Journal list menu
Human phytanoyl-CoA dioxygenase domain-containing 1 (PHYHD1) is a putative oxygen sensor associated with RNA and carbohydrate metabolism
Abstract
Human phytanoyl-CoA dioxygenase domain-containing 1 (PHYHD1) is a 2-oxoglutarate (2OG)-dependent dioxygenase implicated in Alzheimer's disease, some cancers, and immune cell functions. The substrate, kinetic and inhibitory properties, function and subcellular localization of PHYHD1 are unknown. We used recombinant expression and enzymatic, biochemical, biophysical, cellular and microscopic assays for their determination. The apparent Km values of PHYHD1 for 2OG, Fe2+ and O2 were 27, 6 and > 200 μm, respectively. PHYHD1 activity was tested in the presence of 2OG analogues, and it was found to be inhibited by succinate and fumarate but not R-2-hydroxyglutarate, whereas citrate acted as an allosteric activator. PHYHD1 bound mRNA, but its catalytic activity was inhibited upon interaction. PHYHD1 was found both in the nucleus and cytoplasm. Interactome analyses linked PHYHD1 to cell division and RNA metabolism, while phenotype analyses linked it to carbohydrate metabolism. Thus, PHYHD1 is a potential novel oxygen sensor regulated by mRNA and citrate.
Abbreviations
2HG, 2-hydroxyglutarate
2OG, 2-oxoglutarate
2OGDD, 2-oxoglutarate-dependent dioxygenase
C-P4H, collagen prolyl 4-hydroxylase
FIH, factor inhibiting HIF
HIF, hypoxia-inducible factor
HIF-P4H, HIF prolyl 4-hydroxylase
IDH, isocitrate dehydrogenase
KDM, histone lysine demethylase
PHYH, phytanoyl-CoA hydroxylase
TCA, tricarboxylic acid
TET, DNA methylcytosine demethylase
2-Oxoglutarate-dependent dioxygenases (2OGDDs) are an enzyme family of > 60 members in humans that share the same reaction mechanism, cofactor and cosubstrates: Fe2+, 2-oxoglutarate (2OG) and molecular oxygen, respectively [[1, 2]]. Vitamin C (or another reducing agent) is not a direct cofactor but supports the reaction [[1, 2]]. 2OGDDs oxidize (often hydroxylate) their substrate, which vary from protein to nucleic and fatty acids, and other small molecules [[1]]. The hydroxylated substrate can undergo further non-enzymatic modifications, such as demethylation [[1]]. Well-studied 2OGDDs include the prolyl 4-hydroxylases that act on collagen (C-P4H) or hypoxia-inducible factor (HIF) (HIF-P4Hs/PHDs), the HIF asparaginyl hydroxylase (FIH), the chromatin modifying histone lysine demethylases (KDMs) and the DNA methylcytosine demethylases (TETs) [[1, 2]]. The requirements of different 2OGDDs for the cosubstrates and cofactor vary and low affinity (high Km value) offers potential for regulation when the cosubstrate/cofactor concentration becomes limiting (such as oxygen levels under hypoxia). 2OG is a tricarboxylic acid (TCA) cycle intermediate produced by one of the isocitrate dehydrogenase (IDH) isoenzymes or oxidative deamination of glutamate by glutamate dehydrogenase. The TCA cycle intermediates and structural 2OG analogues fumarate and succinate, as well as R-2-hydroxyglutarate (R-2HG) generated by IDH mutant cancers, and its enantiomer S-2HG accumulating under hypoxia due to promiscuous enzyme catalysis, can inhibit to various extent the catalytic activity of 2OGDDs providing another level of regulation [[2]]. 2OGDDs are also targets for pharmaceutical development and small molecule inhibitors of HIF-P4Hs have been accepted for the treatment of anaemia due to chronic kidney disease [[2]].
The exact substrate and therefore cellular function of ~ 15 2OGDDs is yet to be discovered. These include phytanoyl-CoA dioxygenase domain containing 1 (PHYHD1), a structural analogue of phytanoyl-CoA hydroxylase (PHYH/PAHX), which acts on a fatty acid, resides in peroxisomes and is mutated in the neurological Refsum disease [[3]]. PHYHD1 exists in three forms of which the 291-residue long form A appears to be the catalytically active one [[4]]. However, it does not act on phytanoyl-CoA or other tested CoA-derivatives [[4]]. Structural and initial kinetic analyses confirm that PHYHD1 is a 2OGDD possessing the typical double-stranded β-helix (DSBH) structure and having the conserved Fe2+ and 2OG binding residues hydroxylating a yet unidentified prime substrate in a reaction coupled with decarboxylation of 2OG to CO2 [[4]]. The 2OGDD-catalyzed reaction is ordered; the binding of Fe2+ is followed by binding of 2OG, primary substrate and the last oxygen [[1]]. The hydroxylated product is the first to leave the enzyme, followed by CO2 and succinate as the decarboxylation products of 2OG and finally Fe2+ [[1]]. However, 2OGDDs are also known to catalyze 2OG decarboxylation that is not coupled to substrate hydroxylation in the absence of their prime substrate, and to a small extent even in the presence of the substrate when the active site of the enzyme interacts with a nonhydroxylatable sequence [[5, 6]]. In the uncoupled reaction ascorbate acts as an alternative oxygen acceptor [[6, 7]]. This reaction is known to obey the same kinetics to the substrate hydroxylation-coupled reaction for 2OG, and due to the ordered reaction [[1]] also likely for Fe2+ which binds before 2OG. However, since oxygen binds last [[1]], its binding affinity may be influenced by substrate and therefore the uncoupled kinetics may not fully reflect binding affinity of O2 for a substrate [[7]].
PHYHD1 has been associated with Alzheimer's disease, some cancers and immune cell functions [[8-14]]. Recently, its genomic variants were associated with altered levels of the nucleotide metabolites guanine, 5-methylcytidine and 5-methyl uridine [[15, 16]]. We set out to study here the kinetic and inhibitory determinants of PHYHD1, such as its activity towards candidate substrate, sensitivity to hypoxia and inhibition by 2OG analogues, its subcellular localization, interactions with other proteins and its contribution to cellular metabolism to gain more understanding of its cellular role and functions.
Results
Analysis of the molecular and biophysical properties and catalytic activity of the purified recombinant PHYHD1
His-tagged human PHYHD1 was successfully expressed as a full-length recombinant protein and purified using a Ni-NTA affinity chromatography and a size exclusion chromatography (SEC) method. SDS-PAGE analysis of the elution fractions indicated a correct size protein of purity of ~ 80% (Fig. 1A). The SEC profile of PHYHD1 showed one major peak (Fig. 1A). The calculated molecular mass of PHYHD1 by SEC-multi angle light scattering was 35 kDa (Fig. S1), while its theoretical monomeric molecular mass is 33 kDa. These results suggest that PHYHD1 exists as a monomer in solution.
Since the PHYHD1 substrate is unknown, we exploited here the uncoupled reaction to study catalytic properties of PHYHD1 (Fig. 1B). The catalytic activity of PHYHD1 was thus measured based on the scintillation of 14CO2 generated as a result of decarboxylation of 2-oxo[1-14C]glutarate (Fig. 1B). The activity of PHYHD1 was shown to be dependent on time and pH of the reaction buffer (Fig. 1C,D). Based on these, the following experiments were carried out for 10 min of reaction time (Fig. 1C) and at pH 5.8 (Fig. 1D).
Analysis of the kinetic properties of PHYHD1 for reaction cofactor and cosubstrates and activity towards candidate substrates
The catalytic activity of 2OGDDs is dependent on the concentration of their cofactor (Fe2+) and cosubstrates (2OG, O2) and their affinities for PHYHD1 were determined enzymatically. PHYHD1 activity was highly dependent on Fe2+ shown by iron chelation (Fig. S2). The activity was restored by adding back iron concentration dependently and to its full extent (Fig. S2). Addition of iron and 2OG followed the Michaelis–Menten kinetics; the average apparent Km values for Fe2+ and 2OG for PHYHD1 being 6 and 27 μm, respectively (Fig. 2A,B, Table 1, Fig. S3). However, the reaction did not become saturated with increasing concentration of O2 suggesting the substrate hydroxylation uncoupled reaction of PHYHD1 has a higher O2 Km value than the atmospheric O2 concentration at sea level (Fig. 2C, Table 1, Fig. S3).
2OG | Fe2+ | O2 | Fumarate | Succinate | R-2HG | S-2HG | Citrate | |
---|---|---|---|---|---|---|---|---|
Km (μm) | IC50 (μm) | |||||||
PHYHD1 | 27 ± 9a | 6 ± 1a | > 200a | 1290 ± 480a | 830 ± 100a | > 5000a | 4030 ± 960a | 380 ± 20c |
KDM4A | 15 [[25]] | <0.1 [[25]] | 55–170 [[21]] | 2300 [[25]] | 800 [[25]] | 160 [[25]] | 290 [[25]] | 800 [[25]] |
KDM4B | 6 [[25]] | <0.1 [[25]] | 150 [[21]] | > 5000 [[25]] | 2300 [[25]] | 150 [[25]] | 450 [[25]] | 1600 [[25]] |
KDM5B | 10 [[25]] | <0.1 [[25]] | 40 [[21]] | > 5000 [[25]] | 1400 [[25]] | 3600 [[25]] | 1600 [[25]] | 800 [[25]] |
KDM6A | 8 [[25]] | <0.1 [[25]] | 200 [[21]] | 3000 [[25]] | 270 [[25]] | 180 [[25]] | 180 [[25]] | 290 [[25]] |
KDM6B | 50 [[25]] | 6 [[25]] | 25 [[21]] | > 5000 [[25]] | 550 [[25]] | 350 [[25]] | 750 [[25]] | 130 [[25]] |
HIF-P4H-1 | 2 [[2]] | 0.05 [[2]] | 230 [[2]] | 120 [[2]] | 830 [[2]] | Not an inhibitor, 210c [[2]] | 630d [[2]] | 6300 [[32]] |
HIF-P4H-2 | 1–270 [[2]] | 0.05 [[2]] | 65–250, > 450 [[2]] | 80 [[2]] | 510 [[2]] | 7900, 300c [[2]] | 420–1150 [[2]] | 4800, 1800d [[32]] |
FIH | 25 [[2]] | 0.5 [[2]] | 90–240 [[2]] | > 100 [[32]] | > 10 000 [[32]] | 1100–1500 [[2, 32]] | 190–300 [[2, 32]] | 850, 110d [[32]] |
C-P4H-I | 20 [[2]] | 2 [[2]] | 40 [[2]] | 190 [[2]] | 400 [[2]] | 1800 [[2]] | 310 [[2]] | 450d [[32]] |
TET2 | 55 [[33]] | 4.8 [[33]] | 0.5–30b [[2, 33]] | 400 [[33]] | 570 [[33]] | 5000 [[33]] | 1600 [[33]] | > 5000 [[33]] |
- a Values are means ± SD of at least 3 independent assays.
- b Value determined in our laboratory.
- c Km, not IC50.
- d Ki, not IC50.
PHYHD1 purified in the absence of Fe2+ did not exhibit any activity without addition of FeSO4 to the reaction mixture (Fig. 3A). This was used as a negative control when testing various substrate candidates, including total RNA, trimethylated histone H3 peptides, phytanoyl-CoA, genomic DNA, methylated DNA, rRNA, methylated rRNA and mRNA, for PHYHD1 (Fig. 3A–C). Our data showed that none of the tested compounds increased 2OG turnover significantly over the level of the uncoupled decarboxylation detected in the presence of Fe2+ suggesting that none of these is the primary substrate of PHYHD1 (Fig. 3A,B). PHYHD1 activity was however inhibited by mRNA and rRNA in a dose dependent manner (Fig. 3B). Among 2OGDDs, this appeared specific for PHYHD1 since increasing concentration of mRNA did not significantly inhibit uncoupled activity of HIF hydroxylating prolyl 4-hydroxylase-2 (HIF-P4H-2) or collagen hydroxylating collagen prolyl 4-hydroxylase-I (Fig. S4). Ligand binding often increases or decreases thermal stability of a protein which can be detected as a shift in nano differential scanning fluorimetry (nanoDSF) [[17, 18]]. Therefore, as a further method to evaluate the interaction of PHYHD1 with RNA, we measured thermostability of PHYHD1 when incubated without or with mRNA/rRNA/methylated rRNA using nanoDSF. The melting curves of PHYHD1 in the presence of mRNA, rRNA or methylated rRNA displayed differences suggesting an interaction (Fig. S5). The calculated shifts in the melting temperature (ΔTm) of PHYHD1 are presented in Fig. 3C. The major change in PHYHD1 thermal stability was detected through interaction with mRNA that followed a concentration-dependent saturation kinetics and gave a Kd of 1.95 ng·μL−1 (Fig. 3C).
The 3D structures of PHYHD1 and PHYH were analyzed and found to share comparable catalytic residues and a similar pocket suggesting alike reaction mechanisms, as also described previously [[4]]. The substrate binding pocket in both PHYH and PHYHD1 structures are lined by hydrophobic residues suggesting that the prime substrate of PHYHD1 could be a complementary hydrophobic molecule [[4]]. In this study, we analyzed the substrate binding pocket of both PHYHD1 and PHYH in detail using the CASTp webserver [[19]]. The solvent accessible (SA) volume of the substrate binding pocket of PHYH and PHYHD1 were estimated to be 669 Å3 and 606 Å3, respectively (Fig. S6). Furthermore, the α3 helix (residues 46–70) in PHYHD1 loops over the pocket making the cavity smaller and more closed compared to PHYH (Fig. S6).
PHYHD1 is susceptible to inhibition by several 2OG analogues, but not by (R)2-hydroxyglutarate or citrate, the latter enhancing its catalytic activity
It is known that 2OGDD enzymes are susceptible to inhibition by the cancer associated and metabolic regulator 2OG analogues [[2]]. Catalytic activity of PHYHD1 was inhibited by fumarate, succinate and (S)-2HG with the calculated average IC50 values being 1290, 830 and 4030 μm, respectively, the IDH mutation associated oncometabolite (R)-2HG inhibiting PHYHD1 activity very poorly, having an IC50 value of > 5000 μm (Fig. 4A–D, Table 1, Fig. S7). Surprisingly, our data showed that the catalytic activity of PHYHD1 was not inhibited by citrate but in fact, it increased 2OG turnover concentration dependently and thus enhanced its activity, the Km for it being 2380 μm (Fig. 4E, Table 1, Fig. S7). Altogether, these data suggest that succinate and fumarate are the most efficient inhibitors of PHYHD1 among the tested 2OG analogues.
PHYHD1 antibody detects endogenous and recombinant PHYHD1 and suggests a heterogeneous subcellular localization
We raised a polyclonal antibody against the purified recombinant human PHYHD1 in rabbit. The antibody specifically detected the endogenous PHYHD1 from HepG2 and murine primary hepatocyte lysates in addition to the recombinant human PHYHD1 by western blotting (Fig. 5A).
To map the subcellular localization of PHYHD1, we used two different approaches. First, we visualized endogenous PHYHD1 by using two different antibodies, the self-made and a commercial one, with immunofluorescence staining in two different human cell lines, HEK293T and HepG2 (Fig. 5B). Confocal microscopy revealed that endogenous PHYHD1 was localized both to nucleus and cytoplasm and its heterogeneous localization was consistent in both cell lines (Fig. 5B). Next, we overexpressed human PHYHD1 fused to mCherry in HEK293T and HepG2 cells. Images of the N- and C-terminal fusion proteins revealed that PHYHD1 overexpression was present in both the nucleus and cytoplasm independent of the tag location and the cell type (Fig. 5C). Thus, the nuclear and cytoplasmic distribution of PHYHD1 was confirmed by both approaches.
BioID links proximal proteins of BirA*-PHYHD1 to cell division and RNA metabolism
To shed light on the cellular function of PHYHD1, we performed a BioID screening experiment to map proteins linked to PHYHD1 functionally, via a biological process or a cellular compartment. For that, we used HEK293T cells transiently expressing either BirA* or BirA*-tagged PHYHD1 in the absence or presence of biotin. Western blot analyses confirmed that BirA*-tagged PHYHD1 was localized mainly to the soluble fraction and the addition of biotin allowed BirA* and BirA*-tagged PHYHD1 to biotinylate proteins (Fig. 6A); the levels of biotinylated proteins in the presence of biotin with BirA*-PHYHD1 were significantly enhanced as compared to BirA* (Fig. 6A). To identify the endogenous proteins promiscuously biotinylated by BirA*-PHYHD1, streptavidin-coupled magnetic beads were used to purify biotinylated proteins which were then analyzed by mass spectrometry (MS) (see details in Materials and Methods). The Database for Annotation, Visualization and Integrated Discovery (DAVID) [[20]] revealed the interactome of PHYHD1 to contain an increased number of proteins linked to cell division and RNA metabolism (Fig. 6B). Majority of the BioID hit proteins appeared to be located in the nucleus or cytoplasm, other locations identified being cytoskeleton, chromatin, spliceosome and microtubule (Fig. 6C).
Glucose and maltose consumption rates are significantly downregulated in cells deficient in PHYHD1
To study the role of PHYHD1 in cellular metabolism, we generated PHYHD1 deficient HEK293T cell lines by CRISPR-Cas9 (Fig. 7A). A phenotype microarray containing various carbon and nitrogen substrates was used to identify metabolites that are efficiently catabolized by the scrambled control (CRISPR negative control) and two clones of PHYHD1 knockout cell lines. In general, we observed that HEK293T cells were able to consume only few of analyzed energy substrates (8 out of 91). Of all the substrates tested, PHYHD1 deficient cells clearly had a lower rate of consumption of glucose and maltose than the scrambled cells (Fig. 7B). A reduction of the consumption rates of adenosine and inosine was also detected in PHYHD1 deficient cell lines, but the decrease did not reach statistical significance (data not shown).
Discussion
The analyses carried out here indicate that PHYHD1 is a 2OGDD with some specific characteristics. The 2OG Km value of PHYHD1, 27 μm, is in the higher end among 2OGDDs suggesting a potential for regulation of the catalytic activity when 2OG levels become limiting [[2]]. The Km value of Fe2+ of PHYHD1, 6 μm, represents an average among 2OGDDs [[2]]. The high O2 Km of PHYHD1 suggests that it could act as a novel oxygen sensor having an affinity for oxygen similar to those for HIF-P4Hs and KDM6A [[2, 21]]. In the reaction mechanism of 2OGDDs, binding of the prime substrate follows binding of molecular oxygen, which binds the last, to the active site [[1]]. We cannot therefore exclude the possibility that in the conditions used here, where the decarboxylation of 2OG uncoupled to substrate hydroxylation was measured, the absence of the prime substrate affected the oxygen Km value [[5, 7]]. It has also been demonstrated for other 2OGDDs that, for example, the length of the primary substrate can have an effect on oxygen Km values. For example, for HIF-P4H-2, the Km value for oxygen is 250 μm for the 19-residue peptide representing the C-terminal hydroxylation site of HIF1α where as it is 100 μm for the 248-residue full-length oxygen-dependent degradation domain with both N- and C-terminal hydroxylation sites [[22]].
Interestingly, among the studied 2OG analogues PHYHD1 was not susceptible to inhibition by R-2HG, suggesting that its inhibition does not contribute to the phenotype for the IDH mutant cancers [[2]]. Succinate and fumarate were the most potent inhibitors of PHYHD1. However, their IC50 values, around 1 mm, were not among the lowest for these compounds among the 2OGDDs suggesting that PHYHD1 might not be the primary target of succinate dehydrogenase or fumarate hydratase deficient cancers accumulating succinate and fumarate, respectively, however, its inhibition may contribute [[2]]. Very surprisingly, non-radioactive citrate did not inhibit but increased the decarboxylation of radioactive 2OG by PHYHD1. This would suggest that citrate, a central regulator of lipid metabolism, acted as an allosteric activator of PHYHD1.
The subcellular localization of the endogenous and recombinant PHYHD1 was found to be heterogeneously nuclear and cytoplasmic when studied in several cell lines and to differ from the localization of the closest 2OGDD homolog, PHYH, residing in peroxisomes [[23]]. This detected heterogeneity of the PHYHD1 localization poses a potential additional level of regulation assuming that the localization of the substrate would be stable. Interestingly, the pH optimum for PHYHD1 was acidic which could suggest a function in peroxisomes whereas the pH in nucleus and cytosol are neutral. We tested here as candidate substrates for PHYHD1 a large variety biomolecules of different classes and cellular compartments including nucleic acids, the metabolites of which were recently associated with its genomic variants [[15, 16]]. We report here an interaction of PHYHD1 with RNA, which in the case of mRNA followed saturation kinetics. However, increasing concentration of mRNA or rRNA decreased PHYHD1 catalytic activity suggesting that they acted as competitive inhibitors for the catalysis. While 2OG decarboxylation is often accelerated when coupled to prime substrate oxidation, it is difficult to study experimentally. While unlikely, it could be that PHYHD1 catalysis involves rate-limiting 2OG decarboxylation that does not require prime substrate binding. It is also possible that low level oxidation of the candidate substrate(s) was occurring but it was not detected when using 2OG decarboxylation as a readout. To identify substrates unequivocally, analysis of hydroxylation of the prime substrate itself, for example using mass spectrometry, could be tried in the future. For now, the prime substrate of PHYHD1 remains to be identified; our data here suggest that the substrate binding site is smaller than that for PHYH accepting phytanoyl-CoA.
The pathway analysis of BioID hits of PHYHD1 suggested association with proteins involved in cell division and RNA metabolism, specifically with mRNA splicing. Although future experiments will be required to validate these findings, some of these proteins are, at least partially, in line with data showing strong binding of PHYHD1 to mRNA. Data on the subcellular localization of most of the PHYHD1 BioID hit proteins supported the immunofluorescence microscopy detected heterogeneous localization of PHYHD1 in nucleus and cytosol. In addition, some of the identified proteins displayed localization to cytoskeleton or chromosome. Given the importance of the cytoskeleton and chromosome in the trafficking and anchoring of mRNA [[24]], and mRNA transcription, respectively, it can be speculated that PHYHD1 may also be associated with mRNA transport and mRNA transcription. Furthermore, we showed that PHYHD1 deficient cells use glucose and maltose less efficiently than controls linking PHYHD1 function to carbohydrate metabolism. The molecular mechanisms underlying the observed phenomenon and the exact consequences to cellular functions are to be studied in the future; they may involve alteration in metabolism, sensitivity, and/or transport of glucose and/or maltose. Given that cellular metabolism has a central role in cell growth and survival, this observation may be in an alignment with the BioID data, showing interaction of PHYHD1 with proteins of cell division.
Altogether, these data shed more light on the understanding of the role of PHYHD1 in cellular physiology, whereas its prime substrate and in vivo role remain to be discovered.
Materials and methods
Expression and purification of recombinant PHYHD1
The human full-length PHYHD1A plasmid having a sequence coding for a 6XHis tag at N-terminus was a gift from Nicola Burgess-Brown (Addgene plasmid # 38854). The E. coli BL21 (DE3) pLYS strain was transformed with the plasmid and protein was overexpressed with 0.1 mm isopropyl β-d-1-thiogalactopyranoside induction at 22 °C for 16 h. The cells were lysed by sonication in 50 mm HEPES buffer pH 7.5 containing 300 mm NaCl, 5 mm imidazole and an EDTA free protease inhibitor tablet (Sigma-Aldrich, Espoo, Finland). The cell lysates were centrifuged at 30 000 g for 45 min and the filtered supernatant was incubated with Ni-NTA beads at 4 °C for 1 h. The beads were first washed with the lysis buffer followed by a second wash with a buffer containing 20 mm imidazole before eluting the protein with the buffer with 250 mm imidazole. The eluted fractions were loaded to a Superdex 200 Hiload 16/60, 120 mL column (GE Healthcare, Helsinki, Finland) pre-equilibrated with size SEC buffer (25 mm Tris pH 7.5, 150 mm NaCl and 2 mm DTT). When PHYHD1 was purified in the presence of iron, all buffers were supplemented with 10 μm FeSO4. The SEC fractions were analyzed by denaturing SDS-PAGE and the peak fractions were pooled and concentrated up to 10 mg·mL−1 using a 10 kDa Mw cutoff Centricon. Analysis of the recombinant PHYHD1 by SDS-PAGE using image lab 6.0.1 software (Bio-Rad Laboratories, Hercules, CA, USA) suggested purity (abundance) of the PHYHD1 band to be 79% and of molecular mass of 35.4 kDa. The protein was aliquoted and stored at −70 °C until further use.
Kinetic activity assays
The kinetic assays were performed as previously described for other 2OGDDs with slight modifications [[25]]. In brief, the 50 μL reactions consisted of 50 mm buffer, 2 mg·mL−1 BSA (Roche Diagnostics, Espoo, Finland), 60 μg·mL−1 catalase (Sigma-Adrich, Espoo, Finland), 0.1 mm DTT, 2 mm sodium ascorbate, 10% v/v DMSO and 6–10 μm of the PHYHD1 enzyme. The reactions were stopped by adding 100 μL of 1 m KH2PO4, pH 5. The amount of gaseous 14C-labeled CO2 generated from 2-oxo[1-14C] glutarate (11 100 d.p.m.·nmol−1) (Perkin-Elmer, Langen, Germany) during reaction was captured in a filter paper soaked in Soluene-350 attached to the cap of the tube and scintillated by Tri-carb 2900TR (Perkin-Elmer) whereas the not decarboxylated 2-oxo[1-14C] glutarate remained in the liquid phase. In order to define the optimum conditions for PHYHD1-catalyzed reactions, we first determined the optimum reaction time (varying from 0 to 60 min) and pH using buffers of sodium acetate pH 4.8, sodium acetate pH 5.8, Tris pH 6.8, Tris pH 7.8 and Tris pH 8.8. Thereafter, all enzymatic reactions were carried out at 37 °C for 10 min at pH 5.8 at atmospheric O2 concentration at sea level if not indicated otherwise. To determine the apparent Km values for 2OG, Fe2+ and O2, the enzymatic reactions were carried out by varying the concentration of the component in question while keeping the concentration of others constant. The experiment to determine the Km value for O2 was carried out in six different oxygen concentrations in an InVivo400 hypoxia workstation (Baker Ruskinn, Bridgend, UK). The average apparent Km values of at least three independent assays were calculated from Michaelis–Menten saturation curves and Lineweaver-Burk plots using graphpad prism 9, GraphPad Software, San Diego, California USA, www.graphpad.com
Similarly, the IC50 values of the 2OG analogues [fumarate, succinate, R-2HG, S-2HG and citrate (Sigma-Aldrich)] were determined with increasing the analogue concentration while keeping the concentration of other cosubstrates and cofactors constant. The concentration of 2OG was 64 μm. The IC50 value of each 2OG analogue was calculated from the saturation curves. Average of at least three independent assays was calculated.
The candidate substrate screening analysis was done similarly except having the tested substrates in the reaction. Total RNA from HEK293T was isolated and purified with E.Z.N.A. Total RNA Kit I (Omega Bio-Tek, Norcross, Georgia). mRNA was further purified using a magnetic polyA spin mRNA isolation kit (New England Biolab, Ipswich, MA, USA). Histone peptides (H3K27me3 and H3K4me3), 18S rRNA oligonucleotides rRNA462 (5′-CCUGAGAAACGGCUACCACAUCCAAGGAAGGCAGCAG-3′) and rRNA462mC (5′-CCUGAGAAACGGCUACCACAU[2′O-mC]CAAGGAAGGCAGCAG-3′) and phytanoyl-CoA were purchased from Innovagen AB (Lund, Sweden) and Sigma-Aldrich (Espoo, Helsinki), respectively. Genomic DNA (gDNA) was isolated from mouse tissue by an ethanol precipitation method. The double stranded methylated DNA substrate was synthesized by annealing oligonucleotides containing a 5-mC (5′-CTATACCTCCTCAACTT[5-mC]GATCACCGTCTCCGGCG-3′ and 5′-Biotin-CGCCGGAGACGGTGAT[5-mC]GAAGTTGAGGAGGTATAG-3′ (Sigma-Aldrich).
nanoDSF
Protein thermal stability was analyzed with a Prometheus NT.48 instrument (NanoTemper Technologies, Munich, Germany) using tryptophan residues as an intrinsic probe to report protein unfolding. Fluorescence intensity was measured at two wavelengths (330 and 350 nm with excitation at 280 nm). A fluorescence peak around 330 nm represents a hydrophobic environment of tryptophan, and the fluorescence peak around 350 nm indicates a hydrophilic location. A fluorescence ratio at 350 and 330 nm was then used as a proxy for global changes in the tryptophan fluorescence. Samples were filled into standard capillaries (NanoTemper Technologies) with 10 μL of PHYHD1 protein solution at 0.1 mg·mL−1 and the same amount of PHYHD1 protein mixed with mRNA, rRNA or methylated rRNA, respectively, at 0, 0.781, 1.563, 3.125, 6.25, 12.5 ng·μL−1. The fluorescence excitation intensity at 280 nm was optimized to 100%. The temperature was increased from 20 °C to 80 °C at a ramp rate of 1 °C/min. First derivative of the fluorescence intensity ratio and the Tm values were calculated with the PR.ThermControl (NanoTemper Technologies). The Tm shift (ΔTm) was an average of three repeated experiments. The data were further fitted to an ExpAssoc model and analyzed with graphpad prism 9.
Antibody production and western blot analysis
Generation of rabbit antisera against recombinant purified full-length human PHYHD1 protein and purification of the antibody was carried out at Innovagen AB. HepG2 and murine primary hepatocyte whole cell lysates were prepared as described [[26]]. Western blotting analyses were performed using standard protocols with the PHYHD1 primary antibody in two dilutions: 1 : 10 000 and 1 : 200. The secondary anti-rabbit antibody was conjugated to horseradish peroxidase (1 : 5000; Bio-Rad Laboratories). The Pierce ECL system (ThermoScientific, Helsinki, Finland) was used for detection.
Plasmids and transfection to mammalian cells
Mammalian expression vectors encoding human PHYHD1 with a mCherry-tag either in the N- or C-terminus (PHYHD1 ORF clone ID # 100071452) were generated by the Genome Biology Unit supported by HiLIFE and the Faculty of Medicine, University of Helsinki and Biocenter Finland, and verified by sequencing. HEK293T and HepG cells were maintained in DMEM (Sigma-Aldrich) supplemented with 10% fetal bovine serum (Biowest, Nuaillé, France), 1% nonessential amino acids (Sigma-Aldrich) and 1% antibiotics under standard conditions. Transient transfections were performed using Lipofectamine 3000 (Invitrogen) according to the manufacturer's instructions.
Immunofluorescence microscopy
For immunofluorescence microscopic analysis of the subcellular localization of endogenous PHYHD1, HEK293T and HepG2 cells cultured on a μ-plate 24-well black (Ibidi, Gräfelfing, Germany) were rinsed with phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde for 15 min. After permeabilization with 0.5% Triton X-100 in PBS and blocking with 1% BSA in PBS, samples were incubated for 16 h at 4 °C with either the polyclonal self-made anti-PHYHD1 (1 : 200) or a commercial polyclonal anti-PHYHD1 (1 : 200, HPA012115; Sigma-Aldrich) antibody. Cultures were then incubated for 1 h simultaneously with Cy3-conjugated donkey anti-rabbit IgG diluted 1 : 200 (Jackson ImmunoResearch Laboratories, Europe Ltd, Cambridgeshire, UK) and DAPI (1 : 500, D9542; Sigma-Aldrich). Control procedures included omission of the primary antibodies with inclusion of the secondary antibodies. For analysis of the overexpressed PHYHD1, HEK293T and HepG2 cells were transiently transfected with mCherry-tagged human PHYHD1 and 24 h post-transfection the cells were fixed and stained with DAPI (Sigma-Aldrich, D9542, 1 : 1000). All confocal immunofluorescence microscopic images were obtained by a Zeiss LSM 780 confocal microscope using zen 3.5 image software (Carl Zeiss Microscopy GmbH, Jena, Germany). The acquired images were assembled using coreldraw graphics suite 2020.
BioID, LC-ESI-MS/MS analysis and data searches
BioID pull-downs for MS analysis were performed as previously described [[27]]. A BirA-PHYHD1 fusion plasmid (HA-BirA*-PHYHD1-pDEST-N-pcDNA5/FRT/TO) was generated by the Genome Biology Unit supported by HiLIFE and the Faculty of Medicine, University of Helsinki and Biocenter Finland, and verified by sequencing. HA-BirA*-pDEST-N-pcDNA5/FRT/TO was a kind gift from Maria Vartiainen (Addgene # 118375, [[28]]). In brief, HEK293T cells were transfected with either HA-BirA-pDEST-N-pcDNA5/FRT/TO (backbone) or HA-PHYHD1-BirA*-pDEST-N-pcDNA5/FRT/TO plasmid using Lipofectamine 3000 reagent thus expressing BirA* (~ 37 kDa) or BirA*-tagged PHYHD1 (~ 72 kDa). After 24 h of transfection, 50 μm biotin was added to the media for proximity biotinylation. The cells were collected after 24 h and washed 2× with ice-chilled PBS before lysing them in lysis buffer containing 50 mm tris, pH 7.5, 250 mm NaCl, 0.2% SDS, 1 mm EDTA, 1 mm DTT, 1% NP-40, 1.5 mm MgCl2, 1× complete protease inhibitor cocktail tablet and 250 U of Benzonase by agitating for 1 h at 4 °C. Cells were also sonicated and the lysates were used for western blotting using anti-PHYHD1 primary antibody and streptavidin-HRP antibody. Further, the clear supernatant obtained from centrifugation of 16 000 g 30 min was incubated with magnetic streptavidin (MyDynabeads) overnight at 4 °C. The beads were collected and washed twice with 2% SDS in dH2O, followed by a wash with a buffer (50 mm HEPES pH 7.5, 0.1% deoxycholate, 1% NP-40, 500 mm NaCl, 1 mm EDTA), buffer (10 mm Tris pH 8.0, 0.5% deoxycholate, 1 mm EDTA, 250 mm LiCl, 0.5% NP-40), buffer (50 mm Tris pH 7.5, 50 mm NaCl), and buffer (2 m urea/50 mm Tris HCl pH 7.5). Finally, the beads were washed thrice with 50 mm Tris, pH 8. The MS analyses were performed at the Turku Proteomics Facility supported by Biocenter Finland. In brief, on-bead trypsinolysis was followed by peptide desalting with a Sep-Pak C18 96-well plate (Waters, Vantaa, Finland) and vacuum centrifugation to lyophilize peptides, which were stored at −20 °C. The LC-ESI-MS/MS analyses were performed on a nanoflow HPLC system (Easy-nLC1200; Thermo Fisher Scientific, Bremen, Germany) coupled to the Q Exactive HF mass spectrometer (Thermo Fisher Scientific) equipped with a nano-electrospray ionization source. Peptides were first loaded on a trapping column and subsequently separated inline on a 15 cm C18 column (75 μm × 15 cm, ReproSil-Pur 3 μm 120 Å C18-AQ; Dr. Maisch HPLC GmbH, Ammerbuch-Entringen, Germany). The mobile phase consisted of water with 0.1% formic acid (solvent A) or acetonitrile/water [80 : 20 (v/v) with 0.1% formic acid] (solvent B). A 50 min gradient was used to elute peptides (in 28 min from 5% to 21% solvent B, in 22 min from 21% to 36% solvent B, followed by wash stage with 100% of solvent B). Samples were analyzed by a data independent acquisition (DIA) LC–MS/MS method. MS data was acquired automatically by using thermo xcalibur 4.1 software (Thermo Fisher Scientific). Each duty cycle contained one full scan (400–1000 m/z) and 25 DIA MS/MS scans covering the mass range 400–1000 with variable width isolation windows. Direct DIA data analysis was performed with spectronaut software (Biognosys; version 16.0.220606.53000) and the main parameters used in Spectronaut are listed in Table S1. DirectDIA approach was used to identify proteins and label-free quantifications were performed with MaxLFQ. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [[29]] with the dataset identifier PXD041792.
A possibility for a weak signal suppression at the MS2 scan level was suggested, likely due to technical factors. Therefore, we first normalized the peak intensities of each protein to the peak intensity of PHYHD1 within the same sample and then determined the ratios of the normalized peak intensities between different conditions with or without biotin. We then applied a couple of filters to enrich for PHYHD1 dependent biotinylation. First, a filter was applied to select for proteins identified with ≥ 2 unique peptides. Second, a filter was applied to the normalized peak intensities to select for proteins with ratio ≥ 2. Finally, a filter was applied to select for proteins identified with more proteotypic peptides when exogenous biotin was applied to cells expressing the BirA*-PHYHD1 fusion protein versus BirA* alone. Database for Annotation, Visualization and Integrated Discovery (DAVID) annotation tool [[20]] was used to create GO term table and the pie chart of subcellular localization from the hit list.
CRISPR-Cas9 knockout of PHYHD1 in HEK 293T cells
A 20-bp guide sequence 5′-GCACCAGGAATCCATCCTGT-3′ targeting human PHYHD1 (PHYHD1-201 ENST00000308941.9, PHYHD1-202 ENST00000353176.9 and PHYHD1-203 ENST00000372592.8) was designed online using Zhang's laboratory web resource (www.genome-engineering.org). A non-targeting, scrambled sequence from OriGene was used as a negative control. gRNA-encoding oligonucleotides (Sigma-Aldrich) were cloned into the vector SpCas9(BB)-2A-GFP, (PX458), a kind gift from Feng Zhang (Addgene plasmid ID 48138), using standard procedures as described [[30]]. The generation of the control and knockout cells via CRISPR-Cas9-mediated non-homologous end-joining (NHEJ) DNA repair and the screening was performed according to described guidelines [[30]]. In brief, HEK 293T cells were transiently transfected with Lipofectamine™ 2000 (Thermo Fisher Scientific, Finland) and either with the genome editing or the scrambled CRISPR-Cas9 construct according to the manufacturer's instructions. 48 h post-transfection cells were subjected to single-cell-sorting by using fluorescence-based flow cytometry (BD FACSAria™ III cell sorter). The single-cell clones were expanded and screened for frame-shift mutations on genomic and cDNA level. PCR products performed on cDNAs from the single-cell clones were subsequently cloned into pBlueScript II SK (+) (Stratagene, Germany). 15–20 sequences were analyzed per clone by aligning them with the corresponding reference sequences using BLAST and Serial Cloner. For analyses, two single cell-derived knockout clones referred to as PHYHD1 KO#2 and PHYHD1 KO#4 were used.
Phenotype microarray analysis
A scrambled control and two PHYHD1 knockout clones in HEK293T isogenic cell lines were subjected to phenotype microarray assays in an OmniLog instrument (Biolog, Hayward, CA, USA) following the manufacturer's protocol [[31]]. In brief, the cells were incubated in PM-M1 microplates (#13101; Biolog Inc., Hayward, CA, USA) at 50 μL per well (20 000 cells per well) in an RPMI-1640-based media without phenol red (IF-M1, #72301; Biolog Inc., Hayward, CA, USA) supplemented with 1% penicillin/streptomycin, 0.3 mm glutamate and 5% dialyzed FCS (Sigma-Aldrich) in a standard cell culture incubator. After 24 h, 10 μL of Redox Dye Mix MB (6×) (#74352; Biolog Inc., Hayward, CA, USA) was added to each well and the plate was sealed with sterile adhesive film (SealPlate® film, # Z369667; Sigma-Aldrich) to prevent loss of CO2, and incubated at 37 °C in an OmniLog instrument for 8 h to kinetically record formation of purple formazan in the wells. Data analyses were performed with biolog software data analysis® 1.7 following the manufacturer's protocols.
Acknowledgements
We thank J. Müller and E. Lehtimäki for their excellent technical assistance. We also thank Enrico Tatti and Andre Chouankam from Biolog, Inc. for providing us valuable scientific and technical support regarding phenotype microarray assays. We thank Dr. Hamish Pegg for the enlightening proteomics discussions and Dr. Virpi Glumoff for FACS sorting with CRISPR-Cas9 constructs. The use of the facilities and expertise of the Biocenter Oulu biophysical protein analysis core facility facilitated by Dr. Hongmin Tu, a member of Biocenter Finland, is gratefully acknowledged. The authors declare that they have no conflicts of interest regarding the contents of this paper.
Author contributions
TA-N: Formal analysis, Validation, Investigation, Visualization, Methodology, Writing –review and editing. SKS-T: Data Curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing – original draft, review and editing. V-PR: Methodology. EYD: Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing – original draft, review and editing. PK: Conceptualization, Resources, Supervision, Funding acquisition, Project administration, Writing – original draft, review and editing.
Funding
This work was supported by grants from the Academy of Finland (308009, 339900), the S. Jusélius Foundation, the Jane and Aatos Erkko Foundation and the Finnish Cancer Organizations to PK.
Open Research
Peer review
The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1002/1873-3468.14666.
Data accessibility
All data have been provided in the main article or as Supplementary Material. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [[29]] with the dataset identifier PXD041792.