Gel size selection problem (454 marker development)

[maierpa]

Hi everyone,

I will keep my question brief and specific to (hopefully) generate interest, and if you want me to give more background I'm happy to do so.

I am using a modified AFLP protocol to reduce my sample (amphibian) genome size. I hope to sequence 2-5 individuals with this RRL on the 454 Junior platform. My problem is the last step, the selective amplification. For those of you familiar with AFLPs, the protocol is -

1. One-step digestion/ligation - I tried 2 different enzyme combos: EcoRI/MseI, HinDIII/MspI. Adapters are ~13bp.
2. Pre-Amplification - Using EcoRI and MseI primers, ~17bp.
3. Gel size selection and purification - I'm using 1xTBE buffer, 1% gel, 70V (on large gel well), regular agarose (says best for fragments >1kb, but we routinely use it for smaller fragments). I'm purifying with Qiagen gel extraction kit.
4. Selective Amplification - Using longer primers with EcoRI or MseI sequence + MID barcode + fusion primer sequence.

See the 3 gels attached. When I select moderate size fragments, the resulting amplicons are always tiny (1-200bp). Selected size has no relationship with amplicon size. In other words, makes no difference whether I select 500-600 or 900-1000, always comes out 200. We need 400bp. I found one paper describing a downward size bias for my particular taq - Phusion HF - however it's my understanding this taq is standard for 454 library prep. Also, you'll notice wavy bands on all gels - I'm using Gel Red for staining.

What's going on?! Thanks in advance.

[maierpa]

followup

Here's an earlier gel to illustrate the point. This is after the selective amplification, just like gel #3 above.

(To be clear: the gels in the first post are
1. Digestion/Ligation
2. Pre-Amplification
3. Selective Amplification)

In this SA I post-stained with EtBr, instead of prestaining with GelRed. Basically the same result. (The red rectangle shows the expected size, 500-700bp in this case.)

Here are my current ideas:
1. The taq has a huge downward size bias
2. Enough small template is gel selected with the large template to bias PCR
3. My final product is not template at all, but primer-dimer (the fusion primers are 42bp and 51bp)

???

[Uparedes]

AFLP-size selection

Hi,
I have been having the same problems with AFLP, but in my case it is methylation sensitive. I have been cutting with either Eco-MspI and Eco-HpaII.
I have encountered the exact same problem. It looks as if MspI and possibly EcoRI don't produce small enough fragments after restriction digest for some genomes, unless you use ridiculous amounts of enzyme (100-200 molar excess!!) or add spermidine, or some helper oligonucleotides (see papers).

When you tested your genomic DNA, did you test each enzyme independently?
I found either high bands amplified during selective PCR and tiny ones (50-230 bp) but nothing in the middle. The high bands are usually EcoRI-EcoRI products, and the small one are MspI-MspI or HpaII-HpaII. It is worth sequencing the small and large fragments.
A third restriction enzyme which cuts your genome (not methylation sensitive), for for which you don't add any adaptors might help (it would reduce the proportion of fragments of smaller size which will be amplified by your PCR).

Good luck,

Ursula


Effect of cytosine methylation on the cleavage of oligonucleotide duplexes with restriction endonucleases HpaII and MspI. J Ben-Hattar, J Jiricny - Nucleic acids research, 1988 May 11; 16(9): 4160.

Preparation of reduced representation bisulfite sequencing libraries for genome-scale DNA methylation profiling. Nature Protocols 6,468–481 (2011)
doi:10.1038/nprot.2010.190

TE-AFLP: combining rapidity and robustness in DNA fingerprinting
A. W. G. van der Wurff*,Y. L. Chan1, N. M. van Straalen and J. Schouten1
Nucleic Acids Research, Volume 28, Issue 24, Pp. e105