Naturally occurring dual recognition of tRNA^His substrates with and without a universal identity element

ABSTRACT

The extra 5ʹ guanine nucleotide (G-1) on tRNA^His is a nearly universal feature that specifies tRNA^His identity. The G-1 residue is either genome encoded or post-transcriptionally added by tRNA^His guanylyltransferase (Thg1). Despite Caenorhabditis elegans being a Thg1-independent organism, its cytoplasmic tRNA^His (CetRNA_n^His) retains a genome-encoded G-1. Our study showed that this eukaryote possesses a histidyl-tRNA synthetase (CeHisRS) gene encoding two distinct HisRS isoforms that differ only at their N-termini. Most interestingly, its mitochondrial tRNA^His (CetRNA_m^His) lacks G-1, a scenario never observed in any organelle. This tRNA, while lacking the canonical identity element, can still be efficiently aminoacylated in vivo. Even so, addition of G-1 to CetRNA_m^His strongly enhanced its aminoacylation efficiency in vitro. Overexpression of CeHisRS successfully bypassed the requirement for yeast THG1 in the presence of CetRNA_n^His without G-1. Mutagenesis assays showed that the anticodon takes a primary role in CetRNA^His identity recognition, being comparable to the universal identity element. Consequently, simultaneous introduction of both G-1 and the anticodon of tRNA^His effectively converted a non-cognate tRNA to a tRNA^His-like substrate. Our study suggests that a new balance between identity elements of tRNA^His relieves HisRS from the absolute requirement for G-1.

KEYWORDS:

• Aminoacyl
• tRNA synthetase
• Caenorhabditis elegans
• identity element
• protein synthesis
• tRNA
• translation

Introduction

Aminoacyl-tRNA synthetases (aaRSs) belong to an ancient family of enzymes that charge amino acids to their cognate tRNAs and decipher genetic code via base pairing between the anticodon of aminoacyl-tRNA and the codon of mRNA [1]. The accuracy of tRNA aminoacylation is crucial for maintaining the fidelity of translation. AaRSs establish a tRNA’s identity through recognizing certain identity elements present on the tRNA, which could be single nucleotides, base pairs, or post-transcription modifications [2–4]. Identity elements often reside in the acceptor stem or anticodon loop of a tRNA. In eukaryotic cells, the translation process occurs in both the cytoplasm and organelles (e.g. mitochondria and chloroplasts). Hence, there exists at least two distinct sets of aaRSs in each eukaryote [5,6]. For example, humans possess two distinct histidyl-tRNA synthetase (HisRS) genes, one encoding the cytoplasmic enzyme and the other encoding its mitochondrial counterpart [7]. Nevertheless, in some cases, both forms of a given aaRS are encoded by the same gene through alternative transcription or translation, examples of which include genes encoding yeast alanyl-, glycyl-, histidyl-, and valyl-tRNA synthetases [8–11].

In the majority of species, the major identity element of tRNA^His is an extra guanine nucleotide, G-1, at the 5′ end of the acceptor stem, which is positioned across from the discriminator base N73, a minor identity element. The position −1 is normally empty in mature tRNAs, except for tRNA^His. Therefore, the G-1 residue is an exclusive feature to tRNA^His [12]. The essentiality of this residue (and its monophosphate) is evidenced by the widespread existence of two distinct pathways to yield mature tRNAs^His with G-1. In nearly all eubacteria, eukaryotic organelles, and some archaea, this universal identity element is encoded by the tRNA^His genes opposite C73 and is preserved during processing by ribonuclease P [13]. By contrast, In the majority of eukaryotic cytoplasm and some archaea, G-1 is post-transcriptionally added by tRNA^His guanylyltransferase (Thg1) [14], which places G-1 opposite A73 in eukaryotes [15] and opposite C73 in archaea [16]. Thus, in eubacteria, archaea, and eukaryotic organelles, G-1:C73 constitutes an extra base pair in the acceptor stem. However, in eukaryotic cytoplasm, the additional base pair G-1:A73 is actually a non-Watson-Crick mismatch.

Although the G-1 residue is a nearly universal feature for identification of tRNA^His in life, a few exceptions have been identified. These include tRNAs^His recovered from the cytoplasm of Acanthamoeba castellanii [17] and Trypanosoma brucei [18] and some α-proteobacteria such as Caulobacter crescentus [19]. In A. castellanii and T. brucei, G-1 is missing from their cytoplasmic tRNAs^His, and their cytoplasmic HisRSs can efficiently charge yeast tRNA_n^His without G-1 [17,18].

In the worm Caenorhabditis elegans, HARS-1 is the only HisRS gene which provides functions in both compartments. Since no THG1 orthologue exists in this eukaryote [18], this raised the question of whether its cytoplasmic and mitochondrial tRNA^His isoacceptors carry the canonical identity element. A recent report indicated that the cytoplasmic tRNA^His of C. elegans (CetRNA_n^His) retains a genome-encoded G-1 [18]. In contrast, there is no G residue existing in the corresponding position in the tRNA_m^His gene. To gain further insight, we isolated and sequenced the mature CetRNA_m^His and studied the tRNA^His identity mechanism in this eukaryote. Our results showed that the mature CetRNA_m^His lacks G-1. Despite CetRNA_m^His lacking this canonical identity element, it can be efficiently aminoacylated in vivo. Regardless, CeHisRS still preferred CetRNA^His with G-1, but could tolerate CetRNA_n^His without G-1. Moreover, the anticodon has also taken a leading role in tRNA^His identity through evolution. The above findings suggest a pioneering mechanism which allows CeHisRS to recognize tRNA^His isoacceptors with or without G-1.

Results

C. elegans HARS-1 is a dual-functional HisRS gene

Sequence homology search indicated that HARS-1 is the only HisRS gene in the nuclear genome of C. elegans. This gene contained 6 exons (exon 1 to exon 6), with alternative intron splicing yielding two mature transcripts, HARS-1a (containing exons 1, 3, 4, 5, and 6) and HARS-1b (containing exons 2, 3, 4, 5, and 6) (Figure 1(a)). The polypeptides translated from these two transcripts were essentially identical in sequence except for the N-terminal domain. HARS-1a-translated polypeptide carried an N-terminal domain, which shares significant sequence homology with the WHEP domain, while HARS-1b-translated polypeptide carried an N-terminal domain, which is rich in hydroxylated and positively charged residues but devoid of acidic residues, a feature characteristic of a mitochondrial targeting signal (MTS) (Figure 1(a)). Hence, the former (CeHisRS_c) is likely to function in the cytosol, and the latter (CeHisRS_m) in mitochondria. As the MTS of CeHisRS_m is expected to be cleaved off upon being imported into mitochondria, the only difference between CeHisRS_c and the processed CeHisRS_m is the WHEP domain.

Figure 1. C. elegans HisRS isoforms and tRNA^His isoacceptors.

(a) Generation of two CeHisRS isoforms. CeHisRS_c and CeHisRS_m possess essentially the same polypeptide sequence except for the N-terminal domain, with WHEP (residues 1 ~ 63) in CeHisRS_c and MTS (residues 1 ~ 33) in CeHisRS_m. Relative positions of the functional domains in CeHisRS are marked. CD, catalytic domain; ABD, anticodon-binding domain. (b) tRNA^His isoacceptors. Identity elements that are crucial for recognition of SctRNA_n^His by ScHisRS are boxed. The defective TψC arm in CetRNA_m^His is marked with a dashed circle. The small letters in SctRNA_n^His denote the different nucleotides between SctRNA_n^His and CetRNA_n^His.

G-1 is a nearly universal feature that specifies tRNA^His identity. For example, the primary identity element of S. cerevisiae tRNA^His (SctRNA^His) is G-1, and deletion of this nucleotide resulted in an unchargeable tRNA mutant [20,21]. Figure 1(b) shows the nuclear- and mitochondrial-encoded tRNA^His isoacceptors. In contrast to the scenario of E. coli tRNA^His, the discriminator base N73 plays almost no role in SctRNA^His identity [20]. On the other hand, the anticodon has become an important identity element of SctRNA^His. Despite CetRNA_n^His retaining a structure similar to that of SctRNA_n^His, its G-1 is genome encoded, which has so far only been observed in prokaryotes and organelles. Most surprisingly, CetRNA_m^His lacks a G-1 nucleotide at the corresponding position in the mitochondrial genome, and contains U73 instead of C73. Moreover, CetRNA_m^His lacks a TψC arm, which makes it very different from most tRNA^His species.

The C. elegans mitochondrial tRNA^His lacks the canonical identity element

Because no THG1 orthologue exists in C. elegans and its mitochondrial tRNA^His gene does not contain a G nucleotide at position −1, we wondered whether an editing enzyme or pathway exists to salvage the G-1 nucleotide. To take a closer look, we isolated total RNA from a C. elegans culture, circularized the RNA with T4 RNA ligase, and amplified the acceptor stem sequence of tRNA_m^His by RT-PCR using a set of gene-specific primers (Figure 2(a)). The PCR products were cloned into pUC18 and ten of the resultant clones were sequenced. Figure 2(b) shows that almost all of the tRNA_m^His clones (clones 1 ~ 9) contained a CCA tail (nucleotides 74 ~ 76), suggesting that they represent mature tRNA_m^His species. However, none of these clones contained G-1 (clones 1 ~ 10).

Figure 2. Determining the G-1 status of the C. elegans mitochondrial tRNA^His.

(a) Strategy used to determine the G-1 status of CetRNA_m^His. (b) Sequencing data. For clarity, the expected position of G-1 is boxed. (c) In vivo aminoacylation assay. The charged and uncharged tRNAs were separated in an acid 10% polyacrylamide-7 M urea gel, blotted onto a positively charged nylon membrane, and hybridized with specific ³²P-labled probes. The symbols ‘+’ and ‘–’ denote the treatment with alkaline pH. The arrows mark the positions of charged and uncharged tRNAs^His.

To check whether CetRNA_m^His is indeed a natural substrate for CeHisRS, two micrograms of total mitochondrial RNA were isolated from C. elegans under acidic conditions, and a part of the RNA sample was incubated at alkaline pH to hydrolyze aminoacyl-tRNA ester bonds. These RNA samples were electrophoresed through an acid (pH 5.2) 10% polyacrylamide-7 M urea gel and then blotted onto a positively charged nylon membrane. The hybridization was performed with ³²P-labeled oligonucleotide probes specific for the cytoplasmic and mitochondrial CetRNA^His isoacceptors, respectively. As expected, CetRNA_m^His (without G-1) was efficiently charged in vivo (Figure 2(c)). To our knowledge, this is the first example studied to date in which an organellar tRNA^His completely lacks G-1. While this outcome was not totally surprising, it prompted us to ask how CeHisRS can accommodate two such divergent tRNAs^His, one with and the other without the canonical identity element.

C. elegans HiSRs can recognize yeast tRNA^his with G-1:A73 or G-1:C73

Since the native tRNA substrates of CeHisRS are so different in G-1 and N73, we suspected that these two identity elements are not as crucial as previously anticipated. To test this hypothesis, we cloned the genes encoding CeHisRS_c and CeHisRS_m into a yeast expression vector, and transformed the constructs into a yeast strain deleted of THG1 (encoding Thg1) or HTS1 (encoding HisRS). The rescue activities of these constructs are summarized in Figure 3. As described earlier, the G-1 residue in SctRNA_m^His is genome encoded, while the G-1 residue in SctRNA_n^His is post-transcriptionally added. Therefore, no G-1 can be added to SctRNA_n^His in the thg1⁻ strain. 5-FOA and YPG plates were respectively used to test the cytoplasmic and mitochondrial HisRS functions of a gene of interest.

Figure 3. Heterologous rescue activity of C. elegans HisRS isoforms.

(a) Summary of the HisRS constructs and their rescue activities. The symbols ‘+’ and ‘–’ respectively denote positive and negative complementation. (b) Complementation of the genetic loss of yeast THG1 or HTS1. (c) Western blotting. The amount of cellular protein extracts loaded into each well is indicated at the bottom of the blot. Numbers 1 ~ 6 (circled) shown in (b) and (c) represent constructs shown in (a).

As shown in Figure 3(b), CeHisRS_c efficiently restored the growth phenotype of the HTS1 knockout strain on 5-FOA (number 2 in middle panel), and when fused with a yeast MTS, it efficiently rescued the growth defect of the same knockout strain on YPG (number 3 in right panel). Similarly, CeHisRS_m with its own MTS efficiently rescued the growth defect of the HTS1 knockout strain on YPG (number 4), and upon deletion of its native MTS, CeHisRS_m efficiently restored the growth phenotype of the null allele on 5-FOA (number 5). We also noticed that CeHisRS_c fused with a heterologous MTS or CeHisRS_m carrying its own MTS could function in both compartments when overexpressed (numbers 3 and 4), possibly due to incomplete mitochondrial targeting of the fusion enzyme, a scenario often seen in aaRS rescue assays [10,22]. Likewise, overexpressed ScHisRS (carrying its own MTS) could rescue the growth defects of the null allele on both 5-FOA and YPG (number 6). Consistent with a previous observation [18], neither CeHisRS_c or CeHisRS_m were able to complement the genetic loss of THG1 on 5-FOA (left panel), suggesting that CeHisRS cannot charge SctRNA_n^His without G-1 (SctRNA_n^His(∆G-1)). Together, these results suggest that CeHisRS can charge SctRNAs^His with G-1:A73 or G-1:C73, but cannot charge SctRNA^His without G-1.

We next examined the relative protein expression levels of these constructs in yeast using an anti-His₆ tag antibody as a probe. As shown in Figure 3(c), CeHisRS_c had a relatively higher expression level when localized in the cytoplasm than in mitochondria (numbers 2 and 3). A similar scenario was observed for CeHisRS_m (numbers 4 and 5). The expression levels of CeHisRS_c and CeHisRS_m(∆MTS) were similar to that of ScHisRS (numbers 2, 5, and 6). Despite the fact that MTS-CeHisRS_c and CeHisRSm were expressed at a much lower level than their cytoplasmic counterparts (Figure 3(c)), they could provide sufficient rescue activity in both compartments (Figure 3(b)).

Cehisrs is more tolerant than ScHisRS towards tRNA^His without G-1

Consistent with a previous observation [18], our study indicated that CeHisRS poorly recognizes SctRNA_n^His without G-1 (Figure 3). We next tested whether CetRNA^His(∆G-1) is a better substrate for CeHisRS in vivo. Pursuant to this objective, constructs that harbor the genes encoding HisRS and tRNA^His(∆G-1) were co-transformed into the yeast THG1 knockout strain and the growth phenotypes of the resultant co-transformants on 5-FOA were examined (Figure 4(a)). To express tRNA^His in the THG1 knockout strain, the genes encoding SctRNA^His and CetRNA^His were cloned into pRS313 (a low-copy-number yeast shuttle vector for normal expression of the target gene) and pRS423 (a high-copy-number yeast shuttle vector for overexpression of the target gene).

Figure 4. Complementation of the genetic loss of yeast THG1.

(a) Strategy used to determine the rescue activity of an orthogonal or non-orthogonal pair of HisRS/tRNA^His. (b) Co-expression of ScHisRS and various tRNAs^His(∆G-1). (c) Co-expression of CeHisRS and various tRNAs^His(∆G-1). pRS423 is a high-copy-number yeast vector with a 2μ replication origin, while pRS313 is a low-copy-number yeast vector with a CEN replication origin.

As shown in Figure 4(b), overexpression of both ScHisRS and SctRNA_n^His(∆G-1) circumvented the requirement for Thg1 (number 1), a scenario that was reported earlier [23]. Conceivably, overexpression of both the enzyme and substrate enhances the reaction rate to a level that can sustain the growth of the knockout strain. When ScHisRS was overexpressed and SctRNA_n^His(∆G-1) was non-overexpressed, no growth was observed (number 2). A similar result was obtained when CetRNA_n^His(∆G-1) was used as the substrate (numbers 3 and 4). However, we did notice that non-overexpressed CetRNA_n^His(∆G-1), but not SctRNA_n^His(∆G-1), conferred a very weak yet distinguishable complementation (numbers 2 and 4). No growth was observed when the tRNA substrate was replaced with CetRNA_m^His (lacking G-1) (number 5).

We next tested whether CeHisRS is more tolerant than ScHisRS towards tRNA^His(∆G-1). As shown in Figure 4(c), overexpression of both CeHisRS_c and SctRNA_n^His(∆G-1) effectively complemented the genetic loss of THG1 (number 1). However, the complementation turned negative if the tRNA substrate was not overexpressed (number 2). In contrast to SctRNA_n^His(∆G-1), CetRNA_n^His(∆G-1) appeared to be a more favorable substrate for CeHisRS_c. Overexpressed CeHisRS_c was able to restore the growth phenotype of the null allele even when the substrate CetRNA_n^His(∆G-1) was not overexpressed (numbers 3 and 4). No growth was observed when the tRNA substrate was replaced with CetRNA_m^His (which lacks both G-1 and a TψC arm) (number 5). Considering the fact that CetRNA_m^His requires a species-specific post-transcriptional modification to fold into a proper L-shape structure [24], this outcome was not surprising. Because the mature CeHisRS_m is different from CeHisRS_c by only an N-terminal WHEP domain, we next tested whether the WHEP domain is required for the rescue activity. As expected, the mature CeHisRS_m (or CeHisRS_c deleted of WHEP) retained a rescue activity almost indistinguishable from that of CeHisRS_c (numbers 7 ~ 12). Hence, the WHEP domain is dispensable for the rescue activity of CeHisRS. Collectively, these data suggest that CeHisRS is more tolerant than ScHisRS towards tRNA^His without G-1, and CeHisRS prefers CetRNA_n^His(∆G-1) over SctRNA_n^His(∆G-1).

A change in identity element balance enables C. elegans HiSRs to tolerate tRNA^His without G-1

To determine the steady-state kinetic parameters for tRNA^His aminoacylation, His₆-tagged CeHisRS_c and CeHisRS_m (without MTS) were purified to homogeneity through Ni-NTA affinity chromatography. As one of the native tRNA substrates of CeHisRS lacks the canonical identity element G-1, we wondered whether CetRNA^His has gained an additional identity element during evolution. Moreover, we were prompted to ask whether the WHEP domain is really dispensable for tRNA^His histidylation. To this end, various mutations were introduced into CetRNA_n^His and the aminoacylation efficiencies of the resultant tRNA mutants were determined.

As shown in Table 1, deletion of G-1 from CetRNA_n^His reduced the aminoacylation efficiency (k_cat/K_M) to 5% for CeHisRS_c. Unexpectedly, this deletion reduced not only the k_cat value (75-fold) but also the K_M value (3.5-fold). In contrast, mutation at the discriminator base A73 had only a minor effect on aminoacylation. Mutation of A73 to U, C, or G respectively reduced the aminoacylation efficiency to 94%, 22%, and 11%, suggesting that CeHisRS_c slightly prefers A and U at this position. This finding might account for the presence of A73 and U73 on CetRNA_n^His and CetRNA_m^His, respectively. We next focused on the anticodon GUG (nucleotides 34–36) of CetRNA_n^His. As mutations G34U, U35C, and G36C were previously shown to have the strongest effect on histidylation of SctRNA_n^His by ScHisRS [20], similar mutations were introduced into CetRNA_n^His. G34U, U35C, and G36C respectively reduced the aminoacylation efficiency to 19%, 8% and 2%. Apparently, the importance of specific nucleotides in the anticodon is shifted. ScHisRS relies more on G34 and U35, while CeHisRS relies more on G36.

Table 1. Kinetic parameters for aminoacylation of tRNA_n^His by CeHisRS_c.

CeHisRSc
CetRNAn^His	kcat (min^−1)	KM(μM)	kcat/KM(μM^−1min^−1)	Efficiency (%)
WT	24.0 ± 1.46	4.20 ± 0.72	5.67	100
∆G-1	0.32 ± 0.02	1.21 ± 0.18	0.26	5
A73U	14.3 ± 2.12	2.69 ± 0.73	5.32	94
A73C	6.43 ± 0.96	5.19 ± 1.07	1.24	22
A73G	3.29 ± 0.40	5.20 ± 0.81	0.63	11
G34U	3.17 ± 0.79	2.91 ± 0.92	1.08	19
U35C	3.05 ± 0.30	6.58 ± 1.17	0.46	8
G36C	1.51 ± 0.24	14.1 ± 2.59	0.10	2

We next tested whether the WHEP domain is actually dispensable for tRNA^His aminoacylation. Note that the ‘processed’ or mature CeHisRS_m is essentially equivalent to CeHisRS_c minus the WHEP domain. Table 2 shows that deletion of the WHEP domain from CeHisRS_c had a very limited effect on its histidylation activity. The k_cat/K_M value for histidylation of the WT CetRNA_n^His by CeHisRS_m was close to that by CeHisRS_c (4.23 versus 5.67 μM⁻¹min⁻¹). Moreover, deletion of G-1 and mutation of G36 to C respectively reduced the histidylation efficiency to 9% and2%, while mutation of A73 to U increased the efficiency to 123% (Table 2). These results support our hypothesis that the WHEP domain is dispensable for the histidylation activity of CeHisRS.

Table 2. Kinetic parameters for aminoacylation of tRNA_n^His by CeHisRS_m.

CeHisRSm
CetRNAn^His	kcat (min^−1)	KM(μM)	kcat/KM(μM^−1min^−1)	Efficiency (%)
WT	27.8 ± 4.35	6.56 ± 0.39	4.23	100
∆G-1	0.32 ± 0.02	0.3 ± 0.02	0.36	9
A73U	17.1 ± 2.75	3.28 ± 1.07	5.19	123
G36C	0.17 ± 0.03	2.04 ± 0.63	0.08	2
CetRNAn^Thr
WT	ND	ND	ND
G-1	0.02 ± 0.01	0.63 ± 0.24	0.03	0.7
AGU/GUG	0.02 ± 0.01	0.66 ± 0.16	0.03	0.7
G-1/GUG	2.35 ± 0.25	2.63 ± 0.51	0.89	21
SctRNAn^His
WT	26.3 ± 0.78	11.3 ± 1.02	2.33	55
∆G-1	ND	ND	ND

ND, not determined due to low activity.

As the mature mitochondrial CetRNA^His (without G-1) can be efficiently aminoacylated in vivo (Figure 2(c)), we wondered whether addition of G-1 to this tRNA affects its aminoacylation efficiency. As shown in Figure 5, addition of G-1 to CetRNA_m^His significantly enhanced its aminoacylation efficiency (~8-fold). Thus, CeHisRS prefers CetRNA^His with G-1, regardless of whether it is within the context of cytoplasmic or mitochondrial CetRNA^His.

Figure 5. In vitro aminoacylation assay. The in vitro aminoacylation activity of CeHisRS_m (1 μM) was tested using in vitro-transcribed CetRNAs (5 μM) as the substrates. The relative amounts of [¹⁴C]Histidine that were incorporated into tRNA_m^His were subsequently determined by a liquid scintillation counter. The arrows indicate the tRNA species used. CetRNA_m^Gly served as a negative control.

tRNA^Thr can be converted to a tRNA^His-like substrate by insertion of both G-1 and the anticodon of tRNA^His

So far, we have shown that both G-1 and the anticodon significantly contribute to the specific recognition of CetRNA_n^His. To be sure that these two elements are necessary to establish a tRNA^His identity for CeHisRS, we next examined whether a non-cognate tRNA, CetRNA_n^Thr, can be converted to a chargeable tRNA substrate by insertion of one or both of these two elements. A similar strategy has previously been shown in converting SctRNA^Thr to a tRNA^His-like substrate [25]. As shown in Table 2, the WT CetRNA_n^Thr could not be charged by CeHisRS_m, but insertion of G-1 or changing the anticodon AGU to GUG partially rescued its aminoacylation efficiency (~0.7% relative to the WT CetRNA_n^His). Interestingly, when both of these two elements were introduced into CetRNA_n^Thr, the resultant tRNA became a good substrate for CeHisRS_m, with 21% histidylation efficiency relative to the WT CetRNA_n^His. Thus, both the G-1 residue and anticodon are necessary to specify a tRNA_n^His identity for CeHisRS.

Our rescue assays indicated that CetRNA_n^His(∆G-1) is a better substrate than SctRNA_n^His(∆G-1) for CeHisRS (Figure 4). To obtain quantitative data, SctRNA_n^His and its ∆G-1 mutant were also tested. As expected, the WT SctRNA_n^His was an excellent substrate for CeHisRS, with 55% efficiency relative to the WT CetRNA_n^His. However, deletion of G-1 from SctRNA_n^His yielded a ∆G-1 mutant unsuitable for histidylation by CeHisRS_m; no detectable charging activity was observed for this substrate (Table 2). Thus, CeHisRS can tolerate its own tRNA_n^His without G-1, but not yeast tRNA_n^His without G-1.

The WHEP domain plays little role in the thermal stability and structural flexibility of CeHiSRs

As the only difference between CeHisRS_c and the mature CeHisRS_m is the WHEP domain, we wondered whether this domain regulates the thermal stability of CeHisRS. To this end, purified CeHisRS_c and CeHisRS_m proteins were subjected to circular dichroism (CD) spectroscopy at 222 nm between 10°C and 100°C. As shown in Figure 6, both proteins retained more than 80% of their α-helix structure (compared with the molar ellipticity value determined at 20°C) at temperatures below 40°C, but lost almost half of their α-helix structure when the temperature reached 50°C or higher (Figure 6(a,b)). As a result, CeHisRS_c and CeHisRS_m had a melting temperature of 52°C and 55°C, respectively (Figure 6(c)). This result suggests that the WHEP domain does not play a significant role in the thermal stability of CeHisRS.

Figure 6. Thermal stability and structural flexibility of C. elegans HisRS isoforms.

Melting temperatures of (a) CeHisRS_c and (b) CeHisRS_m were determined at 222 nm using CD spectroscopy between 10℃ and 100℃. (c) Relative CD signal changes of CeHisRS_c and CeHisRS_m. (d) Limited proteolysis of CeHisRS_c and CeHisRS_m. The protein amount used was 4 μg per lane. The arrows indicate CeHisRS_c and CeHisRS_m.

To study the effect of the WHEP domain on the structural flexibility of CeHisRS, limited proteolysis was carried out with CeHisRS/trypsin in a ratio of 10:1. Limited proteolysis is often used to probe the structure and dynamics of proteins. Exposed regions such as loops and flexible regions are more susceptible to the prolific protease. As shown in Figure 6(d), both forms of CeHisRS were rather resistant to trypsin. The relative protein levels of CeHisRS_c and CeHisRS_m remained almost unchanged in high concentrations of trypsin throughout the time period tested. This result suggests that the WHEP domain does not play a significant role in regulating the dynamic structure of CeHisRS.

Discussion

The G-1 residue is the major identity element for recognition of tRNA^His, regardless of whether it is of eukaryotic or prokaryotic origin [20,26,27]. Deletion of this canonical identity element from E. coli and yeast tRNAs^His respectively reduced the aminoacylation efficiency (k_cat/K_M) ~250 and ~740-fold [20,27]. Similarly, deletion of G-1 from human tRNA^His led to a mutant tRNA^His unsuitable for histidylation [7]. In contrast, CeHisRS exhibits relatively high tolerance towards CetRNA_n^His lacking G-1; deletion of this identity element only reduced the aminoacylation efficiency ~15-fold (Table 1,2). Hence, overexpression of CeHisRS was able to rescue a yeast THG1 null allele, regardless of whether CetRNA_n^His(∆G-1) was also overexpressed (Figure 4). In contrast, both ScHisRS and SctRNA_n^His(∆G-1) had to be overexpressed in order to rescue the THG1 knockout strain (Figure 4) [23]. While the importance of specific nucleotides in the anticodon is shifted between CeHisRS and ScHisRS, the anticodon still plays a significant role in tRNA^His recognition by CeHisRS. As a result, a non-cognate tRNA can be converted to a tRNA^His-like substrate for CeHisRS via simultaneous insertion of both the G-1 residue and anticodon of tRNA^His (Table 2).

Consistent with our observations (Figure 4 and Table 2), a recent study showed that CeHisRS strongly prefers SctRNA_n^His with G-1 over SctRNA_n^His without G-1, and that overexpression of CeHisRS alone fails to rescue a yeast THG1 knockout strain. However, their study also showed that CeHisRS can still charge SctRNA_n^His without G-1, albeit to a low level [18]. In contrast, no charging of this tRNA variant by CeHisRS was detected in our assay. A careful examination of the reaction conditions used by these two studies suggested that the slight difference in aminoacylation activity might result from different concentrations of histidine used in the assays. They used 40 µM of histidine (and labeled ³²P at tRNA^His), while we used only 26 µM of histidine (and labeled ¹⁴C at histidine). It is possible that CeHisRS possesses a K_M value for histidine higher than expected, leading to the slight difference in aminoacylation activity.

Interestingly, our data also indicated that SctRNA_n^His without G-1 is a poorer substrate for CeHisRS than CetRNA_n^His without G-1 (Figure 4 and Table 2). A more detailed sequence and structural comparison between these two tRNAs^His indicated that SctRNA_n^His is different from CetRNA_n^His at 16 nucleotide positions, including 6 in the acceptor stem, 1 at the corner of the D stem and anticodon stem, 4 in the anticodon stem, 2 in the variable loop, and 3 in the TΨC stem (shown in Figure 1(b)). Despite the fact that none of these positions have ever been reported to be critical for aminoacylation of tRNA^His by HisRS, it is possible that these nucleotides (or base pairings) play a role in aminoacylation only when the major determinant G-1 is missing. As reported earlier, the nucleotides (or base pairings) in the acceptor stem contribute to the structural flexibility of tRNA [28], while the nucleotides in the variable loop are involved in the overall folding of tRNA [29]. In turn, some of these differences in these two tRNAs may confer a unique conformation or a secondary determinant to CetRNA_n^His without G-1. As C. elegans genetically lacks the G-1-addition enzyme, this ability to tolerate a G-1-deficient CetRNA_n^His variant might provide certain survival advantages under conditions where the G-1 nucleotide of CetRNA_n^His is aberrantly removed during processing by RNase P.

The essentiality of G-1 for tRNA^His identity is evidenced by the prevalence of two distinct mechanisms for acquiring the element, one depending on co-transcriptional incorporation and the other on post-transcriptional addition [3,30,31]. The former pathway occurs in bacteria, eukaryotic organelles, and certain archaea, while the latter pathway occurs in eukarya and some archaea. Nevertheless, certain eukaryotes such as A. castellanii and T. brucei lack a THG1 orthologue [17,18], and thus cannot add G-1 to their cytoplasmic tRNAs^His. As a result, the HisRS_c enzymes of these two eukaryotes can efficiently charge G-1-deficient tRNA^His. On the other hand, plant mitochondria appear to retain both pathways for acquisition of the G-1 residue [32]. Unlike other Thg1-dependent eukaryotes, the slime mold Dictyostelium discoideum possesses up to four THG1 homologues. Deletion of the homologue that is responsible for adding G-1 to the mitochondrial tRNA^His resulted in a slow-growth phenotype rather than cell death [33], suggesting a certain level of tolerance towards tRNA_m^His without G-1.

Despite the fact that both T. brucei and A. castellanii are Thg1 independent and their cytoplasmic HisRSs are only responsible for recognition of tRNA^His isoacceptors without G-1 in vivo [17,18], these two HisRSs can also efficiently charge tRNA^His with G-1 in vitro. Although C. elegans is also a Thg1-independent eukaryote, its cytoplasmic tRNA^His contains a genome-encoded G-1 residue, which has so far only been found in prokaryotes and organelles. Moreover, its mitochondrial tRNA^His isoacceptor lacks the universal identity element (Figure 2), a phenomenon never observed in any organelle. To be dual functional and at the same time be able to charge two highly divergent tRNA^His isoacceptors (one with G-1 and the other without), CeHisRS was forced to evolve from a G-1-dependent enzyme to a G-1-preferred enzyme. It remains unclear how the G-1 residue of tRNA_n^His can be retained during processing by nuclear RNase P and whether losing the G-1 identity element in tRNA_m^His confers any selective advantage to the mitochondrial translational machinery of this organism. In summary, our study provided pioneering insights on how the anticodon might work alongside with G-1 in identity recognition of CetRNA^His.

Materials and methods

Construction of plasmids

Cloning of the gene encoding CeHisRS_c or CeHisRS_m into pTEF1 (a high-copy-number yeast shuttle vector with a strong TEF1 promoter, a LEU2 maker, and a 2μ replication origin) followed a standard protocol. Briefly, a pair of gene-specific primers were used to amplify the gene via a polymerase chain reaction (PCR) using C. elegans cDNA as the template. The forward primer (with an SpeI site) was located immediately upstream of the open reading frame (ORF), while the reverse primer (with an XhoI site) was located immediately upstream of the stop codon. The amplified DNA fragment was digested with SpeI and XhoI and then cloned into the SpeI/XhoI sites of pTEF1 for complementary assays. To fuse a heterologous MTS to HisRS, a DNA sequence encoding amino acid residues 1 ~ 46 of the mitochondrial precursor form of yeast valyl-tRNA synthetase was amplified by a PCR as an XbaI-SpeI fragment and then inserted into the 5ʹ SpeI site of the target gene.

For protein purification, the target genes were cloned into pET21b (an E. coli expression vector with a T7 promoter) following a similar protocol. The plasmids harboring the target genes were transformed into BL21-CodonPlus(DE3), and His₆-tagged proteins were purified to homogeneity through Ni-NTA affinity chromatography as previously described [34]. Western blotting followed a standard protocol and used an HRP-conjugated anti-His₆ tag antibody as a probe [35].

Cloning of the SctRNA^His gene into pRS423 (a high-copy-number yeast shuttle vector with a HIS3 marker and a 2μ replication origin) or pRS313 (a low-copy-number yeast shuttle vector with a HIS3 marker and a CEN replication origin) followed a similar protocol, except that the 5ʹ and 3ʹ ends of the sequence respectively extended 345 bp upstream and 375 bp downstream of the gene. To clone and express the genes encoding CetRNA_n^His and CetRNA_m^His, a ~ 800-bp DNA sequence with the CetRNA_n^His or CetRNA_m^His gene replacing the SctRNA_n^His gene in the aforementioned DNA fragment was in vitro synthesized by a manufacturer (Bio-Basic, Markham ON, Canada). The synthesized DNA sequence was then cloned into pRS423 and pRS313 for expression in yeast. Thus, transcription of CetRNA^His and SctRNA_n^His was driven by the same promoter.

Complementation assay for cytoplasmic activity on 5-FOA

Construction of a haploid HTS1 knockout strain (MATα, hts1::kanMX4, his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) was reported earlier [7]. To examine the cytoplasmic HisRS activity of a gene of interest, a test plasmid carrying the target gene and a LEU2 marker was transformed into the haploid knockout strain and the ability of the resulting transformants to grow in the presence of 5-FOA (1 mg/ml) was tested [36]. Starting with a cell density of 1.0 A₆₀₀, cell cultures were 3-fold serially diluted, and 10-μl aliquots of each dilution were spotted onto the 5-FOA plates. Plates were then incubated at 30°C for 3 days. In the presence of 5-FOA, the transformants evicted the maintenance plasmid (which carried a WT HTS1 gene and a URA3 marker) and therefore failed to grow unless the test plasmid provided a functional cytoplasmic HisRS. To test whether a gene of interest can complement the genetic loss of yeast THG1, a similar protocol was followed using a haploid thg1⁻ strain (MATα, thg1::kanMX4, his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0).

Complementation assay for mitochondrial activity on YPG

To examine the mitochondrial HisRS activity of a gene of interest, a test plasmid (which carried the target gene and a LEU2 marker) and a second maintenance plasmid (which carried a HIS3 marker and an initiator mutant of HTS1 that encodes only cytoplasmic HisRS activity) were co-transformed into the yeast hts1⁻ strain. The first maintenance plasmid (which carried a URA3 marker) was evicted from the co-transformants in the presence of 5-FOA, while the second maintenance plasmid (which carried a HIS3 marker) was retained. Following 5-FOA selection, the survived co-transformants were spotted on yeast extract-peptone-glycerol (YPG) plates, which were then incubated at 30°C for 3 days. Because glycerol is a non-fermentable carbon source, a yeast cell cannot survive in this medium without functional mitochondria. Thus, the co-transformants could not survive on YPG plates unless the test plasmid provided a functional mitochondrial HisRS.

In vitro transcription of tRNA^his

Preparation of the WT and mutant tRNA^His transcripts followed a previously described protocol [27]. Briefly, a synthetic DNA sequence containing both a T7 promoter and a gene encoding tRNA^His (GUG) was cloned into the SmaI site of pUC18 using a set of complementary primers. The transcription template was enriched by PCR amplification of the insert. The in vitro transcription reaction of tRNA^His (with a 5ʹ-GMP) was performed at 37°C for 3 h with 0.3 μM T7 RNA polymerase in 20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 20 mM MgCl₂, 5 mM DTT, 1 mM spermidine, 2 mM of each NTP, and 20 mM GMP. The tRNA^His transcript was purified by a 15% denaturing urea-polyacrylamide gel. After 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 then refolded by heating to 80°C and gradually cooled to room temperature after addition of 10 mM MgCl₂. ~80% of in vitro-transcribed tRNA^His were active in aminoacylation.

Northern blotting

Total mitochondrial RNA was purified from the mitochondrial fractions of C. elegans by using TRIzol (Invitrogen) as previously described [37]. A portion of the RNA sample was incubated at alkaline pH to hydrolyze aminoacyl-tRNA ester bonds. Two micrograms of total mitochondrial RNA were electrophoresed by an acid 10% polyacrylamide-7 M urea gel in 100 mM sodium acetate-EDTA running buffer (pH 5.2), electroblotted onto a positively charged nylon membrane (GeneScreen Plus, PerkinElmer), cross-linked by UV, and followed by probing with specific ³²P-labeled primers: CetRNA_n^His (5ʹ-TGCCTGCGGCTGGAATCG) and CetRNA_m^His (5ʹ-AAGCTCTATATTCTCTTA), respectively [38].

Aminoacylation assay

Kinetic parameters for histidylation of tRNA were determined by measuring the initial rate of charging over the first 4 min [39]. Initial rates of aminoacylation were determined at 25°C in a buffer containing 50 mM HEPES (pH 7.5), 50 mM KCl, 15 mM MgCl₂, 5 mM dithiothreitol, 10 mM ATP, 0.1 mg/ml bovine serum albumin, and 26.25 μM histidine (6.25 μM ¹⁴C-histidine; PerkinElmer, Waltham, MA, USA), with tRNA^His concentrations ranging 0.16 ~ 20 μM and HisRS concentrations ranging 50 ~ 1,000 nM. The specific activity of ¹⁴C-histidine used was 325 mCi/mmol. Reactions were quenched by spotting 10-μl aliquots of the reaction mixture onto Whatman filters (Maidstone, Kent, UK) presoaked in 5% trichloroacetic acid (TCA) and 2 mM histidine. The filters were washed three times for 15 min each in ice-cold 5% TCA before liquid scintillation counting. The chargeable tRNA^His concentration was determined by aminoacylation reactions to a saturated state using a high concentration of enzyme. The kinetic parameters were derived from Lineweaver-Burk plots. Determination of active protein concentrations by active site titration was as previously described [40]. Data were obtained from three independent experiments and averaged. Error values represent 2× standard deviations.

Miscellaneous methods

Circular dichroism (CD) spectroscopy [41], tRNA^His 5ʹ end sequencing [17], and limited proteolysis [42] followed previously described protocols.

Abbreviations

aaRS	=	aminoacyl-tRNA synthetase
cDNA	=	complementary DNA
5-FOA	=	5-fluoroorotic acid
HisRS	=	histidyl-tRNA synthetase
MTS	=	mitochondrial targeting signal
ORF	=	open reading frame
PCR	=	polymerase chain reaction
tRNAnHis	=	nuclear-encoded cytoplasmic tRNAHis
tRNAmHis	=	mitochondrial-encoded mitochondrial tRNAHis
Thg1	=	tRNAHis guanylyltransferase
WT	=	wild-type
YPG	=	yeast extract-peptone-glycerol

Acknowledgments

We are grateful to Prof. Yi-Chun Wu (NTU) for providing C. elegans cosmids and Chieh-Min Chu for careful proofreading of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Ministry of Science and Technology, Taiwan [103-2311-B-008-003-MY3];Ministry of Science and Technology, Taiwan [105-2311-B-008-002-MY3];Ministry of Science and Technology, Taiwan [103-2923-B-008-001-MY3].