Deimmunisation of Enhanced Green Fluorescent Protein

DEIMMUNISATION OF ENHANCED GREEN FLUORESCENT PROTEIN

A thesis submitted to The University of Manchester for the degree of DOCTOR OF PHILOSOPHY in the Faculty of Life Sciences

2014

TIMOTHY JOHN EYES

Contents

Contents ...... 2

List of Figures ...... 8

List of Tables...... 11

List of Abbreviations...... 12

Abstract ...... 14

Declaration ...... 15

Acknowledgements ...... 17

Chapter 1

1. Introduction...... 18

1.1. Green Fluorescent Protein...... 18

1.1.1. Discovery of GFP ...... 18

1.1.1. The Structure of Green Fluorescent Protein...... 30

1.1.2. The Chromophore of Green Fluorescent Protein...... 32

1.1.3. Applications of Fluorescent Protein Technology ...... 40

1.1.3.1. Gene expression reporters...... 40

1.1.3.2. Protein Tagging...... 40

1.1.3.3. Biosensors ...... 41

1.1.3.4. Protein Folding...... 41

1.1.3.5. Whole organism imaging...... 41

1.2. Immunogenicity of GFP...... 43

1.2.1. Recognition of transgenic antigens by the immune system...... 45

1.2.2. Discovery of immunodominant epitopes in EGFP...... 51

1.3. Hypothesis, Aims and Experimental Approach...... 53

1.3.1. Hypothesis ...... 53

1.3.2. Aims ...... 53

1.3.3. Experimental approach ...... 53

Chapter 2

2. Rationale for deimmunisation of EGFP...... 55

2.1. Aims...... 56

2.2. Methods ...... 56

2.2.1. Prediction of immunodominant epitopes in EGFP...... 56

2.2.2. Anchor site alanine scanning...... 56

2.2.3. Homology modelling of EGFP epitope/MHC complex...... 57

2.2.4. Identification of epitope anchor site residues in EGFP structure...... 57

2.3.1. Confirmation of immunodominant epitopes in EGFP...... 58

2.3.1.1. BALB/c H2-k ...... 58

2.3.1.2. C57BL/6 H2-d ...... 58

2.3.2. Anchor site alanine scanning...... 61

2.3.2.1. HYLSTQSAL ...... 61

2.3.2.2. DTLVNRIEL ...... 61

2.3.3. Homology modelling of epitopes ...... 63

2.3.3.1. HYLSTQSAL (H2-Kd) ...... 63

2.3.3.2. DTLVNRIEL ...... 66

2.3.4. Location of anchor site residues within EGFP...... 69

2.3.5. Summary ...... 72

Chapter 3

3. Mutagenesis of Enhanced Green Fluorescent Protein MHC Anchor Sites...... 74

3.1. Methods ...... 74

3.1.1. Construction of EGFP expression vector...... 74

3.1.2. E.coli Transformation and plasmid preparation...... 75

3.1.3. Transformation of expression cell strain BL21(DE3) ...... 75

3.1.4. Site Directed Mutagenesis of MHC anchor sites...... 76

3.1.5. Pilot Recombinant Expression of EGFP in E.Coli...... 78

3.1.6. Pilot IMAC chromatography ...... 78

3.1.7. Analysis of chromatography fractions by SDS PAGE...... 79

3.2. Scaled up expression and chromatography...... 79

3.2.1. Chromatography...... 80

3.2.1.1. IMAC purification using AKTA primeTM system...... 80

3.2.1.2. Size Exclusion Chromatography using AKTA primeTM system...... 80

3.2.1.3. Protein Quantification...... 81

3.2.1.4. Removal of Histidine Tag by Thrombin digestion...... 81

3.2.1.5. Analysis of samples by fluorescence spectroscopy...... 81

3.2.2. Mammalian Expression and Imaging of Y200 variants...... 82

3.2.3. Cloning of Y200 variants into mammalian expression vector...... 82

3.2.4. Transfection and imaging of mammalian cells with EGFP Y200 variants. 83

3.3. Results...... 85

3.3.1. Cloning, Expression, Purification of Enhanced Green Fluorescent Protein.

3.3.2. Histidine Tag Removal...... 92

3.3.3. Site Directed Mutagenesis...... 92

3.3.4. Y200 Mutation Library...... 95

3.3.5. Spectral Characterisation of Y200 variants...... 97

3.3.6. Size exclusion chromatography of Y200 variants...... 101

3.3.7. Mammalian Cell Imaging of Y200 variants...... 101

3.3.8. N121 Mutation Library...... 104

3.4. Summary ...... 109

Chapter 4

4. An examination of the immunogenicity of EGFP and EGFP Y200T in BALB/c mice ...... 111

4.1. Aims ...... 112

4.2. Methods ...... 112

4.2.1. Maintenance of A20 cell line ...... 112

4.2.2. Construction of stable A20 EGFP expressing cell line...... 113

4.2.3. BMDC Culture ...... 113

4.2.4. Transduction of BMDC with EGFP lentivirus ...... 114

4.2.5. Splenocyte isolation and culture ...... 114

4.2.6. Immunisation of Mice ...... 115

4.2.6.1. Recombinant EGFP protein ...... 115

4.2.6.2. Immunisation with A20 cell line and BMDC ...... 115

4.2.7. Assessment of EGFP and membrane marker expression by flow cytometric analysis. 116

4.2.8. ELISA ...... 116

4.2.8.1. Mouse interferon gamma ELISA ...... 116

4.2.8.2. Mouse granzyme B ELISA ...... 117

4.2.8.3. Serum anti-EGFP IgG ELISA ...... 118

4.2.9. Cytotoxic T cell Lymphocyte Assay...... 119

4.2.9.1. Splenocyte responder cell isolation and culture...... 119

4.2.9.2. Stimulator cell preparation...... 119

4.2.10. Sensitisation co-culture...... 120

4.2.11. Radiolabelling of A20 target cells ...... 120

4.2.12. Cytotoxicity Assay ...... 121

4.2.13. Statistical Analysis ...... 122

4.3. Results ...... 123

4.3.1. Generation of a A20/EGFP stable cell line...... 123

4.3.3. Cytotoxicity marker ELISA...... 134

4.3.4. Generation of a cytotoxic immune response to EGFP in BALB/c mice. .. 136

4.3.4.1. Immunisation with recombinant EGFP...... 136

4.3.5. Immunisation with heat denatured recombinant EGFP...... 139

4.3.6. Generation of a cytotoxic response to EGFP by immunisation with bone marrow dendritic cells expressing EGFP...... 143

4.3.7. Generation of an immune response to EGFP by immunisation with A20 cell line expressing EGFP...... 148

4.4. Summary ...... 159

Chapter 5

5. Discussion ...... 160

5.1. Prediction and assessment of MHC class I epitopes in EGFP...... 160

5.2. Recombinant Expression of EGFP in E.coli and library screening...... 164

5.3. Mutagenesis of anchor sites...... 165

5.4. Development of a CTL assay to assess EGFP mediated cytotoxicity ...... 166

5.4.1. EGFP/Y200T eliminates a humoral antibody response to EGFP ...... 169

5.4.2. Future Aims ...... 175

5.4.3. Concluding remarks...... 177

6. References ...... 178

7. Appendices...... 196

7.1. Oligonucleotide primers ...... 196

7.2. Vector Maps ...... 197

List of Figures

Chapter 1

Figure 1.1 The impact of green fluorescent protein in scientific literature ...... 19

Figure 1.2 Images of Aequorea victoria photo organs ...... 21

Figure 1.3 Jablonkski diagram illustrating FRET between Aequorin and GFP...... 23

Figure 1.4 Structure of the chromophore of Aequorea GFP...... 25

Figure 1.5 Absorbtion and emission spectra for (a) wild type GFP and (b) S65T /

EGFP...... 27

Figure 1.6 Primary amino acid sequence and structure of Aequorea victoria Green

Fluorescent Protein...... 31

Figure 1.7 Proposed scheme for the formation of the GFP chromophore...... 33

Figure 1.8 Schematic diagram of the interactions between the chromophore and its surroundings in EGFP...... 34

Figure 1.9. Aequorea GFP mutation map...... 36

Figure 1.10 Chromophore diversity generated in GFP-like proteins...... 39

Figure 1.11 Examples of Fluorescent Protein FRET Biosensors...... 42

Figure 1.12. Antigen presentation pathway for MHC1...... 46

Figure 1.13. Depictions of MHC class I molecule bound to antigen peptide ...... 48

Chapter 2

Figure 2.1 Predictive docking of EGFP epitope HYLSTQSAL with H2-Kd MHC by homology modelling...... 64

Figure 2.2. Model of the HYLSTQSAL Tyr200 anchor site interaction with the

MHC H2-Kd binding pocket...... 65

Figure 2.3 Predictive docking of EGFP epitope DTLVNRIEL with known H2-Db epitope by homology modelling...... 67

Figure 2.4. Model of the DTLVNRIEL Asn121 anchor site residue interaction with the MHC H2-Db binding pocket...... 68

Figure 2.5. Location of the dominant anchor site Tyr200 in EGFP...... 70

Figure 2.6. Location of the dominant anchor site Asn121 in EGFP...... 71

Chapter 3

Figure 3.1. Cloning, plasmid construction and mutagenesis of EGFP gene ...... 86

Figure 3.2. DNA Sequence confirmation of EGFP cloning into pET15b ...... 87

Figure 3.3 Recombinant expression of pET15b EGFP in E.coli ...... 89

Figure 3.4 Chromatography of 1L scale EGFP expression...... 90

Figure 3.5 Thrombin cleavage of His Tag from EGFP ...... 91

Figure 3.7. Excitation and emission spectra of candidate Y200 variants...... 99

Figure 3.8. Size exclusion chromatography of candidate Y200 variants ...... 102

Figure 3.9. Live imaging of HEK293 cells expressing EGFP constructs ...... 103

Figure 3.10. Excitation and emission spectra of candidate N121 variants...... 107

Chapter 4

Figure 4.1 Flow cytometric characterisationof A20 cell line...... 124

Figure 4.2 Flow cytometric analysis of EGFP expressing cell lines...... 125

Figure 4.3 Development of thymidine release cytotoxic T cell lymphocyte assay with a mixed lymphocyte reaction...... 130

Figure 4.4 Development of thymidine release cytotoxic T- cell lymphocyte assay with target A20 cell line...... 131

Figure 4.5 Dose effect of cytotoxic T lymphocyte A20 target cell killing...... 133

Figure 4.6 Secretion of cytotoxcity markers IFNγ and Granzyme B in a mixed lymphocyte reaction...... 135

Figure 4.7Anti-EGFP serum ELISA ...... 137

Figure 4.8 Anti-EGFP serum ELISA IgG isotyping...... 138

Figure 4.9 IgG response to heat denatured EGFP recombinant protein...... 141

Figure 4.10 Cytotoxic response to A20/EGFP in mice treated with denatured EGFP recombinant protein...... 142

Figure 4.11 Flow cytometric characterisation BMDC transduced with EGFP lentivirus...... 145

Figure 4.12 IgG response to EGFP in mice immunised with BMDC expressing

EGFP...... 146

Figure 4.13 Cytotoxic response to A20/EGFP in mice immunised with BMDC expressing EGFP...... 147

Figure 4.14 Cytotoxic response to A20 cell lines in mice immunised with A20 cells expressing EGFP...... 150

Figure 4.15 Washing cells prior to immunisation reduces FBS related IgG response.

...... 151

Figure 4.16 Serum IgG response to EGFP in mice immunised with EGFP expressing cells from tissue culture...... 152

Figure 4.17 Secretion of cytotoxicity markers IFNγ and Granzyme B in mice immunised with A20/EGFP...... 155

Figure 4.19 IgG1 and IgG2a response to EGFP in mice immunised with A20 cells expressing EGFP and EGFP/Y200T...... 158

Chapter 5 Figure 5.1 Proposed model for the elimination of an antibody response to EGFP by the removal of the immundominant MHC class I epitope...... 173

List of Tables

Chapter 1

Table 1.1. The range of commercially available fluorescent proteins ...... 29

Table 1.2. MHC alleles expressed by commonly used inbred mouse strains...... 50

Chapter 2

Table 2.1 Predicted MHC class I epitopes of EGFP in BALB/C mice...... 59

Table 2.2 Predicted MHC class I epitopes of EGFP in C57BL/6 mice ...... 60

Table 2.3 Impact of alanine scanning on predicted immunogenicity of EGFP immunodominant epitopes...... 62

Chapter 3

Table 3.1 Y200 Mutation Library ...... 93

Table 3.2. Impact of the isolated Y200 mutations on the predicted immunogenicity of the H2-Kd immunodominant epitope ...... 96

Table 3.3 Spectral characteristics of candidate Y200 variants ...... 100

Table 3.4 N121 Mutation Library ...... 105

Table 3.5. Impact of the isolated N121 mutations on the predicted immunogenicity of the H2-Db immunodominant epitope ...... 106

Table 3.6. Spectral characteristics of N121 variants ...... 108

WORD COUNT: 34,701

List of Abbreviations

Common amino acid abbreviations are used.

7-AAD 7-aminoactinomycin D Ab antibody ANOVA analysis of variance Amp ampicillin Bp base pair BMDC bone marrow derived dendritic cell(s) DMSO dimethyl sulfoxide BSA bovine serum albumin CD4 cluster of differentiation 4, T helper cell marker CD8 cluster of differentiation 8, cytotoxic T cell marker CTL cytotoxic T cell lymphocyte CV column volume Da dalton DsRed Discosoma red fluorescent protein DTT dithiothreitol EDTA ethylenediaminetetraacetic acid EF1α elongation factor 1 alpha promoter. EGFP enhanced green fluorescent protein (F63L, S65T) ELISA enzyme-linked immunosorbant assay ER endoplasmic reticulum FACS fluorescence activated cell sorting FBS foetal bovine serum FRET fluoresence resonance energy transfer FT flow through FP fluorescent protein GFP green fluorescent protein GM-CSF granulocyte-macrophage colony-stimulating factor g gram(s) h hour(s) HLA human leukocyte antigen IC50 half maximal inhibitory concentration IEDB Immune Epitope Data Base IFN interferon Ig immunoglobulin IL interleukin IMAC immobilised-metal affinity chromatography IPTG isopropyl β-D-1-thiogalactopyranoside kDa kilodalton LB lysogeny broth M Molar MFI mean fluorescent intensity

MHC major histocompatibility complex min minute(s) MW molecular weight MWCO molecular weight cut-off NK cells natural killer T cell(s) OD optical density OPD O-phenylenediamine dihydrochloride PBS phosphate buffered saline PCR polymerase chain reaction PI propidium iodide ROS reactive oxygen species RT room temperature SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SEC size-exclusion chromatography SEM standard error of the mean TAP transporter associated with antigen processing Tc cytotoxic T cell TCR T cell receptor Th T helper cell TMB tetramethylbenzidine Treg T regulatory cell U Units UHP urea hydrogen peroxide UV ultraviolet v/v volume/volume WHO world health organisation w/v weight/volume

Abstract

The University of Manchester

Timothy John Eyes

Doctor of Philosophy

Deimmunisation of Enhanced Green Fluorescent Protein

2014

The discovery and application of GFP (Green Fluorescent Protein) has made a very significant contribution to biological research. Detection of a genetically encoded fluorescent signal has proven to be an extremely powerful tool for monitoring biological events both in vitro and in vivo. Whole organism applications for EGFP (enhanced GFP, the widely used variant of GFP) span a broad range of key disciplines which encompass cell biology, immunology and developmental biology. Furthermore, EGFP has specific applications in monitoring tumour development, as an indicator of genetic modification and cell tracking in vivo. However, there is strong evidence that EGFP elicits an immune response when used in animal models. Stimulation of a host immune response can have very serious consequences for the integrity and reliability of experimental studies, including the rejection of cells expressing EGFP. The aim of this study was to develop an EGFP construct that has reduced immunogenicity in the BALB/c model strain but maintains the fluorescent functionality to the original EGFP. The major histocompatability complex (MHC) class I immunodominant epitopes of EGFP in both BALB/c and C57BL/6 strain mice have been described previously. EGFP was recombinantly expressed in E.coli and saturation mutagenesis was used to generate a library of EGFP variants where the MHC class I anchor site in the dominant epitope is mutated to reduce immunogenic potential, termed ‘deimmunisation.’ The library was screened for fluorescent EGFP variants which possessed anchor site mutations that reduced immunogenicity based on computational prediction. The EGFP Y200T variant was isolated with near identical fluorescent properties to the parent EGFP. BALB/c mice were immunised with a tumour cell line expressing EGFP or EGFP Y200T. The Y200T mutation completely abrogated a humoral antibody response to EGFP in vivo. The data suggests that through a single mutation the immunodominant epitope in EGFP can be disrupted and tolerance enhanced in mice. This presents a novel strategy for improving tolerance to heterologously expressed proteins without the requirement for immunosuppressive intervention.

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

Copyright statement i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-property.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses.

Acknowledgements

I would like to thank my supervisors, Professor Andrew Doig and Professor Ian

Kimber for their guidance and advice. Special acknowledgement goes to Dr Rebecca

Dearman for her supervisory support and Lorna Beresford for her technical assistance.

I am very grateful to the BBSRC and UMI3 for the funding that has enabled me to carry out the project.

I would also like to thank my laboratory group members and all the support staff in the Manchester Institute of Biotechnology, the Michael Smith Building, the

Biological Sciences Facility and the Faculty of Life Sciences.

I am very grateful to all my family and friends who have continually supported me throughout my studies; my parents John and Liz for encouraging my scientific interest from a young age, my brother Ben, my good friend Richard and above all my wife Joanne, who has supported me throughout the highs and lows of my research.

Chapter 1

1. Introduction.

1.1. Green Fluorescent Protein.

1.1.1. Discovery of GFP

The application of green fluorescent protein (GFP) in biological research is often described as revolutionary, likened to the “microscope of the 21st century” [1]. GFP technology has given scientists the ability to detect and visualise the otherwise invisible molecular machinery of the cell.

In recognition of GFP’s significance, the 2008 Nobel Prize for Chemistry was awarded to three key researchers, Osamu Shimomura, Martin Chalfie and Roger Y.

Tsien for their contribution to the identification, cloning and engineering of GFP respectively [2]. As a bench mark of the extent to which GFP has penetrated the scientific community, a quick literature search illustrates how the number of publications has increased over the past 50 years since the chemistry of bioluminescence was first described in jellyfish [3] (see Figure 1.1). A noticeable acceleration in the number of publications occurs in the mid 90’s which coincides with Martin Chalfie’s idea that GFP could be recombinantly expressed in E.coli as a gene expression reporter [4]. Several years later the discovery of red fluorescent protein [5] in coral drove research to generating a broad colour palette of fluoresence and opening up many new application areas. Since then the number of publications has snowballed, approaching 10,000 published papers demonstrating the broad applications and value of fluorescent proteins as research tools. Bearing in mind GFP is now considered a standard research tool the number of studies using GFP without citation is likely to be several fold more.

10000

9000

8000

7000

6000

5000

4000

Total Number Citations Number of Total 3000

2000

1000

0 1990 1995 2000 2005 2010 2015 Year of Publication

Figure 1.1 The impact of green fluorescent protein in scientific literature. The total number of citations containing the search term ‘green fluorescent protein’ since 1990. Publications accumulated per year since 1990, source Science Direct [6].

The story of how GFP went from a somewhat understudied protein from an innocuous jellyfish to being a ubiquitously used and commercially successful molecular biology tool is a good example of how bright ideas and chance collaborations can lead to ingenious applications. In the summer of 1960, Osama

Shimomura was recruited by Professor Frank Johnson at Princeton University.

Shimomura had successfully studied the bioluminescent mechanism from Cypridina

(a marine crustacean) luciferase back in his home country and had been recruited by the US laboratory [7]. On arrival in New Jersey he was given the challenge of determining the bioluminescent machinery that caused a common jellyfish,

Aequorea victoria, to glow green when agitated. These small (7-10cm) jelly fish possessed light emitting organs at the periphery of their mantle, see figure 1.2.

Aequorea victoria were abundant during the Summer periods on the Northeastern

Pacific coast, so a research team had to make the 500km drive to University of

Washington’s Friday Harbour Laboratory every year to gather specimens[7].

Figure 1.2 Images of Aequorea victoria photo organs Individual photo organs in the mantle, under magnification (left), the whole jellyfish in the dark demonstrating bioluminescence (middle) and under visible light (right). Taken from Zimmer [7].

The team managed to collect and process over 10,000 jellies to be able to purify the protein responsible for bioluminescence, named aequorin [8]. Aequorin was found to luminesce with blue light on addition of calcium, but it was not understood why in- vivo it was green. However there was another protein purified alongside aequorin;

“A protein giving solutions that look slightly greenish in sunlight though only yellowish under tungsten lights, and exhibiting a very bright, greenish fluorescence in the ultraviolet of a Mineralite.” [8]

The pigmented protein they had isolated was later named green fluorescent protein

(GFP), but its future potential had yet to be realised and study continued to focus on the mechanism of aequorin luminesence. By 1974, through painstaking collection of by now hundreds of thousands of jellyfish, they had gathered enough protein to determine that the function of aequorin and GFP were infact intertwined [9].

GFP’s photochemical properties were analysed and it was found to have major and minor peak excitation at 395nm (near UV) and 475nm (blue light) respectively with a single emission at 509nm (green light). Aequorin had a peak emission of 472nm and when the two were mixed in the presence of a calcium activator, the emission was equivalent to that of the green emission measured in-vivo. There was a fluorescent radiationless energy transfer (FRET) with the conversion of the blue luminescence from Aequorin to green by GFP and this was responsible for the

Aequorea victoria’s natural green glow (see figure 1.3). Attention then turned to the investigation of how GFP contributed to this phenomenon.

Excited State

Green light Emitted Green light Emitted

Energy Level Blue light

(from Aequorin)

Ground State

Figure 1.3 Jablonkski diagram illustrating FRET between Aequorin and GFP. Red lines indicate the energy level of chromophore electrons. Blue light emitted by chemiluminesence of aequorin is absorbed by GFP, which elevates the chromophore energy state, symbolised by the round marker, from the ground state to an excited state. During energetic relaxation back to the ground state photons are released at a lower energy state and longer wavelength i.e. green light [10].

Through proteolysis of GFP, Shimomura isolated a coloured, but non-fluorescent peptide fragment. The peptide contained the chromophore of GFP which is responsible for its’ fluorescence properties [11], see figure 1.4.

In 1994 Douglas Prasher published the cloning and primary sequence of the GFP gene [12] though the product was not fluorescent. Martin Chalfie of the University of

Colombia, while attending a seminar on bioluminescence, heard about GFP and imagined its potential as a fluorescent genetic tag for his work on C. Elegans [7]. He contacted Prasher about this possibility but at this point it was still suspected that chromophore formation required a specific post translational modification and therefore would not be particularly useful as a single unit genetic tag. Prasher passed the project to Chalfie who discovered the cloned gene needed truncating and shortly afterwards demonstrated GFP expression in E.coli, which fluoresced under UV light in exactly the same manner as in Aequorea victoria. This was a breakthrough as it demonstrated that fluorescence was intrinsic to the GFP protein alone, with no additional cofactors or modifications required [7]. Thus the gene could in theory be introduced into any type of cell and made to fluoresce. In his landmark paper Chalfie demonstrated GFP expression in C.elegans to monitor mechanoreceptor gene expression [4] and shortly after the first successful GFP-fusion protein was announced [13].

Figure 1.4 Structure of the chromophore of Aequorea GFP. The cyclisation of the S65/Y66/G67 sequence forms a 4-(p- hydroxybenzylidene)-5-imidazolone structure. R1 and R2 refer to the rest of the GFP protein, N and C-terminal respectively. The series of single and double bonds within the ringed structure generates an extended π-electron system giving rise to the fluorescence property. Taken from Shimomura [11].

During this groundbreaking research, the question still stood as to how GFP’s chromophore formed if no enzymes were required. Roger Tsien (University of

Califonia) had also been working on GFP in a similar direction to Chalfie. He was able to describe an autocatalytic pathway for the formation of the chromophore. The chromphore is produced from a covalent cyclisation event of core amino acids

Ser65, Tyr66 and Gly67 that form a conjugated electron system [14] (see 2.1.3 ).

Since then only one other protein, histidine ammonium lyase (HAL) has been discovered that auto catalytically forms a cyclic peptide structure in a similar manner to GFP [15, 16].

Tsien’s team continued to use his knowledge of chromophore formation and through targeted mutagenesis produced a blue emission (Blue FP) variant which could then be used alongside GFP for imaging [14]. Tsien found the dual peak absorbtion of GFP was due to the presence of two charge states of the chromophore, neutral (395nm) and anionic (475nm) (Figure 5). The ratio between the two states depended on the proteins environment, pH and concentration. Having a mixture of chromophore species was disadvantagous as it leads to photoisomerisation, a swapping between the two states with a gradual dimming during imaging. Through mutational studies Tsien found that the S65T mutation at the chromophore removed the potential for isomersation and reduced the absorption spectra to a single peak at

484nm, six times brighter than GFP with reduced chromophore maturation time.

This was named enhanced GFP (EGFP) and is the now basis for the most commonly used GFP constructs [17, 18].

ised Absorption / Absorption Emission ised Normal

Figure 1.5 Absorbtion and emission spectra for (a) wild type GFP and (b) S65T / EGFP. The wild type protein shows biomodal absorbtion at 395nm and 475nm.The 395nm absorbtion peak in the wild type protein is eliminated by the S65T mutation (EGFP) leading to improved brightness under blue light excitation. Taken from Piston et al [19].

The crystal structure of GFP was solved [20, 21] and an understanding of the chromphore environment allowed for further rational mutation to produce differently coloured variants [22]. With the advent of a colour palette increasing in size, a red shifted fluorescent protein still evaded protein engineering efforts. Mikhail Matz a researcher at the Russian Academy of Science made the discovery of a red fluorescent protein (RFP), found in the non-bioluminescent coral (Discosoma striata) [5]. This was surprising as the convention was that fluorescent proteins came associated with bioluminescent systems as with Aequorea victoria. Named dsRed, this protein, unlike GFP, formed obligate tetramers which was detrimental to its expression and brightness. Through multiple directed mutations Tsien was able to monomerise dsRed [23] and go on to produce a spectrum of optimised RFP variants

[24] now widely used in research. Through continued research efforts the known

GFP-like protein family has rapidly expanded with fluorescent proteins (FPs) identified in over 80 species of coral, jellyfish and anemone. More recently fluorescent proteins have been identified in higher organisms including marine arthropods [25] and amphioxus [26]. However the majority of commercially used fluorescent proteins are optimised versions of GFP and dsRed that were originally discovered (see table 1.1). Continued efforts to find new FPs will undoubtedly turn up novel properties and increase the scope for further engineering and additional applications.

Protein Excitation Emission Molar Quantum in vivo Relative (Acronym) Maximum Maximum Extinction Yield Structure Brightness (nm) (nm) Coefficient (% of EGFP) Blue Fluorescent Proteins EBFP 383 445 29,000 0.31 Monomer* 27 EBFP2 383 448 32,000 0.56 Monomer* 53 Azurite 384 450 26,200 0.55 Monomer* 43 mTagBFP 399 456 52,000 0.63 Monomer 98 Cyan Fluorescent Proteins ECFP 439 476 32,500 0.4 Monomer* 39 mECFP 433 475 32,500 0.4 Monomer 39 Cerulean 433 475 43,000 0.62 Monomer* 79 CyPet 435 477 35,000 0.51 Monomer* 53 AmCyan1 458 489 44,000 0.24 Tetramer 31 Midori-Ishi Cyan 472 495 27,300 0.9 Dimer 73 TagCFP 458 480 37,000 0.57 Monomer 63 mTFP1 (Teal) 462 492 64,000 0.85 Monomer 162 Green Fluorescent Proteins GFP (w t) 395/475 509 21,000 0.77 Monomer* 48 EGFP 488 509 58,000 0.6 Monomer* 100 Emerald 487 509 57,500 0.68 Monomer* 116 Superfolder GFP 485 510 83,300 0.65 Monomer* 160 Azami Green 492 505 55,000 0.74 Monomer 121 mWasabi 493 509 70,000 0.8 Monomer 167 TagGFP 482 505 58,200 0.59 Monomer* 110 TurboGFP 482 502 70,000 0.53 Dimer 102 AcGFP 480 505 50,000 0.55 Monomer* 82 ZsGreen 493 505 43,000 0.91 Tetramer 117 T-Sapphire 399 511 44,000 0.6 Monomer* 79 Yellow Fluorescent Proteins EYFP 514 527 83,400 0.61 Monomer* 151 Topaz 514 527 94,500 0.6 Monomer* 169 Venus 515 528 92,200 0.57 Monomer* 156 mCitrine 516 529 77,000 0.76 Monomer 174 YPet 517 530 104,000 0.77 Monomer* 238 TagYFP 508 524 64,000 0.6 Monomer 118 PhiYFP 525 537 124,000 0.39 Monomer* 144 ZsYellow 1 529 539 20,200 0.42 Tetramer 25 mBanana 540 553 6,000 0.7 Monomer 13 Orange Fluorescent Proteins Kusabira Orange 548 559 51,600 0.6 Monomer 92 Kusabira Orange2 551 565 63,800 0.62 Monomer 118 mOrange 548 562 71,000 0.69 Monomer 146 mOrange2 549 565 58,000 0.6 Monomer 104 dTomato 554 581 69,000 0.69 Dimer 142 dTomato-Tandem 554 581 138,000 0.69 Monomer 283 TagRFP 555 584 100,000 0.48 Monomer 142 TagRFP-T 555 584 81,000 0.41 Monomer 99 DsRed 558 583 75,000 0.79 Tetramer 176 DsRed2 563 582 43,800 0.55 Tetramer 72 DsRed-Express (T1) 555 584 38,000 0.51 Tetramer 58 DsRed-Monomer 556 586 35,000 0.1 Monomer 10 mTangerine 568 585 38,000 0.3 Monomer 34 Red Fluorescent Proteins mRuby 558 605 112,000 0.35 Monomer 117 mApple 568 592 75,000 0.49 Monomer 109 mStrawberry 574 596 90,000 0.29 Monomer 78 AsRed2 576 592 56,200 0.05 Tetramer 8 mRFP1 584 607 50,000 0.25 Monomer 37 JRed 584 610 44,000 0.2 Dimer 26 mCherry 587 610 72,000 0.22 Monomer 47 HcRed1 588 618 20,000 0.015 Dimer 1 mRaspberry 598 625 86,000 0.15 Monomer 38 dKeima-Tandem 440 620 28,800 0.24 Monomer 21 HcRed-Tandem 590 637 160,000 0.04 Monomer 19 mPlum 590 649 41,000 0.1 Monomer 12 AQ143 595 655 90,000 0.04 Tetramer 11 * Weak Dimer

Table 1.1 The range of commercially available fluorescent proteins. Relative brightness is recorded as the percentage of EGFP. Adapted from Nikon Microscopy [27].

1.1.1. The Structure of Green Fluorescent Protein.

The GFP-like FP (fluorescent protein) family is an exquisite example of the relationship between protein structure and function. The structure of GFP was first determined in 1996 by X-ray crystallography [20, 21] and there are now over 600 structures in the Protein Data Bank (PDB) from 25 different species. GFP possesses a single domain of 239 amino acids (27kDa, Figure 1.2) arranged in a β-barrel.

Identification of the structure has been critical in revealing how the chromophore is formed and how the protein maintains its function. The structure is unique to the

GFP-like family in many ways. All members follow the 11 stranded anti-parallel β- sheet barrel or “can” conformation with a central α-helix running through the core.

See figure 1.6.

Many membrane proteins adopt β-barrel structures, such as membrane spanning porins and ion channels, controlling the passage of metabolites through the core.

However GFP-like proteins are unique in their possession of this central α-helix. It is the core helix where the chromophore is located, encased by the tight packing of the surrounding beta sheets and barrel terminal loops. The shielding serves to protect the chromophore from quenching by interfering water molecules, and is held rigidly in place by neighbouring side chains to prevent fluorescence energy being lost to vibration [20, 21]. The tightly packed structure gives rise to remarkable biophysical stability with complete loss of fluorescence only seen in strong denaturing conditions of 6M Guanidine HCl or less than pH 4 or greater than pH 12 [28].

(a)

(b)

Figure 1.6 Primary amino acid sequence and structure of Aequorea victoria Green Fluorescent Protein. (a) Primary amino acid sequence of EGFP, 238 amino acids. The ‘TYG’ chromophore forming sequence is highlighted. (b) The three dimensional structure of GFP consists of an 11-stranded anti-parallel β-sheet barrel conformation. α-helices are coloured yellow with β-sheet and loop regions coloured green. An α-helix is positioned in the barrel core containing the chromophore. Approximate dimensions shown in nm. Adapted from Day and Davidson [29].

The Chromophore of Green Fluorescent Protein.

The chromophore structure was originally identified from the proteolysis of GFP, giving rise to a hexapeptide fragment that has a similar absorption spectrum to the parent protein [11]. The chromophore itself is formed by post translational modification of a key tri-peptide sequence; Serine 65, Tyrosine 66, Glycine 67 [30] that forms a 4-p-hydroxybezylidene-imidazlinone ring structure.

The chromophore is formed in a two stage reaction, see figure 1.7. Firstly, during protein folding the peptide backbone residues are brought into close proximity, followed by the nucleophilic attack of the glycine amide nitrogen on the carbonyl carbon of the serine [14, 31]. Dehydration of the formed structure then traps the structure in the cyclised state. The second stage is the oxidation of the tyrosine side chain producing the fluorescent product, which is rate limiting, taking around 4 hours and is referred to as the maturation step. The critical peptide sequence Ser,

Tyr, Gly is of course not unique to GFP and occurs many times elsewhere in other proteins. However, it is the cyclisation of the residues that is unique to GFP, catalysed by its specific structural architecture. Residues Ser65 and Gly67 are brought together in close proximity by a structural distortion in the core α-helix, which possesses an 80° kink, or “tight-turn”. Structural torsion in the helix is maintained by a sequence of proline residues (Pro; 54,56,58,75 and 89) that border the helix [32]. Mutations at these residues have been shown to prevent the formation of the chromophore as the helix adopts a more native relaxed structure and no cyclisation event can occur. Residues that contact the chromophore within the core have also been shown to be crucial in its formation, see figure 1.8. Two residues in particular are highly conserved within the GFP-like family, Arginine 96 and

Glutamate 222 [31].

Figure 1.7 Proposed scheme for the formation of the GFP chromophore. The upper two forms of GFP are non-fluorescent. Oxidation of Tyr66 is required to form the two fluorescent states. The neutral (protonated) form is excited at 395 nm (lower right) while the anionic form is excited at 475 nm (lower left). Adapted from Heim et al [14].

Figure 1.8 Schematic diagram of the interactions between the chromophore and its surroundings in EGFP. Important side chain contacts Arginine 96, and Glutamine 222 are highlighted. Hydrogen bonds are drawn as dashed lines. Adapted from Zimmer [7].

Arginine 96 provides electrostatic interaction to the chromophore that stabilises chemical intermediates during the cyclisation/dehydration/oxidation reaction.

Glutamine 222 forms a network of hydrogen bonds with surrounding water molecules and along with Arg 96 serves to stabilise the chromophore in position once formed. In this respect, GFP self-catalyses formation of the chromopore with residues that stabilise and protect the formed moiety after synthesis [31]. Apart from molecular oxygen and water molecules no other substrates or co factors are required for the reaction. The understanding of the GFP structure and how the chromophore forms has enabled the development of new variants for wider technical applications

[33]. Research has focused on directed mutagenesis of residues that form and contact the chromophore to change the spectral characteristics and speed maturation time along with other regions to improve folding times, photostability and reduce oligomerisation [31]. GFP has been heavily engineered and these mutations are summarised in figure 1.9.

Figure 1.9. Aequorea GFP mutation map. A structural map showing the numerous optimising mutations superimposed on to a layout of the GFP polypeptide. β-sheets are numbered and depicted as thin, green cylinders with an arrow pointing towards the C-terminus, whereas α-helices are depicted as grey cylinders. Mutations are colour-coded to represent the variants to which they apply: BFPs (blue), CFPs (cyan), GFPs (green), YFPs (yellow), Sapphire (violet), folding, shared and monomerizing (grey). Most of the mutations are located in the central helix and in β-sheet strands 7, 8 and 10. In general, wavelength- specific mutations occur near the central helix containing the chromophore, whereas folding mutations occur throughout the sequence. Adapted from Shaner et al [33].

Design of spectral variants based on GFP has yielded blue, cyan and yellow fluorescing species but protein engineering efforts to push into the red, longer wavelength end (600-700nm) of the spectrum appears to be beyond GFP’s scope

[31]. However the discovery of red fluorescent protein, dsRed, brought the opportunity to develop such tools. Although dsRed posseses the overall characteristic β-can structure that closely matches GFP [34] it is crucial differences in the amino acid sequence that create a chromophore with a red spectral shift.

The dsRed chromophore forming sequence is composed of Gln, Tyr, Gly in comparison to the Ser, Tyr, Gly equivalent in GFP (Figure 1.10.C). During dsRed chromophore formation, the structure undergoes the same cyclisation>dehydration>oxidation reaction as GFP, forming the imidazoline ring with a green fluorescing intermediate. A second oxidation event of the alpha-carbon and amide nitrogen bond of Gln extends the π-bonding electron system to further include the carboxyl group of neighbouring Phe65 which shifts the emission spectra into the red phase [35].

Because of this extended reaction profile the maturation process takes more time, days rather than hours compared wiht the GFP counterpart [36]. Along with this extended maturation time dsRed was found to be significantly dimmer than GFP and form an obligate tetramer [34]. Oligomerisation is undesirable for genetic tag applications as it can obstruct protein localisation and hinder solubility. A concerted effort was made to produce a monomeric form and through a total of 33 mutations at the β-barrel interface Tsien’s team succeeded [23]. This mRFP (monomeric red fluorescent protein) then formed the basis for further effective red shifted spectral variants.

Through further analysis of other anthozoa species it appears that nature has adopted several other chromophore structures (figure 1.10) which give rise to a variety of spectral properties. All are based on the “XYG” sequence, where Tyr and Gly are conserved and the first position “X” is taken by several other residues which alters the conjugated π-electron system and spectral properties [37].

Photoactivatable and photoswitchable variants exist with particularly useful properties for fluorescence microscopy [7]. Fluorescence can be modulated as the hydroxybenzilidine (tyrosine) ring flips from the inactive trans position to the fluorescent active cis position upon excitation at a specific wavelength (Figure 1.10,

D and E). Chromophores have also been discovered where the trans isomer persists

[38], lacking fluorescence properties but creating coloured pigments (See figure 1.10

G.).

Figure 1.10 Chromophore diversity generated in GFP-like proteins. Several chromophopre types have been identified from jellyfish and coral species giving rise to a range of spectral properties which depends on the amino acid sequence of the chromophore and surrounding side chain contacts. (A) GFP type. (B) Yellow from Zoanthus (button coral). (C) Red from Anthozoa (D). Red kaede photoactivatable. (E) Red photoactivatable Kindling Flourescent Protein (KFP). (F). mOrange, derived from dsRed. (G) Non Fluorescent coloured pigments. Adapted from Zimmer [7].

1.1.2. Applications of Fluorescent Protein Technology

The range of applications for fluorescent proteins in research has rapidly evolved since the first use of GFP as a reporter of gene expression in E.coli [4]. Useful features of the intrinsic catalysis of the chromophore, flexibility in expression systems and biophysical stability make them remarkably well suited tools. A selection of the extensive number of emerging uses are summarised here:

1.1.2.1. Gene expression reporters.

The fluorescent protein gene can be inserted under the control of a promoter of interest to quantitate gene expression [4]. Fluoresence can also be used in a qualitative manner for checking transgene uptake, routinely implemented in fluorescent cell sorting for stem cell and gene therapy [39]. Livet et al used the mosaic like co-expression of several different coloured FPs to delineate neurons in the mouse brain, penning the term “brain-bow” for the striking fluoresence microscopy images produced [40].

1.1.2.2. Protein Tagging.

Fluorescent proteins can be co-expressed with genes of interest to produce a fluorescent fusion product with minimal disruption to native function. Fusions are formed by tagging to the N or C terminus, or even within the target protein sequence

[41] . The list of fusions that have been formed is extensive and this has become a standard technique for cell biology. Technology is now emerging which enables fluorescent protein detection down to a single molecule within a cell [42].

1.1.2.3. Biosensors

Genetically encoded probes have been engineered that use FPs to detect a wide range cellular events such as protein interactions, changes in metabolite concentration, pH and proteolysis. Using the phenomena of fluorescence resonance energy transfer

(FRET), biosensors can be designed which switch emission behaviour depending on the proximity of two fluorescent protein tags [43]. The varied types of FRET biosensor are summarised in figure 1.11.

1.1.2.4. Protein Folding.

Because of GFP’s easily quantified fluorescence it can be employed to measure protein folding or chaperone activity in vitro [44]. GFP is fused to a target protein and the progression of protein folding can be analysed by fluorescence spectrophotometry as the fluorescent chromophore is formed.

1.1.2.5. Whole organism imaging.

Transgenic cells expressing FP can be imaged non-invasively within live animals, with fluorescence emission detected through tissue layers [45]. Hoffman et al are pioneering the use of GFP labelled cell lines to monitor tumour progression in vivo

[46, 47]. While GFP is extremely useful for imaging cells in vitro and sub surface imaging in vivo there is currently a major drive to engineer long wavelength far-red

FPs, emitting ≥ 700nm. At these wavelengths light scattering effects and autofluoresence of cells are minimised to allow deeper transmission through tissues

[48].

Figure 1.11 Examples of Fluorescent Protein FRET Biosensors. Cyan, CFPs and Yellow, YFPs shown. (a) Molecular interaction biosensor. Association of the sensory domains switches on a fluorescence resonsonance energy transfer (FRET) signal. (b) Ligand biosensor. When the ligand binds, a conformational change in the sensory domain brings the fluorescent proteins into close proximity for FRET. (c) Molecular interaction biosensor where the substrate and sensory protein domains are linked within one construct in contrast to separate elements in (a). (d) A fluorescent protein biosensor that detects substrate cleavage operates through the elimination of FRET. Adapted from Gines and Davidson [49].

1.2. Immunogenicity of GFP.

One of the key benefits of GFP-like proteins is their low intracellular cytotoxicity demonstrated in many types of cell from many different species [50]. In the context of fluorescence microscopy, standard artificial fluorescent probes generally show phototoxicity on prolonged excitation as reactive oxygen species are created within the cellular environment [51]. However, in GFP the chromophore is kept within the protective protein scaffold. Although there are some reports of GFP related cytotoxicity at high expression levels [52], it is generally accepted that transgenic

GFP-like proteins are well tolerated by most cell types.

Increasingly FPs are being used in the whole organism setting to get a macro scale view of biological function. FPs are widely used as markers for monitoring tumour metastasis [45, 47], gene therapy [39] and non-invasive imaging techniques [46].

Fully transgenic (Tg) animals that ubiquitously express FPs including zebrafish [53], mouse [54], pig [55], and primate [56, 57] have all been successfully reared for research studies with no evidence of detrimental side-effects. However when FPs are to be used in tissue grafts or virus vectors, expression of FPs by transgenic cell implants in a wildtype host animal can be problematic.

GFP when expressed, a foreign jellyfish protein, has been shown to elicit an immune response in several model animal systems; mice, rats, primates in vivo [58-

67] and also human cells in vitro [68]. In some of these cases this has been exploited to recruit immune activity to GFP expressing tumour cell grafts for vaccine development [63, 68, 69]. For in-vivo applications of FPs, interference by the immune system may affect not only accuracy of results but may, more importantly

43 lead to complete transgenic tissue rejection and generation of immunological memory prohibiting the continued use of GFP as a marker in the animal model [60].

To address the problem several strategies to improve tolerance to GFP expressing cells and transgenic proteins as a whole have been employed. One logical direction is to use transgenic animals that have the transgene integrated into the genome and therefore the transgene product is seen as ‘self’. As previously mentioned several

GFP+ Tg species have been successfully raised. For transplantation studies Steitz et al [66] generated GFP+ BALB/c mice which showed tolerance to GFP+ adenovirus transduced tissue, which was otherwise rejected in the wild type strain. Remy [65] et al generated a GFP+ rat strain which also tolerated GFP transplanted cells and adenovirus delivery of GFP. However transgenic expression results in the recipient’s tissues ubiquitously expressing GFP and presumably the distinction between donor

GFP cells against background fluorescence is lost.

Immunosuppressive drugs, such as cyclosporine have been successfully used to encourage GFP+ tissue transplant acceptance in rabbits [61], dogs [59] and primates

[67]. Prolonged use of immunosuppressant, however, is frequently not appropriate especially in studies where an uncompromised immune response is required.

Adoptive transplantation of bone marrow cells transduced to express GFP has been shown to induce tolerance to the transgene [58] [70] but requires complete prior myeoblative conditioning. Annoi et al [71] demonstrated adoptive transfer of GFP+ antigen presenting cells (APCs) into C57BL/6 mice generated tolerance to CMV virus delivered GFP. Although these methods demonstrate advances in understanding tolerance, these are complex techniques that require a prolonged effort

44 to induce tolerance in a single animal and may not always be appropriate for a given experiment.

There is now also emerging evidence that the related dsRed protein is immunogenic in vivo in mice [72, 73] as well as the bioluminescent luciferase enzyme [74]. If FPs and other transgenic markers are to be increasingly used in animal studies then the issue of associated immunogenicity needs to be addressed to improve and further enable technology of this type.

1.2.1. Recognition of transgenic antigens by the immune system.

In the mammalian immune system the MHC class 1 (major histocompatibility complex) pathway is responsible for constitutively presenting the cell’s “self” protein repertoire and alerting the immune system to any intracellular “non-self” antigens such as viral and intracellular bacterial proteins [75]. Presentation of antigen begins with proteolysis of cytoplasmic proteins via the proteasome pathway.

Proteolytic peptide products are recruited onto the MHC1 complex at the endoplasmic reticulum and transported to the cell surface where they are presented to patrolling T-cells which determine whether the antigen is foreign or self. This process is summarised in figure 1.12.

Figure 1.12. Antigen presentation pathway for MHC1.

Endogenous proteins are degraded by the proteasome into peptides. These target proteins include “self” cytosolic as well as nuclear and foreign proteins, such as viral or transgenic products. The transporter for antigen processing (TAP) then translocates peptides into the lumen of the endoplasmic reticulum (ER). MHC class I heterodimers wait in the ER for complex with target peptides. Peptide binding is required for correct folding of MHC class I molecules and release from the ER and transport to the plasma membrane, where the peptide is presented to the immune system via the T-cell receptor (TCR) of CD8+ T-cells. Adapted from Yewdell et al; [76].

Because of the crucial role MHC class 1 takes in generating an immune response the mechanisms of antigenic peptide interaction have been extremely well studied.

Figure 1.13 shows the structure of the MHC class 1 in complex with an antigen peptide. MHC1 is a heterodimer of three immunoglobulin-like chains of MHC1 (α1-

α3) with β2-microglobulin. The antigenic peptide binds into a groove formed between the α1- α2 chains, often referred to as a “hot-dog in a bun” conformation

[75].

Experimental elution of antigenic peptide from isolated mammalian MHC class 1 complexes has shown that they preferentially binds 9 amino acid long, nonomer peptides and specific amino acids along the peptide are involved in the binding mechanism within the MHC pocket, termed anchor residues [77-79],

MHC loci are highly polymorphic with multiple alleles [75]. In humans the MHC class 1 or HLA (human leukocyte antigen) have well over 8000 alleles [80]. This ensures broad coverage of antigen binding within a population and demonstrates why tissue donor to patient inter-compatibility is a problem. However within an inbreeding population, such as an in-bred mouse strain the number of alleles is limited and therefore better defined. In mice the MHC class 1 is termed H-2, composed of three main allele types; K, D and L, see table 1.2. The haplotype, or set of alleles, is further designated by lowercase d,k or b, which vary between strains.

epitope peptide

α2 α2 α 1

α 1

(b)

β2 microgobulin α2 α 3

(a)

(c)

α 1

Figure 1.13. Depictions of MHC class I molecule bound to antigen peptide. (a): Side view of the MHC molecule domains α1-3 (pink, green and orange respectively), associated with β 2-microglobulin (blue) and bound 9-mer epitope peptide shown in red. (b). Top view; chain α1 and α2 chains form the peptide binding groove. (c): Space fill representation of peptide binding groove. Images rendered in Pymol. PDB: 1HSA, human class I MHC molecule HLA-B27 [81] .

H2 Allele

Class I Class II

Strain Appearance Haplotype K D L IA IE

Balb/c albino d Kd Dd Ld IAd IEd

C3H/He agouti k Kk Dk - IAk IEk

C57BL/6 black b Kb Db - IAb -

CBA agouti k Kk Dk - IAk IEk

Table 1.2. MHC alleles expressed by commonly used inbred mouse strains. In certain inbred mouse strains the MHC haplotype is collectively designated as H2 d, k or b which refers to the set of MHC alleles expressed (K,D,L, IA and IE). Dashes represent a null allele, with no gene product expressed at that locus. Adapted from Proimmune [82].

Due to the central function in immune recognition, a concerted effort has been made to understand the sequence specificity of epitope recognition by the MHC system.

Through sequencing isolated epitope peptides from MHC complexes [78, 83], sequence specificity and anchor site position has been determined with databases of known epitope peptides being created [84].

From these databases it has been possible to design algorithms that will predict probable epitopes within a known protein sequence for a given individual MHC allele. Initially this has been based on natively isolated epitopes [84] or in vitro binding experiments [85] and has now evolved to employing more complex programs, using artificial neural networks [86], training on large synthetic peptide libraries [87] and 3D structural analysis of MHC alleles [88]. In addition to looking at MHC binding, predictive methods for likelihood of proteasomal cleavage [89] and

TAP transport [90] prior to presentation have been designed. The immune epitope database [91] is a regularly updated central server that combines these algorithms to accurately predict MHC epitopes.

Predicting dominancy of MHC epitopes within a protein has proven extremely valuable in understanding the mechanisms of immune cell recognition.

These tools can be applied in designing and developing more effective vaccines where an improved immune response is desired [91]. Counter to this, knowledge of

MHC epitopes can be used to reduce epitope content of a protein and reduce potential immunogenicity. Indeed generation of an immune response to protein based biological therapeutics is of particular concern to the pharmaceutical industry

[92]. Uptake of therapeutic protein by APCs and presentation via MHC class II leads

50 to a humoral anti-drug antibody response that can neutralise efficacy, reduce half- life and create the risk of anaphylaxis on repeated dosing.

The industry is widely employing these predictive methods to screen protein drug candidates for potential MHC class II epitopes. Using rational protein engineering, predicted MHC class II epitope anchor sites can be mutated to reduce epitope content, reducing immunogenic potential whilst avoiding disruption to the function of the protein [85, 93-95], popularly termed ‘deimmunisation’.

1.2.2. Discovery of immunodominant epitopes in EGFP.

Following on from Stripecke et al [69] who demonstrated immunogenicity of EGFP in mice, Gambotto et al [63] and Han et al [64] identified EGFP’s immunodominant

MHC class I epitopes in BALB/c and C57BL/6 mice respectively. Using predictive methods [85] both groups predicted key MHC class I epitopes in the EGFP sequence. Epitopes restricted to other MHC alleles in the mice strains (H2-Dd, H2-

Kb) were not considered by these authors due to the very low MHC binding potential predicted. Synthesized peptides of the predicted epitopes were proven to be immunogenic in in vitro T cell stimulation assays using EGFP sensitised mice. In

BALB/c a (H2-Kd) the dominant epitope was determined as HYLSTQSAL (amino acid 199-207) and in C57BL/6 (H2-Db) DTLVNRIEL (117-125).

Interestingly, if follows that mutation of MHC class I anchor site residues can reduce cytotoxic immunogenicity towards foreign proteins, a mechanism employed in nature by parasites and viruses to evade the immune system. Binding of the peptide to the MHC class 1 molecule is the pre-requisite step in producing an

51 immune response to intracellular antigen. If this can process can be disrupted then it would reduce the likelihood of an immunogenic response being generated.

Dobano et al [96] have shown that mutation of the malarial parasite associated H2-

Kd dominant epitope anchor site (underlined: SYVPSAEQIL) is sufficient to remove protective immunity. Similarly Marques et al [97] demonstrated mutation of Herpes virus H2-Kd dominant epitope anchor site (underlined: GFNKLRSTL) aided immune evasion and increased viral load in BALB/c mice.

With prior knowledge of the immunodominant epitopes in EGFP it is plausible that one could employ the same strategies here to disrupt immunodominant epitopes indicated in model mouse strains, whilst maintaining fluorescence function. Whereas previous mutations of GFP have been employed to improve spectral characteristics or biophysical properties [33], no group has yet used protein engineering to change the immunological profile.

A GFP with reduced immunogenicity would be a valuable tool for research. The work here presents a model to validate the plausibility of engineering out immunogenicity in useful biotechnology reagents, with the distinct advantage of being able to readily detect its fluorescence property. A next generation of FPs is envisaged that combine the unique fluorescent qualities with improved safety and efficacy profile for enabling and expanding in vivo applications.

1.3. Hypothesis, Aims and Experimental Approach.

1.3.1. Hypothesis

Immunogenicity of EGFP can be reduced in BALB/c mice by mutation of anchor sites in the immunodominant MHC class I epitope and maintaining fluorescence functionality.

1.3.2. Aims

The overall aim of this thesis was to develop a functional EGFP variant with reduced immunogenicity in BALB/c mice by site directed mutagenesis.

1.3.3. Experimental approach

Initially the immunodominant epitopes characterised for EGFP in BALB/c and

C57BL/6 mice were assessed [63, 64]. MHC class I anchor sites were determined by epitope database analysis and structural homology modelling.

A recombinant expression and purification system for EGFP was established to provide sufficiently pure EGFP protein for characterisation and immunogenicity studies. EGFP was expressed in E.coli with an N-terminal Histidine tag and soluble protein was isolated via IMAC (immobilised metal affinity chromatography) purification.

Saturation site directed mutagenesis PCR was used to mutate appropriate anchor sites and generate mutation libraries. The libraries were then screened for fluorescent variants which carried epitope destabilising anchor site mutations.

These variants were then recombinantly expressed and fully characterised to ensure comparable florescence properties to parent EGFP protein and a lead candidate was selected.

The final stage was to establish a BALB/c mouse model to demonstrate EGFP immunogenicity. Several types of immunogen were trialled using; recombinant protein, syngenic bone marrow dendritic cells (BMDC) transduced with EGFP lentivirus and a syngenic tumour cell line expressing EGFP. The immunogenic response between the parent EGFP and ‘deimmunised’ variant was compared through a series of in vitro assays;

1. Serum antibody ELISA to EGFP.

2. Cytotoxicity marker ELISA.

3. Cytotoxic T cell lymphocyte assay (CTL)

Chapter 2

2. Rationale for deimmunisation of EGFP.

The accurate prediction of both MHC class I and class II epitopes of proteins is becoming a valuable tool in understanding immunology of allergy, cancer and infectious disease as well as improving vaccine design. In addition, the prediction of

MHC epitopes can be exploited to reduce unwanted immunogenicity for proteins to be used in vivo. Computational tools to predict possible epitopes are trained on databases of known epitopes for each MHC allele which share conserved sequence motifs [83]. These conserved sequences within the motifs are the key amino acids that drive the interaction with the MHC, termed anchor residues.

Identification of MHC class II epitopes in therapeutic proteins is of particular importance to prevent immunogenicity towards the protein product. For a given

MHC class II allele, anchor site residues, within the epitope sequence can be identified and mutated using protein engineering to reduce the likelihood of epitope presentation. However these mutations must not detrimentally affect the function of the protein or produce novel epitopes.

The same rationale can be applied to MHC class I epitopes anchor site for intracellularly expressed proteins, such as EGFP. Gambotto et al [63] and Han et al

[64] have experimentally determined the immunodominant epitopes in EGFP for

BALB/c (H2-Kd) and C57BL/6 (H2-Db) strain mice respectively. In this chapter the identification of the relevant anchor sites and a rationale for mutation to reduce epitope immunogenicity in EGFP is described.

2.1. Aims.

 Prediction of MHC class I epitopes of EGFP.

 Predict the effect of epitope anchor site mutation.

 Homology modelling of EGFP epitope peptide-MHC binding.

 Identification of EGFP epitope anchor sites in parent protein structure.

2.2. Methods

2.2.1. Prediction of immunodominant epitopes in EGFP.

The amino acid sequence for EGFP was screened for MHC class I epitopes by using the Immune Epitope Database analysis resource tool (IEDB, NIH, La Jolla,

California, USA [91]). The sequence (Figure 1.6a) was submitted to the Immune

Epitope Database: MHC class 1 T cell epitope prediction tool. MHC alleles were selected for either BALB/c (H2-Kd,Dd and Ld) or C57Bl/6 (H2-Kb and Db). The search was confined to epitopes of 9 amino acids in length, the most frequent size of an epitope. The ANN (artificial neural network) algorithm [86] was used to determine epitope position and predict the affinity of the epitope to the MHC complex as half maximal inhibitory concentration (IC50).

2.2.2. Anchor site alanine scanning.

Each amino acid residue in the epitope sequences HYLSTQSAL (199-207) and

DTLVNRIEL (117-125) was substituted by alanine at each position (1-9) and submitted into the IEDB: MHC class 1 T cell epitope prediction tool (as for 2.2.1) in order to determine the effect of each side chain on binding affinity. MHC alleles were selected as either H2-Kd for HYLSTQSAL or H2-Db for DTLVNRIEL. The

56 search was confined to epitopes of 9 amino acids. The ANN (artificial neural network) algorithm [86] was used to predict the affinity of the alanine subsitutued peptide to the MHC complex (IC50).

2.2.3. Homology modelling of EGFP epitope/MHC complex.

To visualise the theoretical position of the EGFP immunodominant peptides in complex with the MHC, peptides were modelled onto existing 3D structures by homology modelling. Swiss PDB Viewer (Swiss Institute of Bioinformatics) was used to generate peptide structures for the EGFP peptide sequences (HYLSTQSAL or DTLVNRIEL). These were threaded onto the carbon backbone of known epitope structures. The H2-Kd EGFP epitope sequence HYLSTQSAL (199-207) was threaded onto known influenza nucleoprotein antigen peptide TYQRTRALV (147-

155) in complex with H2-Kd MHC molecule (PDB accession: 2FWO [79]). The H2-

Db EGFP peptide sequence DTLVNRIEL (117-125) was threaded onto known

Sendai virus nucleoprotein antigen peptide FAPSNYPAL (324-322) in complex with

H2-Db MHC molecule (PDB accession: 1QLF [98]). Polar contacts of the dominant anchor site residues Y200 (in HYLSTQSAL) and N121 (in DTLVNRIEL) with the

MHC were calculated using Pymol (Schrodinger LLC, Portland, Oregon, USA).

2.2.4. Identification of epitope anchor site residues in EGFP structure.

Structural modelling was used to determine the position of EGFP epitope anchor sites in the parent EGFP structure and assess their suitability for mutation. The X- ray crystal structure of EGFP (PDB accession 2Y0G [99] ) was modelled using

Pymol and the anchor site residues were identified.

2.3. Results

2.3.1. Confirmation of immunodominant epitopes in EGFP.

2.3.1.1. BALB/c H2-k

The EGFP amino acid sequence was submitted to the IEDB, resulting in 230 results of 9-mer peptides. Table 2.1 shows the 5 highest predicted affinity epitopes in

BALB/c mice for each H2d allele using the ANN method. Peptides with an predicted

IC50 value of ≤ 500nM are considered to be potentially immunogenic, whereas

≥500nM are unlikely to be immunogenic [100].

For H2-Kd the HYLSTQSAL (199-207) peptide demonstrated highest predicted binding with a relatively high affinity of 6nM.

Strong predicted interactions that had IC50 values ≤ 500nM were shown for

NYNSHNVYI (144-152) with 260nM and KFICTTGKL (45-53) with 330nM indicated for H2-Kd. No predicted significant affinity (≤ 5000nM) was shown for any peptide indicated in H2-Dd.

Only one peptide TPIGDGPVL (186-194) was indicated for H2-Ld that had intermediate affinity with an IC50 value of 700nM.

2.3.1.2. C57BL/6 H2-d

Table 2.2 shows the 5 highest predicted affinity epitopes in C57BL/6 mice for each

H2-k allele using the ANN method. For H2-Db DTLVNRIEL (117-125) peptide demonstrated highest predicted binding with an intermediate affinity of 730nM. No other epitopes for either H2-Db or H2-Kb demonstrated significant predicted interactions that had IC50 values ≤ 500nM.

Allele Peptide Position IC50 (nM) H2-Kd HYLSTQSAL 199-207 6 NYNSHNVYI 144-152 260 KFICTTGKL 45-53 330 GYVQERTIF 91-99 1530 DHYQQNTPI 180-188 4390

H2-Dd DGDVNGHKF 19-27 6790 DGPVLLPDN 190-198 7160 PWPTLVTTL 56-64 10050 KLPVPWPTL 52-60 12800 TLTYGVQCF 63-71 24400

H2-Ld TPIGDGPVL 186-194 700 VNFKIRHNI 163-171 2490 DPNEKRDHM 210-218 7480 VLLPDNHYL 193-201 2150 YVQERTIFF 92-100 15400

Table 2.1 Predicted MHC class I epitopes of EGFP in BALB/C mice The EGFP amino acid sequence was submitted to the Immune Epitope Database and screened for MHC class I, 9 amino acid epitopes, using the ANN method. For each H2d allele, the five strongest predicted binding epitopes are shown. Positions within the parent EGFP protein are included. IC50 (nM) represents the predicted affinity value for peptide/MHC complex. Values ≤50nM are considered strong interactions, ≤500nM are considered intermediate, ≤5000nM are considered weak and ≥5000nM are considered negligible. Epitopes are listed in decreasing order of predicted MHC binding.

Allele Peptide Position IC50 (nM) H2-Db DTLVNRIEL 117-125 730 YNSHNVYIM 145-153 2340 VLLPDNHYL 193-201 10600 ELFTGVVPI 6-14 13800 DHMVLLEFV 216-224 14500

H2-Kb VNFKIRHNI 163-171 1550 VNRIELKGI 120-128 2940 GNYKTRAEV 104-112 3350 VNGHKFSVS 22-30 4010 YNYNSHNVY 143-151 5490

Table 2.2 Predicted MHC class I epitopes of EGFP in C57BL/6 mice The EGFP amino acid sequence was submitted to the Immune Epitope Database and screened for MHC class I, 9 amino acid epitopes, using the ANN method. For each H2b allele, the five strongest predicted binding epitopes are shown. Positions within the parent EGFP protein are included. IC50 (nM) represents the predicted affinity value for peptide/MHC complex. Values ≤50nM are considered strong interactions, ≤500nM are considered intermediate, ≤5000nM are considered weak and ≥5000nM are considered negligible. Epitopes are listed in decreasing order of predicted MHC binding.

2.3.2. Anchor site alanine scanning.

The highest affinity epitopes predicted previously in 2.3.1 that were also experimentally proven by Gambotto et al [63] and Han et al [64] were subject to alanine substitution at each position in the 9-mer sequence and the output was submitted to the IEDB to predict the destabilising effect of a positional mutation on

MHC binding affinity. Alanine was selected as a suitable destabilising mutation due to the minimum side chain interaction with the MHC.

2.3.2.1. HYLSTQSAL

Table 2.3 (a) shows that only substitution of the Tyr200 at position 2 in the 9-mer sequence had the most significant effect on predicted binding affinity. Affinity changes significantly from a strong affinity of 6nM in the wild type sequence to a low affinity of 1020nM (>500nM). Substitution at other positions only resulted in minor changes in affinity from the wild type sequence (≤30nM).

2.3.2.2. DTLVNRIEL

Table 2.3 (b) shows that substitution of the Asn121 at position 5 in the 9-mer sequence had the largest effect on predicted binding affinity. The binding affinity changes significantly from intermediate strength at 730nm in the wildtype sequence to a negligible affinity of 25mM (>5mM) [86]. Additionally substitution at other positions P3 (Leu119), P4 (Val120) and C-terminal P9 (Leu125) all had a destabilising effect which resulted in decreased affinity <7000nM. In contrast to this, alanine substitution at positions P1 (Asp117) and P2 (Thr118), increased affinity significantly to 10nM and 40nM respectively. Substitution at positions P6, P7, P8 resulted in minor changes in affinity from the wild type sequence and were all

<900nM.

(a) Position Sequence IC (nM) Substituted 50 Wt HYLSTQSAL 6 1 AYLSTQSAL 6 2 HALSTQSAL 1020 3 HYASTQSAL 6 4 HYLATQSAL 6 5 HYLSAQSAL 7 6 HYLSTASAL 6 7 HYLSTQAAL 13 8 HYLSTQSAL 6 9 HYLSTQSAA 30

(b) Position Sequence IC (nM) Substituted 50 Wt DTLVNRIEL 730 1 ATLVNRIEL 10 2 DALVNRIEL 40 3 DTAVNRIEL 6300 4 DTLANRIEL 1800 5 DTLVARIEL 25800 6 DTLVNAIEL 430 7 DTLVNRAEL 730 8 DTLVNRIAL 840 9 DTLVNRIEA 6700

Table 2.3 Impact of alanine scanning on predicted immunogenicity of EGFP immunodominant epitopes. All amino acids in the immunodominant epitope of EGFP in BALB/c (table a) and C57BL/6 (table b) were sequentially substituted with alanine (highlighted as A). The nine sequences were submitted to the IEDB to determine the predicted binding affinity of each epitope using the ANN method. The wildtype (Wt) sequence is

shown for each epitope. IC50 (nM) represents the predicted affinity value for peptide/MHC complex. Values ≤50nM are considered strong interactions, ≤500nM are considered intermediate, ≤5000nM are considered weak and ≥5000nM are considered negligible.

2.3.3. Homology modelling of epitopes

To model the interaction of the EGFP immundominant epitope peptides with the respective MHC, structural models were generated by threading the peptide sequence onto the sequence of existing epitope structures in complex with MHC.

2.3.3.1. HYLSTQSAL (H2-Kd)

By superposition of HYLSTQSAL onto influenza nucleoprotein epitope

TYQRTRALV, it is evident that the two epitopes have common features with Y at position 2 and T at position 5 in the 9-mer sequence. Figure 2.1 shows a comparison of the two peptides docked in the MHC peptide binding domain. In the experimental structure defined by Mitaksov et al [79] (figure 2.1a) the Y148, T151 and V155 are buried into the MHC and act as anchor sites. In comparison, for the EGFP peptide,

Y200, T203 and L207 are acting as anchor sites. Figure 2.1b shows a predictive model of the anchor residue Y200 of HYLSTQSAL in complex with the MHC.

Y200 is tightly docked into a hydrophobic binding pocket of H2-Kd. Y200 makes edge-to-face pi stacking interactions with Y7, F45, and F99 and makes hydrophobic contacts with V9 and the aliphatic portion of R66 in the MHC. The hydroxyl group of Y200 hydrogen bonds directly to the oxygen of D70 and is in a suitable position to make electrostatic interaction with R97.

R152 R150 (a) A153 L154 Q149

T147 T151 V155

Y148

(b) Q204 S205 S202 L201 H199 A206

T203

Y200 L207

Figure 2.1 Predictive docking of EGFP epitope HYLSTQSAL with H2-Kd MHC by homology modelling. (a) H2-Kd epitope peptide for influenza nucleoprotein TYQRTRALV (grey) in complex with H2-Kd MHC molecule [79]. (b) The immunodominant EGFP epitope HYLSQTSAL (magenta) was threaded onto the carbon backbone of TYQRTRALV. Amino acids in the epitope peptides are labelled. Nitrogen groups are coloured blue and oxygen groups red. The MHC peptide binding domain is represented as a semi- transparent ribbon cartoon model, coloured grey. Image generated with Pymol.

Q204 R66 S202 L201

H199 D70 T203

F45 Y200 F99 R97

Figure 2.2. Model of the HYLSTQSAL Tyr200 anchor site interaction with the MHC H2-Kd binding pocket. The immunodominant EGFP epitope, HYLSQTSAL (magenta) was by docked onto the H2-Kd MHC peptide binding domain, represented as a ribbon (grey). Amino acid side chains of the MHC contacting the Y200 anchor site residue are shown (grey). Nitrogen groups are coloured blue and oxygen groups red. Putative hydrogen bonds with Y200 connecting atoms <3.5 Å apart are also shown as dashed yellow lines. The MHC peptide binding domain is represented as a ribbon, coloured grey. Image generated with Pymol.

2.3.3.2. DTLVNRIEL

The EGFP epitope DTLVNRIEL was threaded onto the known structure of the

Sendai virus nucleoprotein FAPSNYPAL epitope in complex with H2-Db. The two epitopes share common features with Asn at position 5 and Leu at position 9 in the

9-mer sequence. In figure 2.3a the structure of H2-Db determined by Glithero et al

[98] shows the side chains of Asn328 and Leu332 of the FAPSNYPAL epitope are deeply buried into the MHC as anchor site residues. Superposition of the EGFP epitope sequence in figure 2.3b shows equivalent side chains of Asn121 and Leu125 in the same position. In figure 2.4, the interaction of Asn121 with the MHC is modelled. The side chain of Asn121 makes electrostatic contacts with Gln97 and hydrophobic contacts with Trp73. Asn at this position has been shown to be the dominant anchor site residue in H2-Db epitopes [83].

(a) Y329 P330

S327 F324 A331

A325 N328 P326 L332

(b) R122 I123 V120 E124

D117

L125

L119 N121 T118

Figure 2.3 Predictive docking of EGFP epitope DTLVNRIEL with known H2- Db epitope by homology modelling. (a) The Sendai virus nucleoprotein 6 epitope FAPSNYPAL (grey) in complex with H2-Db MHC. (b) The immunodominant EGFP peptide DTLVNRIEL (magenta) was threaded onto the carbon backbone of FAPSNYPAL. Amino acids in the epitope peptides are labelled. Nitrogen groups are coloured blue and oxygen groups red. The MHC peptide binding domain is represented as a semi-transparent ribbon cartoon, coloured grey. Image generated with Pymol.

R122 V120 E124 I123

Q70 N121 W73 L119 L125 Q97

Figure 2.4. Model of the DTLVNRIEL Asn121 anchor site residue interaction with the MHC H2-Db binding pocket. The immunodominant EGFP epitope, DTLVNRIEL (magenta) was by docked onto the MHC H2-Db peptide binding domain, represented as a ribbon (grey). Amino acid side chains of the MHC contacting the N121 anchor site are shown (grey). Nitrogen groups are coloured blue and oxygen groups red. Putative hydrogen bonds between N121 and Q97 are shown as dashed yellow lines. The MHC peptide binding domain is represented as a ribbon, coloured grey. Image generated with Pymol.

2.3.4. Location of anchor site residues within EGFP.

To assess the possible effect of mutation of the anchor site residues, the positions of

Tyr200 and Asn121 were first identified within the known structure of EGFP [99].

Tyr200 is a surface residue and makes π-stacking contacts with neighbouring

Tyr151, figure 2.5. Tyr200 is a surface residue and there are no direct contacts with the chromphore. Asn121 is located in the core of the protein, in proximity to the α- helix where the chromophore is positioned, figure 2.6. Asn121 is packed within three valine groups, Val16, Val68 and Val122 and the guanidinium group forms a hydrogen bond the carbonyl group of Phe64 that precedes the chromophore sequence.

A227

Y200 S202 Y151

N149

Figure 2.5. Location of the dominant anchor site Tyr200 in EGFP. The Tyr200 dominant anchor site residue of the HYLSTQSAL epitope was identified in the EGFP protein structure, show in grey with nitrogen groups coloured blue and oxygen groups red. Neighbouring amino acid side chains (within 4 Å) of Tyr200 are also shown. The chromophore structure is coloured green and the whole protein structure is represented as a ribbon in grey (semi transparent). Image generated with Pymol.

V112 V68 N121

V16

F64

Figure 2.6. Location of the dominant anchor site Asn121 in EGFP. The Asn121 dominant anchor site residue of the DTLVNRIEL epitope was identified in the EGFP protein structure, show in grey with nitrogen groups coloured blue and oxygen groups red. Neighbouring amino acid side chains (within 4 Å) of Asn121 are also shown. The putative hydrogen bond between N121 and F64 is shown as dashed yellow lines. The chromophore structure is coloured green and the whole protein structure is represented as a ribbon in grey (semi transparent). Image generated with Pymol.

2.3.5. Summary

The sequence of EGFP was screened for MHC class I epitopes using the Immune

Epitope Database analysis resource. In BALB/c mice, three high affinity H2-Kd restricted epitopes were identified (IC50<500nM); HYLSTQSAL (199-207),

NYNSHNVYI (144-152) and KFICTTGKL (45-53). No high affinity epitopes were predicted for H2-Dd or H2Ld. Gambotto et al [63] used similar predictive methods

[85] with the same epitopes being identified. They showed experimentally both

HYLSTQSAL and KFICTTGKL bound H2-Kd in vitro whereas NYNSHNVYI did not stably bind the MHC complex. Furthermore HYLSTQSAL was shown to be the only epitope peptide that would induce IFNγ secretion and CTL killing in splenocytes from mice immunised with EGFP.

In C57bL/6 mice (H2-b), only one epitope sequence DTLVNRIEL (117-125) was predicted, with intermediate affinity (IC50 = 730nM). Han et al also predicted this epitope in EGFP [64]. They showed that this was able to stably bind H2-Db MHC in vitro and induce IFNγ secretion in splenocytes from mice immunised with EGFP. No epitopes were predicted for H2-Kb.

Based on the predictive methods used here and previous published work [63, 64]

HYLSTQSAL and DTLVNRIEL are suitable candidates as immundominant epitopes in BALB/c and C57BL/6 mice respectively and can therefore be considered as suitable targets for deimmunisation.

Both epitopes were subject to alanine screening to predict the effect of mutation on the known anchor sites. In HYLSTQSAL, mutation of Tyr200 to Ala resulted in a significant IC50 decrease from 6nM to 1020nM, indicating substitution at this residue could destabilise binding. A number of binding studies have shown that Tyr at this position is the dominant anchor site residue [101-103], where conservative mutations

(Y>F or W) decrease binding and non conservative mutations (Y>A or T) completely eliminate binding.

In DTLVNRIEL, mutation of Asn121 to Ala also decreased binding significantly from 730nM to 2.6mM, rendering a binding event improbable. Ideally the minimum number of mutations would be made to EGFP by targeting these anchor sites alone with the highest impact on reducing binding affinity and a minimum effect on parent protein structure and function.

These dominant anchor site positions, Tyr200 and Asn121, were then mapped onto the known structures of MHC-epitope complexes by homology modelling. Both anchor site residues drive structure specific interactions within the MHC binding cleft which demonstrates why they are highly conserved between known epitopes for each MHC allele [83].

If Tyr200 and Asn121 are to be selected as plausible targets for mutation to disrupt epitope binding then consideration must also be given to the effect on the structure of the parent protein EGFP. Both anchor sites were indentified in the known structure of EGFP [99]. Tyr200 is a surface residue with no major contacts to the chromophore region making it a good candidate for mutation, without disrupting

EGFP function. However Asn121 is a protein core residue and makes contacts with amino acid residues associated with the chromophore. Discovery of a suitable mutation at this point could be more challenging as it may affect protein folding, chromophore environment and fluorescence property.

Chapter 3

3. Mutagenesis of Enhanced Green Fluorescent Protein MHC Anchor Sites.

3.1. Methods

3.1.1. Construction of EGFP expression vector.

EGFP containing vector pGL3 was a kind gift from the laboratory of Dean Jackson

(University of Manchester). The EGFP gene was isolated by designing gene specific primers to amplify EGFP DNA by PCR and incorporate suitable restriction enzyme sites suitable for cloning into pET15b E.coli expression vector (Invitrogen). Primer sequences and vector map can be found in the appendices. To amplify the EGFP gene a three step PCR reaction was designed. The PCR reaction mix contained 70ng template DNA (pGL3), 500nM forward and reverse oligonucleotide primers, 200nM deoxynucleotide triphosphate mix, 2 units of proof reading Phusion polymerase

(NEB) in manufacturer’s high fidelity buffer with a total reaction volume of 50µL.

The PCR cycle consisted of denaturation at 98°C for 1 minute, 32 cycles of 98°C for

30 seconds, 70°C annealing for 30 seconds, 72°C extension for 30 seconds and a final 10 minute polishing step at 72°C and performed using a Techne TC-412 thermal cycler. PCR products were visualised by agarose gel electrophoresis. 5µL of reaction product were loaded onto a 1 % TBE agarose gel (200mM Tris base,

180mM Borate, 5mM EDTA, pH 8.3).and run for 30m at 100V in TBE buffer. An endonuclease restriction enzyme digest was performed on the EGFP amplification

PCR product and pET15b. The double digest reaction mix consisted of 20 units of

BamHI and 20 units NdeI (both from NEB), including 1µg of EGFP PCR product or

20µg of pET15b, 0.1mg/mL BSA with NEB buffer 4 (NEB) to a final volume of

50µL.The reaction was incubated at 37°C overnight. The digest products were separated by agarose gel electrophoresis and isolated using a Qiaquick gel extraction

74 kit (Qiagen). The concentration of purified DNA was then measured using a

Nanodrop ND-1000 spectrophotometer. The ligation reaction was performed using the Quick ligation kit (NEB) with EGFP and pET15b DNA in a 3:1 insert to vector molar ratio and 2000 units of T4 DNA ligase. The reaction was incubated at 25°C for 1 hour. The DNA was then purified from the reaction mix using a Qiaquick PCR purification kit (Qiagen) and concentration measured using a Nanodrop spectrophotometer.

3.1.2. E.coli Transformation and plasmid preparation.

The insertion of the EGFP gene into the vector was confirmed through DNA sequencing of the plasmid. 5ng of plasmid DNA was added to 20µL of DH5α chemically competent cells (Invitrogen) on ice and heat shock performed at 42°C for

30 seconds then returned to ice. 0.5mL of liquid SOC medium was added and 100µL of cell suspension was plated out onto selective LB agar with 100µg/mL ampicillin.

The plate was incubated overnight at 37°C. Several colonies were picked and each used to inoculate 10mL liquid LB plus 100µg/mL and cultured on a shaker incubator overnight at 37°C, 200rpm. The cell pellet was isolated by centrifugation at 200 x g and plasmid DNA was prepared using a Qiagen plasmid midi kit (Qiagen). Plasmid

DNA was sent to GATC Biotech Ltd (London, UK) for Sanger method DNA sequencing using the sequencing primer for the T7 promotor located on pET15b.

The DNA sequence for EGFP was confirmed by clustalW alignment with the known

EGFP DNA sequence (NCBI accession: AFA52654.1).

3.1.3. Transformation of expression cell strain BL21(DE3)

5ng pET15b/EGFP DNA was added to 20µL of chemically competent BL21(DE3)

(EMD Millipore) expression cells on ice and heat shock performed at 42°C for 30

75 seconds then returned to ice. 0.5mL of liquid SOC medium was added and 100µL of cell suspension plated out on to LB agar with 100µg/mL ampicillin. The plate was incubated overnight at 37°C. Glycerol stocks were prepared by picking single colonies which were used to inoculate 5mL of liquid LB broth containing 100µg/mL ampicillin and incubated overnight. This culture was then used to dilute a second liquid LB culture 50 fold. When the culture reached log growth phase (OD 600nm =

0.6) the culture was then cooled on ice and diluted 1 in 5 into sterile glycerol in cryogenic tubes snap frozen in liquid nitrogen and then stored at -80°C.

3.1.4. Site Directed Mutagenesis of MHC anchor sites.

MHC anchor sites Y200 and N121 were mutated using site directed mutagenesis

PCR of the parent construct. For saturation mutagenesis of the target amino acid codon the degenerate code ‘NNS’ was implemented where N is any nucleotide and S is G or C nucleotides, encompassing all possible amino acids. For additional directed mutagenesis of N121, primers were designed to incorporate codons for alanine and aspartate. Mutagenesis primers were designed using the QuickChange primer design tool (Aglient). Primer sequences can be found in the appendices.

The PCR reaction mix contained 25ng template pET15b/EGFP plasmid DNA,

100µM forward and reverse oligonucleotide primers, 200nM deoxynucleotide triphosphate mix and2 units of Phusion polymerase (NEB) in high fidelity buffer in a total reaction volume of 50µL. The PCR cycle was: denaturation at 98°C for

1minute, 25 cycles of 98°C for 30 seconds, 65°C primer annealing for 1 minute,

72°C extension for 10 minutes and a final 10 minute polishing step at 72°C. 5µL of

PCR product was run on a 1% agarose gel to confirm amplification. To remove methlyated template DNA, PCR product was incubated with Dpn1 endonuclease

(NEB). The enzyme reaction contained 45µL of PCR reaction product, 3µL of Dpn1 and 5µL of buffer 4 (NEB) and was incubated at 37°C overnight.

The digested PCR product was then transformed into competent DH5α cells using the heat shock method described in 2.1.2. The entire transformation cell mix was transferred to 10mL of sterile liquid LB with 100µg/mL ampicillin and incubated at

37°C overnight. The resulting library of plasmid containing mutations was then prepared from this culture using the QIAprep Miniprep kit. DNA concentration was confirmed by Nanodrop spectrophotometer analysis. 5ng of library plasmid DNA was then used to transform Bl21(DE3) Rosetta 2 (Merck Millipore) following the method described in 2.1.3. The Rosetta strain possesses a number of additional low usage codon tRNAs to enhance heterologous expression in E.Coli. 50µL of transformation mixture was spread onto a sterile nitrocellulose membrane layered onto LB agar with 100µg/mL ampicillin. The plate was incubated overnight at 37°C.

The next day the nitrocellulose membranes with growing colonies attached were transferred to new LB agar plates containing 0.1mM isopropyl β-D-1- thiogalactopyranoside (IPTG) to induce expression of EGFP. These were incubated for 4 hours at 37°C. The library of transformants was screened using a blue light transilluminator (Clare Chemical Research). Fluorescent colonies were picked and used to inoculate 5mL of sterile LB and plasmid DNA prepared using the QIAprep

Miniprep kit. DNA was sent to GATC Biotech Ltd (London, UK) for DNA sequencing as per method 2.1.2.

3.1.5. Pilot Recombinant Expression of EGFP in E.Coli.

To enable characterisation of EGFP constructs a protein expression and purification system was developed. A sample of glycerol stock of BL21(DE3)/pET15b/EGFP was streaked onto a LB agar with 100µg/mL ampicillin. The plate was incubated overnight at 37°C. A single colony was picked and added to starter culture of 5mL of sterile liquid LB broth containing 100ug/mL ampicillin on a shaker incubator at

200rpm, 37°C overnight. The next day 1ml of culture was added to 50mL sterile liquid LB media with 100µg/mL ampicillin. The culture was incubated on a shaker incubator at 37°C, 200 rpm. When the culture reached log growth phase (OD 600nm

= 0.6), IPTG (Formedium) was added to a final concentration of 1.0mM. The culture was then incubated at 18°C, 200 rpm overnight. To harvest the cells the culture was added to a 50mL centrifuge tube and centrifuged at 1000rpm for 10 minutes (Sorvall

Legend RT). The cell pellet was then resuspended in 5mL lysis buffer (50mM Tris,

150mM NaCl, 5mM Imidazole, 0.1mg/ml hen egg white lysozyme, 1U of benzonaze, pH7.0) on ice and incubated for 30 minutes. The lysate was then sonicated on ice using a sonicator probe (Vibra cell VCX130) 6 times at 50% amplitude for 5 second pulses with 30 second pauses inbetween pulses. The homogenate was then centrifuged at high speed (17000 x g, Beckman Avanti JE) for

30 minutes. The soluble fraction supernatant was aspirated and stored.

3.1.6. Pilot IMAC chromatography

EGFP with N-terminal histidine Tag was isolated from cell lysate by immobilised metal affinity chromatography (IMAC). A gravity flow purification column

(Thermo) was prepared by adding 1mL of NTA-agarose (Qiagen). This was equilibrated with 3 CV (column volumes) of binding buffer (50mM Tris, 150mM

NaCl, 5mM Imidazole, pH 8.0). The soluble protein fraction was then aspirated onto the column and allowed to flow through. The column was then washed with 5 CV of binding buffer. The protein was then eluted by adding 1 CV of IMAC elution buffer

(50mM Tris, 150mM NaCl, 1M imidazole, pH 8.0) at a time and eluted fractions were collected.

3.1.7. Analysis of chromatography fractions by SDS PAGE.

5µL (1-100µg of protein expected) of elution samples were diluted in LDS running buffer (Expedeon) with 100mM dithiothreitol (Sigma) and heated at 95°C for

5minutes. 5µL samples were loaded onto a 10% tris-glycine acrylamide gel.

Electrophoresis was run at 100V for 45 minutes. The gel was then stained using

Instantblue protein stain (Expedeon).

3.2. Scaled up expression and chromatography.

To enable increased yield amounts (mg) of protein to be generated for characterisation and immunisation studies the expression system was scaled up accordingly. As per protocol 2.1.5 a colony of BL21(DE3)/pET15b/EGFP was picked and added to starter culture of 50mL of liquid LB broth containing 100µg/mL ampicillin on a shaker incubator at 200rpm, 37°C overnight. The next day the culture was added to 1L sterile liquid LB media with 100µg/mL ampicillin. The culture was incubated on shaker incubator at 37°C, 200rpm. When the optical density at 600nm reached 0.6 IPTG was added to a final concentration of 1.0mM. The culture was then incubated at 18°C, 200rpm overnight. The following day the cells were harvested from the culture by centrifugation. The culture was added to a 1L centrifuge flasks and centrifuged at 1000rpm for 10 minutes (Beckman Avanti J-30). The cell pellet was then washed by resuspension in 50mL of cold PBS and centrifuged again. The

79 washed cell pellet was then weighed. 5mL of lysis buffer per 1g of cell pellet was added and incubated for 30 minutes on ice. The lysate was then sonicated on ice using a sonicator probe (Vibra cell) 6 times at 50% amplitude for 5 second pulses with 30 second pauses. The homogenate was then centrifuged at high-speed

17,000xg for 30 minutes (Beckman Avanti JE).

3.2.1. Chromatography.

Due to the larger volumes of lysate, automated FPLC was used to enable purification protein using a 2 step method of IMAC and size exclusion chromatography to ensure purity and reproducibility.

3.2.1.1. IMAC purification using AKTA primeTM system.

A HisTrap FF column (GE Healthcare) and AKTA prime system were equilibrated

with IMAC buffer (see method 2.1.6). Cell lysate was applied to the column by

syringe. The column was then attached to the AKTA prime system and washed with

25mL (5 CV) of IMAC buffer. The column was then washed with 25mL of 5%

IMAC elution buffer (55mM imidazole). At this point the protein was then eluted by

setting a 30mL gradient to 100% IMAC elution buffer. Eluted fractions were

analysed by SDS PAGE.

3.2.1.2. Size Exclusion Chromatography using AKTA primeTM system.

Protein samples were first concentrated to 10mg/mL using an Amicon Ultra 15 spin

concentrator (10 KDa cut off). Before loading, samples were centrifuged at high

speed to remove any particulates (10,000xg for 10 minutes at 4°C). A Superdex 75

10/300 GL column (GE Healthcare) was connected to an AKTA prime system and

equilibrated with SEC buffer (20mM Tris, 150mM NaCl, pH 7.4). 500uL of protein

sample was loaded into a 500µL injection loop and injected onto the column.

Separation was carried out at 0.5ml/min flow rate, collecting 0.5mL fractions. EGFP

has a retention volume of approximately 12mL on the Superdex 75 column.

3.2.1.3. Protein Quantification.

A BCA assay kit (Pierce) was used to quantify protein concentration in samples.

Samples were added to BCA reagent and read on a 96 well microtitre plate at

562nm using a Biotek Synergy HT plate reader. A BSA standard was used in the

range 0-1000µg/mL to plot a standard calibration curve. To quantitate EGFP by

spectrophotometry, EGFP has an recorded extinction coefficient of 58000 cm-1 M-1

at an absorbance of 488nm [31]. Samples of protein (<1mg/mL) were loaded into a

10mm quartz cuvette and analysed using a Cary 50 UV-Visible spectrophotometer.

3.2.1.4. Removal of Histidine Tag by Thrombin digestion.

For immunisation studies it was necessary to remove the ancillary N-terminal

histidine affinity tag from EGFP. This was performed by thrombin protease

digestion at the site specific sequence encoded by the pET15b vector. 1 unit of

thrombin (GE Healthcare) was added per 100µg/mL of protein after IMAC

purification. To confrim successfully digested protein, the digest mixture was

repassed though a NTA agarose column (see method 3.1.6). Thrombin was then

separated from the EGFP fraction by the subsequent SEC chromatography step.

3.2.1.5. Analysis of samples by fluorescence spectroscopy.

EGFP and the variants spectral characteristics were analysed using a Cary Eclipse fluorescence spectrophotometer. Protein samples (10-100µg/mL) were loaded into a

81 fluorescence cuvette. Fluorescence profiles were conducted by a pre-scan method where peak absorption and emission wavelengths were recorded.

The extinction coefficients of EGFP and variants were calculated by measuring the absorbance value at 488nm of EGFP solution of known protein concentration

(derived from methods 2.4.2.3). Using the Beer’s law equation the extinction coefficient was calculated. Quantum yield was calculated by generating EGFP samples and a fluorescein standard (Sigma) in 0.1M NaOH that showed equal absorbance at 488nm. Emission profiles were measured in the range 490nm-650nm.

The area under the curve of the EGFP sample emission profile was then calculated as a percentage of the fluorescein standard designated a quantum yield value of 0.85

[104]. Calculations of curve area were performed using Microsoft Excel.

3.2.2. Mammalian Expression and Imaging of Y200 variants.

3.2.3. Cloning of Y200 variants into mammalian expression vector.

Y200 variants, Y200R, Y200E and Y200T were cloned into suitable mammalian expression vector, pcDNA3.1 (Invitrogen) for expression and imaging. PCR oligonucleotide primers were designed to amplify EGFP from the plasmid clones and incorporate a 5’ KPN1 restriction enzyme digest site with Kozak sequence

(ACC) and 3’ BamHI site. Primer sequences can be found in the appendices. The

PCR reaction mix contained 5ng template pET15b/EGFP variant DNA, 500nM forward and reverse oligonucleotide primers, 200nM deoxynucleotide triphosphate mix and 2 units of proof reading Phusion polymerase (NEB) in manufacturers high fidelity buffer in a total reaction volume of 50µL. The PCR cycle consisted of denaturation at 98°C for 1 minute, 32 cycles of 98°C for 30 seconds, 70°C annealing

82 for 30 seconds, 72°C extension for 30 seconds and a final 10 minute polishing step at

72°C, performed using a Techne TC-412 thermal cycler.

An endonuclease restriction enzyme digest was performed on the EGFP amplification PCR product and pcDNA3.1. The double digest reaction mix consisted of 20 units of BamHI and 20 units Kpn1 (both from NEB), including 200ng of EGFP

PCR product or 20µg of pcDNA3.1 and 0.1mg/mL BSA with NEB buffer 4 (NEB) in a final volume of 50µL.

The reaction was incubated at 37°C overnight. The digest products were separated by agarose gel electrophoresis and extracted using a Qiaquick gel extraction kit

(Qiagen). The concentration of purified DNA was then measured using a Nanodrop

ND-1000 spectrophotometer.

The ligation reaction was performed using the Quick ligation kit (NEB) with EGFP and pcDNA3.1 DNA in a 3:1 insert to vector molar ratio and 2000 units of T4 DNA ligase. The reaction was incubated at 25°C for 1 hour. The DNA was then purified from the reaction mix using a Qiaquick PCR purification kit (Qiagen) and concentration measured using a Nanodrop spectrophotometer. The purified ligation product was then used to transform DH5α cloning E.coli strain from which plasmid was isolated and the DNA sequence confirmed as per method 2.1.2.

3.2.4. Transfection and imaging of mammalian cells with EGFP Y200 variants.

pcDNA3.1 (Invitrogen) was selected as a suitable vector to transfect human HEK293 cells to express EGFP. A vector map is included in the appendices. 2x 106 HEK293 cells were seeded onto onto a glass bottomed 35mm tissue culture dish in 2mL tissue culture medium; RPMI 1640 (Sigma) and 10% foetal bovine serum (PAA) with 100 units/mL of penicillin and 100 µg/mL of streptomycin (Sigma). Cells were incubated

83 at 37°C and 5% CO2 overnight. The cells were transfected next day when they had reached 60-80% confluence. 2µg of plasmid DNA was added to 200µL JetprimeTM buffer with 4µL JetprimeTM transfection reagent (Polyplus). This was added to the cells and the culture media was replaced after 4 hours incubation. Cells were imaged the next day after an overnight incubation.

Images were acquired on an AS MDW live cell imaging system (Leica) using a

20x/0.5 objective (HC Plan Fluotar), with EGFP filter set (Chroma 59222) and a

Precise LED fluorescent light source. The images were collected using a Coolsnap

HQ Photometrics camera and Image Pro 6.3 (Media Cybernetics) Imaging software with 0.1 second acquisition exposure.

3.3. Results.

3.3.1. Cloning, Expression, Purification of Enhanced Green Fluorescent

Protein.

Initially a suitable recombinant expression system for wild type EGFP was designed to enable isolation of protein, mutation of MHC anchor sites and characterisation of protein fluorescence. The source EGFP gene was orginally located in a mammalian expression vector (pGL3) and therefore needed to be cloned into a suitable E.coli expression vector. pET15b was selected as it contains an N-terminal 6x polyhistine tag (His Tag) for affinity purification which can be removed from the protein by peptide sequence specific cleavage with thrombin protease. PCR primers were designed to amplify EGFP and incorporate restriction enzyme sites to enable ligation into pET15b. Figure 3.1 shows that at both annealing temperatures tested, 60°C and

70°C, the EGFP gene can be specifically amplified by PCR. The EGFP PCR product was then digested using restriction enzymes and ligated into pET15b vector (Figure

3.1 B). EGFP gene insertion integrity was confirmed by DNA sequencing, figure 3.2, demonstrating 100% sequence identity with the known EGFP sequence (NCBI accession: AFA52654.1), in frame with the transcription start site. Expression of

EGFP was readily confirmed by transformation of pET15b/EGFP into the

BL21(DE3) strain and screening resulting colonies for green fluorescence under blue light (Figure 3.3 A). To ensure over expression and purification of EGFP was plausible, a pilot scale expression system was tested. EGFP expression was induced in overnight 50mL cultures of expression E.coli strain BL21(DE3) and protein isolated from the cell lysate using IMAC chromatography.

Annealing Temperature (a) 60° 70° (b) EGF pET15b C C P 6Kb - 6Kb 3Kb - - 1Kb 1Kb - -

N121 Y200

6Kb 3Kb - - 1Kb -

Figure 3.1. Cloning, plasmid construction and mutagenesis of EGFP gene. 1% agarose gel electrophoresis performed in TBE buffer. (a) PCR amplification of the EGFP gene from the pGL source plasmid EGFP bands. (b) Restriction digest products of EGFP and pET15b. (c) PCR products of saturation site directed mutagenesis of MHC anchor site residues.

His Tag Thrombin Cleavage EGFP Start Site

Figure 3.2. DNA Sequence confirmation of EGFP cloning into pET15b. Sequence alignment of constructed pET15b/EGFP with EGFP DNA sequence (NCBI accession: AFA52654.1).

Purification was successful with isolation of a single protein species at approximately 29kDa on SDS PAGE analysis. To enable production of larger yields in mg quantities for immunisation studies the expression culture volume was scaled up. 1L cultures of BL21(DE3)/pET15b/EGFP were induced with IPTG and incubated overnight. Automated protein purification using the AKTA prime system was used to enable purification of larger volumes of cell lysate and improve reproducibility. Using a 5mL HisTrap FF IMAC column, high concentrations of His

Tag EGFP protein were captured and purified. Figure 3.4-A shows SDS PAGE analysis of HisTrap elutions. A single major species of protein at 29kDa representing

EGFP was isolated.

EGFP expression and purification were favourable and at this stage high yields of protein between 200-400mg per litre of culture of protein were commonly recorded.

To ensure purity of the protein for downstream experiments, including immunisation studies the IMAC purified protein was then separated by SEC. A GE Life Sciences

Superdex 75 10/300 GL column was chosen as it resolves molecular species between

3kDa and 70kDa to isolate IMAC purified EGFP from other impurities by molecular weight. Figure 3.4 B shows SDS PAGE analysis of SEC elutions. A single major species of protein began eluting at approximately 12mL, representing 29kDa EGFP.

These elutions were pooled and protein concentration measured for downstream experiments.

(a)

(b)

MW Marker Total FT Wash Elution (kDa)

250 -

130 - 100 -

70 -

55 -

30 - 25 -

Figure 3.3 Recombinant expression of pET15b EGFP in E.coli. (a) Petri dish of LB agar with BL21(DE3) E.coli colonies induced to express pET15b/EGFP visualised under blue light excitation with amber filter. (b) SDS PAGE gel of gravity flow IMAC chromatography fractions. Total: total protein loaded, FT: flow through after load. Eluted protein with band at 29KDa equates to EGFP with His Tag.

(a)

MW Total FT Wash Elution Marker s (kDa) 170 - 130 - 100 - 70 - 55 -

40 -

35 -

25 -

(b)

MW Marker Total Elution (kDa) s 170 - 130 - 100 - 70 - 55 -

40 -

35 -

25 -

Figure 3.4 Chromatography of 1L scale EGFP expression. (a) SDS PAGE gel of FPLC IMAC EGFP wild type chromatography fractions. (b) SDS PAGE gel of FPLC SEC EGFP wild type fractions. Total: total sample loaded, FT: flow through after load.

MW 4° 25°C 30°C Marker C (kDa) 1U 2.5U 5U 1U 2.5U 5U 1U 2.5U 5U Contro l 40 -

35 -

25 -

Figure 3.5 Thrombin cleavage of His Tag from EGFP. SDS PAGE gel of Thrombin digested EGFP product. 100µg/ml EGFP was incubated with 1/2.5/5 Units of Thrombin at different incubation temperatures for overnight. Thrombin cleavage of His Tag can be seen to be effective in all conditions by decreased molecule weight from undigested control sample from 29kDa to 27kDa.

3.3.2. Histidine Tag Removal.

For downstream applications it was a requirement that the N-terminal His Tag and ancillary sequence be removed to leave only the EGFP protein sequence. A thrombin digest was performed on the protein preparation after HisTrap Purification. Figure

3.5 shows an SDS PAGE gel showing successful removal of the His Tag identified by a drop in molecular weight of the main protein band of EGFP after treatment of thrombin under varying concentrations. To remove background thrombin enzyme the protein digest was subsequently purified by SEC.

3.3.3. Site Directed Mutagenesis.

Site directed mutagenesis was used to mutate the selected Y200 and N121 MHC anchor sites of the immundominant epitopes of EGFP as discussed in chapter 2.

Mutagenic primers complimentary to the anchor site regions were designed to incorporate random codons at these sites and amplify the whole pET15b/EGFP vector during a PCR cycle. Of the two tested annealing temperatures, 65°C yielded more specific product with less incomplete amplification for Y200 (Figure 3.1 C).

PCR of N121 was acceptable at both annealing temperatures. The linear PCR products were then transformed into DH5α cloning E.Coli strain to enable recombination and complete circularisation of the plasmid. The generated mutation plasmid libraries were then used to transform expression strain Bl21(DE3) Rosetta 2 and colonies were screened for fluorescence. Fluorescing colonies were picked and plasmid DNA purified for sequencing. Sequences were analysed by ClustalW alignment with wild type EGFP DNA sequence to determine the mutation at the relevant position.

Y200 Amino Library Codon Acid Clone # Wt TAC Y 1.01 CTC L 1.02 GTC V 1.03 TGG W 1.04 CTG L 1.05 GTG V 1.06 GAG E 1.07 TAC Y 1.08 TTC F 1.09 TGG W 1.10 TGG W

1.11 TGG W 1.12 AGG R 2.01 CTC L 2.02 GTG V 2.03 AGG R 2.04 TTG L 2.05 TAC Y 2.06 TAC Y 2.07 TTG L 2.08 CGG R 2.09 ACG T 2.10 ATC I Table 3.1 Y200 Mutation Library. Sequencing results and corresponding coding amino acid of isolated fluorescent Y200 variants.

100 Experimentally Isolated 90 Phylogenetically Conserved

70 60 50 40 Frequency % Frequency 30 20 10 0 E F H I K L R T V W Y Amino Acid at position Y200

Figure 3.6. Frequency of fluorescent Y200 mutations. Y200 mutations isolated from the muation library were compared to known amino acids at the same position from the PFAM alignment of 248 known GFP-like proteins from Cnidaria (PF01353).

3.3.4. Y200 Mutation Library.

Table 3.1 shows results of sequencing and figure 3.6 shows the frequency of mutations within the library. Mutagenesis successfully generated a diverse library of mutations from the parent sequence.

To put the mutations into context, in figure 3.6 these were compared to the occurrence of amino acids at the same position in a structural alignment of all identified GFP-like proteins of Cnidaria species from the PFAM database. Briefly, the PFAM database groups protein sequences based on sequence alignment of domains and groups proteins as families with high accuracy [105]. The raw GFP clan

(n=306, accession number: PF01353) was manually screened for repetitive entries of

GFP (e.g. GFP fusion proteins) and non Cnidaria sequences resulting in an adapted database of 248 sequences. On analysis of the alignment, non-polar, hydrophobic residues are preferred and maintained at the equivalent Y200 position with high conservation of Y and F. Correspondingly, a number of conservative changes (F, W) and semi-conservative changes (L, I and V) to non-polar residues were isolated from the Y200 mutation library. The mutation screen isolated a number of silent mutations to Y. This could be contributed to presence of background undigested parent plasmid template or indeed a selection pressure to maintain Y for function. Interestingly, less conservative changes to polar residues R, E and T were also isolated. To analyse whether these mutation have reduced immunogenicity of the HYLSTQSAL, MHC epitope the variant peptide sequences were resubmitted to the IEDB (as per method

2.2.1). The MHCI binding predictions were made using the IEDB analysis resource and ANN tool to give the most reliable prediction as previously discussed. Results of this are shown in table 3.2. Mutations of Y200 to changes R, E and T decreased the predicted immunogenicity to the greatest extent.

Y200 Mutation Peptide Predicted IC50 (nM)

Y (Wt) HYLSTQSAL 6

F HFLSTQSAL 30

I HWLSTQSAL 130

L HLLSTQSAL 330

V HILSTQSAL 360

W HVLSTQSAL 540

E HRLSTQSAL 860

R HELSTQSAL 860

T HTLSTQSAL 990

Table 3.2. Impact of the isolated Y200 mutations on the predicted immunogenicity of the H2-Kd immunodominant epitope. Calculations performed by using the IEDB, ANN method. IC50 (nM) represents the predicted affinity value for peptide/MHC complex. Values ≤50nM are considered strong interactions, ≤500nM are considered intermediate, ≤5000nM are considered weak and ≥5000nM are considered negligible. Listed in order of decreasing predicted MHC binding.

The IC50 affinity values, the concentration of epitope peptide at which 50% is in complex with the MHC molecule, significantly decreased from 6nM in the wildtype

HYLSTQSAL peptide to 860nM, 860nM and 990nM in Y200E, Y200R and Y200T respectively. This notable drop in predicted affinity is likely due to the nature of these non-conservative mutations. The polar side chains are likely less compatible with epitope peptide docking with the MHC where the Y200 aromatic ring docks into a hydrophobic pocket (Figure 3.2). The three variants Y200R, Y200E, Y200T were therefore selected as candidates for lowering the H2-Kd epitope immunogenicity and carried forward for further characterisation studies.

3.3.5. Spectral Characterisation of Y200 variants.

The three candidate variants Y200R, Y200E, Y200T were expressed in E.coli and purified (as per method 2.1.5). Fluorescent properties of the variants were analysed using a fluorescence spectrophotometer and data shown was representative of 3 separate protein preparations. Figure 3.7 shows an overlay of excitation (A) and emission (B) spectra compared to wild type EGFP. All three variants show very similar spectra compare to the wild type. There are very minor differences in peak absorption and emission compared to wild type (Table 3.3) but these variations do not significantly impact on fluorescence functionality. Small variations in spectra like this have been previously reported for several other mutations of GFP [31].

Extinction coefficients were calculated by measuring the optical absorbance at

488nm with a known concentration of pure sample. The recorded value for the wild type EGFP (58000 ± 3000 M-1 cm-1) was in line with typical recorded values from other authors [31]. Interestingly all three variants show elevated extinction coefficients compared to the wild type, suggesting an improved efficiency of the

97 chromophore in the variants to absorb light quanta. Quantum yield, which is a measure of the ratio of photons absorbed to emitted during fluorescence is measured in comparison to a fluorescein standard. For convenience quantum yield can be multiplied by the extinction coefficient to give an comparable value for brightness and expressed as a percentage of the wild type. All three candidate Y200 variants showed equivalent brightness within error demonstrating that the mutations have in no way diminished fluorescent function, and have only served to improve fluorescence properties of the molecule.

(a)

Wt 100.00 Y200R Y200E 80.00 Y200T

60.00

40.00 RelativeFlouresence

20.00

0.00 350.00 400.00 450.00 500.00 550.00 Wavelength (nm)

(b)

100.00 Wt Y200R Y200E 80.00

Y200T

60.00

40.00 RelativeFlouresence 20.00

0.00 480.00 530.00 580.00 630.00 Wavelength (nm)

Figure 3.7. Excitation and emission spectra of candidate Y200 variants. Comparison of Y200 variants to wild type (Wt) EGFP. (a) Absorption spectra with fixed emission at 510nm. (b) Emission spectra with excitation fixed at 490nm. Equivalent 490nm absorbing samples were prepared in 50mM Tris, 150mM NaCl, pH 8.0. Spectra were corrected to generate equal peak excitation and emission. 99

Brightness Relative Peak Peak Extinction Quantum Y200 ((EXT*Q Brightness Excitation Emission Coefficient Yield Variant Y) to Wt (nm) (nm) M-1 cm-1 (EXT)* (QY)* /1000) EGFP (%)

EGFP Wt 490 509 58000 ± 3000 0.68 ± 0.01 39 ± 2 100 ± 5 Y200R 486 509 65000 ± 3500 0.68 ± 0.02 44 ± 3 113 ± 8 Y200E 488 511 64000 ± 2000 0.71 ± 0.03 45 ± 3 112 ± 5 Y200T 488 509 62000 ± 3000 0.69 ± 0.02 43 ± 3 108 ± 8

Table 3.3. Spectral characteristics of candidate Y200 variants. Fluorescent characteristics of purified protein was analysed by fluorescence spectrophotometry. Brightness is calculated from the product of extinction coefficient and quantum yield. Relative brightness is in comparison to wild type EGFP (Wt). For extinction coefficient and quantum yield values the mean ± SD for each protein was determined from 3 measurements from 3 independent protein preparations.

100

3.3.6. Size exclusion chromatography of Y200 variants. Under standard physiological conditions wildtype EGFP is a monomer [106] and is eluted during size exclusion chromatography as a single peak representing a monomeric species. Analysis of size exclusion chromatograms (Figure 3.8) shows that the EGFP/Y200 variants have identical, single peak, elution profiles and retention volumes at approximately 12mL, demonstrating that the mutations have not significantly altered size behaviour or formation of undesirable multimers.

3.3.7. Mammalian Cell Imaging of Y200 variants. To test whether the EGFP variants expressed efficiently in mammalian cells and showed no detrimental effects the variants were transiently transfected into HEK293 cells and imaged. EGFP wildtype and EGFP/Y200T variant gene were amplified by

PCR and cloned into the pcDNA3.1 vector by restriction digest. HEK293 cells were cultured and transfected with either pcDNA3.1/EGFP or pcDNA3.1/EGFP/Y200T.

Figure 3.9 shows fluorescence microscopy images of HEK293 cells successfully expressing EGFP wild type and EGFP/Y200T.

101

120.00 Wt Y200R 100.00 Y200E

Y200T

80.00

60.00

Absorbance 280nm 280nm Absorbance 40.00

20.00

0.00 5.00 10.00 15.00 20.00 Retention Volume / mL

Figure 3.8. Size exclusion chromatography of candidate Y200 variants. The size behaviour of Y200 variants was analysed by size exclusion chromatography on a Superdex 75 10/300 GL column. Chromatography was run in 50mM Tris, 150mM NaCl, pH 8.0. Absorbance at 280nm represents protein concentration.

102

(a) (b)

Figure 3.9. Live imaging of HEK293 cells expressing EGFP constructs. HEK293 cells were transiently transfected with pcDNA3.1 EGFP (a,b) or EGFP Y200T (c,d) and imaged at 24hrs under bright field (a,c) and with fluorescence microscopy (b, d). 20x magnification.

103

3.3.8. N121 Mutation Library. For the H2-Db allele (C57BL/6 mice), N121 was determined as the anchor residue

(2.3.4). The N121 site was subject to saturation mutagenesis and a library generated using the same method as performed with Y200. Unlike the Y200 library, N121 did not show similar diversity of fluorescent isolates. Table 3.4 shows that 75% of all mutations isolated were silent, resulting in Asn. Only N121S was isolated from the library, however the colonies exhibited weak fluorescence, indicating low expression levels despite appropriate codon usage for clone 1.02. To confirm the high frequency of silent mutations was not an effect of the presence of template DNA and the mutagenesis PCR had worked, controls of two weakly fluorescent and non- fluorescent colonies were also selected by eye and sequenced. Non-conservative mutations identified in these clones resulting in reduction (N121>G/W) and loss

(N121>R/P) of fluorescence. Due to the lack of diversity obtained an alternative approach was taken by performing directed mutagenesis of N121 to alanine and aspartate, as both mutation have significant impact on decreasing the predicted MHC binding of the dominant epitope peptide and considered conservative changes (table

3.5). On isolation of the mutants N121D did not exhibit fluorescence and investigation was discontinued. N121S and N121A were taken forward for characterisation studies. In figure 3.10 both N21A and N121S display similar major peak absorption and emission spectra compared to wildtype EGFP. However N121A shows a secondary minor excitation peak which is equivalent to the unmodified native GFP which has a major peak excitation at 395nm and therefore diminished brightness and quantum yield under 488nm, blue light (see chapter 1).

104

Amino Library # Codon Acid

WT AAC N 1.01 AAC N 1.02 AGC S* 1.03 TCC S* 1.04 AAC N 1.05 AAC N 1.06 AAC N

2.01 AAC N 2.02 AAC N 2.02 AAC N 2.02 AAC N 2.02 AAC N 2.02 AAC N 3.01 AAC N 3.02 AAC N 3.03 AAC N 3.04 AAC N 3.05 AAC N 3.06 AAC N

4.01 AAC N 4.02 AAC N 4.03 GGG G* 4.04 TGG W* 4.05 CGG R** 4.06 CCC P** N121A GCC A N121D GAC D**

Table 3.4 N121 Mutation Library. Sequencing results and corresponding coding amino acid of isolated fluorescent N121 variants. * Isolated from weakly fluorescent and ** non-fluorescent colonies.

105

N121 Peptide IC (nM) Mutation 50

N (Wt) DTLVNRIEL 730

S DTLVSRIEL 20000

A DTLVDRIEL 20000

D DTLVARIEL 26000

Table 3.5. Impact of the isolated N121 mutations on the predicted immunogenicity of the H2-Db immunodominant epitope. Calculations performed by using IEDB, ANN method. IC50 (nM) represents the predicted affinity value for peptide/MHC complex. Values ≤50nM are considered strong interactions, ≤500nM are considered intermediate, ≤5000nM are considered weak and ≥5000nM are considered negligible. Listed in order of decreasing predicted MHC binding.

106

(a) Wt 100.00 N121A N121S 80.00

60.00

40.00

20.00

Relative Fluoresence (RFI) Intensity Fluoresence Relative 0.00 300.00 350.00 400.00 450.00 500.00 550.00 Wavelength (nm)

(b)

100.00 Wt

N121A 80.00 N121S

60.00

40.00

20.00

Relative Fluoresence (RFI) Intensity Fluoresence Relative 0.00 480.00 530.00 580.00 630.00 Wavelength (nm)

Figure 3.10. Excitation and emission spectra of candidate N121 variants. Spectral profiles of purified protein were compared to wild type EGFP. (a) Absorption spectra with fixed emission at 510nm. (b) Emission spectra with excitation fixed at 490nm. Equivalent 490nm absorbing samples were prepared in 50mM Tris, 150mM NaCl, pH 8.0. Spectra were corrected to generate equal peak excitation and emission.

107

Peak Peak N121 Excitation Emission Variant (nm) (nm) EGFP WT 490 509 N121A 491 (398) 510 N121S 491 509

Table 3.6. Spectral characteristics of N121 variants. Fluorescent characteristics of purified protein was analysed by fluorescence spectrophotometry.

108

3.4. Summary

To form the basis of a study into the immunogenicity of EGFP, purified recombinant

EGFP protein was initially required. The optimisation of an appropriate E.coli expression and purification system for production and characterisation of EGFP protein was successful. The expression system generated high yields of soluble protein at high purity suitable for downstream analysis and immunogenicity testing.

Site directed mutagenesis of the target MHC anchor residues was employed to screen for functional fluorescent variants with mutations that reduce predicted immunogenic potential. Exploiting EGFP’s inherent fluorescent property, screening for functional variants was made simple and effective.

Using saturation mutagenesis three Y200 variants were identified, indicated for the

H2-Kd immunodominant epitope with significantly reduced predicted MHC binding where IC50 > 500nM indicating low immunogenic potential [86] whilst maintaining comparable fluorescence properties compared to the parent EGFP protein. Small shifts in optical properties were noted in the variants obtained, with slightly elevated extinction coefficients. Calculations were performed on three separate protein preparations for each variant. Small changes in accuracy of measuring protein concentration by BCA method for example can lead to changes in calculated extinction coefficient. Within the limits of accuracy for the methods these differences in spectral properties are therefore deemed insignificant and it was key that the mutation had not at worst diminished fluorescence.

However, isolation of N121 variants comparable to parent EGFP proved more complex. N121 is a core side chain in close proximity to the chromophore (see

2.3.3.2). It is very likely that mutations at this point are less tolerated compared to

109

Y200. Mutations at this position may have led to detrimental effects on the neighbouring chromophore and protein folding resulting in a loss of fluorescence. At this point, given time restrictions it was therefore decided that investigations into

N121 (H2-Db) variants ceased and focus was placed onto the Y200 mutations (H2-

Kd) to focus the in vivo studies on to a single mouse strain.

110

Chapter 4

4. An examination of the immunogenicity of EGFP and EGFP Y200T in

BALB/c mice

Transplanted EGFP expressing cells have previously been shown to be immunogenic in BALB/c strain mice [60, 63, 69, 70, 107]. The immune response to EGFP has been detected as a function of a number of different endpoints, using various methods including cytometric detection of EGFP cell loss in vivo, in vitro cytotoxicity assessment, cytokine ELISA, ELISpot (for enumeration of cytokine secreting cells) and serum antibody response to EGFP. The cytotoxic T lymphocyte assay (CTL) is a key method for identifying an antigen specific cytotoxic CD8+ response through measurement of cell killing in vitro. Previously this has been carried out by immunising mice with stable EGFP expressing tumour cell lines or antigen presenting cells transduced in vitro with virus to express EGFP in order to allow the development of antigen-specific CTL in vivo [69, 70]. After priming with

EGFP in vivo, the splenocytes are isolated and co-cultured in vitro with non- proliferating EGFP expressing cells to expand the antigen specific CD8+ population.

After a set period of expansion the effector cells are then rechallenged with radiolabelled EGFP target cells and specific target cell killing can be measured by release of radiolabel.

Having selected Y200T as the lead candidate EGFP variant with reduced predicted immunogenicity an in vivo BALB/c mouse model was established using several different routes of exposure to immunise mice with EGFP/Y200T and downstream tests to assess immunogenicity.

111

4.1. Aims

 Establishment of stable EGFP expressing cell lines as target cells for CTL

assay.

 Development of a CTL assay and ELISA for cytotoxicity markers (interferon

gamma [IFNγ] and granzyme B.

 Development of ELISA methods for the detection of EGFP antibody

responses.

 Assessment of immune response to EGFP following:

o Immunisation of mice with recombinant EGFP.

o Immunisation of mice with BMDC/EGFP.

o Immunisation of mice with A20/EGFP cell line.

4.2. Methods

4.2.1. Maintenance of A20 cell line

A20, a BALB/c mouse B-cell lymphoma cell line, was obtained from ATCC. A20 cells were cultured in RPMI 1640 (Sigma) and 10% (v/v) heat inactivated foetal bovine serum (FBS, Gibco), 50µM 2-mercaptoethanol (Gibco) supplemented with

2mM L-glutamine (Gibco), 100 units/mL of penicillin and 100µg/mL of streptomycin (Sigma). Cells were cultured in 25cm2 or 75cm2 vented tissue culture flasks (Corning) at 37°C in an atmosphere of 5% CO2, 95% humidity. Cultures were passaged every 2-3 days; viable cells were counted by 0.05% trypan blue exclusion and seeded at 1x105 cell per mL and maintained between 1x105 and 2x106 cells per mL. A20 cells were maintained for up to 10 passages. A PCR based test for

112

Mycoplasma (Biological Industries) was routinely used to ensure cultures were free of infection.

4.2.2. Construction of stable A20 EGFP expressing cell line

Lentivirus with EGFP or EGFP/Y200T under expression of the elongation factor-1- alpha (EF1α) promoter and puromycin selection marker were generated by

Geneocopeia (Rockville, MD, USA). 1x105 A20 cells in 0.5mL of culture medium were seeded into a well of a 12 well tissue culture plate and incubated 37°C, 5% CO2 overnight. The next day the media was replaced with 0.5mL fresh culture medium containing 1µL of virus particle supernatant (5x106 particles) with 5ng/mL of polybrene (Sigma) and incubated at 4°C for 2 h. The next day the 0.5mL culture was expanded to 5mL of fresh complete media in a 25cm2 culture flask and incubated for

2 days. At day 5 the cells were reseeded at 2x105 cells/mL in complete medium with

1µg/mL puromycin antibiotic selection agent (Sigma). Cells were continuously cultured in the presence of 1µg/mL antibiotic for a further 14 days until EGFP expressing cells had been enriched > 90% as analysed by flow cytometry using a

FACScalibur flow cytometer (BD).

4.2.3. BMDC Culture

Bone marrow was flushed from the tibia and femur of a 6-8 week old female

BALB/c mice (Harlan Olac) with PBS (Sigma). Harvested cells were centrifuged at

200 x g for 5 min and a viable cell count was performed by exclusion of 0.05%

Trypan blue (Sigma). Cells were seeded at 2x106 cells per 10mL of RPMI 1640

(Sigma) and 10% heat inactivated FBS (Gibco), 50µM 2-mercaptoethanol (Gibco) supplemented with 2mM L-glutamine (Gibco), 100 units/mL of penicillin,

100µg/mL of streptomycin (Sigma) and 20ng/mL mouse GM-CSF (granulocyte

113 macrophage colony-stimulating factor, Miltenyi Biotec) onto a petri dish. The cells were incubated at 37°C, 5% CO2, 95% humidity. On day 3 of incubation the culture was supplemented with 10mL of fresh media. On day 6 10ml of old media was aspirated and the culture was supplemented with 10mL fresh media. On day 8 the cells were harvested from the plate using a Pasteur pipette and transferred to a 50mL

Falcon tube and centrifuged at 200 g. Viable cells were counted by 0.05% trypan blue exclusion.

4.2.4. Transduction of BMDC with EGFP lentivirus

BMDC culture was performed as above. On day 1 of culture, 1mL of complete medium containing 10µL lentivirus supernatant (5x107 particles) and 50µg/mL polybrene was added to the dish and incubated for 2 h at 4°C, then incubated at

37°C, 5% CO2 overnight. On day 2 the media was then replaced and supplemented with 10mL of additional media and refreshed again at day 6. Cells were harvested as above at day 8 and analysed for EGFP expression by flow cytometry using a

FACScalibur flow cytometer (BD).

4.2.5. Splenocyte isolation and culture

Spleens were removed from 6-8 week old BALBC/c or C57BL/6 mice (both strains supplied by Harlan Olac), transferred to a petri dish and washed in RPMI 1640

(Sigma) and 10% heat inactivated FBS(Gibco), 50µM 2-mercaptoethanol (Gibco) supplemented with 2mM L-glutamine (Gibco), 100 units/mL of penicillin,

100µg/mL of streptomycin (Sigma). The spleen was passed through stainless steel gauze to generate a single cell suspension and centrifuged at 200 g for 5 min. The cells were then resuspended in 3mL of 0.85% (w/v) ammonium chloride and gently

114 shaken for 5 m. The cells were then washed twice by centrifugation at 200 g and resuspensed in media. Viable cells were counted by 0.05% trypan blue exclusion.

4.2.6. Immunisation of Mice

Female BALB/c and C57BL/6 mice were obtained from Harlan Olac and used between 6-8 weeks old.

4.2.6.1. Recombinant EGFP protein

Mice were immunised by intraperitoneal injection with 250µL of 1mg/mL EGFP protein in PBS (Sigma) at day 0 and day 7. Mice were sacrificed at Day 14 and blood was removed, allowed to clot and serum isolated. For the heat denatured protein studies, 1mL of 1mg/mL of EGFP protein was heated to 95°C for 5 min then fully resuspended in PBS (Sigma) prior to immunisation.

4.2.6.2. Immunisation with A20 cell line and BMDC

For BMDC studies mice were immunised with 1x106 BMDC in 250µL PBS by

subcutaneous or intraperitoneal injection followed by a repeat injection at day 7. For

tumour cell line studies the A20 cell line was treated with mitomycin C to arrest

proliferation and tumourgenicity. A20, A20/EGFP or A20/EGFP/Y200T cells were

harvested from culture and resuspended in PBS with 25µg/mL mitomycin C

(Sigma) and incubated at 37°C for 30 min. After incubation the cells were washed

with 3 times with PBS to remove trace mitomycin C. Mice were immunised by

subcutaneous or intraperitoneal injection with 1x106 or 1x107 A20 cells in 250µL

PBS followed by a repeat injection at day 7 and sacrificed at either day 10 or day

14.

115

4.2.7. Assessment of EGFP and membrane marker expression by flow

cytometric analysis.

For cytometry analysis of A20 and BMDC cells a FACScalibur (BD) instrument was used. Cells were harvested from culture and viability checked by trypan blue (sigma) exclusion. Cells were adjusted to approximately 1x106 cells/mL and washed in

FACS buffer. For cell maker analysis, after washing, 200µL/well of cells was aliquoted into individual wells of a round bottomed 96 well plate. The plate was then centrifuged at 200 x g at 4°C. A20 cells were then stained with allophycocyanin

(APC) conjugated antibodies for MHC class I (2µg/mL purified mouse anti-mouse

H2-Kd) with isotype control (2µg/mL purified mouse IgG2a κ) and MHC class II

(2µg/mL purified rat anti-mouse I-A I-E) with isotype control (2µg/mL purified rat

IgG2b κ) in FACS buffer (all antibodies from Biolegend). Cells were incubated on ice for 30min and then washed three times by centrifugation at 200 x g and resuspension in FACs buffer. On their final wash cells were resuspended in sodium azide buffer. Flow cytometric analyses were carried out using the FACScalibur machine, CellQuest Pro and FlowJo software. Cells (10,000) were acquired and viable cells were determined by exclusion of 10 μg/ml propidium iodide (PI; Sigma) or 10 μg/ml 7-amino-actinomycin D (7-AAD; BD).

4.2.8. ELISA

4.2.8.1. Mouse interferon gamma ELISA

IFN γ mouse ELISA (R&D systems) was carried out according to the following manufacturer’s protocol. A 96-well immunoassay plate (Maxisorb, Nunc) was coated with 4 μg/mL IFN γ capture antibody (R&D Systems), in PBS (Sigma) and incubated overnight at 4oC. Plates were washed 3 times with wash buffer (PBS

116 containing 0.05% Tween-20 [Sigma]). Wells were then blocked with 1% BSA/PBS for 1 hour. A serial dilution of standard recombinant mouse IFN γ (2000pg/mL to

62.5pg/mL) and samples were diluted using RPMI-1640 media containing 10% FBS.

Samples and standards were added at 100 μL/well to plates in duplicate and triplicate respectively and incubated for 2 h at room temperature. Wash steps were then repeated and 100 μL of 0.6 μg/mL of biotinylated detection antibody was added to all wells in reagent diluent buffer (20 mM Trizma base, 150 mM NaCl, 0.1% BSA,

0.05% Tween 20, pH 7.4) and incubated at 4oC overnight. Wash steps were repeated, and 100µL Streptavidin horse raddish peroxidase HRP conjugate (R&D systems) was added to all wells in reagent diluent buffer and incubated for 30 min at room temperature. Wash steps were repeated and 100uL of tetramethylbenzidine

(TMB, Sigma) substrate was added to each well and incubated in the dark for 15 min. The reaction was stopped with 50 μL/well 2M HCl. Plates were read at 450 nm using a ELx800 plate reader (Biotek) and data acquired with Gen5 software (Biotek).

4.2.8.2. Mouse granzyme B ELISA

Mouse granzyme B ELISA (eBioscience) was carried out according to the following manufacturer’s protocol. A 96-well immunoassay plate (Maxisorb, Nunc,) was coated with a 1 in 250 dilution of granzyme B capture antibody (eBioscience) in

PBS (Sigma) and incubated overnight at 4oC. Plates were washed 3 times with wash buffer (PBS containing 0.05% Tween-20 [Sigma]). Wells were then blocked with assay diluent buffer (1% BSA/PBS) for 1 hour. A serial dilution of standard recombinant mouse granzyme B (5ng/mL to 0.16ng/mL) and samples were diluted using RPMI-1640 media containing 10% FBS. Samples and standards were added

100 μL/well to plates in duplicate and triplicate respectively and incubated for 2 h at

117 room temperature. Wash steps were then repeated and 100 μL of 1 in 250 of detection antibody (eBioscience) was added to all wells in assay diluent buffer and incubated at 4oC overnight. Wash steps were repeated, and 100µL avidin HRP

(eBioscience) was added to all wells in assay diluent buffer and incubated for 30 min at room temperature. Wash steps were repeated and 100µL of tetramethylbenzidine

4.2.8.3. Serum anti-EGFP IgG ELISA

Blood was collected from mice by cardiac puncture, allowed to clot and stored overnight at 4°C and centrifuged at 1000 g and serum supernatant aspirated. A 96- well immunoassay plate (Maxisorb, Nunc,) was coated with 50µL of recombinant

EGFP (50µg/mL in PBS [Sigma]) and incubated at 4°C overnight. Plates were washed 3 times with wash buffer (PBS containing 0.05% Tween-20 [Sigma]). Wells were then blocked with 2% BSA/PBS at 37°C for 30 min. Serum samples were serially diluted in 1% BSA/PBS and were added 100 μL/well and incubated for 3 h at room temperature. Wash steps were then repeated. 100 μL of either 1 in 4000 of sheep anti mouse IgG HRP,1 in 2000 rat anti mouse IgG1 HRP or 1 in 1000 rat anti mouse IgG2a HRP (all from AbD Serotec) detection antibody diluted in 1%

BSA/PBS was added to each well and incubated for 2 h at room temperature. O- phenylenediamine dihydrochloride (OPD; Sigma) substrate was formulated in 25 mL citrate-phosphate buffer (47 mM Na2HPO412H2O [Fisher Scientific], 26 mM Citric

Acid [Sigma] pH5), 0.04 g OPD, 250 μL urea hydrogen peroxide (UHP; 1 g tablet

[Sigma] dissolved in 25 mL citrate phosphate buffer). Wash steps were repeated and

100 L/well of OPD substrate was added and incubated in the dark for 15 m. The 118 reaction was stopped with 50 μL/well 0.5 M citric acid (Sigma). Plates were read at

450 nm using a ELx800 plate reader (Biotek) and data acquired with Gen5 software

(Biotek). Serum titres were determined as the serum dilution that gave an a bsorbance reading at ≥ three times the background level of naive mouse serum.

4.2.9. Cytotoxic T cell Lymphocyte Assay.

A cell cytotoxicity assay to measure in vitro CD8+ cell killing of target cells was developed based on the previously described 3H thymidine release assay methods reported by P.Matzinger [108, 109].

4.2.9.1. Splenocyte responder cell isolation and culture.

Spleens were removed from 6-8 week old BALB/c or C57BL/6 mice (both strains supplied by Harlan Olac), transferred to a petri dish and washed in RPMI 1640

(Sigma) and 10% heat inactivated FBS (Gibco), 50µM 2-mercaptoethanol (Gibco) supplemented with 2mM L-glutamine (Gibco), 100 units/mL of penicillin,

100µg/mL of streptomycin (Sigma). The spleen was passed through stainless steel gauze to generate a single cell suspension and centrifuged at 200 x g for 5 min. Red blood cells were removed by resuspension in 3mL of 0.85% (w/v) ammonium chloride and gently shaken for 5 min. The cells were then washed twice by centrifugation at 200 x g and resuspended in media. Viable cells were counted by

0.05% trypan blue exclusion.

4.2.9.2. Stimulator cell preparation

Either A20 cell line or splenocytes were used as stimulator cells. Cells were adjusted to 1x107/mL in culture media (as above) with 25µg/mL mitomycin C (Sigma) and incubated in the dark at 37°C for 30 min. After incubation the cells were thoroughly

119 washed 4 times by centrifugation and resuspended in culture media to remove trace mitomycin C. Viable cells were counted by 0.05% trypan blue exclusion.

4.2.10. Sensitisation co-culture.

Stimulator cells and responder cells were diluted in culture media and mixed to a stimulator:responder ratio of either 10:1, 1:1 or 1:10, maintaining final densities of

2x106/mL. Cells were either plated onto 12 well plates or 10mL bulk cultures in upright T25 vented flasks (Corning). Recombinant mouse IL2 (R&D systems) was added to a final concentration of 20U/mL (WHO standard units). Cultures were incubated for 5 days in at 37°C in an atmosphere of 5% CO2, 95% humidity. Effector cells were harvested by pipetting. Effector cell density was calculated as the original number of responder cells cultured.

4.2.11. Radiolabelling of A20 target cells

A20 cells were seeded into 6 well culture plates at 2x105 /mL in culture medium and incubated overnight at 37°C in an atmosphere of 5% CO2, 95% humidity. Cultures were then radiolabelled by addition of tritiated thymidine (Perkin Elmer) to a final concentration of 185kBq/mL and cultures were incubated overnight. The following day cells were harvested by pipetting and washed twice by centrifugation at 200 g and resuspended in culture media. A viable cell count was performed on unlabelled cells which were cultured in parallel by 0.05% trypan blue exclusion.

120

4.2.12. Cytotoxicity Assay

Effector cells were adjusted to 4x107 cells/mL and 200 L added to each well of a round bottomed 96 well plate. The effectors were diluted 2-fold in media across the plate to generate a serial dilution. Radiolabelled target cells were then added to the effectors, 2x104 cells in 100 L per well. Plates were then centrifuged at 200 g for 5 min and incubated in an atmosphere of 5% CO2, 95% humidity for 4 h. Cells were harvested onto a glass fibre filter mat (Whatman) using a cell harvester (Skatron).

Filter mat samples were dissolved in Scintisafe scintillation cocktail (Fisher) for 1 h and counted using a TriCarb 1900 scintillation counter (Perkin Elmer). Parallel assays were run with unlabelled target cells to generate supernatants suitable for

Granzyme B and IFNγ ELISA.

Percentage specific lysis was calculated as;

T - E x 100 = % specific lysis T

Where T = total labelled target control in counts per minute (cpm) and E = experimental sample in cpm.

121

4.2.13. Statistical Analysis

Statistical analysis was performed using GraphPad Prism 4 software (GraphPad

Software Inc., La Jolla, CA, USA). For experiments where repeats were conducted, mean values and plus/minus standard error of the mean (SEM) were plotted as error bars. Data were analysed by using either an unpaired Student’s T test, or two way

ANOVA.

122

4.3. Results

4.3.1. Generation of a A20/EGFP stable cell line.

The A20 B-cell cell line (BALB/c) was selected as a suitable CTL stimulator and target cell type for the 3H thymidine release assay based on a previous application

[110]. Initially the A20 cell line was characterised by flow cytometry to ensure the presence of class I and class II MHC at the cell surface (Figure 4.1). The A20 cell population was > 99% positive for both MHC class I and II compared with isotype controls. In each case a sharp well defined peak was recorded, indicating a relatively homogenous population. The A20 cell line was transduced with HIV based lentivirus carrying EGFP or EGFP/Y200T under an El4α promoter. To select stable transformants, cells were cultured under antibiotic selection conditions for several passages until > 90% cells showed EGFP expression. Figure 4.2 shows the flow cytometric analysis of both EGFP and EGFP/Y200T transduced cells. Both transduction and selection protocols resulted in > 90% fluorescent cells, with the

EGFP transduced cells displaying a sharper peak indicating greater homogeneity of expression. In addition to characterising the number of fluorescent cells, the level of fluorescence per cell (mean fluorescence intensity; MFI) was recorded and the MFI of the A20/EGFP/Y200T cells was some five times lower than that recorded for the

A20/EGFP cells.

123

(a) (b)

Figure 4.1 Flow cytometric characterisationof A20 cell line. A20 cells were analysed for expression of (a) MHC class I and (b) class II markers by flow cytometry.Cells were stained with APC conjugated anti-MHC class I and class II antibodies. 1x104 cells were acquired per sample analysed. Cell viability was determined by PE exclusion. A Gate (black line) was set to include the MHC positive staining population (red line) and exclude the isotype stained control (grey line).

124

Cell Line % EGFP MFI A20 0 - A20/EGFP 95.3 940.0 A20/EGFP/Y200T 90.4 190.0

Figure 4.2 Flow cytometric analysis of EGFP expressing cell lines. A20 cells were transduced with lentivirus particles carrying EGFP or EGFP/Y200T. Cells were cultured for two weeks under 1µg/mL puromycin selection. Cells were analysed by flow cytometry with 1x104 cells acquired. Viability was checked with PE exclusion. A gate (black line) was set to include the EGFP and EGFP/Y200T positive population (red and blue lines) and exclude autofluoresence of the A20 wild type cell control (grey line). EGFP positive cells (%) and mean fluorescence intensity (MFI) were recorded.

125

4.3.2. Development of a Cytotoxic T-cell Lymphocyte (CTL) Assay.

To assess whether a cell mediated EGFP immune response had been induced in immunised animals the development of a CTL assay was required. The assay is based on the cell killing properties of CTL cells that recognise antigen expressed by target cells (i.e. EGFP). On antigen recognition and activation the CTL release cytotoxic perforins and granzyme B serine proteases which enter the target cell and lead to apoptosis. This killing activity can be measured in vitro. CTL are first generated by in vivo priming with antigen expressing cells. Antigen responder CTL are obtained from the spleen of the animal. These responder cells are then co cultured in vitro with antigen presenting stimulator cells at an optimised ratio and density (stimulation phase). Stimulator cell division is blocked prior to co-culture by treatment with mitomycin C (a DNA cross linker). Exposure to antigen during this stimulation phase allows expansion of the antigen specific CTL population. Addition of recombinant IL2 (interleukin 2), a stimulatory T-cell cytokine, can help to further increase the expansion of effector cells. After a 5 day incubation, the nonadherent

CTL effectors are harvested. CTL are then challenged with radioactively labelled target cells (cells in which 3H thymidine has been incorporated into cell DNA) expressing the desired antigen (killing phase). The CTL activity is determined by measuring degradation of labelled DNA in target cells that have undergone CD8+ induced apoptosis. Rather than measuring release of radioactivity as in the traditional

51Cr release assay [69], in which the labelled chromium is released by dead and dying cells, this assay measures the amount of labelled DNA retained in viable target cells that have not been killed by CTL [109]. When the effector/target cell mixture is harvested, if target cells have been killed, the thymidine labelled DNA fragments

126 are washed through the fibre glass filters, leading to a reduction in radioactivity in the sample. If targets have not been killed by the CTL, the labelled DNA will be intact and will retained by the filters. The percentage lysis can be calculated by comparing the amount of 3H thymidine bound to the filters in the presence and absence of effector cells, i.e. in the absence of effectors, targets will be viable and the maximum amount of label will be recorded

In order to optimise conditions and to provide for a positive control, initially an allogeneic assay was conducted in which naive BALB/c splenocytes and mitomycin treated C57BL/6 splenocytes were utilised in a mixed lymphocyte reaction (MLR) co-culture. In these experiments there is no in vivo priming step, but allogeneic effectors were first expanded in vitro by culture of BALB/c effectors with allogeneic

C57BL/6 target cells (treated with mitomycin C) in the stimulation phase of the assay. After 5 days of culture the BALB/c effectors were harvested and cultured with

3H thymidine labelled C57BL/6 target splenocytes (killing phase). Cell killing was determined by the loss of labelled cells as measured by scintillation counting. Due to the lack of histocompatibility between the cell types, co-culture of BALB/c and

C57BL/6 should generate a strong CTL response, thus this experimental protocol served as a positive control for detecting cell lysis and secreted markers during a

CD8+ cytotoxic response. Figure 4.3 shows the effect of varying the stimulator:responder ratios during the stimulation phase (initial co-culture) on the final cytotoxic response. The level of thymidine labelling in the control untreated target cells was approximately 9000 cpm. A net decrease of 3H thymidine labelled cells compared to the target cell alone control indicates cell killing. Incubation of cells in 1% Triton detergent to disrupt cell membranes indicates the amount of label that can be readily released upon cell destruction. In the presence of Triton,

127 radioactivity was reduced to approximately 3000cpm, indicating that there is not

100% loss of label as residual 3H thymidine DNA is retained on the filtermats during harvesting. There was no decrease in thymidine recorded for effector cells sensitised at a 10:1 stimulator: responder ratio, whereas decreased thymidine label, indicative of target cell killing, was recorded for effector cells at ratios of 1:1 and 1:10, with cpm of ~6000 and 5000 recorded respectively. There was no apparent effect of addition of recombinant IL-2 at any stimulator:responder cell ratio. The cpm data were used to calculate the percentage cell killing as displayed in figure 4.3 (b).

Approximately 70% cell death was recorded in the presence of detergent.

Sensitisation at a 1:1 stimulator:responder ratio resulted in ~25% cell death, whereas at a 1:10 ratio, ~40% cell lysis was recorded. These data suggest that atios <1:1 of stimulators:responders are more suitable than ratios 10:1 for generating a positive

MLR.

The next stage of assay development was to ensure the A20 cell line would be suitable as a target cell in the CTL assay. Naive C57BL/6 (allogeneic) splenocytes were isolated and co cultured with non-proliferating A20 cells in varying ratios. At the end of the co-culture effector cells were harvested and rechallenged with labelled

A20 target cells. Figure 4.4 (a) illustrates that the A20 cells were successfully and stably labelled with 3H thymidine (levels of ~ 40000 cpm in control cells) and loss of label could be detected in the presence of Triton (levels of thymidine reduced to

~5000cpm) which indicated effective cell killing. Figure 4.4 (b) shows percentage killing at various effector: target ratios compared with target cells alone. In contrast to the previous MLR where the C57BL/6 cells were the target and ratios of <1:1 were most effective, only the ratio of 10:1 showed any net killing effect. Presence of

128 recombinant mouse IL2 served to increase cell killing from 44% in the absence of cytokine to ~71% in the presence of 100U/mL IL2.

129

(a) rIL 2

1 5 0 0 0 0 U /m L 1 0 U /m L 1 0 0 U /m L

1 0 0 0 0

p c

5 0 0 0

l n 1 1 0 o : : 1 r o 0 1 : t it 1 1 n r o T C t % e 1 S tim u la to r:R e s p o n d e r R a tio g r a T

(b) rIL 2 1 0 0 0 U /m L

1 0 U /m L s

i 1 0 0 U /m L

i f

i 5 0

l n 1 1 0 o : : 1 r o 0 1 : t it 1 1 n r o T C t % e 1 g r S tim u la to r:R e s p o n d e r R a tio a T

Figure 4.3 Development of thymidine release cytotoxic T cell lymphocyte assay with a mixed lymphocyte reaction. Splenocytes from a BALB/c mouse were sensitised with mitomycin C treated target splenocytes from C57LB/6 in a 5 day culture, at varying stimulator:responder ratios in the presence of recombinant mouse IL2 (rIL2). On day 5 effector cells were recovered and rechallenged with 3H thymidine labelled C57LB/6 target cells at 100:1 effector to target cell ratio. (a) 3H thymidine was measured by scintillation counting. Target cells were lysed in 1% triton as a positive control for cell killing. (b) % specific cell lysis was calculated compared with target cell control. Mean of three separate wells per treatment group (±SEM). cpm; scintillation counts per minute.

130

(a) rIL 2

6 0 0 0 0 0 U /m L 1 0 U /m L 1 0 0 U /m L

4 0 0 0 0

p c

2 0 0 0 0

l n 1 1 0 o : : 1 r o 0 1 : t it 1 1 n r o T C t % e 1 S tim u la to r:R e s p o n d e r R a tio g r a T

(b) rIL 2

1 0 0 0 U /m L

1 0 U /m L s

i 1 0 0 U /m L

i f

i 5 0

l n 1 1 0 o : : 1 r o 0 1 : t it 1 1 n r o T C t % e 1 g r S tim u la to r:R e s p o n d e r R a tio a T

Figure 4.4 Development of thymidine release cytotoxic T- cell lymphocyte assay with target A20 cell line. Splenocytes from a C57LB/6 mouse were sensitised with mitomycin C treated A20 target cell line (BALB/c) in a 5 day culture, at varying effector to target ratios in the presence of recombinant murine IL2 (rIL2). On day 5 effector cells were recovered and rechallenged with 3H thymidine labelled C57LB/6 target cells at 100:1 effector to target cell ratio. (a) 3H thymidine released compared with unchallenged control target cells. Target cells were lysed in 1% triton as a positive control. (b) Percentage specific cell lysis was calculated compared with target control. Mean of three separate wells (±SEM). cpm; scintillation counts per minute.

131

In the experiments described to date, a single effector:target ratio of 100:1 has been used in the final stage of the assay when the extent of killing is determined; in subsequent experiments, the impact of altering the effector:target ratio in this killing phase was examined. Thus, effector cells isolated after the stimulation phase were plated out as a serial dilution prior to addition of labelled target cells, resulting a range of final ratios varying from 1:1 to 1:100. Figure 4.5 shows a clear relationship between percentage cell killing and the ratio of C57BL/6 effectors to the A20 targets in an allogeneic culture, with a ~70% killing at a 100:1 ratio and 10% killing at a 2:1 ratio. No net cell killing was detected in the syngeneic culture that was run in parallel at any effector:target cell ratio. This result confirmed that A20 tumour cells were not subject to natural killer (NK) associated cell killing by the syngeneic effectors.

132

8 0 A llo g e n e ic

S y n g e n e ic s

i 6 0

i f

i 4 0

p S

2 0 %

1 1 0 0 0 . 1 0 0 0 1 0 1

E ffe c to r : T a rg e t R a tio

Figure 4.5 Dose effect of cytotoxic T lymphocyte A20 target cell killing. Splenocytes from either allogeneic C57LB/6 (closed circle) or syngeneic BALB/c (closed square) strain mice were sensitised by culture with mitomycin C treated A20 target cell line (derived from BALB/c strain mouse) in a 5 day culture with a 10:1 stimulator:responder ratio in the presence of 10U/mL recombinant murineIL2. On day 5 effector cells were recovered and rechallenged in a serial dilution with 3H thymidine labelled A20 target cells from 1:1 to 100:1 effectors: targets. The percentage specific cell lysis was calculated compared with target control. Mean of three separate wells (±SEM).

133

4.3.3. Cytotoxicity marker ELISA.

As alternative endpoints to cell lysis in the CTL assay, ELISA was used to detect secreted markers of cytotoxicity. Both the cytokine IFNγ [111] and the serine protease granzyme B [112] are secreted by antigen activated CD8+ cells and the production of these molecules was assessed in the culture supernatants isolated following the stimulation phase. As displayed in figure 4.6, syngeneic (BALB/c versus A20) and allogeneic (BALB/c versus C57BL/6 and C57BL/6 versus A20) cell configurations were cultured in the stimulation phase. There was no detectable secretion of either IFN- or granzyme B by target cells cultured alone. Effectors and target cells alone resulted in very low level expression of both molecules

(<0.1ng/mL). Somewhat higher background levels of secretion of both IFNγ and granzyme B (<1ng/mL) were detected after syngeneic culture (BALB/c splenocytes with the A20 cell line). To demonstrate a positive cytotoxic reaction, C57BL/6 splenocytes were challenged with either the A20 cell line or BALB/c splenocytes

(both resulting in an MLR). Elevated levels of IFNγ to 2.6ng/mL and 4.2ng/mL were detected in the A20 and mixed splenocyte reactions, respectively. Correspondingly increased levels of granzyme B, 13.6ng/mL and 20.9ng/mL, were detected in the

A20 and mixed splencoyte reactions, respectively.

134

(a) 5 S Y N G E N E IC

4 M LR

L 3

/ g

n 2

0 0 L 2 L L P 2 O O A S A R R s 6 s T T v / v N N C L 6 / B / O O L C B 7 C L 5 B R T A C 7 E 5 O B s C T G v C R C E A / F T B F L E A B

(b) 2 5 S Y N G E N E IC

2 0 M LR

L 1 5

/ g

n 1 0

0 0 L 2 L L P 2 O O A S A R R s 6 s T T v / v N N C L 6 / B / O O L C B 7 C L 5 B R T A C 7 E 5 O B s C T G v C R C E A / F T B F L E A B

Figure 4.6 Secretion of cytotoxcity markers IFNγ and Granzyme B in a mixed lymphocyte reaction.

Splenocytes from either syngeneic BALB/c or allogeneic C57LB/6 (MLR) were sensitised with mitomycin C treated A20 cell line (BALB/c) or C57LB/6 splenocyte (SPL) targets in a 5 day culture at a 10:1effector to target ratio in the presence of 10U/mL recombinant mouse IL2. Effectors and targets were also each cultured alone (effector and target controls). Supernatants were analysed by sandwich ELISA for secretion of both (a) mouse IFNγ and (b) mouse granzyme B. Mean of two replicate wells with mean and range shown.

135

4.3.4. Generation of a cytotoxic immune response to EGFP in BALB/c mice.

Having optimised the CTL assay with the A20 cell lines as suitable targets, the next stage was to demonstrate an immunological response to EGFP by in vivo priming. A range of immunisation strategies were employed; using recombinant EGFP protein,

BMDC expressing EGFP and A20/EGFP cell line.

4.3.4.1. Immunisation with recombinant EGFP.

An ELISA was developed to assess the induction of a serum IgG response to EGFP.

BALB/c mice were immunised with EGFP or EGFP/Y200T purified proteins by intraperitoneal injection. Figure 4.7 shows the EGFP specific IgG response using the

EGFP parent protein (a) or the EGFP/Y200T variant as antigen in the ELISA (b).

There was little binding of control serum from naïve (non-immunised) mice to either antigen, indicating that there is no naturally occurring antibody that recognises

EGFP. Immunisation with both EGFP and the EGFP/Y200T proteins resulted in vigorous IgG antibody production, with similar antibody dilution profiles recorded.

Furthermore, there was a high degree of cross reactivity, such that sera derived from mice immunised with EGFP parent develop antibodies that cross reacted with the

EGPF/Y200T, and sera from EGFP/Y200T immunised mice bound the wild type protein to an equivalent extent. For 1 mouse, the IgG1 and IgG2a subclass distribution was analysed by ELISA. Again, control serum from naïve mice did not display any antibody binding, whereas Figure 4.8 shows that both IgG1 and IgG2a specific EGFP antibodies were present in the serum. IgG1 was present at a higher titre compared with IgG2a, indicative of a dominant helper T cell type 2 (Th2) response.

136

(a) 2 .5 E G F P

2 .0 E G F P /Y 2 0 0 T

N aive m

n 1 .5

1 .0

. O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 S e ru m D ilu tio n

(b) 2 .0 E G F P

1 .5 E G F P /Y 2 0 0 T

m N aive

0 5

4 1 .0

. O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 S e ru m D ilu tio n

Figure 4.7Anti-EGFP serum ELISA. BALB/c mice were immunised intraperitoneally with 250µg EGFP (closed circles) or EGFP/Y200T (closed squares), or were untreated (naïve; closed triangles). Serum IgG response to (a) EGFP and (b) Y200T recombinant protein antigen was measured by ELISA. Mean of 2 mice per treatment group with mean and range shown. Naive serum from a single mouse was included as a negative control.

137

2 .0 Ig G 1 N a iv e (Ig G 1 ) 1 .5

Ig G 2 a m

n N a iv e (Ig G 2 a )

0 5

4 1 .0

. O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0

S e ru m D ilu tio n

Figure 4.8 Anti-EGFP serum ELISA IgG isotyping. A single BALB/c mouse was immunised with 250µg EGFP or was untreated (naïve). Serum IgG1(closed circles) and IgG2a (closed triangles) response to recombinant EGFP protein was measured by ELISA. Naive serum from a single mouse was included as a negative control.

138

4.3.5. Immunisation with heat denatured recombinant EGFP

It has been previously reported that antigen specific CD8+ mediated cytotoxicity can be induced by immunisation with heat denatured recombinant protein [113]. This scenario is counter to the understanding that circulating soluble protein would normally be processed by antigen presenting cells and elicit an CD4+ and antibody response. Although the mechanism is not fully understood this suggested a possible approach that entailed a simple route for immunisation of mice with recombinant protein that could be utilised for testing and optimising the established CTL method for EGFP-specific responses.

Recombinant EGFP was heat denatured by heating to 95°C and used to immunise

BALB/c strain mice (n=3). To identify the type of reaction that had been produced,

IgG antibody binding to both native and heat treated EGFP protein antigen was measured by ELISA and compared with responses in serum derived from immunisation with the native protein (Figure 4.9). There was no background binding of naive serum to either the native or to the heat denatured antigen. Immunisation with the denatured protein did induce detectable IgG antibody in all 3 mice, with similar binding profiles displayed regardless of the antigen (native or denatured protein) used in the ELISA. In contrast, the native EGFP provoked a more vigorous antibody response than did the denatured protein, but such was only evident when the native protein was used as antigen. This indicates that during heat denaturation structural epitopes may have been lost that would have existed in the native protein and are able to provoke a more vigorous B cell response.

Following the same immunisation regimen with heat denatured EGFP, the CTL assay was carried out. Splenocytes from the mice immunised with heat denatured

EGFP were co-cultured with A20/EGFP cells to sensitise the effectors to the EGFP

139 antigen. After the stimulation phase, effectors were then rechallenged with either radiolabelled A20 target cells as a control or radiolabelled A20/EGFP cells. Figure

4.10 shows recorded scintillation counts for the CTL assay. There was no significant loss of signal in the A20/EGFP target treatment group compared to the control. In fact, there was a small increase in radiolabel recorded in each case where target and effector cells were co-cultured. This indicates that no cytotoxic response has been observed using this route for immunisation.

140

(a)

2 .0 d E G F P E G F P 1 .5

N aive

0 5

4 1 .0

. O 0 .5

0 .0 1 0 0 1 0 0 0 1 0 0 0 0 S e ru m D ilu tio n

(b) 2 .5 d E G F P

2 .0 E G F P

N aive m

n 1 .5

1 .0

. O 0 .5

0 .0 1 0 0 1 0 0 0 1 0 0 0 0 S e ru m D ilu tio n

Figure 4.9 IgG response to heat denatured EGFP recombinant protein.BALB/c strain mice were immunised intraperitoneally with 250µg/mL heat denatured EGFP (dEGFP, closed circles) or native EGFP (closed squares), or were untreated (naïve) (closed triangles). (a) IgG response to heat denatured EGFP adsorbed onto ELISA plates. (b) IgG response to native EGFP adsorbed onto ELISA plates. Mean of 3 mice in dEGFP group shown (±SEM), Data of single mice for the EGFP and naive groups.

141

4 0 0 0 0

3 0 0 0 0

m p

c 2 0 0 0 0

1 0 0 0 0

l 0 l o o P r 2 r F n t A t o n n G it o o /E r 0 T C C 0 2 % P A 2 F 1 A G /E 0 2 A

Figure 4.10 Cytotoxic response to A20/EGFP in mice treated with denatured EGFP recombinant protein. BALB/c mice were immunised intraperitoneally with 250µg/mL heat treated EGFP (n=4). Splenocytes were stimulated with mitomycin C treated A20/EGFP target cell line in a 5 day culture with a 10:1 effector to target ratio in the presence of 10U/mL recombinant mouse IL2. On day 5 effector cells were recovered and rechallenged with 3H thymidine labelled A20 or A20/EGFP target cells. 3H thymidine labelled cells were measured by scintillation counting. A20 and A20/EGFP target controls were analysed in the absence of effectors. A20 target cells were lysed in 1% triton as a positive control. Cpm; scintillation counts per minute.

142

4.3.6. Generation of a cytotoxic response to EGFP by immunisation with bone marrow dendritic cells expressing EGFP.

Dendritic cells (DCs) are professional antigen presenting cells that naturally take up circulating protein which are then presented as antigens to passing T cells via MHC class I and II. Therefore DCs provide a potentially potent mechanism to drive an immune response to both recombinant and transgenically expressed protein. Indeed,

Stripecke et al [69] demonstrated an EGFP specific in vitro cytotoxic response to an

EGFP expressing cell line following immunisation of BALB/c strain mice with

BMDCs transduced with an EGFP containing retrovirus.

In an attempt to replicate this experiment, BMDCs were isolated from BALB/c strain mice, transduced with EGFP lentivirus and expanded the culture in vitro with recombinant mouse GM-CSF. Cells were measured for fluorescence at day 8 of culture. Typical expression of EGFP was around 20% of total BMDC cell population, as per figure 4.11 (representative experiment). BALB/c strain mice were then immunised with 1x106 BMDC/EGFP (total cell population) and given a repeat dose at day 7. On sacrifice at day 14, there was no detectable serum IgG antibody response to EGFP in the BMDC treated mice, figure 4.12. The integrity of the

ELISA was confirmed by concurrent analysis of serum from negative control (naïve) mice and from mice immunised by intraperitoneal injection of recombinant EGFP protein. Following expansion with nonproliferating A20/EGFP target cells, splenocytes from the DC-immunised mice were challenged with target cells in the

CTL assay. As specificity controls, A20 targets as well as A20/EGFP and

A20/Y200T target cells were utilised. Figure 4.13 (a) shows that significant cell killing was observed, with a steady increase in cell lysis as the effector:target cell

143 ratio increased, reaching a maximum of ~40% lysis. However, this was replicated across all A20, A20/EGFP and A20/EGFP/Y200T targets tested and was not specific for EGFP expression alone. This extent of target cell killing in a syngeneic system had not been observed previously with effectors derived from either naïve mice or from EGFP immunised mice. It was hypothesised the BMDC per se were eliciting a cytotoxic response, possibly as a result of presentation of, and subsequent immune responses being generated to culture media related components such as FBS protein.

As a subsequent control mice were then immunised with native BMDC according to the same protocol, splenocytes expanded in the same way and used in the CTL with the three target cells. The same non-specific cell killing effect was observed, Figure

4.13 (b), with all three target cells being lysed to ~40% at high (100:1) effector:target ratios. This indicated that the BMDCs were indeed driving a cytotoxic immune response not related to the EGFP target protein.

144

Cell Line % EGFP MFI BMDC 0.00 -

BMDC/EGFP 21.8 421

Figure 4.11 Flow cytometric characterisation BMDC transduced with EGFP lentivirus. 2x106 BMDC were transduced with 5x107 EGFP lentivirus particles and cultured for 8 days. Cells were analysed by flow cytometry with 1x104 cells acquired. Viability was checked with PE exclusion. Gates were set to exlude untreated BMDC cells (representing background autofluorescence) for EGFP fluorescence.% EGFP positive and mean fluorescence intensity (MFI) were recorded.

145

2 .0 B M D C /E G F P rE G F P 1 .5

N aive

0 5

4 1 .0

. O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 Ig G T itr e

Figure 4.12 IgG response to EGFP in mice immunised with BMDC expressing EGFP. BALB/c mice were immunised twice subcutaneously with 1x106 cells transduced with EGFP lentivirus (closed circles) 7 days apart or with 250µg recombinant EGFP intraperitoneally (closed squares). Control serum from untreated (naive) mice was analysed also (closed triangles). Serum was isolated 14 days after initiation of immunisation and IgG response to EGFP measured by ELISA. Mean of 3 mice per group (±SEM).

146

(a) 6 0 A 2 0

A 2 0 /E G F P s

i 4 0 s

y A 2 0 /E G F P /Y 2 0 0 T

i f

i 2 0

p S

0 %

-2 0

1 0 0 1 0 1 E ffe c to r : T a rg e t R a tio

( b)

6 0 A 2 0

A 2 0 /E G F P

i s

y 4 0 A 2 0 /E G F P /Y 2 0 0 T

c e

p 2 0

1 0 0 1 0 1

E ffe c to r : T a rg e t R a tio

Figure 4.13 Cytotoxic response to A20/EGFP in mice immunised with BMDC expressing EGFP. BALB/c mice were immunised with subcutaneous injection of (a) 1x106 BMDC transduced with EGFP lentivirus and (b) 1x106 untreated BMDC. Splenocytes from the mice were isolated and stimulated with non-proliferating (mitomycin C treated) A20/EGFP target cell line in a 5 day culture with a 10:1 stimulator:responder ratio in the presence 10U/mL recombinant mouse IL2. On day 5 effector cells were recovered and rechallenged in a serial dilution with 3H thymidine labelled A20 (black, closed squares), A20/EGFP (green, closed triangles) or A20/EGFP/Y200T (red, closed circles) target cells. Percentage specific cell lysis was calculated compared to target cell control. Mean of three separate replicates (±SEM).

147

4.3.7. Generation of an immune response to EGFP by immunisation with A20 cell line expressing EGFP.

Given the problems with using BMDC, including the transfer of other antigen and the fact that for each immunisation a new batch of BMDC had to be prepared, alternative immunisation protocols were explored. The A20/EGFP cell line that had been developed for use as targets in the CTL assay has been used previously in vivo in BALB/c strain mice in order to visualise cells (Stripecke et al [69]) and more recently Bosilic et al [60] have both reported in vivo and in vitro cytotoxicity of

EGFP labelled tumour cells in BALB/c mice. In order to replicate these methods the

A20/EGFP cell line was used as an immunogen. Due to constraints of the animal licence with regards to permissible adverse effects, the A20 cells were first treated with mitomycin C to arrest proliferation and prevent tumorigenicity, whilst allowing continued presentation of EGFP antigen by the MHC system. Mice were immunised subcutaneously with 1x106 attenuated A20/EGFP cells with a repeat dose at day 7.

At day 14 mice were sacrificed and splenocytes were challenged in vitro with

A20/EGFP cells in a CTL assay. Figure 4.14 (a) demonstrates that as with the

BMDC immunisation, cell killing was observed across all A20 cell types and was not specific to EGFP expression. Again, maximal lysis was observed with the highest effector:target cell ratio, and ~40% lysis was achieved. It was again assumed that A20 cells, which have the cellular machinery for antigen presentation, are transferring and presenting a culture medium related antigen, as previous studies

(figure 4.5) had shown no evidence of background lysis or NK cell killing of the

A20 cell line with syngeneic effectors in a CTL assay.

In an attempt to reduce the potential effect of culture medium components acting as antigens, A20 cells were incubated in PBS during mitomycin C treatment and

148 washed extensively 4 times in PBS prior to immunisation. The A20/EGFP cell immunisation and the CTL effector expansion protocol was repeated as before.

Figure 4.14 (b) shows that there was no net killing of any of the three A20 target cells, demonstrating that the washing procedure had removed the non-specific killing effect, however, no specific A20/EGFP killing was shown. Indeed, particularly for the A20/Y200T target cells, there was an apparent increase in total label incorporated in the target cells in the presence of the effectors, particularly at high effector: target ratios although such was very variable. To confirm the presence of a culture medium related immune response following immunisation with A20 cells, serum samples from mice immunised with A20/EGFP cells with and without extensive prior washing were analysed for IgG antibody responses to FBS, as FBS represents the most concentrated proteins found in the tissue culture medium. Figure

4.15 indicates that there was indeed a significant antibody response to FBS generated in the mice immunised with cells from tissue culture in the absence of extensive washing, versus the background response recorded in naïve (untreated) mice.

Extensive washing of the cells prior to immunisation markedly reduced the antibody response to FBS but it was not completely eliminated.

149

(a) 6 0 A 2 0

A 2 0 /E G F P s

i 4 0 s

y A 2 0 /E G F P /Y 2 0 0 T

i f

i 2 0

p S

0 %

-2 0

1 0 0 1 0 1 E ffe c to r : T a rg e t R a tio

(b)

1 0 A 2 0

s 0 A 2 0 /E G F P

i s

y A 2 0 /E G F P /Y 2 0 0 T L

-1 0

i c

e -2 0

% -3 0

-4 0

1 0 0 1 0 1 E ffe c to r : T a rg e t R a tio

Figure 4.14 Cytotoxic response to A20 cell lines in mice immunised with A20 cells expressing EGFP. A BALB/c mouse was immunised by subcutaneous injection of (a) 1x106 A20/EGFP cells and (b) 1x106 A20/EGFP cells washed 4 times prior to immunisation to reduce carryover of foetal bovineserum from culture medium. Splenocytes from the mice were isolated and sensitised with non-proliferating A20/EGFP target cell line in a 5 day culture with a 10:1 sensitiser:responder cell ratio in the presence 10U/mL recombinant mouse IL2. (a) On day 5 effector cells were recovered and rechallenged in a serial dilution with 3H thymidine labelled A20 (black, closed squares), A20/EGFP (green, closed triangles) or A20/EGFP/Y200T (red, closed circles) target cells. % specific cell lysis was calculated by comparison to a target cell alone control. Mean of three separate wells (±SEM) shown.

150

2 .0

A 2 0 /E G F P m in . w a s h 1 .5

A 2 0 /E G F P m a x . w a s h

m n

0 N aive 5

4 1 .0

. O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 S e ru m D ilu tio n

Figure 4.15 Washing cells prior to immunisation reduces FBS related IgG response. BALB/c mice were immunised intraperitoneally twice, 7 days apart with either 1x106 A20/EGFP cells minimally washed once with PBS (closed diamond) or A20/EGFP cells washed four times in PBS (closed triangle) prior to immunisation, naive mouse serum (open circles). Serum was isolated on day 14 and IgG response to FBS was measured by ELISA. Data represent a single mouse per treatment group.

151

1 .5 rE G F P A 2 0 A 2 0 /E G F P

m 1 .0

n B M D C /E G F P 0

5 N M S

. D

. 0 .5 O

0 .0 1 0 1 0 0 1 0 0 0 Ig G T itr e

Figure 4.16 Serum IgG response to EGFP in mice immunised with EGFP expressing cells from tissue culture. BALB/c mice were immunised twice intraperitoneally, 7 days apart, with either 250µg recombinant EGFP (black closed symbols), 1x106 A20 cells (red symbols), 1x106 x A20/EGFP (green symbols), or subcutaneously with 1x106 BMDC/EGFP (blue symbols). Naive mouse serum (open circles) negative control is also displayed. Serum was isolated on day 14 and IgG response to EGFP was measured by ELISA. Data represent a single mouse per treatment group.

152

Although the CTL assays did not indicate an EGFP related cytotoxic response, serum samples from mice immunised in these EGFP cell studies were screened for a humoral anti-EGFP IgG response. Figure 4.16 shows that a specific EGFP antibody response was produced in mice immunised with A20/EGFP cells but not with the

A20 control cells, with a similar level of antibody induced to that observed for the control serum derived from mice immunised with the recombinant protein. No anti-

EGFP IgG was detected in mice treated with BMDC/EGFP from the previous study

(figure 4.12) where levels were identical to naïve mouse serum.

Production of EGFP antibody by immunisation with A20/EGFP cells demonstrated that an antigen specific immune response can be elicited by this method and thus a final study was designed.

153

To test immunogenicity of the EGFP/Y200T construct, groups of mice (n=3) were immunised with 1x107 A20, A20/EGFP or A20/EGFP/Y200T cells injected intraperitoneally, with a repeat dose at day 7 and sacrificed on day 10, as per Tian et al [70]. Splenocytes were challenged with A20/EGFP cells in vitro, mirroring the stimulation phase of the CTL assay and culture supernatants were analysed for indicators of cytotoxicity after 5 days, IFNγ and granzyme B.

In figure 4.17(a) there was no significant change in IFNγ release above the effector control amongst all the three treatment groups, between 1 and 2 ng/mL. In figure

4.17(b) although there was considerable inter-animal variation, levels of granzyme B were somewhat elevated in the CTL reactions compared to previous studies with effectors from naive syngeneic mice (Figure 4.6b). Challenge of effectors from all treatment groups with the A20/EGFP cells in vitro tended to result in higher levels of secreted granzyme B (30-35ng/mL) compared to the same effectors challenged with the A20 or A20/EGFP/Y200T (20-23ng/mL) although this was not statistically significant. The elevated levels of granzyme B in all treatment groups indicates that there is a cytotoxic response, but that this is not EGFP specific and could be perhaps targeted to a tumour related antigen expressed by the A20 cell line, enhanced by in vivo priming.

154

(a)

4 G ro u p # 1 : A 2 0 G ro u p # 2 : A 2 0 /E G F P 3

G ro u p # 3 : A 2 0 /E G F P /Y 2 0 0 T

L m

/ 2

g n

. 0 P T . 0 P T . 0 P T ) t 2 0 t 2 0 t 2 0 0 n F n F n F 2 o A G 0 o A G 0 o A G 0 2 2 2 (A C /E Y C /E Y C /E Y r / r / r / . 0 P 0 P 0 P t to 2 to 2 to 2 n A F A F A F o c G c G c G fe fe fe C f /E f /E f /E t E 0 E 0 E 0 e g 2 2 2 r A A A a T

(b) 6 0 G ro u p # 1 : A 2 0 G ro u p # 2 : A 2 0 /E G F P

4 0 G ro u p # 3 : A 2 0 /E G F P /Y 2 0 0 T

g n 2 0

Figure 4.17 Secretion of cytotoxicity markers IFNγ and Granzyme B in mice immunised with A20/EGFP. BALB/c strain mice were immunised intraperitoneally twice, 7 days apart, with either 1x107 A20 (group#1), A20/EGFP (group#2) and A20/EGFP/Y200T (group#3) cells. On day 10, splenocytes from the mice were sensitised with the non- proliferating A20, A20/EGFP or A20/EGFP/Y200T target cell line in a 5 day culture with a 10:1 sensitiser:responder ratio in the presence 10U/mL recombinant mouse IL2. Supernatants were analysed by sandwich ELISA for secretion of (a) mouse IFNγ and (b) mouse granzyme B. Effector and target cells alone were included as controls (cont). Mean of three mice per group shown (±SEM). The statistical significance of differences between treatment groups were assessed by one-way ANOVA but did not reach significance (at the level of p<0.05).

155

Anti-EGFP antibody responses from serum derived from the same animals were assessed by ELISA (Figure 4.18a). In Group#2, immunised with A20/EGFP cells, 2 of 3 mice showed anti-EGFP response with serum titres of 1 in 512, 1 in 256 and 1 in 32 (Figure 4.18b). None of the mice in group#1, A20 immunised control or group

#3, immunised with A20/EGFP/Y200T,showed an anti-EGFP antibody response above background levels. Anti-EGFP antibody specificity was also checked against

Y200T recombinant protein (Figure 4.18b). Antibody titres showed near identical cross reactivity between the parent EGFP and Y200T variant. The antibody responses were then isotyped. Figure 4.19 shows that antibodies generated were mainly IgG1 (a) subclass with a minor IgG2a component (b) indicating a preferential

Th2 response.

156

(a)

rE G F P N aive 1 .5 A 2 0 M o u s e # 1 A 2 0 M o u s e # 2 A 2 0 M o u s e # 3

1 .0 m

n A 2 0 /E G F P M o u s e # 1 0

5 A 2 0 /E G F P M o u s e # 2

D A 2 0 /E G F P M o u s e # 3

O 0 .5 A 2 0 /E G F P /Y 2 0 0 T M o u s e # 1 A 2 0 /E G F P /Y 2 0 0 T M o u s e # 2

0 .0 A 2 0 /E G F P /Y 2 0 0 T M o u s e # 3 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 S e ru m D ilu tio n

(b) 1 0 0 0 0 E G F P

E G F P /Y 2 0 0 T e

r 1 0 0 0

b A

1 0 0

P 0 P T F 2 F 0 A 0 G . G 2 E E r 1 / Y # 0 / p 2 P u A F o r . G 2 /E G # 0 p 2 u o A r . 3 G # p u o r G

Figure 4.18 IgG response to EGFP in mice immunised with A20 cells expressing EGFP and EGFP/Y200T. (a) BALB/c mice were immunised intraperitoneally with either 1x107 A20 (blue symbols), A20/EGFP (red symbols), A20/EGFP/Y200T cells (green symbols) at day 0 and day 7 (n=3 per group). Serum was isolated 10 days after first immunisation and IgG response to EGFP protein was measured by ELISA. Serum samples from a mouse immunised with 250µg/mL recombinant EGFP protein (rEGFP, black closed circle) and naive mouse serum (black line) were included as controls. (b) Anti-EGFP antibody (Ab) titre to EGFP (closed green symbol) or EGFP/Y200T (closed red symbol) was determined as the serum dilution that gave an absorbance reading of three times background level of naive serum. The statistical significance of differences between treatment groups was assessed by unpaired Student’s t-test but did not reach significance (at the level of p<0.05).

157

(a) 2 .5 rE G F P G ro u p # 1 A 2 0 P o o l 2 .0

G ro u p # 2 A 2 0 /E G F P M o u s e # 1 m

n 1 .5 G ro u p # 2 A 2 0 /E G F P M o u s e # 2 0

5 G ro u p # 2 A 2 0 /E G F P M o u s e # 3

1 .0

D G ro u p # 3 A 2 0 /E G F P /Y 2 0 0 T P o o l O

0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 S e ru m D ilu tio n

(b) 2 .0 rE G F P G ro u p # 1 A 2 0 P o o l

1 .5 G ro u p # 2 A 2 0 /E G F P M o u s e # 1 m

n G ro u p # 2 A 2 0 /E G F P M o u s e # 2 0

5 1 .0 G ro u p # 2 A 2 0 /E G F P M o u s e # 3

D G ro u p # 3 A 2 0 /E G F P /Y 2 0 0 T P o o l O 0 .5

0 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 S e ru m D ilu tio n

Figure 4.19 IgG1 and IgG2a response to EGFP in mice immunised with A20 cells expressing EGFP and EGFP/Y200T. BALB/c mice were immunised intraperitoneally with either 1x107 A20 (blue symbols), A20/EGFP (red symbols), A20/EGFP/Y200T cells (green symbols) at day 0 and day 7 (n=3 per group) or 250µg/mL recombinant EGFP protein (rEGFP, black closed circle) and naive mouse serum (black line). Serum was isolated 10 days after first immunisation. Serum for group #1 and group #3 was pooled. (a) IgG1 and (b) IgG2a response to EGFP protein was measured by ELISA.

158

4.4. Summary

A thorough investigation to compare the immunogenicity of both EGFP and the

EGFP/Y200T variant in BALB/c mice was carried out. A CTL assay was developed employing lysis of a radiolabelled target cell line and quantification of markers of cytotoxicity as the endpoints. Several strategies were used with the aim to measure a

CD8+ response towards EGFP by immunisation with recombinant protein, antigen presenting cells and tumour cells expressing EGFP. However specific cytotoxicity towards EGFP cells by these methods was not detected in vitro.

Although cytotoxicity was not measurable, immunisation with the EGFP expressing cell line demonstrated the ability to generate a strong anti-EGFP IgG antibody response in vivo. This response was equivalent to a humoral response when immunised with recombinant EGFP protein. Encouragingly when mice were immunised with the EGFP/Y200T variant cell line, this did not raise an antibody response above the background signal. This indicates that a humoral immune response towards EGFP cells has been successfully abrogated by the Y200T mutation.

159

Chapter 5

5. Discussion

5.1. Prediction and assessment of MHC class I epitopes in EGFP.

The computational prediction of MHC class I and II epitopes is a valuable tool in understanding immunology. The discovery that for a given MHC allele, short peptide epitopes were bound based on a particular sequence motif provided scope to predict epitopes within a given protein sequence [78]. The canonical amino acids that form these motifs within the epitopes generate the dominant interactions to the

MHC, and are therefore termed anchor site residues. Structural studies have shown the key role these anchor site side chains play in the interaction with the MHC [79].

By implementing the existence of these motifs, several epitope prediction algorithms have been designed which are trained on sequence databases of peptides eluted from

MHC and in vitro MHC-peptide binding experiments. However, prediction of an immunodominant epitope cannot be 100% accurate as the multiple stages of processing, display and recognition lead to the attrition of potential epitopes. It has been stated that for a given antigen as few as 1 in 2000 potential epitope peptides may be responsible for an immunogenic response [114].

For the MHC class I display of a foreign antigen, firstly the expression level and abundance of protein antigen in the cytoplasm will have a direct influence on the frequency of epitopes being presented [114]. The protein must undergo proteolysis in the proteasome resulting in peptides of adequate length for antigen processing, too short peptides (<7 amino acids) will not be further processed as antigens [115]. The next stage is delivery of the peptide to the ER by the TAP transporter. During recruitment onto the TAP transporter longer peptides are further cleaved by aminopeptidases [75] and the TAP transporter itself has specificity for hydrophobic

160

C-terminal amino acids on the peptides [114]. The epitope peptide must then be stably recruited onto the appropriate MHC. In humans, each individual possesses 6

MHC class I alleles, of over 8000 different alleles within the population, each with a particular epitope specificity that must be characterised to accurately predict sequence specificity [75]. However, for model mouse strains, through interbreeding the diversity of MHC class I is restricted and the alleles are well characterised;

BALB/c with 3 alleles and C57BL/6 with 2 (Table 1.2).

MHC class I can accommodate sequences between 8-10 amino acids in length, with a preference for 9 amino acids to fit into the binding cleft [78]. MHC class II specificity is even more diverse, accommodating 13-18 amino acids with more anchor points making binding more complex to predict [75]. The MHC structure without peptide is inherently unstable and it is the affinity of the epitope peptide complex that maintains a stable structure [116]. This binding is driven by the anchor site residues in the epitope sequence which have high structural complimentarity to the MHC cleft [73]. The epitope sequence and occurrence of suitable anchor site residues therefore directs the affinity. Affinity values of IC50 above 500nM have been shown to not provide sufficiently stable binding and are therefore unlikely to be presented [100]. When a single MHC complex reaches the cell membrane, it will be one of approximately 1x105 MHC complexes at the surface, including MHC class I expressing ‘self’ peptides from endogenous protein [75]. To allow detection of the foreign antigen, the immune system must also have a suitable T cell repertoire primed to engage with the epitope [100]. Of the few epitopes that pass though the complete process it is often found that a single epitope dominates the T cell response and are hence termed immunodominant [114]. The existence of a single, high affinity, dominant epitope will drive the initial CD8 response, leading to bias

161 towards the proliferation of T cell clones specific to the epitope. This dominancy overtakes the response to other processed epitopes, with lower affinity, termed subdominant epitopes, which will have a reduced immunogenic potential [114].

Even though the predictive tools mentioned have proven as effective methods in predicting epitope processing and MHC class I binding they can only be used alongside supporting experimental methods to determine the true in vivo immunogenicity of epitopes. These methods can be divided into either in vitro MHC binding assays to confirm binding predictions and more in depth in vivo studies to confirm immunogenic potential. Stabilisation of MHC class I by synthetic peptide epitopes can be measured using TAP deficient cells lines [63, 64]. Epitope affinity can be accurately measured using synthetic peptide libraries to bind recombinant

MHC class I and stability analysed using biophysical techniques such as equilibrium dialysis [85, 101] and circular dichroism [117]. These purely in vitro methods can then be followed up with in vivo investigations with the immunisation of relevant animal models with parent protein or epitope peptides. Epitope peptides can then be screened to check for T cell stimulation (measuring relevant cytokine production though ELISA or ELISpot e.g. IFNγ) [95] or induction of CTL activity ex vivo [63].

In this manner, Gambotto et al [63] predicted the potential epitopes of EGFP in

BALB/C mice using a predictive algorithm [85] and crucially confirmed a single,

9-mer, H2-Kd restricted immunodominant epitope; HYLSTQSAL (199-207), through MHC stabilisation, T cell stimulation and CTL killing assays. No H2-Dd epitopes were considered due to low predicted binding. Subsequently Han et al [64] confirmed the H2-Db restricted immundominant epitope DTLVNRIEL (117-125) in

C57BL/6 mice through a similar strategy.

162

The discovery of a single dominant epitope in EGFP in the two model strains of mice presents an opportunity to reduce immunogenicity through mutation of the anchor site residues, as routinely conducted for MHC class II epitopes in therapeutic proteins [93, 95]. In this study, updated epitope prediction algorithms were applied to analyse the EGFP sequence for potential epitopes. The results for the dominant epitopes agreed with the original studies, identifying HYLSTQSAL and

DTLVNRIEL as high affinity epitopes [63, 64]. Notably, for BALB/c, Gambotto et al [63] did not screen for epitopes restricted to the H2Ld allele. This is plausibly due to either oversight or restricted epitope data on the allele at that time. Screening in this study presented a single possible H2Ld epitope TPIGDGPVL, albeit with an intermediate affinity (IC50 = 700nM, table 2.1), and it was noted whether this might act as a possible subdominant epitope to HYLSTQSAL.

To assess the potential for ‘deimmunisation’ of the two dominant epitopes, the anchors site residues can be identified from the conserved motifs of known epitope peptides [85]. To ascertain the predicted effect of mutation at the anchor sites in the epitopes it was deemed useful to analyse mutations using the same eptiope prediction algorithm. Alanine was substituted at each amino acid position in the sequence. In the H2-Kd restricted epitope HYLSTQSAL, the anchor site Tyr200 at position 2 was the most sensitive to mutation, resulting in a large drop in affinity from 6nM to

1020nM, a 170 fold change. In the H2Db restricted epitope DTLVNRIEL, Asn121 at position 5 was the most sensitive to mutation, resulting in a drop from 730nM to

25mM which is beyond the known affinity for epitope peptides.

Alignment of the epitope sequences with known epitope structures (2.3.3) confirmed that Tyr 200 and Asn 121 would play major roles in the interaction with the MHC

163 molecule. Both Tyr200 and Asn 121 were therefore selected as suitable anchor sites for mutation to disrupt epitope binding.

GFP has been extensively engineered with numerous gain of function mutations employed by random or directed mutagenesis approaches to enhance spectral properties (Figure 1.9) [14, 118], none of which have involved Tyr200 or Asn121.

Although the structure of GFP seems to be stable enough to accommodate mutation at various sites, to ascertain the potential effects of mutations in the epitope on the parent protein EGFP the locations were identified (2.3.4). The Tyr200 side chain is at the surface on a β-strand of the GFP barrel structure and was identified as a suitable site for mutation. Asn121 however occupies a more sensitive site in proximity to the chromphore where non-conservative mutation was predicted to possibly affect folding and fluorescence function of EGFP.

One peptide identified for H2-Db TLTYGVQCF (63-71) contains the EGFP chromphore sequence (TYG) and although predicted affinity was negligible

(2.4mM) it is unknown whether the modified peptide would be accommodated by any MHC allele at all, given the unique post translation modification.

5.2. Recombinant Expression of EGFP in E.coli and library screening.

EGFP proved to be an extremely useful protein in this study as a model protein for transgene MHC class I deimmunisation. It expresses well in E.Coli with relatively high yields of soluble protein (~200mg/L of culture) and can be purified efficiently using a fusion tag for characterisation studies. The mutagenesis and screening system worked well due to EGFP easily detectable fluorescence property. Throughput of screening mutations could be increased and made quantitative in future studies using a multiwell plate format and fluoresence plate reader. Indeed Cantor et al [95] used

164

GFP as a marker protein in a high throughput screen for the deimmunsaiton of MHC class II epitopes in a bacterial enzyme, L-asparginase. To encompass all possible amino acids at the sites of mutation, Cantor et al [95] also employed saturation mutagenesis which was adopted in the strategy used in this study. This increases the sequence space that is explored during one mutagenesis PCR reaction rather than having to perform individual reactions of directed mutation to each of the twenty amino acids. During screening of the library only fluorescent, functional mutations are selected, which are then subsequently determined by DNA sequencing, leaving non functional mutations aside.

5.3. Mutagenesis of anchor sites.

Mutagenesis of the Tyr200 H2-Kd restricted anchor site residue successfully generated a library of fluorescent variants with reduced MHC binding potential.

Three variants had predicted IC50 values sufficiently above the 500nM cut off for significant affinity; Y200R, Y200E and Y200T. All three proteins demonstrated comparable fluorescence function to the parent protein. Y200T was selected as the favoured candidate due to the highest impact on epitope binding (IC50=990nM).

Mutagenesis of the Asn121 H2-Db restricted anchor site residue did not yield a diverse library of fluorescent variants. The majority of functional mutations isolated were silent, matching the wildtype sequence. Two fluorescent clones were identified;

N121S from the mutagenesis library and N121A was generated by directed mutagenesis. Bacterial colonies expressing the proteins were visibly less fluorescent than the wild type. It is likely that due to the proximity of Asn121 next to the chromophore (2.3.4), mutation of the side chain has resulted in disruption of the

165 local hydrogen bonding network which affects the efficiency of protein folding and chromophore function. The fluorescence spectra of N121A demonstrated a secondary absorption peak at approximately 390nm. This is similar to the absorbtion the original wild type GFP (figure 1.5), indicating that the chromophore has shifted to a charged (anionic) state due to a disruption in the hydrogen bonding network which results in a reduction in fluorescence efficiency.

Asn121 therefore appears to be sensitive to mutation. As an alternative to disruption of the dominant anchor residue N at position 5 for H2-Db , it is plausible that other anchor sites could be mutated accordingly to disrupt MHC binding. Table 2.3b illustrates that alanine substitution at positions 3 and C-terminal position 9 in the epitope reduces the IC50 values significantly from 730nM to 6300nM and 6700nM respectively. Position 3 is occupied by Leu119 and the C-terminal position 9 is

Leu125 and neither residues are in contact with the chromophore making them better candidates for mutation that Asn121. Fahnestock et al [101] demonstrated that non- conservative mutation (L>D) at the C-terminal anchor site in a H2-Kd restricted flu epitope was sufficient to completely disrupt binding to the same degree as targeting the main anchor residue (Y at position 2).

5.4. Development of a CTL assay to assess EGFP mediated cytotoxicity

To demonstrate cytotoxicity towards the EGFP expressing cells in vitro, a cytotoxic

T cell lymphocyte assay was developed. Design of the assay was complex, with several variables to optimise. Key to the CTL assay is optimisation of a suitable responder:sensitiser ratio and cell densities to efficiently generate effectors in vitro

166 during the stimulation phase. Depending on the cell types used and the antigens involved different ratios are required for optimal responses [108].

The method was successfully demonstrated using non histocompatible (allogenic) cell types to measure cytotoxic cell killing. To act as antigen target cells in the assay, stable cell lines expressing EGFP or EGFP/Y200T were generated by transduction with lentivirus with ≥ 90% expression levels.

To begin to examine immunogenicity of EGFP in BALB/c mice it was shown that

EGFP has the potential to generate a humoral IgG antibody response when administered intraperitoneally, without the requirement for adjuvant. Immunisation with the Y200T variant, in which an MHC class I immunodominant epitope anchor site residue has been mutated, did not significantly change the specificity of the antibody response to either parent EGFP or EGFP/Y200T. The similarity in humoral response and cross-reactivity between the parent and variant are not surprising.

Because a B cell response to antigen is both adaptive and polyclonal in nature a single amino acid change of an antigen’s surface residue (such as Y200) is unlikely to impact on antibody generation and recognition significantly. Changes in multiple

MHC class II epitope anchor sites, however, have been shown to reduce CD4 help and therefore limit downstream antibody generation and are utilised as such for dampening immunogenicity of administered therapeutic proteins in human patients

[12-14]. It is clear then that the single mutation, Y200T, does not affect a CD4 response to recombinant EGFP.

Initially, generation of a cytotoxic response towards EGFP was attempted in vivo by immunisation with heat denatured recombinant protein. Although a humoral IgG response was detected to denatured protein, no specific cytotoxic activity to

EGFP/cells was observed.

167

The focus was then shifted to delivery of antigen by cell mediated methods.

Professional antigen presenting BMDCs isolated from mice were transduced with lentivirus to express EGFP. Efficiency of transduction was low, with relatively high amounts of virus titre (1x108 virus particles per 2x106 BMDC) required to obtain a minority population (20%) of EGFP expressing cells. Mice were then immunised with BMDC/EGFP, but this method did not generate any antibody response or specific cytotoxic activity to EGFP cells in vitro, possibly due to the limited production and isolation of EGFP positive BMDCs. In contrast to the study here,

Stripecke et al [69] successfully demonstrated an EGFP specific CTL response by immunisation with BMDC/EGFP in BALB/c mice. However, Annoni et al [71] demonstrated tolerance to EGFP could be induced by prior immunisation with APCs from EGFP+ transgenic BALB/c mice, without the requirement for myeloblative conditioning. Because of the potential of DCs to induce tolerance rather than eliciting an immune response, caution should be raised when using DCs in adoptive transfer. This may have been the reason no CTL or humoral response towards EGFP was observed through immunisation with EGFP+ DCs in this study.

The next stage involved using the attenuated (non-proliferating) A20/EGFP cell line as an immunogen. Mice were immunised with mitomycin C treated A20 cells, however cytotoxicity assays were confounded by the carryover of tissue culture components acting as antigens in vivo and may have caused unspecific cell killing during the CTL assay in vitro. Indeed immunogenicity towards culture components during adoptive cell therapies has been proven to be an issue in the clinical setting

[119, 120].

Improved cell washing prior to immunisation was then employed to reduce the presence of contaminants such as FBS and reduce this effect. Despite this it was not

168 possible to directly demonstrate EGFP related cytotoxicity through CTL assay or cytotoxicity marker ELISA. Stripecke et al [69] and Tian et al [70] have both shown in vitro cytotoxic responses to breast carcinoma BM185/EGFP cells in BALB/c by the 51Cr release based CTL method. However, Stripecke et al also co-expressed

CD80 in the BM185 cell line used during in vitro stimulation which may have served to improve effector generation. Indeed the tumour cell type used may also be key in the level of response to cellular antigen response seen, especially as it is common for tumours to demonstrate immunosuppressive behaviour or locate to particular immune privileged sites [121].

5.4.1. EGFP/Y200T eliminates a humoral antibody response to EGFP

Although the CTL assay did not detect cell killing by immunisation with A20/EGFP cells, it was possible to detect an antibody response to EGFP by ELISA. Measuring serum antibody provided a straight forward method to assess immunogenicity of the cellular expressed EGFP and EGFP/Y200T constructs.

In a more in depth study, mice were immunised with A20/EGFP and two out of three mice showed an IgG response to EGFP (Figure 4.18). The antibody isotype generated was predominantly IgG1 in nature, indicating a preferential Th2 response, whereas immunisation with recombinant protein alone demonstrated a combination of IgG1 and IgG2a (Figure 4.2). IgG2a has been shown to be elevated in bacterial infection [122] and elevation of the isotype could be contributed to the presence of residual E.coli related contaminants in the recombinant EGFP protein used for immunisation.

In the group treated with A20/EGFP/Y200T cells, none of the mice developed detectable anti-EGFP IgG antibody. This indicates that the Y200T mutation has

169 served to reduce humoral immunogenicity of EGFP when expressed by transplanted cells.

One concern is the A20/EGFP/Y200T cell line demonstrated fivefold lower fluorescence intensity compared to A20/EGFP (figure 4.4). From expression in

E.coli the purified protein demonstrated highly similar fluorescence characteristics to the parent EGFP protein. This loss in fluoresence may have been due to decreased expression levels through the selection of a non optimal codon for threonine (ACG) in the mutation library. A reduction in EGFP expression levels could have impacted the frequency of epitiope presentation and the strength of the immune response.

Eixarch et al [62] demonstrated that decreased EGFP expression in bone marrow cells can reduce both humoral and cell mediated immunogenicity to EGFP in

C57BL/6 mice. Over 90% of EGFP/Y200T cells here were detectable by fluorescence above background and the Y200T protein would have been abundant, under constitutive expression by the EL4α promoter within the high density (1x107) of cells introduced. The codon for threonine in the Y200T mutation, ACG is used at frequency of 10% in mammals [123]. Optimisation of this to ACC, which has a usage frequency of 35% may help to improve expression.

As the A20/EGFP/Y200T strain yields approximately 20% of the fully folded fluorescent protein compared to A20/EGFP the difference in protein expression can be controlled for through the cell number immunised. Initially, in figure 4.16, a study was performed where mice were immunised with 1x106 A20/EGFP cells to successfully generate an antibody response. In a subsequent study, figure 4.18, mice were immunised with 1x107 A20/EGFP/Y200T cells and no immune response was observed. Even if there is fivefold less EGFP protein with Y200T, immunisation

170 with tenfold the number of cells does still not raise an antibody response in comparison to a significant response with the A20/EGFP strain. Although the Y200T construct without codon optimisation is a ‘dimmer’ version of EGFP, it would still be a useable fluorescent marker for cell sorting applications and imaging as brightness level still exceed many other fluorescent proteins available (Table 1.1).

The complete absence of an IgG response to A20/EGFP/Y200T variant suggests a

CD4 T-helper type 2 response has been reduced and therefore a B cell response to produce antibody has not been enlisted. Other authors have reported humoral IgG response to EGFP in EGFP labelled tumours in BALB/c mice in conjunction with cytotoxic responses to EGFP cells [60, 107]. These investigations suggest that non- secretory cellular antigens, such as EGFP, can enlist both helper T cell and B cell responses alongside a cytotoxic CD8 immune response. CD4 cells have been shown to provide an important role in maintaining a CD8 response (Th1) and activating an antibody response (Th2) in tumour immunity [124].

It is presumed therefore that disabling the dominant MHC class I epitope here has reduced or eliminated CD8 cytotoxicity in the first instance, preventing cell lysis and the release of EGFP protein into circulation for APC uptake and downstream antibody generation. This requires that if the dominant EGFP epitope is disabled it is not substituted by a subdominant epitope. In support of this Gambotto et al [63] have shown that none of the other predicted H2-Kd epitope peptides were able stimulate T cells or CTL in mice sensitised to EGFP, which suggests removal of the immunodominat epitope would be sufficient to completely abrogate a CTL response.

The schematic in Figure 5.1 summarises the proposed model.

171

In agreement to the suggested model, deimmunsiation of H2-Kd dominant epitopes has been shown to eliminate CTL effect towards herpesvirus [97] and malarial antigens in BALB/c mice [96], without subdominant epitopes to H2-Kd or other class

I alleles (H2-Db, H2Ld) generating a cytotoxic response.

Because inbred BALB/c mice have a limited MHC class I repertoire, with only 3 alleles, for a given protein antigen there will only be a very limited number of epitopes that will have suitable MHC binding characteristics with immunogenic potential [114]. If the dominant epitope is removed by mutation then there is only limited scope for another suitable epitope to intervene and gain dominancy.

In other inbred strains of mice such as C57BL/6 mice the immunogenicity towards

EGFP expressing cells is controversial. Although Stripecke et al [69] and Han et al

[64] have characterised EGFP related immunogenicity in this strain, more recently

Skelton et al [125], Denaro et al [126]and Tian et al [70] were unable to measure a cytotoxic response to a syngenic EGFP transduced T cell lymphoma cell line (EL-4).

This could be due to the characteristics of the tumour used and difference in immunisation protocol, but there is mounting evidence to suggest EGFP has a weaker propensity to elicit an immune response in C57BL/6. Interestingly, this low immunogenic potential coincides with the observation that the dominant H2-Db epitope affinity is weak with an IC50 of 700nM (Table 2.2) and no other potential epitopes were identified in H2-Db or H2-Kb.

172

Y (e) Y Y B Cell

(d) CD4 cell MHC II

MHC II

(a) APC

MHC II CD8 cell

(f) (c) Epitope Removal MHC I

(b)

EGFP+ cell Cell killing

EGFP EGFP epitope Y Anti-EGFP IgG Figure 5.1 Proposed model for the elimination of an antibody response to EGFP by the removal of the immundominant MHC class I epitope. (a) A CD8 cell is activated by the immundominant EGFP epitope bound to MHC class I on a transplanted EGFP+ cell. (b) CD8 activation leads to a cytotoxic response and destruction of the cell. EGFP protein is released. (c) EGFP is taken up by antigen presenting cells (APC) and eptiope presented by MHC class II (d) CD4 cells with a cognate receptor to the EGFP epitope help to activate EGFP antigen specific B cells. (e) Activated B cells mature and proliferate leading to secretion of anti-EGFP antibody. (f) If the dominant MHC class I epitope is removed this will inhibit cell killing in the first instance and prevent the cascade of events (a-e) leading to the elimination of the humoral response towards EGFP.

173

This adds weight to the principle that a limited MHC repertoire reduces the potential immunogenicity of a given antigen. This can therefore be exploited to reduce immunogenicity toward antigens by removal of class I epitopes in inbred populations, such as model mouse strains. Successful removal of dominant MHC class I epitopes of a particular antigen in an outbred population would be more complex. As the number and combination of alleles is diverse this increases the number of potential epitopes for a given individual, such as in humans. A strategy to reduce immunogenicity of an antigen by MHC class I deimmunisation would have to focus on the most commonly occurring MHC alleles within a specific population, such as the established method applied for MHC class II epitope removal in therapeutic proteins [93, 95].

Gene therapy, the delivery of novel or replacement of dysfunctional genes with a therapeutic endpoint holds great promise in stably treating genetic disease, having overcome the initial safety barrier with viral transduction [127]. Presumably the frequency of ‘non self’ MHC class I epitopes will have to be considered in the protein products delivered by these methods to improve tolerance to transformed cells and prevent cell rejection.

174

5.4.2. Future Aims

The demonstration that a reduced immune response to EGFP in BALB/c mice, by a single mutation Y200T, is very promising. However due to time restrictions further studies could not be conducted and the limited number of animals used has limited statistical analysis of the significance. An appropriate extension to the investigation would be to continue the immunisation experiments with the A20 (EGFP and

EGFP/Y200T) cell line with an increased number of animals to allow for biological variation.

Due to restrictions under the animal license used in this study, only non- tumourigenic cells could be used in vivo. The tumour cells were rendered non- proliferating by treatment with mitomycin C prior to immunisation, viability would have been limited to a few days in vivo and this may have also reduced the potency of an EGFP related CTL response. Several authors studying EGFP immunogenicity of tumour cells have used proliferating tumour cells to demonstrate EGFP specific

CTL responses in vivo [60, 63, 64, 69, 70] and it would be useful to replicate this through collaboration with a group working with a suitable EGFP tumour model.

The deimmunisation of the H2-Kd epitope is BALB/c specific, so a control experiment conducted in an alternative strain would confirm whether an immune response is maintained with EGFP/Y200T, though use of C57BL/6 would have to be carefully considered as immunogenicity of EGFP in this strain is uncertain.

As EGFP is adaptable to protein fusions, it would be plausible to ‘knock in’ the

HYLSTQSAL epitope at the terminus of the EGFP/Y200T protein or as a mini gene on the lentivirus insert, to act as a reversion control. If the reduction in immune response is due to the loss of the Y200 anchor site specifically, reintroduction of the dominant epitope should engage an immunogenic response.

175

Though a large amount of optimisation was attempted the CTL assay was unable to demonstrate specific cell killing. Because GFP labelled cells can be easily detected by fluorescence in vitro, CTL methods using GFP labelled target cells have been developed as alternatives using a fluorescence plate reader or cytometry to quantify

GFP cell killing [128, 129]. Measuring GFP cell killing through these methods would be highly suited to this study and avoids the use radiolabels which can affect target cell viability.

Several spectral variants of EGFP exist; EYFP (enhanced yellow), ECFP (enhanced cyan) and EBFP (enhanced blue) [130]. It would be interesting to see if the Y200T mutation is compatible with these variants to broaden the colour palette for in vivo applications.

Immunogenicity of useful transgene products is not confined to EGFP. Both DsRed

(red fluorescent protein, [72, 73]) and luciferase [74] have been shown to be immunogenic in mouse models and the H2-Kd and H2-Db epitopes have been determined respectively. It would be of interest to see if anchor sites of these epitopes could be disrupted to improve their immunogenic profiles for in vivo use.

176

5.4.3. Concluding remarks.

A strategy for the deimmunisation of EGFP in BALB/c mice is presented here. The data herein suggests that an immune response to a transgenically expressed product can be reduced by mutation of anchor sites within the dominant MHC class I epitope. This is a ‘bottom up’ approach to reducing the immunogenicity of transgene products through protein engineering rather than immune suppression or other tolerance inducing regimens. After further investigation it is hoped a deimmunised

EGFP construct will be available for use in vivo. However caution should still be exercised in using the standard EGFP in vivo due to its immunogenic potential.

177

Chapter 6

6. References

1. Minderhout, V., Glowing Genes: A Revolution in Biotechnology (Zimmer,

Marc). Journal of Chemical Education, 2006. 83(2): p. 215-null.

2. Ehrenberg, M., Scientific Background on the Nobel Prize in Chemistry 2008;

The green fluorescent protein: discovery, expression and development. The

Royal Swedish Academy of Sciences, 2008.

3. Davenport, D. and J.A.C. Nicol, Luminescence in Hydromedusae.

Proceedings of the Royal Society of London. Series B - Biological Sciences,

1955. 144(916): p. 399-411.

4. Chalfie, M., Y. Tu, G. Euskirchen, W.W. Ward, and D.C. Prasher, Green

fluorescent protein as a marker for gene expression. Science, 1994.

263(5148): p. 802-805.

5. Matz, M.V., A.F. Fradkov, Y.A. Labas, A.P. Savitsky, A.G. Zaraisky, M.L.

Markelov, and S.A. Lukyanov, Fluorescent proteins from nonbioluminescent

Anthozoa species. Nat Biotech, 1999. 17(10): p. 969-973.

6. Science Direct URL:http://www.sciencedirect.com [2014].

7. Zimmer, M., GFP: from jellyfish to the Nobel prize and beyond. Chemical

Society Reviews, 2009. 38(10): p. 2823-2832.

8. Shimomura, O., F.H. Johnson, and Y. Saiga, Extraction, Purification and

Properties of Aequorin, a Bioluminescent Protein from the Luminous

Hydromedusan, Aequorea. Journal of Cellular and Comparative Physiology,

1962. 59(3): p. 223-239.

178

9. Morise, H., O. Shimomura, F.H. Johnson, and J. Winant, Intermolecular

energy transfer in the bioluminescent system of Aequorea. Biochemistry,

1974. 13(12): p. 2656-2662.

10. Science Blogs URL:

http://scienceblogs.com/purepedantry/2007/11/two_photon.php. [2011].

11. Shimomura, O., Structure of the chromophore of Aequorea green fluorescent

protein. FEBS Letters, 1979. 104(2): p. 220-222.

12. Prasher, D.C., V.K. Eckenrode, W.W. Ward, F.G. Prendergast, and M.J.

Cormier, Primary structure of the Aequorea victoria green-fluorescent

protein. Gene, 1992. 111(2): p. 229-233.

13. Wang, S. and T. Hazelrigg, Implications for bcd mRNA localization from

spatial distribution of exu protein in Drosophila oogenesis. Nature, 1994.

369(6479): p. 400-403.

14. Heim, R., D.C. Prasher, and R.Y. Tsien, Wavelength mutations and

posttranslational autoxidation of green fluorescent protein. Proceedings of

the National Academy of Sciences of the United States of America, 1994.

91(26): p. 12501-12504.

15. Baedeker, M. and G.E. Schulz, Autocatalytic Peptide Cyclization during

Chain Folding of Histidine Ammonia-Lyase. Structure, 2002. 10(1): p. 61-67.

16. Schwede, T.F., J. Retey, and G.E. Schulz, Crystal Structure of Histidine

Ammonia-Lyase Revealing a Novel Polypeptide Modification as the Catalytic

Electrophileâ€ ,â€¡. Biochemistry, 1999. 38(17): p. 5355-5361.

17. Cubitt, A.B., R. Heim, S.R. Adams, A.E. Boyd, L.A. Gross, and R.Y. Tsien,

Understanding, improving and using green fluorescent proteins. Trends in

Biochemical Sciences, 1995. 20(11): p. 448-455.

179

18. Brejc, K.a., T.K. Sixma, P.A. Kitts, S.R. Kain, R.Y. Tsien, M. Orm, and S.J.

Remington, Structural basis for dual excitation and photoisomerization of

the Aequorea victoria green fluorescent protein. Proceedings of the National

Academy of Sciences of the United States of America, 1997. 94(6): p. 2306-

2311.

19. Piston, D.W., Campbell, R.E., Day, R.N. and Davidson, M.W., URL:

http://zeiss-campus.magnet.fsu.edu/articles/probes/fpintroduction.html

[2014].

20. Ormö, M., A.B. Cubitt, K. Kallio, L.A. Gross, R.Y. Tsien, and S.J.

Remington, Crystal Structure of the Aequorea victoria Green Fluorescent

Protein. Science, 1996. 273(5280): p. 1392-1395.

21. Yang, F., L.G. Moss, and G.N. Phillips, The molecular structure of green

fluorescent protein. Nat Biotech, 1996. 14(10): p. 1246-1251.

22. Wachter, R.M., M.-A. Elsliger, K. Kallio, G.T. Hanson, and S.J. Remington,

Structural basis of spectral shifts in the yellow-emission variants of green

fluorescent protein. Structure, 1998. 6(10): p. 1267-1277.

23. Campbell, R.E., O. Tour, A.E. Palmer, P.A. Steinbach, G.S. Baird, D.A.

Zacharias, and R.Y. Tsien, A monomeric red fluorescent protein. Proceedings

of the National Academy of Sciences of the United States of America, 2002.

99(12): p. 7877-7882.

24. Shaner, N.C., R.E. Campbell, P.A. Steinbach, B.N.G. Giepmans, A.E.

Palmer, and R.Y. Tsien, Improved monomeric red, orange and yellow

fluorescent proteins derived from Discosoma sp. red fluorescent protein. Nat

Biotech, 2004. 22(12): p. 1567-1572.

180

25. Masuda, H., Y. Takenaka, A. Yamaguchi, S. Nishikawa, and H. Mizuno, A

novel yellowish-green fluorescent protein from the marine copepod,

Chiridius poppei, and its use as a reporter protein in HeLa cells. Gene, 2006.

372: p. 18-25.

26. Baumann, D., M. Cook, L. Ma, A. Mushegian, E. Sanders, J. Schwartz, and

C.R. Yu, A family of GFP-like proteins with different spectral properties in

lancelet Branchiostoma floridae. Biology Direct, 2008. 3(1): p. 28.

27. Piston, D.W., Patterson, G.H., Lippincott-Schwartz, J., Claxton, N.S. and

Davidson, M.W.,

URL:http://www.microscopyu.com/articles/livecellimaging/fpintro.html

[2012].

28. Ward, W.W. and S.H. Bokman, Reversible denaturation of Aequorea green-

fluorescent protein: physical separation and characterization of the

renatured protein. Biochemistry, 1982. 21(19): p. 4535-4540.

29. Day, R.N. and Davidson, M.W., URL: http://zeiss-

campus.magnet.fsu.edu/articles/probes/jellyfishfps.html [2012].

30. Cody, C.W., D.C. Prasher, W.M. Westler, F.G. Prendergast, and W.W.

Ward, Chemical structure of the hexapeptide chromophore of the Aequorea

green-fluorescent protein. Biochemistry, 1993. 32(5): p. 1212-1218.

31. Tsien, R.Y., THE GREEN FLUORESCENT PROTEIN. Annual Review of

Biochemistry, 1998. 67(1): p. 509-544.

32. Barondeau, D.P., C.D. Putnam, C.J. Kassmann, J.A. Tainer, and E.D.

Getzoff, Mechanism and energetics of green fluorescent protein

chromophore synthesis revealed by trapped intermediate structures.

181

Proceedings of the National Academy of Sciences of the United States of

America, 2003. 100(21): p. 12111-12116.

33. Shaner, N.C., G.H. Patterson, and M.W. Davidson, Advances in fluorescent

protein technology. J Cell Sci, 2007. 120(24): p. 4247-4260.

34. Yarbrough, D., R.M. Wachter, K. Kallio, M.V. Matz, and S.J. Remington,

Refined crystal structure of DsRed, a red fluorescent protein from coral, at

2.0-Ã… resolution. Proceedings of the National Academy of Sciences of the

United States of America, 2001. 98(2): p. 462-467.

35. Gross, L.A., G.S. Baird, R.C. Hoffman, K.K. Baldridge, and R.Y. Tsien, The

structure of the chromophore within DsRed, a red fluorescent protein from

coral. Proceedings of the National Academy of Sciences of the United States

of America, 2000. 97(22): p. 11990-11995.

36. Shaner, N.C., P.A. Steinbach, and R.Y. Tsien, A guide to choosing

fluorescent proteins. Nat Methods, 2005. 2: p. 905 - 909.

37. Labas, Y.A., N.G. Gurskaya, Y.G. Yanushevich, A.F. Fradkov, K.A.

Lukyanov, S.A. Lukyanov, and M.V. Matz, Diversity and evolution of the

green fluorescent protein family. Proceedings of the National Academy of

Sciences of the United States of America, 2002. 99(7): p. 4256-4261.

38. Prescott, M., M. Ling, T. Beddoe, A.J. Oakley, S. Dove, O. Hoegh-Guldberg,

R.J. Devenish, and J. Rossjohn, The 2.2 Å Crystal Structure of a Pocilloporin

Pigment Reveals a Nonplanar Chromophore Conformation. Structure, 2003.

11(3): p. 275-284.

39. Wahlfors, J., S. Loimas, T. Pasanen, and T. Hakkarainen, Green fluorescent

protein (GFP) fusion constructs in gene therapy research. Histochemistry

and Cell Biology, 2001. 115(1): p. 59-65.

182

40. Livet, J., T.A. Weissman, H. Kang, R.W. Draft, J. Lu, R.A. Bennis, J.R.

Sanes, and J.W. Lichtman, Transgenic strategies for combinatorial

expression of fluorescent proteins in the nervous system. Nature, 2007.

450(7166): p. 56-62.

41. Biondi, R.M., M. Veron, P.J. Baehler, and C.D. Reymond, Random insertion

of GFP into the cAMP-dependent protein kinase regulatory subunit from

Dictyostelium discoideum. Nucl. Acids Res., 1998. 26(21): p. 4946-4952.

42. Yu, J., J. Xiao, X. Ren, K. Lao, and X.S. Xie, Probing Gene Expression in

Live Cells, One Protein Molecule at a Time. Science, 2006. 311(5767): p.

1600-1603.

43. Miyawaki, A., J. Llopis, R. Heim, J.M. McCaffery, J.A. Adams, M. Ikura,

and R.Y. Tsien, Fluorescent indicators for Ca2+based on green fluorescent

proteins and calmodulin. Nature, 1997. 388(6645): p. 882-887.

44. Waldo, G.S., B.M. Standish, J. Berendzen, and T.C. Terwilliger, Rapid

protein-folding assay using green fluorescent protein. Nat Biotech, 1999.

17(7): p. 691-695.

45. Yang, M., E. Baranov, P. Jiang, F.-X. Sun, X.-M. Li, L. Li, S. Hasegawa, et

al., Whole-body optical imaging of green fluorescent protein-expressing

tumors and metastases. Proceedings of the National Academy of Sciences of

the United States of America, 2000. 97(3): p. 1206-1211.

46. Hoffman, R.M., Advantages of multi-color fluorescent proteins for whole-

body and in vivo cellular imaging. Journal of Biomedical Optics, 2005.

10(4): p. 041202-041202-10.

183

47. Hoffman, R.M. and T.A. Eric, In Vivo Cell Biology of Cancer Cells

Visualized with Fluorescent Proteins, in Current Topics in Developmental

Biology. 2005, Academic Press. p. 121-144.

48. Shu, X., A. Royant, M.Z. Lin, T.A. Aguilera, V. Lev-Ram, P.A. Steinbach,

and R.Y. Tsien, Mammalian Expression of Infrared Fluorescent Proteins

Engineered from a Bacterial Phytochrome. Science, 2009. 324(5928): p.

804-807.

49. Gines, T.B. and Davidson, M.W. URL: http://zeiss-

campus.magnet.fsu.edu/tutorials/spectralimaging/fretbiosensors/index.html

[2012].

50. Chudakov, D.M., S. Lukyanov, and K.A. Lukyanov, Fluorescent proteins as

a toolkit for in vivo imaging. Trends in Biotechnology, 2005. 23(12): p. 605-

613.

51. Murphy, D.B. and M.W. Davidson, Imaging Living Cells with the

Microscope, in Fundamentals of Light Microscopy and Electronic Imaging.

John Wiley & Sons, Inc. p. 357-387.

52. URL: http://www.patentstorm.us/patents/5972638.html.

53. Gong, Z., H. Wan, T.L. Tay, H. Wang, M. Chen, and T. Yan, Development of

transgenic fish for ornamental and bioreactor by strong expression of

fluorescent proteins in the skeletal muscle. Biochemical and Biophysical

Research Communications, 2003. 308(1): p. 58-63.

54. Yang, M., J. Reynoso, M. Bouvet, and R.M. Hoffman, A transgenic red

fluorescent protein-expressing nude mouse for color-coded imaging of the

tumor microenvironment. Journal of Cellular Biochemistry, 2009. 106(2): p.

279-284.

184

55. Liu, Z., J. Song, Z. Wang, J. Tian, Q. Kong, Z. Zheng, Z. Yin, et al., Green

fluorescent protein (GFP) transgenic pig produced by somatic cell nuclear

transfer. Chinese Science Bulletin, 2008. 53(7): p. 1035-1039.

56. Sasaki, E., H. Suemizu, A. Shimada, K. Hanazawa, R. Oiwa, M. Kamioka, I.

Tomioka, et al., Generation of transgenic non-human primates with germline

transmission. Nature, 2009. 459(7246): p. 523-527.

57. Chan, A.W.S., K.Y. Chong, C. Martinovich, C. Simerly, and G. Schatten,

Transgenic Monkeys Produced by Retroviral Gene Transfer into Mature

Oocytes. Science, 2001. 291(5502): p. 309-312.

58. Andersson, G., B.M.W. Illigens, K.W. Johnson, D. Calderhead, C. LeGuern,

G. Benichou, M.E. White-Scharf, and J.D. Down, Nonmyeloablative

conditioning is sufficient to allow engraftment of EGFP-expressing bone

marrow and subsequent acceptance of EGFP-transgenic skin grafts in mice.

Blood, 2003. 101(11): p. 4305-4312.

59. Beagles, K., L. Peterson, T. Neff, and H.-P. Kiem, 412. Cyclosporine Inhibits

the Induction of GFP Specific Immune Responses after Transplantation of

Lentivirally Transduced GFP Expressing Hematopoietic Stem Cells in Dogs.

Mol Ther, 2004. 9(S1): p. S158-S158.

60. Bosiljcic, M., M.J. Hamilton, J.P. Banath, N.E. LePard, D.C. McDougal, J.X.

Jia, G. Krystal, and K.L. Bennewith, Myeloid Suppressor Cells Regulate the

Lung Environmentâ€”Letter. Cancer Research. 71(14): p. 5050-5051.

61. Doi, K., J. Kong, J. Hargitai, S.P. Goff, and P. Gouras, Transient

Immunosuppression Stops Rejection of Virus-Transduced Enhanced Green

Fluorescent Protein in Rabbit Retina. J. Virol., 2004. 78(20): p. 11327-

11333.

185

62. Eixarch, H., A. Gomez, E. Kadar, M. George, N. Martinez, C. Espejo, J.

Petriz, R. Gimeno, and J. Barquinero, Transgene Expression Levels

Determine the Immunogenicity of Transduced Hematopoietic Grafts in

Partially Myeloablated Mice. Mol Ther, 2009. 17(11): p. 1904-1909.

63. Gambotto A, D.G., Cicinnati V, Kenniston T, Steitz J, Tüting T, Robbins PD,

DeLeo AB., Immunogenicity of enhanced green fluorescent protein (EGFP)

in BALB/c mice: identification of an H2-Kd-restricted CTL epitope. Gene

Therapy, 2000. 7(23): p. 2036-40.

64. Han, W.G.H., W.W.J. Unger, and M.H.M. Wauben, Identification of the

immunodominant CTL epitope of EGFP in C57BL/6 mice. Gene Ther, 2008.

15(9): p. 700-701.

65. Remy, S., L. Tesson, C. Usal, S. Menoret, V. Bonnamain, V. Nerriere-

Daguin, J. Rossignol, et al., New lines of GFP transgenic rats relevant for

regenerative medicine and gene therapy. Transgenic Research. 19(5): p. 745-

763.

66. Steitz, J., A.C. Soloff, S.M. Barratt-Boyes, S.M. Alber, S.C. Watkins, H.

Okada, and A. Gambotto, Balb/c EGFP mice are tolerant against

immunization utilizing recombinant adenoviral-based vectors encoding

EGFP: A novel model for the study of tolerance mechanisms and vaccine

efficacy. Molecular Immunology. 47(5): p. 1149-1153.

67. Tarantal, A.F., F. Giannoni, C.C. I Lee, J. Wherley, T. Sumiyoshi, M.

Martinez, C.A. Kahl, et al., Nonmyeloablative Conditioning Regimen to

Increase Engraftment of Gene-modified Hematopoietic Stem Cells in Young

Rhesus Monkeys. Mol Ther.

186

68. Re, F., R. Srinivasan, T. Igarashi, F. Marincola, and R. Childs, Green

fluorescent protein expression in dendritic cells enhances their

immunogenicity and elicits specific cytotoxic T-cell responses in humans.

Experimental Hematology, 2004. 32(2): p. 210-217.

69. Stripecke R, C.V.M., Skelton D, Satake N, Halene S, Kohn D., Immune

response to green fluorescent protein: implications for gene therapy. Nature

Gene Therapy, 1999. 6(7): p. 1305-12.

70. Tian, C., J. Bagley, J. Kaye, and J. Iacomini, Induction of T cell tolerance to

a protein expressed in the cytoplasm through retroviral-mediated gene

transfer. The Journal of Gene Medicine, 2003. 5(5): p. 359-365.

71. Annoni, A., M. Battaglia, A. Follenzi, A. Lombardo, L. Sergi-Sergi, L.

Naldini, and M.-G. Roncarolo, The immune response to lentiviral-delivered

transgene is modulated in vivo by transgene-expressing antigen-presenting

cells but not by CD4+CD25+ regulatory T cells. Blood, 2007. 110(6): p.

1788-1796.

72. Davey, G.M., S.N. Mueller, C. van Vliet, M. Gigowski, A. Zaid, B. Davies,

F.R. Carbone, and W.R. Heath, Identification of a MHC I-restricted epitope

of DsRed in C57BL/6 mice. Molecular Immunology, 2013. 53(4): p. 450-452.

73. Miletic, H., Y.H. Fischer, T. Giroglou, M.A. Rueger, A. Winkeler, H. Li, U.

Himmelreich, et al., Normal Brain Cells Contribute to the Bystander Effect in

Suicide Gene Therapy of Malignant Glioma. Clinical Cancer Research, 2007.

13(22): p. 6761-6768.

74. Limberis MP, B.C., Wilson JM., Identification of the murine firefly

luciferase-specific CD8 T-cell epitopes. Gene Therapy, 2009. 16(3): p. 441-7.

187

75. Kindt, T., B. Osborne, and R. Goldsby, Kuby Immunology. 2006: W. H.

Freeman.

76. Yewdell, J.W., E. Reits, and J. Neefjes, Making sense of mass destruction:

quantitating MHC class I antigen presentation. Nat Rev Immunol, 2003.

3(12): p. 952-961.

77. Engelhard, V.H., Structure of peptides associated with MHC class I

molecules. Current Opinion in Immunology, 1994. 6(1): p. 13-23.

78. Falk, K., O. Rotzschke, S. Stevanovie, G. Jung, and H.-G. Rammensee,

Allele-specific motifs revealed by sequencing of self-peptides eluted from

MHC molecules. Nature, 1991. 351(6324): p. 290-296.

79. Mitaksov, V. and D.H. Fremont, Structural Definition of the H-2Kd Peptide-

binding Motif. Journal of Biological Chemistry, 2006. 281(15): p. 10618-

10625.

80. Immuno Polymorphism Database URL:

http://www.ebi.ac.uk/ipd/imgt/hla/stats.html [2014].

81. Madden, D.R., J.C. Gorga, J.L. Strominger, and D.C. Wiley, The three-

dimensional structure of HLA-B27 at 2.1 Ã… resolution suggests a general

mechanism for tight peptide binding to MHC. Cell, 1992. 70(6): p. 1035-

1048.

82. Proimmune

URL:http://www.proimmune.com/ecommerce/page.php?page=MHC_alleles.

[2014].

83. Rammensee, H.-G., T. Friede, and S. Stevanovic, MHC ligands and peptide

motifs: first listing. Immunogenetics, 1995. 41(4): p. 178-228.

188

84. Rammensee, H.G., J. Bachmann, N.P.N. Emmerich, O.A. Bachor, and S.

Stevanović, SYFPEITHI: database for MHC ligands and peptide motifs.

Immunogenetics, 1999. 50(3): p. 213-219.

85. Parker, K.C., M.A. Bednarek, and J.E. Coligan, Scheme for ranking potential

HLA-A2 binding peptides based on independent binding of individual peptide

side-chains. J Immunol, 1994. 152(1): p. 163-175.

86. Lundegaard, C., K. Lamberth, M. Harndahl, S. Buus, O. Lund, and M.

Nielsen, NetMHC-3.0: accurate web accessible predictions of human, mouse

and monkey MHC class I affinities for peptides of length 8-11. Nucleic Acids

Research, 2008. 36(suppl 2): p. W509-W512.

87. Sidney, J., E. Assarsson, C. Moore, S. Ngo, C. Pinilla, A. Sette, and B.

Peters, Quantitative peptide binding motifs for 19 human and mouse MHC

class I molecules derived using positional scanning combinatorial peptide

libraries. Immunome Research, 2008. 4(1): p. 2.

88. Algonomics presentation URL: http://www.e-i-p.eu/wp-

content/uploads/2013/02/Stas_EIP_Leiden_2009.pdf [2012].

89. Nielsen, M., C. Lundegaard, O. Lund, and C. KeÅŸmir, The role of the

proteasome in generating cytotoxic T-cell epitopes: insights obtained from

improved predictions of proteasomal cleavage. Immunogenetics, 2005. 57(1-

2): p. 33-41.

90. Peters, B., S. Bulik, R. Tampe, P.M. van Endert, and H.-G. Holzhutter,

Identifying MHC Class I Epitopes by Predicting the TAP Transport

Efficiency of Epitope Precursors. The Journal of Immunology, 2003. 171(4):

p. 1741-1749.

189

91. Kim, Y., J. Ponomarenko, Z. Zhu, D. Tamang, P. Wang, J. Greenbaum, C.

Lundegaard, et al., Immune epitope database analysis resource. Nucleic

Acids Research. 40(W1): p. W525-W530.

92. Chirino, A.J., M.L. Ary, and S.A. Marshall, Minimizing the immunogenicity

of protein therapeutics. Drug Discovery Today, 2004. 9(2): p. 82-90.

93. De Groot, A.S. and W. Martin, Reducing risk, improving outcomes:

Bioengineering less immunogenic protein therapeutics. Clinical

Immunology, 2009. 131(2): p. 189-201.

94. Scott, D.W. and A.S. De Groot, Can we prevent immunogenicity of human

protein drugs? Annals of the Rheumatic Diseases. 69(Suppl 1): p. i72-i76.

95. Cantor, J.R., T.H. Yoo, A. Dixit, B.L. Iverson, T.G. Forsthuber, and G.

Georgiou, Therapeutic enzyme deimmunization by combinatorial T-cell

epitope removal using neutral drift. Proceedings of the National Academy of

Sciences. 108(4): p. 1272-1277.

96. Dobano, C., A. McTague, A. Sette, S.L. Hoffman, W.O. Rogers, and D.L.

Doolan, Mutating the anchor residues associated with MHC binding inhibits

and deviates CD8+ T cell mediated protective immunity against malaria.

Molecular Immunology, 2007. 44(9): p. 2235-2248.

97. Sofia, M., A. Marta, G.S. Philip, and J.P. Simas, A Single CD8+ T Cell

Epitope Sets the Long-Term Latent Load of a Murid Herpesvirus. PLoS

Pathogens, 2008. 4(10).

98. Glithero, A., J. Tormo, J.S. Haurum, G. Arsequell, G. Valencia, J. Edwards,

S. Springer, et al., Crystal Structures of Two H-2Db/Glycopeptide Complexes

Suggest a Molecular Basis for CTL Cross-Reactivity. Immunity, 1999. 10(1):

p. 63-74.

190

99. Royant, A. and M. Noirclerc-Savoye, Stabilizing role of glutamic acid 222 in

the structure of Enhanced Green Fluorescent Protein. Journal of Structural

Biology. 174(2): p. 385-390.

100. Paul, S., D. Weiskopf, M.A. Angelo, J. Sidney, B. Peters, and A. Sette, HLA

Class I Alleles Are Associated with Peptide-Binding Repertoires of Different

Size, Affinity, and Immunogenicity. The Journal of Immunology. 191(12): p.

5831-5839.

101. Fahnestock, M.L., J.L. Johnson, R.M.R. Feldman, T.J. Tsomides, J. Mayer,

L.O. Narhi, and P.J. Bjorkman, Effects of Peptide Length and Composition

on Binding to an Empty Class I MHC Heterodimer. Biochemistry, 1994.

33(26): p. 8149-8158.

102. Maryanski, J.L., P. Romero, A. Van Pel, T. Boon, F.R. Salemme, J.-C.

Cerottini, and G. Corradin, The identification of tyrosine as a common key

residue in unrelated H-2Kd restricted antigenic peptides. International

Immunology, 1991. 3(10): p. 1035-1042.

103. Romero, P., G. Corradin, I.F. Luescher, and J.L. Maryanski, H-2Kd-

restricted antigenic peptides share a simple binding motif. The Journal of

Experimental Medicine, 1991. 174(3): p. 603-612.

104. Patterson, G.H., S.M. Knobel, W.D. Sharif, S.R. Kain, and D.W. Piston, Use

of the green fluorescent protein and its mutants in quantitative fluorescence

microscopy. Biophysical Journal, 1997. 73(5): p. 2782-2790.

105. Punta, M., P.C. Coggill, R.Y. Eberhardt, J. Mistry, J. Tate, C. Boursnell, N.

Pang, et al., The Pfam protein families database. Nucleic Acids Research.

40(D1): p. D290-D301.

191

106. Chalfie, M., GREEN FLUORESCENT PROTEIN. Photochemistry and

Photobiology, 1995. 62(4): p. 651-656.

107. Brown, D.M., T.L. Fisher, C. Wei, J.G. Frelinger, and E.M. Lord, Tumours

can act as adjuvants for humoral immunity. Immunology, 2001. 102(4): p.

486-497.

108. Wonderlich, J., G. Shearer, A. Livingstone, and A. Brooks, Induction and

Measurement of Cytotoxic T Lymphocyte Activity, in Current Protocols in

Immunology. 2001, John Wiley & Sons, Inc.

109. Matzinger, P., The JAM test A simple assay for DNA fragmentation and cell

death. Journal of Immunological Methods, 1991. 145(1â€“2): p. 185-192.

110. Glimcher, L.H., K.J. Kim, I. Green, and W.E. Paul, Ia antigen-bearing B cell

tumor lines can present protein antigen and alloantigen in a major

histocompatibility complex-restricted fashion to antigen-reactive T cells. The

Journal of Experimental Medicine, 1982. 155(2): p. 445-459.

111. Schoenborn, J.R., C.B. Wilson, and W.A. Frederick, Regulation of Interferon

Gamma During Innate and Adaptive Immune Responses, in Advances in

Immunology. 2007, Academic Press. p. 41-101.

112. Bots, M. and J.P. Medema, Granzymes at a glance. Journal of Cell Science,

2006. 119(24): p. 5011-5014.

113. Speidel, K., W. Osen, S. Faath, I. Hilgert, R. Obst, J. Braspenning, F.

Momburg, G.J. Hämmerling, and H.-G. Rammensee, Priming of cytotoxic T

lymphocytes by five heat-aggregated antigens in vivo: Conditions, efficiency,

and relation to antibody responses. European Journal of Immunology, 1997.

27(9): p. 2391-2399.

192

114. Yewdell, J.W. and J.R. Bennink, IMMUNODOMINANCE IN MAJOR

HISTOCOMPATIBILITY COMPLEX CLASS I-RESTRICTED T

LYMPHOCYTE RESPONSES. Annual Review of Immunology, 1999. 17(1):

p. 51-88.

115. York, I.A. and K.L. Rock, ANTIGEN PROCESSING AND PRESENTATION

BY THE CLASS I MAJOR HISTOCOMPATIBILITY COMPLEX1. Annual

Review of Immunology, 1996. 14(1): p. 369-396.

116. Williams, A., C.A. Peh, and T. Elliott, The cell biology of MHC class I

antigen presentation. Tissue Antigens, 2002. 59(1): p. 3-17.

117. Morgan, C.S., J.M. Holton, B.D. Olafson, P.J. Bjorkman, and S.L. Mayo,

Circular dichroism determination of class I MHC-peptide equilibrium

dissociation constants. Protein Science, 1997. 6(8): p. 1771-1773.

118. Crameri, A., E.A. Whitehorn, E. Tate, and W.P.C. Stemmer, Improved Green

Fluorescent Protein by Molecular Evolution Using DNA Shuffling. Nat

Biotech, 1996. 14(3): p. 315-319.

119. Sakamoto, N., K. Tsuji, L.M. Muul, A.M. Lawler, E.F. Petricoin, F. Candotti,

J.A. Metcalf, et al., Bovine apolipoprotein B-100 is a dominant immunogen

in therapeutic cell populations cultured in fetal calf serum in mice and

humans. Blood, 2007. 110(2): p. 501-508.

120. Spees, J.L., C.A. Gregory, H. Singh, H.A. Tucker, A. Peister, P.J. Lynch, S.-

C. Hsu, J. Smith, and D.J. Prockop, Internalized Antigens Must Be Removed

to Prepare Hypoimmunogenic Mesenchymal Stem Cells for Cell and Gene

Therapy[ast]. Mol Ther, 2004. 9(5): p. 747-756.

193

121. Donnou, S., C. Galand, V. Touitou, xe, rie, Saut, C. s-Fridman, Z. Fabry, and

S. Fisson, Murine Models of B-Cell Lymphomas: Promising Tools for

Designing Cancer Therapies. Advances in Hematology. 2012: p. 13.

122. Wilson, M.E. and R.G. Hamilton, Immunoglobulin G subclass response of

localized juvenile periodontitis patients to Actinobacillus

actinomycetemcomitans Y4 lipopolysaccharide. Infection and Immunity,

1992. 60(5): p. 1806-1812.

123. Gene Infinity URL:http://www.geneinfinity.org/sp/sp_codonusage.html

[2014].

124. Wang, R.-F., The role of MHC class II-restricted tumor antigens and CD4+

T cells in antitumor immunity. Trends in Immunology, 2001. 22(5): p. 269-

276.

125. Skelton D, S.N., Kohn DB., The enhanced green fluorescent protein (eGFP)

is minimally immunogenic in C57BL/6 mice. Nature Gene Therapy, 2001.

8(23): p. 1813-4.

126. Denaro, M.e.a., EGFP-transduced EL-4 cells from tumors in C57BL/6 mice.

Gene Therapy, 2001. 8(23): p. 1.

127. Yla-Herttuala, S., Endgame: Glybera Finally Recommended for Approval as

the First Gene Therapy Drug in the European Union. Mol Ther. 20(10): p.

1831-1832.

128. Chen, K., L. Chen, P. Zhao, L. Marrero, E. Keoshkerian, A. Ramsay, and Y.

Cui, FL-CTL assay: Fluorolysometric determination of cell-mediated

cytotoxicity using green fluorescent protein and red fluorescent protein

expressing target cells. Journal of Immunological Methods, 2005. 300(1-2):

p. 100-114.

194

129. van Baalen, C.A., D. Kwa, E.J. Verschuren, M.L. Reedijk, A.C.M. Boon, G.

de Mutsert, G.F. Rimmelzwaan, A.D.M.E. Osterhaus, and R.A. Gruters,

Fluorescent Antigenâ€“Transfected Target Cell Cytotoxic T Lymphocyte

Assay for Ex Vivo Detection of Antigen-Specific Cell-Mediated Cytotoxicity.

Journal of Infectious Diseases, 2005. 192(7): p. 1183-1190.

130. Patterson, G., R.N. Day, and D. Piston, Fluorescent protein spectra. Journal

of Cell Science, 2001. 114(5): p. 837-838.

195

7. Appendices

7.1. Oligonucleotide primers

7.1.1. EGFP specific primers incorporating 5’ BamHI and 3’ NdeI sites

EGFP Forward: GGGCTCCATATGGTGAGCAAGGGCGAGGAG

EGFP Reverse: GCCATGGGATCCTTACTTGTACAGCTCGTC

7.1.2. Saturation mutagenesis primers.

Y200NNS Forward GCTGCCCGACAACCACNNSCTGAGCACCCAGTCCG

Y200NNS Reverse CGGACTGGGTGCTCAGSNNGTGGTTGTCGGGCAGC

N121NNS Forward GGGCGACACCCTGGTGNNSCGCATCGAGCTGAA

N121NNS Reverse CCTTCAGCTCGATGCGSNNCACCAGGGTGTCGCCC

7.1.3. Directed mutagenesis primers for Asn121.

N121D Forward CGACACCCTGGTGGACCGCATCGAGCT N121D Reverse GCTGTGGGACCACCTGGCGTAGCTCG N121A Forward GCGACACCCTGGTGGCCCGCATCGAGCTGA N121A Reverse CGCTGTGGGACCACCGGGCGTAGCTCGACT

7.1.4. EGFP specific primers incorporating 5’ KpnI and 3’ BamHI sites

EGFP Forward CGCGGCGGTACCATGGTGAGCAAGGGCGAG EGFP Reverse GCAGCCGGATCCTTACTTGTACAGCTCGTC

196

7.2. Vector Maps

7.2.1. pET 15b vector map for expression in E.Coli.

197

7.2.2. pcDNA 3.1 vector map for expression in mammalian cells.

198

7.2.3. Lentivirus vector map; CS-EGFP-Lv156. EGFP expression is under regulation by EF1α promoter with puromycin antibiotic selection marker.

199