The Escherichia coli serS gene promoter region overlaps with the rarA gene

Deletion of the entire gene encoding the RarA protein of Escherichia coli results in a growth defect and additional deficiencies that were initially ascribed to a lack of RarA function. Further work revealed that most of the effects reflected the presence of sequences in the rarA gene that affect expression of the downstream gene, serS. The serS gene encodes the seryl aminoacyl-tRNA synthetase. Decreases in the expression of serS can trigger the stringent response. The sequences that affect serS expression are located in the last 15 nucleotides of the rarA gene.

The Escherichia coli RarA protein is an ATPase in the AAA+ superfamily [42,43]. The rarA gene encodes a 447-amino-acid polypeptide with a predicted monomeric mass of 49594 kDa. The protein is part of a highly conserved family. It is absent in archaea but highly conserved from bacteria through eukaryotes, sharing about 40% identity and 56-58% similarity with its Saccharomyces cerevisiae (Mgs1) and Homo sapiens (WRNIP1) homologs [42,43]. In E. coli, RarA shares 25% amino acid identity with two other proteins: RuvB, a Holliday Junction helicase, and DnaX, a subunit of DNA polymerase III replisome. DnaX encodes for Tau (τ) and Gamma (γ) components of DNA polymerase III clamp loader complex, placing RarA in the clamp loader AAA+ clade [42,43].
In E. coli chromosome, the rarA gene is located at 20.21 centisomes (location 937,994>939,337). The rarA gene is immediately upstream of an essential gene, serS, a serine-tRNA ligase, located at 20.24 centisomes. SerS is among the 20 aminoacyl-tRNA synthetases (aaRSs) or tRNA-ligases present in the cell. aaRSs are the charging portals of tRNAs. They generate a covalent linkage between an amino acid and its cognate tRNA to form an aminoacyl-tRNA complex. The Ribosome acts on this charged tRNA complex and transfers its attached amino acid onto the growing peptide chain-thereby fostering the translation process in the cell. SerS aminoacylates tRNA Ser and tRNA Sec with serine [44,45]. serS is mainly regulated by its promoter (serSp1) with a transcription start located at 939,365 th position after the end of rarA gene [46] (Fig 1).
aaRSs manage the growth and the stringent response in the cell by directly controlling the two interdependent cellular processes: (1) the flux of protein synthesis, and (2) the levels of uncharged tRNAs. The first process-the flux of protein synthesis-is directly dependent on the amount of tRNA aminoacylated by aaRSs. Modifications in aaRSs production impedes cell growth [47,48]. Globally, high levels of uncharged tRNA slow translation kinetics and thereby slow cell growth-both in bacteria and eukaryotes [49][50][51]. In bacteria, these high levels of uncharged tRNA are detected by the (p)ppGpp synthetase-RelA-which in response induces a stringent response and affects the cell growth [52][53][54].

PLOS ONE
SerS is notable as it is inhibited by serine hydroxamate, a small molecule often used by investigators to induce the stringent response [55]. We have found that the complete deletion of the rarA gene slows cell growth, impedes SOS induction, and rescues DNA damage sensitivity of several repair-deficient cells, effects we initially attributed to a lack of RarA function. This initial conclusion was in error. All of these phenotypes disappear when a slightly more modest rarA deletion is used that deletes more than 90% of the coding sequence, all but the last 41 codons. This suggests that regulatory sequences that affect serS expression may be embedded in the rarA coding sequence. Keeping a small portion of the rarA gene, that which encodes C-terminal of RarA, is vital for optimal growth of the E. coli cell. A -35 segment of the serS promoter or some equivalent regulatory sequence appears to be located in the last 15 nucleotides of the rarA gene.

Strain construction
All strains are E. coli MG1655 derivatives and are listed in Table 1. Some of the rarA strains (rarAN406, rarAN430, rarAN437 and rarAN442) were made using galK+ recombineering method. ΔrarA (EAW98) and all other strains were constructed using Lambda red recombination as described by Datsenko and Wanner [56]. Kanamycin resistance of these strains was removed using FLP recombinase when required [57]. All chromosomal mutations were confirmed using Sanger sequencing. Standard transformation protocols were used to generate strains harboring the indicated plasmids as listed in Table 1.

Plasmid construction
All plasmids were sequenced to confirm the correct mutation(s)/insertion(s) following their construction. pBAD-serS is a pBAD/myc-His A Nde + wt serS. pEAW1176 was constructed by amplifying the wildtype serS gene containing NdeI and BamHI restriction cut sites from the E. coli MG1655 genome in a PCR. pBAD/myc-His A Nde was cut with NdeI and BglII enzymes (BglII creates compatible sticky ends with BamHI), while the PCR product was cut with NdeI and BamHI enzymes. The PCR product was ligated into the pBAD/myc-His A. pEAW1012 is a derivative of pRC7 plasmid (a lac+ mini-F low copy derivative of pFZY1) that expresses a WT copy of the rarA gene. All plasmid sequences are provided in the S1 File.

Growth curve
A single colony of each indicated strains was inoculated into 3 mL of LB media (10 g L −1 tryptone, 5 g L −1 yeast extract, 10 g L −1 sodium chloride, and 1.1mL 1N NaOH), which was incubated overnight for 16 h at 37˚C with the orbital shaking at 200 rpm. Overnight cultures were diluted 1:100 in LB medium. In total, 30 μL of the overnight culture was used to inoculate 3 mL of fresh LB. 100 μl of each culture was poured in a clear bottom 96 well plate (Corning).
Cultures were grown at 37˚C with continuous orbital shaking at 205 cpm in a BioTek Synergy 2 plate reader. OD 600 values were taken every 10 minutes for over the course of 800 minutes.
OD 600 values were normalized by subtracting out the OD 600 value of only LB media. All growth curves represent averaged values from the three biological replicate experiments.

Growth competition assays
Growth competition assays were conducted as previously described [58] using a method originally described by Lenski [59]. The ΔaraBAD ΔParaB marker was included on either wild type (MG1655) or mutant (ΔrarA) in separate experiments to control for any effect the marker may have had on cell fitness.

SOS induction assay
Plasmid expressing the Green Fluorescent Protein (GFP) under the regulation of recN promoter (pEAW903) was used in this assay. First, either empty vector (pQBi63) or pEAW903 was transformed into the appropriate strains (WT, ΔrarA, or rarAΔN406) and the transformants were selected on Amp100 (100 μg/ml) plates. The transformants were then grown in 3 ml of LB + Amp100 medium overnight at 37˚C. The next day, the cultures were diluted in 1:1000 ratio in LB + Amp100 broth and 150 μl of sample were poured into each well of a 96-well plate (Corning Incorporated/Costar) and put into the plate reader (Synergy H1 Hybrid Reader by BioTek). The samples were allowed to grow for 1000 mins at 37˚C, with OD 600 and GFP fluorescence (488/515 nm) recorded at every 10 minutes. Relative fluorescence was calculated by normalizing the fluorescence reading to the OD 600 of the culture.

Analysis of cell shape: Bright field microscopy
All cells were grown overnight at 37˚C and the saturated culture diluted in 1:100 ratio and grown in LB media till O.D. reaches 1.0. 200 μL of culture were then pelleted down, resuspended in 1XPBS buffer and incubated with 2 μl of FM-64 dye (0.33 M) on ice for at least 30 mins. For imaging, 2 μl of this mixture were loaded onto 0.16 mm thick borosilicate glass made coverslips (Azer scientific) and sandwiched with 1% agarose gel pad. For all measurements of cell size and filamentation, wide-field microscopy was conducted on a STORM/TIRF inverted microscope ECLIPSE Ti-E (Nikon) (100× objective). Images using DIA and dsRed filters were collected on an ORCA Flash 4.0 Hamamatsu camera. A bright-field and dsRed image (at 100 ms and 50 ms exposure respectively) were taken at multiple fields of view to determine the cell shape and length. For analysis, all images were imported into MicrobeJ, an ImageJ plugin, to outline cells. Selected cells were manually filtered for any outliers. All strains were imaged in triplicates and the cell size of each strain is averaged compiling each repeat.

Results
It is well documented that RarA-family proteins are involved in the maintenance of genome stability in cells from bacteria to human, but the precise function of these proteins and mechanism of action remains an enigma.  (Fig 2A), and then studied the effect of this deletion on cell fitness. In the following discussion, ΔrarA refers to the complete deletion of the gene encoding RarA protein.

Complete deletion of rarA causes a growth defect and reduced cell size of MG1655 E. coli cells
Using a plate reader, we noted and compared the growth of the ΔrarA strain to wild type cells at 37˚C every 10 mins for 18hrs. The ΔrarA cells grew more slowly than wild type cells. (Fig  2B). To document the growth defect of the ΔrarA mutant with a different and more sensitive method, we carried out a direct competition assay between the wild type strain and the ΔrarA strain, using an approach developed by Lenski and colleagues [59]. (Fig 2C). Wild type or mutant cells were modified to carry a neutral Ara-mutation (which confers a red color on colonies when grown on tetrazolium arabinose (TA) indicator plates) to permit color-based scoring of mixed populations. Overnight cultures of the ΔrarA strain were mixed in a 50/50 ratio with isogenic wild type cells carrying the Ara-mutation, or vice versa. The mixed culture was then diluted and grown up again on successive days, with plating to count red and white colonies occurring once each day. Earlier work [59,63] demonstrated that the Ara − mutation does not affect growth rates by itself. We found that the wild type cells outgrew the ΔrarA cells and dominated the mixed cultures almost completely within 24 hours (Fig 2C). We further investigated the phenotypic dissimilarities between ΔrarA and WT cells using bright field microscopy. We observed that ΔrarA cells were substantially smaller than rarA+ cells (1.58 μm [SEM = 0.01] versus 2 μm [SEM = 0.01] in length) (Fig 2D). RarA is well documented as a vital player in the DNA recombination and repair process. Suppressing the DNA repair system exerts stress response in the cell. With the data collected above, we presumed that the complete deletion of rarA impedes the damage tolerance capability of the cells that results in a significant growth impact.

PLOS ONE
The Escherichia coli serS gene promoter region overlaps with the rarA gene

Phenotypic defects of ΔrarA cells are attributed to lower expression of the serS gene
For further affirmation of the results obtained above, we performed a complementation test. The pRC7 plasmid carrying a wild type copy of rarA along with ampicillin marker is employed in this study. We incorporated this plasmid into a ΔrarA strain and tested its growth rate. Surprisingly, we observed no rescue of the growth defect even in the presence of wild-type copy of rarA on the plasmid (Fig 2E). This observation signals that the growth defect observed earlier is not directly associated with the absence of rarA but might be an outcome of that deletion on other growth-related genes. Based on the genomic location position of rarA, we hypothesized that the growth defect might be ascribed to the defect in the closest downstream essential gene-serS. It is well documented that the addition of SHX (serine hydroxamate), an inhibitor of the serRS gene, causes growth defects in E. coli cells even under the nutrient rich conditions [55]. Changes in the levels of serS are expected to alter the levels of charged to uncharged tRNA ratios and thereby the cell growth. Decreased cell viability of a ΔrarA strain could be a result of the decreased serS levels in the cell. To test this hypothesis, we incorporated a plasmid overexpressing serS in ΔrarA cells. Overproduction of serS rescued the growth defect of ΔrarA cells (Fig 2F). This result signals the presence of promoter element/regulatory sequence for the serS gene within the coding region of a rarA gene. Removal of that segment impacts the level of SerS in the cell.
Using the multi-genome browser of Ecocyc, we next searched for the orthologs of rarA in a broad range of organisms and then mapped the extent to which those orthologs have maintained their genetic context relative to E. coli. It revealed that the positioning of the serS generight downstream of the rarA gene locus-exists only in γ-proteobacteria class of the Proteobacteria phylum (Fig 3). Conservation of this proximity between rarA and serS genes across this class indicates their possible interconnection in other organisms of this class as well.

Identification of promoter/regulatory sequence for serS gene in rarA sequence
We next aimed to identify the segment within the rarA gene that is controlling the serS expression under normal conditions. We constructed various rarA mutations differing in the number of nucleotides deleted from the N-terminus of the rarA to figure out the minimum region of rarA required to remain intact to mitigate the growth defect of ΔrarA cells. The GalK + recombineering method was used instead of the Datensko Warner method to avoid any effect of the kanamycin cassette sequence on the serS expression. We created four variants-rarAΔN406, rarAΔN430, rarAΔN437, and rarAΔN442, leaving 41, 17, 10, and 5 C-terminal codons intact respectively, and studied their growth and cell morphology profiles (Fig 4A). Interestingly, none of these mutants showed any growth adversity like ΔrarA (Fig 4B and 4C). However, the creation of complete deletion of rarA via galK recombineering method failed.

PLOS ONE
The Escherichia coli serS gene promoter region overlaps with the rarA gene This indicates that there exists a possible promoter or regulatory sequence for serS within the last 5 codons of rarA-deletion of which hampers the serS expression and thereby the growth of the cell. The serS gene is mainly regulated by its promoter (serSp1) with a transcription start located at 28 nucleotides downstream from the end of rarA gene [46,64]. Tracing back its possible promoter region, we suspected that the -10 region for this serSp1 is located at~14 nucleotide from the rarA gene end. This overlaps with the previously identified σ 70 promoter of serS [64]. No information about the -35 region of this promoter has been presented previously. The most likely -35 hexamer sequence for serSp1 was analyzed using the Salis lab promoter calculator [65]. We suspected that the -35 is located (although there is no good consensus -35 there) within the last~15 nucleotides of the rarA gene (Fig 1). A -35 segment of the serS promoter or some equivalent regulatory sequence within this last segment of rarA gene makes the complete deletion of rarA an infeasible option in the E. coli cell.

Consequences of ΔrarA on the damage sensitivity of cells compromised with defects in other DNA repair systems
We tested the drug sensitivity of ΔrarA cell alone and it in combination with other repair systems. The absence of a complete rarA sequence itself does not increase the cell sensitivity to DNA damaging agents like UV or NFZ. Removal of a RecA-loading system like RecF or RecO, however, increases cells' sensitivity to UV irradiation, as observed previously [66][67][68][69][70]. Interestingly, deletion of rarA in a ΔrecF or ΔrecO background rescues their damage sensitivity. Complete deletion of rarA in Δpol IV background also decreases the sensitivity of pol IV-cells to both UV and NFZ induced damages (S1 Fig). Moreover, we observed that the SOS response is also induced in ΔrarA cells, both with and without external damaging conditions (UV treatment). The induction was much higher than a WT cell (S2 Fig). Interestingly, none of these results were replicated when the ΔrarAN406 background was used instead of ΔrarA.
With all these observations, we confirmed that the phenotype observed upon the complete deletion of rarA is attributed to the decreased levels of serS in the cell. A decreased level of serS could cause a stringent response which activates the level of ppGpp. High levels of ppGpp act by rescuing the stalled RNA polymerases. The rescue of Δpol IV/ ΔrecF/ ΔrecO cells' drug sensitivity upon deletion of ΔrarA may be due to the rescue of stalled RNA polymerases, an outcome of the action of high levels of ppGpp in ΔrarA cells. High SOS levels could also be a repercussion of this same phenomenon. Deletion of all but the last 41 codons of rarA eliminates all these phenotypes.

Discussion
The major conclusion of this work is straightforward. Genetic elements affecting the expression of the serS gene are embedded in the final five codons of the upstream rarA gene. Upon complete deletion of rarA, we had documented a variety of phenotypic effects (supplementary data) that we initially attributed to a loss of rarA function. These disappeared when we made use of rarA deletions that encompasses most but not all of the gene. We now attribute the effects to changes in serS expression, possibly reflecting some aspect of a stringent response.
The serS promoter element that is within the rarA gene has not been identified precisely. The region in question is positioned so as to potentially include a -35 region for the promoter. However, no -35 consensus is evident. Cox.