Structural and Transglutaminase Substrate Properties of the Small Proline-rich 2 Family of Cornified Cell Envelope Proteins*

The small proline-rich (SPR) proteins are components of the cornified cell envelope of stratified squamous epithelia and become cross-linked to other proteins by transglutaminases (TGases). The SPR2 family is the most complex, as it consists of several differentially expressed members of the same size. To explore their physical and cross-linking properties, we have expressed in bacteria a human SPR2 family member, and purified it to homogeneity. By circular dichroism, it possesses no α or β structure but has some organized structure associated with the central peptide repeat domain. The TGase 1, 2, and 3 enzymes expressed in epithelia use the recombinant SPR2 protein as a complete substratein vitro, but with widely differing kinetic efficiencies, and in different ways. With TGase 1, only one glutamine on the head domain and one lysine on the tail domain were used for limited interchain cross-linking. With TGase 3, multiple head and tail domain residues were used for extensive interchain cross-linking. The total usage of glutamine and lysine residues in vitro by TGase 3 was similar to that seen in earlier in vivo studies. We conclude that SPR2 proteins are cross-linked in epithelia primarily by the TGase 3 enzyme, a minor extent by TGase 1, and probably not by TGase 2.

Small proline-rich (SPR) 1 proteins are a group of about 12 closely related proteins that are expressed in terminally differentiating mammalian stratified squamous epithelia (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). To date, three classes of SPRs have been identified: SPR1 (two members in all species examined so far), SPR2 (8 -11 members), and a single member of the SPR3 class. All are built according to a common plan. They possess (head) amino-and (tail) carboxyl-terminal domains that are enriched in Gln, Lys, and Pro residues of sequences that are characteristic of and highly conserved between members within each class. The SPR name in fact is based on a distinctive central domain consisting of a series of peptide repeats of either eight (SPR1 and SPR3) or nine (SPR2) residues that are enriched in Pro residues and of sequences that vary between the three classes, but members within each class are also highly conserved. The numbers of repeats range from as few as three in all human SPR2 proteins and the smallest mouse SPR2 protein 2 to more than 23 in SPR3 proteins (6,11,12), so that the total protein size ranges from 6 kDa to Ϸ25 kDa.
Recent experiments have documented that SPR proteins serve as structural protein precursors of the cornified cell envelope (CE), which is a specialized structure that affords essential barrier function to stratified squamous epithelia (24 -26). It is assembled at the cell periphery and constitutes a layer of protein of uniform thickness (about 15 nm) and density (about 7 kDa/nm 2 ) in different epithelia (27). The CE is made insoluble by extensive cross-linking of several defined structural proteins including the SPRs by disulfide bonds and N ⑀ -(␥-glutamyl)lysine or N 1 ,N 8 -bis(␥-glutamyl)spermidine isopeptide bonds formed by the action of transglutaminases (TGases) (24 -26, 28). TGases catalyze the usually irreversible formation of an intra-or more usually intermolecular isopeptide bond between acyl donor Gln and acceptor Lys residues (29 -31). Unambiguous evidence that SPRs serve as CE precursors and are in fact good substrates for TGases in vivo comes from direct sequencing analyses. Many peptides recovered from limited proteolysis experiments of CEs isolated from human foreskin epidermis (32)(33)(34), cultured epidermal keratinocytes (35,36), 3 and mouse forestomach epithelium (12) contained recognizable SPR sequences cross-linked through one or more isopeptide bonds to other CE proteins. Examination of these peptides revealed that Gln and Lys residues on only the head and tail sequences of the SPRs were involved in the cross-links (12,(33)(34)(35)(36). Furthermore, the data indicated that the SPRs functioned primarily as cross-bridging proteins, by mediating links between other proteins, often loricrin in the epidermis (33,34,36), or loricrin, trichohyalin, and themselves in the forestomach CEs (12). They also form cross-links between themselves in the CEs of cultured keratinocytes that may serve as a model system for other internal stratified squamous epithelia (36).
However, little information is available to date on the biophysical and biochemical properties of the SPR proteins. Although individual members of the SPR family can function as both amine donors and acceptors for the TGase 1 enzyme in vitro (8), few data are available on which TGase(s) are responsible for cross-linking them to CE structures in epithelia in vivo. Three TGases are commonly expressed in epithelia, including the TGase 1, TGase 2, and TGase 3 enzymes (24 -26). In this paper, we have initiated a detailed analysis of their structures and properties. We have expressed in bacteria a member of the human SPR2 class, examined some of its structural features, and explored its substrate properties by the three TGase enzymes. Our new data reveal that, although the three enzymes can cross-link SPR2 in vitro, we can conclude that the TGase 3 enzyme is used almost exclusively in epithelia in vivo.

Bacterial Expression of Recombinant a Human SPR2 in the pET-11a
System-A full-length cDNA encoding human SPR2 (clone 930 of Ref. 1) was obtained as a generous gift from C. Backendorf. Following addition of appropriate linkers, it was inserted into the pET-11a bacterial expression vector (Novagen, Madison, WI) and transformed into the host Escherichia coli B strain BL 21/DE3 pLys S (Novagen). Cultures (0.5-1 liter volume) were grown in LB medium supplemented with 50 g/ml of ampicillin to A 600 nm of 0.6, and protein expression was induced by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h. In some cultures, L-[ 35 S]cysteine (0.5 Ci/ml) was added immediately prior to induction.
Purification of Recombinant SPR2 Protein-Following induction, bacteria were pelleted at 10,000 ϫ g for 30 min. Fresh or previously frozen (Ϫ70°C) pellets were lysed in a buffer of 100 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM dithiolthreitol, 1 mM EDTA, and the protease inhibitor 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF) (0.2 mM), and pelleted at 10,000 ϫ g for 30 min. The bacterial lysates were treated with DNase I, made to 8 M urea, and dialyzed against several changes of 100 volumes of 25 mM sodium citrate (pH 3.6), 1 mM dithiolthreitol, 1 mM EDTA, and 1 mM AEBSF. Under these conditions, most of the bacterial proteins precipitated; however, as the SPR2 protein remained soluble under these conditions, it was enriched to about 60% purity. Purification to homogeneity was achieved using a Pharmacia fast protein liquid chromatography system on a 0.5 ϫ 5-cm Mono-S column (Amersham Pharmacia Biotech Inc.) equilibrated in the citrate buffer with a 0 -1.0 M NaCl gradient, and was eluted with 0.12 M salt.
The fractions were analyzed on 4 -20% SDS-polyacrylamide gels (Novex, San Diego, CA) with Coomassie stain, or by Western blotting. In the latter, PDVF or nitrocellulose membranes were placed under and on the top of the gel to be blotted since the very basic SPR2 transferred to both sides. A specific antibody was raised in rabbits against the COOH-terminal epitope for SPR2 proteins. 2 The enhanced chemiluminescence detection was performed with the Super Signal CL-HRP substrate system (Pierce). Since SPR2 is enriched in Cys residues, it was also easily monitored by autoradiography following labeling with [ 35 S]cysteine, whereas bacterial proteins contain few Cys residues. The concentrations and purity of purified recombinant SPR2 were determined by amino acid analysis.
Cross-linking of Recombinant Human SPR2-Human full-length TGase 1 was expressed in baculovirus, and the particulate fraction was isolated as described (37). In control experiments, we have found that this enzyme has the same specific activity, and cross-links loricrin in vitro in exactly the same way, as the native keratinocyte or bacterially expressed full-length human TGase 1 enzymes used in previous studies (37). Guinea pig liver TGase 2 was obtained commercially (Sigma). Guinea pig epidermal TGase 3 was isolated and activated by dispase as described previously (38). For in vitro cross-linking studies using the recombinant human SPR2 as a complete TGase substrate, the purified SPR2 was equilibrated by dialysis into a buffer of 50 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM DTT, and 1 mM EDTA. In analytical crosslinking experiments, 35 g of 35 S-labeled SPR2 (about 4000 dpm) were utilized in a 200-l reaction volume. In order to standardize the reactions for comparisons of the efficiency of the three enzymes, the same amount of enzymatic activity was used for each enzyme (39 -41). These activities were measured by [ 14 C]putrescine (Amersham Pharmacia Biotech, specific activity 118 mCi/mmol) incorporation into succinylated casein, and the amount of TGase 1, 2, and 3 that incorporated 4 pmol of putrescine/min into the casein was used. The TGases were activated by addition of 5 mM Ca 2ϩ to the mixture, and incubated at 37°C. Aliquots were withdrawn at selected times and the reactions were stopped by the addition of EDTA (to 10 mM). The cross-linked products were separated on 4 -20% polyacrylamide gels, blotted on nitrocellulose membrane and analyzed by autoradiography, or by specific antibodies against the three TGases (40). In preparative experiments (for isolation of peptides for sequencing) with TGases, 100 g of recombinant human SPR2 were reacted in a volume 250 l for up to 18 h, by which time in control experiments the cross-linking had proceeded to completion.
Kinetic Studies of Putrescine Incorporation into Recombinant Human SPR2-Kinetic reactions were performed using the same three enzymes and buffer conditions as used above. Five concentrations of unlabeled recombinant SPR2 protein (2.5, 3, 4, 5, and 7 M), three different concentrations of [ 3 H]putrescine (0.07, 0.10, and 0.14 mM), and standardized amounts of each enzyme determined as described above, were used. These reactions contained a large molar excess of putrescine in order to achieve linearity of the kinetic parameters (43), and to favor the use of only one putrescine amino group, rather than to permit cross-linking of the SPR2 proteins by putrescine or to themselves. Aliquots of the reactions were stopped by addition of EDTA and unlabeled putrescine (10 mM each, final concentration), and were bound to polyvinylidene difluoride membranes by filtration through a dot-blotting apparatus. Following extensive washing in 20% aqueous ethanol containing 10 mM unlabeled putrescine, the blots were cut up to count each sample. The assay procedures utilized here were modified from previous studies since the SPR2 proteins remained soluble in trichloroacetic acid. Kinetic parameters were calculated as described previously (39 -41).
Peptide Mapping and Amino Acid Sequencing-Control samples of uncross-linked human recombinant SPR2 were digested with trypsin (1:30 by weight, Sigma, sequencing grade) for 6 h at 37°C. The peptides were resolved on a Phenomex ODS reverse phase column (2.1 ϫ 250 mm) containing 0.08% trifluoroacetic acid and with a gradient of 5-65% acetonitrile over 70 min. The 6-and 18-h preparative cross-linking reaction mixtures of recombinant human SPR2 proteins with the TGase 1, TGase 2, and TGase 3 enzymes were stopped by addition of EDTA (10 mM), digested with trypsin, and peptides were separated under identical conditions. Many of the peaks were shifted to later elution times due to the insertion of cross-links. The peaks were collected and sequenced on a Porton LF3000 gas phase sequencer as described previously (33,34). With the TGase 3 enzyme, peptides were poorly resolved due to extensive cross-linking; in these experiments, 1-min fractions across the broad peak were collected for sequencing (40). Where possible, amino acid analyses of peaks were performed prior to sequencing (33,34). The amounts of isodipeptide cross-link inserted into the SPR2 protein were determined by total proteolytic digestions as described previously (42).
Physical and Cross-linking Studies Using Synthetic Peptides-The following peptides were synthesized based on published sequences (6) and purified by HPLC. Human SPR2 head domain: SYQQQQCKQPC-QPPPVCPT; human SPR2 tail domain: QQCQQKCPPVTPSPPCQP-KCPPKSK; and human SPR2 central domain repeat (PKCPEPCPP) 3 and (PKCPEPCPP) 6 . Kinetic constants using these peptides as substrates for putrescine incorporation were determined as described previously (41).
Circular Dichroism Spectra of Recombinant SPR2 Proteins-Recombinant human SPR2 or the above synthetic peptides were equilibrated (1 mg/ml) into either phosphate-buffered saline or the same buffer as used for the in vitro TGase cross-linking reactions, and examined using a Jasco 600 spectropolarimeter. The sample holders were thermostated before and during measurements, using external circulation. Solutions were maintained at 20, 45, or 60°C for 1 h, or cooled to 20°C for 1 h before measurement. Solutions were equilibrated into 5 M guanidine hydrochloride in buffer by dialysis for 16 h prior to measurement.

Expression and Purification of Recombinant Human SPR2
Protein-Many reports have implicated or indicated that SPR proteins are substrates for TGases in vivo and in vitro, yet important questions concerning their structural and biochemical properties remain unresolved. Moreover, the available in vivo cross-linking data from this laboratory have provided no information on which enzymes are used to cross-link the SPR proteins, and their specificities. In this study, we have expressed a human SPR2 protein in bacteria using the pET-11a system. Bacterial pellets were made to 8 M urea and dialyzed directly into 25 mM citrate buffer pH 3.6, in which the SPR2 was very soluble, but the bacterial proteins were largely insoluble, thereby effecting a high degree of purification in one step. A final complete purification was possible by chromatography on a Mono-S FPLC column (Fig. 1). The maximal yield was only on the order of 2-5 mg/liter, which we were unable to improve by modifications to the culture media, time of induction of expression, etc., suggesting that the SPR2 protein may be toxic. In this way, it behaves similarly to loricrin, with which it shares charge and some sequence homologies (1, 2, 6, 43). As expected from its unusually basic pI value, the expressed SPR2 migrated on SDS gels with an apparent molecular mass of about 10 kDa, rather higher than that calculated from its known sequences (6 kDa) (Fig. 1, inset).
Circular Dichroism Spectra of the Recombinant SPR2 Protein-CD spectra were performed to evaluate the secondary structure of the recombinant human SPR2 protein. It possessed some organized structure in phosphate buffered saline or the TGase enzyme assay buffer at 20°C but there was essentially no ␣ or ␤ structure present ( Fig. 2A). Next, we assessed the overall structural properties of the SPR2 protein as a function of temperature (Fig. 2B) and guanidine hydrochloride (Fig. 2C). It was significantly unfolded by heating at 45°C (Fig. 2B, line 1) and 60°C (line 2), but the signals were normalized when returned to 20°C (line 3; cf. line of Fig. 2A), indicating refolding of the protein. Similarly, the SPR2 protein could be reversibly denatured following treatment with 5 M guanidine hydrochloride (Fig. 2C, line 2), but renatured on its removal (dotted line). We performed additional experiments with synthetic peptides. A 19-residue and a 25-residue peptide corresponding exactly to the head and tail domains of human SPR2 proteins generated only very weak CD signals, suggesting little structure (Fig. 2D, lines 1 and 2, respectively). On the other hand, peptides containing three (line 3) or six (line 4) nine-residue repeats characteristic of the central domain of human SPR2 proteins generated CD spectra typical of the intact recombinant protein. Together, these data suggest that only the central peptide repeat domain of SPR2 proteins possesses an organized structure. Furthermore, the reversible denaturation experiments imply that upon isolation from the bacteria the recombinant human SPR protein had folded into a stable configuration and is therefore appropriate for use in further biochemical experiments. The apparent absence of structural order of the end domain sequence motifs seems consistent with their preferred use in cross-linking by the TGases (see below).

Recombinant SPR2 Is a Complete Substrates For Epidermal
TGases-In order to begin to explore which TGase enzyme(s) might be involved in the cross-linking of SPR2 proteins, we used 35 S-labeled recombinant protein. Fig. 3 shows that three TGase enzymes expressed in stratified squamous epithelia cross-link the recombinant human SPR2 protein, indicating that it is a complete TGase substrate in the sense of providing both acyl donor Gln and acceptor Lys residues. Moreover, the data indicate that the three enzymes operate with clearly dif- ferent rates and patterns of reaction (Fig. 3). In this study, we used the same amount of enzyme activity for each enzyme and time (18 h), which together were determined in control studies to allow for complete reaction. The TGase 3 enzyme crosslinked the SPR2 protein most readily giving rise to short oligomeric products, and indeed, by excision of the 35 S-labeled bands from the SDS gels, we determined that 93% of the protein had been oligomerized. With the TGase 1 enzyme, 53% of the SPR2 protein remained as monomer, about 5% as short oligomers, and another 38% as very long oligomers that could not enter the SDS gel. In the case of TGase 2 cross-linking, about 50% remained as a monomer, 5% formed short oligomers, and another 40% formed large oligomers. By Western blotting analyses, however, this high molecular weight material in fact mostly was due to autocatalytic cross-linking of the SPR2 protein to the enzyme by use of its own surface Gln or Lys residues (Fig. 3). Similar phenomena have been reported previously for the TGase 2 (44, 45) and related factor XIIIa enzymes (46).
Kinetics of TGase 1, 2, and 3 Cross-linking of Recombinant SPR2-In order to obtain quantitative information on crosslinking of the recombinant human SPR2 protein with the three TGase enzymes, kinetic constants were determined. By using high concentrations of putrescine we have suppressed the possibility of putrescine oligomerization of SPR2. However, estimates of kinetic values are complicated by the fact that SPR2 protein serves as a complete substrate, and that especially in the case of TGase 3, multiple Gln residues are utilized (see below). Thus the values obtained represent average data for multiple Gln residues used. There were significant differences in the kinetic efficiencies (k cat /K m ) of the reactions between the three enzymes. The TGase 3 enzyme cross-linked the SPR2 protein 3-6 times more efficiently than TGases 1 or 2. In fact, these data for SPR2 are very similar to those for loricrin (41) ( Table I). We also determined some kinetic parameters for the TGase 1 and 3 enzymes using as substrates synthetic peptides corresponding to the head and tail domains of human SPR2 proteins (Table I). For the TGase 3 enzyme, the kinetic efficiency for the tail domain was 70% of that for the intact SPR2 protein, but for the TGase 1 enzyme, the efficiency was Ͻ10%. Likewise, the head domain peptide was used by TGase 3 with similar efficiency, and about 20 times more so than TGase 1. Together, these data are consistent with Fig. 3, and suggest that cross-linking by the TGase 3 enzyme is substantially more efficient. However, neither enzyme could incorporate significant amounts of putrescine into the synthetic peptides corresponding to central domain sequences (data not shown).
Amino Acid Sequencing Analyses of Cross-linking Reactions by the TGase Enzymes in Vitro-In order to obtain more specific information on the different proclivities of the three TGases for cross-linking the recombinant SPR2 protein, we next explored the nature and sequence locations of the crosslinks inserted in vitro. Following preparative-scale cross-linking reactions, the products were digested to completion with trypsin, and the peptides resolved by HPLC. By sequence analyses of diminished, shifted or new peaks (in comparison to non-cross-linked recombinant SPR2), we were able to obtain quantitative information on each of the Gln and Lys residues of the entire SPR2 protein.
The TGase 1 enzyme caused the disappearance of most of the peptide eluted at 22 min containing residues 1-8 (Fig. 4B). The major new peptide eluted at 24 min contained sequences 1-8, but with a Glu residue at position 6 instead of Gln. This probably arose due to partial reaction with the enzyme: the first step involves complex formation of a donor Gln residue with the active site; the second step involves transfer to an amine acceptor (such as a protein ⑀-NH 2 group of a Lys residue) to form an isopeptide bond (29,30). If released from the TGase enzyme before nucleophilic transfer, that is, transferred to water instead, the Gln residue is deamidated to Glu. This is consistent with the data of Table I that the TGase 1 reaction is of low efficiency. The second major new peak eluted at 25 min contained sequences of this peptide cross-linked to a dipeptide sequence Ser-Lys (residues 70, 71) through Lys-71. The majority of the other major peaks did not change. Note that the terminal tryptic peptides containing residues 66 -69, and 70 -71 were not resolved on this chromatogram. As mentioned about half of the latter was recovered in the 25 min peak. Several minor peptides were also apparent later in the chro-   (Fig. 4B), but the amounts were Ͻ5% of the total, indicating trace amounts only of cross-linking involving other Gln and Lys residues of the SPR2 protein. Thus only Gln-6 and Lys-71 of SPR2 are partially used (about 0.5 mol/mol) for crosslinking by the TGase 1 enzyme (Fig. 4D, upper numbers). With the TGase 3 enzyme, however, all but one peptide peak, eluted at 30 min, was shifted after cross-linking (Fig. 4C). This peptide contained residues 22-30 and 31-40 which are located in the central domain, was undiminished in amount, and its sequences were not encountered in the large unresolved peak at the end of the chromatogram, so that it was not used at all in cross-linking (Fig. 4C). Sequencing of the unresolved peak revealed the lost head domain and tail domain peptides (including the tryptic peptide containing terminal residues 70 -71) indicating that all of Lys and Gln residues used in crosslinking were located in the head and tail domains, and the amounts lost (3.1 mol in each domain) (Fig. 4D, lower numbers) were comparable to that measured as isolated isodipeptide. In this case, most of the Gln donor residues were located on the head domain, and acceptor Lys residues were located on the tail domain. In contrast to the TGase 1 enzyme, these data indicate that multiple Gln and Lys residues were used on each end domain simultaneously.
We also performed sequencing experiments on samples of the recombinant SPR2 protein that had been cross-linked for only 6 h by the TGase 1 or 3 enzymes. These data yielded nearly identical data for the utilization of Gln and Lys residues, except that the amounts were about 50%, indicating incomplete reaction (data not shown). This means there was no time-dependent preference for cross-linking sites.
For both the TGase 1 and 3 reactions at 6 and 18 h, we separated the monomer and oligomer bands by excision from SDS-polyacrylamide gels. After total enzymic proteolytic digestion, we found that Ͼ90% of the isodipeptide was present in the oligomeric bands (data not shown). That is, there was little evidence for intrachain cross-linking of the monomer band. However, we cannot rigorously exclude the possibility of some intrachain cross-linking of the oligomeric bands.
Preparative cross-linking with the TGase 2 enzyme was complicated by the finding of extensive cross-linking to the enzyme itself (Fig. 3), and because the total amount of cross-linking was very small in comparison to the other TGases. Thus we were unable to obtain useful sequence information for the very limited degree of this reaction.
Based on the percent utilization data of Fig. 4D, we can conclude that Ͼ90% of the cross-linking of the recombinant SPR2 protein is performed by the TGase 3 enzyme involving multiple head and tail domain residues. The TGase 1 enzyme used only two residues with high specificity although in modest amounts. We can conclude that the TGase 2 enzyme is probably not significantly involved in the cross-linking of SPR2.
Correlation of in Vitro and in Vivo Cross-linking of Recombinant Human SPR2-We have generated a large body of data of cross-links formed in vivo that were recovered in proteolysis experiments of CEs isolated from human foreskin epidermis (33,34) or cultured foreskin keratinocytes (36). 3 A total of 94 involved the several individual SPR2 proteins likely to be expressed in the epidermis 2 (Fig. 5B). In addition, we have described 34 peptides involving 45 occurrences of SPR2 sequences from mouse forestomach CEs (12) (Fig. 5C). Analysis of the mouse and human SPR2 data are valid as the locations of and sequences around the Gln and Lys residues on the head and tail domains have been conserved. Comparisons of these human and mouse in vivo data with those generated in the present in vitro experiments reveal close similarities in the patterns of utilization of the Gln and Lys residues on both head and tail domains. Furthermore, none involved central domain sequences. In effect, these data show that the recombinant SPR2 protein is treated the same way by the TGase 3 enzyme in vitro as occurs in vivo in human foreskin epidermis or cultured keratinocytes, and the forestomach epithelium. DISCUSSION Amino acid sequencing analyses have demonstrated previously that members of each of the three known classes of SPR proteins are widely used for the assembly of the CE of stratified squamous epithelia, a specialized structure which is of critical importance for barrier function (12,(33)(34)(35)(36). Although those studies have provided a wealth of information on which residues are used for cross-linking by TGases in vivo, and the potential structural and functional consequences of the observed cross-linking (36), it was not clear which enzyme(s) were involved, and whether the enzymes display substrate and or sequence specificities. To begin to address these questions, in this paper we have developed methods for the expression and purification of a member of the human SPR2 class, and have used it to explore in vitro its cross-linking properties by three of the TGases widely expressed in stratified squamous epithelia.
We have generated three types of data, which together suggest that the recombinant human SPR2 protein used here had correctly folded following isolation from bacteria. First, by CD analyses, the limited degree of structural order of the proteins could be reversibly recovered on addition and then removal of guanidine hydrochloride or heat treatment (Fig. 2). Second, we observed for the TGase 1 and 3 enzymes a high degree of sequence specificity of Gln and Lys residues employed for crosslinking, which would not have been expected for a denatured or improperly folded protein (40,47). Third, the usage of Gln and Lys residues in the recombinant SPR2 protein observed in vitro was similar to that observed previously in protein sequencing data of in vivo human and mouse CE samples (Fig. 5). Therefore, the data obtained from the in vitro experiments are likely to be biologically relevant in vivo. In this regard, we have found that several other epidermal structural proteins can be refolded following extraction and purification by use of denaturing or non-physiological solvents, including the keratin intermediate filament chains (48), filaggrin (49), loricrin (39), and now an SPR2 protein.
Our data of Fig. 3 indicate that the recombinant human SPR2 protein was used as a complete substrate by the TGase 1 and 3 enzymes present in the epidermis and other stratified squamous epithelia, but by the TGase 2 enzyme only very weakly, if at all. These data confirm earlier studies which have demonstrated that a variety of SPR proteins of different species serve as TGase substrates in in vitro reactions (5,8,11). Notably, however, we show here that the SPR2 protein was crosslinked most readily by the TGase 3 enzyme in vitro ( Fig. 3; Table I), a process that has not been examined by other investigators. Furthermore, we demonstrate in this study that the TGase 1 and 3 enzymes cross-link the SPR2 with significantly different rates and at different residue positions to form interchain oligomeric cross-linked products.
We found that all cross-linking in vitro in both the intact recombinant SPR2 protein (Figs. 3 and 4) as well as synthetic peptides (Table I) involved use of Gln and Lys residues located in head and tail domain sequences, in excellent agreement with in vivo studies (12,(33)(34)(35)(36) (Fig. 5). One possible explanation is that the head and tail domains have little or no organized structure (Fig. 2D) and thus those Gln and Lys residues are more accessible to the TGases, unlike the Gln and Lys residues Note that unlike human the various members of mouse SPR2 class vary in the numbers of central domain peptide repeats. Several are likely to be expressed in epidermal keratinocytes (see Footnote 3). Thus sequence positions in C are numbered from the conserved tail, denoted here as E. Moreover, the pattern of utilization of Gln and Lys residues of the recombinant SPR2 protein in vitro was remarkably similar to that seen in vivo for human foreskin epidermis and cultured human keratinocytes, as well as for mouse forestomach epithelium (Fig. 5). For virtually every head and tail domain Gln and Lys residue, the relative proportions were retained. Of the total utilization of the Gln6 residue, about 55% could be used by the TGase 1 enzyme, accounting for about Ϸ10% of the total Gln residue usage of SPR2. From these data, we may conclude therefore that almost all cross-linking of SPR2 in vivo is performed by the TGase 3 enzyme, with only a minor degree by TGase 1.
Because about 3 mol of cross-link/mol of SPR2 protein are inserted in vivo and in vitro, this means that multiple Gln and Lys residues were used on each end domain simultaneously. Most of the in vivo examples clearly revealed that the SPR2 proteins served as interchain cross-bridgers between other CE proteins. A few examples may have involved intrachain links between head and tail sequences within a single SPR2 protein, as well as interchain links to involucrin (36). While the TGase 3 enzyme formed short interchain oligomers (dimers-tetramers) in vitro (Fig. 3), we cannot exclude the possibility that some intrachain cross-links between head and tail domain sequences were formed as well.
Additional work now will be necessary to explore the crosslinking properties of the other members of the SPR family. Together with the present work, such data should allow further insights into the proposed role of SPRs as modulators of the biomechanical properties of the CEs and epithelia in which they are expressed (12).