50 G E N E R A L M E T H O D S FOR H A N D L I N G PROTEINS A N D E N Z Y M E S [6]

[6] Q u a n t i t a t i o n o f P r o t e i n
By CaRISTA M. STOSCHECK

Introduction

Many methods for estimating protein concentration are available and

the appropriate choice of method depends on five major criteria: the
amount of protein available to assay, the concentration of the protein, the
specificity of the assay, the presence of chemicals which may interfere
with the assay, and the ease and reliability of performing the assay. An
approximate range of sensitivity is given for each assay. It should be
emphasized that this range is approximate since the sensitivity of each
assay is highly dependent on the type of protein being measured and the
assay volume. On the average, microtiter plate assays are 10 times more
sensitive than the indicated range. Only those assays that are easy to
perform, require simple instrumentation, and are highly sensitive will be
discussed although there are many other excellent methods. ~-9 Methods
to concentrate samples or to eliminate interfering reagents are available
and will also be discussed. Protein assay kits are available from Bio-Rad
(Richmond, CA), Pierce (Rockford, IL), and Sigma (St. Louis, MO).

Ultraviolet Absorption Methods

Ultraviolet light absorption methods have several advantages: (1) they

can be performed directly on the sample without the addition of any
reagents, (2) they can be performed very rapidly since no incubations are
required, and (3) the relationship between protein concentration and ab-
sorbance is linear. Despite its technical simplicity there can be many

E. Layne, this series, Vol. 3, p. 447.

2 G. L. Peterson, this series, Vol. 91, p. 95.
3 D. H. Campbell, J. S. Garvey, N. E. Cremer, and D. H. Sussdorf, in "Methods in
Immunology," 2nd Ed., p. 61. Benjamin, Reading, Massachusetts, 1970.
4 G. Kresze, in "Methods of Enzymatic Analysis" (H. U. Bergmeyer, ed.), 3rd Ed., Vol. 2,
p. 84. Verlag Chemic, Deerfield Beach, Florida, 1983.
5 S. J. Jackson and E. L. McCandless, Anal. Biochem. 90, 802 (1978).
6 S. B. Sheffield, D. Graft, and H. P. Li, Anal. Biochem. 166, 49 (1987).
7 G. Krystal, Anal. Biochem. 167, 86 (1987).
g V. Neuhoff, K. Philipp, H. Zimmer, and S. Mesecke, Hoppe-Seyler's Z. Physiol. Chem.
360, 1657 (1979).
9 E. C. Butcher and O. H. Lowry, Anal. Biochem. 76, 502 (1976).

Copyright © 1990 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 182 All rights of reproduction in any form reserved.
[6] QUANTITATION OF PROTEIN 51

1.2

0.6
0
e-
..Q
8if) 0.3

<
0

o4i B

200 240 280 320 360

Wavelength (nm)
Fro. ]. Ultraviolet spectrum of proteins and nucleic acids. Bovine immunoglobulin G (1),
bovine serum albumin (B), and gelatin (G) were weighed and solubilized at 1 mg/ml in the
following buffer: 0.01% Brij 35, 0.1 M K2SO4, and 5 m M KH2PO4, pH 7. The scans of a 1
mg/ml solution are shown in the inset of (A) and the scans of a 15/~g/ml solution are shown
in (A). RNA and DNA w e r e weighed and solubilized at 10/~g/ml and the wavelength scans
are shown in (B).

pitfalls if the principles of this technique are not understood. Several

excellent reviews are available on this topic) °,11
Proteins actively absorb light in the ultraviolet region with two max-
ima, 280 and 200 nm (see Fig. 1). Absorption spectroscopy involves the
absorption of a photon by an electron. Only those photons with a certain
energy level can be absorbed as defined by the difference in energy be-
tween the orbital of the unexcited electron and a higher energy orbital.
This is why there are absorption maxima. Photons with higher energy

to j. R. Little and H. Donahue, Methods Immunol. lmmunochem. 2, 163 (1968).

n j. W. Donovan, in "Physical Principals and Techniques of Protein Chemistry," Part A
(S. J. Leach, ed.), Academic Press, New York, 1969.
 52 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

have shorter wavelengths. Thus, electrons that are excited at 280 nm have
absorbed less energy than those at 200 nm. Less energy is required for the
electrons which absorb at 280 nm because these electrons lie within aro-
matic rings which stabilize the excited state due to resonance. Amino
acids which have aromatic rings are phenylalanine, tryptophan, histidine,
and tyrosine. It should be noted that those proteins with few of these
amino acid residues would be expected to have little absorbance at 280
nm. This is shown to be the case for gelatin (Fig. 1). In addition to
secondary structure, the tertiary structure of a protein can also play a role
in its absorbance spectrum because interactions between different amino
acids can further stabilize electron excited states. Consequently, condi-
tions such as buffer pH, polarity, and ionic strength that alter tertiary
structure can alter the absorbance spectrum of a protein. Buffers and
buffer components can also interact directly with certain amino acids
resulting in the stabilization or destabilization of electron orbitals. Al-
though there is a high level of variability of absorbance at 280 nm from
protein to protein, this wavelength has been found to be convenient for
protein estimation because fewer chemicals absorb at this wavelength
than at shorter wavelengths.
The peptide bond absorbs photons below 210 nm. Because of the large
number of peptide bonds in a protein, this is a highly sensitive area of the
protein spectrum. Although protein conformation and some absorption by
tryptophan and tyrosine residues occurs in this region, less variability
between proteins is observed than at 280 nm. The disadvantage of this
region is that many chemicals also absorb, especially those which contain
double bonds between carbons or carbon and oxygen. However, buffers
can be chosen carefully so that this highly sensitive region can be used
(Table I).

TABLE I
CONCENTRATION LIMITS OF CHEMICALS IN PROTEIN ASSAYSa

Concentration limits

UVg
Enhanced Colloidal
Substance b copper~ BCA d Dye ~ gold f 280 nm 205 nm

Acids and bases

HC1 0.1 M 0.1 M 0.1 M >1 M 0.5 M
NaOH 0.1 M 0.1 M 0.1 M >1 M 25 m M
PCA >1.25% <1% 10% 1M
TCA >1.25% <1% 10% <1%
 TABLE I (continued)

Concentration limits

UVe
Enhanced Colloidal
Substance h copper" BCA d Dye e gold f 280 nm 205 nm

Buffers
Acetate 0.2 M 0.6 M 0.1 M 10 mM
Ammonium sulfate >28 mM 20% 1M <1 M >50% 9%
Borate 10 mM >100 mM
Citrate 2.5 mM <1 mM 50 mM 5% <10 mM
Glycine 2.5 mM 1M 0.1 M 0.1 M 1M 5 mM
HEPES 2.5/xM 100/~M 100 mM 20 mM <20 mM
Phosphate 250 mM 250 mM 2M 100 mM 1M 50 mM
Tris 250 mM 0.1 M 2M 0.5 M 40 mM
Detergents
Brij 35 1% 1% 1%
CHAPS 1% 10% <0.1%
Deoxycholate 625/xg/ml 0.25% 0.30% 0.1%
Digitonin 10%
Lubrol PX 1% 10%
Octylglucoside 1% 10%
SDS 1.25% 1% 0.10% 0.10% 0.10% 0.10%
Triton X-100 0.25% 1% 0.10% 1% 0.02% <0.01%
Triton X-100(R) >10% 2%
Tween 20 0.10% 1% 1% 0.30% 0.10%
Reductants
Dithiothreitol 50 p.M < 1 mM 1M 1 p~M 3 mM 0.1 mM
2-Mercaptoethanol 1.8/zM <1% 1M 10/zM 10 mM <10 mM
Miscellaneous
DNA/RNA 0.2 mg 0.1 mg 0.25 mg 10 ng 1 ~g
DMSO >6.2% 5% 20% 20% < 10%
EDTA 125/zM 10 mM 0.1 M 10 mM 30 mM 0.2 mM
Glycerol 25% 10% 100% 10% 40% 5%
KC1 30 mM < t 0 mM 1M 100 mM 50 mM
NaC1 1.75 M 1M 5M <1 M >1 M 0.6 M
Sucrose 50 m M 40% 1M 1M 2M 0.5 M
Urea >200 m M 3M 6M >1 M <0.1 M

a This table is a general guide. Test buffer mixtures as described in the text. More complete listings
of chemicals may be obtained from Refs. 2, 24, 25, and 30. Figure preceded by (<) or (>) symbols
indicate that the tolerable limit for the chemical is unknown but is, respectively, less than or
greater than the amount shown. Blank spaces indicate that data were unavailable.
b PCA, Perchloric acid; TCA, trichloroacetic acid; HEPES, N-2-hydroxyethylpiperazine-N'-2-
ethanesulfonic acid; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]propanesulfonic acid;
SDS, sodium dodecyl sulfate; (R), reduced; DMSO, dimethyl sulfoxide; EDTA, ethylenediamine-
tetraacetic acid.
c Figures indicate the final concentration of the chemical in the assay.
a Figures indicate the concentration of the chemical in a 50-/~1 sample.
e Figures indicate the concentration of the chemical in a 25-/~1 sample.

YFigures indicate the concentration of the chemical in a 10-p.1 sample.

g Figures indicate the final concentration of the chemical which does not produce an absorbance of
0.5 over water.
54 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

Avoid storing buffers in plastic containers since some plastics leach

plasticizers which absorb at UV wavelengths. Detergents can also be
troublesome since many absorb UV light. In some cases hydrogenated
forms of the detergent are available where the carbon double bonds have
been converted to single bonds, thus reducing their absorption in the UV
range. For example, Calbiochem (La Jolla, CA) and Aldrich (Milwaukee,
WI) now offer hydrogenated Triton X-100.

Method 1: Absorbance at 280 nm

Range: 20 to 3000 tzg

Procedure
1. Turn on the UV lamp of the spectrophotometer and warm up the
machine (usually 15 min). Adjust the wavelength to 280 nm.
2. Zero the spectrophotometer using the buffer in which the protein is
dissolved as a blank.
3. Measure the absorbance of the protein solution.
4. For unknown or protein mixtures, use the following formula for a
rough estimate (when using a cuvette with a path length other than 1 cm,
divide the absorbance reading at 280 nm by the path length in centi-
meters):
Concentration (mg/ml) = absorbance of protein at 280 nm (1)
For a protein whose absorbance coefficient is known, its value may be
found in the reference cited below. 12 The absorbance coefficient is usu-
ally given a s A I cm
me'/ml, ,11%
i l l era, o r E M (molar extinction coefficient). The
absorbance coefficient is dependent on pH and ionic strength, so be sure
to match these parameters with those in the tables. Protein concentration
is then determined by the following formula when using a cuvette with a
1-cm path length:
Concentration (mg/ml) = absorbance/Al cmmg/ml
Concentration (%) = absorbance/A l~m (2)
Concentration (M) = absorbance/Er~
Percentage protein and molarity can easily be converted to milligrams per
milliliter protein by the following formula:
Concentration (mg/ml) = percentage protein/10
= molarity/protein molecular weight (3)

12 D. M. Kirschenbaum, in "Handbook of Biochemistry and Molecular Biology" (G. D.

Fasman, ed.), 3rd Ed., Vol. 2, p. 383. CRC Press, Cleveland, Ohio, 1976.
[6] QUANTITATION OF PROTEIN 55

C o m m e n t s . If the buffer or protein solution is cold, the outside of the

cuvette may need to be wiped between each reading with a lint-free wiper
and the readings made quickly after placing the cold solution into the
cuvette, because atmospheric moisture may condense on the outside of
the cuvette producing an erroneously high reading. Warming of cold solu-
tions also can release gas in the form of bubbles which will also produce
erroneously high readings. Alternatively, use warmed solutions.
ff the absorbance of the protein solution is greater than 2, dilute the
protein sample in the same buffer as the original solution.
Particulates or lipid droplets cannot be present in the protein solution
since they have a strong tendency to scatter light at short wavelengths.
The range of sensitivity given above is extremely rough since it is
dependent on the type of protein being measured. For a more accurate
estimate of protein concentration use Method 2 or 3 if the buffer condi-
tions permit.
The absorbance coefficient iw~cm)
,tl% for some commonly used proteins
are as follow: bovine serum albumin, 6.3; bovine, human, or rabbit im-
munoglobulin G, 13.8; and chicken ovalbumin, 7 ) 2
Nucleic acids absorb strongly at 280 nm. Thus, crude cell extracts
containing RNA and/or DNA will produce erroneously high protein esti-
mates. To correct for nucleic acid content perform the following steps in
addition to those outlined above: Zero the cuvette containing the buffer at
260 nm. Next, place the solution containing the protein in the cuvette and
read the absorbance. To determine the protein concentration of the solu-
tion, use the following formula (A is absorbance which is directly fol-
lowed by the wavelength, in nanometers, at which it was obtained)13:
Protein concentration (mg/ml) = 1.55A280 - 0.76A260 (4)

M e t h o d 2: A b s o r b a n c e at 205 nm 14

R a n g e : 1 to 100 tzg
Procedure. The procedure is the same as for Method 1 except for
adjusting the spectrophotometer to 205 nm and including 0.01% Brij 35 in
the buffer to prevent protein losses onto vessel surfaces. Proteins stick to
both plastic and glass surfaces in both concentrated and dilute solutions,
but losses are proportionately higher in dilute solutions. Proteins absorb
much more strongly at 205 nm and there is less variability from protein to
protein 7 (see Fig. 1). Use Eq. (5) to estimate protein.

i3 W. E. Groves, F. C. Davis, Jr., and B. H. Sells, Anal. Biochem. 22, 195 (1968).
J4 R. K. Scopes, Anal. Biochem. 59, 277 (1974).
56 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

Concentration (mg/ml) = 31(absorbance at 205 nm) (5)

Comments. See Method 1 for general comments.
At normal cellular concentrations of proteins and nucleic acids, nu-
cleic acids contribute relatively little toward the absorbance at 205 nm.
This method is highly dependent on the accuracy of the wavelength
setting of the spectrophotometer because 205 nm is on the shoulder of the
protein peak (see Fig. 1). It is also dependent on the absence of stray light
(this parameter is part of the mechanics of the spectrophotometer itself).
These shortcomings can be overcome by using a 10 /zg/ml solution of
BSA as a standard and determining protein concentration from the
straight line between the buffer blank and the absorbance of the BSA.
Only one standard point is needed since the relationship between protein
concentration and absorbance is linear.
Protein concentration can also be determined at 210 nm. However,
this wavelength is less sensitive and tends to vary more with buffer condi-
tions than at 205 nm. The extinction coefficients for 1 mg/ml concentra-
tions of proteins range between 20 and 24 at 210 nm. 14

Method 3: Determination of Extinction Coefficient for Protein of

Unknown Concentration 14
Range: 20 to 3000/xg
Procedure
I. Determine the absorbance of the protein at 205 and 280 nm as
outlined above. The protein will have to be diluted about 30-fold for the
205-nm reading.
2. Use the absorbances in the following formula. Do not forget to
multiply the reading at 205 nm by the dilution factor before using the
formula.
E205nm = 27 + 120(A28o/A2os) (6)
3. The concentration of the protein is then determined using Eq. (2)
and the value derived in Eq. (7).
4. Now that the concentration of the protein is known, the extinction
coefficient at 280 nm can be determined.
E2s0nm = concentration/Az8o (7)
Comments. These determinations are dependent on the protein having
an "average" amino acid composition. Abnormal phenylalanine content
will especially skew the result.
Scopes claims an accuracy of 2% for most proteins. ~4
16] QUANTITATION OF PROTEIN 57

Colorimetric Assays
When measuring protein concentrations it is best to aim in the middle
of the range of sensitivity of the assay because the assays sometimes
become nonlinear at both very high and very low levels. It is a good idea
to also assay your protein solution following a 1 : 1 dilution to ascertain
that your sample does indeed lie on the linear portion of the curve. Also,
the presence of unknown factors that interfere with the assay may be
revealed by an unexpected result. Sometimes the assay can be linearized
by plotting the log of the absorbance against the log of the amount of
protein. 15
All colorimetric assays have the disadvantage that different proteins
produce different absorbances and thus, unless the same protein as that
being assayed is used as the standard, the protein values obtained are
relative rather than actual values. For this reason it is imperative that
when such protein values are reported in the literature, the assay used and
the protein used as the standard are identified.
Protein standards should be assayed at the same time and in the same
solution as the unknowns to take into account interactions of the buffer
with the assay reagents, reagent decomposition, time differences, and
temperature changes.
In many applications, the solution containing the protein contains a
mixture of components. To test whether these components will interfere
with the protein assay, perform the assay in the usual manner adding the
following reagents as samples: water for the blank, the buffer, a protein
standard dissolved in water, and the same amount of protein dissolved in
the buffer. If the reagent interferes only somewhat with the assay this
interference can usually be corrected for by running the standard curve in
the presence of the reagent. If the reagent drastically interferes with the
assay or the interfering reagent is present in varying amounts, it can be
removed as described below.

Method 4: Enhanced Alkaline Copper (Lowry) Protein Assay 16

Range: 2 to 100 tzg

Reagents
2× Lowry concentrate: Dissolve 20 g Na2CO3 in 260 ml water, 0.4 g
CuSO4 • 5H20 in 20 ml water, and 0.2 g sodium potassium tartrate in
20 ml water and mix the solutions to form the copper reagent. Dis-
15 C. E. Stauffer, Anal. Biochem. 69, 646 (1975).
16 O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. Randall, J. Biol. Chem. 193, 265
(1951).
58 GENERAL METHODS FOR H A N D L I N G PROTEINS AND ENZYMES [6]

solve 10 g sodium dodecyl sulfate (SDS) in 100 ml water to make a 1%

solution. Dissolve 4 g NaOH in I00 ml water to make a I M solution.
Immediately before use, mix 3 parts copper reagent with 1 part SDS
and 1 part NaOH. This reagent is stable for 2 to 3 weeks. If a white
precipitate forms warm the solution to 37°; if the precipitate is black,
discard the solution
0.2 N Folin reagent: Mix I0 ml 2 N Folin reagent with 90 ml water. This
solution is stable for several months at room temperature if stored in
an amber bottle
Procedure
1. To a 400-/zl sample add 400/zl of the 2× Lowry concentrate and
incubate at room temperature for a minimum of 10 min.
2. Add 200/zl of the 0.2 N Folin reagent and vortex immediately after
each addition. This rapid mixing is important since the reagent decom-
poses rapidly. Incubate for an additional 30 min at room temperature.
Careful timing of each sample is unnecessary providing the samples can
be read within I0 min of each other because the color is relatively stable.
3. Turn on the spectrophotometer 15 min before use. Using glass or
polystyrene cuvettes read the absorbances at 750 nm. If the absorbances
are too high, the absorbances may be read at 500 nm. Figure 2 shows the
standard curves for three different proteins.

I
B
E
~- 0.6
C)
u')
G

0.4

o
t--

-Q 0 . 2
o
m
..Q
<
! !
0 215 5=0 75 10 0
protein (~lg)
FIG. 2. Enhanced copper protein assay standard curves for three different proteins.
Bovine immunoglobulin G (I), bovine serum albumin (B), and gelatin (G) were weighed and
solubilized at 2 mg/m] in water. Aliquots of these solutions were used to produce the
standard curves shown above using Method 4.
[6] QUANTITATION OF PROTEIN 59

Comments. The above technique is modified from the original in order

to make it less sensitive to interfering buffer components and more sensi-
tive to low levels of protein. It is similar to the assays reported in Refs. 2
and 17. Despite the improvements, the BCA assay, described below, will
probably replace the Lowry since it requires that only one reagent be
added to the sample, has less protein-to-protein variability, and is linear
over a greater protein range.
The critical step for obtaining sensitive and reproducible results is to
vortex the samples immediately after the addition of the Folin reagent.
This is because the Folin reagent is reactive for only a very short time
after addition. Also examine the tubes for thorough mixing. There should
be no yellow gradient from top to bottom and the solution should appear
homogeneous.
At room temperature, the period of time for incubating a soluble pro-
tein with the copper solution is not critical and can vary from 5 min to
several hours without affecting the final absorbance.
Heating the protein with the alkaline copper solution to 100° for 20
min, as suggested in Ref. 18, is reported to equalize the reaction to differ-
ent proteins. The color decreases for the majority of proteins but for
gelatin, a relatively unreactive protein, the color increases by 50% after 10
min of heating.
Diluting the Folin reagent to 0.2 N increases the sensitivity of the
assay by about 40% over the original procedure. 19 An additional 20%
increase in sensitivity can be obtained by adding the Folin reagent in two
portions, vortexing between each addition. 2°
The assay can be accelerated by the addition of 100 /~1 20 mM
dithiothreitol 3 min after the addition of the Folin reagent. The assay can
be read 5 min after the addition of the thiol with a concomitant 50%
increase in sensitivity. 21
Special care must be given when the solution is transferred into cu-
vettes because bubbles may form, producing erroneously high readings.
Pipette these solutions gently with large-mouth pipets to reduce this
problem.
The 2 × Lowry reagent may be diluted with water to accommodate
smaller sized samples. Also, to accommodate samples dissolved in
NaOH, the Lowry reagent may be made as above, but replacing the

17 M. A. K. MarkweU, S. M. Haas, L. L. Bieber, and N. E. Tolbert, Anal. Biochem. 87, 206

(1978).
18 T. E. Dorsey, P. W. McDonald, and O. A. Roels, Anal. Biochem. 78, 156 (1977).
19 G. M. Oostra, N. S. Mathewson, and G. N. Catravas, Anal. Biochem. 89, 31 (1978).
20 H. H. Hess, M. B. Lees, and J. E. Derr, Anal. Biochem. 85, 295 (1978).
21 E. Larson, B. Howlett, and A. Jagendorf, Anal. Biochern. 155, 243 (1986).
60 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

NaOH with water, as long as the final NaOH concentration remains the
same.
An amplification technique is available to increase sensitivity32
Mechanism. Under alkaline conditions Cu 2÷ forms a complex with the
peptide bonds of proteins and becomes reduced to Cu ÷. The Cu ÷ as well
as the R groups of tyrosine, tryptophan, and cysteine residues then react
with the Folin reagent. The reagent reacts by first producing an unstable
product which is slowly reduced to become molybdenum/tungsten
blue. 21,z3,24Certain predictions can be made from this mechanism. Agents
which acidify the solution (e.g., strong acids or buffers, high ammonium
sulfate), chelate the copper (i.e., EDTA), or cause the reduction of copper
(e.g., mercaptoethanol, dithiothreitol, phenols) will interfere with the as-
say. Proteins will produce different color intensities depending primarily
on their tyrosine and tryptophan content.

Method 5: Bicinchoninic Acid (Smith) Protein Assay ~4

Range: 0.2 to 50 Izg

Reagents
Reagent A: I g sodium bicinchoninate (BCA), 2 g N a 2 C O 3 , 0.16 g
sodium tartrate, 0.4 g NaOH, and 0.95 g NaHCO3 brought to 100 ml
with water. Adjust the pH to 11.25 with 10 M NaOH
Reagent B: 0.4 g CuSO4" 5H20 in 10 ml water. Solutions A and B are
stable. The solutions are available from Pierce (Rockford, IL)
SWR (standard working solution): Mix 100 vol reagent A with 2 vol
reagent B. This solution is stable for I week and should be green in
color
Procedure
1. Add 1 ml SWR to the 20-/xl sample and mix. Incubate 30 min at 60°.
Cool the samples to room temperature and turn on the spectrophotome-
ter. If the samples have been incubated at 60° the readings will be stable
for at least I hr.
2. Read the absorbance at 562 nm using glass or polystyrene cuvettes.
Standard curves for three different proteins are shown in Fig. 3.

2z M. G. Sargent, Anal. Biochem. 163, 476 (1987).

23 G. Legler, C. M. Muller-Platz, M. Mentges-Hettkamp, G. Pflieger, and E. Julich, Anal.
Biochem. 150, 278 (1985).
24 p. K. Smith, R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, M. D. Proven-
zano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. Klenk, Anal. Biochem. 150,
76 (1985).
[6] QUANTITATION OF PROTEIN 61

E
= I
B
~D
to 2
"~ G
O
c

O
,.D
<

o 1'o 2'o a'0 4'0 5b

protein (IJg)
FIG. 3. Bicinchoninic acid protein assay standard curves for three different proteins.
Bovine immunoglobulin G (I), bovine serum albumin (B), and gelatin (G) were weighed and
solubilized at 2 rng/ml and 0.1 mg/ml in water. Aliquots of these solutions and were used to
produce the standard curves shown above using Method 5.

Comments. Sensitivity of the assay can be increased by incubating the

samples for longer periods of time. Conversely, if the color begins to get
too dark the heating can be stopped earlier.
The assay may also be performed at room temperature. Incubation at
temperatures lower than 60° results in diminished sensitivity and greater
protein-to-protein variability. This is thought to be a result of the greater
reactivity of the peptide bonds with the copper solution at high tempera-
tures.
A micromethod is also available in order to measure dilute protein
solutions. Five hundred microliters of the reagent is added to 500/zl of the
protein solution. The mixture is incubated at 60° for 1 hr, cooled to room
temperature, and the absorbance read at 562 nm. The reagent is made as
follows:
Reagent A: 8 g Na2CO3 • H20, 1.6 g NaOH, 1.6 g sodium tartrate in 100
ml water and adjusted to pH 11.25 with 10 M NaOH
Reagent B: 4 g BCA in 100 ml water
Reagent C: 0.4 g CuSO4.5H20 in 10 ml water
Working solution: Mix 1 vol reagent C with 25 vol reagent B. Then add
26 vol reagent A to the C/B mixture
These reagents are available from Pierce as a kit.
Mechanism. The BCA assay is related to the Lowry assay in that it
depends on the conversion of Cu 2÷ to Cu ÷ under alkaline conditions (see
Mechanism under Method 4). In contrast to the Lowry assay, the BCA
62 G E N E R A L M E T H O D S FOR H A N D L I N G PROTEINS A N D E N Z Y M E S [6]

e--
o 0.8

.. 0.6

o 0.4
c

0.2
0

I I I I
0 5 10 15 20
protein (ug)
FIG. 4. Coomassie Blue protein assay standard curves for three different proteins. Bo-
vine immunoglobulin 13 (I), bovine serum albumin (B), and gelatin (13) were weighed and
solubilized at 2 mg/ml and 0.1 mg/ml in water. Aliquots of these solutions were used to
produce the standard curves shown above using Method 6.

assay utilizes BCA as the Cu + capture reagent instead of the Folin-

Ciocalteau reagent. This substitution has the advantage that BCA is rela-
tively stable under alkaline conditions (the Folin-Ciocalteau reagent is
unstable at alkaline pH). Thus, the BCA can be incorporated into the
alkaline copper solution allowing for a one-step procedure.

Method 6: Coomassie Blue (Bradford) Protein A s s a y 25

Range: 0.2 to 20 tzg
Dye Reagent. Dissolve 100 mg Serva Blue G (Serva, Westbury, NY)
in a mixture of 100 ml 85% phosphoric acid and 50 ml 95% ethanol. After
the dye has completely dissolved, bring the volume to 1 liter with cold
water.
Procedure
1. Warm up the spectrophotometer 15 min before use.
2. Place 20-/.d samples into the tubes. Add 50/~1 1 M NaOH to the
sample (alternatively, the NaOH may be mixed into the reagent itself at
50/zl/ml to allow one pipetting step).
3. Add 1 ml of the dye reagent and incubate 5 rain.
4. Measure the absorbance at 590 nm in glass or polystyrene cuvettes.
Figure 4 shows the standard curves for three different proteins using
this assay.

25 M. M. Bradford, Anal. Biochem. ?3, 248 (1976).

[6] QUANTITATION OF PROTEIN 63

Comments. It is best to use disposable polystyrene cuvettes with this

assay since the dye tends to stick to the surface of the cuvette and the
absorbance of the sample changes with pipetting. If using a single cuvette,
rinse it with the reagent before putting in the blank to zero the spectropho-
tometer and pipette each sample gently. The cuvette may be rinsed with
methanol between samples. Cuvettes may be cleaned by soaking in 0.1 M
HC1 for a few hours or rinsing with concentrated detergent followed by
water and acetone.
At high protein values the level of free dye becomes significantly
depleted. The linearity of the assay can be improved by using the formula
A595 - A465 to take into account depletion of the free dye. 26
The dye reagent does not solubilize cell membranes; therefore whole
cells or membrane particulates must be solubilized with alkali or deter-
g e n t . 27
The Serva Blue G has a greater dye content than Coomassie Blue G
and produces less difference in color yield from protein to protein. 28
After several hours of incubation, macroscopic precipitates can be
observed. The presence of precipitates is not deleterious to the assay as
long as they are resuspended before measuring the absorbance. Because
of the colloidal nature of the reagent, gentle mixing of the reagent prior to
use and minimal handling of the assay samples is necessary for reproduc-
ible results.
The assay described above is a modification of the original. The im-
provements allow the solubilization of membrane proteins and lead to less
protein-to-protein variation in color yield. 29
Mechanism. The absorbance maximum of the dye in an acidic solution
shifts from 465 to 595 nm after adding protein due to stabilization of the
anionic form of the dye by both hydrophobic and ionic interactions. The
dye principally reacts with arginine residues and to a lesser extent with
histidine, lysine, tyrosine, tryptophan, and phenylalanine residues. 3°

Method 7: Colloidal Gold Protein Assay 31

Range: 20 to 640 ng
Reagent. Bring 80 ml water to a gentle boil in a scrupulously clean
250-ml flask containing a stir bar. Add 100/~1 40 mg/ml chloroauric acid

26 j. C. Bearden, Jr., Biochim. Biophys. Acta 533, 525 (1978).

27 G. O. Gogstad and M. Krutnes, Anal. Biochem. 126, 355 (1982).
28 S. M. Read and D. H. Northcote, Anal. Biochem. 116, 53 (1981).
29 C. M. Stoscheck, Anal. Biochem., in press (1990).
30 S. J. ~ompton and C. G. Jones, Anal. Biochem. 151, 369 (1985).
31 C. M. Stoscheck, Anal. Biochem. 160, 301 (1987).
64 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

with stirring. After the gold has dissolved rapidly add 1 ml 40 mg/ml
trisodium citrate dihydrate with stirring. Boil the solution for 30 min with
refluxing. Allow the colloid to cool to room temperature and add 32/zl
25% Tween 20 and mix. Add 400/.tl 1 M citric acid to the gold mixture.
The final solution should be a clear red. A cloudy solution cannot be used
and is usually a result of contaminated glassware or reagents. Store at 4 °.

Procedure
1. Place 1 ml gold reagent into 1-ml polyethylene microfuge tubes.
2. Make protein dilutions in 0.01% Tween 20 immediately before use.
Do not allow standards to sit for prolonged periods of time or freeze them
as the proteins will adsorb to their vessels.
3. Add 10/~l of each protein dilution and unknowns into the dye and
vortex immediately. Incubate 15 min.
4. Read the absorbance at 560 nm, using the blank to zero the spectro-
photometer.
5. Plot absorbance at 560 nm against nanograms protein. Figure 5
shows the standard curve for three different proteins.
Comments. The manufacturer (Janssen, Piscataway, NJ) has changed
the formulation of Aurodye which was used in the original publication 31 so
that it can no longer be used for the protein assay. Diversified Biotech
(Newton Centre, MA) offers colloidal gold specially formulated for pro-
tein assays.

E
c- B I
Q
co 0 . 4
L~ C

o
c 0.2
(o
..Q
0
CO
"Q I I I
<~ 0 160 320 480 640
Protein (ng)
FIG. 5. Colloidal gold protein assay standard curves for three different proteins. Bovine
immunoglobulin G (I), bovine serum albumin (B), and gelatin (G) were weighed and solubi-
lized at 2 mg/ml and diluted in 0.01% Tween 20 in order to produce the above concentrations
in a 10-/zl volume. Method 7 was used to produce the standard curves.
[6] QUANTITATION OF PROTEIN 65

Overloading the system will produce erroneously low protein values.

Such samples have a dark blue or gray color or a visible precipitate. At
even higher protein levels the colloid will turn red again. Use greater
dilutions in these cases to obtain the most accurate results. Be sure to
measure two different concentrations of the same solution to make sure
you are on the side of the curve which exhibits proportionate increases in
absorbance.
Proteins tend to stick to glass and plastic surfaces and at very low
concentrations losses from this process can become highly significant.
The addition of small amounts of detergents such as 0.001% Tween and
for certain proteins, NaC1, decrease protein adsorption. Dilutions of the
standards can be easily made on a piece of Parafilm.
The gold reagent contains detergent which will produce bubbles upon
too vigorous mixing or pipetting. These bubbles will cause inappropriately
high absorbance readings.
If reagents are present that interfere with the assay, a related, solid-
phase assay in which the protein is first spotted on nitrocellulose and then
the nonprotein components washed away may be used. 32
Mechanism. The binding of protein to the colloidal gold causes a shift
in its absorbance proportional to the amount of protein added to the
assay. The pH of the solution must be acidic in order for the proteins to
carry a positive charge and interact with the negatively charged colloid. 33

General Instructions

Cuvettes
Glass cuvettes may be used for wavelengths above 320 nm. Although
quartz cuvettes may be used for the entire ultraviolet to visible spectrum,
they are generally not used for visible wavelength readings because of
their cost. Also, in the Coomassie Blue protein assay, the dye sticks to
quartz more readily than glass. Disposable styrene cuvettes may be used
at wavelengths above 340 nm and acrylic cuvettes may be used at 275 to
350 nm. Disposable cuvettes are especially useful for Coomassie Blue and
colloidal gold-based protein assays since the dye-protein complexes stick
to cuvette surfaces. When using disposable semimicrocuvettes there may
be problems in precision unless a single cuvette is left in place.
Cuvettes are fragile and easily scratched. Glass Pasteur pipets should
never be inserted into nondisposable cuvettes. Bio-Rad style D polyethyl-

32 j. B. Hunter and S. M. Hunter, Anal. Biochem. 164, 430 (1987).

33 M. Horisberger and J. Rosset, J. Histochem. Cytochem. 25, 295 (1977).
66 GENERAL METHODS FOR HANDLING PROTEINS AND ENZYMES [6]

ene pipets (Cat. #223-9523) are convenient, nonscratching pipets that will
fit into standard and semimicrocuvettes. Cuvettes should be washed im-
mediately after use to prevent proteins from adhering to the inside surface
of the cuvette. A cuvette washer available at most supply houses will
minimize cuvette breakage. Always handle cuvettes by their sides to
prevent fingerprints from increasing the absorbance of the sample. If
fingerprints do get on the transmitting surfaces these may be wiped away
with 70% v/v ethanol using a lintless wiper.
Make sure the springs that hold the cuvettes in place are properly
adjusted, i.e., the cuvette should not wobble in its holder. One remedy for
a high level of variability produced by simply removing the cuvette from
its holder and replacing it is to simply leave the cuvette in place and
remove and replace solutions with a pipet.

Protein Standards
Most methods of protein quantitation compare an unknown quantity
to a standard curve of a " k n o w n " protein amount. Absolute protein
quantitation can be tricky for several reasons. First, different proteins will
react to varying extents in different protein assays. For example, l0 ~g
gelatin will produce only 69% as much absorbance as 10/zg BSA in the
Lowry assay (Fig. 2). In preparing standards, salts and moisture can add
to the weight of a protein powder. Different pH and ionic strengths can
alter the UV absorbance of proteins. For these reasons, protein values in
many publications are actually relative values and thus it is very impor-
tant that information be provided about the standard protein and assay
used.
Commonly, bovine serum albumin is used as a protein standard. Tt'iis
is because it has been used so much in the past that it makes present and
future protein determinations readily comparable. It is also easy to handle
and inexpensive. Immunoglobulin G is used less frequently, but produces
a more average absorbance for both the dye and enhanced copper protein
assays. Standards can be purchased from any of the vendors mentioned
above. Alternatively, standards can be prepared as follows: the powder
should be stored at - 2 0 ° in a desiccator. Allow the jar to warm up to room
temperature before weighing. Weigh out enough protein for accurate mea-
surement. Dissolve in water to make a 2 mg/ml solution. Aliquot into
0.5-ml fractions and store in a freezer at - 2 0 °. Commercial standards
should also be aliquoted and frozen.
When possible, the same protein as that being determined should be
used as a standard. As a general rule the protein should be extensively
dialyzed against deionized, distilled water, lyophilized, and kept desic-
[6l QUANTITATION OF PROTEIN 67

cated. It should be noted that not all proteins are soluble in water. Some
require the presence of salts, acid, base, or an organic solvent. The Merck
Index 34 lists some proteins and the methods to solubilize them. Informa-
tion on the solubility of specific proteins can be obtained from the manu-
facturer or vendor.

Removal of Interfering Substances

Neutralization. Strong acids, bases, or buffers should be neutralized
or adjusted to the optimal pH of the assay system (Lowry, pH 10; BCA,
pH 11; dye, pH 1; gold, pH 3) by the addition of HC1 or NaOH. EDTA,
which interferes with the copper-based assays by chelating the copper,
may be counteracted with the addition of equimolar CaCI2. The CaClz
precipitates in alkaline solutions and must be removed by centrifugation
before taking absorbance readings. Thiols, such as dithiothreitol, mercap-
toethanol, and cysteine, can be neutralized by the addition of 8 tzl 2 M
iodoacetic acid dissolved in 2.5 M NaOH. 35Mercaptoethanol, by virtue of
its volatility, can simply be evaporated. 36
Precipitation. Buffers, lipids, reducing agents, and certain detergents
can be removed by precipitating the protein using TCA, PCA, or acetone
and resuspending the protein in a suitable solution for the chosen assay.
The presence of detergents sometimes interferes with protein precipita-
tion depending on the detergent and the procedure used. Certain deter-
gents (e.g., deoxycholate, Triton X-100) precipitate under acidic pH con-
ditions and thus cannot be removed by the addition of TCA or PCA. Acid
precipitations result in pellets which are difficult to dissolve unless a
coprecipitant is used. Sonication and incubation in 1 M NaOH will pro-
mote dissolution of the pellet. These techniques are described in greater
detail in Refs. 2, 16, and 37-39.
Dialysis and Ultrafiltration. The time required for dialysis is inversely
dependent on pore size and volume. Using small volumes and large pore
sizes, dialysis can take just a few hours. Ultrafiltration is still faster and
can take as little as 30 min to perform. Caution must be used if detergents
are present since they can form high-molecular-weight micelles and their
capacity to pass through membranes depends on their micellar molecular
weight rather than the molecular weight of a single molecule. Dialysis and

34 M. Windholz, "The Merck Index," 10th Ed. Merck & Co., Rahway, New Jersey, 1983.
35 E. Ross and G. Schatz, Anal. Biochem. 54, 304 (1973).
36 H. P. S. Makkar, O. P. Sharma, and S. S. Negi, Anal. Biochem. 104, 124 (1980).
37 A. Bensdoun and D. Weinstein, Anal. Biochem. 70, 241 (1976).
3s I. Polacheck and E. Cabib, Anal. Biochem. 117, 311 (1981).
39 D. Wessel and U. L. Flugge, Anal. Biochem. 138, 141 (1984).
68 G E N E R A L M E T H O D S FOR H A N D L I N G PROTEINS A N D E N Z Y M E S [7]

ultrafiltration with and without detergents is discussed in greater detail in

Refs. 40-44.

Instrumentation
The choice of the spectrophotometer purchased depends on the types
of protein assays planned. The most versatile instrument is one with both
visible and ultraviolet light sources. The Spectronics 20 spectrophotome-
ter is convenient for routine assays using light in the visible range since
Fisher disposable 13 x 100 mm culture tubes can be used as disposable
cuvettes. A relatively new, convenient innovation is the microwell plate
reader which automates spectrophotometric readings. All of the colori-
metric procedures described above have been modified for these plate
readers, i.e., Lowry, 45 BCA, 46 Bradford, 47 colloidal gold. 4s The microwell
plate reader does not have ultraviolet light capabilities. It should be noted
at what point the detector can no longer measure absorbance in a linear
fashion. This level is usually about 2.

Acknowledgment
Supported by the Veterans Administration.

40 K. C. Retz and W. J. Steele, Anal. Biochem. 79, 457 (1977).

41 A. J. Furth, H. Bolton, J. Potter, and J. D. Priddle, this series, Vol. 104, p. 318.
42 L. M. Hjeimeland and A. Chrambach, this series, Vol. 104, p. 305.
43 A. Helenius, D. R. McCaslin, E. Fries, and C. Tanford, this series, Vol. 61, p. 734.
44 A. J. Furth, Anal. Biochem. 109, 207 (1980).
45 H. J. Fryer, G. E. Davis, M. Manthorpe, and S. Varon, Anal. Biochem. 153, 262 (1986).
46 M. G. Redinbaugh and R. B. Turley, Anal. Biochem. 153, 267 (1986).
47 M. G. Redinbaugh and W. H. Campbell, Anal. Biochem. 147, 144 (1985).
4s T. Ciesiolka and H. Gabius, Anal. Biochem. 168, 280 (1988).

[7] C o n c e n t r a t i o n o f P r o t e i n s a n d R e m o v a l o f S o l u t e s
B y THOMAS POHL

The removal of low-molecular-weight solutes and concentration of

protein solutions are the most frequently performed operations in protein
purification and analysis. The diversity of applications and the multitude
of techniques available today render the selection of the optimal method
in a particular case a difficult task for the investigator. There is no gener-

Copyright © 1990by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 182 All rights of reproduction in any form reserved.