E. coli has roughly 1000X less RNA per cell than mammalian cells do.  So it seems hopeful that you can merely multiply the number of suggested cells by a factor of 1000 to obtain your ends with E. coli bacterium - if indeed the assumption is correct that the unidentified cell type used as an example in the product literature was indeed mammalian.  In your case, use 1010 cells instead of 107... and proceed as usual. RNAzol prides itself on being able to separate gDNA away from your aqueous RNA fraction...so it is up to you as to how gingerly you gather those RNA-containing water layers. You don't have the luxury of intron-exon differences between gDNA and RNA in your model - so it is a large concern.  E.g., why not DNase-treat just in case? Technique is all.

Hi Jack M Gallop, I also thought to do the same as you suggested i.e. to increase the number of cells. But as a result I got gDNA band on my gel. I guess this happened because the ratio of reagent and number of cells is very important. Eventually when we take large number of cells, RNAzol solution doesn't have sufficient amount of guanidine thiocyanate etc to precipitate all gDNA. There is also a possibility that the RNAzol is not sufficient to lyse this much cells, so some cells won't even break up.

I said about mammalian cells, because my friend uses this for RNA isolation from mammalian cells, and it works perfectly fine with 107 cells according to the protocol. Do you know some one who uses RNAzol method for isolating RNA from E.coli ?

Have you compared TRIzol to RNAzol in your application?  Perhaps TRIzol may still be better.  Sounds like you need to find the golden ratio (which can be done in one shot just using different amounts of cells, etc).  When you tried this before, how far did you increase the (# of cells)/(mL RNAzol)? 1010cells/mL? There is a point at which you can get this to work.  Probably also depends on how you measure the number of cells you're using.  Turbidity assays are off by quite a bit sometimes, and O,D. measurements at 600 nm or so are non-linear after an O.D. of ~0.8 or so...  Good Luck~!

I have used several methods for bacterial RNA extraction and I always finish the protocol treatment samples with DNase (Ambion) because always you can find contaminating DNA. Depending on your downstream applications it will be more or less important to remove it. So, I always treat RNA with DNase and then check samples by doing a PCR expecting to find no band, what would suggest that total DNA was removed.

Hi Dr Jack and Anna Fabrega, thanks a lot for your valuable suggestions. I am now able to extract enough amount of RNA by combining few steps of trizol method with RNAzol protocol (optimized the golden ratio also....thanks Dr Jack), that too without gDNA contamination (at least gDNA is not visible in gel picture). Anna, my downstream work includes converting to cDNA and then qRTPCR, although the RNAzol claims that this extracted RNA can be used directly for qRTPCR, shall I go ahead for qRTPCR or include DNase treatment also. I am doubtful because I have slightly modified the protocol.

Hi Ashish, every time that I have to do an RT-PCR I treat samples with DNase. Now I'm using an automated system for RNA extraction, before I used the RNeasy Mini Kit (Qiagen), and previously Trizol or an homolog, and I always decontaminated samples by DNase treatment, particularly if you want to do an RT-PCR, because just a small amount of DNA can be amplified and vary your results.

If you prefere you can check first your samples, just do a PCR from your RNA extractions to see if you get a band or not, or if it is intense or just a small band. Then, you could think about the need of further DNAse treatment.