Identification of an amino acid region supporting specific methionyl-tRNA synthetase: tRNA recognition

J. Mol. Biol. (1989) 208, 429443

Identification of an Amino Acid Region Supporting Specific Methionyl-tRNA Synthetase : tRNA Recognition Patrice Mellot, Yves Mechulam, Daniel Le Corre Sylvain Blanquet and Guy Fayatt Laboratoire de Biochimie UA 240 du CNRS, Ecole Polytechnique 91128 Palaiseau CEDEX, France (Received

9 November

1988, and in revised form

17 January

1989)

Site-directed nuclease digestion and nonsense mutations of the Escherichia coli metG gene were used to produce a series of C-terminal truncated methionyl-tRNA synthetases. Genetic complementation studies and characterization of the truncated enzymes establish that the methionyl-tRNA synthetase polypeptide (676 residues) can be reduced to 547 residues without significant effect on either the activity or the stability of the enzyme. The truncated enzyme (M547) appears to be similar to a previously described fully active monomeric form of 64,000 M, derived from the native homodimeric methionyl-tRNA synthetase (2 x 76,000 &I,) by limited trypsinolysis in vitro. According to the crystallographic three-dimensional structure at 2.5 A resolution of this trypsin-modified enzyme, the polypeptide backbone folds into two domains. The former, the N-domain, contains a crevice that is believed to bind ATP. The latter, the C-domain, has a 28 C-residue extension (520 to 547), which folds back, towards the N-domain and forms an arm linking the two domains. This study shows that upon progressive shortening of this C-terminal extension, the enzyme thermostability decreases. This observation, combined with the study of several point mutations, allows us to propose that the link made by the C-terminal arm of M547 between its N and C-terminal domains is essential to sustain an active enzyme conformation. Moreover, directing point mutations in the 528-533 region, which overhangs the putative ATP-binding site, demonstrates that this part of the C-terminal arm participates also in the specific complexat’ion of methionyl-tRNA synthetase with its cognate tRNAs.

tRNA synthetase from Bacillus stearothermophilus, a dimer of the a2 type exhibiting half-of-the-sites activity (Fersht, 1987, and references therein), extensive enzymologic studies and genetic engineering approaches have shown that the formation of a 1 : 1 specific tRNA : enzyme complex required an intact dimeric enzyme structure (Dessenet al., 1982: Bedouelle & Winter, 1986; Carter et al.. 1986). Though Escherichia coli native methionyl-tRNA synthetase is also a homodimer (2 x 76,000 M,), it can he irreversibly dissociated upon removal of about 130 residuesfrom its C terminus by controlled trypsinolysis (Cassio & Waller, 1971). The resulting monomer (64,000 M,) is fully active for tRNA aminoacylation. The three-dimensional structure of this proteolysed active 64,000 J4, monomeric fragment is presently solved at 2.5 A (1 A = 91 nm) resolution (Zelwer et al., 1982; Brunie et al.. 1987).

1. Introduction The specific recognition and aminoacylation of a tRNA species by its cognate aminoacyl-tRNA synthetase play a key role in the translation process. How a synthetase discriminates among the cellular tRNAs remains an intriguing question. Genetic engineering combined with structural analysis seems now one of the most powerful approaches to decipher the basis of this specificity at the atomic level. To date, high-resolution X-ray crystallographic data have been reported for two bacterial and aminoacyl-tRNA synthetases: methionyl tyrosyl-tRNA synthetases. In the case of tyrosylt Author

to whom

all

correspondenceshould be

addressed. (x)42-2836/r39/180429-15

$03.00/O

429

0

1989

Academic

Press

Limited

430

I’. Mellot

The catalytic mechanisms of tryptic monomeric (MTStt) and native dimeric methionyl-tRNA synthetases (MTS) have been extensively studied and compared. The analyses performed could not differentiate between these two forms at the level of the amino acid activation step (Rlanquet et al., 1974; Hyafil et al., 1976). Moreover, both native and tryptic enzymes can interact specifically with methionine-specific transfer RNAs, with a maximum of one tRNA bound per polypeptide chain (Blanquet et al., 1973a). However, the native dimer binds two tRNA molecules in an anti-co-operative manner. Such behaviour is thought to be able to regulate the turnover of the synthetase in the tRNA aminoacylation reaction, by coupling the rate of release of a bound aminoacylated tRNA to the concentration of free unaminoacylated tRNA (Blanquet et al., 1976). Detailed analyses have demonstrated that anti-co-operativity does not arise from asymmetric assembly of the enzyme subunits and the conclusion has been reached that trypsin-modified methionyltRNA synthetase is functionally a fully representative monomeric model of the native enzyme (Rlanquet et al., 1979, and references therein). To achieve a model of the MTS : tRNA complex, our approach is to identify the points of contact, between these two macromolecules by using sitedirected mutagenesis. In order to simplify the problem, the study of a monomeric form of the enzyme appeared preferable. Firstly, it was therefore necessary to reproduce, by genetic engineering, a truncated MTS similar to the tryptic fragment (560( Ifr 25) residues). Using the set of previously described specialized metG vectors expressing methionyl-tRNA synthetase (Dardel et al., 1984; Hire1 et al., 1988), several tenths of trun(sated forms of MTS were characterized in the present study. It is concluded that the minimal size preserving the activity of the enzyme corresponds, indeed, to that of the trypsin-modified MTSt. As assessed by its ability to support growt’h of a metG lnutant, this truncated monomeric MTS posesses all the elements for discriminating among cellular tRNAs. Secondlv. introduction of short deletions and point mutations in the C-terminal end of this kuncated enzyme demonstrates that the 528-547 region is involved in the stability of the enzyme structure. Tnterestingly, it is shown that a part of this region, the 528-533 segment, participates also in the formation of the specific complex between methionyl-tRNA synthetase and each of its cognate tRNAs.

2. Materials and Methods (a) Strains

and

general

techniques

E. coli strain JMlOlTr vative various

of JMlOl plasmids

(Hire1 et al., 1988), a recA deri(Messing, 1983), was used to host the and Ml3 phages studied. Strain

t Abbreviations used; MT%, tryptic monomeric methionyl-tRNA synthetase; MTS, native dimeric methionyltRNA synthetase; 3-D, three-dimensional; bp, basepair(s); PAGE, polyacrylamide gel electrophoresis.

et al l’a11803.7 was constructed bv conjugation of l’allX03.5 (Dardel et aZ., 1984) with the P+ strain G(‘4585 (Huisman rt al., 1982), selecting for episome transfer. l’allH03.7 is therefore an F’ recA strain carrying t,he r&G83 allele. Plasmids and phages carrying met{: @NAY6 and M13metC) are described elsewhere (Hire1 et (cl., I!+#). General genetic techniques were as described by Miller (1972) and Davis et (~1. (1980). Recombinant DNA techniques were as described by Maniatis et al. (1982) and Silhavy et al. (1984). When required, 1)NA reatrict,ion fragments were purified by high-performance liquid chromatography (Schmitter et al., 1986). Nucleotides, DNA restriction and modification enzymes were purchased from Boehringer-Mannheim, Pharmacia France. or Uenofit. (b) Oligonucleotide site-directed mutagenesis Oligonucleotides were automatically synthesized by the /?-cyanoethyl-phosphoramidite method using a Pharmacia gene assembler. Oligonucleotides were purified by ionexchange chromatography on a Mono-Q HR5/5 (fast protein liquid chromatography, Pharmacia) followed by dialysis and concentration. Uracil-containing M13metG or M13M547 single-stranded DNA, prepared according to the method of Kunkel (1985), was used as a template fol oligonucleotide site-directed mutagenesis. @5 pmol of phosphorylated oligonucleotide was annealed to @17 pmol of M13metG single-stranded DNA in a SO ~1 mixture (10 mw-Tris.HCl (pH 8). 10 mM-MgCl,) by a 5 min incubation at 55°C followed by cooling to 30°C. Synthesis and ligation of the complementary strand was achieved bv adding @2 mM of each dNTP, 0.2 mM-ATP, 1 unit of i. coli DNA polymerase T (Klenow fragment) and 10 units of T4 DNA ligase in a final volume of 50 ~1. After I h of incubation at 30”(‘, portions of the reaction mixture were used to transfect competent ,JMlOlTr cells. Mutant phages were screened by dot-blot hybridization analysis (Wallace et al., 1981) using the mutagenic oligonucleotide as a probe. Ml3 phages carrying the desired mutations were finally characterized by 1)NA srqtwnc*ing of the entire metC gene. The mut,ant proteins were named according to the following rules: the letter .M is followtd by the number of residues of the polypeptidr as exprt~sst~d from each of the modified genes; when a IJoint rnutat,ioll was added. t’he preceding code was completed by the amino acid transition followed by its position on t,h(J tnaturt, polypept,ide. For instance. M547YA531 dt$inrs it rncathionyl-tRNA synt,hetast: of 547 residurs. in which tht> tyrosine at position 531 was replaced by an alaninr residue. According to this procedure, the following mutant M13metG phages were constructed: M 13M551 (opal caodorr at position 553 of the DNA coding sequence. ~orresponding to the 552nd residue of the amino acid sccluem~t~ because the initiator methionine is post-translationally removed). M 13M547 (opal codon at posit,ion 549). M13M542 (opal caodon at position 544), M13MFi37 (opal codon at position 539), M13M533 ( 0 I )d,* 1 codon al) position 535), 1MlSM532 (opal rodon at position ~34). ITsing M13M547 as a template for mutagenesis, M13M547VL543 was constructed by substitution of the Val codon at position 544 by a Leu one ((:TG + TTA): M13M547YF531, M13M547YV531 and M13M547YX531 resulted from the substitution of the Tyr codon (TAT) at position ,531 by a l’he (TTT). Val (UTA) or Ala (U(:A) codon, respectively. (c) Plasmid Phage nuclease

MlSmetGPB, digestion was

constructions

used as constructed

a substrate for RaL31 t)y inserting the 439

Structure-Activity

Relationships

base-pairs (bp) Pstl-BumHI fragment internal to m&G in the corresponding sites of Ml3mp8. Phasmid pBSmetG was constructed by cloning a 2400 bp XbaI-BmaI fragment from MlYmetG between the XbaI and Hind11 sites of Bluescript M13-KS (Stratagene, U.S.A.). This facilitates the exchange of any region of the structural part of the me@ gene with its mutated counterpart by making unique the P&I, BumHI and Hind111 sites. pBSM537, pBSM547VL543, pBSM547 and pBSM547in53 1 (see Results) were derived from pBSmetG after replacement of its PstI--BumHI fragment by the equivalent mutated fragments from M13M537, M13M547VL543. M13M547 and M13M547in531, respectively. (d)

Isolation

of Bal32 deletions site

around of MlSmetGPB

the

unique

BamHI

BamHI-restricted M13metGPB replicative form DNA (20 pg) was treated with 1.2 units of BaE31 exonuclease (Bethesda Research Laboratories) at 30°C in a 200 pl reaction volume containing 12 mM-CaCl,, 12 mM-MgCl,. @2 M-NaCl. 20 mmTris. HCl (pH S), and 1 mm NaEDTA. The reaction was stopped after 3 min, 3.5 min or 4 min, by adding EGTA (pH 8) to a final concentration of 15 mM. After extractions with phenol and ether followed by precipitation with ethanol, the DNA was repaired using the Klenow fragment of DNA polymerase I. The DNA was then recircularized in the presence of a BamHI synthetic linker (5’-CCGGATCCGG-3’). and used to transfect JMlOlTr. Phage clones were analysed by DNA restriction mapping and sequencing (Sanger et al.. 1980). Sixty clones were cahosen for further analysis. For this purpose, the truncated PstI-BamHI fragments were transferred between the corresponding sites of the pNAV6 shuttle vertor. to reconstitute the 5’.terminal end of met