The Story of Free Use GFP (fuGFP)

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 .