Differential Expression of the Lysosome-Associated Membrane Proteins in Normal Human Tissues

Archives of Biochemistry and Biophysics Vol. 365, No. 1, May 1, pp. 75– 82, 1999 Article ID abbi.1999.1147, available online at http://www.idealibrary.com on

Differential Expression of the Lysosome-Associated Membrane Proteins in Normal Human Tissues 1 Koh Furuta,* ,2 Xiao-Ling Yang,† Juei-Suei Chen,† Stanley R. Hamilton,* ,‡ ,3 and J. Thomas August† ,‡ ,4 *Department of Pathology, †Department of Pharmacology and Molecular Sciences, and ‡Department of Oncology, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205-2196 Received November 24, 1998, and in revised form February 4, 1999

The lysosome-associated membrane proteins LAMP-1 and LAMP-2 have closely related structures, with 37% sequence homology, and are major constituents of the lysosomal membrane. Their roles are unknown, but they are thought to be structural or functional components of the lysosomal membrane. Recent reports suggest that despite their similar structure and common localization, LAMP-1 and LAMP-2 may have different functions. In our further study of these two molecules, the presence of LAMP-1 and LAMP-2 in a variety of human tissues was analyzed by immunohistochemistry, and their localization was compared to that of cathepsin D, a lysosomal hydrolase. the tissue content of LAMP-1 and LAMP-2 and their respective mRNAs were also analyzed by Northern and Western blotting. The LAMP molecules were detected by immunohistochemistry primarily in metabolically active cells, with a cytoplasmic distribution similar to that of cathepsin D and consistent with their predominant localization in lysosomes. However, there were marked differences in the intensity of staining and, in some cases, the localization of the three proteins. For example, there was much stronger staining for LAMP-2 than LAMP-1 in brain tissue and prostate ductal cells. These differences in localization were consistent with the results obtained

1

This work was supported in part by the Clayton Fund and Grant CA62924 from the National Cancer Institute, National Institutes of Health. 2 Current address: Department of Medical Technology 1, School of Health Sciences, University of Occupational and Environmental Health, Yahatanishi-ku, Kitakyushu, Japan 807-8555. 3 Current address: Division of Pathology and Laboratory Medicine, University of Texas MD Anderson Cancer Center, Box 73, Room R4.1400, 1515 Holcombe Blvd., Houston, TX 77030-4095. 4 To whom correspondence should be addressed at Department of Pharmacology and Molecular Sciences, Biophysics Building, Room 311, The Johns Hopkins University School of Medicine, 725 North Wolfe St., Baltimore, MD 21205-2185. Fax: 410-955-1894. 0003-9861/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

in Western blotting of protein extracted from the tissues. The pattern of mRNA expression was similar in all of the examined tissues, with a single mRNA identified for LAMP-1 and two splice variant forms seen for LAMP-2. Our studies of these molecules in human tissues support the conclusion that the expression of the molecules is independently controlled in some tissues, suggesting that the molecules may have independent as well as similar functions. © 1999 Academic Press Key Words: LAMP-1; LAMP-2; cathepsin D; lysosomal membrane proteins.

Two lysosome-associated membrane proteins, LAMP-15 and LAMP-2, are type 1 glycoproteins that are localized primarily on the periphery of the lysosome and are recognized as major constituents of the lysosomal membrane (1–7). While the majority of the molecules are associated with lysosomes, a small fraction (,5%) of LAMP-1 and LAMP-2 molecules are also expressed on the cell surface of a variety of cultured cells (1, 4, 7). Both glycoproteins are composed of a polypeptide core of ;40 kDa that consists of (a) an intraluminal sequence with two similar domains separated by a serine/proline-rich hinge region, each domain having four cysteines that form two disulfide bridges, (b) a transmembrane domain, and (c) a nine-amino acid cytoplasmic sequence with a carboxyl terminal Y-X-X-hydrophobic residue lysosomal membrane targeting sequence (8, 9). These molecules are among the most heavily glycosylated cellular proteins, with approximately 50% of their mass being carbohydrate (3, 10, 11). Thus, it is not surprising that they are the major cellular proteins associated with the changes in glycosylation that accompany malignant transformation (11). 5

Abbreviations used: LAMP, lysosome-associated membrane protein; PBS, phosphate-buffered saline; ABC, avidin– biotin peroxidase complex; RACE, rapid amplification of cDNA ends; DIG, digitonin. 75

76

FURUTA ET AL.

Although they are encoded by separate genes, LAMP-1 localized to chromosome 13q34 and LAMP-2 to Xq24-25 (12), the two proteins are highly similar in primary structure, with approximately 37% sequence homology (11, 13–17). LAMP-2 has been found to occur as splice-variant molecules, which are encoded by at least two or three transcripts that produce variant transmembrane and cytoplasmic domains in chicken (18), mouse (19), and human cells (20). Other proteins with sequence similarity to LAMP-1 and LAMP-2 are CD68 (21) and DC-LAMP (22. CD68 and DC-LAMP, which mapped to chromosomes 17p13, and 3q26-27, respectively (22), contain only domain two of the intraluminal sequence, with a mucin-like chain substituted for the N-terminal domain 1. One of the ongoing questions concerning the role of these proteins is whether they have a common function as structural elements of the lysosomal membrane or whether they have distinct functional roles. While the LAMP proteins have assumed importance as markers of lysosomes in a wide variety of biological studies, their biological functions remain unclear. One speculation is that LAMP-1 is a “house-keeping” protein (23, 24) and that LAMP-2 has some additional or distinct role. Clearly, the molecules are integral components of the lysosomal membrane, rendered resistant to proteolytic degradation by their heavy glycosylation and disulfide structure (8). However, other functional roles for the molecules cannot be excluded, and LAMP-2 has been described as a receptor for the uptake and degradation of cytoplasmic proteins (25). The possibility of distinct functional roles for the two molecules, despite their close sequence similarity, is strengthened by the fact that they are encoded on different chromosomes (12). Recent reports of specific expression of mRNA and protein variants of LAMP-2 have raised additional questions about possible differences in the localization and function of the LAMP proteins (18, 26). Examination of chicken LAMP-2 mRNA in different tissues has shown tissue-specific expression of LAMP-2 variants, with almost exclusive expression of one of the forms in the brain (18). In addition, different levels of expression of the LAMP-2 variants at the cell surface have been attributed to differences in the COOH-terminal amino acid residues (26). An alternatively spliced mRNA encoding human LAMP-2 has also been found to be uniquely overexpressed in human muscle, again suggesting the possibility of multiple functions (20). Results of several other studies have also suggested specific cellular localization of LAMP-2: the cell-surface expression of LAMP-2 and CD63 (LAMP-3) on platelets as detected in studies of in vivo platelet activation (27), the expression of LAMP-2 in human genital organs (28), and the identification of LAMP-2 as the mouse macrophage differentiation antigen Mac-3 (2).

Although previous immunohistochemical and electron microscope studies (11) uniformly support the conclusion that the LAMP molecules have a predominant steady-state localization in the lysosome, it has now become apparent that some human tissue cells may differ in their expression of LAMP proteins and lysosomal hydrolases. For example, the azurophilic granules of neutrophilic leukocytes and their precursors in the bone marrow, which are defined as primary lysosomes i.e., membrane-bound organelles containing acid hydrolases), show only low expression of the LAMP molecules (29). Moreover, it has recently been reported that DC-LAMP is expressed only in lymphoid organs and dendritic cells, and not in some other epithelial or mesenchymal tissues that would be expected to be rich in lysosomes (22). In a further comparison of LAMP-1 and LAMP-2, we have now conducted an immunohistochemical and molecular analysis of the proteins that specifically addresses the relative degree of expression of LAMP-1 and LAMP-2 in various human organs. The question we addressed is whether expression of the two proteins is coordinate, with comparable levels of the two proteins in the cells of particular tissues, or whether the two proteins vary in their distribution or expression. We have also compared the expression of the LAMP molecules to that of cathepsin D, an acidic lysosomal protease that is conventionally used as an indicator of cellular lyososomes (30, 31). The results indicate that while LAMP-1 and LAMP-2 are localized primarily in lysosomes, their concentration and that of cathepsin D in different tissues is markedly variable, suggesting that, in at least some tissues, the expression of the LAMP proteins may be independently controlled. MATERIALS AND METHODS Tissues and cell lines. Nine normal human tissues, cerebral cortex, colonic mucosa, kidney cortex, liver, lung, pancreas, prostate, spleen, and uterine myometrium, were obtained from surgical specimens at The Johns Hopkins Hospital following operation for organ disease. The aforementioned nine “normal” tissues were retrieved from portions of each resected specimen that were judged normal by gross observation. Additionally, each retrieved tissue was confirmed normal under microscopic observation. The fresh frozen tissues were divided into three aliquots for immunohistochemistry, Western gels, and Northern gels. The frozen tissues for immunohistochemistry were further processed to formaldehyde-fixed and paraffin-embedded specimens. The human primary colorectal carcinoma tissue used as a control for immunohistochemistry was provided by the frozen tissue bank in the Department of Pathology of The Johns Hopkins University School of Medicine. Positive and negative control samples for Western and Northern gels were 293 cells (human embryonal kidney) and NIH3T3 cells (mouse NIH Swiss embryonal fibroblast). Monoclonal antibodies. The anti-human LAMP-1 mouse monoclonal antibody H4A3 (7) was used at a 1:150 dilution for immunohistochemistry and at 1:500 dilution for Western blots, anti-human LAMP-2 mouse monoclonal antibody H4B4 (7) at 1:100 dilution for immunohistochemistry and at 1:500 for Western blots, and anticathepsin D mouse monoclonal antibody (Triton Diagnostics, Alam-

TISSUE EXPRESSION OF THE LYSOSOME-ASSOCIATED MEMBRANE PROTEINS eda, CA) at 0.25 mg/ml for immunohistochemistry. Anti-human CD44 mouse monoclonal antibody U9M2 (a gift from Dr. James Hildreth, Department of Pathology and Molecular Sciences, The Johns Hopkins University School of Medicine) (32) at 5 mg/ml and anti-human p53 mouse monoclonal antibody DO7 (Signet, Dedham, MA) at 1:50 dilution were used as nonrelevant control antibodies for immunohistochemistry. Immunohistochemistry. Serial sections of formaldehyde-fixed and paraffin-embedded tissues (5 mm) were treated with xylene for 30 min to remove paraffin, rehydrated for 5 min each with 100, 95, and 70% ethanol, treated for 10 min with 3% hydrogen peroxide in methanol to eliminate endogenous peroxidase activity, and treated for 15 min in a microwave oven (33) with 0.05 M glycine–HCl, pH 3.5 (34), for antigen retrieval. The sections were then incubated with 5% normal horse serum with 0.01% Triton-X in phosphate-buffered saline (PBS) at pH 7.4 for 20 min to eliminate nonspecific background immunostaining. Sections were then incubated with the H4A3, H4B4, U9M2, or DO7 antibody diluted in PBS containing 5% normal horse serum and 0.01% Triton-X for 24 h at room temperature. The sections were next treated for 1 h at room temperature with affinitypurified biotinylated horse anti-mouse IgG (BA-2000; Vector Lab., Burlingame, CA) diluted 1:200, followed by incubation in avidin– biotin peroxidase complex (35) (Vectastatin Elite ABC reagent, Vectastatin ABC kit standard, PK6100, Vector Lab.) for 1 h at room temperature. ABC was visualized by incubating with an immunopure metal-enhanced diaminobenzidine substrate kit (Pierce, Rockford, IL). A brown reaction product appeared after 1 to 3 min, at which time the reaction was terminated by transferring the sections to water. After counterstaining with methyl green (Sigma, St. Louis, MO), sections were dehydrated, and a coverslip was attached with Permount (Fisher Scientific, Pittsburgh, PA). Positive and negative controls included 5-mm sections from formaldehyde-fixed and paraffin-embedded primary colorectal carcinoma and substitution of the primary antibody with other antibodies of different specificity but of the same origin and isotype (anti-CD44 mouse monoclonal antibody U9M2 and an anti-p53 mouse monoclonal antibody DO7). Western gel analysis. Approximately 50 mg of frozen tissue was homogenized in 1.0 ml of a buffer solution containing 50 mM Tris– HCl, pH 8.0; 150 mM sodium chloride; 0.02% sodium azide; 0.5 mM phenylmethylsulfonyl fluoride (Sigma); 10 mg aprotinin; 20 mg/ml leupeptin; 10 mg/ml pepstatin A (Boehringer-Mannheim, Indianapolis, IN); 20 mg/ml trypsin inhibitor (Sigma); 10 mM ethylene glycol bis(b-aminoethyl ether) N,N,N9,N9-tetraacetic acid (Sigma); 10 mM EDTA (Sigma); 10 mg/ml trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane (Sigma); and 0.1% Triton X-100 (Sigma). Protein lysates were boiled for 5 min in sample buffer containing 2-mercaptoethanol (Bio-Rad, Hercules, CA). The proteins, 20 mg each, were electrophoretically resolved by 9% sodium dodecyl sulfate–polyacrylamide gel electrophoresis at 35 mA at room temperature and transferred into a polyvinylidene difluoride membrane (Immobilon PVDF, Millipore, Bedford, MA). After blocking with 5% dry milk overnight at 4°C, membranes were incubated for 1 h with either 1:500 diluted anti-LAMP-1 (H4A3) or 1:500 diluted anti-LAMP-2 (H4B4), washed three times with wash solution (PBS containing 1% dry milk and 0.1% Tween 20), and incubated 1 h with horseradish peroxidaseconjugated sheep anti-mouse Ig (Amersham, Arlington Heights, IL), followed by five washes with wash solution. The proteins were visualized by an enhanced chemiluminescence Western blotting detection system (Amersham) and recorded on X-ray film (KodakXAR, KODAK, Rochester, NY). Synthesis of human LAMP 2A and LAMP 2B cDNA probes. Human colon cancer RNA was used in the rapid amplification of cDNA ends (RACE) PCR procedure with a 39-Amplifinder RACE Kit (Clontech, Palo Alto, CA) following the manufacture’s protocol. The total RNA was reverse-transcribed with use of the NN 21 oligo(dT)(59-cct ctg aag gtt cca gaa tcg ata gga att ctt ttt ttt ttt ttt ttt tgc agc at-39)(N 5 A, T, C, or G; N 21 5 A, G, or C). The reaction mixture (20

77

ml) contained 1 mg of RNA, 0.5 oligo(dT), 0.5 mM dNTP mixture, 62.5 mM Tris–HCl, pH 8.3, 11.25 mM dithiothreitol, 11.25 mM MgCl 2, and 12.5 units of reverse transcriptase (from avian myeloblastosis virus). The two upstream, human LAMP-2-specific, nested oligonucleotides used for RACE PCR were cat ggt gtg ctt ccg cct ctt and gtg ggg atg atc aga atg gtc. A downstream anchor-primer oligonucleotide was 59-ctg gtt cgg ccc acc tct gaa ggt tcc aga atc gat ag-39. Amplification steps were performed with AmpliTaq DNA polymerase in a volume of 50 ml containing 103 polymerase buffer (10 mM Tris–HCl, pH 8.3, 2.0 mM MgCl 2, 50 mM KCl), 0.2 mM dNTP mixture, 2.5 units of AmpliTaq DNA polymerase, and 0.2 mM each appropriate primer. The first amplification was performed using the primer with sequence cat ggt gtg ctt ccg cct ctt and the anchor primer at 95°C for 1 min, 60°C for 1 min, and 72°C for 2 min over 30 cycles in a thermal cycler (Ericomp, San Diego, CA). The second amplification was performed using the primer with sequence gtg ggg atg act aga atg gtc and anchor primer under identical conditions for 25 cycles. The four 1060- 1250- 1350- and 1450-bp bands obtained from RACE were excised from a low-melting agarose gel, and each fragment was ligated into the TA cloning vector (Invitrogen, San Diego, CA). Colonies with inserts were selected by plating Escherichia coli DH5alfa cells (Gibco BRL, Gaithersburg, MD) which had been transformed with the ligation mixture on a Luria broth plate containing ampicillin and 5-bromo-4-chloro-3-indoyl b-D-galactopyranoside. Between 30 and 75% of the white colonies contained the desired insert. Plasmid DNA was isolated by use of a QIAGEN Plasmid Mini Kit (Qiagen, Chatsworth, CA). Restriction analysis of the fragment with KpnI and ApaI was used to identify the inserts in the desired orientation. The four complete sequences of the insert were determined by manual sequencing. The insert from the 1060-bp band was found to be identical to the reported hLAMP-2B sequence (20), and the 1250-, 1350-, and 1450-bp bands were identical to the reported hLAMP-2A (36). Northern analysis. Total cellular RNA from each human tissue, including colon cancer tissue, was isolated using the standard guanidium isothyiocyanate method (37) and quantitated by UV absorption. RNA (10 mg/lane) was electrophoresed through a 1.5% gel containing 0.4 M formaldehyde and transferred to positively charged nylon membranes (Boehringer-Mannheim) by capillary action. The RNA blots were probed with DNA fragments prepared by PCR amplification using a PCR DIG Probe Synthesis Kit (Boehringer-Mannheim). The mRNAs of hLAMP-1 (forward primer, ttc tca aca tca acc cca aca; reverse primer, cac agt cgg caa ttc cta caa) were analyzed using the 249-bp probe, and mRNA of hLAMP-2A (standard form) (forward primer, cac aag gaa agt att cta cag; reverse primer, cac cat cat gga tat gag) and hLAMP-2B (variant form) (forward primer, cac aag gaa agt att cta cag; reverse, gga tat cag act ctg taa cac) were analyzed by using a cocktail composed of 166-bp probe for hLAMP-2A and a 154-bp probe for hLAMP-2B. Integrity, equal loading, and uniform transfer of each RNA sample were assessed by ethidium bromide staining and visualization of the 28S and 18S ribosomal bands. In no case did the intensity of the ribosomal bands between tissue specimens differ by more than 50%.

RESULTS

LAMP-1, LAMP-2, and Cathepsin D in Normal Human Tissues Immunohistochemical analysis of the expression of LAMP-1, LAMP-2, and cathepsin D in nine normal human tissues was performed to compare their levels of expression and localization. The same dilutions of antibody, chosen experimentally for optimum staining, were used with all tissues. Clear immunoreactivity was seen in each of the tissues with the three mono-

78

FURUTA ET AL.

clonal antibodies, primarily in metabolically active cells, but in other cells as well. The cellular staining pattern was chiefly granular and cytoplasmic, in keeping with the lysosomal localization of the proteins. No cell-surface staining was observed. In many cases the relative degree of staining of the three proteins varied greatly among the various tissues (Fig. 1). The antiCD44 and anti-p53 control antibodies showed appropriate patterns of reactivity, which differed from those of the LAMP and cathepsin D patterns. Brain. The neuron cell bodies, microglia, and endothelial cells were stained extensively by the LAMP-2 antibody and to a lesser degree by the cathepsin D antibody. In contrast, staining of LAMP-1, while similar to staining of LAMP-2 in endothelial cells, was less intense in neurons and microglia (Figs. 1A–1C). Colon. The localization and intensity of staining of tissue macrophages in the upper lamina propria was similar for the three antibodies, but the epithelial cells showed greater staining for the LAMP proteins than for cathepsin D (Figs. 1D–1F). Kidney. Staining of the glomerulus was greater for LAMP-1 than for LAMP-2 and both were more intensely stained than cathepsin D. In the proximal tubule, cathepsin D showed less intense staining than did LAMP-1 and LAMP-2; there was no significant difference in staining between LAMP-1 and LAMP-2. In the distal tubule, LAMP-2 showed less intense staining than did LAMP-1 and cathepsin D; there was no significant difference between LAMP-1 and cathepsin D (Figs 1G–1I). Prostate. The basic localization of the antigens was similar, but the intensity of staining of the ductal cells was much greater with LAMP-2 than that with the other two proteins (Figs. 1J–1L). Liver. There was little difference in the staining of hepatocytes and Kupffer cells with the three antibodies, except for an apparently slight dominance of LAMP-2 staining in the hepatocytes (Figs. 1M–1O). Lung. The alveolar macrophages were stained extensively by each of the three antibodies. Alveolar epithelial staining was most intense with anti-LAMP-1 (Figs. 1P–1R). Pancreas. There was remarkable staining of islet cells with both LAMP-1 and LAMP-2 antibodies compared to only slight staining for cathepsin D. This finding suggests that the lysosomes of islet cells are functionally different from those of other cells, as evidenced by the paucity of cathepsin D. In comparison, acinar and ductal cells were similarly stained with each of the three antibodies (Figs. 1S–1U). Uterus. The staining intensity of the myometrium was greater for LAMP-1 than LAMP-2 and least for cathepsin D (Figs. 1V–1X).

Spleen. Lymphocytes and endothelial cells were similarly stained by each of the three antibodies (not illustrated). Western Blot Analysis of Normal Human Tissues An alternate explanation for the differences in staining intensity by immunohistochemistry is that the different patterns resulted from some consistent variation in the staining procedure and did not reflect the concentrations of the proteins. For this reason, despite the reproducibility of the immunohistochemistry results, Western gel analysis of LAMP-1 and LAMP-2 in the human tissue samples was performed to verify that the staining intensity of the various proteins by immunohistochemistry corresponded to the level of expression of the proteins (Fig. 2). In all tissues except the brain, there were strong protein bands of 110 to 140 kDa, representing the mature form of the protein. A partially glycosylated precursor form representing various highmannose oligosaccharides of about 92 kDa and the nonglycosylated core protein of about 45 kDa (2, 3, 7, 38, 39) were also present. The positive control cell line, human embryonal kidney 293, showed only the mature forms of LAMP-1 and LAMP-2; as expected, the murine NIH 3T3 cells were negative for both molecules. In every case the degree of expression of the two LAMP proteins directly corresponded to the intensity of tissue immunoreactivity: The dominance of LAMP-2 protein in brain and prostate tissues, the low level of both proteins in uterine tissue, and the high levels of protein in the Western blots of pancreas and kidney were analogous to the strong staining patterns in these tissues. Brain tissue differed from the other eight tissues in containing very little, if any, LAMP-1 and in showing previously uncharacterized molecular weight variants of the mature LAMP-2, which are possibly indicative of different glycosylation patterns or variant proteins. The greater degree of LAMP-2 immunoreactivity than of LAMP-1 in the brain by Western blotting corresponded to the greater tissue staining with the LAMP-2 antibody. Northern Blot Analysis Other studies have shown the presence of splice variant forms of LAMP-2 (2A and 2B) in a variety of human tissues (20). We therefore analyzed the tissue specimens in this study by Northern hybridization with DIG-labeled cDNA probes. The hLAMP-1 cDNA corresponded to residues Ile to Lys of the luminal domain of the protein; the cDNA of hLAMP-2A corresponded to residues Thr to Glu of the carboxyl-terminus of the protein; and the cDNA of hLAMP-2B corresponded to

TISSUE EXPRESSION OF THE LYSOSOME-ASSOCIATED MEMBRANE PROTEINS

79

FIG. 1. Expression of LAMPs and cathepsin D in various human tissues. Immunohistochemical analysis of LAMP-1, LAMP-2,l and cathepsin D expression in human tissues was performed as described under Materials and Methods to compare the levels and localization of these different proteins. Staining of LAMP-1 is shown in the left column, LAMP-2 in the middle column, and cathepsin D in the right column. The tissues are as labeled. Expression of LAMP-1, LAMP-2, and cathepsin D generally occurred at the same tissue sites, but there were major differences in the apparent level of protein expression. Expression of LAMP-2 was greater than that of LAMP-1 in neurons in the cerebral cortex (B) and in prostate duct epithelium (K). Reduced staining of cathepsin D compared to the LAMP antibodies occurred in epithelial cells in the colonic mucosa (F), glomerulus cells in the renal cortex (I), prostate duct epithelium (L), islet cells in the pancreas (U), and myometrium in the uterus (X). Staining of cathepsin D in neurons in the cerebral cortex (C) was lower than that of LAMP-2 (B) but exceeded that of LAMP-1 (A). Brain: neuron (4) and endothelial cells (2). Colon: tissue macrophage in upper lamina propria (4) and epithelial cells (*). Kidney: glomerulus (*), proximal tubules (4), and distal tubules (2. Liver: hepatocyes (*) and Kupffer cells (4). Pancreas: islet cells (n) and acinar cells (*).

80

FURUTA ET AL.

DISCUSSION

FIG. 2. Western gel analysis of LAMP proteins from normal human tissues. A, hLAMP-1; B, hLAMP-2. Western blotting was carried out using human LAMP-1-specific (H4A3) or LAMP-2specific (H4B4) monoclonal antibodies and horseradish peroxidase-conjugated second antibody as described under Materials and Methods. Strongly immunoreactive protein bands of 110 to 140 kDa represent the mature form of the protein. The protein of about 92 kDa corresponds to a partially glycosylated precursor form representing various forms of high-mannose oligosaccharides, and the form of about 45 kDa represents the nonglycosylated core protein.

residues Thr to the noncoding region of the carboxylterminus of the protein. As expected, only one form of hLAMP-1 mRNA, a 2.455-kb transcript, was observed (Fig. 3). The fainter staining lower molecular weight bands were not present in 293 or NIH3T3 cells and are considered to be ribosomal RNA contaminants. hLAMP-2 mRNAs were present in two forms, one at about 4.0 kb (hLAMP-2B, the variant form) and another at 1.9 kb (hLAMP-2A, the standard form). Because of possible degradation of RNA in the various tissues, the differences in the levels of intensity among the tissues were not considered to be meaningful. Nevertheless, we saw no apparent qualitative differences in LAMP-1 or LAMP-2 mRNA expression that could be correlated with the observed different patterns of immunohistochemical staining. The murine NIH3T3 LAMP-1 and LAMP-2 mRNAs were also detected by the human LAMP cDNA sequences; we believe this occurred because of sequence similarity between mouse and human LAMP-1 (75–95% homology; Accession Nos. JO3881 and JO4182), LAMP-2A (74 – 87%; Accession Nos. JO5287 and X77196), and mouse LAMP-2A and human LAMP-2B (74 – 82%; Accession Nos. JO5287 and S79873) (40).

In this study we have found that human LAMP-1 and LAMP-2 and lysosomal cathepsin D are each predominantly localized to metabolically active cells of epithelial and mesenchymal origin, in accord with previous studies of the LAMP proteins (3, 20) and of the tissue expression of cathepsin D (30). The remarkable new finding in our study is that while the localization of LAMP-1 and LAMP-2 was generally similar to that of cathepsin D, the levels of expression of the three proteins varied independently in the specific cells of particular tissues, for example, the strong staining of LAMP-2 despite a virtual absence of LAMP-1 in brain neurons and the stronger staining of LAMP-2 in prostate ductal cells. There also were differences in the staining of the LAMP proteins compared to that of cathepsin D, for example, the strong staining of the LAMP proteins relative to cathepsin in the glomerulus of the kidney and the islet cells of the pancreas. These observations, taken together with the reported absence of the LAMP proteins in the primary lysosomes of human neutrophilic leukocytes (29), suggest structural or functional differences in the role of the LAMP proteins in some lysosomes. These findings are also in accord with a previous report (41) that LAMP-2 is barely

FIG. 3. Northern analysis of LAMP mRNAs from normal human tissue. A, hLAMP-1; B, hLAMP-2A (1868 bp) and hLAMP-2B (4006 bp). Tissues were analyzed for LAMP-specific mRNA by Northern hybridization with DIG-labeled cDNA probes. Only one form of LAMP-1 mRNA, a 2.455-kb transcript, was observed. hLAMP-2 mRNAs were present in two forms, one at about 4.0 kb (hLAMP-2B, standard form) and another at 1.9 kb (hLAMP-2A, variant form). The cDNA probes, used to identify the various mRNA transcripts are described under Materials and Methods.

TISSUE EXPRESSION OF THE LYSOSOME-ASSOCIATED MEMBRANE PROTEINS

detectable in murine embryonal carcinoma cell lines compared to a high level of LAMP-1 expression, but increases to a level comparable to that of LAMP-1 during retinoic acid-induced differentiation of the cells. The case for independent expression of the two LAMP proteins was also recently strengthened by the discovery of a novel lysosome-associated membrane glycoprotein, DC-LAMP, that is present only in lymphoid organs and is induced upon dendritic cell maturation and transiently expressed in a compartment that also contains MHC class II proteins (22). This accumulating evidence shows that while all of different forms of the lysosomal membrane glycoproteins [LAMP-1 (CD107), LAMP-2 (CD107), LAMP-splice variants, CD68, and DC-LAMP] share a common lysosomal membrane targeting signal and steady-state localization in lysosomes, their level of expression and localization differs markedly in may organs. With the genes for these proteins located on different chromosomes, it appears likely that the proteins diverged early in evolution and have distinct functions. The 17-kb gene encoding avian LAMP-1 contains nine exons and eight introns (24). Northern blot analysis of LAMP-1 RNA from several avian tissues has shown the same mRNA species in all tissues, indicating a lack of alternative splicing; in contrast, LAMP-2 in avian tissues is present in at least three splice variant forms that encode three different transmembrane and cytoplasmic domains, with one of the forms predominating in the brain (18). Human tissues have been found to contain at least two different forms of LAMP-2, with marked differences in the tissue expression of the two forms: The 4006-bp hLAMP-2B mRNA predominates in brain and muscle, and the 1868-bp hLAMP-2A mRNA is the predominant form in placenta, lung, and liver (20). Our current study showed a similar result: high levels of hLAMP-2B mRNA in brain and uterine myometrium and of hLAMP-2A mRNA in kidney, liver, lung, pancreas, and prostate gland. Some of the differences in the expression of LAMP-2 in the various tissues may therefore be attributed to the expression of the splice variant forms of the molecule. The functional and structural attributes of the different forms represent an area for future investigation. ACKNOWLEDGMENTS We thank Dr. Minoru Fukuda for providing cDNAs of hLAMP-1 and hLAMP-2 and Jeffrey J. Floyd and Rahj Robinson for technical assistance. We also thank Drs. Masaru Himeno and Naotaka Hamasaki for assistance.

81

REFERENCES 1. Hughes, E. N., and August, J. T. (1981) J. Biol. Chem. 256, 664 – 671. 2. Chen, J. W., Murphy, T. L., Willingham, M. C., Pastan, I., and August, J. T. (1985) J. Cell. Biol. 101, 85–95. 3. Chen, J. W., Chen, G. L., D’Souza, M. P., Murphy, T. L., and August, J. T. (1986) Biochem. Soc. Symp. 51, 97–112. 4. Lippincott-Schwartz, J., and Fambrough, D. M. (1987) Cell 49, 669 – 677. 5. Fukuda, M., Viitala, J., Matteson, J., and Carlsson, S. R. (1988) J. Biol. Chem. 263, 18920 –18928. 6. Carlsson, S. R., Roth, J., Piller, F., and Fukuda, M. (1988) J. Biol. Chem. 263, 18911–18919. 7. Mane, S. M., Marzella, L., Bainton, D. F., Holt, V. K., Cha, Y., Hildreth, J. E., and August, J. T. (1989) Arch. Biochem. Biophys. 268, 360 –378. 8. Arterburn, L. M., Earles, B. J., and August, J. T. (1990) J. Biol. Chem. 265, 7419 –7423. 9. Guarnieri, F. G., Arterburn, L. M., Penno, M. B., Cha, Y., and August, J. T. (1993) J. Biol. Chem. 268, 1941–1946. 10. Hughes, E. N., and August, J. T. (1982) Proc. Natl. Acad. Sci. USA 79, 2305–2309. 11. Fukuda, M. (1991) J. Biol. Chem. 266, 21327–21330. 12. Mattei, M. G., Matterson, J., Chen, J. W., Williams, M. A., and Fukuda, M. (1990) J. Biol. Chem. 265, 7548 –7551. 13. Fambrough, D. M., Takeyasu, K., Lippincott-Schwarz, J., and Siegel, N. R. (1988) J. Cell. Biol. 106, 61– 67. 14. Chen, J. W., Cha, Y., Yuksel, K. U., Gracy, R. W., and August, J. T. (1988) J. Biol. Chem. 263, 8754 – 8758. 15. Viitala, J., Carlsson, S. R., Siebert, P. D., and Fukuda, M. (1988) Proc. Natl. Acad. Sci. USA 85, 3743–3747. 16. Cha, Y., Holland, S. M., and August, J. T. (1990) J. Biol. Chem. 265, 5008 –5013. 17. Howe, C. L., Granger, B. L., Hull, M., Green, S. A., Gabel, C. A., Helenius, A., and Mellman, I. (1988) Proc. Natl. Acad. Sci. USA 85, 7577–7581. 18. Hatem, C. L., Gough, N. R., and Fambrough, D. M. (1995) J. Cell Sci. 108, 2093–2100. 19. Gough, N. R., Hatem, C. L., and Fambrough, D. M. (1995) DNA Cell. Biol. 14, 863– 867. 20. Konecki, D. S., Foetisch, K., Zimmer, K. P., Schlotter, M., and Konecki, U. L. (1995) Biochem. Biophys. Res. Commun. 215, 757–767. 21. Holness, C. L., and Simmons, D. L. (1993) Blood 81, 1607– 1613. 22. deSaint-Vis, B., Vincent, J., Vandenabeele, S., Vanbervliet, B., Pin, J.-J., Aı¨t-Yahia, S., Patel, S., Mattei, M.-G., Banchereau, J., Zurawski, S., Davoust, J., Caux, C., and Lebecque, S. (1998) Immunity 9, 325–336. 23. Himeno, M., Noguchi, Y., Sasaki, H., Tanaka, Y., Furuno, K., Kono, A., Sakaki, Y., and Kato, K. (1989) FEBS Lett. 244, 351– 356. 24. Zot, A. S., and Fambrough, D. M. (1990) J. Biol. Chem. 265, 20988 –20995. 25. Cuervo, A. M., and Dice, J. F. (1996) Science 273, 501–503. 26. Gough, N. R., and Fambrough, D. M. (1997) J. Cell Biol. 137, 1161–1169. 27. Kannan, K., Divers, S. G., Lurie, A. A., Chervenak, R., Fukuda, M., and Holcombe, R. F. (1995) Eur. J. Haematol. 55, 145–151.

82

FURUTA ET AL.

28. Aumueller, G., Renneberg, H., and Hasilik, A. (1997) Cell. Tissue Res. 287, 335–342. 29. Cieutat, A. M., Lobel, P., August, J. T., Kjeldsen, L., Sengelov, H., Borregaard, N., and Bainton, D. F. (1998) Blood 91, 1044 – 1058. 30. Reid, W. A., Valler, M. J., and Kay, J. (1986) J. Clin. Pathol. 39, 1323–1330. 31. Rochefort, H., Capony, F., and Garcia, M. (1990) Cancer Metastasis Rev. 9, 321–331. 32. Guo, M. M. L., and Hildreth, J. E. K. (1993) J. Immunol. 151, 2225–2236. 33. Shi, S. R., Key, M. E., and Kalra, K. L. (1991) J. Histochem. Cytochem. 39, 741–748. 34. Taylor, C. R., Shi, S. R., Chaiwun, B., Young, L., Imam, S. A., and Cote, R. J. (1994) Hum. Pathol. 25, 263–270.

35. Hsu, S.-M., Raine, L., and Fanger, H. (1981) J. Histochem. Cytochem. 29, 577–580. 36. Konecki, D. S., Foetisch, K., Schlotter, M., and Lichter-Konecki, U. (1994) Biochem. Biophys. Res. Commun. 205, 1–5. 37. Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 162, 156 –159. 38. Chen, J. W., Pan, W., D’Souza, M. P., and August, J. T. (1985) Arch. Biochem. Biophys. 239, 574 –586. 39. D’Souza, M. P., and August, J. T. (1986) Arch. Biochem. Biophys. 249, 522–532. 40. Altschul, S. F., Madden, T. L., Scha¨ffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389 –3402. 41. Amos, B., and Lotan, R. (1990) J. Biol. Chem. 265, 19192– 19198.