I am prepping B. subtilis cultures for RNAseq, and have been comparing different storage and lysis methods.  

I have prepped most of my samples using RNAprotect based on what I understood were well-established protocols from the literature and previous work with E. coli.  However, in pilot samples I happened to discover that when I skip the RNAprotect step and dispatch live cultures directly into Trizol-LS, and then store at -80C, I get much better total RNA yields (>20ug) compared to RNAprotect/-80C storage/Trizol treatment (~5 ug). These values are post-DNAse treatment, and DNAse eliminated about 30-35% of the measured pre-treatment nucleic acid in both cases.

The only downside is that I have to use a much larger volume of Trizol-LS since it's volume has to be proportional to the whole culture rather than the pelleted cells... and the fact that the samples I have already collected and stored at -80C were stored in RNAprotect, not Trizol.  But I am curious why trizol would be so much more effective at recovering nucleic acid from live cells versus dead-but-well-preserved cells, if anyone knows.

UPDATE:  I made a bad assumption when I was comparing these methods - I assumed that bacteria in exponential phase grown in my plate reader would lyse as easily as cells grown to the same density in a test tube.  Turns out the cells grown in my plate reader are much tougher, and do not lyse well in Trizol regardless of whether or not they have been frozen yet or not -  but thankfully lyse just great with a bead beater.