The Small Regulatory RNA Spot42 Inhibits Indole Biosynthesis to Negatively Regulate the Locus of Enterocyte Effacement of Enteropathogenic Escherichia coli

The locus of enterocyte effacement is necessary for enteropathogenic Escherichia coli (EPEC) to form attaching and effacing (A/E) lesions. A/E lesions are characterized by intimate bacterial adherence to intestinal cells and destruction of microvilli, which leads to diarrhea. Therefore, studies interrogating the regulation of the locus of enterocyte effacement (LEE) are critical for understanding the molecular epidemiology of EPEC infections and developing interventional strategies. Hitherto, most studies have centered on protein-based regulators, whereas the role of small regulatory RNAs remains underappreciated. Previously, we identified the first sRNAs—MgrR, RyhB, and McaS—that regulate the LEE of EPEC. This study was undertaken to identify additional sRNAs that impact the LEE. Our results suggest that the catabolite-responsive sRNA, Spot42, indirectly controls the LEE by inhibiting synthesis of its inducer, indole. Spot42 base-pairs with the tnaCAB mRNA and presumably destabilizes the transcript, thereby preventing expression of the regulatory and structural proteins that are involved in the import and hydrolysis of tryptophan into indole. The absence of intracellular indole leads to reduced transcription of the LEE1-encoded master transcriptional activator Ler, thereby maintaining the LEE in its silenced state and delaying A/E lesion morphogenesis. Our results highlight the importance of riboregulators that synchronize metabolic and virulence pathways in bacterial infection.


Introduction
Small regulatory RNAs (sRNAs) constitute a structurally and mechanistically assorted class of regulatory nucleic acids that confer numerous advantages upon their host bacterium [1,2]. For instance, sRNAs, due to their small size and the fact that they forgo translation, are metabolically inexpensive to synthesize and maintain, thereby shortening the response time for cells to acclimate and adapt to environmental fluctuations [3,4]. sRNAs also enhance the regulatory and phenotypic range by fine-tuning the transcriptional output [1,2,4]. Additionally, sRNA-regulated circuits are more tolerant to mutational events. A subclass of sRNAs, termed trans-encoded sRNAs, are encoded distantly from their regulated targets. These sRNAs share limited complementary to their target mRNAs and, thus, typically require assistance from a chaperone protein to facilitate base-pairing interactions [1]. One of the most frequently used chaperones is the molecular matchmaker Hfq. The holoprotein functions as a homohexameric toroid [5]. Hfq binds to a sRNA at one surface and an mRNA on the other, bringing the sRNA-mRNA pair in proximity to facilitate base-pairing [1,2,6]. Base-pairing of the sRNA to the mRNA can result in a varying range of regulatory outcomes including transcriptional or

DNA Manipulations
Polymerase chain reactions (PCRs), restriction digestions, ligations, cloning, and transformations were performed following standard protocols [38]. Chromosomal modifications in E. coli were engineered by lambda red-mediated recombineering essentially as described in our previous paper [4].

Screen to Identify Hfq-Dependent sRNA Regulators of Indole Biosynthesis
Electrocompetent EPEC was transformed with individual members of a plasmid library, each of which overproduces a solitary Hfq-dependent sRNA of E. coli, under the transcriptional control of an IPTG-inducible promoter. The plasmid library was a generous gift from Susan Gottesman, in whose lab it was engineered [39]. Thereafter, the EPEC transformant library was screened to identify sRNAs that regulate indole biosynthesis. Briefly, an axenic colony of each EPEC transformant was inoculated in LB broth supplemented with ampicillin and cultured overnight at 37 • C under shaking conditions. The next day, each transformant was sub-cultured 100-fold in LB broth supplemented with ampicillin and IPTG and grown to an OD 600 of 1.2-1.4. One mL aliquot of each culture was assayed for indole biosynthesis by addition of 5 drops of Kovac's reagent. Kovac's reagent reacts with indole to produce a red colored dye, rosindole. Rosindole production was visually monitored over a period of 20 s, after which the cultures were photographed.

Beta-Galactosidase Assay
Beta-galactosidase assays on the translational fusion (P araBAD -tnaA'-'lacZ) containing reporter strains were performed as described in our previous paper [19], with the exception that bacteria were grown to an optical density of 1.1-1.4 in LB supplemented with ampicillin, arabinose, and IPTG. The same conditions, as described above, were employed for the transcriptional fusion (P LEE1 -LEE1'-lacZ + ) containing reporter strains with the omission of arabinose from the medium.

RNA Isolation and qRT-PCR
RNA isolation and qRT-PCR were performed as described in our previous paper [19]. The steady-state mRNA levels were normalized with respect to the housekeeping 16S rRNA transcript.

Preparation of Cell Lysates for Western Blotting
Bacterial cultures were processed for western blotting as described previously [19]. The abundance of tryptophanase protein was measured by using a polyclonal anti-tryptophanase antibody, at a 4000-fold dilution, which was purchased from Assaypro (St. Charles, MO. Cat # 33517-05111). Protein loading was controlled for by staining for the total protein transfer using Ponceau S stain.

Overproduction of Spot42 Inhibits Indole Biosynthesis by Down-Regulating the tnaCAB mRNA
We used a genetic screen to identify sRNA regulators of indole biosynthesis as described in materials and methods. EPEC transformants were grown to the desired optical density and qualitatively assayed for indole synthesis by the addition of Kovacs reagent, which reacts with indole to produce the red colored pigment, rosindole [40]. Thus, the intensity of the red color directly correlates with the amount of indole biosynthesis. It was observed that, of all the sRNAs, Spot42 maximally inhibited indole biosynthesis, evident from the reduced synthesis of rosindole ( Figure 1A). Next, we interrogated the molecular basis of the resulting phenotype. Consistent with the observed reduction in indole levels, overproduction of Spot42 negatively regulated the steady-state levels of tnaC, tnaA, and tnaB ( Figure 1B). The observed reduction in transcript levels resulted in reduced protein expression as the abundance of tryptophanase was also reduced in the overproducer ( Figure 1C). Collectively, these results suggest that overexpression of the Hfq-dependent sRNA Spot42 inhibits indole production by negatively regulating the transcript levels of tnaCAB. overproducer ( Figure 1C). Collectively, these results suggest that overexpression of the Hfq-dependent sRNA Spot42 inhibits indole production by negatively regulating the transcript levels of tnaCAB.
Overproduction of Spot42 inhibits indole synthesis by repressing the tnaCAB mRNA. Overnight grown cultures of EPEC (pBR-plac) or EPEC (pSpot42) were sub-cultured in Luria-Bertani (LB) medium supplemented with ampicillin and IPTG and grown under static conditions to an OD600 of 1.2-1.4, after which Kovacs' reagent was added to assay for indole production (A). For total RNA and protein extraction, overnight cultures of EPEC (pBR-plac) and EPEC (pSpot42) were sub-cultured identically, with the exception that cultures were grown under shaking conditions. The steady-state levels of tnaC, tnaA, and tnaB transcripts were determined by qRT-PCR and normalized with respect to the 16S rRNA transcript (B), whereas the tryptophanase protein level was quantified by western blotting (C). Each experiment was repeated on at least two separate occasions with replicates being used in each experiment. Similar results were obtained in each trial. The depicted results are representative of one such trial. For the qRT-PCR, error bars depict standard deviation. Student's t-test was used to assay for statistical significance of the difference in the means between the wild type strain containing pBR-plac and the Spot42 overexpressor. A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005 and * denotes a p-value of < 0.05. For Western blotting, protein loading was controlled for by assaying for total protein transfer using Ponceau S stain following electroblotting. Overnight grown cultures of EPEC (pBR-plac) or EPEC (pSpot42) were sub-cultured in Luria-Bertani (LB) medium supplemented with ampicillin and IPTG and grown under static conditions to an OD 600 of 1.2-1.4, after which Kovacs' reagent was added to assay for indole production (A). For total RNA and protein extraction, overnight cultures of EPEC (pBR-plac) and EPEC (pSpot42) were sub-cultured identically, with the exception that cultures were grown under shaking conditions. The steady-state levels of tnaC, tnaA, and tnaB transcripts were determined by qRT-PCR and normalized with respect to the 16S rRNA transcript (B), whereas the tryptophanase protein level was quantified by western blotting (C). Each experiment was repeated on at least two separate occasions with replicates being used in each experiment. Similar results were obtained in each trial. The depicted results are representative of one such trial. For the qRT-PCR, error bars depict standard deviation. Student's t-test was used to assay for statistical significance of the difference in the means between the wild type strain containing pBR-plac and the Spot42 overexpressor. A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005 and * denotes a p-value of < 0.05. For Western blotting, protein loading was controlled for by assaying for total protein transfer using Ponceau S stain following electroblotting.

The tnaC-tnaA Intergenic Region Is Sufficient for Spot42-Dependent Repression
Because Spot42, like the vast majority of Hfq-dependent sRNAs, exerts its effect by base-pairing to target mRNAs, next, we aligned the sRNA with the tnaCAB mRNA to identify potential regions of complementarity [41]. IntaRNA revealed a 15-nucleotide tract on Spot42 that was complementary to the intergenic region of tnaC and tnaA, situated downstream of the Rho-dependent transcription terminator of tnaC (Figure 2A) [32]. This suggests that Spot42 may directly base pair to the tnaCAB mRNA and affect mRNA stability and/or translation. Interestingly, the predicted base-pairing region of Spot42 has previously been shown to be important in antisense regulation of multiple mRNAs in E. coli [37]. To test for antisense regulation, the predicted intergenic region of tnaC-tnaA along with 45 nucleotides of the downstream tnaA open reading frame (ORF) were recombineered upstream of a truncated 'lacZ gene that lacks its native 5 UTR and translation initiation codon ( Figure 2B). Recombineering generates a single copy tnaA'-'lacZ translational fusion whereby the lacZ gene is under the posttranscriptional regulatory elements of tnaA and transcriptional control of the araBAD promoter (P araBAD -tnaA'-'lacZ) ( Figure 2B) [34]. Overexpression of Spot42 in this reporter E. coli strain diminished β-galactosidase activity from the minimal P araBAD -tnaA'-'lacZ fusion by 4-fold ( Figure 2C), while having no effect on the control P araBAD -ler'-'lacZ fusion (data not shown). These results suggest that Spot42 specifically targets the 5 leader region of tnaA to repress gene expression. To further validate duplex formation, we used a previously engineered Spot42 mutant allele, Spot42-II, containing a trinucleotide substitution in the predicted base-pairing motif [37]. However, this mutant allele was fully proficient in repressing the tnaA'-'lacZ fusion (data not shown), suggesting that perhaps a larger polynucleotide mutation was necessary for any observable perturbation in regulation by Spot42. Therefore, we engineered a mutant allele, designated Spot42-mut4, in which the entire base-pairing region (15 nucleotides) was substituted with a scrambled sequence (Figure 2A). The Spot42-mut4 allele is not predicted to base pair to the wild type tnaA'-'lacZ fusion. Consistent with this prediction, Spot42-mut4 did not repress β-galactosidase activity from the P araBAD -tnaA'-'lacZ fusion ( Figure 2C), confirming the essentiality of the polynucleotide tract in antisense regulation of the tnaA'-'lacZ translational fusion by Spot42.

Spot42 Represses Transcription from the LEE1 Promoter
The metabolite indole activates transcription from the LEE1 promoter in EPEC [31]. The first gene in the LEE1 operon encodes Ler, which is the master regulator that spatiotemporally synchronizes transcription from the other LEE operons (LEE2-5) to stimulate morphogenesis of A/E lesions. Because indole levels were down-regulated in the Spot42 overproducer, we hypothesized that transcriptional activity from LEE1 would be diminished. Consistent with this prediction, overexpression of Spot42 negatively regulated β-galactosidase activity by 3-fold in a reporter E. coli strain harboring a chromosomal P LEE1 -ler'-lacZ + transcriptional fusion ( Figure 3). Because transcription from the LEE1 promoter is reduced, one would anticipate a correlative reduction in expression of Ler-activated targets and A/E lesion biogenesis.
In summary, we have demonstrated that the Hfq-dependent sRNA Spot42 base-pairs to the tnaCAB mRNA to globally silence gene expression from the entire transcript presumably by destabilizing it. The reduction in mRNA levels, in turn, leads to reduced expression and steady-state levels of the structural and regulatory proteins involved in indole biosynthesis (Figure 4). In the absence of indole, the LEE1 promoter is not optimally active and Ler transcription is reduced (Figure 4), which, in turn, is expected to delay and diminish the morphogenesis of A/E lesion. . Spot42 base-pairs to the tnaCA intergenic region to exert its silencing effect. IntaRNA revealed a region of complementarity between Spot42 and the intergenic region tnaC-tnaA, located immediately downstream of the Rho-dependent transcription terminator of tnaC (A). A polynucleotide substitution in Spot42, which generates the Spot42-mut4 allele, is predicted to abolish base-pairing and the ensuing negative regulation of tnaCAB by Spot42 (A). The tnaCA intergenic region along with 45 nucleotides of the tnaA ORF were fused to a truncated lacZ gene, which lacks its native 5 UTR and 9 of the N-terminal codons, to generate a tnaA'-'lacZ translational fusion. This fusion is transcriptionally driven by the heterologous P araBAD promoter (B). The tnaA'-'lacZ reporter strains harboring pBR-plac or pSpot42 were grown in LB supplemented with ampicillin, arabinose, and IPTG. Overexpression of Spot42 repressed β-galactosidase activity from the minimal P araBAD -tnaA'-'lacZ translational fusion and mutation of the predicted base-pairing region of spf prevented Spot42 from repressing the tnaA'-'lacZ fusion (C). Each experiment was repeated on at least two separate occasions with replicates being used in each experiment. Similar results were obtained in each trial. The depicted results are representative of one such trial. Error bars depict standard deviation. Student's t-test was used to assay for statistical significance of the difference in the means between EPEC(pBR-plac)/EPEC(pSpot42) and EPEC(pSpot42)/EPEC(pSpot42-mut4) pairs. A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005. Figure 3. Spot42 represses transcription of ler and Ler-regulated targets. E. coli strain MC4100 LEE1'-lacZ + transformants containing pBR-plac or pSpot42 were grown in LB broth containing ampicillin and IPTG to an OD600 of 1.1-1.4 and assayed for β-galactosidase activity. Overexpression of Spot42 repressed β-galactosidase activity from the LEE1'-lacZ + fusion. Each assay was conducted on at least two separate occasions with replicate samples being used in each trial. Student's t-test was used to assay for statistical significance of the difference in the means between the LEE1'-lacZ + (pBR-plac) vs LEE1'-lacZ + (pSpot42). A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005. The nutritional status of a bacterium can impact its virulence potential. For instance, the quality and/or quantity of a carbon source, such as sugars, dictate the molecular epidemiology of a bacterial infection [42,43]. Reciprocally, prominent metabolic regulators, such as CRP, Cra, KdpE, and CsrA, Overexpression of Spot42 repressed β-galactosidase activity from the LEE1'-lacZ + fusion. Each assay was conducted on at least two separate occasions with replicate samples being used in each trial. Student's t-test was used to assay for statistical significance of the difference in the means between the LEE1'-lacZ + (pBR-plac) vs. LEE1'-lacZ + (pSpot42). A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005. Figure 3. Spot42 represses transcription of ler and Ler-regulated targets. E. coli strain MC4100 LEE1'-lacZ + transformants containing pBR-plac or pSpot42 were grown in LB broth containing ampicillin and IPTG to an OD600 of 1.1-1.4 and assayed for β-galactosidase activity. Overexpression of Spot42 repressed β-galactosidase activity from the LEE1'-lacZ + fusion. Each assay was conducted on at least two separate occasions with replicate samples being used in each trial. Student's t-test was used to assay for statistical significance of the difference in the means between the LEE1'-lacZ + (pBR-plac) vs LEE1'-lacZ + (pSpot42). A p-value of < 0.05 was considered to be statistically significant. ** denotes a p-value < 0.005. The nutritional status of a bacterium can impact its virulence potential. For instance, the quality and/or quantity of a carbon source, such as sugars, dictate the molecular epidemiology of a bacterial infection [42,43]. Reciprocally, prominent metabolic regulators, such as CRP, Cra, KdpE, and CsrA, The nutritional status of a bacterium can impact its virulence potential. For instance, the quality and/or quantity of a carbon source, such as sugars, dictate the molecular epidemiology of a bacterial infection [42,43]. Reciprocally, prominent metabolic regulators, such as CRP, Cra, KdpE, and CsrA, that perceive environmental carbon sources and/or monitor the intracellular metabolic state, moonlight as virulence regulators in diverse pathogens [25,44,45]. The catabolite-responsive Hfq-dependent sRNA Spot42, specified by the spf gene, is a well-studied riboregulator that is present in both pathogenic and nonpathogenic strains of E. coli [37,46]. The sRNA enables its host bacterium to switch between different carbon sources [37,47]. Preferred carbon sources, such as glucose, induce transcription of Spot42 in a CRP-dependent manner [48]. Spot42, in turn, base-pairs to transcripts to repress the expression of proteins involved in the uptake and/or metabolism of secondary carbon sources, thereby enabling the bacterium to utilize the preferred carbon source [37]. Upon depletion of the preferred carbon source, cAMP levels are elevated, which, in turn, interacts with CRP and activates it [49,50]. The CRP-cAMP holoprotein represses transcription of spf, thereby derepressing mRNAs involved in metabolism of secondary carbon sources [37,48]. Thus, CRP and Spot42 are vital regulators involved in carbon catabolite repression (CCR) [51]. Whereas CRP is the primary transcriptional factor involved in CCR in E. coli [51], Spot42 fine-tunes CCR by posttranscriptional control of the synthesized mRNAs [37,52].
The tnaCAB catabolic operon houses the regulatory and structural gene products that enables its host bacterium to utilize tryptophan as a carbon and energy source [28,53]. Additionally, in EPEC, the hydrolytic product of tryptophan, indole, also induces transcription from the LEE1 promoter, thereby affecting bacterial virulence. Thus, the gene products encoded within the tnaCAB operon bridge cellular metabolism with virulence in EPEC. As expected, in E. coli, tnaCAB is transcriptionally activated by CRP [54,55]. The promoter architecture of the tnaCAB operon is preserved between E. coli and EPEC, suggesting that the CRP-dependent transcriptional regulation is intact in this A/E pathogen [25]. Our discovery that the polycistronic tnaCAB mRNA is also subject to antisense regulation by Spot42 suggests that CCR exerts a multi-tiered regulatory control to fine-tune gene expression from the tnaCAB operon. Based on our results, it is reasonable to hypothesize that Spot42 levels are likely to be higher in EPEC when the bacterium is in an environment where the expression from the LEE is not needed, such as outside of the mammalian host. Upon entry into the small intestine, Spot42 levels are likely to be down-regulated. This down-regulation would lead to derepression of the tnaCAB mRNA and stimulate indole biosynthesis. Indole production, in turn, would activate ler and prime the LEE regulatory cascade to culminate with the biogenesis of A/E lesions. Experiments to test this hypothesis are underway in our lab.
In summary, the presented results expand the repertoire of genes, co-regulated by both CRP and Spot42, that bridge bacterial metabolism and virulence. Moreover, these results also highlight a recurring, but underappreciated, theme in the growing body of work on A/E pathogens, about the importance of regulatory RNAs in modulating pathogenetic pathways.