Identification of an Orthogonal Peptide Binding Motif for Biarsenical

Bioconjugate Chem. 2007, 18, 1259−1265

1259

Identification of an Orthogonal Peptide Binding Motif for Biarsenical Multiuse Affinity Probes Baowei Chen, Haishi Cao, Ping Yan, M. Uljana Mayer, and Thomas C. Squier* Cell Biology and Biochemistry Group, Biological Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352. Received December 18, 2006; Revised Manuscript Received April 5, 2007

Biarsenical multiuse affinity probes (MAPs) complexed with ethanedithiol (EDT) permit the selective cellular labeling of proteins engineered with tetracysteine motifs, but are limited by the availability of a single binding motif (i.e., CCPGCC or PG tag) that prevents the differential labeling of coexpressed proteins. To overcome this problem, we have used a high-throughput peptide screen to identify an alternate binding motif (i.e., CCKACC or KA tag), which has a similar brightness to the classical sequence upon MAP binding, but displays altered rates and affinities of association that permit the differential labeling of these peptide sequences by the red probe 4,5-bis(1,3,2-dithiarsolan-2-yl)-resorufin (ReAsH-EDT2) or its green cognate 4′,5′-bis(1,3,2-dithoarsolan-2-yl)fluorescein (FLAsH-EDT2). The utility of this labeling strategy was demonstrated following the expression of PG- and KA-tagged subunits of RNA polymerase in E. coli. Specific labeling of two subunits of RNA polymerase in cellular lysates was achieved, whereby ReAsH-EDT2 is shown to selectively label the PG-tag on RNA polymerase R-subunit prior to the labeling of the KA-tag sequence of the β-subunit of RNA polymerase with FlAsH-EDT2. These results demonstrate the ability to selectively label multiple individual proteins with orthogonal sequence tags in complex cellular lystates with spectroscopically distinct MAPs, and indicate the absolute specificity of ReAsH to target expressed proteins with essentially no nonspecific binding interactions.

INTRODUCTION (MAPs1)

Biarsenical multiuse affinity probes based on a xanthene scaffold permit the selective labeling of expressed proteins that contain a tetracysteine motif and have been used for the affinity isolation of individual proteins and associated complexes, for the biophysical characterization of protein structure and binding interactions, and for cellular imaging and turnover measurements of protein expression (1-10). Specific labeling is accomplished through the chelation of each arsenic with a pair of vicinal cysteines within a tetracysteine tag sequence, which is uncommon within endogenously expressed proteins (4). Important to these applications has been (1) the ability to stabilize the arsenics through their reversible association with ethanedithiol (EDT) in order to enhance membrane permeability, decrease nonspecific binding, and protect the arsenic from oxidation and polymerization, and (2) the identification of an optimal 9-12 amino acid high-affinity binding motif involving a tetracoordinate linkage between the biarsenical and the vicinal thiols associated with a PG-tag sequence (i.e., a CCPGCC tag) (11, 12). The commercially available red probe 4,5-bis(1,3,2-dithiarsolan-2-yl)-resorufin (ReAsH-EDT2) and its green cognate 4′,5′-bis(1,3,2-dithoarsolan-2-yl)fluorescein (FlAsH-EDT2) both bind with high affinity to the PG-tag sequence, permitting the use of competing dithiol reducing agents to minimize nonspecific binding (12). * Correspondence should be addressed to: Thomas C. Squier, Pacific Northwest National Laboratory; P.O. Box 999; Mail Stop P7-53; Richland, WA 99352; e-mail: [email protected]; Tel: (509) 3762218; FAX: (509) 372-1632. 1 Abbreviations: β-ME, beta-mercaptoethanol; DTT, dithiotreitol; EDT, 1,2-ethanedithiol; FlAsH-EDT2, 4,5-bis(1,3,2-dithioarsolan-2-yl)fluorescein also known as fluorescein arsenical helix binder; HEPES, N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid); MAP, multiuse affinity probe; ReAsH-EDT2, 4,5-bis(1,3,2-dithioarsolan-2-yl)resorufin; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TCEP, tris(carboxyethyl)phosphine.

To date, only a single high-affinity peptide motif has been identified, preventing the simultaneous use of ReAsH-EDT2 and FlAsH-EDT2 to label multiple tag sequences for structural measurements of protein dynamics or cellular colocalization. In fact, current approaches involve the pairing of a tetracysteine tag with a cyanofluorescent protein (CFP) derived from the jellyfish Aequorea Victoria (13), whose relatively large size and slow folding kinetics have the potential to perturb both protein structure and cellular localization. ReAsH and FlAsH are synthesized as ethanedithiol adducts in order to protect the arsenic from oxidation and polymerization. In the initial study of the binding kinetics, the starting material was not this commercial form, but rather the arsenoxide (11). Since the reaction does not involve the thiol exchange necessary for the binding of the EDT chelates, the reported binding rates may overestimate the rates of reaction using the stable ReAsH-EDT2 and FlAsH-EDT2 compounds. Furthermore, it is commonly suggested that addition of 1,2-dithiol antidotes such as 1,2ethanedithiol (EDT) or 2,3-dimercaptopropanol (BAL) into the reaction mixture can facilitate the reduction of nonspecific binding to endogenous cysteine-rich proteins (12, 14), but the binding kinetics under conditions of competing dithiol ligand have not been reported. To extend the applicability of MAPs, we have used a surfacebound peptide library to identify useful protein-encoded tags in addition to CCPGCC that would permit the specific labeling of distinct tag sequences with both ReAsH-EDT2 and FlAsHEDT2. Through one round of selection, in which the two amino acids between the vicinal cysteines were substituted combinatorially, we identified an alternate binding sequence with the amino acids KA located between two pairs of vicinal cysteines (i.e., KA-tag), which has a similar brightness to the classical sequence upon MAP binding, but displays altered rates and affinities of association that permit the differential labeling of these peptide sequences engineered into the R- and β-subunits

10.1021/bc0603900 CCC: $37.00 © 2007 American Chemical Society Published on Web 06/15/2007

1260 Bioconjugate Chem., Vol. 18, No. 4, 2007

Chen et al.

of RNA polymerase by the red probe ReAsH-EDT2 and its green cognate FlAsH-EDT2.

MAP + Peptide Tag f MAP-Peptide complex + 2 EDT (3)

EXPERIMENTAL PROCEDURES

The dissociation constant, Kd, can be described as

Materials. 4,5-Bis(1,3,2-dithiarsolan-2-yl)-resorufin (ReAsHEDT2 or Lumio Red) and 4′,5′-bis(1,3,2-dithoarsolan-2-yl)fluorescein (FLAsH-EDT2 or Lumio Green) were synthesized as previously described (3, 11) or, in the case of FlAsH-EDT2, obtained commercially under the trade name Lumio Green (Invitrogen, Carlsbad, CA), N-(2-Hydroxyethyl)piperazine-N′(2-ethanesulfonic acid) (HEPES) and tris(carboxyethyl)phosphine (TCEP) were obtained from Sigma (St. Louis, MO). 2-Mercaptoethanol (β-ME) and 1,2-ethanedithiol (EDT) were obtained from Aldrich (Milwaukee, WI). The peptides, AREACCPGCCK-CONH2 (peptide PG) and AREACCKACCKCONH2 (peptide KA) were obtained from SynPep Corporation (Dublin, CA). The concentration of tetracysteine-containing peptide was measured as previously described (15). All other chemicals were the purest grade commercially available. Peptide Arrays. The peptide library was from SigmaGenosys (The Woodlands, Texas), which contained 208 different 15 amino acid acetylated peptides (5 nmol per spotted peptide) that were immobilized on a cellulose membrane through the C-terminus with the basic structure Ac-YYYYWDCCXXCCKAA. The YYYY sequence at the N-terminus was constrained to be GGGG, AAAA, or the originally proposed AREA sequence (11), while XX was systematically varied. Highaffinity binding sequences were identified following overnight incubation with FlAsH-EDT2 (5 nmol). Sequences of each peptide and identified peptides that became very bright upon binding FlAsH-EDT2 are provided in the Supporting Information. Peptide-Tag Labeling. Peptide tags (1 µM) were placed in 50 mM HEPES (pH 7.5), 140 mM KCl, 1 mM β-ME, and 1 mM TCEP at room temperature for 1 h to reduce any disulfide bonds prior to incubation at 4 °C with either ReAsH-EDT2 or FLAsH-EDT2. Fluorescence emission spectra of ReAsH-peptide or FlAsH-peptide complexes were measured using a Fluoro Max-2 fluorometer (SPEX, Edison, NJ), with excitation and emission slits set at 5 nm. In all cases, the sample temperature was 25 °C. Apparent rate constants were calculated from the second-order reaction

Ft [L]t)0 × (Fmax - Ft)

) kt

(1)

where Fmax is the fluorescence maximum, Ft is the fluorescence at time t, [L]t)0 is the concentration of fluorescent probe at time zero, k is the second-order rate constant, and t is the time of reaction. For equilibrium binding measurements, the extent of labeling was assessed following 24 h, and the concentration of peptide that is in association with the fluorescent probe, i.e., [FP]bound, can be calculated such that

[FP] )

[

]

|(F - F0)| ∆F × [F]total ) × [F]total (2) ∆Fmax |(Fmax - F0)|

where F0 is the initial fluorescence, Fmax is the fluorescence observed in the presence of saturating concentrations of peptide, and F is the observed fluorescence at a particular peptide concentration. [F]total is the concentration of FlAsH-EDT2 or ReAsH-EDT2 in solution. Determination of the binding affinities assumes a ligand exchange mechanism whereby the binding of the vicinal cysteines within the target peptides to the two arsenics on the MAPs involves the displacement of the bound EDT, as described by the following equilibrium.

Kd )

[F][P] [FP][EDT]2

(4)

where [F] and [P] represent the concentration of free dye and free peptide, respectively. [EDT] represents the concentration of ethanedithiol, which must be in large excess of the dye or peptide concentrations in order to calculate an apparent dissociation constant, Kapp, as given by

Kapp ) Kd[EDT]2 ) [F][P]/[FP]

(5)

When the reaction is at equilibrium, then

[F]total ) [F] + [FP]

(6)

[P]total ) [P] + [FP]

(7)

where [P]total is the concentration of added peptide. After rearranging, the concentration of the dye-peptide complex [FP] can be solved, where

[FP]2 - ([F]total + [P]total + Kapp)[FP] + [F]total[P]total ) 0 (8) [FP] ) {([F]total + [P]total + Kapp) - [([F]total + [P]total + Kapp)2 - 4[F]total[P]total]1/2}/2 (9) Substitution of [FP] (see eq 2) yields

∆F ) ∆Fmax {([F]total + [P]total + Kapp) - [([F]total + [P]total + Kapp)2 - 4[F]total[P]total]1/2}/2[F]total (10) Cloning and Expression of RNA Polymerase Subunits. A C-terminus 52 amino acid sequence was appended onto the R- or β-subunit of RNA polymerase (SO0256 or SO0257) from Shewanella oneidensis cloned into the pBAD/DTOPO 202 vector (Invitrogen) containing in the following order a tetracysteine tag (either AREACCPGCCK or AREACCKACCK) adjacent to a V5 epitope (KGGRADPAFLYKVVINSKLEGK PIPNPLLGL) and His6 sequence at the C-terminus, as previously described (5). The pBAD-TOPO expression system contains the promoter of the araBAD (arabinose) operon, which can be upregulated by arabinose and downregulated by glucose to permit optimal expression. Plasmids were maintained by addition of 20 µg/mL kanamycin. Following transformation of E. coli, cells were grown in flasks in standard Luria-Bertani growth media. Cells were grown until the optical density was 0.4-0.6, and PG-tagged R- or KA-tagged β-subunits were expressed near native levels (about 0.5 pM). Following overnight growth at 30 °C, cells were collected by centrifugation and lysed. Cell lysates (5 mg/mL) in 50 mM HEPES (pH 7.5), 140 mM NaCl, 2 mM β-ME, and 5 mM TCEP were labeled in the dark (due to the photoinstability of ReAsH; see Figure S1 in Supporting Information) with either FlAsHEDT2 or ReAsH-EDT2 (5 µM) for 20 min. Alternatively, equal amounts of lysate expressing either the PG-tagged R- or KAtagged β-subunits were mixed prior to their sequential labeling, which involved first addition of ReAsH-EDT2 (5 µM) for 20 min followed by the addition of FlAsH-EDT2 (5 µM) for an additional 20 min. Proteins were then separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-

Orthogonal Peptide Binding Motifs for Multicolor Measurements

Figure 1. Maximal fluorescence intensities (left panel) and selected emission spectra (right panel) for FlAsH-EDT2 (1 µM) in the presence of indicated thiol reagents (500 µM) or tetracysteine-containing peptide motifs (1 µM) in 50 mM HEPES (pH 7.5), 140 mM KCl, 1.0 mM EGTA, 1.0 mM β-ME, and 1.0 mM TCEP. Inset shows dependence of emission spectra on added DTT. Excitation was at 500 nm.

PAGE) and protein bands were visualized with GELCode Blue stain reagent (Pierce, Rockford, IL); the fluorescence was observed by Kodak IS2000MM illuminator with selected filters for excitation and emission.

RESULTS Identification and Fluorescence Properties of Multiuse Affinity Probes Bound to Target Peptide. Currently, two commercially available biarsenical dyes, i.e., ReAsH-EDT2 and FlAsH-EDT2, permit the labeling of tagged proteins in cells, but cannot be used in multicolor experiments due to the current limitation of a single class of tetracysteine peptide motifs that bind these probes with similar affinities. To circumvent this problem, we have used a combinatorial peptide screen to identify new peptide motifs that differentially bind these commercially available MAPs. Briefly, this involved selection of a cellulosederivatized peptide library containing 208 different peptide motifs for optimal candidate sequences that preferentially bind FlAsH-EDT2 to form a highly fluorescent complex. Following overnight incubation with limiting amounts of FlAsH-EDT2, 18 very bright peptides were identified that fall into 2 sequence classes corresponding to the previously identified CCPGCC motif (PG-peptide) (11, 12) and a motif with a basic amino acid at the first position between the vicinal cysteines (i.e., KA-peptide) (see Supporting Information). Following identification of these candidate peptides, we have measured the binding kinetics and affinities of ReAsH-EDT2 and FlAsH-EDT2 to these two peptide motifs. Since all prior measurements of MAP binding to target peptides involved unchelated MAP (i.e., no bound EDT), we have also measured the apparent affinities and reaction rates for both FlAsH-EDT2 and ReAsH-EDT2 against both the classical PG-peptide and the newly introduced KA-peptide in the presence of both EDT (suggested to enhance labeling specificity) and other thiols that are either present within cells or have previously been suggested to reduce nonspecific binding (9). As previously described (4, 11), there is a large (i.e., >20fold) increase in the fluorescence of FlAsH-EDT2 or ReAsHEDT2 upon binding the PG-peptide (Figure 1) (see Supporting Information for photostability data). In the presence of a large excess of 1,2-ethanedithiol (EDT), 2,3-dimercapto-1-propanesulfonic acid (DMPS), cysteine, or glutathione, there is little change in the fluorescence. A similar large increase in fluorescence is observed upon binding to the KA-peptide or the dithiol DTT (albeit with a substantially reduced affinity; Kd ) 0.1 mM; Figure 1), whose intensities are approximately 91% and 38% of that associated with binding to the PG-peptide. These latter

Bioconjugate Chem., Vol. 18, No. 4, 2007 1261

results indicate that DTT cannot be used as a reducing agent in the application of biarsenical dyes due to substantially higher background fluorescence. A much smaller increase is observed in the presence of β-mercaptoethanol (β-ME), which has been suggested to reduce nonspecific binding in cell imaging measurements with reduced toxicity in comparison to the application of EDT (9). The fluorescence enhancement upon binding to DTT further suggests the possible structural underpinning associated with the large fluorescence enhancements associated with binding to target peptides in comparison to other thiols. Our results suggest a relationship between the ring size between the arsenic atoms in FlAsH and the sulfur ligands. On the basis of the structure of thiol reagents, EDT and DMPS can form a complex with the arsenic in FlAsH to form a rigid five-membered ring that has a very low fluorescence. In contrast, following derivatization with DTT, a seven-membered ring is formed that has an intermediate fluorescence. Upon liganding to either PGtag or KA-tag sequences, a nine-membered ring can form that is associated with the large fluorescence enhancement (Figure 1). Consistent with this model, there is a small increase in fluorescence for FlAsH in the presence of 0.5 mM β-ME that is significantly higher than observed in the presence of either EDT or DMPS. These results are consistent with the known hydrogen bonding character between proximal β-ME that can form upon the liganding of two β-ME with the arsenic atoms of FlAsH to form a transient nine-membered ring. Thus, our results demonstrate that binding of the small mono- and dithiols (i.e., β-ME and DTT) to FlAsH result in an enhanced fluorescence that, in the case of DTT, is comparable to that following peptide tag binding and correlates strongly with increases in ring size. We suggest that arsenic is capable of quenching fluorescein fluorescence, and that ability, in turn, is modulated by the ring strain induced by the arsenic liganding. This is in contrast to prior suggestions that attributed the low fluorescence of FlAsH-EDT2 in comparison to the peptide bound dye to the greater mobility of the arsenic atoms bound to the small ligands that would act to “quench the fluorescence of fluorescein” (http://www.invitrogen.com/content/sfs/manuals/ lumiogreendetection_man.pdf). Rates of ReAsH- or FlAsH-Peptide Complex Formation. Under conditions in which there is a large molar excess of PGor KA-peptide tag relative to ReAsH-EDT2 or FlAsH-EDT2, there is a linear relationship between the pseudo-first-order constant associated with fluorescent probe binding and the concentration of peptide, which is indicative of a reaction mechanism involving second-order kinetics (data not shown). Using equimolar stoichiometries of fluorescent probes and peptides, we find that ReAsH-EDT2 preferentially binds to the PG-peptide with a nearly 50% rate enhancement in comparison to the KA-peptide (Figure 2; Table 1). In comparison, the rates of FlAsH-EDT2 binding to these peptide tags are very similar. These results indicate that ReAsH-EDT2 labeling is sensitive to the conformation of the peptide binding motif in a manner that is not possible for the bigger FlAsH-EDT2 moiety. Binding of Tetracysteine-Containing Peptides with ReAsHEDT2 and FlAsH-EDT2. Peptide binding to ReAsH-EDT2 or FlAsH-EDT2 was measured using associated increases in the fluorescence intensity upon formation of the dye-peptide complex in the absence and presence of excess amounts of the competing dithiol EDT, whose presence diminishes the apparent affinity and permits an accurate determination of an apparent binding affinity (see eq 3 in Experimental Procedures). In all cases, it is apparent that FlAsH-EDT2 binds to all peptide sequences with a higher affinity than ReAsH-EDT2, and that the presence of EDT results in a reduction in the apparent affinity. In the absence of unliganded EDT, binding to both the

1262 Bioconjugate Chem., Vol. 18, No. 4, 2007

Chen et al.

Figure 2. Binding kinetics of ReAsH-EDT2 or FlAsH-EDT2 (1 µM) to peptides (1 µM) AREACCPGCCK-CONH2 (peptide PG) (solid line; closed circles) or AREACCKACCK-CONH2 (peptide KA) (dashed line; open circles). In all cases, the reaction mixture contained 50 mM HEPES (pH 7.5), 140 mM KCl, 1.0 mM EGTA, 1.0 mM β-ME, and 1.0 mM TCEP. Lines represent fits to eq 1, where for ReAsH-binding to PG and KA peptides, k ) 12.5 ( 0.1 × 103 M-1 s-1 and 8.47 ( 0.07 × 103 M-1 s-1. For FlAsH-binding to PG and KA peptides, k ) 124 ( 8 × 103 M-1 s-1 and k ) 143 ( 8 × 103 M-1 s-1 (KA peptide). Fluorescence was measured at 530 and 608 nm with excitation at 500 and 593 nm for FlAsH and ReAsH, respectively. Table 1. Binding Rates and Apparent Affinities of FlAsH-EDT2 and ReAsH-EDT2 to Peptide Tags complex

k (M-1 s-1)a

Kapp (µM)b

ReAsH-Peptide PG ReAsH-Peptide KA FlAsH-Peptide PG FlAsH-Peptide KA

12 500 ( 1000 8470 ( 700 124 000 ( 8000 143 000 ( 8000

1.3 ( 0.2 3.3 ( 1.2 0.24 ( 0.09 0.85 ( 0.23

a The second-order rate constant obtained from fitting the data in Figure 2 using eq 1 in Experimental Procedures. b Apparent dissociation constant in presence of 200 µM EDT obtained from fitting data in Figure 3 to eq 10 in Experimental Procedures, where Kd ) Kapp/[EDT]2.

Figure 3. Equilibrium binding of ReAsH-EDT2 (A, B) or FlAsH-EDT2 (C, D) to peptides AREACCPGCCK-CONH2 (peptide PG) or AREACCKACCK-CONH2 (peptide KA) in the absence (O) or presence (b) of 200 µM EDT. Buffer contained 50 mM HEPES (pH 7.5), 140 mM KCl, 1.0 mM EGTA, 1.0 mM β-ME, and 1.0 mM TCEP at 25 °C. Dashed and solid lines represent fits to eq 10 in Experimental Procedures, assuming that respective concentrations of unliganded EDT in solution are negligible and 200 µM.

PG- and KA-peptide is essentially complete in the presence of equimolar amounts of FlAsH-EDT2 and either peptide (Figure 3), consistent with prior estimates regarding the high-affinity binding of the unliganded bis-arsenoxide FlAsH probe to this

Figure 4. SDS-PAGE of cellular lysates corresponding to total protein gel-code blue protein stain (A) or fluorescence signals (B) associated with ReAsH (lanes 1, 3, and 5; λex ) 555 nm, λem ) 600 nm) or FlAsHlabeling (lanes 2, 4, and 6; λex ) 465 nm, λem ) 535 nm) of cellular lysates containing the R-subunit of RNA polymerase containing an engineered PG-tag (i.e., RpoA), the β-subunit of RNA polymerase containing an engineered KA-tag (i.e., RpoB), or equimolar amounts of both tagged proteins (i.e., RpoA+B). Labeling protocols are fully described in Experimental Procedures. Relative mobilities of molecular mass standards shown at left, and positions of RpoA, RpoB, and SlyD on SDS-PAGE obtained from authentic standards are indicated.

peptide, whose Kd has been estimated to be approximately 10 pM (11). A quantitative determination of differences in binding affinities is possible in the presence of excess EDT, as the presence of EDT reduces the apparent dissociation constant (i.e., Kapp ) Kd × [EDT]2). The reduction in the apparent dissociation constant (Kapp) in the presence of EDT permits the accurate measurement of differences in binding affinities for the different peptides, and in the presence of excess EDT, apparent dissociation constants can be obtained (Table 1). In the presence of 200 µM EDT, differences in FlAsH-EDT2 binding to the PGand KA-peptides are revealed, where complete binding of FlAsH-EDT2 requires a 2- and 6-fold molar excess of the PGand KA-peptide (Figure 3). Likewise, complete binding of ReAsH-EDT2 is observed in the presence of an 8- and 20-fold molar excess of the PG- and KA-peptides (data not shown for the KA-peptide). These differences reflect 3-4-fold differences in binding affinities (Table 1), and suggest a strategy to selectively label these peptide motifs with these complementary dyes. Selective Labeling of Tagged Subunits in RNA Polymerase. Tagged R- and β-subunits of RNA polymerase (RNAP) complex were respectively expressed in E. coli with the PG- and KA-tags located near their C-termini at nearly native levels of expression (about 0.5 pM), essentially as previously described (5, 16). The specificity of ReAsH-EDT2 and FlAsH-EDT2 binding to individual proteins was determined following expression of either the PG-tagged R-subunit (RpoA) or KA-tagged β-subunit (RpoB) of RNA polymerase (Figure 4). Prior to labeling, cell lysates (5 mg protein/mL) were incubated with either ReAsH-EDT2 or FlAsH-EDT2 (5 µM) in buffer containing 50 mM HEPES (pH 7.5), 140 mM NaCl, 2 mM β-mercaptoethanol, and 5 mM TCEP for 20 min prior to separation and visualization of labeling specificity using SDSPAGE. ReAsH-EDT2 and FlAsH-EDT2 both label RpoA tagged with the PG-tag, which migrates as a 40 kDa band on SDS-PAGE (Figure 4). However, while ReAsH-EDT2 specifically labels RpoA, we find that FlAsH-EDT2 binds to a number of additional proteins, with apparent molecular masses 28 kDa, 39 kDa, and 100 kDa. These results indicate that both ReAsH-EDT2 and FlAsH-EDT2 permit the labeling of expressed proteins in cellular lysates, but that ReAsH-EDT2 has the required affinity and selectivity for applications involving the fluorescent labeling of low abundance proteins. Complementary measurements

Orthogonal Peptide Binding Motifs for Multicolor Measurements

assessed the ability of both ReAsH-EDT2 and FlAsH-EDT2 to bind the KA-peptide tag engineered onto the β-subunit of RNA polymerase (RpoB). Consistent with measurements using the PG-tag on RpoA, we find that ReAsH-EDT2 and FlAsH-EDT2 both label RpoB, which migrates with an apparent mass of about 160 kDa (Figure 4). However, in this application, ReAsH-EDT2 and FlAsH-EDT2 bind several additional proteins, including the prominent band at about 28 kDa corresponding to the thiolrich chaperone SlyD protein (17). Multicolor Labeling of RNA Polymerase Subunits. Differences in the affinities and kinetics of ReAsH-EDT2 binding to the PG- and KA-peptides suggest a strategy to differentially label subunits of RNA polymerase. As a proof of principle, lysate expressing PG-tagged R- and KA-tagged β-subunits of RNA polymerase (i.e., RpoA and RpoB) were mixed, and the R- and β-subunits were then sequentially labeled. First, ReAsHEDT2 (5 µM) is used to modify the PG-tag sequence on the R-subunit of RNA polymerase (RpoA), which is possible due to the 50% rate enhancement and 3-fold increase in binding affinity (i.e., 1/Kapp) of the PG-tag in comparison to the KAtag (Table 1). Second, following the selective labeling of the PG-tag sequence, FlAsH-EDT2 (5 µM) was used to selectively label the KA-tag sequence on the β-subunit of RNA polymerase (RpoB). Using this strategy, specific labeling of two subunits of RNA polymerase in cellular lysates from E. coli was achieved (Figure 4), whereby ReAsH-EDT2 specifically labels the RNA polymerase R-subunit RpoA engineered with a PG-tag in the presence of an equimolar mixture of RpoA and RpoB (engineered with the KA-tag). Following labeling of RpoA, the KAtag sequence of the RNA polymerase β-subunit RpoB can then be labeled with FlAsH-EDT2. These results demonstrate the ability to differentially label multiple proteins engineered with orthogonal sequence tags in complex cellular lystates with spectroscopically distinct MAPs.

DISCUSSION Summary of Results. Using a combinatorial approach involving a peptide array, we have identified an orthogonal peptide tag sequence (i.e., CCKACCK-CONH2 or KA-tag) that in combination with the classical PG-tag sequence permits the use of the commercially available cell permeable MAPs ReAsHEDT2 (Lumio Red) and FlAsH-EDT2 (Lumio Green) for multicolor measurements in which different proteins engineered with either tag can be differentially labeled. Upon binding to either peptide-tag, similar levels of brightness are observed (Figure 1). In addition, we have for the first time characterized the affinities and rates of binding of the commercially available and stable forms of these dyes. As expected, the chelation of ReASH or FlAsH with EDT reduces the rates of protein modification, where complete labeling requires approximately 10 min (Figure 2). Further, the commonly suggested inclusion of EDT to reduce nonspecific labeling (12) results in a substantial reduction in the apparent binding affinity (Figure 3). Indeed, in our hands, labeling of tagged proteins in cellular lysates is minimal in the presence of amounts of EDT suggested to be necessary to reduce nonspecific binding. Rather, we find that optimal amounts and specificities of labeling are observed in the presence of both 2 mM β-ME and 5 mM TCEP (but no competing EDT) (Figure 4). Under the latter conditions, we have demonstrated the ability to specifically modify a PG-tagged R-subunit of RNA polymerase (i.e., RpoA) with ReAsH-EDT2 in the presence of an equimolar amount of the β-subunit of RNA polymerase engineered with the KA-tag (see lane 5 in Figure 4). This specificity of labeling permits the subsequent modification of the KA-tag on the β-subunit using FlAsH-EDT2 for multicolor measurements.

Bioconjugate Chem., Vol. 18, No. 4, 2007 1263

Genetic Tagging and Cell Permeable Multiuse Affinity Probes. To understand the expression, localization, and fate of proteins in live cells, current approaches emphasize the engineering of GFP mutants with improved properties and altered colors that provide the basic tools to investigate more complex processes in live cells (18-20). However, because of the large size of GFP and its slow folding kinetics, it has the potential to alter the function, binding interactions, and localization of tagged proteins (21). Furthermore, in bacteria GFP labels are often incompatible with cellular export systems associated with the trafficking of proteins into the periplasm. The development of small molecular probes, i.e., the biarsenical dyes FlAsH and ReAsH, have largely overcome these prior limitations (3, 4, 11, 22, 23). The specific ligand for the labeling reaction is the small tetracysteine binding motif (i.e., CCXXCC), whose sequence is readily engineered within the gene encoding the protein of interest. Indeed, the small tag can be placed within the protein of interest to minimize the physical interference associated with N- or C-terminal tags, and to allow direct functional measurements of protein conformational change and binding interactions (1, 2, 4). Recent improvements in the brightness and selectivity of FlAsH-EDT2 analogs offer the potential to image low-abundance proteins in living cells (24, 25). Nevertheless, the prior recognition of only a single binding motif has limited the application of multiuse affinity probes to imaging the cellular locations or turnover of individual proteins (8, 20), while the identification of protein-protein interactions requires the use of Aequoreaderived fluorescent proteins such as YFP (13). In this respect, the identification of orthogonal tags using a combinatorial screen that permit the simultaneous use of both commercially available MAPS (i.e., ReAsH-EDT and FlAsH-EDT2) to selectively modify the R- and β-subunits of RNA polymerase tagged, respectively, with PG- and KA-tags extends the range of possible applications of these cell-permeable reagents to selectively label multiple proteins to assess protein-protein interactions using, for example, fluorescence resonance energy transfer or colocalization single-molecule measurements in cell lysates. This strategy permits the validation of protein-protein interactions identified using high-throughput screens and mass spectrometry, when specific pairwise interactions are not certain. Thus, upon tagging multiple proteins, their association in a protein complex can be directly assessed using, for example, single-molecule methods that measure the association between tagged proteins to validate their interactions without the need to try to purify specific binding partners. In addition, these reagents permit fluorophore-assisted light inactivation (FALI) to selectively knock-out targeted protein activities to aid in the identification of their functions as well as the modulation of metabolic networks (26-28). However, additional specificity for cellular measurements will require the introduction of new biarsenical probes with different spacings between the arsenic ligands, providing greater labeling specificity. Practical Considerations. Observed rates of binding for the commercially available and more stable reagents ReAsH-EDT2 or FlAsH-EDT2 to tag sequences are substantially slower than reported for the unliganded probes, which bind to peptide tags on the time scale of seconds, and require approximately 10 min for stoichiometric labeling due to the requirement for ligand exchange (Figure 2). Furthermore, the high-affinity association of the unliganded dyes with peptide tags (Kd ) 10 pM) is substantially reduced in the presence of added EDT (Figure 3) (11). In the presence of 200 µM EDT, which has previously been suggested to minimize nonspecific binding (12), we find that apparent binding affinities are in the micromolar range (Table 1). Indeed, under the latter conditions, we observe low stoichiometries of peptide-tag modification in cellular lysates

1264 Bioconjugate Chem., Vol. 18, No. 4, 2007

(data not shown), limiting the ability of these reagents to image highly abundant proteins. In contrast, using other reducing reagents (e.g., β-mercaptoethanol and TCEP), we are able to selectively modify tagged proteins with nearly equimolar stoichiometries of either ReAsH-EDT2 or FlAsH-EDT2, permitting both enhanced detection of low-abundance proteins and multicolor experiments involving the differential labeling of PGand KA-tags engineered onto different proteins. In this application, we take advantage of the substantially larger rates of labeling and binding affinity of ReAsH-EDT2 binding against the PG-peptide tag in comparison to the KA-tag (Figure 4). Following the stoichiometric labeling of PG-tagged protein (RpoA) with ReAsH-EDT2, it is then possible to label the KAtagged protein (RpoB) with the higher-affinity FlAsH-EDT2 probe with minimal cross-reactivity. Conclusions and Future Directions. The identification of orthogonal peptide tags permits the simultaneous application of the commercially available and stable biarsenical dyes ReAsH-EDT2 and FlAsH-EDT2 to differentially label multiple proteins, providing a path forward for their use in multicolor measurements relating to protein associations and colocalization. Specificity of labeling is dependent on the reduction of peptide tags with the cell-permeant reducing agents β-mercaptoethanol and tris(carboxyethyl)phosphine (TCEP), which act to enhance binding to peptide tags without significantly affecting the binding affinity. These differences in binding specificities were demonstrated using a single peptide scaffold in which vicinal cysteines were separated by two amino acids whose sequence was altered, indicating that modest conformational differences in the binding sequence of the peptide tag can promote differences in binding specificity. These results suggest that further variation of the scaffold to vary the spatial separation between the biarsenical linkages within the probes will provide the required specificity to allow the generation of a family of cell-permeant affinity probes that permit simultaneous labeling and isolation of multiple tagged proteins, using peptide tags with variable spacing between the vicinal cysteine binding ligands.

ACKNOWLEDGMENT This work was supported by the U.S. Department of Energy, Office of Science Genomics: GTL project. Pacific Northwest National Laboratory is operated for the U.S. Department of Energy by Battelle Memorial Institute under contract DE-AC0576RLO 1830. Supporting Information Available: Peptide sequences using an initial screen to identify orthogonal binding sequences for FlAsH-EDT2 and ReAsH-EDT2, as well as their photostability in the presence of room light levels (Figure S1). This material is available free of charge via the Internet at http:// pubs.acs.org.

LITERATURE CITED (1) Chen, B., Mayer, M. U., Markillie, L. M., Stenoien, D. L., and Squier, T. C. (2005) Dynamic motion of helix A in the aminoterminal domain of calmodulin is stabilized upon calcium activation. Biochemistry 44, 905-914. (2) Chen, B., Mayer, M. U., and Squier, T. C. (2005) Structural uncoupling between opposing domains of oxidized calmodulin underlies the enhanced binding affinity and inhibition of the plasma membrane Ca-ATPase. Biochemistry 44, 4737-4747. (3) Griffin, B. A., Adams, S. R., Jones, J., and Tsien, R. Y. (2000) Fluorescent labeling of recombinant proteins in living cells with FlAsH. Methods Enzymol. 327, 565-578. (4) Griffin, B. A., Adams, S. R., and Tsien, R. Y. (1998) Specific covalent labeling of recombinant protein molecules inside live cells. Science 281, 269-272.

Chen et al. (5) Mayer, M. U., Shi, L., and Squier, T. C. (2005) One-step, nondenaturing isolation of an RNA polymerase enzyme complex using an improved multi-use affinity probe resin. Mol. Biosys. 1, 53-56. (6) Thorn, K. S., Naber, N., Matuska, M., Vale, R. D., and Cooke, R. (2000) A novel method of affinity-purifying proteins using a bisarsenical fluorescein. Protein Sci. 9, 213-217. (7) Shi, L., Chen, B., Wang, Z., Elias, D. A., Mayer, M. U., Gorby, Y. A., Ni, S., Lower, B. H., Kennedy, D. W., Wunschel, D. S., Mottaz, H. M., Marshall, M. J., Hill, E. A., Beliaev, A. S., Zachara, J. M., Fredrickson, J. K., and Squier, T. C. (2006) Isolation of a high-affinity functional protein complex between OmcA and MtrC: Two outer membrane decaheme c-type cytochromes of Shewanella oneidensis MR-1. J. Bacteriol. 188, 4705-4714. (8) Gaietta, G., Deerinck, T. J., Adams, S. R., Bouwer, J., Tour, O., Laird, D. W., Sosinsky, G. E., Tsien, R. Y., and Ellisman, M. H. (2002) Multicolor and electron microscopic imaging of connexin trafficking. Science 296, 503-507. (9) Langhorst, M. F., Genisyuerek, S., and Stuermer, C. A. (2006) Accumulation of FlAsH/Lumio Green in active mitochondria can be reversed by beta-mercaptoethanol for specific staining of tetracysteine-tagged proteins. Histochem. Cell Biol. 125, 743747. (10) Panchal, R. G., Ruthel, G., Kenny, T. A., Kallstrom, G. H., Lane, D., Badie, S. S., Li, L., Bavari, S., and Aman, M. J. (2003) In vivo oligomerization and raft localization of Ebola virus protein VP40 during vesicular budding. Proc. Natl. Acad. Sci. U.S.A. 100, 1593615941. (11) Adams, S. R., Campbell, R. E., Gross, L. A., Martin, B. R., Walkup, G. K., Yao, Y., Llopis, J., and Tsien, R. Y. (2002) New biarsenical ligands and tetracysteine motifs for protein labeling in vitro and in vivo: synthesis and biological applications. J. Am. Chem. Soc. 124, 6063-6076. (12) Martin, B. R., Giepmans, B. N., Adams, S. R., and Tsien, R. Y. (2005) Mammalian cell-based optimization of the biarsenical-binding tetracysteine motif for improved fluorescence and affinity. Nat. Biotechnol. 23, 1308-1314. (13) Hoffmann, C., Gaietta, G., Bunemann, M., Adams, S. R., Oberdorff-Maass, S., Behr, B., Vilardaga, J. P., Tsien, R. Y., Ellisman, M. H., and Lohse, M. J. (2005) A FlAsH-based FRET approach to determine G protein-coupled receptor activation in living cells. Nat. Methods 2, 171-176. (14) Stroffekova, K., Proenza, C., and Beam, K. G. (2001) The proteinlabeling reagent FLASH-EDT2 binds not only to CCXXCC motifs but also non-specifically to endogenous cysteine- rich proteins. Pflugers ArchiV. 442, 859-866. (15) Gill, S. C., and von Hippel, P. H. (1989) Calculation of protein extinction coefficients from amino-acid sequence data. Anal. Biochem. 182, 319-326. (16) Verma, S., Xiong, Y., Mayer, M. U., and Squier, T. C. (2007) Remodeling of bacterial RNA polymerase supramolecular complex in response to environmental conditions. Biochemistry 46, 30233035. (17) Scholz, C., Eckert, B., Hagn, F., Schaarschmidt, P., Balbach, J., and Schmid, F. X. (2006) SlyD proteins from different species exhibit high prolyl isomerase and chaperone activities. Biochemistry 45, 2033. (18) Tsien, R. Y. (1998) The green fluorescent protein. Annu. ReV. Biochem. 67, 509-544. (19) Zacharias, D. A., and Tsien, R. Y. (2006) Molecular biology and mutation of green fluorescent protein. Methods Biochem. Anal. 47, 83-120. (20) Giepmans, B. N., Adams, S. R., Ellisman, M. H., and Tsien, R. Y. (2006) The fluorescent toolbox for assessing protein location and function. Science 312, 217-224. (21) Andresen, M., Schmitz-Salue, R., and Jakobs, S. (2004) Short tetracysteine tags to beta-tubulin demonstrate the significance of small labels for live cell imaging. Mol. Biol. Cell 15, 56165622. (22) Zhang, J., Campbell, R. E., Ting, A. Y., and Tsien, R. Y. (2002) Creating new fluorescent probes for cell biology. Nat. ReV. Mol. Cell Biol. 3, 906-918. (23) Gaietta, G. M., Giepmans, B. N., Deerinck, T. J., Smith, W. B., Ngan, L., Llopis, J., Adams, S. R., Tsien, R. Y., and Ellisman, M. H. (2006) Golgi twins in late mitosis revealed by genetically encoded

Bioconjugate Chem., Vol. 18, No. 4, 2007 1265

Orthogonal Peptide Binding Motifs for Multicolor Measurements tags for live cell imaging and correlated electron microscopy. Proc. Natl. Acad. Sci. U.S.A. 103, 17777-17782. (24) Cao, H., Chen, B., Squier, T. C., and Mayer, M. U. (2006) CrAsH: a biarsenical multi-use affinity probe with low non-specific fluorescence. Chem. Commun. (Cambridge, U.K.) 2601-2603. (25) Spagnuolo, C. C., Vermeij, R. J., and Jares-Erijman, E. A. (2006) Improved photostable FRET-competent biarsenical-tetracysteine probes based on fluorinated fluoresceins. J. Am. Chem. Soc. 128, 12040-12041. (26) Yan, P., Xiong, Y., Chen, B., Negash, S., Squier, T. C., and Mayer, M. U. (2006) Fluorophore-assisted light inactivation of calmodulin

involves singlet-oxygen mediated cross-linking and methionine oxidation. Biochemistry 45, 4736-4748. (27) Surrey, T., Elowitz, M. B., Wolf, P. E., Yang, F., Nedelec, F., Shokat, K., and Leibler, S. (1998) Chromophore-assisted light inactivation and self-organization of microtubules and motors. Proc. Natl. Acad. Sci. U.S.A. 95, 4293-4298. (28) Tour, O., Meijer, R. M., Zacharias, D. A., Adams, S. R., and Tsien, R. Y. (2003) Genetically targeted chromophore-assisted light inactivation. Nat. Biotechnol. 21, 1505-1508. BC0603900