CroS_R391, an ortholog of the λ Cro repressor, plays a major role in suppressing polV_R391-dependent mutagenesis

Abstract

When subcloned into low-copy-number expression vectors, rumAB, encoding polV_R391 (RumA’₂B), is best characterized as a potent mutator giving rise to high levels of spontaneous mutagenesis in vivo. This is in dramatic contrast to the poorly mutable phenotype when polV_R391 is expressed from the native 88.5kb R391, suggesting that R391 expresses cis-acting factors that suppress the expression and/or the activity of polV_R391. Indeed, we recently discovered that SetR_R391, an ortholog of λcI repressor, is a transcriptional repressor of rumAB. Here, we report that CroS_R391 an ortholog of λ Cro, also serves as a potent transcriptional repressor of rumAB. Levels of RumA are dependent upon an interplay between SetR_R391 and CroS_R391, with the greatest reduction of RumA protein levels observed in the absence of SetR_R391 and presence of CroS_R391. Under these conditions, CroS_R391 completely abolishes the high levels of mutagenesis promoted by polV_R391 expressed from low-copy-number plasmids. Furthermore, deletion of croS_R391 on the native R391 results in a dramatic increase in mutagenesis, indicating that CroS_R391 plays a major role in suppressing polV_R391 mutagenesis in vivo. Inactivating mutations in CroS_R391 therefore have the distinct possibility of increasing cellular mutagenesis that could lead to the evolution of antibiotic resistance of pathogenic bacteria harboring R391.

Abbreviated Summary

When subcloned from R391 into low-copy-number expression vectors, polV_R391 (RumA’₂B) encoded by rumAB, is a potent mutator giving rise to high levels of spontaneous mutagenesis in vivo. This is in contrast to the poorly mutable phenotype when polV_R391 is expressed from the native R391. Here, we report that CroS_R391 an ortholog of λCro, serves as a potent transcriptional repressor of rumAB that plays a major role in suppressing polV_R391 mutagenesis in vivo.

Graphical Abstract

1 |. INTRODUCTION

The accelerating emergence of drug resistant pathogenic micro-organisms is a critical global public health concern (https://www.who.int/news-room/fact-sheets/detail/antibiotic-resistance). Antibiotic over-use and prophylaxis, in both humans and livestock, for the treatment of disease-causing bacteria has led to the selection and increased prevalence of so-called “super-bugs” that have often acquired resistance to multiple classes of antimicrobial compounds (Wendlandt et al., 2015). Estimates suggest a ten-fold increase in mortality rates from untreatable bacterial infections within the next 30 years (https://amr-review.org/sites/default/files/AMR%20Review%20Paper%20-%20Tackling%20a%20crisis%20for%20the%20health%20and%20wealth%20of%20nations_1.pdf). The urgency of this impending crisis has recently led to a concerted effort to identify new synthetic compounds and natural products that might represent novel families of antibiotics (Bhattarai et al., 2020). In addition, the genetic mechanisms that give rise to antibiotic resistance, such as stress-induced mutagenesis and broad-host-range drug-resistance genetic elements (i.e., plasmids, genomic islands and integrative and conjugative elements [ICE]) are being extensively studied as potential targets for mitigating the development of resistance.

E. coli possesses five DNA polymerases, three of which, pol II (encoded by polB), pol IV (encoded by dinB) and polV (encoded by umuDC) are induced as part of the SOS response to stress and DNA damage (Simmons et al., 2008). Expression of these polymerases has previously been shown to contribute to the evolution of antibiotic drug resistance in many pathogenic strains of bacteria (Cirz et al., 2006a; Cirz et al., 2006b; Cirz & Romesberg, 2006; Cirz et al., 2007). PolV alone is responsible for up to a hundred-fold increase in spontaneous mutagenesis after SOS-induction (Sweasy et al., 1990; Fijalkowska et al., 1997). As a consequence, the activity of polV is subject to multiple levels of regulation (Goodman et al., 2016). In addition to transcriptional repression by the SOS repressor, LexA (Bagg et al., 1981), the UmuD protein has to undergo activated RecA (RecA*)-mediated posttranslational cleavage to UmuD’ (Burckhardt et al., 1988; Nohmi et al., 1988; Shinagawa et al., 1988), so that it can interact with UmuC (Woodgate et al., 1989) and generate polV (UmuD’₂C) (Tang et al., 1999). Cleavage of UmuD to UmuD’ also leads to a change in the spatial location of polV inside the cell (Robinson et al., 2015). polVs activity is further increased through an interaction with RecA and ATP to generate the polV Mutasome (polV Mut) (Jiang et al., 2009). Last, but not least, intracellular levels of UmuD, UmuD’ and UmuC are normally kept to a minimum through targeted proteolysis (Frank et al., 1996; Gonzalez et al., 1998; Gonzalez et al., 2000).

polV orthologs are found in many strains of bacteria, as well as bacteriophage, self-transmissible R-plasmids, and integrating conjugative elements (ICEs) (Pinney, 1980; Upton & Pinney, 1983; Ho et al., 1993; McLenigan et al., 1999). Given that these orthologs are located on mobile genetic elements and are likely to find themselves in very different genetic environments, one might expect alternate mechanisms of regulation to keep their activity in check until needed.

A good example is the LexA-regulated polV ortholog (encoded by rumAB) found on R391. R391 and a closely related ICE, SXT, exhibit 95% identity over 65 kb of their sequences (Hochhut et al., 2001). However, one notable difference between R391 and SXT is that rumB is inactivated in SXT due to the insertion of an ISCR2 element into the gene and as a consequence, in contrast to R391, SXT does not possess an active polV.

SetR_SXT a λ cI-like repressor, has been shown to regulate genes involved in conjugation, integration and excision of the SXT element (Beaber et al., 2004; Poulin-Laprade & Burrus, 2015; Poulin-Laprade et al., 2015). Previously, we examined the role of the related SetR_R391 protein for its role in regulating polV_R391 (Gonzalez et al., 2019). We found that the regulatory region of the rumAB_R391 operon contains a single site that is highly similar to the known multiple SetR_SXT 14 bp operator sequences (Figure 1), leading us to suggest that SetR_R391 could be involved in the repression of the rumAB_R391 genes. We found that co-expression of rumAB_R391 and setR_R391 from the same low-copy plasmid reduced the levels of mutagenesis observed in a lexA defective SOS-induced strain indicating that SetR_R391 can regulate the rumAB_R391 genes (Gonzalez et al., 2019). Moreover, SetR_R391 protein was shown to specifically bind to the 14 bp operator sequence that overlaps with the −35 promoter element upstream of rumAB (Figure 1). However, we found that plasmid pRLH421, which contains ~ 21.5 kb of R391, including both setR_R391 and rumAB, nevertheless exhibited very high levels of mutagenesis in an SOS-induced strain [recA718 lexA51(Def)]. We argued that, in this SOS-induced strain, the activated RecA co-protease (encoded by recA718) cleaves the SetR_R391 protein, in a similar fashion to LexA cleavage, inactivating SetR_R391 and allowing expression of the rumAB operon. Indeed, when a non-cleavable allele of setR_R391 was cloned into pRLH421, we observed a significant decrease in mutagenesis (Gonzalez et al., 2019). Additional experiments revealed, however, that intact R391 exhibits even lower levels of mutagenesis in strains in which RecA is in a constitutively and highly activated state (recA730), suggesting that there is at least one additional negative regulator of the rumAB_R391 genes encoded by R391 that is not affected by activated RecA (unpublished observations).

FIGURE 1.

Cartoon of the rumAB promoter region. A single SetR/CroS binding site is shown in blue color. This site partially overlaps with the −35 promoter element, shown in gold. A LexA binding site is shown in green, which partially overlaps with the −10 promoter element (also shown in gold). The ribosome binding site (RBS) is shown in purple and the first two codons of RumA are shown in red.

R391 also expresses a protein that is phylogenetically related to Cro-like transcriptional repressor proteins (Figure S1) (Boltner et al., 2002). While the CroS_R391 protein shows some phylogenetic divergence from related Cro-like proteins, it is 100% identical to the CroS_SXT protein (Figure S1). Previous studies on the CroS_SXT and SetR_SXT proteins revealed that they both bind to four conserved operator sequences; OL, O1, O2 and O3 (albeit with differing affinities) in the croS-setR intergenic region which regulates their own expression from divergent promoters (Poulin-Laprade & Burrus, 2015). In SXT, it is known that CroS_SXT also regulates the setCD_SXT genes and, similar to SetR_SXT, is involved in regulating conjugative transfer (Poulin-Laprade & Burrus, 2015; Poulin-Laprade et al., 2015). The interplay between the ICE SetR_SXT and CroS_SXT proteins is therefore reminiscent of the λcI and λCro proteins that govern the transition between lysogenic and lytic pathways during the bacteriophage life cycle (Takeda et al., 1977; Johnson et al., 1978; Svenningsen et al., 2005).

We hypothesized that like SetR_R391, CroS_R391 would function as a transcriptional regulator of the rumAB_R391 operon and thereby play a role in regulating rumAB_R391-mediated mutagenesis. To test this hypothesis, we have constructed a series of plasmids expressing SetR_R391 and/or CroS_R391 and investigated their effects on rumAB expression and polV_R391-dependent mutagenesis in vivo. We report here, that CroS_R391 is a transcriptional repressor of rumAB and may be the major factor that suppresses polV_R391 activity on R391.

2 |. RESULTS

2.1 |. Comparison of polV_R391-dependent mutagenesis when rumAB is expressed from R391, or sub-cloned onto a low-copy-number vector.

Our studies examined the multifaceted nature of the regulation of the R391 mutagenic response. Regulation of the mutagenic activity of polV_R391 when expressed in its native genetic environment promotes minimal levels of mutagenesis (Pinney, 1980), even in strains in which the LexA repressor is inactivated and RecA is constitutively activated (previous unpublished observations and Figure 2). However, when rumAB was subcloned onto a low-copy-number plasmid, there was a dramatic increase in polV_R391 mutator activity, especially in strains with mutations in recA (recA718 and recA730) that lead to a constitutive RecA* phenotype (Ho et al., 1993) (Figure 2). Indeed, polV_R391 is the most potent polV ortholog characterized to date (Kulaeva et al., 1995; Mead et al., 2007). The fact that polV_R391 appears inactive when expressed from the native R391 indicates that there is likely to be cis-acting factor(s) expressed from R391 that normally act to suppress the potent mutator activity of polV_R391.

FIGURE 2.

Spontaneous mutagenesis promoted by R391, or pRW290, in different genetic backgrounds. Cells were plated on minimal low histidine agar plates as described in section 4.3; Qualitative analysis of spontaneous reversion of the hisG4(Oc) allele. His⁺ revertants appear as creamy white colonies against the dark background behind the agar plate. As observed, the 88.5kb R391 promotes low levels of spontaneous mutagenesis in all genetic backgrounds. This is in contrast to pRW290, which only expresses the rumAB_R391 operon from a low-copy-number vector. Although the rumAB operon is subject to transcriptional regulation by the LexA repressor there is little difference in mutagenesis between lexA⁺ (RW120) and lexA51(Def) (RW546) strains. Mutagenesis increases significantly when RecA is partially activated for co-protease functions (recA718; MVG114), or fully activated for co-protease functions (recA730; RW578).

2.2 |. CroS_R391 plays a major role in suppressing polV_R391-dependent mutagenesis.

The rumAB operon was originally subcloned in 1993 into the low-copy vector pGB2 (Churchward et al., 1984) as a partial EcoRI digest of R391 (Ho et al., 1993). R391 is normally chromosomally located and integrated into the 5’ end of prfC (Hochhut et al., 2001). However, the 21.5 kb insert cloned into pGB2 to generate pRLH421 is clearly of episomal origin, since it contains both 5’ and 3’ ends of the linear R391. Unfortunately, sequence analysis of the insert in pRLH421 (Genbank: U13633) reveals that the croS_R391 gene is truncated at an internal EcoRI site in the gene (Figure S2). We suspected that the croS_ΔC truncation could explain our disparate mutagenesis results with pRLH421 (high levels), versus intact R391 (low levels), if CroS_R391 does indeed repress the rumAB_R391 genes. To test this notion, we constructed a series of pRLH421-derived plasmids with combinations of wild-type croS_R391 and setR_R391, and/or deletions of croS_R391 and setR_R391 while still retaining the rumAB operon (Table 1) (Figure S2).

Table 1.

Plasmids used in this study

Plasmid	Relevant Characteristics	Source or reference
R391	Integrated into prfC	(Ho et al., 1993)
R391ΔcroS	Integrated into prfC and with a deletion of croSR391	This study
pRW290	Low copy number, SpcR, ~2 kb fragment of R391 expressing rumAB	(Szekeres et al., 1996)
pRLH421	Low-copy-number, SpcR, ∼21.5 kb fragment of R391	Genbank:U13633 (Gonzalez et al., 2019)
pJM1355	Low-copy-number, SpcR, ∼21.5 kb fragment of R391 with ΔcroSR391 setR+R391	This study
pJM1356	Low-copy-number, SpcR, ∼21.5 kb fragment of R391with croS+R391 setR+R391	This study
pJM1359	Low-copy-number, SpcR, ∼21.5 kb fragment of R391 with ΔcroSR391 ΔsetRR391	This study
pJM1360	Low-copy-number, SpcR, ∼21.5 kb fragment of R391 with croS+R391 ΔsetRR391	This study
pJM1365	Low-copy-number, SpcR, with ΔcroSR391 setR+R391	This study
pJM1366	Low-copy-number, SpcR, with croS+R391 setR+R391	This study
pJM1367	Low-copy-number, SpcR, with ΔcroSR391 ΔsetRR391	This study
pJM1368	Low-copy-number, SpcR, with croS+R391 ΔsetRR391	This study
pCC1Bac™	Single-copy-number, CmR	Genscript
pJM1378	pCC1-based, CmR, with rumABR391	This study
pJM1467	pCC1-based, CmR, with recA-promoter::rumABR391	This study
pRed/ET	Plasmid encoding the red gene cluster (redγβα) and recA TetR	(Wang et al., 2006)
pFLPe	Plasmid encoding the enhanced FLP recombinase, AmpR	Gen-H

The various plasmid iterations of croS_R391 and setR_R391 were transformed into MVG114 [ΔumuDC596::ermGT, lexA51(Def), recA718, hisG4(Oc)] in order to analyze RumAB-dependent spontaneous mutagenesis utilizing the histidine reversion assay. In the lexA51(Def) background, the RecA718 protein is in a partially activated state (RecA*) (McCall et al., 1987) and promotes significant levels of polV-dependent spontaneous mutagenesis in the absence of DNA damage (Sweasy et al., 1990). As with our earlier findings (Ho et al., 1993), pRLH421 gave very high levels of mutagenesis (Figure 3). Similarly, both the ΔcroS_R391 / setR⁺_R391 construct (pJM1355) and the double ΔcroS_R391 / ΔsetR_R391 construct (pJM1359) also gave high levels of mutagenesis revealing that the rumAB operon is not appreciably down-regulated in the recA718 lexA51(Def) background. However, in strains harboring plasmid constructs that express croS_R391 the level of mutagenesis is significantly reduced. While the levels of mutagenesis with the croS⁺_R391 / setR⁺_R391 construct (pJM1356) are reduced 2.7-fold (c.f. pJM1355) to 4.2-fold (c.f. pRLH421), the croS⁺_R391 / ΔsetR_R391 construct (pJM1360) results in a 50- to 75-fold reduction in mutagenesis (Figure 3).

FIGURE 3.

RumAB_R391-dependent mutagenesis is regulated by croS_R391.

The histidine reversion assay was performed on the E. coli strain MVG114 or MVG114 strains transformed with pRLH421 or various croS_R391 or setR_R391 wild-type or deletion combinations. (wt/ΔC) represents MVG114 transformed with the plasmid pRLH421 which contains ~ 21.5 kb of R391 and harbors a C-terminal deletion of the croS_R391 gene. MVG114 was transformed with plasmids pJM1355, pJM1356, pJM1359 and pJM1360 and the croS_R391 or setR_R391 genotypes are indicated. Error bars indicate standard error of the mean (SEM).

2.3 |. CroS_R391 repression can operate in trans and is specific for the rumAB promoter.

To demonstrate that the regulation of rumAB-dependent mutagenesis was specifically due to CroS_R391 and not some other factor encoded in the R391 DNA cloned in the pRLH421 derivatives described above, we deleted ~ 20.3 kb from ScaI to SmaI of the R391 DNA leaving only the various croS_R391 and setR_R391 operons (Figure S2). pJM1378 (Table 1), a pCC1-derivative (Epicenter/Genscript), carrying the rumAB_R391 operon, including the rumAB promoter, was transformed alone, or in combination, with either pJM1365 (ΔcroS_R391 / setR⁺_R391), pJM1366 (croS⁺_R391 / setR⁺_R391), pJM1367 (ΔcroS_R391 / ΔsetR_R391) and pJM1368 (croS⁺_R391 / ΔsetR_R391) into MVG114 [ΔumuDC596::ermGT, lexA51(Def), recA718, hisG4(Oc)]. Again, utilizing the histidine reversion mutagenesis assay, we found that rumAB-dependent mutagenesis was lower when expressing both wild-type CroS_R391 and SetR_R391 (pJM1366) corresponding to our above finding with plasmid pJM1356 (croS⁺_R391 / setR⁺_R391) (Figure 4A & Figure S3). In contrast, when setR_R391 is deleted such that only CroS_R391 is expressed (pJM1368), rumAB-dependent mutagenesis is virtually eliminated (Figure 4A & Figure S3), again in agreement with the results with pJM1360 (croS⁺_R391 / ΔsetR_R391). Furthermore, we found that in all strains where there is no CroS_R391 expressed (pJM1378 alone, or with pJM1365, or pJM1367), there is a very high level of spontaneous mutagenesis. These results confirm that regulation of rumAB-dependent mutagenesis occurs during an interplay between SetR_R391 and CroS_R391 and that this regulation can operate in trans.

FIGURE 4.

CroS_R391 trans-regulation of RumAB_R391-dependent mutagenesis.

(a). Spontaneous Histidine reversion mutagenesis assays utilizing MVG114 strains harboring pJM1378 alone (−/−), a pCC1-derivative (copy-control plasmid) carrying the rumAB_R391 operon, were transformed with low-copy pGB2-derivatives (pJM1365, pJM1366, pJM1367 and pJM1368) carrying various iterations of the croS_R391 / setR_R391 operon (Table 1). The genotypes of croS_R391 or setR_R391, either wild-type, or deleted, are indicated. (b). Western blot analysis using an anti-CroS antibody indicating that only strains harboring a plasmid with a wild-type croS_R391 gene express any CroS protein. (c). Western blot using an anti-RumA antibody indicating that strains that express the CroS protein have significantly reduced levels of the RumA protein. However, strains that express only SetR show no reduction in the level of RumA protein. Numbers reported for the levels of RumA and RumA’ are relative to RumA in track 1.

Knowing that there is a single SetR_R391 binding site upstream of the rumAB_R391 operon (Figure 1) (Gonzalez et al., 2019) and that the four SetR binding sites in the setR-croS intergenic region of the SXT element are also bound by CroS_SXT (Poulin-Laprade & Burrus, 2015), we wanted to show that the CroS_R391 repression of rumAB_R391 was dependent on the rumAB promoter region. We therefore replaced the rumAB_R391 promoter region in pJM1378 with the promoter region of the E. coli recA gene, to create plasmid pJM1467 (recA-promoter::rumAB_R391) (Table 1). As before, pJM1467 was transformed alone, or in combination, with either pJM1365 (ΔcroS_R391 / setR⁺_R391), pJM1366 (croS⁺_R391 / setR⁺_R391), pJM1367 (ΔcroS_R391 / ΔsetR_R391) and pJM1368 (croS⁺_R391 / ΔsetR_R391) into MVG114 and histidine reversion mutagenesis was performed (Figure 5A & Figure S3). Most plasmid combinations gave uniformly high levels of mutagenesis. The exception was pJM1467 together with pJM1368 (croS⁺_R391 / ΔsetR_R391) that gave ~85% of the level of mutagenesis observed with the other plasmid combinations. It should be emphasized that this is in dramatic contrast to when RumAB_R391 is expressed from the native SetR/CroS binding site-containing rumAB promoter, where CroS_R391 expressed from pJM1368 eliminated virtually all RumAB-dependent mutagenesis (c.f. Figs. 4A, 5A and S3).

FIGURE 5.

CroS_R391 regulation of RumAB_R391-dependent mutagenesis is dependent on the rumAB_R391 promoter region.

(a). Spontaneous Histidine reversion mutagenesis assays utilizing MVG114 strains harboring pJM1467 (−/−), a pCC1-derivative carrying the rumAB_R391 operon under the control of the recA promoter, were transformed with low-copy pGB2-derivatives (pJM1365, pJM1366, pJM1367 and pJM1368) carrying various iterations of the croS_R391-setR_R391 operon (Table 1). The genotypes of croS_R391 or setR_R391, either wild-type, or deleted, are indicated. The histogram illustrates the mean colony count for each indicated strain (n = 5). Error bars represent the standard error of the mean (SEM). An unpaired two-tailed t test was used to compare the mean colony counts for the ΔsetR ΔcroS and ΔsetR croS⁺ strains. * = p < 0.05. (b). Western blot using an anti-RumA antibody indicating that the level of RumA expressed from the E. coli recA promoter do not change appreciably in the presence, or absence, of SetR or CroS. Numbers reported for the expression levels of RumA and RumA’ are relative to RumA in the left-hand lane.

2.4 |. CroS_R391 regulation of RumA_R391 protein level is specific to the rumAB_R391 promoter.

In order to demonstrate that the CroS_R391 protein regulates mutagenesis by repressing the rumAB genes, we performed western blot analysis using rabbit anti- CroS_R391 (this study) and anti-RumA_R391 antibodies. Whole cell protein extracts were made from MVG114 strains harboring pJM1378 (rumAB) alone, or in combination, with either pJM1365 (ΔcroS_R391 / setR⁺_R391), pJM1366 (croS⁺_R391 / setR⁺_R391), pJM1367 (ΔcroS_R391 / ΔsetR_R391) or pJM1368 (croS⁺_R391 / ΔsetR_R391). As expected, only the strains harboring the pJM1366 (croS⁺_R391 / setR⁺_R391) and pJM1368 (croS⁺_R391 / ΔsetR_R391) constructs expressed any CroS_R391 protein (Figure 4B) and the level of CroS was identical in the presence or absence of SetR, suggesting that the phenotypic differences observed with the two plasmids (Figure 4A) is dependent on the presence of the SetR protein, and not the level of CroS protein. When compared to strains lacking CroS, the levels of RumA_R391/RumA’_R391 proteins are significantly reduced in the strains harboring the pJM1366 (croS⁺_R391 / setR⁺_R391) and pJM1368 (croS⁺_R391 / ΔsetR_R391) constructs (Figure 4C). Moreover, in the strain harboring pJM1368, which expresses only CroS_R391, the level of RumA_R391 protein is less than in the strain harboring pJM1366, despite expressing the same amount of CroS (Figure 4B) and the RumA’_R391 protein is close to the limits of detection (Figure 4C). These findings demonstrate that the CroS_R391 protein does indeed regulate the expression of the RumA_R391 protein and that its regulation can be modulated in the presence of SetR_R391. In addition, in the absence of the SetR_R391 protein (pJM1368), CroS_R391 protein may also have some inhibitory effect on RecA-mediated cleavage of RumA_R391 to RumA’_R391 (see below for more discussion).

Next, we wanted to show that the CroS_R391 regulation of RumA protein expression is specific to the rumAB promoter region. Plasmid pJM1467 (recA-prom::rumAB_R391), which expresses RumAB_R391 from the E. coli recA promoter, was co-transformed into MVG114 along with either pJM1365 (ΔcroS_R391 / setR⁺_R391), pJM1366 (croS⁺_R391 / setR⁺_R391), pJM1367 (ΔcroS_R391 / ΔsetR_R391), or pJM1368 (croS⁺_R391 / ΔsetR_R391). Western blot analysis using the anti-RumA antibodies was performed (Figure 5B). We found that the levels of RumA_R391/RumA’_R391 proteins are enhanced when expressed from the E. coli recA promoter as compared to RumA_R391/RumA’_R391 protein expressed from the native rumAB_R391 promoter. However, unlike the previous results, the levels of RumA_R391 protein expressed from the recA promoter are not affected by the presence of the CroS_R391 protein (pJM1366 and pJM1368) indicating that CroS_R391 transcriptional repression is specific to the rumAB_R391 promoter. Again, in the absence of the SetR_R391 protein (pJM1368), we observed a reduction in cleavage of RumA_R391 to RumA’_R391 (Figure 5). We believe that the reduction in RumA cleavage is due to an indirect inhibitory effect of CroS_R391 on the spontaneous generation of RecA* in the recA718 strain, since significantly more RumA cleavage was observed in the same recA718 strain exposed to the DNA damaging agent, Mitomycin C, or in the highly proficient RecA*-forming recA730 strain (−/+ Mitomycin C) (Figure S4).

2.5 |. Expression of RumA/RumA’ in a recA⁺ lexA⁺ strain expressing SetR +/− and CroS +/− after antibiotic-induced SOS induction

Based on the known interplay between CroS and SetR (Poulin-Laprade & Burrus, 2015; Poulin-Laprade et al., 2015), we hypothesized that rumAB would be the most repressed upon conditions that activate R391 transfer, namely at low SetR concentrations and high CroS levels, after damage- or stress-induced induction of the SOS response. We were therefore interested in assaying expression of RumA in a wild-type recA⁺ lexA⁺ strain (RW520) harboring the various setR-croS plasmids, that also encode the rumAB_R391 operon, (pJM1355, pJM1356, pJM1359 and pJM1360) in which the SOS response was stress-induced by treatment with the antibiotic, Ciprofloxacin, for up to 3 hours (Figure 6). Ciprofloxacin-mediated induction of the SOS response was followed by western blot analysis of the LexA-regulated RecA and RumA proteins. With all plasmid combinations, a strong induction of RecA was observed after treatment with Ciprofloxacin, indicating that the chromosomally-encoded LexA repressor had been inactivated in vivo (Figure 6). The lack of any effect of the various setR-croS plasmids on RecA expression also indicates that recA is not negatively regulated by either SetR or CroS. In contrast, the timing of the induction of RumA_R391 was dependent upon the presence, or absence, of SetR, or CroS (Figure 6). For example, in strains harboring pJM1359 (ΔcroS_R391 / ΔsetR_R391), where rumA is only regulated by LexA, there was a time dependent induction of RumA, which peaked around 1 hr after Ciprofloxacin treatment, followed by conversion of RumA to RumA’ 2–3 hrs post treatment. By comparison, the peak of RumA induction in the presence of pJM1355 (ΔcroS_R391 / setR⁺_R391), was around 2 hrs, suggesting that the presence of SetR delays RumA expression by an hour. However, there was no effect on the conversion of RumA to RumA’ which nevertheless occurred 2–3 hrs post treatment. (Fig, 6). In contrast, no Ciprofloxacin-induced expression of RumA was observed in the presence of pJM1360 (croS⁺_R391 / ΔsetR_R391), consistent with our earlier observations (Figure 4C). Interestingly, in the presence of pJM1356 (expressing both croS⁺_R391 / setR⁺_R391), expression of RumA peaked at 1hr post Ciprofloxacin treatment, from that point on, RumA levels decrease and by 3 hrs, they are barely detectable. Such observations can readily be explained by the interplay and hierarchy of the three transcriptional repressors. We know that LexA is cleaved just a few minutes after DNA damage, so as to induce the 40+ protein SOS regulon (Fernández de Henestrosa et al., 2000). In contrast, SetR, which we have previously shown to be cleaved much slower that LexA in vitro (Gonzalez et al., 2019), would be expected to be cleaved and inactivated by RecA* in vivo around the 1–2 hr point, which would then allow CroS_R391 sole access to the SetR/CroS binding site in the rumAB promoter, thereby effectively eliminating residual expression of the RumA protein. The fact that RumA is detected in the croS⁺_R391 / setR⁺_R391 strain, but not the croS⁺_R391 / ΔsetR_R391 strain, implies that SetR_R391 normally competes with CroS_R391 in vivo and to some degree blocks its access to the SetR-CroS binding site in the rumAB promoter. Such a scenario explains why polV_R391, encoded by the stably integrated R391 promotes such low levels of mutagenesis after DNA damage, or in constitutively activated RecA* strains, where both LexA and SetR repressors are inactivated and rumAB is repressed solely by CroS (Figure 2).

FIGURE 6.

Expression of RecA, RumA and RumA’ in wild-type recA⁺ lexA⁺ cells after exposure to the SOS-inducing antibiotic, Ciprofloxacin.

Western blot analysis was performed on whole-cell protein extracts from RW520 (recA⁺ lexA⁺) harboring pRLH421 derivatives (pJM1355, pJM1356, pJM1359 or pJM1360), with various croS_R391 or setR_R391 wild-type, or deletion combinations as indicated. To induce the SOS response, cells were treated with 30 ng ml⁻¹ Ciprofloxacin for various times, as indicated in the figure. Levels of RecA and RumA/ RumA’ were detected using affinity purified polyclonal rabbit antibodies to RecA and RumA proteins. The number reported for the level of RumA or RumA’ are relative to a cross-reacting band in the same track.

2.6 |. Deletion of croS_R391 on R391 leads to enhanced mutagenesis

Since R391 stably integrates into the E. coli chromosome at the 5′ end of the prfC gene (Hochhut et al., 2001), we were able to construct a croS_R391 deletion mutant in the MVG114 genetic background and compare mutagenesis between the R391 and the R391ΔcroS strains. First, we examined mutagenesis using a galK2(Oc) reversion papillation assay in which orange-red Gal⁺ mutant “papillae” grow up within bacterial colonies plated on MacConkey-galactose agar media. Representative colonies from the R391 and R391ΔcroS strains are shown in Figure 7. The R391 strains gave 0 – 3 Gal⁺ mutant papillae per colony, whereas the R391ΔcroS gave 30 – 50 Gal⁺ mutant papillae per colony. Second, we utilized a rifampicin mutagenesis assay in which cells are plated on LB agar plates containing 100 μg ml⁻¹ rifampicin. R391 ΔcroS strains exhibited a 10-fold increase in rifampicin resistant mutagenesis, as compared to the strains harboring the wild-type R391 (Figure 8). Both of these results demonstrate that the CroS protein, expressed from its native locus within an intact R391, down-regulates rumAB and tightly controls mutagenesis in strains carrying the R391 element.

FIGURE 7.

galK2(Oc) reversion papillation assay of R391, or R391 ΔcroS strains.

R391, or R391 ΔcroS were moved into MVG114 by conjugal transfer. Cells were plated onto MacConkey-galactose agar media and grown for 8 days at 37°C. Pictures of representative colonies from each strain were taken showing the appearance of Gal+ papillae indicating the level of mutagenesis occurring within the colonies. The MVG114/R391 ΔcroS colonies contain approximately 10 −50 times the number of revertant papillae, when compared to the MVG114/R391 colonies.

FIGURE 8.

Rifampicin mutagenesis assay of R391 and R391ΔcroS strains.

MVG114, MVG114/R391 and MVG114/R391ΔcroS cultures were started from with approximately 1000 cells, or less, and grown over night to stationary phase. Cells were plated onto LB agar plates containing rifampicin to select for rifampicin resistant mutants. Appropriate dilutions were plated to LB agar plates to determine viable counts and the frequency of mutagenesis to rifampicin resistance was calculated. Error bars represent the standard error of the mean (SEM). The MVG114/R391ΔcroS strain exhibits an ~10-fold higher frequency of mutagenesis to rifampicin resistance than the MVG114/R391 strain.

3 |. Discussion

3.1 |. Unprecedented transcriptional regulation of R391 encoded rumAB

Previous studies have shown that chromosomally encoded E. coli polV is tightly regulated via a combination of transcriptional, posttranslational and spatial regulation (Goodman et al., 2016). It is therefore likely that polV orthologs are also subject to strict regulation. Indeed, since the mid ‘90s, when the R391 rumAB operon was sequenced, it has been known that rumAB-encoded polV_R391 is regulated by the LexA transcriptional repressor (Kulaeva et al., 1995). Recently, we reported that the rumAB operon is also regulated by the λcI-like transcriptional repressor, SetR_R391 (Gonzalez et al., 2019). Here, we provide in vivo data that is compelling and consistent with CroS_R391 as acting as a third and potentially the most potent transcriptional regulator of the rumAB operon. For example, when expressed in cis- or trans- with rumAB, CroS_R391 completely inhibits polV_R391-dependent mutagenesis (Figs. 3 and 4). The lack of polV_R391 mutagenesis in vivo is attributed to extremely low-level expression of RumA (Figure 4) and RumB (unpublished observations) from its native promoter in the presence of CroS_R391. The most reasonable explanation for such a phenotype, is the unhindered access that CroS_R391 has to the single SetR-CroS binding site in the promoter region of the rumAB operon (Figure 1), where it acts as a strong transcriptional repressor of the rumAB operon. Interestingly, expression of RumA actually increases when CroS_R391 and SetR_R391 are co-expressed (Figs. 4 & 6) implying that SetR_R391 may at least partially block CroS_R391 access and binding to the SetR-CroS binding site, thereby preventing it from acting as a potent transcriptional repressor. Indeed, expression of RumA in a wild-type strain after antibiotic-induced SOS induction appears to result from an interplay of all three transcriptional repressors, LexA, SetR_R391 and CroS_R391, such that extremely low levels of polV_R391 are only present, even after full induction of the SOS response (Figure 6).

polV_R391 is a potent mutator DNA polymerase when uncoupled from its normal regulatory pathways (Figure 2). Its unprecedented regulation by three separate transcriptional repressors, two of which (LexA and SetR) are cleaved and inactivated after DNA damage, therefore only allows for the very limited expression of the highly error-prone DNA polymerase (Figure 6), before the third (non-cleavable) repressor (CroS), curtails RumAB expression, and provides a mechanism whereby the cell returns to a resting state, with low levels of cellular mutagenesis.

3.2 |. Inactivation of croS_R391 increases the potential of enhanced mutagenesis

Deletion of croS_R391 on the intact R391 stably integrated into the E. coli genome allowed us to evaluate the effect of the CroS_R391 protein on repression of the rumAB_R391 operon and mutagenesis from R391 in its native locus. Utilizing both a qualitative galK2(Oc) reversion papillation assay (Figure 7), and a quantitative assay rifampicin resistance assay (Figure 8), we found that mutagenesis is significantly increased in MVG114 strains harboring R391 ΔcroS as compared with strains harboring wild-type R391. These results confirm our assertion that the CroS_R391 protein is the major R391-encoded negative regulator of the rumAB_R391 operon and indicate that there are likely no additional rumAB_R391 regulators encoded on the full-length R391.

R391 is widely distributed in enterobacteriaceae that are opportunistic pathogens which cause a variety of infections in humans (Song et al., 2013; Bie et al., 2017; Fang et al., 2018; Kong et al., 2020; Slattery et al., 2020). Given that CroS_R391 appears to be the “master regulator” to switch off rumAB expression, naturally occurring inactivating mutations in croS therefore have the potential of increasing the development and proliferation of polV_R391-dependent antibiotic resistance in a wide range of pathogenic microorganisms.

4 |. EXPERIMENTAL PROCEDURES

4.1 |. Bacterial strains and plasmids

Plasmids are listed in Table 1. Bacterial strains are listed in Table 2.

Table 2.

E. coli strains used in this study

Strain	Relevant Genotype	Source or reference
RW120a	ΔumuDC595::cat hisG4(Oc) galK2(Oc)	(Ho et al., 1993)
RW546a	ΔumuDC595::cat lexA51(Def) hisG4(Oc) galK2(Oc)	(Fernández de Henestrosa et al., 2000)
RW578a	ΔumuDC595::cat lexA51(Def) recA730 hisG4(Oc) galK2(Oc)	(Mead et al., 2007)
RW520a	ΔumuDC596::ermGT hisG4(Oc) galK2(Oc)	LGI stocks
RW584a	ΔumuDC596::ermGT lexA51(Def) recA730 hisG4(Oc) galK2(Oc)	(McDonald et al., 2012)
MVG114a	ΔumuDC596::ermGT lexA51(Def) recA718 hisG4(Oc) galK2(Oc)	(Gonzalez et al., 2019)
MVG114a	R391	This study
RW1766b	ΔdinB61::ble	This study
MVG114a	R391 ΔcroS	This study

^a:Full genotype: thr-1 araD139 Δ(gpt-proA)62 lacY1 tsx-33 glnV44 rpsL31 xyl-5 mtl-1 argE3 thi-1 sulA211

^b:ΔdinB61::ble derivative of MG1655 (F⁻ λ- rph-1)

Previously, R391 DNA (formerly known as the IncJ plasmid, R391) was isolated in its unintegrated episomal form in an E. coli strain harboring the recA718 allele (RW96) (Ho et al., 1993). This episomal R391 DNA was partially digested with EcoRI and an ~ 21.5 kb fragment was cloned into the low copy number vector pGB2 (Churchward et al., 1984) to generate pRLH421 (Genbank U13633) (Gonzalez et al., 2019). Unfortunately, the pRLH421 plasmid contains a truncation of the croS_R391 gene in the divergent operon with setR_R391 at an internal EcoRI within the croS_R391 gene. To reconstruct the croS_R391 gene a fragment, designated “croS complete NarI-PmeI-Bsu36I”, was synthesized (Genscript) that includes from the NarI site to the EcoRI site of pGB2, the complete croS_R391 open reading frame, the upstream promoter sequences and a Bsu36I site replacing the start of the setR_R391 gene. The croS complete fragment was subcloned into pRLH421 from the NarI (in pGB2) to PmeI (in the promoter region) to create an intact croS⁺_R391 setR⁺_R391 operon (pJM1356). Further, the croS_R391 complete fragment was subcloned into pRLH421 from the NarI to Bsu36I (in setR_R391) to create a croS⁺_R391 / ΔsetR_R391 operon within only 222 bp of the 3’ end of the setR_R391 gene (pJM1360). Another fragment designated “croS deletion NarI-PmeI-Bsu36I”, was synthesized (Genscript) from the NarI site to the first 5 bases of the EcoRI site of pGB2, the promoter sequences upstream of the croS_R391 start ATG and a Bsu36I site replacing the start of the setR_R391 gene. The croS deletion fragment was subcloned into pRLH421 from the NarI to PmeI to create a ΔcroS_R391 / setR⁺_R391 operon (pJM1355). Further, the croS_R391 deletion fragment was subcloned into pRLH421 from the NarI to Bsu36I to create a ΔcroS_R391 / ΔsetR_R391 operon (pJM1359).

To create plasmids that carry just these four iterations of the croS_R391 / setR_R391 operon, the plasmids pJM1355, pJM1356, pJM1359 and pJM1360 were digested with ScaI and SmaI and re-ligated, which deletes ~ 20.3 kb of R391, leaving only the croS_R391 / setR_R391 region including about 200 bases downstream of the end of the setR_R391 gene. These plasmids are designated pJM1365 (ΔcroS_R391 / setR⁺_R391), pJM1366 (croS⁺_R391 / setR⁺_R391), pJM1367 (ΔcroS_R391 / ΔsetR_R391) and pJM1368 (croS⁺_R391 / ΔsetR_R391).

The rumAB operon including the rumAB promoter region were cloned into the copy control plasmid, pCC1 (Genscript), to generate pJM1378. In addition, the rumAB operon fused to the E. coli recA promoter region was cloned into pCC1 to generate pJM1467 (recA-prom:rumAB) (Genscript). These low-copy pCC1Bac plasmids are compatible with the pGB2-based croS_R391 / setR_R391 plasmids described above, such that the various iterations of the croS_R391 / setR_R391 operon can be co-expressed with the rumAB operon.

4.2 |. Generation of R391 ΔcroS.

RW1766/R391 was constructed by inoculating an individual colony of RW1766 and RW120/R391 into 5 mL LB and incubating overnight at 37°C without shaking. The next morning, the culture was streaked on to plates containing Zeocin (25 μg ml⁻¹) and Kanamycin (50 μg ml⁻¹) to select for isolates of RW1766 that had acquired R391 via bacterial conjugation.

The marker-less removal of the croS gene from the R391 element of E. coli RW1766/R391 was performed according to Zhang et al. (Zhang et al., 1998) (see Figure S5). In a first recombination step, a linear knock out cassette, generated with primers listed in Table 3, was used to replace the croS gene via Red/ET recombination by using plasmid pRed/ET. This resulted in a chloramphenicol-resistant intermediate strain with the genotype ΔcroS::FRT-Cm^R-FRT (Figure S5). In a second recombination step the selection marker was removed in a FLP recombinase-mediated fashion, leaving a single FRT site at the former croS locus (genotype: ΔcroS::FRT). The intermediate and final clones were analyzed by Sanger-sequencing using primers listed in Table 3 covering the complete modified and adjacent regions.

Table 3.

Oligonucleotides used in this study

Name	Sequence	Source
P1a	CTCGCTTTTGTGTTTACGTAAAATGCTGACCATTATTTGCCCTTAATCCTAAATTAACCCTCACTAAAGGGCGG	BioSpring
P2a	CAGTTTAAACTGTAACATTCAATCTTGTAACAGTTATTATTGTAACAGGAGACGGTTAGCACGGAGTTCATTAGGGCTC	Biospring
cp01b	GCGATGCCATACCAATGAAGTCG	BioSpring
cp02b	CAACGTAACCAATCGGCAGTCG	BioSpring
cp03b	CCAGTGACTTCACGCCACTCC	BioSpring
cp04b	CAGGATCAACAACGATCACTGCC	BioSpring

^a:Used in the generation of linear knock out cassette.

^b:Used in the amplification of the modified region and subsequent DNA sequencing

New strains containing wild-type R391, or R391 ΔcroS, were made by conjugal transfer of the ICE from RW1766 (Table 2) to the desired recipient strain by selecting on plates containing Streptomycin (100 μg ml⁻¹) and Kanamycin (50 μg ml⁻¹).

4.3 |. Qualitative analysis of spontaneous reversion of the hisG4(Oc) allele

The E. coli strain MVG114 was transformed with various plasmid constructs expressing CroS_R391, SetR_R391 and RumAB, either alone, or in various iterartions (Table1). To assay for reversion of the hisG4(Oc) allele, three to five isolates of each strain were grown overnight at 37°C in LB medium containing the appropriate antibiotic(s). Five hundred microliters of the cultures were centrifuged, and the pellet resuspended in an equal volume of SM buffer. One hundred microliters of the resuspended pellets were spread on each of five low-histidine minimal plates (Davis and Mingioli minimal agar plates (Davis & Mingioli, 1950) plus glucose (0.4% wt/vol); agar (1.0% wt/vol); proline, threonine, valine, leucine, and isoleucine [all at 100 μg ml⁻¹); thiamine (0.25 μg ml⁻¹); and histidine (1 μg ml⁻¹)]. After incubating the plates for four days at 37°C, the His⁺ mutant colonies were counted and averaged between the independent cultures and standard error of the mean (SEM) calculated.

4.4 |. galK2(Oc) reversion papillation assay

The galK2(oc) papillation is a visual reversion mutagenesis assay that was previously used to identify genes from R-plasmids that encode orthologs of the E. coli umuDC genes including rumAB from R391 (Ho et al., 1993). Briefly, ~ 50– 75 E. coli cells harboring the galK2(Oc) allele are plated onto MacConkey-galactose agar media and grown for 8 days at 37°C. The number of small orange-red Gal⁺ papillae that grow up within a bacterial colony are noted from multiple colonies. Comparison of the number of papillae allows the assessment of the level of mutagenesis in the MVG114 strains harboring R391 versus MVG114 harboring R391 ΔcroS.

4.5 |. Rifampicin mutagenesis assay

Selection for rifampicin resistance is another generally used mutagenesis assay. Base-pair substitution mutations arising within the central 202 bp of the rpoB gene can give rise to resistance to the rifampicin antibiotic. Five milliliter cultures of MVG114/R391 and MVG114/R391 ΔcroS were started, in triplicate, from an initial inoculum containing ~ 1000 viable cells and grown for 24 h at 37°C. One hundred microliters of these cultures were spread on five LB agar plates containing 100 μg ml⁻¹ rifampicin. In addition, these cultures were serially diluted and appropriate volumes were plated to LB agar plates to determine viable cell counts. Subsequently, frequencies of rifampicin mutations arising within the cultures was calculated.

4.6 |. Western blot analysis of CroS_R391, RumA_R391 and RecA proteins

E. coli cultures were grown in Luria-Bertani media at 37°C until exponential phase (OD₆₀₀ ~0.5). For the experiments shown in Figure 6, Ciprofloxacin (30 ng ml⁻¹) was added to the culture and cells harvested by centrifugation at subsequent time points (as indicated in Figure 6). For all other experiments, undamaged cells were harvested at an OD₆₀₀ ~0.5. The cell pellet was resuspended in NuPage LDS sample buffer (Novex) and freeze-thawed to produce whole cell extracts. Cell extracts were electrophoresed on NuPage 4–12% Bis-Tris gels (Novex). Proteins were transferred to an Invitrolon PVDF membrane (Novex) which was probed with appropriate dilutions of affinity purified rabbit anti-CroS_R391, anti-RumA_R391 or anti-E. coli RecA antibodies and subsequently probed with an appropriate dilution of Goat Anti-Rabbit IgG (H+L)-AP Conjugate (Bio-Rad). Using the CDP-Star chemiluminescent assay (Tropix), the CroS_R391, RumA_R391/A’_R391 or RecA proteins were visualized on Carestream Biomax XAR film after various exposure times. Digital images were also captured using an Alpha Innotech FluorChem HD2. These images were then imported as .tif files into LI-COR Biosciences Image Studio Lite software, where band density was quantified using the Data Analysis tool. Relative protein levels for bands of interest were calculated by normalizing band density to that of non-specific bands within each lane, then expressed relative to a cross-reacting reference band.

4.7 |. Overexpression and purification of CroS.

The gene encoding R391 croS was codon optimized for expression in E. coli and chemically synthesized (Genscript) as a 309 bp NdeI-PstI fragment and cloned into pUC57 (Genscript). The NdeI-PstI fragment was subsequently sub-cloned into the same sites of pCF2. pCF2 is a medium copy plasmid that expresses IPTG-inducible glutathione-S transferase (GST). The construct also contains a PreScission (GE Healthcare) protease site immediately downstream of the GST protein and upstream of the unique NdeI site. When the target gene is cloned into the NdeI site of pCF2, a GST-fusion protein is generated with a PreScission site immediately upstream of the target protein. CroS was initially purified as a GST-CroS fusion protein as a custom service by scientists at Eurofins, as previously described (Poulin-Laprade & Burrus, 2015) and the GST-affinity tag subsequently removed after PreScission protease treatment (Eurofins).

4.8 |. CroS Antibodies

Polyclonal antibodies to the purified CroS protein were raised in rabbits as a custom service (Covance) and affinity purified. These antibodies are highly specific with very few cross-reacting bands in western blots of E. coli extracts lacking CroS.

Abbreviations:

TLS

Translesion DNA synthesis

ICE

Integrating Conjugative Element

polV

DNA polymerase V

RecA*

Activated RecA-nucleoprotein filament

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding authors upon reasonable request.