(Some rethinking and editing done June 15.)
So now we know that purines aren't depleted in sBHI, even when the culture gets dense. At least, they're not depleted enough to inactivate PurR. But we also know that about 1% of the cells in dense cultures become competent, and that transcription of all of the CRP-S-regulated genes is increased. (Not necessarily increased in all cells, and maybe fully increased in 1% of the cells.)
If we ignore our less-than-compelling data about added nucleotides and effects of knocking out purR, the simplest interpretation would be that this competence is caused by rising levels of cAMP, due to depletion of preferred sugars (fructose?) from the medium. Maybe the only-1% activation would be because stochastic fluctuations in the metabolic signals cause only 1% of cells to activate sxy transcription. Or maybe sxy mRNA in only 1% of the cells doesn't form its translation-preventing secondary structure.
If the only-1% activation is caused by stochastic fluctuations in cAMP levels (just below and just above some threshold), then adding lots of cAMP should overcome this and make all cells competent. But adding 1 mM cAMP in early log just causes the 1% competence to happen sooner (~ 45 minutes after cAMP addition), and this competence doesn't increase when the culture with cAMP reaches the cell density that normally induces 1% competence. (This analysis should be repeated - it was done ages ago and not that well.)
After showing that H. influenzae's only PTS sugar uptake system is fructose-specific, and that the PTS system regulates cAMP levels, a former grad student tested the effect of 0.5% fructose on development of competence in sBHI. Surprisingly, she found no effect; she also found only a probably-insignificant 10-fold effect in MIV. She also constructed a lacZ-based cAMP reporter gene, and used it to detect changes in intracellular cAMP during growth and competence induction. She found that cAMP levels went up several-fold in late log, and went up similarly when early-log cells were transferred to MIV. This supports the above hypothesis that cAMP levels are not what limits late-log competence to 1% of cells (i.e. cAMP levels are plenty high, and something else is preventing most cells from becoming competent).
This suggest that the lack of full competence induction in late-log sBHI cultures might instead be due to borderline effects (of nucleotide pools?) on translatability of sxy mRNA, at a time when sxy transcription has been fully induced by high cAMP.. That would be consistent with the phenotype of hypercompetent-sxy mutant cells, whose sxy mRNA is always translatable - they show ~1% competence in early log and full competence in late log or after addition of cAMP. (We would then attribute the only-1% competence in early log to borderline effects (of nucleotide pool depletion?)on sxy transcription (threshold concentrations of cAMP?).
This post was motivated by my finding that purR- cells are much less competent in late log than purR+ cells (not yet replicated), and that this effect is much weaker in hypercompetent-sxy mutants. The microarray analysis (the previous post) shows that the wildtype competence induction isn't due to an increase in the activity of the PurR-repressed genes in late log, so any difference between wildtype and purR mutant should be because these genes are on in the mutant but off in wildtype cells. And as the main thing these genes do is produce purine nucleotides, the simplest interpretation is that having a better supply of purine nucleotides reduces production of Sxy protein in late log, and thus that the supply of purine nucleotides affects sxy expression. The difference in the hypercompetent-sxy mutants would then suggest that the secondary structure of sxy mRNA is responsible for sxy's sensitivity to purine nucleotide availability.
(One issue I need to keep in mind is the difference between a modest effect on all cells and a dramatic effect on a few cells. Recent single-cell observations from other labs are showing much more dramatic cell-to-cell differences than most microbiologists have been assuming.)
So what am I hypothesizing? When wildtype cells grow to high density in sBHI, cAMP levels become high in most or all cells. This cAMP causes high transcription of sxy, but most of the transcripts are not translated because their mRNA has folded into an inhibitory secondary structure. In a small fraction of the cells, the supply of purine nucleotides simultaneously falls low enough to prevent this structure from forming during transcription, thus allowing their sxy transcripts to be efficiently translated; the combination of Sxy and cAMP then causes these cells to express competence genes and become competent. In hypercompetent-sxy mutants, the supply of nucleotides doesn't matter because the mutations have destabilized the sxy mRNA secondary structure, so all the cells with high cAMP express Sxy protein and become competent when the culture becomes dense (only some in early log). In the purR mutant, the levels of purine nucleotides rarely fall below the threshold, so very few cells become competent in late log even though high cAMP levels are causing sxy transcription. And in the purR- hypercompetent-sxy double mutants, the cAMP-induced sxy transcripts are efficiently translated regardless of levels of purine nucleotides.
Continuing the hypothesis -- effects in MIV: When wildtype log-phase ('early-log') cells are abruptly transferred to MIV, cAMP levels rise sharply at the same time as the cells are suddenly cut off from the nucleotide precursors they've been getting from the culture medium. The high cAMP induces sxy transcription and the lack of purine precursors induces the purine regulon and causes a sudden fall in the supply of purine nucleotides (because the enzymes to synthesize them from scratch ('de novo') have yet to be made). This drop in purine nucleotides allows efficient sxy translation, so most of the cells become competent. Hypercompetent-sxy mutants are expected to experience the same fall of nucleotides and rise of cAMP, and they reach the same level of competence. What about MIV competence in the purR knockout? The purR- cells have already made the enzymes for synthesizing purine nucleotides de novo, and our original prediction was that withdrawal of the external supply of precursors wouldn't increase translation of sxy mRNA (i.e. these cells would respond to MIV like wildtype cells do to high density in sBHI). But this prediction wasn't met; the purR cells responded to MIV like wildtype cells do. An alternative hypothesis is that de novo synthesis of purine nucleotides in the purR mutant isn't enough to compensate for the sudden removal of the exogenous supply of precursors (the salvage pathways make a bigger contribution than the de novo pathway). It's also possible that the de novo pathway is subject to feedback regulation that limits its contribution when salvage is active.
Experiments to do? OK, I think I've argued myself right back to where I was a couple of weeks ago, but on a slightly more solid footing. What does the above hypothesis predict that we can test? (1.) First I need to replicate the late log competence results, testing the purR+ and purR- cells in both wildtype and hypercompetent cells. (2.) The late-log competence effect of the purR knockout should go away in a purR purH double mutant. The purR+ purH single mutant should also become competent normally in both sBHI and MIV. So, if the late-log effects are reproducible, I think we can show that they are due to changes in purine nucleotide pools that are sensed by the sxy mRNA secondary structure.
But can we show that this effect is also responsible for Sxy production in MIV? The effects of adding purine nucleotides, nucleosides and precursors are not straightforward and I'm going to assume that this is because of complications in the various salvage pathways. So I want to find a less unnatural way to manipulate nucleotide levels in the cell. But constitutively expressing the PurR-repressed genes didn't do it. What else could we do? Might another purine-supply gene be limiting at the time of transfer to MIV, one that isn't repressed by PurR? Could we prevent cells from using purine salvage in sBHI? They might then grow slowly and not respond to transfer to MIV.
Analysis of purine gene regulation
The legend on the right shows the time each sample was collected from the culture. In this time course, part of the culture was transferred to MIV at T=0 (green bars); the rest remained growing in sBHI for another 130 minutes (blue bars). The first two bars (light blue) for each gene are cells in sBHI 70 minutes and 30 minutes before the transfer; the bars are the same for the two parts of the time course.
The Y-axis is the ratio in mRNA between RNA isolated at that time and a reference sample prepared by mixing the RNAs from all 9 time points. This means that the height of the bars tells us nothing about the absolute expression level of the gene, just about whether its expression differs at different times. The genes not regulated by PurR have the same medium-high amount of RNA at all points (the high 30-minute MIV value for purU is due to an error in one spot of the array). Genes whose expression is sometimes low and sometimes high have low ratios when their expression is low, and higher-than-average ratios when it is high.
The upper graph shows that, when cells were transferred into MIV, expression of the nine PurR-regulated genes was strongly induced. This was known from the analysis we did in 2002 (see Redfield et al. 2005), and is also expected, because the cells have suddenly been moved from a medium with abundant purines to a medium with no purines, and the sudden lack of guanine and hypoxanthine will inactivate PurR.
The lower graph shows that, when the cells continued growing in sBHI, none of these genes were induced. Even though the culture would have become quite dense by the last time point, the purine genes were just as repressed as they were when the culture was still very dilute. This tells me that the medium still has lots of purines when cells stop growing (they must run out of something else), and that the moderate competence that develops at this late-sBHI stage doesn't depend on release of rec2 from PurR repression.
I didn't measure cell densities at these time points, but in similar experiments the density at T= -70 minutes would have been about 10^8, and at T=130 minutes about 7x10^9, only slightly below the stationary phase density of ~10^10. The cells at T=130 would have just reached their peak transformation frequency of about 10^-4, lower than the ~3x10^-3 of MIV-induced cells but much higher than the 10^-8 of dilute cells in sBHI.
Next steps with the old array data
What do I do with the array Background values? I was thinking I should divide by them, but I looked online and found that the usual practice is to subtract each spot's Background value from its Signal value.
I also need to figure out how to properly compare the effect of the purR knockout on purine genes to the(lack of) effect in the sBHI time course arrays. All the slides in the time course used the same 'reference' RNA (a pool of all the time point RNAs), but the purR analysis used RNA from purR+ cells in mid-log + cAMP.
Once I've figured this out I can do the next step with the array data: I'd like to compare rec2's amount of induction in late log (relative to other CRP-S genes) to the difference in its relative expression caused by the purR knockout. Maybe I'd better try saying this another way: I want to get expression ratio data for rec2 and some other CRP-S genes, both from the time course arrays and from the purR+/purR- array. In late log in sBHI (the time course), the CRP-S promoters are a bit more active than in early log, but PurR is keeping the purine genes off. In the purR+/purR- microarray, the expression of the CRP-S genes is probably comparable to the time course, but the purine genes are on full strength. If PurR doesn't normally repress rec2, the ratio of rec2 RNA to the other CRP-S RNAs should be the same in the time course arrays and the purR+/purR- array. If PurR does normally repress rec2, the ratio should be higher in the purR+/purR- array.
I don't want to go to the trouble of extracting the data for all the CRP-S genes, so I'll just do rec2, HI1182/1183, and dprA. No,,wait, the post-doc says he can easily write an R script to do this, so I'll wait until he's finished writing his NIH fellowship application. In the meantime I'll fix all the flaws in my purine gene analysis.
I also need to figure out how to properly compare the effect of the purR knockout on purine genes to the(lack of) effect in the sBHI time course arrays. All the slides in the time course used the same 'reference' RNA (a pool of all the time point RNAs), but the purR analysis used RNA from purR+ cells in mid-log + cAMP.
Once I've figured this out I can do the next step with the array data: I'd like to compare rec2's amount of induction in late log (relative to other CRP-S genes) to the difference in its relative expression caused by the purR knockout. Maybe I'd better try saying this another way: I want to get expression ratio data for rec2 and some other CRP-S genes, both from the time course arrays and from the purR+/purR- array. In late log in sBHI (the time course), the CRP-S promoters are a bit more active than in early log, but PurR is keeping the purine genes off. In the purR+/purR- microarray, the expression of the CRP-S genes is probably comparable to the time course, but the purine genes are on full strength. If PurR doesn't normally repress rec2, the ratio of rec2 RNA to the other CRP-S RNAs should be the same in the time course arrays and the purR+/purR- array. If PurR does normally repress rec2, the ratio should be higher in the purR+/purR- array.
I don't want to go to the trouble of extracting the data for all the CRP-S genes, so I'll just do rec2, HI1182/1183, and dprA. No,,wait, the post-doc says he can easily write an R script to do this, so I'll wait until he's finished writing his NIH fellowship application. In the meantime I'll fix all the flaws in my purine gene analysis.
More thinking about purine pools and PurR, followed by a RESULT!
What I want to find out (big picture for H. influenzae):
One good condition is early log phase growth in rich medium: cAMP is low, purines are high regardless of PurR activity, and Sxy is as off as it gets. Another is early log phase growth in rich medium with added cAMP: purines are still high, but Sxy expression is intermediate. Under these conditions, seeing higher rec2 expression in the purR mutant would be evidence that PurR normally represses rec2.
Starvation in MIV is not so good a condition: exogenous purines are gone, but the amount of endogenous purines may differ between PurR+ and PurR-. This could reduce Sxy expression in the purR mutant and thus activation of the CSP-S-controlled rec2 promoter. So seeing no change in rec2 expression would not be good evidence that PurR doesn't repress rec2. However, seeing that the purR knockout increased rec2 expression would be good evidence of repression, but might underestimate its magnitude. We could also just test whether supplementing MIV with either of PurR's co-repressors, guanine and hypoxanthine, affects rec2 expression - our preliminary experiments suggest that they don't affect transformation frequencies.
We could also test the hypercompetent sxy mutants, because we think that their Sxy expression in MIV may be less affected by nucleotide pools. Here, seeing that the purR knockout didn't change rec2 expression would be stronger evidence that PurR doesn't repress rec2. Analysis in MIV will be cleanest once we have a purH mutation to eliminate endogenous purines. In this background, seeing no effect of the purR knockout on rec2 expression would be unambiguous evidence that PurR doesn't repress rec2.
For question 2, we can measure Sxy protein expression directly with Western blots or with Laura's lacZ fusions, measure mRNA levels of Sxy-regulated genes with realtime PCR, measure DNA uptake, or measure transformation frequencies. One of Laura's fusions eliminates the sxy mRNA secondary structure, and this may be very useful as a control. We want to manipulate the cell's supply of nucleotides for transcription, which we can do by knocking out purH (eliminating endogenous synthesis of purine nucleotides), by knocking out purR (making endogenous synthesis of purine nucleotides constitutive) and by providing exogenous nucleotides, nucleosides or bases. We don't know how big the effects of the latter two interventions would be.) There's also a defined medium for H. influenzae, which can be supplemented with different amounts of inosine or other purine precursors. Hypercompetent sxy mutations may change the sensitivity of Sxy expression to nucleotides. If we're measuring transformation frequencies we also need to control for PurR-mediated effects on rec2 expression.
What culture conditions and genetic combinations should we compare, and what should we measure? Although not knowing whether PurR represses rec2 will initially limit interpretation of the results, it makes sense to initially measure the transformation frequencies of the culture samples that are collected for mRNA and protein measurements. We may want to leave the lacZ fusion analysis for later, as a corroboration.
We don't yet have a purH mutant, so we can't knock out endogenous production of purine nucleotides. However we can figure out how big a contribution endogenous synthesis makes in wildtype cells under different conditions, by comparing old microarray data on expression of PurR-regulated genes during growth in sBHI with the expression induced by transfer to MIV. We also have one array comparing PurR+ and PurR- cells, grown in sBHI + cAMP to an OD of about 0.5, so we can compare fully-derepressed to repressed expression.
LATER: OK, I've gone through the Imagene files of one of our microarray time courses (9 slides) and our one purR+/purR- microarray, manually pulled out all the genes with 'pur' and 'pyr' names into a new Excel file (by sorting by gene name), and examined their expression ratios throughout growth in sBHI and after transfer to MIV. The result is very clear: genes regulated by PurR are not induced at all in sBHI, even when the culture gets close to stationary phase. The genes are strongly induced by transfer to MIV and by the knockout of purR. (I can't show graphs here because I've forgotten how to force Excel to give me sensible axes. I'll post it tomorrow.)
This means that our purR knockout cells should have much higher expression of the purine biosynthetic genes at the time they're transferred to MIV, and much higher expression of them in late log. Furthermore, if PurR does regulate rec2, rec2 should normally be PurR-repressed in late log, with repression only released in MIV and in the purR knockout.
- Does PurR regulate rec2 expression?
- Do nucleotide pools regulate Sxy expression?
One good condition is early log phase growth in rich medium: cAMP is low, purines are high regardless of PurR activity, and Sxy is as off as it gets. Another is early log phase growth in rich medium with added cAMP: purines are still high, but Sxy expression is intermediate. Under these conditions, seeing higher rec2 expression in the purR mutant would be evidence that PurR normally represses rec2.
Starvation in MIV is not so good a condition: exogenous purines are gone, but the amount of endogenous purines may differ between PurR+ and PurR-. This could reduce Sxy expression in the purR mutant and thus activation of the CSP-S-controlled rec2 promoter. So seeing no change in rec2 expression would not be good evidence that PurR doesn't repress rec2. However, seeing that the purR knockout increased rec2 expression would be good evidence of repression, but might underestimate its magnitude. We could also just test whether supplementing MIV with either of PurR's co-repressors, guanine and hypoxanthine, affects rec2 expression - our preliminary experiments suggest that they don't affect transformation frequencies.
We could also test the hypercompetent sxy mutants, because we think that their Sxy expression in MIV may be less affected by nucleotide pools. Here, seeing that the purR knockout didn't change rec2 expression would be stronger evidence that PurR doesn't repress rec2. Analysis in MIV will be cleanest once we have a purH mutation to eliminate endogenous purines. In this background, seeing no effect of the purR knockout on rec2 expression would be unambiguous evidence that PurR doesn't repress rec2.
For question 2, we can measure Sxy protein expression directly with Western blots or with Laura's lacZ fusions, measure mRNA levels of Sxy-regulated genes with realtime PCR, measure DNA uptake, or measure transformation frequencies. One of Laura's fusions eliminates the sxy mRNA secondary structure, and this may be very useful as a control. We want to manipulate the cell's supply of nucleotides for transcription, which we can do by knocking out purH (eliminating endogenous synthesis of purine nucleotides), by knocking out purR (making endogenous synthesis of purine nucleotides constitutive) and by providing exogenous nucleotides, nucleosides or bases. We don't know how big the effects of the latter two interventions would be.) There's also a defined medium for H. influenzae, which can be supplemented with different amounts of inosine or other purine precursors. Hypercompetent sxy mutations may change the sensitivity of Sxy expression to nucleotides. If we're measuring transformation frequencies we also need to control for PurR-mediated effects on rec2 expression.
What culture conditions and genetic combinations should we compare, and what should we measure? Although not knowing whether PurR represses rec2 will initially limit interpretation of the results, it makes sense to initially measure the transformation frequencies of the culture samples that are collected for mRNA and protein measurements. We may want to leave the lacZ fusion analysis for later, as a corroboration.
We don't yet have a purH mutant, so we can't knock out endogenous production of purine nucleotides. However we can figure out how big a contribution endogenous synthesis makes in wildtype cells under different conditions, by comparing old microarray data on expression of PurR-regulated genes during growth in sBHI with the expression induced by transfer to MIV. We also have one array comparing PurR+ and PurR- cells, grown in sBHI + cAMP to an OD of about 0.5, so we can compare fully-derepressed to repressed expression.
LATER: OK, I've gone through the Imagene files of one of our microarray time courses (9 slides) and our one purR+/purR- microarray, manually pulled out all the genes with 'pur' and 'pyr' names into a new Excel file (by sorting by gene name), and examined their expression ratios throughout growth in sBHI and after transfer to MIV. The result is very clear: genes regulated by PurR are not induced at all in sBHI, even when the culture gets close to stationary phase. The genes are strongly induced by transfer to MIV and by the knockout of purR. (I can't show graphs here because I've forgotten how to force Excel to give me sensible axes. I'll post it tomorrow.)
This means that our purR knockout cells should have much higher expression of the purine biosynthetic genes at the time they're transferred to MIV, and much higher expression of them in late log. Furthermore, if PurR does regulate rec2, rec2 should normally be PurR-repressed in late log, with repression only released in MIV and in the purR knockout.
Free software for looking at Imagene files of microarray data?
We have a lot of old microarray data but no longer have a license for Genespring. I can use Excel to look at the individual Imagene files (Cy3 or Cy5 for each array hybridization), but this is very cumbersome, because I have to look up the intensity for each gene in each file separately.
Does anyone know of free software that will take the Cy3 and Cy5 Imagene files and show you the Cy3/Cy5 ratio for each gene? Preferably something that will run on a Mac?
We could try to again take advantage of Genespring's free 30-day trial, but I'd actually prefer something that was less trouble to set up. I don't need to have a genome view, or to have the software know the genome sequence that underlies the genes. And I don't want to do any fancy clustering analysis - I just want to see the intensity ratios.
Does anyone know of free software that will take the Cy3 and Cy5 Imagene files and show you the Cy3/Cy5 ratio for each gene? Preferably something that will run on a Mac?
We could try to again take advantage of Genespring's free 30-day trial, but I'd actually prefer something that was less trouble to set up. I don't need to have a genome view, or to have the software know the genome sequence that underlies the genes. And I don't want to do any fancy clustering analysis - I just want to see the intensity ratios.
Reissues by Wiley are reducing the usefulness of Google Scholar
Has anyone else noticed that Google Scholar searches for recent articles are bringing up a lot of old articles that appear to have been reissued online by Wiley?
The screenshot above is of the first two hits from a search for papers published since 2008. Both hits appear to be to papers published in 2008. The screenshot below shows what you get when you click on the second hit. It's a paper that was published in 1977, but that has somehow been 'republished' online in 2008. The link to the journal issue ("Volume 74 Issue 2") takes me to the correct 1977 issue, but the "Published online: 28 June 2008" is very misleading. The first hit has the same problem; it's to a paper that was originally published in 1985.
This isn't an isolated problem. Lately I've been seeing it a lot. I can't find any way to restrict my searches to genuinely recent articles, nor to exclude papers with the interscience.wiley.com tag. All I can do is avoid clicking on hits with that tag.
I looked at the Wiley-Interscience web site. They're bragging about a pilot project with Google and CrossRef to reissue old articles online; maybe this is the problem.
I looked at the Wiley-Interscience web site. They're bragging about a pilot project with Google and CrossRef to reissue old articles online; maybe this is the problem.
Next day: Quick and helpful response from Google Scholar:
Hello Rosie,
Unfortunately, yes, it looks like Wiley's abstracts occasionally have online dates instead of publication dates.
Thanks for bringing this to our attention. We'll contact Wiley but it will take some time for this to get resolved. In the meantime, a workaround when searching for recent articles is to add -site:wiley.com to your query, like so:
http://scholar.google.com/scholar?q=thymidine+uptake+coli+-site:wiley.com amp;as_sdt=2000&as_ylo=2008&as_vis=0
Sincerely,
The Google Scholar Team
What Amy did
Amy was an undergraduate working in our lab about 8 years ago. I remember her mainly for her smarts and persistence, and because she continued to visit us after moving to a different lab for her undergraduate project, and kept in touch for years afterwards (Amy, where are you now?)
Amy did quite a bit of work relevant to the effects of purine nucleotides on competence. The most immediately useful has been a big poster she drew, showing all the components (genes, enzymes, intermediates of purine metabolism in H. influenzae (bases, nucleosides and nucleotides, and both de novo synthesis and salvage pathways. I've just stuck it up on the corkboard outside my office, for handy reference.
Most of her time was spent constructing a fusion of lacZ to a control gene, surE, which is expressed in competent cultures but not regulated by PurR, and her very last experiment was to put her surE::lacZ fusion plasmid into H. influenzae and freeze the strains. We might use this plasmid as a control for effects of purine supplementation on general gene expression.
She also did a lot of tests of the effects of nucleotide, nucleoside and base supplementation. She repeated our experiments with nucleotides, finding that adding either 5 mM AMP and GMP to MIV reduced transformation more than 1000-fold, and that CMP and UMP had no effect. In one experiment AMP reduced cfu/ml by about 5-fold but in another it had no effect. She then checked effects on transcription, using lacZ fusions to comA, rec2 and sxy. With both AMP and GMP this showed about 10-fold reductions for rec2, 5-fold reductions for comA, and very little effect on sxy. CMP and UMP caused only less than 2-fold reductions. She used cells in sBHI as a control; all fusions were much lower than in any MIV treatment. She also tested the effects of purine bases on transcription; hypoxanthine and guanine had little or no effect (confirming our previous transformation assays), but inosine reduced comA and rec2 transcription by about 5-fold and 8-fold respectively. Inosine hadn't been previously tested by us, but decades earlier it had been reported to inhibit competence development in MIV (Miller and Huang 1970). The effects on AMP and GMP were smaller than we had previously reported, so she modified her protocol a bit (putting filtered cells directly into the supplemented MIV), and saw larger effects (for comA: >20-fold for AMP and ~10-fold for GMP).
Some of her first experiments tested whether nucleotide supplementation of sBHI affected cell growth. At 5 mM, all four nucleotides slowed growth dramatically; AMP halted it. (Adenine toxicity has been previously studied, in old E. coli experiments.) 1 mM nucleotides were much less toxic. The effect is unlikely to be due to unbalanced pools, as supplementing with all four nucleotides at 1 mM each caused similar growth inhibition to the single nucleotides.
Amy did quite a bit of work relevant to the effects of purine nucleotides on competence. The most immediately useful has been a big poster she drew, showing all the components (genes, enzymes, intermediates of purine metabolism in H. influenzae (bases, nucleosides and nucleotides, and both de novo synthesis and salvage pathways. I've just stuck it up on the corkboard outside my office, for handy reference.
Most of her time was spent constructing a fusion of lacZ to a control gene, surE, which is expressed in competent cultures but not regulated by PurR, and her very last experiment was to put her surE::lacZ fusion plasmid into H. influenzae and freeze the strains. We might use this plasmid as a control for effects of purine supplementation on general gene expression.
She also did a lot of tests of the effects of nucleotide, nucleoside and base supplementation. She repeated our experiments with nucleotides, finding that adding either 5 mM AMP and GMP to MIV reduced transformation more than 1000-fold, and that CMP and UMP had no effect. In one experiment AMP reduced cfu/ml by about 5-fold but in another it had no effect. She then checked effects on transcription, using lacZ fusions to comA, rec2 and sxy. With both AMP and GMP this showed about 10-fold reductions for rec2, 5-fold reductions for comA, and very little effect on sxy. CMP and UMP caused only less than 2-fold reductions. She used cells in sBHI as a control; all fusions were much lower than in any MIV treatment. She also tested the effects of purine bases on transcription; hypoxanthine and guanine had little or no effect (confirming our previous transformation assays), but inosine reduced comA and rec2 transcription by about 5-fold and 8-fold respectively. Inosine hadn't been previously tested by us, but decades earlier it had been reported to inhibit competence development in MIV (Miller and Huang 1970). The effects on AMP and GMP were smaller than we had previously reported, so she modified her protocol a bit (putting filtered cells directly into the supplemented MIV), and saw larger effects (for comA: >20-fold for AMP and ~10-fold for GMP).
Some of her first experiments tested whether nucleotide supplementation of sBHI affected cell growth. At 5 mM, all four nucleotides slowed growth dramatically; AMP halted it. (Adenine toxicity has been previously studied, in old E. coli experiments.) 1 mM nucleotides were much less toxic. The effect is unlikely to be due to unbalanced pools, as supplementing with all four nucleotides at 1 mM each caused similar growth inhibition to the single nucleotides.
More puzzles/questions about purine regulation of competence
Why does adding purine nucleotides to MIV repress competence more strongly than adding the corresponding nucleosides? And why does adding purine bases have little effect?
Does adding these nucleotides, nucleosides and bases affect transcription of the purine biosynthetic genes in the same way that it affects transcription of Sxy-dependent competence genes?
Does adding purine nucleotides to MIV repress only transcription of CRP-S competence genes (and the purine biosynthetic genes), or is it a more general effect on transcription?
How big a contribution does endogenous synthesis of purines make to the purine pools , relative to the level maintained in log phase by salvage from sBHI?
Why did we conclude that adding purine nucleotides etc. to sBHI makes cells grow poorly?
- Did Amy do experiments to address this? Look in her notebooks.
Does adding these nucleotides, nucleosides and bases affect transcription of the purine biosynthetic genes in the same way that it affects transcription of Sxy-dependent competence genes?
- Answer by measuring transcriptional effects on one or more PurR-repressed gene as well as on a Sxy-regulated gene.
Does adding purine nucleotides to MIV repress only transcription of CRP-S competence genes (and the purine biosynthetic genes), or is it a more general effect on transcription?
- Answer by measuring transcriptional effects on CRP-regulated genes that are Sxy-independent.
How big a contribution does endogenous synthesis of purines make to the purine pools , relative to the level maintained in log phase by salvage from sBHI?
- Answer by comparing transcription of a PurR-regulated gene in sBHI (log and late-log) and MIV, in wildtype cells and a purH mutant (no endogenous purine synthesis). First I think we'd need to make the purH mutant, but I should check our strain records in case we already did this.
Why did we conclude that adding purine nucleotides etc. to sBHI makes cells grow poorly?
- First repeat this to get solid data.
Purine regulation of competence
At today's lab meeting the Research Associate presented her analysis of all the work we've done over the past >10 years on how purines affect competence. It's time to get this all sorted out. The plan is to first repeat a few basic experiments and then either set it all permanently aside, or do the additional work to clarify the effects and publish a paper.
Here are links to my previous posts about this (most recent first, oldest last):
Because we have anecdotal (not well documented) reports that adding purine nucleotides made the cells sick, I think experiments that use added purines need to carefully control for non-specific effects on growth and transcription. However, the effects of knocking out purR aren't so simple either, as the purine biosynthetic genes will then all be constitutively induced, changing the effect of transfer to MIV.
Here are links to my previous posts about this (most recent first, oldest last):
- Overlooked evidence that purine pools regulate competence
- Does PurR regulate rec2 and does this matter?
- Our vindication is on hold...
- Make the bacteria do the work
- Vindication!
- Maybe PurR does repress competence genes after all
- Does PurR repress any competence genes?
- How sxy expression is regulated
- By changing culture conditions: We grow cells in a purine-rich medium, sBHI, where the purine biosynthetic genes are normally kept off by the purine repressor PurR. We think that purines are probably depleted at the end of log phase (we could check for induction of purine biosynthesis genes in our old microarray time course data), and we know they are when we transfer cells into the 'MIV' starvation medium that induces competence.
- By adding purine nucleotides, nucleosides or bases to BHI or MIV.
- By knocking out the gene for PurR so that the biosynthetic genes are always fully on.
- We can use real-time PCR or microarrays or Northern blots to measure the amount of mRNA produced from purine biosynthetic genes, from sxy, and from Sxy-regulated genes, and from control genes that should be independent of changed purine pools (e.g. housekeeping genes, CRP-induced genes (with CRP-N site promoters)).
- We can use gene fusions to the lacZ gene to indirectly measure the amount of transcription of sxy, rec2 and comA by assaying beta-galactosidase. We could do this for other genes too but we'd need to first construct the necessary lacZ fusions.
- We can use antibodies to measure the amount of Sxy protein produced. We may have an OK antibody for pilin too.
- We can measure DNA uptake using radioactive DNA.
- We can measure transformation frequencies using genetically marked DNA. The cells used for these assays can have been competence-induced with MIV, or they can be at different stages of growth in sBHI.
- We can add cyclic AMP (cAMP) to sBHI or MIV to maximally induce the sxy promoter (and all the Sxy-independent genes regulated by CRP).
- We can use hypercompetent-mutant versions of sxy, in which the 5' end of the mRNA does not maintain the expression-inhibiting secondary structure that we suspect responds to changes in purine pools.
Because we have anecdotal (not well documented) reports that adding purine nucleotides made the cells sick, I think experiments that use added purines need to carefully control for non-specific effects on growth and transcription. However, the effects of knocking out purR aren't so simple either, as the purine biosynthetic genes will then all be constitutively induced, changing the effect of transfer to MIV.
- Does PurR directly repress any competence genes? The only candidate with a reasonably good possible PurR binding site in its promoter is rec2. We can easily use real-time PCR to measure rec2 mRNA in purR+ and purR- cells. The problem is to keep the cells' supply of purines constant, so any effects of purine pools on sxy expression are controlled. I think the best solution is to have both purR+ and purR- cells carry a mutation knocking out the last step of the purine biosynthetic pathway (purH, HI0887), so that both would be totally dependent on the external supply of purines. An independent test would be to mutate the putative PurR binding site in the rec2 promoter - we could then look for effects on rec2 mRNA and on competence.
- Does the secondary structure of sxy mRMA make sxy expression sensitive to purine pools? I think our best tools for investigating this are the hypercompetence mutations in sxy. We could measure sxy mRNA, Sxy protein, and/or expression of any CRP-S genes in normal and hypercompetent cells. Ideally we would also do these experiments in a purH knockout. As a control for hypercompetence effects we could use our other set of hypercompetent mutants, with mutations in murE, but this control would be slightly compromised by our ignorance of how these mutations cause Sxy overexpression.
- We need to use PCR to confirm that our purR mutation is correct, because the original stock died and was recreated by transformation.
- I want to repeat the experiments I described in my post about Overlooked evidence that purine pools regulate competence. Then I only tested one condition, and not carefully. The former grad student also examined this in a single sketchy time course - this showed a smaller effect. So I think I should do time courses of transformation using wildtype, purR, sxy-hypercompetent, and the double sxy/purR mutant, all growing in rich medium. If I again see that the purR knockout dramatically reduces late-log competence, and that the sxy hypercompetence mutations eliminate/reduce this effect, then I'll be confident that the previously described effect of adding purine nucleotides was not due to some general toxicity due to unbalanced growth.
- I'll let someone else do the MIV-competence assays. We have good data showing that the purR knockout doesn't affect MIV competence, and that the purR mutant is not significantly less sensitive to addition of purine nucleotides that its purR+ parent. (However we also have good data showing that the purR knockout reduced expression of the CRP-S gene comA by about 50-fold, which is consistent with the hypothesized nucleotide pool effect of purR on sxy expression. And the knockout didn't reduce expression of rec2, consistent with the nucleotide pool effect being compensated for by loss of PurR repression of rec2.) If the hypothesis that nucleotide pools limit sxy expression is true, we predict that hypercompetence mutations in sxy will make competence in MIV much less sensitive to added purine nucleotides.
- I think it will be hard to use transformation assays to detect the effect of the hypothesized repression of rec2 by PurR, because so many other genes are also needed. We would need to find conditions where the amount of Rec2 is limited by PurR and Rec2 in turn limits the transformation frequency. Hmm, I think that's likely to happen only in a hypercompetent mutant in rich medium, and then maybe only if sxy transcription is induced with cAMP. So maybe I can also do the time courses ± cAMP.
Competent Bacillus subtilis
My new prep of competent B. subtilis worked well. The transformation frequencies to Trp+ and Met+ were 1.4x10^-2 and 4x10^-3 respectively. The frequency of Trp+ Met+ double transformants was 2.9x10^-4, indicating that about 20% of the cells are competent. The higher transformation frequencies may also be because I'm using a new prep of the donor DNA, not the 20-year-old prep I used the first time.
I froze most of these cells, and will use some today in more tests of how cells and beads bind to my poly-L-lysine coated coverslips.
I froze most of these cells, and will use some today in more tests of how cells and beads bind to my poly-L-lysine coated coverslips.
Coverslip tests
The B. subtilis cells stuck nicely to coverslips spotted with various concentrations of poly-L-lysine. Unfortunately (but not unexpectedly) so did the polystyrene beads.
This morning I'll test whether briefly soaking the coverslip in a solution of the protein bovine serum albumin (BSA) will prevent the beads from sticking. BSA is widely used as a general 'blocking' agent - by sticking to surfaces it occupies all the sites that a valuable molecule or particle might otherwise stick to. And because it's just a protein with no significant biological activity, it shouldn't harm the cells or interfere with their interaction with DNA. We probably even have a BSA stock in the freezer.
If this works, then I can test whether some of the cells in the competent cell prep that are already stuck on the coverslip will bind to the beads. I hope to see that no cells bind to the streptavidin-coated beads, but that some cells do bind to beads that have been incubated with DNA and then washed.
And then I'm going snowshoeing on one of the local mountains!
This morning I'll test whether briefly soaking the coverslip in a solution of the protein bovine serum albumin (BSA) will prevent the beads from sticking. BSA is widely used as a general 'blocking' agent - by sticking to surfaces it occupies all the sites that a valuable molecule or particle might otherwise stick to. And because it's just a protein with no significant biological activity, it shouldn't harm the cells or interfere with their interaction with DNA. We probably even have a BSA stock in the freezer.
If this works, then I can test whether some of the cells in the competent cell prep that are already stuck on the coverslip will bind to the beads. I hope to see that no cells bind to the streptavidin-coated beads, but that some cells do bind to beads that have been incubated with DNA and then washed.
And then I'm going snowshoeing on one of the local mountains!
Progress towards the tweezers experiments
On Wednesday I did a second test of ways to attach cells to coverslips.
This time I had pre-cleaned the coverslips by soaking them overnight in acid-ethanol (1% HCl 70% ethanol), and then air-dried them. I spotted solutions of poly-L-lysine on two of these; one goth 20 µl of the 0.01% solution recommended for dipping coverslips into, and one got 20 µl of the 0.1% stock solution. I let the spots sit for 5-10 minutes and then removed most of the liquid with a pipetman, and left the coverslips to air-dry. Then I spotted a bit of dense B. subtilis cell culture onto each coverslip. overlapping the spot of poly-L-lysine. I left the cells on the coverslip for about 5 minutes, and then rinsed each coverslip 5 times 1 second in a large volume of saline. Results: many many cells were stuck on each coverslip in the area that had been coated with poly-L-lysine. Both concentrations worked equally well.
Next I'll try using more dilute solutions, and try dipping coverslips in the solution, with and without rubbing until they're dry,as one experienced user suggested. I'll use the new competent cells I've just made. And I'll test whether the styrene beads (with and without DNA) stick to the treated coverslips, and whether the beads stick to the competent cells.
I also had one slide that had been incubated in sealed plastic tube with a small volume of the HDMS (?) silane, which is very volatile and should deposit vapour on the coverslip. I took several other coverslips and spotted HDMS on them and let it dry, and then spotted the cell culture on, and washed them as I did with the poly-L-lysine coverslips. Results: No cells stuck at all.
Today I'm making more B. subtilis cells competent. This time I'm freezing some of them, so I don't have to go through the competence ritual every time I want to test anything. And I'm also transforming them with my new prep of DNA, hoping to get a higher transformation frequency. And I'm plating some of the transformants on minimal plates with no casamino acids, with and without methionine, so I'll be able to score the frequencies of Met+ and Met+ Trp+ transformants, and thus calculate the fraction of the cells that are competent. (I'll need to wait 2 days for the colonies on these plates to get big enough to count.)
On Wednesday I also visited my physics collaborators. I attended their lab meeting - I think we'll start using their format on alternate weeks, with everyone given a short time (5-10 minutes) to update the others on their progress. The other weeks we'll continue with the one-person-presents format. And I had a good discussion with the PI, so now I know more about the practical issues. And she gave me a couple of reviews, which I haven't started reading yet.
Got to go DNase-treat and plate my transformants...
E. coli doesn't have sex
Yesterday I gave another talk, this one to the Biodiversity Centre in-house series. It was an evolution-focused version using my standard "Do bacteria have sex?" framework. This morning I realized that there's a point I should be making every time I use this framework.
Rather than describe the whole talk I've posted the slides to Slideshare (you can see them here). But you won't need to look at them to understand the point of this post. Early in these talks I always explain that, in bacteria, transfer of chromosomal genes between close relatives can occur by any of three 'parasexual' processes - conjugation, transduction and transformation. Incorporation of transferred alleles into the chromosome requires homologous recombination, mediated by an assortment of enzymes in the recipient cell. Then I describe the evidence that conjugation and transduction occur as unselected side effects of genes on parasitic genetic elements (plasmids and phages respectively) - the selected functions of these genes are infectious transfer of the parasites. And the evidence that all the enzymes that contribute to recombination have been selected for functions in DNA replication and repair, with no evidence of selection for the recombination effects.
At this point I have always next explained that transformation is thus the only bacterial process that still could have evolved for a sexual function (to promote recombination of chromosomal genes). But today I realized that I should first point out that this means that E. coli doesn't have sex, because E. coli is one of the many bacteria that don't have natural transformation.
Rather than describe the whole talk I've posted the slides to Slideshare (you can see them here). But you won't need to look at them to understand the point of this post. Early in these talks I always explain that, in bacteria, transfer of chromosomal genes between close relatives can occur by any of three 'parasexual' processes - conjugation, transduction and transformation. Incorporation of transferred alleles into the chromosome requires homologous recombination, mediated by an assortment of enzymes in the recipient cell. Then I describe the evidence that conjugation and transduction occur as unselected side effects of genes on parasitic genetic elements (plasmids and phages respectively) - the selected functions of these genes are infectious transfer of the parasites. And the evidence that all the enzymes that contribute to recombination have been selected for functions in DNA replication and repair, with no evidence of selection for the recombination effects.
At this point I have always next explained that transformation is thus the only bacterial process that still could have evolved for a sexual function (to promote recombination of chromosomal genes). But today I realized that I should first point out that this means that E. coli doesn't have sex, because E. coli is one of the many bacteria that don't have natural transformation.
DNA is biotin-tagged and binds to beads
The other day I cut H. influenzae chromosomal DNA with XhoI and separately with EcoRI. I inactivated the enzymes with heat (65°C for 20 minutes) and then incubated the DNAs with Klenow polymerase, dA, dG and dC, and biotin-tagged dU. Both of the restriction enzymes generate sticky ends with 5' overhangs that are 4 nucleotides long and include a dA, so the Klenow should have filled these ends in, incorporation the biotinylated dU to pair with the dA.
The only easy way I could think of to test whether the Klenow reaction had worked was to incubate the DNA with streptavidin-coated styrene beads (the same ones I'll use with the laser tweezers) and then run the mixture in an agarose gel. If the DNA has biotin ends, it will stick to the beads, which I think are too big to enter the gel. (I know the size of the beads (1.2 µ or 2.1 µ, depending on which vial I use); I'm just not certain of the exclusion limit of an 0.8% agarose gel.)
Anyway, the photo above shows that this worked. The top (1) and bottom (6) lanes are size standards. Lanes 3 and 5 are beads (1.2 µ ad 2.1 µ respectively) mixed with EcoRI-cut DNA that wasn't biotinylated. Lanes 2 and 4 are the same beads mixed with the biotinylated DNAs. You can see that a lot of the biotinylated DNA stayed in the well, just as it should. I don't know why the rest of the biotinylated DNA is smeared.
So tomorrow I'll get back to making the cells competent and persuading them to stick to a coverslip.
The only easy way I could think of to test whether the Klenow reaction had worked was to incubate the DNA with streptavidin-coated styrene beads (the same ones I'll use with the laser tweezers) and then run the mixture in an agarose gel. If the DNA has biotin ends, it will stick to the beads, which I think are too big to enter the gel. (I know the size of the beads (1.2 µ or 2.1 µ, depending on which vial I use); I'm just not certain of the exclusion limit of an 0.8% agarose gel.)
Anyway, the photo above shows that this worked. The top (1) and bottom (6) lanes are size standards. Lanes 3 and 5 are beads (1.2 µ ad 2.1 µ respectively) mixed with EcoRI-cut DNA that wasn't biotinylated. Lanes 2 and 4 are the same beads mixed with the biotinylated DNAs. You can see that a lot of the biotinylated DNA stayed in the well, just as it should. I don't know why the rest of the biotinylated DNA is smeared.So tomorrow I'll get back to making the cells competent and persuading them to stick to a coverslip.
Note added later: I just did the math, to calculate how many beads and how many DNA fragments I mixed together. There would have been about 10^8 beads (fewer of the 2.1 µ and more of the 1.2 µ) and about 1 µg of DNA. EcoRI cuts H. influenzae DNA to an average fragment size of 6 kb, so 1 µg is about 2 x 10^11 DNA fragments. Thus the mix I loaded on the gels had about 2000 DNA fragments per bead. If about 10% of the DNA stayed in the wells with the beads, that's about 200 DNA fragments per bead, which should be plenty for the preliminary uptake experiments.
Maybe uptake needs nicking, not kinking
On Friday I talked with a physicist who does single-molecule studies on DNA helicases, using a magnetic trap system. He gave me lots of good ideas about the physics of DNA uptake.
He thought that kinking of intact double-stranded DNA was indeed physically difficult, perhaps more difficult than I had realized, and suggested that instead one strand of the DNA might be nicked at the USS to facilitate kinking and thus passage of the molecule through the narrow secretin pore.
He thought that a (hypothetical) nickase might be small enough to pass through the pore along with the nicked/kinked DNA, at least if the pseudopilus wasn't still filling the pore. This would let the nickase remain covalently attached to the ends it had created (as topoisomerases do), and let it use stored bond energy (from the bond it broke) to reseal the nick in the periplasm. I said that resealing shouldn't be needed, but he thought that resealing (reversing the nicking reaction) might be the only way that the nickase could release the DNA. Resealing would also explain how the levels of supercoiling might be preserved, as was seen for some lightly supercoliled plasmids*. I told him about the postulated secreted ligase in the competence regulon, and he thought this was a reasonable candidate for a nickase.
One thing we need to do is carefully measure the uptake of plasmids with known amounts of supercoiling, looking for any evidence that supercoiling changes. Another is to make our own knockout of the periplasmic ligase and carefully test uptake of linear and supercoiled DNAs. And another is to test whether artificially nicking a plasmid stimulates uptake in the same way that an uptake sequence does.
He also suggested that 'tethered particle motion' analysis might be useful for investigating uptake. We would attach a bead to one end of a DNA fragment that had a single USS in a defined position (we already have this construct) and incubate this with competent cells attached to a cover slip. With suitable hardware and software, the position of the tethered bead in the x and y dimensions can be precisely monitored. If there is binding/initiation but no ongoing uptake of DNA, the bead will wander through a hemisphere whose center is the cell and whose radius is the length of the DNA from the point of binding/initiation. If the cell is taking up DNA, the radius will get progressively smaller.
Tethered particle analysis is supposed to be a lot less fussy than optical tweezers work. I'll have to think about what we might learn from it, and find out whether my sabbatical physics lab has the setup.
* A new thought about the experiment where supercoiling was preserved: The plasmids in question weren't really significantly supercoiled - their supercoiling levels were just what's trapped when linear DNA fragments are ligated into covalently closed circles. So the observation that the plasmid pool had the same distribution of supercoiling after uptake as before might just be because the plasmids were cut and then religated, again trapping random supercoils in the same proportions as before.
He thought that kinking of intact double-stranded DNA was indeed physically difficult, perhaps more difficult than I had realized, and suggested that instead one strand of the DNA might be nicked at the USS to facilitate kinking and thus passage of the molecule through the narrow secretin pore.
He thought that a (hypothetical) nickase might be small enough to pass through the pore along with the nicked/kinked DNA, at least if the pseudopilus wasn't still filling the pore. This would let the nickase remain covalently attached to the ends it had created (as topoisomerases do), and let it use stored bond energy (from the bond it broke) to reseal the nick in the periplasm. I said that resealing shouldn't be needed, but he thought that resealing (reversing the nicking reaction) might be the only way that the nickase could release the DNA. Resealing would also explain how the levels of supercoiling might be preserved, as was seen for some lightly supercoliled plasmids*. I told him about the postulated secreted ligase in the competence regulon, and he thought this was a reasonable candidate for a nickase.
One thing we need to do is carefully measure the uptake of plasmids with known amounts of supercoiling, looking for any evidence that supercoiling changes. Another is to make our own knockout of the periplasmic ligase and carefully test uptake of linear and supercoiled DNAs. And another is to test whether artificially nicking a plasmid stimulates uptake in the same way that an uptake sequence does.
He also suggested that 'tethered particle motion' analysis might be useful for investigating uptake. We would attach a bead to one end of a DNA fragment that had a single USS in a defined position (we already have this construct) and incubate this with competent cells attached to a cover slip. With suitable hardware and software, the position of the tethered bead in the x and y dimensions can be precisely monitored. If there is binding/initiation but no ongoing uptake of DNA, the bead will wander through a hemisphere whose center is the cell and whose radius is the length of the DNA from the point of binding/initiation. If the cell is taking up DNA, the radius will get progressively smaller.
Tethered particle analysis is supposed to be a lot less fussy than optical tweezers work. I'll have to think about what we might learn from it, and find out whether my sabbatical physics lab has the setup.
* A new thought about the experiment where supercoiling was preserved: The plasmids in question weren't really significantly supercoiled - their supercoiling levels were just what's trapped when linear DNA fragments are ligated into covalently closed circles. So the observation that the plasmid pool had the same distribution of supercoiling after uptake as before might just be because the plasmids were cut and then religated, again trapping random supercoils in the same proportions as before.
Is analysis of 'functional design' bad science?
Yesterday I gave a talk about my research to a three-lab group of evolutionary biologists at the university across town. The audience was quite skeptical of both my results and my conclusions. At the time this was quite stimulating (the question period went on for half an hour), but it left a bad taste in my mouth. I think I'm going to send them the email below.
Subject: Well, that was interesting...
Hi everyone,
Although I quite enjoyed Friday's extended discussion of my methods and my conclusions, I think the critiques were largely unfounded.
I was unfamiliar with the term 'functional design' as a method of investigating adaptation, and B—'s negative attitude to it left me wondering if I had misunderstood his explanation. But after reading up on it I agree that my approach is indeed analysis of functional design - I'm using a detailed study of a phenomenon to draw inferences about how selection has acted on it - and I don't understand the objections. Surely this is a sensible first step in understanding the evolution of any attribute or process. Before undertaking complex and time-consuming analyses using the comparative method or Lenski-style laboratory evolution experiments, we should always closely examine the attribute or behaviour we're interested in, to see if such studies are warranted, and to clarify the hypotheses that such studies would test.
Here's a made-up example: Suppose we want to know why lizards of some species sometimes rapidly nod their heads up and down. The behaviour is too specific to just be chance. Should we immediately start traveling the world to find out which species nod their heads on a diagonal, and which with a circular motion? And should we populate an island with a mix of lizards, assigning a series of grad students to see if the head-nodders slowly increase in frequency? Of course not. First we should find out the circumstances where head-nodding occurs, and the events that precede and follow it. If only males nod their heads, and only when females are present, and if copulation is always preceded by nodding, then we would sensibly hypothesize that head-nodding is a courtship behaviour. And we would test this hypothesis directly, perhaps by presenting females with plastic model males that did or didn't nod, or by seeing whether males prevented from nodding failed to reproduce. If these experiments left some important questions unanswered, more difficult studies might be warranted.
The same logic applies for molecular and microbiological phenomena, where we need to bring an evolutionarily-knowledgeable mind to the phenomenon in question.
Here's a real example. In this case, initial observations suggested selection for recombination, but this was not supported by more thorough analysis. The E. coli recBCD genes were first discovered because mutations in them dramatically reduced the efficiency of homologous recombination in conjugation experiments. That is, when the conjugative plasmid F is used to transfer DNA from a 'donor' cell into a recBCD-mutant cell, the DNA rarely recombines with the homologous sequences in the recipient chromosome. Initially researchers assumed that the function of these genes was to promote recombination, and mechanistic studies found that the RecBCD complex acted sometimes as a nuclease (degrading one DNA strand) and sometimes as a helicase (unwinding two DNA strands). This seemed to make sense: because homologous recombination is initiated by single DNA strands, these activities increase the recombination frequency.
Because the researchers assumed that these genes evolved for recombination, at first they didn't pay much attention to another phenotype of the recBCD mutants - most of the cells in a mutant colony or culture were dead. And the lethality wasn't associated with recombination of DNA from donor cells, but occurred under conditions where no gene transfer could happen. Years later, when they got around to investigating the cause of the lethality, they found that recBCD mutants were severely defective for DNA replication, because the nuclease and helicase activities are needed to repair double-strand breaks in DNA and to restart replication forks after they encounter DNA damage. Furthermore, the recombination-promoting activities appear to be exactly the same as those that facilitate DNA replication and repair.
Many molecular biologists still treat the two consequences of RecBCD action (recombination and replication/repair) as equally important, but as evolutionary biologists we know that the strength of selection matters. The selection coefficients for the recombination consequence can be estimated by the frequency with which conjugation transfer of chromosomal alleles occurs, multiplied by the frequency with which transferred alleles will increase recipient fitness, and divided by the frequency with which they will decrease recipient fitness. This is going to be a very small number, both because transfer of chromosomal genes by conjugation is rare in natural populations and because evolution-of-sex theory shows that recombination's benefits are typically small. A generous estimate of its upper limit might be 0.0001, but it could be negative if transferred alleles are more often harmful than beneficial, or if the incoming DNA triggers harmful repair reactions. On the other side, although the selection coefficient for the repair/replication consequence of RecBCD activity hasn't been directly measured, in lab cultures it's probably somewhere between 0.1 and 0.3.
So, is it worth doing comparative-method and lab-evolution studies of the hypothesized recombination benefit of the recBCD genes? I say no, for two reasons. First, the hypothesis is unfounded. The above analysis (of functional design) has identified a very large selective force arising from the need for DNA replication and repair, and provides no evidence of selection for the ability to promote recombination. Second, because the replication/repair selection is so strong, if the comparative-method and lab-evolution studies did find any apparent effects of selection for recombination, we would have to suspect that these arose from confounding effects of selection for replication and repair. Excluding these effects would be very difficult given the relative magnitudes of the selection forces.
I should clarify that the molecular analysis of RecBCD and of other proteins that influence recombination is not my work but that of many molecular biologists, and that these experts share the conclusion that these proteins primary (and perhaps only) selected function is DNA replication and repair. That is, over evolutionary time, any selection arising from recombination must have been far weaker than selection arising from replication and repair. The mechanistic information about conjugation and transduction that I provided in my talk was not my work either, and although I'm the only one who has formalized the evolutionary considerations, the conclusions are also shared by these experts. When I began independent research I chose to focus on transformation because the limited information available about its mechanism and regulation was not enough to support any conclusions about its function. Now I think there is.
I wonder if the fundamental problem the group had with my talk was that you found my conclusion distasteful (F— literally so, judging by the faces he was making). To clarify, my conclusion was that, in bacteria, no processes have been selected for the ability to promote homologous recombination between different lineages. I'm not saying that homologous recombination is never beneficial for bacteria - clearly sometimes it is and sometimes it isn't. And I'm not saying that lateral gene transfer between distant lineages has never been beneficial - analysis of genome sequences has revealed many examples where it has. But the benefits of recombination have not fed back on the processes responsible for it in any way that can be easily detected.
Thanks for inviting me. The discussion after my talk was stimulating, and I'd be happy to continue it either by email or in another talk.
Rosie
p.s. I'm attaching a pdf of a old review I wrote.
p.p.s. F—'s suggestion that I should qualify my unpalatable conclusion by adding something like 'the available data suggests' smacks of the kind of 'teach the controversy' strategy used by creationists. All scientific conclusions are based on the available data, and mentioning this serves only to imply that in this case the available data is weak or otherwise suspect.
Subject: Well, that was interesting...
Hi everyone,
Although I quite enjoyed Friday's extended discussion of my methods and my conclusions, I think the critiques were largely unfounded.
I was unfamiliar with the term 'functional design' as a method of investigating adaptation, and B—'s negative attitude to it left me wondering if I had misunderstood his explanation. But after reading up on it I agree that my approach is indeed analysis of functional design - I'm using a detailed study of a phenomenon to draw inferences about how selection has acted on it - and I don't understand the objections. Surely this is a sensible first step in understanding the evolution of any attribute or process. Before undertaking complex and time-consuming analyses using the comparative method or Lenski-style laboratory evolution experiments, we should always closely examine the attribute or behaviour we're interested in, to see if such studies are warranted, and to clarify the hypotheses that such studies would test.
Here's a made-up example: Suppose we want to know why lizards of some species sometimes rapidly nod their heads up and down. The behaviour is too specific to just be chance. Should we immediately start traveling the world to find out which species nod their heads on a diagonal, and which with a circular motion? And should we populate an island with a mix of lizards, assigning a series of grad students to see if the head-nodders slowly increase in frequency? Of course not. First we should find out the circumstances where head-nodding occurs, and the events that precede and follow it. If only males nod their heads, and only when females are present, and if copulation is always preceded by nodding, then we would sensibly hypothesize that head-nodding is a courtship behaviour. And we would test this hypothesis directly, perhaps by presenting females with plastic model males that did or didn't nod, or by seeing whether males prevented from nodding failed to reproduce. If these experiments left some important questions unanswered, more difficult studies might be warranted.
The same logic applies for molecular and microbiological phenomena, where we need to bring an evolutionarily-knowledgeable mind to the phenomenon in question.
Here's a real example. In this case, initial observations suggested selection for recombination, but this was not supported by more thorough analysis. The E. coli recBCD genes were first discovered because mutations in them dramatically reduced the efficiency of homologous recombination in conjugation experiments. That is, when the conjugative plasmid F is used to transfer DNA from a 'donor' cell into a recBCD-mutant cell, the DNA rarely recombines with the homologous sequences in the recipient chromosome. Initially researchers assumed that the function of these genes was to promote recombination, and mechanistic studies found that the RecBCD complex acted sometimes as a nuclease (degrading one DNA strand) and sometimes as a helicase (unwinding two DNA strands). This seemed to make sense: because homologous recombination is initiated by single DNA strands, these activities increase the recombination frequency.
Because the researchers assumed that these genes evolved for recombination, at first they didn't pay much attention to another phenotype of the recBCD mutants - most of the cells in a mutant colony or culture were dead. And the lethality wasn't associated with recombination of DNA from donor cells, but occurred under conditions where no gene transfer could happen. Years later, when they got around to investigating the cause of the lethality, they found that recBCD mutants were severely defective for DNA replication, because the nuclease and helicase activities are needed to repair double-strand breaks in DNA and to restart replication forks after they encounter DNA damage. Furthermore, the recombination-promoting activities appear to be exactly the same as those that facilitate DNA replication and repair.
Many molecular biologists still treat the two consequences of RecBCD action (recombination and replication/repair) as equally important, but as evolutionary biologists we know that the strength of selection matters. The selection coefficients for the recombination consequence can be estimated by the frequency with which conjugation transfer of chromosomal alleles occurs, multiplied by the frequency with which transferred alleles will increase recipient fitness, and divided by the frequency with which they will decrease recipient fitness. This is going to be a very small number, both because transfer of chromosomal genes by conjugation is rare in natural populations and because evolution-of-sex theory shows that recombination's benefits are typically small. A generous estimate of its upper limit might be 0.0001, but it could be negative if transferred alleles are more often harmful than beneficial, or if the incoming DNA triggers harmful repair reactions. On the other side, although the selection coefficient for the repair/replication consequence of RecBCD activity hasn't been directly measured, in lab cultures it's probably somewhere between 0.1 and 0.3.
So, is it worth doing comparative-method and lab-evolution studies of the hypothesized recombination benefit of the recBCD genes? I say no, for two reasons. First, the hypothesis is unfounded. The above analysis (of functional design) has identified a very large selective force arising from the need for DNA replication and repair, and provides no evidence of selection for the ability to promote recombination. Second, because the replication/repair selection is so strong, if the comparative-method and lab-evolution studies did find any apparent effects of selection for recombination, we would have to suspect that these arose from confounding effects of selection for replication and repair. Excluding these effects would be very difficult given the relative magnitudes of the selection forces.
I should clarify that the molecular analysis of RecBCD and of other proteins that influence recombination is not my work but that of many molecular biologists, and that these experts share the conclusion that these proteins primary (and perhaps only) selected function is DNA replication and repair. That is, over evolutionary time, any selection arising from recombination must have been far weaker than selection arising from replication and repair. The mechanistic information about conjugation and transduction that I provided in my talk was not my work either, and although I'm the only one who has formalized the evolutionary considerations, the conclusions are also shared by these experts. When I began independent research I chose to focus on transformation because the limited information available about its mechanism and regulation was not enough to support any conclusions about its function. Now I think there is.
I wonder if the fundamental problem the group had with my talk was that you found my conclusion distasteful (F— literally so, judging by the faces he was making). To clarify, my conclusion was that, in bacteria, no processes have been selected for the ability to promote homologous recombination between different lineages. I'm not saying that homologous recombination is never beneficial for bacteria - clearly sometimes it is and sometimes it isn't. And I'm not saying that lateral gene transfer between distant lineages has never been beneficial - analysis of genome sequences has revealed many examples where it has. But the benefits of recombination have not fed back on the processes responsible for it in any way that can be easily detected.
Thanks for inviting me. The discussion after my talk was stimulating, and I'd be happy to continue it either by email or in another talk.
Rosie
p.s. I'm attaching a pdf of a old review I wrote.
p.p.s. F—'s suggestion that I should qualify my unpalatable conclusion by adding something like 'the available data suggests' smacks of the kind of 'teach the controversy' strategy used by creationists. All scientific conclusions are based on the available data, and mentioning this serves only to imply that in this case the available data is weak or otherwise suspect.
Overlooked evidence that purine pools regulate competence
The research associate has prodded me into looking back at old experiments that used our knockout of the purR gene, and I've just realize that one of them provides strong evidence that purine pools regulate competence.
PurR is the purine repressor - it keeps expression of the genes needed to synthesize purines off unless purine pools are depleted. We initially knocked it out because we'd found that adding purines to starvation medium prevented induction of competence genes, and we thought that the competence genes might be directly repressed by PurR. If that were true then knocking out purR should increase competence under otherwise non-inducing conditions, but it didn't.
We now think that the purine effect we saw arises mainly because purine pools control the translation of the sxy mRNA. (PurR may also repress one competence gene, rec2; it has what looks like a strong PurR binding site in its promoter, but so far we don't have good evidence of repression.) The graph below shows how transformation frequencies are changed by introducing the purR knockout into wildtype cells and into five different strains with mutations in sxy that we expect to make sxy expression less sensitive to purine pools. The cells are not being starved, but are in fairly dense cultures in rich medium, a condition where competence is weakly induced. I forget what I was expecting at the time I did this experiment - I think I constructed the strains for a grad student who planned to look at levels of sxy mRNA or protein.
The blue bars are the transformation frequencies of purR+ cells, and the pink and green bars the transformation frequencies of two independent purR- isolates. The left most group is sxy+ cells, and the five other groups are different sxy hypercompetent mutants.
Let's assume that the only direct effect of the purR knockout is on the purine genes. Because there is no repressor in this mutant, these genes are transcribed at their maximum rate all the time, even when cells are growing in rich medium containing an abundant supply of purines. Thus the purine pools in the purR- cells are never depleted, even when the cells are transferred to the competence-inducing starvation medium.
The first thing to notice about the results is that the sxy+ purR- cells transform 100-fold worse than the wildtype parent. Thus preventing purine pools from becoming depleted dramatically reduces the induction of competence. So we really need to now directly compare the expression of competence genes in these two strains. I predict that they will all be down, except maybe rec2. We have done one microarray comparing purR+ to purR- cells; I wonder if the data is still accessible now that our expensive GeneSpring license has expired?
The second thing to notice is that the sxy mutant strains are much less sensitive to the effect of the purR knockout. Remember that these are not sxy knockouts; all but the first (RR699) have wild-type Sxy amino acid sequences. Rather these strains carry mutations that weaken the secondary structure of sxy mRNA, a structure that we have shown to directly limit the amount of Sxy protein produced from the mRNA. So, this new result shows that cells with a weakened sxy secondary structure are insensitive to the state of purine pools!
This was a quick-and-dirty experiment. I did it once and then froze the cells away. Now it's time to thaw them out and do some careful experiments of both transformation frequencies and gene expression. If we can also clarify the rec2 situation we'll have a nice paper.
Progress
I'm labeling with biotin-dUTP the restriction-digested MAP7 DNA that I want to attach to my streptavidin-coated beads. I wasn't sure what length range of DNA fragments would be best, so I've cut some with XhoI (mean length ~12 kb, ~4 µ) and some with EcoRI (mean length ~6 kb, ~ 2 µ).
After I've purified the biotin-tagged DNA (must get rid of ALL of the free biotin), I'll test its ability to bind to the beads by mixing some DNA with some beads and running the mix in an agarose gel. DNA that's bound to the beads should stay in the well rather than entering the gel, because the beads will be bigger than the gel mesh size. Well, the beads will be 1 µ or 2 µ in diameter (I'll try both, one with the XhoI-cut DNA and one with the EcoRI-cut DNA), and I'm quite sure that's much bigger than the mesh size of a 0.8% agarose gel. As controls I'll use the same DNA digests without the biotin tagging.
I'm also making more B. subtilis DNA for my transformation tests, because the first prep didn't have much DNA in it. I'll pool both preps and repurify by another round of phenol extraction and spooling.
Tomorrow I'm at SFU (the university across town), talking to their evolution group about why I'm working in a physics lab, and picking the brain of the Physics Dept's colloquium speaker.
After I've purified the biotin-tagged DNA (must get rid of ALL of the free biotin), I'll test its ability to bind to the beads by mixing some DNA with some beads and running the mix in an agarose gel. DNA that's bound to the beads should stay in the well rather than entering the gel, because the beads will be bigger than the gel mesh size. Well, the beads will be 1 µ or 2 µ in diameter (I'll try both, one with the XhoI-cut DNA and one with the EcoRI-cut DNA), and I'm quite sure that's much bigger than the mesh size of a 0.8% agarose gel. As controls I'll use the same DNA digests without the biotin tagging.
I'm also making more B. subtilis DNA for my transformation tests, because the first prep didn't have much DNA in it. I'll pool both preps and repurify by another round of phenol extraction and spooling.
Tomorrow I'm at SFU (the university across town), talking to their evolution group about why I'm working in a physics lab, and picking the brain of the Physics Dept's colloquium speaker.
B. subtilis progress
Most of the B. subtilis culture and transformation problems have been solved. What appeared to be growth of the trp- and met-requiring strain Y on minimal-glucose agar with no tryptophan or methionine turned out to be tiny crystals (of CaPO4?) that were seeded in the agar by the act of streaking over it. I now realize that I shouldn't have added CaCl2 to the agar at all.
The only problem remaining is that strain Y can grow moderately well on plates without added methionine, because the defined medium recipe says to add a small amount of 'casamino acids' to the agar. Casamin acids is the mix of amino acids obtained by breaking down the milk protein casein with acid; this destroys the tryptophan but the mixture is about 1% methionine by weight. The genotype of strain Y shows that it should only need methionine and tryptophan, and strain W shouldn't need any amino acid at all. My first test of medium without any casamino acids had no growth of either strain, but maybe I messed up. I've now tested using half and one sixth as much casamino acids as the recipe specifies (cells grow fine), and now I'm testing one tenth as much and none again.
I put strain Y through the B. subtilis competence ritual: Grow )shaking vigorously) in the minimal glucose medium plus casamino acids and a bit of yeast extract until growth has been slowing for 90 minutes, then dilute tenfold into more of the same medium with extra magnesium and calcium and shake for another hour, then mix cells with DNA of strain W. I definitely got transformants to Trp+, but the frequency was quite low. (I couldn't tell about transformation to Met+ because of the background growth due to the casamino acids.) The low frequency could be because the cells weren't very competent or because the strain W DNA was very old (it's been in the fridge for somewhere between 10 and 20 years ).
So now I'm going back into the lab to inoculate a culture of strain W so I can make some fresh DNA tomorrow.
The only problem remaining is that strain Y can grow moderately well on plates without added methionine, because the defined medium recipe says to add a small amount of 'casamino acids' to the agar. Casamin acids is the mix of amino acids obtained by breaking down the milk protein casein with acid; this destroys the tryptophan but the mixture is about 1% methionine by weight. The genotype of strain Y shows that it should only need methionine and tryptophan, and strain W shouldn't need any amino acid at all. My first test of medium without any casamino acids had no growth of either strain, but maybe I messed up. I've now tested using half and one sixth as much casamino acids as the recipe specifies (cells grow fine), and now I'm testing one tenth as much and none again.
I put strain Y through the B. subtilis competence ritual: Grow )shaking vigorously) in the minimal glucose medium plus casamino acids and a bit of yeast extract until growth has been slowing for 90 minutes, then dilute tenfold into more of the same medium with extra magnesium and calcium and shake for another hour, then mix cells with DNA of strain W. I definitely got transformants to Trp+, but the frequency was quite low. (I couldn't tell about transformation to Met+ because of the background growth due to the casamino acids.) The low frequency could be because the cells weren't very competent or because the strain W DNA was very old (it's been in the fridge for somewhere between 10 and 20 years ).
So now I'm going back into the lab to inoculate a culture of strain W so I can make some fresh DNA tomorrow.
The joys (?) of benchwork
So yesterday I streaked out the B. subtilis strain I'm planning to use as a control in my tweezers experiments, and its wildtype parent, first on rich medium (LB) and then (from that, after about 8 hours) on minimal plates with and without the supplements the control strain needs. This morning I found that the parent (call it W) had grown on all the plates, but the control strain (call it Y) had only grown on LB, when it should have grown on the plate with both supplements.
So I wondered whether my supplements (solutions of tryptophan and methionine) might have been harmed by being autoclaved rather than filter-sterilized. So I quickly made up more, this time filtering them, and streaked the right amounts onto the agar of supplemented plates like those that Y hadn't grown on, and then streaked both W and Y on them, from single colonies on the LB plates. Six hours later I could see tiny colonies of both strains on all the new plates. Not just the plate with both supplements, or just on the new-methionine plate, as might be expected if it was the original tryptophan stock that had been ruined by autoclaving. No, strain Y grew on the plate with the new methionine, and on the plate with the new tryptophan, and on the plate with both, as if it was the wildtype strain.
What control did I wish I had done - streaking both Y and W again on the original plate with both old supplements. Maybe I had accidentally streaked strain W in place of strain Y. So I streaked them on that. And then I made some new minimal plates, this time leaving out the trace amount of casamino acids the minimal recipe called for (maybe that had enough tryptophan and methionine???). I spread some of these plates with the new supplements, and streaked W and Y on plates with and without.
Then I wondered if I had somehow taken cells from the wrong vial of my 20-year-old- stocks, so I streaked from the vials again onto LB and old and new minimal plates. Could I have gotten the Y genotype wrong? No, a Google search for its genotype found a page specifying trp and met requirements.
Here's hoping the plates make more sense tomorrow morning.
So I wondered whether my supplements (solutions of tryptophan and methionine) might have been harmed by being autoclaved rather than filter-sterilized. So I quickly made up more, this time filtering them, and streaked the right amounts onto the agar of supplemented plates like those that Y hadn't grown on, and then streaked both W and Y on them, from single colonies on the LB plates. Six hours later I could see tiny colonies of both strains on all the new plates. Not just the plate with both supplements, or just on the new-methionine plate, as might be expected if it was the original tryptophan stock that had been ruined by autoclaving. No, strain Y grew on the plate with the new methionine, and on the plate with the new tryptophan, and on the plate with both, as if it was the wildtype strain.
What control did I wish I had done - streaking both Y and W again on the original plate with both old supplements. Maybe I had accidentally streaked strain W in place of strain Y. So I streaked them on that. And then I made some new minimal plates, this time leaving out the trace amount of casamino acids the minimal recipe called for (maybe that had enough tryptophan and methionine???). I spread some of these plates with the new supplements, and streaked W and Y on plates with and without.
Then I wondered if I had somehow taken cells from the wrong vial of my 20-year-old- stocks, so I streaked from the vials again onto LB and old and new minimal plates. Could I have gotten the Y genotype wrong? No, a Google search for its genotype found a page specifying trp and met requirements.
Here's hoping the plates make more sense tomorrow morning.
Experiments begun
I'm starting the laser tweezers work with Bacillus subtilis cells rather than H. influenzae cells, for several reasons. B. subtilis cells are much bigger than H. influenzae cells, and also tougher. I've worked with them before and know how to make them competent. They take up DNA without any sequence specificity, so I can use them with the same bead-attached H. influenzae DNA that I'll use for the H. influenzae cells. And they've previously been used for optical tweezers measurement of DNA uptake. So they're an excellent positive control for the H. influenzae experiments I want to do.
Today I streaked out the wildtype and auxotroph (trp- met-) strains from some 20-year-old slants; they're growing up fine. And I made the defined media needed for competence induction, and poured some plates with and without tryptophan and methione, and streaked both strains on all the kinds of plates to test that I've made them correctly. I found a DNA stock from the wildtype strain in the fridge. It's about 19 years old but I expect it will still be fine. If the transformations don't work well I'll make fresh DNA.
Today I streaked out the wildtype and auxotroph (trp- met-) strains from some 20-year-old slants; they're growing up fine. And I made the defined media needed for competence induction, and poured some plates with and without tryptophan and methione, and streaked both strains on all the kinds of plates to test that I've made them correctly. I found a DNA stock from the wildtype strain in the fridge. It's about 19 years old but I expect it will still be fine. If the transformations don't work well I'll make fresh DNA.