Ultrasensitive native fluorescence detection of proteins with miniaturized polyacrylamide gel electrophoresis by laser side-entry excitation

Hui ZhangEdward S. Yeung

Ames Laboratory-USDOE andDepartment of Chemistry,Iowa State University,Ames, IA, USA

Received January 13, 2006Accepted May 4, 2006

Research Article

Ultrasensitive native fluorescence detectionof proteins with miniaturized polyacrylamidegel electrophoresis by laser side-entryexcitation

Direct detection of separated proteins inside polyacrylamide gels has many advan-tages compared to staining methods. Ultrasensitive native fluorescence detection ofproteins with miniaturized 1-D and 2-D PAGE was achieved with laser side-entry exci-tation. The detection limit for R-phycoerythrin protein spots in 1-D SDS-PAGE with532 nm excitation was as low as 15 fg, which corresponds to only 40 000 molecules.The average detection limit of six standard native proteins was 5 pg per band with275 nm excitation. The dynamic range spanned more than three orders of magnitude.By using the same detection setup, approximately 150 protein spots from 30 ng of totalEscherichia coli extraction were detected on a 0.8 cm61 cm gel in 2-D separation. Thesignificant improvement in sensitivity for laser side-entry excitation comes from higherexcitation power and lower background level compared with other excitation modes.

Keywords: 1-D and 2-D gel electrophoresis / Fluorescence detection / Laser side-entry excitation / SDS-PAGE DOI 10.1002/elps.200600020

1 Introduction

Gel electrophoresis is one of the most frequently usedtechniques for the separation of complex biopolymermixtures such as DNAs and proteins [1–3]. For decadesSDS-PAGE has been the core analytical tool for theseparation of protein molecules, estimation of molecularmass, and assessment of sample purity [4]. Since 2-Delectrophoresis was introduced by O’Farrell [5], themethod has not changed significantly. First, proteins areseparated according to their pI’s in the first dimension byIEF. Disulfide bonds in the proteins are then cleaved bythiol-reducing agents, the protein subunits are dis-sociated and the polypeptide chains are saturated withSDS. The resulting complexes with similar charge-to-mass ratios are then separated by their molecular sizes inthe second dimension [6].

Sensitive detection of the separated proteins is always achallenge in SDS-PAGE. CBB dyes are the most popularprotein-staining materials due to simplicity and reliability.Sensitivity is limited by the high background value of0.1 mg per protein spot [6]. Silver staining is much moresensitive than CBB staining, with nanogram levels ofdetection ability [7]. However, it suffers from some inher-ent drawbacks such as low dynamic range, tediousstaining/destaining steps, and the use of toxic chemicals[6]. Many new dyes and fluorescence scanners are nowcommercially available to facilitate sensitive detectionbased on fluorescence staining. Some fluorescence dyessuch as SYPRO Ruby are capable of detecting as low as1 ng of protein per band with three orders of magnitude ofdynamic range [7]. In all cases, high background is anintrinsic shortcoming and a subsequent destaining step isrequired. Radioactive labeling (with 14C, 3H, etc.) is themost sensitive method, but very long exposure times areneeded for the detection of low amounts of proteins [5].Moreover, samples are limited to those which could beincorporated with radioactive nuclei [6], and environmen-tal hazards are always an issue.

In proteomics, gels are often analyzed by MS afterseparation [2]. Unfortunately, the coupling of gel separa-tion and MS suffers from the presence of dyes or fluores-

Correspondence: Professor Edward S. Yeung, Ames Laboratory-USDOE and Department of Chemistry, Iowa State University, Ames,IA 50011, USAE-mail: [email protected]: 1515-294-0266

Abbreviations: BPE, B-phycoerythrin; PAA, polyacrylamide; RPE,R-phycoerythrin

Electrophoresis 2006, 27, 3609–3618 3609

© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

3610 H. Zhang and E. S. Yeung Electrophoresis 2006, 27, 3609–3618

cence labels required to first visualize the proteins [8, 9].Direct detection methods are more desirable than stain-ing methods as intact proteins are preserved. Directdetection of protein spots from a thin polyacrylamide(PAA) gel by MS was demonstrated at 10-ng range [10].However, MS in proteomics is mainly used for analyzingpeptide fragments digested from individual protein spots,instead of characterizing all the proteins separated in agel.

Among the 20 common amino acids, tryptophan, phen-ylalanine, and tyrosine have significant UV absorptionbetween 250 and 300 nm. Proteins that contain trypto-phan or tyrosine groups fluoresce between 300 and400 nm when excited into these absorption bands [11].Database search performed by Roegener et al. [12]showed that more than 99% of the 1 million proteins withmolecular mass larger than 10 kDa have at least oneamino acid residue of tryptophan or tyrosine. This makesdirect detection methods based on both UV absorptionand native fluorescence of proteins possible [11–16]. Forexample, Yamamoto et al. [14] detected microgram-levelprotein bands by measuring the 280 nm UV absorbanceof proteins in PAA gels. Roegener et al. [12] expanded afrequency-tripled Ti:Sapphire laser (operated at 280 nm)and excited PAA gels from the top for direct fluorescencedetection of proteins. Detection limits of 5–10 ng for var-ious protein bands were obtained with 35 mW/cm2 exci-tation power. Sluszny and Yeung [16] coupled microelec-trophoresis with native fluorescence detection with topexcitation by a UV lamp. The detection limit achieved inthat project was as low as 40 pg per protein spot.

Side-entry excitation has been suggested as a moreefficient way for DNA detection in slab gels, than topexcitation [17]. In this work, the sensitivity of proteindetection by native fluorescence was improved to thepicogram level with 275 nm laser side-entry excitation.Furthermore, femtogram-level detection limit wasachieved for R-phycoerythrin (RPE), B-phycoerythrin(BPE), and Alexa Fluor-532-labeled proteins by 532 nmNd:YAG laser side-entry excitation.

2 Materials and methods

2.1 Chemicals

IEF and SDS-PAGE: dry IEF PhastGel, DTT, urea, andCHAPS were purchased from Amersham Biosciences(Piscataway, NJ). Acrylamide, bisacrylamide, bromophe-nol blue, 1.5 M Tris-HCl solution (pH 8.8), and 0.5 M Tris-HCl solution (pH 6.8) were purchased from BioRad (Her-cules, CA). SDS, Tris, glycine, TEMED, b-mercaptoetha-nol (5% v/v), ammonium persulfate, and glycerol were

obtained from Sigma (St. Louis, MO). Agarose M waspurchased from Amersham. The sample buffer (Laemmlibuffer) contained 62.5 mM Tris-HCl at a pH of 6.8, 25%glycerol, 2% SDS, and 0.01% bromophenol blue. Thegel-running buffer contained 25 mM Tris, pH 8.3, 192 mMglycine, and 0.1% SDS. All the water used in the experi-ment was obtained from a Milli-Q water purification sys-tem (Millipore, Billerica, MA).

Protein samples: Alexa Fluor-532 protein-labeling kit,RPE, and BPE were purchased from Invitrogen-MolecularProbes (Carlsbad, CA). Protein mixtures containing 2 mg/mL each of the six native proteins, lysozyme (Mr 14 400),trypsin inhibitor (Mr 21 500), carbonic anhydrase (Mr

31 000), ovalbumin (Mr 45 000), serum albumin (Mr

66 200), and phosphorylase b (Mr 97 000) were purchasedfrom BioRad. Standard proteins of BSA (Mr 66 000) andconalbumin (Mr 77 000) were purchased from Sigma.Proteins were dissolved to a final concentration of 2 mg/mL and stored at -207C until use. Escherichia coli (E.coli)protein sample used for 2-D separations was purchasedfrom BioRad.

2.2 Gel electrophoresis

2.2.1 Gel cassette

Quartz gel cassettes for both 1-DE and 2-DE (Fig. 1) weremanufactured by NSG Precision Cells (Farmingdale, NY).The cassettes were 15 mm-long rectangular tubes with15 mm61 mm inner cross-section. All surfaces were flatand polished. For 1-D SDS-PAGE experiments, six0.5 mm holes were drilled in one quartz gel cassette byCiDRA (Wallingford, CT). The six holes were 1 mm awayfrom the edge of top plate and the separation betweenholes was 2.2 mm.

2.2.2 Gel composition

For both 1-DE and 2-DE, the stacking gel was composedof 4% T and 3.3% C at a pH of 6.8 and the resolving gelwas 12% T and 3.3% C at a pH of 8.8, except for theexperiments with 532 nm laser excitation in which the re-solving gel concentration was 8% Tand 3.3% C at a pH of8.8.

Gels for 1-D SDS-PAGE were cast as follows. First, para-film was stretched to cover the bottom of the quartz gelcassette in order to prevent leaking. Monomer solution ofthe resolving gel was poured 3 mm-to-top into the sealedgel cassette. Isopropanol was added on top of the re-solving gel solution to make the gel surface flat. The gelwas polymerized for 30 min in the dark. Then six fused-silica capillaries (1.5 cm long, 360 mm outer diameter, and


Electrophoresis 2006, 27, 3609–3618 Proteomics and 2-DE 3611

Figure 1. Gel-running chamber.The gel-running chamber hasdimensions of 50 mm665 mm.Volume of the buffer containerswas 2 mL each. Inset: quartz gelcassette for 1-D SDS-PAGE.Dimensions of separation gelwere 15 mm615 mm61 mm.Six 0.5 mm diameter injectionholes were drilled on the topplate of the cassette with2.2 mm separation. Gel cas-settes for 2-D separation werethe same except that no injec-tion holes were drilled.

75 mm inner diameter; Polymicro Technologies, Phoenix,AZ) were inserted through each of the six drilled holes toabout 0.2 mm away from the bottom plate of gel cassette.An aluminium holder with six holes that were designed tomatch those in the quartz gel cassette was used to holdthe capillaries and to keep them straight. The stacking gelwas then poured into the gel cassette and polymerized for30 min. The capillaries were then pulled out and injectionwells were made with dimensions of 0.8 mm depth and0.4 mm diameter. Prior to injection, polymerized gels werephotobleached for 10 min by a 254 nm mercury lamp(Spectroline, Westbury, NY). Gels for 2-D separationswere cast the same way as those for 1-D separationexcept that no injection wells were made.

2.2.3 Sample loading and electrophoresis of 1-DSDS-PAGE

BSA and conalbumin protein were labeled with AlexaFluor-532 fluorescence dye according to the manu-facturer’s instruction. The two labeled proteins were thenmixed with RPE and BPE and diluted to the desired con-centration with H2O. Such a mixture was ready for sampleloading in experiments with 532 nm laser excitation. Pro-tein mixtures used in experiments with 275 nm laserexcitation were prepared following Laemmli’s protocol:

premixed protein standard solution from BioRad wasdiluted to the designated concentration with H2O. Then,protein mixtures were mixed with the sample buffer (finalconcentration: 50%) and b-mercaptoethanol (final con-centration: 5%) and heated to 1007C for 10 min. Sampleswere then cooled down and were ready for injection. Themicroinjector was made by connecting a short fused-silica capillary (2 cm long, 150 mm outer diameter, and50 mm inner diameter; Polymicro Technologies) with a0.5-mL syringe (Hamilton, Reno, NV), as shown in Fig. 2. Aprotein solution of 0.08 mL was injected into each injectionwell, made as described previously. Gels were then putinside the running chamber. Melted agarose solution(0.5%, in H2O) was added in-between the gel cassetteand running chamber to prevent current leakage. Elec-trophoresis was carried out horizontally at a constantvoltage of 200 V for 5 min. The whole quartz gel cassettewith the separated proteins inside was taken out from therunning chamber, rinsed with water, and dried with air.The gel was imaged inside the quartz gel cassette.

2.2.4 2-D SDS-PAGE separations

IEF was performed according to the manufacturer’sinstructions. Dry sample proteins extracted from E. coliwere dissolved in 8 M urea, 2 mM DTT, and 2% CHAPS



Figure 2. Microinjector made by connecting a 0.5 mLsyringe with a short fused-silica capillary. A 3-D transla-tional stage was employed for precise positioning.

and stored at 2707C until use. Dry IEF gels were cut into12 mm61 mm strips. The IEF strip was soaked for 1 h inrehydration solution containing 6 M urea, 2% CHAPS,0.3% w/v DTT, and 0.002% bromophenol blue. Variousamounts of E. coli protein, ranging from 30 ng to 2 mgwere added to the rehydration buffer solution. After therehydration step, high voltage was applied across the12 mm gel with the following protocol: 40 V for 10 min,80 V for 5 min, 150 V for 5 min, 200 V for 30 min, and 300 Vfor 20 min. Subsequently, IEF strips were equilibrated for15 min in a solution containing 50 mM Tris-HCl (pH 8.8),6 M urea, 30% glycerol, 2% SDS, 1% DTT, and 0.002%bromophenol blue. The second-dimensional separation(i.e., SDS-PAGE) was performed similarly to the 1-Dexperiments after the IEF strips were delivered on top ofthe stacking gels. The gels were then rinsed and imaged.

2.3 Fluorescence detection

The detection setup is presented in Fig. 3. The gel insidethe quartz cassette was placed horizontally on a holder.Either the 275 nm laser line isolated from an argon-ionlaser (Model 2045, Spectra-Physics, Mountain View, CA)or the 532 nm Nd:YAG laser (Uniphase, San Jose, CA)was used to excite the gels from the side. The laser beam

Figure 3. Experimental setup for fluorescence detectionof proteins in miniaturized gels with laser side-entry exci-tation.

was expanded horizontally by the first cylindrical lens(40.0 mm625.4 mm, f = 3 cm, CVI Laser, Albuquerque,NM) and then focused vertically by the second cylindricallens (25.4 mm in diameter, f = 15 cm, CVI Laser) into a thinsheet. The dimensions of the laser beam at the gel wereapproximately 15 mm wide and 300 mm thick. The laserpassed through the gel perpendicular to the side of the gelcassette to ensure maximum transmittance of light throughthe gel. The thin laser beam was prevented from hitting thetop or the bottom plate of the quartz gel cassette.

2.3.1 UV fluorescence detection

For experiments with UV laser, the 275 nm laser line wasisolated from other laser lines and plasma emission by aprism. All UV mirrors and UV cylindrical lenses were pur-chased from the CVI laser. The average irradiation powermeasured at the side entrance of the gel was 60–120 mW/cm2. In order to reduce the background light, the setup waskept in a closed box. The irradiated gel was imaged by a105 mm UV camera lens (f = 4.5, Nikon, Japan) onto a backilluminated, 16 bit CCD camera (TE/CCD-512-TKB, Prince-ton Instruments, Princeton, NJ), which had an approximatequantum efficiency of 40% for UV light. A long-pass colorglass filter (WG-305, Schott Glass, Duryea, PA) and a band-pass filter (330WB80, Omega Filters, Brattleboro, VT) wereused to select the fluorescence of native proteins. Theexposure of the CCD was set at 10–60 s according to thefluorescence intensity emitted from the gels.



2.3.2 Visible fluorescence detection

Experiments with 532 nm laser side-entry excitation werebased on a similar detection setup as shown in Fig. 3except that all the optics were designed for visible light.All the mirrors were Nd:YAG High Energy Laser Mirrors(CVI). Two cylindrical lenses had the same specificationsas those for the UV experiments except that no UV coat-ing was applied. The average irradiation power measuredat the side entrance of the gel was 50–200 mW/cm2. Theirradiated gel was imaged by a 28 mm Nikkor camera lens(AF28/1.4D). A 532 holographic notch filter (Kaiser OpticalSystems, Ann Arbor, MI) and a band-pass filter(580WB80, Omega Filters) were used to select the proteinfluorescence. The CCD exposure was set at 1–60 saccording to the fluorescence intensity emitted from thegels. Laser top-entry excitation was also carried out byexpanding the laser beam with a symmetric convex lens(5 mm diameter, f = 4.5 mm, Melles Griot, Rochester, NY).The enlarged laser beam was then introduced onto thegel cassette from the top at an excitation power of about15 mW/cm2. The same filters and camera lens as in theexperiments with 532 nm side-entry excitation were usedfor imaging.

2.4 Data processing

Image analysis was performed with the Winview/32(Princeton Instruments) and Pdquest software (BioRad).Quantification was based on the integrated fluorescenceintensity of each protein spot after background subtrac-tion and normalization to the excitation power.

2.5 Safety concerns

Care should be taken when handling acrylamide mono-mers as they are toxic. Appropriate protective eyewear isrequired for working with both the lasers used in this re-search.

3 Results and discussion

3.1 General concern and optimization of thesystem

3.1.1 Injection

Compared with conventional gels, miniaturized gels canprovide higher speeds of separation, higher sensitivity,and smaller quantities of reagents consumed [16]. How-ever, sample loading for miniaturized gels is quite chal-lenging. In normal SDS-PAGE experiments, injection

wells for sample introduction are in the order of severalmillimeters and several tens of microliters of sample areinjected into each well. The miniaturized gels here are only15 mm615 mm61 mm. Various approaches have beensuggested for sample delivery into ultrathin or miniatur-ized slab gels. For example, Stein et al. [18] fabricated0.1 mL injection wells with a sharp comb made by sparkerosion of a razor blade. Zheng et al. [19] applied a spe-cially designed ceramic comb to form rectangularlyshaped wells, which enabled the use of a regular micro-pipette for sample delivery. Alternatively, Liu and Swee-dler [20] and Hietpas et al. [21] used a capillary to deliversamples electrokinetically into ultrathin gels. Our groupdeveloped a method to cast small injection wells byimmersing a capillary tip inside the gel during polymeri-zation. Then, 0.1 mL of sample solution was deliveredthrough a capillary splitter assembly [16]. Here, we sim-plify the injection process by replacing the capillary split-ter with a homemade microinjector. Precise and repro-ducible liquid deliveries in the range of 20–100 nL wereachieved.

3.1.2 Optimization of laser beam

Various excitation sources could be utilized for thedetection of biopolymers separated in gels, includinglamps, lasers, light emitting diodes, etc. Among them,lasers bear many advantages due to high excitationpower, low divergence, and monochromatic character-istics. Lasers have been applied to excite both nativeproteins and labeled proteins in gels [12, 22]. In bothcases lasers were introduced from the top and did notprovide comparable sensitivity as classic staining meth-ods. Also, the gel image had to be assembled from sev-eral sequential images because only partial irradiation ofthe whole gel was achieved due to difficulties in focusingand concerns about excitation power loss. For the minia-turized gels, the laser beam could be expanded easily toexcite the whole gel without significant power decrease.Introducing the laser beam by side entry has been appliedto DNA separations [17], but no analogous proteindetection has been reported before.

Ideally, the laser beam should be able to excite all theproteins inside the entire gel area efficiently and it shouldnot hit the top or bottom plates of the gel cassette to givea high background. So the laser beam was expandedhorizontally by the first cylindrical lens to match the gelwidth, as shown in Fig. 3. Meanwhile, the distance be-tween the gel and the first cylindrical lens should not betoo far in order to prevent over-expansion and excitationpower loss. The expanded laser beam had to be focusedvertically to fit the 1 mm thick gel. Two parameters are



involved, namely a thin beam of 2 w0 (Gaussian beamradius) at the focal plane and a large depth of focus (DOF),which represents the distance in the direction of propa-gation where the laser beam thickness is ,

ffiffiffi

2p

times ofthe beam waist. Compromise was made since both couldnot be achieved simultaneously. With the optics used inour experimental setup, the laser beam was maintained toless than 0.3 mm in thickness for at least 2 cm long dis-tance.

3.1.3 Absorption by gel matrix and fluorescencebackground

One big issue for side-entry excitation is the attenuationof the laser beam as it travelled through the gel matrix.Considering the low amount of protein loading (pg to nglevels), the laser excitation power was not affected muchas a result of absorption by proteins. Monomers of bothacrylamide and PAA gels are totally transparent to the532 nm laser. Hence proteins loaded in different lanesexperience the same excitation power in experimentswith 532 nm laser excitation. For experiments at 275 nm,PAA became nontransparent for UV light. Acrylamidemonomer also has strong absorption below 300 nm dueto the C–C double bonds. During polymerization, ab-sorbance in the 260–280 nm region decreases graduallyas the polymer chain increases. However, even the 12%PAA gel polymerized from the best acrylamide monomerwe could obtain commercially, still had absorption ofabout 0.45 for 1 cm beam path at 275 nm. Normalizationwith respect to the laser power loss is thus necessary, asdescribed in detail later.

Similarly, PAA gel matrix has negligible fluorescencebackground under 532 nm excitation, but the gel matrixfluoresces considerably under a 275 nm excitation. It wasobserved that brief photobleaching under a 254 nm mer-cury lamp could decrease the fluorescence backgroundof the gel matrix to a moderate level. The residual fluo-rescence background is mostly in the 300–360 nm region,which falls in the same region as the native fluorescenceof proteins. So there are no suitable spatial filters toselectively block such a background light. Fortunately,the excitation power provided by the side-entry config-uration was so high that as low as 20 pg per protein bandcould still be detected over fairly high background levels,as presented vide infra.

3.2 Visible fluorescence detection in1-D SDS-PAGE with 532 nm laser side-entryexcitation

1-D SDS-PAGE experiments with visible fluorescencewere carried out to demonstrate the superior sensitivityprovided by laser side-entry excitation. Phycobiliproteinslike RPE and BPE found in cyanobacteria and severalgroups of eucaryotic algae [23] have strong visible fluo-rescence in the 550–650 nm region when excited by 450–600 nm visible light. Two native proteins (BSA and con-albumin) were labeled with the Alexa Fluor-532 dye togive visible fluorescence at similar wavelengths. Labeledproteins were mixed with RPE and BPE, and then dilutedto the desired concentration and injected into each lane.As shown in Fig. 4A, all four protein spots were easilyvisible in the 0.8 pg lane. The S/N ratios of 200 fg labeledBSA, labeled conalbumin, BPE, and RPE were 15,

Figure 4. Visible fluorescencedetection in 1-D SDS-PAGE witha 532 nm laser. Alexa Fluor-532-labeled BSA and conalbuminwere mixed with RPE and BPEat the same concentration. Dif-ferent amounts of each proteinwere loaded into each lane asindicated on top of the images.

Images were taken of the same gel by: (A) 532 nm laser side-entry excitation; the laser entered from the right side withexcitation power of 100 mW/cm2; and (B) 532 nm laser top-entry excitation with excitation power of 15 mW/cm2. Bothimages were taken with 10 s exposure, and both images were inverted.



3, 20, and 40, respectively. The calculated detection limitsfor each of the four proteins were 40, 200, 30, and 15 fg atS/N = 3, which correspond to only 107000, 533 000,80 000, and 40 000 of protein molecules each. It is worthmentioning that the concentration of the two labeled pro-teins was calculated by assuming 100% labeling andrecovery efficiency from the labeling reaction, which inpractice could be much lower. For comparison, laser top-entry excitation was also carried out after beam expansionby a convex lens. Protein spots below the 1 pg level werenot visible, as shown in Fig. 4B. With the same detectionsetup as employed by Sluszny and Yeung [16], 530–550nmlight was also selected from a 500 W Hg (Xe) lamp (Oriel,Stamford, CT) and utilized to excite the gels from the top.The sensitivity achieved was even lower than laser topexcitation, presumably due to the lower excitation power.

The excitation power density in laser side-entry excitationwas much higher than top-entry excitation, as the laserwas focused more tightly. A careful comparison betweenthe two gel images in Fig. 4 also indicates that after nor-malization with respect to the excitation power, the side-entry configuration provides lower background levels thanthe top-entry configuration. A major part of backgroundlight in the gel images came from reflection and scatteringfrom the gel surface. By focusing the laser beam into a thinsheet and applying side-entry excitation, such reflectionand scattering were decreased effectively. The noise level,which was dominated by the fluctuation of backgroundlight in this case, was decreased as well. This combinationled to lower detection limits for side-entry excitation.

3.3 Calibration for heterogeneous excitationpower

The PAA matrix has negligible absorption at 532 nm sothat excitation power is homogeneous across the gel.However, it has considerable absorption at 275 nm. As a

result, excitation power for different locations of the geldecreases exponentially as the laser travels through it, asis clearly shown in Fig 5A. By fitting the background levelto an exponential decay model, absorption by the PAA gelmatrix could be accounted for. This result agreed wellwith measurements made with a spectrometer. Controlexperiments were carried out by loading the trypsin in-hibitor protein into each lane as internal standards. Theintegrated fluorescence signal of each protein spot wasplotted as shaded columns in the histogram of Fig. 5B.Clearly, the fluorescence intensity of protein spots had thesame decreasing trend as the background and could befitted to the same exponential decay model. By assigningdifferent correction factors to different pixels in the raw gelimages, excitation power at different locations of the gelcould be calibrated. The process was further simplified byassigning one correction factor for each protein spot.Such a calibration for the control experiments proved tobe valid and effective, as the blank columns show inFig. 5B. All the 1-D SDS-PAGE gels with 275 nm laserside-entry excitation were calibrated by this procedure.

3.4 Native fluorescence detection of1-D SDS-PAGE with 275 nm laser side-entryexcitation

The 1-D SDS-PAGE separation of the six standard pro-teins is presented in Fig. 6. In order to prevent saturationof the CCD, gels with higher protein loading were imagedwith shorter exposure times than those with lower load-ing. Figure 6 emphasizes the superior detection cap-abilities of native fluorescence detection with laser side-entry excitation compared to those of standard stainingmethods or other fluorescence detection methods. All thesix protein spots (20 pg each) could be clearly visualizedin Fig. 6B. For example, the S/N of the 20 pg carbonicanhydrase spot was about 18, indicating a detection limit

Figure 5. (A) Background levelof a blank gel excited at 275 nmwith the laser entering from theright side. (B) Fluorescenceintensity of five trypsin inhibitorprotein spots with 275 nm laserside-entry excitation. Shadedcolumns represent the data fromthe raw image, and blank col-umns represent the normalizeddata. Protein loading for eachlane was 1.0 ng except lane 5which was 0.8 ng. Bottom: rawgel image of five trypsin inhibitorprotein spots.



Figure 6. Native fluorescencedetection in 1-D miniaturizedgels with 275 nm laser side-entry excitation. Samples weredivided into two groups: (A) highrange, 16–0.8 ng per proteinband and (B) low range, 800–20 pg per protein band. Theamounts of the protein bands

were indicated on top of both images. For both gels, the laser entered from the right side with excitation power of 90 mW/cm2. The high-range gel image was taken with 10 s exposure and the low-range gel was taken with 30 s exposure. Bothimages were inverted and the contrasts of images were adjusted separately for proper visualization.

Figure 7. Calibration plots for1-D SDS-PAGE separationswith UV laser side-entry excita-tion. Each point represents theaverage of three replicates.

of less than 4 pg (S/N = 3). All the six proteins gave aver-age detection limit of about 5 pg, which is about 4 and 2orders of magnitude better than those of the Coomassiestain and silver staining, respectively.

The detection limit for proteins with UV fluorescence isconsiderably higher than those with visible fluores-cence. The major difference is that the very low back-ground level in visible fluorescence detection is notachievable with UV laser excitation. The PAA gel matrix

has a high fluorescence background under the 275 nmexcitation. In addition, the filters are much less effectivein resolving protein fluorescence from the backgroundcompared to those used in the experiments with 532 nmlaser excitation.

Figure 7 presents the calibration plots for the 1-DSDS-PAGE separations. The high-range data points of0.8–16 ng correspond to Fig. 6A, and the low-rangedata points of 20 pg–0.8 ng correspond to Fig. 6B. The



fluorescence signal of each protein spot was calcu-lated by subtracting the adjacent background levelfrom the total signal. Calibration was applied to cor-rect for the attenuation of excitation laser power, asdiscussed previously. The 0.8 ng lane is present inboth high-range gels and low-range gels, so proteinsin this lane were used as internal standards to accountfor the different exposure times. Figure 7 shows thatthe system exhibits a large and linear dynamic range,from the pg region to an upper limit of 16 ng. Thisfeature is also better than the corresponding dynamicrange obtained by silver staining and Coomassie dyes[6, 24].

3.5 Native fluorescence detection of 2-D proteingels with 275 nm laser side-entry excitation

In previous work, the feasibility of miniature gels to sepa-rate complex protein mixtures was demonstrated [16].Miniaturized gels of 2 cm61.5 cm provided similar reso-lution as commercial mini-gel (7 cm67 cm) but with bet-ter sensitivity than conventional staining methods. About115 protein spots were detected with 0.1 mg of proteinloading when the gel was irradiated by UV lamp via top-entry excitation [16]. Figure 8 presents the 2-D separationof E. coli protein extracts with 1 mg, 0.25 mg, and 30 ngloading. Images in the 2-D experiments were not cali-brated for laser excitation power as in 1-D experiments,but there is no reason that such a calibration could not bedone in the same way for quantitative analysis. Within the10 mm68 mm gel area, the number of protein spotsdetected with the Pdquest software were 280, 280, 250,210, and 150 for 1 mg, 0.25 mg, 0.125 mg (data not shown),65 ng (data not shown), and 30 ng of protein loading,respectively. Protein loading was calculated by multi-plying the protein concentration in the rehydration solu-tion by the volume of solution taken up by the dry IEF gel,which was about 6 mL for 12 mm61 mm gel strips. Theseresults underscore the high sensitivity of native proteinfluorescence detection with laser side-entry excitation.

4 Concluding remarks

Native fluorescence detection of separated protein spotsin gel electrophoresis by side-entry excitation has manyadvantages over staining methods, such as high speed,high sensitivity, elimination of tedious staining/destainingor labeling steps, and compatibility with subsequentmass spectrometric analysis. The method was applied toboth 1-D and 2-D separations of native proteins andAlexa Fluor-532 labeled proteins. In 1-D SDS-PAGE, asimple microinjector was designed for precise delivery of

Figure 8. Native fluorescence detection of proteins in2-D gels by 275 nm laser side-entry excitation. (A) 1 mg E.coli protein extract; the image was taken with 60 mW/cm2

excitation power and 30 s exposure; (B) 0.25 mg E. coliprotein extract; the image was taken with 90 mW/cm2

excitation power and 30 s exposure; and (C) 30 ng E. coliprotein extract; the image was taken with 120 mW/cm2

excitation power and 30 s exposure. All the images wereinverted.

less than 0.1 mL of sample solution. Reproducible resultswere obtained for various protein loadings. The detectionlimit for six native proteins and two Alexa Fluor-532-labeled proteins were as low as 15 fg. E. coli proteinextracts were separated by 2-DE. On a 10 mm68 mmarea, as many as 280 and 150 protein spots were detect-ed for 1 mg and 30 ng protein loading, respectively.



When compared with other excitation modes, includinglamp top-entry excitation and laser top-entry excitation,laser side-entry excitation provided higher excitationpower and lower background levels. As a result, signifi-cant improvements in sensitivity were achieved. Althoughmanual operation of the system was used in the presentstudy, automation of the 2-D separation was presentedrecently by Xu et al. [25]. Throughput will be improved aswell as sensitivity, if such automated separation is cou-pled to this fluorescence detection method with laserside-entry excitation.

We thank Dr. Chanan Sluszny, Aoshuang Xu, and GuoxinLu for helpful discussions. E. S. Y. thanks the Robert AllenWright Endowment for Excellence for support. The AmesLaboratory is operated for the US Department of Energyby Iowa State University under Contract No. W-7405-Eng-82. This work was supported by the Director of Science,Office of Basic Energy Sciences, Division of ChemicalSciences.

5 References

[1] Laemmli, U. K., Nature 1970, 227, 680–685.[2] Rabilloud, T., Proteomics 2002, 2, 3–10.[3] Guttman, A., Ronai, Z., Electrophoresis 2000, 21, 3952–3964.[4] Ganesh, G., Kumar, T. K. S., Pandian, S. T. K., Yu, C., J. Bio-

chem. Biophys. Methods 2000, 46, 31–38.[5] O’Farrell, P. H., J. Biol. Chem. 1975, 250, 4007–4021.[6] Rabilloud, T., Proteome Research: Two-Dimensional Gel

Electrophoresis and Identification Methods, Springer-Verlag,Berlin 2000.

[7] Berggren, K., Chernokalskaya, E., Steinberg, T. H., Kemper,C. et al., Electrophoresis 2000, 21, 2509–2521.

[8] Rabilloud, T., Strub, J. M., Luche, S., Van Dorsselaer, A.,Lunardi, J., Proteomics 2001, 1, 699–704.

[9] Richert, S., Luche, S., Chevallet, M., Van Dorsselaer, A. etal., Proteomics 2004, 4, 909–916.

[10] Loo, R. R. O., Stevenson, T. I., Mitchell, C., Loo, J. A.,Andrews, P. C., Anal. Chem. 1996, 68, 1910–1917.

[11] Koutny, L. B., Yeung, E. S., Anal. Chem. 1993, 65, 183–187.

[12] Roegener, J., Lutter, P., Reinhardt, R., Bluggel, M. et al.,Anal. Chem. 2003, 75, 157–159.

[13] Hogan, B. L., Yeung, E. S., Appl. Spectrosc. 1989, 43, 349–350.

[14] Yamamoto, H., Nakatani, M., Shinya, K., Kim, B. H., Kakuno,T., Anal. Biochem. 1990, 191, 58–64.

[15] Kazmin, D., Edwards, R. A., Turner, R. J., Larson, E., Star-key, J., Anal. Biochem. 2002, 301, 91–96.

[16] Sluszny, C., Yeung, E. S., Anal. Chem. 2004, 76, 1359–1365.

[17] Chen, D. H., Peterson, M. D., Brumley, R. L., Giddings, M. C.et al., Anal. Chem. 1995, 67, 3405–3411.

[18] Stein, A., Hill, S. A., Cheng, Z. Q., Bina, M., Nucleic AcidsRes. 1998, 26, 452–455.

[19] Zheng, J. J., Odake, T., Kitamori, T., Sawada, T., Anal. Chem.1999, 71, 5003–5008.

[20] Liu, Y. M., Sweedler, J. V., Anal. Chem. 1996, 68, 3928–3933.

[21] Hietpas, P. B., Bullard, K. M., Gutman, D. A., Ewing, A. G.,Anal. Chem. 1997, 69, 2292–2298.

[22] Guttman, A., Csapo, Z., Robbins, D., Proteomics 2002, 2,469–474.

[23] Apt, K. E., Collier, J. L., Grossman, A. R., J. Mol. Biol. 1995,248, 79–96.

[24] Bradford, M. M., Anal. Biochem. 1976, 72, 248–254.

[25] Xu, A. H., Sluszny, C., Yeung, E. S., J. Chromatogr. A 2005,1087, 177–182.


Documents

Ultrasensitive native fluorescence detection of proteins with miniaturized polyacrylamide gel electrophoresis by laser side-entry excitation