Your lab should switch to Golden Gate cloning. Now!

One of the likely barriers to entry for labs to switch from molecular cloning reliant on Type IIP (palindromic) restriction enzymes (e.g. EcoRI, BamHI, etc.) to Golden Gate cloning (which utilizes Type II S (staggered) restriction enzymes) might hinge on the cost of purchasing synthesized parts. Given that it is now possible to order synthetic parts for $0.12-0.15 per basepair, labs would be wise to reconsider their position. Two key aspects that I have come to greatly appreciate during the last year of using Golden Gate cloning on an exclusive basis are the tremendous time savings and the remarkable success rate of Golden Gate. When I previously gene jockeyed with Type IIP enzymes it was necessary to select between 3 to 15 bacterial colonies to identify the clone containing the plasmid with the desired result/structure (which became a huge amount of work when the miniprep method was three solution lysis coupled with phenol/chloroform extraction and ethanol precipitation, as I had to do during graduate school). With Golden Gate I generally only have to select one colony (maybe one of 10 experiments require a 2nd colony). The decrease in wasted miniprep reagents and supplies represents a real savings of precious lab resources. The decreased amount of time necessary to “find” the desired clone translates directly to increased productivity and allows for efforts to be focused elsewhere. One other important aspect: The “cost” of synthesis will pay for itself. Given the modular nature of parts, where the DNA molecules are now standardized to follow the “syntax” agreed upon by the plant biology community, you only need to have a part synthesized once. When you have the part, you have the part. It can be easily shared. The same is true for transcriptional units. In our own lab we are already starting to share transcriptional units between projects (selection and scorable markers constructed for use in one dicot can often be utilized in other dicots, the same is true for tools made for monocots). The decrease in redundancy means that commonly used transcriptional units save money and time because they simply don’t have to be rebuilt. Yes, a lab (or a group of labs) has to make a concerted effort and investment in the beginning when converting over to Golden Gate cloning, but it will be absolutely worth it after even just a short period of time. While I am certainly a better molecular biologist because of the skills I had to learn to be able to clone with Type IIP enzymes, the fact is that it would be absolutely foolish to revert. I certainly will retain my skills, but I won’t be going back.

Well expressed scorable markers are helpful for identifying bacterial colonies containing (or not containing) a fragment of interest.

I am building a system which I will eventually deploy for use in constructing DNA molecules that I will use in plants. I have engineered a lacZ promoter driven fluorescent protein which is working quite well. Clones which do not contain the fluorescent protein are very easy to identify (see the white arrows; in the upper panel is the brightfield image, while the lower panel is the same colonies but with illumination/filtration to visualize the presence of the fluorescent protein, which in this experiment happens to be tandem Tomato), so I am quite pleased with this result. More about this tool later.brightfield v fluor tdtom

The importance of controls, proper annotation, and methodically working through unexpected results…

For one of my projects I prepared the turbo Green Fluorescent Protein molecule we have in our Golden Gate parts collection and installed this into a plasmid (pUC57-Kan) containing a fragment I had synthesized. My goal was to produce an inducible scorable marker (which should be possible with the lacZ promoter) to easily identify plasmids which contain the tGFP sequence in the targeted interval. I spread some of the transformation recovery broth on a solid LB plate supplemented with kanamycin, IPTG (to induce the lacZ promoter), and X-gal (ordinarily utilized as the substrate for blue/white selection in E. coli, although there shouldn’t be a β-galactosidase in my plasmids, but the plates were already in the refrigerator). I decided to include an uninduced control plate and delivered a portion of the recovery broth to a plate supplemented with kanamycin (no IPTG or X-gal).

tGFP cloning into ST1750-20

Since my synthesized fragment does not contain a β-galactosidase, I was quite surprised to see blue colonies on the induction plate. I became concerned since my project plan includes using for a different region of the plasmid  blue/white selection via the lacZ promoter driven β-galactosidase . tGFP in ST1750-20 panelIt seemed like the GFP expression might really be higher in cells taken from the induction plate (in the upper most micrograph there was no GFP visible in cells containing a plasmid which does not contain the GFP gene), but the blue color was completely unexpected. I really needed to figure out what was going on.

After a bike ride and a shower it occurred to me that I had unwittingly fooled myself with my typical plasmid annotation process. When I receive plasmid sequences from the synthesis company, I usually start the annotation process by first opening the molecule file in SnapGene viewer, and asking the program to find features. This can be helpful, but in this instance, the program did not “see” the partial β-galactosidase remaining in the pUC57-Kan plasmid after the synthesis company had installed my synthetic piece.tGFP in ST1750-20 lacZ issue The original β-galactosidase sequence in pUC57-Kan is 345 basepairs (uppermost cartoon), resulting in a protein of 115 amino acid residues. The synthesized fragment in pUC57-Kan results in the elimination of a portion of the original β-galactosidase sequence, but an open reading frame of 273 basepairs remains (shown as a black arrow in the middle cartoon, as well as the black arrow downstream of the tGFP in the lower cartoon) that could produce a partial β-galactosidase of 91 amino acid residues.

B-gal alignmentAlignment of the two protein sequences reveals that there is a methionine in the modified version that can act as a start codon for the partial β-galactosidase. In spite of the highly modified N terminus (with a deletion of 22 amino acids, as well as 7 non-silent mutations) the modified β-galactosidase protein, which is in frame with the tGFP, is sufficiently functional to cleave the X-gal substrate to produce the blue precipitant. Fortunately, the tGFP modified fragment will be liberated from this initial construct by SapI digestion, and the destination plasmid does not contain a β-galactosidase gene, but I will for sure continue to include proper controls when building the next construct in this series.