Small RNAs Repress Expression of Polysaccharide Utilization Loci of Gut Bacteroides Species

TEXT

It has been known for decades that many human gut bacteria have the ability to utilize a wide variety of plant polysaccharides (PS) and animal glycans for their metabolism and growth (1, 2). The members of Bacteroides, a genus with numerous species abundant in the human colon, are especially adept at utilizing the vast array of PS and glycans that arrive to the colon from the diet or are present on the mucins overlying the epithelial layer. Early studies analyzing genes necessary for starch utilization in Bacteroides thetaiotaomicron led to the identification of the sus locus (3–6), the first characterized locus of a Bacteroides species encoding a cluster of genes necessary to utilize a complex carbohydrate. It was not until the publication of the first Bacteroides genome sequence in 2003 (7) that the vast number of these polysaccharide utilization loci (PUL), as they were named (8), was appreciated. The genome of the B. thetaiotaomicron type strain contains 88 PUL with an average of 10 genes per locus (9), with each locus dedicated to the utilization of a dietary PS or host glycan. All Bacteroides PUL encode SusC and SusD orthologs. SusD proteins are outer surface lipoproteins that bind intact PS or their partially digested oligosaccharides, allowing them to be transported to the periplasm via the outer membrane SusC porin-like proteins. Most PUL also encode additional products involved in PS utilization, including surface and periplasmic glycoside hydrolases and regulatory proteins. In the 13 years since the publication of the first Bacteroides genome sequence, a wealth of information has been acquired regarding the process of utilization of plant and animal glycans. In addition, the regulation of PUL gene expression has been extensively studied. In this issue of the Journal of Bacteriology, Cao et al. report on a new regulatory mechanism whereby antisense small RNAs (sRNA) repress expression of genes of their cognate PUL (10). This regulatory mechanism is shown to be common to PULs of Bacteroides fragilis involved in the utilization of host glycans and also to be a conserved feature of other human gut Bacteroides.

PULs typically encode one of two main types of regulatory systems: hybrid two-component systems (HTCS) or extracytoplasmic function (ECF) sigma factor/anti-sigma factor systems. The Bacteroides species use one of these systems to sense when a particular PS is present and to induce the expression of the PUL to utilize that particular PS. HTCS span the inner membrane with the periplasmic sensor domain typically recognizing an oligosaccharide component of the PS (11) that has been transported to the periplasm by the SusC ortholog. The signal is then transmitted via phosphorelay to the cytoplasmic DNA binding domain of the HTCS to upregulate expression of the cognate PUL. In the ECF σ/anti-σ factor system, the inner membrane anti-σ factor sequesters the ECF sigma factor until it receives a signal indicating that its cognate substrate is available, causing it to release the sigma factor to initiate transcription of the PUL. Therefore, both of these systems respond to substrate availability and upregulate expression of their cognate PUL. In B. thetaiotaomicron, ECF σ/anti-σ factors predominate in PULs involved in utilization of host O-glycans rather than dietary PS (9).

As the colon has a constant supply of host glycans and variable dietary polysaccharides, Bacteroides species must be able to quickly adapt to the nutrient availability. Bacteroides species can simultaneously utilize numerous distinct PS but are also able to prioritize certain PS/glycans over others and repress expression of lower-priority PUL even in the presence of its substrate (15). B. thetaiotaomicron preferentially utilizes dietary PS rather than host glycans, and the PUL for utilization of certain O-glycans are not transcribed when both O-glycans and dietary PS are present (12). This then raises the issue of how one PUL is transcriptionally repressed when a higher-priority substrate is present. Single monosaccharides such as fructose and galactose, components of dietary PS, were shown to repress expression of PUL dedicated to O-glycan utilization in B. thetaiotaomicron (12). In that study, the intergenic region between the gene encoding the anti-σ factor and the susC ortholog of an O-glycan utilization locus was shown to be necessary for repression of the PUL when higher-priority PSs were present. The current report by Cao et al. (10) provides a new insight revealing that such intergenic regions carry sRNA involved in repression of their cognate PUL.

Researchers in the Smith laboratory previously published a study characterizing a unique PUL of B. fragilis termed Don, which was shown to encode genes involved in the harvest and utilization of N-linked glycans from host proteins (13). The Don PUL is a seven-gene operon that begins with genes encoding an ECF-σ factor and an anti-σ factor (donA and donB), followed by a 182-bp intergenic gap and then the susC and susD orthologs donC and donD, followed by donE and donG, encoding predicted glycoside hydrolases involved in N-deglycosylation of host glycoproteins (13). The Don locus is highly induced in defined medium with mucin glycans or transferrin as the sole carbohydrate source. Cao et al. sought to better study the regulation of Don gene expression and performed a differential transcriptome sequencing (differential RNA-seq) analysis of primary transcripts of B. fragilis grown in medium with added glucose, a medium that does not induce expression of Don genes. This analysis revealed the presence of an ∼125-nucleotide (nt) RNA transcribed from the opposite strand in the 182-bp intergenic gap between donC and donB encoding the upstream anti-σ factor. Further analysis of their RNA-seq data revealed 14 other sRNA sized between 78 and 128 nt transcribed in the intergenic regions upstream of and divergently from the susC orthologs in other PUL of B. fragilis. It was noted that 14 of these 15 PUL contained genes encoding the ECF σ-factor and anti-σ factor regulators rather than the HTCS. As B. fragilis has 52 PUL, nearly 30% of their PULs synthesize antisense sRNA.

To understand the role of these sRNA in the regulation of their cognate PUL, Cao et al. studied the sRNA from the Don locus, termed DonS. A mutation of donS led to a 5-fold increase in donC expression under inducing conditions. Overexpression of donS repressed donC expression more than 400-fold under inducing conditions. In addition, the 10-kb Don transcript was not detected by Northern blot analysis, and the bacterium lost the ability to deglycosylate transferrin and the ability to grow with transferrin as the carbon source. They further showed that two other sRNA of B. fragilis PULs similarly repressed expression of their cognate susC orthologs. In addition, the investigators showed that a subset of PULs of three other abundant gut Bacteroides species, including B. thetaiotaomicron, have a similar genetic organization, with an antisense sRNA transcribed between the anti-σ factor gene and susC ortholog, and they confirmed the existence of four sRNA in the PULs of B. thetaiotaomicron. Therefore, this level of PUL regulation by sRNA is conserved in gut Bacteroides.

Using protein modeling software, putative functions were assigned to the glycoside hydrolases of the 14 other PUL of B. fragilis with sRNA and those of the other Bacteroides species. The glycoside hydrolases were limited to the digestion of host-derived polymers rather than dietary polysaccharides. This is consistent with previous findings revealing that PULs encoding ECF-σ/anti-σ factors are predominantly dedicated to the utilization of host-derived glycans (9).

This leaves two main questions for future studies. The first is exactly how these sRNA repress expression of their cognate PUL, and the second is whether these sRNA are involved in catabolite repression when a higher-priority PS is present. The promoters of these sRNA are somewhat conserved, as they overlap the start codon of the susC orthologs and are predicted to be constitutively transcribed. The overall model for DonS repression is that as inducing substrate concentrations decline, transcript quantity from the constitutive transcription of the antisense sRNA will exceed the amount of Don mRNA, preventing its accumulation. One proposed putative mechanism is the formation of duplex RNA between DonS and the Don transcript that targets it for degradation. A second proposed mechanism is that of transcriptional interference by collision of RNA polymerases initiating from divergent promoters (the donS promoter and the promoter upstream of donA). The first sRNA described in Bacteroidetes, RteR, was shown to inhibit transfer of a Bacteroides conjugative transposon. The data suggest that RteR interacts with the operon encoding the transfer (tra) genes and effectively terminates transcription within the operon (14). The proposal of a similar mechanism of transcriptional termination by these sRNA is supported by the finding that the intergenic region between the anti-σ factor gene and the susC ortholog of a B. thetaiotaomicron O-glycan utilization PUL is where the PUL transcript is destabilized (12). As the sRNA promoters overlap the start codon of their susC orthologs, the investigators propose that DonS may have a role, in addition to affecting transcript level, in repressing donC translation.

Whether these sRNA have a role in catabolite repression is still unclear. In the study by Cao et al., the sRNA repression effect occurred in inducing mucin glycan medium without added glucose or other repressing monosaccharides, demonstrating that the repression by these sRNA is independent of inhibiting substrate. However, the ability of glucose to repress Don expression was not as severe in the donS mutant as it was in the wild type. Whether this finding reflects independent additive effects of DonS and glucose repression or whether DonS plays a role in catabolite repression is still to be determined. In summary, the study by Cao et al. reveals a fascinating level of PUL regulation that is conserved in gut Bacteroides and represses transcription of PUL involved in utilization of host glycans.