Vol. 171, No. 5

Genetic Analysis of the rnc Operon of Escherichia coliHOWARD E. TAKIFF, SU-MIN CHEN, AND DONALD L. COURTt*

Laboratory of Molecuilar Oncology, National Cancer Institute-Frederick Cancer Research Facility,Frederick, Maryland 21701

Received 10 November 1988/Accepted 8 February 1989

RNase III, an Escherichia coli double-stranded endoribonuclease, is known to be involved in maturation ofrRNA and regulation of several bacteriophage and Escherichia coli genes. Clones of the region of the E. colichromosome containing the gene for RNase III (rnc) were obtained by screening genomic libraries in A withDNA known to map near rnc. A phage clone with the rnc region was randomly mutagenized with a ATnJOelement, and the insertions were recombined onto the chromosome, generating a series of strains with ATn10insertions in the rnc region. Two insertions that had Rnc- phenotypes were located. One of them lay in the rncgene, and one was in the rnc leader sequence. Polarity studies showed that rnc is in an operon with two othergenes, era and recO. The sequence of the recO gene beyond era indicated it could encode a protein of -26kilodaltons and, like rnc and era, had codon usage consistent with a low level of expression. Experiments usingantibiotic cassettes to disrupt the genes rnc, era, and recO showed that era is essential for E. coli growth butthat rnc and recO are dispensable.

RNase III is an Escherichia coli endoribonuclease thatcleaves specific double-stranded RNA structures (12, 17). Itis known to initiate maturation of rRNA from 30S precursorRNA (11, 49) and has been implicated in the posttranscrip-tional processing of the mRNA transcripts of the early genesof bacteriophage T7 (15) as well as some phage T3 (26),phage lambda (27, 57), and E. coli genes (14, 32, 45, 51, 52).RNase III has been shown to process mRNA at the 5' sidesof genes either to increase or to decrease gene expression.Such processing may enhance expression by removing basepairing that blocks the ribosome-binding site (16) or reduceexpression by cleaving sequences required for ribosomebinding (45, 51). RNase III has also been shown to decreaseA int gene expression by processing at the 3' end of the intgene transcript (21, 23).Although RNase III is involved in varied and important

cellular processes, strains containing a missense mutation,rnc-105 (29), in the gene encoding RNase III show only a

moderate reduction in growth rate and alteration in theexpression of a few identified and unidentified proteins (14,20, 44). Ribosomes-from mutant strains have 23S rRNA thatis incompletely matured yet nevertheless appears to befunctional (31). It is conceivable, however, that RNase III isessential for E. coli, but the rnc-105 mutation does not causea lethal defect. In these studies, we demonstrate that RNaseIII is not essential for survival of E. coli.The rnc gene has been cloned previously, and the se-

quence of a -5-kilobase-pair (kbp) segment of DNA contig-uous to rnc has been reported (1, 37, 38, 43, 55, 58). Geneticstudies reported here show that rnc, era, and recO form an

operon. The Era protein binds GTP and contains a GTP-binding motif similar to that of the yeast RAS proteins (1).Recent genetic studies in another laboratory (P. T. Morri-son, S. T. Lovett, L. Gilson, and R. D. Kolodner, submittedfor publication) have determined that the open reading frame

* Corresponding author.t Present address: Laboratory of Chromosome Biology, BRI-

Basic Research Program, National Cancer Institute-Frederick Can-cer Research Facility, P.O. Box B, Frederick, MD 21701.

beyond era encodes the recO gene. Gene disruption exper-iments show that era is essential but that rnc and recO are

dispensable for growth. Subsequent reports (S.-M. Chen, H.Takiff, and D. Court, manuscript in preparation) will de-scribe the expression, purification, and biochemical proper-ties of Rnc and Era proteins.

MATERIALS AND METHODS

Materials. Restriction endonucleases and DNA-modifyingenzymes were purchased from Bethesda Research Labora-tories, Inc. (Gaithersburg, Md.), New England BioLabs,Inc. (Beverly, Mass.), and Boehringer Mannheim Biochem-icals (Indianapolis, Ind.). Nick translation kits were pur-chased from Bethesda Research Laboratories or BoehringerMannheim. The Sequenase kit was obtained from U.S.Biochemical Corp. (Cleveland, Ohio). Enzymes and kitswere used as recommended by the manufacturers.

Strains and genetic techniques. the E. coli and bacterio-phage X strains used are shown in Table 1. The generalgenetic, bacteriological, and phage techniques used are

described by Miller (41) or Silhavy et al. (47), as are standardrecipes for LB, TB, and M56 media. P1 transductions were

performed by using Plvir as described by Miller (41) exceptthat cells were spread on a nitrocellulose filter (BA85;Schleicher & Schuell, Inc., Keene, N.H.) on an LB plate,and the filter was transferred to an LB-antibiotic plate after2 h of incubation. Lysogens were made as described else-where (47), using tester phages W30 and W248 (Table 1) forX lysogens and 110 for X imm21 lysogens.Auxotrophs were screened on M56 glucose minimal

plates. Each plate was spread with 100 ,ul of hypoxanthine(7.5 mg/ml; purl), glycine (1%; glyA), or nicotinic acid(0.25%; nadB), as appropriate. Antibiotics were used in thefollowing final concentrations (micrograms per milliliter):ampicillin, 50; kanamycin, 25; chloramphenicol, 10; and

tetracycline, 12.5. For growth of tetracycline-dependent(Tetd) strains in liquid medium, tetracycline was used at 3.0,ug/ml.Development of lambda clones and recombinants. Two

different lambda phage libraries of the E. coli genome

2581

JOURNAL OF BACTERIOLOGY, May 1989, p. 2581-25900021-9193/89/052581-10$02.00/0Copyright C 1989, American Society for Microbiology

2582 TAKIFF ET AL.

TABLE 1. Bacterial strains and phages used

Strain Genotype Source or reference

E. coliDC7

DC546DC547HT5HT6HT21HT22HT27HT30HT31HT115HT119HT120HT121HT142HT142-1HT142-2HT142-3HT143HT143-1HT146HT148HT149HT150HT150-1HT150-2HT150-3HT150-4HT200HT210HT222N4903N6496NK6042PA3306RB132TAP135

W3110

rpsL his relAX (imm2' defective prophage)N4903 glyA::TnSN4903 rnc-105 glyA::TnSDC546 (X+)DC547 (A+)N4903 rnc-18::ATnJON4903 rnc40::ATnJON4903 rnc-14::ATnJON4903 TD1-13::ATnJON4903 TD1-17::ATnlOW3110 rnc-14::ATnJOW3110 rnc-18::ATnJOW3110 rnc40::ATnJOHT120 (X+)W3110 (XTD1 imm21 era::cat)HT142(pACS21)HT142(pACS1)HT142(PACS4)W3110 (XTD1 imm21 rnc::kan)HT143(PACS1)W3110 era::cat (pACS1)W3110 (XTD1 imm2' recO::kan)W3110 era::cat (pACS4)TAP135(ATD1 imm21 rnc::kan)HT150(pACS21)HT15O(pACS1)HT150(PJL6)HT150(pCE31)W3110 rnc::kan (pACS1)W3110 recO::kanTAP135 rnc::kan (pCE31)his ilvA relAl Strr pglA8rnc-105 glyA::TnSnadB: :TnlOpurI66gal zxx::TnJOA4HH1O4 (pNK217)W3110 galK lacd bioA AlacZ M15 [X c1857 Nam7 Nam53

(int-cIII)A(cro-J)A]

Phage ACharon 32Charon 28XTD1XTD1 imm2'XTD1 imm2' era::catXTD1 imm21 rnc::kanXTD1 imm2' recO::kanA+X b2 cIX h80 cIX imm434X imm434 cI Ram54 Ram6OX imm21 hy5 cI

7

N4903 + P1 N6496N4903T Pl N6496This workThis workThis workThis workThis workThis workThis workW3110 + P1 HT27W3110 + P1 HT21W3110 +Pr HT22This workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workThis workNIH' collectionNIH collectionN. Kleckner4856T. Patterson

4

446This workThis workThis workThis workThis workWl of the NIH collectionW30 of the NIH collectionW248 of the NIH collectionG3 of the NIH collectionG37 of the NIH collectionB10 of the NIH collection

a NIH, National Institutes of Health, Bethesda, Md.

(provided by F. Blattner) were screened by plaque hybrid-ization (36) to the nick-translated probe pTD101 (13, 38).XTD1 and XTD2 were selected from a Charon 32 library ofthe E. coli genome prepared by partial digestion of the E. colichromosome with EcoRI (35). ATD3 was selected from aCharon 28 library prepared by digestion of the E. colichromosome with BamHI (46). XTD1 imm21 recombinantswere obtained after infection by XTD1 of strain DC7 (7),which contains an imm21 defective prophage. The XTD1

imm" recombinants were selected on the immX lysogenicstrain HT5. XTD1 is defective for cI repressor and deletedfor att. TD1 A imm2' is c+ and deleted for att.The era::cat, rnc::kan, and recO::kan insertions were

recombined from plasmids to XTD1 imm2' by growing thephage on W3110 strains containing the insertion-bearingplasmids. The resulting lysates were then screened for phageable to make antibiotic-resistant lysogens with fresh W3110cells. Similarly, the ATnJO insertions were recombined from

J. BACTERIOL.

RNase III OPERON 2583

the chromosome to XTD1 imm2' by growing the phage oninsertion-bearing strains and screening the lysates for phageable to make tetracycline-resistant lysogens.

Construction of plasmids. For plasmid constructions (seeFig. 3), standard procedures for recombinant DNA work andtransformations were performed as described by Maniatis etal. (36). pBR322 was obtained from Boehringer Mannheim,and pUC19 was obtained from Bethesda Research Labora-tories. Plasmid pUC4K, which contains the gene (kan) forkanamycin resistance (Kanr), was obtained from Pharmacia,Inc. (Piscataway, N.J.). pTAP1, which contains the chlor-amphenicol acetyltransferase (cat) gene cassette, was ob-tained from T. Patterson (unpublished data). pTD101, whoseSalI-to-BamHI insert contains the genes lep, lepA, and rnc,was obtained from W. Wickner (13). Plasmid pACS1, for-merly termed pSB (1), was made by subcloning the 4.3-kbpEcoRI fragment from XTD1 into the EcoRI site of pBR322 inthe orientation shown in Fig. 3. pACS21 was constructed bycutting pACS1 with BamHI, ligating the ends, and isolatingampicillin-resistant (Amp) and tetracycline-sensitive (Tets)transformants. Similarly, pACS3 and pACS4 were made bycutting pACS1 with NruI and AvaI, respectively. pACS22was made by cutting pACS21 with BssH2 and ligating todelete a 40-bp BssH2 fragment from within rnc. pACS2 wasmade by cutting pACS1 at a BstXI site within rnc, treatingwith T4 polymerase in the presence of the four deoxynucle-otide triphosphates, and then ligating, thereby creating a4-bp deletion. pACS5 was constructed from a partial EcoRVdigest of plasmid pACS1 ligated with the cat gene cassettecut from plasmid pTAP1 with HinclI. Transformants wereselected as chloramphenicol resistant (Cm9 and tetracyclineresistant (Tet') to ensure that the cat gene would be insertedinto the EcoRV site in era (1,712 bp beyond the EcoRI site inlep) and that the EcoRV site in pBR322 tet would remainintact. Similarly, pKKF3 was constructed from a partialNruI digest of plasmid pACS1 ligated with the kan gene cutfrom plasmid pUC4K with HincII (obtained from K.Kawakami). By selecting for Kanr and Tetr, a plasmid wasobtained with the kan gene inserted in an NruI site in recO468 bp beyond the presumed termination codon for era and2,678 bp beyond the EcoRI site in lep (see Fig. 4). pACS23was constructed by cutting pACS21 with BssH2 to delete the40-bp fragment in rnc (beginning 798 bp beyond the EcoRIsite in lep), blunting the ends with Klenow polymerase, andligating with the HinclI kan gene cassette from pUC4K.pACS6 was constructed by cutting both pACS23 and pACS1with SalI and BstXI. The BstXI-kan-amp-SalI fragment ofpACS23 was ligated to the BstXI-era-SalI fragment ofpACS1. pCE31 was constructed by inserting the era genecontaining the BstXI-to-NruI fragment from pACS1 intopJL6 (34) and then removing all of the rnc sequence to a ClaIsite in the third codon of era and replacing it with a syntheticoligonucleotide containing a strong ribosome-binding site(Chen et al., in preparation).

Plasmids for sequencing the ATnJO insertions were madeby digesting with PstI the DNA from XTD1 imm21 phagecarrying the insertions. Note that the PstI sites are located inthe rnc leader and in era. The resulting 4.3-kbp PstI frag-ment, carrying the entire ATnJO insertion, was ligated intopUC19.XTD::ATn1O insertions and transfer of ATnJO to the chro-

mosome. Phage clones XTD1, XTD2, and XTD3 were grownon the defective TnJO transposon A16A17TnJO (ATnJO)donor strain RB132 (56). Approximately 1 in 106 phage fromsuch infections of RB132 carries a stable ATnJO insertion inits DNA (40), which confers Tetr. The three XTD lysates

prepared on RB132 were used to infect overnight cultures ofthe A lysogen HT6 at a multiplicity of 1 phage per bacterium.The infected cultures were spread on LB-tetracycline plates,and Tetr colonies were selected. These colonies maintainedthe ATnJO within the prophage location, as judged by thefrequency of Tetr phage isolated by spontaneous induction.To transfer just the ATnJO from the phage to the chromo-

some of E. coli, an overnight culture of the nonlysogenicstrain N4903 was infected with a XTD1 lysate prepared onRB132, and the infected cultures were spread on LB-tetra-cycline plates. To reduce phage viability and increase ho-mologous recombination between the phage and the E. coli,the phage for this infection were preirradiated with UV light(approximately 6 x 103 ergs/mm2; Blak-Ray J-225 shortwaveUV meter).

Tests for the Rnc- phenotype in recombinants or duringcomplementation. Temperate lambdoid phages normallymake turbid plaques on E. coli; on rnc mutants, however,these phages form clear plaques (2). Plaque morphology wasdetermined by spotting 10 to 20 ,ul of a lysate of X or X imm434onto a freshly poured lawn of the test bacterium in TB topagar on a TB agar plate. For the Tetd strains, tetracyclinewas added to the top agar to a final concentration of 6.25,ug/ml. Incubation was at 32 or 37°C. Rnc- strains also makeincreased levels of polynucleotide phosphorylase (45), whichcan be seen on a sodium dodecyl sulfate-polyacrylamide gelof an Rnc- cell lysate (47) as a band running at approxi-mately 84 kilodaltons. A third test for Rnc- used herecompares the sizes of rRNA produced; rnc+ strains have23S and 16S species, whereas rnc- mutants have 30S, 23S,and 16S species, as determined by gel electrophoresis (17).

Physical mapping of the ATn1O insertions. Bacterial DNAfrom the insertion strains was isolated as described bySilhavy et al. (47). Southern blots of the bacterial DNAdigested with various restriction endonucleases were hybrid-ized (36) to nick-translated phage or plasmid probes contain-ing DNA from the rnc region. The ATnJO insertions weremapped by using BamHI, EcoRI, HincII, HindIII, and PstIdigests of chromosomal DNA from the insertion strains. Thecat cassette disruptions of era were confirmed by usingEcoRV digests, and the kan cassette disruptions of rnc wereconfirmed by using BstXI.DNA sequencing. Sequencing was performed with the

Sequenase kit (U.S. Biochemical Corp.) and [32P]dCTP,using miniprep plasmid DNA. Locations of the ATnJO inser-tions were determined by sequencing from synthetic oligo-nucleotides hybridized within each end of the ATnJO ele-ment, priming synthesis into the E. coli DNA. The DNAsequence of recO was determined from both strands ofplasmid pACS1 or pACS3, using a series of synthetic oligo-nucleotide primers.

Haploid segregation experiments. W3110 strains containingXTD1 imm21 prophages with an antibiotic resistance inser-tion (rnc::kan, era::cat, or recO::kan) were spread on TBplates at a concentration of 5 x 103 cells per plate with 2 x106 A imm434 Ram54 Ram6O phage and incubated at 39°C.Lysogenic cells will be lysed and killed by the incoming Ramphage; lysis requires complementation by the intact R geneon the prophage. There will be some cells in the originalculture (0.1%) that have spontaneously lost the prophage(and its intact R gene) by homologous recombination andcellular segregation (see Fig. 5). These form colonies resis-tant to lysis by phage Ximm434 R. Thus, there is a selectionfor haploid segregants from the partial diploid lysogens.

Test for the RecO phenotype. The recO mutants are moresensitive than the wild type to UV irradiation (33). For

VOL. 171, 1989

2584 TAKIFF ET AL.

Complementationof rnc

17V

glyA -/Pur,EE

13V

recO era

14 40vvrnc lop lepA

B E

E

B

B 16

E

E 6.0

E

4.3 E 4.2

-H pTD101S

XTD3B

XTD2 1

EXTD1

I

E

pACS1I

FIG. 1. XTD1, XTD2, and XTD3 phage clones. These three clones were obtained as described in Materials and Methods by screening E.coli genomic libraries with the nick-translated plasmid probe pTD101. The plaque test for RNase III was used to determine whether thesephage, as stable lysogens, complement the rnc-lO5 mutation (see Materials and Methods). Four ATnJO insertions, rnc13::ATnJO,rnc-14::zATnJO, TD-17::ATnJO, and rnc-40::zTnlO, are designated 13, 17, and 40, and their locations (V) are shown relative to genes nadB,glyA, and purI. Linkage of glyA is 55% to TD-17 and 45% to m-nc. Linkage of TD-17 to rnc is 65%, thereby indicating the order of genes shownabove. Three-factor crosses with these markers support this order. The numbers 6.0, 4.2, and 4.3 are the sizes in kilobase pairs of the threetandem EcoRI fragments of the bacterial insert of XTD1; 16 is the size in kilobase pairs of the BamHI insert in XTD3. Restriction sites: BamHl(B), EcoRI (E), Sall (S).

testing, bacterial colonies were streaked on an LB agar

plate, and part of the streak was irradiated with UV at dosessufficient to kill recO mutants but not isogenic wild-typestrains (33).

RESULTS

The rnc gene. Clones of the gene (rnc) for RNase III wereobtained by hybridizing phage libraries of the E. coli genomewith DNA known to map near rnc. The rnc gene maps at 55min, near lep on the E. coli chromosome (5). PlasmidpTD101, which contains the genes lep and lepA (13, 38), wasused as a hybridization probe to screen lambda librariescontaining segments of the E. coli genome. Three phageclones, XTD1, XTD2, and XTD3, whose bacterial DNAinserts hybridized to the probe (Fig. 1), were obtained.Derivatives of these phage clones were used to lysogenizestrain HT6, carrying the rnc-105 mutation (Fig. 1; see

below). XTD1 and XTD3 complemented the defects of thernc mutant (Materials and Methods), whereas XTD2 did not.Therefore, RNase III could be functionally mapped to in-

clude the 1.4-kbp EcoRI-to-BamHI fragment contained on

XTD1 and XTD3 but not XTD2. This fragment was carried on

pTD101, which also complemented inc-105.ATnJO insertions in the rnc region. To facilitate the study

of rnc, a series of ATnJO insertions in the rnc region was

isolated. XTD1 was mutagenized with a mini-TnJO transpo-son, A16A17TnIO (zXTnlO) (see Materials and Methods), andtwo techniques were used to transfer the ATnJO elementsfrom the bacterial segment of XTD1 to the chromosome. Byone method, the ATnJO insertions were recombined directlyfrom the ATnJO-mutagenized phage DNA onto the chromo-some of N4903, using selection for Tetr. XTD1 lacks a phagerepressor gene and cannot form a lysogen. Therefore, afterinfection, all Tetr colonies analyzed contained simple insertsof the /Tn1O recombined from the bacterial segment carried

on the phage into the homologous region of the chromosome(see Materials and Methods). Figure 1 shows the relativelocations in the chromosome of insertions TD1-13::ATnJO,rnc-14:::ATnlO, and TD1-17::ATnJO, as determined bySouthern blot hybridization (data not shown). P1 transduc-tion was used to transfer the ATn1O insertions to otherstrains and also to determine linkage frequencies betweenthese insertions and the genes rnc, glyA, purI, and nadB.Strains with insertion rnc-14::/TnJO were phenotypicallyRNase III defective (Rnc-), and the ATnJO element was

located within the rnc gene, as analyzed by both Southernblot mapping and sequencing (see below).

Physical mapping and the linkage frequencies betweenTD-17::zATnJO, glyA, and rnc mutations (see legend to Fig. 1)indicated the order of the lep-rnc-era-recO segment relativeto glyA (Fig. 1). Other linkage frequencies (data not shown)for putrL and nadB agreed with this order and with frequen-cies published by others (3, 25, 54).

In the second method for screening and isolating ATnJOinsertions within the rnc gene, the zXTnJO insertions, withtheir XTD1 carrier, were integrated into an rnc mutant host(HT6) by selecting for Tetr. Strain HT6 is already lysogenicfor a wild-type X phage which provides A repressor and DNAhomology. The cI XTD1: :ATnJO phage can then stablyintegrate by homologous recombination into either prophageor bacterial DNA. The rnc gene on the integrated phageshould complement the rnc mutant host unless a ATnlOinsertion affects rnc expression. A total of 48 Tetr coloniesisolated by this method were diploid for rnc, and most were

Rnc+. Five, however, remained Rnc-, and the insertions intwo of these, rnc-18 and rnc-40, were transduced by P1 tothe original N4903 strain as well as to strain W3110 byselection for Tetr. Some of the transductants isolated were

Rnc- and nonlysogenic for X. Southern blot hybridizationshowed that both insertions (rnc-18 and rnc40) were located

+

r-,AY I nadBS E B

+

+

J BACTERIOL

E

RNase III OPERON 2585

381 CGTGGCTAAC GACATCCCCC GTCGTTGTGT ATAGAATATT CCCCCGAAGT TTAAGGTTGG 440

441 CACCTCCAGG TTGCCACGGC ACACGAAACA GCGTTGGTCC CCATATACCG GTAAACTGAA 500

501 ACTGCAGCGA AGCAGTTAGC AGAACCATGT ATCAGGTC TGTTTCGTGT GCTGAATTGT 580Pstl A40

561 TGACGCATTT ATTTATTGGT ATCGCATGAA CCCCATCGTA ATTAATCGGC TTCAACGGAE 620Hinc[l

621 GCTGGGCTAC ACTTTTAATC ATCAGGAACT GTTGCAGCAG GCATTAACTC ATCGTAGTGC 680A14

681 CAGCAGTAAA CATAACGAGC GTTTAGAATT TTTAGGCGAC 720IA

RNC 105FIG. 2. Locations of Rnc- ATnJO insertions. Dots are placed over the -10 region (43) of the rnc promoter, and asterisks are under the

rnc translation start. Two long arrows denote a putative mRNA stem; sites for restriction enzymes HincIl and PstI are underlined. The twoboxes labeled A40 and A14 indicate the 9-bp duplications surrounding the insertions rnc40 and rnc-14, respectively. The rnc-105 pointmutation is also shown (43). Base number 381 indicates the position in base pairs beyond the EcoRI site in lep.

between a PstI site and a HincII site that lay 85 and 28 bp,respectively, upstream of the rnc translational start site.Therefore, these insertions mapped outside of the structuralrnc gene yet appeared to block expression of RNase III.Sequence locations of ATn1O insertion sites. The three

insertions, rnc-14::ATnJO, rnc-18::ATnJO, and rnc40::ATnJO, which affect expression of the rnc operon wererecombined from the chromosome onto XTD1 imm2' andsubcloned into plasmids; their insertion sites were deter-mined by sequencing the junction fragments (Materials andMethods; Fig. 2). In rnc-14, the ATnIO was found to bewithin rnc, 43 bp from the translation start site. In bothrnc-18 and rnc40, the ATnJO was inserted at the same siteand in the same orientation, 54 bp upstream of the initialATG for rnc. Subsequent analyses of rnc-18 and rnc40yielded identical results, and only those for rnc40 will bediscussed.ATnJO polarity on an essential gene. When the rnc40::

ATnJO insertion was transduced by P1, growth of the hap-loid, nonlysogenic transductants became Tetd. The originalTetr, diploid parental lysogen had not been dependent ontetracycline for growth. When grown on TB-tetracycline(12.5 ,ug/ml) plates at 37°C, the haploid transductants formedcolonies that were slightly smaller than the wild-type colo-nies; without tetracycline, colonies were not visible afterovernight incubation. Even in the presence of tetracycline,the strain remained Rnc-. This insertion therefore appearedto have polar effects on rnc and a gene essential for bacterialgrowth, which, as shown below, is not rnc. Note that asimilar polar effect was not found for the rnc-14 insertion.Both polarity defects of rnc40 were complemented by

plasmid pACS1, which contains the 4.3-kbp EcoRI fragmentof the rnc region (Fig. 3). The rnc defect, but not the growthdefect, is complemented by plasmid pACS21, which con-tains just the EcoRI-BamHI fragment. The growth defect iscomplemented by plasmids containing an intact era gene,which is just downstream of rnc (1); an intact rnc gene is notsufficient for growth complementation (Fig. 3).Because the DNA sequence from the region beyond era

revealed an open reading frame to which the gene recO hasbeen mapped (Fig. 1 and 4), it was conceivable that thegrowth defect was caused by polar effects on recO. How-ever, the growth defect was complemented by plasmidpACS4, which contains all of era but not recO. Thus, theATnJO insertions just in front of rnc had a polar effect on theera gene, and bacterial growth appeared to depend on theexpression of era.

era is essential; rnc and recO are dispensable. Each of thethree genes rnc, era, and recO was interrupted by insertionof an antibiotic resistance cassette as shown in Fig. 5 for theera gene. A cat gene cassette was inserted into the era geneto form plasmid pACS5 and then recombined onto phageXTD1 imm21 (see Materials and Methods). The resultingphage can form a stable, immune lysogen but lacks anattachment site; therefore, Cmr lysogens are formed only byhomologous recombination at the rnc era recO chromosomalsegment. This lysogenic strain will be diploid only for thehomologous bacterial segment carried by the phage (Fig. 5).During growth of the lysogenic strain, approximately 0.1%of bacteria are spontaneously cured of the prophage (andpartial diploidy) by homologous recombination. The pro-phage can recombine out in two ways, leaving the chromo-some either with the intact era gene or with the copy of eradisrupted by the cat gene cassette. In general, if the genedisrupted by the cassette is nonessential for survival, thenboth Cms and Cmr haploid segregants should be found.However, if the gene is essential or the cassette has polareffects on an essential gene, then haploid segregants thatmaintain the cassette will fail to form colonies and only Cmssegregants will be recovered.

In the experiment using the era::cat insertion, haploidsegregants that had lost the prophage were selected (seeMaterials and Methods) and tested for sensitivity to chlor-amphenicol (Fig. 5). No Cmr segregants were isolated unlessthe strain carried a plasmid with an intact copy of the eragene. Southern blotting confirmed the physical map ofW3110 and HT142 and showed that a Cmr segregant, HT146,isolated in the presence of plasmid pACS1, carried the catgene inserted in the single chromosomal copy of era (datanot shown). This finding suggests that either era is essentialor the insertion cassette is polar on an essential genedownstream. Since plasmid pACS4, which lacks distalgenes, also allows segregation of era::cat (5 Cmr of 12haploid segregants), era is the essential gene. Note that inthese latter Cmr segregants (i.e., HT149), the strain becameUV , a phenotype indicating a polarity on recO.To test whether recO encodes an essential function not

fully affected by upstream insertions in era, a haploidsegregation ekperiment (see above) was performed withHT148. This strain harbors a prophage with a kanamycinresistance cassette (kan) inserted at an NruI site within recO(Fig. 3). Of the haploid segregants from strain HT148, 78%were resistant to kanamycin and sensitive to UV light, whichindicated that recO is not essential (Table 2).

VOL. 171, 1989

2586 TAKIFF ET AL.

/'co N A N

Compl. Compl. rB recO era rncera rnc 0RV B B eNRvVB

"E B B NA Bs RV B C BX Bs H E+ + pACS1- t

+ - pACS2

+ + pACS3

+ + pACS4 CAT

- + pACS5 KAN+a _ pACS6

+ b _ pACS7 KAN

+ + pKKF3

- + pACS21- - pACS22 KAN- - pACS23 V

+ - pCE31 %t w.. pJL6FIG. 3. Plasmid complementation of rnc and era defects. The plasmids shown (see Materials and Methods for constructions) were

transformed into the Tetd strain HT120. Transformants were then tested for the ability to complement rnc by the RNase III plaque test (seeMaterials and Methods). Complementation of era was determined by the ability of transformants to form visible colonies on TB plates withouttetracycline after incubation for 24 h at 37°C. Symbols: x, former BstXI site that was cut, repaired with Klenow polymerase, and joined byligation; 41, deletion of a 40-bp BssH2 fragment from within rnc. Abbreviations for restriction sites: A, AvaI; B, BamHI; Bs, BssH2 (notethere are two BssH2 sites within rnc, 40 bp apart); BX, BstXI; C, ClaI; E, EcoRI; H, HincII; N, NruI; RV, EcoRV; S, Sall. Superscripts:a, the kan insertion in rnc is incompletely polar on era such that it will complement HT120 on plasmid pACS6 but not in single copy on a XTD1imm2' rnc::kan prophage; b, plasmid pACS7 complements era poorly such that tetracycline-independent growth of HT120 is slower than withpACS1 but faster than with no plasmid. Complementation by pCE31 is in strain HT121.

2131 CGCACTGCGC AGTCTCGGTT ACGTTGACGA TCTTTAAGAG TAACTCCGAT GGAAGGCTGG 2190

2191 CAGCGCGCAT TTGTCCTGCA ETACGCCCG TGGAGCGAAA CCAGCCTGAT GCTGGACGTC 2250

2251 TTCACGGAGG AATCGGGGCG CGTGCGTCTG GTTGCCAAAG GCGCACGCTC TAAACGCTCT 2310

2311 ACCCTGAAAG GTGCATTACA GCCTTTCACC CCTCTCTTGC TACGTTTTGG CGGGCGTGGC 2370

2371 GAAGTCAAAA CGCTGCGCAG TGCTGAAGCC GTCTCGCTGG CGCTGCCATT AAGCGGTATC 2430

2431 ACGCTTTACA GCGGTCTGTA CATCAACGAA CTTCTCTCCC GCGTACTGGA ATACGAGACG 2490

2491 CGCTTCTCTG AACTTTTTTT CGATTACTTG CACTGCATTC AGTCTCTTGC AGGGGTCACT 2550

2551 GGTACGCCAG AACCCGCGCT GCGCCGCTTT GAACTGGCAC TGCTCGGGCA TCTGGGTTAT 2610

2611 GGCGTCAATT TTACCCATTG TGCGGGTAGC GGCGAGCCGG TAGATGACAC CATGACGTAT 2670

2671 CGTTATCGCG AAGAAAAAGG GTTTATCGCA AGCGTCGTTA TCGACAATAA AACGTTCACC 2730

2731 GGAAGGCAGT TAAAAGCGTT AAACGCACGG GAATTTCCTG ACGCAGACAC ACTGCGCGCC 2790

2791 GCGAAACGCT TTACCCGCAT GGCGCTTAAG CCGTATCTTG GCGGTAAACC TTTAAAGAGC 2850

2851 AGGGAACTGT TCCGGCAGTT TATGCCTAAG CGAACGGTGA AAACACATTA TGAAfG.GA 2910

FIG. 4. DNA sequence of the region beyond era, encoding recO. The termination codon for era is marked by the first boxed TAG. PossibleATG initiation codons for recO are indicated by asterisks. Base number 2131 indicates the position in base pairs beyond the EcoRI site in lep(37, 38). The putative termination codons for recO, TGATGA, are also boxed. The sequence of era shown differs from the published sequence(1) by the addition of one G at base 2223.

J. BACTERIOL.

RNase III OPERON 2587

-F

W3110 (TD1 Ximm21 era::cat)

cat__ XIMM21 1

recO era rnc

Plasmid

NonepACS21 (mc)

pACS1 (rnc era)

l Chromosome

recO era rnc

HaploidSegregants

CmS CmR

25 036 014 22

FIG. 5. Results of the haploid segregation experiment to determine an era requirement. The W3110 (XTD1 imm2' era::cat) lysogen,HT142, was transformed with pACS21 (rnc) and pACS1 (rnc era). Symbols: Er!, bacterial chromosomal homology to DNA cloned on XTD1;, phage DNA flanked by the homology. Haploid segregants were isolated as described in Materials and Methods and tested for sensitivity

to chloramphenicol.

Finally, insertion of a kan cassette into the rnc gene wasexamined for effects on cell viability. Of 35 haploid segre-gants isolated from the culture of strain HT143, a X immTD1 (rnc: :kan) lysogen of W3110, none were Kanr (Table 2).This finding suggests that either rnc is an essential gene orthe insertion is polar on an essential gene. Kanr segregantsstill could not be obtained from HT143 when rnc alone wassupplied on a plasmid (pACS21) but could be isolated whenboth rnc and era were provided (pACS1). Although thisresult shows the polarity of the kan insertion on the essentialera gene, it does not reveal whether rnc itself is essential.To test specifically whether rnc is essential, the haploid

segregation experiment for rnc::kan was performed withplasmid pCE31, which expresses era under the control of thelambda PL promoter and contains no homology to the rncgene. Previous experiments had shown that the leakyexpression of era from the repressed PL promoter of pCE31is sufficient to complement a strain with the rnc40::ATnJOinsertion (HT121) and allow tetracycline-independentgrowth. Therefore, expression from the repressed PL pro-moter of pCE31 at 32°C was used to compensate for thepolar effects of rnc::kan on era.

Haploid segregants from strain HT150, with the pJL6vector (HT150-3) or pCE31 (150-4), were examined for

TABLE 2. Segregation of antibiotic resistance cassettesin haploid survivorsa

Diploid parent Plasmid No. of haploidsegregants

W3110 (XTD1 imm21 Kans Kanrrnc: :kan)

HT143 None 35 0HT143-1 pACS1 (rnc era) 5 2

TAP135 (XTD1 imm21rnc: :kan)

HT150 None 19 0HT150-1 pACS21 (rnc) 36 0HT150-2 pACS1 (rnc era) 18 13HT150-3 pJL6 58 0HT150-4 pCD31 (era) 20 17

W3110 (XTD1 imm21recO: :kan)

HT148 None 4 14

a Haploid segregation experiments with HT150 were performed at 32°C; allothers were performed at 39°C.

sensitivity to kanamycin. With the pJL6 vector, only Kanshaploid segregants were obtained, as expected. However,when era was supplied from plasmid pCE31, 46% of thehaploid segregants were Kanr (Table 2). Southern blottingconfirmed that the Kanr segregants contained a single dis-rupted copy of rnc. This result demonstrates that rnc is notrequired for E. coli growth but that rnc::kan is polar on theexpression of the downstream, essential era gene.

DISCUSSIONTo learn about RNase III and its role in E. coli, we sought

to clone the associated gene and study its expression. Aphage carrying the rnc region was isolated from a library ofthe E. coli genome, and the genes were subcloned ontoplasmids. By mutagenizing this phage with a ATnJO elementand transferring the insertions onto the chromosome, aseries of strains with ATnJO insertions in the rnc region wasobtained. These insertions allowed rnc to be mapped andtranscriptionally oriented with respect to surrounding genesand also yielded genetic evidence for an essential gene, era.The rnc-era-recO region has been sequenced; it can en-

code peptides in the range of 25 kilodaltons for the rnc geneproduct, 35 kilodaltons for the era gene product, and 26kilodaltons for the recO gene product. We have expressedand purified the protein products of both rnc and era (Chenet al., in preparation). The sequence of recO shows twopossible in-frame ATG initiation codons (Fig. 4), one ofwhich lies within the carboxy-terminal end of the era gene.Evidence indicates that rnc and era are part of the same

operon. The termination codon of rnc and the initiationcodon for era overlap. Expression of both rnc and era isblocked by a ATnJO insertion, rnc40, upstream of rnc.Growth of strains carrying this insertion depends on thepresence of tetracycline, and this dependence is comple-mented by plasmids containing a functional era gene. Inaddition, the haploid segregation experiment showed that akan cassette inserted in rnc has a polar effect on eraexpression. This polar effect could be caused either bytranscriptional termination before era or by a lack of trans-lational coupling of rnc and era (44).The UV sensitivity of strain HT149, with a cat insertion in

era, suggests that recO is part of the rnc-era operon.Furthermore, preliminary information suggests that genesdistal to recO are not part of this operon (data not shown).Like rnc and era, recO has the poor codon usage character-istic of genes normally expressed at low levels. Both rnc andera are expressed at low levels (Chen et al., in preparation).

VOL. 171, 1989

2588 TAKIFF ET AL.

The rnc40::ATnI0 insertion is located in the leader tran-script, within a putative stem-loop structure (Fig. 2) that isprocessed by RNase III (J. Bardwell and D. Court, unpub-lished data). It exhibits a polar effect, blocking expression ofboth rnc and era, but allows expression of era in thepresence of tetracycline, perhaps from a transcript originat-ing in the ATnJO element. This element contains two diver-gent tetracycline-induced genes, tetA and tetR (8, 9, 24), thatare oriented in rnc40 such that the promoter for the tetRgene reads in the direction of rnc. Tetracycline and heatedchlortetracycline, a gratuitous inducer of the tet operon (10),allow growth of these insertion strains. We do not know whyrnc is not also dependent on the presence of tetracycline, butthis difference may be related to the regulation of the rncoperon or to levels of expression required for each pheno-type. If, for example, the expression of both proteins fromthe tetR promoter were reduced 10-fold compared with thelevel of the normal rnc promoter, this low level might beadequate to allow for full function of Era but not of RNaseIII.Experiments in which the genes of the rnc operon were

disrupted with antibiotic cassettes and ATnJO transposonsshow that era is essential for colony growth but that rnc andrecO are not. The results for era agree with those presentedin a recent report by March et al. (39), who could not isolatea disrupted era gene at 42°C. Morrison et al. (submitted)have shown that insertion elements located throughout therecO gene affect UV sensitivity and recombination but notnormal cell viability. The results for m-c are consistent withthe phenotype of strains carrying the rnc-14::ATnJO inser-tion. This insertion is located within the rnc structural geneand therefore should constitute a null mutation, yet itsphenotype is similar to that of rnc-105 mutant strains. We donot understand why this particular insertion, although in thesame orientation as rnc40::ATnJO, is not polar on eraexpression. One possibility could be that translation resumesin the distal segment of the rnc gene, allowing sufficientexpression of era. We hope that further studies to under-stand the different effects of these two insertions will help usto learn about the normal expression of the operon.

If RNase III is not essential, then its processing of mRNAand rRNA must also be dispensable or accomplished byother RNases (18). It has been shown that in rnc-105 strains,processing of the 30S rRNA precursor is slowed but 16SrRNA is completely and accurately processed (30), whereas23S rRNA is incompletely matured (31).Nashimoto and Uchida (43) found cold-sensitive lethal

mutations that map in rnc. They propose that these muta-tions may prevent E. coli growth at low temperatures byimpairing ribosome assembly. We have tested rnc mutantstrains carrying the rnc-105 mutation, the ATnJO insertionsrnc-14 and rnc40, and rnc::kan (HT222) and found that allcan grow at temperatures as low as 18°C. Therefore, mostrnc mutations, including nulls, fail to cause a cold-sensitivephenotype. Nashimoto et al. also isolated another mutationin rnc, rev-3, which suppressed a temperature-sensitivemutation in ribosomal protein S12 (42). Stern et al. (50)recently showed that S12 binds to sequences in the 5' end of16S rRNA and suggested that this interaction may be anearly step in ribosome assembly. RNase III, in addition tocleaving 16S rRNA, may play a structural role in ribosomeassembly by interacting with S12. Although our results showthat RNase III is not essential for ribosome assembly even at18°C, the particular cold-sensitive rtzc mutation isolated byNashimoto et al. may cause RNase III to block or poison

assembly at lower temperatures, at which the process, atleast in vitro, is less efficient (53).The Era protein. The function of Era remains unknown.

Although the protein was named because of its proposedamino acid homology to the yeast RAS I protein, subsequentcomputer-assisted analysis (A. Barber, J. Bardwell, H.Takiff, D. Court, and J. Maizel, manuscript in preparation)suggests that it is unwarranted at this stage to assume anevolutionary relationship between the Era and RAS pro-teins. Nevertheless, the purified Era clearly binds GTP andGDP (39; Chen et al., in preparation).GTP-binding proteins are currently subjects of consider-

able research interest. Besides the proteins of the RASfamily, whose functions are largely unknown (6), there areother types of GTP-binding proteins whose functions arebetter understood. Members of one class, the G proteins, actas signal transducers, altering the activity of specific en-zymes in response to other protein interactions (19). OtherGTP-binding proteins serve as translation initiation andelongation factors, mediating interactions of mRNAs, AAtRNAs, and ribosomes in protein synthesis (22, 28). Byanalogy, Era could modulate the activity of an enzyme suchas RNase III in response to some unknown regulatorysignals or might be involved somehow in the translationalmachinery. We hope that current studies, using strain HT120and a recently isolated temperature-sensitive era mutation(T. Inada, K. Kawakami, S.-M. Chen, H. E. Takiff, D. L.Court, and Y. Nakamura, unpublished data), will help toelucidate the function of this protein.

ACKNOWLEDGMENTS

We thank R. Crouch for help in beginning our study of RNase IIIand its purification, T. Patterson, Y. Nakamura, J. Bardwell, and M.Davis for many helpful discussions, and N. Costantino for excellenttechnical help. We are grateful to K. Kawakami for providingplasmid pKKF3, to J. Bardwell for providing preliminary data onrnc promoter mapping, and to P. T. Morrison and R. Kolodner forproviding their recO sequence for comparison before publication.We thank S. Adhya for critical reading of the manuscript and LaurelCox for patient and skillful typing.

This research was sponsored in part by the National CancerInstitute under Health and Human Services contract NO1-CO-74101with Bionetics Research, Inc. H.T. was a recipient of a BurroughsWellcome senior research fellowship from 1984 to 1987.

LITERATURE CITED

1. Ahnn, J., P. E. March, H. E. Takiff, and M. Inouye. 1986. AGTP-binding protein of Escherichia coli has homology to yeastRAS proteins. Proc. Natl. Acad. Sci. USA 83:8849-8853.

2. Altuvia, S., H. Locker-Giladi, S. Koby, 0. Ben-Nun, and A. B.Oppenheim. 1987. RNase III stimulates the translation of theclI gene of bacteriophage X. Proc. Natl. Acad. Sci. USA84:6511-6515.

3. Apostolakos, D., and E. A. Birge. 1979. A thermosensitive pdxJmutation affecting vitamin B6 biosynthesis in Escherichia coliK-12. Curr. Microbiol. 2:39-42.

4. Bachmann, B. J. 1972. Pedigrees of some mutant strains ofEschericliia coli. Bacteriol. Rev. 36:525-557.

5. Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12,edition 7. Microbiol. Rev. 47:180-230.

6. Barbacid, M. 1987. ras genes. Annu. Rev. Biochem. 56:779-827.

7. Bear, S. E., D. L. Court, and D. I. Friedman. 1984. Anaccessory role for Escherichia coli integration host factor:characterization of a lambda mutant dependent upon integrationhost factor for DNA packaging. J. Virol. 52:966-972.

8. Beck, C. F., R. Mutzel, J. Barbe, and W. Muller. 1982. A

J. BACTERIOL.

RNase III OPERON 2589

multifunctional gene (tetR) controls TnlO-encoded tetracyclineresistance. J. Bacteriol. 150:633-642.

9. Bertrand, K. P., K. Postle, L. V. Wray, Jr., and W. S. Reznikoff.1984. Construction of a single-copy promoter vector and its usein analysis of regulation of the transposon TnJO tetracyclineresistance determinant. J. Bacteriol. 158:910-919.

10. Bochner, B. R., H. C. Huang, G. L. Schieven, and B. N. Ames.1980. Positive selection for loss of tetracycline resistance. J.Bacteriol. 143:926-933.

11. Bram, R. J., R. A. Young, and J. A. Steitz. 1980. The ribonu-clease III site flanking 23S sequences in the 30S ribosomalprecursor RNA of E. coli. Cell 19:393-401.

12. Crouch, R. J. 1974. Ribonuclease III does not degrade deoxy-ribonucleic acid-ribonucleic acid hybrids. J. Biol. Chem. 249:1314-1316.

13. Date, D., and W. Wickner. 1981. Isolation of the Escherichiacoli leader peptidase gene and effects of leader peptidaseoverproduction in vivo. Proc. Natl. Acad. Sci. USA 78:6106-6110.

14. Dennis, P. P. 1984. Site specific deletions of regulatory se-quences in a ribosomal protein-RNA polymerase operon inEscherichia coli. J. Biol. Chem. 259:3202-3209.

15. Dunn, J. J., and F. W. Studier. 1973. T7 early RNAs andEscherichia coli ribosomal RNAs are cut from large precursorRNAs in vivo by ribonuclease III. Proc. Natl. Acad. Sci. USA70:3296-3300.

16. Dunn, J. J., and F. W. Studier. 1975. Effect of RNase IIIcleavage on translation of bacteriophage T7 messenger RNAs.J. Mol. Biol. 99:487-499.

17. Gegenheimer, P., and D. Apirion. 1981. Processing of procary-otic ribonucleic acid. Microbiol. Rev. 45:502-541.

18. Gegenheimer, P., N. Watson, and D. Apirion. 1977. Multiplepathways for primary processing of ribosomal RNA in Esche-richia coli. J. Biol. Chem. 252:3064-3073.

19. Gilman, A. G. 1987. G proteins: transducers of receptor-gener-ated signals. Annu. Rev. Biochem. 56:615-649.

20. Gitelman, D. R., and D. Apirion. 1980. The synthesis of someproteins is affected in RNA processing mutants of Escherichiacoli. Biochem. Biophys. Res. Commun. 96:1063-1070.

21. Gottesman, M., A. Oppenheim, and D. Court. 1982. Retroregu-lation: control of gene expression from sites distal to the gene.Cell 29:727-728.

22. Grunberg-Manago, M. 1980. Initiation of protein synthesis asseen in 1979, p. 445-477. In G. Chambliss, G. R. Craven, J.Davies, K. Davis, L. Kahan, and M. Nomura (ed.), Ribo-somes-structure, function, and genetics. University ParkPress, Baltimore.

23. Guarneros, G., C. Montanez, T. Hernandez, and D. Court. 1982.Posttranscriptional control of bacteriophage X int gene expres-sion from a site distal to the gene. Proc. Natl. Acad. Sci. USA79:238-242.

24. Hillen, W., K. Schollmeier, and C. Gatz. 1984. Control ofexpression of the TnlO-encoded tetracycline resistance operonII. Interaction of RNA polymerase and TET repressor with thetet operon regulatory region. J. Mol. Biol. 172:185-201.

25. Houlberg, U., B. Hove-Jensen, B. Jochimsen, and P. Nygaard.1983. Identification of the enzymatic reactions encoded by thepurG and purI genes of Escherichia coli. J. Bacteriol. 154:1485-1488.

26. Hughes, J. A., L. R. Brown, and A. J. Ferro. 1987. Nucleotidesequence and analysis of the coliphage T3 S-adenosylmethio-nine hydrolase gene and its surrounding ribonuclease III proc-essing sites. Nucleic Acids Res. 15:717-729.

27. Hyman, H. C., and A. Honigman. 1986. Transcription termina-tion and processing sites in the bacteriophage XPL operon. J.Mol. Biol. 189:131-141.

28. Kaziro, Y. 1978. The role of guanosine 5'-triphosphate inpolypeptide chain elongation. Biochim. Biophys. Acta 505:95-127.

29. Kindler, P., T. U. Keil, and P. H. Hofschneider. 1973. Isolationand characterization of a ribonuclease III deficient mutant ofEscherichia coli. Mol. Gen. Genet. 126:53-69.

30. King, T. C., and D. Schlessinger. 1983. S1 nuclease mapping

analysis of ribosomal RNA processing in wild type and proc-essing deficient Escherichia coli. J. Biol. Chem. 258:12034-12042.

31. King, T. C., R. Sirdeshmukh, and D. Schlessinger. 1984. RNaseIII cleavage is obligate for maturation but not for function ofEscherichia coli pre-23S rRNA. Proc. Natl. Acad. Sci. USA81:185-188.

32. King, T. C., R. Sirdeshmukh, and D. Schlessinger. 1986. Nucle-olytic processing of ribonucleic acid transcripts in procaryotes.Microbiol. Rev. 50:428-451.

33. Kolodner, R., R. A. Fishel, and M. Howard. 1985. Geneticrecombination of bacterial plasmid DNA: effect of RecF path-way mutations on plasmid recombination in Escherichia coli. J.Bacteriol. 163:1060-1066.

34. Lautenberger, J. A., D. Court, and T. S. Rapas. 1983. High-levelexpression in Escherichia coli of the carboxyl-terminal se-quences of the avian myclocytomatosis virus (MC29) v-mycprotein. Gene 23:75-84.

35. Loenen, W. A. M., and F. R. Blattner. 1983. Lambda Charonvectors (Ch 32, 33, 34, and 35) adapted for DNA cloning inrecombination-deficient hosts. Gene 26:171-179.

36. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

37. March, P. E., J. Ahnn, and M. Inouye. 1985. The DNAsequence of the gene (rnc) encoding ribonuclease III of Esche-richia coli. Nucleic Acids Res. 13:4677-4685.

38. March, P. E., and M. Inouye. 1985. Characterization of the lepoperon of Escherichia coli. J. Biol. Chem. 260:7206-7213.

39. March, P. E., C. G. Lerner, J. A. Ahnn, X. Cui, and M. Inouye.1988. The Escherichia coli Ras-like protein (Era) has GTPaseactivity and is essential for cell growth. Oncogene 2:539-544.

40. Maurizi, M. R., P. Trisler, and S. Gottesman. 1985. Insertionalmutagenesis of the lon Gene in Escherichia coli: lon is dispens-able. J. Bacteriol. 164:1124-1135.

41. Miller, J. H. 1972. Experiments in molecular genetics. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

42. Nashimoto, H., A. Miura, H. Saito, and H. Uchida. 1985.Suppressors of temperature-sensitive mutations in a ribosomalprotein gene, rpsL (S12), of Escherichia coli K12. Mol. Gen.Genet. 199:381-387.

43. Nashimoto, H., and H. Uchida. 1985. DNA sequencing of theEscherichia coli ribonuclease III gene and its mutations. Mol.Gen. Genet. 201:25-29.

44. Oppenheim, D. S., and C. Yanofsky. 1980. Translational cou-pling during expression of the tryptophan operon of Escherichiacoli. Genetics 95:785-795.

45. Portier, C., L. Dondon, M. Grunberg-Manago, and P. Regnier.1987. The first step in the functional inactivation of the Esche-richia coli polynucleotide phosphorylase messenger is a ribonu-clease III processing at the 5' end. EMBO J. 6:2165-2170.

46. Rimm, D. L., D. Horness, J. Kucera, and F. R. Blattner. 1980.Construction of coliphage lambda Charon vectors with BamHIcloning sites. Gene 12:301-309.

47. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984.Experiments with gene fusions. Cold Spring Harbor Labora-tory, Cold Spring Harbor, N.Y.

48. Silver, P., and W. Wickner. 1983. Genetic mapping of theEscherichia coli leader (signal) peptidase gene (lep): a newapproach for determining the map position of a cloned gene. J.Bacteriol. 154:569-572.

49. Sirdeshmukh, R., and D. Schlessinger. 1985. Ordered processingof Escherichia coli 23S rRNA in vitro. Nucleic Acids Res.13:5041-5054.

50. Stern, S., T. Powers, L. M. Changchien, and H. F. Noller. 1988.Interaction of ribosomal proteins S5, S6,S11, S12, S18 and S21with 16S rRNA. J. Mol. Biol. 201:683-695.

51. Takata, R., T. Mukae, and K. Hori. 1987. RNA processing byRNase III is involved in the synthesis of Escherichia colipolynucleotide phosphorylase. Mol. Gen. Genet. 209:28-32.

52. Talkad, V., D. Achord, and D. Kennel. 1978. Altered mRNAmetabolism in ribonuclease III-deficient strains of Escherichiacoli. J. Bacteriol. 135:528-541.

VOL. 171, 1989

2590 TAKIFF ET AL.

53. Traub, P., and M. Nomura. 1969. Structure and function ofEscherichia coli ribosomes. VI. Mechanism of assembly of 30 s

ribosomes studied in vitro. J. Mol. Biol. 40:391-413.54. Trjtz, G. J., T. S. Matney, and R. K. Gholson. 1970. Mapping of

the nadB locus adjacent to a previously undescribed purinelocus in Escherichia coli K-12. J. Bacteriol. 102:377-381.

55. Watson, N., and D. Apirion. 1985. Molecular cloning of the gene

for the RNA-processing enzyme RNase IIl of Escherichia coli.Proc. Natl. Acad. Sci. USA 82:849-853.

56. Way, J. C., M. A. Davis, D. Morisato, D. E. Roberts, and N.

Kleckner. 1984. New TnlO derivatives for transposon mutagen-esis and for construction of lacZ operon fusions by transposi-tion. Gene 32:369-379.

57. Wilder, D. A., and H. A. Lozeron. 1979. Differential modes ofprocessing and decay for the major N-dependent RNA tran-script of coliphage X. Virology 99:241-256.

58. Wolfe, P. B., W. Wickner, and J. M. Goodman. 1983. Sequenceof the leader peptidase gene of Escherichia co/i and the orien-tation of leader peptidase in the bacterial envelope. J. Biol.Chem. 258:12073-12080.

J. BACTERIOL.