Wednesday, May 6, 2015

Sxy-regulated genes

I've been working to identify all Sxy-regulated genes using the following procedure:

Briefly, I found genes downregulated in sxyx relative to KW20 and genes upregulated in sxy1 relative to KW20. I took the intersection of these two sets to be Sxy-dependent genes. In total, 36 genes were identified; a list of their IDs can be found here.

Prior to this, 26 competence genes were thought to exist. This list contains 25 of those. The single stranded binding protein (ssb) is missing from the list. This is not surprising because this gene is essential (knocking it out leads to a non-viable strain) and is only weakly regulated by Sxy.

Unsuprisingly, sxy appears on the list:

Virtually no reads align to the sxy gene in the sxyx samples as expected. Interestingly, it appears the sxy-1 strain has higher levels of sxy transcript. Not entirely sure why; is Sxy self-regulating?

As for the 25 known Sxy-dependent gene (Note that MIV data is being shown on the left and sBHI data on the right):
(slightly higher resolution versions: here and here)

Ten other genes were identified. I'll first go though the ones that appear to be false positives.

Likely false positives

HI0609.2 tRNA-Arg:

This gene doesn't appear to behave similarly to any of the competence genes. It looks like the only reason it was flagged was because it's significantly higher than sxyx in MIV at M3 (probably due to the one sample that is an outlier) and slightly (but significatly) higher in sxy1 at the last sBHI timepoint.

HI1281 pseudogene:

The expression pattern is actually very similar to the tRNA above, so for the same reasons, I'll consider this another false positive.

HI1572 pseudogene:

It appears that one of the KW20 replicates behaves oddly in MIV for this gene but other than that, the expression appears to be mostly independent of Sxy.

More interesting results  

HI0984 dprB

This gene is downstream of dprA and both are likely part of the dprABC operon. dprA is a known competence gene and the graph above suggests that dprB may also be Sxy-dependent. This paper claims that dprABC are all coordinately expressed and competence inducible but our microarray paper claims that dprBC showed little induction by Sxy and were not coordinately expressed with dprA. So I looked closely at the actual RNA-seq reads:

 It appears that dprA is the only gene in the operon that is Sxy-dependent and dprBC appear to be expressed independent of Sxy. The reason dprB was flagged by this analysis is likely because some of the reads from dprA probably overlapped and "contaminated" drpB (i.e. the dprB read counts were inflated when dprA was expressed).

HI0654, HI0656, HI0657, HI0658

This appears to be an operon adjacent to HI0659-HI0660 and moderately regulated by Sxy. The microarray paper also identified this operon, but noted that these genes lacked a CRP-S site and could not identify a terminator between HI0659 and HI0658, suggesting that this may be due to readthrough. It appears to be a good explanation but it comes short of showing that this operon doesn't have a competence-specific role. I would like to see if this these two operons (HI0658 and the toxin-antitoxin operons) were always adjacent in species that have both. I believe Hailey has looked into this, so I'll look into what she has found later.


I think this one is really interesting and will require a much more in-depth look. There appears to be a very large difference between sxy1 and KW20 for this operon. As for the sxy knockout, there doesn't appear to be such a dramatic difference. However, the sxyx graphs above are for cells in MIV. In for cells in sBHI it looks more like this:

This suggests that this operon is upregulated in the sxy-1 mutant and downregulated in the sxy knockout (compared to KW20) only in sBHI. I need to look into this one a little more, maybe it will get its own blogpost.

HI0352 lic3A

This one is really weird. In sxy-1, it is clearly expressed at a lower level. No difference can be seen between wild type and sxyx though (though there does appear to be a replicate that is an outlier). Again, this is sxyx in MIV though. Looking at cells in sBHI:

In sBHI, it appears that any deviation from wildtype sxy causes this gene to be downregulated. This gene encodes an alpha-2,3-sialyltransferase which adds N-acetylneuraminic acid (Neu5Ac) to a lactose moiety on the LPS. More information can be found in this paper. I think this gene also deserves its own blog post.


The RNA-seq data verifies 25 of the known 26 competence genes with this analysis. dprB and an operon next to HI0659 may be Sxy-regulated, but I would tend to agree with the conclusion from the microarray paper that this is caused by readthrough (and dprA reads leaking into dprB). The genes
HI1456, HI1457 and HI0352 appear to behave very differently when Sxy protein levels are high/low - but this appears to be limited to cells growing in sBHI. These genes demand further investigation.


Rosie Redfield said...

Odd that the baseline expression of dprB is so much higher than that of dprA.

andrea wang said...

The RNA-seq data verifies 25 of the known 26 competence genes with this analysis.

Candy Swift said...

Can you provide analyzing tools for denovo sequencing?