- Email: [email protected]
The uvrB gene of Escherichia coli has both lexA-repressed and lexA-independent promoters
Cell, Vol. 28. 523-530,
March
1982,
Copyright
0 1982
by MIT
The uvrB Gene of Escherichia coli Has Both IexA -Repressed and IexA -Independent Promoters Gwendolyn B. Sancar,* Aziz Sancar, * John W. Little+ and W. Dean Rupp* * Department of Therapeutic Radiology and Department of Molecular Biophysics and Biochemistry Yale University School of Medicine New Haven, Connecticut 06510 + Department of Molecular and Medical Microbiology University of Arizona College of Medicine Tucson, Arizona 65724
Summary We have found that the uvrB gene of Escherichia coli is transcribed from at least two promoters, which we call Pl and P2. Transcription from Pl begins with an ATP at + 1, and transcription from P2 begins primarily with a GTP at -31. A binding site for the /exA protein (LEXA), located between the -35 sequence and Pribnow box of P2, regulates transcription from this promoter. In vitro, LEXA inhibits transcription from P2 but has no detectable effect on transcription from Pl. A third promoter, P3, was also detected at -341; transcription from P3 is toward uvr6 but terminates in vitro in the region of the LEXA binding site. The binding of LEXA to P2 inhibits transcription from the P3 promoter even though several hundred nucleotides separate the two promoters. The data suggest that a transcribing RNA polymerase stalls when it reaches the repressor-operator complex but remains bound to the DNA, causing a jamming of RNA polymerases between P3 and the repressor-operator complex at P2. The physiological significance of P3 is unknown. Introduction Damage to DNA in E. coli by far ultraviolet light (200300 nm) induces a variety of cellular responses, including filamentation, increased radioresistance and mutability, some of which help repair the damage or ensure survival of the cell (for review, see Witkin, 1976). The coordinated expression of these functions has been called the “SOS response,” while the term SOS genes is used to designate the genetic loci that are induced under these conditions (Radman, 1974; Witkin, 1976). Included among the SOS genes are /exA, recA (postreplication repair and recombination), sfiA (cell division), urn& (induced mutagenesis) and a number of din genes whose functions are not known (Witkin, 1976; Kenyon and Walker, 1960; Huisman and D’Ari, 1981; Bagg et al., 1981). One property SOS genes have in common is that their expression is controlled by the recA and IexA genes (for review see Witkin, 1976). The essential features of the interaction between the recA and IexA
genes have been described (Roberts et al., 1977; Brent and Ptashne, 1980 and 1981; Craig and Roberts, 1980; Little et al., 1980 and 1981). The LEXA protein is the repressor of both the IexA and recA genes; under normal conditions neither gene is expressed at high levels. It is believed that DNA damage results in the production of a signal, such as oligonucleotides or single-stranded DNA, which activates the protease activity of the endogenous recA protein to cleave LEXA, activating both recA and IexA transcription. Following repair and recovery, the absence of an appropriate activator results in the loss of recA protease activity, causing the accumulation of LEXA and, consequently, the reestablishment of repression. Presumably other SOS genes are also repressed by LEXA, although this has not been proved directly. In addition to the SOS genes listed above, evidence has been presented that the nucleotide excision repair genes uvrA (Kacinski et al., 1981; Kenyon and Walker, 1981) and uvrB (Fogliano and Schendel, 1981) are also induced during the SOS response and thus may be part of the recA-IexA regulon. These results were unexpected as it had been reported previously that mutations in the recA gene do not abolish the activity of the excision repair system (Howard-Flanders and Boyce, 1966). In an attempt to resolve this paradox, we have undertaken a detailed study of the nucleotide sequence organization and expression of the excision repair genes. In this paper we present the results of our analysis of the uvrB gene. Results Restriction Map of the uvrB Promoter-Operator Region We reported previously that uvrB is a monocistronic gene approximately 2 kb in length, which codes for a protein of molecular weight 84,000 (Sancar et al., 1981). We showed that the uvrB gene is inactivated when cut with Eco RI and presented evidence that the Eco RI site is very near the translational start of the gene. To obtain a detailed restriction map of the promoter-operator region of uvrB, we isolated several restriction fragments that carry the Eco RI site of the gene. Among these a 540 bp Bst NI fragment was used for the experiments reported here. The location of the 540 bp Bst NI fragment relative to uvr6 on the plasmid pDR1494 (Sancar et al., 1981) and the recognition sites for various restriction endonucleases on the 540 bp Bst NI fragment are shown in Figure 1. Sequence of the uvrB Promoter-Operator Region To find the uvrB promoter we determined the sequence of the entire 540 bp Bst NI fragment with the exception of about 10 nucleotides at each end (see Figure 1 for the sequencing strategy used). The sequence, shown in Figure 2, contains a number of
Ceil
524
A+T-rich regions and several potential promoter regions. Of particular interest is a region that is similar to the operator-promoter regions of the ret and IexA genes. The sequence CAGTATAATT beginning at -46 is homologous to the recA sequence from - 16 to -7; in the recA gene this sequence contains the
Pribnow box (Sancar et al., 1980). in addition, the sequence of the 540 bp Bst Ni fragment from -59 to -44 is similar to the consensus sequence CTGTAT...C...CAG of the LEXA binding sites in the recA and IexA genes (Little et al., 1981). These observations suggest that the region around -50 contains a LEXA-controlled operator of the uvrB gene.
LEXA Footprinting To determine whether LEXA interacts with the putative LEXA binding site in the 540 bp Bst Ni fragment, we incubated end-labeled DNA from the 540 bp Bst Ni fragment with DNAase I in the presence and absence of LEXA, then separated the products on a DNA sequencing gel (Galas and Schmitz, 1978; Johnson et al., 1979). LEXA protected, from DNAase I attack, the region between -63 and -33 on the anticoding strand and between -67 and -40 on the coding strand (Figure 3). Enhanced cutting was observed at two T residues within this region. I
540 bp
1
Figure 1. Restriction Map of pDR1494 and the 540 bp Bst NI Fragment Carrying the ovr8 Promoter, and the Strategy Used to Sequence the Fragment The vector (pBR322) is indicated by the cross-hatched line and the poiy (dA-dT) connectors with + + + + Arrows indicate the regions occupied by the uvrB gene and the distal portion of bioD as well as the direction of transcription of these genes. The single circle designates the promoter region of the uvrB gene. Numbering is relative to the Eco RI site of the vector as in Sancar et al. (1981). Arrows below the enlarged map of the 540 bp Bst NI fragment indicate the direction of sequencing and the lengths of the sequences read. Both strands of the Hae Ill-Eco RI fragment carrying the uvrB promoter were sequenced.
in Vitro Transcription of uvrB and Determination the Transcrlption Start Sites
of
Having found a LEXA binding site in the 540 bp Bst Ni fragment, we wanted to determine whether in vitro transcription of uvr8 is controlled by binding of repressor to this site. As is seen in Figure 4A, E. coli RNA poiymerase produces not only one, but three discrete transcripts with lengths of 130 bases, 160 bases and 290 bases, which we call RNA-l, RNA-2 and RNA-3, respectively. Cutting the DNAase I fragment with Eco RI shortens the length of both RNA-l 53
-400
JI GAAbACGAGGCAA6CGAGAGAATkGcGGCTTG~ACGCGAA CTTTTGTGCTCCGTTCGCTCTCTTATGCGCCGAACGTGCGCTT
GCCGCCCCCTATCCCCGACCTCTGTCAATAGGTGA
-300 -250 -200 TT~CGTTAkAGliCCGCTCA~~TATCTT~TATTTTTT~CCGCTTAGA~~TGC~T~GCAGTCACT~~CA~CAT~TCTTGCCAT;9AAACTGTCA~CACTCAT AACCGCAATTTCTGCCGAGTTTCTTTATAGAAAATAMAAA TTGGCCAATCTATTTACGTTACCGTCAGTGACTTGTCCGTAGAGAACGGTATTTTGACAGTAGTGAGTA -150 -100 CTkACAAATGT~ AAAAAAGCCeTTGCTTTGGEGATNLCCC~TAAGGCCGGlicTTTTATCT~GCCACAGAG~AAATTTTGCICATGATTGA~A~~A~T~~C~TGT GAACTGTTTACAATTTTTTCGGCAACGAAACCCCTATTGGGCCATn:CGCCCTCAAAATAGAGCGGTGTCTCATTTM~CGAGTACT~CTGTCGCCTCAAATGCGACA
t GAkTGT@TTTfATCCA CTTGACAAAAAAATAG I +50 +100 G~CTCATGAGT~CCGTTC~CTGAATTC~TTTTAITGGCC CTGAGTACTCATTTGGCAAGTTTGACTTAAGCGAAAATTTGGAAGACCCCTAGTCGGTCTCCGCTAAGCTGCAGAGCTTCTCCCCGACCTTCTACCGG Figure
2. Sequence
of the 540 bp Bst NI Fragment
Containing
the uvrS Operator-Promoter
Region
The base sequence of the 540 bp Bst NI fragment in Figure 1 is given along with the regulatory regions described in the text. The transcription initiation sites for RNA-l, RNA-2 and RNA-3 are denoted by Sl , 52 and S3 with the numbering of nucieotides relative to Sl . Pribnow boxes and -35 sequences are enclosed in boxes (dashed for Sl , solid for S2 and S3). Single bars above and below the line indtcate the region protected by LEXA from DNAase I attack: circles denote nucieotides at which cutting was enhanced in the presence of LEXA. Potential ribosome binding sites (Steitz. 1979) near ATG codons on the uvr6 message are overscored with two lines. (Since both fMet codons at +34 and +121 are in the same open reading frame, we do not know which of these sites defines the start of translation of uvri3 in viva.)
uvf6
Promoters
525
A E
G
T + C
C
DNaroI -+ IrxA IsxA
A
G T
E
E
C
DNoae I + 1w.A IexA
-IrxA
+IsxA
-lexA
+lexA
RNA-3*
-60.
RNA-2 -40. -30.
-40.
RNA- I
-30.
Figure 4. Autoradiograms of RNAs Transcribed bp Est NI Fragment in the Presence or Absence
in Vitro from the 540 of LEXA
(A) Transcripts produced In the absence of hepartn and the absence (lane 1) or presence (lane 2) of LEXA. (6) Transcripts produced In the presence of heparin and the absence (lane 1) or presence (lane 2) of LEXA. Transcriptlon, gel electrophoresis and autoradiography were performed as described in Experimental Procedures. Exposure time for the autOradiOQrafII shown In B was approximately ten times longer than for the autoradiogram shown in A.
Figure
3. LEXA Footprintmg
The 1 QB bp Hae Ill-Eco RI fragment of the 540 bp Bst NI fragment was labeled at the Eco RI site and treated with DNAase I in the presence or absence of LEXA. Fragments not treated with DNAase I were used for sequence analysis (Maxam and Gilbert, 1 QBO) and run in wells adjacent to the DNAase l-treated samples. The region protected by LEXA Is enclosed in brackets: asterlsks indicate nucleotides at which cutting by DNAase I is enhanced in the presence of LEXA. (A) Anticoding strand, 5’ end labeled. (B) COditIg strand, 3’ end labeled.
and RNA-2 by approximately 75 bases, but has no effect on RNA-3 (data not shown). Thus RNA-l and RNA-2 are runoff transcripts that originate upstream from the Eco RI site and extend into uvr8. The direction of transcription of RNA-1 and RNA-2 was confirmed, and the precise transcriptional start sites (Sl and S2) were determined by RNA sequence analysis of these transcripts (Figure 5). RNA-l is initiated at an A residue, which we have numbered + 1 (Figure 2); RNA-2 begins with a G at position -31 and at low frequency at an A at -29 (not shown) and extends through Sl . From this finding, we conclude that uvr8 is read from two separate promoters. Separate experiments showed that the 5’ end of RNA-3 lies at -341 and extends toward uvrB (Figure 6). When LEXA was added to the transcription reaction prior to the addition of polymerase. two separate
effects were observed. First, LEXA strongly inhibits synthesis of the uvrB transcript RNA-P, but not that of RNA-1 (Figure 4A). Thus, although Sl and S2 are separated by only 31 bp, transcription from their respective promoters, Pl and P2, is regulated in different manners. Second, and more surprising, transcription of RNA3 is also inhibited by LEXA. Because the sequence 5’ to 53 is not protected by LEXA (not shown), we considered the possibility that LEXA inhibited transcription of RNA-3 by an indirect mechanism. From the length of RNA-3 we would predict that transcription terminates in the region of the proposed LEXA binding site; preliminary data (not shown) indicate that termination occurs within the stretch of Ts centered at position -53. This suggested that LEXA might inhibit transcription of RNA-3 by blocking the site at which termination and release of RNA-3 is effected in the absence of LEXA, leading to accumulation on the template of stalled polymerase molecules bearing progressively shorter nascent transcripts. To test this hypothesis we transcribed the 540 bp Bst NI fragment in the presence of heparin to limit transcription to one initiation per promoter. As is seen in Figure 4B, transcription under these conditions and in the presence of LEXA results in the disappearance of RNA-3 and the appearance of a new major transcript (RNA-3’) approximately 25 nucleotides shorter than RNA-3. In a related experiment, it was found that cutting the 540 bp Bst NI fragment with Msp I (an isoschizomer of Hpa
Cell 526
-ENZ G
0”.
A
U+A
CA UU
G
OH-
A
U+A U*C
UK
A: i”
$6
GA C
t E G
EG A U
C
G A A
c c
U
G
U A
U
A c
U
A
c
c ATTP G 4
Figure
5. Nucleotide
Sequence
of the 5’ End of RNA-1
RNA-i (A) and RNA-2 (6) were labeled sequenced as described in Experimental tions were obtained with RNAase 1 (lane U2 (lane A). RNAase Phy M (lane U+A) U+C). In B, the position of r3*P-ATP. nucleotide. is shown.
and RNA-2
at their 5’ ends, purified and Procedures. Partial dlgesG). base (lane OH-). RNAase and RNAase 8. cereus (lane used as a marker for the first
% uCA noA AG AG CGi A CA G G A G
C
II) abolished the ability of LEXA to inhibit transcription from the promoter used for synthesis of RNA-3 (data not shown). These results support the conclusion that LEXA does not directly affect initiation of transcription of RNA-3. Discussion We have sequenced the operator-promoter region of the uvrB gene of E. coli and have shown that in vitro the gene is transcribed from two different promoters, which we call Pl and P2. The LEXA binding site, located at positions -40 to -63 in the DNA sequence (Figure 21, functions as the operator for P2; in the presence of LEXA, transcription from P2 is inhibited while that from Pi is unaffected. We have also detected a third promoter, P3, which lies 320 bp upstream from P2; transcription from P3 is directed toward uvr6 and terminates in vitro in the region of the LEXA binding site. The physiological role of this promoter and its transcript is unknown (see below). Several aspects of the DNA sequence within the
Figure 6. Nucleotide Sequence of the 5’ End of RNA 3 All procedures and symbols are as described in the 1srgend to Figure 4 with the exception of the -Enz lane, which shows lat Beled tra lnscript untreated with either RNAase or base.
uvrB Promoters 527
540 bp Bst NI fragment containing the uvrB promoters deserve further discussion. In E. coli two specific regions on the DNA prior to the transcriptional start site are important for promoter function: the Pribnow box canonical sequence TATAATA is located approximately 10 bp upstream, while the canonical sequence TTGACA is found 35 bp upstream (for recent review see Rosenberg and Court, 1979). Within the uvrt3 gene the Pribnow box of Pl, TAAAATT, begins at -11 in the DNA sequence and the -35 sequence, TTGGCA, begins at -34; the analogous structures for P2 begin at -43 and -69, respectively (Figure 2). Pi and P2 are actually overlapping promoters, as transcription from P2 initiates at a G at position -31 within the -35 sequence of PI. Transcription from P2 also but rarely initiates at the A at position -29; this may be due to infrequent usage of a second Pribnow box, TAATTTG, beginning at -41, as has been proposed for the aroH gene of E. coli (Zurawski et al., 1981). The Pribnow boxes of Pl and P2 as well as the -35 sequence of Pl show close homology to the canonical sequences; however, the -35 sequence of P2 is a poor match to the consensus sequence. Despite this fact, transcription from Pl and P2 in the absence of LEXA is approximately equally efficient in vitro in 100 mtvl KCI, while in 180 mM KCI, transcription from P2 is increased relative to that from Pl (unpublished observations). Possibly other features of the DNA sequence surrounding P2, for example the stretch of T residues at -56 to -50, increase the efficiency of this promoter. In the presence of LEXA, transcription from P2 of uvrB is inhibited. We have shown that LEXA binds to a region on the DNA encompassing the Pribnow box and extending to the -35 sequence of P2 (Figure 2). This corresponds to the region bound by LEXA in the recA operator (Little et al., 1981) and by araC protein in the araC operator (Lee et al., 1981 I. Although only several base pairs separate the LEXA binding site (as defined by protection experiments) in uvr8 from the -35 sequence of Pl (Figure 21, transcription from Pl is unaffected by LEXA in vitro. This suggests that for Pl sequences upstream from the -35 sequence are not required for RNA polymerase binding. Our data indicate that Pl and P2 are authentic promoters of the uvrB gene, but the role of P3 and RNA-3 in the expression of the uvrl3 gene is unclear. It is unlikely that transcription from P3 is an artifact of the in vitro system used, as transcription from P3 is as efficient as transcription from Pl . Examination of the DNA sequence corresponding to RNA-3 reveals that this transcript does not encode a tRNA (Gauss and Sprinzl, 1981), 5s RNA (Erdmann, 1981) or a polypeptide longer than 31 amino acids. Although in vitro transcription of RNA-3 terminates within or near the LEXA binding site even in the absence of LEXA, it is possible that termination does not occur in vivo, as it may be modulated in ways not yet understood.
One interesting outcome of our studies on transcription from P3 is the answer to the question of what happens when actively transcribing RNA polymerase encounters a repressor-operator complex. Our results indicate that, in vitro and in the absence of a proper termination signal, polymerase stalls on the template near the complex. The discovery of a LEXA binding site in one of the uvrB promoters supports the prediction that most SOS genes have a LEXA regulated operator (Little et al., 1981; Brent and Ptashne, 1981). Since it is of interest to find a consensus sequence for a LEXA binding site, in Figure 7 we compare the sequences of the LEXA binding sites that have been reported to date. Although the prototypic LEXA binding site, the recA operator, has a perfect dyad symmetry with the exception of 2 bp in the middle, the two LEXA binding sites of the IexA gene and the LEXA binding site controlling P2 of the uvr6 gene are considerably different. The only base pairs that constitute dyad symmetry around the center of the binding sites in all four operators are the CTGs in the top and bottom strands, as indicated in Figure 7. While a CTG palindrome separated by ten base pairs may be necessary for recognition by LEXA, other portions of the DNA sequence must also affect binding, as such a palindrome would be expected to occur once in every 4096 base pairs. In addition, although all known LEXA binding sites possess the CTG palindrome and the conserved bases or base pairs shown in Figure 7, the strength of LEXA binding to these operators varies considerably (Brent and Ptashne, 1981). Other functionally important features of the LEXA binding site should become apparent from sequence studies of other SOS genes and their operator mutants. Identification of a LEXA-controlled operator in uvr6 1
TATGAGCAT
recA
t
Co”sens”s sequence
5*--$aT;qT-$-Y$Ym-R--3’
Figure 7. Comparison uvrB Genes of E. coli
of LEXA Binding
Sites
in the recA,
/exA and
The sequences are aligned around the center of symmetry of the LEXA binding sites. The dyad symmetry common to all four binding sites is indicated by boxes, and the consensus sequence is indicated below. The data for the LEXA binding sites in recA and lexA are from Little et al. (1981). Y indicates pyrimidine and R indicates purine. A/ T indicates that either an A or a T is found in this position.
Cell 528
explains the in vivo observation that insertion of a Mud (iac, amp) phage genome into the uvrB gene makes /I-galactosidase synthesis inducible by irradiation with ultraviolet light (Fogliano and Schendel, 1981) and demonstrates that induction is directly regulated by LEXA. The significance of a LEXA-independent promoter in uvrB is unknown; however, it is clear that no all excision repair genes contain both LEXA-dependent and LEXA-independent promoters, as LEXA inhibits essentially all transcription from the uvrA promoter (our unpublished observations). We have recently demonstrated that the ratio of uvrA protein to uvrB protein in recA cells is approximately 1:7 (Sancar et al., 1981); it is possible that the LEXA-independent promoter of uvr6 is responsible for this relatively high amount of uvrB protein. Whether this reflects the stoichiometry of uvrA and uvrB proteins in active ultraviolet-irradiated endonuclease or is the result of involvement of uvrB protein in some other cellular process requiring constitutive synthesis (Shizuya and Dykhuizen, 1972; Morimyo and Shimazu, 1976) is not known. Alternatively, it is possible that a different repressor binds to Pl of uvrB and is inactivated subsequent to ultraviolet irradiation, or that LEXA inhibits transcription from Pl in vivo but not in vitro as has been observed for galR repression of a promoter in the gal operon of E. coli (DiLauro et al., 1979). Further studies on the control of expression of the uvr genes in vivo and in vitro should answer these questions. Although it has been known for some time that E. coli possesses an inducible repair system (Weigle, 1953) whose expression requires the recA gene (Kneser, 1968) and is part of the SOS response, experiments designed to test the relationship between inducible repair and excision repair have given conflicting results. Radman and Devoret (1976), Witkin (1976) and Boyle and Setlow (1970) have obtained evidence indicating that inducible repair and excision repair are independent, whereas Mount et al. (19761, Rothman et al. (1979) and Castellazzi et al. (1980) have detected either low level or no inducible repair in uvr- cells under SOS-inducing conditions. The presence of LEXA-controlled operator in both the uvrA and uvrB genes provides strong evidence that the SOS response includes induction of the uvr genes. However, even in uvrAand uvrB- cells there is probably a residual level of inducible repair (Mount et al., 1976; Rothman et al., 1979), and there is no decrease in mutagenesis induced by ultraviolet irradiation (Witkin, 1976). Inducible repair therefore seems to have two components: one that is due to induction of the uvrA and uvrB genes and is errorfree, and another that requires the umuC gene product and is error-prone (Kato and Shinoura, 1977). Experimental
procedures
Bacterial Strains and Plaamida The E. coli K-l 2 derivative CSR603
frecA 7 uvrA6
phrl) (Sancar
and
Rupert. 1978) was used as the host pDR1494 (Sancar et al.. 1981).
strain
for the
uvr8
plasmid
Preparation of Plaamid DNA, ReWlction Enzyme Analyaia and Purltkatlon of DNA Fragments Followlng chloramphenicol-induced amplification, plasmid DNA was prepared from Sarkosyl lysates by centrifugation in ethidlum bromide-cesium chloride gradients, phenol extraction and ethanol precipitation. Restriction endonucleases Eco RI. Bst NI. Hae Ill. Hin 11 and Fnu DII were obtained from New England BioLabs and were used in accordance with the manufacturer’s speclflcetiona. DNA restriction fragments were analyzed on 6% acrylamide gels and stained with ethidium bromide for visualization with ultraviolet light; Hae Ill fragments of +X1 74 RFII DNA were used as molecular weight standards (Sanger et al., 1978). Fragments were recovered from acrylamide gels as described previously (Sancar and Rupp, 1979) or by electroelution (McDonell et al.. 1977). DNA Sequencing The 540 bp Bst NI fragment of pDRl494 or restriction fragments derived from it were labeled at either the 5’ or 3’ ends (Maxam and Gilbert, 1980; Schwarz et al., 1978), cut with a second restriction enzyme and purified by electrophoreais in acrylamide gels, then sequenced by the method of Maxam and Gilbert (19130). Preparation of LEXA LEXA was prepared as described by Little et al. (I 961) except that the phosphocellulose chromatography was replaced by passage through Affigel-501 (see Brent and Ptashne, 1961). The preparation was greater than 95% pure, and the protein concentration was 0.79 mg/ml. In Vitro Transcription Transcription inhibition experiments were performed essentially as described by Little et al. (1981). The reaction mixture consisted of 40 mM Tris-HCI (pH 8.0), 100 mM KCI. 10 mM MgCl2, 50 fl Na? EDTA (pH 8.0). 1 mM dithiothreitol and 100 ng of template In a total volume of 40 ~1. To this mixture we then added either 4 gl LEXA (in 20 mM Tris [pH 7.41. 10% sucrose, 1 mM dithiothreitol and 100 $vf Nan EDTA) or 4 Al buffer (without LEXA) in control tubes. Following incubation for IO min at 37°C 1 nl RNA polymerase (Enzo Biochemical& 1.75 units/Al) was added, and incubation was continued; 10 min later, 10 Al of 5x nucleotide mix were added to give a final concentration of 150 pM ATP, GTP and UTP and 35 j&f &‘P-CTP (280 Ci/mmole). In transcription experiments in which heparin was used, 1 ~1 of heparln (5 mg/ml in H20) was added to the reaction mix 1 min prior to addition of nucleotides, and the RNA polymerase concentration was decreased to one third of that used in the absence of heparin. Transcription was terminated 10 min after the addition of nucleotides by diluting the reaction mix with an equal volume of 90% formamide containing 50 mM Na, EDTA, 0.5% SDS, 20% sucrose, bromophenol blue and xylene cyanol. The mix was heated at QOOC for 2 min. then analyzed on 6% acrylamide gels in 9 M urea and exposed to Kodak X-Omat R film for autoradiography. For precise determination of the lengths of transcripts, a portion of the transcription reaction was brought to a concentration of 0.2% in SDS, extracted with phenol, precipitated and treated with 1 M glyoxal (Carmichael and McMaster. 1960). Glyoxylated RNA was analyzed on 6% acrylamide gels in 10 mM sodium phosphate buffer (pH 7.0). Glyoxylated Hae Ill fragments of @Xl 74 RFII DNA labeled at the 5’ ends with 3’P were used as molecular weight standards. 5’-End Labeling of Transcripts and RNA Sequence Analysis RNA-3 and RNA-l were labeled at the 5’ end by in vitro transcription. The reaction mixture consisted of 100 mM KCI, 40 mM Tris-HCI (pH 8.0). 10 mM MQCI., 1 mM dithiothreitol. 50 PM EDTA. 150 AM CTP. GTP and UTP, 15 PM r3*P-ATP (7000 Ci/mmole). 300 ng template and 7 ~1 of RNA polymerase in a total volume of 70 PI. After incubation for 1 hr at 37’C. the reaction was terminated by addition of SDS to 0.2%. followed by phenol extraction in the presence of 25 j.t~ E. coli tRNA and ethanol precipitation. The pellet was dissolved in 5 pl of
uvrB Promoters 529
water, 50 cl of formamide solution were added and the labeled transcripts were separated by electrophoresis in 6% acrylamide and 9 M urea eels. FOliOWing autoradiography, RNA was excised tom the gel and eluted OVemiQht at 4°C in 15 mM NaCI. 1.5 mM sodium citrate, 200 mM sodium acetate (PH 7.0) and 0.3% diethylpyrocarbonate. E. coli tRNA (25-50 PQ) was added, and the RNA was then precipitated in ethanol, extracted with phenol and dissolved in water. Labeling of RNA-2 was achieved by treating purified dephosphorylated RNA with T4 polynucleotide kinase (Pederson and Haeeltine, 1980). The 540 bp Bst NI fragment was transcribed in vitro under the conditions described above for 5’-end labeling except that all nucieotides were present at 150 CM and no labeled precursor was added. Transcription was terminated after 20 min by addition of formamidedye mixture, and the unlabeled transcripts were separated on a 6% acrylamide and 9 M urea gel; transcripts labeled with u-3*P-CTP were run in adjacent lanes as markers. Unlabeled RNA-2 was excised, &ted and ethanol-precipitated as described above except that carrier RNA was omitted. The RNA was dissolved in 10 cl of 10 mM Tris-HCI (pH 8.0) and 1 mM Nan EDTA and was dephosphorylated by the addition of 2.8 units bacterial alkaline phosphatase (Bethesda Research Laboratories). then incubated at 40°C for 1 hr. The endlabeling reaction was initiated by addition of 1 fl each of 100 mM KH,PO, (pH 9.5). 100 mM MQ&. 50 mM dithiothreitol and 1.3 cl T4 poiynucleotide kinase (Bethesda Research Laboratories, IO U/PI). The entire mixture was transferred to a tube containing 100 pmole lyophilized r3’P-ATP (7000 Ci/mmole) and incubated at 37’C for 4.5 hr. after which the RNA was precipitated in 2 M ammonium acetate, 2.5 volumes of ethanol and 10 PQ E. coli tRNA. The transcript was further purified by electrophoresis in a 6% acrylamide and 9 M urea gel, excised, eiuted and precipitated as described above. We sequenced 6’+nd-labeled transcripts by enzymatic methods. The reaction conditions for ribonucleases Tl (Sanyo). U2 and PhyM (P-L Biochemical@ and the base-generated ladder were essentially as described (Donis-Keller et al., 1977; Don&-Keller. 1980). except that carrier RNA was added only when necessary to bring the RNA concentration to 1.5-2 PQ per 25 PI reaction, and digestion with U2 was performed at pH 3.5 (DaSQupta et al., 1980). Digestion with 8. cereus ribonuclease (P-L Biochemicals) followed the procedure of Lockard et al. (1978) except that a single 15 min incubation was used. Digestion products were analyzed on 20% or 10% acrylamide gels in 8 M urea (Maxam and Gilbert, 1980). LEXA Footprinting DNAase I digestion of end-labeled 540 bp Bst NI fragment in the presence of LEXA was used to determine the location of the LEXA binding site within this fragment (footprinting. Galas and Schmltz, 1979; Johnson et al.. 1979). The experiments were performed essentially as described for the recA and /exA operators (Little et al., 1981). The 198 bp Hae Ill-Eco RI fragment at about 10 nM was incubated in the presence or absence of LEXA (monomer concentration, 100 nM) at 22°C for 10 min in a reaction mixture COntaininQ 20 mM Tris-HCI (pH 7.4). 50 mM NaCi. 2.5 mM MQCI,, 1.5 mM CaCIZ. 1 mM dithiothreitol, 100 CM Nan EDTA and 0.1% bovine serum albumin. DNAase I (Worthington) was added to a concentration of 10 n&ml. and incubation was continued for another 10 min. The samples were treated and analyzed as described by Galas and Schmitz (1978). Acknowledgments This work was supported by grants from the National Institutes of Health and a grant from the American Cancer Society. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received
November
19, 1981:
revised
December
A., Kenyon,
C. J. and Walker,
G. C. (1981).
lnducibility
in Es&e-
Boyle, J. M. and Setiow. R. B. (1970). Correlations between host-cell reactivation. ultraviolet reactivation and pyrimidine dimer excision in the DNA of bacteriophage X. J. Mol. Biol. 51, 131-l 44. Brent, R. and Ptashne, M. (1980). The /exA gene product represses its own promoter. Proc. Nat. Acad. Sci. USA 77, 1932-l 938. Brent, R. and Ptashne. M. (1981). Mechanism of action gene product. Proc. Nat. Acad. Sci. USA, 78, 4204-4208.
of the /exA
Carmichael, G. G. and McMaster. G. K. (1980). The analysis of nucleic acids using glyoxal and acridine orange. Meth. Enzymol. 65, 380-391. Castellazzi. M.. Jacques, M. and George, deoxyribonucleic acid repair in Escherichia 143, 703-709.
J. (1980). tif-stimulated co/i K-l 2. J. Bacterial.
Craig, N. L. and Roberts, J. W. (1960). E. co/i recA protein-directed cleavage of X repressor requires polynucleotide. Nature 283, 28-30. Dasgupta. R.. Ahlquist, P. and Kaesberg, P. (1980). Sequence of the 3’ untranslated region of brome mosaic virus coat protein messenger RNA. Virology 104, 339-346. DiLauro, R., Taniguchi. T., Musso. R. and decrombrugghe, B. (1979). Unusual location and function of the operator in the Escherichia co/i galactose operon. Nature 279, 494-500. Donis-Keller. A residues 3133-3142.
H. (1980). Phy M: an RNase activity specific for U and useful in RNA sequence analysis. Nucl. Acids. Res. 8.
Don&Keller, H., Maxam, A. M. and Gilbert, W. (1977). Mapping adenines. guanines. and pyrimidines in RNA. Nucl. Acids. Res. 4, 2527-2538. Erdmann. sequences
V. A. (1981). Collection of published 5S and 5.85 and their precursors. Nucl. Acids Res. 9, r25-r42.
FOQiianO. M. and Schendel, P. (1981). Evidence the uvn3 operon. Nature 289, 196-l 98. Galas, D. and Schmitz, method for the detection Acids Res. 5, 3157-3170.
RNA
for the inducibility
of
A. (1978). DNase footprinting: a simple of protein-DNA binding specificity. Nucl.
Gauss, D. H. and Sprinzi. M. (1981). Nucl. Acids Res. 9, rl -r23.
Compilation
of tRNA sequences.
Howard-Flanders, P. and Boyce, R. P. (1966). DNA repair and genetic recombination: studies on mutants of Escherichia co/i defective in these processes. Radiation Res. Suppl. 6, 158-l 84. Huisman, 0. and D’Ari, R. (1981). An inducible DNA replication-cell division coupling mechanism in E. co/i. Nature 290. 797-799. Johnson, between repressor.
A., Meyer, B. J. and Ptashne, M. (1979). Interactions DNA-bound repressors govern regulation by the h phage Proc. Nat. Acad. Sci. USA 78, 5061-5065.
Kacinski. B. M., Sancar. A. and Rupp. W. D. (1981). A general approach for purifying proteins encoded by cloned genes without using a functional assay: isolation of the UVRA gene product from radiolabelled maxicells. Nucl. Acids Res. 9, 4495-4508. Kato. T. and Shinoura, Y. (1977). Isolation and characterization mutants of Escherichia co/i deficient in induction of mutations ultraviolet light. Mol. Gen. Genet. 156, 121-131.
of by
Kenyon. C. J. and Walker, G. C. (1980). DNA-damaging agents stimulate gene expression at specific loci in Escherichia co/i. Proc. Nat. Acad. Sci. USA 77, 281 Q-2823. Kenyon. C. J. and Walker, G. C. (1981). Expression uvrA gene is inducible. Nature 289, 808-810.
of the E. co/i
Kneser, tivation
and UV-reac-
H. (1968). Relationship between K-reactivation of bacteriophage X. Virology 36, 303-305.
Lee, N. L.. Gielow. W. 0. and Wallace, R. G. (1981). Mechanism of araC autoregulation and the domains of two overlapping promoters, PC and Pero. in the L-arabinose regulatory region of Escherichia co/i. Proc. Nat. Acad. Sci. USA 78, 752-756.
30. 1981
References Bagg.
gene product required for UV and chemical mutagenesis richia co/i. Proc. Nat. Acad. Sci. USA 78, 5749-5753.
of a
Little, J. W.. Edmiston,
S. H., Pacelli.
L. 2. and Mount,
D. W. (1980).
Cell 530
Cleavage of the fscherichia co/i leti protein Proc. Nat. Acad. Sci. USA 77, 3225-3229.
by the recA
protease.
Little, J. W., Mount, D. W. and Yanisch-Perron, C. R. (1961). Purified /exA protein is a repressor of the recA and /exA genes. Proc. Nat. Acad. Sci. USA, 78,4199-4203. Lockard, R. E.. Abner-Deweerd, B., Heckman. J. E., MacGee. J.. Tabor. M. W. and RajBhandary. U. L. (1978). Sequence analysis of 5’ t3*P) labelled mRNA and tRNA using polyacrylamide gel electre phoresis. Nucl. Acids Res. 5, 37-56. Maxam, A. M. and Gilbert, W. (1980). with base-specific chemical cleavages.
Sequencing end-labelled DNA Meth. Enzymol. 65499-560.
McDonell, M. W., Simon, M. N. and Studier, F. W. (1977). Analysis of restriction fragments of T7 DNA and determination of molecular weights by electrophoresis in neutral and alkaline gels. J. Mol. Biol. 110, 119-146. Morimyo, M. and Shimazu, Y. (1976). Evidence that the gene uvr8 is indispensable for polymerase I deficient strain of Escherichia co/i K12. Mol. Gen. Genet. 147, 243-250. Mount, D. W.. Kosel, C. K. and Walker, A. (1976). Inducible. errorfree DNA repair in tsl recA mutants of E. co/i. Mol. Gen. Genet. 146, 37-41. Pederson. F. S. and Hasaltine, W. A. (1960). A micromethod detailed characterization of high molecular weight RNA. Meth. zymol. 84, 660-687.
for En-
Radman, M. (1974). Phenomenology of an inducible mutagenic DNA repair pathway in Escherichia co/i: SOS repair hypothesis. In Molecular and Environmental Aspects of Mutagenesis, L. Prakash. F. Sherman. M. Miller, C. Lawrence and H. W. Tabor, eds. (Springfield: Charles C Thomas), pp. 126-l 42. Radman, M. and Devoret, R. (1971). UV-reactivation h in excision repair deficient hosts: independence and attachment regions. Virology 43, 504-506.
of bacteriophage of red functions
Roberts, J. W., Roberts, C. W. and Mount, D. W. (1977). Inactivation and proteolytic cleavage of phage h repressor in vitro in an ATPdependent reaction. Proc. Nat. Acad. Sci. USA 72, 2330-2334. Rosenberg, M. and Court, D. (1979). Regulatory sequences involved in the promotion and termination of RNA transcription. Ann. Rev. Genet. 13, 319-353. Rothman. R. H., Margossian. L. J. and Clark, reactivation of phage lambda in recF, recL, uvrA. of E. co/i K-i 2. Mol. Gen. Genet. 169, 279-287.
A. J. (1979). Wand uvr8 mutants
Sancar. A. and Rupert. C. S. (1978). Determination lecular weights from ultraviolet sensitivities. Nature
of plasmid mo272, 471-472.
Sancar. A. and Rupp, W. D. (1979). Physical Proc. Nat. Acad. Sci. USA 76, 3144-3148.
map of the recA
gene.
Sancar, A., Stachelek. C., Konigsberg. W. and Rupp. W. D. (1980). Sequences of the recA gene and protein. Proc. Nat. Acad. Sci. USA 77, 261 l-261 5. Sancar, A., Clarke, N. D., Griswold, J.. Kennedy, W. J. and Rupp. W. D. (1981). Identification of the uvr8 gene product. J. Mol. Biol. 748, 63-76. Sanger, F.. Coulson, A. R., Friedmann, Brown, N. L.. Fiddes. J. C., Hutchinson. Smith, M. J. (1978). The nucleotide @Xl 74. J. Mol. Biol. 125, 225-248.
T.. Air, G. N.. Barrel& B. G., C. A. II, Slocombe. P. M. and sequence of bacteriophage
Schwarz, E., Schere, G., Hobom, and Kossel. H. (1978). Nucleotide sequence of cfo, c/l, and part of the 0 gene in phage DNA. Nature 272, 410-414. Shizuya, H. and Dykhuizen, D. (1972). Conditional lethality of deletions which include uvf8 in strains of Escherichia co/i lacking deoxynucleic acid polymerase I. J. Bactertol. 112, 676-881. Steib, J. A. (1979). RNA-RNA interactions during polypeptide chain initiation. In Ribosomes, G. Chambliss, G. R. Craven, J. Davies, L. Kahan and M. Nomura, eds. (Baltimore: University Park Press), pp. 349-399. Weigle. J. J. (1953). Induction of mutation Nat. Acad. Sci. USA 39, 628-636.
in a bacterial
virus.
Proc.
Witkin. E. M. (1976). Ultraviolet mutagenesis and inducible repair in Escberichia co/i. Bacterial. Rev. 40, 869-907.
DNA
Zurawski, G.. Gunsalus. R. B., Brown, K. D. and Yanofsky. C. (1981). Structure and regulatton of afofi, the structural gene for the tryptophan-repressible 3-deoxy-arabino-heptulosonic acid-7-phosphate synthetase In Escherichia co/i. J. Mol. Blol. 145. 47-73. Note
Added
in Proof
We have now established that the sequence of the amino terminus of uvrB protein is Ser-Lys-Pro-Phe-Lys-Leu-Asn-Ser-Ala-Phe, indicating that translation of uvrB message begins at the first fMet codon at position +34. Recently, van den Berg et al. (Nucl. Acids Res. 9, 5623-5643. 1981) also have identified two promoters of the uvr8 gene, which correspond to the promoters named Pl and P2 in this report.
March
1982,
Copyright
0 1982
by MIT
The uvrB Gene of Escherichia coli Has Both IexA -Repressed and IexA -Independent Promoters Gwendolyn B. Sancar,* Aziz Sancar, * John W. Little+ and W. Dean Rupp* * Department of Therapeutic Radiology and Department of Molecular Biophysics and Biochemistry Yale University School of Medicine New Haven, Connecticut 06510 + Department of Molecular and Medical Microbiology University of Arizona College of Medicine Tucson, Arizona 65724
Summary We have found that the uvrB gene of Escherichia coli is transcribed from at least two promoters, which we call Pl and P2. Transcription from Pl begins with an ATP at + 1, and transcription from P2 begins primarily with a GTP at -31. A binding site for the /exA protein (LEXA), located between the -35 sequence and Pribnow box of P2, regulates transcription from this promoter. In vitro, LEXA inhibits transcription from P2 but has no detectable effect on transcription from Pl. A third promoter, P3, was also detected at -341; transcription from P3 is toward uvr6 but terminates in vitro in the region of the LEXA binding site. The binding of LEXA to P2 inhibits transcription from the P3 promoter even though several hundred nucleotides separate the two promoters. The data suggest that a transcribing RNA polymerase stalls when it reaches the repressor-operator complex but remains bound to the DNA, causing a jamming of RNA polymerases between P3 and the repressor-operator complex at P2. The physiological significance of P3 is unknown. Introduction Damage to DNA in E. coli by far ultraviolet light (200300 nm) induces a variety of cellular responses, including filamentation, increased radioresistance and mutability, some of which help repair the damage or ensure survival of the cell (for review, see Witkin, 1976). The coordinated expression of these functions has been called the “SOS response,” while the term SOS genes is used to designate the genetic loci that are induced under these conditions (Radman, 1974; Witkin, 1976). Included among the SOS genes are /exA, recA (postreplication repair and recombination), sfiA (cell division), urn& (induced mutagenesis) and a number of din genes whose functions are not known (Witkin, 1976; Kenyon and Walker, 1960; Huisman and D’Ari, 1981; Bagg et al., 1981). One property SOS genes have in common is that their expression is controlled by the recA and IexA genes (for review see Witkin, 1976). The essential features of the interaction between the recA and IexA
genes have been described (Roberts et al., 1977; Brent and Ptashne, 1980 and 1981; Craig and Roberts, 1980; Little et al., 1980 and 1981). The LEXA protein is the repressor of both the IexA and recA genes; under normal conditions neither gene is expressed at high levels. It is believed that DNA damage results in the production of a signal, such as oligonucleotides or single-stranded DNA, which activates the protease activity of the endogenous recA protein to cleave LEXA, activating both recA and IexA transcription. Following repair and recovery, the absence of an appropriate activator results in the loss of recA protease activity, causing the accumulation of LEXA and, consequently, the reestablishment of repression. Presumably other SOS genes are also repressed by LEXA, although this has not been proved directly. In addition to the SOS genes listed above, evidence has been presented that the nucleotide excision repair genes uvrA (Kacinski et al., 1981; Kenyon and Walker, 1981) and uvrB (Fogliano and Schendel, 1981) are also induced during the SOS response and thus may be part of the recA-IexA regulon. These results were unexpected as it had been reported previously that mutations in the recA gene do not abolish the activity of the excision repair system (Howard-Flanders and Boyce, 1966). In an attempt to resolve this paradox, we have undertaken a detailed study of the nucleotide sequence organization and expression of the excision repair genes. In this paper we present the results of our analysis of the uvrB gene. Results Restriction Map of the uvrB Promoter-Operator Region We reported previously that uvrB is a monocistronic gene approximately 2 kb in length, which codes for a protein of molecular weight 84,000 (Sancar et al., 1981). We showed that the uvrB gene is inactivated when cut with Eco RI and presented evidence that the Eco RI site is very near the translational start of the gene. To obtain a detailed restriction map of the promoter-operator region of uvrB, we isolated several restriction fragments that carry the Eco RI site of the gene. Among these a 540 bp Bst NI fragment was used for the experiments reported here. The location of the 540 bp Bst NI fragment relative to uvr6 on the plasmid pDR1494 (Sancar et al., 1981) and the recognition sites for various restriction endonucleases on the 540 bp Bst NI fragment are shown in Figure 1. Sequence of the uvrB Promoter-Operator Region To find the uvrB promoter we determined the sequence of the entire 540 bp Bst NI fragment with the exception of about 10 nucleotides at each end (see Figure 1 for the sequencing strategy used). The sequence, shown in Figure 2, contains a number of
Ceil
524
A+T-rich regions and several potential promoter regions. Of particular interest is a region that is similar to the operator-promoter regions of the ret and IexA genes. The sequence CAGTATAATT beginning at -46 is homologous to the recA sequence from - 16 to -7; in the recA gene this sequence contains the
Pribnow box (Sancar et al., 1980). in addition, the sequence of the 540 bp Bst Ni fragment from -59 to -44 is similar to the consensus sequence CTGTAT...C...CAG of the LEXA binding sites in the recA and IexA genes (Little et al., 1981). These observations suggest that the region around -50 contains a LEXA-controlled operator of the uvrB gene.
LEXA Footprinting To determine whether LEXA interacts with the putative LEXA binding site in the 540 bp Bst Ni fragment, we incubated end-labeled DNA from the 540 bp Bst Ni fragment with DNAase I in the presence and absence of LEXA, then separated the products on a DNA sequencing gel (Galas and Schmitz, 1978; Johnson et al., 1979). LEXA protected, from DNAase I attack, the region between -63 and -33 on the anticoding strand and between -67 and -40 on the coding strand (Figure 3). Enhanced cutting was observed at two T residues within this region. I
540 bp
1
Figure 1. Restriction Map of pDR1494 and the 540 bp Bst NI Fragment Carrying the ovr8 Promoter, and the Strategy Used to Sequence the Fragment The vector (pBR322) is indicated by the cross-hatched line and the poiy (dA-dT) connectors with + + + + Arrows indicate the regions occupied by the uvrB gene and the distal portion of bioD as well as the direction of transcription of these genes. The single circle designates the promoter region of the uvrB gene. Numbering is relative to the Eco RI site of the vector as in Sancar et al. (1981). Arrows below the enlarged map of the 540 bp Bst NI fragment indicate the direction of sequencing and the lengths of the sequences read. Both strands of the Hae Ill-Eco RI fragment carrying the uvrB promoter were sequenced.
in Vitro Transcription of uvrB and Determination the Transcrlption Start Sites
of
Having found a LEXA binding site in the 540 bp Bst Ni fragment, we wanted to determine whether in vitro transcription of uvr8 is controlled by binding of repressor to this site. As is seen in Figure 4A, E. coli RNA poiymerase produces not only one, but three discrete transcripts with lengths of 130 bases, 160 bases and 290 bases, which we call RNA-l, RNA-2 and RNA-3, respectively. Cutting the DNAase I fragment with Eco RI shortens the length of both RNA-l 53
-400
JI GAAbACGAGGCAA6CGAGAGAATkGcGGCTTG~ACGCGAA CTTTTGTGCTCCGTTCGCTCTCTTATGCGCCGAACGTGCGCTT
GCCGCCCCCTATCCCCGACCTCTGTCAATAGGTGA
-300 -250 -200 TT~CGTTAkAGliCCGCTCA~~TATCTT~TATTTTTT~CCGCTTAGA~~TGC~T~GCAGTCACT~~CA~CAT~TCTTGCCAT;9AAACTGTCA~CACTCAT AACCGCAATTTCTGCCGAGTTTCTTTATAGAAAATAMAAA TTGGCCAATCTATTTACGTTACCGTCAGTGACTTGTCCGTAGAGAACGGTATTTTGACAGTAGTGAGTA -150 -100 CTkACAAATGT~ AAAAAAGCCeTTGCTTTGGEGATNLCCC~TAAGGCCGGlicTTTTATCT~GCCACAGAG~AAATTTTGCICATGATTGA~A~~A~T~~C~TGT GAACTGTTTACAATTTTTTCGGCAACGAAACCCCTATTGGGCCATn:CGCCCTCAAAATAGAGCGGTGTCTCATTTM~CGAGTACT~CTGTCGCCTCAAATGCGACA
t GAkTGT@TTTfATCCA CTTGACAAAAAAATAG I +50 +100 G~CTCATGAGT~CCGTTC~CTGAATTC~TTTTAITGGCC CTGAGTACTCATTTGGCAAGTTTGACTTAAGCGAAAATTTGGAAGACCCCTAGTCGGTCTCCGCTAAGCTGCAGAGCTTCTCCCCGACCTTCTACCGG Figure
2. Sequence
of the 540 bp Bst NI Fragment
Containing
the uvrS Operator-Promoter
Region
The base sequence of the 540 bp Bst NI fragment in Figure 1 is given along with the regulatory regions described in the text. The transcription initiation sites for RNA-l, RNA-2 and RNA-3 are denoted by Sl , 52 and S3 with the numbering of nucieotides relative to Sl . Pribnow boxes and -35 sequences are enclosed in boxes (dashed for Sl , solid for S2 and S3). Single bars above and below the line indtcate the region protected by LEXA from DNAase I attack: circles denote nucieotides at which cutting was enhanced in the presence of LEXA. Potential ribosome binding sites (Steitz. 1979) near ATG codons on the uvr6 message are overscored with two lines. (Since both fMet codons at +34 and +121 are in the same open reading frame, we do not know which of these sites defines the start of translation of uvri3 in viva.)
uvf6
Promoters
525
A E
G
T + C
C
DNaroI -+ IrxA IsxA
A
G T
E
E
C
DNoae I + 1w.A IexA
-IrxA
+IsxA
-lexA
+lexA
RNA-3*
-60.
RNA-2 -40. -30.
-40.
RNA- I
-30.
Figure 4. Autoradiograms of RNAs Transcribed bp Est NI Fragment in the Presence or Absence
in Vitro from the 540 of LEXA
(A) Transcripts produced In the absence of hepartn and the absence (lane 1) or presence (lane 2) of LEXA. (6) Transcripts produced In the presence of heparin and the absence (lane 1) or presence (lane 2) of LEXA. Transcriptlon, gel electrophoresis and autoradiography were performed as described in Experimental Procedures. Exposure time for the autOradiOQrafII shown In B was approximately ten times longer than for the autoradiogram shown in A.
Figure
3. LEXA Footprintmg
The 1 QB bp Hae Ill-Eco RI fragment of the 540 bp Bst NI fragment was labeled at the Eco RI site and treated with DNAase I in the presence or absence of LEXA. Fragments not treated with DNAase I were used for sequence analysis (Maxam and Gilbert, 1 QBO) and run in wells adjacent to the DNAase l-treated samples. The region protected by LEXA Is enclosed in brackets: asterlsks indicate nucleotides at which cutting by DNAase I is enhanced in the presence of LEXA. (A) Anticoding strand, 5’ end labeled. (B) COditIg strand, 3’ end labeled.
and RNA-2 by approximately 75 bases, but has no effect on RNA-3 (data not shown). Thus RNA-l and RNA-2 are runoff transcripts that originate upstream from the Eco RI site and extend into uvr8. The direction of transcription of RNA-1 and RNA-2 was confirmed, and the precise transcriptional start sites (Sl and S2) were determined by RNA sequence analysis of these transcripts (Figure 5). RNA-l is initiated at an A residue, which we have numbered + 1 (Figure 2); RNA-2 begins with a G at position -31 and at low frequency at an A at -29 (not shown) and extends through Sl . From this finding, we conclude that uvr8 is read from two separate promoters. Separate experiments showed that the 5’ end of RNA-3 lies at -341 and extends toward uvrB (Figure 6). When LEXA was added to the transcription reaction prior to the addition of polymerase. two separate
effects were observed. First, LEXA strongly inhibits synthesis of the uvrB transcript RNA-P, but not that of RNA-1 (Figure 4A). Thus, although Sl and S2 are separated by only 31 bp, transcription from their respective promoters, Pl and P2, is regulated in different manners. Second, and more surprising, transcription of RNA3 is also inhibited by LEXA. Because the sequence 5’ to 53 is not protected by LEXA (not shown), we considered the possibility that LEXA inhibited transcription of RNA-3 by an indirect mechanism. From the length of RNA-3 we would predict that transcription terminates in the region of the proposed LEXA binding site; preliminary data (not shown) indicate that termination occurs within the stretch of Ts centered at position -53. This suggested that LEXA might inhibit transcription of RNA-3 by blocking the site at which termination and release of RNA-3 is effected in the absence of LEXA, leading to accumulation on the template of stalled polymerase molecules bearing progressively shorter nascent transcripts. To test this hypothesis we transcribed the 540 bp Bst NI fragment in the presence of heparin to limit transcription to one initiation per promoter. As is seen in Figure 4B, transcription under these conditions and in the presence of LEXA results in the disappearance of RNA-3 and the appearance of a new major transcript (RNA-3’) approximately 25 nucleotides shorter than RNA-3. In a related experiment, it was found that cutting the 540 bp Bst NI fragment with Msp I (an isoschizomer of Hpa
Cell 526
-ENZ G
0”.
A
U+A
CA UU
G
OH-
A
U+A U*C
UK
A: i”
$6
GA C
t E G
EG A U
C
G A A
c c
U
G
U A
U
A c
U
A
c
c ATTP G 4
Figure
5. Nucleotide
Sequence
of the 5’ End of RNA-1
RNA-i (A) and RNA-2 (6) were labeled sequenced as described in Experimental tions were obtained with RNAase 1 (lane U2 (lane A). RNAase Phy M (lane U+A) U+C). In B, the position of r3*P-ATP. nucleotide. is shown.
and RNA-2
at their 5’ ends, purified and Procedures. Partial dlgesG). base (lane OH-). RNAase and RNAase 8. cereus (lane used as a marker for the first
% uCA noA AG AG CGi A CA G G A G
C
II) abolished the ability of LEXA to inhibit transcription from the promoter used for synthesis of RNA-3 (data not shown). These results support the conclusion that LEXA does not directly affect initiation of transcription of RNA-3. Discussion We have sequenced the operator-promoter region of the uvrB gene of E. coli and have shown that in vitro the gene is transcribed from two different promoters, which we call Pl and P2. The LEXA binding site, located at positions -40 to -63 in the DNA sequence (Figure 21, functions as the operator for P2; in the presence of LEXA, transcription from P2 is inhibited while that from Pi is unaffected. We have also detected a third promoter, P3, which lies 320 bp upstream from P2; transcription from P3 is directed toward uvr6 and terminates in vitro in the region of the LEXA binding site. The physiological role of this promoter and its transcript is unknown (see below). Several aspects of the DNA sequence within the
Figure 6. Nucleotide Sequence of the 5’ End of RNA 3 All procedures and symbols are as described in the 1srgend to Figure 4 with the exception of the -Enz lane, which shows lat Beled tra lnscript untreated with either RNAase or base.
uvrB Promoters 527
540 bp Bst NI fragment containing the uvrB promoters deserve further discussion. In E. coli two specific regions on the DNA prior to the transcriptional start site are important for promoter function: the Pribnow box canonical sequence TATAATA is located approximately 10 bp upstream, while the canonical sequence TTGACA is found 35 bp upstream (for recent review see Rosenberg and Court, 1979). Within the uvrt3 gene the Pribnow box of Pl, TAAAATT, begins at -11 in the DNA sequence and the -35 sequence, TTGGCA, begins at -34; the analogous structures for P2 begin at -43 and -69, respectively (Figure 2). Pi and P2 are actually overlapping promoters, as transcription from P2 initiates at a G at position -31 within the -35 sequence of PI. Transcription from P2 also but rarely initiates at the A at position -29; this may be due to infrequent usage of a second Pribnow box, TAATTTG, beginning at -41, as has been proposed for the aroH gene of E. coli (Zurawski et al., 1981). The Pribnow boxes of Pl and P2 as well as the -35 sequence of Pl show close homology to the canonical sequences; however, the -35 sequence of P2 is a poor match to the consensus sequence. Despite this fact, transcription from Pl and P2 in the absence of LEXA is approximately equally efficient in vitro in 100 mtvl KCI, while in 180 mM KCI, transcription from P2 is increased relative to that from Pl (unpublished observations). Possibly other features of the DNA sequence surrounding P2, for example the stretch of T residues at -56 to -50, increase the efficiency of this promoter. In the presence of LEXA, transcription from P2 of uvrB is inhibited. We have shown that LEXA binds to a region on the DNA encompassing the Pribnow box and extending to the -35 sequence of P2 (Figure 2). This corresponds to the region bound by LEXA in the recA operator (Little et al., 1981) and by araC protein in the araC operator (Lee et al., 1981 I. Although only several base pairs separate the LEXA binding site (as defined by protection experiments) in uvr8 from the -35 sequence of Pl (Figure 21, transcription from Pl is unaffected by LEXA in vitro. This suggests that for Pl sequences upstream from the -35 sequence are not required for RNA polymerase binding. Our data indicate that Pl and P2 are authentic promoters of the uvrB gene, but the role of P3 and RNA-3 in the expression of the uvrl3 gene is unclear. It is unlikely that transcription from P3 is an artifact of the in vitro system used, as transcription from P3 is as efficient as transcription from Pl . Examination of the DNA sequence corresponding to RNA-3 reveals that this transcript does not encode a tRNA (Gauss and Sprinzl, 1981), 5s RNA (Erdmann, 1981) or a polypeptide longer than 31 amino acids. Although in vitro transcription of RNA-3 terminates within or near the LEXA binding site even in the absence of LEXA, it is possible that termination does not occur in vivo, as it may be modulated in ways not yet understood.
One interesting outcome of our studies on transcription from P3 is the answer to the question of what happens when actively transcribing RNA polymerase encounters a repressor-operator complex. Our results indicate that, in vitro and in the absence of a proper termination signal, polymerase stalls on the template near the complex. The discovery of a LEXA binding site in one of the uvrB promoters supports the prediction that most SOS genes have a LEXA regulated operator (Little et al., 1981; Brent and Ptashne, 1981). Since it is of interest to find a consensus sequence for a LEXA binding site, in Figure 7 we compare the sequences of the LEXA binding sites that have been reported to date. Although the prototypic LEXA binding site, the recA operator, has a perfect dyad symmetry with the exception of 2 bp in the middle, the two LEXA binding sites of the IexA gene and the LEXA binding site controlling P2 of the uvr6 gene are considerably different. The only base pairs that constitute dyad symmetry around the center of the binding sites in all four operators are the CTGs in the top and bottom strands, as indicated in Figure 7. While a CTG palindrome separated by ten base pairs may be necessary for recognition by LEXA, other portions of the DNA sequence must also affect binding, as such a palindrome would be expected to occur once in every 4096 base pairs. In addition, although all known LEXA binding sites possess the CTG palindrome and the conserved bases or base pairs shown in Figure 7, the strength of LEXA binding to these operators varies considerably (Brent and Ptashne, 1981). Other functionally important features of the LEXA binding site should become apparent from sequence studies of other SOS genes and their operator mutants. Identification of a LEXA-controlled operator in uvr6 1
TATGAGCAT
recA
t
Co”sens”s sequence
5*--$aT;qT-$-Y$Ym-R--3’
Figure 7. Comparison uvrB Genes of E. coli
of LEXA Binding
Sites
in the recA,
/exA and
The sequences are aligned around the center of symmetry of the LEXA binding sites. The dyad symmetry common to all four binding sites is indicated by boxes, and the consensus sequence is indicated below. The data for the LEXA binding sites in recA and lexA are from Little et al. (1981). Y indicates pyrimidine and R indicates purine. A/ T indicates that either an A or a T is found in this position.
Cell 528
explains the in vivo observation that insertion of a Mud (iac, amp) phage genome into the uvrB gene makes /I-galactosidase synthesis inducible by irradiation with ultraviolet light (Fogliano and Schendel, 1981) and demonstrates that induction is directly regulated by LEXA. The significance of a LEXA-independent promoter in uvrB is unknown; however, it is clear that no all excision repair genes contain both LEXA-dependent and LEXA-independent promoters, as LEXA inhibits essentially all transcription from the uvrA promoter (our unpublished observations). We have recently demonstrated that the ratio of uvrA protein to uvrB protein in recA cells is approximately 1:7 (Sancar et al., 1981); it is possible that the LEXA-independent promoter of uvr6 is responsible for this relatively high amount of uvrB protein. Whether this reflects the stoichiometry of uvrA and uvrB proteins in active ultraviolet-irradiated endonuclease or is the result of involvement of uvrB protein in some other cellular process requiring constitutive synthesis (Shizuya and Dykhuizen, 1972; Morimyo and Shimazu, 1976) is not known. Alternatively, it is possible that a different repressor binds to Pl of uvrB and is inactivated subsequent to ultraviolet irradiation, or that LEXA inhibits transcription from Pl in vivo but not in vitro as has been observed for galR repression of a promoter in the gal operon of E. coli (DiLauro et al., 1979). Further studies on the control of expression of the uvr genes in vivo and in vitro should answer these questions. Although it has been known for some time that E. coli possesses an inducible repair system (Weigle, 1953) whose expression requires the recA gene (Kneser, 1968) and is part of the SOS response, experiments designed to test the relationship between inducible repair and excision repair have given conflicting results. Radman and Devoret (1976), Witkin (1976) and Boyle and Setlow (1970) have obtained evidence indicating that inducible repair and excision repair are independent, whereas Mount et al. (19761, Rothman et al. (1979) and Castellazzi et al. (1980) have detected either low level or no inducible repair in uvr- cells under SOS-inducing conditions. The presence of LEXA-controlled operator in both the uvrA and uvrB genes provides strong evidence that the SOS response includes induction of the uvr genes. However, even in uvrAand uvrB- cells there is probably a residual level of inducible repair (Mount et al., 1976; Rothman et al., 1979), and there is no decrease in mutagenesis induced by ultraviolet irradiation (Witkin, 1976). Inducible repair therefore seems to have two components: one that is due to induction of the uvrA and uvrB genes and is errorfree, and another that requires the umuC gene product and is error-prone (Kato and Shinoura, 1977). Experimental
procedures
Bacterial Strains and Plaamida The E. coli K-l 2 derivative CSR603
frecA 7 uvrA6
phrl) (Sancar
and
Rupert. 1978) was used as the host pDR1494 (Sancar et al.. 1981).
strain
for the
uvr8
plasmid
Preparation of Plaamid DNA, ReWlction Enzyme Analyaia and Purltkatlon of DNA Fragments Followlng chloramphenicol-induced amplification, plasmid DNA was prepared from Sarkosyl lysates by centrifugation in ethidlum bromide-cesium chloride gradients, phenol extraction and ethanol precipitation. Restriction endonucleases Eco RI. Bst NI. Hae Ill. Hin 11 and Fnu DII were obtained from New England BioLabs and were used in accordance with the manufacturer’s speclflcetiona. DNA restriction fragments were analyzed on 6% acrylamide gels and stained with ethidium bromide for visualization with ultraviolet light; Hae Ill fragments of +X1 74 RFII DNA were used as molecular weight standards (Sanger et al., 1978). Fragments were recovered from acrylamide gels as described previously (Sancar and Rupp, 1979) or by electroelution (McDonell et al.. 1977). DNA Sequencing The 540 bp Bst NI fragment of pDRl494 or restriction fragments derived from it were labeled at either the 5’ or 3’ ends (Maxam and Gilbert, 1980; Schwarz et al., 1978), cut with a second restriction enzyme and purified by electrophoreais in acrylamide gels, then sequenced by the method of Maxam and Gilbert (19130). Preparation of LEXA LEXA was prepared as described by Little et al. (I 961) except that the phosphocellulose chromatography was replaced by passage through Affigel-501 (see Brent and Ptashne, 1961). The preparation was greater than 95% pure, and the protein concentration was 0.79 mg/ml. In Vitro Transcription Transcription inhibition experiments were performed essentially as described by Little et al. (1981). The reaction mixture consisted of 40 mM Tris-HCI (pH 8.0), 100 mM KCI. 10 mM MgCl2, 50 fl Na? EDTA (pH 8.0). 1 mM dithiothreitol and 100 ng of template In a total volume of 40 ~1. To this mixture we then added either 4 gl LEXA (in 20 mM Tris [pH 7.41. 10% sucrose, 1 mM dithiothreitol and 100 $vf Nan EDTA) or 4 Al buffer (without LEXA) in control tubes. Following incubation for IO min at 37°C 1 nl RNA polymerase (Enzo Biochemical& 1.75 units/Al) was added, and incubation was continued; 10 min later, 10 Al of 5x nucleotide mix were added to give a final concentration of 150 pM ATP, GTP and UTP and 35 j&f &‘P-CTP (280 Ci/mmole). In transcription experiments in which heparin was used, 1 ~1 of heparln (5 mg/ml in H20) was added to the reaction mix 1 min prior to addition of nucleotides, and the RNA polymerase concentration was decreased to one third of that used in the absence of heparin. Transcription was terminated 10 min after the addition of nucleotides by diluting the reaction mix with an equal volume of 90% formamide containing 50 mM Na, EDTA, 0.5% SDS, 20% sucrose, bromophenol blue and xylene cyanol. The mix was heated at QOOC for 2 min. then analyzed on 6% acrylamide gels in 9 M urea and exposed to Kodak X-Omat R film for autoradiography. For precise determination of the lengths of transcripts, a portion of the transcription reaction was brought to a concentration of 0.2% in SDS, extracted with phenol, precipitated and treated with 1 M glyoxal (Carmichael and McMaster. 1960). Glyoxylated RNA was analyzed on 6% acrylamide gels in 10 mM sodium phosphate buffer (pH 7.0). Glyoxylated Hae Ill fragments of @Xl 74 RFII DNA labeled at the 5’ ends with 3’P were used as molecular weight standards. 5’-End Labeling of Transcripts and RNA Sequence Analysis RNA-3 and RNA-l were labeled at the 5’ end by in vitro transcription. The reaction mixture consisted of 100 mM KCI, 40 mM Tris-HCI (pH 8.0). 10 mM MQCI., 1 mM dithiothreitol. 50 PM EDTA. 150 AM CTP. GTP and UTP, 15 PM r3*P-ATP (7000 Ci/mmole). 300 ng template and 7 ~1 of RNA polymerase in a total volume of 70 PI. After incubation for 1 hr at 37’C. the reaction was terminated by addition of SDS to 0.2%. followed by phenol extraction in the presence of 25 j.t~ E. coli tRNA and ethanol precipitation. The pellet was dissolved in 5 pl of
uvrB Promoters 529
water, 50 cl of formamide solution were added and the labeled transcripts were separated by electrophoresis in 6% acrylamide and 9 M urea eels. FOliOWing autoradiography, RNA was excised tom the gel and eluted OVemiQht at 4°C in 15 mM NaCI. 1.5 mM sodium citrate, 200 mM sodium acetate (PH 7.0) and 0.3% diethylpyrocarbonate. E. coli tRNA (25-50 PQ) was added, and the RNA was then precipitated in ethanol, extracted with phenol and dissolved in water. Labeling of RNA-2 was achieved by treating purified dephosphorylated RNA with T4 polynucleotide kinase (Pederson and Haeeltine, 1980). The 540 bp Bst NI fragment was transcribed in vitro under the conditions described above for 5’-end labeling except that all nucieotides were present at 150 CM and no labeled precursor was added. Transcription was terminated after 20 min by addition of formamidedye mixture, and the unlabeled transcripts were separated on a 6% acrylamide and 9 M urea gel; transcripts labeled with u-3*P-CTP were run in adjacent lanes as markers. Unlabeled RNA-2 was excised, &ted and ethanol-precipitated as described above except that carrier RNA was omitted. The RNA was dissolved in 10 cl of 10 mM Tris-HCI (pH 8.0) and 1 mM Nan EDTA and was dephosphorylated by the addition of 2.8 units bacterial alkaline phosphatase (Bethesda Research Laboratories). then incubated at 40°C for 1 hr. The endlabeling reaction was initiated by addition of 1 fl each of 100 mM KH,PO, (pH 9.5). 100 mM MQ&. 50 mM dithiothreitol and 1.3 cl T4 poiynucleotide kinase (Bethesda Research Laboratories, IO U/PI). The entire mixture was transferred to a tube containing 100 pmole lyophilized r3’P-ATP (7000 Ci/mmole) and incubated at 37’C for 4.5 hr. after which the RNA was precipitated in 2 M ammonium acetate, 2.5 volumes of ethanol and 10 PQ E. coli tRNA. The transcript was further purified by electrophoresis in a 6% acrylamide and 9 M urea gel, excised, eiuted and precipitated as described above. We sequenced 6’+nd-labeled transcripts by enzymatic methods. The reaction conditions for ribonucleases Tl (Sanyo). U2 and PhyM (P-L Biochemical@ and the base-generated ladder were essentially as described (Donis-Keller et al., 1977; Don&-Keller. 1980). except that carrier RNA was added only when necessary to bring the RNA concentration to 1.5-2 PQ per 25 PI reaction, and digestion with U2 was performed at pH 3.5 (DaSQupta et al., 1980). Digestion with 8. cereus ribonuclease (P-L Biochemicals) followed the procedure of Lockard et al. (1978) except that a single 15 min incubation was used. Digestion products were analyzed on 20% or 10% acrylamide gels in 8 M urea (Maxam and Gilbert, 1980). LEXA Footprinting DNAase I digestion of end-labeled 540 bp Bst NI fragment in the presence of LEXA was used to determine the location of the LEXA binding site within this fragment (footprinting. Galas and Schmltz, 1979; Johnson et al.. 1979). The experiments were performed essentially as described for the recA and /exA operators (Little et al., 1981). The 198 bp Hae Ill-Eco RI fragment at about 10 nM was incubated in the presence or absence of LEXA (monomer concentration, 100 nM) at 22°C for 10 min in a reaction mixture COntaininQ 20 mM Tris-HCI (pH 7.4). 50 mM NaCi. 2.5 mM MQCI,, 1.5 mM CaCIZ. 1 mM dithiothreitol, 100 CM Nan EDTA and 0.1% bovine serum albumin. DNAase I (Worthington) was added to a concentration of 10 n&ml. and incubation was continued for another 10 min. The samples were treated and analyzed as described by Galas and Schmitz (1978). Acknowledgments This work was supported by grants from the National Institutes of Health and a grant from the American Cancer Society. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received
November
19, 1981:
revised
December
A., Kenyon,
C. J. and Walker,
G. C. (1981).
lnducibility
in Es&e-
Boyle, J. M. and Setiow. R. B. (1970). Correlations between host-cell reactivation. ultraviolet reactivation and pyrimidine dimer excision in the DNA of bacteriophage X. J. Mol. Biol. 51, 131-l 44. Brent, R. and Ptashne, M. (1980). The /exA gene product represses its own promoter. Proc. Nat. Acad. Sci. USA 77, 1932-l 938. Brent, R. and Ptashne. M. (1981). Mechanism of action gene product. Proc. Nat. Acad. Sci. USA, 78, 4204-4208.
of the /exA
Carmichael, G. G. and McMaster. G. K. (1980). The analysis of nucleic acids using glyoxal and acridine orange. Meth. Enzymol. 65, 380-391. Castellazzi. M.. Jacques, M. and George, deoxyribonucleic acid repair in Escherichia 143, 703-709.
J. (1980). tif-stimulated co/i K-l 2. J. Bacterial.
Craig, N. L. and Roberts, J. W. (1960). E. co/i recA protein-directed cleavage of X repressor requires polynucleotide. Nature 283, 28-30. Dasgupta. R.. Ahlquist, P. and Kaesberg, P. (1980). Sequence of the 3’ untranslated region of brome mosaic virus coat protein messenger RNA. Virology 104, 339-346. DiLauro, R., Taniguchi. T., Musso. R. and decrombrugghe, B. (1979). Unusual location and function of the operator in the Escherichia co/i galactose operon. Nature 279, 494-500. Donis-Keller. A residues 3133-3142.
H. (1980). Phy M: an RNase activity specific for U and useful in RNA sequence analysis. Nucl. Acids. Res. 8.
Don&Keller, H., Maxam, A. M. and Gilbert, W. (1977). Mapping adenines. guanines. and pyrimidines in RNA. Nucl. Acids. Res. 4, 2527-2538. Erdmann. sequences
V. A. (1981). Collection of published 5S and 5.85 and their precursors. Nucl. Acids Res. 9, r25-r42.
FOQiianO. M. and Schendel, P. (1981). Evidence the uvn3 operon. Nature 289, 196-l 98. Galas, D. and Schmitz, method for the detection Acids Res. 5, 3157-3170.
RNA
for the inducibility
of
A. (1978). DNase footprinting: a simple of protein-DNA binding specificity. Nucl.
Gauss, D. H. and Sprinzi. M. (1981). Nucl. Acids Res. 9, rl -r23.
Compilation
of tRNA sequences.
Howard-Flanders, P. and Boyce, R. P. (1966). DNA repair and genetic recombination: studies on mutants of Escherichia co/i defective in these processes. Radiation Res. Suppl. 6, 158-l 84. Huisman, 0. and D’Ari, R. (1981). An inducible DNA replication-cell division coupling mechanism in E. co/i. Nature 290. 797-799. Johnson, between repressor.
A., Meyer, B. J. and Ptashne, M. (1979). Interactions DNA-bound repressors govern regulation by the h phage Proc. Nat. Acad. Sci. USA 78, 5061-5065.
Kacinski. B. M., Sancar. A. and Rupp. W. D. (1981). A general approach for purifying proteins encoded by cloned genes without using a functional assay: isolation of the UVRA gene product from radiolabelled maxicells. Nucl. Acids Res. 9, 4495-4508. Kato. T. and Shinoura, Y. (1977). Isolation and characterization mutants of Escherichia co/i deficient in induction of mutations ultraviolet light. Mol. Gen. Genet. 156, 121-131.
of by
Kenyon. C. J. and Walker, G. C. (1980). DNA-damaging agents stimulate gene expression at specific loci in Escherichia co/i. Proc. Nat. Acad. Sci. USA 77, 281 Q-2823. Kenyon. C. J. and Walker, G. C. (1981). Expression uvrA gene is inducible. Nature 289, 808-810.
of the E. co/i
Kneser, tivation
and UV-reac-
H. (1968). Relationship between K-reactivation of bacteriophage X. Virology 36, 303-305.
Lee, N. L.. Gielow. W. 0. and Wallace, R. G. (1981). Mechanism of araC autoregulation and the domains of two overlapping promoters, PC and Pero. in the L-arabinose regulatory region of Escherichia co/i. Proc. Nat. Acad. Sci. USA 78, 752-756.
30. 1981
References Bagg.
gene product required for UV and chemical mutagenesis richia co/i. Proc. Nat. Acad. Sci. USA 78, 5749-5753.
of a
Little, J. W.. Edmiston,
S. H., Pacelli.
L. 2. and Mount,
D. W. (1980).
Cell 530
Cleavage of the fscherichia co/i leti protein Proc. Nat. Acad. Sci. USA 77, 3225-3229.
by the recA
protease.
Little, J. W., Mount, D. W. and Yanisch-Perron, C. R. (1961). Purified /exA protein is a repressor of the recA and /exA genes. Proc. Nat. Acad. Sci. USA, 78,4199-4203. Lockard, R. E.. Abner-Deweerd, B., Heckman. J. E., MacGee. J.. Tabor. M. W. and RajBhandary. U. L. (1978). Sequence analysis of 5’ t3*P) labelled mRNA and tRNA using polyacrylamide gel electre phoresis. Nucl. Acids Res. 5, 37-56. Maxam, A. M. and Gilbert, W. (1980). with base-specific chemical cleavages.
Sequencing end-labelled DNA Meth. Enzymol. 65499-560.
McDonell, M. W., Simon, M. N. and Studier, F. W. (1977). Analysis of restriction fragments of T7 DNA and determination of molecular weights by electrophoresis in neutral and alkaline gels. J. Mol. Biol. 110, 119-146. Morimyo, M. and Shimazu, Y. (1976). Evidence that the gene uvr8 is indispensable for polymerase I deficient strain of Escherichia co/i K12. Mol. Gen. Genet. 147, 243-250. Mount, D. W.. Kosel, C. K. and Walker, A. (1976). Inducible. errorfree DNA repair in tsl recA mutants of E. co/i. Mol. Gen. Genet. 146, 37-41. Pederson. F. S. and Hasaltine, W. A. (1960). A micromethod detailed characterization of high molecular weight RNA. Meth. zymol. 84, 660-687.
for En-
Radman, M. (1974). Phenomenology of an inducible mutagenic DNA repair pathway in Escherichia co/i: SOS repair hypothesis. In Molecular and Environmental Aspects of Mutagenesis, L. Prakash. F. Sherman. M. Miller, C. Lawrence and H. W. Tabor, eds. (Springfield: Charles C Thomas), pp. 126-l 42. Radman, M. and Devoret, R. (1971). UV-reactivation h in excision repair deficient hosts: independence and attachment regions. Virology 43, 504-506.
of bacteriophage of red functions
Roberts, J. W., Roberts, C. W. and Mount, D. W. (1977). Inactivation and proteolytic cleavage of phage h repressor in vitro in an ATPdependent reaction. Proc. Nat. Acad. Sci. USA 72, 2330-2334. Rosenberg, M. and Court, D. (1979). Regulatory sequences involved in the promotion and termination of RNA transcription. Ann. Rev. Genet. 13, 319-353. Rothman. R. H., Margossian. L. J. and Clark, reactivation of phage lambda in recF, recL, uvrA. of E. co/i K-i 2. Mol. Gen. Genet. 169, 279-287.
A. J. (1979). Wand uvr8 mutants
Sancar. A. and Rupert. C. S. (1978). Determination lecular weights from ultraviolet sensitivities. Nature
of plasmid mo272, 471-472.
Sancar. A. and Rupp, W. D. (1979). Physical Proc. Nat. Acad. Sci. USA 76, 3144-3148.
map of the recA
gene.
Sancar, A., Stachelek. C., Konigsberg. W. and Rupp. W. D. (1980). Sequences of the recA gene and protein. Proc. Nat. Acad. Sci. USA 77, 261 l-261 5. Sancar, A., Clarke, N. D., Griswold, J.. Kennedy, W. J. and Rupp. W. D. (1981). Identification of the uvr8 gene product. J. Mol. Biol. 748, 63-76. Sanger, F.. Coulson, A. R., Friedmann, Brown, N. L.. Fiddes. J. C., Hutchinson. Smith, M. J. (1978). The nucleotide @Xl 74. J. Mol. Biol. 125, 225-248.
T.. Air, G. N.. Barrel& B. G., C. A. II, Slocombe. P. M. and sequence of bacteriophage
Schwarz, E., Schere, G., Hobom, and Kossel. H. (1978). Nucleotide sequence of cfo, c/l, and part of the 0 gene in phage DNA. Nature 272, 410-414. Shizuya, H. and Dykhuizen, D. (1972). Conditional lethality of deletions which include uvf8 in strains of Escherichia co/i lacking deoxynucleic acid polymerase I. J. Bactertol. 112, 676-881. Steib, J. A. (1979). RNA-RNA interactions during polypeptide chain initiation. In Ribosomes, G. Chambliss, G. R. Craven, J. Davies, L. Kahan and M. Nomura, eds. (Baltimore: University Park Press), pp. 349-399. Weigle. J. J. (1953). Induction of mutation Nat. Acad. Sci. USA 39, 628-636.
in a bacterial
virus.
Proc.
Witkin. E. M. (1976). Ultraviolet mutagenesis and inducible repair in Escberichia co/i. Bacterial. Rev. 40, 869-907.
DNA
Zurawski, G.. Gunsalus. R. B., Brown, K. D. and Yanofsky. C. (1981). Structure and regulatton of afofi, the structural gene for the tryptophan-repressible 3-deoxy-arabino-heptulosonic acid-7-phosphate synthetase In Escherichia co/i. J. Mol. Blol. 145. 47-73. Note
Added
in Proof
We have now established that the sequence of the amino terminus of uvrB protein is Ser-Lys-Pro-Phe-Lys-Leu-Asn-Ser-Ala-Phe, indicating that translation of uvrB message begins at the first fMet codon at position +34. Recently, van den Berg et al. (Nucl. Acids Res. 9, 5623-5643. 1981) also have identified two promoters of the uvr8 gene, which correspond to the promoters named Pl and P2 in this report.