ANALYTICALBIOCHEMISTRY 177,41-45 (1989)

Protein Analysis of Mammalian Cells in Monolayer

Culture Using the Bicinchoninic Assay
R. C. Goldschmidt and H. K. Kimelberg”
Division of Neurosurgery and *Departments of Biochemistry and Pharmacology/Toxicology, Albany Medical College,
Albany, New York 12208

Received July 21,1988

choninic acid (BCA)’ protein assay (2) affords a signifi-

We have applied the bicinchoninic acid (BCA) protein cantly improved methodology for the fast and efficient
assay to rat brain primary astrocyte monolayer cul- analysis of large numbers of samples and takes less time
tures growing in multiwell culture plates. The BCA than the Lowry method. However, while this method has
method provides a more rapid and sensitive procedure been described for use with soluble protein samples and
with greater stability of color than is obtained using the standards (3-5), there has been no description of the ad-
Lowry method. Also, large numbers of samples can be aptation of this method for the measurement of protein
read rapidly at the available wavelengths on an en- in mammalian cell cultures.
zyme-linked immunosorbent assay microtiter plate We describe here an adaptation of the BCA assay that
reader. We found, however, artifactually high readings allows for the fast and efficient analysis of the protein
when using isotonic buffered sucrose to wash the cul- contents of mammalian cells growing as monolayer cul-
tures followed by sodium hydroxide to solubilize the cell tures in multiwell plates. This method possessesboth
protein. Such a procedure is commonly used for wash-
increased sensitivity and greater ease of use compared
ing monolayer cell cultures in transport and binding
to the Lowry procedure. The present method makes use
studies. This effect was found to be due to hydrolysis of
of microtiter plates and readers to measure color genera-
sucrose to the reducing sugar glucose. Use of Triton X-
100 eliminated this problem, but this agent only solubi- tion, and while the use of plate readers has been de-
lized about 80% of the protein that could be solubilized scribed for both the Lowry assay (6) and the BCA assay
with sodium hydroxide. Furthermore, the high viscos- with soluble proteins (5,7-g), the requirements of pro-
ity of Triton X-100 makes it more difficult to use. We tein solubilization in cell cultures and some aspects of
found that washing the cells with isotonic mannitol so- the BCA method itself necessitated significant modifi-
lution followed by solubilization with sodium hydrox- cations for successful use in monolayer cell cultures.
ide gave reliable results. The sensitivity and speed of
this method makes it suitable for multiple protein de- METHODS
terminations in experiments using large numbers of
cell culture samples. 0 1999 Academic Press, he. Cell cultures. Primary astrocyte monolayer cultures
were grown in 12-well plates as previously de-
scribed (10).
Protein determination. Protein standards (human
Physiological, pharmacological, and biochemical serum albumin, HSA) were prepared in either 1 N NaOH
studies of mammalian cells growing in culture usually or 1% Triton X-100. Cell cultures were washed five times
involve measuring protein as a convenient indicator of with either cold (4°C) sucrose wash (0.29 M sucrose, 0.01
the amount of cells present. Until recently, the method M Tris, 0.5 mM Ca(N03)2, pH 7.4), cold mannitol wash
of choice for protein analysis was that of Lowry et al. (similar to sucrose wash except that 0.29 M mannitol re-
(1). This method, while effective, has some drawbacks, placed sucrose), or cold phosphate-buffered saline (PBS;
namely, its limited protein concentration assay range,
the instability of the chromophore generated, and the 1 Abbreviations used: BCA, Bicinchoninic acid, HSA, human serum
inhibition of the assay by many substances used in tissue albumin; PBS, phosphate-buffered saline; ELISA, enzyme-linked im-
culture analysis. The recent introduction of the bicin- munosorbent assay; ECM, extracellular matrix.

0003-2697/89 $3.00 41
Copyright 0 1989 by Academic Press, Inc.
All rights of reproduction in any form reserved.
42 GOLDSCHMIDT AND KIMELBERG

TABLE 1 plate reader (Fig. 1, filled squares; slope, 0.0079). This is

Extraction and Assay Conditions due to the less than optimal wavelength for the chromo-
phore of the absorbance filter available in the reader
Extraction Reagent (630 nm; optimal, 750 nm) and the shorter light path
Method solution Diluent (ml) through the wells of the microtiter plates. The standard
BCA method, run on neutralized Lowry HSA standards
Lowry 1NNaOH Hz0 Lowry (1.2)
Alkaline (standard) BCA 1 NNaOH Hz0
and read in the plate reader (Fig. 1, filled circles; slope,
BCA (2)
Neutralized (standard) 0.018), resulted in an almost 50% increase in sensitivity
BCA 1 N NaOH ~NHCI BCA (2) (i.e., the slope of the standard curve) relative to the
Neutralized (mini-)BCA 1 N NaOH ~NHCI BCA (1) Lowry/Beckman procedure. Results of the standard
Triton-(mini-)BCA 1% Triton Hz0 BCA (1) BCA analysis using HSA standards prepared only in
Note. Aliquots (100 ~1) of the extraction solution were diluted with
Hz0 were superimposable on those of the neutralized
100 ~1 of the diluents and the volumes of the reagents shown added curve (Fig. 1, open triangles; slope, 0.018), indicating
prior to measurement of absorbance (see Methods for further details). that preparation of protein in NaOH and subsequent
neutralization did not affect the assay. Further optimi-
zation of the method was obtained by halving the volume
137 mM NaCl, 27 mM KCl, 1.5 mM KH2P04, 8.1 mM of BCA reagent added to standard solutions, which we
Na2HP04, pH 7.4). The culture wells were aspirated to term “mini”-BCA assay (Fig. 1, open circles; slope,
remove as much of the final wash solution as possible. 0.031). The concentration effects resulted in both sig-
One milliliter of one of the extraction solutions shown nificantly increased sensitivity within the protein range
in Table 1 was then added to each well. After a 15-min examined and increased cost effectiveness of the re-
incubation, the plates were placed in a Cole-Parmer ul- agent. This modification did not affect the accuracy of
trasonicator (Model 8845-3) and sonicated for 1 min. Al- the assay as R values were typically greater than 0.998
iquots (100 ~1) were then diluted with 100 ~1 of either for the curve. Standard BCA analysis of non-neutralized
HSA samples resulted in a significantly lower slope rela-
HZ0 or 1 N HCl (Table 1) and processed by either the
Lowry or the BCA assay (standard or mini, see below). tive to the neutralized standard assay (Fig. 1, open
Standard curves were obtained by preparing and dilut- squares; slope, 0.011).
ing samples of HSA so that conditions were identical to Use of BCA assay in astrocyte cultures. In examining
those of the different assays described in Table 1. various methods of extraction and analysis of protein, it
Lowry assay samples were prepared and incubated at was found that sucrose washing followed by Triton ex-
45°C for 30 min as described (1) and read in either a
Beckman 25 spectrophotometer (in glass l-ml cuvettes)
at 750 nm or in a Dynatech MR600 ELISA microtiter
plate reader (in 250-~1 aliquots in NUNC 96-well flat
bottom microtiter plates) at 630 nm. BCA assays were
performed by adding 2 (standard assay) or 1 ml (mini
assay) of the BCA reagent to the different final extract
samples (Table 1) or standards, incubating at 60°C for
30 min, and reading them either in the Beckman 25 spec-
trophotometer at 562 nm or in the Dynatech plate reader
at 550 nm.
Materials. BCA reagent concentrate was obtained
from Pierce Scientific. Sucrose, mannitol, and Triton X-
100 were purchased from Sigma; Folin-Ciocalteau re- 0 5 10 15 20 25

agent for Lowry assays was purchased from Fisher. HSA h)

Graphics were generated with Sigma Plot (Jandel Scien- FIG. 1. Comparison of standard curves for human serum albumin.
tific) using a World Computer AT-compatible computer. For the BCA method, standards were prepared under alkaline condi-
tions (Cl), under alkaline conditions and neutralized (O,O), or in Hz0
(A). All Lowry assay standards (A#) were in 0.5 N NaOH. The sam-
RESULTS ples were analyzed by the Lowry or BCA methods as indicated in the
Optimization of the BCA microplate assay using soluble graph and described in Table 1 and Methods. Absorbance values were
measured in either a Dynatech ELISA microplate reader or a Beck-
protein standards. The chromophore generated using
man 25 spectrophotometer (shown in parentheses in graph). All
the Lowry method gave approximately twofold greater ELISA-measured data are the means of quadruplicate aliquots of the
readings when assayed in a Beckman 25 spectrophotom- same samples. The percentage standard errors (%SE) were approxi-
eter (Fig. 1, filled triangles; slope, 0.013), compared to a matelv _ 0.5%. All R values were zreater than 0.998.
 PROTEIN ANALYSIS IN MONOLAYER CULTURES 43

TABLE 2 TABLE 3

Analysis of the Protein Content of Primary Astrocyte Effect of Washing Conditions on Protein Content of Primary
Cultures by Different Methods Astrocyte Cultures Using either Lowry or BCA Assay

Protein (rg per well) Washing solution Analysis Protein (mean pg per well)
No. Extraction Analysis Experiment A Experiment B
Sucrose Lowry 71.46 + 2.76
1” NaOH Lowry 77.60 * 2.09 78.18 f 2.48 PBS Lowry 62.99 f 2.97
1.
1” NaOH BCA 286.70 k 10.40 569.95 k 84.79 Sucrose BCA 248.28 + 7.97
2.
3. 1” Triton BCA 58.60 zk 2.70* 58.31 f 3.s9* PBS BCA 47.61 f 1.48
4. 2” NaOH Lowry N.D. 12.85 f 0.79
Note. Three-week-old primary astrocyte cultures were washed with
Note. Three-week-old cultures were washed five times with cold su- either cold sucrose wash or PBS as described under Methods. Wells
crose wash (see Methods) and primary extractions performed with ei- were extracted with 1 N NaOH. After 15 min the plates were sonicated
ther 1 ml of 1 N NaOH or 1 ml or Triton X-100 (see Table 1). Experi- for 1 min, loo-p1 aliquots were removed and either diluted (Lowry) or
ments A and B represent two different experiments using the same neutralized (BCA), and the samples were analyzed by either the Lowry
culture. Aliquots (100 ~1) of the same primary NaOH extracts were or the mini-BCA assay. Values are the mean for n = 6 wells + SE (see
assayed by either the Lowry (row 1) or the mini-BCA method (row 2). Methods for further experimental details).
In row 3, loo-p1 aliquots of primary Triton extracts were analyzed by
the BCA method. In addition, in row 4 of experiment B, the wells that
had been extracted with Triton were subsequently rewashed with cold
sucrose wash and reextracted with 1 ml of 1 N NaOH (indicated as 2” It seemed likely that the source of the problem lay in
NaOH), and 100-J aliquots analyzed by the Lowry method. For each
value, n = 12 wells (i.e., one tray) and results are given as means f SE.
the use of sucrose as a washing medium, since alkaline
N.D., Not done. conditions can lead to the hydrolysis of sucrose to the
* Significantly different (P < 0.001) from 1” NaOH/Lowry assay. reducing sugar glucose, which can interfere in the BCA
assay (2). This was supported by the finding that wash-
ing with PBS gives protein values with the BCA method
traction was satisfactory for the BCA method (Table 2, after extraction with 1 N NaOH that were comparable to
row 3) although it did give values that were 75% of the those of the Lowry method (Table 3). The cause of the
mean value found for the same cultures by the Lowry sucrose-induced artifact is presumably the hydrolysis by
method (Table 2, row 1). We further observed that after NaOH of the residual sucrose remaining in the well fol-
Triton extraction of cells, significant amounts of protein lowing final aspiration. Confirmation of the interference
remained in the culture wells (Table 2, row 4, Experi- by sucrose exposed to alkaline conditions is shown by
ment B). This was determined by washing off the re- the data in Table 4, where both 50 mM sucrose in 1.0 N
maining Triton with sucrose wash, reextracting the NaOH (final concentration) and 50 mM glucose in
wells with NaOH (2” extraction), and analyzing the 2” NaOH or Hz0 generate large amounts of color (50 mM
extract by the Lowry method. By combining the values sucrose in Hz0 generates no chromophore, data not
from the 1” Triton and 2” NaOH extracts, we found that shown). The lower absorbance due to sucrose in 1 N
18% of the total culture protein was resistant to extrac- NaOH is possibly due to incomplete hydrolysis. This,
tion by Triton. Total protein values obtained in this plus the varying amounts of residual sucrose remaining
manner (71.3 pg per well) were also not significantly
different from those obtained by the 1” NaOH/Lowry
method from the same culture (Table 2, row 1, Experi-
ment B) . TABLE 4
BCA analysis of sucrose-washed and NaOH-extracted Efficacy of Different Sugar Solutions in Chromophore
cultures proved totally unsatisfactory since the method Generation in the BCA Assay
generated artifactually high protein values (Table 2, row
2, and Table 3). Protein values by the BCA analysis after Sugar Solution Absorbance units
NaOH solubilization were four to seven times larger
than those found with the NaOH/Lowry procedure 50 mM sucrose 1 N NaOH 1.3
50 mM mannitol 1 N NaOH 0.0
(compare rows 1 and 2 in Table 2). Furthermore, the 50 mM glucose 1 N NaOH >2.1
wide discrepancies in the protein values between repli- 50 mM glucose Hz0 >2.1
cate experiments (Table 2, row 2, Experiments A and B)
indicated a severe methodological problem, since both Note. Different hexoses were prepared in the solutions shown above.
One hundred microliters of the alkaline solutions was neutralized with
sets of experiments used the same culture preparation
100 ~1 of 1 N HCl. Aqueous glucose was diluted with 100 ~1 of H20. All
and gave close agreement when assayed by the Lowry samples were analyzed with the mini-BCA method. Glucose values
method (Table 2, row l), or the BCA method on Triton- were greater than 2.1 OD units and so were beyond machine reading
extracted samples (Table 2, row 3). limits (blank = neutralized 1 N NaOH).
44 GOLDSCHMIDT
in the culture wells following aspiration would account
for the widely discrepant values between replicate deter-
minations on the same cultures in Table 2 (row 2). Man-
nitol (50 InM in 1 N NaOH) failed to generate chromo-
phore (Table 4). When mannitol is substituted for su-
crose in the washing medium (see Methods), NaOH can
be used to solubilize the cultures for subsequent analysis
using the mini-BCA method (Table 3). Using a mannitol
wash, we again found that Triton extracted 82% (79.7
+ 1.8 Kg per well) of the protein that was solubilized by
NaOH (96.8 f 2.9 pg per well) in the same cultures (data
not shown). The difference between these values was sig-
nificant at a level of P < 0.001. This ratio of Triton-ex-
tractable protein to NaOH-extractable protein is identi-
cal to that seen in Table 2. Parallel analysis from the
same extraction solution of mannitol-washed and
NaOH-extracted primary astrocyte cultures by both the
Lowry and the BCA mini method found no significant
differences in protein content values (data not shown).
DISCUSSION
For many studies using cell cultures, quantification of
data based on the protein content of cultures is widely
used. Because of the large numbers of samples that are
usually processed, any procedure that improves speed,
sensitivity, or ease of protein analysis is useful. The
method of Lowry et al. (1) has been the most widely used
assay for protein analysis. The recently developed BCA
assay (2), a modified Lowry method, provides an alterna-
tive that is superior to the Lowry method in sensitivity,
speed, and stability of the chromophore. However, there
has been no description of the use of the BCA method
for protein analysis in cell cultures. For transport and
binding studies, cell cultures are washed extensively to
remove unincorporated material and the cells are then
often solubilized in NaOH prior to analysis of protein
by the Lowry method. Detergent solubilization has also
been used in situations when neutral pH conditions are
required. However, these detergents are not compatible
with the Lowry assay (2); they are difficult to pipet due
to their viscosity and they may not be as effective as
NaOH in solubilizing matrix proteins.
The method of washing of cells prior to the extraction
procedure is an important variable and it is critical for
transport or binding studies that nonbound extracellular
material, usually radiolabelled tracers, be removed from
cultures prior to solubilization or further processing. It
has been shown that the use of physiological ionic buff-
ers for washing, i.e., PBS, can cause both significant de-
creases in cellular protein (up to 17%) and decreases in
the content of various electrolytes, with as little as 10 s of
washing (11). These authors proposed the use of isotonic
sucrose as a washing medium since it eliminated the de-
creases in protein and ionic content of cultures seen af-
AND KIMELBERG
ter washing with PBS. Sucrose washing, NaOH extrac-
tion, and Lowry analysis of culture wells is now a widely
used and successful procedure.
The data presented in Fig. 1 show that the mini-BCA
assay, utilizing microtiter plates and less reagent volume
than the standard BCA method, was almost three times
more sensitive than the standard Lowry assay as defined
by the steepness of the slope of the standard curve and
the ability to discern differences in protein content of
closely grouped samples. Also, the lower color density of
the Lowry-derived chromophore made the use of micro-
titer plates and readers less accurate. In addition, the
BCA assay is faster than the standard Lowry assay.
However, the use of sucrose wash followed by NaOH sol-
ubilization of the cells was found to be totally unsatisfac-
tory for the BCA assay (see Tables 2 and 3). This ap-
pears to be most likely due to the generation of the
reducing sugar glucose from the alkaline-induced
hydrolysis of sucrose and the subsequent reduction of
Cu2+ to Cul+ by glucose. The basis of the BCA assay is
thought to involve the initial reduction of Cu2+ to Cul+
due to the formation of a peptide bond-Cu2+ complex
and Cul+ then complexes with two BCA molecules to
generate an intense color (2). Clearly, the reduction of
Cu2+ to Cu’+ by glucose derived from the hydrolysis of
sucrose constitutes a major interference. The use of Tri-
ton X-100 for extraction eliminated this problem; we as-
sume that sucrose is not hydrolyzed under these condi-
tions. However, Triton X-100 is not as effective a solubi-
lizing agent as NaOH. It leaves significant amounts of
protein (18% in our experiments) in the wells relative to
the amount of protein extracted using NaOH (see Table
2). This Triton X-loo-resistant protein is most likely
the extracellular matrix (ECM) generated by the glial
cells in culture ( 12).
Another way to avoid the problems due to sucrose
washing is to change the wash conditions. However, the
use of PBS is to be avoided because it has been reported
to cause decreases in cell electrolyte and protein content
in cultures (11). We found that substitution of mannitol
for sucrose in the washing medium is a good alternative.
Mannitol is a nonreducing alcohol sugar, it is nonionic,
and has no inherent capability to induce chromophore
generation in the BCA assay. It thus allows for the use
of NaOH solubilization with BCA analysis. Further-
more, as with sucrose, monolayer cultures are relatively
impermeable to mannitol(13).
In conclusion, the mini-BCA method of protein analy-
sis is a significant improvement over the Lowry method,
especially when assaying a large number of cell culture
samples. It is two- to threefold faster, almost threefold
more sensitive, has a wider range, its chromophore is
stable, and the fewer pipetting steps leads to potentially
fewer errors. NaOH solubilization is preferable since it
solubilizes all cellular-derived protein including the pre-
 PROTEIN ANALYSIS IN MONOLAYER CULTURES 45

sumed matrix protein and has increased speed and accu- 3. Lonnerdal, B., Woodhouse, L. R., and Glazier, C. (1987) J. Nutr.
117,1385-1395.
racy in pipetting over Triton X-100 due to the high vis-
4. Keller, R. P., and Neville, M. C. (1986) Clin. Chem. 32,120-123.
cosity of Triton X-100 (and other detergents).
5. Sorensen, K., and Brodbeck, V. (1986) Experientia 42,161-162.
6. Fryer, H. J. L., Davis, G. E., Manthorpe, M., and Varon, S. (1986)
ACKNOWLEDGMENTS Anal. Biochem. 153,262-266.
We thank Drs. Asrar Malik and Peter Del Vecchio for helpful dis- 7. Redinbaugh, M. G., and Turley, R. B. (1986) Anal. Biochem. 153,
cussions and material support and Susan Goderie for her skilled tech- 267-271.
nical help. We also thank Mrs. E. P. Graham for typing assistance. 8. Hinson, D. L., and Webber, R. J. (1988) BioZ’echniques 6,14-19.
This work was supported by Grants NS19492 and NS23750 to H. K.
9. Lane, R. D., Federman, D., Flora, J. L., and Beck, B. L. (1986) J.
Kimelberg and Grant HL07529-07 to A. B. Malik.
Immunol. Methods 92,261-270.
10. Frangakis, M. V., and Kimelberg, H. K. (1984) Neurochem. Res.
REFERENCES 9,1685-1693.
1. Lowry, D. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. 11. Sanui, H., and Rubin, A. H. (1979) J. Cell. Physiol. 100,215-226.
(1951) J. Biol. Chem. 193,265-275. 12. Reid, L. M., and Jefferson, D. M. (1984) in Mammalian Cell Cul-
2. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gart- ture (Mather, J. P., Ed.), pp. 239-280, Plenum, New York.
ner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., 13. Vasquez, B., Ishibashi, F., and Howard, B. V. (1982) In Vitro 18,
Olson, B. J., and Klenk, D. C. (1985) Anal. Biochem. 150,76-85. 643-649.