Distinct Developmental Functions of Prostasin (CAP1/PRSS8) Zymogen and Activated Prostasin*

The membrane-anchored serine prostasin (CAP1/PRSS8) is essential for barrier acquisition of the interfollicular epidermis and for normal hair follicle development. Consequently, prostasin null mice die shortly after birth. Prostasin is found in two forms in the epidermis: a one-chain zymogen and a two-chain proteolytically active form, generated by matriptase-dependent activation site cleavage. Here we used gene editing to generate mice expressing only activation site cleavage-resistant (zymogen-locked) endogenous prostasin. Interestingly, these mutant mice displayed normal interfollicular epidermal development and postnatal survival, but had defects in whisker and pelage hair formation. These findings identify two distinct in vivo functions of epidermal prostasin: a function in the interfollicular epidermis, not requiring activation site cleavage, that can be mediated by the zymogen-locked version of prostasin and a proteolysis-dependent function of activated prostasin in hair follicles, dependent on zymogen conversion by matriptase.

Prostasin (also known as channel-activating protease-1, CAP1, and PRSS8) is a glycosylphosphatidylinositol-anchored trypsin-like serine protease that is widely expressed in epithelial tissues, including both the interfollicular and the follicular compartments of the epidermis. Loss-of-function genetic studies in mice have uncovered critical functions of prostasin in both the formation of epidermal barrier formation and the formation of whiskers and pelage hair (1,2). Prostasin is synthesized as an inactive proform (zymogen) that is converted to a catalytically active protease by a single endoproteolytic cleavage in the conserved activation cleavage site (3)(4)(5)(6). Prostasin zymogen conversion in the epidermis requires the membrane-activated serine protease matriptase (7)(8)(9). This observation, combined with the observation that both matriptase-deficient and prostasin-deficient mice display identical epidermal phenotypes, led to the formulation of the hypothesis that prostasin exerts its functions in this tissue as part of a matriptase-prostasin cell surface zymogen activation cascade (8). Several observations made during the last decade suggest, however, that prostasin may also execute biological functions independent of its own proteolytic activity. For example, in a reconstituted Xenopus oocyte system, prostasin can activate the epithelial sodium channel (ENaC) by inducing proteolytic cleavage of its ␥ subunit to release an inhibitory domain. However, this activation, which can be inhibited by the broad-spectrum serine protease inhibitor, aprotinin, can also be efficiently executed by mutant prostasin variants that lack the catalytic histidine-aspartate-serine triad (10 -12). Likewise, catalytically inactive prostasin mutants can stimulate the activation of protease-activated receptor-2 in a reconstituted mammalian cell-based system (13). Strong support for a non-proteolytic function of prostasin in vivo has been gained from the observation that mis-expressed catalytically inactive prostasin induces severe skin pathology in transgenic mice (13,14), and in particular, by our recent demonstration that mice expressing only catalytically inactive endogenous prostasin, unlike prostasin null mice, display normal long-term survival (15).
The above findings have raised a number of unanswered mechanistic questions regarding prostasin and its functions in epidermal development. We addressed four of these questions in the present study. (a) Does prostasin need activation site cleavage to perform its epidermal functions, analogous to trypsin-like serine protease-like growth factors, such as hepatocyte growth factor and macrophage-stimulating protein (16,17)? (b) Do prostasin zymogen and activated two-chain prostasin have different functions in distinct epidermal compartments? (c) Does matriptase exert its essential epidermal functions solely through the conversion of the prostasin zymogen? (d) Does prostasin zymogen stimulate matriptase activation in the epidermis? indicate the cleavage site): 5Ј-TGCCGTCATCCAGCCacgca-TCACCGGTGGTGGCAGTG-3Ј. The linearized donor DNA and ZFN mRNA were microinjected into the male pronucleus of FVB zygotes, which were implanted into pseudopregnant mice. The offspring were screened for the point mutation using the following primers: WT forward, 5Ј-CCGTCATCCAGC-CACGC-3Ј; MUT forward, 5Ј-CCGTCATCCAGCCCCAG-3Ј; WTϩMUT reverse, 5Ј-TAGATCTGGACTGACCCATG-3Ј. All mice positive for the mutation were subsequently analyzed by direct sequencing using primers located upstream of the potential insert and downstream of the desired mutation, excluding PCR amplification of randomly inserted donor DNA. Primers were as follows: forward, 5Ј-GAGTTCTGCAAGGCTGACGTGG-3Ј; reverse, 5Ј-TAGATCTGGACTGACCCATG-3Ј. RNA Preparation and RT-PCR-Tissues were collected from newborn mice, snap-frozen in liquid nitrogen, and ground to a fine powder with mortar and pestle, and RNA was purified using the RNeasy mini kit (Qiagen, Hilden, Germany). Reverse transcription and PCR amplification were performed using a High Capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA), per the manufacturer's instructions. First strand cDNA synthesis was performed using an oligo(dT) primer. The primer pair utilized for Prss8 RT-PCR was as follows: 5Ј-TTCCGCAAGTTCACCTACC-3Ј and 5Ј-CGGGC-CGGCCATGCTTTACG-3Ј. The annealing temperature for this primer set was 60°C. Expression levels were compared with S15 mRNA levels in each sample.

Experimental Procedures
Protein Extraction from Mouse Tissue-The tissues were dissected, snap-frozen on dry ice, and stored at Ϫ80°C until homogenization. The tissues were homogenized in ice-cold lysis buffer containing 1% Triton X-100, 0.5% sodium deoxycholate in PBS plus Protease Inhibitor Cocktail (Sigma) and incubated on ice for 10 min. The lysates were centrifuged at 20,000 ϫ g for 20 min at 4°C to remove the tissue debris, and the supernatant was used for further analysis. The protein concentration was measured with a standard BCA assay (Pierce).
Western Blotting-Samples were mixed with 4ϫ LDS sample buffer (NuPAGE, Invitrogen) containing 7% ␤-mercaptoethanol and boiled for 10 min, unless otherwise indicated. The proteins were separated on 4 -12% BisTris NuPAGE gels and transferred to 0.2-m pore size PVDF membranes (Invitrogen). The membranes were blocked with 5% nonfat dry milk in TBS containing 0.05% Tween 20 (TBS-T) for 1 h at room temperature. The individual PVDF membranes were probed with primary antibodies diluted in 1% nonfat dry milk in TBS-T overnight at 4°C. Antibodies used for mouse lysates were sheep anti-human matriptase (AF3946, R&D Systems, Minneapolis, MN) and mouse anti-human prostasin (catalogue number 612173, BD Transduction Laboratories). ␣-Tubulin was used as a loading control (9099S, Cell Signaling, Beverly, MA). The next day, the membranes were washed three times for 5 min each in TBS-T and incubated for 1 h with alkaline phosphatase-conjugated secondary antibodies (Thermo Scientific). After three 5-min washes with TBS-T, the signal was developed using nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate solution (Pierce).
Immunohistochemistry-Samples from mice were fixed in 4% paraformaldehyde in PBS for 24 h, embedded into paraffin, and sectioned. Tissue sections were cleared with xylene substitute (SafeClear, Fisher Scientific), rehydrated in a graded series of alcohols, and boiled in sodium citrate buffer, pH 6, for 20 min for antigen retrieval. The sections were blocked with 2.5% BSA in PBS and incubated overnight at 4°C with mouse anti-human prostasin (1:100, BD Transduction Laboratories, catalogue number 612173) or rat-anti-mouse antibodies against Ki-67 (clone TEC-3, Dako, Carpentaria, CA). Bound antibodies were visualized using a biotin-conjugated anti-mouse secondary antibody (1:1000, Vector Laboratories, Burlingame, CA) and a VECTASTAIN ABC kit (Vector Laboratories) using 3,3-diaminobenzidine as the substrate (Sigma-Aldrich).
Transepidermal Fluid Loss Assay-Transepidermal fluid loss assay was performed exactly as described (18).
Simple Western Size Separation of Proteins-The Peggy Sue capillary electrophoresis system (Protein Simple, Wallingford, CT) was used to quantify the amount of the different matriptase forms present in skin lysates from newborn mice. Briefly, snapfrozen skin from newborn mice was lysed in T-PER lysis buffer (Pierce) with a final concentration of 0.5 mM of the protease inhibitor 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) (Sigma). The lysates were spun at 15,000 ϫ g for 15 min at 4°C, and the supernatant was used for the analysis. A final concentration of 1 g of total protein per l of sample was analyzed with the antibodies sheep anti-human matriptase (R&D Systems, catalog number 3946) and ␤-actin as a loading control (Abcam, ab6276-100) according to the manufacturer's instructions for size separation of proteins.

Results and Discussion
Mice Expressing Zymogen-locked Endogenous Prostasin-We generated mice expressing zymogen-locked prostasin by introducing c.362-363 GC3 AG dinucleotide substitution into exon 3 of the Prss8 gene by template-guided repair of a ZFN-induced double strand DNA break in FVB/NJ zygotes (Fig. 1A). This dinucleotide change generated a Prss8 allele (hereafter referred to as Prss8 zym ) that encodes a prostasin in which Arg-44 is substituted by Gln, thus rendering the mutant prostasin refractory to activation site cleavage, as previously demonstrated in vitro (13). Introduction of the dinucleotide change was verified by sequencing analysis of DNA from mice bred to homozygosity for the mutated allele (Fig. 1B). RT-PCR analysis of mRNA isolated from skin, gastrointestinal tract, lung, and kidney of Prss8 zym/zym mice showed that the introduction of the dinucleotide substitution (and the two synonymous single nucleotide substitutions; see "Experimental Procedures") did not affect transcription of the Prss8 zym allele (Fig. 1C). Western blotting analysis of protein extracts from skin, kidney, and lungs of newborn Prss8 zym/zym mice and wild-type (Prss8 ϩ/ϩ ) littermates showed that the mutant prostasin was expressed at levels similar to wild-type prostasin (Fig. 1D).
As expected, whereas prostasin from Prss8 ϩ/ϩ mice presented in these tissues as two bands, representing, respectively, the prostasin zymogen and activated prostasin (8), only the prostasin zymogen was found in tissues from Prss8 zym/zym mice (Fig. 1D, compare lanes 1-3 with lanes 4 -6). Antibody speci-ficity was demonstrated by the absence of these two immunoreactive proteins in the same tissues from Prss8 Ϫ/Ϫ mice analyzed in parallel (Fig. 1D, lane 7). Furthermore, immunohistochemical analysis of skin, kidney, and lung from newborn Prss8 zym/zym mice and Prss8 ϩ/ϩ littermates revealed a cellular location of the mutant prostasin that was indistinguishable from wild-type prostasin (Fig. 1, compare E with F, H with I,  and K with L). Again, antibody specificity was verified by paral-lel staining of identical tissues from Prss8 Ϫ/Ϫ mice (Fig. 1, G, J,  and M).
Abnormal Whisker and Pelage Hair Development in Mice Expressing Zymogen-locked Endogenous Prostasin-Although mice expressing only zymogen-locked prostasin displayed normal interfollicular epidermal development and postnatal survival, obvious defects were apparent in hair follicle development. Whiskers, which are present in wild-type mice at birth, were absent in newborn Prss8 zym/zym pups (Fig. 2F). When the whiskers erupted later in postnatal development, the whiskers were kinked and curly (Fig. 2G). Furthermore, pelage hairs were markedly sparser in adult Prss8 zym/zym mice (Fig. 2H), correlating with a reduced proliferation of hair follicle cells as compared with littermate controls (Fig. 2, I-N, compare I with J and K with L; results were quantified in M and N).
Epidermal Matriptase Zymogen Conversion Is Only Modestly Stimulated by Wild-type and Zymogen-locked Prostasin-Prostasin and matriptase were previously proposed to be part of a single epidermal proteolytic cascade in which matriptase is upstream of prostasin (7-9). Prostasin, however, is upstream of matriptase in other epithelia (6), and zymogen-locked prostasin supports matriptase auto-activation in cell-based overexpression systems (13). We, therefore, next directly examined the status of matriptase activation in the epidermis of newborn Prss8 Ϫ/Ϫ , Prss8 zym/zym , and Prss8 ϩ/ϩ pups. We separated skin protein lysates by reducing SDS-PAGE and visualized the matriptase zymogen and activated matriptase by Western blotting using an antibody directed against the serine protease domain. Importantly, activated matriptase was readily found in skin lysates from mice of all genotypes, showing that prostasin is dispensable for epidermal matriptase activation. The matriptase levels varied somewhat between different skin preparations, but we observed no consistent genotype-related difference in overall matriptase expression levels. However, an immunoassay by ProteinSimple, utilizing capillary electrophoresis, which gives a quantitative measure of the amount of both the latent and the cleaved form of matriptase, showed a small increase in the ratio of activated to latent matriptase in Prss8 ϩ/ϩ and Prss8 zym/zym skin extracts, as compared with Prss8 Ϫ/Ϫ skin extracts (Fig. 2, P and Q). This decrease, however, is unlikely to cause any of the epidermal phenotypes observed in prostasin null mice or in mice expressing zymogen-locked prostasin, as the level of activated matriptase in Prss8 Ϫ/Ϫ and Prss8 zym/zym skin extracts exceeds that observed in matriptase heterozygote (St14 ϩ/Ϫ ) mice, which display no epidermal phenotype (18,19).
The relationship between matriptase and prostasin in the epidermis was initially believed to be a simple one, with matriptase activating prostasin, and prostasin executing the epidermal functions of the cascade through the proteolytic cleavage of specific substrates of unknown identity (8). The findings in this study combined with our recent phenotypic characterization of mice expressing only catalytically inactive prostasin, however, necessitate a significant revision of this model, at least regarding interfollicular epidermal development (18,19). First, the essentially normal development of the interfollicular epidermis of mice expressing zymogen-locked or catalytically inactive prostasin shows that prostasin requires neither zymogen conversion nor catalytic activity to execute its essential functions in this epidermal compartment. It follows from this observation that matriptase must exert its essential functions in interfollicular epidermal development essentially independently of prostasin zymogen conversion, by cleavage of unidentified substrates or through a non-catalytic mechanism. Likewise the persistence of activated matriptase in the epidermis of prostasin null mice or mice expressing zymogen-locked prostasin shows that prostasin activation of matriptase is of limited importance in the epidermis, although the experimental approach used does not exclude the possibility that matriptase activation in some epidermal sub-compartments may be significantly stimulated by, or even be dependent on, prostasin.
of the co-expression of matriptase and prostasin in hair shaftforming keratinocytes of the hair follicle, and the strict requirement of matriptase for prostasin zymogen conversion, it is reasonable to assume that matriptase and prostasin form a more "conventional" proteolytic cascade in this compartment whereby matriptase activates prostasin and prostasin cleaves specific substrates to promote hair morphogenesis.
In summary, our study has demonstrated that prostasin is unique among trypsin-like serine proteases in that it has dual essential in vivo functions as a zymogen that are non-enzymatic, as well as essential proteolytic functions once it is converted to its two-chain enzymatically active conformation.