The "Green Fluorescent Protein" (GFP), discovered back in 1962, has become the most widely used "tag" for proteins whose localization in living cells is examined microscopically, be it in eukaryotic, archaeal or bacterial ones. Not least because GFP tagged proteins, that is, gene fusions of the relatively small GFP (~27 kDa) with the protein of interest, retain their functionality in many, if not most cases. Here, here and here we presented studies that would have been impossible without the use of GFP-tagged proteins. We were thus thrilled to learn (on Twitter, glad we joined!) that skilled artisans among us microbiologists are working on further improvements of this virtually indispensable tool. We asked Nick Coleman to tell us about the work in his lab prior to publication.
by Nick Coleman and Mark Somerville
It is perhaps ironic that for a story about open-source science, the tale of Free Use GFP (fuGFP) begins with a commercial collaboration. The Coleman lab was contracted by a small industry partner to develop green fluorescent bacteria for quality control applications. We planned initially to use superfolder GFP (sfGFP) for this job, since it's the gold standard GFP for most purposes. sfGFP was available in the iGEM Parts kit, and thus we assumed that it was an open source gene.
It turns out that sfGFP is patent-protected (lucky we checked!) and that that commercial use requires that a licence fee be paid to the patent-holders. This was a problem for our commercial partnership since the sfGFP licence fees were more than we were being paid to do this work! This was a possible deal-breaker for the project. Luckily the Coleman lab research assistant Mark Somerville was not afraid of a little legal legwork and he dived deep into the sfGFP patent.
Our best option seemed to be to mutate the sfGFP amino acid sequence to make it >20% different to the protein sequence claimed by the patent, while still (somehow!) maintaining its fluorescence and superfolding properties. Interestingly, the patent claim with respect to amino acid identity is referenced to GFPmut3, and then separately claims the 'superfolder' mutations.
We consulted our industry partner and they were happy to pay a little more to enable us to pursue a non-patented version of superfolding GFP. Their intellectual property was in the downstream applications, and thus they didn't really care where the GFP came from or how we got it or subsequently managed it. They were supportive of making our proposed new GFP open-source.
fuGFP as single-copy gene in chromosomes of E. coli, P. putida, and M. smegmatis. Strains were grown on blood agar plates. Illumination was from a hand-held long wave UV lamp. Source. Frontpage: E.coli expressing different fluoroproteins and chromoproteins from corals and jellyfish. Photos by Mark Somerville. Source
We first designed four different synthetic genes with approx. 20% amino acid difference to GFPmut3, based on conservative mutations with reference to alignments of various fluoroproteins. These synthetic variants had very weak fluorescence. This was disappointing… but we did not give up! Mark had an inspired idea that we could use DNA shuffling to recombine the sequences… maybe this would yield something better?
So we shuffled the four synthetic GFPs, recloned the resulting hybrids, and screened for fluorescence. Surprisingly, this yielded several strongly-fluorescing colonies. Sequencing revealed that the best-performing of these were all unfortunately still "within patent", but with a little sleight-of-hand we managed to push this under the desired 80% aa identity cutoff by digesting and ligating different sections of these genes, and by a few further conservative sequence tweaks.
The final gene, designated fuGFP, has 76% amino acid identity to GFPmut3, and is thus safely outside the claim of the sfGFP patent. It maintains the superfolder mutations, and is thus fast-folding and very bright. Interestingly, fuGFP absorbs light best in the long-wave UV (not blue like sfGFP). We believe this is due to a random mutation in the chromophore (TYG→SYG). This may be advantageous since it gives clearer separation between excitation and emission maxima.
Publication is still some way off, we need to do much more biochemical and biophysical characterisation of the fuGFP protein. But in the meantime, we will give fuGFP immediately to anyone who wants it for any purpose. While we respect the patent process and the idea of intellectual property, we believe that in the case of these incredibly useful fluoroproteins that the patenting of the best-performing variants is harmful to scientific progress. We say "F.U." to the idea of patenting genes. DNA wants to be free.
Nick Coleman is Associate Professor of Microbiology at the School of Life and Environmental Sciences, Faculty of Science, University of Sydney, Australia. Mark Somerville is a research assistant in the Coleman lab. See the lab website for more information about their research. Nick is on twitter: @Colemanomonas .
Comments