dead, a New Escherichia coli Gene Encoding a Presumed

Transcription

1 JOURNAL OF BACTERIOLOGY, June 1991, p /91/ $02.00/0 Copyright 1991, American Society for Microbiology Vol. 173, No. 11 dead, a New Escherichia coli Gene Encoding a Presumed ATP-Dependent RNA Helicase, Can Suppress a Mutation in rpsb, the Gene Encoding Ribosomal Protein S2 W. MARK TOONE,"2 KENNETH E. RUDD,3 AND JAMES D. FRIESEN'2* Department of Genetics, The Hospitalfor Sick Children, 555 University Avenue, Toronto, Ontario M5G IX8,1* and Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ontario MSS 1A8,2 Canada, and Laboratory of Bacterial Toxins, Division of Bacterial Products, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, Maryland Received 11 January 1991/Accepted 23 March 1991 We have cloned and sequenced a new gene from Escherichia coli which encodes a 64-kDa protein. The inferred amino acid sequence of the protein shows remarkable similarity to eif4a, a murine translation initiation factor that has an ATP-dependent RNA helicase activity and is a founding member of the D-E-A-D family of proteins (characterized by a conserved Asp-Glu-Ala-Asp motif). Our new gene, called dead, was cloned as a gene dosage-dependent suppressor of temperature-sensitive mutations in rpsb, the gene encoding ribosomal protein S2. We suggest that the DeaD protein plays a hitherto unknown role in translation in E. coli. Cellular processes in which RNA plays an essential role are likely to involve proteins whose function is the unwinding of RNA-RNA (or RNA-DNA) duplexes. Until recently only a few RNA helicases have been characterized biochemically. These include the Escherichia coli transcription termination factor Rho (7), the simian virus 40 large T antigen (53), the eukaryotic translation initiation factor eif4a (45), and the nuclear protein P68 (17), which shares similarity with both eif4a and the simian virus 40 large T antigen. Recently a number of gene products have been identified which, on the basis of their homology with eif4a, are presumed to be ATP-dependent RNA helicases. These gene products, which make up the so-called D-E-A-D family of proteins (34), are thought to function primarily in events that occur posttranscriptionally. This growing family includes the following: the Drosophila vasa gene product (28) and the murine spermspecific protein PL10 (32), both of which are thought to be involved in temporal or tissue-specific mrna translation; the yeast proteins PRP5 (14) and PRP-16 (11) and yeast mitochondrial protein MSS116 (54), which participate in pre-mrna splicing; and the yeast translation initiation factors TIF1/2 (35) and SPB4 (49) and E. coli SrmB (41), which play roles in mrna translation and ribosome assembly, respectively. The E. coli ribosome is the prototypical ribonucleoprotein particle. As with other ribonucleoprotein particles, there is a growing recognition that at least some of the catalytic activity associated with the ribosome resides in ligandinduced conformational changes in its RNA component, with the ribosomal proteins serving as modulators of the RNA structure. Translation of mrna in E. coli requires not only ribosomes, consisting of 53 ribosomal proteins and 3 rrnas, but also trnas, aminoacyl-trna synthetases, and trna-modifying enzymes. In addition, translation requires proteins transiently associated with the ribosome, including initiation factors (IF-1, IF-2, and IF-3), elongation factors (EF-G, EF-Tu, and EF-Ts), and release factors (RF-1, RF-2, and RF-3 [23]). One would assume that the repertoire of * Corresponding author. proteins necessary for translation in vivo might include the proteins whose function is the unwinding of RNA duplexes. These might facilitate conformational changes in rrna, trna-rrna interactions, and the melting of mrna secondary structure. One D-E-A-D box protein in E. coli, SrmB, has been implicated in the assembly of ribosomes (41). There are many D-E-A-D box proteins in the yeast Saccharomyces cerevisae, and several are known to be associated directly with ribonucleoprotein particles, such as the ribosome and the spliceosome (13). Using a strategy designed to find temperature-sensitive mutations affecting amino acid biosynthesis, we have isolated two E. coli strains with temperature-sensitive lesions in rpsb, the gene encoding ribosomal protein S2. Our aim was to isolate extragenic suppressors of these mutations in anticipation that these might turn up new genes involved in the function or assembly of the translational apparatus. Here we report the isolation of a new gene, called dead, which is located at 68.8 min on the E. coli map, in a region that contains genes involved primarily in transcription and translation. This gene, when overexpressed from a high-copynumber plasmid, is able to suppress the phenotype in two rpsb mutants. The predicted dead gene product has remarkable similarity to eif4a and other members of the D-E-A-D family of ATP-dependent RNA helicases. The similarity to a eukaryotic translation factor and its ability to suppress a defect in S2 suggest that the dead gene product may be a hitherto unknown component in the function or assembly of the translational apparatus. MATERIALS AND METHODS Bacterial strains and media. The following bacterial strains were used in this work: JF368 (relc metf reca ilv rpsl [43]); JF6015 (JF368 rpsb2131); JF6020 (JF368 rpsb2132); CSR603, used for maxicell analysis (thr-l leub6 proa2 phr-j recal arge3 thi-l uvra6 ara-14 lac YI galk2 xyl-5 mtl-l rpsl31 tsx-33 supe44 [50]); JF2419, used to map the temperature-sensitive mutations relative to dapd2 (thi-j relal and dapd2 [10]); JF2420, used to map the temperature-sensitive mutations relative to dapd4 (thi-j relal and dapd4 [10]); 3291

2 3292 TOONE ET AL. RL425 (Hfr P4X rpsb [ts25] metb [6]); CAG12105, used to map the temperature-sensitive mutations relative to a TnJO insertion at 3.5 min (MG1655 zad-3094::tnlokan [57]); and CAG12204, used for Hfr mapping in the region between 6 and 90 min (KL227 btub3192::tnlokan [57]). LB and M9 minimal media were prepared as described previously (39). AT agar was prepared as described (47). When necessary, antibiotics were supplemented to the following concentrations: ampicillin, 100,ug/ml; tetracycline, 20,ug/ml; kanamycin, 100,ug/ml. dapd expression in the dapd host strains was determined by growth on LB agar lacking diaminopimelic acid as described previously (10). Mutagenesis of JF368 was performed by using 50 jig of nitrosoguanidine per ml as described previously (39). Details of plasmids used in this study are given in the text. Genetic mapping. Recombination-proficient derivatives of E. coli strains JF6015 and JF6020 were constructed by transformation with plasmid pdr1453, which carries a wildtype copy of the reca gene (51). Genetic mapping of the temperature-sensitive mutation in each strain was then carried out by conjugation with Hfr strains and bacteriophage P1 transduction, by using the strategy and strains described previously (57). Mapping plasmid DNA inserts to the chromosome. Plasmid DNA inserts were mapped to the chromosome by hybridizing a 32P-labeled restriction fragment from the inserts to a DNA blot of an orthogonal field agarose gel electrophoresis (OFAGE) gel, kindly provided by Cassandra Smith (58). The OFAGE DNA blot contained NotI-digested DNA from three strains of E. coli (58). The E. coli DNA NotI restriction fragments have been correlated with the E. coli genetic map (58). This allowed localization of the inserts to a relatively small region of the chromosome. The results showed that the inserts reside on the 1,000-kb NotI A fragment that corresponds to a region between 60 and 80 min on the E. coli map (58). Further mapping was done by scanning the highresolution restriction map of the E. coli chromosome, in the region indicated (26), comparing the E. coli restriction map with that of the plasmid inserts. A region at coordinate 3150 (26) showed a good match to the restriction map of the inserts in pmt1 and pmt2. The chromosomal localization of the insert was confirmed by hybridization of a labeled restriction fragment from the insert to a DNA blot from the phage carrying the suspect region of the chromosome. Hybridizations to DNA blots were carried out as described previously (38) in the presence of 50% formamide, 5x SSC (lx SSC is 0.15 M NaCl plus M sodium citrate), and 8.25% dextran sulfate at 42 C; the insert-derived restriction fragment was radiolabeled by the random-primer technique (15). The blots were washed twice for 30 min in 0.1X SSC-0.1% sodium dodecyl sulfate (SDS) at 65 C. Maxiceli analysis. Maxicell analyses were performed as described previously (50). E. coli CSR603 (50) was transformed with various plasmids and UV irradiated, and the plasmid-encoded proteins were labeled with [35S]methionine. Samples were analyzed on a sodium dodecyl sulfate- 15% polyacrylamide gel. DNA sequence analysis. Restriction fragments from plasmid pmt1 were cloned into Bluescript vectors psk and pks. DNA sequences were determined by the method of Sanger et al. (52). Procedures, including construction of overlapping unidirectional deletions, preparation of singlestranded DNA templates, and dideoxy sequencing, were performed by using reagents and methods provided by the manufacturer (United States Biochemicals). Comparisons with the GenBank data base were performed by using the J. BACTERIOL. FASTP program (36) provided in the International Biotechnologies sequence analysis package. RNA extraction and primer extension. RNA was extracted from various strains of E. coli by using the hot-phenol method (4). The oligonucleotide 5'-CAGATCTGCAAAAG TGGTTTCGA-3' was 5' end labeled with polynucleotide kinase as described previously (38). Primer extension analyses were performed on 10 jig of RNA (as measured by optical density and formaldehyde gel analysis) by using the labeled oligonucleotide. Primer extension was performed as described previously (18). Recombinant DNA methods. Plasmid DNA preparation, transformation of E. coli, and agarose gel electrophoretic analyses were performed by conventional procedures (38). Restriction endonuclease digestions and ligations were carried out as recommended by the manufacturers (Bethesda Research Laboratories and New England BioLabs). Nucleotide sequence accession number. The nucleotide sequence presented here has been submitted to the GenBank data base under accession no. M RESULTS Strategy and isolation of temperature-sensitive mutants. Histidine biosynthesis in E. coli is regulated at several levels, including an attenuation mechanism that is sensitive to the levels of trnahis (25) and a metabolic control mechanism in which ppgpp is a positive effector of transcription (61). Because of the complexity of these mechanisms, one would expect mutational alterations of normal his operon expression to affect a wide variety of cellular processes. Previous studies have already proven this to be the case. Mutations altering histidine biosynthesis have been found in the gyra and gyrb genes, which encode the subunits of DNA gyrase (48); hist, which encodes a trna-modifying enzyme (33) that affects many species of trna; and spot, which encodes an enzyme that degrades ppgpp (47). All of these mutant genes have pleiotropic effects that extend beyond their His phenotype, yet their effect on his operon expression has helped to elucidate the larger role their gene products play in the cell. We have taken advantage of the histidine biosynthetic system in looking for temperature-sensitive mutations that allow growth on the toxic histidine analog 3-amino-1,2,4- triazole (AT). AT inhibits the imidazoleglycerol-phosphate dehydratase activity, one of the two activities associated with the product of the hisb gene, causing a reduction in synthesis of histidine (64). We have isolated a number of temperature-sensitive mutants, which at the permissive temperature (30 C) were able to grow on AT plates containing 20 mm AT. The his operon is regulated positively by ppgpp, as are a number of other amino acid biosynthetic operons (61). A strain of E. coli, JF368, was used; this contains a relc mutation, coding for an altered ribosomal protein, Lii (43), which effectively blocks the rela-dependent pathway for the synthesis of ppgpp and lowers the basal level of ppgpp (unpublished observation). The lower basal levels of ppgpp in relaxed (Rel-) mutants renders them more sensitive to inhibition by AT and facilitates selection of AT-resistant mutants (47). The spot gene product is involved in degradation of ppgpp (2); consequently, mutations in the spot gene raise the basal level of ppgpp and thereby increase his operon expression. To minimize the probability of isolating spot mutations, we performed all mutagenesis with strain JF368 that had been transformed with pga1, a multicopy

3 VOL. 173, 1991 E. COLI RNA HELICASE 3293 minutes on E. coil man map I I I zed-3094:tn loken dopd ( rpsb) JF6015 ts ATr pdb9 pmt4 pmt5 lp~rpso ~Arps B 'IT pga209 I D I tsf Suppression,*** 1% of TS phenotype FIG. 1. Mapping and complementation of the temperature-sensitive (TS) mutation in E. coli JF6015 and JF6020. Bacteriophage P1 cotransduction frequencies are given as percent cotransduction. plasmid carrying the spot gene (2). In this strain, SpoT activity does not increase in a gene-dosage-dependent manner (2). Strain JF368(pGA1) was mutagenized with 50 pug of nitrosoguanidine per ml (39). AT-resistant mutants were selected on M9 minimal medium supplemented with 19 amino acids (minus histidine) and 20 mm AT. The parental strain, JF368(pGA1), grows poorly on 5 mm AT and not at all on 20 mm AT. Colonies that grew on 20 mm AT agar at 30 C were tested for temperature-sensitive growth on AT agar and LB agar at 42 C. AT-resistant colonies arose at a frequency of 10-6, and approximately 1% of these showed a temperaturesensitive phenotype. Two independently isolated temperature-sensitive mutants, JF6015(pGA1) and JF6020(pGA1), were selected for further analysis. Coreversion studies for the AT resistance and temperature sensitivity showed that the two phenotypes (in both JF6015 and JF6020) coreverted at a high frequency (30%), indicating that the two phenotypes were probably due to a single mutation. Similar coreversion frequencies of two phenotypes that are due to a single mutation have been seen in other systems (8) and are much higher than one would expect for mutations in two different genes. The locus causing the temperature-sensitive and AT resistance phenotypes was mapped genetically. At the same time, a suppressor analysis was undertaken by transforming JF6015 and JF6020 with a pbr322-based wild-type E. coli gene library. For the purposes of further manipulation, both JF6015(pGA1) and JF6020(pGA1) were cured of pga1 by the method of Maloy and Nunn (37), selecting for loss of tetracycline resistance. Curing the cells of pga1 had no effect on either the temperature sensitivity or the AT resistance (data not shown). The temperature-sensitive mutations map to rpsb, the gene encoding ribosomal protein S2. Since both mutant strains i w I I I co~ Q- were Rec-, all genetic mapping was done in the presence of pdr1453, a pbr322-based plasmid carrying a wild-type copy of reca (51). Transformation of the mutants with this plasmid caused them to become UV resistant, as was expected, but did not affect the AT resistance or temperature-sensitive phenotypes. The approximate map location of the temperature-sensitive mutations was determined by using a set of Hfr strains with kanamycin resistance markers located at known distances from the respective origin of transfer and then screening for linkage of the temperaturesensitive phenotype with the Kanr marker (57). The temperature-sensitive mutations in both JF6015 and JF6020 were mapped to between 90 and 6 min on the E. coli map. A set of isogenic strains of E. coli, each with a single Kanr marker inserted into the chromosome, was used to localize further the temperature-sensitive mutations (57). Bacteriophage P1 transduction showed linkage to a Kanr marker (zad-3094:: TnlOkan) located at 3.5 min. Furthermore, cotransduction of the temperature-sensitive mutation with dapd, an auxotrophic marker for diaminopimelic acid dependence, indicated that in both strains the temperature-sensitive lesion resides very near the rpsb operon (67% cotransduction). The two genes, rpsb and tsf, in the rpsb operon code for ribosomal protein S2 and EF-Ts, respectively. Wild-type clones of these two genes were analyzed for their ability to complement the temperature-sensitive and AT resistance phenotypes. Plasmid pdb9 (1, 5), carrying rpsb and tsf, and plasmid pmt4, carrying only rpsb, were both able to complement the temperature-sensitive and AT resistance phenotypes of strains JF6015 and JF6020 (Fig. 1). Taken together, these data indicate that the mutation responsible for both phenotypes resides in the gene coding for ribosomal protein S2 İsolation of gene-dosage-dependent suppressors. Both JF6015 and JF6020 were transformed with a wild-type, +

4 3294 TOONE ET AL. J. BACTERIOL. mtr dead ORF pnp rpso p35 p I 5B InfB nusa pl5a mety WEI mm ME I it I I.m I I A-I I I 1 tp P t P2 1t t t P Po P-1 1 i ] EcoI I Eco supgression of TS ghenotude pmtl + pmt2 + pmt3-15 EL-.O I OOObp FIG. 2. Organization of genes in the region of dead. All genes are transcribed counterclockwise, and the locations of known promoters (P) and terminators (t) are shown. The expansion of the dead region indicates the position of the major transcription start site for dead, as well as relevant restriction sites. The location of the 1-kb EcoRI-EcoRV restriction fragment used as a probe to hybridize to Kohara phage is also indicated. Plasmid clones including the region and their ability to suppress the rpsb temperature-sensitive (TS) phenotype are shown. Kohara phages 13A9 and E46113 were used to localize the inserts and are located at coordinates 3150 on the map of Kohara et al. (26). pbr322-based E. coli library (9) to isolate plasmid-borne genes that could suppress the temperature-sensitive phenotype. We expected to find plasmids that carried a wild-type version of the mutated gene (rpsb) and hoped as well to find plasmids that carried genes unlinked to the original temperature-sensitive mutation but that in increased copy could also suppress the temperature-sensitive phenotype. Unexpectedly, the initial pool of clones included only plasmids containing DNA inserts which, when mapped back to the chromosome, were shown to be unlinked to the temperature-sensitive mutation at 4 min. In total, five plasmids carrying DNA inserts from three separate loci, all unlinked to the temperature-sensitive lesion at 4 min, have been characterized. Two clones in particular, pmt1 and pmt2, which share common DNA restriction fragments, completely suppressed the temperature-sensitive phenotype and partially suppressed the AT resistance phenotype (at 30 C, even in the presence of pmt1 or pmt2, the strains are slightly more resistant to AT than is the parent) of JF6015 and JF6020. These clones were selected for further analysis. Physical mapping of suppressor plasmids. The DNA fragments inserted in plasmids pmt1 and pmt2 were localized to a region of the E. coli chromosome corresponding to 68.8 min on the E. coli map (3). This was done first by hybridizing a 1.4-kb EcoRI probe common to both plasmids to a NotI digest of E. coli DNA that had been electrophoresed on an OFAGE gel (58; data not shown). This localized the clones to a 1,000-kb NotI fragment corresponding to between 60 and 80 min on the E. coli map (58). Comparison of the restriction maps of the plasmid inserts with the Kohara et al. map of the E. coli chromosome (26) showed that the clones reside immediately downstream of the pnp gene (46) (Fig. 2). This location was confirmed by hybridization to DNA from Kohara bacteriophages known to map in the 68.8-min region. As illustrated in Fig. 2, a 1.0-kb EcoRI-EcoRV fragment showed hybridization to Kohara phage 13A9 but not to Kohara phage E4G11S (26). Sequence analysis (see below) indicated that part of the insert in pmt1 overlaps with the pnp gene, confirming the location of this insert. Maxicell and sequence analysis of suppressor plasmids. Maxicell analysis was carried out with cells carrying plasmids pmt1 and pmt2 in order to identify the plasmidencoded gene products. Both plasmids were shown to encode a protein of approximately 64 kda (Fig. 3). The primary structure of this protein was inferred by obtaining the nucleotide sequence of the entire insert of pmt1 (Fig. 4). As shown in Fig. 2, part of this insert overlaps with sequences published previously for the pnp gene (46) and then continues counterclockwise on the E. coli chromosome for a further 2,705 bp, subsequently overlapping with the mtr gene on the 3' end (22). An open reading frame (ORF) which can encode a protein with a molecular mass of 64 kda (Fig. 4) was found 1,113 bp downstream from the 3' end of the pnp gene. We have named this new gene dead (see below). The cluster of genes immediately upstream of dead consists of nusa, a postinitiation transcription factor; infb, encoding two forms of translation initiation factor IF2B; two ORFs, coding for 15- and 35-kDa proteins; rpso, coding for ribosomal protein S15; and pnp, coding for polynucleotide phosphorylase (20). The gene immediately downstream of dead is mtr, which encodes a tryptophan-specific permease (22). The intergenic distance between pnp and dead is 1,113 bp. Regnier et al. (46) observed the beginning of an ORF, preceding the dead ORF, 108 bp downstream of pnp and continuing past the region they sequenced. Our sequencing data indicate that the stop codon (TGA) for the ORF identified by Regnier et al. (46) lies at position -777 (in Fig. 4). This would create a protein with a predicted molecular mass of approximately 10 kda. No protein of this molecular mass was visible in our maxicell analysis (Fig. 3). However, it is possible that the ORF requires the pnp promoter for expression, and, since neither of the plasmids that we isolated contains a complete copy of the pnp gene, we do not

5 VOL. 173, :.:,. :.. :... : --,:.:.::.:: '... :. w",g,. ",a,.:.: : ::. : ::-_ i... ::... - t F X. i1.....: :: FIG. 3. Maxicell analysis of plasmids carrying the dead gene. The analysis was performed by using the procedure of Sancar et al. (50). Lysates containing plasmid-encoded proteins labeled with [35S]methionine were analyzed on an SDS-15% polyacrylamide gel (27). Lanes: 1, E. coli CSR603, untransformed; 2, CSR603 transformed with pbr322; 3 and 4, CSR603 transformed with pmt1 and pmt2, respectively. The only plasmid-encoded protein bands consistently seen in CSR603 transformed with pmt1 and pmt2 are the pbr322-encoded 29-kDa P-lactamase band and the 64-kDa DeaD band (arrow). necessarily expect to see expression. The sequence of the region downstream of dead containing the mtr gene has recently been published (22). The TGA stop codon of dead lies at position 93 of this sequence. A potential G+C-rich stem-loop structure followed by seven thymidine residues can be seen starting at position 363 in the leader region of mtr (267 nucleotides downstream of the TGA of dead). This sequence resembles a rho-independent terminator and most probably is the transcriptional terminator for dead. Primer extension analysis with an oligonucleotide that hybridizes 62 bp 3' of the putative ATG initiation codon of dead showed a major reverse transcriptase stop at nucleotide -173 (Fig. 5). A second reverse transcriptase stop was also seen further upstream, near position -650, and may represent a second promoter (data not shown). No primer extension products were seen extending into the pnp cistron, indicating that dead is probably regulated independently from the genes upstream. Sequence analysis in the region of the reverse transcriptase stop at position -173 revealed a good -10 consensus 5 bp from the putative transcription start and a -35 consensus (21) 17 bp upstream of the -10 (Fig. 4). By comparing the primer extension analysis shown in lane 3 of Fig. 5 with the data shown in either lane 1 or lane 2, it is evident that a strain transformed with pmt1 greatly E. COLI RNA HELICASE 3295 overproduces the dead transcript. It is reasonable to assume, although it has not been demonstrated, that this dosage-dependent transcription results in overproduction of the protein. There is no obvious Shine-Dalgarno (SD) (56) ribosomebinding site upstream of the first ATG of dead. However, there is a 6-nucleotide sequence, TGATCA, which could hybridize with 16S rrna immediately 5' to the anti-sd sequence. This site (called a TP site for translation initiationpromoting site) has been shown to be part of the translation initiator region in many E. coli genes (63). The anti-tp site on 16S rrna is located immediately 5' to the anti-sd and in the case of dead may be able to circumvent the lack of an SD. There are examples of other genes that lack an obvious SD sequence, such as dnag, rpsa, rpsk, and malb (63). All contain a TP site within 50 bp 5' to the ATG. In the case of dnag the sequence AUC (predicted to be part of a TP site) at position -10 may functionally replace the SD sequence. Perhaps the TP site also acts as a ribosome-binding site for dead. It should also be noted that there is another ATG 18 codons downstream from the proposed start codon of dead. There is a consensus SD sequence (GAGG) 12 nucleotides upstream of this ATG. At this point we cannot be sure which start codon (or both) is used in the translation of the dead transcript. dead can suppress other temperature-sensitive alleles of rpsb. We tested to see whether dead could suppress other rpsb(ts) alleles. The only other rpsb(ts) allele that has been characterized (both genetically and biochemically) resides in E. coli RL425 (6). Transforming RL425 with either pmt1 or pmt2 (Fig. 2) showed good suppression of the temperaturesensitive phenotype on LB agar plates at 42 C (data not shown). Vector plasmid alone (pbr322) or pmt3 (Fig. 2) failed to suppress the temperature-sensitive phenotype. As expected, the temperature-sensitive mutation in RL425 is also complemented by the same rpsb-carrying plasmids (shown in Fig. 1) that complement the temperature-sensitive mutations in JF6015 and JF6020. The observation that dead, on a high-copy-number plasmid, can suppress a known temperature-sensitive allele of rpsb supports our data mapping the temperature-sensitive mutations in JF6015 and JF6020 to the rpsb gene. DeaD shares sequence similarity with murine eif4a. The primary structure of the 64-kDa protein (Fig. 4) was compared with the GenBank data base (issue 63) by using the Putsell FastP/FastN programs. This search revealed extensive similarity to the inferred amino acid sequence of the murine translation initiation factor 4A (eif4a) and to the srmb gene product of E. coli (Fig. 6). Both of these proteins are members of a family of proteins known as the D-E-A-D family (named after an Asp-Glu-Ala-Asp motif that is conserved in all members of the family [34]). eif4a, an ATPdependent RNA helicase, is part of the mrna cap-binding structure (eif4f) and probably functions by denaturing RNA secondary structure, thereby eliminating it as an impediment to translation (45, 59). By association, all members of the D-E-A-D family are thought to be ATP-dependent RNA helicases (34). The region over which the D-E-A-D family shows similarity encompasses 388 of the entire 408 amino acids of the eif4a protein (40). Among the D-E-A-D box proteins listed in Fig. 6, DeaD shares the greatest similarity with eif4a (51%; 36% identity) and the least with MSS116 (41%, using the conservative substitutions listed in the legend to Fig. 6; 25% identity). Protein secondary-structure predictions for DeaD and other D-E-A-D box proteins (32) predict a pre-

6 3296 TOONE ET AL. J. BACTERIOL GTGATTCTGGCACGTATGGAACAAATCCTTGCCAGTCGGGCTTTAACCGATGACGAACGCGCACAGCTTTTATATGAGCG -802 GGAGTGTTGTATGATAGTCTCGGTCFUj3GGCATTAGCGCGTAACGATTTTTCGCAAGCGCTGGCAATCCGACCGGATAT -722 CCTGAAGTATTCAATTACTTAGCCATATATTAACGCAGGCAGGCAATTTTGATGCTGCCTATGAAGCGTTTGTTCTGTAC -642 TGAGCTTGATCCAACTTACAACTACGCGCACTTGATCGCGGGATCGCATTATATTACGGCGGTCGTGACAAGTTAGCGCA -562 AGATGATCTGCTGGCGTTTAACTTGACGATCCCAATGATCCTTTCCGTAGTCTGTGGCTTTATCTCGCCGAGCAGAAGCT -482 CGATGAGAAGCAGGCTAAAGAAGTGTTGAAACAGCACTTCGAAAAATCGGATAAGGAACAGTGGGGATGGAACATTGTCG -402 AGTTCTACCTGGGCAACATTAGCGAACAAACGTTAATGGAAAGGCTCAAGGCGGACGCAACGGATAACACCTCGCTCGCT -322 GAGCATCTCAGTGAAACCAACCTTCTATTTAGGTAAGTACTACCTAAGTCTGGGGGATTTGGACAGCGCCACGGCACTGT -242 TCAACTGGAGGTTGCCAACAACGTTCATAACTTTGTTr.GAr-CCGATACGCATTGTTGGALTAITzGCTCCTGGGCCAGG -162 ACCAAGATGACCTGGCAGAATCGGACCAGCAATAGCTGACGTACACATCAGCCCGTAATCTTTTTTGATTGCCATCACCT -82 AACGGGTGAGGGCGTTGTTGTTCGTTAATACACCTACTTTGAGCCGGTTCACACTTTTCAATGAAAATTGCTaIAT&-TT -2 TC ATG ATG AGT TAT GTA GAC TGG CCG CCA TTA ATT TTG AGG CAC ACG TAC TACC ATG GCT M M S Y V D W p p L I L R H T y y M A GAA TTC GAA ACC E F E T GAT CTG GGT TAC D L G Y GGC CGC GAC GTT G R D V CCT CTG TTG CAG P L L Q ACC CGC GAA CTG T R E L GGC GTA AAT GTG G V N V CGT CAG GGG CCG R Q G P GGC ACT CTG GAC G T L D CGC ATG GGC TTC R M G F ACC GCT CTG TTC T A L F AAA GAG CCG CAG K E P Q AGC TAC TGG ACT S Y W T GAA GAT TTT GAT E D F D GAA GCT CTT GAG E A L E CTG CGT GAA CAG L R E Q GAC GTT GCA GCC D V A A CCG ATG GAT TCT P M D S GGC CGC GCG CTG G R A L ACT ATG AAG CTG T M K L CGT CTG GAA AAA R L E K TAC CGC GCA CTG Y R A L ACT CTG GCT GCG T L A A GAT GGC CCG ATG D G P M CGT AAC GAC CGT R N D R GTT GGC GAT ATG V G D M CAT ATC GTT GGT H I V G AAG CTG TTT GCT K L F A AAC ACT TTA CGC N T L R ACT T GAA E CTG L TTT F AAA K GGT G GCA A CCA p ATG M AAT CTT GAT N L D GCG GTA CAG A V Q GTT GCT CTG V A L CAG ATC GTT Q I V CTC TCT AAA L S K L S ATC GAA GAC GTT GAA I E D V E TCT GCA ACC ATG CCG S A T M P GAA GTG CGC ATT CAG E V R I Q GTC TGG GGT ATG CGC V W G M R GCG GCG ATT ATC TTC A A I I F CGT AAC GGC TAC AAC R N G Y N ACA CTG GAA CGC CTG T L E R L CGT GGC CTG GAC GTT R G L D V GAG TCT TAC GTT CAC E S Y V H CTG TTC GTT GAG AAC L F V E N ACT ATT CCG GAA GTA T I P E V TTC GCC GCT AAA GTA F A A K V CTG AGC AAA ATT GAG L S K I Q GCA CTG CTG AAA ATG A L L K M CGT CCG AAA CGT GAA R P K R E GGC CCG CGT GGT GAC G P R G D CAG CTG TAC CGC ATT Q L Y R I GCG ATT GCT AAC GAA A I A N E TCT CAC TCC ACC ATC S H S T I GCA CTC GCA TTC TCA A L A F S GAT CTG GGC CTG AAG GCT CCT D L G L K A P TCT CCA ATT CAG GCA GAG TGT S P I Q A E C GCC CAG ACG GGG AGC GGA AAA A Q T G S G K CCT GAG CTG AAA GCA CCA CAG P E L K A P Q GTT GCT V A TAC GGC Y G GAA GCA ATG ACG GAT E A M T D GGC CAG CGT TAT GAC G Q R Y D GTC GGT ACT CCG GGC CGT CTG V G T P G R L CTG AGC GGT CTG GTT CTG GAT G L V L D ACC ATT ATG GCG CAG T I M A Q GAA GCG ATT CGT CGC E A I R R TCC AGC GTG ACT ACC S S V T T AAA AAC GAA GCA CTG K N E A L GTT CGT ACC AAA AAC V R T K N AGC GCC GCG CTG AAC S A A L N AAA GAT GGT CGT CTG K D G R L GAG CGT ATC AGC CTG E R I S L CGT ATC GGT CGT ACC R I G R T CGC GAG CGT CGT CTG R E R R L GAA CTG CCG AAC GCA E L P N A CAG CAG CAG CTG GAA Q Q Q L E CCG ACT GCT GAA GGT P T A E G GCA CAG GGT GAA CGT A Q G E R TTC CGT GAC CGT GAT F R D R D CGT GAA GAT CGT CCG R E D R P GAA GTG GGC CGC GAT E V G R D GGC GAC ATC AGC AGC G D I S S GAA CTG CCG AAA GTA E L P K V ACA AGC CGA TGA ACA T S R * ATC CTT I L ATT CCA I P ACT GCA T A ATT CTG I L TTC TCT F S GTG CAA V Q GAA GCC E A CAT CT( G H L GCA TTC A F GTG CT( G V L AAA CA( C K H TTA CGC L R CTT AAC L N CTG AAT L N TCT TTA S L GCA CCG A P ATG CGC M R GCC CTG A L CTG GAC CAC CT( G AAA CGT L D H LI K R GAA GCT E A ATC CCG I P ATT ACC I T CGT CCT R P GAC GAAk D E GAA GG' T E G CGC CG( C R R GAC ATCC D I ATG CTG M L CAT CAG H Q TTT ATG F M AGC CAG S Q GTA CGT TTC CTCG V R F LI GAA GCG E A GCG ACT A T CTG GA. A L E GTG GCT V A GGT GAC ATG AA( CAG GCG G D M N Q A GAC ATC CTG AT" T GCG ACC D I L I A T GTA GTT V V GGT CGT G R CTG CGC L R GAA CTG E L AAC TA( N Y GCG GG" T A G AAC AT" T N I CTA GGCC L G GAT ATC D I CGT GCT R A GAA CGT E R AAA CGC K R AGC AGC S S GAT CTG GAT CAA D L D Q GAA GAG CTG GA' T E E L DI CTC GAA L E ACT CTG T L ATC GT. A CCG CCA I V p p GAC CGT D R CGT CGT R R GAT GGT D G CGT TAC R Y GGT CCCG G P GAA CG" T E R GTT GAPA V E ATT GG' T I G CGC GAT R D CGT GAT R D GTT CGT V R AAC ATC N I. TGC C"GG GTG AAG TGC TGC C R V K C C,TGCAGTTACTGGGCGATGCACAGC 1745 GCATACTGGCGGTGAGCGTCGTGGCGGTGGTCGTGGTTTCGGTGGCGAACGTCGTGAAGGCGGTCGTAACTTCAGCGGT

7 : VOL. 173, 1991 E. COLI RNA HELICASE 3297 FIG. 4. Nucleotide and predicted amino acid sequence of the dead gene. The putative ATG is labeled + 1, and the sequence beginning at -882 starts where that described by Regnier et al. (46) ended. The stop codon for an ORF starting downstream of pnp is boxed. The -10 and -35 sequences for an E. coli promoter are underlined. The translation initiator region (with homology to a TP site [63]) 11 bp upstream of the first ATG is also underlined. ponderance of a-helical structure throughout the protein, D-E-A-D box proteins have domains that extend beyond the with a consensus helix-turn-helix motif, normally characteristic of DNA-binding proteins (42), found in the conserved SrmB, P68, and PRP5, has highly charged carboxy-terminal central region of similarity with eif4a. DeaD, as well as a-tdv-arglda-v region (a stands for aliphatic). This domains. These charged regions are similar to domains motif has been observed previously in a number of D-E-A-D found in other proteins that are implicated in RNA binding, box proteins (32), and with the addition of new proteins it such as the 70K Ul small nuclear ribonucleoprotein particle appears now to be conserved in all members of this family. protein, and may be important for either protein-protein or As noted by Leroy et al. (32), this observation indicates that protein-nucleic acid recognition (44). Another motif embedded in the charged carboxy-terminal domain of DeaD is an the helix-turn-helix motif may act as a general nucleic acid-binding structure that is important for RNA-binding arginine-rich consensus (RPRRERRD), which is found in a proteins as well as DNA-binding proteins. Most of the number of proteins thought to bind to regions of RNA A G C T *_ ww.... _S. t -._ A a _F G _. C _-.::.*-- ish:f G 5, *:: A -i -^. T +1 _F G C - G C N A T -..: C G -.i- :. : *-w- -* \ C G - - J.. - -i - ialik --i us;-:. -iii -. ins rf -229 TTGCCAACAACGTTCATAACTTTGTTGAGCACCGATACGCATTGTTGAATTATCGCTCCTGGGCCAGGACCAAGATGACC -149 TGGCAGAATCGGACCAGCAATAGCTGACGTACACATCAGCCCGTAATCTTTTTTGATTGCCATCACCTTAACGGGTGAGG -70 GCGTTGTTGTTCGTTAATACACCTACTTTGAGCCGGTTCACACTTTTCAATGAAAATTGTCTGATCAATTC ATG ATG M M 7 AGT TAT GTA GAC TGG CCG CCA TTA ATT TTG AGG CAC ACG TAC TAC ATG GCT GAA TZQ_GA S Y V D W P P L I L R H T Y Y M A E F E 66 ArC ACT TTT.CA (AT CT, GGC CTG AAG GCT CCT ATC CTT GAA GCC CTT AAC GAT CTG GGT T T F A D L G L K A P I L E A L N D L G FIG. 5. Primer extension analysis of the dead transcript. The major transcription start site was mapped by extension of a 20-nucleotide primer, complementary to the underlined sequence, with reverse transcriptase (18) and analyzed on a 5% polyacrylamide-urea sequencing gel. The dideoxy sequencing ladder was generated by using the same oligonucleotide. The sequence of the actual 5' end of the mrna that is complementary to the extended products is shown on the right. Lanes: 1, primer extension analysis with RNA isolated from the parent strain, JF368; 2, primer extension analysis with RNA isolated from one of the mutants, JF6015; 3, primer extension analysis with RNA isolated from strain JF6015 transformed with pmt1.

8 3298 TOONE ET AL. J. BACTERIOL. DEAD E. coli NH2-28aa... LGLKAPILEALNDLGYEKPSPIQAECIPHLL--NGRDVLGMAQTG EIF-4A mouse NH2-36aa... MNLSESLLRGIYAYGFEKPSAIQQRAILPCI--KGYDVIAQAQSG PL 10 mouse NH2-183aa.. VEMGEIIMGNIELTRYTRPTPVQKHAIPIIK--EKRDLMACAQTG VASA fruit fly NH2-249aa.. ADLRDIIIDNVNKSGFKIPTPIQKCSIPVIS--SGRDLMACAQTG P68 human NH2-78aa... ANFPANVMDVIARQNFTEPTAIQAQGWPVAL--SGLDMVGVAQTG TIF1/2 yeast NH2-26aa... MELDENLLRGVFGYGFEEPSAIQQRAIMPII--EGHDVLAQAQSG MSS116 yeast NH2- lloaa.. GVLDKEIHKAITRMEFPGLTPVQQKTIKPILSSEDHDVIARAKTG SRMB E.coli NH2-8aa... LELDESLLEALQDKGFTRPTAIQAAAIPPAL--DGRDVLGSAPTG a a a F PT aq a G DaG AQTG S A DEAD SGKTAAFSLPLLQNL-----DPELKA PQILVLAPTRELAVQVA---EAMTDFSKHMRGV EIF-4A TGKTATFAISILQQI-----ELDLKA TQALVLAPTRELAQQIQ---KVVMALGDYMGAS PL10 SGKTAAFLLPILSQIYTDGPGEALRAHKENGKYGRRKQYPISLVLAPTRELAVQIY---EEARFKSYRSRVR VASA SGKTAAFLLPILSKLLEDPHELELGR PQVVIVSPTRELAIQIF---NEARKFAFESYLK P68 SGKTLSYLLPAIVHINHQPFLERGDG PICLVLAPTRELAQQVQ---QVAAEYCRACRLK TIF1/2 TGKTGTFSIAALQRI-----DTSVKA PQALMILAPTRELALQIQ---KVVMALAFHMDIK MSS116 TGKTFAFLIPIFQHL-INTKFDSQYM VKAVIVAPTRDLALQIEAEVKKIHDMNYGLKKY SRMB TGKTAAYLLPALQHL-LDFPRKKSGP PRILILTPTRELAMQVS---DHARELAKHTHLD TGKTA F ap a P railatrplaaqa S G DEAD NVVALYGGQRY---DVQLRAL-RQGPQIVVGTPGRLLDHLK-RGTLDLSKLSGLVLDEADEMLRMGFIEDVE EIF-4A CHAC-IGGTNV---RAEVQKLQMEAPHIIVGTPGRVFDMLN-RRYLSPKYIKMFVLDEADEMLSRGFKDQIY PL10 PCVV-YGGADI---GQQIRDLER-GCHLLVATPGRLVDMME-RGHIGLDFCKYLVLDEADRMLDMGFEPQIR VASA IGIV-YGGTSF---RHQNECITR-GCHVVIATPGRLLDFVD-RTFITFEDTRFVVLDEADRMLDMGFSEDMR P68 STCI-YGGAPK---GPQIRDLER-GVEICIATPGRLIDFLE-CGKTNLRRTTYLVLDEADRMLDMGFEPQIR TIF1/2 VHAC-IGGTSF---VEDAEGL-R-DAQIVVGTPGRVFDNIQ-RRRFRTDKIKMFILDEADEMLSSGFKEQIY MSS116 SRMB DEAD EIF4A PL10 VASA P68 TIF1/2 MSS116 SRMB DEAD EIF4A PL10 VASA P68 TIF1/2 MSS116 SRMB ACVSLVGGTDFRAAMNKMNKL-R--PNIVIATPGRLIDVLEKYSNKFFRFVDYKVLDEADRLLEIGFRDDLE IATIF-GGVAY---MNHAEVFSE-NQDIVVATTGRLLVQYIKEENFDCRAVETLILDEADRMLDMGFAQDIE GG a R jtpgraad a VLDZAD ML agy a TIMAQI PEGHQTALISATMPEA-IRRITRRFMKEP----QEVRIQSSVTTRPDISQSYWTVWG-M DIFQKL NSNTQVVLLSATMPSD-VLEVTKKFMRDP-----IRILVKKEELTLEGIRQFYINVER-E RIVEQD---TMPPKGVRHTMMFSATFPKE-IQMLARDFLDEY----IFLAVGRVGSTSENITQKVVWVEE-A RIMTHV-----TMRPEHQTLMFSATFPEE-IQRMAGEFLKNY----VFVAIGIVGGACSDVKQTIYEVNK-Y KIVDQI RPDRQTLMWSATWPKE-VRQLAEDFLKDY----IHINIGALELSANHNILQIVDVCH-D QIFTLL PPTTQVVLLSATMPND-VLEVTTKFMRNP----VRILVKKDELTLEGIKQFYVNVEE-E TISGILNEKNSKSADNIKTLLFSATLDDKVQKLANNIMNKKECLFLDTVDKNEPEAHERIDQSVVISEKFAN HIAGET RWRKQTLLFSATLEGDAIQDFAERLLEDP----VEVSANPSTRERKKIHQWYYRADD-L I a QTaa SAT P a a Ia a a a a Q V RKNEALVRFL----EAEDFDAAIIFVRTKNATLEVAEAL-ERN--GYN-SALNGDMNQALREQTLERLKDGR EWKLDTLCDLY---ETLTITQAVIFLNTRRKVDWLTEKM-HAR--DFTVSAMHGDMDQKERDVIMREFRSGS DKRSFLLDLLN---ATGKDSLILVFVETKKGADSLEDFL-YHE--GYACTSIHGDRSQRDREEALHQFRSGK AKRSKLIEILS---EQADGT--IVFVETKRGADFLASFL-SEK--EFPTTSIHGDRLQSQREQALRDFKNGS VEKDEKLIRLMEEIMSEKENKTIVFVETKRRCDELTRKM-RRD--GWPAMGIHGDKSQQERDWVLNEFKHGK EYKYECLTDLY---DSISVTQAVIFCNTRRKVEELTTKL-RND--KFTVSAIYSDLPQQERDTIMKEFRSGS SIFAAVEHIKKQIKERDSNYKAIIFAPTVKFTSFLCSILKNEFKKDLPILEFHGKITQNKRTSLVKRFKKDE EHKTALLVHLL---KQPEATRSIVFVRKRERVHELANWL-REAG-GINNCYLRGEMVQGKRNEAIKRLTEGR b a L aaa Tbb a A a GD Q t atb G LDILIATDVAARGLDVERISLVVNYDIPMDSESYVHRIGRTGRAGRAGRALLFVENRERRLLRNIERTMKLT DEAD EIF-4A SRVLITTDLLARGIDVQQVSLVINYDLPTNRENYIHRIGRGGRFGRKGVAINMVTEEDKRTLRDIETFYNTS PL10 SPILVATAVAARGLDISNVKHVINFDLPSDIEEYVHRIGRTGRVGNLGLATSFFNERNINITKDLLDLLVEA VASA MKVLIATSVASRGLDIKNIKHVINYDMPSKIDDYVHRIGRTGCVGNNGRATSFFDPEKDRAIAADLVKILEG P68 APILIATDVASRGLDVEDVKFVINYDYPNSSEDYIHRIGRTARSTKTGTAYTFFTPNNIKQVSDLISVLREA TIF1/2 SRILISTDLLARGIDVQQVSLVINYDLPANKENYIHRIGRGGRFGRKGVAINFVTNEDVGAMRELEKFYSTQ MSS116 SGILVCTDVGARGMDFPNVHEVLQIGVPSELANYIHRIGRTARSGKEGSSVLFICKDELPFVRELEDAKNIV SRMB VNVLVATDVAARGIDIPDVSHVFNFDMPRSGDTYLHRIGRTARAGRKGTAISLVEAHDHLLLGKVGRYIEEP ata TDVGAGLDa V V= DaP s YaBRIGRTGR Gb G A F a DEAD IPEVELPNAELL AA 175aa- COOH A EIF4A IEEMPLNVADLI PL10 KQEVPSWLENMA 85aa- COOH VASA SGQTVPDFLRTC 38aa- COOH P68 NQAINPKLLQLV 139aa- COOH TIF1/2 IEELPSDIATLL laa- COOH MSS116 IAKQEKYEPSEE 153aa- COOH SlRMB IKARVIDELRPK 62aa- COOH FIG. 6. Alignment of DeaD and seven other members of the helicase family, eif4a, PL10, VASA, P68, TIF1/2, MSS116, and SrmB. Pairwise computer alignments (using the amino acid single-letter code) were altered slightly to give optimal matches for the eight sequences. Identical residues and conservative substitutions in six of the eight proteins are listed in boldface under each row, as are the branched-chain aliphatic amino acids, I, L, and V (a), and the basic (b) and acidic (a/c) amino acids. Conservative substitutions used to arrive at percent similarities are grouped as follows: ILVM, DE, RK, GA, TS, and YF.

9 VOL. 173, 1991 secondary structure, including the XN protein and the HIV- Tat protein (30). DISCUSSION Isolation of mutants with an altered ribosomal protein. We have cloned a gene from Escherichia coli which encodes a protein that shares significant similarity with a group of proteins called the D-E-A-D family. This is a growing family of proteins whose founding member is the murine protein translation initiation factor eif4a. dead was cloned on the basis of its ability, when present on a high-copy-number plasmid, to suppress temperature-sensitive mutations in rpsb, the gene encoding ribosomal protein S2. Ribosomal protein S2 binds late in the assembly process of the 30S subunit, and its binding is stimulated by S3 as well as other ribosomal proteins. S2 is among the group of proteins, which also includes S9, S10, and S14, that are necessary for efficient incorporation of ribosomal protein Si into the ribosome (29). Previous studies have shown that a temperature-sensitive rpsb mutant (RL425) has not only reduced levels of S2 but also reduced levels of ribosomal protein Si (6). Si has an RNA helix-unwinding activity, is important for the binding of mrna to the ribosome, and influences the affinity of ribosomes for the different mrna initiation sequences (60, 62). Therefore, the phenotypes associated with our temperature-sensitive mutants may be caused either directly by a mutated S2 or indirectly by the inability of an altered S2 to cooperate in the incorporation of another protein, such as Si, into the 30S subunit. Although some ribosomal proteins appear to be dispensable under some conditions, most are essential for ribosomal assembly and protein synthesis; thus, it is not surprising that mutations in rpsb give rise to a temperature-sensitive growth phenotype. However, how a mutation in rpsb causes the cell to become resistant to the histidine analog AT is still a matter of conjecture. One possibility, which is consistent with results obtained for all other AT-resistant mutants, is that the rpsb mutations enhance histidine biosynthesis, thereby rendering the cell resistant to AT. Since the mechanism of his operon attenuation involves the stalling of ribosomes at "hungry" histidine codons in the his leader region (33), it is conceivable that a defect in the translation machinery could cause contextual stalling of ribosomes at these sites, thus mimicking a decrease in concentration of trnahis. Alternatively, the mutation in rpsb may slow the growth of cells to a point that results in increased ppgpp, causing an increase in his operon expression. The strains used in this study carry a relc mutation that effectively blocks the rela-dependent pathway for ppgpp synthesis. Any increase in ppgpp levels would have to come from the rela-independent pathway (12). A second possibility is that a mutant rpsb, even at the permissive temperature, causes a decrease in synthesis of a subset of proteins that may be particularly sensitive to alterations in the translational apparatus. Perhaps one of these is a membrane protein necessary for transport of AT into the cell. The situation could be reminiscent of a search for extragenic suppressors of conditional-lethal mutations that affect the protein export pathway in E. coli. Many extragenic suppressors of both secy(ts) and seca(ts) protein secretion mutants map to genes whose products are part of the translational apparatus (31). These include a number of ribosomal protein genes, trna genes, and the gene for initiation factor 2 (31, 55). One mutation, which was isolated as an extragenic suppressor of seca(ts) and which maps to E. COLI RNA HELICASE 3299 rpso (secc), the gene encoding ribosomal protein S15, itself blocks the synthesis of transported proteins (16). The exact mechanisms underlying these observations are not clear; however, they suggest an interplay between the protein export system and the protein synthesis system. Therefore, a defect in a ribosomal protein could specifically affect the synthesis of membrane proteins, which in turn may cause AT resistance; this could also be responsible for the conditional-lethal phenotype seen in our strains. Germane to this hypothesis is the fact that we have characterized two other gene-dosage-dependent suppressors of the rpsb(ts) mutants, one of which maps very near a gene whose product has been implicated in regulation of membrane protein synthesis (unpublished observations). A gene-dosage-dependent suppressor of temperature-sensitive rpsb mutants encodes a putative RNA helicase. We have isolated a number of plasmid clones that can suppress the temperature-sensitive phenotype of rpsb. Two of these plasmids carry a gene-dosage-dependent suppressor, called dead. The dead gene product has extensive similarity to a family of proteins called the D-E-A-D family (34). The D-E-A-D family is a subset of a larger superfamily of proteins that consists of ATP-dependent proteins involved in nucleic acid recombination and/or replication (24, 54). To date, only P68 and eif4a have been shown biochemically to have RNA unwinding activity; however, the high degree of similarity among these proteins suggests that RNA binding and unwinding is a characteristic function of the entire family. The ability of dead to suppress a defect in a ribosomal protein suggests that it has a role in translation. Other D-E-A-D family members known to play a role in translation are the mouse translation initiation factor 4A, its yeast analogs TIF1/2 (35), and SPB4 (44) and E. coli SrmB (41) proteins, both of which are thought to function in ribosome assembly. In E. coli an srmb clone suppresses a temperature-sensitive mutation in rplx, the gene encoding ribosomal protein L24 (41). This mutation causes a defect in ribosome assembly that can be overcome by overproducing srmb. Nishi et al. (41) suggest that srmb, when overexpressed, may stabilize an unstable ribosome assembly product. DeaD shows no more similarity to SrmB (48%) than to some of the other eukaryotic proteins in this family (Fig. 6); therefore, it is unlikely that the two proteins share the same function in the cell. In fact, we have shown that dead, when overexpressed, cannot suppress the rplx temperature-sensitive mutation and that srmb, when overexpressed, cannot suppress the rpsb temperature-sensitive mutation (data not shown). Models for DeaD function. What is the normal role of DeaD in E. coli? Given its similarity to eif4a and its ability to suppress a mutation in a ribosomal protein, it is possible that DeaD has a role in translation. Overproduction of dead on a high-copy-number plasmid was shown to overproduce the dead transcript (Fig. 5) and presumably leads to an increase in the amount of DeaD protein. Increasing the cellular concentration of DeaD may either enhance its natural role or allow it to take on secondary functions which it might not normally perform. Figure 7 illustrates two possible models for the natural role of DeaD and the way in which increasing the amount of DeaD allows suppression of the temperature-sensitive mutations. As is shown in Fig. 7A, DeaD may be part of the normal translation machinery, but it may be rate limiting. Simply increasing the concentration of the protein may enhance the efficiency of translation in the mutants, coun-

10 3300 TOONE ET AL. J. BACTERIOL. A) mrna B) site FIG. 7. Models for how dead, in high copy number, might function to complement a mutant ribosomal protein that is defective in translation. (A) DeaD might function directly in translation, perhaps in the ejection of noncognate deacylated trna from the ribosome (a function postulated to be associated with the translation factor W [19]). Overproducing DeaD may improve the efficiency of the translation cycle at one point, counterbalancing the effect of a mutant ribosomal protein. Alternatively, DeaD may be acting in a manner similar to that postulated for SrmB, participating in the ribosome assembly process. Overproducing DeaD may stabilize an unstable assembly intermediate, either allowing ribosomes to form in the absence of S2 or allowing a mutant protein to be incorporated into the assembling ribosome. (B) DeaD may function as a gene-specific translation factor. By binding to specific mrnas, unwinding RNA secondary structure, and allowing access to ribosomes, DeaD would facilitate increased synthesis of certain gene products. These products may include the mutant rpsb gene itself or other genes whose expression is altered as a result of the mutation. Assuming that underproduction of certain genes contributes to the temperature-sensitive phenotype, overproducing DeaD would restore translation of these genes to a level that would allow growth at 42 C. terbalancing the detrimental effect of a mutant ribosomal protein S2. For example, the mutant S2 causes a decrease in the amount of ribosome-associated Si (whose association with the ribosome is partially dependent on S2 [29]); perhaps increasing the concentration of a second ribosome-associated RNA helicase could compensate for this deficiency. It is of interest that Ganoza et al. (19) have isolated a ribosomeassociated factor, called W, which has an approximate molecular mass of 62 kda and is necessary for efficient translation of MS2 phage RNA in vitro. W is thought to act after translation initiation by stimulating protein chain elongation, perhaps by removing noncognate, deacylated trna from the ribosome (19). In the second model (Fig. 7B), DeaD could be a translation accessory factor that activates or enhances the translation of specific genes, perhaps by denaturing RNA secondary structure. Its effect on rpsb may then be direct or indirect. An RNA helicase may be necessary for efficient translation of the rpsb mrna. Increasing the amount of DeaD may allow ISD more S2 to be made, effectively compensating for a defect in the S2 protein. Indirectly, overproducing DeaD may enhance the expression of proteins that are undersynthesized (which may cause or contribute to the temperature-sensitive phenotype as well as the AT resistance phenotype) in the rpsb mutants, thereby correcting the temperature-sensitive phenotype. This is somewhat similar to the role that eif4a is proposed to play in the internal initiation of translation and in the translation of naturally uncapped mrnas, such as poliovirus and encephalomyocarditis virus. According to this model, proposed by Sonenberg (59), free eif4a (not complexed with eif4f) and eif4b (which shows affinity for the AUG) bind to single-stranded mrna, allowing ribosome binding and internal initiation of translation. As an alternative to the above models, DeaD may act on the ribosome at an earlier stage in a manner similar to that postulated for SrmB, stabilizing an unstable intermediate in the assembly of ribosomes. The existence of at least two D-E-A-D proteins, facilitating similar processes, may

11 VOL. 173, 1991 indicate that there are at least two stages in ribosome assembly that require enzymatically modulated conformational changes in rrna. ACKNOWLEDGMENTS We thank Cassandra Smith for providing us with an E. coli OFAGE blot, A. Wittinghofer for the E. coli library, J. Schnier for sending us the srmb clone, and Sharon Durfy for assistance in mapping plasmid clones back to the chromosome. We thank our colleagues Michael Drebot, Jacques Archambault, Deborah Field, and Richard Collins for helpful suggestions and critical reading of the manuscript. This work was supported by the Natural Sciences and Engineering Research Council of Canada (grant A8060). REFERENCES 1. An, G., D. S. Bendiak, L. A. Mamelack, and J. D. Friesen Organization and nucleotide sequence of a new ribosomal protein operon in Escherichia coli containing the genes for ribosomal protein S2 and elongation factor Ts. Nucleic Acids Res. 9: An, G., J. Justesen, R. J. Watson, and J. D. Friesen Cloning the spot gene of Escherichia coli: identification of the spot gene product. J. Bacteriol. 137: Bachmann, B. J Linkage map of Escherichia coli K-12, edition 8. Microbiol. Rev. 54: Barry, G., C. Squires, and C. L. Squires Attenuation and processing of RNA from the rpljl-rpobc transcription unit of Escherichia coli. Proc. Natl. Acad. Sci. USA 77: Bendiak, D. S., and J. D. Friesen Organization of genes in the four minute region of the Escherichia coli chromosome: evidence that rpsb and tsf are cotranscribed. Mol. Gen. Genet. 181: Bollen, A., R. Lathe, A. Herzog, D. Denicourt, J. P. Lecocq, L. Desmarez, and R. Lavalle A conditionally lethal mutation of Escherichia coli affecting the gene coding for ribosomal protein S2 (rpsb). J. Mol. Biol. 132: Brennan, C. A., A. J. Dombroski, and T. Platt Transcription termination factor Rho is an RNA-DNA helicase. Cell 48: Brown, S Mutations in the gene for EF-G reduce the requirement for 4.5S RNA in the growth of E. coli. Cell 49: Brune, M., R. Schumann, and F. Wittinghofer Cloning and sequencing of the adenylate kinase gene (adk) of Escherichia coli. Nucleic Acids Res. 13: Bukhari, A. I., and A. L. Taylor Genetic analysis of diaminopimelic acid- and lysine-requiring mutants of Escherichia coli. J. Bacteriol. 105: Burgess, S., J. R. Couto, and C. Guthrie A putative ATP binding protein influences the fidelity of branchpoint recognition in yeast splicing. Cell 60: Cashel, M., and K. E. Rudd Stringent response, p In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 2. American Society for Microbiology, Washington, D.C. 13. Chang, T., J. Arenas, and J. Abelson Identification of five putative yeast RNA helicase genes. Proc. Natl. Acad. Sci. USA 87: Dalbadie-McFarland, G., and J. Abelson PRP5: a helicase-like protein required for mrna splicing in yeast. Proc. Natl. Acad. Sci. USA 87: Feinberg, A. P., and B. Vogelstein A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity (addendum). Anal. Biochem. 137: Ferro-Novick, S., M. Honma, and J. Beckwith The product of gene secc is involved in the synthesis of exported proteins in E. coli. Cell 38: Ford, M. J., I. A. Anton, and D. P. Lane Nuclear protein with sequence homology to translation initiation factor eif4a. E. COLI RNA HELICASE 3301 Nature (London) 332: Fouser, L. A., and J. D. Friesen Mutations in a yeast intron demonstrate the importance of specific conserved nucleotides for the two stages of nuclear mrna splicing. Cell 45: Ganoza, M. C., C. Cunningham, and R. M. Green Isolation and point of action of a factor from Escherichia coli required to reconstruct translation. Proc. NatI. Acad. Sci. USA 82: Grunberg-Manago, M Regulation of the expression of aminoacyl-trna synthetases and translation factors, p In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 2. American Society for Microbiology, Washington, D.C. 21. Hawley, D. K., and W. McClure Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res. 11: Heatwole, V. M., and R. L. Somerville Cloning, nucleotide sequence, and characterization of mtr, the structural gene for a tryptophan-specific permease of Escherichia coli K-12. J. Bacteriol. 173: Hershey, J. W. B Protein synthesis, p In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 1. American Society for Microbiology, Washington, D.C. 24. Hodgman, T. C A new superfamily of replicative proteins. Nature (London) 333: (Erratum, 333:578.) 25. Johnston, H. M., W. M. Barnes, F. G. Chumley, L. Bossi, and J. R. Roth Model for regulation of the histidine operon of Salmonella. Proc. Natl. Acad. Sci. USA 77: Kohara, Y., K. Akiyama, and K. Isono The physical map of the whole E. coli chromosome: application of a new strategy for the rapid analysis and sorting of a large genomic library. Cell 50: Laemmli, U. K Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227: Lasko, P. F., and M. Ashburner The product of the Drosophila gene vasa is very similar to eukaryotic initiation factor-4a. Nature (London) 335: Laughrea, M., and P. B. Moore On the relationship between the binding of ribosomal protein S1 to the 30S subunit of Escherichia coli and the 3' terminus of 16S RNA. J. Mol. Biol. 121: Lazinski, D., E. Grzadzielska, and A. Das Sequencespecific recognition of RNA hairpins by bacteriophage antiterminators requires a conserved arginine-rich motif. Cell 59: Lee, C. A., and J. Beckwith Suppression of growth and protein secretion defects in the Escherichia coli seca mutants by decreasing protein synthesis. J. Bacteriol. 166: Leroy, P., P. Alzari, D. Sassoon, D. Wolgemuth, and M. Fellous The protein encoded by a murine male germ cell-specific transcript is a putative ATP-dependent RNA helicase. Cell 57: Lewis, J. A., and B. N. Ames Histidine regulation in Salmonella typhimurium. XI. The percentage of trnahis charged in vivo and its relation to the repression of the histidine operon. J. Mol. Biol. 66: Linder, P., P. F. Lasko, M. Ashburner, P. Leroy, P. J. Neilson, K. Nishi, J. Schnier, and P. P. Slominski Birth of the D-E-A-D box. Nature (London) 337: Linder, P., and P. Slominski An essential yeast protein, encoded by duplicated genes TIFI and TIF2 and homologous to the mammalian translation initiation factor eif4a, can suppress a mitochondrial missense mutation. Proc. Natl. Acad. Sci. USA 86: Lipman, D. J., and W. R. Pearson Rapid and sensitive protein similarity searches. Science 243: Maloy, S. R., and W. D. Nunn Selection for loss of

12 3302 TOONE ET AL. tetracycline resistance by Escherichia coli. J. Bacteriol. 145: Maniatis, T., E. F. Fritsch, and J. Sambrook Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 39. Miller, J. H Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 40. Nielson, J. P., and H. Trachsel The mouse protein synthesis initiation factor 4A gene family includes two related functional genes which are differentially expressed. EMBO J. 7: Nishi, K., F. Morel-Devilie, J. W. B. Hershey, T. Leighton, and J. Schnier An eif4a-like protein is a suppressor of an Escherichia coli mutant defective in 50S ribosome subunit assembly. Nature (London) 336: (Erratum, 340:246, 1989.) 42. Pabo, C. O., and R. T. Sauer Protein-DNA recognition. Annu. Rev. Biochem. 53: Parker, J., R. J. Watson, J. D. Friesen, and N. P. Fill A relaxed mutant with an altered ribosomal protein Lli. Mol. Gen. Genet. 144: Query, C. C., R. C. Bentley, and J. D. Keene A common RNA recognition motif identified within a defined Ul RNA binding domain of the 70K Ul snrnp protein. Cell 57: Ray, B. K., T. G. Lawson, J. C. Kramer, M. H. Cladaras, J. A. Grifo, R. D. Abramson, W. C. Merrick, and R. E. Thach ATP-dependent unwinding of messenger RNA structure by eukaryotic initiation factors. J. Biol. Chem. 260: Regnier, P., M. Grunberg-Manago, and C. Portier Nucleotide sequence of the pnp gene of Escherichia coli encoding polynucleotide phosphorylase. J. Biol. Chem. 262: Rudd, K. E., B. R. Bochner, M. Cashel, and J. R. Roth Mutations in the spot gene of Salmonella typhimurium: effects on his operon expression. J. Bacteriol. 163: Rudd, K. E., and R. Menzel his operons of Escherichia coli and Salmonella typhimurium are regulated by DNA supercoiling. Proc. NatI. Acad. Sci. USA 84: Sachs, A. B., and R. W. Davis Translation initiation and ribosome biogenesis: involvement of a putative rrna helicase and RPL46. Science 247: Sancar, A., A. M. Hack, and W. D. Rupp Simple method for the identification of plasmid-encoded proteins. J. Bacteriol. 137: Sancar, A., and W. D. Rupp Physical map of the reca gene. Proc. Natl. Acad. Sci. USA 76: Sanger, F., S. Nicklen, and A. R. Coulson DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. J. BACTERIOL. USA 74: Scheffner, M., R. Knippers, and H. Stahl RNA unwinding activity of SV40 large T antigen. Cell 57: Seraphin, B., M. Simon, A. Boulet, and G. Faye Mitochondrial splicing requires a protein from a novel helicase family. Nature (London) 337: Shiba, K., K. Ito, and T. Yura Suppressors of the secy24 mutation: identification and characterization of additional ssy genes in Escherichia coli. J. Bacteriol. 166: Shine, J., and L. Dalgarno The 3' terminal sequence of Escherichia coli 16S ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. USA 71: Singer, M., T. A. Baker, G. Schnitzler, S. M. Deischel, M. Goel, K. J. Jaacks, A. D. Grossman, J. W. Erickson, and C. Gross A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli. Microbiol. Rev. 53: Smith, C. L., J. G. Econome, A. Schutt, S. Klco, and C. R. Cantor A physical map of the Escherichia coli K12 genome. Science 236: Sonenberg, N Cap-binding proteins of eukaryotic messenger RNA: functions in initiation and control of translation. Prog. Nucleic Acid Res. Mol. Biol. 35: Steitz, J. A., K. U. Sprague, R. C. Yuan, M. Laughrea, P. B. Moore, and A. J. Wahba RNA-RNA and protein-rna interactions during the initiation of protein synthesis, p In H. J. Vogel (ed.), Nucleic acid-protein recognition. Academic Press, Inc., New York. 61. Stephens, J. C., S. W. Artz, and B. N. Ames Guanosine 5'-diphosphate 3'-diphosphate (ppgpp): positive effector for histidine operon transcription and general signal for amino acid deficiency. Proc. Natl. Acad. Sci. USA 72: Szer, W., J. M. Hermoso, and M. Boublik Destabilization of the secondary structure of RNA by ribosomal protein S1 from Escherichia coli. Biochem. Biophys. Res. Commun. 70: Thanaraj, T. A., and M. W. Pandit An additional ribosome-binding site on the mrna of highly expressed genes and a bifunctional site on the colicin fragment of 16S rrna from Escherichia coli: important determinants of the efficiency of translation-initiation. Nucleic Acids Res. 17: Winkler, M. E Biosynthesis of histidine, p In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 1. American Society for Microbiology, Washington, D.C.