11 March, 2009

Not-So-Mad Science: Shiga Toxin Increases EHEC Intestinal Colonization

This post is part of Cyber Journal Club with Science Bear and R.E.S.E.A.R.C.H.E.R.S..

Figure A: EHEC + other stuff on sorbitol MacConkey Agar. Most EHEC cannot ferment sorbitol and on SMAC agar it will grow translucent. Other stuff pictured here can ferment sorbitol and as such is pink because it has raised the local pH and tripped the pH indicator in MacConkey plates.

(For background, I've discussed Shiga toxin and EHEC in greater detail before.)

EHEC = enterohemorrhagic Escherichia coli O157:H7
Stx = Shiga toxin

ResearchBlogging.orgBriefly, EHEC is a food-borne enteric pathogen that has been evolving over the past 2 decades from a minor poop-tainted beef-borne pathogen to a more severe and flexible pathogen that is still found in beef, but now also on spinach (even growing inside the spinach). Primary sequalae of EHEC infection are intestinal cramping and diarrhea progressing to hemorrhagic colitis. This may usually be resolved with careful and selective use of antibiotics (antibiotics that cause any DNA stress will just make it worse). However, in some cases EHEC infection progresses to hemolytic uremic syndrome (HUS), which includes hemolytic anemia (lysed RBCs), uremia (acute renal failure), and thrombocytopenia (low platelet count). HUS can involve microangiopathic inflammation of the kidneys, bowels, and CNS, which can result in long-term dysfunction.

Epidemiologically, HUS is closely associated with infection by Stx-producing E. colis. As such, and because Stx is directly cytotoxic to in vitro renal cell cultures, Stx has been implicated as a primary causative agent of HUS. Stx is a AB5 N-glycosidase that shuts down protein synthesis by binding to and disabling ribosomes. Stx is encoded by a lambdoid prophage integrated into the EHEC chromosome that is inducible by DNA stress, specifically RecA in the SOS response. RecA cleaves the cI/LexA repressor on the lambdoid genes and allows transcription. Shiga toxin is thought to be released primarily when the induced host cell is lysed by prophage proteins. It has been observed that Stx+ EHEC results in more severe pathologies than Stx- EHEC.

Robinson et al (below) has found that EHEC adherence (in vitro) and colonization (in vivo) are increased in the prescence of Stx2. They used the Relatively Straightforward Common Biotech Method:
Step 1: Break it and compare that to that that ain't broke, then supplement the broken stuff to the level of that which ain't broke and compare again. If breaking it resulted in nothing, or less, happening, then what you broke might be causing that effect to happen in the first place and you may develop sweaty palms.
Step 2: If supplementing what's broken to non-broken levels restores the effect, then what you broke is probably causing that effect and you may start fidgeting uncontrollable.
Step 3: But, if you've also broken other related stuff and not seen the same pattern, then you can say that what you broke is causing it to happen (to the best of our current knowledge!) and you may then do a happy marshmallow dance of victory.
So Robinson et al took a normal Stx+ EHEC (strain 86-24) and made a mutant Stx- EHEC (strain TUV86-2, checked culture supernatant for cytotoxin activity: negative), then made them both into glowy mutants by adding constitutive GFP (this was smart for reasons we'll come to shortly). Then they grew up some HEp-2 cells in culture and let both strains loose on it in separate wells, which resulted in more 86-24 sticking than TUV86-2 (step 1). This same pattern was also observed in mice infected with both strains: 86-24 colonized the gut better than TUV86-2. They also checked microcolony formation and the degree of actin condensation (both characteristics of EHEC adherence), which was important because it demonstrates that TUV86-2 wasn't fundamentally different than 86-24 and so could be compared. So naturally, Robinson et al's next logical step was to add some Stx2 to the TUV86-2 to see if it could restore 86-24 adherence and colonization. It worked (step 2). Interestingly, in TUV86-2 + Stx2 groups, the proportion of HEp-2 cells to which the bacteria adhered remained the same, but the bacteria/cell increased (not significantly). And to prove that step 2 worked, they also repeated this with a neutralizing anti-Stx2 mAb and compared that to normal mouse serum. The TUV86-2 + Stx2 + anti-Stx2 mAb had similar adherence and colonization patterns to TUV86-2 alone. The normal mouse serum had no effect on the effect of Stx2.

So what was the Stx2 doing in the lumen that was causing increased adhesion? They looked at bacterial adhesion factors (Tir, intimin, et al) after incubation with Stx2 and found no change (however, they used Western blot for this; I'd have accepted this more readily if they'd also done a bacterial mRNA profile as it is possible that the incubation time was not long enough for a signal change in transcription to become evident at the level of proteins; also: step 3). So they looked at putative adhesion factors on the host HEp-2 cells, namely: nucleolin. Incubation of HEp-2 cells with Stx2 resulted in a dose- (and, kinda, time-) dependent increase in the cell surface expression of nucleolin, which was statistically significant even with just 10ng/ml Stx2 for 2h (for contrast, 100ng/ml Stx2 for 2h resulted in a significant increase of nucleolin). However, I would have again have preferred to see some mRNA analysis of the host cells. Not because I suspect there could be a significant effect from an incubation time artifact (at least, not after 24h), but because the data would have been very interesting. Is Stx2 increasing transcription of nucleolin or just an increase in the surface display thereof? mRNA data could have answered that. Also, why not wash the crap out of the cells to remove any residual media-borne Stx2 and check to see if there is any Stx2 inside the cells (either with a GFP-tagged Stx2 or ELISA on cell lysate)? Stx2 has been shown to enter gastrointestinal epithelial cells (in culture) nonspecifically through macropinocytosis, but is it doing something specific once its in there, if it's even getting in (because the macropinocytosis has only been observed to happen in some specific cell culture lines)? Is Stx2 reacting with a surface HEp-2 receptor or is it some internal effect? Is Stx2 binding nucleolin itself and causing autoaggregation at the surface (FRET!)?

All in all, an interesting paper, but for me it raised many more questions than it answered.
Robinson, C. (2006). Shiga toxin of enterohemorrhagic Escherichia coli type O157:H7 promotes intestinal colonization Proceedings of the National Academy of Sciences, 103 (25), 9667-9672 DOI: 10.1073/pnas.0602359103


Ambivalent Academic said...

HAHAHAHA - I like your methodology so much that I would like to share it with readers at my place.

Would you mind if I reposted that bit with proper citation?

Toaster Sunshine said...

I wouldn't mind at all. So by all means, go ahead!

drdrA said...

Toaster- I respectfully submit that looking at mRNA levels isn't very useful in this case. After all adhesion happens at the protein level. So, whether or not there is more/ less transcript for a particular adhesin at a given time, if there is the same amount of that particular adhesin on the surface of the cell... seems pretty uninformative.

I should probably read the paper!!

Toaster Sunshine said...

Stx2 inhibits protein synthesis by acting as an N-glycosidase on the ribosomes. Therefore, if Stx2 is having an effect inside the HEp-2 cells, one would think that the mRNA->protein turnover would be affected and there would be a large discrepancy between nucleolin mRNA and nucleolin protein. However, if Stx2 were effecting an increase in nucleolin from outside the cell, then one wouldn't see so much of a discrepancy. It's not so much that I completely distrust their quantitation of nucleolin, I just want to know where the effect of Stx2 is happening.

Your point is quite valid though and what I'm asking for is probably beyond the scope of that particular paper.

drdrA said...

I should have made clear that I was primarily referring to Intimin and Tir levels- these are bacterial proteins and their production shouldn't be affected by STX. Seeing mRNA data for those isn't really useful if you already have protein levels.

For Nucleolin- yes, that's of course interesting.