WO2011147894A1 - Chimeric mhc class ii proteinpeptide - Google Patents

Abstract

The invention relates to a chimeric protein, which comprises an α-chain and ß-chain of a MHC-class II protein, a linker and an epitope of interest, wherein the epitope is linked to the α-chain via the linker.

Description

CHIMERIC MHC CLASS II PROTEINPEPTIDE

DESCRIPTION

The invention relates to a chimeric protein comprising, an a-chain and β-chain of a MHC-class II protein, a linker and an epitope of interest, wherein the epitope is linked to the a-chain via the linker. Furthermore, the invention also relates to dimers and multimers of the chimeric protein and to modified antigen presenting cells.

BACKGROUND OF THE INVENTION

The central function of the immune system is to remove pathogens and eliminate diseased cells. The immune system can only function properly, if antigens associated with offending cells and microbes are distinguishable from antigens normally present in the body. The interaction between classes of molecules present on cellular elements of the immune system, including B lymphocytes, T lymphocytes, and professional antigen presenting cells ("APCs") such as macrophages and dendritic cells is responsible for the distinction between self and non-self.

T cells are mainly involved in cell mediated immunity whereas B cells are involved in the generation of antibody-mediated immunity. The two most important types of T cells involved in adaptive cellular immunity are a3-CD8+ cytotoxic T lymphocytes (CTL) and CD4+ T helper lymphocytes. CTL are important mediators of cellular im- munity against many viruses, tumors, some bacteria and some parasites because they are able to kill infected cells directly and secrete various factors which can have powerful effects on the spread of infectious organisms. CTLs 5 recognise epitopes derived from foreign intracellular proteins, which are 8-10 amino acids long and which are presented by class I major histocompatibility complex (MHC) molecules (in humans also called human lymphocyte antigens - HLAs). T helper cells enhance and regulate CTL responses and are necessary for the establishment of long-lived memory CTL. They also inhibit infectious organisms by secreting cytokines such as IFN-y. T helper cells recognize epitopes derived mostly from extracellular proteins which are preferably 12-25 amino acids long and which are presented by class II MHC molecules. B cells, or more specifically the antibodies they secrete, are impor- tant mediators in the control and clearance of mostly extracellular organisms. Antibodies recognise mainly conformational determinants on the surface of organisms, for example, although sometimes they may recognise short linear determinants.

The central event of T cell mediated adaptive immunity is the interaction of the T cell receptor ("TCR"), which is expressed on the surface of T lymphocytes and elements of the major histocompatibility complex ("MHC"). In order to respond to a non-self antigen, a T cell must encounter the antigen in the context of a self MHC molecule. There are two classes of MHC molecules, referred to as "class I" and "class II" antigens. Cytotoxic T lymphocytes (which bear CD8 surface antigen, and are referred to as "CD8+ T lymphocytes") will only kill cells bearing a foreign antigen in the context of a self MHC class I molecule. Analogously, helper T lymphocytes (which bear CD4 surface antigen, and are referred to as "CD4+ T lymphocytes") will only proliferate in response to a foreign antigen in the context of a self MHC class II molecule on the surface of an APC. MHC class I molecules are comprised of a heavy a-chain and a 32-microglobulin light chain. In contrast, MHC class II molecules are heterodimers comprised of o and β-chains, each having two domains and being of approximately the same length. The DNA regions have been mapped e.g. for mouse and man. The human MHC genes are also designated as the "HLA complex" (for Human Leukocyte Anti- gen). The genes encoding class I molecules are at the A, B and C loci in man and the K, D, and L loci in mouse. Class II molecules are encoded at the DP, DQ and DR regions in man and the l-A and l-E regions in mouse. At each region, a multitude of alleles have been identified.

While class I MHC molecules typically present peptides derived from proteasomal degradation of cytoplasmic proteins, the majority of the ligands of class II MHC (MHC II) stem from extracellular pathogens.

After internalization (e. g. by endocytosis), the pathogens are fragmented by proteases and loaded onto class II MHC molecules in a process catalyzed by the chaper- one HLA-DM. Processing and loading take place in a dedicated endosomal com- partment (MCll vesicle), that also harbour MHC class II molecules which at this point are bound to peptides derived from the invariant chain (also termed li). These li derived peptides (also termed CLIP) vary in length between 14 and 23 amino acids and are bound promiscuously and with varying affinities to distinct MHCII alleles. Exchange of these so-called placeholder ligands against pathogen-derived peptides of higher affinity is catalyzed by the MHC-like molecule HLA-DM, an exchange catalyst with an optimum activity at the acidic pH of the vesicle. However, the structural mechanism underlying CLIP peptide exchange is still incompletely understood, due to a lack of information on the dynamics of conformational equilibria at the atomic level. This contrasts the ample structural information on stably formed class II peptide complexes with more than thirty unique class II peptide complexes already deposited at the protein data bank. Evidence is mounting that CLIP function extends beyond its placeholder properties: stable MHC ll-CLIP complexes are up-regulated on dendritic cells upon maturation and modulate the activation of T cells specific for exogenous foreign antigen. The continued presentation of CLIP on the surface of antigen presenting cells requires the active maintenance of peripheral tolerance against self. MHC alleles susceptible to autoimmune diseases, in contrast, are associated with low affinities for, and thus low surface levels of, CLIP. Taken together, the presence of the self-peptide CLIP on the cell surface seems to shape the T-cell repertoire in vivo, which, as observed for other peptides as well, has additionally been shown to be influenced by N- terminal length variations. The MHC II— peptide binding mode might account for these observations. Peptides are assumed to bind in a conserved N- to C-terminal orientation by two governing principles: (i) a set of hydrogen bonds between the peptide backbone and MHC side chains; and (ii) the geometrically and chemically favorable accommodation of four to five peptide side chains by defined MHC surface pockets. The observation that sin- gle peptide-MHC complexes can elicit functionally distinct immune responses has been attributed to distinct peptide binding registers or the existence of different conformational isomers of the MHCII itself. However, because X-ray crystallography might not capture the dynamic rearrangements that accompany alternative complex formation, the molecular basis for these findings is still poorly understood. After loading, the MHC class II molecules are transported to the cell surface and presented to CD4+ T cells, which upon recognition induce the antigen-specific immune response. On the cell surface some of these peptides are displayed for up to 100 h. A considerable fraction of the peptide-MHC molecules, however, dissociates during presentation, so that also "empty" MHC molecules are always present on class II MHC expressing cells. As a safeguard mechanism class II MHC molecules inactivate rapidly after dissociation of the ligand. They convert into a "non-receptive" state, which is characterized by the inability of the MHC molecule to bind new pep- tide ligands.

The interaction between MHC-peptide complexes expressed on antigen presenting cells and T cell receptors expressed on T cells leads to various T cell functions including proliferation and cytokine secretion, differentiation toward various cell subsets, anergy and apoptosis (see Davis et al. (1998) AmJ. Rev. Immunol. 16:523- 544). Various attempts have been made to mimic these immunomodulatory effects with soluble MHC-peptide complexes. For example, MHC molecules have been extracted from cell membranes and subjected to peptide elution followed by exchange for specific peptides in vitro. (Nag et al. (1996) Cell. Immunol. 170:25-33). Also, MHC molecules have been produced recombinantly and then loaded with peptides in vitro (Abastado et al. (1995) J. Exp. Med. 182: 439-447). Other attempts include the production of genetically engineered, covalently-linked peptide/ MHC chimeras (U.S. Pat. No. 5,869,270). A major disadvantage of the state of the art is that they are recognized by cognate TCRs with low avidity.

Prior to the present invention, MHC-peptide complexes have proven to be difficult to produce and, therefore, expensive to make in large quantities. The peptides, derived synthetically or in vivo, did not bind to MHC in a manner that allowed for a stable and reliable complex required for medical uses. Furthermore, the peptide is only bound in canonical orientation. Expression constructs are described in the state of the art, where a chimeric MHC-peptide molecule is expressed. However, the peptide is coupled to the beta-chain of the MHC molecule and not to the alpha-chain. MHC- peptide complexes are currently produced by isolating MHC protein from large numbers of cells presenting membrane-bound MHC protein and mixing in solution the isolated MHC protein with purified peptides. Separation of the MHC protein is a laborious, multi-step process and requires a large number of cells to obtain sufficient amounts of MHC protein. In addition, considerable effort is required to obtain purified peptide to mix with the MHC protein. Moreover, the association between individual peptides and MHC has been shown to be unstable. Particular MHC-peptide complexes and methods of making them have been suggested by various investiga- tors, including US 5,260,422 or US 5,194,425. However, they have only disclosed the use of soluble, as opposed to membrane bound, MHC-peptide complexes. Moreover, methods to produce such complexes suffered from the unpredictable and unstable association of peptides with MHC. Furthermore, it is not described in the state of the art, that a peptide can be bound to the MHC class II molecule in an inverted orientation. The peptide is always coupled to the beta-chain of the MHC class II molecule.

SUMMARY OF THE INVENTION

It was an objective of the invention to provide an alternative presentation of anti- genie peptides by MHC class II molecules.

This problem is solved by the features of the independent claims. Preferred embodiments of the present invention are provided by the dependent claims.

The invention concerns a chimeric protein comprising, an a- and β-chain of an MHC-class II protein, a linker and an epitope of interest, wherein the epitope is linked to the a-chain via the linker. It was very surprising, that the invention allows the inverted presentation of epitopes or antigens or peptides by MHC class II proteins, which is due to the binding of the epitope to the a-chain. The term "epitope" refers to a peptide or peptide fragment or protein or protein segment or in general to a amino acid sequence, that is bound in the MHC class II binding pocket and pre- sented to T cells. The state of the art assumes that the epitopes or peptides are loaded into the binding pocket of the MHC class II molecule in only one orientation. Therefore it was very surprising, that the loading can occur in the inverted orientation as well (inverted to the normal orientation described in the state of the art). The peptide inversion leads to stark differences in the display of a peptide or a peptide fragment or a epitope or a amino acid sequence pointing in direction of the T cell receptor. The MHC class II complex comprising the inverted epitope is not the same as the normal MHC class II complex, where the epitope is bound in the normal orientation. Consequently, different T cells are activated and different T cell clones can be selected, which will lead to an increased number of T cell variants recognizing the epitope. The epitope, which is loaded in the inverted orientation by binding the epitope at the a-chain of the MHC class II complex, can lead to different biological effects. Skewing of the orientational epitope repertoire could consequently affect the observed modulation of the CD4+ T cell population. A further clear indication for a potential role of the orientational equilibria is that the terminal expansions of T cells are relevant in regard to the maintenance of the critical hydrogen bonds that determine epitope orientation. Hence, distinct subtypes of T cells are produced by epi- tope variants very likely to be bound in distinct orientations in regard to the MHC class II molecule.

Furthermore, several autoimmune antigens display a certain degree of symmetry, a prerequisite for orientational diversity. For example, two antigenic peptides of the autoimmunity inducing myelin basic protein (MBP) are of the sequence

85'ENPVVHFFKNIVTP'98 or 88VHFFKNIVPRTP 9. Molecular modeling predicts that the individual MHC binding pockets can well accommodate the 85-98 peptide in both orientations. In contrast, the 88-99 variant should preferably bind in the non- canonical orientation, since only then the critical hydrogen bond network can be maintained. Consequently, the two MBP peptides evoke distinct T cell subsets and show that a potentially reoriented (=inverted) MHC class II binding peptide

(=epitope) can induce the production of autoreactive T cells. While different registers might well contribute to the observed T cell diversity, it is proposed that reorientation of peptide plays an equally important role in engaging different T cell receptors.

Furthermore, the invention will also be important for practical aspects of repertoire analysis and peptide design. In so far as peptides fulfill the key requirement of hydrogen bond conservation and as long as anchor residues fit with similar geometry into the individual pockets, peptides might be bound in either orientation. The predictive power of MHC class II binding prediction servers is moderate compared to class I binding. Taking into account orientational inversion can now be used to reassess existing programs and to design novel search algorithm that take into account peptide inversion. Furthermore peptides can now be designed that exclusively bind in either of the two orientations, potentially leading to distinct T cell responses within disease-relevant settings.

In a preferred embodiment, the epiotope is an autoreactive peptide or a binding variant thereof. The following are particular non-limiting examples of peptides and proteins that comprise epitopes relevant to particular autoimmune diseases. Where specific peptides are recited, since the epitopes are contained within the recited peptides, the boundaries of the peptides may be extended by amino acid sequences occurring in the parent protein, for example by at least five amino acids. Further, the following peptides may be comprised in larger peptides, provided that they retain at least some (preferably, at least 25 percent) of the immunogenic activity which they exhibit in the recited peptides. The present invention provides for the use of peptide fragments of proteins toward which an autoimmune response is generated in an autoimmune disease. Such peptide fragments are between 5 and 200, preferably between 5 and 100, and more preferably between 5 and 30 amino acids in length. Where a peptide is found to comprise an epitope, it is understood that the determination of he minimum amino acid sequence comprising the epitope could be achieved using standard techniques, so that peptide fragments of those recited herein may be used according to the invention. Similarly, it is understood that epi- tope-containing peptides may be comprised in larger peptides, either naturally occurring or engineered. The protein is preferably linked to the amino terminus of the a-chain. It is also preferred that the linker is a chemical crosslinker or a linker pep- tide.

The peptide linker preferably comprises a repetitive (Gly₄Ser)_n sequence

(n=preferably 2 or 3) but may contain any other natural or unnatural amino acid. In one class of molecules a protease cleavage site (thrombin, TEV, prescission protease) is introduced into the linker to allow peptide cleavage and release at a later time point. In another class of molecules, the peptide is linked after in vivo folding or in vitro refolding of the MHC-peptide complex by sortase mediated reaction. Here the MHC-interacting peptide - or its modified variants containing fluorescence or Magnetic Resonance active moieties - is synthesized or recombinantly expressed with a C-terminal LPETG sequence to allow sortase-mediated ligation of the N- terminus of the a-chain that has been modified to contain three N-terminal glycines. The overall length of the linker will be preferably in the range of 12-18 amino acids or span a distance of 25-40 A in case of non-amino acid based linkers.

The chemical crosslinker is preferably a polyalkylene glycol, preferably polyethylene glycol. Coupling of polyethylene glycol (PEG) to biologically active molecules termed "PEGylation" is used in the delivery of biologically active molecules, usually proteins and small molecules. PEGylation increases the size and molecular weight of proteins and small molecules resulting in the extension of their half-life in plasma. In general, PEGylation may alter the physicochemical properties of the proteins and therapeutic molecules resulting in decreased bioactivity of the parent proteins and organic therapeutic molecules. Polyalkylene glycols and the use thereof to modify proteins are known in the art. In accordance with the present invention, a difunc- tional polyalkylene glycol, preferably a difunctional polyethylene glycol (PEG) is used to attach the epitope to the amino terminus of the a-chain of the MHC class II molecule or to cross link the chimeric molecule to form a dimer or multimer.

The molecular weight of the PEG is not limited but is preferably from 1000 to 5000, and more preferably about 3500. Diamino-PEG having a molecular weight of 3400 is preferred in accordance with the present invention. The chimeric molecules may be covalently or noncovalently linked through enzymatic methods known to the skilled person. Also, leucine zippers or other dimeriza- tion domains can be utilized for linkage as it is known for the skilled person.

Furthermore, it is preferred to form a dimer comprising two chimeric proteins. Preferably such hetero- or homodimers are joined together into multimers of at least three chimeric proteins. Especially preferred are tetramers. This may be accomplished by fusing one or both MHC class II components to a multimerizing structure, which tends to form multimers with similar or dissimilar structures. In particular nonlimiting embodiments, one unit comprises at least a portion (comprising at least one constant region domain) of an immunoglobulin constant region to act as a mul- timerizing structure. Preferably, an immunoglobulin hinge region or portion thereof may be incorporated between the constant region portion and the MHC class II chain portion. The result is a dimer or multimer of chimeric proteins. It is itself referred to herein as a dimer because it consists of two like units bound together and functionally has two MHC class II epitope structures for associating with TCRs. Such multimers may be joined together to form larger complexes.

The chimeric proteins are preferably joined together by disulfide bonds. However, chemical crosslinking can also be preferred.

The invention also relates to an expression construct comprising: a. a promoter sequence, b. a nucleotide sequence encoding the a-chain and β-chain of a MHC class II protein, c. a nucleotide sequence encoding an epitope, and d. a nucleotide sequence encoding a linker, whereby the promoter sequence is operably linked to the sequence encoding the o chain or β-chain and wherein the sequence encoding the epitope is linked to the sequence encoding the amino terminus a-chain through the linker sequence.

Furthermore, the expression construct preferably contains a nucleotide sequence encoding a Kozak sequence (for eucaryotic expression, a nucleotide sequence en- coding a targeting signal for the endoplasmic reticulum (for eucaryotic expression), a nucleotide sequence encoding a tag, e. g. a Flag-tag, a nucleotide sequence encoding a trans-membrane sequence (for eukaryotic expression), a nucleotide sequence encoding an N-terminal affinity or labelling tag.

The invention also relates to a vector, or use thereof, which has been modified so that a DNA molecule that encodes a construct has been inserted into the vector. This construct is subsequently expressed from the vector (post-transduction), so that the epitope and the MHC class II molecule is presented on the surface of a transfected or transformed cell cells. The invention therefore relates to modified antigen presenting cells that induce immune tolerance, or repression of an immune response, towards the antigen of interest, which is encoded by a nucleic acid sequence that has been inserted into the vector as described herein. The modification of the vector with the epitope/MHC class II molecule-encoding nucleic acid molecule can occur via methods known to one skilled in the art, such as recombinant DNA technology, involving for example standard cloning techniques such as restriction digestion and ligation or recombination-based approaches used to insert the antigen-coding DNA. References in this document to vectors which encode and express an antigen to be presented relate to vectors that have been modified, so that a DNA sequence, which encodes an antigen of interest, can been inserted.

The invention can preferably be used for in vivo or in vitro or in situ expression. Furthermore, the invention relates to a modified antigen presenting cell, characterised in that the cell is expressing the construct, wherein the epitope of interest is presented via the MHC class II molecule on the surface of the modified antigen presenting cell. The construct or a derivative thereof, which encodes an MHC class II molecule and an epitope of interest, is preferably be used for the transformation or transfection of antigen presenting cells and the production of modified antigen-presenting cells.

It is preferred to use the modified cell for the induction of immune tolerance and/or the suppression and/or inhibition of immune reactions, preferably antigen-specific immune reactions directed against the antigen encoded by the exogenous nucleic acid and/or unwanted adaptive immune reactions mediated by CD4+ T cells, CD8+ T cells, B cells and/or antibodies.

Furthermore, it is also preferred to use the modified cell for the labelling and/or selection of T cells. The ability of a chimeric molecule to selectively bind to a T cell of for example a patient may be determined by standard laboratory methods which detect T

cell/molecule binding. As one example, binding of the chimeric molecule to a T cell may be detected by detectably labelling a component of the chimeric molecule, and then detecting the presence oft he label, e.g., labelling the chimeric molecule with a radioactive label, a fluorophore, a fluorescent label or any other marker known in the arts. The invention also relates to a pharmaceutical composition comprising modified antigen presenting cells according and at least one pharmaceutically acceptable carrier. The chimeric molecules oft he invention may also be administered in a suitable pharmaceutical carrier. The chimeric molecules oft he invention may be stored frozen and/or in ly- ophilized form prior to use. The chimeric molecules of the invention may be comprised in formulations which further include stabilizing agents.

The invention also concerns an amino acid sequence and corresponding polypeptide encoded by the nucleic acid molecule and/or sequence according to SEQ ID NO. 1 to SEQ ID NO. 27, fragments or parts of the sequences that preferably relate to MHC class II coding regions, linker, targeting and antigenic sequences that code for functional pro- teins, and/or homologues thereof. In another embodiment of the invention the amino acid sequence and corresponding polypeptide molecule comprises an additional analogue amino acid or another structure (lipid and/or sugar residues).

According to another preferred embodiment of the invention, the amino acid sequence encoding a peptide is selected from the group comprising: a) a peptide or a peptide encoded by the nucleic acid sequence according to SEQ ID NO 1 to SEQ ID NO. 27 or part thereof; b) a peptide comprising an amino acid sequence having sufficient homology to be functionally analogous to an amino acid sequence in accordance with a); c) a peptide according to an amino acid sequence a) or b) which is modified by deletions, additions, substitutions, translocations, inversions and/or insertions and functionally analogous to an amino acid sequence in accordance with a) or b).

In a distinctive embodiment of the invention the peptide which has sufficient homology to be functionally analogous to a peptide or amino acid sequence encoded by the nucleic acid sequence or amino acid sequences according to SEQ ID NO 1 to SEQ ID NO. 27 or part thereof has at least 40% homology thereto.

In another preferred embodiment said amino acid sequences have at least 60%, preferably 70%, more preferably 80%, especially preferably 90% homology to a peptide or amino acid sequence encoded by the nucleic acid sequence or a nucleotide acid se- quence according to SEQ ID NO 1 to SEQ ID No. 27 or part thereof.

Methods for the inhibition, blocking, suppression, activation or over-expression of an antigen-presenting molecule or co-stimulatory molecule involved in presentation of antigen are known to one skilled in the art. For example blocking or suppression of an antigen-presenting molecule or co-stimulatory molecule involved in presentation of antigen could be carried out via an antibody targeting the protein, genetic modification of the cell, such as deletions or other targeted mutations, or the use of anti-sense RNA approaches, such as siRNA, in order to silence the protein before translation of its mRNA. The activation or over-expression could be achieved via administration of activating antibodies, the administration of active protein to the cell or by over expression of the par- ticular protein in an expression vector, or by transformation of any other exogenous nucleic acid coding for the protein to be increased. The invention also relates to a nucleic acid sequence, wherein the sequence is selected from the group consisting of: a) a nucleic acid molecule according to SEQ ID NO 1 to SEQ ID NO. 27 or complementary nucleotide sequences thereof, b) a nucleic acid molecule hybridizing with a nucleotide sequence according to a) under stringent conditions, c) a nucleic acid molecule comprising a nucleotide sequence having sufficient homology to be functionally analogous to a nucleotide sequence according to a), d) a nucleic acid molecule which, as a consequence of the genetic code, is degenerated into a nucleotide sequence according to a), and e) a nucleic acid molecule in accordance with a nucleotide sequence according to a), which is modified and functionally analogous to a nucleotide sequence according to a) as a result of deletions, additions, substitutions, translocations, inversions and/or insertions.

The invention is also transferable to the design of a peptide library genetically linked to an MHCII a-chain. This includes a proper relation to nucleotide sequences created by limited/restricted random mutagenesis which will be inserted into suitable expression vectors, e.g. pCDNA-based vectors for eukaryotic expression or the pET vector series for expression in E. coli. Similarly, a functional system for presentation of the library by phage will be part of the invention. Phage display libraries encoding degenerate, inversely bound peptides enable the biopanning against purified or cell-bound T cell receptor complexes. In particular, a stably with the library transfected cell line -insect cells, fibroblasts etc.- will be modified to function as quasi-antigen presenting cell, meaning the co-expression of co-receptors will allow detection by and stimulation of T cells.

A peptide library will ideally yield inverted peptide orientations when presented by one of the alleles listed below. Potential applications for libraries are in the diagnostic field as they might serve in identifying pre-selected/-enriched T cell clones in relation to particular diseases. Possible read outs for stability of individual pMHC clones from the library would not necessarily be given T cell-dependent but rather from determining their surface stability or half-life time by FACS or exchange experiments. The invention also relates to a subsequent utilization of identified new epitopes via the aforementioned library. This includes confirmation of binding and orientation by free peptides in a linker-independent manner when either co-refolded with recombinant proteins of any mentioned MHCII allele or by surface loading/exchange of new epitope se- quences on professional APCs or MHCII-transfected cell cultures.

Ambiguous orientational equilibria of new epitopes will be circumvented by implementation of biasing anchor residues into wild-type sequences or so far unrelated artificial sequences. This has been realized by introducing a tryptophane instead of methionine at the P9 position of the CLIP peptide (Figure XX4), while a tryptophane at P9 promotes the canonical orientation

To verify particular antigen orientations for known epitopes the invention also relates to orientation-dependent antibodies (Ab), including both possible directins of peptide display. This also includes the derived nucleotide sequence of an approached clone if the Ab was created via yeast or phage display as long as in-house production can be managed. Further, any modification of the mentioned antibodies as fluorophore- linkages, enzyme-, chemical tag-, or any utilizable moiety-linkages (e.g.Gd, crypto- phane) will be part of the invention to use the Ab as direct or indirect detection tool of epitope orientations when displayed both by APCs or any cell culture or in vitro assays like ELISAs. Given this, the invention also includes potential usage of orientation-dependent Abs in in-vitro or in-vivo studies where subsequent applications are the site-resoluted imaging or display of presented epitopes as in infected tissues or at sites of progressive autoreactive destruction... by magnetic resonance spectroscopy and tomography. DETAILED DESCRIPTION OF THE INVENTION

The chimeric molecule of the invention is referred to herein as a protein, however, such "chimeric proteins" as defined herein may comprise nonprotein components, including, but not limited to, carbohydrate residues, chemical crosslinking agents, lipids, etc. The term "loaded" (particularly "loaded MHC molecule" or similar phrase), as used herein, refers to an empty MHC molecule (class II) which includes a presenting pep- tide or epitope bound to the peptide binding groove or cleft of the MHC molecule, preferably so that the loaded MHC molecule can modulate the activity of T cells.

Antigenic peptides associate with an MHC protein by binding to a peptide binding site of an MHC protein. As used herein, the term "peptide binding site" refers to the portion of an MHC protein capable of binding peptide. Peptide binding sites can be internal binding sites (e.g., peptide binding grooves) or external binding sites (e.g., binding sites on the external surface of an MHC protein). The conformation of a peptide binding site is capable of being altered upon binding of an antigenic peptide to enable proper alignment of amino acid residues important for TCR binding to the MHC protein and or peptide.

The domain organization of class I and class II proteins form the peptide binding site. In one embodiment of the present invention, a peptide binding site includes a peptide binding groove. A peptide binding groove refers to a portion of an MHC protein which forms a cavity in which a peptide can bind. A peptide binding groove of a class I protein can comprise portions of the a1 and a2 domains. A binding groove of a class II protein can comprise portions of the a1 and β1 domains capable of forming two (β-pleated sheets and two a helices. In one embodiment, an MHC segment of a epitope-MHC molecule of the present invention can comprise at least a portion of a class II MHC protein. According to the present invention, "at least a portion" refers preferably to a portion of an MHC class II protein capable of forming a peptide binding site or capable of forming a binding site upon addition of another chain of an MHC class II protein.

The introduction of nucleic acids into a cell can be carried out via transfection, transduction, transformation or any other process of genetic modification or trans- formation. This can take place naturally, as occurs when a virus infects cells, or artificially. Methods of artificial transfection include but are not limited to chemical methods, including calcium phosphate precipitation, DEAE-dextran complexation and lipid-mediated DNA transfer; physical methods, including electroporation, microinjection, and biolistic particle delivery (gene gun); and using recombinant, lab manipu- lated viruses as vectors. The term "transformation" relates to the bringing of exogenous nucleic acids into the cell via either natural or chemical methods. The MHC class II molecule comprises at least the minimal structure necessary to interact with both the epitope and a T cell receptor ("TCR"). Generally, this will require the presence of some or all of the extracellular domains of the a- and/or β- chains of a DP, DQ or DR antigen. If a portion of an a- and/or β-chain is used, it is particularly desirable to determine that the final construct is capable of selectively binding to a TCR of interest.

The term "epitope of interest", as used herein, refers to a molecular structure which serves as the part of an antigen which is recognized by an antigen receptor, such as a TCR. The epitope is "of interest" if it is part of an antigen associated with an infec- tious agent, a self antigen relevant to an autoimmune disease or alloreactive epitopes associated with graft rejection or graft versus host (GVH) disease.

Preferred autoantigens of the present invention include, but are not limited to, antigens which are associated with an autoimmune disease. Preferred infectious agents of the present invention include, but are not limited to, bacteria, viruses, and eu- karyotic parasites. Preferred animal parasites include protozoan parasites, helminth parasites (such as nematodes, cestodes, trematodes, ectoparasites and fungi. Preferred allergens of the present invention include, but are not limited to plant, animal, bacterial, parasitic allergens and metal-based allergens that cause contact ensitivity. More preferred allergens include weed, grass, tree, peanut, mite, flea, and cat anti- gens. Preferred toxins of the present invention include, but are not limited to, staphylococcal enterotoxins, toxic shock syndrome toxin, retroviral antigens, streptococcal antigens, mycoplasma, mycobacterium, and herpes viruses. Retroviral antigens include antigens derived from human immunodeficiency virus.

There is no limitation as to the epitopes which may be used according to the inven- tion, provided that the epitope of interest is able to associate with the MHC elements of the construct so as to be able to bind to a T cell receptor (such binding may be evaluated, for example, by fluorescence activated cell sorting analysis). The epitope may be a peptide, may be a molecule comprising amino acids as well as non-amino acid components, or may be a molecule which lacks amino acid residues altogether. The epitope may be comprised in a larger molecule which may or may not further comprise amino acid residues. The epitope may be recognized as implicated in the causation of an autoimmune disease, or it may be a newly discovered epitope. Suitable antigenic peptides of the present invention include peptides comprising at least a portion of an antigen selected from a group comprising autoantigens, infectious agents, toxins, allergens, or mixtures thereof. Examples of epitopes of interest which may be attached to the linker or the MHC class II molecule according to the invention preferably comprise those derived from glutamic acid decarboxylase 65 (associated with insulin dependent diabetes mellitus); myelin basic protein (associated with multiple sclerosis); human cartilage glucoprotein 39 (associated with rheumatoid arthritis); wheat gliadin (associated with celiac disease); and acetyl choline receptor (associated with myasthenia gravis). The person skilled in the arts knows, that antigenic peptides that have not been discovered will also be compatible with the present invention. Any peptide or epitope comprises a sequence of amino acids, that can be translated into DNA, which can be linked via a linker to the o chain of the MHC class II protein of the present invention.

The term "MHC class II element" refers to a MHC class II molecule or portions thereof which is a functional antigen presenting molecule; as such, the MHC class II element preferably comprises an a-chain component and a β-chain component. The a- and β-chain components preferably comprise all or part of the extracellular domains of the complete a- and β-chain proteins. The MHC class II element may be human or non-human. Examples of human MHC class II elements include DP, DQ and DR molecules and portions thereof, for which numerous alleles are known. Particularly preferred are MHC class II elements associated with particular autoimmune conditions: for example, DR3, DQw2 and DR4, DQw3 (associated with insulin dependent diabetes mellitus (IDDM)); DR4, DQw3 and DRI, DQwl (associated with rheumatoid arthritis); DR2, DQwl (associated with multiple sclerosis); DR3, DQw2 and DR7, DQw2 (associated with celiac disease); DR4, DQw3, and DR6, DQw (associated with pemphigus vulgaris); DR8 and DR5 (associated with pauciarticular juvenile rheumatoid arthritis); DR3, Dqw2, and DR2, DQwl (associated with systemic lupus erythematosis); DR3 (associated with Sjogren's syndrome); DR2, DQwl (associated with narcolepsy), DR3, DQw2 (associated with Graves' disease); and DR3,DQw2 (associated with dermatitis herpetiformis). The person skilled in the arts knows, that the invention comprises human and mouse MHC genes and protein products. However, the invention can also be adapted to any MHC genes of preferably mammals. The invention also relates to a expression construct for the expression of the chimeric protein. In specific nonlimiting embodiments of the invention, nucleic acids encoding the chimeric protein may be prepared as follows. For example, as a first step, where the sequence of an MHC class II gene is known, oligonucleotide prim- ers may be designed so that nucleic acid sequence encoding one or more external domains of a first MHC subunit may be obtained by reverse transcription - polymerase chain reaction (RT-PCR) of mRNA obtained from cells expressing the MHC class II gene. In particular, oligonucleotide primers used in the RT-PCR reaction may be designed to (optionally) incorporate nucleic acid sequence encoding the epitope of interest, nucleic acid sequence encoding other peptide linkers, where appropriate, and/or nucleic acid sequence containing one or more restriction enzyme cleavage site(s) that facilitate splicing together of the individual parts of the construct. The primers should also be designed such that nucleic acid and/or encoded peptide sequences necessary or desirable for the proper processing of the final chimeric protein product are included, such as, but not limited to, a leader sequence and polyadenylation sequences. For expression, the foregoing constructs may be operatively linked to a suitable promoter element. It may further be desirable to incorporate other elements that favour expression and/or secretion, such as transcriptional start and stop signals, translational start and stop sites, ribosome binding sites, and leader sequences. Such elements are known in the art.

The nucleic acid encoding the epitope of interest may be introduced into constructs encoding the a- and β-chain using genetic engineering techniques, wherein the epitope is bound to the a-chain, preferably to the amino terminus of the a-chain of the MHC class II protein. Furthermore, a peptide linker is provided between the epitope and the MHC region; adjusting the length of the peptide linker may optimize epi- tope/MHC interactions.

The nucleic acid sequences encoding the MHC portions, the epitope and the linker, may be joined (for example, by cleavage at specific sites using one or more restriction enzymes followed by ligation) and introduced into an appropriate expression vector. The ligated nucleic acids are referred to as expression construct. Suitable expression vectors include prokaryotic expression vectors, baculovirus expression vectors, eukaryotic expression vectors, bacterial expression vectors and yeast expression vectors. Suitable expression systems include mammalian cells, insect cells, yeast cells and bacterial cells. However, it can also be preferred to express elements of the chimeric protein in separated expression vectors. The separated expression vectors may either be cointroduced into a cell for protein synthesis (for example, by cotransfection where markers allow for the selection of cells containing both constructs), or the elements may be produced separately and then allowed to associate into the desired chimeric protein. Molecular weight, degree of glycosyla- tion, and disulfide linkage of the resulting chimeric protein may be evaluated and confirmed by standard techniques known to the person skilled in the art.

However, it can also be preferred that the epitope of interest is added to MHC class II molecule later, meaning after expression of the MHC class II molecule. The epitope can be expressed separately and added to the amino terminus of the a-chain of the MHC class II protein by especially chemical conjugation or peptide bonding.

The chimeric protein of the invention may be further modified to adjust its biological properties according to its intended purpose. For example, but not by way of limita- tion, half-life may be increased by PEGylation of either amino acid or carbohydrate residues, or glycosylation or any other protein modification technique known to the skilled person. As another example, the chimeric protein may be covalently or non- covalently linked to another molecule having biological activity which may act as a cytokine or as a toxin.

An "autoimmune disorder" herein is a disease or disorder arising from and directed against an individual's own tissues or a co-segregate or manifestation thereof or resulting condition there from. Examples of autoimmune diseases or disorders include, but are not limited to arthritis (rheumatoid arthritis, juvenile-onset rheumatoid arthritis, osteoarthritis, psoriatic arthritis, and ankylosing spondylitis), psoriasis, dermatitis including atopic dermatitis, chronic idiopathic urticaria, including chronic autoimmune urticaria, polymyositis/ dermatomyositis, toxic epidermal necrolysis, scleroderma (including systemic scleroderma), sclerosis such as progressive systemic sclerosis, inflammatory bowel disease (IBD) (for example, Crohn's disease, ulcerative colitis, autoimmune inflammatory bowel disease), pyoderma gangrenosum, erythema nodosum, primary sclerosing cholangitis, episcleritis), respiratory distress syndrome, including adult respiratory distress syndrome (ARDS), meningitis, IgE- mediated diseases such as anaphylaxis and allergic and atopic rhinitis, encephalitis such as Rasmussen's encephalitis, uveitis or autoimmune uveitis, colitis such as microscopic colitis and collagenous colitis, glomerulonephritis (GN) such as membranous GN (membranous nephropathy), idiopathic membranous GN, membranous proliferative GN (MPGN), including Type I and Type II, and rapidly progressive GN, allergic conditions, allergic reaction, eczema, asthma, conditions involving infiltration of T cells and chronic inflammatory responses, atherosclerosis, autoimmune myocarditis, leukocyte adhesion deficiency, systemic lupus erythematosus (SLE) such as cutaneous SLE, subacute cutaneous lupus erythematosus, lupus (including nephritis, cerebritis, pediatric, non-renal, discoid, alopecia), juvenile onset (Type I) diabetes mellitus, including pediatric insulin-dependent diabetes melli- tus (IDDM), adult onset diabetes mellitus (Type II diabetes), multiple sclerosis (MS) such as spino-optical MS, immune responses associated with acute and delayed hypersensitivity mediated by cytokines and T-lymphocytes, tuberculosis, sarcoidosis, granulomatosis including lymphomatoid granulomatosis, Wegener's granulomatosis, agranulocytosis, vasculitis (including large vessel vasculitis (including polym- yalgia rheumatica and giant cell (Takayasu's) arteritis), medium vessel vasculitis (including Kawasaki's disease and polyarteritis nodosa), CNS vasculitis, systemic necrotizing vasculitis, and ANCA-associated vasculitis, such as Churg-Strauss vasculitis or syndrome (CSS)), temporal arteritis, aplastic anemia, Coombs positive anemia, Diamond Blackfan anemia, hemolytic anemia or immune hemolytic anemia including autoimmune hemolytic anemia (AIHA), pernicious anemia, pure red cell aplasia (PRCA), Factor VIII deficiency, hemophilia A, autoimmune neutropenia, pancytopenia, leukopenia, diseases involving leukocyte diapedesis, CNS inflammatory disorders, multiple organ injury syndrome, antigen-antibody complex mediated diseases, anti-glomerular basement membrane disease, antiphospholipid antibody syndrome, allergic neuritis, Bechet's or Behcet's disease, Castleman's syndrome, Goodpasture's syndrome, Reynaud's syndrome, Sjogren's syndrome, Stevens- Johnson syndrome, pemphigoid such as pemphigoid bullous, pemphigus (including vulgaris, foliaceus, and pemphigus mucus-membrane pemphigoid), autoimmune polyendocrinopathies, Reiter's disease, immune complex nephritis, chronic neuropa- thy such as IgM polyneuropathies or IgMmediated neuropathy, thrombocytopenia (as developed by myocardial infarction patients, for example), including thrombotic thrombocytopenic purpura (TTP) and autoimmune or immune-mediated thrombocytopenia such as idiopathic thrombocytopenic purpura (ITP) including chronic or acute ITP, autoimmune disease of the testis and ovary including autoimune orchitis and oophoritis, primary hypothyroidism, hypoparathyroidism, autoimmune endocrine diseases including thyroiditis such as autoimmune thyroiditis, chronic thyroiditis (Hashimoto's thyroiditis), or subacute thyroiditis, autoimmune thyroid disease, idiopathic hypothyroidism, Addison's disease, Grave's disease, polyglandular syn- dromes such as autoimmune polyglandular syndromes (or polyglandular endocri- nopathy syndromes), paraneoplastic syndromes, including neurologic paraneoplastic syndromes such as Lambert-Eaton myasthenic syndrome or Eaton-Lambert syndrome, stiff-man or stiff-person syndrome, encephalomyelitis such as allergic encephalomyelitis, myasthenia gravis, cerebellar degeneration, limbic and/or brain- stem encephalitis, neuromyotonia, opsoclonus or opsoclonus myoclonus syndrome (OMS), and sensory neuropathy, Sheehan's syndrome, autoimmune hepatitis, chronic hepatitis, lupoid hepatitis, chronic active hepatitis or autoimmune chronic active hepatitis, lymphoid interstitial pneumonitis, bronchiolitis obliterans (non- transplant) vs NSIP, GuillainBarre syndrome, Berger's disease (IgA nephropathy), primary biliary cirrhosis, celiac sprue (gluten enteropathy), refractory sprue, dermatitis herpetiformis, cryoglobulinemia, amyotrophic lateral sclerosis (ALS; Lou Gehrig's disease), coronary artery disease, autoimmune inner ear disease (AIED) or autoimmune hearing loss, opsoclonus myoclonus syndrome (OMS), polychondritis such as refractory polychondritis, pulmonary alveolar proteinosis, amyloidosis, giant cell hepatitis, scleritis, a non-cancerous lymphocytosis, a primary lymphocytosis, which includes monoclonal B cell lymphocytosis (e.g., benign monoclonal gammopathy and monoclonal gammopathy of undetermined significance, MGUS), peripheral neuropathy, paraneoplastic syndrome, channelopathies such as epilepsy, migraine, arrhythmia, muscular disorders, deafness, blindness, periodic paralysis, and chan- nelopathies of the CNS, autism, inflammatory myopathy, focal segmental glomerulosclerosis (FSGS), endocrine ophthalmopathy, uveoretinitis, autoimmune hepatolo- gical disorder, fibromyalgia, multiple endocrine failure, Schmidt's syndrome, adrena- litis, gastric atrophy, presenile dementia, demyelinating diseases, Dressler's syndrome, alopecia areata, CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, and telangiectasia), male and female autoimmune infertility, ankylosing spondolytis, mixed connective tissue disease, Chagas' disease, rheumatic fever, recurrent abortion, farmer's lung, erythema multiforme, post-cardiotomy syndrome, Cushing's syndrome, bird-fancier's lung, Alport's syndrome, alveolitis such as allergic alveolitis and fibrosing alveolitis, interstitial lung disease, transfusion reaction, leprosy, malaria, leishmaniasis, kypanosomiasis, schistosomiasis, ascariasis, aspergillosis, Sampter's syndrome, Caplan's syndrome, dengue, endocarditis, endomyocardial fibrosis, endophthalmitis, erythema elevatum et diutinum, erythroblastosis fetalis, eosinophilic faciitis, Shulman's syndrome, Fel- ty's syndrome, flariasis, cyclitis such as chronic cyclitis, heterochronic cyclitis, or Fuch's cyclitis, Henoch-Schonlein purpura, human immunodeficiency virus (HIV) infection, echovirus infection, cardiomyopathy, Alzheimer's disease, parvovirus infection, rubella virus infection, post-vaccination syndromes, congenital rubella infection, Epstein-Barr virus infection, mumps, Evan's syndrome, autoimmune gonadal failure, Sydenham's chorea, post-streptococcal nephritis, thromboangitis ubiterans, thyrotoxicosis, tabes dorsalis, and giant cell polymyalgia. The treatment of autoimmune disorders via immune suppression is therefore one embodiment of the present invention.

As used herein, the term "allergy" or "allergies" refers to a disorder (or improper reac- tion) of the immune system. Allergic reactions occur to normally harmless environmental substances known as allergens; these reactions are acquired, predictable, and rapid. Strictly, allergy is one of four forms of hypersensitivity and is called type I (or immediate) hypersensitivity. It is characterized by excessive activation of certain white blood cells called mast cells and basophils by a type of antibody known as IgE, resulting in an ex- treme inflammatory response.

Common allergic reactions include eczema, hives, hay fever, asthma, food allergies, and reactions to the venom of stinging insects such as wasps and bees. Mild allergies like hay fever are highly prevalent in the human population and cause symptoms such as allergic conjunctivitis, itchiness, and runny nose. Allergies can play a major role in conditions such as asthma. In some people, severe allergies to environmental or dietary allergens or to medication may result in life-threatening anaphylactic reactions and potentially death. The treatment of allergies via immune suppression is therefore one embodiment of the present invention.

As used herein, the term "gene therapy" relates to the introduction and expression in an animal (preferably a human) of an exogenous sequence (e.g.; a therapeutic gene or cDNA sequence coding for a therapeutic protein, for example a full and functional sequence for a gene, for which the endogenous gene of the animal or human contains a mutation) to supplement, replace or inhibit a target gene, or to enable target cells to produce a protein that has a prophylactic or therapeutic effect toward any given condition or illness.

The term "protein therapy", "protein therapeutics" or "therapeutical proteins" refer to proteins and/or peptides and their administration in the therapy of any given condition or illness. Therapeutical proteins relate to any proteins or peptides, such as therapeutic antibodies, cytokines, enzymes or any other protein, that is administered to a patient. Examples of protein therapy relate to treatment of hemophilia via administration of plasma-derived or recombinant clotting factor concentrates, the treatment of cancer or cardiovascular disease using monoclonal antibodies or the treatment of metabolic or lysosomal disease by enzyme replacement therapy.

As used herein "pharmaceutical composition" means a composition comprising a modified antigen presenting cell preparation and at least one ingredient that is not an active ingredient whereby the composition can be safely and effectively used as a product to obtain or achieve a desired outcome. The term "pharmaceutical composition" as used herein means compositions which result from the combination of individual components which are themselves pharmaceutically acceptable. For example, where intravenous administration is foreseen, the components are suitable or acceptable (in both quality and quantity) for intravenous administration. The modified APCs of the present invention can be administered to mammals, namely humans and livestock, by numerous routes, such as intravenously, subcutaneously or intramuscularly. The dose administered may be between about 1 ,000 and about 1000,000,000 cells per dose, or other amounts understood by a person of ordinary skill in the art to be therapeutically effective as a therapy to treat or prevent symptoms of unwanted immune reactions.

The disclosure presented herein is directed towards a pharmaceutical composition which can be administered through a variety of routes including intravenously, subcutaneously, intramuscularly or directly into or onto the affected individual. When the pharmaceutical composition is delivered via an injection, the injection of the modified APC cell composition can occur as a single injection or multiple injections at any location inside or outside the body and the injection(s) can occur in a single day or over multiple days. The daily dose is administered to a subject wherein the daily amount of the modified APC cell preparation delivered to the subject from the pharmaceutical composition is about that which is therapeutically effective for treating symptoms associated with unwanted immune reactions. Additionally, the pharmaceutical composition may optionally include additional components such as salts, stabilizers and antimicrobials without departing from the spirit and scope of the claimed invention. The pharmaceutical composition of the present invention contains a modified APC cell preparation and a pharmaceutically acceptable carrier.

Although the invention has been described with respect to specific embodiments and examples, it should be appreciated that other embodiments utilizing the concept of the present invention are possible without departing from the scope of the invention. The present invention is defined by the claimed elements, and any and all modifications, variations, or equivalents that fall within the true spirit and scope of the underlying principles. EXAMPLES

Example 1 : Crystal Structures of HLA-DR1/CLIP Reveal Bidirectional Binding

The crystal structures of the human MHCII molecule HLA-DR1 in complex with the two CLIP length variants

CLIP106-120 and CLIP102-120 was determined (see Fig. 1A for sequences; num- bering refers to the p35 form of invariant chain). The latter exhibited a canonical peptide alignment showing all of the hallmarks of a conventional MHC-peptide complex [Figs. 1 B (Right) and C, and 2A; Table S1]: the pocket P1 of the MHC molecule is occupied by the first anchor residue near the N terminus of the peptide, and side chains 4, 6, and 9 protrude into the corresponding surface depressions of the MHC binding groove. In contrast, the short CLIP variant was found to bind in an inverse orientation, with the C-terminal methionine accommodated in P1 and the N- terminal methionine in P9 [Fig. 1 B(Left) and C, and Table S1]. Whereas co-refolded HLA-DR1/CLIP102-120 crystallized at room temperature, crystals showing the inverted orientation only grew at the same temperature when HLA-DR1 was refolded without any peptide and subsequently loaded with CLIP106-120. However, when changing the conditions to refolding HLA-DR1 in the presence of CLIP106-120 and crystallizing the complex at low temperatures the structure of canonically aligned CLIP106-120 bound to the MHC (Fig. S1 ) could be determined. In-depth comparison of the three structures analyzed reveals as a striking feature the preservation of a dense hydrogen bond network in both orientations (Fig. S2A). This particular feature of peptide-MHC complexes is mediated by MHC side chains that interact with the peptide backbone and, hence, is peptide-sequence independent. Surprisingly, almost no rearrangement of MHC side chains is necessary to preserve the network in the inverted orientation, and the pseudosymmetry of the polyproline type II helix ensures the availability of CO and NH peptide backbone atoms as hydrogen bond- ing partners (Fig. S2 A and B and Table S2). In the reverse peptide orientation an additional H-bond between the side chain carbonyl group of aN62 and the backbone NH group of P5 of the peptide can be formed (Fig. 2A and Fig. S2A). The central threonine at P5 acts as point of symmetry, with methionine occupying P1 and P9 and promiscuous binding of either alanine or proline at positions P4 and P6 (Fig. 1 A and C). A explanation for the observed peptide inversion is revealed by a closer analysis of the hydrogen bond network around P-2 of the MHC. At this position CLIP102-120 is engaged in three hydrogen bonds with the MHCII residues aF51 and aS53. In the canonical HLA-DR1/CLIP106-120 complex this stabilizing interface is missing. However, upon peptide inversion these hydrogen bonds are rees- tablished (Fig. 2). The only other available MHCII-CLIP structures are of longer CLIP variants that are able to form these critical N-terminal hydrogen bonds and display a canonical peptide orientation.

Example 2: Peptide Inversion Monitored by NMR Chemical Shift Analysis of HLADR1/ CLIP106-120 To ensure that bidirectional peptide alignment is not due to constraints imposed by the crystallization process, analysis of MHC-ligand complexes was carried out in solution. The only spectroscopic method to yield such information at the atomic level is NMR. Using an optimized refolding procedure, the NMR spectra of high quality for several HLADR1/ peptide complexes (Fig. S3 A and B) were obtained and assigned to the backbone chemical shifts for the 15N/13C/2H isotope-labeled β-chain of HLA- DR1/CLIP. The NMR spectra obtained from 15N-3-chain-labelled MHC molecules co-refolded with CLIP106-120 showed chemical shift changes over time that exclusively map to the peptidebinding β1 domain (Fig. S3 B and C). The initial spectrum superimposed well with that obtained from the CLIP102-120 complex, indicating that both assemblies indeed represent the same canonical binding mode observed in the crystal structures of MHCrefolded in the presence of peptide (Fig. S3A). Realignment of the peptide is then paralleled by time-dependent changes observed in the NMR spectra (Fig. S3B). Large NH backbone chemical shift differences (Fig. S3C) map to residues in the peptide binding groove, as for example βΥ78 and βΗ81 , which experience a different chemical environment in the presence of canonical or flipped CLIP, as is seen in the corresponding crystal structures (Fig. 3). Thus, this thermodynamically slightly more stable complex (Table S3) is suggested to be identical with the crystallized inverted complex obtained from HLA-DR1 refolded in isolation and loaded with CLIP106-120 a posteriori. In support, the NMR spectra of the latter assembly and the equilibrated corefolded sample are fully su- perimposable (Fig. S4).

Example 3: Spin-Labeled CLIP106-120 Confirms the Inverted Orientation The chemical shift analysis is in full agreement with peptide inversion relative to the MHC binding groove, yet it only provides circumstantial evidence. To provide more direct confirmation of the noncanonical binding mode in solution, a NMR experiments was performed with CLIP106-120 carrying an N-terminally attached TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl) moiety (Fig. 4A). Atoms up to a distance of 15-20 A of the unpaired electron should experience enhanced relaxation leading to a clearly increased line-width of the corresponding NMR resonances. Therefore, the TEMPO group should affect residues of the MHC molecule that accommodate the N-terminal methionine of the peptide. TEMPO-CLIP106-120 was loaded a posteriori to directly obtain the thermodynamically preferred complex. Dramatic changes in NMR signal intensities for a single contiguous region within the MHC molecule (Fig. 4B and C) was observed, comprising residues β54-68 that surround the P9 pocket of the antigen binding site (Fig. 4D). It was concluded that the thermodynamically stable MHCII-CLIP106-120 complex detected in solution contains a peptide in the reverse orientation corresponding to the crystal structure shown in Fig. 1 B (Left). HLA-DM-Catalyzed Peptide Inversion and Exchange Observed by NMR. CLIP inversion from the kinetically trapped to the thermodynamically more stable complex proceeds over a period of days at 37 °C, with a half-maximum inversion after =40 ± 6 h. It was assumed that CLIP cleaved from the invariant chain in the late endosome corresponds to the kinetically trapped variant found in our solution NMR studies (16). Peptide inversion would then not occur in a physiologically relevant time frame. In vivo, however, peptide exchange is enhanced by the natural catalyst HLA-DM. It was therefore important to determine whether CLIP inversion could also be catalyzed. To further investigate this phenomenon, the finding that peptide reorientation was observed for the HLA-DR1/CLIP102-120 complex was capitalized as well. This process was extremely slow and came to a halt at 50% inversion, indicating the same thermodynamic stability for both peptide alignments (Fig. S5 and Table S3). However, when adding equimolar amounts of purified HLA-DM to freshly refolded HLA-DR1 complexed with CLIP102-120, it catalyzed the 50% exchange within the experimental dead time of 30 min (Fig. 5A). Furthermore, it was determined whether both CLIP orientations would allow efficient exchange to a more tightly binding antigen. The influenza hemagglutinin-derived peptide (HA306-318) was added in excess to inverted HLA-DR1/CLIP106-120 or canonically bound HLA-DR1/ CLIP102- 120 in the presence of HLA-DM (Fig. 5B). Exchange was rapid for both samples, and the resultant spectra were fully superimposable. Therefore, as illustrated in the scheme in Fig. S6A, HLA-DM catalyzes both flipping of CLIP and, independent of the CLIP orientation, exchange against HA.

Example 4: Peptides, Loading of MHC, and Peptide Exchange by HLA-DM Peptides used in this study were synthesized in house or obtained from EMC micro- collections using Fmoc-based solid phase chemistry. The 2,2,6,6- tetramethylpiperidine-1 -oxyl (TEMPO)-labeled CLIP106-120 was synthesized from a TEMPO-4-carboxylic acid forming an amide bond to 2,3-diamino-propionic acid (DAPA), which itself replaces Lys106 in the natural CLIP sequence. Sequences from N terminus to C terminus are as follows: HA306-318

(PKYVKQNTLKLAT),CLIP106-120 (KMRMATPLLMQALPM), CLIP102-120 (KPVSKMRMATPLLMQALPM), and TEMPOCLIP106-120 (X-DAPA- MRMATPLLMQALPM), with X as the TEMPO moiety and DAPA as 2,3-di-amino- propionic acid and Ac-FR-NH2 (MHC loading enhancer). All peptides were purified by HPLC and analyzed by HPLC-ESI-MS. Peptides were used for loading in 10-fold molar excess during protein refolding and 10- to 20-fold molar excess for loading a posteriori using the dipeptidic MHC loading enhancer Ac-FRNH2 (usually 2-10 mM) to accelerate the reaction. To exchange CLIP106-120 or CLIP102-120 for HA306- 318 HLADM was added to isolated HLA-DR1/CLIP complexes in stoichiometric amounts followed by addition of HA in 1 :1-5:1 stoichiometry. No significant spontaneous exchange of CLIP for HA was observed by NMR in the absence of HLA-DM.

Example 5: HLA-DR1 with covalently linked peptides is more stable than complexes with free peptide In addition to restricting the peptide binding direction to the flipped mode, covalently linking the peptide to the MHC alpha-chain also increases the lifetime of single pMHC by reducing the ability to exchange the peptide ligand against other higher affine peptides. Figure A1 shows that the peptide ligand can be exchanged with the help of the natural peptide ligand exchange factor HLA-DM but the kinetics of this exchange are much slower as compared to the complex of HLA-DR1 with the free peptide.

Example 6: Crystal structure of CLIP106-120 linked to the alpha chain of HLA- DR1 shows identical binding mode as HLA-DR1/CLIP106-120, flipped The crystal structure of HLA-DR1/CLIP106-120 with the peptide linked to the alpha- chain of HLA-DR1 was determined to univocally demonstrate the flipped binding mode of a peptide ligand attached to the MHC in this way. The structure was determined to a resolution of 2.2 A (see Table below).

Tab. A1 : data collection and refinement statistics for the structure of HLA- DR1/CLIP106-120,a-linked and CLIP-tryptophan variants

HLA- HLA-

HLA- DR1/CLIP106

DR1 /CLIP 102 -120, DR1 /CLIP 106

- 120, M107W -120, a- linked

M1 15W

Data

collection

Space group C2221 P21 C222

Cell

dimensions

66.2, 145.11,

a, b, c (A) 73.78, 90.85, 70.68, 81.04,

138.79 80.41 108.78

, β, γ (°) 90, 90, 90 90, 91.83, 90 90, 90, 90

mol/asu 1 2 1

Resolution

34.7-1.34 34.41-2.20

34.46-2.36

(A)* (1.37-1.34) (2.42-2.36) (2.26-2.20)

4.9 (50.5) 10.7 (69.3) 7.9 (56.1)

Ι/ σΙ* 15.1 (2.0) 12.9 (2.2) 15.8 (2.9) Completeness

98.2 (83.8) 98.5 (98.5) 99.9 (100)

(%)*

Redundancy* 4.0 (3.0) 4.1 (4.1) 4.5 (4.5)

Refinement

34.46-

34.7-1.34 34.41-2.20

Resolution (A) 2.36(2.44- (1.375-1.340 (2.26-2.20)

2.36)

No. reflections 97450 36912 27008

Rwork 1 Rfree 13.88 / 17.27 18.3/23.7 18.7/22.9

No. atoms 3920 6619 3287

Protein 3357 6205 3052

Ligand/ion 6 34 -

Water 557 380 235

5-factors

Protein 16.8 28.5 33.3

Ligand/ion 29.4 41.6 -

Water 29.9 29.6 35.9

R.m.s.

deviations

Bond 0.015

0.008 0.007

lengths (A)

Bond angles 1.523

1.085 1.026

(°)

Ramachandra

n

Favored (%) 97.8 98.0 98.0

Outlier (%) 0.2 0 0

The overlay with the structure for HLA-DR1/CLIP106-120, flipped (PDB-ID: 3PGC) clearly shows that the binding mode of the peptide is identical (see Figure A2). No change of the binding register or the direction of the peptide is imposed by the introduction of the linker. The linker itself is not visible in the structure and therefore the linker seems to retain its flexibility also in the peptide bound state and is not expected to interfere with recognition of the pMHC by a TCR. Example 7: Full length versions of the alpha-chain linked pMHC are folded properly in and surface-presented on cells

Full length versions HLA-DR1 with CLIP106-120 covalently linked to the alpha and also beta chain each containing the two extracellular domains plus the transmem- brane region and a cytoplasmic tail were transiently expressed in murine fibroblast cells which are devoid of any endogen MHC II. The alpha-chain linked HLA- DR1/CLIP was expressed on the cell-surface and could be detected by the HLA-DR conformer specific monoclonal antibody L243. The expression level of the covalently linked pMHC was higher than for the wt MHC without any peptide (see Figure A3). Example 8: Fixation of peptide binding direction by introduction of binding direction biasing peptide anchors.

The two outermost pockets of the MHC antigen binding site in HLA-DR1 pocket 1 and 9 are asymmetric in their size. The P1 pocket is much larger and preferably large, aromatic side chains fit into it. In contrast the P9 pocket also favors binding of hydrophobic side chains although its size is much smaller und usually only small hydrophobic side chains are accommodated. This asymmetry enables the fixation of the direction in which peptide ligands bind that can bind in both direction due to their favorable, symmetric P1/P9 anchors (e.g. as described for CLIP with two methionines as main anchors for peptide binding). Here, these methionines were replaced by tryptophans that should only be accommodated in the P1 but not P9 pocket. Replacement of met107 with tryptophan fixes the peptide in the canonical binding mode, whereas replacement of Met1 15 with tryptophan leads to the reversed binding of the peptide. To proof this hypothesis the crystal structures of HLA- DR1/CLIP102-120,M107W and HLA-DR1/CLIP106-120,M1 15W were determined (see Figure A4 and Table A1 ).

Example 9: Construct for cellular tests: H2 IAg7-a-chain fused to Insulin B9- 23. The targeting sequence for the Endoplasmatic reticulum was displaced to the very N-terminus. Via the EcoRI and Xhol-restriction sites this construct can be inserted into the commercially available pCDNA3.1 vector.

FASTA-Format: SEQ ID NO. 1:

AAAAAGAATTCGCCACCATGCCGTGCAGCAGAGCTCTGATTCTGGGGGTCCTCGCCCTGAACACCATG

CTCAGCCTCTGCGGAGGTtcccacctggtggaggctctctacctggtgtgtggggagcgtggcGGGGG

CGGGGGCTCA GGAGGTGGAGGTTCTCTGGTGCCGCGCGGCAGCGGAGGTGGAGGTTCT

CACGTAGGCTTCTATGGTACAACTGTTTATCAGTCTCCTGGAGACATTGGCCAGTACACA

CATGAATTTGATGGTGATGAGTTGTTCTATGTCGACTTGGATAAGAAGAAAACTGTCTGG

AGGCTTCCTGAGTTTGGCCAATTGATACTCTTTGAGCCCCAAGGTGGACTGCAGAACATA

GCAGCAGAAAAACACAACTTGGGAATCTTGACTAAGAGATCTAATTTCACCCCAGCTACC

AATGAGGCTCCTCAAGCGACTGTGTTCCCCAAGTCCCCTGTGCTGCTGGGTCAGCCCAAC

ACCCTTATCTGCTTTGTGGACAACATCTTCCCACCTGTGATCAACATCACATGGCTCAGA

AATAGCAAGTCAGTCACAGACGGCGTTTATGAGACCAGCTTCCTCGTCAACCGTGACCAT

TCCTTCCACAAGCTGTCTTATCTCACCTTCATCCCTTCTGATGATGACATTTATGACTGC

AAGGTGGAGCACTGGGGCCTGGAGGAGCCGGTTCTGAAACACTGGGAACCTGAGATTCCA

GCCCCCATGTCAGAGCTGACAGAAACTGTGGTGTGTGCCCTGGGGTTGTCTGTGGGCCTT

GTGGGCATCGTGGTGGGCACCATCTTCATCATTCAAGGCCTGCGATCAGGTGGCACCTCC

AGACACCCAGGGCCTTTATGACTCGAGAAAAA

EcoRI 5' and Xhol 3' :

SEQ ID NO. 2: GAATTC,

SEQ ID. NO. 3: CTCGAG

Kozak :

SEQ ID NO. 4: GCCACC

Targeting sequence:

SEQ ID NO. 5:

ATGCCGTGCAGCAGAGCTCTGATTCTGGGGGTCCTCGCCCTGAACACCATGCTCAGCCTCTGCGGAGG T

InsulinB9-23

SEQ ID NO. 6: tcccacctggtggaggctctctacctggtgtgtggggagcgtggc

GGGGS-stretch

Seq ID NO. 7: GGGGGCGGGGGCTCAGGAGGTGGAGGTTCT

SEQ ID NO. 8: GGAGGTGGAGGTTCT

Thrombin-Cleavage site

SEQ ID NO. 9: CTGGTGCCGCGCGGCAGC

Resulting protein sequence:

Gene Bank-Format:

SEQ ID NO. 10:

MPCSRALILG VLALNTMLSL CGGSHLVEAL YLVCGERGGG GGSGGGGSLV PRGSGGGGSHVGFYGTTVYQ SPGDIGQYTH EFDGDELFYV DLDKKKTVWR LPEFGQLILF EPQGGLQNIA AEKHNLGILT KRSNFTPATN EAPQATVFPK SPVLLGQPNT LICFVDNIFP PVINITWLRN SKSVTDGVYE TSFLVNRDHS FHKLSYLTFI PSDDDIYDCK VEHWGLEEPV LKHWEPE

Linker-sequence including Thrombin cleavage site

SEQ ID NO. 1 1 :

GG GGSGGGGSLV PRGSGGGGS Example 10: Construct for expression in bacteria: H2 IAg7-a-chain fused to Insulin B9-23. This construct does not encode the Targeting sequence as well as the C-terminal Transmembrane domain.

Via the EcoRI and Xhol-restriction sites this construct can be inserted into the com- mercially available pET24d vector.

FASTA-Format:

SEQ ID NO. 18:

AAAAAGAATTCatgtcccacctggtggaggctctctacctggtgtgtggggagcgtggcGGGGGCGGG

GGCTCA GGAGGTGGAGGTTCTCTGGTGCCGCGCGGCAGCGGAGGTGGAGGTTCT

CACGTAGGCTTCTATGGTACAACTGTTTATCAGTCTCCTGGAGACATTGGCCAGTACACA

CATGAATTTGATGGTGATGAGTTGTTCTATGTCGACTTGGATAAGAAGAAAACTGTCTGG

AGGCTTCCTGAGTTTGGCCAATTGATACTCTTTGAGCCCCAAGGTGGACTGCAGAACATA

GCAGCAGAAAAACACAACTTGGGAATCTTGACTAAGAGATCTAATTTCACCCCAGCTACC

AATGAGGCTCCTCAAGCGACTGTGTTCCCCAAGTCCCCTGTGCTGCTGGGTCAGCCCAAC

ACCCTTATCTGCTTTGTGGACAACATCTTCCCACCTGTGATCAACATCACATGGCTCAGA

AATAGCAAGTCAGTCACAGACGGCGTTTATGAGACCAGCTTCCTCGTCAACCGTGACCAT

TCCTTCCACAAGCTGTCTTATCTCACCTTCATCCCTTCTGATGATGACATTTATGACTGC

AAGGTGGAGCACTGGGGCCTGGAGGAGCCGGTTCTGAAACACTGGGAACCTGAGTGACTCGAGAAAAA

EcoRI 5' and Xhol 3' :

SEQ ID NO. 12: GAATTC

and SEQ ID NO. 13: CTCGAG

InsulinB9-23

SEQ ID No. 14: atgtcccacctggtggaggctctctacctggtgtgtggggagcgtggc GGGGS-stretch:

SEQ ID NO. 15: GGGGGCGGGGGCTCAGGAGGTGGAGGTTCT

SEQ ID NO. 16: GGAGGTGGAGGTTCT

Thrombin-Cleavage site

SEQ ID No. 17: CTGGTGCCGCGCGGCAGC

Resulting protein sequence:

SEQ ID NO. 19:

MSHLVEALYL VCGERGGGGG SGGGGSLVPR GSGGGGSHVG FYGTTVYQSP GDIGQYTHEF

DGDELFYVDL DKKKTVWRLP EFGQLILFEP QGGLQNIAAE KHNLGILTKR SNFTPATNEA

PQATVFPKSP VLLGQPNTLI CFVDNIFPPV INITWLRNSK SVTDGVYETS FLVNRDHSFH

KLSYLTFIPS DDDIYDCKVE HWGLEEPVLK HWEPE

Linker-sequence including Thrombin cleavage site

SEQ ID NO. 20: VCGERGGGGG SGGGGSLVPR GSGGGGS Example 11 : Construct for cellular tests: HLA-DR1 -a-chain fused to CLIP 106- 120. The targeting sequence for the Endoplasmatic reticulum was displaced to the very N-terminus.

Via the Hindi 11 and Xbal-restriction sites this construct can be inserted into the commercially available pCDNA3.1 vector.

SEQ ID NO. 21:

aacttaagcttgccaccatggccataagtggagtccctgtgcta

ggatttttcatcatagctgtgctgatgagcgctcaggaatcatgggctAAGATGCGCATG

GCAACACCTCTTCTCATGCAAGCACTCCCCATGGGAGGTGGAGGTTCAGGAGGTGGAGGT TCAGGAGGTGGAGGTTCAatcaaagaagaa

catgtgatcatccaggccgagttctatctgaatcctgaccaatcaggcgagtttatgttt gactttgatggtgatgagattttccatgtggatatggcaaagaaggagacggtctggcgg cttgaagaatttggacgatttgccagctttgaggctcaaggtgcattggccaacatagct gtggacaaagccaacctggaaatcatgacaaagcgctccaactatactccgatcaccaat gtacctccagaggtaactgtgctcacgaacagccctgtggaactgagagagcccaacgtc ctcatctgtttcatcgacaagttcaccccaccagtggtcaatgtcacgtggcttcgaaat ggaaaacctgtcaccacaggagtgtcagagacagtcttcctgcccagggaagaccacctt ttccgcaagttccactatctccccttcctgccctcaactgaggacgtttacgactgcagg gtggagcactggggcttggatgagcctcttctcaagcactgggagtttgatgctccaagc cctctcccagagactacagagaacgtggtgtgtgccctgggcctgactgtgggtctggtg ggcatcattattgggaccatcttcatcatcaagggagtgcgcaaaagcaatgcagcagaa cgcagggggcctctgtaaTCTAGAaaaaa

Appropriate restriction sites and Kozak sequences and GGGGS- stretches were introduced into the sequence.

CLIP 106-120

SEQ ID NO. 22: AAGATGCGCATG

GCAACACCTCTTCTCATGCAAGCACTCCCCATG

Resulting protein sequence:

SEQ ID NO. 23:

MAISGVPVLG FFI IAVLMSA QESWAKMRMA PLLMQALPM GGGGSGGGGS GGGGSIKEEH

VI IQAEFYLN PDQSGEFMFD FDGDEIFHVD MARKETVWRL EEFGRFASFE AQGALANIAV

DKANLEIMTK RSNYTPITNV PPEVTVLTNS PVELREPNVL ICFIDKFTPP WNVTWLRNG

KPVTTGVSET VFLPREDHLF RKFHYLPFLP STEDVYDCRV EHWGLDEPLL KHWEFDAPSP

LPETTENWC ALGLTVGLVG IIIGTIFIIK GVRKSNAAER RGPL

Targeting sequence

CLIP 106-120

Linker-sequence :

SEQ ID NO. 24: GGGGSGGGGS GGGGS

Example 12: Construct for expression in bacteria: HLA DR1 -a-chain fused to CLIP 106-120. This construct does not encode the Targeting sequence as well as the C-terminal Transmembrane domain. Via the Ndel and Bam HI -restrict! on sites this construct can be inserted into the commercially available pET1 1 a vector.

FASTA-Format:

SEQ ID NO. 25:

AAAAACATATGAAGATGCGCATGGCAACACCTCTTCTCATGCAAGCACTCCCCATG GGAGGTGGAGGTTCAGGAGGTGGAGGTTCAGGAGGTGGAGGTTCA

ATCAAAGAAGAACATGTGATCATCCAGGCCGAGTTCTATCTGAATCCTGACCAATCAGGCGAGTTTAT GTTTGACTTTGATGGTGATGAGATTTTCCATGTGGATATGGCAAAGAAGGAGACGGTCTGGCGGCTTG AAGAATTTGGACGATTTGCCAGCTTTGAGGCTCAAGGTGCATTGGCCAACATAGCTGTGGACAAAGCC AACCTGGAAATCATGACAAAGCGCTCCAACTATACTCCGATCACCAATGTACCTCCAGAGGTAACTGT GCTCACGAACAGCCCTGTGGAACTGAGAGAGCCCAACGTCCTCATCTGTTTCATCGACAAGTTCACCC CACCAGTGGTCAATGTCACGTGGCTTCGAAATGGAAAACCTGTCACCACAGGAGTGTCAGAGACAGTC TTCCTGCCCAGGGAAGACCACCTTTTCCGCAAGTTCCACTATCTCCCCTTCCTGCCCTCAACTGAGGA CGTTTACGACTGCAGGGTGGAGCACTGGGGCTTGGATGAGCCTCTTCTCAAGCACTGGGAGTTTGATG CTCCAAGCCCTCTCCCAGAGACTACAGAGAACTAAGGATCCAAAAA

Appropriate restriction sites and Kozak sequences and GGGGS-stretches were introduced into the sequence.

Resulting protein sequence:

SEQ ID NO. 26:

MKMRMATPLL MQALPMGGGG SGGGGSGGGG SIKEEHVIIQ AEFYLNPDQS GEFMFDFDGD EIFHVDMAKK ETVWRLEEFG RFASFEAQGA LANIAVDKAN LEIMTKRSNY TPITNVPPEV TVLTNSPVEL REPNVLICFI DKFTPPWNV TWLRNGKPVT TGVSETVFLP REDHLFRKFH YLPFLPSTED VYDCRVEHWG LDEPLLKHWE FDAPSPLPET TEN

Linker-sequence :

SEQ ID No. 27: GGGG SGGGGSGGGG S

MATERIALS AND METHODS USED IN THE EXAMPLES

Expression and Purification of Proteins

Recombinant bacterial HLA-DR1 was produced and refolded in milligram amounts as described previously (1 , 2). Because inclusion bodies of the individual a- and β- chains were purified under denaturing conditions the HLA-DR1 heterodimer should be peptide free until addition of purified synthesized class ll-associated invariant chain peptides (CLIP) or HA peptide. In case of co-refolding, all peptides were used at 10-fold molar excess during the folding reaction. A posteriori loading of empty refolded HLA-DR1 was achieved by a 10- to 20-fold molar excess of the peptide in the presence of an MHC loading enhancer. Isotope-labeled HLADR1 molecules were obtained by adapting the growth protocol to defined media (M9) supplemented with 1 mg/mL 15N-NH4CI for NH-backbone fingerprint spectra [heteronuclear single quantum coherence (HSQC)]. For triple resonance-based spectra cells were grown on M9 that contained 1 mg/mL 15N-ND4CI, 2 mg/mL 13C-D7-glucose and >99% of D20 for expression of the β-chain and 14N-ND4CI, 2 mg/mL 12C-D7-glucose and >99% of D20 for expression of the a-chain. Cell growth was adapted to severed conditions in steps of low-glucose-M9 (50 ml_), 50% D20 (300 ml_, starting at OD600 = 0.2), and finally fully deuterated medium (2 L, starting at OD600 0.2-0.4). Each step required growth overnight at 37 °C. All M9 components were D20- washed before use. Cells obtained from the partially deuterated culture were washed with the fully deuterated medium before the last step of adaptation. Typical yields from these expressions were 30-80 mg protein per liter of cell culture. For obtaining selective amino acid-labeled protein we used the auxotroph DL39 (DE3) strain (kindly provided by Volker Dotsch, University of Frankfurt, Germany) which is depleted in transaminase activity. All proteinogenic amino acids were added to M9- based medium according to their approximate natural abundance in Escherichia coli and to a total of 4 g amino acids/L. Before induction with isopropyl-3-D- thiogalactopyranosid isotope- labeled amino acids were added in proper excess (3). In particular we used Gly, Phe, Leu, and Val in individual cultures on a protonated background. For NMR analysis purified proteins were dialyzed against PBS (pH 5.8) in 90% 1-120/10% D20 unless otherwise mentioned. Co-refolded peptide-MHC complexes were measured immediately after affinity purification, whereas samples prepared from "empty" protein subsequently loaded with peptides were either used immediately to monitor CLIP orientations or after an additional gel filtration step. Flag-HLA-DM was obtained and purified from stably transfected S2 cells cultured in conditioned Sf900 medium supplemented with 5% FCS and 1 % PenStrep as previously described (4). Cell cultures were adjusted to a density of 4-5 ^χ 106/mL before induction with 1 mM cupric sulfate and then grown for 4- 5 d at 27 °C. Supernatants were collected by centrifugation, filtrated, and loaded onto anti-Flag-coupled sepha- rose repetitively. Elution was achieved with 0.15 mg/mL Flag peptide or using 100 mM glycine (pH 3.0). The protein was immediately buffer exchanged to the required conditions by dialysis or gel filtration (Superdex 200; GE Healthcare). Crystallization of HLA-DR1 with Variants of CLIP

For the HLA-DR1/CLIP106-120,flipped structure determination, empty HLA-DR1 was loaded with 20-fold molar excess of peptide in the presence of 2 mM Ac-FR- NH2 for enhanced complex formation (5). Unbound peptide was removed during buffer exchange against 20 mM 2- (N-morpholino)ethanesulfonic acid (Mes) (pH 6.4)/50 mM NaCI using a Superdex 200 column (GE Healthcare). For the canonical HLA-DR1/CLIP structures the MHC was co-refolded and purified as described for the NMR sample preparations, and subsequently the complex was transferred to 20 mM Mes (pH 6.4)/ 50 mM NaCI by gel filtration. Protein was concentrated to 10 mg/ ml. and subjected to crystallization by vapor diffusion in sitting drop wells. Initial crystals for HLA-DR1/CLIP106-120,flipped appeared after 5 d under several conditions at room temperature. X-ray diffraction data were collected from a single crystal grown in 25% (wt/vol) PEG 1500 and 0.1 M malonate/imidazole/boric acid buffer at pH 6.0. Crystals for HLA-DR1/CLIP106-120,canonical grew at 4 °C in 20% (wt/vol) PEG 3350, 0.2Msodium citrate, and O.I MBisTris propane (pH 6.5) after 3 d. A single crystal was used for data acquisition. Data for HLA-DR1/CLIP102-120 were also collected from a single crystal grown at room temperature after 4 d in 20% (wt/vol) PEG 3350 and 0.1 M magnesium formate. Data Collection and Refinement

Crystals were prepared for data collection by flash freezing in reservoir solution containing 20% glycerol. X-ray diffraction data were collected at 100 K on beamline BL14.1 at the BESSY synchrotron (Berlin, Germany). Data from the HLA- DR1/CLIP106-120,flipped crystal belonged to space group P3121 , data from the HLA-DR1/CLIP106-120,canonical crystal to space group P43212, and data from the HLA-DR1/ CLIP102-120 crystal to space group P212121 . Data were processed using XDS (6). Initial phases were obtained by molecular replacement with Phaser (7) based on the HLA-DR1 a- and β- chain from PDB 2G9H (8). The structures were refined using PHENIX (9) (HLA-DR1/CLIP106-120) and with REFMAC5 (10) (HLA- DR1/CLIP102-120), interspersed with manual model building in COOT (1 1 ). Initially only the MHC molecule was modelled, and peptide was added at a later stage. For HLA-DR1/ CLIP106— 120,flipped two structures were found in the asymmetric unit. For both molecules no electron density was observed for loop β105—1 12. Eighteen surface residues could only be modelled up to the C3-atom. Electron density was also clearly present for the whole peptide except for the side chain of Lys106. For HLADR1/ CLIP106-120, canonical two structures per asymmetric unit were found, too. In this case application of noncrystallographic symmetry improved the initial maps. No density was observable for loop β107-1 12 and for nine surface residues. In the HLA-DR1/ CLIP102-120 structure the whole a- and β-chain could be modelled, as well as residues 103-1 19 of the CLIP peptide. Crystallographic data collection, processing, and refinement statistics are given in Table S1 . All figures were generated using PyMOL (DeLano Scientific). Determination of Peptide-MHC Thermal Stability.

The HLA-DR1 melting points (TM) were measured with the variants of CLIP for either co-refolded or a posteriori-loaded samples in a citrate/ phosphate buffer system at pH 5.8. As in previous studies (1 ) we used a fluorescence-based assay (Sypro Orange; Molecular probes) to determine the thermostability of proteins (Table S3). High-Field NMR Spectroscopy of MHC Class II Molecules

NMR spectra were acquired on Bruker AV700MHzor 900MHzAvance spectrometers equipped with triple-resonance cryoprobes. To decrease spectral complexity only the DR3 subunit was labelled with 15N. 1 H-15N HSQC spectra of HLA-DR1 (14N- DRa/15NDR3) in complex with one of the peptides were recorded at 310 K and pro- tein concentrations of 50 μΜ-250 μΜ. For assignment of the DR3 backbone resonances we used 15N-13C-2H labeled DR3 in complex with 2H-labeled DRa. We started from a 125-μΜ sample containing the influenza virus HA peptide (306-318) that was very stable over time, recording the following spectra at 290 K and at neutral pH: HNCA, HNcoCA, HNCO, HNcaCO, HNCACB, HNcoCACB, and a 15N- filtered NOESY. Initial assignments were facilitated by selectively (15N-Gly, 15N-

Phe, 15NLeu, or 15N-Val, respectively) labeled HLA-DR1/HA samples. HSQC spectra of all these selectively labeled samples were recorded at concentrations of 50- 100 μΜ at neutral pH. Assignments for HLA-DR1/HA could subsequently be transferred to the conditions used for HLA-DR1/CLIP complexes (pH 5.8 and 310 K). For HLA-DR1/CLIP106-120 and CLIP102-120 β-chain assignments, samples were measured at pH5.8 and concentrations of 360 μΜ and 280 μΜ, respectively. For CLIP102-120 we recorded HNCA, HNcoCa, HNCO, and HNcaCO and 15N-filtered NOESY spectra at 310 K. For HLA-DR1/CLIP106-120 we recorded HNCA and HNcoCA spectra at 310 K. These two spectra were used for approximately 80% assignments of the flipped orientation, whereas the spectra for the canonical orientation were nearly identical to those of HLA-DR1/CLIP102-120 (Fig. S3A). The latter assignment could be transferred to HLA-DR1/TEMPO-CLIP106-120. Missing HLA- DR1/CLIP β-chain assignments were caused either by proline residues or the few cases of spectral overlap. The region of β102-1 15 could not be assigned unambiguously, presumably owing to line-broadening of these resonances (unless they are in the overlap region). Processing of spectra used the Topspin (Bruker) and

SPARKY (University of California, San Francisco) software packages. All interpretation of data including assignments was made with CCPNMR Analysis software (12). Chemical shift differences were calculated using the formula Δδ = (6(δΗ)2 + (δΝ)2) 0.5. Relaxation enhancement was measured by determining the ratio of peak heights corrected for deviations based on concentration differences. Epitope map- pings were made for residues showing significant chemical shift perturbation or from large changes in intensity ratios. Significance was defined as a value above (chemical shift) or below (relaxation) the mean value ± SD unless denoted differently.

DESCRIPTION OF THE FIGURES

Fig.1 Crystal structures of HLA-DR1 bound to two CLIP length vari- ants

Fig. 2 Lack of hydrogen bonds is the driving force in CLIP106-120 reorientation

Fig. 3 Dynamic behavior of the HLA-DR1/CLIP106-120 complex

Fig. 4 Spin-labeled CLIP106-120 proves the inverted orientation in solution

Fig. 5 HLA-DM catalyzes CLIP interconversion and peptide exchange

Fig. S1 Crystal structure of HLA-DR1/CLIP106-120, canonical

Fig. S2 Preservation of a conserved hydrogen bond network in the reversed peptide-binding orientation

Fig. S3 Comparison of 1 H-15N correlation spectra of different HLA-

DR1/CLIP complexes Fig. S4 Posterior loading of empty HLA-DR1 with CLIP106-120 immediately produces the flipped state

Fig. S5 Flipped fractions of CLIP106-120 and CLIP102-120 match in their chemical shifts Fig. S6 Summary of the experimental observations and impact of

CLIP inversion for T-cell recognition

Fig. A1 Crystal structure of CLIP106-120 linked to the alpha chain of

HLA-DR1

Fig. A2 Soluble HLA-DR1 either loaded a posteriori with free

CLIP106-120 or HLA-DR1

Fig. A3 Full length versions of the alpha-chain linked pMHC

Fig. A4 Pocket P1 of HLA-DR1

Fig. A5 Generalized scheme for cloning and expression

Fig. A6 Principle of genetic or chemical linkage of peptides to HLA-DR subunits

Fig. A7 SDS-PAGE

Fig. A8 Spectra of CLIP-linked DR1 versions

Fig. A9 Spectral overlay

Fig. A10 Spectral overlay of 1 H-15N-HSQCs Fig. 1 : Crystal structures of HLA-DR1 bound to two CLIP length variants. (A) Amino acid sequences of CLIP variants used in this study. (B) Left: Crystal structure of HLA-DR1 bound to CLIP106-120 showing the peptide fragment in an atypical inverted orientation, as indicated by the arrow. Right: Analogous view onto the crystal structure of HLA-DR1 bound to CLIP102-120 in the canonical orientation. Both structures are shown as semitransparent surface (white) top view onto

α1 β1 domains, with the ribbon presentation embedded. The relative positions of the membrane distal domains are indicated. Peptides are displayed as sticks. (C) Upper: Lateral cut of the binding groove of HLA-DR1 (peptide removed), highlighting the antigen-binding pockets P1 and P9. Lower: 2Fo-Fc electron density maps (contoured at 1 o) of peptides derived from the crystal structure of H LA-DR1 /CLIP 106— 120,flipped and HLA-DR1/CLIP102-120 are shown from the same perspective inside the binding groove. Both peptides appear with two methionine side chains sequestered into the respective P1 and P9 pockets, as indicated.

Fig. 2: Lack of hydrogen bonds is the driving force in CLIP106-120 reorientation. (A-C) Magnifications of the region harboring the canonical peptide's N terminus of crystal structures of HLA-DR1 in complex with the peptides CLIP102-120 (A, blue), canonical CLIP106-120 (B, violet) or CLIP106-120 in the inverted orientation (C, red). For canonical CLIP106-120 three H-bonds from the HLA-DR1 a-chain residues Phe51 and Ser53 remain unsaturated, whereas they are replaced in the inverted CLIP106-120 orientation (compare B and C). The peptide ligand is shown in yellow.

Fig. 3: Dynamic behavior of the HLA-DR1/CLIP106-120 complex. (A) Close-up view from crystal structures of HLA-DR1/CLIP106-120 with canonical (blue) and inverted (red) peptide orientation. Residues Tyr78 and His81 located in the helix of β1 are in close contact to CLIP but experience different chemical environments in the two orientations. (B) Two regions of 1 H-15N heteronuclear single quantum coherence spectral (HSQC) overlays of HLA-DR1 (14Na/15N3)/CLIP106-120 reflect the peptide flipping sensed by these two amino acids. The blue spectrum was measured directly after sample preparation by corefolding, whereas the red one was recorded from the same sample after 120 h at 37 °C. Fig. 4: Spin-labeled CLIP106-120 proves the inverted orientation in solution. (A) Scheme of the primary sequence of CLIP106-120 carrying an N-terminal TEMPO moiety as site-specific paramagnetic relaxation enhancer. (B) CLIP106-120 carrying an N-terminal spin label orwild-type CLIP106-120 was loaded to empty HLA-DR1 (only β-chain labeled to decrease complexity). Superimposed are 1 H-15N HSQC- NMR spectra showing HLA-DR1 loaded with wild-type CLIP106-120 (blue) and

TEMPO-CLIP106-120 (red). Several resonances display significantly reduced peak intensities. (C) Intensity ratios of all unambiguously assigned resonances from both 1 H-15N-HSQC spectra were plotted to the HLA-DR3 sequence. P indicates a proline invisible in the HSQC spectrum, and asterisks mark missing assignments. (D) Epitope mapping of residues that show significantly reduced HSQC-peak intensities according to C.

Fig. 5: HLA-DM catalyzes CLIP interconversion and peptide exchange. (A) Super- imposed regions of 1 H-15N heteronuclear single quantum coherence (HSQC) spectra showing HLA-DR1 (14Na/15N3)/CLIP102-120 when freshly co-refolded (blue) or at indicated time points (red). Representative resonances indicate an increasing population of inverted CLIP102-120 over time. Rapid interconversion of the same resonances is seen when equimolar amounts of HLA-DM were added before spec- tral acquisition (Right). (B) Superimposed 1 H-15N-HSQC spectra showing HLA- DR1 (14Na/15N3)/CLIP102-120 representing canonical CLIP (blue), HLA- DR1/CLIP106-120 representing inverted CLIP (red), and spectra of these two samples after HLA-DM-catalyzed exchange against HA306-318 (light and dark gray, respectively). The latter two spectra superimpose well, indicating the formation of HLA-DR1/HA306-318 from both HLA-DR1/CLIP complexes.

Fig. S1 : Crystal structure of HLA-DR1/CLIP106-120, canonical. Structure is shown as semitransparent surface (white) top view onto the α1 β1 domains, with the ribbon presentation embedded. Peptide is displayed as sticks positioned in the canonical orientation indicated by the arrow. Fig. S2: Preservation of a conserved hydrogen bond network in the reversed pep- tide-binding orientation. (A) Comparison of individual hydrogen bonds in the canonical and flipped orientation. CLIP peptide side chains are only shown up to the C3- atom. The side chains of residues forming hydrogen bonds to CLIP106-120 are depicted as ball and stick model, with residues from the canonical structure in blue and from the inverted structure in red. Hydrogen bonds are indicated as dashed lines and are labeled according to the CLIP peptide position within the MHC binding groove. The position in the flipped orientation is marked with an asterisk. See Table S2 for a detailed summary. (B) Overview of HLA-DR1 residues involved in the pep- tide-MHC hydrogen bond network. The α1 β1 domain of HLADR1 bound to the ca- nonical CLIP106-120 is shown as gray loop structure for better orientation. Both CLIP106-120 peptides are represented as loop structure. The color code is as in A. Fig. S3: Comparison of 1 H-15N correlation spectra of different HLA-DR1/CLIP complexes. (A) 1 H-15N HSQC overlay of HLA-DR1/CLIP106-120 (blue) and HLA- DR1/CLIP102-120 (red) recorded directly after co-refolding. The CLIP102-120 complex serves as a reference for a canonical peptide alignment because the sam- pie was directly used for crystallization. Only the β-subunit has been 15N-labeled to decrease complexity. Spectra are largely identical. (B) 1 H-15N HSQC overlay of HLADR1/CLIP106-120 when either measured freshly co-refolded (blue) or after 120 h at 37 °C (red) as for Fig. 3 (main text). Inset: A set of residues that are either significantly affected by peptide interconversion (mapping to certain parts of β1 ) or show little NH-chemical shift differences (predominantly located in β2 beginning with valine 95; C). (C) NH-chemical shift differences in HSQC spectra of co-refolded HLA-DR1/CLIP106-120 observed after 120 h at 37 °C (B) are plotted as combined changes of 15N and 1 H chemical shifts (Δδ) against the sequence for HLA-DR1 β- chain residues that are assigned unambiguously in both orientations. The unas- signed loop region of β101—1 15 is not displayed proportionally..

Fig. S4: Posterior loading of empty HLA-DR1 with CLIP106-120 immediately produces the flipped state. Excerpt of 1 H-15N HSQC superpositions showing selected residues of HLA-DR^ in complex with CLIP106-120. The freshly co-refolded sample shows the canonical orientation (blue). The red spectrum shows the flipped state measured after 5 d. The green spectrum was recorded from refolded empty and subsequently loaded HLA-DR1 . The latter peaks show identical chemical shifts as the flipped HLA-DR1 sample. The three glycines represent residues from both ends of the HLA-DR1 binding cleft. Only the β-subunit has been 15N-labeled to decrease complexity. Fig. S5: Flipped fractions of CLIP106-120 and CLIP102-120 match in their chemical shifts, indicating structural homology. Spectral regions of HLA-DR1 1 H-15N HSQC spectra acquired for the co-refolded complex of CLIP102-120 after 150 d (gray) or for the co-refolded complex of CLIP106-120 fresh (blue) or after 5 d (red). The CLIP102-120-complexed protein shows the aforementioned two populations, each matching one of the distinct states of CLIP106-120. Thus, both CLIP complexes undergo peptide reorientation in solution but with significantly different propensities for the final state. Only the β-subunit has been 15N-labeled to decrease complexity. Fig. S6: Summary of the experimental observations and impact of CLIP inversion for T-cell recognition. (A) Shown is a schematic representation of the dynamic HLA- DR1-peptide exchange observed in our study. Canonically (blue) and inversely (red) bound CLIP (as cylinder) exchanges for high affine HA that is bound in a single canonical orientation (dark gray). The CLIP color code indicates the orientation. The interconversion is facilitated by HLA-DM (light gray), enabling both HLA-DR1/CLIP complexes to replace ligand for HA306-318 within a physiologically relevant time frame. (B) Electrostatic surfaces of peptide-MHC complexes as they are produced by the reactions in A. Electrostatic surfaces are shown for HLA-DR1/CLIP106-120 structures and of HLA-DR1 in complex with HA (PDB: 1 DLH). Only the α1 β1 domains of the MHCII are shown as top view, indicating the surface that would be seen by the combining site of the T-cell receptor. The peptide derived surface is indicated by a black line, emphasizing the drastic charge redistribution induced by CLIP inversion. Blue and red correspond to electrostatic surface potentials of ±2 kT/e. Surface potentials were calculated using the program APBS [Baker, NA, Sept D, Joseph S, McCammon JA (2001 ) Electrostatics of nanosystems: Application to microtubules and the ribosome. Proc Natl Acad Sci USA 98:10037-10041].

Fig. A1 : Crystal structure of CLIP106-120 linked to the alpha chain of HLA-DR1 shows identical binding mode as HLA-DR1/CLIP106-120,flipped. The structure of HLA-DR1/CLIP106-120, a-linked (blue) was superposed onto the structure of HLA- DR1/CLIP106-120, flipped (red) (PDB-ID: 3PGC). For clarity only the αΐ/βΐ-domains are shown. The bound peptide is represented by sticks. The amino- and carboxy termini are indicated.

Fig. A2: Soluble HLA-DR1 either loaded a posteriori with free CLIP106-120 or HLA- DR1 with CLIP106-120 covalently attached to the a- or β-chain were loaded with FITC-labeled HA306-318 in the presence or absence of the natural peptide exchange catalyst HLA-DM in PBS, pH 5.8. Loading was followed by fluorescence polarization. Data points are the mean ± SD of triplicates.

Fig. A3: Full length versions of the alpha-chain linked pMHC are folded properly in and surface-presented on cells. Murine fibroblast cells were transfected with full length HLA-DR1 containing the sequence for CLIP106-120 plus an additional (Gly4Ser)3-linker. The expression of the pMHC was analyzed by fluorescence- activated cell sorting 2 days later. Cells were stained with FITC-labeled a-DR anti- body L243. Bars represent the average ± SD of three independent experiments. Left, overall HLA-DR1 positive cells from transfection. Right, HLA-DR1 expression levels shown as mean of fluorescence level of all HLA-DR1 +-cells.

Fig. A4: Pocket P1 of HLA-DR1 allows accommodation of tryptophan in canonical and flipped binding orientation. The lateral cut through the surface representation of HLA-DR1 reveals the asymmetrical size of the outermost peptide binding pockets P1 and P9. A, crystal structure of CLIP102-120,M107W (yellow sticks) canonically bound to HLA-DR1 . B, CLIP106-120,M1 15W reversely bound to HLA-DR1.

Fig. A5 shows a generalized scheme for the cloning and expression of eukaryotic and procaryotic Peptide-linker-MHC constructs.

Fig. A6 shows a principle of genetic or chemical linkage of peptides to HLA-DR sub- units via their N-termini. The proof of principle for guaranteeing a fixed orientation has been carried out with CLIP106-120A6.

Fig. A7 shows a SDS-PAGE The CLIP-DR1 complexes show different SDS resis- tance in dependence of peptide orientation (compare control lanes). This is maintained in CLIP-DR1 fusion constructs as alpha-linked CLIP106-120 gives an inverted peptide orientation (see comparison with control verified by NMR) and beta- linked CLIP represents canonical peptide. Note that only non-boiled samples yield heterodimer bands which acts as a control for unspecific contaminating bands. Fig- ure is a composite. Different band sizes for subunits are based on differences in molecular weights of DR1 subunits when linked to peptides. DTT induced band shifts indicate correct formation of disulfides.

Fig. A8 shows spectra of CLIP-linked DR1 versions. ¹ H-¹⁵N-HSQC spectra of CLIP- linked DR1 versions when either measured via alpha- or beta subunit fusion. For better comparison with assigned spectra labeling was carried out in the unmodified chain, however, the full protein is present when spectra are acquired. Both spectra represent nicely folded proteins including homogenous peptide complexation.

Fig. A9 shows a spectral overlay of ¹H-¹⁵N-HSQCs recorded from samples of HLA- DR1 , labeled in alpha, when either bound to fused CLIP (blue) or canonical free CLIP peptide as control. Small differences in chemical shifts express steric perturbations mediated by the linker. The overall arrangement of resonances proves an iden- tical canonical peptide orientation for beta-linked CLIP (in contrast to alpha-linked CLIP.

Fig. A10 shows a spectral overlay of 1 H-15N-HSQCs recorded from samples of HLA-DR1 , labeled in beta, when either bound to fused CLIP (blue) or inverted free CLIP peptide as control. Small differences in chemical shifts express steric perturbations mediated by the linker. The overall arrangement of resonances proves an identical flipped peptide orientation for alpha-linked CLIP (in contrast to beta-linked CLIP, see previous figure). Data are entirely in line with SDS based read outs.

Claims

ims:

1. A chimeric protein comprising, an a-chain and β-chain of a MHC-class II protein, a linker and an epitope of interest, wherein the epitope is linked to the a-chain via the linker.

2. The protein according to claim 1 , wherein the epiotope is an autoreactive peptide.

3. The protein according to claim 1 or 2, wherein the linker is a chemical

crosslinker or a linker peptide.

4. The protein according to at least one of the preceding claims, wherein the epitope is linked to the amino terminus of the a-chain.

5. The protein according to at least one of the preceding claims, wherein the chemical crosslinker is polyethylene glycol.

6. A dimer comprising two chimeric proteins according to claim 1 to 5.

7. The dimer according to claim 6, wherein the two chimeric proteins are joined by disulfide bonds.

8. The dimer according to claim 6, wherein the two chimeric proteins are non- covalently associated via leucine zippers.

9. The dimer of claim 6, wherein the two chimeric proteins are joined by chemical crosslinking.

10. A multimer comprising at least three chimeric proteins according to claim 1 to 5.

1 1 . An expression construct comprising: a. a promoter sequence, b. a nucleotide sequence encoding the a-chain and β-chain of a MHC class II protein, c. a nucleotide sequence encoding an epitope, and d. a nucleotide sequence encoding a linker, whereby the promoter sequence is operably linked to the sequence encoding the a-chain or β-chain and wherein the sequence encoding the epitope is linked to the sequence encoding the amino terminus a-chain through the linker sequence.

12. A modified antigen presenting cell, characterised in that the cell is expressing the construct according to claim 1 1 , wherein the epitope of interest is presented via the MHC class II molecule on the surface of the modified antigen presenting cell.

13. Use of the construct according to claim 1 1 or a derivative thereof, which encodes an MHC class II molecule and an epitope of interest, for the transformation or transfection of antigen presenting cells and the production of modified antigen-presenting cells.

14. Use of modified antigen presenting cells according to claim 12 for the induction of immune tolerance and/or the suppression and/or inhibition of immune reactions, preferably antigen-specific immune reactions directed against the antigen encoded by the exogenous nucleic acid and/or unwanted adaptive immune reactions mediated by CD4+ T cells, CD8+ T cells, B cells and/or antibodies.

15. Use of modified cells according to claim 12 for the labelling and/or selection of T cells.

16. Pharmaceutical composition comprising modified antigen presenting cells according to any of the preceding claims and at least one pharmaceutically acceptable carrier.