Copyight 6 1995 by the Genetics Society of America
Suppression of RecJ Exonuclease Mutants of Escherichia coli by Alterations in DNA Helicases I1 (uvrD)and IV (heZD) Susan T. Lovett and Vincent A. Sutera, Jr. Department of Biology and Rosenstiel Basic Medical Sciences Center, Brandeis University, Waltham, Massachusetts 02254- 91 10 Manuscript receivedJune 16, 1994 Accepted for publication January 23, 1995 ABSTRACT The recj gene encodes a single-strand DNA-specific exonuclease involved in homologous recombination. We have isolated a pseudorevertant strain in which recjmutant phenotypes were alleviated. Suppression of recjwas due to at least three mutations, two of which we have identified as alterations in DNA helicase genes. A recessive amber mutation, “uwD517,,,” at codon 503 of the gene encoding helicase I1 was sufficient to suppress recj partially. The uwD517,, mutation does not eliminate u w D function because it affects W survival onlyweakly; moreover, a uvrD insertion mutation could not replace uwD517,, as a suppressor. However, suppression may result from differential loss of UWD function: mutation rate in a uwD517,, derivative was greatly elevated, equal to that in a uznD insertion mutant. The second cosuppressor mutation is an allele of the helD gene, encoding DNA helicase IV, and could be replaced by insertion mutations in helD. The identity of the third cosuppressor “srjD” is not known. Strains carryingthe three cosuppressor mutations exhibited hyperrecombinationalphenotypes including elevated excision ofrepeated sequences. To explain recjsuppression,we propose that loss of antirecombinational helicase activity. by. the suppressor mutations stabilizes recombinational intermediates formed in the absence of re$ ”
I
N the bacterium Escherichia coli, two exonucleases participate in genetic recombination: the RecBCD and RecJ nucleases (reviewed in WEST1993) . Exonucleases may promotegeneticexchangeat several different steps. The RecA protein of E. coli, which pairs and transfers strands between homologous DNA molecules in vitro, requires that one of the DNA partners in the strand exchange reaction must be at least partially single-stranded ( CASSUTO et al. 1980; CUNNINGHAM et al. 1980). Therefore, proteins such as the RecBCD nuclease/helicase complex and the RecJ exonuclease could function presynaptically to create the singlestranded regions in duplex DNA, which RecA requires for strandinvasion. In addition,single-strand DNA exonucleases such as RecJmay facilitate recombination after joint molecules have been formed. During the branch-migration phase of RecAstrand-transfer in vitro, RecJ enhances the rate of strand-assimilation into heteroduplex, presumably by degrading a competitor strandfor pairing (CORFWITE-BENNETT and LOVETT 1995). As a final postsynaptic step, exonucleases may be required to “trim” unpaired DNA before ligation. This study focuses on the genetic function of the recJ gene of E. coli. The r e g gene encodes a 5 ’ to 3’ DNA exonuclease with a very strong preference for singlestrand DNA ( LOVETIand KOLODNER1989). Strains mutant in the recJgene are highly defective for recombinaCorrapunding authm: Susan T. Lovett, Department of Biology and Rosenstiel Basic Medical Sciences Center, Brandeis University, Waltham, MA 022549110. E-mail:
[email protected] Genetics 140: 27-45 (May, 1995)
tion of duplicated genes carried on plasmids ( KOLODNER et al. 1985). TherecJgene is also essentialfor the minor recBCD-independentpathways of recombination measured in conjugational and transductional crosses (LoVETT and CLARK1984) . Mutations inrecJare highly synergisticwith those inrecBW: recBC recJ mutants are extremely recombination deficient, UV sensitive and are partially inviable (LOVETTand CLARK1984). To investigate the role of RecJ in genetic recombination in vivo, we have isolated second-site suppressor mutations of various recJ mutations. We report here the isolation of a suppressor strain that carries at threemutations, which together suppress any allele of re$ One of these “srj” (for suppressor of ZecJ) mutations we show to be an allele of the UWD gene encoding DNA helicase I1 ( HICKSONet al. 1983), anda second cosuppressor mutation is an allele of the heZD gene, encoding DNA helicase IV (WOODand MATSON 1989) . The existence of the third mutation, srjD, is implied by genetic analysis but its identity is yetunknown. The phenotypes of these mutants suggest that helicase I1 (and perhaps helicase IV) have antirecombinational properties and that recombination in the absence of RecJ may be especially prone to reversal by the action of these antirecombinational helicases. MATERIALS AND METHODS Bacterial strains and media: The majority of strains used in this study are derived from AB1157 and listed in Table 1. Strains were grown routinelyin LB medium ( WIUETTS et al. 1969), with plate media containing 1.5%agar. Plate minimal
Lovett S. T.
28
and V. A. Sutera
TABLE 1 Eschericia coli K-12strains Strain AB1 157 and derivatives AB1157" AM207 JC7623 JC8679 JClO990 JC12123 JC12159 JC13015 JC13018 RDK1230 RDK1445 RDK1541 RDK1658 RDK1687 RDK1688 RDK1690 RDKl769 RDK1811
Genotype
RDK2121
recR252::Tn 10-9 recB21 recC22 sbcBl5 sbcC2Ol recB21 recC22 ~bcA.23~ re8332::Tn' (PhiX') red284 ::Tn 10 recB21 recC22 sbcA23 lysA recB21 recC22 sbcA23 red284 : :Tn 10 recB.21 recC22 sbcA23 red77 srl-300::Tn 10 red56(supE+) serAl zgb224 ::Tn 10 recO1504::Tn5 recB2l recC22 sbcA23 recO1504::Tn5 recB2l recC22 sbcA23 recTl01::Tn 10' recB21 recC22 sbcA23 red56 srl-300::Tn 10 recB21 recC22 sbcA23 recQ61::Tn3 recB2l recC22 sbcA23 aroA354 ompF::Tn5 recB2l recC22 sbcA23 helD104 uvrD517,,mjD7 smAl zgb224::TnlO recB.21 recC22 sbcA23 helD104 uwD517,,mjD7 zgb 224::TnIO recB2l recC22 sbcA23 red77 helD104 uwD517,,,srjD7 red56 srl-300::Tn 10 recB21 recC22 sbcA23 recJ77 helD104 uwD517,,mjD7 re8332::Tn3 recB21 recC.22 sbcA23 helD104 uwD517,,srjD7 recJ284: :Tn 10 recB21 recC22 sbcBl5 sbcC201 red2051 ::Tnl0-9
RDK2128
recJ2051 ::Tn 10-9
RDK2 144
recB21 recC22 sbcA23 recJ2003::Tn 10-9
SR1277 STL282 STL658 STL660
uvrD254 ::Tn5 recB.21 recC22 sbcA23 re8332::Tn3 recB2l recC22 sbcA23 recR252::Tn 10-9 recB2l recC22 sbcA23 red77 helDlO4 uwD517,,,srjD7 recTl01::Tn 10 recB2l recC22 sbcA23 recJ77 helD104 uwD517,,,srjD7 recR252::Tn 10-9 recB21 recC22 sbcA23 red77 helD104 uvrD517,,mjD7 recO1541::Tn5 lacZ: : bla recB21 recC22 sbcA23 helDlO2 ::Tn 10-9 recB21 recC22 sbcA23 helD102::Tn10-9 red284 ::Tn 10 recB21 recC22 sbcA23 helDl03: :Tn 10-9 recJ284::Tn 10 recB.21 recC22 sbcA23 helDl01::Tn 10-9 recB.21 recC22 sbcA23 helDlO3::Tnl0-9 recB21 recC22 sbcBl5 sbcC201 helDlO2: :Tn 10-9 helDlO2: :Tn 10-9 recB21 recC2.2 sbcA.23 helD104 uwD517,,srjD7
RDKl812 RDK1813 RDK1814 RDK1819
STL662 STL664 STL695 STL811 STL814 STL815 STL834 STL837 STL863 STL864 STL941 STL964 STL1028 STL1029
recB21 recC22 sbcA23 zcc-282::Tn 1W D 3 4 recB2l recC22 sbcA23 recJ.2051: : Tnl0-9 recB21 recC22 sbcA23 rec12051:: Tn10-9 helD104 uwD517,,sriD7
Derivation BACHMANN (1987) MAHDI and LLOYD (1989) HONI and CLARK(1973) GILLENet al. (1981) BLANAR et al. (1984) LOVETT and CLARK(1984) LOVETT and CLARK(1984) LOVETT and CLARK (1984) LOVETT and CLARK (1984) R. KOLODNER R. KOLODNER KOLODNERet al. (1985) Km' transductant P1 RDK1541 X JC8679 Tc' transuctant P1 KF1053 X JC8679 Tc' W transductant P1 RDKl230 X JC8679 Ap' transductant P1 KD1996 X JC8679
-d
Tc' Ser- transductant P1 RDK1445 X STL1051" Tc' Ser+ transductant P1 RDK1445 X STLlO51 Tc' W transductant P1 RDK1230 X STL1051" Ap' transductant P1 JClO990 X STL1051" Ser+ W transductant P1 JC12123 X RDK1811" Km' Aps W pRDK161 transformant of JC7623 K m ' Ser+ transductant P1 RDK2121 X RDKl445 K m ' Aps W pRDK163 transformant of JC8679 N. SARGENTINI Ap' transductant P1 JClO990 X JC8679 K m ' transductant P1 AM207 X JC8679 Tc' transductant P1 KF'1053 X STL1051' Km' transductant P1 AM207
X
STLlO51'
K m ' transductant P1 RDK1541 X STLlO51"
LOVETT et al. (1993) Km' Aps pSTL45 transformant of JC8679 Tc' transducant P1 JC12123 X STL811 Tc' transducant P1 JC12123
X
STL837
K m ' Aps pSTL42 transformant of JC8679 Km' Aps pSTL46transformant of JC8679 Km' transductant P1 STL811 X JC7623 Km' transductant P1 STL811 X AB1157 Ser+ Tcs transductant P1 AB1157 X RDK1812' Tc' Pyr- transductant P1 DC305 X JC8679 Km' transductant P1 RDK2128 X JC8679 Km' transductant P1 RDK2128 X STL1051"
recJ Suppression
TABLE
29
1
Continued Genotype STLlO3l STL1036 STL1051 STL1066 STLll62 STL1163 STL1495 STL1496 STL1498 STL1499 STL1500 STL1523 STL1524 STL1526 STL1527 STL1543 STL1544 STL1548 STL1549 STL1550 STL1551
recB2l recC22 sbcA23 helD102::Tn10-9 recQ61: :T n 3 recB2l recC22 sbcA23 recJ2003: :Tn 10-9 helDlO4 uwD517.,s@7 recB21 recC22 sbcA23 recJ77 helD104 uwD517,,sr3D7 recB2l recC22 sbcA23 recJ77 helD104 uwD517,,srjD7 pyrD34 zcb222: :Tn 10 recB2l recC22 sbcA23 lacZ: : bla' tetA,,, recB2l recC22 sbcA23 helDI04 uwD517,,srjD7 lacZ: :blu+ tea,,,, recB2l recC22 sbcA23 metE3079::TnlO recB2I recC22 sbcA23 helDl02: :Tn 10-9 metE3079: :Tn 10 recB21 recC22 sbcB15 sbcC201 helD102::Tn10-9 metE3079::TnlO helD102::Tn10-9 metE3079::TnlO recB21 recC22 sbcA23 uwD254: :T n 5 recB2l recC22 sbcA23uwD254: :T n 5 recQ61: :T n 3 recB21 recC22 sbcA23 uurD254: :T n 5 helD102::TnIQ9 uwD254 ::T n 5 recB21 recC22 sbcB15 sbcC201uvrD254::T n 5 recB2I recC22 sbcBI5 sbcC2Ol uwD254: :Tn5 helDlO2::Tn10-9 ::Tn 10-9 uwD254 ::Tn5 hlD102
STL1661
recQ61: :T n 3 recB2I recC22 sbcBl5 sbcC201 recQ61::T n 3 recQ61: :T n 3 uwD254 ::T n 5 recB2I recC22 sbcB15 sbcC201 recQ61: :T n 3 uvrD254: :T n 5 recB21 recC22 sbd23 uwD254: :Tn5 helD102::TnlO-9 recQ61::Tn3 helDlO2: :Tn I 0-9 recQ61: :T n 3 recB2l recC22 sbcBl5 sbcC201 helDl02: :Tn 10-9 recQ61: :Tn? recB2I recC22 sbcB15 sbcC201 helD102::Tnl09 recQ61: :T n 3 uwD254: :T n 5 uurD254::Tn5 helD102::Tnl0-9recQ61::Tn3 recB21 recC22 sbcA23p y # : :T n 5 recB21 recC22 sbcA23 helD104 uwD517,,sr~D7pr#: :T n 5 recB2I recC22 sbcA23 i l ~ 3 1 6 4:Tn : 10-kan metE3079: :Tn 10 recB21 recC22 sbcA23 helD104
STL1665
recB2l recC22 sbcA23 uvrD517,,
STL1672 STL1674 STL1681
recB2l recC22 sbcA23 uwD517,,recJ284::Tn10 recB21 recC22 sbcA23helDl04 recJ284: :Tn 10 recB21 recC22 sbcA23 uwD517,,,,pyrF::T n 5
STL1683
recB21 recC22 sbcA23 uwD517,,zcc-282: :Tn IO pyrD34 recB21 recC22 sbcA23 helD104 uwD517,,
STLl609 STL1611 STL1613 STL1615 STLl617 STL1639 STL1640 STL1651
STL1685 STL1686 STL1688
recB21 recC22 sbcA23 helDl04 uwD5I 7, re4284 ::Tn I O recB21 recC22 sbcA23 heLD104 uurD517,, srjD7 recJ284 ::Tn 10
Derivation
Ap' transductant P1 RDKl690 X STL811 K m ' transductant P1 RDK2144 X STL1051' MitoC' "revertant" of JC13018' Tc' Pyr- transductant P1 CY307 X STL1051"
Ap' transductant PI STL695 X JC8679 Ap' transductant P1 STL695 X STL941' Tc' transductant P1 CAG18491 X JC8679 Tc' transductant P1 CAG18491 X STL811 Tc' transductant P1 CAG18491 X STL863 Tc' transductant PI CAG18491 X STL864 K m ' transductant P1 SR1277 X JC8679 Ap' transductant P1 RDK1690 X STL1500 Met+ uv" transductant P1 SR1277 X STL1496f Km' transductant PI SR1277 X AB1157 K m ' transductant P1 SR1277 X JC7623 Met+ uv" transductant P1 SR1277 X STL1498f Met' uv" transductant P1 SR1277 X STL1499f Ap' transductant P1 RDK1690 X AB1157 Ap' transductant P1 RDK1690 X JC7623 Ap' transductant P1 RDKl690 X STL1526 Ap' transductant P1 RDKl690 X STL1527
Ap' transductant PI STL1523 X STL1524f Ap' transductant P1 RDKl690 X STL864 Ap' transductant P1 RDK1690 X STL863 Ap' transductant PI STL1523 X STL1543f Ap' transductant P1 STL1523 X STL1544f K m ' Ura- transductant P1 DB6507 X JC8679 Krd Ura- transductant P1 DB6507 X STL941 Km' Tc' transductant P1 CAG18599 X STL1495 P y r 'Tcs transductant P1 STLlO5l X STL964" Ilv' Met+ transductant P1 STL1051 X STL1651' Tc' transductant PIJC12123 X STL1665' Tc' transductant P1 JC12123 X STL1661" K m ' Ura- transductant P1 DB6507 X STL1665' reTc' Pyr- transductant P1 DC305 X STL1665" P y r 'Tcs transductant P1 STLlO5l X STL1683" Tc' transductant P1 JC12123 X STL1685' Tc' transductant P1 JC12123 X STL1051"
S.
T. Lovett and V. A. Sutera
30
TABLE 1 Continued
Strain
Genotype
Derivation
3 recJ284::TnIO recB21 recC22 sbcA2? recQ61::Tn recB21 ~recC22sbcA23 uwD254::Tn5 recJ284::TnIO recB21 recC22 sbcA23 uwD254: :Tnt recQ61: :Tne recJ284::TnIO STLl745 recB21 recC22 sbcA23 uwD254::Tn5 helD102:: TnlO-9 recJ284::TnIO STLl746 recB2l recC22 sbcA2? helD104 uwD517.,srjD7 sup,?? STLl749 recB2I recC22 sbcA2? recJ284:: TnlO sup,?? STL1751 sup,?? sbcA23 recC22 recB2I STLl766 recB21 recC22 sbcA2? helDlO4 uwD517,,s?JD7 recJ284 ::Tn 10 sup,?? STL1789 recB21 recC22 sbcA2? helD102:: TnIO-9 recQ61::Tnjr Tc' uvrD254::Tn5 recJ284::TnIO STLl790 recB21 recC22 sbcA23 helD102:: TnIO-9rec@l::Tn? reg284::Tn IO STL1854 recB21 recC22 sbcA2? uvrD254 ::Tn5 supE+ sup+ STLl742 STLl743 STLl744
Other strains CAG18491 CAG18557 fadAB?I CAG18599 ilv-3164 CY307 DB6507 DC305 DPBlOl HMSl74 JC158 KD1996 recQ61 KF'1053" "55
mtE3079: :Tn IO 65::Tn IO-kan : :Tn 10-kan Hfr PO3 rcb222: :TnlOpyrD34 r e a l spoT1 mtBl f y F : :Tn5 recAl? leuB6 A(@t+roA)62 thi-1 lacy1 ara-I4 xyl-5 mtl-1 qsL20 hsdS20 supE44 zcc-282::TnIOpyrD34 galK2 mal41 xyl-7B.mtl-2 rpsLl18 himD451: :Tn 10-9A(1acjno)rpsL recA 1 riy hsdR Hfr PO1 serA6 thi-1 rel4l lacZ22 ::Tn? ilv-145 t q C 3 pro mtl-1 thi maMl ara-9 gal= lac-114 rpsL ton pol412 Hfr PO45 recB2l recC22 sbc-I1 I ::Tn5 recTlO1: :Tn IO thr-300 ilv-318 lacAU169 araD ompF: :Tn5
Tc' transductant P1 JC12123 X RDK1690 Tc' transductant P1 JC12123 X STL1500 Tc' transductant P1 JC12123 X STL1523 Tc' transductant JC12123 P1
X
ST11524
sup+ revertantSTL941" of sup+ revertant of JC13018 sup' revertant JC8679 of sup' revertant STL1688" of transductant P1 JC12123 X STL1609 Tc' transductant P1 JC12123
X
STL1031
revertantSTL1500 of C. GROSS c. GROSS C. GROSS B. BACHMANN J. HABER BACHMANN S. COHEN R. KOLODNER CLARK (1963) NAKAYAMA et al. (1984)
FOUTESet al. (1983) M. SWANEN
a Genotype of AB1157 and derived strains, unless otherwise indicated, includes F-A - thi-1 hzsG4 A(@tproA)62 axE2 thr-I leuB6 kd K51 Dl ara-14 lacy1 galK2 xyl-5mtl-1tsx-?3supE44 rpsL?l rac-. 'JC86% and its derivates are rat+. Originally described by FOUTESet al. (1983) as "recElO1 ::TnlO", this mutation is an insertion in the linked recTgene (R. D. KOLODNER,personal communication). Km' Aro- transductant of STLlO51 with P1 grown on RDKl768. RDKl768 was a K m ' transductant produced from SY503 with P1 donor "55. (gal aroA?54 supE strain obtained from M. SWANEN) 'ThehelDIO4 mutation was originally assigned the designation "srjB2"; uvrD517,,,was designated "srjC6." The identity and map position of srjD is not known but its existence is inferred by the genetic characterization described in the text. f Because both uvrD254: :Tn5 and helD102: :TnlO-9confer Km', the presence of the helD allele was confirmed by transductional backcross into recipient DC305, showing cotransduction of Km', Tcs and P y r ' phenotypes. The presence of the uvrD allele is inferred by U V phenotype of these strains.
medium consisted of 56/2 salts ( WILLETTSet ai. 1969) supplemented with 0.2% glucose, 1 pg/ml thiamine and 50 pg/ml of the appropriate required amino acids. Uracil or thymine, when required, were added to 50 pg/ml. Streptomycin (Sm) , tetracycline (Tc) , kanamycin ( K m ) and rifampin (Rif)were used at concentrations of 100,15,30 and100 pg/ml, respectively. Strains carrying re&: :Tn?, recQ: :Tn? or lac ::bla+ tetA, were selected on media containing 30 pg/ml ampicillin (Ap) ; selection for plasmids conferring Ap resistance employed media containing 100 pg/ml ampicillin. For selection of suppressor mutants, LB plates containing Mitomycin C (Sigma) at 2 pg/ml were employed. P1 phage were grown in plate lysates on LCTG medium, LB plates containing 2 mM CaC12, 50 pg/ml thymine, 0.2% glucose and 1.2% agar (top
agar contained only 0.7% agar). P1 transductions were performed as in WILLE~TSet d. 1969. Reversion of supE44 was selected by plating of AJC1929 (N7 N53 cZ26) and AJC1930 ( h80 N7 N53 cZ26) on appropriate strains. Survivors ofphage infection were then screened for plating ofA1929,A1930, AcZ26 and Avir to identify the sup' derivatives. Plasmids: Restrictionenzymes and T4 DNAligasewere purchased from New England Biolabs, Inc. A 5.4kb BamHI fragment contained thehelDgene was cloned into the BamHI site of plasmid pBS Ks- (Stratagene, Inc.) from A DNA prepared from phage 223 of the Kohara "miniset" library (KOHARA et al. 1987), producing plasmid pSTL3O. The orientation of the fragment is such that the helD gene is not transcribed by the lac promoter on this plasmid. Plasmid
recJ Suppression pSTL34 is the PstI BamHI helD+ fragment cloned into vector pT7-6 (TABOR 1990). Plasmids pSTL35 is an EcoRI deletion of pSTL34, which has the Cterminal portion of helD deleted from the unique EcoRI site within helD to the EcoRI site of the vectorpolylinker.Plasmids containing the u d region from PCR-amplified chromosomal DNA from JC8679 ( u d ’ ) or STL1051 (uvrD5170,,,) were cloned into low-copy vector pWSK29 ( WANGand KUSHNER1991). PCR primers were: 5’ CGGCGAGATCTTTACATGTTGG and 5 ’ GGCTCTAGATACTGAAGATGG. Amplification was performed with Tag polymerase (Promega) and the supplied buffer supplemented with2.5 mM MgC12 for 25cyclesof95” 1 min, 50” 1 min and 72” 2 min. The PCR product was purified by glass-bead extraction ( USBioclean fromUS Biochemicals) from agarose gels,cleavedwith XbuI and BglII and ligated into vector pWSK29 cut withXbaI and BamHI.Two uwD+ (pSTL101, pSTL102) and three ud517,, (pSTL103, pSTL104, pSTL105) plasmidisolatesweresaved. No differencesin subsequent complementation analysis were observedamong the independent uwD+ or uvrD517,, isolates although only data for pSTLlOl and pSTL105 are shown. All five of these plasmids exhibited complementation of the W sensitivity conferred by the uwD254::Tn5mutation when transformed into STL1526. Isolation of TnIO-9 (element “9”of WAY et al. 1988) insertion mutants of plasmids are discussed briefly here and in more detail below.PlasmidspRDKl61 and pRDK163 are Tnl09insertion mutants ( recJ2051 and recJ2003, respectively) ofpJC763 ( LOVEIT and CLARK 1985). PlasmidspSTL42, pSTL45 and pSTL46 are Tn109insertion mutants (helD101, helD102 and helD103, respectively) of pSTL30. Plasmid DNA was purified by the alkaline-SDS method ( BIRNBOIM and DOLY 1979) as modified by (AUSUBEL et al. 1989). Lambda phage DNA was prepared as described ( SILHAWet al. 1984). We used the TSS method ( CHUNG et al. 1989) or electroporation et al. 1988) for introducing plasmid DNA into strains. (DOWER Recombination and UV-survivalassays: Patch tests for conjugational recombination proficiency and W sensitivity (CLARKand MARGULIES 1965) were used to score rec mutations in strain constructions and in mapping experiments. Recombination in Hfr crossesand UV survival werequantified as described in LOVETT et al. ( 1988) from, in most cases, LB cultures grown to mid-log stage (OD590 = 0.4). Selections and duration of the matings are indicated in the tables in the text. For the experiment using the strains carrying various RecF pathwaymutations, nonaerated ( “standing”) overnight cultures were tested. Serial dilutions were performed in 56/ 2 buffer and plated on appropriate selective media. Notethat in matings with rec mutants, extended transfer can result in expression of the rec+ allele in the recipient and complementation of the mutant phenotype, producing a false baseline level of recombination. Apparent recombination values can also be depressed by defects in viability or failure to initiate conjugation. Because recmutantsof E.coli show a concomitant defect in W repair, we therefore always determine the Wsurvival phenotype of mutant strains to compare with their conjugational recombination capacity. Isolation of Tn2&9 insertion mutations: Transposition of the defective TnlO element carryingkanamycinresistance (“element 9” as described in WAYet al. 1988) into plasmid pJC763 and pSTL30 were performed as described (WAYet al, 1988). Plasmids that carried the Tn109 element wereselected from the population by subsequent retransformation into HMS174 conferring both Ap and Km resistance. The location and orientation of insertions were determined by restriction analysis ofpurified plasmid DNA. All the insertions isolated in the chromosomal helD region of plasmid pSTL30 had the same orientation of the transposon, with the kan gene transcribed in the opposite orientation to helD. Insertions in
31
recJgene of pJC763 included reg2003 and recJ2051 oriented with kun reading in the same direction as re$ Transformation of Tn2&9 mutationsto the E. coli chromosome: Plasmid DNA carrying the recJ2003,recJ2051,helDlO1, helD102 and he1103::TnlO-9mutationswas subjected to digestion with the restriction enzymes (EcoRI, PstIand Sun for the recJalleles;EcoRI SalI for the helD alleles). This cleaved DNA was transformed into strain JC7623(recB21recC22sbcB15 sbcC201) or JC8679 ( recB21recC22sbcA23), selecting Km‘. The presence of the recJalleleswas confirmed by the W and Rec phenotype of the resulting transformants, by transductional linkage of Km‘ with SerA+ in crosses into RDK1445 et al. 1989) using and by Southern blot analysis (AUSUBEL digoxygenin-labeled pJC763 probe (kit purchased from Boehringer Mannheim) . The presence ofhelD insertion in the appropriate location in the E. coli chromosome was confirmed similarlyby Southern blot analysis with pSTL30probe and by genetic linkage to pyrD in transductional crosses. Sequence analysis: Variousrestrictionfragmentswere cloned into M13 phage mp18 or mp19. DNA sequence was determined using Sequenase 2.0 (US Biochemical). The sequence of various TnlO-9 insertion sites was determined by cloning the B a d 1 SalI fragments of the recJ2051 and recJ2003 plasmid mutants into the M13 sequencing vector mp18 and the BamHIEcoRI fragment of the helD plasmid mutant (helDlOl and helD102) into mp18 vector. The universal -40 primer was used to obtain DNA sequence from the B a d 1 site of the TnlO-9elements into the adjoining DNA. To determine the location of recJ284::TnlO, the B a d 1 SaZI fragment from plasmid pJC760 (LOVETT and CLARK1985) was cloned into mp19. A synthetic primer complementary to the end of TnlO ( 5 ’ GACAAGATGTGTATCTACC) was used to obtain DNA sequence across the insertion site. The recJ77 allele was sequenced from PCR-amplified chromosomal DNA from strain RDK1789(LOVETT et al. 1988) with Tag polymerase (Promega) . PCR primers were 5 ’ AGCAATGGCACACITGTTCCG and 5’ TTAAGCCGTAACCGTCTTCGA. The PCR fragment was digested with EcoRI KpnI and cloned into M13 mp18 and mp19 for sequencing. Sequence was determined for two independent isolates. The UWDlocus was amplified from chromosomal DNA purified fromJC8679 (uvrD+) or STL1051 (uvrD517am),cleaved withPstI and EcoRI and cloned into M13 vectors mp18 or mp19with subsequent dideoxy sequencing as above.PCR primers were: 5 ’ AGTTGTGGCTT-MCAAGCCGC and 5 ’ TTCACCTGCTTCCAGTGCCGC and amplification was performed with Tag polymerase ( Promega) under the conditions described previously. Five clones derived from two independent PCR reactions for both uwD+ and uvrD517., template were sequenced to verify the presence of the uvrD517,, C to T transition at codon 503 of UvrD. Excision and mutationassays: All excision and mutation rates were calculated by method of the median as described in LEAand COULSON 1949. The py#::Tn5mutation was introduced into appropriate strain backgrounds byP1 transductionfrom DB6507, selecting Km resistance on platescontaining additional uracil (50 pg/ml) . Precise excision was determined by resuspending entire overnightcolonies on LB+Ura medium into LB+Ura broth with subsequent overnight growth. Cultures were collected, washed twice in 56/2 buffer and plated on 56/2 minimal medium lacking Ura. The average number of cells in these cultures was determined for the cultures by serial dilution and plating on LB + Ura medium. The excision rate was calculated usingthe median number of Ura+ and the mean number of cells in the cultures (LEA and COULSON 1949). Nearly precise excision ofa Tn 10 derivative was assayed by introducing plasmid pHV857 ( D’ALENCON et al. 1994) into
32
and Lovett S. T.
appropriate strain background by transformation, selecting Km‘. Excisionofthe miniTnlMan element (marked with a Km-resistance gene and lacking the transposase) restores transcription of the bla gene and therefore Ap resistance to the cell. Similar to procedure above, the number of Ap‘ colonieswas determined from independent cultures inoculated from entire colonies grown an additional 2 hr. The excision rate was calculated from the median number of Ap‘ colonies in the cultures and the mean number of Tc‘ (total) cells in the cultures (LEAand COULSON1949). A 787-bp tetA chromosomal duplicationwas transduced into several genetic backgrounds by P1 transduction selecting Ap‘. Excision of the repeat was measured by determining themedian number ofTc‘cellsforseveral independent cultures inoculated from entire individual colonies as described previously LOVE^ et al. 1993). Mutation to Rif rwas determined by inoculating 100-1000 cfu of an overnight culture in seven tubes. After overnight growth, samples from each independent culture were plated on LB irifampin, and total viable cells were determined by dilution with 56/2 buffer and plating on LB. Colonieswerecounted after overnight growth and the mutation ratewas calculated from the median numberof Rif colonies and the mean number of viable cells in the culture (LEA and Cou1.30~1949). RESULTS
Isolation of a recfiuppressor strain: The recJ77 allele is a strong allele of recJ ( LOVETTand CLARK1984). When present in recB21 recC22 sbcA23 genetic background, recJ77 causes an approximate30-fold reduction in conjugational recombination frequencies, extreme sensitivity to W light and Mitomycin C (MitoC) and partial inviability. We selected spontaneous MitoC‘ derivatives and screened these for concomitantresistance to UV and elevated recombination frequencies in Hfr crosses. We obtained such “revertants” at frequencies of -5 X 10”’ for the recJ77strain. One such derivative, STL1051, exhibited partial alleviation ofthe recJmutant phenotype (Table 2 ) .A transductional backcross using P1 grown on STL1051 and a lysA recBC sbcA recipient JC12159 revealed that STL1051 still carries the recJ77 allele cotransducing with lysA:16 Rec- W s among 86 Lys’ transductants. The suppressed phenotype must therefore be due to a second-site suppressor mutation, which we call srj (for “suppressor of ref’). In fact, as we show below, three loci are requiredfor the full suppressive effect. Two of the srj loci are alleles of the DNA helicase genes, helD and uvrD, but the identity of the third locus, srjD is not yet known. Allelenonspecificityofsuppression: Several insertion recJalleles wereintroduced by P1 transduction into the srj suppressor strain. Three insertion mutations, recJ284::T n 10, recJ2003:Tn 1@9 and recJ2051: :T n 10-9, were suppressed to about the same extent as that seen for recJ77allele (Table 2) . The same was seen for other recJ alleles, recJ153 and recJ154 (data not shown). The locations of these recJ mutations were determined by DNA sequence analysis. The recJ77 allele is caused by a transition mutation GGT (Gly) to GAT (Asp) at codon 78 of the redgene. The recJ284 insertion is 434 nucleo-
V. A. Sutera
tides downstream of the presumed start codon, whereas recJ2003 insertion is more N-terminal, 124 nucleotides downstream of the start. The recJ2051 insertion is quite C-terminal, 1252 nucleotides downstream of the start. The fact that all alleles of re$ including insertion mutations, can be suppressed in this genetic background suggests that suppression results from “bypass” of recJ function. We can not, however, rule out the possibility that truncated forms of RecJ protein from these insertion mutants retain partial activity and that thesuppression enhances the effectiveness ofthe mutantproteins. Mapping s r j loci-evidence for two cosuppressors: Various Hfr crosses were performed with STL1051 to map the suppressor loci. Inheritance of the srj+ allele from the donor should remove suppression and cause the strain to become W s and Rec-. Inheritance patterns were complex with two different regions of the E. coli chromosome affecting suppression: one between gal and his (which we initially called srjB and will show is helD) and the second (which we initially called srjC and is, in fact, uvrD) near argE (data not shown). P1 transductional crosses established a position for srjB at 22 min (Table 3 ) . Markers closest to srjB were inherited mostpoorly in these crosses (datanot shown) -this may be due to dominance of srjB’ over srjB in the transductant,reversing the suppressive effect and lowering recombinational inheritance. Coinheritance between srjB and other markers was somewhat less than expected from the genetic map. This may be explained by the poor recovery of srj+ recombinants due to their reduced viability relative to srjB- transductants or possibly reestablishment of a srjB suppressor mutation in the transductants. Other transductional crosses located mjCnear 86 min on the E. coli chromosome (Table 4 ) . Again, srj+ derivatives showed poor viability and may be underrepresented among the transductants. These mapping data reveal the location of two mutations, srjB and srjC, both required for suppression of recJin strain STL1051. The low frequency at which such suppressors were isolated is consistent with this hypothesis. We note that increased genetic instability conferred by one of the mutations (see below) may have facilitated the accumulation of the second cosuppressor mutation. Also, further analysis (see below) suggests yetanother cosuppressor mutation (“srjD7”) is required to see the full suppressive effect. Identification of hem, the gene for helicase IV, as wjB: Fragments corresponding to the 22 min region were cloned from the ordered Kohara lambda library (KOHARAet al. 1987) into plasmids and screened for complementation of the Srj phenotype in strain STL1051. One plasmidpSTL34, derived from phage 223, carried a 3.2-kb PstI BamHI fragmentand was shown to convert STL1051 to W sensitivity comparable with nonsuppressed recJ strains (Figure 1) . This plasmid therefore carried the srjB’ allele, which, at least
recJ Suppression
33
TABLE 2 recj-allele specificity of srj-reJ suppression
Strain JC8679 JC13018 STL1051" JC13015 STLl688 RDK2 144 STL1036 STL1028 STL1029
W survival
Relative RF"
Genotype recB21 recC22 sbcA23 0.0053 recB21 recC22 sbcA23 recJ77 0.37 recB2l recC22 sbcA23 recJ77 helD104 uvrD517,, srjD7 recB21 recC22 sbcA23 reg284 ::Tn 10 0.23 recB2l recC22 sbcA23 recJ284::Tn 10 helDlO4 uwD517.,,,srjD7 recB21 recC22 sbcA23 recJ2003: :Tn 1 0-9 0.089 recB2l recC22 sbcA23 recJ2003: :Tnl0-9 helD104 uwD517,,,,srjD7 recB2l recC22 sbcA23 recJ2051: :Tn 10-9 0.13 recB2l recC22 sbcA23 recJ2051: :Tnl0-9 helD104 uvrD517,,srjD7
(%Ib
(1) 0.035 0.24 0.01 1 0.30 0.01 1 0.18 0.026 0.39
15
.E i?l
0.00001
0.000011 -
0.00001
0
10
UV dose ( J / d )
20
0
10
uv dose (J/&)
20
0
10
20
UV dose ( J / d )
FIGURE4.-Effects of various helicase mutations uwD254::Tn5, recQ61::Tn3 and he1D::TnlO-9 in combination on UV survival. Helicase mutation combinations in ret+ ( A ) , recBC sbcA ( B ) , and recBC sbcBC ( C ) genetic backgrounds include hP (~),h~D(A),recQ((.),uwD(A),helDrecQ(4),helDuwD(O),recQuvrD(O),helDrecQuvrD(O).Datashownarethe averages of two to three determinations. These strains are listed in Table 10 and their derivations are given in Table 1.
by helD mutations in recBC sbcA strains (although no had little effect in recBCsbcA strains. These synergistic effect on recBC sbcBC strains) and a synergistic effect on genetic effects illustrate a measure of overlapping funcrecombination by mutations in helD and UWD in the tion of these DNA helicases for W repair. Our analysis recBCsbcBC genetic background ( 100-fold reduction). did not show evidence for overlapping function in geOther of our conjugal crosses with Hfrs with differing netic recombination. origins of transfer including those that transfer the heZD The recJ284: :TnlOallele was introduced into the variregion very late have failed to reveal anydefect in conjuous combinations of helicase mutants in the recBrecC gation recombination in heZD strains (data not shown). The explanation for this discrepancy is not known but TABLE 11 may include allele differences, extragenic mutations or Conjugational recombination of UV survival of helicase experimental differences. e @in combination with mutations in heD, uvrD and r The most striking effect seen in our analysis was the recJ in re& recC sbcA genetic background lesser effect of uvrD mutations on W survival in recBC uv sbcA or recBCsbcB genetic backgrounds as compared Genotype: recB2l red22 sbcA23 Relative survival with the more extreme reductionof W survival in rec+ red284 : :Tn 1O+ RF (%I" genetic background (Figure 4). This suppression of the UV sensitivity in the former genetic backgrounds is 0.020 0.0036 hel' helD102::TnI0-9 0.0018 0.035 due to the action of re@: mutations in UWD and recQ uvrD254: :Tn5 0.024 0.0024 were strongly synergistic in both recBCsbcA and recBC 0.029 helD102::Tnl0-9 uwD254::Tn5 0.00091 sbcB strains, reducing UV survival to thelevels conferred 0.15 recQ61: :Tn3 0.047 by uvrD alone in the ret+ background. This backgroundrecQ61::Tn3 helDlO2: :Tn 10-9 0.018 0.014 specific synergism may be because recQ an DNAdam~ecQ61: :Tn3 uwD254::Tn5 0.19 0.0052 age inducible gene ( IRINOet aZ. 1986), is expressed recQ61: :Tn3 helDlO2: :Tn 10-9 constitutively in recBCsbcA and recBCsbcBC strains; in 0.0048 0.034 uwD254: :Tn5 recBC+ strains, levels of recQ expression may not be sufIsogenic helicase mutants in the recBC sbcA genetic backficient to 'Ompensate for loss Of uvrD. Higher levels Of ground are described in Table 1. Swains include Jc1.3015, expression of DNAdamage inducible genes have been STL814,STL1743,STL1745, STLl742, STLl790, STLl744, observed in recBC mutant strain backgrounds ( M U STLl789. Matings wereperformed for 0.5 hr. Selections were for Thr' Leu+ [Ser' Sm'] with Hfr donor JC158. Inheritance and BELK1982; LLOYDet al. 1983) . An additional mutafrequencies are expressed relative to thatdetermined for tion in helD also enhanced somewhat the sensitivityof JC8679 (recBC s b d ) performed in parallel, which was 1-4%. recQmutants in these backgrounds' An mutaBoth recombination and W survival values are averages for tion in helD enhancedthe UV sensitivity of UWD mutwo to three experiments. tants in the recBC sbcBCgenetic background slightly but "At 10 J/m2.
recJ Suppression
s b d genetic background. Mutations in UWDand helD did not confer any suppression of recJphenotypic effects in tests of recombinational proficiency and UV survival (Table 11 ) .This supportsour observations in the analysis of the uvrD heZD mjsuppressor mutations, that some residual function of UvrD is required for RecJ suppression. Our data confirmed the results of KUSANO et al. (1994) that inactivation of the recQ helicase partially suppresses the recombination deficiency and UV sensi-
45
tivity conferred byrecJ mutations. This suppression is neither enhanced nor inhibited by introduction of mutations in UWD.However, suppression ofrecJbyre@ required a functional helD gene. This distinguishes the recQmodeof suppression (helicase IV dependent) from the helD UWDsrjD suppression (helicase IV independent) andsuggests that, although bothtypes ofsuppression involve DNA helicase genes, they are mechanistically distinct.
