VIROLOGY
77,
306-318
(19771
Isolation and Characterization of Specialized Transducing Bacteriophages for the recA Gene of Escherichia co/i KEVIN Department
of Biophysics
and
McENTEE Theoretical
Biology Chicago, Accepted
WOLFGANG
AND
and Department Illinois 60637 October
EPSTEIN
of Biochemistry,
University
of Chicago,
29,1976
Lysogens ofEscherichia coli have been isolated in which hc1857 is integrated near the recA locus in a gene required for sorbitol utilization (srl). These secondary-site lysogens produce lysates containing plaque-forming specialized transducing phages which carry part of the srl locus and the recA gene. Genetic evidence, DNA heteroduplex examination, and restriction enzyme analysis demonstrate that the bacterial DNA substitution occurs in the b2 region of the transducing i\. One such recA transducing phage, designated AprecA, contains approximately 3% less DNA than the parental A.
hprecA (McEntee, 1976). This phage was obtained from a strain in which bacteriophage A is integrated near the recA gene in a locus (srl) affecting sorbitol utilization. In this paper, we describe the isolation of lysogens in which he1857 is integrated into the srl locus and present a biophysical characterization of the XprecA phage, which carries part of the srl locus in addition to the recA gene.
INTRODUCTION
The product of the recA gene of Escherichia coli performs an important role in several cellular processes. Homologous recombination in recA mutants is reduced to less than 0.001% of that in recA+ strains (Clark, 1973). Furthermore, recA- mutants are exceedingly sensitive to killing by uv or X irradiation and alkylating agents such as mitomycin C (Clark and Margulies, 1965). Repair of DNA is impaired in these mutants and much of the damaged DNA is hydrolyzed to acid-soluble material (Clark et al., 1966). In recAmutants, uv treatment neither produces mutations (Miura and Tomizawa, 1968) nor induces prophages (Hertman and Luria, 1967). These pleiotropic properties result from mutations in a single gene near min 58 on the revised linkage map of E. coli (Bachmann et aZ., 1976). The isolation of amber (Mount, 1971) and temperaturesensitive (Lloyd et al., 1974) recA- mutations indicates that the gene product is a protein. Although much information exists on the physiology and genetics of recA- mutants, nothing is known about the biochemical properties and function of the recA gene product. To identify the recA gene product and study it biochemically, we have isolated a specialized recA transducing phage,
MATERIALS
AND
METHODS
Bacterial strains and bacteriophages. The strains and phages used in this study are listed in Table 1. Bacteriological media. KML complex medium and K115 minimal medium have been described (Epstein and Kim, 1971). Solid minimal medium contains 10 g of the appropriate sugar per liter, except that sorbitol (Pfanstiehl) minimal medium contains 3 g/liter. Nitrofurantoin (Sigma) plates contain 2 mg of drug/liter of KML medium. Indicator medium consists of 10 g of the appropriate sugar and 40 g of MacConkey agar base (Difco) per liter. Isolation of A inserted into the srl locus. A slight variation of the method of Shimada et al. (1972) was employed to isolate lysogens with A inserted in the srl region. Strain KM2134 was grown to a density of 7 x 10’ cells/ml in KML medium supple306
Copyright All rights
0 1977 by Academic Press, Inc. of reproduction in any form reserved.
ISSN
0042-6822
recA
TRANSDUCING TABLE
BACTERIALAND
1
BACTERIOPHAGE
Relevant Strains” AB2462 AB2463 KM100 KM2134 C6OOS’ KL16-99 KM197 KM21352146
307
PHAGES
STRAINS
genotype
Sourceb
KLF8/MA50 KA197 KM3000 KM601
F-pro arg his thr leu recAl2 supE strA thi F- pro arg his thr leu recAl3 supE strA thi F- srl-1 cysC 1ysA argA mtl strA thi F- mtl 1ysA argA thi (gal-atth-bio) deletion F- thr leu lac tonA supE thi strA Hfr KL16 recA1 thi F- srl-1 thyA pheA cysC thi mtl nalA KM2134 lysogenized at srl; at least several are independent tyrA pyrD thi his trp thyA recA1 mtl xyl malA A' galK strAIF143 Same as KLF43iKL259, but with the F143-1 episome pyrB argG metB 1ysA his leu mtl xyl malA A’ galK lac tonA supEiF160 thr leu lac mtl thi ma1 A' cysD lysAlF108 Hfr KL16 thi pheA KM2136(A at srl) (hpgal8)lFB F- thr leu recA1 strA tonA lac supE
KM2199 KM155 H560
F- mtl recA99 (gal-atth-bio) E. coli C his (P2) F+ endol polA strA
KLF43/KL259 KLF43-A/KL259 NF306
Phages hcI857 hcI857susS7 AC1 hcIh80 hpgal8 Apro Aimm434b515b519
See Table
deletion
K. B. Low
CGSC CGSC CGSC CGSC CGSC KL16-99 x thyA of C6OO.S’
R. M. R. R. suppressors
and are
is from the laboratory University.
mented with biotin (0.1 mg/liter) and maltose (4 g/liter). The cells were harvested, washed once with 0.17 M NaCl, and resuspended in 0.2 vol of 10 mM MgSO,. After shaking at 37” for 60 min, the cells were infected with A~1857 at a multiplicity of approximately 8. The suspension was chilled on ice for 10 min and then incubated with agitation for 45 min at 30”. The cells were diluted into KML medium and, after several generations of growth, were spread on sorbitol MacConkey plates which had been seeded with hc1 and hcIh80 (approximately log of each phage). After 2-3 days incubation at 30”, the plates were
mutant
strA
4
a All bacterial strains lack nonsense wise indicated. @ Where no source is shown, the strain c Coli Genetics Stock Center at Yale
CGSC’ CGSC
sensitive collection
Gayda E. Gottesman Weisberg Weisberg
to streptomycin or was constructed
and
A unless for this
otherwork.
examined and white colonies were picked to sorbitol and to glucose indicator plates. Colonies which gave a negative sorbitol indicator reaction, but gave a positive glucose reaction, were saved for further testing. Sorbitol selection and scoring. The linkage of recA to srl-1 was determined by Plkc transduction using srl+ as the selected marker. Misleading results can be obtained in these crosses when selecting or scoring the srl marker. Strains which are mtl+ srl- will grow on sorbitol, although more slowly than srl+ strains. This problem can be avoided by using mtl- strains
308
McENTEE AND EPSTEIN
as Plkc donors. When this is not possible, the distinction between mtl+ srl- and srl+ can be made on MacConkey indicator plates. The former class of recombinants is white or pink on sorbitol MacConkey medium, while the latter appear red. Another difficulty in the selection is cross-feeding of the srl~ recipients by srl+ recombinants. In addition, the frequency of unstable heterozygotes obtained in the srl+ selection was unusually high and contributed to the low linkage measurements obtained in early experiments. Upon repurification to obtain pure srl+ recombinants, Plkc cotransduction of srl-1 with several recA alleles ranges from 70 to 90% (see Results). Determination of prophage orientation. The orientation of the prophage was determined by chromosome mobilization. Strains KM21352146 are gal- because of the gal-atth-bio deletion. These A lysogens were transduced to gal+ with Apgal8. Because both gal homology and attA are deleted in these recipients, gal+ transductants result from homologous recombination between Apgal8 and the prophage. The location of Apgal8 at srl was confirmed by Plkc transduction of the gal+ derivatives to srl+. All srl+ transductants (80/80) lost the ability to ferment galactose. To mobilize the chromosome from this transposed gal region, the F8 episome was introduced. The resulting F’ strains are capable of mobilizing chromosomal markers from two separate origins: One of these sites is determined by homology of the gal region of F8 with Apgal8, while the other is determined by residual homology of F8 with the chromosome near nadA. Mobilization of markers near srl was tested by selecting for phe+ nalA and thy+ nalA recombinants in 30-min matings at 30” with strain KM197. Chromosome mobilization by the F8 episome in Apgal8 transductants is a general method for determining the orientation of A at unusual prophage sites not near gal. Other genetic techniques. Transduction with Plkc and isolation of A’ strains were performed as described by Miller (1972). Matings were performed by mixing 0.2 ml of a recipient culture in KML medium containing 4 x 10” cells/ml with an equal
volume of donor culture at the same cell density, incubating without agitation at 30 for a sufficient length of time (usually 30 to 60 min) to obtain the desired recombinants, and plating serial dilutions on selective medium. To isolate cured derivatives of strains with A integrated at srl, the strains were made A’ to prevent killing by reinfection. The Ar derivatives were grown in KML medium at 30” to log cells/ml and volumes of 1 ml were spread on sorbitol MacConkey plates incubated at 42”. The EDTA sensitivity of phages was tested as described by Shimada et al. (1973). Measurements of lysogenization frequency were performed as described by Shimada et al. (1975). UV survival measurements were done as previously described (McEntee, 1976). Heteroduplex mapping. Phage were purified in CsCl step density gradients. Disruption of the phage, denaturation of the DNA, formamide spreading, and shadowing were done essentially as described by Davis et al. (1971), except that for spreading the hypophase was 17% formamide and the hyperphase was 50% formamide. An RCA EMU-4 electron microscope was used to photograph heteroduplex molecules. +X174 RF molecules were added to all preparations. Molecules were traced at a final magnification of at least 200,000 and measured with a Keuffel and Esser map measurer. The locations of the b515 and 6519 deletions, att, and the imm434 substitution are from Davidson and Szybalski (1971). EcoRI restriction enzyme analysis. CsCl purified phage were disrupted by dialysis against a solution containing 50% formamide, 100 mM Tris-Cl (pH 8.5), and 10 mM EDTA for 16 hr at room temperature followed by dialysis against a solution containing 100 miJ4 Tris-Cl (pH 7.5), 0.1 mM EDTA overnight at 4”. Phage DNA (2 pug) was digested at 37” for 2 hr in a total mixture (0.1 ml) which contained 6 mM Tris-Cl (pH 7.8), 6 mM MgCl*, 1 mM dithiothreitol, and 3 units of EcoRI endonuclease. The exonuclease-free enzyme was the gift of K. Agarwal. The reaction was terminated by addition of EDTA to 50 mM and heating for 2 min at 80”. Each sample
recA
TRANSDUCING
was made 20% (w/v) in glycerol and 0.02% (w/v) in bromophenol blue. Electrophoresis in 0.7% agarose slab gels (20 x 20 cm) was done at 4” for 16 hr with 40 mM Trisacetate (pH 7.6) buffer at an applied field of approximately 2 V/cm. Gels were stained for 20 min in a solution of 1 pg/ml of ethidium bromide and photographed under uv light. The molecular weight of the restriction fragment unique to AprecA (fragment R) was calculated from the mobilities of h restriction fragments of known molecular weight (Thomas and Davis, 1975). RESULTS
A Insertion
into the srl Locus
Several mutants defective in sorbitol metabolism were isolated and characterized in this laboratory (W. Epstein and E. Jackson, unpublished observations). One of these mutations, d-1, was located at min 58 on the E. coli genetic map (Bachmann et al., 1976). The gene order cysCsrl-recA is deduced from three-factor transductional crosses (Table 2, Fig. 1). The donor recA allele is inherited at high frequency when srl+ is the selected trait (Table 2). In several experiments, cotransduction frequencies of 70-95% have been observed for several recA mutations and srl-1 in this laboratory and in the laboratory of A. J. Clark (A. J. Clark and M. Kotewicz, personal communication). We have exploited this high linkage between recA and d-1 to isolate a recA transducing phage from a secondary-site lysogen in which A is integrated into the srl locus. Strain KM2134 was lysogenized with A (see Materials and Methods). Approximately 1 lysogen in 2000 did not ferment sorbitol. Because sorbitol is a substrate for the phosphoenolpyruvate-de-
Donor
AB2462 AB2462 KL16-99
(recAl2) (recA1)
KM100 KM100 KM100
Plkc
COTRANSDUCTION
Recipient
Selected marker
(cysC (cysC (cyst
d-1 1 d-1 1 srl-1)
cyst+ srl+ srl+
309
PHAGES
pendent phosphotransferase system (PTS) and A is known to integrate frequently into the pts locus (Shimada et al., 19731, it was expected that many phenotypically sorbito1 negative (Srll) lysogens would be pts mutants. The Sri- lysogens were scored on glucose MacConkey medium to distinguish pts- mutants from the true srl- mutants. Approximately 5% of the lysogens tested were positive for growth on glucose (pts+). In three experiments, a total of 12 srl- mutants was obtained. Six of these lysogens were arbitrarily selected for further study. Three genetic tests were performed to map the sorbitol mutation of these lysogens and to demonstrate that the sorbitol defect is the result of A insertion into the srl locus: (i) Complementation tests were performed with four episomes which extend from 1ysA to different points in the cysC-pheA region. All lysogens tested were complemented for growth on sorbitol by F143 and F108 but not by the F143-1 or F160 episomes (Fig. 1). This complementation pattern is identical to that observed with the srl-1 point mutation. (ii) Hfr strain KA197 was mated with three of the srl-1 lysogens. This Hfr strain transfers the srl region early; all (77177) srl+ strrecombinants obtained in a 30-min mating at 30” had lost A immunity and temperature sensitivity. (iii) The srl- mutation and A immunity are 100% linked in transductional crosses as well. When a mtl~ srl+ strain was used as donor, all (3201320) srl+ recombinants had lost A immunity and were capable of growth at high temperature. However, when Plkc grown on a mtl+ srl+ strain was used, a fraction of the Sri’ transductants retained the prophage. All of these A-immune transductants were shown to be mtl+ srl- by testing on indica-
TABLE 2 OFCYSC,
d-l,
AND recA
Number scored
160 75 64
Recombinants srlrecA -
srlrecA +
srl+ recA -
sr1+ recA L
0
149
8 56 57
3 19 7
310
McENTEE
tor media (see Materials Orientation
AND
EPSTEIN
and Methods).
pheA + recombinants per 0.1 ml of mating mixture were obtained, but no thy+ recombinants were obtained. Spot matings confirmed that cysC+ is not transferred early in crosses with strain KM3000, consistent with the order shown in Fig. 2.
of h in the srl Locus
Chromosome mobilization by the F8 episome in a hpgal8 transductant of strain KM2137 was used to determine the orientation of the resident prophage in the strain (see Materials and Methods). The prophage orientation shown in Fig. 2 predicts that FS will transfer as early markers pheA and other genes counterclockwise to the phage. In a 30-min mating at 30” between strain KM3000 and thyA- pheArecipient strain KM197, approximately 300
--
IysA \
IhyA argA 1 /
61
60
cyst
srl
red \/
In 59
56
--077
tyrA , 57
isolation of srl and Phages
pheA
y 56--
F 143
F 143-I F 160
______ - _____ - - - - - - ---
Transducing
Thermal induction of strain KM2136 produces low-titer lysates, approximately lop4 phagellysogen. Phages capable of transducing srl were detected among plaque-forming phage by plating the lysate on strain KMlOO. Two hundred forty turbid plaques on this host were picked and cloned on sorbitol MacConkey plates seeded with XCI. Fifty-two independent lysogens were Srl+ as judged by a positive indicator reaction as well as by their ability to grow on sorbitol minimal medium. Three of the Srl+ lysogens were grown in liquid. Following thermal induction, each of the lysogens produced plaque-forming transducing phages which complement the srl-1 mutation of strain KMlOO. Typically, lo-20 transducing phages were released per lysogenic cell. Specialized recA transducing derivatives were identified among the plaqueforming phage by picking turbid plaques formed on several recA- hosts (AB2462, AB2463, and KM6011 and testing for
- 0.81
F IO6
recA
_--
FIG. 1. A genetic map of the recA region of E. c&i. The gene locations and minute scale are from the revised E. coli linkage map (Bachmann et al., 1976). Linkage data shown are from Table 2. The length of each of the episomes is based upon complementation of the markers shown; dashed segments represent uncertainty in the extent of the episome. The representation of F143-1 is based upon the failure of this episome to complement d-1, recA, pheA, and tyrA mutations.
XprecA +
4
thyA argA cysC b2 J ”
A R cl ‘,,’ In\
IhyA orgA cysC
J b2 J
A R cl int
.--------F8
SJ ,_.~_..~
srl-I recA pheA I”+ ’ -f”
srl-I recA pheA / A R cl I”+ , . . . . . . ~.... _,-.._i ..__._.
;
FIG. 2. Orientation of h integrated at srl and mobilization of represents the orientation of the prophage at srl. Connected arrows which generate hprecA. The formation of a double lysogen incorporating tion by FS of chromosomal markers to the right of the prophages, are including the location and orientation of gal on this phage has been orientation of the gal region on the F8 episome is from Low (1972).
chromosomal markers. The top line at the top indicate the sites of cutting hpgul8, and the resultant mobilizashown below. The structure of hpguld reported by Nissley et al. (1971). The
recA
TRANSDUCING
growth in the presence of the radiomimetic agent nitrofurantoin (2 pg/ml). Approximately 1% of the turbid centers tested produced colonies on the nitrofurantoin medium. Several lysogens isolated in this way were tested for sensitivity to killing by uv irradiation. All were found to be uv resistant and recombination proficient in transductional crosses (McEntee, 1976) and in conjugal crosses with Hfr strains which do not transfer the recA+ gene. Nine independent Ret+ lysogens were grown in liquid and thermally induced; each of the strains produced high-frequency transducing lysates (phage titer of log/ml) which conferred recombination proficiency and resistance to uv to recAmutants. No segregation of recombination proficiency and uv resistance was detected among several dozen phages tested from these lysates. Both uv resistance and recombination ability were lost simultaneously upon curing the recA- lysogen at 42”. A second screening method, based upon the ability of the phages to restore recombination to recA- hosts, was used to isolate four additional recA transducing derivatives, The low-frequency transducing lysate was plated on strain AB2462 and lysogens of this recA- mutant were tested for the ability to form thr+ leu+ proi str- recombinants in conjugal crosses. Each of TABLE RESTORATION
OF uv
Host
RESISTANCE
AND
Relevant
RECOMBINATION
311
PHAGES
the four Ret+ lysogens isolated by this technique produces phage which restore uv and nitrofurantoin resistance to recA strains in addition to complementing the recombination defect. This result is consistent with the notion that the recombination and repair functions of recA are due to the expression of a single gene. Isolation of Amber Phage hprecA99
3 BY AprecA99
IN
sup+ supE supE
(B) Recombinationb KL 16-99
recA1
sup+
AB2462
recAl2
supE
SUPPRESSING
recA
STRAINY
Prophage AcI857
recA1 recAl2 recA1
Transducing
Mount (1971) has described a mutation in the recA gene which is suppressed by nonsense suppressors. Strain amber KM2199, which carries the recA99 mutation, was lysogenized with hcI857susS7 and a Sri- lysogen was obtained as previously described. From this strain, a AprecA99 transducing phage was isolated. This phage complements the uv sensitivity and recombination deficiency of recAstrains which carry amber suppressors (AB2462, AB2463, KM601) but fails to complement the recA- defects of sup+ recA- strains (Table 3). All phage isolated which transduce the recA gene, representing at least four independent isolations, complement the srl-1 mutation of strain KMlOO. In the remainder of this paper, we present a genetic and biophysical characterization of one deriva-
genotype
(A) uv resistancea KL 16-99 AB2462 KM601
recA
Survival after 4 x 10-G 5 x 10-S 3 x 10-S
0 0
ADred
hored 200
erglmm’ of irradiation 7 x 10-S 0.59 0.40 0.45 0.20 0.32
bgl recombinantsIlO 5 pro+ recombinantsIlO 159
Plkc 469 Plkc 132
n The recA host lysogenic for the indicated phage was grown in K115 medium containing 50 mg of required amino acids per liter and irradiated as described under Materials and Methods. The dose delivered was approximately 200 erg/mmz. Results are presented as a fraction of the initial viable cell titer that survived irradiation. b The recAhost lysogenic for the indicated phage was grown in KML medium at 30”, infected with Plkc at a multiplicity of infection between 0.1 and 0.3, plated on appropriate selective medium, and incubated at 30”.
312
McENTEE
tive of he1857 designated hprecA, transduces the recA + gene. Site of Lysogenization by hprecA
AND
which
Specialized transducing phages carrying hybrid attachment sites (AOP’ or POA’) may insert into the host chromosome via homologous recombination dependent upon the function of the host recA gene or by int promoted site-specific recombination at the bacterial attA site. Phage integration by homologous recombination depends upon the extent of homology between the transducing variant and the bacterial chromosome, while int-dependent insertion of the phage depends upon the structure of the particular hybrid attachment site carried by the transducing derivative (Shimada et al., 1975). Three lines of evidence suggest that the principal site of AprecA insertion is at or near atth: (1) A AprecA lysogen of gal- recAstrain AB2462 was transduced to gal + with PMc grown on strain KMlOO. Among 160 gal+ recombinants, seven (4%) had lost A immunity and regained the uv sensitivity of the parental recA- strain. Since attA is cotransduced with gal by Plkc, this result is consistent with insertion of XprecA at or near at& (2) Evidence which argues against a prophage location in the recA region of the E. coli chromosome was obtained in conjugal crosses between a AprecA lysogen of strain KL16-99 and recipient strain KMlOO. Selecting cysC+ str- recombinants results in 37% coinheritance of the donor recA- mutation. This linkage is identical to that observed when the nonlysogenic donor strain KL16-99 is used. The linkage value is not altered when the recipient is lyso-
EPSTEIN
genized with A to prevent zygotic induction. The normal transmission of the donor recA- allele and the failure to detect zygotic induction in matings of 30-min duration indicate that the prophage is not transferred as an early marker in conjugal crosses with AprecA lysogens of strain KL16-99. Since the recA gene is transferred as an early marker by this Hfr strain, we conclude that AprecA does not lysogenize by homologous recombination in the recA region of this strain. (3) The frequency with which AprecA lysogenizes a host is markedly affected by the presence of the bacterial atth region. Infection of strain KM100 (attA+) by AprecA results in 27% lysogenization. Under identical conditions (see Methods), A infection yields approximately 50% lysoFor atth deletion strain genization. KM2134, the lysogenization frequencies are 0.4 and 0.2% for AprecA and A, respectively. Thus AprecA, like A, shows significantly lower lysogenization efficiency when the atth site is deleted. This evidence strongly suggests that AprecA lysogenizes in the atth region of the E. coli chromosome. Structure
TABLE TEST
Phage recA + (KM1001 AC1857 hpFOh
AprecA
(1.0) (1.0) (1.0)
4 OF ApprecA
OF SPI PHENOTYPE
Host
of AprecA
The observation that AprecA lysogenizes at or near attA implies that the int gene of the phage is intact. To test for loss of genes in the early region of the phage, the plating efficiency of AprecA on a p&A- strain and on a P2 lysogenic host was examined (Table 4). The plating efficiency of AprecA is similar to Ac1857; it is phenotypically Spi+ and plates with normal efficiency on a polA1 mutant, indicating that the gam
(efficiency
of plating
1.0 CO.002 0.8 which
to recA+)”
P2 lysogen
recA
(AB2462)
* Relative plating efficiency was determined at 37”. h The hpro is a plaque-forming pro transducing phage gam, i.e., Spi- (R. Weisberg, personal communication).
relative
is deleted
(KM1551
polA (H560)