- Email: [email protected]
Nucleoside triphosphate synthesis catalysed by adenylate kinase is ADP dependent
ABB Archives of Biochemistry and Biophysics 444 (2005) 195–199 www.elsevier.com/locate/yabbi
Nucleoside triphosphate synthesis catalysed by adenylate kinase is ADP dependent Martin Willemoe¨s a,*, Mogens Kilstrup b a
b
Department of Biological Chemistry, Institute of Molecular Biology and Physiology, University of Copenhagen, Sølvgade 83H DK-1307 Copenhagen, Denmark Microbial Physiology and Genetics Group, BioCentrum, Technical University of Denmark, Building 301, DK-2800 Lyngby, Denmark Received 5 September 2005, and in revised form 3 October 2005 Available online 2 November 2005
Abstract Adenylate kinase (Adk) that catalyses the synthesis of ADP from ATP and AMP has also been shown to perform an ATP dependent phosphorylation of ribo- and deoxynucleoside diphosphates to their corresponding nucleoside triphosphate; ATP + (d)NDP M ADP + (d)NTP. This reaction, suggested to occur by the transfer of the c-phosphoryl from ATP to the nucleoside diphosphate, is overall similar to that normally carried out by nucleoside diphosphate kinase (Ndk). Accordingly, Adk was proposed to be responsible for residual Ndk-like activity measured in a mutant strain of Escherichia coli, where the ndk gene was disrupted. We present data supporting a mechanism for the synthesis of nucleoside triphosphates by Adk that unlike the previously suggested mechanism mentioned above are in complete agreement with the current knowledge about the Adk enzyme and its various catalytic properties. We propose that nucleoside triphosphate synthesis occurs by b-phosphoryl transfer from ADP to any bound nucleoside diphosphate. Our results point to the fact that the proposed Ndk-like mechanism of Adk originated from an erroneous interpretation of data, in that contamination of ATP preparations with AMP and ADP was not taken into account. Our results also address the proposed role of Adk in restoring a normal growth rate of mutant strains of E. coli lacking Ndk. These mutant strains apparently, in spite of a mutator phenotype, are able to synthesise nucleoside triphosphates by alternative pathways to maintain the same growth rate as the wildtype. 2005 Elsevier Inc. All rights reserved. Keywords: Adenylate kinase; Nucleoside diphosphate kinase; Adenylate energy charge; Nucleoside triphosphate synthesis; Nucleotide metabolism
Adenylate kinase (Adk)1 catalyses the important phosphorylation of AMP to ADP in the purine nucleotide synthesis pathway according to Eq. (1) ATP þ AMP $ 2ADP
ð1Þ
The equilibrium constant of this reaction is defined in Eq. (2) h 2 K eq ¼ ADP =½ATP½AMP ð2Þ
*
Corresponding author. E-mail address: [email protected] (M. Willemoe¨s). 1 Abbreviations used: Adk, adenylate kinase; DTT, dithiothreitol; Ndk, nucleoside diphosphate kinase; Ap 5 A, P 1 ,P 5 -di(adenosine-5 0 ) pentaphosphate. 0003-9861/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2005.10.003
where Keq is close to unity [1,2] and the reaction will therefore readily proceed in both directions being either ADP- or ATP and AMP synthesis. Adk has two distinct nucleotide binding sites designated by their role in the forward reaction; the ATP binding site, that has a low stringency for the nucleobase moiety, and the AMP binding site, which is highly specific for the adenine base. Both sites can bind ADP in the reverse reaction [3–5]. Purification of a residual nucleoside diphosphate kinase (Ndk) activity from the cell extract of a mutant Escherichia coli strain, where the ndk gene had been inactivated, showed that this activity resided with Adk. From a product profile analysis, it was concluded that Adk had the property to transfer the c-phosphoryl of ATP to all tested purine
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
and pyrimidine deoxy- or ribonucleoside diphosphates. A similar analysis of Adk from rabbit and chicken suggested that this activity was a general property of Adk [6]. Indeed, Ndk-like activity has been shown in the same type of analysis by a separate group, to also be associated with Adk from Mycobacterium tuberculosis [7]. The formulation of the mechanism for the Ndk-like reaction requires that Adk allows for binding of the substrate nucleotide diphosphate to the AMP binding site [6] even though this site is very specific for the nucleotides AMP and ADP, as mentioned above. Another intriguing property of the Ndk-like reaction was the inability of GTP to function as a c-phosphoryl donor as opposed to the phosphorylation of AMP to ADP [6]. Despite the apparent support by data for the suggested mechanism for the Ndk-like reaction of Adk, some puzzling questions regarding this activity remain that were not addressed in the original work. Since the Ndk-like reaction will form ADP, if it proceeds as suggested, i.e., ATP + (d)NDP fi ADP + (d)NTP, this would immediately equilibrate with ATP and AMP by the normal action of Adk. ADP (bound in the AMP site) can then combine on the enzyme with (d)NDP (bound in the low stringency ATP site) to form AMP and (d)NTP via the analogues to the reverse reaction of Adk. The formed AMP can then be converted back to ADP by the normal Adk activity, and the Ndklike activity would not be needed for further NTP synthesis. In the present work, we have re-investigated the synthesis of nucleoside triphosphates from the corresponding diphosphate as catalysed by Adk. We show that the previously suggested mechanism for the Ndk-like activity of Adk is most likely based on an artifact of the assay procedure by which this property of the enzyme was identified and investigated. Experimental procedures Chemicals, including all nucleotides, were purchased from Sigma–Aldrich. Chicken muscle Adk was obtained from Sigma–Aldrich and was dissolved in 50 mM Hepes, pH 7.5, 100 mM KCl, and 2 mM DTT to a concentration of 2000 U/ml. Radiolabelled nucleotides were obtained from Amersham ([5-3H]CTP), Perkin–Elmer ([a-32P]CTP) or NEN ([8-14C]ADP). Polyethyleneimine cellulose sheets for thin layer chromatography were obtained from Merck. All assays were performed at 37 C in 100 ll containing 50 mM Hepes, pH 7.5, 100 mM KCl, 2 mM DTT, and 1 mM MgCl2. Adk was added to a final concentration of 0.4 U/ml. The concentration of nucleotides varied as described under Results. At the given time points, 10 ll of the incubation was transferred to 5 ll of 0.33 M formic acid and immediately spotted on PEI-cellulose sheets. The relevant marker nucleotides were added to the spots and dried. The chromatograms were developed in 1.3 M LiCl and 1 M acetic acid [8]. When 3H- or 14C-labelled nucleotides were used, the relevant spots identified by the
marker nucleotides were cut out and transferred to scintillation vials. Two milliliters of 2 M ammonia was added to each vial to dissolve the nucleotide bound to PEI-cellulose sheet. The vials were left to stand for 1 h when scintillation liquid (EcoScint A, National Diagnostics) was added and the vials shaken to mix scintillation liquid with the water phase. Radioactivity was quantitated on a Packard Tri-carb 2200CA scintillation counter. When 32 P labeled nucleotide was used, the radioactivity in the spots was quantitated directly on a Packard Instant Imager. Results Effect of AMP on the CTP synthesis catalysed by Adk In the experiment shown in Fig. 1, we analysed the rate of CTP synthesis from (commercial analytical grade) ATP and radioactive CDP. As seen, there is a long lag period before CTP synthesis commences when only ATP is added to the incubation. However, at increasing concentrations of added AMP, the reaction is greatly stimulated. This apparent superposition of the two equilibria, involving cytidine and adenine nucleotides, led us to study the influence of CDP on the adenine nucleotide equilibrium of the Adk reaction (Eq. (1)). Influence of CTP synthesis on the equilibrium concentrations of adenine nucleotides in the Adk catalysed reaction When ADP was added as substrate for the Adk reverse reaction (Figs. 2A and D), the conditions were found to result in immediate equilibration of AMP, ADP, and ATP. In Figs. 2B and C, the addition of CDP to the reaction mixture is seen to result in a drastic time dependent displacement of AMP, ADP, and ATP concentrations concomitant with CTP synthesis. However, as shown in Fig. 2D, the adenine nucleotides are in equilibrium throughout the entire time span of the assay.
100 80
CTP, mM
196
60 40 20 0 0
500 1000 Time, s
1500
Fig. 1. AMP dependent synthesis of CTP catalysed by Adk. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 100 lM ATP, 200 lM [5-3H]CDP, and AMP added at (circles) 0 lM, (squares) 10 lM or (triangles) 20 lM.
A 100
C 100
Nucleotides, mM
Nucleotides, mM
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
80 60 40 20
80
80
60
60
40
40
20
20
0
500
0
1000 1500
B 100
D
500
1000 1500
2
Keq
40
500
1000 1500
0
60 120 Time, s
20 0
0 500
0
40
20 0
1000 1500
60
1
0
500
80
80 60
0
B 100
3 CTP, mM
0
Nucleotides, mM
A 100
0
0
197
1000 1500
0
500
20
C
1000 1500
Time, s
15
Fig. 2. Equilibration of adenyl nucleotides during CTP synthesis catalysed by Adk. Assays were performed as described in Experimental procedures. The initial [8-14C]ADP concentration was 100 lM and CTP was added at (A) 0 mM, (B) 0.2 mM, and (C) 1 mM; circles represent AMP, squares represent ADP, and triangles represent ATP. (D) The equilibrium constant, Keq, defined in Eq. (2), is calculated from the data of (A) circles, (B) squares, and (C) triangles.
The time course of CTP synthesis at two different CDP concentrations was determined with ADP as the phosphate donor, under similar conditions as for the experiments in Figs. 2B and C. The results are presented in Fig. 3A, where the rate of CTP synthesis is seen to increase with the concentration of CDP. Fig. 3B shows the results from an experiment where a similar adenine nucleotide concentration to that in Fig. 3A was added as equimolar amounts of ATP and AMP. As shown, the rates of CTP synthesis are identical whether ADP (Fig. 3A) or ATP and AMP (Fig. 3B) were added. This is a clear confirmation of the rapid Adk dependent equilibration of the adenine nucleotides shown in Fig. 2. As a control that we were actually measuring an Adk dependent activity, we determined the rate of CTP synthesis in the absence and presence of P1,P5-di(adenosine-5 0 ) pentaphosphate (Ap5A), the classic bi-substrate analog and potent inhibitor of Adk [9]. From Fig. 3C, it is evident that no CTP synthesis occurs at even 1 lM Ap5A. We conclude that CTP synthesis reported in the present work is derived from an Adk dependent reaction. Synthesis of CDP from CTP by the reverse reaction in the presence of AMP or ADP To investigate the reversibility of the Adk dependent nucleoside triphosphate synthesis, we also measured CDP synthesis from CTP. In Fig. 4, the time dependency on
5 0 180
Fig. 3. CTP synthesis catalysed by Adk. Assays were performed as described in Experimental procedures. The initial adenyl nucleotide concentrations were (A) 100 lM ADP or (B) 50 lM each of ATP and AMP. In both (A and B) circles represent initial 0.2 mM [5-3H]CDP and squares represent initial 1 mM [5-3H]CDP. (C) Ap5A inhibition of CTP synthesis by Adk. The initial substrate concentrations were 0.1 mM ADP and 0.2 mM [5-3H]CDP. Ap5A was present at concentrations of (circles) 0 lM or (squares) 1 lM.
CDP synthesis is shown in the presence of AMP or ADP. From these results, it is clear that AMP by far is best at stimulating the rate of CDP synthesis. CTP synthesis using GTP as a phosphoryl donor In agreement with the previous result of Lu and Inouye [6], it is clear from Fig. 5 that GTP is not a good phosphoryl 50 40
CDP, mM
CTP synthesis from ADP or equimolar amounts of ATP and AMP
10
30 20 10 0 0
500
1000 Time, s
1500
Fig. 4. Reverse reaction of CTP synthesis as catalysed by Adk. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 1 mM [a-32P]CTP and either (circles) 100 lM AMP or (squares) 100 lM ADP.
198
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199 100
CTP, mM
80 60 40 20 0 0
1500
3000
4500
Time, s
Fig. 5. Adk dependent CTP synthesis using GTP as phosphoryl donor. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 100 lM GTP and 200 lM [5-3H]CDP. AMP was added to a concentration of (circles) 0 lM or (squares) 10 lM.
donor for CTP synthesis as catalysed by Adk. However, the addition of a low concentration of AMP to the reaction greatly stimulates CTP synthesis, as also observed for the ATP dependent reaction (Fig. 1). Discussion Adk was previously identified as the enzyme that substituted for Ndk in a E. coli ndk strain, and was found to catalyse the transfer of the c-phosphoryl from ATP to deoxy- or ribonucleoside diphosphates [6]. The proposed reaction scheme can be written as follows: ATP þ ðdÞNDP $ ADP þ ðdÞNTP
ð3Þ
However, this mechanism that overall appears to be similar to that of Ndk, except for the preference for ATP as phosphoryl donor, involves the violation of several known restricted properties of Adk as mentioned in the introduction. Although, we were readily able to reproduce the previously reported results [6] (data not shown), our additional work clearly showed that a mechanism for Adk dependent nucleoside triphosphate synthesis as illustrated by Eq. (3) is unlikely. Instead, it appears that the transfer of the c-phosphoryl of ATP to any (d)NDP occurs via the following reactions that violate none of the known catalytic restrictions for Adk and agrees with long known stimulation of the Adk reaction by various nucleoside diphosphates [3]: ATP þ AMP $ 2ADP ADP þ ðdÞNDP $ AMP þ ðdÞNTP
ð4Þ
Because commercial ATP is always contaminated with small amounts of AMP, the reaction will proceed even if AMP is not deliberately added to the incubation, and ADP will be synthesised via the Adk reaction (compare with Eq. (1)). ADP bound in the AMP site can then donate the b-phosphoryl to any (d)NDP bound to the ATP site, which is much more tolerant to the nucleobase moiety of the bound nucleotide. By this mechanism, only trace amounts of AMP are needed initially, since AMP will be recycled and the concentration increased during the time course of the reaction.
The scheme is consistent with the observed acceleration of CTP synthesis in the absence of added AMP and the activation of CTP synthesis by addition of AMP is seen in Fig. 1. Furthermore, Eq. (4) is also in agreement with the observation that AMP by far is the best substrate over ADP for the reverse reaction of CDP synthesis from CTP (Fig. 4). Since AMP and ADP were added at equimolar concentrations in the two experiments in Fig. 4, the result obtained with ADP can be understood in terms of the equilibrium for the Adk reaction. The initial AMP concentration under this condition will be about one-third compared to the situation where only AMP was added. Relevant to the above arguments is our demonstration in Fig. 2 that Adk catalyses the equilibration between ATP, ADP, and AMP at rates greatly exceeding the rate of CTP synthesis (Figs. 3A and B). Finally, we also offer an explanation for the puzzling inability of GTP to act as phosphoryl donor in the Ndk-like reaction of Adk (Fig. 5). Commercial GTP, unlike ATP, is not likely to contain AMP. However, when AMP is added to the reaction, CTP is formed (Fig. 5). The reason for this observation can now be understood on the basis of the mechanism in Eq. (4); GTP can, by replacing ATP, phosphorylate AMP to ADP that in turn will transfer the b-phosphoryl group to CDP yielding CTP as shown in Fig. 5. The reaction shown by Eq. (4), where ADP acts as the phosphoryl donor, is also in agreement with the observed ability of Adk to phosphorylate thiamine diphosphate to its triphosphate [10,11] and further illustrates the relaxed binding specificity in the ATP binding site of Adk. The problem, not realised when the original mechanism for the Ndk-like activity of Adk was forwarded in the work of Lu and Inouye [6], is the role of trace impurities of AMP working as a ‘‘coenzyme’’ for the reaction. The nature of the more trivial mechanism for Adk dependent nucleoside triphosphate synthesis identified in this work addresses important questions as to whether Adk will be solely responsible in vivo for restoring the normal growth-rate of an E. coli strain lacking the Ndk [12]. Adk was both identified as the residual Ndk activity in this mutant E. coli strain and subsequently characterised based on the same erroneous assay. Although, as we show in this work, Adk is able to synthesise nucleoside triphosphates following the scheme outlined in Eq. (4), this activity will in vivo be strictly governed by the intracellular concentration of all participating nucleotides, and most importantly not only ATP but also ADP and AMP. In this respect, the interplay between the Adk catalysed reactions is by far more complicated than the relatively simple equilibration between nucleoside di- and triphosphates carried out by the normal Ndk enzyme. Due to the broad substrate specificity and the ‘‘neutral’’ phosphorylated enzyme intermediate, that can donate a c-phosphoryl from any nucleoside triphosphate to any nucleoside diphosphate [13–15], Ndk can optimally balance deoxy- and ribonucleoside di- and triphophosphate ratios according to the supply and demand of nucleotide metabolism.
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
In an E. coli strain devoid of Ndk enzyme, all deoxy- or ribonucleoside diphosphates will compete for the ADP pool for synthesis of nucleoside triphosphates. In other words, the synthesis of nucleoside triphosphates in this mutant strain will be governed by the adenylate energy charge [16,17], that is, if Adk is the sole source of residual Ndk-like activity. Therefore, we find it plausible that other reversible metabolic reactions that will yield, or equilibrate, between nucleoside di- and triphosphates, all together, plays a role in sustaining the total synthesis of nucleoside triphosphates [18]. Bearing in mind that cells lacking Ndk activity clearly must have severe changes in the pools of the various nucleotides compared to wild type cells as confirmed by the mutator phenotype of the mutant strains [12]. Acknowledgment This work was supported by the Danish Natural Science Research Council. References [1] I.A. Rose, Proc. Natl. Acad. Sci. USA 61 (1968) 1079–1086. [2] S. Su, P.J. Russell Jr., J. Biol. Chem. 243 (1968) 3826–3833.
199
[3] W.E. Criss, T.K. Pradhan, Methods Enzymol. 51 (1978) 459–467. [4] K.K. Tsuboi, Methods Enzymol. 51 (1978) 467–473. [5] K. Miura, S. Inouye, K. Sakai, H. Takaoka, F. Kishi, M. Tabuchi, T. Tanaka, H. Matsumoto, M. Shirai, T. Nakazawa, A. Nakazawa, J. Biol. Chem. 276 (2001) 13490–13498. [6] Q. Lu, M. Inouye, Proc. Natl. Acad. Sci. USA 93 (1996) 5720–5725. [7] L.S. Meena, P. Chopra, R.S. Bedwal, Y. Singh, Biotechnol. Appl. Biochem. 38 (2003) 169–174. [8] K. Randerath, E. Randerath, J. Chromatogr. 16 (1964) 111–125. [9] G.E. Lienhard, I. Secemski, J. Biol. Chem. 248 (1973) 1121–1123. [10] H. Shikata, Y. Egi, S. Koyama, K. Yamada, T. Kawasaki, Biochem. Int. 18 (1989) 943–949. [11] H. Shikata, S. Koyama, Y. Egi, K. Yamada, T. Kawasaki, Biochem. Int. 18 (1989) 933–941. [12] Q. Lu, X. Zhang, N. Almaula, C.K. Mathews, M. Inouye, J. Mol. Biol. 254 (1995) 337–341. [13] J. Munoz-Dorado, N. Almaula, S. Inouye, M. Inouye, J. Bacteriol. 175 (1993) 1176–1181. [14] J. Munoz-Dorado, S. Inouye, M. Inouye, J. Biol. Chem. 265 (1990) 2707–2712. [15] R.L. Ratliff, R.H. Weaver, H.A. Lardy, S.A. Kuby, J. Biol. Chem. 239 (1964) 301–309. [16] C.C. Glembotski, A.G. Chapman, D.E. Atkinson, J. Bacteriol. 145 (1981) 1374–1385. [17] A.G. Chapman, L. Fall, D.E. Atkinson, J. Bacteriol. 108 (1971) 1072– 1086. [18] A. Kuroda, A. Kornberg, Proc. Natl. Acad. Sci. USA 94 (1997) 439– 442.
Nucleoside triphosphate synthesis catalysed by adenylate kinase is ADP dependent Martin Willemoe¨s a,*, Mogens Kilstrup b a
b
Department of Biological Chemistry, Institute of Molecular Biology and Physiology, University of Copenhagen, Sølvgade 83H DK-1307 Copenhagen, Denmark Microbial Physiology and Genetics Group, BioCentrum, Technical University of Denmark, Building 301, DK-2800 Lyngby, Denmark Received 5 September 2005, and in revised form 3 October 2005 Available online 2 November 2005
Abstract Adenylate kinase (Adk) that catalyses the synthesis of ADP from ATP and AMP has also been shown to perform an ATP dependent phosphorylation of ribo- and deoxynucleoside diphosphates to their corresponding nucleoside triphosphate; ATP + (d)NDP M ADP + (d)NTP. This reaction, suggested to occur by the transfer of the c-phosphoryl from ATP to the nucleoside diphosphate, is overall similar to that normally carried out by nucleoside diphosphate kinase (Ndk). Accordingly, Adk was proposed to be responsible for residual Ndk-like activity measured in a mutant strain of Escherichia coli, where the ndk gene was disrupted. We present data supporting a mechanism for the synthesis of nucleoside triphosphates by Adk that unlike the previously suggested mechanism mentioned above are in complete agreement with the current knowledge about the Adk enzyme and its various catalytic properties. We propose that nucleoside triphosphate synthesis occurs by b-phosphoryl transfer from ADP to any bound nucleoside diphosphate. Our results point to the fact that the proposed Ndk-like mechanism of Adk originated from an erroneous interpretation of data, in that contamination of ATP preparations with AMP and ADP was not taken into account. Our results also address the proposed role of Adk in restoring a normal growth rate of mutant strains of E. coli lacking Ndk. These mutant strains apparently, in spite of a mutator phenotype, are able to synthesise nucleoside triphosphates by alternative pathways to maintain the same growth rate as the wildtype. 2005 Elsevier Inc. All rights reserved. Keywords: Adenylate kinase; Nucleoside diphosphate kinase; Adenylate energy charge; Nucleoside triphosphate synthesis; Nucleotide metabolism
Adenylate kinase (Adk)1 catalyses the important phosphorylation of AMP to ADP in the purine nucleotide synthesis pathway according to Eq. (1) ATP þ AMP $ 2ADP
ð1Þ
The equilibrium constant of this reaction is defined in Eq. (2) h 2 K eq ¼ ADP =½ATP½AMP ð2Þ
*
Corresponding author. E-mail address: [email protected] (M. Willemoe¨s). 1 Abbreviations used: Adk, adenylate kinase; DTT, dithiothreitol; Ndk, nucleoside diphosphate kinase; Ap 5 A, P 1 ,P 5 -di(adenosine-5 0 ) pentaphosphate. 0003-9861/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2005.10.003
where Keq is close to unity [1,2] and the reaction will therefore readily proceed in both directions being either ADP- or ATP and AMP synthesis. Adk has two distinct nucleotide binding sites designated by their role in the forward reaction; the ATP binding site, that has a low stringency for the nucleobase moiety, and the AMP binding site, which is highly specific for the adenine base. Both sites can bind ADP in the reverse reaction [3–5]. Purification of a residual nucleoside diphosphate kinase (Ndk) activity from the cell extract of a mutant Escherichia coli strain, where the ndk gene had been inactivated, showed that this activity resided with Adk. From a product profile analysis, it was concluded that Adk had the property to transfer the c-phosphoryl of ATP to all tested purine
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
and pyrimidine deoxy- or ribonucleoside diphosphates. A similar analysis of Adk from rabbit and chicken suggested that this activity was a general property of Adk [6]. Indeed, Ndk-like activity has been shown in the same type of analysis by a separate group, to also be associated with Adk from Mycobacterium tuberculosis [7]. The formulation of the mechanism for the Ndk-like reaction requires that Adk allows for binding of the substrate nucleotide diphosphate to the AMP binding site [6] even though this site is very specific for the nucleotides AMP and ADP, as mentioned above. Another intriguing property of the Ndk-like reaction was the inability of GTP to function as a c-phosphoryl donor as opposed to the phosphorylation of AMP to ADP [6]. Despite the apparent support by data for the suggested mechanism for the Ndk-like reaction of Adk, some puzzling questions regarding this activity remain that were not addressed in the original work. Since the Ndk-like reaction will form ADP, if it proceeds as suggested, i.e., ATP + (d)NDP fi ADP + (d)NTP, this would immediately equilibrate with ATP and AMP by the normal action of Adk. ADP (bound in the AMP site) can then combine on the enzyme with (d)NDP (bound in the low stringency ATP site) to form AMP and (d)NTP via the analogues to the reverse reaction of Adk. The formed AMP can then be converted back to ADP by the normal Adk activity, and the Ndklike activity would not be needed for further NTP synthesis. In the present work, we have re-investigated the synthesis of nucleoside triphosphates from the corresponding diphosphate as catalysed by Adk. We show that the previously suggested mechanism for the Ndk-like activity of Adk is most likely based on an artifact of the assay procedure by which this property of the enzyme was identified and investigated. Experimental procedures Chemicals, including all nucleotides, were purchased from Sigma–Aldrich. Chicken muscle Adk was obtained from Sigma–Aldrich and was dissolved in 50 mM Hepes, pH 7.5, 100 mM KCl, and 2 mM DTT to a concentration of 2000 U/ml. Radiolabelled nucleotides were obtained from Amersham ([5-3H]CTP), Perkin–Elmer ([a-32P]CTP) or NEN ([8-14C]ADP). Polyethyleneimine cellulose sheets for thin layer chromatography were obtained from Merck. All assays were performed at 37 C in 100 ll containing 50 mM Hepes, pH 7.5, 100 mM KCl, 2 mM DTT, and 1 mM MgCl2. Adk was added to a final concentration of 0.4 U/ml. The concentration of nucleotides varied as described under Results. At the given time points, 10 ll of the incubation was transferred to 5 ll of 0.33 M formic acid and immediately spotted on PEI-cellulose sheets. The relevant marker nucleotides were added to the spots and dried. The chromatograms were developed in 1.3 M LiCl and 1 M acetic acid [8]. When 3H- or 14C-labelled nucleotides were used, the relevant spots identified by the
marker nucleotides were cut out and transferred to scintillation vials. Two milliliters of 2 M ammonia was added to each vial to dissolve the nucleotide bound to PEI-cellulose sheet. The vials were left to stand for 1 h when scintillation liquid (EcoScint A, National Diagnostics) was added and the vials shaken to mix scintillation liquid with the water phase. Radioactivity was quantitated on a Packard Tri-carb 2200CA scintillation counter. When 32 P labeled nucleotide was used, the radioactivity in the spots was quantitated directly on a Packard Instant Imager. Results Effect of AMP on the CTP synthesis catalysed by Adk In the experiment shown in Fig. 1, we analysed the rate of CTP synthesis from (commercial analytical grade) ATP and radioactive CDP. As seen, there is a long lag period before CTP synthesis commences when only ATP is added to the incubation. However, at increasing concentrations of added AMP, the reaction is greatly stimulated. This apparent superposition of the two equilibria, involving cytidine and adenine nucleotides, led us to study the influence of CDP on the adenine nucleotide equilibrium of the Adk reaction (Eq. (1)). Influence of CTP synthesis on the equilibrium concentrations of adenine nucleotides in the Adk catalysed reaction When ADP was added as substrate for the Adk reverse reaction (Figs. 2A and D), the conditions were found to result in immediate equilibration of AMP, ADP, and ATP. In Figs. 2B and C, the addition of CDP to the reaction mixture is seen to result in a drastic time dependent displacement of AMP, ADP, and ATP concentrations concomitant with CTP synthesis. However, as shown in Fig. 2D, the adenine nucleotides are in equilibrium throughout the entire time span of the assay.
100 80
CTP, mM
196
60 40 20 0 0
500 1000 Time, s
1500
Fig. 1. AMP dependent synthesis of CTP catalysed by Adk. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 100 lM ATP, 200 lM [5-3H]CDP, and AMP added at (circles) 0 lM, (squares) 10 lM or (triangles) 20 lM.
A 100
C 100
Nucleotides, mM
Nucleotides, mM
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
80 60 40 20
80
80
60
60
40
40
20
20
0
500
0
1000 1500
B 100
D
500
1000 1500
2
Keq
40
500
1000 1500
0
60 120 Time, s
20 0
0 500
0
40
20 0
1000 1500
60
1
0
500
80
80 60
0
B 100
3 CTP, mM
0
Nucleotides, mM
A 100
0
0
197
1000 1500
0
500
20
C
1000 1500
Time, s
15
Fig. 2. Equilibration of adenyl nucleotides during CTP synthesis catalysed by Adk. Assays were performed as described in Experimental procedures. The initial [8-14C]ADP concentration was 100 lM and CTP was added at (A) 0 mM, (B) 0.2 mM, and (C) 1 mM; circles represent AMP, squares represent ADP, and triangles represent ATP. (D) The equilibrium constant, Keq, defined in Eq. (2), is calculated from the data of (A) circles, (B) squares, and (C) triangles.
The time course of CTP synthesis at two different CDP concentrations was determined with ADP as the phosphate donor, under similar conditions as for the experiments in Figs. 2B and C. The results are presented in Fig. 3A, where the rate of CTP synthesis is seen to increase with the concentration of CDP. Fig. 3B shows the results from an experiment where a similar adenine nucleotide concentration to that in Fig. 3A was added as equimolar amounts of ATP and AMP. As shown, the rates of CTP synthesis are identical whether ADP (Fig. 3A) or ATP and AMP (Fig. 3B) were added. This is a clear confirmation of the rapid Adk dependent equilibration of the adenine nucleotides shown in Fig. 2. As a control that we were actually measuring an Adk dependent activity, we determined the rate of CTP synthesis in the absence and presence of P1,P5-di(adenosine-5 0 ) pentaphosphate (Ap5A), the classic bi-substrate analog and potent inhibitor of Adk [9]. From Fig. 3C, it is evident that no CTP synthesis occurs at even 1 lM Ap5A. We conclude that CTP synthesis reported in the present work is derived from an Adk dependent reaction. Synthesis of CDP from CTP by the reverse reaction in the presence of AMP or ADP To investigate the reversibility of the Adk dependent nucleoside triphosphate synthesis, we also measured CDP synthesis from CTP. In Fig. 4, the time dependency on
5 0 180
Fig. 3. CTP synthesis catalysed by Adk. Assays were performed as described in Experimental procedures. The initial adenyl nucleotide concentrations were (A) 100 lM ADP or (B) 50 lM each of ATP and AMP. In both (A and B) circles represent initial 0.2 mM [5-3H]CDP and squares represent initial 1 mM [5-3H]CDP. (C) Ap5A inhibition of CTP synthesis by Adk. The initial substrate concentrations were 0.1 mM ADP and 0.2 mM [5-3H]CDP. Ap5A was present at concentrations of (circles) 0 lM or (squares) 1 lM.
CDP synthesis is shown in the presence of AMP or ADP. From these results, it is clear that AMP by far is best at stimulating the rate of CDP synthesis. CTP synthesis using GTP as a phosphoryl donor In agreement with the previous result of Lu and Inouye [6], it is clear from Fig. 5 that GTP is not a good phosphoryl 50 40
CDP, mM
CTP synthesis from ADP or equimolar amounts of ATP and AMP
10
30 20 10 0 0
500
1000 Time, s
1500
Fig. 4. Reverse reaction of CTP synthesis as catalysed by Adk. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 1 mM [a-32P]CTP and either (circles) 100 lM AMP or (squares) 100 lM ADP.
198
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199 100
CTP, mM
80 60 40 20 0 0
1500
3000
4500
Time, s
Fig. 5. Adk dependent CTP synthesis using GTP as phosphoryl donor. Assays were performed as described in Experimental procedures. The initial nucleotide concentrations were 100 lM GTP and 200 lM [5-3H]CDP. AMP was added to a concentration of (circles) 0 lM or (squares) 10 lM.
donor for CTP synthesis as catalysed by Adk. However, the addition of a low concentration of AMP to the reaction greatly stimulates CTP synthesis, as also observed for the ATP dependent reaction (Fig. 1). Discussion Adk was previously identified as the enzyme that substituted for Ndk in a E. coli ndk strain, and was found to catalyse the transfer of the c-phosphoryl from ATP to deoxy- or ribonucleoside diphosphates [6]. The proposed reaction scheme can be written as follows: ATP þ ðdÞNDP $ ADP þ ðdÞNTP
ð3Þ
However, this mechanism that overall appears to be similar to that of Ndk, except for the preference for ATP as phosphoryl donor, involves the violation of several known restricted properties of Adk as mentioned in the introduction. Although, we were readily able to reproduce the previously reported results [6] (data not shown), our additional work clearly showed that a mechanism for Adk dependent nucleoside triphosphate synthesis as illustrated by Eq. (3) is unlikely. Instead, it appears that the transfer of the c-phosphoryl of ATP to any (d)NDP occurs via the following reactions that violate none of the known catalytic restrictions for Adk and agrees with long known stimulation of the Adk reaction by various nucleoside diphosphates [3]: ATP þ AMP $ 2ADP ADP þ ðdÞNDP $ AMP þ ðdÞNTP
ð4Þ
Because commercial ATP is always contaminated with small amounts of AMP, the reaction will proceed even if AMP is not deliberately added to the incubation, and ADP will be synthesised via the Adk reaction (compare with Eq. (1)). ADP bound in the AMP site can then donate the b-phosphoryl to any (d)NDP bound to the ATP site, which is much more tolerant to the nucleobase moiety of the bound nucleotide. By this mechanism, only trace amounts of AMP are needed initially, since AMP will be recycled and the concentration increased during the time course of the reaction.
The scheme is consistent with the observed acceleration of CTP synthesis in the absence of added AMP and the activation of CTP synthesis by addition of AMP is seen in Fig. 1. Furthermore, Eq. (4) is also in agreement with the observation that AMP by far is the best substrate over ADP for the reverse reaction of CDP synthesis from CTP (Fig. 4). Since AMP and ADP were added at equimolar concentrations in the two experiments in Fig. 4, the result obtained with ADP can be understood in terms of the equilibrium for the Adk reaction. The initial AMP concentration under this condition will be about one-third compared to the situation where only AMP was added. Relevant to the above arguments is our demonstration in Fig. 2 that Adk catalyses the equilibration between ATP, ADP, and AMP at rates greatly exceeding the rate of CTP synthesis (Figs. 3A and B). Finally, we also offer an explanation for the puzzling inability of GTP to act as phosphoryl donor in the Ndk-like reaction of Adk (Fig. 5). Commercial GTP, unlike ATP, is not likely to contain AMP. However, when AMP is added to the reaction, CTP is formed (Fig. 5). The reason for this observation can now be understood on the basis of the mechanism in Eq. (4); GTP can, by replacing ATP, phosphorylate AMP to ADP that in turn will transfer the b-phosphoryl group to CDP yielding CTP as shown in Fig. 5. The reaction shown by Eq. (4), where ADP acts as the phosphoryl donor, is also in agreement with the observed ability of Adk to phosphorylate thiamine diphosphate to its triphosphate [10,11] and further illustrates the relaxed binding specificity in the ATP binding site of Adk. The problem, not realised when the original mechanism for the Ndk-like activity of Adk was forwarded in the work of Lu and Inouye [6], is the role of trace impurities of AMP working as a ‘‘coenzyme’’ for the reaction. The nature of the more trivial mechanism for Adk dependent nucleoside triphosphate synthesis identified in this work addresses important questions as to whether Adk will be solely responsible in vivo for restoring the normal growth-rate of an E. coli strain lacking the Ndk [12]. Adk was both identified as the residual Ndk activity in this mutant E. coli strain and subsequently characterised based on the same erroneous assay. Although, as we show in this work, Adk is able to synthesise nucleoside triphosphates following the scheme outlined in Eq. (4), this activity will in vivo be strictly governed by the intracellular concentration of all participating nucleotides, and most importantly not only ATP but also ADP and AMP. In this respect, the interplay between the Adk catalysed reactions is by far more complicated than the relatively simple equilibration between nucleoside di- and triphosphates carried out by the normal Ndk enzyme. Due to the broad substrate specificity and the ‘‘neutral’’ phosphorylated enzyme intermediate, that can donate a c-phosphoryl from any nucleoside triphosphate to any nucleoside diphosphate [13–15], Ndk can optimally balance deoxy- and ribonucleoside di- and triphophosphate ratios according to the supply and demand of nucleotide metabolism.
M. Willemoe¨s, M. Kilstrup / Archives of Biochemistry and Biophysics 444 (2005) 195–199
In an E. coli strain devoid of Ndk enzyme, all deoxy- or ribonucleoside diphosphates will compete for the ADP pool for synthesis of nucleoside triphosphates. In other words, the synthesis of nucleoside triphosphates in this mutant strain will be governed by the adenylate energy charge [16,17], that is, if Adk is the sole source of residual Ndk-like activity. Therefore, we find it plausible that other reversible metabolic reactions that will yield, or equilibrate, between nucleoside di- and triphosphates, all together, plays a role in sustaining the total synthesis of nucleoside triphosphates [18]. Bearing in mind that cells lacking Ndk activity clearly must have severe changes in the pools of the various nucleotides compared to wild type cells as confirmed by the mutator phenotype of the mutant strains [12]. Acknowledgment This work was supported by the Danish Natural Science Research Council. References [1] I.A. Rose, Proc. Natl. Acad. Sci. USA 61 (1968) 1079–1086. [2] S. Su, P.J. Russell Jr., J. Biol. Chem. 243 (1968) 3826–3833.
199
[3] W.E. Criss, T.K. Pradhan, Methods Enzymol. 51 (1978) 459–467. [4] K.K. Tsuboi, Methods Enzymol. 51 (1978) 467–473. [5] K. Miura, S. Inouye, K. Sakai, H. Takaoka, F. Kishi, M. Tabuchi, T. Tanaka, H. Matsumoto, M. Shirai, T. Nakazawa, A. Nakazawa, J. Biol. Chem. 276 (2001) 13490–13498. [6] Q. Lu, M. Inouye, Proc. Natl. Acad. Sci. USA 93 (1996) 5720–5725. [7] L.S. Meena, P. Chopra, R.S. Bedwal, Y. Singh, Biotechnol. Appl. Biochem. 38 (2003) 169–174. [8] K. Randerath, E. Randerath, J. Chromatogr. 16 (1964) 111–125. [9] G.E. Lienhard, I. Secemski, J. Biol. Chem. 248 (1973) 1121–1123. [10] H. Shikata, Y. Egi, S. Koyama, K. Yamada, T. Kawasaki, Biochem. Int. 18 (1989) 943–949. [11] H. Shikata, S. Koyama, Y. Egi, K. Yamada, T. Kawasaki, Biochem. Int. 18 (1989) 933–941. [12] Q. Lu, X. Zhang, N. Almaula, C.K. Mathews, M. Inouye, J. Mol. Biol. 254 (1995) 337–341. [13] J. Munoz-Dorado, N. Almaula, S. Inouye, M. Inouye, J. Bacteriol. 175 (1993) 1176–1181. [14] J. Munoz-Dorado, S. Inouye, M. Inouye, J. Biol. Chem. 265 (1990) 2707–2712. [15] R.L. Ratliff, R.H. Weaver, H.A. Lardy, S.A. Kuby, J. Biol. Chem. 239 (1964) 301–309. [16] C.C. Glembotski, A.G. Chapman, D.E. Atkinson, J. Bacteriol. 145 (1981) 1374–1385. [17] A.G. Chapman, L. Fall, D.E. Atkinson, J. Bacteriol. 108 (1971) 1072– 1086. [18] A. Kuroda, A. Kornberg, Proc. Natl. Acad. Sci. USA 94 (1997) 439– 442.