Okay, not to overwhelm with Streptococcus biology, but I mentioned this new paper in the comments to this post, and had to share a bit of the results, because 1) it's just cool, and 2) it directly stems from some of the research I did for my dissertation. (Always a bonus when someone else can actually *use* stuff you slaved over).
A bit of background: I noted that Streptococcus pyogenes is a tough pathogen to study. Strains of the bacterium may vary in the presence or expression of virulence genes, and even when one insight into the regulation of these genes is uncovered in one strain, another strain may be regulated completely differently. Therefore, it's often difficult to extrapolate results obtained from working on one strain out to the species as a whole. (For example, a locus we found to control expression of the M protein at a transcriptional level was found by others to work at the post-transcriptional level in an isolate of a different serotype.)
However, one phenomenon has been routinely noted. I mentioned here that the M protein of Streptococcus pyogenes is a determinant of serotype. This is a surface-expressed protein, but isolates can vary widely in the amount of M protein present on the bacterial surface. Isolates that express little M protein are therefore difficult to serotype--not enough M protein on the surface to react with the anti-M antibodies. It was discovered early on that one way isolates could be made to express more M protein was to passage them through a mouse: inject them into the animal, and recover them from the blood or organs. (They also could be passaged in vitro through human blood, but this typically did not work as well). It wasn't known exactly why this worked, however.
Fast-forward about 70 years. Work done in my PI's lab further investigated the phenomenon mentioned above. Mice were inoculated either intraperitoneally (IP--injected into the stomach area) or via a mouse skin air sac (MSAS). In the latter, 100 microliters of a bacterial solution was injected just under the skin, along with 900 microliters of air. This created almost a little balloon on the side of the mouse; the bacteria were injected just under the top layer of skin, but still had to cross additional layers of tissue in order to obtain access to the bloodstream and other organs. For some reason, this latter method of injection led to a huge increase in M protein expression after just a single mouse passage, whereas it took about 14 passages IP (and about 9 through human blood) to obtain the same level of expression.
Other phenotypes were affected as well. SpeB, a cysteine protease associated with invasive disease in some studies (and negatively correlated in other studies), was turned off in isolates recovered from the MSAS. In an M1 serotype isolate, a gene called sph (encoding a protein called protein H) was only found to be expressed in isolates which had been mouse-passaged. Another gene encoding a protein involved in immune system evasion (the C5a peptidase, encoded by the scpA gene) was also up-regulated. Overall, it was found that the MSAS-passaged isolates had an expression profile that was associated with invasive disease--and indeed, when these MSAS-passed isolates were injected into naive mice, they produced lesions that were statistically larger than their un-passaged relatives.
Finally, as far as we could tell, this phenomenon was irreversible. Once these isolates were expressing large amounts of M protein, they did so for as many passages in broth or agar culture as we examined them for. They also did not revert to producing the SpeB protease. The change appeared to be permanent. What we never did identify, however, was exactly what was causing this change. We looked at the regulator of the M protein ( a gene called mga, and found that after mouse passage, the M protein could be expressed even if the mga gene was disrupted. We also looked at the regulator of SpeB expression (a gene called rgg), but sequencing of that gene and the region between rgg and speB were identical in parental and MSAS-passaged isolates. We even tried subtractive hybridization, to check for areas that were present in one and absent in the other--no dice.
This brings me to the new paper. In it, Sumby et al. examined the "transcriptomes" (the mRNA produced, essentially) of several pharyngeal ("strep throat") isolates, and others taken from invasive infections. They found that these transcriptomes differed at approximately 10% of the total.
Looking at these further, they found that the transcriptomes from invasive isolates were similar to our MSAS-passaged isolates described above: no (or greatly reduced) SpeB, high M protein, etc., while the pharyngeal isolates were more similar to the parental isolates: producing SpeB, less M protein. (They differed in expression of several other putative virulence factors as well). When they sequenced these, they found a frameshift mutation in another regulatory gene, named covS.
Was this enough to change expression of all these genes? It was already known that covS played a central role in gene regulation in S. pyogenes, so it seemed plausible. To check, they complemented the bacteria with the frameshifted covS with a plasmid carrying a normal gene--and found that indeed, it made the invasive isolates "revert" to a pharyngeal-type transcriptome. So it appears that this deletion is enough to cause the phenotype we'd seen. Being a deletion, it also explains why it's irreversible.
I mentioned previously that both the manifestations of streptococcal disease, as well as the serotypes of S. pyogenes present in the population have changed over the years. While the incidence of rheumatic fever has decreased in developed countries, the incidence of other severe disease caused by S. pyogenes--such as necrotizing fasciitis and streptococcal toxic shock-like syndrome--has increased. (Indeed, as I was drafting this, an episode of Grey's Anatomy featured a case of necrotizing fasciitis--resulting in debridement of much of the patient's leg tissue). Is this increase in severe disease manifestations due to something as seemingly minor as a 7-base pair mutation in a regulatory gene? Are isolates with this deletion becoming more common? There are a number of archived isolates which can be examined for this, so expect to hear more about it in the future.