Transmembrane and Extracellular Domains of Syndecan-1 Have Distinct Functions in Regulating Lung Epithelial Migration and Adhesion*

Background: Syndecan-1 regulates migration and adhesion in the lung epithelium. Results: The transmembrane domain modulates focal adhesion disassembly, whereas the extracellular domain regulates cell adhesion. Conclusion: Lung epithelial migration is governed by the syndecan-1 transmembrane domain and is independent of the extracellular domain control of cell adhesion. Significance: Syndecan-1 is a receptor that uses multiple domains to facilitate its biological function. Syndecan-1 is a cell surface proteoglycan that can organize co-receptors into a multimeric complex to transduce intracellular signals. The syndecan-1 core protein has multiple domains that confer distinct cell- and tissue-specific functions. Indeed, the extracellular, transmembrane, and cytoplasmic domains have all been found to regulate specific cellular processes. Our previous work demonstrated that syndecan-1 controls lung epithelial migration and adhesion. Here, we identified the necessary domains of the syndecan-1 core protein that modulate its function in lung epithelial repair. We found that the syndecan-1 transmembrane domain has a regulatory function in controlling focal adhesion disassembly, which in turn controls cell migration speed. In contrast, the extracellular domain facilitates cell adhesion through affinity modulation of α2β1 integrin. These findings highlight the fact that syndecan-1 is a multidimensional cell surface receptor that has several regulatory domains to control various biological processes. In particular, the lung epithelium requires the syndecan-1 transmembrane domain to govern cell migration and is independent from its ability to control cell adhesion via the extracellular domain.

Syndecan-1 is a cell surface proteoglycan that can organize co-receptors into a multimeric complex to transduce intracellular signals. The syndecan-1 core protein has multiple domains that confer distinct cell-and tissue-specific functions. Indeed, the extracellular, transmembrane, and cytoplasmic domains have all been found to regulate specific cellular processes. Our previous work demonstrated that syndecan-1 controls lung epithelial migration and adhesion. Here, we identified the necessary domains of the syndecan-1 core protein that modulate its function in lung epithelial repair. We found that the syndecan-1 transmembrane domain has a regulatory function in controlling focal adhesion disassembly, which in turn controls cell migration speed. In contrast, the extracellular domain facilitates cell adhesion through affinity modulation of ␣ 2 ␤ 1 integrin. These findings highlight the fact that syndecan-1 is a multidimensional cell surface receptor that has several regulatory domains to control various biological processes. In particular, the lung epithelium requires the syndecan-1 transmembrane domain to govern cell migration and is independent from its ability to control cell adhesion via the extracellular domain.
The syndecan family of proteoglycans is composed of four mammalian members with unique distribution and function (1,2). Syndecan-1-4 are type I transmembrane proteins and, like all proteoglycans, are decorated with glycosaminoglycan side chains (3). The distinct tissue and cellular distribution of the various syndecan family members could account for their non-overlapping functions, but increasing data also suggest that the core protein has unique features that tailor specific roles (4 -11).
The syndecan proteins are highly conserved across species. However, individual syndecan proteoglycans have distinct features compared with other family members. The ectodomain is the most unique region across the different syndecan proteins, varying in length and sequence homology (1). Although all of the syndecans have heparan sulfate glycosaminoglycans attached to the extracellular domain, syndecan-1 and syndecan-3 also have chondroitin sulfate side chains. The short cytoplasmic tail contains a variable stretch of sequences (V region) flanked by two highly conserved regions (C1 and C2). Finally, the single-pass transmembrane domain, although relatively conserved among family members, has enough differences to confer unique functions (12).
Although found on most cells in the body, syndecan-1 is most abundant on epithelial and plasma cells (13). Shedding of the syndecan-1 ectodomain is an important mechanism for controlling biological processes such as inflammation (14 -17). However, syndecan-1 also functions as a cell surface receptor and governs in vivo processes such as wound repair, gut barrier function, and lipid metabolism (18 -23). In contrast to homeostatic functions, malignant cells also co-opt syndecan-1 to regulate cell invasion, angiogenesis, and tumorigenesis (8,10,11,24).
Syndecan-1 has cell-and tissue-specific functions. A prime example is the fact that syndecan-1 augments epithelial cell migration in skin and cornea but restrains migration in lungs (18 -20, 25). The divergent actions of syndecan-1 are thought to be largely controlled by the cellular context, where the appropriate co-receptor is presented along with syndecan-1 to create a signaling complex that can then transduce extracellular clues to modulate cellular function (26). Accordingly, the multiple domains of syndecan-1 have different effects in regulating these processes. Although some studies have demonstrated that the extracellular domain facilitates cell spreading (4, 9, 10), others have found that the syndecan-1 cytoplasmic domain is the requisite portion (27)(28)(29).
Our previous work showed that syndecan-1 regulates both cell adhesion and migration (18). Here, we used wild-type and mutant syndecan-1 constructs to map out the relevant domains of the core protein that govern the phenotypic response in the lung epithelium. We demonstrate that the syndecan-1 extracellular domain controls cell adhesion through ␣ 2 ␤ 1 integrin affinity modulation. However, the extracellular domain by itself has no effect on cell migration and requires the transmembrane domain to control migration speed and focal adhesion disassembly.
Migration Assay-Monolayers of B2b shRNA.scr and B2b shRNA.hSdc1 cells were plated on No. 1.5 chambered coverglass (Corning, Union City, CA) coated with rat tail type I collagen (2 g/cm 2 ; BD Biosciences). Monolayers were wounded with a sterile P100 pipette tip, and migration was observed under a Nikon TiE inverted widefield fluorescence microscope, which has a humidified chamber to maintain cells at 37°C and 5% CO 2 . Differential interference contrast images were obtained every 10 -20 min for up to 12 h using a CFI 60ϫ/1.49 NA Apo TIRF oil immersion objective. Migration speed was determined by manually measuring the distance traveled by the cell front over time using Nikon Elements AR software. In transient transduction experiments, cells expressing mutant syndecan-1 were identified by coexpression of eGFP.
Focal Adhesion Disassembly Assays-B2b shRNA.scr and B2b shRNA.hSdc1 cells stably expressing paxillin-eGFP were created previously (31). Cells were transduced with an AAV vector to coexpress mouse syndecan-1 and mCherry. To determine the focal adhesion disassembly rate in migrating cells, monolayers of cells were injured and observed by total internal reflection fluorescence (TIRF) microscopy using a CFI 60ϫ/1.49 NA Apo TIRF oil immersion objective. Additionally, focal adhesion disassembly was induced with a nocodazole washout assay (32,33). The fluorescence intensity of the focal adhesion was measured over the entire life span (NIH ImageJ). All data were normalized so that the fluorescence intensity was a percentage of the maximum intensity for the individual focal adhesion. The rate of focal adhesion disassembly was determined for mCherry-expressing cells as described previously (31,32,34).
Fluorescence Recovery after Photobleaching-A Nikon A1R scanning laser confocal microscope using a CFI 60ϫ/1.40 NA Plan Apo VC oil immersion objective fitted with a Tokai Hit stage-top incubator to maintain cells in a humidified chamber at 37°C and 5% CO 2 was used for fluorescence recovery after photobleaching experiments. Cells were transduced with an AAV vector to coexpress mouse syndecan-1 and mCherry. Migrating B2b shRNA.scr and B2b shRNA.hSdc1 cells stably expressing paxillin-eGFP had leading edge focal adhesions selectively photobleached in mCherry-expressing cells. With the 405 and 488 nm lasers set to 50% power, a 1-s pulse was applied over the entire focal adhesion, which would consistently photobleach the focal adhesion to 80% of the prebleaching intensity. The fluorescence recovery was measured with Nikon Elements AR software, and the half-time of recovery was determined as described previously (31). Integrin Activation Experiments-Assays measuring cell adhesion to type I collagen were performed following published protocols (18). So that data from independent experiments could be compared, all adhesion data were normalized to the percentage of adherent B2b shRNA.scr cells. Additionally, cells stably expressing mutant mouse syndecan-1 were stained with activation state-specific antibodies (clones 12G10 and TS2/16) to determine the affinity state of the ␤ 1 integrin subunit. Flow cytometry was performed with a Guava bench-top flow cytometer (Millipore, Billerica, MA). (18). Here, we used shRNA to suppress syndecan-1 expression in lung epithelial cells and found that cells with suppressed syndecan-1 expression (B2b shRNA.hSdc1 ) had a faster migration speed than cells expressing a control shRNA (B2b shRNA.scr ) To map the domains of the syndecan-1 core protein that control cell migration, we expressed wild-type and mutant syndecan-1 cDNAs into B2b shRNA.hSdc1 cells and compared their migration speed with that of control B2b shRNA.scr and B2b shRNA.hSdc1 cells (eGFP expression only). In contrast to fulllength syndecan-1 (mSdc1, 311-amino acid core protein), mSdc1-CT1 lacked the cytoplasmic domain (amino acids 1-279), and mSdc1-GPI lacked both the transmembrane and cytoplasmic domains (amino acids 1-252) ( Fig. 2A) (10). The entire syndecan-1 extracellular domain was present in all of the constructs. The expression levels of full-length and mutant mSdc1 proteins were similar under all conditions (supplemental Fig. S1).

Syndecan-1 Transmembrane Domain Restrains Cell Migration-Syndecan-1 restrains migration of lung epithelial cells
We found that B2b shRNA.hSdc1 cells migrated faster than B2b shRNA.scr cells (Fig. 2B, Table 1, and supplemental Video 2). Both mSdc1 and mSdc1-CT1 slowed migration of B2b shRNA.hSdc1 cells to a similar rate compared with B2b shRNA.scr cells. Only mSdc1-GPI, which contains the entire syndecan-1 ectodomain but is anchored to the cell membrane by a glycosylphosphatidylinositol domain (11,(35)(36)(37), lacked the ability to slow cell migration ( Table 1), indicating that the transmembrane region is the crucial domain regulating lung epithelial migration.
To ensure that the glycosylphosphatidylinositol anchor did not have any inhibitor actions on syndecan-1, we used a mutant syndecan-1 that had polyleucine substitutions in the transmembrane domain (mSdc1-pL) (supplemental Fig. S2) (38). In concordance with the data from the mSdc1-CT1 and mSdc1-

Syndecan-1 Domains in Migration and Adhesion
GPI mutants, the mSdc1-pL mutant did not slow migration of B2b shRNA.hSdc1 cells.

Syndecan-1 Transmembrane Domain Controls Focal
Adhesion Disassembly-Cell migration speed is largely determined by the focal adhesion turnover rate (32,37,39). Our previous work demonstrated syndecan-1 restrains cell migration by slowing focal adhesion disassembly (31). Therefore, we evaluated the disassembly of focal adhesions in migrating cells (Fig.  3). Kymographs of individual focal adhesions best represented the turnover through time (Fig. 3A). Moreover, the intensity of the focal adhesion was measured, and the downward slope of intensity correlated with the disassembly rate (Fig. 3B) (31,32,37). Migrating B2b shRNA.hSdc1 cells were compared with B2b shRNA.scr cells and had a significantly shorter focal adhesion life span and faster disassembly ( Table 2), consistent with our prior observations (31). Whereas mSdc1-GPI had no effect, mSdc1 and mSdc1-CT1 both reversed the focal adhesion life span and disassembly differences in B2b shRNA.hSdc1 cells to become more similar to those in B2b shRNA.scr cells, which trends with the pattern seen with the migration speed.
Microtubules target focal adhesions and induce disassembly (32,33). Forced disassembly of focal adhesions occurred within minutes of microtubule polymerization after nocodazole washout (Fig. 4, A-E). The intensity of the focal adhesion was measured (Fig. 4F), and the disassembly rate was calculated ( Table  2). B2b shRNA.hSdc1 cells had faster disassembly of focal adhesions compared with B2b shRNA.scr cells. These findings are congruous with the focal adhesion disassembly rate of migrating cells (31). Moreover, the transmembrane domain once again was the relevant domain that could reverse the accelerated focal adhesion disassembly rate in B2b shRNA.hSdc1 cells ( Table 2).

Syndecan-1 Transmembrane Domain Facilitates Exchange of Adhesion Complex
Proteins-Cycling of adhesion complex proteins (e.g. paxillin and focal adhesion kinase) between focal adhesions and the cytosol controls migration speed by altering the stability of the focal adhesions (31)(32)(33)(34)40). Fluorescence recovery after photobleaching was used to evaluate the exchange rate of adhesion complex proteins (Fig. 5, A and B). Consistent with our previous observations (31), the dynamic cycling of paxillin was faster in B2b shRNA.scr cells than in B2b shRNA.hSdc1 cells ( Table 3). The kinetics of adhesion complex protein exchange was reversed and similar to those in B2b shRNA.scr cells only when mSdc1 and mSdc1-CT1 were introduced into B2b shRNA.hSdc1 cells.
Syndecan-1 Ectodomain Is Required for Integrin Activation-Changes to the integrin affinity state can modulate cell migration (32,41). Because syndecan-1 regulates both cell migration and ␣ 2 ␤ 1 integrin activation (18), we evaluated the relevant domains of syndecan-1 that control ␣ 2 ␤ 1 integrin activation. Using activation state-specific antibodies, we quantified the level of the high affinity ␤ 1 integrin subunit (Fig. 6). The high  OCTOBER 12, 2012 • VOLUME 287 • NUMBER 42 JOURNAL OF BIOLOGICAL CHEMISTRY 34931 affinity ␤ 1 subunit was less abundant in B2b shRNA.hSdc1 cells compared with B2b shRNA.scr cells, replicating our prior findings (18). However, in contrast to cell migration, where the transmembrane domain regulated migration speed and focal adhesion dynamics, integrin activation was controlled by the syndecan-1 ectodomain. Indeed, mSdc1, mSdc1-CT1, and mSdc1-GPI were all capable of increasing the level of the activated ␤ 1 integrin subunit in B2b shRNA.hSdc1 cells compared with B2b shRNA.scr cells (Fig. 6).

DISCUSSION
After epithelial injury, cell-cell and cell-matrix interactions form a contextual framework vital to guiding and optimizing collective cell migration (42,43). Although the cells in an epithelial surface migrate in concert while maintaining their cellular interconnections, each individual cell retains characteristics of single-cell migration (44). Syndecan-1 restrains cell migration and facilitates the activation of ␣ 2 ␤ 1 integrin (18). Here, we have demonstrated that syndecan-1 regulates cell migration independent of the ability to control the ␣ 2 ␤ 1 integrin activation state. The transmembrane domain is the critical portion of the syndecan-1 core protein that controls cell migration by controlling focal adhesion disassembly. In contrast, cell adhesion is  a B2b shRNA.hSdc1 and B2b shRNA.hSdc1 ϩ mSdc1-GPI cells were significantly longer than B2b shRNA.scr cells by one-way ANOVA and Bonferroni post hoc analysis (p Ͻ 0.001; n ϭ three independent experiments). A minimum of five focal adhesions were measured for each condition in each experiment.

Syndecan-1 Domains in Migration and Adhesion
controlled by the syndecan-1 ectodomain through integrin affinity modulation. Various functions have been identified for the syndecan-1 extracellular domain. Angiogenesis and cancer cell invasion are regulated by the syndecan-1 ectodomain (8,10,11,45). Additionally, the syndecan-1 extracellular domain is required for cell adhesion and spreading (4, 8 -10). Affinity modulation of the ␣ v ␤ 3 and ␣ v ␤ 5 integrins also requires the extracellular domain (5,7,8). Moreover, the heparan sulfate side chains are required to engage the matrix, but distinct portions of the syndecan-1 core protein are required for its cell-specific functions (5,27,28). Although heparan sulfate can bind collagen, this interaction is most likely not the main determinant of cell adhesion (9,10,46). These findings are consistent with ours, which indicate that the syndecan-1 extracellular domain mediates the adhesive properties of the cells. Furthermore, at least in the lung epithelium, syndecan-1 facilitates cell adhesion through affinity modulation of ␣ 2 ␤ 1 integrin (18).
Syndecan-1 is a multidimensional protein that coordinates several events for a specific cellular function. For example, ␣ v ␤ 3 integrin and insulin-like growth factor receptor-1 coalesce through syndecan-1 to form a fully functional ternary complex (47). Furthermore, murine B82L fibroblasts send out filopodial projections upon syndecan-1 ligation but also require ␣ v ␤ 5 integrin engagement for a spread phenotype (7). Both the extracellular and transmembrane domains are necessary to induce cell spreading and polarization in Raji lymphoid cells (38). Additionally, the syndecan-1 extracellular domain can augment cell adhesion but in itself is not sufficient to promote cell spreading and invasion (9,10).
Our data show that syndecan-1 truncated at the first cytoplasmic amino acid but retaining the endogenous transmembrane domain (i.e. mSdc1-CT1) is functionally conserved. In contrast, substitutions with a glycosylphosphatidylinositol anchor or with polyleucines in the transmembrane domain both lose the ability to control cell migration. Because the transmembrane domain is believed to facilitate dimerization of the syndecan receptor (1), syndecan-1 may require the transmembrane domain to dimerize in mediating its effects on cell migration. Of particular note, glycine-to-leucine mutations within the transmembrane domain of syndecan-2 and syndecan-4 abrogate both dimerization and effects on migration (48).  Unlike syndecan-2, -3, and -4, which form strong homo-and heterodimers, the syndecan-1 transmembrane domain has relatively weak interactions with itself and other members of the syndecan family (12). Thus, alternatively, the transmembrane domain may not necessarily mediate syndecan-1 function through dimerization but through organizing into distinct cell surface domains. Indeed, Raji cells require lipid rafts for syndecan-1-mediated spreading (38). Syndecan-1 may coalesce in specific membrane pools, bringing it in proximity to other coreceptors to form a multimeric complex required to carry out its cellular function. In fact, the GxxxG sequence within the transmembrane domain of syndecan-1 is found in many transmembrane proteins (49). Thus, syndecan-1 is primed for interactions with other transmembrane proteins carrying a similar GxxxG motif. Moreover, flanking sequences help determine the strength of homo-and heterodimeric bonds, adding specificity of these interactions to the various syndecans (12,50,51). Certainly, the transmembrane and extracellular domains may function in combination to bring together co-receptors on the cell surface. Upon injury, syndecan-1 is shed from epithelial surfaces (17,18,52), and the loss of the syndecan-1 ectodomain could break down these complexes and alter the intracellular signals to induce a migratory phenotype.
Cell must balance their matrix adhesiveness to maximize traction forces without compromising cell body translocation (35)(36)(37). The ability of a cell to migrate can be regulated by changes to the integrin affinity state (32,41). However, our findings demonstrate that alteration to cell adhesion by itself does not affect cell migration. Although these findings indicate the necessity of the transmembrane domain in cell migration, we cannot exclude the additional requirement of the extracellular domain. Our data only show that the extracellular domain in isolation does not regulate cell migration. Interestingly, the extracellular domain enhances cell adhesion, whereas the transmembrane and cytoplasmic domains slows migration when syndecan-1 is exogenously expressed in fibrosarcoma cells, which lack endogenous syndecan-1 expression (53). These results further substantiate the notion that cells use multiple syndecan-1 domains to fully transduce signals from the extracellular matrix into the cell.
Cell migration speed can be controlled by altering the focal adhesion turnover rate (39). Cells can also alter the optimum substrate concentration by regulating focal adhesion dynamics (32,37). Syndecan-1 restrains cell migration by attenuating focal adhesion disassembly, which prolongs focal adhesion life span and slows migration speed (31). Here, we have shown that the syndecan-1 transmembrane domain regulates cell migration. Moreover, our fluorescence recovery after photobleaching experiments demonstrate that the transmembrane domain controls the cycling of adhesion complex proteins, which regulates focal adhesion turnover. Our findings are consistent with the fact that faster exchange of focal adhesion proteins slows migration by attenuating focal adhesion turnover (31,34,40,54,55).
In sum, syndecan-1 most likely interacts with one or more co-receptors to form a signaling complex that transduces signals to augment cell adhesion and restrain cell migration. Although the extracellular domain is necessary for integrin affinity modulation, the transmembrane domain is crucial for controlling focal adhesion disassembly and migration speed in wounded lung epithelial cells. These data support the notion that syndecan-1 coordinates co-receptor complex formation through various domains of its core protein to mediate cellular function. The specific functional outcome is largely determined by the cellular context and regulated coexpression of signaling receptors.