|Home | About | Journals | Submit | Contact Us | Français|
Intrinsic protein fluorescence may interfere with the visualization of proteins after SDS-polyacrylamide electrophoresis. In an attempt to analyze tear glycoproteins in gels, we ran tear samples and stained the proteins with a glycoprotein-specific fluorescent dye. The fluorescence detected was not limited to glycoproteins. There was strong intrinsic fluorescence of proteins normally found in tears after soaking the gels in 40% methanol plus 1–10% acetic acid and, to a lesser extent, in methanol or acetic acid alone. Nanograms of proteins gave visible native fluorescence and interfere with extrinsic fluorescent dye detection. Poly-L-lysine, which does not contain intrinsically fluorescent amino acids, did not fluoresce.
An important part in protein electrophoresis studies is the visualization of protein separation on SDS-polyacrylamide gels. Often, fluorescent dyes (extrinsic fluorophores) are used to stain proteins, due to their significant advantages in terms of sensitivity and dynamic range over other staining techniques, such as classical colloidal Coomassie blue staining and silver staining.1 These fluorescent dyes are also usually compatible with mass spectrometric analysis of proteins after gel electrophoresis. Fluorescent dyes can be used for general protein staining2 or specifically for glycoprotein3 or phosphoprotein4,5 staining on SDS-polyacrylamide gels.
Almost all proteins contain tryptophan (Trp), tyrosine (Tyr), or phenylalanine (Phe) in their sequence,6 and the aromatic nucleus of these residues is a natural, or intrinsic, fluorophore, which is the source of intrinsic protein fluorescence. This has been well-known in protein chemistry for many years7 and has been used to visualize protein bands/spots in SDS-polyacrylamide gels recently.6,8 However, the intrinsic protein fluorescence may also interfere with protein detection when extrinsic fluorescent dyes were used to stain the gels. We encountered this problem when using Pro-Q Emerald 300 (an extrinsic fluorophore) to selectively stain glycoproteins3 in tears after electrophoresis in SDS-polyacrylamide gels. The intrinsic protein fluorescent bands was visible using different imaging systems or on different gels, indicating that the fluorescence was not due to a specific gel or a specific imaging instrument. This paper presents the results of our investigation of the interference and should be a warning of the pitfall to other researchers.
Bovine serum albumin (BSA), lysozyme from chicken egg white, β-lactoglobulin from bovine milk, lactoferrin from bovine colostrum, human serum albumin (HSA), and poly-L-lysine hydrobromide (MW 150–300 kDa) were purchased from Sigma Chemical Co. (St. Louis, MO). Coomassie brilliant blue R-250 was from Bio-Rad Laboratories (Hercules, CA). Pro-Q Emerald 300 glycoprotein gel stain and CandyCane glycoprotein molecular weight standard were from Molecular Probes (Invitrogen) (Eugene, OR).
Open eye basal tears were collected from one of the authors (YA) using a blunt glass capillary tube from the outer canthus of the eye.9 The tear sample was stored at −80ºC.
Individual protein solutions (BSA, lysozyme, β-lactoglobulin, lactoferrin, or HSA in a concentration range of 0.004–2.5 mg/mL) in Milli-Q water or a 1:10 diluted tear sample in Milli-Q water (18 μL) were mixed with 4X sample buffer (6 μL) containing 0.125 M Tris·HCl, 2% w/v SDS, 40% v/v glycerol, and 0.8% w/v bromophenol blue, pH 6.8. Then a 20-μL aliquot of the mixture was loaded onto a well of a pre-cast NuPAGE 4–12% w/v Bis-Tris 1.0-mm minigel (Invitrogen, Carlsbad, CA). The stacking gel contained 4% w/v acrylamide/Bis-Tris. Electrophoresis was carried out at 100 V in NuPAGE MES SDS running buffer (50 mM Tris base, 50 mM MES, 0.1% w/v SDS, 1 mM EDTA, pH 7.3) until the dye front reached the end of the gel. Bio-Rad 4–12% Bis-Tris gels (Bio-Rad) and 10% gels cast in our laboratory using standard procedures11 were also examined.
The tear protein gels were stained with Pro-Q Emerald 300 according to the manufacturer’s instructions. Briefly, after electrophoresis, the gels were fixed in 50% v/v methanol and 5% v/v acetic acid for 45 min at ambient temperature. After washing with 3% v/v glacial acetic acid twice, the carbohydrates were oxidized by incubating the gels with the periodic acid solution for 30 min at ambient temperature. The gels were then washed three times as above and stained with Pro-Q Emerald 300 solution for 100 min at ambient temperature in the dark. After washing the gels twice as above, the fluorescent bands were visualized using UV transilluminator 2000 with 302-nm lamps (Bio-Rad) and a Gel-Doc 2000 (Bio-Rad). Images of gels were taken with a 540 nm long pass filter (SciTech, Preston, Victoria, Australia) for the emission.
After electrophoresis, each sample lane was cut and then soaked at ambient temperature for 18 h in one of the following solutions: Milli-Q water, 5%, 10%, 20%, or 40% (v/v) acetic acid solution, 20%, 40%, or 80% (v/v) methanol solution, or 40% methanol with 1%, 5%, or 10% (v/v) acetic acid. Then the fluorescence of protein bands was directly visualized as above.
After taking images, the gels were soaked in Coomassie brilliant blue R-250 solution (0.1 % v/v in 50% v/v methanol and 10% v/v acetic acid) for 18 h at ambient temperature with gentle shaking. Then the gels were de-stained using a solution containing 50% v/v methanol and 10% v/v acetic acid 3–4 times during an 8 h period. Gel images were taken using a Bio-Rad GS-800 Calibrated Densitometer (Bio-Rad).
Tear samples, 1:10 diluted in Milli-Q water, were separated in SDS-polyacrylamide gels, and then the gels were stained with Pro-Q Emerald 300. Three clear bands could be seen using the UV transilluminator and the 540-nm long-pass filter (Figure 1, lane 1). The molecular weights of these bands were 78, 16, and 14 kDa, corresponding to lactoferrin, lipocalin, and lysozyme, respectively, which are the most abundant proteins in human tears.9 The result was unexpected because lysozyme is not a glycoprotein.10 The lysozyme in the CandyCane glycoprotein molecular-weight standard also gave a fluorescent band (Figure 1, lane 3, 14 kDa), together with all other proteins in the standard (1.25 μg each protein). According to the information sheet from the manufacturer (Invitrogen, Eugene, OR), the 14-kDa and 29-kDa markers in the standard are not glycoproteins and should not fluoresce with the stain used.
To demonstrate the specificity of the staining, we ran another gel of the tear sample and found that the three native fluorescent bands could be visualized after fixing the gel in the methanol (50% v/v) acetic acid (5% v/v) solution, before glycoprotein staining (Figure 1, lane 2). The result indicates that the proteins fluoresce without staining and the fluorescent bands of the tear sample in the gel stained with Pro-Q Emerald 300 are just tear proteins, not necessarily glycoproteins. The intrinsic protein fluorescence was visible using either Invitrogen NuPAGE Bis-Tris gels or Bio-Rad Bis-Tris gels, or our self-cast gels (images not shown).
The emission wavelength of intrinsic protein fluorescence in solution is in the range of 300–400 nm,7 whereas the emission wavelength of Pro-Q Emerald 300 is 480–600 nm in SDS-polyacrylamide gels. Although a 500-nm long-pass filter was recommended by the stain kit manufacturer to capture Pro-Q Emerald 300 fluorescence, we used a 540-nm long-pass filter. This filter should capture at least half of the Pro-Q Emerald 300 emission and remove all the intrinsic protein fluorescence (Product Information for Pro-Q Emerald 300 Glycoprotein Gel Stain Kit, Invitrogen). We tried to get the intrinsic protein fluorescent image without any filter, but no band could be detected due to very high background. The fluorescent images captured with the filter could be explained by the red shift of the emission of intrinsic protein fluorescence within the environment of the SDS-polyacrylamide/protein/gel matrix, as reported by others.11
BSA was used as a sample protein to investigate the native protein fluorescence in detail. The protein (38 μg in each well) was run in an SDS-polyacrylamide gel. Very weak fluorescent bands could be detected in the gels using the UV transilluminator immediately after electrophoresis (images not shown). Then each lane was cut and soaked in solutions containing different concentrations of methanol or/and acetic acid for 18 h at ambient temperature. Figure 2 shows the image of the gel pieces on the UV transilluminator. Gel pieces in Milli-Q water and 20% v/v and 40% v/v methanol solutions showed very weak bands. A higher concentration (80% v/v) of methanol made the gel shrink and turn opaque on the UV transilluminator. The gel pieces in acetic acid solution showed stronger fluorescence of the protein bands. Among the different concentrations, 20% v/v acetic acid gave the strongest fluorescence. The overall strongest fluorescent protein bands were seen in the gel pieces soaked in solutions containing both methanol and acetic acid. We tested 40% v/v methanol with different concentrations of acetic acid. No obvious difference of fluorescence intensity was seen among the 1% v/v, 5% v/v, and 10% v/v acetic acid concentrations used in the mixture.
To demonstrate whether the fluorescence detected in the gels is intrinsic protein fluorescence, the fluorescence of poly-L-lysine, which does not contain any intrinsically fluorescent amino acids (Trp, Tyr, or Phe), was measured under similar conditions. Due to its positive charge, the peptide cannot migrate into the gel.12 Instead, 10 μL (25 μg) of BSA, lysozyme, and the peptide solutions in water was dotted onto a semi-dried gel (37°C for 1 h) and the gel was dried for another hour at 37°C. Then the gel was soaked in 40% v/v methanol + 10% v/v acetic acid solution for 2 h at ambient temperature. The fluorescent spots were visualized under the above conditions. Strong fluorescent spots could be seen for BSA and lysozyme but not for poly-L-lysine (Figure 3). Coomassie blue staining of the gel afterward clearly showed an intensive poly-L-lysine spot, indicating that the lack of fluorescent peptide spot was not due to the diffusion of the polypeptide out of the gel during soaking. This result supports the hypothesis that the native fluorescence of unstained proteins comes from the intrinsic fluorophores.
To determine the visualization limits of different proteins, different amounts (64 ng, 320 ng, and 1600 ng) of lysozyme, BSA, HSA, lactoferrin, and β lactoglobulin were loaded onto gels and run as described above. After soaking the gel in 40% v/v methanol and 10% v/v acetic acid solution, the native fluorescent bands were visualized as above. Then the gels were stained with Coomassie brilliant blue R-250. The results (Figure 4) showed that the fluorescence detection was slightly less sensitive than Coomassie brilliant blue R-250 staining, but as little as 320 ng of BSA, HSA, lactoferrin, and lysozyme could be clearly visualized. For both fluorescence detection and Coomassie brilliant blue staining, the sensitivity for β-lactoglobulin was slightly lower than for other proteins.
It may be possible that an even smaller amount of protein would interfere with extrinsic fluorescent dye visualization if the conditions were more favorable. Under the conditions we used in this paper, the fluorescence of abundant proteins in a sample imposes a significant problem for investigators who use extrinsic fluorescent dyes to detect specific proteins. Researchers should be aware of this phenomenon and be careful when explaining results obtained from extrinsic fluorescent dye staining of SDS-polyacrylamide gels.
The simplicity and reasonable sensitivity of the procedure used in this study to visualize intrinsic protein fluorescent bands directly on SDS-polyacrylamide gels also make it useful to quickly establish the types of proteins in biological samples, and any changes that might occur to these proteins during pathology.
A study is underway to define the intrinsic protein fluorescence in SDS-polyacrylamide gels. When more characteristics of the fluorescence are obtained, we may be able to find a solution for the problem by changing the staining conditions or using a proper emission filter, or make it more sensitive so that it may become a common technique to visualize proteins on such gels.
We thank Brad Walsh for his valuable opinions about the work and his comments on the manuscript. This work was partly sponsored by an unrestricted grant from Allergan, Inc.