Do you have any marker which is specific for your cells? Then you can enrich it using antibody coupled magnetic beads and columns (Miltenyi has a wide range of products here) wash the cells carefully on the column and the elute only your cells of interest and work on with them. This is widely practized in immunology. You could also sort the cells using FACS.

If you are interested in gene expression, then you can separate your host mRNA from the bacterial RNA by using polyA tail-based techniques (different kits available).

Thanks.

Christian, wouldn't this isolation process lead to changes in gene expression? How much time does it take. I am interested in differential gene expression between two groups, as it occurs in the peritoneal cavity (where there is considerable stimulation occurring.)

Using polyA tails is the way. However, the separation is not usually necessary for mRNA analysis (e.g. uARRAY or qPCR). We were doing analysis of host mRNA transcriptome profile and there were no problems even when the samples contained large amount of bacterial RNA. In majority of microarray analysis cases, the host mRNA is amplified using polydT-T7 promoter primers. Just be sure that your samples contains enough of host RNA (e.g. by analysis of proportion of host 18S and 28S rRNA vs bacterial rRNA). Good luck.

I would do the same, bacterial RNA is not polyadenylated!

Separating the cells prior to the analyzes is the best way to go but if you are concerned about changes in gene expression I agree with Rastislav. Look at the amplification step for the array platform you plan to use. If it uses poly-dT it will amplify the mammalian RNA preferentially. For real-time PCR the same applies. Use poly-dT for your primer during the RT reaction. Also make sure that your primers will be specific for the mammalian gene sequence that you are interested in. You could use random hexamers in the RT also but then you DO have to be ABSOLUTELY sure your primers only detect the mammalian sequence and not the bacterial sequence.

Thank you all for the good information.

I will be using Affymetrix Mouse Gene St 1.0 arrays, so i do not believe any specific primers will be used.

The poly-a purification seems like the way to go, can anyone suggest a kit? Will these kits retain the rRNA? I'm under the impression that they (microarray core) use the rRNA peaks to normalize the amount of input RNA across samples.

@Evan: Ok, for gene expression analysis I wouldn't use it. It will take too long and the cells will change their gene expression pattern. The poly-A route should work pretty good as well.

Just for the sake of correctness, as I don't think that this will influence the outcome: bacterial mRNA can be and often is polyadenylated. Not always, and not every mRNA, but there is an ubiquitous mechanism of RNA polyadenylation in all prokaryotes (both bacteria and Archaea). See for example this review: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3041983/

Other than that, I would do a spike-in experiment with clean RNA samples mixed with bacterial (compared with clean host RNA) to see whether a poly-A route is necessary at all.

January has a point. I agree that low fraction of polyadenylated bacterial mRNA should not influence the experiment. To be sure 100% you can try this scenario :

- isolate total RNA from cultures of major bacteria found in your samples

- process it with a RNA amplification polydT-T7 promoter-based procedure (affymetrix has these types of 3'IVT kits )

- see the outcome.

Already the result of the amplification should yield a low or no amount of aRNA.

Normally all this should be discussed in your genomic core centre.

Ewans, to answer your last question : No rRNA will be present after polyA purification procedure. Please see an attached image of total RNA electrophoresis where two smaller peaks correspond to bacterial rRNA.Two higher are 18S and 28S.

Just to enrich the discussion, it is also known that eukaryotic transcriptome contains non- polyadenylated mRNA (eg. histones). We were surprised to find out that these normally non-polyadenylated eukaryotic mRNA to be expressed in a particular group using microarray (done with polydT-T7 promoter amplification!). Hopefully we will publish results soon :).

Btw I'm looking for a post-doc position in a genomic core centre or genomics oriented research team where I would have an opportunity to learn to work with NGS. If you known about something interesting, please be kind to let me know :).

I've attached a picture of my bioanalyzer results for some samples (labeled EC#), to add a face to all this. You'll notice a range in their appearance, is there anything else that is cause for concern?

Evan, you've got 3 groups of samples. Two samples are heavily contaminated, two moderately and 3 are with very low or no contamination. So actually, you are quite lucky, and probably you will see some kind of "dose" dependent response of leukocytes on the presence bacteria. Upregulation of immflamation-related genes such as interleukins is highly probable in the transcriptome profile. Do you have a CFUs/ml to each samples? Could be quite useful to support. What is your reference ?

I would suggest to continue with the RNA amplification/labeling and if the yield is ok the hyb should be ok as well. I was working with samples that were similar to your second group and everything was fine. I was using different microarrays than you (Agilent). The reference was non infected tissue.

Good luck.