INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

4-Coumarate:CoA ligase (4CL; EC 6.2.1.12) represents the branch point enzyme of plant phenylpropanoid metabolism, catalyzing the activation of 4-coumaric acid and various other hydroxylated and methoxylated cinnamic acid derivatives to the corresponding CoA esters in a two-step reaction. During the first step (Eq. 1), coumarate and ATP form a coumaroyl-adenylate intermediate with the simultaneous release of pyrophosphate (PP_i), a reaction which is reversible (Eq. 2).

(1)

(2)

In the second step (Eq. 3), the coumaroyl group is transferred to the sulfhydryl group of CoA, and AMP is released (Knobloch and Hahlbrock, 1975; Becker-André et al., 1991).

(3)

On the basis of the presence of a highly conserved peptide motive, the putative nucleotide-binding domain, 4CLs, luciferases, fatty acyl-CoA synthetases, acetyl-CoA synthetases, and non-ribosomal peptide synthetases have been grouped in one superfamily of adenylate-forming enzymes (Fulda et al., 1994). A second, independent group of adenylate-forming proteins make up the aminoacyl-tRNA synthetases (Pavela-Vrancic et al., 1994). In the presence of ATP and Mg²⁺, all enzymes listed above form an enzyme-bound acyl-adenylate intermediate, which is esterified with either CoA (4CL, acetyl-CoA synthetases, and fatty acyl-CoA synthetases), or the enzyme-bound cofactor 4'-phosphopantetheine (non-ribosomal peptide synthetases), or tRNAs (aminoacyl-tRNA synthetases).

Several of the enzymes forming an acyl-adenylate intermediate with concomitant release of PP_i can in a second type of reaction also catalyze the formation of uncommon mononucleoside and dinucleoside polyphosphates such as adenosine 5'-tetraphosphate (p₄A; Eq. 4), adenosine 5'-pentaphosphate (p₅A; Eq. 5), diadenosine 5',5"-P¹,P⁴-tetraphosphate (Ap₄A; Eq. 6), and diadenosine 5',5"-P¹,P⁵-pentaphosphate (Ap₅A; Eq. 7; for reviews, see McLennan, 2000; Sillero and Günther Sillero, 2000).

(4)

where P₃ is tripolyphosphate.

(5)

where P₄ is tetrapolyphosphate.

(6)

(7)

Such capacity was demonstrated for certain aminoacyl-tRNA synthetases (Plateau et al., 1981; Goerlich et al., 1982; Plateau and Blanquet, 1982; Jakubowski, 1983), some acetyl-CoA and fatty acyl-CoA synthetases (Guranowski et al., 1994; Fontes et al., 1998), firefly luciferase (Guranowski et al., 1990), and more recently also for a non-ribosomal peptide synthetase from Bacillus brevis (Dieckmann et al., 2001). Corresponding to this broad range of enzymes catalyzing reactions that in vitro lead to the synthesis of (di)nucleoside polyphosphates, appreciable amounts of these compounds, p_nN and Np_nN', have been found to also naturally occur in a variety of tissues and organisms, including bacteria, fungi, and animal cells (Garrison and Barnes, 1992; Kisselev et al., 1998; McLennan, 2000). They presumably occur ubiquitously in all organisms, however, their presence in plants has not yet been demonstrated.

Although (di)nucleoside polyphosphate synthesis is typically promoted by the absence or an insufficient concentration of the respective end acceptor (e.g. CoA, 4'-phosphopantetheine, and tRNAs) the (di)adenosine polyphosphate product spectra observed are specific for each enzyme. For example, most aminoacyl-tRNA synthetases can catalyze the formation of Ap₄A, whereas mammalian aminoacyl-tRNA synthetases with specificity for Trp only produce Ap₃A (Kisselev et al., 1998). Adenosine 5'-polyphosphates (e.g. p₄A) and diadenosine 5'-polyphosphates (e.g. Ap₃A and Ap₄A) have been considered to play important roles in regulating cellular processes such as sporulation, stress responses, cell proliferation, and apoptosis in various biological systems (Kisselev et al., 1998; Nishimura, 1998). In Brewer's yeast (Saccharomyces cerevisiae), p₄A and p₅A accumulate during sporulation, whereas these compounds were not detectable during its vegetative growth (Jakubowski, 1986). Further studies demonstrated that yeast acetyl-CoA synthetase can synthesize p₄A and p₅A and therefore might be the enzyme responsible for the accumulation of these adenosine 5'-polyphosphates (Guranowski et al., 1994). In bacteria (Escherichia coli, Synechococcus sp.), yeast (S. cerevisiae), and animal cells (fruitfly [Drosophila melanogaster] and Artemia franciscana) the concentrations of diadenosine polyphosphates (Ap_nA) rapidly increased in response to heat shock, oxidative, or heavy-metal ion stress, suggesting that these compounds could act as alarmones (Brevet et al., 1985; Baltzinger et al., 1986; Miller and McLennan, 1986; Coste et al., 1987; Johnstone and Farr, 1991; Pàlfi et al., 1991). Several proteins were identified to directly bind diadenosine polyphosphates, including the stress-related proteins DnaK and GroEL from E. coli (Johnstone and Farr, 1991). In higher eukaryotes, especially mammalian cells, it has been observed that appreciable physiological and pathological effects are associated with changes of diadenosine polyphosphate levels, in particular with alterations of the Ap₃A to Ap₄A ratio (Kisselev et al., 1998). In agreement with the suggested importance of the Ap₃A to Ap₄A ratio, the human Ap₃A hydrolase, Fhit, appeared to be involved in the protection of cells against tumorigenesis (Siprashivili et al., 1997).

Phenylalanyl- and seryl-tRNA synthetases from yellow lupin seeds are the only plant enzymes known to synthesize Ap₃A and Ap₄A (Jakubowski, 1983). However, the existence of highly specific enzymes that catalyze the hydrolysis of (di)nucleoside polyphosphates, such as (asymmetrical) dinucleoside tetraphosphatase (EC 3.6.1.17), dinucleoside triphosphatase (EC 3.6.1.29), and nucleoside 5'-tetraphosphatase (EC 3.6.1.14), indicates that these nucleotide derivatives may exist and have a function in planta (Jakubowski and Guranowski, 1983; Guranowski et al., 1996, 1997). In this report, we demonstrate that Arabidopsis 4CL2, a key enzyme of plant secondary metabolism, can catalyze the synthesis of various mono- and dinucleoside polyphosphates. The basic requirements and kinetic parameters of these reactions are presented, and their possible biological roles are discussed.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Synthesis of Adenosine 5'-Polyphosphates

Having identified At4CL2 as a member of the large family of adenylate-forming enzymes (Stuible et al., 2000), we wanted to test whether At4CL2 and two mutant variants derived from this enzyme are also capable to catalyze the synthesis of p₄A or p₅A. The respective enzymes were incubated with either P₃ or P₄ as potential adenylate acceptors in the presence of their cognate phenolic acid substrate (a cinnamic acid derivative), ATP and Mg²⁺, whereas CoA was omitted from the incubation mixture. For product analysis, aliquots were sampled at different time points, subjected to thin layer chromatography and subsequently visualized under short wavelength UV. Various At4CL2 variants (the wild-type enzyme and mutants derived thereof) support the synthesis of p₄A and p₅A in the presence of P₃ or P₄, respectively. In Figure 1, this is demonstrated for the mutant M293P/K320L. The identity of the reaction products, p₄A and p₅A, was verified by their comigration with authentic standards and their susceptibility to yeast exopolyphosphatase, scPPX1, which degrades nucleoside polyphosphates to ATP and orthophosphate (Guranowski et al., 1998). Addition of inorganic pyrophosphatase to the reaction mixture had no significant effect on the rates of p₄A and p₅A production and was therefore not included in the standard assay.

View larger version (54K): [in this window] [in a new window]   Figure 1.   Synthesis of p4A and p5A by the At4CL2 mutant M293P/K320L from Arabidopsis. Product synthesis from ATP and P3 or P4 in the absence (no FA) and presence of 62 µM ferulic acid (+FA) was monitored by withdrawing 3-µL aliquots of the reaction mixtures at the times indicated. After 120 min, the reactions were stopped by heating (3 min at 100°C) and further incubated at 30°C in the presence of 1 ng of scPPX1 (exopolyphosphatase from Brewer's yeast), and two additional aliquots were withdrawn at the times indicated. Samples were subjected to thin-layer chromatography in solvent system I.

For the At4CL2 wild-type enzyme, nucleoside polyphosphate synthesis was strictly dependent on the presence of a cinnamic acid derivative, such as coumarate or caffeate, which are efficiently converted to their corresponding thiol-esters by the enzyme in the standard 4CL reaction (Ehlting et al., 1999). In the presence of 100 µM coumarate, a specific activity of 6.2 nkat mg¹ protein for p₄A synthesis was determined, whereas in its absence p₄A was undetectable. Ferulate and cinnamate, in contrast to coumarate and caffeate, are known to represent only poor substrates for the At4CL2 wild-type enzyme and correspondingly were considerably less efficient in stimulating p₄A synthesis (Table I). Sinapate, which is not converted to the CoA ester by At4CL2, does also not support p₄A synthesis. In Table I, the kinetic properties of the At4CL2 wild-type catalyzed reactions leading to the synthesis of p₄A or the CoA esters of various cinnamic acid derivatives are compared. Interestingly, apparent K_m values for coumarate and caffeate are about 10-fold lower in the reaction leading to p₄A in comparison with the corresponding values for thiol-ester formation.

The At4CL2 mutant enzyme carrying two amino acid substitutions, M293P and K320L, was analyzed in detail for its capacity to catalyze adenosine 5'-polyphosphate formation. The At4CL2 mutant M293P/K320L is a gain of function variant of At4CL2, which in contrast to the wild-type enzyme is capable to efficiently activate ferulate and cinnamate due to its enlarged and more hydrophobic substrate binding pocket (Stuible and Kombrink, 2001; K. Schneider, K. Witzel, E. Kombrink, and H.-P. Stuible, unpublished data). In addition to its broader substrate utilization spectrum, the double mutant exhibits higher conversion rates than the wild-type enzyme, with several cinnamic acid derivatives as substrates during CoA ester formation. In accordance with these properties, the double mutant also displayed high in vitro activity of (di)adenosine polyphosphate synthesis. Under optimal conditions and in the presence of 62 µM ferulate the specific activities for p₄A and p₅A synthesis were 7.9 and 0.8 nkat mg¹, respectively. Moreover, without added ferulate, a very low but detectable amount of p₄A was synthesized, accumulating at a 25-fold lower rate than in the presence of ferulate (Fig. 1; Table II). The rates of p₄A and p₅A synthesis were constant over a time period of at least 40 min when accumulation of radiolabeled products originating from [³H]ATP was quantified by scintillation counting (Fig. 2). When different cinnamic acid derivatives were compared, ferulate and coumarate were found to be the best inducers of p₄A synthesis, whereas caffeate and cinnamate were less efficient (Table II). In contrast, sinapate, Phe, and Tyr, which are no substrates for thiol-ester formation by the At4CL2 mutant M293P/K320L, also did not support p₄A synthesis. Table II summarizes the kinetic properties of p₄A synthesis by the At4CL2 mutant M293P/K320L and relates these values to the kinetic properties of thiol-ester formation with CoA. As observed for the wild-type enzyme, the apparent K_m values for the different cinnamic acid derivatives were at least 10-fold lower for p₄A synthesis in comparison with CoA ester formation. However, the relative reaction rates with each of the phenolic substrates were comparable for both reactions.

View larger version (9K): [in this window] [in a new window]   Figure 2.   Time course of p4A and p5A synthesis catalyzed by the At4CL2 mutant M293P/K320L from Arabidopsis. Product accumulation was monitored by thin-layer chromatography as described in "Materials and Methods." Synthesis of p4A was monitored in the presence of 5 mM [3H]ATP and 10 mM P3 and either in the presence of 62 µM ferulic acid () or in its absence (). Synthesis of p5A proceeded in the presence of 5 mM [3H]ATP, 10 mM P4, and 62 µM ferulate (). The chromatogram was developed in solvent system I and radioactivity in p4A or p5A spots determined by scintillation counting.

The K_m values of all other components participating in p₄A formation by the At4CL2 mutant M293P/K320L were determined under standard conditions, with fixed concentrations of ferulate (62 µM), P₃ (10 mM), and ATP (5 mM), respectively. The K_m for the adenylate donor ATP was 150 µM when determined with [³H]ATP concentrations ranging from 37 to 310 µM. This value is of the same order of magnitude as the K_m observed for the 4CL-catalyzed standard reaction (CoA ester formation), typically ranging from 50 to 250 µM for enzymes from different plants (Knobloch and Hahlbrock, 1975; Voo et al., 1995; Stuible et al., 2000). The K_m for the adenylate acceptor P₃ was 4.9 mM, as estimated from measurements with P₃ concentrations varying from 0.62 to 10 mM.

To get some insight into the reaction mechanism leading to (di)adenosine polyphosphate synthesis by At4CL2, we extended our analysis to the mutant enzyme K540N. The Lys residue 540 has been suggested to function as the active site of adenylate formation by stabilizing the negatively charged pentavalent transition state (Conti et al., 1997). In agreement with this assumption, the At4CL2 variant carrying the K540N substitution was incapable of activating cinnamic acid derivatives to the corresponding CoA esters (Stuible et al., 2000). Likewise, this mutant is also incapable of catalyzing the synthesis of (di)adenosine polyphosphates (not shown), indicating that adenylate formation is an essential step during both CoA ester formation and (di)adenosine polyphosphate synthesis.

Special Requirements for Adenosine 5'-Polyphosphate Synthesis

The pH dependence of p₄A synthesis catalyzed by At4CL2 was monitored in various buffers, including acetate, MES, HEPES, CHES, covering a pH range from 3.6 to 9.5. Maximum activity was observed between pH 6 and 8, with half-maximal activity at pH 4.5 and 9 (results not shown). 4CL-catalyzed reactions converting cinnamic acid derivatives to CoA esters typically have a pH optimum ranging from pH 7.5 to 8.5 (Knobloch and Hahlbrock, 1975; Knobloch and Hahlbrock, 1977; Voo et al., 1995). Various divalent cations were tested for their capacity to support p₄A synthesis, and maximum activity was obtained with 5 mM Mg²⁺ (100%), whereas lower rates were observed with Mn²⁺ (54%), Co²⁺ (52%), Ni²⁺ (47%), and Zn²⁺ (15%), all tested at 5 mM. Ca²⁺ had no stimulating effect.

Synthesis of Dinucleoside Polyphosphates

In additional experiments, we analyzed the capacity of the At4CL2 mutant M293P/K320L to support the synthesis of dinucleoside polyphosphates, such as Ap₃A, Ap₄A, dAp₄dA, and Ap₅A. When the enzyme was supplied with 5 mM ATP, which served as both adenylate donor and acceptor, 5 mM MgCl₂ and 62 µM ferulate, accumulation of Ap₄A could be observed by thin-layer chromatography (Fig. 3A). The identity of the product was confirmed by comigration with the authentic standard, its resistance to alkaline phosphatase, and its susceptibility to Ap₄A hydrolase from L. angustifolius, which converts Ap₄A to ATP plus AMP (Fig. 3). Maximum rates of Ap₄A synthesis were observed between pH 6 and 7 (not shown). The kinetic constants of Ap₄A synthesis were estimated with [³H]ATP concentrations ranging from 0.3 to 5 mM and yielded an apparent K_m for ATP of 4.1 mM and V_max of 1.2 nkat mg¹. When ATP was replaced by dATP, the At4CL2 mutant M293P/K320L also catalyzed the synthesis of dAp₄dA at a comparable rate (Fig. 3B). With p₄A as the only nucleotide present in the reaction mixture, the enzyme additionally supported the synthesis of Ap₅A (not shown). However, no Ap₃A synthesis could be detected in a reaction mixture containing 5 mM ADP as potential adenylate acceptor, in addition to the standard ingredients (5 mM ATP, 5 mM MgCl₂, and 62 µM ferulate).

View larger version (44K): [in this window] [in a new window]   Figure 3.   Synthesis of diadenosine tetraphosphate (Ap4A) and di(deoxyadenosine) tetraphosphate (dAp4dA) by the At4CL2 mutant M293P/K320L from Arabidopsis. Product accumulation was monitored by withdrawing 3-µL aliquots of the reaction mixtures at the times indicated. After 16 h, the reactions were stopped by heating (3 min, 100°C) and further incubated at 30°C in the presence of 2 units of alkaline phosphatase (AP) from calf intestine and after its inactivation, with 50 ng of (asymmetrical) Ap4A hydrolase (Ap4A-ase) from Lupinus angustifolius, and additional aliquots were withdrawn at the times indicated. Samples were subjected to thin-layer chromatography in solvent system II as described in "Materials and Methods." A and dA at the arrows represent adenosine and deoxyadenosine, respectively.

The At4CL2 wild-type enzyme also catalyzed the synthesis of Ap₄A, albeit at a 10-fold lower rate than the At4CL2 double mutant. The specific activity determined in the presence of 100 µM coumarate as activator was 0.12 nkat mg¹.

Mono- and Dinucleoside Polyphosphates as AMP-Moiety Donors

In the following experiments, we analyzed whether (di)nucleoside polyphosphates can also serve as adenylate donors in At4CL2-catalyzed reactions, thereby leading to regeneration of ATP at the expense of the unusual nucleotide derivatives. In fact, these reactions are the reverse of the reactions characterized above (compare Eq. 2 with Eqs. 4-7). As shown in Figure 4, the At4CL2 mutant M293P/K320L efficiently catalyzed the transfer of the AMP moiety from p₄A onto PP_i resulting in the accumulation of ATP. Ap₄A and Ap₅A could also serve as adenylate donors, when polyphosphates such as PP_i, P₃, and P₄ were supplied as acceptors. Incubation of the At4CL2 mutant M293P/K320L with [³H]Ap₄A and P₃ led to the accumulation of both p₄A and ATP (Fig. 5). The K_m for Ap₄A in this reaction was 64 µM and the V_max 0.53 nkat mg¹; the pH optimum was between pH 5.5 and 6.5, and the optimum MgCl₂ concentration, 7.5 mM (not shown). The transfer of the AMP moiety from [³H]Ap₄A to the acceptors PP_i or P₄, yielding two molecules of either ATP or ATP plus p₄A, occurred with lower efficiency (not shown). The relative reaction rates determined in the presence of 10 mM each of PP_i, P₃, or P₄ were 100:5.3:1. All of the above reactions were strictly dependent on the presence of a suitable phenolic compound as cocatalyst and hence were tested in the presence of 62 µM ferulate.

View larger version (71K): [in this window] [in a new window]   Figure 4.   Synthesis of ATP from p4A and PPi catalyzed by the At4CL2 mutant M293P/K320L from Arabidopsis. Time course of the reaction was monitored as described in "Materials and Methods." The chromatogram was developed in solvent system I.

View larger version (84K): [in this window] [in a new window]   Figure 5.   Synthesis of p4A and ATP from Ap4A and P3 by the At4CL2 mutant M293P/K320L from Arabidopsis. Time course of the reaction was monitored as described in "Materials and Methods." The chromatogram was developed in solvent system II.

Finally, the AMP moiety originating from Ap₄A could also be transferred to orthophosphate (P_i). Using [³H]Ap₄A as adenylate donor and 5 mM potassium phosphate as acceptor, equal amounts of radioactivity appeared in the products ADP and ATP that were generated by At4CL2 mutant M293P/K320L in the presence of 62 µM ferulate. This result clearly shows that the reaction proceeds via a feruloyl-adenylate intermediate from which the AMP moiety is transferred onto P_i. A symmetric, P_i-independent cleavage of Ap₄A resulting in two ADP molecules can therefore be excluded.

When incubated with Ap₅A, instead of Ap₄A, and the same oligophosphate acceptors as above (e.g. PP_i, P₃, and P₄), the At4CL2 mutant M293P/K320L catalyzed the synthesis of the corresponding products: Two molecules of p₄A were generated from Ap₅A and P₃ (Fig. 6), whereas p₄A and p₅A originated from Ap₅A and P₄ (not shown).

View larger version (82K): [in this window] [in a new window]   Figure 6.   Synthesis of p4A from Ap5A and P3 by the At4CL2 mutant M293P/K320L from Arabidopsis. Time course of the reaction was monitored as described in "Materials and Methods." The chromatogram was developed in solvent system I.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Here, we demonstrate that 4CL2 from Arabidopsis, in addition to its bona fide function of generating CoA esters of cinnamic acid derivatives, is also capable to catalyze the in vitro synthesis of (di)adenosine polyphosphates from ATP and (nucleoside) polyphosphates. This novel reaction strictly depends on the presence of at least traces of an authentic phenolic substrate. The assumption that a cinnamoyl-adenylate intermediate acts as AMP donor during catalysis is strengthened by the observation that the At4CL2 mutant enzyme K540N, which carries a point mutation in the catalytic center responsible for adenylate formation (Stuible et al., 2000), is incapable to synthesize (di)adenosine polyphosphates. Similar results were recently obtained by studying the enzymatic properties of the heterologously expressed adenylation domain of tyrocidine synthetase A, which also efficiently catalyzes the formation of various mono- and dinucleoside polyphosphates (Dieckmann et al., 2001). The adenylation domain of this non-ribosomal polypeptide synthetase shares significant structural and sequence similarities with At4CL2 (Stuible et al., 2000).

Formation of (di)adenosine polyphosphates has been reported for several, but not all, enzymes that perform catalysis via an adenylate intermediate. For example, highly purified acetyl-CoA synthetases from E. coli (kindly donated by Dr. Alan J. Wolfe), bovine heart, and lupin seeds (A. Guranowski and A. Biryukov, unpublished data) did not catalyze the synthesis of p₄A, Ap₄A, or Ap₃A, although all preparations were fully active to carry out the acetylation of CoA. On the other hand, acetyl-CoA synthetases purified from yeast (A. Guranowski, unpublished data) and some, but not all, commercially available preparations of yeast acetyl-CoA synthetase (A. Guranowski, M.A. Günther Sillero, and A. Sillero, unpublished data) were able to synthesize p₄A, Ap₃A, and Ap₄A. These examples indicate that the capacity of an enzyme to synthesize (di)nucleoside polyphosphates is difficult to predict and therefore has to be verified experimentally in each single case.

On the basis of their different requirements for low M_r AMP donors functioning as cocatalytic factor, such as fatty acids, luciferin, amino acids or cinnamic acid derivatives, adenylate-forming enzymes exhibiting the capacity to synthesize (di)adenosine polyphosphates synthesis can be further subclassified. One group of enzymes, including most of aminoacyl-tRNA synthetases (Plateau et al., 1981; Goerlich et al., 1982) and firefly luciferase (Guranowski et al., 1990) strictly require their cognate acid substrate for diadenosine polyphosphate synthesis. In contrast, members of the second group, composed of lysyl-tRNA synthetase from E. coli (Zamecnik et al., 1966; Goerlich et al., 1982), seryl-tRNA and phenylalanyl-tRNA synthetases from yellow lupin (Jakubowski, 1983), and fatty acyl-CoA synthetase from Pseudomonas fragi (Fontes et al., 1998), can synthesize measurable amounts of (di)adenosine polyphosphates in the absence of their primary AMP acceptor, such as fatty acids or amino acids. However, even with the latter group of enzymes, a stimulation of p₄A synthesis by cognate acid substrates seems to be possible, as previously demonstrated for acetyl-CoA ligase from Brewer's yeast (Guranowski et al., 1994).

From the above observations, the existence of two alternative molecular mechanisms for (di)adenosine polyphosphate synthesis can be proposed, one operating via an enzyme-bound acyl~adenylate (aforementioned Eqs. 1 and 4-7), the other operating via an enzyme~adenylate (Eqs. 8 and 9)

(8)

(9)

Adenosine 5'-polyphosphate synthesis catalyzed by the At4CL2 wild-type enzyme obviously functions according to the mechanism described for example by Equations 1 and 4, because its activity was absolutely dependent on the presence of a cinnamic acid derivative.

Although p₄A is the most prominent nucleotide derivative synthesized by At4CL2 wild-type and mutant M293P/K320L, a significant production of different dinucleoside polyphosphates has also been observed. The At4CL2 variant M293P/K320L catalyzes the synthesis of Ap₄A and dAp₄dA at comparable rates (Fig. 3). The capacity to synthesize these nucleotide derivatives at comparable rates has previously been observed for the lysyl-tRNA synthetase from E. coli (Plateau and Blanquet, 1982), firefly luciferase (Günther Sillero et al., 1991), and acyl-CoA synthetase from P. fragi (Fontes et al., 1998). At4CL2 wild-type and the mutant M293P/K320L were also able to catalyze the synthesis of Ap₅A. In this reaction, p₄A either acts as an acceptor of adenylate donated by ATP, or alternatively, functions both as an AMP-donor and acceptor when representing the only available nucleotide. ADP is obviously not recognized as an AMP acceptor because neither enzyme variant produced Ap₃A, thereby resembling properties of firefly luciferase, which synthesized Ap₄A but not Ap₃A (Günther Sillero et al., 1991). In contrast, several aminoacyl-tRNA synthetases produced Ap₃A (Ap₃N) and Ap₄A (Ap₄Ns) at comparable rates (Zamecnik et al., 1966; Plateau and Blanquet, 1982; Jakubowski, 1983; Kitabatake et al., 1987).

Finally, we observed that At4CL2 was also capable to catalyze reverse reaction, i.e. the (re)use of the adenylate moieties originating from adenosine 5'-polyphosphates and diadenosine polyphosphates for the generation of ATP. This activity may connect the degradation of unusual nucleotide derivatives with the regeneration of ATP. Comparable enzymatic activities have previously been reported for some aminoacyl-tRNA synthetases (Zamecnik et al., 1966; Plateau et al., 1981; Goerlich et al., 1982; Led et al., 1983; Guédon et al., 1987) as well as for firefly luciferase (Momsen, 1978; Günther Sillero et al., 1991; Ortiz et al., 1993). The functional significance of these reactions remains to be demonstrated but it could be related to (a) recycling the energy stored in the PP_i linkages of (di)nucleoside polyphosphates back to ATP, (b) necessary degradation of important signaling molecules, or (c) efficient removal of otherwise toxic compounds that accumulate as byproducts of adenylate-transfer reactions.

4CL represents a key enzyme of the general phenylpropanoid metabolism, which connects primary metabolism with various product-specific pathways by supplying appropriate mixtures of substrates for these subsequent reactions (Dixon and Paiva, 1995; Douglas, 1996). Due to its position at a metabolic branch point, 4CL represents a prime target for regulatory control mechanisms. In this context, the observation that At4CL2 is capable to synthesize and reuse unusual nucleotide derivatives, which may function as intracellular messengers is obviously intriguing. Interestingly, 4CL can support the synthesis of Ap₄A but does not transfer the AMP-moiety onto ADP to yield Ap₃A. This is contrasted by properties of some other plant enzymes, such as phenylalanyl-tRNA synthetase from yellow lupin, which can produce both Ap₃A and Ap₄A (Jakubowski, 1983). Whether these differences in catalytic properties of different adenylate-forming enzymes contribute to the Ap₃A to Ap₄A ratio in plant cells and whether this ratio has a regulatory function or determines cell fate, such as apoptosis versus proliferation as previously suggested for some mammalian model systems (Vartanian et al., 1997), remains to be elucidated. Nevertheless, the idea that unusual nucleotide derivatives may have a regulatory function in plant metabolism is strengthened by the recent identification of Arabidopsis enzymes involved in (p) ppGpp synthesis. Characterization of the respective proteins suggest a function of (p) ppGpp in mediating stress-induced defense responses in plants (van der Biezen et al., 2000).

In conclusion, the data obtained in this study clearly show that, in the absence of CoA, 4CL can function as a p_nA and Ap_nA synthetase, provided that alternative adenylate acceptors such as oligophosphates (P_n) or adenosine oligophosphates (p_nA) are present at sufficient concentrations. In fact, 4CL represents the first plant enzyme, in addition to aminoacyl-tRNA synthetases, for which the capacity to synthesize these unusual mono- and diadenosine polyphosphates has been demonstrated. Thus, our results shine new light on an old enzyme, 4CL, which may have an important function beyond the well-established role in phenylpropanoid metabolism.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Chemicals and Enzymes

Coenzyme A, mono- and dinucleotides, buffers, and salts were from Sigma-Aldrich (St. Louis), [2,8-³H]ATP (25 Ci mmol¹) was from ICN (Irvine, CA), and [³H]Ap₄A (20 Ci mmol¹) was from Moravek Biochemicals (Brea, CA).

Recombinant At4CL2 proteins, the wild-type form, and the two mutants M293P/K320L and K540N were expressed and purified as previously described (Stuible et al., 2000; Stuible and Kombrink, 2001). The following enzymes were used as analytical tools: yeast inorganic pyrophosphatase (Sigma-Aldrich), calf intestinal alkaline phosphatase (Promega, Madison, WI), recombinant exopolyphosphatase from yeast, scPPX1 (kindly supplied by Dr. S. Liu, Department of Biochemistry, Beckman Center, Stanford, CA; Guranowski et al., 1998), and recombinant Ap₄A hydrolase from Lupinus angustifolius (kindly donated by Drs. K. Gayler and D. Maksel, University of Melbourne; Maksel et al., 2001).

Chromatographic Media and Solvent Systems

Thin-layer chromatography was performed with aluminum sheets precoated with silica gel containing fluorescent indicator (catalog no. 5554, E. Merck, Darmstadt, Germany). For monitoring of the synthesis of mononucleoside polyphosphates, the chromatograms were developed in dioxane:ammonia:water mixed at the ratio 6:1:6 (v/v/v), (system I) and for monitoring of the synthesis of dinucleoside polyphosphates that ratio was 6:1:4 (v/v/v, system II).

Enzyme Assays

The synthesis of adenosine 5'-polyphosphates p₄A and p₅A was determined in an assay mixture (50 µL final volume) containing 100 mM HEPES/KOH (pH 7.0), 5 mM MgCl₂, 5 mM ATP, 10 mM polyphosphate (P₃ or P₄), and either 100 µM coumarate and 12.5 µg of At4CL2 wild-type enzyme or 62 µM ferulate and 3.75 µg of the ferulate-preferring At4CL2 mutant M293P/K320L. Incubations were carried out at 30°C. At appropriate times, usually 10 to 120 min, 3-µL aliquots were withdrawn and spotted onto thin-layer plates. Chromatograms were developed for 90 min in solvent system I, dried, and visualized/photographed under short wave UV light. For quantitative assays, the mixture was supplied with [³H]ATP (2.5 µCi); at appropriate times 5-µL aliquots were withdrawn for analysis by thin-layer chromatography, and radioactivity in p₄A or p₅A spots was determined by scintillation counting. For estimation of kinetic constants, pH optima, etc., the conditions varied appropriately.

The synthesis of diadenosine 5',5-P¹,P⁴-tetraphosphate, Ap₄A, was assayed in an assay mixture (50 µL final volume) containing 100 mM HEPES/KOH (pH 7.0), 5 mM MgCl₂, 5 mM ATP, 62 µM ferulic acid, and 7.5 µg of the At4CL2 mutant M293P/K320L. Incubations were carried out at 30°C and lasted up to several hours. Five-microliter aliquots were withdrawn and spotted onto thin-layer plates, and chromatograms were developed in solvent system II. To monitor the synthesis of dAp₄dA or Ap₅A, dATP or p₄A substituted ATP. When the activity of the At4CL2 wild-type was determined, 100 µM coumarate was used instead of ferulate.

The transfer of the adenylate moiety from mono- or dinucleoside 5'-polyphosphates onto various acceptors, was monitored in a reaction mixtures (50 µL final volume) containing 100 mM HEPES/KOH (pH 7.0), 5 mM MgCl₂, 62 µM ferulate, an adenylate donor (1 mM p₄A, 0.5 mM Ap₄A, or 1 mM Ap₅A), an adenylate acceptor (10 mM PP_i, 5 mM P₃, or 5 mM P₄), and 7.5 µg of the At4CL2 mutant M293P/K320L. The time course of product accumulation in these reactions was monitored as above on thin-layer chromatograms, which were developed either in system I or II.

Synthesis of CoA esters of different cinnamic acid derivatives by At4CL2 was determined by the spectrophotometric assay as previously described (Ehlting et al., 1999; Stuible et al., 2000).