating GDH to be the predominant method of
46 ammonium assimilation in the gastrointestinal tract, where ammonium is rarely found in limiting concentrations.
47 Chapter 3. Nitrogen utilization patterns of B. thetaiotaomicron and genome wide
transcriptional response to extracellular ammonium
3.1 Introduction
Ammonium assimilation in gut microbes is the major contributor to bacterial
nitrogen acquisition. Under normal feeding conditions, ammonium is continuously
released from the breakdown of proteins and urea, and is subsequently responsible for
providing between 42% and 100% of the microbial nitrogen (Bach et al., 2005; Firkins,
Yu, & Morrison, 2007; Nocek & Russell, 1988; Pilgrim, Weller, Gray, & Belling, 1970;
Wallace et al., 1997). Lactating dairy cattle can survive on ammonium salts and urea as
the sole nitrogen sources without exogenous protein supplementation (Virtanen, 1966).
Under these feeding conditions, the animals were sustain ned throughout lactation and
were reliant exclusively on microbial proteins to supply the amino acid requirements of
the host. Bacteria within mammalian hindguts assimilate ammonium to a similar
capacity, as characterized within the rabbit cecum. In total, 55% of cultivable species
demonstrate the ability to utilize ammonium as the sole nitrogen source (Forsythe &
Parker, 1985). While most members of Bacteroides genus preferentially utilize
ammonium over other nitrogen sources, the machinery employed to acquire nitrogen
remain poorly understood.
Mechanisms of ammonium assimilation have been well studied in the
proteobacteria Escherichia coli, Klebsiella, and Salmonella, resulting in the derivation of an “enteric model” (Van Heeswijk et al., 2013). The proposed model describes two assimilatory pathways, differentially utilized under specific exogenous ammonium concentrations, termed a low affinity pathway and the high affinity pathway. At
48 + physiological pH (6.5-7.5), ammonium exists primarily in the protonated form (NH4 ), but
1% also exists in the deprotonated gaseous form of ammonia (NH3), with the potential of traversing the cell membrane via diffusion. Under excess ammonium, bacteria employ the low affinity system. Composed of a single enzyme, glutamate dehydrogenase (GDH) is responsible for the incorporation of ammonium into an α-
ketoglutarate (α-KG) forming glutamate, with a high Km for ammonium (~1 mM). In the process, the GDH consumes a reducing equivalent in the form of NADPH.
Alternatively, under limiting ammonium concentrations (<1 mM), gaseous ammonia is too low to sufficiently traverse the membrane and a transport system is induced, the ammonium transporter AmtB. This trimeric membrane protein facilitates the internalization of ammonium into the cytoplasm. Within the cytoplasm, glutamine synthetase (GS) appends ammonium onto glutamate, using an ATP, and releases glutamine. Subsequently, glutamate synthase (GOGAT) transfers the amine from glutamine onto α-KG, while oxidizing NADPH, to produce two glutamates.
The dominant gut phylum Bacteroidetes presents challenges in conforming to the presented enteric paradigm. First, in Bacteroides and Prevotella species, glutamate dehydrogenase activity increases when cells are subjected to ammonium limitation
(Baggio & Morrison, 1996; Z. Wen & Morrison, 1996, 1997; Yamamoto et al., 1984;
Yamamoto, Saito, et al., 1987). Second, there is disagreement as to how glutamate
synthetases are regulated in Prevotella species. Under limiting concentrations of
ammonium, glutamine synthetase activity has been demonstrated to increase in
Prevotella bryantii B14 (Z. T. Wen et al., 2003), while more recently Prevotella ruminicola 23 has been demonstrated to increase transcription of GSIII-2 when grown in excess ammonium (J. N. Kim et al., 2012; J. N. Kim et al., 2017). To date, little
49 information is known as to the regulatory mechanisms for ammonium assimilation in these organisms.
As described in Chapter 2, the multiplicity of genes adds further complication.
Definitively identifying which gene or genes are responsible for ammonium assimilation is required, in addition to validating the preferred pathway under specific conditions.
Towards such goals, previous publications used transcriptional analysis in order to identify gene products responsive to ammonium concentration. Prior to the RNA-seq and microarray technologies, qualitative assays (gene fusion reporter and northern blot analysis) describe increased transcription of glutamine synthetase type 3 from
Bacteroides and Prevotella species under ammonium limitation (Abratt et al., 1993; Z.
T. Wen et al., 2003). The transcription of the low affinity gdhA gene of Prevotella bryantii
B14 is inversely regulated to the high affinity glnN genes. Under ammonium-limited
condition, the gdhA is decreased in transcription and significantly decreased when grown on peptides, while it is increased during ammonium excess (Z. Wen & Morrison,
1996, 1997). Unlike Prevotella bryantii B14, the gdhA of B. thetaiotaomicron is constitutively transcribed and not responsive to ammonium concentration, based on northern blot analysis (Baggio & Morrison, 1996). To rectify the differences regarding
Bacteroidetes ammonium assimilation patterns, more detailed genetic techniques and more accurate quantification of transcript abundance are required to describe the precise mechanisms of ammonium assimilation under well-defined environmental conditions.
Continuous cultures possess the primary advantage of control over nutrient concentrations, including ammonium, at a predetermined growth rate. Using organisms like Escherichia coli, Klebsiella aerogenes, and Salmonella typhimurium, analyses of
50 organisms grown in continuous cultures describe multiple aspects of ammonium assimilation: (1) how ammonium assimilatory pathways are employed; (2) how ammonium is transported; (3) and what are the mechanisms for regulation (Atkinson,
Blauwkamp, Bondarenko, Studitsky, & Ninfa, 2002; Cole, Coleman, Compton,
Kavanagh, & Keevil, 1974; M. Kim et al., 2012; Senior, 1975; Soupene et al., 2002;
Straight & Ramkrishna, 1994; Wang, Lai, Ouyang, & Tang, 2011). In ruminal organisms, including Prevotella bryantii B14, use of continuous cultures allowed for the quantification of ammonium saturation constants, describing the high affinity for ammonium of predominant ruminal isolates (Schaefer et al., 1980).Global transcriptional analysis of Mycobacterium smegmatis grown in continuous culture identified global regulatory patterns controlled by nitrogen status (Petridis, Benjak, & Cook, 2015).
The primary goal of this study was to determine the nitrogen utilization patterns
of Bacteroides thetaiotaomicron. The working hypothesis invokes GDH enzymes expressed constitutively, while preferentially employing the GS/GOGAT system based on external ammonium concentrations. In support of this hypothesis, nutritional studies performed on B. thetaiotaomicron determine the preferential utilization of ammonium for cell nitrogen. The transcriptional patterns of B. thetaiotaomicron have been shown to be highly dynamic and acutely responsive to environmental nutrient concentrations
(Martens et al., 2011). Our hypothesis predicts that B. thetaiotaomicron will have specific transcriptional responses when grown under ammonium limitation as compared to ammonium excess conditions. Initially, batch-grown bacterial cultures utilizing ammonium, peptides, or amino acids as nitrogen sources confirms the preference of B. thetaiotaomicron towards ammonium. Subsequently, continuous culture techniques provide controlled and consistent growth rates of B. thetaiotaomicron with constant
51 ammonium concentrations. Continuous cultures stabilized at a limiting ammonium concentration (<0.05 mM) and excess ammonium concentration (>15 mM) were
sampled for global transcriptional responses toward substrate concentrations.
3.2 Materials and Methods
Materials. Anaerobic growth was performed in a vinyl anaerobic chamber from
Coy Laboratory Products (Gass Lake, MI). Purification of RNA was performed using
Qiagen RNeasy Mini Kit with RNAprotect Bacteria Reagent (Qiagen, Germantown MD).
All other reagents were obtained from Sigma Aldrich.
Bacterial strains and growth conditions. B. thetaiotaomicron VPI-5482 Δtdk
was obtained from Dr. Eric Martens of the University of Michigan, Ann Arbor. Anaerobic
bacteria were stored in agar slants in the vapor phase (-110°C) of a liquid N2
cryostorage system as well as 30% glycerol stocks at -80°C. All other materials were
obtained from Sigma Aldrich. For batch cultures, B. thetaiotaomicron was grown in an
anaerobic cabinet with a headspace of 95% CO2 / 5% H2. Initially, isolated colonies
were grown in Brain Heart Infusion Supplemented medium (BHIS, Appendix A:Table 1),
preconditioned via three passages in chemically defined media (CDM, Appendix
A:Table 2) in the respective final nitrogen source, and finally inoculated into CDM
medium with either ammonium sulfate (1 or 10mM) or a nitrogen equivalent
concentration of peptides (Bacto Tryptone, 0.25% or 0.025%). After cells were
preconditioned in the specific nitrogen condition, cells were sub-cultured at 20-fold
dilution into the corresponding nitrogen condition and absorbance measurements were
taken at 600nm using a Spectronic 20D+. Cultures were sampled for ammonium
concentration by centrifuging (5 min. @ 16,000 x g) a 200 µl aliquot, the supernatant
was stored at -20°C until analyzed.
52
Table 2. Growth parameters of B. thetaiotaomicron VPI-5482 on non-limiting concentrations of various nitrogen sources.a
Growth Rate Doubling Max OD600 (μ) time (hr) Non-limiting Ammonium 0.346 ± 0.05 2.05 ± 0.33 1.15 ± 0.019 Peptides 0.291 ± 0.09 2.23 ± 0.46 1.32 ± 0.006 Amino acids 0.355 ± 0.05 1.99 ± 0.28 0.981 ± 0.011 Limiting (single-N) Peptides 0.514 ± 0.18 1.53 ± 0.63 0.445 ± 0.005 Ammonium 0.386 ± 0.14 2.13 ± 0.79 0.892 ± 0.018 Amino acids 0.249 ± 0.10 3.21 ± 1.33 0.277 ± 0.019 Casein 0.193 ± 0.04 3.76 ± 0.84 0.410 ± 0.011 Limiting (mixed-N) + Peptides / NH4 0.417 ± 0.15 2.12 ± 1.58 0.623 ± 0.011 + AA / NH4 0.306 ± 0.11 2.58 ± 1.05 0.572 ± 0.012
Casein / NH4+ 0.319 ± 0.1 2.38 ± 0.78 0.727 ± 0.003
aGrowth experiments were performed in triplicates, results presented as mean ± SD
53 Continuous culture. B. thetaiotaomicron VPI-5482 Δtdk was grown for 20 days in a continuous culture system in 20% CO2 / 75% N2 / 5% H2 headspace at a dilution
rate of 0.166 h-1. A modified CDM medium (Appendix A:Table 2) was formulated and
contained high (10 mM) and low (1 mM) concentrations of ammonium sulfate. The
culture vessels were constructed of cell culture flasks (Wheaton Industries Inc., Millville
NJ). Modified CDM (Appendix A:Table2) was fed at a dilution rate of 0.17 h-1 with Tygon
tubing using a peristaltic pump (Cole-Palmer, Vernon Hills IL).
The continuous cultures were run in two consecutive phases, culture vessels #1
and #2 were fed for 10 days with low ammonium medium while vessels #3 and #4 were
fed with high ammonium medium. At day 10, the input media were switched and
vessels #1 and #2 were fed high ammonium medium while vessels #3 and #4 were fed
low ammonium medium. One mL of outflow was collected daily for absorbance (600
nm) and pH measurements. The culture was then centrifuged to remove cell debris and
the supernatant was stored at -20°C until further analysis. Samples were collected for
transcriptomic analysis at days 7, 8, 9, 17, 18, and 19.
Ammonia assay. Ammonium was quantified using a classic colorimetric assay
by Chaney and Marbach (Chaney & Marbach, 1962). Briefly, two reagents were
prepared and stored in the dark: Reagent A, 50 g/L phenol and 0.25 g/L sodium
nitroprusside; and Reagent B, 25 g/L sodium hydroxide and 2.1 g/L sodium
hypochlorite. Prior to the assay, samples were centrifuged (5 min. @ 16,000 x g) to
remove any precipitate or cells. A sample volume of ten microliters was added to one
milliliter of Reagent A and vortexed and then one milliliter of Reagent B was added and
vortexed in quick succession. The samples were incubated at room temperature for 30
54 minutes and absorbance at 630 nm was taken. Samples were compared to a standard
+ curve of ammonium chloride to determine final NH4 concentrations.
Glucose and VFA analysis. To quantify glucose, a Dionex ICS-5000+ Ion
Chromatography System (Thermo Fisher, Waltham MA) was used to separate sugars on an Dionex CarboPac PA100 (Thermo Fisher, Waltham MA) and the glucose was detected using pulsed amperometric detector. The retention time and peak area were compared to a standard curve of known glucose concentrations.
To quantify volatile fatty acid concentrations, the samples were subjected to high performance anion exchange chromatography using a Shimadzu Prominence HPLC system (Kyoto, Japan) and an Aminex HPX-87H (Biorad, Hercules CA) column to separate specific VFAs (formate, propionate, ethanol, succinate, acetate), which were detected with a refractive index detector. The resulting retention times and peak areas were compared to those of known standards and concentrations.
Transcriptional analysis. During Day 7, 8, 9, 17, 18, and 19, 2.5 mL of cells were collected in two volumes of RNAprotect Bacteria Reagent (Qiagen, Germantown
MD). Cells were pelleted by centrifugation at 10,000 x g for 10 minutes and resuspended in lysis buffer (10 mM Tris, 1 mM EDTA, 1 mg/ml lysozyme, 0.1 mg/ml
Proteinase K) for 10 minutes. Total microbial RNA was extracted using the RNeasy Mini
Kit from Qiagen following the manufacturer’s protocol with an on-column DNase treatment step. The RNA quality was assessed using an Agilent 2100 Bioanalyzer
(Agilent, Santa Clara CA) with all samples achieving a RNA Integrity Number (RIN) of 9 or higher. The High-Throughput Sequencing and Genotyping Unit of the Roy J. Carver
Biotechnology Center at the University of Illinois Urbana-Champaign performed high throughput sequencing using an Illumina 4000 sequencing platform.
55 Resulting sequence data was analyzed for differentially expressed genes following a previously published protocol (Anders et al., 2013). Briefly, reads were filtered for quality using Trimmomatic (Bolger, Lohse, & Usadel, 2014). The reads were aligned to the genome using BowTie2 (Langmead & Salzberg, 2012). Reads mapping to gene features were counted using htseq-count (Anders, Pyl, & Huber, 2015).
Differential expression analysis was performed using the edgeR package in R and the
TMM method was used for library normalization (Robinson, McCarthy, & Smyth, 2010).
The reads per kilobase per million mapped reads (RPKM) values were converted to transcripts per million (TPM) values as previously described (Wagner, Kin, & Lynch,
2012). Coverage data was visualized using Integrated Genome Viewer (IGV)
(Thorvaldsdottir, Robinson, & Mesirov, 2013).
3.3 Results
Growth of B. thetaiotaomicron on various nitrogen sources. In order to quantify the contribution of different nitrogen sources towards nitrogen acquisition and resulting growth kinetics, a medium was formulated with no additional confounding nitrogen sources with the removal of methionine, histidine, and cysteine (Appendix
A:Table 2). Initial growth experiments in B. thetaiotaomicron VPI-5482 on three nitrogen sources (Tryptone, ammonium sulfate, and Casamino Acids) present in excess (20 mM
+ NH4 equivalent) demonstrated a preference for peptides in the form of Tryptone (Figure
10). Peptides provided a maximal growth rate of 0.291 ± 0.09 and a maximal optical
density of 1.32 ± 0.01. Growth on ammonium sulfate was slightly slower with a growth
rate of 0.346 ± 0.05, while Casamino acids produced the slowest growth rate and lowest
optical density, 0.355 ± 0.05 and 0.981 ± 0.011 respectively. The quantification of
ammonium in the media revealed a small reduction in the culture containing only
56
Figure 10. Growth of B. thetaiotaomicron VPI-5482 on non-limiting concentrations of various nitrogen sources. Linear plots depicting growth of B. thetaiotaomicron in non-limiting concentrations of ammonium sulfate (10 mM), Tryptone (0.233 %), and Casamino Acids (0.233 %). Optical density measurements were made using a Spectronic 20D+. Experiments were performed in triplicates and bars represent standard deviation.
57 ammonium sulfate, but the samples containing Tryptone and Casamino acids also contained trace levels of ammonium that were depleted during exponential growth
(Figure 11).
To further investigate the differential consumption of nitrogen sources, growth on limiting concentration of ammonium sulfate, Tryptone, Casamino acids and casein were
+ conducted at equal nitrogen content (2 mM NH4 equivalent, Figure 12). Ammonium was rapidly utilized as a nitrogen source for growth with a high growth rate and the highest maximal density within 24 hours, 0.33 h-1 and 0.892, respectively. Peptides as
expected produced the most rapid growth rate, 0.51 h-1, but only for a short period of time resulting in a much lower growth yield. Although growth occurs on casein,
Tryptone, and Casamino acids, they were not efficiently incorporated into cell N, as the cell densities at 24 hours (0.28-0.45) were significantly lower than cell density for growth on ammonium (0.89). The heterogenous nature of proteins, peptides and amino acids likely contributed to the low level of growth.
Nitrogen sources were mixed at 1:1 ratios with ammonium sulfate and either peptides, casein, or Casamino acids at a final nitrogen content equivalent to 2 mM ammonium (Figure 13). Culture of B. thetaiotaomicron containing Casamino acids or
casein supplemented with ammonium grew similarly to cultures with ammonium as the
sole nitrogen source. In culture containing both peptide and ammonium, growth was
more similar to ammonium as the sole nitrogen source (Table 2). Ammonium
concentrations were measured to quantify ammonium consumption (Figure 14). During
growth on various mixed nitrogen sources, ammonium was consumed at a rate similar
to cultures grown solely on ammonium, except for casein with ammonium. After
ammonium depletion, cell growth decreased drastically to rates similar to the cultures
58
Figure 11. Ammonium concentration of B. thetaiotaomicron VPI-5482 growing on non-limiting concentrations of various nitrogen sources. Concentration of ammonium in the culture supernatant of growth curves of Figure 9. Ammonium was quantified using the chromogenic Chaney-Marbach method and absorbance measurements were compared to those of known ammonium concentrations. Experiments were performed in triplicates and bars represent standard deviation.
59
Figure 12. Growth of B. thetaiotaomicron VPI-5482 on limiting concentrations of individual and mixed nitrogen sources. Logarithmic (A) and linear (B) plots depicting growth of B. thetaiotaomicron with limiting and isonitrogenous concentrations of individual nitrogen sources. Individual N-sources (A, C) include ammonium sulfate (▲, 1 mM), Tryptone (●, 0.0233%), Casamino Acids (♦, 0.0233%), and casein (□, 0.0233%). Experiments were performed in triplicates and bars represent standard deviation.
60
Figure 13. Growth of B. thetaiotaomicron VPI-5482 on limiting concentrations of mixed nitrogen sources. Logarithmic (A) and linear (B) plots depicting growth of B. thetaiotaomicron with limiting and isonitrogenous concentrations of mixed nitrogen sources. Individual N-source includes ammonium sulfate (▲, 1 mM). Mixed nitrogen sources include ammonium sulfate (0.5 mM) with Tryptone (0.0116%) (■), ammonium sulfate (0.5 mM) with Casamino Acids (0.0116%) (▼), and ammonium sulfate (0.5 mM) with casein (0.0116%) (○). Experiments were performed in triplicates and bars represent standard deviation.
61
Figure 14. Ammonium concentration of B. thetaiotaomicron VPI-5482 on limiting concentrations of individual and mixed nitrogen sources. Concentration of ammonium in the culture supernatant described in Figure 12 and Figure 13, quantified using the chromogenic Chaney-Marbach method and absorbance measurements were compared to those of known ammonium concentrations. Experiments were performed in triplicates and bars represent standard deviation.
62 growing on the corresponding non-ammonium nitrogen source.
Continuous culture growth on limiting and non-limiting ammonium.
Although batch culture methods are commonly used for analysis of transcriptional regulation to different substrates, this technique is not adequate for the transcriptional analysis in response to varying concentrations of the same substrate. In order to identify transcriptionally controlled genes involved in ammonium assimilation, continuous culture techniques were employed to reproducibly measure bacterial growth parameters and transcriptional responses on non-limiting and limiting ammonium concentrations. To remove potential effects of cell density or accumulation of metabolic end products, glucose was decreased to 0.1% w/v in the culture medium which provided a consistent cell density in all culture conditions. At a 6-hour doubling time, continuous cultures were maintained for 10 days at either limited or non-limited ammonium and subsequently shifted to the other condition and maintained for another 10 days. The daily monitored growth parameters included optical density at 600 nm, pH, and ammonium, glucose, and VFA concentrations.
Cultures maintained a consistent cell density, between 0.6 and 0.8, with minor fluctuations and a steady pH of ~6.9 (Figure 15). Ammonium concentrations in the limiting culture vessels were maintained below 50µM and the non-limiting cultures were between 16 and 18 mM. Glucose concentrations were also near limiting, consistently below 0.5 mM for both culture conditions. The VFA concentrations were found to be relatively consistent but contained unexpected variation during the second stage (Figure
16). For example, cultures switched from low to high had decreased succinate concentrations from 0.4 to 0.2 mg/ml and the propionate increased from 0.15 to 0.2 mg/ml. The culture switched from high to low had a large increase of succinate, from 0.2
63 Figure 15. Growth parameters of B. thetaiotaomicron VPI-5482 during growth in continuous culture. Analysis of growth characteristic and physiochemical parameters of continuous cultures of B. thetaiotaomicron grown for ten days under ammonium limitation (green circle, 2mM ammonium input) and switched to ammonium excess (20 mM input ammonium) for 10 days or grown under ammonium excess (beige square) and switched to ammonium limitation. (A) Ammonium concentration in the supernatant was calculated using the Chaney Marbach method. (B) Absorbance was measured for cell density using Beckman DU7500 at 600 nm wavelength. (C) Glucose concentration was measured using HPLC -PAD for visualization and quantification of glucose. (D) The pH of fresh effluent, as measured by a Fisher Scientific Accumet Basics AB15 pH meter, was monitored throughout growth. Experiments were performed in duplicate and bars represent standard deviation.
64 Figure 16. Production of short chain fatty acids by B. thetaiotaomicron VPI- 5482 grown for 20 days in continuous culture. Analysis of growth characteristic and physiochemical parameters of continuous cultures of B. thetaiotaomicron grown for ten days under ammonium limitation (, 2mM ammonium input) and switched to ammonium excess (20 mM input ammonium) for 10 days or grown under ammonium excess () and switched to ammonium limitation. (A) Acetate, (B) succinate, and (C) propionate were quantified using HPLC-RI to visualize and quantify the fermentation products. Experiments were performed in duplicate and bars represent standard deviation.
65 to 0.6 mg/ml and a drop in propionate from 0.15 to 0.1 mg/ml.
Differential expression RNA-Seq analysis of B. thetaiotaomicron in limiting vs. non-limiting ammonium. At day 7, 8, 9, 17, 18, and 19 of continuous growth, RNA was extracted from continuously grown cells and subjected to high throughput sequencing using an Illumina HiSeq4000 platform. Sequencing reads received from the
Roy J. Carver Biotechnology Center were trimmed of adaptor sequences and contained excellent quality scores for all 395,724,605 reads. After alignment of the reads to the reference genome, the average genome coverage per condition was 397 X and 358 X for limiting and non-limiting ammonium, respectively (Table 3). The average percentage of mapped reads to the genome was 98.65% and 96.3% for limiting and non-limiting ammonium, respectively, reflecting the purity of the cultures and quality of the extracted
RNA. Differential expression analysis was conducted using edgeR and TPM in order to compare gene expression levels within a sample as well as between samples.
Due to the high sample number (n=12), the highly statistically significant results could be filtered based on a very strict false discovery rate (FDR) cutoff of 0.0002, which allows for only one falsely discovered differentially expressed gene in the genome.
Global transcriptional analysis revealed a limited number of genes with significant differential expression levels (fold change > 4) between limiting and excess ammonium.
In Figure 17, the log2 fold changes of each gene was plotted onto the genome, and red
dots describe the genes with low FDRs and >4 fold differential expression. A large
majority of the statistically significant differentially expressed genes were found to be
involved in nitrogen metabolism. One locus was especially responsive to limiting
ammonium concentrations, and was bioinformatically identified as the GS/GOGAT
66
Table 3. Qualitative assessment of read mapping to the closed genome of B. thetaiotaomicron strain VPI-5482 from Illumina HiSeq4000 sequencing results.a
% reads mapped Genome % reads mapped Condition Sample Name Total reads to genomeb Coverage to genes Excess NH4+ H1_17 20,063,862 99.8 469.4 85.7 H1_18 13,760,555 99.8 322.1 87.1 H1_19 11,681,392 99.5 272.7 87.1 H2_17 19,594,409 97.8 448.9 85.4 H2_18 17,705,594 99.7 413.8 86.2 H2_19 13,821,028 99.4 321.9 86.5 H3_7 14,637,261 82.3 282.9 74.8 H3_8 18,251,603 99.8 426.8 85.6 H3_9 14,758,791 98.7 341.6 86.8 H4_7 19,578,699 99.4 456 87.4 H4_8 11,510,924 99.5 268.5 88.3 H4_9 14,641,400 80.0 275 70.7 Average 15.8x106 ± 3.1x106 96 ± 7 358 ± 79 84 ± 6 Limiting L1_7 21,261,290 98.1 489.4 86.3 NH4+ L1_8 15,571,590 99.4 363.4 87.4 L1_9 11,211,915 98.9 260.3 88.9 L2_7 31,177,919 99.7 728.5 88.2 L2_8 18,454,639 98.9 427.8 86.6 L2_9 13,356,565 99.0 310 88.9 L3_17 20,491,883 99.5 478.7 87.4 L3_18 13,253,385 99.4 309.2 88.3 L3_19 13,829,863 99.5 323.1 88.4 L4_17 21,999,006 99.7 514.8 87.9 L4_18 11,286,580 99.7 264.1 87.4 L4_19 13,824,452 92.1 298.9 81 Average 17.1x106 ± 5.8x106 99 ± 2 397 ± 137 87 ± 2 Total 395,724,605 97.7 9068 86.1
a Average read length is 250nt b Reads mapped after ribosomal RNA trimming
67
Figure 17. Global transcriptional regulatory changes of B. thetaiotaomicron strain VPI-5482 grown in continuous culture with limiting vs. excess ammonium. RNA- Seq analysis performed for B. thetaiotaomicron genes differentially transcribed under ammonium limitation as compared to ammonium excess. Transcriptional fold changes were plotted for each gene locus across the entire genome and plasmid of B. thetaiotaomicron VPI-5482. Red dots describe genes significantly differentially transcribed between the two conditions compared. Fold change cutoff set at four-fold increase or decrease. False discovery rate limited to one gene, i.e. FDR value <0.0002.
68 encoding locus (Table 4 and Table 5). All of the genes depicted in Figure 18A were between 146- and 395-fold more highly expressed under ammonium limitation as compared to non-limiting ammonium. Specifically, the amtB gene was the most
transcriptionally responsive gene, with 394.5-fold higher expression under ammonium
limitation. The glutamine synthetase type 3, glnN1, was the next highest
transcriptionally responsive gene with 322.6-fold increased expression followed by the
glutamate synthase subunits, gltB and gltD, with 283.2- and 275.2-fold increased
expression, respectively. The top nine most differentially expressed genes exist in two
putative operons encoding the high affinity pathway, lysine biosynthetic genes and
arginine synthase. Outside of these two operons, there was a log decrease in
differential expression, highlighting the importance of the GS/GOGAT pathway, AmtB,
and glutamine/glutamate consuming lysine and arginine biosynthetic genes as the
primary drivers of ammonium assimilation under nitrogen limitation.
Other genes involved in amine transfer were differentially expressed and may play
important roles in ammonium/nitrogen acquisition. For instance, carbamoyl synthase genes (carAB), hexophosphate aminotransferase (glmS), and amidophosphoribosyl transferase (purF) are positioned in a putative operon adjacent to the GS/GOGAT system. These genes were 4- to 7-fold more highly expressed under ammonium limitation. The large gap in fold changes illuminates the GS/GOGAT locus as the primary driver of ammonium assimilation under ammonium limited conditions.
For ammonium assimilation genes associated with non-limiting ammonium
concentration, the glutamine synthetase type 1, glnA, was the most responsive gene
with 9.45 fold upregulation. It is unclear why this gene was more highly expressed under
69
Table 4. Genes of B. thetaiotaomicron with significant transcriptional changes during continuous growth in limiting vs. excess ammonium.
Transcript Abundance (TPM) Locus Fold Limiting Tag Gene Predicted function Changea FDRb NH4+ Excess NH4+ BT_0544 amtB Ammonium transporter 394.5 1.51E-265 9785 ± 2937 26.25 ± 9.93 BT_0543 glnN1 Glutamine synthetase 3 322.6 6.68E-226 22630 ± 5301 76.22 ± 38.64 BT_0552 gltD Glutamate synthase SS 283.2 5.67E-131 3772 ± 495.8 15.40 ± 14.20 BT_0553 gltB Glutamate synthase LS 275.2 1.49E-120 15393 ± 1564 64.53 ± 57.70 BT_0551 asnB Asparagine synthetase 266.7 2.02E-131 4381 ± 629.8 18.91 ± 16.94 BT_0545 glnK PII regulatory protein 266.4 1.21E-201 4049 ± 1268 16.48 ± 9.62 BT_0546 Hypothetical 252.2 8.56E-227 7905 ± 2837 33.04 ± 16.24 BT_0548 dapF Diaminopimelate epimerase 176.9 1.09E-198 2786 ± 627.4 16.77 ± 10.02 BT_0547 dapL LL-DAP aminotransferase 146.0 9.39E-165 5261 ± 1048 39.35 ± 25.07 BT_2779 Coq3 O-methyltransferase 8.35 4.07E-28 82.35 ± 71.05 12.26 ± 10.63 BT_2178 hypothetical 7.25 1.06E-55 419.8 ± 182.4 58.72 ± 15.16 BT_0557 carB carbamoyl phosphate synthetase LS 6.72 3.15E-61 3288 ± 672.5 520.9 ± 101.0 D-glycero-D-manno-heptose 1- BT_0474 6.54 1.16E-04 11.57 ± 20.48 1.521 ± 1.107 phosphate kinase BT_0554 glmS hexosephosphate aminotransferase 6.10 2.79E-11 2205 ± 218.6 518.8 ± 798.7 BT_2778 sigma factor 5.98 6.80E-24 52.24 ± 41.79 9.617 ± 6.710 BT_1564 7-cyano-7-deazaguanine reductase 4.76 2.88E-38 136.4 ± 41.88 30.04 ± 6.206 BT_1566 aluminum resistance protein 4.66 1.79E-28 185.5 ± 63.96 42.17 ± 11.10 BT_0480 glycosyltransferase 4.50 9.63E-05 22.56 ± 29.27 4.539 ± 1.629 BT_0555 purF Amidophosphoribosyl transferase 4.43 6.96E-10 2137 ± 220.7 595.1 ± 757.2 BT_1565 Hypothetical 4.37 1.65E-32 134.0 ± 42.25 32.76 ± 5.886 BT_3572 Hypothetical 4.35 1.25E-06 76.69 ± 88.13 16.23 ± 4.588 BT_0556 carA carbamoyl phosphate synthase SS 4.33 1.51E-09 1530 ± 313.7 412.5 ± 495.4 BT_3279 SusC homolog 4.30 4.51E-29 110.8 ± 43.70 25.5 ± 7.826 BT_3278 anti-sigma factor 4.28 2.32E-26 51.9 ± 25.04 12.32 ± 4.206 BT_1563 phosphohydrolase 4.14 2.94E-39 111.6 ± 28.26 28.70 ± 5.653 BT_3573 Hypothetical 4.04 4.95E-06 25.96 ± 30.42 5.888 ± 1.420 BT_4014 restriction endonuclease 0.220 3.27E-08 6.683 ± 8.594 111.9 ± 119.4 BT_4013 restriction endonuclease 0.218 5.57E-11 35.31 ± 36.18 537.3 ± 568.4 BT_2131 Hypothetical 0.205 8.06E-09 63.93 ± 31.71 507.8 ± 628.4 BT_2409 SusC homolog 0.193 2.65E-11 4979 ± 4999 10217 ± 2234 BT_0785 glnA Glutamine synthetase 1 0.106 4.32E-187 72.74 ± 12.99 719.2 ± 76.38
a Comparison of Limiting/Excess ammonium. Fold change cutoff was >4 fold. b FDR cut off is very strict (0.0002)
70 Table 5. Relative expression and transcript abundance of the ammonium assimilation genes of B. thetaiotaomicron strain VPI-5482 grown in continuous culture with limiting vs. excess ammonium.
Gene Locus_tag Fold Change P-Value Non-limiting NH4+ Limiting NH4+ (TPM ± SD) (TPM ± SD) amtB BT_0544 394.5 3.12E-269 26.25 ± 9.9 9785 ± 2937 glnN1 BT_0543 322.6 4.15E-229 76.21 ± 38.7 22630 ± 5301 gltD BT_0552 283.2 1.06E-133 15.40 ± 14.2 3772 ± 495.9 gltB BT_0553 275.2 3.08E-123 64.53 ± 57.7 15393 ± 1564 glnK BT_0545 266.4 1.00E-204 16.48 ± 9.6 4049 ± 1268 gdhA BT_1970 3.22 7.57E-34 1596 ± 142.0 4823 ± 482.3 glnN2 BT_4339 2.96 7.68E-24 294.4 ± 41.16 824.2 ± 271.5 gdhB BT_1973 0.65 1.46E-3 636.1 ± 172.5 392.7 ± 85.15 glnA BT_0785 0.106 5.37E-190 719.2 ± 76.38 72.74 ± 12.99
71 ammonium excess. It is unlikely that this gene is primarily responsible for ammonium assimilation under non-limiting conditions as gene deletion analysis (presented in
Chapter 4) reveals this gene to be non-essential under all conditions tested.
As a result of calculating TPM for each gene, comparison of expression level between different genes of different sizes can be made within a sample. The comparison of TPM values can be used as a metric for estimating the relative contribution of a gene’s transcripts to the mRNA pool under a specific condition. For instance, under excess ammonium, the gdhA has the highest expression level of the ammonium assimilatory genes at 1596 ± 142 TPM. The second most expressed ammonium assimilatory gene was glnA, with 719 ± 76 TPM, likely for production of glutamine. To identify the contribution of individual glutamine synthetases towards ammonium assimilations, glnA was the most highly expressed during non-limiting
conditions, followed by glnN2 at 294 ± 41 TPM, and the glnN1 has very low expression
at 76 ± 39 TPM. However, under ammonium-limited conditions, the glnN1 becomes the
most highly expressed gene in the cell with 22,630 ± 5301 TPM. The expression of the
other ammonium assimilation genes were also extremely high, including gltD, amtB,
gdhA, glnK, and gltB at 15394, 9785, 4823, 4049, and 3772 TPM, respectively.
3.4 Discussion
Through growth experiments performed in batch culture, a differential utilization
nitrogen sources is clear based on the corresponding growth characteristics. While B.
thetaiotaomicron rapidly and efficiently metabolizes ammonium, growth on excess
concentration of peptides results in a more rapid growth rate and a higher maximal
optical density. Through bioinformatic analysis of the genome, four predicted peptide
72 transporters are encoded by BT_0580, BT_1086, BT_4012, and BT_4385. While the exact peptides transported by these putative transporters are unknown, the bacterium clearly displays the capacity to acquire and metabolize peptides. A related organism,
Prevotella albensis, decreases proteolytic activity in the presence of exogenous ammonium (Sales-Duval et al., 2002; Sales et al., 2000). Peptide utilization is not mutually exclusive of ammonium utilization as the consumption of trace ammonium in the medium occurs during the logarithmic phase of growth, which may imply co- metabolism of extracellular peptides and ammonium. This is a critical characteristic, as the peptide concentrations in the mammalian host are continuously fluctuating (and potentially limiting) while the ammonium concentration is comparatively stable (Chen,
Sniffen, & Russell, 1987; Visek, 1978; Wolpert et al., 1970; O. Wrong et al., 1965).
When the concentration of peptides was limited and compared to the isonitrogenous concentration of ammonium (2 mM), drastic differences in utilization become apparent. While peptides provide a more rapid growth rate initially, ammonium grown cultures reach a higher optical density with a more consistent growth rate.
Previous studies have described the preferential use of hydrophilic peptides, as opposed to hydrophobic peptides (Chen, Strobel, et al., 1987). The preferential consumption of hydrophilic peptides may provide a rapid growth rate but the imbalance in peptide composition prevent maximal utilization or cell yield.
Metabolism of amino acids as a nitrogen source was poor in relation to the
ammonium, which is a common feature in gut bacteria (Chen, Strobel, et al., 1987;
Cotta & Russell, 1982). Annotation of the genome sequence revealed several predicted
amino acid transporters for cysteine (BT_3407/3408), glycine betaine & proline
73 (BT_1749-1751), aminobenzyl-glutamate (BT_0751), alanine, glutamate/γ- aminobutyrate antiporter (BT_2573), aspartate/alanine antiporter, and glutamate
(BT_1223). These may be responsible for some growth achieved with amino acids, but the precise functions of the transporters are difficult to predict. With cysteine as the sole nitrogen and sulfur source, B. thetaiotaomicron was able to grow unhindered, while both
glutamate and glutamine were unable to foster growth (data not shown). While peptides
are transiently available in the gut, and ammonium is ever-present, amino acids are
rapidly de-aminated and do not accumulate, therefore growth on amino acids is not
essential for survival in the gastrointestinal tract (Wright & Hungate, 1967).
One of the most useful techniques in the study of concentration-dependent
ammonium assimilation is the employment of continuous cultures. Through our
continuous culture approach, a transcriptional analysis of ammonium responsive genes
was possible at extracellular ammonium concentrations of below 50 µM. After stabilizing
all other extracellular stimuli that could influence transcription of ammonium assimilatory
genes, including pH, OD, glucose, and VFAs, the transcriptionally responsive genes
would encompass only genes responsive to extracellular ammonium. A global
transcriptional analysis exposed the selective expression of a single genetic locus, the
High-affinity Ammonium Assimilation Locus (HAA Locus) (Figure 18A), to limiting
ammonium concentrations. Analogous to the genetic organization of Bacteroidales PUL
systems, the machinery employed to ammonium limitation is co-localized on the
genome and expressed in two putative operons. This is the first evidence for the entire
High affinity system being employed by a gut Bacteroides species towards the
assimilation of ammonium when ammonium is in limiting concentrations.
74
Figure 18. RNA-Seq coverage map of the B. thetaiotaomicron VPI-5482 ammonium assimilation loci with representative samples from limiting and non-limiting conditions. Reads from RNA-Seq have been plotted as genome coverage onto the ammonium assimilation loci encoding all the predicted genes in the high affinity pathway (A, B C) and low affinity pathway (D). Genes depicted on the X- axis and transcripts on the Y-axis. Integrated Genome Viewer used to make transcript coverage maps of representative samples for limiting ammonium (L1_7) and excess ammonium (H1_17).
75 As the prevailing theory of Bacteroidales ammonium assimilation invokes the
GDH under all environmental ammonium concentrations, we propose that the high affinity system must assimilate ammonium when ammonium becomes limiting. Several
studies describe NAD(P)H-dependent GDH to be constitutively active or even elevated under ammonium limitation in Prevotella and Bacteroides species, including P.
ruminicola, P. brevis, P. bryantii, B. fragilis and B. thetaiotaomicron (Baggio & Morrison,
1996; Z. Wen & Morrison, 1996, 1997; Yamamoto et al., 1984; Yamamoto, Saito, et al.,
1987). In B. fragilis, a nitrogen limiting continuous culture showed increase glutamine
synthetase activity, but addition of the glutamine synthetase inhibitor, L-methionine
sulfoximine, did not impact growth (Yamamoto et al., 1984). Unfortunately, the authors
added cysteine in the medium which contributes to a confounding nitrogen source that
could potentially support growth. Further evidence for the importance of glutamine
synthetase provided by increased transcription of glutamine synthetase in Prevotella bryantii B14 when ammonium is limiting, while P. ruminicola 23 has been demonstrated
to up-regulate the transcription of the high affinity system when ammonium was in
excess (J. N. Kim et al., 2012; J. N. Kim et al., 2017; Z. T. Wen et al., 2003). The primary limitation in interpreting these studies is the addition of cysteine to the medium, which can be used as a source of nitrogen. In contrast to the previous work, we
minimized the variance of environmental nutrients between the two conditions being compared and also increased the biological replication. Through the careful control of external metabolites, the high affinity system dominated the transcript pool, with the glnN1, gltB, and amtB being the first, third, and fifth most highly expressed genes in the cell, respectively.
76 In enteric proteobacterial organisms, the glutamine synthetase is modestly regulated via transcription and the post-translational modification is a dominant mode of
regulation. This mechanism is different to our results, where transcription of the gln1
and other genes of the high affinity system are strongly regulated via transcription.
Organisms like Escherichia, Salmonella, and Klebsiella only encode one glutamine
synthetase responsible for glutamine biosynthesis under all external ammonium
concentrations as well as assimilation of ammonium when concentration becomes
limiting. To increase the metabolic flexibility of a single enzyme, these organisms
possess greater constitutive transcription of the glnA gene and increased transcription
when environmental ammonium concentrations become limiting. For example, in E. coli,
a reporter gene (lacZ) was fused to the glnA or a glnK promoter to assess
transcriptional regulation in limiting vs. non-limiting concentrations of ammonium
(Atkinson et al., 2002). While the glnK promoter was essentially silent under ammonium
excess the glnA promoter was still active. When the organism was exposed to
ammonium limitation, the glnA promoter experiences a ~10-fold increase in the expression; the glnK promoter became as transcriptionally active as the glnA promoter.
In our work, the glnA was functionally expressed under all conditions but modestly up regulated when ammonium became limiting. In contrast, the glnK promoter, representing ammonium transport and expression of the high affinity system,
experienced dramatic up-regulation under ammonium limitation. Additionally, the
transcriptional response of the glutamine synthetase gene from M. smegmatis is also
modest in their response to ammonium concentration, approximately 7.5 fold higher in
ammonium-limited as compared to ammonium-excess continuous cultures (Petridis et
77 al., 2015). The high affinity system of B. thetaiotaomicron is consistent with the transcriptional regulation of the glnK promoter of E. coli, with the novel addition of a glutamine synthetase specific for the high affinity system selectively expressed under ammonium limitation.
The gut Bacteroidales encode three GS genes, which may play specific roles of glutamine biosynthesis under different environmental conditions. The glnN2 is constitutively transcribed and the glnA is expressed preferentially under ammonium excess. One of these two genes may be primarily responsible for glutamine biosynthesis under ammonium excess conditions and the other solely responsible for ammonium assimilation under limiting conditions.
If the High-affinity Ammonium Assimilation locus is solely responsible for ammonium assimilation, several additional genes are novel to this paradigm. For example, the dapFL or asnB genes supplement the GS/GOGAT in the incorporation of ammonium. These genes are massively up-regulated when ammonium is limiting and utilize the products of the GS/GOGAT pathway. We propose the DapLF and AsnB consume glutamate and glutamine, respectively, in order to drive the assimilatory reaction of GS/GOGAT forward, either by preventing product inhibition or forcing the forward reactions by depleting the reaction products. The AsnB product, asparagine, can be used for protein synthesis, and the meso-diaminopimelate (mDAP) produced by
DapLF can subsequently be used for lysine biosynthesis or as a precursor for peptidoglycan biosynthesis. The dapLF and asnB genes of Prevotella ruminicola 23 were also upregulated in response to ammonium along with the high affinity pathway (J.
N. Kim et al., 2017). The high conservation of these genes within the Bacteroides
78 species (Figure 5) emphasizes their importance for ammonium assimilation in the high affinity system of Bacteroides. Our results lead us to propose a restructuring of the
enteric paradigm of ammonium assimilation by exposing “accessory” or “secondary”
enzymes mechanistically linked to the GS/GOGAT.
Through differential gene expression analysis, the understanding of ammonium
assimilation under strongly limiting conditions is illuminated, suggesting B.
thetaiotaomicron modifies the classical enteric paradigm with notable differences and
novel features. Most notably, only one of the GS enzymes, GlnN1, is selectively
transcribed in response to ammonium limitation. Additionally, the dapLF and asnB
genes appear to be key accessory genes to the high affinity system. While ammonium
limitation clearly elicits a selective transcriptional response, it is still unclear which genes
are crucial for ammonium assimilation under higher ammonium concentrations that are
commonly found in the gastrointestinal tract. Therefore, further analyses involving
genomic manipulations and gene deletions are needed to identify essential genes for
ammonium acquisition.
79 Chapter 4. Probing the genetic machinery of ammonium assimilation through
phenotypic and competitive fitness analyses
4.1 Introduction
Through host digestion and microbial metabolism, ammonium is a readily available source of nitrogen for commensal microorganisms to assimilate into microbial protein. The concentration of ammonium in the gut is rarely limiting and ranges between
2-50 mM (Russell & Strobel, 1987; Visek, 1978; Wolpert et al., 1970; O. Wrong et al.,
1965). The enteric paradigm of ammonium assimilation describes the employment of
two competing catalytic pathways, coordinated by intricate regulatory mechanisms. The
high affinity and the low affinity pathway are active under specific environmental
conditions, dictated by the environmental ammonium concentration.
To uncover the catalytic and regulatory machinery, genetically tractable model
organisms are mutated, through random or targeted mutagenesis, resulting in amino
acid auxotrophies when catalytic genes are disrupted. The ensuing phenotypes from
gene disruptions correlate to deficiencies in one of the two ammonium assimilation
pathways, elucidating the anabolic pathways employed in response to specific
environmental cues. If phenotypic genetic analyses in phylogenetically diverse
microorganisms repeat similar finding, it demonstrates that the mechanism of
ammonium assimilation is conserved throughout the three domains of life.
According to the enteric model of ammonium assimilation, the conditions commonly encountered in the mammalian gastrointestinal tract require only a single enzyme from the low affinity pathway, glutamate dehydrogenase (GDH). The GDH
80 enzyme reductively aminates ammonium, via NADPH as a cofactor, onto an α- ketoglutarate releasing glutamate. Organisms manifest a glutamate auxotrophy when
GDH is impaired if glutamate synthase activity (GOGAT, a high affinity pathway enzyme which produces glutamate) is additionally removed (Berberich, 1972; Dendinger, Patil, &
Brenchley, 1980; Harth, Maslesa-Galic, Tullius, & Horwitz, 2005). If GDH (NADH- dependent) is utilized for the catabolism of glutamate, as in Mycobacterium bovis, glutamate auxotrophy develops without a functional GOGAT, and deleting the catabolic
GDH causes a growth defect when grown on glutamate as the sole nitrogen source
(Viljoen, Kirsten, Baker, van Helden, & Wiid, 2013). Additionally, deletion of gdh alone results in a sensitivity to acid or nitric oxide stress (Gallant, Viljoen, van Helden, & Wiid,
2016).
Analysis of mutant strains additionally can lead to the identification of pathway regulation. In Klebsiella aerogenes, a single mutant in GDH does not impact growth when ammonium is limiting, <1mM, revealing the preferential employment of GDH at non-limiting concentrations of ammonium (Brenchley & Magasanik, 1974). In
Salmonella and Klebsiella, there is differential regulation of GDH under limiting ammonium concentrations. With decreasing ammonium, gdhA is repressed in
Klebsiella, while Salmonella does not repress enzymatic GDH activity (Brenchley,
Baker, & Patil, 1975; Brenchley, Prival, & Magasanik, 1973).
When environmental ammonium is present at limiting concentrations (<0.7 mM), two enzymes assimilate ammonium in an energy dependent manner, glutamine synthetase and glutamate synthase. Initially, the glutamine synthetase incorporates ammonium onto glutamate while consuming an ATP and releasing a glutamine.
81 Subsequently, glutamate synthase consumes a glutamine, an α-ketoglutarate, and an
NADPH to produce two glutamate. In organisms encoding only one glutamine synthetase gene, i.e. K. aerogenes and E. coli, loss of a functional glnA causes glutamine auxotrophy, incapable of producing endogenous glutamine from ammonium for the intracellular amino acid pools (Bender & Magasanik, 1977; MacNeil, MacNeil, &
Tyler, 1982). Deletion of glnA can impact virulence, protease expression, and even
ethanol production in Streptococcus pneumoniae, Bacillus subtilis, and Clostridium thermocellum respectively (Abe, Yasumura, & Tanaka, 2009; Kloosterman et al., 2006;
Rydzak et al., 2017). Furthermore, growth of a GOGAT-deficient Salmonella
typhimurium strain in ammonium-limited continuous culture results in spontaneous mutations. Here, an increase in the glutamate dehydrogenase and partial impairment of the glutamine synthetase allowed the mutant to outcompete the parental mutant strain and maintain intracellular glutamate and glutamine concentrations, ultimately restoring normal growth kinetics (Yan, 2007).
Many organisms encode multiple genes with the same predicted function, including Saccharomyces, Mycobacterium, and gut microorganisms like Prevotella and
Bacteroides. In organisms with potential genetic redundancy, gene deletion analysis has revealed the singular gene responsible for ammonium assimilation. For example,
Saccharomyces cerevisiae has three glutamate dehydrogenase genes, only one of which is required for ammonium assimilation (Tang, Sieg, & Trotter, 2011).
Mycobacterium tuberculosis has four glutamine synthetase genes, and it is only when glnA1 is absent that the organism is incapable of growth with ammonium as the sole nitrogen source (Harth et al., 2005). This analysis directly applies to gut organisms like
82 Bacteroides and Prevotella, which have multiple glutamine synthetase, as well as glutamate dehydrogenase, genes.
The application of the enteric paradigm towards gut Bacteroidales has been confounded by two contrary research findings. First, Bacteroides and Prevotella rely on
glutamate dehydrogenase for ammonium acquisition under excess and limiting
ammonium concentrations, although they do encode the high affinity pathway (Z. Wen
& Morrison, 1997; Yamamoto et al., 1984; Yamamoto, Saito, et al., 1987). Secondly,
they encode multiple genes within each pathway, commonly with two glutamate
dehydrogenases and three glutamine synthetases. It remains unclear as to which
enzymes are strictly required for uptake and growth on ammonium and which enzymes
are conditionally expressed in response to different extracellular or intracellular nitrogen
conditions.
In order to assess the contribution of a particular gene to the ammonium
assimilation of Bacteroides, the genes predicted to be associated with these pathways
were systematically deleted from the genome. Of the two GDH genes encoded within B.
thetaiotaomicron, only loss of the putative NADPH-dependent glutamate dehydrogenase, gdhA (BT_1970), results in a growth defect. When genes annotated as glutamine synthetases were deleted, the glnN2 also resulted in a growth defect.
Competitive fitness analysis was performed between the parental strain and mutants lacking either one of the glutamine synthetase (glnN1, glnN2, or glnA) or the gdhA. The
second glutamine synthetase type 3 (glnN2) displayed a mild fitness defect whether
ammonium was limiting or in excess, but the gdhA mutant resulted in a more severe
fitness defect. Overall, gene deletion analysis has uncovered the crucial role of the
83 NADPH-dependent gdhA, as well as provided evidence for the physiological importance of a second glutamine synthetase, glnN2.
4.2 Materials and Methods
Materials, bacterial strains, and plasmids. B. thetaiotaomicron VPI-5482 Δtdk
and E. coli strain S17-1 λpir was obtained from Dr. Eric Martens from the University of
Michigan, Ann Arnor. Anaerobic bacteria were stored in agar slants in the vapor phase
(-110°C) of a liquid N2 cryostorage system as well as 30% glycerol stocks at -80°C. All
E. coli strains were maintained as 30% glycerol stocks in -80°C. An E. coli strain S17-1
λpir harboring pExchange-tdk was also obtained from Dr. Eric Martens and stored at -
80°C. The complementation plasmid, pNBU2-erm, was obtained from Dr. Michael
Fishbach, and the barcoding plasmids, set of five pNBU2-tet plasmids, were obtained from Dr. Patrick Degnan. All DNA was stored in -20°C freezers. Reagents used for PCR include Phusion High Fidelity DNA Polymerase (New England Biolabs, Ipswich MA).
Quantitative PCR was performed using KAPA SYBR FAST qPCR Master Mix (KAPA
Biosystems, Boston MA). Purification of DNA was performed using either QIAprep Spin
Miniprep Kit or Qiagen DNeasy Blood & Tissue Kit (Qiagen, Germantown MD). All other materials were obtained from Sigma Aldrich.
Bacterial growth conditions. For batch cultures, B. thetaiotaomicron was grown in an anaerobic cabinet with a headspace of 95% CO2 & 5% H2. Initially, isolated
colonies were grown in Brain Heart Infusion Supplemented medium (BHIS, Appendix
A:Table 1), and inoculated into chemically defined medium (CDM, Appendix A:Table 2)
with either ammonium sulfate (1 or 10mM) or an isonitrogenous concentration of
peptides (Bacto Tryptone, 0.25% or 0.025%, Appendix A:Table 2). After cells were pre-
84 conditioned in the specific nitrogen condition through three successive daily passages into fresh medium, cells were sub-cultured at a 20-fold dilution into Balch tubes with the
corresponding nitrogen condition and absorbance measurements were taken at 600 nm
using a Spectronic 20D+.
Genomic deletions. Gene deletions were performed as described in Koropatkin
et al. (Koropatkin et al., 2008) and depicted in Figure 19. Briefly, a deletion plasmid
(pExchange-tdk) was constructed through sewing PCR to contain two fused 750bp
segments of DNA from upstream and downstream the gene of interest and represents
the final genotype desired (Table 6). The plasmid was sequenced at the William Keck
Center for Comparative Genomics to confirm the correct insert. Subsequently, E. coli
S17-1 λpir harboring the deletion plasmid of interest and B. thetaiotaomicron Δtdk were
grown to an OD of 0.4 and 5 ml of culture was centrifuged and re-suspended together in
1 ml of TYG and plated on BHIS agar, for the conjugal transfer of the suicide plasmid
(pExchange-tdk) into B. thetaiotaomicron (Figure 19B). After 24 hours of incubation at
37°C, the lawn of cells were anaerobically scraped and re-suspended in BHIS medium,
serially diluted, and plated on BHIS agar supplemented with erythromycin (25 μg/ml)
(Figure 19C). The resulting isolated colonies were re-streaked and secondary isolated
were picked and grown to saturation in BHIS without supplementation. The eight
saturated cultures from eight separate isolated colonies were pooled, serially diluted,
and plated on BHIS supplemented with 5-flouro-2-deoxyuridine (FUdR, 200 μg/ml)
(Figure 19D). Isolated colonies were re-streaked and secondary isolates were picked
and grown to saturation. Saturated cultures were stored in 40% glycerol at -80°C.
85 Table 6. List of primers used for construction of deletion plasmids.
Plasmid Region Forward Primer Reverse Primer pExchange-tdk- Upstream AAAAAGATATCCTATTCTTGGTCCCCGTATC CTCCATTCAGCATTTCTTTTAATACCTAATTT ΔglnN1 GGTAAATATGG GATTTGTTGTTAAATATGTTAGAGTAAAAAT TCCTTTTTTAG Downstream CAACAAATCAAATTAGGTATTAAAAGAAAT AAAGGATCCTTATATTCGCAAAAGTATTGAA GCTGAATGGAGAATATCTTCATTC ATGGACTTCGTTAAAAAATATAGAATATATG pExchange-tdk- Upstream NNNGTCGACTTATAATGGATACACAAGTGG GTTAGAGTAAAAATTCCTTTTTTAGCAATGA ΔamtB GTTCTGCGTTTC TTATCTTATTTAATACCTGTTTTTATAATATT ATCATCCGAAAGGAAATCTG Downstream AAAACAGGTATTAAATAAGATAATCATTGC NNNGGATCCGGACGTGCGGCATACAAAC TAAAAAAGGAATTTTTACTCTAACATATTTA ACAACAAATCAAATTAGGTATTATG pExchange-tdk- Upstream AAAGTCGACGTTATATCTGGCGTGGTCAGGA GCCGGTTGCCGGATACCTTAAATATTAGTGT ΔglnK TTTAGG TGACCTGTAAAAAGAAGAGGTATAGGG Downstream CACTAATATTTAAGGTATCCGGCAACCGGC AAAGGATCCGAACCTTCTTCTCCGTTCATCA AAGG GCC pExchange-tdk- Upstream AAAGTCGACGCACCTCGGTATCGGTACTACT GTTCCTTGCTGTAAGGGTCTTTCTAGCTTTAT ΔgltBD TTTAAAG TGTATGATAATACCTATTTGATGTATCAAAT TGTTC Downstream CAATAAAGCTAGAAAGACCCTTACAGCAAG AAAGGATCCAAAAGTAATGTCCCGGAAGAA GAACACCATGAAG ATGGTTCG pExchange-tdk- Upstream AAAGTCGACAGGTGCCCGGTATCACCG CAGAATGCTTTTATTGCCTTCTATCGGTTTGT ΔglnA TGTTAGTGAACCAAAGGTAAG Downstream CTAACAACAAACCGATAGAAGGCAATAAA AAAGGATCCAGTGAACCACTCTCCTATTATA AGCATTCTGCTTATAAAAATACTACAACTAA TCAGCCAATGTG AAGAGG pExchange-tdk- Upstream AAAGTCGACCTCGATGTTGTCGTACTCCAAC CTATAGCTTTCTTATATTTTGCTGAAATCTTT ΔgdhA TTCG TTTTCTTTTATTGGTTGATTAAATTTCGATTAC CTATTTCTTTTATG Downstream CAACCAATAAAAGAAAAAAAGATTTCAGCA AAAGGATCCTAGAATTTCATGGCAAAACCG AAATATAAGAAAGCTATAGTCATTTTTAAGA CGGATATCAC TAAAGGCTCC pExchange-tdk- Upstream AAAGTCGACAGAATACGTATTCAAATGCGC CAAATACAACTAAACAAAGTCTCTTGATATC ΔgdhB TCAGGTC ATATTTTAAAAGGTTATTCTGATAAAATGTA AACTGCTCACAC Downstream CAGAATAACCTTTTAAAATATGATATCAAG AAAGGATCCGGTCAGCACACTTCCC AGACTTTGTTTAGTTGTATTTGTATTTGTGGT TTGC pExchange-tdk- Upstream AAGCGGCCGCAAAAAAGAATATCGAAAAG GCAATCTTCAGGAGTTTTAAATCGTCTTCTTT ΔglnN2 CAAGGGAATTAAGCTGC TTATCGGTTGATGGCG Downstream CGATAAAAAGAAGACGATTTAAAACTCCTG AAAGGATCCGATCAGACGTTCCGTAGCGAA AAGATTGCACGAAAGAACCTAC CTGTG
86 In order to confirm the desired phenotype, the strains were grown on FUdR-
containing or erythromycin-contained BHIS plates. Only strains that are resistant to
FUdR and sensitive to erythromycin were retained for further analysis. Genomic DNA
was purified from selected cultures and the genotype was verified using three PCR
reactions: (1) the presence of B. thetaiotaomicron (susCD-specific primers); (2) the
absence of the pExchange-tdk plasmid (oriR6K-specific primers); and (3) the length of the excised region (UR forward and DR reverse).
Complementation strain construction. Each constructed deletion strain was complemented with the corresponding gene in trans using a pNBU2-erm plasmid harboring the deleted gene and the putative promoter, 250bp region upstream of the gene. First, a complementing plasmid was constructed by cloning the gene and promoter of interest with primers listed in Table 7. Amplicons and plasmids were digested, ligated and transformed into E. coli S17-1 λpir. The plasmid was sequenced at the William Keck Center for Comparative Genomics to confirm the correct insert. The E. coli strain harboring the plasmid was conjugated as previously described into the corresponding B. thetaiotaomicron deletion mutant. The resulting lawns from conjugation were anaerobically scraped and resuspended in BHIS. Serial dilutions were plated on BHIS supplemented with erythromycin (25 μg/ml) and incubated at 37°C for
48 hours. The resulting isolated colonies were re-streaked on BHIS supplemented with erythromycin (25 μg/ml) and secondary isolates were picked and grown to saturation in
BHIS supplemented with erythromycin and cultures were stored in 40% glycerol at -
80°C. As the pNBU2 plasmid can integrate into two locations on the genome, att1 and att2 specific primers were used to discern the location of the insertion and only att1-
87 Table 7. List of primers used for construction of complementation plasmids.
Plasmid Region Forward Primer Reverse Primer pNBU2-erm- promoter AAAGCGGCCGCCATATCGACCTATTGTAAAA GTTAAATATGTTAGAGTAAAAATTCCTTTAGA ΩglnN1 GCGAAAATTC TATGACTTTTTGATTTAACTTTTGATGAAAAGA TG gene GTTAAATCAAAAAGTCATATCTAAAGGAATTT AAAGGATCCTTATCTTACAAACAACAGTTCTC TTACTCTAACATATTTAACAACAAATCAAATT TGTATTTCGGCAG AGGTATTATG pNBU2-erm- promoter AAAGCGGCCGCCATATCGACCTATTGTAAAA CCGAAAGGAAATCTGGAGATATGACTTTTTGA ΩamtB GCGAAAATTC TTTAACTTTTGATGAAAAGATG gene GTTAAATCAAAAAGTCATATCTCCAGATTTCC AAAGGATCCTTAGTTATAAGCAGATTCTCCGT TTTCGGATGATAATATTATAAAAAC GTTCGTAGATG pNBU2-erm- promoter AAAGCGGCCGCCATATCGACCTATTGTAAAA GAAGAGGTATAGGGAGGAGATATGACTTTTT ΩglnK GCGAAAATTC GATTTAACTTTTGATGAAAAGATG gene GTTAAATCAAAAAGTCATATCTCCTCCCTATA AAAGGATCCTTATCGTTCCTGTTCGGCGTTATA CCTCTTCTTTTTACAGG GAGAG pNBU2-erm- AAAGCGGCCGCCGTTAAATAATATATTATTTT AAAGGATCCTTACTTCACCAAATAGTGATCTA ΩgltBD AAGCTCAAAAAGAAACAAATTACAATCAGG TGTCCGCAG pNBU2-erm- AAAGCGGCCGCAAAGAATTGAATGGTGCTGA AAAGGATCCTTATCCACAGTGGAAATACTTAT ΩglnA ATATGTAGGCC GTACCAGTTCC pNBU2-erm- AAAGCGGCCGCTTGACGGTATGACGATAGTA AAAGGATCCTTATACGATACCCTGAGCCATCA ΩgdhA ATGCTTGTCATTTC TAGCTTTG pNBU2-erm- AAAGCGGCCGCTAAATCTCATTTGCTATCATC AAAGGATCCTTAAATAACGCCTTGTCCCATCA ΩgdhB TTCTTTATAAAGTTTATAATAGGTAATAAGTA TAGCG ATAGG pNBU2-erm- AAAGCGGCCGCTGGCTGCAGAAATCGACCGA AAAGGATCCTTATTTGGTGAACAAGAGTTCCC ΩglnN2 C TATACTTGGG pNBU2-erm- gdhA AAAGCGGCCGCTTGACGGTATGACGATAGTA GAAGATGATAGCAAATGAGATTTATTATACG ΩgdhAB ATGCTTGTCATTTC ATACCCTGAGCCATCATAGCTTTG gdhB CTCAGGGTATCGTATAATAAATCTCATTTGCT AAAGGATCCTTAAATAACGCCTTGTCCCATCA ATCATCTTCTTTATAAAGTTTATAATAGGTAAT TAGCG AAGTAATAG
88 inserted strains were retained (Table 8).
Barcoded strain construction. Barcoding of mutant strain was employed for quantitative-PCR analysis of competition assays between selected mutant and wild type strains. Plasmids for barcoding utilize the pNBU2-tet series previously described
(Martens et al., 2008). The plasmids used in this study are listed in Table 8. Barcoded
strain construction is analogous to complementation strain construction with the use of
tetracycline (2 μg/ml) instead of erythromycin. Plasmid insertion location was confirmed
and only insertions at the att1 site were retained.
Competition Assays. Barcoded mutants were initially streaked for isolated
colonies and inoculated into 10 ml of BHIS medium and grown overnight. One ml of
saturated culture was transferred into a 1.5 ml tube, centrifuged at 10,000 x g for 5
minutes and cells were washed three times in chemically defined medium lacking a
nitrogen source. The cells were ultimately suspended in 1 ml of medium lacking
nitrogen and optical density was measured. Cells were diluted to a final OD of 0.75 with
equal amounts of both wild type and mutant. Cell mixtures were inoculated (100 μl) into
a 96 deep-well plate containing 1 ml of medium with either 1 mM ammonium sulfate or
10 mM ammonium sulfate and incubated at 37°C for 24 hours and each competition
was run in quadruplicate. After 24 hours, 10 μl of saturated culture was inoculated into
fresh medium. The saturated culture was stored at -80°C until the end of the
experiment. The 24-hour incubations and passages were continued for a total of 5 days.
The stored cultures were thawed at room temperature and DNA was extracted using the
HotSHOT method (Truett et al., 2000). Briefly, 10 μl of sample was transferred to a lysis solution (25 mM NaOH, 0.2 mM EDTA) and heated at 99°C for 30 minutes and
89 Table 8. List of assorted primers used for mutant construction, sequencing and qPCR analysis.
Name Purpose Forward Primer Reverse Primer Att1 Confirmation CCTTTGCACCGCTTTCAACG TCAACTAAACATGAGATACTAGC Att2 Confirmation TATCCTATTCTTTAGAGCGCAC GGTGTACCTGGCATTGAAGG Orir6k Confirmation AGGGCTTCTCAGTGCGTTAC GCCGTTGGATACACCAAGGA susCD Confirmation ACAGACCGTAACCGAACAGG TAGCTGCCGGAAAACCGTAG pExchseq Insert Sequencing GTTGTGGACAACAAGCCAGG AGGGAGTGCTACCTCCGTG pNBU2 Confirmation GATGTAACGCACTGAGAAGCCCTTAG CAAGCGGTCGGTTCGGAAATAGAAC BC01 qPCRWT ATGTCGCCAATTGTCACTTTCTCA CACAATATGAGCAACAAGGAATCC BC04 qPCR543 CTCCATAAAGGCGCATACCGACTA CACAATATGAGCAACAAGGAATCC BC06 qPCR785 GATTACGGCGTGATAGATTGGTGT CACAATATGAGCAACAAGGAATCC BC11 qPCR1970 ATGCCGCGGATTTATTGGAAGAAG CACAATATGAGCAACAAGGAATCC BC14 qPCR4339 GGCACGCCATTCTTCATCTAACTG CACAATATGAGCAACAAGGAATCC
90 subsequently transferred to a neutralization solution (20 mM Tris-HCl, pH 5). The resulting DNA was diluted 4-fold and 5 μl was used in a 20 μl quantitative PCR. The qPCR was performed using KAPA SYBR FAST (KAPA Biosystems, Boston MA) in a
Roche Lightcycler (Indianapolis IN) following the manufacturers protocol using the primers listed in Table 8 and the cycling method listed in Table 8. The DNA ratios were calculated by correlating ΔΔCt values to a standard curve of known genomic DNA concentrations.
4.3 Results
To elucidate the contribution of predicted ammonium assimilatory genes to their
respective pathway, gene deletion mutants were screened to identify phenotypic growth
defects or differences. A double recombination event of an engineered pExchange-tdk
plasmid is able to excise the gene of interest and replace the genomic locus with a
clean markerless deletion (Figure 19). Using this method, eight single deletion strains
were obtained in the genes of interest, including three glutamine synthetases
(ΔBT_0543 or ΔglnN1, ΔBT_4339 or ΔglnN2, and ΔBT_0785 or ΔglnA), two glutamate
dehydrogenases (ΔBT_1970 or ΔgdhA, and ΔBT_1973 or ΔgdhB), the glutamate
synthase complex (ΔBT_0552/BT_0553 or ΔgltBD), the ammonium transporter
(ΔBT_0544 or ΔamtB), and the P-II regulatory protein (ΔBT_0543 or ΔglnK). In addition,
a double mutant was constructed for both glutamate dehydrogenase genes, gdhA and
gdhB. To achieve this double mutant, B. thetaiotaomicron ΔgdhA was used as a
conjugation recipient for E. coli S17-1 λpir harboring the pExchange-tdk ΔBT_1973
plasmid. The resulting double deletion strain in both GDH encoding genes was termed
91
Figure 19. Schematic diagram showing deletion of the ammonium transporter gene (amtB) in B. thetaiotaomicron. (A) Diagram of suicide plasmid pExchange-tdk. Diagram of steps during construction of mutants including the conjugation (B) and selection for the first (C) and second (D) recombination events. Abbreviations: bt_tdk, B. thetaiotaomicron thymidine kinase gene; UR, upstream region; DR, downstream region; ermG, erythromycin resistance gene; bla, ampicillin resistance gene; FUdR, 5-flouro-2- deoxyuridine.
92 B. thetaiotaomicron ΔgdhAB. In total, nine deletion strains were constructed, in an effort to identify the genomic features contributing to ammonium assimilation.
To reveal whether a particular gene is conditionally essential, each of the mutant
strains was grown in a chemically defined medium (Appendix A: Table 2) supplied with
a single nitrogen source. When peptides (supplied as Tryptone) were provided, every
mutant construct showed normal growth parameters similar to B. thetaiotaomicron wild-
type (Figure 20A). These results are expected as the predominant nitrogen source
during mutant construction is derived from peptides and protein digest in BHIS. In
contrast, when ammonium is provided as the sole nitrogen source (Figure 21), three
mutants showed a strong impairment in growth. Each mutant was grown in four different
conditions of either carbon excess or limitation and nitrogen excess or limitation.
Glucose was supplied at either 0.5% w/v or 0.1% w/v and ammonium was
supplemented at either 20 mM or 2 mM. First, the B. thetaiotaomicron deficient in the
NADP(H)-dependent glutamate dehydrogenase (ΔgdhA, ΔBT_1970) was markedly
compromised in growth. When glucose was supplied in excess or glucose and
ammonium were both limiting, the organism experienced a longer lag phase but
eventually reached an optical density similar or slightly lower than wild-type. When
glucose was limiting with an excess of ammonium, the ΔgdhA strain was not able to
grow. In addition, the double mutant, ΔgdhAB was also impaired to a similar extent as the ΔgdhA strain. Lastly, the deletion mutant of the second glutamine synthetase,
ΔglnN2, also showed impairment in growth under all conditions but was able to recover in cell density to a similar OD as wild type. The remaining strains all showed a normal growth phenotype similar to the wild-type B. thetaiotaomicron.
93
Figure 20. Growth of B. thetaiotaomicron mutant strains on limiting concentration of peptides. All mutant strains of B. thetaiotaomicron, including (A) deletion strains, (B) complementation strains, and (C) barcoded strains were grown with a nitrogen-limiting concentration of peptides (0.025% w/v Tryptone). Growth was performed in triplicate and error bars denote standard deviation from the mean. Absorbance was monitored at 600nm using Spectronic D20+ spectrophotometer.
94
Figure 21. Growth of B. thetaiotaomicron deletion strains on varying concentrations of glucose and ammonium. The B. thetaiotaomicron wild type and deletion strains were grown with varying concentrations of ammonium and glucose. The WT and mutant strains were grown in combinations of either high (0.5% w/v, A and B) or low (0.1% w/v, C and D) concentrations of glucose and high (20 mM, A and C) or low (2 mM, B and D) concentrations of ammonium. Growth was performed in triplicate and error bars denote standard deviation from the mean. Absorbance was monitored at 600nm using Spectronic D20+ spectrophotometer.
95 To confirm that the observed phenotypes are not a result of unanticipated genetic interactions, like polar effects, complementation with the targeted genes and their corresponding promoter region was performed in trans using the pNBU2-erm system
(Martens et al., 2008). A stable integration of pNBU2-erm plasmids occurs at the att elements within the serine-tRNA synthetase gene, of which B. thetaiotaomicron has two copies. Only genotypes containing the integration at the att1 region were reserved for analysis. The genes encoded within the High Affinity Assimilatory Locus were expressed as a putative operon and therefore sewing PCR reactions were used to engineer a promoter region predicted by the RNA-seq data. For the glnN1, amtB, and glnK genes, the promoter region upstream of dapF was used. The ΔgdhAB double mutant was also engineered by sewing the gdhA gene and putative promoter to the gdhB gene and promoter.
The deletion strains and their complemented counterparts were grown in parallel to determine if wild-type growth could be restored (Figure 22). In relation to the deletion
strains without phenotypic deviations, the complementation with the gene targets did not
influence the growth characteristics, and the growth remained as for the wild type. For
the three deletion strains with impaired growth phenotypes, B. thetaiotaomicron ΔgdhA,
ΔgdhAB, and ΔglnN2, the normal phenotypes have been restored and the
complemented strains grew as the wild-type B. thetaiotaomicron under all conditions
tested. These results confirm that the lack of the NADPH-dependent glutamate dehydrogenase and the second glutamine synthetase type 3 is solely responsible for the impaired growth phenotype. This illustrates the importance of gdhA and glnN2, irrespective of whether the genes are in the native locations or present in trans.
96
Figure 22. Growth of B. thetaiotaomicron complementation strains on varying concentrations of glucose and ammonium. The B. thetaiotaomicron complementation strains were grown with varying concentrations of ammonium and glucose. The mutant strains were grown in combinations of either high (0.5% w/v, A and B) or low (0.1% w/v, C and D) concentrations of glucose and high (20 mM, A and C) or low (2 mM, B and D) concentrations of ammonium. Growth was performed in triplicate and error bars denote standard deviation from the mean. Absorbance was monitored at 600nm using Spectronic D20+ spectrophotometer.
97 Additionally, this provides evidence that the promoter is indeed within 250 bp upstream of the genes of interest.
With three different genes encoding glutamine synthetases, each gene likely plays a unique metabolic role within the bacterium. In order to elucidate the impact of each of the GS genes towards the fitness of the organism, competition analyses was performed using five barcoded mutants, B. thetaiotaomicron WT ΩBC01, ΔglnN1
ΩBC04, ΔglnA ΩBC06, ΔgdhA ΩBC11, and ΔglnN2 ΩBC14. The B. thetaiotaomicron
ΔgdhA ΩBC11 strain was included in the competition analyses to confirm the impaired phenotype. Each mutant was barcoded using pNBU2-tet barcoding plasmids (Table 8).
Prior to performing the competitions, all constructed strains were grown under the same conditions as their unbarcoded counterparts to confirm their phenotypic characteristics (Figure 23). As predicted, the B. thetaiotaomicron WT ΩBC01, ΔglnA
ΩBC06, and ΔglnN1 ΩBC04 strains grew as the wild type strain of B. thetaiotaomicron.
The B. thetaiotaomicron ΔgdhA ΩBC11 retained its impaired growth phenotype, growing just as the unbarcoded B. thetaiotaomicron ΔgdhA strain. Unfortunately, B. thetaiotaomicron ΔglnN2 ΩBC14 strain had a slightly lessened impairment in growth as compared to the non-barcoded strain.
For the competition assays, barcoded strains were grown to saturation in BHIS, washed using fresh medium, mixed in equal proportion based on absorbance measurements, and inoculated into minimal medium containing excess or limiting concentrations of ammonium and 0.5% w/v glucose. Competition assays were performed as batch cultures with daily transfers into fresh medium over five days. Cell
98
Figure 23. Growth of B. thetaiotaomicron barcoded strains on varying concentrations of glucose and ammonium. The B. thetaiotaomicron barcoded strains were grown with varying concentrations of ammonium and glucose. The mutant strains were grown in combinations of either high (0.5% w/v, A and B) or low (0.1% w/v, C and D) concentrations of glucose and high (20 mM, A and C) or low (2 mM, B and D) concentrations of ammonium. Growth was performed in triplicate and error bars denote standard deviation from the mean. Absorbance was monitored at 600nm using Spectronic D20 spectrophotometer.
99 populations were monitored daily via sampling during passage and, the relative
proportion of B. thetaiotaomicron WT ΩBC01 (barcoded wild-type) to each deletion
mutant was quantified using qPCR with barcode specific primers.
Relative proportions of each deletion strain to wild-type are plotted over course of
+ the experiment when grown in high and low NH4 concentrations (Figure 24A, 20 mM;
Figure 24B, 2 mM). The competition between B. thetaiotaomicron ΔglnN1 ΩBC04 or B.
thetaiotaomicron ΔglnA ΩBC06 and B. thetaiotaomicron WT ΩBC01 reveal no competitive inhibition and the deletion mutants remain as a large proportion of the population, >10% after five days. The greatest defect in fitness is expectedly attributed to the B. thetaiotaomicron ΔgdhA ΩBC11 strain, which was rapidly outcompeted within three days at which point the mutant strain was below the limit of detection,
corresponding to a complete growth defect. In opposition, B. thetaiotaomicron ΔglnN2
ΩBC14 shows only a moderate fitness defect that was more pronounced when
ammonium was in excess. Under ammonium excess conditions, the ΔglnN2 ΩBC14
mutant was outcompeted after five days of growth. While in ammonium limitation, the
mutant was still detectable at 5 days of growth, but still drastically depressed. It is
unclear whether the fitness defect results from a deficiency in ammonium assimilation or
in intracellular metabolism.
To identify a phenotype for the deletion of the highly expressed glnN1 glutamine
synthetase, competition assays were additionally conducted against B. thetaiotaomicron
ΔglnA ΩBC06, ΔgdhA ΩBC11, and ΔglnN2 ΩBC14 (Figure 25). Results of these
competitions confirm the results from the WT competitions. Since B. thetaiotaomicron
ΔglnN1 ΩBC04 grows as wild type, there was no competition between this mutant and
100
Figure 24. Competition between selected B. thetaiotaomicron barcoded deletion strains versus the barcoded wild-type strain when cultured on excess and limiting ammonium. Ratio of barcoded mutant strain(ΔBT_0543, ×; ΔBT_0785, ♦; ΔBT_1970, ●; ΔBT_4339, ■) when cocultured with barcoded wild-type B. thetaiotaomicron in chemically defined medium with ammonium at (A) 20 mM or (B) 2 mM concentrations. Cultures in stationary phase were collected every 24 hours and DNA was extracted using the HotSHOT method. Proportions were quantified using qualitative PCR using specific primers towards the integrated pNBU2-tet barcode region. The ΔΔCt values were compared to those of known DNA concentrations. Experiments were performed in triplicate and error bars denote standard deviation from the mean. Asterisk denotes population below the limit of detection.
101
Figure 25. Competition between selected B. thetaiotaomicron barcoded deletion strains versus the B. thetaiotaomicron ΔglnN1 ΩBC04 strain when cultured on excess and limiting ammonium. Proportion of barcoded mutant strain (ΔBT_0785, ♦; ΔBT_1970, ●; ΔBT_4339, ■) when cocultured with barcoded B. thetaiotaomicron ΔBT_0543 in chemically defined medium with ammonium at (A) 20 mM or (B) 2 mM concentrations. Cultures in stationary phase were collected every 24 hours of growth and DNA was extracted using the HotSHOT method. Relative proportions were quantified using qualitative PCR using specific primers towards the integrated pNBU2- tet barcoded region. The ΔΔCt values were compared to those of known DNA concentrations. Experiments were performed in triplicate and error bars denote standard deviation from the mean. Asterisk denotes population below the limit of detection.
102 B. thetaiotaomicron ΔglnA ΩBC06 while the B. thetaiotaomicron ΔgdhA ΩBC11 and
ΔglnN2 ΩBC14 strains were outcompeted at three and five days respectively.
4.4 Discussion
Bacteria are well adapted to respond to ammonium concentrations with the most efficient mechanisms of assimilation with various enzymatic mechanisms that are fine- tuned and specific to the environmental conditions. As Chapter 3 detailed the mode of ammonium acquisition when concentrations are strongly limiting, the results presented here illuminate the mechanism for ammonium assimilation in B. thetaiotaomicron during growth in non-limiting concentrations of ammonium. The importance of glutamate dehydrogenase has been widely reported for Bacteroides and Prevotella species, implicating this singular mechanism as the sole mode of ammonium assimilation in
Bacteroidales. Activity of GDH is high when ammonium is in excess and activity increases when ammonium concentrations decrease (Z. Wen & Morrison, 1996, 1997;
Yamamoto, Abe, et al., 1987; Yamamoto et al., 1984; Yamamoto, Saito, et al., 1987).
Addition of glutamine synthetase inhibitors does not impair the growth of B. fragilis with ammonium as the sole nitrogen source (Yamamoto et al., 1984). Interestingly, Baggio and Morrison constructed a strain of B. thetaiotaomicron with an insertion in the gdhA gene, but no phenotypic deficiencies were observed, possibly due to the addition of cysteine in the medium (Baggio & Morrison, 1996).
If ammonium is the only source of nitrogen present in the medium, the NADPH- dependent glutamine dehydrogenase, gdhA, is responsible for the assimilation of ammonium when concentrations are in excess. Of all the assimilatory genes, gdhA had the high transcript level when ammonium was in excess, 1596 ±142 (Table 5). Once
103 deleted from the genome, a strong negative impact on growth was observed and the barcoded mutant, B. thetaiotaomicron ΔgdhA ΩBC11, had a drastic decrease in fitness
(Figure 21 and Figure 24). Similarly, phenotypic trends have been observed in other model organisms, including Saccharomyces, Klebsiella, and Salmonella, as the loss of an anabolic glutamate dehydrogenase gene impairs growth when ammonium is provide in excess (Berberich, 1972; Brenchley et al., 1975; Brenchley & Magasanik, 1974;
Brenchley et al., 1973; Dendinger et al., 1980; Harth et al., 2005). Taken together with the dramatic loss of fitness observed, our results affirm the importance of Bacteroides’ glutamate dehydrogenase, responsible for ammonium assimilation under non-limiting concentrations. The extreme conservation of this gene in all of the Bacteroides points to its essentiality towards ammonium assimilation and a primary finding of our experimentation.
Additionally, the predicted catabolic gdhB showed no phenotype under varying ammonium concentrations. This gene is homologous to the NADH dependent GDH of
B. fragilis, which is responsive to the concentration of peptides (Abrahams & Abratt,
1998; Abrahams et al., 2001). In line with the RNA-seq results in Chapter 3, the gdhB likely does not participate in ammonium assimilation and is involved in glutamate catabolism. Further evidenced by the inability of gdhB to compensate for the loss of gdhA, and if both gdh genes are simultaneously deleted, the organisms maintains the phenotypic characteristics of the gdhA single mutant strain.
Due to the phylogenetically conserved multiplicity of glutamine synthetases, the role of each gene towards the assimilation of ammonium remains unclear. Historically, screening of phenotypic traits of gene deletions or inactivating mutations has been
104 widely used to assess the mechanisms of nitrogen metabolisms. For example, many organisms including E. coli, S. typhimurium, M. tuberculosis and M. bovis, S. cerevisiae,
K. aerogenes have contributed to the identification genes involved in ammonium assimilation through genetic analysis. In organisms like M. tuberculosis and S. cerevisiae which possess a multiplicity in ammonium assimilation genes, genetic mutational analyses directly identifies genes responsible for nitrogen utilization (Harth et al., 2005; Tang et al., 2011). The gut Bacteroidales encode a multiplicity in glutamine synthetase genes, with two GS type 3 genes and one GS type 1 gene.
The only deletion strain of a glutamine synthetase to possess growth defect is B. thetaiotaomicron ΔglnN2. Whether or not ammonium or glucose are in excess or limiting, the GlnN2 plays an important role in cellular metabolism, without which a growth is impaired (Figure 21). Under non-limiting concentrations of ammonium, glnN2 has higher levels of the transcript compared to glnN1, 294.4 ± 41.2 vs. 76.2 ± 38.7, respectively (Table 5). We propose that the GlnN2 functions in the biosynthesis of glutamine when ammonium is in excess. This finding is a novel proposition for the enteric paradigm of ammonium assimilation. The proteobacterial system employs a single glutamine synthetase, regulated via covalent modification to fine tune the activity for glutamine production and/or ammonium assimilation. Under ammonium limitation, proteobacterial glnA is up regulated transcriptionally and GlnA activity is increased through de-adenylylation to promote ammonium assimilation. In ammonium excess, the activity is decreased through adenylylation but continues to synthesize glutamine (Leigh
& Dodsworth, 2007; Reitzer, 2003). In contrast, the Bacteroidales exploit two separate genes for glutamine synthesis under different environmental conditions. As discussed in
105 Chapter 3, the glnN1 is selectively expressed under ammonium limitation for
assimilation and glutamine synthesis. While during ammonium excess, the glnN2 is
employed for glutamine biosynthesis while glnN1 is repressed.
Most surprisingly, the mutant lacking glnN1 did not have any defect when grown
with low ammonium. In addition to a lack of phenotypic difference, the barcoded mutant,
B. thetaiotaomicron ΔglnN1 ΩBC04, did now show a lack of fitness compared to wild
type in high or low concentrations of ammonium. In continuous culture, the ammonium
concentration falls below 50 µM, therefore expression of the high affinity pathway may not be achieved in batch culture, at least until late log phase of growth. The threshold concentration of ammonium, which elicits expression of the high affinity pathway in
Bacteroides, is still unknown, and further testing is required to identify the conditions under which loss of the glnN1 becomes deleterious. In Bacteroides cellulosilyticus, the high affinity pathway genes (BcellWH2_05255 or amtB, BcellWH2_05244 or glnN1,
BcellWH2_05271 or gltB), including the glnN1 homolog, was shown to provide a fitness
benefit within murine gastrointestinal tract if the mice were fed arabinoxylan (Wu et al.,
2015). Future employment of continuous culture techniques using mutant strains
deficient in genes of the high affinity pathway may enable deeper understanding of B.
thetaiotaomicron deals with the challenges associated with assimilation of ammonium
under nitrogen limitation.
In continuous culture, the GS type 1, glnA, was identified to be preferentially
expressed in ammonium-excess conditions, with a ~10 fold upregulation. The B.
thetaiotaomicron ΔglnA maintains a normal growth phenotype, as well as retaining WT
fitness, indicating that the GS1 is not responsible for ammonium assimilation under
106 + excessive NH4 growth conditions. The Bacteroidales glnA is distinct from the glnA of
Proteobacteria, with the B. thetaiotaomicron and E. coli K12 homologs displaying 25%
identity and 41% similarity at the amino acid level. The only biochemically characterized
Bacteroidetes’ glutamine synthetase type 1 belongs to P. ruminicola 23, which did not
possess any glutamine biosynthesis activity, but did possess γ-glutamyl transferase
activity, indicating recognition of glutamine as a substrate and inability to assimilate
ammonium (J. N. Kim et al., 2012). Our results, compounded with the lack of evidence
for glutamine biosynthetic activity, suggest that GlnA is not involved in ammonium
assimilation, and may function in other cellular processes. This is analogous to the
finding of M. tuberculosis, as GlnA2 was involved in biosynthesis of isoglutamine and D-
glutamine for synthesis of cell wall precursors (Harth et al., 2005).
The use of genetic tools to manipulate the genome of Bacteroides
thetaiotaomicron is a primary advantage of exploring ammonium assimilation in this
organism. Classically, genetic mutants and phenotypic screenings illuminated the genes
involved and their roles in ammonium assimilation in model organism. Using genomic
manipulation and phenotypic screening, we illuminate the NADPH-dependent glutamate
dehydrogenase as the dominant mechanism of ammonium assimilation, and the glnN2
as the biosynthetic source of glutamine, as Bacteroides species encounter
concentration of ammonium similar to those found in the human colon.
107 Chapter 5. Discussion and Future Perspectives
5.1 Introduction
In mammalian systems, nitrogen is in constant flux between the host animal and its bacterial symbionts. In the gastrointestinal tract, most of the nitrogenous compounds, including proteins and urea, are degraded into ammonium, an abundant molecule readily available for bacterial nitrogen acquisitions. Ammonium assimilation is an essential process providing fitness to gut microorganisms. Protein degradation and urea hydrolysis provide the majority of the bioavailable ammonium in the gut and ruminant animals are further supplemented with ammonium salts because of their capacity to transform inorganic nitrogen into host-available microbial proteins. Pioneering work by
Arturi Virtanen demonstrated this capacity by sustaining lactating dairy cows with only urea and ammonium salts as the sole nitrogen sources (Virtanen, 1966). Microbial ammonium assimilation is therefore capable of supplying all of the amino acid requirements of the ruminant. In monogastric animals, nitrogen cycles in the colon can negatively influence the health of the host. Increased ammonium production can occur from high protein consumption and compound metabolic dysfunctions (i.e. liver disease), culminating in hyperammonemia which affects various systems in the host (E.
Kim et al., 2013; Marckmann, Osther, Pedersen, & Jespersen, 2015; Bone, Tamm, &
Hill, 1976; Chung, Fulk, & Slein, 1975).
The Current Enteric Paradigm. The current model of ammonium assimilation in enteric microbes is founded on the early research in Escherichia coli, Klebsiella aerogenes, and Salmonella typhimurium (Van Heeswijk et al., 2013). Environmental
108 ammonium concentration dictates which of two possible pathways is employed for the
acquisition of ammonium. At high environmental concentrations, the quantity of
deprotonated ammonia is sufficient to permeate through the cell membrane and satisfy the bacterial nitrogen requirement. Cytoplasmic ammonium is then reductively aminated onto α-ketoglutarate forming glutamate, in turn, establishing a direct link between the nitrogen cycle and central carbon metabolism. The resulting glutamate participates in numerous transaminase reactions, supplying the amino group to a majority of the amino acids in the cell. The enzyme catalyzing the reductive amination is glutamate dehydrogenase (GDH), which is commonly referred to as the low affinity pathway due to its relatively high Km for ammonium (>1 mM).
Under decreased ammonium concentrations (<<1 mM), ammonium needs
assistance through the membrane via an ammonium transporter, AmtB, accompanied
by a regulatory protein, GlnK. After transport, a high affinity pathway (Km ~0.1 mM) assimilates ammonium, in an energy dependent manner, through two catalytic enzymes. First, a glutamine synthetase (GS) consumes an ATP and appends ammonium onto glutamate, forming glutamine. Second, glutamate synthase (GOGAT) consumes an NADPH and transfers an amine from the glutamine onto α-ketoglutarate, releasing two glutamates.
Model Proteobacteria have further elucidated an elegant regulatory network to control the expression and activity of the two pathways. Covalent modification through an ATase and a UTase modulates the activity of the high affinity pathway, fine tuning ammonium transport and glutamine synthetase activity. Several transcriptional regulators, including an ammonium sensing NTRI and NTRII, add further intricacy to the
109 complex regulation system. The primary pitfall of generalizing the enteric model from
Proteobacteria to all gastrointestinal bacteria is the lack of supportive research in
ammonium assimilation by dominant gut bacteria, primarily Bacteroidetes and
Firmicutes. Of the regulatory enzymes uncovered in organisms like E. coli and
Klebsiella aerogenes, no homologs exist in the dominant gut phylum Bacteroidetes.
Additionally, the multiplicity of catalytic enzymes creates uncertainty toward which enzymes contribute towards ammonium assimilation. Finally, prior research in dominant gastrointestinal microbes has evidenced several contradictions to the current model, which need to be resolved to understand how dominant gut bacteria assimilate ammonium.
5.2 Major finding
In this research, several methodologies are employed to uncover the true mechanisms of ammonium assimilation in dominant gut microorganisms within the genus Bacteroides (Figure 26). The human gut isolate Bacteroides thetaiotaomicron serves as a model for probing the genes employed by Bacteroides and gut Prevotella for the acquisition of ammonium. A bioinformatics investigation of sequenced genomes of Bacteroides and Prevotella reveals the conservation of assimilatory genes, supporting the concept of Bacteroides thetaiotaomicron as a good model for ammonium assimilation in gut Bacteroidales. A whole-genome transcriptional study reveals vital genes differentially expressed in response to ammonium concentration. Finally, the deletion of genes putatively involved in ammonium assimilation reveals the essential components required by Bacteroides to effectively assimilate ammonium.
5.2.1 Bioinformatic Approach
110
Figure 26. Schematic diagram of ammonium assimilation in Bacteroides thetaiotaomicron VPI-5482 under ammonium limitation and ammonium excess.
111 The bioinformatics analysis of Bacteroides and Prevotella ammonium assimilation compared the genes encoded within the Bacteroides thetaiotaomicron genome to those of other sequenced Bacteroides and Prevotella genomes. A collection of Mauve alignments constructed a database for comparison of any B. thetaiotaomicron gene or locus to other Bacteroides genomes, and similarly, a database was constructed for the Prevotella ruminicola 23 genome aligned with other sequenced Prevotella genomes. Using desired genes as a reference, manually scanning though the alignments results in a collection of homologs without annotation biases. In addition, the
Mauve alignments serve as a useful proprietary tool for future analyses.
The comparative genomic analysis confirmed widespread gene multiplicity in
Bacteroides ammonium assimilation genes. This type of genome-wide analysis validates the use of B. thetaiotaomicron as a model for enteric ammonium assimilation.
A large majority of sequenced gut Bacteroides encode three glutamine synthetases genes (glnA, glnN1, and glnN2) and two glutamate dehydrogenases genes (gdhA and gdhB), homologous to those of B. thetaiotaomicron. In addition, the gene organization maintains some level of conservation throughout the genus. In particular, the locus encoding the high affinity pathway, from BT_0543 (glnN1) to BT_0553 (gltB), shows ordered conservation. Altered organization of the genes still maintain a locus encoding the glutamine synthetase, ammonium transporter and P-II regulator, and a second locus encoding the glutamate synthase and asparagine synthase. These finding strengthens the importance of using B. thetaiotaomicron as a model organism, and demonstrate how future findings are applicable to other Bacteroidales within the mammalian gastrointestinal tract.
112 5.2.2 Transcriptional Evidence
The transcriptional plasticity and responsiveness of Bacteroides and Prevotella
species has been reproducibly demonstrated, especially in regards to polysaccharide utilization loci. The advantages of employing RNA-Seq include a large dynamic range and unbiased view of gene transcripts. Continuous cultures allowed for precise control of growth rate, pH, and concentrations of ammonium, glucose, and fermentation product. Using this method, only ammonium concentrations differed between cultures, and ammonium was maintained at concentrations below 50µM, under ammonium limiting conditions. A large sampling number provides excellent statistical power to minimize false positives.
The global transcriptional study revealed the locus encoding the high affinity pathway to be selectively responsive to ammonium concentration. Under ammonium limitation, the high affinity locus encoded the nine most differentially expressed genes, which were also some of the most highly expressed genes in the cell. In accordance with the enteric paradigm, the high affinity pathway was selectively expressed under ammonium-limiting conditions. A global transcriptional analysis utilizing continuous culture techniques exposed a strong transcriptional control of glnN1, which was over
320-fold up-regulated when ammonium was limiting. Under ammonium limitation, the glnN1 gene was also the most highly expressed gene in the cell. In addition to the glnN1 and glutamate synthase (gltBD), accessory enzymes were also high upregulated, including dapLF and asnB, which utilize the products of GS/GOGAT and drive the assimilatory reaction forward. In comparison to the high affinity locus, the transcriptional activity of the genome was relatively unresponsive, with over a log decrease in fold-
113 change values. Some of the mildly responsive genes are responsible for downstream metabolism of glutamate and glutamine, products of the high affinity pathway. As predicted by its localization with other high affinity pathway genes and confirmed through RNA-seq, glnN1 is the primary glutamine synthetase responsible for ammonium assimilation under ammonium limiting conditions.
5.2.3 Evidence from Genetic Manipulations
The primary advantage to using B. thetaiotaomicron is the ability to interrogate the genome through genetic manipulations, historically exploited in other model organisms leading to discoveries of key genes in ammonium assimilation. A loss in a majority of ammonium assimilation genes had no effect on growth, with the exception of the putative NADPH-dependent glutamate dehydrogenase (gdhA) and a glutamine synthetase type 3 (glnN2). Without gdhA or glnN2, growth is severely impaired, a phenotype which is restored by an in trans complementation.
Manipulating the genome of B. thetaiotaomicron provided an opportunity to assess the impact of individual genes towards the fitness of the bacterium. Utilizing the single mutants of each glutamine synthetase gene and gdhA, a 5-day competition assay between individual mutants and the parental strain recapitulated the extreme fitness defect of gdhA and glnN2. The glnN2 deletion mutant was slowly outcompeted by the wild-type barcoded strain after five days. Since glnN1 is highly repressed under ammonium excess, glutamine biosynthesis is performed by glnN2 when ammonium is in excess. Since glnA deletion does not import a fitness defect, the function of this enzyme is still unknown and further experimentation is required.
114 5.3. Implication to the Enteric Paradigm
Through extensive computational, transcriptional, and genetic analyses, the
enteric paradigm of ammonium assimilation is further developed to model the
mechanisms of gut Bacteroidales. If ammonium is readily available, at colonic
concentrations of 2-50 mM, the glutamate dehydrogenase, GdhA, is responsible for the
acquisition of ammonium, and transcription of the GS/GOGAT pathway is strongly
inhibited. In contrast, when ammonium is limiting, a single genetic locus encoding the
entire GS/GOGAT system is highly up-regulated. Through the research, we uncover
accessory enzymes, and intricate details applicable to Bacteroides and Prevotella
species, which provide a greater understanding of the predominant mechanism of
ammonium assimilation in the human and other mammalian gut ecosystems.
5.3.1 Fate of excess extracellular ammonium
Under normal physiological conditions, the luminal concentration of ammonium
ranges from 2-50mM, corresponding to non-limiting or ammonium-excess conditions. At
physiological pH of ~7, gaseous ammonia (NH3, ~1% of total ammonium/ammonia) can
passively diffuse through the cell wall. Based on calculation from Boogerd et al., the
passive diffusion of ammonia is capable of providing a sufficient flux of ammonium to facilitate normal growth (Boogerd et al., 2011). Enteric organisms repress ammonium transporters at higher ammonium concentration (Soupene et al., 1998; Soupene et al.,
2002). Analogously, the B. thetaiotaomicron ammonium transporter, amtB, was suppressed in ammonium excess continuous culture but was the most differentially expressed gene and the 5th most abundant transcript when ammonium was limiting
(Table 5).
115 Once in the cytoplasm, the NADPH-dependent glutamate dehydrogenase is the
primary mechanism of assimilation in excess ammonium conditions as evidenced by the
deletion mutant of gdhA (Figure 21). In Bacteroides and Prevotella species, the GDH
activity is active under ammonium excess and increases when ammonium becomes limiting (Baggio & Morrison, 1996; Z. Wen & Morrison, 1996, 1997; Yamamoto, Abe, et al., 1987; Yamamoto et al., 1984; Yamamoto, Saito, et al., 1987). This is evidenced in our transcriptional profiles, although transcription is high in ammonium excess (TPM of
1596), the transcripts increase three fold when ammonium becomes limiting (TPM of
4823). In competition assays against barcoded wild type B. thetaiotaomicron ΩBC01,
the B. thetaiotaomicron ΔgdhA ΩBC11 strain is outcompeted very rapidly, in 3 day
(Figure 24). Although previously described, the transcriptional response of gdhA to
ammonium limitation is not an implicit feature of the enteric paradigm. For example,
Klebsiella GDH is depressed under ammonium limitation, while Salmonella GDH activity
is constitutive (Brenchley et al., 1975; Brenchley et al., 1973). Therefore, increasing
GDH during ammonium limitation may be a unique feature to Bacteroidales, with three
potential explanations.
First, the constitutive expression of GDH allows this enzyme to participate,
although to a more limited capacity, in the assimilation of ammonium under limiting
concentrations, potentially conserving ATP. Second, the intracellular enzyme
concentrations of GDH could be maintained for rapid acclimation to normal ammonium
concentrations. Limiting concentrations of ammonium may be a transient state in the
colon/mammalian gut. In this scenario, gdhA repression could hinder the transition from
transient ammonium limitation to ammonium excess. Lastly, since α-ketoglutarate is the
116 substrate, this enzyme is well-positioned to provide rapid metabolic response to
changes in ammonium concentration. Since α-ketoglutarate is a key metabolite for both
carbon and nitrogen metabolism, the concentration of this molecule increases when
ammonium or carbon are limiting (Chubukov & Sauer, 2014; Rabinowitz & Kimball,
2007; Radchenko, Thornton, & Merrick, 2010; Reyes-Ramirez, Little, & Dixon, 2001;
Yan, Lenz, & Hwa, 2011; Yuan et al., 2009; Zhang, Wei, & Ye, 2013). The ammonium
transport regulatory protein, GlnK, is able to sense α-ketoglutarate concentrations and
modulate ammonium transport accordingly to prevent excess ATP expenditure and
futile cycling (Adler, Purich, & Stadtman, 1975; da Rocha et al., 2013; Gerhardt et al.,
2012; Jiang & Ninfa, 2007; Radchenko et al., 2010). As a result, the kinetic parameters
are an important factor in understanding the impact on ammonium assimilation and
regulatory mechanisms.
Most importantly, the transcription of the high affinity assimilatory locus (HAAL) is
strongly repressed under ammonium excess, with an average TPM of 8440 under
ammonium excess and only 34.1 TPM under ammonium limitation. More importantly, the glutamine synthetase type 3 that exists outside of the locus, glnN2, follows a similar
transcriptional pattern to gdhA, increasing transcription 3-fold under ammonium
limitation (Table 5). The type 1 glutamine synthetase, glnA, is actually 9.43 fold more
transcriptionally active in ammonium excess. Although the glnA is preferentially
expressed under ammonium excess, deletion of this gene did not influence growth. In
contrast, the deletion of glnN2 inhibited growth on ammonium as well as decreased the
fitness of the mutant strain relative to B. thetaiotaomicron WT ΩBC01and ΔglnN1
ΩBC04 strains.
117 If the HAAL, encoding glnN1 is strongly repressed under ammonium excess, the organism still requires a biosynthetic source of glutamine. Therefore, either glnA or glnN2 must be responsible for glutamine synthesis. Based on the deletion analysis, removal of glnA does not impact growth while glnN2 is essential to maintain a normal growth phenotype. Additionally, there has been no biochemical evidence for glutamine biosynthetic activity in the glnA homologs of Bacteroidetes, with the only known biochemical characterized gene only possessing glutamyl-transferase activity (J. N. Kim et al., 2012). Lastly, the conservation of glnN2 homologs is more stringent, if only slightly, in Bacteroides and glutamine synthetase type 3 genes are more conserved than glutamine synthetase type 1 genes in Prevotella (Figure 5, Figure 7, and Figure 9).
Cumulatively this leads to the hypothesis that Bacteroidales employ glutamine synthetase type 3 enzymes for the assimilation of ammonium and selectively employ either glnN1 and glnN2 genes under specific conditions. The GlnN1 selectively assimilates ammonium under ammonium excess and GlnN2 functions in the biosynthesis of glutamine.
Additional enzymes exist that can incorporate free ammonium into cell metabolites, which warrant attention. For instance, the biosynthesis of meso-2,6- diaminopimelate (mDAP), a precursor for lysine and peptidoglycan synthesis, employs one of two potential catalytic pathways, either meso-diaminopimelate dehydrogenase
(Ddh) or the aminotransferase pathway (DapLF, discussed in section 5.3.2) (Born &
Blanchard, 1999). Importantly, the ddh, encoded by BT_1979, is located upstream of gdhA and is not differentially regulated in response to ammonium (1.62-fold increase under ammonium limitation). The Ddh catalyzes the reductive amination of L-2-amino-6-
118 oxopimelate, consuming a molecule of ammonia and NADPH, catalytically analogous to glutamate dehydrogenase. Employment of Ddh utilizes and NADPH for direct assimilation of ammonium into mDAP in lieu of utilizing the DapLF pathway which would consume glutamate. It is unclear why the Ddh pathway is preferred as it is equally energetically inexpensive. Most likely, the use of free ammonium prevents consumption of important intracellular nitrogen-containing metabolites. Therefore, the intracellular glutamate pools are unencumbered. Prior research demonstrates the importance of maintaining intracellular glutamate and glutamine pools, as S. typhimurium GOGAT- negative strain mutates both GDH and GS to maintain normal intracellular glutamate/glutamine pools (Yan, 2007)
The second enzyme capable of directly assimilating ammonium is involved in asparagine biosynthesis. The production of asparagine can be produced by two enzymes in B. thetaiotaomicron, asparagine synthase A (AsnA, encoded by BT_2129) or asparagine synthase B (AsnB), originally characterized in E. coli (Humbert & Simoni,
1980). While AsnB utilizes glutamine as the amide donor, AsnA catalyzes the amidation of aspartate using ammonia and ATP. First the aspartate is activated by formation of an aminoacyl-AMP, followed by nucleophilic attack by ammonia. Under ammonium excess, the intracellular concentration must be sufficient to inhibit employment asnB (discussed in section 5.3.2), and ammonium is directly assimilated onto aspartate, instead of utilizing AsnB which would consume glutamine and therefore the extra ATP used to produce glutamine. In E. coli, the AsnA is responsible for asparagine synthesis under ammonium excess but down-regulated under ammonium limitation (Felton, Michaelis, &
119 Wright, 1980). In contrast, B. thetaiotaomicron increases expression under ammonium
limitation 3-fold, following the regulatory pattern of gdhA.
All three gene coding for enzymes predicted to assimilate ammonium, asnA, ddh,
and gdhA, are transcribed at a high level when ammonium is in excess, TPMs of 629.4
± 80.3, 959.0 ± 133, and 1596 ± 142 respectively. Based on the importance of
glutamate as the amide donor for most aminotransferase reactions, especially
glutamine, which is the nitrogen donor for purine and pyrimidine synthesis, GDH is likely
to be the primary mechanism of ammonium assimilation into microbial protein. Since
mDAP is predominantly used for cell wall and lysine biosynthesis, while asparagine is
not an amide donor for aminotransferase reactions, these genes are crucial for cell
growth and are likely employed to take advantage of higher ammonium concentrations
to circumvent utilizing the intracellular nitrogen metabolites. Further analyses are
necessary to determine the contribution of each enzyme to intracellular nitrogen.
5.3.2 Fate of limiting extracellular ammonium
If ammonium is depleted from the environment, although not commonly found in
mammalian gut systems, bacteria require a pathway to force the small amount of
ammonium available into bacterial biomass. The pKa of ammonium/ammonia only allows 1% of the total to exist in the freely diffusible gaseous form, mandating a transporter under limiting-ammonium concentrations. Additionally, the binding affinity of glutamate dehydrogenase towards ammonium is too low to maintain the forward reaction (Sharkey & Engel, 2008). Under these circumstances, additional catalytic machinery as well as higher energy input are required to assimilate assimilation.
120 Ammonium transport. Early work on ammonium transport described integral
transmembrane proteins found adjacent to glnK, encoding the P-II regulatory protein,
from Saccharomyces cerevisiae (Marini, Vissers, Urrestarazu, & Andre, 1994),
Arabidopsis thaliana (Ninnemann, Jauniaux, & Frommer, 1994), Corynebacterium
glutamicum (Siewe et al., 1996), Escherichia coli (van Heeswijk et al., 1996), and
Rhizobium etli (Tate, Riccio, Merrick, & Patriarca, 1998). In parallel, the glnK gene is high transcribed under ammonium limiting conditions and commonly identified in tandem with amtB, and even exists as a fusion gene in gut Firmicutes, i.e.
Ruminococcus albus (Arcondeguy et al., 2001; J.N. Kim et al., 2014). The amtB gene encoding the transporter is selectively expressed when ammonium concentrations are low (Boogerd et al., 2011; M. Kim et al., 2012; Soupene et al., 2002). Regulation of the transporter occurs via GlnK, essentially acting as a plug inserted into the AmtB channel to prevent excess ammonium influx.
In the B. thetaiotaomicron transcriptional profile, amtB along with glnK are strongly up-regulated in the ammonium limited continuous cultures, and are likely responsible for the procurement of exogenous ammonium. It is still unclear which species of ammonium/ammonia is transported, whether ammonium is deprotonated prior to transport, or if protonated ammonium itself is moved across the membrane. In the enteric paradigm, an unmodified GlnK inserts a loop into AmtB to prevent ammonium transport, and when GlnK is uridylylated via a UTase, it can no longer bind to AmtB and ammonium enters the cytoplasm (Van Heeswijk et al., 2013). The ammonium transporter of Ruminoccus albus is linked to the GlnK in a single polypeptide, evidencing a tight interaction between these two proteins. Duplication of a
121 glnK gene in E. coli has elaborated deeper levels of regulation to occur based on glnB, the glnK homolog. The GlnB protein additionally regulates ATase, covalent modifier of glutamine synthetase. The ATase and UTase genes are entirely absent in
Bacteroidetes, but a method of regulating the GlnK activity is still important to regulate intracellular ammonium levels, especially when the bacterium encounters high levels of ammonium after nitrogen starvation. Important intracellular metabolites (ATP, ADP, and
α-ketoglutarate) have been shown to moderate interactions between GlnK and AmtB, fine-tuning the transport of ammonium (Durand & Merrick, 2006; Gruswitz, O'Connell, &
Stroud, 2007; Radchenko et al., 2010).
The gene directly adjacent to glnK, BT_ 0546, encodes a hypothetical protein that is co-expressed with the high affinity pathway, but shows no homology to any known protein. This protein is likely periplasmic or extracellular, as it encodes a signal peptide, causing it to be the only protein in the high affinity locus to be secreted.
Additionally, this protein is conserved with the glnK/amtB genes (Figure 5). These point to a role in the transport of ammonium. A tempting hypothesis is to speculate an interaction with the AmtB at the periplasmic face, possibly facilitating the transport of ammonium or preventing it. No known homolog exists and characterization of this protein may lead to novel features of ammonium transport, especially in the
Bacteroides.
Catalytic enzymes. Most convincingly, our transcriptional results highlight the high affinity assimilatory locus (HAAL) as critical for ammonium assimilation under ammonium limitation. Encoded within this locus, the classical GS/GOGAT enzymes populate the high affinity pathway, complemented with meso-diaminopimelate and
122 arginine biosynthetic genes. As most of the genome is relatively transcriptionally
unresponsive to environmental ammonium concentration, the HAAL is over 2 orders of
magnitude more expressed in ammonium limited continuous culture as compared to
ammonium excess. Two separate putative operons are transcribed within this locus.
First, the GS3-containing operon encodes the glutamine synthetase (glnN1),
ammonium transporter (amtB), and regulatory P-II (glnK), a hypothetical protein
(BT_0456), and two biosynthetic genes dapL and dapF. The glutamine synthetase
encoded here is the primary enzyme responsible for appending free ammonium onto an
intracellular metabolite, glutamate. While GlnN2 is responsible for glutamine
biosynthesis when ammonium is high, B. thetaiotaomicron increases glnN1 in response
to ammonium limitation. Prior research describing multiplicity in glutamine synthetases
was performed in Mycobacterium tuberculosis, which only employs the high affinity
system (GS/GOGAT) to assimilate ammonium, regardless of environmental nitrogen
conditions (Harth et al., 2005). In this system, only one glutamine synthetase is required
for ammonium assimilation, under all ammonium concentrations. In contrast, B.
thetaiotaomicron employs glnN1 only under ammonium limitation, while glnN2 is
constitutively expressed for glutamine synthesis.
The dapL and dapF genes are employed to produce m-DAP, a precursor for cell wall and lysine synthesis, utilizing glutamate as the amine donor (Born & Blanchard,
1999). These genes perform the function of Ddh (described in Section 5.3.1), which utilizes ammonium directly as the amine source. These genes have been implicated to be important in ammonium assimilation under ammonium excess in Prevotella ruminicola 23 (J. N. Kim et al., 2017). The rationale behind appending the dapLF genes
123 to the high affinity pathway may be two-fold. First, the diaminopimelate dehydrogenase
may not possess the catalytic capacity for sufficient mDAP production from decreased intracellular ammonium concentrations under limiting conditions. Therefore, utilization of glutamate, an abundant intracellular metabolite, is critical to maintain enough mDAP
production for sufficient peptidoglycan and lysine synthesis. Second, DapLF may drive
the GOGAT reaction towards increased glutamate production (and ultimately driving
ammonium assimilation) by consuming the products of the GS/GOGAT pathway and
releasing any potential product inhibition.
Second, a three-gene operon upstream of the glnN1-encoding operon codes for the GOGAT (gltBD) but also has an additional biosynthetic gene, the asparagine synthetase asnB. The asnB consumes glutamine, ATP, and aspartate in order to produce asparagine. As discussed in the previous section, the AsnA is responsible for asparagine production directly utilizing ammonium, but under ammonium limitation, the asnB is strongly up-regulated. In opposition to AsnA, AsnB consumes glutamine
(produced consuming an ATP) and an additional ATP, which is more energy expensive.
One hypothesis as to why B. thetaiotaomicron encodes both genes is that the affinity of
AsnA towards ammonium is not sufficiently high enough when ammonium concentrations become limiting. Additionally, AsnB consumes glutamine and forces the reaction of ammonium assimilation forward.
In addition to the classical high affinity pathway of glutamine synthetase and glutamate synthase, the dapLF and asnB are also expressed in response to low ammonium concentrations, and should be considered as complementary enzymes to ammonium assimilation. The consumption of glutamine and glutamate by these proteins
124 may contribute to decreased product inhibition, maintaining high rates of ammonium
assimilation at very low concentration of ammonium, and sustaining intracellular amino
acid pools. Further experimentation will surely focus on the mechanism of
transcriptional regulation, whether there is transcriptional activation, repression, and the
transcriptional regulators employed.
5.3.3 Relating to the host
Ultimately, the balance between the ammonium assimilation pathways is a
response to environmental conditions. Within the gastrointestinal tract, Bacteroides
species encounter ammonium concentrations in excess. Ammonium is released from
microbial degradation of dietary proteins, endogenous host-secreted protein, and urea.
Increases in protein intake lead to an increase in colonic ammonium levels (Cummings,
Hill, Bone, Branch, & Jenkins, 1979; Geypens et al., 1997; Kelsay, Behall, & Prather,
1978; Silvester, Bingham, Pollock, Cummings, & O'Neill, 1997). Since concentrations of
ammonium are in excess, the NADPH glutamate dehydrogenase would be the primary mechanism of ammonium assimilation in the gut. Indeed, Forsythe and Parker describe
ammonium assimilation in the rabbit caecum to be assimilated into α-ketoglutarate by
NADPH dependent GDH activity (Forsythe & Parker, 1985). In addition, we demonstrate
the highly conserved GdhA to be the primary gene from Bacteroides to accomplish this
task.
As we demonstrated in Chapter 3, B. thetaiotaomicron is able to utilize peptides
and ammonium simultaneously (Figure 13). This is likely to be occurring in the colon as
well, evidenced by the presence of peptide transporters in B. thetaiotaomicron. While
peptide concentrations fluctuate in the gastrointestinal tract, ammonium concentrations
125 are less affected; therefore, gut organisms likely employ ammonium assimilatory
mechanisms continuously and co-metabolize peptides whenever available. As
demonstrated in R. albus 8, when grown on peptides without ammonium, the
ammonium assimilatory genes are upregulated as the organism attempts to co utilized
both nitrogen sources (J.N. Kim et al., 2014). Fibrolytic microbes play a major role in
regulating ammonium levels in the gut. A decrease in colonic ammonium levels occurs
during increased dietary fiber intake (Kawasaki, Min, Li, Hasegawa, & Sakaguchi,
2015). Gut Bacteroides likely play a pivotal role by ammonium assimilation during fiber
degradation in the colon. Modulation of the microbiome is recommended during the
treatment of hyperammonemia, through either antibiotics, laxatives, fiber, and/or
decreased protein intake (Al Sibae & McGuire, 2009; Alfadhel et al., 2016; Iwasa et al.,
2005; Liu et al., 2004; Riordan & Williams, 1997). Even addition of fiber to swine diets
decreases ammonium concentrations in fecal excretions (Philippe, Laitat, Wavreille,
Nicks, & Cabaraux, 2015). When ammonium concentrations are high, GDH is pivotal for ammonium assimilation, but the high affinity pathway must also play an important role as gut Bacteroides and Prevotella species conserve the HAAL.
Wu et al. performed a key experiment shedding light on the importance of the
HAA Locus in Bacteroides by performing an in vivo competition assay of transposon mutagenesis libraries of Bacteroides cellulosilyticus in a simplified community (Wu et al., 2015). They report that when arabinoxylan is supplemented with a high fat/high simple sugar (HF/HS) diet, the HAA Locus of B. cellulosilyticus provides a fitness benefit when compared to an unsupplemented HF/HS diet. This provides evidence for conditions in the mammalian gastrointestinal tract leading to an ammonium-limited
126 state. Since western diets primarily contain easily digestible nutrients (high in fat, high in sugar), colonic microbiota experience fluctuations in nitrogen availability. If excess fiber is ingested, ammonium may be depleted by the colonic bacteria and the high affinity pathway of ammonium assimilation can become important factors for nitrogen acquisition.
Additionally, Bacteroides species have been identified in infant undernutrition and are positively associated with weight loss (Charbonneau et al., 2016; Santacruz et al.,
2009). During malnutrition, nitrogen becomes less available for the host, decreasing the nitrogen availability for the colonic microbiota. Organisms capable of utilizing ammonium in the most energy efficient manner would have a growth advantage especially when nitrogen and carbon supply are low with a high level of competition.
Therefore, Bacteroides species also up-regulate the NADPH dependent GDH as well as the high affinity pathway, previously described and highlighted by our RNA-Seq results
(Yamamoto, Abe, et al., 1987; Yamamoto et al., 1984; Yamamoto, Saito, et al., 1987).
This allows the organism to minimize energy expenditure for the assimilation of ammonium and maintain its colonization state.
Microbes existing in the gastrointestinal tract must also survive outside of the body, regardless of whether or not it succeeds in colonizing a new host. Therefore,
Bacteroides species must survive in wastewater and other environments after leaving the colon. In this manner, Bacteroides are commonly used as markers for fecal contamination of water sources (Converse, Blackwood, Kirs, Griffith, & Noble, 2009). In ex vivo environments, nitrogen availability is less consistent than in the host and organisms need high metabolic plasticity, especially in regards to carbon and nitrogen
127 acquisition. While wastewater commonly contain mM ranges of ammonium, other environmental conditions may vary, i.e. contaminated freshwater (Henze, 2002). Here, the organism is required to modulate the ammonium assimilation pathways between the low affinity GDH and the high affinity GS/GOGAT systems in order to survive in unpredictable environments. The transcriptional response of B. thetaiotaomicron reveals its robust capacity for surviving in nutrient limited environments, with ammonium concentrations in the low µM range.
5.4 Future Prospective
Uncovering the mechanism for ammonium assimilation under ammonium excess and ammonium-limited condition establishes a starting point for future lines of research.
Many questions remain, especially in the domain of regulatory mechanisms, protein biochemistry, cell physiology, and host interactions. Through our work, we have established a foundation for understanding the mechanisms prominent gut
Bacteroidales employ towards ammonium assimilation, a reference for future investigations into dietary manipulations and therapeutic strategies.
Since ammonium is incorporated into intracellular metabolites, the specific pathway employed for assimilation impacts α-ketoglutarate, glutamate, and glutamine.
Modulation of enzymatic activities, performed here through gene deletions, can have broad effects on metabolite concentrations, fluxes, and subsequent cell growth. The next phase of experimentation will identify the concentration of intracellular metabolites involved in ammonium assimilation and nitrogen cycling. Identifying metabolite concentrations within various strains under different conditions will not only illuminate the pathway employed for ammonium assimilation, but will describe how the bacterium
128 controls nitrogen cycles in the cell. One hypothesis predicts the bacterium to modulate
enzymatic activities (potentially through transcriptional regulation) to maintain
intracellular concentration of glutamate and α-ketoglutarate, as in S. typhimurium (Yan,
2007). Alternatively, metabolite concentrations may change in response to changes in
genes employed for ammonium assimilation. For example, the deletion strain of B.
thetaiotaomicron ΔgdhA may have a decreased concentration of glutamate compared to the wild-type due to deficiencies in enzymatic activities. Metabolite analyses will be performed in efforts to complete this research for publication.
While the mechanism of regulation is elaborated and well-detailed in proteobacterial systems, there are no homologs of the regulatory genes within the dominant gut bacteria. The extension of the enteric paradigm to Bacteroidetes only applies to the functional genes, and the regulatory mechanisms are completely unknown. The covalent modification of the glutamine synthetase is a key step in regulation in many organisms, and whether Bacteroides species participate in this type of regulation remains unknown. Fortunately, the genetic manipulation of B. thetaiotaomicron allows for testing a wide variety of hypotheses through the introduction of His-tagged or labeled genes. If phosphorylated molecules are used to tag the glutamine synthetase, then growing the organism in isotopic 32P and purifying proteins
to identify labeling will provide evidence for the modification.
The transcriptional regulators of Proteobacteria have no homologs in
Bacteroidetes. The NTRI/NTRII system of E. coli directly senses intracellular ammonium
and regulates the transcription of the GS/GOGAT system. Bacteroides commonly employ two component systems and hybrid two component systems to regulate PULs,
129 and it is possible for one such system to regulate the expression of the high affinity
locus. If cells are cultured in continuous systems under ammonium-limitation and
ammonium-excess, the cell extracts can be passed over biotinylated probes designed based on the promoter region identified from the RNA-seq profiles. Using protein pull-
down assays, it is possible to identify whether the regulatory proteins bind to activate or
repress transcription. Further biochemical characterization of the transcriptional
regulators can reveal the ligand(s), either ammonium itself or other intracellular
metabolites, which activates the regulator.
To understand the intracellular catalytic machinery, it is important to elucidate
how each protein catalyzes their respective reactions and interact with other cellular
components. The catalytic parameters, such as Km and kcat, detail the impact to
metabolite fluxes and amino acid biosynthesis through each pathway within their
respective environmental conditions. Regulatory mechanisms between pathways may
occur through protein-protein interactions, modulating enzymes activity, most evidently
between AmtB and GlnK. Protein-protein interactions can also reveal if any functional
proteins bind and impact any potential regulatory/transcriptional proteins. Understanding
the impact of DapLF and AsnB towards the high affinity pathway can be ascertained
through identifying the biochemical properties of each protein at various metabolite
concentrations.
Ultimately, the goal is to understand how an organism lives and interacts with the
host. Each of the genes involved in the ammonium assimilation pathway provide some
level of fitness to the organism, otherwise the genes would have been deleted through
evolutionary selection. Understanding the framework in which the proteins facilitate
130 survival should be done in vivo. Mouse models, although not an ideal substitution for the native host, can elucidate the context in which a particular gene provides the bacterium with an advantage. To determine the most beneficial genes for host colonization, gnotobiotic mice can be inoculated with a mixed population of barcoding genetic variants produced in our work, followed by tracking the population through quantitative-
PCR.
After a thorough investigation into the regulatory properties of ammonium assimilation in Bacteroides, therapeutic bacteria can be synthetically engineered to relieve disease symptoms associated with elevated ammonium levels. For example, in cases of hyperammonemia, hepatic disease, and urea cycle dysfunction, supplementation with a probiotic organism capable of efficient ammonium assimilation would be able to forcibly assimilate ammonium rapidly and store the nitrogen as cell biomass. Alternatively, a strain can be engineered that is capable of assimilating the ammonium into molecules/metabolites that would be unavailable for future bacterial degradation in the colon, i.e. intracellular protein or amino acid polymers. Ultimately, the potential toxic effects of systemic ammonium poisoning can be alleviated. Modulation of the microbial community could also be enhanced through nutraceutical agents to moderate Bacteroides populations through means other than polysaccharide supplementation, and potentially through nitrogen source modulation.
5.5 Conclusion
Nitrogen metabolism in gut systems remains poorly studied in spite of its importance for microbial growth and its implications for the metabolism of the host. The current study expands the enteric paradigm to dominant gut microorganisms by
131 extensive comparative genomics, whole genome transcriptome analysis, and selective
deletion of genes relevant to nitrogen assimilation. Through our work, we have
confirmed the mechanisms employed by Bacteroides species to assimilate ammonium
under varying environmental conditions. As Bacteroides spp. dominate in the human gut
and have been correlated with both healthy and disease states, the mechanistic
understanding of its interaction with energy and nutrient resources provides a sound
basis for precise therapeutic modulation when dysbiosis occurs. In this sense, nitrogen
assimilation, the coupling engine between carbon and nitrogen metabolisms, is the
process required for amino acids production, and therefore biomass generation.
Our main finding, which places Bacteroides thetaiotaiomicron in line with other
enteric species with regard to the mechanism of nitrogen assimilation, opens interesting
ecological questions in relation to differences in the regulation of the enzymes relevant
to the process previously observed in Prevotella in the context of a different
environment, the bovine rumen. Previous studies suggested that Bacteroides and
Prevotella might be anti-correlated in population-level metagenomic surveys
(Gorvitovskaia, Holmes, & Huse, 2016; Hjorth et al., 2017). When species of both genera co-occur in the gut, one or the other predominates, raising questions about their niches and interactions. One of the driving factors regulating biological function in a given environment is resource competition. In this context, differential regulation of nitrogen assimilation genes might be fundamental to predicting ecological success.
Nevertheless, whether the observed differential regulation responds to the environment or is species-dependent remains to be elucidated.
132 The advance in the knowledge illustrated though the present work lays the foundation for future comparative mechanistic studies regarding nitrogen assimilation in other dominant gut Bacteroidetes, and sheds light onto the expression and regulation of an expanded catalog of genes implicated in nitrogen metabolism in the model species
Bacteroides thetaiotaomicron.
133 References
Abe, S., Yasumura, A., & Tanaka, T. (2009). Regulation of Bacillus subtilis aprE
Expression by glnA through Inhibition of scoC and sigmaD-Dependent degR
Expression. J. Bacteriol., 191(9), 3050-8.
Abou Akkada, A. R., & Blackburn, T. H. (1963). Some observations on the nitrogen
metabolism of rumen proteolytic bacteria. J. Gen. Microbiol., 31, 461-9.
Abrahams, G. L., & Abratt, V. R. (1998). The NADH-dependent glutamate
dehydrogenase enzyme of Bacteroides fragilis Bf1 is induced by peptides in the
growth medium. Microbiology, 144 ( Pt 6), 1659-67.
Abrahams, G. L., Iles, K. D., & Abratt, V. R. (2001). The Bacteroides fragilis NAD-
specific glutamate dehydrogenase enzyme is cell surface-associated and
regulated by peptides at the protein level. Anaerobe, 7(3), 135-42.
Abratt, V. R., Mbewe, M., & Woods, D. R. (1998). Cloning of an EF-P homologue from
Bacteroides fragilis that increases B. fragilis glutamine synthetase activity in
Escherichia coli. Mol. Gen. Genet., 258(4), 363-72.
Abratt, V. R., Zappe, H., & Woods, D. R. (1993). A reporter gene vector to investigate
the regulation of glutamine synthetase in Bacteroides fragilis Bf1. J. Gen.
Microbiol., 139(1), 59-65.
Adler, S. P., Purich, D., & Stadtman, E. R. (1975). Cascade control of Escherichia coli
glutamine synthetase. Properties of the PII regulatory protein and the
uridylyltransferase-uridylyl-removing enzyme. J. Biol. Chem., 250(16), 6264-72.
134 Al Sibae, M. R., & McGuire, B. M. (2009). Current trends in the treatment of hepatic
encephalopathy. Ther. Clin. Risk Manag., 5(3), 617-26.
Alfadhel, M., Mutairi, F. A., Makhseed, N., Jasmi, F. A., Al-Thihli, K., Al-Jishi, E., . . .
Ben-Omran, T. (2016). Guidelines for acute management of hyperammonemia in
the Middle East region. Ther. Clin. Risk Manag., 12, 479-87.
Allison, M. J. (1970). Nitrogen metabolism of ruminal microorganism. . Cambride, UK:
Oriel Press.
Almassy, R. J., Janson, C. A., Hamlin, R., Xuong, N. H., & Eisenberg, D. (1986). Novel
subunit-subunit interactions in the structure of glutamine synthetase. Nature,
323(6086), 304-9.
Anders, S., McCarthy, D. J., Chen, Y., Okoniewski, M., Smyth, G. K., Huber, W., &
Robinson, M. D. (2013). Count-based differential expression analysis of RNA
sequencing data using R and Bioconductor. Nat. Protoc., 8(9), 1765-86.
Anders, S., Pyl, P. T., & Huber, W. (2015). HTSeq-a Python framework to work with
high-throughput sequencing data. Bioinformatics, 31(2), 166-9.
Anderson, K. L., & Salyers, A. A. (1989). Biochemical evidence that starch breakdown
by Bacteroides thetaiotaomicron involves outer membrane starch-binding sites
and periplasmic starch-degrading enzymes. J. Bacteriol., 171(6), 3192-8.
Andrade, S. L., & Einsle, O. (2007). The Amt/Mep/Rh family of ammonium transport
proteins. Mol. Membr. Biol., 24(5-6), 357-65.
Andriamihaja, M., Davila, A. M., Eklou-Lawson, M., Petit, N., Delpal, S., Allek, F., . . .
Blachier, F. (2010). Colon luminal content and epithelial cell morphology are
135 markedly modified in rats fed with a high-protein diet. Am. J. Physiol.
Gastrointest. Liver Physiol., 299(5), G1030-G7.
Arcondeguy, T., Jack, R., & Merrick, M. (2001). P(II) signal transduction proteins, pivotal
players in microbial nitrogen control. Microbiol. Mol. Biol. Rev., 65(1), 80-105.
Aspeborg, H., Schrader, J., Coutinho, P. M., Stam, M., Kallas, A., Djerbi, S., . . . Teeri,
T. T. (2005). Carbohydrate-active enzymes involved in the secondary cell wall
biogenesis in hybrid aspen. Plant Physiol, 137(3), 983-97.
Atasoglu, C., Valdes, C., Walker, N. D., Newbold, C. J., & Wallace, R. J. (1998). De
novo synthesis of amino acids by the ruminal bacteria Prevotella bryantii B14,
Selenomonas ruminantium HD4, and Streptococcus bovis ES1. Appl. Environ.
Microbiol., 64(8), 2836-43.
Atkinson, M. R., Blauwkamp, T. A., Bondarenko, V., Studitsky, V., & Ninfa, A. J. (2002).
Activation of the glnA, glnK, and nac promoters as Escherichia coli undergoes
the transition from nitrogen excess growth to nitrogen starvation. J. Bacteriol.,
184(19), 5358-63.
Bach, A., Calsamiglia, S., & Stern, M. D. (2005). Nitrogen metabolism in the rumen. J.
Dairy Sci., 88, E9-21.
Baggio, L., & Morrison, M. (1996). The NAD(P)H-utilizing glutamate dehydrogenase of
Bacteroides thetaiotamicron belongs to enzyme family I, and its activity is
affected by trans-acting gene(s) positioned downstream of gdhA. J. Bacteriol.,
178(24), 7212-20.
Bartley, E. E., Davidovich, A. D., Barr, G. W., Griffel, G. W., Dayton, A. D., Deyoe, C.
W., & Bechtle, R. M. (1976). Ammonia toxicity in cattle. I. Rumen and blood
136 changes associated with toxicity and treatment methods. J. Anim. Sci., 43(4),
835-41.
Batshaw, M. L., Tuchman, M., Summar, M., & Seminara, J. (2014). A longitudinal study
of urea cycle disorders. Mol. Genet. Metab., 113(1-2), 127-30.
Belitsky, B. R., & Sonenshein, A. L. (1998). Role and regulation of Bacillus subtilis
glutamate dehydrogenase genes. J. Bacteriol., 180(23), 6298-305.
Bender, R. A., & Magasanik, B. (1977). Regulatory mutations in the Klebsiella
aerogenes structural gene for glutamine synthetase. J. Bacteriol., 132(1), 100-5.
Bennet, S. M., Ohman, L., & Simren, M. (2015). Gut microbiota as potential
orchestrators of irritable bowel syndrome. Gut Liver, 9(3), 318-31.
Bennett, B. D., Kimball, E. H., Gao, M., Osterhout, R., Van Dien, S. J., & Rabinowitz, J.
D. (2009). Absolute metabolite concentrations and implied enzyme active site
occupancy in Escherichia coli. Nat. Chem. Biol., 5(8), 593-9.
Berberich, M. A. (1972). A glutamate-dependent phenotype in E. coli K12: the result of
two mutations. Biochem. Biophys. Res. Commun., 47(6), 1498-503.
Bernlohr, R. W., Schreier, H. J., & Donohue, T. J. (1984). Enzymes of glutamate and
glutamine biosynthesis in Bacillus licheniformis. Curr. Top. Cell. Regul., 24, 145-
52.
Birkett, A., Muir, J., Phillips, J., Jones, G., & O'Dea, K. (1996). Resistant starch lowers
fecal concentrations of ammonia and phenols in humans. Am. J. Clin. Nutr.,
63(5), 766-72.
137 Blauwkamp, T. A., & Ninfa, A. J. (2002). Physiological role of the GlnK signal
transduction protein of Escherichia coli: survival of nitrogen starvation. Mol.
Microbiol., 46(1), 203-14.
Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: a flexible trimmer for
Illumina sequence data. Bioinformatics, 30(15), 2114-20.
Bone, E., Tamm, A., & Hill, M. (1976). Production of urinary phenols by gut bacteria and
their possible role in causation of large bowel cancer. Am. J. Clin. Nutr., 29(12),
1448-54.
Boogerd, F. C., Ma, H., Bruggeman, F. J., van Heeswijk, W. C., Garcia-Contreras, R.,
Molenaar, D., . . . Westerhoff, H. V. (2011). AmtB-mediated NH3 transport in
prokaryotes must be active and as a consequence regulation of transport by
+ GlnK is mandatory to limit futile cycling of NH4 /NH3. FEBS Lett., 585(1), 23-8.
Born, T. L., & Blanchard, J. S. (1999). Structure/function studies on enzymes in the
diaminopimelate pathway of bacterial cell wall biosynthesis. Curr. Opin. Chem.
Biol., 3(5), 607-13.
Brenchley, J. E., Baker, C. A., & Patil, L. G. (1975). Regulation of the ammonia
assimilatory enzymes in Salmonella typhimurium. J. Bacteriol., 124(1), 182-9.
Brenchley, J. E., & Magasanik, B. (1974). Mutants of Klebsiella aerogenes lacking
glutamate dehydrogenase. J. Bacteriol., 117(2), 544-50.
Brenchley, J. E., Prival, M. J., & Magasanik, B. (1973). Regulation of the synthesis of
enzymes responsible for glutamate formation in Klebsiella aerogenes. J. Biol.
Chem., 248(17), 6122-8.
138 Broderick, G., & Craig, W. M. (1989). Metabolism of peptides and amino acids during in
vitro protein degradation by mixed rumen organisms. J. Dairy Sci., 72(10), 2540-
8.
Broderick, G. A., & Balthrop Jr., J. E. (1979). Chemical inhibition of amino acid
deamination by ruminal microbes in vitro. J. Anim. Sci., 49(9), 1101-11.
Bryant, M. P., & Robinson, I. M. (1962). Some nutritional characteristics of predominant
culturable ruminal bacteria. J. Bacteriol., 84(4), 605-14.
Bryant, M. P., Small, N., Bouma, C., & Chu, H. (1958). Bacteroides ruminicola n. sp.
and Succinimonas amylolytica; the new genus and species; species of succinic
acid-producing anaerobic bacteria of the bovine rumen. J. Bacteriol., 76(1), 15-
23.
Camarena, L., Poggio, S., Garcia, N., & Osorio, A. (1998). Transcriptional repression of
gdhA in Escherichia coli is mediated by the Nac protein. FEMS Microbiol. Lett.,
167(1), 51-6.
Carroll, D. J., Barton, B. A., Anderson, G. W., & Smith, R. D. (1988). Influence of protein
intake and feeding strategy on reproductive performance of dairy cows. J. Dairy
Sci., 71(12), 3470-81.
Castorph, H., & Kleiner, D. (1984). Some properties of a Klebsiella pneumoniae
ammonium transport negative mutant (Amt-). Arch. Microbiol., 139(2-3), 245-7.
Chalupa, W. (1975). Rumen bypass and protection of proteins and amino acids. J.
Dairy. Sci., 58(8), 1198-218.
Chalupa, W. (1976). Degradation of amino acids by the mixed rumen microbial
population. J. Anim. Sci., 43(4), 828-34.
139 Chaney, A. L., & Marbach, E. P. (1962). Modified reagents for determination of urea and
ammonia. Clin. Chem., 8, 130-2.
Charbonneau, M. R., O'Donnell, D., Blanton, L. V., Totten, S. M., Davis, J. C., Barratt,
M. J., . . . Gordon, J. I. (2016). Sialylated Milk Oligosaccharides Promote
Microbiota-Dependent Growth in Models of Infant Undernutrition. Cell, 164(5),
859-71.
Chen, G., Sniffen, C. J., & Russell, J. B. (1987). Concentration and estimated flow of
peptides from the rumen of dairy cattle: effects of protein quantity, protein
solubility, and feeding frequency. J. Dairy Sci., 70(5), 983-92.
Chen, G., Strobel, H. J., Russell, J. B., & Sniffen, C. J. (1987). Effect of hydrophobicity
of utilization of peptides by ruminal bacteria in vitro. Appl. Environ. Microbiol.,
53(9), 2021-5.
Chubukov, V., & Sauer, U. (2014). Environmental dependence of stationary-phase
metabolism in Bacillus subtilis and Escherichia coli. Appl. Environ. Microbiol.,
80(9), 2901-9.
Chung, K. T., Fulk, G. E., & Slein, M. W. (1975). Tryptophanase of fecal flora as a
possible factor in etiology of colon cancer. J. Natl. Cancer Inst., 54(5), 1073-8.
Cole, J. A., Coleman, K. J., Compton, B. E., Kavanagh, B. M., & Keevil, C. W. (1974).
Nitrite and ammonia assimilation by anaerobic continuous cultures of Escherichia
coli. J. Gen. Microbiol., 85(1), 11-22.
Collins, S. M., Surette, M., & Bercik, P. (2012). The interplay between the intestinal
microbiota and the brain. Nat. Rev. Microbiol., 10(11), 735-42.
140 Colombo, G., & Villafranca, J. J. (1986). Amino acid sequence of Escherichia coli
glutamine synthetase deduced from the DNA nucleotide sequence. J. Biol.
Chem., 261(23), 10587-91.
Consortium, T. H. M. (2012). Structure, function and diversity of the healthy human
microbiome. Nature, 486(7402), 207-14.
Converse, R. R., Blackwood, A. D., Kirs, M., Griffith, J. F., & Noble, R. T. (2009). Rapid
QPCR-based assay for fecal Bacteroides spp. as a tool for assessing fecal
contamination in recreational waters. Water Res., 43(19), 4828-37.
Cotta, M. A., & Russell, J. B. (1982). Effect of peptides and amino acids on efficiency of
rumen bacterial protein synthesis in continuous culture. J. Dairy Sci., 65(2), 226-
34.
Coulton, J. W., & Kapoor, M. (1973a). Purification and some properties of the glutamate
dehydrogenase of Salmonella typhimurium. Can. J. Microbiol., 19(4), 427-38.
Coulton, J. W., & Kapoor, M. (1973b). Studies on the kinetics and regulation of
glutamate dehydrogenase of Salmonella typhimurium. Can. J. Microbiol., 19(4),
439-50.
Cummings, J. H., Hill, M. J., Bone, E. S., Branch, W. J., & Jenkins, D. J. (1979). The
effect of meat protein and dietary fiber on colonic function and metabolism. II.
Bacterial metabolites in feces and urine. Amer. J. Clin. Nutr., 32(10), 2094-101.
da Rocha, R. A., Weschenfelder, T. A., de Castilhos, F., de Souza, E. M., Huergo, L. F.,
& Mitchell, D. A. (2013). Mathematical model of the binding of allosteric effectors
to the Escherichia coli PII signal transduction protein GlnB. Biochemistry, 52(15),
2683-93.
141 Darrow, R. A., & Knotts, R. R. (1977). Two forms of glutamine synthetase in free-living
root-nodule bacteria. Biochem. Biophys. Res. Commun., 78(2), 554-9.
De Palma, G., Blennerhassett, P., Lu, J., Deng, Y., Park, A. J., Green, W., . . . Bercik, P.
(2015). Microbiota and host determinants of behavioural phenotype in maternally
separated mice. Nat. Commun., 6, 7735.
Den Besten, G., van Eunen, K., Groen, A. K., Venema, K., Reijngoud, D. J., & Bakker,
B. M. (2013). The role of short-chain fatty acids in the interplay between diet, gut
microbiota, and host energy metabolism. J. Lipid Res., 54(9), 2325-40.
Dendinger, S. M., Patil, L. G., & Brenchley, J. E. (1980). Salmonella typhimurium
mutants with altered glutamate dehydrogenase and glutamate synthase
activities. J. Bacteriol., 141(1), 190-8.
Detsch, C., & Stulke, J. (2003). Ammonium utilization in Bacillus subtilis: transport and
regulatory functions of NrgA and NrgB. Microbiology, 149(11), 3289-97.
Dodd, D., Kiyonari, S., Mackie, R. I., & Cann, I. K. (2010). Functional diversity of four
glycoside hydrolase family 3 enzymes from the rumen bacterium Prevotella
bryantii B14. J. Bacteriol., 192(9), 2335-45.
Dodd, D., Kocherginskaya, S. A., Spies, M. A., Beery, K. E., Abbas, C. A., Mackie, R. I.,
& Cann, I. K. (2009). Biochemical analysis of a b-D-xylosidase and a bifunctional
xylanase-ferulic acid esterase from a xylanolytic gene cluster in Prevotella
ruminicola 23. J. Bacteriol., 191(10), 3328-38.
Dodd, D., Mackie, R. I., & Cann, I. K. (2011). Xylan degradation, a metabolic property
shared by rumen and human colonic Bacteroidetes. Mol. Microbiol., 79(2), 292-
304.
142 Dodd, D., Moon, Y. H., Swaminathan, K., Mackie, R. I., & Cann, I. K. (2010).
Transcriptomic analyses of xylan degradation by Prevotella bryantii and insights
into energy acquisition by xylanolytic bacteroidetes. J. Biol. Chem., 285(39),
30261-73.
Durand, A., & Merrick, M. (2006). In vitro analysis of the Escherichia coli AmtB-GlnK
complex reveals a stoichiometric interaction and sensitivity to ATP and 2-
oxoglutarate. J. Biol. Chem., 281(40), 29558-67.
Eckburg, P. B., Bik, E. M., Bernstein, C. N., Purdom, E., Dethlefsen, L., Sargent, M., . . .
Relman, D. A. (2005). Diversity of the human intestinal microbial flora. Science,
308(5728), 1635-8.
Edmands, J., Noridge, N. A., & Benson, D. R. (1987). The actinorhizal root-nodule
symbiont Frankia sp. strain CpI1 has two glutamine synthetases. Proc. Natl.
Acad. Sci. U.S.A., 84(17), 6126-30.
Felton, J., Michaelis, S., & Wright, A. (1980). Mutations in two unlinked genes are
required to produce asparagine auxotrophy in Escherichia coli. J. Bacteriol.,
142(1), 221-8.
Ferguson, J. D., Blanchard, T., Galligan, D. T., Hoshall, D. C., & Chalupa, W. (1988).
Infertility in dairy cattle fed a high percentage of protein degradable in the rumen.
J. Am. Vet. Med. Assoc., 192(5), 659-62.
Firkins, J. L., Yu, Z., & Morrison, M. (2007). Ruminal nitrogen metabolism: perspectives
for integration of microbiology and nutrition for dairy. J. Dairy Sci., 90, E1-16.
Fisher, S. H. (1999). Regulation of nitrogen metabolism in Bacillus subtilis: vive la
difference! Mol. Microbiol., 32(2), 223-32.
143 Fisher, S. H., & Wray Jr., L. V. (2002). Mutations in the Bacillus subtilis glnRA operon
that cause nitrogen source-dependent defects in regulation of TnrA activity. J.
Bacteriol., 184(16), 4636-9.
Forchhammer, K. (2008). PII signal transducers: novel functional and structural insights.
Trends Microbiol., 16(2), 65-72.
Forsythe, S. J., & Parker, D. S. (1985). Nitrogen metabolism by the microbial flora of the
rabbit caecum. J. Appl. Bacteriol., 58(4), 363-9.
Gallant, J. L., Viljoen, A. J., van Helden, P. D., & Wiid, I. J. (2016). Glutamate
dehydrogenase is required by Mycobacterium bovis BCG for resistance to
cellular stress. PLoS ONE, 11(1), e0147706.
Gerhardt, E. C., Araujo, L. M., Ribeiro, R. R., Chubatsu, L. S., Scarduelli, M., Rodrigues,
T. E., . . . Huergo, L. F. (2012). Influence of the ADP/ATP ratio, 2-oxoglutarate
and divalent ions on Azospirillum brasilense PII protein signalling. Microbiology,
158(Pt 6), 1656-63.
Geypens, B., Claus, D., Evenepoel, P., Hiele, M., Maes, B., Peeters, M., . . . Ghoos, Y.
(1997). Influence of dietary protein supplements on the formation of bacterial
metabolites in the colon. Gut, 41(1), 70-6.
Gibson III, F. C., Onderdonk, A. B., Kasper, D. L., & Tzianabos, A. O. (1998). Cellular
mechanism of intraabdominal abscess formation by Bacteroides fragilis. J.
Immunol., 160(10), 5000-6.
Gorvitovskaia, A., Holmes, S. P., & Huse, S. M. (2016). Interpreting Prevotella and
Bacteroides as biomarkers of diet and lifestyle. Microbiome, 4, 15.
144 Griswold, K. E., & Mackie, R. I. (1997). Degradation of protein and utilization of the
hydrolytic products by a predominant ruminal bacterium, Prevotella ruminicola
B1(4). J. Dairy Sci., 80(1), 167-75.
Gruswitz, F., O'Connell, J., 3rd, & Stroud, R. M. (2007). Inhibitory complex of the
transmembrane ammonia channel, AmtB, and the cytosolic regulatory protein,
GlnK, at 1.96 A. Proc. Natl. Acad. Sci. U.S.A., 104(1), 42-7.
Gunka, K., & Commichau, F. M. (2012). Control of glutamate homeostasis in Bacillus
subtilis: a complex interplay between ammonium assimilation, glutamate
biosynthesis and degradation. Mol. Microbiol., 85(2), 213-24.
Harth, G., Maslesa-Galic, S., Tullius, M. V., & Horwitz, M. A. (2005). All four
Mycobacterium tuberculosis glnA genes encode glutamine synthetase activities
but only GlnA1 is abundantly expressed and essential for bacterial homeostasis.
Mol. Microbiol., 58(4), 1157-72.
Henze, M. (2002). Wastewater treatment : biological and chemical processes (3rd ed.).
Berlin ; New York: Springer.
Hill, R. T., Parker, J. R., Goodman, H. J., Jones, D. T., & Woods, D. R. (1989).
Molecular analysis of a novel glutamine synthetase of the anaerobe Bacteroides
fragilis. J. Gen. Microbiol., 135(12), 3271-9.
Hjorth, M. F., Roager, H. M., Larsen, T. M., Poulsen, S. K., Licht, T. R., Bahl, M. I., . . .
Astrup, A. (2017). Pre-treatment microbial Prevotella-to-Bacteroides ratio,
determines body fat loss success during a 6-month randomized controlled diet
intervention. Int. J. Obes. (Lond).
145 Hong, P. Y., Iakiviak, M., Dodd, D., Zhang, M., Mackie, R. I., & Cann, I. (2014). Two
new xylanases with different substrate specificities from the human gut bacterium
Bacteroides intestinalis DSM 17393. Appl. Environ. Microbiol., 80(7), 2084-93.
Hsiao, E. Y., McBride, S. W., Hsien, S., Sharon, G., Hyde, E. R., McCue, T., . . .
Mazmanian, S. K. (2013). Microbiota modulate behavioral and physiological
abnormalities associated with neurodevelopmental disorders. Cell, 155(7), 1451-
63.
Huergo, L. F., & Dixon, R. (2015). The Emergence of 2-Oxoglutarate as a Master
Regulator Metabolite. Microbiol. Mol. Biol. Rev., 79(4), 419-35.
Humbert, R., & Simoni, R. D. (1980). Genetic and biomedical studies demonstrating a
second gene coding for asparagine synthetase in Escherichia coli. J. Bacteriol.,
142(1), 212-20.
Iakiviak, M., Mackie, R. I., & Cann, I. K. (2011). Functional analyses of multiple lichenin-
degrading enzymes from the rumen bacterium Ruminococcus albus 8. Appl.
Environ. Microbiol., 77(21), 7541-50.
Iwasa, M., Nakao, M., Kato, Y., Kobayashi, Y., Takagi, K., Kaito, M., & Adaci, Y. (2005).
Dietary fiber decreases ammonia levels in patients with cirrhosis. Hepatology,
41(1), 217-8.
Jayakumar, A., Epstein, W., & Barnes, E. M., Jr. (1985). Characterization of ammonium
(methylammonium)/potassium antiport in Escherichia coli. J. Biol. Chem.,
260(12), 7528-32.
146 Jiang, P., & Ninfa, A. J. (2007). Escherichia coli PII signal transduction protein
controlling nitrogen assimilation acts as a sensor of adenylate energy charge in
vitro. Biochemistry, 46(45), 12979-96.
Jones, E. A., Smallwood, R. A., Craigie, A., & Rosenoer, V. M. (1969). The
enterohepatic circulation of urea nitrogen. Clinical Science, 37(3), 825-36.
Joyner, A. E., Jr., & Baldwin, R. L. (1966). Enzymatic studies of pure cultures of rumen
microorganisms. J. Bacteriol., 92(5), 1321-30.
Kabel, M. A., Yeoman, C. J., Han, Y., Dodd, D., Abbas, C. A., de Bont, J. A., . . .
Mackie, R. I. (2011). Biochemical characterization and relative expression levels
of multiple carbohydrate esterases of the xylanolytic rumen bacterium Prevotella
ruminicola 23 grown on an ester-enriched substrate. Appl. Environ. Microbiol.,
77(16), 5671-81.
Kalmokoff, M., Zwicker, B., O'Hara, M., Matias, F., Green, J., Shastri, P., . . . Brooks, S.
P. (2013). Temporal change in the gut community of rats fed high amylose
cornstarch is driven by endogenous urea rather than strictly on carbohydrate
availability. J. Appl. Microbiol., 114(5), 1516-28.
Kane, J. F., Wakim, J., & Fischer, R. S. (1981). Regulation of glutamate dehydrogenase
in Bacillus subtilis. J. Bacteriol., 148(3), 1002-5.
Karlsson, F. H., Fak, F., Nookaew, I., Tremaroli, V., Fagerberg, B., Petranovic, D., . . .
Nielsen, J. (2012). Symptomatic atherosclerosis is associated with an altered gut
metagenome. Nat. Commun., 3, 1245.
147 Kawasaki, K., Min, X., Li, X., Hasegawa, E., & Sakaguchi, E. (2015). Transfer of blood
urea nitrogen to cecal microbial nitrogen is increased by fructo-oligosaccharide
feeding in guinea pigs. Anim. Sci. J., 86(1), 77-82.
Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., . . .
Drummond, A. (2012). Geneious Basic: an integrated and extendable desktop
software platform for the organization and analysis of sequence data.
Bioinformatics, 28(12), 1647-9.
Kelsay, J. L., Behall, K. M., & Prather, E. S. (1978). Effect of fiber from fruits and
vegetables on metabolic responses of human subjects I. Bowel transit time,
number of defecations, fecal weight, urinary excretions of energy and nitrogen
and apparent digestibilities of energy, nitrogen, and fat. Amer. J. Clin. Nutr.,
31(7), 1149-53.
Khademi, S., O'Connell, J., 3rd, Remis, J., Robles-Colmenares, Y., Miercke, L. J., &
Stroud, R. M. (2004). Mechanism of ammonia transport by Amt/MEP/Rh:
structure of AmtB at 1.35 A. Science, 305(5690), 1587-94.
Kim, E., Coelho, D., & Blachier, F. (2013). Review of the association between meat
consumption and risk of colorectal cancer. Nutr. Res., 33(12), 983-94.
Kim, J. N., Cann, I. K., & Mackie, R. I. (2012). Purification, characterization, and
expression of multiple glutamine synthetases from Prevotella ruminicola 23. J.
Bacteriol., 194(1), 176-84.
Kim, J. N., Henriksen, E. D., Cann, I. K., & Mackie, R. I. (2014). Nitrogen utilization and
metabolism in Ruminococcus albus 8. Appl. Environ. Microbiol., 80(10), 3095-
102.
148 Kim, J. N., Mendez-Garcia, C., Geier, R. R., Iakiviak, M., Chang, J., Cann, I., & Mackie,
R. I. (2017). Metabolic networks for nitrogen utilization in Prevotella ruminicola
23. Sci. Rep., 7(1), 7851.
Kim, M., Zhang, Z., Okano, H., Yan, D., Groisman, A., & Hwa, T. (2012). Need-based
activation of ammonium uptake in Escherichia coli. Mol. Syst. Biol., 8, 616.
Kirk, J. M., Woodward, C. L., Ellis, W. C., & Ricke, S. C. (2000). Glutamine synthetase
and protease enzyme activities and growth response of ruminal bacterium
Prevotella ruminicola strain B14 to nitrogen source and concentration. J. Environ.
Sci. Health B, 35(1), 103-20.
Kleiner, D. (1982). Ammonium (methylammonium) transport by Klebsiella pneumoniae.
Biochim. Biophys. Acta., 688(3), 702-8.
+ Kleiner, D. (1985). Energy expenditure for cyclic retention of NH3/NH4 during N2 fixation
by Klebsiella pneumoniae. FEBS Lett., 187(2), 237-9.
Kleiner, D., & Castorph, H. (1982). Inhibition of ammonium (methylammonium) transport
in Klebsiella pneumoniae by glutamine and glutamine analogues. FEBS Lett.,
146(1), 201-3.
Kloosterman, T. G., Hendriksen, W. T., Bijlsma, J. J., Bootsma, H. J., van Hijum, S. A.,
Kok, J., . . . Kuipers, O. P. (2006). Regulation of glutamine and glutamate
metabolism by GlnR and GlnA in Streptococcus pneumoniae. J. Biol. Chem.,
281(35), 25097-109.
Kohn, R. A., & Allen, M. S. (1995). In vitro protein degradation of feeds using
concentrated enzymes extracted from rumen contents. Anim. Feed Sci. Technol.,
52, 15-28.
149 Kopecny, J., & Wallace, R. J. (1982). Cellular location and some properties of
proteolytic enzymes of rumen bacteria. Appl. Environ. Microbiol., 43(5), 1026-33.
Koropatkin, N. M., Martens, E. C., Gordon, J. I., & Smith, T. J. (2008). Starch catabolism
by a prominent human gut symbiont is directed by the recognition of amylose
helices. Structure, 16(7), 1105-15.
Kumada, Y., Takano, E., Nagaoka, K., & Thompson, C. J. (1990). Streptomyces
hygroscopicus has two glutamine synthetase genes. J. Bacteriol., 172(9), 5343-
51.
Langmead, B., & Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2.
Nat. Methods, 9(4), 357-9.
Leigh, J. A., & Dodsworth, J. A. (2007). Nitrogen regulation in bacteria and archaea.
Annu. Rev. Microbiol., 61, 349-77.
Ley, R. E., Backhed, F., Turnbaugh, P., Lozupone, C. A., Knight, R. D., & Gordon, J. I.
(2005). Obesity alters gut microbial ecology. Proc. Natl. Acad. Sci. U.S.A.,
102(31), 11070-5.
Liaw, S. H., & Eisenberg, D. (1994). Structural model for the reaction mechanism of
glutamine synthetase, based on five crystal structures of enzyme-substrate
complexes. Biochemistry, 33(3), 675-81.
Liaw, S. H., Kuo, I., & Eisenberg, D. (1995). Discovery of the ammonium substrate site
on glutamine synthetase, a third cation binding site. Protein Sci., 4(11), 2358-65.
Lin, H. C., & Visek, W. J. (1991). Colon mucosal cell damage by ammonia in rats. J.
Nutr., 121(6), 887-93.
150 Ling, J. R., & Armstead, I. P. (1995). The in vitro uptake and metabolism of peptides
and amino acids by five species of rumen bacteria. J. Appl. Bacteriol., 78(2), 116-
24.
Liu, Q., Duan, Z. P., Ha, D. K., Bengmark, S., Kurtovic, J., & Riordan, S. M. (2004).
Synbiotic modulation of gut flora: effect on minimal hepatic encephalopathy in
patients with cirrhosis. Hepatology, 39(5), 1441-9.
Mackie, R. I., & White, B. A. (1990). Recent advances in rumen microbial ecology and
metabolism: potential impact on nutrient output. J. Dairy Sci., 73(10), 2971-95.
MacNeil, T., MacNeil, D., & Tyler, B. (1982). Fine-structure deletion map and
complementation analysis of the glnA-glnL-glnG region in Escherichia coli. J.
Bacteriol., 150(3), 1302-13.
Magasanik, B. (1989). Regulation of transcription of the glnALG operon of Escherichia
coli by protein phosphorylation. Biochimie, 71(9-10), 1005-12.
Marckmann, P., Osther, P., Pedersen, A. N., & Jespersen, B. (2015). High-protein diets
and renal health. J. Ren. Nutr., 25(1), 1-5.
Marini, A. M., Vissers, S., Urrestarazu, A., & Andre, B. (1994). Cloning and expression
of the MEP1 gene encoding an ammonium transporter in Saccharomyces
cerevisiae. EMBO J., 13(15), 3456-63.
Martens, E. C., Chiang, H. C., & Gordon, J. I. (2008). Mucosal glycan foraging
enhances fitness and transmission of a saccharolytic human gut bacterial
symbiont. Cell Host Microbe, 4(5), 447-57.
151 Martens, E. C., Lowe, E. C., Chiang, H., Pudlo, N. A., Wu, M., McNulty, N. P., . . .
Gordon, J. I. (2011). Recognition and degradation of plant cell wall
polysaccharides by two human gut symbionts. PLoS Biol., 9(12).
Martinelle, K., & Haggstrom, L. (1997). On the dissociation constant of ammonium:
Effects of using an incorrect pKa in calculations of the ammonia concentration in
animal cell cultures. Biotechnol. Tech., 11(8), 549-51.
Mathis, R., Gamas, P., Meyer, Y., & Cullimore, J. V. (2000). The presence of GSI-like
genes in higher plants: support for the paralogous evolution of GSI and GSII
genes. J. Mol. Evol., 50(2), 116-22.
Mathison, G. W., & Milligan, L. P. (1971). Nitrogen metabolism in sheep. Br. J. Nutr.,
25(3), 351-66.
McNulty, N. P., Wu, M., Erickson, A. R., Pan, C., Erickson, B. K., Martens, E. C., . . .
Gordon, J. I. (2013). Effects of diet on resource utilization by a model human gut
microbiota containing Bacteroides cellulosilyticus WH2, a symbiont with an
extensive glycobiome. PLoS Biol., 11(8).
Meers, J. L., Tempest, D. W., & Brown, C. M. (1970). 'Glutamine(amide):2-oxoglutarate
amino transferase oxido-reductase (NADP); an enzyme involved in the synthesis
of glutamate by some bacteria. J. Gen. Microbiol., 64(2), 187-94.
Meier-Wagner, J., Nolden, L., Jakoby, M., Siewe, R., Kramer, R., & Burkovski, A.
(2001). Multiplicity of ammonium uptake systems in Corynebacterium
glutamicum: role of Amt and AmtB. Microbiology, 147(Pt 1), 135-43.
152 Miranda-Rios, J., Sanchez-Pescador, R., Urdea, M., & Covarrubias, A. A. (1987). The
complete nucleotide sequence of the glnALG operon of Escherichia coli K12.
Nucleic Acids Res., 15(6), 2757-70.
Moon, Y. H., Iakiviak, M., Bauer, S., Mackie, R. I., & Cann, I. K. (2011). Biochemical
analyses of multiple endoxylanases from the rumen bacterium Ruminococcus
albus 8 and their synergistic activities with accessory hemicellulose-degrading
enzymes. Appl. Environ. Microbiol., 77(15), 5157-69.
Moreau, M. C., Ducluzeau, R., & Raibaud, P. (1976). Hydrolysis of urea in the
gastrointestinal tract of "monoxenic" rats: effect of immunization with strains of
ureolytic bacteria. Infect. Immun., 13(1), 9-15.
Mowbray, S. L., Kathiravan, M. K., Pandey, A. A., & Odell, L. R. (2014). Inhibition of
glutamine synthetase: a potential drug target in Mycobacterium tuberculosis.
Molecules, 19(9), 13161-76.
Newbold, C. J., Wallace, R. J., & McKain, N. (1990). Effects of the ionophore tetronasin
on nitrogen metabolism by ruminal microorganisms in vitro. J Anim Sci, 68(4),
1103-9.
Nili, N., & Brooker, J. D. (1995). A defined medium for rumen bacteria and identification
of strains impaired in de novo biosynthesis of certain amino acids. Lett. Appl.
Microbiol., 21(2), 69-74.
Ninfa, A. J., & Jiang, P. (2005). PII signal transduction proteins: sensors of alpha-
ketoglutarate that regulate nitrogen metabolism. Curr. Opin. Microbiol., 8(2), 168-
73.
153 Ninnemann, O., Jauniaux, J. C., & Frommer, W. B. (1994). Identification of a high
+ affinity NH4 transporter from plants. EMBO J., 13(15), 3464-71.
Nocek, J. E., & Russell, J. B. (1988). Protein and energy as an integrated system.
Relationship of ruminal protein and carbohydrate availability to microbial
synthesis and milk production. J. Dairy Sci., 71(8), 2070-107.
Ohland, C. L., & Jobin, C. (2015). Microbial activities and intestinal homeostasis: A
delicate balance between health and disease. Cell. Mol. Gastroenterol. Hepatol.,
1(1), 28-40.
Overbeek, R., Olson, R., Pusch, G. D., Olsen, G. J., Davis, J. J., Disz, T., . . . Stevens,
R. (2014). The SEED and the Rapid Annotation of microbial genomes using
Subsystems Technology (RAST). Nucleic Acids Res., 42(Database issue), D206-
14.
Pan, F. L., & Coote, J. G. (1979). Glutamine synthetase and glutamate synthase
activities during growth and sporulation in Bacillus subtilis. J. Gen. Microbiol.,
112(2), 373-7.
Petridis, M., Benjak, A., & Cook, G. M. (2015). Defining the nitrogen regulated
transcriptome of Mycobacterium smegmatis using continuous culture. BMC
Genomics, 16, 821.
Phibbs, P. V., Jr., & Bernlohr, R. W. (1971). Purification, properties, and regulation of
glutamic dehydrogenase of Bacillus licheniformis. J. Bacteriol., 106(2), 375-85.
Philippe, F. X., Laitat, M., Wavreille, J., Nicks, B., & Cabaraux, J. F. (2015). Effects of a
high-fibre diet on ammonia and greenhouse gas emissions from gestating sows
and fattening pigs. Atmos. Environ., 109, 197-204.
154 Pilgrim, A. F., Weller, R. A., Gray, F. V., & Belling, C. B. (1970). Synthesis of microbial
protein from ammonia in the sheep's rumen and the proportion of dietary nitrogen
converted into microbial nitrogen. Br. J. Nutr., 24(2), 589-98.
Pittman, K. A., & Bryant, M. P. (1964). Peptides and other nitrogen sources for growth
of Bacteroides ruminicola. J. Bacteriol., 88(2), 401-10.
Prins, R. A., van Rheenen, D. L., & van't Klooster, A. T. (1983). Characterization of
microbial proteolytic enzymes in the rumen. Antonie Van Leeuwenhoek 49, 585-
95.
Qin, J., Li, R., Raes, J., Arumugam, M., Burgdorf, K. S., Manichanh, C., . . . Ehrlich, S.
D. (2010). A human gut microbial gene catalogue established by metagenomic
sequencing. Nature, 464(7285), 59-65.
+ Raabe, W. (1990). Effects of NH4 on the function of the CNS. Adv. Exp. Med. Biol.,
272, 99-120.
Rabinowitz, J. D., & Kimball, E. (2007). Acidic acetonitrile for cellular metabolome
extraction from Escherichia coli. Anal. Chem., 79(16), 6167-73.
Radchenko, M. V., Thornton, J., & Merrick, M. (2010). Control of AmtB-GlnK complex
formation by intracellular levels of ATP, ADP, and 2-oxoglutarate. J. Biol. Chem.,
285(40), 31037-45.
Reitzer, L. (2003). Nitrogen assimilation and global regulation in Escherichia coli. Annu.
Rev. Microbiol., 57, 155-76.
Reyes-Ramirez, F., Little, R., & Dixon, R. (2001). Role of Escherichia coli nitrogen
regulatory genes in the nitrogen response of the Azotobacter vinelandii NifL-NifA
complex. J. Bacteriol., 183(10), 3076-82.
155 Richardson, A. J., McKain, N., & Wallace, R. J. (2013). Ammonia production by human
faecal bacteria, and the enumeration, isolation and characterization of bacteria
capable of growth on peptides and amino acids. BMC Microbiol., 13, 6.
Riordan, S. M., & Williams, R. (1997). Treatment of hepatic encephalopathy. N. Engl. J.
Med., 337(7), 473-9.
Robinson, M. D., McCarthy, D. J., & Smyth, G. K. (2010). edgeR: a Bioconductor
package for differential expression analysis of digital gene expression data.
Bioinformatics, 26(1), 139-40.
Rogers, T. E., Pudlo, N. A., Koropatkin, N. M., Bell, J. S., Moya Balasch, M., Jasker, K.,
& Martens, E. C. (2013). Dynamic responses of Bacteroides thetaiotaomicron
during growth on glycan mixtures. Mol. Microbiol., 88(5), 876-90.
Rogowski, A., Briggs, J. A., Mortimer, J. C., Tryfona, T., Terrapon, N., Lowe, E. C., . . .
Bolam, D. N. (2015). Glycan complexity dictates microbial resource allocation in
the large intestine. Nat. Commun., 6.
Romagnolo, D., Polan, C. E., & Barbeau, W. E. (1994). Electrophoretic analysis of
ruminal degradability of corn proteins. J. Dairy Sci., 77(4), 1093-9.
Rosario, C. J., & Bender, R. A. (2005). Importance of tetramer formation by the nitrogen
assimilation control protein for strong repression of glutamate dehydrogenase
formation in Klebsiella pneumoniae. J. Bacteriol., 187(24), 8291-9.
Russell, J. B. (1983). Fermentation of peptides by Bacteroides ruminicola B14. Appl.
Environ. Microbiol., 45(5), 1566-74.
Russell, J. B., & Strobel, H. J. (1987). Concentration of ammonia across cell
membranes of mixed rumen bacteria. J. Dairy Sci., 70(5), 970-6.
156 Rydzak, T., Garcia, D., Stevenson, D. M., Sladek, M., Klingeman, D. M., Holwerda, E.
K., . . . Guss, A. M. (2017). Deletion of Type I glutamine synthetase deregulates
nitrogen metabolism and increases ethanol production in Clostridium
thermocellum. Metab. Eng., 41, 182-91.
Sales-Duval, M., Lucas, F., & Blanchart, G. (2002). Effects of exogenous ammonia or
free amino acids on proteolytic activity and protein breakdown products in
Streptococcus bovis, Prevotella albensis, and Butyrivibrio fibrisolvens. Curr.
Microbiol., 44(6), 435-43.
Sales, M., Lucas, F., & Blanchart, G. (2000). Effects of ammonia and amino acids on
the growth and proteolytic activity of three species of rumen bacteria: Prevotella
albensis, Butyrivibrio fibrisolvens, and Streptococcus bovis. Curr. Microbiol.,
40(6), 380-6.
Santacruz, A., Marcos, A., Warnberg, J., Marti, A., Martin-Matillas, M., Campoy, C., . . .
Sanz, Y. (2009). Interplay between weight loss and gut microbiota composition in
overweight adolescents. Obesity (Silver Spring), 17(10), 1906-15.
Schaefer, D. M., Davis, C. L., & Bryant, M. P. (1980). Ammonia saturation constants for
predominant species of rumen bacteria. J. Dairy Sci., 63(8), 1248-63.
Schumacher, M. A., Chinnam, N. B., Cuthbert, B., Tonthat, N. K., & Whitfill, T. (2015).
Structures of regulatory machinery reveal novel molecular mechanisms
controlling B. subtilis nitrogen homeostasis. Genes Dev., 29(4), 451-64.
Schwalm, N. D., 3rd, & Groisman, E. A. (2017). Navigating the Gut Buffet: Control of
Polysaccharide Utilization in Bacteroides spp. Trends Microbiol.
157 Senior, P. J. (1975). Regulation of nitrogen metabolism in Escherichia coli and
Klebsiella aerogenes: studies with the continuous-culture technique. J. Bacteriol.,
123(2), 407-18.
Sharkey, M. A., & Engel, P. C. (2008). Apparent negative co-operativity and substrate
inhibition in overexpressed glutamate dehydrogenase from Escherichia coli.
FEMS Microbiol. Lett., 281(2), 132-9.
Shatters, R. G., & Kahn, M. L. (1989). Glutamine synthetase II in Rhizobium:
reexamination of the proposed horizontal transfer of DNA from eukaryotes to
prokaryotes. J. Mol. Evol., 29(5), 422-8.
Shawcross, D., & Jalan, R. (2005). Dispelling myths in the treatment of hepatic
encephalopathy. Lancet, 365(9457), 431-3.
Shen, T. C., Albenberg, L., Bittinger, K., Chehoud, C., Chen, Y. Y., Judge, C. A., . . .
Wu, G. D. (2015). Engineering the gut microbiota to treat hyperammonemia. J.
Clin. Invest., 125(7), 2841-50.
Siewe, R. M., Weil, B., Burkovski, A., Eikmanns, B. J., Eikmanns, M., & Kramer, R.
(1996). Functional and genetic characterization of the (methyl)ammonium uptake
carrier of Corynebacterium glutamicum. J. Biol. Chem., 271(10), 5398-403.
Silvester, K. R., Bingham, S. A., Pollock, J. R., Cummings, J. H., & O'Neill, I. K. (1997).
Effect of meat and resistant starch on fecal excretion of apparent N-nitroso
compounds and ammonia from the human large bowel. Nutr. Cancer, 29(1), 13-
23.
Sommer, F., & Backhed, F. (2013). The gut microbiota - masters of host development
and physiology. Nat. Rev. Microbiol., 11(4), 227-38.
158 Soupene, E., He, L., Yan, D., & Kustu, S. (1998). Ammonia acquisition in enteric
bacteria: physiological role of the ammonium/methylammonium transport B
(AmtB) protein. Proc. Natl. Acad. Sci. U.S.A., 95(12), 7030-4.
Soupene, E., Lee, H., & Kustu, S. (2002). Ammonium/methylammonium transport (Amt)
proteins facilitate diffusion of NH3 bidirectionally. Proc. Natl. Acad. Sci. U.S.A.,
99(6), 3926-31.
Southern, J. A., Parker, J. R., & Woods, D. R. (1986). Expression and purification of
glutamine synthetase cloned from Bacteroides fragilis. J. Gen. Microbiol.,
132(10), 2827-35.
Stevenson, R., & Silver, S. (1977). Methylammonium uptake by Escherichia coli:
+ evidence for a bacterial NH4 transport system. Biochem. Biophys. Res.
Commun., 75(4), 1133-9.
Stevenson, R. M. (1979). Amino acid uptake systems in Bacteroides ruminicola. Can. J.
Microbiol., 25(10), 1161-8.
Straight, J. V., & Ramkrishna, D. (1994). Modeling of bacterial growth under multiply-
limiting conditions. Experiments under carbon-limiting or/and nitrogen-limiting
conditions. Biotechnol. Progress, 10(6), 588-605.
Sun, Y., & O'Riordan, M. X. (2013). Regulation of bacterial pathogenesis by intestinal
short-chain fatty acids. Adv. Appl. Microbiol., 85, 93-118.
Suzuki, A., & Knaff, D. B. (2005). Glutamate synthase: structural, mechanistic and
regulatory properties, and role in the amino acid metabolism. Photosynth. Res.,
83(2), 191-217.
159 Takahashi, M., Benno, Y., & Mitsuoka, T. (1980). Utilization of ammonia nitrogen by
intestinal bacteria isolated from pigs. Appl. Environ. Microbiol., 39(1), 30-5.
Tang, Y., Sieg, A., & Trotter, P. J. (2011). 13C-metabolic enrichment of glutamate in
glutamate dehydrogenase mutants of Saccharomyces cerevisiae. Microbiol.
Res., 166(7), 521-30.
Tate, R., Riccio, A., Merrick, M., & Patriarca, E. J. (1998). The Rhizobium etli amtB
+ gene coding for an NH4 transporter is down-regulated early during bacteroid
differentiation. Mol. Plant. Microbe Interact., 11(3), 188-98.
Team, R. D. C. (2004). R: a language and environment for statistical computing Version
2.0.1. R Foundation for Statistical Computing: Vienna, Austria.
Thorvaldsdottir, H., Robinson, J. T., & Mesirov, J. P. (2013). Integrative Genomics
Viewer (IGV): high-performance genomics data visualization and exploration.
Brief Bioinform., 14(2), 178-92.
Tománková, O., & J., K. (1995). Prediction of feed protein degradation in the rumen with
bromelain. Anim. Feed Sci. and Technol., 53, 71-80.
Tomasello, G., Tralongo, P., Damiani, P., Sinagra, E., Di Trapani, B., Zeenny, M. N., . . .
Leone, A. (2014). Dismicrobism in inflammatory bowel disease and colorectal
cancer: changes in response of colocytes. World J. Gastroenterol., 20(48),
18121-30.
Trompette, A., Gollwitzer, E. S., Yadava, K., Sichelstiel, A. K., Sprenger, N., Ngom-Bru,
C., . . . Marsland, B. J. (2014). Gut microbiota metabolism of dietary fiber
influences allergic airway disease and hematopoiesis. Nat. Med., 20(2), 159-66.
160 Truett, G. E., Heeger, P., Mynatt, R. L., Truett, A. A., Walker, J. A., & Warman, M. L.
(2000). Preparation of PCR-quality mouse genomic DNA with hot sodium
hydroxide and tris (HotSHOT). Biotechniques, 29(1), 52, 4.
Valentine, R. C., Shapiro, B. M., & Stadtman, E. R. (1968). Regulation of glutamine
synthetase. XII. Electron microscopy of the enzyme from Escherichia coli.
Biochemistry, 7(6), 2143-52. van Heeswijk, W. C., Hoving, S., Molenaar, D., Stegeman, B., Kahn, D., & Westerhoff,
H. V. (1996). An alternative PII protein in the regulation of glutamine synthetase
in Escherichia coli. Mol. Microbiol., 21(1), 133-46.
Van Heeswijk, W. C., Westerhoff, H. V., & Boogerd, F. C. (2013). Nitrogen assimilation
in Escherichia coli: putting molecular data into a systems perspective. Microbiol.
Mol. Biol. Rev., 77(4), 628-95.
van Rooyen, J., Belrhali, H., Abratt, V., & Sewell, B. T. (2011). Proteolysis of the type III
glutamine synthetase from Bacteroides fragilis causes expedient crystal-packing
rearrangements. Acta Crystallogr Sect F Struct Biol Cryst Commun, 67(Pt 3),
358-63.
van Rooyen, J. M., Abratt, V. R., Belrhali, H., & Sewell, B. T. (2010). Crystallization of
recombinant Bacteroides fragilis glutamine synthetase (GlnN) isolated using a
novel and rapid purification protocol. Protein Expr. Purif., 74(2), 211-6.
Van Rooyen, J. M., Abratt, V. R., Belrhali, H., & Sewell, T. (2011). Crystal structure of
Type III glutamine synthetase: surprising reversal of the inter-ring interface.
Structure, 19(4), 471-83.
161 van Rooyen, J. M., Abratt, V. R., & Sewell, B. T. (2006). Three-dimensional structure of
a type III glutamine synthetase by single-particle reconstruction. J. Mol. Biol.,
361(4), 796-810.
Varel, V. H., & Bryant, M. P. (1974). Nutritional features of Bacteroides fragilis subsp.
fragilis. Appl. Microbiol., 28(2), 251-7.
Viljoen, A. J., Kirsten, C. J., Baker, B., van Helden, P. D., & Wiid, I. J. (2013). The role
of glutamine oxoglutarate aminotransferase and glutamate dehydrogenase in
nitrogen metabolism in Mycobacterium bovis BCG. PLoS ONE, 8(12), e84452.
Virtanen, A. I. (1966). Milk production of cows on protein-free feed. Science, 153(3744),
1603-14.
Visek, W. J. (1978). Diet and cell growth modulation by ammonia. Am. J. Clin. Nutr.,
31(10), S216-S20.
Vogt, S. L., Pena-Diaz, J., & Finlay, B. B. (2015). Chemical communication in the gut:
Effects of microbiota-generated metabolites on gastrointestinal bacterial
pathogens. Anaerobe, 34, 106-15.
Wacker, T., Garcia-Celma, J. J., Lewe, P., & Andrade, S. L. (2014). Direct observation
+ of electrogenic NH4 transport in ammonium transport (Amt) proteins. Proc. Natl.
Acad. Sci. U S A, 111(27), 9995-10000.
Wagner, G. P., Kin, K., & Lynch, V. J. (2012). Measurement of mRNA abundance using
RNA-seq data: RPKM measure is inconsistent among samples. Theory Biosci.,
131(4), 281-5.
Wahren, A., & Gibbons, R. J. (1970). Amino acid fermentation by Bacteroides
melaninogenicus. Antonie Van Leeuwenhoek, 36(1), 149-59.
162 Wallace, R. J. (1996). Ruminal microbial metabolism of peptides and amino acids. J.
Nutr., 126(4 Suppl), 1326S-34S.
Wallace, R. J., & Cotta, M. A. (1988). Metabolism of nitrogen-containing compounds. In
The Rumen Microbial Ecosystem., 217-49.
Wallace, R. J., Onodera, R., & Cotta, M. A. (1997). Metabolism of nitrogen containing
compounds: Chapman & Hall.
Walser, M., & Bodenlos, L. J. (1959). Urea metabolism in man. J. Clin. Invest., 38,
1617-26.
Wang, L., Lai, L., Ouyang, Q., & Tang, C. (2011). Flux balance analysis of ammonia
assimilation network in E. coli predicts preferred regulation point. PLoS ONE,
6(1), e16362.
Wen, Z., & Morrison, M. (1996). The NAD(P)H-dependent glutamate dehydrogenase
activities of Prevotella ruminicola B14 can be attributed to one enzyme (GdhA),
and gdhA expression is regulated in response to the nitrogen source available for
growth. Appl Environ Microbiol, 62(10), 3826-33.
Wen, Z., & Morrison, M. (1997). Glutamate dehydrogenase activity profiles for type
strains of ruminal Prevotella spp. Appl. Environ. Microbiol., 63(8), 3314-7.
Wen, Z. T., Peng, L., & Morrison, M. (2003). The glutamine synthetase of Prevotella
bryantii B14 is a family III enzyme (GlnN) and glutamine supports growth of
mutants lacking glutamate dehydrogenase activity. FEMS Microbiol. Lett., 229(1),
15-21.
Wexler, A. G., & Goodman, A. L. (2017). An insider's perspective: Bacteroides as a
window into the microbiome. Nat. Microbiol., 2, 17026.
163 Wexler, H. M. (2007). Bacteroides: the good, the bad, and the nitty-gritty. Clin.
Microbiol. Rev., 20(4), 593-621.
Williams, K. J., Bennett, M. H., Barton, G. R., Jenkins, V. A., & Robertson, B. D. (2013).
Adenylylation of mycobacterial Glnk (PII) protein is induced by nitrogen limitation.
Tuberculosis (Edinb), 93(2), 198-206.
Winkler, F. K. (2006). Amt/MEP/Rh proteins conduct ammonia. Pflugers Arch, 451(6),
701-7.
Wolin, M. J., Miller, T. L., & Stewart, C. S. (1997). Microbe-microbe interactions. The
Rumen Microbial Ecosystem, 2nd ed., 467-91.
Wolpert, E., Phillips, S. F., & Summerskill, W. H. (1970). Ammonia production in the
human colon. Effects of cleansing, neomycin and acetohydroxamic acid. N. Engl.
J. Med., 283(4), 159-64.
Wozny, M. A., Bryant, M. P., Holdeman, L. V., & Moore, W. E. (1977). Urease assay
and urease-producing species of anaerobes in the bovine rumen and human
feces. Appl. Environ. Microbiol., 33(5), 1097-104.
Wray Jr., L. V., Zalieckas, J. M., & Fisher, S. H. (2001). Bacillus subtilis glutamine
synthetase controls gene expression through a protein-protein interaction with
transcription factor TnrA. Cell, 107(4), 427-35.
Wright, D. E., & Hungate, R. E. (1967). Amino acid concentrations in rumen fluid. Appl.
Microbiol., 15(1), 148-51.
Wrong, O., Metcalfe-Gibson, A., Morrison, R. B., Ng, S. T., & Howard, A. V. (1965). In
vivo dialysis of faeces as a method of stool analysis. I. Technique and results in
normal subjects. Clin. Sci., 28(2), 357-75.
164 Wrong, O. M., Vince, A. J., & Waterlow, J. C. (1985). The contribution of endogenous
urea to faecal ammonia in man, determined by 15N labelling of plasma urea. Clin.
Sci. (Lond), 68(2), 193-9.
Wu, M., McNulty, N. P., Rodionov, D. A., Khoroshkin, M. S., Griffin, N. W., Cheng, J., . .
. Gordon, J. I. (2015). Genetic determinants of in vivo fitness and diet
responsiveness in multiple human gut Bacteroides. Science, 350(6256),
aac5992.
Wyatt, K., White, H. E., Wang, L., Bateman, O. A., Slingsby, C., Orlova, E. V., &
Wistow, G. (2006). Lengsin is a survivor of an ancient family of class I glutamine
synthetases re-engineered by evolution for a role in the vertebrate lens.
Structure, 14(12), 1823-34.
Yamamoto, I., Abe, A., & Ishimoto, M. (1987). Properties of glutamate dehydrogenase
purified from Bacteroides fragilis. J. Biochem., 101(6), 1391-7.
Yamamoto, I., Abe, A., Saito, H., & Ishimoto, M. (1984). The pathway of ammonia
assimilation in Bacteroides fragilis. J.Gen. Appl. Microbiol., 30(6), 499-508.
Yamamoto, I., Saito, H., & Ishimoto, M. (1987). Regulation of synthesis and reversible
inactivation in vivo of dual coenzyme-specific glutamate dehydrogenase in
Bacteroides fragilis. J. Gen. Microbiol., 133(10), 2773-80.
Yamashita, M. M., Almassy, R. J., Janson, C. A., Cascio, D., & Eisenberg, D. (1989).
Refined atomic model of glutamine synthetase at 3.5 A resolution. J. Biol. Chem.,
264(30), 17681-90.
Yan, D. (2007). Protection of the glutamate pool concentration in enteric bacteria. Proc.
Natl. Acad. Sci. U.S.A., 104(22), 9475-80.
165 Yan, D., Lenz, P., & Hwa, T. (2011). Overcoming fluctuation and leakage problems in
the quantification of intracellular 2-oxoglutarate levels in Escherichia coli. Appl.
Environ. Microbiol., 77(19), 6763-71.
Yuan, J., Doucette, C. D., Fowler, W. U., Feng, X. J., Piazza, M., Rabitz, H. A., . . .
Rabinowitz, J. D. (2009). Metabolomics-driven quantitative analysis of ammonia
assimilation in E. coli. Mol. Syst. Biol., 5, 302.
Zhang, C., Wei, Z. H., & Ye, B. C. (2013). Quantitative monitoring of 2-oxoglutarate in
Escherichia coli cells by a fluorescence resonance energy transfer-based
biosensor. Appl. Microbiol. Biotechnol., 97(18), 8307-16.
166 Appendix A. Media Composition
Table 9. Brain heart infusion supplemented (BHIS) medium
Ingredient Concentration (L-1) Concentration (M) Brain heart infusion 37 g Hemin chloride 1.2 mg 1.9 µM L-Cysteine Hydrochloride 0.63 g 4 mM Sodium Carbonate 4 g 38 mM Gentamycin a (Gent) 200 mg Erythomycin a (Erm) 25 mg Tetracycline a (Tet) 2 mg 5-flouro-2-deoxyuridine a (FUdR) 200 mg a used during construction of B. thetaiotaomicron gene deletion strains.
167 Table 10. Composition of chemically defined medium (CDM) used for the culture of B. thetaiotaomicron. Ingredient Concentration (g/L) Concentration (M) D-Glucose 5 g 27.8 mM High 1.32 g 10 mM Ammonium Sulfatea Low 0.132 g 1 mM High 2.5 g Tryptonea Low 0.25 g High 2.5 g Casamino Acidsa Low 0.25 g Sodium Chloride 0.875 g 15 mM Dipotassium Hydrogen Phosphate 0.871 g 5 mM Monopotassium Hydrogen Phosphate 0.68 g 5 mM Sodium Carbonateb 4 g 38 mM Sodium Bicarbonatec 2 g 38 mM Sodium Sulfide 0.5 g 6.4 mM Magnesium (II) Chloride Heptahydrate 20 mg 0.1 mM Calcium (II) Chloride Dihydrate 8 mg 54 µM Iron (II) Sulphate Heptahydrate 0.4 mg 1.4 µM Hemin Chloride 1.2 mg 1.9 µM Vitamin K3 (Menadione) 1 mg 5.8 µM Vitamin B12 10 µg 7.3 nM a Used either ammonium sulfate or Tryptone or Casamino acids as mentioned. b Used sodium carbonate for 95%CO2 headspace in batch cultures. c Used sodium carbonate for 20%CO2 headspace in continuous cultures.
168
Appendix B. Two New Xylanases with Different Substrate Specificities from the
Human Gut Bacterium Bacteroides intestinalis DSM 17393
Pei-Ying Hong, Michael Iakiviak, Dylan Dodd, Meiling Zhang, Roderick I. Mackie, and
Isaac Cann
Manuscript published in Applied Environmental Microbiology 2014 Apr;80(7):2084-93. Contributions include protein purification, SDS-PAGE, CD spectroscopy, enzyme reactions, TLC, HPAEC.
B.1 Abstract
Xylan is an abundant plant cell wall polysaccharide and is a dominant component of dietary fiber. Bacteria in the distal human gastrointestinal tract produce xylanase enzymes to initiate the degradation of this complex heteropolymer. These xylanases typically derive from glycoside hydrolase (GH) families 10 and 11; however, analysis of the genome sequence of the xylan-degrading human gut bacterium Bacteroides intestinalis DSM 17393 revealed the presence of two putative GH8 xylanases. In the current study, we demonstrate that the two genes encode enzymes that differ in activity.
The xyn8A gene encodes an endoxylanase (Xyn8A), and rex8A encodes a reducing-
end xylose-releasing exo-oligoxylanase (Rex8A). Xyn8A hydrolyzed both xylopentaose
(X5) and xylohexaose (X6) to a mixture of xylobiose (X2) and xylotriose (X3), while
Rex8A hydrolyzed X3 through X6 to a mixture of xylose (X1) and X2. Moreover, rex8A is
located downstream of a GH3 gene (xyl3A) that was demonstrated to exhibit β-
xylosidase activity and would be able to further hydrolyze X2 to X1. Mutational analyses
of putative active site residues of both Xyn8A and Rex8A confirm their importance in
catalysis by these enzymes. Recent genome sequences of gut bacteria reveal an
169
increase in GH8 Rex enzymes, especially among the Bacteroidetes, indicating that these genes contribute to xylan utilization in the human gut.
B.2 Introduction
Xylan is an abundant plant cell wall polysaccharide that consists of a β-(1,4)- linked xylosyl backbone with various degrees of polymerization and substitution. Xylans are the major polysaccharides in cereal-derived food products, fruits, and vegetables that are consumed daily by humans. The hydrolysis and fermentation of plant cell wall polysaccharides, inclusive of xylan, are important metabolic processes that contribute approximately 10% of the caloric requirements in human hosts (1). The fermentation process yields volatile fatty acids, including butyrate, which is a compound essential for maintaining the integrity of colonic epithelial cells (2-4). Due to the heterogeneous nature of xylan, its degradation requires a combination of different enzymes, including endoxylanases, β-xylosidases, α-l-arabinofuranosidases, α-glucuronidases, ferulic acid esterases, and acetylxylan esterases (5). Humans do not possess these enzymes; therefore, degradation of xylan in the intestine is a function exclusively undertaken by commensal microorganisms (6,7).
Among the human gut microbiota, Bacteroides spp. are the most numerically dominant xylanolytic bacteria (8, 9). Past studies have identified Bacteroides eggerthii,
B. ovatus, B. fragilis, B. vulgatus, B. intestinalis, B. cellulosilyticus, and B. xylanisolvens as the major xylanolytic members of this genus (10-17). Many of the Bacteroides spp. possess xylanases, which are required to depolymerize xylans by cleaving the long polymeric chain into shorter chains of xylose units. The genome sequences of
Bacteroides ovatus ATCC 8483 and Bacteroides intestinalis DSM 17393 revealed that
170
these organisms possess the most highly expanded repertoire of glycoside hydrolase and polysaccharide lyase genes among all gut bacteria sequenced to date (18). These genes are arranged in polysaccharide utilization loci (PULs) that are specifically regulated at the transcriptional level during growth with the cognate polysaccharides
(19). In B. ovatus ATCC 8483, a relatively large number of genes are regulated at the transcriptional level during growth on xylan and these genes are highly expressed in vivo (20), indicating that they are important for xylan degradation by this bacterium.
Despite the large number of genes induced by xylan, biochemical evidence to define the substrate specificities of these enzymes is lacking. This information is particularly important in helping to define the metabolic potential of these abundant gut bacteria.
The majority of xylanases that have been studied derive from the glycoside hydrolase
(GH) families 10 and 11, with a relative minority belonging to GH families 5, 8, 30, and
43 (21, 22). Compared to xylanases in the GH10 and GH11 families, the substrate preference and hydrolysis product profiles of xylanases in GH families 5, 8, 30, and 43 have not been extensively studied. As of January 2014, the CAZy database has 729 entries in the GH8 family, with a total of 56 enzymes listed as characterized. Among these entries, six have been shown to degrade xylan, including endoxylanases (23–26) and reducing-end xylose-releasing exo-oligoxylanases (27, 28). The genome map of B. intestinalis DSM 17393 revealed the presence of two GH family 8 genes
(BACINT_04210 and BACINT_00927) (Fig. 27). BACINT_04210 is located in a polysaccharide utilization locus (PUL) consisting of 11 genes (BACINT_04220 to
BACINT_ 04210). BACINT_00927 is located downstream of a predicted GH3
171
Figure 27. Genomic context for the two Bacteroides intestinalis GH8 genes. (A) The xyn8A gene (BACINT_04210) is located within a large xylan-specific polysaccharide utilization locus (BACINT_04220 to BACINT_04210). An integrase gene (BACINT_04209) and an xynR transcriptional regulator homolog (BACINT_04208) are also located in close proximity to the xyn8A gene. (B) The rex8A gene (BACINT_00927) is not located within a polysaccharide utilization locus but is preceded immediately by xyl3A (BACINT_00926), which encodes a GH family 3 β-xylosidase.
172
glycosidase (BACINT_00926). The genomic context of these genes indicates possible roles in xylan degradation.
In this study, the protein coding sequences for BACINT_00926, BACINT_00927, and
BACINT_04210 were cloned, and the proteins were expressed in Escherichia coli and purified to near homogeneity. It is hypothesized that these three genes are involved in xylan degradation, and therefore, the activities of the three proteins against xylo- oligosaccharides and natural xylans were evaluated. The important catalytic residues in the two GH8 enzymes were also evaluated by site-directed mutagenesis. Results from this study reveal that BACINT_04210 encodes an endoxylanase (Xyn8A),
BACINT_00926 encodes a β-xylosidase (Xyl3A), and BACINT_00927 encodes a reducing-end xylose-releasing exo-oligoxylanase (Rex8A). Xyl3A cleaves xylobiose released by Rex8A, thus representing an alternative xylan-degrading pathway in gut bacteria involving GH8 and GH3 enzymes.
B.3 Materials and Methods
Materials and strains. Bacteroides intestinalis DSM 17393 (29) was obtained from the DSMZ (Braunschweig, Germany). Escherichia coli XL-10 Gold competent cells and E. coli BL-21 CodonPlus(DE3) RIL competent cells were obtained from Agilent
(Santa Clara, CA). Medium viscosity wheat arabinoxylan (WAX) and xylo- oligosaccharides were obtained from Megazyme (Bray, Ireland). All other reagents were obtained from Sigma-Aldrich or Fisher Scientific.
Gene cloning, expression, and protein purification. B. intestinalis DSM 17393 genomic DNA was extracted using the UltraClean Soil DNA isolation kit from Mo-Bio
173
(Carlsbad, CA) according to the manufacturer's protocol. The concentrations of total
DNA were quantified using the Qubit dsDNA BR assay kit (Invitrogen, Grand Island,
NY). Oligonucleotide primers used for amplifying xyl3A, rex8A, and xyn8A (Table 11)
were engineered to include 5′ and 3′ extensions for subsequent ligation-independent
cloning (LIC). Signal peptide cleavage sites were predicted at the N terminus of each
protein using SignalP v4.1 (http://www.cbs.dtu.dk/services/SignalP/) (30). Thus, to
ensure that the protein accumulates within the E. coli cells, the forward primers were
designed to amplify the genes beginning with the codon immediately downstream of the
predicted peptidase cleavage site. The coding sequences for these three genes were
amplified by PCR using the PicoMaxx high-fidelity PCR mix from Agilent, and the
resulting amplicons were purified using the Wizard DNA purification kit (Promega,
Madison, WI). The purified amplicons were digested using the exonuclease activity of
T4 DNA polymerase, annealed with a similarly digested pET-46b vector (EMD
Chemicals, Darmstadt, Germany), and introduced into E. coli XL10 Gold competent
cells by chemical transformation. Individual colonies were picked and cultured overnight
in lysogeny broth (LB) supplemented with ampicillin (100 μg/ml), and plasmid DNA was
purified using a Plasmid Minikit from Qiagen (Valencia, CA). The cloned inserts were
then sequenced to confirm the integrity of the genes (W. M. Keck Center for
Comparative and Functional Genomics at the University of Illinois).
The recombinant pET-46 EK/LIC plasmids containing the cloned genes (xyl3A,
rex8A, or xyn8A) were introduced into E. coli BL-21 CodonPlus(DE3) RIL chemically competent cells (Agilent, Santa Clara, CA) by the heat shock transformation method and cultured overnight on LB agar plates supplemented with ampicillin (100 μg/ml) and
174
Table 11. Primers used in this study Gene use and Orientation Sequence(5′–3′) Desired name mutation Cloninga xyl3A Forward GACGACGACAAGATGCAACCTCCCTACAAAAACCC (BACINT_00926) Reverse GAGGAGAAGCCCGGTTATTTAACAATGACTGGTATCGC C rex8A Forward GACGACGACAAGATGGACCCGACAAAGCCCTGGGATA AAG (BACINT_00927) Reverse GAGGAGAAGCCCGGTTACTTTTCAGGGAAGATGATGC GGTAG xyn8A Forward GACGACGACAAGATGCATCCTGTTCAGGAAGACAGTAG TGGGG (BACINT_04210) Reverse GAGGAGAAGCCCGGTTACTTGATAATCCGGAAATTGCC ACTGACATGC Mutagenesisb rex8A E90A Forward CATGATGTCCGCACCGCAGGTATGTCTTACGGA Glu90Ala rex8A D148A Forward GGCCCCGCCTCCGCCGGAGAACTTTACT Asp148Ala rex8A D286A Forward GATGCTTTCCGCTTCGCTTCTTGGCGTGTACCG Asp286Ala Forward AATCAGGATGTACGTACAGCAGGAATGTCCTATG Glu104Ala xyn8A E104A GAATG Forward AGGAGCCAAGTTGCGCTTCTGCTGGTGAAATTTA Asp164Ala xyn8A D164A TTTTATAACT Forward CAAGAAGATATCAGTTTGCTGCTCTTCGCTGTGC Asp303Ala xyn8A D303A CAT a Underlined sequences indicate the incorporated T4 exonuclease digestion sites from the pET-46b Ek-LIC cloning kit. b Underlined sequences indicate the substituted codon. Primers were designed using the Agilent QuikChange Primer Design tool
(http://www.genomics.agilent.com/primerDesignProgram.jsp).
175
chloramphenicol (50 μg/ml) at 37°C. After 12 h, a single colony was used to inoculate
fresh LB (10 ml) supplemented with the same antibiotics and cultured with vigorous
aeration at 37°C for 8 h. The culture was then diluted into 1 liter of fresh LB
supplemented with ampicillin (100 μg/ml) and chloramphenicol (50 μg/ml) in 2.8-liter
Fernbach flasks, and the cultures were incubated at 37°C with vigorous aeration by
shaking at 250 rpm. When the culture reached an optical density at 600 nm (OD600) of
0.3, isopropyl β-d-thiogalactopyranoside (IPTG) was added to a final concentration of
0.1 mM, and the culture was incubated at 16°C for an additional 16 h. The cells were
then harvested by centrifugation (4,000 × g, 15 min, 4°C). The cell pellets were
resuspended in 30 ml of ice-cold lysis buffer (50 mM Tris-HCl, 300 mM NaCl, pH 7.5)
and ruptured by two passages through an EmulsiFlex C-3 cell homogenizer from
Avestin (Ottawa, Canada). The cell lysates were clarified by centrifugation at 12,000 × g for 30 min at 4°C. The recombinant proteins were then purified from the clarified lysates using Talon metal affinity resin (ClonTech, Mountain View, CA) according to the supplier's protocol with slight modifications to the binding (50 mM Tris-HCl, 300 mM
NaCl, pH 7.5) and elution (50 mM Tris-HCl, 300 mM NaCl, 250 mM imidazole, pH 7.5) buffers. Aliquots of eluted fractions were analyzed by sodium dodecyl sulfate- polyacrylamide gel electrophoresis (SDS-PAGE) according to Laemmli's method (31), and protein bands were visualized by staining with Coomassie brilliant blue G-250. The protein concentrations were calculated by absorbance spectroscopy at 280 nm using a
NanoDrop 1000 instrument (Thermo Scientific, Waltham, MA) with extinction coefficients of 120,670 M−1 cm−1, 117,940 M−1 cm−1, and 103,625 M−1 cm−1 for Xyl3A,
Rex8A, and Xyn8A, respectively.
176
Size exclusion chromatography. The quaternary structures of Xyn8A and
Rex8A were analyzed by size exclusion chromatography using a Superdex 200 10/300
GL size exclusion column affixed to an AKTAxpress FPLC unit. One hundred microliters
of Xyn8A (1 mg/ml), Rex8A (1 mg/ml), or a gel filtration standard mixture (Bio-Rad,
Hercules, CA) was loaded onto the column preequilibrated with a buffer composed of 50
mM sodium phosphate-150 mM NaCl. The pH of the buffer was adjusted to 6.0 for
Rex8A and 6.5 for Xyn8A. The proteins were eluted in the same buffer at a flow rate of
0.5 ml/min. A calibration curve of molecular mass versus retention time was constructed
with the gel filtration standards, and the apparent molecular masses of the two proteins
were calculated by comparison of experimental retention times with calibration
standards.
CD spectroscopy. Circular dichroism (CD) spectra were obtained using a J-815
spectropolarimeter from Jasco (Easton, MD). The enzymes were exchanged into CD
buffer (50 mM sodium phosphate, pH 6.0 for Rex8A or pH 6.5 for Xyn8A) using a
HiPrep 26/10 desalting column affixed to an AKTAxpress FPLC unit. The proteins were
then diluted in CD buffer to a final concentration of 1 μM, and the far-UV CD spectra
were recorded from 260 to 190 nm with a wavelength step of 0.1 nm. The secondary-
structure contents of Xyn8A and Rex8A were calculated using the DichroWeb online
circular dichroism analysis server with the reference set 4 optimized for 190 to 240 nm
(32).
Evaluation of hydrolysis of xylo-oligosaccharides. The capacities of Xyn8A,
Rex8A, and Xyl3A to hydrolyze xylo-oligosaccharides were tested with xylose (X1), xylobiose (X2), xylotriose (X3), xylotetraose (X4), xylopentaose (X5), and xylohexaose
177
(X6). Substrates (5 mM) were diluted in phosphate reaction buffer (50 mM sodium
phosphate, 150 mM NaCl, pH 6.0 for Rex8A or pH 6.5 for Xyn8A and Xyl3A); reactions
were initiated by the addition of enzyme (0.5 μM final concentration), and the mixtures
were incubated at 37°C for 16 h. Preliminary experiments revealed pH optima of 6.0 for
Rex8A and 6.5 for Xyn8A and Xyl3A (Fig. 28). Reaction mixtures were terminated by
heating at 99°C for 10 min, and 1-μl aliquots of samples were spotted onto silica gel
thin-layer chromatography (TLC) plates (60 , 250-μm thickness) (Whatman,
Piscataway, NJ). The products of hydrolysis Åwere then resolved by one ascent with a mobile phase composed of n-butanol, acetic acid, and water (10:5:1). The plates were then dried, and products were visualized by spraying with a mixture of sulfuric acid
(10%, vol/vol), orcinol (0.1%, wt/vol), and methanol (50%, vol/vol) and developed by heating at 80°C for 15 min.
To evaluate the hydrolysis patterns over time, enzyme concentrations that gave near-complete hydrolysis over the 3-h reaction period for each enzyme-substrate combination were first determined. For Rex8A, the enzyme concentration used was 50 nM (X3 through X6), whereas for Xyn8A the concentration was 500 nM (X4), 20 nM (X5),
or 2 nM (X6). Reaction mixtures were prepared with substrate (1 mM) and enzyme, incubated at 37°C, and were then taken at 10-, 30-, 60-, 120-, 180-, and 360-min intervals for heat inactivation. The products of hydrolysis were first diluted 50-fold and then analyzed using high-performance anion exchange chromatography (HPAEC) with a System Gold high-performance liquid chromatography (HPLC) instrument (Beckman
Coulter, Fullerton, CA). The instrument was fitted with a CarboPac PA1 guard column (4 by 50 mm) and a CarboPac PA1 analytical column (4 by 250 mm) from Dionex
178
Figure 28. Optimal parameters for B. intestinalis Xyn8A, Rex8A, and Xyl3A activity. (A) To determine the optimal pH (i) for Xyn8A, xylopentaose (2 mM) was incubated with Xyn8A (25 μM) at 37°C for 5 minutes at various pHs, ranging from 4.5 to 8.0. To determine the optimal temperature (ii) for Xyn8A, xylopentaose (2 mM) was incubated with Xyn8A (25 μM) at various temperatures ranging from 25-65°C for 5 minutes in phosphate buffer (pH6.5). (B) To determine the optimal pH (i) for Rex8A, xylotriose (2 mM) was incubated with Rex8A (25 μM) at 37°C for 5 minutes at various pHs, ranging from 4.5 to 8.0. To determine the optimal temperature (ii) for Rex8A, xylotriose (2 mM) was incubated with Rex8A (25 μM) at various temperatures ranging from 25-65°C for 5 minutes in phosphate buffer (pH6.0). Hydrolysis products were separated using high performance anion exchange chromatography (HPAEC) and detected using a pulsed amperometric detector (PAD). The hydrolysis products were quantified by comparing the peak area and retention time to those of commercially available standards. (C) To determine the optimal pH (i) for Xyl3A, para-nitrophenol xylofuranoside (2 mM) was incubated with Xyl3A (250 μM) at 37°C for 10 minutes at various pHs, ranging from 4.0 to 8.5. Release of para-nitrophenol was monitored at 400nm every 31seconds using a Synergy II microplate reader. To determine the optimal temperature for Xyl3A, para-nitrophenol xylofuranoside (1 mM) was incubated with Xyl3A (250 μM) at various temperatures ranging from 25-65°C for 5 minutes in phosphate buffer (pH6.5). Release of para-nitrophenol was monitored at 400nm continuously using a Cary UV-Vis spectrophotometer.
179
Corporation (Sunnyvale, CA) and with a Coulochem III electrochemical detector (ESA
Biosciences, Chelmsford, MA). Peak retention times and peak areas from sample chromatograms were compared to those obtained using commercially available xylo- oligosaccharides analyzed as standards.
Evaluation of hydrolysis of natural xylans. The capacity of the enzymes to hydrolyze natural xylans (i.e., wheat arabinoxylan [WAX] and oat spelt xylan [OSX]) was assessed by dissolving WAX or OSX (2.5 mg/ml) in phosphate buffer (50 mM, 150 mM
NaCl,; pH 6.0 for Rex8A or pH 6.5 for Xyn8A; final volume, 100 μl). Reactions were initiated by the addition of enzyme (final concentration, 0.5 μM), and reaction mixtures were incubated at 37°C for 16 h. The reaction mixtures were then heat inactivated at
99°C for 10 min and centrifuged for 10 min at 15,000 × g, and hydrolysis products in the supernatants were evaluated by TLC as described above. The concentrations of reducing sugars were quantified by the PAHBAH (para-hydroxybenzoic acid hydrazide) method as previously described (33).
Site-directed mutagenesis. Mutagenesis was performed using the QuikChange
Lightning Multi Site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA).
Mutagenic primers were designed with the QuikChange Primer Design tool (Agilent
Technologies) (Table 11). Mutations were prepared according to the QuikChange protocol with either the rex8A or the xyn8A pET-46b plasmid as the DNA template. The residual parent plasmid was then digested by incubation with DpnI overnight at 37°C, and the resulting DNA was transformed into E. coli XL10-Gold ultracompetent cells by the heat shock method. Colonies were picked and cultured, and plasmids were purified using a Plasmid Minikit (Qiagen, Valencia, CA). The purified plasmid DNA was then
180
sequenced to ensure that the appropriate mutations were introduced and that the rest of
the gene sequences remained unchanged. Expression and purification of the mutant
recombinant proteins were performed as described above for the wild-type (WT)
proteins.
Hydrolysis of pNP-linked sugars. The enzyme-catalyzed hydrolysis of para- nitrophenyl (pNP)-linked monosaccharide substrates was assayed by using a thermostated Synergy II multimode microplate reader (BioTek Instruments Inc.,
Winooski, VT). A library of pNP substrates was screened for activity as described previously (34). The substrates (1 mM) in 100 μl phosphate buffer (50 mM sodium phosphate, 150 mM NaCl, pH 6.5) were incubated at 37°C in the presence or absence of Xyl3A (0.1 μM for pNPXyl; 0.5 μM for pNPAra; 1 μM for pNPGal and pNPGlu) for 30 min, and the amount of pNP release was determined by continuously monitoring the absorbance at 400 nm. The path length correction feature of the instrument was employed to convert the absorbance values recorded to correspond to those for a 1-cm path length. The extinction coefficient for pNP at pH 6.5 and at a wavelength of 400 nm was measured to be 3.179 mM−1 cm−1.
B.4 Results
Identification of two xylan-specific GH8 genes in Bacteroides intestinalis.
Genomic analysis of Bacteroides intestinalis DSM 17393 revealed the presence of two
GH8 genes, BACINT_04210 and BACINT_00927, which encode the proteins Xyn8A
and Rex8A, respectively. The genomic context for these genes supports their predicted
roles in xylan degradation. To illustrate, BACINT_04210 is located in a polysaccharide
181
utilization locus (PUL) consisting of 11 genes (BACINT_04220 to BACINT_04210).
Included in this PUL is a large gene cluster consisting of two tandem repeats of susC and susD orthologs (xusC and xusD) followed by a hypothetical gene (BACINT_04216) and a GH10 endoxylanase (BACINT_04215) (Fig. 27A). Downstream are a hypothetical protein (BACINT_04214), a GH5 endoxylanase (xyn5A, BACINT_04213) (17), a family
1 carbohydrate esterase (BACINT_04212), a two-domain CE6/GH95 gene
(BACINT_04211), and the GH8 gene BACINT_04210. Following this cluster is a predicted integrase, an observation that suggests that this locus may be part of an integrative and conjugative element. Divergently transcribed relative to the gene cluster described above is a homolog of xynR, a hybrid two-component system regulator that regulates xylanase gene expression in Prevotella bryantii B14 (35). The second GH8 gene (BACINT_00927) is located directly downstream of a predicted GH3 glycosidase
(Fig. 27B).
Both of these proteins, Xyn8A and Rex8A, possess putative signal peptides with predicted signal peptidase II cleavage sites, followed by GH8 domains. Amino acid sequence alignments of the GH8 domains demonstrated 39% amino acid sequence identity (Fig. 29).
Cloning, expression, and purification of Xyn8A and Rex8A. To determine whether or not the two GH8 genes encode proteins with redundant biochemical properties, they were individually cloned into an E. coli expression vector. The two proteins were then expressed as soluble recombinant hexahistidine fusion proteins for subsequent purification via cobalt immobilized metal affinity chromatography (IMAC). In
182
Figure 29. Amino acid sequence alignment for Rex8A (BACINT_00927) and Xyn8A (BACINT_04210). The amino acid sequences for the two proteins were aligned using ClustalW with the BLOSUM 62 similarity matrix and visualized using ESPript (http://espript.ibcp.fr/ESPript/ESPript/). Black boxes indicate amino acid identity and white boxes indicate amino acid similarity. Amino acids predicted to be important for catalytic activity and subsequently targeted for site-specific mutagenesis are indicated with an asterisk. Arrows indicate predicted signal peptide cleavage sites.
183
a single IMAC step, the two recombinant proteins were purified to near homogeneity as
assessed by SDS-PAGE (Fig. 30A).
Recombinant Xyn8A and Rex8A have similar secondary structures. To
compare the secondary structural compositions of the two recombinant proteins, CD
spectra in the far-UV region (190 to 260 nm) were collected (Fig. 30B). The raw data
were then uploaded onto the DichroWeb server, and secondary structural elements
were predicted by comparison with a standardized data set. The recombinant proteins
both yielded CD spectra indicating the presence of α-helix, β-sheet, and β-turn secondary structural elements. Comparison of the DichroWeb results using the Student t test revealed no significant differences in α-helices, β-sheets, β-turns, or unordered regions between the two proteins (Table 12). These data suggest that when expressed as recombinant proteins, Xyn8A and Rex8A have similar overall secondary structural properties.
Xyn8A and Rex8A are both monomers in solution. To evaluate whether the two proteins exhibited similar quaternary structures and to rule out whether the two proteins exist as aggregates, their apparent molecular weights were determined by gel filtration analysis. Both proteins eluted from the column well after the size exclusion limit, indicating that neither of the two proteins existed as large aggregates in solution.
The elution volumes of the two proteins revealed apparent molecular weights slightly lower than those demonstrated by SDS-PAGE. This discrepancy suggests that the proteins may have different shapes from those of the proteins in the calibration standards. Nevertheless, these results indicate that both proteins exist as monomers in solution (Fig. 30C).
184
Figure 30. The two Bacteroides intestinalis GH8 genes encode xylan-degrading enzymes with distinct properties. (A) Purification of the two GH8 proteins, Xyn8A and Rex8A. The proteins were produced in their recombinant forms in E. coli, purified by immobilized metal affinity chromatography, resolved by 12% SDS-PAGE, and stained with Coomassie brilliant blue G250. (B) Predicted secondary structure of the two GH8 proteins. Circular dichroism (CD) spectra were obtained for Rex8A and Xyn8A in the far-UV region (190 to 260 nm). (C) Gel filtration chromatography. The sizes of purified Rex8A and Xyn8A were determined by size exclusion fast-protein liquid chromatography (FPLC). The molecular masses of Rex8A and Xyn8A were calculated from the retention time of the peak absorbance by comparison with calibration standards having known molecular masses. (D) Hydrolysis of model xylans. Rex8A and Xyn8A were incubated with wheat arabinoxylan (WAX) or oat spelt xylan (OSX) for 16 h, and the products of hydrolysis were analyzed by thin-layer chromatography (TLC). (E and F) Hydrolysis of xylo-oligosaccharides (XOS). Rex8A (E) or Xyn8A (F) was incubated with XOS of different lengths, and the products were analyzed by TLC.
185
Table 12. Analysis of CD spectra using DICHROWEBa
Protein (ORF number) % α-helixb % β-sheetb % β-turnb % Unordered Xyn8A 71 ± 6 9 ± 3 9 ± 4 10 ± 3 (BACINT_04210) Rex8A 81 ± 2 6 ± 1 6 ± 2 6 ± 2 (BACINT_00927)
a CD spectra were recorded in the far-UV range utilizing a J-815 CD spectropolarimeter. The buffer composition was 50 mM sodium phosphate, pH 5.5. The spectra were recorded from 190 to 260 nm at a scan rate of 50 nm/s and a 0.1-nm wavelength step with five accumulations. Data are from three independent experiments and are presented as means ± standard deviations of the means. The spectra were uploaded onto the DICHROWEB online server and analyzed as described in Materials and Methods. b Student’s t-test, P < 0.01
186
Xyn8A and Rex8A exhibit different biochemical properties. The two proteins
were incubated with either wheat arabinoxylan (WAX), oat spelts xylan (OSX), or xylo-
oligosaccharides (XOS), and their capacity to degrade the substrates was assessed by
analyzing the hydrolysates by TLC. In the absence of enzyme, no oligosaccharides
were present in either the WAX or the OSX mixture (Fig. 30D). However, upon the
addition of Xyn8A, several spots appeared, corresponding to shorter xylo-
oligosaccharides being released, an activity characteristic of endoxylanase enzymes
(Fig. 30D). The spots did not clearly comigrate with the standard XOS, which is likely explained by the fact that these products are xylo-oligosaccharides substituted with
arabinosyl side chains. In contrast, no detectable degradation of the two
polysaccharides was noted for Rex8A.
Both enzymes degraded XOS (Fig. 30E and Fig. 30F); however, Xyn8A exhibited
no detectable activity toward X2 and X3. Upon incubation with xylotetraose, a mixture of
X3, X2, and X1 was observed, and a mixture of X3 and X2 was seen following incubation
with xylopentaose and xylohexaose. In contrast, Rex8A completely converted X3
through X6 to a mixture of X2 and X1. Most notably, the two enzymes displayed
differences in the product distribution profiles for X5, with Xyn8A cleaving X5 into X2 and
X3, whereas Rex8A converted X5 to X1 and X2. This result suggests that the two
enzymes have different cleavage site preferences, with Xyn8A cleaving the middle
glycosidic linkage within X5, and Rex8A cleaving one of the terminal glycosidic linkages
in X5, producing X4 and X1 and then converting X4 to X3 and X1 and so on.
To further evaluate the hydrolytic activity of Rex8A with XOS, the enzyme was incubated with X3 through X6, and the concentrations of X1 through X6 were followed
187
over time by HPLC. These experiments confirmed results from the TLC plates, which
showed that the predominant products at the final time point of the reaction were X1 and
X2 for all substrates tested (Fig. 31). Rex8A converted all of the X3 to a mixture of X1
and X2 by 180 min of incubation (Fig. 31A). Rex8A hydrolyzed all of the X4 substrate after 180 min of incubation by first accumulating X1 and X3 and further hydrolyzing the
accumulated X3 to X2 and X1 (Fig. 31B). With X5 as a substrate, there was an initial
accumulation of X4 and X1 and a subsequent hydrolysis of X4 to accumulate X3 (Fig.
31C). Similarly, in the hydrolysis of X6, X5 first accumulated during 20 min of incubation
and was subsequently hydrolyzed to X4 and X3 and finally to X1 and X2 by the end of the
180-min incubation (Fig. 31D). At a relatively high concentration of Xyn8A (500 nM), X4 was converted first to X3 and X1, and then the X3 was eventually hydrolyzed to X2 and
X1 (Fig. 32A). The enzyme exhibited much higher activity with longer oligosaccharides; therefore, the amount of enzyme added was decreased to observe the initial hydrolysis products (X5, 20 nM; X6, 2 nM). Under these conditions, X5 was converted to
stoichiometric amounts of X3 and X2, whereas X6 was cleaved, accumulating mainly X3
(Fig. 32B and C).
Taken together, the activities seen for Xyn8A with polysaccharides and XOS indicate that this enzyme is an endoxylanase, whereas the activity for Rex8A was clearly very different from that of Xyn8A and is consistent with the activity seen for
reducing-end xylose-releasing exo-oligoxylanases, which have been demonstrated for
GH8 proteins (27, 28).
Site-specific mutagenesis to elucidate catalytic residues. To predict amino
acid residues important for catalysis, three-dimensional homology models were
188
Figure 31. Hydrolysis of XOS by Rex8A (BACINT_00927). Rex8A was incubated with xylotriose (A), xylotetraose (B), xylopentaose (C), and xylohexaose (D), and products were analyzed at designated intervals using high-performance anion exchange chromatography (HPAEC). Abbreviations are as follows: X6, xylohexaose; X5, xylopentaose; X4, xylotetraose; X3, xylotriose; X2, xylobiose; and X1, xylose). Concentrations were determined by comparison to calibration curves constructed with known concentrations of sugars. Results are presented as means and standard deviations from three individual experiments.
189
Figure 32. Hydrolysis of XOS by Xyn8A (BACINT_04210). Xyn8A was incubated with xylotetraose (A), xylopentaose (B), and xylohexaose (C), and products were analyzed at designated intervals using high-performance anion exchange chromatography (HPAEC). Abbreviations are as follows: X6, xylohexaose; X5, xylopentaose; X4, xylotetraose; X3, xylotriose; X2, xylobiose; and X1, xylose. Concentrations were determined by comparison to calibration curves constructed with known concentrations of sugars. Results are presented as means and standard deviations from three individual experiments.
190
constructed for Rex8A and Xyn8A with the reducing-end xylose-releasing exo-
oligoxylanase (Rex) from Bacillus halodurans (PDB accession number 1WU4) and the
cold-adapted endoxylanase from Pseudoalteromonas haloplanktis Xyn8A (PDB
accession number 1H12), respectively. These structures revealed that the three
catalytic residues proposed previously for REX (27, 36) (Glu90, Asp148, Asp286) were
conserved in both Rex8A and Xyn8A (Fig. 33A). Therefore, to evaluate whether these
residues (Glu104, Asp164, and Asp303 for Xyn8A and Glu90, Asp148, and Asp286 for
Rex8A) were also important for Xyn8A and Rex8A, these three residues were changed
to alanine by site-directed mutagenesis.
The three mutants for each protein were expressed in E. coli and purified by
cobalt IMAC as described for the wild-type protein. All six proteins were expressed in
the soluble fraction, indicating that the mutations did not cause the proteins to form
insoluble aggregates. Similar to the wild-type protein, the mutants were purified after a
single chromatography step (Fig. 33B). The proportions of secondary structural
elements between the WT and the mutants were determined and compared by circular-
dichroism (CD) spectroscopy to ensure that the change in enzyme activity was not
caused by gross structural differences. Comparison of the α-helix, β-sheet, β-turn, and unordered regions using the Student t test revealed no significant differences between the wild-type and mutant forms of Rex8A and Xyn8A (Table 12). These data demonstrate that mutation of these active-site residues did not appreciably alter the secondary structures of the proteins.
Following 16 h of incubation at 37°C with X6, only the D164A mutant of Xyn8A
exhibited minor residual activity compared to the wild type, with a trace amount of X3
191
Figure 33. Mutational analysis reveals residues important for catalysis in Xyn8A (BACINT_04210) and Rex8A (BACINT_00927). The residues are Glu104, Asp164, and Asp303 for Xyn8A. The residues are Glu90, Asp148, and Asp286 for Rex8A. (A) Three-dimensional homology modeling. Homology models were built for Xyn8A and Rex8A using the ModWeb server (http://modbase.compbio.ucsf.edu/ModWeb20- html/modweb.html) with Pseudoalteromonas haloplanktis Xyn8A (PDB accession no. 1H12) and Bacillus halodurans Rex (PDB accession no. 1WU4) as the templates, respectively. (B) Purification of mutants. Three mutants for each Xyn8A and Rex8A were expressed as recombinant proteins, purified by cobalt affinity chromatography, and analyzed by SDS-PAGE. (C) Hydrolysis of xylohexaose. Wild-type and mutant proteins were incubated with xylohexaose (X6) for 16 h, and the reaction products were analyzed by thin-layer chromatography.
192
visualized on the TLC plate (Fig. 33C). For the three Rex8A mutants, no activity was detected (Fig. 33C). These results therefore show that these three carboxylate- containing residues are important for catalysis in both Rex8A and Xyn8A.
Identification of Xyl3A as a GH3 β-xylosidase. Rex8A was observed to hydrolyze X3 through X6 to a mixture of xylose and xylobiose (Fig. 30F). Given that rex8A is located immediately downstream of a predicted GH3 glycosidase gene, it is possible that the xylobiose produced by Rex8A would be converted to xylose by this
GH3 enzyme. Because GH3 glycosidases have a very broad substrate range, it was decided to clone this GH3 glycosidase gene and study its biochemical properties. The protein was produced recombinantly in E. coli as described for Xyn8A and Rex8A and purified to near homogeneity by cobalt IMAC (Fig. 34A). Biochemical activity assays with a library of para-nitrophenyl (pNP)-linked substrates revealed that the enzyme had highest activity with pNP-β-d-xylopyranoside as a substrate (Fig. 34B). These results suggest that the enzyme is a β-xylosidase. To further confirm this with natural substrates, the enzyme was incubated with XOS and the capacity to hydrolyze these substrates was assessed by TLC. Following incubation of the purified protein with XOS ranging in length from X2 through X6, the sole product visualized was xylose (Fig. 34C).
These results indicate that BACINT_00926 encodes a GH3 β-xylosidase, and the associated enzyme was therefore named Xyl3A.
B.5 Discussion
A recent analysis of human gut microbiome reference genomes revealed that of all gut bacterial strains, Bacteroides intestinalis DSM 17393 exhibited the highest
193
Figure 34. BACINT_00926 encodes a GH3 β-xylosidase, Xyl3A. (A) Expression and purification of Xyl3A. Xyl3A was expressed as a recombinant protein, purified by cobalt affinity chromatography, and analyzed by SDS-PAGE. (B) Activity with artificial substrates. Purified Xyl3A was screened with a library of pNP-linked glycans using a continuous spectrophotometric assay, and the specific activities for a subset of substrates are shown. Abbreviations are as follows: pNPAra, pNP-α-l- arabinofuranoside; pNPGal, pNP-β-d-galactopyranoside; pNPXyl, pNP-β-d- xylopyranoside; pNPGlu, pNP-β-d-glucopyranoside. (C) Hydrolysis of XOS. Xyl3A was incubated with XOS ranging in length from xylobiose (X2) to xylohexaose (X6) for 16 h, and products of hydrolysis were analyzed by thin-layer chromatography. Specific activity (mIU/mg) is displayed in units of nmol substrate consumed per minute per mg of protein.
194
representation of individual glycoside hydrolase (GH) and polysaccharide lyase (PL)
families and the second highest total number of GH and PL genes (18). The number of
genes in the B. intestinalis genome targeted toward degradation of the plant cell
structural polysaccharide xylan is particularly high. Coupled with previous studies of this
organism (17, 37), the current data demonstrate that B. intestinalis DSM 17393 encodes at least nine endoxylanase enzymes deriving from three different GH families (i.e.,
GH10, GH5, and GH8).
Expansion of GH families is a characteristic of many gut bacteria, most notably those from the Bacteroidetes phylum (18, 38, 39). The current study highlights this
process by revealing that B. intestinalis DSM 17393 harbors two copies of genes
encoding xylan-specific GH8 enzymes with completely distinct biochemical properties.
Xyn8A is a GH8 endoxylanase enzyme that targets longer xylan fragments, hydrolyzing
them to shorter xylo-oligosaccharides that can subsequently be degraded by side-chain-
cleaving enzymes and β-xylosidases. Rex8A, on the other hand, is likely a reducing-end
xylose-releasing exo-oligoxylanase that releases xylose from the reducing end of xylo-
oligosaccharides, liberating fermentable monosaccharides. The discovery of these two
enzymes broadens our understanding of xylan degradation by gut bacteria and provides
insight into the highly dynamic genomic assemblages that gut bacteria possess to
capture energy from dietary polysaccharides.
Rex8A forms the core of a group of GH8 enzymes that derive largely from
bacteria isolated from human gastrointestinal tract and rumen sources (Fig. 35A).
Despite sharing high amino acid sequence similarity, the polypeptide domain
architectures vary remarkably among these different organisms. Another human gut
195
Figure 35. Expansion and diversity of Rex8A homologs in the gut microbiome. (A) Phylogenetic analysis of Xyn8A (BACINT_04210) and Rex8A (BACINT_00927). The amino acid sequences for GH8 proteins annotated by the CAZy database were aligned with Xyn8A and Rex8A using ClustalW, and a neighbor-joining tree was constructed using CLC Genomics Workbench v5.0 software. Each alignment was resampled 100 times, and the bootstrap values are indicated on the internal branches. The branch length is reported as the expected number of substitutions per amino acid position. GIT, gastrointestinal tract. (B) Protein domain structure of Rex8A homologs. Domain architectures were predicted using dbCAN (48). Signal peptides and lipoprotein signal sequences were predicted using SignalP v4.1 (30) and LipoP v1.0 (43), respectively. Organism abbreviations: Bcel, Bacteroides cellulosilyticus DSM 14838; Bi, Bacteroides intestinalis; Bole, Bacteroides oleiciplenus YIT 12058; Begg, Bacteroides eggerthii DSM 20697; Buni, Bacteroides uniformis CL03T12C37; Pber, Prevotella bergensis DSM 17361; Pcop, Prevotella copri DSM 18205; Dmos, Dysgonomonas mossii DSM 22836; Prum, Prevotella ruminicola 23; Dgad, Dysgonomonas gadei ATCC BAA-286; Mpal, Mucilaginibacter paludis DSM 18603; Slin, Spirosoma linguale DSM 74; Fjoh, Flavobacterium johnsoniae UW101; Bhal, Bacillus halodurans C-125; Bsp, Bacillus sp. 10403023; Pgol, Parabacteroides goldsteinii CL02T12C30; Fsuc, Fibrobacter succinogenes subsp. succinogenes S85; Chut, Cytophaga hutchinsonii ATCC 33406; Ccel, Clostridium cellulolyticum H10; Acel, Acetivibrio cellulolyticus; Athe, Anaerolinea thermophila UNI-1; Phal, Pseudoalteromonas haloplanktis; Cthe, Clostridium thermocellum ATCC 27405.
196
bacterium, Bacteroides eggerthii, possesses a carbohydrate esterase (CE) family 15 domain at the N terminus, while Prevotella copri, Prevotella bergensis, and Bacteroides uniformis possess N-terminal GH43/CBM6 modules (Fig. 35B). These proteins are likely to be bifunctional enzymes with two distinct active sites. CE family 15 proteins have been demonstrated to cleave the methyl ester linkage in 4-O-methyl-glucuronyl methyl esters (40, 41), an activity that is important for fermenting the glucuronic acid component of heteroxylans. GH43/CBM6 proteins are commonly found to have either β- xylosidase or arabinofuranosidase activities, both of which are associated with the degradation of xylan fragments. Therefore, it is likely that these alternative forms of GH8
Rex enzymes have evolved to possess additional enzymatic activities that are important for xylan degradation.
The rex8A gene is located immediately downstream of the xyl3A gene, which encodes a GH3 β-xylosidase. This unique pairing of two genes encoding enzymes with activity against xylo-oligosaccharides is deserving of further discussion. Rex enzymes possess an active site with a minimum binding requirement for three xylose sugars that is closed off at the terminal reducing end. Catalysis then occurs between the sugars occupying the −1 and +1 subsites, liberating xylose and the remaining oligosaccharide chain (36). Importantly, the minimum requirement of binding 3 sugars for catalysis means that the shortest XOS that this enzyme can cleave is xylotriose, and therefore xylobiose is not cleaved by these enzymes. On the other hand, GH3 glycosidases have a short coin-shaped active site that contains two subsites (42); therefore, enzymes from this family efficiently cleave xylobiose (34). These two enzymes have complementary
197
activities that may be important for the mechanism of xylan degradation employed by this organism.
The three proteins described in this study each possess an N-terminal signal peptide with a signal peptidase cleavage II site, suggesting that they are transferred across the inner membrane and anchored to a lipid moiety (43). The so-called “+2” rule implies that prolipoproteins containing an aspartic acid at the +2 position relative to the signal peptidase II cleavage site results in retention of the polypeptide on the inner membrane within the cytoplasm, whereas a serine residue directs the protein to the outer membrane (44, 45). Xyn8A, Rex8A, and Xyl3A all possess serine residues at the
+2 position, which suggests that these proteins are all localized to the outer membrane.
Since GH8 Rex enzymes represent a recently discovered group of enzymes, relatively few studies have evaluated the catalytic residues for these enzymes. In the current study, we identified amino acid residues (E90, D148, and D286) that were absolutely conserved among the GH8 Rex and Xyn enzymes and made the corresponding mutations in Xyn8A and Rex8A. These mutations had very large effects on the activities of the two enzymes, despite CD spectroscopy demonstrating no significant change in secondary structure (Table 12). These data confirm results from previous studies that show residues corresponding to E90, D148, and D286 as having roles as catalytic acid, pKa modulator, and catalytic base, respectively (46).
In summary, this study reports distinct activities for two GH8 enzymes that are present in B. intestinalis DSM 17393, a bacterium that is endemic to the human gut. Our findings reiterate the previous observation that this enzyme family contains members with considerable differences in their substrate specificity (47). Furthermore, our results
198
suggest a xylan degradation pathway active in gut Bacteroidetes that involves
endoxylanases, a reducing-end oligo-xylanase, and a β-xylosidase. These results could
be of importance in understanding the pathways of xylan degradation present in other
gut microorganisms harboring GH8 enzymes.
B.6 References
1. Bergman E. 1990. Energy contributions of volatile fatty acids from the gastrointestinal
tract in various species. Physiol. Rev. 70:567–590
2. Hague A, Singh B, Paraskeva C. 1997. Butyrate acts as a survival factor for colonic
epithelial cells: further fuel for the in vivo versus in vitro debate. Gastroenterology
112:1036–1040. 10.1053/gast.1997.v112.agast971036
3. Roediger WE. 1980. The colonic epithelium in ulcerative colitis: an energy-deficiency
disease? Lancet ii:712–715
4. Scheppach W, Bartram HP, Richter F. 1995. Role of short-chain fatty acids in the
prevention of colorectal cancer. Eur. J. Cancer 31A:1077–1080
5. Dodd D, Cann IK. 2009. Enzymatic deconstruction of xylan for biofuel production.
Glob. Change Biol. Bioenergy 1:2–17. 10.1111/j.1757-1707.2009.01004.x
6. Davison A, Blaxter M. 2005. Ancient origin of glycosyl hydrolase family 9 cellulase
genes. Mol. Biol. Evol. 22:1273–1284. 10.1093/molbev/msi107
7. Morris SC. 2003. Life's solution: inevitable humans in a lonely universe. Cambridge
University Press, Cambridge, United Kingdom
199
8. Hespell RB, Whitehead TR. 1990. Physiology and genetics of xylan degradation by
gastrointestinal tract bacteria. J. Dairy Sci. 73:3013–3022. 10.3168/jds.S0022-
0302(90)78988-6
9. Chassard C, Goumy V, Leclerc M, Del'homme C, Bernalier-Donadille A. 2007.
Characterization of the xylan-degrading microbial community from human faeces.
FEMS Microbiol. Ecol. 61:121–131. 10.1111/j.1574-6941.2007.00314.x
10. Chassard C, Delmas E, Lawson PA, Bernalier-Donadille A. 2008. Bacteroides
xylanisolvens sp. nov., a xylan-degrading bacterium isolated from human faeces.
Int. J. Syst. Evol. Microbiol. 58:1008–1013. 10.1099/ijs.0.65504-0
11. Salyers AA, Vercellotti JR, West SE, Wilkins TD. 1977. Fermentation of mucin and
plant polysaccharides by strains of Bacteroides from the human colon. Appl.
Environ. Microbiol. 33:319–322
12. Robert C, Chassard C, Lawson PA, Bernalier-Donadille A. 2007. Bacteroides
cellulosilyticus sp. nov., a cellulolytic bacterium from the human gut microbial
community. Int. J. Syst. Evol. Microbiol. 57:1516–1520. 10.1099/ijs.0.64998-0
13. Crittenden R, Karppinen S, Ojanen S, Tenkanen M, Fagerstrom R, Matto J, Saarela
M, Mattila-Sandholm T, Poutanen K. 2002. In vitro fermentation of cereal dietary
fibre carbohydrates by probiotic and intestinal bacteria. J. Sci. Food Agric.
82:781–789. 10.1002/jsfa.1095
14. Weaver J, Whitehead TR, Cotta MA, Valentine PC, Salyers AA. 1992. Genetic
analysis of a locus on the Bacteroides ovatus chromosome which contains xylan
utilization genes. Appl. Environ. Microbiol. 58:2764–2770
200
15. Whitehead TR. 1995. Nucleotide sequences of xylan-inducible xylanase and
xylosidase/arabinosidase genes from Bacteroides ovatus V975. Biochim.
Biophys. Acta 1244:239–241. 10.1016/0304-4165(95)00051-C
16. Mirande C, Mosoni P, Bera-Maillet C, Bernalier-Donadille A, Forano E. 2010.
Characterization of Xyn10A, a highly active xylanase from the human gut
bacterium Bacteroides xylanisolvens XB1A. Appl. Microbiol. Biotechnol.
87:2097–2105. 10.1007/s00253-010-2694-0
17. Dodd D, Moon YH, Swaminathan K, Mackie RI, Cann IK. 2010. Transcriptomic
analyses of xylan degradation by Prevotella bryantii and insights into energy
acquisition by xylanolytic Bacteroidetes. J. Biol. Chem. 285:30261–30273.
10.1074/jbc.M110.141788
18. Kaoutari AE, Armougom F, Gordon JI, Raoult D, Henrissat B. 2013. The abundance
and variety of carbohydrate-active enzymes in the human gut microbiota. Nat.
Rev. Microbiol. 11:497–504. 10.1038/nrmicro3050
19. Martens EC, Koropatkin NM, Smith TJ, Gordon JI. 2009. Complex glycan
catabolism by the human gut microbiota: the Bacteroidetes Sus-like paradigm. J.
Biol. Chem. 284:24673–24677. 10.1074/jbc.R109.022848
20. Martens EC, Lowe EC, Chiang H, Pudlo NA, Wu M, McNulty NP, Abbott DW,
Henrissat B, Gilbert HJ, Bolam DN, Gordon JI. 2011. Recognition and
degradation of plant cell wall polysaccharides by two human gut symbionts.
PLoS Biol. 9:e1001221. 10.1371/journal.pbio.1001221
21. Collins T, Gerday C, Feller G. 2005. Xylanases, xylanase families and extremophilic
xylanases. FEMS Microbiol. Rev. 29:3–23. 10.1016/j.femsre.2004.06.005
201
22. Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, Henrissat B. 2009.
The Carbohydrate-Active EnZymes database (CAZy): an expert resource for
glycogenomics. Nucleic Acids Res. 37:D233–D238. 10.1093/nar/gkn663
23. Yoon K-H, Yun HN, Jung KH. 1998. Molecular cloning of a Bacillus sp. KK-1
xylanase gene and characterization of the gene product. Biochem. Mol. Biol. Int.
45:337–347
24. Elleuche S, Piascheck H, Antranikian G. 2011. Fusion of the OsmC domain from
esterase EstO confers thermolability to the cold-active xylanase Xyn8 from
Pseudoalteromonas arctica. Extremophiles 15:311–317. 10.1007/s00792-011-
0361-8
25. Brennan Y, Callen WN, Christoffersen L, Dupree P, Goubet F, Healey S, Hernandez
M, Keller M, Li K, Palackal N, Sittenfeld A, Tamayo G, Wells S, Hazlewood GP,
Mathur EJ, Short JM, Robertson DE, Steer BA. 2004. Unusual microbial
xylanases from insect guts. Appl. Environ. Microbiol. 70:3609–3617.
10.1128/AEM.70.6.3609-3617.2004
26. Lee CC, Kibblewhite-Accinelli RE, Wagschal K, Robertson GH, Wong DW. 2006.
Cloning and characterization of a cold-active xylanase enzyme from an
environmental DNA library. Extremophiles 10:295–300. 10.1007/s00792-005-
0499-3
27. Honda Y, Kitaoka M. 2004. A family 8 glycoside hydrolase from Bacillus halodurans
C-125 (BH2105) is a reducing end xylose-releasing exo-oligoxylanase. J. Biol.
Chem. 279:55097–55103. 10.1074/jbc.M409832200
202
28. Lagaert S, Van Campenhout S, Pollet A, Bourgois TM, Delcour JA, Courtin CM,
Volckaert G. 2007. Recombinant expression and characterization of a reducing-
end xylose-releasing exo-oligoxylanase from Bifidobacterium adolescentis. Appl.
Environ. Microbiol. 73:5374–5377. 10.1128/AEM.00722-07
29. Bakir MA, Kitahara M, Sakamoto M, Matsumoto M, Benno Y. 2006. Bacteroides
intestinalis sp. nov., isolated from human faeces. Int. J. Syst. Evol. Microbiol.
56:151–154. 10.1099/ijs.0.63914-0
30. Petersen TN, Brunak S, von Heijne G, Nielsen H. 2011. SignalP 4.0: discriminating
signal peptides from transmembrane regions. Nat. Methods 8:785–786.
10.1038/nmeth.1701
31. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature 227:680–685. 10.1038/227680a0
32. Whitmore L, Wallace BA. 2004. DICHROWEB, an online server for protein
secondary structure analyses from circular dichroism spectroscopic data. Nucleic
Acids Res. 32(Suppl 2):W668–W673. 10.1093/nar/gkh371
33. Powell JC, Lever M. 1972. A new automated procedure for the colorimetric
determination of glucose. Biochem. Med. 6:543–547. 10.1016/0006-
2944(72)90008-7
34. Dodd D, Kiyonari S, Mackie RI, Cann IK. 2010. Functional diversity of four glycoside
hydrolase family 3 enzymes from the rumen bacterium Prevotella bryantii B14. J.
Bacteriol. 192:2335–2345. 10.1128/JB.01654-09
35. Miyazaki K, Miyamoto H, Mercer DK, Hirase T, Martin JC, Kojima Y, Flint HJ. 2003.
Involvement of the multidomain regulatory protein XynR in positive control of
203
xylanase gene expression in the ruminal anaerobe Prevotella bryantii B14. J.
Bacteriol. 185:2219–2226. 10.1128/JB.185.7.2219-2226.2003
36. Fushinobu S, Hidaka M, Honda Y, Wakagi T, Shoun H, Kitaoka M. 2005. Structural
basis for the specificity of the reducing end xylose-releasing exo-oligoxylanase
from Bacillus halodurans C-125. J. Biol. Chem. 280:17180–17186.
10.1074/jbc.M413693200
37. Dodd D, Mackie RI, Cann IK. 2011. Xylan degradation, a metabolic property shared
by rumen and human colonic Bacteroidetes. Mol. Microbiol. 79:292–304.
10.1111/j.1365-2958.2010.07473.x
38. Flint HJ, Bayer EA, Rincon MT, Lamed R, White BA. 2008. Polysaccharide
utilization by gut bacteria: potential for new insights from genomic analysis. Nat.
Rev. Microbiol. 6:121–131. 10.1038/nrmicro1817
39. Koropatkin NM, Cameron EA, Martens EC. 2012. How glycan metabolism shapes
the human gut microbiota. Nat. Rev. Microbiol. 10:323–335.
10.1038/nrmicro2746
40. Topakas E, Moukouli M, Dimarogona M, Vafiadi C, Christakopoulos P. 2010.
Functional expression of a thermophilic glucuronyl esterase from Sporotrichum
thermophile: identification of the nucleophilic serine. Appl. Microbiol. Biotechnol.
87:1765–1772. 10.1007/s00253-010-2655-7
41. Spanikova S, Biely P. 2006. Glucuronoyl esterase—novel carbohydrate esterase
produced by Schizophyllum commune. FEBS Lett. 580:4597–4601.
10.1016/j.febslet.2006.07.033
204
42. Varghese JN, Hrmova M, Fincher GB. 1999. Three-dimensional structure of a barley
beta-D-glucan exohydrolase, a family 3 glycosyl hydrolase. Structure 7:179–190.
10.1016/S0969-2126(99)80024-0
43. Juncker AS, Willenbrock H, Von Heijne G, Brunak S, Nielsen H, Krogh A. 2003.
Prediction of lipoprotein signal peptides in Gram-negative bacteria. Protein Sci.
12:1652–1662. 10.1110/ps.0303703
44. Seydel A, Gounon P, Pugsley AP. 1999. Testing the ‘+2 rule' for lipoprotein sorting
in the Escherichia coli cell envelope with a new genetic selection. Mol. Microbiol.
34:810–821. 10.1046/j.1365-2958.1999.01647.x
45. Yamaguchi K, Yu F, Inouye M. 1988. A single amino acid determinant of the
membrane localization of lipoproteins in E. coli. Cell 53:423–432. 10.1016/0092-
8674(88)90162-6
46. Collins T, De Vos D, Hoyoux A, Savvides SN, Gerday C, Van Beeumen J, Feller G.
2005. Study of the active site residues of a glycoside hydrolase family 8
xylanase. J. Mol. Biol. 354:425–435. 10.1016/j.jmb.2005.09.064
47. Pollet A, Schoepe J, Dornez E, Strelkov SV, Delcour JA, Courtin CM. 2010.
Functional analysis of glycoside hydrolase family 8 xylanases shows narrow but
distinct substrate specificities and biotechnological potential. Appl. Microbiol.
Biotechnol. 87:2125–2135. 10.1007/s00253-010-2659-3
48. Yin Y, Mao X, Yang J, Chen X, Mao F, Xu Y. 2012. dbCAN: a web resource for
automated carbohydrate-active enzyme annotation. Nucleic Acids Res.
40:W445–W451. 10.1093/nar/gks479
205
Appendix C. Functional and modular analyses of diverse endoglucanases from
Ruminococcus albus 8, a specialist plant cell wall degrading bacterium
Pei-Ying Hong, Michael Iakiviak, Dylan Dodd, Meiling Zhang, Roderick I. Mackie, and
Isaac Cann
Manuscript published Scientific Reports 6, Article number: 29979 (2016)
C.1 Abstract
Ruminococcus albus 8 is a specialist plant cell wall degrading ruminal bacterium capable of utilizing hemicellulose and cellulose. Cellulose degradation requires a suite of enzymes including endoglucanases, exoglucanases, and β-glucosidases. The enzymes employed by R. albus 8 in degrading cellulose are yet to be completely elucidated. Through bioinformatic analysis of a draft genome sequence of R. albus 8, seventeen putatively cellulolytic genes were identified. The genes were heterologously expressed in E. coli, and purified to near homogeneity. On biochemical analysis with cellulosic substrates, seven of the gene products (Ra0185, Ra0259, Ra0325, Ra0903,
Ra1831, Ra2461, and Ra2535) were identified as endoglucanases, releasing predominantly cellobiose and cellotriose. Each of the R. albus 8 endoglucanases, except for Ra0259 and Ra0325, bound to the model crystalline cellulose Avicel, confirming functional carbohydrate binding modules (CBMs). The polypeptides for
Ra1831 and Ra2535 were found to contain distantly related homologs of CBM65.
Mutational analysis of residues within the CBM65 of Ra1831 identified key residues required for binding. Phylogenetic analysis of the endoglucanases revealed three distinct subfamilies of glycoside hydrolase family 5 (GH5). Our results demonstrate that
206
this fibrolytic bacterium uses diverse GH5 catalytic domains appended with different
CBMs, including novel forms of CBM65, to degrade cellulose.
C.2 Introduction
Cellulose is the main structural component of the plant cell wall and represents a
large source of renewable carbon (1). As a structural polysaccharide, cellulose is
recalcitrant to degradation due to the high degree of crystallinity formed by extensive
hydrogen bonding, as well as the stability of the β-1,4 glycosidic linkage. During
synthesis, the hydrogen-bonding network may be disrupted, forming amorphous regions
that are more accessible to enzymatic attack (2).
The degradation of cellulose mainly requires the concerted action of three
classes of enzymes, namely endoglucanases, exoglucanases, and β-glucosidases, which may also harbor cellodextrinase activity. Endoglucanases (EC 3.2.1.4) randomly cleave β-1,4 linkages of cellulose to yield shorter chain products. Exoglucanases (EC
3.2.1.91) bind to the glucan chain ends and cleave off cellobiose, the repeating unit of cellulose. Thus, exoglucanases release cellobiose as the main end products. The cellobiose and cello-oligosaccharides released by the endoglucanase and the
exoglucanase are further hydrolyzed into glucose monomers by β-glucosidases (EC
3.2.1.21) or cellodextrinases (EC 3.2.1.74). Though rare, processive endoglucanases are able to cleave the linkages in glucan chains in a random manner, as described for the commonly known endoglucanases, and then continue to processively release cellobiose, as described for exoglucanases (3,4). Synergy in degradation of cellulose to
soluble sugars or reducing ends can be detected in reactions containing β-
glucosidases, endoglucanases, and exoglucanases (5,6).
207
Cellulolytic organisms commonly possess multiple enzymes within each category
of the enzymes described above (7). Furthermore, multiple glycoside hydrolase (GH)
catalytic modules can exist in a single polypeptide chain with other non-catalytic
accessory modules, such as carbohydrate-binding modules (CBM). The modular
arrangement of the polypeptide may improve the capacity of an enzyme to hydrolyze its
substrate by increasing affinity and proximity to the substrate (8). Carbohydrate-binding
modules can exhibit specific or non-specific binding to substrates (8,9), and due to their
unique characteristics, CBMs have been adapted to technologies such as affinity tags
and analysis of carbohydrate content of a substrate (10). Recent reports suggest a
greater diversity of CBMs than is currently known, and this is evidenced by the
identification and biochemical characterization of a new CBM family designated CBM65
(11).
Ruminal microbes have evolved to degrade plant cell wall polysaccharides.
Genome sequences from the major fibrolytic ruminal bacteria, Fibrobacter
succinogenes S85, Prevotella bryantii B14, Prevotella ruminicola 23, and Ruminococcus
albus (strains 7 and 8) reveal multiple putative carbohydrate active enzymes implicated
in the degradation of cellulose and hemicelluloses. Furthermore, analyses of rumen
derived metagenomic sequences have led to the discovery of unique enzymes, proving
the rumen as a rich source of fibrolytic enzymes (12).
Ruminococcus albus is a dominant cellulolytic bacterium in the rumen, and it is able to degrade and ferment both cellulose and hemicellulose (13). During growth of R. albus 7 on cellulose in continuous culture, soluble products are rapidly utilized resulting
208
in trace amount of soluble sugars in the culture medium, suggesting depolymerization of
cellulose as the rate-limiting step (14). Adhesion of R. albus cells to cellulose is
important for degradation and utilization of the substrate, as adhesion defective mutants
cannot effectively utilize cellulose (15,16). Previous work on an R. albus 8 adhesion
defective mutant identified two proteins, Cel9B and Cel48A, as the potential cause for
decreased cellulose hydrolysis (16). These two proteins contain a novel CBM found only in R. albus. The unique CBM, which binds to various polysaccharides, as well as the bacterial cell wall, has been designated CBM37. The R. albus CBM37 is proposed to function by shuttling the catalytic domain, to which it is attached, between the bacterial cell wall and the polysaccharide substrate (17).
Although several enzymes from related strains have been cloned and partially
characterized (Table 13), no cellulolytic enzymes from R. albus 8 have been analyzed in
detail. In the present report, seven endoglucanases from R. albus 8 were biochemically
characterized to determine their contribution to hydrolysis of cellulose. Furthermore,
examination of the binding capacity of the endoglucanases reflected the predicted
functionality of their CBMs, and truncational analysis revealed a potentially novel CBM
present in Ra1831, one of the endoglucanases. Mutational analysis of the Ra1831 CBM
uncovered the residues involved in substrate recognition and binding. The results
presented here provide a deeper understanding of the enzymes employed by one of the
major plant cell wall degrading organisms in the rumen, and the insights gained should
also be of utility to efforts aimed at plant cell wall depolymerization for conversion to
value added products, including biofuels.
209
Table 13. R. albus genes previously cloned, expressed, and demonstrated to exhibit cellulose degrading activity
Cellulases Accession # R. Secreted GH CBM Familyb albus (predicted)a familyb strain EGI (CelA) AAA26469 F-40 Yes GH 5 EgIV BAA32286 F-40 No GH 5 (CelB) EgV BAA92146 F-40 Yes GH 5 (CelC) Cel5D BAA92430 F-40 Yes GH 5 2xCBM4_9 Cel9Ac BAB64431 F-40 Yes GH 9 CBM3 EgA AAA26467 SY3 No GH 5 EgB CAA38693 SY3 Yes GH 5
a Signal peptides were predicted using SignalP
(http://www.cbs.dtu.dk/services/SignalP/) b Domain identity was predicted using Pfam (http://pfam.sanger.ac.uk/) c Proteins not biochemically characterized
210
C.3 Materials and Methods
Materials. Genomic DNA was extracted from a culture of R. albus 8 stored in a culture collection at the Department of Animal Science at the University of Illinois at
Urbana-Champaign. Escherichia coli strains, JM109 and BL-21-CodonPlus (DE3) RIPL, were purchased from Stratagene (La Jolla, CA). The pET-46 Ek/LIC vector kit was from
Novagen (San Diego, CA). The pGEM-T Easy Vector system was purchased from
Promega (Madison, WI). QIAprep Spin Miniprep kit was from Qiagen (Valencia, CA).
Cello-oligosaccharides and medium viscosity wheat arabinoxylan (WAX) were purchased from Megazyme (Bray, Ireland). All other reagents were of the highest possible purity and were purchased from Sigma-Aldrich (St. Louis, MO).
Gene cloning. A bioinformatic search of the R. albus 8 draft genome sequence
(accession no. NZ_ADKM00000000) revealed 17 genes as putatively involved in cellulose degradation. These genes include two open reading frames (ORFs) for putative GH3 β-glucosidases (ra2003 and ra3307) and three ORFs coding for putative
GH9 exoglucanases (ra2259, ra3055, ra3241). There were 12 genes that were predicted to encode endoglucanases; ten GH5 (ra0185, ra0259, ra0325, ra0711, ra0903, ra1553, ra1831, ra2461, ra2535, ra2979), one GH9 (ra2876), and one GH48
(ra2561) polypeptides.
The genes for the putative endoglucanases were cloned using primers designed to delete the N-terminal region coding for signal peptides to prevent secretion and rather enhance accumulation of the recombinant protein in the cell (Table 14). Signal peptides were predicted using the SignalP v3.0 online server (www.cbs.dtu.dk/services/SignalP/)
(Table 14).
211
Table 14. Primers used in cloning R. albus 8 functional endoglucanases, their predicted signal peptides, and characteristic of recombinant proteins.
MW Extinction Forward primer Accession Signal peptide (kDa) coefficient Gene a -1 -1 a number (N → C terminal) (M cm ) Reverse primer 5’- GACGACGACAAGATGGCTTT 82.2 148,795 MKTKSRRALTLS ZP_08160 CGAGAAAGATGCCAAACAGACC -3’ ra0185 TVCAMALSCLAF 335.1 5’- GAGGAGAAGCCCGGTTCAGT AEPTRAEA TGCTCTTTGTGATGACCAAGTCTG -3’ 5’- GACGACGACAAGATGCTGAAAA 36.7 62,630 ZP_08160 AACTAAAGGTTATCGGAGGAAGG -3’ ra0259 No signal peptide 389.1 5’- GAGGAGAAGCCCGGTTCATTTG TTTTTCTCCTCTTCTGCAAAGTCC -3’ 5’- GACGACGACAAGATGGATAA 44.9 78,965 ZP_08158 GGACAGCAAGACAGAAAGCAGG -3’ MKLKKFAALFIAA ra0325 005.1 5’- GAGGAGAAGCCCGGTTC AMSVGTLSACG ATGAAGCGGAAGCAGGGTC -3’ 5’- GACGACGACAAGATGGCAACA 73.5 202,515 MNKVFKRMTAA ZP_08159 TCAGCAGTGAATGACACCAATG -3’ ra0903 ASVAALTLSFAA 293.1 5’- GAGGAGAAGCCCGGTTTACTTTACA ASMPAVVTASA GTGATAGTCACAGCGTTCTTGATAGC-3’ 5’- GACGACGACAAGATGCTGC 72.0 171,340 MKKTIQGRVVSA ZP_08160 CTTCAGCAGTCATCGAAGC -3’ ra1831 LSAAAIALTGMSS 422.1 5’- GAGGAGAAGCCCGGTTTACGGAAC LPSAVIEA TGTAACTATTACAGCGTTCTTGATCG -3’ 5’- GACGACGACAAGATGGAAAAATCTG 70.5 161,145 MNGGISKKIIAAF ZP_08158 ACAAAGACGCAGTTTCAAAGAAGTC-3’ ra2461 VALAAVIIALVVLL 504.1 5’- GAGGAGAAGCCCGGTTCAAA VVTNFSGG GCCCCTTAGCTTTTGAAAGCG -3’ 5’- GACGACGACAAGATGAAAGAC 71.5 120,460 MKVTFKRMLSLA ZP_08158 GTGTCAAAAATGACACCCTTCG -3’ ra2535 AAGAMTLSAMPV 565.1 5’- GAGGAGAAGCCCGGTTTATCC LTASA TGCAATAGTTTTGTCGGGGAGC -3’ a ProtParam (http://web.expasy.org/protparam/) was used to predict molecular weights and extinction coefficients
212
Each gene analyzed in this study was amplified using the PicoMaxx high fidelity
PCR kit (Stratagene) and cloned into the pET-46b vector using the ligation independent
cloning technique following the protocols of the manufacturer. Recombinant plasmids
were transformed into E. coli JM109 cells via electroporation, plated, and grown for
16 hours at 37 °C on lysogeny broth (LB) agar plates supplemented with ampicillin
(100 μg/mL). A single colony from each transformation was inoculated into LB medium
(10 mL) containing ampicillin (100 μg/mL) and grown to saturation. The cells were
subsequently centrifuged (3220 × g, 15 min, 4 °C) and plasmids were extracted from the
resulting cell pellet using the QIAprep Spin Miniprep kit (Qiagen, Valencia, CA). The
foreign DNA inserts in the recombinant plasmids were sequenced to confirm the
correctness of the gene sequence (W. M. Keck Center for Comparative and Functional
Genomics at the University of Illinois at Urbana-Champaign). Ra2876 was unable to be
cloned and excluded from further analysis.
Gene expression and protein purification. For gene expression, the pET-46b
vector with the correct DNA insert was transformed into E. coli BL-21 CodonPlus (DE3)
RIPL cells by the heat shock method and grown overnight at 37 °C on LB agar plates supplemented with ampicillin (100 μg/mL) and chloramphenicol (50 μg/mL). After
16 hours, five colonies were used to inoculate 10 mL of fresh LB medium supplemented
with 100 μg/mL of ampicillin and 50 μg/mL of chloramphenicol and grown at 37 °C for 6–
8 hours with vigorous aeration. The pre-cultures were then added to fresh LB medium
(1 L), supplemented with the two antibiotics at the same concentrations as described
above, and grown with vigorous aeration at 37 °C. At OD600 of 0.3, isopropyl β-D-thio-
galactopyranoside (IPTG) was added to each culture at a final concentration of 0.1 mM,
213
and the temperature was decreased to 16 °C. Following 16 hours of culturing, cells were
pelleted by centrifugation (4,000 × g, 15 min, 4 °C) and re-suspended in 30 mL of
binding buffer (50 mM Tris-HCl, 300 mM NaCl, pH 7.5). The cell suspensions were
subjected to two passages through an EmulsiFlex C-3 cell homogenizer (Avestin,
Ottawa, Canada), and the cell debris in the cell lysates were removed by centrifugation
(20,000 × g, 30 min, 4 °C).
To purify the recombinant proteins, Talon Metal Affinity Resin (Clontech) was used according to the protocol of the manufacturer with a modified binding/wash buffer
(50 mM Tris-HCl, 300 mM NaCl, pH 7.5) and elution buffer (50 mM Tris-HCl, 300 mM
NaCl, 150 mM imidazole, pH 7.5). Ra2561 was found to be insoluble and excluded from further analysis. Partially purified proteins were screened for catalytic activity by spotting
5 μL of protein to a substrate infused (0.25%) agarose petri dish containing either CMC, glucomannan, or xyloglucan. The plate was incubated at 37 °C for 16 hours and subsequently stained with Congo Red and destained with 1 M NaCl. Thin layer chromatography was also employed to screen for activity as described below (Fig. 36).
To further purify the seven proteins capable of degrading cellulose, elution
fractions from the Talon resin containing Ra0185, Ra0259, Ra0325, Ra0911, Ra1831,
Ra2461 or Ra2535 were subjected to size-exclusion chromatography (HiLoad 16/60
Superdex 200 prep grade column, GE Healthcare) on an AKTAxpress fast protein liquid
chromatograph equipment (FPLC, GE Healthcare). The chromatography was
developed with a Tris buffer (50 mM Tris-HCl, pH 7.0). Elution fractions of Ra0259 and
Ra0325 were dialyzed into protein storage buffer (50 mM Tris, 150 mM NaCl, pH 7.5) and stored at 4°C until used for enzymatic reactions. For Ra0185, Ra0911, Ra1831,
214
Figure 36. Thin-layer chromatography screening of R. albus 8 endoglucanases. Cellulosic substrates, CMC (0.5% w/v) (A,C) and PASC (0.5% w/v) (B,D), were incubated with partially purified putative endoglucanases (5% v/v) for 16 hours. The soluble products were separated by thin layer chromatography and visualized by spraying with methanolic orcinol.
215
Ra2461, and Ra2535, anion exchange chromatography (5 mL HiTrap Q HP column, GE
Healthcare) was performed after gel filtration to achieve purification. Proteins were
bound to the column in the same buffer used for gel filtration, and eluted with the
binding buffer supplemented with 1 M NaCl (50 mM Tris-HCl, 1 M NaCl, pH 7.0). Elution fractions were pooled and dialyzed into protein storage buffer and stored at 4 °C until used for enzymatic reactions. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used to assess protein purity. The protein bands were visualized by staining with Coomassie brilliant blue G-250, as described previously
(26). Purified proteins were quantified by measuring absorbance at 280 nm (A280nm)
using a NanoDrop 1000 (Thermo Scientific, Waltham, MA) and the molecular mass and
the theoretical extinction coefficient of each protein (Table 14).
Analysis to assign function to regions of unknown function in the R. albus
8 endoglucanases. To identify the function of unknown regions of Ra1831 and the
CBM65 from Ra2535, truncational mutants were produced and tested for carbohydrate
binding activity. The R. albus 8 Ra1831 TM1 contained approximately the C-terminal
half of the entire polypeptide, including the region of unknown function as well as the
CBM37. The TM2 truncational mutant contained only the region of unknown function.
The Ra2535 TM1 truncational mutant contained the C-terminal half of the polypeptide, including the putative CBM65 and dockerin domains. Compared to the TM1 mutant, the
Ra2535 TM2 mutant lacked the dockerin-like sequences, and the TM3 contained only the putative CBM65. The primers used to create the truncated derivatives are shown in
Table 14. The DNA sequences encoding the regions of unassigned function in Ra1831
(or TM1 and TM2) were cloned into a modified pET-28a vector as described previously
216
(27). Briefly, primers were designed to include a 5′-NdeI and 3′-XhoI restriction sites to facilitate cloning (Table 15). After amplification using PicoMaxx high fidelity PCR kit, the
PCR products were ligated into the pGEM-T vector following the protocol of the manufacturer. The inserts were excised from the plasmids by digesting with NdeI and
XhoI and then ligated into a modified pET-28a vector already digested with NdeI and
XhoI. The inserts in the recombinant plasmids were sequenced to confirm the correctness of each DNA insert (W. M. Keck Center for Comparative and Functional
Genomics at the University of Illinois). The expression and purification of Ra1831 TM1 and TM2 were carried out as described above for the full length protein. The regions of
Ra2535 designated TM1, TM2, and TM3 were cloned, expressed, and the polypeptides purified as described above for the full length protein. Final protein concentrations were calculated using the molecular mass and theoretical extinction coefficients (Table 15).
Mutational analysis of Ra1831 TM2 to identify key residues for binding to substrate. In order to identify amino acid residues that may be involved in binding to substrates, an amino acid sequence alignment was carried out with Ra1831 TM1 and its related sequences in the NCBI. Mutations were made in the pET-46b containing the non-mutated Ra1831 TM2 using QuikChange Lightning Site-Directed Mutagenesis Kit
(Agilent, Santa Clara) with the primers listed in Table 16. Mutated plasmids were transformed into E. coli JM109 by electroporation and grown overnight at 37 °C on LB agar plates supplemented with ampicillin (100 μg/mL). Individual colonies from each mutagenic PCR were inoculated into LB, with ampicillin, and plasmids were extracted using the QIAprep Spin Miniprep kit (Qiagen, Valencia, CA). Plasmids containing the expected mutations were identified by DNA sequencing, and the proteins were
217
Table 15. Primers used in cloning truncational mutants of Ra1831 and Ra2535 and characteristic of recombinant proteins.
MW Extinction Forward Primer (kDa) a coefficient Gene (M-1 cm-1) a Reverse Primer 5’- CATATGGGCTGCACCTGGTATCCTAACTCTG -3’ 32.2 80,790 ra1831 TM1 5’- CTCGAGTTACGGAACTGTAACTATTACAGCGTTCTTG -3’ 5’- CATATGGGCTGCACCTGGTATCCTAACTCTG -3’ 20.6 53,860 ra1831 TM2 5’- CTCGAGTTAGCCTGAAACAGGTACCGCATAC -3’ 5’-GACGACGACAAGATGTCTCTGCCGATGGTAAACAAGCTGATC -3’ 34.7 46,300 ra2535 TM1 5’- GAGGAGAAGCCCGGTTTATCCTGCAATAGTTTTGTCGGGGAGC - 3’ 5’- GACGACGACAAGATGTCTCTGCCGATGGTAAACAAGCTGATC -3’ 27.2 44,810 ra2535 TM2 5’- GAGGAGAAGCCCGGTTAACCGCTGATGTCGGCGG -3’ 5’- GACGACGACAAGATGTCTCTGCCGATGGTAAACAAGCTGATC -3’ 12.0 10,430 ra2535 TM3 5’- GAGGAGAAGCCCGGTTAGCCGTAGGCAGACCTGTATG -3’
Bolded and underlined sequence indicate the NdeI site
Bolded sequence indicate the XhoI site a ProtParam (http://web.expasy.org/protparam/) was used to predict molecular weights and extinction coefficients
218
Table 16. Primers used in site directed mutagenesis of ra1831 TM2 and characteristic of recombinant proteins.
MW Extinction Forward primer a (kDa) coefficient Gene -1 -1 a (M cm ) Reverse primer ra1831 5’- CGTAAATATAGGTTCTGACGCGGTAGAGCTTTACAAGAATC -3’ TM2 20.435 48,360 W433A 5’- GATTCTTGTAAAGCTCTACCGCGTCAGAACCTATATTTACG -3’ ra1831 5’- CAAGTTCGTACTCGTTGCCGATTCCGATCAGGCAC -3’ TM2 20.458 52,370 Y470A 5’- GTGCCTGATCGGAATCGGCAACGAGTACGAACTTG -3’ ra1831 5’- GCAGCTACGACACAGCAAACCGCGTAGCTTCC -3’ TM2 20.435 48,360 W488A 5’- GGAAGCTACGCGGTTTGCTGTGTCGTAGCTGC -3’ 5’- CTTACATAAAGGAATTCTATGCAGATGACCTGACAAGCGCTATC - ra1831 3’ TM2 20.458 52,370 5’- GATAGCGCTTGTCAGGTCATCTGCATAGAATTCCTTTATGTAAG - Y507A 3’ ra1831 5’- GAATTCTATTATGATGCCCTGACAAGCGCTATC -3’ TM2 20.506 53,860 D509A 5’- GATAGCGCTTGTCAGGGCATCATAATAGAATTC -3’ ra1831 5’- CTCTACATCAGCGCTGCCACGAGTTCGCTG -3’ TM2 20.520 53,860 T532A 5’- CAGCGAACTCGTGGCAGCGCTGATGTAGAG -3’ ra1831 5’- GACTGCTTACGGTCTGGCTGCGGTACCTGTTTC -3’ TM2 20.458 52,370 Y542A 5’- GAAACAGGTACCGCAGCCAGACCGTAAGCAGTC -3’ ra1831 5’- GACTGGGTAGAGCTTGCCAAGAATCCTAACG -3’ TM2 20.458 52,370 Y437A 5’- CGTTAGGATTCTTGGCAAGCTCTACCCAGTC -3’ ra1831 5’- GCTCCACCATTGAAGCAGCGAAGAACTTCACAGTATC -3’ TM2 20.435 48,360 W448A 5’- GATACTGTGAAGTTCTTCGCTGCTTCAATGGTGGAGC -3’ ra1831 5’- GAACTTCACAGTATCAAACGCGAAGAACTATGTAAACAAG -3’ TM2 20.435 48,360 W456A 5’- CTTGTTTACATAGTTCTTCGCGTTTGATACTGTGAAGTTC -3’ ra1831 5’- CACCCGAGCTGGTTCTCGCAGACGGCAGCTACGACAC -3’ TM2 20.493 53,860 Q481A 5’- GTGTCGTAGCTGCCGTCTGCGAGAACCAGCTCGGGTG -3’ ra1831 5’- GAGCTGGTTCTCCAGGCAGGCAGCTACGACACATG -3’ TM2 20.506 53,860 D482A 5’- CATGTGTCGTAGCTGCCTGCCTGGAGAACCAGCTC -3’ a ProtParam (http://web.expasy.org/protparam/) was used to predict molecular weights and extinction coefficients
219
expressed and purified using the protocol described above for Ra1831 TM2. Each mutated derivative of Ra1831 TM2 was purified by immobilized metal affinity chromatography (Talon resin) and gel filtration with the buffers described above for the two methods. The proteins were quantified based on their molecular mass and theoretical extinction coefficient as listed in Table 16.
Determination of specific activities of R. albus 8 endoglucanases on phosphoric acid swollen cellulose (PASC). The specific activities of Ra0185,
Ra0259, Ra0325, Ra0903, Ra1831, Ra2461, and Ra2535 were determined at 37 °C in a buffer composed of 50 mM sodium phosphate (pH 6.5) and 150 mM NaCl. Phosphoric acid swollen cellulose (PASC) was prepared following the method of Wood (28). Each protein (0.5 μM of Ra0185, Ra0325, Ra0903, Ra1831, Ra2461, and Ra2535 and
1.0 μM of Ra0259) was incubated with PASC (0.5% w/v, final concentration) for 0, 3, 6, and 9 min, and the samples were boiled to stop the reaction. The concentration of soluble reducing ends released was quantified with the para-hydroxybenzoic acid hydrazide (PAHBAH) method as described previously (29). Cellobiose, representing the repeating unit of cellulose, was used as the standard in generating a prediction equation to estimate the reducing ends released by each enzyme on the PASC. Specific activity was calculated as the amount of cellobiose equivalent released per time per enzyme concentration.
Kinetic parameters were estimated using a range of PASC concentrations from
0.08% w/v to 1.2% w/v. This range of substrate concentration was used due to a lack of sensitivity below 0.08% w/v and concentrations above 1.2% w/v were too viscous. The methods were described in our previous reports (30,31).
220
End products from hydrolysis of phosphoric acid swollen cellulose (PASC).
Hydrolysis of PASC by Ra0185, Ra0259, Ra0325, Ra0903, Ra1831, Ra2461, and
Ra2535 were determined by incubating each enzyme (0.5 μM) with PASC (0.5% w/v, final concentration) in the reaction buffer (50 mM sodium phosphate, 150 mM NaCl, pH
6.5) for 16 hours at 37 °C. The resulting concentration of reducing ends was quantified using the PAHBAH method with the standard curve generated as described above. Thin layer chromatography (TLC) was employed to identify the products of hydrolysis from
PASC. The TLC (plate dimensions 10 cm × 20 cm) method followed the procedure described in our earlier report (25).
Quantitative analysis of end products of hydrolysis. To quantitatively assess products released from the enzymatic hydrolysis, high performance anion exchange chromatography (HPAEC) was used to separate the mono- and oligo-saccharides produced from substrate hydrolysis. The methods were as described in our previous reports (30,31). The identity and concentration of mono- and oligo-saccharides produced from enzymatic hydrolysis of polysaccharides were determined by comparison of peak retention times and peak areas to those of known oligosaccharides with known concentrations.
Binding assays. To determine the functions of the putative carbohydrate-binding modules within Ra0185, Ra0903, Ra1831, and Ra2461, each enzyme (1 mg/ml) was incubated with Avicel (10% w/v) for 1 hour at 4 °C. After the incubation, each reaction mixture was centrifuged (25,000 × g, 10 minutes, 4 °C) to pellet the Avicel and any enzyme or protein bound to the model crystalline cellulose. The residual proteins
(present in the supernatant) were analyzed as the unbound fraction for comparison.
221
After three washes in the reaction buffer (50 mM sodium phosphate, 150 mM NaCl, pH
6.5), the protein associated with the pelleted Avicel was eluted by boiling with SDS- buffer and designated the bound fraction. The bound and unbound fractions (protein remaining in solution) were visualized via SDS-PAGE. During analysis of Ra0185, the
SDS-PAGE analysis found only a small portion of Ra0185 with the correct size, but several bands corresponding to cleaved or degraded products (data not shown), potentially due to the prolonged incubation with Avicel. Therefore, this protein was omitted from this analysis.
To determine the binding of recombinant proteins to soluble polysaccharides, non-denaturing gel electrophoresis was used. Polyacrylamide gels (6%) were infused with carboxymethyl cellulose (CMC) or wheat arabinoxylan (WAX) at a concentration of
0.25% w/v. Each recombinant protein or bovine serum albumin (BSA as control) (2 μg) was loaded into wells and electrophoresed for 2 hours and 15 minutes at 80 V. To visualize the protein bands, the gels were stained with Coomassie brilliant blue G-250.
The migration and banding patterns were compared between carbohydrate-infused gels and gels without substrate electrophoresed in juxtaposition.
Analysis of substrate binding of mutant CBMs by Isothermal titration calorimetry (ITC). Isothermal titration calorimetry was performed at 25 °C using a VP-
ITC micro-calorimeter (MicroCal Inc., Northampton, UK). Each of the Ra2535 and
Ra1831 truncational mutants was diluted to 50 μM in reaction buffer (50 mM sodium phosphate, 150 mM NaCl, pH 6.5) and 28 successive injections of 10 μl of oligosaccharide (10 mM cellopentaose or xylopentaose in reaction buffer) were administered at 300 second intervals. A non-linear regression, using a single site
222
binding model (MicroCal Origin software), was used to analyze the data collected, and the thermodynamic parameters were calculated with both the Gibbs free energy equation (ΔG = ΔH − TΔS) and the relationship ΔG = −RT lnKa.
Phylogenetic analysis. Representative GH5 from each subfamiliy, excluding
GH5_29, GH5_35, and GH5_50 due to low numbers of representatives, were retrieved and trimmed to contain only the catalytic domain, using dbCAN
(csbl.bmb.uga.edu/dbCAN/annotate.php) and Pfam (pfam.sanger.ac.uk/). ClustalW was used to obtain the phylogenetic tree and the circular cladogram was drawn using
MEGA6 (www.megasoftware.net/).
C.4 Results
Domain analysis of endoglucanases from R. albus 8. In order to identify cellulose degrading enzymes in R. albus 8, we analyzed the genome of the bacterium by bioinformatics and biochemical approaches. A search through the draft genome sequence of this bacterium yielded 17 genes predicted to code for enzymes involved in cellulose hydrolysis. The genes were expressed and their products were subjected to an initial enzymatic screen with cellulosic substrates. Seven polypeptides exhibited endoglucanase activities and were selected for further analysis.
In Fig. 37A, the domain architecture of the seven polypeptides demonstrated as endoglucanases (Ra0185, Ra0259, Ra0325, Ra0903, Ra1831, Ra2461, and Ra2535) in this study are presented. All of the endoglucanases possess GH5 catalytic domains; however, the presence of accessory modules and their organization varied among the polypeptides as depicted in Fig. 37A. Two endoglucanases lacked any accessory domains, Ra0259 and Ra0325. The R. albus 8 Ra0903 and Ra1831 both possess GH5
223
Figure 37. Domain architectures and purification of 7 glycoside hydrolase family 5 enzymes from Ruminococcus albus 8. (A) Domain architecture of 7 putative glycoside hydrolases predicted to be involved in cellulose hydrolysis. Domain boundaries were predicted using Pfam (http://pfam.sanger.ac.uk). Regions in light grey are signal peptides which were predicted using SignalP (http://www.cbs.dtu.dk/services/SignalP). Regions in dark grey display homology to dockerin sequences. Abbreviations: CBM, carbohydrate binding module; GH, glycoside hydrolase; Ig, immunoglobulin. (B) SDS-PAGE analysis of 7 putative endoglucanases purified using cobalt affinity chromatography and gel filtration. The R. albus 8 Ra0185, Ra0903, Ra1831, Ra2461 and Ra2535 were further purified using anion exchange chromatography. Lane L, molecular mass markers; Lane 1, Ra0185; Lane 2, Ra0259; Lane 3, Ra0325; Lane 4, Ra0903; Lane 5, Ra1831; Lane 6, Ra2461; Lane 7, Ra2535.
224
domains at the N-terminus and CBM37 domains at the C-terminus. The Ra0185 and
Ra2535 enzymes possess dockerin-like sequences as well as CBMs C-terminal to the
GH5 domain. Finally, Ra2461, contains a CBM2 N-terminal to the GH5 domain.
All the genes were cloned without the signal peptide to facilitate accumulation of
the gene products in the cytoplasm of the E. coli strain used for protein production.
Furthermore, a hexa-histidine tag (His-tag) in the plasmid was fused to the N-terminus of the recombinant protein to facilitate purification of each protein by affinity chromatography. The proteins were heterologously expressed in E. coli BL-21
CodonPlus RIPL and were purified to near homogeneity using immobilized metal affinity chromatography (IMAC), followed by size exclusion chromatography. Further purification was required for Ra0185, Ra0903, Ra1831, Ra2461, and Ra2535, and this was accomplished through anion exchange chromatography (Fig. 37B). For two of the
endoglucanases, Ra0185 and Ra0903, some lower molecular weight bands were
always present after purification, likely corresponding to degradation products of the
respective enzymes.
Quantification of PASC hydrolysis by the R. albus 8 enzymes. Each
predicted endoglucanase was screened for hydrolysis of cellulosic substrates through
overnight incubations on two different substrates, soluble carboxymethyl cellulose
(CMC) and the mostly amorphous cellulosic substrate phosphoric acid swollen cellulose
(PASC). Each of the enzymes depicted in Fig. 37A was capable of releasing products
from the cellulosic substrates, and the end products of hydrolysis were separated and
detected by TLC (Fig. 36).
225
Due to the higher release of end-products from PASC, the amorphous substrate was used to characterize the hydrolysis patterns of each endoglucanase (Fig. 38A). The endoglucanases Ra0185, Ra0259, Ra0325, Ra1831, and Ra2535 released reducing sugars at relatively similar rates, i.e., between 24.7 and 56.6 mmol of cellobiose equivalent per mmol of enzyme per minute. The enzyme with the highest activity,
Ra0903 was able to release reducing sugars at 102.8 mmol cellobiose equivalents per mmol enzyme per min, while the enzyme with the lowest activity, Ra2461, released cellobiose equivalents at a rate of 21.6 mmol per mmol enzyme per min.
Endoglucanases are expected to release a mixture of longer chain oligosaccharides, ranging from cellotriose to cellohexaose (18). Interestingly, all of the endoglucanases from R. albus 8 were found to release mixtures of shorter chain oligosaccharides, including cellotriose, cellobiose and small amounts of glucose (Fig.
38B). Similar results were observed when the substrate was digested with all of the endoglucanases as a mixture, although in this case more glucose was released.
Consistently, however, cellobiose was released as the largest proportion of the end products (Fig. 38B).
It was our prediction that the short chain oligosaccharides are the products of the prolonged (16 hours) hydrolytic reactions, i.e., it was possible that the oligosaccharides released from cellulose hydrolysis were initially longer chain oligosaccharides. Further incubation would then allow the longer chains to be hydrolyzed into the shorter chain oligosaccharides and glucose. To test this hypothesis, short-term hydrolytic reactions were conducted using representatives of the proteins with varying modular structures, including Ra0259 (no signal peptide, no accessory domains), Ra0325 (with signal
226
Figure 38. Comparison of cellulose hydrolysis by the endoglucanases of R. albus 8. (A) Specific activity (mmole of cellobiose equivalent released per minute per mmole of enzyme) of endoglucanases on phosphoric acid swollen cellulose (PASC) determined using the para-hydroxybenzahydrazide (PAHBAH) reducing sugar assay. (B) Hydrolysis of PASC (0.5% w/v) by individual endoglucanases and their mixture (0.5 µM each) was conducted for 16 hours at 37°C. Samples were analyzed by separating the soluble products via HPAEC-PAD. Products were identified by comparison of the retention time to those of glucose, cellobiose, and cellotriose. The concentrations of products were determined by comparison of the peak area to a standard curve created using known concentrations of oligosaccharide standards.
227
peptide but no accessory domains) and Ra2535 (with signal peptide and accessory
domains). All three enzymes (0.5 μM) released longer chain oligosaccharides (degree of
polymerization greater than 3), when incubated with PASC (0.5% w/v), as well as
shorter chains (DP less than 4) during the initial time points (Fig. 39). As the reaction
progressed, the proportion of shorter chain oligosaccharides increased with concomitant
decreases in the proportion of the longer chain oligosaccharides (Fig. 39).
To estimate the kinetic parameters of the R. albus 8 endoglucanases, release of
cellobiose equivalents were measured following incubation with increasing
concentrations of PASC. In general, the endoglucanase activities did not follow
Michaelis-Menten kinetics within the substrate range tested. The kinetic parameters could only be estimated for two of the endoglucanases, Ra0903 and Ra2461, which interestingly had the highest and lowest specific activities on PASC, respectively (Fig.
38A). In agreement with the specific activity data, the estimated turnover number (kcat) for Ra0903 was 1.7 s−1 whereas the value was 0.4 s−1 for Ra2461. However, the
estimated Km of Ra0903 for PASC was four times (5.6 mg/ml) that of Ra2461
−1 (1.3 mg/ml), leading to similar estimated catalytic efficiencies (kcat/Km) of ~0.3 mg ml
s−1 for the two enzymes (Fig. 40).
Phylogenetic analysis of GH5 modules of R. albus 8 endoglucanases. To
determine whether R. albus 8 has expanded its cellulose degrading capacity by using
the same family of GH5 to evolve different endoglucanases, we conducted a
phylogenetic analysis of the GH5 catalytic modules characterized in this study. As
shown in Fig. 41, the R. albus 8 GH5 modules fell within three distinct clusters: two in
GH5_1, one in GH5_2 and four in GH5_4. These findings support the concept of
228
Figure 39. Time course analysis of hydrolysis of phosphoric acid swollen cellulose by Ra0259, Ra0325 and Ra2535. PASC was incubated with endoglucanase (0.5 µM) for 0, 5, 10, 20, and 30 minutes. The soluble products were separated and detected using HPAEC-PAD. Representative HPAEC-PAD chromatographs are shown for Ra0259 (A), Ra0325 (B), and Ra2535 (C). Products were identified by comparison of retentions times to those of commercially available cello-oligosaccharides.
229
Figure 40. Enzyme kinetic analysis for individual endoglucanases. Individual endoglucanases (0.5 µM) were incubated with varying concentrations of PASC. The velocity of each reaction was determined by detecting release of soluble reducing sugars over time by using the PAHBAH reducing sugar assay. Reaction velocities were plotted against PASC concentrations and where applicable fitted to the Michaelis- Menten equation. A, B, C, D, E, F, and G represent Ra0185, Ra0259, Ra0325, Ra0903, Ra1831, Ra2461, and Ra2535, respectively.
230
Figure 41. Phylogenetic analysis of GH5 catalytic domains of R. albus 8 endoglucanases. Representative GH5 protein sequences from each GH5 subfamiliy, as identified by Aspeborg et al. (Aspeborg et al.) (excluding GH5_29, GH5_35, GH5_50 due to low number of representatives) were retrieved from the NCBI (http:www.ncbi.nlm.nih.gov) and trimmed to contain only the catalytic domain by using dbCAN (csbl.bmb.uga.edu/dbCAN/annotate.php) and Pfam (pfam.sanger.ac.uk/). Protein sequences were aligned using ClustalW. The circular cladogram phylogenetic tree was built using MEGA6 (www.megasoftware.net/).
231
expansive and increased diversity of GH5 containing enzymes in order to accommodate
glycan diversity.
Functional analysis of carbohydrate-binding modules. Putative CBMs are
encoded within five of the seven endoglucanases (Fig. 37A). To determine the functions
of the CBMs within the polypeptide chain, individual endoglucanases were incubated
with Avicel and the bound (B) and unbound (U) protein fractions were determined (Fig.
42A). The endoglucanase Ra0185 was found to be unstable, as several degradation
products were always seen during purification. This endoglucanase was, therefore,
excluded from the carbohydrate binding analysis. Two of the endoglucanases, Ra0259
and Ra0325, were predominantly found in the unbound fraction, i.e., not associated with
cellulose. As these two endoglucanases were not predicted to encode CBMs, it is
reasonable that they were unable to bind to the cellulose. The four remaining
endoglucanases, Ra0903, Ra1831, Ra2461, and Ra2535 were found predominantly in the bound fraction. Although initially we could not find a CBM of similar sequence with the region now identified as a CBM, subsequent search of the literature indicated that
Ra2535 harbors a member of a recently identified CBM family assigned to family 65
(11). So far, this is the only CBM65 homolog identified in the partial genome sequence
of R. albus 8. The first characterized member of this CBM family was identified in
Eubacterium cellulosolvens, another cellulose degrading Firmicute found in the rumen
(11). A number of amino acid residues determined to be responsible for binding to
substrate in the E. cellulosolvens homolog are not present in the CBM65 of Ra2535.
Therefore, further binding analysis with other substrates was conducted with
Ra2535. Migration of the protein on a non-denaturing polyacrylamide gel and a similar
232
Figure 42. Binding of soluble and insoluble polysaccharides by the R. albus 8 endoglucanases. (A) Individual endoglucanases (1 mg/ml) were incubated with Avicel (10% w/v) at 4°C for one hour. The unbound fraction or protein remaining in the supernatant (U) and the bound fraction (B) were separated by centrifuging the insoluble cellulose and washing away the proteins that did not adhere to the cellulose pellet. SDS-PAGE analyses and Coomassie staining were used to observe the bound and unbound proteins. Ra0185 was unstable, and therefore it was excluded from the analysis. (B) Affinity non-denaturing gel electrophoresis was performed to qualitatively analyze the capacity of Ra2535 to bind to insoluble polysaccharide (Lane 2, 4 and 6). Migration through a non-denaturing gel without polysaccharide (Blank, Lane 1 and 2) was compared to migration through non-denaturing gels containing carboxymethyl- cellulose (CMC, Lane 3 and 4) or wheat arabinoxylan (WAX, Lane 5 and 6). BSA (Lane 1, 3 and 5) was used as a standard for comparison of protein migration.
233
gel infused with carboxymethyl-cellulose, a soluble cellulose derivative, or wheat
arabinoxylan, was investigated (Fig. 42B). The CMC infused gel results were
inconclusive, as the migration of Ra2535 as well as that of BSA were retarded.
However, the migration of Ra2535 was retarded on WAX compared to the non-
denaturing gel without substrate, while BSA was not, indicating that the polypeptide
binds to the hemicellulose WAX. Interestingly, when Ra2535 was tested for hydrolytic
activity against WAX, there was release of reducing ends (data not shown), suggesting
that the enzyme also has hydrolytic activity on the xylan portion of the plant cell wall.
Truncational mutants reveal substrate binding range for the CBM65 of
Ra2535 and a novel CBM in R. albus 8 Ra1831. To further explore the range of substrates bound by the CBM65 of Ra2535, truncational mutants were made from the polypeptide. The endoglucanase Ra1831 also contains a large region without known function, while also possessing a sequence homologous to CBM37. The regions of interest in the two endoglucanases were either cloned alone or with their C-terminal sequences. In Fig. 43A, the modular architectures of the wild-type proteins and the truncational mutants are presented, and the recombinant proteins were purified to near homogeneity (Fig. 43B–D).
Detailed characterization of the Ra2535 truncational mutants were performed,
including binding of Ra2535 TM3 to insoluble substrates (Fig. 44). As expected, the protein was able to bind tightly to the crystalline cellulose Avicel. Although Ra2535 TM3 shows homology to the CBM65 from E. cellulosolvens (EcCBM65), the Ra2535 TM3 protein was also able to bind weakly to mannan and arabinan. These results differed from the results of the binding studies with the E. cellulosolvens CBM65. Several key
234
Figure 43. Truncational analyses for identification of CBMs in Ra1831 and Ra2535. (A) Domain architecture of the truncational mutants created from Ra1831 and Ra2535. Coloring and domain predictions were performed as in Fig. 37. Abbreviations: CBM, carbohydrate binding module; GH, glycoside hydrolase. (B) SDS-PAGE analysis of truncational mutants of Ra1831 purified using cobalt affinity chromatography and gel filtration. Lane L, molecular mass markers; Lane 1, Ra1831 TM1; Lane 2, Ra1831 TM2. (C) SDS-PAGE analysis of truncational mutants of Ra2535 purified using cobalt affinity chromatography and gel filtration. Lane L, molecular mass markers; Lane 1, Ra2535 TM1; Lane 2, Ra2535 TM2. (D) SDS-PAGE analysis of truncational mutants of Ra2535 CBM65 purified using cobalt affinity chromatography and gel filtration. Lane L, molecular mass markers; Lane 1, Ra2535 TM3.
235
Figure 44. Substrate binding characteristics of the CBM65 from endoglucanase Ra2535 to insoluble substrate. Purified Ra2535 TM3 (0.5 mg/ml) was incubated with Avicel, β-1,4 mannan, or debranched arabinan (10 mg/ml) at 4°C for one hour. The unbound fraction or protein remaining in the supernatant was separated from the bound fraction (B) as described in the text. An SDS-PAGE analysis was carried out to visualize (B) and (U) proteins to insoluble substrates.
236
residues identified in the E. cellulosolvens CBM65 differ from the CBM found in R. albus
8. The residues essential to the binding activity of EcCBM65 include W55, Q106, and
Q110. In Ra2535, a serine corresponds to W55 and a lysine corresponds to Q110,
while there is a gap in alignment corresponding to Q106 (Fig. 45A). Alignment of
EcCBM65 and the Ra2535 TM3 with similar sequences, retrieved from the Genbank
database, suggests that there are related CBMs yet uncharacterized. Binding to soluble
substrates was explored using electrophoresis with non-denaturing affinity gels infused
with soluble polysaccharides. As observed for the E. cellulosolvens CBM65, the TM3 of
R. albus 8 was able to bind tightly to xyloglucan and soluble wheat arabinoxylan and lichenin, although we observed that there was also some retardation in migration of the
BSA used as the control in this experiment. Weaker binding was identified for konjac glucomannan (KJM), and no binding was seen for galactan and laminarin (Fig. 46), while the results for binding to CMC were inconclusive.
Isothermal titration calorimetry was also conducted (Table 17) on Ra2535 TM1
and TM2. The affinity constants of Ra2535 TM1 and TM2 for cellopentaose were
similar, i.e., 1.68 × 105 M−1. The identical affinity constants of Ra2535 TM1 and TM2
reveal the presence of a single region responsible for binding to the oligosaccharide.
The number of cellopentaose molecules bound (n) during each binding event was
approximately 1 for Ra2535 TM1 and TM2, respectively, suggesting a single binding
site within each peptide for cellopentaose. As Ra2535 TM1 and TM2 had the same
affinity for cellopentaose, the comparative analysis between cello- and xylopentaose
was conducted using Ra2535 TM2 alone. The binding affinity of Ra2535 TM2 for
xylopentaose was approximately 15-fold lower than that for cellopentaose. Thus, the
237
Figure 45. Amino acid sequence alignments of the analyzed CBMs in Ra2535 and Ra1831 with homologous sequences in the NCBI. (A) The R. albus 8 CBM65 (ECG03689) aligned with its homologous sequences 4AEM, CCO04515, CDC66773, CCO04360, WP_022748873, ADZ83488, and ACZ98591 from E. cellulosolvens, Ruminococcus sp. 80/3, Ruminococcus sp. CAG:57, Ruminococcus sp. 80/3, L. bovis, C. lentocellum DSM 5427, C. ruminicola, respectively. Asterisk denotes residues shown to be essential for EcCBM65 but are not conserved in the R. albus 8 CBM65. Conserved and similar amino acids are shaded in black and gray, respectively. (B) Homologs of the Ra1831 CBM were identified through a search of the NCBI database and aligned using ClustalW. Identical amino acids are highlighted in black and similar amino acids are highlighted in gray. Accession numbers of EGC01628, EXM38945, WP_024858347, ADU21423, BAK27255, and ACZ98591 correspond to sequences from R. albus 8, R. albus SY3, R. albus, R. albus 7, Streptococcus gallolyticus, and Cellulosyliticum ruminicola, respectively. Asterisk denotes residues selected for site-directed mutagenesis of Ra1831 TM2.
238
Figure 46. Substrate binding characteristics of the CBM65 from endoglucanase Ra2535 to soluble substrates. Native affinity gel electrophoreses were performed to qualitatively analyze the binding of Ra2535 TM3 to soluble polysaccharides (Lane 2, 4, 6, 8, 10, 12, 14, and 16). Migration through a native gel without polysaccharide (Native, Lane 1 and 2) was compared to migration through native gels containing carboxymethyl-cellulose (CMC, Lane 3 and 4), wheat arabinoxylan (WAX, Lane 5 and 6), tamarind xyloglucan (Lane 7 and 8), lichenin (Lane 9 and 10), konjac glucomannan (KGM, Lane 11 and 12), galactan (Lane 13 and 14), and laminarin (Lane 15 and 16). BSA (Lane 1, 3, 5, 7, 9, 11, 13, and 15) was used as a standard for comparison.
239
Table 17. Binding parameters of Ra2535 truncational mutants for cello- and xylo- oligosaccharides determined by ITC
Relative K x103 ΔG ΔH TΔS Mutant Substrate N a affinity (M-1) (kJ) (kJ) (kJ) (%) TM1a Cellopentaose 1.28 ± 0.01 168 ± 6.00 -7.13 ± 0.23 -14.9 ± 0.23 -7.78 100
TM2b Cellopentaose 1.37 ± 0.01 168 ± 6.56 -7.12 ± 0.18 -15.3 ± 0.18 -8.24 100
Xylopentaose 1.12 ± 0.28 10.1 ± 1.35 -5.46 ± 1.35 -18.3 ± 1.35 -12.39 6
a Data from 1 replicate b Data from 3 replicates
240
ITC results were in agreement with the values obtained for the EcCBM65 (11).
Ra1831 also contains a large region with no sequence homology to known
modules in plant cell wall polysaccharide degrading enzymes. In order to determine
whether this region also binds to polysaccharides, a set of truncational mutants was
made and tested for their affinity for polysaccharides and cellopentaose. In Fig. 47A,
non-denaturing PAGE was used to determine whether the truncational mutants bind to
CMC or WAX. Here also BSA was used as a control. When compared to gels without
polysaccharides, the CMC infused gel retarded the migrations of Ra1831 TM1 and
TM2. The WAX-infused gels displayed similar results. However, these results were not
convincing since retardation was also observed for the BSA control in the presence of
the soluble substrates. To determine if the mutants are capable of binding to insoluble
cellulose, Ra1831 TM1 and TM2 were incubated with crystalline cellulose, and the
bound (B) and unbound (U) fractions were compared. In Fig. 47B, the resulting SDS-
PAGE gels are shown. Both the Ra1831 TM1 and TM2 bound to the crystalline cellulose (Fig. 47B lanes 3 and 5) revealing that the unknown region within Ra1831 binds to insoluble cellulose. To obtain quantitative information on the new CBM, ITC was performed for both truncational mutants, with cellopentaose as the substrate.
Figure 47C shows a representative ITC binding isotherm for Ra1831 TM2. The representative ITC data for Ra1831 TM1 was not very different from that of TM2 (results not shown). In each case, the binding was weak and binding parameters were not deducible from the ITC results. The ITC analysis for Ra1831 TM2 with xylopentaose suggested no binding affinity; however, the binding to cellopentaose, although weak, shows that this region contains a new CBM. The amino acid sequence alignment in Fig.
241
Figure 47. Identification of a novel CBM in Ra1831. (A) Affinity non-denaturing PAGE (6%) analysis that compared the migration of Ra1831 TM1 and TM2 in the absence ( lanes 2 and 8, respectively) or presence of carboxymethyl cellulose (CMC, lanes 4 and 10, respectively) and wheat arabinoxylan (WAX, lanes 6 and 12, respectively). BSA (lanes 1, 3, 5, 7, 9, and 11) was used as control. (B) SDS PAGE (12%) showing the unbound (U) and bound (B) fractions of Ra1831 TM1 and TM2 after incubation of 1 mg/ml of protein with 10% w/v of Avicel for one hour at 4oC with shaking. In Lane 1 are the molecular mass markers. (C) Representative binding isotherms of ITC of Ra1831 TM2 (50 µM) with cellopentaose as the substrate at 25 oC.
242
45B shows that this CBM is conserved in endoglucanases found in homologous proteins in R. albus strains and also other cellulolytic bacteria.
Mutational analysis reveal key residues involved in binding of Ra1831 TM2 to substrate. Due to the low sequence homology of Ra1831 TM2 to the characterized
CBM65, we conducted site-directed mutagenesis to determine whether substrate binding is mediated through the same amino acid residues demonstrated for EcCBM65.
The R. albus 8 Ra1831 TM2 was aligned with EcCBM65 and the aromatic residues that were conserved in the two proteins were mutated to alanine. A second alignment was constructed between Ra1831 TM2 and Ra2535 CBM65 and amino acids that were conserved were mutated to alanine in Ra1831 TM2. The mutations made included the following: W433A, Y470A, W488A, Y507A, D509A, T532A, Y542A, Y437A, W448A,
W456A, Q481A, and D482A (Fig. 45). Mutants were expressed and purified to homogeneity in parallel with the non-mutated Ra1831 TM2 (Fig. 48A). To analyze the binding capacity of the mutated derivatives of Ra1831 TM2, protein migration through non-denaturing polyacrylamide gels with and without infusion of substrates, including
WAX, glucomannan, and xyloglucan, were determined (Fig. 48B–E). The non-mutated
Ra1831 TM2 and all of the mutants migrated to a similar position in the non-denaturing gel without substrate (Fig. 48B). The migrations on non-denaturing gels infused with glucomannan (Fig. 48D) and xyloglucan (Fig. 48E) by Ra1831 TM2 and 10 of the mutants were clearly retarded. However, the mutants W448A and W488A were impaired in their capacity to bind to both substrates and also to wheat arabinoxylan (Fig.
48C). In addition, the mutants Y437A, Y470A, Q481A, D482A, D509A, and T532A also showed some defective binding to WAX (Fig. 48C).
243
Figure 48. Site directed mutagenesis of Ra1831 TM2 and binding of the mutants to soluble polysaccharides. (A) SDS PAGE (12%) of Ra1831 TM2 and its site- directed mutants purified through cobalt affinity chromatography and gel filtration. Abbreviations; Lane M, molecular mass markers. (B-E) Affinity non-denaturing PAGE (6%) comparing the migration of Ra1831 TM2 and mutants with its single amino acid mutants (2 µg) in the absence (B) or presence of soluble polysaccharides (0.2%), including wheat arabinoxylan (C), glucomannan (D), and xyloglucan (E). Abbreviations; Lane M, BSA as migration standard; Lane 1, Ra1831 TM2 WT; Lane 2, Ra1831 TM2 W433A; Lane 3, Ra1831 TM2 Y470A; Lane 4, Ra1831 TM2 W488A; Lane 5, Ra1831 TM2 Y507A; Lane 6, Ra1831 TM2 D509A; Lane 7, Ra1831 TM2 T532A; Lane 8, Ra1831 TM2 Y542A; Lane 9, Ra1831 TM2 Y437A; Lane 10, Ra1831 TM2 W448A; Lane 11, Ra1831 TM2 W456A; Lane 12, Ra1831 TM2 Q481A; Lane 13, Ra1831 TM2 D482A.
244
C.5 Discussion
R. albus is a major plant cell wall degrading bacterium residing in the rumen. To increase our understanding of the mechanism of cellulose deconstruction by this prominent ruminant microbe, genes predicted to encode endoglucanases were cloned and expressed, and their gene products screened for cellulolytic activity. The polypeptides capable of cellulose hydrolysis were further biochemically characterized.
Typical endoglucanases cleave amorphous regions within cellulose and release long chain cello-oligosaccharides (18). The endoglucanases in this study were able to release shorter products including glucose, cellobiose and cellotriose from cellulosic substrates. This degradation strategy supports the documented substrate preference for
R. albus B199, which preferentially ferments cellobiose or cellodextrins over glucose
(19). Cellobiose phosphorylase (EC 2.4.1.20) activity, where cellobiose is cleaved using phosphate instead of water, producing glucose-1-phosphate and glucose or cello- oligosaccharides (DP = n − 1), has been detected in the cytoplasm of R. albus B199.
The activity of the R. albus B199 cellobiose phosphorylase was four fold higher in cells grown on cellodextrins and cellobiose versus glucose (20). Using the cellobiose phosphorylase eliminates the use of one ATP in glycolysis, by bypassing hexokinase and harnessing phosphoglucomutase to convert glucose-1-phosphate to glucose-6- phosphate.
The putative endoglucanases of R. albus 8 exhibited diverse modular architectures. Over half of the endoglucanases contained CBMs that are located C- terminally to the catalytic domain. In contrast, Ra2461 contained a CBM that was N- terminal to the GH catalytic domain. Although the specific activity of this enzyme was
245
the lowest, kinetic parameters showed the estimated catalytic efficiency to be similar to
that of Ra0903, the endoglucanase with the highest specific activity. We postulate that
the endoglucanases characterized in the present study enhance cellulose
deconstruction through their diverse domain architectures.
To investigate the phylogenetic relationship among the GH5 domains within the
R. albus 8 endoglucanases, a phylogenetic tree was constructed using sequences from
the GH5 catalytic domains of the seven enzymes, in addition to representative
sequences from all of the GH5 subfamilies used for a similar analysis by Aspeborg et al.
(21). The enzymes from this study clustered with three different subfamilies. The GH5_1
subfamily is known to contain primarily endo-β-1,4-glucanases and clustered with the two R. albus 8 endoglucanases, Ra0903 and Ra2461. The largest reported subfamily,
GH5_2, clustered with only one R. albus 8 endoglucanase, Ra0259, which lacks a signal peptide and is not multimodular, as reported for many members of this subfamily
(21). The GH5_4 subfamily is a versatile group of enzymes that includes enzymes that cleave xyloglucan, lichenin and xylan. The R. albus 8 enzymes clustering with this family are also promiscuous in their substrate specificities, since Ra0185, Ra0325,
Ra1831, and Ra2535 exhibited the ability to hydrolyze diverse glycan substrates including xyloglucan and glucomannan (Table 18).
The endoglucanases of R. albus 8 were compared to those of other fibrolytic ruminal bacteria. A prominent cellulolytic bacterium Fibrobacter succinogenes S85 harbors many glycoside hydrolases for cellulose deconstruction. Similar to R. albus 8,
F. succinogenes S85 contains enzymes from GH5_4 and GH5_2. There are four
members of each of the two subfamilies, including Cel5H within GH5_2. Previous
246
Table 18. Screening of R. albus 8 putative family 5 endoglucanases for enzymatic activity on different glucans.
Ra0185 Ra0259 Ra0325 Ra0903 Ra1831 Ra2461 Ra2535
CMC ++a ± ++ + + + +
Glucomannan ++ - ++ ± ++ + ++
Xyloglucan ++ - ++ ± + ± ++
Enzymes were spotted on agar plate infused with polysaccharide and incubated overnight at 37°C. Activity was detected by the Congo red assay as described in the manuscript. a(-, No halo; ±, small unclear halo; +, clear halo; ++, large halo).
247
reports describe synergy between FsCel5H and other endoglucanases including Cel9B,
Cel8B, and Cel51A (22). Based on our analysis of the draft genome sequence, R. albus
8 also contains homologous genes encoding GH8 and GH51 proteins, suggesting an
analogous cellulose degradation strategies may have evolved in the two prominent
cellulolytic ruminal bacteria.
Some bacteria within the genus Prevotella and Bacteroides are well known
fibrolytic organisms in gut environments, and they function primarily in hemicellulose
degradation. The prominent human gut bacterium, Bacteroides ovatus ATCC 8483,
contains a single homolog (30% identity) of the gene encoding Ra0325. This gene is
located within a polysaccharide utilization locus (PUL) identified as a xyloglucan
degradation locus (23). Prevotella bryantii B14, from the rumen, contains two homologs
(36 and 34% identity) of Ra0325 and a single GH9 (EFI71705, EFI72295,and
EFI73198, respectively). However, these genes are not localized in a defined PUL,
unlike in B. ovatus. In both bacteria (B. ovatus and P. bryantii), the homologous GH5
enzymes lack CBMs, which likely minimizes their versatility in plant cell polysaccharide
degradation.
A well-known cellulolytic soil bacterium also belonging to the phylum Firmicutes,
Clostridium thermocellum ATCC 27405, is known for expressing a large multi-protein complex, the cellulosome. The C. thermocellum cellulosomes degrade cellulose as well as hemicellulose. The C. thermocellum homologs of the R. albus 8 GH5_4 proteins include CelE and CelH, and the homologs of the R. albus 8 GH5_1 proteins include the
C. thermocellum CelG, CelB, CelO, and a predicted endoglucanase (accession number
CAK18680). The homologous enzymes of C. thermocellum lack CBMs, but all contain
248
dockerin domains for incorporation into the cellulosome. The C. thermocellum CelO encodes a cellobiohydrolase and contains a CBM3 at the N-terminus, a modular organization similar to Ra2641, which has a CBM2 at the N-terminus (24). The GH5_1 domains of Ra2461 and CtCelO share 49% amino acid sequence identity.
Carbohydrate-binding modules are commonly found within endoglucanases, as they improve affinity of an enzyme for insoluble substrates, essentially increasing enzyme concentration on the surface of the cellulose microfibril (8,10). The lifestyle of
R. albus 8 requires this bacterium to adhere to insoluble plant material in order to hydrolyze and ferment the cellulose and hemicellulose. The extracellular enzymes used by R. albus 8 are thus expected to possess CBMs to ensure fitness in the rumen. There are a total of nine CBMs encoded within the R. albus 8 endoglucanases biochemically characterized in this study. These cellulose-binding CBMs fall within four different families (2, 4, 37, and 65). In addition, three of these families (4, 37, and 65) of CBMs are known to bind to xylan. Thus, these diverse CBMs likely ensure binding of the endoglucanases to the major polysaccharides present in plant cell walls, the main energy source for ruminants.
Due to the variety and complex arrangements of CBMs, the unidentified region within the endoglucanase Ra1831 was hypothesized to contain a new CBM. The truncational analysis that delineated the unknown regions in the polypeptides identified these modules as encoding CBMs capable of binding to cellulosic oligosaccharides.
The CBM from Ra2535 has 24% amino acid sequence identity to the recently identified member of CBM family 65. The CBM65 from E. cellulosolvens has been characterized and key amino acid residues have been identified (11). Through a comprehensive
249
screening of polysaccharide substrates, the Ra2535 CBM65 was also shown to bind to a wide range of polysaccharides including mannan and arabinan (Fig. 44). The differences in some of the key residues may account for the differences in substrate recognition between the R. albus 8 CBM65 and its homolog in E. cellulosolvens (11).
Mutational analysis revealed two residues (W448 and W488) already shown to be essential in EcCBM65 as also critical in Ra1831 TM2, hence providing evidence for a similar binding mechanism in the R. albus CBM65. In addition, we uncovered a novel
CBM (Ra1831 TM2) in this study. The new CBM does not show significant sequence homology to other characterized proteins in the NCBI database. We anticipate that as more genome sequences of carbohydrate/polysaccharide degrading organisms become available in the future, the sequence of the new CBM may be used to identify novel carbohydrate active enzymes to which it is anchored.
In a previous report, we identified and presented the contribution of several enzymes that enable R. albus 8 to completely deconstruct complex hemicellulose (25).
In the present work, we present biochemical evidence for the mechanism that is likely used by R. albus 8 to initiate depolymerization of cellulose. We look forward to characterizing the necessary enzymes that lead to further degradation of the oligosaccharides released by the R. albus 8 endoglucanases. In addition to enhancing our understanding of plant cell wall degradation by a prominent ruminal bacterium, the results from this work are also important to the prospecting of enzymes of biomass degradation for biofuel production.
250
C.6 Acknowledgments
We thank Anton F. Evans, Beth Mayer, Prashanth Singanallur, Sara
Westergaard, and Robert Beverly for technical assistance. This research was supported
in part by the Energy Biosciences Institute.
C.7 References
1 Spano, L. A., Medeiros, J. & Mandels, M. Enzymatic hydrolysis of cellulosic
wastes to glucose. Resource Recovery and Conservation 1, 279-294, (1976).
2 Mansfield, S. D., Mooney, C. & Saddler, J. N. Substrate and Enzyme
Characteristics that Limit Cellulose Hydrolysis. Biotechnology Progress 15, 804-
816, (1999).
3 Reverbel-Leroy, C., Pages, S., Belaich, A., Belaich, J. P. & Tardif, C. The
processive endocellulase CelF, a major component of the Clostridium
cellulolyticum cellulosome: purification and characterization of the recombinant
form. J. Bacteriol. 179, 46-52 (1997).
4 Gilad, R. et al. CelI, a noncellulosomal family 9 enzyme from Clostridium
thermocellum, is a processive endoglucanase that degrades crystalline cellulose.
J. Bacteriol. 185, 391-398 (2003).
5 Irwin, D. C., Spezio, M., Walker, L. P. & Wilson, D. B. Activity studies of eight
purified cellulases: Specificity, synergism, and binding domain effects.
Biotechnol. Bioeng. 42, 1002-1013, (1993).
6 Barr, B. K., Hsieh, Y. L., Ganem, B. & Wilson, D. B. Identification of two
functionally different classes of exocellulases. Biochemistry 35, 586-592, (1996).
251
7 Koeck, D. E., Pechtl, A., Zverlov, V. V. & Schwarz, W. H. Genomics of cellulolytic
bacteria. Curr. Opin. Biotechnol. 29, 171-183, (2014).
8 Herve, C. et al. Carbohydrate-binding modules promote the enzymatic
deconstruction of intact plant cell walls by targeting and proximity effects. Proc.
Natl. Acad. Sci. U. S. A. 107, 15293-15298, (2010).
9 Xu, Q. et al. A novel family of carbohydrate-binding modules identified with
Ruminococcus albus proteins. FEBS Lett. 566, 11-16, (2004).
10 Shoseyov, O., Shani, Z. & Levy, I. Carbohydrate binding modules: biochemical
properties and novel applications. Microbiol. Mol. Biol. Rev. 70, 283-295, (2006).
11 Luis, A. S. et al. Understanding how noncatalytic carbohydrate binding modules
can display specificity for xyloglucan. J. Biol. Chem. 288, 4799-4809, (2013).
12 Hess, M. et al. Metagenomic discovery of biomass-degrading genes and
genomes from cow rumen. Science 331, 463-467, (2011).
13 Weimer, P. J., Waghorn, G. C., Odt, C. L. & Mertens, D. R. Effect of diet on
populations of three species of ruminal cellulolytic bacteria in lactating dairy
cows. J. Dairy Sci. 82, 122-134, (1999).
14 Pavlostathis, S. G., Miller, T. L. & Wolin, M. J. Fermentation of insoluble cellulose
by continuous cultures of Ruminococcus albus. Appl. Environ. Microbiol. 54,
2655-2659 (1988).
15 Miron, J., Morag, E., Bayer, E. A., Lamed, R. & Ben-Ghedalia, D. An adhesion-
defective mutant of Ruminococcus albus SY3 is impaired in its capability to
degrade cellulose. J. Appl. Microbiol. 84, 249-254 (1998).
252
16 Devillard, E. et al. Ruminococcus albus 8 mutants defective in cellulose
degradation are deficient in two processive endocellulases, Cel48A and Cel9B,
both of which possess a novel modular architecture. J. Bacteriol. 186, 136-145
(2004).
17 Ezer, A. et al. Cell surface enzyme attachment is mediated by family 37
carbohydrate-binding modules, unique to Ruminococcus albus. J. Bacteriol. 190,
8220-8222, (2008).
18 Bras, J. L. et al. Structural insights into a unique cellulase fold and mechanism of
cellulose hydrolysis. Proc. Natl. Acad. Sci. U. S. A. 108, 5237-5242, (2011).
19 Thurston, B., Dawson, K. A. & Strobel, H. J. Cellobiose versus glucose utilization
by the ruminal bacterium Ruminococcus albus. Appl. Environ. Microbiol. 59,
2631-2637 (1993).
20 Lou, J., Dawson, K. A. & Strobel, H. J. Cellobiose and cellodextrin metabolism by
the ruminal bacterium Ruminococcus albus. Curr. Microbiol. 35, 221-227 (1997).
21 Aspeborg, H., Coutinho, P. M., Wang, Y., Brumer, H., 3rd & Henrissat, B.
Evolution, substrate specificity and subfamily classification of glycoside hydrolase
family 5 (GH5). BMC Evol Biol 12, 186, (2012).
22 Qi, M., Jun, H. S. & Forsberg, C. W. Cel9D, an atypical 1,4-beta-D-glucan
glucohydrolase from Fibrobacter succinogenes: characteristics, catalytic
residues, and synergistic interactions with other cellulases. J Bacteriol 190, 1976-
1984, (2008).
23 Larsbrink, J. et al. A discrete genetic locus confers xyloglucan metabolism in
select human gut Bacteroidetes. Nature 506, 498-502, (2014).
253
24 Zverlov, V. V., Velikodvorskaya, G. A. & Schwarz, W. H. A newly described
cellulosomal cellobiohydrolase, CelO, from Clostridium thermocellum:
investigation of the exo-mode of hydrolysis, and binding capacity to crystalline
cellulose. Microbiology 148, 247-255 (2002).
25 Moon, Y. H., Iakiviak, M., Bauer, S., Mackie, R. I. & Cann, I. K. Biochemical
analyses of multiple endoxylanases from the rumen bacterium Ruminococcus
albus 8 and their synergistic activities with accessory hemicellulose-degrading
enzymes. Appl. Environ. Microbiol. 77, 5157-5169, (2011).
26 Laemmli, U. K. Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature 227, 680-685 (1970).
27 Dodd, D., Kiyonari, S., Mackie, R. I. & Cann, I. K. Functional diversity of four
glycoside hydrolase family 3 enzymes from the rumen bacterium Prevotella
bryantii B14. J. Bacteriol. 192, 2335-2345, (2010).
28 Wood, T. M. Preparation of Crystalline, Amorphous, and Dyed Cellulase
Substrates. Method. Enzymol. 160, 19-25 (1988).
29 Lever, M. A new reaction for colorimetric determination of carbohydrates. Anal.
Biochem. 47, 273-279, (1972).
30 Ye, L. et al. Molecular and biochemical analyses of the GH44 module of
CbMan5B/Cel44A, a bifunctional enzyme from the hyperthermophilic bacterium
Caldicellulosiruptor bescii. Appl. Environ. Microbiol. 78, 7048-7059, (2012).
31 Su, X., Mackie, R. I. & Cann, I. K. Biochemical and mutational analyses of a
multidomain cellulase/mannanase from Caldicellulosiruptor bescii. Appl. Environ.
Microbiol. 78, 2230-2240, (2012).
254