Adaptation of a eukaryote-like ProRS to a prokaryote-like tRNA^Pro

Abstract

Prolyl-tRNA synthetases (ProRSs) are unique among aminoacyl-tRNA synthetases (aaRSs) in having two distinct structural architectures across different organisms: prokaryote-like (P-type) and eukaryote/archaeon-like (E-type). Interestingly, Bacillus thuringiensis harbors both types, with P-type (BtProRS1) and E-type ProRS (BtProRS2) coexisting. Despite their differences, both enzymes are constitutively expressed and functional in vivo. Similar to BtProRS1, BtProRS2 selectively charges the P-type tRNA^Pro and displays higher halofuginone tolerance than canonical E-type ProRS. However, these two isozymes recognize the primary identity elements of the P-type tRNA^Pro―G72 and A73 in the acceptor stem―through distinct mechanisms. Moreover, BtProRS2 exhibits significantly higher tolerance to stresses (such as heat, hydrogen peroxide, and dithiothreitol) than BtProRS1 does. This study underscores how an E-type ProRS adapts to a P-type tRNA^Pro and how it may contribute to the bacterium's survival under stress conditions.

Graphical Abstract

Graphical Abstract.

Introduction

Aminoacyl-tRNA synthetases (aaRSs) are a group of essential enzymes responsible for attaching amino acids to their cognate tRNAs, forming aminoacyl-tRNAs. These charged tRNAs are subsequently delivered to ribosomes, where they participate in the deciphering of the genetic code by matching tRNA anticodons with mRNA codons (1). Therefore, the precise aminoacylation of tRNA by aaRSs greatly contributes to the fidelity of protein translation. AaRSs identify their cognate tRNAs by recognizing a specific set of identity elements or determinants inherent to the tRNA molecule. These identity elements may include individual nucleotides, complementary base pairs, or post-transcriptional modifications within the tRNA molecule (2–4), which most often reside in the acceptor stem or anticodon loop of tRNA. For example, alanine tRNA (tRNA^Ala) carries a primary identity element, G3:U70, in the acceptor stem (5,6), while histidine tRNA (tRNA^His) carries an extra G at position −1 as the identity element (7,8).

Prolyl-tRNA synthetase (ProRS) belongs to class II synthetases and is responsible for attaching proline to the 3′-terminus of tRNA^Pro (9). Class II aaRSs contain three conserved sequence motifs (motifs 1, 2 and 3) in their catalytic domains (10). The primary structure of ProRS comprises a catalytic domain followed by an anticodon-binding domain. Based on phylogenetic studies and sequence alignments, ProRS can be categorized into two distinct forms: the eukaryote/archaeon-like type (E-type) and the prokaryote-like type (P-type) (11,12). The former is distinguished by an additional C-terminal extension domain attached to the anticodon-binding domain, while the latter is characterized by an extra insertion domain (with approximately 180 residues) located between the motifs 2 and 3 of the catalytic domain (11). The insertion domain is associated with the editing process, whereas the C-terminal extension domain is involved in zinc binding. Almost all E-type ProRSs feature a highly conserved Tyr residue at their C-termini, which has been shown to interact with and stabilize the aminoacylation active site (13). In addition, a unique Ybak domain is attached N-terminally to the catalytic domain of yeast ProRS (14), and GluRS and ProRS are fused to form a bifunctional synthetase, GluProRS, in metazoan (15).

Previous studies show that human and E. coli ProRSs are each specific for their own cognate tRNAs^Pro and cannot cross-acylate their respective tRNAs. While recognition of the anticodon of tRNA^Pro has remained unchanged between the two distinct ProRSs, changes in recognition of the acceptor stem have occurred. In E. coli, G35 and G36 in the anticodon, along with G72 and A73 in the acceptor stem play a crucial role in recognition of tRNA^Pro by ProRS (16). Sequence analysis indicates that G72 and A73 are conserved in all bacterial tRNAs^Pro. In contrast, in eukarya and archaea, only G35 and G36 in the anticodon play a crucial role in recognition of tRNA^Pro by ProRS. The nucleotides at positions 72 and 73 have diverged to CC in eukaryotic tRNA^Pro and CA in archaeal tRNA^Pro. C72/C73 play little role in the recognition of eukaryotic tRNA^Pro, while C72/A73 play a minor role in the recognition of archaeal tRNA^Pro (17,18). In addition to the primary identity elements situated in the acceptor stem and anticodon, evidence suggests that certain minor determinants are embedded in the D-arm and anticodon-arm (18,19).

Most bacteria harbor one P-type ProRS and one or several P-type tRNAs^Pro. However, it has been estimated that approximately 22% of bacteria possess an E-type ProRS (20). There are numerous reports of genomes with more than one gene for the same aaRS enzyme. The presence of more than one aaRS gene of the same enzyme is usually the result of horizontal gene transfer or gene duplication (21,22). Bacillus thuringiensis, a Gram-positive facultative aerobic, spore-forming soil bacterium widely used as an insecticide (23), is among the bacteria that contain two ProRS genes of distinct origins. One gene encodes the P-type ProRS (designated herein as BtProRS1), while the other encodes the E-type ProRS (designated herein as BtProRS2) (Figure 1A and Supplementary Figure S1). Although this bacterium encodes four tRNA^Pro isoacceptors, they all belong to the P-type tRNA^Pro (with G72 and A73 in the acceptor stem) (Figure 1B). This finding prompted us to ask whether BtProRS2 also charges the P-type tRNA^Pro and whether the acquisition of an additional ProRS confers benefits to the growth or survival of the bacterium. Our results show that, despite their distinct evolutionary origins, both BtProRS1 and BtProRS2 are specific for the P-type tRNA^Pro. Moreover, BtProRS2 exhibits significantly higher tolerance toward heat, DTT and H₂O₂ compared to BtProRS1. Thus, continued maintenance of an extra ProRS might contribute to the survival of the bacterium under stress conditions.

Figure 1.

Two BtProRSs and one BttRNA^Pro. (A) Domain organization of ProRS. INS, insertion domain; MTS, mitochondrial targeting signal. A highly conserved Y residue exists at the C-terminus of the E-type ProRS. (B) Cloverleaf structure of tRNA^Pro. Primary identity elements of tRNA^Pro are highlighted in yellow. SctRNA_n^Pro, S. cerevisiae nuclear-encoded cytoplasmic tRNA^Pro; SctRNA_m^Pro, S. cerevisiae mitochondrial-encoded mitochondrial tRNA^Pro; BttRNA^Pro, B. thuringiensis tRNA^Pro. (C) Relative expression levels of proS1 and proS2. Total RNA was isolated from B. thuringiensis cells grown under various conditions. Relative levels of proS1 and proS2 mRNAs were determined by a RT-PCR. Densitometric quantification was performed on the PCR product image using ImageJ software, and the normalized values (with 16S rRNA as the internal control) were plotted. Quantitative data were obtained from three independent experiments and averaged. The red and blue bars denote BtproS1 and BtproS2, respectively.

Materials and methods

Bacterial strains and culture conditions

B. thuringiensis ATCC 13367 was used in this study. This bacterium was grown in Luria–Bertani (LB) broth (10 g/l peptone, 5 g/l NaCl, 5 g/l yeast extract) and LB solid plates with 15 g/l agar at 30°C. Total RNA of B. thuringiensis was isolated from cells grown in LB broth under various conditions for 4 hours: normal (30°C); cold (20°C), hot (37°C), acidic (pH 5.0), and alkaline (pH 9.0). For preparation of total cellular extracts of B. thuringiensis for halofuginone (HF) inhibition assay, cells were grown in 4 mL of LB broth at 30°C and harvested by centrifugation after 24 h incubation by shaking (200 rpm). The cell pellets were washed three times with ice-cold PBS buffer (10 mM, pH 7.8) and then resuspended in 1 ml lysis buffer containing 50 mM HEPES (pH 7.2), 25 mM KCl, 12 mM MgCl₂, 2 mM DTT, 2 mM PMSF and 2 μl protease inhibitor cocktail (100×). The suspension was further disrupted with glass beads (212–300 μm; Sigma) in ice for 10 minutes. Cell debris was removed by centrifugation at 13 000 rpm for at least 15 min, and the supernatant was transferred to another clean tube and stored at −80°C for further experiments.

Construction of plasmids

Cloning of B. thuringiensis proS1 (BtproS1) (UJP58595.1, GenBank: CP074714.1, locus tag: JRY14_21045) into pTEF1-C-His₆ (a high-copy-number yeast shuttle vector with a strong TEF1 promoter, a C-terminal His₆ tag, a LEU2 marker, and a 2μ replication origin) for yeast rescue assays followed a standard protocol. In brief, a set of gene-specific primers was designed to amplify the open reading frame of proS1 (without its native stop codon) as an NdeI-XhoI fragment using polymerase chain reaction (PCR), with B. thuringiensis genomic DNA as the template. Cloning of B. thuringiensis proS2 (BtproS2) (UJP60857.1, GenBank: CP074714.1, JRY14_04045) into pTEF1-N-His₆ (a yeast vector with an N-terminal His₆ tag) followed a similar strategy. For protein purification, proS1 (or its mutant) was cloned in pRIC72 (an E. coli expression vector with a T7 promoter and a C-terminal His₆ tag), while proS2 (or its mutants) was cloned in pET19b (an E. coli expression vector with a T7 promoter and an N-terminal His₆ tag). The plasmid construct was transformed into an E. coli expression strain, BL21-CodonPlus(DE3), and the target gene expression was induced with 1 mM IPTG. The resulting N- or C-terminally His₆-tagged protein was purified to homogeneity through Ni-NTA column chromatography (24). Western blotting was conducted following a standard protocol, using an HRP-conjugated anti-His₆ tag antibody as the probe (25).

Rescue of the genetic loss of yeast PROS1 on 5-FOA

To perform functional complementation assays, two yeast haploid knockout strains (PROS1⁻ and PROS2⁻) were constructed in this research following standard protocols (26,27). To determine whether a heterologous ProRS can functionally substitute for yeast cytoplasmic ProRS, a test plasmid carrying the target gene was introduced into the haploid PROS1 knockout strain that harbors a maintenance plasmid carrying a wild-type (WT) yeast PROS1 gene and a URA3 marker. Cultures with an initial optical density (A₆₀₀) of 1.0 underwent three-fold serial dilutions, and 10-μl aliquots from each dilution were spotted onto a 5-FOA (1 mg/ml) plate (28). The plate was then subsequently incubated at 30°C for 3–5 days. As 5-FOA can be converted to a toxic compound by yeast in the presence of URA3, transformants must evict the maintenance plasmid in order to survive on this plate. As a result, transformants can only grow on 5-FOA plate when the test plasmid encodes a functional ProRS that can charge yeast cytoplasmic tRNA^Pro to a significant level.

Rescue of the genetic loss of yeast PROS2 on YPG

To determine whether a heterologous ProRS can functionally substitute for yeast mitochondrial ProRS, a test plasmid containing the target gene was introduced into the haploid PROS2 knockout strain that harbors a maintenance plasmid with a WT yeast PROS2 gene and a URA3 marker. The resulting transformants were then plated on a medium containing 5-FOA (1 mg/ml) to evict the maintenance plasmid (28). Subsequent to the 5-FOA selection, the surviving transformants were further spotted onto a YPG plate and incubated at 30°C for 3–5 days to assess their growth capabilities. As glycerol is the only carbon source in YPG, yeast needs functional mitochondria to metabolize this non-fermentable organic compound through oxidative phosphorylation. Therefore, transformants can only grow on a YPG plate when the test plasmid encodes a functional ProRS that can charge yeast mitochondrial tRNA^Pro to a significant level.

Reverse-transcription (RT)-PCR

To determine the relative levels of specific proS1 and proS2 mRNAs expressed in B. thuringiensis, RT-PCR were carried out following the manufacturer's protocols (Invitrogen, Carlsbad, CA). Total RNA was isolated from the bacterium after growth under various conditions for 4 hours, and DNase I treatment was applied to remove contaminating DNA. Aliquots (3 μg) of the RNA were then reverse-transcribed into single-stranded complementary (c)DNA using an oligo-dT primer. After RNase H treatment, the single-stranded cDNA products were amplified by PCR using two sets of gene-specific primers: one set (KAN17 and KAN18) for proS1 and the other set (KAN19 and KAN20) for proS2. KAN17 (a forward primer) (5′-AGCAAAGACCTCGTTTCGGC) is complementary to nucleotides +425 to +444 of proS1; KAN18 (a reverse primer) (5′-TGCACGGAAATTCAAGCCAC) is complementary to nucleotides +563 to +582 of proS1; KAN19 (a forward primer) (5′-TATCCCAGAGAGTTTATTGC) is complementary to nucleotides +219 to +237 of proS2; KAN20 (a reverse primer) (5′-GCACAATTTTTGAAAAGTGC) is complementary to nucleotides +360 to +379 of proS2. As a control, genomic DNA from B. thuringiensis was isolated and used as a template for PCR amplification of proS1 and proS2. The relative mRNA levels of proS1 and proS2 were normalized to those of 16S rRNA.

Preparation of tRNA^Pro transcripts

In vitro transcription of tRNA^Pro followed a previously described protocol (6). The transcription template was enriched by PCR amplification of the insert, containing a T7 promoter and the designated tRNA gene. The in vitro transcription reaction for tRNA^Pro was carried out with 0.3 μM T7 RNA polymerase at 37°C for 3 hours in a buffer comprising 20 mM Tris–HCl (pH 8.0), 150 mM NaCl, 20 mM MgCl₂, 5 mM dithiothreitol (DTT), 1 mM spermidine, and 2 mM of each NTP. The tRNA^Pro transcript was then purified using an 8 M urea–10% polyacrylamide gel electrophoresis. Following ethanol precipitation and vacuum-drying, the tRNA pellet was dissolved in 1 × TE buffer (20 mM Tris–HCl, pH 8.0, and 1 mM EDTA) and refolded by heating to 80°C, gradually cooling to room temperature after the addition of 10 mM MgCl₂. Approximately 80% of the in vitro-transcribed tRNA^Pro were active in aminoacylation.

Aminoacylation assay

Aminoacylation assay was carried out in a buffer containing 50 mM HEPES (pH 7.2), 0.2 mg/ml BSA, 25 mM KCl, 12 mM MgCl₂, 2 mM DTT, 2 mM 2-mercaptoethanol, 1 mM spermine, 4 mM ATP, 10 μM in vitro-transcribed tRNA^Pro, and 21.13 μM proline (1.13 μM ³H-proline; PerkinElmer, Waltham, MA, USA) (29). To stop the reactions, 10-μl aliquots of the reaction mixture were applied to Whatman filters (Maidstone, Kent, UK) pre-soaked in 5% trichloroacetic acid (TCA) and 2 mM proline. Subsequently, the filters were washed three times in ice-cold 5% TCA, 15-min each, before liquid scintillation counting. The data represent the averages obtained from three independent experiments. Active protein concentrations were determined by active site titration as previously described (30).

To determine the inhibitory effects of H₂O₂, DTT, urea and GnHCl on BtProRSs, the tested enzymes (at a concentration of 1000 nM) were pretreated with H₂O₂ (0–32 mM), DTT (0–80 mM), urea (0–3200 mM), and GnHCl (0–3200 mM) for 10 min at 20°C. The pretreated enzymes were then added to the aminoacylation buffer (at a final concentration of 100 nM) and the mixture was incubated at 30°C for 10 min before liquid scintillation counting.

Kinetic parameters for the prolylation of tRNA were determined by quantifying the initial rate of charging over the first 2 min. The aminoacylation assay was carried out at 30°C with BttRNA^Pro concentrations ranging from 1 to 32 μM and BtProRS concentrations ranging from 100 to 1000 nM. The parameters were derived from Lineweaver–Burk plots. Error values represent standard deviations. Data were obtained from three independent experiments and averaged (29,31).

HF inhibition assay

For the HF inhibition assay, HF (0.1 nM to 10 μM) was mixed with the tested enzyme (at a final concentration of 100 nM) or 3 μg of total cellular extracts in a 50-μl aminoacylation buffer and incubated at 30°C for 10 min. Reactions were stopped as described above. The relative aminoacylation activity was determined by the ratio of the reaction rate at different inhibitor concentrations to the one with no inhibitor. This can be represented by the following equation:

where is the relative activity as a function of the inhibitor concentration , is the reaction rate at different inhibitor concentrations, and is the reaction rate when no inhibitor is present.

Since the reaction rate is determined by the rate of product (Pro-tRNA^Pro) production, in an isozyme mixture (B. thuringiensis total extract), this can be considered as the sum of each enzyme's catalytic rate:

where is the total product (Pro-tRNA^Pro) production rate, and and are the contributions of BtProRS1 and BtProRS2, respectively. Therefore, the relative aminoacylation activity of this mixture () is

Here, is defined as the activity proportion of BtProRS1. and are the relative aminoacylation activity curves of BtProRS1 and BtProRS2, respectively.

The curves and were obtained by fitting the results of individual BtProRS1 and BtProRS2 in this research, respectively. The ‘’ value was obtained by fitting the mixture's result using the parameters of the individual and curves.

Results

Both BtProRS1 and BtProRS2 are constitutively expressed in vivo

Unlike most other bacteria, the soil bacterium B. thuringiensis encodes two ProRSs of different evolutionary origins, but it only encodes the P-type tRNA^Pro (carrying bacterial-specific G72 and A73 in the acceptor stem) (Figure 1AB). BtProRS1 possesses an insertion domain (of 190 amino acid residues) situated between the motifs 2 and 3 of the catalytic domain, while BtProRS2 possesses a unique C-terminal extention domain (of 76 amino acid residues) with a highly conserved Tyr at its C-terminus. To investigate whether B. thuringiensis proS1 (which encodes ProRS1) and proS2 (which encodes ProRS2) are both constitutively expressed in vivo, we carried out an RT-PCR using B. thuringiensis cDNA as the template. As shown in Figure 1C, both proS1 and proS2 were expressed under normal growth conditions (defined herein as growth at 30°C and pH 7.0), with proS1 having an expression level ∼4-fold higher than that of proS2. To explore whether the expression of proS2 is stress-inducible, we selected various environmental stresses commonly encountered in nature for this purpose, including cold (20°C), heat (37°C), acidity (pH 5.0), and alkalinity (pH 9.0). The results showed that the expression of both genes was slightly enhanced (∼2-fold) under hot and acidic conditions, but none of the selected stresses significantly induced the expression of proS2. It is important to note that this result was obtained under specific laboratory conditions using LB medium. Therefore, we cannot rule out the possibility that certain stimuli might induce the expression of proS2.

To gain deeper insight, we conducted mass spectrometry analysis to determine the relative protein expression levels of BtProRS1 and BtProRS2 under different stress conditions, including heat (42°C), cold (20°C), and oxidative stress (2 mM H₂O₂). Mass spectrometry data showed that the ratio of BtProRS1 to BtProRS2 remained almost constant under these conditions (Supplementary Figure S2), suggesting that BtProRS2 is not inducible under the conditions tested.

Both BtProRS1 and BtProRS2 are specific for the P-type tRNA^Pro

To explore the prolylation activity of BtProRS1 and BtProRS2, these two enzymes were purified (as BtProRS1-His₆ and His₆-BtProRS2) through Ni-NTA column chromatography and assayed using in vitro-transcribed BttRNA^Pro as the substrate. To determine the optimal temperature for aminoacylation by BtProRSs, we carried out the assay between 20°C and 50°C. As shown in Figure 2A, the aminoacylation activity of BtProRS1 was optimal at 30°C and increasing temperature to 40°C reduced the aminoacylation activity up to ∼8-fold. The enzyme was essentially inactive at 50°C, suggesting a heat-sensitive nature for the enzyme. In contrast, BtProRS2 showed relatively high tolerance toward heat. BtProRS2 was rather active at 30°C, but it was even more active at 40°C. The activity decreased only ∼2-fold when the temperature dropped to 50°C (Figure 2B). Based on these results, we chose 30°C as the temperature for all subsequent aminoacylation assays involving these two enzymes.

Figure 2.

Aminoacylation activities of BtProRS1 and BtProRS2. (A) Aminoacylation of BttRNA^Pro by BtProRS1 at various temperatures. (B) Aminoacylation of BttRNA^Pro by BtProRS2 at various temperatures. (C) Aminoacylation of TgtRNA_n^Pro by various ProRSs. (D) Aminoacylation of TttRNA^Pro by various ProRSs. Aminoacylation by BtProRSs (100 nM) was carried out at 30°C, while aminoacylation by TgProRS (100 nM) and TtProRS (20 nM) was carried out at 37°C.

To further verify the tRNA preferences of these two enzymes, we chose an E-type tRNA^Pro, Toxoplasma gondii cytoplasmic tRNA^Pro (TgtRNA_n^Pro), and a P-type tRNA^Pro, Thermus thermophilus tRNA^Pro (TttRNA^Pro), as the substrates (Supplementary Figure S3). T. gondii tRNA_n^Pro was chosen because it shares similar nucleotide sequences and identity elements with other canonical E-type tRNAs^Pro, while T. thermus tRNA^Pro was a well-studied substrate for ProRS. As shown in Figure 2C, D, both enzymes effectively charged TttRNA^Pro, but neither charged TgtRNA_n^Pro to a detectable level. This result suggests that, despite their distinct evolutionary origins, both enzymes are specific for the P-type tRNA^Pro (with G72 and A73).

BtProRS1 and BtProRS2 can rescue the genetic loss of yeast PROS2, but not PROS1

As BtProRS1 and BtProRS2 specifically charge the P-type tRNA^Pro in vitro (Figure 2), we were curious about their functional potential to substitute for yeast cytoplasmic and mitochondrial ProRSs (ScProRS_c and ScProRS_m, respectively) in vivo. To this end, we constructed two yeast haploid KO strains (PROS1⁻ and PROS2⁻) following the basic principles depicted in Figure 3A. PROS1 and PROS2 are the yeast nuclear genes respectively encoding ScProRS_c (an E-type enzyme) and ScProRS_m (a P-type enzyme) (Figure 1A). Yeast nuclear-encoded cytoplasmic tRNA^Pro (SctRNA_n^Pro) belongs to the E-type tRNA^Pro (with C72 and C73 in the acceptor stem), while yeast mitochondrial-encoded mitochondrial tRNA^Pro (SctRNA_m^Pro) belongs to the P-type tRNA^Pro (with G72 and A73 in the acceptor stem) (Figure 1B).

Figure 3.

Rescue of the genetic loss of yeast PROS1 or PROS2. (A) Construction of haploid yeast PROS1⁻ and PROS2⁻ strains. (B) Summary of the ProRS constructs and their rescue activities. The symbols ‘+’ and ‘–’ respectively denote positive and negative complementation. Growth on 5-FOA and YPG respectively indicates complementation of the cytoplasmic and mitochondrial ProRS activities. (C) Western blotting. Expression of N- or C-terminally His₆-tagged ProRSs from the plasmids was probed with an HRP-conjugated anti-His₆ tag antibody. Numbers 1∼9 (circled) represent constructs shown in (B).

To verify the growth phenotypes of these two haploid KO strains, we first tried the positive controls ScProRS_c and ScProRS_m. As shown in Figure 3B, ScProRS_c effectively rescued the growth defect of the PROS1 KO strain but not the PROS2 KO strain, regardless of whether a heterologous MTS was fused to the enzyme. It is likely that, even with an MTS, a minor portion of the enzyme still remained in the cytoplasm, presumably due to the overexpression of the enzyme and the overloading of the mitochondrial import machinery, a scenario often seen in aaRS complementation (24,32). In contrast, ScProRS_m (deleted of its native MTS) effectively rescued the growth defect of the PROS2 KO strain, but not the PROS1 KO strain, regardless of whether a heterologous MTS was fused to the enzyme. It is likely that ScProRS_m (lacking its native MTS) carried a cryptic MTS in its polypeptide sequence, enabling a small fraction of the enzyme to be imported into mitochondria for functioning, a scenario frequently observed in aaRS complementation (6,24).

As anticipated, neither BtProRS1 nor BtProRS2 successfully rescued the genetic loss of the yeast PROS1 KO strain, regardless of whether a heterologous MTS was fused to the enzymes. However, the fusion of an MTS to each of these two enzymes enabled the fusion enzymes to rescue the growth defect of the PROS2 KO strain. This result suggests that BtProRS1 and BtProRS2 can charge the yeast mitochondrial tRNA^Pro to a substantial level. In addition, it implies that no cryptic MTS is embedded in the polypeptide sequences of the bacterial enzymes, and therefore, a heterologous MTS is required for efficient import into mitochondria. These findings align with the in vitro data mentioned above and reinforce the conclusion that both BtProRS1 and BtProRS2 are specific for the P-type tRNA^Pro. Western blotting confirmed the proper expression of all the tested constructs in the yeast KO strains (Figure 3C).

BtProRS1 and BtProRS2 recognize the same identity elements in BttRNA^Pro

Previous research indicated that G72 and A73 in the acceptor stem, and to a lesser extent G35 and G36 in the anticodon, serve as the identity elements for the P-type tRNA^Pro (16). In contrast, recognition of the E-type tRNA^Pro depends mainly on the anticodon, without direct involvement of the bases in the acceptor stem (18). To examine whether G35, G36, G72, and A73 of BttRNA^Pro also play as the identity elements, base substitutions were made at these nucleotides and the resulting tRNA mutants were used as the substrates for aminoacylation. In our assays, BttRNA^Pro ΔC1 was used to achieve high yields of tRNA transcripts. Previous research showed that deletion of C1 from tRNA^Pro has only minimal effects on aminoacylation (16). Similarly, deletion of C1 from BttRNA^Pro resulted in only 2-fold increase in its specificity constant (k_cat/K_M) (Table 1).

Table 1.

Kinetic parameters for aminoacylation of BttRNA^Pro variants by BtProRSs

BttRNAPro	BtProRS1		BtProRS2
	k cat/KM (x 10−3 μM−1 s−1)	Loss of specificity (x-fold)	k cat/KM (x 10−3 μM−1 s−1)	Loss of specificity (x-fold)
ΔC1	27 ± 6.0	1	22 ± 4.2	1
ΔC1-G35→C35	0.9 ± 0.2	30	3.4 ± 0.7	6.5
ΔC1-G35→U35	3.0 ± 0.4	9	3.6 ± 0.9	6.1
ΔC1-G35→A35	2.0 ± 0.3	13.5	2.3 ± 0.1	9.6
ΔC1-G36→C36	0.5 ± 0.1	54	1.3 ± 0.3	17
ΔC1-G36→U36	1.0 ± 0.2	27	3.7 ± 1.0	5.9
ΔC1-G36→A36	3.2 ± 0.8	8.4	2.5 ± 0.6	8.8
ΔC1-G72→C72	0.2 ± 0.01	135	2.3 ± 0.5	9.6
ΔC1-G72→U72	0.1 ± 0.01	270	0.9 ± 0.2	24
ΔC1-G72→A72	2.0 ± 0.4	13.5	8.2 ± 1.8	2.7
ΔC1-A73→C73	0.5 ± 0.06	54	0.8 ± 0.3	27.5
ΔC1-A73→U73	0.4 ± 0.08	67.5	2.5 ± 0.5	8.8
ΔC1-A73→G73	0.3 ± 0.02	90	1.3 ± 0.2	17

The specificity constants (k_cat/K_M) for BtProRS1 and BtProRS2 towards WT BttRNA^Pro (with C1) are 14 ± 2.0 and 13 ± 4.0, respectively.

As shown in Table 1, when BtProRS1 was used as the test enzyme, mutation at G35 reduced the aminoacylation efficiency (k_cat/K_M) by 9 to 30-fold, with G35C having the strongest effect; mutation at G36 reduced the aminoacylation efficiency by 8.4- to 54-fold, with G36C having the most significant effect; mutation at G72 reduced the aminoacylation efficiency by 13.5- to 270-fold, with G72U having the strongest effect; and mutation at A73 reduced the aminoacylation efficiency by 54 to 90-fold, with A73G having the strongest effect. Thus, all these four nucleotides play a role in BttRNA^Pro recognition by BtProRS1, with the following order of significance: G72 > A73 > G36 > G35. A similar scenario was seen when BtProRS2 was used as the test enzyme. Mutation at G35 reduced the aminoacylation efficiency 6.1- to 9.6-fold, with G35A having the strongest effect; mutation at G36 reduced the aminoacylation efficiency 5.9- to 17-fold, with G36C having the strongest effect; mutation at G72 reduced the aminoacylation efficiency 2.7- to 24-fold, with G72U having the strongest effect; and mutation at A73 reduced the aminoacylation efficiency 8.8- to 27.5-fold, with A73C having the strongest effect (Table 1). We noticed that mutations in these identity elements appear to have a stronger impact on BtProRS1 than on BtProRS2. It is likely that BtProRS2 recognizes additional identity elements existing elsewhere in the tRNA (18,19). Individual k_cat and K_M values were reported in Supplementary Table S1.

Recognition of the identity elements G72/A73 by BtProRS2 through a distinct mechanism

An earlier study demonstrated that R144 in the motif 2 loop of E. coli ProRS plays a key role in recognition of the primary identity elements G72 and A73 in the acceptor stem of tRNA^Pro (19). Sequence alignment unveiled that this R residue is highly conserved in the motif 2 loop of the P-type ProRSs, including BtProRS1. As expected, mutation of this conserved R to A (resulting in R144A) in the motif 2 loop of BtProRS1 resulted in a drastic decrease in aminoacylation efficiency, with a remarkable 270-fold reduction (Table 2). This finding reinforces the critical role of this conserved R residue in recognition of the P-type tRNA^Pro.

Table 2.

Kinetic parameters for aminoacylation of BttRNA^Pro by BtProRS1 and BtProRS2 variants

Enzymes	BttRNAPro ΔC1
	k cat/KM (x 10−3 μM−1 s−1)	Loss of specificity (x-fold)
BtProRS1
WT	27 ± 6.1	1
R144A	0.1 ± 0.02	270
BtProRS2
WT	23 ± 3.2	1
T148A	5.2 ± 0.3	4.4
T149A	0.2 ± 0.07	115
R150A	0.04 ± 0.01	575

As expected, the P-type-specific R residue is absent from the motif 2 loop of BtProRS2 (Supplementary Figure S4). This raised the question of how this E-type enzyme recognizes the bacterial P-type tRNA^Pro. To assess whether motif 2 loop of BtProRS2 is important for aminoacylation of BttRNA^Pro, we targeted the amino acid residues (T148, T149, and R150) that are highly diverged from P-type ProRSs. Mutation of T148 to A only reduced the aminoacylation efficiency (k_cat/K_M) 4.4-fold. Surprisingly, mutation of T149 or R150 to A dramatically reduced the aminoacylation efficiency, with reductions of 115 and 575-fold, respectively (Table 2 and Supplementary Table S2). Thus, T149 and R150 in the motif 2 loop of BtProRS2 play an important role in tRNA aminoacylation.

BtProRS2 is much more resistant than a typical E-type ProRS toward HF

HF, a synthetic derivative of febrifugine (FF), is one of the most potent antimalarials (33). It binds to the L-Pro and tRNA^Pro sites in Plasmodium falciparum cytoplasmic ProRS (which belongs to the E-type ProRS) (34). In contrast, HF is much less effective (∼100-fold) against a P-type ProRS (35). This raises the question of whether the E-type BtProRS2 still maintains its high sensitivity toward HF.

To gain more insight into HF binding by ProRS, we performed sequence alignment and structural prediction using AlphaFold2 (36). The HF binding pockets of BtProRS1 and BtProRS2 were individually superimposed on that of Plasmodium falciparum cytoplasmic ProRS (PfProRS_c) (PDB entry 4YDQ). Structural predictions indicated that the amino acid residues likely to bind HF are highly conserved in the catalytic site of BtProRS2 (Supplementary Figure S5A) but poorly conserved in the catalytic site of BtProRS1 (Supplementary Figure S5B). Significant differences were observed in many residues in BtProRS1, such as H110, M202, Y441 and I443 (Supplementary Figure S5B). Despite this, HF inhibition assay demonstrated that BtProRS2 has a half-maximal inhibitory concentration (IC₅₀) value only 4-fold lower than that of BtProRS1 (with IC₅₀ values of 170 nM for BtProRS2 and 680 nM for BtProRS1) (Figure 4). To trace whether both enzymes contribute to the in vivo prolylation activity, we next determined the IC₅₀ value of HF for the total protein extracts of B. thuringiensis. Our result showed that the IC₅₀ value of HF for the total cellular extracts was around 400 nM, implying the collaborative involvement of both isozymes in the overall aminoacylation activity of B. thuringiensis, with an activity ratio of 2:1 for BtProRS1 to BtProRS2 in total extracts extracted from cells grown under normal conditions (see Materials and Methods for details on calculation).

Figure 4.

Inhibition of BtProRS1 and BtProRS2 by HF. Aminoacylation assays were carried out at 30°C for 10 min with varying HF concentrations (0.1 nM to 10 μM). To determine the relative contributions of BtProRS1 and BtProRS2 in vivo, total protein extracts were isolated from a cell culture of B. thuringiensis and 3 μg of the cellular extracts was added into a 50-μl aminoacylation buffer.

BtProRS2 exhibits a higher tolerance toward stresses than does BtProRS1

Hydrogen peroxide (H₂O₂) can induce oxidative stress, leading to varied effects on aaRS function across different organisms (37). Dithiothreitol (DTT) is a powerful reducing agent that induces acute ER stress by disrupting the redox conditions required to form disulfide bridges in proteins (38). Urea and guanidine hydrochloride (GnHCl) can induce denaturation of proteins (39,40). To assess whether BtProRS1 and BtProRS2 exhibit different sensitivities toward these stresses and denaturants, we pretreated the enzymes with these agents for 10 min before analyzing their aminoacylation activity.

Our results showed that BtProRS2 exhibits greater tolerance than BtProRS1 to H₂O₂ treatment, with IC₅₀ values of 2.6 mM for BtProRS1 and larger than 32 mM for BtProRS2 (Figure 5A). A similar scenario was observed for DTT, with IC₅₀ values of 27 mM for BtProRS1 and larger than 80 mM for BtProRS2 (Figure 5B). In contrast, both enzymes exhibited similar sensitivities to the denaturants urea and GnHCl. The IC₅₀ values of urea were 230 mM for BtProRS1 and 180 mM for BtProRS2 (Figure 5C), while the IC₅₀ values of GnHCl were 270 mM for BtProRS1 and 200 mM for BtProRS2 (Figure 5D).

Figure 5.

Tolerance of BtProRS1 and BtProRS2 to H₂O₂, DTT, urea, and GnHCl. BtProRS1 and BtProRS2 were pretreated with (A) H₂O₂ (0–32 mM), (B) DTT (0–80 mM), (C) urea (0–3200 mM) and (D) GnHCl (0–3200 mM) for 10 min at 20°C before being tested for aminoacylation activities. Aminoacylation was carried out by adding the pretreated enzymes (at a final concentration of 100 nM) to the aminoacylation buffer for 10 min at 30°C.

To further evaluate the stress tolerance of BtProRS1 and BtProRS2, we cloned the genes encoding these two proteins and integrated them into the chromosome of the B. subtilis 168 strain. The resulting strains were designated as Bs-s (amyE::spec), Bs-Bt1 (amyE::spec-BtproS1), and Bs-Bt2 (amyE::spec-BtproS2). Upon obtaining these strains, we assessed the impact of these two proteins on the stress endurance of the resulting strains through growth curve measurements conducted under normal (37°C), cold (20°C), hot (45°C), and oxidative stress (2 mM of H₂O₂) conditions. Our results showed that the introduction of BtProRS2 did not significantly alter the growth phenotypes of B. subtilis under the conditions used (Supplementary Figure S6). However, unexpectedly, the introduction of BtProRS1 led to an increase in its endurance toward cold temperature. This suggests that BtProRS2 alone may not suffice to enhance the bacterium's survival capacity under the conditions tested, indicating the potential involvement of additional factors in determining its resilience.

Discussion

BtProRS2 is an E-type enzyme but has adapted to the P-type tRNA^Pro

E. coli and human ProRSs recognize their respective cognate tRNAs^Pro through distinct mechanisms. In E. coli, the key identity elements for recognition of tRNA^Pro include G72 and A73 in the acceptor stem, and, to a lesser extent, anticodon G35 and G36 (16). Despite the highly conserved nature of C72 and C73 among eukaryotic cytoplasmic tRNAs^Pro, the human cytoplasmic ProRS disregards these two nucleotides in the acceptor stem. Instead, the human enzyme appears to contact the backbone rather than the bases of the acceptor stem (41). As a result, human cytoplasmic ProRS recognizes only anticodon G35 and G36 as the primary identity elements (18). A similar scenario has been observed in the archaeal system, where only anticodon G35 and G36 are recognized as the primary identity elements (17). It is believed that sequence elements within the respective synthetases have coevolved with their cognate tRNAs to accommodate changes in the acceptor stem throughout evolution (19). Like a P-type ProRS, BtProRS2 exhibited a strong preference for the P-type tRNA^Pro (Figures 2 and 3). Mutation at G72 and A73 of BttRNA^Pro drastically reduced aminoacylation by both enzymes (Table 1 and Supplementary Table S1). We summarize the identity elements of the tRNA^Pro recognized by E. coli, human, M. jannaschii and B. thuringiensis in Figure 6.

Figure 6.

Primary identity elements of tRNA^Pro. Acceptor stem (boxed) and anticodon (circled) nucleotide positions examined in this and previous work are indicated. Arrows denote nucleotides that serve as specificity determinants for the respective cognate ProRS. The absence of an arrow denotes that the position is not a primary determinant. The size of an arrow indicates the relative effect of the determinant: a small arrow denotes a minor effect, while a larger arrow denotes a major effect on the k_cat/K_M value of aminoacylation.

BtProRS2 recognizes G72 and A73 through a distinct mechanism

Specific elements within the motif 2 loop of many class II synthetases, including E. coli ProRS, have been shown to participate in recognition of the acceptor stem. For example, the sequence RPR, which is highly conserved in the P-type ProRS, is critical for aminoacylation of tRNA^Pro by E. coli ProRS. Substitution of the first R (R144) in this sequence reduces the k_cat/K_M value of aminoacylation by more than 1000-fold, but has little effect on amino acid activation, suggesting a specific role in tRNA recognition. Cross-linking experiments further show that R144 is proximal to G72 in the acceptor stem of E. coli tRNA^Pro. In contrast, the amino acid residue (K1084) at the corresponding position in human cytoplasmic ProRS does not play a similar role; mutation at this amino acid leads to less than 2-fold reduction in aminoacylation efficiency (19).

Although BtProRS2 carries a sequence feature typical of an E-type ProRS, its motif 2 loop sequence has diverged from the motif 2 loop sequences of eukaryotic cytoplasmic E-type ProRSs. Instead of having the eukaryote-specific sequence FKHPQ, its motif 2 loop possesses the sequence TTR (residues 148–150) at the corresponding positions (Supplementary Figure S4). Mutation at T149 and R150 alone resulted in a dramatic decrease in aminoacylation efficiency (Table 2 and Supplementary Table S2), lending further support to the coevolution hypothesis. From this perspective, it is interesting to mention that Thermus thermophilus ProRS belongs to the E-type ProRS but is specific for the P-type tRNA^Pro. The motif 2 loop of this enzyme carries the sequence RTR instead of the eukaryote-specific sequence FKHPQ at the corresponding positions (17) (Supplementary Figure S4). However, the mechanism by which T. thermophilus ProRS recognizes its cognate tRNA is yet to be determined.

BtProRS2 is much more resistant to HF than a typical E-type ProRS

HF is a synthetic derivative of FF, a natural compound extracted from the Chinese herb Dichroa febrifuga and commonly used in the treatment of fevers and malaria. Over the past few years, several groups have attempted to use HF to target the cytoplasmic ProRSs of Plasmodium falciparum (a parasite that causes malaria) and Toxoplasma gondii (a parasite that causes toxoplasmosis), both of which belong to the E-type ProRS (34). The IC₅₀ values of HF for P. falciparum, T. gondii, and human ProRSs are 9.0 nM, 5.41 nM, and 10.65 nM, respectively (34,42). However, HF is relatively ineffective (∼100-fold) against a P-type ProRS (such as Pseudomonas aeruginosa ProRS) (35). Despite the fact that BtProRS2 carries a sequence feature typical of an E-type ProRS, it behaves more like a P-type enzyme, with an IC₅₀ value only 4-fold lower than that for BtProRS1 (Figure 4). Such a feature may reflect the structural accommodation of the enzyme to the P-type tRNA^Pro.

Acquisition of an E-type ProRS may confer a survival advantage to the bacterium

An extra copy of an aaRS gene can be acquired through horizontal gene transfer or intra-genome duplication. These evolutionary events may occur independently or simultaneously to promote the expansion of aaRSs with novel functions and to facilitate adaptation to environmental changes, such as variations in amino acid supply, metal ions, temperature fluctuations, and exposure to toxic substances. Thus, coexistence of two copies of an aaRS in a bacterium normally reflects the need for distinct enzymatic features. While bacteria typically rely on either ileS1 (encoding IleRS1) or ileS2 (encoding IleRS2) as standalone housekeeping genes, Priestia megaterium (and many other Bacilli) possess two IleRS enzymes, PmIleRS1 and PmIleRS2. Notably, PmIleRS2 demonstrates resistance to the antibiotic mupirocin, in contrast to the sensitivity of PmIleRS1 to this antibiotic (43). Further evidence implies that the acquisition and maintenance of IleRS1 may serve to promote fast translation in Bacilli and other fast-growing bacteria (43). Similarly, Agrobacterium radiobacter K84 produces the antibiotic agrocin 84, which is transformed into the toxin TM84 upon entering A. tumefaciens. TM84 inhibits leucyl-tRNA synthetase (LeuRS) using a unique tRNA-dependent mechanism (44). To prevent self-suicide, the TM84-producing organism expresses a resistant LeuRS called AgnB2, which has evolved to resist TM84 by reducing its binding affinity to tRNA^Leu, but also reduces the effect of tRNA^Leu on the inhibitor's binding for the enzyme, without affecting its overall aminoacylation catalytic activity (45).

Another scenario was found in Bacillus subtilis, where a second, specialized tyrosyl-tRNA synthetase (TyrZ) is induced to protect cells against high concentrations of D-Tyr. Unlike the housekeeping enzyme TyrS, TyrZ exhibits significantly higher selectivity for L-Tyr over D-Tyr. This selective behavior prevents the bacterium from the misincorporation of D-Tyr into proteins (46). E. coli lysyl-tRNA synthetase exemplifies another interesting scenario (47). E. coli contains a constitutive (LysS) and an inducible lysyl-tRNA synthetase (LysU). LysU is more thermostable, but is more error-prone than LysS (48). Interestingly, LysU can also synthesize a number of adenyl dinucleotides (in particular AppppA). These dinucleotides appear to play as modulators of the heat-shock response and stress response (49).

Our phylogenetic analysis showed that most bacteria within the Bacillus genus possess a second ProRS gene of a distinct origin (Supplementary Figure S7). Although some species of Bacillus have lost this E-type ProRS later during evolution, most have opted to retain this enzyme (Supplementary Figure S7). Retaining the second ProRS perhaps allows the bacterium to endure stresses such as heat and oxidative stress. Another conceivable explanation is that the presence of BtProRS2 might serve as a means to withstand the potential antibacterial effects of natural substances, possibly generated by B. thuringiensis itself or other soil bacteria. There are at least three known natural compounds that potentially inhibit the activity of ProRSs: FF, cispentacin, and phosmidosine. Cispentacin, produced by Bacillus cereus L450-B2 and Streptomyces setonii No. 7562, shares structural similarities with proline and exhibits moderate efficacy in inhibiting fungal ProRS as an antifungal agent (50). On the other hand, phosmidosine, produced by the soil bacteria Streptomyces sp., resembles proline-AMP and shows inhibitory effects on tumor growth (51,52). It is proposed that continued maintenance of this E-type ProRS in B. thuringiensis confers a survival advantage to the bacterium under harsh environments.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

National Science and Technology Council, Taiwan [NSTC112-2311-B008-001]; Taipei Veterans General Hospital [VGHUST113-G4-1-1]. Funding for open access charge: National Science and Technology Council, Taiwan [NSTC112-2311-B008-001].

Conflict of interest statement. None declared.