Characterization of dystroglycan binding in adhesion of human induced pluripotent stem cells to laminin-511 E8 fragment

Human induced pluripotent stem cells (hiPSCs) grow indefinitely in culture and have the potential to regenerate various tissues. In the development of cell culture systems, a fragment of laminin-511 (LM511-E8) was found to improve the proliferation of stem cells. The adhesion of undifferentiated cells to LM511-E8 is mainly mediated through integrin α6β1. However, the involvement of non-integrin receptors remains unknown in stem cell culture using LM511-E8. Here, we show that dystroglycan (DG) is strongly expressed in hiPSCs. The fully glycosylated DG is functionally active for laminin binding, and although it has been suggested that LM511-E8 lacks DG binding sites, the fragment does weakly bind to DG. We further identified the DG binding sequence in LM511-E8, using synthetic peptides, of which, hE8A5-20 (human laminin α5 2688–2699: KTLPQLLAKLSI) derived from the laminin coiled-coil domain, exhibited DG binding affinity and cell adhesion activity. Deletion and mutation studies show that LLAKLSI is the active core sequence of hE8A5-20, and that, K2696 is a critical amino acid for DG binding. We further demonstrated that hiPSCs adhere to hE8A5-20-conjugated chitosan matrices. The amino acid sequence of DG binding peptides would be useful to design substrata for culture system of undifferentiated and differentiated stem cells.

heterotrimeric isoforms have been identified in various tissues and cell culture media. Of these isoforms, laminin-511 (LM511; α5, β1, γ1) is widely distributed in the basement membranes of foetal and adult tissues 10 . LM511 is also present in blastocysts and contacts the inner cell mass that is the origin of ES cells 11 . Lack of the α5 chain does not influence the proliferation of the inner cell mass, suggesting that LM511 is not involved in stemness during early development 12 . However, because LM511 is useful for stem cell culture, LM511 is still considered to be involved in maintaining the pluripotency of the inner cell mass.
The interaction of cells with laminins is mediated by various receptors, including integrins and non-integrins 9 . Intact LM511 is bound by integrin α3β1, α6β1, α6β4, α7β1, dystroglycan (DG), and Lutheran/basal cell adhesion molecule (Lu/B-CAM) [13][14][15][16][17] . Cell adhesion to α5-containing laminins is mediated through the binding of integrin α3β1/α6β1 and Lu/B-CAM to the LG1-3 modules, and the preferential binding of dystroglycan to sites in the LG4-5 modules 15,16,18 . Consistent with previous studies, LM511-E8, which contains the LG1-3 modules, exhibits potent cell attachment activity, mediated through integrin α3β1, α6β1, α6β4, and α7β1 19,20 . In human stem cells, integrin α6β1 is a major isoform at cell surfaces 7,21,22 . The fragment containing the integrin α6β1-binding site, enables superior adhesion of single cell-dissociated cultures of hESCs and hiPSCs 7 . Therefore, it is well-known that the adhesion of hiPSCs to LM511-E8 is mainly mediated through integrin α6β1. However, it is unknown whether non-integrin receptors are expressed in stem cells and play roles in pluripotency and differentiation of the cells, cultured on laminins and their fragments.
Defined media and substrata are required for the culture system in order to effectively exploit stem cells for clinical use. In this study, in order to define the adhesion of cultured hiPSCs to LM511-E8, we analyzed the expression of DG in hiPSCs, examined the biochemical properties of DG in hiPSCs cultured on LM511-E8, and characterised the DG binding sequences in LM511-E8 using synthetic peptides. We also explored the possibility that DG-binding peptides could serve as culture substrata. Our results provide useful information to design substrata for culture system of undifferentiated and differentiated stem cells.

Analysis of DG expression in hiPSCs.
Dystroglycan is a non-integrin receptor that is mainly expressed in muscle and nervous system 23 . Previous studies have demonstrated DG expression in many other cell types, including mouse ES cells [24][25][26] . Therefore, we examined the expression of DG in hiPSCs. Two hiPSC clones were used in this study: 201B7 and 454E2, derived from human dermal fibroblasts and dental pulp cells, respectively 27,28 . Immunocytochemistry was first performed to analyse DG expression in hiPSCs cultured in different systems (Fig. 1A). Two monoclonal antibodies, VIA4 and IIH6, were used for the immunostaining. Both antibodies recognized the fully glycosylated α-DG 29 . IIH6 also detects the laminin binding glycoepitope on α-DG, and hypoglycosylation results in the absence of epitopes, for this antibody. VIA4 antibody staining showed that DG was uniformly detected in 201B7 colonies, cultured on either feeder cells or LM511-E8. DG positive cells usually expressed Nanog, which is an undifferentiated state maker. We next performed flow cytometric analysis using hiPSCs cultured on LM511-E8. The results showed that 201B7 cells mostly expressed DG that was recognized by both antibodies (Fig. 1B). The results indicate that hiPSCs express the laminin binding glycoepitope on α-DG. Similar results were observed in 454E2 cells (Fig. S1). However, human dermal fibroblasts (HDFs), which are of hiPSC origin, did not express DG.
Additionally, we performed biochemical analysis of DG expression in hiPSCs. After SDS-PAGE separation, α-DG retains its laminin binding property on PVDF membranes 30 . We performed immunoblotting and laminin overlay assay to examine if DG expressed in hiPSCs, exhibits binding affinity. DG was partially purified from the cell lysates using wheat germ agglutinin (WGA) agarose as described in previous study 31 . Immunoblotting using IIH6 antibody showed that the highly glycosylated α-DG, from both hiPSC clones, migrated broadly at 170-240 kDa (Fig. 2). The laminin overlay assay showed that α-DG of hiPSCs was functionally active. Because MsLM111 overlay assay was more sensitive than the immunoblotting, the band of α-DG migrated more broadly at 150-250 kDa. Moreover, we examined the binding of LM511 that is used for stem cell culture. Although, LM511 bound to α-DG on the membrane, the bands were narrower than those of MsLM111.
The binding of DG to LM511-E8. LM511-E8 now serves as a cell culture substratum that maintains the pluripotency of hESCs and hiPSCs 7 . High expression of DG in hiPSCs allowed us to examine the binding of the DG to LM511-E8. Solid phase binding assays to LM511-E8 were performed using a soluble recombinant DG (MsDG-Fc) composed of mouse α-DG fused to human IgG 1 Fc 32 . Because purified MsDG-Fc was unstable, we used conditioned medium containing MsDG-Fc, for the solid phase binding assays. Conditioned medium, containing Fc, was used as control. LM-111 was used as a positive control in the overlay assay. We first compared the DG binding affinity between α1 and α5 chains, using LM-111 and -511. Both molecules were recognized to similar extents by MsDG-Fc (Fig. 3). Although, the substitution of laminin β1 with laminin β2 in LM-521, promotes survival of human pluripotent stem cells without ROCK inhibitor 33 , the β2 chain did not influence the binding of DG. MsDG-Fc bound to the plate coated with 60 nM of LM-511-E8, however, DG failed to bind to it at 20 nM concentration. These results indicate that although domains of LM-511, other than the E8 region, harbour the major DG binding sites, LM511-E8 does contain DG binding sites.
DG binding activity of human laminin α5 chain peptides. As shown above, the solid phase binding assay indicated that DG binding sites are localized on LM511-E8. Our previous study also showed that DG binds to two synthetic peptides derived from the laminin α2 chain LG4-5 modules 34 . In this study, we synthesized a series of peptides covering the amino acid sequences of LM511-E8 to identify DG binding sites (Figs 4 and S2-3). Peptides derived from the E8 region of laminin α5, β1 and γ1 chains were synthesized. The length of peptides were generally 12-13 amino acids, and they overlapped with neighbouring peptides by four amino acids. The glutamate or glutamic acid at the N-terminus of peptides form pyroglutamine 35 . To avoid the reaction, one amino acid was added at the N-terminus of glutamate or glutamic acid. In addition, to avoid the impact of disulphide bonds, cysteine residues were omitted. The series of peptides were dissolved in PBS (−), coated on 96-well ELISA plates, and tested for their DG binding activity. Of the peptides assayed, hE8A5-20, hA5G1, hA5-G4, hA5-G29, www.nature.com/scientificreports www.nature.com/scientificreports/ and hA5-G47 exhibited DG binding activity at 10 μg/mL (Figs 4 and S2). The series of peptides derived from laminin β1 and γ1 did not bind to DG (Figs S3 and S4). Moreover, we examined the dose-dependency of DG binding activity. hE8A5-20 and hA5G-29 derived from the laminin coiled-coil (LCC) domain and LG2 module, respectively, exhibited strong DG binding activity in a dose-dependent manner (Fig. 5A). The remaining active peptides revealed moderate DG binding activity dose-dependently. hiPSC attachment activity of DG binding peptides. We next examined if hiPSCs adhered to DG binding peptides using peptide-coated 96-well ELISA plates as well as solid phase binding assays. Of the five peptides, only hE8A5-20 exhibited high binding activity for attachment of hiPSCs (Fig. 5B). For culturing hiPSCs, cell adhesive peptides must be coated on cell culture dishes or plates. Although, coating of cell culture plates with the hE8A5-20 peptide posed considerable challenge, it did not exhibit cell attachment activity. Cell adhesive peptides often require mechanical support for cell culture. To address this issue, we have previously developed cell www.nature.com/scientificreports www.nature.com/scientificreports/ adhesive peptide-chitosan matrices for cell culture 36 . The peptide-chitosan matrices often improve cell adhesive activity of peptides. Chitosan matrices are natural polymers and have been used for medical applications, such as suture thread 37 . As shown in Fig. 5C, DG binding peptides were conjugated to chitosan-matrices. Of the five peptide-chitosan matrices, hE8A5-20-conjugated matrices exhibited activity for attachment of hiPSCs (Fig. 5D). Unfortunately, the activities of other peptides could not be improved on chitosan-matrices. Furthermore, although hiPSCs were cultured on hE8A5-20-conjugated matrices, the cells did not proliferate on the matrices (data not shown).
The active core sequence of the hE8A5-20 peptide in DG binding activity. Because hE8A5-20 exhibited DG binding activity that mediates hiPSC attachment, the structural requirements were determined using systematically truncated N-terminal and C-terminal peptide derivatives of hE8A5-20 (Fig. 6A). hE8A5-20e (LLAKLSI), an N-terminal truncated peptide of hE8A5-20, retained activity, whereas hE8A5-20f (LAKLSI), with a deletion of the C-terminal leucine from hE8A5-20e, did not reveal activity. hE8A5-20i (KTLPQLLAKLS), with a deletion of the C-terminal isoleucine from hE8A5-20, abolished DG binding activity. These results show that the seven-amino acid sequence, LLAKLSI, is critical for hE8A5-20's DG binding activity. Furthermore, we synthesized Ala-substituted hE8A5-20 peptides to identify crucial amino acid residues for DG binding activity (Fig. 6B). When the Lys at the 9th position was substituted with Ala, DG binding activity of hE8A5-20 was significantly decreased. Therefore, these results indicate that the 9th Lys is critical for hE8A5-20's DG binding activity.

Discussion
In this study, we characterised the adhesion of hiPSC to LM511-E8, which is generally used in stem cell culture. We first demonstrated the expression of DG, a non-integrin receptor, in hiPSCs cultured on either feeder cells or LM511-E8. HDFs, from which the hiPSCs originate, did not express DG, indicating that it was induced during cellular reprogramming. DG was originally identified in skeletal muscle as a component of the dystrophin-glycoprotein complex that interacts with the actin cytoskeleton and connects it to the muscle basement membrane 23 . DG is also expressed in various tissues and can interact with utrophin in non-muscle cells 38 . Because utrophin was expressed in hiPSCs (Fig. S4), it is likely that DG serves as a linker protein between utrophin and LM511-E8 in these cells.
The major DG binding sites of laminin (α5LG4-5 modules) are absent in LM511-E8. Therefore, the possibility of LM511-E8 binding to DG was not previously considered. In this study, although the overlay assay using LM511-E8 did not exhibit DG binding (data not shown), the solid phase binding assay indicated that LM511-E8 is capable of DG binding. When hiPSCs are cultured on whole LM511 with strong DG binding property, the self-renewal ability and pluripotency are maintained, similarly to when cultured on the E8 fragment only 3,7 . This suggests that because the attachment of hiPSCs to the E8 region mediated via integrin α6β1 is enough for maintaining stemness, the DG binding is not required for self-renewal ability and pluripotency of stem cells. However, because the function of DG-E8 binding is unclear in hiPSCs culture, it remains a possibility that the binding influences hiPSCs behaviour. Recently, Nguyen et al. demonstrated that whole laminin-511 with strong DG binding properties promote the endothelial cell differentiation of hESCs when induced with growth factors 39 . On the other hand, the E8 fragments with weak DG binding properties support the induction of cell differentiation, such as midbrain dopaminergic neurons and forebrain oligodendrocyte precursor cells from hiPSCs 40,41 . Thus, although the induction efficiency of cell differentiation is unclear between the whole molecule and E8 fragment of LM511, the difference of DG binding affinity may be one of cues in cell fate determination. Moreover, it is well known that differentiated cells appear at a low frequency in long-term stem cell cultures 7,42 . We cannot exclude a possibility that the DG binding influences the pluripotency of hiPSCs cultured on LM511-E8. To completely define the cell adhesion of hiPSCs to the substrata, a better strategy would be abolishing the binding activity of DG in LM511-E8.
DG is translated from a single gene followed by posttranslational cleavage, to give rise to the α and β subunits, the two non-covalently associated proteins 24 . α-DG, a heavily glycosylated protein, acts as a binding subunit for  43,44 . The laminin overlay assays suggest that DG of hiPSCs contains O-mannosyl glycans comprising GlcA-Xyl repeats. The O-mannosyl glycans are sequentially modified by several enzymes 45 . Of these, LARGE is a glycosyltransferase-like protein that generates polymers of GlcA-Xyl at the final DG modification step; these are required for the binding of DG to laminin. DG becomes functionally active when cellular reprogramming induces the biosynthetic pathway for O-mannosyl glycans.
We have previously reported that it is possible to define the DG binding sequences within laminin using synthetic peptides, and to further identify the amino acids critical for DG binding 34 48 . Integrin α6β1 binds to the bottom face of laminin α5LG1-3 modules with the laminin γ1-tail. We explored the location of DG binding sequences in the spatial structure of LM511-E8. 2688 KTLPQLLAKLSI 2699 of hE8A5-20 is located in the α-helix of the LCC domain. Because the side chain of K 2696 , critical for DG binding, faces outward (Fig. 7), it could interact with the carbohydrate chains of DG. DG binding sites are localized in laminin globular (LG) domains of laminins, agrin, perlecan, neurexin, pikachurin, and slit, which are currently known as DG ligands 32,[49][50][51] . The DG binding site in the α5LCC domain that we identified, may provide new insight into the interaction of multiple receptors for laminin-511. The amino acid sequences of the remaining peptides are partially and fully buried in the α5LG1-3 modules. Proteolytic modification exposes cryptic sites with biological activity within larger molecules 52,53 . These DG binding sequences may also be cryptic sites that can be exposed under certain circumstances.
Although hE8A5-20-conjugated chitosan matrices exhibited adhesive activity for hiPSCs, they were unable to maintain cell growth. Because integrin α6β1 mediates potent cell attachment to LM511-E8, the self-renewal of hiPSCs would be mainly mediated through the receptor. Several groups reported that laminins with strong DG binding properties are key molecules for the differentiation of hESCs and hiPSCs 39,54,55 . However, it is still unclear as to how cell adhesion to laminins is involved in the commitment to differentiate. hE8A5-20-conjugated chitosan matrices may be useful to clarify the mechanism of cellular differentiation. www.nature.com/scientificreports www.nature.com/scientificreports/ The development of transplantation therapy and drug discovery using hiPSCs requires modulation of cell differentiation. We found that functionally active DG is expressed in hiPSCs. Although the function of DG in stem cells is unclear, it is likely that the binding of DG mediates the differentiation of cells grown on laminins. We have produced the recombinant LM511-E8 that abolishes the binding activity of DG. The recombinant proteins will clarify the biological relevance of DG binding in hiPSC culture. Our approach also revealed that synthetic peptides mimic DG binding to laminin and mediate the adhesion of hiPSCs. The DG binding peptides and their amino acid sequences could be useful to design substrata for culturing differentiated cells, derived from hiPSCs, for cell-based therapy.

Reagents.
Human recombinant laminin-111 (LM111), laminin-511 (LM511), and laminin-521 (LM521) were purchased from BioLamina (Sundbyberg, Sweden). iMatrix-511 (LM511-E8) and mouse EHS laminin (MsLM111) was from Nippi (Tokyo, Japan) and BD Biosciences (San Jose, CA), respectively. Cell culture. Two clones of hiPSCs, 201B7 and 454E2, were purchased from the RIKEN BioResource Center (Tsukuba, Japan). As described in Takahashi et al. 27 , the hiPSCs were maintained on feeder cells in Primate ES medium (ReproCELL, Yokohama, Japan), supplemented with basic fibroblast growth factor (bFGF, ReproCELL). For passaging of conventional colony cultures, hiPSCs were washed once with PBS (−), followed by incubation with Dissociation Solution for human ES/iPS Cells (ReproCELL), at 37 °C for 5 minutes. The detached colonies were collected and suspended in the growth medium. After a couple of passages, the hiPSCs were transferred to a feeder-free culture system, comprising LM511-E8 8 ; culture dishes (Thermo Fisher Scientific, Waltham, MA) were coated with LM511-E8 (according to the manufacturer's instructions) for preparing the feeder-free culture system. The colonies of hiPSCs were dissociated into single cells with cell dissociation buffer (Thermo Fisher Scientific). After incubation at 37 °C for 5 minutes, the cell dissociation buffer was removed. The dissociated cells were suspended in StemFit AK02 (ReproCELL) with Y-27632, and counted using LUNA Automated Cell Counter (Logos Biosystems, Gyeonggi-do, South Korea). About 6 × 10 4 live cells were plated onto LM511-E8-coated 60 mm dishes; Y-27632 was used only at the time of plating 56 . The medium was replaced with StemFit AK02, without Y-27632, on the following day. The medium was changed every other day, until the cells reached 80-90% confluency. hiPSCs were subcultured every 6-8 days.
Human dermal fibroblasts (HDFs) were purchased from Cell Applications (San Diego, CA). The cells were maintained in DMEM containing 10% FBS, 0.1 mg/mL streptomycin and 100 U/mL penicillin G.
Immunocytochemistry. For immunocytochemistry, hiPSCs were cultured on lumox multiwell plates (Sarstedt, Numbrecht, Germany) with feeder cells or coated with LM511-E8. After culturing for 4-7 days, the colonies of hiPSCs were doubly stained with the antibodies listed in Table 1. Briefly, the cells were fixed with 4% paraformaldehyde/PBS (−), and the unreacted aldehydes were blocked with 0.1 M Glycine/PBS (−). The fixed cells were permeabilized with 1% Triton-X100/PBS (−) and blocked in 10% normal goat serum. After blocking, the cells were incubated with primary antibodies. Secondary antibodies conjugated to Alexa 488 or 594 were used in these experiments. Hoechst 33258 was used for counter staining of the nuclei. After several washes, the cells were mounted in 90% glycerol, containing 0.1xPBS (−) and 1 mg/mL p-phenylenediamine. Images were captured using BZ-X700 (Keyence, Osaka, Japan).  Laminin overlay assays. Laminin overlay assays were performed on PVDF membranes using mouse EHS laminin (MsLM111) and human laminin-511 (LM511), as previously described, with slight modifications 31 . Briefly, PVDF membranes were blocked with TBS (+) (10 mM Tris-HCl, 150 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , pH7.5) containing 5% non-fat dry milk for 1 hour at room temperature. The membranes were incubated with 6 μg/mL of laminin at 4 °C overnight, in TBS (+) containing 3% BSA. The bound laminin was probed with anti-laminin-111 polyclonal antibody, followed by incubation with anti-rabbit IgG antibody conjugated with horseradish peroxidase. TBS (+) was used for antibody dilution and membrane washing.
Preparation of conditioned media containing MsDG-Fc. cDNAs encoding mouse α-DG fused with human IgG 1 Fc (MsDG-Fc) and control Fc expression vectors were constructed in our previous studies 18,32 . The recombinant proteins were expressed using the Expi293 Expression System according to the manufacturer's instructions. The conditioned media (CM) were clarified through 0.22 μm pore filters and dialyzed against TBS (+). MsDG-Fc CM and Fc CM were diluted 1:3 with TBS (+) and used for solid phase binding assays.
Solid phase binding assays. Solid phase binding assays were carried out with recombinant laminins coated onto high protein-binding capacity 96-well ELISA plates (IWAKI, Tokyo, Japan). Plates were blocked with PBS (−) containing 1% BSA and incubated with MsDG-Fc at room temperature for 1 hour. After washing with TBS (+), the bound MsDG-Fc was detected with a biotinylated anti-human IgG Fc antibody (Jackson ImmunoResearch). After further steps of washing, the bound antibodies were detected by addition of streptavidin-conjugated horseradish-peroxidase, followed by addition of 1 mg/mL o-phenylenediamine and 0.012% H 2 O 2 . The absorbance was measured at 450 nm with Multiskan GO microplate Spectrophotometer (Thermo Fisher Scientific).
Synthetic peptides and cell attachment assays. The 9-fluorenylmethoxycarbonyl (Fmoc)-based solid-phase method with a C-terminal amide was used for the synthesis of all peptides, as described previously 57 . The purity and identity of the synthetic peptides were verified by analytical HPLC and electrospray ionization mass spectrometry at the Central Analysis Centre, Tokyo University of Pharmacy and Life Sciences. Cell attachment assays using synthetic peptides were performed as previously described with slight modifications 58 . Briefly, the synthetic peptides were dissolved in PBS (−) at 1 mM concentration, and 50 μL of this solution was added to each well of the high protein-binding capacity 96-well ELISA plates (IWAKI, Tokyo, Japan). After coating overnight at 4 °C, the plates were blocked with 1% BSA in PBS (−). The dissociated hiPSCs were suspended in 0.1% BSA in DMEM with Y-27632, plated at 1.0 × 10 4 cells/50 μL/ well. After incubation at 37 °C for 1 hour, the adhered cells were stained with Diff-Quik (International Reagents Corp., Kobe, Japan). The number of attached cells was counted under a microscope.
Preparation of peptide chitosan matrices. Maleimidobenzoyloxy (MB)-chitosan was prepared as previously described 59 . For conjugation to chitosan membrane, a cysteine residue was added at the N-terminus, and two glycine residues were used as a spacer between the cysteine and the DG binding peptide sequences. Various amounts of peptides (0.5-50 nmol/well) in 1% trifluoroacetic acid (in Milli-Q water) and an equal amount of 1% NaHCO 3 solution, were added into the wells and incubated for 2 hours. After conjugating the peptides, plates were washed three times with PBS (−) and blocked by the addition of 1% BSA in PBS (−) for 1 hour. The peptide chitosan matrices were used for cell attachment assays.