Uptake and Transformation with "Biorupted" samples

To get an idea of how uptake and transformation would work with different sized donor chromosome fragments, I took MAP7 DNA and sheared it in a “bioruptor”. I ended up with several samples with different size distributions, three of which I used for a pair of experiments:

LARGE: >40 kb (unsonicated)
MEDIUM: 1-10kb (1 X 10 min sonication)
SMALL: 100-400 bp (5 X 10 min sonication)

My naïve assumption was that % uptake would go down as the fragment size decreased, since fewer fragments would contain “uptake signal sequences” (USS), which have an average density in the genome of ~1kb.

I also thought that transformation rates would also go down for smaller fragments, but would not necessarily correlate that well with uptake, since additional steps of translocation and recombination could also potentially influence the efficiency of transformation. So for example, transformation might drop off more quickly than uptake, if degradation would affect smaller fragments more than larger fragments. (This seemed to be the case in Pifer and Smith, 1985.)

Keeping in mind that these are just one-off experiments and need to be repeated (like pretty much every experiment I’ve reported in this blog), the above predictions look like they’re true, but I’m not certain if my reasons are necessarily correct...

First, I’ll show the uptake data. I end-labeled the MEDIUM and SMALL donor fragments using Klenow and did a simple uptake experiment (comparing total radiolabel to that in cell pellets after 30 min of uptake) using wild type cells and as donors, either MEDIUM or SMALL chromosome fragments, and either saturating (500 ng) or sub-saturating (100 ng) amounts of input DNA per 0.5 ml of wild-type competent cells. Here’s the data:
So clearly, several-fold less chromosomal DNA is taken up when smaller fragments (100-400 bp) are used than when larger fragments (1kb-10kb) are used. As explained above, one reasonable explanation for this is that fewer fragments in the SMALL sample contain USS, so maybe only 1/10 to 1/2 will have a USS, whereas in the MEDIUM sample most fragments will contain at least 1 USS.

But there is an alternative explanation that I’d like to be able to distinguish (but am pretty sure I can’t with this one experiment). It could be that a given competent cell only takes up a fixed number of DNA fragments, independent of fragment size. So since the SMALL sample is composed of ~10-100X more fragments per unit mass, it could be that I’ve simply saturated the system with fragments in the case of SMALL, but not in the case of MEDIUM. This issue was addressed by Deich and Smith, 1980, and they concluded that indeed this was the case (that the number of molecules taken up was independent of fragment size), but while they do mention USS, they do not bring up USS density as a potential reason for their data.

I’d hoped that by doing a second DNA concentration (100 ng) that I might get hints as to which of the two above models is correct (or if they are both correct and both contribute to the observation), but I don’t really think I can say too much without repeating this several times and getting some error bars on that graph. Furthermore, I’m not really entirely sure what the expectations are for the two models. I’ll have to think on this some more....
-----
Okay, what about transformation rates using my “biorupted” fragments? Below are two graphs reporting the transformation rates of the KanR and NovR alleles from MAP7 to KW20 for the three different DNA pools (LARGE, MEDIUM, and SMALL):
(Note: In the case of the NovR/CFU SMALL sample, the number reported is actually the limit of detection, so NovR/CFU(small) is less than 4.4e-6. I didn't observe any NovR transformants for the SMALL fragments, despite having a decent limit of detection.)

First, I'll look at the difference between the LARGE and MEDIUM fragments. Both markers showed ~6-fold decrease in transformation rate in the briefly sonicated sample, compared to the large intact fragments. Possible explanations:

1. Less USS per fragment: I doubt this is a major reason for the difference. While a larger percentage of fragments are expected to contain no USS in the MEDIUM sample, it shouldn’t be that large of a difference, since the mean density of USS motifs is ~1kb.
2. Degradation by translocation or cytosolic nucleases: This seems to be a reasonable explanation. From an old set of experiments using a defined plasmid donor, Pifer and Smith, 1985 estimated that an average of ~1.5 kb of a leading 3’ end is degraded during translocation. Maybe the medium-sized fragments simply don’t survive translocation as well as large taken up fragments.
3. Recombination efficiency: Maybe both the LARGE and MEDIUM fragments make it into the cytosol, but homology search and recombination are much better for larger fragments.

Things look a little more interesting when looking at the change between MEDIUM and SMALL fragments: While the KanR rate only changed modestly (less than 2-fold), the NovR rate went down below my limit of detection. Possible explanations:

1. Distance to USS: The nearest USS to the NovR allele is more than 4oo bp away, but less than 400 bp for the KanR allele: I like this explanation. I really really need to figure out the identity of these antibiotic resistance alleles. We’re pretty sure it’s a mutation in the gyrB gene, but I don’t know the actual change. I looked at gyrB and it does contain a USS core motif and two other core motifs a few hundred bases before the start codon, but the gene is ~2.5 kb, so the actual gyrB mutation could easily be too far away from these USS. When I looked at the putative gene responsible for KanR (the ribosomal S7 gene), there was a single USS near the start, but none within. Again, without knowing the causative lesion, I can’t tell whether this is within the size distribution of the SMALL fragments.
   
2. Differences in degradation rates at the two loci: This is possible. The other thing these data suggest is that, despite the ~1.5 kb average degradation reported by Pifer and Smith, 1985, there’s still plenty of small fragments that can recombine, since the KanR rates between MEDIUM and SMALL are not really dramatically different.
   
3. Recombination signals: Also possible. And probably the hardest to tell, since the other effects need to be canceled out.

The main take-away (assuming that the results replicate) is that not only do different markers transform at different rates, the change in transformation rate for different sized-fragments also varies for different markers. The underlying reasons for this are probably interesting. So what next? I need to repeat this, but next time I’d like to:

1. Extend this to additional markers to see if this variability also applies to other loci.
2. Measure linkage between Kan and Nov. I’ve previously seen the known linkage between Kan and Nov using large fragments, but would expect linkage to vanish for small fragments when the KanR and NovR alleles never share the same fragment.

And again, I really need to know what the lesions are that are responsible for the MAP7 antibiotic resistances. I’ve looked around a fair amount, but it seems that many antibiotic resistances can be produced by mutations in more than one different gene, so narrowing it down isn’t that straightforward.