Comparative Gene Expression Profiles Following UV Exposure in Wild-Type and SOS-Deficient Escherichia coli

Corresponding author: Justin Courcelle, Department of Biological Science, P.O. Box GY, Mississippi State University, Mississippi State, MS 39762., jcourcelle{at}biology.msstate.edu (E-mail)

Communicating editor: P. L. FOSTER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

The SOS response in UV-irradiated Escherichia coli includes the upregulation of several dozen genes that are negatively regulated by the LexA repressor. Using DNA microarrays containing amplified DNA fragments from 95.5% of all open reading frames identified on the E. coli chromosome, we have examined the changes in gene expression following UV exposure in both wild-type cells and lexA1 mutants, which are unable to induce genes under LexA control. We report here the time courses of expression of the genes surrounding the 26 documented lexA-regulated regions on the E. coli chromosome. We observed 17 additional sites that responded in a lexA-dependent manner and a large number of genes that were upregulated in a lexA-independent manner although upregulation in this manner was generally not more than twofold. In addition, several transcripts were either downregulated or degraded following UV irradiation. These newly identified UV-responsive genes are discussed with respect to their possible roles in cellular recovery following exposure to UV irradiation.

IRRADIATION of growing Escherichia coli cultures with ultraviolet light (UV) produces DNA lesions that at least transiently block the essential processes of replication and transcription. A large amount of work has demonstrated that the cell responds to this stress by upregulating the expression of several genes that function to repair the DNA lesions, restore replication, and prevent premature cell division. A number of other genes are known to be upregulated, yet remain functionally uncharacterized. The changes in gene expression in response to DNA damage produced by UV and some other environmental agents have been collectively termed the SOS response, after the international distress signal (RADMAN 1974 and reviewed in FRIEDBERG et al. 1995 ; KOCH and WOODGATE 1998 ).

Many of the DNA damage-induced genes are negatively regulated by the LexA repressor protein, which binds to a 20-bp consensus sequence in the operator region of the genes, suppressing their expression (BRENT and PTASHNE 1981 ; LITTLE et al. 1981 ). Derepression of these genes occurs when the RecA protein binds to single-stranded regions of DNA created at replication forks when they are blocked by DNA damage. RecA bound to single-strand DNA becomes conformationally active, serving as a coprotease to cleave the LexA repressor. As the cellular concentration of LexA diminishes, the genes normally suppressed by LexA are more frequently transcribed (SASSANFAR and ROBERTS 1990 and references therein; FRIEDBERG et al. 1995 ). An interesting feature of the LexA/RecA regulatory circuit is that the timing, duration, and level of induction can vary for each LexA-regulated gene, depending upon the location and binding affinity of the LexA box(es) relative to the strength of the promoter. As a result of these properties, some genes may be partially induced in response to even endogenous levels of DNA damage, while other genes appear to be induced only when high or persistent DNA damage is present in the cell. In fact, the SOS response may represent a continuum in the monitoring of environmental stress, rather than simply operating as an emergency switch following acute injury.

The first systematic search for damage-inducible (din) genes was carried out by KENYON and WALKER 1980 by randomly inserting a lac reporter gene into the E. coli chromosome to identify promoters that were upregulated following DNA damage in a recA/lexA-dependent fashion. Using this same technique, subsequent studies identified additional din genes and in some cases identified genes previously characterized to be involved in the recovery from DNA damage (BAGG et al. 1981 ; FOGLIANO and SCHENDEL 1981 ; HUISMAN and D'ARI 1981 ; KENYON and WALKER 1981 ; SHURVINTON and LLOYD 1982 ; LLOYD et al. 1983 ; SIEGEL 1983 ; BONNER et al. 1990 ; IWASAKI et al. 1990 ; OHMORI et al. 1995B ). Analysis of the known din genes revealed a 20-bp consensus LexA-binding motif, or "SOS box," shared by these genes in their promoter/operator regions (WALKER 1984 ), which has been used in more recent studies to systematically search and identify additional lexA-regulated genes (LEWIS et al. 1994 ; OHMORI et al. 1995A ; FERNANDEZ DE HENESTROSA et al. 2000 ). These studies in total have identified 31 genes under lexA/recA control.

Other genes have been reported to be upregulated following DNA damage but are believed to be independent of the lexA regulon. In some cases, the induction is thought to be dependent on recA, but independent from the LexA repressor. In other cases, genes have been shown to be upregulated independently from both recA and lexA (FRIEDBERG et al. 1995 ; KOCH and WOODGATE 1998 ). The mechanism of regulation in these cases is not understood. An additional, although as yet unexplored, possibility is that some genes may be repressed or their transcripts may be degraded in response to DNA damage.

It was of interest to us not only to learn whether additional genes can be regulated in a LexA-dependent manner but also to determine whether other cellular responses to UV irradiation exist that are lexA independent. The lexA1 allele encodes an amino acid change at a position that is essential for the cleavage and inactivation of LexA (SLILATY and LITTLE 1987 ). Thus, in lexA1 mutants, the LexA1 concentration remains high and LexA-regulated genes are not induced, even in the presence of high levels of activated RecA. Therefore this system should allow an analysis of gene expression that occurs independent of the LexA repressor.

The changes in gene expression in the entire genome can be measured simultaneously using high-density cDNA microarrays (SCHENA et al. 1995 ). DNA microarrays contain PCR-amplified DNA fragments of known and predicted genetic sequences that are printed on the surface of a glass slide. Through the comparative hybridization of two cellular RNA preparations, the relative difference between transcript levels of any gene in these preparations can be determined. Using the complete sequence of the E. coli genome (BLATTNER et al. 1997 ), DNA microarrays were prepared containing PCR products corresponding to 95.5% (4101 out of 4295) of all annotated open reading frames in the E. coli genome (KHODURSKY et al. 2000 ). We utilized these microarrays to follow the changes in gene expression occurring during the first hour following UV irradiation in the wild-type strain, MG1655, and in an isogenic lexA1 mutant.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

Bacteria:
Strain MG1655 was used as the wild-type strain in this study since its genome has been completely sequenced (BLATTNER et al. 1997 ). The MG1655 lexA1(Ind-) malB::Tn9 was constructed by P1-mediated transduction of the lexA1 allele from strain GC2281 (TADDEI et al. 1995 ) into strain MG1655. Transfer of the lexA1 allele was verified by resistance to chloramphenicol and hypersensitivity to UV irradiation.

Growth and irradiation:
Cells were grown in Davis medium plus 0.4% glucose. Cultures were inoculated at a 1:200 dilution from a fresh overnight culture into 200 ml Davis media and incubated in a 1-liter Erlenmeyer flask at 37° in a New Brunswick Scientific (Edison, NJ) model G76 gyrotory water bath at 220 rpm to midlog (OD₆₀₀ 0.4, 2 x 10⁸ cells/ml). A 15-W germicidal lamp (254 nm, 0.66 J/m²/sec at the sample position) provided the UV irradiation. A total of 70 ml of culture was placed into a 15-cm-diameter glass petri dish and irradiated for 60 sec with gentle agitation. Two 65-ml unirradiated samples were also agitated in a 15-cm petri dish but were not exposed to UV. A total of 70 ml of irradiated culture (in a 500-ml Erlenmeyer flask) and 30 ml of unirradiated culture (in a 250-ml Erlenmeyer flask) were then returned to the shaking water bath for the duration of the time course. At the appropriate times, 10-ml samples were placed into 20 ml of ice-cold NET (100 mM NaCl, 10 mM Tris, 10 mM EDTA), pelleted by centrifugation, washed with 1 ml cold NET, repelleted, and frozen at -80°. The limited availability of microarray chips constrained this experiment to a single time course containing seven samples (five irradiated, two unirradiated) for each strain.

Microarray procedures:
Relative mRNA levels were determined by parallel two-color hybridization to cDNA microarrays representing 4101 open reading frames (ORFs) representing 95.5% of E. coli ORFs according to BLATTNER et al. 1997 . cDNA arrays were manufactured as described in MGuide at http://cmgm.stanford.edu/pbrown/mguide/index.html. Total mRNA was extracted from 2–5 x 10⁹ cells using QIAGEN (Chatsworth, CA) RNeasy spin columns. A total of 25–30 µg of total RNA was labeled with Cy-3-dUTP (or Cy-5-dUTP) in a standard reverse transcriptase (RT) reaction by Superscript II (+) (GIBCO BRL, Gaithersburg, MD) with 1 µg of random hexamer (Pharmacia, Piscataway, NJ) primers. Following purification through Microcon-30 (Millipore, Bedford, MA) (MGuide), Cy-3- and Cy-5-labeled cDNA were combined with SSC (2.5x final), SDS (0.25%), and 40 µg of E. coli rRNA (Boehringer Mannheim, Indianapolis) in a final volume of 16 ml and hybridized to a DNA microarray for 5 hr at 65°. Slides were washed as described in MGuide and scanned using an AxonScanner (Axon Instruments, Foster City, CA; GenPix 1.0) at 10 mM per pixel resolution. Acquired 16-bit TIFF images were analyzed using ScanAlyze software, which is publicly available at http://rana.stanford.edu/software/.

Comparative measurements of transcript abundance:
Time course samples were analyzed directly by comparing the abundance of each gene's transcripts relative to the t₀ sample. RNA samples taken during the time course were labeled with Cy-5, and RNA from the t₀ sample was labeled with Cy-3.

Sequence analysis:
Nucleotide sequences in the regions of induced genes were examined using the COLIBRI program provided by the Pasteur Institute at http://genolist.pasteur.fr/colibri/. Regions surrounding induced genes were searched for the consensus sequence CTG(N)₁₀CAG, allowing for one mismatch. Matching sequences that fell between -400 and +100 bp of a start codon were then examined for their heterology index. The heterology index was determined as reported in LEWIS et al. 1994 on the basis of the formula developed by BERG and VON HIPPLE (1988). Heterology index = ln[], where n(consensus) refers to the number of times that the most common, or consensus, base occurs at a given position in the set of known binding sites, and n_(actual) refers to the number of times that the base being analyzed occurs at the same position in the set of known binding sites. n values for each position of the 20-bp LexA binding site were determined using the known LexA-binding sites shown in Fig 1A and their respective complementary sequences.

View larger version (496K): In this window In a new window Download PPT slide   Figure 1. Changes in gene expression within the E. coli genome following UV irradiation. The average change in transcript levels in the irradiated samples compared to unirradiated samples for each gene along the E. coli chromosome is plotted sequentially. Open squares, wild type; solid circles, lexA1. The location of genes on the chromosome, in kilobase pairs, is indicated along the top of each graph. The average change in transcript levels was calculated as the (average change in irradiated samples)/(average change in unirradiated samples). The data for all plotted genes represent an average of between 3 and 5 irradiated time points and at least one unirradiated time point. Time points for which the PCR or hybridization failed, or the fluorescent signal generated by the unirradiated sample was <30% above the background level of fluorescence, were not included in the averages.

Nomenclature:
All genes are named according to the Rudd system at http://bmb.med.miami.edu/ecogene/ecoweb (RUDD 2000 ). In cases where we found no corresponding Rudd gene for the open reading frame examined, the original identification numbers of Blattner, b#### (BLATTNER et al. 1997 ), were used.

Raw data:
The raw data from these experiments are available for download at the following web address http://www2.msstate.edu/~jcc129.

	RESULTS AND DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

We examined the response of E. coli strain MG1655 following a dose of 40 J/m² (254 nM). Previous studies in our laboratory have shown that exposing an exponentially growing culture of E. coli to 40 J/m² of UV produces approximately one cyclobutane pyrimidine dimer on each strand per 6 kb of DNA (MELLON and HANAWALT 1989 ; CROWLEY and HANAWALT 1998 ). This dose transiently inhibits both replication and transcription, and induces a strong SOS response (COURCELLE et al. 1997 ; CROWLEY and HANAWALT 1998 ). More than half of the cells survive and genomic replication fully recovers within 45 min following UV irradiation. Within that time, most of the DNA lesions have also been repaired (MELLON and HANAWALT 1989 ; COURCELLE et al. 1999 ).

To examine the changes in gene expression in response to this dose of UV irradiation, we compared samples of total RNA taken 5, 10, 20, 40, and 60 min after irradiation to samples made just prior to irradiation. To control for UV-independent changes, total RNA preparations from nonirradiated samples at 20 and 60 min were also prepared. This analysis was carried out with the wild-type MG1655 strain as well as the isogenic lexA1 derivative and represents the changes in transcript levels of each gene from up to seven independent comparative hybridizations for each cell line, which were observed within the same experiment.

The average change in transcript level in the irradiated samples compared to those in the unirradiated samples for each gene along the E. coli chromosome is plotted sequentially in Fig 1. In some cases, no data were plotted for a gene, because either the PCR reaction failed during microarray construction or the fluorescent signal in the unirradiated samples was too low for reliable detection. However, raw data for any or all genes are available for downloading at the web addresses indicated in MATERIALS AND METHODS or upon request to the authors.

Genes induced in a LexA-dependent manner following UV irradiation:
Twenty-six functional LexA-binding regions controlling at least 31 genes have been previously demonstrated to be functionally active following irradiation. At the time at which these bacterial microarrays were constructed, 3 of these genes, ysdAB, dinQ, and dinS, had not yet been identified as open reading frames (FERNANDEZ DE HENESTROSA et al. 2000 ) and were not included in our analysis. The time courses observed for all other genes/operons known to be regulated by LexA and that contain LexA-binding sites are plotted in Fig 2A. For most of the LexA-regulated genes, the level and timing of the induction observed in our experiments are in good agreement with previous observations. In confirmation of previous studies, we find that recN, recA, and sulA are heavily induced within the first 5 min of irradiation whereas the uvrD induction is much less robust (CASAREGOLA et al. 1982 ; ARTHUR and EASTLAKE 1983 ; SALLES and PAOLETTI 1983 ; PICKSLEY et al. 1984 ; SANDLER 1994 ). umuCumuD are also known to be strongly induced; however, full induction of these genes is not observed until 20 min after UV irradiation (WOODGATE and ENNIS 1991 ). We were unable to assay ftsK induction in wild-type cells due to a problem amplifying the ftsK fragment when constructing the bacterial microarray. However, some lexA-dependent induction is observed in lolA, possibly representing some transcriptional readthrough from the ftsK gene. In some cases, we observed co-upregulation of the neighboring ORFs that are transcribed in the same orientation. This effect can clearly be seen in the induction of dinB transcription, which also renders an increase in yafN, yafO, and yafP mRNA. Similarly, yebF and yebE are also induced along with LexA-regulated yebG. In the case of recN, the downstream genes smpA and smpB also appear to be upregulated following UV irradiation. However, for b2619 and b2618 it is actually the antisense strand that would be transcribed following UV irradiation if this induction represents transcriptional readthrough. The actual mechanism of such coregulation could be produced by: (1) transcriptional readthrough resulting from inefficient transcriptional termination; (2) the actual operon spanning across the entire group of neighboring genes; or (3) a regional effect conferred through protein factors or DNA structural conformations within the region under consideration.

View larger version (63K): In this window In a new window Download PPT slide   Figure 2. Transcriptional induction following UV irradiation in the genes surrounding known LexA boxes. The change in transcript levels for the indicated gene is plotted over time. The arrows indicate the direction of transcription within the operon relative to the LexA box. Arrows pointing left are transcribed on the minus strand and arrows pointing right are transcribed on the plus strand. The locations and distances, in base pairs, of the LexA box from the initial ATG codon are indicated in the boxes. The graphs of genes that are directly adjacent on the chromosome are joined together. The location of the gene(s) on the chromosome, in kilobase pairs, is indicated along the top of each plot. Solid squares, irradiated wild type; open squares, unirradiated wild type; solid circles, irradiated lexA1; open circles, unirradiated lexA1. Time points in which the PCR or hybridization failed, or the fluorescent signal generated by the unirradiated sample was <30% above the background level of fluorescence, are not plotted.

Some of the transcripts from documented LexA-regulated genes, dinG, molR, uvrD, and uvrA, did not significantly rise following UV irradiation. However, in each of these cases, the samples of these transcripts in the unirradiated (control) culture were significantly decreased during the time course. The reason for this observation is unclear. However, both initial and unirradiated samples were "mock" UV treated by gentle agitation for 60 sec in a 15-cm glass petri dish and it is possible that some genes were affected by this treatment. Importantly, however, when comparing the net change in irradiated and unirradiated samples, the lexA-dependent induction of these genes is clearly evident: 1.77-, 1.78-, 2.51-, and 3.85-fold increases, respectively.

Of the reported LexA-regulated genes, we did not detect significant induction of either hokE or ssb in our experiment. recA/lexA-dependent transcription from hokE has previously been shown to occur in the E. coli strain RW118 following mitomycin C treatment (FERNANDEZ DE HENESTROSA et al. 2000 ). However, the SOS induction of ssb is less clear. Although a plasmid-encoded ssb has been shown to be slightly upregulated following SOS induction, no induction of the chromosomally encoded ssb has been reported and SOS induction does not lead to higher levels of SSB protein (BRANDSMA et al. 1983 ; PERRINO et al. 1987 ).

If the LexA-binding site is located within the operator regions of two diverging transcripts, it is possible that a single site may regulate the transcription of both operons. This is presumed to be the case at uvrA and ssb and also appears to occur between ybiA and dinG as well as umuCD and hylE.

In addition to the previously reported lexA-regulated genes, we observed several other genes that appeared to be upregulated in a LexA/RecA-dependent manner (Fig 2B). However, to determine whether these genes are directly under LexA control will require further investigation. Some of these genes do appear to have potential candidate LexA boxes (Table 1B). One method of predicting whether a LexA-like sequence will bind LexA is to examine its heterology index (HI), which is a value derived from a mathematical formula ranking the relatedness of a potential sequence to that of known LexA-binding sequences (LEWIS et al. 1994 ). Low HI values predict that a potential sequence is more likely to bind LexA protein. Previous studies found that lexA sequences with an HI value of <15 generally bound LexA (LEWIS et al. 1994 ) and recently this method was used to identify seven new lexA-regulated genes (FERNANDEZ DE HENESTROSA et al. 2000 ). However, there are exceptions to the predictions from HI values. On the basis of earlier calculations that were based upon a smaller set of lexA sequences, FERNANDEZ DE HENESTROSA et al. 2000 found that although dinJ had an HI value of 7.06, it did not bind LexA. Yet ybfE, which they calculated to have an HI value of 14.07, did bind LexA (LEWIS et al. 1994 ). This observation suggests that all factors comprising a functional LexA box have not yet been identified. Therefore, we have recalculated the HI values using all 28 functional LexA sequences found on the chromosome and we have included any potential LexA-binding sequence with an HI value of <20 in Table 1B.

While some of these newly identified genes appear to have potential LexA-binding sequences, many of the induced genes do not, suggesting that in some cases the regulation may occur indirectly. Indirect lexA-dependent induction of these genes might occur through regulatory proteins that are themselves under lexA regulation, or through other regulatory proteins that are inactivated by a similar RecA-catalyzed proteolysis. The activated form of RecA is also known to induce proteolytic cleavage of other proteins containing "LexA-like" cleavage motifs such as those found in UmuD, the plasmid-encoded MucA, and the repressor proteins of several bacteriophage (LITTLE 1984 ; PERRY et al. 1985 ). This possibility is especially attractive considering that several of the newly identified genes, borD, the lit-intE region, and ogrK, share homology with cryptic prophage genes. The borD gene product is a homolog of the phage Bor protein, a lipoprotein expressed during lysogeny that is present in the outer membrane. The borD gene product shares homology to other bacterial virulence proteins and its expression increases E. coli survival in animal serum (BARONDESS and BECKWITH 1990 , BARONDESS and BECKWITH 1995 ).

lit, intE, and several genes of unknown function in this same region are expressed at relatively late times following UV exposure (Fig 3). lit encodes a protease specific for elongation factor EF-tu. Expression of lit is induced at late times following phage T4 infection and prevents late phase phage amplification through its EF-tu proteolysis (GEORGIOU et al. 1998 ). Following T4 infection, the endogenous lit gene has been suggested to trigger an apoptotic-like death of the infected cell, thereby thwarting the reproduction of the virus and precluding the widespead infection of the population. Further characterization is required to know whether this activity or any of the genes in this region affect recovery following DNA damage.

View larger version (50K): In this window In a new window Download PPT slide   Figure 3. Genes that displayed the largest LexA-dependent transcriptional induction following UV irradiation are plotted. The change in transcript levels for the indicated gene is plotted as in Fig 2. Solid squares, irradiated wild type; open squares, unirradiated wild type; solid circles, irradiated lexA1; open circles, unirradiated lexA1.

ogrK is a prophage gene from a phage P2. Ogr has been shown to regulate late P2 gene transcription through an interaction with the host RNA polymerase (WOOD et al. 1997 ). Additionally, an ogr-like gene, regC in Serratia marcescens, has been shown to be induced following mitomycin C treatment in an SOS-dependent manner (JIN et al. 1996 ).

grxA is a glutoredoxin that acts as a hydrogen donor for the E. coli ribonucleotide reductases. Several thioredoxin and glutoredoxin genes in E. coli are coregulated with ribonucleotide reductase gene expression (PRIETO-ALAMO et al. 2000 ). From this perspective, it is interesting to note that the ribonucleotide reductase genes nrdA and nrdB are among the strongest lexA-independent induced genes following UV exposure (Fig 4). grxA is also induced in an oxyR-dependent manner under some conditions (PRIETO-ALAMO et al. 2000 ).

View larger version (38K): In this window In a new window Download PPT slide   Figure 4. Representation of genes displaying a LexA-independent transcriptional induction following UV irradiation. The change in transcript levels for the indicated gene is plotted as in Fig 2. (A) The largest LexA-independent inductions. (B) Typical profiles of genes with an early UV-dependent transcriptional induction. (C) Typical profiles of genes with a slow UV-dependent transcriptional induction. Solid squares, irradiated wild type; open squares, unirradiated wild type; solid circles, irradiated lexA1; open circles, unirradiated lexA1.

glvB induction is unusual in that glvB lies in the middle of a predicted operon and encodes a portion of a protein transport system. Nevertheless, glvB induction was also observed following exposure to gamma irradiation (data not shown) and may be driven from an alternative promoter or alternative open reading frame in the region.

In the case of yigN, a previous study has demonstrated that it contains a functional LexA binding site; however, no further increase in yigN expression was observed following treatment with mitomycin C in wild-type, lexA51 (deficient), or lexA1 (uninducible) E. coli cultures. We have no clear explanation of this difference. However, alternative promoters proximal to yigN could have allowed for full expression to occur in these previous studies prior to mitomycin treatment since yigN appeared to be heavily expressed under all conditions in that study (FERNANDEZ DE HENESTROSA et al. 2000 ).

ibpA (hslT) and ibpB (hslS), encoding heat-inducible chaperonins, were also induced in a LexA-dependent manner.

There has been little functional characterization of the remaining induced genes. yoaA shares homology with other ATP-dependent helicases. The gene products of ydeT, ydeS, and ydeR share homology to other fimbrial proteins. b1169 ycgH has homology with other ATP-binding subunits of transport systems. Both ydeO and ydiW have motifs that suggest they may function as transcriptional regulators. No significant homology between ybiN, yqgC, yhiJL, or yifL and any other characterized proteins has been reported.

Genes induced independently of LexA following UV irradiation:
The time courses of the largest lexA-independent inductions are plotted in Fig 4A. Most striking is precisely how few lexA-independent changes occur following UV exposure. In general, lexA-independent inductions, with the exception of nrdA, nrdB, and yeeF, are in the range of twofold effects. Furthermore, many of these lexA-independent profiles appear to rise very rapidly (within the first 5 min) and then either subside or plateau. A large number of genes and regions were observed to be regulated in this manner and some generalizations are apparent from both Fig 1 and Table 2. Many proteins associated with the replication machinery are slightly induced following UV irradiation. Additionally, several genes associated with purine and pyrimidine metabolism seem to be upregulated in a similar manner. Although not dramatic, these results are particularly impressive considering that these nucleotide metabolism genes are often found in very small operons spaced throughout the genome. Other categories of genes that appeared to be upregulated included heat-shock or chaperone proteins as well as several of the genes involved in RNA metabolism. It is notable that nearly half of the genes that were upregulated have had little or no functional characterization.

Loss of gene expression following UV irradiation:
Whereas several studies have focused upon the need to upregulate certain gene products following UV irradiation, it has remained relatively unexplored, yet very possible, that repression or even active degradation of some gene transcripts will also be an important factor in cellular recovery. The bacterial microarray offers an opportunity to address this very question. Indeed, repression was observed in a large number of genes following UV irradiation. However, our results do not allow us to determine whether the decrease in a given transcript represents diminished transcription or accelerated degradation in response to UV irradiation. Nevertheless, a large number of genes in the wild-type, but not in the lexA1, samples were reduced in their transcript levels at the 5-, 10-, and 20-min time points following irradiation. This observation may suggest that some inhibition of transcription or degradation of transcripts occurs in a lexA-dependent manner, and it may point to a lexA-dependent mechanism for inhibition of transcription or enhanced degradation of these transcripts. Due to the small sample size of this experiment, we are inclined to interpret these findings cautiously. However, further investigation of these observations is clearly warranted.

The phenomenon of downregulation is also interesting to consider with respect to DNA repair. Actively transcribed genes in E.coli are repaired preferentially compared to nontranscribed genes (MELLON and HANAWALT 1989 ). When the lac operon is actively transcribed at the time of UV irradiation, repair of the transcribed strand occurs within 5 min after irradiation. While it is assumed that this response allows for the rapid transcriptional recovery of expressed genes, little is known about the actual inhibition or recovery of transcription for operons other than lac in E. coli.

Diminished transcript levels were clearly evident in some operons. Several different temporal profiles were observed (Fig 5). In some cases, exemplified by the gat operon, the decrease in transcript levels following UV irradiation was observed in both wild-type and lexA1 cells (Fig 5A). In the wild-type cells, the time at which transcripts recovered to preirradiation levels varied but generally occurred prior to 40 min postirradiation. Interestingly, whereas transcripts recovered to pretreatment levels in the wild-type cells, transcripts of several genes failed to recover in lexA1 mutants within the period observed in these experiments, suggesting that lexA-regulated genes have an important role in this transcriptional recovery (Fig 5A).

View larger version (45K): In this window In a new window Download PPT slide   Figure 5. Representation of genes that displayed a reduced level of transcripts following UV irradiation. The change in transcript levels for the indicated gene is plotted as in Fig 2. (A) A typical operon in which the transcript from both wild-type cells and lexA1 mutants was observed to be reduced. (B) An operon in which the transcript from wild-type cells was observed to be severely reduced at the 5' end. (C) Three operons in which transcripts were reduced in wild-type cells throughout the operon.

The gat operon, controlling galactitol uptake and metabolism, is one of several operons involved in the metabolism of different carbon sources, which were found to have reduced transcript levels following UV irradiation. Other carbon metabolism operons whose transcript levels decreased include fru, man, wwb, glg, and mal . Our cultures were grown with glucose as the sole carbon source. It may be interesting to know how these metabolic pathways are regulated when cells are grown in the presence of their respective carbon sources.

A second form of repression profile is exemplified by the rfa operon, which encodes gene products involved in lipid synthesis in the membrane. The rfa operon displayed a loss of transcript following UV irradiation in wild-type cells, but not in the lexA1 mutant. The loss of transcript was more severe at the 5' end of the operon, which may at least partially reflect the fact that most RNA degradation in E. coli is believed to occur 3'–5' (Fig 5B).

A third pattern of transcript reduction can be seen in nrdHIEFproVWX and fruAKB operons shown in Fig 5C. In these cases, loss of the wild-type transcripts occurred uniformly throughout the operons.

The most severely reduced transcripts from operons for which we obtained signal are listed in Table 3. Of interest to note is the repression of operons such as minCDE following UV irradiation. These genes are required to induce and regulate septum formation prior to cell division and it is interesting to consider the loss of these transcripts with respect to the induction of other genes that regulate cell division such as sulA and ftsK. Also worth noting is that minC, which is severely repressed, actually contains a LexA-like box, which is predicted to be a very good LexA-binding sequence (FERNANDEZ DE HENESTROSA et al. 2000 ). Dramatic reductions in transcript levels at predicted LexA boxes were also observed for rfaJ and metE (Fig 5B and Table 3). While it is impossible from these experiments to determine if the repression is directly due to a LexA regulation, we did observe cases where functional LexA boxes appeared to result in inhibition of gene expression following UV irradiation. The loss of transcription from araDAB following UV irradiation seems likely to be due to the upregulation of polB, which is located proximal to this operon (Fig 1 and Table 3). Thus, although LexA serves as a negative regulator for polB, it essentially behaves as a positive regulator of araDAB expression.

The study we have presented should serve as a starting point for follow-up projects. It provides some generalization with respect to the role of LexA in the regulation (up and possibly down) of genes in response to one type of environmental stress. Surprisingly, it shows us that, in the absence of LexA regulation, there are no other major responses to UV irradiation at the level of transcriptional regulation. It remains to be determined whether the minor changes are significant in terms of the overall stress response. The values and raw data for any or all of the genes can be retrieved either via the web locations listed in MATERIALS AND METHODS or upon request.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We thank Nicholas Cozzarelli and David Botstein for initiating and supporting experiments with bacterial microarrays. We thank C. Richmond, Y. Wei, and F. Blattner for their help in preparing a library of E. coli ORFs. J.C. is currently being supported by a European Molecular Biology Organization fellowship ALTF 265-1999. A.K. is a PMMB/NSF Postdoctoral Fellow (grant DMS-9406348). B.P. was supported by the National Institutes of Health (NIH), Molecular Biophysics training grant GM08295-11. The work was supported by NIH grants to Nicholas R. Cozzarelli and David Botstein. P.O.B. was supported by the Howard Hughes Medical Institute and NIH grant HG00983. P.O.B. is an associate investigator at the Howard Hughes Medical Institute. Research by P.C.H. is supported by an Outstanding Investigator grant CA44349 from the National Cancer Institute.

Manuscript received October 4, 2000; Accepted for publication January 29, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

ARTHUR, H. M. and P. B. EASTLAKE, 1983  Transcriptional control of the uvrD gene of Escherichia coli. Gene 25:309-316[Medline].

BAGG, A., C. J. KENYON, and G. C. WALKER, 1981  Inducibility of a gene product required for UV and chemical mutagenesis in Escherichia coli. Proc. Natl. Acad. Sci. USA 78:5749-5753[Abstract/Free Full Text].

BARONDESS, J. J. and J. BECKWITH, 1990  A bacterial virulence determinant encoded by lysogenic coliphage lambda. Nature 346:871-874[Medline].

BARONDESS, J. J. and J. BECKWITH, 1995  bor gene of phage lambda, involved in serum resistance, encodes a widely conserved outer membrane lipoprotein. J. Bacteriol. 177:1247-1253[Abstract/Free Full Text].

BERG, O. G. and P. H. VON HIPPEL, 1988  Selection of DNA binding sites by regulatory proteins. II. The binding specificity of cyclic AMP receptor protein to recognition sites. J. Mol. Biol. 200:709-723[Medline].

BLATTNER, F. R., G. R. PLUNKETT, C. A. BLOCH, N. T. PERNA, and V. BURLAND et al., 1997  The complete genome sequence of Escherichia coli K-12 (comment). Science 277:1453-1474[Abstract/Free Full Text].

BONNER, C. A., S. HAYS, K. MCENTEE, and M. F. GOODMAN, 1990  DNA polymerase II is encoded by the DNA damage-inducible dinA gene of Escherichia coli. Proc. Natl. Acad. Sci. USA 87:7663-7667[Abstract/Free Full Text].

BRANDSMA, J. A., D. BOSCH, C. BACKENDORF, and P. VAN DE PUTTE, 1983  A common regulatory region shared by divergently transcribed genes of the Escherichia coli SOS system. Nature 305:243-245[Medline].

BRENT, R. and M. PTASHNE, 1981  Mechanism of action of the lexA gene product. Proc. Natl. Acad. Sci. USA 78:4204-4208[Abstract/Free Full Text].

CASAREGOLA, S., R. D'ARI, and O. HUISMAN, 1982  Quantitative evaluation of recA gene expression in Escherichia coli. Mol. Gen. Genet. 185:430-439[Medline].

COURCELLE, J., C. CARSWELL-CRUMPTON, and P. C. HANAWALT, 1997  recF and recR are required for the resumption of replication at DNA replication forks in Escherichia coli. Proc. Natl. Acad. Sci. USA 94:3714-3719[Abstract/Free Full Text].

COURCELLE, J., D. J. CROWLEY, and P. C. HANAWALT, 1999  Recovery of DNA replication in UV-irradiated Escherichia coli requires both excision repair and recF protein function. J. Bacteriol. 181:916-922[Abstract/Free Full Text].

CROWLEY, D. J. and P. C. HANAWALT, 1998  Induction of the SOS response increases the efficiency of global nucleotide excision repair of cyclobutane pyrimidine dimers, but not 6-4 photoproducts, in UV-irradiated Escherichia coli. J. Bacteriol. 180:3345-3352[Abstract/Free Full Text].

FERNANDEZ DE HENESTROSA, A. R., T. OGI, S. AOYAGI, D. CHAFIN, and J. J. HAYES et al., 2000  Identification of additional genes belonging to the LexA regulon in Escherichia coli. Mol. Microbiol. 35:1560-1572[Medline].

FOGLIANO, M. and P. F. SCHENDEL, 1981  Evidence for the inducibility of the uvrB operon. Nature 289:196-198[Medline].

FRIEDBERG, E. C., C. W. GRAHAM and W. SIEDE, 1995 DNA Repair and Mutagenesis. ASM Press, Washingtion, DC.

GEORGIOU, T., Y. N. YU, S. EKUNWE, M. J. BUTTNER, and A. ZUURMOND et al., 1998  Specific peptide-activated proteolytic cleavage of Escherichia coli elongation factor Tu (comment). Proc. Natl. Acad. Sci. USA 95:2891-2895[Abstract/Free Full Text].

HUISMAN, O. and R. D'ARI, 1981  An inducible DNA replication-cell division coupling mechanism in E. coli. Nature 290:797-799[Medline].

IWASAKI, H., A. NAKATA, G. C. WALKER, and H. SHINAGAWA, 1990  The Escherichia coli polB gene, which encodes DNA polymerase II, is regulated by the SOS system. J. Bacteriol. 172:6268-6273[Abstract/Free Full Text].

JIN, S., Y. CHEN, G. E. CHRISTIE, and M. J. BENEDIK, 1996  Regulation of the Serratia marcescens extracellular nuclease: positive control by a homolog of P2 Ogr encoded by a cryptic prophage. J. Mol. Biol. 256:264-278[Medline].

KENYON, C. J. and G. C. WALKER, 1980  DNA-damaging agents stimulate gene expression at specific loci in Escherichia coli. Proc. Natl. Acad. Sci. USA 77:2819-2823[Abstract/Free Full Text].

KENYON, C. J. and G. C. WALKER, 1981  Expression of the E. coli uvrA gene is inducible. Nature 289:808-810[Medline].

KHODURSKY, A. B., B. J. PETER, M. B. SCHMID, J. DERISI, and D. BOTSTEIN et al., 2000  Analysis of topoisomerase function in bacterial replication fork movement: use of DNA microarrays. Proc. Natl. Acad. Sci. USA 97:9419-9424[Abstract/Free Full Text].

KOCH, W. H., and R. WOODGATE, 1998 The SOS response, pp. 107–134 in DNA Damage and Repair: DNA Repair in Prokaryotes and Lower Eukaryotes, edited by J. A. NICKOLOFF and M. F. HOEKSTRA. Humana Press, Totowa, NJ.

LEWIS, L. K., G. R. HARLOW, L. A. GREGG-JOLLY, and D. W. MOUNT, 1994  Identification of high affinity binding sites for LexA which define new DNA damage-inducible genes in Escherichia coli. J. Mol. Biol. 241:507-523[Medline].

LITTLE, J. W., 1984  Autodigestion of lexA and phage lambda repressors. Proc. Natl. Acad. Sci. USA 81:1375-1379[Abstract/Free Full Text].

LITTLE, J. W., D. W. MOUNT, and C. R. YANISCH-PERRON, 1981  Purified lexA protein is a repressor of the recA and lexA genes. Proc. Natl. Acad. Sci. USA 78:4199-4203[Abstract/Free Full Text].

LLOYD, R. G., S. M. PICKSLEY, and C. PRESCOTT, 1983  Inducible expression of a gene specific to the RecF pathway for recombination in Escherichia coli K12. Mol. Gen. Genet. 190:162-167[Medline].

MELLON, I. and P. C. HANAWALT, 1989  Induction of the Escherichia coli lactose operon selectively increases repair of its transcribed DNA strand. Nature 342:95-98[Medline].

OHMORI, H., E. HATADA, Y. QIAO, M. TSUJI, and R. FUKUDA, 1995a  dinP, a new gene in Escherichia coli, whose product shows similarities to UmuC and its homologues. Mutat Res 347:1-7[Medline].

OHMORI, H., M. SAITO, T. YASUDA, T. NAGATA, and T. FUJII et al., 1995b  The pcsA gene is identical to dinD in Escherichia coli. J. Bacteriol. 177:156-165[Abstract/Free Full Text].

PERRINO, F. W., D. C. REIN, A. M. BOBST, and R. R. MEYER, 1987  The relative rate of synthesis and levels of single-stranded DNA binding protein during induction of SOS repair in Escherichia coli. Mol. Gen. Genet. 209:612-614[Medline].

PERRY, K. L., S. J. ELLEDGE, B. B. MITCHELL, L. MARSH, and G. C. WALKER, 1985  umuDC and mucAB operons whose products are required for UV light- and chemical-induced mutagenesis: UmuD, MucA, and LexA proteins share homology. Proc. Natl. Acad. Sci. USA 82:4331-4335[Abstract/Free Full Text].

PICKSLEY, S. M., P. V. ATTFIELD, and R. G. LLOYD, 1984  Repair of DNA double-strand breaks in Escherichia coli K12 requires a functional recN product. Mol. Gen. Genet. 195:267-274[Medline].

PRIETO-ALAMO, M. J., J. JURADO, R. GALLARDO-MADUENO, F. MONJE-CASAS, and A. HOLMGREN et al., 2000  Transcriptional regulation of glutaredoxin and thioredoxin pathways and related enzymes in response to oxidative stress. J. Biol. Chem. 275:13398-13405[Abstract/Free Full Text].

RADMAN, M., 1974 Phenomenology of an inducible mutagenic DNA repair pathway in Escherichia coli: SOS repair hypothesis, pp. 128–142 in Molecular and Environmental Aspects of Mutagenesis, edited by L. PRAKASH, F. SHERMAN, M. MILLER, C. LAWRENCE and H. W. TABOR. Charles C. Thomas, Springfield, IL.

RUDD, K. E., 2000  EcoGene: a genome sequence database for Escherichia coli K-12. Nucleic Acids Res. 28(1):60-64[Abstract/Free Full Text].

SALLES, B. and C. PAOLETTI, 1983  Control of UV induction of recA protein. Proc. Natl. Acad. Sci. USA 80:65-69[Abstract/Free Full Text].

SANDLER, S. J., 1994  Studies on the mechanism of reduction of UV-inducible sulAp expression by recF overexpression in Escherichia coli K-12. Mol. Gen. Genet. 245:741-749[Medline].

SASSANFAR, M. and J. W. ROBERTS, 1990  Nature of the SOS-inducing signal in Escherichia coli. The involvement of DNA replication. J. Mol. Biol. 212:79-96[Medline].

SCHENA, M., D. SHALON, R. W. DAVIS, and P. O. BROWN, 1995  Quantitative monitoring of gene expression patterns with a complementary DNA microarray (comment). Science 270:467-470[Abstract/Free Full Text].

SHURVINTON, C. E. and R. G. LLOYD, 1982  Damage to DNA induces expression of the ruv gene of Escherichia coli. Mol. Gen. Genet. 185:352-355[Medline].

SIEGEL, E. C., 1983  The Escherichia coli uvrD gene is inducible by DNA damage. Mol. Gen. Genet. 191:397-400[Medline].

SLILATY, S. N. and J. W. LITTLE, 1987  Lysine-156 and serine-119 are required for LexA repressor cleavage: a possible mechanism. Proc. Natl. Acad. Sci. USA 84:3987-3991[Abstract/Free Full Text].

TADDEI, F., I. MATIC, and M. RADMAN, 1995  cAMP-dependent SOS induction and mutagenesis in resting bacterial populations. Proc. Natl. Acad. Sci. USA 92:11736-11740[Abstract/Free Full Text].

WALKER, G. C., 1984  Mutagenesis and inducible responses to deoxyribonucleic acid damage in Escherichia coli. Microbiol Rev 48:60-93[Free Full Text].

WOOD, L. F., N. Y. TSZINE, and G. E. CHRISTIE, 1997  Activation of P2 late transcription by P2 Ogr protein requires a discrete contact site on the C terminus of the alpha subunit of Escherichia coli RNA polymerase. J. Mol. Biol. 274:1-7[Medline].

WOODGATE, R. and D. G. ENNIS, 1991  Levels of chromosomally encoded Umu proteins and requirements for in vivo UmuD cleavage. Mol. Gen. Genet. 229:10-16[Medline].

This article has been cited by other articles:

				J. Serment-Guerrero, M. Brena-Valle, and J. J. Espinosa-Aguirre In vivo role of Escherichia coli single-strand exonucleases in SOS induction by gamma radiation Mutagenesis, July 1, 2008; 23(4): 317 - 323. [Abstract] [Full Text] [PDF]

				I. Tagkopoulos, Y.-C. Liu, and S. Tavazoie Predictive Behavior Within Microbial Genetic Networks Science, June 6, 2008; 320(5881): 1313 - 1317. [Abstract] [Full Text] [PDF]

				D. Gawel, P. T. Pham, I. J. Fijalkowska, P. Jonczyk, and R. M. Schaaper Role of Accessory DNA Polymerases in DNA Replication in Escherichia coli: Analysis of the dnaX36 Mutator Mutant J. Bacteriol., March 1, 2008; 190(5): 1730 - 1742. [Abstract] [Full Text] [PDF]

				R. P. da Rocha, A. C. de Miranda Paquola, M. do Valle Marques, C. F. M. Menck, and R. S. Galhardo Characterization of the SOS Regulon of Caulobacter crescentus J. Bacteriol., February 15, 2008; 190(4): 1209 - 1218. [Abstract] [Full Text] [PDF]

				R. E. de Cristobal, P. A. Vincent, and R. A. Salomon A Combination of sbmA and tolC Mutations in Escherichia coli K-12 Tn10-Carrying Strains Results in Hypersusceptibility to Tetracycline J. Bacteriol., February 15, 2008; 190(4): 1491 - 1494. [Abstract] [Full Text] [PDF]

				M. Abella, S. Campoy, I. Erill, F. Rojo, and J. Barbe Cohabitation of Two Different lexA Regulons in Pseudomonas putida J. Bacteriol., December 15, 2007; 189(24): 8855 - 8862. [Abstract] [Full Text] [PDF]

				K. Inoue, S. Basu, and M. Inouye Dissection of 16S rRNA Methyltransferase (KsgA) Function in Escherichia coli J. Bacteriol., December 1, 2007; 189(23): 8510 - 8518. [Abstract] [Full Text] [PDF]

				M. Abella, S. Rodriguez, S. Paytubi, S. Campoy, M. F. White, and J. Barbe The Sulfolobus solfataricus radA paralogue sso0777 is DNA damage inducible and positively regulated by the Sta1 protein Nucleic Acids Res., November 29, 2007; 35(20): 6788 - 6797. [Abstract] [Full Text] [PDF]

				L. M. Pollard, Y. K. Chutake, P. M. Rindler, and S. I. Bidichandani Deficiency of RecA-dependent RecFOR and RecBCD pathways causes increased instability of the (GAA{middle dot}TTC)n sequence when GAA is the lagging strand template Nucleic Acids Res., November 29, 2007; 35(20): 6884 - 6894. [Abstract] [Full Text] [PDF]

				E. Rotman and A. Kuzminov The mutT Defect Does Not Elevate Chromosomal Fragmentation in Escherichia coli Because of the Surprisingly Low Levels of MutM/MutY-Recognized DNA Modifications J. Bacteriol., October 1, 2007; 189(19): 6976 - 6988. [Abstract] [Full Text] [PDF]

				M. Ni, S.-Y. Wang, J.-K. Li, and Q. Ouyang Simulating the Temporal Modulation of Inducible DNA Damage Response in Escherichia coli Biophys. J., July 1, 2007; 93(1): 62 - 73. [Abstract] [Full Text] [PDF]

I. Bjedov, C. N. Dasgupta, D. Slade, S. Le Blastier, M. Selva, and I. Matic Involvement of Escherichia coli DNA Polymerase IV in Tolerance of Cytotoxic Alkylating DNA Lesions in Vivo Genetics, July 1, 2007; 176(3): 1431 - 1440. [Abstract] [Full Text] [PDF]