A Biuret Method for Determination Protein in Normal Urine John Savory,

of

Pin H. Pu, and F. William Sunderman, Jr.

A biuret method has been developed which provides quantitative measurements of protein in normal urine without interference from drugs or pigments. This method is intended for use in monitoring clinical trials of new drugs-to detect nephrotoxicity. Protein is precipitated from duplicate samples of urine by addition of cold ethanolic phosphotungstic acid. The protein precipitates are separated by centrifugation and washed with ethanol. Protein from one of the duplicate samples is dissolvel in biuret reagent. Protein from the second sample is dissolved in an alkaline tartrate reagent which is identical to the biuret reagent, excepting that copper sulfate has been omitted. After 20 mm., the differential absorbance of the two samples is measured at 540 m. The limit of sensitivity for detection of protein in urine is 0.5 mg./100 ml. The coefficient of variation of replicate analyses of protein in normal urine is 4.2%. The recovery of protein added to urine averages 103 ± 3%. Analyses of urinary protein by the biuret procedure provide close correlation with measurements by an amido black staining method. Systematic search has failed to reveal interference from urinary pigments, compounds, or drugs which are normally or occasionally encountered in hospitalized patients. In 24-hr. collections of urine from 28 healthy adults, the protein concentrations averaged 6.2 mg./100 ml. (range 3.0-12.2), and the protein excretions averaged 77 mg./day (range 40-150).

R.

INTEREST IN QUANTITATIVE METHODS foi’ deterniination of prourine from lmealthy subjects has been stimulated by tile revised regulations of the Food amid Drug Administration regarding clinical trials of new drugs (1). The protocols for many clinical trials currently specify analyses of urinary protein, inasmuch as increased excretion of protein in urine is one of time most dependable indications of nephrotoxicity (2). Fnfortunately, the methods employed for measECENT

teill

From

ill

the

Pathology

Department,

University

of

Florida

College

of

Medicine,

Gainesville,

Fla.32601. Supported by U. S. Atomic Energy Commission Grant AT-(40-1).3461; by American Cancer Society Grant E.374B; amid by U. 5. Public Health Service Research Grant CA-08783-02 from the National Cancer Institute. Presented at the scientific sessions of the American Association of Clinical Chemists at the American Association for the Advancement of Science meeting, New York, N. Y., Dec. 27, 1967. Received for publication Feb. 2, 1968; accepted for publication Mar. 31, 1968. 1160

Vol. 14. No. 12, 1968

URINARY

PROTEIN

1161

urements of urinary protein are often subject to interference from drugs and drug metabolites. For exanmple, colorimetric determinations of protein by the Folin-Lowry phenol reaction may be affected by streptomycin, sulfonamides, salicylates, p-aminosalicyclic acid, phenacetin,and chiorpromazine (3). Similarly, turbidimetric estimatioims of protein by the sulfosalicyclicacid procedures are subject to interference from organic iodine compounds, and from i-butyl-3-carboxyphenylsulfonylurea, an excretory product of tolbimtamide (3). The most sensitive methods for determinatioll of urinary protein involve drying specimens of urine on strips of filter paper and staining the dried proteins with amido black (4, 5), bromphenol blue (6), or light green (7) dyes. The dye which becomes bound to the proteins is either estimated directly by densitometry or is eluted and measured by colorimetry. Although interference from drugs is apparently not a serious problem with the dye-binding methods, such methods are prone to marked day-to-day variability and are inconvenient for routine use in clinical laboratories. Moreover, the dye-bindilmg methods are subject to systematic errors owing to differencesin time relative affinities of the protein fractions for amido black or bromphenol blue (8). Several investigators have employed the biuret reaction for qllantitative measurements of urinary proteins in patients with a pathologic process (9-14), since the biuret reaction is relativelyspecificfor proteins and yields equivalent chromogenicity with albumin and globulin fractions (15). The biuret reaction has been reported by Piscator (16) to be applicable to measurements of protein in normal urine, and has been used successfully by him for the detectiolm of renal toxicity in industrial workers who are exposed to cadmium. The I’iscator procedure (16) has been employed in our laboratory during the past 2 years; it has undergone a lmumber of modifications which have improved its sensitivity and precision and which Imave Inilmilnized immterference from drugs and pigments.

Method Principle

After centrifugation of urine to remove sediment, the protein is precipitated from duplicate 20-ml. samples of urine by addition of an equal volume of cold ethanolic phosphotungstic acid. After ii mm. in an ice bath to ensure complete precipitation, the protein precipitates are separated by centrifugation and washed with cold ethanol. Protein from one of the duplicate samples is dissolved in 4 ml. of biuret reagent. Protein from the second sample is dissolved in 4 ml. of an alkaline tartrate reagent which is identical to the biuret reagent, except

1162

SAVORY

that copper sulfate ture, tile differential 540 m.

fT

AL.

Clinical

Chemistry

has been omitted. After 20 mm. at room temperaabsorbance of the two samples is measured at

Reagents EthanoliC phosphotungstie aCid Transfer 50 ml. of cone. hydrochloric acid (s.g. 1.19), 60 ml. of distilled water, and 770 ml. of 95c/ (v/v) ethanol into a 1-L. beaker. Add 15 gm. of phosphotmmgstic acid and dissolve. The solution is filtered and stored in the refrigerator.

Ethanol,

absolute

Sodium

GlIb ride

Store solution,

in the refrigerator. 0.85%

(u/v)

Biuret reagent Add 9.6 gm. of potassinm sodium tartrate (KNaC4H1O.4H00) to a 2-L. graduate cylinder with ground-glass stopper and dissolve in 400 ml. of distilled water. Add 2.4 gin, of cupric sulfate pentahvdrate (CuSO4.5H00) and dissolve. Slowly add 360 ml. of 2.5 N sodium hydroxide followed by 1.0 gin. of potassium iodide. Dilute the solution to 1500 ml. with distilled water. The rcagent is stored in a plastic bottle and is stable for several months at room temperature. Alkaline tartrate reagent This solution is prepared in exactly the same manner as the biuret reagent, except that the cul)ric sulfate is omitted. Protein standards The concentration of protein nitrogen in a sample of pooled human serum is determined by the Kjeldahl procedure (17). The total protein concentration is calculated usilmg time factor of 6.54 gm. of protein per gram of nitrogen (18). The serum protein standard is stored in small ampoules at -10#{176}.An appropriate dilution of the protein standard is made with 0.85% (w/v) sodium chloride solution to achieve a flumal concentration of 10-15 mg. of protein per 100 ml. The dilute protein standard is stored at 4#{176} and is stable for 1 week.

SpecialApparatus Centrifuge tubes, 50 ml., with ground-glass stoppers (Cat. No. 45168, Kimble Glass Co., Toledo, Ohio) These tubes fit centrifuge cups (No. 367), with rubber cushions (No. 575), metal shields (No. 572), aimd trunnion rings (No. 366)-International Equipment Co., Boston, Mass. Procedure The aliquot

volume of the 24-hr. of

Urille

is centrifuged

collection at

of urine is measured. 2000 rpm for 15 mill.

A 50-mi. The

super-

Vol. 14, No. 12. 1968

URINARY

1163

PROTEIN

liatant urine is decanted into a beaker alldis used for the analysis. The protein concentration in tilecentrifuged urine sample is estimated by means of Albustix test strips (Ames Co.). If the Albustix reaction result is read as “negative” or “trace,” 20 ml. of urine is taken for analysis. If the Albustix test gives qualitative reaction results of 1+, 2+, 3+, or 4+, the following volumes of urine should be taken: 10, 5, 2, and 0.5 ml., respectively. The requisite volumes of urine are transferred into duplicate 50-mi. centrifuge tubes and are diluted (if necessary) to 20 ml. with 0.85% (w/v) sodium chloride solution. Into two additional pairs of 50-ml. centrifuge tubes are transferred (in duplicate) 20 ml. of sodium chloride solution (blank samples) and 20 ml. of dilute protein standard solution (standard samples). The centrifuge tubes are placed in an ice bath for 5-10 mm. A 20-ml. portion of cold ethaumolic phosphotungstic acid reagent is added to each of the tubes. The contents of tile tubes are mixed, and the tubes are allowed to stand in the ice bath for 15 mm., at the end of which time they are centrifuged at 2000 rpm for 15 mm. Tile supernatant liquids are discarded by decantation, and the tubes are inverted to drain onto filter paper. Cold ethanol (10 ml.) is added to the tubes, and the protein pellets are dispersed by means of a Vortex rotary mixer. The tubes are centrifuged at 2000 rpm for 10 mm., and the ethanol is discarded by decantation. The tubes are again inverted to drain Ollto filter paper. Biuret reagent (4 ml.) is added to one of each pair of duplicate tubes; 4 ml. of alkaline tartrate reagent is added to the remaining tubes. The protein precipitatesare dissolved by agitation with a Vortex rotary mixer. Time tubes are allowed to stand at room temperature for 20 mm., which is the time required for maximum color development. The color is stable for at least 2 hr. Time contents of each tube are transferred to a Coleman spectrophotometer cuvet (19 mm. diameter). An adapter is placed in the bottom of a Coleman cuvet holder so that the absorbance of a 4-mi. volume can be measured. A Coleman Junior spectrophotometer is adjusted to zero absorbance at 540 m, using the cuvet which contains the blank sample aimd alkaline tartrate reagent. The absorbarices of the remaining samples are then measured and recorded. Calculations Protein

cone.

(nmg./100

ml.)

=

X

-

std.

cone.

where: =

absbi’banc#{234}of urille

sample

treated

with

biuret

reagent

1164

SAVORY

=

A8 A88

=

A88

=

=

absorbance reagent absorbance absorbance reagent absorbance

of urine of protein of protein of blank

fT AL.

saml)le

treated

standard standard sample

Clinical

treated treated treated

with

alkaline

Chemistry

tartrate

with biuret reagent with alkaline tartrate with

biuret

reagent

Results Modifications of Piscator Method

The following are the principal modifications of the Piscator method (16) which were introduced in the present method: (1) The volume of urine was increased from 2 to 20 ml., in order to increase the sensitivity of the method. (2) Protein precipitation was performed in an ice bath instead of at room temperature, since it was found that the yield of protein from seven normal samples of urine averaged 6.1% (range 2.6-10.2%) higher when precipitation was performed at 00 rather than at 25#{176}. Prolongation of the precipitation period from 15 mm. to 0.5, 1, 2, 4, or 18 hr. did not ilmcrease the yield of protein. (3) The protein precipitate was washed once instead of twice, since a second ethanol washing did not significantly affect the analytic results. (4) Spectrophotometry was performed instead of at 330 m, in order to avoid violet spectrophotometer almd to avoid absorb light at 330 m1z. (5) Pigment preparation of urine “blanks” which samples, except that copper sulfate is

at the 540 m absorption peak the imecessity of using an ultrainterference from drugs which interference was prevented by are identical to the urine biuret omitted.

Sensitivity, Precision, and Recovery Studies

The limit of detection for urine protein (i.e., the protein concentration which gives a corrected absorbance value of 0.01) was 0.5 mg./100 1111. As illustrated in Fig. 1, the calibration curve was linear with protein concentrations ranging up to at least 27 mg./100 ml. The coefficient of variation of duplicate analyses of 24-hr. urine collections from 28 healthy adults was 4.2% (mean concentration 6.2 mg./100 ml.; standard deviation of duplicates ± 0.26). Measurements of recovery of serum protein added to seven normal urine sanlples in a concentration of 15 mg./100 ml. averaged 103% (standard deviation ± 3) with a range of 98-109%. Preservation of Urine Specimens

Urine specimens were stored in a refrigerator at 4-10#{176} during the 24-hr. collection period and until the time of analysis. No preserva-

Vol. 14, No. 12, 1968

URINARY

1165

PROTEIN

tives were added. The stability of urinary protein concentration was studied by repeated analyses of a specimen of urine which was kept for 1 month at 4-10#{176}. Protein determinations were performed three times each week for 4 weeks. The 12 measurements yielded a mean 0.6-

0.5-

Fig. curve

1. for

of urinary biuret

Calibration

0.4-

measurementa

protein

by the

procedure.

0.3-

0.2

0.

G

I

0

5

0

15

20

Protein

1mg 1100 ml)

25

30

protein concentration of 39.3 mg./100 ml. (S.D. ± 0.5) with a range of 38.5-40.5 mg./100 ml. No consistent increase or decrease in protein concentration was observed during the period of observation. The stability of protein concentration in urine which was kept frozen at -15#{176} was investigated by performing five replicate analyses of a urine specimen before and after freezing for 1 month. The mean protein concentration was 40.5 mg./100 ml. before freezing and 40.7 mg./100 ml. after freezing. Comparisons with Other Methods

Concentrations of protein lii 24-hr. urine collections from 14 healthy adults were measured by the biuret procedure and by the amido black method of Kaltwasser et al. (5). As shown in Fig. 2, urinary protein levels measured by the amido black method averaged 8% higher than those measured by the biuret procedure. The correlation coefficient was 0.96, and the standard error of estimate was 0.77. Protein concentrations in 24-hr. urine collections from 28 healthy adults were measured by the biuret procedure and by the sulfosalicyclic acid method of Poortmans and van Kerchove (4). As shown in Fig. 3, there was apparently a random scattering of results of protein measurements by these two technics.

1166

SAVORY

fT AL.

Clinkal

Chemistry

Normal Values for Urinary Protein

Collections of urine (24 hr.) female adults who were judged history and physical examination.

were obtained from 14 male and 14 to be healthy on the basis of medical The mean concentration of urinary

14

20

Fig. Slope

of Regression -1.08

2.

Comparison

of

protein concentration measurements in 24-hr. urine coflections from healthy adults by biuret method and by amido black procedure Kaitwasser et at. Broken tine indicates

o

culated solid

regression line

theoretical

14 the the of (5). cal-

line;

represents

the

relationship

of

x=y. 0.

0

2 Protein

4

roy

6

00 roll Biuret

8

10

2

14

Procedure

protein in the males was 6.4 mg./100 ml. (range 3-12), and the mean concentration in the females was 6.0 mg./100 ml. (range 4-10). The mean excretion of urinary protein in the males was 79 mg./day (range 48-150), and the mean excretion in the females was 76 mg./day (range 40-131). There was no significant difference between the urinary protein concentrations and excretions in the two sexes, as determined by Student’s t test. Frequency distribution graphs for urine protein in the combined group of 28 healthy adults are illustrated in Fig. 4. The urinary protein concentrations (milligrams per 100 ml.) are plotted in the graphs on the left side of the figure, and the urinary protein excretions (milligrams per day) are plotted on the right side of the figure. As shown in Fig. 4, the logarithmic distribution graphs adhered more closely to time gaussian conformation than did time arithmetic graphs. Therefore, the ±2 S.D. limits computed after logarithmic transformations are recommended for use as the normal ranges of values.

Vol. 14, No. 12, 1968

URINARY

1167

PROTEIN

Normal Values for Urinary Albumin

Fractionations of protein in urine collections from tile 28 healthy adults were performed after concentration by ultrafiltration and cellulose acetate electrophoresis, as described by Suimderman et al. (19). The mean percentage of albumin was 37.7 (S.l). ± 5.8) with a range of 27-50%. Albumin excretions in urine were computed by multiplying the total protein values (determined by the biuret Procedure) by tile percentages of albumin (measured by electrophoresis). Time mean excretion of albumin was 29.0 mg./day (S.D. ± 12.0) with a range of 18-52 mg./day. Interfering

Drugs and Pigments

In an endeavor to detect sources of interference, time biuret procedure was employed for measurements of protein in 24-ilr. collections of urine from patients on the medical, surgical, pediatric, and psychi14

2

Fig.

3. Comparison

protein

of

concentration

measurements

urine collections healthy adults

in

#{176}-

24-hr. from 28 by the

biuret procedure and by the sulfosalicylic acid method of Poortmans and van

Kerchove

0


..

Os

‘cb

I.-

U,

a’O C a

o:

Ut

4) 0.

4)

a’ E

a’ .0

.

CJ N-

-

a

C’)

-v

1.

En q.,

N-

.0

0i,

.2

Ic

00 V

-

a0

U,

0

.2

U,

.2

#{176}

.2

a

(54

‘0(5_i Ii’) +1

ci

Is

,oI

E 0 0

II

cs_i H

“U)

1’< (s_i

IIJ

o

0

no -‘

On 5.

5-,

5..

5.

-

S.-

a,

4) 0.

E

U)

a’