- Email: [email protected]
Development of a charcoal paper adenosine triphosphate:pyrophosphate exchange assay: Kinetic characterization of NEDD8 activating enzyme
Analytical Biochemistry 394 (2009) 24–29
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
Development of a charcoal paper adenosine triphosphate:pyrophosphate exchange assay: Kinetic characterization of NEDD8 activating enzyme Frank J. Bruzzese *, Christopher A. Tsu, Jingya Ma, Huay-Keng Loke, Dongyun Wu, Zhi Li, Olga Tayber, Lawrence R. Dick Discovery, Millennium Pharmaceuticals, Cambridge, MA 02139, USA
a r t i c l e
i n f o
Article history: Received 2 April 2009 Available online 12 July 2009 Keywords: APP-BP-1/UBA3 ATP Charcoal paper E1 Inorganic pyrophosphate NEDD8 activating enzyme Ubiquitin-like modifier
a b s t r a c t Ubiquitin activating enzyme (UAE, UBE1, or E1) and seven known homologous ‘‘E1s” initiate the conjugation pathways for ubiquitin and 16 other ubiquitin-like modifiers (ULMs) found in humans. The initial step catalyzed by E1s uses adenosine triphosphate (ATP) to adenylate the C terminus of the appropriate ULM and results in the production of inorganic pyrophosphate (PPi). The mechanism of these enzymes can be studied with assays that measure the rate of ULM-dependent ATP:PPi exchange. The traditional method follows the initial velocity of [32P]PPi incorporation into ATP by capturing the nucleotide on activated charcoal powder to separate it from excess [32P]PPi and then measuring [32P]ATP in a scintillation counter. We have modified the method by using charcoal paper to capture the nucleotide and a phosphorimager to quantify the [32P]ATP. The significant increase in throughput that these modifications provide is accomplished without any sacrifice in sensitivity or accuracy compared with the traditional method. To demonstrate this, we reproduce and extend the characterization of the NEDD8 activating enzyme. Ó 2009 Elsevier Inc. All rights reserved.
The discovery of an energy-dependent ubiquitin activating enzyme [1] (UAE, UBE1, or E1)1 was an important milestone in the elucidation of the ubiquitin proteasome system (UPS) [2] and this system’s role in intracellular protein turnover. Since that time, many other biological roles for the posttranslational modification by ubiquitin have been described [3]. In addition, a family of homologous ubiquitin-like modifiers (ULMs) has been discovered and a stunning diversity of biological regulation is believed to derive from conjugation of these ULMs to their target substrates. The activating enzymes that set these tasks in motion also constitute an evolutionarily related family that shares a relationship of homology and whose molecular mechanisms are similar [4,5]. The approval of the proteasome inhibitor bortezomib for the treatment of relapsed refractory multiple myeloma in 2003 [6], and its recent approval as a frontline treatment for this disease [7], validates the UPS as a target of cancer chemotherapeutics. Our interest in the ULMs and their activating enzymes stems from a desire to identify other points of intervention, both within and
* Corresponding author. Fax: +1 617 551 8905. E-mail address: [email protected] (F.J. Bruzzese). 1 Abbreviations used: UAE (UBE1 or E1), ubiquitin activating enzyme; UPS, ubiquitin proteasome system; ULM, ubiquitin-like modifier; NAE, NEDD8 activating enzyme; ATP, adenosine triphosphate; PPi, inorganic pyrophosphate; AMP, adenosine monophosphate; SPA, scintillation proximity assay; DTT, dithiothreitol; TCA, trichloroacetic acid; SAE, SUMO activating enzyme; UV, ultraviolet; tRNA, transfer RNA. 0003-2697/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2009.07.011
outside of the UPS, that may lead to the discovery of better cancer drugs. NEDD8 is a ULM whose conjugation pathway is initiated by a heterodimeric activating enzyme known as APP-BP-1/UBA3 [8]. We refer to it simply as NAE (NEDD8 activating enzyme). The most well-described biological role for modification by NEDD8 is to regulate the activity of the ‘‘cullin RING” family of ubiquitin ligases [9]. In this way, the NEDD8 conjugation pathway intersects the UPS by regulating the turnover of a specific subset of proteins degraded through the UPS. Recently, we have initiated clinical trials with a first-in-class small molecule inhibitor of NAE for the treatment of cancer [10]. Discovery and development of this experimental drug required a detailed understanding of how it inhibits NAE as well as its selectivity for NAE versus other ULM activating enzymes. In turn, this necessitated the development of direct and generic in vitro assays to study the mechanism of activating enzymes and inhibitors thereof. Activating enzymes form a ULM–adenylate intermediate during the course of their catalytic cycle and, thus, can be studied by assaying adenosine triphosphate (ATP):pyrophosphate (PPi) exchange [1,5,11–22]. ULM activation is generally thought to proceed via the three-step mechanism described for UAE (Scheme 1) [5]. Briefly, in step 1, the E1 enzyme binds one molecule of ATP and one molecule of its partner ULM, catalyzing the adenylation of the ULM C terminus, followed by the release of PPi. In step 2, the E1 catalytic cysteine attacks the ULM–adenylate, forming the
25
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
ULM–thioester and releasing adenosine monophosphate (AMP). Once AMP is released, the thioester form of the E1 catalyzes a second round of ULM adenylation in step 3, resulting in the formation of a ternary complex containing a tightly bound ULM–adenylate in addition to the ULM–thioester. In this assay, one follows the incorporation of radiolabeled PPi into ATP, which in this case occurs by the reversal of steps 1 and 3 (Scheme 1). In the traditional method, the radiolabeled ATP formed during the exchange reaction is separated from the excess radiolabeled PPi by absorption onto activated charcoal. During recent years, several groups have improved on or completely removed the use of charcoal as a means of capturing labeled ATP. Examples of methods to increase throughput, sensitivity, and/or reliability of ATP radiometric assays include the use of 96-well filter plates, column or thin layer chromatography, high-performance liquid chromatography, and streptavidin-coated scintillation proximity assay (SPA) beads [12–22]. We have modified the original method by using charcoal paper to capture the nucleotide and a phosphorimager to quantify the bound [32P]ATP. The significant increase in throughput by incorporation of a 96-well format expands the utility of the assay. More importantly, this new format requires no special instrumentation or equipment beyond that commonly available in most biochemical laboratories, and the advances are realized without sacrificing sensitivity or accuracy as compared with the traditional method. To demonstrate this, we reproduce and extend the characterization of NAE [11].
Materials and methods Reagents [32P]PPi and [a-32P]ATP were obtained from PerkinElmer (Boston, MA, USA). ATP, AMP, PPi, dithiothreitol (DTT), Tris, trichloroacetic acid (TCA), MgCl2 activated charcoal (Fluka Norit A), and bovine erythrocyte ubiquitin were obtained from Sigma (St. Louis, MO, USA). All recombinant E1 and ULM proteins were generated in-house from human gene sequences. For the E1 enzymes, the expression vectors were constructed by subcloning to pFastBac or using Gateway technology into pDEST (Invitrogen, Carlsbad, CA, USA). Baculoviruses were generated with the Bac-to-Bac Expression System (Invitrogen) and expressed in Sf9 insect cells. These include N-terminal His6-tagged UAE (UBE1X), N-terminal His6-tagged UBE1L (codon optimized), N-terminal His6-tagged UBA4, N-terminal His6-tagged UBA5, N-terminal His6-tagged UBA6, and N-terminal His6-tagged ATG7. The NAE and SUMO activating enzyme (SAE) heterodimers were expressed in Sf9 insect cells by coinfection of the a and b subunit baculovirus constructs. NAE: N-terminal His6-tagged NAEa (APP-BP-1) and untagged NAEb (UBA3). SAE: untagged SAEa (AOS1) and N-terminal His6tagged SAEb (UBA2). The ULM proteins NEDD8, SUMO1, SUMO2, ISG15, FAT10, Urm1, Ufm1, MAPLC3A, GABARAP, and GATE-16 were generated untagged or with N-terminal FLAG tags by subcloning into pT7-7 or using the Gateway technology into pDEST vectors and expressed in Escherichia coli. Proteins were purified by affinity (nickel–nitrilotriacetic acid agarose, Qiagen, Valencia, CA, USA) or conventional chromatography using standard buffers. Stock concentrations of E1 enzymes, ULMs, AMP, and ATP were estimated spectrophotometrically based on their calculated molar absorption coefficients [23] using a Cary 50 Bio ultraviolet (UV)– visible spectrophotometer (Varian, Palo Alto, CA, USA).
assay used a Schleicher & Schuell Minifold-I Dot-Blot System (Whatman Cat. No. 10447900) connected to a GAST Manufacturing vacuum pump (model MOA-P109CA, Benton Harbor, MI, USA). Cherenkov counting was performed on a Beckman LS6500 scintillation counter (Fullerton, CA, USA). Thermal equilibration and incubation of the reaction solutions were carried out in 1.5-ml microcentrifuge tubes using an Eppendorf Thermomixer (Westbury, NY, USA) for the charcoal powder method, whereas a VWR convection oven (West Chester, PA, USA) was used for the sealed 96-well plates in the charcoal paper assay. Autoradiographs were analyzed using BAS-MS2025 or BAS-MS2040 image plates and a BAS-2500 phosphorimager with accompanying Image Reader 1.8 and Multi Gauge 3.1 software (Fuji Film Life Sciences, Woodbridge, CT, USA). Standard assay conditions Reactions and component dilutions were performed in assay buffer (50 mM Tris–HCl [pH 7.5], 10 mM MgCl2, and 1 mM DTT). Assays were run in triplicate for 30 min at 37 °C in a final volume of 50 ll assay buffer including 12.5 nM NAE, 1 mM ATP, and 1 mM PPi (containing 50 cpm/pmol [32P]PPi). Reactions were initiated with the addition of 5 lM NEDD8 to a final concentration of 1 lM. Control reactions were carried out omitting NEDD8. All reactions were stopped with 500 ll of 5% (w/v) TCA and 10 mM PPi. (All TCA/PPi stop and wash solutions were made fresh daily and kept at 4 °C until used.) Other E1 enzymes (20 nM) and ULMs (1 lM) were assayed under identical conditions. Variations are noted in the text. Quenched reaction samples were analyzed at room temperature using either charcoal powder or paper methods. Samples analyzed using activated charcoal powder were processed essentially as described by Haas and Rose [5] with the modification that the concentration of PPi in the 2% (w/v) TCA wash solution was increased from 4 to 10 mM. The Cherenkov count values were corrected for the background control counts (measured in the absence of ULM) and converted to picomoles of ATP using the slope of an [a-32P]ATP standard curve constructed as described in Fig. 1. For samples analyzed using activated charcoal paper, the paper was first prepared by soaking in wash solution (2% TCA and 10 mM PPi) for 10 min. Presoaking of the charcoal paper in 10 mM PPi was found to improve spot resolution and reduce background counts by blocking the binding of unreacted [32P]PPi. The charcoal paper was then placed between the filter support and sample well plates of the dot-blot manifold, and the apparatus was assembled. An initial vacuum was applied for several seconds to remove excess solution from the wells. Quenched reactions and controls were transferred to the dot-blot wells using a multichannel pipette. Vacuum was applied to completely draw the samples through the wells and was maintained throughout the washing process. The sample wells were washed in situ with 1 ml of the wash solution, and the charcoal paper was removed from the manifold, placed in a flat-bottom dish, and rinsed with 100 ml of wash solution six times for 5 min with agitation. After washing, the charcoal paper was briefly dried
Step ESH + ATP + Ub 1 AMP-Ub
ESH
AMP-Ub
ESH
+ PPi
ES-Ub + AMP
1
2
Equipment B-Safe activated charcoal paper (Cat. No. 10320163) was obtained from Whatman (Florham Park, NJ, USA). The charcoal paper
ES-Ub + ATP + Ub
AMP-Ub
ES-Ub
+ PPi
3
Scheme 1. Three step mechanism detailing UAE activation of ubiquitin.
26
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
Fig. 1. Construction of an ATP standard curve. First, 500 ll of 1 mM ATP (50 cpm/ pmol [a-32P]ATP) and 1 mM PPi were serially diluted 1:3 in assay buffer. Then 50-ll volumes in triplicate were diluted with 500 ll of 5% TCA and 10 mM PPi. Samples were processed using either the charcoal paper (phosphorimager) (A) or charcoal powder protocols (Cherenkov counting) (B). The data are plotted on a logarithmic scale for illustrative purposes.
before being covered in plastic wrap and exposed to an imaging plate for 1 h. Autoradiographs were visualized and quantified using a phosphorimager. The count values were corrected for the background control counts (measured in the absence of ULM) and converted to picomoles of ATP using the slope of an [a-32P]ATP standard curve constructed as described in Fig. 1. The data for the ATP, PPi, and AMP titrations were plotted as velocity versus substrate concentration. The kinetic constants Vmax, kcat, and Km for ATP and PPi titrations were calculated by nonlinear least squares fit of the plotted data using the Michaelis–Menten equation, whereas K1/2 for AMP was calculated from a fit of the plot to the Hill equation. All fits used the program SigmaPlot (version 10.0, Systat Software, San Jose, CA, USA).
powder shows a substantial loss of sensitivity compared with the paper. The ability to measure low amounts of ATP with the new method is an advantage when working at the low enzyme concentrations that one typically employs in inhibition studies. The dayto-day variations in the slopes of the ATP standard curves were between 5% and 10%, further demonstrating that the charcoal paper detection method is robust and reliable. To further quantify the nucleotide binding capacity of the charcoal paper, we performed a series of measurements in which increasing concentrations of cold ATP were spiked with a fixed amount of [a-32P]ATP, quenched, and blotted using the standard protocol. Based on the loss of [32P]ATP signal upon saturation of the paper, it was determined that the binding capacity was approximately 75,000 pmol (a standard reaction volume of 50 ll containing 1.5 mM ATP [data not shown]). For experiments requiring total nucleotide concentrations greater than 1.5 mM, the volume of the quenched solution blotted onto the paper can be reduced accordingly. This adjustment is not required with charcoal powder because its binding capacity is more than 500,000 pmol per 300 ll of 10% charcoal slurry (data not shown). We also evaluated the potential for signal variability in the charcoal paper assay as a function of load volume. Using a fixed amount of [a-32P]ATP in 50 ll of reaction buffer quenched with 500 ll of 5% TCA and 10 mM PPi, we measured the signal over several increasing load volumes (62, 125, 250, and 500 ll). Counts of bound [32P]ATP were a linear function of the volume of quench solution added to the charcoal paper (data not shown). This suggests that there is no loss of binding efficiency in loading the larger volumes. We next measured the sensitivity and linearity of enzymatic [32P]ATP generation for the powder and paper assay protocols as a function of NAE concentration (Fig. 2). The data showed a high correlation between 2.5 and 40 nM NAE (Fig. 2) with r2 values of 1.0 for both methods. Analysis of the lines yielded slopes that are in excellent agreement with one another, namely 0.39 ± 0.014 and 0.37 ± 0.02 pmol of ATP/min/nM NAE for the powder and paper methods, respectively. The charcoal paper and powder methods were in good agreement with one another. The significant increase in throughput afforded by the use of charcoal paper and phosphorimaging was accomplished without any sacrifice in sensitivity or accuracy compared with the traditional method. The only significant drawback of the paper method is the lower ATP binding capacity of the charcoal. However, this can easily be taken into account. Having
Results Demonstration of signal linearity To establish a measure of comparison between the two detection methods, we performed two experiments. First, a standard curve was constructed using [a-32P]ATP. Based on previously reported enzyme activity, the standard curve reflected the range of ATP amounts generated under the assay conditions used [11]. As shown in Fig. 1, the standard curves display excellent linearity between 7.5 and 16,000 pmol of ATP by both scintillation counting and autoradiographic analysis. Dissection of the ATP standard curves reveals that the charcoal paper method is linear from 2.5 to 16,000 pmol of ATP and deviates from linearity only at ATP amounts approaching 50,000 pmol. At 50,000 pmol of ATP, the raw image spot is ‘‘saturated” but can be easily corrected for by decreasing exposure times. Conversely, the charcoal powder ATP standard curve reveals excellent linearity between 7.5 and 50,000 pmol of ATP. However, at low amounts of ATP, the charcoal
Fig. 2. NAE titration. The concentration of NAE was varied from 2.5 to 40 nM in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi). The reaction was initiated by the addition of NEDD8 (1 lM final). The reaction time was 20 min at 37 °C. The plots of the converted data using the charcoal powder (closed circles) and charcoal paper (open circles) protocols are shown.
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
27
Fig. 3. ATP and PPi Km determination. ATP titration: ATP was serially diluted in assay buffer containing 1 mM PPi (50 cpm/pmol [32P]PPi) and 12.5 nM NAE. PPi titration: PPi (50 cpm/pmol [32P]PPi) was serially diluted in assay buffer containing 1 mM ATP and 12.5 nM NAE. Both reactions were initiated with 1 lM NEDD8 (final). The reaction time was 30 min at 37 °C. (A and C) Raw phosphorimages. (B and D) Converted data plotted as a function of ATP (B) and PPi (D) concentrations. Kinetic constants were derived from nonlinear regression fits.
AMP dependence of ATP:PPi exchange
Table 1 Substrate kinetic constants obtained by Michaelis–Menten analysis. 1
Substrate
Vmax (pmol/min)
kcat (s
)
ATP PPi
18 ± 0.4 20 ± 0.1
0.49 ± 0.01 0.54 ± 0.004
Km (lM) 175 ± 14 11 ± 0.4
Haas and Rose [5] found that AMP had an unusual effect on the ATP:PPi exchange rate for UAE. In particular, they observed that AMP affected the ATP:PPi exchange reaction within a limited range of concentrations and was without effect above or below this region. They surmised that this ‘‘titration” effect reflected a shift be-
validated the charcoal paper method, we proceeded with the characterization of the NAE kinetic mechanism. Titration of ATP and PPi The rate of ATP:PPi exchange as a function of ATP concentration is shown in Fig. 3A and B. In the absence of NEDD8, the reaction is negligible, verifying the ULM dependence of ATP:PPi exchange. At 1 lM NEDD8 and 1 mM PPi, the rate of ATP:PPi exchange displayed hyperbolic kinetics over the range of 4–2000 lM ATP (Fig. 3B). The kinetic parameters Vmax, kcat, and Km were 18 ± 0.4 pmol/min, 0.49 ± 0.01 s 1, and 175 ± 14 lM, respectively (Table 1). The rate of ATP:PPi exchange as a function of PPi concentration is shown in Fig. 3C and D. At 1 lM NEDD8 and 1 mM ATP, the rate of ATP:PPi exchange displayed hyperbolic kinetics over the range of 2– 1000 lM PPi. Nonlinear regression analysis yielded values of 20 ± 0.1 pmol/min, 0.54 ± 0.004 s 1, and 11 ± 0.4 lM for Vmax, kcat, and Km, respectively (Table 1). The kcat value of approximately 0.5 s 1, as measured for both the ATP and PPi titrations, was in good agreement with that observed previously (1 s 1) for the NEDD8 titration of NAE [11].
Fig. 4. AMP K1/2 determination. AMP was serially diluted in assay buffer containing 1 mM ATP, 100 lM PPi (50 cpm/pmol [32P]PPi), and 12.5 nM NAE. The reaction was initiated with NEDD8 (1 lM final). The reaction time was 30 min at 37 °C. Converted data are plotted as a function of AMP concentration.
28
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
tween two parallel ATP:PPi exchange pathways represented by steps 1 and 3 of Scheme 1. That AMP binding directly influenced the equilibrium position of step 2 linking steps 1 and 3 was an important observation because it provided evidence for their three-step model and helped to explain the order of substrate binding for UAE. Because no measurement has been reported previously for the AMP dependence of NAE, we decided to test the hypothesis that NAE behaved in a mechanistically similar fashion to UAE. The titration profile is shown in Fig. 4. AMP affects the NAE ATP:PPi exchange reaction within a limited concentration range and is without effect above or below this region, an observation that is consistent with the proposed mechanism for UAE [5]. A simple fit of the data yields a K1/2 of 28 lM, the inflection point defining the concentration of mononucleotide required for halfmaximal change in rate [5]. This result clearly suggests that NAE follows the same kinetic mechanism as UAE. NEDD8 dependence of ATP:PPi exchange We next examined the rate of the ATP:PPi exchange reaction as a function of NEDD8 concentration (Fig. 5). The increase in rate is hyperbolic to a maximum value of 28 pmol/min at approximately 1.25 lM NEDD8. Above 1.25 lM, the rate decreases to 6 pmol/ min at 40 lM NEDD8. This observation is consistent with that reported previously for NAE as being a formally random addition mechanism but with the preferential binding of ATP as the leading substrate and NEDD8 as the trailing substrate [11]. Previously, Haas and Rose had shown that UAE followed a strictly ordered mechanism of binding [5]. In that case, as the concentration of ubiquitin was increased above 1–2 lM, the rate of the reaction approached zero. Use of ATP:PPi exchange for ULM specificity and other E1 activating enzymes Recently, Groettrup and coworkers reviewed the eight known human E1 enzymes and their corresponding ULMs [4]. Here we addressed the diversity of this family in two experiments. First, we investigated the specificity of NAE activity across a panel of ULMs. Second, we measured the initial rates for all of the known E1 enzymes and their corresponding ULMs in the ATP:PPi exchange assay—NAE:NEDD8; SAE:SUMO1; UAE:ubiquitin; ATG7:MAPLC3A; UBA4:Urm1; UBA5:Ufm1; UBA6:ubiquitin; and UBE1L:ISG15—under identical conditions. The results are shown in Fig. 6A and B,
Fig. 5. NEDD8 dependence of ATP:PPi exchange. NEDD8 was serially diluted in assay buffer. The reactions were initiated by adding the diluted NEDD8 at the stated final concentrations to 12.5 nM NAE in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi). The reaction time was 30 min at 37 °C. Converted data are plotted as a function of NEDD8 concentration.
Fig. 6. (A) ULM specificity of NAE. The reactions of 20 nM NAE in assay buffer containing 1 mM ATP, 1 mM PPi, and 50 cpm/pmol [32P]PPi were initiated with 1 lM (final) of the corresponding ULM. The reaction time was 30 min at 37 °C. (B) Activity of E1:ULM pairs. The reactions of 20-nM E1 enzymes in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi) were initiated with 1 lM (final) of the corresponding ULM. The reaction time was 30 min at 37 °C.
respectively. NAE did not support robust PPi exchange with any of the ULMs tested except NEDD8. NEDD8’s closest related ULM, ubiquitin, was 25 times less active—30 pmol/min for NEDD8 versus 1.2 pmol/min for ubiquitin. Although clearly selective for NEDD8, we also examined the potential for noncognate ULM binding to NAE, thereby inhibiting NEDD8 activation. To test this hypothesis, we conducted a parallel experiment to that described above where 1 lM NEDD8 was added to each of the ULMs tested. Analysis of the data revealed no inhibition of PPi exchange (data not shown). Accordingly, noncognate ULMs do not bind to NAE. Conversely, the data shown in Fig. 6B reveal that the PPi exchange assay is amenable to studying other E1 enzymes and their cognate ULMs. (Although several of the E1 enzymes are known to activate more than one ULM, only those listed were assayed.) The eight known human E1 enzymes were tested. All showed at least fivefold activity versus the ULM control, ranging from 0.06 to 30 pmol/min. That E1 enzymes undergo PPi exchange in the adenylation of their corresponding ULM is consistent with their function. These results demonstrate the general utility of the charcoal paper assay for measuring the ULM specificity and activity of various E1 enzymes, and by extension other enzymes, that catalyze ATP:PPi exchange. Conclusion Although many alternative assays have been proposed to measure ATP:PPi exchange, the traditional method using powder char-
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
coal has remained the standard. We have shown that the charcoal paper assay described here represents a significant improvement over the previously published method incorporating powder charcoal absorption. The paper assay affords increased sample throughput and sensitivity while maintaining the simplicity of the powder assay. The results demonstrate the utility of this technique in determining kinetic constants associated with enzymes undergoing ATP:PPi exchange. We showed that the data generated with the charcoal paper assay compare favorably with the historical data for NAE. Moreover, we showed that NAE binds to and is affected by AMP in a manner reminiscent of UAE and, therefore, is consistent with the proposed E1 activating mechanism. These findings demonstrate that classic experiments to resolve substrate binding and catalysis, specificity, order of substrate binding, inhibitor mechanism of action, and the like can now be readily and routinely answered for enzymes undergoing ATP:PPi exchange, placing the charcoal paper assay into the enzymologist’s tool box of everyday experimental protocols. In addition, although the eight E1 enzymes are known to possess a diverse set of biological activities and functions, we showed that all share formation of a similar ULM–adenylate intermediate and can be readily studied by the ATP:PPi exchange assay. The premise of these results argues for applications to other enzymes that exhibit ATP:PPi exchange reactions such as aminoacyl transfer RNA (tRNA) synthetases [18,22]. The potential for using the charcoal paper assay to study other enzymes within this class with minimal modifications, if any, makes this assay a very attractive alternative to the more sophisticated techniques that may require custom substrate labeling, overtly expensive reagents, equipment, and the like to move between studying different enzymes. The advantages of the new method are evident—ease of use, increased sampling, decreased processing time, excellent signal-to-background ratio, signal linearity, versatility, and decreased cost. With the added advantage of the 96-well plate format, automation of the charcoal paper assay is possible. Our approach in the design of the charcoal paper ATP:PPi exchange assay has been one of utility for mechanistic analysis, substrate profiling, and inhibitor characterization. Acknowledgments The authors thank Arthur Haas for his insightful and critical review of the data and methods described in this article. We also thank Michael Bembenek for his assistance during the early developmental stages of the assay, and we acknowledge James Brownell, Mark Rolfe, and William Mallender for their comments on the manuscript. References [1] A. Ciechanover, H. Heller, R. Katz-Etzion, A. Hershko, Activation of the heatstable polypeptide of the ATP-dependent proteolytic system, Proc. Natl. Acad. Sci. USA 78 (1981) 761–765.
29
[2] A. Hershko, A. Ciechanover, The ubiquitin system, Annu. Rev. Biochem. 67 (1998) 425–479. [3] C.M. Pickart, Ubiquitin enters the new millennium, Mol. Cell 8 (2001) 499–504. [4] M. Groettrup, C. Pelzer, G. Schmidtke, K. Hofmann, Activating the ubiquitin family: UBA6 challenges the field, Trends Biochem. Sci. 33 (2008) 230–237. [5] A.L. Haas, I.A. Rose, The mechanism of ubiquitin activating enzyme: a kinetic and equilibrium analysis, J. Biol. Chem. 257 (1982) 10329–10337. [6] R.C. Kane, P.F. Bross, A.T. Farrell, R. Pazdur, VELCADE: U.S. FDA approval for the treatment of multiple myeloma progressing on prior therapy, Oncologist 8 (2003) 508–513. [7] J.F. San Miguel, R. Schlag, N.K. Khuageva, M.A. Dimopoulos, O. Shpilberg, M. Kropff, I. Spicka, M.T. Petrucci, A. Palumbo, O.S. Samoilova, A. Dmoszynska, K.M. Abdulkadyrov, R. Schots, B. Jiang, M.V. Mateos, K.C. Anderson, D.L. Esseltine, K. Liu, A. Cakana, H. Van de Velde, P.G. Richardson, VISTA Trial Investigators, bortezomib plus melphalan and prednisone for initial treatment of multiple myeloma, N. Engl. J. Med. 359 (2008) 906–917. [8] L. Gong, E.T. Yeh, Identification of the activating and conjugating enzymes of the NEDD8 conjugation pathway, J. Biol. Chem. 274 (1999) 12036–12042. [9] Z.Q. Pan, A. Kentsis, D.C. Dias, K. Yamoah, K. Wu, NEDD8 on cullin: building an expressway to protein destruction, Oncogene 23 (2004) 1985–1997. [10] T.A. Soucy, P.G. Smith, M.A. Milhollen, A.J. Berger, J.M. Gavin, S. Adhikari, J.E. Brownell, K.E. Burke, D.P. Cardin, C.A. Cullis, S. Critchley, A. Doucette, J.J. Garnsey, J.L. Gaulin, R.E. Gershman, A.R. Lublinsky, A. McDonald, H. Mizutani, U. Narayanan, E.J. Olhava, S. Peluso, M. Rezaei, M.D. Sintchak, T. Talreja, T. Traore, S. Vyskocil, G.S. Weatherhead, J. Yu, J. Zhang, L.R. Dick, C.F. Claiborne, M. Rolfe, J.B. Bolen, S.P. Langston, An inhibitor of NEDD8-activating enzyme as a novel approach to treat cancer, Nature 458 (2009) 732–736. [11] R.N. Bohnsack, A.L. Haas, Conservation in the mechanism of NEDD8 activation by the human AppBp1–Uba3 heterodimer, J. Biol. Chem. 278 (2003) 26823– 26830. [12] A.R. Fersht, R.S. Mulvey, G.L.E. Koch, Ligand binding and enzymatic catalysis coupled through subunits in tyrosyl–tRNA synthetase, Biochemistry 14 (1975) 13–18. [13] M.M. Madden, W. Song, P.G. Martell, Y. Ren, J. Feng, Q. Lin, Substrate properties of ubiquitin carboxyl-terminally derived peptide probes for protein ubiquitination, Biochemistry 47 (2008) 3636–3644. [14] K. Beebe, W. Waas, Z. Druzina, M. Guo, P. Schimmel, A universal plate format for increased throughput of assays that monitor aminoacyl transfer RNA synthetase activities, Anal. Biochem. 368 (2007) 111–121. [15] I. Dusha, Z.G. Dénes, A simplified assay of enzymes catalyzing ATP– pyrophosphate exchange reactions, Anal. Biochem. 81 (1977) 247–250. [16] R. Alvarez, D.V. Daniels, A single column method for the assay of adenylate cyclase, Anal. Biochem. 187 (1990) 93–108. [17] L.G. Otten, M.L. Schaffer, B.R.M. Villiers, T. Stachelhaus, F. Hollfelder, An optimized ATP/PPi-exchange assay in 96-well format for screening of adenylation domains for applications in combinatorial biosynthesis, Biotechnol. J. 2 (2007) 232–240. [18] R. Jørgensen, T.M.M. Søgaard, A.B. Rossing, P.M. Martensen, J. Justesen, Identification and characterization of human mitochondrial tryptophanyl– tRNA synthetase, J. Biol. Chem. 275 (2000) 16820–16826. [19] D.L. Purich, R.D. Allison, Isotope exchange methods for eludiating enzymatic catalysis, in: D.L. Purich (Ed.), Contemporary Enzyme Kinetics and Mechanism, Academic Press, San Diego, 1996, pp. 335–378 (Chapter 12). [20] K.T. Hughs, Radiometric assays, in: R. Eisenthal, M.J. Danson (Eds.), Enzyme Assays, Oxford University Press, Oxford, UK, 2002, pp. 79–102 (Chapter 3). [21] S.E.H. Syed, High performance liquid chromatographic analysis, in: R. Eisenthal, M.J. Danson (Eds.), Enzyme Assays, Oxford University Press, Oxford, UK, 2002, pp. 103–140 (Chapter 4). [22] T.N.C. Wells, J.W. Knill-Jones, T.E. Gray, A.R. Fersht, Kinetic and thermodynamic properties of wild-type and engineered mutants of tyrosyl– tRNA synthetase analyzed by pyrophosphate-exchange kinetics, Biochemistry 30 (1991) 5151–5156. [23] S.C. Gill, P.H. von Hippel, Calculation of protein extinction coefficients from amino acid sequence data, Anal. Biochem. 182 (1989) 319–326.
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
Development of a charcoal paper adenosine triphosphate:pyrophosphate exchange assay: Kinetic characterization of NEDD8 activating enzyme Frank J. Bruzzese *, Christopher A. Tsu, Jingya Ma, Huay-Keng Loke, Dongyun Wu, Zhi Li, Olga Tayber, Lawrence R. Dick Discovery, Millennium Pharmaceuticals, Cambridge, MA 02139, USA
a r t i c l e
i n f o
Article history: Received 2 April 2009 Available online 12 July 2009 Keywords: APP-BP-1/UBA3 ATP Charcoal paper E1 Inorganic pyrophosphate NEDD8 activating enzyme Ubiquitin-like modifier
a b s t r a c t Ubiquitin activating enzyme (UAE, UBE1, or E1) and seven known homologous ‘‘E1s” initiate the conjugation pathways for ubiquitin and 16 other ubiquitin-like modifiers (ULMs) found in humans. The initial step catalyzed by E1s uses adenosine triphosphate (ATP) to adenylate the C terminus of the appropriate ULM and results in the production of inorganic pyrophosphate (PPi). The mechanism of these enzymes can be studied with assays that measure the rate of ULM-dependent ATP:PPi exchange. The traditional method follows the initial velocity of [32P]PPi incorporation into ATP by capturing the nucleotide on activated charcoal powder to separate it from excess [32P]PPi and then measuring [32P]ATP in a scintillation counter. We have modified the method by using charcoal paper to capture the nucleotide and a phosphorimager to quantify the [32P]ATP. The significant increase in throughput that these modifications provide is accomplished without any sacrifice in sensitivity or accuracy compared with the traditional method. To demonstrate this, we reproduce and extend the characterization of the NEDD8 activating enzyme. Ó 2009 Elsevier Inc. All rights reserved.
The discovery of an energy-dependent ubiquitin activating enzyme [1] (UAE, UBE1, or E1)1 was an important milestone in the elucidation of the ubiquitin proteasome system (UPS) [2] and this system’s role in intracellular protein turnover. Since that time, many other biological roles for the posttranslational modification by ubiquitin have been described [3]. In addition, a family of homologous ubiquitin-like modifiers (ULMs) has been discovered and a stunning diversity of biological regulation is believed to derive from conjugation of these ULMs to their target substrates. The activating enzymes that set these tasks in motion also constitute an evolutionarily related family that shares a relationship of homology and whose molecular mechanisms are similar [4,5]. The approval of the proteasome inhibitor bortezomib for the treatment of relapsed refractory multiple myeloma in 2003 [6], and its recent approval as a frontline treatment for this disease [7], validates the UPS as a target of cancer chemotherapeutics. Our interest in the ULMs and their activating enzymes stems from a desire to identify other points of intervention, both within and
* Corresponding author. Fax: +1 617 551 8905. E-mail address: [email protected] (F.J. Bruzzese). 1 Abbreviations used: UAE (UBE1 or E1), ubiquitin activating enzyme; UPS, ubiquitin proteasome system; ULM, ubiquitin-like modifier; NAE, NEDD8 activating enzyme; ATP, adenosine triphosphate; PPi, inorganic pyrophosphate; AMP, adenosine monophosphate; SPA, scintillation proximity assay; DTT, dithiothreitol; TCA, trichloroacetic acid; SAE, SUMO activating enzyme; UV, ultraviolet; tRNA, transfer RNA. 0003-2697/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2009.07.011
outside of the UPS, that may lead to the discovery of better cancer drugs. NEDD8 is a ULM whose conjugation pathway is initiated by a heterodimeric activating enzyme known as APP-BP-1/UBA3 [8]. We refer to it simply as NAE (NEDD8 activating enzyme). The most well-described biological role for modification by NEDD8 is to regulate the activity of the ‘‘cullin RING” family of ubiquitin ligases [9]. In this way, the NEDD8 conjugation pathway intersects the UPS by regulating the turnover of a specific subset of proteins degraded through the UPS. Recently, we have initiated clinical trials with a first-in-class small molecule inhibitor of NAE for the treatment of cancer [10]. Discovery and development of this experimental drug required a detailed understanding of how it inhibits NAE as well as its selectivity for NAE versus other ULM activating enzymes. In turn, this necessitated the development of direct and generic in vitro assays to study the mechanism of activating enzymes and inhibitors thereof. Activating enzymes form a ULM–adenylate intermediate during the course of their catalytic cycle and, thus, can be studied by assaying adenosine triphosphate (ATP):pyrophosphate (PPi) exchange [1,5,11–22]. ULM activation is generally thought to proceed via the three-step mechanism described for UAE (Scheme 1) [5]. Briefly, in step 1, the E1 enzyme binds one molecule of ATP and one molecule of its partner ULM, catalyzing the adenylation of the ULM C terminus, followed by the release of PPi. In step 2, the E1 catalytic cysteine attacks the ULM–adenylate, forming the
25
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
ULM–thioester and releasing adenosine monophosphate (AMP). Once AMP is released, the thioester form of the E1 catalyzes a second round of ULM adenylation in step 3, resulting in the formation of a ternary complex containing a tightly bound ULM–adenylate in addition to the ULM–thioester. In this assay, one follows the incorporation of radiolabeled PPi into ATP, which in this case occurs by the reversal of steps 1 and 3 (Scheme 1). In the traditional method, the radiolabeled ATP formed during the exchange reaction is separated from the excess radiolabeled PPi by absorption onto activated charcoal. During recent years, several groups have improved on or completely removed the use of charcoal as a means of capturing labeled ATP. Examples of methods to increase throughput, sensitivity, and/or reliability of ATP radiometric assays include the use of 96-well filter plates, column or thin layer chromatography, high-performance liquid chromatography, and streptavidin-coated scintillation proximity assay (SPA) beads [12–22]. We have modified the original method by using charcoal paper to capture the nucleotide and a phosphorimager to quantify the bound [32P]ATP. The significant increase in throughput by incorporation of a 96-well format expands the utility of the assay. More importantly, this new format requires no special instrumentation or equipment beyond that commonly available in most biochemical laboratories, and the advances are realized without sacrificing sensitivity or accuracy as compared with the traditional method. To demonstrate this, we reproduce and extend the characterization of NAE [11].
Materials and methods Reagents [32P]PPi and [a-32P]ATP were obtained from PerkinElmer (Boston, MA, USA). ATP, AMP, PPi, dithiothreitol (DTT), Tris, trichloroacetic acid (TCA), MgCl2 activated charcoal (Fluka Norit A), and bovine erythrocyte ubiquitin were obtained from Sigma (St. Louis, MO, USA). All recombinant E1 and ULM proteins were generated in-house from human gene sequences. For the E1 enzymes, the expression vectors were constructed by subcloning to pFastBac or using Gateway technology into pDEST (Invitrogen, Carlsbad, CA, USA). Baculoviruses were generated with the Bac-to-Bac Expression System (Invitrogen) and expressed in Sf9 insect cells. These include N-terminal His6-tagged UAE (UBE1X), N-terminal His6-tagged UBE1L (codon optimized), N-terminal His6-tagged UBA4, N-terminal His6-tagged UBA5, N-terminal His6-tagged UBA6, and N-terminal His6-tagged ATG7. The NAE and SUMO activating enzyme (SAE) heterodimers were expressed in Sf9 insect cells by coinfection of the a and b subunit baculovirus constructs. NAE: N-terminal His6-tagged NAEa (APP-BP-1) and untagged NAEb (UBA3). SAE: untagged SAEa (AOS1) and N-terminal His6tagged SAEb (UBA2). The ULM proteins NEDD8, SUMO1, SUMO2, ISG15, FAT10, Urm1, Ufm1, MAPLC3A, GABARAP, and GATE-16 were generated untagged or with N-terminal FLAG tags by subcloning into pT7-7 or using the Gateway technology into pDEST vectors and expressed in Escherichia coli. Proteins were purified by affinity (nickel–nitrilotriacetic acid agarose, Qiagen, Valencia, CA, USA) or conventional chromatography using standard buffers. Stock concentrations of E1 enzymes, ULMs, AMP, and ATP were estimated spectrophotometrically based on their calculated molar absorption coefficients [23] using a Cary 50 Bio ultraviolet (UV)– visible spectrophotometer (Varian, Palo Alto, CA, USA).
assay used a Schleicher & Schuell Minifold-I Dot-Blot System (Whatman Cat. No. 10447900) connected to a GAST Manufacturing vacuum pump (model MOA-P109CA, Benton Harbor, MI, USA). Cherenkov counting was performed on a Beckman LS6500 scintillation counter (Fullerton, CA, USA). Thermal equilibration and incubation of the reaction solutions were carried out in 1.5-ml microcentrifuge tubes using an Eppendorf Thermomixer (Westbury, NY, USA) for the charcoal powder method, whereas a VWR convection oven (West Chester, PA, USA) was used for the sealed 96-well plates in the charcoal paper assay. Autoradiographs were analyzed using BAS-MS2025 or BAS-MS2040 image plates and a BAS-2500 phosphorimager with accompanying Image Reader 1.8 and Multi Gauge 3.1 software (Fuji Film Life Sciences, Woodbridge, CT, USA). Standard assay conditions Reactions and component dilutions were performed in assay buffer (50 mM Tris–HCl [pH 7.5], 10 mM MgCl2, and 1 mM DTT). Assays were run in triplicate for 30 min at 37 °C in a final volume of 50 ll assay buffer including 12.5 nM NAE, 1 mM ATP, and 1 mM PPi (containing 50 cpm/pmol [32P]PPi). Reactions were initiated with the addition of 5 lM NEDD8 to a final concentration of 1 lM. Control reactions were carried out omitting NEDD8. All reactions were stopped with 500 ll of 5% (w/v) TCA and 10 mM PPi. (All TCA/PPi stop and wash solutions were made fresh daily and kept at 4 °C until used.) Other E1 enzymes (20 nM) and ULMs (1 lM) were assayed under identical conditions. Variations are noted in the text. Quenched reaction samples were analyzed at room temperature using either charcoal powder or paper methods. Samples analyzed using activated charcoal powder were processed essentially as described by Haas and Rose [5] with the modification that the concentration of PPi in the 2% (w/v) TCA wash solution was increased from 4 to 10 mM. The Cherenkov count values were corrected for the background control counts (measured in the absence of ULM) and converted to picomoles of ATP using the slope of an [a-32P]ATP standard curve constructed as described in Fig. 1. For samples analyzed using activated charcoal paper, the paper was first prepared by soaking in wash solution (2% TCA and 10 mM PPi) for 10 min. Presoaking of the charcoal paper in 10 mM PPi was found to improve spot resolution and reduce background counts by blocking the binding of unreacted [32P]PPi. The charcoal paper was then placed between the filter support and sample well plates of the dot-blot manifold, and the apparatus was assembled. An initial vacuum was applied for several seconds to remove excess solution from the wells. Quenched reactions and controls were transferred to the dot-blot wells using a multichannel pipette. Vacuum was applied to completely draw the samples through the wells and was maintained throughout the washing process. The sample wells were washed in situ with 1 ml of the wash solution, and the charcoal paper was removed from the manifold, placed in a flat-bottom dish, and rinsed with 100 ml of wash solution six times for 5 min with agitation. After washing, the charcoal paper was briefly dried
Step ESH + ATP + Ub 1 AMP-Ub
ESH
AMP-Ub
ESH
+ PPi
ES-Ub + AMP
1
2
Equipment B-Safe activated charcoal paper (Cat. No. 10320163) was obtained from Whatman (Florham Park, NJ, USA). The charcoal paper
ES-Ub + ATP + Ub
AMP-Ub
ES-Ub
+ PPi
3
Scheme 1. Three step mechanism detailing UAE activation of ubiquitin.
26
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
Fig. 1. Construction of an ATP standard curve. First, 500 ll of 1 mM ATP (50 cpm/ pmol [a-32P]ATP) and 1 mM PPi were serially diluted 1:3 in assay buffer. Then 50-ll volumes in triplicate were diluted with 500 ll of 5% TCA and 10 mM PPi. Samples were processed using either the charcoal paper (phosphorimager) (A) or charcoal powder protocols (Cherenkov counting) (B). The data are plotted on a logarithmic scale for illustrative purposes.
before being covered in plastic wrap and exposed to an imaging plate for 1 h. Autoradiographs were visualized and quantified using a phosphorimager. The count values were corrected for the background control counts (measured in the absence of ULM) and converted to picomoles of ATP using the slope of an [a-32P]ATP standard curve constructed as described in Fig. 1. The data for the ATP, PPi, and AMP titrations were plotted as velocity versus substrate concentration. The kinetic constants Vmax, kcat, and Km for ATP and PPi titrations were calculated by nonlinear least squares fit of the plotted data using the Michaelis–Menten equation, whereas K1/2 for AMP was calculated from a fit of the plot to the Hill equation. All fits used the program SigmaPlot (version 10.0, Systat Software, San Jose, CA, USA).
powder shows a substantial loss of sensitivity compared with the paper. The ability to measure low amounts of ATP with the new method is an advantage when working at the low enzyme concentrations that one typically employs in inhibition studies. The dayto-day variations in the slopes of the ATP standard curves were between 5% and 10%, further demonstrating that the charcoal paper detection method is robust and reliable. To further quantify the nucleotide binding capacity of the charcoal paper, we performed a series of measurements in which increasing concentrations of cold ATP were spiked with a fixed amount of [a-32P]ATP, quenched, and blotted using the standard protocol. Based on the loss of [32P]ATP signal upon saturation of the paper, it was determined that the binding capacity was approximately 75,000 pmol (a standard reaction volume of 50 ll containing 1.5 mM ATP [data not shown]). For experiments requiring total nucleotide concentrations greater than 1.5 mM, the volume of the quenched solution blotted onto the paper can be reduced accordingly. This adjustment is not required with charcoal powder because its binding capacity is more than 500,000 pmol per 300 ll of 10% charcoal slurry (data not shown). We also evaluated the potential for signal variability in the charcoal paper assay as a function of load volume. Using a fixed amount of [a-32P]ATP in 50 ll of reaction buffer quenched with 500 ll of 5% TCA and 10 mM PPi, we measured the signal over several increasing load volumes (62, 125, 250, and 500 ll). Counts of bound [32P]ATP were a linear function of the volume of quench solution added to the charcoal paper (data not shown). This suggests that there is no loss of binding efficiency in loading the larger volumes. We next measured the sensitivity and linearity of enzymatic [32P]ATP generation for the powder and paper assay protocols as a function of NAE concentration (Fig. 2). The data showed a high correlation between 2.5 and 40 nM NAE (Fig. 2) with r2 values of 1.0 for both methods. Analysis of the lines yielded slopes that are in excellent agreement with one another, namely 0.39 ± 0.014 and 0.37 ± 0.02 pmol of ATP/min/nM NAE for the powder and paper methods, respectively. The charcoal paper and powder methods were in good agreement with one another. The significant increase in throughput afforded by the use of charcoal paper and phosphorimaging was accomplished without any sacrifice in sensitivity or accuracy compared with the traditional method. The only significant drawback of the paper method is the lower ATP binding capacity of the charcoal. However, this can easily be taken into account. Having
Results Demonstration of signal linearity To establish a measure of comparison between the two detection methods, we performed two experiments. First, a standard curve was constructed using [a-32P]ATP. Based on previously reported enzyme activity, the standard curve reflected the range of ATP amounts generated under the assay conditions used [11]. As shown in Fig. 1, the standard curves display excellent linearity between 7.5 and 16,000 pmol of ATP by both scintillation counting and autoradiographic analysis. Dissection of the ATP standard curves reveals that the charcoal paper method is linear from 2.5 to 16,000 pmol of ATP and deviates from linearity only at ATP amounts approaching 50,000 pmol. At 50,000 pmol of ATP, the raw image spot is ‘‘saturated” but can be easily corrected for by decreasing exposure times. Conversely, the charcoal powder ATP standard curve reveals excellent linearity between 7.5 and 50,000 pmol of ATP. However, at low amounts of ATP, the charcoal
Fig. 2. NAE titration. The concentration of NAE was varied from 2.5 to 40 nM in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi). The reaction was initiated by the addition of NEDD8 (1 lM final). The reaction time was 20 min at 37 °C. The plots of the converted data using the charcoal powder (closed circles) and charcoal paper (open circles) protocols are shown.
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
27
Fig. 3. ATP and PPi Km determination. ATP titration: ATP was serially diluted in assay buffer containing 1 mM PPi (50 cpm/pmol [32P]PPi) and 12.5 nM NAE. PPi titration: PPi (50 cpm/pmol [32P]PPi) was serially diluted in assay buffer containing 1 mM ATP and 12.5 nM NAE. Both reactions were initiated with 1 lM NEDD8 (final). The reaction time was 30 min at 37 °C. (A and C) Raw phosphorimages. (B and D) Converted data plotted as a function of ATP (B) and PPi (D) concentrations. Kinetic constants were derived from nonlinear regression fits.
AMP dependence of ATP:PPi exchange
Table 1 Substrate kinetic constants obtained by Michaelis–Menten analysis. 1
Substrate
Vmax (pmol/min)
kcat (s
)
ATP PPi
18 ± 0.4 20 ± 0.1
0.49 ± 0.01 0.54 ± 0.004
Km (lM) 175 ± 14 11 ± 0.4
Haas and Rose [5] found that AMP had an unusual effect on the ATP:PPi exchange rate for UAE. In particular, they observed that AMP affected the ATP:PPi exchange reaction within a limited range of concentrations and was without effect above or below this region. They surmised that this ‘‘titration” effect reflected a shift be-
validated the charcoal paper method, we proceeded with the characterization of the NAE kinetic mechanism. Titration of ATP and PPi The rate of ATP:PPi exchange as a function of ATP concentration is shown in Fig. 3A and B. In the absence of NEDD8, the reaction is negligible, verifying the ULM dependence of ATP:PPi exchange. At 1 lM NEDD8 and 1 mM PPi, the rate of ATP:PPi exchange displayed hyperbolic kinetics over the range of 4–2000 lM ATP (Fig. 3B). The kinetic parameters Vmax, kcat, and Km were 18 ± 0.4 pmol/min, 0.49 ± 0.01 s 1, and 175 ± 14 lM, respectively (Table 1). The rate of ATP:PPi exchange as a function of PPi concentration is shown in Fig. 3C and D. At 1 lM NEDD8 and 1 mM ATP, the rate of ATP:PPi exchange displayed hyperbolic kinetics over the range of 2– 1000 lM PPi. Nonlinear regression analysis yielded values of 20 ± 0.1 pmol/min, 0.54 ± 0.004 s 1, and 11 ± 0.4 lM for Vmax, kcat, and Km, respectively (Table 1). The kcat value of approximately 0.5 s 1, as measured for both the ATP and PPi titrations, was in good agreement with that observed previously (1 s 1) for the NEDD8 titration of NAE [11].
Fig. 4. AMP K1/2 determination. AMP was serially diluted in assay buffer containing 1 mM ATP, 100 lM PPi (50 cpm/pmol [32P]PPi), and 12.5 nM NAE. The reaction was initiated with NEDD8 (1 lM final). The reaction time was 30 min at 37 °C. Converted data are plotted as a function of AMP concentration.
28
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
tween two parallel ATP:PPi exchange pathways represented by steps 1 and 3 of Scheme 1. That AMP binding directly influenced the equilibrium position of step 2 linking steps 1 and 3 was an important observation because it provided evidence for their three-step model and helped to explain the order of substrate binding for UAE. Because no measurement has been reported previously for the AMP dependence of NAE, we decided to test the hypothesis that NAE behaved in a mechanistically similar fashion to UAE. The titration profile is shown in Fig. 4. AMP affects the NAE ATP:PPi exchange reaction within a limited concentration range and is without effect above or below this region, an observation that is consistent with the proposed mechanism for UAE [5]. A simple fit of the data yields a K1/2 of 28 lM, the inflection point defining the concentration of mononucleotide required for halfmaximal change in rate [5]. This result clearly suggests that NAE follows the same kinetic mechanism as UAE. NEDD8 dependence of ATP:PPi exchange We next examined the rate of the ATP:PPi exchange reaction as a function of NEDD8 concentration (Fig. 5). The increase in rate is hyperbolic to a maximum value of 28 pmol/min at approximately 1.25 lM NEDD8. Above 1.25 lM, the rate decreases to 6 pmol/ min at 40 lM NEDD8. This observation is consistent with that reported previously for NAE as being a formally random addition mechanism but with the preferential binding of ATP as the leading substrate and NEDD8 as the trailing substrate [11]. Previously, Haas and Rose had shown that UAE followed a strictly ordered mechanism of binding [5]. In that case, as the concentration of ubiquitin was increased above 1–2 lM, the rate of the reaction approached zero. Use of ATP:PPi exchange for ULM specificity and other E1 activating enzymes Recently, Groettrup and coworkers reviewed the eight known human E1 enzymes and their corresponding ULMs [4]. Here we addressed the diversity of this family in two experiments. First, we investigated the specificity of NAE activity across a panel of ULMs. Second, we measured the initial rates for all of the known E1 enzymes and their corresponding ULMs in the ATP:PPi exchange assay—NAE:NEDD8; SAE:SUMO1; UAE:ubiquitin; ATG7:MAPLC3A; UBA4:Urm1; UBA5:Ufm1; UBA6:ubiquitin; and UBE1L:ISG15—under identical conditions. The results are shown in Fig. 6A and B,
Fig. 5. NEDD8 dependence of ATP:PPi exchange. NEDD8 was serially diluted in assay buffer. The reactions were initiated by adding the diluted NEDD8 at the stated final concentrations to 12.5 nM NAE in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi). The reaction time was 30 min at 37 °C. Converted data are plotted as a function of NEDD8 concentration.
Fig. 6. (A) ULM specificity of NAE. The reactions of 20 nM NAE in assay buffer containing 1 mM ATP, 1 mM PPi, and 50 cpm/pmol [32P]PPi were initiated with 1 lM (final) of the corresponding ULM. The reaction time was 30 min at 37 °C. (B) Activity of E1:ULM pairs. The reactions of 20-nM E1 enzymes in assay buffer containing 1 mM ATP and 1 mM PPi (50 cpm/pmol [32P]PPi) were initiated with 1 lM (final) of the corresponding ULM. The reaction time was 30 min at 37 °C.
respectively. NAE did not support robust PPi exchange with any of the ULMs tested except NEDD8. NEDD8’s closest related ULM, ubiquitin, was 25 times less active—30 pmol/min for NEDD8 versus 1.2 pmol/min for ubiquitin. Although clearly selective for NEDD8, we also examined the potential for noncognate ULM binding to NAE, thereby inhibiting NEDD8 activation. To test this hypothesis, we conducted a parallel experiment to that described above where 1 lM NEDD8 was added to each of the ULMs tested. Analysis of the data revealed no inhibition of PPi exchange (data not shown). Accordingly, noncognate ULMs do not bind to NAE. Conversely, the data shown in Fig. 6B reveal that the PPi exchange assay is amenable to studying other E1 enzymes and their cognate ULMs. (Although several of the E1 enzymes are known to activate more than one ULM, only those listed were assayed.) The eight known human E1 enzymes were tested. All showed at least fivefold activity versus the ULM control, ranging from 0.06 to 30 pmol/min. That E1 enzymes undergo PPi exchange in the adenylation of their corresponding ULM is consistent with their function. These results demonstrate the general utility of the charcoal paper assay for measuring the ULM specificity and activity of various E1 enzymes, and by extension other enzymes, that catalyze ATP:PPi exchange. Conclusion Although many alternative assays have been proposed to measure ATP:PPi exchange, the traditional method using powder char-
Charcoal paper ATP:PPi exchange assay / F.J. Bruzzese et al. / Anal. Biochem. 394 (2009) 24–29
coal has remained the standard. We have shown that the charcoal paper assay described here represents a significant improvement over the previously published method incorporating powder charcoal absorption. The paper assay affords increased sample throughput and sensitivity while maintaining the simplicity of the powder assay. The results demonstrate the utility of this technique in determining kinetic constants associated with enzymes undergoing ATP:PPi exchange. We showed that the data generated with the charcoal paper assay compare favorably with the historical data for NAE. Moreover, we showed that NAE binds to and is affected by AMP in a manner reminiscent of UAE and, therefore, is consistent with the proposed E1 activating mechanism. These findings demonstrate that classic experiments to resolve substrate binding and catalysis, specificity, order of substrate binding, inhibitor mechanism of action, and the like can now be readily and routinely answered for enzymes undergoing ATP:PPi exchange, placing the charcoal paper assay into the enzymologist’s tool box of everyday experimental protocols. In addition, although the eight E1 enzymes are known to possess a diverse set of biological activities and functions, we showed that all share formation of a similar ULM–adenylate intermediate and can be readily studied by the ATP:PPi exchange assay. The premise of these results argues for applications to other enzymes that exhibit ATP:PPi exchange reactions such as aminoacyl transfer RNA (tRNA) synthetases [18,22]. The potential for using the charcoal paper assay to study other enzymes within this class with minimal modifications, if any, makes this assay a very attractive alternative to the more sophisticated techniques that may require custom substrate labeling, overtly expensive reagents, equipment, and the like to move between studying different enzymes. The advantages of the new method are evident—ease of use, increased sampling, decreased processing time, excellent signal-to-background ratio, signal linearity, versatility, and decreased cost. With the added advantage of the 96-well plate format, automation of the charcoal paper assay is possible. Our approach in the design of the charcoal paper ATP:PPi exchange assay has been one of utility for mechanistic analysis, substrate profiling, and inhibitor characterization. Acknowledgments The authors thank Arthur Haas for his insightful and critical review of the data and methods described in this article. We also thank Michael Bembenek for his assistance during the early developmental stages of the assay, and we acknowledge James Brownell, Mark Rolfe, and William Mallender for their comments on the manuscript. References [1] A. Ciechanover, H. Heller, R. Katz-Etzion, A. Hershko, Activation of the heatstable polypeptide of the ATP-dependent proteolytic system, Proc. Natl. Acad. Sci. USA 78 (1981) 761–765.
29
[2] A. Hershko, A. Ciechanover, The ubiquitin system, Annu. Rev. Biochem. 67 (1998) 425–479. [3] C.M. Pickart, Ubiquitin enters the new millennium, Mol. Cell 8 (2001) 499–504. [4] M. Groettrup, C. Pelzer, G. Schmidtke, K. Hofmann, Activating the ubiquitin family: UBA6 challenges the field, Trends Biochem. Sci. 33 (2008) 230–237. [5] A.L. Haas, I.A. Rose, The mechanism of ubiquitin activating enzyme: a kinetic and equilibrium analysis, J. Biol. Chem. 257 (1982) 10329–10337. [6] R.C. Kane, P.F. Bross, A.T. Farrell, R. Pazdur, VELCADE: U.S. FDA approval for the treatment of multiple myeloma progressing on prior therapy, Oncologist 8 (2003) 508–513. [7] J.F. San Miguel, R. Schlag, N.K. Khuageva, M.A. Dimopoulos, O. Shpilberg, M. Kropff, I. Spicka, M.T. Petrucci, A. Palumbo, O.S. Samoilova, A. Dmoszynska, K.M. Abdulkadyrov, R. Schots, B. Jiang, M.V. Mateos, K.C. Anderson, D.L. Esseltine, K. Liu, A. Cakana, H. Van de Velde, P.G. Richardson, VISTA Trial Investigators, bortezomib plus melphalan and prednisone for initial treatment of multiple myeloma, N. Engl. J. Med. 359 (2008) 906–917. [8] L. Gong, E.T. Yeh, Identification of the activating and conjugating enzymes of the NEDD8 conjugation pathway, J. Biol. Chem. 274 (1999) 12036–12042. [9] Z.Q. Pan, A. Kentsis, D.C. Dias, K. Yamoah, K. Wu, NEDD8 on cullin: building an expressway to protein destruction, Oncogene 23 (2004) 1985–1997. [10] T.A. Soucy, P.G. Smith, M.A. Milhollen, A.J. Berger, J.M. Gavin, S. Adhikari, J.E. Brownell, K.E. Burke, D.P. Cardin, C.A. Cullis, S. Critchley, A. Doucette, J.J. Garnsey, J.L. Gaulin, R.E. Gershman, A.R. Lublinsky, A. McDonald, H. Mizutani, U. Narayanan, E.J. Olhava, S. Peluso, M. Rezaei, M.D. Sintchak, T. Talreja, T. Traore, S. Vyskocil, G.S. Weatherhead, J. Yu, J. Zhang, L.R. Dick, C.F. Claiborne, M. Rolfe, J.B. Bolen, S.P. Langston, An inhibitor of NEDD8-activating enzyme as a novel approach to treat cancer, Nature 458 (2009) 732–736. [11] R.N. Bohnsack, A.L. Haas, Conservation in the mechanism of NEDD8 activation by the human AppBp1–Uba3 heterodimer, J. Biol. Chem. 278 (2003) 26823– 26830. [12] A.R. Fersht, R.S. Mulvey, G.L.E. Koch, Ligand binding and enzymatic catalysis coupled through subunits in tyrosyl–tRNA synthetase, Biochemistry 14 (1975) 13–18. [13] M.M. Madden, W. Song, P.G. Martell, Y. Ren, J. Feng, Q. Lin, Substrate properties of ubiquitin carboxyl-terminally derived peptide probes for protein ubiquitination, Biochemistry 47 (2008) 3636–3644. [14] K. Beebe, W. Waas, Z. Druzina, M. Guo, P. Schimmel, A universal plate format for increased throughput of assays that monitor aminoacyl transfer RNA synthetase activities, Anal. Biochem. 368 (2007) 111–121. [15] I. Dusha, Z.G. Dénes, A simplified assay of enzymes catalyzing ATP– pyrophosphate exchange reactions, Anal. Biochem. 81 (1977) 247–250. [16] R. Alvarez, D.V. Daniels, A single column method for the assay of adenylate cyclase, Anal. Biochem. 187 (1990) 93–108. [17] L.G. Otten, M.L. Schaffer, B.R.M. Villiers, T. Stachelhaus, F. Hollfelder, An optimized ATP/PPi-exchange assay in 96-well format for screening of adenylation domains for applications in combinatorial biosynthesis, Biotechnol. J. 2 (2007) 232–240. [18] R. Jørgensen, T.M.M. Søgaard, A.B. Rossing, P.M. Martensen, J. Justesen, Identification and characterization of human mitochondrial tryptophanyl– tRNA synthetase, J. Biol. Chem. 275 (2000) 16820–16826. [19] D.L. Purich, R.D. Allison, Isotope exchange methods for eludiating enzymatic catalysis, in: D.L. Purich (Ed.), Contemporary Enzyme Kinetics and Mechanism, Academic Press, San Diego, 1996, pp. 335–378 (Chapter 12). [20] K.T. Hughs, Radiometric assays, in: R. Eisenthal, M.J. Danson (Eds.), Enzyme Assays, Oxford University Press, Oxford, UK, 2002, pp. 79–102 (Chapter 3). [21] S.E.H. Syed, High performance liquid chromatographic analysis, in: R. Eisenthal, M.J. Danson (Eds.), Enzyme Assays, Oxford University Press, Oxford, UK, 2002, pp. 103–140 (Chapter 4). [22] T.N.C. Wells, J.W. Knill-Jones, T.E. Gray, A.R. Fersht, Kinetic and thermodynamic properties of wild-type and engineered mutants of tyrosyl– tRNA synthetase analyzed by pyrophosphate-exchange kinetics, Biochemistry 30 (1991) 5151–5156. [23] S.C. Gill, P.H. von Hippel, Calculation of protein extinction coefficients from amino acid sequence data, Anal. Biochem. 182 (1989) 319–326.