Selective splitting of 3'-adenylated dinucleoside polyphosphates by specific enzymes degrading dinucleoside polyphosphates.

Several 3'-[(32)P]adenylated dinucleoside polyphosphates (Np(n)N'p*As) were synthesized by the use of poly(A) polymerase (Sillero MAG et al., 2001, Eur J Biochem.; 268: 3605-11) and three of them, ApppA[(32)P]A or ApppAp*A, AppppAp*A and GppppGp*A, were tested as potential substrates of different dinucleoside polyphosphate degrading enzymes. Human (asymmetrical) dinucleoside tetraphosphatase (EC 3.6.1.17) acted almost randomly on both AppppAp*A, yielding approximately equal amounts of pppA + pAp*A and pA + pppAp*A, and GppppGp*, yielding pppG + pGp*A and pG + pppGp*A. Narrow-leafed lupin (Lupinus angustifolius) tetraphosphatase acted preferentially on the dinucleotide unmodified end of both AppppAp*A (yielding 90% of pppA + pAp*A and 10 % of pA + pppAp*A) and GppppGp*A (yielding 89% pppG + pGp*A and 11% of pG + pppGp*A). (Symmetrical) dinucleoside tetraphosphatase (EC 3.6.1.41) from Escherichia coli hydrolyzed AppppAp*A and GppppGp*A producing equal amounts of ppA + ppAp*A and ppG + ppGp*A, respectively, and, to a lesser extent, ApppAp*A producing pA + ppAp*A. Two dinucleoside triphosphatases (EC 3.6.1.29) (the human Fhit protein and the enzyme from yellow lupin (Lupinus luteus)) and dinucleoside tetraphosphate phosphorylase (EC 2.7.7.53) from Saccharomyces cerevisiae did not degrade the three 3'-adenylated dinucleoside polyphosphates tested.

Synthesis and purification of 3¢-[ 32 P]adenylated Np n Ns. 3¢-[ 32 P]Adenylated Ap 3 A, Ap 4 A, Gp 3 G and Gp 4 G were synthesized enzymatically by the use of poly(A) polymerase from E. coli (Sillero et al., 2001).The incubation mixture (0.05 ml) contained 0.02 mM [a-32 P]ATP, 60 mCi/ml), 1 mM of the Np n N and 3.7 units of poly(A) polymerase.After 3.5 h incubation at 37°C, the mixture was treated with 0.5 ml (0.5 units) of shrimp alkaline phosphatase to degrade, during a 60 min incubation, the remaining [ 32 P]ATP and 3¢-adenylated ATP, pppAp*A.Subsequently, the mixture was applied as a 5-6 cm band on a thin-layer silica gel aluminum plate containing fluorescent indicator (Merck), and chromatography was carried out for 120 min in dioxane/ammonium hydroxide/water (6:1:6, by vol.).In this system, each 3¢-adenylated derivative migrated slightly faster than its "core" Np n N (Sillero et al., 2001).The 3¢-[ 32 P]adenylated Np n Ns were localized by autoradiography, eluted from the silica gel with water, and used as potential substrates of Np n N¢-degrading enzymes.The preparation of AppppAp*A was slightly contaminated with Ap*A.Enzyme assays.For assaying the asymmetrically acting Ap 4 A hydrolases and the Ap 3 A hydrolases, the incubation mixtures (25 ml) contained 50 mM Hepes/KOH (pH 7.6), 0.02 mM dithiothreitol, 5 mM MgCl 2 , appropriate 3¢-[ 32 P]adenylated Np n N and the enzyme under investigation.For assaying yeast Ap 4 A phosphorylase, the above mixture was supplemented with 5 mM phosphate and, in the assay of (symmetrical) Ap 4 A hydrolase, 5 mM MgCl 2 was replaced with 0.1 mM CoCl 2 (Guranowski et al., 1983).The reaction mixtures were subjected to thin-layer chromatography.Aliquots of 5 ml were spotted on silica gel plates at the indicated times of incubation at 30°C, the chromatograms developed for 110 min in dioxane/ammonium hydroxide/water (6:1:6, by vol.), and the radioactive compounds detected by autoradiography and quantified with the help of an InstanImager.In pilot experiments, an amount of enzyme completely converting 12 nmoles of "core" Np n N in less than 30 min was used.Based on that, concentrations of the indicated enzymes were appropriately adjusted in order to show progress of substrate degradation on the autoradiograms.
The approach followed to determine the substrate specificity of the enzymes tested here was similar in all cases, i.e., a radiolabeled substrate was incubated with the specified enzyme and, at different times of incubation, aliquots were taken and subjected to thinlayer chromatography.The nature of the radioactive compounds was deduced mainly from their chromatographic position and coelution with standards.From the radioactive products generated, it could be inferred whether the cleavage of the dinucleotide took place at the phosphoanhydride bond located in position 1, 2 or 3, counting from the unmodified (nonadenylated) nucleotide end.The results obtained are presented in Figs.1-3 and summarized in Table 1.When AppppAp*A (Fig. 1A) or GppppGp*A (Fig. 2A) were treated with (asymmetrical) dinucleoside tetraphosphatase from narrow-leafed lupin, cleavages at positions 3 (preferential) and 1 were observed in both cases (Table 1).In contrast, random cleavage at positions 1 and 3 was obtained when AppppAp*A (Fig. 1B) or GppppGp*A (Fig. 2B) were treated with dinucleoside tetraphosphatase from human placenta.Exhaustive treatment of GppppG*A (Fig. 2B) with the latter enzyme resulted in complete disappearance of the substrate.
Such a difference in the preference of cleavage of the asymmetrical substrates exerted by the plant and human/animal types of (asymmetrical) dinucleoside tetraphosphatases can be explained by the differences in topography of the substrate binding sites between these two subgroups of Ap 4 A hydrolases whose three-dimensional structures have been revealed only recently (Swarbrick et al., 2000;Bailey et al., 2002;respectively).Although the plant and animal Ap 4 A hydrolases belong to the same Nudix protein family, they have low sequence similarity outside the Nudix sequence motif.Human and Caenorhabditis elegans Ap 4 A hydrolases are very similar (Abdelghany et al., 2001) and analysis of the three-dimensional structure of the latter reveals much more space in its substrate binding site than one can observe in such a site of the lupin counterpart.This is probably the reason why human Ap 4 A hydrolase tolerated such a bulky substituent as a nucleotide residue in its potential substrates and bound the investigated asymmetrical substrates randomly, whereas the plant (lupin) enzyme clearly preferred to interact with the asymmetrical substrates from their unmodified end.
The (symmetrical) dinucleoside tetraphosphatase from E. coli cleaved at position 2 (Table 1) with either AppppAp*A (Fig. 1C) or  The reaction mixtures contained about 0.4 mM GppppGp*A, the indicated enzyme and other components as described in Materials and Methods.
ApppAp*A was not a substrate of the reactions catalyzed by either yellow lupin dinucleoside triphosphatase (Fig. 3A) or by the human Fhit protein (not shown), at the same experimental conditions in which unlabeled 0.5 mM ApppA was completely transformed to pA + ppA in less than 30 min.In contrast, ApppAp*A was a substrate for E. coli (symmetrical) Ap 4 A hydrolase that can also hydrolyze NpppN to ppN and pN (Guranowski et al., 1983).In the chromatographic system used we observed one of the products, ppAp*A (Fig. 3A).Comigration of ApppApA with pApA does not allow one to determine whether two pairs of products accumulate: the observed ppAp*A + pA and the alternative one, ppA + pAp*A.Both (asymmetrical) Ap 4 A hydrolases and yeast Ap 4 A phosphorylase did not degrade ApppAp*A, in the same way as the core ApppA was not a substrate of these enzymes (Guranowski & Blanquet, 1985;Jakubowski & Guranowski, 1983;La¿ewska et al., 1993).This study shows that both Ap 4 A phosphorylase, and the human (Fhit protein) and yellow lupin Ap 3 A hydrolases are rather strict concerning the 3¢-adenylation of their respective substrates, Ap 4 A and Ap 3 A. Similarly, previous studies showed that 2¢-(deoxy)adenylated Ap 3 A and Ap 4 A were not substrates of the human Ap 3 A hydrolase (Guranowski et al., 2000).In contrast, both the symmetrical and asymmetrical Np 4 N¢ hydrolases are able to cleave their 3¢-adenylated substrates Ap 4 A or Gp 4 G, in line with previous findings showing the susceptibility of 2¢-deoxyadenylated Ap 4 A to the hydrolysis catalyzed by (asymmetrical) dinucleoside tetraphosphatases from human (Guranowski et al., 2000) or lupin (Maksel et al., 2001).Altogether, these results show that both types of Ap 4 A-degrading enzymes tolerate such a bulky substituent as adenylate at the 3¢ or 2¢ position of their substrates.The substrate specificity of rat liver (asymmetrical) dinucleoside tetraphosphatase was previously tested using AppppA, AppppddA and ddAppppddA as substrates (Sillero et al., 1997).The main conclusion was that lack of only one of their 3¢-OH residues greatly diminishes the rate of catalysis of the enzyme, as reflected by the similarity between the actual velocities observed with equal concentrations of AppppddA and ddAppppddA.With AppppddA, the products of the reaction were preferentially pA and pppddA, i.e. the enzyme cleaved the substrate at position 1.It is clear that more investigation is needed to elucidate the mechanism of catalysis of the enzymes cleaving specifically this type of dinucleotides.Unfortunately, 3¢-adenylated substrates were available in limited amounts which precluded detailed kinetic studies.The work presented here, although largely qualitative, widens nevertheless the spectra of substrate specificities of the dinucleoside polyphosphate-cleaving enzymes.Since none of the eukaryotic Np n N¢-degrading enzymes investigated here was able to degrade 3¢-adenylated dinucleoside triphosphates, Np 3 NpAs, involvement of those enzymes in mRNA decapping is unlikely.The decapping is controlled by different, very specific enzymes (Liu et al., 2002;Milone et al., 2002).The reaction mixtures contained about 0.5 mM ApppAp*A, the indicated enzyme and other components as described in Materials and Methods.

Figure 1 .
Figure 1.Analysis of enzymatic cleavage of AppppA[ 32 P]A (AppppAp*A) by thin-layer chromatography.The reaction mixtures contained about 0.2 mM AppppAp*A, the indicated enzyme and other components as described in Materials and Methods.Accumulation of radioactivity close to the chromatogram origin, observed mostly in lanes C, resulted from [ 32 P]orthophosphate that had been liberated from the ppAp*A product due to the action of phosphodiesterase and nucleotidase/phosphatase which apparently contaminated the partially purified preparation of the E. coli hydrolase.

Table 1 .
3¢-Adenylated dinucleoside polyphosphates as substrates of enzymes degrading dinucleoside polyphosphates a Cleavage site 1, 2 or 3 refers to the first, second or third (where applicable) phosphoanhydride bond, counting from the unmodified end of the dinucleotide; b No cleavage was observed; N.A., not assayed; not applicable; p* corresponds to 32 P-labeled products.The investigated compounds were substrates neither for lupin nor human (Fhit protein) dinucleoside triphosphatase (EC 3.6.1.29).