Cloning of the Human Prolyl 4-Hydroxylase α Subunit Isoform α(II) and Characterization of the Type II Enzyme Tetramer

Prolyl 4-hydroxylase (proline hydroxylase, EC1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens. The vertebrate enzyme is an α2β2tetramer, the β subunit of which is identical to protein disulfide-isomerase (PDI, EC 5.3.4.1). We report here on cloning of the recently discovered α(II) subunit from human sources. The mRNA for the α(II) subunit was found to be expressed in a variety of human tissues, and the presence of the corresponding polypeptide and the (α(II))2β2 tetramer was demonstrated in cultured human WI-38 and HT-1080 cells. The type II tetramer was found to represent about 30% of the total prolyl 4-hydroxylase in these cells and about 5–15% in various chick embryo tissues. The results of coexpression in insect cells argued strongly against the formation of a mixed α(I)α(II)β2 tetramer. PDI/β polypeptide containing a histidine tag in its N terminus was found to form prolyl 4-hydroxylase tetramers as readily as the wild-type PDI/β polypeptide, and histidine-tagged forms of prolyl 4-hydroxylase appear to offer an excellent source for a simple large scale purification of the recombinant enzyme. The properties of the purified human type II enzyme were very similar to those of the type I enzyme, but theK i of the former for poly(l-proline) was about 200–1000 times that of the latter. In agreement with this, a minor difference, about 3–6-fold, was found between the two enzymes in the K m values for three peptide substrates. The existence of two forms of prolyl 4-hydroxylase in human cells raises the possibility that mutations in one enzyme form may not be lethal despite the central role of this enzyme in the synthesis of all collagens.

Prolyl 4-hydroxylase (proline hydroxylase, EC 1.14.11.2) catalyzes the hydroxylation of proline in -Xaa-Pro-Gly-triplets in collagens and other proteins with collagen-like sequences. The enzyme plays a central role in the synthesis of all collagens, as the 4-hydroxyproline residues formed in the reaction are essential for the folding of the newly synthesized collagen polypeptide chains into triple helical molecules. The vertebrate enzyme is an ␣ 2 ␤ 2 tetramer in which the ␣ subunits contribute to most parts of the two catalytic sites (for reviews, see Refs. [1][2][3]. The ␤ subunit is identical to the enzyme protein disulfide-isomerase (PDI, EC 5.3.4.1) 1 and has PDI activity even when present in the prolyl 4-hydroxylase tetramer (4 -6). The PDI polypeptide also has several other functions (1-3, 7, 8).
Prolyl 4-hydroxylase had long been assumed to be of one type only, with no isoenzymes (1-3), but recently an isoform of the ␣ subunit, termed the ␣(II) subunit, was cloned from mouse tissues (9). Correspondingly, the previously known ␣ subunit is now called the ␣(I) subunit. The ␣(II) subunit was found to form an (␣(II)) 2 ␤ 2 tetramer with the PDI/␤ subunit when the two polypeptides were coexpressed in insect cells. The properties of the new type II enzyme were found to be very similar to those of the type I tetramer, with the distinct difference that it was inhibited by poly(L-proline) only at very high concentrations (9).
The ␣ subunit of prolyl 4-hydroxylase cloned from the nematode Caenorhabditis elegans (10) has been found to have features of both types of mouse ␣ subunit, suggesting that C. elegans may have only one type of ␣ subunit (9). This forms active prolyl 4-hydroxylase in insect cell coexpression experiments with either the C. elegans or the human PDI/␤ polypeptide, but surprisingly, the enzymes containing the C. elegans ␣ subunit are ␣␤ dimers (10,11).
We report here that the existence of ␣(II) subunit mRNA is not limited to the mouse, as a corresponding mRNA is expressed in a variety of human tissues. All the data so far available on the existence of the type II prolyl 4-hydroxylase tetramer are based on insect cell coexpression experiments, but we now demonstrate that this enzyme is indeed present in cultured human fibroblasts and represents about 30% of their total prolyl 4-hydroxylase activity. We also studied whether the ␣(I) and ␣(II) subunits can form a mixed ␣(I)␣(II)␤ 2 tetramer, and whether any differences are found between the type I and II enzymes in their K m values for various peptide substrates, as the two mouse enzymes differ so markedly from each other with respect to inhibition by poly(L-proline). A new affinity purification procedure was developed that is based on the use of a histidine tag in the N terminus of the PDI/␤ polypeptide, and this makes it possible to obtain large amounts of any form of the recombinant enzyme by very simple steps.

MATERIALS AND METHODS
Isolation of cDNA Clones-Screening of a human lung gt10 cDNA library (CLONTECH) with BT 14.1, a cDNA clone for the mouse ␣(II) subunit (9), as a probe yielded one positive clone, H9, among 600,000 recombinants. Rescreening of the same library with H9 as a probe gave 8 positive clones out of 600,000 recombinants. Four of them, L121, L142, L21, and L22, were characterized further.
Nucleotide Sequencing, Sequence Analysis, and Northern Blot Analysis-The nucleotide sequences were obtained by the dideoxynucleotide chain termination method (12) with T7 DNA polymerase (Pharmacia). Vector-specific or sequence-specific 17-mer primers synthesized in an Applied Biosystems DNA Synthesizer (Dept. of Biochemistry, University of Oulu) were used, and the sequences were determined for both strands. DNASIS and PROSIS version 6.00 sequence analysis software (Pharmacia) was used to compile the sequence data.
A human multitissue Northern blot (CLONTECH) containing 2 g of poly(A) ϩ RNA per sample isolated from various human tissues was hybridized under the stringent conditions suggested in the manufacturer's instructions. The probes used were 32 P-labeled cDNA clones encoding either the whole human ␣(II) subunit or PA-58 and PA-59 (13) encoding almost all of the human ␣(I) subunit. The autoradiography time was 3 days.
Construction of Baculovirus Transfer Vectors and Generation of Recombinant Baculoviruses-To construct a baculovirus transfer vector for the human ␣(II) subunit, a 5Ј fragment was amplified from the -DNA of L142. The cDNA-specific primers used were ␣IIiN5Ј (5Ј-TCAGGCGGCCGCGACAGCCAGACACTTCCCTC-3Ј), containing an artificial NotI site, and ␣IIiE3Ј (5Ј-GGTGAAGAATTCGGCCTGCAC-3Ј), containing a natural EcoRI site. Polymerase chain reaction was performed under the conditions recommended by the supplier of the Taq polymerase (Promega), and the reactions were cycled 27 times as follows: denaturation at 94°C for 1 min, annealing at 62°C for 1 min, and extension at 72°C for 3 min. The product was digested with NotI and EcoRI restriction enzymes to give a fragment that extended from nucleotide 166 to 251. The next fragment was prepared by digesting L142 with EcoRI and SacI. The resulting EcoRI-SacI fragment, covering nucleotides 252-873, was ligated together with the above NotI-EcoRI fragment to the NotI-SacI site of the pBluescript vector (Stratagene) and termed HuNS-BS. Clone H9 was digested with SacI and PstI to give a fragment encompassing nucleotides 874 -1426. Clone L21 was digested with PstI, and the resulting fragment, covering nucleotides 1427-2021, was ligated together with the SacI-PstI fragment to the SacI-PstI site of pBluescript, and the construct was termed I3ЈSPP-BS. The NotI-SacI fragment from HuNS-BS, and the SacI-PstI fragment from I3ЈSPP-BS were then ligated to the NotI-PstI site of pBluescript, and the construct was termed ␣(II)human-BS. Finally, the ␣(II)human-BS was digested with NotI and EcoRV, and the resulting fragment was ligated to the NotI-SmaI site of the transfer vector pVL1392 (14).
A histidine affinity tag was generated in the N terminus of the human PDI/␤ polypeptide. Six histidine codons were created by polymerase chain reaction downstream of the codon coding for the last amino acid of the signal peptide in a full-length cDNA for the human PDI/␤ polypeptide (4). The cDNA was digested with EcoRI and BamHI and ligated to pVL1392.
Spodoptera frugiperda Sf9 insect cells (Invitrogen) were cultured in TNM-FH medium (Sigma) supplemented with 10% fetal bovine serum (BioClear) at 27°C, either as monolayers or in suspension in spinner flasks (Techne Laboratories, Princeton, NJ). The pVL constructs were cotransfected into Sf9 insect cells with a modified Autographa californica nuclear polyhedrosis virus DNA using the BaculoGold transfection kit (Pharmingen). The resultant viral pools were collected 4 days later, amplified, and plaque-purified (15). The recombinant viruses were termed ␣(II) and His-PDI. The viruses human ␣59(I), coding for the ␣(I) subunit, and human PDI/␤, coding for the PDI/␤ polypeptide, have been described previously (16).
Expression and Analysis of Recombinant Proteins-Sf9 insect cells were cultured as above. To produce an enzyme tetramer, the ␣59(I) or ␣(II) virus and the PDI/␤ or His-PDI virus were used in a ratio of 1:1 or 2:1, and when attempting to produce an ␣(I)␣(II)␤ 2 tetramer, the ␣59(I), ␣(II), and PDI/␤ viruses were used in the ratio 1:1:2 or 1:1:1. The cells were harvested 72 h after infection, washed with a solution of 0.15 M NaCl and 0.02 M phosphate, pH 7.4, homogenized in a solution of 0.1 M glycine, 0.1 M NaCl, 10 M dithiothreitol, 0.1% Triton X-100, and 0.01 M Tris, pH 7.8, and centrifuged at 10,000 ϫ g for 20 min. The resulting supernatants were analyzed by 8% SDS-PAGE or nondenaturing 7.5% PAGE and assayed for enzyme activity. The cell pellets were further solubilized in 1% SDS and analyzed by 8% SDS-PAGE. Western blot analysis was performed with a monoclonal antibody L7B to the human ␣(I) subunit or a monoclonal antibody K4 to the mouse ␣(II) subunit. These monoclonal antibodies were generated by immunizing mice with denatured recombinant ␣(I) or ␣(II) polypeptides that had been purified by SDS-PAGE. The antibodies recognize the ␣(I) and ␣(II) subunit isoforms, respectively, both as native and denatured proteins from man, mouse, and chicken, but show no cross-reactivity between isoforms. NaCl, 50 M dithiothreitol, 0.1% Triton X-100, 0.01% soybean trypsin inhibitor, and 20 mM Tris-HCl, pH adjusted to 7.5 at 4°C, and centrifuged at 10,000 ϫ g for 30 min. The resulting supernatants were analyzed by nondenaturing 8% PAGE followed by Western blotting using ECL. Total prolyl 4-hydroxylase activity was measured in aliquots of the supernatants, and the type II enzyme activity in aliquots of the supernatants that had been passed through small poly(Lproline) columns.
Assay of Type II Prolyl 4-Hydroxylase Activity in Tissues of 17-Dayold Chick Embryos-Calvaria, sternum, tendon, and liver tissues from 17-day-old chick embryos and whole chick embryos were homogenized in a solution of 0.1 M glycine, 0.2 M NaCl, 50 M dithiothreitol, 0.1% Triton X-100, 0.01% soybean trypsin inhibitor, and 20 mM Tris-HCl buffer, pH adjusted to 7.5 at 4°C, and centrifuged at 10,000 ϫ g for 30 min. Aliquots of the supernatants were then used to assay the enzyme activities as above.
Protein Purification and N-terminal Amino Acid Sequence Analysis-The type II prolyl 4-hydroxylase resulting from coinfection with the viruses ␣(II) and PDI/␤ was first purified by a procedure consisting of anion exchange chromatography on a DEAE-cellulose column (DE52 Whatman) and two gel filtrations. . The 3-ml fractions containing most of the prolyl 4-hydroxylase activity were pooled, concentrated by ultrafiltration (Millipore), and applied to a 1.5 ϫ 90-cm Ultrogel AcA34 (Serva) column, equilibrated, and eluted with a solution of 0.1 M glycine, 0.25 M NaCl, 10 M dithiothreitol, and 0.01 M Tris, pH 7.8. Fractions of 3.5 ml were collected, and their absorbance at 280 nm was measured. Several fractions were analyzed by 8% SDS-PAGE and assayed for prolyl 4-hydroxylase activity, and a sample pool was combined and concentrated as above for a second gel filtration on a 2.5 ϫ 95-cm Bio-Gel A-0.5m fine (Bio-Rad) column equilibrated and eluted as above. Fractions of 2.5 ml were collected, analyzed by 8% SDS-PAGE, and assayed for prolyl 4-hydroxylase activity.
The type I and II prolyl 4-hydroxylase tetramers and the PDI/␤ polypeptide containing the histidine affinity tag were also purified by a procedure consisting of a metal chelate affinity chromatography and a gel filtration step. Insect cells expressing type I or type II prolyl 4-hydroxylase or the His-PDI polypeptide were harvested 72 h after infection, washed with a solution of 0.15 M NaCl and 0.02 M phosphate, pH 7.4, homogenized in a 0.3 M NaCl, 0.1 M glycine, 0.1% Triton X-100, and 0.01 M Tris buffer, pH 7.8, and centrifuged at 10,000 ϫ g for 20 min. The supernatant was applied to a ProBond column (Invitrogen) equilibrated with a 0.3 M NaCl, 0.1 M glycine, and 0.01 M Tris buffer, pH 7.8. Unbound material was washed off with the same buffer, and the bound proteins were eluted with a 100 mM imidazole, 0.3 M NaCl, 0.1 M glycine, 10 M dithiothreitol, and 0.01 M Tris buffer, pH 7.8. Fractions containing the eluted proteins were pooled and applied to a 2.5 ϫ 95-cm Bio-Gel A-0.5m fine (Bio-Rad) gel filtration column, equilibrated, and eluted with a 0.3 M NaCl, 0.1 M glycine, 10 M dithiothreitol, and 0.01 M Tris buffer, pH 7.8. Fractions of 3 ml were collected, their absorbance at 230 and 280 nm was measured, and they were analyzed by 8% PAGE under nondenaturing conditions.
Fractions containing the prolyl 4-hydroxylase tetramer or the PDI/␤ dimer/monomer were pooled and analyzed by 8% SDS-PAGE and 7.5% nondenaturing PAGE. Protein concentrations were estimated with the Bio-Rad protein assay kit (Bio-Rad) according to the manufacturer's instructions.
For the N-terminal sequencing the type II prolyl 4-hydroxylase was run on 8% SDS-PAGE under reducing conditions and blotted onto a polyvinylidene difluoride membrane. The membrane was stained with 0.1% Coomassie R-250 in 50% methanol, and the band corresponding to the ␣(II) subunit was cut off. The N-terminal sequence was determined using an Applied Biosystems model 477A on-line 120 liquid pulse protein sequencer.
Enzyme Activity Assays-Prolyl 4-hydroxylase activity was assayed by a method based either on the hydroxylation-coupled decarboxylation of 2-oxo[1-14 C]glutarate or on the formation of hydroxy[ 14 C]proline in protocollagen, a biologically prepared, [ 14 C]proline-labeled protein substrate consisting of nonhydroxylated pro-␣ chains of chick type I procollagen (17). K m values were determined by varying the concentration of one substrate in the presence of fixed concentrations of the second while the concentrations of the other substrates were held constant (18). The pure type II prolyl 4-hydroxylase tetramer was used as an enzyme source for the K m determinations.

RESULTS
Isolation of cDNA Clones-To isolate cDNA clones for the human ␣(II) subunit, a human lung cDNA library was screened using BT14.1, a cDNA for the mouse ␣(II) subunit (9), as a probe. One positive clone, H9, which codes for the central region of the cDNA, was obtained. Rescreening of the same library with H9 as a probe gave 8 positive clones, 4 of which were characterized further.
Amino Acid Sequence of the Human ␣(II) Subunit and Its Comparison with the Human ␣(I) and Mouse ␣(II) Subunit Sequences-The cDNA clones cover 2194 nucleotides of the corresponding mRNA (these cDNA sequences are not shown, but have been deposited in the GenBank /EMBL Data Bank with accession number U90441) and encode a 535-amino acid polypeptide, including a putative signal peptide of 21 residues (Fig. 1). Sequencing of the N terminus of the polypeptide indicated that its first amino acid is glutamate (Fig. 1). The first amino acid of the mouse ␣(II) subunit was previously thought to be tryptophan (9), based on the computational parameters of von Hejne (19), but it now seems more likely that this is likewise glutamate, as such a result would also be compatible with the computational parameters.
The human ␣(II), human ␣(I), and mouse ␣(II) polypeptides are very similar in size, ␣(I) being three amino acids longer (Fig. 1). All three polypeptides contain two potential attachment sites for asparagine-linked oligosaccharides, but the position of the C-terminal site in the ␣(II) and ␣(I) subunits differs by 4 residues (Fig. 1). The five cysteine residues present in the human, mouse, and chick ␣(I) and the C. elegans ␣ subunit are all conserved in the human ␣(II) subunit, but the latter contains an additional cysteine, which is also present in the mouse ␣(II) subunit (Fig. 1).
The derived amino acid sequence of the human ␣(II) subunit shows 64% identity and 81% similarity to the sequence of the human ␣(I) subunit and 93% identity and 96% similarity to the sequence of the mouse ␣(II) subunit. The identity of the subunits is not distributed equally, being highest within the Cterminal domains (Fig. 1).
Expression of the ␣(II) and ␣(I) Subunit mRNAs in Various Human Tissues-Expression of the two types of ␣ subunit mRNA in various human tissues was studied by Northern hybridization. The sizes of the mRNAs for the human ␣(II) and ␣(I) subunits are 2.3 and 3.0 kilobases, respectively (Fig. 2). The mRNA for the ␣(II) subunit was found to be expressed in a variety of tissues, but distinct differences were found relative to the expression pattern of that for the ␣(I) subunit, in that the relative expression level of the latter was much higher in the skeletal muscle, liver and kidney (Fig. 2).
The ␣(I) and ␣(II) Subunits Do Not Form ␣(I)␣(II)␤ 2 Tetramers-To study whether the ␣(I) and ␣(II) subunits can form ␣(I)␣(II)␤ 2 mixed tetramers in addition to the (␣(I)) 2 ␤ 2 type I and (␣(II)) 2 ␤ 2 type II tetramers (1)(2)(3)9), both types of ␣ subunit were expressed in insect cells together with the PDI/␤ polypeptide. The (␣(I)) 2 ␤ 2 tetramer is effectively inhibited by poly(Lproline) (1, 2) and becomes bound to poly(L-proline) affinity columns (20). The (␣(II)) 2 ␤ 2 tetramer differs distinctly from the type I enzyme in that it is inhibited by poly(L-proline) only at very high concentrations (9). It thus can be expected that the type II enzyme will not become bound to poly(L-proline) columns. If a mixed ␣(I)␣(II)␤ 2 enzyme existed, it could be expected either to become bound to poly(L-proline), due to the presence of the ␣(I) subunit, or to remain unbound.
Sf9 insect cells were coinfected with baculoviruses coding for either the ␣(I) subunit or the ␣(II) subunit together with viruses coding for the PDI/␤ polypeptide, and a third set of cells was infected with all three viruses. The cells were harvested 72 h after infection, and Triton X-100-soluble proteins were analyzed by PAGE performed under nondenaturing conditions. When the cells were coinfected with viruses coding for either the ␣(I) subunit or the ␣(II) subunit together with a virus coding for the PDI/␤ polypeptide, a type I or type II enzyme tetramer was formed, the mobilities of these two types of tetramer being essentially identical (Fig. 3A, lanes 1 and 4). An enzyme tetramer was likewise formed when the cells were infected with all three viruses (Fig. 3A, lane 7).
Western blotting with monoclonal antibodies specific to the isolated ␣(I) and ␣(II) subunits was used to distinguish between the types of tetramer. The antibody to the ␣(I) subunit stained the type I tetramer (Fig. 3B, lane 1) but not the type II tetramer (Fig. 3B, lane 4), whereas the antibody to the ␣(II) subunit stained the type II tetramer (Fig. 3C, lane 4) but not the type I tetramer (Fig. 3C, lane 1).
The type I tetramer became efficiently bound to a poly(Lproline) affinity column, as no enzyme could be detected in the column effluent (Fig. 3, A and B, lanes 2), and could be eluted with poly(L-proline) (Fig. 3, A and B, lanes 3). The type II tetramer was found in the column effluent (Fig. 3, A and C, lanes 5), and no additional amounts could be eluted from the column with poly(L-proline) (Fig. 3, A and C, lanes 6). When the tetramer formed during infection with viruses coding for both types of ␣ subunit was studied as above, a Coomassie-stained enzyme band was seen in both the column effluent (Fig. 3A, lane 8) and the eluate (Fig. 3A, lane 9). The band in the column effluent could be stained by the antibody to the ␣(II) subunit (Fig. 3C, lane 8) but not that to ␣(I) (Fig. 3B, lane 8), thus ruling out the presence of ␣(I)␣(II)␤ 2 in the effluent. The band in the column eluate could be stained by the antibody to the ␣(I) subunit (Fig. 3B, lane 9) but not to ␣(II) (Fig. 3C, lane 9) thus ruling out the presence of ␣(I)␣(II)␤ 2 in the eluate. The absence of the ␣(I)␣(II)␤ 2 tetramer cannot be due to the low level of infection of Sf9 cells by three viruses simultaneously, as we have recently shown that these cells can be efficiently infected by three viruses (21). -1080) were homogenized, and Triton X-100soluble proteins were analyzed by PAGE performed under nondenaturing conditions followed by Western blotting using ECL with monoclonal antibodies specific to the ␣(II) and ␣(I) subunits (as demonstrated above). Bands corresponding to both types of enzyme tetramer were found in extracts from both cell types (Fig. 4). Quantification of the Western blots using ECL which had been standardized against known amounts of both types of enzyme tetramer indicated that the amount of type II enzyme is lower than that of type I enzyme in both cell types, being about 30% of total prolyl 4-hydroxylase (details not shown).

Ratio of Type II Enzyme Activity to Total Prolyl 4-Hydroxylase Activity in Human WI-38 Fibroblasts and HT-1080 Cells
and Certain Chick Embryo Tissues-Total prolyl 4-hydroxylase activity, i.e. the sum of the type I and type II enzyme activities, was measured in Triton X-100 extracts from cell or tissue homogenates using proline-labeled protocollagen as a substrate. The activity of the type II enzyme was measured by passing aliquots of Triton X-100 extracts through poly(L-proline) affinity columns and determining the enzyme activity in column effluents. The values were corrected for dilution, and the type I enzyme activity was estimated by subtracting the type II activity from the total activity. The type II enzyme activity was found to represent about 30% of the total enzyme activity in confluent cultures of human lung fibroblasts and HT-1080 cells (details not shown). The corresponding percentage in Triton extracts from homogenates of 17-day-old whole chick embryos was 10%, and the percentages in Triton extracts from homogenates of chick embryo tissues were 13% in calvaria, 8% in sternum, 10% in tendon, and 5% in liver (details not shown).
Purification of the Human Type II Prolyl 4-Hydroxylase Tetramer-As the type II enzyme does not become bound to poly(Lproline), this enzyme could not be purified by the standard affinity column procedure (20,22) developed for type I. It was therefore initially purified using an ion-exchange chromatography procedure consisting of DEAE-cellulose chromatography and two gel filtrations. The enzyme purified by this procedure was pure as judged by Coomassie staining of 8% SDS-PAGE performed under reducing conditions (Fig. 5A) and 7.5% PAGE performed under nondenaturing conditions (Fig. 5B).
To make the purification simpler and more efficient, a histidine affinity tag was constructed to the N terminus of the mature PDI/␤ polypeptide. Sf9 cells were then coinfected with viruses coding for either the ␣(I) or the ␣(II) subunit and the His-PDI polypeptide, and the Triton X-100-soluble proteins were studied by PAGE performed under nondenaturing conditions. Both types of ␣ subunit were found to form an enzyme tetramer with the His-PDI polypeptide (Fig. 6, A and B, lanes 1) as efficiently as with the wild-type PDI/␤ polypeptide ( Table  I). The Triton X-100 extracts were then applied to a ProBond column (Invitrogen), the unbound material was washed off with the column equilibration buffer, and the bound proteins were eluted with imidazole. No enzyme tetramer was found in the flow-through fractions (Fig. 6, A and B, lanes 2), whereas both types of enzyme tetramer and the His-PDI dimer and monomer were present in the column eluates (Fig. 6, A and B,  lanes 3). The enzyme tetramers (Fig. 6, A and B, lanes 4) could then be separated from the His-PDI polypeptide by gel filtration, during which the latter was converted from dimers to monomers (Fig. 6A, lane 5). The final preparations of both types of enzyme tetramer were also pure when analyzed by SDS-PAGE and Coomassie staining (Fig. 6C).
Catalytic Properties of the Human Type II Prolyl 4-Hydroxylase Tetramer-In agreement with data on mouse type II prolyl 4-hydroxylase (9), the K m values of the human type II enzyme for Fe 2ϩ , 2-oxoglutarate, and ascorbate were identical to those of the human type I enzyme (Table II). A marked difference was found between the two human enzymes in inhibition by poly(L-proline); however, the K i values of the type II enzyme for poly(L-proline), M r 7700, being about 200 times greater than those of the type I enzyme and those for M r 44,000 about 1000 times greater (Table II). Small but significant differences were found between the human type II and type I enzymes in their K m values for three peptide substrates, in that all these values were 3-6 times greater in the case of the type II enzyme than for the type I enzyme (Table II). Additional experiments (details not shown) demonstrated that the type II enzyme, like the type I enzyme, did not catalyze any formation

FIG. 5. Analysis of the purified human type II prolyl 4-hydroxylase by SDS-PAGE under reducing conditions (A) and PAGE under nondenaturing conditions (B).
The type II enzyme shown in lanes 2 was purified by a procedure consisting of DEAE-cellulose chromatography and two gel filtrations, and the human type I enzyme shown in lanes 1 by a poly(L-proline) affinity column procedure (20,22). Both gels were analyzed by Coomassie staining. The locations of the ␣ and ␤ subunits on the SDS-PAGE are shown by the short and long arrows, respectively. The human type I and type II prolyl 4-hydroxylase tetramers are indicated by the arrow ␣ 2 ␤ 2 . of 3-hydroxyproline when [ 14 C]proline-labeled protocollagen was used as a substrate and the hydrolyzed reaction products were separated using an amino acid analyzer as described previously (17). DISCUSSION The data reported here indicate that the existence of an mRNA for the ␣(II) subunit of prolyl 4-hydroxylase is not limited to the mouse (9), as an mRNA coding for a highly similar ␣(II) subunit was also found in human tissues. Furthermore, the present data indicate that the ␣(II) subunit is translated into the corresponding polypeptide in human cells. The ␣(II) subunit mRNA was found to be expressed in a variety of tissues, but distinct differences were found in the expression patterns of the ␣(II) and ␣(I) subunit mRNAs between tissues.
Quantification of the proportions of the two types of prolyl 4-hydroxylase tetramer by Western blotting in cultured human WI-38 lung fibroblasts and HT-1080 fibrosarcoma cells indicated that the type II tetramer represents about 30% of the total enzyme protein in these two cell types. Correspondingly, about 30% of the total prolyl 4-hydroxylase activity in extracts from these two cell types was found in the flow-through fractions of poly(L-proline) affinity columns, suggesting that the (␣(II)) 2 ␤ 2 tetramer represents about 30% of the total prolyl 4-hydroxylase activity. The type II prolyl 4-hydroxylase is also likely to be present in chick embryo tissues, as a fraction of the total enzyme activity was found to pass through the poly(Lproline) affinity column in the case of all the chick embryo tissues studied. Nevertheless, the proportion of type II enzyme activity may be lower in chick embryo tissues than in human fibroblasts, about 5-15% of total prolyl 4-hydroxylase activity. This percentage agrees with early reports indicating that up to at least 80% of the total prolyl 4-hydroxylase activity present in crude extracts from whole chick embryos is bound to a poly(Lproline) affinity column (20), and that at least 80% of the total prolyl 4-hydroxylase activity present in extracts from whole chick embryo homogenates is inhibited by poly(L-proline) (23).
The present insect cell expression data argue strongly against the presence of a protein containing the ␣(I) and ␣(II) subunits in a single molecule. No information is currently available on sequences in the ␣ subunits that are involved in the ␣ 2 ␤ 2 tetramer assembly, but the C-terminal domains of the ␣ subunits, which show the highest degrees of amino acid sequence identity between the ␣(I) and ␣(II) subunits and the C. elegans ␣ subunit (9, 10), are known to contain residues involved in the binding of all the cosubstrates to a catalytic site (24,25). It seems probable that the regions involved in tetramer assembly contain some sequences that prevent incorporation of the ␣(I) and ␣(II) subunits into the same molecule.
Although no information is available on the sequences in the ␣ subunits that are critical for tetramer assembly, several observations suggest that in the case of the PDI/␤ subunit some such sequences are located close to the C-terminal domain of the polypeptide (11,26). The present data demonstrate that the N terminus of the PDI/␤ polypeptide is not critical for tetramer assembly, as the His-PDI polypeptide was found to form an active prolyl 4-hydroxylase as readily as the wild-type PDI/␤ polypeptide. Our additional experiments have demonstrated that the His-PDI polypeptide also efficiently forms an ␣␤ dimer with the C. elegans prolyl 4-hydroxylase ␣ subunit. 2 The histidine-tagged forms of prolyl 4-hydroxylase appear to offer an excellent source of the enzyme for simple large scale purification in experiments such as attempts at crystallization.
Poly(L-proline) has been regarded as a highly effective competitive inhibitor with respect to the polypeptide substrate of prolyl 4-hydroxylases from all the vertebrate sources studied, and an efficient polypeptide substrate for all plant prolyl 4-hydroxylases (1-3). It is therefore highly surprising that the human and mouse (9) type II prolyl 4-hydroxylases are inhibited by poly(L-proline) only at very high concentrations. This property of the type II enzyme agrees with that reported for crude preparations of prolyl 4-hydroxylase from the nematode Ascaris lumbricoides (27) and for the recombinant C. elegans prolyl 4-hydroxylase ␣␤ dimer (10,11). As these findings suggest that distinct differences are likely to exist in the structures of the peptide binding sites of various prolyl 4-hydroxylases, a comparison was made here between the K m values of the human type I and type II enzymes for three peptide substrates: the polypeptide (Pro-Pro-Gly) 10 , which is the most commonly  6. Analysis of the histidine tag affinity purification of human type I and type II prolyl 4-hydroxylases by PAGE analysis under nondenaturing conditions (A and B) and SDS-PAGE under reducing conditions (C). The type I enzyme is shown in panel A and type II enzyme in panel B under nondenaturing conditions. Lanes 1 show proteins soluble in a buffer containing 0.1% Triton X-100, lanes 2 flow-through fractions of the ProBond column, and lanes 3 fractions eluted with a buffer containing 100 mM imidazole. Lanes 4 show enzyme tetramers and lane 5 the His-PDI monomer after gel filtration. In panel C, lanes 1 and 2 show the type I and type II prolyl 4-hydroxylases, respectively. The locations of the ␣ and ␤ subunits are shown by the short and long arrows, respectively. It may be noted that the histidine tag increases the size of the PDI/␤ polypeptide, and therefore this polypeptide is poorly separated from the ␣ subunits in SDS-PAGE. The human type I and type II prolyl 4-hydroxylase tetramers are indicated by the arrow ␣ 2 ␤ 2 . a Values are given as dpm/10 g of Triton X-100-soluble cell protein.
used synthetic peptide substrate for prolyl 4-hydroxylase, the pentapeptide Gly-Val-Pro-Gly-Val, which has been introduced as a model peptide for elastin (28), a protein that also contains 4-hydroxyproline (1-3), and protocollagen, a biologically prepared collagenous substrate for the enzyme. Small but significant differences were found between the type I and type II enzymes in these experiments, in that the K m values for all three peptide substrates with the type II enzyme were about 3-6 times those with the type I enzyme. Nevertheless, these differences are very small when compared with the at least 200 -1000-fold differences between their K i values for poly(L-proline).
Mutations have been characterized in the genes for many types of collagen and for lysyl hydroxylase, a collagen hydroxylase closely related to prolyl 4-hydroxylase in its catalytic properties (3, 29 -32). No mutations have been identified in the gene coding for the ␣(I) subunit of prolyl 4-hydroxylase, and due to the central role of this enzyme in the synthesis of all collagens, such mutations have generally been assumed to be lethal. The present data indicating the presence of two isoforms of prolyl 4-hydroxylase ␣ subunit in human tissues raises the possibility, however, that mutations in the gene coding for one type may not be lethal, especially if cells are capable of upregulating the expression of the other type in cases when one type is inactive.