Copyright 0 1990 by the Genetics Society of America

Physical Analysis of Spontaneous and Mutagen-Induced Mutantsof Escherichia coli K-12 Expressing DNA Exonuclease VI11 Activity

Suresh K. Mahajan,’ Charles C. Chu? David K. Willis,” Ann Templin and AlvinJ. Clark Department of Molecular and CellBiology, University of Calqornia, Berkeley, Calgornia 94720 Manuscript received January25, 1989 Accepted for publication February28, 1990

ABSTRACT We have mapped the extents of two deletion sbcA which result in production of DNA exonuclease VI11 (ExoVIII). One ,sbcA8, deletes about 140 kb of DNA which includes most of the Rac prophage and thetrg gene. Western blot analysis showsthat the protein produced is larger than wild type ExoVIII. The nucleotide sequenceshows that a translational gene fusion has occurred. The N-terminal 294 codons of recE have been deleted and the remaining C-terminal codons have been fused to the N-terminal portion of another reading frame we call SfcA. Analysis of the protein sequence encoded by SfcA shows an 83% similarity with rat and mouse NADP-linked malic enzyme. We discuss the possibility that SfcA is identical tomaeA which encodes NAD-linked malic enzyme from Escherichia coli. Restriction analysis ofa second deletion,sbcA81, by Southern blot technique indicates that about105 kb of DNA have been deleted anda transcriptional gene fusion has occurred between recE and the regulatory region of an E. coli chromosomal gene. We also examined eight other sbc mutations that result in ExoVIII production. Five have no effect on restriction nucleotide fragment sizes detected by complementarity to X rev as probe. These are presumed point mutations. Three seem to produce additional restriction nucleotide fragments complementary to X rev. The possible nature of these sbc mutations is discussed.

UTATIONS suppressing recB21 and/or recC22, on the E. coli genetic map (FOUTSet al. 1983; WILLIS M called sbc mutations, have beendetected in four et al. 1983), and is the defective prophage Rac (LOW different genetic backgrounds Escherichiaof coli K-12. 1973). Rac carries the recE gene (GILLENet al. 1977; Such backgroundscan be nicknamed to avoid writing KAISERand MURRAY1979) and orij (DIAZand PRIT- the complete genotype each time theyare referred to. CHARD 1978; DIAZ,BARNSLEY and PRITCHARD1979). In this case the backgrounds were designated EndoI-, Thus the 1157 background carries a deletion muta- Hfr, Su- and 1157 (TEMPLIN,KUSHNER and CLARK tion which removes recE and hence cannot host recE- 1972). In the first three genetic backgrounds the sbc expressing suppressor mutations. mutationslead to detectable levels ofexonuclease Because the sbc suppressors fell into the recE-ex- VI11 (ExoVIII), an ATP-independent DNA exonucle- pressing and nonexpressing groups, TEMPLIN,KUSH- ase not found in the original strains (BARBOURet al. NER and CLARK(1972) called the first sbcA and the 1970; TEMPLIN,KUSHNER and CLARK1972). The second sbcB. sbcB was found to be the structural gene authors interpreted this to indicate that the structural for DNA exonucleaseI (KUSHNER, NACAISHIand gene for ExoVIII(called re&) had become expressed CLARK,1972, 1974). sbcA, however, has not been so as a result of the suppressor mutation. In the fourth easy to characterize because different types of muta- background (1 157) sbc mutations do not result in recE tions lead to expression of recE, possibly by different expression. KAISER and MURRAY(1979) found that mechanisms. Of five sbcA mutations examined,KAISER several strains with the 1157 background lacked a 27- and MURRAY(1 980) found thattwo, sbcAI and sbcA23, kb elementof DNA present in the otherbackgrounds. showed no change in Rac detectable by the Southern Where it is present, this element is partly complemen- blot method and presumed them to be point muta- tary toand can recombine with phagelambda tions. sbcAl was later found to be cis-acting and dom- (GOTTESMANet al. 1974; KAISER and MURRAY1979). inant to sbcA+ (FOUTSet al. 1983) indicating the pos- The element comprises the region from 29.6 to 30.2 sibility that it alters an operator or promoter forrecE. One mutation, sbcA8, consisted of a large deletion of ’ Present address: and Agriculture Division, Bhabha most of Rac, includingattR and theimmunity repres- Atomic Research Center, Bombay 400 085, India. sor region (KAISERand MURRAY1980). Because the ‘ Present address: Laboratory of Immunology, National Institute of Al- lergy and Infectious Diseases, Bethesda, Maryland 20892. leftward promoter and operator, presumed to control ’ Present address: ARS/USDA Plant Disease Resistance Unit and Depart- ment of Plant Pathology,University of Wisconsin,Madison, Wisconsin recE transcription by analogy with phage X, were also 53706. absent, KAISERand MURRAY(1 980) hypothesized that

Genetics 125: 261-273 (June, 1990) 262 S. K. Mahajan et al. sbcA8 fused recE to a promoter from a gene outside were transferred to nitrocellulose filter paper by the proce- Rac. To explain recessivenessof sbcA8 to sbcA+ (LLOYD dure of SMITHand SUMMERS (1 980).Hybridization of radio- active Xrev probe DNA to the immobilized restriction frag- and BARBOUR 1974), KAISER and MURRAY (1980) ments wasby the method of WAHL,STERN and STARK further hypothesized that an operator anda repressor (1979). gene controlled the promoter to whichrecE was fused. Sequencing the endpoints of sbcA8 Samples of 1.8 pg Furthermore, they hypothesized that sbcA8 had de- chromosomal DNA fromJC5412 (sbcA8) and1.2 pg leted the repressor gene. Two final mutations,sbcA5O pBR322 DNA were digested separately with HindIII, mixed, treated with T4 DNA ligase and used to transform and sbcA51 were associated with EcoRI nuclease diges- JC5519 (recB21recC22) to resistance to ampicillin (50 pg/ tion fragments of Rac which could have been derived ml) and mitomycin (1 pg/ml). Six transformants were ob- from tandemly arranged multiple copiesof Rac, either tained of which one, JC15901, was stably MitCR UVK and onthe chromosome or inthe cytoplasm, or from Ret+. This transformant contained a12.4-kb plasmid having monomeric circular copies of excised prophage (KAI- an 8.1-kb HindIII fragment inserted into the HindIII site of pBR322. The plasmid was named pSKM1. pSKM2 was SER and MURRAY 1980).Additional unusual EcoRI constructed by digesting pSKM1DNA with EcoRI and fragments of Racmade it seem possible that small BamHI, treating the mixture withligase and using it to deletions occurred in someof the copiesof Rac in the transform DH 1 to ampicillin resistance. Subsequent restric- sbcA51 strain (KAISERand MURRAY1980). The phys- tion enzyme analysis of plasmidDNA from fiveof the ical nature of sbcA5O and sbcA51 could not be conclu- transformants revealed one 5.2 kb plasmid which contained a 1.2 kb EcoRI-BamHI fragment inserted between the EcoRI sively determinedpartly because of thenumerous and BamHI sites of pBR322. The strain containing pSKM2 copies of Rac apparently present and partly because was named JC15903. neither sbcA+ strain from which the two mutants were pSKM1 and pSKM2 were used as sources ofDNA to derived was examined. make M 13 strains for sequencing (Figure la). In this paper we extend the studies of KAISERand Most phages were made by digesting plasmid DNA withone or two restriction , adding the mixture to appro- MURRAY(1 980)by determining thatsbcA8 is a trans- priately digested phage DNA, treating the mixture with lational fusion of part of recE with part of another ligase, transfecting JM103 with the mixture and screening geneapproximately 140 kb distant. We also deter- surviving cells in colorless plaques for the desired phage by mine the characteristics of another deletion, sbcA81, restriction digests of DNA minipreps. Techniques used were and hypothesize that this produces a transcriptional those of MESSING (1983). Enzymes used were those whose sites mark the termini of cloned fragments(Figure la). fusion of recA to a promoter about 105 kb distant. Phages JCM 1591 1to JCM 15916 were derived frompSKM 1 Additionally we examine the Rac prophage in mutants and MI3 mp8. Phages JCM15919 to JCM15923 were de- carrying sbcA1, sbcA23, and six other sbcA mutations rived from pSKM2 and M13 mp8.JCM15927 was con- not examinedby KAISERand MURRAY(1 980). Finally structedfrom pSKM2 and M13 mp9. Finally, phages we discuss the appropriateness of labeling as sbcA all JCM 15658 andJCM 15659 were derived respectively from gel-isolated 505 bp and 684 bp BamHI-EcoRI fragments of mutations which suppressrecB and recC mutations and JCM15914 RF DNA and M13 mp8. lead to ExoVIII activity. Template ssDNA generated from the above phages was employed in the chain termination method (SANGERet al. MATERIALS AND METHODS 1977) to obtain the sequence. The extents of each insert actually sequenced are illustrated in Figure 1b. The univer- Bacterialstrains, abbreviations and general genetic sal primer was used except where noted. Three oligonucle- methods: All bacterial strains analyzed in this work are otide primers were also used. prJC22 has the sequence 5' derivatives of Escherichia coli K-12. The derivations and ACGATCGGCAATGCCCAT 3'. It has two mismatches to relevant genotypic markers of most are listed in Table 1. 5' ACCGATCGGAATGCCCAT 3' from coordinates1000 Genotype symbols are those used by BACHMANN(1983). to 983 of the final sequence because it was designed with Question marks indicate that allele and plasmid numbers reference to an earlier inaccurate version of the sequence. have not been stated or that there is a question about the prJC23 has the sequence 5' TTGAGGAGTTAAAATGAA presence of an allele. We also used DH 1, recAI, endAl 3' corresponding to coordinates 172 to 189 andprJC27 has gyrA96 thi-1 hsdR17 supE44 relAI(?) F-X- (MANIATIS, the sequence 5' TCTGCTTTATACTTG 3' correspond- FRITSCHand SAMBROOK1982) and JM103, de (lac-proAB) ing to coordinates 389 to 403 of the final sequence. thi-? supE? rpsL? endA? sbcBI5 hsdR4 [F? traD36 proAB+ lac' de(lacZ)M15], [PI] (MESSING 1983). Tet abbreviates tetra- RESULTS cycline and MitC abbreviates mitomycin C. Superscripts R and S indicate resistance and sensitivity respectively. Trans- Deletion mutations: KAISER and MURRAY(1980) ductionprocedures were described by WILLETTSand showed that ansbcA8 strain JC5412 carried a deletion MOUNT(1 969). Xrev Sam7 cI857 (GOTTESMANet al. 1974) was obtained from M. GOTTESMANas lysogen N2262. of most of Rac. We have confirmed their observation Isolation of bacterial DNA and detection of restriction and have also determined the amount of Rac deleted fragments complementary to hrev DNA: Isolation of bac- by sbcA81 in strain JC9625. Figure 2a shows a South- terial chromosomal DNA, in vitro labeling of Xrev with "P, ern blot of DNA from each mutant and itswild type conditions for restriction endonuclease digestion of chro- ancestordigested with EcoRI. Southern blots of mosomal DNA, and methods forthe separation of fragments of agarose gel electrophoresis were as previously described HindIII-cutDNA and BamHI-cut DNA are shown (GILLEN,WILLIS and CLARK 1981).Restriction fragments respectively in Figure 3a and Figure 3b. Xrev DNA Analysis of ExoVIII+ Mutants 263

TABLE 1 Bacterial strains

Relevant genotype Genetic Strain Sex background sbcA RAC Other Reference AB1 157 F- 1157 Deletion - BACHMANN(1972) JC55 19 F- 1157 Deletion - WILLETTSand CLARK(1 969) JD8675 F- 1157 23 + GILLEN,WILLIS and CLARK(198 1) JC8679 F- 1157 23 + GILLEN,WILLIS and CLARK(198 1) JC4695 F- su- + + TEMPLIN,KUSHNER and CLARK(1972) JC7608 F- su- 19 + TEMPLIN,KUSHNER and CLARK(1 972) JC7609 F- su- 20 + TEMPLIN,KUSHNER and CLARK(1 972) JC9625 F- su- 81 de-8 1‘ GILLEN(1974) KL16 Hfr(P045) Hfr + + BACHMANN(1 972) JC5029 Hfr(P045) Hfr + + BACHMANN(1972) JC54 12 Hfr(P045) Hfr 8 de-8‘ BACHMANN(1972) JC5491 Hfr(P045) Hfr + + WILLETTSand CLARK(1 969) JC7659 Hfr(P045) H fr 21 + TEMPLIN,KUSHNER and CLARK(1 972) JC7660 Hfr(P045) Hfr 22 + TEMPLIN,KUSHNER and CLARK (1972) JC7661 Hfr(P045) Hfr 23 + TEMPLIN,KUSHNER and CLARK(1 972) JC5172 F- EndoI- 2 + BARBOURet al. (1 970) JC5 174 F- EndoI- I + BARBOURet al. (1970) JC5 176 F- EndoI- 6 + BARBOURet al. (1970) JC6720 F- EndoI- + + CAPALW-KIMBALLand BARBOUR(1 97 1) IC6722 F- EndoI- + + CAPALDO-KIMBALLand BARBOUR (1 97 1)

a The genotype of AB1 157 is thr-1 ara-14 leuB6 proA2 lacy1 tsx-33 supE44 galK2his-4 rpsL31 (Sd)mtl-I xyl-5 argE3 thi-I. ‘Same as AB1 157 except recB2l recC22. ‘ Same asJC5519 except JC8675 is his+ recE+ and JC8679 is his-328 recE+. The mutant genotypes of JC4695 and its derivatives are his-318 leu-307 trpEY82Y rpsL321, thi-1 recB2l sup+. The letters “de” are used in place of an upper case Greek delta to indicate deletion The hyphen stands for a parenthetical expression indicating the length of the deletion. For de-81 the expression is “(rcC-trg)”and for de-8 the expression is “(recE-sJcA).” /The mutant genotype of KL16 is thi-I, rel-I, X-. Same as KL16 except thr-300, ilv-318, spc-300. ’ Same as JC5029 except recB21. ‘ Same as JC5029 except recB2l recC22. Genotvae is F- lambda? Rac+ deo-27 lacZ4 subE44 pal-44 endAl recB21 thi-l and the $a allele from E. coli C [+Xs] (CAPALDO-KIMBALL ,. 3” and BARBOUR197 1; B. J. BACHMANN,personal communication). ‘Same as JC6722 except recC22 (In addition to recBZI). was used as a probe. This phage carries about 8 kb map of the two deletion mutant chromosomes (Figure derived from the ends of Rac prophage as shown by 4, b and c). An approximate location for the other the shaded areas in the restriction map of Rac and endpoint of each deletion was deduced by examining flanking chromosome regions (Figure 4a). Comparing an EcoRI HindIII PstIrestriction map of 10 min the restriction map with the bands visible in the DNA (about 400 kb) of the chromosome from trp to man of the mutants shows that those from the middle and (BOUCH~1982) for the appropriate spacing of EcoRI right hand part of the Rac prophage are missing, e.g., and HindIII restriction sites. Southern blots of PstI the 13.6- and 4.3-kb EcoRI fragments, the 4.0 and digested DNAs from the mutant andwild type strains 26.5-kb HindIII fragments and the 13.0- and 7.1-kb were essentially consistent with the restriction maps BamHI fragments. By contrastbands from the left as drawn in Figure4 (data not shown). sbcA8 was hand part of Rac prophage are retained, e.g., the 2.8- deduced to be a deletion of about 140 kb (Figure 4b) kb EcoRI and 3.6-kb BamHI fragments. and sbcA81 a deletion of about 105 kb (Figure 4c). The approximate location of one endpoint of each These deductions lead to two predictionsboth of deletion of Raccan be deduced from theband pattern which have been experimentally confirmed. to lie between an EcoRI and a HindIII cleavage site in First of all both deletions should have removedtrg, the left hand portion of the phage (bold arrows in a gene lying at 247 kb (BOUCH~et al. 1982) from the Figure 4a). In JC5412 DNA the 2.8-kb EcoRI frag- zero point on the mapof BOUCH~(1982). To test this ment is retainedbut the 7.6-kb HindIIIfragment we transduced thetwo deletion strains with P 1 grown changes to an8.1-kb fragment (Figures2a and 3a). In on strain RP3342, a trg-2::TnIO mutant (HARAYAMA, JC9625 DNA the 2.8-kb EcoRI fragment is retained PALVA and HAZELBAUER 1979).Table 2 shows that (Figure 2a) but the 7.6 kb HindIII fragment changes neither deletion strain produced TetR transductants to the mobility of a 16.5 kb fragment (Figure 3a). while their point sbcA mutant relatives (see below) did. These changes have been incorporatedin a restriction This indicates that homologous DNA for approxi- 264 S. K. Mahajan et al.

rq5 B 41 I I I I 1 c?

a) Phage Strains

15912 15914 4 I 15911 I15916 15915 - 15919 + 15921

9 15922 15920

15659 15923 15925 .. 15658 - --15927 b) Sequence Obtained

pdC23 I-” -+ prJC27 b+- w I-+ FIGURE1.-M 13 phage strains and their use in determining the nucleotide sequence of sbcA8. A 2892-bp fragment bounded by EcoRl sites from pSKMl is shown in the upper part of the figure for reference. Relevant sites and deletion junction site are marked. A shaded box indicates pBR322 DNA. The line is E. coli DNA and the empty box is Rac DNA. Coordinates refer to the first base of the Hind111 recognition site as coordinate 1. Numbers in parentheses refer to coordinate system for Rac prophage sequence containing rrcE obtained by CHU,TEMPLIN and CLARK(1989). (a) MI3 phage strains: Arrows indicate extent of fragment subcloned into M13 vectors. The arrowhead indicates orientation of insertion into M13 vectors, where the arrowhead indicates the end of the fragment distal to the universal primer site. Above or below the arrows are phage numbers (minus the JCM prefix) denoting the phage containing this portion of DNA. JCM15923 was created by inserting an approximately 1.3-kb ScaI fragment from pSKM2 containing about 0.5 kb of pRR322 DNA into appropriately digested M 13mp8 RF DNA. (b) Nucleotide sequence obtained from phage strains. Arrows indicate extents sequenced in a 5’ to 3‘ direction, where the arrowhead symbolized the sequence most distal from the primer site. Arrows labeled above by prJC22, prJC23, or prJC27 indicate that these sequences were obtained by utilizing these primers. Otherwise, the universal primer was utilized. mately 1 min (40 kb) on both sides of trg is missing The second prediction is that plasmids carrying fromthe deletion mutants. We also comparedthe fragments deduced to be deleted should hybridizenot transduction recipient-abilities of the sbcA8 and with restriction nuclease-digested DNA from the mu- sbcA23 strains and found thatthey inherited iZv+ from tants. Figure 5 shows that this prediction was verified. the same P1 lysate at about equal frequencies (Table Three plasmids were used which contained DNA hy- 2). bridizable to restrictionfragments spanning more Qin 17.51 ~17.5~ .- ”.~,..%.. ;: - Qin 15.2- -15.2 Rac 13.6./- \13.6 Rae 8.1- -8.1 - -7.2 Qsr’ 6.5- l6.5 Qsr’5.8 f Rac 4.55~ r4.551 -4.3 - Rac 4.3- -4.0 Qsr’

ROC 2.8- -2.8 - 2.6-

2.2-

1.85- FIGURE2.-Autoradiographs showing EcoRI generated fragments of chromosomal DNA that contain homology to y2P-labeledXrev DNA. lhe strains analyed itre listed. The fragment sizes are given in kb. The EcoRI fragments which belong to Rac are 13.6 kb, 8.1 kb, 4.5.5 kb, 4.3 kb, and 2.8 kb. Fragments which belong to two other prophages (KAISER 1980; &PION, KAISER and DAMRLY-CHAUDIERE1983) are as follows: Qsr”13.8 kb (not indicated). 6.5 kb (Hfr and Su- genetic backgrounds), 5.8 kb (Endol- background), 4.0 kb (1 157 background), and Qin-I 5.2 kb, 17.5 kb. than 60 kb of the chromosome in the region con- of recE occupies the left hand 1982 nucleotidesof this cerned. sbcA81 DNA has lost part of the DNA corre- fragment, as the maps in Figure 4 are drawn. Because sponding to a 13.4-kb EcoRI and a 6.9-kb HindIII sbcA8 reduces the size of the EcoRI fragment to 1.85 fragment hybridizable to pLN54. Thus the deletion kb, one endpoint must lie within the coding portion endpoint is between the HindIII sites at 267.47 and of recE. Because no HindIII orPstI sites can lie in this 274.37 on the map of BOUCH~: (1982).sbcA8 DNA EcoRI fragment, given the sizes of HindIII and PstI has lost not only the fragments lost by sbcA81 DNA fragments observed (Figure 3a, and data not shown), but several others as well, retaining only some frag- and because the BamHI site approximately 540 nucle- ments hybridizable to pBS2O. This puts the deletion otidesfrom the EcoRI site is notdeleted, we can endpoint between 299.20 and 310.05 on the map of further refine the deduction to predict that the end- BOUCH~:(1 982). point lies between 167.12 and 168.43, using the map Gene fusions deduced for both sbcA8 and sbcA8l: coordinates of BOUCH~: (1982,see the black box in The nucleotide sequencehas been determined for the Figure 4b). These are the equivalents of codons 41 2.45-kb EcoRI-Hind111 fragment of Rac within which and 441 of recE, respectively, indicating that sbcA8 is lie the endpoints of sbcA8 and sbcA8I (CHU,TEMPLIN likely to be a translational gene fusion. and CLARK 1989).The N-terminal half (approximate) The same type of calculation was done for sbcA81. 266 S. K. Mahajan et al.

Su- Hfr nn

Hfr 1157 su- EndoI- I I

Q, (u (D 0 c ln J V Rac 28.,?26.57 Y 7 13.0-

7.1- 13.0- Qsr' 1O.d

Roc Qsr' 7.1 4.2-

Qsr' 4.8- 3.6- WII

Rac 4.0-

Qsr' 3.1-

Rac 2.5-

FIGURE%-(a) Composite autoradiograph showing HindIII generated chromosomal DNA fragments that are homologous to ""P-labeled Xrev DNA. Fragment sizes are given in kb and the strains analyzed are listed at the top. The prophage fragment sizes are as follows: Rac- 26.5 kb, 7.6 kb, 4.0 kb, 2.5 kb: Qin-28.0 kb: Qsr'-10.3 kb (Hfr and Su- backgrounds), 9.6 kb (Endot- background, not indicated), 7.1 kb (1 157 background), 4.8 kb and 3.1 kb. (b) Autoradiograph showing Xrev homologous BamHI restriction fragments. Strains analyzed are listed at the topand the sizes of Rac prophage restriction fragments are indicated in kb. In this case the rangewithin which the endpointsmust and UV and mitomycin C sensitivity of recB and recC fall is quite small because HindIII and EcoRI sites, mutations. This not only showed that pSKM1 con- which must be deleted,lie within 0.15 kbof a PstI site tained a functional (presumed hybrid) recE gene but that must be retained (BOWCHI?1982). Since the new that expression of this gene on themulticopy plasmid EcoRI fragment formed by sbcA81 was measured at was not prevented by the chromosomal repressor hy- 7.2 kb (Figure 2a) and since the distance from the pothesized by KAISER and MURRAY(1980). The se- retained PstI to EcoRI site is 4.8 kb (BOUCHI?1982), quences of two portions of this fragment were deter- 2.4 2 0.15kb must becontributed by the EcoRI- mined by the strategyoutlined in MATERIALS AND HindIII fragment of Rac. This is more than enough METHODS. One portion was the 1.85-kb EcoRI frag- to include the entire amino terminal coding portion ment containing the endpoints of sbcA8 (Figure 4b); of recE (CHU,TEMPLIN and CLARK 1989) andleads to the otherwas the approximately 1.1-kb EcoRI-Hind111 the deduction that sbcA81 is a transcriptional gene fragment between the bold arrows in Figure 4b. The fusion. combined sequenceis presented in Figure 6. The first Nucleotide sequenceof sbcA8: The 8.1-kb HindIII 1765bp are not contained in the portion ofRac fragment containing endpoints of sbcA8 (Figure 4b) previously sequenced(CHU, TEMPLINand CLARK was cloned as described in MATERIALS AND METHODS, 1989). The remainder of the sequence is identical to using pBR322as vector. The resulting plasmid the recE sequence beginning at recE codon 294. This pSKM 1 was able to suppress recombinationdeficiency is within the range predicted in the previous section. Analysis of ExoVIIIf Mutants 267

(a) sbcA+ Prophage Roc I N ? 0 cu0

"- "- "- "- 6.45 15.6 1.34.6 PstI ,1,8, 1n.o.-

8.1 1 , 13.6 4.55 , 4.3 EcoRI 1 2.8 ' 7.6 t2.5, 4.0 , 26.5 Hind I11 I i I 4.2 13.0 7.1 13.0 3.6 4.2 Barn H I I

Portions of ROC in Xrev recE Regionwithin which deletionend points arelocated 6.45 ,1,8, 8.6 Pst I I Restrictionsites 8.1 , 2.81.85 P PstI EcoRI 1 E EcoRI I 8.1 4s' HindI11 H HindIII 3.6 Barn H I -

16.5 HindIII k I 3.6 Barn H I c"( FIGURE4,"Physical map of chromosomal DNA near the Rac prophage showing restriction sites for EcoRI, HindIII, and PstI. Coordinates are taken from the map of BOUCH~(1982). Coordinates of the Rac prophage are deduced from the work of WILLISet al. (1983). Shaded boxes indicate the extent of Rac DNA found in Xrev. Solid boxes indicate an areaof uncertainty for the deletion end points. Lines underneath the maps indicate the length of restriction fragments detected in the strains carrying sbcA alleles; numbers are the sizes measured in kb. Only those restriction sites detectable by Southern blot using Xrev as probe are shown; further sites distal to Rac (e.g., E at 314.6 in b and E at 280.97 in c) would be undetectable. Location of the BarnHI sites is taken from the work of WILLISet al. (1983). (a) KL16 (&A+) and JC4695 (sbcA+), (b) JC5412(sbcA8) and (c) JC9625 (sbcA8I). n.a. means not available. Arrows are explained in text.

Fused to the recE reading frame is an ORF of 447 positions with the Shine-Dalgarnosequence (SHINE codons. Within the first 67 codons of this ORF are 3 and DALGARNO 1974) henceand could be a ribosome ATG (Met) and 2 GTG (Val) codons at which trans- binding site. We call the ORF from this ATG codon lation could potentially initiate (Figure 6). We think sfA. Translation of sfA would substitute 429 codons the most likely initiation codon is the ATG at nucle- for 293 codons of recE. The differential 136 codons otides479-81 because the sequence5'AAGAGTS' would be expected to add about 14.5 kDa in mass to terminates7 bp upstream, matches in 4 out of 6 recE protein.A Western blot of proteinextracted 268 S. K. Mahajan et al.

TABLE 2 linked malic enzyme, must be considered provisional Absence of TetR transductantsimplies inability by deletion sbcA partly because we do not know how large a protein mutants to inherit trg-2::TnlUfollowing PI transduction the wild type allele encodes. If it is really maeA, then only about eighteen codons would have been deleted Hfr backgroundb Su- background‘ Selected and the amino acid composition would notbe ex- sbcA8 sbcA23 sbcA81 sbcA20property”sbcA81 sbcA23 sbcA8 pected to change very much from that shown in Table TetR <0.4 51 <0.4 10 3. If, however, it is really maeB then about 1 13 codons IlV+ 140 47 N Dd ND would have been deleted. Inthis case the low amounts Numbers of transductants are measured per lofi input phage of five amino acids might be redressed by the missing plaque-forming units. codons but the already large amountsof tyrosine and A P1 lysate was produced on strain RP3342. TetR(tetracycline resistance) implies inheritance of trg-2::TnlO and llv+ (isoleucine tryptophanmight increase. Consequently we favor and valine independence) implies inheritance of the wild-type allele maeA as the true identity of $A. of ilu-?I8. Point mutations: Those recE-expressing mutants * Strains JC5412 and JC7661 wererecipients. ‘ Strains JC9625 and JC7609were recipients. whichshow no differencesfrom their ancestors in Not done. DNA fragments complementary to Xrev are likely to carry sbcA point mutations, although very small dele- from sbcA8 mutant JC5412 shows a protein of about tions or insertions (<50 bp) or internal Rac inversions 150 kDa compared to proteinfrom an sbcA point would not produce detectable fragment size changes mutant of about140 kDa (data not shown). This and cannot be ruled out. Two Su- background mu- confirms our finding of an unequal substitution and tants (JC7608[sbcA19], JC7609[sbcA20]) andthree is consistent with the putative initiation codon at nu- EndoI-background mutants (JC5 172[sbcA2], cleotides 479-81 butdoes not rule out translation JC5174[sbcAl] and JC5176[sbcA6]) fall into this cate- initiation at one of the other potential initiation co- gory and show all of the Rac restriction fragments dons. containing homology to Xrev (Figure 2, a, b and c). Possible identity of SfcA: We screened 4206 trans- The results with JC5 I74 are identical to those ob- lated nucleotide sequences of GENBANK (December tained by KAISER and MURRAY(1 980). All five mu- 1987) using the tfastp program (with parameter tants in this group were obtained after mutagenesis “ktup” set at two) to detect similarities to sfcA. Two with NTG or showed very high optimized scores indicating a high (N-methyl-N’-nitro-N-nitrosoguanidine) EMS (ethyl methanesulfonate), which is consistent degree of similarity. Both were NADP-linked malic enzymes (EC 1.1.1.40), one from mouse liver (MAG- with our hypothesis that they carry sbcA point muta- tions. NUSON et ai. 1986), the other from rat liver (BAGCHI et al. 1987). They showed 43% identical residues to Othermutants: KAISER and MURRAY(1980) ob- the 429 aminoacids encoded by sfcA (data notshown). served an extra 2.45-kb EcoRI band with DNA from If conservative amino acid changes were accepted, the sbcA23 mutant JC7661 and hypothesized that it was sequences were 83% similar (data not shown). Only derived by juxtaposition of attL and attR elements of four small gaps in the pairs of sequences were neces- the Rac prophage. We have also examined DNA from sary to achieve this degree of similarity. JC7661 and find an extra band which we measure to In E.coli two malic enzymes have been reported: be 2.6 kb (Figure 2a). In additionwe analyzed JC766 1 NADP-linked malic enzyme (EC 1.1.1.40) andNAD- andJC5491 DNAs withHind111 andfound that linked malic enzyme (EC 1.1.1.38) (KATSUKI et al. JC7661 DNA contains an extra fragment with a size 1967). These enzymes have beenpurified (SPINA, of approximately 19 kb (Figure 3a). This fragment BRIGHTand ROSENBLOOM,1970; YAMAGUCHI,TO- could also have been formed by juxtaposition of the KUSHIGE and KATSUKI 1973)and their amino acid two ends of the prophage.KAISER and MURRAY(1 980) compositions determined. Table 3 shows a compari- pointed out thatDNA from JC8679,a transconjugant son of SfcA product with the two E. coli malic enzymes. inheriting Rac prophage and sbcA23 from JC7661, Lower amounts of lysine, histidine, proline, alanine did not show the extra 2.6-kb EcoRI band. We have and valine and higher amounts of tyrosine and tryp- confirmed that (Figure 2a) and have additionally tophan make sfcA protein more similar to the NAD- shown the absence of the 19-kb Hind111 band (Figure linked enzyme. E. coli mutants lacking each of the 3a).JC8679, however, is a second generation enzymes have been detected and characterized phe- transconjugant. It was derived by a second conjuga- notypically (HANSEN andJUNI 1975) butmapping the tion eventin which his-?28 was transmitted to JC8675, mutant genes was not reported. The genes encoding the original Rac+ sbcA23 transconjugant(GILLEN, NAD-linked and NADP-linked malic enzymes are WILLISand CLARK 1981). We have found that DNA known as maeA and maeB, respectively (B. BACHMANN, from JC8675 cleaved with EcoRI or Hind111 also lacks personal communication). We hypothesize that SfcA is the 2.6-kb and 19-kb bands (data notshown). identical to maeA. JC7661, whose DNA contains theextra 2.6-kb Identification of sfcA as maeA, the gene for NAD- EcoRI and 19-kb Hind111 bands, was derived along Analysis of ExoVIII+ Mutants 269

W X wx;wwxwx X wswxwxw

t (0 lu (b) Plasmidprobes I pL N54 pLN49 pBSeO pLN49 pLN54

13.4 4.9 24.4 ,4.6 I 6.1 (c) EcoRIfragments detected (kb) I I I I I sbcA+ + + + ++ sbcA8 - - - ++ sbcA8I - + + ++

(d) Hind111fragments I 23.0 , 6.9 , 8.8 I , 13.6 , 12.0 I 8.1 detected(kb) 1 I 11 I I sbcA+ ? + + + + + sbcA8 ? - - - - + sbcA81 ? - + + + + (e) Endpoint range inferred: ~bcA8 299.20-310.05 sbcA8I 267.47-274.37 FIGURE5."Summary of data on the endpoints of deletion mutations sbcA8 and sbcA8l. (a) Diagram of the region of the chromosome tested. This region includes the right hand portions of the maps in Figure 4, b and c. Coordinates of restriction sites are taken from BOUCH~ (1982). H stands for Hind111 and E stands for EcoRI. More restriction sites are shown here than in corresponding places in Figure 4, b and c. This is intended to be a complete restriction map while Figure 4, b and c are probe-dependent Southern blot fragment maps. (b) The names of the plasmid probes are listed under rectangles indicating the chromosomal region carried. Plasmids pLN54 and pLN49 are derivatives of pBR325 (BOLIVAR1978) and were provided by J. P. BOUCH~in strains LNll75 and LN1170, respectively. pBS2O is also derived from pBR325 and was supplied by J. P. BOUCHF in the C600 background. (c) and (d) Sizes (in kb) of the restriction fragments detected by the plasmid probes are listed above a line which showsthe extent of the fragment onthe map in (a). + and - indicate (respectively) that the fragment was or was not detected in restriction digests of DNA from the strain listed at the left. ? indicates that the expected fragment was observed only indistinctly with wild type DNA so that we cannot be certain of its absence in mutant DNA. DNA was isolated from JC5029 (sbcA+),JC5412 (sbcAB),JC4695 (sbcA+)and JC9625 (sbcA8l). (e) Coordinates indicate the restriction sites shown in(a) between which the deletion endpoints must lie based on the evidence in (c) and (d). with five other strains by mutagen treatment of Hfr They thought thefirst most likely while we prefer the strain JC549 1(TEMPLIN, KUSHNER and CLARK1972). second and third forreasons presented in the DISCUS- We have tested EcoRI cleaved DNA from JC7660, SION. At this point, however, we need to point out another mutagen-induced mutant and find it has the that JC5491 DNA does not show these extra bands same extra 2.6-kb EcoRI band (Figure 2a). The same (Figures 2a and 3a). Thus we have no evidence that extra EcoRI band is present in DNA from JC7659, an ancestral sbcA+ strain JC5491 has a tandem array of sbcA mutant of JC5491 selected for MitCR after no chromosomal Rac prophages, in contradiction to the mutagenic treatment (Figure 2a); however in this case reference to our data made by KAISERand MURRAY the 2.6-kb band is weak and a much stronger extra (1980, see p. 560 second column). We think it is likely 2.2-kb band is present (Figure 2a). Hind111 digests of that the extra bands in JC7659, JC7660 and JC7661 JC7659 DNA likewise show a relatively weak 19-kb are in some way related to the sbcA mutations each and a new strong 13-kb band (Figure 3a). has suffered. The presence of the extra bands in JC7661 DNA was interpreted by KAISER and MURRAY(1 980) to DISCUSSION indicate either of three conditions of Rac DNA: (1) Expression of ExoVIII activity is accomplished by a tandemly arranged multiple copies in thechromo- variety of mutations. Thinking that a simple repres- some, (2) circular or linearmultimers in the cyto- sion system might be operating (BARBOURet al. 1970; plasm, and (3) circular monomers in the cytoplasm. TEMPLIN,KUSHNER and CLARK (1972))called these 270 S. K. Mahajan et al. A 1 AAGCTTCTGATAACGACATC CTCAGTGATA TCTACCAGCA AACGATCAAT TATGTGGTCA GTGGCCAGCA CCCTACGCTT

81 TAAGGTGCTATGCTTGATCG GCAACCTAAT TTAGGGGTTT AGCACGTGTT TCTTCGCTAC GGCGATGTTG TCCTTAAAAC

161 TAGCTACAGGTCGAACCGTCAGGCACGTCATATTCTTGGA CTGGACCATA AAATTTCTAA

241 CCAGCGCAAAATAGTTACCG AAGGTGACAA ATCCAGCGTA GTAAATAACC CAACCGGCAG AAAACGCCCC GCTGAAAAGT

321 AATTCATAACCATCAGTCCT CAATGACGAT T ARACACCATTCC CTGC ~TCCTCT!CT TTGTTTTT-

*** SerLys Asp Asp Lys Ser Fro Fro Gly mt Asp Ile GlnLys Arg Val 400 CLTGAGGCCGACGCCCTGGC GG TAA AGC AAA GACGAT AM AGC CCCCCA GGG ATGGAT ATT CAA AAA AGA GTG

SerAsp l6t Glu Fro LysThr Lys Lys Gln Arg Ser Leu Tyr Ile Fro TyrAla Gly Fro ValLeu Leu 473 AGTGAC ATG GAA CCA AAA ACA AAA AAA CAG CGTTCG CTT TAT ATC CCT TAC GCT GGC CCTGTA CTG CTG

GluPhe Pro Leu Leu Asn Lys Gly Ser Ala Fhe Ser Met GluGlu Arg Arg Asn Phe Asn Leu Leu Gly 542 GAA TTT CCG TTGTTG AAT AAA GGCAGT GCC TTC AGC ATG GAA GAA CGCCGT AAC TTC AAC CTGCTG GGG

LeuLeu Pro Glu W1 ValGlu Thr Ile GluGlu Gln Ala Glu Arg Ala Trp Ile GlnTyr Gln Gly Fhe 611TTA CTG CCG GAA GTGGTC GAA ACCATC GAA GAA CAA GCG GAA CGAGCA TGG ATC CAG TAT CAG GGA TTC

LysThr Glu Ile AspLys His Ile TyrLeu Arg Asn Ile Gln Asp Thr Asn Glu Thr Leu Fhe Tyr Arg 680 AAA ACC GAA ATC GAC AAA CAC ATCTAC CTG CGT AAC ATC CAG GAC ACT AAC GAA ACC CTCTTC TAC CGT

LeuVal Asn Asn His LeuAsp Glu Met Met ProVal Ile TyrThr Pro Thr Val Gly Ala Ala Cys Glu 749 CTGGTA AAC AAT CATCTT GAT GAG ATGATG CCT GTT ATT TAT ACCCCA ACC GTC GGC GCA GCCTGT GAG

ArgPhe Ser Glu Ile Tyr Arg Arg Ser Arg Gly Val Phe Ile SerTyr Gln Asn Arg His Asn Met Asp 818 CGTTTT TCT GAG ATCTAC CGC CGT TCA CGC GGC GTG TTT ATCTCT TAC CAG AAC CGG CAC AATATG GAC

Asp Ile LeuGln Asn Val Fro Asn His Asn Ile LysVal Ile ValVal Thr Asp Gly Glu Arg Ile Leu 887 GATATT CTG CAA AAC GTGCCG AAC CATAAT ATT AAA GTGATT GTG GTG ACT GAC GGT GAA CGC ATT CTG

GlyLeu Gly Asp Gln Gly Ile Gly Gly Met Gly Ile Pro Ile GlyLys Leu Ser Leu Tyr Thr Ala Cys 956 GGG CTT GGTGAC CAG GGC ATC GGC GGG BTI; GGCATT AAA CTGTCG CTC TAT ACC GCC TGT

GlyGly Ile Ser Pro AlaTyr Thr Leu Fro ValVal Leu Asp Val Gly Thr Asn Asn Gln Gln LeuLeu 1025 GGCGGC ATC AGC CCG GCG TAT ACC CTTCCG GTGGTG CTG GAT GTC GGA ACG AAC AAC CAA CAG CTGCTT

AsnAsp Pro Leu Tyr Met Gly TrpArg Asn Pro ArgIle Thr Asp Asp Glu Tyr Tyr Glu Fhe Val Asp 1094 AAC GATCCG CTG TAT ATG GGCTGG CGT AAT CCG CGTATC ACT GAC GAC GAA TACTAT GAA TTCGTT GAT

GluPhe Ile GlnAla Val Lys Gln Arg Trp Fro AspVal Leu Leu Gln Phe Glu Asp Phe Ala Gln Lys 1163 GAA TTT ATC CAG GCT GTG AAA CAACGC TGG CCA GAC GTG CTG TTG CAG TTT GAAGAC TTTGCT CAA AAA

AsnAla Met Pro LeuLeu Asn Arg Tyr Arg Asn Glu Ile Cys SerPhe Asn Asp Asp Ile Gln Gly Thr 1232 AAT GCG ATG CCGTTA CTT AAC CGCTAT CGC AAT GAA ATTTGT TCT TTT AAC GAT GAC ATT CAG GGC ACT

AlaAla Val Thr Val Gly Thr Leu Ile AlaAla Ser Arg Ala Ala Gly Gly Gln Leu Ser Glu Lys Lys 1301 GCG GCG GTAACA GTC GGC ACA CTG ATCGCA GCA AGC CGC GCG GCAGGT GGT CAG TTA AGC GAGAAA AAA

Ile ValPhe Leu Gly Ala Gly Ser Ala Gly Cys Gly Ile AlaGlu Met Ile IleSer Gln Thr Gln Arg 1370 ATCGTC TTC CTT GGC GCA GGT TCA GCG GGA TGC GGC ATT GCC GAA ATGATC ATCTCC CAG ACC CAG CGC

GluGly Leu Ser Glu Glu Ala Ala Arg Gln Lys Val Phe Met ValAsp Arg Fhe Gly Leu Leu Thr Asp 1439 GAA GGA TTA AGC GAG GAA GCG GCG CGG CAG AAA GTCTTT ATG GTCGAT CGC TTT GGC TTGCTG ACT GAC

FIGURE6A Analysis of ExoVIII+ Mutants 27 1

B Lys Met ProAsn Leu Leu Pro PheGln Thr Ly5 Leu Val Gln Lys Arg Glu Asn Leu Ser Asp Trp Asp 1508 AAG ATG CCGAAC CTG CTG CCTTTC CAGACC AAA CTG GTG CAGAAG CGC GAA AAC CTC AGTGAC TGG GAT

ThrAsp Ser Asp Val Leu Ser Leu Leu Asp Val Val Arg Asn Val Lys Pro Asp Ile Leu Ile Gly Val 1577 ACC GACAGC GAT GTG CTG TCA CTG CTG GAT GTGGTG CGC AAT GTA AAA CCA GAT ATT CTG ATT GGC GTC

SerGly Gln Thr Gly Leu Phe Thr Glu Glu Ile Ile Arg Glu Met His Lys His Cys ProArg Pro Ile 1646 TCA GGACAG ACC GGG CTG TTT ACGGAA GAG ATC ATC CGT GAG ATG CAT AAA CAC TGT CCG CGT CCG ATC

Val Met ProLeu Ser Asn Pro Thr Ser Arg Val Glu Ala Thr Pro Gln Asp Ile Ile Ala Glu Asn Lys 1715 GTG ATG CCG CTG TCT AAC CCG ACG TCA CGC GTGGAA GCC ACA CCG CAG GAC IATT ATC GCT GAAAAT AAA

ProPro Phe Ser Val Phe Arg Asp Lys PheIle Thr Met ProGly Gly Leu Asp Tyr Ser Arg Ala Ile 1784 CCG CCC TTTTCT GTT TTC CGC GAC AAA TTC ATC ACC ATG CCT GGCGGG CTG GAT TATTCC CGC GCC ATC

ValVal Ala Ser Val Lys Glu Ala Pro Ile Gly Ile GluVal Ile Pro Ala His ValThr Glu Tyr Leu 1853 GTG GTT GCG TCC GTA AAA GAA GCA CCA ATT GGG ATC GAG GTC ATCCCC GCG CAC GTC ACT GAA TAT CTG FIGURE6.-DNA sequence of the region containing sbcA8. The entire sequence of the 1.07-kb HzndlII-EcoRI fragment followed by a partial sequence of the 1.85-kb EcoRI fragment up to the deletion junction plus 156 bp of recE sequence is presented 5' and 3'. The first base of the Hind111 recognition site was arbitrarily designated as coordinate 1, resulting inEcoRl sites at coordinates 1151 and 2861. Nucleotide coordinates are provided on the sides of the sequence. The ORF containing the +A-RecE fusion protein is shown translated above the DNA sequence. Five potential translational start sites are illustrated by bold lettering of Met or Val. *** indicates a termination codon. The fusion site is represented as a vertical line between sJcA (up tonucleotide 1765) and recE sequences (nucleotide 1766 andbeyond). A dyad symmetry is represented by bold lettering of the DNA sequence. The locations of primers prJC22, prJC23, and prJC27 are illustrated by underlining of the DNA sequence prJC22 corresponds with two mismatches to the sequence complementary to that underlined from coordinates 1000-983. prJC23 and prJC27 correspond to sequences underlined from 172-189 and 389-403, respectively.

TABLE 3 extended these results by showing that sbcA1, sbcA2, Comparison of the translation product from the partial sbcA6, sbcA19 and sbcA20 fall into the point mutant sequence of SfcA with E. coli malic enzymes class, and by showing that sbcA8,sbcA21, sbcA22, sbcA23 and sbcA8I affect the structure of the Rac Number of residues per subunitPercent of total residues prophage. The main purpose of this discussion is to NADP- NAD- NADP- NAD- NADP- NAD- NADP- examine the evidence to see whether or not thecom- Amino linked linkedlinked Amino SfcA linkedlinked SfcA acid enzymeacidenzyme product enzyme enzyme product mon gene symbol sbcA for all these mutations is ap- LYS 39 17 18 7.2 3.9 4.2 propriate. His 612 5 2.2 1.0 1.4 If a simple repression system were operative one Arg 28 22 26 5.2 4.8 6.1 might expect tofind bothoperator and repressor gene Asx 46 46 49 8.6 10.4 11.3 mutations which derepress recE. FOUTSet al. (1983) 58 59Glx 58 53 10.7 13.3 12.3 Thr 21 23 23 3.8 5.1 5.4 found that sbcA1 and sbcA6 were dominant and cis- Ser 27 23 20 5.1 5.2 4.7 acting to sbcA+. Both are characteristics expected of 5 5 8 5 1.4 1.1 1.2 operatorand not repressor mutations (JACOB and cysTYr 11 14 14 2.0 3.0 3.3 MONOD 1961 ; JACOB,ULLMANN and MONOD1964). Phe 18 17 17 3.2 3.8 4.0 7 41 7 0.2 1.5 0.9 Thus FOUTSet al. (1983) hypothesized that sbcA was TrpPro 33 24 23 6.1 5.5 5.4 an operator of recE (ie., recEo) although they could GlY 39 33 31 7.2 7.3 7.2 not rule out otherpossibilities such as that sbcA1 and Ala 59 34 23 11.0 7.7 5.4 sbcA6 create new promotersnot repressed by the Val 47 30 30 8.7 6.7 7.0 Met 16 14 13 2.9 3.0 3.0 hypothesized Rac repressor. Ile 34 32 31 6.3 7.2 7.2 The basic nomenclature question can then be re- Leu 45 42 43 8.2 9.5 10.0 stated as follows. Is there anyevidence that some Total447 542 429 100.0 100.0 100.0 mutations which are called sbcA mutations do not affect the operatorof recE? For point mutationssbcA2, sbcA mutations, as if theyaffected oneregulatory sbcA19 and sbcA20 there is no such evidence. Muta- gene. KAISER and MURRAY (1980)found that some tions sbcA8 and sbcA81, althoughlarge deletions sbcA mutations (sbcA8,sbcA23, sbcA5O and sbcA5I) rather than point mutations, delete the operator of affected the structure of the Rac prophage, as ex- recE and hence are justifiably called sbcA mutations pected for deletions and tandemmultiplications, while even though such nomenclature is not standard and sbcA1 did not affect the structure, as expected for a lacks importantinformation. Mutations sbcA21, point mutation. In this paper we have confirmed and sbcA22 and sbcA23 differ from the othersin that they 2’72 S. K. Mahajan et al. produce new 2.6-kb EcoRI and 19-kb HindIII bands. transposition of an elementof the approximatesize of For these mutations the question becomes whether or TnlOOO (GUYER 1978).Accordingly the extra bands not the new bands signify that a sequence other than would include transposon DNA. If confirmed by fur- the operator of recE has been affected. ther experiment, sbc-21 would join the class of muta- The new 2.6-kb EcoRI band was first detected by tion expressing recE by insertion, which contains four- KAISERand MURRAY(1 980) in DNA from sbcA23 Hfr teen Tn5 and IS50 insertion mutations (FOUTSet al. strain JC7661 and in DNA from sbcA5O and sbcA51 1983; WILLISet al. 1983). Alternativehypotheses, also strains. sbcA5O is a spontaneous mutation which oc- under consideration, are that the extra bands repre- curred in a thyA-polA1 F- strain which was transduced senteither an aberrantly excised form of Rac or to thyA+recB21 (STRIKEand EMMERSON1972). sbcA51 deletion mutants of normally excised Rac which rep- is amutation selected forconferring resistance to licate more from orij. methyl methane sulfonate on a recB21 HfrH deriva- In apparent contradiction we have found here that tive (STRIKEand EMMERSON 1974). WhethersbcA51 sbc-23, after transmission from its original Hfr genetic is spontaneous or mutagen induced is thus open to background to the AB1 157 genetic background, is question. KAISERand MURRAY(1 980)suggested that not associated with theextra bands of DNA. To the new 2.6-kb EcoRI band in the three sbcA mutants explain this we note that the cross involved transfer was due to tandem multiplication of Rac prophage in of the Rac prophage into a naive cytoplasm and that the chromosome although they advanced two other Rac+ sbc-23 recombinants were very rare (GILLEN possibilities. We have adopted as working hypothesis 1974). We assume that JC8679 and its predecessor theother possibilities because they involve excised JC8675 carry a spontaneous mutationinactivating the prophage which could undergo limited multiplication int or xis genes in addition to sbc-23. We have not yet in the cytoplasm. This is consistent with our hypothesis been able to test our assumption. that the presence of the extra bands reflects coordi- nate derepression of recE and int(xis) genes of Rac so The authors are grateful toJ. P. BOUCHEfor his gift of plasmids andinformation relative tothem and the restriction site map; that there is a high frequency of prophage excision. STEVEN SANDLERfor improvements in Figure 1; WARRENGISH and Excised prophage might multiply somewhat because CONRADHALLING for help with computer-aided analysis of se- of the orij sequence(DIAZ and PRITCHARD1978; quence; and LESLIESATIN and MURTYMADIRAJU for aid in the DIAZ,BARNSLEY and PRITCHARD 1979). sequencing of sbcA8. C.C.C. was supported through the Department of Genetics at the University of California at Berkeley by National The nomenclature question then becomes whether Institutes of Health (NIH) Predoctoral Institutional Training Grant coordinate derepression of recE and int(xis) could be GM07127. This work was supported by grant NP-237 from the due to operatormutations in one genetic background American Cancer Society and NIH grant A105371 from the Na- and not another. The answer is “yes” if there are tional Institute of Allergy and Infectious Diseases. differences in the Rac prophage in different back- grounds. For example, equal amounts of recE tran- LITERATURE CITED scription could result in different amounts of down- BACHMANN,B. J., 1972Pedigrees of somemutant strains of stream int(xis) transcription if anintervening tran- Escherichia coli K-12. Bacteriol. Rev. 36: 525-557. scription terminator were less powerful in one BACHMANN,B. J., 1983 Linkage map of Escherichia coli. Microbiol. background than another. There is, however, an al- Rev. 47: 180-230. ternative possibility, namely that sbc mutations which BAGCHI,S., L. S. WISE,M. L. BROWN,D. GREGMAN, H.S. SUH and C. S. RUBIN, 1987 Structure and expression of murine malic cause coordinate derepression of recE and int(xis) may enzymemRNA, differentiation-dependent accumulation of lie in the repressor generather than therecE operator. two forms of malic enzyme mRNA in 3T3-Ll cells. J. Biol. Since prophage excision is not complete, we would Chem. 262: 1558-1565. expect that the mutant repressors would retain some BARBOUR,S. D., H. NAGAISHI,A. TEMPLINand A. J. CLARK, activity. As a result we think it preferable to leave 1970 Biochemical and genetic studies of recombination pro- ficiency in Escherichia coli, 11. Rec+ revertants caused by indirect open the possibility that sbc-21, sbc-22, sbc-23, sbc-50 suppression of Rec- mutations. Proc. Natl. Acad. Sci. USA 67: and sbc-51 are not in sbcAby substituting a hyphen 128-135. for the cistron designation in their symbols. BOLIVAR,F., 1978 Construction and characterization of new clon- Regardless of the genewhich they affect, sbc-22 and ing vehicles. 111. Derivatives of plasmid pBR322carrying unique EcoRI sites for selection of EcoRI generated recombi- sbc-23 are likely to be point mutations because they nant DNA molecules. Gene 4: 121-136. do not affect the internal structure of Rac. The same BOUCH~,J. P., 1982 Physical map of a 470 X lo3 base-pair region may not be true for the spontaneous mutant JC7659 flanking the terminusof DNA replication in the Escherichia coli carrying sbc-21 because strong 2.2-kb EcoRI and 13- K-12 . J. Mol. Biol. 154 1-20. kb Hind111 bands are found in addition to the 2.6-kb BOUCHE,J. P., J. P. GELUNGE,J. LOURARN,J. M. LOURARN andK. KAISER,1982 Relationships between the physical and genetic EcoRI and 19-kb HindIII bands. We are puzzled by maps of a 470 X 10’ base-pair region around the terminus of this result, but our working hypothesis is that sbc-21, Escherichia coli K12 DNA replication. J. Mol. Biol. 154: 21-32. as a spontaneousmutant, may have been produced by CAPALDO-KIMBALL,F., and S. D. BARBOUR, 1971 Involvement of Analysis of ExoVIII+ Mutants 273

recombination genes in growth and viability of E. coli K-12. J. suppressor. Proc. Natl. Acad. Sci. USA 71: 3593-3597. Bacteriol. 106: 204-212. LLOYD,R. G., AND S. D. BARBOUR, 1974The genetic location of CHU,C. C., A. TEMPLINand A. J. CLARK, 1989 Suppression of a the sbcA gene of Escherichia coli. Mol. Gen. Genet. 134: 157- frameshift mutation in the recE gene of E. coli K12 occurs by 171. gene fusion. J. Bacteriol. 106 2101-2109. LOW,B., 1973 Restoration by the rac locus of recombinant form- DIAZ, R.,P. BARNSLEYand R. H. PRITCHARD,1979 Location and ing ability in recB- and red- merozygotes of Escherichia coli K- characterisation of a new replication origin in the E. coli K12 12. Mol. Gen. Genet. 122: 119-130. chromosome. Mol. Gen. Genet. 175: 151-157. MAGNUSON,M. A., H. MORIOKA,M. F. TECCEand V. M. NIKODEM, DIAZ,R., and R. H. PRITCHARD,1978 Cloning of replication 1986 Coding nucleotide sequence of rat liver malic enzyme origins from the E. coli K12. Nature 275: 561-564. mRNA. J. Biol. Chem. 261: 1183-1 186. ESPION,D., K. KAISER and C. DAMBLY-CHAUDIERE,1983 A third MANIATIS,T., E.F. FRITSCH and J. SAMBROOK, 1982Molecular defective lambdoid prophage of Escherichia coli K12 defined Cloning. Cold Spring Harbor Laboratory, Cold Spring Harbor, by the X derivative, X gin 11 1. J. Mol. Biol. 170 61 1-633. N.Y. FOUTS,K. E., T. WASIE-GILBERT,D. K. WILLIS,A. J. CLARK andS. MESSING, J., 1983 New M13 vectors for cloning. Methods En- D. BARBOUR,1983 Genetic analysisof transposon-induced zymol. 100 20-78. mutations of the Rac prophage in Escherichia coli K-12 which SANGER,F., G. J. AIR, B. G. BARRELL,N. L. BROWN,A. R. COULSON, affect expression of recE. J. Bacteriol. 156: 718-726. J. C. FIDDESS, C.A. HUTCHISON111, P. M. SLOCOMBEand M. GILLEN,J. R., 1974 The RecE pathway of SMITH,1977 Nucleotide sequence of bacteriophage OX174 in Escherichia coli. Ph.D. thesis, University of California at DNA. Nature 265: 687-695. Berkeley. SHINE,J., and L. DALGARNO,1974 The 3”terminal sequence of GILLEN,J. R., A. E. KARU,H. NAGAISHI and A. J. CLARK, E. coli 16s RNA. Proc. Natl. Acad. Sci. USA 71: 1342-1346. 1977 Characterization of the deoxyribonuclease determined SMITH,G. E., and M. D. SUMMERS, 1980The bidirectional trans- by lambda reverse as exonuclease VI11 of Escherichia coli. J. fer of DNA and RNA to nitrocellulose or diazobenzyloxy- Mol. Biol. 113: 27-41. methyl-paper. Anal. Biochem. 109 123-129. GILLEN,J. R., D. K. WILLIsand A. J. CLARK,1981 Geneticanalysis SPINA,J., JR., H. BRIGHTand J. ROSENBLOOM,1970 Purification of the RecE pathway of genetic recombination in Escherichia and properties ofL-malic enzyme from Escherichia coli. Bio- coli K-12. J. Bacteriol. 145: 521-532. chemistry 9: 3794-3801. GOTTESMAN,M. M., M. E.GOTTESMAN, S. GOTTESMANand M. STRIKE,P., and P. T. EMMERSON,1972 Coexistence of polA and GELLERT, 1974 Characterization of bacteriophage X reverse recB mutations of Escherichia coli in the presence of sbc, a as an Escherichia coli phage carrying uniquea set of host-derived mutation which indirectly suppresses recB. Mol. Gen. Genet. recombination functions. J. Mol. Biol. 88471-487. 116: 177- 180. GUYER,M. S., 1978 The 76 sequence of F is an insertion sequence. STRIKE,P., and P. T. EMMERSON,1974 Degradation ofDNA J. Mol. Biol. 126: 347-365. following ultraviolet irradiation of Escherichia coli K12 mutants HANSEN,E. J., and E. JUNI, 1975 Isolation of mutants of Esche- lacking DNA polymerase I and exonuclease V.Mol. Gen. richia coli lacking NAD- and NADP-linked malic enzyme ac- Genet. 130 39-45. tivities. Biochem. Biophys. Res. Commun. 65: 559-566. TEMPLIN,A., S. R. KUSHNER and A. J.CLARK, 1972 Genetic HARAYAMA,S., E. T. PALVA and G. L. HAZELBAUER, analysis of mutations indirectly suppressing recB and recC mu- 1979 Transposon-insertion mutants of Escherichia coli K-12 tations. Genetics 72: 205-21 5. defective in acomponent common to galactose and ribose WAHL,G. M., M. STERNand G. R. STARK, 1979Efficient transfer chemotaxis. Mol. Gen. Genet. 171: 193-203. of large DNA fragments from agarose gels to diazobenzyloxy- JACOB, F., and J. MONOD,1961 On the regulation of gene activity. methyl-paper and rapid hybridization using dextran sulfate. Cold Spring Harbor Symp. Quant. Biol. 26 193-21 1. Proc. Natl. Acad. Sci. USA 76 3683-3687. JACOB, F., A. ULLMANN andJ. MONOD, 1964 Le promoteur PI& WILLETTS,N. S., and A. J. CLARK, 1969 Characteristics of some ment gknktique neckssaire i I’expression d’un operon. C. R. multiply recombination-deficient strains of Escherichia coli. J. Acad. Sci. 258 3 125-3 128. Bacteriol. 100: 231-239. KAISER,K., 1980The origin of Q-independent derivatives of WILLETTS,N. S., and D. W. MOUNT, 1969 Genetic analysis of phage X. Mol. Gen. Genet. 179: 547-554. recombination-deficient mutants of Escherichia coli K-12 car- KAISER,K., and N. E. MURRAY,1979 Physical characterization of rying rec mutations cotransducible with thyA. J. Bacteriol. 100: the “Rac prophage” in E. coli K12. Mol. Gen. Genet. 175: 159- 923-943. 174. WILLIS,D. K., K. E. FOUTS,S. D. BARBOURand A. J. CLARK, KAISER, K., and N. E. MURRAY,1980 Onthe nature of sbcA 1983 Restriction nuclease and enzymatic analysis of transpo- mutations in E. coli K12. Mol. Gen. Genet. 179 555-563. son-induced mutations of the Rac prophage which affect KATSUKI, H., K. TAKEO, K. KAMEDA and S. TANAKA, expression and function of recE in Escherichia coli K-12. J. 1967 Existence of twomalic enzymes in Escherichia coli. Bacteriol. 156 727-736. Biochem. Biophys. Res. Commun. 27: 331-336. YAMAGUCHI,M., M. TOKUSHICEand H. KATSUKI,1973 Studies KUSHNER,S. R., H. NAGAISHIand A. J. CLARK,1972 Indirect on regulatory functions of malic enzymes 11. Purification and suppression of recB and recC mutations by exonuclease I defi- molecular properties of nicotinamide adenine dinucleotide- ciency. Proc. Natl. Acad. Sci. USA 69: 1366-1370. linked malic enzyme from Escherichia coli. J. Biochem. 73: 169- KUSHNER,S. R., H. NAGAISHIand A. J. CLARK,1974 Isolation of 180. exonuclease VIII: the enzyme associated with the sbcA indirect Communication editor: J. R. ROTH