| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,1
,2
* Department of Biochemistry and
Biocenter Oulu, University of Oulu, Oulu, Finland; and
Department of Biochemistry and Molecular Biophysics and Department of Molecular and Cellular Biology, University of Arizona, Tucson, Arizona, USA
2Correspondence: Department of Biochemistry and Biocenter Oulu, University of Oulu, P.O. Box 3000, FIN-90014 Oulu, Finland. E-mail: kalervo.hiltunen{at}oulu.fi
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: HsHTD2 • RNase P
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. These transcripts can be divided into two different types according to their further processing. The first type is characterized by an initial polycistronic transcript that is modified by 3' end cleavage and trans-splicing to form a monocistronic mRNA (2)
. This kind of transcript modification is used by the nematode Caenorhabditis elegans (3)
, flatworms (4)
, and primitive chordates (5)
. The other type is characterized by a bicistronic transcription unit that specifies an mRNA encoding two different proteins and is translated in that form after transport to the cytosol. In mammals, the first bicistronic gene cluster was described by Lee (6)
. In several documented cases, the bicistronic transcripts encode proteins with related function. Many of these genes produce an alternative shorter version of the mRNA containing only the 3' open reading frame (ORF), which makes an internal AUG codon a functional translational start site. However, at least three truly bicistronic transcripts have been characterized in humans to date. They encode 1) two subunits of molybdopterin synthase (7)
; 2) growth and differentiation factor 1 (GDF-1) and a trans-membrane protein of unknown function (UOG-1; ref. 6
); and 3) SmN spliceosomal protein together with a protein of unknown function encoded by the SNURF gene (8)
. However, in higher eukaryotes, it is under debate whether transcripts harboring two ORFs represent true bicistronic translation templates or whether they are only structurally bicistronic but functionally monocistronic (1)
.
Ribonuclease P (RNase P) is an essential endonuclease that acts early in the tRNA biogenesis pathway. It catalyzes cleavage of the leader sequence of precursor tRNAs, generating the mature 5' end of tRNAs. RNase P activities have been identified in bacteria and archaea as well as eucarya and their organelles. All characterized RNase P enzymes are ribonucleoproteins and consist of an essential RNA subunit and one or more protein subunits. In humans, the nuclear RNase P complex is composed of the H1 RNA subunit and 10 protein subunits: RPP14, RPP20, RPP21, RPP25, RPP29, RPP30, RPP38, RPP40, hPOP5, and hPOP1 (9)
. Yeast nuclear RNase P shows a similar but not identical composition, as homologs for several human subunits are not present (10)
. Yeast mitochondria have their own RNase P complex for mitochondrial tRNA maturation, which consists of a mitochondrially encoded RNA subunit and a nuclear encoded protein subunit. The human mitochondrial RNase P has not been purified to homogeneity, and its nature is subject to debate. While one group claimed that this enzyme is lacking an RNA subunit (11)
, another group reported the presence of an RNA identical to the RNA present in the nuclear RNase P complex (12)
. There is no mitochondrially encoded RNase P RNA in mammals.
Among the recently recognized aspects of mitochondrial functions, in yeast as well as humans, is the ability to synthesize fatty acids in a malonyl-CoA dependent manner. This mitochondrial fatty acid synthesis (FAS) pathway follows the bacterial type II fatty acid synthesis (FASII) mode, with separate polypeptides carrying out the individual reactions. In contrast, eukaryotic cytosolic type I FAS is performed by multifunctional proteins. The mitochondrial pathway is best described in the yeast Saccharomyces cerevisiae (for review see Hiltunen et al., ref. 13
), where deletion of any member of the mitochondrial FAS pathway leads to a respiratory deficient phenotype and lack of cytochromes, indicating that this pathway is essential for mitochondrial function. The role and function of the mitochondrial FAS pathway are not yet fully understood but have been suggested to provide octanoyl-ACP as a substrate for lipoic acid synthesis (14)
. To date, acyl carrier protein, malonyl transferase (15)
, β-ketoacyl synthase (16)
, and enoyl reductase (17)
of the human mitochondrial FAS II pathway have been characterized, but obvious candidates for 3-hydroxyacyl-thioester dehydratase (18)
have evaded identification. Here, we report the isolation of human cDNAs encoding both RPP14 of the RNase P complex and mitochondrial 3-hydroxyacyl thioester dehydratase. The cDNAs encode a bicistronic transcript that has been evolutionarily conserved.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Yeast was grown on either rich YPD (1% yeast extract, 2% peptone, and 2% D-glucose), YPG (3% glycerol), or synthetic complete (SC) media [Sigma-Aldrich, Helsinki, Finland; or Bio101/QBiogene, Carlsbad, CA, USA (SCD 2% D-glucose, SCG 3% glycerol)] or SC media lacking one or more nutrients. Yeast cells were transformed as described previously (19)
.
|
The screen for human mitochondrial dehydratase cDNAs
We previously reported the generation of a respiratory-deficient mutant yeast strain (W15368B htd2-1) harboring a defective mitochondrial 3-hydroxyacyl thioester dehydratase (18)
. This strain was transformed with human cerebellum and/or kidney cDNA libraries in the pMETtPGK3-1/2/3 vector (20)
using the Liac/SS carrier DNA/PEG method (21)
. After transformation, the cells were plated on SC-Met–Ura– plates with 3% glycerol and grown for 6–7 days at 30°C, and candidate clones were selected for their ability to rescue growth of the respiratory deficient phenotype of the htd2-1 mutant on glycerol as a sole carbon source. Large-scale transformation with cerebellum cDNA, representing 
1.8 x 106 transformants, yielded six colonies after 7 days of growth on glycerol. Transformation with the kidney library (2.8x106 transformants) yielded seven colonies after 6 days of growth. To confirm that complementation was due to a plasmid-borne factor, these transformants were then plated on media containing 5-fluoro-orotic acid (5'-FOA) to select for loss of the plasmids and were subsequently tested for loss of respiratory competence. Three colonies of six cerebellum library isolates and all seven of the kidney isolates passed the 5'FOA test. Transformants that were not able to grow on glycerol after 5'FOA treatment were selected for further tests. The plasmid DNA was isolated from the corresponding original transformant yeast clones and introduced into E. coli TOP10 cells by transformation. Plasmid DNA was isolated using a NucleoSpin plasmid kit (Macherey-Nagel, Düren, Germany) and transformed back into yeast strains htd2-1 and 
htd2 to confirm the ability to complement the mutations on glycerol. After isolation of two clones (A and B) from each transformant and retransformation of the yeast strains, only four (1A, 1B, 5A, and 5B) were able to complement both yeast mutations. Correspondingly, plasmid isolates 3A, 3B, and 4B from the kidney library were able to complement both mutations. Isolated plasmid-DNAs were digested with BstXI and NotI restriction enzymes to release the cDNA inserts and analyzed on an 0.8% agarose gel; 1A, 1B, 5A, 5B, and 4B had an insert with an approximate size of 1.5 kb, while isolates 3A and 3B displayed a different restriction pattern, with two fragments of 1 kb and 250 bp. The isolated plasmids were sequenced by using seqp1.MET3 (5'-CTCTCTGTCGTAACAGTTGT-3') and seqp2.PGK1 (5'-GGCAATTCCTTACCTTCCAA-3') primers. The plasmids were later sequenced completely by using primers seqp3.MET3 (5'-CTGTTAGGATCCTATAAAGGC-3') and seqp4.PGK1 (5'-TGCTCTTAACCCAAAGAATG-3').
Yeast complementation
To test whether the 3' ORF of RPP14 mRNA was sufficient to complement the yeast dehydratase knockout or point mutations, the corresponding human 3' ORF was amplified by polymerase chain reaction (PCR) from the cerebellum cDNA library using the primers Hshtd2yEP352f (5'-GCTCTAGAATGTTCCCACTAATTTCCAGCCAT-3') and Hshtd2yEP352r (5'-CCGCTCGAGTCAGGATTTGGAAGCTTCTGG-3') introducing an XbaI restriction site 5' of the initiation ATG and an XhoI restriction site 3' of the stop codon, respectively. The digested PCR product was ligated into the S. cerevisiae expression vector pYE352-CTA1 (22)
, replacing the CTA1 ORF and resulting in plasmid pYE352-Hshtd2. Plasmids pYE352-Hshtd2, pYE352-htd2, and pYE352-Cta1 were transformed into yeast strains htd2-1 and htd2 on SC-Ura– with 2% glucose. The transformants were moved to synthetic complete plates with 3% glycerol and grown at 30°C for 6–7 days.
Expression and purification of recombinant HsHTD2 and hydratase 2 activity assay
For hydratase 2 activity assays, HsHTD2 was produced as a recombinant protein fused to the C terminus of the maltose binding protein without the predicted mitochondrial targeting signal (mature HsHTD2). The DNA sequence encoding mature HsHTD2 was amplified by PCR from the library clone, using primers matHsHTD2-MBP5'AseI (5'-CAGACATTAATACAGTCTGTCTGAACCTGCCA-3') and HsHTD2-MBP3'BamHI (5'-GACGGATCCTCAGGATTTGGAAGCTTCTGG-3'). The product was purified, cut with AseI and BamHI, and ligated into NheI/BamHI digested plasmid pMAL-c2x-pET23a, a gift from Lloyd Ruddock. The recombinant protein was expressed in E. coli BL21(DE3)pLysS strain containing a RARE plasmid. LB medium supplemented with 50 µg/ml carbenicillin and 34 µg/ml chloramphenicol was used for expression experiments. A 10 ml portion of overnight culture of E. coli cells containing pMAL-c2x-pET23a- matHsHTD2 plasmid was used to inoculate 1 L of culture. The cells were allowed to grow at 37°C under aerobic conditions until an OD600 of 0.4 was reached. The expression of the recombinant protein was induced by the addition of isopropyl-1-thio-β-D-galactopyranoside (IPTG) to a final concentration of 0.1 mM. The incubation continued at 18°C for 20 h, and the cells were harvested, washed with PBS, and stored at –70°C until used.
A bacterial cell pellet (3.6 g wet weight) was suspended in 36 ml of 20 mM Tris, 200 mM NaCl, 1 mM EDTA, 1 mM DTT, 1 mM benzamidine hydrochloride hydrate (BA), and 0.1 mM phenylmethylsulfonyl fluoride (PMSF) pH 7.4, and the cells were broken by sonication. The mixture was centrifuged at 9000 g for 30 min at 4°C, and the supernatant was applied to a column containing amylose resin (6 ml, New England Biolabs, Beverly, MA, USA) equilibrated with column buffer (20 mM Tris, 200 mM NaCl, 1 mM EDTA pH 7.4). The column was washed with column buffer, and the bound protein was eluted with column buffer containing 10 mM maltose. The pooled fractions of MBP+Hshtd2p were concentrated and applied to a size exclusion chromatography column Superdex 200 HR 10/30 (Amersham Pharmacia Biotech AB, Uppsala, Sweden) equilibrated with 30 mM HEPES, 200 mM NaCl, and 1 mM sodium azide (NaN3) pH 7.4. Proteins were eluted at a flow rate of 500 µl/min, and the protein purification was monitored by SDS-PAGE. Hydratase 2 activity was measured in the purified sample as complex formation of ketoacyl-CoA with Mg2+ (23)
using trans-2-hexenoyl-CoA and trans-2-decenoyl-CoA as substrates and recombinant (3R)-hydroxyacyl-CoA dehydrogenase (consisting of both domains A and B) from Candida tropicalis as an auxiliary enzyme (24)
.
Mutagenesis of active site residues
Plasmid pYE352-Hshtd2 was used as template in PCR amplification for constructing mutations D62A and H67A in the HsHTD2 gene. The H67A mutation was created by using the QuikChange multisite-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) with primer Hshtd2H67Af (5'-GTTAATCCTTTG[b]GCTTTGAATGAAGAC-3'). The H67A mutant plasmid was used as template for the D62AH67A double mutation, which was also made with the QuikChange site-directed mutagenesis kit (Stratagene) using primers Hshtd2D62Af (5'-CCTTCTCAGAATTAACAGGGGCTGTTAATCCTTTG-3') and Hshtd2D62Ar (5'-CAAAGGATTAACAGCCCCTGTTAATTCTGAGAAGG-3'). In a similar way, the D62A mutant was created by using the original pYE352-Hshtd2 plasmid as template.
Localization studies
For localization studies, RPP14 and HsHTD2 with and without a mitochondrial signal sequence were amplified by PCR using DyNAzyme EXT polymerase (Finnzymes, Espoo, Finland). For HsHTD2, the primers were Hshtd2GFPmitf (with a mitochondrial targeting signal, 5'-CCGCTCGAGATGTTCCCACTAATTTCC-3') or Hshtd2GFPf (without a mitochondrial targeting signal, 5'-CCGCTCGAGATGACAGTCTGTCTGAACCTG-3') and Hshtd2GFPr (5'-TAACCGCGGGGATTTGGAAGCTTCTGG-3') with XhoI and SacII restriction sites at the 5' and 3' ends of the cDNA sequence, respectively. The resulting PCR product was digested with XhoI and SacII, gel-purified, and subcloned inframe with the 5' end of the green fluorescent protein (GFP) cDNA in the expression vector pEGFP-N1 (Clontech, Mountain View, CA, USA), resulting in pEGFP-N1-Hshtd2mit and pEGFP-N1-Hshtd2. The pEGFP-N1-Rpp14 construct was cloned exactly in the same way using primers HsRpp14GFPf and HsRpp14GFPr. HeLa cell transfection and fluorescence microscopy were done as described previously(17)
.
Northern blot analysis
For Northern blot analysis, a 32P-labeled probe was synthesized by the random priming method using [
-32P]dCTP (3000 Ci/mmol), PCR amplified RPP14 or HsHTD2 cDNA, and the Random Primed DNA labeling kit (Roche Diagnostics, Espoo, Finland). A human Multiple Tissue Northern blot (BD Biosciences Clontech, Palo Alto, CA, USA) containing poly(A+) RNA from various human tissues was hybridized with purified radiolabeled probe. The hybridization was done in ExpressHyb (BD Biosciences Clontech) solution according to manufacturer’s recommendations. The blot was analyzed using a phosphorimager (Molecular Imager FX, Bio-Rad Laboratories, Richmond, CA, USA).
Bioinformatics and construction of phylogenetic tree
The BLAST and PSI-BLAST tools (25)
were used in DNA homology searches and in discovering identities to the coding region of human mitochondrial dehydratase in the Genbank and TrEMBL/SwissProt databases. Additional vertebrate homologs were retrieved by BLASTX searches of the Genbank redundant and EST databases followed by identification of ORFs using the Translate tool at ExPASy. The authenticity of the identified mitochondrial dehydratase candidates was further established by reciprocal homology searches using every retrieved dehydratase sequence as the query.
A multiple alignment was then constructed based on the Pfam MaoC dehydratase family (accession number Pf01575). A representive set of sequences chosen from the complete family alignment was biased toward sequences with characterized functions or structures. The sequences of mitochondrial dehydratases homologous to the human sequences as well as the previously identified fungal mitochondrial dehydratases (18)
were aligned to the MaoC dehydratase using the ClustalX 1.8 program (26)
with default settings. Secondary structure information from the Protein Structure Database was annotated, and the resulting alignment was manually inspected. The final alignment included 60 sequences and 107 positions around the conserved fingerprint of the MaoC dehydratase family. A subset of 31 sequences and 105 positions was used to construct phylogenetic trees. Distance, parsimony, and maximum likelihood analysis including 100 bootstrap replicates, respectively, was performed using the programs PROTDIST, NEIGHBOR, FITSCH, PROTPARS, PROML SEQBOOT, and CONSENSE of the PHYLIP package (27)
.
Information on transcripts, genomic sequences, and exon-intron organization of chordate mitochondrial dehydratases was retrieved from the Ensembl database at the EBI and the Genbank and UniGene databases at the National Center for Biotechnology Information (NCBI). In cases where a RPP14/HTD2 transcript could not be identified unambiguously by keyword searches, sequences were identified by BLAST searches. Alternative transcripts were sought by mapping of transcripts and genomic sequences to the NCBI EST database, as well as the construction of contigs from retrieved ESTs.
Expression of 3' HsHtd2 in vivo
To test if the bicistronic transcript without a promoter can complement the yeast dehydratase knockout or point mutations, the entire mRNA was amplified by PCR from the cerebellum cDNA library plasmid 1A using the primers HsRpp14RFPf (5'-CTAGCTAGCGGTGTGATCAGCACTGGA-'3) and seqp2.PGK1, which introduced an NheI restriction site 5' of the initiation ATG. An XhoI restriction site at the 3' prime end of the insert was used. The digested PCR product was ligated into the S. cerevisiae expression vector YCplac33, resulting in plasmid HsRpp14cDNA1A+YCplac33. Plasmid HsRpp14cDNA1A+YCplac33, the original human cerebellum cDNA library plasmid 1A, and the YCplac33 vector were transformed into yeast strains htd2-1 and htd2 and plated on SC-Ura– with 2% glucose. The transformants were transferred to synthetic complete plates with 2% glucose or 3% glycerol and grown at a temperature above 30°C for 7 days.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
yielded four individual cDNA clones (see Materials and Methods) encoding the human RPP14 gene. Sequence analysis revealed that three clones (plasmids 1A, 1B, 5A, 5B, and 4B) contained the complete RPP14 transcript (28)
, whereas one clone (plasmid isolates 3A and 3B) contained only the 3' end of the same transcript lacking the upstream RPP14 coding region. The coding sequence of human RPP14 is located on chromosome 3 (3p14.3), and the gene size is over 30 kb with 6 exons.
Figure 2
shows the genomic structure of RPP14. The inserts of 1A, 1B, 4B, 5A, and 5B covered all six exons, whereas clones 3A and 3B contained only exon 6 with a short internal deletion and an additional Alu-repeat that is located downstream of the gene.
|
|
RPP14 mRNA contains two ORFs
We noticed that exon 6 of the RPP14 mRNA contained an additional large ORF. Since only this second 3' ORF was present in all the isolated clones, we reasoned that the RPP14 mRNA might be bicistronic, encoding two proteins on the same transcript, and that the second 3' ORF may encode the protein that complemented the mitochondrial dehydratase deficiency in our yeast screen. The novel ORF was 507 bp long, encoding a protein of 168 amino acids. This protein was predicted to belong to the MaoC dehydratase family based on an NCBI-conserved domain search (25)
and displayed homology to prokaryotic acyl dehydratases, yeast peroxisomal Fox2p, and mammalian MFE2 and FAS1. This protein also contains the "hydratase-2" fingerprint (29)
. The MITOPROT II mitochondrial prediction program identified the first 17 N-terminal amino acids of the dehydratase candidate as a potential mitochondrial targeting signal, and the protein was predicted to be localized in mitochondria with 67% probability (30)
. The predicted molecular mass for the protein is 18.5 kDa, which is similar to bacterial dehydratases (31)
and about one-half of the size of S. cerevisiae Htd2p. Taken together, the evidence suggested that we had identified an excellent candidate for human 3-hydroxyl-ACP dehydratase in the mitochondrial FAS pathway.
To confirm that the dehydratase activity is encoded by the 3' ORF of the biscistronic transcript, only this part of the RPP14 mRNA was cloned into the pYE352 vector. The ability of this construct to complement both the yeast dehydratase knockout and htd2-1 mutations was tested on glycerol plates. As seen in Fig. 1
, the results clearly show that the dehydratase activity in this bicistronic transcript is encoded only by this second ORF. In contrast, the 5' ORF coding for the RPP14 protein, without the dehydratase region, was unable to complement the mutations (results not shown). Based on these results, the protein encoded by the 3' ORF of the RPP14 mRNA was named human mitochondrial 3-hydroxyacyl-thioester dehydratase 2 or HsHTD2.
Linkage of RPP14 and HTD2 is highly conserved in vertebrates
When we examined the relationship of RPP14 and HTD2 in other eukaryotes, we found a striking conservation of the bicistronic arrangement throughout evolution (Fig. 2
). The linkage between RPP14 and HTD2 on a common transcript has remained unchanged from fishes and frogs to birds and humans, indicating that the organization of RPP14 and HTD2 as two successive ORFs encoded by the same mRNA appears to be conserved in vertebrates. The exon-intron organization also seems to be largely the same within this group. The ORF coding for RPP14 is interrupted by several introns, whereas the ORF for HTD2 is always contained within the last exon. In some fish, such as zebrafish or green puffer fish, there is evidence for two alternatively spliced forms of the RPP14 transcript, one of which is shortened due to the removal of an additional intron within the last exon sequence. The removal of this short intron results in the inframe fusion of the ORFs for RPP14 and HTD2 and a transcript coding for a single fusion protein. Searches of EST databases for zebrafish and green puffer fish revealed both transcript forms. Thus, the bicistronic relationship between the two genes is conserved over at least 400 million years of vertebrate evolution.
Similar to Rosenblad et al. (32)
, we were not able to identify an RPP14 transcript from the tunicate Ciona intestinalis, but BLAST searches identified the potential HTD2 homologue in the genome of this organism. Again, the ORF for HTD2 is predicted to be downstream of another ORF coding for a conserved hypothetical protein that is homologous to human hypothetical protein LOC58493. Further analysis of C. intestinalis ESTs supported the presence of the putative bicistronic transcript. In other metazoans such as Drosophila melanogaster, homologs of RPP14 and HTD2 are coded by two different genes located on different chromosomes (data not shown).
Purification and activity assays of HsHTD2
Attempts to express HsHTD2 or His-tagged variants in E. coli resulted in aggregation of the polypeptides in inclusion bodies. However, when HsHTD2 was expressed as a fusion to maltose binding protein, a soluble fusion protein was obtained. The recombinant protein was purified from E. coli lysate with amylose resin and further purified by gel filtration. The kcat-value of (3R)-specific hydratase 2 activity of the recombinant protein was 0.24 ± 0.07 s–1 (n=4) when using trans-2-hexenoyl-CoA substrate and 0.36 ± 0.05 s–1 (n=3) with trans-2-decenoyl-CoA substrate. When the (3S)-specific dehydroxygenase is used as an auxiliary enzyme, the reaction did not occur, indicating that the reaction is (3R) specific. The native molecular mass of recombinant MBP+HsHTD2 protein was found to be 230 kDa by size exclusion chromatography on a Superdex 200 HR column, whereas based on SDS-polyacrylamide gel electrophoresis analysis, the polypeptide was estimated to have a molecular mass of 59 kDa, indicating that the recombinant protein is a tetramer.
The active site of HsHTD2
Figure 3
A shows the sequence of human mitochondrial dehydratase aligned with mitochondrial dehydratases and other representative groups of the MaoC dehydratase family. The region around the hydratase 2 motif is presented, encompassing the most highly conserved aspartate/aspargine and histidine residues (indicated by arrows) important for the catalytic activity of the protein family. In the Candida tropicalis hydratase 2 of the multifunctional enzyme type 2, the active site residues Asp808 and His813 are located in this hydratase 2 fingerprint region (33)
. As can be seen in Fig. 3A
, these critical amino acids are highly conserved in all members of the MaoC dehydratase family and also can be found in the hydratase 2 fingerprint of human mitochondrial 3-hydroxyacyl-thioester dehydratase 2 (Asp62 and His67). These residues are important for the activity of the enzyme because mutation of these amino acids to alanine in the HsHTD2+pYE352 construct prevents the complementation of the htd2-1 mutation (not shown) and the htd2 knockout (Fig. 3B
).
|
HsHTD2 localizes to mitochondria
To investigate the subcellular localization of human 3-hydroxyacyl-thioester dehydratase, the ORF of HsHTD2 was ligated into the vector pEGFP-N1 and the resulting HsHTD2-GFP fusion protein was expressed in HeLa cells. The mitochondria were stained with MitoTracker Red, and the nuclear DNA was counterstained with DAPI. The cells transfected with pEGFP-N1-HsHTD2 show a well-defined punctate fluorescence pattern that superimposes with MitoTracker staining (Fig. 4
A). In contrast, a diffuse uniform cellular fluorescent signal, not superimposable with MitoTracker staining, was observed in the cells transfected with a plasmid lacking the 51 nucleotides encoding the first 17 N-terminal amino acids of HsHTD2 (Fig. 4B
). Thus, this experiment confirmed that HsHTD2 is localized to mitochondria and that the N-terminal region of the full-length protein is required for mitochondrial targeting. The localization of human RPP14 in HeLa cells was investigated using a construct overexpressing an inframe fusion of GFP to the C terminus of RPP14. We found that RPP14, expressed at high levels as a GFP fusion protein, showed localization throughout the cell without any specific punctate pattern in the nucleus (Fig. 4C
). This agrees well with indirect immunofluorescence results for the localization of an overexpressed RPP14–GFP fusion protein previously obtained by Jarrous et al. (34)
. In that study, the localization of RPP14 was shown to be nucleolar in a subpopulation of cells.
|
Transcription of RPP14 and HsHTD2 in human tissues
Transcription of RPP14 and HsHTD2 in different human tissues was investigated by probing a Northern blot first with a probe hybridizing only to RPP14 and, after stripping, with a probe specific for HsHTD2 (Fig. 5
). Also, separate blots were hybridized with either RPP14 or HsHTD2 probes alone. In all cases, the probes hybridized to a band with an approximate size of 1.3 kb. With the RPP14-specific probe, we also detected a band of 1.5 kb, which was not detected with the HsHTD2 probe. We were not able to detect any transcripts that contained HsHTD2 alone in our Northern blot analysis, which would correspond to isolate 3A/B obtained in our library screen. Analysis of human EST databases provided evidence for two alternative forms of exon 1, probably due to alternative promoter usage. Additionally, some ESTs revealed splicing variants lacking exon 2. Nevertheless, all the resulting transcripts differed only in their 5' ends and appear to be bicistronic. Therefore, the identity of the 1.5 kb transcript recognized solely by the RPP14 probe remains unclear.
|
The tissue-specific distribution of transcripts was identical with both probes, with the highest transcript abundance in heart and liver tissues and lower levels present in skeletal muscle, spleen, kidney, and placenta. These results corresponded well with the transcription patterns reported for human β-ketoacyl synthase and enoyl reductase of the human mitochondrial FAS pathway (16
, 17)
.
Phylogenetic analysis of the MaoC family of dehydratases
Phylogenetic analysis of representative members identified several subgroups of the MaoC family of dehydratases (Fig. 6
). As expected, hydratases of multifunctional enzymes type 2 and dehydratases of eukaryotic FAS1 both form discrete clades backed up by solid bootstrap support. Human mitochondrial dehydratases and other homologous eukaryotic enzymes form a branch with bacterial PhaJ hydratases that are involved in polyhydroxyalkanoate synthesis, and thereby energy storage, in bacteria (35)
. Unexpectedly, fungal mitochondrial dehydratases form a clade that is separate from the other eukaryotic mitochondrial dehydratases. This subfamily also includes bacterial enzymes such as the Pseudomonas sp. L1 Ich1 protein (Q14DY6_9PSED), which is annotated as itaconyl-CoA hydratase, involved in biodegradation of the tulip allergen tulipalin A. Well-conserved, uncharacterized sequences can also be found in many groups of proteobacteria, the chloroflexus group, as well as some actinomycetes but not in E. coli or other enterobacteria. The distinct position of the fungal mitochondrial dehydratases and their bacterial Ich1 homologs as a separate subfamily is strongly supported by bootstrapping and was also reflected in homology searches. BLASTP searches of the redundant Genbank database using the yeast Htd2p or the Pseudomonas Ich1 sequence as the query retrieved fungal mitochondrial dehydratases as well as bacterial Ich1 homologs, but not nonfungal eukaryotic dehydratases or bacterial PhaJ, MaoC, or NodN sequences. This result suggests that fungal mitochondrial dehydratases may have a different bacterial phylogenetic origin with respect to the other eukaryotic enzymes. It is noteworthy that the bacterial sister groups of both the fungal and the nonfungal mitochondrial dehydratases do not appear to be directly involved in bacterial FAS.
|
MaoC enzymes are implicated in both the degradation of aromatic amines and polyhydroxyalkanoate synthesis (36
, 37)
. They are related to the nodN gene product, which is involved in the production of the root hair deformation factor during the interaction of Rhizobia and leguminous plants. Phylogenetic analysis placed these sequences in a sister clade to the nonfungal dehydratase/PhaJ branch, but bootstrap support is very weak. The sequences of the fungal HDT2/Ich1 and the MaoC/NodN groups have diverged more rapidly than the other groups (Figs. 3A
and 6
). This could be the consequence of functional adaptation of the proteins. For the same reason, and due to the limited sequence information of the alignment, the position of these two groups with respect to the other branches could not be established unambiguously. Phylogenetic and bootstrap analyses indicate that fungal dehydratases could also form a sister group to the FAS1 branch or the MFE-2 branch, but notably, there are no indications that fungal dehydratases could group with the nonfungal enzymes.
Expression of 3' HsHTD2 in vivo
We took several approaches to determine if HsHTD2 is translated from the bicistronic transcript in human cells but failed to obtain clear results. To test if expression of HsHTD2 in yeast was dependent on the presence of the MET3 promoter, we cloned the original bicistronic cDNA into a promoterless single copy YCplac33-vector. This construct complemented the respiratory deficient phenotype of htd2 and the htd2-1 point mutation only very weakly compared the original library plasmid 1A. These results indicate that, although yeast is able to weakly transcribe the RPP14/HsHTD2 cDNA sequence in the absence of a bona fide yeast promoter, a genuine yeast promoter is required to obtain full complementation by the bicistronic cDNA.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
encodes a mitochondrial 3-hydroxyacyl-ACP dehydratase: 1) only library clones from the human cerebellum or kidney cDNA libraries containing the complete bicistronic cDNA of RPP14 or its last exon alone were able to complement the yeast htd2-1 mutant strain respiratory deficient phenotype; 2) the ORF encoding only HsHTD2, but not the ORF encoding RPP14 alone, was sufficient to complement both the yeast dehydratase knockout and the htd2-1 mutation on nonfermentable medium; 3) HsHTD2 belongs to the MaoC dehydratase family based on an NCBI-conserved domain search, and it contains the hydratase-2 fingerprint and accordingly shows similarity to prokaryotic acyl dehydratases and the hydratase-2 domains of Fox2p and mammalian MFE2 and FAS1; 4) an HsHTD2-GFP fusion chimera specifically localizes to mitochondria in HeLa cells, and this localization is dependent on the presence of the predicted N-terminal mitochondrial targeting sequence of HsHTD2; and 5) purified recombinant HsHTD2 is enzymatically active, catalyzing the reaction trans-2-enoyl(CoA/ACP) + H2O 
(3R)-hydroxyacyl(CoA/ACP). The bicistronic nature of the RPP14 transcript, which has remained unchanged during the evolution of vertebrates, has not been recognized previously.
Linkage between mitochondrial FAS and tRNA processing
Bicistronic transcripts in eukaryotes, by analogy to bacteria, are thought to encode proteins of related function (2)
. Although this presumption makes intuitive sense, to our knowledge it has been conclusively shown in only one case in higher eukaryotes (7)
. We do not know yet the molecular link between RNA processing and mitochondrial FAS. There is, however, growing evidence obtained in yeast that supports the notion of a functional link between these two pathways (unpublished results). One of the major roles of RNase P is the cleavage of the 5' leader sequences of tRNAs during tRNA maturation. In this context, it is intriguing to note that it was shown nearly 15 years ago that mutations in the yeast LIP5 gene, encoding lipoic acid synthase, cause defects in 5' processing of mitochondrially encoded tRNAs (38)
. The role of mitochondrial FAS has been proposed to be the production of octanoyl-ACP for lipoic acid synthesis. In yeast it has been shown that mutants with lesions in this pathway contain only 5–10% of the wild-type lipoic acid level (39
, 40)
. In light of the clear connection between mitochondrial FAS and RNA processing in yeast, it seems unlikely that RPP14 and HsHTD2 are in the same transcript just by coincidence, as this arrangement has remained conserved throughout evolution from fish to mammals and, therefore, appears to have been under positive selection. The nature of the molecular link between these two processes remains to be elucidated.
How is mitochondrial dehydratase translated from a bicistronic transcript?
An additional issue is the question of whether both RPP14 and HsHTD2 are translated from a single transcript. The bicistronic transcript described here is essentially identical to the published RPP14 cDNA clone (28)
. However, the identification of the cDNAs complementing the yeast deletion strain containing only the HsHTD2 ORF may indicate the presence of alternative transcripts or splicing products lacking the RPP14 ORF. A search of human EST databases yielded 107 clones derived from RPP14 but did not provide evidence for monocistric transcripts encoding only HsHTD2. Therefore, the short cDNA variant observed in the screen may represent an artifact that arose during library preparation. Nevertheless, the presence of the 1.5 kb transcript in our Northern blots recognized by the RPP14 probe, but not the HsHTD2 probe, raises the possibility that a monocistrinic transcript exists for human RPP14.
If HsHTD2 is translated from the bicistronic transcript, then a scenario in which translation of the second ORF is achieved through leaky translation initiation (41)
at the RPP14 start codon and scanning for an alternative start site can most likely be excluded, as there are nine additional ATGs preceding the HsHTD2 start codon. It is possible that HsHTD2 translation is regulated in mammals in a manner analogous to yeast GCN4. This is achieved by regulation of translation of the GCN4 message via modulation of translational initiation at four small ORFs upstream of the GCN4 ORF involving scanning for further start codons to reinitiate translation (42)
. In the case of the RPP14/HsHTD2 transcript, three ATGs can be found after the RPP14 ORF stop codon and before the HsHTD2 translation initiation site. An alternative scenario is the use of an internal ribosomal entry site (IRES) to translate the second gene. The poliovirus and encephalomyocarditis virus (EMCV) trancripts were the first to be described to use IRES elements (43
, 44)
. Later it was shown that cellular mRNAs can be translated despite inhibition of cap-dependent translation, potentially through the use of an IRES element (45)
. On the other hand, the very existence of IRESs in higher eukaryotes is still under debate (1)
. Our attempts to shed light on the question of translation of HsHTD2 from the bicistronic transcript by in vitro translation or detection of fluorescent reporters in HeLa cells were unsuccessful. The results of our complementation studies in yeast with MET3-driven and promoterless constructs can be interpreted in two ways. It is possible that HsHTD2 is being translated from mRNAs that are initiated downstream of the RPP14 initiator AUG. Yeast initiates transcription often at multiple sites (46)
. Another possibility is that RPP14 and HsHTD2 are both being translated from the same bicistronic mRNAs. Hence, the mechanism of translation of the dehydratase in mammalian cells, as well the regulation of this process, remain intriguing issues that need to be addressed in the future.
The central outcome of the current work is the identification of the 3' ORF of RPP14 cDNA encoding HsHTD2, the missing component of the mitochondrial FAS pathway in vertebrates. The data also show that humans and other eukaryotes may have recruited a dehydratase from a different evolutionary origin than fungi for this pathway.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication May 15, 2007. Accepted for publication August 23, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Feng, A. Witkowski, and S. Smith Down-regulation of Mitochondrial Acyl Carrier Protein in Mammalian Cells Compromises Protein Lipoylation and Respiratory Complex I and Results in Cell Death J. Biol. Chem., April 24, 2009; 284(17): 11436 - 11445. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gurvitz, J. K. Hiltunen, and A. J. Kastaniotis Heterologous Expression of Mycobacterial Proteins in Saccharomyces cerevisiae Reveals Two Physiologically Functional 3-Hydroxyacyl-Thioester Dehydratases, HtdX and HtdY, in Addition to HadABC and HtdZ J. Bacteriol., April 15, 2009; 191(8): 2683 - 2690. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Hiltunen, M. S. Schonauer, K. J. Autio, T. M. Mittelmeier, A. J. Kastaniotis, and C. L. Dieckmann Mitochondrial Fatty Acid Synthesis Type II: More than Just Fatty Acids J. Biol. Chem., April 3, 2009; 284(14): 9011 - 9015. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Schonauer, A. J. Kastaniotis, J. K. Hiltunen, and C. L. Dieckmann Intersection of RNA Processing and the Type II Fatty Acid Synthesis Pathway in Yeast Mitochondria Mol. Cell. Biol., November 1, 2008; 28(21): 6646 - 6657. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gurvitz, J. K. Hiltunen, and A. J. Kastaniotis Function of Heterologous Mycobacterium tuberculosis InhA, a Type 2 Fatty Acid Synthase Enzyme Involved in Extending C20 Fatty Acids to C60-to-C90 Mycolic Acids, during De Novo Lipoic Acid Synthesis in Saccharomyces cerevisiae Appl. Envir. Microbiol., August 15, 2008; 74(16): 5078 - 5085. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
