Patent Application
20090117153
A critical threshold density of cell surface MHC/peptide complexes is required to prime naïve CD8 T cells and to activate cytolytic CD8 effectors. Although the affinity of the peptide for the MHC is a major factor in determining the immunogenicity of any particular peptide antigen, immunodominant peptides in CD8 T cell responses to pathogens are not always the tightest class I binders (1). This is because immunodominance is also influenced by levels of donor protein in the cytosol, by the relative efficiency of processing particular epitopes, and by pathogen interference with antigen presentation. Moreover, the T cell repertoire has a major impact on determining which peptides are immunodominant in CD8 T cell responses (1-3). The fact that certain antigenic peptides are difficult to process and/or are relatively weak binders has considerable implications on vaccination strategies to pathogens and tumors (4). This is particularly the case in primary CD8 T cell responses that require cross priming. In the case of cross-priming, professional APCs which have acquired antigen exogenously, rather than by endogenous synthesis in the cytosol, will only be able to activate T cells if class I/peptide complexes have a sufficient half-life (5).
Tetramers and other multivalent staining reagents used to enumerate CD8 T cell populations in response to pathogens or tumors are also highly influenced by class I peptide binding affinity. Such reagents are typically made with soluble recombinant class I heavy chains expressed in bacteria and refolded with synthetic peptides (6). The production and stability of a tetramer is determined by the affinity of the peptide for the MHC, thus limiting the study of T cells specific for lower-affinity complexes.
Class I major histocompatibility complex (MHC) proteins serve a critical role in the adaptive immune response by binding short peptide fragments intracellularly and presenting them at the cell surface for surveillance by cytotoxic T lymphocytes (CTLs) (69-72). Structural studies of human and murine MHC class I proteins have revealed that peptides of 8-10 residues in length are presented to T cell receptors (TCRs) in the context of a narrow groove formed by two antiparallel α-helices positioned above an eight-strand, antiparallel β-sheet (73-77). This peptide-binding platform is comprised of the α1 and α2 domains of a polymorphic heavy chain (HC), which also contains an immunoglobulin-like α3 domain. The surface expression of peptide-MHC complexes requires the association of this heavy chain with an invariant, light chain, β2-microglobulin (β2m), which forms extensive contacts with both the α3 domain and the peptide-binding platform.
The ability of MHC proteins to activate T cells is critically dependent on the amount of peptide-MHC expressed at the cell surface (58, 60, 78). However, certain MHC-presented epitopes that have the capacity to activate the immune response are not expressed efficiently at the plasma membrane. This is because the cell-surface density of these protein complexes can be influenced by a variety of factors such as, for example, efficiency of antigen processing (79, 80), specificity of peptide translocation into the ER (81), interference by viral pathogens (1), and the kinetic stability of the peptide-MHC complex itself (82). Although T cell responses to such peptide-MHC complexes tend to be subdominant, the efficient induction of such responses has been shown to be crucial in achieving a broader and more effective antiviral and antitumor immunity (7, 9, 83-85). We have engineered peptide-MHC complexes as single chain trimers (SCTs) (10, U.S. patent application Ser. No. 11/397,377 filed Apr. 4, 2006). These engineered proteins comprise an antigenic peptide followed by a flexible linker that connects the C terminus of the peptide to the N terminus of β2-microglobulin (β2m), and another flexible linker, which connects the C terminus of β2m to the N terminus of the heavy chain (10). In some configurations, SCTs can have the format: secretion signal sequence—antigenic peptide—linker 1—mature β2-microglobulin—linker 2—mature class I heavy chain (
To characterize peptide binding by an SCT, we have focused most of our attention on the SCT construct of the well-characterized Kb/OVA complex. This SCT was found to have remarkable qualities including extended cell surface half-life and the ability to stimulate potently T cells that recognize native Kb/OVA (10) We also chose the Kb/OVA complex because of the large number of immunological reagents that are available. We found this SCT is 1000-fold more refractory to exogenous peptide binding compared to Kb loaded with endogenous peptides. However, the SCT was more susceptible to exogenous peptide binding compared to native Kb/OVA (11).
Initial studies with these prototypic SCTs have shown that when transfected they assemble rapidly in the ER, show extended surface half-life, and are refractory to binding of exogenous peptide in comparison to Kb loaded with endogenous peptides (10).
The ability of SCTs to retain binding of a single peptide has been exploited for the production of multimeric staining reagents, which have been used for T cell expansion and diagnostics (86-89). This property of SCTs also affords novel opportunities for the production of tumor vaccines due to the feet that some CTL epitopes recognized by cancer patents appear to bind their respective MHCs with relatively weak affinity (90). Furthermore, since the single-chain peptide is preprocessed and preloaded, an SCT expressed following vaccination with DNA comprising a sequence encoding the SCT can be a potent in vivo stimulator of CD8 T cells (13, 16), SCT technology is therefore a highly efficacious method for tumor vaccination. In strong support of this conclusion, incorporation of a tumor-specific peptide into an SCT was recently shown to elicit robust CTL responses and complete protection against a lethal tumor challenge in a mouse model (4, 16). Lastly, SCTs have been successfully used as novel probes for the study of NK cell biology (17, 18) and T cell activation (19).
In spite of these exciting and diverse applications of SCTs, certain limitations regarding their general applicability have been recently noted. One such limitation is that SCTs constructed with relatively low affinity peptides do not assemble as efficiently nor do they prevent the binding of exogenous peptide as effectively as SCT formed with relatively high affinity peptides. These limitations were presumed to reflect impaired F pocket anchoring resulting from the linker extension (11).
Although SCT tetramers retain the ability to bind antigen-specific T cells (11, 102, 103), there are certain limitations to using SCT material. First, excess peptide cannot be used during in vitro refolding of solubilized E. coli inclusion bodies of SCTs. This means that refolding efficiency may not be optimal for some SCT constructs. The second limitation is a problem with expression in E. coli. Endogenous bacterial methionyl aminopeptidase activity must cleave initiator N-formyl methionine (fMet) off of the SCT to reveal the correct antigen sequence for binding in the SCT groove. However, this cleavage event is not universal; it depends on the size of amino acid side chain which follows the initiator fMet (104, 105). Thus expression of SCTs for tetramers is limited to those that have antigenic peptide sequences that allow for fMet cleavage.
The present inventors disclose molecules designated herein as “disulfide traps.” A disulfide trap comprises an MHC antigen peptide covalently attached to an MHC class I heavy chain. The covalent linkage between the antigen peptide and the MHC class I heavy chain comprises a disulfide bond, which extends between a pair of oxidized cysteines (i.e., a cystine). The cysteines comprising the disulfide bond comprise a first cysteine and a second cysteine. A first cysteine, is comprised by a linker extending from the carboxy terminal of an MHC antigen peptide, and a second cysteine is comprised by an MHC class I heavy chain, in particular an MHC class I heavy chain which has a non-covalent binding site for the antigen peptide, in some configurations, a disulfide trap can comprise one contiguous polypeptide chain as well as a disulfide bridge. In other configurations, a disulfide trap can comprise two contiguous polypeptide chains which are attached via the disulfide bridge as the only covalent linkage.
In various configurations of a disulfide trap, the sequence extending from the carboxy terminal of the peptide can comprise at least one amino acid in addition to the cysteine, including one or more glycines, one or more, alanines, and/or one or more serines. In some configurations, the sequence extending form the carboxy terminal of the peptide can comprise a carboxy-terminal cysteine.
In various configurations, the second cysteine can be a mutation in the MHO class I heavy chain. In some aspects, the mutation can be a cysteine which substitutes for an amino acid of the MHC class I heavy chain, or a cysteine addition to the MHC class I heavy chain. In various configurations, the second cysteine can be situated from about 1 to about 100 amino acids from the amino terminal of the MHC class I heavy chain, in some aspects, the second cysteine can be a Y84C substitution (i.e., a substitution of tyrosine-84 of a MHC class I heavy chain with a cysteine). In other aspects, the second cysteine can be a T80C substitution (i.e., a threonine-80 to cysteine substitution) or an A86C substitution (i.e., an alanine-86 to cysteine substitution).
The present inventors have also developed disulfide trap single chain trimer (dtSCT) molecules. The inventors have found that a disulfide bond can effectively trap an antigen peptide in the class I groove of an SCT if the SCT comprises a first cysteine in a Gly-Ser linker extending between the C-terminus of the peptide and the β2-microglobulin, and a second cysteine in a proximal heavy chain position. In various configurations, a disulfide trap such as a dtSCT does not succumb to high concentrations of competitor peptide, even when the dtSCT is based on a low-affinity complex. The present inventors have also obtained similar results for dtSCT's comprising either Kb/OVA sequences or a second MHC allele, H-2Ld, known for its relatively poor pep tide-binding capacity (14, 15).
In some configurations, a disulfide trap can comprise any antigen that can bind a corresponding MHC class I heavy chain or MHC class I-like antigen presenting molecule such as CD1 (Altamirano, M. M., et. al., Proc. Natl. Acad. Sci. 98: 3288-3293, 2001). In some aspects, an antigen peptide sequence can be that of a peptide which can be presented by an MHC class I molecule, in various configurations, an antigen peptide sequence can comprise from about 8 to about 15 contiguous amino acids. In some configurations, the antigen peptide sequence can comprise 9 contiguous amino acids. In various aspects, a peptide sequence can be that of a protein fragment, wherein the protein is a pathogen protein or a cellular protein, such as, for example, a protein expressed by a cancer cell. In some aspects, an antigen can comprise an antigen peptide such as that of an HLA-A restricted peptide or HLA-B restricted peptide, such as an HLA-A*0201-restricted peptide. In some aspects, an antigen peptide can have a sequence as set forth in Table 1:
In some aspects, the MHC class I heavy chain can be an HLA-A, an HLA-B, an HLA-C, an HLA-E, an HLA-F, or an HLA-G class I heavy chain, and an antigen can comprise an antigen peptide corresponding to the MHC class I heavy chain, such as an HLA-A, an HLA-B, an HLA-C, an HLA-E, an HLA-F, or an HLA-G restricted peptide, respectively.
In some aspects, the present inventors have developed methods for producing disulfide trap molecules including dtSCT's. These methods include expressing a nucleic acid vector in a suitable host cell, wherein the vector comprises a promoter and/or an IRES operably linked to a sequence encoding a polypeptide comprising the primary amino acid sequence of a disulfide trap (including a dtSCT). Following translation of an mRNA, a disulfide bridge forms between the cysteines of the nascent polypeptide chain. A disulfide trap synthesized within a host cell can be recognized at the cell surface by both antibodies and T cells specific for the peptide-receptor complex. In other aspects, a disulfide trap can be produced by separately expressing a) an antigen peptide comprising an extension at its carboxy terminus, wherein the extension comprises a cysteine, and b) an MHC class I heavy chain comprising a cysteine.
In some aspects, the present inventors disclose vaccines against tumors and pathogens such as bacteria and viruses. These vaccines comprise either a disulfide trap, a nucleic acid comprising a sequence encoding a disulfide trap, or a combination thereof.
In some aspects, the present inventors disclose multivalent MHC/peptide complexes comprising a disulfide trap. In various applications, such multivalent complexes can be used for enumerating populations of T cells during infections or malignancies, or a probes for identifying cells such as subsets of T cells. Such probes can be used, for example, to monitor T cell-specific immune responses during infection or malignancy.
To extend the novel applications of SCT for vaccines and probes for pathogen surveillance, we took a structure-based approach to engineer three consecutive SCT designs that are characterized here biophysically and functionally in the context of their high resolution crystal structures. Progressive generations of SCTs show dramatic improvements in linker accommodation, C-terminal peptide anchoring, and improved refractoriness to exogenous peptide binding. These modifications in SCT design were made without disrupting the MHC fold as determined by structural comparisons, highly sensitive T cells assays, and expressing SCTs as tetramers to stain pathogen-specific T cells.
Although the data presented here use the well-defined Kb-Ova model system, SCT approaches clearly extend to other mouse and human class I peptide-MHC complexes. To date we and others have reported SCT constructs with human HLA-A2, -B27, -E or mouse H2-Kb, -Ld, -Db each bound by several different respective peptide ligands (18). This general applicability of the SCT format is in large part reflected by the highly conserved atomic basis of C terminal peptide anchoring. Thus we fully expect that the new SCT designs will also extend to other mammalian class I peptide-MHC complexes.
SCTs have now undergone three generations of design. In our initial protein engineering studies we determined the order in which to connect each component into the SCT and the optimal length of each spacer allowing efficient surface presentation (10). In the second generation we re-engineered the peptide-binding groove to accommodate the linker between the antigenic peptide and β2m (11). Lastly, in the third generation, disclosed herein, we introduce a disulfide trap. Without being limited by theory, a disulfide trap in the F pocket is believed to enhance the association of the peptide for the MHC within the SCT format. Our structural data show that the Y84A and Y84C mutations restructure the Kb groove to open a channel allowing the linker to freely extrude from the peptide-binding platform. In addition, experimental electron density maps for the SCTY84C-PBL2C protein show the presence of a disulfide bond between CysPBL2 and Cys84. We also present experimental evidence that the presence of this disulfide bridge increases the thermal stability of the SCT format and effectively prevents binding of competitor peptides. Consequently, such disulfide traps have many applications, such as, for example, as vaccines against pathogens and tumors, or for staining reagents to enumerate pathogen-specific T cells. In particular, disulfide traps are noteworthy for physiologically important antigenic peptides that are not tight binders.
We have tested the capacity of a disulfide trap to compensate for poor peptide binding, using a disulfide-trapped SCT constructed with a weakly binding variant of Ova (Ova5y, SIINYEKL) (50, 78) We show that introduction of the disulfide trap in the Ova5y-based SCT significantly improves its surface expression and completely prevents exogenous peptides exchange of the weaker Ova5y analog. These results indicate that the disulfide-trap is able to enhance the association of weakly binding peptides within the SCT format. In addition, disulfide traps were successfully employed to make a SCT with an Ld heavy chain and the QL9 peptide (31). Ld-QL9 complexes have poor surface stabilities and have proven to be difficult to refold as recombinant proteins. The disulfide-trapped form of Ld-QL9 is resistant to peptide competition akin to the one observed with Kb-based SCTY84C-PBL2C, indicating that disulfide trap technology can be applied to different MHC allelic forms, including ones with peptides of relatively poor binding affinities.
Induction of subdominant T cell responses has been shown to be critically important in producing a broadly effective immunity against tumors (84, 90) and pathogens (7, 9, 83). Such T cell responses are typically the result of less efficient antigen processing or weaker peptide-MHC association. Due to their preassembled nature SCT proteins bypass antigen processing, express efficiently at the cell surface, and exhibit extended cell surface half-fifes (10, 11). These properties make SCT proteins ideal for vaccine design. Recently, two studies on the use of SCT in DNA vaccinations for cancer prevention have been reported. In one of these studies a DNA construct encoding a breast cancer epitope derived from mammoglobin-A presented in the context of HLA-A2 as a SCT was used successfully in vaccination of doubly transgenic HLA-A2+/hCD8+ mice to induce the specific expansion of epitope-positive CTLs (16). While in the second study, C57BL/6 mice vaccinated with SCT proteins presenting an immunodominant epitope derived from HPV-16 E6 protein in the context of H-2Kb exhibited significantly increased E6 peptide-specific responses compared to mice immunized, with DNA encoding E6 protein. Moreover 100% of the mice vaccinated with the SCT encoding DNA were protected against lethal challenge with E6-expressing tumors (4). It should be noted that these vaccine applications were done with first generation SCTs. SCTs incorporating disulfide traps offer further advantages for vaccines due to their improved stability and refractoriness to exogenous peptide binding.
In a second application, disulfide traps can be used for multimeric staining reagents to enumerate antigen-specific, CD8+ T cells. Evidence for this conclusion is that tetramers with certain suboptimally binding peptides have been difficult to make or are particularly unstable once made (6). We disclose herein tetramers with only a disulfide-trapped peptide rather than an entire dtSCT. More specifically we determined that a soluble heavy chain with a Y84C mutation, produced in E. coli, can be refolded in the presence of a synthetic peptide with a GC extension. As reported herein, this approach can produce, in some configurations, disulfide trap tetramers, such as, for example, tetramers with disulfide-trap Kb-Ova. Such tetramers can stain T cells generated to the native Kb-Ova complex even in the context of a pathogen infection.
Some non-limiting advantages of using disulfide-trap tetramers over single-chain-trimer-(SCT)-based tetramers are disclosed herein. Although SCT tetramers retain the ability to bind antigen-specific T cells (11, 102, 103), there are certain limitations to using SCT material. First, excess peptide cannot be used during in vitro refolding of solubilized E. coli inclusion bodies of SCTs. This means that refolding efficiency may not be optimal for some SCT constructs. The second limitation is a problem with expression in E. coli. Endogenous bacterial methionyl aminopeptidase activity must cleave initiator N-formyl methionine (fMet) off of the SCT to reveal the correct antigen sequence for binding in the SCT groove. However, this cleavage event is not universal; it depends on the size of amino acid side chain which follows the initiator fMet (104, 105). Thus expression of SCTs for tetramers is limited to those that have antigenic peptide sequences that allow for fMet cleavage. Both of these limitations are obviated with the use of disulfide-trap technology. Excess peptide can be added to drive formation of the class I complexes during the in vitro refolding, and there is no limitations on the peptide sequence. This represents a more versatile approach for generating covalently linked MHC reagents, and can dramatically improve their shelf life and quality of staining, especially of those utilizing weaker-binding peptides. Such reagents allow for a more comprehensive T cell enumeration for diagnostics and study of T cell immune responses to pathogens and tumors.
The present inventors have constructed disulfide trap molecules, such as SCTs which incorporate a disulfide trap. In various configurations, these novel dtSCT constructs can better accommodate the linker in the SCT groove, and can restore F pocket anchoring. The crystal structures of these SCT constructs along with their functional characterization indicate they can be widely applicable over a broader range of peptide affinities, greatly facilitating SCT-based approaches for pathogen surveillance.
The methods and compositions described herein utilize laboratory techniques well known to skilled artisans and can be found in laboratory manuals such as Sam brook, J., et al. Molecular Cloning: A Laboratory Manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 2001; Spector, D. L. et al, Cells: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1998; and Harlow, E Using Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1999.
A disulfide trap disclosed herein comprises an MHC antigen peptide covalently attached to an MHC class I heavy chain. The MHC antigen peptide can be any MHC antigen peptide, such as, for example, an antigen peptide disclosed in U.S. patent application Ser. No. 11/397,377 filed Apr. 4, 2006. The covalent linkage between the antigen peptide and the MHC class I heavy chain comprises a disulfide bond, which extends between a pair of oxidized cysteines (i.e., a cystine). The cysteines comprising the disulfide bond can comprise a first cysteine and a second cysteine. A first cysteine can be comprised by an oligopeptide, comprising the antigen peptide sequence and, extending from the carboxy terminal of antigen peptide, an extension comprising a cysteine. In various configurations, the sequence extending from the carboxy terminal of the peptide can comprise at least one amino acid in addition to the cysteine, such as, for example, one or more glycines, one or more, alanines, and/or one or more serines.
In various configurations, a second cysteine can be comprised by an MHC class I heavy chain, in particular an MHC class I heavy chain which has a non-covalent binding site for the antigen peptide, such as an MHC class I heavy chain disclosed in U.S. patent application Ser. No. 11/397,377.
In various configurations, the second cysteine can be a mutation in the MHC class I heavy chain. In some aspects, the mutation can be a cysteine which substitutes for an amino acid of the MHC class I heavy chain, or a cysteine addition to the MHC class I heavy chain. In various configurations, the second cysteine can be situated from about 1 to about 100 amino acids from the amino terminal of the MHC class I heavy chain. In some aspects, the second cysteine can be a Y84C substitution (i.e., a substitution of tyrosine-84 of a MHC class I heavy chain with a cysteine). In other aspects, the second cysteine can be a T80C substitution (i.e., a threonine-80 to cysteine substitution) or an A86C substitution (i.e., an alanine-86 to cysteine substitution). In some configurations, a disulfide trap can further comprise a Y84A substitution (i.e., a tyrosine-84 to alanine substitution).
The present inventors have also developed disulfide trap single chain trimer (dtSCT) molecules. The inventors have found that a disulfide bond can effectively trap an antigen peptide in the class I groove of an SCT if the SCT comprises a first cysteine in the Gly-Ser linker extending between the C-terminus of the peptide and the β2-microglobulin, and a second cysteine in a proximal heavy chain position. A disulfide trap such as a dtSCT does not succumb to high concentrations of competitor peptide, even when the dtSCT is based on a low-affinity complex. Similar results have been for dtSCTs comprising either Kb/OVA sequences or a second MHC allele, H-2Ld, known for its relatively poor peptide-binding capacity (14, 15).
In some aspects, the present inventors have developed methods for producing disulfide trap molecules including dtSCT's. These methods include expressing a nucleic acid vector in a suitable host cell, wherein the vector comprises a promoter and/or an IRES operably linked to a sequence encoding a polypeptide comprising the primary amino acid sequence of a disulfide trap (including a dtSCT). Following translation of an mRNA, a disulfide bridge forms between the cysteines of the nascent polypeptide chain. A disulfide trap synthesized within a host cell can be recognized at the cell surface by both antibodies and T cells specific for the peptide-receptor complex.
The inventors herein disclose the engineering of a disulfide bond to lock peptide into the MHC class I binding groove. This disulfide trap was introduced into class I molecules expressed as single chain trimers, or SCTs. Some SCTs have been crystallized and high resolution structures analyzed. These structures verify the native class I conformation of the SCTs, as well as provide an anatomical basis for our biochemical and functional characterizations of the SCTs.
Some SCTs are class I MHC molecules expressed in the format: peptide—[G4S]3—mature β2m—[G4S]4—mature class I heavy chain (
Some other SCTs include the heavy chain substitution Y84A, which opens the peptide binding groove where the linker exits (11). The crystal structure of an SCT comprising Y84A shows that the linker adopts a more extended conformation compared to SCTs comprising Y84, In SCTs comprising Y84, the linker arches out and around heavy chain residues in the F pocket. This linker protrusion explains the improved detection of the latter by mAb 25-D1.16 (11) and by the TCR staining reagent m6α, in the case of the Ld SCT (
Although functional SCTs have been constructed from a number of different mouse and human class I MHC/peptide complexes, SCTs constructed with high affinity peptides are clearly superior. The reason for this is now clear. As noted above, SCTs have disrupted F pocket anchoring and are thus dependent upon rebinding of the linker-attached peptide. Consequently, SCTs made with lower affinity peptides have higher steady state levels of peptide-empty conformers at the cell surface (see, e.g.
The data presented herein confirm that the engineering of the disulfide bond was achieved at heavy chain position 84 and the second position on the linker extending from the C-terminus of the peptide. Migration of the dtSCT in non-reducing SDS-PAGE, as well as the ability of dtSCTs to exclude formidable concentrations of high-affinity competitor peptide, is consistent with formation of the disulfide bond in cells. This conclusion is further confirmed by electron density data from dtSCT crystals. In addition, we demonstrate herein that dtSCTs of two distinct MHC/peptide complexes are recognized by antibodies, TCR reagents, and T cells specific for native class I/peptide complexes. Accordingly, the disulfide-trap approach presented here befits other H-2 and HLA complexes. Moreover, we show herein that dtSCT construction offers great utility for studying lower affinity MHC/peptide complexes, as weak Kb/peptide and Ld/peptide complexes can both be expressed with a disulfide trap and, furthermore, exclude competitor peptides. As revealed by crystal structure, cysteine residues in the dtSCT provide not only a disulfide bond, but also better linker accommodation and, unexpectedly, additional hydrogen bonds. These hydrogen bonds improve peptide anchoring in the F pocket of the dtSCT groove, compared to SCTs that do not comprise a disulfide trap.
Recent, applications of SCTs have begun to provide unique insights into our understanding of the complex role of MHC class I molecules in lymphocyte development and function (17-19). In a seminal study of the role of class I in natural killer (NK) cell development, Kim et al. examined NK cells from mice that expressed the OVA.β2m.Kb SCT as a transgene, but not β2m or other class I genes (17). These experiments established that expression of a single MHC class I molecule selectively licenses NK functional activity only in NK precursors with an inhibitory receptor specific for this MHC class I molecule. The SCT also promises to play a pivotal role in studies of CD8 T cell development. To study the development of CD8 T cell in the context of a single class I/peptide complex, we have expressed the OVA.β2m.Kb SCT as a transgene and bred to a Kb-, Db-, β2m-deficient background. Importantly, the covalently attached OVA excludes the binding of endogenous peptides, as demonstrated by the strong primary CTL response of T cells from these mice when cultured with cells expressing Kb loaded with endogenous peptides. However, previous studies of CD4 T cell development established that it is imperative to absolutely ablate endogenous peptide binding in order to make definitive conclusions about the role of peptide in thymic selection (61-63). Thus, transgenic mice that express a disulfide-trap OVA.β2m.Kb SCT provide important tools for approaching a key outstanding question in αβ T cell development, i.e., the relationship between the MHC-bound selecting peptide in the thymus vs. the activating peptide in periphery.
Disulfide traps, including dtSCTs, have clinical applications. There are already four reports demonstrating that vaccination with plasmid DNA encoding an SCT leads to generation of specific antibodies and/or CTL. In the report by Yu et al., DNA encoding the OVA.β2m.Kb SCT was used to vaccinate BALB/c mice (10). BALB/c mice vaccinated with plasmid encoding SCT were found to generate anti-Kb/OVA antibody, demonstrating that the SCTs are expressed as intact structures by DNA vaccination. Furthermore, in a recent report by Primeau et al. (13), syngeneic immunization of plasmid DNA encoding SCT was found to elicit antigen-specific. CTLs. Extending these findings are two exciting recent studies showing that SCT DNA vaccination is highly effective for priming T cells in clinically relevant model systems. In one of these studies, mice doubly transgenic for HLA-A2 and human CD8 were vaccinated with DNA encoding an SCT that included HLA-A2 and a breast cancer-associated peptide derived from mammaglobin (16). This vaccination with SCT DNA was reported by Jaramillo et al. to induce a significant expansion of CTLs capable of specific detection of A2 positive human breast cancer cells. Another recent study by Huang et al. tested SCT DNA vaccination in a mouse model of human papillomavirus (HPV)-induced tumors such as cervical cancer (4). The HPV oncoprotein E6 is responsible for malignant transformation and is consistently expressed in HPV-associated tumors. Peng et al. identified an immunodominant CTL epitope, of the E6 protein that binds to the mouse class I molecule Kb (64) and this E6 epitope was then incorporated into an SCT (E6p.β2m.Kb). In this study, B6 mice were vaccinated with plasmid DNA encoding either the SCT or the intact E6 protein alone, DNA vaccination of B6 mice with the SCT elicited high levels of CTL, whereas DNA vaccination with only E6 displayed little response over background. Furthermore, DNA vaccination with SCT protected mice against a lethal tumor challenge, whereas vaccination with DNA encoding E6 or OVA-β2m-Kb SCT did not. These experimental findings represent the first evidence that SCTs can have a significant advantage over protein-based vaccine approaches in a clinically relevant model system. The success of SCTs was credited to the fact they are antigen processing-independent, do not need to compete with endogenous peptide, and are more stable at the cell surface due to their covalent nature (4). Each of these DNA vaccination approaches used first generation SCTs. Although SCT-based DNA vaccines show great promise, they can be improved even further by incorporation of the disulfide trap. Certain vaccines can elicit T cell responses to peptide determinants that bind relatively poorly to the MHC, in which case the use of dtSCTs can have a significant impact on the stability of antigenic determinants, and thus the level of vaccine protection.
Disulfide traps such as dtSCTs can be used as probes for cell staining. For example, a dtSCT multimer can be used as a high sensitivity FACS staining reagent, in comparison, tetramers for several defined CD8 T cell epitopes have been difficult to construct or are very unstable once constructed (65). Furthermore, in vivo applications of recombinant MHC molecules to modulate CD8 responses can require reagents with longer in vivo half-lives and greater thermal stability (66-68). dtSCT crystal structure reveals that an analogous disulfide trap can be incorporated into soluble, recombinant class I molecules without the SCT linkers. This can be done by refolding the class I heavy chain harboring a Y84C mutation with recombinant β2m and a synthetic peptide having a GC extension. This approach works for at least Kb/OVA complexes: the disulfide trap forms properly upon in vitro refolding, and dtMHC tetramers bind T cells specific for the native complex. Accordingly, the present teachings include the construction of stable multimers of lower-affinity MHC/peptide complexes and make possible more comprehensive analyses off cell responses to pathogens and tumors.
The following examples are illustrative, and are not intended to be limiting.
The following materials and methods were used in the examples presented herein.
Cell Lines and Antibodies—Triple knockout fibroblasts (Kb−/− Db−/− β2m−/−; 3KO) are a transformed murine embryo fibroblast line derived from 3KO obtained from Stephen Jennings, (Louisiana State University Health Sciences Center). LM1.8 cells are L cells (H-2k) transfected with ICAM (22) and were obtained from Dr. Phillipe Kourilsky (INSERM, Institut Pasteur, Paris, France). Monoclonal antibodies (mAbs) include the following: B8-24-3, which recognizes folded Kb (American Type Culture Collection); Y3, which recognizes folded H-2K molecules (23); 25-D1.16 (a gift of Dr. Jonathan Yewdell, National Institutes of Health), which recognizes Kb+SIINFEKL peptide (24); and 64-3-7, which recognizes open (peptide-free) forms of Ld and other class I molecules tagged with the 64-3-7 epitope (25).
In some experiments. HeLa cells (human cervical carcinoma) transfected with 64-3-7 epitope-tagged HLA-B27 were used (37). Transfection of HeLa cells with cDNAs encoding A2 and B27 SCTs was performed using FuGene6 (Roche Diagnostics, Indianapolis, Ind.). Stable transfectants were selected and maintained in 0.6-1.2 mg/mL geneticin (Life Technologies, Grand Island, N.Y.), identified by flow cytometry, and cloned by limiting dilution. All cells were maintained in RPMI 1640 (Life Technologies) supplemented with 10% FCS (HyClone Laboratories, Logan, Utah), 2 mM L-glutamine, 0.1 mM nonessential amino acids, 1.25 mM HEPES, 1 mM sodium pyruvate, and 100 U/ml penicillin/streptomycin (all from the Tissue Culture. Support Center, Washington University School of Medicine, St. Louis, Mo.).
In some experiments, antibodies mAb BB7.2 and mAb ME-1 were used for immunoprecipitation and flow cytometry. BB7.2 is directed against folded A2 molecules (109), while ME-1 is directed against folded B27 (110). mAb HCA2 (a gift of H. Ploegh, Harvard Medical School, Boston, Mass.) recognizes unfolded A2 molecules (111). Open/peptide empty conformers of human class I molecules or SCT constructs were also detected by mAb 64-3-7 epitope-tagging (25, 31, 54-56). This antibody was also used for immunoprecipitations and immunoblots of epitope-tagged A2 and B27 and the corresponding SCT constructs.
Peptides—The OVA257-264 peptide (SIINFEKL) (26), SIYR peptide (SIYRYYGL) (27), VSV8 peptide (RGYVYQGL) (28), the QL9 peptide (QLSPFPFDL) (29-31), and the MCMV pp89 peptide (YPHFMPTNL) (32) were synthesized on an Applied Biosystems Model 432A peptide synthesizer. Peptides were dissolved directly in culture media and incubated with cells for 5 hours or overnight before cytofluorometry or CTL assays.
DNA Constructs and Retroviral Transduction—Constructs were generated using standard techniques and confirmed by DNA sequence analysis. SCTs such as the OVA.β2m.Kb SCT sequence were described previously (10, and U.S. patent application Ser. No. 11/397,377 which is incorporated by reference herein in its entirety), and the QL9.β2m.Ld SCT was constructed, similarly as follows. In amino-to-carboxy terminal order, the SCT comprises a secretion signal sequence such as that of β2mb, followed by an antigen peptide sequence, then a first linker of 15 residues, (G4S)3. An alanine residue can occupy the fourth position of the first linker (see Table 2), which can correspond to a convenient restriction site in a DMA construct encoding the SCT. This first linker is followed by the mature β2-microglobulin β2-mb sequence, the second linker of 20 residues, (G4S)4, then the mature class I sequence. A dtSCT (such as QL9.β2m.Ld SCT) can also comprise a Y84A mutation, first described in the OVA.β2m.Kb SCT (11). This mutation and the cysteine mutations shown in Table 2 can be introduced by site-directed mutagenesis (QuikChange II XL, Stratagene).
Retroviral expression vectors pMSCV-IRES-hygromycin and pMSCV-IRES-neomycin were constructed in our lab (33) and used for concomitant expression of class I SCTs and drug resistance genes. Retrovirus-containing supernatants were generated using the Vpack vector system (Stratagene, La Jolla, Calif.) for transient transfection of 293T cells to generate ecotropic virus for infection of 3KO and B6/WT3 cells or amphitropic virus for infection of LM1.8 cells.
In some experiments, PCR-generated inserts with appropriate DNA restriction sites flanking mature HLA-A0201 (A2) and -B2705 (B27) heavy chain sequences replaced the mature H2-Kb sequence of the previously constructed murine SCT (10) in the pIRESneo vector (Clontech Laboratories, Palo Alto, Calif.). The mature human β2m sequence was spliced into these constructs in place of the mouse β2m sequence. Thus, HLA SCTS retained the subunit order and linker lengths [G4S]3 and [G4S]4 found in the original SCT design (10) (
SCT Expression and Purification—The SCT constructs (
The formed SCT proteins were purified on a Superdex75 (Pharmacia, Piscataway N.J.) size exclusion column using a running buffer containing 20 mM HEPES pH 7.5, 150 mM NaCl, and 0.01% NaN3. The fractions containing the refolded proteins were pooled, diluted threefold in buffer containing 20 mM Tris pH 8.5, loaded on an anion exchange MonoQ column (Pharmacia, Piscataway N.J.), and eluted with a NaCl gradient (0 mM to 400 mM NaCl over 30 ml). Prior to crystallization the purified SCT proteins were exchanged in buffer containing 20 mM HEPES pH 7.5, and 20 mM NaCl. Typically 800 μg of purified complex were obtained from a 500 ml refolding reaction.
Flow Cytometry—Viable cells, gated by forward and side scatter, were analyzed on a FACSCalibur (BD Biosciences), and data (10,000 events per sample) were analyzed using CELLQuest software (BD Biosciences). Staining with anti-class I mAbs was visualized using phycoerythrin-conjugated goat anti-mouse IgG (BD Biosciences). Soluble, recombinant m6α T cell receptor (TCR) was used for flow cytometry directly from the supernatant of insect cells (34-36). Secondary reagents used to detect binding of the m6α TCR included biotinylated anti-mouse TCR β chain monoclonal antibody H57 (BD Biosciences) and phycoerythrin-conjugated streptavidin (BD Biosciences). In some experiments, Flow cytometric analyses were performed using a FACSCalibur (BD Biosciences, San Jose, Calif.). Dead cells and debris were excluded from analysis on the basis of forward-angle and side-scatter light gating. A minimum of 10,000 gated events was collected for analysis. Data were analyzed using Cell Quest software (BD Biosciences). For surface staining, 5×105 cells per sample were incubated on ice in microliter plates with culture supernatant from the appropriate hybridoma. After washing. PE-conjugated goat anti-mouse IgG (BD Pharmingen, San Diego, Calif.) was used to visualize class I staining. In some experiments, flow cytometry followed incubation (1×106 cells/mL) with exogenous peptides TAX (A2-specific) (107) or MCMV (H2-Ld specific) (32) at indicated concentrations.
Immunoprecipitation and Western blotting—Immunoprecipitations were performed essentially as described previously (20). Briefly, cells were treated with 1.0% NP-40 (IGEPAL CA-630, Sigma) lysis buffer including 20 mM iodoacetamide and protease inhibitors. Postnuclear lysates were mixed with EDS sample buffer (Invitrogen), and 2-mercaptoethanol (Sigma) was either added to a final concentration of 1%, or excluded for nonreducing gels. Western blotting was performed after SDS-PAGE separation of precipitated proteins as described previously (37), Rabbit serum for blotting Kb was generated in our lab by peptide immunization using an amino acid sequence from the cytoplasmic tail of Kb. Biotin-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch) was used as a secondary staining reagent followed by streptavidin-horseradish peroxidase (Zymed Laboratories). Specific proteins were visualized by chemiluminescence with the ECL System (Amersham Biosciences). In some experiments. Cells were lysed in PBS+1.0% NP-40 (Sigma, St. Louis, Mo.), 20 mM iodoacetamide, protease inhibitors, and a saturating concentration of precipitating mAb. After lysis for 30 min on ice, post nuclear lysates were incubated with protein A-Sepharose (Amersham Pharmacia Biotech, Uppsala, Sweden) for 1 h. Beads were washed four times in PBS+0.1% detergent, and bound proteins were eluted by boiling in 1× SDS-PAGE sample buffer. For nonreducing gels, 2-ME was excluded. For some samples, endoglycosidase H treatment followed immunoprecipitation. Bound proteins were eluted from protein A-Sepharose by boiling in 10 mM TrisCl, pH 6.8+0.5% SDS+1% 2-ME. Eluates were mixed with an equal volume of 100 mM sodium acetate, pH 5.4, and either digested (or mock-digested) at 37° C. for >1 h with 1 mU endoglycosidase H (ICN Pharmaceuticals. Costa Mesa, Calif.) that was reconstituted in 50 mM sodium acetate, pH 5.4. In some experiments, Immunoblotting was performed following SDS-PAGE separation of precipitated proteins and transfer to Immobilon P membranes (Millipore, Bedford, Mass.). Membranes were blocked (1 h to overnight) with PBS+10% dried milk+0.05% Tween 20. Primary Abs were added and incubated for 1 h, followed by washing in PBS+0.05% Tween 20. As a second step, membranes were incubated for 1 h with biotin-conjugated anti-mouse Ig (Caltag Laboratories, San Francisco, Calif.). In some cases, biotin-conjugated mAb 64-3-7 was used, which obviated the second step. After washing, HRP-conjugated streptavidin (Zymed Laboratories) was added for 1 h, followed by three washes. Specific proteins were visualized by chemiluminescence using the ECL system (Amersham, Boston, Mass.).
CTL assays—LM1.8 target cells transduced with the indicated class I constructs were incubated with or without 10 μM exogenous peptide and then labeled with Na51CrO4 for 1 hour. OT-1 or 2C T cells were plated at various concentrations onto 96-well microtiter plates and incubated with target cells for 4 h at 37° C. in 5% CO2. Radioactivity in supernatants was measured in an Isomedic γ-counter (ICN Biomedicals). The mean of triplicate samples was calculated, and percentage 51Cr release was determined as follows: % 51Cr release=100×((experimental 51Cr release−control 51Cr release)/(maximum 51Cr release−control 51Cr release)), where experimental 51Cr release represents counts from target cells mixed with effector cells, control 51Cr release represents counts from target cells in medium alone, and maximum 51Cr release represents counts from target cells lysed with 5% (v/v) Triton X-100 (Sigma).
In some experiments, HeLa cell targets expressing A2 and B27 SCTs were tested against specific CTL lines. Protocols for the generation of CTLs and chromium-release assays to measure specific lysis have been described (112). The RR10 CTL was generated from a patient immunized with G280-9V pulsed DC as described (113). A CMV pp65-specific CTL line was generated from a CMV seropositive healthy donor using purified CD8+ T cells, peptide-pulsed irradiated autologous DC, and 10 U/mL IL-2 added on day 2. Primed CD8+ T cells were repeatedly stimulated with antigen to generate CTL lines. Peptide competition assays were performed by incubating targets with competitor peptide at the indicated concentrations. On the following day, targets were trypsinized, washed twice, and used in a standard chromium release assay.
N15 hybridoma assays—The TCR− murine T cell hybridoma 58α−β− (38) transfected with the α and β chains for the N15 TCR was obtained from Dr. Hsiu-Ching Chang (Dana-Farber Cancer Institute, Harvard Medical School). This cell line is also transfected with CD3ζ and CD8αβ cDNAs to optimize TCR expression and recognition of class I (39). Biological activity of IL-2 secreted by the M15 hybridoma was used to detect Kb/VSV8 complexes.
In some experiments, B6/WT3 cells or 3KO cells (5×104/200 μl/well) expressing endogenous Kb or Kb SCT constructs were incubated for 1 hour at 37° C. with the indicated concentrations of VSV8 peptide in a flat-bottom 96-well plate. The cells were washed, fixed in 1% paraformaldehyde for 15 minutes at room temperature, then washed again. The N15 hybridoma cells (105/200 μl/well) in fresh media were then added to the plate and cultured for 24 hours. Supernatants were harvested and frozen at −80° C. for at least 1 hour to lyse trace cells that may have carried over. CTLL-2 cells were washed and added to supernatants at a final cell density of 104/200 μl/well in a 96-well plate. After 18-24 h of incubation, Alamar blue (BioSource International) was added at 20 μl/well, and relative amounts of IL-2 in each supernatant were determined by fluorescence on a multi-detection microplate reader (Bio-Tek Instruments).
In some experiments, LM1.8 cells (2×104/200 μL/well) expressing native Kb or the SCT constructs were incubated for 24 hours in a 96-well plate with N15 hybridoma cells (2×105/200 μL/well) in the continuous presence of the indicated concentrations of VSV8 (RGYVYQGL) peptide. Supernatants were harvested and frozen at −80° C. for at least 1 hour to lyse trace cells that may have carried over. CTLL-2 cells were washed and added to supernatants at a final cell density of 5×103/200 μL/well in a 96-well plate. Triplicate wells were run for each sample. After 18-24 hours of incubation, Almar blue (BioSource International, Camarillo Calif.) was added at 20 μL/well, and relative amounts of IL-2 in each supernatant were determined by fluorescence in a multi-detection microplate reader (Bio-Tek Instruments, Winooski Vt.).
Thermal Denaturation Studies Purified SCT proteins were dialyzed against 10 mM K2HPO4/KHPO4 buffer pH=7.5, and 150 mM. NaCl. Protein concentration was determined by absorption at 280 nm of denatured protein in buffer containing 8M Urea, 25 mM K2HPO4/KHPO4, pH=7.5. SCT proteins were diluted to 475 nM in 20 mM K2HPO4/KHPO4 buffer pH=7.5, and 150 mM NaCl. CD spectra at 10° C. were collected between 260 and 200 nm at 0.5 nm increments on Jasco-8.10 instrument equipped with Peltier temperature controller using 1-cm path-length cuvettes. Four spectra were averaged for the final spectrum of each sample. The thermal denaturation profiles for each SCT protein and native H2 Kb-Ova were monitored by the change in CD signal at 220 nm as a function of temperature. Thermal scan data were collected at a 1.0° C. interval from 10 to 70° C. with a temperature ramp rate at 50° C./hour. All measurements were made at least four times and averaged. Thermal denaturation curves were scaled from 0% to 100% to provide plots of the percent of signal change versus temperature. The T1/2 is the temperature at which the CD signal change is half of the total signal change.
Crystallization and Data Collection—Diffraction-quality crystals of each of the SCT proteins were obtained by hanging-drop, vapor-diffusion method. Protein at 6 to 8 mg/ml was equilibrated at 20° C. against 13% (w:v) PEG 10,000 and 100 mM MES pH 6.2 for SCTWT; 14% PEG 6,000 and 1.00 mM MES pH 6.4 for SCTY84A; 13% PEG 6,000 and 100 mM MES pH 6.3 for SCTY84C-PBL2C. Small crystals obtained in these drops were used to microseed protein hanging drops equilibrated against similar conditions with marginally lower concentrations of PEG. Larger crystals appeared overnight and grew over three to four weeks. Just prior to data collection, crystals were briefly soaked in crystallization buffer to which Ethylene Glycol was added to a final concentration of 20% (v:v) as cryoprotectant for liquid nitrogen flash cooling. X-ray diffraction data for each SCT crystal were collected on the ID-19 beamline at APS. A total of 360 frames were collected for each SCT crystal each frame representing a 1.0° oscillation range. These data were indexed and integrated using DENZO (HKL Suite, HKL Research, Inc. Charlottesville Va.) in the primitive monoclinic lattice for all SCT proteins with nearly identical cell dimensions (Table 3), and scaled and merged using SCALEPACK (HKL Suite, HKL Research, Inc., Charlottesville Va.). Wilson scaling was applied to the final output structure factor amplitudes [Collaborative Computing Project 4 (CCP4), Daresbury Laboratory, Warrington UK (92)].
Structure Determination and Refinement Initial phase estimates for each SCT protein were obtained by rigid body refinement of the atomic coordinates of H-2Kbm8 (PDB 1RJY) against each of the data sets. Extensive model building was performed with the macromolecular modeling program O (O version 6.22, Uppsala Software Factory, Sweden) using 2FOFC, FO-FC, and 2FO-FC composite omit maps [CMS, Yale University, New Haven Conn. (93)]. Atomic refinement was done employing simulated annealing, energy minimization, and restrained B-factor refinement protocols as implemented in CNS and CCP4. Refinement statistics of the final models are listed in Table 3.
Computational Analysis—Graphical structure representations were primarily created using Ribbons (94). Molecular surfaces of the peptide-binding grooves were generated using InsightII (Biosym Technologies, San Diego Calif.). Superpositions and r.m.s.d, calculations between the different MHC proteins were obtained using Lsqkab (CCP4). R.m.s.d. values between the aligned peptide main chain, and side chain atoms were calculated using CNS (93). HBPLUS (95) was used to catalogue contacting atoms and putative hydrogen bonds.
OT-1 CTL assays LM 1.8 target cells transfected with the indicated constructs were incubated with or without 10 μM exogenous peptide and then labeled with Na51CrO4 for an hour. OT-1 cells were plated at various concentrations onto 96-well microliter plates and incubated with target cells for 4 hours at 37° C. In 5% CO2. Radioactivity in supernatants was measured in an Isomedic γ-counter (ICN Biomedicals, Costa Mesa Calif.). The mean of triplicate samples was calculated, and percentage of 51Cr release was calculated as follows: % of 51Cr release=100×(experimental 51Cr release—control 51Cr release)÷(maximum 51Cr release−control 51Cr release). Experimental 51Cr release represents counts from target cells mixed with effector cells, control 51Cr release represents counts from target cells in medium alone, and maximum 51Cr release, represents counts from target cells lysed with 5% (v:v) Triton X-100 (Sigma. St. Louis Mo.)
Tetramer construction—Peptides SIINFEKL and SIINFEKLGC were used for H-2 Kb and H-2KbY84C heavy chains respectively. Irrelevant Kb peptide SIYRYYGL (27) was used for construction of negative control H-2Kb tetramers. Peptides were obtained commercially (EZBiolab Inc., Westfield Ind.) or synthesized on an Applied Biosystems Model 432A peptide synthesizer. The BirA biotinylation sequence was included at the C-terminus of the H-2Kb ectodomain. In addition. H-2Kb and H-2KbY84C tetramer constructs included the mutation C121S, which removes the unpaired cysteine of Kb and thus simplifies its refolding and purification with disulfide-trapped peptide. Refolding and purification were performed as described above. At this point, samples were submitted for mass spectrometric analysis to confirm integrity and identity of the recombinant MHC molecules. Biotinylation with BirA ligase was carried out overnight at ambient temperature according to the manufacturer's instructions (Avidity, Boulder Colo.). To remove free biotin, MHC molecules were again purified by SEC, then tetramerized by addition of fluorochrome-conjugated streptavidin at a molar ratio of 4 molecules of MHC to 1 molecule of either phycoerythrin-(PE-)streptavidin or allophycocyanin-(APC-)streptavidin (both from BD Biosciences. San Jose Calif.). The MHC concentration of all stock tetramer preparations was 4.3±0.2 μM, and tetramers were used at identical dilutions for staining experiments. The specificity of each of the tetramers was confirmed using positive-control transgenic T cells.
Infection of mice with Listeria and tetramer staining—B6 mice were infected by tail vein injection of 5×103 cfu of Listeria monocytogenes expressing ovalbumin (L.m.-Ova, a kind gift of Hao Shen, University of Pennsylvania) in a volume of 200 μL pyrogen-free saline. After 7 days, splenocytes were harvested and red blood cells lysed. Splenocytes were stained with the indicated tetramers for 30 minutes on ice, then with FITC-conjugated anti-CD8 (BD Biosciences) for 20 minutes. Propidium iodide staining was used to exclude dead cells. Splenocytes were also prepared from (uninfected) OT-1 transgenic mice and stained in the same manner. Cells were analyzed on a FACSCalibur (BD Biosciences), and data were analyzed using CELLQuest software (BD Biosciences). More than 15,000 CD8+ events were collected per staining in order to reliably determine the frequency of tetramer-positive CD8 T cells.
This example illustrates disulfide bond engineering strategy.
As in our previous characterization of SCT molecules (10, 11), we again chose the Kb/OVA system, for which there are excellent reagents to monitor both MHC conformation and peptide occupancy (24, 40). Our approach was to introduce cysteine residues by site-directed mutagenesis, one in the heavy chain of Kb and one in the first linker (
To select the exact positions for cysteine mutations, we superimposed the crystal structures of Kb/OVA (41) and the MHC class II molecule 1-Ek bound to a hemoglobin peptide (Hb64-76) (42). This latter structure features the Hb64-76 peptide covalently attached with a Gly-Ser linker to the β chain of class II (43, 44). The linker extends from the open class II groove from the C-terminal residue of the Hb64-76 peptide. Thus, overlaying these structures provided a framework for determining distances between Kb heavy chain residues and specific positions along the first SCT linker (
This example illustrates characterization of SCT cysteine variants
Retroviral transduction was used to introduce each of the SCT variants into Kb−/− Db−/− β2m−/− (3KO) fibroblasts (20). This permitted the unambiguous analysis of the transduced class I constructs. Cell surface expression of SCTs with disulfide traps was also confirmed in wildtype cells (see
Our next objective was to determine biochemically which constructs actually had the additional disulfide bond, for it was unknown a priori how engineered cysteine residues would interact with ER thioreductases (46-48). In non-reducing SDS-PAGE, the presence of disulfide bonds can reduce the radius of gyration of denatured proteins, causing them to migrate faster through the gel matrix. Thus the variant OVA.β2m.Kb SCT constructs were compared in reduced vs. non-reduced SDS-PAGE gels (
For a number of reasons, we focused our subsequent analyses on the Y84C, L2C disulfide bond position. First of all, we had previously engineered position 84 in the SCT heavy chain, and T cells recognized the Y84A mutant of the SCT, perhaps even more efficiently (11). Secondly, position 84 is highly conserved, increasing the likelihood that engineering at that position would easily transfer to other class I molecules (see
This example illustrates design of some HLA SCT molecules.
In some experiments, three different SCTs were used, each with disease relevance. The first SCT was constructed with the peptide, G280-9V (YLEPGPVTV), bound to A2. This peptide is derived from the melanocyte-specific gp100 protein, which is expressed in most melanomas (106, 114). Adoptive transfer of cultured tumor-infiltrating lymphocytes specific for gp100 has been shown to cause tumor regression (90). The second SCT was constructed with the peptide, FAX (LLFGYPVYV) also bound to A2. This peptide is derived from p40tax protein, the major antigenic protein of the human T-cell leukemia virus (HTLV), which causes adult T-cell leukemia (107, 115). The third SCT was constructed with the peptide, NP383-391, (SRYWAIRTR) bound to B27, and is a fragment of the influenza A nucleoprotein (108). These antigens were chosen for their predominant roles in CTL responses to melanoma, HTLV, or influenza, respectively. However, the HLA-B27 allele is also of interest due to its association with arthritic disease (116). SCTs were constructed for each of these peptide/MHC complexes to test efficiency of cell surface expression, folding of the molecules (by monitoring conformational changes serologically), and T cell recognition.
This example illustrates efficient folding, cell surface expression, and T cell recognition of HLA SCTs
In these experiments, HeLa cells were transfected with native HLA heavy chains or SCTs of the complexes A2/G280-9V, A2/TAX or B27/NP383-391. The SCT constructs exhibited high-level surface expression, comparable to the native HLA molecules, as detected in flow cytometry experiments using mABs specific for the respective folded HLA molecules (
This example illustrates the level of steady-state fldig of the HLA SCTs.
In these experiments, to quantify the amount of open MHC conformers, the 64-3-7 epitope tag was introduced into both native B27 and the NP383-391.β2m.B27 SCT (25, 37, 54-56). HeLa cells expressing these constructs were lysed and analyzed by immunoprecipitation and blotting (
This example illustrates that the SCT peptide moieties are not proteolytically cleaved and “re-presented” in the context of the HLA SCT binding groove.
In these experiments, we found that the G280-9V.β2m.A2 SCT migrated more slowly than the β2m.A2 covalently linked dimer when analyzed by immunoprecipitation and blotting, (
This example illustrates that HLA SCTS are recognized by cytotoxic T cells specific for these clinically relevant HLA complexes.
In these experiments, targets expressing the NP383-391.β2m.B27 SCT (
This example illustrates that HLA SCTs are susceptible to peptide exchange.
In these experiments, we first measured the ability of the TAX.β2m. A2SCT to exclude a competitor peptide with comparable affinity for A2. Increasing concentrations of the CMV pp65 peptide were incubated with TAX.β2m.A2SCT-expressing targets, which were then subjected to a cytotoxicity assay using TAX- or pp65-specific CTLs (
This example illustrates serological and biochemical analyses of disulfide-trap HLA SCT.
In these experiments, mAb 64-3-7 was used to monitor the peptide binding state of MHC class I molecules serologically. This mAb defects a peptide-empty or “open” conformation of class I heavy chains that bear the appropriate epitope for this reagent (25, 37, 54-56). The conserved sequence recognized by 64-3-7 is naturally found in H2-Ld, but a single point mutation (R48Q) was all that was required to recreate the 64-3-7 epitope in HLA-A2. We tested the functionality of the 64-3-7 epitope in HeLa cells expressing the β2m.A2 dimer with the intent to use this approach later to measure the open conformers of A2 SCTs with and without the disulfide trap.
Hence, in some experiments, HeLa cells expressing 64-3-7 epitope-tagged β2m.A2 were incubated overnight with the A2-binding TAX peptide, control peptide, or no peptide. Specific peptide induced expression of folded A2 conformers was defected by BB7.2 (
This example illustrates cell surface levels of open and closed conformers of epitope-tagged G280-9V.β2m.A2 SCT with or without the disulfide trap.
Using flow cytometry, we compared cell surface levels of open and closed conformers of 64-3-7 epitope-tagged G280-9V,β2m.A2 SCT either with or without the disulfide trap (
This example illustrates formation of a disulfide trap.
To directly test whether the disulfide trap was formed, a biochemical approach was taken. Considering the placement of the engineered disulfide bond, the radius of gyration of the denatured SCT molecule is expected to be significantly shortened under nonreducing SDS-PAGE conditions, which would result in faster migration through polyacrylamide. This was indeed the case (
This example illustrates T cell recognition of the dtSCT
Given that T cells are highly sensitive to perturbations in the MHC/peptide structure and/or conformation, it was imperative to determine whether the mutations in the dtSCT constructs affected T cell recognition. The dtSCT and each of the previous versions of the OVA.β2m.Kb SCT were transduced into LM1.8 cells (H2k fibroblast expressing ICAM) and tested in chromium-release assays as targets for Kb/OVA-reactive OT-1 T cells (
As further evidence of the integrity of the interaction between TCR and SCT molecules, we have found that SCT tetramers and conventional tetramers detect the same polyclonal antigen-specific T cells after pathogen infection. The same holds true for tetramers built with the engineered disulfide-trap—they retain the ability to bind the same T cells as conventional tetramers.
This example illustrates both T cell recognition of the disulfide-trap HLA SCT and exclusion of competitor peptides
In these experiments, to determine whether the disulfide-trapped SCT retains its native conformation for antigen presentation, a CTL line specific for the A2/G280-9V complex was used (
This example illustrates that a disulfide trap functionally retains an antigenic peptide.
In these experiments, to test whether the disulfide trap functionally retains the antigenic peptide, a T cell assay was used to monitor the intrusion of competitor peptides into the SCT peptide binding groove. For these assays, a CTL line specific for A2/CMVpp65 was used to monitor exogenous peptide binding by the G2809V.β2m.A2SCT (
A remarkable finding is the difference the disulfide trap made to the susceptibility of the G2809V.β2m.A2SCT to binding exogenous pp65 (
This example illustrates that dtSCT molecules exclude high affinity competitor peptides.
A crucial test of the utility of the dtSCT was to measure its capacity to exclude competing peptides. We had previously found that the SCT format without the disulfide trap excluded high-affinity competitor peptides to a great extent (10), about 1000 times more effectively than native Kb loaded with endogenous peptides. This exclusion of competitor peptides was enhanced by the Y84A mutation in the SCT that opens the groove and allows for better linker accommodation (11). To test the dtSCT construct, we incubated cells expressing the dtSCT with increasing amounts of exogenous high-affinity competitor, the Kb-binding SIYR peptide (27). Reactivity of mAb 25-D1.16 was used to specifically monitor the loss of Kb/OVA complexes. In control cells expressing the Y84A SCT, half of the Kb/OVA complexes were displaced by the addition of 1.25 μM competitor (
However, this assay does not detect very low levels of competitor binding, which could be sufficient for T cell recognition. Thus, as a complementary approach, we also monitored exogenous peptide binding using a gain of recognition T cell assay. For this assay we used the highly sensitive T cell hybridoma N15. The N1.5 hybridoma, derived from the TCR murine hybridoma 58α-β- (38), expresses the N15 TCR, which is specific for the VSV8/Kb complex, as well as CD8 α and β chains (39). As shown in
The example illustrates that a poor-binding OVA analog can be tethered to the MHC groove using the dtSCT approach.
The above results demonstrate that a dtSCT can exhibit stronger binding of the covalently attached peptide compared to an SCT without the disulfide trap. Accordingly, a dtSCT can be used to enhance presentation of relatively poor-binding peptides. As a demonstration, we investigated the previously described OVA analog SIINYEKL, or OVAp5Y which binds poorly to Kb (50-52). Howarth et al. showed that the OVAp5Y analog was still recognized in the context of Kb by the 25-01.16 mAb (50). Furthermore, they demonstrated that the surface half-life of Kb/OVAp5Y complexes was reduced 3-fold and efficiency of expression more than 75-fold compared to native Kb/OVA complexes. Therefore, using this OVA analog as a model of a poor-binding peptide for Kb, we constructed an OVAp5Y.β2m.Kb SCT and an OVAp5Y.β2m.Kb dtSCT (i.e., with and without a disulfide trap, respectively (Table 2)). Although the steady-state cell surface expression level of the OVAp5Y SCT was approximately 10-fold lower than the OVA-based SCT, the surface expression of OVAp5Y SCT was improved three-fold by the addition of the disulfide trap, demonstrating that the disulfide trap facilitates de novo folding and/or increases cell surface stability of SCTs constructed with weak-binding peptides. In peptide competition experiments, the OVAp5Y.β2m.Kb Y84A SCT was readily displaced with the SIYR peptide (
The Kb SCT and dtSCT constructed with the weak-binding OVA analog were also tested in N15 hybridoma assays (
This example illustrates that a second, low-affinity H-2 complex can be secured when expressed as a dtSCT.
To test the applicability of the dtSCT technology on an additional class I/peptide complex, and to further test its ability to enhance presentation of a lower affinity complex, we employed the Ld/QL9 complex. The QL9 peptide (QLSPFPFDL) is derived from an endogenous dehydrogenase and binds to the FI-2Ld (29-31), a class I allele which is unique due to its relatively weak, association with peptide and β2m (15, 53). Importantly for this study, the Ld/QL9 complex is relatively unstable based on its limited cell surface half-life (about ½ hr) and the difficulty of constructing Ld/QL9 tetramers. Thus, the Ld/QL9 complex was ideal for testing the stabilizing effects on class I molecules conferred by the single-chain format itself and by the introduction of a disulfide bond. A second advantage of studying the Ld molecule is the ability to measure directly relative amounts of peptide-occupied versus peptide-empty class I conformers at the cell surface with mAb 64-3-7, which specifically detects open (peptide-empty) class I molecules (25, 31, 54-56). A third reason to construct SCTs of the Ld/QL9 complex was its well-defined reactivity to the 2C TCR (31, 57, 58) and the availability of a recombinant high-affinity 2C-derived mutant TCR designated m6α (34-36), which is an excellent peptide-specific Ld/QL9 staining reagent for flow cytometry.
In these experiments native Ld and QL9.β2m.Ld SCTs with or without the disulfide trap were subcloned individually into a retroviral vector for expression in B6-derived fibroblasts. As a control, a QL9.β2m.Ld SCT with the Y84A mutation was also included. This mutation opens the groove and reduces binding of exogenous peptides (11). As shown in
Peptide occupancy determines the half-life of class I molecules at the cell surface (59). Ld, being a relatively poor peptide binder, is thus highly inducible at the cell surface by incubation with exogenous ligands (15). Given this we reasoned that susceptibility to peptide induction would be informative for comparing the relative peptide occupancy of the Ld/QL9 SCT complexes. As shown in
We next wanted to extend these findings with Ld/QL9 SCTs with more physiologically relevant TCR based assays. It should first be noted that 2C T cells detected target cells expressing QL9.β2m.Ld dtSCT or QL9 peptide-fed targets comparably (
The strong reactivity of m6α TCR for the dtSCT allowed us to determine the extent to which disulfide trapping of the QL9 peptide to Ld excluded competitor peptides relative to native Ld/QL9 complexes. Cells expressing native Ld were fed exogenous QL9 peptide overnight. The next morning these cells, along with cells expressing each generation of QL9.β2m.Ld SCT, were incubated with increasing concentrations of competitor peptide YPHFMPTNL from murine cytomegalovirus (MCMV) pp89 (32), which binds to Ld with similar affinity to QL9 (60). The cells were then stained with recombinant m6α TCR and analyzed by flow cytometry to specifically detect loss of the Ld/QL9 epitope (
This example illustrates SCT design rationale
In these experiments, the initial SCT, designated here SCTWT, was designed by Sinking the C terminus of Ova peptide (SIINFEKL, Ovalbumin, residues 257-264) using a (Gly4-Ser)3 linker to the N terminus of β2m and linking the C terminus of β2m to the N terminus of the Kb heavy chain using a (Gly3-Ser)5 linker (
As a structural framework for the design of the next SCT generations we overlapped the peptide-binding platform of the class II MHC I-Ek protein bound to a hemoglobin-derived peptide onto the α1α2 domain of Kb-Ova (42,96). The I-Ek peptide is attached to the class II β chain via a flexible linker analogous to the one used in our SCT design. Based on the superposition of the peptide-binding platforms the Hb peptide and its linker were docked onto the peptide-binding groove of Kb. This modeling exercise allowed us to determine which residues in Kb impede optimal linker accommodation. We selected Tyr84 and mutated it to Ala (
Next, we sought to introduce a stabilizing disulfide bridge between the peptide and the MHC to replace the loss of C-terminal anchoring. Our overlap model allowed us to determine potential distances between the linker and the MHC for disulfide bond engineering. Several disulfide bridges were introduced involving the conserved MHC residues Thr80, Tyr84, and Asn86 and the first four linker residues. Each of these constructs were introduced in Kb−/−, Db−/−, and β2m−/− (3KO) fibroblasts (20) and their cell-surface expression was detected by flow cytometry. In addition, in vivo disulfide bond formation for each construct was also confirmed. For our structural and functional characterizations, we selected a SCT construct with a disulfide between a cysteine introduced in place of Tyr84 and a cysteine introduced at the second position in the peptide-β2m linker (
This example illustrates structure determination.
In these experiments, each of the SCT constructs (
Each of the SCT molecules crystallized in the primitive monoclinic space group, P21, with two molecules per asymmetric unit and nearly identical cell dimensions (Table 3). Initial phase estimates were obtained by rigid-body refinement of the coordinates of H-2Kbm8 (PDB 1RJY, peptide and water molecules omitted), which was crystallized in the same space group and unit cell dimensions. After initial refinement easily interpretable electron density was seen for the Ova and the linker residues immediately C-terminal Ova that improved upon further building and refinement cycles. Diffraction data to 2.00 Å (SCTWT), 2.00 Å (SCTY84A), and 1.80 Å (SCTY84C-PBL2C) were used for refinement of the final atomic models, which have R factors of 21.5% (Rfree 25.3%) for SCTWT, 21.0% (Rfree 25.2%) for SCTY84A, and 20.8% (Rfree 23.9%) for SCTY84C-PBL2C.
The electron density maps for each of the SCT molecules were of good to excellent quality. No ambiguities were seen for the main chain and side chains of the Ova peptide in each of the complexes. Interpretable density was seen for all of the residues in the first linker in the SCTWT molecule, for residues 1 through 7 and 11 through 15 in the first linker of SCTY84A, and residues 1 through 7 and 12 through 15 in the first linker of SCTY84C-PBL2C. In each SCT this linker does not adopt any secondary structural elements and is seen in different conformations in each molecule in the asymmetric unit. Clear evidence for disulfide bond formation was seen in both molecules in the asymmetric unit for SCTY84C-PBL2C. No interpretable electron density was seen for the β2m-heavy-chain linker and for residues 278 through 280 at the C terminus of the heavy chain in any of the single-chain structures.
This example illustrates structural similarities between SCTs and Native Kb-Ova.
In these experiments, structural alignment of all of the atoms of α1α2 domains of the SCTs to those of native Kb bound to Ova yielded overall pair-wise r.m.s.d values of 1.14 Å (Cα r.m.s.d.=0.58 Å) for SCTWT, 1.32 Å (Cα r.m.s.d.=0.77) for SCTY84A, and 1.32 (Cα r.m.s.d, 0.80) for SCTY84C-PBL2C indicating that only minimal structural perturbations of the antigen presenting platform are created in each of the SCT proteins. These initial results let us to believe that each of the SCT proteins would retain their native fold in vivo and that the introduced engineering perturbations would minimally affect T cell recognition. To test this hypothesis each of the SCT constructs was introduced into the LM1.8 cell line and tested as targets in chromium-release assay against the Kb-Ova reactive OT-1 T cells (
To visualize any structural differences between the SCTs and native Kb we aligned the main chain residues of the peptide binding platforms of all four structures and this alignment is shown as a ribbon diagram in B. Most of the main chain differences are found in solvent-exposed, flexible loops and the C-terminal regions of the α1α2 domains. Some minor differences were also seen in the C-terminal and N-terminal portions of the α1 and α2 helix, respectively with SCTY84C-PBL2C showing the greatest deviation from Kb-Ova. In light of our functional data, however, we believe that these differences have little is any affect of TCR binding.
We also compared the conformations of Ova in each of the SCT constructs to that of Ova bound to native Kb.
This example illustrates accommodation of linker residues—MHC class I proteins typically present antigenic peptides of 8-10 residues in length. This is in part due to the confined nature of their peptide binding grooves. To engineer a class I MHC that would favorably accommodate a C-terminal peptide linker, the conserved Tyr84 residue was mutated to an alanine to form SCTY84A in order to open the Kb-groove. In SCTY84C-PBL2C we mutated this residue to Cys to maintain the open Kb groove and to establish a disulfide bridge between the MHC and a Cys residue introduced in the second linker position.
To visualize the structural changes in the binding grooves of each SCT construct we calculated solvent accessible surfaces (97) for a spherical probe with a radius of 1.4 Å for their α1α2 domains (
This example illustrates F-Pocket Hydrogen Bonding Interactions
One of the hallmarks of peptide binding by MHC class I proteins is the anchoring of the terminal regions of the peptide through conservative hydrogen bonding networks in the A and F pockets of the MHC (99). Since the SCT format and the mutations that we introduced could closely impact on anchoring of the C-terminus of the peptide we analyzed the bonding established in the F pockets of the SCT proteins (
This example illustrates stability of SCTY84C-PBL2C over SCTWT and SCTY84A
An important goal of SCT design was to achieve a peptide-MHC assembly that is kinetically stable and resists peptide exchange by competitor peptides. Previous characterization of SCT constructs has shown that SCTWT proteins excluded tight-binding competitor peptides to a greater extent than native Kb (10) loaded with endogenous peptides and this exclusion was enhanced by the Y84A mutation (11). However, the Ova portion of both SCTWT and SCTY84A was competed by high concentrations of endogenous peptide. To similarly test the SCTY84C-PBL2C protein, exogenous peptide binding to this construct was monitored by using a gain of recognition T cell assay (
We also tested the thermostability of purified, recombinant SCTs by following their denaturation profiles as a function of temperature using CD spectroscopy (
This example illustrates generation of disulfide-trapped staining reagents.
In these experiments, generation of peptide-MHC tetramers in which the peptide is permanently attached to the class I heavy chain through a disulfide trap without a β2-microglobulin (β2m) is demonstrated. To generate these complexes, Kb heavy chain carrying the Y84C mutation and a C-terminal biotinylation sequence was refolded with β2m and a modified Ova peptide. This peptide was C-terminally extended to include a Gly-Cys sequence for disulfide bridge formation analogous to the one designed in SCTY84C-PBL2C. The formed disulfide-trap MHC complex was purified using size exclusion and anion exchange chromatographies. The formation of the engineered disulfide was confirmed by non-reducing SDS-PAGE and electrospray mass spectrometry of the intact disulfide-trap complex, which revealed the expected masses of unattached β2m and a single mass peak of the Kb heavy chain disulfide bonded to the extended Ova peptide. The pure complex was biotinylated at the C-terminal end of the heavy chain by incubation with BirA ligase. This biotin-labeled product was incubated with chromophore-conjugated streptavidin to generate the staining tetramer through the tight association of streptavidin to four biotinylated MHC proteins.
To demonstrate the feasibility and validity of the disulfide-trap tetramers, we tested their T cell binding specificity. The disulfide-trap tetramers were compared directly with conventional Kb-Ova tetramers at the same molar concentration of MHC. First, we confirmed tetramer binding to T cells from OT-1 transgenic mice (
This example illustrates that T cells that respond to the native class I complex also recognize the disulfide bond engineered class I tetramers.
In these experiments, B6 mice were, infected with a strain of Listeria monocytogenes that expresses ovalbumin. During this infection the mice mount a vigorous response in which a large fraction (5-10%) of CD8 T are specific for the Ova peptide presented by Kb (101). To determine the extent of overlap in CD8 T cell populations that recognize each type of MHC complex, tetramers of both conventional and disulfide-trap configurations were used to stain splenocytes from infected mice (
All publications, patents, patent applications and other references cited in this application are herein incorporated by reference in their entirety as if each individual publication, patent, patent application or other reference were specifically and individually indicated to be incorporated by reference.
90, β (°)
/R
)
Statistics as defined in SCALEPACK
Values in parentheses are for data in the highest resolution shell
Statistics as defined in CNS
R.m.s.d. values calculated based on an alignment of main-chain atoms of the indicated complexes or domains using Lsqkab in CCP4
indicates data missing or illegible when filed
This application claims priority to Provisional Application Ser. No. 60/840,521 filed on Aug. 28, 2006, which is incorporated herein by reference in its entirety.
This work was supported at least in part by National Institutes of Health grants AI055849 and AI27568.
| Number | Date | Country | |
|---|---|---|---|
| 60840521 | Aug 2006 | US |
