Keratan Sulfate Modification of CD44 Modulates Adhesion to Hyaluronate*

, CD44 alternative splicing has been implicated in the regulation of CD44 function. CD44 undergoes significant posttranslational modification in all cells, but the functional consequences of these modifications are poorly understood. In the current study, we have dem- onstrated that keratan sulfate modification of CD44 significantly modulates its ability to bind to hyaluronate. We observed naturally occurring differences in CD44 keratan sulfate substitution between two clonal vari- ants of the KM12 human colon carcinoma cell line. CD44 on the highly metastatic KM12L4 clone is more heavily substituted with keratan sulfate than CD44 on the poorly metastatic KM12C6 clone. Moreover, CD44H on KM12L4 bound to hyaluronate poorly compared to CD44H on KM12C6. Removal of keratan sulfate from CD44 greatly enhanced CD44-mediated cell adhesion to hyaluronate. Removal of keratan sulfate from CD44H- immunoglobulin fusion proteins also enhanced their adhesion to hyaluronate. The influence of glycosaminogly- can substitution on CD44 function was specific to keratan sulfate substitution; treatment to remove chon- droitin sulfate, heparan sulfate, or hyaluronate did not affect CD44-mediated cell adhesion to hyaluronate. Use of site-directed CD44 H cDNA mutants with arginine changed to alanine at position 41 indicated that keratan sulfate modification of CD44 modulates hyaluronate adhesion through its B loop domain. These findings sug- gest that keratan sulfate modification of CD44 may play an important regulatory role in

CD44 is the principal cell surface receptor for hyaluronate (1)(2)(3) and has been implicated in a wide variety of processes, including cell motility (4,5), growth control (6), tumor metastasis (5,(7)(8)(9), and lymphocyte activation (10 -12). Much interest has been devoted to the extensive alternative splicing of CD44 mRNA. Several CD44 isoforms arise from mRNA alternative splicing of at least 10 exons encoding a portion of the extracellular domain (13,14). The predominant CD44 isoform detected in many normal tissues is CD44H, an isoform encoded by a transcript that does not contain any of the central alternatively spliced exons (15). Inclusion of one or more of the alternatively spliced exons generates individual CD44 isoforms.
While differences in CD44 alternative splicing between cells may result in different cell behavior, cell type-specific posttranslational modification of CD44 may also alter their phenotype. CD44 undergoes extensive post-translational modification, including N-and O-linked glycosylation and substitution with high molecular weight glycosaminoglycans (16 -20). We have recently demonstrated that the same CD44H isoform expressed on two clonal variants of a human colon carcinoma cell line display very different functional characteristics (21). CD44H reintroduced by stable transfection back into the poorly metastatic KM12C6 colon carcinoma clone binds hyaluronate and mediates a reduction in both in vitro and in vivo growth. In contrast, CD44H transfected into the highly metastatic KM12L4 colon carcinoma clone does not bind hyaluronate and does not mediate reduction in either in vitro or in vivo growth. These results indicate that subtle differences exist between the KM12C6 and KM12L4 cells that alter the ability of CD44H expressed on their cell surface to bind hyaluronate and modulate cell growth.
In the report presented herein, we have examined how CD44H glycosaminoglycan substitution influences CD44H function. We report that CD44H is more heavily substituted with keratan sulfate when expressed on KM12L4 cells than on KM12C6 cells. Moreover, this difference in keratan sulfate substitution significantly modulates CD44H function. Removal of keratan sulfate from cell surface CD44 or from CD44Himmunoglobulin fusion proteins (CD44H receptorglobulins) greatly enhances their adhesion to hyaluronate. Use of sitedirected CD44H mutants that are unable to bind hyaluronate because of an amino acid substitution in the B loop domain indicates that keratan sulfate substitution modulates hyaluronate binding through this domain. The dramatic impact of this regulatory mechanism on CD44 function indicates that it is an additional mechanism, which, together with alternative splicing, regulates the function of CD44.

MATERIALS AND METHODS
Cell Lines-The human colon carcinoma cell lines KM12L4 and KM12C6 were generous gifts from Dr. Isaiah Fidler (M. D. Anderson Cancer Center, Houston, TX) and have been described previously (22). Human colon carcinoma cell line HT29 was obtained from the American Type Culture Collection (Rockville, MD). The SW620 human colon carcinoma cell line was a generous gift from Dr. Lee Ellis (M. D. Anderson Cancer Center, Houston, TX). Cells were grown in Dulbecco's modified Eagle's medium/Ham's F-12 supplemented with 8% fetal calf serum.
These cell lines were also transfected with three constructs, and the resulting transfectants have been characterized previously (21). Briefly, cells designated with the suffix ⌬H were transfected with CD44H fulllength cDNA in the pRC/CMV vector (Invitrogen, San Diego, CA), and * This work was supported in part by National Institutes of Health Grants CA64454 (to K. K. T.), CA55735 (to I. S.), and DK433351 (for core facilities). 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  these cells express CD44H in addition to high molecular weight CD44 isoforms. A mutant form of CD44H with arginine 41 changed to alanine, thereby destroying its affinity for hyaluronate, is expressed on the cell surface of transfectants designated with the suffix ⌬41R/A. Control transfectants were transfected with the vector only (no insert) and are designated with the suffix ⌬neo. Transfectants were grown in Dulbecco's modified Eagle's medium/Ham's F-12 with glutamine (Life Technologies, Inc.) and G418 (Sigma) added to a final concentration of 500 g/ml for KM12 and HT29 transfectants and 1500 g/ml for SW620 transfectants.
Antibodies and Receptorglobulins-The mAb 1 F10-44-2 (Biodesign International, Kennebunk, ME) is directed against epitopes common to all CD44 isoforms. The mAb BU52 (Binding Site, Inc., San Diego, CA) also is directed against epitopes common to all CD44 isoforms. The mAb BRIC 205, which effectively blocks CD44 hyaluronate binding, was a kind gift from Dr. D. J. Anstee (Bristol, United Kingdom) (23).
The CD44H receptorglobulin was prepared as described previously (1). Briefly, oligonucleotide-primed amplification of cDNA sequences of CD44H was performed by polymerase chain reaction. The oligonucleotide primers were designed to encode endonuclease restriction sites to facilitate subsequent cloning into Ig vectors digested with the same restriction enzymes. CD44-Ig constructs were introduced into COS cells by the DEAE-dextran method, and supernatants were harvested 5-8 days post-transfection. Receptorglobulins purified on protein A-Sepharose beads (Repligen, Cambridge, MA) were eluted with 0.1 M citric acid, pH ϭ 3.0, dialyzed overnight, and purified protein concentration was determined using the BCA assay (Pierce).
Metabolic Labeling and Enzymatic Digestion-Tissue culture cells were starved in sulfate-free minimum essential medium (Life Technologies, Inc.) for 6 h and then incubated with 10% dialyzed fetal calf serum and 200 Ci/ml Na 35 SO 4 (1000 Ci/mmol, DuPont NEN) for 12 h. Cells harvested with 5 mM EDTA in PBS were washed with PBS and lysed in a buffer containing 1% Nonidet P-40, 50 mM Tris (pH ϭ 8.0), 150 mM NaCl, 0.1% SDS, 0.5% deoxycholate, 10 g/ml leupeptin, 100 units/ml aprotinin, and 100 g/ml phenylmethylsulfonyl fluoride. Nuclei were removed by centrifugation, and lysates were precleared with protein G-agarose beads (Oncogene Science, Cambridge, MA). Lysates were then incubated with protein G-agarose coated with anti-CD44 mAb F10-44-2 for 1 h at 4°C. The beads were washed with lysis buffer. Immunoprecipitated CD44 was treated with 1 unit/ml keratanase from Pseudomonas species (Sigma catalog no. K2876), 0.5 unit/ml chondroitin ABC lyase from Proteus vulgaris (Sigma catalog no. C2905), or 0.2 unit/ml heparitinase I from Flavobacterium heparinum (Sigma catalog no. H8991) for 1 h at 37°C in PBS. After counting radioactivity of each supernatant for immunoprecipitates treated with enzyme, the beads were washed with PBS and eluted by boiling. Immunoprecipitates were analyzed by 8% SDS-PAGE under reducing conditions. The gels were dried and subjected to autoradiography.
Surface Labeling and Immunoprecipitation of CD44 -Cell surface proteins were labeled using NHS-LC-biotin (Pierce) in PBS for 1 h at 4°C. Excess biotin was then washed away with PBS, and cells were lysed in a lysis buffer containing 0.25% Triton X-100, and proteinase inhibitors. After removal of nuclei, lysates were precleared and immunoprecipitation of CD44 was performed as described above. Immunoprecipitates were analyzed by 8% SDS-PAGE and electroblotted. After blocking the filters, specific proteins were detected using horseradish peroxidase-conjugated streptavidin and an enhanced chemiluminescence system (Amersham Corp.). To determine the pattern of keratan sulfate substitution on CD44, immunoprecipitated CD44 was treated with keratanase under the same conditions as described above.
Western Blot for Receptorglobulin-One g of receptorglobulins were conjugated with protein G-agarose. After washing, receptorglobulinconjugated agarose beads were incubated with or without keratanase (1 unit/ml). The beads were washed with PBS and separated in nonreducing conditions on acrylamide gels, and then transferred to a nitrocellulose filter by electroblotting at 4°C. The filters were blocked for 1 h in PBS containing 5% dry milk, washed in PBS containing 1% dry milk and 0.2% Tween 20, and incubated with mAb F10-44-2 for 1 h at room temperature. Filters were again washed and then incubated with horseradish peroxidase-conjugated anti-mouse antibody (Amersham Corp.) for 30 min. Filters were then washed in TBST (150 mM NaCl, 10 mM Tris, pH ϭ 8, 0.05% Tween 20), and specific proteins were detected using an enhanced chemiluminescence system (Amersham Corp.).
Adhesion Assay-96-well flat-bottomed plates (Corning, Corning, NY) were coated with hyaluronate (Sigma) at 1 mg/ml, or heat-denatured BSA (Sigma) in PBS at 10 mg/ml overnight at 4°C. The plates were washed with PBS, and nonspecific sites were blocked with 10 mg/ml BSA in PBS for 2 h at 37°C. Cells were detached from plates with 5 mM EDTA in PBS and 1 ϫ 10 5 cells in a single cell suspension were added to each well. For some experiments, the cell suspensions were treated with 1 unit/ml keratanase, 0.5 unit/ml chondroitin ABC lyase, 0.2 unit/ml heparitinase, or 2 units/ml hyaluronidase type IV-S from bovine testes (Sigma catalog no. H3884) at 37°C for 30 min and then washed twice in PBS. Adhesion was allowed to proceed for 1 h at 4°C. The plates were inverted and centrifuged at 150 ϫ g for 4 min; unattached cells were aspirated. The number of viable cells was estimated using a colorimetric assay that depends on the reduction by living cells of tetrazolium salt, MTT, to form a blue formazan salt (24,25). Briefly, the adherent cells were placed in RPMI 1640 without phenol red containing 0.5 mg/ml MTT (Sigma) for 2 h at 37°C. The medium was then removed and formazan crystals were solubilized with 50 l of Me 2 SO. After vigorously shaking the plate, the optical density of each well was measured using an automatic plate reader (Anthos HT2) with a 550 nm measurement wavelength and a 650 nm reference wavelength. The percent specific adhesion was determined by calculating the ratio of OD 550/650 of adherent cells to the OD 550/650 of all of the cells initially seeded. All experiments were performed in triplicate.
An enzyme-linked immunosorbent assay was used to measure the adhesion of receptorglobulins to hyaluronate. 96-well flat-bottomed plates were coated with hyaluronate or BSA in PBS overnight at 4°C. The plates were washed, and nonspecific sites were blocked with BSA. Receptorglobulin (5 g/ml) was added to the hyaluronate-or BSAcoated plates and incubated at room temperature for 2 h. For some experiments, receptorglobulin was treated with keratanase (1 unit/ml) for 30 min. The plates were washed with PBS containing Tween 20 (0.1%) four times. Horseradish peroxidase-conjugated anti-human immunoglobulin Fc (Sigma) was added to the plates and incubated at room temperature for 1 h. The plates were washed with PBS containing Tween 20 four times and incubated with o-phenylenediamine (0.4 mg/ ml) (Sigma) in citrate phosphate buffer containing H 2 O 2 at 37°C. The reaction was stopped by addition of 2 N H 2 SO 4 , and the optical density was measured at wavelength 492 nm.
Immunofluorescence-Cells were detached from plates in 5 mM EDTA in PBS, washed with PBS, and then treated with or without keratanase (1 unit/ml) for 30 min. Cells were washed in PBS and then incubated with BU52 (5 g/ml) in 1% BSA in PBS at 4°C for 30 min. Cells were washed in PBS and then incubated with fluorescein-labeled anti-mouse mAb (Sigma) for 30 min at 4°C. Cells were washed in PBS, resuspended in PBS, and analyzed on a FACScan™ (Becton-Dickinson Co., Mountain View, CA).

Differences between KM12 Clones in CD44
Glycosaminoglycan Substitution-Poorly metastatic KM12C6 cells and the highly metastatic KM12L4 cells express predominantly high molecular weight CD44 alternative splice variants and do not express CD44H (21). CD44H expressed on the cell surface of these two cell lines after CD44H cDNA transfection differ in their functional properties. Only CD44H expressed on KM12C6⌬H transfectants binds to hyaluronate and reduces in vitro and in vivo growth. CD44H expressed on KM12L4⌬H transfectants binds to hyaluronate poorly and does not reduce in vitro or in vivo growth. Both cell lines substitute CD44 with sulfated glycosaminoglycans, and the molecular mass of CD44H expressed by the two KM12 clones appears similar by immunoprecipitation analysis (Fig. 1A). Treatment with keratanase and chondroitinase resulted in both a shift in mobility and a decrease in the intensity of the band corresponding to CD44H. This is in contrast to treatment with heparitinase, which did not result in any changes in mobility or band intensity in KM12C6⌬H. Results from quantitation of 35 SO 4 -labeled oligosaccharide released after treatment with keratanase correlated with the degree of mobility shift seen by PAGE analysis (data not shown). These data suggest that both KM12C6⌬H and KM12L4⌬H cell lines substitute CD44H with keratan sulfate and chondroitin sulfate.
However, these two cell lines differ in their degree of keratan Effect of Keratanase Treatment on CD44 Adhesion to Hyaluronate-We next examined if these specific differences in CD44H post-translational modification between the two KM12 clonal variants resulted in differences in CD44H function. Cells treated with keratanase were tested for adhesion to either hyaluronate or BSA. Significantly enhanced hyaluronate binding was detected after treatment of KM12L4⌬neo (control transfectant) and KM12C6⌬neo cells with keratanase (Fig. 2) To examine the possibility that keratanase treatment enhanced cell adhesion to hyaluronate through an increase in cell surface CD44, we performed fluorescence-activated cell sorting analysis with mAb BU52 to measure cell surface CD44 before and after keratanase treatment (Fig. 3). No immediate changes in cell surface CD44 were detected, indicating that keratanase treatment modified pre-existing cell surface CD44 and did not induce an up-regulation of CD44 expression.
Effect of Removal of Additional Glycosaminoglycans on CD44 Function-We next examined if enhanced hyaluronate adhesion could be detected after treatment to remove chondroitin sulfate, hyaluronate, or heparan sulfate. CD44H cDNA transfectants and control transfectants were treated with chondroitin ABC lyase, hyaluronidase, heparitinase, or keratanase and then tested for adhesion to hyaluronate (Fig. 4) with heparitinase treatment, and a very minimal increase in KM12L4⌬H cell binding was noted after treatment with chondroitin ABC lyase, hyaluronidase, and heparitinase. These slight increases were not nearly as significant in magnitude as the increase in hyaluronate binding detected after removal of keratan sulfate. Clearly, keratan sulfate removal had the most dramatic impact on CD44-mediated adhesion, and this effect was not seen after treatment with chondroitin ABC lyase, hyaluronidase, or heparitinase.
Effect of Keratanase Treatment on CD44 Receptorglobulin Adhesion to Hyaluronate-While the enhanced hyaluronate adhesion measured in keratanase-treated intact cells was significantly inhibited by anti-CD44 mAb BRIC 205, it remained possible that the principal effect of keratanase treatment was not on CD44 itself, but rather on other extracellular proteins associated with CD44 that influence its interaction with hyaluronate. Therefore, to further examine the influence of keratan sulfate substitution on CD44 adhesion to hyaluronate, we examined the effect of keratanase treatment on CD44H receptorglobulins. The CD44H receptorglobulin, which is a fusion protein consisting of the CD44H extracellular domain and human IgG1, has been described previously (1,26). Purified CD44H receptorglobulin was treated with keratanase and then tested for adhesion to hyaluronate. The molecular mass of CD44H receptorglobulin decreased after keratanase treatment (Fig. 5A), and the resulting CD44H receptorglobulin demonstrated enhanced adhesion to hyaluronate (Fig. 5B). These data indicate that treatment modifies CD44 itself and thereby results in enhancement of its binding to hyaluronate. The CD44H receptorglobulin used in this study was produced in COS cells, which posttranslationally modify CD44H differently than the colon carcinoma cell lines used in this study. Nonetheless, data from these experiments clearly support the functional importance of CD44H keratan sulfate substitution, because similar to the colon carcinoma cell lines examined, COS cells substitute CD44H with keratan sulfate.  (26,27). Changing amino acid 41 from arginine to alanine in this domain completely abolishes its affinity for hyaluronate (26). If differences in CD44H keratan sulfate substitution influence its interaction with hyaluronate through the B loop domain, then treatment of a mutant CD44H with arginine 41 changed to alanine should not enhance its binding to hyaluronate. Conversely, the finding that keratanase treatment of this mutant CD44H does enhance its hyaluronate binding would suggest that removal of keratan from CD44H is KM12 clones transfected with a site-directed mutant CD44H cDNA to express CD44H with arginine at position 41 changed to alanine have been described previously. As expected, KM12L4⌬41R/A and KM12C6⌬41R/A cells demonstrate the same hyaluronate binding characteristics as do KM12L4⌬neo and KM12C6⌬neo cells (Fig. 6). These results confirm that the mutated CD44H expressed by the KM12L4⌬41R/A and KM12C6⌬41R/A cells bind to hyaluronate poorly. Treatment of these cells with keratanase did not enhance their binding to hyaluronate to any greater extent than observed in the control cells. In other words, the keratanase treatment enhanced hyaluronate binding of only the high molecular weight CD44 isoforms, and not of the site-directed mutant CD44H. These results strongly suggest that keratan substitution on CD44H modulates its interaction with hyaluronate through its B loop domain.

Identification of CD44 Domain Responsible for Enhanced
Effect of CD44 Glycosaminoglycan Substitution in other Human Colon Carcinomas-Our initial impetus for the studies presented herein was to examine naturally occurring differences in CD44H post-translational modification between two KM12 clones that differ in metastatic potential. We then studied one additional human colon carcinoma cell line (HT29) to determine if the apparent effects of CD44H keratan sulfate substitution could be observed in a colon carcinoma from a completely separate origin. Treatment of CD44 immunoprecipitates from HT29 cells to remove glycosaminoglycans revealed a substitution pattern similar to that seen in KM12L4 cells (data not shown). HT29 cells modify CD44 with keratan sulfate. And similar to our findings in the KM12 clones, re-moval of keratan sulfate from CD44 enhanced hyaluronate binding in HT29 cells (Fig. 7). This enhancement was detected both in mock-transfected HT29⌬neo cells, which express only high molecular weight CD44, and in the HT29⌬H cells transfected to express CD44H. The enhancement could be blocked with mAb BRIC 205 (data not shown). Treatment with keratanase did not increase cell surface CD44 expression (data not shown). No enhancement in hyaluronate affinity was detected in the HT29 transfectants after treatment with chondroitin ABC lyase, hyaluronidase, or heparitinase. Therefore, the effects of keratan sulfate modification of CD44H on HT29 colon carcinoma cells are similar to the effects seen with the KM12 colon carcinoma cells. Specifically, keratan sulfate modification of CD44 reduces its adhesion to hyaluronate.

DISCUSSION
The diversity of biological functions attributed to CD44 may result from its role as a cell surface adhesion molecule that binds to hyaluronate. Given the broad tissue distribution of CD44, it is reasonable to assume that CD44 interaction with extracellular matrix is tightly regulated. This hypothesis is supported by our finding that CD44H expressed on two individual clonal variants of a single colon carcinoma cell line drastically differ in their ability to bind hyaluronate. It has also been reported previously that hyaluronate binding by CD44 expressed on murine T cells is transiently activated during an in vivo immune response (28). Potential mechanisms that regulate CD44 interaction with extracellular matrix include: 1) alternative splicing, 2) phosphorylation of residues in the cytoplasmic domain, 3) interaction of the cytoplasmic domain with intracellular proteins, 4) posttranslational modification by glycosylation or glycosaminoglycan substitution, 5) interaction with other cell surface proteins, 6) interaction with extracellular ligands, and 7) masking or shedding of cell surface CD44 (for review, see Ref. 29).
Several tumor types express different CD44 alternative splice isoforms compared to their normal tissue counterparts FIG. 5. Influence of keratan sulfate modification on CD44H receptorglobulin adhesion to hyaluronate. CD44H receptorglobulin before and after keratanase treatment was separated by SDS-PAGE, transferred to nitrocellulose, and detected with mAb F10-44-2. The molecular weight of CD44H receptorglobulin decreased after keratanase treatment (A). CD44H receptorglobulin adhesion to hyaluronate or BSA before and after keratanase treatment was measured and found to be significantly greater after removal of keratan sulfate (B  (7, 30 -34); therefore, CD44 alternative splicing has received the most attention in studies of CD44 function. Our studies have focused on a distinct and separate mechanism that influences CD44 function, namely keratan sulfate substitution on the CD44 protein. CD44 is known to undergo extensive posttranslational modification, including N-and O-linked glycosylation, and substitution with several glycosaminoglycans (16,17,19,20). Recently it was reported by Jackson and colleagues that only CD44 isoforms that contain exon v3 are modified with heparan sulfate or chondroitin sulfate on Namalwa lymphoma cells (18). These authors reported that CD44H and other isoforms lacking exon v3 were not modified with either heparan sulfate or chondroitin sulfate. In contrast, we have detected keratan sulfate substitution on CD44H, which does not contain exon v3, in three different colon carcinoma cell lines, and demonstrated its influence on CD44 function. Furthermore, the degree of keratan sulfate modification of CD44 differs between different tumor cell lines.
Little is understood about the functional consequences of CD44 glycosaminoglycan substitution. Chondroitin sulfate substitution on CD44 enhances melanoma motility and invasive properties in vitro (35,36). It has recently been reported that heparan sulfate modification of exon v3 containing CD44 isoforms allows presentation of heparin-binding growth factors such as basic fibroblast growth factor and epidermal growth factor (19). However little else is known about the functional consequences of CD44 glycosaminoglycan substitution, especially in epithelial cells. Our finding that two clonal variants of a colon carcinoma cell line, KM12L4 and KM12C6, differ in their CD44H function (21) led us to the current series of investigations in which we report that naturally occurring differences in keratan sulfate modification of CD44 between KM12L4 and KM12C6 cells account for their differences in hyaluronate adhesion.
The mechanism by which keratan sulfate modification of CD44 alters its adhesion to hyaluronate remains unclear. CD44 contains a cluster of basic residues in the B loop region that is responsible for the majority of its adhesion to hyaluronate (27), and mutation of arginine 41 drastically reduces its adhesion to hyaluronate (26). Our data using mutant CD44H indicate that keratan sulfate substitution modulates the interaction of this specific region with hyaluronate. CD44 binding to hyaluronate can also be influenced by treatment with mAb IRAWB 14 (37) or by inclusion of alternatively spliced exons in the membrane proximal region (18,38). In concert, these data suggest that CD44 keratan sulfate substitution may modulate adhesion to hyaluronate via changes in protein conformation.
Although we have demonstrated that naturally occurring differences between cells in their addition of keratan sulfate to CD44H modulates its binding to hyaluronate, the actual sites on CD44H containing the keratan sulfate substitutions are not known.
Serine residues as part of a S-G-X-G sequence preceded by several acidic residues appear to serve as sites for chondroitin sulfate and heparan sulfate substitution (39); however, no clear consensus sequence exists for keratan sulfate substitution. A hexapeptide motif repeated in the keratan sulfate-enriched region of bovine cartilage proteoglycan has been suggested as a possible consensus sequence for keratan sulfate substitution (40), but CD44H does not contain this sequence. The amino acid sequence A-P-S-P-G, which was deduced to contain the keratan sulfate linkage site in pig nucleus pulposus proteoglycan (41), is also not contained in CD44H. Keratan sulfate is linked to bovine fibromodulin through asparagine-glycosidic linkages (42). These linkages do not appear to be the major site of keratan sulfate substitution on CD44H, because growth of KM12 and HT29 cells in tunicamycin does not reduce keratan sulfate substitution (data not shown). Consequently, we are unable to use site-directed mutagenesis of CD44H cDNA to inhibit keratan sulfate substitution on CD44H in colon carcinoma cells.
The enhancement in hyaluronate adhesion by KM12 and HT29 cells occurred nearly immediately after treatment with keratanase. This rapid increase in hyaluronate adhesion suggests that keratanase treatment modified preexisting CD44 molecules, rather than induced new CD44 expression. This conclusion is supported by fluorescence-activated cell sorting data that demonstrates no change in cell surface CD44 after keratanase treatment. This conclusion is also supported by the receptorglobulin adhesion data that demonstrate enhanced CD44H receptorglobulin adhesion to hyaluronate after treatment with keratanase. The rapid influence of keratanase treatment combined with the receptorglobulin data also suggest that new protein synthesis is not required for the enhanced hyaluronate adhesion. This regulatory mechanism is distinct from that of phorbol 12-myristate 13-acetate-inducible binding of lymphocytes to hyaluronate, in which new protein synthesis is essential (43).
In conclusion, we have demonstrated that CD44 function on colon carcinoma cells is modulated by posttranslational modification. Specifically, we have identified differences in keratan sulfate substitution on CD44 that markedly modulate its interaction with hyaluronate through its B loop domain. The impact of keratan sulfate substitution on CD44 function is significant and may be functionally as important as CD44 alternative splicing. Closer examination of this CD44 regulatory mechanism in investigations of development, tumor metastasis, and lymphocyte function will likely reveal it to be an important regulator for these biological processes. Adhesion of CD44H cDNA transfectants (HT29⌬H) and control transfectants (HT29⌬neo) to hyaluronate was measured after treatment with medium, 0.5 unit/ml chondroitin ABC lyase (chondroitinase), 2 units/ml hyaluronidase, 0.2 unit/ml heparitinase, or 1 unit/ml keratanase at 37°C for 1 h. Data are presented as the mean Ϯ S.D. of triplicate experiments. Only treatment with keratanase enhanced adhesion to hyaluronate. Ⅺ, medium; u, chondroitinase; o, hyaluronidase; s, heparitinase; f, keratanase.