Purification of Protease from a Mixture of Exfoliative Toxin and Newborn-Mouse Epidermis

Abstract

Although the role of exfoliative toxin in staphylococcal scalded-skin syndrome has been suggested to be that of a serine protease, it has not been demonstrated to show proteolytic activity. Our purpose was to purify a proteolytic enzyme from a mixture of exfoliative toxin and newborn-mouse epidermis. We used gel filtration and ion-exchange and hydroxyapatite chromatography with a high-pressure liquid chromatography system. A casein-hydrolyzing enzyme was isolated from the mixture. The molecular mass of the enzyme was confirmed to be 20 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Subcutaneous injection of the purified enzyme into newborn mice reproduced the epidermal splitting that is seen in staphylococcal scalded-skin syndrome. These results suggest that exfoliative toxin does not work as a protease itself but that some reaction between exfoliative toxin and an epidermal component(s) first produces a protease, after which epidermal splitting occurs.

Staphylococcal scalded-skin syndrome (SSSS) occurs in infants and is characterized by extensive reddening and exfoliation of the skin. In 1970, Melish and Glasgow (10) established an experimental model of SSSS using newborn mice injected with Staphylococcus aureus isolated from patients. Soon several workers isolated and characterized an exotoxin of S. aureus which was responsible for the skin exfoliation in SSSS (1, 4, 11). The exotoxin was named “exfoliative toxin” (ET).

ET is classified into two serotypes, A and B (ETA and ETB), which differ in their molecular masses, levels of heat stability, and antigenic specificities (19). ETA is encoded by a bacterial chromosome, whereas ETB is encoded by a plasmid (20). Both genes have been sequenced (7, 13), and the ETA gene has been cloned in Escherichia coli (15).

Using a light microscope, the lesion responsible for the exfoliation caused by ET can be observed histologically as a cleft in the granular layer of the epidermis (8). Electron-microscopic examination confirmed that the desmosomes were disrupted and that a fluid-filled cleft was formed between the cells (8). Lillibridge et al. (8) reported that the initial change of epidermal splitting was a disappearance of the small organelles called “bubbles” between the granular cells, with interepidermal edema occurring secondarily, followed by disruption of the desmosomes. They suggested that a hydrolytic enzyme in the bubbles caused disruption of the desmosomes.

It has been postulated that ETA and ETB might themselves be, or are inducers of, proteolytic enzymes causing disruption of the desmosomes.

After establishment of the amino acid sequences of ETA and ETB (7, 13), ETA was found to be about 25% identical with staphylococcal V8 protease. Replacement of the Ser-195 residue by a cysteine residue inactivates the toxicity of ETA (14). Vath et al. (18) reported that the X-ray crystal structure of ETA consisted of two crystal forms. ETA's structure indicates that it belongs to the chymotrypsin-like family of serine proteases and that it cleaves protein substrates at sites after acidic residues. On the other hand, Sakurai et al. (16) reported that the Tyr-157 and Tyr-159 residues are involved in the toxicity of ETB. They also reported that ETA contained the same amino acid residues.

These findings suggest that ET functions as a proteolytic enzyme. However, the active site of ET is not clear, and addition of protease inhibitors to both in vitro and in vivo models of epidermolysis failed to inhibit epidermal splitting (12). Moreover, ET has never been demonstrated to show proteolytic activity either in vitro or in vivo.

We reported that casein-hydrolyzing activity appeared in a mixture of ET and newborn-mouse epidermis (17). As other authors have noted, this suggests that the epidermal splitting seen in SSSS is caused by a protease.

Here, we report chromatographic purification of a casein-hydrolyzing enzyme from a mixture of ET and newborn-mouse epidermis and confirmation of a relationship between this enzyme and the epidermal splitting seen in SSSS. The molecular mass of the enzyme was found to be 20 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

MATERIALS AND METHODS

Animals and preparation of epidermis.

Skin was obtained from 1- to 2-day-old mice (Jla-ddy strain). The epidermis was separated from the skin, which had been stored at 4°C after incubation in 2.4 M NH₄Cl (pH 9.5) solution (3).

Purification of ET.

S. aureus, phage group 2 strain, which had been isolated from a patient with bullous impetigo, was cultured in broth containing 10 g of yeast extract (Difco Laboratories, Detroit, Mich.), 17 g of Trypticase (Becton Dickinson and Co., Cockeysville, Md.), 5 g of NaCl, and 2.5 g of K₂HPO₄ per liter of purified water at 37°C for 12 h. Ten milliliters of cultured sample was added to 500 ml of the same broth and incubated at 37°C for 48 h in an atmosphere containing 10% CO₂. After centrifugation at 17,700 × g for 20 min at 4°C (model CR20B2 centrifuge; Hitachi, Tokyo, Japan), ET was partially purified from the culture supernatant by the method described by Kondo et al. (5) by using ammonium sulfate precipitation.

The precipitate was dissolved in 10 mM Tris-HCl buffer (pH 7.7) and dialyzed against the same buffer at room temperature for 12 h. After concentration to 2 ml by ultrafiltration with an Amicon (Beverly, Mass.) YM3 membrane, the sample was applied to a Sephadex G-50 (Pharmacia Fine Chemicals, Uppsala, Sweden) column (27 by 90 mm) which had been equilibrated with 10 mM Tris-HCl buffer (pH 7.7) containing 0.1 M NaCl. This sample was eluted with the same buffer at a flow rate of 20 ml/h. Fractions (5 ml each) of ET were collected, combined, and concentrated by ultrafiltration.

Determination of protein concentration.

Protein was determined by the method of Lowry et al. (9).

Preparation of casein-hydrolyzing enzyme from a mixture of ET and newborn-mouse epidermis.

Approximately 4.1 g (wet weight) of epidermis was suspended in 4 ml of 10 mM Tris-HCl buffer (pH 7.7) containing 9.6 mg of the partially purified ET and incubated at 37°C for 12 h. The epidermis was removed by centrifugation at 950 × g for 15 min at 4°C and passed through a Millipore filter (pore size, 0.22 μm). The supernatant was concentrated to 2 ml by ultrafiltration with an Amicon YM3 membrane.

The sample was applied to a Sephadex G-50 (Pharmacia Fine Chemicals) column (27 by 90 mm) which had been equilibrated with 10 mM Tris-HCl buffer (pH 7.7) containing 0.1 M NaCl and eluted with the same buffer at a flow rate of 20 ml/h. Fractions (5 ml each) were collected and assayed for casein-hydrolyzing activity. The active fractions were combined and concentrated.

The sample was charged onto a gel filtration high-pressure liquid chromatography (HPLC) column (BIO-Gel SEC 30XL, 7.8 by 300 mm; Bio-Rad, Hercules, Calif.) which had been equilibrated with 10 mM Tris-HCl buffer (pH 7.7) containing 0.1 M NaCl. Using an HPLC system (model 800; Bio-Rad), the sample was eluted with the same buffer at a flow rate of 1 ml/min and the fractions positive for casein-hydrolyzing activity were collected, combined, and concentrated.

The sample was applied to an ion-exchange HPLC column (Bio-Gel diethylaminoethyl-5-pw, 7.5 by 75 mm; Bio-Rad) which had been equilibrated with 10 mM Tris-HCl buffer (pH 7.7) containing 0.15 M NaCl. Using a linear gradient from 0.15 to 1 M NaCl at a flow rate of 1 ml/min, the sample was eluted. While monitoring the concentration of NaCl and the absorbance value at 280 nm, we collected, combined, and concentrated the fractions positive for casein-hydrolyzing activity.

After we changed the buffer to one containing 1 mM phosphate (pH 6.8), the sample was applied to a hydroxyapatite HPLC column (model MP; Kanto Chemical, Tokyo, Japan), equilibrated with the same buffer, and eluted with a linear gradient from 1 to 50 mM sodium phosphate at a flow rate of 0.5 ml/min. While monitoring the concentration of sodium phosphate and the absorbance value at 280 nm, we collected the fractions positive for casein-hydrolyzing activity.

Assay for casein-hydrolyzing activity.

Samples (500 μl) were incubated with 100 μl of casein (2.5 mg of Hammarsten casein; E. Merck, Darmstadt, Germany) at 37°C for 1 h. The reaction was stopped by addition of 67 μl of 50% trichloroacetic acid. After cooling for 15 min, samples were centrifuged at 950 × g for 10 min at 4°C. We assayed for soluble amino peptides from the casein in the supernatant using the method described by Lowry et al. (9). The optical density at a wavelength of 500 nm was measured with a spectrophotometer (model UV-120-02; Shimadzu, Tokyo, Japan). We prepared blanks by adding casein just after the addition of trichloroacetic acid. A 0.1-unit increase in the corrected absorbance value at 500 nm was calculated as 1 enzyme unit, and the specific activity was expressed as enzyme units per milligram of protein.

Gel electrophoresis.

SDS-PAGE was performed according to the method described by Laemmli (6) with a microslab gel electrophoresis apparatus (Marisol, Tokyo, Japan). The polyacrylamide concentration of the gel was 12.5%. The sample was first boiled with 2% SDS and 5% 2-mercaptoethanol for 2 min, and 30 μl was then electrophoresed. SDS-PAGE was carried out at 10 mA for 1 h at room temperature. The gel was stained with silver nitrate.

Western blotting.

After SDS-PAGE, the sample and ET were transferred onto a polyvinylidene difluoride membrane (Sequi-Blot; Bio-Rad). After being blocked with 2% bovine serum albumin, the membrane was reacted with rabbit anti-ET immunoglobulin G (IgG) antibody for 1 h at 37°C. The antibody which was bound to the protein band was reacted with peroxidase-goat anti-rabbit IgG antibody and stained with 3,3′-diaminobenzidine tetrahydrochloride.

Reproduction of epidermal splitting.

One hundred microliters of purified sample (36 ng) or partially purified ET (0.32 μg/μl) was injected under the dorsal skin of newborn mice. The mice were held at 37°C, and epidermal splitting was confirmed by rubbing the skin directly each hour. To evaluate the histological appearance, specimens from the same site were collected, fixed with formalin, and stained with hematoxylin and eosin. We observed the stained specimens under a light microscope.

RESULTS

Figure 1 shows the profile of the gel filtration HPLC. Casein-hydrolyzing activity was detected between fractions 23 and 29. When only ET was charged, ET appeared between fractions 20 and 23. When only newborn-mouse epidermis was charged, no casein-hydrolyzing activity was found in any fraction and epidermal splitting could not be reproduced.

FIG. 1.

Profile of gel filtration HPLC. Active fractions from Sephadex G-50 column chromatography were applied to a column which had been equilibrated with 10 mM Tris HCl buffer (pH 7.7) containing 0.1 M NaCl. The sample was eluted with the same buffer at a flow rate of 1 ml/min. Each fraction was collected and assayed for casein-hydrolyzing activity at a wavelength of 500 nm.

Casein-hydrolyzing activity was eluted between fractions 38 and 40 in ion-exchange HPLC, and the concentration of NaCl was calculated to be about 0.25 to 0.56 M (Fig. 2). When only ET was applied, it did not adsorb to the column under these conditions.

FIG. 2.

Profile of ion-exchange HPLC. Active fractions from gel filtration HPLC were applied to a column which had been equilibrated with 10 mM Tris-HCl buffer (pH 7.7) containing 0.15 M NaCl. Using a linear gradient from 0.15 to 1 M NaCl at a flow rate of 1 ml/min, the sample was eluted. The concentration of NaCl was monitored. Each fraction was collected and assayed for casein-hydrolyzing activity at a wavelength of 500 nm.

The enzyme was eluted between fractions 26 and 27 in hydroxyapatite HPLC, and the concentration of phosphate buffer was calculated to be about 3.27 mM (Fig. 3).

FIG. 3.

Profile of hydroxyapatite HPLC. Active fractions from ion-exchange HPLC were applied to a column which had been equilibrated with 1 mM phosphate buffer (pH 6.8) and eluted with a linear gradient from 1 to 50 mM sodium phosphate at a flow rate of 0.5 ml/min. The concentration of sodium phosphate was monitored. Each fraction was collected and assayed for casein-hydrolyzing activity at a wavelength of 500 nm.

Table 1 compiles the results of purification of the casein-hydrolyzing enzyme from the mixture of ET and newborn-mouse epidermis. We isolated 5,140 enzyme units from the mixture containing 15,990 enzyme units. The sample yield was 34.1%, and the specific activity was 80,125 U/mg. In addition, the ET we used was confirmed to be ETB on the basis of its heat stability before these examinations.

TABLE 1.

Results of purification

Method	Total enzyme units	Total wt (mg)	Sp act (U/mg)	Yield (%)
Mixture of ET and epidermis	15,990	137	116.6	100.0
Gel filtration HPLC	7,821	4.3	1,818.9	48.9
Ion-exchange HPLC	6,509	0.39	16,819.6	40.7
Hydroxyapatite HPLC	5,449	0.07	80,125	34.1

In SDS-PAGE of the enzyme fractions, a single band was detected (Fig. 4). Its molecular mass was calculated to be about 20 kDa.

FIG. 4.

Result of SDS-PAGE. The polyacrylamide concentration of the gel was 12.5%. The sample was first boiled with 2% SDS and 5% 2-mercaptoethanol for 2 min, and 30 μl was then electrophoresed. SDS-PAGE was carried out at 10 mA for 1 h at room temperature. The gel was stained with silver nitrate. A single band was detected in the gel. Molecular masses of marker proteins are shown at the left.

In Western blot analysis, the sample cross-reacted with anti-ET antibody (Fig. 5).

FIG. 5.

Result of Western blotting. The sample (A) and ET (B) were transferred onto a polyvinylidene difluoride membrane. After being blocked with 2% BSA, the membrane was reacted with rabbit anti-ET antibody at 37°C for 1 h. The antibody which was bound to the protein band was reacted with peroxidase-goat anti-rabbit IgG antibody and stained with 3,3′-diaminobenzidine tetrachloride. The sample cross-reacted with anti-ET antibody.

Twelve hours after injection of the enzyme to a newborn mouse, epidermal splitting was observed on the dorsal skin (Fig. 6). Figure 7 shows the histological appearance of the dorsal skin. Splitting occurred in the granular layer of the epidermis. Figure 8 shows the histological appearance of the dorsal skin 12 h after injection of ET. The splitting occurred at the same location.

FIG. 6.

Result at 12 h after injection of the sample into a newborn mouse. Epidermal splitting of the dorsal skin is observed.

FIG. 7.

Histological appearance of Fig. 5 specimen. The specimen shows a cleft in the granular layer.

FIG. 8.

Histological appearance after injection of ET. The specimen shows a cleft in the granular layer, which corresponds to that in Fig. 7.

DISCUSSION

The most important problem regarding purification of the enzyme is whether or not it was contaminated by ET. The gel filtration HPLC profile of the sample was quite different from that of ET. In ion-exchange HPLC, the sample was eluted successfully and ET was not adsorbed to the column. Furthermore, the result of SDS-PAGE showed a single band, which was different from the result with ET. These results indicate that the casein-hydrolyzing enzyme in the mixture of ET and newborn-mouse epidermis was definitely not ET and that ET did not contaminate the sample.

Epidermal splitting caused by injection of the purified enzyme agreed perfectly with that caused by injection of ET itself. It was also confirmed histologically. After all, epidermal splitting in SSSS is an enzyme-induced phenomenon. However, ET itself does not work as a protease. Some reaction between ET and an epidermal component(s) first produces a proteolytic enzyme, and then epidermal splitting occurs.

Many authors have pointed out a similarity between ET and serine protease. In Western blot analysis, our isolated proteolytic enzyme cross-reacted with anti-ET antibody. The result suggests that ET is a proenzyme and that our isolated proteolytic enzyme is an active form of ET. We think that reaction with an epidermal component(s) changes the structure of ET. We need to further determine its characteristics, such as its amino acid sequence.

We were able to cause epidermal splitting by dorsal injection of the enzyme into mice, but the substrate of the enzyme in vitro was casein. With what kind of substrate the enzyme reacts in vivo and which component(s) of the intracellular junction it degrades are also unclear. We hope to solve these problems in future experiments.