Advertisement

β1 Integrins with Individually Disrupted Cytoplasmic NPxY Motifs Are Embryonic Lethal but Partially Active in the Epidermis

      β1 Integrin adhesion is believed to require binding of talins and kindlins to the membrane proximal and distal NPxY motifs of the β1 cytoplasmic tail, respectively. To test this hypothesis, we substituted the membrane proximal and distal tyrosines (Y) of the β1 tail with alanine (A) residues (β1 Y783A; β1 Y795A) in the germline of mice. We report that β1 Y783A or β1 Y795A substitutions blocked talin or kindlin binding, respectively, and led to β1 null-like peri-implantation lethality. Expression of β1 Y783A or β1 Y795A in the epidermis, however, resulted in skin blister and hair follicle phenotypes that were considerably milder than those observed with β1 integrin gene deletion or a β1 double Y-to-A substitution (β1 YY783/795AA). In culture, defects in adhesion, spreading, and migration were more severe with the β1 Y783A than with the β1 Y795A substitution despite markedly reduced β1 Y795A integrin surface levels owing to diminished protein stability. We conclude that regulation of β1 integrin adhesion through talins and kindlins may differ substantially between stably adherent keratinocytes and cells of the developing embryo, and that β1 cytoplasmic NPxY motifs contribute individually and independent of each other to β1 function in keratinocytes.

      Abbreviations

      EB
      embryoid body
      fl
      floxed
      K5
      keratin 5
      KI
      knock-in
      LN
      laminin

      Introduction

      Integrins are adhesion receptors that bind extracellular matrix proteins and counter receptors. When integrins bind their ligand, they cluster and recruit a large number of adaptor and signaling proteins to their cytoplasmic domain to finally form cell-extracellular matrix anchoring structures called focal adhesions. Focal adhesions provide a large signaling platform, which decodes the physical and chemical qualities of the extracellular environment (
      • Legate K.R.
      • Wickström S.A.
      • Fässler R.
      Genetic and cell biological analysis of integrin outside-in signaling.
      ;
      • Moser M.
      • Legate K.R.
      • Zent R.
      • et al.
      The tail of integrins, talin, and kindlins.
      ). All integrins are composed of an α- and a β-subunit. The β1 integrin subunit is ubiquitously expressed and dimerizes with 12 integrin α-chains (
      • Meves A.
      • Stremmel C.
      • Gottschalk K.
      • et al.
      The Kindlin protein family: new members to the club of focal adhesion proteins.
      ). Not surprisingly, deletion of the β1 integrin gene in mice is embryonic lethal at peri-implantation (
      • Fassler R.
      • Meyer M.
      Consequences of lack of beta 1 integrin gene expression in mice.
      ).
      The β1 integrin cytoplasmic tail contains two key adaptor protein-binding sites: the membrane proximal W(x)4NPIY motif that binds talins and the membrane distal TT(x)2NPKY motif that binds kindlins (
      • Moser M.
      • Nieswandt B.
      • Ussar S.
      • et al.
      Kindlin-3 is essential for integrin activation and platelet aggregation.
      ;
      • Meves A.
      • Geiger T.
      • Zanivan S.
      • et al.
      [beta]1 integrin cytoplasmic tyrosines promote skin tumorigenesis independent of their phosphorylation.
      ). Binding of talins to the membrane proximal motif induces conformational changes in the extracellular portion of β1 integrin that increases its affinity for extracellular matrix (
      • Wegener K.L.
      • Partridge A.W.
      • Han J.
      • et al.
      Structural basis of integrin activation by talin.
      ;
      • Ye F.
      • Hu G.
      • Taylor D.
      • et al.
      Recreation of the terminal events in physiological integrin activation.
      ). It is therefore believed that the interaction of the β1 tail with talin preceeds integrin binding to extracellular matrix in a process referred to as integrin activation or inside-out signaling. More recently, we reported that integrin activation in platelets, leukocytes, and epithelial cells such as primitive endoderm and intestinal epithelial cells requires not only the binding of talins but also of kindlins to the membrane distal NPxY motif (
      • Moser M.
      • Nieswandt B.
      • Ussar S.
      • et al.
      Kindlin-3 is essential for integrin activation and platelet aggregation.
      ). These findings were in contrast to reports attributing no or only a minor role to kindlins in integrin inside-out signaling (
      • Ye F.
      • Hu G.
      • Taylor D.
      • et al.
      Recreation of the terminal events in physiological integrin activation.
      ).
      To directly test the role of talin and kindlin binding to the β1 integrin tail in vivo, we compared phenotypes in mice and cells carrying a Y-to-A substitution in the membrane proximal (β1 Y783A) or distal (β1 Y795A) β1 NPxY motif. Both mutations lead to peri-implantation lethality resembling the β1-null phenotype. Interestingly, however, targeted expression of single Y-to-A mutations in the epidermis leads to a phenotype that is significantly milder than the β1-null phenotype, indicating residual activity of the β1 Y783A and β1 Y795A integrins. A β1-null-like phenotype is only achieved with the simultaneous mutation of both tyrosine residues (β1 YY783/Y795AA). These findings indicate that the functional consequences of perturbed talin and kindlin binding to β1 integrin NPxY motifs differ between cell types.

      Results

      Single β1 Y-to-A mutations result in peri-implantation lethality

      Mice lacking β1 integrin expression or carrying homozygous β1 YYAA mutations die at the peri-implantation stage (
      • Fassler R.
      • Meyer M.
      Consequences of lack of beta 1 integrin gene expression in mice.
      ;
      • Chen H.
      • Zou Z.
      • Sarratt K.L.
      • et al.
      In vivo beta1 integrin function requires phosphorylation-independent regulation by cytoplasmic tyrosines.
      ;
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ). To assess the in vivo phenotype of single β1 cytoplasmic Y-to-A mutations and to analyze the role of talin and kindlin binding to the β1 integrin cytoplasmic tail, we generated mice with nonconservative tyrosine (Y) to alanine (A) mutations in the membrane proximal and distal β1 integrin NPxY motifs (β1 Y783A; β1 Y795A). The mutant knock-in (KI) alleles (KIneo+; Figure 1a) were confirmed by Southern blotting and PCR on genomic DNA derived from tail biopsies (Figure 1b and c). Deletor-Cre-mediated removal of the neomycin cassette yielded mice with heterozygous β1 Y783A or Y795A mutations, respectively. They were healthy and fertile. Progeny of heterozygous crossings yielded wild-type and heterozygous but no homozygous offspring (Figure 1d).
      Figure thumbnail gr1
      Figure 1β1 Y-to-A mutations cause a β1-null-like phenotype in embryonic stem (ES) cells. (a) Partial map of β1 wild-type (wt) and knock-in (KI) alleles before (KIneo+) and after (KIlox) neo deletion through Cre. Asterisk: site of point mutagenesis. Triangles: loxP sites. (b) Southern blotting identified homologous recombination. (c) PCR genotyping of KIlox mice using loxP site flanking primers. (d) Numbers and genotypes of offspring from heterozygous crossings. (e) Bright-field images of E7.5 embryos. (f) ES cell colonies on feeder cells. (g, h) Bright-field and immunofluorescence images of EBs on day 5 of suspension culture. (g) Arrows: endoderm cells. Arrowheads: BM. Dotted line: central cavity lining. (h) EBs were stained with antibodies against β1 integrin (red) and laminin 111 (green), and nuclei were counterstained with 4′6-diamidino-2-phenylindole (DAPI; blue). Bar=100μm. (i) Expression of integrin subunits on ES cells by fluorescence-activated cell sorting (mean±SD; n=4; *P<0.05, ***P<0.0001 vs. control). (j) Quantification of the 9EG7 epitope by FACS (mean±SD; n=4). (k) Quantification of cell adhesion (mean±SD; n=4; ***P<0.0001 vs. control). LN111, laminin 111.
      To assess the embryonic defects of mice with single β1 cytoplasmic Y-to-A mutations, we performed timed heterozygous matings and analyzed the gross morphology of embryos at embryonic day (E) 7.5. Compared with wild-type embryos, β1 Y783A and Y795A embryos were severely malformed, indicating death at peri-implantation (Figure 1e). To better define the cause of death during early development, we established ES cell cultures from homozygous β1 Y783A and Y795A blastocysts. Similar to β1-null and β1 YY783/795AA (YYAA) ES cells, single β1 Y-to-A ES cells adhered less to feeder cells (Figure 1f). Next, we generated embryoid bodies (EBs) (
      • Montanez E.
      • Piwko-Czuchra A.
      • Bauer M.
      • et al.
      Analysis of integrin functions in peri-implantation embryos, hematopoietic system, and skin.
      ). Wild-type EBs gave rise to an outer layer of endoderm, followed by the conversion of an undifferentiated core into a layer of pseudo-stratified primitive ectoderm with a central cavity on day 4–6 (Figure 1g) (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). β1 Y-to-A and β1 YYAA EBs formed compact aggregates with significant detachment of the endoderm, abnormally assembled and often discontinuous basement membrane, and absent cavity formation, which resembled the defects of β1-null EBs (Figure 1g and h) (
      • Li S.
      • Harrison D.
      • Carbonetto S.
      • et al.
      Matrix assembly, regulation, and survival functions of laminin and its receptors in embryonic stem cell differentiation.
      ).
      Fluorescence-activated cell sorting analysis revealed normal levels of β1 integrin on ES cells carrying the membrane proximal β1 Y783A mutation (Figure 1i). As expected, β1 integrin surface levels on ES cells carrying the membrane distal Y795A mutation were significantly reduced (Figure 1i). This is due to the impaired recruitment of SNX17 to the β1 Y795A tails in early and recycling endosomes, which prevents lysosomal β1 degradation (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). β1 heterodimerization partners α5 and α6 integrin were also reduced, whereas αvβ3 integrin levels were normal (Figure 1i). As expected, all β1 Y-to-A mutated integrins had reduced extracellular ligand binding activity as assessed by the reduced availability of 9EG7 epitope during fluorescence-activated cell sorting (Figure 1j) and reduced adhesion to fibronectin and laminin (LN) 111 (Figure 1k).

      Compound heterozygous β1Y783A/Y795A mice die at peri-implantation

      The availability of the β1 Y783A and Y795A mouse strains allowed the generation of compound heterozygous (β1Y783A/Y795A) animals that enable kindlin binding to the β1 Y783A and talin binding to the β1 Y795A allele and thus testing whether talin and kindlin cooperate with each other in a trans configuration (
      • Moser M.
      • Legate K.R.
      • Zent R.
      • et al.
      The tail of integrins, talin, and kindlins.
      ). Intercrosses of the β1 Y783A strain with the Y795A strain revealed no live offspring. The dissection of decidua chambers at different stages of pregnancy showed that the β1Y783A/Y795A compound mice died at the peri-implantation stage with a phenotype that resembled the homozygous β1 Y783A or β1 Y795A mutations (Figure 2a). β1Y783A/Y795A ES cells were less adhesive compared with wild-type (Figure 2b), and EBs derived from β1Y783A/Y795A ES cells showed defects similar to β1-null EBs (Figure 2c and d). β1 integrin surface levels (Figure 2e), 9EG7 epitope availability (Figure 2f), and cell adhesion to fibronectin and LN111 (Figure 2g) were reduced to a similar extent as observed with the homozygous β1 Y795A mutation.
      Figure thumbnail gr2
      Figure 2Absence of β1 integrin trans activation in compound heterozygous β1Y783A/Y795A embryos. (a) Bright-field images of E7.5 embryos. (b) Embryonic stem (ES) cell colonies on feeder cells. (c, d) Bright-field (c) and (d) immunofluorescence images of embryoid bodies (EBs) on the 5th day of suspension culture. EBs were stained with antibodies against β1 integrin (red) and laminin 111 (green), and nuclei were counterstained with 4′6-diamidino-2-phenylindole (DAPI; blue). Bar=100μm. (e) Expression of integrin subunits on ES cells determined by fluorescence-activated cell sorting (mean±SD; n=4; *P<0.05, ***P<0.0001 vs. control). (f) Integrin activation on ES cells measured by 9EG7 binding (mean±SD; n=4). (g) Quantification of cell adhesion (***P<0.0001 vs. control). LN111, laminin 111.

      Single β1 Y-to-A integrin is partially active in keratinocytes

      To test whether the β1 Y-to-A mutations also lead to β1-null-like phenotypes in other cell types, we decided to express the β1 Y783A or β1 Y795A mutations in epidermal keratinocytes. Epidermal keratinocytes were chosen because they predominantly express β1 integrin, lack β3 integrin, and express low levels of αv integrin. To achieve hemizygous targeted expression of the knock-in alleles in the epidermis, we intercrossed a floxed (fl) β1 allele with either a β1 Y-to-A allele or a non-floxed wild-type β1 allele (control) and ablated the floxed β1 gene using a keratin 5(K5) promoter-driven Cre recombinase (Figure 3a). The floxed β1 integrin allele was designed to couple genomic β1 integrin deletion to the activation of a lacZ gene (
      • Brakebusch C.
      • Grose R.
      • Quondamatteo F.
      • et al.
      Skin and hair follicle integrity is crucially dependent on beta 1 integrin expression on keratinocytes.
      ). Successful β1 integrin deletion was identified by lacZ expression in the epidermis and hair follicles of hemizygous wild-type and β1 Y-to-A skin (Figure 3b and c). Deletion of homozygous β1fl/fl integrin by K5-Cre (
      • Brakebusch C.
      • Grose R.
      • Quondamatteo F.
      • et al.
      Skin and hair follicle integrity is crucially dependent on beta 1 integrin expression on keratinocytes.
      ) or hemizygous expression of β1 YYAA in the epidermis (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ) led to a near complete loss of terminal hair in mice and reduced weight gain. In contrast to this marked phenotype we did not detect strong defects in β1 Y783A or β1 Y795A epidermis except a patchy hair loss over the dorsal midline of the skull in β1 Y783A epidermis (Figure 3d). This finding suggests that the activities of the mutant integrins are not or only marginally affected by impaired talin or kindlin binding.
      Figure thumbnail gr3
      Figure 3Single β1 Y-to-A integrin is partially active in keratinocytes. (a) Approach for achieving hemizygous β1 Y-to-A integrin expression in the epidermis. Exon structure of β1 integrin is depicted. Site of point mutation is highlighted by an asterisk. LoxP sites are shown as triangles. Recombination of β1fl/fl loxP sites by Cre results in LacZ transcription. Detection of LacZ activity by (b) immunohistochemistry or (c) flow cytometry. (d) Gross morphology and histological analysis of mice and the epidermis with single or double β1 Y-to-A mutations at postnatal day 14. Dotted lines: dermoepidermal junction. Asterisks: subepidermal split. White arrowheads: nonlinear subepidermal LN332 deposition. H&E, hematoxylin and eosin; PM, point mutation.
      To further test whether the β1 Y783A or the β1 Y795A integrins are partially active in the epidermis, we analyzed skin histologically (Figure 3d). β1-null epidermis is characterized by subepidermal blistering, patchy deposition of LN332 in the papillary dermis, impaired hemidesmosome assembly, and dermal inflammatory infiltrates that induce hyperproliferation of basal keratinocytes (
      • Brakebusch C.
      • Grose R.
      • Quondamatteo F.
      • et al.
      Skin and hair follicle integrity is crucially dependent on beta 1 integrin expression on keratinocytes.
      ). The same phenotype was seen in β1 YYAA epidermis (Figure 3d) (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ). By comparison, single β1 Y-to-A epidermis was less affected. We observed linear deposition of α6 integrin at the dermoepidermal junction, demonstrating proper localization of hemidesmosomes, and normal proliferation in the skin as indicated by the single layer of K5-expressing basal keratinocytes and normal numbers of Ki67-positive cells (wild-type: 6±3; β1 Y783A: 8±3; β1 Y795A: 8±3; β1 YYAA: 18±1 Ki67-positive cells per high-powered field). Laminin 332 deposition was focally shaggy and abnormal, but not to the extent observed in β1 YYAA epidermis. Interestingly, we failed to observe subepidermal blistering, indicating the ability of basal keratinocytes to firmly adhere to the underlying basement membrane. As expected from the focal hair loss in vivo, we saw impaired hair follicle morphogenesis in β1 Y783A but not in β1 Y795A skin; enlarged hair follicle stumps were seen intermixed with normally shaped thin hair follicles extending deep into the subcutaneous fat. In β1 YYAA and in β1-null epidermis, all hair follicles were enlarged and abnormal. These findings suggest that keratinocytes expressing either the β1 Y783A or the β1 Y795A substitution must possess residual β1 activity.

      Distinct consequences of β1 Y783A and β1 Y795A mutations on β1 function in keratinocytes

      The subtle defects of β1 Y783A or the β1 Y795A epidermis in vivo prompted us to analyze β1 integrin functions ex vivo. In line with findings from mutant ES cells, flow cytometry revealed normal expression of β1 Y783A and markedly reduced levels of β1 Y795A and β1 YYAA on freshly isolated primary keratinocytes (Figure 4a). Despite the normal β1 Y783A levels, keratinocytes exhibited defects in adhesion (Figure 4b) and spreading that were more pronounced compared with β1 Y795A keratinocytes (Figure 4c). β1 YYAA keratinocytes did not adhere (Figure 4b) (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ). β1 Y783A keratinocytes did not grow into confluent monolayers but instead became large and flat and rapidly initiated terminal differentiation (Figure 4b). In contrast, β1 Y795A keratinocytes underwent normal spreading, proliferated, and consistently formed confluent monolayers in culture. Scratch-wounding assays demonstrated that β1 Y795A keratinocytes were able to close the in vitro wound, although slower than wild type (Figure 4d and e). Single-cell tracking revealed that β1 Y795A keratinocytes migrated directionally (Figure 4f) but with reduced speed when compared with wild type (Figure 4g). As β1 Y783A keratinocytes were unable to form monolayers in vitro, we were unable to perform scratch assays.
      Figure thumbnail gr4
      Figure 4Adhesion and migration of β1 Y-to-A keratinocytes. (a) Epidermis-derived single-cell suspensions of normal and β1 Y-to-A epidermis were assayed for β1 integrin surface expression (mean±SD; n=4; ***P<0.0001 vs. control). (b) Adhesion of normal and β1 Y-to-A keratinocytes to fibronectin- and collagen-coated tissue culture plastic over the indicated time period. (c) Spreading of normal and β1 Y-to-A keratinocytes on fibronectin- and collagen-coated tissue culture plastic (mean±SD; n=3; *P<0.05 vs. control). (d) Scratch wounding of a confluent monolayer of normal and β1 Y795A keratinocytes. (e) Quantification of cell front migration into a scratch wound (mean±SD; n=3; *P<0.05, ***P<0.0001 vs. control). (f) Assessment of migratory directionality. (g) Quantification of keratinocyte migratory velocity and accumulated distance (mean±SD; n=3; *P<0.05 vs. control).

      Distinct consequences of β1 Y783A and β1 Y795A mutations on β1 function in fibroblasts

      β1 Y783A and β1 YYAA keratinocytes poorly adhered and did not grow in culture owing to terminal differentiation. To further characterize the cellular phenotype of single and double β1 Y-to-A mutations, we decided to reconstitute β1-null fibroblasts with wild-type, β1 Y783A, β1 Y795A, or β1 YYAA integrins by retroviral cDNA transduction (Figure 5a) (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). In contrast to keratinocytes, fibroblasts express β3 integrin, which partially compensates for the loss in β1 integrin-mediated adhesion. Fluorescence-activated cell sorting analysis showed reduced β1 integrin surface levels in β1 Y795A cells, whereas the levels of other β1 Y-to-A variants were comparable to wild-type β1 integrin surface levels (Figure 5b). Similar to primary keratinocytes, β1 Y795A and wild-type fibroblasts were well spread, whereas β1 Y783A and β1 YYAA fibroblasts resembled the more rounded β1-null cells (Figure 5c). Interestingly, the Y783A mutation significantly impaired internalization of cell surface β1 integrin (Figure 5d), whereas β1 integrin stability was only reduced in Y795A and YYAA cells (Figure 5e). In single-cell migration assays on fibronectin, all β1 Y-to-A mutations exhibited reduced cell velocity (Figure 5f and g). However, the velocity defect was least pronounced with the membrane distal β1 Y795A mutation.
      Figure thumbnail gr5
      Figure 5Effect of β1 Y-to-A mutations on β1 integrin stability, internalization, and migration. (a) Scheme depicting the generation of wild-type (wt) and β1 Y-to-A expressing fibroblasts. (b) β1 Integrin surface levels as determined by fluorescence-activated cell sorting (mean±SD; n=3; *P<0.05). (c) Phase-contrast images of β1 integrin–reconstituted fibroblasts. (d) Quantification of cell surface β1 integrin internalization (mean±s.e.m.; n=5; **P<0.005; ***P<0.001 vs. control). NS, not significant. (e) Quantification of cell surface β1 integrin stability (mean±SD; n=3). Quantification of (f) fibroblast velocity and (g) accumulated distance (mean±SD; n=30 cells from three movies; ***P<0.0001 vs. control).
      In summary, the phenotypes of the β1 Y-to-A mutations were similar in fibroblast and keratinocytes. Specifically, β1 Y783A mutations had a greater impact on spreading and migration than β1 Y795A mutations. The latter, however, significantly impaired integrin half-life.

      Talin and kindlin recruitment is affected by the β1 cytoplasmic Y-to-A mutations

      Subtle defects of β1 Y783A or β1 Y795A epidermis suggested that residual talin and kindlin binding to β1 integrin tails occurs in keratinocytes but not in ES cells. Therefore, we tested to which degree protein binding to β1 integrin tails is retained in the presence of single (β1 Y783A, β1 Y795A) or double (β1 YYAA) β1 Y-to-A mutations in keratinocytes versus ES cells. Pull-down experiments were performed with synthesized peptides corresponding to wild-type and β1 Y-to-A mutated β1 integrin cytoplasmic tails. Bait was provided as SILAC-labeled ES cells or keratinocyte protein lysate. Quantification of precipitated protein was by mass spectrometry (
      • Meves A.
      • Geiger T.
      • Zanivan S.
      • et al.
      [beta]1 integrin cytoplasmic tyrosines promote skin tumorigenesis independent of their phosphorylation.
      ) (Figure 6a). We found that β1 YYAA peptide completely lost the ability to recruit proteins from both keratinocyte and ES cell lysates when compared with a scrambled peptide sequence (Figure 6b–e). Specifically, binding of talins and kindlins was completely abolished in keratinocytes (Figure 6c) and ES cells (Figure 6e). Selective disruption of single NPxY motifs by β1 Y-to-A substitutions prevented binding of a specific subset of proteins, most notably talins or kindlins to the membrane proximal or distal NPxY motifs, respectively (Figure 6f and g). This observation was made with keratinocyte (
      • Mathew S.
      • Lu Z.
      • Palamuttam R.J.
      • et al.
      β1 integrin NPXY motifs regulate kidney collecting-duct development and maintenance by induced-fit interactions with cytosolic proteins.
      ) and ES cell lysates.
      Figure thumbnail gr6
      Figure 6β1 Integrin cytoplasmic Y-to-A mutations impair protein recruitment in keratinocytes and embryonic stem (ES) cells. (a) Amino-acid sequence of synthesized β1 integrin peptides. Loss in protein binding vs. wild-type peptide in keratinocytes using (b) scrambled and (c) β1 YYAA peptides. Loss in protein binding vs. wild-type peptide in ES cells using (d) scrambled and (e) β1 YYAA peptides. Scatter plots highlight kindlins (red circles) and talins (green circles). Quantification of talin and kindlin binding to β1 Y-to-A peptides relative to wild-type using (f) keratinocyte or (g) ES cell lysates.

      Discussion

      Single disrupted β1 cytoplasmic NPxY motifs have been predicted to prevent talin and kindlin binding to β1 tails and therefore to result in β1 integrin loss of function (
      • Moser M.
      • Legate K.R.
      • Zent R.
      • et al.
      The tail of integrins, talin, and kindlins.
      ). Consistent with established models of integrin activation (
      • Ye F.
      • Kim C.
      • Ginsberg M.H.
      Reconstruction of integrin activation.
      ), we found that both single β1 Y-to-A mutations were early embryonic lethal and resulted in β1-null-like in vivo phenotypes and ex vivo EB defects. The same observations were made for compound heterozygous β1Y783A/Y795A mice, supporting a role for both adaptor proteins on the same β1 integrin tail and thus lack of β1 integrin trans activation through talins and kindlins (
      • Moser M.
      • Legate K.R.
      • Zent R.
      • et al.
      The tail of integrins, talin, and kindlins.
      ). Unexpectedly, the epidermis exhibited a β1-null phenotype only when both β1 NPxY motifs were disrupted, whereas single Y-to-A mutations resulted in subtle defects. Mice harboring single β1 Y-to-A mutations in either the talin or kindlin binding sites retained a fully attached and normally differentiated epidermis with minor abnormalities in basement membrane structure, indicating only minor impairment of β1 integrin function.
      Talin binding has long been viewed as the “on-switch” for integrins by inducing a high-affinity state for extracellular ligand (
      • Tadokoro S.
      • Shattil S.J.
      • Eto K.
      • et al.
      Talin binding to integrin beta tails: a final common step in integrin activation.
      ;
      • Ye F.
      • Hu G.
      • Taylor D.
      • et al.
      Recreation of the terminal events in physiological integrin activation.
      ). The absence of this interaction was thought to yield integrins unable to adhere and signal. Our study, however, shows that models of β1 integrin regulation that predict activity based on talin recruitment are oversimplified. It becomes increasingly clear that the regulation of integrin activity is multi-facetted. In addition to talins, kindlins are required for affinity regulation in cells that need rapid integrin adhesion (
      • Montanez E.
      • Ussar S.
      • Schifferer M.
      • et al.
      Kindlin-2 controls bidirectional signaling of integrins.
      ;
      • Moser M.
      • Nieswandt B.
      • Ussar S.
      • et al.
      Kindlin-3 is essential for integrin activation and platelet aggregation.
      ;
      • Ussar S.
      • Moser M.
      • Widmaier M.
      • et al.
      Loss of Kindlin-1 causes skin atrophy and lethal neonatal intestinal epithelial dysfunction.
      ). Other mechanisms such as mechanical force (
      • Shi Q.
      • Boettiger D.
      A novel mode for integrin-mediated signaling: tethering is required for phosphorylation of FAK Y397.
      ;
      • Friedland J.C.
      • Lee M.H.
      • Boettiger D.
      Mechanically activated integrin switch controls alpha5beta1 function.
      ;
      • Schiller H.B.
      • Friedel C.C.
      • Boulegue C.
      • et al.
      Quantitative proteomics of the integrin adhesome show a myosin II-dependent recruitment of LIM domain proteins.
      ), clustering (
      • Roca-Cusachs P.
      • Gauthier N.C.
      • Del Rio A.
      • et al.
      Clustering of α5β1 integrins determines adhesion strength whereas αvβ3 and talin enable mechanotransduction.
      ), chemical modifications (
      • Isaji T.
      • Kariya Y.
      • Xu Q.
      • et al.
      Functional roles of the bisecting GlcNAc in integrin-mediated cell adhesion.
      ), compartmentalization (
      • Larkin D.
      • Treumann A.
      • Murphy D.
      • et al.
      Compartmentalization regulates the interaction between the platelet integrin αIIbβ3 and ICln.
      ), trafficking (
      • Margadant C.
      • Monsuur H.N.
      • Norman J.C.
      • et al.
      Mechanisms of integrin activation and trafficking.
      ), and stability (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ) exert important additional spatiotemporal control of integrin function. It appears likely that the latter become particularly important to cells that do not require rapid regulation of adhesion. In contrast to platelets or leukocytes, basal keratinocytes are stably exposed to ligand of the basement membrane, and integrin-extracellular matrix bonds may form and dissolve continuously without affinity modulation by adaptor proteins (
      • Boettiger D.
      Mechanical control of integrin-mediated adhesion and signaling.
      ). Stable adhesions may form primarily as a consequence of adhesion strengthening, a process that involves mechanical force (
      • Schiller H.B.
      • Friedel C.C.
      • Boulegue C.
      • et al.
      Quantitative proteomics of the integrin adhesome show a myosin II-dependent recruitment of LIM domain proteins.
      ) or clustering of integrins (
      • Paszek M.J.
      • Boettiger D.
      • Weaver V.M.
      • et al.
      Integrin clustering is driven by mechanical resistance from the glycocalyx and the substrate.
      ). These processes also require adaptor protein binding to β1 integrin NPxY motifs (
      • Zhang X.
      • Jiang G.
      • Cai Y.
      • et al.
      Talin depletion reveals independence of initial cell spreading from integrin activation and traction.
      ;
      • Feng C.
      • Li Y.F.
      • Yau Y.H.
      • et al.
      Kindlin-3 mediates integrin αLβ2 outside-in signaling, and it interacts with scaffold protein receptor for activated-C kinase 1 (RACK1).
      ). However, in contrast to affinity modulation, disruption of either β1 NPxY is insufficient to block their roles completely.
      Although effective affinity modulation requires binding of talins and kindlins, e.g., in platelets (
      • Moser M.
      • Nieswandt B.
      • Ussar S.
      • et al.
      Kindlin-3 is essential for integrin activation and platelet aggregation.
      ) or ES cells (
      • Montanez E.
      • Ussar S.
      • Schifferer M.
      • et al.
      Kindlin-2 controls bidirectional signaling of integrins.
      ), stabilization of the β1 integrin–ligand interaction in cells that continuously adhere to the extracellular matrix such as keratinocytes occurs—at least to a certain extent—in the absence of direct talin or kindlin binding. Although both β1 NPxY motifs have unique properties such as the regulation of β1 protein stability through a distal NPxY-SNX17 interaction (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ;
      • Margadant C.
      • Kreft M.
      • de Groot D.J.
      • et al.
      Distant roles of talin and kindlin in regulating integrin α5β1 function and trafficking.
      ), each motif couples to the actin cytoskeleton through its respective focal adhesion–based adaptor protein talin or kindlin, and thus may be used independently to strengthen the integrin–ligand interaction through catch bonds, i.e., integrin–ligand bonds that increase in stability and lifetime through acto-myosin-mediated pulling (
      • Boettiger D.
      Mechanical control of integrin-mediated adhesion and signaling.
      ). Hence, deletion of a single NPxY motif would not eliminate but only reduce the ability of a cell to regulate integrin adhesion through force. In contrast, deletion of both NPxY motifs would eliminate all β1 integrin adhesion, as was found in this study.
      In summary, β1 integrin adhesion in the absence of either talin or kindlin binding is possible but significantly delayed and functionally compromised. Although the epidermis is well equipped to compensate for some of the consequences of deficient β1 integrin function, e.g., through the hemidesmosomal α6β4 integrin, other types of adhesion receptors such as dystroglycans, or other yet unknown cell type–specific adaptor proteins, this may not be the case for the developing embryo, which depends on the mechanical and signaling support of β1 integrin, as well as rapid affinity regulation.

      Materials and methods

      Generation of mice

      Mice with β1 YYAA and β1 Y795A mutations were reported previously (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ;
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). Mouse genomic DNA used to generate the targeting vector for the β1 Y783A knock-in and the targeting strategy were as described (
      • Fassler R.
      • Meyer M.
      Consequences of lack of beta 1 integrin gene expression in mice.
      ;
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ). Compound heterozygous β1Y783A/Y795A ES cells and embryos were generated by inter-crossing heterozygous β1 Y783A with β1 Y795A animals. All animal studies were approved by the Regierung von Oberbayern.

      Flow cytometry

      Flow cytometry of freshly isolated keratinocytes and ES cells was as previously described (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ;
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). The following antibodies were used: β1 integrin PE (102207, HMB1-1, BioLegend, San Diego, CA; 1:400), β1 integrin 9EG7 (550531, 9EG7, BD Pharmingen, San Jose, CA; 1:100), β3 integrin PE (12-0611, 2C9.G3, eBioscience, San Diego, CA; 1:400), α5-integrin PE (557447, 5H10-27, BD Pharmingen; 1:400), α6-integrin PE (555736, GoH3, BD Pharmingen), and αv-integrin PE (551187, RMV-7, BD; 1:400). Dilutions were as previously described (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ).

      Immunohistochemistry

      The following antibodies were used for immunohistochemistry: β1 integrin (MAB1997, Millipore, Billerica, MA), LN111 (ab30320, Abcam, Cambridge, MA), LN332 (a kind gift from M. Aumailley, University of Cologne, Cologne, Germany), α6 integrin–FITC (BD Biosciences, San Jose, CA), and keratin 5 (ab24647, Abcam). β-Galactosidase activity was determined as previously described (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ).

      SILAC-based peptide pull-downs

      Pull-downs were performed as previously described (
      • Meves A.
      • Geiger T.
      • Zanivan S.
      • et al.
      [beta]1 integrin cytoplasmic tyrosines promote skin tumorigenesis independent of their phosphorylation.
      ).

      ES cells and EBs

      ES cells were isolated and cultured as previously described (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). EBs were generated as described previously (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ).

      Isolation of primary keratinocytes, adhesion, spreading, and migration assays

      Primary keratinocytes were isolated from P21 mice and grown to confluence as previously described (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ;
      • Meves A.
      • Geiger T.
      • Zanivan S.
      • et al.
      [beta]1 integrin cytoplasmic tyrosines promote skin tumorigenesis independent of their phosphorylation.
      ). Adhesion of ES cells and primary keratinocytes to fibronectin, LN, and collagen I and cell spreading were performed and analyzed as described previously (
      • Czuchra A.
      • Meyer H.
      • Legate K.R.
      • et al.
      Genetic analysis of beta1 integrin “activation motifs” in mice.
      ;
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ). Cell wounding was performed as previously described (
      • Lorenz K.
      • Grashoff C.
      • Torka R.
      • et al.
      Integrin-linked kinase is required for epidermal and hair follicle morphogenesis.
      ). Live-cell recordings were performed immediately after wounding for 12hours at 37°C and 5% CO2. At least four independent scratch-wound experiments were used for calculations. Single-cell tracking of cells within the leading edge was performed using the MetaMorph software (Molecular Devices, Sunnyvale, CA), choosing 15 cells each in at least three independent experiments.

      β1 Y-to-A fibroblast lines

      Point mutations of β1 integrin cDNA (β1 Y783A, β1 Y795A, β1 YY783/795AA) were introduced by site-directed mutagenesis. For stable expression in fibroblasts, β1 integrin cDNA was cloned into the retroviral expression vectors pCLMFG or pLZRS. Viral particles were concentrated from cell culture supernatant as described previously (
      • Pfeifer A.
      • Kessler T.
      • Silletti S.
      • et al.
      Suppression of angiogenesis by lentiviral delivery of PEX, a noncatalytic fragment of matrix metalloproteinase 2.
      ) and used for infection.

      Integrin stability and internalization

      This was determined as previously described (
      • Böttcher R.T.
      • Stremmel C.
      • Meves A.
      • et al.
      Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
      ).

      ACKNOWLEDGMENTS

      This work was funded by the Max Planck Society, the DFG (SFB 914), and the Mayo Clinic.

      REFERENCES

        • Boettiger D.
        Mechanical control of integrin-mediated adhesion and signaling.
        Curr Opin Cell Biol. 2012; 24: 592-599
        • Brakebusch C.
        • Grose R.
        • Quondamatteo F.
        • et al.
        Skin and hair follicle integrity is crucially dependent on beta 1 integrin expression on keratinocytes.
        EMBO J. 2000; 19: 3990-4003
        • Böttcher R.T.
        • Stremmel C.
        • Meves A.
        • et al.
        Sorting nexin 17 prevents lysosomal degradation of [beta] 1 integrins by binding to the [beta] 1-integrin tail.
        Nat Cell Biol. 2012; 14: 584-592
        • Chen H.
        • Zou Z.
        • Sarratt K.L.
        • et al.
        In vivo beta1 integrin function requires phosphorylation-independent regulation by cytoplasmic tyrosines.
        Genes Dev. 2006; 20: 927-932
        • Czuchra A.
        • Meyer H.
        • Legate K.R.
        • et al.
        Genetic analysis of beta1 integrin “activation motifs” in mice.
        J Cell Biol. 2006; 174: 889-899
        • Fassler R.
        • Meyer M.
        Consequences of lack of beta 1 integrin gene expression in mice.
        Genes Dev. 1995; 9: 1896-1908
        • Feng C.
        • Li Y.F.
        • Yau Y.H.
        • et al.
        Kindlin-3 mediates integrin αLβ2 outside-in signaling, and it interacts with scaffold protein receptor for activated-C kinase 1 (RACK1).
        J Biol Chem. 2012; 287: 10714
        • Friedland J.C.
        • Lee M.H.
        • Boettiger D.
        Mechanically activated integrin switch controls alpha5beta1 function.
        Science. 2009; 323: 642-644
        • Isaji T.
        • Kariya Y.
        • Xu Q.
        • et al.
        Functional roles of the bisecting GlcNAc in integrin-mediated cell adhesion.
        Methods Enzymol. 2010; 480: 445-459
        • Larkin D.
        • Treumann A.
        • Murphy D.
        • et al.
        Compartmentalization regulates the interaction between the platelet integrin αIIbβ3 and ICln.
        Br J Haematol. 2009; 144: 580-590
        • Legate K.R.
        • Wickström S.A.
        • Fässler R.
        Genetic and cell biological analysis of integrin outside-in signaling.
        Genes Dev. 2009; 23: 397-418
        • Li S.
        • Harrison D.
        • Carbonetto S.
        • et al.
        Matrix assembly, regulation, and survival functions of laminin and its receptors in embryonic stem cell differentiation.
        J Cell Biol. 2002; 157: 1279-1290
        • Lorenz K.
        • Grashoff C.
        • Torka R.
        • et al.
        Integrin-linked kinase is required for epidermal and hair follicle morphogenesis.
        J Cell Biol. 2007; 177: 501-513
        • Margadant C.
        • Kreft M.
        • de Groot D.J.
        • et al.
        Distant roles of talin and kindlin in regulating integrin α5β1 function and trafficking.
        Curr Biol. 2012; 22: 1554-1563
        • Margadant C.
        • Monsuur H.N.
        • Norman J.C.
        • et al.
        Mechanisms of integrin activation and trafficking.
        Curr Opin Cell Biol. 2011; 23: 607-614
        • Mathew S.
        • Lu Z.
        • Palamuttam R.J.
        • et al.
        β1 integrin NPXY motifs regulate kidney collecting-duct development and maintenance by induced-fit interactions with cytosolic proteins.
        Mol Cell Biol. 2012; 32: 4080-4091
        • Meves A.
        • Geiger T.
        • Zanivan S.
        • et al.
        [beta]1 integrin cytoplasmic tyrosines promote skin tumorigenesis independent of their phosphorylation.
        Proc Natl Acad Sci USA. 2011; 108: 15213-15218
        • Meves A.
        • Stremmel C.
        • Gottschalk K.
        • et al.
        The Kindlin protein family: new members to the club of focal adhesion proteins.
        Trends Cell Biol. 2009; 19: 504-513
        • Montanez E.
        • Piwko-Czuchra A.
        • Bauer M.
        • et al.
        Analysis of integrin functions in peri-implantation embryos, hematopoietic system, and skin.
        Methods Enzymol. 2007; 426: 239-289
        • Montanez E.
        • Ussar S.
        • Schifferer M.
        • et al.
        Kindlin-2 controls bidirectional signaling of integrins.
        Genes Dev. 2008; 22: 1325-1330
        • Moser M.
        • Legate K.R.
        • Zent R.
        • et al.
        The tail of integrins, talin, and kindlins.
        Science. 2009; 324: 895-899
        • Moser M.
        • Nieswandt B.
        • Ussar S.
        • et al.
        Kindlin-3 is essential for integrin activation and platelet aggregation.
        Nat Med. 2008; 14: 325-330
        • Paszek M.J.
        • Boettiger D.
        • Weaver V.M.
        • et al.
        Integrin clustering is driven by mechanical resistance from the glycocalyx and the substrate.
        PLoS Comput Biol. 2009; 5: e1000604
        • Pfeifer A.
        • Kessler T.
        • Silletti S.
        • et al.
        Suppression of angiogenesis by lentiviral delivery of PEX, a noncatalytic fragment of matrix metalloproteinase 2.
        Proc Natl Acad of Sci USA. 2000; 97: 12227-12232
        • Roca-Cusachs P.
        • Gauthier N.C.
        • Del Rio A.
        • et al.
        Clustering of α5β1 integrins determines adhesion strength whereas αvβ3 and talin enable mechanotransduction.
        Proc Natl Acad of Sci USA. 2009; 106: 16245-16250
        • Schiller H.B.
        • Friedel C.C.
        • Boulegue C.
        • et al.
        Quantitative proteomics of the integrin adhesome show a myosin II-dependent recruitment of LIM domain proteins.
        EMBO Rep. 2011; 12: 259-266
        • Shi Q.
        • Boettiger D.
        A novel mode for integrin-mediated signaling: tethering is required for phosphorylation of FAK Y397.
        Mol Biol Cell. 2003; 14: 4306-4315
        • Tadokoro S.
        • Shattil S.J.
        • Eto K.
        • et al.
        Talin binding to integrin beta tails: a final common step in integrin activation.
        Science. 2003; 302: 103-106
        • Ussar S.
        • Moser M.
        • Widmaier M.
        • et al.
        Loss of Kindlin-1 causes skin atrophy and lethal neonatal intestinal epithelial dysfunction.
        PLoS Genet. 2008; 4: e1000289
        • Wegener K.L.
        • Partridge A.W.
        • Han J.
        • et al.
        Structural basis of integrin activation by talin.
        Cell. 2007; 128: 171-182
        • Ye F.
        • Hu G.
        • Taylor D.
        • et al.
        Recreation of the terminal events in physiological integrin activation.
        J Cell Biol. 2010; 188: 157-173
        • Ye F.
        • Kim C.
        • Ginsberg M.H.
        Reconstruction of integrin activation.
        Blood. 2012; 119: 26-33
        • Zhang X.
        • Jiang G.
        • Cai Y.
        • et al.
        Talin depletion reveals independence of initial cell spreading from integrin activation and traction.
        Nat Cell Biol. 2008; 10: 1062-1068