Maybe I’m going through PCR withdrawal? I was doing a lot of cloning towards the end of my PhD and would often think of side experiments I would love to do if I could get my hands on the reagents (i.e. rat brain.) Ironically, now that I have all the brain slices I could want, it’s hard to get away from the ephys rig long enough to clone stuff. Still, one thing that always fascinated me was studying non-ionotropic receptors in frog oocytes. You accomplish this by co-expressing** a GPCR and a g-protein inwardly rectifying potassium channel, or GIRK, (e.g. the human H3 receptor + GIRK1+GIRK4) so that the GPCR’s activity can be indirectly monitored via the downstream GIRK response. That’s cool, but if you don’t want to optimize your coexpression system, an alternative approach is to create chimeric receptors: a receptor that is part g-protein of interest, part GIRK. Plus, I recently read a great review on “Tethering-based chemogenetic approaches for the modulation of protein function in live cells” which features cool biochemical approaches (bioorthogonal chemistry, genetic code expansion, cysteine modification) and some nice genetic approaches (validated with patch-clamp!) This post is part molecular biology review, part exercise in experimental design*** for optimizing fusion constructs.

Polymerase Incomplete Primer Extension (PIPE), or incomplete PCR (iPCR)

Figure 1. The chain reaction cycles of a PCR experiment. For dramatic/artistic effect, I’ve exaggerated the incomplete-ness of elongation. Created/adapted with BioRender.com

A traditional PCR experiment amplifies a gene; simple enough. Your protocol is straightforward too: melt your DNA (Step 0), then initiate repeating cycles (18-35X) of melting (Step 1), annealing (Step 2), and elongation (Step 3) or alternatively, just cycle the melting and elongations steps (2-step PCR.) Your final step before the end of the reaction will be usually be a final, extra-long elongation step, just to make sure that all of your amplicons are completely elongated (see Figure 1.)

Say I decide to toss out the final elongation step though; what happens? Obviously, I risk my amplicons not being blunt, but if I’m doing sequence and ligation independent cloning, I’d be chewing away the 5’ and 3’ DNA ends anyway. With this in mind, I can skip an intermediary cloning step (T4 DNA polymerase) and simply run an incomplete PCR; I just have a decision to make on quickly I want to proceed. If I feel like living life on the edge and need this done yesterday, statistically I can transform without DpnI digestion and DNA clean up (in my experience, IF I used < 5 pg of template plasmids for PCR). Assuming I get a decent yield (and good transformation efficiency for my desired backbone), I’d use a small volume of my iPCR to transform into bacteria (for in vivo assembly, I swear by Zymo Mix n’ go competent cells.) The fact that I’m using small amounts of the total PCR, added to the fact that I’m “diluting” my original plasmid (more like reducing the likelihood of transforming my starting material) means that I have a pretty good chance of getting my desired ligation in a colony, I just have to make sure to set up a colony PCR in parallel when inoculating minipreps (i.e. dip the inoculating loop into the PCR mix before dipping into LB/TB/whatever) so I know which samples I should send for sequencing the next day. I personally love Thermo’s 2X SuperFi mastermix (aliquot for fewer freeze-thaws tho) and can confirm you can successfully run 6-10 uL colony PCRs. This paper is a great resource for deciding which cloning method works best for what situation.

Sequence and Ligation Independent cloning (SLIC)

I love SLIC and think it’s super cool (see my original post.) Plus, it works beautifully with Zymo Mix’n Go cells: no more shaking at 37*C for 30+ min like chemical transformation (and no 2-mercaptoethanol smell), and is less stressful than electroporation**** (no current arc anxiety!!) It’s also appealing because once you get it off the ground, it’s easy, which is important. Not because I’m lazy, but because “tedious-ness” is a serious barrier to doing science. Say I wanted to test/optimize the best linker for a fusion construct. Traditionally, I would have to engineer (via PCR) restriction enzyme sites into the “insert” and the “backbone”, restriction digest each of them, then do a DpnI digest, clean up again, ligate for a few hours, then transform. And that’s IF the receptor isn’t chewed up by common restriction enzymes (RIP human GluN2B; ask me how I know :’()

SLIC also lets you correct engineering oversights, like limited/non-existent restriction enzyme sites, and lets you design modular workflows. If you need things done faster and money’s not an object, then you technically don’t even need to reamplify genes with new primers in the case of Gibson Assembly; according to NEB, their HiFi 2X mastermix is capable of using ssDNA (e.g. a primer!) to fuse two non-homologous ends together. Since you can usually get primers <60bp synthesized and delivered next-day (thank you IDT), you can have an easy time fusing distinct proteins and quickly optimize the spacer/linker sequences.

Site-directed mutagenesis (SDM)

SDM is a staple molecular biology technique. In the beginning of my PhD, I wanted to try a few mutations to the GluN1 and GluN2 receptors, so I looked through the old lab protocols for guidance. The tried and true lab method was Agilent’s QuickChange lightning SDM kit, but for a few reasons, I had trouble getting it to work: I was still relatively new to cloning, the laboratory’s molecular biologist who optimized the protocols was gone, and I had way too many things to do at the time (this was just before I did a summer sabbatical at UIC.)

One year later though, I had introduced new reagents (SuperFi Polymerase, Zymo Mix’n Go cells) and methods into lab (Geneious for sequence analysis, SLIC), so I was much more comfortable trying SDM again. At the same time, my molecular biology skills were giving me a new appreciation for quality by design (QbD.)

QbD was formally discussed in my drug discovery and regulatory science classes in the context of target product profiles (TPPs), quality assurance (QA) and product development. This article describes QbD as “… a systematic approach to product design and development. As such, it enhances development capability, speed, and formulation design. Furthermore, it transfers resources from a downstream corrective mode to an upstream proactive mode.” Maybe I’m reaching, but that sounds like a good molecular biology philosophy; better than “redesign your primers and try again” anyway.

What does this have to do with site-directed mutagenesis? Well by this time, I had a very consistent, very simple method of cloning: SuperFi polymerase. Rather than try to get the old mutagenesis protocol working, I decided to look into a new one; something with as few moving parts as possible. Again, SuperFi Polymerase came in clutch, with two ways to introduce mutations: with overlapping primers and non-overlapping primers.

Figure 2. A schematic from ThermoFisher showing you how to design overlapping primers for site directed mutagenesis using SuperFi II polymerase. Notice that your point mutation primers will be <30bp. I’ve gotten 72 bp primers from an SDM kit before.

Figure 3. A schematic from ThermoFisher showing you how to design non-overlapping, phosphorylated primers for mutagenesis using SuperFi II polymerase. Consider this approach for more challenging mutations.

For simple point mutations, overlapping primers (Figure 2) work great; changing a single base in the middle of a pair of complimentary primers won’t kill your PCR (i.e. prevent annealing) and its easy enough to do, even without DNA analysis software like Geneious or SnapGene. For modest insertions (20-40 bp) and most deletions, the non-overlapping method works well; design phosphorylated primers that aren’t complementary to each other that carry the insertion/deletion. The fact that they’re phosphorylated means that you can proceed with cloning via the traditional blunt-end route. While phosphorylated primers are a little more expensive and may take longer to synthesize, you could always just phosphorylate the primers yourself. Plus, Superfi II has a universal annealing temperature of 60°C!

Of course, if you’re comfortable, you can have fun mixing and matching methods. And as always, determine how fast you need to proceed: do you DpnI digest, clean up, nanodrop, and run it on a gel before transforming, or is visualizing the correct band on an agarose gel (without DNA clean up) good enough?

Why does this matter?

These techniques are cool, but like my old mentor would always [sarcastically] ask: “DO you really need to reinvent the wheel?” especially for simple point mutations and deletions? The answer is no, you don’t HAVE to. Maybe spending weeks/months learning and optimizing new techniques isn’t worth saving you a few days of routine work (for long running projects, I strongly disagree.) I get what he’s trying to say, but I think a better way to say it is “How do you need to get to where you’re going?” and it depends. Traditional cloning strategies are ok for occasional experiments, e.g. maybe you know a lot about the structure/function of your protein and/or have a very specific question to ask. However, what happens if you need to do a bunch of novel modifications, e.g. chimeric/tethered receptors? That’s going to require a lot of optimization and testing, and in this case, traditional cloning strategies complicate things with extra steps.

Reviewer 2 might ask “Are chimeric/tethered receptors even useful if they aren’t endogenously expressed in animals like mice/rats?” I don’t know, are DREADDS/optogenetic tools useful in neuroscience?

* Nowadays BRET seems like the workhorse of GPCR biology/drug discovery? Although fundamentally the idea of this blog remains (method of choice would probably gibson/golden gate cloning tho.) Also beware of PAINS.

** Agonist binding to the G-protein won’t produce a change in current, but activation of a GPCR will cause the alpha and beta-gamma subunits to dissociate, activating the GIRK, which will cause a change in current.

*** I can be GLP compliant, I swear! These are all theoretical/pre-pre IND studies.

**** I’ve read that chemical transformation is not as efficient as electroporation with respect to transformation efficiency. During grad school, I was in the “I’m good with >10 colonies” camp of molecular biologists, but then again I was mutating my plasmids myself. If you’re making antibody phage display libraries and need diversity, you better get to arc-ing.