Analytical Biochemistry 418 (2011) 102–110

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/locate /yabio

A spectrophotometric assay for conjugation of ubiquitin and ubiquitin-likeproteins q

Christopher E. Berndsen, Cynthia Wolberger ⇑Department of Biophysics and Biophysical Chemistry, Howard Hughes Medical Institute and the Johns Hopkins University School of Medicine, Baltimore, MD 21202, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 26 April 2011Received in revised form 23 June 2011Accepted 24 June 2011Available online 2 July 2011

Keywords:Ubiquitin conjugationUbiquitin-like proteinsUbc13–Mms2Ubiquitin-activating enzyme

0003-2697/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.ab.2011.06.034

q This work was supported by the Howard Hughessupported by a Ruth Kirchstein Fellowship from theMedical Science (F32GM089037).⇑ Corresponding author. Address: Johns Hopkins U

725 N. Wolfe Street, 714 WBSB, Baltimore, MD 21205E-mail address: cwolberg@jhmi.edu (C. Wolberger

1 Abbreviations used: Ub, ubiquitin; Ubl, ubiquitinresonance energy transfer; DTT, dithiothreitol; SDS–PApolyacrylamide electrophoresis; ATP, adenosine trimonophosphate; BSA, bovine serum albumen; EDTAacid; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesu

Ubiquitination is a widely studied regulatory modification involved in protein degradation, DNA damagerepair, and the immune response. Ubiquitin is conjugated to a substrate lysine in an enzymatic cascadeinvolving an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitinligase. Assays for ubiquitin conjugation include electrophoretic mobility shift assays and detection of epi-tope-tagged or radiolabeled ubiquitin, which are difficult to quantitate accurately and are not amenableto high-throughput screening. We have developed a colorimetric assay that quantifies ubiquitin conjuga-tion by monitoring pyrophosphate released in the first enzymatic step in ubiquitin transfer, the ATP-dependent charging of the E1 enzyme. The assay is rapid, does not rely on radioactive labeling, andrequires only a spectrophotometer for detection of pyrophosphate formation. We show that pyrophos-phate production by E1 is dependent on ubiquitin transfer and describe how to optimize assay conditionsto measure E1, E2, and E3 activity. The kinetics of polyubiquitin chain formation by Ubc13–Mms2 mea-sured by this assay are similar to those determined by gel-based assays, indicating that the data producedby this method are comparable to methods that measure ubiquitin transfer directly. This assay is adapt-able to high-throughput screening of ubiquitin and ubiquitin-like conjugating enzymes.

� 2011 Elsevier Inc. All rights reserved.

Protein ubiquitination is a reversible posttranslational modifi-cation that regulates diverse pathways in eukaryotic cells by con-trolling protein stability, activity, binding, and intracellulartargeting [1,2]. Biological processes that are regulated by ubiquiti-nation include proteasomal protein degradation, DNA damage sig-naling, the inflammatory response, and transcription [2–4].Ubiquitin, and ubiquitin-like modifiers (Ubl)1 such as SUMO andNEDD8, are small proteins that are attached to substrates via an iso-peptide linkage between the C-terminus of ubiquitin (or the Ubl) anda substrate lysine [1,2]. Substrates can be modified by a single ubiq-uitin or by a polyubiquitin chain, in which one ubiquitin is cova-lently linked to the next via one of its seven surface lysines [5–7].Ubiquitin and Ubls are conjugated to substrate lysines by thesequential activity of three enzymes: an E1-activating enzyme, an

ll rights reserved.

Medical Institute. C.E.B. wasNational Institute of General

niversity School of Medicine,, USA. Fax: +1 410 614 8648.).-like protein; FRET, FörsterGE, sodium dodecyl sulfate–

phosphate; AMP, adenosine, ethylenediaminetetraacetic

lfonic acid.

E2 ubiquitin-conjugating enzyme, and an E3 ligase [2,8]. The humangenome encodes several E1 enzymes, 35 E2 enzymes, and more than500 E3 enzymes [3,4,9,10]. Our understanding of which E2 and E3enzymes work in concert to conjugate ubiquitin to specific sub-strates is incomplete and remains an active area of research [4,9,11].

The first step in the ubiquitination cascade is the ATP-depen-dent charging of the E1-activating enzyme (Fig. 1). The E1 usesATP to adenylate the C-terminus of ubiquitin, forming ubiquitin–AMP and pyrophosphate. The ubiquitin–AMP ester bond is thencleaved by an active site cysteine within E1, resulting in a thioesterbond between the C-terminus of ubiquitin and the sulfhydryl ofcysteine [12–14]. Ubiquitin is then transferred from the active sitecysteine of the E1 to an active site cysteine in the E2 [2,4]. The E3ligase then binds to both the E2 and the substrate, promoting iso-peptide bond formation with a substrate lysine [3]. Members of theHECT family of E3 ligases first transfer the ubiquitin from the E2 toan active site cysteine in the E3 before conjugating the ubiquitin tothe substrate, while the RING domain E3 ligases promote directtransfer from the E2 to the substrate [3,8]. Substrate specificity isthought to be intrinsic to the specific combination of E2 and E3,although the number of functional combinations of E2s and E3sis not known. Studies of the mechanism and specificity of ubiquitinand Ubl transfer enzymes are important for understandingubiquitin and Ubl signaling in the cell, and for identifying targetsfor therapeutic intervention.

E1

ATP

AMP + PPi

E2HS

S

E3HN

NH Pyrophosphatase

2 Pi Phosphomolybdate

E2Ubiquitin

Substrate

Quench reaction with acidAdd molybdate

Measure absorbanceProcess Data

xx

x

time

Ub

trans

ferre

d

Fig.1. Schematic of molybdenum blue assay. The ubiquitination cascade (bottom) involves the sequential action of E1, E2, and E3 enzymes, leading to ubiquitination of thesubstrate lysine. In the first step, the ubiquitin-activating enzyme, E1, adenylates ubiquitin, forming ubiquitin–AMP and releasing pyrophosphate. To assay ATP consumption(top), the pyrophosphate product is broken down by pyrophosphatase into two phosphate molecules. Addition of ascorbic acid and ammonium molybdate in hydrochloricacid halts the reaction progress and initiates phosphomolybdate formation. The solution is then developed by the addition of citric acid and sodium arsenite and thephosphomolybdate formed measured by absorbance spectroscopy.

A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110 103

The past 10 years have seen growing interest in the ubiquitinconjugation machinery and in the proteasome as targets for drugdiscovery. The effectiveness of the proteasome inhibitor, Bortezo-mib, in treating multiple myeloma and mantle lymphoma hasspurred efforts to target upstream events in the ubiquitin pathwayby inhibiting E1, E2, and E3 enzymes [15]. Nutlin, an inhibitor ofthe RING E3, Mdm2, appears to be effective in the treatment ofmultiple myeloma [16]. Recently, an inhibitor of the E1-activatingenzyme for the ubiquitin-like molecule, Nedd8, entered Phase I tri-als for the treatment of cancer [9]. These two examples offer thepromise of success in targeting the ubiquitin conjugation pathwayby therapeutics. However, progress in developing and screeningsmall molecule inhibitors of E1, E2, and E3 enzymes has been lim-ited by the lack of a simple, high-throughput assay for ubiquitina-tion using native substrates [16].

The most widely used activity assay for ubiquitin or Ubl transferis the band shift assay, in which ubiquitinated and unmodifiedproteins are separated by gel electrophoresis and detected byone of a variety of methods, including staining or Western blot[17–19]. However, these assays make it difficult to quantitate pro-tein modification accurately because of variability or nonlinearityin the signal from stained or antibody-bound proteins. The attach-ment of epitopes or fluorescent tags to ubiquitin or substratespermits accurate quantification and enhances sensitivity; however,adding an epitope tag such as hexahistidine, glutathione-S-trans-ferase (GST), green fluorescent protein (GFP), Myc, or HA to ubiqui-tin or to the substrate has the potential to disrupt interactionsbetween the substrate and the transfer enzymes [20–22]. Ubiquitinradiolabeled with 125[I] or 32[P] is also used [23–25], but requiresspecial handling due to the hazards of working with radioactivity,particularly gamma emitters like 125[I]. In addition, radiolabelingmust be performed regularly because of the short half-lives ofthese isotopes. None of the gel shift assays are amenable to high-throughput formats, both because of the time needed to performthe assay and the variability in signal between gels. Ubiquitinationand Ubl transfer assays that rely on Förster resonance energy

transfer (FRET) between GFP or fluorescently labeled ubiquitinand Ubls have also been developed for real-time, quantitative mea-surement of ubiquitination [22,26–30]. However, the FRET assayrequires attaching a fluorophore to ubiquitin or to the substrate,which may interfere with the function of ubiquitin transfer en-zymes and may preclude the use of native substrates [21]. In addi-tion to methods that directly measure the ubiquitinated product,coupled enzyme methods that monitor ATP cleavage by E1 haveused radiolabeled ATP and quantitated the phosphate released[31,32]. However, none of the above methods are amenable tolarge-scale or high-throughput screening.

We have developed a spectrophotometric method for measur-ing ubiquitin and Ubl transfer that does not require tagged sub-strates or gel electrophoresis and can be used to quantitate eachstep within the ubiquitin transfer cascade. The assay couples ubiq-uitin transfer by E2 and E3 to the formation by E1 of pyrophos-phate, a product of the charging reaction, which is broken downto phosphate by pyrophosphatase. The inorganic phosphate isquantitated by measuring the absorbance of molybdenum blue, areduced complex of phosphate and molybdate that absorbs visiblelight. All reagents required for the assay are easily obtainable andthe assay only requires a spectrophotometer capable of measuringabsorbance between 600 and 850 nm. With the use of a plate read-er, the assay can be readily applied to high-throughput screeningof compounds that target ubiquitin or Ubl transfer enzymes.

Materials and methods

Chemicals and reagents

All chemicals were purchased from Sigma–Aldrich (St. Louis, MO,USA) or Fisher Scientific (Pittsburgh, PA, USA) and used without fur-ther purification. Inorganic pyrophosphatase from Escherichia coli(I5907) was purchased from Sigma–Aldrich and resuspended in20 mM Hepes, pH 7.5, 100 mM NaCl, aliquoted, frozen, andthen stored at �20 �C. We note that, while pyrophosphatase from

104 A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110

Saccharomyces cerevisiae can be substituted for the bacterial en-zyme, the yeast enzyme has measurable ATPase activity and pro-duces higher background absorbance values in the molybdenumblue assay compared to the bacterial enzyme (data not shown).We therefore recommend using the bacterial enzyme.

Protein expression and purification

Saccharomyces cerevisiae Ubc13 and Mms2 were expressed inE. coli and purified as previously described by Eddins et al. [33].Hexahistidine tagged, human E1 enzyme was expressed in E. coliusing a pET32 vector. Cells were grown at 37 �C to an OD of 0.3,shifted to 16 �C, and induced with IPTG overnight. Cells were lysedand then loaded onto a nickel affinity column and eluted withimidazole. Peak fractions were pooled, dialyzed into buffer con-taining 50 mM NaCl, loaded onto a mono Q column (GE Health-care) and eluted with a 0–400 mM NaCl gradient. Peak fractionswere pooled, concentrated, and purified by size-exclusion chroma-tography on a Superdex 200 column to 20 mM Tris, pH 7.8,100 mM NaCl, and 7.5 mM b-mercaptoethanol (GE Healthcare).Fractions containing E1 were pooled, concentrated, and stored at�80 �C until use. A gel showing the purity of recombinant E1 isshown in Supplementary Fig. 1.

Wild-type and mutant human ubiquitin were expressed inE. coli and purified following the protocol of Pickart and Raasi[20,34]. Ubiquitin mutants are denoted in the text with super-scripts; UbK63R or UbDG75,DG76, for example, indicate mutation oflysine 63 to arginine or the deletion of glycine 75 and glycine 76from ubiquitin.

Saccharomyces cerevisiae Rad5 RING domain (residues 846–1196) was expressed in E. coli from a pMalc2xHV plasmid contain-ing an N-terminal hexahistidine tag followed by a tobacco etchvirus (TEV) protease cleavage site. The RING domain was purifiedby nickel affinity chromatography and then dialyzed with 1 mgof TEV protease overnight to remove the affinity tag. Dialyzed pro-tein was then concentrated and further purified by size-exclusionchromatography on a Superdex 75 column in 20 mM Tris, pH 7.5,100 mM NaCl, and 5 mM b-mercaptoethanol. Fractions containingthe Rad5 RING domain were pooled, concentrated, and stored at�80 �C until use.

Solutions for molybdenum blue assay

The quench and developer solutions were modified from condi-tions described in Gawronski and Benson [35] and McCain andZhang [36]. The quench solution contains 8% L-ascorbic acid and0.67% ammonium molybdate in 0.67 N HCl. This solution is madefresh from a 2:3 dilution of a solution of 12% ascorbic acid in 1 NHCl with 2% ammonium molybdate in water. The ammoniummolybdate/ascorbic acid solution turns yellow on mixing withammonium molybdate and is stable for only a few days in the darkat 4 �C (the solution will eventually turn green or brown and signalwill deteriorate significantly). The developer solution contains 2%sodium citrate-trihydrate, 2% sodium (meta)arsenite in 2% aceticacid. Acetic acid is added last. This solution is stable for about aweek at room temperature.

Generating a standard curve

A 1 mM sodium phosphate stock solution used for standardswas made in buffer containing 20 mM Hepes, pH 7.5, 100 mMNaCl. This stock is diluted to a final phosphate concentration be-tween 0 and 200 lM in 20 mM Hepes, pH 7.5, 100 mM NaCl in afinal volume of 50 lL. To each standard, 150 lL of the quench solu-tion is added and the quenched reactions are incubated at roomtemperature. After 5 min, 150 lL of the developer solution is added

and the absorbance of the sample is measured at 850 nm. Theabsorbance of each standard is then plotted as absorbance versusthe concentration of phosphate (in mM or lM). The data are thenfitted to a linear equation to obtain the slope in units ofabs �mM�1 or abs � lM�1. Absorbance values from ubiquitinationreactions are then converted to micromolar phosphate using theslope of the phosphate standard curve:

absorbance850 nm=slope of standard curve ðabs=lMÞ¼ lM phosphate formed: ð1Þ

Since there are two phosphates generated from every pyrophos-phate molecule, the concentration of pyrophosphate formed (usu-ally in lM) is then divided by 2:

lM phosphate formed=2 ¼ lM pyrophosphate formed

¼ lM ubiquitin transferred: ð2Þ

It is important to generate the standard curve using the sameinstrument as that used to conduct the experiments, since instru-ment-to-instrument variations, measurement path length, andother factors can affect the absolute absorbance measurements.

Molybdenum blue assay of ubiquitin conjugation

For the assays described in this paper, reactions contained50 mM Hepes, pH 7.5, 100 mM NaCl, 1 mM DTT, 10 mM MgCl2,500 U inorganic pyrophosphatase, 0.05–0.6 lM E1 enzyme, 0.1–4lM Ubc13–Mms2 enzyme, 1 lM Rad5 RING domain, 80 lMUbK63R, and 0–1 mM UbDG75,DG76 or Ubd77. Reactions were initiatedby the addition of 2 mM ATP. All reactions were performed at37 �C. At the designated time points, a 50-lL aliquot of the reactionwas removed and quenched with 150 lL of the quench solution.After 5 min at room temperature (23–26 �C), 150 lL of the devel-oper solution was added and the absorbance of the solution wasthen measured at 850 nm in a Cary 50 Bio UV–Vis spectrometer(Agilent Technologies, Santa Clara, CA) or by recording a spectrumfrom 500–900 nm in a POLARStar Omega plate reader (BMGLABTECH GmbH, Offenburg, Germany). If necessary, the volumeof the reaction aliquot can be decreased and the ratios of the ali-quot to quench and developer proportionally reduced. The absor-bance at 850 nm was used to quantitate activity; however, anywavelength between 600 and 875 nm can be used for measure-ment. The amount of pyrophosphate formed should be equivalentto the amount of ubiquitin transferred by E1 to E2. The quantity ofpyrophosphate formed is also equivalent to the amount of ubiqui-tin conjugated to substrate if either E2 or E3 is the rate-limiting en-zyme (see Results and discussion). Each new E2 or E3 enzyme to beassayed will require optimization of conditions (see Results anddiscussion for more details).

Reaction rates in micromolar per minute were determined fromlinear fits of micromolar pyrophosphate formed versus time. Therate of pyrophosphate formation was then plotted as a functionof ubiquitin concentration and fit to the Michaelis–Mentenequation to determine kinetic constants:

v ¼ ðkcat � ½E�t � ½S�Þ=ðKM þ ½S�Þ: ð3Þ

Here v is the initial velocity at a given substrate concentration, kcat

is the apparent first-order rate constant of the enzyme, [E]t = the to-tal concentration of enzyme, [S] is the total concentration of sub-strate, and KM is the concentration of substrate required to reachone-half the maximum velocity of the enzyme.

Fluorescent-ubiquitin assays

UbiquitinK63C was labeled with fluorescein-5-maleimide(Thermo Scientific, Pittsburgh, PA) in PBS buffer for 2 h at room

A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110 105

temperature in the dark. Labeling reactions were dialyzed over-night into 2 L of 20 mM Hepes, pH 7.5, 1 mM EDTA. Ubiquitinwas further purified by cation-exchange chromatography as de-scribed [20]. Fractions that contained ubiquitin were pooled anddialyzed overnight into 2 L of 20 mM Hepes, pH 7.5. The concentra-tion of ubiquitin was determined using the BCA assay kit (ThermoScientific) following the manufacturer’s protocol. Labeled ubiquitinwas stored at 4 �C wrapped in foil.

Band shift assays using 10 lM fluorescently labeled ubiquitin(labeled at cysteine 63) with an additional 70 lM ubiquitinK63R

(total donor ubiquitin concentration = 80 lM) were performedusing the assay conditions and reagents described above for themolybdenum blue assay. At designated time points, reactions werestopped by the addition of 4� NuPAGE SDS–PAGE loading bufferwith dye (NP0007 Invitrogen, Carlsbad, CA). Reactions were sepa-rated by SDS–PAGE and the bands visualized on a Typhoon 9410phosphorimager (GE Healthcare, Waukesha, WI). Band intensitywas measured using Quantity One version 4.6.7 (Bio-Rad, Hercules,CA). Fluorescence intensity was converted to concentration bydividing the band intensity by the slope obtained from a calibra-tion curve of fluorescein ubiquitin standards of known concentra-tion on the same gel. Data were plotted and fitted as describedabove to determine kinetic constants.

Results and discussion

Assay description

We set out to develop a quantitative assay that could be usedwith native substrates for both general activity screening assaysand mechanistic studies. Activation of ubiquitin or an Ubl by E1,the first step in the ubiquitin transfer cascade (Fig. 1), involves for-mation of ubiquitin–AMP and pyrophosphate from ubiquitin andATP. The ubiquitin adenylate is then cleaved by the active site cys-teine of the E1, resulting in a thioester bond between the E1 andthe C-terminus of ubiquitin and AMP. The ubiquitin is then trans-ferred to the active site cysteine of the E2 in a transthiolation reac-tion (Fig. 1). The E3 ligase binds to the E2–ubiquitin conjugate andpromotes transfer of the ubiquitin or Ubl to the substrate lysine.The release of pyrophosphate is therefore, in principle, directly re-lated to the amount of ubiquitin transferred because E1 musttransfer ubiquitin to E2 before it can undergo another round ofATP cleavage. It is therefore possible to use the rate of pyrophos-phate production by E1 as a measure of ubiquitin transfer. Pyro-phosphate is measured by first hydrolyzing it with the enzyme,pyrophosphatase, to produce inorganic phosphate, which is thencomplexed with molybdate in the presence of ascorbic acid toproduce a blue color, sometimes referred to as molybdenum blue[34–37]. The relative absorbance of visible or infrared light byphosphomolybdate is proportional to the amount of phosphategenerated, and hence to the number of times the E1 enzyme ischarged with ubiquitin. The absorbance of phosphomolybdatecompounds has been used to assay the activity of phosphatases,synthetases, and other enzymes that release inorganic phosphate[35,36]. We therefore investigated whether this method could beadapted to measuring substrate ubiquitination by monitoringpyrophosphate release by E1.

To assay ubiquitination using the phosphomolybdate assay, allproteins are first mixed in the absence of ATP and equilibrated at37 �C. The ubiquitination reaction is then initiated with the addi-tion of ATP. At designated times, an aliquot of each reaction isremoved and diluted into a solution of ammonium molybdate,ascorbic acid, and hydrochloric acid as described under Materialsand methods. Hydrochloric acid lowers the pH of the solution,which decreases the catalytic activity of the ubiquitin conjugation

enzymes and effectively quenches the reaction [38,39]. Molybdateforms a complex with phosphate, which is reduced by ascorbic acidand gives rise to the blue color [35,40]. Phosphomolybdate forma-tion is then stopped after 5 min by the addition of a citrate/arsenitesolution, which competes for free molybdate and enhances the col-or stability of the phosphomolybdate [41]. The phosphomolybdateformed at each time point was then quantified by measuring theabsorbance at 850 nm. The absorbance is divided by 2, since twophosphates are produced for every ubiquitin transferred becauseATP is cleaved to AMP and pyrophosphate. These numbers are cor-rected for background ATP hydrolysis that is not dependent on theactivity of E1 or E2. The concentration of phosphate is then deter-mined from the corrected absorbance by comparison with thestandard curve (see Materials and methods).

We developed the assay using the enzymes that synthesizeK63-linked polyubiquitin, which is synthesized in yeast by theE2, Ubc13, in complex with Mms2, a noncatalytic subunit that con-fers the specificity for K63 of the acceptor ubiquitin [21,33,42,43].Ubc13–Mms2 can synthesize free polyubiquitin chains in the ab-sence of the RING E3 ligase, Rad5 [42,44,45]. When added to thein vitro reaction in the absence of other substrates, Rad5 stimulatesthe rate of free chain synthesis by Ubc13–Mms2 [45]. It is thereforepossible to monitor the rate of the unstimulated E1/E2 reaction, aswell as the E3-stimulated reaction.

We first determined whether the rate of ATP cleavage by E1 wasincreased during free chain synthesis by Ubc13–Mms2, whichwould confirm that ATP cleavage by E1 is coupled to transfer ofubiquitin to E2. We carried out the reactions under conditions thatwould lead to formation of diubiquitin only by using 80 lM UbK63R,which can only serve as a donor ubiquitin, and saturating amountsof UbD77, which can only serve as an acceptor because D77 pre-vents the ubiquitin from binding to and reacting with the E1[20]. Fig. 2 shows the time course of pyrophosphate producedduring diubiquitin formation by Ubc13–Mms2, which is repeatedlycharged by E1 for each ubiquitin transferred. A control reactioncontaining the E1, ubiquitin, pyrophosphatase, and ATP, which cor-rects for background ATP hydrolysis that is not due to ubiquitintransfer to E2, is nearly 100-fold lower than with Ubc13–Mms2(1.2 lM/min versus 0.014 lM/min). We verified that pyrophospha-tase was saturating by assaying the rate of pyrophosphate forma-tion over a broad range of pyrophosphatase concentrations andfound that it did not change significantly at pyrophosphatase con-centrations greater than 20 U (Supplementary Fig. 2). We note thatthere is significant ATPase activity at pyrophosphatase concentra-tions greater than 500 U (data not shown) and recommendcontrolling for this activity by assaying phosphate formation inthe absence of E1.

We next measured the rate of pyrophosphate production in thepresence of a RING domain E3 ligase to show that pyrophosphateproduction can reflect the activity of an E3. In the presence ofthe RING domain of yeast Rad5, we measured a rate of 2.9 lM/min, a 3-fold stimulation in the rate of pyrophosphate production(Fig. 2). This is comparable to a previously reported 4-fold stimula-tion in the rate of ubiquitin chain formation by Ubc13-Mms2 andintact Rad5, which was assayed using band shift assays [45]. Thesedata indicate that measuring phosphate production is a viableassay of E2 and RING E3 activity.

Different sets of enzymes involved in ubiquitin and Ubl modifi-cation require distinct conditions for assaying activity. In order todetermine whether some of these conditions might affect molyb-denum blue color formation, we determined how different assaybuffers, salts, and additives affect assay results, thereby establish-ing the range of conditions under which the assay can be used.Absorbance of a phosphomolybdate standard was compared inthe presence of a variety of buffers, salts, reducing agents, andreaction additives such as BSA and EDTA, and the absorbance

E1Ubc13-Mms2 (E2)RING domain (E3)

E1Ubc13-Mms2 (E2)

E1

diubiquitin

ubiquitin

E1time (min.) 0 2 5 12 15 22 0 2 5 12 15 22

E1 + E2 0 2 5 12 15 22

E1 + E2 + E3

Fig.2. Pyrophosphate production during diubiquitin synthesis. The plot shows pyrophosphate formation during diubiquitin synthesis by Ubc13–Mms2 in the absence (opencircles) and presence (black circles) of the RING domain of the Rad5 E3. The control (black triangles) shows the background signal in the absence of Ubc13–Mms2. Reactionscontained 0.6 lM Ubc13–Mms2, 1 lM Rad5 RING domain, 0.3 lM E1, 80 lM ubiquitinK63R (the donor ubiquitin), and 800 lM ubiquitinD77 (the acceptor ubiquitin). Datapoints are an average of three independent measurements and the error bars are the standard deviation of the experiments. Lower panel shows a gel of diubiquitin synthesisusing fluorescently labeled ubiquitin.

106 A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110

was compared over a broad range of concentrations (Supplemen-tary Fig. 3). Phosphomolybdate absorbance was unaffected by mostreaction components tested except for phosphate buffers, highconcentrations of DTT (>20 mM), Hepes buffer (>100 mM), and bo-vine serum albumin (>3 mg/mL; this also showed significant pre-cipitation). For best results, phosphate standards should beobtained using buffers that are identical to reaction conditions.

Phosphomolybdate absorbs most strongly between 600 and900 nm; therefore, any spectrophotometer or plate reader thatcan measure absorbance between these wavelengths can be usedfor this assay (Fig. 3A). We find that the absorbance of phosphomo-lybdate at 600, 710, and 850 nm is linear and highly reproduciblefrom 5 to 200 lM phosphate (Fig. 3B). Phosphomolybdate absorbsmost strongly around 850 nm, but any wavelength between 600and 875 nm is suitable for assaying ubiquitin transfer andproduces similar results (Fig. 3C). The blue color is stable at allthree wavelengths for greater than 4 h in the presence of ubiquitin,with a �3% loss in signal (data not shown).

Application of the molybdenum blue assay to high-throughputinvestigations of ubiquitin conjugation

The molybdenum blue assay is easily adapted to a multiwellplate format for high-throughput measurement of ubiquitin conju-gation using a plate reader. Possible applications include screeningfor activity by different E2–E3 pairs, potential substrates, or theeffect of small molecule inhibitors of E1, E2, or E3 enzyme activity.The assay is particularly well suited to screening for E2–E3 pairsand ubiquitin/Ubl targets because it can be used with native en-zymes and substrates.

To demonstrate the application of the molybdenum blue assayto high-throughput screening, we used the assay in a plate formatto measure the effects of several different ubiquitin mutations onK63-linked polyubiquitin formation by Ubc13–Mms2. Assayscontaining 0.5 lM E1, 2 lM E2 (Ubc13–Mms2), and 100 or

500 lM of each mutant ubiquitin were incubated at 37 �C for60 min (Fig. 4). For each mutant ubiquitin, control reactions mea-sured pyrophosphate production by E1 plus ubiquitin and ubiqui-tin alone in separate wells in the plate. Reactions were quenchedafter 60 min and the absorbance was measured in a plate readeras described under Materials and methods. The results in Fig. 4show that Ubc13–Mms2 formed ubiquitin chains with all ubiquitinmutants tested except the K63R substitution, which cannot serveas an acceptor ubiquitin for K63-linked polyubiquitin chain syn-thesis [21]. Substitutions at residues 8, 20, 42, 44, or 57 of ubiquitindecreased chain formation by factors ranging from 3- to 10-fold ascompared to wild-type ubiquitin. Importantly, the relative effectsof each ubiquitin mutation as assayed by the phosphomolybdateplate assay agree well with band shift assays of ubiquitin chain for-mation (Supplementary Fig. 4). The effects of mutations at residues8, 42, and 44 of ubiquitin agree with structural studies showingthat these residues are important for acceptor ubiquitin bindingto Mms2 [33,46]. Residues 20 and 57 of ubiquitin are proximalto the acceptor ubiquitin binding surface in Ubc13, suggesting thatthese substitutions disrupt this interaction [33]. These data showthat it is possible to rapidly screen the effects of ubiquitin muta-tions on ubiquitin transfer using the molybdenum blue assay.

Establishment of the rate-limiting enzyme

For quantitative investigations using the molybdenum blue as-say, it is essential to carry out the assay under conditions in whichthe enzyme of interest (E1, E2, or E3) is rate limiting. Under theseconditions, the rate of pyrophosphate formation will reflect therate-limiting step in the enzymatic cascade. To address whetherisopeptide bond formation by Ubc13–Mms2 was the rate-limitingstep in our assays, we determined how the rate of pyrophosphateformation changed with increasing concentrations of E1 and E2.Reaction rates are dependent on both the concentration of sub-strate and the concentration of enzyme [47]. Therefore, if E2 is rate

Fig.3. Absorbance of phosphomolybdate. (A) Absorbance spectrum of phosphomo-lybdate from 100 lM (solid line), 200 lM (dotted line), and 500 lM (dashed line)sodium phosphate in 20 mM Hepes, pH 7.5, and 100 mM NaCl. Spectra were takenof 200-lL solutions in a 96-well plate using a POLARStar Omega plate reader. (B)Linearity of phosphomolybdate absorbance at 600 nm (filled circles), 710 nm (opencircles), and 850 nm (filled triangles). (C) Molybdenum blue assay of time course ofUbc13–Mms2 ubiquitin chain formation measured at 600 nm (filled circles),710 nm (filled triangles), and 850 nm (open circles).

A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110 107

limiting, increasing the concentration of E2 2-fold will increase therate of pyrophosphate formation by 2-fold as well. A smaller

change would indicate that E1 or, in the molybdenum blue assay,pyrophosphatase is rate limiting. Since the rates of catalysis willlikely be different for each combination of E1, E2, and E3 enzymes,it is therefore important to determine the concentration rangewhere the activity of the enzyme of interest is rate limiting if themolybdenum blue assay is to be used to determine kinetic con-stants. Once the concentrations at which E2 is the rate-limitingstep have been established, the steady-state kinetics of ubiquitintransfer can be measured with the confidence that the measuredrates of phosphate production correlate with isopeptide bond for-mation catalyzed by E2. To maximize signal-to-noise and, hence,accuracy, assays should utilize the maximum enzyme concentra-tion that is still rate limiting.

As an example of how the rate-limiting concentration of E2 isdetermined, we measured the rate of pyrophosphate formationat concentrations of E2 ranging from 0 to 4 lM and at two differentconcentrations of E1. Fig. 5 shows a plot of the rate of pyrophos-phate formation as a function of Ubc13/Mms2 concentration. Therate of pyrophosphate formation varies linearly with E2 concentra-tion between 0 and 2 lM Ubc13–Mms2 at both 0.3 and 0.6 lM E1(shown by the gray box in Fig. 5), indicating that ubiquitin conju-gation by Ubc13–Mms2 is the rate-limiting step under these con-ditions. At concentrations of Ubc13–Mms2 greater than 2 lM,however, this relationship no longer strictly holds, suggesting thatthe E2 is no longer entirely rate limiting.

Steady-state kinetics of diubiquitin formation by Ubc13–Mms2

In order to compare the results of the molybdenum blue assaywith a conventional ubiquitin conjugation assay, we comparedthe steady-state kinetics of diubiquitin formation by Ubc13–Mms2 as measured using a band shift assay and the molybdenumblue assay. Formation of diubiquitin only was ensured by utilizinga donor ubiquitin with a K63R substitution, which blocks chain for-mation by Ubc13–Mms2, and an acceptor ubiquitin lacking thetwo C-terminal glycine residues, which cannot be activated byE1. The rate of diubiquitin formation was assayed by the two meth-ods at several concentrations of the acceptor ubiquitin. In themolybdenum blue assay, absorbance measurements were cor-rected for background ATP hydrolysis by subtracting from eachmeasurement the absorbance in a matched reaction lacking E2.Data were plotted as the rate (in s�1) versus concentration of ubiq-uitin and fit to the Michaelis–Menten equation to calculate kineticconstants. The saturation plot shown in Fig. 6 displays the curvesfrom the molybdenum blue and gel assays along with the calcu-lated rate constants. The turnover numbers for E2 and Michaelisconstant determined from the molybdenum blue (kcat = 0.033 ±0.004 s�1 and KM = 243 ± 63 lM, R2 = 0.95) and gel shift assay(kcat = 0.034 ± 0.005 s�1 and KM = 271 ± 93 lM, R2 = 0.94) are invery good agreement. These values are comparable to those previ-ously reported by Carlile et al. (kcat = 0.02 s�1 and KM = 212 lM)using a gel assay with 125[I]ubiquitin [45]. Taken together, our re-sults indicate that kinetic data obtained using the molybdenumblue assay are comparable to kinetic data obtained by othermethods.

Assay controls and caveats

The key to successfully measuring ubiquitin conjugation in amolybdenum blue assay is to control for background ATPase activ-ity that could artificially inflate or deflate the apparent rate of ubiq-uitin transfer. While we have explained the importance ofdetermining the rate-limiting enzyme when using the molybde-num blue assay, good laboratory practice and additional controlsmay be necessary to increase accuracy. The background level ofphosphate released can be affected by the purity of the enzymes,

Fig.4. High-throughput screen for the effect of ubiquitin mutations on polyubiquitin chain formation. The effect of ubiquitin mutations on polyubiquitin chain formation byUbc13–Mms2. Wells in row 1 contain E1, Ubc13–Mms2, and ubiquitin, row 2 contains E1 and ubiquitin, and row 3 contains ubiquitin alone. Protein concentrations used were2 lM Ubc13–Mms2, 0.5 lM E1, 500 lM ubiquitin (except for no-lysine ubiquitin which was at 100 lM). Reactions were incubated for 1 h at 37 �C and quenched and theabsorbance was measured. Assay was performed in duplicate, with representative data shown.

Fig.5. Rate-limiting concentration range of Ubc13–Mms2. Rate of pyrophosphateformation in lM/s at 0, 0.4, 0.8, 1.2, 2, and 4 lM Ubc13–Mms2 at 0.3 lM E1 (filledcircles) and 0.6 lM E1 (open circles). Gray box indicates concentrations whereubiquitin transfer by Ubc13–Mms2 is directly correlated with pyrophosphateproduction by E1.

Fig.6. UbiquitinDG75,DG76 saturation curves for Ubc13–Mms2. Reactions contained0.3 lM Ubc13–Mms2, 0.2 lM E1, 0–950 lM ubiquitinDG75,DG76, and 80 lM ubiq-uitin K63R. For molybdenum blue assay reactions (black circles), 500 U ofpyrophosphatase was present and for the band shift assay (open circles), 10 lMubiquitin labeled with fluorescein at position 63 and with 70 lM unlabeledubiquitinK63R. Each rate is an average of 3–5 separate experiments. Error bars arenot shown for clarity.

108 A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110

substrates, and reagents used in the assay. In general, a backgroundvalue or rate reflecting ATP cleavage by solvent, pyrophosphatase,and E1 (if E1 is not the enzyme under investigation) should be sub-tracted from all measurements. This value can be from a singlereaction which measures the combined enzymatic and nonenzy-matic ATP cleavage rates or from separate reactions independently

measuring the rate of ATP cleavage by solvent, pyrophosphatase,and E1. Fresh ATP stocks should be used to reduce the backgroundphosphate signal that is due to the slow hydrolysis of ATP duringprolonged storage. Signs of ATP degradation will initially appear

A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110 109

as elevated background and can lead to lowered rates because E1 isinhibited by AMP [12]. We typically make a large 100 mM stock ofATP, which is aliquoted and stored at �20 �C and discarded after6 months. For highly quantitative measurements, the E1, E2, andE3 enzymes and ubiquitin should be >95% pure to reduce the like-lihood of contaminating ATPase activity. The background rate ofphosphate formation in the absence of substrate due to a contam-inate can be subtracted from the overall rate of phosphate forma-tion or this activity can be removed through further purification.For a more qualitative activity screen, for example, in high-throughput screens of different E2–E3 pairs, it is possible to usepartially purified proteins provided that the appropriate controlsfor the ATPase activity of the contaminates are performed.

The molybdenum blue assay correlates pyrophosphate produc-tion with total ubiquitin transferred. Therefore, any transfer ofubiquitin, including automodification of E3 enzymes, modificationof multiple sites within the same substrate, and ubiquitin chainformation, will be counted in the final absorbance measurement.If measurement of heterogeneous modifications is not desired, itis important to establish the order of modification (i.e., sub-strate > E3 > chains) and the time scale in which the desired mod-ification occurs. In the simplest case, in which the substrate is thefirst protein modified, the assay length can be kept sufficientlyshort such that nearly all the ubiquitin conjugation events involvethe product of interest. In cases where the substrate is not prefer-entially modified first or where mixed modification is unavoidable,we recommend combining the molybdenum blue assay with aband shift assay. Aliquots of the reaction should be taken beforeinitiation and after the last time point for separation by SDS–PAGE.The gel should be stained for total protein or blotted for the ubiq-uitin or Ubl to identify the modified proteins. For each protein, thepercentage of the product of interest is determined and used topartition the total rate of ubiquitin transfer measured in themolybdenum blue assay into the component rates of modification.

In addition to background formation of ATP, any reagents addedto the reaction must not act by altering the activity of pyrophos-phatase or affecting the chemical reaction to form phosphomolyb-date. This is especially important when screening chemicallibraries for inhibitors or activators of ubiquitin conjugation. Sincethis interference with the assay is not a likely event, follow-up con-trols can be carried out only on compounds that show an effect inthe initial high-throughput screen.

Conclusions

We have developed a straightforward, colorimetric assay forubiquitin conjugation that can be used in kinetic studies as wellas in high-throughput screens for the effects of mutations, smallmolecules, different E2–E3 pairs, and the presence of different po-tential substrates. The assay couples release of pyrophosphate byE1 to the transfer of ubiquitin by E2 and/or E3. The pyrophosphateis quantitated by breaking it down with pyrophosphatase intophosphate, which is combined with molybdate to form a bluephosphomolybdate complex, sometimes referred to as molybde-num blue. The assay is rapid, simple, safe, and as inexpensive toperform as gel shift assays, while as quantitative as radioactiveand FRET assays that have been previously used to investigateubiquitin conjugation [22,23,26–29,45,48]. Additional advantagesof the assay are the ability to use native, untagged ubiquitin as wellas native substrates. Since ATP cleavage by an E1 is universal foractivation of ubiquitin and ubiquitin-like proteins such as SUMOand NEDD8, the molybdenum blue assay is applicable to these sys-tems, as well. This assay will open the door to in-depth screeningof therapeutics targeting ubiquitin conjugation and mechanisticstudies of E2 and E3 enzymes.

Acknowledgments

We thank Dr. Reuven Wiener for the generous gift of the mutantubiquitin proteins and the RING domain of Rad5. We also thank Dr.Tao Wang for additional ubiquitin mutants and Anthony DiBellofor purified human E1.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ab.2011.06.034.

References

[1] M. Hochstrasser, Origin and function of ubiquitin-like proteins, Nature 458(2009) 422–429.

[2] C.M. Pickart, M.J. Eddins, Ubiquitin: structures, functions, mechanisms,Biochim. Biophys. Acta 1695 (2004) 55–72.

[3] R.J. Deshaies, C.A. Joazeiro, RING domain E3 ubiquitin ligases, Annu. Rev.Biochem. 78 (2009) 399–434.

[4] D.M. Wenzel, K.E. Stoll, R.E. Klevit, E2s: structurally economical andfunctionally replete, Biochem. J. 433 (2010) 31–42.

[5] Z.J. Chen, L.J. Sun, Nonproteolytic functions of ubiquitin in cell signaling, Mol.Cell 33 (2009) 275–286.

[6] D. Komander, M.J. Clague, S. Urbe, Breaking the chains: structure and functionof the deubiquitinases, Nat. Rev. Mol. Cell Biol. 10 (2009) 550–563.

[7] W. Li, Y. Ye, Polyubiquitin chains: functions, structures, and mechanisms, Cell.Mol. Life Sci. 65 (2008) 2397–2406.

[8] A.D. Capili, C.D. Lima, Taking it step by step: mechanistic insights fromstructural studies of ubiquitin/ubiquitin-like protein modification pathways,Curr. Opin. Struct. Biol. 17 (2007) 726–735.

[9] L. Bedford, J. Lowe, L.R. Dick, R.J. Mayer, J.E. Brownell, Ubiquitin-like proteinconjugation and the ubiquitin–proteasome system as drug targets, Nat. Rev.Drug Discov. 10 (2011) 29–46.

[10] B.A. Schulman, J.W. Harper, Ubiquitin-like protein activation by E1 enzymes:the apex for downstream signalling pathways, Nat. Rev. Mol. Cell Biol. 10(2009) 319–331.

[11] D. Rotin, S. Kumar, Physiological functions of the HECT family of ubiquitinligases, Nat. Rev. Mol. Cell Biol. 10 (2009) 398–409.

[12] A.L. Haas, I.A. Rose, The mechanism of ubiquitin activating enzyme. A kineticand equilibrium analysis, J. Biol. Chem. 257 (1982) 10329–10337.

[13] A.L. Haas, J.V. Warms, A. Hershko, I.A. Rose, Ubiquitin-activating enzyme.Mechanism and role in protein–ubiquitin conjugation, J. Biol. Chem. 257(1982) 2543–2548.

[14] A.L. Haas, J.V. Warms, I.A. Rose, Ubiquitin adenylate: structure and role inubiquitin activation, Biochemistry 22 (1983) 4388–4394.

[15] M.D. Petroski, The ubiquitin system, disease, and drug discovery, BMCBiochem. 9 (Suppl. 1) (2008) S7.

[16] A.G. Eldridge, T. O’Brien, Therapeutic strategies within the ubiquitinproteasome system, Cell Death Differ. 17 (2010) 4–13.

[17] A.A. Yunus, C.D. Lima, Purification and activity assays for Ubc9, the ubiquitin-conjugating enzyme for the small ubiquitin-like modifier SUMO, MethodsEnzymol. 398 (2005) 74–87.

[18] A.A. Yunus, C.D. Lima, Purification of SUMO conjugating enzymes and kineticanalysis of substrate conjugation, Methods Mol. Biol. 497 (2009) 167–186.

[19] A. Hershko, E. Eytan, A. Ciechanover, A.L. Haas, Immunochemical analysis ofthe turnover of ubiquitin–protein conjugates in intact cells. Relationship to thebreakdown of abnormal proteins, J. Biol. Chem. 257 (1982) 13964–13970.

[20] C.M. Pickart, S. Raasi, Controlled synthesis of polyubiquitin chains, MethodsEnzymol. 399 (2005) 21–36.

[21] R.M. Hofmann, C.M. Pickart, In vitro assembly and recognition of Lys-63polyubiquitin chains, J. Biol. Chem. 276 (2001) 27936–27943.

[22] M.D. Boisclair, C. McClure, S. Josiah, S. Glass, S. Bottomley, S. Kamerkar, I.Hemmila, Development of a ubiquitin transfer assay for high throughputscreening by fluorescence resonance energy transfer, J. Biomol. Screen. 5(2000) 319–328.

[23] A.L. Haas, Purification of E1 and E1-like enzymes, Methods Mol. Biol. 301(2005) 23–35.

[24] D.C. Scott, J.K. Monda, C.R. Grace, D.M. Duda, R.W. Kriwacki, T. Kurz, B.A.Schulman, A dual E3 mechanism for Rub1 ligation to Cdc53, Mol. Cell 39(2010) 784–796.

[25] P. Tan, S.Y. Fuchs, A. Chen, K. Wu, C. Gomez, Z. Ronai, Z.Q. Pan, Recruitment of aROC1–CUL1 ubiquitin ligase by Skp1 and HOS to catalyze the ubiquitination ofI [kappa] B [alpha], Mol. Cell 3 (1999) 527–533.

[26] T.L. Gururaja, T.R. Pray, R. Lowe, G. Dong, J. Huang, S. Daniel-Issakani, D.G.Payan, A homogeneous FRET assay system for multiubiquitin chain assemblyand disassembly, Methods Enzymol. 399 (2005) 663–682.

[27] S.F. Martin, M.H. Tatham, R.T. Hay, I.D. Samuel, Quantitative analysis of multi-protein interactions using FRET: application to the SUMO pathway, Protein Sci.17 (2008) 777–784.

110 A spectrophotometric assay for ubiquitin conjugation / C.E. Berndsen, C. Wolberger / Anal. Biochem. 418 (2011) 102–110

[28] M.W. Kao, L.L. Yang, J.C. Lin, T.S. Lim, W. Fann, R.P. Chen, Strategy for efficientsite-specific FRET-dye labeling of ubiquitin, Bioconjug. Chem. 19 (2008) 1124–1126.

[29] C.B. Carlson, R.A. Horton, K.W. Vogel, A toolbox approach to high-throughputTR-FRET-based SUMOylation and DeSUMOylation assays, Assay Drug Dev.Technol. 7 (2009) 348–355.

[30] K.E. Wee, Z. Lai, K.R. Auger, J. Ma, K.Y. Horiuchi, R.L. Dowling, C.S. Dougherty, J.I.Corman, R. Wynn, R.A. Copeland, Steady-state kinetic analysis of humanubiquitin-activating enzyme (E1) using a fluorescently labeled ubiquitinsubstrate, J. Protein Chem. 19 (2000) 489–498.

[31] C.M. Pickart, I.A. Rose, Functional heterogeneity of ubiquitin carrier proteins,Prog. Clin. Biol. Res. 180 (1985) 215.

[32] I.A. Rose, J.V. Warms, A specific endpoint assay for ubiquitin, Proc. Natl. Acad.Sci. USA 84 (1987) 1477–1481.

[33] M.J. Eddins, C.M. Carlile, K.M. Gomez, C.M. Pickart, C. Wolberger, Mms2–Ubc13covalently bound to ubiquitin reveals the structural basis of linkage-specificpolyubiquitin chain formation, Nat. Struct. Mol. Biol. 13 (2006) 915–920.

[34] A.B. Datta, G.L. Hura, C. Wolberger, The structure and conformation of Lys63-linked tetraubiquitin, J. Mol. Biol. 392 (2009) 1117–1124.

[35] J.D. Gawronski, D.R. Benson, Microtiter assay for glutamine synthetasebiosynthetic activity using inorganic phosphate detection, Anal. Biochem.327 (2004) 114–118.

[36] D.F. McCain, Z.Y. Zhang, Assays for protein-tyrosine phosphatases, MethodsEnzymol. 345 (2002) 507–518.

[37] J.K. Heinonen, R.J. Lahti, A new and convenient colorimetric determination ofinorganic orthophosphate and its application to the assay of inorganicpyrophosphatase, Anal. Biochem. 113 (1981) 313–317.

[38] A.D. Capili, C.D. Lima, Structure and analysis of a complex between SUMO andUbc9 illustrates features of a conserved E2–Ubl interaction, J. Mol. Biol. 369(2007) 608–618.

[39] M.C. Rodrigo-Brenni, S.A. Foster, D.O. Morgan, Catalysis of lysine 48-specificubiquitin chain assembly by residues in E2 and ubiquitin, Mol. Cell 39 (2010)548–559.

[40] O.H. Lowry, J.A. Lopez, The determination of inorganic phosphate in thepresence of labile phosphate esters, J. Biol. Chem. 162 (1946) 421–428.

[41] E.S. Baginski, P.P. Foa, B. Zak, Determination of phosphate—study of labileorganic phosphate interference, Clin. Chim. Acta 15 (1967) 155–158.

[42] R.M. Hofmann, C.M. Pickart, Noncanonical MMS2-encoded ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin chains forDNA repair, Cell 96 (1999) 645–653.

[43] C. Tsui, A. Raguraj, C.M. Pickart, Ubiquitin binding site of the ubiquitin E2variant (UEV) protein Mms2 is required for DNA damage tolerance in the yeastRAD6 pathway, J. Biol. Chem. 280 (2005) 19829–19835.

[44] M.D. Petroski, X. Zhou, G. Dong, S. Daniel-Issakani, D.G. Payan, J. Huang,Substrate modification with lysine 63-linked ubiquitin chains through theUBC13–UEV1A ubiquitin-conjugating enzyme, J. Biol. Chem. 282 (2007)29936–29945.

[45] C.M. Carlile, C.M. Pickart, M.J. Matunis, R.E. Cohen, Synthesis of free andproliferating cell nuclear antigen-bound polyubiquitin chains by the RING E3ubiquitin ligase Rad5, J. Biol. Chem. 284 (2009) 29326–29334.

[46] M.J. Lewis, L.F. Saltibus, D.D. Hau, W. Xiao, L. Spyracopoulos, Structural basisfor non-covalent interaction between ubiquitin and the ubiquitin conjugatingenzyme variant human MMS2, J. Biomol. NMR 34 (2006) 89–100.

[47] A. Fersht, Structure and Mechanism in Protein Science: A Guide to EnzymeCatalysis and Protein Folding, Freeman, New York, 1999.

[48] Y.P. Kim, Z. Jin, E. Kim, S. Park, Y.H. Oh, H.S. Kim, Analysis of in vitroSUMOylation using bioluminescence resonance energy transfer (BRET),Biochem. Biophys. Res. Commun. 382 (2009) 530–534.