Eur. J . Biochem. 207, 1 - 11 (1992) (0 FEBS 1992

The polyanion-binding domain of cytoplasmic Lys-tRNA synthetase from Saccharomyces cerevisiae is not essential for cell viability

Ricardo MARTINEZ and Marc MIRANDE

Laboratoire d’Enzymologie du Centre National de la Recherche Scientifique, Gif-sur-Yvette, France

(Received November 1, 1991/February 5 , 1992) - EJB 91 1477

Cytoplasmic Lys-tRNA synthetase (LysRS) from Saccharomyces cerevisiae is a dimeric enzyme made up of identical subunits of 68 kDa. By limited proteolysis, this enzyme can be converted to a truncated dimer without loss of activity. Whereas the native enzyme strongly interacts with polyanionic carriers, the modified form displays reduced binding properties. KRSl is the structural gene for yeast cytoplasmic LysRS. It encodes a polypeptide with an amino-terminal extension composed of about 60-70 amino acid residues, compared to its prokaryotic counterpart. This segment, containing 13 lysine residues, is removed upon proteolytic treatment of the native enzyme. The aim of the present study was to probe in vivo the significance of this amino-terminal extension. We have constructed derivatives of the KRSI gene, encoding enzymes lacking 58 or 69 amino-terminal residues and, by site-directed mutagenesis, we have changed four or eight lysine residues from the amino-terminal segment of LysRS into glutamic acids. Engineered proteins were expressed in vivo after replacement of the wild-type KRSl allele. The mutant enzymes displayed reduced specific activities ( 2 - 100-fold). A series of carboxy-terminal deletions, encompassing 3,lO or 15 amino acids, were introduced into the LysRS mutants with modified amino-terminal extensions. The removal of three residues led to a 2-7-fo1d increase in the specific activity of the mutant enzymes. This partial compensatory effect suggests that interactions between the two extreme regions of yeast LysRS are required for a proper conformation of the native enzyme. All K R S l derivatives were able to sustain growth of yeast cells, although the mutant cell lines displaying a low LysRS activity grew more slowly. The expression, as single-copy genes, of mutant enzymes with a complete deletion of the amino- terminal extension or with four Lys + Glu mutations, that displayed specific activities close to that of the wild-type LysRS, had no discernable effect on cell growth. We conclude that the polycationic extensions of eukaryotic aminoacyl-tRNA synthetases are dispensable, in vivo, for aminoacylation activities. The results are discussed in relation to the triggering role in in situ compartmentalization of protein synthesis that has been ascribed to the polypeptide-chain extensions that characterize most, if not all, eukaryotic aminoacyl-tRNA synthetases.

In Succharomyces cerevisiae, cytoplasmic and mitochon- drial Lys-tRNA synthetase (LysRS) are encoded by two dis- tinct genes (Mirande and Waller 1988; Gatti and Tzagoloff, 1991). The cytoplasmic enzyme is a dimer made of identical subunits of 68 kDa. Proteolytic, yet fully active, forms of this enzyme can be generated by controlled cleavage with elastase and papain (Cirakoglu and Waller, 1985). It was shown that the elastase-modified form results from the removal of 57 amino acids located at the amino-terminus of the protein (Mirande and Waller, 1988). Escherichia coli LysRS (lysS and lysU) species differ from the yeast enzyme by the virtual ab- sence of this polypeptide segment (LevEque et al., 1990). The yeast mitochondrial LysRS (Gatti and Tzagoloff, 1991) also possesses an amino-terminal extension, compared to the E. coli enzymes. This polypeptide segment, which does not ex- hibit sequence similarity with the corresponding region of the

Correspondence to M. Mirande, Laboratoire d’Enzymologie, Cen- tre National de la Recherche Scientifique, F-91190 Gif-sur-Yvette, France

Abbreviations. PCR, polymerase chain reaction; LysRS, Lys- tRNA synthetase.

Enzymes. Lys-tRNA synthetase (EC 6.1.1.6).

yeast cytoplasmic LysRS, should be involved in mitochondria1 targeting.

The native yeast cytoplasmic LysRS interacts strongly with polyanionic carriers (Cirakoglu and Waller, 1985). This prop- erty was lost upon elastase treatment, generating an enzyme similar in size to its prokaryotic counterpart. The polyanion- binding property was ascribed to the clustering of lysine resi- dues within the amino-terminal polypeptide extension that characterizes the yeast enzyme (Mirande and Waller, 1988). The biological function, in vivo, of the additional segment of yeast LysRS, dispensable in vitro for the aminoacylation function, was not deciphered.

As a general rule, yeast cytoplasmic aminoacyl-tRNA synthetases are larger than their homologous bacterial en- zymes (Schimmel, 1987). Comparison of the known primary structures of bacterial and yeast aminoacyl-tRNA synthetases revealed that eukaryotic enzymes possess a polyanion-binding domain (Mirande and Waller, 1988; Mirande, 1991). In all cases so far examined, it was shown that this polypeptide extension is unrelated to catalysis (Cirakoglu and Waller, 1985; Ludmerer and Schimmel, 1987; Walter et al., 1989; Eriani et al., 1991). To test for the functional significance of

2

these domains, in vivo approaches were undertaken. Internal deletions within the large amino-terminal extension of yeast Gln-tRNA synthetase (Ludmerer and Schimmel, 1987) or a complete deletion of the polypeptide extension of yeast Met- tRNA synthetase (Walter et al., 1989) did not affect cell vi- ability. These results established that the polyanion-binding domains are not essential in vivo. With regard to the general occurrence of these polypeptide-chain extensions in eukary- otic aminoacyl-tRNA synthetases, one can speculate that they fulfill an important, yet unknown, biological function. A trig- gering role in spatial organization of the protein-biosynthesis machinery, through electrostatic interactions with subcellular structures, has been proposed (Cirakoglu and Waller, 1985).

To probe in vivo the functional significance of the amino- terminal extension of yeast LysRS, we have performed a series of deletions and mutations within the polyanion-binding do- main of this enzyme. We first established that a small carboxy- terminal deletion is required to partially compensate for a large amino-terminal deletion. The effect of a complete de- letion of the polypeptide-chain extension, or of lysine to glutamate mutations within the polycationic domain, was then monitored in vivo, after insertion/replacement of the wild-type KRSI allele.

MATERIALS AND METHODS

Materials

Restriction endonucleases and DNA-modifying enzymes were purchased from Boehringer (Mannheim), New England Biolabs, Pharmacia or Perkin-Elmer Cetus, and used as rec- ommended by the suppliers. Radionucleotides and 14C- labeled amino acids were from Amersham and Commissariat a I'Energie Atomique, respectively. Antibodies to the papain- modified form of yeast cytoplasmic LysRS were prepared by P. Kerjan.

MI3 and pUC vectors (Yanisch-Perron et al., 1985) were used for subcloning, and amplified in E. coli JMlOlTr [A(lac,pro), supE, thi-I, recA 56, srl300:: TnlO, F( t raD36, proAB, l a d q , lucZ AM15) ; a gift from Y. Mechulam, Luhora- toire de Biochimie, Ecole Polytechnique].

The selectable S. cerevisiae genes TRPI and LEU2 were taken from plasmids pEMBLYr25 (Baldari and Cesareni, 1985) and pUC9-LEU2 (pUC9 bearing a 2.2-kb yeast DNA fragment containing the LEU2 gene, gift from D. Thomas, Lahorutoire d'Enzymologie, Gif-sur-Yvette). The shuttle vec- tor pEMBLYe23 (Baldari and Cesareni, 1985) contains the selectable genes URA3 and hlu. The haploid yeast strain CC499-7A (leu2, trpl, ura3, arg'), a segregant from a cross of MG409 (MATol, argl'; Delforge et al., 1975) by CC454-14B (MATu, leu.2, trpl , ura3), is a gift from Y. Surdin-Kerjan (Laboratoire d'Enzyrnologie, Gif-sur-Yvette).

Analytical procedures

Media for growth of S. cerevisiae were prepared according to Sherman et al. (1986). Crude extracts from exponentially growing yeast cells were prepared as previously described (Mirande and Waller, 1988). Aminoacyl-tRNA synthetases were assayed by the aminoacylation of tRNA (Kellermann et al., 1982). Crude yeast tRNA (Boehringer) was used (tRNA1,YS; AZ6, = 1 corresponds to 60 pmol). For the determi- nation of K,,, for tRNA in the aminoacylation reaction, tRNA concentrations of 0.08 - 5 pM were used; concentrations refer to lysine-acceptor tRNA. K,,, and V,,,,,/E values were deduced

from double-reciprocal Lineweaver-Burke plots. Protein con- centration was determined according to Gornall et al. (1949).

SDSjPAGE was performed according to Laemmli (1970) on 10% acrylamide gels. Western blots were conducted ac- cording to Harlow and Lane (1 988), using peroxidase-labeled goat anti-(rabbit IgG) antibodies (Biosys, Compeigne).

Standard procedures were used for DNA manipulations and transformation of E. coli (Sambrook et al., 1989). Plasmids were amplified in E. coli JMlOlTr and purified by centrifugation on CsCl gradients. The relevant nucleotide se- quence of the constructs was verified by the dideoxy- 'nucleotide-chain-termination method (Sanger et al., 1977) using KRSI-specific oligonucleotide primers and the T7 sequencing kit from Pharmacia. Yeast cells were transformed by the lithium chloride method (Ito et al., 1983).

Polymerase chain reaction (PCR) experiments were performed with 5 - 10 ng plasmid DNA and 100 pmol of each oligonucleotide primer, in a Perkin-Elmer Cetus thermocycler, as recommended by the manufacturer. Reaction mixtures were subjected to 25 cycles of 1 min at 9 4 T , 2 min at 50°C and 3 min at 72°C.

Total RNA was electrophoresed on formaldehyde/agarose gels (Sambrook et al., 1989), transferred on nylon membranes (NY 13 N from Schleicher & Schuell) by vacuum blotting (VacuGene, Pharmacia) and probed with [32P]DNA frag- ments, radiolabeled by random oligonucleotide priming (Hodgson and Fisk, 1987).

Determination of the in vivo tRNA aminoacylation level

Cells exponentially growing in 1 1 YPG medium (A,,, = 1 .O) were poured over 500 ml ice (- 20°C) composed of 0.1 M sodium acetate, pH 4.5. After centrifugation, cells were resus- pended in 10 ml ice-cold 50 mM sodium acetate, pH 4.5, and 150 mM NaC1. An equal volume of phenol equilibrated in the same buffer was added. The mixture was shaken vigorously for 30 s and chilled on ice for 30 s; this treatment was repeated 10 times. After centrifugation, the aqueous phase was submit- ted to a second phenol-extraction procedure followed by chloroform extraction. After ethanol precipitation, the RNA was dissolved in 1.0 ml 10 mM sodium acetate, pH 4.5, and 10 mM MgCI2. A fraction (0.5 ml) wdS oxidized with a 200- fold molar excess of sodium periodate for 15 min (at 25 "C, in the dark) followed by a further 10-min incubation in the presence of 2 M ethylene glycol. After ethanol precipitation of the two samples (oxidized and non-oxidized RNA fractions), RNA pellets were washed by two additional precipitations with ethanol after dissolving in 50 mM sodium acetate, pH 4.5, and 150 mM NaC1. RNA was dissolved in 1.0 m12 M Tris/acetate, pH 8.0; deacylation was conducted at 35 "C for 30 min. After ethanol precipitation, the RNA was taken up in 0.1 ml H 2 0 . Aminoacylation of tRNA from the periodate- treated and untreated fractions was performed as described above for the aminoacyl-tRNA synthetase assay, using satu- rating amounts of purified LysRS and Val-tRNA synthetase from S. cerevisiae (gifts from P. Kerjan and G. Bec, Laboratoi- re d'Enzymologie, Gif-sur-Yvette) or Met-tRNA synthetase from sheep liver (Lazard et al., 1985).

Construction of the integrative plasmid pYK

Yeast genomic DNA fragments were taken from phages LLysl (ILys9), ILys8 and LLys17 (Mirande et al., 1986). The combined overlapping sequences (Fig. 1) account for 5180 nucleotides (Mirande and Waller, 1988; Martinez et al., 1991).

3

hLysl7 ! hLys8

I

: BS B s H H E E H Bt E Bs B E E Sa Sn S Bs Bs

I KRSl u N 4

hLys8 pUC18 1- Hpal +Sac1 1- PstI + Smal 2- T4 DNA POI. 4dXTPs 2- T4 DNA poi, 4dXTPs

1 1 Ligation

SP

pYK03

, 1 I

I pEMBLYr25

2- DNA pol (Klenow). 4dXTPs 1- BstEIl

SP

1. Bsml 2. T4 DNA pol, 4dXTPs 3- Ligation of a Xhol linker I +

pYK04

hLys9 p U C l 8 1- EcoRl 1- BamHl 2- DNA pol (Klenow), 4dXTPs 3 Sall 3- Sall

2- DNA pol (Klenow). 4dXTPs

L Ligation 1 Sm

pYKO6

I pUC9-LEU2 I SnaBl I 1- Sall + Smal

Sm

pYK07

i Insertion of the Sphl-Sail fragment of pYKM into (Sphl-Sal1)digested pYK07

ORFX TRPl KRS I LEU2 PMRP

Fig. 1. Construction of the integrative plasmid pYK. The restriction map of the KRSl locus is shown with the three sets of cloned DNA fragments iLys8, ILysl7 and ILysl (1Lys9). The inserted arrows show the direction of transcription of the cloned genes. The nucleotide numbering refers to the position of the initiating ATG of the KRSI gene. The HpaI - Sac1 fragment of 1Lys8 (- 1501 to 2094) was cloned into pUC18 digested with PstI and SmaI to g,ive pYK03. The EcoRI -BglII fragment (853 bp) from pEMBLYr25 carrying the selectable gene TRPI was introduced into the unique BstEII site (position -561) of pYK03. A XhoI linker (5’-CCTCGAGG-3’) was inserted into the BsmI site (position -272). The resulting plasmid, pYK04, contains the 5’ region of the KRSI locus. The SaA-EcoRI fragment from iLys9 (positions 1294-3659) was cloned into pUC18 digested with BarnHI and SalI to give pYK06. pYK07 results from the jmertion of the selectable gene LEU2 (Sun --ma1 fragment of 2225 bp from pUC9-LEU2) into the unique SnaBI site (position 1921) of pYK06. The complete KRSl locus was constructed by insertion of the SphI-Sun fragment of pYK04 into pYK07, digested with SphI and SalI, to give pYK. Relevant sites for EcoRI (E), HpaI (H), BstEII (Bt), BsmI (Bs), BnmHI (B), SalI (Sa), SnaBI (Sn), Sac1 (S), SphI (Sp), SmaI (Sm) and XhoI (X) are indicated. Conversion of protruding 5’ or 3’ termini to blunt-ended molecules by E. coli DNA polymerase I (Klenow fragment) or by bacteriophage T4 DNA polymerase was accomplished in the presence of the four deoxynucleotides (4dXTPs).

Plasmids pYK04 and pYK07 containing the 5’ and 3’ regions of the KRSl locus, respectively, were constructed as described in Fig. 1. The complete KRSI locus, with TRPl and LEU2 flanking the coding region of KRSl on the 5’ and 3‘ sides, respectively, was constructed by combining pYK04 and

pYK07 to give pYK (Fig. 1). The unique Xhol (position - 279) and BarnHI (position 204) sites allowed replacement of the 5‘ coding sequences of KRSl by a corresponding fragment bearing mutations or deletions in the amino-terminal exten- sion of LysRS. The pYK insert of 8250 bp, isolated after

4

MYS8 M13mp8

EcoRl EmRl

1 1 Ugation

E X B E I L/ Ml3YKNTer

KRSl

Mys8 M13mp19

Sal + Sac1 Sall t Sac1

I I Ligation

il- BamHl 12- Ba131 exonuclease digistion i3- DNA pol (Klenowj. 4dYTPs i4- EcoRl i5- Size selection i b b .,.....;: MlBmplB . -... 1.111111111....

EmRl and Smal digestion I / E X B

M13YK69NTer

ATG ‘PCT UVL CAA GAT AAT GTC AAA GCC GCC Gu: GAT CC net SBL G l n G l n Asp Asn V a l Lys A l a A l a Gly Asp 1 2 3 4 5 6 1 8 9 1 0 69

Fig. 2. Construction of M13YK69NTer, encoding a truncated amino- terminal fragment of LysRS. MI 3Y KNTer carries the EcoRI - EcoRI fragment from i.Lys8 (positions -391 to 386 in Fig. 1). It contains an XhoI linker introduced into the unique BsmI site (position -272), as described in Fig. 1 . After linearization with BumHI (position 204), M13YKNTer plasmids weredigested with Ed31 exonuclease for vari- ous time intervals. The pool of plasmids was treated with Klenow enzyme and digested with EcoRI. Fragments corresponding to the 5’ rcgion of KRSZ were isolated by high-performance size-exclusion chromatography (Schmitter et al., 1986) and cloned into M13mp18 digested with EcoRI and SmaI. The extent of Bum1 digestion was monitored by restriction mapping and DNA sequencing. The resulting amino-terminal sequence of LysRS from the selected plasmid M13YK69NTer is indicated. Relevant sites for EcoRI (E), BurnHI (B) and XhoI (X) are shown.

digestion with SphI and SmaI, was used to replace the wild- type chromosomal allele of KRSl by the one-step gene-re- placement procedure (Rothstein, 1983) (cf. Fig. 4). Allele re- placement was confirmed by Southern analysis (Southern, 1975).

Construction of the replicative plasmid pYe-YK

Plasmid pYe-YK harbors the BsmI-SnaBI fragment of KRSl (positions -272 to 1921) from ALys8 inserted into the BumHI site of pEMBLYe23, containing the 2-p replication origin and the selectable gene URA3. pYK02 contains the Bsml - SacI fragment of KRSl (positions - 272 to 2094) from iLys8 inserted into pUC18 (Martinez et al., 1991). The Eco- RI - RumKI fragment of M13YK58NTer (see below) was in- serted into pYK02 digested with PstI (in the polylinker of pUC18, upstream from the BsmI site of KRSI) and BamHI (position 204) to give pY K15. The BsmI - SnaBI fragment of pYK15 was cloned into pEMBLYe23 digested with BamHI to give pYe-YK58. The two multi-copy plasmids pYe-YK and pYe-YKS8 allowed overexpression, in S . ceuevisiae, of the native or amino-terminal truncated (Alall - Ala57) forms of LysRS, respectively.

Creation of amino-terminal deletions or mutations

The amino-terminal alterations of LysRS were con- structed in an M13mp8 derivative, M13YKNTer (Fig. 2).

TAG Rev KRSl I Uni

Sa c10 s C15

1 1- PCR between primers - Rev and C3 -Rev and C10 -RevandC15

2- Sall and Spel digestion

1. PCR between primers Uni and CO 2- Spel and Sac1 digestion

I c sa Se Se - IS 1 G

Fig. 3. Construction of deletions in the 3’-coding region of KRSI. The Sall- SacI fragment of ILys8 (positions 1294 - 2094 in Fig. 1 ; stop codon of KRSI at position 1774) was cloned into M13mp19. The 3‘ non-coding region of KRSI was amplified with Taq DNA polymerase between the M13 universal sequencing primer (Uni; 17 bases; Pharmacia) and oligonucleotide CO (5‘-AACTAGTCTAAAAGATA- TTAGAAC-3’) that introduced a SpeI site at the TAG stop codon. The SpeI -Sac1 fragment was isolated. The 3’ coding region of KRSl was amplified between the MI3 reverse sequencing primer (Rev; 17 bases; Pharmacia) and the oligonucleotides C3 (5’-AACTAGTT-

GCTTCAAAG-3’) or C15 (5‘-AACTAGTTAGTTGGGAACAA- TAAGA-3’) that introduced a SpeI site at the TAG stop codon and led to the removal of 3, 10 or 15 carboxy-terminal residues preceding Asn591 (cf. Fig. 8). The three corresponding Sun -SpeI fragments were isolated. Relevant sites for Sun (Sa), SacI (S) and SpeI (Sc) are indicated.

CTTTTTGACTTCCTCTC-3‘), C10 (5’-AACTAGTTAACATCAG-

M13YK69NTer, encoding a truncated amino-terminal frag- ment of LysRS, containing the first 10 residues of LysRS, a Gly residue and Asp69, was constructed as described in Fig. 2. The XhoI-BamHI fragment of pYK (Fig. 1) was replaced by the XhoI-BumHI fragment of M13YK69NTer to give pYK69. To construct M13YK58NTer or pYK58, M13YK- 69NTer or pYK69 were linearized by BamHI and religated in the presence of the two complementary oligonucleotides

5’ - GATCCATCCAAAAAAAAAACAGATTTGTTCG CTGACCTC-3’

5’ - GATCGAGGTCAGCGAACAAATCTGTTTTTTT TTTGGATG-3‘

that restore the amino acid sequence over Ser58 - Leu68 (cf. Fig. 5).

Site-directed mutagenesis was performed according to the method of Nakamaye and Eckstein (1986). The oligonucleo- tides used were as follows:

K4E : 5‘ - CGAACAAATCTGTTTCTTCTTCGGAAGC CGGTTCTGGTTGAGCGGCAGC-3’

K8E: 5’ - TGGTTGAGCGGCAGCTTCTTCGGCGGC TTCTTCAGCTTCG ACTTGTCT-3' Single-stranded M13YKNTer was mutagenized with K4E to give M13YKNTer: K4E (Lys55,59,60,61 --+ Glu) and further submitted to a second step of mutagenesis with K8E to give MI 3 YKNTer: K8E (Lys44,45,48,49,55,59,60,61 --t Glu). The XhoI - BamHI fragment of pYK (Fig. 1) was replaced by

5

KRSl locus (chromosome)

modified K R S l (plasmid)

c-)

.........

.. .... .. modified KRSI (chromosome)

I k b 1 ........ .......

Fig. 4. One-step replacement of the wild-type KRSI chromosomal allele by a plasmid-borne KRSl derivative. The construction of the pY K plasmid family is described under Materials and Methods. The KRSl gene (open box) is flanked by the two selectable genes TRPl and LEU2 (dotted boxes). The KRSl flanking open reading frames, indicated by vertical hatched boxes, are not disrupted by the insertion of the TRPl and LEU2 genes. The horizontal arrows indicate the direction of transcription of the genes. The SphI - SmaI fragments of pYK and derivatives were used to transform the haploid yeast strain CC499-7A (leu2, t rp l , ura3, arg'). Leu+ and Trp+ transformants arise from a double- recombination event, upstream of TRPl and LEU2, and result in the replacement of the KRSl chromosomal allele by the plasmid-borne copy. Stars indicate positions where mutations were introduced into the KRSI gene. The haploid strains obtained after transformation with the plasmids pYKn are designated CCR-YKn; n refers to the mutation introduced into the KRSI gene. Only relevant restriction sites are indicated; BstEII and SnaBI sites were lost upon insertion of the TRPl and LEU.? selectable genes, respectively.

the XhoI-BamHI fragments of M13YKNTer:K4E or M13- YKNTer:K8E to give pYK:K4E and pYK:K8E.

Creation of carboxy-terminal deletions

The SalI-Sac1 fragment of pYK15 was removed and the plasmid was religated in the presence of the SpeI - Sac1 fragment and one of the Sun - SpeI fragments generated by PCR amplification, as described in Fig. 3. The BsmI - SnaBI fragments from the resulting plasmids, pYKl5C3, pYKl5C10 and pYK15C15, were inserted into the multi-copy plasmid pEMBLYe23 linearized by BamHI to give pYe-YK58C3, pYe-YK58C10 and pYe-YK58C15.

Integrative plasmids harboring amino- and carboxy-ter- mind modifications were constructed. The SalI - SnaBI frag- ment of pYK06 (Fig. 1) was replaced by the corresponding Sun-SnaBI fragment of pYKl5C3 to give pYK06C3. The SalI - Smal fragment of pYK07C3, obtained after insertion of the LEU2 gene (see above) into the SnaBI site of pYK06C3, was used to replace the SalI - Smal fragments of the integra- tive plasmid pYK (Fig. l ) and derivatives to give pYKC3, pYK58C3, pYK69C3, pYK:K4EC3 and pYK:K8EC3.

RESULTS

General strategy for expression of KRSl derivatives as single- copy genes

In S. cerevisiae, KRSl is an essential, single-copy gene (Martinez et al., 1991). The aim of the present study was to probe in vivo the functional significance of the polycationic amino-terminal extension of yeast cytoplasmic LysRS. We created in-frame deletions and Lys + Glu mutations within that segment of the KRSl gene product. In order to accurately monitor the associated phenotypes, it was necessary to express KRSI derivatives in conditions as close as possible to that of the wild-type KRSI allele in order to produce the various KRSI gene products at the same level and in a null back- ground. We chose to introduce mutant alleles of KRSI into chromosomal DNA at the KRSI locus. The insertion/replace- ment of KRSI alleles was accomplished by the one-step dis- ruption procedure, illustrated in Fig. 4. Plasmid pYK (Fig. 1) and derivatives thereof were used to transform to Leu + and

Trp' the haploid yeast strain CC499-7A. Southern-blot analysis verified that the replacement occurred via a double- recombination event, upstream of TRPI and LEU2. It results in a complete replacement of the wild-type KRSI allele by the plasmid-borne allele. The strain CCR-YK, obtained after transformation of CC499-7A with a pYK fragment, harbored a wild-type KRSI gene flanked by the TRPI and LEU2 genes. CC499-7A and CCR-YK displayed the same LysRS activity (Table 3). The insertion of the two selectable genes had no influence on the level of expression of KRSI. Thus, this pro- cedure enabled us to test the effect on cell growth of modifi- cations introduced into the KRSl gene expressed as a single- copy gene.

Construction and analysis of amino-terminal mutations

Earlier experiments have shown that the removal of 57 amino acids from the amino-terminus of the purified enzyme by elastase treatment has no effect on LysRS activity (Cirakoglu and Waller, 1985; Mirande and Waller, 1988). Another proteolytic product with a slightly lower molecular mass was also obtained with papain, without loss of activity (Cirakoglu and Waller, 1985). Comparison of the primary structures of E. coli and yeast LysRS revealed an amino- terminal extension of about 60 - 70 residues in the yeast en- zyme. We have created two in-frame deletions encompassing amino acids 11 - 57, named YK58 and mimicking the elastase- modified enzyme, and amino acids 11 - 68, named YK69 and mimicking the papain-modified enzyme. The corresponding amino-terminal sequences of the truncated proteins are indi- cated in Fig. 5. In order to ensure identical post-translational amino-terminal processing to all constructs, the first 10 resi- dues were retained. In addition to. these deletion mutants, we changed the net charge of the amino-terminal extension of yeast LysRS by site-directed mutagenesis of lysine codons into glutamate codons. Mutant YK:K4E results from four Lys + Glu substitutions (Lys55,59,60,61-+ Glu) and Y K : K8E has four additional mutations (Lys44,45,48,49,55, 59,60,61 + Glu; Fig. 5).

The wild-type KRSl gene of CC499-7A was replaced by the mutant alleles encoding YK58, YK69, YK:K4E or YK : K8E enzymes yielding strains CCR-YK58, CCR-YK69, CCR-Y K : K4E and CCR-YK : K8E, respectively. Appropri-

6

1 10 20 30 40 50 4 6 0 70 tl SQQD NUKA RR E GURN LHLDEATGEIIU SKS E L U R I K Q RQU ERKKR R K R R R Q P K P R S . C U T D L F RDL DPS Q . I . HHHHHCHHHHHHHHHHHCHCHHCCHHHHHHHHHHHHHHHHHHHHHHHHHHHHHCCHHHHHHHHHHHHCCCCC

Y K

YK58

YK69

1 1 0 60 70 IISQQDNUKRAgdpS.CUTDLFRDLDPS0.

1 10 70 I lSQQDNUKRRgDPSQ

1 1 0 20 30 40 50 60 70 Y K : K 4 E IISQQDNUt;RRREGUflNLHLDERTGEIIUSKSELlil;RII;QRQUEAlil;ARJit;RARQP~PRS~~~TDLFRDLDPSQ . . .

1 10 20 30 40 50 60 70 flSQQDNUKRRREGURNLHLDERTGEflUSKSEL~R I ~ ; P R Q U E A ~ ~ R R ~ ~ R A R Q P ~ P R S ~ ~ ~ T D L F R D L D P S Q . . . Y K : K8E

Fig. 5. Amino-terminal sequence of native and modified forms of yeast LysRS. The primary structure of the amino-terminal extension of native yeast LysRS (YK) is shown with its corresponding predicted secondary structure (Garnier et al., 1978; H, cr-helix; C , coil). The point of cleavage by elastase, residues Ala57-Ser58, is indicated by an arrow. Lysine residues are underlined. The sequences of the two truncated polypeptides (YK58 and YK69) are shown with the amino acids added by the construction indicated by lower-case letters. The locations of Lys + Glu substitutions within the sequence of the mutant enzymes YK:K4E and YK: K8E are marked by dots. The numbering refers to the position in the predicted amino acid sequence for the native polypeptide, as deduced from the nucleotide sequence of the cloned KRSl gene (Mirande and Waller, 1988), taking into account the initiating Met.

Table 1. Activity of amino- and carboxy-terminal modified LysRS. LysRS activities were measured in crude extracts of yeast cells by the tRNA aminoacylation reaction. One unit corresponds to the formation of 1 nmol Lys-tRNA/min at 25°C. Relative activities and k,,, values are relative to YK, corrected for respective amounts of the KRSJ allele gene products detected by Western-blot analysis.

Chromosome- Mutation encoded enzyme

Specific activity in Relative K, (tRNA) Relative k,,, crude extract activity (VmaxiE)

Y K YK58 YK69 YK : K4E YK:KXE.

YKC3 YK58C3 YK69C3 YK : K4EC3 Y K : K8EC3

wild-type LysRS dAla11 +Ah57 d Aka1 1 ---* Leu68 Lys55,59,60,61 -+ Glu Lys44,45,48,49,55,59,60,61 + Glu dClu588,589,590 YK58 and AGlu588,589,590 YK69 and AGlu588,589,590 YK:K4E and AGlu588,589,590 YK:K8E and dGlu588,589,590

Uimg total protein 1.68 0.07 0.29 1.04 0.03

2.04 0.21 0.52 1.71 0.15

Y o 100

18 62

122

31 102

0.9"

0.3"

2.5"

2"

PM 2.5 3.4 3.1 2.1 2.8 3.4 2.8 2.7 2.6 4.0

1 0.01 a

0.2 0.6 0.002"

1.5 0.02" 0.4 0.7 0.015"

Corrected values

ate insertion of the modified alleles, and removal of the wild- type allele, was monitored by Southern mapping and also by sequencing genomic DNA from CCR-YK:K4E and CCR- YK : K8E after PCR amplification with suitable primers (data not shown). Each of these four KRSl derivatives supported growth of a wild-type KRSl-deletion strain. This indicates that the amino-terminal extension of LysRS is not absolutely required for viability of S. cerevisiae.

LysRS activity was measured in crude extracts from the various cell lines (Table 1). The specific activity of YK: K4E was only slightly altered compared to YK. YK69 or YK58 and YK:K8E had reduced specific activities (5 - 100-fold). The affinity displayed by the various derivatives of LysRS for polyanionic carriers was analysed upon chromatography on heparin-Ultrogel under conditions described previously (Cirakoglu and Waller, 1985). Whereas the native enzyme (YK) is strongly retained on this carrier, its elution requiring 300 mM KCI, this affinity was attenuated in the case of YK:K4E(230 mM),YK:K8E(170mM)andYK58orYK69 (1 10 mM). This indicates that the polyanion-binding charac-

ter of the native LysRS from yeast is mainly due to the lysine residues clustered in its amino terminal extension.

In order to determine whether the decrease in LysRS ac- tivity of some constructions could be due to a decreased tran- scriptional efficiency or to an instability of the transcript, we performed a Northern-blot analysis with mRNA isolated from exponentially growing cells (Fig. 6). A probe specific for actin mRNA was used as an internal standard. The KRSl- specific probe revealed approximately the same amount of transcript for all constructs, with the exception of the two strains exhibiting a large decrease in LysRS activity (CCR- YK58 and CCR-YK: K8E) that displayed a slight increase in KRSl mRNA level. We conclude that the reduced level of activity is not due to a transcriptional event.

To monitor the amount of the LysRS protein in crude extracts from the various cell lines, we performed a Western- blot analysis with antibodies directed to the papain-modified enzyme, that is to a fragment of the protein which is present in all constructs. As shown in Fig. 7, the amount of the enzymes YK58 and YK : K8E is increased about fivefold, whereas the

7

KRSl -

Actin -

Fig. 6. Analysis of the transcripts generated by modified KRSI alleles. Total RNA from yeast strains CC499-7A. CCR-YK, CCR-YK69, CCR-YK58, CCR-YK:K4E and CCR-YK:K8E was prepared and analysed by Northern blot. Hybridization with a 32P-labeled probe specific for KRSZ (EcoRI -EcoRI fragment from ILys8; positions 386- 1009 in Fig. 1) reveals that the steady-state level of the KRSZ transcripts is not reduced by deletions or mutations in the 5' coding region of KRSZ. In order to accurately quantify relative amounts of KRSZ transcripts, hybridization was also performed with a yeast actin probe (1100-bp HindIll-Xhol fragment from the ACTZ gene of S. crvevisiae; Gallwitz and Sures, 1980) as an internal standard.

LysRS +

Fig. 7. Analysis of LysRS expressed by modified KRSI alleles. Crude extracts (50 pg total protein) from strains CCR-YK, CCR-YKC3,

YK:K4E, CCR-YK :K4EC3, CCR-YK: K8E and CCR-YK : K8EC3 were analysed by Western blotting using antibodies directed to papain- modified LysRS from yeast. The antibody preparation reacts with additional proteins present in the extracts. The equal background levels observed in each lane is an internal standard for normalisation of the amount of total protein contained in the samples.

CCR-YK69, CCR-YK69C3, CCR-YK58, CCR-YK58C3, CCR-

expression levels of YK69 and YK: K4E are similar to that of the native enzyme (YK). The LysRS mutants YK:K8E and YK : K8EC3 displayed a lower electrophoretic mobility, com- pared to the wild-type enzyme YK. The reason for this behav-

CCCCCCHHHHHHHHHHHHH 550 560 570 580 590

VGLPPTGGWGCGIDRLRflFLTDSNTIREULLFPT~KPDULREEUKKEEEN Y K

H G L P P T R G L G l G l O R f l U f l L F T U l L F P R n R P U K Ec

VGIlPPUGGFGLGIDRLCflLFCDKKRIEEULPFGWDDUNRQ Y m I t

............. .....

............ 0 . . 550 560 570 5 8 0

VGLPPTGGWGCGIORLRflFLTDSNTIREULLFPTLKPDULREEUKKN Y K C 3

550 560 570 580 V G L P PTG G U G C G I DR L A fl F L TDS NT I REU L L F P T L K P D UEI Y K C l O

550 560 570 I V G L P P T G G U G C G I ORLRDFLTOSNT I REULLFPTN Y K C 1 5

Fig. 8. Carboxy-terminal sequences of native and modified forms of LysRS. The primary structure of the carboxy-terminus of native yeast LysRS (YK) is shown with, indicated bclow, the carboxy-terminal sequences of the homologous enzymes from E. coli (Ec, from the IysS gene; Levtque et al., 1990) and yeast mitochondria (Ymit; Gatti and Tzagoloff, 1991). Identical residues are indicated by black dots. The dotted box represents the so-called motif 3, characteristic of class- 2 aminoacyl-tRNA synthetases (Eriani et al., 1990). The predicted secondary structure (Gamier et al., 1978; H, a-helix; C, coil) for the carboxy-terminal extension of the yeast enzyme is shown above the amino acid sequence. The carboxy-terminal sequences of the trunc- ated proteins with 3 (YKC3), 10 (YKC10) or 15 (YKC15) deleted amino acids are indicated.

iour is not known, but may be due to the clustering of acidic residues that could impair the binding of SDS molecules or to post-translational modification of the mutant enzymes.

We conclude that the replacement of the four lysine resi- dues Lys44,45,48,49 by glutamate residues, or at least one of these, or the deletion of 57 amino acids in the amino-terminus severely impairs the specific activity of LysRS. Michaelian parameters in the aminoacylation reaction, corrected for the relative abundance of the LysRS derivatives as detected in the corresponding crude extracts by Wcstern-blot analysis, are summarized in Table 1. The decrease in enzyme activity is correlated to an alteration of the relative catalytic efficiency (V,,,,,/E) of the mutants; the apparent K, value for tRNA is not affected.

The results we obtained with YK58, and to a lesser extent with YK69, are at variance with those previously reported following in vitro proteolysis of the purified enzyme (Cirakoglu and Waller, 1985). In that study, it was shown that the truncated forms of LysRS, obtained after elastase or papain treatment, remain fully active. The alignment of the primary structures of LysRS from E. coli and yeast (Levique et al., 1990) reveals, in addition to the large amino-terminal extension, the presence of a short carboxy-terminal extension in the yeast enzyme (Fig. 8). It is composed of 11 amino acids, comprising eight charged residues. These observations prompted us to investigate the possibility that an active trunc- ated enzyme could have been generated in the proteolytic treatment by a double-cleavage event at the amino- and carboxy-termini.

Construction and analysis of carboxy-terminal deletions

To test the possibility that the short carboxy-terminal ex- tension of yeast LysRS may impair the catalytic activity of the amino-terminal truncated enzymes YK58 and YK69 de- scribed above, we submitted corresponding crude extracts to

8

L * c z ul

z II)

z

c 0

w Elastase/Extract (by m a s s )

Fig. 9. Activation of LysRS activity of YK58 and YK69 enzymes upon elastase treatment. Crude extracts (1 mg/ml) from strains CCR-YKSB (7 ) and CCR-Y K69 (m) were incubated at 25 "C for 60 min in 10 mM potassium phosphate, pH 7.5, 20 mM 2-mercaptoethanol and 10% glycerol, with increasing amounts of elastase. After addition of 2 mM diisopropyl fluorophosphate to inactivate elastase, initial rates of tRNA aminoacylation were measured after appropriate dilution. Re- sults are expressed as percentage activities relative to activity re- covered in the control sample incubated in the absence of protease.

an elastase treatment. A twofold increase in LysRS activity was observed after a 60-min incubation, at 25 "C, in the pres- ence of a twofold excess (by mass) of protease (Fig. 9). Since this increase in activity was not observed when an extract of CCR-YK was submitted to the same treatment (data not shown), we envisaged that a carboxy-terminal deletion could be able to compensate for the deleterious effects induced by an amino-terminal deletion.

A series of carboxy-terminal modifications were intro- duced in the LysRS mutant YK58. We have created carboxy- terminal deletions encompassing 3 (Glu588 - Glu590 +YK58C3), 10 (Leu581 -Glu590+YK58C10) or 15 resi- dues (Leu576 - Glu590 + YK58C15). The resulting carboxy- terminal sequences are indicated in Fig. 8. The corresponding yeast episomal plasmids, derived from the vector pEMBLYe23, harboring URA3 as a selectable marker and the KRSl allele for YK58, were used to transform to Ura+ the strain CCR-YK58.

LysRS activity was measured in crude extracts from the various transformants (Table 2). The specific activity of YK58C3 was increased fourfold, compared to YK58. The deletion of 15 carboxy-terminal residues resulted in total inac- ti\ ation of LysRS (specific activity similar to that determined in strain transformed with the control plasmid pEMBLYe23 ; Table 2). We verified by Western-blot analysis that the re- combinant proteins were actually produced in vivo (results not shown). Whereas the plasmid-encoded species YK58, YK58C3 and YK58C10 were produced in equal amounts, YK58C3 5 was also overproduced but to a lower extent.

The C3 mutation, corresponding to a deletion of the three

Table 2. LysRS activity in a yeast strain transformed with plasmid- encoded YK58 derivatives displaying deletions in the carboxy-terminal- coding region. Strain CCR-YK58 was transformed by the episomal plasmid pEMBLYe23 or by derivatives harboring various KRSl alleles. LysRS activity was measured in crude extracts of yeast cells by the tRNA aminoacylation reaction. 1 U corresponds to the forma- tion of l nmol Lys-tRNA/min at 25°C.

Episomal plasmid Plasmid-encoded Specific activity LysRS derivative in crude extract

p EMB LY e23 none pYe-YK wild-type LysRS pYe-YKS8 d A l a l l + Alas7 pYe-Y KS8C3 dAlal1 + Alas7 and

AGlu588 + Glu590 pYe-YK58C10 AAlalI+ Ala57 and

ALeuSX1 +GluS90 pYe-YK58ClS dAlal1 + Aka57 and

dLeu576 + Glu590

U/mg total protein

10.7 0.071

0.68 2.64

0.73

0.035

parameters for the mutant enzymes were determined (Table 1). The C3 deletion led to a 2 - 7-fold increase in LysRS activity of the amino-terminally modified enzymes (Table 1) and restored normal growth rates of the mutant yeast strains (see below). The results established that a small carboxy-terminal deletion partially compensates for a deletion in the amino-terminal extension of yeast LysRS.

Effect of amino- and carboxy-terminal deletions on growth rate

KRSl derivatives were tested for their ability to ensure growth of the transformed yeast cells. Since modified KRSI alleles were expressed at the proper place of the endogenous KRSI gene, we could accurately measure the effect of allele replacement on the growth phenotype. As shown in Table 3, there was no difference in the growth phenotype of the native (CC499-7A, CCR-YK) and modified, but active constructions

K4EC3). The constructions that lead to a fivefold (CCR- YK69) to 50-fold (CCR-YK58 and CCR-YK: K8E) decrease in LysRS activity, as measured in vitro, grow more slowly (Table 3). This is certainly the result of poor in vivo aminoacylation activity. The two strains CCR-YK58C3 and CCR-YK:K8EC3 display a growth rate close to that of the wild-type strain, demonstrating the compensatory effect of the C3 mutation.

Thus, the removal of the polycationic extension of LysRS or the mutation of four lysine codons into glutamate codons, leading to a much weaker affinity for polyanionic carriers, has no discernable effect on cell growth. The relative abundance and the in vivo aminoacylation level for some tRNA has been determined in several mutant strains. The results shown in Table 3 for tRNALys, tRNA""' and tRNAMe' suggest that the mutations do not significantly affect the in vivo level of aminoacylated tRNA.

(CCR-YKC3, CCR-YK69C3, CCR-YK:K4E, CCR-YK:

residues Glu588 - Glu590, was introduced into the series of integration plasmids pYK and derivatives, which were used to transform the haploid strain CC499-7A to Leu+ and Trp', to yield strains CCR-YKC3, CCR-YK58C3, CCR-YK69C3, CCR-YK : K4EC3 and CCR-YK: K8EC3. The modified en- zymes, which were the sole source of LysRS activity, support- ed growth of yeast cells. Specific activities and Michaelian

DISCUSSION

Cytoplasmic aminoacyl-tRNA synthetases from the yeast S. cerevisiae have lysine-rich polypeptide-chain extensions, as compared to their bacterial or yeast mitochondria1 counter- parts. The aim of the present study was to delineate the seg-

9

Table 3. Growth rate and in vivo tRNA aminoacylation level for yeast strains harboring derivatives of KRSI. LysKS activity in a crude extract of yeast strain CC499-7A was measured by the tRNA aminoacylation reaction. 1 U corresponds to the formation of 1 nmol Lys-tRNA/min at 25°C. Data from other yeast strains are taken from Table 1. Growth rates were determined in rich medium (YPC). Total tRNA, isolated as described in Materials and Methods, was aminoacylated with saturating amounts of purified aminoacyl-tRNA synthetases. The in vivo aminoacylation level of tRNA corresponds to the ratio of the amino acid acceptance determined for the periodate-treated (oxidized) RNA fraction to that obtained for the untreated fraction.

Strain Specific activity Growth Amount of tRNA in crude extract time

LYS Val

~ ~

In vivo aminoacylated tRNA

Met LY s Val Met

CC499-7A CCR-YK CCR-YKC3

CCR-YK58 CCR-Y K58C3

CCR-Y K69 CCR-YK69C3 CCR-YK:K4E CCR-YK: K4EC3

CCR-YK: K8E CCR-YK: K8EC3

U/mg total protein 1.94 1.68 2.04

0.07 0.21

0.29 0.52 1.04 1.71

0.03 0.15

min

130 130 130 255 155

3 39 130 130 130 210 150

PmoIlA2,o

100 f 10

125 f 10 90 f 10

-

-

-

105 f 5

100 * 10

120 * 20 95 f 10

-

-

100 + 15

65 f 10 90 + 10

95 f 15

1 1 0 f 1 5

70 f 5 120 f 10

-

-

-

-

45 * 5

45 * 5 45 & 5

45 * 10

60& 10 60 * 10 50 & 5

-

-

-

-

70 * 10

60 f 5 80 f 10

70 f 5

70 f 5

70f 10 80 f 10

-

-

-

-

80 f 10

95 f 5 8Of10

70 f 10

9 0 f 5 9 5 & 5 90 f 10

-

-

-

-

75 i 10

95 * 5 70+ 10

95 + 5

9 0 1 10

90+ 10 9 0 1 10

-

-

-

ment of yeast LysRS required for catalytic activity. In ad- dition, to address the physiological significance of the extra sequences, we designed a strategy allowing testing of various truncated or mutant proteins in a null background.

Previous in vitro studies (Cirakoglu and Waller, 1985) have suggested that yeast cytoplasmic LysRS is composed of a prokaryotic-like catalytic domain and of a non-catalytic do- main responsible for the polyanion-binding properties dis- played by the native eukaryotic enzyme. The latter domain was attributed to the amino-terminus of the protein (Mirande and Waller, 1988). The present deletion analysis gives further insight into the structural organization of this enzyme. The removal of 58 or 69 amino-terminal residues significantly af- fected the catalytic efficiency in the aminoacylation reaction. In addition, whereas the four lysines located at amino acids 55 - 61 could be exchanged for glutamate residues without impairing the catalytic activity, one or more of the four ad- ditional substitutions in the YK:K8E mutant affect the ac- tivity of the enzyme to a large extent. The residual activity is sufficient, however, to maintain cell viability.

The finding that a carboxy-terminal deletion partially compensates for deletions or mutations in the amino-terminal extremity of LysRS is noteworthy. This conclusion is drawn from two types of evidence. (a) The deletion of the three acidic residues Glu588,589,590 increases the catalytic activity of the mutant polypeptides, as assessed by in vitro activity measure- ments. However, the finding that the YK58C3 enzyme is only 2.5% as active as the native enzyme suggests that the in vitro elastase-modified, fully active protein (Cirakoglu and Waller, 1985) is cleaved between Leu581 and Glu588. In that study, no data concerning the sequence of the carboxy-terminal ex- tremity of the truncated enzyme was presented. (b) A carboxy- terminal deletion actually acting as a compensatory mutation is also demonstrated by the finding that the growth phenotype of the corresponding double-mutant strains (Table 3) is indis- tinguishable (YK69C3, YK: K4EC3) from that of the wild- type strain. Moreover, the C3 mutation largely compensates the growth defects of the mutants YK58 and YK:K8E, although the corresponding double mutants were less than

1 % as active as the native enzyme. While such a low level of activity may be non-limiting for growth, the latter results also suggests that the in vivo level of activity produced by YK58C3 or YK: K8EC3 may be close to that o f the native YK enzyme and that the altered enzymes are unstable in vitro. In this context, it must be remembered that in crude extracts from htsl-1 or mesl strains grown at permissive temperatures, no His-tRNA (Natsoulis et al., 1986) or Met-tRNA (Chatton et al., 1987) synthetase activities could be detected, respectively, despite normal growth phenotypes. Conversely, the in vivo level of activity produced by the two mutants YK58 and YK:K8E may be too low to ensure cell viability without overproduction of the modified proteins, as assessed by the Western-blot experiment shown in Fig. 7. This effect is most certainly the result of an increase in the KRSl steady-state mRNA level (Fig. 6). This could be accomplished through transcriptional induction of KRSI, a low level of LysRS ac- tivity leading to a constitutive derepression of the GCN4 regu- latory pathway (R. Martinez and M. Mirande, unpublished results).

Deletion analyses in the amino-terminal extensions of other aminoacyl-tRNA synthetases from S. cerevisiue have been reported. Whereas the complete extension from Asp- tRNA (Eriani et al., 1991) and Met-tRNA (Walter et al., 1989) synthetases could be removed without affecting activity, partial in-frame deletions in the large extension of Gln-tRNA synthetase (Ludmerer and Schimmel, 1987) generated en- zymes with apparently reduced specific activities. LysRS is a class-2 aminoacyl-tRNA synthetase (Eriani et al., 1990) and is evolutionary related to Asp-tRNA synthetase (Gampel and Tzagoloff, 1989). Class-2 enzymes are characterized by three conserved sequence motifs (Eriani et al., 1990), the active site, composed of antiparallel P-strands, being formed in part by motifs 2 and 3, assessed from the crystal structures of Asp- tRNA (Ruff et al., 1991) and Ser-tRNA (Cusack et al., 1990) synthetases. Motif 3 is located close to the carboxy-terminus. Site-directed mutagenesis of presumed essential amino acids from motif 3 (Anselme and Hartlein, 1991; Gatti and Tzagoloff, 1991) had shown the functional importance of this

10

V

L

Fig. 10. Helical-wheel representation of the amino- and carhoxy-ter- minal extensions of LysRS and design of the helical pair. The predicted r-hclicnl segmcnts Glu25 -Ah52 (N-ter) and Asp579 - Asn591 (C- ter) of yenst cytoplasmic LysRS are shown according to the graphic technique of Shiffer and Edmundson (1967). Salt bridges between oppositely charged residues that could contribute to the stabilization of the helical dimer, distributed over four helix turns, are indicated. The two helices run parallel to one another, amino acids in front are marked in larger scripts The cluster of lysine residues is indicated by d shaded arca

region for catalysis. Moreover, the deletion of as few as three carboxy-terminal residues from yeast Asp-tRNA synthetase resulted in a twofold decrease in aminoacylation activity (Eriani et al., 1991).

Our results establish the important role of the’carboxy- terminus of yeast cytoplasmic LysRS for its catalytic activity. The deletion of 15 residues results in total inactivation of the enzyme. However, we showed that the removal of three residues had no significant effect on the activity of the enzyme (mutant YKC3) but partially compensated for amino-ter- minal deletions or mutations. In order to propose a rational explanation to these findings, two observations are note- worthy. (a) Yeast cytoplasmic LysRS displays, in addition to a large amino-terminal extension, a small carboxy-terminal segment which is absent from the homologous enzymes from yeast mitochondria or E. coli (Fig. 8). (b) The crystal structure of the yeast Asp-tRNA synthetase/tRNAAsp complex has been solved in part (Ruff et al., 1991). Although the amino-terminal region, comprising the polycationic extension, is not yet re- fined, i t can be observed that the carboxy-terminus folds back to the amino-terminal domain. One can speculate that in the native yeast LysKS the positioning of the polycationic segment enables hydrophilic and ionic interactions between the carboxy- and amino-terminal extensions.

As previously observed, the lysine-rich segments of Lys-, Asp-. Thr- and Val-tRNA synthetases are predicted as a- helices (Mirande and Waller, 1988; Lorber et al., 1988; Chatton et al., 1988), leading to the segregation of lysine residues on one side of the helix. This anisotropic distribution could generate a high-affinity site for polyanionic carriers and designates the lysine-rich domain of these enzymes as a physiologically relevant biological interface. Taking into account the secondary-structure predictions for the amino- and carboxy-terminal extensions of yeast cytoplasmic LysRS (Figs 5 and 8), we tentatively proposed that the two cc-helices made of amino acid residues Glu25 - Ah52 and Asp579 -

Asn591 interact via ionic residues located on one side of these helices (Fig. 10). As discussed by Cohen and Parry (1990), interchain salt links between adjacent helices could contribute to the stabilization of an a-helix. In this context, the removal of the amino-terminal extension of LysRS could alter the conformation of the carboxy-terminal domain. According to the helix packing schematized in Fig. 10, the polyanion-bind- ing site of LysRS might be defined by lysine residues located on the opposite side of the amino-terminal helix (marked by a shaded area in Fig. 10).

Assuming that these interactions do exist, one could expect that the exchange of the two lysine residues 45 and 49 for glutamate residues (YK : K8E) would destabilize this struc- ture, the removal of residue Glu590 (YK:K8EC3) leading to a partial compensatory effect. In addition, the removal of the amino-terminal domain (YK58 and YK69), generating a protruding carboxy-terminus composed of several charged residues, could alter the functional conformation of the active site, made in part by motif 3 which lie adjacent to the carboxy- terminal helix (Fig. 8). The finding that the larger amino- terminal deletion (YK69 and YK69C3) is less deleterious than the smaller deletion (YK58 and YK58C3) also suggests that the presence of the protruding polypeptide segment Ser58 - Leu68, comprising five charged residues, destabilize the con- formation of the catalytic domain. In order to test for the influence of the remaining ten residues at the amino-terminal extremity of YK58 (Fig. 5), a derivative of YK58 displaying only four additional residues was constructed (data not shown). N o increase in specific activity was observed.

Constructions represent the sole source of LysRS activity, therefore it can be inferred that the mutant enzymes YK69C3 or YK: K4E and YK: K4EC3 are able to sustain normal cell growth. It appears, therefore, that the amino-terminal polycationic extension is, per se, dispensable in vivo. This is consistent with previous observations that alleles encoding truncated forms of Gln-tRNA or Met-tRNA synthetases, car- ried on multicopy plasmids, sustain cell growth (Ludmerer and Schimmel, 1987; Walter et al., 1989). In this work, we also show that a single copy of a modified allele is sufficient to provide the biosynthetic machinery with aminoacylated tRNALys. With regard to the hypothesis of a spatial organiza- tion of the translational machinery, through electrostatic in- teractions involving the polyanion-binding domains of eukaryotic aminoacyl-tRNA synthetases (Mirande, 1991, for a review), it was reasonable to speculate that overproduction of a truncated synthetase or of the corresponding tRNA could palliate the free diffusion of the enzyme. However, we ob- served that YK69C3 or YK : K4EC3 constructions, expressed as single-copy genes, had no discernable effect on cell growth or on the in vivo level of aminoacylated tRNA. It is also noteworthy that in vitro less efficient LysRS variants (YK58 or Y K : K8E) display growth limitations but normal levels of aminoacylated tRNA. Since the primary defect is a poor aminoacylation activity, it could be envisaged that an adaptative mechanism allows for compensation for the de- creased tRNALy” aminoacylation ability by lowering the growth rate.

The data presented in this study provide no compelling evidence in favor of a specific role for the polyanion-binding domains that characterize eukaryotic aminoacyl-tRNA synthetases and indicate that the amino-terminal domain of LysRS is individually dispensable for cell viability. These re- sults suggest that a triggering role in in situ compartmentaliza- tion of protein synthesis may not be the actual function of these polycationic chain extensions. However, although no

particular phenotype was observed when mutant strains were grown in standard medium, a more detailed study (e.g. behavior on different carbon sources, relative length of the various phases of the cell division cycle, morphology of the cells throughout the division cycle, temperature-sensitive growth, etc.) might be required to make apparent a phenotype otherwise too discrete. The strains obtained during this work, namely CCR-YK58C3 and CCR-YK:K4EC3, could be a valuable tool for such studies. In addition, one can argue that deletion of the extensions from several aminoacyl-tRNA synthetases could be required to enlarge phenotypic effects. Further studies should provide more insights into the function of these domains. Since the polyanion-binding domains are a common denominator for all eukaryotic aminoacyl-tRNA synthetases, it would be surprising if their occurrence had no functional meaning.

The authors acknowledgc the constant support of Jean-Pierre Waller in whose laboratory this work was carried out. We thank Yolande Surdin-Kerjan and Dominique Thomas for the gift of yeast strains and plasmids and for valuable suggestions, Picrre Kerjan for providing antibodies to papain-modified LysRS and Marie-Therkse Latreille for skillful technical assistance. This work was supported by grants from the Centre Nutiand de la Recherche Scientifique (UPR 2401) and from the Association pour la Recherche sur le Cancer. R. M. gratefully acknowledges a scholarship from the Dina Surdin Foundation.

REFERENCES Anselme, J . & Hartlein, M. (1991) FEBS Lett. 280, 163-166. Baldari, C. & Cesareni, G. (1985) Gene (Amst.) 35, 27-32. Chatton, B., Winsor, B., Boulanger, Y. & Fasiolo, F. (1987) J . B id .

Chatton, B., Walter, P., Ebel, J . P., Lacroute, F. & Fasiolo, F. (1988)

Cirakoglu, B. & Waller, J . P. (1985) Eur. .I. Biochem. 149, 353-361. Cohen, C. & Parry, D. A. D. (1990) Proteins 7, 1 -15. Cusack, S., Berthet-Colominas, C., Hartlein, M., Nassar, N. &

Chem. 262, 15094-15097.

J . Biol. Chem. 263, 52 - 57.

Leberman. R. (1990) Nature 347. 249 - 255. Delforge, J., Messenguy, F. & Wiame, J. M. (1975) Eur. J . Biochem

57.231 -239. Eriani, G., Delarue, M., Poch, O., Gangloff, J . & Moras, D. (1990)

Eriani, G., Prevost, G., Kern, D., Vincendon, P., Dirheimer, G. &

Gallwitz, D. & Surcs, 1. (1980) Proc. Nut1 Acad. Sci USA 77, 2546-

Gampel, A. & Tzagoloff, A. (1989) Proc. ,Vat1 Acad. Sci USA 86,

Nature 347, 203 -206.

Gangloff, J. (1991) Eur. J . Biochem. 200, 337-343.

2550.

6023 - 6027.

Garnier, J., Osguthorpe, D. J. & Robson, B. (1 978) J . Mol. Biol. 120,

Gatti, D. L. & Tzagoloff, A. (1991) J . Mol. Bid. 218, 557-568. Gornall, A. G., Bardawill, C. J. & David, M. M. (1949) J . Bid. Chern.

Harlow, E. & Lane, D. (1988) Antibodies: a laboratory manual, Cold

Hodgson, C. P. & Fisk, R. Z. (1987) Nucleic Acids Res. 15, 6295. Ito, N. Fukuda, Y., Murata, K. & Kimura, A. (1983) J . Bacteriol.

Kellermann, O., Tonetti, II., Brevet, A,, Mirande, M., Pailliez, J . P. &

Laemmli, U. K . (1970) Nature 227, 680-685. Lazard, M., Mirande, M. & Waller, J . P. (1985) Biochemistry 24,

Lkveque, F., Plateau. P., Dessen, P. & Blanquet, S. (1990) Nucleic

Lorber, B., Mejdoub, H., Reinbolt, J. , Boulanger, Y. & Giege, R.

Ludmcrer, S. W. & Schimmel, P. (1987) J . Bid. Chem. 262, 10807-

Martinez, R., Latreille, M. T. & Mirande, M . (1991) Mol. Gen. Genet.

Mirande, M., LeCorre. D., Riva, M. & Waller, J . P. (1986) Biochimie

Mirande, M. & Waller, J . P. (1988) J . B i d . Chem. 263, 18443 - 18451. Mirande, M. (1991) Prog. Nucleic Acid Res. Mol. Biol. 40, 95- 142. Nakamaye, K. L. & Eckstein, F. (1 986) Nurleic Acids Res. 18,9679 -

Natsoulis, G., Hilger, F. & Fink, G. R. (1986) Cell 46, 235-243. Rothstein, R. J . (1983) Methods Enzymol. 101, 202-211. Ruff, M., Krishnaswamy, S., Boeglin, M., Poterszman, A., Mitschler,

A., Podjarny, A., Rees, B., Thierry, J . C. & Moras, D. (1991) Science 252, 1682 - 1689.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: u laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl Acad. Sci

Schimmel, P. (1987) Annu. Rev. Biochem. 56,125-158. Schmitter, J. M., Mechulam, Y., Fayat, G. & Anselme, M. (1986) J .

Chromatogr. 378,462 - 466. Sherman, F., Hicks, J . B. & Fink, G. R. (1986) Methods in yeast

genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

97-120.

177, 751 -766.

Spring Harbor Laboratory, Cold Spring Harbor, N Y .

153, 163-168.

Waller, J. P. (1982) J . Biol. Chem. 257, 1 1 041 - I 1 048.

50YY -5106.

Acids Res. 18, 305-312.

(1988) Eur. J . Biochem. 174, 155-163.

10813.

227. 149-154.

(Paris) 68, 1001 -1007.

9698.

USA 74, 5463 - 5461.

Shiffer, M. & Edmundson, A. B. (1967) Biophys. J . 7, 121 -135. Southern, E. (1975) J . Mol. Bid. 98, 503-517. Walter, P., Weygand-Durasevic, I., Sanni, A,, Ebel, J . P. & Fasiolo,

Yanisch-Perron, C., Vieira, J . & Messing, J. (1985) Gene (Amst . ) 33, F. (1989) J . Biol. Chem. 264,17126-17130.

103- 119.