A Single Polypeptide Acts Both as the @ Subunit of Prolyl4-Hydroxylase and as a Protein Disulfide-Isomerase*

A single polypeptide is shown to act both as the j3 subunit of the proline hydroxylase (EC 1.14.1 1.2) and as a protein disulfide-isomerase (EC 5.3.4.1). When isolated from chick embryos or rat liver, the /3 subunit of prolyl4-hydroxylase and the enzyme protein disul-fide-isomerase have identical molecular weights and peptide maps as produced by digestion with Staphylo-coccus aureus V8 protease. The apparent molecular weights of both proteins isolated from human placental tissue are slightly higher, and the human B subunit and one of its peptides have molecular weights about M, 500 higher than the protein disulfide-isomerase and its corresponding peptide. Experiments with polyclonal and monoclonal antibodies also suggest a structural identity between the two proteins. The B subunit isolated from the prolyl4-hydroxylase tetramer has pro- tein disulfide-isomerase activity similar to protein di-sulfide-isomerase itself, and even the @ subunit when present in the prolyl4-hydroxylase tetramer has one-half of this activity. Proline hydroxylase (EC 1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens by the hydroxylation of certain proline residues in peptide linkages. The active enzyme is a tetramer (aZpz) peroxidase Antibodies-Polyclonal antibodies to human placental prolyl 4- hydroxylase and to purified human placental and chick embryo protein disulfide-isomerase were prepared in rabbits (13). Two monoclonal antibodies to human placental prolyl 4-hydroxylase, termed 3G3 and 5B5, were specific for the @ subunit of the enzyme (22).

A single polypeptide is shown to act both as the j3 subunit of the proline hydroxylase (EC 1.14.1 1.2) and as a protein disulfide-isomerase (EC 5.3.4.1). When isolated from chick embryos or rat liver, the / 3 subunit of prolyl4-hydroxylase and the enzyme protein disulfide-isomerase have identical molecular weights and peptide maps as produced by digestion with Staphylococcus aureus V8 protease. The apparent molecular weights of both proteins isolated from human placental tissue are slightly higher, and the human B subunit and one of its peptides have molecular weights about M, 500 higher than the protein disulfide-isomerase and its corresponding peptide. Experiments with polyclonal and monoclonal antibodies also suggest a structural identity between the two proteins. The B subunit isolated from the prolyl4-hydroxylase tetramer has protein disulfide-isomerase activity similar to protein disulfide-isomerase itself, and even the @ subunit when present in the prolyl4-hydroxylase tetramer has onehalf of this activity.
Proline hydroxylase (EC 1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens by the hydroxylation of certain proline residues in peptide linkages. The active enzyme is a tetramer (aZpz) consisting of two types of monomer with molecular weights of about 64,000 ( a subunit) and 60,000 (p subunit). These monomers do not have prolyl4-hydroxylase activity (for reviews, see Refs. 1-4). The enzyme protein is found in cells not only as the active tetramer but also in a monomeric form that corresponds to the / 3 subunit in a number of molecular properties (1-7). This monomer has been shown to act as a precursor of the p subunit in the assembly of the enzyme tetramer (see Ref. 4). It has not been known, however, whether this monomeric protein also has some additional functions, since it is present in many cells in large excess over the active tetramers (1-7 rangement of disulfide bonds in various proteins in vitro (for reviews, see Refs. 8-10), including prolyl 4-hydroxylase (11) and type I and I1 procollagens (12). Protein disulfide-isomerase, with a reported molecular weight of about 57,000, has been suggested as the in vivo catalyst of disulfide bond formation (8-13). We have recently found' that the cDNA sequences of the p subunit of human prolyl 4-hydroxylase are very similar to those recently reported for rat protein disulfide-isomerase (14), the degree of homology between these two proteins being in the whole coding region 84% at the level of nucleotide sequences and 94% at the level of cDNA-derived amino acid sequences. Furthermore, Southern blot analyses of human genomic DNA with a cDNA probe for the prolyl4hydroxylase / 3 subunit indicated the presence of only one gene containing these sequences.' We here report comparisons between the / 3 subunit isolated from the prolyl4-hydroxylase tetramer and the purified protein disulfide-isomerase from the same species to study the structural and functional relationship of these two proteins. Our data suggest that a single polypeptide acts as both these proteins and thus the excess ( 3 subunit protein of prolyl 4-hydroxylase present in various cells (1-7) is the enzyme protein disulfide-isomerase.

EXPERIMENTAL PROCEDURES
Isolation of Enzymes-Human and chick embryo prolyl 4-hydroxylase was purified to homogeneity by affinity chromatography on poly(L-proline) linked to agarose followed by ion-exchange chromatography on DEAE-cellulose and gel filtration (15,16). To isolate the j3 subunit, the enzyme tetramer was incubated in the presence of 10 mM dithiothreitol at 37 "C for 30 min in 50 mM Tris-HC1 buffer, pH 7.5. The dithiothreitol was removed by centrifugation in Sephadex G-25 column (17) and the 6 subunit was isolated by concanavalin Aagarose chromatography (6). Protein disulfide-isomerase was isolated from human placentas and chick embryos by ammonium sulfate fractionation (5540%) followed by two ion-exchange chromatographies on CM-Sephadex C-50 and DEAE-Sephacel (10). Gel filtration on Sephacryl S-300 was used as a final purification step to give a homogeneous protein (11).
Assay of Enzyme Activities-Prolyl4-hydroxylase activity was assayed by measuring the stoichiometric decarboxylation of 2-oxo-[I-"Clglutarate in the reaction (see Ref. 18) and protein disulfideisomerase activity by the ribonuclease-linked method (10).
Electrophoretical Method.-SDS-PAGE2 was performed under reducing conditions using 10% acrylamide as the separating gel (19). Linear peptide maps from purified proteins were obtained using the method of Cleveland et al. (20). About 40 pg of chick embryo-or human placental prolyl4-hydroxylase and 20 pg of chick embryo-or human placental protein disulfide-isomerase were analyzed by SDS-PAGE and the gel was stained briefly with Coomassie Brilliant Blue. The j 3 subunit of prolyl 4-hydroxylase and the protein disulfideisomerase were then cut from the gel and placed in wells of a second SDS-PAGE consisting of a 4.5% stacking gel and a 16% separating gel. The slices were surrounded with a solution of 125 mM Tris-HC1 (pH 6.8), 0.1% SDS, and 20% glycerol and then overlayered with 20 pl of the above solution containing 10% glycerol, 0.04% bromphenol blue, and 4 pg of 5 ' . aureus V8 protease. Current was applied until the dye just passed the interphase between the stacking and separating gels and was continued after a pause of 45 min. The gels were then stained with either Coomassie Brilliant Blue (19) or silver (21) to visualize the peptides.

6447
Prolyl Hydroxylase p Subunit Is a Protein Disulfide-Isomerase from human placental tissue were separated by SDS-PAGE and blotted onto nitrocellulose, and the blot was incubated with antibodies. Bound antibody was recognized with anti-IgG antibody linked to peroxidase (see Refs. 13 and 22).
Antibodies-Polyclonal antibodies to human placental prolyl 4hydroxylase and to purified human placental and chick embryo protein disulfide-isomerase were prepared in rabbits (13). Two monoclonal antibodies to human placental prolyl 4-hydroxylase, termed 3G3 and 5B5, were specific for the @ subunit of the enzyme (22).

RESULTS AND DISCUSSION
The molecular weights, as determined by SDS-PAGE, for the prolyl 4-hydroxylase @ subunit and the protein disulfideisomerase polypeptide chain were found to be both slightly higher (by about M , 1500) in the case of human placental tissue than in chick embryos (Fig. 1A). Protein disulfideisomerase and the prolyl 4-hydroxylase @ subunit purified from rat liver had the same apparent molecular weight as their counterpart from chick embryos (not shown). The molecular weights determined for the two polypeptides from chick embryo and rat liver were identical, while those from human placental tissue differed by about M, 500, protein disulfide-isomerase having a slightly lower molecular weight (Fig. 1A). The peptide maps for the two peptides from chick embryos produced by digestion with S. aureus V8 protease were identical and those from human placental tissue almost identical, except for one peptide which consistently showed a slightly lower apparent molecular weight (by about M, 500) in protein disulfide-isomerase than in the @ subunit (Fig. 1B).
The reasons for this minor difference are unknown, but it may be due to differences in glycosylation or some other posttranslational modification of the proteins in the placental tissue.
Polyclonal antibodies prepared against the human prolyl4hydroxylase stained human protein disulfide-isomerase on Western blots and those prepared to human and chick embryo protein disulfide-isomerase stained the human @ subunit (Fig.  2). In addition, two different monoclonal antibodies to the human @ subunit stained the human protein disulfide-isomerase (Fig. 2). Polyclonal antibodies to human prolyl4-hydroxylase, immunoadsorbed by human protein disulfide-isomerase, did not stain either human protein disulfide-isomerase or the @ subunit of human prolyl 4-hydroxylase on Western blots. This result demonstrates that there were no @ subunitspecific epitopes which were not shared by protein disulfideisomerase and thus supports the view of a structural identity between the two proteins.
Various polyclonal and monoclonal antibodies which were inhibitory against the activity of the human prolyl4-hydroxylase tetramer did not inhibit the activity of human protein disulfide-isomerase, even though the ratio of concentration of the antibody to concentration of the enzyme protein was higher in the cases of the protein disulfide-isomerase than prolyl4-hydroxylase activity assays (Table I). Such antibodies included two polyclonal antibodies prepared to either human or chick protein disulfide-isomerase that slightly inhibited the activity of human prolyl 4-hydroxylase (Table I)   analyzed in separate gels. A, prolyl4-hydroxylase from chick embryos ( l a n e 2 ) and from human placenta ( l a n e 5) and protein disulfideisomerase from chick embryos ( l a n e 3 ) and from human placenta ( l a n e 6) were run in SDS-polyacrylamide gel electrophoresis under reducing conditions using 10% acrylamide as the separating gel and stained with Coomassie Brilliant Blue. M, markers (lanes 1 and 4 ) are shown in the figure. B, linear peptide map from the @ subunit of chick embryo prolyl 4-hydroxylase ( l a n e 21, chick embryo protein disulfide-isomerase ( l a n e 3), the ,fI subunit of human placental prolyl 4-hydroxylase ( l a n e 4 ) and human placental protein disulfide-isomerase ( l a n e 5). The proteins from human tissue were stained with Coomassie Brilliant Blue and those from chick embryos by silver. The arrow shows a polypeptide which has a slightly higher M , in the human ,fI subunit than in the human protein disulfide-isomerase. M, markers (lanes 1 and 6) are also shown.

2-
Thus the antibodies may bind at various sites of this large catalytic center. The catalytic site of protein disulfide-isomerase is simpler and probably comprises a small area of the polypeptide chain (14). Accordingly, only antibodies which bind to this small area may inhibit the protein disulfideisomerase activity.
A mixture of the a and @ subunits of the dissociated human prolyl4-hydroxylase tetramer and the pure @ subunit isolated from such samples had protein disulfide-isomerase activity (Table 11), which was in both cases identical in amount to that of the protein disulfide-isomerase when expressed per mg of the @ subunit. The intact prolyl4-hydroxylase tetramer a Protein Disulfide-Isomerase 6449

4-hydroxylase
Protein disulfide-isomerase activity was assayed as described under "Experimental Procedures," and 120 ng of each protein was used in the assay. Bovine serum albumin was used as a control. The values are given as means f S.D., and the number of observations is shown in parentheses.

FIG.
2. Western blot analysis of the @ subunit of prolyl 4hydroxylase and protein disulfide-isomerase. The odd-numbered lanes contained human prolyl 4-hydroxylase and the evennumbered lanes human protein disulfide-isomerase. Lanes I and 2 were incubated with monoclonal antibody 5B5 and lanes 3 and 4, with monoclonal antibody 3G3, both to human prolyl4-hydroxylase; lanes 5 and 6, with polyclonal antibody to human protein disulfideisomerase, lanes 7 and 8, with polyclonal antibody to chick embryo protein disulfide-isomerase; and lanes 9 and 10, with polyclonal antibody to human prolyl 4-hydroxylase. The minor band below protein disulfide-isomerase in lane 6 is a minor contaminant often present in human placental protein disulfide-isomerase.

TABLE I Effect of prolyl 4-hydroxyhe and protein disulfide-isomerase
antibodies on the enzyme activities Prolyl 4-hydroxylase and protein disulfide-isomerase activities were measured as described under "Experimental Procedures": 100 ng of human prolyl4-hydroxylase and 5 pg of antibody IgG were used in the prolyl 4-hydroxylase assay, while 150 ng of human protein disulfide-isomerase and 25 pg of antibody IgG were used in the protein disulfide-isomerase assay. The enzymes were incubated with antibody for 5 min a t 30 "C before assaying the activity. The enzyme activities are expressed as percentages of the values obtained with the corresponding amounts of nonimmune IgG. was also found to have protein disulfide-isomerase activity, but in this case about one-half of that recorded above when expressed per mg of the 0 subunit (Table 11). It has been proposed that protein disulfide-isomerase has two catalytic siteslone polypeptide chain, both catalytic sites involving a Trp-Cys-Gly-His-Cys-Lys sequence (14). These data suggest that one of the two catalytic sites of the protein disulfide- isomerase activity is obscured in the prolyl 4-hydroxylase tetramer structure.
Our results demonstrate that a single polypeptide possesses two entirely different enzymatic functions. The polypeptide has protein disulfide-isomerase activity both when it is present as the monomer and when it is present in the prolyl 4hydroxylase tetramer. In the latter case it also catalyzes the 4-hydroxylation of proline residues, contributing together with the a subunit to the structure of the catalytic center of the prolyl4-hydroxylase tetramer.