The putative zinc-binding domain (pZBD) in Escherichia coli glutamyl-tRNA synthetase (GluRS) is known to correctly position the tRNA acceptor arm and modulate the amino acid-binding site. However, its functional role in other bacterial species is not clear since many bacterial GluRSs lack a zinc-binding motif in the pZBD. From experimental studies on pZBD-swapped E. coli GluRS, with Thermosynechoccus elongatus GluRS, Burkholderia thailandensis GluRS and E. coli glutamyl-queuosine-tRNAAsp synthetase (Glu-Q-RS), we show that E. coli GluRS, containing the zinc-free pZBD of B. thailandensis, is as functional as the zinc-bound wild-type E. coli GluRS, whereas the other constructs, all zinc-bound, show impaired function. A pZBD-tinkered version of E. coli GluRS that still retained Zn-binding capacity, also showed reduced activity. This suggests that zinc is not essential for the pZBD to be functional. From extensive structural and sequence analyses from whole genome database of bacterial GluRS, we further show that in addition to many bacterial GluRS lacking a zinc-binding motif, the pZBD is actually deleted in some bacteria, all containing either glutaminyl-tRNA synthetase (GlnRS) or a second copy of GluRS (GluRS2). Correlation between the absence of pZBD and the occurrence of glutamine amidotransferase CAB (GatCAB) in the genome suggests that the primordial role of the pZBD was to facilitate transamidation of misacylated Glu-tRNAGln via interaction with GatCAB, whereas its role in tRNAGlu interaction may be a consequence of the presence of pZBD.

INTRODUCTION

Zinc plays an important structural and functional role in a number of aminoacyl-tRNA synthetases (aaRSs). For example, zinc is intimately related with cognate ligand recognition in Escherichia coli threonyl-tRNA synthetase [1] and cysteinyl-tRNA synthetase [2,3]. Similarly, a retrovirus like zinc-coordinating motif is important in tRNA recognition in E. coli alanyl-tRNA synthetase [4]. Chemical or mutational modification of the zinc-coordinating ligands in isoleucyl-tRNA synthetase (of E. coli and Thermus thermophlilus) affects the enzyme efficiency [5,6]. The crystal structure of Methanosarcina bakeri seryl-tRNA synthetase contains a Zn2+ ion in its active site; mutations of the zinc-coordinating ligands lead to the inactivation of the enzyme [7]. Removal of Zn2+ from T. thermophlilus methionyl-tRNA synthetase from many bacteria results in substantial loss of activity of the enzyme, although the zinc-binding motif (ZB-motif) lies far away from the active site [8].

The presence of a Zn2+ in glutamyl-tRNA synthetase (GluRS) was first identified in the bacterium E. coli (Ec-GluRS) by X-ray fluorescence spectroscopy [9]. The removal of GluRS-bound Zn2+ reduced enzyme activity accompanied by a conformational change. Depletion of Zn2+ was inhibited in the presence of ATP, suggesting a close relationship between Zn2+ and the ATP-binding pocket. Partial proteolysis of Ec-GluRS followed by atomic absorption spectroscopic studies revealed the presence of Zn2+ in the N-terminal region of Ec-GluRS [9]. Extensive proteolytic analyses also showed that the ZB-motif is present in the fragment 98–138 of Ec-GluRS whereas Extended X-Ray Absorption Fine Structure data indicated that the co-ordination sphere of Zn2+ in Ec-GluRS consists of three sulfur and one nitrogen atoms [10]. Based on cysteine and histidine mutants of Ec-GluRS, present in the residue stretch 98-138, it was suggested that residues C98, C100, C125 and H127 coordinate Zn2+ ion in Ec-GluRS [10]. The pZBD of Ec-GluRS, containing the ZB-motif 98CxCx24CxH, belongs to the SWIM domain family that binds to DNA or protein [11,12]. An earlier study showed that the C100Y mutant of Ec-GluRS exhibits decreased glutamylation efficiency along with a reduced L-glutamic acid binding in presence of tRNAGlu (L-glutamic acid binding remained unaffected in absence of tRNAGlu) without causing significant changes in the protein structure [12]. Close interaction of the putative zinc-binding domain (pZBD) with tRNAGlu is also reflected in the tRNAGlu-bound structure of T. thermophilus GluRS (Tt-GluRS; Figure 1) [13].

The pZBD in T. thermophilus GluRS and its interaction with tRNAGlu

Figure 1
The pZBD in T. thermophilus GluRS and its interaction with tRNAGlu

(A) tRNAGlu-bound structure of T. thermophilus GluRS (pdb ID: 2dxi). (B) The pZBD (residues 101–150), defined by two helices H4 and H5 and capped by two β-strands E3 and E4 [orientations of pZBD in panels (A and B) are different for clarity]. Key H-bond interactions are highlighted: between tRNAGlu (Ade73 and Cyt74) and pZBD residues (i, ii and iii), between tRNAGlu and pZBD Ser181 (iv and v) and between Tyr122 and Ala126 (vi).

Figure 1
The pZBD in T. thermophilus GluRS and its interaction with tRNAGlu

(A) tRNAGlu-bound structure of T. thermophilus GluRS (pdb ID: 2dxi). (B) The pZBD (residues 101–150), defined by two helices H4 and H5 and capped by two β-strands E3 and E4 [orientations of pZBD in panels (A and B) are different for clarity]. Key H-bond interactions are highlighted: between tRNAGlu (Ade73 and Cyt74) and pZBD residues (i, ii and iii), between tRNAGlu and pZBD Ser181 (iv and v) and between Tyr122 and Ala126 (vi).

Despite demonstrated structural and functional importance of the bound Zn2+ ion in Ec-GluRS, sequence analysis on a small set of bacterial GluRS showed that the putative ZB-motif CxCx24CxH (identified in E. coli GluRS) is absent from many bacteria, especially the third and fourth co-ordinating ligands [12]. This is corroborated from known crystal structures of bacterial GluRS, where Zn2+ is present in only one, GluRS from Borrelia burgdorferri (Bb-GluRS), out of six available crystal structures (Table 1), is zinc-bound. The role played by Zn2+ ion in bacterial GluRS therefore is poorly understood. Although it was proposed that the pZBD in different bacterial GluRS might have different local conformation and affinity for metal ions other than zinc [12], the proposal has not been examined in detail. We have experimentally studied a number of Ec-GluRS pZBD-chimeric constructs, with and without a ZB-motif, to understand the functional role of Zn2+. In addition, as part of our ongoing work on sequence analysis of a large database of bacterial GluRS from whole genome sequences [14], we have examined the occurrence of pZBD and ZB-motifs in GluRS across different bacterial phyla in conjunction with analyses of available GluRS structures. We show that a number of extant bacterial GluRS lack a pZBD. Our results imply a broader understanding of the functional role of the pZBD with or without bound Zn2+.

Table 1
GluRS and Glu-Q-RS structures in protein data bank
Protein Organism PDB ID Status of zinc zinc-binding motif 
Bacterial GluRS B. burgdorferri 4gri Present CXCX20YX3
 T. thermophilus 2cuz, 2cvo, 2cv1, 2cv2, 1j09, 1n75, 1n77, 1n78, 2dxi Absent — 
 T. elongatus 2cfo Absent CXCX20YX3
 M. tuberculosis 2ja2 Absent — 
 B. thailandensis 4g6z Absent — 
Bacterial GluRS1 T. maritima (TM1351) 3afh Absent — 
Bacterial GluRS2 T. maritima (TM1875) 2o5r Absent — 
Archaeal GluRS M. thermautotrophicus 3aii Present CXCX14CXC 
Bacterial Glu-Q-RS E. coli 1nzj, 4a91 Present CXCX11YX3
Protein Organism PDB ID Status of zinc zinc-binding motif 
Bacterial GluRS B. burgdorferri 4gri Present CXCX20YX3
 T. thermophilus 2cuz, 2cvo, 2cv1, 2cv2, 1j09, 1n75, 1n77, 1n78, 2dxi Absent — 
 T. elongatus 2cfo Absent CXCX20YX3
 M. tuberculosis 2ja2 Absent — 
 B. thailandensis 4g6z Absent — 
Bacterial GluRS1 T. maritima (TM1351) 3afh Absent — 
Bacterial GluRS2 T. maritima (TM1875) 2o5r Absent — 
Archaeal GluRS M. thermautotrophicus 3aii Present CXCX14CXC 
Bacterial Glu-Q-RS E. coli 1nzj, 4a91 Present CXCX11YX3

EXPERIMENTAL

Cloning and purification of pZBD-chimeras of Ec-GluRS

A previously reported [15] Ec-GluRS encoding plasmid was used for the construction of pZBD-chimeras. Four sets of oligonucleotides were designed and were purchased from Integrated DNA Technologies. Using the above Ec-GluRS plasmid as a template, two separate sets of PCR reactions were performed with appropriate oligonucleotide combinations. The overall construction scheme along with the adequate oligonucleotide sequences is shown in Supplementary Figure S1. The resultant PCR products were cloned separately in the pETSUMO2 vector, as described previously [15]. All constructs were confirmed by DNA-sequencing (Applied Biosystems) and purified as described previously [15].

Overproduction and purification of E. coli tRNAGlu

E. coli tRNAGlu was overproduced in vivo using plasmid pKR15 (kindly provided by Professor Jacques Lapointe, Université Laval, Québec, Canada). E. coli DH5α containing pKR15 was grown at 37°C for ∼16 h in Luria broth media. The cells were harvested by centrifugation at 2700 g for 15 min and re-suspended in buffer containing 50 mM sodium acetate, 10 mM MgCl2 and 0.1 mM EDTA (pH 6). The RNA was then extracted in the aqueous layer using phenol–chloroform procedure (phenol was buffered with Tris/HCl, pH 7). The RNA was precipitated from the aqueous layer by using excess of pre-chilled isopropyl alcohol followed by incubation at −20°C for ∼20 h. The resulting white precipitate was separated by centrifugation at 16000 g at 4°C. The precipitate was dissolved in a buffer containing 20 mM HEPES (pH 7), 10 mM MgCl2, 100 mM NaCl and was loaded on to a 5 ml of HiTrap Q HP column (GE-healthcare). The column was washed with a buffer containing 20 mM HEPES, pH 7, 10 mM MgCl2 and 200 mM NaCl. Finally, purified tRNAGlu was eluted with a NaCl gradient (300 mM, 1 M) in a buffer containing 20 mM HEPES, pH 7 and 10 mM MgCl2. Fractions containing purified tRNA were confirmed by 10% urea/PAGE and were extensively dialysed in sterile water. The purified tRNA samples were then lyophilized and dissolved in diethylpyrocarbonate treated Mili-Q water. The acceptor activity of the purified tRNAGlu was measured to be ∼1.1 nmol/D260.

Structural studies by CD and fluorescence spectroscopy

Far-UV CD (200—260 nm) studies were performed using a Jasco J-815 spectro-polarimeter at 25°C in 50 mM phosphate buffer (pH 7.5) containing 100 mM NaCl using a cuvette of 2-mm path-length and a protein concentration of 10 μM. The Near-UV CD spectra (260—330 nm) of the protein samples (∼10 μM) were recorded in the same machine but with a 10-mm path-length cuvette. Steady state fluorescence spectra (310–400 nm) of the protein samples (∼5 μM each) were recorded in a Hitachi F7000 spectrofluorimeter, using a cuvette of path-length of 5 mm (excitation wavelength: 295 nm). All of the above experiments were carried out at 25°C. Protein concentrations were determined spectrophotometrically (absorption at 280 nm) for this and all following studies.

Binding affinities of tRNAGlu and ATP using fluorescence spectroscopy

Binding of substrates (tRNAGlu and ATP) with Ec-GluRS and the pZBD-chimeras were monitored by tryptophan-fluorescence quenching experiments in a Hitachi F7000 spectro-fluorimeter. All binding experiments were performed in 20 mM HEPES buffer, pH 7.5, containing 5 mM MgCl2, with an enzyme concentration of 0.5 μM (for Ec-tRNAGlu titrations) and 2 μM (for ATP titrations) at 25°C. Binding of Ec-tRNAGlu was monitored by single point titration as described before [16]. The excitation and emission wavelengths were 295 and 340 nm respectively. The resulting binding isotherms were analysed using standard equations assuming a 1:1 binding stoichiometry, as described earlier [17].

Glutamylation assay

For the glutamylation assay, uniformly labelled [3,4-3H] L-glutamic acid was purchased from Perkin–Elmer, with a specific radioactivity is 50 Ci/mmol. In vitro glutamylation assay of the wild-type and the pZBD-chimeras of Ec-GluRS were carried out with 4 μM of the previously isolated Ec-tRNAGlu in 50 mM HEPES (pH 7.5), 0.1 mM unlabelled L-glutamic acid, 16 mM MgCl2, 2 mM ATP, 0.8 mM β-mercaptoethanol and [3,4-3H] L-Glu (0.5 μl of stock per 100 μl of assay buffer) at 37°C using a methodology described earlier [16]. Kinetic parameters (Km and kcat) associated with the glutamylation reactions were determined with respect to L-glutamic acid (50-300 μM) at 37°C as described earlier [18].

Measurement of zinc-content

The zinc-contents of Ec-GluRS and pZBD-chimeras were measured using a spectroscopic method [19] with methyl methanethiolsulfonate (MMTS) as the cysteine modifier [20] and 4-(2-pyridylazo) resorcinol (PAR) as the zinc-sensitive probe that displays εmax at 416 and 500 nm in zinc-free and zinc-bound states respectively. Aliquots of 10 μM proteins were incubated with 50 μM PAR in 50 mM Tris/HCl (pH 7) buffer containing 500 mM NaCl and the absorbances at 500 nm were recorded (as background). Protein-bound zinc was released by adding 100 μM MMTS (from DMSO stock solution) and the time course of Zn–(PAR)2 formation was monitored by measuring the absorbance at 500 nm as a function of time for 10 min. Concentrations of protein-bound zinc were estimated by correlating the background-subtracted absorbance at 500 nm with a standard curve where the protein in the above protocol was replaced by known amounts of ZnSO4. All experiments were carried out at 25°C.

Database construction and sequence analysis

A total of 212 bacterial GluRS sequences (Supplementary Figure S2; Supplementary Table S1) were analysed from a database published earlier [14]. In addition, 61 bacterial Glu-Q-RS sequences (Supplementary Figure S3; Supplementary Table S2), 31 archaeal GluRS sequences (Supplementary Figure S4; Supplementary Table S3), 22 eukaryal GluRS and 11 glutamyl-prolyl-tRNA synthetase (GluProRS) sequences (Supplementary Figure S5; Supplementary Table S4) were compiled from Kyoto Encyclopedia of Genes and Genomes database [21]. Using a methodology published earlier, sequences were aligned using PROMALS3D [22], with default parameters using X-ray structures of bacterial GluRS, archaeal GluRS and bacterial Glu-Q-RS. Structural alignments were performed using by MATRAS [23] with default parameters.

RESULTS

Analysis of pZBDs and ZB-motifs in bacterial GluRS

Among the six available structures pZBDs present in bacterial GluRS (Table 1) only Bb-GluRS is zinc-bound, chelated by the CxCx20–21Yx3C motif (bold letters indicate co-ordinating residues). Ec-GluRS, for which only a preliminary crystallization report is available [15], also displays an identical ZB-motif CxCx20Yx3C (the histidine mentioned in the introduction section appears after the motif as: CxCx20Yx3CxH). The bound Zn2+ in Ec-GluRS was shown to be functionally important [9,10,12]. Yet, pZBDs in many bacterial GluRS sequences lack a characteristic ZB-motif [12]. We investigated the presence (or absence) of ZB-motifs and the length distribution of pZBDs in a database of 212 bacterial GluRS sequences (Table 2; Figure 2; Supplementary Figure S2). A ZB-motif was present in about 44% (93) cases in the database. As shown in Figure 2(A), the length of pZBDs mostly varied between 46 and 54 residues. However, for pZBDs without a ZB-motif, two minor populations, with significantly shorter lengths (seven each in the range 31–36 and 10–17), were also observed.

Table 2
A Summary of ZB-motifs in GluRS and Glu-Q-RS
 Protein (total) Group (total) ZB-motif 
Bacteria GluRS (212) I (88) CxCx20–21Yx3
  II (5) CxCx20Yx3
  III (105) No ZB-motif (long pZBD
  IV (7) No ZB-motif (short pZBD
  V (7) pZBD-deleted 
 Glu-Q-RS (61) I (46) CxCx11–28Yx3
  II (15) No ZB-motif 
Archaea GluRS (37) I (7) CxCx14CxC 
  II (16) CxCx14CxH 
  III (14) No ZB-motif 
Eukarya GluRS (22) I (11) CxCx20Yx3
  II (1) CxCx20Cx3
  III (10) No ZB-motif 
 GluProRS (11) 11 No ZB-motif 
 Protein (total) Group (total) ZB-motif 
Bacteria GluRS (212) I (88) CxCx20–21Yx3
  II (5) CxCx20Yx3
  III (105) No ZB-motif (long pZBD
  IV (7) No ZB-motif (short pZBD
  V (7) pZBD-deleted 
 Glu-Q-RS (61) I (46) CxCx11–28Yx3
  II (15) No ZB-motif 
Archaea GluRS (37) I (7) CxCx14CxC 
  II (16) CxCx14CxH 
  III (14) No ZB-motif 
Eukarya GluRS (22) I (11) CxCx20Yx3
  II (1) CxCx20Cx3
  III (10) No ZB-motif 
 GluProRS (11) 11 No ZB-motif 

Analysis of pZBD in GluRS and Glu-Q-RS

Figure 2
Analysis of pZBD in GluRS and Glu-Q-RS

(A) Length distributions of pZBDs in bacterial GluRS and representative structures of canonical length bacterial pZBD with zinc. (B) A canonical length pZBD with zinc. (C) A canonical length pZBD without zinc. (D) A short pZBD without zinc. (E) pZBD-deleted bacterial GluRSs (eco: E. coli; nse1: N. sennetsu GluRS1; min: M. infernorum; amu: A. muciniphila; poy: Onion yellows phytoplasma; pml: P. mali; acl: A. laidlawii; phm: P. mikurensis). (F) Length distributions of pZBDs in bacterial Glu-Q-RS. (G) Length distributions of pZBDs in archaeal GluRS. (H) Length distributions of pZBDs in eukaryal GluRS/GluProRS.

Figure 2
Analysis of pZBD in GluRS and Glu-Q-RS

(A) Length distributions of pZBDs in bacterial GluRS and representative structures of canonical length bacterial pZBD with zinc. (B) A canonical length pZBD with zinc. (C) A canonical length pZBD without zinc. (D) A short pZBD without zinc. (E) pZBD-deleted bacterial GluRSs (eco: E. coli; nse1: N. sennetsu GluRS1; min: M. infernorum; amu: A. muciniphila; poy: Onion yellows phytoplasma; pml: P. mali; acl: A. laidlawii; phm: P. mikurensis). (F) Length distributions of pZBDs in bacterial Glu-Q-RS. (G) Length distributions of pZBDs in archaeal GluRS. (H) Length distributions of pZBDs in eukaryal GluRS/GluProRS.

We classified bacterial GluRSs into five groups (Table 2) based on a multiple sequence alignment of the entire database (Supplementary Figure S2). The first group (88 GluRS) contained the canonical ZB-motif CxCx20–21Yx3C. The second group (five GluRS) contained a modified ZB-motif CxCx20Yx3H. The third and the largest group (105 GluRS) did not display any discernable ZB-motif. Available crystal structures [2426] show that the core architecture of pZBDs remains unaltered with (Figure 2B) or without a bound Zn2+ ion (Figures 2B and 2C). The seven GluRS sequences, with length 31–36 (marked as ‘short pZBD’ in Figure 2A), form the outlier fourth group. The crystal structure of one pZBD in this category from Thermatoga maritima GluRS1 (Tm-GluRS1) [26] shows that the short pZBD can adopt a non-canonical folded structure (Figure 2D). Surprisingly, all members of the fifth group (seven with length 10–17) had their pZBDs deleted (marked as ‘deleted pZBD’ in Figure 2A), except for the first and the last (E3 and E4; see Figure 1B) β-strands (see Figure 2E for pZBD sequence alignment).

pZBDs and ZB-motifs in bacterial Glu-Q-RS and GluRS from archaea and eukarya

For comparison, we also analysed representative pZBDs from Glu-Q-RS (Supplementary Figure S3), a catalytic-domain-only paralogue of bacterial GluRS [27]. The crystal structure of E. coli Glu-Q-RS [27,28], is also bound to a Zn2+ ion (inset to Figure 2F), ligated by three cysteine and one tyrosine residues and present in the motif CxCx11–28Yx3C. This motif is present in 46 (out of 61) Glu-Q-RS sequences. The length distribution of pZBDs with a ZB-motif showed a broad distribution around 45 with a sharp peak at 25 (corresponding to E. coli Glu-Q-RS); pZBDs without a ZB-motif also show a broad distribution.

The N-terminal domain of GluRS containing the pZBD is thought to be of ancient origin and is homologous in bacteria, archaea, as well as in eukaryotes [29]. We analysed pZBDs from representative archaea and eukaryotic GluRSs. The ZB-motif in archaeal GluRS, CxCx14CxH or CxCx14CxC (present in Methanothermobacter thermautotrophicus GluRS whose structure [30] is shown in the inset to Figure 2G), is slightly different from that present in bacterial GluRS. The length distributions of archaeal pZBD, with (23 out of 37) or without a ZB-motif, are very similar to each other, with a peak at 53 (Figure 2G). Out of a total of 33 eukaryal GluRS sequences examined here (including 11 GluProRS), 11 show the bacteria-like ZB-motif CxCx20Yx3C, whereas one showed the archaeal motif CxCx20Cx3C. The lengths of all pZBDs are restricted mostly to 54 residues (Figure 2H).

Phylum-specific distribution of pZBD in bacterial GluRS

The appearance of myriad pZBDs in bacterial GluRS calls for their origin and evolutionary history. We addressed this by projecting the ZB-motifs on to a previously published phylogenetic tree of the parent GluRS sequences [14]. The ZB-motif-annotated GluRS phylogenetic tree shows that specific motifs mostly appear in monophyletic clades with a phylum-specific preference (Figure 3). The canonical motif (CxCx20–21Yx3C; group-I Table 2) is associated with specific proteobacterial classes like γ (GluRS, GluRS1 and GluRS2; where GluRS1 and GluRS2 refer to two copies of non-identical GluRS present in some bacteria [31] where GluRS1 is more specific to tRNAGlu whereas GluRS2 is specific to tRNAGln), δ (GluRS) and α (especially GluRS that cluster with non-proteobacteria) and non-proteobacterial phyla like fusobacteria (GluRS), firmicutes (GluRS), hyperthermophilic bacteria (GluRS). The canonical motif also appears in spirochetes, green non-sulfur bacteria (GluRS), tenericutes, chlamydae and acidobacteria, although not in a strict sense. A slightly altered motif (CxCx20Yx3H; group-II, Table 2) occurs in three cyanobacteria (GluRS), one spirochaete (GluRS) and one firmicute (GluRS). Seven GluRSs with short pZBD (group-IV, Table 2) appear mostly as GluRS1 in hyperthermophilic bacteria in addition to two occurrences in verrucomicrobia (GluRS). The seven pZBD-deleted GluRSs sequences (group-V, Table 2; Figure 2E), three tenericutes (GluRS), two verrucomicrobia (GluRS), one plancomycets (GluRS) and one α-proteobacterim (GluRS1), mostly form monophyletic groups.

Phylogenetic tree of bacterial GluRS with and without ZB-motifs

Figure 3
Phylogenetic tree of bacterial GluRS with and without ZB-motifs

Projection of ZB-motifs (present in pZBDs) on to the phylogenetic tree of bacterial GluRS. The abbreviations stand for bacterial phyla and classes: proteobacteria (α, β, δ, ε, γ), hyperthermophilic bacteria (ht), acidobacteria (ad), spirochaetes (sp), bacteroidetes (ba), chlamydiae (ch), fusobacteria (fu), deinococcus-thermus (dt), (ns), green sulfur bacteria (gs), planctomycetes (pl), verrucomicrobia (ve), cyanobacteria (cy), actinobacteria (ac), firmicutes (fi), tenericutes (te). Suffixes 1 and 2 stand for GluRS1 and GluRS2 respectively (no suffix signifies canonical GluRS).

Figure 3
Phylogenetic tree of bacterial GluRS with and without ZB-motifs

Projection of ZB-motifs (present in pZBDs) on to the phylogenetic tree of bacterial GluRS. The abbreviations stand for bacterial phyla and classes: proteobacteria (α, β, δ, ε, γ), hyperthermophilic bacteria (ht), acidobacteria (ad), spirochaetes (sp), bacteroidetes (ba), chlamydiae (ch), fusobacteria (fu), deinococcus-thermus (dt), (ns), green sulfur bacteria (gs), planctomycetes (pl), verrucomicrobia (ve), cyanobacteria (cy), actinobacteria (ac), firmicutes (fi), tenericutes (te). Suffixes 1 and 2 stand for GluRS1 and GluRS2 respectively (no suffix signifies canonical GluRS).

Superposition of pZBDs of known GluRS structures with and without a ZB-motif

The absence of ZB-motif in about ∼ 60% cases suggests that the role of Zn2+ is not universally important in bacterial GluRS. By co-ordinating four non-contiguous amino acids, the bound Zn2+ is expected to play a structural role in ZB-motif-containing GluRS. What about structures lacking a bound Zn2+ despite displaying a ZB-motif [25] or those with a disrupted ZB-motif? Superposition of the pZBDs from known GluRS structures (Figures 4A and 4B) showed that the core fold, defined by two consecutive helices (H4 and H5) capped by two β-strands (E3 and E4), is maintained in almost all structures, with or without zinc. Minor exceptions are: (i) E. coli Glu-Q-RS [27,28] where the first helix H4 is short (two-turn) and the second helix H5 is lost, (ii) Burkholderia thailandensis GluRS (Bt-GluRS) [32] where the first helix H4 is disordered in the crystal structure and the second helix H5 is missing, (iii) Mycobacterium tuberculosis GluRS (Mt-GluRS) [33] and Tt-GluRS [34] where the first helix H4 is short (three-turn) with a one-turn helix between the two longer helices H4 and H5 and, (iv) T. maritima GluRS2 (Tm-GluRS2) [26] which also displays a one-turn helix between the two longer helices H4 and H5, the latter displaced compared with Bb-GluRS. The overall conservation of the core fold suggests that bacterial GluRS without Zn2+ may exploit an alternative strategy to maintain the local structural scaffold.

Superposition of GluRS pZBDs and the design of Ec-GluRS pZBD-chimeras

Figure 4
Superposition of GluRS pZBDs and the design of Ec-GluRS pZBD-chimeras

(A) Structural superposition and (B) structure-based multiple sequence alignment of pZBDs from bacterial GluRS with known crystal structures (Table 1). Structural superposition as shown here was restricted to the first and the last β-strands only (E3 and E4). RMSD and the number of superposed atoms for pZBD from 2o5r, 2ja2, 2cuz, 4a91, 2cfo and 4g6z (against pZBD of 4gri) were 0.804 Å (22), 0.884 Å (37), 0.939 Å (32), 0.333 Å (23), 0.805 Å (54) and 0.576 Å (16) respectively. (C) Sequence of pZBDs of Ec-GluRS and four pZBD-chimeras used in this work. The grey-highlighted sequence stretches were grafted into the corresponding sequence blocks in Ec-GluRS pZBD in the chimeras. The four zinc-binding residues are highlighted in panels (B) and (C).

Figure 4
Superposition of GluRS pZBDs and the design of Ec-GluRS pZBD-chimeras

(A) Structural superposition and (B) structure-based multiple sequence alignment of pZBDs from bacterial GluRS with known crystal structures (Table 1). Structural superposition as shown here was restricted to the first and the last β-strands only (E3 and E4). RMSD and the number of superposed atoms for pZBD from 2o5r, 2ja2, 2cuz, 4a91, 2cfo and 4g6z (against pZBD of 4gri) were 0.804 Å (22), 0.884 Å (37), 0.939 Å (32), 0.333 Å (23), 0.805 Å (54) and 0.576 Å (16) respectively. (C) Sequence of pZBDs of Ec-GluRS and four pZBD-chimeras used in this work. The grey-highlighted sequence stretches were grafted into the corresponding sequence blocks in Ec-GluRS pZBD in the chimeras. The four zinc-binding residues are highlighted in panels (B) and (C).

Designing the pZBD-chimeras of Ec-GluRS

To understand the role played by Zn2+ in bacterial GluRS, we focused on Ec-GluRS and constructed four chimeric versions (Figure 4C) where part of the Ec-GluRS pZBD was replaced by a stretch of residues from other bacterial GluRS pZBDs. The pZBD can best be defined from a structural superposition of known GluRS structures as the domain anchored by a two-stranded (E3 and E4) β-sheets (Figures 1B and 4). In the first chimera [Ec(Te)-GluRS], the pZBD of Ec-GluRS was replaced by the corresponding pZBD from Thermosynechococcus elongatus GluRS (Te-GluRS) whose structure [25] is devoid of a bound Zn2+ despite containing the modified ZB-motif CxCxnYx3H. In the second chimera [Ec(EQRS)-GluRS] the pZBD of Ec-GluRS was replaced by the corresponding pZBD from E. coli Glu-Q-RS (Ec-EQRS) whose zinc-bound structure [28] contains the ZB-motif CxCxnYx3C. In the third chimera [Ec(Bt)-GluRS], the pZBD of Ec-GluRS was replaced by a 19-residue stretch from the pZBD of Bt-GluRS which contains a disrupted ZB-motif CxMx20Yx3W and whose structure is devoid of a bound Zn2+ ion [32]. The 19-residue stretch starts from the residue preceding the fourth zinc-co-ordinating cysteine residue in Ec-GluRS and continues until the last β-strand of the pZBD fold. In most other GluRS (Figure 4B), this stretch adopts a helical structure but not in Bt-GluRS. It is also this stretch where the pZBDs of Ec-GluRS and Bt-GluRS differ the most. In addition, a pZBD-disrupted Ec-GluRS was constructed [Ec(ΔH4)-GluRS] where the ZB-motif of Ec-GluRS was left untouched but the first helix was shortened by grafting the corresponding stretch from Ec-EQRS in which the first helix is only two-turn long. All four variants of Ec-GluRS, along with the wild-type enzyme, were overexpressed and purified for structural and functional studies.

The structural integrity of Ec-GluRS remains unaltered by pZBD perturbations

Far-UV CD spectroscopy is an important tool to study secondary structure of the polypeptide whereas near-UV CD spectroscopy is a tool to gauge tertiary structure of proteins. Intrinsic protein fluorescence is a complementary tool that can report the solvent accessibility tryptophan residues in proteins. To identify the overall structural perturbation in pZBD-chimeras CD and fluorescence spectroscopy were performed and compared with that of Ec-GluRS.

The far-UV CD spectra of wild-type Ec-GluRS, Ec(Bt)-GluRS, Ec(Te)-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS are comparable (Figure 5A), an indication of very similar secondary structure contents. The fact that secondary structural content of Ec-GluRS is comparable to that of the pZBD-chimeras with shorter pZBDs [Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS] possibly indicates that the longer stretch of pZBD in Ec-GluRS, not present in Ec(EQRS)-GluRS or Ec(ΔH4)-GluRS, may actually be devoid of canonical secondary structure. In summary, the secondary structures of all three pZBD-chimeras of Ec-GluRS are similar to that of wild-type Ec-GluRS. The near-UV CD spectra of Ec-GluRS and the pZBD-chimeras are shown in Figure 5(A). The nature and intensity of the near-UV CD spectra are also comparable to the wild-type protein, indicating that the tertiary structure of Ec-GluRS remains almost unaltered after pZBD-swapping.

Solution conformational studies of Ec-GluRS and pZBD-chimeras

Figure 5
Solution conformational studies of Ec-GluRS and pZBD-chimeras

(A) CD spectra of Ec-GluRS variants in phosphate buffer (pH 7.5) containing 100 mM NaCl at 25°C. (B) Fluorescence spectra of Ec-GluRS variants under identical conditions as panel (A).

Figure 5
Solution conformational studies of Ec-GluRS and pZBD-chimeras

(A) CD spectra of Ec-GluRS variants in phosphate buffer (pH 7.5) containing 100 mM NaCl at 25°C. (B) Fluorescence spectra of Ec-GluRS variants under identical conditions as panel (A).

Apart from Bt-ZBD, which has nine tryptophan residues, all pZBDs in the chimeras and in Ec-GluRS have eight tryptophan residues each. The emission maxima in tryptophan fluorescence spectra of all chimeras and wild-type Ec-GluRS were identical (λmax=334 nm; Figure 5B). The relative intensities were also comparable indicating that that replacement pZBD of Ec-GluRS does not affect the overall solvent exposure of tryptophan residues. Therefore, it can be concluded from the spectroscopic analysis that the overall secondary structure and compactness of wild-type Ec-GluRS remains unaltered in the chimeric constructs studied in the present paper.

Association of GluRS with tRNAGlu but not with ATP is sensitive to pZBD perturbations

Efficient glutamylation demands optimal binding of substrates (ATP, tRNAGlu and L-glutamic acid) by GluRS. Binding of tRNAGlu induces conformational changes in GluRS that stimulates the binding of L-glutamic acid leading to the productive binding of ATP [16,34,35]. Substrate binding is known to quench tryptophan fluorescence [17]. Therefore, the extent of substrate binding to Ec-GluRS can be monitored by fluorescence titration experiments. The binding affinities of ATP and Ec-tRNAGlu towards pZBD-chimeras of Ec-GluRS were measured from tryptophan fluorescence detected titrations and compared with that of wild-type Ec-GluRS.

It was shown by Liu et al. [9] that 1,10-phenanthroline-mediated zinc depletion and inactivation of Ec-GluRS is inhibited in the presence of ATP. Even at an elevated temperature of 50°C ATP strongly hindered the depletion of zinc, indicating a close association between pZBD and the ATP-binding pocket [9]. Therefore it is instructive to demonstrate how pZBD-swapped versions of Ec-GluRS bind ATP. The dissociation constants of ATP (Kd), derived from tryptophan fluorescence titration (Table 3), shows that affinity of ATP to Ec-GluRS (28.0 μM) is comparable to its pZBD-variants [Ec(Bt)-GluRS: 39.5 μM; Ec(Te)-GluRS: 32.0 μM; Ec(EQRS)-GluRS: 35.5 μM; Ec(ΔH4)-GluRS: 28.0 μM]. Thus, swapping of pZBD of Ec-GluRS by pZBDs from other GluRSs or even with the pZBD of Ec-Glu-Q-RS does not alter the ATP-binding activity.

Table 3
Kinetic assay parameters and zinc-content of Ec-GluRS and its variants
Ec-GluRS constructs kcat (s−1Km (L-glutamic acid; 10−6 M) kcat/Km (L-glutamic acid; M−1·s−1Activity loss Kd (ATP) (10−6 M) Kd (tRNAGlu) (10−9 M) [Zn2+]/[protein]* (molar ratio) 
wt Ec-GluRS 5.3±0.1 65.4±6.7 8.1×104 – 28.0±2.8 62.5±6.3 0.93±0.01 
Ec(Bt)-GluRS 6.6±0.7 139.4±3.7 4.7×104 ∼2 39.5±0.7 91.0±7.0 0.05±0.01 
Ec(Te)-GluRS 0.20±0.03 250.5±4.0 8.0×102 ∼100 32.0±4.2 222±5.6 1.08±0.08 
Ec(EQRS)-GluRS (1.5±0.2) × 10−2 284.6±3.0 53 ∼1500 35.5±0.7 385±19 1.09±0.05 
Ec-(ΔH4)-GluRS (8.5±0.1) × 10−3 441.5±9.2 19 ∼4200 28.0±1.4 499±8.4 1.00±0.05 
Ec-GluRS constructs kcat (s−1Km (L-glutamic acid; 10−6 M) kcat/Km (L-glutamic acid; M−1·s−1Activity loss Kd (ATP) (10−6 M) Kd (tRNAGlu) (10−9 M) [Zn2+]/[protein]* (molar ratio) 
wt Ec-GluRS 5.3±0.1 65.4±6.7 8.1×104 – 28.0±2.8 62.5±6.3 0.93±0.01 
Ec(Bt)-GluRS 6.6±0.7 139.4±3.7 4.7×104 ∼2 39.5±0.7 91.0±7.0 0.05±0.01 
Ec(Te)-GluRS 0.20±0.03 250.5±4.0 8.0×102 ∼100 32.0±4.2 222±5.6 1.08±0.08 
Ec(EQRS)-GluRS (1.5±0.2) × 10−2 284.6±3.0 53 ∼1500 35.5±0.7 385±19 1.09±0.05 
Ec-(ΔH4)-GluRS (8.5±0.1) × 10−3 441.5±9.2 19 ∼4200 28.0±1.4 499±8.4 1.00±0.05 

*Under identical experimental conditions, [Zn2+]/[Ec-GlnRS]=0.06±0.01.

Abbreviation: wt, wild-type.

The pZBD of Ec-GluRS has been proposed to interact with the acceptor arm of Ec-tRNAGlu and impart higher selectivity of L-glutamic acid over other non-cognate amino acids by assisting in reorganizing the active site of GluRS [12]. The fluorescence titration derived Kd values (Table 3) of Ec-tRNAGlu dissociation from Ec(Bt)-GluRS (91.0 nM) and Ec(Te)-GluRS (222 nM) are about 1.4- and 3.5-times higher than that of wild-type Ec-GluRS (62.5 nM). This indicates that pZBD of Te-GluRS is not as compatible as the pZBD of Bt-GluRS towards Ec-tRNAGlu. The Kd values (Table 3) of Ec-tRNAGlu dissociation from Ec(EQRS)-GluRS (385 nM) and Ec(ΔH4)-GluRS (499 nM) are ∼6- and 8-fold higher than that of Ec-GluRS. Compared with Ec-GluRS, the free energies of Ec-tRNAGlu-binding (ΔG0=−RTlnKd) to Ec(Bt)-GluRS, Ec(Te)-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS are unfavourable by 0.22, 0.75, 1.07 and 1.23 kcal/mol (1 cal ≡ 4.184 J) respectively. Overall, the experiments showed that tRNAGlu binding to GluRS is sensitive to pZBD perturbations.

Except Ec(Bt)-GluRS all pZBD-chimeras show 100-fold or more weaker catalytic efficiency

Glutamylation of tRNAGlu by GluRS is the exclusive pathway for Glu-tRNAGlu synthesis. In this reaction, L-glutamic acid is first activated by GluRS in presence of ATP to form the adenylate complex. This is followed by the catalytic step where the acceptor stem of tRNAGlu is glutamylated. In the light of the weaker affinities of tRNAGlu for most pZBD-chimeras, one would expect poor glutamylation efficiencies. However, the effect of the pZBD-perturbations on the catalytic step (kcat) may not necessarily be reflected in the weaker Kd values. The kinetic parameters (kcat and Km) for tRNAGlu glutamylation of Ec-GluRS and pZBD-chimeras were measured by varying L-glutamic acid concentration.

The Km variation of the pZBD-mutants showed a trend very similar to that of Kd (GluRS–tRNAGlu) variation. Compared with the Km of wild-type Ec-GluRS (65.4 μM), the Km values of the pZBD-chimeras (Table 3) were enhanced: Ec(Bt)-GluRS (2.1-fold), Ec(Te)-GluRS (3.8-fold), Ec(EQRS)-GluRS (4.3-fold) and Ec(ΔH4)-GluRS (6.7-fold). In terms of free energies, the association process is unfavourable by 0.45, 0.79, 0.87 and 1.13 kcal/mol respectively, for the chimeras. The kcat variation of the pZBD-mutants (Table 3), on the other hand, showed a slightly different trend. Although kcat values for Ec-GluRS (5.3 s−1) and Ec(Bt)-GluRS (6.6 s−1) were comparable, kcat values of other chimeras were order of magnitude smaller: Ec(Te)-GluRS (26-fold), Ec(EQRS)-GluRS (353-fold) and Ec(ΔH4)-GluRS (623-fold). The overall efficiency of the glutamylation step is reflected in the ratio kcat/Km (Table 3). Compared with Ec-GluRS (kcat/Km=8.1×104 M−1 s−1), the kcat/Km ratios for Ec(Te)-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS were lower by ∼ 100-fold, 1500-fold and 4200-fold respectively. On the other hand, the kcat/Km ratios for Ec-GluRS and Ec(Bt)-GluRS (kcat/Km=4.7×104 M−1 s−1) were comparable.

In summary, the above experimental results suggest that the pZBD in GluRS not only affects the local interaction with tRNAGlu acceptor arm (thus modulating the active site) but also affects the catalytic step of glutamylation reaction indicating some distant communication with other part of the protein. Among the three pZBD-variants, favourable long-range communication between this domain and the rest of the protein is most severely hampered in Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS, both with a shorter H4 helix than Ec-GluRS (Figure 4C). This suggests that the H4 helix of Ec-GluRS pZBD is important for glutamylation of tRNAGlu. Interestingly tRNAGlu–H4 helix interaction was also observed in the crystal structure of Tt–GluRS–tRNAGlu complex (Figure 1).

Except Ec(Bt)-GluRS all pZBD-chimeric constructs contain zinc

The zinc-content of Ec-GluRS and all pZBD chimeras were spectroscopically determined from the absorbance of zinc-bound PAR at 500 nm (Figure 6A) [19]. The experimentally determined protein–zinc molar ratios are summarized in Table 3. Moles of zinc, released from each mole of protein, were 0.93 (Ec-GluRS), 1.09 [Ec(EQRS)-GluRS], 1.08 [Ec(Te)-GluRS] and 1.0 [Ec(ΔH4)-GluRS], confirming that the pZBDs of these proteins are zinc-bound. This is not unexpected since the presence of zinc in Ec-GluRS was reported earlier [9,10,12]. Similarly, the pZBD of Ec(EQRS)-GluRS is expected to bind zinc since EQRS–pZBD contains a ZB-motif and Ec-EQRS crystal structure also contains zinc [27,28] Likewise, the pZBD of Ec(ΔH4)-GluRS is almost identical with that in Ec-GluRS except that the first helix H4 in pZBD of Ec(ΔH4)-GluRS is shortened without disturbing the four potential zinc-co-ordinating residues. Therefore, that it also released one equivalent of zinc is not surprising. The crystal structure of Te-GluRS does not contain zinc despite having a modified ZB-motif [25] Therefore, the presence of zinc in Ec(Te)-GluRS was a little unexpected.

Zinc-content of Ec-GluRS and pZBD-chimeras

Figure 6
Zinc-content of Ec-GluRS and pZBD-chimeras

(A) UV-vis spectra of pure Ec(Te)-GluRS (10 μM) and its mixture with PAR (50 μM) and MMTS (100 μM) in Tris/HCl buffer (pH 7) containing 500 mM NaCl at 25°C. (B) Time evolution of zinc-bound PAR (absorbance at 500 nm) upon addition of PAR (50 μM) to a mixture of MMTS (100 μM) and Ec-GluRS variants (∼10 μM) in Tris/HCl buffer (pH 7) containing 500 mM NaCl at 25°C.

Figure 6
Zinc-content of Ec-GluRS and pZBD-chimeras

(A) UV-vis spectra of pure Ec(Te)-GluRS (10 μM) and its mixture with PAR (50 μM) and MMTS (100 μM) in Tris/HCl buffer (pH 7) containing 500 mM NaCl at 25°C. (B) Time evolution of zinc-bound PAR (absorbance at 500 nm) upon addition of PAR (50 μM) to a mixture of MMTS (100 μM) and Ec-GluRS variants (∼10 μM) in Tris/HCl buffer (pH 7) containing 500 mM NaCl at 25°C.

Although it contains one equivalent of zinc, the kinetics of zinc-release by Ec(Te)-GluRS was different from others (Figure 6B). It took Ec-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS about 5 min to release about 90% of the bound zinc. In contrast, almost all the bound zinc in Ec(Te)-GluRS was released within a minute. The zinc-chelator MMTS reacts with cysteine thiol groups to release zinc [20], therefore the presence of only two cysteines in Ec(Te)-GluRS, as opposed to three in the rest [Ec-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS] explains why zinc-release was much faster in Ec(Te)-GluRS. This also makes Ec(Te)-GluRS a weaker binder of zinc than the rest. Interestingly, Bacillus subtilis GluRS, with an identical ZB-motif as Ec(Te)-GluRS (CxCxnYx3H) also binds to zinc but with a weak affinity as evident from the fact that only 0.6 moles of zinc (per mole of protein) was found to be protein-bound [9]. We presume that the apparently loosely bound zinc in Ec(Te)-GluRS was lost during the crystallization process due to the presence of mild chelators (like sodium citrate and citric acid) in the crystallization cocktail and as a result the crystal structure is devoid of any bound zinc.

In contrast with the above results, the chimera Ec(Bt)-GluRS released 0.05 moles of zinc per mole of protein. This number is comparable to the control (0.06 mole of zinc), E. coli glutaminyl-tRNA synthetase (GlnRS), which is known to not bind zinc [9]. That Ec(Bt)-GluRS does not bind zinc is expected since Ec(Bt)-GluRS contains a disrupted ZB-motif CxCx20Yx3W.

DISCUSSION

Natively zinc-bound Ec-GluRS does not require zinc to be active

The aim of the present study is to revisit the functional and structural role of the pZBD, present in the N-terminal catalytic domain of bacterial GluRS (Figure 1). Previous experimental work on Ec-GluRS, where the ZB-motif in pZBD was disrupted or the Zn2+ was removed, showed that the Zn2+ ion plays an important functional role [9,10]. We revisited this conclusion by constructing three chimeric versions of Ec-GluRS where pZBDs from Te-GluRS, Bt-GluRS and Ec-Glu-Q-RS replaced the pZBD of Ec-GluRS (Figure 4C). The fourth chimera had a pZBD with shortened H4 helix. Despite being zinc-bound, Ec(Te)-GluRS, Ec(EQRS)-GluRS and Ec(ΔH4)-GluRS showed impaired enzymatic efficiencies. On the other hand, Ec(Bt)-GluRS showed comparable enzymatic efficiency as the wild-type GluRS despite lacking the canonical ZB-motif or a bound Zn2+. The results clearly demonstrate that Zn2+ ion, by itself, is not required for the function of Ec-GluRS. The dispensability of zinc in bacterial GluRS was already known since a large number of extant (therefore functional) bacterial GluRSs do not contain zinc or a ZB-motif (Table 2). In this context, our conclusion may not seem new, but in the context of Ec-GluRS, which natively contains a ZB-motif and which was earlier shown to be zinc-dependent for proper functioning, the conclusion shows that zinc is dispensable even for bacterial GluRSs that are known to contain zinc. This is in addition to zinc being dispensable for many extant bacteria that do contain a ZB-motif.

This not to say that Zn2+ doesn't play an important role in GluRSs that contain a zinc-binding motif (like Ec-GluRS). Our results show that a zinc-bound pZBD may be replaced by a structurally compatible pZBD that does not contain a zinc-binding motif without impairing the tRNA aminoacylation efficiency of the parent enzyme. In case of Ec(Bt)-GluRS, compatibility between pZBDs of E. coli (γ-proteobacterium) and B. thailandensis (β-proteobacterium) is not unexpected because of their close evolutionary origin (phylum: proteobacteria) and therefore their pZBDs are replaceable. On the other hand, the pZBDs of T. elongatus, a non-proteobacterium (cyanobacterium) and E. coli Glu-Q-RS, a paralogue of GluRS, cannot substitute for the pZBD of Ec-GluRS.

Alternate structural and functional strategies adopted by GluRS without zinc-bound pZBD

Structural superposition of GluRS pZBDs of a number of bacterial GluRSs (Figure 4) showed excellent overlap of the core fold even though several of the structures lacked a bound Zn2+ ion. This indicates that zinc-devoid pZBD must use an alternate strategy that provides the intra-chain locking induced by a bound Zn2+ ion. An analysis of four zinc-devoid pZBDs, from Bt-GluRS, Tt-GluRS, Mt-GluRS and Te-GluRS, showed a number of long-range H-bonds (more than four residue separation in the sequence; Supplementary Figure S6), mediated by side chains. In addition, a number of hydrophobic stacking, π–π stacking and cation–π interactions were also observed. We also identified a conserved cation–π interaction in all four structures, between two semi-conserved residues (Supplementary Figure S2), an arginine present in the second β-strand E4 and a tyrosine in the loop joining the helices H4 and H5 that co-ordinates zinc in zinc-bound pZBDs (Figure 7A). Interestingly, the tyrosine–arginine cation–π interaction is also present in Bb-GluRS and Ec-Glu-Q-RS, both containing a bound Zn2+ ion in their pZBDs. In both, the phenolic oxygen atom of tyrosine participates in zinc-binding whereas the ring π-cloud is found to be stacked with the guanidium moiety of arginine (Figure 7B). The pZBD of archaeal GluRS from M. thermautotrophicus is also zinc-bound but all four co-ordinating side chains are cysteine residues, with the arginine-interacting conserved tyrosine in bacterial GluRS replaced by one of the cysteines. Despite the lack of a tyrosine residue, the conserved arginine in the second β-strand E4 still participates in an alternate cation–π interaction where phenol, present in the first helix H4, is the alternate π-partner (Figure 7C).

Conserved cation–π interaction in pZBDs of bacterial GluRS

Figure 7
Conserved cation–π interaction in pZBDs of bacterial GluRS

(A) Superimposition of zinc-devoid pZBDs from Mt-GluRS (2ja2), Tt-GluRS (2cuz), Bt-GluRS (4g6z) and Te-GluRS (2cfo). (B) Superimposition of zinc-containing pZBDs from Bb-GluRS (4gri) and Ec-Glu-Q-RS (4a91). (C) pZBD of M. thermautotrophicus GluRS (3aii). Cation–π interactions in each panel [arginine–tyrosine in panels (A and B) and arginine–phenylalanine in panel (C) are highlighted by yellow arrows]. The annotated H-bond in panel (A) refers to a conserved H-bond between the phenolic oxygen atom of tyrosine and the (i + 4)-th backbone nitrogen atom.

Figure 7
Conserved cation–π interaction in pZBDs of bacterial GluRS

(A) Superimposition of zinc-devoid pZBDs from Mt-GluRS (2ja2), Tt-GluRS (2cuz), Bt-GluRS (4g6z) and Te-GluRS (2cfo). (B) Superimposition of zinc-containing pZBDs from Bb-GluRS (4gri) and Ec-Glu-Q-RS (4a91). (C) pZBD of M. thermautotrophicus GluRS (3aii). Cation–π interactions in each panel [arginine–tyrosine in panels (A and B) and arginine–phenylalanine in panel (C) are highlighted by yellow arrows]. The annotated H-bond in panel (A) refers to a conserved H-bond between the phenolic oxygen atom of tyrosine and the (i + 4)-th backbone nitrogen atom.

In addition to playing a structural role, the conserved cation–π interaction in pZBDs of bacterial GluRS, with or without a bound Zn2+ ion, may also be functionally important. In the only tRNAGlu-bound structure of Tt-GluRS (Figure 1), the arginine (Arg116 in Tt-GluRS) side chain forms a H-bond with the backbone phosphate of tRNAGlu (Ala73). The arginine side chain, locked via the cation–π interaction, is conformationally predisposed for a fruitful tRNAGlu interaction. In pZBDs with a bound Zn2+, where the phenolic oxygen atom of tyrosine is one of the co-ordinating ligands, the interaction becomes multi-layered. The tyrosine ring, immobilized due to co-ordination with Zn2+ ion, in turn, restricts the mobility of the tRNAGlu-interacting arginine side chain due to cation–π interaction. Interestingly, in all four Zn2+-devoid pZBDs, the phenolic oxygen atom of the tyrosine participates in a conserved H-bond interaction with the (i+4)-th backbone nitrogen atom (Tyr123–Trp127 in PDB ID: 4g6z; Tyr121–His125 in PDB ID: 2cfo; Tyr122–Ala126 in PDB ID: 2cuz; Tyr132–Asp136 in PDB ID: 2ja2). In place of Zn2+-ligation, this can be viewed as an alternate strategy for immobilizing the tyrosine side chain. This explains why Zn2+ is dispensable in bacterial GluRS [36,37]. The loss of Zn2+ ion within the same protein family is rarely observed and when Zn2+ is lost, the associated surrounding secondary structure/loops are also often lost [36]. For bacterial GluRS, however the loss of Zn+2 does not lead to an overall structural or functional change. It is interesting to note that a zinc-co-ordinating sphere, formed by two cysteines and two histidines, is also absent from many bacterial Ros DNA-binding domains despite the maintenance of a functional fold [37].

Is facilitation of indirect glutaminylation the core function of pZBD in bacterial GluRS?

The pZBD of GluRS is functionally coupled with the glutamylation of tRNAGlu [9,10,12]. Therefore, the deletion of the pZBD in seven extant bacterial GluRSs (Figure 2E), all wild-type and therefore functional, is surprising. It is possible that the tRNAGlu in these bacteria, especially nucleotides in the acceptor stem, are different to compensate for the deleted pZBDs. We compared the tRNAGlu sequences in bacteria for which the corresponding GluRSs (both pZBD-containing and pZB-deleted) formed a monophyletic cluster. A projection of the seven pZBD-deleted GluRS on to a phylogenetic tree of bacterial GluRS (Figure 3) showed that three pZBD-deleted GluRSs of tenericutes, Acholeplasma laidlawii, Phytoplasma mali and Onion yellows phytoplasma OY-M are monophyletic with three pZBD-containing GluRSs from tenericutes and four pZBD-containing GluRSs from firmicutes. Similarly, two pZBD-deleted GluRSs from verrucomicrobia and one from planctomycetes, Methylacidiphilum infernorum, Akkermansia muciniphila and Phycisphaera mikurensis, are monophyletic with pZBD-containing GluRSs from planctomycetes. The remaining pZBD-deleted GluRS (GluRS1) from α-proteobacteria, Neorickettsia sennetsu, appears in an isolated clade. Except for the nucleotide identity variation at 11–24 base-pair at the D-helix (AU compared with UA) in selected cases, no systematic differences were observed between tRNAGlu of bacteria with or without pZBD in the monophyletic GluRS clusters (Supplementary Figure S7). The D-helix is distant from the pZBD-contacting tRNA acceptor stem. Therefore, it can be concluded that tRNAGlu of pZBD-containing and pZBD-deleted GluRS are not significantly different.

The curious case of these seven pZBD-deleted bacterial GluRS becomes more interesting since no such deletion was observed in archaeal and eukaryal GluRSs (Supplementary Figures S3 and S4). Therefore, one may assume that some bacteria-specific evolutionary stress might be responsible for the deletion of this primordial domain. In addition to tRNAGlu, bacterial GluRS is evolutionarily and functionally connected with tRNAGln and the heterotrimeric GatCAB. GluRS glutamylates tRNAGln in bacteria that lack the tRNAGln-specific GlnRS, yielding the misacylated product GlutRNAGln. The misacylated tRNAGln is further processed to Gln-tRNAGln by GatCAB through a ternary transamidosome complex [3840]. This ancestral route for the production of Gln-tRNAGln can become redundant when the bacterial genome contains functional copies of tRNAGln-specific GlnRS. In addition, in genomes with two copies of GluRS (GluRS1 and GluRS2), only GluRS2 may participate in the indirect production of Gln-tRNAGln, making GluRS1 strictly tRNAGlu-specific [14,31,41,42].

The genomes of all but one pZBD-deleted GluRS containing bacteria (M. infernorum, the exception, is discussed later) contain either GlnRS or GluRS2 indicating that the indirect glutaminylation pathway, via tRNAGln–GluRS–GatCAB or tRNAGln–GluRS1–GatCAB interaction, may be redundant in them. Of course, one could argue that the mere appearance of GlnRS or GluRS2 in the genome may not necessarily mean that these are functional and that GluRS (or GluRS1) in these bacteria may still, along with GatCAB, produce Gln-tRNAGln by the indirect route. However, independent genomic signatures clearly indicate that either the tRNAGln–GatCAB or the GluRS–GatCAB or both the interactions are disrupted in these six bacteria. Three genomes (A. laidlawii, P. mali and Onion yellows phytoplasma) lack the gatCAB gene whereas tRNAGln in five cases (except N. sennetsu) lack the characteristic U1-A72 signature (Supplementary Figure S8) required for productive GatCAB recognition [43]. In summary, the indirect glutaminylation pathway seems to be redundant in bacteria that contain pZBD-deleted GluRS. The connection between the two becomes evident from the previously solved structure of transamidosome in T. maritima [40] where GluRS directly contacts GatCAB through the pZBD.

Correlations mentioned above are consistent with the following evolutionary scenario. The pZBD of GluRS, present at the N-terminal catalytic domain, is of primordial origin. It appeared before the bacteria-eukarya branching when all GluRS were non-discriminatory, charging both tRNAGlu and tRNAGln (the pZBD is present in extant GluRS from eukarya, bacteria and archaea). There was an evolutionary pressure against its deletion until the genome acquired an alternative route of Gln-tRNAGln synthesis (acquisition of GlnRS or GluRS2). However, even in the presence of an alternative route of Gln-tRNAGln synthesis, the pZBD may have been retained, since over time, it also played a role in optimizing GluRS–tRNAGlu interaction. Only in rare cases, where deletion of pZBD did not significantly impair GluRS–tRNAGlu interaction, it got deleted. This explains why pZBD plays an important role in GluRS–tRNAGlu interaction, yet, it is deleted in some extant bacterial GluRS. We hypothesize that the core function of pZBD in bacterial GluRS is gatB interaction and not tRNAGlu interaction.

The only bacterium that does not fit this model is M. infernorum. Even without GlnRS or GluRS2, it contains a pZBD-deleted GluRS (its tRNAGln A1-U72 signature also indicates the presence of an active glutaminylation route via GatCAB interaction). M. infernorum is an autotroph with an unusually small genome of 2.3 mbp [44] which must be a later adaptation to its unique ecological niche. It is likely that GlnRS was lost during genome streamlining and an alternate mechanism of GluRS–gatB interaction, skirting the use of pZBD, evolved. For example, GluRS from M. infernorum and Methylacidiphilum fumariolicum SolV [45], another thermoacidophilic methanotroph of the phylum verrucomicrobia, also without GlnRS and with a pZBD-deleted GluRS, contain conserved cysteine and tyrosine residues in the neighbourhood of the canonical pZBD (Supplementary Figure S9). These unique signatures may give rise to alternate structural features that compensate for the absence of pZBD. This needs to be further explored by structural and functional studies.

AUTHOR CONTRIBUTION

All authors designed the experiments and analysed the data. Nipa Chongdar and Saumya Dasgupta performed the experiments. Nipa Chongdar, Saumya Dasgupta and Gautam Basu performed the sequence and structural analysis. All authors contributed in writing the final manuscript.

Ajit B. Datta acknowledges support from Wellcome Trust–DBT Alliance Intermediate Fellowship. Preliminary work on protein zinc-content was done by Ishita Jalan.

FUNDING

This work was funded by a grant from Council of Scientific and Industrial Research [grant number 37(1494)/11/EMR-II (to G.B.)]; and the Wellcome Trust–DBT Alliance Intermediate Fellowship [grant number 500241-Z-11-Z (to A.B.D.)].

Abbreviations

     
  • Bb-GluRS

    B. burgdorferri GluRS

  •  
  • Ec-GluRS

    E. coli GluRS

  •  
  • GatCAB

    glutamine amidotransferase CAB

  •  
  • GlnRS

    glutaminyl-tRNA synthetase

  •  
  • Glu-Q-RS

    glutamyl-queuosine-tRNAAsp synthetase

  •  
  • GluRS

    glutamyl-tRNA synthetase

  •  
  • MMTS

    methyl methanethiolsulfonate

  •  
  • Mt-GluRS

    M. tuberculosis GluRS

  •  
  • PAR

    4-(2-pyridylazo) resorcinol

  •  
  • pZBD

    putative zinc-binding domain, Te-GluRS, T. elongatus GluRS

  •  
  • Tm-GluRS1

    T. maritima GluRS1

  •  
  • ZB-motif

    zinc-binding motif

References

References
1
Sankaranarayanan
 
R.
Dock-Bregeon
 
A.C.
Romby
 
P.
Caillet
 
J.
Springer
 
M.
Rees
 
B.
Ehresmann
 
C.
Ehresmann
 
B.
Moras
 
D.
 
The structure of threonyl-tRNA synthetase-tRNA(Thr) complex enlightens its repressor activity and reveals an essential zinc ion in the active site
Cell
1999
, vol. 
97
 (pg. 
371
-
381
)
[PubMed]
2
Newberry
 
K.J.
Hou
 
Y.-M.
Perona
 
J.J.
 
Structural origins of amino acid selection without editing by cysteinyl-tRNA synthetase
EMBO J.
2002
, vol. 
21
 (pg. 
2778
-
2787
)
[PubMed]
3
Zhang
 
C.-M.
Christian
 
T.
Newberry
 
K.J.
Perona
 
J.J.
Hou
 
Y.-M.
 
Zinc-mediated amino acid discrimination in cysteinyl-tRNA synthetase
J. Mol. Biol.
2003
, vol. 
327
 (pg. 
911
-
917
)
[PubMed]
4
Miller
 
W.T.
Schimmel
 
P.
 
A retroviral-like metal binding motif in an aminoacyl-tRNA synthetase is important for tRNA recognition
Proc. Natl. Acad. Sci. U.S.A.
1992
, vol. 
89
 (pg. 
2032
-
2035
)
[PubMed]
5
Zhou
 
L.
Rosevear
 
P.R.
 
Mutation of the carboxy terminal zinc finger of E. coli isoleucyl-tRNA synthetase alters zinc binding and aminoacylation activity
Biochem. Biophys. Res. Commun.
1995
, vol. 
216
 (pg. 
648
-
654
)
[PubMed]
6
Landro
 
J.A.
Schmidt
 
E.
Schimmel
 
P.
Tierney
 
D.L.
Penner-Hahn
 
J.E.
 
Thiol ligation of two zinc atoms to a class I tRNA synthetase: evidence for unshared thiols and role in amino acid binding and utilization
Biochemistry
1994
, vol. 
33
 (pg. 
14213
-
14220
)
[PubMed]
7
Bilokapic
 
S.
Maier
 
T.
Ahel
 
D.
Gruic-Sovulj
 
I.
Söll
 
D.
Weygand-Durasevic
 
I.
Ban
 
N.
 
Structure of the unusual seryl-tRNA synthetase reveals a distinct zinc-dependent mode of substrate recognition
EMBO J.
2006
, vol. 
25
 (pg. 
2498
-
2509
)
[PubMed]
8
Nureki
 
O.
Kohno
 
T.
Sakamoto
 
K.
Miyazawa
 
T.
Yokoyama
 
S.
 
Chemical modification and mutagenesis studies on zinc binding of aminoacyl-tRNA synthetases
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
15368
-
15373
)
[PubMed]
9
Liu
 
J.
Lin
 
S.X.
Blochet
 
J.E.
Pézolet
 
M.
Lapointe
 
J.
 
The glutamyl-tRNA synthetase of Escherichia coli contains one atom of zinc essential for its native conformation and its catalytic activity
Biochemistry
1993
, vol. 
32
 (pg. 
11390
-
11396
)
[PubMed]
10
Liu
 
J.
Gagnon
 
Y.
Gauthier
 
J.
Furenlid
 
L.
L'Heureux
 
P.J.
Auger
 
M.
Nureki
 
O.
Yokoyama
 
S.
Lapointe
 
J.
 
The zinc-binding site of Escherichia coli glutamyl-tRNA synthetase is located in the acceptor-binding domain. Studies by extended X-ray absorption fine structure, molecular modeling, and site-directed mutagenesis
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
15162
-
15169
)
[PubMed]
11
Makarova
 
K.S.
Aravind
 
L.
Koonin
 
E.V.
 
SWIM, a novel Zn-chelating domain present in bacteria, archaea and eukaryotes
Trends Biochem. Sci.
2002
, vol. 
27
 (pg. 
384
-
386
)
[PubMed]
12
Banerjee
 
R.
Dubois
 
D.Y.
Gauthier
 
J.
Lin
 
S.-X.
Roy
 
S.
Lapointe
 
J.
 
The zinc-binding site of a class I aminoacyl-tRNA synthetase is a SWIM domain that modulates amino acid binding via the tRNA acceptor arm
Eur. J. Biochem.
2004
, vol. 
271
 (pg. 
724
-
733
)
[PubMed]
13
Sekine
 
S.
Nureki
 
O.
Shimada
 
A.
Vassylyev
 
D.G.
Yokoyama
 
S.
 
Structural basis for anticodon recognition by discriminating glutamyl-tRNA synthetase
Nat. Struct. Biol.
2001
, vol. 
8
 (pg. 
203
-
206
)
[PubMed]
14
Dasgupta
 
S.
Basu
 
G.
 
Evolutionary insights about bacterial GlxRS from whole genome analyses: is GluRS2 a chimera?
BMC Evol. Biol.
2014
, vol. 
14
 pg. 
26
 
[PubMed]
15
Chongdar
 
N.
Dasgupta
 
S.
Datta
 
A.B.
Basu
 
G.
 
Preliminary X-ray crystallographic analysis of an engineered glutamyl-tRNA synthetase from Escherichia coli
Acta Crystallogr. Sect. F Struct. Biol. Commun.
2014
, vol. 
70
 (pg. 
922
-
927
)
16
Dasgupta
 
S.
Saha
 
R.
Dey
 
C.
Banerjee
 
R.
Roy
 
S.
Basu
 
G.
 
The role of the catalytic domain of E. coli GluRS in tRNAGln discrimination
FEBS Lett.
2009
, vol. 
583
 (pg. 
2114
-
2120
)
[PubMed]
17
Roy
 
S.
 
Fluorescence quenching methods to study protein-nucleic acid interactions
Methods Enzymol.
2004
, vol. 
379
 (pg. 
175
-
187
)
[PubMed]
18
Dasgupta
 
S.
Manna
 
D.
Basu
 
G.
 
Structural and functional consequences of mutating a proteobacteria-specific surface residue in the catalytic domain of Escherichia coli GluRS
FEBS Lett.
2012
, vol. 
586
 (pg. 
1724
-
1730
)
[PubMed]
19
Hunt
 
J.B.
Neece
 
S.H.
Ginsburg
 
A.
 
The use of 4-(2-pyridylazo)resorcinol in studies of zinc release from Escherichia coli aspartate transcarbamoylase
Anal. Biochem.
1985
, vol. 
146
 (pg. 
150
-
157
)
[PubMed]
20
Smith
 
D.J.
Maggio
 
E.T.
Kenyon
 
G.L.
 
Simple alkanethiol groups for temporary blocking of sulfhydryl groups of enzymes
Biochemistry
1975
, vol. 
14
 (pg. 
766
-
771
)
[PubMed]
21
Kanehisa
 
M.
Goto
 
S.
 
KEGG: kyoto encyclopedia of genes and genomes
Nucleic Acids Res.
2000
, vol. 
28
 (pg. 
27
-
30
)
[PubMed]
22
Pei
 
J.
Kim
 
B.-H.
Grishin
 
N.V.
 
PROMALS3D: a tool for multiple protein sequence and structure alignments
Nucleic Acids Res.
2008
, vol. 
36
 (pg. 
2295
-
2300
)
[PubMed]
23
Kawabata
 
T.
 
MATRAS: a program for protein 3D structure comparison
Nucleic Acids Res.
2003
, vol. 
31
 (pg. 
3367
-
3369
)
[PubMed]
24
Nureki
 
O.
Vassylyev
 
D.G.
Katayanagi
 
K.
Shimizu
 
T.
Sekine
 
S.
Kigawa
 
T.
Miyazawa
 
T.
Yokoyama
 
S.
Morikawa
 
K.
 
Architectures of class-defining and specific domains of glutamyl-tRNA synthetase
Science
1995
, vol. 
267
 (pg. 
1958
-
1965
)
[PubMed]
25
Schulze
 
J.O.
Masoumi
 
A.
Nickel
 
D.
Jahn
 
M.
Jahn
 
D.
Schubert
 
W.-D.
Heinz
 
D.W.
 
Crystal structure of a non-discriminating glutamyl-tRNA synthetase
J. Mol. Biol.
2006
, vol. 
361
 (pg. 
888
-
897
)
[PubMed]
26
Ito
 
T.
Kiyasu
 
N.
Matsunaga
 
R.
Takahashi
 
S.
Yokoyama
 
S.
 
Structure of nondiscriminating glutamyl-tRNA synthetase from Thermotoga maritima
Acta Crystallogr. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
813
-
820
)
[PubMed]
27
Campanacci
 
V.
Dubois
 
D.Y.
Becker
 
H.D.
Kern
 
D.
Spinelli
 
S.
Valencia
 
C.
Pagot
 
F.
Salomoni
 
A.
Grisel
 
S.
Vincentelli
 
R.
, et al 
The Escherichia coli YadB gene product reveals a novel aminoacyl-tRNA synthetase like activity
J. Mol. Biol.
2004
, vol. 
337
 (pg. 
273
-
283
)
[PubMed]
28
Blaise
 
M.
Olieric
 
V.
Sauter
 
C.
Lorber
 
B.
Roy
 
B.
Karmakar
 
S.
Banerjee
 
R.
Becker
 
H.D.
Kern
 
D.
 
Crystal structure of glutamyl-queuosine tRNAAsp synthetase complexed with L-glutamate: structural elements mediating tRNA-independent activation of glutamate and glutamylation of tRNAAsp anticodon
J. Mol. Biol.
2008
, vol. 
381
 (pg. 
1224
-
1237
)
[PubMed]
29
Lamour
 
V.
Quevillon
 
S.
Diriong
 
S.
N'Guyen
 
V.C.
Lipinski
 
M.
Mirande
 
M.
 
Evolution of the Glx-tRNA synthetase family: the glutaminyl enzyme as a case of horizontal gene transfer
Proc. Natl. Acad. Sci. U.S.A.
1994
, vol. 
91
 (pg. 
8670
-
8674
)
[PubMed]
30
Nureki
 
O.
O'Donoghue
 
P.
Watanabe
 
N.
Ohmori
 
A.
Oshikane
 
H.
Araiso
 
Y.
Sheppard
 
K.
Soll
 
D.
Ishitani
 
R.
 
Structure of an archaeal non-discriminating glutamyl-tRNA synthetase: a missing link in the evolution of Gln-tRNAGln formation
Nucleic Acids Res.
2010
, vol. 
38
 (pg. 
7286
-
7297
)
[PubMed]
31
Salazar
 
J.C.
Ahel
 
I.
Orellana
 
O.
Tumbula-Hansen
 
D.
Krieger
 
R.
Daniels
 
L.
Söll
 
D.
 
Coevolution of an aminoacyl-tRNA synthetase with its tRNA substrates
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
13863
-
13868
)
[PubMed]
32
Baugh
 
L.
Gallagher
 
L.A.
Patrapuvich
 
R.
Clifton
 
M.C.
Gardberg
 
A.S.
Edwards
 
T.E.
Armour
 
B.
Begley
 
D.W.
Dieterich
 
S.H.
Dranow
 
D.M.
, et al 
Combining functional and structural genomics to sample the essential Burkholderia structome
PLoS One
2013
, vol. 
8
 pg. 
e53851
 
[PubMed]
33
Paravisi
 
S.
Fumagalli
 
G.
Riva
 
M.
Morandi
 
P.
Morosi
 
R.
Konarev
 
P.V.
Petoukhov
 
M.V.
Bernier
 
S.
Chênevert
 
R.
Svergun
 
D.I.
, et al 
Kinetic and mechanistic characterization of Mycobacterium tuberculosis glutamyl-tRNA synthetase and determination of its oligomeric structure in solution
FEBS J.
2009
, vol. 
276
 (pg. 
1398
-
1417
)
[PubMed]
34
Sekine
 
S.
Shichiri
 
M.
Bernier
 
S.
Chênevert
 
R.
Lapointe
 
J.
Yokoyama
 
S.
 
Structural bases of transfer RNA-dependent amino acid recognition and activation by glutamyl-tRNA synthetase
Struct. Lond. Engl. 1993
2006
, vol. 
14
 (pg. 
1791
-
1799
)
35
Sekine
 
S.-I.
Nureki
 
O.
Dubois
 
D.Y.
Bernier
 
S.
Chênevert
 
R.
Lapointe
 
J.
Vassylyev
 
D.G.
Yokoyama
 
S.
 
ATP binding by glutamyl-tRNA synthetase is switched to the productive mode by tRNA binding
EMBO J.
2003
, vol. 
22
 (pg. 
676
-
688
)
[PubMed]
36
Torrance
 
J.W.
Macarthur
 
M.W.
Thornton
 
J.M.
 
Evolution of binding sites for zinc and calcium ions playing structural roles
Proteins
2008
, vol. 
71
 (pg. 
813
-
830
)
[PubMed]
37
Baglivo
 
I.
Russo
 
L.
Esposito
 
S.
Malgieri
 
G.
Renda
 
M.
Salluzzo
 
A.
Di Blasio
 
B.
Isernia
 
C.
Fattorusso
 
R.
Pedone
 
P.V.
 
The structural role of the zinc ion can be dispensable in prokaryotic zinc-finger domains
Proc. Natl. Acad. Sci. U.S.A.
2009
, vol. 
106
 (pg. 
6933
-
6938
)
[PubMed]
38
Curnow
 
A.W.
Hong
 
K.W.
Yuan
 
R.
Kim
 
S. i
Martins
 
O.
Winkler
 
W.
Henkin
 
T.M.
Söll
 
D.
 
Glu-tRNAGln amidotransferase: a novel heterotrimeric enzyme required for correct decoding of glutamine codons during translation
Proc. Natl. Acad. Sci. U.S.A.
1997
, vol. 
94
 (pg. 
11819
-
11826
)
[PubMed]
39
Lapointe
 
J.
Duplain
 
L.
Proulx
 
M.
 
A single glutamyl-tRNA synthetase aminoacylates tRNAGlu and tRNAGln in Bacillus subtilis and efficiently misacylates Escherichia coli tRNAGln1 in vitro
J. Bacteriol.
1986
, vol. 
165
 (pg. 
88
-
93
)
[PubMed]
40
Ito
 
T.
Yokoyama
 
S.
 
Two enzymes bound to one transfer RNA assume alternative conformations for consecutive reactions
Nature
2010
, vol. 
467
 (pg. 
612
-
616
)
[PubMed]
41
Rogers
 
K.C.
Söll
 
D.
 
Discrimination among tRNAs intermediate in glutamate and glutamine acceptor identity
Biochemistry
1993
, vol. 
32
 (pg. 
14210
-
14219
)
[PubMed]
42
Becker
 
H.D.
Kern
 
D.
 
Thermus thermophilus: a link in evolution of the tRNA-dependent amino acid amidation pathways
Proc. Natl. Acad. Sci. U.S.A.
1998
, vol. 
95
 (pg. 
12832
-
12837
)
[PubMed]
43
Nakamura
 
A.
Yao
 
M.
Chimnaronk
 
S.
Sakai
 
N.
Tanaka
 
I.
 
Ammonia channel couples glutaminase with transamidase reactions in GatCAB
Science
2006
, vol. 
312
 (pg. 
1954
-
1958
)
[PubMed]
44
Hou
 
S.
Makarova
 
K.S.
Saw
 
J.H.
Senin
 
P.
Ly
 
B.V.
Zhou
 
Z.
Ren
 
Y.
Wang
 
J.
Galperin
 
M.Y.
Omelchenko
 
M.V.
, et al 
Complete genome sequence of the extremely acidophilic methanotroph isolate V4, Methylacidiphilum infernorum, a representative of the bacterial phylum verrucomicrobia
Biol. Direct
2008
, vol. 
3
 pg. 
26
 
[PubMed]
45
Khadem
 
A.F.
Wieczorek
 
A.S.
Pol
 
A.
Vuilleumier
 
S.
Harhangi
 
H.R.
Dunfield
 
P.F.
Kalyuzhnaya
 
M.G.
Murrell
 
J.C.
Francoijs
 
K.-J.
Stunnenberg
 
H.G.
, et al 
Draft genome sequence of the volcano-inhabiting thermoacidophilic methanotroph Methylacidiphilum fumariolicum strain SolV
J. Bacteriol.
2012
, vol. 
194
 (pg. 
3729
-
3730
)
[PubMed]

Author notes

1

These authors contributed equally to the article.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/) which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Supplementary data