A balance between cell shedding at the skin surface and keratinocyte proliferation in the basal layer of the epidermis is necessary for the maintenance of a functional normal stratum corneum (SC) assuring an efficient physiochemical barrier for our organism against a hostile environment. The desquamation process is a precisely controlled cascade of events, the molecular mechanisms of which are only known in part and have to be put correctly in place to resolve the puzzle. It is now generally accepted that SC cohesion is essentially mediated through desmosome derived structures, called corneosomes or corneodesmosomes (Serre et al, 1991;Walsh and Chapman, 1991). The entry of corneocytes into the stratum disjunctum is characterized by a dramatic loss of nonperipheral corneodesmosomes, resulting in an intercorneocyte adhesion restricted to the overlapping edges. The selective degradation of the nonperipheral corneodesmosomes and the preservation of the edge structures might indicate that, at this stage, the lateral structures are protected against proteolysis, possibly by distinct carbohydrate entities (Walsh and Chapman, 1991).
The key event that eventually results in desquamation (cell shedding) is the complete proteolysis of all corneodesmosomes. Among the corneodesmosomal proteins, desmoglein I, a member of the cadherin adhesion molecules family, is believed to be a major player, even though its degradation remains difficult to identify (Rawlings et al, 1995). Other corneodesmosomal components like desmocollins (King et al, 1993), desmoplakins (Bernard et al, unpublished results), and corneodesmosin (Serre et al, 1991) are progressively degraded during desquamation (Lundström et al, 1994). Corneodesmosin has recently been demonstrated to be the preferred substrate of two SC serine proteases (Simon et al, 2001).
Walsh and Chapman (1991) described desquamation as a sequential action of glycosidases and proteases on corneodesmosomes. They demonstrated that (i) the addition of exogenous proteases induced no major changes in corneodesmosome structure and composition, (ii) the addition of exogenous glycosidases alone, or added after proteases treatment, caused only a mild fragmentation of the adhesive structures, and (iii) only glycosidase treatment and the subsequent addition of exogenous proteases caused dramatic degradation of the corneodesmosomal structures. These findings indicate that sugar moieties are present, protecting the corneodesmosomal protein structures from proteolysis. Walsh and Chapman further demonstrated that endoglycosidases were more effective than exoglycosidases. We recently succeeded in finding evidence for the first time of the existence of endoglycosidase activity in the human SC (Bernard et al, 2001).
Based on zymographic analysis, several laboratories have described the presence of proteases in the intercellular regions of the SC: (i) a chymotrypsin-like protease named stratum corneum chymotryptic enzyme (Egelrud and Lundström, 1991); (ii) a trypsin-like protease named stratum corneum tryptic enzyme (Egelrud and Lundström, 1990;Suzuki et al, 1994;Ekholm et al, 2000); (iii) a protease of the cysteine protease family named stratum corneum thiol protease (SCTP) (Watkinson, 1999); and (iv) the aspartic protease cathepsin D (Horikoshi et al, 1998;1999). The purpose of this study was to investigate if proteases other than the above are present in human SC and might be a part of the desquamation puzzle.
Materials and Methods
Materials
Pre-made sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) homogeneous or gradient gels, RainbowTM molecular weight markers, a silver staining kit, a gel filtration G75 HR10/30 column, a MonoS HR5/5 column, and a chromatography molecular weight standard protein kit were from Amersham Pharmacia Biotech (Uppsala, Sweden). Taq polymerase and second antibody antirabbit coupled to peroxidase were from Sigma (St. Louis, MO). The RNA preparation kit RNeasy was obtained from Qiagen (Hilden, Germany). Oligonucleotides were synthesized by Genset (Evry, France). Enzchek protease assay kit "green fluorescence" and Z-(Phe-Arg)2R110 (ZPA) substrate were from Molecular Probes (Eugene, OR). Mouse monoclonal antibody G36.19 against corneodesmosin was a gift from G. Serre (Toulouse, France). Sheep antihuman cathepsin L antibody cat. no. 1911-0507 was from Biogenesis (Poole, U.K.), human cathepsin L cat. no. 194973 from ICN (Orsay, France), and antirabbit biotinylated secondary antibody from Roche Diagnostics (Meylan, France). Vivaspin™ and Sartocon micro™ ultrafiltration filters were from Sartorius (Göttingen, Germany). Gelstar™ was from BMA (Rockland, ME).
Extraction of SC proteases
SC proteins were collected by the so-called scraping method. The scraping was carried out on the anterior surface of the leg. Before scraping, the area was washed with 200 ml of a buffer composed of 50 mM sodium phosphate buffer, pH 7, 5 mM ethylenediamine tetraacetic acid (EDTA), 150 mM NaCl, and 0.1% Tween 20. Thereafter, the washed area was scraped with the edge of a microscope slide and continuously rinsed with the above buffer. The buffer containing the corneocytes was collected in a container placed underneath the leg and passed first over a Whatman No. 4 paper, thereafter over a Millipore 0.45 
m filter, and finally over a Millipore 0.22 m filter. The resulting SC extract was concentrated to 6 ml by tangential ultrafiltration on a Sartocon micro™ ultra filter with a cut-off of 10 kDa at 4°C, applying a counter pressure of 1 bar and a filtrate exit flow rate of 2 ml per min using the K.BL apparatus (Sartorius). Protein content was evaluated with the Bio-Rad protein assay, using bovine serum albumin as a standard. The average protein concentration of the extract was about 0.5 mg per ml, i.e., a total of 3 mg per volunteer. Each experiment was performed with a pooled extract from at least three volunteers. Informed consent was obtained from all participants and the non-evasive sampling technique didn't require the approval of an ethical committee.
Gel filtration
Purification was performed at 4°C on a Biologic™ chromatography station (Bio-Rad, Marnes la Coquette, France). 250 l of the extract were directly loaded on a G75 HR10/30 column, previously equilibrated with the extraction buffer (flow rate of 0.5 ml per min), where Tween 20 was substituted with the low-ultraviolet-absorbing form of Triton X100 (buffer GF). Protein elution was monitored by recording the OD280nm and 150 l fractions were collected, 14 min after injection of the sample, into 96-well plates ready for the analyses of enzymatic activities or electrophoretic analysis. The apparent molecular weights were calculated with the calibration curve obtained after injection of a mixture of standard proteins from the low molecular weight protein kit (Amersham Pharmacia Biotech).
Protease assays and zymograms
To detect the caseinolytic activities, we used the Enzchek kit from Molecular Probes, applying the procedure proposed by the manufacturer, maintaining a 10 l/200 l sample to substrate buffer ratio. Incubations were performed at 37°C for 18 h. pH studies were performed with the following buffers: buffer A (acetate buffer 0.1 M, pH 4.0–5.75) or buffer B (phosphate buffer 0.1 M, pH 5.75–8.25) complemented with 5 mM EDTA and 0.1% Triton X100 each.
Protease activities in the first 48 fractions obtained after gel filtration were measured with buffers A and B at pH intervals of 0.25. In parallel the same experiment was performed using the buffer containing 5 mM cysteine. To exclude a possible buffer effect on the fluorescence detection, a standard curve with the fluorescent reaction product Bodipy-FL (BPFL) was run with each buffer. Activities are expressed in fmoles released BPFL per min per l sample. Detection of fluorescence was performed on a Biolumin™ fluorescent plate reader from Bio-Rad (exc. 485 nm/em. 535 nm). Cathepsin-L-like protease activities were detected using the substrate ZPA previously described byAssfalg-Machleidt et al (1992) at a final substrate concentration of 10 M after a 3 h incubation time with a 10 l/200 l sample to substrate buffer ratio. The inhibition profiles of the caseinolytic activities, detected at pH 5.5, were studied using the following inhibitors: 1.4 M E64, 1 M leupeptine, 2.5 M chymostatine, 1 M CA074, 1 M pepstatine, and 6 M aprotinine.
The gelatin zymography was performed as described byWatkinson (1999), except that the acrylamide concentration used was 12.5% instead of 11%.
Cathepsin-L-like zymography was developed in our laboratory. After horizontal SDS-PAGE separation on prefabricated gelbond PAG film-supported gel (12.5% acrylamide homogeneous gels or gradient G8-18 gels) of the samples diluted 1: 2 in Laemmli buffer without dithiothreitol (DTT), a re-naturation step for 30 min at room temperature in buffer A containing 2.5% Triton X100 was introduced. Thereafter the gel was soaked for 1 min in buffer A with 5 mM cysteine and 10 M ZPA. Detection of the fluorescence was performed on a Fluorimager™ laser scanner (Amersham Pharmacia Biotech).
Corneodesmosin degradation
Fraction no. 40, obtained after gel filtration chromatography and exhibiting the highest caseinolytic cysteine-dependent activity, was dialyzed against 0.1 M acetate buffer, pH 5. The dialyzed fraction (100 l), adjusted to 5 mM cysteine, was added to 1 mg dry weight of acetonic SC powder prepared as previously described (Méhul et al, 2000). Two more assays were run in parallel, one with 1 M leupeptine and another with 1.4 M E64. The control did not contain proteases. After 24 h at 37°C, the acetonic powder was centrifuged at 10,000
g for 10 min and the resulting pellet was extracted with 100 l of boiling Laemmli buffer containing 200 mM DTT. After centrifugation (10,000g for 10 min), the supernatant was recovered and assayed for protein content using the Bio-Rad protein assay. Proteins were adjusted to 1 mg per ml before Western blot analysis. Mouse monoclonal antibody G36.19, directed against corneodesmosin (Serre et al, 1991;Simon et al, 1997) was used after 1: 12,500 dilution in 25 mM Tris-HCl, pH 7.2, 150 mM NaCl, 0.05% (vol/vol) Tween 20 containing 0.5% nonfat milk (TBSTL). Corneodesmosin degradation was assessed by quantification of the 43 kDa, 33 kDa, and 19 kDa SC degradation products of the native cytosolic 52 kDa protein.
Preparation of extracts for Western blotting
Reconstructed human epidermis after 13 d of culture or SC or plantar SC obtained from human volunteers was homogenized in Tris buffered saline (TBS), pH 7.2, containing 0.5% (vol/vol) Triton X100 on an ice bed (50 l buffer per 1 mg tissue weight) with a polytron for 1 min before centrifugation (10,000g for 10 min). The protein contents of the respective supernatants were quantified using the Bio-Rad protein assay. Samples were adjusted to a protein concentration of 0.5 mg per ml before performing the protease assays or Western blot analysis.
Reconstructed human epidermis
Reconstruction of human epidermis with normal human keratinocytes was performed as described previously (Tinois et al, 1991). The reconstructed epidermis was examined at different time intervals during its reconstruction (days 1–13).
Western blotting
The proteins were submitted to SDS-PAGE on prefabricated gelbond PAG film-supported gels (12.5% acrylamide). The molecular weight standard used was a 1: 1 mix of low and high molecular weight prestained Rainbow™ standards from Amersham Pharmacia Biotech. Proteins were transferred to a PVDF membrane (Immobilon P, Millipore, Saint Quentin en Yvelines, France) in the presence of 20 mM Tris-HCl, pH 8.3, 192 mM glycine, 0.1% (wt/vol) SDS, and 20% (vol/vol) methanol using a trans-blot cell (Bio-Rad) with a constant voltage of 60 V for 4 h. The filters were blocked with TBSTL at room temperature for 1 h. Thereafter, the filters were incubated overnight at 4°C with the primary antibody. The filters were then washed twice for 5 min in TBSTL and incubated at room temperature for 2 h with the second antibody coupled to peroxidase, diluted in TBSTL. The filters were washed twice for 5 min in TBS before the detection procedure. The bands were visualized using the ECL+substrate system (Amersham Pharmacia Biotech) on a FluorSmax™ imager (Bio-Rad). Pre-stained Rainbow™ molecular weight markers served as standards and molecular weights were calculated using Quantity-One™ software (Bio-Rad).
Immunoprecipitation
SC protein extract (500 l) was incubated overnight at 4°C with 10 l of a sheep antihuman cathepsin L antibody (2.6 mg per ml). The resulting immunocomplex was precipitated using protein G (protein G kit from Roche Molecular). After centrifugation of the protein G agarose beads, 250 l of the supernatant was applied to the gel filtration column. A control without antibody was run in parallel. The functionality of the immunoprecipitation procedure was checked by incubating 0.25 g of purified human cathepsin L with 500 l extraction buffer following the same protocol as for the SC protein extract. Cysteine protease caseinolytic activity was detected as described above.
Purification and identification of the SCTP
The purification steps were performed at 4°C on a Biologic™ chromatography station (Bio-Rad). SC soluble proteins from three healthy donors were obtained as described above (9 mg protein in 18 ml buffer). The buffer was exchanged with econo-Pac™ 10DG columns (Bio-Rad) in cation exchange buffer: 50 mM acetate buffer pH 5.0, 350 mM NaCl, 1 mM EDTA, 0.1% reduced Triton X100 (buffer A). The resulting protein extract (9 mg protein in 24 ml buffer) was loaded on a Mono S™ HR5/5 column (Amersham Pharmacia Biotech) equilibrated with the same buffer at a flow rate of 1 ml per min. The column was washed with five column volumes of buffer A. Elution was performed using buffer A containing 550 mM NaCl. Proteins were eluted with a linear NaCl gradient ranging from 0 to 100% in 40 column volumes. The flow rate was 1 ml per min. Fractions (0.42 ml) were collected starting at the beginning of the gradient. Protein elution was monitored by recording the OD280nm.
After gel chromatography, the eluted proteins were submitted to SDS-PAGE on prefabricated gelbond PAG film-supported gels (G8-18) using a 1: 2 dilution with 2 Laemmli sample buffer without DTT and silver stained with the silver nitrate kit (Amersham Pharmacia Biotech). SCTP activity was revealed using the gelatin and the ZPA zymography methods. SCTP of the most active fractions were pooled, precipitated (10% trichloroacetic acid) and sequenced (Edmann sequencing) as previously described (Bernard et al, 2001).
Human cathepsin L2 antibodies
The anti-cathepsin L2 polyclonal antibody was produced by CovalAb (Lyon, France) by immunizing rabbits with selected peptides from the cathepsin L2 sequence. One peptide was retained as potentially immunizing: DBcat15CE (CFRNQKFRKGKVFRE amino acid residues 94–108 from Swissprot sequence 060911). The rabbit were immunized with a proprietary protocol (CovalAb) and the resulting antibodies were affinity purified using a peptide immobilized column. Antibody concentration obtained after the immuno-purification step was 0.5 mg per ml. For Western blot analysis, the DBcat15CE antibody was diluted 1: 1000.
Reverse transcription polymerase chain reaction (RT-PCR) analysis
Total RNA from reconstructed human epidermis was prepared using an RNeasy kit. Briefly, two samples of the reconstructed epidermis (1 cm2 each) were homogenized in 300 l of lysis buffer (included in the kit) and the cell debris was removed using a QIA shredder column. cDNA was produced from purified total RNA using the first-strand cDNA synthesis kit. Expression analysis of cathepsin L2, loricrin, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was performed by RT-PCR using equivalent amounts of total RNA (0.1 g per assay), taq polymerase, and different primers (for cathepsin L2, 5'-ctcacacattggggtagctggc-3' and 5'-cttcccaaatctgtggattg-3'; for loricrin, 5'-accacggaggcgaaggagtt-3' and 5'-ctggggttgggaggtagttg-3'; for GAPDH, 5'-aatcccatcaccatcttcca-3' and 5'-gtcatcatatttggcaggtt-3') under the following conditions: 1 cycle for 2 min at 95°C, 25–30 cycles (94°C for 30 s, 55°C for 30 s, 72°C for 30 s), and 1 cycle at 72°C for 2 min in parallel. The resulting PCR products were visualized by Gelstar™ staining after separation on 2%–3% (wt/vol) agarose gels.
Histochemistry
Thin cryosections (5 m) of skin from plastic surgery were obtained with a Microm HM500M microtome. Sections were fixed in acetone at -20°C for 5 min and air dried. Immunohistochemical detection of cathepsin L2 was performed with streptABComplex/HRP kit from Dako and AEC detection kit from Sigma according to the manufacturer's instructions except that the incubation steps were extended to 1 h using DB15CE diluted 1: 2 in phosphate-buffered saline/Tween 20 0.05% (PBST) as primary antibody and an antirabbit biotinylated secondary antibody from Roche Molecular diluted 1: 200 in PBST. Nuclei were counterstained with hematoxylin.
Results and Discussion
Caseinolytic activity of an SC extract
Gel filtration of the SC protein extract resulted in the separation of two distinct caseinolytic activities, one around pH 8.0 and the large majority at pH 5–5.5 (Figure 1a–c), a pH characteristic for the surface of the SC (Öhman and Vahlquist, 1994). As proteases involved in the desquamation process are likely to have an acidic pH optimum, we decided to further characterize the acidic activity. Based on the calculated molecular weight, the responsiveness to class specific inhibitors, and the zymograph profile, we distinguished four distinct groups within the acidic active part of the extract.
Figure 1.
Caseinolytic activities of SC proteases separated by gel filtration. The caseinolytic activities were detected with the Enzchek kit (Molecular Probes) used at pH ranging from 4 to 8.25 on fractions after gel filtration of 125 g protein extracted from human SC. Very low activities were detectable at basic pH (a, arrow 1 on two-dimensional view). The highest activities were detectable at acid pH split into three groups (a): group A with an 80-90 kDa molecular weight and pH optimum at 5.5; group B with a 40 kDa molecular weight and pH optimum between 5.25 and 5.5; group C with a 25 kDa molecular weight and pH optimum between 5.25 and 5.5. After addition of cysteine, a new caseinolytic group D was revealed with a molecular weight of 28 kD and pH optimum of 5-5.5 (b). Compared to the results without cysteine, group C is masked by this new group. Group A and B proteases are still visible. Data obtained without cysteine (a) were subtracted from the data obtained with cysteine (b). This differential analysis shows the group D activity (c).
Group A, with a calculated average molecular weight of 80 kDa (Figure 1a), and completely inhibited by the classic aspartic protease inhibitor pepstatin (Figure 2.3), was considered to comprise part of the cathepsin-E-like family. Even though, to our knowledge, this type of protease activity has not yet been described in the SC, we did not perform any further characterization, as its caseinolytic activity was very low.
Figure 2.
SC caseinolytic protease profiles obtained at pH 5.5 and the effect of specific inhibitors. Group A proteases were essentially affected by 1 M pepstatin (3). Group B proteases were totally inhibited by 1 M pepstatin (3) and partially inhibited by 6 M aprotinin (1). Group C was strongly inactivated by 6 M aprotinin (1) and to a lesser extent by 2.5 M chymostatin (2).
Group B, characterized by an average molecular weight of 40 kDa (Figure 1a), a total inhibition by pepstatin (Figure 2.3), and a broad pH optimum, was supposed to represent the cathepsin D activity. The very high caseinolytic activity in our SC extract was not surprising, asHorikoshi et al (1998;1999) described cathepsin D as the major SC proteolytic activity, possibly implicated in desquamation.
Group C, exhibiting a molecular weight of 24 kDa (Figure 1a), with a sensitivity to chymostatin and aprotinin (Figure 2.1, 2.2), comprised the chymotrypsin-like activity. As it perfectly converted (data not shown) the synthetic substrate S2586 (Egelrud, 1993), we assumed that this group represents mainly the acidic activity of the chymotryptic enzyme (SCCE), described byHansson et al (1994).
Group D attracted our particular interest for various reasons. It represented the main caseinolytic activity in the SC extract studied. It revealed its activity only in the presence of cysteine (Figure 1b, c), required the presence of mild reducing agents, and exhibited a distinct inhibitory response to specific inhibitors. Almost insensitive to pepstatin (1 M) and aprotinin (6 M) (Figure 3.5, 3.4), the activity was completely inhibited by E64 (1.4 M) (Figure 3.1) and not affected by the specific cathepsin B inhibitor CA074 (1 M) described byInubushi et al (1994) (Figure 3.6). This particular inhibition profile excluded the presence of cathepsin B, and its sensitivity to leupeptin (1 M) and chymostatin (2.5 M) (Figure 3.2, Figure 3.3) ruled out a cathepsin-H-like activity (Schwartz and Barett, 1980). So far, all of the above data suggested the presence of a cathepsin-L-like protease of the papaïn-type family. The fact that cathepsin-L-specific antibodies failed to immunoprecipitate the proteolytic activity in group D, however, clearly indicated that this activity cannot be attributed to cathepsin L itself but is exerted by another member of this family (data not shown).
Figure 3.
SC caseinolytic cysteine-dependent (group D) protease profiles at pH 5.5 and effect of specific inhibitors. The curves presented were obtained by subtracting the activities obtained with and without cysteine. Group D protease activity was tested in the presence of (1) 1.4 M E64, (2) 1 M leupeptin, (3) 2.5 M chymostatin, (4) 6 M aprotinin, (5) 1 M pepstatin, (6) 1 M CA074.
We therefore first checked whether the recently described SCTP (Watkinson, 1999), another member of the cathepsin L family, is involved. Superimposing the caseinolytic and gelatinolytic activity, detected in the fractions of the SC extract, showed that SCTP activity did not match group D activity (Figure 4).
Figure 4.
Comparison of SCTP (gelatinase) and group D cysteine-dependent protease activity after gel filtration chromatography. Human SC extract separated by gel filtration chromatography was analyzed for its cysteine-dependent activity. The activity detected on the gelatin zymogram (SCTP) did not correlate with the caseinolytic dependent activity (group D).
Full figure and legend (16K)At this stage of investigation, we decided to name the protease present in group D, stratum corneum cathepsin-L-like protease (SCCL). Due to the uniform response to specific inhibitors we were convinced that we were dealing with a single protease. To evaluate whether this new protease is implicated in the degradation of corneodesmosomes, we assayed corneodesmosin degradation by SCCL. Exposing corneocytes to purified SCCL followed by a Western blot immunodetection of the characteristic 33 kDa corneodesmosin fragment (Simon et al, 2001) revealed an efficient degradation of corneodesmosin (Figure 5). As expected from the inhibition profile (see Figure 3), leupeptin reduced, and E64 completely inhibited, SCCL-dependent degradation of corneodesmosin. Interestingly the amount of intact corneodesmosin increased in the presence of E64, indicating that this inhibitor probably blocks other cysteine proteases present in this extract. This observation can be explained by E64 inhibition of endogenous cysteine proteases associated with the corneocytes.
Figure 5.
Western blot analysis of corneodesmosins (antibody G36.19) extracted from human SC after exposure to semi-purified group D protease at pH 5 plus 5 mM cysteine. Control, SC incubated without addition of protease;+D, with protease and without inhibitor;+D+L, with protease and with leupeptin 1 M;+D+E64, with protease and with E64 1.4 M. The major SC forms of corneodesmosin are also indicated by their molecular weight (kDa).
Attempts to purify SCCL protease for further characterization by gel filtration, cation exchange, and hydrophobic interaction chromatography enabled us to dramatically increase its specific activity but did not provide sufficient material to sequence the protein. Efforts are in progress to obtain sufficient material to sequence SCCL, a protease particularly interesting because of its surprising high specific activity.
During our detailed analysis of the various protease activities, we performed experiments to make sure that SCCL activity is distinct from SCTP activity. The gelatin zymograms and caseinolytic profiles always identified SCTP and SCCL activity in distinct but neighboring fractions (data not shown). To provide unequivocal evidence that what we called SCCL is distinct from SCTP, the gelatinolytic activity was purified. We succeeded in a one-step identification by cation exchange chromatography (Figure 6). The most active fractions detected by gelatin zymogram and ZPA zymogram, i.e., fractions 32–37, were pooled and concentrated to perform another ZPA zymogram with the objective of obtaining sufficient material for sequencing. Figure 7 clearly shows that SCTP-related activity is associated with the 35 kDa protein. The band was cut, trypsin digested, and submitted to Edmann sequencing. To our surprise, the resulting peptide sequence revealed a 100% homology with the recently described cathepsin L2, also named cathepsin U or V protease (Adachi et al, 1998;Santamaria et al, 1998;Itoh et al, 1999;Kinoshita et al, 2001). These results indicate that SCTP and cathepsin L2 are two distinct names for the same enzyme. Cathepsin L2 is expressed in colorectal and breast cancer cell lines as well as in tumors of diverse origins, including ovarian and renal carcinomas (Santamaria et al, 1998), and in testis, thymus, and cornea (Itoh et al, 1999;Kinoshita et al, 2001).
Figure 6.
Purification of SCTP by cation exchange chromatography. 5 mg of protein extracted from human SC were loaded on a cation exchange chromatography column at pH 5. Proteins were eluted with a gradient ranging from 350 to 550 mM NaCl in 40 column volumes. Fractions (0.42 ml) were collected in a 96-well plate. Fractions 7 to 54 were analyzed for SCTP activity by gelatin zymography (GZ) and ZPA zymography (ZPAZ). Proteins of these fractions were separated by SDS-PAGE (Eph) on a G8-18 gradient and silver stained. Arrows indicate the molecular weights calculated for the detected activities and the molecular weight markers for the SDS-PAGE gel.
Full figure and legend (39K)Figure 7.
Coupling zymography and protein staining for the identification of purified SCTP. Fractions 32-37 from the cation exchange step supporting SCTP activity (Figure 6) were pooled and concentrated to 250 l. (S), untreated sample; (TCA), trichloroacetic acid precipitated sample. The same gel was used in (1) to detect its cathepsin-L-like activity performing a ZPA zymogram (ZPAZ) and (2) Coomassie blue stained to reveal the band supporting the enzyme activity (indicated by double-headed arrows) Arrows on the right indicate the molecular weights of the protein associated with the activity (35 kDa) and of two other stained proteins. The 35 kDa (SCTP) band was then cut and sequenced.
A detailed analysis of the cathepsin L2 sequence together with the results obtained in Western blots using a specific antibody revealed some discrepancy with earlier reports concerning the processing of cathepsin L2 and the putative active form of the enzyme.Adachi et al (1998) reported that the full-length cDNA encodes a peptide of 334 amino acids with a potential peptide signal (aa 1–17) and a putative activation peptide (aa 18–113) plus two potential N-glycosylation sites.Brömme et al (1999) described a recombinant heterogeneous pro-form of 38 and 45 kDa and an autocatalytic processing, resulting in a 31 and a 35 kDa mature form. Even though the latter result confirms the activation site predicted by Adachi, our results clearly demonstrate a distinct processing in the epidermis: (i) sequence analysis of the purified catalytic active form of cathepsin L2 comprised the aa sequence 30–33 and 39–49 (see Table I) and (ii) an antibody recognizing the cathepsin L2 aa sequence 94–108 decorated both the pro-form and the active form of the enzyme. Furthermore, this polyclonal antibody decorated cathepsin L2 in Western blots and did not cross-react with purified cathepsin L (data not shown). This antibody was used to further characterize the epidermal forms of cathepsin L2.
Table I - Amino acid sequences obtained by Edmann sequencing of the purified SCTP and sequence chosen for the generation of antibody.
Full table Western blot analysis
A comparative Western blot analysis of cathepsin L2 expression in SC extracts obtained from plantar SC and nonpalmoplantar SC (Figure 8) confirmed the presence of the 39 kDa pro-form and the mature 34 and 35 kDa forms in nonpalmoplantar SC. In addition, a 33 kDa band, probably a further degradation product, was detected. The plantar SC only contained the 39 kDa inactive pro-form of the enzyme. This is the reason why Watkinson, who analyzed plantar SC only for its catalytic gelatinolytic enzyme activity, could not detect the presence of this enzyme and concluded that it was not expressed in plantar SC.
Figure 8.
Western blot analysis using the polyclonal anti-cathepsin L2 antibody DBcat15CE. Plantar SC and SC protein extracts (3.7 g per each lane) were run on a 12.5% acrylamide SDS gel. After transfer to a PVDF membrane, immunoreactive proteins were detected with the ECL method. Arrows on the right indicate the calculated molecular weights. Arrows on the left indicate the position of two molecular weight markers.
A more detailed Western blot analysis combined with a gelatin zymogram of the fractions obtained after gel filtration of the SC extract revealed the presence of two bands (Figure 9), one with a calculated molecular weight of 39 kDa and devoid of catalytic activity in fractions 25–30, and the second with a molecular weight of 35 kDa and catalytic activity evidenced in fractions 29–39. These results clearly show that our antibody recognizes the 39 kDa pro-form as well as the 35 kDa active form of epidermal cathepsin L2. An additional set of experiments was performed to characterize the pattern of cathepsin L2 expression in human epidermis.
Figure 9.
Comparative analysis by Western blot with DBcat15CE antibody and gelatin zymography of SC fractions obtained by gel filtration. (WB), Western blot revealing the 39 kDa inactive form of cathepsin L2 and its 35 kDa active form; (GZ), gelatin zymography revealing the 35 kDa protein supporting cathepsin L2 activity (identical with the 35 kDa in the Western blot). Arrows on the right indicate the calculated molecular weights of the detected bands.
Full figure and legend (17K)Cathepsin L2 expression and epidermal differentiation
The reconstructed human epidermis model was used to analyze cathepsin L2 expression during keratinocyte differentiation. RT-PCR analysis was performed at different time intervals during the reconstruction of the epidermis, days 1-13 (Figure 10). The time course clearly evidenced that cathepsin L2 expression is independent of keratinocyte differentiation. Experiments to identify at which level of the stratified epidermis the pro-form of cathepsin L2 becomes activated are in progress. Immunolocalization of cathepsin L2 in normal human skin with the DB15CE antibody confirms the presence of cathepsin L2 throughout the living layers of the epidermis (Figure 11). The reduced staining at the intermediate layers remains to be explained. The negative staining of the SC for the enzyme is less surprising, as it is known that immunostaining in this part of the epidermis is often hampered by the strong hydrophobic environment.
Figure 10.
RT-PCR analysis of cathepsin L2 expression during the reconstruction of human epidermis. The figure shows the band intensity of specific amplicons generated by a set of primers corresponding to a differentiation marker (loricrin) and a housekeeping gene (GAPDH) compared with cathepsin L2 (CatL2) (see Materials and Methods).
Full figure and legend (10K)Figure 11.
Immunolocalization of cathepsin L2 in human epidermis. The rabbit polyclonal antibody DBcat15CE was used as the primary antibody. Nuclei were counterstained with hematoxylin. E, epidermis; D, dermis. Scale bar: 100 m.
In conclusion, the analysis of the SC extract from human skin revealed the presence of a new cysteine protease, which we named SCCL, exhibiting a surprisingly high specific activity. Moreover, the purification, sequencing, and detailed analysis at the molecular level of another cysteine protease present in the SC extract demonstrated that what was described earlier to be the "stratum corneum thiol protease" (SCTP) is not a distinct epidermal enzyme but identical with cathepsin L2. A view of these latest results in relation to other existing data concerning proteases involved in the desquamation process is presented in Table II. The fact that SCCL degrades corneodesmosin, a key component of the corneodesmosomes, indicates a possible active function of this enzyme in the desquamation process. Its purification and further characterization are in progress.
References
| 1. | Adachi W, Kawamoto S, Ohno I, Nishida K, Kinoshita S, Matsubara K & Okubo K. Isolation and characterization of human cathepsin V: A major proteinase in corneal epithelium. Invest Ophthalmol Vis Sci (1998) 39: 1789–1796. | PubMed | ISI | ChemPort | |
| 2. | Assfalg-Machleidt I, Rothe G, Klingel S, Banati R, Mangel W, Valet G & Machleidt W. Membrane permeable fluorogenic rhodamine substrates for selective determination of cathepsin L. Biol Chem Hoppe Seyler (1992) 373: 433–440. | PubMed | ChemPort | |
| 3. | Bernard D, Méhul B, Delattre C, Simonetti L, Thomas-Collignon A & Schmidt R. Purification and characterization of the endoglycosidase heparanase 1 from human plantar stratum corneum: A key enzyme in epidermal physiology? J Invest Dermatol (2001) 117: 1266–1273. | Article | PubMed | ISI | ChemPort | |
| 4. | Brömme D, Zhenqiang L, Barnes M & Mehler E. Human cathepsin V functional expression, tissue distribution, electrostatic surface potential, enzymatic characterization and chromosomal localization. Biochemistry (1999) 38: 2377–2385. | Article | PubMed | ISI | ChemPort | |
| 5. | Egelrud T. Purification and preliminary characterization of stratum corneum chymotryptic enzyme: a proteinase that may be involved in desquamation. J Invest Dermatol (1993) 101: 200–204. | Article | PubMed | ISI | ChemPort | |
| 6. | Egelrud T & Lundström A. The dependence of detergent-induced cell dissociation in non-palmo-plantar stratum corneum on endogenous dermatology. J Invest Dermatol (1990) 95: 456–459. | Article | PubMed | ISI | ChemPort | |
| 7. | Egelrud T & Lundström A. A chymotrypsin-like proteinase that may be involved in desquamation in plantar stratum corneum. Arch Dermatol Res (1991) 283: 108–112. | Article | PubMed | ISI | ChemPort | |
| 8. | Ekholm IE, Brattsand M & Egelrud T. Stratum corneum tryptic enzyme in normal epidermi: A missing link in the desquamation process? J Invest Dermatol (2000) 114: 56–63. | Article | PubMed | ISI | ChemPort | |
| 9. | Hansson L, Strömqvist M, Bäckman A, Wallbrandt P, Carlstein A & Egelrud T. Cloning expression, and characterization of stratum corneum chymotryptic enzyme: A skin-specific human serine proteinase. J Biol Chem (1994) 269: 19420–19426. | PubMed | ISI | ChemPort | |
| 10. | Horikoshi T, Arany I & Rajaraman S et al. Isoforms of cathepsin D and human epidermal differentiation. Biochimie (1998) 80: 605–612. | Article | PubMed | ISI | ChemPort | |
| 11. | Horikoshi T, Igarashi S, Uchiwa H, Brysk H & Brysk MM. Role of endogenous cathepsin d-like and chymotrypsin-like proteolysis in human epidermal desquamation. Br J Dermatol (1999) 141: 453–459. | Article | PubMed | ISI | ChemPort | |
| 12. | Inubushi T, Kakegawa H, Kishino Y & Katunuma N. Specific assay method for the activities of cathepsin l-type cysteine proteinases. J Biochem (1994) 116: 282–284. | PubMed | ISI | ChemPort | |
| 13. | Itoh R, Kawamoto S, Adachi W, Kinoshita S & Okubo K. Genomic organization and chromosomal localization of the human cathepsin L2 gene. DNA Res (1999) 6: 137–140. | Article | PubMed | ChemPort | |
| 14. | King IA, Tabiowo A, Purkis P, Leigh I & Magee AI. Expression of distinct desmocollin isoforms in human epidermis. J Invest Dermatol (1993) 100: 373–379. | Article | PubMed | ISI | ChemPort | |
| 15. | Kinoshita S, Adachi W, Sotozono C, Nishida K, Yokoi N, Quantock AJ & Okubo K. Characteristics of the human ocular surface epithelium. Prog Retin Eye Res (2001) 20: 639–673. | Article | PubMed | ISI | ChemPort | |
| 16. | Lundström A, Serre G, Haftek M & Egelrud T. Evidence for a role of corneodesmosine, a protein which may serve to modify desmosomes during cornification, in stratum corneum cell cohesion and desquamation. Arch Dermatol Res (1994) 286: 369–375. | PubMed | ISI | ChemPort | |
| 17. | Méhul B, Bernard D, Simonetti L, Bernard MA & Schmidt R. Identification and cloning of a new calmodulin-like protein from human epidermis. J Biol Chem (2000) 275: 12841–12847. | PubMed | |
| 18. | Öhman H & Vahlquist A. In vivo studies concerning a pH gradient in human stratum corneum and upper epidermis. Acta Derm Venereol (1994) 74: 375–379. | PubMed | ISI | ChemPort | |
| 19. | Rawlings A, Harding C, Watkinson A, Banks J, Ackerman C & Sabin R. The effect of glycerol and humidity on desmosome degradation in stratum corneum. Arch Dermatol Res (1995) 287: 457–464. | Article | PubMed | ISI | ChemPort | |
| 20. | Santamaria I, Velasco G, Cazorla M, Fueyo A, Campo E & Lopez-Otin C. Cathepsin L2, a novel human cysteine proteinase produced by breast and colorectal carcinomas. Cancer Res (1998) 58: 1624–1630. | PubMed | ISI | ChemPort | |
| 21. | Schwartz WN & Barett AJ. Human cathepsin H. Biochem J (1980) 191: 487–497. | PubMed | ISI | ChemPort | |
| 22. | Serre G, Mils V & Haftek M et al. Identification of late differentiation antigens of human cornified epithelia, expressed in re-organized desmosomes and bound to cross-linked envelope. J Invest Dermatol (1991) 97: 1061–1072. | Article | PubMed | ISI | ChemPort | |
| 23. | Simon M, Montézin M, Guerrin M, Durieux JJ & Serre G. Characterization and purification of human corneodesmosin, an epidermal basic glycoprotein associated with corneocyte-specific modified desmosomes. J Biol Chem (1997) 272: 31770–31776. | Article | PubMed | ISI | ChemPort | |
| 24. | Simon M, Jonca N & Guerrin M et al. Refined characterization of corneodesmosin proteolysis during terminal differentiation of human epidermis and its relationship to desquamation. J Biol Chem (2001) 276: 20292–20299. | Article | PubMed | ISI | ChemPort | |
| 25. | Suzuki Y, Nomura J, Koyama J & Horii I. The role of proteases in stratum corneum: involvement in stratum corneum desquamation. Arch Dermatol Res (1994) 286: 249–253. | Article | PubMed | ISI | ChemPort | |
| 26. | Tinois E, Tillier J & Gaucherand M. Tardy M, Thivolet J: In vitro and post-transplantation differentiation of human keratinocyte growth on the human type IV collagen film of a bilayered dermal substitute. Exp Cell Res (1991) 193: 310–319. | Article | PubMed | ISI | ChemPort | |
| 27. | Walsh A & Chapman SJ. Sugars protect desmosome and corneosome glycoproteins from proteolysis. Arch Dermatol Res (1991) 283: 174–179. | Article | PubMed | ISI | ChemPort | |
| 28. | Watkinson A. Stratum corneum thiol protease (SCTP): A novel cysteine protease of late epidermal differentiation. Arch Dermatol Res (1999) 291: 260–268. | Article | PubMed | ISI | ChemPort | |
Acknowledgments
We thank Dr. Jacques d'Alayer from the Laboratoire de Microséquençage des Protéines, Institut Pasteur, Paris, for the protein sequencing. We are also grateful to M. Turine, E. Boivin, and Dr. de Lacharrière for providing the human epidermis and Dr. Michel Kermici for helpful discussions.
