CLIN. CHEM.
32/8, i55i-i 544 (1986)
UrinaryProteinas Measuredwitha PyrogallolRed-MolybdateComplex,
Manuallyand in a Hitachi726 AutomatedAnalyzer
Nobuko Watanabe,’ Sachiko KameI,1 Aklyukl Ohkubo,1Manabu Yamanaka,’ Susumu Ohsawa,2 Kazuo
Maklno,3 and Kunlaki Tokuda3
in this new method for determining urinary protein, the
reaction is compiete within 10 mm at 37 #{176}C.
This method is
applicable to automated as well as manual measurements.
Proteinconcentration and absorbanceat 600 nm are linearly
related throughout a wide range of concentrations, 10 to
16 000 mg/L. However,the chromogenicityof the y-globulins
in this method is 70% of that of albumin,as estimatedfrom
resultsby a biuret method.Within-runCVs were <3.3%; the
day-to-day CV was 2.9%. Errors due to interfenng compo-
nents in urine are <2%. The normalrangefor urinaryprotein
as measured by this method was from 28 to 141 mg/day.
Results by this method (y) and by a tnchloroacetic acid-
biuret method ) correlated well (n = 80, r = 0.995; y =
0.99x - 2.9).
AdditionalKeyphraaes:colorimetry results for albumin and
globullnscompared ‘ biuret method compared reference in-
teival
Many methods for measuring protein in urine have been
reported, including some automated ones (1-4). However,
none is completely satisfactory because the range of linear-
ity in the calibration curves in these methods is too narrow
or because repetitive measurements are disturbed by the
shift of the baseline caused by (dye) reagents adhering to the
wall of the cuvet. To developa method applicableto auto-
mated analyses, we studiedthe color reaction between a
pyrogallol red-molybdate complex and protein, originally
reported by Fujita et al. (5). This method is based on
measuring the shift in the absorption spectrum of the
complex that occurs when the pigments are bound to protein
molecules:the absorbance peak shifts from 480 to 600 nm at
pH 2.5. Here we report a modification of the method of
Fujita et al. and describe its application to a discrete
analyzer, the Hitachi 726.
Materials and Methods
As the standard proteins, we used albumin (Cohn Frac-
tion V; Sigma Chemical Co., St. Louis, MO), human’
globulin (The Green Cross Corp., Osaka, Japan), and Bence
Jones protein (Behringwerke AG, Marburg, F.R.G.). The
protein concentrations of the standard solutionswere deter-
mined by a biuret method in a sarc analyzer (Technicon
Instruments Corp., Tarrytown, NY). Pyrogallol red (PR)
‘The Central Laboratory, University of TokyoHospital,Faculty
ofMedicine,University of Tokyo,7-3-1, Hongo,Bunkyo-ku,Tokyo.
Division of Laboratory Medicine, University Hospital,
School of Medicine, Chiba University, 1-8-1 Inohana, Chiba-shi,
Chiba.
3Wako Pure Chemical Industries, Ltd. 3-10, Doshomachi,Higa-
shi-ku, Ohsaka-shi,Ohsaka.
Received August 26, 1985; acceptedMarch 7, 1986.
was obtained from Dojin Kagaku Kenkyusho, Co., Kuma-
moto, Japan. Succinic acid, sodium molybdate, sodium ben-
Downloaded from https://academic.oup.com/clinchem/article-abstract/32/8/1551/5652690 by guest on 25 February 2020
zoate, sodium oxalate, and methanol were all from Wako
Pure Chemical Industries, Ltd., Osaka, Japan. All reagents
were “special” grade. Urine control material was obtained
from Cooper Biomedical, Inc., Diagnostics Division, Mal-
vern, PA.
Reagents
PR solution (1.5 mmol/L): Dissolve 60mg ofPR in 100 mL
of methanol.
Molybd ate (10 mmol/L) solution: Dissolve 0.24 g of disodi-
um molybdate’ 2 1120 in 100 mL of distilled water.
PR-molybdate reagent: Dissolve 5.9 g of succinic acid,
0.14 g of sodium oxalate, and 0.5 g ofsodiumbenzoatein 900
mL of distilled water, add 40 mL of the PR solution and 4
mL of the molybdate solution, adjust to pH 2.5 with 0.1
mol/L HC1 solution, and dilute to 1 L with distilled water.
Procedures
Manual measurements: Mix, separately, 50 ,uL of urine
sample or of a standard 1 g/L albumin solutionwith 3.0 mL
of the PR.-molybdate reagent and let stand for 20 mm at
room temperature. Measure the absorbance of each at 600
nm vs the reagent and calculate the protein concentrations
of the samplesby comparing with a standard curve previ-
ously prepared from data on the albumin standard solu-
tions.
Automated assay with the Hitachi 726: Dilute the urine
samples fivefoldwith distilled water. In the machine, 20 L
of the diluted sample is mixed with 1.25 mL of PR.-
molybdate reagent. The sample movesfrom this point to the
absorbance-measurement position in 10 miii and 48 a, the
temperature being maintained at 37#{176}C.The absorbance is
measured at 600 nm. The concentrations of protein are
calculated by comparison with a standard curve.
Results
Absorption Spectrum
Figure 1 shows the absorption spectra of the reagent and
of the reagent plus albumin, and the difference spectrum.
When the pigment binds to albumin, the absorptionpeak for
the pigment shifts from 467 rim to 598 nm. The absorbance
of the difference spectrum at 600 nm is proportional to the
concentration of albumin.
Analytical Variables
Buffer. To optimize the buffer used in the PR-molybdate
reagent for absorbance at 600 rim and color stability, we
studied acetate, succinate, phthalate, citrate, and glycine
buffers at pH 2.0-4.0. We found the sodiumsuccinate/suc-
CLINICALCHEMISTRY,Vol. 32, No. 8, 1986 1551
 Effect of temperature. The absorbanceof the PR.-molyb-
date reagent at 600 mn decreased when the reaction tem-
perature was increased from 12#{176}C to 25#{176}C,
but did not
change between 25 and 37#{176}C. The absorbance at 600 tim of
the reaction mixture also remained unchanged throughout
a range of temperature from 25 to 37#{176}C.
In the succeeding
experiments we measured samples at 37#{176}C.
Effect of pH on absorbance. The absorbance at 600 nm of
an albumin-reagent mixture was maximum when the pH of
the mixture (which is also the pH of the reagent)was
between 2.5 and 3.0. However, the absorbance of a y-
globulin-reagent mixture exhibited the maximum at a pH
between 2.25 and 2.5 (Figure 3). We selected pH 2.5 as the
optimal pH for the succinate buffer.

Downloaded from https://academic.oup.com/clinchem/article-abstract/32/8/1551/5652690 by guest on 25 February 2020

400
Wavelength (nm)
Fig. 1. Absorptionspectra of (A) reagent,( a mbctureof 4 g/L albumin
solution and the reagent,(C) the difference between the mixture and
the reagent
cinicacid buffer (50 mmol/L, pH 2.5) to be the best for this
purpose, and we used it in the succeeding experiments.
Optimizing the concentration of the PR-molybdate rea-
gent. When the concentration of molybdate in the PR-
molybdate reagent was fixed at 40 .anol/L, this method was
most sensitivefor determinationof protein in the range of
PR concentrations between 60 and 120 pmol/L (Figure 2).
Because the absorbance at 600 rim of the reagent with 40
pmol of molybdate per liter increases linearly in proportion
to the concentration of PR, the lower the concentration of PR
the better is the spectral resolution in this method. Similar-
ly, when the concentration of PR in the reagent was 60
.imol/L, the assay was most sensitive with the concentration
of molybdate between 40 and 160 mo1/L (Figure 2). To
minimize the absorption of the reagent and attain better
Effect of chelating substances in urine. When urine sam-
ples that were negative for protein by a sulfosalicylic acid
method were mixed with PR-molybdate reagent, the absor-
bance of the mixture at 600 rim decreased below that of the
reagent. Substances in urine that chelate molybdate suppos-
edly are responsible for this effect. The absorbance of PR-
molybdate reagentat 600 nm decreased when oxalate was
added to the reaction mixture (Figure 4). The decrease in
absorbance became larger as the concentration of oxalate in
the reaction mixture was increased up to 1 mmol/L, at which
point the effect of oxalate became essentially maximum. The
endogenous oxalate in urine, if less than 3 mmol/L (less
than 10 imoi/L in the reaction mixture), or citrate, if less
than 2.5 mmol/L (less than 8 Mznol/L in the reaction
mixture), did not decrease the absorbance further in the
presence of 1 mmol of added oxalateper liter. To eliminate
negative errors caused by chelating substances in urine, we
add 1 mmol of oxalate to the reagent solution per liter.
Time course of the reaction between protein and the
complex. The reactionwas followed at 37#{176}C by the change in
absorbance at 600 rim after a sample solution was mixed
with the PR.-molybdate reagent. The absorbance stopped
0.3
E
C

resolution, we decided on 60 pmol of PR and 40 tmol of 0

0
molybdate solution per liter as the best composition of this 100
reagent. a
U

0.4
I
Albumin 0.1
0.3 50

0.2

0.1
,Reagent

1.0 2.0 3.0 4.0 5.0

I
0 01 0:2 0.3 pH
PR conc. (m moIJL) Molybdate conc. I,, n,l/L)
Fig. 3. Effectof pH on the absorbanceat 600 nm of an albuminor y
Fig.2. Effectoftheconcentrationof reagentson the absorbanceat 600 globulin solution
nm of the reaction mixturewithan albumin solution at pH 2.5 (A) Medure ofalbumin solutionandthereagentvs thereagent;( mbdureof
Left:Effectofvariousconcentrations
of PR ata molybdate
concentration
of 40 globulinsolution and reagent vs the reagent; (C) the calculated ratio of the
ano4JLRht: Effectofvariousconcentrationsofmolybdateata PRconcentra- absorbance ofBto that of A. expressedaspercent(right-,’andordinate,albumin:
tionof 60 imol& 100%)

1552 CLINICALCHEMISTRY, Vol.32, No. 8, 1986

 0.31 Table 1. SubstancesTestedfor PotentialInterference
with Measurementof ProteinConcentration
Concn Prot&n concn
Substance tasted,g/L observed, g/l.
None 2.17
0 2 Chlorpromazine
Caffeine 0.2
0.6 2.19
2.17
Ampicillin 0.4 2.14

o
C)
C
\ L-DOpa
Cefazolin sodium
0.03
2
2.16
2.17
1

0.1 B Latamoxef
Gentamicinsulfate
sodium
Phosphate,inorg. 20.2
8 2.17
2.23
2.18
Ca 2 2.19
Mg2 4 2.18

Downloaded from https://academic.oup.com/clinchem/article-abstract/32/8/1551/5652690 by guest on 25 February 2020

_____ ________________
_______________________________________
Creatinine
Urea
Glucose
3
30
40
2.19
2.19
2.21
0 UrIc acid 4 2.15
4- 0.51 2 5 10 Citrate, sodium 0.5 2.13
Oxalate (m mol/L) Oxalate, sodium 0.3 2.17
Ascorbate, sodium 5 2.15
Fig. 4. Effect ofthe concentration
of oxalateon the absorbanceat 600 HCI, 2 moi/L 0.04a 2.12
nm of: (A) the mixtureof an albuminsolution plusreagentvs the NH3, 0.5 mot/L 0.06 2.26
reagent,( the reagentvs water,and (C) a urinesample plusreagent
vs the reagent amL

increasing in 8 mm. The change in absorbance at 600 rim

Table 2. ReproducIbility
was routinely measured 10 mm after mixing.
Linearity. Using albumin solutions as test samples, we SD
observed linearity throughout for concentrations from 10 to n mg/I. CV,%
16 000 mgIL-.-i.e., from 0.032 to 51 ,ug/mL of reagent Within-run
(Figure 5). 40 89 3.0 3.3
Interfering materials. Table 1 summarizes our interfer- 20 590 9.5 1.6
ence study. Strong acids or alkalis gave negative or positive 40 2080 26.9 1.3
20 4110 42.9 1.0
errors through the shifting of the pH of the reaction mixture
down or up. However, this hardly can be expected to happen Between-run
in actual urine tests. The substances ordinarily present in 40 800 23.3 2.9
urine apparently would affect the measurement by less than 40 1430 43.3 3.0
2%.
Reproducibility. Four different samples were used to
study reproducibility of this method (Table 2). The within- 3.25 g/L). The observed values were 97-102% of the calcu-
run CVs were <3.3%. The day-to-day CVs for the control lated values for albumin, and 69-71.5% for y-globulin.
urine and the in-house pooled urine were 2.9% (n = 40,1= Stability of the reagents. When the reagents are protected
0.8 gIL) and 3.0% (n = 40, 1 = 1.43 g/L), respectively. from light, they can be used, stored at room temperature, for
Analytical-recovery study. We added 1 mL of 1 or 2 g/L as long as six months.
solutions of human serum albumin or y-globulin to 9 mL of
each of two urine samples (protein concentrations: 2.08 g/L, Comparisons with a Biuret Method
Substantial differences are seen in measurement for vari-
15.0 ous proteins as measured with assays based on dye-binding
J
capacity of protein. In contrast, the concentrations of albu-
B min as measured by the present method were the same as
C
a
those determined with the biuret method. But the concen-
4)
0 trations of y-globulin as determined by the present method
0. were only 70% as great as those measured by the biuret
4-
0 method, Bence Jones protein of the X type 52% and of the K
4-)
type 68%, and hemoglobin 98%, respectively. For urine that
a contained either Bence Jones protein or hemoglobin, values
4-)
C
a obtained by the present method were 59 to 98% as great as
C
C
0
those determined by the biuret method. However, for other
C) urine samples collected from patients, the values deter-
mined by the present method (y) correlated well with those
(x) obtained by the biuret method: n = 80, r = 0.995, y =
Dilution Dilution 0.999x - 2.9, range 50-5800 mg/L.
Normal Range for Urinary Protein in Adults
Fig. 5. Unearity of assay of thismethod
Ordinate:concentrationsof protein.Abscissa: dilution facrs of the albumin We used urine samples collected from healthy adults (37
solutions.
Concentrationsof albumin:
(A) 0-15 g/L; ( 0-1.22 g/L; (C) 0-0.1 g/L men, 23 women, ages 24-56 years) to assess the normal

CLINICAL CHEMISTRY, Vol. 32, No. 8, 1986 1553

referenceintervalfor urinary protein as determined by this
method. It was 28-141 mg/day for each sex.
Discussion
Protein concentration in patients’ urine ranges from al-
most null to about 60 g/L. Many substances in urine can
interfere with the reaction between protein and reagent.
Moreover, many kinds of dye show a strong tendency to
stick to glassware, including cuvettes. For these reasons, no
really satisfactory colorimetric method has yet been devised
for assay of protein in urine, especially in an automated
analyzer system.With the present method protein concen-
trations in urine ranging from 10 to 16000 mg/L can be
measured, without dilution of the sample. The reagent does
not adsorb onto the wall of cuvettes, so this method is
applicable to an automated analyzer of the discrete type.
Negative errors due to possible interferents, oxalate and
other chelating substances in urine, could be eliminated by
prior addition of oxalate to the reagent to give a 1 mmol/L
final concentration.
The difference in the reactivity for various kinds of
protein in this method was compared with that in a biuret
method. Any current methods determine the concentration
of protein in body fluid by measuring one kind of physico-
chemical property of protein, so the results obtained by
different methods on different bases do not usually agree
CLIN. CHEM. 32/8, 1544-1556 (1986)
PolyethyleneGlycolCan Be ValidlyOmittedfromChromogenicPeptide
SubstrateAssayfor AntithrombinIll
H.-J. Kolde
To investigate whether the characteristics of a commercial
test kit for antithrombin Ill (Berichrom#{174}
Antithrombin Ill) could
be influenced by surfactants such as Tween-80 or polyethyl-
ene glycol (PEG), we performed some experiments with the
original kit reagents and with the reagents dissolved in
surfactants. Neither the reliability of the calibration curvenor
the data for precision and assay kinetics were amended by
the addition of either PEG to the (human) thrombin reagent or
Tween-80 to the chromogenic substrate. In the same test
system, assays with some other chromogenic substrates and
with bovine thrombin showed comparable behavior. Evident-
ly, if one follows the working scheme proposed for this kit, the
use of surfactants is not warranted.
AddItIonal Keyphrase: coagulation assays
In a recent article Grain and Jespersen (1) described the
influence of polyethylene glycol (PEG) and surfactants on
chromogemc substrate assays for antithrombin ifi.’ These
Behringwerke AG, D-3550 Marburg, F.R.G.
‘Nonstandard abbreviations:pNA, 4-nitroaniline; CHG, cyclo-
hexylglycine; Toe, tosylsulfonyl-; But, aipha-aminobutyric acid; Pip,
t-pipecolinicacid;and PEG, polyethylene glycol.
Received March 13, 1986; acceptedMay 27, 1986.
1554 CLINICALCHEMISTRY, Vol. 32, No. 8, 1986
with each other. Thus we think that this method correlate
well with the biuret method in measuring protein concen-
trations in urine samples collected routinely from patients.
Reproducibility and recovery studies proved this method
to be adequate for clinical use. Even with the problem of
different reactivity of various proteins to the reagent being
unsolved, the present method is thought to be useful for
automated measurement of urinary protein in the clinical
laboratory.
References
1. Lustenberger P, Launay-Godard A, Cornu G, Bernard S. Auto-
masation en analyse centrifuge et enflux continu du dosage des
proteines totales urinaires. Clin Chim Acta 1978;84:293-303.
Downloaded from https://academic.oup.com/clinchem/article-abstract/32/8/1551/5652690 by guest on 25 February 2020
2. Joern WA, Schmoele L. Urinary protein measurement by the
Coomassie blue dye-binding method adapted to the ABA-l00 bi-
chromaticanalyzer [Letter]. Clin Chem 198 1;27:1305
3. Sano K, Kanamori K, ShibaA, Nakao M. Automatic assay of
urinary protein using CoomassieBrilliant Blue G-250. Anal Bio-
chem 1981;113:197-201.
4. Watanabe N, Kamei 5, Ohkubo A, Yamanaka M. Automatic
method for determination of protein in urine with a Hitachi 726
analyzer: a modification of Bradford’smethod. Aiial Chem Speci-
men 1984;7:27-31 (in Japanese).
5. Ftjita Y, Mon I, Kitano S. Color reaction between pyrogallol
red-molybdenum(VI) complex and protein. Bunseki Kagaku
1983;32:E379-E386.
assays make use of the ability of heparin to stimulate the
inhibition of serine proteases such as thrombin (EC 3.4.21.5)
or Factor Xa (EC 3.4.21.6). Most procedures are based on the
original method described by Abildgaard et al. (2) and
usually involve bovine thrombin. In a former paper von
Voorthuizen and Kluft (3) reported improvingthe linearity
of the assay calibration curve for a commercial kit involving
bovine thrombin and the chromogenic substrate S-2238 (H-
D-Phe-Pip-Arg-pNA) by adding PEG or Tween-80. These
additives minimized the deviation of the zero point (throm-
bin blank) from the standard curve, which improved the
reliability of the results, because with most automated
instruments the enzyme blank must be included in calculat-
ing the results.
Gram and Jesperson (1) performed a detailedstudy of
several additives, using a modification of the procedure of
Abildgaard et al. (2)and the chromogenic peptide substrates
S-2238, TH1 (H-D-CHG-Ala-Arg-pNA), Chromozym#{174}TH
(Tos-Gly-Pro-Arg-pNA), and H-D-CHA-But-Arg-pNA. The
latter substrate is included in the test kit Berichrom#{174}
Antithrombinifi (Behringwerke AG, Marburg, F.R.G.). The
Berichrom kit differs from the original method (2) in several
respects (Table 1) and does not contain PEG. The character-
istics of this assay have been described elsewhere (4, 5). To
determine whether PEG might also have a significant