murE mutagenesis planning

 (Edited on March 1, after more thinking and planning.)

I want to create a pool of cells with random point mutations in the H. influenzae murE gene, and to select and screen this pool of cells for hypercompetent mutants.  I'm going to do this by mutagenizing the DNA with the chemical mutagen ethyl methanesulfonate (EMS) in vitro and then transforming it into cells, rather than mutagenizing cells.

One unanticipated benefit of the in vitro method is that the mutation spectrum is better.  With in vivo mutagenesis, EMS produces mainly  GC-to-AT transition mutations by alkylating guanines in DNA, creating O-6 ethylguanine which mispairs with T instead of C during DNA replication. (info from Wikipedia). But the in vitro work found a much less biased distribution, with 42% GC-to-AT transitions, 34% AT-to-GC transitions, and 24% GC-toCG transversions.


Plan:

Step 1. Cut chromosomal DNA of strain RR797 RR805 with the restriction enzymes KpnI and BglII. This strain contains the wild type allele of murE (oops, no, this strain has the murE749 hypercompetence allele! The strain I want is RR805.), and has a chloramphenicol resistance cassette inserted about 5 kb away from the site of the known murE mutations,. This digest creates an 8 kb fragment that contains both the CAT cassette and the wild type murE allele.

Here's the map:

This pre-digestion step could probably be omitted if necessary, because random fragmentation of the DNA will accomplish almost as much. But it shouldn't hurt, and it might double the frequency of cotransformation.   But I just looked at some old cotransformation data, and I see 60-70% linkage (selecting for CmR gives the linked murE allele), which is very good

Step 2. Soak this DNA in an EMS solution for 1 hr.

Step 3. Wash the DNA and transform it into competent wildtype cells.  Use about 100 ng DNA per ml, so that each cell is likely to recombine only a single DNA fragment.  As a control, transform the same cells with DNA from the chloramphenicol-resistant murE749 strain.

Step 4. Select for chloramphenicol resistance, to enrich for cells that have recombined in murE.  This will also confirm that the level of DNA damage was not so high as to limit transformation.  I should be able to get many thousands of independent transformants.

Step 5. Pool chloramphenicol resistant colonies, creating separate pools from independent sets of transformants.  Aim for about 5 pools.  Freeze some of the cells of each pool.  Make a pool for the control transformants too.

How many colonies should be in each pool? I want enough colonies per pool that each is likely to contain at least one hypercompetent mutant - how many colonies will this be?  I know of three mutations that produce hypercompetence, which would let me predict the minimum expected frequency of hypercompetent colonies if I knew the frequency of mutations in the DNA and the degree of linkage in the transformation.  I can measure linkage by doing colony assays on the control transformation.  The enrichment can increase the frequency of hypercompetence by 1000-fold, if all the mutants are as hypercompetent as the ones we have.  So if the frequency of hypercompetence in the chloramphenicol-resistant transformants is 1/1000, I should put at least 1000 colonies in each pool.  If it's less, I should put more. 

Step 6. Grow the pooled cells in sBHI at low density for a few hours, then transform with cloned or PCR'd NovR DNA (or a different marker?).  Plate on nov plates.  Do this with the control murE749 transfornation too.

Step 7. Screen individual NovR colonies for hypercompetence by touching them to nov plates and then resuspending the rest of the cells in sBHI containing MAP7 DNA and plating on Kan (or Nov?) plates.  Do only 10 colonies per pool, or 1/1000 as many colonies as went into the pool?  I expect most of the control colonies to be hypercompetent.

Step 8. For each pool, pick one or two high-transformation colonies from their toothpicked plate, and retest their competence with a simple time course.

Step 9.  PCR and sequence the murE genes from5 or 10 of the confirmed hypercompetent mutants (depending on how many I get, of course).  Are the known mutations present?  New mutations?


First we should test different levels of mutagenesis:  

The protocol we have (Lai et al.) says to use 1 µg DNA in 20 µl 10 mM EMS for 1 hr; this gave 5-6 mutations per kb in the clones they sequenced.  It also reduced the transformation efficiency of the plasmid insert they mutagenized to about 60%.  If they carefully standardized the amounts of DNA, this reduction should have been a direct consequence of DNA damage and repair processes, since they were not selecting for function of their mutagenized insert.

5-6 mutations per kb sounds pretty good for us (but see next post, which suggests we want fewer), since about half of them will be silent, but I think we should first try a wide range of concentrations.  For the cell mutagenesis (many years ago) I used 50 mM for 45 min and 80 mM for 30 min (RR expt # 181), but we want much heavier mutagenesis here.  So here let's try 0, 2, 5, 10, 20, 50, and 100 mM - that's 7 DNA samples to do transformations with.

Two assays for the extent of mutagenesis:  

1. (To identify an optimal concentration) Mutations creating low-level resistance to novobiocin: Mutagenize any novS DNA (e.g. RR805) and transform into KW20 and select for low-level novobiocin resistance (1 µg/ml rather than 2.5), to check the efficacy of the mutagenesis.  There should be an optimal dose of EMS, above which the frequency of nov resistance drops because the DNA is too damaged to recombine or contains too many mutations that block gene function.

2. (To identify concentrations that are too high) Mutations that inactivate the CAT cassette:  Mutagenize RR805 DNA and transform KW20 to chloramphenicol resistance. At some EMS dose the transformation frequency will decrease because the DNA is too damaged to recombine or contains too many mutations that block gene function.  (This test could also be done with any point mutation creating antibiotic resistance.)

What we know about the competence-regulon gene comM

The grad student of an upstairs colleague has been doing a lot of excellent work on the Rhodobacter capsulatus homologs of some H. influenzae competence genes, because he has discovered that they are also needed for gene transfer by GTA, the phage-related 'gene transfer agent'.

One of the genes he's looking at is comM.  ComM is predicted to be a cytoplasmic protein, a member of the YifB subfamily of AAA-ATPase proteins.  Here's a review about the AAA+ superfamily.  These proteins have a very diverse range of activities, so it's hard to make any prediction about a likely function for ComM from looking at its relatives.

ComM was originally studied in H. influenzae, by Michelle Gwinn and Jean-Francois Tomb in Ham Smith's lab.  They reported that their comM mutant had normal DNA uptake but reduced transformation (down about 300-fold).  It had normal expression of a lacZ fusion to another competence gene, indicating that it didn't affect regulation of competence.  (It also had reduced phage recombination, but we still don't know what this assay means.)

To find out why transformation was reduced, they followed the fate of end-labelled DNA fragments. The kinetics were like those of both wildtype cells and a rec1 mutant (rec1 is the H. influenzae homolog of E. coli's recA; it's absolutely needed for homologous recombination). So the authors concluded that the comM knockout does not affect the transport of DNA into the cytoplasm.  But their data doesn't distinguish between an effect on DNA degradation (indirectly preventing recombination) and a direct effect on recombination.

We've independently created a comM knockout; its DNA uptake is also normal, and its transformation is also down, but only about 20-fold.  We haven't done anything more to evaluate its phenotype.



*Interestingly, Gwinn et al. commented that "In addition, HI1117 has homology to a magnesium chelatase gene of Rhodobacter capsulatus, bchI, involved in bacteriochlorophyll biosynthesis (1) and to related genes from other photosynthetic organisms."

Grant proposal's done! What experiment shall I do?

I clicked 'Submit' on my grant proposal last night; my immediate teaching responsibilities are light, and there's nothing else big on my plate, so now I get to start doing experiments again!

I think the most fun thing to do will be to join the sabbatical visitor and the co-op tech in doing mutant hunts for hypercompetent strains.  They're mutagenizing the rpoD gene and screening for new mutations that cause hypercompetence, and I can use the same methods on the murE gene.

This old post describes what we know about the relationship between murE and competence.  Well, what we used to know, because now we have new RNA-seq data that will tell us how transcription changes.  Basically, we have four independent mutants that all cause very similar extreme-hypercompetence phenotypes.  murE749 is the main one we've studied.  Some lacZ-fusion analyses indicate that it acts by causing overexpression of genes in the competence regulon (we looked at two genes) and one low-quality microarray appeared to confirm this and (maybe?) show some overexpression of sxy (the regulatory protein that controls expression of the competence regulon).

We assume that the other murE mutations act the same way.  But we have absolutely no idea how the mutations cause the phenotype.  MurE is an essential cytoplasmic enzyme in the pathway that synthesizes the cell wall.  The mutants all grow normally (though we haven't done a BioScreen run), and are not unusually sensitive to any simple test of cell-wall function.

One big part of the puzzle is how the mutations change the protein. The diagram above shows that three of the mutations change a poorly conserved amino acid (at position 435); these changes wouldn't be expected to have any serious impact on the enzyme's catalytic function.  So how do they have such a big impact on cometence?

On the other hand, the mutation in murE751 changes the strongly conserved leucine at position 361 to a very different amino acid (serine).  Leucine is hydrophobic but serine is polar, so they make very different interactions with their surroundings.  Because this leucine is highly conserved we think it must play an important role in the enzyme's catalytic function.  This would explain how the mutation can have a big effect on competence, but leaves us instead wondering why it doesn't have a big effect on cell growth.

I need to do several things:

1. Update my reading to find out what's been learned about MurE function since we published our paper way back in 2000.
2. Dig into the new RNA-seq data to see what it tells us about RNA changes in the murE749 mutant.  This will require finally learning some R and/or getting help from other lab members.
3. Isolate new murE mutations that also cause hypercompetence.

Lots of fun!

Yes, I'm still here

The last-chance-for-everyone CIHR grant proposal deadline is Friday at 8:30 am!  After that's in, I promise to get back to the bench and back to proper research blogging.