THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 25, Issue of September 5, pp. 16818-16824,1991 Printed in U. S. A.

Cloning, Sequencing, and Expression of a cDNA Encoding Rat LIMP 11, a Novel 74-kDa Lysosomal Membrane Protein Related to the Surface Adhesion Protein CD36*

(Received for publication, February 1, 1991)

Miguel A. Vega$, Bartolome Segui-Real*, Jose Alcalde Garcia$, Carmela Cales8, Fernando Rodriguez$, Joel Vanderkerckhovev, and Ignacio V. Sandoval$ From the SCentro de Biologia Molecular, Facultad de Ciencias, Universidad Autonoma de Madrid, 28049 Madrid, Spain, the SCentro de Biomedicina, Consejo Superior de Investigaciones Cientificas, Madrid, Spain, and the TLaboratorium voor Genetica, Rijksuniversiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium

LIMP I1 is a glycoprotein expressed in the membrane of lysosomes and secretory granules with lysosomal properties. Sequence analysis of a CNBr-cleaved pep- tide allowed the synthesis of a 47-mer oligonucleotide that was used to screen a rat liver cDNA library in Xgtll. This resulted in isolation of a 2-kilobase cDNA containing 1,434 bases encoding the entire protein. The deduced amino acid sequence indicates that LIMP I1 consists of 478 amino acid residues. The segment spanning residues 4-6 to 26 constitute an uncleavable signal peptide. LIMP I1 possesses a hydrophobic amino acid segment near the carboxyl end, that together with the uncleaved signal peptide may anchor the protein to the membrane through two distant segments. The ma- jor portion of the protein resides on the luminal side and displays 11 potential N-glycosylation sites and 5 cysteine residues. Two short cytoplasmic tails, 2-4 and 20-21 amino acids long, correspond to the NH2- and COOH-terminal ends of the protein, respectively. Transfection of COS cells with the cDNA of LIMP I1 resulted in expression of the protein and its transport to lysosomes. Comparison of the entire sequence to various data bases of known proteins revealed exten- sive homology between LIMP I1 and the cell surface protein CD36 involved in cell adhesion. No significant homology was detected with the two families of lyso- somal membrane proteins A and B, recently described.

Lysosomes are organelles involved in degradation of extra- cellular and intracellular material (de Duve, 1963; Kornfeld and Mellman, 1989). Soluble and lysosomal membrane pro- teins are synthesized in ribosomes bound to the endoplasmic reticulum, co-translationally translocated through its mem- brane and transported to and through the Golgi complex together with proteins resident in the Golgi, targeted to secre- tory granules, and bound for the cell surface (Sly and Fisher, 1985; Skudlarek et al., 1984; Lewis et al., 1985; Barriocanal et

*This work was done in part with grants from the Comisi6n Interministerial para Ciencia y Tecnologia del Ministerio Espaiiol de Educaci6n y Ciencia and the Belgium National Fund for Scientific Research. 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 this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) M68965.

4 Supported by fellowships from the Ministerio Espaiiol de Edu- caci6n y Ciencia.

al., 1986; Lippincott-Schwartz and Fambrough, 1986; Chen et al., 1985). Most likely, sorting of lysosomal proteins occurs in the trans-Golgi network simultaneously with their packing into transporting vesicles coated with clathrin ( i e . primary lysosomes) (Lewis et al., 1985; Barriocanal et al., 1986; Lip- pincott-Schwartz and Fambrough, 1986; Chen et al., 1985; Geuze et al., 1984; Willingham et al., 1983). After uncoating, the vesicles are targeted to the endosomal network (Hopkins et al., 1990) and autophagic vacuoles (Dunn, 1990). As a result, secondary lysosomes are formed. Digestion of bound extra- cellular and intracellular material occurs in secondary lyso- somes, which functionally can be considered as a terminal compartment (de Duve, 1963; Kornfeld and Mellman, 1989). A few cells of the immune system also contain secretory granules with lysosomal properties, which, upon stimulation, discharge their content to the extracellular medium (Henkart, 1985).

In mammalian cells, the mechanisms involved in the sorting of soluble and membrane lysosomal proteins are different. Soluble proteins are sorted and transported from the Golgi complex by a mannose 6-phosphate signal introduced into asparagine-linked high mannose carbohydrates (Kornfeld and Mellman, 1989; von Figura and Hasilik, 1986; Kornfeld, 1987; Nolan and Sly, 1987; Robbins, 1988; Roth, 1988; Pfeffer, 1988). Membrane proteins are delivered to lysosomes by a mannose 6-phosphate-independent mechanism (Lewis et al., 1985; Barriocanal et al., 1986; Sandoval et al., 1989). Two of these proteins, lamp-1 and lysosomal acid phosphatase, have been shown to contain in their cytoplasmic tails a tyrosine involved in their transport to lysosomes (Williams and Fu- kuda, 1990; Peters et al., 1990).

Comparison of cDNA sequences and study of the distribu- tion of proteins coded by manipulated cDNAs have been powerful tools to characterize the signals involved in protein sorting (Garoff, 1985). Recently, the cDNAs of acid phospha- tase, and of two related families of lysosomal-associated mem- brane proteins (LAMPS), A and B, have been characterized (Pohlmann et al., 1988; Chen et al., 1988; Howe et al., 1988; Viitala et al., 1988; Fukuda et al., 1988; Noguchi et al., 1989; Cha et al., 1990).

Here we report the cloning and sequence of the cDNA of LIMP 11, a rat lysosomal membrane glycoprotein unrelated to proteins of classes A and B. LIMP I1 expressed in COS cells is specifically transported to lysosomes. Sequence com- parisons indicate that LIMP I1 is related with CD36, a surface protein involved in cell adhesion (Asch et al., 1987; Barnwell et al., 1985; Bernstein et al., 1982; Knowles et al., 1984; McGregor et al., 1989; Oquendo et al., 1990).

16818

Homology between Lysosomal LIMP 11 and Surface CD36 16819

EXPERIMENTAL PROCEDURES

Purification, Cleavage, and Amino Acid Sequencing of Rat LIMP ZZ-LIMP I1 (Barriocanal et al., 1986) was purified from rat liver. Eight livers were homogenized in 3 volumes of 20 mM Tris, 130 mM NaCI, pH 7.4 (buffer A) and centrifuged for 15 min at 10,000 X g, 4 "C, to remove debris and nuclei, and the supernatant was centri- fuged for 1 h at 100,000 X g, 4 "C. The resultingpellet was resuspended in 500 ml of buffer A plus 1% Triton X-100 and solubilized with continuous rotation for 1 h at 4 "C. The solubilized material was centrifuged for 30 min at 100,000 X g, 4 'C, and the resulting super- natant was passed over an affinity chromatography column contain- ing 500 pg of monoclonal antibody anti-LIMP I1 29G10 (Barriocanal et al., 1986) coupled to 0.25 ml of Sepharose 4B-CL. The column was washed with 5 volumes of buffer A plus 0.1% Triton X-100 containing 0.5 M NaCl, 2 volumes of water, and eluted with 0.25 M glycine-HC1, pH 2.75. The eluted protein was resolved from small contaminants by 12.5% SDS-PAGE,' eluted, and cleaved with CNBr. The resulting fragments were resolved by 15% SDS-PAGE, blotted onto glass fiber coated with Polybrene (Vandekerckhove et al., 1985), and stained with fluorescamine, and a 10-kDa peptide was NH,-terminal-micro- sequenced using a gas-phase sequenator (Applied Biosystems Inc., Model 478). The natural NH, terminus of the protein was also sequenced from the undigested protein using the same procedure.

Oligonucleotide Synthesis and cDNA Cloning-A randomized 47- mer oligonucleotide (5'-GGG CAC TCA CAG G/CA/TG TCA AAG

designed from the NHz-terminal amino acid sequence of the IO-kDa LIMP I1 CNBr-derived peptide (certainty factor 86%, stringent tem- perature 68 "C) (Lathe, 1985), was synthesized. The oligonucleotide was labeled at its 5' end with [y-32P]ATP (3000 Ci/mmol, Amersham Corp.) and T4 polynucleotide kinase and used as a hybridization probe to screen a X g t l l rat liver cDNA library (kindly provided by Dr. Frank Gonzalez, NIH, Bethesda, MD). A total of 9 X lo5 phage plaques on Escherichia coli strain Y1090 lawn cells were screened. The plaques were blotted onto nitrocellulose filters (Schleicher and Schuell) and prehybridized for 5 h at 48 "C in 6 X SSC, 0.05% sodium pyrophosphate, 0.1% SDS, 5 X Denhardt's solution, and 200 pg/ml salmon sperm DNA. Hybridization was performed at 42 "C for 20 h in the same buffer containing the radiolabeled probe (lo6 cpm/l3-cm nitrocellulose filter). The filters were washed for 30 min at 23 "C in 5 X SSC, 0.05% sodium pyrophosphate. Positive clones were sub- cloned to purity, and the cDNA inserts were tested by Southern blots using the same oligonucleotide probe.

DNA Sequencing and Computer-Assisted Data Analysis-cDNA from phage clone LIMP 11-31 was purified and subcloned into the EcoRI site of the Bluescript vector. DNA constructs used for sequenc- ing were derived by digestion with either exonuclease Bal-31 or restriction enzymes. Specific oligonucleotide primers were used when necessary. Sequencing was performed by the dideoxy chain terminator method (Sanger et al., 1977).

Hydrophobicity plots were generated using the Kyte and Doolittle algorithm (Kyte and Doolittle, 1982). PIR protein and EMBL-DNA data bases were scanned for searching homologies between LIMP I1 and known proteins using the FASTA and TFASTA programs, re- spectively (Pearson and Lipman, 1988). Sequence alignment between CD36 and LIMP I1 was carried out according to the LFASTA program (Pearson and Lipman, 1988).

Construction of Mutant t(431-478)-LZMP ZI-The mutant was constructed by introducing a stop codon before the second transmem- brane domain of LIMP 11, after nucleotide 1520, using the 22-mer oligonucleotide (5"GTGATTAACTGAACTTTGATTG).

DNA Transfection and Expression of LIMP ZZ in COS Cells-The 2-kilobase pair cDNA insert from clone LIMP 11-31 containing the entire coding sequence was ligated into the EcoRI site of the expres- sion vector pcEXV-3 (Miller and Germain, 1986). Plasmid was puri- fied by CsCl gradients and transfected into COS cells by the DEAE- dextran procedure (Lopata et al., 1984).

The cellular localization of the expressed protein was studied by double immunofluorescence using antibody 29G10 (Barriocanal et aL, 1986) and a rat monoclonal antibody against human lamp-1 (i.e. rat LIMP 111) (Hughes and August, 1982) reacting with the protein expressed constitutively in the lysosomal membranes of COS cells. Characterization of the translational product was performed by two- dimensional isoelectric focusing/PAGE analysis of the untreated and

' The abbreviations used are: SDS, sodium dodecyl sulfate; PAGE,

ACC CAG/T GTG/C CCA ATC/T TGC/G AG/AC/G ACC AT -3'),

polyacrylamide gel electrophoresis.

neuraminidase-treated polypeptide immunoprecipitated from cells metabolically labeled with 1 mCi of [35S]methionine (1,000 Ci/mmol, Amersham Corp.) overnight.

The localization and anchor of t(431-478)-LIMP I1 to cellular membranes was studied in COS cells transfected with the plasmid vector pcEXV-3 containing the cDNA encoding the truncated pro- tein. The localization of the protein was studied by immunofluores- cence microscopy as described above. The anchor of the protein to membranes was examined in transfected cells metabolically labeled by incubation with 1 mCi of [35S]methionine (1,000 Ci/mmol, Amer- sham Corp.) for 3 h and chased for 3 h with normal medium. Following that, the cells were washed with phosphate-buffered saline and dis- rupted by nitrogen cavitation, and cytosol and membranes were separated by centrifugation at 150,000 X g for 1 h, at 4 "C. The membrane pellet was treated with 0.1 M Na,C03 for 1 h at 4 "C and collected by centrifugation at 150,000 X g for 1 h at 4 "C. The Na2C03- washed membranes were then treated with 2% Triton X-100 for 45 min at 4 "C and centrifuged for 1 h at 150,000 X g. The cytosol, Na2CO3 supernatant, and the pellet and supernatant resulting from treating the Na2C03-treated membranes with Triton X-100 were analyzed for their content in LIMP I1 by immunoprecipitation. Im- munoprecipitation was performed using an anti-LIMP I1 rabbit pol- yclonal antibody bound to protein A-Sepharose. The immunoprecip- itates were run on SDS-PAGE and analyzed by autoradiography.

RESULTS

Cloning and Sequence of LIMP 11-Screening of a rat liver cDNA library in X g t l l with the mouse monoclonal anti-LIMP I1 antibody 29G10 and a rabbit polyclonal antibody raised against the protein purified by affinity chromatography failed to detect any LIMP 11-positive clones.

For further screening of the library, both the natural amino terminus of the protein (NHz-ARCCFYTAGTLS) and the amino terminus of a 10-kDa CNBr-derived peptide (NHz- MVLQNGTWVFDSCECPPLPVYIQ) were determined. A 47-mer degenerated oligonucleotide (5'-GGG CAC TCA CAG G/CA/TG TCA AAG ACC CAG/T GTG/C CCA ATC/T TGC/G AG/AC/G ACC AT -3') intended to hybridize the bases encoding the first 16 amino acids of the CNBr peptide was designed. With this oligonucleotide, 15 positive clones were isolated. Clones 3 and 31 produced the strongest positive signals after hybridization under high stringency conditions. However, clone 3 was not template for the T7 polymerase when the oligonucleotide was used as a primer. In contrast, under the same conditions, the cDNA of clone 31 was effi- ciently read.

Fig. 1 shows the nucleotide sequence of the 2-kilobase pair cDNA insert of clone 31. The putative methionine codon closest to the 5' end of the insert initiated the longest (1434 nucleotides) possible open reading frame. This reading frame encoded the first 12 microsequenced amino acids of the nat- ural amino terminus of the protein, and 20 of the 23 amino- terminal residues of the CNBr peptide, thus confirming the identity of the cDNA (Fig. 1). The cDNA contained 5'- and 3"untranslated flanking sequences, 230 and 274 nucleotides long, respectively. A consensus polyadenylation sequence (AATAAA) was found in the untranslated 3' sequence, at position 1910.

The initiator methionine is not found in the mature protein (Fig. 1). The first four NHz-terminal amino acids of the mature protein (Ala,Arg,Cys,Cys), as determined by protein microsequencing, are followed by a hydrophobic stretch of 21 amino acids, as shown by hydropathy analysis (Fig. 2). Its hydrophobicity and location strongly suggest that amino acids 6 to 26 constitute the core of an uncleaved signal peptide (Figs. 1 and 2). Admitting that the polypeptide stripped of the initiator methionine is not proteolytically processed further, the mature polypeptide would be 477 amino acids long and would have a molecular weight of 53,000. This calculated molecular weight is in close agreement with the 47,000 of the

16820 Homology between Lysosomal LIMP 11 and Surface CD36

FIG. 1. Nucleotide and deduced amino acid sequences of LIMP I1 cDNA. Nucleotides are numbered, a t the right of each line, starting with the first base following the linker used for cDNA cloning. The deduced protein se- quence is shown below the DNA se- quence with the amino acids numbered at the right of each line; the white arrow indicates the natural amino terminus of the protein; - - -, amino acid sequences obtained by microsequencing; #, cysteine residues; *, potential asparagine-linked glycosylation sites; the hydrophobic core of the signal peptide and the stop trans- fer signal are boned; the black arrow in- dicates the stop codon; =, polyadenyl- ation signal in the 3"untranslated se- quence.

90

180

2 70 13

364 43

450 73

540 103

630 133

720 163

810 193

900 223

990 25 3

1080 283

1170 313

126D 343

1350 373

1440 403

1530 533

1620 c63

1710 4 78

1800

1890

1938

Endo-F-treated LIMP I1 determined by SDS-PAGE (Barrio- canal et al., 1986).

Hydropathy analysis revealed another strong hydrophobic region between residues 433 and 457 close to the carboxyl terminus (Fig. 2). This region could act as a stop transfer signal anchoring the protein to the membrane.

Between the signal peptide and the stop transfer signal was a segment 406 amino acids long (Fig. 1). All of the 11 potential asparagine-linked (Am-X-Ser or Asn-X-Thr) glycosylation sites were found spread throughout this segment (Figs. 1 and 2). Its location between the signal peptide and the stop trans- fer signal, overall hydrophobicity, high content in carbohy- drates, and complete resistance to digestion when intact ly- sosomes were incubated with proteinase K (Barriocanal et al., 1986) makes this region a strong candidate for the luminal domain. Interestingly, the 5 cysteine residues present in this domain were clustered in the carboxyl-terminal half (Fig. 1).

The 20-21 amino acids that followed the stop transfer signal would then constitute a cytoplasmic tail (see "Discussion").

Expression of LIMP 11 in COS Cells-Transfection of COS cells with rat LIMP I1 cDNA resulted in exclusive expression of the protein in the membranes of vesicles clustered around the nucleus (Fig. 3A) . The vesicles were identified as lyso- somes by the co-distribution of the transfected LIMP I1 with the endogenous lysosomal membrane protein LIMP I11 (Fig. 3, A and B ) .

When proteins metabolically labeled by incubating the cells with [35S]methionine were immunoprecipitated using the spe- cific monoclonal anti-LIMP I1 antibody 29G10, a polypeptide with an apparent molecular weight of 74,000, and a PI of 5.8 was immunoprecipitated (Fig. 4, top). Treatment of the pro- tein with neuraminidase resulted in a small shift in PI, 6.4 (Fig. 4, bottom), indicating that the protein contained a sig- nificant, albeit small, amount of sialic acid residues. Both the molecular weight and PI of the protein immunoprecipitated from COS cells were identical with the ones of LIMP I1 isolated from NRK cells (Barriocanal et aZ., 1986) and rat liver (not shown). This result indicated that LIMP I1 was

Homology between Lysosomal LIMP 11 and Surface CD36 16821

FIG. 2. Hydropathy plot of LIMP 11. The hydropathy of 11 consecutive amino acids was calculated using the al- gorithm of Kyte and Doolittle (Kyte and Doolittle, 1982). Each point represents a single amino acid. Potential asparagine- linked glycosylation sites, *, predicted membrane spanning regions, boxes.

* * * * * * * - 4 ~ ' " " " " " " ' " " " " " '

6 25 44 63 82 101 126130158177 105215 234253272291 310323 348 367 3864054244.43 462

Residue number

2 0 0 -

26 -

FIG. 3. Expression of LIMP I1 in COS cells. The expression and cellular localization of LIMP I1 was studied by immunofluores- cence microscopy in cells stained with the mouse monoclonal anti- rat LIMP I1 antibody 29G10 (A, rhodamine channel), and a rat monoclonal anti-human LIMP I11 antibody ( B , fluorescein channel). Arrows indicate the exclusive staining of lysosomes with the anti- LIMP 111 antibody in untransfected as well as in the transfected cell. Bars, 12 pm.

200-

98-

normally processed in COS cells. The Uncleaved Signal Peptide Anchors LIMP 11 to the

Membrane-Transfection of COS cells with t(431-478)- LIMP 11, a polypeptide containing the leader sequence and luminal domain of LIMP 11, resulted in expression of the protein in the endoplasmic reticulum and plasma membrane, as shown by immunofluorescence microscopy (Fig. 5, panel A ) . This localization was consistent with slow transport of the newly synthesized protein from the endoplasmic reticulum and its presentation on the cell surface. I t also suggested that LIMP I1 was anchored to the membrane through the un- cleaved signal peptide. To further confirm this, transfected cells were metabolically labeled with ["S]methionine, and the distribution of "S-labeled t(431-478)-LIMP I1 in the soluble and membrane cellular fractions was studied by immunopre- cipitation using a polyclonal anti-LIMP I1 antibody (Fig. 5 , panel B ) . Both the truncated protein and endogenous LIMP I1 were quantitatively recovered in the microsomal fraction. Moreover, after treating the membranes with 0.1 M NazCOa,

68-

2 0

26-

pH

FIG. 4. Biochemical characterization of LIMP I1 expressed in COS cells. Cells were metabolically labeled with ["SS]methionine overnight, immunoprecipitated, treated without (top panel) and with (bottom panel) neuraminidase, and analyzed by two-dimensional iso- electric focusing/PAGE.

16822 Homology between Lysosomal LIMP 11 and Surface CD36 I 2 3 - 8: - 84

- 48

ur -36 " - f

A B

FIG. 5. t(431-478)-LIMP I1 expressed in COS cells is firmly bound to membranes and is localized in the endoplasmic retic- ulum and plasma membrane. The cellular distribution of truncated LIMP I1 was studied by immunofluorescence microscopy using anti- body 29G10, as shown in panel A. Note the strong staining of both the endoplasmic reticulum (extended throughout the cytoplasm and wrapping the nucleus) and plasma membrane. Bars, 12 pm. The expressed protein was analyzed by immunoprecipitation using the polyclonal anti-LIMP I1 antibody, panel B. Lune I , supernatant of microsomal membranes washed with 0.1 M NazCOs. Lunes 2 and 3, supernatant and pellet, respectively, of the NazCOn washed mem- branes treated with 2% Triton X-100. The 74-kDa protein ( a ) , identified as endogenous LIMP 11, was the only polypeptide found in untransfected cells reacting with the anti-LIMP I1 polyclonal anti- body. The 69-kDa protein ( b ) , found only in transfected cells, exhib- ited the molecular mass expected for t(431-478)-LIMP 11. The 36- kDa polypeptide (c) is a major degradation product of the truncated protein, as shown by its increase in extracts from transfected cells that when stored at 4 "C for prolonged times showed a parallel decrease in t(431-478)-LIMP 11.

most of the truncated protein and all of the endogenous LIMP I1 were recovered with the washed membranes. Since only integral membrane proteins remain bound to membranes treated with Na2COs, this result indicated that the truncated protein was firmly fastened to the membranes through the uncleaved signal peptide. Treatment of the Na2C03 membrane pellet with 2% Triton X-100 resulted in complete solubiliza- tion of the truncated and intact forms of LIMP I1 as corre- sponds to membrane proteins.

LIMP IZ Protein Is Structurally Related to CD36-A search into the EMBL DNA data base using the TFASTA algorithm revealed that LIMP I1 protein showed an extensive homology with the human cell surface protein CD36 (Oquendo et al., 1990). The alignment of their primary amino acid sequences is shown in Fig. 6. It can be seen that besides the large number of positions displaying conserved substitutions, both proteins shared 33.9% of the amino acids. This identity was found spread throughout the entire sequence of the luminal domain. Like LIMP 11, CD36 displayed two hydrophobic regions near the amino and carboxyl ends. Similarly, CD36 also has an uncleaved signal peptide (Tandon et al., 1989a). There are 10 potential N-glycosylation sites in CD36, 3 of them are con- served in LIMP 11, and the rest are in close proximity. Furthermore, the cysteine residues a t positions 4, 245, 274, 312,318,329, and 458 of LIMP I1 are preserved in CD36. All these data strongly suggest that both proteins possess an identical topology and a similar structure.

DISCUSSION

In an effort to identify the signal(s) involved in the sorting and transport of lysosomal membrane proteins from the Golgi complex to lysosomes and characterize their structure and function, we have cloned the cDNA encoding the rat lysoso- mal protein LIMP 11.

Clone LIMP I1 isolated from a rat cDNA library in Xgtll contains the full coding sequence of the protein as shown by: the presence of the codon corresponding to the initiator methionine in the 5' end; the fact that the initiation codon is followed by the sequence of oligonucleotides encoding the

LIMP I1

CD36

LIUP 11

CD36

LIUP I1

CD36

LIMP I1

CD36

LIUP I1

CD36

LIMP I1

CD36

LIMP11

CD36

LIMP I1

CD36

# # 10 20 30 40 50 6 0

YIQFYFPHVPNPEEIIQGEIPL-LEEVGPY~R-ELPJlKFGENGTl'ISAVTNWtYIFER

Y R Q F W I P D V Q N P P ~ S S N I Q ~ Q R G P ~ R ~ F ~ ~ D ~ D N ~ S F I 4 ~ G A I F E P

* 7 0 8 0 9 0 100 110 1 2 0

. . . . . . . . . ... . . . . . . . . . . ............................................... 7 0 9 0 100 1 1 0 1 2 0

1 3 0 1 4 0 150 1 6 0 1 7 0 1 8 0 N Q S V G D P T V D L I R T I N X P L C T V V E I U a a P P U ( E I 1 E ~ ~ ~ G Y ~ ~

S L S V G T E A D N F T - V L H W S H I Y Q N Q W Q ~ I L N S L I H Y R D P F . . . . . . . . . . . . ........... .................................. ... :...

1 3 0 * 140 150 160 170 180

1 9 0 200 210 220 2 3 0 240 # l S L V H I P R P D V S P N K L P Y E R G ~ D G E ~ L T G E D N Y L N P P

LSLVP-YP--VmVGLFYPYNNTADGVYKWNGKDNISWAIIDTYKGKFlNLSYWESH-CDU . . . . . . . . . . . . . . . . . . . . ............................................... ....

1 9 0 2 0 0 210 . 2 3 0 2 4 0 #

2 6 0 270 # 2 8 0 2 9 0 300 3 1 0 I N G T D G D S P H P L I S K D E T L Y I F ~ D F C R S ~ I T F S S F ~ G ~ A ~ Y ~ ~ I ~ S S E N A G

I N G T D M S F P P P V E K S Q ~ F F S S D I C R S I Y A ~ S ~ ~ G I ~ ~ ~ F ~ ~ N P D . . . . . . . . . . . . . . . . . . . . . . . . . . . ........................................................

+250 260 270 # 2 8 0 2 9 0 300

# 315 # 4 I 340 350 3 6 0 - F C - - - - - I P E G N C I I D A G V L N V S I C X N G A P l I W S A I K ~ P N K E E H E S W

~ C P C T E R I I S K N ( I T S Y G V L 1 S K C K E G R P W I S L P H P L Y A Y L # # 320*# 330 # 340 3 5 0 3 6 0 3 7 0

. . . . . . . . . . . . . . . . . . . . . . . . ............................................

370 380 390 400 4 1 0 4 420 D I H P L T G I I ~ G A K R P Q I N D F - ~ N I ~ ~ ~ S ~ D ~ ~ Q ~ S V I

D I E P I T G P T I 4 F A I ( R ~ ~ I V N L G V K P S E K I Q ~ ~ I W I ~ ~ T I G D E ~ ~ S Q V . . . . . . . . . . . . . . . . . . . . . . . ................................................

380 390 400 410 *420 430

STDEGTADEIUPLIRT 470

440 450 460 # # 470

FIG. 6. Aligned amino acid sequences of LIMP 11 and CD36. Amino acids are numbered above or below the lines. CD36 data are from Oquendo et al., 1990. Gaps (- - -) have been introduced to optimize alignments. Cysteine residues, #;potential asparagine-linked glycosylation sites, *. Hydrophobic cores of the signal peptides and stop transfer signals are enclosed into boxes. Double dots appear above or below identical residues. Single dots indicate conserved residues.

uncleaved signal peptide preserved in the mature protein; the occurrence after the reading frame of the termination codon TGA; and finally, the expression and transport of the protein to lysosomes when the cDNA is transfected into COS cells.

With respect to the 5' and 3' sequences that flank the coding sequence and could be important parts of the mRNA, it is important to note that the poly(A) tail is missing from the 3'-untranslated sequence, and that the size of the mRNA, as determined by Northern blot (not shown), exceeds in 0.4 kb the length of the cDNA insert contained in clone LIMP 11. Whereas it is clear that the 3"untranslated sequence contained in clone LIMP I1 is incomplete, determination of its entire sequence as well as that of the 5"untranslated sequence must await either sequencing of new complementary cDNA clones or study of the gene. Nevertheless, it is clear that truncation of the flanking sequences does not interfere with translation since the protein is expressed in COS cells transfected with the LIMP I1 cDNA.

The 74-kDa rat LIMP I1 has been biochemically character- ized in our laboratory together with four other lysosomal membrane proteins: the 35-50-kDa LIMP I, 90-100-kDa LIMP 111, 110-kDa LIMP IV (Barriocanal et al., 1986), and 80-kDa LIMP V (Bonifacino et al., 1990). Comparison of partial amino acid sequences has revealed that LIMP I11 and LIMP IV belong to the A and B classes of lysosomal mem- brane proteins (Chen et al., 1988; Howe et al., 1988; Viitala et al., 1988; Fukuda et al., 1988; Noguchi et al., 1989; Cha et al., 1990), respectively.2 Both classes of proteins possess between

* I. V. Sandoval, unpublished results.

Homology between Lysosomal LIMP 11 and Surface CD36 16823

380 and 396 amino acids and 16 and 20 asparagine sites for glycosylation. They have two homologous luminal domains, each with 4 cysteine residues, separated by a segment 20-35 amino acids long, rich in proline and serine (class A) or proline and threonine (class B). Furthermore, they have short, 10- 11-amino-acid-long cytoplasmic tails displaying 19% identity and conserved Gly-Tyr pairs. Class A proteins are rat lgp 120 (which is the same as rat LGP 107), mouse and human LAMP- 1, and chicken LEP-100. Rat lgp 110 (equivalent to rat LGP 96), mouse lgp 110, and human LAMP-2 constitute class B.

Comparison of primary structures reveals that LIMP I1 is not a member of the lysosomal membrane protein classes A and B. It is also relevant that, whereas proteins A and B have cleavable signal peptides, the signal peptide of LIMP I1 is not cut.

The transport of t(431-478)-LIMP I1 to the plasma mem- brane, its partitioning with Na2C03-treated membranes, and its solubilization by Triton X-100, all indicate that the protein is anchored to the membrane through the uncleaved signal peptide. This result suggests that the signal peptide also anchors the intact LIMP I1 to the membrane (see below). Further studies are needed to determine the boundaries of the signal peptide. It is particularly important to establish if cysteines 4, 5, and 458 are or are not embedded into the membrane. Pairing between cysteines 458 and 1 of the 2 cysteines located at the NHp-terminal segment might bring the two transmembrane domains together. Furthermore, since paired cysteines are known to become extremely hydrophobic, after formation of a disulfide bond they are likely to be embedded into the membrane. It is also noteworthy that LIMP I1 is palmitylated in uiuo.' Palmitic acid is frequently post-translationally incorporated into cysteine residues and plays an important role in anchoring proteins to membranes. In either case, cysteine pairing, palmitylation, or both, would increase the stability of the protein anchor to the membrane.

The luminal domain of LIMP I1 does not show the splitting into two homologous segments characteristic of proteins A and B and has a different distribution of cysteine residues and asparagine sites for glycosylation. The clustering of all 5 cysteine residues within the carboxyl half of the domain suggests its division into discrete structural regions.

The distinct hydrophobic segment, residues 433-458, that follows the luminal domain is a strong candidate to act as a stop transfer signal and to anchor the protein to the mem- brane. Both the existence of this and evidence showing that the uncleaved signal peptide is inserted to the membrane indicate that LIMP I1 is anchored to the membrane through two points. If this is the case, most likely the protein will be constituted by one luminal, two transmembrane, and two cytoplasmic domains. According to this model (Fig. 7), the luminal domain would comprise the larger part of the protein, residues 27-432.

To our knowledge the only protein insofar described dis- playing this topology is the bacterial aspartate receptor (Russo and Koshland, 1983). The lysosomal membrane is known to contain numerous receptors involved in the transport of amino acids and other digestion-derived products to the cy- toplasm. Although LIMP I1 and the bacterial aspartate recep- tor show no obvious sequence homology, this finding raises the possibility that LIMP I1 is itself a receptor, and, by extension, that other receptors might display similar topolo- gies.

Although no apparent similarity exists between the cyto- plasmic tails of proteins A, B, and acid phosphatase (Chen et al., 1988; Howe et al., 1988 Viitala et al., 1988; Fukuda et al., 1988; Noguchi et al., 1989; Cha et al., 1990; Pohlmann et al.,

0 Lumen

N H ~ - A 0 R R

0

s T o

R E O A

Ea Cytoplasm

* P L l R ~ . c w n

FIG. 7. Cartoon depicting the main features of LIMP 11. The dotted box illustrates the membrane. Cysteine residues, C. Circles represent oligosaccharide chains attached to the potential glycosyla- tion sites.

1988), the Tyr contained in their cytoplasmic tails has been shown to play a role in their sorting and/or transport to lysosomes (Williams and Fukuda, 1990; Peters et al., 1990). However, examination of the two cytoplasmic tails of LIMP I1 reveals that they do not contain any tyrosine residues.

Data available in the literature indicate that the traffic of lysosomal membrane proteins in the cell is indeed a very complex process and may involve more than one sorting signal. Whereas the proteins after moving from the endo- plasmic reticulum to and through the Golgi complex appear to be preferentially and directly transported to lysosomes (D'Souza and August, 1986; Barriocanal et al., 1986), small amounts of some of them can be detected on the cell surface (Lippincott-Schwartz and Fambrough, 1987; Bonifacino et al., 1990). Moreover, lysosomal protein carriers (Brown et al., 1986; Goda and Pfeffer, 1988; Geuze et al., 1988; Griffiths et al., 1988; Jin et al., 1989) and lysosomal membrane proteins exposed on the cell surface (Bonifacino et al., 1990) are recycled for further use. To fulfill all these needs, the path- ways through which proteins move must possess sorting points. In these probably reside devices operating on specific sorting signals contained in the proteins and their carriers (Lobe1 et al., 1989; Glickman et al., 1989; Narendra et al., 1989).

Sequence comparison studies have revealed the unexpected property that LIMP I1 is evolutionarily related with the surface glycoprotein CD36. The possibility should be consid- ered that LIMP I1 and CD36 belong to a new family of proteins. The identity between the primary structures of the two proteins is confined to, and spread throughout, the extra- cytoplasmic tracts. CD36 appears to be involved in the at- tachment of cells to the extracellular matrix through its interaction with collagen (Tandon et al., 198913) and perhaps with thrombospondin (Asch et al., 1987; McGregor et al., 1989). Furthermore, it recognizes a component of the mem- brane of red blood cells infected with Plusmodium falciparum (Ockenhause et al., 1989). It would be interesting to study if LIMP 11, in spite of its intracellular location, has conserved some of these properties. With regard to this, it is remarkable that changes in the expression and processing of surface proteins related to lysosomal membrane proteins A and B have been correlated with decreased cell adherence and trans- formation (Chen et al., 1988).

In contrast to the similar luminal domains, the two trans- membrane and cytoplasmic regions show no obvious homol-

16824 Homology between Lysosomal LIMP II and Surface CD36

ogy. These dissimilarities, the different cellular location of the two proteins and the probable transport of CD36 to the plasma membrane by the default pathway, are consistent with a role of the cytoplasmic tail(s) and/or the transmembrane domain(s) in the sorting and transport of LIMP I1 and, by analogy of other membrane proteins, to lysosomes. These possibilities could be tested by studying the transport of hybrid proteins constructed by exchanging domains between CD36 and LIMP 11.

Acknowledgments-We thank Dr. T. August for the gift of mono- clonal anti-lamp-1 antibody and the Fundaci6n Ram6n Areces for its support. We acknowledge the technical assistance of J . Van Damme and M. Puype.

REFERENCES

Asch, A. S., Barnwell, J., Silverstein, R. L., and Nachman, R. L.

Barnwell, J. W., Okenhouse, C. F., and Knowles, D. M. (1985) J.

Barriocanal, J., Bonifacino, J., Yuan, L., and Sandoval, I. V. (1986)

Bernstein, I. D., Andrews, R. G., Cohen, S. F., and McMaster, B. E.

Bonifacino, J. S., Yuan, L., and Sandoval, I. V. (1990) J . Cell Sci. 92,

Brown. W. J.. Goodhouse. J.. and Farauhar. M. G. 11986) J. Cell Biol.

(1987) J. Clin. Inuest. 7 9 , 1054-1061

Zmmunol. 135,3494-3497

J. Biol. Chem. 261 , 16755-16763

(1982) J. Zmmunol. 128,876-881

701-712 . .

103; 1235-1247 ~, . .

Cha. Y.. Holland. S. M.. and Aumst. T. (1990) J. Biol. Chem. 265. 5008-5013 ’

Chen, J. W., Pan, W., D’Souza, M. D., and August, T. (1985) Arch. Biochem. Biophys. 2 3 9 , 574-589

Chen, J. W., Cha, Y., Yuksel, K. U., Gracy, R. W., and August, T. (1988) J. Biol. Chem. 263,8754-8758

de Duve, C. (1963) Ciba Found. Symp. 1-35 D’Souza, M. P., and August, T. (1986) Arch. Biochem. Biophys. 2 4 9 ,

Dunn, W. A. (1990) J. Cell Biol. 110 , 1923-1933 Fukuda, M., Viitala, J., Matteson, J., and Carlsson, S. R. (1988) J.

Garoff, H. (1985) Annu. Reu. Cell Biol. 1, 403-445 Geuze, H. J., Slot, J. W., Serous, G. J. A. M., Hasilik, A., and von

Figura, K. (1984) J. Cell Biol. 9 8 , 2047-2054 Geuze, H. J., Stoorvogel, W., Strous, G. J., Slot, J. W., Bleekemolen,

J. E., and Mellman, I. (1988) J. Cell Biol. 107,2491-2501 Glickman, J. N., Conibear, E., and Pearse, B. M. F. (1989) EMBO J.

Goda, Y., and Pfeffer, S. R. (1988) Cell 55, 309-320 Griffiths, G., Hoflack, B., Simons, K., Mellman, I., and Kornfeld, S.

Henkart, P. A. (1985) Annu. Reu. Zmmunol. 3,31-58 Hopkins, C. R., Gibson, A., Shipman, M., and Miller, K. (1990)

Nature 346,335-339 Howe, C. L., Granger, B. L., Hull, M., Green, S. A., Gabel, C. A.,

Helenius, H., and Mellman, I. (1988) Proc. Natl. Acad. Sci. U. S.

Hughes, E. N., and August, T. T. (1982) J. Biol. Chem. 257 , 3970-

Jin, M., Sahagian, G., and Snider, M. (1989) J. Bwl. Chem. 2 6 4 ,

Knowles, D. M., Tolidjian, B., Marboe, C., D’Agati, V., Grimes, M.,

~ I

522-532

Biol. Chem. 263,18920-18928

8,1041-1047

(1988) Cell 52,329-341

A. 8 5 , 7577-7581

3977

7675-7680

and Chess, L. (1984) J. Immunol. 132 , 2170-2173 Kornfeld, S. (1987) FASEB J. 1,462-468 Kornfeld, S., and Mellman, I. (1989) Annu. Rev. Cell Biol., 5, 483-

Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 157 , 105-132 Lathe, R. (1985) J. Mol. Biol. 183 , l -12 Lewis, S., Green, S. A., Marsh, H., Vihko, P., Helenius, H., and

Lippincott-Schwartz, J., and Fambrough, D. M. (1986) J. Cell Biol.

Lippincott-Schwartz, J., and Fambrough, D. M. (1987) Cell 49,669- 677

Lobel, P. K., Fujimoto, K., Ye, R. D., Griffiths, G., and Kornfeld, S. (1989) Cell 5 7 , 787-796

Lopata, M., Clevelan, D., and Sollner-Webb, B. (1984) Nucleic Acids Res. 12,5707

McGregor, J. L., Catimel, B., Parmentier, S., Clezardin, P., Dechav- anne, M., and Leung, L. L. K. (1989) J. Biol. Ckm. 264,501-506

Miller, J., and Germain, R. N. (1986) J. Exp. Med. 164, 1478-1489 Noguchi, Y., Himeno, M., Sasaki, H., Tanaka, Y., Kono, A., Sakaki,

Y., and Kato, K. (1989) Biochem. Biophys. Res. Commun. 164, 1113-1120

Nolan, C. M., and Sly, W. S. (1987) Adu. Exp. Med. Biol. 225, 199- 212

Ockenhause, C. F., Tandon, N. N., Magowan, C., Jammieson, G. A., and Chulay, J. D. (1989) Science 243, 1469-1471

Oquendo, P., Hundt, E., Lawler, J., and Seed, B. (1990) Cell 58,95- 101

Pearson, W. R., and Lipman, D. J. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,2444-2448

Peters, C., Braun, M., Weber, B., Wendland, M., Schmidt, B., Pohl- mann, R., Waheed, A., and von Figura, K. (1990) EMBO J. 9,

525

Mellman, I. (1985) J. Cell Biol. 100,1839-1847

102,1593-1605

3497-3506 Pfeffer, S. R. (1988) J. Membr. Biol. 103 , 7-16 Pohlmann, R., Krentler, C., Schmidt, B., Schroeder, W., Lorkowski,

G., Culley, J., Mersmann, G., Geier, C., Waheed, A., Gottschalk, S., Grzeschik, K. H., Hasilik, A., and von Figura, K. (1988) EMBO

Robbins, A. R. (1988) in Protein Transfer and Organelle Biogenesis,

Roth, R. A. (1988) Science 2 3 9 , 1269-1271 Russo, A. F., and Koshland, D. E., Jr. (1983) Science 220,1016-1020 Sandoval, I. V., Chen, J. W., Yuan, L., and August, T. (1989) Arch.

Biochem. Biophys. 271,157-167 Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad.

Sci. U. S. A. 74,5463-5467 Skudlarek, M. D., Novak, E. K., and Swank, R. T. (1984) in Lysosomes

in Biology and Pathology (Dingle, J. T., Dean, R. T., and Sly, W. S., eds) Vol. 7, pp. 17-43, Elsevier Science Publishing Co., New York

J. 7,2343-2350

pp. 463-520, Academic Press, Orlando, FL

Sly, W. S., and Fisher, H. D. (1985) J. Cell Biochem. 18, 67-85 Tandon, N. N., Lipsky, R. H., Burgess, W. H., and Jamieson, G. A.

Tandon, N. N., Kralisz, U., and Jamieson, G. A. (198913) J. Biol.

Vandekerckhove, J., Bauw, G., Puype, M., Van Damme, J., and

Viitala, J., Carlsson, S. R., Siebert, P. D., and Fukuda, M. (1988)

von Figura, K., and Hasilik, A. (1986) Annu. Reu. Biochem. 55,167-

Williams, M. A., and Fukuda, M. (1990) J. Cell Biol. 111,955-966 Willingham, M. C., Pastan, I. H., and Sahagian, G. G. (1983) J .

(1989a) J. Biol. Chem. 264,7570-7575

Chem. 264, 7576-7583

Montagu, M. (1985) Eur. J. Biochem. 152,9-19

Proc. Natl. Acad. Sci. U. S. A. 8 5 , 3743-3747

193

Histochem. Cytochem. 31 , 1-11