The Predatory Bacterium Bdellovibrio bacteriovorus Aspartyl-tRNA Synthetase Recognizes tRNAAsn as a Substrate

The predatory bacterium Bdellovibrio bacteriovorus preys on other Gram-negative bacteria and was predicted to be an asparagine auxotroph. However, despite encoding asparaginyl-tRNA synthetase and glutaminyl-tRNA synthetase, B. bacteriovorus also contains the amidotransferase GatCAB. Deinococcus radiodurans, and Thermus thermophilus also encode both of these aminoacyl-tRNA synthetases with GatCAB. Both also code for a second aspartyl-tRNA synthetase and use the additional aspartyl-tRNA synthetase with GatCAB to synthesize asparagine on tRNAAsn. Unlike those two bacteria, B. bacteriovorus encodes only one aspartyl-tRNA synthetase. Here we demonstrate the lone B. bacteriovorus aspartyl-tRNA synthetase catalyzes aspartyl-tRNAAsn formation that GatCAB can then amidate to asparaginyl-tRNAAsn. This non-discriminating aspartyl-tRNA synthetase with GatCAB thus provides B. bacteriovorus a second route for Asn-tRNAAsn formation with the asparagine synthesized in a tRNA-dependent manner. Thus, in contrast to a previous prediction, B. bacteriovorus codes for a biosynthetic route for asparagine. Analysis of bacterial genomes suggests a significant number of other bacteria may also code for both routes for Asn-tRNAAsn synthesis with only a limited number encoding a second aspartyl-tRNA synthetase.


Introduction
Bdellovibrio bacteriovorus HD100 preys on other Gramnegative bacteria by inserting into the host's periplasm where B. bacteriovorus grows and replicates, taking advantage of the nutrient rich environment of the host cell [1]. Because this predatory process kills the host, B. bacteriovorus is being studied as a living antibiotic for therapeutic, agriculture, and waste treatment purposes [2][3][4][5][6]. Based on its genome, the bacterium was predicted to be missing biosynthetic pathways for nine of the proteinogenic amino acids including Asn, likely making B. bacteriovorus protein synthesis dependent on host degradation products [7].
However, B. bacteriovorus codes for only one AspRS [7]. We therefore predicted the lone B. bacteriovorus AspRS is nondiscriminating in order to facilitate GatCAB synthesis of Asn on tRNA Asn . We demonstrate that the B. bacteriovorus AspRS can readily form Asp-tRNA Asn as the first step in tRNA-dependent Asn biosynthesis. By analyzing bacterial genomes, we found a significant number of bacteria may also encode both routes for Asn-tRNA Asn synthesis including additional species with a second AspRS. However, only a limited number of bacteria encode AsnRS, GlnRS, GatCAB, and only one AspRS but neither Asn synthetase like B. bacteriovorus. Over-production and purification aaRSs B. bacteriovorus aspS (Bd3311) was cloned between the NdeI and BamHI restriction sites in pET28a to be N-terminally His 6tagged. B. bacteriovorus AspRS was overproduced using the autoinduction method [23] and purified by nickel-affinity chromatography in the same manner as the S. aureus AspRS following manufacturer's protocols (Qiagen) [24]. The purified enzyme was dialyzed, concentrated, and stored as described [24]. The enzyme preparation was determined.95% pure by Coomassie-stained polyacrylamide gel [24]. The Legionella pneumophila aspS was chemically synthesized (Life Technologies, GeneArt) and then subcloned between the NdeI and BamHI sites in pET28a and overproduced and purified as described for the S. aureus AspRS [24].

General
The B. bacteriovorus asnC (Bd1054) was chemically synthesized with optimized codons for overproduction in E. coli (Life Technologies GeneArt). The optimization increased the number of codons, 52% to 90%, in the GeneArt's top codon class (90-100) based on frequency of codon usage in E. coli [25]. The gene was then subcloned into pET28a between the NdeI and BamHI sites to be N-terminally His 6 -tagged. AsnRS was over-produced as described previously for the S. aureus homolog [24] using the autoinduction method [23] and purified by nickel-affinity chromatography following manufacturer's protocols (Qiagen) with a buffer of 50 mM Tris-HCl, pH 7.6 with 10 mM MgCl 2 and 300 mM NaCl. The purified AsnRS was dialyzed in 50 mM Tris-HCl, pH 7.6 with 10 mM MgCl 2 , 30 mM NaCl, and 50% glycerol, and then concentrated and stored as described [24].
In vitro transcription, tRNA folding, and 32 P labeling The tRNA genes were in vitro transcribed and the resultant tRNA was purified by chromatography as described [22]. The tRNAs were heated to 95uC for 5 min and slowly cooled to room temperature to refold with MgCl 2 added to a final concentration of 5 mM at 65uC. Samples were stored at 220uC and 32 P-labeled as described previously using the E. coli CCA-adding enzyme [8].
Methanothermobacter thermautotrophicus tRNA Gln was in vitro transcribed, purified, and folded as described previously [26] and 32 P-labeled as described previously using the E. coli CCA-adding enzyme [8].

P-based tRNA aminoacylation assay
The aminoacylation activities of the aaRSs were monitored using the established 32 P-based assay [8,[27][28][29][30]. The AspRS reactions contained 50 mM HEPES-KOH, pH 7.2, 30 mM KCl, 15 mM MgCl 2 , 5 mM DTT, 4 mM L-Asp, and 4 mM ATP. The AsnRS reactions contained 50 mM HEPES-KOH, pH 7.5, 30 mM KCl, 15 mM MgCl 2 , 5 mM DTT, 4 mM L-Asn, and 4 mM ATP. For plateau aminoacylation of tRNA, reactions were carried out at 37uC with 1.0 mM 32 P-labeled tRNA, 11.0 mM tRNA, and 3.0 mM enzyme. Steady-state kinetic studies with 5 nM AspRS were carried out at 37uC with 0.055-1.0 mM 32 Plabeled tRNA, and 0-10.0 mM tRNA over 6 min. Steady-state kinetic studies with 5 nM AsnRS were carried out at 37uC with 0.055-1.0 mM 32 P-labeled tRNA Asn , and 0-12.0 mM tRNA Asn over 6 min. Reaction mixtures and enzymes were pre-incubated for 30 sec at 37uC. Reactions were started by the addition of enzyme and repeated three to four times. Time points were quenched, digested, separated by TLC, processed and analyzed as described previously [8,24,29,30]. The activity of the L. pneumophila AspRS was measured in the presence of 0.1 mM 32 P-labelled tRNA, 50 mM HEPES-KOH, pH 7.2, 30 mM KCl, 15 mM MgCl 2 , 5 mM DTT, 4 mM L-Asp, and 4 mM ATP over 5 minutes. Reactions were started by the addition of L. pneumophila AspRS to a final concentration of 10 nM at 37uC. Time points were quenched, digested, separated by TLC, processed and analyzed as described previously [8,29].

E. coli trpA34 in vivo assay
The B. bacteriovorus aspS was cloned into pCBS2 between the NdeI and BglII restriction sites (pCBS2-Bb-aspS) [31]. In a similar fashion the L. pneumophila aspS was subcloned into pCBS2 (pCBS2-Lp-aspS). Following transformation into E. coli trpA34 cells, the cultures were grown and assayed as described previously on M9 minimal media agar plates with or without Trp with minor adjustments [24]. Briefly, cultures were grown overnight at 37uC in LB in the presence of ampicillin (100 mg/ml). The overnight culture was used to inoculate 5 mL of M9 minimal media supplemented with ampicillin (100 mg/ml) and all twenty amino acids (20 mg/ml each). The cultures were then grown shaking at 37uC for four hours. The samples were adjusted to the same O.D. 600 by diluting with M9 minimal media before 5 mL of adjusted culture was spun down at 1,5006g for 5 min, and washed three times with M9 minimal media supplemented with ampicillin (100 mg/ml). After washing, the samples were resuspended in 0.2 mL of M9 minimal media supplemented with ampicillin (100 mg/ml) before spotting 2 mL of culture on M9 minimal media agar ampicillin (100 mg/ml) plates with or without L-Trp (20 mg/ ml), and supplemented with the other 19 amino acids (20 mg/ml each).

E. coli JF448 in vivo assay
The B. bacteriovorus aspS and gatCAB (operon of Bd0058, Bd0059, Bd0060) were fused into an artificial operon as described previously [31]. The artificial operon was subcloned into the pCBS2 plasmid between the NdeI and BglII restriction sites (pCBS2-Bb-aspS-gatCAB) and transformed into E. coli JF448 cells. The cells were grown and assayed as described previously on M9 minimal media agar plates with or without Asn [24]. Briefly, cultures were grown overnight at 37uC in LB in the presence of ampicillin (100 mg/ml). The overnight culture was used to inoculate 5 mL of M9 minimal media supplemented with ampicillin (100 mg/ml) and all twenty amino acids (20 mg/ml each). The cultures were then grown shaking at 37uC for four hours before being spun down at 1,5006g for 5 min, and washed three times with M9 minimal media supplemented with ampicillin (100 mg/ml). After washing, the samples were resuspended in 1 mL of M9 minimal media supplemented with ampicillin (100 mg/ml) and then diluted to an O.D. 600 of 0.45. The samples were then diluted 100-fold in M9 minimal media before spotting 2 mL of culture on M9 minimal media agar ampicillin (100 mg/ml) plates with or without L-Asn (20 mg/ml), and supplemented with the other 19 amino acids (20 mg/ml each).

Bioinformatic survey of bacterial genomes
Bacterial genomes representing 547 different genera were analyzed for genes encoding AsnRS (asnS), GlnRS (glnS), GatCAB (gatC, gatA, gatB), AsnA (asnA), AsnB (asnB), and AspRS (aspS). Genes were searched either in the UniProt (http:// www.uniprot.org) or KEGG: Kyoto Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/) databases. Sequences were then compared to known relevant enzymes by BLAST to validate the presence of the relevant active sites and domain architecture. The E. coli CFT AsnRS (AAN79540), GlnRS (AAN79239), AsnA (AAN83104), and AsnB (AAN79222), the H. pylori J99 GatA (AAD06348) and GatB (AAD06184), and the Deinococcus radiodurans discriminating AspRS (AAF10918) and ND-AspRS (AAF10623) sequences were used for the analysis. This was of particular importance to distinguish AsnA, which lacks an anticodon-binding domain, from its orthologs, AspRS and AsnRS [32]. Bacterial-type AspRS sequences were distinguished from archaeal-type AspRS sequences by the presence of a GAD insertion domain specific to bacterial AspRSs [33]. When no gene was initially identified for an enzyme, a tBLASTn search was performed with an enzyme sequence from a related organism. In addition, the tRNA Asn isoacceptors of these bacteria were analyzed for the presence of a U1-A72 base pair. The tRNA isoacceptors sequences studied were either from the Genomic tRNA Database (http://lowelab.ucsc.edu/GtRNAdb/) or the KEGG database (http://www.genome.jp/kegg/). The results of the survey are detailed in Table S1.

In vitro aminoacylation of tRNA Asp and tRNA Asn
Bacterial GatCAB recognizes the U1-A72 base pair present in many bacterial tRNA Asn isoacceptors [34,35]. B. bacteriovorus tRNA Asn has a U1-A72 base pair (Table S1) meaning if aspartylated, the tRNA could serve as a substrate for GatCAB. However, the presence of tRNA Asn with a U1-A72 base pair and GatCAB does not necessarily mean B. bacteriovorus encodes the two-step pathway for Asn-tRNA Asn formation. For example, Lactobacillus delbruekii bulgaricus encodes tRNA Asn with a U1-A72 pair along with GatCAB (Table S1) but does not synthesize Asn on tRNA Asn as it lacks a ND-AspRS and uses only AsnRS to form Asn-tRNA Asn [36].  For B. bacteriovorus to encode the two-step pathway for Asn-tRNA Asn formation, the organism must code for a ND-AspRS along with GatCAB and tRNA Asn with a U1-A72 base pair. Given the presence of GatCAB in B. bacteriovorus despite encoding GlnRS and AsnRS, and the absence of both asparagine synthetases (AsnA and AsnB) to synthesize Asn [7], we predicted the lone B. bacteriovorus AspRS aspartylates tRNA Asn to enable the bacterium to synthesize Asn in a tRNA-dependent manner.
To determine if the B. bacteriovorus AspRS is non-discriminating, we overproduced the enzyme in E. coli and purified it to homogeneity [24]. The recombinant enzyme was readily able to aspartylate both tRNA Asp and tRNA Asn to similar levels ( Figure 1A). Discriminating AspRS enzymes typically prefer tRNA Asp to tRNA Asn by a factor of 500-2,250 [22,37]. In contrast and similar to other ND-AspRS enzymes [22,24,37,38], the B. bacteriovorus AspRS preferred tRNA Asp as a substrate by only 3fold ( Table 1). The difference in catalytic efficiency by AspRS was attributed to an increased k cat with tRNA Asp as a substrate ( Table 1). The B. bacteriovorus AsnRS also readily uses tRNA Asn as a substrate, reaching a similar aminoacylation plateau ( Figure 1B). The tRNA Asn was a better substrate for AsnRS by 3-fold with a higher k cat compensating for an increased K M relative to AspRS (Table 1).

B. bacteriovorus aspS rescues E. coli Trp auxotroph
To establish whether B. bacteriovorus AspRS also uses tRNA Asn as a substrate in a cellular context where there exists competition from other aaRSs and modified tRNA isoacceptors, we used the established E. coli trpA34 complementation assay [24,31,39]. Tryptophan synthetase alpha subunit (TrpA) is required for Trp synthesis in E. coli. The trpA34 strain is a Trp auxotroph due to mutation of codon 60 from an essential Asp codon to an Asn codon [40]. Production of a ND-AspRS in the strain rescues the phenotype, because the missense suppressor Asp-tRNA Asn formed by the ND-AspRS allows decoding of the mutant Asn codon with Asp and production of active TrpA [24,31,39]. Consistent with our in vitro results demonstrating the B. bacteriovorus AspRS readily uses tRNA Asn as a substrate, the trpA34 strain with the B. bacteriovorus aspS was able to grow in the absence of Trp (Figure 2).

B. bacteriovorus aspS with gatCAB rescues E. coli Asn auxotroph
We predicted B. bacteriovorus encodes a ND-AspRS so the bacterium could synthesize Asn in a tRNA-dependent manner using GatCAB. Bacterial GatCABs readily amidate Asp-tRNA Asn to Asn-tRNA Asn in vitro [8,12,[15][16][17]21]. To verify co-production of B. bacteriovorus AspRS and GatCAB in vivo leads to Asn synthesis, we used the established E. coli JF448 system [22,41]. The JF448 strain is an Asn auxotroph due to mutation of both Asn synthetase genes in E. coli [41] and the phenotype can be rescued by introducing the tRNA-dependent route for Asn biosynthesis [22,24]. Consistent with our hypothesis, co-production of the B. bacteriovorus AspRS and GatCAB enabled the JF448 strain to grow in the absence of Asn in the media (Figure 3).

Bioinformatic analysis
To determine how common it is for a bacterium to encode AsnRS, GlnRS, and GatCAB, we surveyed genomes from 547 different bacterial genera (Table S1). The three enzymes are encoded together in 68 different genera (Table 2). Like B. bacteriovorus, only 18 genera coded for all three while not encoding an Asn synthetase in their genomes. They represent a diverse range of bacteria from the d-proteobacteria, the Deinococcus-Thermus, Bacteroidetes, and Verrucomicrobiae clades. All these bacteria have a tRNA Asn with a U1-A72 base pair required for recognition by bacterial GatCAB, consistent with these bacteria possibly synthesizing Asn on tRNA Asn . A second AspRS is encoded in five of the 18 genomes, all from the Deinococcus-Thermus phylum (Tables S1 and S2). This second AspRS in this clade is of the archaeal type and may stabilize GatCAB at higher growth temperatures [42]. Similar to B. bacteriovorus, only 13 bacterial genera encode GlnRS, AsnRS, GatCAB and one AspRS but neither Asn synthetase (Table S1).
The c-proteobacteria L. pneumophila belongs to the group encoding GlnRS, AsnRS, one AspRS, and GatCAB but neither Asn synthetase. The lack of an Asn synthetase suggested its AspRS recognizes tRNA Asn as the first step in tRNA-dependent Asn biosynthesis as we hypothesized for B. bacteriovorus. We therefore tested whether the L. pneumophila AspRS could use tRNA Asn as a substrate both in vivo (Fig. 4A), using the E. coli trpA34 assay, and in vitro (Fig. 4B). Like the B. bacteriovorus AspRS, the L. pneumophila was able to aspartylate tRNA Asn suggesting this cproteobacteria potentially encodes the two-step pathway for Asn-tRNA Asn formation. The L. pneumophila AspRS has about a 1.6-  fold preference for tRNA Asp over tRNA Asn , similar to other ND-AspRS enzymes [22,24,37,38].
Beyond the Deinococcus-Thermus phylum, an additional 15 bacterial genera encode an extra AspRS (Table S2). C. acetobutylicum is unique in encoding three AspRSs and an additional GatCAB [18]. All 15 genomes encoded GatCAB and a tRNA Asn with a U1-A72 base pair. The majority of the bacteria in this group (10 out of 15) lack an AsnRS and the indirect route with GatCAB is the only means for Asn-tRNA Asn synthesis. Primarily, these bacteria are Actinobacteria in the order Actinomycetales. Of those with GatCAB and AsnRS, four are Firmicutes in the Clostridiaceae family and all encode at least one Asn synthetase. However, in the case of C. acetobutylicum, the AsnB is split into two halves and does not appear to be functional under normal physiological conditions with Asn synthesis being tRNAdependent [18]. One Actinomycetales, Amycolatopsis mediterranei RB, also encoded an additional AspRS with AsnRS and GatCAB. It also in its genome has four asnB genes encoding the glutaminedependent Asn synthetase (AsnB).
We also examined how many other bacteria potentially use both Asn-tRNA Asn biosynthetic pathways. AsnRS and GatCAB are encoded in 174 bacterial genera with 163 also coding for a tRNA Asn with a U1-A72 base pair (Table 2). In 46 of these bacteria, Asn synthesis could only be tRNA-dependent as they lack AsnA and AsnB, as was found in Staphylococcus aureus [24]. The other 117 genera encode at least one Asn synthetase. Of these, 49 also have a GlnRS suggesting GatCAB may be used for tRNAdependent Asn synthesis including other d-proteobacteria. Bacteriovorax marinus was an exception among the d-proteobacteria as it coded for AsnRS and GlnRS but not GatCAB. It is also possible that instead the GatCAB is for Gln-tRNA Gln formation but to date no bacteria are known to encode both routes for Gln-tRNA Gln synthesis. The other 68 genera lack a GlnRS and retain GatCAB for Gln-tRNA Gln though that does not exclude GatCAB from also being used for Asn-tRNA Asn formation in these bacteria. As noted previously, AsnRS is present in all prokaryotes with AsnA, the ammonia-dependent Asn synthetase, [8,32].

Discussion
The B. bacteriovorus AspRS is non-discriminating readily able to form Asp-tRNA Asn . The ND-AspRS may provide the bacterium the ability synthesize Asn in a tRNA-dependent manner using GatCAB, providing a potential functional role for the amidotransferase in an organism with both AsnRS and GlnRS [19]. Thus, unlike the initial prediction after the B. bacteriovorus genome was sequenced [7], the bacterium does potentially encode an Asn biosynthetic pathway. Also, B. bacteriovorus with a ND-AspRS, GatCAB, and AsnRS could encode both routes for Asn-tRNA Asn formation in addition to the direct route for Gln-tRNA Gln synthesis similar to Deinococcus radiodurans and Thermus thermophilus [10,12,18,[20][21][22]. However, unlike those Representative genomes from 547 different bacterial genera were analyzed for the presence of genes coding for AsnA, AsnB, AsnRS, GlnRS, and GatCAB. The results are detailed in Table S1. In parentheses is the number of bacterial genera with a tRNA Asn isoacceptor containing a U1-A72 base pair. doi:10.1371/journal.pone.0110842.t002 two species that acquired an additional AspRS from archaea [10,12,[20][21][22], B. bacteriovorus has only one AspRS. Acquisition of an additional AspRS in bacteria is rare. The second AspRS may provide an advantage in particular environmental niches. In the case of T. thermophilus, GatCAB binding to the archaeal-type ND-AspRS stabilizes the amidotransferase at elevated temperatures and the second AspRS may be an adaptation to a thermophilic environment [42]. The second AspRS in other thermophiles may also serve the same purpose in addition to allowing the bacteria to synthesize Asn in a tRNAdependent manner. For C. acetobutylicum, the additional AspRS enzymes and GatCAB have been linked to the organism switching from acidogenesis to solventogenesis [18,43].
Why B. bacteriovorus may still retain the tRNA-dependent route for Asn production despite coding for AsnRS is not clear. The organism is unable to synthesize eight other proteinogenic amino acids [7]. During growth away from a host cell, basal protein synthesis in B. bacteriovorus with those eight amino acids requires recycling them from protein degradation [7]. Recycling Asn residues may be problematic as Asn residues in polypeptides are susceptible to deamidation [44,45]. ND-AspRS and GatCAB could provide B. bacteriovorus the means to compensate for deamidation of Asn residues by synthesizing Asn from Asp on tRNA Asn . The route would also provide direct coupling of Asn synthesis for use in translation [9]. In addition, the indirect pathway would allow B. bacteriovorus to convert Asp to Asn from hosts that underutilize Asn. Conversely, B. bacteriovorus may retain AsnRS to take advantage of the Asn provided by a host as well as to efficiently recycle non-deamidated Asn residues following protein degradation as has been hypothesized in T. thermophilus, D. radiodurans, and S. aureus [10,12,22,24].
It is also possible one of the B. bacteriovorus Asn-tRNA Asn pathways is non-functional. However, it should be noted asnS expression increased significantly during B. bacteriovorus growth in a host cell like the other aaRS genes including AspRS with only minimal increased expression of the GatCAB genes [46]. The increased expression of the aaRS genes during the growth phase was after initial predation by B. bacteriovorus as expression levels were unchanged 30 min. after host infection like the GatCAB genes [47]. Interestingly, host-independent B. bacteriovorus cultures grown in the presence of peptone and tryptone exhibited increases in the expression of not only aspS and asnS but also the genes for GatCAB [47]. The available gene expression results are consistent with a role for GatCAB in B. bacteriovorus when freeliving and AsnRS during growth when Asn is present either from a host or in the media. Such a scenario would be similar to what is hypothesized in S. aureus, that encodes a ND-AspRS, GatCAB, and AsnRS but lacks a GlnRS and either Asn synthetase [24]. The S. aureus GatCAB is predicted for Gln-tRNA Gln formation and tRNA-dependent Asn biosynthesis, while AsnRS is predicted for growth in Asn-rich environments like the human body [24].
Interestingly, L. pneumophila like B. bacteriovorus is capable of free-living and growth in a host cell [48]. The presence of a ND-AspRS along with GlnRS, AsnRS and GatCAB in L. pneumophila raises the possibility it may also encode both routes for Asn-tRNA Asn formation. In L. pneumophila, the genes for AsnRS, ND-AspRS, and GatCAB are all up regulated during post-exponential and transmissive growth phases, suggesting a role in the organism's life cycle [49].
Two distinct routes for Asn-tRNA Asn formation in bacteria like B. bacteriovorus were likely acquired in a stepwise manner. Both AspRS and GatCAB were likely present in the last universal common ancestor (LUCA) while AsnRS evolved from a duplication of an archaeal AspRS [32,50,51]. Therefore, it has been hypothesized that Asn-tRNA Asn formation in LUCA was via the indirect pathway using a ND-AspRS and GatCAB [50,51]. Under this scenario, early bacteria likely also used a ND-AspRS and GatCAB for Asn-tRNA Asn formation [51]. Consistent with that hypothesis, the phylogenies of the GatCAB subunits and bacterialtype AspRSs suggest they were vertically inherited in bacteria [51,52]. AsnRS was later likely acquired in different bacterial lineages via horizontal gene transfer from archaea [32,50,52].
In some bacterial species like L. delbruekii bulgaricus, the presence of AsnRS to aminoacylate tRNA Asn appears to have lessened the selective pressure for AspRS to recognize tRNA Asn as a substrate, facilitating the AspRS to evolve specicity for just tRNA Asp [36]. Accordingly, the role of the L. delbruekii bulgaricus GatCAB is for only Gln-tRNA Gln synthesis [36]. In B. bacteriovorus, AspRS retained its relaxed tRNA specificity along with GatCAB after acquisition of AsnRS possibly for tRNAdependent Asn biosynthesis as was found in S. aureus [24]. Given the above and the presence of GatCAB and AsnRS in most dproteobacteria, the ancestral d-proteobacteria likely encoded GatCAB for Asn-tRNA Asn formation before acquiring AsnRS to directly attach the Asn to tRNA Asn . As many d-proteobacteria, like the predatory bacterium Myxococcus xanthus [53] and the metalreducing anaerobic Geobacter metallireducens [54], encode one or more asparagine synthetase (AsnA and or AsnB) along with AsnRS and GlnRS, the retention of GatCAB in these bacteria maybe vestigial. However, it may be beneficial for these bacteria to retain both routes for Asn-tRNA Asn formation as hypothesized for B. bacteriovorus, L. pneumophila, and S. aureus [24].
It is unclear how many other bacteria encode both routes for Asn-tRNA Asn formation. Based on our genomic analysis, about 30% of bacterial genera surveyed could code for both routes as they encode GatCAB, AsnRS, and tRNA Asn isoacceptors with a U1-A72 base pair. In the few bacteria like B. bacteriovorus that also encode GlnRS but neither Asn synthetase, it is likely GatCAB was retained for tRNA-dependent Asn production. Similarly, those bacteria like S. aureus without a GlnRS and both Asn synthetases may also use GatCAB for tRNA-dependent Asn biosynthesis in addition to Gln-tRNA Gln formation [24]. However, the majority of bacteria that encode AsnRS and GatCAB also code for at least one Asn synthetase (AsnA and/or AsnB); in total, 21% of all bacterial genera analyzed. Since these organisms have alternate means to synthesize Asn, GatCAB may not be for Asn-tRNA Asn formation as their AspRS enzymes might be specific for tRNA Asp . For example, L. delbruekii bulgaricus AspRS does not aspartylate tRNA Asn and uses GatCAB only for Gln-tRNA Gln formation [36]. Clarifying the tRNA specificity of the AspRSs in other bacteria encoding both AsnRS and GatCAB would establish how many other bacteria code for both routes for Asn-tRNA Asn formation, providing the foundation to better understand how Asn metabolism and protein synthesis are integrated into bacterial physiology and adaptation to certain environmental niches.

Supporting Information
Table S1 Genomic analysis of bacterial genomes from 547 different genera for genes related to Asn-tRNA Asn , Gln-tRNA Gln , and Asn synthesis.

(XLSX)
Table S2 Genomic analysis of bacterial genomes encoding at least two AspRS enzymes for genes related to Asn-tRNA Asn , Gln-tRNA Gln , and Asn synthesis. (XLSX)