Therapeutic monoclonal antibodies generated against tumour associated antigens (TAA) are highly specific cancer treatment. However, they are limited to the antigens presented on the cell surface whereas most carcinogens are intracellular. The increasing number of tumours refractory to the current anti-cancer antibody therapies, raised concerns regarding the suitability of targeting cell surface antigens. Intracellular TAAs are degraded into peptide fragments and displayed on the surface of malignant cells in the context of MHC class I where they can be recognised by cytotoxic T cell lymphocytes. To target this family of carcinogens, TCR mimic (TCRm) monoclonal antibodies have been developed with the unique specificity for peptides presented by MHC class I.
In spite of their vast applications, development of TCRm Abs is technically demanding. One efficient method is hybridoma technology which involves immunisation of mice with tetramers. At present fully human MHC class I tetramers are used which provokes the immune response against the whole tetramer and leads to the production of undesired antibodies targeting structural parts. To tackle this problem, we generated chimeric tetramers in which structural parts of human HLA-A*0201 were substituted with murine counterparts. Three chimeric constructs have been engineered previously in the lab and α3 domain of heavy chain was replaced with H-2Db and H-2Dd for use in C57BL/6 and Balb/c mice, respectively. The constructs were amplified using primers designed to engineer restriction sites into the final sequence. Products were initially cloned into TOPO vector followed by ligation into similarly digested expression vector, pET9C. Plasmids containing desired fusion proteins were transformed into DH5α strain of E.coli prior transforming the expression strain, BL21 (DE3). Sufficient amount of protein was expressed and extracted using IPTG induction. Monomers were then refolded in the presence of human β2m and standard 'Flu' peptide (GILGFVFTL) and biotinylated. Correctly refolded monomers were isolated by FPLC and concentrated. After tetramerization, the viability of the chimeric tetramers was demonstrated by FACS analysis.
1- Introduction
Therapeutic antibodies in cancer treatment
Generation of mouse derived hybridomas by Kohler and Milstein in 1975 and development of Monoclonal antibodies (mAbs) as a highly specific cancer therapy tackled the redoubtable challenge of differentiating tumour cells from healthy cells and lightened up a path towards the development of selective therapeutic modality for cancer. To date, more than 400 anticancer mAbs, comprising approximately 25% of all biotechnology products, are under investigation. Half of them are anticancer monoclonal antibodies and a few have been approved for clinical use 1,4.
Monoclonal antibodies can be divided into four main categories: murine, chimeric, humanized and human (except for murine mAbs, other types have human Fc portion). In general, mAbs can employ various direct and indirect functions for destruction of tumour cells. They can mediate antibody-dependent cellular cytotoxicity (ADCC) by recruiting effector cells of the immune system through engagement of Fc portion with FcγRI (CD64), FcγRII (CD32), the B isoform of FcγRIII (CD16) on neutrophils and the A isoform of FcγRIII (CD16) on NK cells1,2,3. ADCC can induce tumour destruction and increase antigen presentation and subsequently activate tumour associated T cell responses. Moreover, complement dependent cytotoxicity (CDC) can be induced by activating the complement system which in turn enhances ADCC by the release of chemotactic factors (C5a and C3a). Human IgG1, IgG3 and murine IgG2a are the potent isoforms in induction of ADCC and CDC. Unmodified mAbs can also mediate tumour cell killing by inhibiting angiogenesis, inducing apoptosis and blocking relevant receptors particularly activated growth factor receptors 1,2.
Furthermore, to increase the specificity of aggressive treatments (radiotherapy and chemotherapy) and improving the efficiency of immunotherapy, mAbs can be conjugated to radioactive isotopes, toxins, cytotoxic drugs and cytokines to directly target tumour cells. Although murine IgG2a has a weak ability to induce CDC and ADCC, their immunogenicity and short circulating half-life have limited murine mAbs' application in cancer therapy and the immune responses prevent repeated administration being very successful. Generally they can only be employed for targeting radioactive elements or cytotoxic agents to tumour cells 1,59.
Regardless of the mode of action, human IgG1 is a preferred antibody isoform due to its multiple potential functions for mediating tumour cell death and relatively longer half life 1.
Unmodified monoclonal antibodies:
Chimeric, humanized and human IgG1 mAbs which use ADCC and CDC as effector functions are of commercial and clinical interest. Rituximab (Rituxan), a chimeric IgG1 mAb against CD20 was the first mAb to be approved in the clinical setting as an effective treatment for diffuse large B cell lymphoma, leukemias, transplant rejection and some autoimmune disorders. It has believed to induce ADCC, CDC and apoptosis by altering intracellular calcium levels1,2,3,5. Alemtuzumab (Campath), a humanized IgG1 mAb against CD52 uses ADCC and CDC as main effector mechanisms while trastuzumab (Herceptin) mainly down regulates human epidermal growth factor 2 (HER2) and uses ADCC and CDC as an alternative mechanism. Studies showed that it can improve survival in advanced breast cancer and is a viable agent in patients with late-stage, metastatic endometrial carcinomas overexpressing HER2/neu 6,7. Altering the function of costimulatory molecules such as CD40 and CD137 is another mechanism by which mAbs can regulate tumour growth.
Modified monoclonal antibodies
Monoclonal antibodies can be used to selectively deliver cytotoxic agents. Radioisotopes (90Y and 131I), immunotoxins which consist of proteins (Pseudomonas exotoxin, Staphylococcus enterotoxin, neocarzinostatin, ricin and gelonin) and small molecules (vinblastine, methotrexate, doxorubicin, calicheamicin and maytansine) can be conjugated to the monoclonal antibodies and directly mediate tumour destruction. 90Y ibritumomab tiuxetan, 131I tositumomab and 131I ch-TNT are approved mAbs for non-Hodgkin's lymphoma and lung cancer. Gemtuzumab ozogamicin is an immunotoxin used to treat acute myelogenous leukemia1,3, 4.
Antibody-cytokine fusion therapy has been evolved to precisely activate the anti-tumour immune responses. Although antibody-IL-2 fusion proteins are the most investigated drugs in this category, other cytokines such as GM-CSF, IL-12, TNF-α, IFN-γ and LT-α have been scrutinized with limited success3,4.
Tumour Associated Antigens (TAA)
Anti-cancer monoclonal antibodies can be targeted against tumour cell-surface proteins, antigens associated with tumour stroma, vasculature and ligands. These targets should be of specific nature to be suitable for mAbs development. They should be expressed on the surface of tumour compartments and should not be internalized rapidly if ADCC and CDC are to be performed. Conversely, internalization is crucial for the cytotoxic activity of immunotoxins. Highly specific expression in target tissue, having a role in disease pathogenesis to minimise escape variants and active expression throughout disease stages and metastatic lesions are of other characteristics. They should not be shed or secreted into the circulation to lead the mAbs to the tumour site. Although various glycoproteins, glycolipids and carbohydrates were targeted, some were used with greater frequency (summarized in table ...) 1,2.
Currently generation of anti-cancer monoclonal antibodies are limited to the intact antigens presented on the surface of tumour cells. Although many successful mAbs have been developed, the number of cancer patients resistant to current medications is increasing (70%). Several reasons have addressed this issue. First, free TAAs may shed from the tumour and engage the antibody binding sites which lead to a drop in the number of active antibodies and subsequently their tumour cell killing ability. Moreover, most targeted antigens are complex molecules whereas only a single epitope is being recognised by antibodies which confine their efficacy. On the other hand, many potential targets associated with tumour genesis e.g. p53 are intracellular and are thus not accessible to antibody targeting by conventional methods. Therefore, it has been suggested that cell surface antigens are not ideal therapeutic targets for generation of mAbs.
The immune system is configured to allow the sentry of intracellular milieu through the peptides displayed on MHC class I molecules. These cytosolic oligopeptides derived from malignant transformation and intracellular pathogens or improperly presented and heavily expressed antigens, are generated by proteasomal and non-proteasomal pathways, displayed on MHC class I complex and recognised by CD8+ T cell lymphocytes. Since B cell lymphocytes recognise tertiary structure of proteins without any MHC restriction, development of antibodies recognising peptides presented by MHC molecules on the cell surface is not straightforward. One possibility is targeting peptide-MHC complexes by TCRs, but their low affinity and minimal stability limited their application8. In contrast, antibodies have higher affinities and are easier to handle. Thus, several approaches were performed to target this family of antigens as the next generation of antigens for development of anti-cancer monoclonal antibodies to direct immune responses toward intracellular antigens that are not themselves secreted or displayed on the cell surface. Telomerase catalytic subunit (hTERT), melanoma differentiation antigen gp100, epithelial cell associated mucin (MUC1/CanAg), MAGE 10 , NY-ESO-1 9 are examples of intracellular antigens presented by human MHC molecules HLA-A1 or HLA-A2 2,4,12.
MHC class I - structure, antigen processing and presentation
Major histocompatibility complex (MHC) class I molecule - known as Human Leukocyte Antigen (HLA) in humans and "H-2" in mice - formed of α chain which is polymorphic and encoded within the MHC region of chromosome 6 and non polymorphic β2 microglobulin chain encoded in chromosome 15 in human. The α chain spans the membrane and comprises three domains, α1, α2 and α3. The tertiary folded structure of α1 and α2 domains form a highly polymorphic peptide binding groove on the surface of MHC class I molecules which accommodates the presented peptide fragment of 8 to 10 amino acids with similar anchor residues. The α3 domain interacts with β2 microglobulin and facilitates the stability of the molecules but neither of them is involved in the formation of the peptide binding cleft (Fig 2). CD8 molecules are cell surface markers of cytotoxic T cells and recognise MHC class I complexes. They interact mainly with an invariable part of α3 domain and the base of α2 domain.
MHC class I presents a wide variety of peptides generated from incomplete, mutated or unfolded self proteins or foreign antigens on the cell surface to be recognized by CD8+ cytotoxic T cell lymphocytes and if necessary initiate an immune response. This allows the surveillance of the entire intracellular milieu by the adaptive immune system. The bound peptides are the integral part of the MHC molecules to avoid peptide exchange at the surface of the cells. They are mainly generated as N-extended peptides by proteasomal processing of intracellular antigens and undergo further sequential trimming in the cytosol and ER. Dendritic cells, to the limited extent, display extracellular proteins on MHC type I by cross presentation 7,32,33.
MHC tetramer production and its applications
Recombinant major histocompatibility complex (MHC) multimers, coupled with flow cytometry, have been developed to identify antigen specific T cells. Multimers differ in their valency, the applied expression system and the peptide loading strategy. Tetramers, by far the most popular reagents, are conventionally generated by refolding soluble MHC α chain monomers with β2m. Soluble MHC monomers are then biotinylated using biotin ligase and converted into tetravalent by adding the correct ratio of streptavidin or avidin. Tetramers are conjugated to fluorochromes (PE and APC) to be visualised and enumerated by fluorescently activated cell sorting (FACS) analysis 19, 13.
The expression system used for the generation of MHC class I and II are generally different due to the difference in their refolding processes 20. MHC class I is produced in bacterial cells such as Esherichia coli and refolded conveniently in vitro while refolding process of MHC type II is cumbersome therefore eukaryotic cells such as baculoviruses infected insect cells 21or Drosophila cell transfectants is used 22.
There are several techniques for loading the antigenic peptide in different stages. In the case of MHC class II molecules, peptides can be genetically linked to one of the MHC chains, while peptides are included during in vitro production or after generation of MHC class I monomer or even multimers 20.
HLA multimer technology has helped to study the frequency, phenotype and function of antigen specific T cells 13. Tetramers are able to identify antigen specific T cells targeting viral, tumour, and transplantation antigens with delicate sensitivity. Several types of viral infections have been studied particularly frequently such as EBV, CMV, LCMV, HCV, VSV, influenza virus, parvovirus B19 and HIV. Among bacterial infections Listeria monocytogenes dominated. Analysis of anti-tumour responses and evaluating post immunotherapy responses in patients suffering from cancer as well as estimation of the heterogeneity of memory repertoire are other areas of research 13.
Although tetramers were first designed as diagnostic tools, their ability as tools to manipulate T cell response has made them a promising therapeutic strategy. Potential clinical applications are selection and expansion of desired T cells (Adoptive T cell transfer) and removal of unwanted T cells following hematopoietic stem cell transplantation (reduction of GVHD). MHC tetramers are able to identify and tolerise autoreactive T cells in autoimmune disorders for example autoreactive CD4+ Tcells specific for GAD65 555-567(minitope) in type I diabetes 26. Another novel strategy was applied by Dimopoulos and colleagues. They joined MHC tetramer technology and intracellular cytokine staining (ICS) to identify antigen presentation specificity of tumour cells for the development of cancer vaccine targets with no need for Tcell cloning, T cell culture and no bias was seen by IFN γ production of unknown cells18.
T-cell receptor mimic (TCRm) antibodies and their therapeutic potential
Therapeutic vaccines for treatment of cancer and particular types of viruses are developed to induce T cell mediated immune responses. To measure the potency of such vaccines, several culture-based qualitative and semi-quantitative assays such as LDA (Limiting dilution assay), ELISPOT and ICS have been developed but the underestimation of the immune responses was the main limitation. Since the concentration of MHC bound to the specific antigen on the surface of vaccine treated antigen presenting cells (APCs) directly correlates with the intensity of the cytotoxic responses, monoclonal antibodies with the unique specificity for antigen specific MHC restricted T cells (TCR mimic antibodies) were employed as a reliable tool to detect presentation and intensity of CTL responses directed against tumours and viral infections17. They can elucidate structural and functional MHC-peptide-TCR interactions in detail, quantify the number of MHC bound to the specific peptide on the surface of antigen presenting cells and localise APCs within the normal and diseased tissues. Their ability to inhibit MHC-Peptide-TCR interactions, introduce their possible role in regulation of autoimmune disorders in vivo 8.
Recombinant TCR mimic antibodies can be isolated from a phage display library by expression in E.coli cells 8or generating in hybridomas30. Several studies were combine genetic immunisation with hybridoma technology to obtain high affinity monoclonal antibodies.
Although the application of TCR mimic antibodies in treatment of tumours was first confined to delivering toxins and drugs to the tumour site, Wittman and colleagues in their cutting edge research have showed that TCRm antibodies could potentially activate components of the innate immune system and kill specific tumour cell lines. They have developed a murine IgG2a TCR mimic antibody targeted intracellular GVL peptide from human chorionic gonadotropin β (hCGβ) bound to HLA-A2 and demonstrated that it can mediate tumour cell lysis by both complement dependent cytotoxicity (CDC) and antibody dependent cellular cytotoxicity (ADCC) in a human breast cell line carcinoma in vitro. They were also revealed the in vivo prophylactic ability of the 3.2 G1 TCRm antibody as the implantation and growth of MDA-MB-231tumour cells were inhibited in nude mice28.
Recently, Verma and colleagues have provided the proof of the therapeutic potential of this antibody as it can impede the growth of MDA-MB-231 and MCF-7 tumours in orthotopic models of the breast cancer or even eliminate them at the highest dose without attacking the normal tissue of the breast27.
Difficulties in generating T-cell receptor mimic antibodies
In spite of their vast implications in diagnosis and therapy of tumours, viral infections, autoimmunity and transplantation, the production of monoclonal antibodies with precise specificity for T cell receptors are technically demanding. As explained earlier one efficient method for generation of TCR mimic antibodies is using hybridoma technology which involves the immunisation of HLA/A*0201 transgenic mice with the specific tumour or viral peptide fragments contained in human MHC Class I tetramer. It was demonstrated that immunisation with human MHC molecules results in development of unwanted antibodies against structural regions which are not involved in TCR recognition. Since human MHC α3 domain interacts with low affinity with the murine CD8 molecule, the vast majority of undesired antibodies are targeted to this region35,36,37. This complicates the isolation of TCR mimic antibodies.
Moreover, assessing the efficacy of vaccines and evaluating different vaccination regimens in HLA-A2 transgenic mice needs high HLA-A2 restricted cytotoxic T cells and high staining efficacy which could not be obtained with poor interaction of human MHC class I α3 domain of tetramers with murine CD8 molecules35,36,37.
Chimeric tetramers - potential in CTL evaluation and promoting TCRm generation
Generation of chimeric human-murine MHC class I tetramers in which the human α3 domain is substituted with the murine counterpart can increase the affinity of MHC-I CD8 interactions without reflecting on the peptide binding groove. These recombinant tetramers are able to increase the staining efficiency leading to more accurate evaluation of CTL responses in transgenic mice.
To date, a few successful attempts have been made to produce chimeric tetramers. Engineering of tetramers containing a non-specified H-2D murine α3 domain by Ren and colleagues was the proof of concept. Furthermore, Choi and colleagues produced a human-murine chimeric HLA-A2 tetramer containing the murine MHC allele H-2Kb α3 domain coupled with human β2 microglobulin (A2Kb). The recombinant A2Kb tetramers were used for the evaluation of CTL responses in HHD mice (H-2Dd-/-/β2m-/-) vaccinated with "DNA-prime recombinant vaccinia virus (rVV) boost strategy". They showed higher staining capability for cytotoxic T cells compared to unmodified A2 tetramers34.
HHD molecules - a human HLA-A*0201 α1/α2 linked to the mouse H-2Db α3 transmembrane and cytoplasmic domains - is a successful chimeric monochain devised by Pascolo and colleagues in the University of Edinburgh. A human β2 microglobulin is covalently bound to its heavy chain by a 15 amino acid linker. Another chimeric monochain with the same heavy chain coupled with the murine β2 microglobulin (MHD) was engineered as well. Their efficiency was evaluated in H-2Dd-/-/β2m-/- double knockout mice (HHD) and compared with the fully human HHH monochain. The expression of HHD and HHH molecules on the cell surface was far better than MHD constructs. The superior interaction of mice CD8 with the mice α3 part leads to the better recognition of cytotoxic responses by HHD chimeric monochains. These data proposed that development of tetramers with more murine character may not function as viable as HHD for detection of cytotoxic responses 38.
In the case of TCR mimic monoclonal antibodies chimeric tetramers could direct the generation of antibodies towards the peptide-binding regions of the MHC molecule by making the structural parts of the molecule less immunogenic. Therefore, sparing unwanted antibodies and make the purification of practical antibodies straightforward.
Recombinant HHD is a valuable target in development of monoclonal antibodies as it derived from the most common human HLA, HLA-A2 and its functionality has been confirmed for other applications. Moreover, generation of further chimeric tetramers with the same human heavy chain and different murine compartment for use in common murine expression systems in development of monoclonal antibodies, Balb/c mice, can improve the efficiency of TCR mimic antibody generation ( next chapter).
Chimeric tetramers for use in Balb/c mice
Balb/c mice, an albino inbred widespread lab strain, are commonly used to generate monoclonal antibodies as they develop plasmocytomas when injected with mineral oil. Neethling and colleagues generate TCR mimic antibodies by immunization of this strain with peptide-HLA-A*0201 complex and Quil-A adjuvant17. The same procedure was utilized by Weidanz via "eIF4G (720)-HLA-A0201 complex"30. Timusk developed TCRm antibodies by genetic immunization of the same strain39.
Successful employment of Balb/c mice by Wittman for the development of TCR mimic antibodies after immunisation with fully human tetramers "HLA-A2-hCG 47-55 peptide (designated GVL/A2) complex" showed the practicability of this strain27,28. Although HHD chimeric monomer seems to be a perfect construct for the generation of TCR mimic monoclonal antibodies but HHD strain of mice have not been applied for this intention. As a result TCRm antibody development could be hampered as it needs optimisation of hybridoma fusion techniques in this strain. On the other hand, HHD construct might not be compatible with Balb/c mice for generation of monoclonal TCR mimic antibodies.
Consequently combination of these two systems - HHD constructs and Balb/c strain of mice- and production of chimeric tetramers compatible with both strains could increase the chance of generating TCR mimic monoclonal antibodies.
Antibody repertoire screening
To identify antibodies generated against the peptide binding groove, HLA-A*0201 control tetramers were refolded using influenza A matrix protein-derived peptide ("Flu Peptide", amino acid sequence GILGFVFTL). Flu peptide is a widespread standard peptide for generation of control tetramers. Since the structural regions are similar between control tetramers and those developed using tumour associated antigens, antibodies that can bind to control tetramers are considered nonspecific and removed from the repertoire 50.
Aims of the dissertation
As discussed in detail earlier, the aim of this project is therefore to generate HHD and Balb/c compatible chimeric MHC Class I tetramers by substituting structural parts of MHC class I (α3 and β2 microglobulin) with their mouse counterparts.
To achieve this, the chimeric constructs' and mouse β2m sequences were amplified by PCR and cloned into TOPO vector followed by sub-cloning into pET9C plasmid vector. The expression vector were first transformed into DH-5α to produce high levels of plasmid DNA followed by transformation into BL21 (DE3) for protein expression. Monomers were refolded in the presence of antigenic peptide and β2 microglobulin, biotinylated, tetramerized and analysed by FACS. Different steps are summarised in figure........
These tetramers could then be used to generate therapeutic TCR mimic monoclonal antibodies in C57BL/6 and Balb/c mice against various tumour antigens; potentially with improved efficiency due to decreased immunogenicity of the tetramer.
Materials and Methods:
Templates
Three chimeric constructs were available. HHD construct containing human HLA-A2.1 (α1 and α2 domains) and mouse H-2Db (α3 domain) for use in C57BL/6 mice was a generous gift from Dr. F. Lemonnier, Pasteur Institute, Paris. The construct has been modified by removing β2 microglobulin, transmembrane domain and cytoplasmic tail in Nuffield Department of Clinical Laboratory Sciences. Balb/c compatible constructs, MM and LS, have been previously engineered by stitch PCR using human HLA-A*0201 and mouse H-2Dd cDNA as a template. H-2Dd RNA, derived from Balb/c mouse splenocytes, was undergone reverse transcription and attached to the human HLA-A*0201 α1α2 domain at different transition points to ensure efficient recombinant protein production. Regarding the MM construct the exon junction point in the coding sequence between the α2 and α3 domains was selected as a transition point while the LS construct contained slightly more human component and the transition point was three amino acids downstream of the predicted junction.
Primer design
Two oligonucleotide primers were designed manually for HHD and Balb/c compatible chimeric fragments. To maintain the specificity and efficiency of amplifications and reduce mispriming events, the melting temperatures were calculated by Suggs and co-workers formula as Tm=2°C (A+T ) + 4°C (G + C) and were kept around 72 degree centigrade. For optimal results the 3' end of the primers and the templates were perfectly paired. Interaction between forward and reverse primers as well as self complementarity of primers like stem loops were avoided. G or C nucleotides were chosen at a 3' end to increase the priming efficiency. A stop codon was introduced to the end of the reverse primers prior to the restriction sites6.
In order to clone the PCR products the restriction enzyme sites of Nde I or BamHI that does not cut within the templates and located at the convenient site for insertion of the fragments in the expression plasmid vector were introduced at the end of forward and reverse primers respectively. Six additional nonspecific bases 5' to the recognition sequence of the restriction enzymes were added to the length of the non-template specific portion of the primer to cut efficiently. Another similar reverse primer was designed adding Biotin tag followed by a stop codon and Bgl II restriction enzyme site to its 3'end.
Primers commercially synthesised using the (Invitrogen) Custom Oligonucleotide facility. Alignments were verified by using DNA strider software package. 100µM stock was prepared and heated at 65°C for 10 minutes. A working stock of 4µM in 200µl purified H2O was made for each primer.
Polymerase Chain Reaction
For high-fidelity amplification of DNA fragments, recombinant Pfu DNA Morepolymerase from Thermococcus species KOD with proofreading (3´-5´ exonuclease) activity supplied with the 10x buffer was purchased from Promega (Cat. M774B) as a cloned product. The PCR was carried out in a reaction volume of 100 μl in a PTC-200 thermal cycler for 35 cycles with an optimal annealing temperature of 57°C. 1µl Pfu DNA polymerase mixed with 10µl 10x buffer, 72µl nuclease free dH2O, 5µl primers complementary to the 3´ends of each of the sense and antisense strand of the template and 5µl buffered deoxyribonucleotide (dNTP). The conditions were as follows:
5 minutes initial denaturation at 95°
35 cycles of:
Denaturation step: 30 seconds at 95°
Annealing step: 1 minute at 57°
Elongation step: 2 minutes at 72°
A final 10 minutes Final elongation at 72°
Gel extraction of desired DNA
The 'Promega Wizard SV Gel and PCR Clean-Up Kit' was used for gel extraction and PCR product purification (Technical Bulletin 308 Wizard SV Gel and PCR clean-up system, 2009). The desired band was identified under a long-wavelength UV lamp and physically excised with a scalpel. The removed band was weighted, vortexed and heated to 65°C with equal amount of Membrane washing solution containing guanidine isothiocyanate. The mixture was decanted into the membrane-containing filter unit and incubated for 1 minute prior to spinning at 13,000g. DNA bound to silica membrane in the presence of chaotropic salts. The filter was washed twice with the membrane wash solution (10 mM potassium acetate, 80% ethanol, 16.7 µM EDTA) 700µl and 500µl respectively followed by centrifugation. Elution was done using 30µl nuclease free water pH:7.5.
Preparation of Chemically Competent strains of E. coli
BL21 (DE3) and DH5α strains of E.coli were made chemically competent by the following protocol: 5ml LB (Luria-Bertani medium) was inoculated with a single colony picked from a freshly streaked agar plate grown for 16 hours at 37°c with vigorous shaking (250 rpm). The culture was diluted 1/100 into 100ml sterile fresh LB with aeration. The viable cell concentration was monitored closely until the OD600 reached 0.3- 0.4 (mid logarithmic phase). The solution were decanted into 50mL falcon tubes and cooled on ice for 30 minutes followed by centrifuge at 4000 g for 10 minutes at 4°c. The pellet was then resuspended in 10mL sterile chilled 0.1M CaCl2 and incubated on ice for 30 minutes. The cells were harvested and resuspended in 5ml 0.1M CaCl2 containing 20% glycerol and incubate on ice for 30 minutes. The solution containing chemically competent E.coli cells were instantly frozen in combination of dry ice and methanol before storage in -80°C.
TOPO cloning and DNA mini preparation
One microlitre Salt Solution (supplied with the vector) was added to 4.7μl purified PCR product and 0.5μl TOPO cloning vector (pCR-Blunt II TOPO Invitrogen Cat. No. K2700-20) prior to incubation at room temperature for 30 minutes. Ligation reactions were incubated on ice for 30 minutes with 50μl One Shot Top 10 competent cells, followed by heat shock at 42° for 30 seconds, cooling on ice for 2 minutes. 250µl of S.O.C. medium was added at the room temperature and shook horizontally (200rpm) for an hour at 37°C. The transformation was spread on pre-warmed kanamycin selective agar plate and incubated at 37°C over night.
Six universal tubes each containing 4ml of terrific broth with 50ïg/ml Kanamycin were inoculated with single colonies and grown at 37° for 16 hours on a 225rpm shaker. Bacterial cells were harvested by centrifuge at 8000 rpm for 3 minutes at room temperature. DNA was extracted by alkaline lyses of bacterial cells using QIAprep spin miniprep kit (QIAgen). Pelleted cells were re-suspended in 250µl buffer P1 containing RNase A, Tris.Cl and EDTA by vortexing. 250µl Buffer P2 containing NaOH and 1% SDS followed by 350µl neutralization buffer N3 containing guanidine hydrochloride and acetic acid dispersed through viscous bacterial lysate. The mixture was applied to a mini-column, DNA adsorbed onto the silica in the presence of high salt. Mixtures were spun at maximum speed for 10 minutes and sequentially washed by 500µl and 750µl buffer PB and PE (containing 96% ethanol). DNA was eluted by 50µl nuclease free water pH: 7.5 and recovered by centrifuge for 1 min at the top speed. The DNA concentration quantified on a (Thermo Scientific) Nanodrop 2000 spectrophotometer.
Vector digestion
The inserts were cleaved out of the vector by digestion with the appropriate restriction enzymes in 50µl reactions: 25µl construct, 5μl 10x (New England Biolabs) NEBuffer 3 Buffer, 2μl of each restriction enzyme, volume adjusted to 50μl using nuclease free water and reacted at 37° for 5 hours. The insert were separated by Ethidium bromide agarose gel electrophoresis and gel extracted. Smaller digests retained reagents rations and conditions.
Conventional Cloning method
The conventional cloning method was performed applying different concentrations of the vector and the LS insert to maximise the possibility of the successful ligation. The reactions were set up under the following conditions: 2µg of digested purified pET9C (see next chapter), Nde1/BamHI digested LS fragments to vector in a molar ratio of 1:1, 2:1 and 3:1, 1µl Bacteriophage T4 DNA ligase and 10x ligase buffer. The reactions were transformed into DH5α, and plated on kanamycin resistant selective plates (as described in the following chapters)
Modified pET9C vector preparation
pET9C plasmid vector was purchased from Novagen. 80µl solution containing 5µg vector was digested over night at 37°C with 5µl Nde1, 5µl BamHI , 10µl 10x REact 4 Buffer (Invitrogen) and 5µl nuclease free water. To prevent re-ligation of complementary ends of the linearized vector, the phosphate group from the 5' end of vector was removed using 2µl Shrimp Alkaline Phosphatase (SAP) and kept in 37°C for 60 minutes. The mixture is then heated for 15 minutes at 65°c to stop the reaction and completely inactivate SAP. The reaction was run on a 100ml 1% agarose gel and bands corresponding to pET9C extracted using the (Promega) Wizard SV Gel and PCR Clean Up Kit.
Sub-cloning of fragments into pET9C
Ligation was performed by adding 3-fold molar excess of the inserts (6µl) to 2µl Nde1/ BamHI digested (NdeI/BglII digestion for MM constructs) pET9C vector. 1 μl 10x T4 Ligase Buffer (Invitrogen) and 1μl T4 DNA ligase (Invitrogen) was added to the mixture to bind them covalently and kept for 24 hours at room temperature, after which the ligation blend was transformed into competent E.coli DH5α cells, using the same protocol as described for TOPO cloning. Negative transformation control was performed using 6μl dH2O as a substitute of the vector. 2.5μl 100pg/μl of a high copy number E. coli plasmid cloning vector pUC19 containing an ampicillin resistant gene was replaced with the expression vector to apply as a positive control and plated on ampicillin contained selective plates.
Glycerol stock preparation
Fifteen mililitres of LB containing 50μg/ml kanamycin was inoculated with one Colony from each transformation plate and grown for 16 hours at 37° on a 225rpm shaker. Aliquots of 0.5ml sterile 30% glycerol combined with 0.5ml cell culture were prepared and stored as 15% glycerol stock at -80° for future use.
Transformation into BL21
Fifty micro litres of the expression strain of E.coli, BL21 (DE3), was transformed by 0.5µl purified pET9C plasmids encompassing MM, LS and HHD fragments, the procedure was the same as one applied for TOP 10 and DH-5α cells.
Small scale protein expression
After overnight incubation of 1ml 50μg/ml kanamycin LB with three freshly grown colonies, 50μl (1/100 dilution) of each sample were decanted into 5ml 50μg/ml kanamycin LB; cultured for 2.5 hours at 37° on a 225rpm shaker until the optical density (OD550) reached 0.5- 0.6. 1mL uninduced samples were separated before induction for an additional 3 hours with 1mM (4μl 1M added) isopropyl-β-D-thiogalactopyranoside (IPTG). The cell pellets were collected by centrifuge at 13,000rpm for 1 minute at 4°C and stored at -20°. Positive control was performed by inoculating 1mL 100μg/ml ampicillin LB medium with 0.5µl glycerol stock BL21/A2010.
50 mM Tris.Cl, 2% SDS, 0.1% bromophenol blue and 10% glycerol were combined to make 1x SDS gel loading buffer. 3.09 g dithiothreitol (DTT) was dissolved in 20mL of 0.01 M sodium acetate and filtered. DTT were added to the 1x SDS gel loading buffer with 1/10 ratio. Each pellet were then resuspended in 100µl of 1x SDS gel loading buffer (DTT added) and centrifuged at 13000 rpm for 1 minute at room temperature after heating up for 3 minutes at 100°c.
Optimisation of protein expression
Screening post-induction time intervals
After overnight culture of the colonies in 1ml 100µl/ml kanamycin LB medium, 100µl were decanted into 10ml fresh 100µl/ml kanamycin LB medium. When the optical density reached (A550) 0.5 - 0.6, the induction with 1mM IPTG carried out and bacterial cells harvested every hour for 6 hours.
IPTG titration
To optimise the protein yield, different concentrations of IPTG were tested by transferring 200µL of the overnight culture into 20ml 100μg/ml kanamycin LB. In mid log phase of growth, 1ml un-induced culture was separated and the remaining solution was divided into 2ml aliquots and induced with different concentrations of IPTG from 0.001-5mM prior to 2 hours incubation.
Overnight Express media
The auto-induced 'Overnight express instant TB media' was tried to assess the protein expression. 15ml 100µl/ml kanamycin of the overnight express media was inoculated with one colony of the transformation plates and cultured overnight in 225rpm shaker at 37°c. 1ml of the culture was decanted, centrifuged for 1 minute at maximum speed and the pellet was prepared for SDS- PAGE as described in the next chapter.
SDS-polyacrylamide gel electrophoresis (SDS-PAGE)
To verify the expression of the chimeric MHC monomers and estimate the quantity of the expressed protein, SDS polyacrylamide gel electrophoresis was carried out. 10% SDS polyacrylamide resolving gel 0.75mm thick was prepared by mixing 1.9ml H2O, 1.7ml 30% acrylamide mix, 1.3ml 1.5Tris (PH:8.8), 0.05ml 10% SDS, 0.05ml 10% ammonium persulfate (APS), 0.002ml TEMED prior to adding 5% SDS polyacrulamide stacking gel (1.4 H2O, 0.33 acrylamide,0.25 1.0 Tris (PH:6.8), 0.02 10% SDS, 0.02 10% APS, 0.002 TEMED). After polymerization, the samples were loaded in the wells of the stacking gel and run in a discontinuous buffer system containing 1x SDS running buffer.
The gel was stained with coomassie brilliant blue for an hour and followed by destaining by 30% methanol: 10% acetic acid solution for further 4 hours. For transillumination and easy visualization, acetate sheets were used to dry the gels.
Large scale expression of the chimeric MHC monomers
For high-throughput protein expression, 50ml 100µl/ml kanamycin LB media was inoculated with a picked colony from BL21 transformation plates and grown at 37°c overnight. Overnight cultures were used for aseptic inoculation of sterile 100µl/ml kanamycin LB (0.001 ratio) and shook for 3 hours in 130rpm shaker at 37°c. The optical density (OD550) was measured by Genova spectrophotometer and kept between 0.5- 0.8. The constructs were induced by the ideal concentration attained in optimization process; 0.005mM IPTG for MM and LS constructs and 0.01mM IPTG for HHD.
Protein extraction
Several methods were tested to get high-throughput protein yield:
Mechanical disruption of E.coli cells using sonication:
The induced cultures were grown at 37°c on 200rpm shaker for 2 hours prior to centrifuge at 4000rpm for 20 minutes at 4°c. The pellets were resuspended in 8ml cold PBS buffer per each litre of culture. 10 30-second cycles of sonication with 60-minute intervals were performed on ice in a Branson Sonifier® 250 (VWR, Leuven, Belgium) to lyse the cells and shear DNA. Once the samples were no longer viscous, 15000rpm centrifuge at 4°c for 20 minutes were carried out followed by three rounds of manual homogenisation in 25ml of Triton wash buffer (50Mm Tris PH:8.0, 100mM NaCl, 0.1% NaN3, 1mM EDTA, 1mM DTT and 0.5% Triton X-100) separated by 10-minute rounds of centrifuge. The pellet were homogenised in 20ml wash buffer with no Triton. After spinning, 10ml of 8M Urea containing 16ml 10M Resin treated (Amberlite) Urea, 0.1M NaH2PO4, 0.01M Tris PH:8.0, 0.1mM EDTA, 0.1mM DTT and 1.9ml dH2O was used to homogenise the pellet followed by overnight rolling at 4°c. The soluble protein was recovered by centrifugation at 15000rpm for 20 min at 4 °C, the protein concentration of supernatant was analysed by Biorad protein assay (Cat no 500-0006); protein solution was diluted 1:10 in urea buffer, 1ml 1:5 diluted Biorad assay reagent was added to different concentrations (0,2,4 and 8µl) of protein and optimal density (OD595) was measured at 595nm on a Genova spectrophotometer. As SDS gel electrophoresis showed incomplete purification, chemical method of protein purification was performed.
Chemical cell wall perforation for releasing expressed proteins:
'BugBuster' reagent (Novagen, 70750-3) was used to chemically extract protein from bacterial cultures. Cells were harvested from the culture by centrifuge at 10,000g for 10 minutes at 4°C and weighted. The wet cell paste was resuspended in 5ml BugBuster reagent per gram pellet to lyse bacterial cells and dissolve soluble proteins. 1µl/ml Benzonase was added and gently shook at room temperature for 20 minutes to reduce the viscosity of the solution by degrading liberated DNA. To remove soluble protein fraction, the solution was centrifuged at 16000g for 20 minutes at 4°C. After pellet resuspension in the same amount of BugBuster, rLysozyme was added to the final concentration of 1 KU/ml (25µl 5mg/ml lysozyme per 1ml suspension) and incubated for 5 minutes at room temperature. rLysozyme hydrolyzes N-acetylmuramide linkages and perforates the cell wall, hence improve the extraction. 5ml 1/10 H2O diluted BugBuster per gram original pellet weight was added and centrifuged at 5,000xg for 15 minutes at 4° followed by three rounds of washing with 10ml 1/10 diluted BugBuster and centrifuge at 5000g for 15 minutes at 4°C. The pellet was resuspended in 10ml 1/10 diluted BugBuster and centrifuged for 15 minutes at 16000g at 4°C. Using a manual homogeniser, the inclusion bodies were resuspended in 8M urea solution. Protein was quantified at 595nm on a Genova spectrophotometer using Biorad protein assay reagent and verified on 10% SDS gel.
Refolding of inclusion bodies
Refolding buffer was prepared adding 42.14g L-Arginine, 2ml 0.5M EDTA, 1ml Protease Inhibitor cocktail, 25ml 2M Tris and dH2O to the final volume of 500ml. After stirring at 4° for 30 minutes the solution was supplemented with 0.15g oxidised Glutathione and 0.77g reduced Glutathione followed by a further 30 minutes stirring at 4oC. 100μl of 50mg/ml "Flu" peptide GILGFVFTL was added drop by drop. 12.5mg of human β2m inclusion bodies and 7.5mg chimeric MHC class I α chain were added respectively via 25G needle over a 30 minute period and stirred for 24 hours at 4°C. A further 7.5mg of α chain was added, followed by 48 hours stirring at 4°C.
Resultant solution was centrifuged at 4000rpm for 20 minutes at 4°C and the supernatant concentrated using a YM-10 ultrafiltration 76mm membrane (Milipore, 13642) in an Amicon Stirred Ultrafiltration cell 8400 (Amicon, 5124) to 50ml using N2 gas. After overnight stirring at 4°C, further concentration was performed using Amicon Ultra 15 Millipore Filter Tube and desalinated on a GE Healthcare PD10 desalination column and eluted with 3.5ml of 10mM Tris pH:8 to the final volume of 5ml.
Biotinylation
To generate MHC tetramers, refolded monomers were enzymatically biotinylated by adding 875μl Biomix A, 875μl Biomix B, 100μl Biotin and 8μl 1mg/ml BirA (Biotin Ligase BirA500, Avidity LLC reagents) to the final concentration and incubated at room temperature for 24 hours.
Fast Protein Liquid Chromatography (FPLC)
Refolded monomers were isolated using a ÄKTA Purifier UPC10 Fast Protein Liquid Chromatography (FPLC) machine (GE Healthcare) with a HiLoad 26/60, Superdex 75 prep grade column 51. FPLC peaks samples were collected; first and second peaks were concentrated using Amicon Ultra-15 30K Millipore filter tubes and 3K tubes were used for third and forth peaks. To identify each peak 10% SDS-PAGE was applied and refolded monomers were stored in -80 in 50µg aliquots.
Tetramerization
Biotinylated chimeric MHC class I monomers were conjugated by addition of 49µl 190µg/ml extravidin-R-phycoerythrin (Sigma-Aldrich). In the first step 25µl extravidin was mixed with 50µg monomers and gently shook at 4° C for 30 minutes. The remnant 24µl extravidin was divided into 3 equal shots and added subsequently in 3 steps divided by 30-minute rotatary shaking at 4°C. Once prepared, the tetramers were gently shook at 4° C overnight.
FACS analysis of tetramers
Fluorescently activated cell sorting (FACS) analysis was performed on 'Flu' peptide-specific T cell lines there were generated from HLA-A*0201 positive peripheral blood mononuclear cells (PBMCs). T cells were harvested by centrifuge at 1500 rpm for 5 minutes at 4°C and resuspended in 50µl FACS wash buffer containing 48.9µl PBS, 0.1 5% sodium azide and 1µl fetal calf serum. Cells were co-stained with 0.5µl anti - CD3- allophycocyanin (APC) conjugated antibody and 2µl of each tetramer for 30 minutes. Positive and negative staining were also performed by adding HLA-A2 tetramers and FACS wash buffer, respectively. Cells were washed with FACS wash buffer followed by centrifuge at 1500 rpm for 5 minutes and fixation by 0.5ml 1% formaldehyde in PBS. Analysis was carried out on FACSCalibur machine by CellQuest software. The data was further analysed using Flowjo software.
