Pseudouridine, the fifth-most abundant nucleoside in RNA, is not metabolized in mammals, but is excreted intact in urine. The purpose of the present work was to search for an enzyme that would dephosphorylate pseudouridine 5′-phosphate, a potential intermediate in RNA degradation. We show that human erythrocytes contain a pseudouridine-5′-phosphatase displaying a Km ≤ 1 μM for its substrate. The activity of the partially purified enzyme was dependent on Mg2+, and was inhibited by Ca2+ and vanadate, suggesting that it belonged to the ‘haloacid dehalogenase’ family of phosphatases. Its low molecular mass (26 kDa) suggested that this phosphatase could correspond to the protein encoded by the HDHD1 (haloacid dehalogenase-like hydrolase domain-containing 1) gene, present next to the STS (steroid sulfatase) gene on human chromosome Xp22. Purified human recombinant HDHD1 dephosphorylated pseudouridine 5′-phosphate with a kcat of 1.6 s−1, a Km of 0.3 μM and a catalytic efficiency at least 1000-fold higher than that on which it acted on other phosphate esters, including 5′-UMP. The molecular identity of pseudouridine-5′-phosphatase was confirmed by the finding that its activity was negligible (<10% of controls) in extracts of B-cell lymphoblasts or erythrocytes from X-linked ichthyosis patients harbouring a combined deletion of the STS gene (the X-linked ichthyosis gene) and the HDHD1 gene. Furthermore, pseudouridine-5′-phosphatase activity was 1.5-fold higher in erythrocytes from women compared with men, in agreement with the HDHD1 gene undergoing only partial inactivation in females. In conclusion, HDHD1 is a phosphatase specifically involved in dephosphorylation of a modified nucleotide present in RNA.
- haloacid dehalogenase-like hydrolase domain-containing 1 (HDHD1)
- RNA metabolism
Pseudouridine, the most abundant modified nucleoside present in RNA, is formed post-transcriptionally in tRNAs, rRNAs and small nuclear RNAs through isomerization of uridine by pseudouridine synthases . Free pseudouridine, which results from the breakdown of RNAs, is metabolized in Escherichia coli by a specific kinase (YeiC), which phosphorylates the 5′ carbon, and a specific glycosidase (YeiN) that hydrolyses the unique C–C glycosidic bond of 5′-PsiMP (pseudouridine 5′-monophosphate) [2–4]. Homologues of the kinase and the glycosidase are found in many bacteria and eukaryotes. In eukaryotes, the two enzymes belong to a single bifunctional protein. Analysis of genomes indicates that the gene encoding this bifunctional protein is absent from mammalian genomes, accounting for the fact that pseudouridine is not metabolized in man, but is excreted intact in urine. Its urinary excretion is increased in cancerous patients, and serves therefore as a biomarker for cancer [5–7]. The reason for this rise is not well understood, but it is probably due to an increase in RNA turnover. To avoid random incorporation in RNAs, pseudouridine is not reutilized but excreted from cells, as are most other modified nucleosides .
Little is known about the way free pseudouridine is formed. It is likely that breakdown of RNA leads to the formation of both 3′-PsiMP and 5′-PsiMP, which are dephosphorylated before leaving the cells. The availability of purified E. coli pseudouridine-5′-kinase allowed for the easy synthesis of radiolabelled 5′-PsiMP, which we used to check for the presence of a phosphatase acting on this substrate. We found that human erythrocyte extracts indeed contain a rather specific phosphatase for this substrate. The purpose of the present study was to characterize this phosphatase and determine its molecular identity.
Reagents, of analytical grade whenever possible, were from Sigma–Aldrich, Acros, Roche Applied Science, PerkinElmer or Merck. DEAE-Sepharose, Q-Sepharose, HisTrap, HisSpinTrap, Nap-5, PD-10 and Superdex-200 16/60 columns were purchased from GE Healthcare. Pseudouridine and its N1-methyl derivative were from Berry and Associates. Other nucleosides and nucleotides were from Sigma–Aldrich. Vivaspin-15 centrifugal concentrators were from Vivascience. Dowex 1-X8 (100–200 mesh) was purchased from Acros. Enzymes were purchased from Sigma–Aldrich, Roche Applied Science, Finnzymes or Fermentas. B-cell lymphoblast cell lines from patients were obtained from the ECACC (European Collection of Cell Culture) and control B-cell lymphoblast cell lines were kindly provided by Dr Nathalie Demotte, Ludwig Institute for Cancer Research, Brussels.
Preparation of 5′-PsiMP
Unlabelled 5′-PsiMP was synthesized by phosphorylation of pseudouridine with purified YeiC (as described in ). For the synthesis of [32P]5′-PsiMP, 100 μM pseudouridine was incubated for 20 min at 30 °C with 0.5 μg of YeiC in a solution (final volume of 100 μl) containing 25 mM Hepes, pH 7.1, 25 mM KCl, 1 mM MgCl2, 10 μM Mg-ATP and 5 × 107 c.p.m. of [γ-32P]ATP. The incubation was stopped by the addition of 25 μl of ice-cold 10% (w/v) HClO4. After neutralization with K2CO3, the salts were removed by centrifugation (10000 g for 10 min) and the supernatant was diluted with water. The sample (1 ml) was applied on to a 1-ml Q-Sepharose column to purify [32P]5′-PsiMP; this compound was eluted with a NaCl gradient at neutral pH. 3′-UMP was prepared by digestion of polyuridylic acid with pancreatic RNAse A and purified by chromatography on the anion exchanger AG1-X8 using a NaCl gradient at neutral pH.
Measurement of phosphatase activities
Unless otherwise stated, enzymatic assays were performed at 30 °C in a medium containing 25 mM Hepes, pH 7.1, 25 mM KCl, 1 mM MgCl2, 1 mM DTT (dithiothreitol), 0.1 mg/ml BSA, and the indicated concentration of substrate. Assays using unlabelled substrates were performed in a total volume of 50 μl and stopped by adding 100 μl of 10 mM HCl. Inorganic phosphate was then assayed as described by Itaya and Ui . Assays with 32P-labelled 5′-PsiMP were performed in a solution containing, unless otherwise stated, 1 μM 5′-PsiMP (50000 c.p.m.). The reaction was stopped by the addition of 100 μl of ice-cold 5% (w/v) trichloroacetic acid. The samples were centrifuged at 10000 g for 10 min and [32P]Pi was isolated as a phosphomolybdic complex  before radioactivity was counted.
Partial purification of 5′-PsiMPase (pseudouridine-5′monophosphatase) from human erythrocytes
A total of 300 ml of packed erythrocytes were washed three times with 150 mM NaCl and diluted in 1200 ml of lysis buffer (5 mM Hepes, pH 7.1, 1 mM DTT, 1 μg/ml leupeptin and 1 μg/ml antipain). The haemolysate was centrifuged at 11000 g for 20 min. The supernatant (1300 ml) was applied to a DEAE–Sepharose column (200 cm3) equilibrated with buffer A (20 mM Hepes, pH 7.1, 1 mM DTT, 1 μg/ml leupeptin and 1 μg/ml antipain). The column was washed with 400 ml of buffer A and protein was eluted with a 0–0.5 M NaCl gradient in 900 ml of buffer A. Fractions of 7 ml were collected. Fraction 62 (3.5 ml), which contained the peak of 5′-PsiMPase activity (see the Results section), was concentrated to 1 ml and applied on to a Superdex-200 16/60 column equilibrated with buffer B (25 mM Hepes, pH 7.1, 200 mM NaCl, 1 mM DTT, 1 μg/ml leupeptin and 1 μg/ml antipain). Fractions of 1 ml were collected.
Expression and purification of putative phosphatases
For HDHD1 (haloacid dehalogenase-like hydrolase domain-containing 1) a 5′ primer containing the initiator ATG codon (ACCTCGCATATGGCGGCGCCCCCGC) in a NdeI site (in bold) and a 3′ primer containing the stop codon (ACCTCGGGATCCTCACTCATAGGAGGGCAAACCA) flanked by a BamHI site (in bold) were used to PCR-amplify cDNA from human brain with Pwo polymerase. An ~700-bp product was obtained. For PDXP (pyridoxal phosphatase) we used a 5′ primer containing the initiator ATG codon (CCATATGGCGCGCTGCGAGAGGCTG) in an NdeI site (in bold) and a 3′ primer containing the stop codon (AGGATCCTCAGTCCTCCAACCCCTCTGTC) near to a BamHI site (in bold) to PCR-amplify cDNA from human liver with Pwo polymerase. An ~900-bp product was obtained, which was subcloned into pBlueScript. For PHOSPHO2 (phosphatase, orphan 2), a forward primer containing the start codon (ACCTCGCATATGAAAGTTCTGTTGGTGTTTGAC) in a NdeI site (in bold) and a reverse primer with the stop codon (ACCTCGGGATCCTCACATCTTTATTAGAAATTGTAAA) flanked by a BamHI site (in bold) were used to PCR-amplify cDNA from human brain with Pwo polymerase. An ~700-bp product was obtained. For NT5C3L (5′-nucleotidase, cytosolic III-like) a 5′ primer containing the initiator ATG codon (ACCTCGCATATGGCGGAGGAGGTAAGCAGC) in a NdeI site (in bold) and a 3′ primer containing the stop codon (ACCTCGCTCGAGTTAAGAGCCCTGGAGCTCCA) flanked by a XhoI site (in bold) were used to PCR-amplify cDNA from mouse brain with Phusion polymerase. An ~900-bp product was obtained.
PCR products were incubated with the indicated restriction enzymes and the fragments were ligated into pET-15b. The recombinant plasmids were used to transform E. coli BL21pLysS and their inserts were checked by sequencing. The resulting bacteria were grown in LB (Luria–Bertani) medium containing 100 mg/l ampicillin. The cultures were grown at 37 °C until the D600 reached 0.5–0.6. They were then cooled on ice for 20 min and IPTG (isopropyl β-D-thiogalactopyranoside) was added to a final concentration of 0.4 mM. After overnight incubation at 37 °C, the cells were collected by centrifugation (10000 g for 10 min), resuspended in a buffer containing 20 mM Hepes, pH 7.1, 0.5 mM PMSF, 5 mM EDTA, 5 μg/ml leupeptin, 5 μg/ml antipain and 1 mg/ml lysozyme, and submitted to three cycles of freezing and thawing. The bacterial extracts were incubated on ice for 1 h with 0.5 mg/ml DNaseI in the presence of 10 mM MgSO4, and were then centrifuged at 10000 g for 30 min. The resulting supernatants were used to purify the His6-tagged recombinant proteins by chromatography on HisTrap HP columns (1 ml), as previously described .
Preparation of B-cell lymphoblast and red blood cell extracts for activity measurements
B-cell lymphoblasts were cultured in IMDM (Iscove's modified Dulbecco's medium) (Life Technologies) supplemented with 10% (v/v) FBS (fetal bovine serum), 0.24 mM L-asparagine, 0.55 mM L-arginine, 1.5 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin. They were collected by centrifugation of 8-ml cultures at 2000 g for 10 min at 4 °C. Cell pellets were washed twice with ice-cold PBS and resuspended in 300 μl of buffer C (25 mM Hepes, pH 7.1, 0.5 mM PMSF, 1 mM DTT, 5 μg/ml leupeptin and 5 μg/ml antipain). After two cycles of freezing and thawing, extracts were submitted to a 10 min centrifugation (10000 g for 10 min) and the 5′-PsiMPase activity was measured in the supernatants.
Blood was collected in EDTA tubes and centrifuged at 2000 g and 4 °C for 10 min. The plasma was discarded and the buffy coat was isolated and washed twice with 150 mM NaCl. Contaminating erythrocytes were eliminated by two washings with 50 mM NaF, and the leucocytes were then washed twice with 150 mM NaCl and stored at −70 °C before DNA extraction. The erythrocyte pellet was washed twice with 150 mM NaCl and stored at −70 °C before use. One volume of washed erythrocytes was diluted in three volumes of buffer C. After a 10 min centrifugation (10000 g), 5′-PsiMPase activity was assayed in the supernatant.
All experiments with human cells have been performed in accordance with the Declaration of Helsinki (2000) of the World Medical Association and have been approved by the Ethical Committee of the Faculty of Medicine of the University of Louvain. Human blood samples were obtained with informed consent.
DNA extraction and verification of the absence of the HDHD1 gene
Genomic DNA was extracted from frozen leucocytes isolated from 5 ml of EDTA/blood or from B-cell lymphoblasts using the QIAamp DNA Blood Mini Kit (Qiagen). Exons 1 and 3 of the HDHD1 gene were PCR-amplified. Primers (sequences are available from the corresponding author on request) for the amplification were designed in such a way as to avoid amplification of pseudogenes related to HDHD1.
5′-PsiMPase activity in red blood cells
Using [32P]5′-PsiMP as a substrate, we found that human red blood cell extracts catalysed the dephosphorylation of this nucleotide. The dephosphorylation of radiolabelled 5′-PsiMP was inhibited by unlabelled 5′-PsiMP at concentrations in the micromolar range (apparent Ki of ≈1 μM), suggesting that the phosphatase involved in this reaction had a low Km for its substrate. Computation of these results indicated indeed that the phosphatase acting on 5′-PsiMP had a Km of approx. 1 μM and a Vmax of ≈7 nmol/min per mg of haemoglobin. Interestingly, the phosphatase activity measured on [32P]5′-PsiMP was not inhibited by 5′-UMP at a concentration of 300 μM, i.e. 300-fold higher than the Km for 5′-PsiMP. This indicated that 5′-UMP was at best a poor substrate for this enzyme. As 5′-UMP is a close structural analogue of 5′-PsiMP, differing mainly in its accessible portion, through the replacement of a protonated carbon by a protonated nitrogen in structurally equivalent positions on the uracil base, these findings suggested that the phosphatase might be specific for 5′-PsiMP.
5′-PsiMPase was partially purified by chromatography on DEAE–Sepharose and gel filtration on Superdex-200 16/60 columns. The 5′-PsiMPase nearly co-eluted with a peak of 5′-UMPase (uridine-5′-phosphatase) activity in the anion-exchange step (Figure 1A), but was largely separated from it in the subsequent gel filtration step (Figure 1B). Its apparent molecular mass, as determined by comparison with the elution profile of molecular mass markers, was 26 kDa, whereas the enzyme that hydrolysed 5′-UMP had a molecular mass of ≈35 kDa, in agreement with it corresponding to 5′-UMPase (also designated cytosolic 5′-nucleotidase III and encoded by the NT5C3 gene ). Further characterization of the partially purified enzyme confirmed the low Km for 5′-PsiMP of ≈0.8 μM (results not shown) and indicated that the activity was dependent on the presence of Mg2+ and inhibited by Ca2+ and by vanadate (Figure 2).
Vanadate inhibits phosphatases that form a phosphoenzyme during catalysis . Among these, dependency on Mg2+ and inhibition by low concentrations of Ca2+ point to phosphatases of the HAD (haloacid dehalogenase) family . These enzymes, which share structural homology , form a phospho-aspartate residue during their catalytic cycle . They comprise many specific phosphatases acting on different substrates, such as L-serine 3-phosphate , purine and pyrimidine 5′-nucleotides , N-acetylneuraminate 9-phosphate  and glucose 1,6-bisphosphate , as well as eukaryotic phosphomannomutase  and prokaryotic β-phosphoglucomutase . They are characterized by the presence of three conserved motifs, the first one comprising a characteristic N-terminal DXDXT/V motif in which the first aspartate residue serves as a phosphate acceptor during catalysis [14,15]. Owing to the presence of these motifs, they are easily identified by BLAST searches.
5′-PsiMPase corresponds to HDHD1
By performing BLAST searches, we identified four human proteins of the HAD family for which the function was unknown and which had a molecular mass compatible with that of the human erythrocyte 5′-PsiMPase (Table 1). These proteins were expressed in E. coli as fusion proteins with a His6 tag at their N-terminus. They were purified to homogeneity (Figure 3A) and their phosphatase activity was determined with 5′-PsiMP as a substrate (Figure 3B). Only the protein encoded by the HDHD1 gene was found to have a substantial 5′-PsiMPase activity, whereas the other three proteins were either inactive or poorly active on this substrate.
To check the specificity of this enzyme, the activity of HDHD1 was tested on different phosphate esters, which included classical and modified nucleoside monophosphates, as well as several other derivatives (Table 2). The assay was based on the formation of Pi as determined with the sensitive method of Itaya and Ui , and in the case of 5′-PsiMP also through the formation of radiolabelled Pi from [32P]5′-PsiMP. With the first method we found that the phosphatase activity of HDHD1 had already reached its Vmax with the lowest concentration of 5′-PsiMP that we used (20 μM), indicating that the Km of HDHD1 must be at least one order of magnitude lower than this concentration. Accordingly, the Km value determined with the radiochemical method amounted to 0.3 μM, whereas the Vmax was similar to that obtained with the chemical phosphate assay.
The chemical phosphate assay was suitable to determine the kinetic properties for all other tested substrates, because the Km was in all cases higher than the lowest concentration of substrate used (20 μM). Using the catalytic efficiency as a criterion, all tested substrates were poorer by at least three orders of magnitude compared with 5′-PsiMP. Interestingly 3′-AMP was the second best substrate in our series and was much better than 5′-AMP. 3′-UMP was also better than 5′-UMP, although it was not as good as 3′-AMP. It would have been of interest to test 3′-PsiMP. However, this compound is not commercially available to date. We tried to prepare it by digestion of synthetic polypseudouridylic acid (5-mer; Eurogentec), but this polymer was apparently resistant to hydrolysis by pancreatic RNAse A (results not shown). None of the modified nucleoside monophosphates that we tested were a good substrate for HDHD1. As mentioned below HDHD1 is related to plant FMNPase (flavin mononucleotide phosphatase). The catalytic efficiency of human HDHD1 on this substrate was, however, more than 30000-fold lower than that observed with 5′-PsiMP.
Eukaryotic rRNA comprises N1-methylpseudouridine, a precursor for the hypermodified nucleoside 1-methyl-3-(3-amino-3-carboxypropyl)-pseudouridine . Our attempts to prepare N1-methylpseudouridine 5′-phosphate from commercial N1-methylpseudouridine failed, because this compound was not phosphorylated by E. coli pseudouridine kinase (results not shown). We found, however, that, whereas pseudouridine inhibited the HDHD1 activity measured either with 5′-PsiMP or 3′-AMP (Figures 4A and 4B), 1-methylpseudouridine barely affected this activity, suggesting that the methyl group on N1 prevents recognition of pseudouridine by the catalytic site of HDHD1. No inhibition was observed with uridine and adenosine, which were used as controls.
5′-PsiMPase activity in X-linked ichthyosis and according to gender
The HDHD1 gene is present on chromosome X next to the STS (steroid sulfatase or arysulfatace C) gene  which encodes the enzyme that is deficient in X-linked ichthyosis, a disorder that is usually caused by a submicroscopic deletion in the X chromosome encompassing both the STS gene and the HDHD1 gene . We obtained B-cell lymphoblasts or erythrocytes from four patients with X-linked ichthyosis and verified that they had a deletion of the HDHD1 gene by showing that no amplification product could be obtained when their DNA was submitted to PCR reactions with appropriate primers. As shown in Figure 5, B-cell lymphoblasts from three patients and erythrocytes from a fourth one were devoid of the low Km 5′-PsiMPase activity, confirming that HDHD1 indeed encodes this enzyme.
The STS and HDHD1 genes are present in a region of chromosome X that partially escapes X-inactivation, as confirmed by activity measurements of the STS activity in female subjects compared with male subjects [23,24] and, for the HDHD1 gene, by analysis of its mRNA level with microarrays . As shown in Figure 5(B), the mean pseudouridine phosphatase activity of erythrocyte extracts was approx. 50% higher in female subjects compared with male subjects.
Sequence comparison and distribution of HDHD1
Homologues of HDHD1 are widely distributed. Proteins sharing more than 40% identity are indeed found in virtually all eukaryotes, including vertebrates, invertebrates, plants and fungi. The HDHD1 gene is encoded by a multi-exon gene present next to the STS gene on chromosome X in man and several other mammals (cow, horse and dog), but is on an autosome in birds and fishes. The mouse and rat homologues are encoded by a single-exon gene on chromosome 18 in both cases. This gene presumably resulted from a retrotransposition event. Arabidopsis thaliana and other plants have two homologues of HDHD1. One of them, more distant (34% sequence identity with the human enzyme; Figure 6), forms a fusion protein with riboflavin kinase and was shown to act as an FMNPase. We speculate that the other protein, which shares 50% sequence identity with HDHD1 (Figure 6), catalyses the hydrolysis of 5′-PsiMP. As mentioned above, human HDHD1 has an extremely low FMNPase activity.
The alignment displayed in Figure 6 shows that the HDHD1 proteins possess the three motifs known to be conserved in phosphatases of the HAD family, although the usual valine or threonine residue at the end of the DXDXT/V motif is replaced by a leucine residue. These motifs comprise the two aspartate residues of the first motif, the threonine (or serine) residue of the second motif and a lysine and an aspartate residue in the third motif. These residues are shared with A. thaliana FMNPase.
Identification of 5′-PsiMPase as a product of HDHD1
In the present study we show that human erythrocytes contain a phosphatase that acts with very high affinity on the modified nucleotide 5′-PsiMP and much less on classical nucleoside monophosphates and other phosphate esters. We identified this enzyme as being the product of the HDHD1 gene (also known as HDHD1A or GS1), a putative phosphatase of the HAD family for which the function has long remained unknown. This identification is based on the findings that: (i) human 5′-PsiMPase has kinetic properties typical of phosphatases of the HAD family  (inhibition by vanadate, Mg2+-dependency and inhibition by Ca2+, which replaces Mg2+ in the catalytic site and perturbs the orientation of the nucleophilic carboxylate ); (ii) gel filtration shows it to have a size in agreement with that of HDHD1; (iii) recombinant HDHD1 catalyses the hydrolysis of 5′-PsiMP with a Km of 0.3 μM, similar to the 1 μM value observed with the enzyme purified from erythrocytes, and shows a much lower affinity for 5′-UMP than for 5′-PsiMP; (iv) 5′-PsiMPase activity is absent from human cells obtained from male X-linked ichthyosis patients, in which the HDHD1 gene is deleted due to partial deletion of the X-chromosome; and (iv) 5′-PsiMPase activity is higher in erythrocytes of women than of men, consistent with the HDHD1 gene being, like the STS gene, more highly expressed in cells from women than from men, due to incomplete inactivation of the second copy .
Specificity of HDHD1
The HAD family is the most important family of phosphatases acting on low-molecular-mass compounds. Some of the enzymes from this family are specific, in the sense that they act on one substrate with a catalytic efficiency that is at least three or four orders of magnitude higher than on other substrates (see [27, 28] for phosphoethanolamine/phosphocholine phosphatase; see  for N-acetylneuraminate phosphatase).
The study of the substrate specificity of HDHD1 indicates that this enzyme dephosphorylates a series of phosphate esters, but acts on 5′-PsiMP with a much (>1000-fold) higher catalytic efficiency than on all other phosphate esters that we could test. It is 105-fold less active on 5′-UMP, which is structurally very close to 5′-PsiMP, given it essentially differs through the presence of a protonated nitrogen (N1 in pseudouridine) instead of a protonated carbon (C5 in uridine). It is likely that specific recognition of 5′-PsiMP involves hydrogen bonding with this N1-proton. This is further indicated by the conclusion that N1-methyl-pseudouridine 5′-phosphate is probably not a good substrate for the enzyme, as deduced from the finding that N1-methyl-pseudouridine is a much poorer inhibitor of HDHD1 activity than non-methylated pseudouridine. The catalytic efficiency observed when HDHD1 acts on 5′-PsiMP (4.2×106 s−1·M−1) favourably compares with the values observed with other specific phosphatases of the HAD family acting on their physiological substrate: 0.75 × 106 s−1·M−1 for phosphoethanolamine phosphatase ; 0.62 × 106 s−1·M−1 for N-acetylneuraminate phosphatase ; 0.68 × 106 s−1·M−1 for E. coli phosphoserine phosphatase ; and 3 × 106 s−1·M−1 for glucose-1,6-bisphosphatase . Taken together, these findings suggest that HDHD1 may be a ‘specific’ 5′-PsiMPase.
Remarkably, the second best substrate that we found for HDHD1 is 3′-AMP, which, although being 1300-fold less active than 5′-PsiMP (based on the catalytic efficiency), is approx. 600-fold better than 5′-AMP. The finding that the activity on 3′-AMP was inhibited much more by pseudouridine than by other nucleosides confirmed that this activity was contributed by HDHD1 and not by any contaminating phosphatase originating from E. coli in our purified preparation of HDHD1. 3′-UMP was also found to be a better substrate than 5′-UMP, although in this case, the ratio of the two catalytic efficiencies amounted to ≈3. We interpret this as being due to the fact that non-optimal substrates do not bind to the catalytic site in the exact same manner as the best substrate(s). Therefore the presence of a bulkier base (as in 3′- and 5′-AMP) may mean that a 3′-phosphate is better positioned for hydrolysis than a 5′-phosphate, explaining that 3′-AMP is a better substrate than 5′-AMP (yet both of are poor substrates compared with 5′-PsiMP). It would have been interesting to test 3′-PsiMP, but this could not be done due to the lack of availability of this substrate. Our expectation is that 3′-PsiMP would be a poorer substrate than 5′-PsiMP; correct positioning of the uracil moiety of pseudouridine in the catalytic site of HDHD1 presumably means that only a 5′-phosphate can be in the best position to be attacked by the nucleophilic aspartate residue.
Role in RNA degradation
The role of HDHD1 is most probably to degrade 5′-PsiMP and maybe also its 2′-O-methyl derivative, which is also present in rRNA . Experiments in cell-free systems indicate that 5′-PsiMP is phosphorylated further to pseudouridine di- and tri-phosphate . This conversion probably involves uridine/cytidine kinases, which have a broad substrate specificity  and nucleotide diphosphate kinases, which act on a variety of nucleotide diphosphates . The accumulated pseudouridine 5′-triphosphate can be utilized instead of UTP by various enzymes, such as RNA polymerases and UDP-glucose pyrophosphorylase, as suggested by Goldberg and Rabinowitz [34,35]. The presence of an enzyme that acts on 5′-PsiMP with high affinity is presumably an efficient means to prevent aberrant incorporation of pseudouridine into RNA or nucleotide diphosphate sugars.
The role of many phosphatases of the HAD family remains unknown. Kuznetsova et al.  studied the catalytic activity of a series of phosphatases of the HAD family encoded by the E. coli genome and found that most of these enzymes acted with a low catalytic efficiency on a series of phosphate esters. We speculate that some of them may act with a much higher catalytic efficiency on 5′-PsiMP or on other modified nucleoside 5′- or 3′-monophosphate resulting from the degradation of RNAs. This dephosphorylation is presumably a prerequisite for the excretion of the modified nucleosides from cells.
Potential role of HDHD1 deficiency in X-linked ichthyosis
The HDHD1 gene was first identified (as the GS1 gene) because of its close proximity to the STS gene on human chromosome X . Inactivating mutations or, more often, deletion of the latter gene is responsible for X-linked ichthyosis, which is characterized by scaly skin on the scalp, trunk and limbs, often accompanied by corneal opacities, delayed birth, difficult labour, undescended testis and a higher frequency of testis tumours . Most of these pathological findings are the consequence of the defect in the sulfatase activity and the consequent changes in the concentration of oestriol, oestrol-sulfate and cholesterol sulfate. To the best of our knowledge, cryptorchidism and testis tumours have not been reported in cases with specific mutations (point mutations, intragenic deletions) of the STS gene or in multiple sulfatase deficiency, a disorder due to a deficiency in the enzyme that converts cysteine into formylglycine in the catalytic site of sulfatases and leads to inactivity of all sulfatases, including steroid sulfatase. We may therefore not exclude that the absence of 5′-PsiMPase activity plays a role in the pathophysiology of cryptorchidism or in the development of testis cancer. Although we have no simple hypothesis to explain this potential link, it would be interesting to examine more systematically if 5′-PsiMPase activity is absent in X-linked ichthyosis patients with cryptorchidism or testis cancer.
Alice Preumont and Emile Van Schaftingen conceived the study. Alice Preumont, Rim Rzem and Didier Vertommen collected the experimental data under the supervision of Emile Van Schaftingen. Alice Preumont, Rim Rzem and Emile Van Schaftingen wrote the article.
This work was supported by the Directorate General Higher Education and Scientific Research, French Community of Belgium; the Fund for Medical Scientific Research; and the Interuniversity Attraction Poles Programme, Belgian Science Policy (Network P6/05). A.P. is a Fonds de la Recherche Scientifique-Télévie fellow.
We thank J. Fortpied, M.-F. Vincent, S. Marie, N. Demotte and P. van der Bruggen for their helpful suggestions. We also thank Professor Tennstedt and one of his patients for providing us with red blood cells.
Abbreviations: DTT, dithiothreitol; FMNPase, flavin mononucleotide phosphatase; HAD, haloacid dehalogenase; HDHD1, HAD-like hydrolase domain-containing 1; NT5C3L, 5′-nucleotidase, cytosolic III-like; PDXP, pyridoxal phosphatase; PHOSPHO2, phosphatase orphan 2; 5′-PsiMP, pseudouridine 5′-monophosphate; 5′-PsiMPase, pseudouridine-5′-monophosphatase; STS, steroid sulfatase; 5′-UMPase, uridine-5′-phosphatase
- © The Authors Journal compilation © 2010 Biochemical Society