Human minor histocompatibility Ags (mHag) present significant barriers to successful bone marrow transplantation. However, the structure of human mHag and the basis for antigenic disparities are still largely unknown. Here we report the identification of the gene encoding the human mHag HA-2 as a previously unknown member of the class I myosin family, which we have designated MYO1G. The gene is located on the short arm of chromosome 7. Expression of this gene is limited to cells of hemopoietic origin, in keeping with the previously defined tissue expression of the HA-2 Ag. RT-PCR amplification of MYO1G from different individuals led to the identification of two genetic variants, designated MYO1GV and MYO1GM. The former encodes the peptide sequence previously shown to be the HA-2 epitope (YIGEVLVSV), whereas the latter shows a single amino acid change in this peptide (YIGEVLVSM). This change has only a modest effect on peptide binding to the class I MHC-restricted element HLA-A*0201, and a minimal impact on recognition by T cells when added exogenously to target cells. Nonetheless, as detected using either T cells or mass spectrometry, this amino acid change results in a failure of the latter peptide to be presented at the surface of cells that express MYO1GM endogenously. These studies have thus identified a new mHag-encoding gene, and thereby provide additional information about both the genetic origins of human mHag as well as the underlying basis of an Ag-positive vs Ag-negative state.
Minor histocompatibility Ags (mHag)3 provoke transplantation immunity and function as targets of graft-versus-host disease (GVHD) in bone marrow transplant (BMT) in HLA-identical donor/recipient pairs (reviewed in Ref. 1). This necessitates life-long pharmacological immunosuppression of organ and BMT recipients. Although >50 different mHag loci have been defined among inbred strains of mice (2), the number in humans is still unknown. The genetics and tissue distribution of several human mHag have been studied by using T cells isolated from patients suffering from GVHD after BMT (3, 4, 5). mHag have been shown to be peptides derived from cellular proteins that are presented by class I MHC molecules and recognized by MHC-restricted T cells (6, 7). However, because of the difficulty in identifying such peptides, the chemical structures of mHag in both humans and mice remain largely unknown.
We previously used a combination of peptide fractionation, T cell Ag reconstitution, and mass spectrometry to identify the first amino sequence of an autosomally encoded human mHag, known as HA-2 (7). This Ag is restricted by HLA-A*0201, present in 95% of the HLA-A*0201+ population, and expressed only on cells of hemopoietic origin (3, 5). The peptide sequence of HA-2 (YIGEVLVSV) did not yield an exact match in genetic databases available at the time, although it was highly homologous to a sequence present in several genes of the class I myosin family, one of the classes of nonfilament-forming myosins thought to play a role in intracellular transport and locomotion (8). The nearest human match, YIGSVLISV, was present in MYO1E and MYO1F (9). However, the ubiquitous tissue expression of these class I myosins was inconsistent with the possibility that either was an allelic homolog of HA-2. Failure to identify the gene encoding HA-2 has stood as an impediment to understanding the basis for its antigenicity and to the development of molecular methods to screen for its presence or absence.
In the present study, we relied on newly available genetic information from the Human Genome Project to identify a candidate genomic DNA sequence encoding the HA-2 epitope. Using this information, we have established that this candidate sequence does encode HA-2, and that it represents a new member of the class I human unconventional myosin family that is expressed only in cells of hemopoietic origin. We have also defined its nonantigenic allelic homolog and established the basis for its differential antigenicity.
Materials and Methods
Cell lines and culture
The origin and maintenance of the HA-2-specific CD8+ CTL clone 5H17 has been previously described (5). The HLA-A*0201+ EBV-transformed B lymphoblastoid cell lines (EBV-BLCL) Rp (HLA-A*0101, A*0201, B8, B27), JY (HLA-A*0201, B*0702 homozygous), Blk (HLA-A*0201, B44 homozygous), and Maja (HLA-A2, B35, C4) have all been described previously (7). C1R-A2 and JK-A2 are HLA-A*0201+ stable transfectants of C1R (10) and Jurkat (11), respectively. Other transformed cell lines used and their tissues of origin are described in the text and the figure legends. All transformed cell lines were cultured in RPMI 1640 containing 5% FBS, 4 mM HEPES, 0.125% SerXtend (Irvine Scientific, Santa Ana, CA), and 3 mM l-glutamine. To maintain the expression of the HLA-A*0201 gene in the C1R and Jurkat transfectants, this medium was supplemented with either 200 μg/ml Hygromycin (Mediatech, Herndon, VA) or 300 μg/ml G418, respectively.
PCR amplification and cloning of the HA-2 coding sequence
Poly(A)+ mRNA was prepared from 1 × 107 cells using the QuickPrep mRNA Purification kit (Amersham Pharmacia Biotech, Piscataway, NJ). Alternatively, total RNA was prepared from patient PBMC and normal tissue samples with the RNAzol method (Cinaa/Biotecx Laboratories, Houston, TX). First-strand cDNA was synthesized using an oligo(dT) primer as per the manufacturer’s instructions (First Strand cDNA Synthesis kit; MBI Fermentas, Hanover, MD). Amplifications were performed on cDNA using 0.5 μM the forward and reverse primers shown in Fig. 2 (Research Genetics, Huntsville, AL) in 1.5 mM MgCl2, 0.2 mM dNTPs, and 2.5 U Taq polymerase in 1× PCR buffer (all obtained from Life Technologies, Rockville, MD). Cycle parameters were: initial denaturation at 94°C for 2 min; 36 cycles of denaturation at 94°C for 1 min, annealing at 60.5°C for 1 min, extension at 72°C for 1 min, and final extension at 72°C for 10 min. PCR products were gel purified and isolated with the Wizard PCR-Preps DNA purification system (Promega, Madison, WI). Purified fragments were cloned into the pT-Adv plasmid using the AdvanTAge cloning kit (Clontech Laboratories, Palo Alto, CA), and the inserts were sequenced at the University of Virginia Biomolecular Core Facility using M13 forward and reverse primers. At least seven individual clones were sequenced and analyzed bidirectionally for each sample.
PCR amplification for RFLP and tissue-specific expression analyses
For RFLP analysis, PCR fragments corresponding to bases 1–206 in Fig. 2 were amplified from cDNA using the forward primer 5′-AAGCTTTTCGAGAAGGGCCGCATCTA-3′ and the reverse primer 5′-GAATTCGAGATGACGATGCAGGTGTC-3′ under the conditions described above. Samples were digested at 37°C for 12 h with either Hsp92 II (Promega) or NlaIII (New England Biolabs, Beverly, MA), and analyzed on a 4% agarose gel. The MYO1GV derived sequence was not digested under these conditions and ran as a fragment of 218 bp, whereas the MYO1GM sequence gave fragments of 163 and 55 bp. For tissue-specific expression, PCR fragments corresponding to bases 1–334 in Fig. 2 were amplified from cDNA using the forward primer 5′-AAGCTTTTCGAGAAGGGCCGCATCTA-3′ and the reverse primer 5′-GAATTCCACAGGTGGACTTGAGCAGC-3′. Primers specific for glucose-6-phosphate dehydrogenase (Stratagene, La Jolla, CA) were used to amplify a 358-bp positive control fragment. The conditions used were as above except that the MgCl2 concentration was adjusted to 1.0 mM, and the annealing temperatures used were 66.2°C and 61.6°C for MYO1G and glucose-6-phosphate dehydrogenase, respectively.
Fluorescence in situ hybridization
Three cloned PCR fragments representing bases 1–1246, 978–1397, and 1215–1900 were isolated from bacterial cultures with the Wizard Plus Minipreps DNA Purification System (Promega), and inserts were excised by digestion of the pT-Adv plasmid with BstXI. The products were separated by gel purification, and the excised inserts were isolated using the Wizard PCR Preps DNA Purification System (Promega). The fragments were labeled with biotin-14-dATP using the BioNick labeling system (Roche Molecular Biochemicals, Indianapolis, IN), and standard methodology was used for hybridization and detection of this probe (14).
Peptides were synthesized on an AMS 422 multiple peptide synthesizer (Gilson, Middleton, WI) using solid-phase FMOC chemistry and Wang resins. Sequences of all synthetic peptide structures were confirmed by tandem mass spectrometry (MS/MS).
51Cr release assays were conducted as previously described (7).
Class I MHC peptide affinity binding assays
The HLA-A*0201 peptide binding assay was conducted as previously described (15) using the iodinated indicator peptide FLPSDYFPSV.
Preparation of HLA-A*0201-associated peptides
HLA-A*0201 molecules were immunoaffinity purified from JY (HLA-A*0201, B*0702 homozygous) or JK-A2 (HLA-A*0201) cells using the mAb BB7.2 (16), and their associated peptides were extracted as previously described (7, 17). Iodoacetamide was omitted from the protease inhibitor mixture to avoid potential amidocarboxy methylation of free cysteine residues. Peptides were separated from class I heavy chains and β2-microglobulin by elution in 10% acetic acid and passage through a 5-kDa-cutoff filter. Peptides were fractionated using a HAISIL C18 column (2.1 × 40 mm, 5-μm particles, 300-Å pore size; Higgins, Winter Park, FL) on an Applied Biosystems 130A HPLC (Foster City, CA) using a gradient of acetonitrile (HPLC grade; Mallinckrodt, Paris, KY) with trifluoroacetic acid (HPLC grade; Applied Biosystems) as a modifying agent. Those fractions expected to contain the synthetic peptide of interest (either YIGEVLVSV or YIGSVLISV) were established by separating a mixture of synthetic peptides including HA2V and HA2M in a parallel HPLC fractionation experiment.
Targeted MS/MS analysis of HLA-A*0201-associated peptides
Data-dependent MS/MS analysis was performed using an LCQ mass spectrometer (Thermo-Finnigan, San Jose, CA) equipped with sheathless nanoflow HPLC electrospray ionization as previously described (18). MS/MS data were acquired only on masses within 3-atomic-mass-unit (amu) windows centered on either 978.6 amu (the +1 mass of YIGEVLVSV) or 1010.6 amu (the +1 mass of YIGEVLVSM). Dynamic exclusion was not enabled for these experiments, such that multiple MS/MS spectra of the same mass were taken, increasing the quality of the data for masses of interest. The identity and amount of these peptides were established by comparison of fragmentation spectra with those of synthetic peptides.
Fourier transform mass spectrometry of HLA-A*0201-associated peptides
Mass spectrometric data were acquired on a home-built Fourier transform ion cyclotron resonance mass spectrometer (FTMS) (19) equipped with a nano-HPLC microelectrospray ionization source. Nano-HPLC columns were 50 μm inside diameter fused silica packed with ∼8 cm of 5-μm diameter reversed phase beads. An integrated microESI emitter tip (∼1-μm diameter) was located a few millimeters from the column bed. Samples were loaded onto the column and eluted into the FTMS using a gradient of 0–60% B in 32 min and 60–100% B in the next 3 min, where solvent A is 0.1 M acetic acid (Sigma, St. Louis, MO) in NANOpure water (Barnstead, Dubuque, IA), and solvent B is 0.1 M acetic acid in 70% acetonitrile. Full scan mass spectra, over a mass-to-charge (m/z) range 300 ≤ m/z ≤ 2500, were acquired at a rate of 1 scan/s.
Streptolysin O peptide transport assay
In vitro assays of TAP-mediated peptide transport were performed as previously described (20), with modifications. T1 cells (1 × 106/sample) were permeabilized on ice for 15 min with streptolysin O (15 U/ml; Murex, Norcross, GA) and incubated for 5 min at 37°C with 100 ng of the reporter peptide TVNKTERAY (21) (radiolabeled with Na125I using the chloramine T method; Ref. 22), 10 μl of 100 mM ATP, and indicated dilutions of competitor peptides. The reporter peptide contains an N-linked glycosylation site (Asn-X-Thr/Ser), and will become glycosylated after translocation by TAP into the endoplasmic reticulum (ER). Glycosylated reporter peptide was isolated using Con A Sepharose (Pharmacia Biotech, Uppsala, Sweden), eluted with 0.2 M methyl α-d-mannopyranoside (Sigma), and quantitated on a gamma counter. Reporter peptide transport in TAP-negative T2 cells was assessed as a negative control. Samples were tested in duplicate except for T2 negative control and T1 cells with no inhibitor, tested in triplicate.
Identification of the gene encoding HA-2
The peptide sequence of HA-2 (YIGEVLVSV) was used to search the high throughput genomic sequencing database, and a precise match was identified in nucleotides 56,571–56,545 of GenBank accession no. AC004847 (Fig. 1,A).4 Additionally, the protein sequence encoded by nucleotides 56,598–56,388 of AC004847 was 44% identical and 67% homologous to amino acids 42–111 of MYO1E, which had previously been identified as containing a peptide that matched HA-2 at 7/9 residues (7, 9). We next searched AC004847 using 50- to 200-bp overlapping stretches of the MYO1E cDNA sequence and low stringency parameters, and regions of homology were further delimited based on putative splice sites. This analysis led to the identification of 10 putative exons each with amino acid sequence homology of greater than 40% to the corresponding region of MYO1E (labeled 1–10 in Fig. 1,B). The apposition of these exons defined a putative partial cDNA corresponding to amino acids 42–214, 260–285, 358–497, 513–533, and 575–647 of MYO1E (Fig. 1 C). Therefore, we hypothesized that AC004847 contained a previously unidentified unconventional myosin that was the source of the HA-2 mHag.
To establish the validity of this hypothesis, RT-PCR was performed on poly(A) mRNA isolated from the HA-2+ cell lines Rp and C1R-A2 using primers that would selectively amplify overlapping segments of this putative message (Fig. 2). Each primer combination generated a product of the expected size (data not shown), which was then sequenced. This analysis established 1900 bp of a cDNA sequence that encompassed the 10 predicted exons, as well as four additional exons (labeled I-IV in Fig. 1 B). The amplified cDNA that was sequenced reproducibly differed from the AC004847-derived sequence by three nucleotides (one each in exons 8, 9, and 10) of 1900, none of which occurred within the region encoding the HA-2 epitope. We think it most likely that these three nucleotides represent sequencing errors in AC004847.
In Fig. 2, the HA-2 epitope is encoded by nucleotides 25–51, which lie entirely within a single exon, and represents amino acids 9–17 of the corresponding coding sequence. The translated cDNA sequence also includes three sequence elements that are conserved in all or most myosins: the ATP-binding sequence GESGAGKT (23) at amino acids 70–77; the consensus sequence EAFGNART (9) at amino acids 114–121; and the actin-binding motif PFYVRCIKPNE (23) at amino acids 564–574. It should also be noted that the 43 nucleotides at the 3′ end of the sequence in Fig. 2 may represent one of two alternatively spliced transcripts. This sequence is identical with that of a putative exon that matches to nucleotides 46,192–46,150 of AC004847. However, three human expressed sequence tags (GenBank accession nos. AA310652, AW340433, and AA824566) match our cDNA sequence for 60–250 nucleotides to the 5′ side of this sequence, but are then juxtaposed to AC004847 residues 45,807–45,598. Nonetheless, the results of this analysis established convincingly that a cDNA encoding an unconventional myosin and containing the HA-2 sequence was expressed in HA-2+ cells.
The gene encoding HA-2 encodes a novel human class I myosin located on chromosome 7
By comparison with the sequences of other human unconventional myosins, the 617 amino acids encoded by the cDNA in Fig. 2 represent ∼89% of the head region of the protein. To determine the relationship of the HA-2-encoding myosin to other known human unconventional myosins, we constructed an unrooted phylogenetic tree, once using that region spanned by the known HA-2 sequence and once using the same spanned region, but with gaps produced in HA-2 by insertions in other sequences removed. Both analyses gave similar results, and the results of the latter alignment are shown in Fig. 3. This analysis demonstrates that this new gene falls definitively within the MYO1 clade, and is most closely related to MYO1C. In accordance with the nomenclature in this field, we propose that this new gene be named MYO1G.
We also established the chromosomal location of this gene by fluorescence in situ hybridization using a pooled probe consisting of three segments covering the entire sequence shown in Fig. 2. Ninety one percent of 22 metaphase spreads showed hybridization to chromosome 7 band p12p13, with no evidence for additional loci on other chromosomes (data not shown). These findings establish MYO1G as the first human myosin, as well as the first mHag-encoding gene, to be localized to this chromosome.
MYO1G exists in two allelic forms
All of the RT-PCR-amplified molecular clones derived from the cell lines Rp and C1R-A2 contained precisely the same sequence in the interval encoding the HA-2 epitope. To determine the basis for differential expression of this minor Ag, we RT-PCR amplified and sequenced the region of MYO1G known to encode the HA-2 epitope from several additional cell lines. Approximately 50% of the molecular clones from the HA-2+ cell lines JY, Blk, and Maja, and all of the clones derived from the HA-2− cell line JK-A2, contained a single G-to-A transition at nucleotide 49 of the sequence in Fig. 2. This was the only variation observed in over 100 independent clones from 14 cell lines of hemopoietic origin. This change alters the sequence of the HA-2 epitope from YIGEVLVSV to YIGEVLVSM. We designated the sequence YIGEVLVSV as HA-2V and YIGEVLVSM as HA-2M, and the respective genes as MYO1GV and MYO1GM.
To confirm the genotypes inferred from direct sequencing, we took advantage of the fact that the restriction endonucleases Hsp92II and NlaIII specifically recognize the sequence CATG, which is present only in HA-2M. In keeping with this, the RT-PCR product from C1R-A2 (HA-2V/V by sequencing) representing nucleotides 1–206 of the sequence in Fig. 2, was not cut by either Hsp92II or NlaIII, whereas that obtained from JK-A2 (HA-2M/M by sequencing) was completely cut into two fragments of 163 and 55 bp by both enzymes (Fig. 4, top). Digestion of the RT-PCR products from Blk, JY, and Maja (HA-2V/M by sequencing) resulted in a mixture of both cleaved and uncleaved fragments. These results established definitively that the HA-2+ cell lines Rp and C1R-A2 are homozygous for HA-2V, whereas JY, Blk, and Maja are heterozygous. The HA-2− cell line JK-A2 is homozygous for HA-2M.
To confirm that MYO1GV and MYO1GM represent two alleles of the same gene, we performed genotyping experiments on material derived from family “VR”, which had previously been typed for HA-2 expression using the CTL clone 5H17 (Fig. 4, bottom). Four family members whose lymphoid cells were recognized by 5H17 CTL were found to be heterozygous or homozygous for HA-2V, whereas two that were phenotypically HA-2− were homozygous for HA-2M. Similar results were seen with a second, unrelated family (data not shown). This genealogy also confirmed the autosomal, Mendelian inheritance of the HA-2 gene, as both parents are heterozygous at this locus and produced offspring in all three allelic combinations.
Tissue-specific expression of the HA-2 gene
HA-2+ cells of hemopoietic origin are well recognized by 5H17 CTL, but cells derived from other tissue sources are not (3). Thus, we examined whether MYO1G expression was restricted to hemopoietic cells. We used RT-PCR to amplify nucleotides 1–334 of the sequence in Fig. 2 (the primer pair used does not discriminate between MYO1GV and MYO1GM). Strong amplification of the HA-2-specific fragment was consistently seen in lymphocytic lines of both B cell (Rp, C1R-A2, JY, Blk, and 18LCL) and T cell (Jurkat) origin (Fig. 5). Interestingly, however, MYO1G was not expressed in K562 erythroleukemia cells, which represent erythroblasts, and had not been assessed for HA-2 expression using CTL because they are class I MHC negative. Conversely, no amplification of the HA-2-specific fragment was observed in melanoma (18 Mel, DM93), embryonal kidney (293K), melanocyte (MC292, MC293), proximal tubule epithelial cells (PTEC29), HUVEC (HUVE59, HUVE62), and two samples of epidermal keratinocytes (KC25, KC45). Very weak signals were seen in lung carcinoma (VBT2), ovarian carcinoma (COV413), colorectal carcinoma (SW480), osteosarcoma (Tk-), fibroblasts (GM0568), and two other samples of epidermal keratinocytes (KC8, KC42). However, because HA-2-specific CTL fail to recognize these nonhemopoietic cells (3), this low level of gene expression does not result in peptide presentation at the cell surface. Thus, high-level expression of MYO1G is restricted to hemopoietic cells, consistent with the expression of the HA-2 epitope.
HA-2-specific CTL recognize both forms of the HA-2 peptide similarly
To more clearly define the basis for the HA-2+ and HA-2− phenotypes, we compared the recognition of HA-2V and HA-2M by the HA-2-specific CTL clone 5H17. Surprisingly, when these peptides were exogenously pulsed onto T2 (HA-2−, HLA-A*0201+) target cells, we found that both were recognized at a similar peptide concentration (Fig. 6). The difference in recognition between the two allelic forms of the peptide ranged from 2- to 10-fold in five independent experiments (data not shown). This suggested that differences in the endogenous processing and presentation of these two peptides, rather than T cell specificity, accounted for the failure of cells that expressed MYO1GM only to be recognized by 5H17 CTL.
HA-2 peptide presentation at the cell surface
The cell surface density of peptide achieved under the short-term pulsing conditions used to make CTL targets in the above experiments minimize differences in peptide affinities for class I MHC molecules that might be important during peptide binding in the ER (24). Accordingly, we performed quantitative, cell-free peptide binding assays to determine the affinity of the HA-2V and HA-2M peptides for HLA-A*0201. HA-2V half-maximally inhibited the binding of an iodinated indicator peptide (IC50) at a concentration of 4.4 nM, whereas comparable inhibition by HA-2M required ∼15 times as much peptide (Table I). Thus, substitution of Met for Val at position 9 of HA-2 reduces peptide binding significantly, but this appeared unlikely to be large enough to account for the difference in recognition of HA-2+ and HA-2− cells.
|Peptide .||Origin .||IC50 .|
|YIGSVLISV||MYO1E and MYO1F||6.8|
|Peptide .||Origin .||IC50 .|
|YIGSVLISV||MYO1E and MYO1F||6.8|
Peptide binding affinity for HLA-A*0201 was determined by identifying the concentration necessary to inhibit the binding of a radioiodinated indicator peptide by 50% (IC50) as described in Materials and Methods.
To assess the expression of HA-2V and HA-2M directly,HLA-A*0201 molecules were immunoaffinity purified from either JY (MYO1GV/M heterozygous) or JK-A2 (MYO1GM homozygous) cells and separated by HPLC. Fractions that could have contained either HA-2V or HA-2M peptides were identified based on the elution position of synthetic peptides in parallel HPLC runs, and these fractions were analyzed by mass spectrometry. Next, peptide ions of 978.6 and 1010.5 amu, corresponding to the masses of the +1 ions of HA-2V and HA-2M, respectively, were targeted for MS/MS analysis. The identification of HA-2V was established by comparison of the fragmentation spectrum of a naturally processed peptide with that of synthetic HA-2V. By comparing the magnitudes of the ion current of several fragment ions from naturally processed HA-2V with those of a known quantity of synthetic material, we calculated that HA-2V was present in the JY peptide sample in an amount corresponding to ∼54 copies/cell (Fig. 7). Although JY also expresses mRNA encoding HA-2M (Fig. 4), a similar analysis of fragmentation spectra established that this peptide was not detected at a level above 0.2 copies/cell (data not shown). The absence of the HA-2M peptide was also established by analysis of the JK-A2 extract using a FTMS, which is ∼10× more sensitive than the LCQ. A single mass with an m/z of 1010.52+1 and an appropriate elution time for HA-2M was detected in a single HPLC fraction of this sample (Fig. 8). However, when synthetic HA-2M was added to an aliquot of this sample and a second analysis performed, the endogenous peptide eluted almost a full half-minute before synthetic HA-2M. No peptide of the appropriate mass was detected at this elution position in the original sample. In these experiments, we also did not detect appropriate masses at the elution position of HA-2 M with an oxidized Met residue (25) (data not shown). Based on the amount of peptidic material loaded and an instrument detection limit of ∼6 amol, the detection limit in these experiments is ∼0.04 copies/cell. We conclude that HA-2M peptide was present at less than this level in the JK-A2 extract, despite the expression of mRNA encoding HA-2M in this cell.
To gain additional insight into the failure of HA-2M to be presented at the cell surface, we compared the ability of HA-2V and HA-2M to inhibit TAP-dependent transport of the radiolabeled reporter peptide TVNKTERAY in streptolysin O-permeabilized T1 cells. Somewhat surprisingly, we found that HA-2M was transported more efficiently than HA-2V (Fig. 9). Although it is possible that the actual substrates for TAP transport may correspond to precursors peptides rather than the mature 9 mer epitope, these data are consistent with the hypothesis that factors other than TAP transport determine the differential expression of the HA-2V and HA-2M peptides in association with HLA-A*0201.
In this study, we have identified 1900 bp of sequence representing the gene from which the previously identified mHag HA-2 is derived. This identification confirmed our previous hypothesis that HA-2 originates from a class I myosin protein. Based on a phylogenetic analysis of the known portion of the HA-2-encoding gene compared with the same region of the other 14 human unconventional myosins, we have classified this gene as a new member of the myosin I family, and have named it MYO1G. Although it is relatively homologous to myosin IC, it is the first class I myosin whose expression is largely confined to hemopoietic tissues. It is also the first class I myosin whose gene has been localized to chromosome 7. Our assignment of MYO1G to 7p12p13 is also supported by the identification of a chromosome 7-derived human expressed sequence tag (GenBank accession no. AA078664) that matches nucleotides 375–595 of our sequence at 218 of 221 possible residues (98.6% identity), and by the fact that AC004847 was recently reported to be derived from chromosome 7-enriched material (http://www.nhgri.nih.gov/DIR/GTB/CHR7). This identification process illustrates the impact of information provided by the Human Genome Project to augment the rapid identification of new genes and proteins.
The results of our work also provide an understanding of how HA-2 is expressed as a mHag. By analyzing the sequence of MYO1G in cells from different individuals, we identified only a single polymorphism that encoded either a valine or a methionine at position 9 of the HA-2 peptide sequence. The inheritance of this polymorphism establishes two alleles of MYO1G. Furthermore, there was complete concordance between the presence MYO1GV and detection of HA-2 by CTL, as well as the homozygous presence of MYO1GM and the lack of detection of HA-2. Despite the differences in peptide structure, both HA-2V and HA-2M are recognized similarly by HA-2-specific CTL when the synthetic peptides are exogenously pulsed onto HLA-A*0201+ target cells. In addition, the equilibrium binding of HA-2M to HLA-A*0201 is only ∼12-fold lower than that of HA-2V. Based on the detection of HA-2V at ∼54 copies/cell on JY (HA-2V/M) and assuming that equilibrium binding conditions exist in the ER, we would expect that HA-2M should be detected at roughly 4–8 copies/cell and that such cells should be relatively well-recognized by HA-2-specific CTL. However, we found no HA-2M above a level of 0.04 copies/cell in the material derived from either JY or JK-A2 (HA-2M/M).
One possible explanation for the failure of HA-2M to be presented is that it is transported into the ER relatively poorly by TAP. In fact, we have recently shown this to be the basis for differential expression of another human mHag, HA-8 (26). However, the V-to-M change is relatively conservative, and Met at P9 is not contraindicated for successful TAP transport (27, 28, 29). Furthermore, direct analysis of the ability of these two peptides to be transported by TAP was not consistent with this explanation. A second possibility is that MYO1GV and MYO1GM are processed differently by the proteasome, such that substitution of an M for a V results in either creation or loss of a cleavage site. Although the specific sites recognized by proteasomes are less well defined than are the binding motifs for TAP and HLA-A*0201, this amino acid change is rather minor, and it is not immediately evident that such a switch would have an effect on proteasomal cleavage (30). Finally, it should be noted that our analysis of polymorphism encompasses only the 24 bp immediately upstream of the HA-2 sequence. We have not defined the complete 5′ end of MYO1G or the sequence of any relevant control sequences. Thus, it is possible that as yet unrecognized differences in MYO1GM influence proteolytic cleavage efficiency or proper translation of the gene. In addition, our RT-PCR conditions were not quantitative, such that there might be unrecognized differences in the levels of transcription of the two alleles. However, in this regard it should be noted that we were able to detect very low levels of transcription of MYO1G in a few cell lines, yet these were not recognized by HA-2-specific CTL. Further experimentation will be required to fully resolve this issue.
In keeping with previous work on HA-2 expression (5), we found significant transcription of MYO1G only in cells of hemopoietic origin. Such highly restricted tissue expression suggests great promise for immunotherapy. In addition to those derived posttransplant from the original patient, HA-2-specific T cells have now been generated ex vivo and shown to efficiently lyse EBV-BLCL and malignantly transformed leukemic blasts, while ignoring cells of nonhemopoietic origin (31). Thus, they represent excellent potential adoptive immunotherapy reagents in conjunction with BMT for hemologic malignancies. It is significant that we failed to detect expression of MYO1G in K562 erythroleukemia cells. The identity of the gene in the present work should enable a more thorough examination of the tissues in which it is expressed.
HA-2 is also expressed in nonmalignant leukocytes (PBL, thymocytes, monocytes, dendritic cells) and should thereby present a target on those cells that mediate autoimmune disease in patients with the proper haplotype. Although not yet common practice, studies in both mice and humans have suggested the potential of BMT as a curative treatment for various autoimmune diseases (32, 33, 34, 35, 36, 37). Therefore, it is feasible that HA-2-specific T cells could also be used to treat these diseases in the same way that they might be used to treat leukemia. Finally, the facts that the HA-2 epitope is so prevalent (95% in the HLA-A*0201+ population), and that disparity for this minor H Ag has not been shown to have a positive correlation with the incidence of GVHD, only reinforce the potential therapeutic value of T cells directed against this Ag.
This work was supported by U.S. Public Health Service Grants AI20963 and AI44134 (to V.H.E.), AI33993 (to D.F.H.), and AI39501 (to L.C.E.); by grants from the J. A. Cohen Institute for Radiopathology and Radiation Protection (to E.G.); by the Dutch Cancer Society (to M.W. and T.M.); and by the Leukemia and Lymphoma Society and American Cancer Society Grant RPG-98-036-01-CIM (to L.C.E.). R.A.P. was supported by U.S. Public Health Service Training Grant AI0746, and T.N.G. by the Cancer Research Institute/CIGNA Foundation Fellowship.
Address correspondence and reprint requests to Dr. Victor H. Engelhard, Carter Immunology Center, University of Virginia, P.O. Box 801386, Charlottesville, VA 22908. E-mail address: vhe@Virginia.edu
Abbreviations used in this paper: mHag, minor histocompatibility Ag; GVHD, graft-versus-host disease; BMT, bone marrow transplant; EBV-BLCL, EBV-transformed B lymphoblastoid cell line; MS/MS, tandem mass spectrometry; amu, atomic mass unit; m/z, mass-to-charge; ER, endoplasmic reticulum; FTMS, Fourier transform ion cyclotron resonance mass spectrometer.
The reverse numbering indicates that the HA-2 epitope is encoded on the complementary (reverse) strand of AC004847.