Bacterial scRNA-seq experience?

[nucacidhunter]

Maybe so, but it doesn't appear to be possible in my hands using superscript II and the conditions from the SMART-seq2 protocol. Maybe some proprietary additives in the Clontech buffer makes a big difference?

Using other reverse transcriptase with strand switching activity (SMARTScribe, Maxima H Minus) might give better results. There are some threads in this site as well under single cell RNA-Seq headings that has discussed some issues with SSII.

[AndrewP]

Maybe so, but it doesn't appear to be possible in my hands using superscript II and the conditions from the SMART-seq2 protocol. Maybe some proprietary additives in the Clontech buffer makes a big difference?

[cmbetts]

Template switching definitely doesn't need a 5' cap to work. Clontech's (Takara) SMARTer stranded kits all use template switching on chemically fragmented RNA with random priming. The internal fragments don't have caps and template switch just fine. The CATS method also uses template switching on uncapped RNA fragments. You can even template switch DNA under the right conditions.
There are definitely reasons the scRNA-Seq on bacteria will be extremely challenging, but it's not the lack of 5' caps. Without a polyA tail to select for mRNA, you'd need to do something more like the the SMARTer-stranded pico kits with post library rRNA removal.

[AndrewP]

In case anyone is trying something similar and finds this thread. Template switching is not possible in bacteria because they lack a 5' cap that RTase requires for terminal transferase activity. I didn't think of this because most papers since ~2008 don't mention this fact, but the whole SMARTseq idea originated from finding transcription start sites of 5'capped mRNA.

See this great paper by Pacbio and NEB that find a workaround (unfortunately not useful for my application).

https://www.nature.com/articles/s41467-018-05997-6

[AndrewP]

Why are you working at 37C? Most TS I have seen is around 50C, this might help to prevent any mismatching of your TSO or RT primers.

Most TS is at 42 degC which is the optimal temperature for superscript II, I don't think the 5 degC makes much of a difference, but it fits in with other aspects of my workflow including better initiation with random hexamers.

[seqsuave]

Why are you working at 37C? Most TS I have seen is around 50C, this might help to prevent any mismatching of your TSO or RT primers.

[AndrewP]

Thanks for the advice, I'll give that a try. I should have done that the first time around, but was hesitant because of conditions of the experiment. Any thoughts on RT primer and TSO concentrations? I'm thinking of trying 10 uM random hexamers/decamers and 2.5 uM TSO (there's some hybridization between the RT primer tag and the TSO, but shouldn't be a problem at 42 degC). I'm not concerned with TSO concatenation.

Actually, I have a specific question. Let's say I have a gene that is 500 bp, and a random hexamer hybridizes at the middle and at the end and is extended by the RTase. Is SSII strong enough to extend the end primer and displace the DNA-RNA hybrid that starts in the middle? How about if the random hexamers have 1 or 2 LNA modifications?

[cmbetts]

I'd start with as SMARTseq2 like protocol as fits with your design and optimize from there.
I can't go too much in the black magic parts (former Clontech/Takara employee, and still friendly with them) beyond pointing to the SMARTSeq2 protocol as a good jumping off point if you're working on a novel protocol.

[AndrewP]

Are you using default SSII buffer? MgCl2 levels are very important for TS, and higher than most standard buffers.

Thanks for the feedback cmbetts. I used the standard SSII buffer that has 3 mM MgCl2 because I've noticed some TS methods don't add additional MgCl2. I am thinking of adding 9 mM MgCl2 with 1M Betaine similar to the SMART seq2 protocol. Think that's a good idea?

[cmbetts]

Are you using default SSII buffer? MgCl2 levels are very important for TS, and higher than most standard buffers.

[AndrewP]

Hi all,

I'm trying to perform a scRNA-like library prep in bacteria, so I cannot take advantage of the polyT reverse transcription primers. Instead, I'm using tagged random hexamers/decamers. I also need to use a relatively long template switching oligo that is about 90 nt and contains 20 random bases as an UMI.

So far, I haven't had a lot of luck with high efficiencies (I would estimate I'm generating less than one cDNA fragment per cell).

I'm working on optimizing everything, but I'm hopeful that someone has some insight into what is the limiting step? Lysis is definitely not an issue, but RNA is possibly modified by formaldehyde. RNase contamination is very unlikely

My RT reaction
Superscript II (10U/uL, standard SSII protocol -> is it detrimental to use too much RTase relative to RNA input?)
1 mM dNTP
1 uM TSO
2.5 uM Random hexamers
10 mM DTT
20 ug/mL BSA
Enzymatics RNase inhibitor

Leave 1 hour at room temperature (to facilitate hybridization of random hexamers), then overnight at 37 degC. I know this is unconventional, but my system is quite a bit different than normal...


I appreciate any and all advice/help!