Engineering of a monomeric fluorescent proteinAsGFP499 and its applications in a dual translocationand transcription assay

Aynur Tasdemir1,4, Farid Khan2, Thomas A. Jowitt3,Lucia Iuzzolino1, Stefan Lohmer1, Sabrina Corazza1,5,and Thomas J. Schmidt4,5

1Axxam SpA, Via Olgettina 58, 20132 Milan, Italy, 2Lumophore Limited,20 Oakleigh Avenue, Timperley, Cheshire WA15 6QT, UK and 3Faculty ofLife Sciences, The University of Manchester, Michael Smith Building,Oxford Road, Manchester M13 9PT, UK and 4Institute for PharmaceuticalBiology und Phytochemistry (IPBP), Westfalische Wilhelms-Universtitat,Munster, Hittorfstraße 56, D-48149 Munster, Germany

5To whom correspondence should be addressed. E-mail: thomschm@uni-muenster.de (T.J.S.)/sabrina.corazza.sc@axxam.com (S.C)

The tetrameric green fluorescent protein AsGFP499 fromthe sea anemone Anemonia sulcata was converted into adimeric and monomeric protein by site-directed mutagen-esis. The protein was engineered without prior knowledgeof its crystal structure based on a sequence alignment ofmultiple proteins belonging to the GFP-family. Crucialresidues for oligomerisation of AsGFP499 were predictedand selected for mutation. By introduction of a single sitemutation (S103K) the A/B subunit was disrupted whereastwo substitutions were necessary to separate the A/Csubunit (T159K/F173E). This method can be applied as ageneral predictive method for designing monomeric pro-teins from multimeric fluorescent proteins. The maturationtemperature was optimised to 3788888C by a combination ofSite-directed and random mutagenesis. In cell-basedassays, the NFATc1A (nuclear factor of activated T-cells,subtype 1, isoform A)-AsGFP499 chimera formed massivecytoplasmic aggregates in HeLa cells, which prevented theshuttling of NFATc1A into the nucleus and consequentiallyits transcriptional activity. In contrast, the cells expressingthe NFATc1A in fusion with our engineered dimeric andmonomeric fluorescent mutants were homogeneously dis-tributed throughout the cytoplasm, making these stablecell lines functional in both translocation and trans-criptonal assays. This new dual cellular assay will allowthe screening and discovery of new drugs that targetNFAT cellular processes.Keywords: AsGFP499/fluorescent/GFP/monomeric/NFATc1

Introduction

In the early sixties, the first member of the green fluorescentprotein (GFP) family, AvGFP, was discovered in the jellyfish,Aequorea victoria (Shimomura et al., 1962). It forms a b-canstructure consisting of 11 anti-parallel b-sheets that iscrossed by a buried a-helix (Ormo et al., 1996; F.Yanget al., 1996; T.T.Yang et al., 1996). The chromophore, anextended, conjugated p-system located on the helix, is auto-catalytically formed by the tripeptide sequence, S65/Y66/G67, which gives the protein its fluorescence (Heim et al.,1994; Cubitt et al., 1995). The unusual nature of its

chromophore, unique among known chromoproteins, is thereason why GFP has become a very popular tool in cellbiology as reporter gene, selection marker, fusion tag andbiosensor (F.Yang et al., 1996; T.T.Yang et al., 1996; Tsien,1998; Lippincott-Schwartz and Patterson, 2003).

Currently, the GFP family comprises .100 GFP-like pro-teins with different spectral properties, most of them comingfrom the Anthozoa class (Matz et al., 1999; Lukyanov et al.,2000; Wiedenmann et al., 2000). Contrary to monomericAvGFP, all known Anthozoa GFP-like proteins reported todate form oligomers and exist mostly as tetramers (Bairdet al., 2000; Vrzheshch et al., 2000; Mizuno et al., 2001;Wiedenmann et al., 2002; Shagin et al., 2004). The oligomer-isation does not impair their application as reporter genes,selection markers or biosensors, but limits their use as fusiontags to study, for instance, protein localisation and dynamicsin living cells (Baird et al., 2000; Lauf et al., 2001; Mizunoet al., 2001). A possible solution to this problem is thegenetic engineering of the fluorescent protein (FP) of interestby mutagenesis, creating mutants that form functionalmonomeric variants. For instance in DsRed, isolated fromDiscosoma corals, 33 mutations were needed to disrupt thetetrameric structure and to maintain its fluorescence (Matzet al., 1999; Campbell et al., 2002). Other groups used thisapproach to generate dimeric and monomeric variants fromdifferent FPs by mutating similar key residues (Karasawaet al., 2003; Karasawa et al., 2004; Wiedenmann et al.,2004a,b).

The FP under study, AsGFP499 cloned from Anemoniasulcata, is a protein composed of 228 amino acids(Wiedenmann et al., 2000). Here we present the generationand characterisation of dimeric and monomeric variants ofthe tetrameric AsGFP499 created by site-directed mutagenesis(SDM), which was guided by a multi-sequence alignment.Since most mammalian cells are cultured at 378C, the matu-ration temperature of the FP should be also at 378C to ensureoptimal function in cell-based applications. Therefore, inaddition to monomer design, we took a similar approach asCrameri et al. (1996) who applied random mutagenesis toselect mutants with chromophore maturation at 378C .

Furthermore we demonstrate here the suitability of theAsGFP499 mutants as fusion tags for NFATc1A (nuclearfactor of activated T-cells, subtype 1, isoform A) in translo-cation and functional transcription assays. NFAT transcrip-tion factors are intimately involved in the regulation ofmultiple important biological functions including immuneresponse, development and bone homeostasis (Ranger et al.,1998a,b; Graef et al., 1999; Graef et al., 2001; Crabtree andOlson, 2002; Graef et al., 2003; Horsley et al., 2003; Changet al., 2004; Winslow et al., 2006).

Finally, HeLa cells, expressing the NFATc1A-FP chimerasand the luciferase (Luc) reporter gene under control of theNFAT response element were developed, providing a dual

# The Author 2008. Published by Oxford University Press. All rights reserved.

For Permissions, please e-mail: journals.permissions@oxfordjournals.org

613

Protein Engineering, Design & Selection vol. 21 no. 10 pp. 613–622, 2008Published online August 1, 2008 doi:10.1093/protein/gzn040

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

assay system which allows the detection of both, theNFATc1A translocation and NFATc1A-controlled transcrip-tional activity. This dual cell-based assay format will enablethe discovery of new drugs targeting the NFATc1A signaltransduction pathway.

Materials and methods

Cloning and mutagenesis of bacterial expression vectorsThe coding sequences for EGFP, DsRed2 and humanisedAsGFP499 [hAsGFP499 (Supplementary data, Table S2)] wereamplified from the vectors pEGFP-N1, pDsRed2-N1(Clontech) and pcDNA3-hAsGFP499, respectively. The frag-ments were subcloned into pET28a(þ)-vector (Novagen) andtransformed into electro-competent BL21(DE3)Gold bacteria(Stratagene). Positive clones were confirmed by DNAsequencing.

Site-directed mutagenesisSDM was carried out on the construct pET28a(þ)-hAsGFP499

using the QuikChange SDM kit (Stratagene) according tothe manufacturer’s instructions. The resulting vectors weretransformed into electro-competent BL21(DE3)Gold cells.Mutations were confirmed by DNA sequencing.

Random mutagenesisThe coding sequences of hAsGFP499(S103K/R150E) andhAsGFP499(S103K/T159K/F173E) were cloned into pRSET-B vector (Invitrogen). Random mutagenesis was performedusing the Genemorph random mutagenesis kit II (Stratagene)(Miyazaki, 2003). The resulting vectors were transformedinto electro-competent BL21(DE3)Gold cells and grown at378C. Mutants of interest were sequenced and recloned intopET28a(þ) vector.

Cloning of mammalian expression vectorsFor mammalian cell expression, EGFP in pEGFP-N1 fusionvector (Clontech) was substituted with the coding sequencesof hAsGFP499 and its mutants (pFP-N1). Further, the codingsequence of NFATc1A was amplified without its stop-codonfrom pcDNA3-NFATc1A by PCR. The fragment wasinserted in frame to the FPs to generate the fusion constructspNFATc1A-FP-N1. Positive clones were confirmed by DNAsequencing.

All primer pairs used for cloning are listed in theSupplementary data, Table S3.

Protein expression and purificationFluorescent colonies were picked and grown with antibiotic,overnight on a shaker at 378C in 2 ml liquid LB medium. Fivemillilitre of fresh medium was inoculated with 0.1 ml of theovernight bacterial cultures and grown to A600 of 0.6–0.8.Depending on the FP, expression was induced either at RT, 30or 378C by addition of IPTG to 0.5 mM. After overnightincubation, 4 ml of the cultures were centrifuged and pelletswere lysed in 200 ml BPER-II (Pierce) solution addedwith 30 mM Tris–HCl pH 8.0, 300 mM NaCl, 10 mg/mlDNAse and EDTA-free protease-inhibitor-cocktail (BoehringerMannheim). Finally, the bacterial lysates were centrifuged andthe supernatants were analysed on pseudo-native SDS–PAGE(sodium dodecyl sulphate–polyacrylamide gel electrophoresis).

For large-scale protein purification, 100 ml LB mediumcontaining kanamycin were inoculated with an overnightculture of fluorescent bacteria and grown at 378C until A600

of 0.6–0.8 was reached. Protein production was inducedwith 0.5 mM IPTG. Bacterial pellets were lysed and centri-fuged as described earlier, and subsequent protein purifi-cations of the soluble fractions were performed on 1 mlHis-Trap columns (GE Healthcare) using AKTAexplorer 100(GE Healthcare). Bound proteins were washed with 10 mMTris–HCl þ 150 mM NaCl pH 8.0 buffer containing a gradi-ent of imidazole (10–400 mM). Following protein elution by500 mM imidazole, samples were exchanged into 10 mMTris–HCl pH 7.4, 150 mM NaCl, using Amicon 10 filterunits (Millipore). The protein concentrations were deter-mined using the BCA Assay Kit (Pierce) and further used inbiochemical characterisation.

Sodium dodecyl sulphate–polyacrylamide gel electrophoresisThermally denatured (for SDS–PAGE) and non-denatured[for pseudo-native SDS–PAGE (Baird et al., 2000)] proteinsamples prepared with �5 SDS sample buffer containing200 mM DTT were loaded on 16% Tris-Glycine gels(Invitrogen) and separated in Tris-Glycine-SDS runningbuffer at pH 8.0. SeaBlue plus2 (Invitrogen) and (His)6-tagged DsRed2 (tetramer) and EGFP (monomer) were usedas protein standards. Following separation, gels were illumi-nated with UV light to detect fluorescent bands and thenstained with Coomassie blue.

Size-exclusion chromatographyThe apparent molecular weights (MW) of the purified(His)6-tagged FPs were determined using size-exclusionchromatography (SEC) on a pre-calibrated Superdex 200HR10/30 column (GE Healthcare) connected to AKTAexplorer 100(GE Healthcare) at a flow rate of 0.5 ml/min. Hundred microli-tre purified protein samples at 2 mg/ml were separated using10 mM Tris–HCl buffer containing 150 mM NaCl at pH 7.4,and the absorbance at 280, 400 and 480 nm was measured inreal-time.

SEC coupled to multi-angle laser light scatteringProteins were separated according to size on a 24/30 gel fil-tration column (GE, Amersham, UK) run in 20 mM Tris–HCl pH 7.4, 200 mM NaCl at 0.71 ml/min using a DionexBioLC HPLC (Camberley, UK). Proteins resolved as onepeak with little evidence of aggregation or other speciesbeing present. Proteins were passed through a Wyatt EOS18-angle laser photometer (Wyatt Technology, SantaBarbara, CA, USA) with the 13th detector replaced with aQELS detector (Wyatt Technology) for the simultaneousmeasurement of hydrodynamic radius (Rh). This wascoupled to an Optilab rEX refractive index detector (WyattTechnology), and the Rh, MW moments and concentrationof peaks were analysed using Astra 4.98 (WyattTechnology). Analysis was performed using a differentialrefractive increment of 0.18 ml/g and mass was extrapolatedusing the Zimm plot.

Absorbance spectroscopyA Cary 400 Scan UV-visible absorbance spectrophotometer(Varian) with a 1 cm path length microcuvette was used forabsorbance spectra measurements in PBS (phosphate-buffered

A.Tasdemir et al.

614

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

saline) at 258C. The protein concentration of each AsGFPsample was determined using the BCA protein assay kit(Pierce) using GFPuv (Crameri et al., 1996) as a standard andBeer–Lambert law was applied to calculate the molar extinc-tion coefficient (M21 cm21).

Fluorescence spectroscopyThe Cary Eclipse Fluorescence spectrophotometer (Varian)was used to obtain the fluorescence emission profiles of allproteins. For quantum yield (QY) measurements, EGFP wasused as a standard, such that the AsGFP samples matchedthe absorbances at 490 nm in PBS. After excitation at490 nm, the fluorescence emission spectra peaks for EGFPand for AsGFP samples were integrated using the Caryanalysis software. The QY of the AsGFP samples were cal-culated by direct comparison of the integrated emissionintensity with the known QY of 0.6 for EGFP (Tsien, 1998).

Cell line, culture conditions and transfectionsHeLa cells were cultured in Alpha Minimum EssentialMedium (a-MEM; Gibco-BRL), supplemented with 100 UI/mlpenicillin, 100 mg/ml streptomycin and 10% heat-inactivatedFBS at 378C in 95% humidified air and 5% CO2. Cell transfec-tions were performed using Fugene 6 (Roche).

HeLa/NFAT-Luc reporter cell line generationThe HeLa wt cell line was stably co-transfected with 0.8 mgNFAT response element, pNFAT-TA-Luc and 0.2 mg pPUR(Clontech). Antibiotic resistant clones were selected (2.5 mg/ml) and maintained (0.5 mg/ml) in puromycine containingmedium. NFAT-TA-Luc carrying clones were isolated bylimiting dilutions at 1 cell/well (c/w) in 96-well plates.Confluent clones were replicated using an automatic dispen-sing/diluting device (IGEL; OpalJena GmbH, Jena/Germany). The following day, cells were treated with 1 mMTrichostatin A (TSA; Sigma) overnight at 378C and day afterloaded with Tyrode-buffer. After injection of triton/luciferinmix the Luc activity in relative light units (RLU) wasimmediately recorded using a Lumibox CCD camera-basedluminescence reader designed and built by BayerTechnologies GmbH (Wuppertal, Germany).

NFAT-Luc/NFATc1A-FP cell line generationStable HeLa/NFAT-Luc cells (MOCK) were transfected withpNFATc1A-FPs-N1 fusion vectors and antibiotic resistantclones were selected (2 mg/ml) and maintained (0.8 mg/ml)gentamicin (G418) containing medium. NFATc1A-FP posi-tive clones were isolated by limiting dilutions at 1 c/w in96-well plates. Confluent clones were replicated and follow-ing day incubated with Tyrode-buffer plus 100 mM adeno-sine 50-triphosphate (ATP; Sigma) for 5 h at 378C beforemeasurement of Luc activity.

NFAT-Luc assayMOCK and MOCK/NFATc1A-FP cells were seeded at 10000 c/w density overnight in 96-well plates. The followingday the growth medium was removed and cells were eitherincubated in Tyrode-buffer with different concentrations ofATP (1 mM to 100 nM) for 5 h at 378C or pre-treated 1 hbefore stimulation with 50 mM ATP, with various concen-trations of Cyclosporin A [CsA 5 mM–320 pM (Sigma)] orTacrolimus [FK506 5 mM–10 pM (Sigma)] at 378C. Each

single concentration was tested in quadruplicates. The Lucactivity was then measured. To determine the mean EC50 forATP and IC50 for both CaN (calcineurin) inhibitors, CsA andFK506, RLU values were corrected against the backgroundand normalised. The normalised RLU values were plottedversus the molarity of the substances using the PRISManalysis software (GraphPad, San Diego, CA, USA).

NFAT-translocation assay in living cellsStable MOCK/NFATc1A-FP cells were grown on 35 mmglass bottom dishes at 378C. For maximum gene expression,cells were treated overnight with 1 mM TSA. The followingday, cells were either incubated in Tyrode-buffer alone orTyrode-buffer plus 5 mM CsA or FK506 for 1 h at 378C.Translocation was then initiated by adding 100 mM ATP andfluorescence images were captured every 60 s for 20 minusing a confocal microscope (BioRad Leica).

Supporting online informationMulti-sequence alignment, list of primers, DNA sequence ofhAsGFP499.

Results

Multi-sequence alignmentAt the time of this study, the crystal structure of AsGFP499

was unknown. Therefore, a multi-sequence alignment strat-egy was taken to identify key residues that may be involvedin the subunit interfaces. These residues would be promisingtargets for SDM for the disruption of intermolecular contactsresponsible for oligomerisation. Crystal structures ofGFP-like proteins were retrieved from the protein data bank[PDB-codes (Berman et al., 2000): 1GGX (Wall et al.,2000); 1MOU (Prescott et al., 2003); 1UIS (Petersen et al.,2003); 1XQM (Wilmann et al., 2005); 1XA9 (Remingtonet al., 2005); 1B9C (Battistutta et al., 2000) and 1EMM(Palm et al., 1997)] and based on secondary structureelements these proteins were aligned manually with eachother and with the protein sequences of 122 further GFP-likeproteins of unknown crystal structures (Bairoch, 2000)(Supplementary data, Table S1). Using this alignment, thesecondary structure elements of AsGFP499 were predicted.The predicted subunit interfaces of AsGFP499 were comparedwith the corresponding residues in the crystal structure of thenative tetrameric DsRed (Wall et al., 2000; Yarbrough et al.,2001), which is known to occur at two chemically distinctprotein interfaces. The hydrophobic interface, termed the A/B-interface, is formed between the chains A and B as well asC and D, whereas the hydrophilic A/C-interface is a result ofthe interaction of chains A and C as well as B and D.

Engineering of a dimeric variant of AsGFP499 by SDMIn the predicted A/B-interface of AsGFP499, the residue S103is in a position homologous to T106 in DsRed, which hasbeen previously described to form a hydrogen bond with itscounterpart on chain B (Yarbrough et al., 2001). T106 inDsRed as well as S103 in AsGFP499, according to our align-ment, are centrally located in the A/B-interface and comple-tely surrounded by a number of hydrophobic amino acids (inDsRed: V96; V104; I225; V227, in AsGFP499: V93; V101;L122). The substitution of S103 with lysine could thus lead

A monomeric mutant AsGFP499 acts as biomarker for NFAT-assays

615

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

to a disruption of the A/B-interface, since two positivelycharged residues in close proximity to each other would sig-nificantly weaken the contact. Indeed, the mutation S103K inAsGFP499 led to a fully functional dimeric FP. We termedthis mutant As-Dimer1. Interestingly, we found that the sub-stitution of T106 with lysine in DsRed caused a completeloss of its fluorescence activity (data not shown).

Engineering of a monomeric variant of AsGFP499 by SDMBased on our alignment, the analysis of the predicted A/C-interface of AsGFP499 revealed a more complex interactionnetwork than in case of the A/B-interface. A number of prom-ising mutation sites from our sequence alignment were tar-geted by SDM: (i) a salt-bridge between the residues E97(chain A) and R150 (chain C), also occurring in DsRed(E100, chain A - R153, chain C), where a randomly insertedmutation R153E was included in the mutations leading to amonomeric variant (Campbell et al., 2002); (ii) hydrogenbond(s) between R157 (chain A) and T159 and/or S171, (bothchain C), also observed in the corresponding residues of theeqFP611 from the sea anemone Entacmaea quadricolor,where Y157 forms a hydrogen bond to S171 (Petersen et al.,2003); (iii) a p-stacking between F173 (chain A) and itscounterpart (chain C).

In order to disrupt the salt-bridge, a variety of mutants ofour dimeric variant As-Dimer1 was generated in which eitherof the two residues forming the salt bridge was altered. Manymutants lost their fluorescence activity or still formed stabledimeric proteins. However, we generated the mutant As-pMono1 [AsGFP499(S103K/R150E)], which showed partialmonomeric character, but became fluorescent only at RTwhen expressed in Escherichia coli (E.coli) cultures.

Secondly, we addressed the mentioned hydrogen bond(s)between R157 (chain A) and T159 and/or S171 (both chainC). All three residues were altered by mutagenesis to perturbthese potential hydrogen bond interaction(s), which,however, did not result in any functional monomeric variant.

Attempts to disrupt potential p-stacking interactionbetween F173 (chain A) with its counterpart on chain C byreplacing F173 with the positively charged residue argininedid not alter the dimeric structure (data not shown), butexchange by the negatively charged amino acid glutamateled to a variant with partial monomeric character (data notshown). However, the mutant AsGFP499(S103K/F173E) wasfluorescent when expressed at RT in E.coli cultures, but notat higher temperatures.

A single site mutation on the A/C-interface did notproduce a monomeric FP, however, several As-Dimer1 var-iants were generated by combining mutations targeting twoor all of the residues predicted in the A/C-interface. Finally,the combination of two mutations (T159K/F173E) within theAs-Dimer1 yielded a monomeric variant (As-Mono1) thatstill exhibited fluorescence, but only when expressed at RTin bacteria.

Increasing the maturation temperature of As-Mono1An higher maturation temperature was achieved by subjectingAs-pMono1 and As-Mono1 to three cycles of random muta-genesis and screening for fluorescent bacterial colonies thatgrew at 378C. Although none of the As-Mono1 descendantsrevealed strict monomeric character on pseudo-native gels, onedimeric mutant, As-Dimer2 [AsGFP499(S94T/S103K/R150K/

S171 N/K206R)], a mutant of As-pMono1, was selected forfurther characterisation, because it efficiently developed fluor-escence at 378C. Based on the mutations leading to improvedmaturation temperature of As-Dimer2, a number of variants ofAs-Mono1 were created, which led to the monomeric mutantAs-Mono2 [AsGFP499(S94T/S103K/T159K/F173E)]. Thismutant could be functionally expressed in cells at 378C.

Pseudo-native SDS–PAGEPseudo-native SDS–PAGE of DsRed2 and wt AsGFP499

showed both proteins migrating as tetramers with an apparentMW between 64 and 98 kDa. In contrast, As-Dimer1migrated between DsRed2 and EGFP as an apparent dimer at�50 kDa with a faint band at lower MW. The mutants,As-Mono1 and As-Mono2, ran with similar mobilities asEGFP of �28 kDa for a monomeric protein. In addition,As-pMono1 had both a monomer band and a faint band atthe position of a dimer (shown by an arrow in Fig. 1A, Lane8), indicating partial monomeric/dimeric character. The ran-domly generated mutant As-Dimer2 did not show a sharpband. Its oligomeric nature could not be deduced using thismethod (Fig. 1A, Lane 9). However, all bands were clearlyfluorescent under UV excitation, indicating the formationof the mature chromophore (data not shown). Thermallydenatured SDS–PAGE gel runs produced non-fluorescentbands, with all of the samples appearing as monomeric pro-teins with mobilities between 22 and 36 kDa. The Coomassiestained gel was also used to estimate the purity of thesamples, which was .90%, except for DsRed2 (Fig. 1B,Lane 2).

Size-exclusion chromatographyThe SEC experiments showed that the determined apparentMW of As-Dimer1 (�39 kDa), As-Mono2 (�23 kDa) andEGFP (�27 kDa) are smaller in size than wt AsGFP499

(�78 kDa) and DsRed2 (�79 kDa) (Fig. 2A and B andTable I). Regarding the absorption profiles obtained inSEC-chromatograms for wt AsGFP499, As-Dimer1 andAs-Mono2, As-Mono2 exhibited a decreased absorption at480 nm together with an increased absorption at 400 nm rela-tively to the protein absorption at 280 nm (Fig. 2A).

In agreement with the SDS–PAGE results, As-pMono1displayed here two separated peaks at �39 and 23 kDa.

Fig. 1. SDS–PAGE of purified His-tagged recombinant proteins. Lane 1:Protein MW standard marker (labels express relative MW in kDa); Lane 2:DsRed2; Lane 3: EGFP; Lane 4: AsGFP499; Lane 5: As-Dimer1; Lane 6:As-Mono1; Lane 7: As-Mono2; Lane 8: As-pMono1 and Lane 9:As-Dimer2. (A) Coomassie blue stained pseudo-native SDS–PAGE. (B)Coomassie blue stained SDS–PAGE.

A.Tasdemir et al.

616

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

Interestingly, the peak corresponding to a dimer showedabsorptions at 400 and 480 nm similar to wt AsGFP499 andAs-Dimer1, whereas the monomeric peak was comparable toAs-Mono2 (Fig. 2B and Table I).

SEC coupled to multi-angle laser light scatteringThe absolute MW and the Rh for wt AsGFP499 and itsmutants were determined by SEC-MALLS (SEC coupled tomulti-angle laser light scattering). Wt AsGFP499 was con-firmed to exist as a tetramer (�102 kDa). The dimeric natureof As-Dimer1 (�54 kDa) and As-Dimer2 (�51 kDa), aswell as the monomeric nature of As-Mono2 (�30 kDa), wasconfirmed.

Absorbance spectroscopyThe absorption spectra of all AsGFP samples displayed twodistinct maxima at 482 nm (major peak) and 406 nm (minorpeak). Compared with wt AsGFP499, both As-Dimer1 andAs-Dimer2 showed an increased absorption at 482 nm withrespect to the minor peak. The most notable spectral changeoccurred within As-Mono2, where the 482 nm peak is reducedsuch that its intensity equals the 406 nm peak (Fig. 3A).

Fluorescence spectroscopyThe fluorescence emission spectra of As-Dimer1 and As-Dimer2 protein samples were very similar to wt AsGFP499

with peak maxima at 498 nm, whereas As-Mono2 was slightlyshifted towards longer wavelengths and emitted maximallylight at 500 nm (Fig. 3B). Moreover, the QY of wt AsGFP499,As-Dimer1 and As-Dimer2 were also very similar with a QY�0.5, while As-Mono2 had a QY of 0.4. Nevertheless, allproteins were extremely bright, having marginally less QYcompared with EGFP (Tsien, 1998). Table I gives a summaryof all spectroscopic properties.

Application of wt AsGFP and its mutants in cell-based assaysIt has been frequently reported that oligomeric FPs are notideal fusion partners for labelling target proteins (Baird et al.,2000; Vrzheshch et al., 2000; Lauf et al., 2001; Mizuno et al.,2001; Shagin et al., 2004). In the present work the suitabilityof AsGFP499 mutants as fusion markers was tested by investi-gating the translocation and transcriptional behaviour ofNFATc1A when fused to hAsGFP499 or a mutant of interestin HeLa cells. Therefore, HeLa cells were stably transfectedwith the NFATc1A-FP fusion constructs and an NFAT-Luc

Fig. 2. SEC chromatograms of purified proteins. Absorbance at 280 nm (solid lines), 400 nm (dashed lines) and 480 nm (dotted lines). The chromatograms arenormalised to the peaks at 280 nm. Overlaid plots of AsGFP499 (1); As-Dimer1 (2), As-Mono2 (3) (A); and As-pMono1 with dimeric and monomeric fraction (B).

Table I. Biophysical properties of wt AsGFP499 and its mutants including GFPuv, EGFP and DsRed2

Protein Apparent MW (kDa) Absolute MW (kDa) Rh (nm) Exl1/l2 (nm) Eml1/l2 (nm) 1 (M21 cm21) F

GFPuv 25 31.7 2.1+0.1 395/475 497 30 000 –EGFP 27 – – 490 500 61 000 0.6a

DsRed2 79 107.1 3.2+0.1 – – – –AsGFP499 78 101.6 3.4+0.2 482/406 498 88 000 0.47As-Dimer1 39 54.6 2.6+0.3 482/406 498 53 000 0.49As-Dimer2 39 51.3 2.6+0.3 482/406 498 64 000 0.49As-Mono1 23 – – – – – –As-Mono2 23 30 – 482/406 500 12 000 0.40As-pMono1 39 and 23 – – – – – –

aSee Tsien, 1998.

A monomeric mutant AsGFP499 acts as biomarker for NFAT-assays

617

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

reporter vector carrying three copies of the NFAT-bindingsite derived from the IL-2 promoter and fused to the Lucgene. Cells were stimulated with ATP, which increases intra-cellular Ca2þ concentration by acting on P2 purinoreceptors,inducing inositol 1,4,5-triphosphate (IP3)-mediated release ofCa2þ from intracellular stores (Badminton et al., 1996; Ferrariet al., 1999). Elevated Ca2þ activates the Ca2þ/calmodulin-dependent protein phosphatase CaN, which in turn depho-sphorylates cytoplasmic NFAT and thus promotes its nucleartranslocation. In the nucleus dephosphorylated NFAT binds topromoter regions, containing NFAT response elements andmodulates the transcription of targeted genes (Jain et al.,1993; Luo et al., 1996; Beals et al., 1997; Rao et al., 1997;Klee et al., 1998; Crabtree, 1999; Okamura et al., 2000). Theimmunosuppressive drugs CsA and Tacrolimus (FK506) wereused to inhibit NFATc1A. Both drugs interfere with theNFAT signalling pathway at the level of CaN activation andblock the nuclear shuttling of NFAT proteins, thereby pre-venting their action on target genes (Emmel et al., 1989;Mattila et al., 1990; Randak et al., 1990; Brabletz et al.,1991; Liu et al., 1991; Liu, 1993).

NFAT translocation assayLive-cell images of NFATc1A-hAsGFP499 showed formationof large aggregates in the cytoplasm before stimulation (0 min).After addition of 100 mM ATP the aggregation presumablyprevented an efficient translocation of the fusion protein intothe nucleus (20 min) (Fig. 4). In a control experiment transi-ently transfected cells with NFATc1A-DsRed2 showed similaraggregation events (data not shown). In contrast, fusions ofNFATc1A to As-Dimer1, As-Mono2 and EGFP as a positivecontrol (Kehlenbach et al., 1998, Scott et al., 2001) were homo-genously distributed in the cytoplasm prior to stimulation andtranslocated significantly into the nucleus within 20 min afteraddition of ATP. The translocation of NFATc1A-FPs was alsoinvestigated in presence of 5 mM CsA or FK506. Regardless ofthe FP used, the translocation of NFATc1A-FPs was blockedby both inhibitors and fluorescence was observed mainly in thecytoplasm after stimulation (Fig. 4).

NFAT-Luc transcriptional assayTo assess whether the function as a transcription factor ofNFATc1A is affected by fusion to a FP, its transcriptional

activity was measured using a Luc reporter system. Positivecontrol cells were either stably expressing NFATc1A only oras a fusion with EGFP. The results demonstrated that the celllines harbouring the AsGFP constructs were comparable to thepositive control cells. All cell lines were sensitive to activatorsand inhibitors of the NFAT signalling pathway (Fig. 5).

Extracellular ATP induced Luc activity resulted in NFATtranscription activity in a dose-dependent manner. In all celllines tested, including MOCK cells, the calculated meanEC50 values for ATP was �125 mM, whereas the maximalRLU values obtained by the untagged or differently taggedNFATc1A varied significantly in the order: NFATc1A(�424 000) . NFATc1-EGFP (�250 000) . NFATc1A-As-Dimer1 (�77 100)¼NFATc1A-As-Mono2 (�57 800) ..NFATc1A-hAsGFP499 (�11 650) .. MOCK (�1400).This indicates that the increased Luc activity was dependenton the recombinant expression of NFATc1A either alone oras a fusion with FPs (Fig. 5, Table II).

Fig. 3. Absorbance spectra of AsGFP499 and its mutants at 1 mg/ml (A) and fluorescence emission spectra at 1 mM protein concentration in PBS (B).

Fig. 4. Translocation of NFATc1A from the cytoplasm into the nucleus inliving cells: HeLa/NFAT-Luc cells stably expressing NFATc1A in fusionwith FPs were treated overnight with TSA and translocation was initiated bystimulation with ATP in presence or absence of CsA or FK506. Images weretaken at time 0 and 20 min after stimulation with 100 mM ATP.

A.Tasdemir et al.

618

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

In presence of CsA and FK506, the Luc activity wasreduced in a dose dependent manner. The mean IC50 valuesfor these CaN-blockers were independent from the cell lineused. The calculated mean IC50‘s for CsA and FK506 were�286 and 6 nM, respectively (Fig. 5, Table II). Themeasured IC50‘s for FK506 and CsA were consistent withthe reported inhibitory potency of FK506 (Siekierka et al.,1989; Schwaninger et al., 1993). Noteworthy, all NFATc1Aover-expressing cells showed relatively high basal Lucactivities compared with MOCK cells when incubatedwithout ATP (Fig. 7). The higher the observed basal values,

Fig. 5. Concentration response curves for ATP, CsA and FK506 in HeLa/NFAT-Luc cells stably expressing the NFATc1A alone or in fusion with FPs. Celllines are indicated as follows: MOCK (open diamond), NFATc1A-hAsGFP499 (multiplication sign), NFATc1A-As-Dimer1 (plus sign), NFATc1A-As-Mono2(open square), NFATc1A-EGFP (inverted open triangle) and NFATc1A (open triangle). Cells were treated with various concentrations of ATP (A), CsA (B)and FK506 (C). The NFAT mediated Luc activity is given in RLU in (A), (B) and (C). Each data point represents the mean+SE of four replicates.Normalised RLU values are shown for ATP (D), CsA (E) and FK506 (F).

Table II. In cell-based assays calculated mean EC50 and IC50 values for

ATP, CsA and FK506

Cell line EC50 ATP (mM) IC50 CsA (nM) IC50 FK506 (nM)

MOCK 93 163 2.4NFATc1A 214 270 5.0NFATc1A-EGFP 161 415 8.2NFATc1A-hAsGFP499 145 196 5.4NFATc1A-As-Dimer1 104 302 9.4NFATc1A-As-Mono2 85 303 5.3

A monomeric mutant AsGFP499 acts as biomarker for NFAT-assays

619

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

the higher maximal RLU values were observed by treatmentwith ATP.

Discussion

When engineering an oligomeric protein to generate a mono-meric version, knowledge of the structure allows the appli-cation of rational approaches to design mutants with thedesired properties. Since the structure of AsGFP499 wasunknown at the time of this study, secondary structureelements had to be predicted by multiple sequence alignmentand compared with FPs with known 3D(three-dimensional)-structures. SDM was then used to modify amino acidsthat were predicted to contribute to the intermolecular con-tacts responsible for oligomerisation. A dimeric mutant(As-Dimer1) was obtained by mutation of a single residue atthe hydrophobic A/B-interface, namely the polar residue(S103). The replacement of only hydrophobic residues in theA/B-interface had been applied in other oligomeric FPs,including DsRed monomerisation (Campbell et al., 2002;Karasawa et al., 2003; Wiedenmann et al., 2004a,b).

The disruption of the remaining A/C-interface was morechallenging. It has been proposed that manipulations at thisinterface may cause a decrease of the folding temperature,because the interactions here are likely involved in thefolding process of the b-can structure (Wiedenmann et al.,2005). Indeed, all of our A/C-mutants showed fluorescenceat temperatures ,378C. In addition, we generated mutantsthat lost completely fluorescence or led to higher oligomeri-sation states. The optimisation of the maturation temperatureof As-Mono1 was finally achieved by combining Site-directed and random mutagenesis.

The majority of FPs are known to form a compact, rigidglobular b-can structure, presumably to shield the chromo-phore from solvent (Reid and Flynn, 1997; Zimmer, 2002).SEC-MALLS analysis provided not only data on absolutemolecular mass, but also on apparent hydrodynamic size(Rh), which depends on shape (conformation) as well asmass (Table I). As expected from crystal structures of otherFPs, wt AsGFP499 had a compact and tetrameric arrangementsimilar to DsRed2, with Rh values of �3.2 nm. In the caseof As-Dimer1 and As-Dimer2, the Rh values of 2.6 nm werean indication of its more tightly folded state and its expectedstructural arrangements of side-to side b-barrel cylindricalcontacts (rather than an end to end arrangement). The Rh forAs-Mono2 values could not be fitted due to concentrationrestraints. Nevertheless, As-Mono2 is also expected to be atightly folded protein with a buried chromophore that allowsit to exhibit fluorescence. The unique feature of As-Mono2 isthat it can be excited equally at 406 nm and 482 nm using aUV or blue laser, respectively. The QY of the novel dimerswere �0.5 with As-Mono2 having marginally less QY.However, all proteins were extremely fluorescent and QYwere comparable to EGFP (Table I) (Tsien, 1998).

In retrospect, the recently published X-ray structure ofAsGFP499 (Nienhaus et al., 2006) gave us the opportunity toexamine these targeted interface residues. The crystal struc-ture confirmed that all of the predicted and mutated residueswere important for oligomerisation to maintain interfacialinteractions. Consistent with our multi-sequence alignment,analysis of the A/B-interface revealed that S103 in chain Aforms a hydrogen bond with its counterpart in chain

B. Furthermore, the A/C-interface displays all three predictedinteractions: the salt-bridge between R150 and E97, thep-stack between two phenylalanines, F173 from the A- andC-chain, respectively, and the hydrogen-bond between theamino acids T159 and R157 (Fig. 6A and B).

Spectroscopic and oligomeric studies of the mutantsrevealed some interesting features of AsGFP499 and perhapsFPs in general. In SEC experiments, the absorption profile ofthe dimeric fraction (480 . 400 nm) of As-pMono1[AsGFP499(S103/R150E)] turned out to be different from themonomeric fraction (480 � 400 nm) (Fig. 2A). This suggestsan altered balance between the neutral and anionic protona-tion state of the chromophore depending on the oligomerisa-tion state of the protein. Analysis of the crystal structure ofAsGFP499 supports this hypothesis. Closer inspection of theA/C-interface interactions between the chromophore in chainA and the protein matrix of chain C, showed a network ofhydrogen bonds, linking the phenolic oxygen of the chromo-phore in chain A with histidine 227 in chain C via four inter-vening water molecules (Fig. 6C). Thus, it is possible thatthe positively charged protonated histidine residue stabilisesthe negatively charged deprotonated form of the phenolicchromophore through this ‘water route’. This mechanism ofstabilisation would be consistent with the increased absorp-tion at 480 nm of the A/C-dimer. A chromophoric interactionhas also been reported for DsRed within the A/C-interface(Wall et al., 2000).

Application of almost all cellular assay types for highthroughput screening and high content screening are carriedout by using stable cell lines, which guarantees more dataconsistency and less variability compared with cellular assaytypes based on transient gene expression methods (Wilsonet al., 2002). Therefore a further goal of this study was todevelop a cell line stably co-expressing the NFATc1A-FPchimeras and the NFAT-Luc reporter gene to evaluate its usein translocation and functional assays testing compounds thatinteract with the NFAT-signalling pathway. To our knowl-edge, this is the first dual-assay system in one stable cell linespecifically designed for a single member of the NFAT-family.

Wt AsGFP499 in fusion with NFATc1A cannot be used fortranslocation assays, due to formation of aggregates that pre-vented any visible translocation into the nucleus (Fig. 4). Incontrast, EGFP (Kehlenbach et al., 1998; Scott et al., 2001)and the mutants As-Dimer1 and As-Mono2 fusions translo-cated visibly into the nucleus. In the presence of CaN inhibi-tors, CsA and FK506, the translocation of all fusion productswas efficiently blocked (Fig. 4). This demonstrates that thetranslocation event of the chimeras is dependent onNFATc1A, which is a known substrate for CaN.

Wt AsGFP499 and its derivatives do not interfere with thetranscriptional activity of NFATc1A fusions. The transcrip-tional activity of NFATc1A-As-Dimer1 and NFATc1A-As-Mono2 was significantly increased compared with MOCKcells (55-fold and 40-fold, respectively). It is noteworthy thatan 8-fold Luc activity of NFATc1A-hAsGFP499 with respectto the MOCK cells was detected, although we did not observesignificant nuclear import in our translocation assays. Aplausible explanation would be that very low concentrationsof mono- and dimeric NFATc1A-AsGFP499 might exist atequilibrium with the predominant tetrameric form. These non-tetrameric fusion proteins would be capable of translocating

A.Tasdemir et al.

620

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

into the nucleus and could hence be responsible for theobserved increase of transcription activity. In presence of CaNinhibitors the transcriptional activity of untagged and fusionproducts of NFATc1A was blocked. This is consistent with

our results from the translocation assay and confirms that thetranscription of Luc relies on NFATc1A nuclear import. TheEC50 value of ATP and IC50 values of CsA and FK506 deter-mined demonstrate that the signal transduction mechanism ofNFATc1A is not altered by the chimeras (Fig. 5, Table II).

All NFATc1A over-expressing cell lines showed increasedbasal Luc activity in our functional assay (Fig. 7). Thisphenomenon might be a result of a shifted equilibriumbetween phosphorylated and unphosphorylated NFATc1A inthe cytoplasm. Since NFATc1A needs to be phosphorylatedby specific cytoplasmic kinases to prevent its translocationin non-activated cells (Scott et al., 1997), the endogenousamount of kinase activity in NFATc1A over-expressing cellsmight not be sufficient to maintain the entire population ofNFATc1A in the phosphorylated state. The increased basalvalues would thus be an artefact of over-expression.

In summary, the present investigations led to a successfulconversion of tetrameric AsGFP499 into dimeric and mono-meric variants guided by a rational strategy of sequencealignments. The exchange of only three amino acids led to amonomeric variant of AsGFP499. A fourth mutation had tobe introduced in order to increase maturation temperature ofthe protein. This new FP, As-Mono2, shows different spectralproperties and matures in the temperature range compatiblefor its use as fluorescence marker in cellular assay systems.The findings of this study have thus revealed interesting newdetails on the A/B and A/C subunit interactions responsiblefor oligomer formation of FPs. Furthermore, analysis of theabsorbance characteristics of mono- and dimeric variants ofAsGFP499 yielded new insights in protein-chromophoreinteractions mediated by intervening water molecules, whichmay be a general principle of interaction in some FPs with ahigher oligomerisation state. More detailed theoretical inves-tigations of this phenomenon will be an interesting subjectfor further studies.

Finally, using simultaneous stable transfections of HeLacells with the NFAT-Luc reporter element and NFATc1A-As-Dimer1 or NFATc1A-As-Mono2 fusion constructs, we gener-ated for the first time a cell-based assay system allowing theinvestigation of NFATc1A translocation and transcriptionalactivity within the same cell line. This system, and the per-spective to develop further analogous assays by the fusion of

Fig. 6. Molecular representation of the network of interactions at theinterfaces of AsGFP499, (nomenclature of monomers according to the DsRedcrystal structure that is used throughout as reference). (A) A/B (equivalent toA/D in crystal structure of AsGFP499)-interface, showing the interaction ofserine 103 with its counterpart. (B) A/C (equivalent to A/B in crystal structureAsGFP499)-interface, illustrating important interactions for oligomerisation: thesalt-bridge formed by glutamate 97 and arginine 150, p-stacking betweenF173 and its counterpart, p-cation interaction between F173 and R157 and ahydrogen bond between R157 and T159. (C) A water bridge connects thechromophore of chain C with H227 of chain A.

Fig. 7. The histogram shows the basal values of NFAT-driven Luc responsemeasured in various HeLa cell lines stably co-expressing the NFAT response-element either with NFATc1A or NFATc1A in fusion with FPs.

A monomeric mutant AsGFP499 acts as biomarker for NFAT-assays

621

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from

As-Dimer1 or As-Mono2 with other transcription factors,represents a promising tool in the search for new drugs inter-vening with the function of transcription factors.

Supplementary data

Supplementary data are available at PEDS online.

Acknowledgements

We thank Daniele Carettoni for many fruitful discussions and help for thebiochemical analysis, Paolo Guarnieri for the bioinformatic support,Johannes C. Hermann and Tod Flak for reading the manuscript.

ReferencesBadminton,M.N., Campbell,A.K. and Rembold,C.M. (1996) J. Biol. Chem.,

271, 31210–31214.Baird,G.S., Zacharias,D.A. and Tsien,R.Y. (2000) Proc. Natl Acad. Sci.

USA, 97, 11984–11989.Bairoch,A. (2000) Bioinformatics, 16, 48–64.Battistutta,R., Negro,A. and Zanotti,G. (2000) Proteins, 41, 429–437.Beals,C.R., Clipstone,N.A., Ho,S.N. and Crabtree,G.R. (1997) Genes Dev.,

11, 824–834.Berman,H.M., Westbrook,J., Feng,Z., Gilliland,G., Bhat,T.N., Weissig,H.,

Shindyalov,I.N. and Bourne,P.E. (2000) Nucleic Acids Res., 28, 235–242.Brabletz,T., Pietrowski,I. and Serfling,E. (1991) Nucleic Acids Res., 19,

61–67.Campbell,R.E., Tour,O., Palmer,A.E., Steinbach,P.A., Baird,G.S., Zacharias,D.A.

and Tsien,R.Y. (2002) Proc. Natl Acad. Sci. USA, 99, 7877–7882.Chang,C.P., Neilson,J.R., Bayle,J.H., Gestwicki,J.E., Kuo,A., Stankunas,K.,

Graef,I.A. and Crabtree,G.R. (2004) Cell, 118, 649–663.Crabtree,G.R. (1999) Cell, 96, 611–614.Crabtree,G.R. and Olson,E.N. (2002) Cell, 109, S67–S79.Crameri,A., Whitehorn,E.A., Tate,E. and Stemmer,W.P. (1996) Nat.

Biotechnol., 14, 315–319.Cubitt,A.B., Heim,R., Adams,S.R., Boyd,A.E., Gross,L.A. and Tsien,R.Y.

(1995) Trends Biochem. Sci., 20, 448–455.Emmel,E.A., Verweij,C.L., Durand,D.B., Higgins,K.M., Lacy,E. and

Crabtree,G.R. (1989) Science, 246, 1617–1620.Ferrari,D., Stroh,C. and Schulze-Osthoff,K. (1999) J. Biol. Chem., 274,

13205–13210.Graef,I.A., Mermelstein,P.G., Stankunas,K., Neilson,J.R., Deisseroth,K.,

Tsien,R.W. and Crabtree,G.R. (1999) Nature, 401, 703–708.Graef,I.A., Chen,F., Chen,L., Kuo,A. and Crabtree,G.R. (2001) Cell, 105,

863–875.Graef,I.A., Wang,F., Charron,F., Chen,L., Neilson,J., Tessier-Lavigne,M. and

Crabtree,G.R. (2003) Cell, 113, 657–670.Heim,R., Prasher,D.C. and Tsien,R.Y. (1994) Proc. Natl Acad. Sci. USA, 91,

12501–12504.Horsley,V., Jansen,K.M., Mills,S.T. and Pavlath,G.K. (2003) Cell, 113,

483–494.Jain,J., McCaffrey,P.G., Miner,Z., Kerppola,T.K., Lambert,J.N., Verdine,G.L.,

Curran,T. and Rao,A. (1993) Nature, 365, 352–355.Karasawa,S., Araki,T., Yamamoto-Hino,M. and Miyawaki,A. (2003) J. Biol.

Chem., 278, 34167–34171.Karasawa,S., Araki,T., Nagai,T., Mizuno,H. and Miyawaki,A. (2004)

Biochem. J., 381, 307–312.Kehlenbach,R.H., Dickmanns,A. and Gerace,L. (1998) J. Cell. Biol., 141,

863–874.Klee,C.B., Ren,H. and Wang,X. (1998) J. Biol. Chem., 273, 13367–13370.Lauf,U., Lopez,P. and Falk,M.M. (2001) FEBS Lett., 498, 11–15.Lippincott-Schwartz,J. and Patterson,G.H. (2003) Science, 300, 87–91.Liu,J. (1993) Immunol. Today, 14, 290–295.Liu,J., Farmer,J.D., Jr, Lane,W.S., Friedman,J., Weissman,I. and Schreiber,S.L.

(1991) Cell, 66, 807–815.Lukyanov,K.A., et al. (2000) J. Biol. Chem., 275, 25879–25882.Luo,C., Burgeon,E. and Rao,A. (1996) J. Exp. Med., 184, 141–147.Mattila,P.S., Ullman,K.S., Fiering,S., Emmel,E.A., McCutcheon,M.,

Crabtree,G.R. and Herzenberg,L.A. (1990) EMBO J, 9, 4425–4433.Matz,M.V., Fradkov,A.F., Labas,Y.A., Savitsky,A.P., Zaraisky,A.G.,

Markelov,M.L. and Lukyanov,S.A. (1999) Nat. Biotechnol., 17, 969–973.Miyazaki,K. (2003) Methods Mol. Biol., 231, 23–28.

Mizuno,H., Sawano,A., Eli,P., Hama,H. and Miyawaki,A. (2001)Biochemistry, 40, 2502–2510.

Nienhaus,K., Renzi,F., Vallone,B., Wiedenmann,J. and Nienhaus,G.U.(2006) Biophys. J., 91, 4210–4220.

Okamura,H., Aramburu,J., Garcia-Rodriguez,C., Viola,J.P., Raghavan,A.,Tahiliani,M., Zhang,X., Qin,J., Hogan,P.G. and Rao,A. (2000) Mol. Cell.,6, 539–550.

Ormo,M., Cubitt,A.B., Kallio,K., Gross,L.A., Tsien,R.Y. and Remington,S.J.(1996) Science, 273, 1392–1395.

Palm,G.J., Zdanov,A., Gaitanaris,G.A., Stauber,R., Pavlakis,G.N. andWlodawer,A. (1997) Nat. Struct. Biol., 4, 361–365.

Petersen,J., Wilmann,P.G., Beddoe,T., Oakley,A.J., Devenish,R.J.,Prescott,M. and Rossjohn,J. (2003) J. Biol. Chem., 278, 44626–44631.

Prescott,M., Ling,M., Beddoe,T., Oakley,A.J., Dove,S., Hoegh-Guldberg,O.,Devenish,R.J. and Rossjohn,J. (2003) Structure, 11, 275–284.

Randak,C., Brabletz,T., Hergenrother,M., Sobotta,I. and Serfling,E. (1990)EMBO J, 9, 2529–2536.

Ranger,A.M., Grusby,M.J., Hodge,M.R., Gravallese,E.M., de la Brousse,F.C.,Hoey,T., Mickanin,C., Baldwin,H.S. and Glimcher,L.H. (1998a) Nature,392, 186–190.

Ranger,A.M., Oukka,M., Rengarajan,J. and Glimcher,L.H. (1998b)Immunity, 9, 627–635.

Rao,A., Luo,C. and Hogan,P.G. (1997) Annu. Rev. Immunol., 15, 707–747.Reid,B.G. and Flynn,G.C. (1997) Biochemistry, 36, 6786–6791.Remington,S.J., Wachter,R.M., Yarbrough,D.K., Branchaud,B., Anderson,D.C.,

Kallio,K. and Lukyanov,K.A. (2005) Biochemistry, 44, 202–212.Schwaninger,M., Blume,R., Oetjen,E., Lux,G. and Knepel,W. (1993) J. Biol.

Chem., 268, 23111–23115.Scott,J.E., Ruff,V.A. and Leach,K.L. (1997) Biochem. J., 324, 597–603.Scott,E.S., Malcomber,S. and O’Hare,P. (2001) J. Virol., 75, 9955–9965.Shagin,D.A., et al. (2004) Mol. Biol. Evol., 21, 841–850.Shimomura,O., Johnson,F.H. and Saiga,Y. (1962) J. Cell. Comp. Physiol.,

59, 223–239.Siekierka,J.J., Staruch,M.J., Hung,S.H. and Sigal,N.H. (1989) J. Immunol.,

143, 1580–1583.Tsien,R.Y. (1998) Annu. Rev. Biochem., 67, 509–544.Vrzheshch,P.V., Akovbian,N.A., Varfolomeyev,S.D. and Verkhusha,V.V.

(2000) FEBS Lett., 487, 203–208.Wall,M.A., Socolich,M. and Ranganathan,R. (2000) Nat. Struct. Biol., 7,

1133–1138.Wiedenmann,J., Elke,C., Spindler,K.D. and Funke,W. (2000) Proc. Natl

Acad. Sci. USA, 97, 14091–14096.Wiedenmann,J., Schenk,A., Rocker,C., Girod,A., Spindler,K.D. and

Nienhaus,G.U. (2002) Proc. Natl Acad. Sci. USA, 99, 11646–11651.Wiedenmann,J., Ivanchenko,S., Oswald,F., Schmitt,F., Rocker,C., Salih,A.,

Spindler,K.D. and Nienhaus,G.U. (2004a) Proc. Natl Acad. Sci. USA, 101,15905–15910.

Wiedenmann,J., Vallone,B., Renzi,F., Nienhaus,K., Ivanchenko,S.V.,Roecker,C. and Nienhaus,G.U. (2004b) Genetically Engineered andOptical Probes for Biomedical Applications II, vol. 5329, pp. 23–29.

Wiedenmann,J., Vallone,B., Renzi,F., Nienhaus,K., Ivanchenko,S., Rocker,C.and Nienhaus,G.U. (2005) J. Biomed. Opt., 10, 14003.

Wilmann,P.G., Petersen,J., Devenish,R.J., Prescott,M. and Rossjohn,J.(2005) J. Biol. Chem., 280, 2401–2404.

Wilson,V.S., Bobseine,K., Lambright,C.R. and Gray,L.E., Jr (2002) Toxicol.Sci., 66, 69–81.

Winslow,M.M., Pan,M., Starbuck,M., Gallo,E.M., Deng,L., Karsenty,G. andCrabtree,G.R. (2006) Dev. Cell., 10, 771–782.

Yang,F., Moss,L.G. and Phillips,G.N., Jr (1996) Nat. Biotechnol., 14, 1246–1251.Yang,T.T., Cheng,L. and Kain,S.R. (1996) Nucleic Acids Res., 24, 4592–4593.Yarbrough,D., Wachter,R.M., Kallio,K., Matz,M.V. and Remington,S.J.

(2001) Proc. Natl Acad. Sci. USA, 98, 462–467.Zimmer,M. (2002) Chem. Rev., 102, 759–781.

Received April 21, 2008; revised June 27, 2008;accepted July 2, 2008

Edited by Andreas Kungl

A.Tasdemir et al.

622

by guest on June 29, 2014http://peds.oxfordjournals.org/

Dow

nloaded from