GeneMorph II

Back to Protocols

These notes are assembled from Janet Matsen's 1st creation of an enzyme library, leveraging Yakov Kipnis' extensive experience.

Overview:

1. Mutazyme rxn (GeneMorph II kit)
2. PCR clean-up
3. Digest with NcoI/XhoI (+DpnI if desired)
4. Gel extraction
5. Insert self-ligation test
6. Ligation into backbone

Tips

Mutazyme II rxn

The enzyme in this kit is more sensitive to operating conditions than most polymerases. It can be sensitive to:

• template concentration (specifically inhibitors present)
• adding enzyme into the 10X buffer directly (best to add it at the end & not include it in a master mix).

• First steps:
  • Make sure your amplification primers work using robust polymerase
    • Phusion works well because it has the same extension temperature. OneTaq Quick Load 2X also works at 72°C despite the suggested extension temperature of 68°C.
• Optional controls:
  • omit template. This is useful when you are using a lot of template.
  • Include a well with OneTaq or Phusion as a + control.

• Once you have it working:
  • Run a range of template concentrations

PCR cleanup

• Don't worry about yield loss here (Yakov advice)

Ligation into backbone

I never do mutagenesis experiment with a single reaction. I prepare several reactions with amount of template varying by factor of 10. Typically I also include (at least at the beginning or when I change primers/templates) amplification reactions with some robust polymerases to see if primer design is OK and template is amplifiable. With such huge amount of template that can be directly seen on the gel I would also suggest negative control (all the components except polymerase) because the band you identify as a product may be actually some minor species in your template prep and not an expected product. As an additional way of depleting wild type background you may consider nested PCR using gel purified product as template and standard robust polymerase (it is probably not even necessary to use high-fidelity polymerase). I always prepare several libraries (with different expected level of mutagenesis) and sequence 10-20 clones from each. This allows very reliable estimation of mutation load in each library and especially frequency of wild type sequences. For actual activity screening I obviously take the library/libraries that satisfy whatever criteria I had in mind.

I have been in the situations when GeneMorph kits simply didn't work (bad production batch probably). Mutagenic reactions with lowest possible concentration of template and largest number of cycles (conditions that according to instructions should flood the gene with 20 mutations/kb) barely produced 0.5 mutations/kb. Therefore I prefer scanning range of conditions (easiest being range of template concentrations) at least first time so not wasting time on sequential optimization of conditions. I do several reactions in parallel (taking advantage of using master mixes as much as possible) and preparing as many libraries as I feel necessary or can process. Characterizing the libraries with small scale sequencing. At the next step I pool together libraries I feel similar enough to each other, discarding bad libraries and end up with very few (even single) libraries I intend to screen. Transformation efficiency you want is limited by number of clones you can screen. There is no point of getting 10^9 transformants if you can test only 10^3. Any method reliably and reproducibly giving you 1000 colonies (if you can screen 96) should be OK. Electroporation is obviously a "gold standard", but considering hands-on time with cuvette, electroporator etc I would vote for something heat-shock based. Inoue's method is fine if you feel comfortable making your own cells. Advantage is you are not limited by bacterial strains available from companies. Buying cells is also fine. Since you don't need a lot of transformants you can easily split aliquot of commercial cells into several transformations which make it even more economically feasible. Keep in mind even commercial cells can be bad (I have encountered non-transformable cells from Invitrogen; I have obtained them via UW BioChem store therefore don't know whom to blame). In my opinion, a mindset "don't want to waste something precious" is way too often leads to experimental disaster. When forced to cut down on usage of particularly valuable reagent it is often control reactions that are sacrificed, which is never a good idea at early stages of establishing or learning a new protocol. Try to make as much of a library DNA as necessary to test several conditions and run all necessary controls. Regarding IPTG vs autoinduction. Personally I prefer explicit induction with IPTG since I feel I have more control over time, intensity and homogeneity of induction. It may depend on particular protein/expression strain. If autoinduction produces enough protein - great, stay with it. But sometime problems with amount of expressed/soluble protein can be solved by transferring cultures to lower temperature post-induction, decreasing rate of protein production by decreasing concentration of IPTG, induction at different cell densities etc. You can try first IPTG and if it works you can try autoinduction for comparison. If protein production is not affected, but hands-on time decreases then use autoinduction.

In my opinion Mutazyme is a little bit more finicky than other polymerases and requires more attention to make it work. And while bad kit is definitely a possibility, it is not the time to resort to this explanation just yet. Amount of DNA template you use in the reaction is my main concern for two reasons. First, some polymerases are more sensitive to the excessive amount of DNA in the reaction mixture than others; second, template DNA prep may contain toxic contaminant, which directly or indirectly inhibits Mutazyme (example EDTA or salts from miniprep). Both of this issues can be addressed by titrating the amount of template. Additionally, I have seen that Mg2+ titrations may be very helpful (as in case of other polymerases) to get Mutazyme to work. By the way you don't have to run all this optimization experiments on the large scale (50 uL). 10 ul per reaction is more than enough (5 ul for agarose gel afterwards and another 5 ul diluted 1:100 in water provides enough template for second reamplification with high-fidelity polymerase to get quantity sufficient for reliable cloning). In my opinion buffers and polymerases from different suppliers are not easily interchangeable (unless explicitly stated compositions are identical, which probably never happens). Running Phusion in Mutazyme buffer and getting no amplification is very hard to interpret and not very informative.

Thanks for PPT on autoinduction. Nice summary. But couple of points are worrisome. They mention (slide #12) that not every clone works with the method and you may need to find "best inducing cells". They also mention "occasional" protein degradation (slide #17). However, similar things can be said about IPTG induction as well probably. So if autoinduction works reliably (and you don't lose members of your library as false negatives due to induction/degradation issues) it is great.

By the way do you know how much of your protein expressed in soluble-vs-insoluble form using different conditions? My last questions. When you add 1ul of ON culture I wonder how enzyme gets released and whether you tried the assay with negative controls (cells with no enzyme)?

Hi Janet,

Mutazyme is probably no more fickle than many other polymerases or their blends, the perception is due to the fact that if one particular polymerase doesn't work with some gene it is easier to switch to other polymerase (or just test these polymerases in parallel) than spend time to optimize conditions for a particular one. Since you cannot easily replace Mutazyme (unless you ready to tolerate biased mutation spectrum) you become stuck and frustrated with it slightly more often. Soluble and insoluble fractions of cell lysate look distinct on gel (ratios of band intensities for various proteins is quite different), therefore if you see large portion of unlysed cells contaminating insoluble fraction distinction between soluble/insoluble becomes less pronounced. To make the comparison easier I typically run 3 lanes/sample (total cell lysate/soluble(sup after spin)/insoluble(pellet after spin) I always try to load similar amount of protein on a gel. For "total" and "soluble" it is straightforward, to help solubilize as much of "insoluble" as possible I resuspend pellet in small volume of 8-9 M urea (avoid GuHCl it destroys SDS gels) first and reconstitute to the same volume before taking small aliquot for gel sample. Example: total volume of the lysate 500 uL (take 10 uL for gel="total"), spin (take 10 uL="soluble"), remove supernatant as much as possible, add 20-50 uL 9M urea, resuspend pellet, dilute to 490 uL by water/buffer (take 10 uL="insoluble"). This way theoretically overlay of "soluble" + "insoluble" lanes should give you "total", if corresponding bands do not sum up there is a chance something unaccounted happened. Later you can skip "total" to save space on gel.