Cloning of cdnas Encoding Human Lysosomal Membrane Glycoproteins, h-lamp- 1 and h-lamp-2

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 263, No. 35, Issue of December 15, pp ,1988 Printed in U.S.A. Cloning of cdnas Encoding Human Lysosomal Membrane Glycoproteins, h-lamp- 1 and h-lamp-2 COMPARISON OF THEIR DEDUCED AMINO ACID SEQUENCES* (Received for publication, June 6, 1988) Minoru Fukuda, Juha ViitalaS, Jeanne Matteson, and Sven R. Carlssons From the La Jolla Cancer Research Foundation, Cancer Research Center, La Jolla, California and the Department of Physiological Chemistry, University of Umei, S Ume6, Sweden We have isolated and sequenced cdna clones corresponding to the entire coding sequences of the human lysosomal membrane glycoproteins, lamp- 1 and lamp- 2 (h-lamp-1 and h-lamp-2). The deduced amino acid sequences indicate that h-lamp-1 and h-lamp-2 consist of 416 and 408 amino acid residues, respectively, and suggest that 27 and 28 NHz-terminal residues are cleavable signal peptides. The major portions of both h-lamp-1 and h-lamp-2 reside on the the luminal side of the lysosome and are heavily glycosylated by N- glycans: h-lamp- 1 and h-lamp-2 were found to contain 19 and 16 potential N-glycosylation sites, respectively. The findings are consistent with the results obtained by endo-/3-n-acetylglucosaminidase F treatment of h- lamp- 1 and h-lamp-2 precursors, described in the preceding paper (Carlsson, s. R., Roth, J., Piller, F., and Fukuda, M. (1988) J. Biol. Chen. 263, ). These N-glycosylation sites are clustered into two domains separated by a hinge-like structure enriched with proline and serine in h-lamp-1 or proline and threonine in h-lamp-2. The two domains of h-lamp-1 on each side of the hinge region are homologous to each other, whereas no such homology was detected between the two domains of h-lamp-2. Both proteins have one putative transmembrane domain consisting of 24 hydrophobic amino acids near the COOH terminus, and contain a short cytoplasmic segment composed of 11 amino acid residues at the COOH-terminal end. Com- parison of h-lamp-1 and h-lamp-2 sequences reveal strong homology between the two molecules, particularly in the proximity to the COOH-terminal end. It is the lysosomal membrane. possible that this portion is important for targeting the molecules to lysosomes. These results also suggest that lamp-1 and lamp-2 are evolutionarily related. Comparison of known lamp-1 sequences among different species, on the other hand, show that human lamp-1 has more similarity to lamp- 1 from other species than to human lamp-2. This fact, taken together with the finding that h-lamp-2 lacks repeating domains, sug- gests that lamp- 1 and lamp-2 diverged from a putative ancestor gene in early stages of evolution. These results also suggest that lamp-1 and lamp-2 probably have *This workwas supported by Grant CA28896 and, in part, by Grant CA33000 awarded by the National Cancer Institute (to M. F.), and by Grant B from the Swedish Medical Research Council (to S. R. C.). 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. The nucleotide sequence(s) reported in thispaper has been submitted distinctly separate functions despite the fact that they share many structural features. In the preceding paper, we described the characterization of two different human lysosomal membrane glycoproteins with M, - 120,000, h-lamp-1 and h-lamp-2, isolated from chronic myelogenous leukemia cells (1). The two glycoproteins appear to share some structural features and have the polypeptide portions of about -40 kda. They are heavily glycosylated by N-glycans, and h-lamp-1 and h-lamp-2 were shown to contain 18 and 16 N-glycans, respectively. The most striking feature is that some of the N-glycans attached to lamp-1 and lamp-2 are polylactosaminoglycans (1). We have also succeeded in isolating cdna clones encoding h-lamp-1 and reported its unique structure (2). However, it was also noticed that the NHp-terminal amino acid sequence deduced from those cdnas seemed to be truncated, when compared to that obtained on a purified protein (2). We report here the characterization of cdnas encoding another human lysosomal membrane glycoprotein, h-lamp-2, and the isolation of cdna clones derived from the full-length h-lamp-1 mrna. Comparison of the deduced amino acid sequence of h-lamp-2 with that of h-lamp-1 indicates that human lamp-1 and lamp-2 share common structural features in both gross structures and their COOH-terminal regions. Further analysis, however, reveals that lamp-1 and lamp-2 are distinct molecules which may have different functions in EXPERIMENTAL PROCEDURES Zsolation of the Glycoprotein Fraction Containing Human lamp-1 and lamp-2 and Their Derived Peptides-The glycoprotein fraction, which is a mixture of h-lamp-1 and h-lamp-2, was purified as described previously (1). Peptides were generated by CNBr fragmentation of the glycoprotein fraction after reduction and S-carboxymethylation. Reduced and carboxymethylated glycoprotein (-250pgof protein) was cleaved with CNBr (50 mg/ml in 70% formic acid) for 20 h. After lyophilization, the sample was dissolved in 0.1% trifluo- roacetic acid and filtered through a 0.45-pm Millipore filter. The filtered sample was applied on a Vydac TP214 C18 wide pore reverse phase column (4.6 mm X 20 cm), previously equilibrated with 0.1% trifluoroacetic acid. After washing with 10 ml of 0.1% trifluoroacetic acid, the column was eluted with 40 ml consisting of a gradient (0-60%) of acetonitrile in 0.08% trifluoroacetic acid. Flow rate was 1 ml/ min, and the absorbance of the eluate was monitored at 215 nm (Fig. 1). Three major peaks, obtained after HPLC, were further purified bygel filtration on Sephacryl S-200. Each peak was lyophilized, dissolved in 0.5 ml of 0.1 M NH~HCOS, and applied on a column (1 X to the GenBankTM/EMBL Data Bank accession with number(s) The abbreviations used are: h-lamp-1 and h-lamp-2, human lyand sosome-associated membrane protein-1 and protein-2; SDS, sodium -$ Present address: University of Helsinki, Helsinki, Finland. dodecyl sulfate; HPLC, high performance liquid chromatography

2 Molecular Cloning of Human Lysosomal Membrane Glycoproteins cm) of Sephacryl S-200, equilibrated and eluted with 0.1 M NH4HC03. Flow rate was 6 ml/h, fractions (2 ml) were collected (Fig. 2). Among these peptides, four fractions designated, J1, K1, Kat and Ml, were sequenced by automated Edman degradation using an Applied Biosystems Model 470A gas phase amino acid sequencer. Phenylthiohydantoin amino acid derivatives were identified by HPLC as described (3). Similar peptides were isolated from tryptic digests by Sephacryl S-200 gel filtration followed by reverse phase HPLC. Oligonucleotide Probes-Synthetic oligonucleotides were synthesized on an Applied Biosystems 350A oligonucleotide synthesizer. Sequencing primers were used directly without further purification. For cdna library screening, oligonucleotide probes, which are based on amino acid sequences using most probable codon usage, were purified by reverse phase HPLC. "Exact" oligonucleotide probes based on cdna sequences were used without further purification. cdna Library Screening-Among the peptides sequenced, two peptides (J1 and K,) were chosen for constructing oligonucleotide probes, and oligonucleotides A and B, based on most probable codon usage (4,5), were synthesized. The oligonucleotides were purified on reverse phase HPLC and end-labeled with [T-~*P]ATP (>7000 Ci/ mmol, Du Pont-New England Nuclear) using T4 polynucleotide kinase (Bethesda Research Laboratories) to a specific activity of >2 X IO9 cpm/pg. After purification on a NACS column (Bethesda Research Laboratories), the radioactive oligonucleotides were used for screening of human fetal lung fibroblast cell line (IMR-90) (6) and human placental (7) Xgtll cdna libraries. S1 nuclease was employed in the synthesis of the IMR-90 library, whereas the placental library was made without the use of S1 nuclease. The Xgtll cdna libraries were kindly provided by Drs. Krusius, Ruoslahti, and Millan at the La Jolla Cancer Research Foundation. In situ colony hybridization was performed essentially according to the method of Woods et al. (8). In brief, the filter paper was prehybridized at 60 "C (for cdna probe) or at 42 "C (for oligonucleotide probe) in 6 X SSC, 1 X Denhardts's solution containing 0.25% SDS and 100 pg/ml sheared and denatured salmon sperm DNA (for details of the solutions, see Ref. 9). The filters were then hybridized in 50 mm sodium phosphate, ph 6.8, 6 X SSC, 1 X Denhardt's, 50 pg/ml of Escherichia coli trna for h. The hybridization temperature was 45 "C for probes based on the most probable codon usage, 50 "C for exact oligonucleotide probes, and 60 "C for cdna inserts. The filters were washed at room temperature for 40 min to 1 h in 2 X SSC, 0.1% SDS. If necessary, the filters were then washed further at 50 "C in 6 X SSC, 0.1% SDS for exact oligonucleotide probes or at 60 "C in 0.1 X SSC, 0.1% SDS for cdna insert probes for min. The background radioactivity was monitored by a hand-held Geiger counter (series 900 minimonitor, Mini-Instruments, Ltd., purchased from Research Products International Corp., Mount Prospect, IL). In order to obtain cdnas encoding h-lamp-1 from the placental library, the previously described cdna clone, X-h-lamp-A-4 (see Ref. 2) was used as probe for initial screening. cdna probes were made by random-oligonucleotide-primed extension (10) as described previously (11). For the later stages of screening, differential hybridization to oligonucleotide probes specific to the h-lamp-1 or h-lamp-2 was performed. The oligonucleotide probes were synthesized according to the h-lamp-1 or h-lamp-2 cdna sequence and end-labeled as described above. Nucleotide Sequencing-The h-lamp-1 and h-lamp-2 cdnas were excised from the Xgtll clones by EcoRI digestion and inserted into the EcoRI site of the Bluescript plasmid (Strategene). The sequence was obtained by sequencing double-stranded DNAs by the dideoxy nucleotide method (12). Following use of universal primers, sequencing was continued by the use of sequentially constructed 16- to 23- mer oligonucleotides (2, 11, 13). Since the h-lamp-2 cdna sequence was found to contain two internal EcoRI sites, the h-lamp-2 cdna fragment was excised from the Xgtll clone by KpnI and Sac1 digestion, utilizing the KpnI and Sac1 sites in Xgtll (14). The 3.3-kilobase fragment containing the h- lamp-2 cdna and flanking Xgtll nucleotides was then inserted into the Bluescript plasmid and sequenced. A synthetic oligonucleotide, corresponding to the region 44 to 25 nucleotides upstream from the 5' EcoRI site, was used as a primer for sequencing through the EcoRI sites. All the other procedures have been described previously (2, 11, 13). RESULTS AND DISCUSSION Construction of Oligonucleotide Probes-Initially, the glycoprotein fraction containing both human lamp-1 and lamp- 2 was used to generate peptides, since it was not clear that this fraction contained both glycoproteins. The reverse phase HPLC of the CNBr fragments provided three major peptide peaks (J, K, and M), as shown in Fig. 1. Each peak was further purified by gel filtration on Sephacryl S-200. Fraction J provided one peptide, fraction K provided two peptides, and fraction M provided one peptide (Fig. 2). The amino acid sequence of each peptide was obtained as summarized in Table I. We chose J1 and K1 peptides to construct oligonucleotide probes A and B. It was later found that peptide J1 was derived from lamp-1 protein and K1 was derived from lamp-2 protein (see below). A mixture of two probes was used to screen a Xgtll library of IMR-90 cdna. In the initial screening, a total of nine positive clones were obtained. When these clones were tested for reactivity of oligonucleotide probe A and probe B separately, it was discovered that each clone reacted with only one of the probes (Fig. 3). We therefore tested if the cdna clones could be grouped into two sets. Labeled cdna inserts from clone 5 and clone 9, respectively, were used for hybridization to cdnas from clones 1 to 9. As shown in the right side of Fig. 3, the results clearly indicate that those cdna clones represent two different sets of the molecules and each clone reacts with only those within the same set of cdnas. During these experiments, clone 3 was found to be unrelated to either set and was not tested further. Isolation of cdnas Encoding of the Full-length of h-lump- 1-For the first attempt, cdnas isolated from the IMR-90 library were isolated, based on their hybridization to oligonucleotide A. The predicted amino acid sequence of h-lamp- 1, based on this cdna clone (X-h-lamp-A-4), was described previously (2), and oligonucleotide A was denoted as oligonucleotide 1 in the previous report (2). This glycoprotein is called h-lamp-1 because it was found to have more similarity to the mouse counterpart m-lamp-1 than h-lamp-2, which was isolated by oligonucleotide probe B (as described below). H- lamp-1 was previously called lamp A (2). Although the amino acid predicted from the X-h-lamp-A-4 cdna was a nearly perfect match for the amino acid sequence of the isolated t.' I I I I Ebfm VoClrne (rnll FIG. 1. HPLC of peptides obtained after CNBr fragmentation. The glycoprotein fraction, which contained h-lamp-l and h- lamp-2, was cleaved by CNBr and subjected to reverse phase HPLC employing a Vydac TP214 C18 column, as described under "Experimental Procedures." The elution was performed by increasing concentration of acetonitrile (dotted line). Three peaks (J, K, M) were separately pooled as shown by horizontal bars. Ordinate, absorbance at 215 nm; abscissa, elution volume. 0

3 Molecular Cloning of Human Lysosomal Membrane Glycoproteins E N 0.03 v a, 0 C s O Fraction Number FIG. 2. Sephacryl S-200 gel filtration of peptides which were separated by HPLC. J (A), K ( B ),and b1 ( C ) peptides, isolated after HPLC in Fig. 1, were separately subjected to Sephacryl S-200 gel filtration as described under ExperimentalProcedures. Fractions were pooled as indicated by horizontal bars. Other fractions were found to contain negligible amounts of peptides by amino acid analysis. The vertical arrows indicate the elution positions of the void volume. TABLE I Amino Amino acid sequences of h-lamp-1 and h-lamp-2 and their derived peptides Dash indicates that no signal for amino acid was obtained. Asterisks indicate that the yield of those residues was low. Peptides acid sequences Amino acid residues JI N-VTVTLHDATIQAYLS-SSFSRGETRC.EQDRPP.-P. KI ALQL-ITQDKVASVININP SFVY-LSDTHLFP-AS -FTVRYETT-KTYKTV EKPEAGTY DPAFK AMFMVKNG-GTAC*IMA-FSAAFS LEL-LTDSE-ATC LYAKWQM-FT] Kz MI TsC TsD h-lamp-1 h-lamp-2 h-lamp-1, h-lamp-2, h-lamp-l, h-lamp-2,21-36 h-lamp-2, h-lamp-1, h-lamp-1,l-23 h-lamp-2, 1-23 Determined as S-carboxymethylcysteine. * The residues were obtained as a mixture, except for the first 10 residues, which were obtained for individual proteins (see the preceding paper (1)). Oligonucleotide A Insert From x 9 X ~ ~~ ~~ J Oligonucleotide B m, 9.. o I Insert From x 5 X m *- e * r, FIG. 3. Blot hybridization analysis of cdna clones for human lamp-1 and lamp-2 isolated from the IMR-90 cdna library. Xgtll clones (1-9) and control X g t l l (X) were cut by EcoRI and resolved by electrophoresis in 1% agarose gels. Followingtransfer to nitrocellulose, the blot was sequentially hybridized to oligonucleotide probes and cdna inserts. The hybridization to oligonucleotide probes was performed a t 45 C for A and 50 C for B, and lowstringency wash was performed as described under Experimental Procedures. Inserts from X5 and X9 clones were excised by EcoRI and purified by agarose gel electrophoresis. Those cdna inserts (850 base pairs from X5 and 700 base pairs from X9) were sequentially hybridized to thesame blot. protein, the NH2-terminal sequence deducedfromcdna analysis was different from that obtained on a human lamp1 glycoprotein purified from chronic myelogenous leukemia cells (2). To determine if another cdna could beisolated that more closely matches the protein sequence, a placental cdna library was screened using X-h-lamp-A-4 cdnaas probe. Positive clones were further tested by the oligonucleotides synthesized according to the sequences of h-lamp-1 (nucleotides 334 to 358 in Fig. 4) and h-lamp-:! (nucleotides 146 to 193 in Fig. 6). These oligonucleotides were chosen on the basis that the nucleotide sequence of the NH2-terminal half of the glycoproteins is more characteristic to each gene than thatin the COOH-terminal half. The cdna clones, whichpreferentially reacted with the h-lamp-1 oligonucleotide, were isolated and sequenced. Fig. 4 shows the nucleotide sequence of one of the clones, P-hL1-15B. The amino acid sequence deduced from the phl1-15b clone is consistent with that obtained from purified h-lamp-1 protein. Starting from the initiating methionine, a hydrophobic segment of27 amino acids is present. Following this, the sequence beginning at the alanine residue exactly matches the NH2-terminal sequence of the isolated human lamp-1. The amino acid sequence surrounding residue 1 (15) and the presence of arginine near the initiation methionine are consistent with the peptide segment of 27 amino acids serving as a leader peptide (16). The cdna sequence also includes a large segment of 3 -end untranslated region. These two portions were missing in the previous clone, perhaps due to S1

4 Molecular Cloning of Human Lysosomal Membrane Glycoproteins GAATTCGGGC GGGCTTCTTC GCTGCCGACG TACGACGAGT GGCCGGGCTC TTGCGTCTGG TAACGCGCTG TCTCTAACGC CAGCGCCGTC TCGCGCG CAC TGC GCA CAG ACC ACC CGC AGA CGC CCG GCA GTC CGC AGG CCC AAA CGC GCA CGC GAC CCC GCT CTC CGC ACC GTA CCC GGCGC CTC GGC ATG GCG CCC CGC AGC GCCGG CGA CCC CTG CTG CTG CTA CTG CCT GTT GCT GCT GCT CGG CCT CAT GCA TTG TCG TCA GCA Met Ala Pro Arg Ser Ala Arg Arg Pro Leu Leu Leu Leu Leu Pro Val Ala Ala Ala Arg Pro His Ala Leu Ser Ser Ala TTT ATG GTG AAA AAT GGC AAC GGG ACC GCG TGC ATA ATG GCC AAC TTC TCT GCT GCC TTC TCA GTG AAC TAC GAC ACC AAG AGT GGC Phe Met Val Lys Asn Gly Asn Gly Thr A l a m Ile Met Ala A:n Phe Ser Ala Ala Phe Ser Val Asn Tyr Asp Thr Lys Ser Gly 0 CCC AAG AAC ATG ACC TTT GAC CTG CCA TCA GAT GCC ACA GT GTG CTC AAC CGC AGC TCC TGT GGA AAA GAG AAC ACT TCT GAC CCC Pro Lys A:n Met Thr Phe Asp Leu Pro Ser Asp Ala Thr Val Val Leu A:n Arg Ser SermGly Lys Glu Afn Thr Ser Asp Pro AGT CTC GTG ATT GCT TTT GGA AGA GGA CAT ACA CTC ACT CTC AAT TTC ACG AGA AAT GCA ACA CG TAC AGC GTT CAG CTC ATG AGT Ser leu Val Ile Ala Phe Gly Arg Gly His Thr leu Thr Leu A:n Phe Thr Arg Afn Ala Thr Arg Tyr Ser Val Gln Leu Met Ser TTT GTT TAT AAC TTG TCA GAC ACA CAC CTT TTC CCC AAT GCG AGC TCC AAA GA ATC AAG ACT GT GAA TCT ATA ACT GAC ATC AGG Phe Val Tyr Asn Leu Ser Asp Thr His Leu Phe Pro Asn Ala Ser Ser Lys Glu Ile Lys Thr Val Glu Ser Ile Thr Asp Ile Arg 0 0- GCA GAT ATA GAT AAA AAA TAC AGA TGT GTT AGT GGC ACE CAG GTC CAC ATG AAC AAC GTG ACC GTA ACG CTC CAT GAT GCC ACC ATC Ala Asp Ile Asp Lys Lys Tyr Arg a Val Ser Gly Thr Gln Val His Met Asn Ain Val Thr Val Thr Leu His ASD Ala Thr 11% CAG GCG TA CT TCC AAC AGC AGC TTC AGC AGG GGA GAG ACA CGC TGT GAA CAA GAC AGG CCTCCCACC ACA GCG CCC CCT GC 799 Gln A1 a Tyr Leu Ser A:n Ser Ser Phe Ser Arg Gly G1 u Thr Arg m G 1 u G1 n Asp Arg Fro Ser Pro Thr Thr Ala Pro Pro Al! 176 CCA CCC AGC CCC TCG CCC TCA CCC GTG CCC AAG AGC CCC TCT GTG GAC AAG TAC AAC GTG AGC GGC ACC AAC GGG ACC TGCTG CTG 886 Pro Pro Ser Pro Ser Pro Ser Pro Val Pro Lvs Ser ProlSer Val Asp Lys Tyr Ap Val Ser Gly Thr Asn Gly ThrmLeu Leu 205 GCC AGC ATG GGG CTG CAG CTG AA CTC ACC TAT GAG AGG AAG GAC AAC ACG ACG GTG ACA AGG CTT CTC AAC ATC AAC CCC AAC AAG 973 Ala Ser Met Gly Leu Gln Leu Asn Leu Thr Tyr Glu Arg Lys Asp Asn Thr Thr Val Thr Arg Leu Leu Asn Ile Asn Pro Asn Lys 234 ACC TCG GCC AGC GGG AGC TGC GGC GCC CACT GTG ACT CTG GAG CTG CAC AGC GAG GGC ACC ACC GTCTG CTC TTC CAG TTC GGG 1060 Thr Ser Ala Ser Gly S e r m Gly Ala His Leu Val Thr Leu Glu Leu His Ser Glu Gly Thr Thr Val Leu Leu Phe Gln Phe Gly 263 ATG AAT GCA AG TCT AGC CGG TT TTC CTA CAA GGA ATC CAG TTG AAT ACA AT1 CTT CCT GAC GCC AGA GAC CCT GCC TTT AAA GCT 1147 Met Asn Ala Ser Ser Ser Arg Phe Phe Leu Gln Gly Ile Gln Leu Asn Thr Ile Leu Pro Asp Ala Arg Asp Pro Ala Phe Lys Ala 292 GCC AAC GGC TCCTG CGA GCG CTG CAG GCC ACA GTC GGC AAT TCC TAC AAG TGC AAC GCG GAG GAG CAC GTCGT GTC ACG AAG GCG 1234 Ala A:n Gly Ser Leu Arg Ala Leu Gln Ala Thr Val Gly Asn Ser Tyr L y s a Asn Ala Glu G1 u His Val Arg Val Thr Lys Ala 321 TT TCA GTC AAT ATA TTC AAA GTG TGG GTCA GCTTC AA GT GAA GGT GGC CAG TTT GGC TCT GTG GAG GAG TGT CTG CTG GAC 1321 Phe Ser Val Asn Ile Phe Lys Val Trp Val Gln Ala Phe Lys Val Glu Gly Gly Gln Phe Gly Ser Val Glu GlumLeu Leu Asp AAG AGG AGT CAC GCA GGC TAC CAG ACT ATC TAG CCT GGT GCA CGC AGG CAC AGC AGC TGC AGG GGC CTC TG TCC TTT CTC TGG GCT 1495 Lys Arg Ser His Ala Gly Tyr Gln Thr Ile end 389 TAGGGTCCTG TCGAAGGGGA GGCACACTTT CTGCAAACGTTCTCAAATC TGCTTCATCC AATGTGAAGTCATCTTGCA GCATTTACTA TGCACAACAG AGTAA 1600 CTATCGAAAT GACGGTGTT ATTTTGCTAA CTGGGTTAAA TATTTTGCT ACTGGTTAAA CATTAATATTACCAAAGTA GGATTTTcAG,G_G~G_G_G_G_G_TG_CTC_T_C_ 1705 TGAAGGTCAC TCGTGTGAGG CTGACATGCT CACACATTAC AACAGTAGAG AGGGAAAAT CTAAGACAGA GGAACTCCAG AGATGAGTGT CTGGAGCGGC TTCAG 2125 TTCAGCTTTA AAGGCCAGGA CGCGCGACAC GTGGCTGGCG GCCTCGTTCC AGTGGCGGCA CGTCCTTGGC GTCTCTAATG TCTGCAGCTC AAGGGCTGGC ACTTT 2230 TTTAAATAT AAAATGGTGTATTTTTATTTTTTTTGT AAGTGATTTTGGTCTTCTG TTGACATTCG GGTGATCCTG TTCTGCGCTG TGTACAATGT GAGAT 2335 CGGTGCGTTC TCCTGATGTT TTGCCGTGGC TTGGGGATTG TACACGGGACAGCTCACGT AATGCATTGCTGTAACAAT GT_A_ATAMAAA GCCTCTTTCTTCAA 2440 AAAAACCCCG AATTC 2455 FIG. 4. The nucleotide sequence and amino acid sequence of the human lamp-1 cdna clone, P-hL1-15B, isolated from a placental cdna library. The sequence of the insert for the total coding region, and deduced amino acid sequences, are shown. The amino acid sequence obtained from protein and peptide analysis is underlined. The sequence used for making a probe was underlined with a thick line. The positions that were shown as blank during amino acid sequencing are not underlined. The NH2-terminal amino acid is indicated by a square. Two regions of the sequence corresponding to the serine/proline-rich hinge region (solid line) and the putative transmembrane portion (dashed line) are indicated by boxes. Cysteine residues are circled, and potential N- glycosylation sites are indicated by dots. The repeated sequences in the 3'-untranslated region are underlined by a dotted line, and the consensus sequence for polyadenylation signal is underlined by two dotted lines. ""_

5 18924 Molecular Cloning of Human Lysosomal Membrane Glycoproteins A T G G G G G T G C T C T C T C T G A G G G G G T G G G G G T G C C G C T ~ T C T C T G A G G G G G T G G G G G T G C C G C T C T C T C T G A G G G G G T G G G G G T G C O G C T C T C T C [ G A G G G G I - I T G G G G G T G C C G C T C T I 1905 FIG. 5. Dot matrix analysis of the human lamp-1 3'-untranslated nucleotide sequence (A) and the alignment of the tandemly repeated nucleotide sequence (B). A, the nucleotide sequence is presented both on the horizontal and the vertical axes. Dots, which represent matched sequences, are obtained by the Microgenie (Beckman) program. B, the tandemly repeated sequences (nucelotides 168&1905) are aligned. The nucleotides are numbered as described in Fig. 4. nuclease digestion of nascent cdna during the IMR-90 library construction. All the other structural features are the same as those shown in the cdna clone isolated from the IMR-90 library (2). Thus, we believe that a full-length cdna clone for human lamp-1 has been isolated. This conclusion is supported by the fact that the size of human lamp-1 mrna was found to be about 2.2 kilobases (2). It is noteworthy that the amino acid sequence contains 19 potential N-glycosylation sites, although the human lamp-1 glycoprotein was shown to have 18 N-glycans (1). It is most likely that the N-glycosylation site at residue 352 is the one not utilized, because it is very close to the lipid bilayer (within 3 amino acid residues, approximately). When the presence of repeating sequences was tested in the whole cdna sequence by the matrix method, extensive repeating sequences were detected in the 3'-untranslated region of h-lamp-1 by computer search (Fig. 5A). Each of the repeats is essentially composed of GAGGGGTGGGGGTGCCGCT- CTCTCT, and the sequence is tandemly repeated nine times (Fig. 5B and the dotted line in Fig. 4). It is not known if the repeating sequences play any role in transcription or translation of h-lamp-1. It has been shown that the Ha-ras gene has tandem repeats of 28 base pairs in a position 3' to the coding region (17), and those repeated sequences were shown to have an enhancer activity (18, 19). Furthermore, it has been shown that the length of such tandem repeats are highly variable according to individuals, and those variations serve as a basis for restriction fragment length polymorphisms (20). It is noteworthy, however, that the tandem repeats in the Haras genes are present in introns, whereas the tandem repeats in the h-lamp-1 gene are located in transcribed, 3"untranslated sequences. It willbe interesting to determine if any function is associated with these repeats in the h-lamp-1 mrna. Predicted Primary Sequence and Structure of h-hrnp-2- cdna clones reacting with the oligonucleotide B were isolated, and three independent clones were found as shown in Fig. 3. Among them, clones 2 and 5 were found to contain cdna corresponding to the full-length coding region of h- lamp-2. The sequence of clone 2, I-hL2-2, is shown in Fig. 6B. The identical sequence was obtained from clone 5. This cdna contains nucleotide sequences encoding a 408-amino acid polypeptide, which contains a putative signal sequence and transmembrane domain (Fig. 6B). Analogous to h-lamp- 1, a considerably hydrophobic domain of 28 amino acids is present following the initiation methionine. The sequence following this domain exactly matches the NH2-terminal amino acid sequence determined from purified mature protein, starting from the leucine residue. The amino acid sequence surrounding residue 1 conforms with the leader peptide cleavage site (-3, -1) rule (15), and the presence of 1 arginine residue close to the initiation methionine also supports the idea that the peptide segment of 28 amino acids is probably a leader peptide, which is cleaved from the mature protein (16). The deduced amino acid sequence of h-lamp-2 has a hydrophobic segment (amino acid 346 to residue 369) of 24 amino acids near the COOH-terminus. This sequence is flanked on the COOH-terminal side by a putative cytoplasmic tail of approximately 10 amino acid residues. The amino acid sequence indicates that there are 16 potential N-glycosylation sites in h-lamp-2. Five of the 16 putative glycosylation sites are within the sequenced CNBr fragments. Each of these at positions 4, 10, 21, 30, and 214 gave a blank signal in amino acid sequencing, indicating that these asparagine residues are modified by N-linked saccharides. The partial and complete digestions of h-lamp-2 by endo-p-n-acetylglucosaminidase F provided evidence that h-lamp-2 contains 16-N-linked saccharides, as shown in the preceding paper (l), and the results indicate that all of 16 N-glycosylation sites identified in the deduced amino acid sequence are utilized in the mature protein. Beginning at the amino acid residue 171, a characteristic of a 20-amino acid-long domain with a high content of proline (6/20) and threonine plus serine (8/20) is found in the h- lamp-2 sequence. This segment corresponds to the previously identified hinge region of h-lamp-1, which in turn was found to have a homology to the IgA-al chain hinge structure (2, 21). Furthermore, the hinge-like structure in h-lamp-1 is enriched by proline and serine like the IgA-cy1 chain hinge region, whereas that of h-lamp-2 is enriched by proline and threonine. It is of interest to note that the hinge region of IgD heavy chain is enriched by proline and threonine (22). Both immunoglobulin hinge regions are glycosylated by 0- linked oligosaccharides (23, 24). It is, therefore, most likely

6 Molecular Cloning of Human Lysosomal Membrane Glycoproteins I AGA A Peptide Kl N- Ile Thr Glu Asp Lys Val Ala Ser Val Ile As" Ile Asn Pro Gly -F Oligo R 3'- TAG TGG GTC CTG TTC CAC CGE, CA: TAG TTG TAG TTG GG! :$$- 5 8 GAATTCTGAT GACCCAGGTC CAGCGAGCCT TTTCCCTGGT GTTGCAGCTG TTGTTGTACC GCCGCCGTCG CCGCCGTCGC CGCCTGCTCT GCGGGGTC ATG GTG 104 Met Val -27 FIG. 6. The oligonucleotide probe used for screening the Xgtll library (A) and the nucleotide sequence and deduced amino acid sequence of the human lamp-2 cdna clones (B). A, the peptide sequence and the nucleotide sequence of the probe based on most probable codon usages. The positions indicated by asterisks differ from the actual sequence shown in B. B, the sequence of the insert for the total coding region, I- hl2-2, and deduced amino acid sequences are shown. The denotions are explained in the legend to Fig. 4. TGC TTC CGC CTC TIC CCG GTT CCG GGC TCA GGG CTC GTT CTG GTC TGC CTA GTC CTG GGA GCT GTG CGG TCT TAT GCA TTG GAR CTT 191 QPhe Arg Leu Phe O I P Val Pro Gly Ser Gly Leu Val Leu ValQLeu Val Leu Gly Ala Val Arg Ser Tyr AlaUGlu Leu 3 AAT TTG ACA GAT TCA GAA AAT GCC ACT TGC CT TAT GCA AAA TGG CAG ATG AATTC ACA GTT CGC TAT GAA ACT ACA AAT AAA ACT 278 Afn Leu Thr Asp Ser Glu A:n Ala ThrmLeu Tyr Ala Lyr Trp Gln Met A:n Phe Thr Val Arg Tyr Glu Thr Thr A:n 32 TAT AAA ACT GTA ACC ATT TCA GAC CAT GGC ACT GTG ACA TAT AAT GGAGC AT7 TGT GGG GAT GAT CAG AAT GGT CCC AAA ATA GCA 365 Tyr Lys Thr Val Thr Ile Ser Asp His Gly Thr Val Thr Tyr A:n Gly Ser I l e a G l y Asp Asp Gln A m Gly Pro Lyr Ile Ala 61 GTG CAG TTC GGA CCT GGC TTT TCC TGG ATT GCG AATTT ACC AAGCA GCA TCT ACT TAT TCA AAT GAC AGC GTC TCA TTT TCC TAC 452 Val Gln Phe Gly Pro Gly Phe Ser Trp Ile Ala A:n Phe Thr Lys Ala Ala Ser Thr Tyr Ser Ile Asp Ser Val Ser Phe Ser Tyr 90 AAC ACT GGT GAT A4C ACA ACA TTT CCT GAT GCT GAA GAT AAA GGA ATT Cli ACT GTT GAT GAA CT TTG GCC ATC AGA ATT CCA TTG 539 Asn Thr Gly Asp Afn Thr Thr Phe Pro Asp Ala Glu Asp Lys Gly ile Leu Thr Val Asp Glu Leu Leu Ala Ile Arg Ile Pro Leu 119 AAT GAC CTT TTT AGA TGC AAT AGT TTA TCA ACT TTG GAA AAG AAT GAT GTT GTC CAA CAC TAC TGG GAT GTT CTT GTA CAA GCT TTT 626 A m Asp Leu Phe A r g m Asn Ser Leu Ser Thr Leu Glu Lys Asn Asp Val Val Gln His Tyr Trp Asp Val Leu Val Gln Ala Phe 148 GTC CAA AAT GGC ACA GTG AGC ACA AAT GAG TTC ClG TGT GAT AAA GAC AAA ACT TCA ACA GTGCA CCC ACC ATA CAC ACC ACT GTG 113 Val Gln A:n Gly Thr Val Ser Thr Asn Glu Phe Leu@ Asp Lys Asp Lyr Thr Ser Thr Val AlalPro Thr Ile His Thr Thr Val I17 CCA TCT CCT ACT ACACA CCT ACT CCA AAG GAA AAA CCA GAA CCT GGACC TAT TCA GTT AAT AAT GGC PAT GAT ACT TGT CTG CGT 800 Pro Ser Pro Thr Thr Thr Pro Thr Pro Lys Glu Lys ProlGlu Ala Gly Thr Tyr Ser Val Am Arn Gly A y Asp T h r a L e u Leu 206 GCT ACC ATG GGG CTG CAG CTG AAC ATC ACT CAG GAT AAG GTT GCT TCA GTT ATT AAC ATC AACCC AAT ACA ACT CAC TCC ACA GGC Ala Thr Met Gly Leu Gln Leu Ap Ile Thr Gin Asp Lys Val Ala Ser Val lie Asn Ile Arn Pro A;" Thr Thr His Ser Thr Gly AGC TGC CGT TCT CAC ACT GCT CTA CTT AGA CTC AAT AGC AGC ACC ATT AAG TAT CTA GAC TTT GTC TTT GCT GlG AAA AAT GAAAC SerQArg Ser His lhr Ala Leu Leu Arg Leu A:n Ser Ser Thr Ile Lyr lyr Leu Asp Phe Val Phe Ala Val Lys As" Glu Am CGA TTT TAT CTG AAG GAA GTG AAC ATC AGC ATG TAT TTG GTT AAT GGC TCC GTT TTC AGC ATT GCA AAT AAC AAT CTC AGC TAC TGG Arg Phe Tyr Leu Lys Glu Val A:n Ile Ser Met Tyr Leu Val Am Gly Ser Val Phe Ser Ile Ala Asn As" A:n Leu Ser Tyr Trp ATG CCCCA AGT TC TAT ATG TGC AAC AAA GAG CAG ACT GTT TCA GTG TCT GGA GCA TTT CAG ATA AAT ACC TTT GAT CTA AGG GTT Met Pro Pro Ser Ser Tyr HetmAsn Lys Glu Gln Thr Val Ser Val Ser Gly Ala Phe Gln Ile Asn Thr Phe Asp Leu Arg Val CAG CCT TTC AAT GTG ACA CAA GGA AAG TAT TCT ACA GCT CAA GAC TGC AGT GCA GAT GAC GAC AAC TT CTT GTG GCCjJ&-&cG-Xc Gln Pro Phe A:n Val Thr Gln Gly Lys Tyr Ser Thr Ala Gln AspmSer Ala Asp Asp Asp Asn Phe~-;:V:a:l:Py_o_!lep?a-??! TAG AAT CTG CAA CCGAA end that the 0-glycans found in human lamp-1 and lamp-2, shown in the preceding paper (l), are attached around these hinge regions. This hinge-like structure in h-lamp-2 appears to separate the molecule into two domains. Each domain contains 8 N-linked carbohydrates and 4 cysteine residues. The latter residues probably form disulfide bonds, since the molecule could not be digested well by trypsin unless it was first reduced with 2-mercaptoethanol followed by S-carboxymethylation. Another initiation methionine signal (nucleotides 9-11 in Fig. 6B) is present at 90 nucleotides upstream from the putative initiation methionine, but this sequence is possible due to an artifact of the cdna library construction, considering that it is next to the linker. This upstream methionine would produce very unusual peptides which contain (Arg)9 sequences, if this were the site of translation initiation. In addition, the putative initiation methionine at nucleotides would produce a leader peptide with a reasonable size, whereas the upstream methionine would produce a leader peptide of 58 residues, which is by far larger than reported sizes of leader peptides (16). Alternatively, the upstream initiation codon may not be translated due to its being buried in two consecutive stop codons, TGATGA. Those results, taken together with the findings that the putative initiation methionine produces a leader peptide with a comparable size to that of h-lamp-1, strongly suggest that the ATG at nucleotides codes for the initiation methionine. Further studies on genomic structure will be essential to clarify these points. Homology between Human lump-1 and lamp-2-based on the deduced amino acid sequences of human lamp-1 and lamp- 2, the homology between these proteins was examined. The results are shown in Fig. 7A and can be summarized as follows: 1) all of the cysteine residues are aligned well between human lamp-1 and lamp-2 sequences; 2) six of the N-glycosylation sites are at identical positions and four additional N-glycosylation sites are in close proximity; and 3) the homology is more prominent in the second half of the COOH-terminal TTC domain (40.5% matches) than the first half of the NH2- terminal domain (31.5% matches). In particular, the homology is the most prominent in the transmembrane and cytoplasmic portion. In contrast, the leader peptides, which are hydrophobic, have minimal similarity. In total, the two proteins share 36.7% matched sequences. When homologous sequences between human lamp-1 and lamp-2 were examined, the results shown in Fig.7Bwere obtained. The first homology lies in the hinge-like structure. In fact, this homology extends to the residue 177 (lamp-1) and 186 (lamp-2), when the nucleotide sequences are compared. The second and third homologous sequences are present around the first cysteine residue after the hinge structure. The fourth homologous sequence encompasses the transmembrane and cytoplasmic portion, and this segment is the longest stretch among the homologous sequences. It is plausible that the homologous structures play an important role shared by both proteins. Homology among lump-l Molecules-To determine if human lamp-1 is more similar to lamp-1 from other species or to human lamp-2, the deduced amino acid sequence of human lamp-1 was compared to available mouse lamp-1 (25) and chicken LEPlOO (26) sequences. The results of this comparison are shown in Fig. 8, and it can be seen that the alignment of these sequences is extensive. It was found that the human lamp-1 and mouse lamp-1 sequences have 66.1% identity, while 51.5% identical residues were found in the human lamp- 1 and chicken LEPlOO sequences. It is also evident that the homology is most extensive in the COOH-terminal portions. These results indicate that human lamp-1 is more homologous to related molecules from other species as distant as chicken and mouse than to human lamp-2, suggesting that lamp-1 and lamp-2 diverged relatively early in evolution, but that lamp-1 (and possibly lamp-2) structures have been strongly conserved during evolution. In fact, human lamp-2 was immunologically related to mouse lamp-2 (Z), and some monoclonal antibodies specific to human lamp-2 were found to react with monkey lamp-2.2 M. A. Williams, and M. Fukuda, unpublished results. 1343

7 lamp-1 S L V I A F G R G H T L T L N F T R N A T R Y S V ~ L M S F V Y N L S D T H L F P N A S S K E l K T V E S I T D I R A D I l 2 l lamp-2 K I A V Q U P Q F S W I A F K AUS T O 1 D S V U S U T G& T T U D D E D i G I L[l;lA D E L L A D 1 P L 119 lamp-1 D K K Y R C V S G T q V H M N N V T V T L H D A T I q A Y L S N S S F S R G E T R C E q D R P S P T T A P lamp-2 N 0 L F B N D L S T L E K O i O V p H Y M B V L V U F V QQG T V D T N U F L I D K D K 1 B Y U T lamp-1 P S P S P V P K S P S V D K Y N V S G T N G T C L L A S M G L q L N L T Y E R K D N T T V T R L L N I N P N K T S A S G lamp-2 ] P T T T P T P K E K P E A G T Y S V N N G N D T C L L A TcM G L L Nl I1 T Q D u * 1 A S V I G 1 T n H S T L n an 0 [ T I I H!?!Et;: lamp-1 q A T V G N S Y K C N A E E H V R V T K A F S V N l F K V W V q A F K V E G G q F G S V E ~ C L L D E N S T L I P I A V G lamp-2 Y W M P S m M a K u Q T B S igdqi S a T g D L R ~ P Q U K ~ Y S T ~ A 0 D B~ S A D T D D N F[Vl-l352 * B lamp ProSerProThrThrAlaPro 174 lamp ProSerProThrThrThrPro 184 * lamp AsnGly ThrCysLeuAlaSerMetGlyLeuGlnLeuAsnLeuThr 215 lamp AsnGlyAsnAspThrCysLeuAlaThrMetGlyLeuG~nLeuAsnlleThr 216 * * * lamp AsnIleAsnProAsnLysThr 235 lamp AsnIleAsnProAsnThrThr 231 * lamp LeuIleProIleAlaValGlyGlyAlaLeuAlaGlyLeuValLeuIle ValLeuIleAlaTyrLeuValGlyArgLysArgSerHisAlaGlyTyr 386 lamp LeuValProIleAlaValGlyAlaAlaLeuAlaGly ValLeuIleLeuValLeuLeuAlaTyrPheIleGlyLeuLysHisHisHisAlaGlyTyr 377 * * * * t * * FIG. 7. Alignment of the amino acid sequences of human lamp-1 and lamp-2 (A) and homology of the human lamp-1 and lamp-2 amino acid sequences (B). The alignment of two amino acid sequences and homology between them were tested by using the Microgenie (Beckman) program. In A, identical amino acids are boxed, N-glycan attachment sites are indicated by asterisks, and cysteine residues are circled. In B, the amino acid residues that differ are marked by asterisks. N A S S K E l K T V ED Sh-lamp-l li T TD IqR AK D Ia D NK K uy RAC VwS GD T I QR V H H[ N NY V T V [ KL H Dw A T RI Q Aw Y L S 150 F,~. 8, *lignment of the lamp-l- LEPlOO A T E G K V M V A T Q K S V I Q A I G T E Y R C 1 N S K Y V R M K H V N 1 f S N V T L E A Y P T 149 related molecule from human, mouse, and chicken origins. The :;:E:; 8; imt ~ ~ ~ [ ~ amino acid sequences of human lamp-1 LEPlOO N D T F S A N K T E C R D ti V S T T T V k T T m K H amino acids are boxed. $1 H V F@ R V T ~ K L A F@ S I N V ~ ~ F S K ~ V WQ V( QV Al FD K S V D E RO G Q F ~ ~ V N F Q V T D K LA V N V F N V Q V Q A F K V D ~ D K F G A M

8 FIG. 9. Comparison of the NHaterminal domain ( N) and the COOHterminal domain (0 of human lamp- 1 sequence. The human lamp-1 amino acid sequences of residues 1 to 169 (N domain) and 191 to 354 (C domain) were compared by the Microgenie alignment program. Identical residues are boxed, cysteine residues are circled, and N-glycosylation sites are indicated by asterisks. Note that all cysteine residues are aligned, and they have 22.4% matches in total. Molecular Cloning of Human Lysosomal Membrane Glycoproteins N C A M F M V K N G N G T A C I M A N F S A A F S V N Y D T K S G P K N M T F D L P S D A T V 45 V D K Y! U S G TlFl : i t L n : M G L Q L ; L T[E R D D! T V I R L M N I N P! K 234 R S S C G K E N T S D P S L V I A F G R G H T L T L N F T R N A T R V S V Q L M S F V 91 S GW A: L V T L E[ H S E[ T T V D L F I I G M D S S S R F F I Q G I N V L N C T S A 'd N N N V T V T L H D A T I Q YA L S N S S F S R G E T R C E Q D R P S 169 C K A F S N N I F K V W VUF K V E G G QIGS V(I1 E i L L i E N I T 354 A f COOH FIG. 10. Depicted structure of human lamp-1 (A) and lamp- 2 (B). N-Linked carbohydrate moieties and hinge-like structures are indicated by Y and W, respectively. It is assumed that the N- glycans attached closely to transmembrane are polylactosaminoglycan, and those at the NHe-terminal end are typical complex N-linked saccharides. N-Glycans in the third loop may be of intermediate size. Each presumptive loop is made by a disulfide bond. Cysteine residues connected to each other are tentatively assigned. The majority of the molecules reside in the luminal side of lysosomes. The membrane is indicated by a shaded area. Although this figure does not include 0- glycan attachment sites, it is likely that they are clustered around hinge-like structures as shown in IgA al-chain (23) and IgD &-chain (24). Close inspection of the deduced human lamp-1 amino acid sequence indicated that internal repeats exist between the NH2-terminal half and the COOH-terminal half of the luminal domain, which are separated by a hinge-like structure. As shown in Fig. 9, it is apparent that two domains of human lamp-1 constitute internal repeats with a good alignment of all cysteine residues. Interestingly, attempts to locate internal repeats in the deduced amino acid sequence on human lamp- 2 failed to reveal such a structure, and no alignment of cysteine residues was achieved. These results further support the hypothesis that the lamp-1 and lamp-2 genes have undergone extensive changes since diverging from the putative ancestor gene. These results also suggest that the ancestor gene for lamp-1 and lamp-2 may have evolved from duplica- tion of the precursor gene which would have encoded one domain consisting of about 170 amino acids, including 4 cysteine residues. It is conceivable that a gene segment coding for the hinge-like structure was inserted between duplicated genes and the transmembrane and cytoplasmic portions were added independently at the COOH-terminal of the gene. Studies on the genomic structures of lamp-1 and lamp-2 will be essential to make further hypothesis on the evolution of these related glycoproteins. Depicted Structure of Human lamp-1 and lamp-2-we pro- pose that human lamp-1 and lamp-2 have the structures depicted in Fig. 10, based on their deduced amino acid se- quences. The data discussed above suggest that each glycoprotein contains four loops, with each presumptive loop formed by a disulfide bond and each loop consisting of 36 and 39 amino acid residues. The distance between the first loop (NHz-terminal) and the second loop is the 74 residues in both h-lamp-1 and h-lamp-2. The distance between the third and fourth loop is 69 residues for h-lamp-1 and 63 for h-lamp-2. Alternatively, it is possible that loops are formed between the second and third cysteine and between the sixth and seventh cysteine. In this case, the first loop consists of74 residues (Cys5' to CyslZ7 in lamp-1 or Cys51 to CYS''~ in lamp-2), and the second loop consists of 68 residues (Cys"' to Cys3O0 in lamp-1) or 63 residues (CYS'~~ Cys301 to in lamp-2). The size of loop is in a range that is observed in immunoglobulin superfamily sequences (27). Furthermore, the COOH-terminal end of these putative loops contains a consensus sequence, Tyr-X-Cys, found in immunoglobulin-related domains (27). However, we detected no further homology between lamps and immunoglobulin-related domains in the sequences preceding and following the cysteine residues. If the cysteine residues in lamps are linked to form loops similar to those observed in immunoglobulin superfamily, it would suggest a potential receptor function for lamp molecules (27). It will, therefore, be important to determine how disulfide bonds are formed in lamps in order to understand the functional significance of these molecules. The glycoproteins are divided roughly in half by the presence of proline/serine- or proline/threonine-rich hinge-like regions. The NHf-terminal half consists of 169 amino acid residues, and the COOH-terminal half contains 155 amino acid residues both in h-lamp-1 and h-lamp-2. Both contain one transmembrane domain consisting of24 hydrophobic amino acid residues and an approximately 11-residue cytoplasmic tail. In these aspects, human lamp-1 and lamp-2 have strikingly similar structures. In addition, it is evident that human lamp- 1 and lamp-2 share very homologous structures in the third disulfide loop in Fig. 10, which begins at residue 200 or 201 (Fig. 7B). The functional significance of the third loop is not currently known, but its appearance in two glycoproteins, which seemingly diversified early in evolution, suggests that this structure plays an important role in the synthesis or function of lamp-1 and lamp-2. Similarly, very homologous sequences are present in the transmembrane and cytoplasmic segments. Specific signals have been identified for proteins remaining in the endoplasmic reticulum (28,29), those staying in the Golgi (30), and those transported to the nucleus (31). These signals have been found to reside in the COOH-terminal portion (28, 29, 31) or transmembrane and cytoplasmic regions of specific proteins (30). The data presented here suggest that the homologous transmembrane and cytoplasmic segments of lamp-1 and lamp-2 may serve as a molecular

9 18928 Molecular Cloning signal to target these glycoproteins to lysosomes. The deduced amino acid sequences of human lamp-1 and lamp-2 show that these glycoproteins have some differences in their sites for N-linked saccharides. H-lamp-1 has only one N-glycosylation site close to the transmembrane portion, whereas h-lamp-2 has four N-glycosylation sites close to the transmembrane portion. This difference could be important for the content of polylactosaminoglycans in the lamp-1 and lamp-2. We recently discovered that the distance of glycosylation sites from the membrane may determine the mode of glycosylation (32). When human chorionic gonadotropin a- chain was fused with the transmembrane plus cytoplasmic segment of vesicular stomatitis virus-g protein, the fused protein acquired polylactosaminoglycan, whereas the parent a-chain and vesicular stomatitis virus-g protein express only typical N-linked saccharides (33, 34). We proposed that, by bringing the glycosylation sites close to the transmembrane portion in the chimeric protein, the signal to acquire polylactosaminoglycans was formed (32). If proximity of glycosylation sites to the membrane is involved in the acquisition of polylactosaminoglycan, h-lamp-2 should contain more polylactosaminoglycan than h-lamp-1 since h-lamp-2 contains more glycosylation sites close to the transmembrane portion (see Fig. 10). The data presented in the preceding paper indicate this is the case (l), suggesting that our previous hypothesis applies to a naturally occurring polylactosaminoglycan-containing glycoprotein. It will be interesting to fur- ther test this hypothesis by engineering and expressing altered forms of human lamp-1 and lamp-2. Human lamp-1 and lamp-2 differ in amino acid sequences as described above. It is likely that the amino acid sequence of lamp-2 has been conserved during evolution, although this cannot be tested at this time since cdna for lamp-2 has not been isolated from other species. It is highly possible that lamp-1 and lamp-2 have distinct functions which emerged as soon as eukaryotic cells acquired lysosomes as subcellular compartments. It will be of interest to determine if non- homologous structures contribute to any function of these lysosomal membrane glycoproteins. Acknowledgments-We would like to thank Drs. Chen and August for communicating unpublished results on mouse lamp-1, Dr. Der for useful discussion, Drs. Krusius, Ruoslahti, and Millan for providing cdna libraries, Mat Williamson for amino acid sequencing, John Knight for oligonucleotide synthesis, Dr. Mark Williams for critical reading of the manuscript, and Tami Clevenger for secretarial assistance. REFERENCES 1. Carlsson, S. R., Roth, J., Piller, F., and Fukuda, M. (1988) J. Biol. Chem. 263, Viitala, J., Carlsson, S. R., Siebert, P. D., and Fukuda, M. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, Grant, G. R., Sacchezini, J. C., and Weglus, H. G. (1983) Biochemistry, 22, Lathe, R. (1985) J. Mol. Biol. 183, 1-12 of Human Lysosomal Membrane Glycoproteins 5. Ullrich, A., Coussons, L., Hayflick, J. S., Dull, T. J., Gray, A., Tam, A. W., Lee, Y., Yarden, Y., Liberman, T. A., Schlesinger, J., Downward, J., Mayes, E. L., Whittle, N., Waterfield, M. D., and Seeburg, P. H. (1985) Nature 309, Krusius, T., and Ruoslahti, E. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, Millan, J. L. (1986) J. Biol. Chem. 261, Woods, D. E., Markham, A. F., Ricker, A. T., Goldberger, G., and Colten, H. R. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manuul, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 10. Feinberg, A,, and Vogelstein, B. (1984) Anal. Biochem. 137, Siebert, P. D., and Fukuda, M. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, Siebert, P. D., and Fukuda, M. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, Huynh, T. V., Young, R. A., and Davis, R. W. (1986) in DNA Cloning, (Glover, D. M., ed) pp , IRL Press, Washington, D. C. 15. Von Heijne, G. (1983) Eur. J. Biochem. 133, Sabatini, D.D., Kreisbich, G., Morimoto, T., and Adesnik, M. (1982) J. Cell Biol. 92, Capon, D. J., Chen, E. Y., Levinson, A. D., Seeburg, P. H., and Goeddel, D. V. (1983) Nature 302, Seeburg, P. H., Colby, W. W., Capon, D. J., Goeddel, D. V., and Levinson, A. D. (1984) Nature 312, Krontiris, T. G., DiMartino, N. A., Colb, M., and Parkinson, D. R. (1985) Nature 313, Nakamura, Y., Leppert, M., O'Connell, P., Wolff, R., Holm, T., Culver, M., Martin, C., Fujimoto, E., Hoff, M., Kumlin, E., and White, R. (1987) Science 236, Tsuzukida, Y., Wang, C.-C., and Putnam, F. W. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, Takahashi, N., Tetaert, D., Debuire, B., Lin, L.-C., and Putnam, F. W. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, Baenziger, J., and Kornfeld, S. (1974) J. Biol. Chern. 249, Mellis, S. J., and Baenziger, J. U. (1983) J. Biol. Chem. 258, Chen, J. W., Cha, Y., Yuksel, K. U., Gracy, R. W., and August, J. T. (1988) J. Biol. Chem. 263, Fambrough, D. M., Takeyasu, K., Lippincott-Schwartz, J., and Siegel, N. R. (1988) J. Celt BioZ. 106, Williams, A. F., and Barclay, A. N. (1988) Annu. Reu. Immunol. 6, Paabo, S., Bhat, B. M., Wold, W. S. M., and Paterson, P. A. (1987) Cell 50, Munro, S., and Pelham, H. R. B. (1987) Cell Machamer, C. E., and Rose, J. K. (1987) J. Cell. Biol. 106, Burglin, T. R., and Robertis, E. M. D. (1987) EMBO J. 6, Fukuda. M.. Guan. J.-L., and Rose. J. K. (1988) J. Biol. Chem. 263, ' 33. Reading, C. L., Penhst, E. E., and Ballou, C. E. (1978) J. Biol. Chern Kessler, M. J:, Reddy, M. S., Shah, R. H., and Bahl, 0. P. (1979) J. Biol. Chem. 254,