Prolyl 4-hydroxylase (proline hydroxylase, EC 1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens. The vertebrate enzyme is an alpha2beta2 tetramer, the beta subunit of which is identical to protein disulfide-isomerase (PDI, EC 5.3.4.1). We report here on cloning of the recently discovered alpha(II) subunit from human sources. The mRNA for the alpha(II) subunit was found to be expressed in a variety of human tissues, and the presence of the corresponding polypeptide and the (alpha(II))2beta2 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 alpha(I)alpha(II)beta2 tetramer. PDI/beta polypeptide containing a histidine tag in its N terminus was found to form prolyl 4-hydroxylase tetramers as readily as the wild-type PDI/beta 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 the Ki 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 Km 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.
Timely and successful drug development for rare cancer populations, such as pediatric oncology, requires consolidated efforts in the spirit of shared responsibility. In order to advance tailored development efforts, the concept of multistakeholder Strategy Forum involving industry, academia, patient organizations, and regulators has been developed. In this study, we review the first five pediatric oncology Strategy Forums co-organized by the European Medicines Agency between 2017 and 2020, reflecting on the outcomes and the evolution of the concept over time and providing an outline of how a "safe space" for multistakeholder engagement facilitated by regulators could be of potential value beyond pediatric oncology drug development.
Prolyl 4-hydroxylase (EC 1.14.11.2) catalyzes the posttranslational formation of 4-hydroxyproline in collagens. The vertebrate enzyme is an alpha 2 beta 2 tetramer, the beta subunit of which is a highly unusual multifunctional polypeptide, being identical to protein disulfide-isomerase (EC 5.3.4.1). We report here the cloning of a second mouse alpha subunit isoform, termed the alpha (II) subunit. This polypeptide consists of 518 aa and a signal peptide of 19 aa. The processed polypeptide is one residue longer than the mouse alpha (I) subunit (the previously known type), the cloning of which is also reported here. The overall amino acid sequence identity between the mouse alpha (II) and alpha (I) subunits is 63%. The mRNA for the alpha (II) subunit was found to be expressed in a variety of mouse tissues. When the alpha (II) subunit was expressed together with the human protein disulfide-isomerase/beta subunit in insect cells by baculovirus vectors, an active prolyl 4-hydroxylase was formed, and this protein appeared to be an alpha (II) 2 beta 2 tetramer. The activity of this enzyme was very similar to that of the human alpha (I) 2 beta 2 tetramer, and most of its catalytic properties were also highly similar, but it differed distinctly from the latter in that it was inhibited by poly(L-proline) only at very high concentrations. This property may explain why the type II enzyme was not recognized earlier, as an early step in the standard purification procedure for prolyl 4-hydroxylase is affinity chromatography on a poly(L-proline) column.
Prolyl 4-hydroxylase (EC 1.14.11.2) catalyses the formation of 4-hydroxyproline in collagens. The vertebrate enzymes are α2β2 tetramers while the Caenorhabditis elegans enzyme is an αβ dimer. The β-subunit is identical to protein disulphide isomerase (PDI), a multifunctional endoplasmic reticulum luminal polypeptide. ERp60 is a PDI isoform that was initially misidentified as a phosphatidylinositol-specific phospholipase C. We report here on the cloning and expression of the human and Drosophila ERp60 polypeptides. The overall amino acid sequence identity and similarity between the processed human ERp60 and PDI polypeptides are 29% and 56% respectively, and those between the Drosophila ERp60 and human PDI polypeptides 29% and 55%. The two ERp60 polypeptides were found to be similar to human PDI within almost all their domains, the only exception being the extreme C-terminal region. Nevertheless, when the human or Drosophila ERp60 was expressed in insect cells together with an α-subunit of human prolyl 4-hydroxylase, no tetramer was formed and no prolyl 4-hydroxylase activity was generated in the cells. Additional experiments with hybrid polypeptides in which the C-terminal regions had been exchanged between the human ERp60 and PDI polypeptides demonstrated that the differences in the C-terminal region are not the only reason for the lack of prolyl 4-hydroxylase tetramer formation by ERp60.
Protein disulphide isomerase (PDI; EC 5.3.4.1) is a multifunctional polypeptide that is identical to the β subunit of prolyl 4-hydroxylases. We report here on the cloning and expression of the Caenorhabditis elegans PDI/β polypeptide and its isoform. The overall amino acid sequence identity and similarity between the processed human and C. elegans PDI/β polypeptides are 61% and 85% respectively, and those between the C. elegans PDI/β polypeptide and the PDI isoform 46% and 73%. The isoform differs from the PDI/β and ERp60 polypeptides in that its N-terminal thioredoxin-like domain has an unusual catalytic site sequence -CVHC-. Expression studies in insect cells demonstrated that the C. elegans PDI/β polypeptide forms an active prolyl 4-hydroxylase α2β2 tetramer with the human α subunit and an αβ dimer with the C. elegans α subunit, whereas the C. elegans PDI isoform formed no prolyl 4-hydroxylase with either α subunit. Removal of the 32-residue C-terminal extension from the C. elegans α subunit totally eliminated αβ dimer formation. The C. elegans PDI/β polypeptide formed less prolyl 4-hydroxylase with both the human and C. elegans α subunits than did the human PDI/β polypeptide, being particularly ineffective with the C. elegans α subunit. Experiments with hybrid polypeptides in which the C-terminal regions had been exchanged between the human and C. elegans PDI/β polypeptides indicated that differences in the C-terminal region are one reason, but not the only one, for the differences in prolyl 4-hydroxylase formation between the human and C. elegans PDI/β polypeptides. The catalytic properties of the C. elegans prolyl 4-hydroxylase αβ dimer were very similar to those of the vertebrate type II prolyl 4-hydroxylase tetramer, including the Km for the hydroxylation of long polypeptide substrates.
Prolyl 4-hydroxylase catalyzes the formation of 4-hydroxyproline in collagens. The vertebrate enzymes are α2β2 tetramers, whereas theCaenorhabditis elegans enzyme is an αβ dimer, the β subunit being identical to protein-disulfide isomerase (PDI). We report here that the processed Drosophila melanogaster α subunit is 516 amino acid residues in length and shows 34 and 35% sequence identities to the two types of human α subunit and 31% identity to the C. elegans α subunit. Its coexpression in insect cells with the Drosophila PDI polypeptide produced an active enzyme tetramer, and small amounts of a hybrid tetramer were also obtained upon coexpression with human PDI. Four of the five recently identified critical residues at the catalytic site were conserved, but a histidine that probably helps the binding of 2-oxoglutarate to the Fe2+ and its decarboxylation was replaced by arginine 490. The enzyme had a higherK m for 2-oxoglutarate, a lower reaction velocity, and a higher percentage of uncoupled decarboxylation than the human enzymes. The mutation R490H reduced the percentage of uncoupled decarboxylation, whereas R490S increased the K m for 2-oxoglutarate, reduced the reaction velocity, and increased the percentage of uncoupled decarboxylation. The recently identified peptide-binding domain showed a relatively low identity to those from other species, and the K m of theDrosophila enzyme for (Pro-Pro-Gly)10 was higher than that of any other animal prolyl 4-hydroxylase studied. A 1.9-kilobase mRNA coding for this α subunit was present inDrosophila larvae. Prolyl 4-hydroxylase catalyzes the formation of 4-hydroxyproline in collagens. The vertebrate enzymes are α2β2 tetramers, whereas theCaenorhabditis elegans enzyme is an αβ dimer, the β subunit being identical to protein-disulfide isomerase (PDI). We report here that the processed Drosophila melanogaster α subunit is 516 amino acid residues in length and shows 34 and 35% sequence identities to the two types of human α subunit and 31% identity to the C. elegans α subunit. Its coexpression in insect cells with the Drosophila PDI polypeptide produced an active enzyme tetramer, and small amounts of a hybrid tetramer were also obtained upon coexpression with human PDI. Four of the five recently identified critical residues at the catalytic site were conserved, but a histidine that probably helps the binding of 2-oxoglutarate to the Fe2+ and its decarboxylation was replaced by arginine 490. The enzyme had a higherK m for 2-oxoglutarate, a lower reaction velocity, and a higher percentage of uncoupled decarboxylation than the human enzymes. The mutation R490H reduced the percentage of uncoupled decarboxylation, whereas R490S increased the K m for 2-oxoglutarate, reduced the reaction velocity, and increased the percentage of uncoupled decarboxylation. The recently identified peptide-binding domain showed a relatively low identity to those from other species, and the K m of theDrosophila enzyme for (Pro-Pro-Gly)10 was higher than that of any other animal prolyl 4-hydroxylase studied. A 1.9-kilobase mRNA coding for this α subunit was present inDrosophila larvae. Prolyl 4-hydroxylase (proline hydroxylase, EC 1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens and more than 10 other proteins with collagen-like sequences. This cotranslational and posttranslational modification plays a central role in the synthesis of all collagens because the 4-hydroxyproline residues are essential for formation of the collagen triple helix at body temperature (for reviews, see Refs. 1Prockop D.J. Kivirikko K.I. Annu. Rev. Biochem. 1995; 64: 403-434Crossref PubMed Scopus (1372) Google Scholar, 2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar). The vertebrate type I and type II enzymes are (α(I))2β2 and (α(II))2β2 tetramers, respectively, in which the β subunit is identical to protein-disulfide isomerase (EC5.3.4.1; PDI) 1The abbreviations used are: PDI, protein-disulfide isomerase; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; bp, base pair(s).1The abbreviations used are: PDI, protein-disulfide isomerase; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; bp, base pair(s). and has PDI activity even when present in the tetramer (2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar, 4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar, 6Annunen P. Autio-Harmainen H. Kivirikko K.I. J. Biol. Chem. 1998; 273: 5989-5992Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar). An α subunit of prolyl 4-hydroxylase has also been cloned from the nematode Caenorhabditis elegans (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar). This forms an active enzyme in insect cell coexpression experiments with both the C. elegans and human PDI polypeptides, but surprisingly, the enzymes containing the subunit are αβ dimers (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar, 8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar). The C. elegans PDI polypeptide also forms an active prolyl 4-hydroxylase with the human α subunits, but these enzymes are α2β2 tetramers (8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar). Assembly of a tetramer or dimer must therefore depend on the properties of the α subunit. It is currently unknown whether other nonvertebrate prolyl 4-hydroxylases are likewise dimers rather than tetramers. Prolyl 4-hydroxylase requires Fe2+, 2-oxoglutarate, O2, and ascorbate (2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar). A search for conserved amino acid residues within sequences of several 2-oxoglutarate dioxygenases and prolyl 4-hydroxylase α subunits from various species, together with extensive site-directed mutagenesis studies on the human α(I) subunit, have identified five amino acid residues that are critical at the cosubstrate binding sites of prolyl 4-hydroxylases (4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar, 7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar,9Myllylä R. Günzler V. Kivirikko K.I. Kaska D.D. Biochem. J. 1992; 286: 923-927Crossref PubMed Scopus (57) Google Scholar, 10Lamberg A. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1995; 270: 9926-9931Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar, 11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar). Three of these, His-412, Asp-414, and His-483 (numbered according to the human α(I) subunit), are involved in binding of the Fe2+ atom, whereas Lys-493 binds the C-5 carboxyl group of 2-oxoglutarate (10Lamberg A. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1995; 270: 9926-9931Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar, 11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar). His-501 is an additional critical residue, probably involved in both the binding of the C-1 carboxyl group of 2-oxoglutarate to the Fe2+ atom and the decarboxylation of this cosubstrate (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar). The aim of the present study was to clone the α subunit of prolyl 4-hydroxylase from Drosophila melanogaster and to characterize the corresponding enzyme. Special emphasis was given to comparison of the cDNA-derived amino acid sequence of theDrosophila α subunit with those of the human α(I) and α(II) subunits and the C. elegans α subunit, to identify conserved and nonconserved residues and regions, and to the question of whether the Drosophila enzyme is an α2β2 tetramer or an αβ dimer. The amino acid sequence of the Drosophila α subunit was found to show several distinct differences as compared with those of the other species studied, and the Drosophila enzyme was found to have some unique catalytic properties that appear related to these differences. A sequence homology search in FlyBase indicated the presence of a 179-nucleotide-long sequence (DM59D7T, FlyBase ID no. FBgn0015713) homologous to those of the vertebrate prolyl 4-hydroxylase α subunits. PCR primers PKA5 (5′-AAATGGAGGTTCCACGGCAG-3′) and PKA3 (5′-CCGATCCCATAGTTTGCCAC-CTG-3′) were synthesized based on this sequence and used to obtain a 160-bp PCR product from a cDNA pool generated from D. melanogastermRNA. The purified PCR product was 32P-labeled and used to screen a D. melanogaster larva λgt10 cDNA library (CLONTECH). This screening yielded seven positive clones among 600,000 recombinants, two of which, termed Dα2 and Dα6, were characterized in detail. The nucleotide sequences were determined by the dideoxynucleotide chain termination method (12Sanger K. Nicklen S. Coulson A.R. Proc. Natl. Acad. Sci. U. S. A. 1977; 74: 5463-5467Crossref PubMed Scopus (52599) Google Scholar) with the Sequenase enzyme (U. S. Biochemical Corp.) in an automated DNA sequencer (Applied Biosystems). Vector-specific or sequence-specific primers (Pharmacia) were used. DNASIS and PROSIS (version 6.0) (Pharmacia) were used to analyze the nucleotide and amino acid sequence data. The cleavage site of the signal pepide was predicted using the computational parameters of von Hejne (13von Hejne G. Nucleic Acids Res. 1986; 14: 4683-4690Crossref PubMed Scopus (3692) Google Scholar). An expression construct for theDrosophila α subunit was obtained by digesting the Dα2 clone with EcoRI. This fragment, which covers the whole coding region, 62 bp of the 5′-untranslated region and 12 bp of the 3′-untranslated region, was ligated to the transfer vector pVL1392 (14Luckow V.A. Summers M.D. Virology. 1989; 170: 31-39Crossref PubMed Scopus (235) Google Scholar). Arginine 490 (codon CGT) was converted to histidine (CAT) or serine (AGT) to produce the transfer vectors for the recombinant baculoviruses DαH and DαS, respectively. The mutagenesis steps were performed in a pSP72 vector (Promega) containing the Dα2 clone, using a PCR-based method (QuickChange site-directed mutagenesis kit; Stratagene). To construct a baculovirus transfer vector for theDrosophila PDI polypeptide, a PCR product containing 9 bp of the 5′-untranslated region, the whole coding region, and 227 bp of the 3′-untranslated region, was amplified using sequence-specific (15McKay R.R. Zhum L.Q. Shortridge R.D. Insect Biochem. Mol. Biol. 1995; 25: 647-654Crossref PubMed Scopus (16) Google Scholar) primers DRO1 (5′-GTGACCGCAATGAAATTCCTG-3′) and DRO4 (5′-GCAAAAAGCTTCGATGGCTAC-3′) from a cDNA pool generated fromD. melanogaster mRNA, and the fragment was ligated to the SmaI site of the pVL1393. Sf9 insect cells (Invitrogen) were cultured as monolayers in TNM-FH medium (Sigma) supplemented with 10% fetal bovine serum (Bioclear) at 27 °C. The recombinant baculovirus transfer vectors were cotransfected into Sf9 cells (Invitrogen) with a modified Autographa californica nuclear polyhedrosis virus DNA (BaculoGold, PharMingen) by calcium phosphate precipitation (16Gruenwald S. Heitz J. Baculovirus Expression Vector System, Procedures and Methods Manual. PharMingen, San Diego, CA1994Google Scholar). The resultant viral pools were collected 4 days later, plaque-purified, and amplified (16Gruenwald S. Heitz J. Baculovirus Expression Vector System, Procedures and Methods Manual. PharMingen, San Diego, CA1994Google Scholar). The recombinant viruses coding for the Drosophila α subunit, its R490H and R490S mutants, and the Drosophila PDI polypeptide were termed Dα, DαH, DαS, and Dβ. Other recombinant baculoviruses used were human α(I), C. elegans α, human PDI, and C. elegans PDI coding for the corresponding prolyl 4-hydroxylase subunits (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar, 8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar, 17Vuori K. Pihlajaniemi T. Marttila M. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7467-7470Crossref PubMed Scopus (114) Google Scholar). Sf9 insect cells were cultured as described above, either as monolayers or in suspension in spinner flasks (Techne Laboratories, Princeton, NJ), and were infected at a multiplicity of 5. For the production of an enzyme tetramer/dimer, the Dα, DαH, DαS, human α(I), orC. elegans α, and the human, C. elegans, orD. melanogaster PDI viruses were used in a ratio 1:1. The cells were harvested 3 days 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, 0.1m NaCl, 10 μm dithiothreitol, 0.1% Triton X-100, and 0.01 m Tris, pH 7.8, and centrifuged at 10,000 × g for 20 min at 4 °C. The resulting supernatants were analyzed by 8% SDS-PAGE or nondenaturing 8% PAGE and assayed for enzyme activity. Two μg of poly(A)+RNAs from embryonic, larval, and adult Drosophila(CLONTECH) were run on a 1% formaldehyde gel with molecular weight standards (Promega). The poly(A)+ RNAs were then transferred onto a nitrocellulose filter, and the filter was hybridized under stringent conditions with 32P-labeled Dα2 clone covering the whole coding region of theDrosophila α subunit. Autoradiography time was 9 days. Prolyl 4-hydroxylase activity was assayed by a method based on the hydroxylation-coupled decarboxylation of 2-oxo-[1-14C]glutarate as described previously (18Kivirikko K.I. Myllylä R. Methods Enzymol. 1982; 82: 245-304Crossref PubMed Scopus (324) Google Scholar), except that in the assays involving the Drosophila enzyme the 2-oxoglutarate concentration was increased from 0.1 to 0.2 μmol/ml and the (Pro-Pro-Gly)10·9H2O peptide substrate concentration from 0.1 to 1.5 mg/ml. The same method was used for measuring the uncoupled decarboxylation activity, except that the peptide substrate was omitted and the amount of enzyme was increased about 10-fold. K m values were determined by varying the concentration of one substrate in the presence of fixed concentrations of the second, whereas the concentrations of the other substrates were held constant (19Myllylä R. Tuderman L. Kivirikko K.I. Eur. J. Biochem. 1977; 80: 349-357Crossref PubMed Scopus (163) Google Scholar). The Triton X-100-soluble fraction of the insect cell homogenate was used as the enzyme source for theK m determinations. The amounts of recombinant proteins were compared by densitometry of Coomassie Brilliant Blue-stained bands in nondenaturing PAGE using a BioImage instrument (BioImage, Millipore). The first cDNA clone was a 160-bp PCR product obtained from a D. melanogaster cDNA pool with primers PK5 and PK3, which were based on a 179-nucleotide sequence DM59D7T present in FlyBase. This PCR product was used to screen a D. melanogaster larva λgt10 cDNA library, and one of the 7 positive clones obtained, Dα2, was found to cover 62 bp of the 5′-untranslated sequence, the whole coding region and 12 bp of the 3′-untranslated sequence of the corresponding Drosophilaα subunit mRNA. Dα6 continued 201 bp downward from the 3′-end of Dα2, and thus a total of 213 bp of the 3′-untranslated sequence was characterized. This 3′-untranslated region does not contain the canonical polyadenylation signal AATAAA, but it does contain a GATAAA sequence, which is accompanied 21 bp downstream by a possible poly(A) tail of 4 nucleotides at which the clone ends (these cDNA sequences are not shown but have been deposited in the GenBank/EMBL Data Bank with accession number AF096284). The cDNA clone encodes a 535-amino acid polypeptide. A putative signal peptide is present at its N terminus, the most likely first amino acid of the processed α subunit being glutamate, based on the computational parameters of von Hejne (13von Hejne G. Nucleic Acids Res. 1986; 14: 4683-4690Crossref PubMed Scopus (3692) Google Scholar) and comparison with the N terminus of the processed human α(II) subunit (5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar). Thus the length of the signal peptide is probably 19 amino acids and that of the processed α subunit 516 amino acids (Fig.1). The Drosophila α subunit is very similar in size to the human α(I) and α(II) subunits which have 517 (20Helaakoski T. Vuori K. Myllylä R. Kivirikko K.I. Pihlajaniemi T. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 4392-4396Crossref PubMed Scopus (103) Google Scholar) and 514 (4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar) residues, respectively (Fig. 1), whereas the C. elegans α subunit (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar) is longer, 542 amino acids, mainly because of a 32-residue C-terminal extension (Fig. 1). The Drosophila α subunit also has a C-terminal extension as compared with the vertebrate α(I) and α(II) subunits (3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar), but this extension is much shorter than in the C. elegans α subunit, being only 10 residues (Fig. 1). The Drosophila α subunit sequence has several minor deletions and insertions as compared with the other α subunit sequences, the longest deletion, between residues 225 and 226, being 9 amino acids (Fig. 1). The overall amino acid sequence identities between theDrosophila α subunit (excluding its 10-residue C-terminal extension) and the human α(I) and α(II) subunits are 34 and 35%, respectively, whereas its identity to the C. elegans α subunit is 31%. The identity is highest within the catalytically important (2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar) C-terminal region, residues 385–502 being 51 and 52% identical to those in the human α(I) and α(II) subunits, respectively, and 46% identical to those in the C. elegansα subunit. All three residues that bind the Fe2+ atom and the lysine that binds the C-5 carboxyl group of the 2-oxoglutarate (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar) are conserved, these amino acids being His-402, Asp-404, His-472, and Lys-482 in the Drosophila polypeptide. Surprisingly, however, the fifth critical residue, an additional histidine that is probably involved in both the binding of the C-1 carboxyl group of the 2-oxoglutarate to the Fe2+ atom and the decarboxylation of this cosubstrate (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar), is replaced by Arg-490. The recently identified peptide substrate binding domain (21Myllyharju J. Kivirikko K.I. EMBO J. 1999; 18: 306-312Crossref PubMed Scopus (61) Google Scholar), corresponding to residues 139–229 in the Drosophila α subunit, contains a 9-residue deletion close to its C-terminal end (between residues 225 and 226), whereas residues 139–225 show only 31, 36, and 23% identity to those in the human α(I) and α(II) subunits and the C. elegans α subunit, respectively. The 32-amino acid sequence of residues 230–261 shows a particularly low degree of identity to those in the other α subunits, namely 3.1, 6.3, and 12.5% when compared with those of the human α(I) and α(II) subunits and the C. elegans α subunit (Fig. 1). The sequence of residues 27–60 represents another region of low identity, namely 5.9, 11.8, and 2.9% to the same subunits. The Drosophila α subunit has three potential attachment sites for asparagine-linked oligosaccharide units, but the positions of these sites are different from those of the two sites (4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 20Helaakoski T. Vuori K. Myllylä R. Kivirikko K.I. Pihlajaniemi T. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 4392-4396Crossref PubMed Scopus (103) Google Scholar) present in the human α(I) and α(II) subunits (Fig. 1). The five cysteine residues present in the vertebrate α(I) subunits (2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar, 4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar, 20Helaakoski T. Vuori K. Myllylä R. Kivirikko K.I. Pihlajaniemi T. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 4392-4396Crossref PubMed Scopus (103) Google Scholar) and theC. elegans α subunit (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar) are all conserved in theDrosophila α subunit, whereas the vertebrate α(II) subunits (4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar) contain an additional cysteine not present in the other α subunits. A recombinant baculovirus encoding theDrosophila α subunit was generated and used to infectS. frugiperda insect cells. The cells were harvested 72 h after infection, homogenized in a buffer containing Triton X-100, and centrifuged. The cell pellet was solubilized in 1% SDS, and the 0.1% Triton X-100 soluble and 1% SDS soluble proteins were analyzed by 8% SDS-PAGE under reducing conditions followed by Coomassie Blue staining. Most of the insect cell proteins were soluble in the Triton X-100 containing buffer (details not shown). As with prolyl 4-hydroxylase α subunits from other sources, the recombinant polypeptide formed insoluble aggregates, and efficient extraction of the recombinantDrosophila α subunit required the use of 1% SDS. The polypeptide was found in the form of three main bands (Fig.2, lane 2) with mobilities slightly less than those of the three human α(I) subunit bands (Fig. 2, lane 1). Digestion of the sample with endoglycosidase H produced a single band with a mobility slightly higher than that of the lowest of the three bands present in the untreated sample (Fig. 2, lane 3), suggesting that the original polypeptide was present in tri-, di-, and monoglycosylated forms. This agrees with the presence of three potential N-glycosylation sites in the Drosophilaα subunit (Fig. 1) and indicates that all three sites are used, at least in insect cells. To produce a Drosophila prolyl 4-hydroxylase tetramer or dimer, a cDNA coding for the Drosophila PDI polypeptide was prepared by PCR using primers based on the published sequence (15McKay R.R. Zhum L.Q. Shortridge R.D. Insect Biochem. Mol. Biol. 1995; 25: 647-654Crossref PubMed Scopus (16) Google Scholar). A recombinant baculovirus encoding the Drosophila PDI polypeptide was then generated and used to infect insect cells together with the virus coding for the Drosophila α subunit. The cells were harvested 72 h after infection, and Triton X-100-soluble proteins of the cell homogenate were analyzed by nondenaturing 8% PAGE. A Coomassie Blue-stained band with a mobility slightly higher than that of the enzyme tetramer produced by coinfection of insect cells with recombinant viruses coding for the α(I) subunit of human prolyl 4-hydroxylase and the human PDI polypeptide (Fig. 3 A,lane 1) was seen in samples coinfected with the two types of Drosophila virus (Fig. 3 A,lane 2), whereas no such band was seen in a corresponding sample from cells infected with the Drosophilaα subunit virus alone (not shown). The mobility of thisDrosophila prolyl 4-hydroxylase band is distinctly different from that produced by coinfecting insect cells with baculoviruses coding for the C. elegans α subunit and the human PDI polypeptide (Fig. 3 A, lane 5), shown previously (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar) to represent an αβ dimer. The data thus strongly suggest that the Drosophila prolyl 4-hydroxylase is an α2β2 tetramer. No band corresponding to an enzyme tetramer or dimer could be seen in Coomassie Blue-stained samples from cells coexpressing theDrosophila α subunit and the human PDI polypeptide (Fig.3 A, lane 3), but formation of a small amount of a hybrid enzyme tetramer could be demonstrated by Western blotting with an antibody to the human PDI polypeptide (Fig.3 B, lane 3). Coexpression of the human α subunit with the Drosophila PDI polypeptide produced an enzyme tetramer that could easily be seen by Coomassie Blue staining (Fig. 3 A, lane 4). Prolyl 4-hydroxylase activity of the Triton X-100-soluble proteins was measured by a method based on the hydroxylation-coupled decarboxylation of 2-oxo-[1-14C]glutarate (18Kivirikko K.I. Myllylä R. Methods Enzymol. 1982; 82: 245-304Crossref PubMed Scopus (324) Google Scholar). When the specific activity of the Drosophila enzyme was compared with that of the human type I enzyme, the Coomassie Blue-stained bands in nondenaturing PAGE corresponding to these enzyme tetramers were first studied by densitometry, and the activity levels obtained were corrected (10Lamberg A. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1995; 270: 9926-9931Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar, 11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar) for the differences in the amounts of the two types of enzyme. Initial prolyl 4-hydroxylase activity assays performed under standard conditions (18Kivirikko K.I. Myllylä R. Methods Enzymol. 1982; 82: 245-304Crossref PubMed Scopus (324) Google Scholar) suggested that the specific activity of theDrosophila enzyme was very low, but when the values were corrected for saturating concentrations of 2-oxoglutarate and the (Pro-Pro-Gly)10 peptide substrate (see below), the specific activity of the Drosophila enzyme increased to about one-third of that of the human type I enzyme (details not shown). The rate of uncoupled 2-oxoglutarate decarboxylation, i.e.decarboxylation observed in the absence of the peptide substrate, was 3.7% of the rate of the complete reaction (details not shown, see also Table II below). This percentage is markedly higher than the value of 0.7% measured for the human type I enzyme (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar).Table IIProlyl 4-hydroxylase activity, rate of uncoupled decarboxylation of 2-oxoglutarate and Km for 2-oxoglutarate of the αArg-490 mutant Drosophila enzymesEnzymeProlyl 4-hydroxylase activityUncoupled enzyme activityK m for 2-oxoglutaratedpm/50 μldpm/50 μl% complete reactionμmNoninfected control<50<15DαDβ6960 ± 120260 ± 83.7 ± 0.184 ± 29DαR490HDβ6000 ± 30120 ± 3ap < 0.005.2.0 ± 0.04bp < 0.001.106 ± 27DαR490SDβ1890 ± 400ap < 0.005.260 ± 14013.8 ± 6.8ap < 0.005.325 ± 55ap < 0.005.The values are given in dpm/50 μl of Triton X-100-soluble cell protein, mean ± S.D. The statistical significances of the values obtained with the mutant enzymes were calculated versus the wild-type enzyme.a p < 0.005.b p < 0.001. Open table in a new tab Table IKm values of the Drosophila and human type I and type II prolyl 4-hydroxylases for cosubstrates and the (Pro-Pro-Gly)10substrate and Ki values for poly-(l-proline)Cosubstrate, substrate, or inhibitorConstantK m orK iDrosophilaHuman type IHuman type IIμmFe2+K m3 ± 13 ± 14 ± 12-OxoglutarateK m84 ± 29ap < 0.001 versus type I enzyme; p < 0.002 versus type II enzyme.22 ± 222 ± 2AscorbateK m300 ± 90320 ± 20340 ± 40(Pro-Pro-Gly)10K m260 ± 100ap < 0.001 versus type I enzyme; p < 0.002 versus type II enzyme.21 ± 588 ± 25Poly-(l-proline), M r7,000K i18 ± 8bp < 0.01 versus type I enzyme;p < 0.005 versus type II enzyme.0.6 ± 0.195 ± 31Poly-(l-proline),M r 44,000K i2.9 ± 0.1cp < 0.001 versus type I and type II enzyme.0.02 ± 0.00422 ± 2The values are given as mean ± S.D.a p < 0.001 versus type I enzyme; p < 0.002 versus type II enzyme.b p < 0.01 versus type I enzyme;p < 0.005 versus type II enzyme.c p < 0.001 versus type I and type II enzyme. Open table in a new tab The values are given in dpm/50 μl of Triton X-100-soluble cell protein, mean ± S.D. The statistical significances of the values obtained with the mutant enzymes were calculated versus the wild-type enzyme. The values are given as mean ± S.D. The K m values of the Drosophila enzyme for Fe2+ and ascorbate were very similar to those of the human type I and type II enzymes, whereas the K m for 2-oxoglutarate was about 4-fold (TableI). A major difference was found in theK m for the (Pro-Pro-Gly)10 peptide substrate, because that of the Drosophila enzyme was about 12 times that of the human type I enzyme and 3 times that of the human type II enzyme (Table I). The Drosophila enzyme resembled the vertebrate type II enzymes in being inhibited by poly(l-proline) only at high concentrations, theK i for poly(l-proline)M r 7,000 being about 30-fold relative to that obtained for the human type I enzyme and that for poly-(l-proline) M r 44,000 about 150-fold (Table I). These K i values of theDrosophila enzyme are not as high as those of the human type II enzyme, however, because the latter values are still about 5-fold and 7-fold higher than the Drosophila enzyme values for poly-(L-proline) M r 7,000 and 44,000, respectively (Table I). Arg-490 in the Drosophila α subunit was converted to either histidine or serine, and baculoviruses coding for the mutant α subunits were used to infect insect cells. The cells were harvested and homogenized as above, and the 1% SDS-soluble proteins were analyzed by 8% SDS-PAGE under reducing conditions followed by Coomassie Blue staining. The expression levels of both types of mutant α subunit per μg of protein were found to be similar to that of the wild-type polypeptide (Fig.4 A). The mutant α subunits were then coexpressed in insect cells with theDrosophila PDI polypeptide, and the cells were harvested 72 h after infection. Proteins soluble in a Triton X-100 buffer were analyzed by nondenaturing 8% PAGE and assayed for prolyl 4-hydroxylase activity. The Coomassie Blue-stained bands corresponding to the enzyme tetramer (Fig. 4 B) were studied by densitometry in each experiment, and these values were used to correct the enzyme activity levels for differences in the amounts of the various types of enzyme tetramer, as described previously (10Lamberg A. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1995; 270: 9926-9931Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar, 11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar). The specific activity of the R490H mutant enzyme tetramer was found to be about 90% that of the wild-type enzyme, whereas the specific activity of the R490S enzyme tetramer was only about 30% (TableII). The rate of uncoupled 2-oxoglutarate decarboxylation as a percentage of the rate of the complete reaction was lower with the R490H mutant enzyme than with the wild-type enzyme, whereas the R490S enzyme gave a very high percentage (Table II). The K m of the R490H mutant enzyme for 2-oxoglutarate was similar to that of the wild-type enzyme, whereas that of the R490S enzyme was about 4-fold that of the wild-type enzyme (Table II). Northern analysis of poly(A)+ RNA from embryonal, larval, and adultDrosophila gave a distinct hybridization signal only with the larval RNA (Fig. 5). The size of the mRNA was about 1.9 kilobases, which agrees well with the size of the 1883 nucleotides characterized by cloning and sequencing. The data reported here indicate that the α subunit ofDrosophila prolyl 4-hydroxylase shows about 30–35% amino acid sequence identity with the two types of human and the C. elegans α subunit. The expression level of the mRNA for theDrosophila α subunit was much higher at the larval stage than the embryonic or adult Drosophila. Our data further indicate that the Drosophila enzyme resembles the vertebrate enzymes (2Kivirikko K.I. Myllyharju J. Matrix Biol. 1998; 16: 357-378Crossref PubMed Scopus (235) Google Scholar, 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar, 4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar) rather than the C. elegans prolyl 4-hydroxylase αβ dimer (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar, 8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar) in that it is an α2β2 tetramer. The Drosophila α subunit was found to have several distinct amino acid sequence differences from those of other species in both its catalytic and peptide substrate binding domains. These differences appear to be related to certain unique catalytic properties of the Drosophila enzyme. The most distinct of these differences was that the histidine likely to assist in the binding of the C-1 carboxyl group of 2-oxoglutarate to the Fe2+ atom, and also in the decarboxylation of this cosubstrate (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar), was replaced by Arg-490. Site-directed mutagenesis studies on the human α(I) subunit have indicated that replacement of this histidine by other amino acids, including arginine, leads to a 2–3-fold increase in theK m for 2-oxoglutarate, a marked decrease in the reaction velocity, and a marked increase in the rate of uncoupled 2-oxoglutarate decarboxylation as a percentage of the rate of the complete reaction (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar). The present data on the wild-typeDrosophila enzyme are similar to those obtained with the histidine to arginine mutant human type I enzyme, in that theK m of the former for 2-oxoglutarate was about 4-fold that of the wild-type human enzyme, the reaction velocity was lower, and the percentage of uncoupled decarboxylation was about 5-fold. Mutation of Arg-490 to histidine in the Drosophila α subunit reduced the percentage of uncoupled decarboxylation but did not increase the reaction velocity or reduce the K m for 2-oxoglutarate, whereas mutation of Arg-490 to serine gave a much lower reaction velocity, a much higher percentage of the uncoupled reaction, and an increased K m for 2-oxoglutarate relative to those determined with the wild-type Drosophila enzyme. The data obtained with the R490S mutant Drosophila enzyme agree with those reported for the corresponding human α(I) subunit mutant (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar), whereas the data obtained with the R490H Drosophilamutant resemble those expected on the basis of the human mutants (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar) only in that the rate of uncoupled 2-oxoglutarate decarboxylation was lower than with the wild-type Drosophila enzyme. Thus, other residues in the Drosophila α subunit must influence the environment around residue 490 so that the arginine in this position is much more acceptable than in the human α(I) subunit. The finding that the lysine that binds the C-5 carboxyl group of 2-oxoglutarate (11Myllyharju J. Kivirikko K.I. EMBO J. 1997; 16: 1173-1180Crossref PubMed Scopus (164) Google Scholar) is conserved in the Drosophila α subunit is of interest, as the corresponding position in all other 2-oxoglutarate dioxygenases sequenced so far, including lysyl hydroxylase (22Myllylä R. Pihlajaniemi T. Pajunen L. Turpeenniemi-Hujanen T. Kivirikko K.I. J. Biol. Chem. 1991; 266: 2805-2810Abstract Full Text PDF PubMed Google Scholar, 23Valtavaara M. Szpirer C. Szpirer J. Myllylä R. J. Biol. Chem. 1998; 273: 12881-12886Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 24Passoja K. Rautavuoma K. Ala-Kokko L. Kosonen T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 10482-10486Crossref PubMed Scopus (90) Google Scholar, 25Passoja K. Myllyharju J. Pirskanen A. Kivirikko K.I. FEBS Lett. 1998; 434: 145-148Crossref PubMed Scopus (32) Google Scholar), is occupied by an arginine (9Myllylä R. Günzler V. Kivirikko K.I. Kaska D.D. Biochem. J. 1992; 286: 923-927Crossref PubMed Scopus (57) Google Scholar, 26Roach P.L. Clifton I.J. Fülöp V. Harlos K. Barton G.J. Hajdu J. Andersson I. Schofield C.J. Baldwin J.E. Nature. 1995; 375: 700-704Crossref PubMed Scopus (380) Google Scholar). It thus seems that this feature is specific to prolyl 4-hydroxylases from various species. The relatively low degree of identity between the Drosophilaα subunit and the other α subunits within the peptide binding domain agrees with the finding that the K m of theDrosophila enzyme for the (Pro-Pro-Gly)10peptide substrate is higher than that of any vertebrate prolyl 4-hydroxylase tetramer studied (Table I and Ref. 3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar) or of the C. elegans enzyme dimer (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar, 8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar). It is currently unknown whether this difference is related to possible differences between the sequences of the susceptible prolines in the Drosophila collagens (27Blumberg B. MacKrell A.J. Fessler J.H. J. Biol. Chem. 1988; 263: 18328-18337Abstract Full Text PDF PubMed Google Scholar, 28Fessler J.H. Fessler L.I. Annu. Rev. Cell Biol. 1989; 5: 309-339Crossref PubMed Scopus (145) Google Scholar, 29Yasothornsrikul S. Davis W.J. Cramer G. Kimbrell D.A. Dearolf C.R. Gene (Amst .). 1997; 198: 17-25Crossref PubMed Scopus (0) Google Scholar) and the human collagen substrates. The actual residues in the peptide binding domain that are critical for the binding of (Pro-Pro-Gly)10 and poly-(l-proline) are likewise unknown. The high K m of theDrosophila enzyme for (Pro-Pro-Gly)10 and the high K i values for poly-(l-proline) indicate that the peptide binding properties of theDrosophila enzyme are more closely related to those of the vertebrate type II enzymes (4Helaakoski T. Annunen P. Vuori K. MacNeil I.A. Pihlajaniemi T. Kivirikko K.I. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4427-4431Crossref PubMed Scopus (90) Google Scholar, 5Annunen P. Helaakoski T. Myllyharju J. Veijola J. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1997; 272: 17342-17348Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar) and the C. elegans enzyme (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar, 8Veijola J. Annunen P. Koivunen P. Pihlajaniemi T. Kivirikko K.I. Biochem. J. 1996; 317: 721-729Crossref PubMed Scopus (31) Google Scholar) than to those of the vertebrate type I enzyme (3Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar). A 32-amino acid residue region very close to the C-terminal end of the peptide binding domain was found to have a particularly low degree of identity between the Drosophila α subunit and those from other species. No function is currently known for this region, but the low degree of identity suggests that it may represent a variable interdomain sequence. It is also currently unknown whether the catalytic domain begins around residue 262, the C-terminal end of this variable region, or whether there is an additional domain with an unknown function between this region and the catalytic domain, as the highly conserved C-terminal region begins only around residue 385. The formation of only very small amounts of an enzyme tetramer between the Drosophila α subunit and the human PDI polypeptide is of interest, as the C. elegans α subunit readily forms an active enzyme dimer with the human PDI polypeptide (7Veijola J. Koivunen P. Annunen P. Pihlajaniemi T. Kivirikko K.I. J. Biol. Chem. 1994; 269: 26746-26753Abstract Full Text PDF PubMed Google Scholar) and as the human α(I) subunit readily formed an enzyme tetramer with theDrosophila PDI polypeptide. Thus the Drosophilaα subunit must have some structural feature that acts against tetramer formation with human PDI, and the Drosophila PDI polypeptide must have, in addition to structural features that allow prolyl 4-hdyroxylase formation with both the human andDrosophila α subunits, some specific feature that is not present in the human PDI polypeptide and that is recognized by theDrosophila α subunit. Further work will be required to elucidate these features, as no specific data are available at present on sequences in these polypeptides that are critical for prolyl 4-hydroxylase assembly. We thank Ari-Pekka Kvist for help with the computational work and Riitta Polojärvi, Jaana Träskelin, Liisa Äijälä, and Eeva Lehtimäki for expert technical assistance.
Prolyl 4-hydroxylase (EC 1.14.11.2) catalyzes the formation of 4-hydroxyproline in collagens. The vertebrate enzyme is an alpha 2 beta 2 tetramer, the beta subunit of which is identical to protein disulfide-isomerase (PDI). We report here on the cloning of the catalytically important alpha subunit from Caenorhabditis elegans. This polypeptide consists of 542 amino acids and signal peptide of 16 additional residues. The C. elegans alpha subunit is 25 amino acids longer than the human alpha subunit, mainly because of a 32-amino-acid C-terminal extension present only in the former. The overall amino acid sequence identity between these two alpha subunits is 45%, a 127-amino acid region close to the C terminus being especially well conserved. When the C. elegans alpha subunit was expressed together with the human PDI/beta subunit in insect cells by baculovirus vectors, an active prolyl 4-hydroxylase was formed, but surprisingly this C. elegans/human enzyme appeared to be an alpha beta dimer. The specific activity of this C. elegans/human enzyme was comparable with that of the human enzyme, and most of the other catalytic properties were also highly similar. Nevertheless, the C. elegans/human enzyme was not inhibited by poly(L-proline). The data indicate that the multifunctional PDI/beta subunit can form an active prolyl 4-hydroxylase with alpha subunits having marked differences in their amino acid sequences.
