Hfq is an RNA chaperone that functions as a homohexameric toroidal protein with distinct proximal and distal surfaces that facilitate sRNA-mRNA transactions (De Lay et al., 2013). Structural studies with Hfq reveal that its proximal surface binds to polyuridine tracts located downstream of a stem-loop—a structural feature abundant in Hfq-dependent sRNAs (Valentin-Hansen et al., 2004; De Lay et al., 2013). Meanwhile, the distal surface binds to tandem poly-(A-R-N) repeats, where A, R, and N represent adenine, purine, and any nucleotide respectively. E. coli mRNAs are replete with ARN repeats suggesting that Hfq preferentially associates with mRNAs by using its distal face (Link et al., 2009; De Lay et al., 2013). Hfq can simultaneously utilize its proximal and distal faces and facilitate sRNA-mRNA pairing (Link et al., 2009; De Lay et al., 2013). The relatively relaxed sequence recognition enables Hfq to control numerous cellular processes including virulence (Chao and Vogel, 2010). In both EHEC and EPEC Hfq controls the LEE with varying regulatory outcomes in a pathotype-specific manner (Hansen and Kaper, 2009; Shakhnovich et al., 2009; Kendall et al., 2011; Figure 1). In the EHEC strain EDL933, Hfq globally silences gene expression from the LEE through two independent regulatory pathways (Hansen and Kaper, 2009; Shakhnovich et al., 2009; Figure 1). In the exponential phase, inactivation of hfq stabilizes the grlRA mRNA (Hansen and Kaper, 2009). Because increased expression of grlA is epistatic to grlR this results in transcriptional activation of ler, which, in turn, activates the other LEE-encoded operons and stimulates pedestal formation in the hfq mutant. Meanwhile, in the stationary phase, the effect of Hfq is independent of grlRA since Hfq-dependent repression of the LEE is intact in the grlRA mutant (Hansen and Kaper, 2009). Presumably, this effect involves direct translational repression of ler since a ler'-‘lacZ translational fusion containing the 5′ UTR of ler that is transcribed from a heterologous GrlA-independent promoter is still regulatable by Hfq (Shakhnovich et al., 2009; Figure 1). Curiously, in the related EHEC biotype 86-24, the hfq mutant exhibits a starkly contrasting phenotype compared to the hfq mutant of EHEC EDL933 (Kendall et al., 2011). In EHEC 86-24, loss of hfq globally diminishes gene expression from the LEE in a ler-dependent manner. This suggests that Hfq functions as an activator, rather than a repressor, of the LEE in EHEC 86-24 (Kendall et al., 2011; Figure 1). However, whether the effect is direct or indirect remains to be elucidated. The antagonistic role of Hfq in EDL933 and 86-24 has been attributed to other genotypic differences such as the presence/absence of strain-specific sRNAs that lead to the observed regulatory outcomes. Whereas, the physiological role of Hfq in EHEC virulence has received considerable attention, the role of Hfq in EPEC has only been investigated superficially. In EPEC, inactivation of hfq derepresses the expression of GrlA and the GrlA-regulated LEE genes, suggesting that Hfq has a similar role to that observed in the EHEC strain EDL933 (Hansen and Kaper, 2009; Shakhnovich et al., 2009; Figure 1). However, the molecular details in EPEC have not been addressed.

Ongoing studies have finally illuminated the elusive Hfq-dependent sRNAs that coregulate the LEE in EHEC 86-24. Using RNA sequencing, Gruber and Sperandio identified seven novel EHEC-specific sRNAs (Gruber and Sperandio, 2015). The expression of all but one of these sRNAs was diminished in the hfq mutant. Multiple Hfq-dependent sRNAs—sRNA350, sRNA103, and sRNA56—were shown to activate the LEE by affecting different targets and to varying degrees (Gruber and Sperandio, 2015; Figure 1). For instance, overexpression of sRNA350 globally activated transcription from all the LEE-encoded operons by affecting the master regulator ler (Gruber and Sperandio, 2015; Figure 1). However, the direct target of this riboregulator remains to be determined. Interestingly, the genetic architecture of sRNA350 does not conform to that of prototypical Hfq-dependent sRNAs. Most Hfq-dependent sRNAs are encoded by monocistronic transcription units; however, sRNA350 is encoded by the 3′ UTR of the LEE-encoded cesF gene, which specifies the chaperone for the T3S effector protein EspF (Elliott et al., 2002). Furthermore, sRNA350 does not appear to be posttranscriptionally cleaved and exerts its regulatory effect as part of the cesF transcript (Gruber and Sperandio, 2015). In contrast to sRNA350, sRNA103 and sRNA56 selectively target the LEE4-encoded espA transcript with the former eliciting a stronger regulatory response. However, neither sRNA appears to be complementary to espA, suggesting that the observed regulatory effect is mediated indirectly via an intermediate factor (Gruber and Sperandio, 2015). sRNA103 and sRNA56 also affect other genes scattered elsewhere in the EHEC genome (Gruber and Sperandio, 2015).

Besides EHEC-specific sRNAs, ancestral sRNAs, conserved between non-pathogenic and pathogenic lineages of E. coli, also regulate the LEE in EHEC 86-24. The two conserved Hfq-dependent sRNAs—GlmY and GlmZ—control the expression from the LEE4 and LEE5 operons as well as the non-LEE encoded gene espF_U (Gruber and Sperandio, 2014, 2015). GlmY and GlmZ are paralogous sRNAs that were originally identified as translational activators of the enzyme Glucoseamine-6-phosphate synthase (GlmS; Kalamorz et al., 2007; Reichenbach et al., 2008; Urban and Vogel, 2008; Waters and Storz, 2009). Despite extensive identity, GlmY and GlmZ exert their regulatory effects via distinct mechanisms. Unprocessed GlmZ possesses a seed region that base-pairs to and activates translation from the glmS transcript. GlmY, however, lacks the seed region and therefore does not base-pair to glmS. Rather, GlmY functions indirectly by preventing the processing of GlmZ by the enzyme RapZ, thereby increasing the cellular availability of unprocessed GlmZ to promote translation from glmS (Kalamorz et al., 2007; Reichenbach et al., 2008; Urban and Vogel, 2008; Waters and Storz, 2009). In EHEC, both GlmY and GlmZ destabilize the LEE4 and LEE5 encoded polycistronic transcripts while enhancing translation of espF_U by promoting cleavage in the intergenic region of the espJ-espF_U transcript (Gruber and Sperandio, 2014; Figure 1). GlmZ directly base-pairs to the LEE4 transcript and selectively destabilizes the 3′ segment, containing espADB and the downstream ORFs, while having no effect on the 5′ segment of the transcript that contains sepL (Gruber and Sperandio, 2014). Direct evidence for duplex formation between GlmZ and LEE4 was provided by site-directed mutations within the seed region of GlmZ as well as compensatory mutations in its target site on LEE4. Furthermore, in a subsequent study the same authors clarified the role of GlmY in regulation of the LEE4 operon (Gruber and Sperandio, 2015). Here, they demonstrated that increased gene expression from the LEE and the ensuing A/E lesion formation observed in the ΔglmY mutant is abolished in the ΔglmY ΔrapZ double mutant suggesting that rapZ is epistatic (or downstream) to glmY (Gruber and Sperandio, 2015). This observation, coupled to the fact that overexpression of GlmY represses the LEE4 transcript without affecting the LEE4 promoter activity, suggest that GlmY post transcriptionally represses LEE4 indirectly by binding to and sequestering RapZ from GlmZ. Free GlmZ, in turn, directly binds to and destabilizes the LEE4 transcript thereby reducing pedestal formation. In other words, the GlmY- and GlmZ- dependent regulation of LEE4 occurs in a manner similar to how these paralogs affect the expression of glmS. Interestingly, EspA, EspD, EspB, and some of the other downstream-encoded proteins are structural components of the T3S translocon, whereas SepL, along with SepD, functions as a regulatory switch that promotes the hierarchical secretion of translocators over effectors (Wang et al., 2008). SepL binds to effectors, such as Tir, effectively sequestering them until the maturation of the T3SS (Wang et al., 2008). Thereafter, the SepL/SepD switch triggers the shift from translocator to effector secretion. Perhaps, GlmZ and GlmY are expressed after the assembly of the T3SS when EspA, EspB, and EspD are no longer required but SepL is still needed to synchronize the hierarchical order of effector secretion, including EspF_U. Paradoxically, a counterintuitive discovery made in this study was that GlmY and GlmZ antagonistically regulated targets all of which are required for the morphogenesis of pedestals in EHEC (Gruber and Sperandio, 2014). For instance, the proteinaceous factors encoded within LEE4, LEE5, and espF_U promote A/E lesions in EHEC; however, GlmY and GlmZ negatively regulated LEE4 and LEE5 but positively regulated espF_U. The authors propose an attractive hypothesis that perhaps these regulatory sRNAs limit the uncontrolled expression from the LEE and synchronize it with the non-LEE encoded gene espF_U (Gruber and Sperandio, 2014). This mechanism would ensure that physiologically precise stoichiometric ratios of the architectural and secreted proteins are synthesized, which has been shown to be critical for the formation of A/E lesions and successful infection of the host. However, it remains to be determined if the sRNA-dependent regulation observed in EHEC 86-24 extends to other biotypes of EHEC, which form pedestals by the same mechanism. Interestingly, the linker protein EspF_U, essential for EHEC to form pedestals, is not present in the genome of EPEC and the bacterium relies on a different posttranslational mechanism for pedestal formation (Mellies et al., 2007). This observation suggests that GlmY and GlmZ are unlikely to be functionally equivalent in EHEC and EPEC with regards to pedestal formation. This regulatory divergence between EHEC and EPEC appears to extend to another conserved Hfq-dependent sRNA—DsrA. In all tested strains of EHEC, DsrA activates the transcription of ler in an RpoS-dependent manner. However, DsrA does not affect the LEE in EPEC (Laaberki et al., 2006). This is in stark contrast to conserved proteinaceous transcription factors that regulate the LEE identically between EHEC and EPEC. For instance, the DNA-binding proteins H-NS, Fis, and GrlA modulate the LEE similarly in all A/E pathogens (Mellies et al., 2007; Bhatt et al., 2011). Thus, conserved sRNAs appear to be more malleable to regulatory rewiring in order to elicit strain-specific plastic responses for conserved morphogenetic pathways—a trait that is particularly advantageous in adapting pathogens to different niches. Moreover, these findings also suggest that it would be ill-advised to extrapolate the role of conserved sRNAs in one A/E pathogen based upon its role in another, and that their functions must be experimentally deduced in each member.

Besides trans-encoded sRNAs, at least one cis-encoded sRNA, antisense regulator of ler RNA (arl), has been implicated in regulation of the LEE in the EHEC strain Sakai (Tobe et al., 2014; Figure 1). The arl gene is located downstream of ler but transcribed from the antisense strand. Consequently, Arl exhibits extensive complementarity to the LEE1-encoded ler mRNA. The transcription of arl is stimulated by elevated cytoplasmic levels of iron (Fe²⁺) or hydroxyl (OH⋅) radical but does not require the iron-responsive transcriptional factor Ferric uptake regulator (Fur; Tobe et al., 2014). Arl regulates the ler-encoded LEE1 mRNA posttranscriptionally by specifically targeting the 3′ region of ler, over a region spanning the C-terminal domain of ler as well as the 3′ UTR. This conclusion is based on the observation that Arl-dependent regulation of ler is intact when just the C-terminal coding region of ler and its 3′ UTR is translationally fused to MBP and this chimeric MBP'-‘ler construct is transcriptionally driven by the heterologous lac promoter (Tobe et al., 2014). Moreover, Arl not only destabilizes the LEE1 mRNA but also directly impacts translation completion from the ler ORF. As cytoplasmic iron is depleted the transcription of arl is reduced and this enhances the stability and translation from the ler-encoding LEE1 mRNA, which in turn primes the LEE regulatory cascade that culminates with morphogenesis of A/E lesions (Tobe et al., 2014). These regulatory and phenotypic observations indisputably suggest that Arl posttranscriptionally controls the LEE1 mRNA, presumably by direct base-pairing. However, the role of Arl in EPEC as well as the other EHEC biotypes has not been explored.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.