What's known about T4P and DNA uptake in Neisseria

I've been rereading an important 2002 paper by Finn Erik Aas et al. (Mike Toomey's group; Molecular Microbiology 46:749-760).  They dissect the roles of the Neisseria type IV pilus components PilE (the major pilin), ComP (a minor pilin important for DNA uptake), and PilT (the retraction ATPase).  The results are important and need to be considered in the CIHR grant proposal we're rewriting. 

I was about to try to summarize them here, when I wondered whether I might have already posted anything about this paper.  A quick search for 'Aas' found a whole post from last summer, which I had completely forgotten about!  So here I'm instead going to edit/annotate/rewrite that post.

An important difference between this paper and previous papers is their careful attempts to separate of DNA binding from DNA uptake.  DNA binding is usually measured indirectly  by giving cells radioactively labeled DNA and comparing cell-associated cpm with and without pretreatment with DNase I, which removes DNA that has bound to cells but not been taken inside them.  And that's how it's done in this paper - the 'binding' values are how much of the total cpm is removed by DNase I treatment (i.e. they're determined by subtraction).

The Aas et al. paper  showed that, when 5x10^8 wildtype Neisseria cells are incubated for 30 min with 500 ng of 32P-labeled plasmid DNA containing the Neisseria uptake sequence (the DUS), only about 1-2% of the DNA sticks to the cells (i.e. is still there after 10 min on ice and three washes with cold medium).  They found that more than half of this DNA is taken up during the 30 min incubation, because it wasn't removed when the cells are incubated with DNase for 10 min at room temperature before being washed instead of being stored on ice.  If the DNA doesn't contain a DUS the cells still bind a lot of it (25-100% of the +DUS binding amount), but almost none of the bound DNA is taken up (> 0.01% of the 500 ng).  This says that cells will bind any DNA but can only take it up if it has a DUS.

The minor pilin-like protein encoded by comP is normally expressed at a very low level (so low that it's barely detectable by protein assays), but the level is high enough to show that the ComP protein is incorporated into pili along with the major pilus protein PilE.  Cells lacking PilE don't have visible pili and they didn't bind DNA or take up DNA, but cells lacking ComP have normal-looking visible pili and they bound just about as much DNA as wildtype cells.  But they hardly took up any of this DNA even if it had a DUS (they interacted with DUS+ DNA the way wildtype cells interact with DUS- DNA).

This strongly suggests that the ComP protein recognizes the DUS at the cell surface, but that ComP is not involved in the non-specific DNA binding step.   Consistent with this, overexpressing the normally-scarce ComP protein increased by 20-fold the amount of DNA bound and taken up, and proportionately increased the transformation frequency. This increased uptake was specific for DNA containing a DUS, although a modest increase was also seen for DNA that lacked DUS.  Overexpression also greatly increased the amount of ComP protein in the pili.  I think this result says that ComP can cause DUS-specific DNA binding, on top of the normal non-specific binding.  ComP, when assembled into pili, would be able to bind DUS-containing DNA but not other DNA.

However, the authors found that, when they purified ComP, it did not bind specifically to any DNA.  They tested both 'recombinant' ComP (+ His-tagged) and 'overexpressed' ComP, but they they don't show this data and they don't say how the binding assays were done.  Does this result mean that ComP really doesn't interact with the DUS?  Maybe.  The authors suggest that it acts by modifying some other protein so that it binds the DUS, but another explanation might be that ComP loses its DNA-recognizing function when it's purified away from the T4P complex.

My tentative model:  ComP is processed by the  prepilin peptidase as is the main pilin (PilE), and is assembled along with PilE into the pilus filament, where it forms a very minor component.  Once DNA binds non-specifically to the PilE part of the pilus, it can interact with ComP and, if the DNA has a DUS, be taken up when the pilus is retracted.  If the amount of ComP in the pilus exceeds the non-specific binding capacity of the PilE, DNA will bind directly to ComP (even if it can't be taken up, see below).

What about PilT?  In most bacteria with T4P, PilT provides the power to retract type 4 pili filaments, by disassembling the subunits from the pilus base.  (The exceptions are H. influenzae and its relatives in the family Pasteurellaceae, who don't have a pilT homolog.)  As pili are thought to bind DNA at the cell surface, PilT would then be responsible for pulling the pilus and its DNA into the periplasm.  PilT mutants had already been shown to have the expected phenotype: abnormally large amounts of pili, and unable to take up DNA or be transformed.  We might then expect that having more pili would make pilT mutants bind more DNA than wildtype cells, but Aas et al. found that they bound <1% of the wildtype amount.  They took up about 20% of what they bound, but I think this may be at the detection threshold. 

Overexpressing ComP in a pilT knockout increased the DNA binding (but not uptake) by about 5-fold.  This added binding was DUS-dependent.  I could add this to my tentative model:  The nonspecific binding by PilE requires PilT, but we have no idea why (I'll call this the magic PilT effect).  Specific binding by ComP does not.

They then made transcriptional fusions of pilE and pilT to the E. coli lac promoter (in different strains) so they could keep the genes turned off or turn them on by adding IPTG, as desired, and measured transformation.  When each gene was on throughout the transformation experiment, transformation was close to normal, as expected.  When it was off, there was no transformation, again as expected.  They then tried turning the genes on halfway through the transformation experiment, to find out whether PilT was only needed to pull the pili in

When pilE was off during the DNA incubation step and then turned on after the cells had been washed free of unbound DNA, transformation was down >500-fold.  This is consistent with both nonspecific and specific binding happening only in assembled pili - it's no use making the pili after the DNA has all been washed away.  When pilT was off during the DNA incubation step and then turned on after the cells had been washed free of unbound DNA, transformation was down  >32-fold; this says that some DNA can bind to cells in the absence of PilT, and then be later taken up once PilT becomes available.

Does this mean that the binding that happens before turn-on of PilT is specific for the DUS?

The authors then did the same experiment in cells overexpressing ComP.  When there are no pili (pilE off) my model predicts that no DNA will bind even though there's tons of ComP, so turning pilE back on after the cells are washed shouldn't give any transformants.  In fact it gives some, but not a lot.  When there are pili but no PilT, my model predicts that having tons of ComP will increase DNA binding, and that this will increase transformation frequencies once PilT becomes available.  My model predicts that this binding will be DUS-dependent, but that isn't shown.

Summary of the PilT bit:  Having pilT off while DNA is binding to the pili reduces transformation frequencies 32-fold when most binding is initially nonspecific, because most of this binding is caused by the magic PilT effect, but only 7-fold when more of the binding is by ComP to the DUS, because the magic PilT effect doesn't act on ComP binding to the DUS. 

But uptake requires pilus retraction.  So maybe I should conclude that, in wildtype cells, most DNA initially binds non-specifically to the many pili that have no ComP but have been magicked by PilT.  DNA bound to these pili is not released to the solution, and these pili are not retracted.  But if a DNA fragment has a DUS, the DUS can bind to one of the fewer pili (or pseudopili?) that have ComP (binding to directly to ComP or to another protein).  Once a ComP-containing pilus has bound DNA it can be retracted, pulling the DNA in and initiating uptake.  This pilus might then elongate and bind another DUS, or other ComP-containing pili might be continually assembled and accept DNAs from the PilE pili. 

I'm going to stop here, because the more I read the confuseder I get.  I'll talk this all over with the RA tomorrow morning and see if she can sort me out.

Role of competence in biofilm attachment (part 2)

Yesterday I talked to my colleague down the hall about a possible joint project.  She reminded me that one of her grad students has done quite a bit of work on DNA in Campylobacter jejuni biofilms; the student has a mutant that makes faster/thicker biofilms, and whose biofilms have much more DNA than is usual.  The biofilms also fall apart in the presence of DNase I.  She gave me some papers to read, one by the student and two others about DNA in Pseudomonas biofilms.

Despite quite a few publications about DNA in biofilms, none consider the role of competence.  So I still like the idea of testing whether biofilm formation by C. jejuni and by H. influenzae is enhanced by the ability to bind (and take up?) DNA.

I also did some reading about Campylobacter competence.  There aren't many papers, but there's some nice work characterizing genes needed for DNA uptake, and characterizing how competence varies under different culture conditions.  We'd need to send for a suitable mutant, one that was unable to take up DNA but had no growth defect.  Ideally we'd want to know that the cells couldn't bind DNA, or couldn't turn on the competence genes, but I'll have to do more reading to see if such a mutant is known.

All the H. influenzae biofilm work appears to have been done with 'nontypable' strains (clinical strains lacking a capsule, which are important causes of ear infections), rather than with the Rd strain we normally use.  I was hoping that these experiments might have used Rd as a control, but apparently not.  I'd much prefer to do these experiments with Rd, but we'll have to check whether it readily forms biofilms in culture.  If not, we might be able to use one of the nontypable strains that we've found to be readily transformable, but we'd need to introduce a hypercompetence mutation into it.

The minimal experiment would use glass culture tubes, some of which had been precoated with DNA.  Some DNA sticks to plain glass, although I can't find out how much.  A lot more sticks in the presence of saturating sodium iodide but that's not an option for living cells.  So we could try just filling tubes with moderately concentrated solutions of H. influenzae or C. jejuni DNA, maybe 10 µg/ml, leaving them for a few hours, and then letting the DNA dry on overnight.  (There's evidence that C. jejuni, like H. influenzae, prefers to take up its own DNA, although there's no obvious uptake sequence in its genome.)

We'd probably be wise to rinse the tubes with culture medium in the morning, to remove DNA that wasn't stuck to the glass surface.  Then we'd add dilute cultures of hypercompetent (sxy*) or sxy- H. influenzae, or wildtype or noncompetent C. jejuni, and let the cultures incubate overnight at 37°C.  We would also include cultures with added DNase I as another negative control.  After overnight culture (or shorter), we'd wash out the non-attached cells, stain the biofilm with crystal violet, and measure the staining.

Another experiment would measure whether being competent helps cells attach to an existing biofilm, presumably by attaching to the DNA.  For this experiment we'd grow biofilms in the ordinary way, using whatever strain makes good biofilms.  Then we'd add antibiotic resistant H. influenzae or C. jejuni cells (competent and non-competent), and give them time to attach.  Then we'd wash away all the unattached cells, break up the biofilm by adding DNase (or ???) and plate on antibiotic plates to count the newly-attached cells.  To distinguish binding to DNA from binding to other biofilm components, we could swamp the culture with free DNA - that should abolish DNA-specific attachment.  (This control could also be done for the minimal experiment.)

DNA mediated attachment in biofilms

One questioner after my CIfAR talk asked whether competence genes might have some other function for the cell.  When I raised this at lab meeting yesterday, the post-doc pointed out that competent cells are able to attach to and pull on the DNA that's a major component of biofilms, and we considered whether this might help cells to establish biofilms, to move within biofilms, or to avoid being displaced from biofilms.

We then started considering experiments we might do.  My idea was to precoat glass surfaces with DNA, incubate them with cultures of cells that can form biofilms, and then measure whether the DNA increased the initial steps of biofilm formation.  The lab down the hall studies, among other things, biofilm formation by Campylobacter, which is naturally competent, so we might work with them to test this in both H. influenzae and Campylobacter. 

What controls would we want?  We'd certainly need to test cells that can and cannot express their competence genes.  Competent cells express type IV pili, and these might allow attachment and pulling. independent of their involvement in DNA uptake, but I think doing the experiment with and without DNaseI would control for DNA-independent effects of pili.

The basic assay for biofilm formation is very simple, and the Research Associate's extensive experience turns out to include doing lots of these.  I think it's time to talk to my colleague down the hall, to see if she's interested.

Variation in core genes

I ended my talk at the CIfAR Integrated Microbial Diversity meeting by asking how much variation we should expect in 'core' bacterial genes within a single species, and I considered this more at lab meeting yesterday.  We see a lot of variation in competence phenotypes, due, we assume, to variation in the corresponding genes.  But we don't know if this variation is telling us something particular about how selection acts on DNA uptake and transformation, or is just 'normal'.

The variation could accumulate because loss of competence is sometimes a good thing, at least in the short term, but it could also be accumulate just because mildly deleterious changes are only slowly eliminated by selection.  This is especially likely to be true for genes that are only expressed occasionally.

So the variation in competence might mean that cells only need DNA uptake occasionally.  It would really help if we know more about when cells develop competence int heir natural environment...

Planning my CIfAR talk

OK, the CIHR proposal is presentable and has been sent to the internal reviewers, so now it's time to start thinking seriously about the talk I'll give at the CIfAR Integrated Microbial Diversity meeting on Sunday morning.  The talk's title is "Sporadic Sex? Scary Food?  The biodiversity of bacterial competence", but I'm going to also talk more generally about the genetic diversity of bacteria, especially the diversity within a single species.

I've started making an outline:

1. Genetic diversity in eukaryote species:  Not a lot.  The members of a species typically have very similar chromosomes, with the same genes arranged in the same order.  Different versions of the genes differ by no more than a few %, and there are very few differences in gene content or arrangement.
2. We take this for granted, but it's probably one of the many consequences of sexual reproduction.  Individuals with chromosomes whose genes have been rearranged (by, for example, an inversion) are at a big disadvantage in sexually reproducing species, because such chromosomes don't pair properly with normal chromosomes at meiosis and the resulting gametes often have two versions of some genes and none of others.
3. What about in bacterial species?  I should probably start by considering what we want 'species' to mean in Bacteria.
4. For most eukaryotes, the 'biological species concept' is the most useful way to group individuals into species.  In its simplest form, it says that all individuals that can reproduce sexually with each other should be considered members of the same species.
5. But bacteria don't have eukaryote-type (meiotic) sexual reproduction.  They don't have any process that regularly combines the sets of genes from two individuals and then shuffles them into two 'recombined' sets.  Maybe I'll talk here about what bacteria do have that puts genes into new combinations, or maybe I'll leave this till later.
6. But we can define bacterial species as groups of individuals for which at least 50% (75%?  90%?) of their genes are recently descended from a common ancestor (recent enough that sequence divergence is no more than a few %), and for which most of these genes are syntenic (arranged in the same order).  This definition works for most of the groupings that microbiologists consider for other reasons to deserve being called species.
7. For this 50% or 75% or 90% of the genome, the members of a bacterial species are a lot like a eukaryotic species.
8. But these species are far more genetically diverse than are eukaryote species, because of the other 10% or 25% or 50%.  Many of these sequences are stupendously diverse...

That's as far as I've gotten.  I mustn't forget that I only have 25 minutes to talk, plus 5 for questions...

Improving the CIHR proposal's focus

I've been thinking more about one reviewer's concern that our CIHR proposal was 'unfocused'.  Although part of the problem was that we were proposing to do too much, I think the underlying problem was indeed a genuine lack of focus.

We identified three major gaps in the understanding of DNA uptake by gram negative bacteria, and proposed a series of experiments that would help fill in these gaps.  BUT, we didn't tie these experiments together very well, and we weren't proposing to completely fill any of these gaps.  They were all parts of a big problem, but they weren't the whole solution (that will likely take many years of work by more than one lab).

Ideally, our proposal should identify one well defined problem that our set of experiments will address, and then we should show how all of our experiments will come together to solve it.  I think the way to do this is to focus on first determining whether uptake consists of distinct steps, and then on characterizing the steps as much as possible.  So we'll propose a distinction between the initiation of DNA uptake, an event that happens at one place on the DNA fragment (we think at the USS), and the subsequent uptake, an ongoing process that gradually pulls the entire fragment into the periplasm.

We can further ask whether there is an initial binding step that's distinguishable from the initiation of active uptake.  That is, does the uptake machinery at the cell surface first bind to DNA, either at the USS or first elsewhere and then at the USS, and then kink the DNA and pull the initiating loop across the outer membrane?  The alternative is that binding is not an independent step, but that DNA is kinked and pulled immediately on contacting the uptake machinery.

All of our experiments contribute to clarifying these distinctions, in one way or another.  So we just need to situate them in this context, and emphasize how they reinforce each other.  The second goal, of characterizing the steps, is open-ended, but I don't think this creates a 'focus' problem.

The distinction between binding and initiation has implications for the evolution of uptake specificity too.  We've previously considered two alternative explanations for uptake specificity.  Either the protein responsible for uptake specificity is an add-on, evolved to screen DNA for relatedness before allowing it to be taken up, or the uptake specificity is intrinsic to the initiation process, occurring because specific sequences are easier to process.  Identifying a separate binding step would favour the former.

Identifying the retraction protein

I've added a new (but obvious in retrospect) question to our CIHR proposal:  Which protein creates the retraction force that pulls the DNA across the outer membrane?  

The PilT protein that does this in other bacteria, but H. influenzae and the other Pasteurellaceae have no pilT gene.  We expect that the same retraction force is responsible for both initiation and continuation of uptake, and we can't begin to investigate its function until we know its identity (duh!).  We have a full set of nonpolar knockout mutants, covering all the genes that are implicated in competence by either direct mutations or by their membership in the CRP-S regulon, so we will screen these for the expected phenotype: no transformation, no DNase I-resistant DNA uptake, but binding of DNA to competence-induced cells (DNase I-sensitive radioactivity bound to cells.  

Testing for transformation and for DNA uptake are routine in our lab, but binding will be a bit trickier because we don't know what to expect.  Tests for DNA binding are hard to interpret in normal competent cells, because DNA is taken up soon after being bound.  So 'binding' is measured as the relative difference in cell-associated radioactivity ± treatment with DNase I ((noDNaseCPM - DNaseCPM)/noDNaseCPM).  This isn't very sensitive, because noDNaseCPM isn't much larger than DNaseCPM.  We don't know how much DNA should bind to the surface of cells that can't pull it in, nor how strong this binding should  be.  But DNaseCPM should be negligible (background in these experiments is usually <1% of DNaseCPM), so even if binding is only 10% of uptake  it should be readily detected.

Should this binding be USS-specific?  I think yes, but we won't throw out any uptake-minus mutants that bind DNA non-specifically. 

This screen will be the first set of experiments in the proposal.  The second set of experiments will be the DNA-protein crosslinking, and we will test any candidate retraction proteins for crosslinking to DNA, and test the corresponding knockout mutants for effects on the binding of other proteins.  A genuine retraction knockout should increase the crosslinking of the binding/initiation proteins that act upstream of  (before) it.

What if this screen for retraction mutants doesn't find anything?   We won't conclude that no retraction protein exists, but rather that there is no real binding step distinct from the initial initiation that pulls the first DNA into the periplasm.

Quick, change focus!

(I wrote this two days ago, but forgot to click 'Publish post'.)

The optical tweezers apparatus is now working well, so I'm off across town tomorrow to give it a try, with the expert help of my biophysicist collaborator.  But can I remember what I had learned so far, and what the status of my DNA-on-beads preps is?  No.  So I'm glad I wrote some blog posts and kept a tolerably good notebook.

I can attach biotinylated DNA to streptavidin-coated polystyrene beads.  The problem that the DNA caused the beads to clump together was resolved by agitating the beads better while they were incubating with the DNA (by putting the tubes on the roller in an orientation where they would be turned end-over-end).  However this resulted in beads that were difficult to pellet for washing, a problem that hasn't been solved but can be circumented by washing the beads by filtration and then concentrating them with a disposable protein-concentrator (Amicon).

• I need to find the bead-DNA preps I made and take them home with me tonight.

I can get cells to adhere nicely to coverslips if I rub the coverslips with ~10 µl of poly-L-lysine solution until dry.  It's easy to then assemble the treated coverslips into 'chambers' for the tweezers work.  I mark an 'X' near the center of the coverslip (on the coated side) with a fine-point sharpie; this makes it easy to be sure I'm focusing in the right place with the tweezers apparatus.  To prepare chambers with cells in them, I add the cell suspension to the chamber, incubate it coverslip-side down for 1-2 minutes, and then rinse thoroughly with 5-10 volumes of competence medium (rapping several times to dislodge weakly bound cells) before sealing with candlewax.  The attached cells are still alive and able to grow and divide if given culture medium.

• I will prepare some more coverslips and chambers before I go home tonight, and take them with me.

I have frozen competent cells (H. influenzae and B. subtilis) in the -80 °C freezer at the biophysics lab.

• But I suspect there may be contamination in the fridge stock of competence medium I use to wash the chambers and to wash and resuspend the cells in after thawing (to remove the glycerol antifreeze), so I should take some fresh stock with me and leave it in the -20 °C freezer rather than in the fridge.

Tomorrow the first objectives will be to make sure I can trap cells with the laser, and I can see the cells on the coverslip.  If that works, we can try bringing a bead to a cell and see if it attaches!

Tweezers progress (I've advanced to a new problem!)

The optical tweezers apparatus is finally functioning, so I spent yesterday trying it out.  I can't yet do what I want to do, but enough of the steps are working that I've been able to get to a new problem.

The various optical components have been adjusted, so now I could (sometimes) trap beads at the laser focus.  This is quite fussy; most of the time the beads are pulled into the focus point (the trap) and then spit right out again.  My colleague thinks this is probably because of a bit of astigmatism in the (cheap) laser, and because she's now using an oil-immersion lens rather than a water immersion lens.  I don't understand why the oil immersion lens would be worse, as the oil is specially designed to have the same refractive index as the glass on either side of it (the objective lens and the cover slip) so oil causes less refraction than water.  But anyway, the trap works best when it is within 20 microns of the coverslip/oil/lens.  Unfortunately it's tricky to get this distance because the micrometer used to move the chamber forwards and back has quite a bit of wobble ('hysteresis').

I could also focus on the cells I had attached to the coverslip, and the images looked much clearer than they had before the optics were adjusted.  Again though, the wobble in the micrometer made this somewhat imprecise.  And once I'd trapped a bead I could use the micrometer to move the chamber away from the objective, thus bringing the trap position to the surface of the coverslip where the cells were. 

At this point, if the beads had DNA on them, the cells might attach to the DNA.  But the beads I had brought were too small.  I had made them in a bit of a rush the night before, and although I was quite sure I'd taken the beads from the little bottle labeled "2.1 micron" I must have somehow used the one labeled "1.2 micron".  So I was using 2 micron beads without DNA, taken from an old tube that was in the drawer of the tweezers lab.

But this is where the new problem became evident.  When I brought the beads to the coverslip where the cells were, the beads quickly became stuck to the coverslip.  This wasn't exactly a surprise.  I knew this was likely to be a problem, and had been trying out ways to prevent it a couple of months ago.  But I hadn't found any that worked.  So I need to go back and try some more, both treating the coverslip after the cells have attached, and pre-treating the beads after the DNA has attached.

I may also be able to solve this problem by not petting the beads touch the cover slip in the first place.  In principle I should be able to bring the bead to a position that's still a few microns above the coverslip.  Depending on how far the DNA extends from the bead surface, the cell will be able to contact the DNA but the bead will not contact the coverslip.  The problem here is the poor control of the chamber position (the wobbly micrometer).  My colleague has what appears to be a better micrometer (bought on eBay!), but she's not sure it is compatible with the present setup.

I'm going to make one more try at this before submitting the CIHR grant (due in a month).  This time I'll take more time to carefully prepare my 2.1 micron beads with attached DNA.  The problem of beads sticking to the sides of the tube when being washed hasn't been solved but it can be minimized.  I'll add biotin to the first wash to block the streptavidin, and finally resuspend them in TE with added BSA to reduce nonspecific binding. 

A plan to identify competence proteins that interact with DNA

We're revising the first Specific Aim of our CIHR proposal because one of the reviewers (correctly) thought that it was too unfocused.

Originally we had:

Aim I. The functions of DNA uptake proteins

Q. 1. Which genes are needed for DNA uptake?

Q. 2. Which proteins contact DNA during uptake?

Under Q. 1 we proposed to create nonpolar knockout mutants of every competence gene (all 25 CRP-S genes and 2 other genes implicated in DNA uptake), and characterize their phenotypes.  Under Q. 2 we proposed to systematically test whether they bound DNA, using His-tagging and crosslinking.

The RA has almost completed making the knockouts.  Those of them with cotranscribed downstream genes (i.e. likely to have polar effects) are not nonpolar yet, because the selection for excision of the selective cassette turns out not to work in our lab strain (it's the standard lab strain Rd).  But we expect she'll be able to manually screen for excision (or we can hire an undergrad to do it).  We expect that this will be completed before the proposed start date of the grant. We now describe this in the Preliminary Results section of the proposal, and we've eliminated what was Q. 1.

We now also have a prioritized list of candidate DNA-contacting proteins, based on several attributes.  One is the presence of protein-export signals and other clues in the sequence annotation.  Another is what is known about the functions of homologs.  Another is what is known from the available mutants; mutants are available for only some proteins, and for some of these the mutant is expected to be polar on another gene that could account for its phenotype.

Aim I. Identifying DNA uptake proteins

Q. 1. Which proteins contact DNA during uptake?

Rationale:  Our in vivo crosslinking strategy can succeed where previous in vitro approaches have failed.  We will begin these studies with the two top candidates on our list, secretin and pilin.  In parallel we will use two simple steps to identify additional strong candidate proteins.

Methods:  1. Basic crosslinking assay:

• First put a His-tag on the gene in the chromosome.  Test whether cells with this tag arre still able to take up DNA.  But a test for DNA crosslinking might be worth doing even if the tag does interfere with function, if the protein is still assembled into a pore (secretin) or pilus (pilin).  So we should also test retrieval of the tagged secretin or pilin after formaldehyde crosslinking,with and without reversing the crosslinks before running the SDS-PAGE gel.  This will tell us whether the protein is assembled into its normal complex.
• Incubate the mutant cells with 32P-labelled DNA (probably the 222-bp USS-C fragment).  The standard way to form DNA-protein crosslinks is by adding formaldehyde, but this has the BIG disadvantage of also forming protein-protein crosslinks.  We hope to instead purchase photoaffinity nucleotides and incorporate one or more of these into the DNA we give to the cells.  This will allow us to specifically induce DNA-protein crosslinks (no protein-protein crosslinks) by irradiating the cells with UV (do we need a UV laser?  I think a colleague has one.)  After a very short time (1 minute?) UV to form crosslinks.
• Wash the cells to remove the external DNA, then lyse them and load them on a Ni-NTA resin column to bind the His-tagged protein.  Wash the column to remove everything else.  Elute the protein and check whether (i) the expected protein has eluted and (ii) any radioactivity has eluted.  If the protein is there but the radioactivity isn't, there was no crosslinking between this protein and the DNA.  If radioactivity elutes with the protein, investigate further.
• This experiment needs several controls.  The most important is probably a positive control for crosslinking of the DNA with a known protein.  How about SSB - it binds single-stranded DNA in the cytoplasm?  or DprA - it binds incoming DNA and protects it from nucleases?  A good negative control will be cells that aren't UV'd, as will DNA without the photoreactive nucleotide, and cells without the protein tag.

2.  Identifying other strong candidates (2 methods):

1. We will do simple transformation and DNA uptake assays on all the non-polar mutants.  These are already standard in our lab, and can be done in one term by an undergraduate or M.Sc. student.  Only proteins that are needed for normal DNA uptake will be retained as candidates.
2. We will use formaldehyde crosslinking , followed by gel electrophoresis and HPLC-mass spectromtry to identify proteins that are crosslinked (directly or indirectly) to DNA.  Any proteins not crosslinked will not be strong candidates.  The DNA will be tagged with biotin so that it and all crosslinked proteins can be recovered by attachment to streptavidin-coated magnetic beads (Dynabeads), and after recovery the crosslinks will be reversed so the proteins can be identified.  This is not a very specific test for uptake proteins, as it will also give proteins that bind incoming DNA in the cytoplasm, but missing proteins can be safely excluded.

3.  Characterizing the proteins that crosslink specifically to DNA:

Here;'s where the planning peters out a bit...

What do we do with these proteins once we've found them?  We test them for in vitro DNA binding, using bandshift and Southwestern assays. 

What if secretin and pilin don't crosslink to DNA???? Does this mean that they don't contact DNA at all?  This would be quite surprising.

Bacterial pseudogenes and within-species diversity

Last night Jon Eisen posted about a new paper by Chih-Horng Kuo and Howard Ochman, about the evolutionary fates of bacterial pseudogenes (PLoS Genetics: The Extinction Dynamics of Bacterial Pseudogenes).  I don't (yet) understand their conclusion very clearly, but it ties in well to the issues around the diversity of bacterial competence that I need to sort out for my CIfAR talk next week.

What do we know about within-species genetic diversity in bacteria?  The big issue is core genome and accessory genome.

In most (all?) species, different strains have a core set of genes in common; usually these make up about 80% of each strain's gene set (typical range ~70%-90%).  These core genes are usually syntenic.  They are very similar across the different strains, usually no more than a few percent different in DNA sequence, and almost identical in protein sequence, consistent with recent descent from a common ancestor.  These shared-by-descent genes are what justifies grouping the strains as representatives of a single 'species'.

The rest of each genome gene set comprises genes that are absent from some or most other strains.  It's not just that the alleles of these genes are very divergent, but that the genes have different ancestries.  Many of these accessory genes are in large blocks ('islands') with evidence of a mechanism by which they have been transferred from another distantly related species (e.g. phage, integron or transposon sequences, flanking tRNA genes).  This within-species genetic diversity is not seen in typical eukaryote genomes, perhaps because of the homogenizing effect of meiotic sexual reproduction.

Also unlike most eukaryote genomes, bacterial genomes usually contain only a small amount of non-gene sequences, usually about 10% of the genome.  This is almost entirely intergenic; introns are very rare and usually contain other genes (excisionases and mobilization genes).

What about pseudogenes?  Pseudogenes are DNA sequences that are closely related to functional genes but have mutations that destroy the function.  They are usually identified by comparison with the functional sequence in a close relative ('allele' if in the same species, 'homlog' if in another species).  Although function could be destroyed by mutations that change one or more critical amino acids, these can't be recognized without biochemical characterization of the gene product, and in practice pseudogenes are identified by the presence of a stop codon or indel that would prevent translation into a full-length protein.

Bacteria do have pseudogenes; in 2005 Lerat and Ochman examined 11 genomes from 4 genera and found that1%-8% of the open reading frames were pseudogenes.  Most pseudogenes were unique, defective in one genome and apparently functional in the genomes of close relatives, but some pseudogenes were shared between several Staphylococcus pyogenes strains and between two Vibrio vulnificus strains, and two were shared between the closely related V. vulnificus and V. parahaemolyticus.  Because shared pseudogenes were uncommon the authors concluded that old pseudogenes are rare.

The new paper examines the evolutionary histories of pseudogenes in five strains of Salmonella.  The strains all did have pseudogenes, from 0.3% to 3.7% of their functional genes.  All but 32 of the 378 pseudogenes identified had only a single defect, suggesting that they had arisen recently.  Consistent with recent origin, very few pseudogenes were shared (maybe 3?).  Most were created by small deletions or by point mutations that created stop codons.  The authors don't explicitly consider the core gene/accessory gene distinction, but because the pseudogenes were identified by alignment of not just the gene itself but of the genes flanking it on both sides, I think these are pseudogenes of the core gene set common to all strains, not of accessory genes present in only one or two strains.  (I just emailed the authors to check this.)  Many of the genes have no assigned or suggested function.

Kuo and Ochman ask why there are not more old pseudogenes.  But first I want to consider the basics -  what we might expect to happen after the first mutation happens.  If the functional gene makes an important contribution to fitness, we expect cells with the mutation to die or be quickly outcompeted by other cells, so the mutation will be gone from the population.  These pseudogenes are so short-lived that they are unlikely to be present in sequenced genomes.  If the functional gene makes little or no contribution to fitness in the present environment, the mutant cells may persist and even found a lineage (or, more likely, still go extinct).  The pseudogenes that are detected in sequenced genomes must be of this type.  Because the pseudogene's sequences are no longer under selection for the coding function, additional mutations that change its sequence may be selectively neutral, or they may be beneficial if they eliminate a harmful effect of the pseudogene.  What could such harmful fitness effects be?  The non-functional gene could produce a toxic product, being translated into a defective protein that interfered with the regulation or function of other proteins.  It might be transcribed but not translated, using up transcriptional resources.  Even if it is never transcribed, the cells still has to replicate and maintain this DNA, and it's often thought that bacterial cells have compact genomes because selection favours deletions of nonfunctional DNA that reduce this burden.

Kuo and Ochman conclude that(from the Abstract)

We found that after their initial formation, the youngest pseudogenes in Salmonella genomes have a very high likelihood of being removed by deletional processes and are eliminated too rapidly to be governed by a strictly neutral model of stochastic loss. Those few highly degraded pseudogenes that have persisted in Salmonella genomes correspond to genes with low expression levels and low connectivity in gene networks, such that their inactivation and any initial deleterious effects associated with their inactivation are buffered.

There are two points here, one I agree with and one I don't.  I agree that most pseudogenes are of recent origin, and their results do suggest that genes that are highly expressed and/or well connected are less likely to persist once they become pseudogenes.  The Discussion emphasizes the toxic-product hypothesis, which makes sense.

But I don't agree that deletion must be the reason we see few old pseudogenes in genome sequences.  It's true that deleting a pseudogene will eliminate both any toxic-protein cost and the cost of maintaining the unneeded DNA.  But it doesn't eliminate the cost of the original mutation that created the pseudogene.  Unless we have independent evidence that the DNA of pseudogenes is removed from genomes by deletion, we should probably suspect that instead cells carrying pseudogenes are removed from populations by selection.

Bottom line:

Is the DNA of new pseudogenes quickly lost from genomes by deletion, creating strains that are more fit than those with the pseudogene (but probably not more fit than the ancestor with the functional gene)?  This predicts that sequenced genomes should contain many sites where 'core' genes have been deleted.

Alternatively, are cells containing new pseudogenes quickly lost from populations because the cells compete poorly with cells that retain the functional gene?  This predicts that sequenced genomes will typically all contain the same core genes.

The figure below shows what we might expect to see when comparing 5 closely related genomes under each hypothesis  The orange parts of each bar represent genes that are intact in most genomes but are a pseudogene in one genome.

Back to doing the urgent stuff

For the last couple of months I've been doing experiments that were important but not particularly urgent, examining relationships between culture media, purine genes, and culture growth and competence.  Now it's time to turn to a couple of things that are both important and urgent.  One is preparing a talk about the diversity of natural competence, for a meeting next week.  The other is revising our unsuccessful CIHR grant proposal on DNA uptake for the September 15 submission deadline. 

The meeting is of the Integrated Microbial Biodiversity Program of the Canadian Institute for Advanced Research (CIfAR), and it's in Seattle (yes, not in Canada), from the evening of Thursday August 19 to lunchtime on Sunday August 22.  CIfAR meetings are always excellent: small (by invitation only), focused, very interactive, great food and accommodations.  My title is "Sporadic sex?  Scary food?  The biodiversity of natural competence", and I'll have 30 minutes (25 + questions).  In my talk I want to pull together our published work on the distribution of competence and transformability and integrate it into the bigger picture of why bacteria take up DNA.

The grant proposal is less urgent but more important.  It's already pretty good, but not good enough (it was ranked 10/47 but only 8 were funded).  The reviewers' comments were quite favourable.  Usually that's a good thing, but now it means that we don't have much guidance about what should be changed to make the proposal stronger.  The first step will be to sit down and carefully read the proposal with as fresh a mind as possible.  That should be easy, as I haven't looked at it or thought much about it since the end of February.  Then I'll read the reviewers' comments again.  Then we'll make all the immediate improvements we can, with the goal of having a draft ready for internal review in a week.

But first (most urgent but should be quick) we have to submit by Wednesday the documentation we need to get our Genome BC grant activated on October 1.  I've gathered the various forms and letters, but we're waiting for some information about Illumina sequencing costs from the Genome Sciences Centre.  We need this so we can can decide on the most cost-effective sequencing strategy for our $50,000.  Then we'll get a Statement of Work for this project from the GSC, and finalize the budget.

Preliminary test of viability in MIV (and other stuff)

Yesterday I looked at changes in the numbers of viable cells (cfu, colony-forming units) after cells of different genotypes (wildtype, purH, purR) had been transferred to the starvation medium MIV.  I also did this after transfer of wildtype cells to the defined medium cMMB without inosine.

The experiment had a couple of problems.  I'd wanted to compare survival of sxy- cells, but the only vials of frozen cells we have date from before the big freezer meltdown of 2004, and the culture I started from one vial remained resolutely non-turbid all day.  However I streaked it out and some colonies grew up overnight, and I've now confirmed that they are indeed kanamycin-resistant (the knockout is an insertion of miniTn10kan). So I can test them tomorrow.  I also had intended to use RM- rather than cMMB-, but my new stock of the 10x base for MMB smelled very nasty after autoclaving, and had a large precipitate, so I suspected that the glutathione I'd included had broken down (though the published instructions say to autoclave it).  So I threw it out and made fresh without glutathione, but this has also thrown a big precipitate.  Now I wonder if the problem is that this time I carefully adjusted its pH up before autoclaving, which the instructions don't say to do.  The solution started to become a bit cloudy when the pH got above 7.0, so I adjusted it down a bit, to 6.8, before autoclaving.  I'll try again tomorrow.

Anyway, both the mutant strains survived in MIV just as well as wildtype, with the cfu increasing 2-3-fold in 60 minutes.  I left the cells incubating overnight and plated them again this morning, so we'll  see if there's any differences in longer-term survival.  The wildtype cells transferred to cMMB didn't divide at all, nor did they become even a tiny bit competent. 

Now that I know roughly what to expect, I can do a more detailed test, and include the sxy knockout mutant.

What will we do with our new purH mutant?

I think our purH mutant may turn out, like our cya mutant, to be much more valuable than I had originally thought.  My first graduate student made the cya mutant (defective in adenylate cyclase and thus unable to make cyclic AMP (cAMP) as her M.Sc. project.  When she proposed doing this I thought only that it would let us confirm what was already known, that cAMP induces competence.  But we used it to test other questions about how the regulation works, and it's also been very useful as a conditional competence mutant - it can't transform at all, but transforms fine if given cAMP.

We made the purH mutant to eliminate endogenous purine synthesis so we could test the role of the purine repressor PurR without having to consider its effects on purine nucleotide pools, and so we could manipulate purine nucleotide levels by providing precursors in the medium, without having to worry about the unknown effects of endogenous synthesis.

Now we have the mutant I'm going back through my previous posts to collect the ideas I've had about ways to use it.  I've now checked its competence after induction with MIV starvation medium, and it's not significantly different from the purH+ wildtype strain (only done once so far).

1.  To find out whether PurR represses rec2, we can compare spontaneous competence in the purH mutant and a purH purR mutant.  We haven't constructed this double mutant yet but doing so will only take a day.   We'll do this test during 'late-log' growth in rich medium, a condition where PurR is normally active.  If PurR does repress rec2, we expect the double mutant to have a higher transformation frequency and higher expression of rec2 (assayed by quantitative PCR).  We can also examine rec2 mRNA by quantitative PCR and by using a rec2::lacZ reporter gene strain we have in the freezer.  This would be a good minor project for the post-doc, as it may lead to a publication.

2.  To find out whether sxy translation is controlled by purine pools, we can manipulate purine pools and compare sxy expression (sxy mRNA assay and Sxy protein assay or mRNA assay of any CRP-S gene) in purH mutants carrying either the normal sxy gene or one of our hypercompetent mutants.  If purine pools regulate translation of sxy mRNA, we expect to see the ratio of protein to mRNA increase when purine precursors are withdrawn.

3.  Does purine synthesis by wildtype cells help them survive in MIV?  Does it affect how competence develops at all?  The one experiment I've just done says not, but this needs to be repeated in more detail.

4. Do cells in rich medium get all their purines by salvage?  If so, the purH mutant should grow just as well as wildtype.

5. Does purine synthesis play any role in late-log competence?  If so, the purH mutant might become more competent than wildtype cells.

6.  I only today thought of this - we could also use the purH mutant to do something that's been on the back burner for more than a decade - demonstrate that H. influenzae cells actually do use the DNA they take up as a nutrient.  The experiment would use a hypercompetent-sxy purH double mutant, testing whether the presence of DNA increases survival (and maybe allows growth) in medium lacking a purine precursor.

Experiments to do

1.  Does induction of the competence regulon increase survival even when no DNA is available?

To explain why the CRP-S competence regulon includes genes that are unlikely to play any role in DNA uptake, we've hypothesized that these genes are turned on because they help the cell cope with stalled replication forks until the nucleotide supply can be restored.  If that's right, then a reasonable prediction is that viability should be reduced or growth delayed by that conditions that strongly induce the CRP-S regulon should threaten cell viability or growth, and that preventing CRP-S induction under these conditions will reduce viability or growth.

That is, mutants that can't turn on the CRP-S regulon should lose viability or show delayed growth when they experience competence-inducing conditions, but cells that can turn the genes on should do better.   So I'm going to grow up our sxy-knockout mutant and carefully compare it to sxy+ cells for survival in MIV, the starvation medium that best induces competence.  As a control, I'll compare its growth in the two defined RM media (+ and - inosine).

We now have a new purR knockout that we're sure is a complete knockout (because almost the whole gene has been deleted).  I need to check its competence, so I'll also check how well it survives transfer to MIV.  Maybe I'll also make a purR sxy double knockout and check that.

2.  Does the ability to synthesize purine nucleotides affect competence development?

We now have a purH knockout mutant, which can't synthesize purine nucleotides and can't grow in RM medium unless inosine is provided).  Does it develop normal competence when transferred to MIV from RM+ or from rich medium?  How quickly does it die in MIV?  What if it can't turn on the competence regulon - does it die sooner?

3.  Does the citrulline in MIV make any contribution as a pyrimidine precursor?

From reading the very old papers about H. influenzae's nutrient requirements, I've found out that it can use citrulline to synthesize the amino acid arginine as well as pyrimidine nucleotides.  That is, cells can't synthesize either arginine or pyrimidines from scratch.  If given uracil as a pyrimidine precursor , they must also be given arginine.  But if they're given citrulline they don't need the arginine.  I also hadn't realized that MIV contains citrulline, which cells could use to synthesize pyrimidine nucleotides.  But the concentration is only 6 µg/ml, whereas MMB and the RPMI-based medium both contain 300 µg/ml.

What does MIV do that inosine and cAMP don't?

I've done several more tests of how cAMP and the presence of inosine inluence competence development.  All these experiments used wildtype cells and the 'mixture' medium RM.  In each case I grew the cells in RM plus inosine for at least several hours at low density, so that competence was very low.   I then transferred them to different media that I thought would affect their competence (keeping the cell density about the same), and took samples.

Here's a graph of the results of the latest experiment.   The bottom (blue) line is the negative control: cells were put into fresh RM medium with inosine.  They continued to grow and became a bit ore competent as the culture got denser, but competence was still much lower than that of dense cultures.  The top (purple) line is the positive control: cells were put into the starvation medium MIV, which caused them to become very competent very quickly.  As I recall, cells transferred from rich medium to MIV take longer - I should look up some data.

The red line is cells transferred from RM plus inosine ('RM+') to RM minus inosine ('RM-').  The cells can still grow in this medium (with a variable lag), but there's a rapid (though modest) increase in competence.  (Hmm, I've been thinking that this competence is much higher than that of cells that were growing continuously in RM-, mainly because that was true the first time I did this transfer.  But I need to check my data for continuous growth in RM-, and maybe do more controls.)

Cells became tenfold more competent if they were instead transferred to RM+ with cAMP (teal line), and the combination of RM- and cAMP induced even higher competence (green line), up about 1000-fold from cells in RM+.  But this is still 100-fold lower than the MIV-induced competence.  How come?

The difference is probably because of something that RM- has and MIV lacks.  It shouldn't be the amino acids from casamino acids, as both media contain these (MIV has less), but RM has a lot of other ingredients that come from the RPMI component.  I wonder how much competence would be induced by transfer from RM+ to cMMB without inosine but with cAMP?  Perhaps I'll do a series of tests of adding different RM components to MIV.

Does it or doesn't it? (cAMP induce competence)

Here are the results of yesterday's analysis of competence development.

The cultures were kept at low density by repeatedly diluting them 10-15-fold in fresh medium.  For each medium type, the second sample was taken 70 minutes after the first, and the third sample was taken 95-125 minutes after the second.  Each culture was then diluted twofold into fresh medium with and without cAMP,  and the last two samples (with and without added cAMP) were taken after an additional 50 minutes.

The bottom graph (red bars) shows that cell densities were kept fairly constant by the dilutions (one sBHI sample had lower density than I expected).  The upper graph shows transformation frequencies on a log scale; the stars are there to indicate that the bars for cells in sBHI represent upper limits as there were no transformed colonies.  The middle graph shows the same data on a linear scale to emphasize that cAMP induced competence in RM + inosine just as well as it does in sBHI, but didn't induce competence at all in RM - inosine.

This clean result agrees with one of my previous observations, that log-phase cells in RM are more competent than log-phase cells in BHI, but unfortunately disagrees with another, that cAMP doesn't significantly induce competence in RM + inosine.  So I'd better repeat this experiment tomorrow.   I'll try to keep the cell densities more constant by making more frequent and smaller dilutions, and I'll plate larger volumes so the transformation estimates are more precise (many of them are based on less than 10 colonies).

Luckily I now have plenty of clean glassware.

Time to pause and regroup...

I did lots more transformation assays yesterday, log phase cells plus and minus cAMP, in rich medium and defined medium with and without inosine.  But I can't do any more until I wash and sterilize all the glassware/plasticware I've used.  Right now we don't have any lab slaves to do this for us.