So You Need to Work With Prokaryotic Transcription
I spent about two years running in vitro transcription assays with E. coli extracts before I stopped making the same mistakes repeatedly. The core concept is straightforward, but the details are where people drop the ball. Let me walk through it. Prokaryotic transcription is carried out by a single RNA polymerase, unlike eukaryotes which have three. The enzyme is a holoenzyme composed of a core polymerase (alpha two, beta, beta prime, and omega subunits) plus a sigma factor that handles promoter recognition. Without the sigma factor, the core enzyme can still catalyze RNA synthesis but it binds DNA non-specifically. That distinction matters when you're setting up a reaction because you need to know which sigma factor your system is using. The process breaks into initiation, elongation, and termination. Initiation starts when the sigma factor guides the holoenzyme to a promoter, typically at the -10 region (the Pribnow box with the consensus sequence TATAAT) and the -35 region (TTGACA). The DNA melts to form an open complex, and the first phosphodiester bond forms. Then sigma releases, the core enzyme proceeds into elongation, and RNA grows in the 5' to 3' direction at roughly 40 to 80 nucleotides per second under optimal conditions.
Termination happens through one of two mechanisms. Rho-independent termination relies on a GC-rich hairpin followed by a poly-U tract in the RNA transcript. The hairpin stalls the polymerase and the weak A-U bonds between the transcript and template strand let the RNA fall off. Rho-dependent termination involves the Rho protein, a hexameric helicase that loads onto the RNA at a rut site, chases down the polymerase, and unwinds the RNA-DNA hybrid. I've seen more grad students mess up their termination controls by assuming every gene uses the rho-independent mechanism. It's not true. Some of the most highly expressed genes in E. coli are rho-dependent terminators. Here is a practical problem I ran into that took me three weeks to diagnose. I was running an in vitro transcription reaction using a cloned gene insert and getting full-length product in the expected size range. Great. But when I checked the amount of product on a quantitative basis using a spectrophotometer, the yield was dramatically lower than the literature values for that particular template. I assumed I had a bad enzyme prep. I re-made the reaction from fresh components and got the same thing. It turned out the template I was using had a very efficient rho-independent terminator right at the end of my insert, and the Rho-dependent machinery in the crude extract was reading through that terminator and continuing to transcribe well past my insert into the vector backbone. The product on my gel looked right because the full-length product dominated the band, but I was actually generating a heterogeneous population of RNAs of varying lengths, and the actual yield of my desired transcript was maybe a third of what I thought. The workaround was straightforward: I switched to a template where the terminator was placed downstream of a multiple cloning site that had no additional promoter-readable sequences, and I added a Rho inhibitor like bicyclomycin to the reaction when I needed to study that particular termination event. That cut my confusion in half immediately.
Setting Up a Transcription Reaction
If you're working with purified systems, you need the core RNA polymerase, the appropriate sigma factor, NTPs, a DNA template, and a buffer. A typical 20 microliter reaction contains about 40 millimolar Tris-HCl at pH 7.9, 100 millimolar potassium glutamate, 10 millimolar magnesium chloride, 1 millimolar DTT, 0.5 millimolar each NTP, and roughly 50 to 100 nanograms of supercoiled plasmid template per microliter. Incubate at 37 degrees Celsius for 30 to 60 minutes. Supercoiled templates transcribe more efficiently than linear ones because the topological stress helps with promoter melting. But supercoiled plasmids also tend to form knots and catenanes during the reaction, which can complicate downstream processing. If you need clean product for things like in vitro translation or RNA structure studies, linearize the template downstream of your transcript to prevent the polymerase from running around the circular plasmid and creating long concatemers. That's a standard move but people forget it sometimes. When working with crude extracts, the composition is more complex. An E. coli S30 extract contains the endogenous RNA polymerase along with all the transcription factors, ribonucleases, and metabolic enzymes naturally present in the cytoplasm. The advantage is you get the full physiological complement of proteins. The disadvantage is that RNases are also there, and they will chew up your product if you're not careful. I usually add an RNase inhibitor like RNasin to my reactions, and I keep everything on ice until the moment I start the incubation. The extract itself is typically prepared by lysing cells in a high-salt buffer, spinning down the debris, and storing the supernatant at minus 80 degrees. Freshly prepared extract gives better yields than thawed aliquots that have been through multiple freeze-thaw cycles. One freeze-thaw cycle typically reduces transcription activity by about 15 to 20 percent.
Promoter Selection and Its Consequences
The choice of sigma factor determines which promoters your RNA polymerase will recognize. The primary housekeeping sigma in E. coli is sigma 70, also called RpoD. It recognizes the canonical -10 and -35 boxes I mentioned earlier. But there are alternative sigmas for stress response and stationary phase. Sigma 32 handles heat shock promoters, sigma 54 handles nitrogen metabolism, and sigma 28 handles flagellar genes. If you're cloning a gene under control of a sigma 70 promoter and you're also expressing a stress response gene in the same cell, the endogenous sigma 32 can compete for core polymerase binding and reduce your yield. This is one of those subtle interactions that shows up when your expression drops unexpectedly during a co-culture or dual-expression experiment. Inducible promoter systems are the standard approach for controlled transcription. The T7 promoter system is the most widely used, but T7 RNA polymerase is technically a phage enzyme, not a native prokaryotic one. Native systems like the lacUV5 promoter or the lambda P-L promoter are options if you want to stay within E. coli's own machinery. The lacUV5 promoter is constitutively active at a moderate level and can be induced with IPTG. The lambda P-R promoter requires the N and Q antitermination proteins for read-through, which adds complexity but gives you tight temporal control. The tradeoff is that tighter control usually means more components to manage and more places where something can go wrong.
Common Pitfalls and What Actually Fails
Contamination is the most common issue. A single pipette tip touch can introduce RNases that degrade your template or your product. I've lost entire batches of sensitive in vitro transcripts because someone set the tube down on a bench that had been used for PCR cleanup earlier that day. RNases are everywhere and they're extremely stable. They survive autoclaving. The only reliable way to eliminate them is to use DEPC-treated water and dedicated consumables, or to buy certified RNase-free supplies. DEPC treatment requires a 30-minute incubation followed by thorough autoclaving to remove residual DEPC, which is carcinogenic. I stopped doing my own DEPC treatment years ago and just buy pre-aliquoted RNase-free reagents. It costs more but it saves time and it stops me from second-guessing every failed reaction. Magnesium concentration is another variable that people don't pay enough attention to. The optimal MgCl2 concentration for E. coli transcription is around 10 millimolar, but it shifts depending on the NTP concentrations and the salt environment. Too little magnesium and the polymerase can't catalyze phosphodiester bond formation efficiently. Too much and you get non-specific transcription from weak promoters. I usually titrate magnesium from 5 to 15 millimolar in 2 millimolar increments when I'm setting up a new reaction condition. It takes an extra gel but it prevents a lot of wasted effort later. One thing that surprises people is that prokaryotic transcription is coupled to translation in vivo. The ribosome follows closely behind the RNA polymerase and this coupling actually affects transcription dynamics. Ribosomes prevent premature termination at rho-dependent sites by physically blocking the rut site on the nascent RNA. In an in vitro system without ribosomes, rho-dependent termination can be much more efficient than you'd expect from cellular conditions. If your in vitro results look qualitatively different from what you'd predict from the genomic sequence, the lack of coupling is probably the reason. This is especially relevant when you're using templates derived from native operons that contain internal rho-dependent terminators. Those terminators that are leaky in vivo because of ribosome coverage can become near-complete blocks in a coupled transcription-translation system or completely efficient in a pure transcription system.
Prokaryotic transcription as a system has real limitations that you should factor into your experimental design. The single RNA polymerase means there's no compartmentalization of transcription and RNA processing. There is no splicing in the eukaryotic sense, but there are exceptions. Some bacteriophages and a growing number of bacterial genes encode self-splicing introns or use trans-splicing mechanisms. If you're working with unusual organisms or phage systems, don't assume the transcript is immediately ready for translation. RNA stability is also a major constraint. Prokaryotic mRNAs typically have half-lives ranging from 30 seconds to a few minutes, compared to hours for eukaryotic transcripts. This means your transcription product degrades quickly even in controlled conditions if any RNase activity remains. The degradosome, a multi-enzyme complex containing RNase E, PNPase, and other degradation machinery, is always present in crude extracts and it will process your RNA if given the chance. If you need large-scale production of a specific transcript, the T7 promoter system with purified T7 RNA polymerase and a linearized template is usually the best option. It gives you yields in the range of 50 to 200 micrograms per milliliter of reaction, which is significantly higher than what you get from the native E. coli polymerase system. The main drawback is that T7 RNA polymerase has its own promoter specificity and won't transcribe from native E. coli promoters without engineering them into your template. For studying regulation, native systems are more appropriate despite the lower yields. For structural or functional RNA studies where quantity matters, the T7 system is the practical choice. I typically use the T7 system for anything where I need more than 10 micrograms of clean transcript and switch to the native system only when I need to preserve native promoter architecture or study sigma factor competition. The take away is that prokaryotic transcription is mechanistically simpler than eukaryotic transcription but experimentally it requires attention to template topology, termination mechanisms, and contamination control that newcomers often underestimate. The coupling between transcription and translation, the variety of termination pathways, and the fragility of RNA products are the three areas where the literature tends to gloss over the practical difficulties. If you account for those explicitly in your setup, your results will be more reproducible than most people who treat it as a straightforward enzyme-substrate reaction.
Get the Full Details
