Detection of G Proteins by Affinity Probe Capillary Electrophoresis

Anal. Chem. 2003, 75, 4297-4304

Detection of G Proteins by Affinity Probe Capillary Electrophoresis Using a Fluorescently Labeled GTP Analogue Emily E. Jameson,† Jennifer M. Cunliffe,† Richard R. Neubig,‡,§ Roger K. Sunahara,‡ and Robert T. Kennedy*,†,‡

Department of Chemistry, Department of Pharmacology, and Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan 48109

An affinity probe capillary electrophoresis (APCE) assay for guanine-nucleotide-binding proteins (G proteins) was developed using BODIPY FL GTPγS (BGTPγS), a fluorescently labeled GTP analogue, as the affinity probe. In the assay, BGTPγS was incubated with samples containing G proteins and the resulting mixtures of BGTPγS-G protein complexes and free BGTPγS were separated by capillary electrophoresis and detected with laser-induced fluorescence detection. Separations were completed in less than 30 s using 25 mM Tris, 192 mM glycine at pH 8.5 as the electrophoresis buffer and applying 555 V/cm over a 4-cm separation distance. BGTPγS-Gro peak heights increased linearly with Gro up to ∼200 nM using a 50 nM BGTPγS probe. The detection limit for Gro was 2 nM, corresponding to a mass detection limit of 3 amol. The high speed of the APCE assays allowed reaction kinetics and the dissociation constant (Kd) to be determined. The on-rate and off-rate of BGTPγS to Gro were 0.0068 ( 0.0004 and 0.000 23 ( 0.000 01 s-1, respectively. The half-life of the BGTPγS-Gro complex was 3060 ( 240 s and Kd was 8.6 ( 0.7 nM. The estimates of these parameters are in good agreement with those obtained using established techniques, indicating the suitability of this method for such measurements. Lowering the temperature of the separation improved the detection of the complex, allowing the assay to be performed on a commercial instrument with longer separation times. Additionally, the capability of the technique to detect several G proteins based on their binding to BGTPγS was demonstrated with assays for Gr and Gri1 and for Ras and Rab3A. The suitability of capillary electrophoresis (CE) for separation of noncovalent complexes has enabled the development of a suite of techniques that couple affinity interactions with CE.1-7 Affinity probe CE (APCE) is one such technique in which a fluorescently * To whom correspondence should be addressed: (phone) 734-615-4363. E-mail: [email protected]. † Department of Chemistry. ‡ Department of Pharmacology. § Department of Internal Medicine. (1) Heegaard, N. H. J. Mol. Recognit. 1998, 11, 141-148. (2) Heegaard, N. H.; Kennedy, R. T. Electrophoresis 1999, 20, 3122-3133. 10.1021/ac0342976 CCC: $25.00 Published on Web 07/15/2003

© 2003 American Chemical Society

labeled receptor (the affinity probe) is added to a sample and allowed to bind to a ligand which is the analyte.8-10 Separation and detection of bound and free affinity probe by CE with laserinduced fluorescence detection (LIF) allows determination of analyte concentration directly from the bound peak. APCE assays have large linear dynamic range (LDR), detection limits less than 10 nM, and the potential to detect families of analytes that all exhibit binding to the same affinity probe if the complexes can be resolved. In addition, if the separation is rapid compared to the dissociation time, the method can be used to determine dissociation constants (Kd). APCE has many potential applications including clinical assays, environmental analysis, and drug screening.1,2 Most previous APCE work has focused on the use of an affinity probe, such as an antibody8,9,11 or aptamer,12 that is artificially developed for binding a specific analyte. In addition, proteins have been use to detect drugs.13 Another route to development of affinity probes would be to use known biomolecular interactions. An attractive approach would be to utilize a small natural ligand or cofactor as the affinity probe to selectively label proteins in complex mixtures. By choosing native affinity probes that crossreact with multiple proteins, whole classes of proteins or other targets could be determined in one assay based on their affinity for a single ligand. Such assays could be of use in studies of signal transduction pathways, biochemical investigation of binding, or screening for inhibitors of binding. Despite the considerable potential of such native affinity probes, their use in APCE is relatively unexplored. Examples include use of labeled DNA to detect transcription factors (a CE version of the gel-mobility shift (3) Schmalzing, D.; Buonocore, S.; Piggee, C. Electrophoresis 2000, 21, 39193930. (4) Shimura, K.; Kasai, K.-I. J. Mol. Recognit. 1998, 2, 134-140. (5) Guijy-van Duijn, R. M.; Frank, J.; van Dedem, G. W.; Baltussen, E. Electrophoresis 2000, 21, 3905-3918. (6) Chu, Y.-H.; Avila, L. Z.; Gao, J.; Whitesides, G. M. Acc. Chem. Res. 1995, 28, 461-468. (7) Bao, J. J. Chromatogr., B 1997, 699, 463-480. (8) Hafner, F. T.; Kautz, R. A.; Iverson, B. L.; Tim, R. C.; Karger, B. L. Anal. Chem. 2000, 72, 5779-5786. (9) Shimura, K.; Karger, B. L. Anal. Chem. 1994, 66, 9-15. (10) Schultz, N. M.; Kennedy, R. T. Anal. Chem. 1993, 65, 3161-3165. (11) Attiya, S.; Dickinson-Laing, T.; Cesarz, J.; Giese, R. D.; Lee, W. E.; Mah, D.; Harrison, D. J. Electrophoresis 2002, 23, 750-758. (12) German, I.; Buchanan, D. D.; Kennedy, R. T. Anal. Chem. 1998, 70, 45404545. (13) Tim, R. C.; Kautz, R. A.; Karger, B. L. Electrophoresis 2000, 21, 220-226.

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003 4297

Figure 1. Structure of BODIPY FL GTPγS (BGTPγS).

assay),14,15 fluorescently labeled phosphatidylserine to probe its association with bovine serum albumin,16 and proteases to detect substrates.17 In this work, we demonstrate the potential of APCE for detecting of a class of proteins based on a native affinity interaction by using guanosine 5′-O-(3-thiotriphosphate) BODIPY FL thioether (BGTPγS), a fluorescent guanosine 5′-triphosphate (GTP) analogue (Figure 1), as an affinity probe for guaninenucleotide-binding proteins (G proteins). G proteins are important components of intracellular signaling pathways, coupling cell surface receptors to downstream effectors such as adenylyl cyclase and ion channels.18 They play vital roles in many cellular functions including exocytosis and cell growth.19,20 G proteins are divided into two distinct families. Heterotrimeric G proteins consist of two subunits, R and a βγ dimer, with the R subunit containing the guanine-nucleotide-binding site. In the inactive state, the protein is bound to guanosine 5′-diphosphate (GDP). Binding of agonist to a cell surface receptor promotes exchange of GDP for GTP, activating the protein and causing its dissociation into two componentssGβγ and GR-GTP, both of which can interact with downstream effectors to transmit signal. G protein signaling is terminated by hydrolysis of GTP to GDP via intrinsic GTPase activity of the R subunit. The other family of G proteins is the low molecular weight or Ras-like proteins. These G proteins are monomers that have a similar GDP/GTP exchange cycle for activation and deactivation. G protein analysis is usually performed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE),21-26 filter binding assays,27-29 or fluorescence assays.27,30-36 SDS(14) Xian, J.; Harrington, M. G.; Davidson, E. H. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 86-90. (15) Stebbins, M. A.; Hoyt, A. M.; Sepaniak, M. J.; Hurlburt, B. K. J. Chromatogr., B 1996, 683, 77-84. (16) Hu, S.; Zhang, L.; Dovichi, N. J. J. Chromatogr., A 2001, 924, 369-375. (17) Reif, O. W.; Freitag, R. J. Chromatogr., A 1995, 716, 363-369. (18) Gomperts, B. D.; Kramer, I. M.; Tatham, P. E Signal Transduction; Academic Press: San Diego, 2002; pp 107-125. (19) Gomperts, B. D.; Kramer, I. M.; Tatham, P. E. Signal Transduction; Academic Press: San Diego, 2002; pp 71-105. (20) Neer, E. J.; Clapham, D. E. Nature 1988, 333, 129-134. (21) Yang, C.-Q.; Kitamura, N.; Nishino, N.; Shirakawa, O.; Nakai, H. Biol. Psychiatry 1998, 43, 12-19. (22) Kowluru, A.; Rabaglia, M. E.; Muse, K. E.; Metz, S. A. Biochim. Biophys. Acta 1994, 1222, 348-359. (23) Petit, A.; Geoffroy, P.; Bessette, P.; Prevost, J.; Belisle, S. J. Soc. Gynecol. Invest. 1995, 2, 678-685. (24) Bhullar, R. P.; Haslam, R. J. Biochem. J. 1987, 245, 617-620. (25) Bhullar, R. P.; McCartney, D. G.; Kanfer, J. N. J. Neurosci. Res. 1999, 55, 80-86. (26) Kowluru, A.; Li, G.; Rabaglia, M. E.; Segu, V. B.; Hofmann, F.; Aktories, K.; Metz, S. A. Biochem. Pharm. 1997, 54, 1097-1108. (27) McEwen, D. P.; Gee, K. R.; Kang, H. C.; Neubig, R. R. Anal. Biochem. 2001, 291, 109-117. (28) Takeda, S.; Sugiyama, H.; Natori, S.; Sekimizu, K. FEBS Lett. 1989, 244, 469-472.

4298

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

PAGE, while providing molecular weight information, is plagued by analysis times of several hours and a requirement for relatively large amounts (7-30 µg) of protein. The technique can be tailored specifically for G proteins by the use of radiolabeled GTP to stain the gels;22 however, this method is limited to Ras-like G proteins, as heterotrimers lose their ability to bind GTP as a result of SDS treatment and nitrocellulose transfer.24 Another technique that utilizes radioactivity for the analysis of G proteins is the filter binding assay. In filter binding assays, radiolabeled GTP or GTPγS (guanosine 5′-O-(3-thiotriphosphate)) is allowed to bind to G proteins and the G protein complexes are separated from the free label using a protein-binding filter. Using this method, Kds can be measured28 and reaction rates probed.27 However, the radioactivity requirement and relatively slow analysis times combined with an inability to distinguish among different G proteins (unless additional selectivity is introduced) limits the utility of this technique. Fluorescence assays utilize either intrinsic protein fluorescence or GTP analogues conjugated with BODIPY FL or N-methyl-3′-O-anthraniloyl (MANT) to monitor G protein interactions with effectors, GTPase-activating proteins, and GDP/GTP exchange.27,30-38 The high temporal resolution of optical techniques permits the determination of reaction kinetics and affinity constants, though as with filter binding assays, proteins must be purified prior to being assayed to ensure that the signal observed is from the analyte of interest. Also, simultaneous, independent detection of multiple G proteins is not possible since no separation is involved. We demonstrate that APCE using BGTPγS as the affinity probe can be applied to mixtures of G proteins and can determine multiple analytes at the attomole level in a single assay. The high speed of the assay allows reaction kinetics and binding constants to be determined. (29) Mick, G. J.; Chun, K. Y.; VanderBloomer, T. L.; Fu, C.-L.; McCormick, K. L. Biochimi. Biophys. Acta 1998, 1384, 130-140. (30) Remmers, A. E.; Neubig, R. R. J. Biol. Chem. 1996, 271, 4791-4797. (31) Remmers, A. E.; Posner, R.; Neubig, R. R. J. Biol. Chem. 1994, 269, 1377113778. (32) Ahmadian, M. R.; Hoffman, U.; Goody, R. S.; Wittinghofer, A. Biochemistry 1997, 36, 4535-4541. (33) Lenzen, C.; Cool, R. H.; Prinz, H.; Kuhlmann, J.; Wittinghofer, A. Biochemistry 1998, 37, 7420-7430. (34) Leonard, D. A.; Evans, T.; Hart, M.; Cerione, R. A.; Manor, D. Biochemistry 1994, 33, 12323-12328. (35) Leonard, D. A.; Satoskar, R. S.; Wu, W.-J.; Bagrodia, S.; Cerione, R. A.; Manor, D. Biochemistry 1997, 36, 1173-1180. (36) Ferguson, K. M.; Higashijima, T.; Smigel, M. D.; Gilman, A. G. J. Biol. Chem. 1986, 261, 7393-7399. (37) Lan, K.-L.; Zhong, H.; Nanamori, M.; Neubig, R. R. J. Biol. Chem. 2000, 275, 33497-33503. (38) Remmers, A. E.; Engel, C.; Liu, M.; Neubig, R. R. Biochemistry 1999, 38, 13795-13800.

EXPERIMENTAL SECTION Chemicals. Unless stated otherwise, all chemicals, including proteins, used in the experiments were purchased from Sigma Chemical Co. (St. Louis, MO). Tris-glycine buffer (10×) was purchased from Bio-Rad Laboratories (Hercules, CA). BGTPγS was purchased from Molecular Probes (Eugene, OR). His6-GRo and r-myristolated GRi1 (which we will refer to as GRo and GRi1) were expressed and purified as previously described39,40 and stored at -80 °C until used. All buffers were made with deionized water purified by an E-Pure water system (Barnstead International Co., Dubuque, IA) and filtered prior to use. Anisotropy and Fluorescence Instrumentation. Initial GRo binding studies were performed on a Fluoromax-2 fluorometer (Jobin Yvon, Inc., Edison, NJ). Samples were made to the desired concentration (50 nM BGTPγS with 0-500 nM GRo) in 25 mM Tris, 192 mM glycine, with 1 mM EDTA, and 10 mM MgCl2 at pH 8.5 (TGEM) and immediately analyzed in a manner similar to that previously described.27 Fluorescence was excited at 488 nm, and emission was scanned from 500 to 550 nm. Slit widths were 5 nm for excitation and emission. Ras experiments were executed on a Fluorolog-τ-3 spectrofluorometer (Jobin Yvon, Inc.). Samples (2 µM BGTPγS with 0-3.2 µM Ras) were prepared in 20 mM Tris-HCl, 25 mM EDTA at pH 8.0 and incubated at room temperature for 30 min, at which time MgCl2 was added to 35 mM.30,32-35,38 Fluorescence was excited at 488 nm and emission scanned from 500 to 550 nm. Slit widths were 2 nm for excitation and emission. All experiments were carried out at room temperature. Flow-Gated CE. For rapid CE separations, a CE-LIF equipped with flow-gated injection similar to that described elsewhere was used.41,42 Briefly, a pressurized sample chamber was connected to the flow gate interface by a 35 cm × 50 µm inner diameter (i.d.), 360-µm outer diameter (o.d.) capillary (Polymicro Technologies, Phoenix, AZ). Sample loaded into the chamber was continuously delivered into the flow gate at a rate of 3.4 µL/min. The sample delivery capillary was situated 40 µm across from the separation capillary (9 cm × 10 µm i.d. with a length to detector (Ld) of 4 cm) in the flow gate. Electrophoresis buffer was supplied as a gating flow at 0.4 mL/min by a Series I HPLC pump (LabAlliance, Fisher Scientific, Pittsburgh, PA) equipped with an external pulse dampener. To inject sample onto the electrophoresis capillary, the gating flow was diverted to waste using a highspeed air-actuated six-port valve (Valco Instruments, Houston, TX). Injection voltage was applied across the separation capillary by a high-voltage power supply (CZE 1000R, Spellman High Voltage Electronics, Plainview, NY) to load sample. Once sample was injected, gating flow of electrophoresis buffer was resumed by actuating the gating valve and separation voltage applied (5 kV unless stated otherwise). Serial injections could be performed at a rate limited by separation time. All operations were controlled by a personal computer equipped with a multifunction board (ATMIO-16, National Instruments, Austin, TX) using software written in-house. (39) Lee, E.; Linder, M. E.; Gilman, A. G. Methods in Enzymology; Academic Press: New York, 1994; Vol. 237, pp 146-163. (40) Mumby, S.; Linder, M. E. In Methods in Enzymology; Academic Press: New York, 1994; Vol. 237, pp 254-268. (41) Hooker, T. F.; Jorgenson, J. W. Anal. Chem. 1997, 69, 4134-4142. (42) German, I.; Kennedy, R. T. J. Chromatogr., B 2000, 742, 353-362.

Electrophoresis buffer was 25 mM Tris, 192 mM glycine at pH 8.5 (Tris-glycine). At the beginning of each day, the separation capillary was rinsed with 0.1 M NaOH and Tris-glycine for 2 min each. Detection was accomplished using LIF as described elsewhere.43 The LIF detector used the 488-nm line of an air-cooled 15-mW Ar+ laser (Spectra Physics, Mountain View, CA) as the excitation source. Fluorescence was collected 90° from the excitation source via a 40× microscope objective (Melles Griot, Irvine, CA). The fluorescence emission was spectrally filtered using a 520 ( 10 nm band-pass filter (Corion, Holliston, MA) and spatially filtered through two pinholes (5 mm × 7 mm and 1.5 mm × 4 mm) before entering a photomultiplier tube (PMT) (model E717-21, Hamamatsu Photonics, Bridgewater, NJ). Signal from the PMT was amplified by a Keithley 428 current amplifier and collected using the same computer and software used for instrument control. APCE-LIF Assays. Samples containing BGTPγS and GRo and/ or GRi1 were made to the desired concentration in TGEM as described in the fluorescence and anisotropy studies and were assayed immediately using flow-gated CE-LIF. For Ras-like G protein assays, samples containing BGTPγS and Ras and/or Rab3A were made in a manner similar to that described in the fluorescence and anisotropy studies in Tris-glycine supplemented with 25 mM EDTA and incubated for 30 min at room temperature. Just prior to CE analysis, MgCl2 was added to 35 mM. All samples were introduced onto the separation capillary via electrokinetic injections at 0.5 kV for 0.1 s. Kinetics Experiments. Kinetics experiments were performed using GRo as analyte. Samples were prepared to the desired concentration (50 nM BGTPγS with 0-200 nM GRo) in TGEM and immediately assayed. Samples were introduced onto the separation capillary at 1 kV for 0.5 s. Serial electropherograms were taken for each sample over the course of 15 min with injections made every 30 s. The on-rate (kon) was determined from a plot of the peak height of the complex (which we will refer to as complex peak height) versus time, using GraphPad Prism (version 3.0, GraphPad Software, San Diego, CA) to fit the data to a one-phase exponential association function. The off-rate (koff) was determined by the addition of 20 µM unlabeled GTPγS to a sample containing 50 nM BGTPγS and 200 nM GRo that had already reached equilibrium. The koff was estimated from a plot of complex peak height versus time, using GraphPad Prism to fit the data to a one-phase exponential decay function. Kd Determination. To determine Kd of BGTPγS with GRo, samples consisting of 1 nM BGTPγS and 0-500 nM GRo were prepared in TGEM and assayed by CE-LIF using injection times of 0.5 s at 5 kV. A plot of the concentration of the complex [C] versus [GRo] added was generated, and data were fit to a one-site binding function using GraphPad Prism. Complex concentration was calculated as the difference between the BGTPγS added and the amount of free BGTPγS detected. This calculation was used to prevent inaccuracies in [C] due to imprecise complex height estimations at low [GRo]. Separation Temperature Studies. For studies on the effect of temperature on the separation, a P/ACE MDQ capillary electrophoresis unit (Beckman Coulter Inc., Fullerton, CA) (43) Tao, L.; Kennedy, R. T. Anal. Chem. 1996, 68, 3899-3906.

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

4299

equipped with a LIF detector was used. The LIF detector used the 488-nm line of a 3-mW Ar+ laser for excitation and emission was detected after passing through a 488-nm notch filter and a 520 ( 10 nm band-pass filter. Data acquisition and control were performed using P/ACE 32 Karat software version 5.0 (Beckman) for Windows 2000 on a 2.0-GHz IBM personal computer. Peak heights were calculated using the 32 Karat algorithm. Fused-silica capillaries used for the separation had total length (Lt) of 30 cm, Ld of 20 cm, i.d of 50 µm. and o.d. of 360 µm. Capillaries were rinsed for 5 min at 20 psi with 0.1 M NaOH, water, and Trisglycine (5 min, 20 psi) at the beginning of each day and for 2 min at 20 psi with Tris-glycine after each analysis. Samples (50 nM BGTPγS with 200 nM GRo in TGEM) were incubated at room temperature for 10 min before loading into the autosampler, which was maintained at 4 °C. Samples were injected onto the capillary hydrodynamically at 0.5 psi for 5 s. Electrophoresis buffer was Tris-glycine, and electric field was 1000 V/cm unless stated otherwise. RESULTS AND DISCUSSION APCE-LIF Assay for Gr Subunits. Development of APCE assays requires identification of solution conditions suitable for complex formation and electrophoresis conditions that allow resolution and stability of complex during separation. Initial experiments were directed at determining conditions for complex formation using fluorescence anisotropy. Previous work has shown that complexes of BGTPγS and various GR subunits form in 10 mM Hepes, 1 mM EDTA, and 10 mM MgCl2 at pH 8.0;27 however, this buffer proved undesirable for our experiments as preliminary CE experiments with Hepes buffer showed poor efficiency for the BGTPγS peak (data not shown). Complex formation was evaluated in TGEM, as Tris-glycine buffer had previously been shown to yield high-efficiency CE peaks for noncovalent complexes.44 Anisotropy was found to increase with increasing [GRo] (see Figure 2A) such that anisotropy of sample containing 50 nM BGTPγS and 500 nM GRo was almost 5-fold higher than that of 50 nM BGTPγS. Control experiments performed with trypsin inhibitor instead of GRo had no change in anisotropy, proving that the increase was not the result of viscosity changes or other nonspecific effects of protein in solution. Thus, the increase in anisotropy observed with addition of GRo was indicative of complex formation. The overall fluorescence intensity also increased with increasing [GRo] (Figure 2B), as was previously shown.27 The fluorescence enhancement has been attributed to a conformational change in BGTPγS upon binding to G proteins that relieves quenching of BODIPY FL fluorescence by guanine.27,36 The above results suggested that TGEM buffer was favorable to complex formation. Thus, we evaluated APCE for the detection of GR subunit complexes using TGEM as the sample buffer and Tris-glycine as the electrophoresis buffer. Figure 3 shows electropherograms for samples containing 50 nM BGTPγS and 0-500 nM GRo. As GRo is added to sample containing 50 nM BGTPγS, a new peak is evident at a migration time (tmig) of 14 s, which increases with increasing GRo concentration. Electropherograms of samples containing only GRo resembled the buffer blank and contained no peaks (data not shown). In addition, the peak at tmig (44) Buchanan, D. D.; Jameson, E. E.; Kennedy, R. T. Electrophoresis 2003, 24, 1375-1382.

4300 Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

Figure 2. Determination of BGTPγS-GRo binding via (A) fluorescence anisotropy and (B) fluorescence. Samples contained 50 nM BGTPγS and GRo in TGEM buffer pH 8.5. (A) Fluorescence was excited at 488 nm and collected at 510 nm. (B) Fluorescence was excited at 488 nm.

Figure 3. APCE-LIF assay of GRo. Samples contained 50 nM BGTPγS in TGEM. [GRo] was increased from 0 to 500 nM. Separation conditions were as described in the Experimental Section under Kinetics Experiments. All electropherograms shown represent sample composition 15 min after the addition of GRo. The asterisk (/) marks a BGTPγS impurity that comigrates with the complex. Peak heights are reported as relative fluorescence units (RFU).

) 14 s decreased with the addition of unlabeled GTPγS, which is expected to compete with BGTPγS for binding to the protein (more detail on the competition experiment is provided in the

Figure 4. APCE-LIF assay for GRi1 and GRo. The electropherogram shown is for a sample containing 300 nM BGTPγS with 480 nM GRi1 and 75 nM GRo in TGEM buffer, at 30 min after addition of G proteins. Separation conditions were as described in the Experimental Section under APCE-LIF Assays.

Reaction Kinetics section). From these results, it was concluded that this new peak corresponds to the BGTPγS-GRo complex. The complex peak showed splitting, suggesting the presence of at least two complexes, which may indicate detection of GRo variants. A peak with tmig ) 18 s was also observed to increase with addition of GRo. This peak, which was observed to increase with addition of any G protein we assayed, likely corresponds to a hydrolysis product of BGTPγS as discussed in the Reaction Kinetics section. Identification of such a peak will require application of other methods such as mass spectrometry or immunoaffinity probes. These results indicated that complexes could be formed and detected by CE-LIF. The quantitative aspects of the assay were then evaluated. The complex peak height increased linearly with GRo added up to ∼200 nM (R2 ) 0.97). The LDR is readily extended in the upper range by increasing the concentration of BGTPγS used. Analysis of individual samples had relative standard deviations (RSDs) of 1-5% for complex and BGTPγS peak heights (n ) 10 for each sample), except for the sample containing 10 nM GRo, where RSD for complex peak height was 17%. The limit of detection (LOD) for GRo was 2 nM, corresponding to 3 amol injected, as determined from LOD ) 3σbl/m, where σbl is the standard deviation of the blank signal and m is the slope of the calibration curve. The signal in the blank in this case was due to the presence of a contaminant in BGTPγS that comigrated with BGTPγS-GRo complex. (This peak is marked with a star in the 0 nM GRo electropherogram, Figure 3.) The use of purer BGTPγS should improve the LOD. Even with this limitation, the LOD by APCE was comparable to the LOD of 0.7 nM by anisotropy and 4 nM by fluorescence enhancement. After determining that GRo could be quantified by APCE, the assays were extended to mixtures of GRo and GRi1. Figure 4 shows a representative electropherogram for a sample containing 480 nM GRi1 and 75 nM GRo with 300 nM BGTPγS added as the affinity probe. Peaks corresponding to BGTPγS-GRi1, BGTPγS-GRo, and free BGTPγS are evident and well-resolved from each other. As with GRo, multiple peaks are observed for GRi1, indicating detection of several forms of GRi1. (Peaks due to GRi1 were identified from analyses performed with only that protein added.) APCE-LIF Assay for Ras-like G Proteins. Extension of the method to Ras-like proteins required identification of appropriate

Figure 5. APCE-LIF assay for Ras and Rab3A. Shown is an electropherogram obtained for 2 µM BGTPγS with 2 µM Ras and 2.3 µM Rab3A. Samples were prepared as described in the Experimental Section under APCE-LIF Assays. Separation conditions were the same as in Figure 4.

incubation conditions. The interaction between BGTPγS and Raslike G proteins has not been studied previously; therefore, method development was begun using conditions that have been shown to favor complex formation between Ras and MANT-GTP.30,32-35,38 These conditions involve an initial incubation with EDTA and labeled GTP followed by addition of Mg2+ (see Experimental Section). This protocol is used because Ras is obtained as a complex with GDP with which it forms a nearly irreversible complex in the presence of Mg2+.45 (High affinity for GDP is not observed for all G proteins. GRi, GRo, and GRs bind GDP with lower affinity than GTP. Mg2+thus has little effect on GDP binding for these proteins, although Mg2+ does reduce the rate of GTP dissociation from GRi and GRo by at least 10-fold.) To form a BGTPγS-Ras complex it is necessary to dissociate GDP-Ras by removing Mg2+ with EDTA. Once GDP has dissociated, MgCl2 can be added in the presence of excess BGTPγS to form BGTPγS-Ras complex. Using this approach, anisotropy and fluorescence were observed to increase with increasing Ras concentration in a manner similar to that noted for GRo (data not shown). APCE assays for GR subunits had been successful using Trisglycine as the electrophoresis buffer; therefore, this buffer was utilized for Ras-like G protein assays. Figure 5 shows an electropherogram for sample containing 2 µM Ras and 2.3 µM Rab3A with 2 µM BGTPγS added. As with the GR assays, complexes for different G proteins are resolved in