Expression and characterisation of Green fluorescent protein mutant

The GFP from the jellyfish Aequora victoria is a small protein consisting of 238 amino acids that has gained widespread scientific attention since the demonstration that the GFP gene can be cloned to generate green fluorescence in organisms other than the jellyfish itself. GFP serves as an indispensible biological tool in the determination of gene expression and localisation of expressed proteins in living organisms, and the development of its variants with altered biochemical properties has further extended its utility in the biosciences . The chromophore, formed spontaneously by the residues Serine-Tyrosine-Glysine(Ser65-Tyr-66-Gly67), is responsible for the green fluorescence. Despite several organisms possessing GFP’s, only the GFP from A.victoria has been most extensively studied and characterised . Mutations in its amino acid sequence can result in fluorescent proteins(FPs) with altered excitation and emission spectra compared to the wt protein . Efforts have been focussed towards the development, identification and application of FPs with novel characteristics to directly visualise dynamic biological processes and the organisation of various molecular structures within cells . In fact, a wide variety of FPs have been developed and made available with fluorescence emission profiles that cover almost the entire spectrum of visible light, which includes cyan, orange, red and blue FPs . The development of CFP has led to improved sensitivity of Fluorescence Resonance Energy Transfer(FRET) applications whereby CFP is employed in pairs with a Yellow Fluorescent Protein(YFP), which allows monitoring of multiple biochemical reactions in cells simultaneously .

In this study, we mutated Tyr66 of the UV-excitable GFPuv gene and expressed and purified the mutated protein CFP. We attempted to analyse some of its properties by means of mass spectroscopy and fluorimetry. The findings of this study may encourage further attempts to alter the fluorescence properties of GFP, by undertaking mutagenesis experiments not only within the tripeptide chromophore but also within other regions of the protein.

2. Materials and Methods

2.1 Sub-cloning of the GFPuv gene from the pET23 into the pET28c expression vector

In this study, an improved version of the wild-type GFP gene, called GFPuv, contained in a pET23 plasmid vector was used as the starting construct. An internal Nde1 restriction site was deleted and a new Nde1 site introduced at the beginning of the GFPuv open reading frame for the purpose of sub-cloning into pET28c. Concentration and purity of pET23-GFPuv plasmid DNA was determined by recording its optical density (O.D) at 260nm and 280nm, using a spectrophotometer. The vector (5µg) was digested with restriction enzymes Nde1 (30U) and Hindlll (30U) at an incubation temperature of 37oC for 3 hours and the restricted products were analysed on a 1% agarose gel. The 761base pair (bp) GFPuv DNA fragment was extracted from the gel by using the QIAquick gel extraction kit (Qiagen) and its concentration was determined on a 1.5% agarose gel. The GFPuv insert was then ligated into the 5.4kb pET28c expression vector under the control of a T7 promoter, which was digested with the same enzymes as the GFPuv gene insert to enable directional cloning. The reaction mixture included 40ng of the insert and 40ng of pET28c plasmid to afford an 8 fold molar excess of the insert, as the vector (5.4kb) was approximately 8 times bigger than the insert (0.7kb). Three control reactions. viz. positive control (undigested pET28c), negative control (insert replaced with water) and sterility water control were also set up. The purpose of including positive and negative controls was to distinguish between transformed and untransformed bacterial colonies. Following an overnight incubation at 16oC, the ligation products were transformed into E.coli DH5α host cells by heat-shocking the mixture at 42oC for 30 seconds and then incubating in SOC growth medium at 37oC for 1 hour, followed by plating of the transformants onto LB agar plates containing 50µg/ml kanamycin and subsequent overnight incubation at 37oC, which would allow growth of only successfully transformed cells as the pET28c vector also contained a kanamycin resistance gene. Transformation efficiencies of the agar plates were calculated the following day. Colony PCR (initial denaturation at 95oC for 1min, denaturation at 94oC for 1min, annealing at 55oC for 1min, elongation at 72oC for 2min) of three transformed colonies from the vector and insert ligation plate was carried out in a thermal cycler for 35 cycles to establish the presence of GFPuv insert in the pET28c plasmid. Primers (0.1µM) complementary to the T7 promoter and terminator region were employed in the PCR reaction mix. A single colony from the undigested pET28c transformation plate was used as the positive control in addition to a sterility water control and subjected to the same. PCR products were analysed on a 1% agarose gel.

2.2 Incorporation of a single codon mutation within the GFPuv gene in the pET28c plasmid and identification of the mutation generated

After successful propagation of pET28c-GFPuv in E.coli DH5α host cells, the plasmids were isolated and subjected to site-directed mutagenesis using the Quik-Change® site directed mutagenesis kit (Stratagene), which employs a hot start KOD polymerase (1U/µl). The codon TAT encoding for the Tyrosine66 (Y66) residue was replaced with codon TGG to encode for the amino acid Tryptophan (W66) by designing appropriate forward (125ng) and reverse (125ng) primer pairs for PCR, which was expected to result in the generation of a CFP possessing the Y66W mutation. The thermal cycler was programmed for 24 cycles with initial denaturation at 94oC for 30sec, denaturation at 94oC for 30sec, annealing at 55oC for 1min, elongation at 68oC for 20sec and a final elongation step at 68oC for 10min. The mutated plasmids were then used to transform E.coli XL1 blue supercompetent cells with undigested pET28c DNA as the positive control. Bacteria were subsequently played onto LB-kanamycin plates followed by overnight incubation at 37oC and calculation of transformation efficiency. Plasmid DNA was extracted the following day using the QIAprep Miniprep kit (Qiagen) and its concentration and purity were determined by Nanodrop technology. Incorporation of the desired Y66W mutation was confirmed by DNA sequencing and ClustalW analysis. The mutated plasmids were transformed into the non-mutagenic, competent E.coli BL21 (DE3) host cells for protein overexpression, which contains a chromosomal copy of the T7 RNA polymerase gene under the control of a lacUV5 promoter, in addition to a negative control transformation, followed by plating onto LB-kanamycin plates, overnight incubation at 37oC and calculation of transformation efficiency respectively.

2.3 Induction of protein expression and detection of the expressed recombinant protein

A single transformation colony was inoculated into 3ml of LB-1D media containing 100µg/ml kanamycin and cultured at 37oC for 6 hours in a shaking incubator (300rpm). 1ml of the culture was transferred to a fresh 1.5ml eppendorf tube as the “non-induced sample”. 15µl of the innoculum culture was transferred to a 250ml flask containing 25ml auto-induction medium (SB-5052) and 100µg/ml kanamycin, followed by incubation at 28oC for 20 hours in a shaking incubator (400rpm). 1ml of the culture was transferred to a fresh eppendorf tube as the “total induced sample”. Cells from both samples were harvested by centrifugation for 5 minutes and resuspended in 100µl SDS-PAGE sample buffer. The remaining innoculum culture was fractionated into “soluble” and “insoluble” samples using Bugbuster technology. All 4 samples were analysed on a 12% polyacrylamide gel and visualised by Coomassie blue staining after heating at 95oC for 5min to detect presence of the mutated GFPuv (CFP). Expressed proteins were also electrophoretically transferred onto a nitrocellulose membrane and the blot was probed with a Hisprobeâ„¢-HRP reagent which was then detected using a chemiluminescent kit.

2.4 Purification of the His-tagged CFP and analysis of its properties by mass spectrometry and fluorimetry

The pET28c expression vector also encodes for six adjacent histidine amino acid residues that result in the expression of an N-terminal 6-His tagged CFP, which can be exploited for use in immobilized metal affinity chromatography. The soluble fraction of the bacterial culture, expected to contain the CFP, was subjected to nickel-nitrilotriacetate ( Ni-NTA) affinity chromatography. Non-specifically bound proteins were washed using appropriate wash buffers and the bound target protein (CFP) was finally eluted with 1M imidazole elution buffer. Soluble, unbound, wash and elution fractions were subjected to SDS-PAGE analysis and Coomassie blue staining as described above. Protein concentration of elution fractions was determined by Bradford assay . Molecular weight of the purified CFP was determined by mass spectroscopy and fluorescence emission spectra recorded by fluorimetric analysis.

3. Results

3.1 Sub-cloning of the GFPuv gene from the pET23 into the pET28c expression vector

Purity of the starting construct pET23-GFPuv plasmid DNA was determined by ratio of A260/A280 .i.e. 1.9, indicating virtually no contamination of protein. Concentration of the plasmid DNA was calculated to be 0.215mg/ml, based on the fact that 1mg/ml of DNA has an A260 of 20.

Restriction digestion analysis of the vector revealed two sharp DNA bands, one in the 3-4kb and another in the 0.5-1kb range, indicating that the 0.761kb GFPuv DNA and the remaining 3.57kb vector were successfully isolated(Fig.1). The GFPuv fragment was extracted from the agarose gel(Fig.2) and its concentration determined by comparing its intensity to a Low Range Mass DNA Ladder. A sharp band in the 700-800bp range confirmed successful recovery of 761bp GFPuv DNA with a concentration of 6ng/ml. Ligation of GFPuv insert into pET28c vector followed by transformation of E.coli DH5α host cells resulted in growth of bacterial cells possessing recombined plasmids and undigested pET28c plasmid as white colonies on the LB-kanamycin plates, which indicated successful transformation.

3.2 Incorporation of a single codon mutation within the GFPuv gene in the pET28c plasmid and identification of the mutation generated

The pET28c vector harbouring/propagating in E.coli DH5α was extracted and subjected to site-directed mutagenesis. The primer pair designed to incorporate the desired mutation .i.e. TAT to TGG, is as shown. Sequencing of the plasmid DNA was carried out to confirm presence of intended Y66W mutation. Multiple sequence alignment of wild-type (wt) GFP and mutant CFP showed >95% homology between the two sequences except for replacement of the codon for Y66 from TAT at 198 base position to TGG, thus confirming effective mutagenesis. ClustalW analysis also revealed a shift in the stop codon of CFP by 10 codons as compared to wt GFP. A chromatogram of the sequencing data was also generated that further displayed the region of DNA bearing required mutation. The mutated plasmid was used to transform XL1 blue supercompetent cells, which were plated onto LB-kanamycin plates. Appearance of bacterial colonies after overnight incubation demonstrated successful transformation. Plasmid DNA was extracted from cell culture obtained from the transformed colonies and its concentration determined by Nanodrop technology was reported to be 72.43ng/ml with an A260/A280 ratio of 1.97 representing high purity plasmid preparation. Mutant plasmid DNA was transformed into E.coli BL21 (DE3) cells for protein expression, which grew into colonies on the LB-kanamycin plates, thus demonstrating successful transformation.

3.3 Induction of protein expression and detection of the expressed recombinant protein

pET28c vector allowed expression of CFP under the control of an IPTG/lactose inducible T7lac promoter. SDS-PAGE analysis of the soluble fraction showed a protein band in the 24-33kDa range, indicating presence of CFP. Induced and uninduced fractions showed a large amount of protein on the gel as expected. Presence of the ~29kDa CFP in the soluble fraction was confirmed by Western Blotting, which showed a protein band in the 24-33kDa range (Fig.3). The insoluble fraction lacked any protein bands since it was free of CFP. The induced and uninduced fractions also resulted in protein bands in the same range.

3.4 Purification of the His-tagged CFP and analysis of its properties by mass spectrometry and fluorimetry

6 His-tagged CFP was purified using Ni-NTA affinity chromatography. Purified recombinant CFP appeared as single band between 24-33kDa in SDS-PAGE analysis (Fig.4), which was also in accordance with its calculated molecular mass of 29.6kDa. Concentrations of elution1 and elution2 were calculated to be 0.204mg/ml and 0.104mg/ml respectively, by using the Bradford Assay. Mass Spectroscopy data of the affinity purified protein displayed a molecular mass of 29.597k Da. Fluorescence spectra of the protein at a concentration of 10µg/ml exhibited a fluorescence emission maximum of 2039.1 (a.u) at 486nm when excited at 440nm with a scan rate of 1nm/s(Fig.5).

4. Discussion

The study demonstrates that the GFPuv gene can be successfully mutated and stably expressed in the cells of E.coli, which is deficient in proteases that can degrade the intact recombinant protein . In this study, CFP, a mutant of the wt-GFP with altered fluorescence properties, as discussed in previous studies , has been described and isolated by site-directed mutagenesis. For this purpose, Tyr66, located in the chromophore of the wt-GFP was deliberately mutated to Trp66, giving rise to a cyan shifted fluorescent protein with an emission spectrum peaking at 486nm (Fig.5), in close association with its fluorescence characteristics reported in previous studies . When compared to wt-GFP, the emission spectrum of the protein extract confirmed the substantial variation between the Y66W variant and the wt-protein which exhibits two fluorescence emission maxima at 503 and 508nm. Cyan fluorescence is afforded by the formation of a mutated chromophore in which the usual phenol group of the wt-GFP is replaced by an indole group . Hence, the mutation Tyr66→Trp can be concluded to be responsible for the altered emission spectra of GFP.

DNA sequencing of the coding region also revealed a successful single amino acid mutation, changing the codon of Y66 from TAT to TGG, which results in a tryptophan residue in this region. All experimental transformations yielded positive results, wherein bacterial cells appeared as colonies on LB-kanamycin plates, which indicated a high success rate of the experimental procedures. Absolute sterility was maintained throughout the practical investigation which was confirmed by absence of bacterial colonies on all sterility water control plates.

The coding region of wt-GFP was cloned into the pET28c T7 expression vector and subsequently mutated, resulting in a polyhistidine tagged fusion protein (CFP) expressed in E.coli BL21 (DE3). It was also evident that fusion of the 6-His tag to the N-terminus of GFP did not affect the fluorescence characteristics of the recombinant protein. However, it was observed that the fluorescence intensity emitted by CFP was less than either of the two emission maxima of wt-GFP. This could be due to the possibility that substitution at Tyr66 position affects the formation of the chromophore and makes it less/more efficient which can be confirmed by further investigating the changes involved in the chromophore maturation pathway.

The use of a polyhistidine tag proved beneficial, since it has a high affinity for Ni-NTA resin , which permitted single step purification of the recombinant fusion protein, thus contributing to a more efficient purification approach as affinity chromatography is superior to other purification methods such as ion-exchange chromatography and hydrophobic interaction chromatography which generally afford low purity and poor yield of fusion proteins .

Appearance of a single protein band in the 24-33kDa range was detected in the soluble fraction during Western Blotting which confirmed the presence of the ~29kDa recombinant CFP in the sample and demonstrated effective purification. Absence of a band in the insoluble fraction indicated that all of the protein was expressed in a soluble form. The unintended presence of a similar protein band in the apparent uninduced fraction indicated loading error.

Purified CFP eluted from the Ni-NTA resin depicted an apparent molecular weight between 24-33kDa when subjected to SDS-PAGE analysis, in good agreement with its molecular weight of 29.597kDa as determined by mass spectroscopy.

The development of CFP is particularly valuable as it can be used for multicolour labelling and multiparameter imaging. The findings of this study may prove/prove to be useful in generating significant scientific interest to further investigate the structural and biochemical properties of the CFP relative to the wt-GFP.

CFP can be mutated further to generate more photostable and brighter versions of the protein. The study can be extended by obtaining/deducing X-ray crystallographic/3D structures of the purified protein and understanding the mechanisms involved in the maturation of its chromophore or its photoactivation, which would further advance our knowledge about engineering and developing fluorescent proteins with novel and enhanced desirable characteristics/properties.

Figure 1: Electrophoresis pattern of pET23-GFP DNA digested with restriction enzymes Nde1 and HindIII.

Double digestion products compared on a 1% agarose gel. Lane 1: 0.5µg of 1kb DNA ladder. Lane 2: 4.33kb uncut pET23-GFP. Lanes 3-4: 0.761kb GFP fragment and remaining 3.57kb pET23 vector.

Figure 2: Electrophoresis pattern of the recovered 761bp GFPuv DNA fragment.

5µl aliquot containing the eluted GFP DNA fragment analysed on a 1.5% agarose gel. Lane 1: 761bp GFPuv DNA fragment. Concentration was estimated to be 6ng/ml. Lanes 2-4: 2µl, 5µl and 10µl of Low range DNA mass ladder respectively.

Figure 4: SDS-PAGE analysis after Coomassie blue staining.

Results of SDS-PAGE of CFP purified by Ni-NTA chromatography. Total soluble, unbound samples(lanes1-2), washes 1,2(lanes 3-4) and elutions 1,2(lanes 5-6) demonstrate non-specific binding, indicating possibility of excessive beads in Ni-NTA resin. However, expression of ~29kDa CFP can be observed in lane 5(elution1).

Figure 3: Western Blot analysis.

Western blot analysis of fractionated protein samples. M: Protein molecular weight marker. Lane 1: Protein band indicates loading error. Lane 2: Expected presence of ~29kDa CFP. Soluble expression of ~29kDa CFP can be observed(lane 4) in contrast to the insoluble fraction(lane 3).

Figure 4: SDS-PAGE analysis after Coomassie blue staining.

Results of SDS-PAGE of CFP purified by Ni-NTA chromatography. Total soluble, unbound samples(lanes1-2), washes 1,2(lanes 3-4) and elutions 1,2(lanes 5-6) demonstrate non-specific binding, indicating possibility of excessive beads in Ni-NTA resin. However, expression of ~29kDa CFP can be observed in lane 5(elution1).

Figure 3: Western Blot analysis.

Western blot analysis of fractionated protein samples. M: Protein molecular weight marker. Lane 1: Protein band indicates loading error. Lane 2: Expected presence of ~29kDa CFP. Soluble expression of ~29kDa CFP can be observed(lane 4) in contrast to the insoluble fraction(lane 3).

Figure 5: Fluorescence emission spectrum of CFP.

Graph showing fluorescence emission spectrum of purified soluble cell extract containing 10 µg/ml recombinant CFP and peaking at 486nm, which was recorded at an emission wavelength ranging from 460-520nm. Slit Width used was 2.5nm and scan rate was 1nm/s.

Figure 5: Fluorescence emission spectrum of CFP.

Graph showing fluorescence emission spectrum of purified soluble cell extract containing 10 µg/ml recombinant CFP and peaking at 486nm, which was recorded at an emission wavelength ranging from 460-520nm. Slit Width used was 2.5nm and scan rate was 1nm/s.

Acknowledgement

We are extremely grateful to Dr. Aysha Divan (Institute of Molecular and Cellular Biology, University of Leeds, UK) for suitably organising the laboratory sessions and for providing constant help and support throughout the course of the study, without which this work would not have been possible. We would like to thank our demonstrator, Mr. Blessing, for his expert suggestions and helpful discussions. We also thank Mrs. Jindi and Mrs. Poli for excellent technical assistance. CFP was selected for further characterisation by mass spectroscopy and fluorimetry.