It sounds like you could be using excess glycogen, or the sample itself may be inherently rich in glycogen. Since glycogen is insoluble in isopropanol it can co-precipitate with the RNA. This could cause some problems down the line from a quantitative standpoint, since glycogen can skew the nanodrop data (230nm, see link). You could try increasing the amount of sample used in the trizol extraction and omitting glycogen.

https://assets.thermofisher.com/TFS-Assets/CAD/Product-Bulletins/TN52646-E-0215M-NucleicAcid.pdf

Dave Unis

Interesting. I'm using 30 ug of GlycoBlue. I didn't think that was that excessive.

The samples are bacterial, but I suppose they could be rich in other kinds of sugars. The 260/230 was like 1.97, so not that bad.

Have you tried increasing the centrifugal time/speed during the trizol extraction? That seemed to improve my RIN values, and 260/230. That being said a 260/230 of 1.97 sounds pretty good to me! What's your downstream applications? cDNA and sequencing?

Dave Unis

cDNA for qPCR, and maybe RNA-seq in the future.

If this was my only problem (a pellet smear), I wouldn't mind. The purity (by nanodrop) and the RIN values were excellent. It also looked great on an RNA gel.

But my TRIzol preps are also failing a lot and I don't know why. ~3/4 give me 0 ng/uL, while ~1/4 give me perfect products (like this one).

This seems to only happen with my bacterial samples though. I can mix purified RNA into TRIzol and everything works fine.

Either there's some contaminant inherent to my bacteria, or the lysis is incomplete. I always lyse exactly the same way, so it's hard for me to believe it's lysis.

So I suspect my bacteria have some contaminant (sugar, lipid, RNase, etc), but I don't know what. I thought the pellet smear might be a hint as I've never seen anything like it with alcohol precipitation methods.

Have you tried using a kit, instead of trizol, to extract the RNA?

https://www.takarabio.com/products/nucleic-acid-purification/rna-purification-kits/total-rna-from-cells-and-tissues/nucleospin-rna

I've tried the PureLink RNA mini kit. It takes the aqueous phase of TRIzol and passes that over a column. Same results: Mostly fails (with bacterial samples), but occasionally works perfectly (high yield, high purity, high RIN).

Perhaps try one that avoids trizol all together.

Ultimately, I have to use TRIzol+chloroform for biosafety reasons.

But I can try to prep similar, nonhazardous samples with non-TRIzol methods just for the sake of troubleshooting.

Since it sounds like you're looking to quantify a trancriptional response. You could try running the aqueous phase of trizol through a PolyA kit. https://www.neb.com/protocols/0001/01/01/isolation-of-mrna-s1560

First off, Poly(A) won't work for bacteria. Second, does the smeary-ness of the pellet correlate at all with yield? If not, I wouldn't really imagine they were related. What is a "good" yield for these samples? If it's high, you may not need to add glycogen anyway.

They don't polyadenylate mRNA.

Sorry, forgot that the sample was from bacteria; however, it is still possible to specifically isolate bacterial mRNA https://www.thermofisher.com/us/en/home/references/ambion-tech-support/rna-isolation/tech-notes/purify-bacterial-mrna.html

That being said, poly adenylation of mRNA does occur in bacteria as well: https://www.ncbi.nlm.nih.gov/books/NBK6253/

Article RNA polyadenylation and its consequences in prokaryotes