Communication Vol. 267, No. 17, Issue of June 15, pp. 11669-11672,1992 THE JOURNAL OF BlOLOClCAL CHEMISTRY

0 1992 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U S.A.

Major Histocompatibility Complex (MHC)-encoded HAM2 Is Necessary for Antigenic Peptide Loading onto Class I MHC Molecules*

(Received for publication, January 9, 1992)

Young Yang, Klaus FruhS, James Chambers, James B. Waters, Lin Wu, Thomas Spies§, and Per A. Peterson From the Department of Immunology, Scripps Research Institute, La Jolla, California 92037 and §Division of Tumor Virology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 021 15

The mutant murine lymphoma cell line RMA-S is unable to present endogenous antigens due to its ina- bility to efficiently assemble class I major histocom- patibility complex molecules and antigenic peptides. Therefore, it has been suggested that RMA-S cells are defective either in peptide generation or in peptide transport into the endoplasmic reticulum, where class I major histocompatibility complex molecule assembly is believed to occur. As proteasomes and the putative peptide transporters HAMl and HAM2 have been im- plicated in class I antigen processing, we have inves- tigated their expression in RMA-S and its wild-type counterpart RMA. Both proteasomes and HAMl pro- teins are expressed at similar levels and show identical subcellular distributions in the two cell lines. However, only one copy of the HAM2 gene is present in RMA-S cells, and it contains a point mutation that leads to a premature stop codon. Thus, the HAM2 protein is ab- sent from RMA-S cells. These data demonstrate that HAM2 is essential for peptide loading onto class I molecules.

Although the structure and expression of the genes encod- ing P2-microglobulin and MHC’ class I heavy chains are normal in RMA-S cells (I),’ low cell surface expression of class I molecules has been observed (2, 3). The addition of synthetic, antigenic class I-binding peptides to RMA-S cells has restored surface expression of class I molecules and in- duced assembly of RMA-S-derived class I molecules in vitro

* This work was supported by the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

to the GenBankT“/EMBL Data Bank with accession number(s) The nucleotide sequencefs) reported in this paper has been submitted

M904.59. $ Recipient of a fellowship from the Deutsche Forschungsgemein-

schaft. The abbreviations used are: MHC, major histocompatibility com-

plex; HAM, histocompatibility antigen modifier; MTP, MHC-linked transporter protein; PSF, peptide supply factor; PCR, polymerase chain reaction; ER, endoplasmic reticulum; bp, base pair(s). ’ Y. Yang and P. A. Peterson, unpublished results.

(3, 4). Therefore, the deficient assembly and intracellular transport of class I molecules in RMA-S cells might have been due to lack of class I-binding peptides resulting from defective antigen processing. Recent data have demonstrated that proteasomes (5), which represent the major extralysoso- mal proteolytic system (6), contain two MHC-encoded sub- units (7-11), rendering the proteasome a prime candidate for generating antigenic peptides for class I molecules. Two genes called HAMl and HAM2 (12) in the mouse, which are mem- bers of the traffic ATPases superfamily of transporters (13), have been mapped in close vicinity to the MHC-encoded proteasomal subunits (12, 14). The rat homologues and the human homologues of HAMl and HAM2 are named MTPl and MTPZ (15) and PSFl and PSF2 (16), respectively. Their involvement in the transport of peptides across the ER mem- brane has been discussed (12, 14-20). In the present study, we have investigated the expression of proteasomal subunits and the putative peptide transporters HAMl and HAM2 in RMA-S and its wild-type counterpart RMA.

RESULTS AND DISCUSSION

To examine if proteasomes are different in RMA and RMA- S cells, we immunoprecipitated metabolically labeled protea- somes from the two cell lines and separated the subunits by two-dimensional gel electrophoresis. The polypeptide pat- terns were identical, including the levels, charges, and molec- ular weights of the two MHC-encoded subunits (data not shown), which strongly suggested that proteasomes from RMA and RMA-S do not differ. If peptide generation is normal in RMA-S cells, it seemed likely that peptide transport from the cytoplasm into the ER is defective in RMA-S cells. The human cell line LCL 721.134, with a phenotype similar to that of RMA-S, had previously been shown to revert to the wild-type phenotype following transfection with PSFl (20), the human homologue of HAMl. Therefore, we investigated the expression of HAMl in RMA and RMA-S cells by im- munoprecipitation and immunofluorescence staining using two antisera raised against a peptide corresponding to the carboxyl-terminal region of HAMl and a purified recombi- nant protein containing the entire putative ATP-binding domain of HAM1, respectively. Both antisera specifically recognized an interferon y-inducible 75-kDa protein as re- vealed by immunoprecipitation and sodium dodecyl sulfate- polyacrylamide gel electrophoresis. Immunofluorescence staining occurred in the perinuclear area representing the ER (data not shown). However, no difference was found with respect to size, charge, level of expression, and immunofluo- rescence staining pattern of HAMl between RMA and RMA- S cells. Furthermore, it has been demonstrated that the HAMl mRNA is expressed at similar levels in the two cell lines (14), and sequence analysis of the HAMl cDNA did not reveal any difference (data not shown). This indicates that the sequence and expression level of HAMl is identical in RMA and RMA-S cells, which may explain the findings that neither transfection of RMA-S with the rat homologue MTPl (15) nor transfection with the HAMl-containing cosmid 5.10 (12, 21) restored the surface expression of class I molecules (22).2 Consequently, the RMA-S phenotype must be caused by mutations in another gene.

It has also been reported that the MHC region contains a

11669

11670 HAM2 Is Essential for Peptide 1

6 1

121

181

241

301

361

421

481

541

6 0 1

661

721

781

841

901

ATGGCGCTGTCCTACCTGAGGCCCTGGGTCTCTCTGCTGCTGGCGGACATGGCTTTACTT

GGGTTGCTACMGGATCTCTGGGAAATCTGCTTCCCCAGGGGCTGCCAGGACTCTGGATA G L L O G S L G N L L P P G L P G L W I GAGGGCACCCTGCGACTTGGAGTGCTGTGGGGACTGCT~GTGGGAGAGCTGCTGGGA E G T L R L G V L W G L L K V G E L L G CTTGTGGGGACCCTTCTGCCCTTGCTCTGCCTTGCCACTCCCCTGTTTTTCTCGCTMGA L V G T L L P L L C L A T P L F F S L R

GCTCTGGTGGGAGGCACCGCGAGCACCTCAGTAGTCCGAGTGGCTTCTGCCTCTTGGGGC A L V G G T A S T S V V R V A S A S W G TGGCTGCTGGCTGGCTATGGGGCTGTTGCGCTGAGCTGGGCCGTGTGGGCTGTGCTGAGC W L L A G Y G A V A L S W A V W A V L S CCGGCTGGffiTCCAGGAGMGGMCCAGGCCAGGAG~~ACT~TGMGCGGTTG P A G V Q E K E P G O E N R T L M K R L

CTGMGCTGTCCAGGCCGGACCTGCCTTTCCTCATAGCTGCCTTCTTCTTCCTTGTGGTG L K L S R P D L P F L I A A F F F L V V

GCTGTGTGGGGGGAGACATTMTCCCTCGCTATTCGGGTCGTGTMTTGACATCCTGGGA A V W G E T L I P R Y S G R V I D I L G GGTGATTTCGACCCCGACGCCTTTGCMGCGCGCCATCTTTTTCATGTGCCTGTTCTCTGTT G D F D P D A F A S A I F F M C L F S V G~GCTCCTTCTCTGCAGGCTGTAGAGGAGGCTCCTTCCTCTTCACCATGTCCAGGATC

MCCTGCGGATACGAGAGCAGCTTTTCTCATCTTTGTTGCGCCMGACCTTGGATTCTTC N L R I R E P L F S S L L R P D L G F F CAGGAGAC~CAGGGGAGCTGMCTCGAGGCTGAGCKTGACACCTCTCT~TGAGC P E T K T G E L N S R L S S D T S L M S CGCTGGCTCCCTTTCMTGCCATCCTGCTGCGGAGCCTGGT~GTGGTGGGGCTC R W L P F N A N I L L R S L V K V V G L TACTTCTTCATGCKCAGGTATCGCCCCGACTCACCTKCTCTCCCTGCTGGACCTGCCC Y F F M L P V S P R L T F L S L L D L P CTCACGATAGCAGCTGAGMGGTGTACMCCCCCGCCATCAGGCGGTGCT~GGAGATC

~ A L S Y L R P W V S L L L A D ~ A L L

G S S F S A G C R G G S F L F T ~ S R I

20

40

60

80

100

120

140

160

180

200

220

240

260

280

300

961 CAGGATGCAGTGGCCMGGCGGGGCAGGTGGTGCGCGAGGCGGTffiGAGGGCTGCAGACT

1021 GTGCGMGCTTTGGGGCCGAGGAGCAGGMGT~GCCACTACM~GCCCTGGAGCGA

1081 TGTffiACAGCTGTGGTGGCGCCGAGACCTG~GACGTGTATCTAGTCATACGGAGG

1141 GTGATGGCCTTGGGCATGCGGTGCTGATTCTGMCTGCGGCGTGC~C~TTCTGGCT

1201 GGAGAGGTCACCCGGGGTGGCCTGCTCTCCTTCCTGCTGTACCAGGAGGMGTGGGACM

1261 TATGTCCGGMCCTGGTTTACATGTACGGGGATATGCTG~MCGTGGGCGCTGCTGM

1321 MGGTGTTTTCCTACCTGGACCGAAAGCCGMTCTGCCCCAGCCTGCGATCCTGGCCCCT

1381 CCCTGGCTGGAGGGGCGCGTGGMTTCCMGACGTCTCCTTTTCGTATCCCAGGCGCCCC

1441 GAGMGCCTGTGCTCCAGGGTCTGACGTTCACCCTGCATCCTGGMCGGTGACAGCGTTG

1 5 0 1 G T G P GTGGCCGCCCTGCTGCAGMCCTGTACCAG

1 5 6 1 CCCACTGGGGGCCAGCTGCTGCTGGATGGCGAGCCCCTGACCGAGTATGATCACCACTAC

1621 CTGCACCGCCAGGTGGTTCTGGTGGGGCAGGAGCCTGTGCTGTTCTCGGGTTCTGT~

1681 GACMTATTGCCTATGGCCTGAGGGACTGTGAGGACGCTCMGTGATGGCAGCTGCCCAG

1741 G C G G C C T G T G C A G A C G A C T T C A T A G G G ~ T G A C T M T ~ T ~ C A C A ~ T C G G G

1801 G ~ G G G G C C W T T A G C T G T G G G A C A G M G C M C G T C T G G C C A T T G C C ~

1861 P G A G G C T A C C A G C G C C C T G G A C G C C C A G T G T

1921 GMCAGGCCCTACAGMCTGGAGATCGCAGGGGGACAGGACGATGCTGGTGATTGCCCAC

1981 AGGCTGCACACGGTTCAGMTGCTGAC~GTTCTGGTGCTCMGCAGGGACGTCTGGTG

2041 GAGCATGACCAGCTCAGGGACGGCCAGGATGTCTACGCCCACCTGGTACAGCAGCGGCTG

2101 GAGGCATGA

L T I A A E I V Y N P R H O A V L K E I 3 2 0

Q D A V A K A G Q V V R E A V G G L p T 3 4 0

V R S F G A E E O E V S H Y K E A L E R 3 6 0

C R P L W W R R D L E K D V Y L V I R R 3 8 0

V U A L G M P V L I L N C G V Q O I L A 4 O O

G E V T R G G L L S F L L Y Q E E V G P 4 2 0

Y V R N L V Y W Y G D ~ L S W V G A A E ~ ~ O

K V F S Y L D R K P N L P O P G I L A P 4 6 0

P W L E G R V E F P D V S F S Y P R R P 4 8 0

E K P V L Q G L T F T L H P G T V T A L 5 0 0

V G P N G S G K S T V A A L L O N L Y P 5 2 0

P T G G P L L L D G E P L T E Y D H H Y 5 4 0

L H R P V V L V G P E P V L F S G S V K 5 6 0

D N I A Y G L R D C E D A Q V M A A A Q 5 8 0

A A C A D D F I G E ~ T N G I N T E I G ~ O O

E K G G P L A V G P K O R L A I A R A L 6 2 0

V R N P R V L I L D E A T S A L D A Q C 6 4 0

E Q A L Q N W R S Q G D R T M L V I A H 6 6 0

R L H T V Q N A D P V L V L K P G R L V 6 8 0

E H D Q L R D G O D V Y A H L V P Q R L 7 0 0

E A ' 702

FIG. 1. Nucleotide sequence of the HAM2 coding region. Decoded amino acids are shown below codons in the single-letter code. Numbers on the left and on the right correspond to nucleotide and amino acid positions, respectively. The amino acid sequences corre- sponding to the Walker motifs A and B (29) within the ATP-binding domain (positions 495-637) are underlined. In addition, the computer program KeyBank 7.1 (Intelligenetics) predicts a potential tyrosine kinase phosphorylation site at positions 367-375. Of the two predicted N-glycosylation sites at positions 133-136 and 504-507 probably only the first one is used, as the second one lies in the ATP-binding Walker motif that is facing the cytoplasm. The HAM2 sequence was obtained from cDNA libraries constructed from poly(A+) mRNA of 10* RMA and RMA-S cells, respectively. The cDNA was selected by size (1500 bp), cloned into pcDNAII, and amplified in Escherichia coli DHlaF' according to standard protocols (Invitrogen). Several independent clones were sequenced by dideoxy chain termination on an AB1 373A DNA Sequencer (Applied Biosystems, Inc., Foster City, C A).

gene, HAM2, that is adjacent and homologous to HAMl (12, 14). The role for the HAM2 gene in the transport of peptides to class I molecules is not clear. Although the mRNA of HAM2 is present in RMA-S (14), transfection of RMA-S with MTP2 leads to complete restoration of class I surface expres- sion (22) suggesting a defect in the HAM2 gene. This defect is most likely caused by a point mutation, since RMA-S cells were selected from RMA cells mutagenized by using ethyl

Loading onto Class I Molecules

methanesulfonate (23, 24). Therefore, we isolated and com- pletely sequenced cDNA clones corresponding to HAM2 from RMA and RMA-S cDNA libraries. The sequence of HAM2 from RMA cells, shown in Fig. 1, contains an open reading frame of 702 residues, which encompasses the two short amino acid sequence stretches (positions 495-566 and 605-637) pre- viously published (12). Fig. 2 shows that the HAM2 sequence displays 91% identity with the rat MTP2 sequence (22) and 77% identity with the human PSF2 sequence (25), while the identity with the HAMl sequence (12)* is less pronounced (39%). The HAM2 sequence, like the HAMl sequence, also displays striking homologies to other members of the traffic ATPases superfamily of transporters (not shown) and con- tains the typical a-@-a nucleotide binding motifs (26, 27) (amino acid residues 502-510 and 618-630 in Fig. 1).

The sequence analysis of the HAM2 cDNA from RMA-S revealed only one difference compared with the RMA se- quence at nucleotide position 97, which displayed a C to T transition (Fig. 3). This mutation introduces a premature stop codon. Several independent cDNA clones contained the same mutation at position 97 and no clone displayed the RMA sequence. Furthermore, the same mutation was found by directly sequencing PCR products obtained from reverse tran- scribed mRNA of RMA-S cells. No evidence of a mixed mRNA population consisting of both normal and mutated sequences was found (e.g. in Fig. 3). To investigate whether the RMA-S cells were homo- or hemizygous for the mutation, we amplified a region of the HAM2 gene corresponding to nucleotides 3-151 of HAM2 from RMA and RMA-S genomic DNA. The amplified DNA was separated by electrophoresis with or without prior digestion with AurII. This enzyme should cleave the mutated but not the wild-type HAM2 DNA fragment as the mutation created the recognition sequence CCTAGG for the enzyme AurII. Fig. 4 demonstrates that the HAM2 genomic DNA fragment derived from RMA-S was completely cleaved, while the fragment from RMA remained intact following the enzymatic treatment. The exclusive pres- ence of a point mutation (C to T) in the HAM2 gene of RMA- S was further confirmed by directly sequencing the PCR- amplified HAM2 genomic DNA and by Southern hybridiza- tion analysis of AurII-digested genomic DNA prepared from RMA and RMA-S (data not shown).

Because it is unlikely that RMA-S cells are homozygous for the same point mutation in the HAM2 gene, this strongly suggests that RMA-S cells contain only one copy of the HAM2 gene and that the HAM2 protein is absent in RMA-S cells. On the contrary, Powis et al. (22) showed that a 65-kDa protein, which they identified as the HAM2 protein on the basis of its reactivity with an aMTP2 peptide (the NHp- terminal MTPB sequence MALSHJRPWA(LC)) antiserum, was present in both RMA-S and RMA cells. However, a comparison of the NHp-terminal sequences of MTP2 and HAM2 reveals differences in 3 out of 10 amino acid residues (underlined; see Fig. 2). Furthermore, the HAM2 protein has a predicted molecular mass of 77.4 kDa, calculated from the translated cDNA sequence of HAM2. Therefore, this 65-kDa protein present in RMA-S is unlikely to be the HAM2 protein.

The present data are consistent with the view that the absence of the HAM2 protein in RMA-S cells generates the defect in the assembly and intracellular transport of class I molecules. This would implicate HAM2 as an essential com- ponent of the peptide-loading machinery for class I molecules. This conclusion is supported by the finding that transfection of RMA-S with MTPB cDNA restores cell surface expression of MHC class I molecules (22). Taken together with the observation that the lack of PSFl (the human homologue of HAMl) in the human cell line LCL 721.134 produces a

HAM2 Is Essential for Peptide Loading onto Class I Molecules 11671

HAMS, MTPS, PSFS, and HAMl. FIG. 2. Amino acid comparison of

The sequences were aligned using the computer program SeqEd 1.0. The posi- tions where gaps have been introduced are marked by asterisks. Identical amino acids are indicated by dashes. Numbers corresponding to amino acid positions in HAMl (12), HAM2, MTPZ (22), and PSF2 (25) are indicated on the right.

HAM2 MTPZ PSF2 HAMl

HAM2 MTP2 PSFZ HAMl

HAM2 MTPZ PSF2 HAMl

HAM2 MTP2 PSFZ HAMl

HAM2 MTPZ PSFZ HAMl

HAM2 MTPZ PSFZ HAMl

HAM2 MTP2 PSF2 H A M l

HAM2 MTPZ PSF2 HAMl

HAM2 MTPZ PSFZ HAMl

HAM2 MTPZ PSFZ HAMl

75 75 75 70

135 136 136 144

210 211 211 222

289 2 90 2 90 301

365 366 366 317

443 444 444 455

522 523 523 534

599 600 600 612

674 675 675 691

RMA HAM2 4 . h T C T G C f f C C C C A G G G G C T G I

RMA-S HAM2 1

bP 89 91 93 95 97 99 101 103 FIG. 3. Point mutation in the HAM2 gene of RMA-S. Representative chromatograms from automated sequence analysis of nucleotide

positions 87-105 of HAM2 (see Fig. 1) in RMA and RMA-S. Arrows indicate the nucleotide transition from C to T at position 97. The sources of HAM2 template DNA of RMA and RMA-S for PCR were reverse-transcribed mRNA, genomic DNA, and plasmids from RMA and RMA- S cDNA libraries. The PCR products covering the mutation in RMA-S were sequenced directly. No signal over background for any other nucleotide but T at position 97 was observed for all RMA-S-derived templates.

phenotype similar to that of the RMA-S cells (20), our results among HAMI, HAM2, and other traffic ATPases transport- demonstrate that defects in both MHC-encoded transporter ers has led to the suggestion that HAMl and HAM2 function genes have similar effects. However, it is also evident that as heterodimers, since HAMl and HAM2 each consist of HAMl and HAM2 have distinct functions since they cannot single hydrophobic transmembrane and hydrophilic ATP- complement each other, suggesting that both proteins are binding domains (half of the usual functional unit). Indeed, necessary for class I peptide loading. Structural comparison it has recently been demonstrated that PSFZ is co-immuno-

11672 HAM2 Is Essential for Peptide Loading onto Class I Molccules

FIG. 4. Restriction annlysis of I’(’K-amplified H A M 2 ge- nomic DNA f rom IIMA and KMA-S cells. HAM2 genomic DN.4 covering nucleotide pnsitions :1-1.51 was nmplilied using primers corresponding to positions :3-19 and to the opposite strand of nucle- otides 1.75-151. In hoth RMA (lone 2 ) and HMA-S (lane 4 ) a PCR DNA fragment of approximately 150 base pairs was ohtained. The 150-hp HAM2 DNA fragments of RMA (lone 3) and RMA-S (lane .5) were digested with Aur l l (RioLabs) whose recognition sequence C’CTAGG is only present in the mutated sequence. The restriction digestion results in two fragments of 90 and 60 hase pairs. I m w fi (undigested) and lane 7 (digested with AurI I ) contain a 1:l mixture of the 150-hp HAM2 DNA fragments from RMA and RMA-S. I m w s 1 and H contain molecular weight standard (I-kh ladder, RRL). Fragment sizes are indicated in base pairs on the right. DNA frag- ments were electrophoretically separated on a 6% polyacrylamide gel using Tris-acetate buffer. Template genomic DNA for PCR was prepared from HMA and RMA-S using genomic DNA isolation ki t (BRL).

precipitated with PSFl by using a PSF1-specific antiserum (28). As HAMl apparently has a normal half-life and subcel- lular distribution in RMA-S cells, the two proteins could fold independently of each other. Thus, the possible interaction between HAMl and HAM2 may be a regulated event, which may activate the function of the putative peptide transporter.

~ c k n f J / c ~ / ~ d ~ ! m f ~ f l t s - w e are grateful to Dr. .I. c . Howard for com- municating the MTI’2 sequence prior to puhlication and Drs. M. .Jackson and I,. Karlsson for valuahle discussions and unpublished data.

REFERENCES 1. nhlen. C., Rastin, .I.. Ljunggren. H.-G., Imreh, S., Klein, G.,

Townsend. A. X: Karre, K. (1990) Eur. J. Irnrncrnol. 20, 1874- 1876

2.

3 .

4.

5. 6. - I .

8.

9.

I O . 11.

12.

1.7.