How to test our model of competence regulation

I've been working to clarify a written explanation of how I think competence is regulated.  Now I want to add ideas about how we should test this model.  Below I've pasted the explanation, and I'm going to embellish it with information about the evidence supporting these statements, and the need for additional evidence, in italics.

Effects in rich medium:  When wildtype cells grow to high density in sBHI, cAMP levels become high in most or all cells because the phosphotransferase system senses that preferred sugars are scarce and activates adenylate cyclase. Do we know this from the phenotypes of PTS mutants and from the induction of cAMP-dependent transcription seen in microarrays?  Or have these tests only been done in MIV-treated cultures?  We do know that adding cAMP to log-phase cultures raises competence to dense-culture levels but doesn't increase the level of competence they develop when dense, and that the rise in cAMP is due to the PTS.

The elevated cAMP activates CRP, causing high transcription of sxy, but most of the sxy transcripts are not translated because their mRNA has folded into an inhibitory secondary structure.  In a small fraction of the cells in high-density cultures, this structure either doesn't form or doesn't prevent translation of sxy mRNA, perhaps because random fluctuations in the supply of purine nucleotides have slowed transcription.  The postulated connection between purine nucleotides, rate of transcription, and sxy translation is purely hypothetical at present.  This is the connection I want to find out about.

The combination of high Sxy and high cAMP then causes this subset of cells to transcribe competence genes (in the CRP-S regulon) and become competent.  However, most cells in these cultures don't express enough Sxy protein to turn on competence genes.

When cells grow in rich medium (even at high density) the PurR repressor is active and the purine-biosynthesis genes are off, presumably because of high cytoplasmic concentrations of guanine or hypoxanthine, obtained directly from the medium or by conversion of other purine precursors.  We know from microarrays  that all genes subject to repression by PurR remain off in dense cultures; if H. influenzae's PurR protein uses the same cofactors as E. coli's PurR (the two proteins have similar sequences), this must be because levels of guanine or hypoxanthine are high.

Competence levels are different in dense cultures of hypercompetent-sxy mutants and of purR mutants.  In hypercompetent-sxy mutants we postulate that the mutations that destabilize the sxy mRNA secondary structure make sxy translation insensitive to the supply of nucleotides (again, this is what I want to investigate).  This causes all cells with high cAMP to express enough Sxy protein that they become competent; in dense cultures this is most of the cells.  Even in low-density cultures, many cells with these mutations express enough Sxy to become competent. 

Speculation based on behaviour of suspect purR mutant:  The opposite effect was seen in our present purR mutant (suspect because PCR shows a intact purR gene).  In such a mutant we expect that the constitutive activation of the purine biosynthetic genes will keep levels of purine nucleotides higher than in purR+ cells.  This may reduce the fraction of cells that translate enough Sxy to become competent at high density, even though high cAMP levels are causing sxy transcription, and that's the phenotype of the suspect mutant.

In the purR- hypercompetent-sxy double mutants these effects might cancel each other out, because the sxy transcripts will still be efficiently translated regardless of the levels of purine nucleotides, but the exact effect will depend on the relative strengths of the effects of the two mutations.

The other factor I'm ignoring here is the effect of PurR on expression of rec2.  The rec2 gene has what looks like a strong binding site for PurR, but we don't know if it's real.  This is a simple problem and we ought to quickly get it solved.  I discussed strategies here and here, several months ago, but they're on hold until we have a validated purR mutant (should be soon - the RA returns tomorrow).

Effects in MIV:  When wildtype cells in low-density ('log-phase) cultures are abruptly transferred to the starvation medium MIV, the cells are suddenly cut off from the nucleotide precursors they've been getting from the culture medium.  This causes a rapid fall in the supply of purine nucleotides AMP and GMP, which are continually being consumed by transcription.  cAMP levels also rise sharply on transfer to MIV, inducing sxy transcription, and the shortage of purine nucleotides allows efficient sxy translation so that most cells become competent.  Purine nucleotides cannot be immediately synthesized from scratch ('de novo') because the genes for purine nucleotide biosynthesis have until now been repressed by PurR.  The lack of purine precursors inactivates PurR, so the de novo biosynthesis genes are turned on, at least partially replenishing the supply of AMP and GMP.

What should happen if AMP or GMP is added to the MIV?  (I think these nucleotides are readily interconverted, so only one is needed.)  Even though nucleotides must be converted to nucleosides (have their phosphates removed) for transport across the cell membrane, I hypothesize that they have a more direct (faster?) effect on nucleotide pools than do simpler base precursors such as inosine, guanine, or hypoxanthine, which require extensive biochemical processing to be converted to nucleotides.  (I hypothesize this because the simple precursors have much less effect on competence than the nucleotides.   Is there a way to test this hypothesis?)  If so, supplementing MIV with AMP or GMP prevents the development of competence in most cells because it keeps their cytoplasmic concentrations above the threshold, preventing translation of sxy mRNA.

What about mutants?  Hypercompetent-sxy mutants in MIV are expected to experience the same fall of nucleotides and rise of cAMP, and their insensitivity to nucleotide supply won't matter when the supply is already depleted. This explains why they reach the same level of competence in MIV as wildtype cells. 

What about MIV competence in a purR knockout?  The purR- cells growing in rich medium will have already made the enzymes for synthesizing purine nucleotides de novo, and our original thinking was that this pathway would maintain high enough AMP and GMP concentrations to prevent translation of sxy mRNA when the external supply of  precursors was removed.   Thus we expected these cells to respond to MIV like wildtype cells do to high density in sBHI, with only a small fraction of cells becoming competent.  But this prediction wasn't met; the purR cells became just as competent in MIV as wildtype cells (this wasn't the suspect mutant, but one that had been carefully validated).  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 (i.e. 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, even though the genes are fully expressed.

OK, I'd better stop and make sure I'm not relying on incompatible hypotheses:  On the one hand, I've hypothesized that the salvage pathway that takes up simple purines from the medium and converts them into AMP and GMP is less efficient than the pathway that uses AMP or GMP from the medium to produce AMP and GMP in the cytoplasm.  On the other hand, I've hypothesized that de novo synthesis of AMP and GMP (from scratch) produces substantially less cytoplasmic AMP and GMP than the salvage pathway from simple purines.  These aren't incompatible, but they might be strengthened if I learned more about the biochemical control of these pathways.  For example, might

What to do?  I can't immediately start using the purR and purH knockouts to test these ideas, because these mutants aren't ready to use yet.  Is there another way to test the hypothesis that de novo synthesis is less effective than salvage?

Washing beads free of unattached DNA

So a while back I attached some more DNA to my polystyrene beads, but I didn't get to measuring how much DNA because I was having problems washing the beads+DNA free of the unattached DNA.  I wanted to just pellet the beads+DNA in a microcentrifuge, remove all the supernatant liquid, and resuspend the beads+DNA in fresh TE.  Because the volume of the pellet should be less than 15 µl, repeating this several times should be enough to bring the concentration of free DNA below its initial value of about 50 µg/ml to less than 1 ng/ml, well below the expected concentration of the DNA attached to the beads and thus low enough not to interfere with measuring how much DNA is attached to each bead.

BUT, the beads were difficult to pellet.  Some of them formed a pellet at the bottom of the tube, but the rest seemed to just stick to the sides of the plastic tube, even though they were subject to a centrifugal force of more than 10,000 times gravity.  Because the spinning tubes are oriented at an angle to the centrifugal force, I tried my usual trick of interrupting the spin to rotate the tubes 180°C, but the beads still didn't dislodge.

With this strong force, I had expected that the beads would be packed tightly ino the pellet, but they weren't - they dislodged all too easily while I was trying to remove the liquid above them*.  And the beads that, during centrifugation, behaved as if they were stuck firmly to the sides also dislodged easily when I was removing liquid.  The overall result was that either a lot of the beads had to be removed with the liquid, or a lot of the liquid had to be left with the beads.

One way to prevent things from sticking to surfaces is to coal the surface with a silane.  But pre-treating my microscentrifuge tubes with a silinazing solution ("Sigmacote") didn't make a big difference.

Then I thought of a totally different solution.  The beads are about the same size as bacterial cells, which we routinely collect and wash by filtration rather than by centrifugation.  Maybe I could wash the unwanted DNA from the beads+DNA by filtration.

The photo shows what we use to filter our cells: "Analytical Test Filter Funnels" from Nalgene.  They're sterile, fit on a vacuum flask, and have cellulose nitrate membranes with 0.2 µ pores.  The liquid flows through, and the retained cells are washed by flowing through a washing solution.  The filter funnels snap apart to allow the filter plus the retained cells to be easily retrieved.  We recover the washed cells by simply shaking the filter in a flask or tube with some medium.  The funnel can hold 100 ml of liquid, but we never use the full capacity because the filter becomes clogged if we try to use more than 10 ml of cell culture at an OD of 0.25.

So I tested using a filter funnel to wash some beads.  The first test wasn't very successful; although the filtering and washing seemed to work very well, about half of the beads stayed on the filter even after washing with 1% SDS.  But I tried again, using 2.1 µ beads rather than 1.2 µ, keeping the vacuum weak, making sure that the filter was never sucked dry, and recovering the beads by gently shaking the filter+beads for 15 minutes.  This gave full recovery of the beads, in 3 ml of liquid. 

But I need to concentrate my beads as well as wash them, and 3 ml is much too dilute.  So I still need to test whether I can use filtration to remove most of the unwanted liquid.  Perhaps I can just pipette the 3 ml onto a small disk of membrane filter, with some absorbent material beneath.  Or maybe I can put the filter disk on the filter of another funnel.

*  In hindsight this makes sense; the beads are rigid spheres, not squishy cells, so they can't really be 'packed down'.  However my biophysicist advisor told me she had been instructed to not spin the beads harder than 3,000 times gravity as they would deform.  So I checked the ones I'd spun hard - they all looked spherical under the microscope, but the laser tweezers may be much more sensitive to deformation than my eyes are.  Filtering will eliminate this concern too.

AAARRRGGHHH!!! (triaged)

Our application to NIH is rated unscored.

This means that it was ranked in the bottom half of all submissions, based on the evaluations by the reviewers who read it before the Review Panel meeting, so it was not discussed at the meeting.  The only reviews we'll get will be those prepared by these reviewers before the meeting (we'll get them in a couple of months).  There's probably not much point in working on a resubmission till then.

Later:  And as we're only allowed one resubmission, we need to make sure it's as strong as possible.  Better to take our time and do it right.

Back to thinking about purine regulation of sxy translation

I spent part of yesterday reading over posts from earlier this spring about how purine nucleotides and the PurR repressor might contribute to competence regulation.  I then started editing an old post to make it clearer, but now I'm elevating those edits to 'New Post' status here.  I'm continuing to edit them to make my thinking as clear as possible (more edits July 19).

I think it's important to consider the interdependent effects of at least two factors that affect nucleotide pools (1) the extracellular and intracellular concentrations of purine precursors, especially guanine and hypoxanthine, which control the repressing activity of the PurR repressor, and (2) the intracellular concentrations of purine nucleotides (ATP and GTP) that are available for transcription.  Another issue is the cell-to-cell variation in the amounts of cAMP, sxy mRNA and Sxy protein due to random fluctuations in transcription of the sxy gene, translation of sxy mRNA and the activities of various catalytic enzymes.

Here are the paragraphs I've been working on:

So what am I hypothesizing?  When wildtype cells grow to high density in sBHI, cAMP levels become high in most or all cells because the phosphotransferase system senses that preferred sugars are scarce and activates adenylate cyclase.  This cAMP activates CRP, causing high transcription of sxy, but most of the sxy transcripts are not translated because their mRNA has folded into an inhibitory secondary structure.  In a small fraction of the cells in high-density cultures, this structure either doesn't form or doesn't prevent translation of sxy mRNA, whether by chance or because random fluctuations in the supply of purine nucleotides have slowed transcription.  The combination of high Sxy and high cAMP then causes these cells to express competence genes and become competent.  However, most cells in these cultures don't express enough Sxy protein to turn on competence genes.

In these cells the PurR repressor is active and the purine-biosynthesis genes are off because of high cytoplasmic concentrations of guanine and hypoxanthine, obtained directly from the medium or by conversion of other purine precursors.

Competence levels are different in dense cultures of hypercompetent-sxy mutants and of purR mutants.  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 enough Sxy protein to become competent when the culture becomes dense.  (Even in low-density cultures, many cells with these mutations express enough Sxy to become competent.)  The opposite effect is seen in the purR mutant, where the constitutive activation of the purine biosynthetic genes keeps levels of purine nucleotides relatively high, so few cells translate enough Sxy to become competent at high density, even though high cAMP levels are causing sxy transcription.  These effects might cancel out in the purR- hypercompetent-sxy double mutants, because the sxy transcripts are still efficiently translated regardless of levels of purine nucleotides, but the exact effect will depend on the relative strengths of the effects of the two mutations.

Continuing the hypothesis -- effects in MIV:  When wildtype cells in low-density ('log-phase) cultures are abruptly transferred to the starvation medium MIV, the cells are suddenly cut off from the nucleotide precursors they've been getting from the culture medium.  This causes a rapid fall in the supply of purine nucleotides AMP and GMP.  These nucleotides cannot be immediately synthesized from scratch ('de novo') because the genes have until now been repressed by PurR.  cAMP levels also rise sharply on transfer to MIV, inducing sxy transcription, and the shortage of purine nucleotides allows efficient sxy translation so that most cells become competent.  The lack of purine precursors also inactivates PurR, so the de novo biosynthesis genes are turned on, at least partially replenishing the supply of AMP and GMP.

What would happen if AMP or GMP was added to the MIV?  (I think these nucleotides are readily interconverted, so only one is needed.)  Even though nucleotides must be converted to nucleosides (have their phosphates removed) for transport across the cell membrane, I hypothesize that they have a more direct effect on nucleotide pools than do simpler base precursors such as inosine, guanine, or hypoxanthine, which require extensive biochemical processing to be converted to nucleotides.  Thus supplementing MIV with AMP or GMP maintains their cytoplasmic concentrations above the threshold, preventing translation of sxy mRNA and thus the development of competence in most cells.

What about mutants?  Hypercompetent-sxy mutants in MIV are expected to experience the same fall of nucleotides and rise of cAMP, and their insensitivity to nucleotide supply won't matter when the supply is already depleted. This explains why they reach the same level of competence in MIV as wildtype cells.

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 thinking was that this would maintain high enough AMP and GMP concentrations to prevent translation of sxy mRNA when the external supply of precursors was removed.   Thus we expected these cells to respond to MIV like wildtype cells do to high density in sBHI, with only a small fraction of cells becoming competent.  But this prediction wasn't met; the purR cells became just as competent in MIV as wildtype cells.  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 (i.e. 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, even though the genes are fully expressed.

What to do?  I can't immediately start using the purR and purH knockouts to test these ideas, because these mutants aren't ready to use yet.  Is there another way to test the hypothesis that de novo synthesis is less effective than salvage?

Cells behaving badly

I'm still having problems with the E. coli cells that contain the DNA construct for the advanced tweezers experiments.  The cells grew OK overnight on an LB+Amp50 agar plate but poorly in what I later realized was LB+Amp200 broth (too much ampicillin).  I did a plasmid prep from these cells but got little or no plasmid (a faint smear only).  Yesterday the postdoc inoculated another colony for me, into LB+Amp50, but this also stopped growing at a low density, and a plasmid prep gave no plasmid at all.

The E. coli cells are strain DH5alpha, which should be quite vigorous, and the plasmid is a derivative of pGEM and so should have a high copy number, even with a 12 kb insert.  Perhaps the cells aren't what they're supposed to be.  Maybe there's something wrong with my medium (and the postdoc's medium).  Perhaps there was something wrong with my original LB+Amp plate.  Perhaps the plasmid replicates poorly because it's so big.  Perhaps the plasmid is present in the cells but lost during purification, because the miniprep kit doesn't give good recovery for DNAs bigger than ~10 kb.

Update:  this morning i inoculated colonies from the LB+Amp plate into plain LB and LB+Amp50.  The cells in plain LB seem to be growing faster, so maybe the plasmid is toxic.  I'd better go back and reread the thesis chapter of the M.Sc. student who made it.

Success, and maybe another reason why beads clump

I did a test of bead clumping.  I started with my prep of chromosomal DNA that had been cut with XhoI and biotin-tagged at both ends of the fragments.  I then cut this DNA with EcoRI, which will convert many of the fragments into shorter fragments with biotin-tags at only one end.  Then I incubated this DNA with beads overnight, using either a very concentrated mixture (20 µl each of beads and DNA, in a total volume of 100 µl) or a 50-fold more dilute mixture (40 µl each of beads and DNA, in a total volume of 10 ml).  The next morning I found almost no clumping, even less than at the beginning of the incubation.

But I had made one other change, in addition to the EcoRI digestion, which caused them to be much more vigorously mixed.  Previously I had been placing the tubes containing the mixtures into the slots on our roller wheel,  where they would rotate around their long axes with their bottoms always slightly lower than their tops.  This allowed some of the beads to eventually settle at the bottom of the tube.  This time I taped the tubes onto the sides of the wheel so they would rotate around their short axes, being turned completely upside down and then right side up with every rotation of the wheel.  I don't know if this improved mixing also contributed to the elimination of clumps.

I had done these incubations with both magnetic Dynabeads and polystyrene beads.  About half the Dynabeads went missing somewhere along the line (or maybe I accidentally put in less than I had intended).  I tested using a hemocytometer to directly count the number of beads in a defined volume.  This worked well, even though it's designed for cells much bigger than my beads, and I now know that I have about 2 x 10^7 of each type of bead, coated with DNA.  But I don't yet know how much DNA is on the beads, and I'll have to use up a substantial fraction of the beads for the Picogreen assay to measure this.

I also grew up the strain carrying the DNA construct plasmid.  It grew fine on a LB+Amp (ampicillin) plate, but not in LB +Amp broth.  But I now suspect I may have put too much Amp into the broth.

Why beads clump, and ways to prevent or minimize it

My preps of beads+DNA (biotin-tagged DNA linked to streptavidin-coated polystyrene or magnetic beads) typically contain dense clumps of beads - sometimes more of the beads are in these clumps than are solitary or in small groups.  This is a big problem, partly because the tweezers work requires isolated beads and partly because I can't accurately measure the properties of the preps.

It's pretty clear that the clumping is caused by the biotin-tagged DNA.  Adding DNase I to the preps breaks up the clumps, and there are no big clumps in preps of washed beads (no DNA) or if the beads are incubated with DNA that hasn't been biotin-tagged.  The likely cause is that, if the tagging reaction worked well, both ends of each DNA fragment should have biotin tags.  Such fragments can then bind to two beads, crosslinking them, and may also create DNA loops that entrap other DNA fragments.

I can see several ways to reduce this problem, illustrated in the figure.

1.  One strategy is to use a defined DNA substrate rather than restriction-digested chromosomal DNA, one that can be biotin-tagged at only one end.  A 14 kb plasmid designed for such experiments is available, and the biophysics grad student who worked on this project modified added a H. influenzae uptake sequence to it.  It's in our freezer.  All I need to do is digest with XhoI, biotin-tag the ends, and then digest with EcoRI to snip off the tag at the USS end.  I won't need to purify the long fragment, as the snipped-off bit shouldn't interfere with anything.  The EcoRI site is only 6 bp from the Xho site; luckily the ever-useful New England Biolabs catalog tells me that EcoRI doesn't mind cutting very close to an end.   Because this fragment contains only a single uptake sequence rather than the ~1 uptake sequence per kb of chromosomal DNA, it might not be as useful for the preliminary experiments whose goal is just to get cells to bind to DNA on beads and begin uptake.

2.  Another strategy is to use chromosomal DNA but reduce the number of fragments that are biotinylated at both ends, by cutting with a second restriction enzyme, one that doesn't give the appropriate DNA overhangs for the biotinylation reaction.  The first enzyme should be one that cuts infrequently, giving fragments that are mostly larger than, say, 20 kb.  The second enzyme should be one that cuts more frequently.  Because restriction sites are randomly positioned on chromosomal DNA. this second digest would cut many but probably not all of the doubly-tagged fragments, so I'd need to find the best balance between leaving some uncut and cutting too often, producing very short singly tagged fragments and many untagged fragments.  The presence of untagged fragments is probably not a serious problem, given the high affinity of streptavidin for biotin, but DNA uptake may not be easy to detect if the fragments are too short.

3.  The third strategy is technically simplest, although it's the last one I thought of.  I've been incubating the DNA with the beads in a small volume.  For most kinds of reactions, using a high concentration of reactants is good because it increases the frequency of interactions.  But streptavidin-biotin interactions have such strong affinities that this may not be an issue, especially because I can let the reactions proceed for a long time.  On the other hand, once a doubly tagged DNA fragment has bound a bead at one end, using a dilute mixture will increase the chance that the biotin at the fragment's other end will bind to the same bead before it encounters a different bead.

4. (not shown)  As a stopgap solution I can take the bead+DNA preps that I have and let the big clumps settle out.

CIHR results

We got the score and reviews for our CIHR grant proposal on DNA uptake this morning. 

Score:  4.37

Rank:  #10 out of 47 proposals reviewed

Percentile:  21.28% (yes, they report four significant figures)

Chance of funding:   Borderline


We'll find out whether it's funded in a few weeks.

(If you want to read the proposal, it's here.)

DNA is stably attached to beads

OK, after some delay due to putting the microtiter plate wrong-way-around into the plate-scanning machine, I have the results of the DNA assays of my bead+DNA preps before and after a final wash.

I think the diagram is pretty self-explanatory.  Most important result:  Almost all of of the DNA is firmly attached to the beads.  Even for the magnetic Dynabeads (columns on the right), the volume of DNA-containing supernatant remaining with the washed beads contributes no more than 1% of the total DNA in the -after-wash prep.  The three preps on the right have fewer but larger fragments than the three preps on the left.

Surprisingly, the Picogreen assay detected consistently more DNA on the beads after washing than before washing.  This is probably an artefact of something.  The beads were all originally in TE (10 mM Tris, 1 mM EDTA), and were washed in TE overnight and resuspended in TE.  Might the overnight rolling at 37 °C have better dispersed the DNA so the Picogreen could access it easier?  This seems unlikely to be the explanation as, because of my putting-the plate-in-backwards error, all the samples were mixed with Picogreen and then sat for 24 hr before being read.

Because the tweezers work doesn't need millions of beads, these preps should be enough to go on with.  This characterization also gives me confidence that my biotin-tagging and bead-attaching procedures work reliably.

Next steps:  Repeat the transformation and cell-attachment assays.  But first, clarify the bead-clumping situation.

How much washing is enough?

Today I'm testing my preps of beads with DNA bound to them by streptavidin-biotin linkages.  I want to know how much DNA is on the beads, and to be sure that this DNA is held by a streptavidin-biotin bond at one or both ends, rather than being passively trapped by other bead-DNA linkages.

I had four preps I'm made previously, two to EcoRI-cut DNA and two to XhoI-cut DNA; the average fragment sizes are about 6 kb and about 12 kb respectively.  I also have two new preps of a mixture of both DNAs, bound to magnetic Dynabeads (either  2.8 µ M280 beads or 1 µ T1 beads.  All of the preps were washed three times with about 100:1 volumes of TE, and resuspended in 100 µl of TE.  DNA concentrations have only been measured for one of the EcoRI and one of the XhoI preps.

The goal is to find out whether the intervening incubation and an additional thorough washing (rotating overnight in 10 volumes of TE) will remove more DNA from the beads.

So last night I measured the volume of each bead+DNA prep and took a 5 µl sample, to be assayed today for DNA concentration using the Picogreen assay.  The beads+DNA I did assay had about 250 ng DNA/ml, so this 5 µl was chosen to give enough sensitivity without using up too much of my preps.  Then I added 9 volumes of TE to each prep and put the tubes on the roller wheel in the 37°C incubator overnight.  Now I am going to pellet the beads (centrifuge or magnet), remove the supernatant, and resuspend the beads at their original concentration (= original volume - 5 µl).  Then I'll take 5 µl of the beads+DNA and 25 µl of each supernatant for Picogreen assay.

For assay standards I have lambda DNA at 1000, 100, 10 and 1 ng/µl, and lambda DNA mixed with beads at known concentrations (to check that the beads don't interfere with the assay).

I hope to find that most of the DNA is still associated with the beads. 

Ionic strength?

My physicist collaborator asked me the ionic strength of the MIV medium I'll be using for my optical tweezers experiments.  This is significant because the ions in the medium may shield the DNA from surface charges on the glass slide and cover slip, and so reduce unwanted binding.  I knew roughly what ionic strength is, but I had to look up the formula in Wikipedia.  I then looked up the recipe for 'solution 21'(the main component of MIV), looked up the molecular weights, and did the arithmetic.  The ionic strength of MIV is 0.5 M.  I think this is high enough to give lots of electrostatic shielding. 

While I was at it I also calculated the weight/volume concentration of MIV; it's about 3.5% (amino acids and salts), which I think is not high enough to interfere with the refraction needed by the tweezers.

Did the cells attach to the DNA-coated beads?

Yesterday I scored the results of Wednesday's test.  I had incubated competent H. influenzae cells with magnetic beads, using both DNA-coated beads and control beads that had gone through the same treatment without DNA.  Then I had washed the unattached cells away from the beads, taking samples for plating at every step so I could track the progress of the three washes.  I added DNase I to the beads after the last wash, so I'd be counting the total number of attached cells rather than the number of beads that had cells on them.  Then I plated all the samples, and later counted the colonies.

There were ten times more cells on the DNA-coated beads than on the control beads (about 10^5 vs 10^4).  This is good, but because I used about 10^8 beads, it means that, at best, only about one bead in 1000 had a cell attached.  I hadn't measured how much DNA was on these beads, lazily hoping that they had as much DNA as the similarly treated polystyrene beads.  Now I need to repeat the experiment with better-characterized beads.

Progress without tweezers

Yesterday I went to the biophysics lab, both to attend their weekly group meeting (another very good practice talk by a grad student) and to finally test whether I can see the cells under their microscopes (yes I could).

Their ordinary scope is a Zeiss that has darkfield but not phase contrast.  We couldn't get the darkfield set up properly.  Having now read up a bit on how to set up darkfield, I realize this was probably because we didn't have the 'stop' disk for the 20X objective, and didn't have a special darkfield condenser for the 100X objective.  But the RA showed me how to get OK contrast by shutting down the condenser diaphragm to a pinhole, and with this I was able to see not only B. subtilis cells but the much smaller H. influenzae cells.  This let me check that the cells were attached to the coverslip of the chamber before I put the slide with the chamber into the tweezers apparatus.

The laser component of the apparatus hadn't been realigned yet, but that didn't matter as I just wanted to use the visible-light component to find out whether I could see the cells.  If I hadn't been able to see the cells I might have had to abandon the whole project.  I guess I wasn't proceeding entirely out of optimistic ignorance, because I did know that the biophysics grad student who started this project a few years ago had been able to see H. influenzae cells in the apparatus he was using.  I couldn't get the cells into very clear focus under the 70X water-immersion lens of the tweezers apparatus, I think because the optics were a bit out of alignment.  The cells also weren't very conspicuous because they aren't very refractile (unlike the polystyrene beads).  But I could easily see that they were there.

From now on I'm going to mark the center of the coverslip with an X using a fine-point Sharpie before I assemble each chamber.  This X will give me a easy target to focus on, right in the plane of the attached cells.  I did this today, with some more careful tests of cell attachment, and it worked very well.

Tomorrow I'm going to make more competent B. subtilis cells, and continue testing cell attachment issues, especially how to prevent the beads from sticking to the cover slip.

Planning: Do competent cells bind to DNA-coated magnetic beads?

I want to test whether competent cells will bind to DNA that's attached to the surface of beads.  I'll use magnetic beads for this, rather that the polystyrene beads I'll use with the tweezers, as I can easily separate them from cells.  In principle I just need to mix cells and beads, pull out the beads, and plate them to count the number of beads that have attached cells  But the details need some thought.

In fact I already gave this experiment some thought, but I was derailed by finding that the beads formed big clumps when mixed with competent cells and by running out of competent cells.  I found that the BSA-coated ('MyOne') beads didn't clump so I'll use these, even though they're not the ones recommended for DNA work.  And I've made new cells.  But I didn't think much about some other details.

One is timing.  I think I should allow only a short time for the cells to interact with the beads.  This is more critical for B. subtilis, which can cut bound DNA and so might break lose of the beads.  So I think I'll give the H. influenzae cells 5 min to find the DNA, and the B. subtilis cells only 1 min.  Then I'll chill the tubes on ice before exposing them to the magnet.

Another issue is the absolute and relative concentrations of cells and beads.  I think I should use similar numbers of cells and beads.  My frozen competent H. influenzae cells are at about 10^9 cfu/ml.  The frozen competent B. subtilis cells are more dilute, ~2x10^8/ml.  The MyOne bead stock is even more concentrated, ~10^10 beads/ml, but I only have 1.0 ml and will use only a little bit of this. 

So let's say I coat 10 µl of beads with DNA, and mix them with 100 µl of H. influenzae cells, and after incubation I dilute this to 1 ml with cold wash solution.  Should I plate both the bead fraction and the supernatant ('side-natant'? call it the SN) fraction, and how many washes should I do?  If all the cells are competent and all the beads are DNA-coated, then all the cells might bind to the beads.  If most of the cells aren't competent, or if cells are inefficient at finding and binding to the beads, most of the cells will be in the SN, and plating it will probably not be sensitive enough to detect the small decrease. 

Let's conservatively assume that only 10^6 cells bind to the 10^8 beads, and that about 5% of each SN is carried over with the beads in each wash.  The first SN will still contain ~10^7 cells/ml, and the bead fraction will have 1.5x10^6 cells (~10^6 attached and ~5x10^5 free).  I'll resuspend the beads in 1 ml, mix gently, and collect the bead fraction again.  The second SN will contain about 5x10^5 cells/ml, and the beads will contain about 10^6 (~10^6 attached and 2.5x10^4 free).  If I wash these beads a third time, the SN will contain about 2.5x10^4 cells/ml and the bead fraction will still contain about 10^6 cells.

So I could get away with doing only two washes, but doing a third would give more confidence that the cells are really attached.  This assumes that the cells stay attached to the beads through the washings - this will require mixing them quite gently and keeping them cold, and hoping.

One other complication is that counting the cells on the beads uses up beads.  But I don't want to count more than ~100 colonies, so I could just take 10 µl of each 1 ml resuspended bead fraction for dilution and plating.

Cells attached to glass slides are not only viable but transformable

I bound competent cells (H. influenzae and B. subtilis, separately) onto cover slips in chambers, and then washed in medium containing genetically marked DNA.  After 15 minutes at 37 °C I washed in some medium with DNase I and then medium containing 0.5% low-melt agarose, with and without either nalidixic acid (to select the H. influenzae nalR allele) or tryptophan (to select the B. subtilis trp+ allele).  Then I sealed the chambers' ends with wax from a candle (clumsily) and incubated them overnight.

The results were a bit messy.  The H. influenzae cells started out quite dense (1000-5000 per 40X field of view) and without nalidixic acid they grew into nice microcolonies.  With nalidixic acid they instead formed filaments (i.e. grew without dividing, and then stopped growing).  At the edges of the chamber some large and well-defined microcolonies were present, but these were absent from the central part of the chamber.  So transformation is clearly happening, but I can't estimate a frequency.  The distribution of the NalR microcolonies suggests that oxygen might affect resistance, but I don't know enough about how the drug acts (a gyrase inhibitor) to guess why that would be the case.

The B. subtilis results were worse.  The chambers with and without tryptophan had similar numbers of cells.  Both had substantially more cells than had originally been attached, so there must have been some growth.  But there were no obvious microcolonies.  I might have made an error with the medium.  But many of the cells were moving around, so I suspect that I also need to use a higher concentration of low-melt agarose to block their strong motility.

On a separate topic, a reader suggested using polyethylene glycol (PEG) to block the poly-L-lysine coating and prevent beads from sticking to it once the cells had been bound.  After checking that PEG isn't toxic to cells, I tried washing my chambers with a 1% solution of the PEG type we had on the shelf (PEG 3350, which I think is a moderate chain length for PEG), washing it out, and then washing in some beads and washing them out after 10 minutes.  Result - no significant difference between PEG and no PEG in the numbers of beads bound to the coverslip.  I could try a higher concentration of PEG, but maybe the problem is washing out the PEG before adding the beads.  I'll try adding PEG to the beads as well as to the washing solution.  If this does prevent beads sticking, I'll next need to test whether PEG inhibits transformation.

While I was at it I also tested whether the 16% glycerol that's mixed with the frozen competent cells inhibits cell binding to the coverslips.  I had been conscientiously pelleting the thawed cells and resuspending them in glycerol-free medium before adding them to the chambers, but now I know that cells bind just fine in the presence of glycerol.  The glycerol is subsequently washed out of the chamber along with the non-attached cells, so I won't bother with the centrifuging step any more.

And I tested alternatives to sealing the chamber ends with wax, which is difficult to apply and tends to form big lumps rather than a smooth layer.  (I don't like the risk of getting wax on the microscope lens.)  First I tried a better way of applying the melted wax - rather than using a glass Pasteur pipette with a rubber bulb, I tried using an old Pipetman p200 with a snipped-off tip.  This gave better control, but the wax still formed a big lump when it met the glass.   Parafilm didn't work - a film of liquid quickly seeped under the parafilm.  Melting the parafilm with a heated spatula didn't help.  I also tried using paint from a paint-Sharpie; this only sort-of worked, perhaps because it's a water-based paint.  (But it was cool to look at the paint droplets under the microscope.)