Isolating bacterial mRNA from mixed prokaryotic + eukaryotic total RNA pool

[sweetph3]

My supervisor has already agreed to help a collaborator to sequence total RNA from a pool of plant and bacteria. They want me to do one library set for the plant mRNA (easy enough, capture poly-A mRNA on poly-dT beads), and another library set for the bacterial mRNA. Here are the two options I've come up with for the second bacterial set. I'm hoping someone out there has experience with something like this, and can either give me pointers on these options, or else suggest another option that I haven't thought of. BTW, there are going to be a LOT of samples (time points, triplicates, different temperatures...)

1. Bacterial mRNA can't be pulled out of a mixed pool, so we'd first have to pull out the poly-A plant mRNA (using poly-dT beads from Illumina kit), then pull out the plant rRNA (and plastid RNA) using either RiboMinus Plant kit (Life Technologies, $562 / 8 preps) or RiboZero rRNA Removal kit-Plant Seed/Root (Epicentre, $554 / 6 preps), then pull out the bacterial rRNA using either RiboMinus Transcriptome Isolation Kit - bacteria (Life Technologies, $493 / 12 preps), MICROBExpress bacterial mRNA enrichment kit (Life Technologies, $420 / 20 preps), or RiboZero Magnetic Kit - bacteria (Epicentre, $1968 / 24 preps). This is both labor-intensive and expensive ($91 - $175 per sample).

2. Pull out the poly-A plant mRNA, then make cDNA from this RNA pool (bacterial mRNA, bacterial rRNA, and plant rRNA), use DSN to remove the abundant ribosomal fractions, and make libraries from what remains. This may not be very efficient.

I'll continue to search for other options, but so far the only other one I've found (a kit called mRNA-ONLY) is no longer available, so it probably wasn't very efficient.

Thank you very much for any help!

[sfranzenburg]

Hi,
I basically tried exactly that with gut bacteria from Insects.
I pulled euk. mRNA with Poly A and removed euk. and prok. rRNA with the Ribozero Kits.
The result was disappointing, as still a tiny fraction of reads was prok. mRNA.

So you really have to do the math first, considering several factors:
1. genome size of your host (plant) vs. average bacteria
2. Number of bacterial cells per host cell (e.g. 10:1 as a typical gut ratio)
3. Ratio of rRNA to mRNA (~97:3)
4. Efficiency of the Ribozero Kits

Just do an excel file and assume a fixed number of reads and distribute them to your mixed sample according to your ratios. Then step by step remove what you expect to remove with the kits.
In my case it was like this:
If I assumed 10:1 ratio of bacteria to host cells, rRNA to mRNA of 97:3 and a Ribozero Efficiency of 99% (a claim from Ribozero and probably toooo optimistic), I ended up with ~15% of reads to be bacterial mRNA (in theory). However, the reality was ~1%, which is very easy to get if you go down with the RiboZero Efficiency to more reasonable numbers like 90%...

Conclusion: In order for this to work well, a high ratio of bacterial to eukaryotic cells/DNA is desired. Otherwise, the remaining euk. rRNA (and tRNA, small RNAs etc) will easily outnumber your tiny fraction of prok. mRNA (especially since prok. mRNA has a much shorter lifetime than euk. RNA).

I hope this is helpful for you.

[Simone78]

Hi,
I never worked with bacteria or plants but I can say something about the Ribozero kit: it sucks when working with low-inputs! Maybe in your case this is not an issue.
However, a far better way to get rid of your rRNA which doesn´t require many purification steps is the so-called SD-RNA method (Morlan, Qu & Sinicropi, PLoS One 2012). If the genome of interest has been sequenced you can easily design your rRNA probe set (and it is much cheaper than buying any kit). Clontech and NEB (and maybe even some other companies) have now packaged this in their RiboGone and NEBnext rRNA depletion kits, respectively...asking a dear price for it.
As pointed out above, with Ribozero you´ll get A LOT of reads from rRNA (in Adiconis et al., Nature Methods 2013 they got up to 25% when working with tumor samples and degraded RNA!), while with the SD-RNA method this figure can be as low as 0.1%.
Good luck!

[MLog]

Hi Simone78,

do you have any examples of Bioanalyzer traces of RNA samples after SD-RNA depletion? I tried this protocol (as implemented in NEBNext rRNA removal kit) and got some strange results - very short RNA fragments (samples 3-6). I did a depletion of the same RNAs with Epicentre RiboZero (samples 1,2), and is this case the results were more typical.

[Simone78]

Hi Simone78,

do you have any examples of Bioanalyzer traces of RNA samples after SD-RNA depletion? I tried this protocol (as implemented in NEBNext rRNA removal kit) and got some strange results - very short RNA fragments (samples 3-6). I did a depletion of the same RNAs with Epicentre RiboZero (samples 1,2), and is this case the results were more typical.

Unfortunately I do not have any Bioanalyzer trace. However, I tried the Ribozero kit on FFPE-samples and I lost everything, even the non-rRNA due to the many steps of the protocol. And, since the sequence of the probes for the SD-RNA depletion is published, I thought of doing my own set of probes instead of buying the kit...but I didn´t have time to go on with that project so far!

[markopetek]

I'm also planning to so some metatranscriptomics of insect gut content. Did any of you guys find a good protocol to get rid of the host(eukaryotic) RNA? In my case I would have insect and plant RNA contaminations since the insect is a herbivore.

[Johnwang]

Extract RNA from single pollen

Hi guys, is there any good RNA extraction method for single pollen ? Thanks.

[lsherlin@NuGEN]

It is possible to perform depletion of transcripts after library preparation, which could be a good option if an upfront rRNA pulldown is leaving you with too little RNA material to work with. There are methods that create libraries from total RNA or enriched RNA and perform targeted depletion of transcripts the final library. This process is also amenable to picogram inputs and can be customized for various targets in a mixed system, such as host-pathogen or 'host-symbiote'. For example, targeting host rRNA transcripts can often remove 80% or more of the host transcripts and create more flowcell space for the reads of interest; other prominent transcripts can be depleted as well.

Below are links to the available NuGEN systems that use this approach; there may be others as well to look into.

http://www.nugen.com/products/ngs/ovation-solo-rna-seq-system

http://www.nugen.com/products/ngs/ovation-universal-rna-seq-system