Competence in rich medium?

It was pointed out to me that the last post I made shows that competence genes do not get turned on in rich medium for our KW20 samples. For example:

This is strange; expression of competence genes should increase as cell density increases. As for Sxy, I mentioned in the last post that it seems like sxy-1 had much higher expression levels than KW20. It could be the case that in fact the expression seen is sxy-1 is normal and there is an issue with what is believed to be KW20.

 

Looking at Sxy expression in other strains in sBHI, here is what I get:

As expected, sxyx has no Sxy expression. The three KW20 replicates are very reproducible and appear to be similar to the CRP knockout strain for the two CRP samples we have. sxy-1 and rpoD appear to have similar Sxy levels except at the last timepoint where sxy transcript appears to decrease in rpoD and slightly increase in sxy-1. The two murE replicates also appear to differ greatly. One has higher expression than all other strains and the other appears to only modestly turn on sxy, similar to the KW20 samples but at a higher baseline. I can look into this later, but right now the KW20 samples are the most concerning.

Looking at coverage of CRP and adenylate cyclase, there does not seem to be any sign of a deletion in these genes for the KW20 replicates. On Monday, I'll be sure to take a closer look at the raw reads in IGV.

I've been trying a lot of things to figure out what is going on. Consider the KW20 samples in MIV at time t=0. These data points correspond to cells in sBHI at an optical density of about 0.2 - in other words these are cells between timepoints B1 and B2. The KW20 MIV replicates we have appear to behave as expected in terms of competence development so I am pretty confident that this is how cells in sBHI behave.

I made plots like these for all the competence genes and they suggest that the behaviour of the M0 samples is consistent with the sBHI samples. That is, expression levels of competence genes in M0 is in line with the levels seen in sBHI. Sxy levels in M0 also appear to be similar to B1.

Finally, I wanted to check to see if there was any antibiotic resistance genes in the KW20 samples suggesting that there may be a gene knocked out or something. I took all the unaligned/unmapped reads from the KW20 sBHI samples and pooled them together. Then I used a de novo genome assembly tool to assemble the unmapped reads into contigs. Then I looked for long contigs with high coverage that would indicate some genetic element in the sample that could not find a place in the KW20 reference genome. Most of the hits I got were things liks 23S ribosomal RNA from a whole bunch of different bacteria. It may be interesting to see exactly what kinds of contaminants were present.

But I found some reads that when BLASTed aligned to specR cassettes on plasmids. For example:
http://www.ncbi.nlm.nih.gov/nuccore/814617358?report=graph&from=4760&to=4898
http://www.ncbi.nlm.nih.gov/nuccore/29242913?report=graph&from=3730&to=4374

I should mention that the contigs that aligned to these regions had low coverage of about 4 (that is, only 4 reads in all the pooled samples made up that contig). Although there were ~95000 contigs and I only BLASTed about 10, so it's very probable that there may be more pieces of specR cassettes that I am missing. As a positive control, I took a single toxx sample (which has specR) and did this assembly process and found a specR contig with a coverage of about 40 (which is extremely high compared to the other contigs). In comparison, there are much fewer specR reads in the KW20 samples, suggesting that they are from contaminants or they are in a gene that is not as highly expressed as the toxin is in MIV.

In summary, there is some evidence that the KW20 samples may have spec resistance. We have these cells frozen so it is possible to check. It would also be useful to do a competence timecourse with these cells to see if they have a normal competence phenotype.