Analytical Biochemistry 267, 217221 (1999) Article ID abio.1998.2980, available online at http://www.idealibrary.

com on

A One-Step Method for Protein Estimation in Biological Samples: Nitration of Tyrosine in Nitric Acid
Keith C. Bible, Scott A. Boerner, and Scott H. Kaufmann
Division of Oncology Research, Mayo Clinic and Department of Pharmacology, Mayo Medical School, Rochester, Minnesota 55905

Received September 22, 1998

A number of methods are commonly employed for the determination of protein in biological samples. Unfortunately, several compounds that are constituents of biological buffers interfere with these methods, limiting their application. Previous studies have demonstrated that tyrosine rapidly undergoes nitration in nitric acid to yield 3-nitrotyrosine, which has a max of 358 nm. Utilizing this reaction, we have developed a one-step method for the assessment of protein content in biological samples. Common interfering substances, including SDS, urea, glycerol, ammonium sulfate, and -mercaptoethanol, do not interfere with this method. Because of its simplicity, this reaction might be useful for estimating protein content in a variety of biological samples. 1999 Academic Press Key Words: nitrotyrosine; 3-nitrotyrosine.

The determination of protein in biological samples is a widespread practice. These measurements are utilized to ensure that various samples are equally loaded on SDSpolyacrylamide gels and to provide a basis for comparison of enzyme activities and other analytes. Because of the importance of these measurements, a number of techniques for the estimation of protein content have been devised, including the measurement of absorbance at 260 (1) or 205 nm (2); the method of Lowry et al. (3), which relies upon the generation of a new chromophore upon reaction of an alkaline protein hydrolysate with phosphotungsticphosphomolybdic acid in the presence of Cu 2 (4); the method of Smith et al. (5), which relies upon the reduction of Cu 2 to Cu , which then forms a colored complex with BCA 1; the method of Bradford (6), which relies upon a change in absorbance upon binding of Coomassie blue to basic
Abbreviations used are: BCA, bicinchoninic acid; BSA, bovine serum albumin; Chaps, 3-[3-cholamidopropyl)dimethylammonio]-1propanesulfonate; EDTA, ethylenediammine tetraacetic acid.
0003-2697/99 $30.00 Copyright 1999 by Academic Press All rights of reproduction in any form reserved.
1

and aromatic amino acids under acidic conditions (7); and the method of Bohlen et al. (8), which relies upon the reaction of uorescamine with primary amines to generate a uorescent product. Although each of these methods is useful in a wide variety of settings, each also has limitations. Nucleic acids, neutral detergents of the polyoxyethylenephenol class, SDS, and urea interfere with spectrophotometric determinations at 260 and 205 nm (9). A number of components that are commonly used in biological buffers, including Tris, Hepes, EDTA, neutral detergents, and reducing agents, interfere with the method of Lowry et al. (10). Likewise, Tris, ammonium sulfate, EDTA, and reducing agents interfere with the BCA method, although neutral detergents and SDS do not (5). Finally, many commonly used reagents, including neutral detergents and SDS, interfere with the method of Bradford (6). A variety of strategies have been employed to avoid interfering substances. It has been noted that the precipitation of a protein with deoxycholate and trichloroacetic acid will eliminate many of these interfering substances (10), but the inclusion of this step is laborious when multiple samples are being simultaneously analyzed. Alternatively, samples can be lysed in a buffer that is compatible with a particular assay and then diluted into a second buffer after analysis. For example, cells for SDSpolyacrylamide gel electrophoresis can be lysed in Triton X-100, assayed for protein, and then diluted into SDS sample buffer. This approach, however, can result in spurious results, as illustrated by the recent demonstration that caspase-3 activation in lysates of mitogen-stimulated lymphocytes reects proteolytic cleavage occurring in the cell lysate rather than in the cells prior to lysis (11, 12). Finally, protein can be estimated by solubilizing samples directly in SDS sample buffer, spotting them onto a solid support (e.g., glass or nitrocellulose), washing away interfering substances, reacting the immobilized
217

218

BIBLE, BOERNER, AND KAUFMANN

polypeptides with Coomassie blue, and eluting the Coomassie blue to estimate the protein content (13, 14), a process that is potentially time consuming and laborious. In view of these limitations, there remains a need for alternative methods of protein determination in biological samples. Studies dating to the 1800s have demonstrated that aromatic molecules undergo nitration when treated with nitric acid. In particular, treatment of tyrosine with nitric acid produces 3-nitrotyrosine (15), which is distinguished from the parent compound by the appearance of a new absorbance peak at 358 nm. In the present report, we illustrate how this chemistry can be utilized in a one-step method to determine protein content of biological samples in several different buffers.
MATERIALS AND METHODS

Materials. All reagents were ACS grade unless otherwise indicated. Nitric acid, glycerol, HPLC-grade methanol, and acetic acid were from J. T. Baker (Phillipsburg, NJ). Tris, BSA, bovine insulin, human bronectin, human brinogen, rat albumin, chicken ovalbumin, 3-nitrotyrosine, triethylamine, Triton X-100, Tween 20, Brij 97, n-octyl-glucoside, digitonin, and various amino acids were from Sigma (St. Louis, MO). Urea was from Aldrich (Madison, WI). Electrophoresis grade SDS and -mercaptoethanol were from Bio-Rad (Richmond, CA). SDS sample buffer consisted of 2% (w/v) SDS, 4 M urea (prepared as an 8 M stock and deionized over Bio-Rad AG501-X8 mixed bed resin prior to use), 62.5 mM TrisHCl (pH 6.8 at 20 22C), and 1 mM EDTA with or without 5% (v/v) -mercaptoethanol. Protein determination method. As a consequence of renement of the observations described under Results, the following is the recommended procedure for determining protein content in samples solubilized in SDS sample buffer: 1. Remove an aliquot containing 5100 g of protein. Bring sample volume to 10 l. 2. Add 140 l of 70% nitric acid. 3. Incubate at 20 22C for 2 h. 4. Utilizing suitable microcuvettes, measure absorbance at 358 nm with H 2O as a blank. 5. Compare results to a standard curve constructed with varying amounts of BSA in 10 l SDS sample buffer subjected to the same conditions. Modications of the procedure. Simple modications allowed this procedure to be applied to other settings. When the protein content in whole cells was determined, cells were dissolved directly in nitric acid, incubated at 22C for 24 h, and sampled for absorbance at 358 nm. When perchloric acid or trichloric acetic acid precipitates were sampled, these precipitates were

solubilized directly in nitric acid and assayed in a similar fashion after 2-h incubations at 22C. Analytical methods. Tyrosine and 3-nitrotyrosine were separated by HPLC on a Beckman (Palo Alto, CA) Model 125 dual pump gradient system equipped with 507e autosampler, a 168 diode array detector, and an IBM Personal Computer 350 with Beckman Gold Nouveau software using a Brownlee MPLC. A Brownlee HPLC Newguard C18 precolumn (3.2 mm 15 mm 7 m) and a Beckman Ultrasphere ODS column (4.6 mm 250 mm 5 m) preequilibrated for at least 20 min with mobile phase A [1% (v/v) triethylamine in H 2O (pH 5.0)]. Separation was accomplished by applying the sample in mobile phase A and eluting using a linear gradient from 0 to 20% MeOH over 10 min followed by a linear gradient from 20 to 80% MeOH over the next 10 min, with detection at 254 and 358 nm. To examine the reaction of authentic tyrosine with nitric acid, 1 mg L-tyrosine was incubated in 1 ml nitric acid for 0.524 h at 22C and then diluted 1:20 with H 2O before neutralization with 1 M KOH to pH 7.0 and analysis by HPLC. Reaction of nitric acid with other amino acids. Authentic amino acids were reacted with nitric acid for 2 h at 22C. Absorption spectra were determined using a Beckman DU7400 diode array UV-vis spectrometer using 100- l sample cells and H 2O as a blank. Tissue culture. HL-60 and K562 human leukemia cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum, 100 U/ml penicillin G, 100 g/ml streptomycin, and 2 mM glutamine. Cell size was determined on a Coulter counter equipped with a channel analyzer and calibrated using polystyrene beads of dened diameter (BecktonDickinson, San Jose, CA). Prior to assay for protein, cells were sedimented at 200g, washed onced in serum-free RPMI 1640 medium containing 10 mM Hepes (pH 7.4), and solubilized in nitric acid.
RESULTS

Previous experiments from our laboratory have examined the effects of various agents on cellular accumulation of cytotoxic platinum analogues (16, 17). Because of concern about the efux of platinum analogues from cells during the course of trypsinization and centrifugation, we sought to devise a method for the assessment of protein content after cells were washed in situ and then solubilized directly in nitric acid, one of the matrices used for platinum determination. Examination of the absorption spectrum of these nitric acid lysates revealed a peak at 358 nm that was not present in cells solubilized in SDS sample buffer (Fig. 1A). The absorbance at 358 nm increased as a linear function of cell concentration (Fig. 1B). Comparison of two human leukemia cell lines, HL-60 (diameter, 11 m) and K562

PROTEIN NITRATION

219

FIG. 1. Treatment of cells or L-tyrosine with nitric acid results in an increase in absorbance at 358 nm. (A) Absorption spectrum resulting from treatment of OV202 cells with nitric acid. (B) Relationship between cell number and absorbance at 358 nm in HL-60 and K562 cells solubilized in nitric acid for 24 h. (C) Absorption spectra resulting from treatment of L-tyrosine, L-tryptophan, L-serine, and L-arginine with nitric acid for 1 h. Amino acid concentrations in nitric acid were all 0.1 mg/ml.

(diameter, 15 m), revealed that the slope was two-fold higher in the larger cell line (Fig. 1B). To assess the possibility that the appearance of this new absorbance peak at 358 nm reected a reaction of nitric acid with one or more amino acids, various Lamino acids were incubated with nitric acid as described under Materials and Methods. As indicated in Fig. 1C, L-tyrosine produced a sharp peak at 358 nm and L-tryptophan produced a shoulder at the same wavelength. In contrast, other amino acids, including L-serine, L-arginine, L-phenylalanine, L-threonine, L-alanine, L-proline, L-valine, L-histidine, L-methionine, L-isoleucine, L-glutamic acid, L-aspartic acid, and Lasparagine, failed to yield products that absorbed at 358 nm. In additional experiments, nucleotides (ATP, CTP, GTP, TTP, and UTP), RNA, and DNA also failed to yield products that absorbed at 358 nm. Based upon the activating inuence of the tyrosine hydroxyl group at the ortho position, we hypothesized that treatment of tyrosine with nitric acid would yield 3-nitrotyrosine. Further substitution (nitration) would then be impeded due to the deactivating inuence of the nitro substitutent. Consistent with this hypothesis, treatment of authentic tyrosine with nitric acid resulted in a single product that comigrated with authentic 3-nitrotyrosine by HPLC and had an absorbance maximum at 358 nm (Fig. 2). Incubation for longer periods of time resulted in disappearance of the 3-nitrotyrosine peak and appearance of additional peaks that are thought to reect nitration at additional sites as well as oxidation (data not shown). As a consequence, the absorbance at 358 nm reached a peak between 1 and 6 h and then subsequently declined. To determine whether these observations could be utilized to devise a method for quantitating protein, BSA was solubilized in nitric acid and incubated for varying lengths of time. As indicated in Fig. 3A, absorbance at 358 nm measured after incubation for 0.570 h was a linear function of protein content. Examination

of these data revealed that production of the chromophore was 88% complete within 30 min, reached a maximum at approximately 24 h, and remained stable thereafter (Fig. 3B). To determine whether the same approach could be utilized to quantitate protein under conditions compatible with SDSpolyacylamide gel electrophoresis, aliquots of BSA in SDS sample buffer containing Tris, SDS, urea, and -mercaptoethanol were treated with nitric acid as described under Materials and Methods. As indicated in Fig. 3C, the presence of increasing amounts of protein resulted in increasing absorbance at 358 nm, which was again a linear function of protein content. To conrm that the present assay was useful for measuring polypeptides other than BSA, we assessed absorbance after reaction of ve other polypeptides with nitric acid. Bovine insulin, human brinogen, human bronectin, chicken ovalbumin, and rat albumin all reacted with nitric acid to produce species with absorption maxima at 358 nm. In each case, there was a strong correlation (R 0.99) between protein content and absorbance (data not shown). In subsequent studies, the potential effect of various interfering substances was examined. Urea, SDS, -mercaptoethanol, 20% glycerol, and 50% saturated ammonium sulfate did not interfere with protein determination by the nitric acid method. Likewise, the neutral detergents Tween 20 (1%), Brij 97 (1%), n-octyl -D-glucopyranoside, and digitonin (10 M) did not interfere with this method. In contrast, the presence of 1% Triton X-100 (a phenol derivative) or trace amounts of phenol produced strong absorbance at 358 nm that interfered with the assay. Although 1% Chaps increased absorbance at 358 nm, it did not interfere with the assay as long as the protein samples used to generate standard curves were also solubilized in Chaps. Finally, to assess the reliability of this technique, one of us blindly measured protein content in six coded unknown samples using the Bradford assay (Coo-

220

BIBLE, BOERNER, AND KAUFMANN

FIG. 2. 3-Nitrotyrosine results from the treatment of L-tyrosine with nitric acid. (Top) HPLC chromatogram of L-tyrosine (see text for mobile phases and conditions). (Center) HPLC chromatogram of L-tyrosine reacted with nitric acid for 1 h. (Inset) Absorption spectrum of peak eluting at 8 min. The prominant peak eluting at 2 min corresponds to a contaminant in the nitric acid. (Bottom) HPLC chromatogram of 3-nitrotyrosine. (Inset) Absorption spectrum of peak eluting at 8 min.

massie blue binding), reaction with BCA, and nitration in nitric acid. Results of these experiments indicated that the three methods produce comparable results in assessing protein content in unknown samples (Fig. 3D). Interestingly, however, the correlation between actual and measured (predicted) protein content was slightly higher for the nitric acid method (r 0.9998) than for the BCA assay (r 0.9992) or Bradford assay (r 0.9991).
DISCUSSION

In the present study, we have demonstrated that treatment of biological samples or puried proteins with nitric acid is accompanied by the appearance of a new absorbance peak at 358 nm, that this new peak reects primarily the appearance of 3-nitrotyrosine residues, and that the appearance of this chromophore can be utilized to measure protein in biological samples

dissolved in nitric acid, water, or SDS sample buffer. This one-step assay is potentially useful for the estimation of protein in a variety of settings. The basis for this assay appears to be the formation of 3-nitrotyrosine residues in protein after treatment with nitric acid. Accordingly, proteins with differing tyrosine content would be expected to yield different slopes when utilized as standards for this assay. Although this can present a problem in experiments that require the precise quantitation of a single polypeptide species, it does not appear to be a limitation when this method is used to compare relative protein content in complex biological mixtures. Moreover, when the nitric acid assay was applied to a variety of polypeptides (including BSA, rat albumin, chicken ovalbumin, bovine insulin, human brinogen, and human bronectin), the resulting slopes did not correlate with either tyrosine or tryptophan content of the polypeptides (data not shown), suggesting that tertiary structure or other factors might also inuence the extent of nitration during nitric acid treatment. Similar differences in reactivity of various puried polypeptides have been previously observed with other protein estimation methods as well (5, 6). In contrast to tyrosine, which initially forms a single nitration product and then reacts further to produce additional products with declining absorbance at 358 nm, polypeptides yield nitration products that have a relatively stable absorbance at this wavelength (Fig. 3B). This difference in behavior between free tyrosine and polypeptides might also reect effects of the peptide backbone on reactivity of the tyrosine group in the protein-bound setting. The assay of protein content in detergent-solubilized cells by the nitric acid method requires an incubation period of only 2 h at 22C. In contrast, similar determinations on nitric acid-solubilized cells require incubation periods of at least 24 h for optimal detection of differences between different cell lines (Fig. 1B). Longer incubations may be required for nitric acidsolubilized cells because longer times may be required to complete protein nitration under these conditions. The assay based on this chemistry has a limit of sensitivity of 5 g when performed using 100- l cuvettes (Figs. 3A, 3C, and 3D). This sensitivity is slightly lower than the sensitivities of the methods of Lowry et al. (3), Smith et al. (5), and Bradford (6) performed under similar semimicro conditions. Needless to say, all of these assays are less sensitive than the uorescamine-based assay of Bohlen et al. (8). In contrast to existing assays, the nitration assay is not adversely affected by urea, reducting agents, and most nonionic and ionic detergents. As a consequence, it is possible to solubilize samples directly in SDS sample buffer, remove an aliquot for protein determination, and then adjust the sample volume based on the result of this assay (Fig. 3C). This procedure should

PROTEIN NITRATION

221

FIG. 3. Measurement of absorbance at 358 nm can be used to quantitate protein levels in samples treated with nitric acid. (A) Absorbance (at 358 nm) of varying concentrations of BSA incubated in nitric acid for 30 min, 4 h, 23 h, and 70 h. (B) Time-dependent increase of absorbance at 358 nm for a 1000 g/ml solution of BSA dissolved in nitric acid. (C) Relationship between BSA concentration and absorbance at 358 nm for samples dissolved in electrophoresis sample buffer in the presence and absence of 5% (v/v) -mercaptoethanol. (D) Relationship between actual and measured BSA concentration in six blindly submitted BSA samples assessed by the nitric acid, BCA, and Bradford protein assays. Results in C and D represent the means of results from the assay of triplicate samples. Error bars (hidden by data points) represent 1 sample standard deviation.

circumvent problems such as those discussed in the Introduction. In addition, the present method allows protein estimation in samples solubilized directly in nitric acid or precipitated in denaturing acids such as perchloric acid (Fig. 3A) or triochloracetic acid (17). Accordingly, this method might have widespread applicability to biological experiments.
ACKNOWLEDGMENTS
This work was supported in part by grants from the NIH (CA 67818 and CA 69008) and the Jack Taylor Family Foundation. The secretarial assistance of Deb Strauss is gratefully acknowledged.

REFERENCES
1. Warburg, O., and Christian, W. (1942) Biochem. Z. 310, 384421. 2. Waddell, W. J. (1956) J. Lab. Clin. Med. 48, 311314. 3. Lowry, O. H., Rosebrough, N. J., A. L., F., and Randall, R. J. (1951) J. Biol. Chem. 193, 265275. 4. Herbert, D., Phipps, P. J., and Strange, R. E. (1971) in Methods in Microbiology, Vol. 5B, pp. 242265. 5. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and Klenk, D. C. (1985) Anal. Biochem. 150, 76 85.

6. Bradford, M. M. (1976) Anal. Biochem. 72, 248 254. 7. Compton, S. J., and Jones, C. G. (1985) Anal. Biochem. 151, 369 374. 8. Bohlen, P., Stein, S., Dairman, W., and Udenfriend, S. (1973) Arch. Biochem. Biophys. 155, 213220. 9. Simonian, M. H., and Smith, J. A. (1996) in Current Protocols in Molecular Biology, Chapter 10, Suppl. 35, pp. 10.1.110.1.10, Wiley, New York. 10. Peterson, G. L. (1979) Anal. Biochem. 100, 201220. 11. Miossec, C., Dutilleul, V., Fassy, F., and Diu-Hercend, A. (1997) J. Biol. Chem. 272, 13459 13462. 12. Zapata, J. M., Takahashi, R., Salvesen, G. S., and Reed, J. C. (1998) J. Biol. Chem. 273, 6916 6920. 13. McKnight, G. S. (1977) Anal. Biochem. 78, 86 92. 14. Winterbourne, D. J. (1993) in Biomembrane Protocols, Vol. I, Isolation and Analysis, pp. 197202, Humana, Totowa, NJ. 15. Waser, E., and Lewandowski, M. (1921) Helv. Chim. Acta 4, 657 666. 16. Budihardjo, I. I., Walker, D., Svingen, P. A., Buckwalter, C. A., Desnoyers, S., Shah, G. M., Poirier, G., Ames, M. M., and Kaufmann, S. H. (1998) Clin. Cancer Res. 4, 117130. 17. Bible, K. C., Boerner, S. A., Kirkland, K., Anderl, K. L., Bartelt, Jr., D., Svingen, P. A., Kottke, T. J., Eckdahl, S., Stalboerger, P. G., Jenkins, R. B., and Kaufmann, S. H. (1999) Submitted for publication.