INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Aberrant transcripts of the human FHIT (fragile histidine triad) gene have been detected in nearly 50% of esophageal, stomach, and colon carcinomas (26). Accordingly, deletions in the FHIT gene, which is located at chromosome region 3p14.2, were found associated with many types of common human cancers, including lung (7, 37), skin (36), head and neck (22, 38), breast (20, 23), cervical (9), and colorectal (10) cancers. Actually, the FHIT gene encompasses FRA3B, the most highly inducible fragile site in the human genome (13, 26). The FRA3B breakpoints, including the renal carcinoma-associated one, t(3;8), fall within introns 3, 4, and 5 of the FHIT gene and are likely to be at the origin of the cancer-associated deletions (39). However, although the strong correlation between FHIT alterations and cancer development may indicate a suppressor function of the FHIT gene product, this function has not yet been proven (for reviews, see references 13, 16, and 27).

A marked similarity between the protein product of FHIT and the amino acid sequence of a Schizosaccharomyces pombe enzyme described as a diadenosine 5',5'''-P¹,P⁴-tetraphosphate (Ap₄A) asymmetrical hydrolase (12, 33) was noted early (26). Ap₄A belongs to an unusual family of ubiquitous dinucleoside polyphosphates, the possible functions of which are still debated. At one time, these nucleotides were suspected of being involved in the proliferation activity of animal cells (for a review, see reference 8). However, this hypothesis did not receive support from further studies (for a review, see reference 32). Because their intracellular concentration increases in response to oxidative stresses (for reviews, see references 15 and 28), dinucleoside polyphosphates were also proposed to participate to cellular adaptation. Finally, extracellular functions have been envisaged. For instance, dinucleoside polyphosphates were reported to interfere with cardiovascular and neurotransmission activities (for reviews, see references 2 and 24).

The Fhit protein could be produced from cDNA in Escherichia coli. In agreement with its resemblance to an Ap₄A hydrolase, this protein accelerated the in vitro hydrolysis of various dinucleoside polyphosphates, including Ap₄A (1). The most efficient substrate was diadenosine 5',5'''-P¹,P³-triphosphate (Ap₃A). In light of the putative cancer suppression function of the FHIT product, the observations described above supported the idea of some role of dinucleoside polyphosphates in tumor development. Whether dinucleoside triphosphates rather than dinucleoside tetraphosphates would be involved is still an open question.

In Saccharomyces cerevisiae, Ap₄A is transformed into ATP plus ADP by two Ap₄A phosphorylases encoded by two distinct genes, APA1 and APA2 (30, 31). To recycle Ap₃A, which is fully resistant to the two Ap₄A phosphorylases, the yeast cell produces an Ap₃A hydrolase capable of converting Ap₃A into ADP plus AMP (3). The enzyme also slowly hydrolyzes Ap₄A into ATP plus AMP.

The S. cerevisiae genome sequence data indicates a gene (APH1) product highly similar to the Ap₄A hydrolase of S. pombe (12, 26). To determine whether this protein product behaves like the S. pombe Ap₄A hydrolase or whether it corresponds to the previously described S. cerevisiae Ap₃A hydrolase, we have undertaken its expression in E. coli. In vitro characterization of the obtained protein strongly indicates that it corresponds to the already isolated Ap₃A hydrolase, therefore resembling the Fhit protein. In addition, gene disruption experiments unambiguously show that the APH1 product is the main actor in S. cerevisiae Ap₃N (N = A, C, G, or U) catabolism. This conclusion emphasizes the putative predominance of dinucleoside triphosphates in any FHIT-mediated function.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Yeasts, bacteria, and plasmids. The strains used are listed in Table 1. Bacterial and yeast transformations were performed by electroporation and by the lithium acetate method of Ito et al. (14), respectively. YPD and minimal sporulation media were as described before (34). Plasmid pBluescript SK() was from Stratagene (San Diego, Calif.).

Nucleotides. Ap₄C and Ap₄G were extracted from E. coli PAL2103D (18) as described previously (30). Ap₃C was enzymatically synthesized with purified E. coli lysyl-tRNA synthetase (29). Other nucleotides were from Boehringer (Ap₄A and Ap₅A), Sigma (Ap₃A), or Pharmacia (Ap₃G, Gp₃G, and Gp₄G). [³H]Ap₄A (reference batch TRQ.4405, 159 GBq/mmol, radiochemical purity of >98%) was from Amersham.

Preparation of S. cerevisiae crude extracts. For the preparation of S. cerevisiae crude extracts, cells were grown until the optical density at 650 nm of the culture reached 2 ± 0.5. After centrifugation at 12,000 × g for 15 min, the cell pellet was suspended in 50 mM Tris-HCl buffer (pH 7.8) containing 0.1 mM EDTA, 10 mM 2-mercaptoethanol, and 1 mM phenylmethylsulfonide fluoride (PMSF). Cells were sonicated twice for 5 min at 0°C, and cell debris was removed by centrifugation at 17,000 × g for 15 min. The total amount of protein in the supernatant was determined by using the Bio-Rad protein assay reagent.

Enzymatic assays. Hydrolysis of [³H]Ap₄A was used to monitor enzyme activity; the reaction mixture (100 µl) contained 50 mM Tris-HCl (pH 7.8), 5 mM MgCl₂, 0.1 mM EDTA, 50 µM [³H]Ap₄A (2 GBq/mmol), and a 140-U/ml concentration of alkaline phosphatase from calf intestine (2,000 U/mg; Boehringer). After an incubation ranging from 10 to 90 min at 37°C, the [³H]adenosine produced was counted as described previously (31).

Determination of Ap₄N and Ap₃N concentrations in S. cerevisiae crude extracts. Nucleotides were extracted with perchloric acid from 30-ml aliquots of the cultures, as described previously (31). The sample was then neutralized by K₂CO₃ and diluted fivefold in a Tris-HCl buffer (20 mM, pH 7.8) containing 0.1 mM ZnCl₂ and 1 mM MgCl₂. After digestion of ATP with alkaline phosphatase from calf intestine (3,000 U/mg, enzyme immunoassay grade; Boehringer), two successive luminescence assays were performed (4). In the first one, a 10-µl extract sample was mixed with 100 µl of the ATP-bioluminescence HS mixture from Boehringer. After extinction of the luminescence background due to ATP contamination, snake venom phosphodiesterase (Boehringer) was added at a final concentration of 5 µg/ml and the luminescence was recorded. In this assay, only Ap₄N nucleotides contribute to the signal. In the second assay, 10 µl of sample was mixed with 100 µl of the luminescence mixture containing, in addition, 0.25 mM phosphoenolpyruvate (from Sigma) and 35 U of rabbit muscle pyruvate kinase (Boehringer) per ml. After extinction of the signal caused by contaminating ATP and ADP, phosphodiesterase was added at a final concentration of 5 µg/ml and the luminescence was recorded. Under these conditions, both the Ap₃N and Ap₄N compounds contributed to the light signal. The Ap₃N concentration was deduced from the difference between the two assays (4, 25).

Cloning of the S. cerevisiae APH1 gene. The APH1 gene of S. cerevisiae was amplified by PCR using 25 pmol of the oligonucleotides ATABam (5'-CTTAGAATGCAGCGGATCCTTGGGATTAGC-3') and ATAXho (5'-TTTACTGTTGAGTCTCCTCGAGGAAAGTAG-3') and 3.4 µg of genomic DNA from strain CMY214 (30 cycles; 95°C for 1 min, 40°C for 1 min, 72°C for 2 min). The sequences of the two primers encompass the BamHI and XhoI sites flanking the APH1 gene (Fig. 1). The amplified DNA fragment of 1.3 kbp was purified with the Qiagen Plasmid Mini Kit 100. After digestion by BamHI and XhoI, the purified DNA fragment was inserted into the corresponding sites of plasmid pBluescript SK() to give plasmid p35.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Construction of plasmid p19. The p35-inserted APH1 gene is represented by a rectangle with the intron shaded. Two PCR amplifications were performed with plasmid p35 as the template and either ATA3 plus ATA1, or ATA2 plus ATAXho, as the primer pair. Primers ATA1 and ATA2 are complementary to the exonic sequences on both sides of the intron. The two amplified DNA fragments were mixed and further subjected to PCR with ATA3 and ATAXho as the primers. The resulting DNA was cut with BamHI and XhoI, the sites of which are present in ATA3 and ATAXho, respectively, and inserted into plasmid pBluescript SK(), to give plasmid p19.

Disruption of the APH1 gene in S. cerevisiae. To inactivate the APH1 gene, a 75-bp NsiI-BglII fragment internal to the gene was deleted and substituted by the selectable marker TRP1. For this purpose, the TRP1 marker carried by plasmid pRS314 (35) was amplified by PCR with primers TRPNsi (5'-GAGTGATGCATAAACGACATTACTATATATA-3') and TRPBgl (5'-GTCACAGATCTGGCAAGTGCACAAACAATA-3') (30 cycles; 95°C for 1 min, 45°C for 1 min, 72°C for 2 min). The resulting DNA was digested by NsiI and BglII before insertion into p35.

Recombination of the APH1 gene for overexpression in E. coli cells. Deletion of the intron harbored by the APH1 gene and insertion of a canonical Shine-Dalgarno sequence upstream of this gene was performed as described in the legend to Fig. 1. Briefly, two PCR amplifications were performed with plasmid p35 as the template and either ATA3 (5'-GCGGGATCCCTAGAAAGGAGGTACGATCATGAATAAGCCAATATATTTCAGCA-3') plus ATA1 (5'-ATATTTTGACTTATAGAAAACTTGTTCAGTTACAAG-3') or ATA2 (5'-CAAGTTTTCTATAAGTCAAAATATACGTATGCATTG-3') plus ATAXho as the primer pair (30 cycles; 95°C for 1 min, 50°C for 1 min, 72°C for 2 min). After Qiagen purification, the two resulting DNA fragments were mixed and further subjected to 30 cycles of amplification using ATA3 and ATAXho as primers. The final purified DNA was cut by BamHI and XhoI and inserted into the corresponding sites of plasmid pBluescript SK(), to give plasmid p19. The nucleotide sequence of the insert was verified.

Purification of the APH1 gene product. JM101TR cells harboring plasmid p19 were grown in 125 ml of LB medium containing 100 µg of ampicillin per ml and 1 mM IPTG (isopropyl--D-thiogalactopyranoside). When the optical density at 650 nm reached 2.0, the cells were harvested by a 30-min centrifugation at 12,000 × g. The cell pellet, resuspended in 8 ml of a 50 mM Tris-HCl buffer (pH 7.8) containing 10% glycerol and 0.5 mM PMSF, was sonicated for 4 min at 0°C. After removal of cellular debris by centrifugation (30 min, 12,000 × g), nucleic acids were precipitated by the addition of streptomycin sulfate (3% [wt/vol] final concentration) and centrifuged (30 min, 12,000 × g). The supernatant (7.5 ml) was dialyzed against 1 liter of 20 mM potassium phosphate (pH 7.0)-10% glycerol-0.1 mM PMSF and applied to a DEAE Sephadex column (1 by 5 cm) equilibrated in 20 mM potassium phosphate (pH 7.0). The column was eluted by use of 20 to 200 mM linear gradient of potassium phosphate (80 ml, pH 7.0) at a flow rate of 6 ml/h. During the purification, the activity of the APH1 gene product was monitored through hydrolysis of either Ap₄A or Ap₃A. The fractions containing activity were pooled and concentrated by ammonium sulfate precipitation. At the end of this step, the specific Ap₃A degradation activity of the sample had increased ~16-fold in comparison to that of the crude extract.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Expression of the S. cerevisiae APH1 gene in E. coli. In 1996, the sequencing of the FHIT cDNA revealed homologies between the Fhit protein and two yeast proteins, an Ap₄A hydrolase produced by the S. pombe aph1 gene and the product of an S. cerevisiae gene (26). The latter gene was named APH1 because of its homology to the S. pombe aph1 gene. It was also designated HNT2 by other authors (see accession no. U28374 in the EMBL and GenBank databases) and received the name YDR305c in the systematic genome sequence designation. Two possible in-frame initiator codons are found in the 5' region of the S. cerevisiae APH1 gene; they correspond to either MILSK or MNKPI N-terminal sequences. The second one is likely to be used as the translational start because it corresponds to the initiator methionine in S. pombe. The coding sequence of the APH1 gene is interrupted by a single intron of 89 bp. The 5' boundary of this intron is located 56 bp downstream of the putative translational start. The protein encoded by the APH1 gene is expected to contain 206 amino acids (M_r = 23,541).

Disruption of the APH1 gene. In S. cerevisiae, Ap₄A catabolism is mainly sustained by two Ap₄A phosphorylases produced by the APA1 and APA2 genes (30, 31). Strains fully devoid of Ap₄A phosphorylase activity have been described. In crude extracts of such a strain, a weak Ap₄A-hydrolyzing activity was measurable (3). This hydrolysis could be assigned to an Ap₃A hydrolase having a rather broad specificity. To examine whether the weak Ap₄A hydrolysis described above was associated with the APH1 gene, we have undertaken the knockout of this gene in an apa1 apa2 context and measured dinucleoside tri- and tetraphosphate-hydrolyzing activities in cell extracts.

his ura trp

his⁺ ura trp

his ura⁺ trp

his ura trp⁺

his⁺ ura⁺ trp

his⁺ ura trp⁺

his ura⁺ trp⁺

Ap₃N and Ap₄N concentrations in various aph1 mutants. In the control strain YPALS (APA1 APA2 APH1), Ap₃N and Ap₄N concentrations are equal to 0.4 µM. These concentrations were increased 30- and 2.5-fold, respectively, upon disruption of the APH1 gene only (Table 3). The variation of Ap₄N may appear relatively small in view of the broad specificity of the APH1 product. However, in the YPALS and YPALST strains, two Ap₄A phosphorylases which are likely to control Ap₄N concentration occur, thereby counterbalancing the effect of the absence of Ap₃A hydrolase activity on the cellular pool of these nucleotides.

Conclusions. Altogether, the present results establish that APH1, an analog of FHIT in S. cerevisiae, encodes a protein product involved mainly in Ap₃N catabolism. This product corresponds to the Ap₃A hydrolase already characterized (3). The similarity between the in vitro functions of the APH1- and FHIT-encoded polypeptide products sustains the view that Ap₃N concentration in human cells also might be deregulated by FHIT inactivation.