In pathogenic E. coli strains belonging to different pathotypes, the association between specific virulence factors and the corresponding pathotype has been controversial in some instances. The reason for that is that some of the isolates that belong to the pathotype do not express that virulence factor. This controversy was generated several years ago for the toxin alpha-hemolysin, expressed by several but not all uropathogenic E. coli isolates. As another example, enteroaggregative E. coli (EAEC) show high genetic heterogenicity. EAEC are diarrheagenic pathogens. A consequence of the EAEC genetic heterogeneity is that many virulence determinants are not present in all strains, and the mechanisms by which EAEC cause disease are still poorly understood¹. In fact, a recent genomic analysis of epidemiological EAEC has shown that the only common virulence traits of all the strains analyzed are the global virulence regulator AggR and an adherence factor².
A common genomic feature of most EAEC strains is the presence of a virulence plasmid termed pAA. pAA plasmids encode the above-mentioned AggR transcriptional activator, which is a member of the AraC-XylS family of transcription factors. The genes subjected to AggR regulation are localized both in the pAA2 plasmid and in the bacterial chromosome. AggR-dependent plasmid-encoded virulence factors include, among others, the aggregative adherence fimbriae (AAF), the Aap dispersin and its type I secretion system (T1SS), and the Shigella flexneri virulence protein VirK. The chromosomal virulence determinants regulated by AggR include a type VI secretion system. In our study³ we aimed to better understand the regulation of the aggR gene. We generated aggR::lacZ gene fusions in different locations along the aggR coding sequence. Unexpectedly, clones that contained the cassette containing the lacZ gene immediately downstream the last aggR codon showed a colony phenotype different to that of the wt strain or to that of other clones containing the lacZ cassette in other locations of the aggR gene. A detailed analysis showed that overexpression of the aggR gene and further expression of virulence determinants such as the AAF fimbriae accounted for the observed phenotype.
Next, we investigated the mechanism underlying overexpression of the aggR gene in clones containing DNA insertions in the sequences located downstream of the aggR ORF. This has led to the identification of a novel mechanism modulating AggR levels. Transcripts initiating at the aggR promoter extend beyond the 3’ end of the aggR gene because of the absence of a conventional Rho-independent transcriptional terminator. Transcriptional readthrough renders long transcripts that are subjected to PNPase mediated processing, thus rendering low expression levels of AggR. Alterations of the 3’UTR region by inserting foreign DNA sequences must alter transcripts structure, making them less prone to RNase processing. The consequence is the overexpression of AggR and hence the increased transcription of the AggR regulon.
Upon coming to that point, we realized that we were able to in vitro mimic the in vivo expression levels of this virulence regulator and decided to compare the transcriptomes of wt cells and of cells having altered aggR 3’UTR sequences. Usually, the analysis of the set of genes targeted by a global regulator is performed by comparing the transcriptomes of the wt strain and of the mutant in the gene coding for the regulator. This analysis had been already performed for the aggR gene, and we were curious to check whether those genes belonging to the AggR regulon that were found to be downregulated in an aggR mutant were upregulated in our mutant that expressed high AggR levels. As expected, this was the case. Nevertheless, most of the genes showing the highest upregulation in our mutant were metabolic genes. This has already been found in other studies under conditions of virulence activation⁴. The large number of identified upregulated metabolic genes was puzzling, but we had the patience of analyzing their function each by each, and soon the puzzle was solved. They corresponded to specific metabolic pathways, namely the fatty acid degradation pathway (including the glyoxylate cycle), the ast metabolic pathway for arginine degradation, and the gamma-aminobutyrate utilization pathway. To support the hypothesis of these metabolic pathways contributing to EAEC virulence, we obtained a double mutant derivative of the EAEC strain 042 exhibiting high AggR expression that lacked the ast and fad pathways. We then compared the virulence features of the parental strain with those of the mutant derivative exhibiting high AggR expression and thereafter with those of the ast fad derivative from the latter. Both in vivo and in vitro experiments showed that, whereas the mutant strain exhibiting high AggR levels exhibited higher virulence than the wt 042 strain, it is derivative that lacks the ast and fad metabolic pathways showed decreased virulence. 
The relationship between metabolism and virulence has been deeply investigated in the last years⁵. Usually, this relationship is established because either the bacterial pathogens adapt their metabolism to better compete for the nutrients existing within the host environment⁶, or because the existing host metabolites may trigger virulence induction⁷.
The results from our work and others⁴ show that induction of specific metabolic pathways can be also a result of virulence induction. Which can be the biological meaning for the pathogen? Utilization of nutrients such as fatty acids or amino acids can be considered as a strategy to utilize additional energy sources during the infective process but, as we discuss in our paper, it may also seek to deplete the host environment from specific molecules, such as short-chain fatty acids or arginine, that can be essential for the epithelial gut homeostasis or to properly trigger the inflammatory response. In this context, these metabolic pathways would really play a role as additional virulence factors, allowing pathogen proliferation and/or persistence. A question to be answered is whether AggR-dependent induction of these metabolic pathways is a general feature of EAEC or, as it happens with other EAEC virulence factors, a feature of some but not all EAEC isolates.
References
1.    Yasir, M. et al. Organization and architecture of AggR-dependent promoters from enteroaggregative Escherichia coli. Mol. Microbiol. 111, 534–551 (2019).
2.    Boisen, N. et al. Redefining enteroaggregative Escherichia coli (Eaec): Genomic characterization of epidemiological eaec strains. PLoS Negl. Trop. Dis. 14, 1–19 (2020).
3.    Prieto, A. et al. Modulation of AggR levels reveals features of virulence regulation in enteroaggregative E. coli. Commun. Biol. 4, 1295 (2021).
4.    Rivera-Chávez, F. & Mekalanos, J. J. Cholera toxin promotes pathogen acquisition of host-derived nutrients. Nature 572, 244–248 (2019).
5.    Passalacqua, K. D., Charbonneau, M.-E. & O’Riordan, M. X. D. Bacterial Metabolism Shapes the Host–Pathogen Interface. Microbiol. Spectr. 4, (2016).
6.    Rohmer, L., Hocquet, D. & Miller, S. I. Are pathogenic bacteria just looking for food? Metabolism and microbial pathogenesis. Trends Microbiol. 19, 341–348 (2011).
7.    Rudra, P. & Boyd, J. M. Metabolic control of virulence factor production in Staphylococcus aureus. Curr. Opin. Microbiol. 55, 81–87 (2020).