Prolyl 4‐hydroxylase (EC 18.104.22.168), an α2β2 tetramer, catalyzes the formation of 4‐hydroxyproline in collagens. We converted 16 residues in the human α subunit individually to other amino acids, and expressed the mutant polypeptides together with the wild‐type β subunit in insect cells. Asp414Ala and Asp414Asn inactivated the enzyme completely, whereas Asp414Glu increased the Km for Fe2+ 15‐fold and that for 2‐oxoglutarate 5‐fold. His412Glu, His483Glu and His483Arg inactivated the tetramer completely, as did Lys493Ala and Lys493His, whereas Lys493Arg increased the Km for 2‐oxoglutarate 15‐fold. His501Arg, His501Lys, His501Asn and His501Gln reduced the enzyme activity by 85‐95%; all these mutations increased the Km for 2‐oxoglutarate 2‐ to 3‐fold and enhanced the rate of uncoupled decarboxylation of 2‐oxoglutarate as a percentage of the rate of the complete reaction up to 12‐fold. These and other data indicate that His412, Asp414 and His483 provide the three ligands required for the binding of Fe2+ to a catalytic site, while Lys493 provides the residue required for binding of the C‐5 carboxyl group of 2‐oxoglutarate. His501 is an additional critical residue at the catalytic site, probably being involved in both the binding of the C‐1 carboxyl group of 2‐oxoglutarate and the decarboxylation of this cosubstrate.
Prolyl 4‐hydroxylase (EC 22.214.171.124) catalyzes the hydroxylation of proline in ‐X‐Pro‐Gly‐ sequences in collagens and related proteins. This co‐translational and post‐translational modification plays a central role in the synthesis of all collagens, as 4‐hydroxyproline residues are essential for the folding of the newly synthesized collagen polypeptide chains into triple‐helical molecules. The hydroxylation requires Fe2+, 2‐oxoglutarate, O2 and ascorbate and involves an oxidative decarboxylation of 2‐oxoglutarate (for reviews, see Kivirikko et al., 1989, 1992; Prockop and Kivirikko, 1995).
Prolyl 4‐hydroxylase from vertebrates is an α2β2 tetramer, in which the β subunits are identical to the protein disulfide isomerase (PDI, EC 126.96.36.199, Koivu et al., 1987; Pihlajaniemi et al., 1987; Parkkonen et al., 1988) and the α subunits contribute to most parts of the two catalytic sites (Kivirikko et al., 1989, 1992). A catalytic site appears to comprise a set of separate locations for the binding of the various co‐substrates and the peptide substrate. The Fe2+ is probably coordinated with the α subunit by three side chains (Hanauske‐Abel and Günzler, 1982; Kivirikko et al., 1989, 1992). The 2‐oxoglutarate‐binding site can be divided into at least two distinct subsites: subsite I probably consists of a positively charged side chain of the α subunit, which binds the C‐5 carboxyl group of the 2‐oxoglutarate, while subsite II consists of two cis‐positioned coordination sites of the enzyme‐bound Fe2+ and is chelated by the C‐1‐C‐2 moiety (Hanauske‐Abel and Günzler, 1982; Majamaa et al., 1984; Kivirikko et al., 1989, 1992). The O2 is thought to be bound to the Fe2+ end‐on in an axial position, and the subsequent decarboxylation of 2‐oxoglutarate is thought to lead to the formation of ferryl ion, which hydroxylates a proline residue in the second half‐reaction of the catalytic cycle (Hanauske‐Abel and Günzler, 1982; Kivirikko et al., 1989, 1992; Hanauske‐Abel, 1991). Ascorbate acts as an alternative oxygen acceptor in the uncoupled decarboxylation cycles, in which 2‐oxoglutarate is decarboxylated without subsequent hydroxylation of the peptide substrate (de Jong and Kemp, 1984; Myllylä et al., 1984; Majamaa et al., 1986).
A search for conserved residues within the sequences of several 2‐oxoglutarate dioxygenases and a related dioxygenase, isopenicillin N synthase, indicated weak homology within two histidine‐containing motifs, His‐1 and His‐2, spaced ∼50‐70 amino acids apart (Myllylä et al., 1992), the histidines concerned in the human α subunit sequence (Helaakoski et al., 1989) being His412 and His483. Site‐directed mutagenesis indicated that conversion of either of these histidines to serine inactivated the enzyme completely, while mutation of one additional histidine, His501, inactivated it by ∼96% (Lamberg et al., 1995). As analyses of the related enzyme, isopenicillin N synthase, by a variety of techniques had suggested that all three Fe2+‐binding ligands in that enzyme may be histidines (Jiang et al., 1991; Ming et al., 1991; Randall et al., 1993), it initially was considered possible that these three histidines might provide the three Fe2+‐binding ligands in prolyl 4‐hydroxylase, even though the His501Ser mutation did not inactivate the enzyme completely (Lamberg et al., 1995). Recent determination of the crystal structure of isopenicillin N synthase has indicated, however, that although two of the Fe2+‐binding ligands are histidines, corresponding to those in the His‐1 and His‐2 motifs, the third is an aspartate present in position +2 with respect to the histidine in the His‐1 motif (Roach et al., 1995). A corresponding aspartate residue is found in the α subunits of all prolyl 4‐hydroxylases studied (Helaakoski et al., 1989, 1995; Bassuk et al., 1989; Veijola et al., 1994).
The aim of the present work was to identify in detail all four critical residues involved in the binding of Fe2+ (three ligands) and 2‐oxoglutarate (subsite I) to the catalytic site of human type I prolyl 4‐hydroxylase (Helaakoski et al., 1989) and to determine whether some of these can be replaced by other residues. We also attempted to determine whether His501 is involved in any of these functions.
Aspartate 414 is an essential residue but can be partly replaced by glutamate
Six aspartate residues were converted individually to alanine (Figure 1). These represented all six aspartates present within the highly conserved 127 amino acid region (residues 387‐513) close to the COOH‐terminus of the prolyl 4‐hydroxylase α subunit. Asp414 (circled in Figure 1) is the only residue among these that is conserved between the human and Caenorhabditis elegans α subunits (Helaakoski et al., 1989; Veijola et al., 1994). All the mutant α subunits were expressed in Sf9 cells together with the wild‐type PDI/β polypeptide. The cells were harvested 72 h after infection, homogenized in a buffer containing Triton X‐100 and centrifuged, and the remaining pellets were solubilized in 1% SDS. No differences were found in the mobilities of the various mutant α subunits in SDS‐PAGE analysis (data not shown). The Triton X‐100‐soluble proteins were then analyzed by non‐denaturing PAGE and Coomassie staining and assayed for prolyl 4‐hydroxylase activity by a method based on the hydroxylation‐coupled decarboxylation of 2‐oxo[1‐14C]glutarate. The Coomassie‐stained bands corresponding to the enzyme tetramer were studied by densitometry in each experiment, and the amounts of the various mutants were calculated as percentages of the amount of the wild‐type enzyme. These percentages were then used to correct the enzyme activities of the various mutants to differences in the expression levels. In almost all cases this correction was less than ±20%.
Five of the six mutant α subunits formed an enzyme tetramer with the PDI/β polypeptide (Figure 2A). Asp478Ala was the only mutant that did not form a tetramer, indicating that although this position can be occupied by aspartate, or by asparagine in the C.elegans α subunit (Veijola et al., 1994), alanine in this position will prevent tetramer formation. As the intention here was to characterize the catalytic sites of the enzyme rather than its tetramer assembly, this aspect was not studied any further.
Only one aspartate to alanine mutation was found to have a marked effect on prolyl 4‐hydroxylase activity in the enzyme tetramer. Mutation of the conserved Asp414 (underlined in Figure 1), the residue corresponding to the Fe2+‐binding aspartate in isopenicillin N synthase, inactivated the enzyme completely (Table I). The Asp391Ala mutation had no effect on enzyme activity, while Asp422Ala and Asp442Ala reduced it by about half and Asp419Ala increased it by ∼20% (Table I). The Km values of the Asp419Ala, Asp422Ala and Asp442Ala mutant prolyl 4‐hydroxylases for Fe2+ and 2‐oxoglutarate were identical to those of the wild‐type enzyme (details not shown).
When Asp414 was replaced with asparagine and glutamate, both mutant α subunits formed an enzyme tetramer with the PDI/β polypeptide (Figure 2B). The mutation Asp414Asn eliminated the enzyme activity completely, whereas Asp414Glu reduced it to ∼15% (Table I). The Km of the Asp414Glu mutant prolyl 4‐hydroxylase for Fe2+ was markedly increased relative to the wild‐type enzyme, ∼15‐fold (Table II). The Km for 2‐oxoglutarate was also increased, although not as markedly, ∼5‐fold (Table II). The Vmax determined from the kinetic plots was identical to that obtained with the wild‐type enzyme (details not shown).
Histidines 412 and 483 cannot be replaced by a negatively or positively charged residue
Mutation of the two histidines present in the His‐1 and His‐2 motifs of prolyl 4‐hydroxylase, His412 and His483 (underlined in Figure 1), to serine has been shown previously to inactivate the enzyme completely (Lamberg et al., 1995). Site‐directed mutagenesis of bovine aspartyl (asparaginyl) β‐hydroxylase has shown that substitution of the histidine present in the His‐2 motif of that enzyme with a negatively charged amino acid results in 10‐20% residual activity, whereas substitution with an uncharged or a positively charged residue results in complete inactivation (Jia et al., 1994; McGinnis et al., 1996). We therefore converted the His412 and His483 of the prolyl 4‐hydroxylase α subunit individually to glutamate, and His483 also to arginine. The mutant α subunits were then expressed together with the PDI/β polypeptide in Sf9 cells and analyzed as above. All three mutant α subunits formed enzyme tetramers with the PDI/β polypeptide, as shown by the presence of a Coomassie‐stained band in non‐denaturing PAGE (Figure 2C). None of the His412Glu, His483Arg and His483Glu mutant prolyl 4‐hydroxylases had any enzyme activity, however (Table I).
Lysine 493 is involved in the binding of 2‐oxoglutarate
Attempts to model the Fe2+‐binding region of the human α subunit sequence suggested that the residue that ionically binds the C‐5 carboxyl group of the 2‐oxoglutarate (subsite I) may be Lys493 or Lys498 (A.Haataja and A.Koskinen, unpublished data). Seven lysine residues, five of which are conserved between the human (Helaakoski et al., 1989) and C.elegans (Veijola et al., 1994) α subunit sequences (circled in Figure 1), were therefore converted individually to alanine, and the mutant polypeptides were expressed in Sf9 cells together with the wild‐type PDI/β polypeptide. The Triton X‐100‐soluble proteins were then used for the analyses as above. None of the mutations was found to inhibit tetramer assembly (Figure 3A), and only minor differences were found in the amounts of tetramer formed, as judged from the intensities of the Coomassie‐stained bands in non‐denaturing PAGE (details not shown). Mutation of the conserved Lys493 (underlined in Figure 1) inactivated the enzyme completely (Table III), while mutation of the other conserved lysines either reduced the enzyme activity by ∼15‐30% (Lys365Ala, Lys418Ala and Lys498Ala), or increased it by ∼20% (Lys462Ala). Mutation of the non‐conserved Lys425 had no effect on enzyme activity, but mutation of the other non‐conserved Lys461 increased it by ∼30% (Table III). As in the case of the aspartate mutations, none of these six lysine mutations with slight effects on enzyme activity altered the Km values for Fe2+ or 2‐oxoglutarate (details not shown).
To study the role of Lys493 in more detail, this residue was also converted to arginine and histidine. Neither mutation had any effect on tetramer assembly (Figure 3B), but Lys493Arg reduced the enzyme activity to ∼15% and Lys493His resulted in complete inactivation (Table III). The Km values of the Lys493Arg mutant enzyme for Fe2+, ascorbate and the peptide substrate showed no differences relative to the wild‐type enzyme, but the Km for 2‐oxoglutarate was markedly increased, ∼15‐fold (Table II). The Vmax obtained with the Lys493Arg mutant enzyme was the same as with the wild‐type enzyme (details not shown).
The Ki values of the Lys493Arg mutant enzyme for structural analogs of 2‐oxoglutarate, oxoadipinate and pyridine 2,4‐dicarboxylate, which become bound at both main subsites of the 2‐oxoglutarate‐binding site (Hanauske‐Abel and Günzler, 1982; Majamaa et al., 1984) were also increased, ∼4‐fold (Table IV), whereas the Ki values for oxovalerate and pyridine 2‐carboxylate, which cannot become bound at subsite I due to the lack of a carboxyl group (Majamaa et al., 1984), were the same as for the wild‐type enzyme (Table IV). The Ki for oxaloacetate, which is shorter than 2‐oxoglutarate and therefore reacts poorly at subsite I, was increased ∼2‐fold (Table IV). The Lys493Arg enzyme, like the wild‐type enzyme, was able to utilize 2‐oxoadipinate as a co‐substrate in the hydroxylation of [14C]proline‐labeled protocollagen substrate (details not shown).
Histidine 501 is an important residue but can be partly replaced by arginine or lysine
Previous site‐directed mutagenesis studies have shown that mutation of His501 (underlined in Figure 1) to serine reduces the prolyl 4‐hydroxylase activity of the tetramer to ∼4%, the Km for 2‐oxoglutarate being ∼3‐fold compared with that of the wild‐type enzyme (Lamberg et al., 1995). The role of His501 in the catalytic activity was studied further by converting this residue to arginine, asparagine, glutamine and lysine. None of these four mutations was found to inhibit enzyme tetramer formation when the mutant α subunits were expressed in Sf9 cells together with the PDI/β polypeptide (Figure 4). The mutations His501Asn and His501Gln reduced the enzyme activity of the tetramer to ∼5%, while His501Arg and His501Lys reduced it to ∼10‐15% (Table V). The decreases in Vmax values were similar to those observed in the reaction rates measured under standard assay conditions (details not shown). The Km values for Fe2+ and the peptide substrate of the four mutant tetramers were the same as those of the wild‐type enzyme, whereas the Km values for 2‐oxoglutarate were increased ∼2‐ to 3‐fold (Table II). The Ki of the His501Ser mutant enzyme for pyridine 2‐carboxylate was likewise increased 2‐fold (details not shown).
Rate of uncoupled decarboxylation
Prolyl 4‐hydroxylase also catalyzes uncoupled decarboxylation of 2‐oxoglutarate (Tuderman et al., 1977; Counts et al., 1978; Rao and Adams, 1978) at a rate which is 0.7 ± 0.2% of that of the complete reaction (Table VI). All the mutant prolyl 4‐hydroxylases were tested for this rate by performing the enzyme reaction without the polypeptide substrate, and in almost all cases, including those with a markedly decreased rate of the complete reaction, this percentage did not differ significantly from that of the wild‐type enzyme (some examples are shown in Table VI). However, the His501 mutants differed distinctly from the wild‐type enzyme and all the other mutants, in that the rate of uncoupled decarboxylation decreased much less than the rate of the complete reaction, and thus the percentage of the uncoupled reaction was increased up to ∼12‐fold, the highest percentages being seen with the His501Asn and His501Ser mutants and the lowest with the His501Lys mutant (Table VI).
Since poly(l‐proline), a competitive inhibitor with respect to the polypeptide substrate, enhances the rate of uncoupled decarboxylation (Counts et al., 1978; Myllylä et al., 1984), all the mutants were also tested for uncoupled decarboxylation in the presence of 5 μg/ml of poly(l‐proline), mol. wt 7000. The rate of the uncoupled reaction (as d.p.m./100 μg) increased ∼1.3‐ to 1.6‐fold with the wild‐type enzyme and all the mutants, there being no difference between the His501 mutants and the others (details not shown).
The present data clearly demonstrate that Asp414 is one of the three Fe2+‐binding ligands in the α subunit of human type I prolyl 4‐hydroxylase. Mutation of this residue to alanine or asparagine inactivated the enzyme completely, whereas mutation to glutamate inactivated it by ∼85% under standard assay conditions. The marked increase in the Km of the Asp414Glu mutant enzyme for Fe2+ is in agreement with the proposed function of Asp414 as one of the Fe2+‐binding ligands, and the increase in that for 2‐oxoglutarate, although smaller, is not surprising, as enzyme‐bound Fe2+ constitutes one of the two main binding sites for this co‐substrate (Hanauske‐Abel and Günzler, 1982; Majamaa et al., 1984). It seems very likely that the longer side chain of glutamate distorts the Fe2+‐binding site to such an extent that this influences the binding of 2‐oxoglutarate to its other main site.
The observation that substitution of glutamate for His412 or His483 or arginine for His483 inactivated prolyl 4‐hydroxylase completely differs from recent data on aspartyl (asparaginyl) β‐hydroxylase, which suggests that the histidine present in the His‐2 motif can be replaced by glutamate with retention of partial activity (McGinnis et al., 1996). Due to the complete inactivation, it was not possible to determine whether replacement of either of the two histidines in prolyl 4‐hydroxylase by other amino acids increases the Km for Fe2+. Previous data on the inactivation of prolyl 4‐hydroxylase by diethyl pyrocarbonate and prevention of this inactivation by co‐substrates of the reaction strongly suggest that histidine residues are functional at the catalytic site, probably at the Fe2+‐binding site (Myllylä et al., 1992). Furthermore, determination of the crystal structure of isopenicillin N synthase has demonstrated that two of the Fe2+‐binding ligands in that enzyme are histidines, corresponding to those in the His‐1 and His‐2 motifs of prolyl 4‐hydroxylase, while the third is an aspartate in position +2 with respect to the histidine in the His‐1 motif (Roach et al., 1995). It thus seems very likely that the three Fe2+‐binding ligands in the α subunit of human prolyl 4‐hydroxylase are His412, Asp414 and His483 (Figure 5). Recent site‐directed mutagenesis studies on a closely related enzyme, lysyl hydroxylase, likewise suggest that the three Fe2+‐binding ligands in that enzyme are two histidines and an aspartate in the corresponding positions (Pirskanen et al., 1996), although virtually no other amino acid sequence similarities are found between these two enzymes (Myllylä et al., 1991).
Prolyl 4‐hydroxylase differs from isopenicillin N synthase in that the latter does not utilize 2‐oxoglutarate as a co‐substrate. The present data strongly suggest that the residue that ionically binds the C‐5 carboxyl group of 2‐oxoglutarate (subsite I) is Lys493 (Figure 5). Mutation of this residue to alanine or histidine inactivated the enzyme completely, but had no inhibitory effect on α2β2 tetramer assembly, whereas mutation to arginine led to ∼85% inactivation of the tetramer. The marked increase in the Km of the Lys493Arg enzyme for 2‐oxoglutarate with no changes in the values for other co‐substrates or the peptide substrate indicates that Lys493 is involved in the binding of 2‐oxoglutarate. This residue probably constitutes subsite I of the 2‐oxoglutarate‐binding site, as the Ki values of the Lys493Arg enzyme for oxoadipinate and pyridine 2,4‐dicarboxylate, which become bound at both main subsites of the 2‐oxoglutarate site (Majamaa et al., 1984), were distinctly increased, whereas those for oxovalerate and pyridine 2‐carboxylate, which cannot bind to subsite I, were unaltered. The Ki for oxaloacetate, which reacts poorly or not at all at subsite I (Majamaa et al., 1984), was only slightly increased.
The ascorbate‐binding site of prolyl 4‐hydroxylase also contains the two cis‐positioned coordination sites of the enzyme‐bound iron, and is thus partially identical to the binding site of 2‐oxoglutarate (Majamaa et al., 1986). Nevertheless, ascorbate does not become bound at subsite I of the 2‐oxoglutarate‐binding site. In agreement with this, the Km of the Lys493Arg enzyme for ascorbate was the same as that of the wild‐type enzyme. The Km of the Asp414Glu enzyme for ascorbate was likewise not increased, although that for 2‐oxoglutarate was increased ∼5‐fold, again supporting the view that the only main determinant for ascorbate binding is the enzyme‐bound iron atom.
His501 appears to be an additional important residue, as its mutation to serine has been found previously to inactivate prolyl 4‐hydroxylase by 96% (Lamberg et al., 1995) and its mutation to asparagine or glutamine was now found to inactivate the enzyme by ∼95%, while mutation to lysine or arginine inactivated it by 85‐90%. This residue was clearly not involved in Fe2+ binding, however, as the Km values of all the His501 mutants for Fe2+ were the same as those of the wild‐type enzyme. In agreement with data on the His501Ser mutant enzyme (Lamberg et al., 1995), the Km values of all the His501 mutant enzymes for 2‐oxoglutarate were increased ∼2‐ to 3‐fold, whereas the Km values for ascorbate (Lamberg et al., 1995) and the peptide substrate were not increased significantly. The Ki of the His501Ser enzyme for pyridine 2,4‐dicarboxylate, a competitive inhibitor with respect to 2‐oxoglutarate, was also ∼2.5‐fold (Lamberg et al., 1995). In contrast to the data obtained with the Lys493Arg mutant enzyme, the Ki of the His501Ser enzyme for pyridine 2‐carboxylate was increased to the same extent as the Ki for pyridine 2,4‐dicarboxylate. These data suggest that His501 may play a role in directing the coordination of the C‐1 carboxyl group of 2‐oxoglutarate to the enzyme‐bound Fe2+.
A surprising finding was that the rate of uncoupled decarboxylation of 2‐oxoglutarate as a percentage of the rate of the complete reaction was increased up to ∼12‐fold with all the His501 mutants, but not with any other mutant. This increase in the relative rate of uncoupling was due to a much smaller decrease in the rate of the uncoupled reaction than in that of the complete reaction. This difference cannot be explained by any effect on the binding of the peptide substrate, as the Km values of the His501 mutant enzymes for the peptide substrate in the complete reaction and the rate of enhancement by poly(l‐proline) in the uncoupled reaction were essentially identical to those observed with the wild‐type enzyme. The most likely explanation is that His501 binds through hydrogen bonding to one of the oxygen atoms in the six‐membered oxygen‐bridged ferryl intermediate (Hanauske‐Abel and Günzler, 1982; Hanauske‐Abel, 1991) and thus helps the cleavage of the intermediate to succinate, CO2 and ferryl ion. As the rate of the uncoupled decarboxylation is only 0.7% of that of the complete reaction, the rate of cleavage of the intermediate may be much more critical in the latter, and thus the His501 mutations can be expected to influence more the rate of the complete reaction. His501 is thus proposed to play two roles in the 2‐oxoglutarate‐binding site (Figure 5). It directs the orientation of the C‐1 carboxyl group of 2‐oxoglutarate to the active iron center and accelerates the breakdown of the ferryl intermediate.
Materials and methods
Aspartates 391, 414, 419, 422, 442 and 478 (codon GAC or GAT) and lysines 365, 418, 425, 461, 462, 493 and 498 (codon AAA) in the type I α subunit of human prolyl 4‐hydroxylase were converted individually to alanine (codon GCC, GCT or GCA). Asp414 (codon GAC) was converted further to asparagine (codon AAC) and glutamate (codon GAA), and Lys493 to arginine (codon AGA) and histidine (codon ATT). His412 and His483 (codon CAT) were converted to glutamate (codon GAA), and His483 was also converted to arginine (codon CGT). His501 (codon CAT) was converted to arginine (codon CGT), asparagine (codon AAT), glutamine (codon CAA) and lysine (codon AAA). The mutagenesis steps were performed in a pBluescript vector (Stratagene) containing the full‐length cDNA clone (PA‐59) for the human prolyl 4‐hydroxylase α subunit at the SmaI site (Helaakoski et al., 1989). The mutagenesis was carried out using an oligonucleotide‐directed in vitro system based on the unique site elimination procedure (Pharmacia Biotech Inc.), after which the plasmid was digested with AflII and BamHI. The resulting 1.8 kb AflII‐BamHI cDNA fragment containing the mutant site was then cloned into the AflII‐BamHI‐digested baculovirus transfer vector pVLα59 (Vuori et al., 1992). The sequences were verified by dideoxynucleotide sequencing (Sanger et al., 1977).
Generation of recombinant baculoviruses
The recombinant baculovirus transfer vectors were co‐transfected into Spodoptera frugiperda Sf9 cells with a modified Autographa californica nuclear polyhedrosis virus DNA (PharMingen) by calcium phosphate transfection, and the recombinant viruses were selected (Gruenwald and Heitz, 1993).
Analysis of recombinant proteins in insect cells
Sf9 cells were cultured as monolayers in TNM‐FH medium (Sigma) supplemented with 10% fetal bovine serum (BioClear) at 27°C. Cells seeded at a density 5×106 per 100 mm plate were infected at a multiplicity of 5 with the viruses coding for the wild‐type or mutant α subunit together with a virus coding for the protein disulfide isomerase/β polypeptide (Vuori et al., 1992). 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 0.1 M NaCl, 0.1 M glycine, 10 μM dithiothreitol (DTT), 0.1% Triton X‐100 and 0.01 M Tris buffer, pH 7.8, and centrifuged at 10 000 g for 20 min. Aliquots of the Triton extracts were analyzed by non‐denaturing 8% PAGE and assayed for enzyme activity. The cell pellets were solubilized further in 1% SDS and analyzed by 8% SDS‐PAGE under reducing conditions.
Prolyl 4‐hydroxylase activity was assayed by a method based on the hydroxylation‐coupled decarboxylation of 2‐oxo[1‐14C]glutarate (Kivirikko and Myllylä, 1982). The reaction was performed in a final volume of 1.0 ml, which contained 10 μl of the Triton X‐100 extract as the source of the enzyme, 0.1 mg of (Pro‐Pro‐Gly)10.9H2O as substrate, 0.05 μmol of FeSO4, 0.1 μmol of 2‐oxo[1‐14C]glutarate (100 000 d.p.m.), 1 μmol of ascorbate, 0.3 mg of catalase (Sigma), 0.1 μmol of DTT, 2 mg of bovine serum albumin (Sigma) and 50 μmol of Tris‐HCl buffer adjusted to pH 7.8 at 25°C. The uncoupled reaction was performed as above, except that the amount of the Triton X‐100 extract was increased to 200 μl and the peptide substrate was omitted. Km values were determined as described previously (Myllylä et al., 1977). Formation of 4‐hydroxy‐[14C]proline was studied using [14C]proline‐labeled protocollagen substrate (Kivirikko and Myllylä, 1982). Protein concentrations were determined with a Bio‐Rad protein assay kit (Bio‐Rad). The levels of expression of the wild‐type and mutant prolyl 4‐hydroxylase tetramers were compared by densitometry of the Coomassie‐stained band in non‐denaturing PAGE using a BioImage instrument (BioImage, Millipore).
We thank Professor Ari Koskinen, Department of Chemistry, University of Oulu, for valuable comments and suggestions and Eeva Lehtimäki for her expert technical assistance. This work was supported by grants from the Research Council for Health within the Academy of Finland and from FibroGen Inc., Sunnyvale, CA.
- Copyright © 1997 European Molecular Biology Organization