detection of phosphoproteins on electroblot membranes using a small-molecule organic fluorophore

Teresa Goodman1

Birte Schulenberg1

Thomas H. Steinberg1

Wayne F. Patton2

1Molecular Probes, Inc.,Eugene, OR, USA

2Perkin Elmer LAS,Boston, MA, USA

Detection of phosphoproteins on electroblotmembranes using a small-molecule organicfluorophore

A new formulation of the small-molecule organic fluorophore, Pro-Q Diamond dye,has been developed that permits rapid and simple detection of phosphoproteins di-rectly on polyvinylidene difluoride (PVDF) or nitrocellulose membranes (electroblots).Protein samples are first separated by electrophoresis and then electroblotted to mem-branes, stained and destained, in an analogous manner as typically performed withAmido Black or Ponceau S dye staining of total protein profiles. After staining, blotsare imaged using any of a variety of laser-based gel scanners, xenon-arc lamp-basedgel scanners or charge-coupled device (CCD) camera-based imaging devicesequipped with UV trans- or epi-illumination. The uncomplicated and reliable stainingprotocol delivers results in as little as 1 h and the limit of detection for the stain is typi-cally 2–4 ng of phosphoprotein with a linear dynamic range of approximately 15-fold.Compared with traditional radiolabeling and antibody-based approaches, the newmethod offers significant advantages, including avoidance of radioactivity, no needfor expensive antibodies, no requirement for blocking unoccupied sites on the mem-brane with protein or detergent solutions, no sequence context-specific binding tophosphorylated amino acid residues and the ability to analyze the native, steady-statephosphorylation of proteins obtained directly from tissue specimens or body fluids.Pro-Q Diamond dye binds directly and exclusively to the phosphate moiety, allowingit to detect the broadest spectrum of phosphorylated proteins possible. The stainbinds noncovalently to phosphoproteins and is thus fully compatible with matrix-assisted laser desorption/ionization-time of flight-mass spectrometry (MALDI-TOF-MS) or Edman sequencing. The blot stain is also compatible with standard colori-metric, fluorogenic, and chemiluminescent detection techniques employed in immu-noblotting.

Keywords: Phosphopeptides / Phosphoproteins / Phosphoproteome / Signal transductionDOI 10.1002/elps.200406008

1 Introduction

Phosphorylation is considered a fundamental covalentpost-translational modification that regulates the func-tional status of proteins and peptides and in turn controlsmost cellular phenomena. Methods for determining thephosphorylation status of proteins and peptides are thusimportant with respect to the evaluation of diverse biolog-ical processes including carcinogenesis, signal trans-duction, apoptosis, cell division, cell motility, metabolism,gene regulation, and differentiation. Pro-Q Diamond dyeis a unique fluorescence-based detection system for the

specific and sensitive analysis of protein and peptidephosphorylation status in polyacrylamide gels, on micro-arrays and on polymeric beads, but to date has beenunsuitable for the detection of phosphorylated proteinson polyvinylidene difluoride (PVDF) or nitrocellulose mem-branes due to high nonspecific background staining [1–4].Hence, we have developed a new formulation of the fluor-ophore, referred to as Pro-Q Diamond phosphoproteinblot stain, which allows sensitive fluorescence detectionof phosphoproteins electroblotted from 1-D and 2-D gels.Proteins are electroblotted onto membranes and stainedwith Pro-Q Diamond blot stain using a short and simpleprocedure. After staining, the blots are briefly destainedand signal is detected using any of a variety of laser-based gel scanners, xenon-arc lamp-based gel scanners,or CCD camera-based systems equipped with UV epi- ortransillumination. Pro-Q Diamond blot stain can quantifyphosphoprotein loads ranging from 8 ng to 1 mg on asingle blot. The blot stain binds noncovalently to phos-

Correspondence: Teresa Goodman, Molecular Probes, Inc.,29851 Willow Creek Road, Eugene, OR 97402, USAE-mail: [email protected]: 1541-335-0504

Abbreviations: DDAO, 9H-(1,3-dichloro-9,9-dimethylacridin-2-one-7-yl); PKA, protein kinase A

Electrophoresis 2004, 25, 2533–2538 2533

2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Gen

eral

2534 T. Goodman et al. Electrophoresis 2004, 25, 2533–2538

phoproteins and is thus fully compatible with MALDI-TOF-MS or Edman sequencing. The electroblot may bepost-stained for total protein, using SYPRO Ruby proteinblot stain, allowing protein phosphorylation and expres-sion to be monitored on the same blot. Specific proteinsin the profile can be detected by using antibodies usingstandard colorimetric, fluorogenic or chemiluminescentimmunodetection procedures.

2 Materials and methods

2.1 Materials

Tween 20, chicken egg ovalbumin, bovine a-casein,dephosphorylated bovine a-casein, bovine b-casein, andporcine pepsin were from Sigma-Aldrich (St. Louis, MO,USA). Pro-Q Diamond phosphoprotein blot stain, SYPRORuby protein blot stain, Pro-Q Diamond phosphoproteindestaining solution, Pro-Q Western blot stain kit #1, Pep-permint Stick markers, and DTT were from MolecularProbes (Eugene, OR, USA). NUPAGE gels, LDS loadingbuffer, and MES running buffer were obtained from Invi-trogen (Carlsbad, CA, USA). Membranes used for electro-blotting were Pall PVDF (Pall, Ann Arbor, MI, USA), Immo-bilon-P (Millipore, Bedford, MA, USA) or nitrocellulose(Bio-Rad Laboratories, Hercules, CA, USA). Sodium ace-tate trihydrate and acetonitrile were from EM Science(Gibbstown, NJ, USA).The NDUFS4 (AQDQ subunit)monoclonal antibody was a kind gift of Dr. Roderick ACapaldi’s laboratory (University of Oregon, Eugene, OR,USA). Mowiol was obtained from Calbiochem (La Jolla,CA, USA).

2.2 Fluorescence spectroscopy

Fluorescence excitation and emission spectra wereobtained for Pro-Q Diamond phosphoprotein blot stainusing undiluted stain in a Hitachi FL 4500 spectrophot-ometer (Hitachi, Tokyo, Japan). The emission spectrumof Pro-Q Diamond dye was collected using a fixed excita-tion at 530 nm and scanning emission from 540–700 nm.The excitation spectrum of Pro-Q Diamond dye was col-lected using a fixed emission at 605 nm and scanningemission from 400–600 nm.

2.3 Electrophoresis and electroblotting

Electrophoresis was performed using NUPAGE Gels(12% Tris/glycine, 1 mm) following the manufacturer’sinstructions using a MES running buffer. Gels were elec-troblotted to PVDF or nitrocellulose membranes by stand-

ard procedures [5]. Briefly, following electrophoresis, pro-teins were electroblotted for 90 min with a semidry blot-ting system (The WEP Company, Seattle, WA, USA) at1.25 mA/cm2 using a transfer buffer consisting of 10 mM

Tris, 96 mM glycine, 10% methanol, pH 8.3. After blotting,the membranes were air-dried, prior to staining. Gelswere loaded with 2-fold dilution series of PeppermintStick markers, 800–1000 ng down to 1–2 ng/protein. Sin-gle proteins used for determining the limit of detectionwere also loaded at a 2-fold dilution series starting at a1000 ng/protein and going down to 0.25 ng/protein.

2.4 Staining of phosphoproteins afterelectroblotting

Dried PVDF membranes were rewetted by dipping into100% methanol for a few seconds. For nitrocellulosemembranes this step should be omitted, as they will bedestroyed by the treatment. All steps of staining andwashing were performed using slow agitation on an orbi-tal mixer (50 rpm). The membranes were immersed in7% acetic acid/10% methanol solution (25 mL, 10 min),washed in distilled water (50 mL, four changes), incu-bated in Pro-Q Diamond phosphoprotein blot stain (20–25 mL, 15 min), and washed with 50 mM sodium acetatepH 4, 20% acetonitrile for PVDF membranes. For nitro-cellulose, which is not compatible with acetonitrile, thePro-Q Diamond phosphoprotein gel destaining solution(Molecular Probes) was used (50 mL, 365 min). The nitro-cellulose blots were washed with distilled water for365 min to remove the destaining solution. All blotswere dried before imaging. After imaging, the blots couldbe stained for total protein using SYPRO Ruby proteinblot stain by fixing the dried blots in 45% methanol, 5%acetic acid (50 mL, 10 min), washing in distilled water(50 mL, three65 min), immersing in SYPRO Ruby blotstain (20–25 mL, 15 min), and final washing in distilledwater (50 mL, 365 min). The membranes were then air-dried and imaged as described in Section 2.5.

2.5 Detection of proteins

Following Pro-Q Diamond dye staining, images wereacquired on a Fujifilm FLA 3000 laser scanner (Fuji PhotoFilm Company, Tokyo, Japan), with 532 nm excitation and580 nm bandpass emission filter. SYPRO Ruby dye wasdetected using a 473 nm excitation and 580 nm bandpassemission filter, while 9H-(1,3-dichloro-9,9-dimethyl-acri-din-2-one-7-yl) (DDAO) was detected using 633 nm exci-tation source and a 675 nm long pass filter. Data were dis-played and analyzed using the Image Gauge Analysissoftware.


Electrophoresis 2004, 25, 2533–2538 Fluorescence detection of phosphoproteins on electroblots 2535

2.6 In vitro phosphorylation of complex I

Mitochondria were prepared from bovine heart asdescribed previously [6]. For in vitro phosphorylation, pu-rified mitochondria were diluted to a protein concentra-tion of 5 mg/mL in a buffer containing: 1 mM ATP, 100 mM

Tris-HCl, 1 mM EDTA, pH 7.5, 5 mM MgCl2, 1 mg/mL pep-statin, 1 mg/mL leupeptin, 1 mM PMSF, 20 mM NaF, 60 mM

cAMP, 20 mg/mL oligomycin, and 20 U of the catalyticsubunit of protein kinase A (PKA). After a 30 min incuba-tion at 307C, a final concentration of 1% lauryl maltosidewas added for 20 min on ice while stirring. As a non-phosphorylated control, purified mitochondria were re-suspended at a protein concentration of 5 mg/mL with100 mM Tris-HCl, 1 mM EDTA, pH 7.5, 1 mg/mL pepstatin,1 mg/mL leupeptin, 1 mM PMSF, 1% lauryl maltoside aswell as 20 mM NaF. The solution was incubated for 20 minon ice with stirring, before the extracted membranes werepelleted at 174 0006g for 20 min at 47C, TLA100.2 rotor(Beckman-Coulter, Fullerton, CA, USA). The supernatantwas applied on a sucrose step gradient as described pre-viously [1, 6]. Fraction 2 contained highly purified complexI, which was subjected to a Western blot after chloroform/methanol precipitation [7]. Ten mg of complex I was sub-jected to SDS-PAGE and Western blotting.

2.7 Immunoblotting of complex I

After completing the steps decribed for electroblottingand staining for phosphoproteins and total protein pro-files, the blots were immersed in fresh PBS (25 mL eachwash, 3615 min). Membranes were blocked for 1–2 husing a solution containing TBST (50 mM Tris, 150 mM

NaCl, 0.2% Tween 20) plus 0.25% Mowiol 4–88, and0.5% bovine serum album (BSA), pH 7.5. The primaryantibody against the 18 kDa subunit of complex I wasincubated for 90 min at a concentration of 1 mg/mL inblocking buffer at room temperature. Excess antibodywas washed off with TBST (5 min per wash), and incu-bated with 0.2 mg/mL of the secondary antibody, alkalinephosphatase-conjugated goat anti-mouse antibody, inblocking buffer. The blot was washed again with TBST(165 min), then 3 times with TBS (50 mM Tris, 150 mM

NaCl) to remove excess Tween. The blot was developedwith 1.25 mg/mL DDAO phosphate for 15 min as de-scribed before [8].

2.8 Detection sensitivity and linear dynamicrange

Four different proteins were loaded in a twofold dilutionseries on SDS-polyacrylamide minigels, separated byelectrophoresis, and electroblotted. Each protein sample

was evaluated using triplicate blots. Each protein con-tained a different number of phosphates: ovalbumin(2 phosphate residues), b-casein (5 phosphate residues),a-casein (7–8 phosphate residues) and pepsin (1 phos-phate residue). The limit of detection was determinedwith the Image Gauge Analysis software (Fuji Film) andgraphed.

3 Results

3.1 Spectral characteristics of Pro-Q Diamondphosphoprotein gel stain

The spectral properties of Pro-Q Diamond phosphopro-tein gel stain and phosphoprotein blot stain are nearlyidentical. The fluorescence excitation maximum of bothstains is 555 nm, and the fluorescence emission maxi-mum is 580 nm as described before [4]. Thus, the dye isreadily imaged using laser-based gel scanners equippedwith a 532 nm excitation source, such as the Fuji FLA3000 fluorescent imager or the Bio-Rad FX MolecularImager. Additionally, the stain is readily imaged usingxenon-arc lamp-based CCD camera imaging systemsequipped with appropriate excitation/emission filters,such as the ProXPRESS system (Perkin-Elmer LifeSciences, Cambridge, England). CCD-camera-based gelimaging systems with UV transillumination sources, suchas the Roche Lumi-Imager or Bio-Rad Fluor-S Max ima-ger may also be used to visualize Pro-Q Diamond dye(data not shown) with the same overall results. Longexposure times of 15–30 s are required with these instru-ments, however, to obtain the best dynamic range andvisualize even small amounts of phosphoprotein.

3.2 Specificity of phosphoprotein detectionon blots

To demonstrate the selectivity of Pro-Q Diamond phos-phoprotein blot stain, Peppermint Stick markers wereloaded at approximately 500 ng/band onto an SDS-poly-acrylamide gel. After electrophoresis and electroblottingthe blots were stained with Pro-Q Diamond phosphopro-tein blot stain, imaged and then stained with SYPRORuby protein blot stain and reimaged (Figs. 1A and B).The molecular weight marker standard mixture containstwo phosphoproteins (ovalbumin and b-casein) and fournonphosphoproteins (b-galactosidase, bovine serumalbumin, avidin, and lysozyme). As can be clearly seenin Fig. 1, the blot stain is highly specific for phosphopro-teins, as only ovalbumin and b-casein are stronglystained. Weak, nonspecific staining of the unphosphory-lated proteins was noted, particularly with those of



Figure 1. Sensitivity and specificity of phosphoproteindetection on PVDF membrane using Pro-Q Diamond blotstain. (A) A twofold serial dilution of molecular weightmarkers was stained with Pro-Q Diamond phosphopro-tein blot stain (Fig. 2A). (B) Same blot post-stained withSYPRO Ruby protein blot stain to visualize the total pro-tein profile. The molecular weight standard (Peppermint-Stick markers) contains two phosphoproteins (ovalbuminand b-casein) and four nonphosphoproteins (b-galactosi-dase, bovine serum albumin, avidin, and lysozyme). Theprotein amounts loaded in lane 1 are 800–1000 ng/proteinand decreases to 2 ng in lane 10.

basic isoelectric point. Even proteins with just a singlephosphate, such as pepsin, could be detected readily(see Fig. 2).

3.3 Detection sensitivity and linear dynamicrange

Four different proteins were loaded in a twofold dilutionseries on SDS-polyacrylamide minigels, separated byelectrophoresis, and electroblotted. Each protein samplewas evaluated using triplicate blots and contained a dif-ferent number of phosphates: ovalbumin (2 phosphateresidues) and b-casein (5 phosphate residues), a-casein(7 to 8 phosphate residues) and pepsin (1 phosphate resi-due) (see Fig. 2). The limit of detection was determinedusing the Image Gauge analysis software (Fuji Film) andthe results were graphed to determine the linear dynamicrange of each protein (see Fig. 3). The limit of detection onblots is typically 2–4 ng of phosphoprotein, representingthe last data point that could be detected visually on thescreen and clearly above the background fluorescencesignal. Only b-casein, containing 8 phosphates could bedetected down to 1 ng. The detection limit is influenced toa small degree by the number of phosphates/protein, withpepsin only being detected at 4 ng or higher. However, ithas to be kept in mind, that the overall transfer efficiencyof a given protein plays a greater role than the phosphatecontent. Table 1 summarizes the results for the linear dy-namic range obtained for four different proteins. As canbe seen, the results are somewhat protein-dependent,but not dependent of the number of phosphates per pro-tein.

Figure 2. SensitivityofPro-QDiamond dyedetectionof fourdifferent phosphoproteins transferred onto PVDF. (A)b-Casein; (B) a-casein; (C) pepsin; (D) ovalbumin. All proteinswere loaded in a 2-fold dilution series starting at 1000 ngdown to 0.5 ng per protein per lane as mentioned above thelanes. Proteins were detected using Pro-Q Diamond dye.

Figure 3. Fluorescence signal quantitation of two differ-ent phosphoproteins blotted onto PVDF membrane. Two-fold serial dilutions of various phosphoproteins wereblotted onto PVDF membrane, stained with Pro-Q Dia-mond blot stain, and quantified using the Fuji FLA-3000Fluorescent Imager. (A) Dilution series of ovalbumin(2 phosphate residues/protein); (B) dilution series ofb-casein (5 phosphate residues/protein). Inserts showthe respective linear dynamic ranges. All data pointsreflect an average of three experiments.


Electrophoresis 2004, 25, 2533–2538 Fluorescence detection of phosphoproteins on electroblots 2537

Table 1. Detection limits of different phosphoproteins

Protein name Number ofphosphates

Detectionlimit (ng)

Linear dynamicrange (ng)a)

Pepsin 1 4 4–500Ovalbumin 2 2 2–250b-Casein 5 1 1–62a-Casein 8 2 1–250

a) Each linear dynamic range has an r2 value of 0.99.

3.4 Compatibility of Pro-Q Diamond dyestaining with Western blotting

In addition the blot stain was tested in conjunction with aWestern blot of complex I against the mitochondrial18 kDa subunit (human NDUFS 4), which was reportedto be phosphorylated. Figure 4 clearly demonstrates thespecificity of the stain compared to the total protein stainas well as its full compatibility with antibody detectionmethods using a fluorescent substrate. Upon cAMP de-pendent in vitro phosphorylation of complex I, a newband at 18 kDa can be detected by Pro-Q Diamond dyecompared to the control sample without treatment. Thisprotein has been reported before to be a phosphorylatedsubunit in complex I [9–12] using radioactive phosphate.Whether this protein represents NDUFS 4 or another ofthe small complex I subunits has not been investigatedin this study. Overlays of the NDUFS 4 Western blot withthe Pro-Q Diamond dye stain show that the NDUFS 4

Figure 4. Multiplexed Proteomics technology used on aWestern blot of complex I before and after in vitro phos-phorylation. (A) Detection of phosphoproteins by Pro-QDiamond phosphoprotein blot stain. (B) Detection of totalprotein content using SYPRO Ruby protein blot stain.(C) Western blot detection of NDUFS4 (18 kDa) usingDDAO phosphate. (D) Lanes from C 3 and A 3 are placednext to each other to demonstrate the migration differ-ences indicated by arrows. Dia (Pro-Q Diamond dye sig-nal).

protein migrates at a slightly higher molecular weight,than the Pro-Q Diamond dye detected band (see Fig. 4).This agrees with a most recent study that identifies thePKA phosphorylated subunit as ESSS (or B17) and notas the NDUFS4 protein of complex I [13].

4 Discussion

Previously, Pro-Q Diamond dye has been used to detectphosphorylated proteins after SDS-polyacrylamide gelelectrophoresis, isoelectric focusing, and 2-D gel electro-phoresis [1, 4]. Additionally, the dye has been success-fully used to detect phosphorylated proteins and peptidesimmobilized to polymeric beads or affixed to a range ofmicroarray surfaces, including polyacrylamide-coatedslides (HydroGels, Perkin-Elmer Life Sciences), aminemicroarray substrates, aldehyde microarray substrates,epoxy microarray substrates and poly-L-lysine-coatedslides [2, 3]. The only significant limitation of Pro-Q Dia-mond dye in the context of detecting phosphorylated pro-teins displayed on solid-phase supports was that it couldnot be employed to detect phosphoproteins immobilizedon nitrocellulose or PVDF membranes, due to high non-specific staining of the membrane itself [2–4]. By reformu-lation the dye solution and optimizing the destaining pro-tocols, we have now successfully developed a reagentthat is suitable for fluorescence-based detection of phos-phorylated proteins and peptides on nitrocellulose andPVDF membranes. Like antibodies it can be used todetect phosphoproteins except, that the dye is not de-pendent on the amino acid content of a given protein,thus increasing the chances of finding novel phosphopro-teins. If used in conjunction with standard Western blottechniques, clear target identification is facilitated sincecross-referencing with gels or films is eliminated. Thestain also minimizes the need for expensive antibodiesas well as radioactivity, making it attractive for screeningproteomes for signal transduction pathways.

This research was supported in part by National CancerInstitute grant R33 CA093292, awarded to MolecularProbes, Inc. Pro-Q and SYPRO are registered trade-marks of Molecular Probes, Inc.

Received January 28, 2004

5 References

[1] Schulenberg, B., Aggeler, B., Beechem, J., Capaldi, R., Pat-ton, W., J. Biol. Chem. 2003, 278, 27251–27255.

[2] Martin, K., Steinberg, T., Goodman, T., Schulenberg, B., Kil-gore, J., Gee, K., Beechem, J., Patton, W., Comb. Chem.High-Throughput Screen. 2003, 6, 79–99.



[3] Martin, K., Steinberg, T., Cooley, L., Gee, K., Beechem, J.,Patton, W., Proteomics 2003, 3, 1244–1255.

[4] Steinberg, T., Agnew, B., Gee, K., Leung, W., Goodman, T.,Schulenberg, B., Hendrickson, J., Beechem, J., Haugland,R., Patton, W., Proteomics 2003, 3, 1128–1144.

[5] Towbin, H., Staehelin, T., Gordon, G., Proc. Natl. Acad. Sci.USA 1979, 76, 3116–3120.

[6] Hanson, B., Schulenberg, B., Patton, W., Capaldi, R., Electro-phoresis 2001, 22, 950–959.

[7] Wessel, D., Flügge, U. I., Anal. Biochem. 1984, 138, 141–143.

[8] Pretty On Top, K., Hatleberg, G., Berggren, K., Kemper, C.,Haugland, R., Patton, W., Electrophoresis 2001, 22, 896–905.

[9] Scacco, S., Vergari, R., Scarpulla, R., Technikova-Dobrova,Z., Sardanelli, A., Lambo, R., Lorusso, V., Papa, S. J., Biol.Chem. 2000, 275, 17578–17582.

[10] Sardanelli, A. M., Technikova-Dobrova, Z., Scacco, S. C.,Speranza, F., Papa, S., FEBS Lett. 1995, 377, 470–474.

[11] Sardanelli, A. M., Technikova-Dobrova, Z., Speranza, F.,Mazzocca, A., Scacco, S., Papa, S., FEBS Lett. 1996, 396,276–278.

[12] Technikova-Dobrova, Z., Sardanelli, A. M., Speranza, F.,Scacco, S., Signorile, A., Lorusso, V., Papa, S., Biochemistry2001, 40, 13941–13947.

[13] Chen, R., Fearnley, I., Peak-Chew, S. Y., Walker, J. E., J. Biol.Chem. 2004, in press.


detection of phosphoproteins on electroblot membranes using a small-molecule organic fluorophore

Documents