Deletions of the Aequorea victoria Green Fluorescent Protein Define the Minimal Domain Required for Fluorescence

Abstract

The Green Fluorescent Protein (GFP) from the jellyfish Aequorea victoria is a widely used marker for gene expression and protein localization studies. Dissection of the structure of the protein would be expected to shed light on its potential applications to other fields such as the detection of protease activity. Using deletion analysis, we have defined the minimal domain in GFP required for fluorescence to amino acids 7–229. This domain starts at the middle of the first small α helix at the N terminus of GFP and ends immediately following the last β sheet. Studies of the amino acids at both termini of the minimal domain revealed that positions 6 and 7 at the N terminus are Glu-specific. Change of the Glu residues to other amino acids results in reduction of GFP fluorescence. Position 229 at the C terminus of GFP, however, is nonspecific: the Ile can be replaced with other amino acids with no measurable loss of fluorescence. A total of only 15 terminal amino acids can be deleted from GFP without disrupting fluorescence, consistent with findings of a previous study of GFP crystal structure (Ormo, M., Cubitt, A. B., Kallio, K., Gross, L. A., Tsien, R. Y., Remington, S. J. (1996) Science 273, 1392–1395 and Yang, F., Moss, L. G., and Phillips, G. N., Jr. (1996) Nat. Biotechnol. 14, 1246–1251) that a tightly packed structure exists in the protein. We also generated internal deletions within the loop regions of GFP according to its crystal structure and found that all such deletions eliminated GFP fluorescence. The Green Fluorescent Protein (GFP) from the jellyfish Aequorea victoria is a widely used marker for gene expression and protein localization studies. Dissection of the structure of the protein would be expected to shed light on its potential applications to other fields such as the detection of protease activity. Using deletion analysis, we have defined the minimal domain in GFP required for fluorescence to amino acids 7–229. This domain starts at the middle of the first small α helix at the N terminus of GFP and ends immediately following the last β sheet. Studies of the amino acids at both termini of the minimal domain revealed that positions 6 and 7 at the N terminus are Glu-specific. Change of the Glu residues to other amino acids results in reduction of GFP fluorescence. Position 229 at the C terminus of GFP, however, is nonspecific: the Ile can be replaced with other amino acids with no measurable loss of fluorescence. A total of only 15 terminal amino acids can be deleted from GFP without disrupting fluorescence, consistent with findings of a previous study of GFP crystal structure (Ormo, M., Cubitt, A. B., Kallio, K., Gross, L. A., Tsien, R. Y., Remington, S. J. (1996) Science 273, 1392–1395 and Yang, F., Moss, L. G., and Phillips, G. N., Jr. (1996) Nat. Biotechnol. 14, 1246–1251) that a tightly packed structure exists in the protein. We also generated internal deletions within the loop regions of GFP according to its crystal structure and found that all such deletions eliminated GFP fluorescence. The Green Fluorescent Protein (GFP) 1The abbreviations used are: GFP, Green Fluorescent Protein; EGFP, enhanced GFP; PCR, polymerase chain reaction; CHO, Chinese hamster ovary; FACS, fluorescence-activated cell sorter; Tc, tetracycline. 1The abbreviations used are: GFP, Green Fluorescent Protein; EGFP, enhanced GFP; PCR, polymerase chain reaction; CHO, Chinese hamster ovary; FACS, fluorescence-activated cell sorter; Tc, tetracycline. from the jellyfishAequorea victoria has been widely used as a reporter in the determination of gene expression and protein localization (1Chalfie M. Euskirchen G. Ward W.W. Prasher D.C. Science. 1994; 263: 802-805Crossref PubMed Scopus (5466) Google Scholar, 2Marshall J. Molloy R. Moss G.W. Howe J.R. Hughes T.E. Neuron. 1995; 14: 211-215Abstract Full Text PDF PubMed Scopus (220) Google Scholar, 3Chalfie M. Photochem. Photobiol. 1995; 60: 651-656Crossref Scopus (250) Google Scholar, 4Cubitt A.B. Heim R. Adams S.R. Boyd A.E. Gross L.A. Tsien R.Y. Trends Biochem. Sci. 1995; 20: 448-455Abstract Full Text PDF PubMed Scopus (1196) Google Scholar). GFP cDNA can be expressed in various cells or organisms with an easily detected fluorescence in the absence of any substrate or cofactor (5Heim R. Prasher D.C. Tsien R.Y. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12501-12504Crossref PubMed Scopus (1482) Google Scholar). GFP is a 27-kDa, single-chain polypeptide of 238 amino acids (6Prasher D.C. Eckenrode V.K. Ward W.W. Prendergast F.G. Cormier M.J. Gene (Amst.). 1992; 111: 229-233Crossref PubMed Scopus (1754) Google Scholar). A key sequence of Ser-Tyr-Gly at amino acids 65–67 near the N terminus functions as the GFP fluorophore (7Cody C.W. Prasher D.C. Westler W.M. Prendergast F.G. Ward W.W. Biochemistry. 1993; 32: 1212-1218Crossref PubMed Scopus (572) Google Scholar). These three amino acids undergo spontaneous oxidation to form a cyclized chromophore responsible for the fluorescence of GFP (8Inouye S. Tsuji F.I. FEBS Lett. 1994; 351: 211-214Crossref PubMed Scopus (133) Google Scholar). Enhanced GFP (EGFP) is a mutant of GFP with a 35-fold increase in fluorescence (9Cormack B.P. Valdivia R.H. Falkow S. Gene (Amst.). 1996; 173: 33-38Crossref PubMed Scopus (2499) Google Scholar, 10Haas J. Park E.C. Seed B. Curr. Biol. 1996; 6: 315-324Abstract Full Text Full Text PDF PubMed Scopus (439) Google Scholar, 11Yang T.T. Cheng L. Kain S.R. Nucleic Acids Res. 1996; 24: 4592-4593Crossref PubMed Scopus (349) Google Scholar). This variant has mutations of Ser to Thr at amino acid 65 and Phe to Leu at position 64 and is encoded by a gene with optimized human codons (10Haas J. Park E.C. Seed B. Curr. Biol. 1996; 6: 315-324Abstract Full Text Full Text PDF PubMed Scopus (439) Google Scholar). Recent crystallographic studies of wild-type GFP and the mutant S65T indicate that the GFP tertiary structure resembles a barrel (12Ormo M. Cubitt A.B. Kallio K. Gross L.A. Tsien R.Y. Remington S.J. Science. 1996; 273: 1392-1395Crossref PubMed Scopus (1909) Google Scholar,13Yang F. Moss L.G. Phillips Jr., G.N. Nat. Biotechnol. 1996; 14: 1246-1251Crossref PubMed Scopus (1287) Google Scholar). It consists of 11 antiparallel β sheets and a single central α helix surrounded by the β sheets. The chromophore resides in the center of the barrel, completely shielded from the extral environment. There are three small α helices within the protein, one of which is located at the N terminus. Most of the β sheets are connected by small loops of one to four amino acids. Two larger connecting loops, beginning at positions 128 and 188, contain 15 and 9 amino acids, respectively. The compact structure of GFP presumably contributes to its unusual resistance to proteolysis and denaturation. This structure might also explain the observation that few amino acids can be removed from the termini of GFP without affecting its activity. Based on deletion analysis, Dopf and Horiagon (1996) concluded that amino acids 2–232 are the minimal domain required for the fluorescence of wild-type GFP (14Dopf J. Horiagon T.M. Gene (Amst.). 1996; 173: 39-44Crossref PubMed Scopus (84) Google Scholar). N-terminal deletions of amino acids 2–8 abolished activity. In contrast, the C terminus of GFP is slightly less critical. A deletion of amino acids 233–238 has activity, but the slightly larger deletion of amino acids 226–238 does not. A finer map could likely be established by making single amino acid deletions from both ends. Such defined analysis would provide not only the minimal region of GFP that is required for its fluorescence but also information about key amino acids at the termini of GFP that are required for its structure and function. Here we report a more defined map of the minimal functional domain of GFP based on single amino acid truncation studies of EGFP, and substitution analysis of several important amino acids at the termini of the minimal fluorescent domain. Our study shows that the minimal domain of GFP consists of amino acids 7–229. Additional amino acid deletions at either end dramatically decrease GFP activity. We also found that Glu residues at amino acids 6 and 7 and Ile at amino acid 229 are critical for maintaining GFP activity. Our mutation studies indicate that the amino acid at the N terminus of the GFP minimal domain must be Glu, while the specific amino acid at its C terminus is not critical. Furthermore, internal deletions eliminate GFP fluorescence. Wild-type EGFP and its deletion mutants were made by PCR withPfu DNA polymerase (Stratagene). EGFP cDNA was amplified by using plasmid pEGFP-N1 (CLONTECH) as a template and a 5′ primer with an introduced SacII sequence and a 3′ primer with EcoRI sequence. These two restriction sites were used for cloning of the amplified sequences into the mammalian expression vector pTRE (CLONTECH). To make the N-terminal deletions, the translation initiation codon AUG was introduced into the 5′ primer such that translation would start at nucleotide sequences encoding amino acid 7, 8, or 9. To make the C-terminal deletions, a stop codon was introduced into the 3′ primer such that translation would terminate at amino acid 233, 232, 231, 230, 229, 228, or 227. Site-specific mutations at amino acid 6, 7, or 229 were also created by two-step PCR. First-round PCR was carried out with the primers that contain the mutated nucleotides, and second-round PCR was with the EGFP-5′ and -3′ primers containing SacII andEcoRI sites. An insertion sequence containingBglII and HindIII recognition sites was introduced between amino acids 132 and 133, 137 and 138, and 192 and 193, respectively. The mutants were made according to the homology extension procedure (15Horton R.M. Hunt H.D. Ho S.N. Pullen J.K. Peasse L.R. Gene (Amst.). 1989; 77: 61-68Crossref PubMed Scopus (2634) Google Scholar). First, the insertion sequence was introduced into a pair of primers that were complementary to each other. In the primers, the inserted sequence was flanked by 15 nucleotides of EGFP cDNA sequence on each side. These two primers along with EGFP-5′ and -3′ primers were used for PCR amplification to generate two DNA fragments, corresponding to the N and C termini of GFP. The two DNA fragments were then denatured and renatured to allow them to hybridize at their homologous sequences. The hybrid molecule was then extended to synthesize a full-length EGFP containing the insertion. Internal deletion and mutations were also made by this method. The deletions include Del [76–81], Del [83–88, Del [132–139], and Del [191–195]. Amino acids 71–81 and 83–88 were mutated to Ala, Ala [73–81] and Ala [83–88]. The amplified DNAs were cloned into pTRE for expression in mammalian cells. pTRE is a tetracycline (Tc)-inducible expression vector developed by Gossen and Bujard (16Gossen M. Bujard H. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 5547-5551Crossref PubMed Scopus (4232) Google Scholar). In this expression system, a modified cytomegalovirus (CMV) promoter with a Tc-repressor binding element is regulated by a fusion of the tet-repressor and the herpes simplex virus VP-16, designated as tTA (16Gossen M. Bujard H. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 5547-5551Crossref PubMed Scopus (4232) Google Scholar). It binds to the promoter only in the absence of Tc. Gene expression needs to be carried out in the cells that have been transfected by tTA, such as CHO/tTA (CLONTECH). This system gives a higher level of expression than wild-type CMV-promoter systems (17Ying D.X. Zhu L. Schimke R.T. Anal. Biochem. 1996; 235: 195-201Crossref PubMed Scopus (76) Google Scholar). The constructs were purified with Qiagen columns and transfected into CHO/tTA cells with a liposome transfection reagent (CLONfectin,CLONTECH). 5.0 μg of DNA were incubated with 5 μg of CLONfectin for 30 min and overlaid onto 106 cells for 1 h in serum-free medium. After 16 h of growth in the medium with serum, the cells were examined with a fluorescence microscope. The transfected cells were also analyzed for fluorescence intensity by FACS (Becton Dickinson). The deletion study of Dopf and Horiagon (14Dopf J. Horiagon T.M. Gene (Amst.). 1996; 173: 39-44Crossref PubMed Scopus (84) Google Scholar) showed that the minimal domain of wild-type GFP begins at the N terminus somewhere between amino acids 2 and 8 and ends at the C terminus somewhere between amino acids 226 and 233 (14Dopf J. Horiagon T.M. Gene (Amst.). 1996; 173: 39-44Crossref PubMed Scopus (84) Google Scholar). To further define this map, we made a series of single amino acid deletions from both ends of an enhanced version of GFP. EGFP was used as a target, instead of wild-type GFP, because of its 35-fold higher fluorescence and ease of detection in mammalian cells. The crystal structures of GFP and S65T suggest that mutations in EGFP are unlikely to affect the N- and C-terminal structure of the protein (12Ormo M. Cubitt A.B. Kallio K. Gross L.A. Tsien R.Y. Remington S.J. Science. 1996; 273: 1392-1395Crossref PubMed Scopus (1909) Google Scholar, 13Yang F. Moss L.G. Phillips Jr., G.N. Nat. Biotechnol. 1996; 14: 1246-1251Crossref PubMed Scopus (1287) Google Scholar). As shown in Fig.1 A, we made three N-terminal deletions, δ 6, 7, and 8 and seven C-terminal deletions, δ 233, 232, 231, 230, 229, 228, and 227. All of these deletion mutants were made by PCR and cloned into the mammalian expression vector pTRE. The fluorescence of these mutants was determined by fluorescence microscopy of transiently transfected CHO/tTA cells. Expression from several key constructs was also quantitatively analyzed by FACS. We found that the N-terminal deletion mutant δ 6 has similar fluorescence to the full-length EGFP, but additional deletion of one or two more amino acids dramatically reduced the fluorescence (Fig. 1 B). We also determined the activity of some of these mutants by FACS analysis and found that, after subtraction of the autofluorescence background, δ 6 exhibited 112% activity compared with the full-length EGFP, while δ 7 retained only 12% activity (TableI). These data indicated that the minimal functional region of GFP starts at amino acid 7.Table IFACS analysis of EGFP and mutations in transfected CHO/tTA cellsEGFP mutationsRelative fluorescence intensity% ControlIntact EGFP610100%N-Terminal Deletions Del6683112% Del77612%C-Terminal Deletions Del2291055173% Del22834456%N-Terminal Mutations A6E/A7E34657% K6E/K7E20433% D6E/D7E33455%C-Terminal Mutations L229I696114% K229I670110% D229I54289%Negative Control264%Transfections were performed in CHO/tTA cells using the procedure as described under “Experimental Procedures.” After 24 h, cells were collected and analyzed for fluorescence intensity by FACS Caliber (Becton Dickinson). Untransfected cells were included in the analysis as a negative control. Open table in a new tab Transfections were performed in CHO/tTA cells using the procedure as described under “Experimental Procedures.” After 24 h, cells were collected and analyzed for fluorescence intensity by FACS Caliber (Becton Dickinson). Untransfected cells were included in the analysis as a negative control. Similar analyses were performed with the C-terminal deletions of EGFP. All of the C-terminal mutations were shown to have the same fluorescence as the full-length protein, except for δ 228 and 227 (Fig. 1 C). Surprisingly, the δ 229 mutant exhibits a brighter fluorescence than full-length EGFP (Fig. 1 C and Table I). One possible explanation for this is that some of the amino acids after amino acid 229 may have a negative affect on GFP fluorescence via a non-favorable interaction with residues located in the barrel structure. δ 228 has 56% fluorescence compared with the full-length protein (Table I). δ 227 mutant exhibits the similarly declined fluorescence as δ 228 not analysis with GFP indicated that similar of protein were used in this that in fluorescence are to the in structure than protein not the minimal functional domain of GFP ends at amino acid these N- and C-terminal deletion analyses that amino acids the minimal functional domain of EGFP required for fluorescence. the N-terminal deletion analysis, we the of amino acid 7 in GFP fluorescence. amino acid sequence of GFP a Glu at this There are at two possible that this amino acid might to the fluorescence of GFP. First, the Glu may be for interaction with amino acid to the structure and of GFP. this position may to be by an amino To between these two we the Glu to other amino acid is Glu at amino acid 6, we made single and mutations of these two Glu residues to a amino Ala, or to the amino (Fig. 2 Change of either one of these two Glu residues to Ala or not affect fluorescence compared with EGFP in transfected CHO/tTA cells (Fig. 2 B). both of the Glu residues were to Ala or the activity was reduced dramatically (Fig. 2 and Table I). These results indicate that one of these two Glu at either position 6 or 7, is for GFP fluorescence. we that negative of these two Glu residues is important to GFP structure and function. To these two Glu residues can be replaced by other amino acids, we mutated both of them to We found the fluorescence of this mutant to be reduced to the same as in the Ala and mutants (Fig. 2 and Table I). These suggest that the Glu residues at positions 6 and 7 are important not only in negative but also in maintaining specific One of these two residues might with a within GFP, and the interaction is important to GFP structure and function. to the crystal structure of GFP, these two Glu residues are located in a α helix at the N from amino acids to 9 (12Ormo M. Cubitt A.B. Kallio K. Gross L.A. Tsien R.Y. Remington S.J. Science. 1996; 273: 1392-1395Crossref PubMed Scopus (1909) Google Scholar, 13Yang F. Moss L.G. Phillips Jr., G.N. Nat. Biotechnol. 1996; 14: 1246-1251Crossref PubMed Scopus (1287) Google Scholar). the α helix is in the deletion mutant δ 6, this variant has the same fluorescence as full-length GFP (TableI). we that an α helix structure near the N terminus is not critical for fluorescence. The last amino acid at the C terminus of the GFP minimal domain is a Ile interaction at this position is not an in the GFP structure and function. Ile is a amino is possible that the of the amino acid may be important to GFP structure and function. To the of this single amino acid to GFP structure and we mutated this Ile to Leu and to the or the The fluorescence of these three mutants was examined in CHO/tTA cells and no was found between the fluorescence of the mutants and full-length GFP and Table I). the of the Ile at amino acid position 229 is not critical. to the crystal structures of GFP and S65T (12Ormo M. Cubitt A.B. Kallio K. Gross L.A. Tsien R.Y. Remington S.J. Science. 1996; 273: 1392-1395Crossref PubMed Scopus (1909) Google Scholar, 13Yang F. Moss L.G. Phillips Jr., G.N. Nat. Biotechnol. 1996; 14: 1246-1251Crossref PubMed Scopus (1287) Google the last antiparallel β ends at amino acid Ile at amino acid 229 is from the β the of this structure by the mutations does not for its in GFP function. Based on the of for a specific amino acid at this we that of this position is important to GFP structure and function. to its crystal GFP the two loops, which from amino acids to and from to The regions not in the structure but as to the To these regions are critical to GFP fluorescent activity, we generated two internal deletions, Del and Del one within each of these amino acids are deleted from Del and 5 amino acid from Del [191–195]. These two deletion mutants showed no fluorescence after were transfected into CHO/tTA cells (Table these loop regions are required for either of GFP or for maintaining its structure. We found that these two regions are not only to deletions but also to We three insertion two in between amino acids 132 and and 137 and and one in between 192 and These three also GFP activity not The results suggest that the structures of these two loops are important to GFP structure and fluorescence. Change of their either by or GFP analysis of EGFP and mutations in transfected CHO/tTA cellsEGFP were performed in CHO/tTA cells using the procedure as described under “Experimental Procedures.” After 24 cells were examined under a fluorescence microscope. Open table in a new tab Transfections were performed in CHO/tTA cells using the procedure as described under “Experimental Procedures.” After 24 cells were examined under a fluorescence microscope. There are two small α that from amino acids to and from to To these two small structures are for GFP fluorescence, we first deleted these two We found no fluorescence were examined in CHO/tTA cells (Table The might be to the structure of GFP by such To eliminate such we then mutated each of the amino acids within these two regions to Ala, Ala and Ala [83–88]. The mutations also GFP fluorescence (Table The results indicate these two structures are required for GFP fluorescence. In we have determined that the minimal functional domain of GFP the region from amino acids 7 to Additional amino acid deletions from either terminus of GFP dramatically reduced fluorescence. Our results indicate that a total of 15 amino acids can be deleted from GFP. of a total of 15 amino acids from GFP a form of the reporter that might be for systems the of full-length GFP is such as studies with cells of the reporter may be The form may also be as a for studies. It is that fusion of some to the N or C terminus of GFP for in protein localization studies may the tertiary structure of GFP and decrease its fluorescence (17Ying D.X. Zhu L. Schimke R.T. Anal. Biochem. 1996; 235: 195-201Crossref PubMed Scopus (76) Google Scholar). of the N or C terminus may this to the crystal structure of GFP, a small α helix structure from amino acid to at the N terminus of GFP. Our N-terminal deletion results indicate that this α helix is not critical to GFP structure and because some of N-terminal deletions have the structure while maintaining fluorescence. In to this N-terminal small α the internal two small α are critical for GFP fluorescence. deletions or mutations of these two α GFP activity. The fluorescence of GFP is also to the other internal deletions within the two loop The of GFP fluorescence to such internal deletion mutations suggest the of these loops in maintaining GFP structure and function. There is a from amino acids 229 to following the last β sheet. This but not amino acid 229, can be deleted from GFP without affecting fluorescence. Amino acid 229 is important in maintaining GFP activity, but Ile to other amino acids does not affect GFP activity. position 229 is not of this position is likely to be more important than the of the amino This may a in maintaining the last β sheet. In contrast, positions 6 and 7 are Glu-specific. Change of these amino acids to other amino acids, GFP activity. These results indicate that the Glu residues at position 6 and 7 not only negative but also specific within the GFP structure. The and functional of the Glu residues at amino acids 6 and 7 and Ile at amino acid 229 in GFP needs further We for of and and for of the and