V4 16s Amplification for Sequencing via Illumina Hiseq 2000 -> Multiple bands!

[adw222]

Hello!

We are performing a metagenomics study in which we are characterizing bacterial diversity in clinical bone marrow specimens. I have extracted the DNA using Trizol LS.

We are basing our protocol on Caporaso et al. (2011), using the v4 16s primers 515f/806r. I ordered the primers sans the barcodes and adapters to ensure that we are able to amplify bacterial 16s in our samples prior to committing to this protocol. When I run the 16s PCR on my samples I get multiple bands - the brightest of which is about 100 bp smaller than the amplicon should be. I sanger sequenced this product and it seems to be human mitochondrial DNA. There is also a very faint band of the correct size and a faint band somewhere around 500 bp. I'm very hesitant to move forward with sequencing products containing multiple bands, as I don't want to waste coverage. Has anyone else had this problem with clinical samples?

Many thanks!!

[JamieHeather]

Is it mitochondrial ribosomal RNA? Does it contain sites that could legitimately be primed off your primers? Maybe the microbial or mitochondrial sequences share enough identity.

If the sequence is completely different, you might just generally be getting mis-priming. Can you raise your annealing temperatures any, or maybe try out a touch-down PCR?

[adw222]

Thank you for the advice - I'm going to try touchdown today! I was hesitant at first to troubleshoot the protocol because so many people are using it successfully...

[JamieHeather]

Did the sequences match to mitochondrial rRNA? I'm curious, seems to me it'd make sense that they would be the most likely candidates.

[LVAndrews]

Maybe you've already solved you problems, but I've moved to a slow and cold amplification that is pretty much the ABI procotol for Sanger sequencing for my amplifications with 515/806 primers. As long as I don't overamplify, my bands look good. I would not use more than 20 cycles, but here is the program:

1) 95C 2min
2) 95C 30s
3) 55C 30s
4) 60C 4min
5) goto 2, 19x
6)10C forever

I also have shied away from generic Taq polymerase for fear of creating surprise organisms in my data. My reactions (8uL each, done in triplicate) contain the following:

0.02U/uL Phusion HotStart II polymerase, 1X Phusion HF buffer, 1uM primers (increased to avoid extinction due to degeneracy -- esp the 806 primer), 3.0mM MgCl2, 200nM dNTPs, ~2ng DNA template, 6% glycerol, phenol red to color (for direct gel loading).

I've used touchdown in the past with mixed success. I find long/cold cycling usually gives better results with less non-specific priming. One other thing, when you do start sequencing, consider the results of Berry et al (2011, AEM 77:7846-7849) showing your barcode sequences can generate other spurious products by interacting in an unforeseen way with the rest of the DNA in the sample. First do PCR with untailed 515/806, pool, check on gel, dilute products, then perform tailing PCR with barcoded primers. Our data are looking more consistent with this protocol.

[mrslwszwczk]

Interesting and cool ideas, I like it!