Characterization of two bifunctional Arabdopsis thaliana genes coding ...

Eur. J. Biochem. 266, 848±854 (1999) q FEBS 1999

Characterization of two bifunctional Arabdopsis thaliana genes coding for mitochondrial and cytosolic forms of valyl-tRNA synthetase and threonyl-tRNA synthetase by alternative use of two in-frame AUGs Ginette Souciet, BenoõÃt Menand*, Jaroslava Ovesna², Anne Cosset, AndreÂ Dietrich and Henri Wintz Institut de Biologie MoleÂculaire des Plantes du CNRS, UniversiteÂ Louis Pasteur, Strasbourg, France

We characterized two Arabidopsis thaliana cDNAs coding for class I valyl-tRNA synthetase and class II threonyl-tRNA synthetase. The proteins display characteristics of cytosolic enzymes, yet possess an N-terminal extension relative to their prokaryotic homologs. The proximal part of the N-terminal extension is a mitochondrial-targeting signal. Through transient expression of GFP fusions in tobacco cells, we demonstrated that both genes encode the cytosolic and mitochondrial forms of the enzymes by alternative use of two in-frame initiation codons. A long, mitochondrial form of the enzyme is translated from a first initiation codon at reduced levels because of a poor sequence context and a shorter, cytosolic form is translated from a second in-phase AUG, which is in a better context for translation initiation. Primer extension experiments revealed several transcript ends mapping upstream of the first AUG and between the two AUGs. Distal to the mitochondrial transit peptide both valyl-tRNA synthetase and threonyl tRNA synthetase possess an NH2-appended domain compared with their prokaryotic counterparts. This domain's amphiphilic helix is conserved between yeast and A. thaliana valyl-tRNA synthetase, suggesting an important role in translation. Based on the high structural similarities between yeast and A. thaliana valyl-tRNA synthetase, we propose that the acquisition of bifunctionality of valyl-tRNA synthetase predates the divergence of these two organisms. Keywords: aminoacyl-tRNA synthetase; GFP; mitochondria; translation initiation.

In plants, protein synthesis takes place in three cell compartments: the cytosol, the mitochondria and the plastids. Each compartment requires a complete set of tRNAs and aminoacyltRNA synthetases (aaRSs) to function. Except in some algae, all aaRSs are coded for by nuclear genes and addressed posttranslationally to their respective compartments by means of signal peptides present at their N-terminus. Transfer RNAs are coded for by genes present on all three genomes (nuclear, mitochondrial and plastidial). In general, plastids encode all the tRNAs needed for their own translation, whereas mitochondrial genomes of plants do not code for all the tRNAs required for protein synthesis [1]. It is now well established that a subset of mitochondrial tRNAs is coded for by nuclear genes and is imported into mitochondria [2]. The situation in mitochondria is even more complex, as several mitochondrial tRNAs are coded for by chloroplast genes that have been inserted into the mitochondrial genome during evolution [3,4]. Correspondence to H. Wintz, Institut de Biologie MoleÂculaire des Plantes du CNRS, UniversiteÂ Louis Pasteur, 12, rue du GeÂneÂral Zimmer, 67084 Strasbourg Cedex, France. Fax: +33 3 8861 4442, Tel.: +33 3 8841 7227, E-mail: [email protected] Abbreviations: aaRS, aminoacyl-tRNA synthetase; AlaRS, alanyl-tRNA synthetase; GFP, green fluorescent protein; HisRS, histidyl-tRNA synthetase; ThrRS, threonyl-tRNA synthetase; TL, translation leader; ValRS, valyl-tRNA synthetase. *Present address: Laboratoire du MeÂtabolisme CarboneÂ, DEVM, CEA de Cadarache, 13108 Saint Paul Lez Durance Cedex, France. ²Present address: Research Institute for Crop Production, Department of Molecular Genetics, Prague-6 Ruzyne, Czech Republic. Received 5 August 1999, revised 30 September 1999, accepted 1 October 1999

The translational apparatus in organelles is characterized by its prokaryotic nature. Its components, tRNAs, ribosomal RNAs, ribosomal proteins and aaRSs display striking similarities with their bacterial counterparts. Cytosolic and organelle aaRSs can be distinguished based on their substrate specificities [5]. Although there are some exceptions, the general rule is that organelle, as well as Escherichia coli, tRNAs are better substrates for organelle aaRSs than for cytosolic aaRSs and conversely, cytosolic tRNAs are poor substrates for prokaryotic and organelle enzymes. Because of this substrate specificity, one can predict that, in higher plants, cytosolic aaRSs have to be transported into mitochondria to efficiently aminoacylate tRNAs that are imported from the cytosol. This is the case for alanyl-tRNA synthetase (AlaRS) in Arabdopsis thaliana [6]. It was demonstrated that the same gene codes for the cytosolic and mitochondrial forms of the enzyme by the alternative use of two transcription initiation sites and two translation initiation codons, leading to the synthesis of a short form of the protein that remains in the cytosol and a long form with an N-terminal mitochondrial targeting peptide. A similar expression pattern exists for the yeast aminoacyl-tRNA synthetase genes HTS1 [7,8] and VAS1 [9], which code for histidyl-tRNA synthetase (HisRS) and valyl-tRNA synthetase (ValRS), respectively, and for other genes in yeast and in animals [10,11]. Here, we demonstrate that this expression pattern is also valid for two other A. thaliana genes coding for aaRSs specific for cytosolic tRNAs that are imported into mitochondria [12], namely ValRS and threonyl-tRNA synthetase (ThrRS), respectively. We developed a simple and rapid experimental protocol using the jellyfish Aequorea victoria green fluorescent protein (GFP) to demonstrate in vivo that the cytosolic and mitochondrial forms

q FEBS 1999

A. thaliana gene characterization using two in-frame AUGs (Eur. J. Biochem. 266) 849

of these enzymes are translated from the same transcript by the alternative use of two initiation codons. We also show that A. thaliana ValRS and ThrRS both have an NH2-appended domain that is also present in the yeast proteins. The evolutionary implications are discussed.

M AT E R I A L S A N D M E T H O D S Materials The cDNA library (Strasbourg, France) was obtained from the Arabidopsis Biological Resources Center (Ohio State University) and EST clone Z34023 was from F. Quigley (Grenoble, France). Molecular biology Cloning of DNA fragments in vectors was carried out according to standard procedures [13]. PCR amplifications were performed as described previously [14,15]. RACE amplification was performed using a Marathon Kit (Promega) according to the manufacturer's instructions. Cloning full-length ThrRS cDNA Cloning the full-length cDNA for ThrRS was a two-step procedure. An incomplete cDNA of ThrRS was obtained by

screening an A. thaliana cDNA library (Strasbourg, France) using an EST (accession number Z34023) with homology to known ThrRS genes. Because RACE did not allow isolation of the 5 0 -end of the ThrRS cDNA, the sequence of the 5 0 -end of the gene was obtained from A. thaliana genomic DNA using inverse (I)-PCR as described by Silver [16]. A subfragment of the I-PCR product, containing the first exon and the 5 0 -UTR of ThrRS was PCR amplified using oligonucleotides Th8(KpnI), ACTGCGGTACCATGCTCCTTCGTCTTACGGC and Th9(AflII), GATACATAAGAACATCGGCATC. Digested with KpnI and AflII, the fragment was ligated into the KpnI and AflII sites of the partial ThrRS cDNA clone to reconstruct a full-length cDNA (accession number Y14329). GFP work GFP fusions were made by ligating PCR-amplified presequences in the pCKGFP3A vector [15]. The following oligonucleotides were used: Va1(Afl III), GCCCTAGCAGGTTTTTTGTACATGTCA; Va2(BamHI), TAGGATCCGATGTTCTGGAACCGTGA; Va3(EcoRI), GCGAATTCTTTTTTGTCTATGTCACTAC; Va4(BamHI), TAGGATCCGACATTGTTCTGGAACCG; Th1(EcoRI), ATGAATTCGCCATGCTCCTTCGTCTTACGGCTC;

Fig. 1. Alignment of the NH2-ends of A. thaliana (at) ValRS (A) and ThrRS (B) sequences with their homologs of Saccharomyces cerevisiae (Sc), E. coli (Ec) and Rickettsia prowazeki (Rp). Mitochondrial transit peptides are shaded in light gray, NH2-appended domains are shaded in dark gray. Residues which are conserved between A. thaliana and other organisms are highlighted. The second in-frame methionine is boxed.

850 G. Souciet et al. (Eur. J. Biochem. 266)

Th2(BamHI), CACAGGATCCGCTGCTACTGTTGGAACAGTACAAA; Th3 (EcoRI), ACGAAGAATTCAACAAATTTTGATGCTCCTTCG; Th4(BamHI), CACAGGATCCGCTGCCATTGTTGGAACAGTACAAA. The pairs of primers used for each construct are shown in Fig. 3. For constructs pVal2 and pVal3, the PCR-amplified fragments were digested with AflIII and BamHI and ligated into the NcoI and BamHI sites of pCKGFP3A. For constructs pVal1, pVal4 and pThr1±4, the PCR-amplified fragments were digested with EcoRI and BamHI and ligated into the EcoRI and BamHI sites of pCKGFP3A after removal of the EcoRI/ BamHI fragment containing the translational leader sequence [17]. Expression of GFP constructs in cultured tobacco cells (Nicotiana tabacum var. Xanthi) was carried out as described previously [15]. GFP fluorescence was observed using a GFP long-pass filter set (excitation 460±500, dichroic mirror 505, long-pass emission 510). MitoTracker fluorescence was observed using a TRITC filter set (excitation 540/25, dichroic mirror 565, band-pass emission 605±55). Images were captured with a Power HAD 3DDD video camera (Sony, Tokyo) using Visiol@b 200 (Biocom, Paris) and were prepared using Adobe photoshop. Primer extension Total RNA was extracted from greenhouse-grown A. thaliana rosettes (ecotype Columbia) using a RNeasy Plant Mini Kit (Qiagen). Annealing of primers (Va5, CTTCTTTTTCCGCTCAAGCTC for ValRS and Th5, TTGATCGGATCGTGTGGAAGAGACT or Th7, GAAGAAGAAGTTGTGAAACGACG for ThrRS) was carried out by incubating 2 pmoles of endlabeled primer with 10 mg of total A. thaliana RNA at 90 8C in 50 mm Tris/HCl (pH 8.3), 7 mm MgCl2, 40 mm KCl, 10 mm dithiothreitol, 0.1 mg´mL21 bovine serum albumin for 5 min, followed by rapid cooling on ice. Primer extension was performed using the same buffer containing 1 mm of each dNTP and 200 units of M-MLV reverse transcriptase (Promega) for 1 h at 47 8C. Sequencing was performed according to the method described by Sanger et al. [18]. Computer analyses Sequence alignments were performed using the pileup program of the University of Wisconsin Genetics Computer Group software package. Amphiphilicity predictions were performed using mitoprot [19] and macvector software (Oxford Molecular, Oxford) and predictions of protein intracellular localization were done using psort (http:// psort.nibb.ac.jp).

R E S U LT S Characterization of ValRS and ThrRS cDNAs An A. thaliana ValRS gene (here termed AtSYV1) and the derived cDNA were cloned by Zhang and Somerville from the twn2 mutant, a T-DNA insertion mutant that is affected in early embryogenesis [20]. It was shown that ValRS gene expression in this mutant is affected by insertion of the T-DNA into the 5 0 -UTR of the gene. ValRS is a class I enzyme characterized by the two signature sequences HIGH and KMSKS [21]. High similarity between the cloned A. thaliana ValRS and the yeast and human cytosolic ValRS suggests that the A. thaliana gene

q FEBS 1999

Fig. 2. Transient expression of constructs pVal1 (top) and pThr1 (bottom) in tobacco cells. Contructs were bombarded into N. tabacum cells and expressed for 24 h. GFP fluorescence was observed using a 100 lens and a GFP band-pass filter set (left). The fluorescence of the mitochondrial specific dye (MitoTrackerTM) was observed in the same cells using a TRITC filter set (right). Images were processed using Adobe Photoshop 4.0. Scale bar, 10 mm.

codes for a cytosolic enzyme (Fig. 1A and data not shown). A. thaliana ThrRS (termed AtSYT1), like its homologs in other organisms [22], is a class II enzyme based on the presence of conserved motifs characteristic of this class [21]. High sequence similarity with eukaryotic enzymes suggests that the ThrRS encoded by the gene we isolated is also a cytosolic enzyme (Fig. 1B and data not shown).

Presence of a mitochondrial transit peptide Alignment of the N-termini of A. thaliana ValRS and ThrRS, deduced from the cDNA sequences, with their respective yeast and prokaryotic counterparts (Fig. 1) revealed the presence of an N-terminal extension as in yeast. The most distal part of the N-terminal extension of the yeast ValRS protein is a mitochondrial transit peptide. Interestingly, there is detectable sequence similarity between the yeast mitochondrial transit peptide and the A. thaliana N-terminal extension, suggesting that the AtSYV1 gene also codes for a mitochondrial protein. The N-terminal extension of ThrRS has the characteristics of a mitochondrial targeting peptide according to computer predictions using psort. To test whether the N-terminal peptides of the two aaRSs are functional mitochondrial targeting signals, we fused them to the GFP. The sequence coding for the 47 first amino acids of ValRS was amplified using primers Va1 and Va2 (Fig. 3). The reverse primer (Va2) was designed to remove the ATG codon at position 45 in order to prevent re-initiation of translation. The PCR fragment was cloned into the NcoI and BamHI sites of the pCKGFP3A vector, in place of the translation leader (TL). To amplify the ThrRS presequence, primers Th1 and Th2 were used. The second ATG was replaced by a GTA (Ala) codon. The amplified fragment, coding for the 36 first amino acids was cloned into the EcoRI and BamHI sites of pCKGFP3A. Transient expression of these constructs (constructs pVal1 and pThr1, Fig. 3) in cultured tobacco cells following biolistic transfection has revealed that both presequences are able to deliver GFP into mitochondria (Fig. 2). We can therefore conclude that both genes, the ValRS gene described by Zhang and Somerville [20] and the ThrRS cDNA

q FEBS 1999


Fig. 3. GFP fusion constructs. Nucleotide sequence of GFP fusions around the two in-frame AUGs in the different ValRS (top) and ThrRS (bottom) constructs. Nucleotide positions important for translation initiation are in bold and underlined. Linker amino acids are shown in bold, the first amino acids of the GFP are in italic.

we isolated, code for mitochondrial enzymes, although the proteins are more closely related to cytosolic enzymes. Alternative use of two in-frame initiation codons The sequence alignment in Fig. 1 shows that there is a methionine near the predicted end of the mitochondrial transit peptide in A. thaliana ValRS and ThrRS. This observation led us to investigate whether AtSYV1 and AtSYT1 could be bifunctional genes coding for the mitochondrial and cytosolic forms of ValRS and ThrRS by alternative use of two initiation codons. Initiation of translation is ruled by different factors such as the presence of a m7G cap, the length of the 5 0 -UTR, the secondary structure and the sequence context around the AUG initiation codon [11,23]. In plant mRNAs, purines at positions 23 and +4 (with preference for G at this position), relative to the first nucleotide of the initiation codon, are preferred for the efficient initiation of translation [24,25]. When other nucleotides are present at these positions, translation is poorly initiated, thus allowing the initiation complex to bypass the first AUG and initiate at a downstream AUG which is in a better sequence context. Analysis of the sequence around the two in-frame AUGs of the A. thaliana ValRS and ThrRS mRNAs indicates that the second AUGs are in a better context than the first (Fig. 3). To test in vivo whether the `leaky scanning' model is used in AtSYV1 and AtSYT1 gene expression, we made several GFP fusions where the context surrounding the first AUG was either the natural context (5 0 -UTR of the cDNA) or the vector sequence, which is optimal for translation. The second AUG was either present in its natural sequence context or absent (Fig. 3). Constructs pVal2 and pThr2 have the first and second AUG in the natural context. When these constructs are expressed in tobacco cells, the GFP is localized in the cytosol (Fig. 4) in the same manner as for control GFP (not shown), indicating that most of the translation is initiated at the second AUG. Because of the high intensity of fluorescence in the cytosol, it is not possible to see GFPlabeled mitochondria that could result from translation initiated

at the first AUG codon. To test whether translation initiation occurs at the first AUG when it is in its natural nonoptimal context, we used constructs pVal3 and pThr3, which are identical to constructs pVal2 and pThr2, except that the second AUG has been removed or replaced by another codon. Expression of these constructs in tobacco cells leads to the accumulation of GFP in mitochondria. However, this accumulation is much less intense than with constructs pVal1 and pThr1 and the sensitivity of the video camera had to be increased by a factor of 30 (by integrating 30 images) to obtain the image shown in Fig. 4. This indicates that translation initiation indeed occurs at the first AUG to synthesize the mitochondrial form of ValRS and ThrRS, but to a lesser extent than initiation at the second AUG. This is sufficient to provide

Fig. 4. Expression of GFP fusion constructs in tobacco cells. Constructs were expressed transiently in tobacco suspension cells following particle bombardment. V1, V2 and V3 correspond to constructs pVal2, pVal3 and pVal4, respectively, and T1, T2 and T3 to constructs pThr2, pThr3 and pThr4, respectively. GFP fluorescence was observed using a 100 lens and a GFP long-pass filter set. Scale bars, 10 mm. Chloroplasts identified by the red fluorescence of chlorophyll are visible in panels V4 and T3 (they are out of the focus plane in the other pictures).


q FEBS 1999

efficiently at the first AUG, there is no initiation at the second AUG and it also shows that the sequence context around the first AUG determines the amount of cytosolic enzyme that will be synthesized. Presence of multiple ValRS transcripts with different 5 0 -ends

Fig. 5. Primer extension analysis of ValRS mRNAs. Primer Va5 was end-labelled, annealed to total A. thaliana RNA and used to initiate cDNA synthesis. The primer extension reaction was loaded onto a sequencing gel. The major transcript ends are indicated by arrowheads. (Left) The primer extension products ending upstream of the first AUG and (right) those ending between the first and the second AUG. The end of primer extension product 1 could not be assigned more precisely in the sequence because it falls outside of the ValRS cDNA clone that was used to produce the sequencing ladder and no longer cDNA clone was available. The bottom part of the figure shows the nucleotide sequence of the ValRS cDNA relevant for these experiments. The sequence from which the oligonucleotide used in primer extension was derived is in italic and underlined.

mitochondria with the enzymes needed for their own translation, as mitochondrial aaRSs represent only a fraction of total cellular aaRSs. To demonstrate that the efficiency of translation initiation at the first AUG drives the synthesis of the cytosolic and mitochondrial forms of ValRS and ThrRS, we used constructs pVal4 and pThr4, where both AUGs are placed in a similar favorable context (A at position 23). These constructs lead to the accumulation of GFP in mitochondria with the same intensity as for contructs lacking the second AUG (pVal1 and pThr1). This indicates that, when translation is initiated

An alternative to the `leaky scanning' model to use two inframe AUGs for synthesis of two polypeptides from a single gene is to translate the two polypeptides from two messengers with different 5 0 -ends [7±9]. The mitochondrial form of the enzyme is translated from a messenger including the two initiation codons and the cytosolic enzyme from a shorter version of the same messenger lacking the first AUG. We checked whether several transcripts with different 5 0 -ends are transcribed from the ValRS and ThrRS genes by primer extension using total A. thaliana RNA as a template. Figure 5 shows a primer extension experiment using an antisense oligonucleotide (Va5) complementary to a sequence downstream of the second AUG of the A. thaliana ValRS. Four major transcript ends can be detected, two mapping upstream of the first AUG and two upstream of the second. These results indicate that several AtSYV1 transcripts with different 5 0 -ends are present in A. thaliana. This suggests that, on top of the translational regulation that we demonstrated through the GFP fusion experiments, transcriptional regulation could exist, provided that all the transcripts are functional in translation. However, it should be noted that one major transcript end maps only six nucleotides upstream of the second AUG, which is too short for efficient initiation of translation to occur at this AUG [11]. If this transcript is functional, it could be used to translate a protein from a downstream AUG at a position that lines up with the initiator methionine of the bacterial ValRS (Fig. 1). Similarly, the 5 0 -end of transcript 2 is probably too close to the first AUG for translation to initiate efficiently at this AUG. For ThrRS, we could not conclude whether two transcription initiation sites are used, because reverse transcription of the ThrRS mRNA stops prematurely before the second AUG, possibly due to the presence of a strong secondary structure. This may explain why no full-length cDNA could be isolated from the cDNA library and why the 5 0 -end of the cDNA could not be amplified by RACE. Using an oligonucleotide that anneals between the two AUGs, it was possible to extend reverse transcription 20 nucleotides past the first AUG (data not shown). The presence of conserved NH2-appended domains in yeast and A. thaliana ValRS and ThrRS The sequence alignment in Fig. 1 shows that in the yeast cytosolic/mitochondrial ValRS, distal to the mitochondrial

Fig. 6. Amphiphilic helical structure of the A. thaliana ValRS NH2-appended domain. (Left) A helical wheel projection of the ValRS N-terminal domain between residues 48 and 65 of the cytosolic protein. (Right) The hydrophobic moment of the same region of the protein calculated using a periodicity of 1008.

q FEBS 1999


targeting peptide, and in the yeast cytosolic ThrRS, there is an N-terminal extension compared with the prokaryotic enzymes. The yeast ValRS and ThrRS N-terminal-appended domains are rich in lysine (22%) and are predicted to form an amphiphilic helix [9]. The sequence alignment indicates that this domain is also present in the A. thaliana cytosolic ValRS and that it is predicted to be part of the mature mitochondrial protein. Computer analyses predict that the amphiphilic properties of the yeast NH2-extension are highly conserved in the corresponding domain of the A. thaliana protein (Fig. 6). The A. thaliana ThrRS also has a domain distal to the mitochondrial transit peptide that is absent from the bacterial proteins. This domain is only half the size of the yeast ThrRS NH2-appended domain but sequence similarities suggest that both peptides have a common origin and that part of the domain has been lost in the A. thaliana ThrRS.

DISCUSSION We present evidence that two A. thaliana genes, namely AtSYV1 and AtSYT1 coding for aaRSs specific for tRNAs that are imported into mitochondria, are bifunctional genes that code for both the cytosolic and the mitochondrial form of these enzymes. We demonstrate that `leaky scanning' past a first AUG, which is in an unfavorable sequence context, allows most initiation of translation to occur at a second in-frame AUG, leading to the synthesis of a short cytosolic form of the enzyme. Because the two gene products are targeted to different cell compartments, GFP fusions demonstrated in vivo the alternative use of the two in-frame initiation codons. This approach is simple and rapid and could be used to better characterize the nucleotide positions that are important in translation initiation. There is also evidence that transcriptional regulation of the expression of the two forms of the protein may occur for ValRS. Several AtSYV1 mRNAs with different 5 0 -ends are present in A. thaliana. We mapped two 5 0 -ends upstream of the first AUG and two upstream of the second. The shorter forms of the messenger would allow only for the synthesis of the cytosolic form of the enzyme. Such a pattern of expression may provide the cell with a handle on the regulation of the ratio of mitochondrial vs. cytosolic enzyme to adjust to its needs at different developmental stages. Primer extension results with ThrRS suggest that there is a strong secondary structure downstream of the second AUG that prevents reverse transcriptase extending past this site. A secondary structure downstream of an AUG can act as an enhancer for initiation of translation at the preceding AUG [11]. The pattern of expression of AtSYV1 and AtSYT1 in A. thaliana is reminiscent of the A. thaliana ALATS gene, which codes for the cytosolic and mitochondrial forms of AlaRS [6]. In the case of ALATS, this mode of expression is also correlated with the fact that cytosolic tRNAAla is imported into mitochondria and that it requires a cytosolic AlaRS to be efficiently aminoacylated in the organelles. The A. thaliana glycyl-tRNA synthetase gene, coding for an aaRS specific for another tRNA imported into mitochondria, is also a bifunctional gene [26]. Our data indicate that this type of expression is used for class I (ValRS) and class II (ThrRS and AlaRS) enzymes and that it is probably a general feature of aaRSs specific for tRNAs imported into mitochondria. The use of two in-frame initiation codons to translate a mitochondrial and a cytosolic enzyme from a same gene has been described for yeast aaRSs, namely ValRS (VAS1 gene; [9]) and HisRS (HTS1 gene; [8]). However, in the case of the yeast VAS1 gene, synthesis of the two forms of the enzyme occurs

only by the use of two different promoters that initiate transcription upstream of the first ATG and between the first and the second ATG, respectively. Unlike the plant aaRS mRNAs, both initiation codons are in an optimal sequence context. While, in plants, we have a correlation between the presence of bifunctional aaRS genes and the import of cytosolic tRNAs into mitochondria, this is not the case in yeast [9]. Neither tRNAVal nor tRNAHis is imported into mitochondria. However, a `loose' substrate specificity and/or the structure of mitochondrial tRNAVal and tRNAHis allows these cytosolic enzymes to charge the mitochondrial tRNAs. Our results show that, besides the conserved mode of expression of the yeast and A. thaliana ValRS genes, the structure of the protein in both organisms is also highly conserved. Sequence similarity between yeast and A. thaliana proteins extends into the mitochondrial transit peptide. This is particularly striking, given that the primary structure of mitochondrial targeting peptides varies greatly from one protein to another [27,28]. In addition, both proteins have a conserved lysine-rich N-terminal extension with an amphiphilic domain distal to the mitochondrial transit peptide. This appended domain, absent in the prokaryotic enzymes, is present in the cytosolic and in the mature mitochondrial form of the enzyme, provided that it is not cleaved in mitochondria. The fact that the ValRS gene is bifunctional in yeast and A. thaliana and the high sequence conservation between the two proteins, even in domains that are generally highly variable (the transit peptide), suggest that bifunctionality of the gene predates the divergence of yeast and A. thaliana. If this is true, the cytosolic ValRS must have been present very early in mitochondria, which might allow tRNAVal to be one of the first cytosolic tRNAs to operate in higher plant mitochondria. This is not the case for ThrRS, the N-terminal extension found in yeast cytosolic ThrRS is conserved only partially in the A. thaliana protein and the yeast cytosolic and mitochondrial ThrRSs are coded for by distinct genes [29,30]. Most of yeast aaRSs possess an N-terminal extension compared with their prokaryotic counterparts. These extensions, the role of which is not clearly defined, display polyanion-binding properties [31] and may have amphiphilic properties [9]. They may be involved in non-specific interactions with tRNAs [32,33] or in the formation of complexes with proteins. In the human cytosolic ValRS, an N-terminal extension of 200 residues, which confers hydrophobic properties on the enzyme, is involved in the formation of a complex with elongation factor EF-1H, a complex which is regulated by EF-1a [34]. There is no evidence that the yeast ValRS N-terminal extension has a similar function. Because it is not present in the bacterial enzymes it is possible that it does not have a function in catalysis, but the fact that it is conserved between yeast and A. thaliana strongly suggests that it may have a function, possibly in translation. The question remains whether this sequence is also part of the mature ValRS in mitochondria and whether this eukaryotic-specific feature could play a role in a prokaryotic-like translation system.

ACKNOWLEDGEMENTS This work was supported by the Centre National de la Recherche Scientifique (CNRS), the UniversiteÂ Louis Pasteur and a grant from the Groupement de Recherche et d'Etude des GeÂnomes (GREG). We thank J. Zhang and C. Somerville for the gift of the ValRS cDNA and F. Quigley for providing us with the ThrRS EST clone. We thank Ian Small and Laurence MareÂchal-Drouard for helpful discussions.


q FEBS 1999

REFERENCES 1. MareÂchal-Drouard, L., Weil, J.H. & Dietrich, A. (1993) Transfer RNAs and transfer RNA genes in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 44, 13±32. 2. Dietrich, A., Weil, J.H. & MareÂchal-Drouard, L. (1992) Nuclearencoded transfer RNAs in plant mitochondria. Annu. Rev. Cell Biol. 8, 115±131. 3. Dietrich, A. et al. (1993) The plant mitochondrial transfer RNA population: a mosaic of species with different genetic origins. In The Translational Apparatus, Structure, Function, Regulation, Evolution (Nierhaus, K.H., Franceschi, F., Subramanian, A.R., Erdmann, V.A., Wittmann-Liebold, B., eds), pp. 617±626. Plenum Press, New York, USA. 4. Dietrich, A., MareÂchal-Drouard, L. & Weil, J.H. (1994) The plant mitochondrial transfer RNAs have three different genetic origins. In Plant Molecular Biology, Molecular Genetic Analysis of Plant Development and Metabolism, NATO ASI Series H: Cell Biology, Vol. 81 (Coruzzi, G. & PuigdomeÁnech, P., eds), pp. 333±345. Springer-Verlag, Berlin, Germany. 5. Steinmetz, A. & Weil, J.H. (1986) Isolation and characterization of chloroplast and cytoplasmic transfer RNAs. Methods Enzymol. 118, 212±231. 6. Mireau, H., Lancelin, D. & Small, I. (1996) The same Arabidopsis gene encodes both cytosolic and mitochondrial alanyl-tRNA synthetase. Plant Cell 8, 1027±1039. 7. Chiu, M.I., Mason, T.L. & Fink, G.R. (1992) HTS1 encodes both the cytoplasmic and mitochondrial histidyl-transfer RNA synthetase of Saccharomyces cerevisiae ± mutations alter the specificity of compartmentation. Genetics 132, 987±1001. 8. Natsoulis, G., Hilger, F. & Fink, G.R. (1986) The HTS1 gene encodes both the cytoplasmic and mitochondrial histidine tRNA synthetases of S. cerevisiae. Cell 46, 235±243. 9. Chatton, B., Walter, P., Ebel, J.P., Lacroute, F. & Fasiolo, F. (1988) The yeast VAS1 gene encodes both mitochondrial and cytoplasmic valyltRNA synthetases. J. Biol. Chem. 263, 52±57. 10. Small, I., Wintz, H., Akashi, K. & Mireau, H. (1998) Two birds with one stone: genes that encode products targeted to two or more compartments. Plant Mol. Biol. 38, 265±277. 11. Kozak, M. (1991) Analysis of vertebrate mRNA sequences: intimation of translational control. J. Cell Biol. 115, 887±903. 12. MareÂchal-Drouard, L., Small, I., Weil, J.H. & Dietrich, A. (1995) Transfer RNA import into plant mitochondria. Methods Enzymol. 260, 310±327. 13. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, USA. 14. Wintz, H. (1994) Analysis of heteroplasmy in a cytoplasmic mutant of maize. Plant Physiol. Biochem. 32, 649±653. 15. Menand, B., MareÂchal-Drouard, L., Sakamoto, W., Dietrich, A. & Wintz, H. (1998) A single gene of chloroplast origin codes for mitochondrial and chloroplastic methionyl-tRNA synthetase. Proc. Natl Acad. Sci. USA 95, 11014±11019. 16. Silver, J. (1991) Inverse polymerase chain reaction. In PCR, A Practical Approach (McPherson, M.J., Quirke, P. & Taylor, G.R., eds), pp. 140±141. IRL Press, Oxford, UK. 17. Reichel, C., Methur, J., Eckes, P., Langenkemper, K., Koncz, C., Schell,

18. 19. 20.

21. 22.

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

J., Reiss, B. & Maas, C. (1996) Enhanced green fluorescence by the expression of an Aequorea victoria green fluorescent protein mutant in mono- and dicotyledonous plant cells. Proc. Natl Acad. Sci. USA 93, 5888±5893. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 74, 5463±5467. Claros, M.G. & Vincens, P. (1996) Computational method to predict mitochondrially imported proteins and their targeting sequences. Eur. J. Biochem. 241, 779±786. Zhang, J.Z. & Somerville, C.R. (1997) Suspensor-derived polyembryony caused by altered expression of valyl-tRNA synthetase in the twn2 mutant of Arabidopsis. Proc. Natl Acad. Sci. USA 94, 7349±7355. Eriani, G., Delarue, M., Poch, O., Gangloff, J. & Moras, D. (1990) Partition of tRNA synthetases into two classes based on mutually exclusive sets of sequence motifs. Nature 347, 203±206. Meinnel, T., Mechulam, Y. & Blanquet, S. (1995) Aminoacyl-tRNA synthetases: occurence, structure, and function. In tRNA: Structure, Biosynthesis, and Function (SoÈll, D. & RajBandhary, U., eds), pp. 251±292. American Society for Microbiology, Washington, DC, USA. Kozak, M. (1995) Adherence to the first-AUG rule when a second AUG codon follows closely upon the first. Proc. Natl Acad. Sci. USA 92, 2662±2666. Joshi, C.P., Zhou, H., Huang, X. & Chiang, V. (1997) Context sequences of translation initiation codon in plants. Plant Mol. Biol. 35, 993±1001. Luehrsen, K.R. & Walbot, V. (1994) The impact of AUG start codon context on maize gene expression in vivo. Plant Cell Report 13, 454±458. DucheÃne, A.M. & Dietrich, A. (1997) Isolation and characterization of a glycyl-tRNA synthetase sequence from Arabidopsis thaliana. Plant Physiol. 115, 1730. Glaser, E., SjoÈling, S., Tanudji, M. & Whelan, J. (1998) Mitochondrial protein import in plants. Plant Mol. Biol. 38, 311±338. Von Heijne, G., Steppuhn, J. & Herrmann, R.G. (1989) Domain structure of mitochondrial and chloroplast targeting peptides. Eur. J. Biochem. 180, 535±545. Pape, L.K. & Tzagoloff, L.K. (1985) Cloning and characterization of the gene for the yeast cytoplasmic threonyl-tRNA synthetase. Nucleic Acids Res. 13, 6171±6183. Pape, L.K., Koerner, T.J. & Tzagoloff, A. (1985) Characterization of a yeast nuclear gene (MST1) coding for the mitochondrial threonyltRNA synthetase. J. Biol. Chem. 260, 15362±15370. Cirakoglou, B. & Waller, J.P. (1985) Do yeast aminoacyl-tRNA synthetases exist as soluble enzymes in the cytoplasm. Eur. J. Biochem. 149, 353±361. Whelihan, E.F. & Schimmel, P. (1997) Rescuing an essential enzyme± RNA complex with a non-essential appended domain. EMBO J. 16, 2968±2974. Wang, C.C. & Schimmel, P. (1999) Species barrier to RNA recognition overcome with non-specific RNA binding domains. J. Biol. Chem. 274, 16508±16512. Negrutski, B.S., Shalak, V.F., Kerjan, P., El'skaya, A.V. & Mirande, M. (1997) Functional interaction of mammalian valyl-tRNA synthetase with elongation factor EF-1a in the complex with EF-1H. J. Biol. Chem. 274, 4545±4550.