Skip to main content
PMC home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2014 Sep 2;111(37):13331–13336. doi: 10.1073/pnas.1409916111

Human oxygen sensing may have origins in prokaryotic elongation factor Tu prolyl-hydroxylation

John S Scotti a, Ivanhoe K H Leung a,1,2, Wei Ge a,b,1, Michael A Bentley c, Jordi Paps d, Holger B Kramer e, Joongoo Lee a, WeiShen Aik a, Hwanho Choi a, Steinar M Paulsen c,3, Lesley A H Bowman f, Nikita D Loik a,4, Shoichiro Horita a,e, Chia-hua Ho a,5, Nadia J Kershaw a,6, Christoph M Tang f, Timothy D W Claridge a, Gail M Preston c, Michael A McDonough a, Christopher J Schofield a,7
PMCID: PMC4169948  PMID: 25197067

Significance

The Fe(II)- and 2-oxoglutarate (2OG)-dependent hypoxia-inducible transcription factor prolyl-hydroxylases play a central role in human oxygen sensing and are related to other prolyl-hydroxylases involved in eukaryotic collagen biosynthesis and ribosomal modification. The finding that a PHD-related prolyl-hydroxylase in Pseudomonas spp. regulates pyocyanin biosynthesis supports prokaryotic origins for the eukaryotic prolyl-hydroxylases. The identification of the switch I loop of elongation factor Tu (EF-Tu) as a Pseudomonas prolyl-hydroxylase domain containing protein (PPHD) substrate provides evidence of roles for 2OG oxygenases in both translational and transcriptional regulation. A structure of the PPHD:EF-Tu complex, the first to the authors' knowledge of a 2OG oxygenase with its intact protein substrate, reveals that major conformational changes occur in both PPHD and EF-Tu and will be useful in the design of new prolyl-hydroxylase inhibitors.

Abstract

The roles of 2-oxoglutarate (2OG)-dependent prolyl-hydroxylases in eukaryotes include collagen stabilization, hypoxia sensing, and translational regulation. The hypoxia-inducible factor (HIF) sensing system is conserved in animals, but not in other organisms. However, bioinformatics imply that 2OG-dependent prolyl-hydroxylases (PHDs) homologous to those acting as sensing components for the HIF system in animals occur in prokaryotes. We report cellular, biochemical, and crystallographic analyses revealing that Pseudomonas prolyl-hydroxylase domain containing protein (PPHD) contain a 2OG oxygenase related in structure and function to the animal PHDs. A Pseudomonas aeruginosa PPHD knockout mutant displays impaired growth in the presence of iron chelators and increased production of the virulence factor pyocyanin. We identify elongation factor Tu (EF-Tu) as a PPHD substrate, which undergoes prolyl-4-hydroxylation on its switch I loop. A crystal structure of PPHD reveals striking similarity to human PHD2 and a Chlamydomonas reinhardtii prolyl-4-hydroxylase. A crystal structure of PPHD complexed with intact EF-Tu reveals that major conformational changes occur in both PPHD and EF-Tu, including a >20-Å movement of the EF-Tu switch I loop. Comparison of the PPHD structures with those of HIF and collagen PHDs reveals conservation in substrate recognition despite diverse biological roles and origins. The observed changes will be useful in designing new types of 2OG oxygenase inhibitors based on various conformational states, rather than active site iron chelators, which make up most reported 2OG oxygenase inhibitors. Structurally informed phylogenetic analyses suggest that the role of prolyl-hydroxylation in human hypoxia sensing has ancient origins.

The hypoxia-inducible transcription factor (HIF) is a major regulator of the response to limited oxygen availability in humans and other animals (1–3). A hypoxia-sensing component of the HIF system is provided by 2-oxoglutarate (2OG)-dependent and Fe(II)-dependent oxygenases, which catalyze prolyl-4-hydroxylation of HIF-α subunits, a posttranslational modification that enhances binding of HIF-α to the von Hippel-Lindau protein (pVHL), so targeting HIF-α for proteasomal degradation. The HIF prolyl-hydroxylases (PHDs) belong to a subfamily of 2OG oxygenases that catalyze prolyl-hydroxylation, which also includes the collagen prolyl-3-hydroxylases (CP3Hs) and prolyl-4-hydroxylases (CP4Hs) (4). Subsequently identified prolyl-hydroxylases include the ribosomal prolyl-hydroxylases (OGFOD1 and Tpa1), which catalyze ribosomal protein 23 prolyl-3-hydroxylation in many eukaryotes, and slime-mold enzymes, which catalyze prolyl-4-hydroxylation of Skp1, a ubiquitin ligase subunit (5–9). The HIF-PHD-VHL triad is likely present in all animals, but probably not in other organisms (3). However, structurally informed bioinformatic analyses imply the presence of PHD homologs in bacteria (10, 11), including in Pseudomonas spp, suggesting PHDs may have ancient origins.

Results

Pseudomonas spp. Contain a Functional PHD.

To investigate the role of a putative PHD homolog in Pseudomonas aeruginosa (PPHD), we initially characterized a PPHD insertional mutant strain. Metabolic screening studies revealed that the PPHD mutant strain displays impaired growth in the presence of iron chelators (e.g., 2,2-bipyridine) and produces increased levels of the bacterial virulence factor pyocyanin (SI Appendix, Fig. S1) (12), implying PPHD is involved in iron regulation. Ferrous iron is essential for PHD catalysis and treatment of human cells with iron chelators, including pyocyanin (13), and causes HIF-α up-regulation via PHD inhibition, suggesting related roles for PPHD and the animal PHDs (1).

Elongation Factor Tu Is a PPHD Substrate.

We then investigated possible PPHD substrates. Mass spectrometric (MS) studies identified elongation factor Tu (EF-Tu) as a PPHD binding partner and potential substrate in Pseudomonas putida (SI Appendix, Table S3). This assignment is supported by a proteomic MS study on Shewanella oneidensis, reporting that EF-Tu is modified by a +16-Da mass shift (14); bioinformatics reveal Shewanella oneidensis contains a potential hydroxylase closely related to PPHD (38% identity). We tested a set of 14 synthetic peptides covering the 19 prolines in the EF-Tu sequence as substrates for isolated recombinant PPHDputida (SI Appendix, Table S4). We identified an EF-Tu peptide as undergoing a +16-Da mass shift (Fig. 1B and SI Appendix, Table S4); MSMS analyses demonstrated hydroxylation at Pro54 (Hyp54; SI Appendix, Fig. S2A). Amino acid analyses reveal that PPHDputida catalyzes trans prolyl-4-hydroxylation of EF-Tu, as occurs for PHD-catalyzed HIF and CP4H-catalyzed collagen prolyl-hydroxylation (SI Appendix, Fig. S2B) (1). We did not observe cross hydroxylation reactivity between the PPHD and PHD2 substrates; that is, EF-Tu with PHD2 and the C-terminal oxygen degradation domain (CODD) of HIF-1α (CODD556–574) with PPHD.

Fig. 1.

Fig. 1.

Open in a new tab

Identification of EF-Tu as a PPHDputida substrate. (A) PPHDputida, human HIF (PHDs), and collagen (CP4Hs) prolyl-4-hydroxylases catalyze 4-trans prolyl-hydroxylation. (B) EF-Tu is a PPHDputida substrate; MS spectrum showing a +16-Da peak revealing PPHDputida-dependent prolyl-4-hydroxylation of EF-Tu44–63.

EF-Tu undergoes hydroxylation in the wild-type but not in the PPHD insertional mutant strain, as shown by MS studies (SI Appendix, Fig. S3A and B). Because the human elongation factor 2 (eEF2) interacts with PHD2 (15), we tested peptides derived from the switch I loop of human eEF2 as potential PPHD or PHD2 substrates but did not observe hydroxylation within the limits of detection.

Effect of Prolyl-Hydroxylation on EF-Tu Function.

Kinetic analyses were then carried out using a 20-residue EF-Tu fragment (EF-Tu44–63). Compared with an analogous substrate fragment with PHD2, PPHDputida has a similar apparent Km (47.2 ± 8.5 μM for PPHD/EF-Tu44–63; 36.7 ± 9.0 μM for PHD2/HIF-1α CODD), but a lower apparent Km for O2 (47.5 ± 9.2 μM for PPHD/EF-Tu44–63; 229 ± 60 μM for PHD2/HIF-1α CODD), and higher apparent Km values for 2OG (182 ± 22 μM for PPHD/EF-Tu44–63; 55 ± 11 μM for PHD2/HIF-1α CODD), Fe(II) (6.3 ± 0.7 μM for PPHD/EF-Tu44–63; <1 μM for PHD2/HIF-1α CODD), and l-ascorbate (1,064 ± 95 μM for PPHD/EF-Tu44–63; 54 ± 10 μM for PHD2/HIF-1α CODD; SI Appendix, Fig. S4 A–E) (16–20). Although the results with isolated enzyme may not reflect the cellular kinetics, given that PHD2 kinetics are proposed to reflect its role as a hypoxia sensor, these preliminary differences are of interest. In particular, the higher Km for Fe(II) is notable given the phenotypic results on the PPHD insertional mutant strain, supporting a role for PPHD in iron regulation.

The switch I loop of EF-Tu plays a central role in bacterial translation by coupling GTP hydrolysis with conformational changes from the GTP-bound active state to the GDP-bound inactive state (21, 22). NMR was used to directly monitor kirromycin-induced GTP hydrolysis by EF-Tuputida; we observed no difference in the GTP hydrolysis rate for Hyp54 EF-Tu (0.71 ± 0.02 μM/min) and unmodified EF-Tu (0.76 ± 0.08 μM/min; SI Appendix, Fig. S5A). However, addition of PPHDputida to the EF--Tu GTP hydrolysis reaction decreased the rate of GTP hydrolysis in a metal ion and 2OG-dependent manner [Zn(II) 0.60 ± 0.05 μM/min; Zn(II) and 2OG 0.49 ± 0.16 μM/min; Zn(II) was used as a surrogate for Fe(II) and a PPHDputida inhibitor; SI Appendix, Fig. S5 B–D]. In P. aeruginosa, we observed no difference in the global translation rate between the PPHD mutant strain and the wild-type under standard growth conditions, as measured by [35S]methionine incorporation (SI Appendix, Fig. S5E). The effect of PPHDputida on the GTP hydrolysis rate is less with hydroxylated EF-Tu (0.83 ± 0.08 μM/min) (SI Appendix, Fig. S5B), likely because of the weaker binding of hydroxylated EF-Tu to PPHD.

Using a paramagnetic relaxation enhancement-based NMR technique [with Mn(II) substituting for Fe(II)], we observed that addition of EF-Tuprotein decreased the PPHDputida 2OG Kd by ∼50-fold (8 ± 4 μM) compared with the absence of any substrate or in the presence of EF-Tu44–63 (>400 μM; SI Appendix, Fig. S5F), implying a substantial interaction interface between PPHDputida and EF-Tu. These results implied that very substantial interactions between PPHDputida and EF-Tu likely extend well beyond the immediate active site region. Notably, the observation that PPHDputida addition decreases the rate of GTP hydrolysis and the location of Pro54 on the switch I loop suggests that PPHDputida sequesters the switch I loop. We therefore pursued structural characterization of PPHDputida, EF-Tu, and the PPHD:EF-Tu complex.

Structure of PPHD.

To investigate the structural similarity of PPHDputida to PHD2, we determined a crystal structure of PPHDputida complexed with Mn(II) and N-oxalylglycine (NOG), which are inert substitutes for Fe(II) and 2OG, respectively (2.0-Å resolution; space group P1). Similar to other 2OG oxygenases, PPHDputida contains a double-stranded β-helix (DSBH) core fold comprising eight β-strands (I–VIII) folded into major and minor β-sheets that enclose the metal-/2OG-binding sites (Fig. 2A) (23, 24). Five antiparallel β-strands [β1-β2(I)-β4(III)-β7(VI)-β9(VIII)] form the major β-sheet, and four antiparallel β-strands [β3(II)-β5(IV)-β6(V)-β8(VII)] form the minor β-sheet. Three α-helices (α1–α3) pack against the major β-sheet of the DSBH. A disordered region (residues 47–71) of PPHDputida corresponds to the β2–β3 finger loop of PHD2, which is conformationally flexible and involved in HIF-α binding (10, 25). PPHDputida contains a HXD...H triad of metal-binding residues (His124, Asp126, and His183), which, with the NOG oxalyl group and a water, complete octahedral metal-coordination. Although nonidentical, the PPHDputida active site is very similar to that of PHD2 (10); NOG is positioned between the major and minor β-sheets of the DSBH such that its C-5 carboxylate forms a salt bridge with Arg192 (Arg383PHD2) (Fig. 2 B and C). The clear similarity between PPHDputida and PHD2 supports the proposal of common ancestry for the two hydroxylases (Fig. 2B).

Fig. 2.

Fig. 2.

Open in a new tab

PPHDputida is structurally similar to human PHD2. (A) Ribbons representation of the overall structure of PPHDputida; DSBH-associated β-strands (pale cyan), roman numerals indicate strands in the DSBH, DSBH-extension β-strands (dark blue), α-helices (red), loops (white), NOG (yellow), Mn(II) (pink sphere). (B) Superposition of the PPHDputida(red):NOG (yellow) and human PHD2 (blue):1 (gray) (PDB ID: 2G1M; rmsd, 1.7 Å over 181 Cα atoms) structures. (C) Superimposition of active site residues of PPHDputida (green):NOG (yellow) and human PHD2 (light-brown):1 (gray) (PDB ID: 2G1M) [1: N-(4-hydroxy-8-iodoisoquinolin-3-yl)carbonylglycine].

Structure of a PPHD:EF-Tu Complex.

We determined a P. putida EF-Tu structure (2.3-Å resolution; space group C2). EF-Tuputida and Escherichia coli EF-Tu have very similar overall folds (rmsd, 1.5 Å over 353 Cα atoms) and Mg(II) binding sites in their inactive GDP-bound states (Fig. 3A and SI Appendix, Fig. S6 A and B) (26). The switch I loop of EF-Tuputida (residues 40–62) is disordered; this is notable because the prolyl-hydroxylated regions of HIF-α are also predicted to be disordered in solution (18).

Fig. 3.

Fig. 3.

Open in a new tab

A PPHD:EF-Tu complex structure reveals major conformational changes in both proteins. (A) View from a PPHDputida (magenta):EF-Tu (cyan) complex crystal structure reveals EF-Tu in its GDP-bound state; the EF-Tu switch I loop is located in the PPHD active site positioning Pro54 in its endo-conformation for trans-4-hydroxylation. Six residues (43–48) in the switch I loop are disordered (dotted line). (Inset) details of how PPHD interacts with EF-Tu domain II. (B) The EF-Tu switch I loop is enclosed in the PPHD catalytic site; the electron density |Fo - Fc| map is in green mesh contoured to 3.0 σ. (C) Surface representation of the PPHDputida (magenta):EF-Tu (cyan) complex revealing enclosure of the PPHDputida β2–β3 finger loop and the β5(II)–β6(III) thumb loop around the EF-Tu switch I loop. (D) Superimposition of domains II and III of P. putida EF-Tu (cyan) in complex with PPHD (rmsd, 0.6 Å; PPHD not shown) and E. coli EF-Tu (yellow) (PDB ID: 2BVN) (27) bound to enacyloxin IIA (blue sticks) and a nonhydrolyzable GTP-analog guanosine 5′-(β,γ-imido)triphosphate (GDPNP) (red). (E) Superimposition of E. coli EF-Tu (PDB ID: 1DG1) (26) and P. putida EF-Tu from the PPHDputida:EF-Tu complex; the switch I loop is highlighted (P. putida EF-Tu, blue; E. coli EF-Tu, red). (F) Superimposition of the β5(II)–β6(III) thumb loop as in isolated PPHD (green) and in the PPHDputida (magenta):EF-Tu (cyan) complex reveals reorganization of thumb residues on substrate binding: Phe128PPHD rotates 90° to position its phenyl ring directly behind Pro54EF-Tu; Asp131PPHD forms a salt bridge with Arg133PPHD; Asp132PPHD rotates 180° to salt bridge to Arg234EF-Tu of EF-Tu domain II, allowing Arg134PPHD to hydrogen bond to Pro54EF-Tu.

To obtain a structure of a PPHD:EF-Tu protein–protein complex (Fig. 3 A–C), PPHD and EF-Tu were expressed separately to prevent EF-Tu hydroxylation and then mixed, purified, and crystallized using Mn(II) and NOG to stabilize and inactivate/inhibit PPHD, and Mg(II) and GDP to stabilize EF-Tu. In the resultant PPHD:EF-Tu complex structure (2.7-Å resolution; P31 2 1), EF-Tu is in its translationally inactive GDP-bound conformational state consistent with the observed PPHD-mediated inhibition of GTP hydrolysis. PPHDputida binds EF-Tu between domain I and domains II/III, occupying a position similar to that observed for EF-Tu domain I in its GTP-bound active state relative to EF-Tu domains II/III (Fig. 3D) (27). A Mg(II) ion is present in the EF-Tu nucleotide binding site, as in the EF-Tu structure without PPHDputida. Mn(II) and NOG are present in the PPHDputida active site.

In contrast to the isolated EF-Tuputida structure, in the complex, most EF-Tu switch I loop residues are ordered (16 residues disordered in isolated EF-Tu; 6 in the PPHD:EF-Tu complex). In a structure of isolated E. coli EF-Tu (in its inactive GDP-bound state), all of the switch I loop residues are ordered and located on the surface of EF-Tu domain I, adopting a loop-helix-loop-strand conformation (Fig. 3E) (26). Comparison of the P. putida PPHD:EF-Tu complex crystal structure with that of E. coli EF-Tu (78% identity; rmsd, 1.2 Å over 338 Cα atoms) reveals major differences. Relative to the isolated E. coli EF-Tu structure, in the PPHD:EF-Tu complex structure, the entire EF-Tu switch I loop pivots around residues Gly41EF-Tu and Ile61EF-Tu, such that Pro54EF-Tu moves 22 Å and is positioned at the apex of an extended loop projecting into the PPHDputida active site (Fig. 3E).

In contrast to the isolated PPHDputida structure, in the PPHD:EF-Tu complex, the PPHDputida β2–β3 finger loop (residues 51–74) is ordered and encloses the EF-Tu switch I loop within the active site (Fig. 3C). The residues at the apex of the flexible β2–β3 finger loop of PPHDputida adopt a 310-helix conformation (residues 58–60). An arginine (Arg57) before the 310-helix is positioned to form a salt bridge with Asp130 of the β5(II)-β6(III) “thumb” loop (seven residues, 127–133), forming a tunnel through which the EF-Tu switch I loop is threaded (Fig. 3C). The β5(II)–β6(III) thumb loop of PPHDputida contains three arginines, three aspartates (including Asp130), and one phenylalanine, the backbone and side chains of which make interactions with domain II of EF-Tu and undergo clear conformational changes relative to their positions in the isolated PPHDputida structure (rmsd, 1.4 Å; Fig. 3F and SI Appendix, Fig. S6C).

As indicated by the kinetic analyses in solution, the PPHD:EF-Tu structure reveals substantial enzyme-substrate contacts beyond the active site. PPHD and EF-Tu share a total interaction interface of 1,662 Å2 (24 hydrogen bonds; 10 salt bridges), resulting in a calculated ΔGbinding of ∼-9.7 kcal/mol, of which only 965 Å2 (13 hydrogen bonds; 3 salt bridges) is accounted for by direct interactions of PPHDputida with the EF-Tu switch I loop. Other interactions include positioning the PPHDputida α3-helix into a hydrophobic pocket in the β-barrel of EF-Tu domain II formed by Phe222EF-Tu, Ile224EF-Tu, and Val230EF-Tu, and burying Phe265EF-Tu in a hydrophobic pocket formed by the side chains of Leu101PPHD, Leu103PPHD, and Leu177PPHD (Fig. 3A, Inset). These interactions collectively order the PPHDputida thumb loop, suggesting that analogous interactions between PHD2 and full-length HIFα, not observed in a PHD2:HIF-α fragment structure (25), may occur.

Conservation of Substrate Binding by Prolyl-Hydroxylases.

The positions of the β2–β3 finger loop residues in the complex are strikingly similar to those observed for PHD2 in complex with HIF-α and to a collagen prolyl-4-hydroxylase-related protein in Chlamydomonas reinhardtii (CrP4H) when complexed with a proline-rich substrate (Fig. 4 A and B and SI Appendix, Fig. S6 D and E) (25, 28). In PHD2 and CrP4H, the β2–β3 finger loop (PHD2: 23 residues, 236–258; CrP4H: 24 residues, 76–99) and β5(II)–β6(III) thumb loop (PHD2: 6 residues; 316–321; CrP4H: 15 residues, 146–160) are of similar length and in the same locations as those of PPHDputida (Fig. 4 A and B and SI Appendix, Fig. S7 A and B); however, an analogous Arg57:Asp130 salt bridge that connects the finger and thumb loops is not present in PHD2 or CrP4H.

Fig. 4.

Fig. 4.

Open in a new tab

Pseudomonas PPHD, human PHD2, and C. reinhardtii P4H (CrP4H) share a conserved mechanism for substrate binding. Overall and active site superimpositions of the PPHDputida (magenta):EF-Tu (cyan) complex and both (A) human PHD2 (sand) in complex with HIF-α CODD (residues 556–574; PDB ID: 3HQR) (green) and (B) CrP4H (green) in complex with proline-rich peptide substrate (PDB ID: 3GZE) (yellow) reveal conservation of the roles of the β2–β3 finger loop and the β5(II)–β6(III) thumb loop, the orientation of the substrate backbone in the active site, and other key substrate binding interactions. Note that although CrP4H does not share a structurally conserved Arg133PPHD with PPHD, as does PHD2 (Inset), CrP4H does share a conserved Glu109PPHD (Glu127CrP4H, Thr296PHD2). (Inset) The PPHDputida and PHD2 C termini interact differently with their substrates. In contrast, the PHD2 C terminus interacts directly with HIF-α CODD (25).

The PHD2 C-terminal α-helix is involved in substrate binding and catalysis (16, 25). The PPHDputida C-terminal loop (residues 201–207, corresponding to PHD2 392–408) directly interacts with EF-Tu, as does the PHD2 C-terminal α-helix with HIF-α CODD (Fig. 4A, Inset), suggesting evolutionary conservation of a role for the C terminus in substrate recognition.

Comparison of the substrate complexes reveals that conservation is not limited to the enzymes but extends to the bound substrates. Strikingly, the unrelated substrates all adopt closely related conformations in their active site regions when bound to their respective prolyl-hydroxylase partners; EF-Tu (residues 50–57) complexed with PPHD shares a similar orientation and conformation to that of HIF-α (residues 560–567) complexed with PHD2 (backbone atoms rmsd, 0.1 Å) and to that of a proline-rich peptide complexed with CrP4H (backbone atoms rmsd, 0.1 Å; Fig. 4 A and B and SI Appendix, Fig. S6 D and E). However, there are differences in observed substrate binding for the prolyl-4-hydroxylases (PPHD, PHDs, CP4Hs) and that predicted for prolyl-3-hydroxylases (OGFOD1/Tpa1; SI Appendix, Fig. S7 A and B). The prolyl-4-hydroxylases have a conserved arginine at their active site opening positioned to hydrogen bond to the carbonyl of the substrate proline (Arg134PPHD, Arg322PHD2, Arg161CrP4H). Prolyl-4-hydroxylases also have a conserved tyrosine (Tyr121PPHD, Tyr310PHD2, Tyr140CrP4H), the hydroxyl of which points toward the nitrogen of the substrate proline (Fig. 4 A and B and SI Appendix, Fig. S6 D and E and S7B). In contrast, the prolyl-3-hydroxylases have a smaller hydrophobic leucine at the equivalent position to Tyr121PPHD (Leu156Tpa1, Leu152OGFOD1) (29), which may promote prolyl-3-hydroxylation over prolyl-4-hydroxylation by adjusting the position of the proline within the active site.

In the PPHDputida active site, the pyrrolidine ring of Pro54EF-Tu is observed in a C-4 endo conformation, poised for hydroxylation, as observed in the PHD2:HIF-α and CrP4H:proline-rich peptide complexes (Fig. 4 A and B and SI Appendix, Fig. S6 D and E) (25, 28, 30), and in contrast to the C-4 exo conformation, as observed for prolyl-hydroxylated HIF-1α when complexed with pVHL (1).

Discussion

The combined results suggest that roles for prolyl-hydroxylation are ancient, extending from EF-Tu in bacteria to, at least, collagen, HIF-α, and ribosomal protein 23 in eukaryotes. The results support growing evidence for the importance of 2OG oxygenases, including PHD homologs, in the regulation of translation, as well as of transcription (5–8, 29, 31–35). It is notable that the Jumonji C subfamily 2OG oxygenases also play roles in both histone/chromatin and ribosomal protein modification (5, 36).

Structurally guided bioinformatic analyses are informative with regard to the distribution of prolyl-hydroxylases. Comparison of the topologies of the prolyl-3-hydroxylases and prolyl-4-hydroxylases for which crystal structures are available reveals conservation not only of the DSBH, Fe(II), and 2OG binding sites but also of other functionally important elements including the β2–β3 finger loop and α-helices that pack against the DSBH (SI Appendix, Fig. S7 A and B), enabling assignment of putative prolyl-hydroxylases with a high confidence level. All identified subtypes of prolyl-4-hydroxylases (CP4Hs and PHDs) and prolyl-3-hydroxylases [Leprecans (mammalian collagen prolyl-3-hydroxylases) and OGFOD1s] are ubiquitous in animals, and all but the Leprecans are present in choanoflagellates. However, their presence in prokaryotes is sporadic, and they are apparently absent in archaea (Fig. 5 and SI Appendix, Fig. S8).

Fig. 5.

Fig. 5.

Open in a new tab

Distribution of prolyl-hydroxylases in kingdoms of life. Tracking putative homologs of the CP4Hs, PHDs, Leprecans (CP3Hs), and OGFOD1s in representative organisms throughout the kingdoms of life (for phylogenetic analyses encompassing a wider distribution, see SI Appendix, Fig. S8). CP4Hs and PHDs catalyze prolyl-4-hydroxylation. Human OGFOD1 catalyzes a single prolyl-3-hydroxylation, whereas its yeast homolog Tpa1 catalyzes a double prolyl-3- and prolyl-4-hydroxylation (7).

Given that HIF and collagen exist in animals, but not bacteria, protein prolyl-hydroxylases (and other enzymes catalyzing posttranslational modifications) may have first evolved to catalyze oxidation of highly abundant proteins, such as the switch I loop of EF-Tu, which are part of the translational machinery. Such protein hydroxylases were possibly derived from 2OG oxygenases acting on small molecules (e.g., proline-hydroxylases) (37), some of which have simpler structures than the protein-hydroxylases (23). In eukaryotes, the prolyl-hydroxylases evolved divergently to catalyze hydroxylation of collagen and other proteins involved in translational and transcriptional regulation [e.g., Skp1 and ribosomal protein 23 (6–9)]. Their role in HIF-α hydroxylation evolved in animals, possibly via fusion of a basic helix–loop–helix domain transcription factor with an EF-Tu, collagen, or Skp1-related or Skp1-derived substrate. The observation that PHD2 interacts with, but apparently does not hydroxylate, eEF2, the eukaryotic counterpart to prokaryotic EF-G, an EF-Tu homolog (15), is interesting because it may reflect an ancestral relationship between PPHD and EF-Tu. Competition between HIF-α and eEF2 [and potentially other PHD substrates; e.g., pyruvate kinase M2 (38) or centrosomal protein 192 (39)], may help regulate the hypoxic response in higher animals. Analogous competition is proposed for factor-inhibiting HIF (FIH) catalyzed HIF-1α asparagine hydroxylation, which reduces HIF activity by blocking its interaction with coactivator proteins (1). Factor-inhibiting HIF also takes ankyrin repeat domains as substrates in a manner proposed to enable a robust and tuneable hypoxic response (40).

The conformational changes observed on substrate binding in the PPHD:EF-Tu complex reveal that major induced fit during 2OG oxygenase catalyzed protein-modification can occur. The observed changes will be useful in designing new types of 2OG oxygenase inhibitors based on various conformational states, rather than active site iron chelators, which make up most reported 2OG oxygenase inhibitors, including PHD inhibitors, in clinical trials (41).

The finding that PPHD regulates the production of pyocyanin implies that iron regulation in P. aeruginosa may be important in terms of clinically relevant infections. The role of the PHDs in the hypoxic response has been highlighted in part because of medicinal opportunities relating to the treatment of ischemic and related diseases such as anemia (41). Although regulation of animal PHD activity by iron availability is well-established [cobaltous ion induced erythropoiesis is proposed to be via PHD inhibition (41)], it has had less prominence. P. aeruginosa infections, which often occur in hypoxic niches, such as in the cystic fibrosis lung (42), can induce a HIF-mediated hypoxic response in eukaryotic cells by production of small molecule factors, which have been proposed to be cyanide (43). Pyocyanin induces HIF-α up-regulation in human cells (13), likely via PHD inhibition, at least in part by iron sequestration. Thus, it seems that during some infections, prolyl-hydroxylases in both human and bacterial cells may be involved in “cross-talk” involving regulation limited not only by oxygen availability but also by that of iron and, possibly, 2OG (or other tricarboxylic acid cycle intermediate or related compounds) (44). Therefore, the biological roles of the PHDs and PPHDs likely extend beyond hypoxic sensing in an isolated organism.

From a clinical perspective, the work is relevant to ongoing trials involving PHD inhibitors for anemia treatment (41). At least some PHD inhibitors inhibit other 2OG oxygenases (41, 45). There is also increasing use of metal chelators, including orally active compounds (which inhibit the PHDs by iron chelation) (46), for iron overload diseases. Although our work does not predict what effect, if any, application of clinically applied PHD inhibitors/iron chelators will have on the course of infections by bacteria containing PHD homologs, the potential for effects of PHD inhibitors on infection should be considered.

Materials and Methods

Protein Purification and Crystallization.

Recombinant PPHDputida and EF-Tuputida were produced in E. coli and purified using Ni-affinity and size-exclusion chromatography. For purification of the PPHD:EF-Tu complex, cell pellets containing either PPHDputida or EF-Tuputida were colysed with Mn(II) and NOG to prevent EF-Tu hydroxylation and were purified in the same way. For details of crystallization conditions, data collection and processing, and refinement statistics, see SI Appendix, SI Materials and Methods, and Tables S1 and S2.

P. aeruginosa Phenotypic Analyses.

P. aeruginosa PAO1 wild-type and PPHD insertional knockout mutant strain were grown (28 °C) in various concentrations of 2,2-bipyridine (0, 1, 5, and 25 mM); the optical density (600 nm, OD600) was assayed every 60 min. For pyocyanin quantification, 5-mL cultures of P. aeruginosa wild-type and PA0310 mutant strains grown to stationary phase were extracted with 3 mL CHCl3 and then back-extracted with 1 mL 0.2 HCl, and then the aqueous phase was decanted and its absorbance measured at 520 nm. For global translation rate assays, 1-mL cultures of P. aeruginosa wild-type and PA0310 mutant strains were grown to log phase (OD600 = 0.5), and the rate of incorporation of [35S]methionine (10 μCi/mL) was measured by scintillation counting.

Determination of EF-Tu as a PPHD Substrate.

Hydroxylation assays for PPHDputida contained 10 μM PPHD, 100 μM EF-Tu44–63 (H2N-IVEFDKIDSAPEEKARGITI-CONH2), 50 μM Fe(II) [prepared from (NH4)2Fe(SO4)2], 2 mM 2OG, 4 mM sodium l-ascorbate in 50 mM Hepes at pH 7.5, 500 mM NaCl, and 5% (vol/vol) glycerol and were quantified by MALDI MS. MALDI MS/MS was performed to determine the site of hydroxylation. To determine the stereochemistry of hydroxylation, acid hydrolyzed of Hyp54-modified EF-Tu44–63 was derivatized and compared with the elution times of hydroxyproline standards. To analyze EF-Tu Pro54 hydroxylation, cells were lysed and analyzed by SDS-PAGE, and the EF-Tu-containing band was excised and subjected to proteosomal digestion (ArgC) and liquid chromatography-MS/MS. For purification of recombinant Hyp54 EF-Tu, PPHDputida, Fe(II), 2OG, and sodium l-ascorbic acid were added to the cell lysate at a final concentration of 1 mg/mL, 50 μM, 2 mM, and 4 mM, respectively, and incubated for 1 h at room temperature. The hydroxylation status of Pro54 was then analyzed by trypsinolysis and MSMS analysis.

NMR Experiments.

Measurements of Mn(II) and 2OG Kd to PPHDputida used a Bruker AVII 500 instrument. Mixture (160 μL) contained 50 μM apo (i.e. no metal) PPHDputida, 50 μM MnCl2, 50 μM EF-Tu (full length or EF-Tu44–63 oligopeptide if required), 10 mM MgCl2 and 200 mM NaCl (buffer: ~30 mM Tris-D11/~20 mM HEPES pH 7.5 in 36.25% H2O and 63.75% D2O) (SI Appendix, Fig. S6 F and G). For EF-Tu GTP-GDP turnover assays, we used a Bruker AVIII 700 instrument/5-mm inverse cryoprobe. Incubations (160 μL) contained 10 μM unmodified or Hyp54 EF-Tu, 50 μM kirromycin, 100 μM GTP, 50 μM apo-PPHDputida (if required), 100 μM ZnCl2 (if required), 100 μM 2OG (if required), 10 mM MgCl2, and 200 mM NaCl [buffer: 50 mM HEPES (pH 7.5) in 90% H2O/10% D2O]. Solutions were buffered using ∼30 mM Tris-d11 and ∼20 mM Hepes at pH 7.5 in 36.25% H2O and 63.75% D2O]. For ZnCl2 inhibition experiments, the mixture contained 10 μM apo PPHDputida, 50 μM FeCl2, 500 μM 2OG, 600 μM EF-Tu44–63, 400 μM ZnCl2 (if required), 1.25 mM ascorbate, and 100 mM NaCl buffered using 50 mM Tris-D11 at pH 7.5 (SI Appendix, Fig. S5D).

See SI Appendix, SI Materials and Methods.

Supplementary Material

Supplementary File

Acknowledgments

We thank the Diamond Macromolecular Crystallography beamline staff for technical support; Dr. Christoph Loenarz and Prof. Peter J. Ratcliffe for helpful discussion; and the Rhodes Trust (J.S.S.), the British Heart Foundation (I.K.H.L.), the Wellcome Trust, the European Research Council, the Japan Society for the Promotion of Science (S.H.), and the Biotechnology and Biological Sciences Research Council for financial support.

Footnotes

The authors declare no conflict of interest.

Data deposition: The atomic coordinates have been deposited in the Protein Data Bank, www.pdb.org (PDB ID codes 4J25, 4J0Q, 4IW3), and sequence data deposited in Uniprot, www.uniprot.org/uniprot (accession nos. Q88CM1 and Q9I6I1) and GenBank, www.ncbi.nih.gov/protein (accession nos. AAN70724.1 and AAG03699.1).

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1409916111/-/DCSupplemental.

References

  • 1.Kaelin WG, Jr, Ratcliffe PJ. Oxygen sensing by metazoans: The central role of the HIF hydroxylase pathway. Mol Cell. 2008;30(4):393–402. doi: 10.1016/j.molcel.2008.04.009. [DOI] [PubMed] [Google Scholar]
  • 2.Semenza GL. Oxygen sensing, homeostasis, and disease. N Engl J Med. 2011;365(6):537–547. doi: 10.1056/NEJMra1011165. [DOI] [PubMed] [Google Scholar]
  • 3.Loenarz C, et al. The hypoxia-inducible transcription factor pathway regulates oxygen sensing in the simplest animal, Trichoplax adhaerens. EMBO Rep. 2011;12(1):63–70. doi: 10.1038/embor.2010.170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Myllyharju J, Kivirikko KI. Collagens, modifying enzymes and their mutations in humans, flies and worms. Trends Genet. 2004;20(1):33–43. doi: 10.1016/j.tig.2003.11.004. [DOI] [PubMed] [Google Scholar]
  • 5.Ge W, et al. Oxygenase-catalyzed ribosome hydroxylation occurs in prokaryotes and humans. Nat Chem Biol. 2012;8(12):960–962. doi: 10.1038/nchembio.1093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Singleton RS, et al. OGFOD1 catalyzes prolyl hydroxylation of RPS23 and is involved in translation control and stress granule formation. Proc Natl Acad Sci USA. 2014;111(11):4031–4036. doi: 10.1073/pnas.1314482111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Loenarz C, et al. Hydroxylation of the eukaryotic ribosomal decoding center affects translational accuracy. Proc Natl Acad Sci USA. 2014;111(11):4019–4024. doi: 10.1073/pnas.1311750111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Katz MJ, et al. Sudestada1, a Drosophila ribosomal prolyl-hydroxylase required for mRNA translation, cell homeostasis, and organ growth. Proc Natl Acad Sci USA. 2014;111(11):4025–4030. doi: 10.1073/pnas.1314485111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.van der Wel H, Ercan A, West CM. The Skp1 prolyl hydroxylase from Dictyostelium is related to the hypoxia-inducible factor-alpha class of animal prolyl 4-hydroxylases. J Biol Chem. 2005;280(15):14645–14655. doi: 10.1074/jbc.M500600200. [DOI] [PubMed] [Google Scholar]
  • 10.McDonough MA, et al. Cellular oxygen sensing: Crystal structure of hypoxia-inducible factor prolyl hydroxylase (PHD2) Proc Natl Acad Sci USA. 2006;103(26):9814–9819. doi: 10.1073/pnas.0601283103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Aravind L, Koonin EV. 2001. The DNA-repair protein AlkB, EGL-9, and leprecan define new families of 2-oxoglutarate- and iron-dependent dioxygenases. Genome Biol 2(3):RESEARCH0007.
  • 12.Lau GW, Hassett DJ, Ran H, Kong F. The role of pyocyanin in Pseudomonas aeruginosa infection. Trends Mol Med. 2004;10(12):599–606. doi: 10.1016/j.molmed.2004.10.002. [DOI] [PubMed] [Google Scholar]
  • 13.Ahmad IM, Britigan BE, Abdalla MY. Oxidation of thiols and modification of redox-sensitive signaling in human lung epithelial cells exposed to Pseudomonas pyocyanin. J Toxicol Environ Health A. 2011;74(1):43–51. doi: 10.1080/15287394.2010.514233. [DOI] [PubMed] [Google Scholar]
  • 14.Gupta N, et al. Whole proteome analysis of post-translational modifications: Applications of mass-spectrometry for proteogenomic annotation. Genome Res. 2007;17(9):1362–1377. doi: 10.1101/gr.6427907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Romero-Ruiz A, et al. Prolyl hydroxylase-dependent modulation of eukaryotic elongation factor 2 activity and protein translation under acute hypoxia. J Biol Chem. 2012;287(12):9651–9658. doi: 10.1074/jbc.M111.299180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Flashman E, et al. Kinetic rationale for selectivity toward N- and C-terminal oxygen-dependent degradation domain substrates mediated by a loop region of hypoxia-inducible factor prolyl hydroxylases. J Biol Chem. 2008;283(7):3808–3815. doi: 10.1074/jbc.M707411200. [DOI] [PubMed] [Google Scholar]
  • 17.Ehrismann D, et al. Studies on the activity of the hypoxia-inducible-factor hydroxylases using an oxygen consumption assay. Biochem J. 2007;401(1):227–234. doi: 10.1042/BJ20061151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.McNeill LA, et al. Hypoxia-inducible factor prolyl hydroxylase 2 has a high affinity for ferrous iron and 2-oxoglutarate. Mol Biosyst. 2005;1(4):321–324. doi: 10.1039/b511249b. [DOI] [PubMed] [Google Scholar]
  • 19.Hirsilä M, Koivunen P, Günzler V, Kivirikko KI, Myllyharju J. Characterization of the human prolyl 4-hydroxylases that modify the hypoxia-inducible factor. J Biol Chem. 2003;278(33):30772–30780. doi: 10.1074/jbc.M304982200. [DOI] [PubMed] [Google Scholar]
  • 20.Flashman E, Davies SL, Yeoh KK, Schofield CJ. Investigating the dependence of the hypoxia-inducible factor hydroxylases (factor inhibiting HIF and prolyl hydroxylase domain 2) on ascorbate and other reducing agents. Biochem J. 2010;427(1):135–142. doi: 10.1042/BJ20091609. [DOI] [PubMed] [Google Scholar]
  • 21.Voorhees RM, Schmeing TM, Kelley AC, Ramakrishnan V. The mechanism for activation of GTP hydrolysis on the ribosome. Science. 2010;330(6005):835–838. doi: 10.1126/science.1194460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Villa E, et al. Ribosome-induced changes in elongation factor Tu conformation control GTP hydrolysis. Proc Natl Acad Sci USA. 2009;106(4):1063–1068. doi: 10.1073/pnas.0811370106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Clifton IJ, et al. Structural studies on 2-oxoglutarate oxygenases and related double-stranded beta-helix fold proteins. J Inorg Biochem. 2006;100(4):644–669. doi: 10.1016/j.jinorgbio.2006.01.024. [DOI] [PubMed] [Google Scholar]
  • 24.Aik W, McDonough MA, Thalhammer A, Chowdhury R, Schofield CJ. Role of the jelly-roll fold in substrate binding by 2-oxoglutarate oxygenases. Curr Opin Struct Biol. 2012;22(6):691–700. doi: 10.1016/j.sbi.2012.10.001. [DOI] [PubMed] [Google Scholar]
  • 25.Chowdhury R, et al. Structural basis for binding of hypoxia-inducible factor to the oxygen-sensing prolyl hydroxylases. Structure. 2009;17(7):981–989. doi: 10.1016/j.str.2009.06.002. [DOI] [PubMed] [Google Scholar]
  • 26.Abel K, Yoder MD, Hilgenfeld R, Jurnak F. An alpha to beta conformational switch in EF-Tu. Structure. 1996;4(10):1153–1159. doi: 10.1016/s0969-2126(96)00123-2. [DOI] [PubMed] [Google Scholar]
  • 27.Parmeggiani A, et al. Enacyloxin IIa pinpoints a binding pocket of elongation factor Tu for development of novel antibiotics. J Biol Chem. 2006;281(5):2893–2900. doi: 10.1074/jbc.M505951200. [DOI] [PubMed] [Google Scholar]
  • 28.Koski MK, et al. The crystal structure of an algal prolyl 4-hydroxylase complexed with a proline-rich peptide reveals a novel buried tripeptide binding motif. J Biol Chem. 2009;284(37):25290–25301. doi: 10.1074/jbc.M109.014050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Kim HS, et al. Crystal structure of Tpa1 from Saccharomyces cerevisiae, a component of the messenger ribonucleoprotein complex. Nucleic Acids Res. 2010;38(6):2099–2110. doi: 10.1093/nar/gkp1151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Loenarz C, et al. Evidence for a stereoelectronic effect in human oxygen sensing. Angew Chem Int Ed Engl. 2009;48(10):1784–1787. doi: 10.1002/anie.200805427. [DOI] [PubMed] [Google Scholar]
  • 31.Jia G, Fu Y, He C. Reversible RNA adenosine methylation in biological regulation. Trends Genet. 2013;29(2):108–115. doi: 10.1016/j.tig.2012.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Feng T, et al. Optimal translational termination requires C4 lysyl hydroxylation of eRF1. Mol Cell. 2014;53(4):645–654. doi: 10.1016/j.molcel.2013.12.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Henri J, et al. Structural and functional insights into Saccharomyces cerevisiae Tpa1, a putative prolylhydroxylase influencing translation termination and transcription. J Biol Chem. 2010;285(40):30767–30778. doi: 10.1074/jbc.M110.106864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Lee CY, Yeh TL, Hughes BT, Espenshade PJ. Regulation of the Sre1 hypoxic transcription factor by oxygen-dependent control of DNA binding. Mol Cell. 2011;44(2):225–234. doi: 10.1016/j.molcel.2011.08.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.van den Born E, et al. ALKBH8-mediated formation of a novel diastereomeric pair of wobble nucleosides in mammalian tRNA. Nat Commun. 2011;2:172. doi: 10.1038/ncomms1173. [DOI] [PubMed] [Google Scholar]
  • 36.Klose RJ, Zhang Y. Regulation of histone methylation by demethylimination and demethylation. Nat Rev Mol Cell Biol. 2007;8(4):307–318. doi: 10.1038/nrm2143. [DOI] [PubMed] [Google Scholar]
  • 37.Hausinger RP. FeII/alpha-ketoglutarate-dependent hydroxylases and related enzymes. Crit Rev Biochem Mol Biol. 2004;39(1):21–68. doi: 10.1080/10409230490440541. [DOI] [PubMed] [Google Scholar]
  • 38.Luo W, et al. Pyruvate kinase M2 is a PHD3-stimulated coactivator for hypoxia-inducible factor 1. Cell. 2011;145(5):732–744. doi: 10.1016/j.cell.2011.03.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Moser SC, et al. PHD1 links cell-cycle progression to oxygen sensing through hydroxylation of the centrosomal protein Cep192. Dev Cell. 2013;26(4):381–392. doi: 10.1016/j.devcel.2013.06.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Cockman ME, et al. Posttranslational hydroxylation of ankyrin repeats in IkappaB proteins by the hypoxia-inducible factor (HIF) asparaginyl hydroxylase, factor inhibiting HIF (FIH) Proc Natl Acad Sci USA. 2006;103(40):14767–14772. doi: 10.1073/pnas.0606877103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Rose NR, McDonough MA, King ON, Kawamura A, Schofield CJ. Inhibition of 2-oxoglutarate dependent oxygenases. Chem Soc Rev. 2011;40(8):4364–4397. doi: 10.1039/c0cs00203h. [DOI] [PubMed] [Google Scholar]
  • 42.Schaible B, Schaffer K, Taylor CT. Hypoxia, innate immunity and infection in the lung. Respir Physiol Neurobiol. 2010;174(3):235–243. doi: 10.1016/j.resp.2010.08.006. [DOI] [PubMed] [Google Scholar]
  • 43.Budde MW, Roth MB. The response of Caenorhabditis elegans to hydrogen sulfide and hydrogen cyanide. Genetics. 2011;189(2):521–532. doi: 10.1534/genetics.111.129841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Frezza C, Pollard PJ, Gottlieb E. Inborn and acquired metabolic defects in cancer. J Mol Med (Berl) 2011;89(3):213–220. doi: 10.1007/s00109-011-0728-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Aik W, et al. Structural basis for inhibition of the fat mass and obesity associated protein (FTO) J Med Chem. 2013;56(9):3680–3688. doi: 10.1021/jm400193d. [DOI] [PubMed] [Google Scholar]
  • 46.Cho EA, et al. Differential in vitro and cellular effects of iron chelators for hypoxia inducible factor hydroxylases. J Cell Biochem. 2013;114(4):864–873. doi: 10.1002/jcb.24423. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary File
pnas.1409916111.sapp.pdf (6.1MB, pdf)

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES