V O L U M E 2 8 , N O . 1, J A N U A R Y 1 9 5 6 taken down to dryness on a medium hot plate and cooled. T o the platinum dish are added, by running down the sides, about 5 ml. of perchloric acid and 2 ml of hydrofluoric acid. The dish is swirled t o mix, and again taken down t o dryness and cooled. The residue should now yield to solution in dilute hydrochloric acid. ACKNOWLEDGMENT
The advice and encouragement of R. L. Wilson during the progress of this work are gratefully acknowledged, as well as the courtesy of R. S. Arrandale in permitting its publication. LITERATURE CITED (1) Baker, G.
L., and Johnson, L. H., ANAL. CHEM.26, 465 (1954).
39 Beckman Instruments, Inc., South Pasadena, Calif., Bull. 278, May 1952. Brabson, J. A,, and Wilhide. W. D., ANAL. CHEM.26, 1060 (1954). Close, P., Smith, W. E., and Watson, M. T., Jr., Ibid., 25, 1022 (1953). Close, P., and Watson, M. T., Jr., J . Am. Ceram. Sac. 37, 235 (1954). Gilbert, P. T., Jr., Hawes, R. C., and Beckman, A. 0.. ANAL. CHEM.22, 772 (1950). Mosher, R. E., Bird, E. J., and Boyle, A. J., Ibid., 5, 715 (1950). Scott, W. W., “Standard Methods of Chemical Analysis,” 5th ed., p. 275, Van Nostrand, New York, 1945. RECEIVED for review A4pril30, 1955. Accepted Ootober 17, 1955. Division of Analytical Chemistry, 127th Meeting, ACS, Cincinnati, Ohio, MarchApril 1955.
Colorimetric Determination of Cycloserine, a New Antibiotic L A W R E N C E R. J O N E S Commercial Solvents Corp., Terre Haute, Ind.
A specific method of analysis was needed for cycloserine, a new antibiotic, as an aid in chemical evaluation. Cycloserine reacts with sodium nitrotopentacyanoferroate in slightly acidic medium to give an intense blue-colored complex suitable for quantitative measurement at 625 mp. The color deviates slightly from Beer’s law but is reproducible in the range of 5 to 200 y of cycloserine. The test quantitatively determines the antibiotic in amounts as little as 2 to 3 y and has an accuracy within +2% and a precision within *l%. The method is specific for the cycloserine molecule and has been adapted to all phases of its production and use. The results are in good agreement with the bioassay.
C
YCLOSERINE is the generic name of an antibiotic isolated from a culture of Streptomyces orchidaceus by Harned, Hidy, and LaBaw (3). Preliminary clinical evaluation has demonstrated its effectiveness in pulmonary tuberculosis ( 2 ) and certain genitourinary infections (9). Cycloserine is a cyclic hydroxamic acid derivative of serine, determined to be D-4-amino-3-isoxazolidinone ( 1 , 4, 5 , 7 , 8), and has the formula:
”,
SH*
I
H-C-C=O
A-H
Hz-h
=
H-
A
-C-OH
H,-J
0 ‘’
$
0 ‘’
A specific method of analysis was needed for cycloserine in the presence of amino acids, other antibiotics, degradation products, and compounds of biological origin. A micromethod most suited the several requirements. Therefore, colorimetric methods were examined, because these are the most universally applicable. Cycloserine reacts with sodium nitritopentacyanoferroate in slightly acidic medium to give an intense blue-colored complex suitable for quantitative measurement a t 625 mp. The color develops rapidly and is stable for several hours. It deviates slightly from Beer’s law but is reproducible in the range of 5 to 200 y of cycloserine. The method is specific and sensitive for the cycloserine molecule and has been adapted to all phases of its production and use. The results are in good agreement with the bioassay.
CENTRIFUGE, any laboratory type for 15-ml. tube VACUUM OVEN,60” C. WEIGHINGBOTTLE,glass-stoppered DEsICC~TOR,with phosphorus pentoxide desiccant PIPETS, 1.0, 2.0, 3.0, 4.0, and 10.0 ml. Normax brand or equivalent BURET,50 ml., Normax brand or equivalent REAGENTS
CYCLOSERINE STANDARD. Repeated biological assay, elemental analyses, optical rotation, titration, and other tests indicate that the purity approaches 100%. SODIUMHYDROXIDE, 4.000 and 0.100N solutions ACETIC ACID,3.000 and 1.000N solutions SULFURIC ACID,0.666N solution Na2W04, 10 j = 0.5% aqueous solution SODIUMTUNGSTATE, TUNGSTIC ACID REAGENT. Mix equal volumes of the sodium tungstate and sulfuric acid solutions. This reagent must be made fresh daily. SODIUMNITROPRUSSIDE, N a 2 F e ( C N ) a 0 .2H20. Prepare a fresh 4 f 0.5% aqueous solution every 2 weeks and store in a glass-stoppered brown bottle. CYCLOSERINE COLORREAGENT, Nac[Fe(CN)bNOz]. Mix equal volumes of the sodium nitroprusside and 4 V sodium hydroxide solutions just prior t o use. The reagent is ready for immediate use and must be discarded after one set of determinations because of instability. ACTWATEDCHARCOAL, Darco G-60 or equivalent, PREPARATION OF CALIBRATION CURVE
Add 0.10 to 0.12 grams of the crystalline cycloserine to a tared glass-stoppered weighing bottle and place in’ the 60’ C. oven at 10 to 15 mm. of mercury pressure for 2 hours. Cool in the desiccator. Prepare a solution of cycloserine standard in 0.100N sodium hydroxide solution t o contain 1.0 mg. per ml. (This solution is stable for several weeks if kept refrigerated.) Transfer by means of a buret, 0.0-, 2.5-, 5.0-, 7.5-, 10.0-, 12.5-, 15.0- 17.5-, and 20.0ml. portions of the standard solution to separate 100-ml. volumetric flasks. Dilute each to volume with 0.100N sodium hydroxide. Transfer 1.0 ml. of each dilution into a test tube. Add 3.0 ml. of 1.OOOX acetic acid and 1.0 ml. of color reagent and m x . Allow the solution to stand a t room temperature for 10 minutes. Transfer to a 1-em. Corex cell and read the absorbance a t 625 mp, using the solution containing 0.0 ml. of standard as the blank. Plot concentration against absorbance on linear graph paper. The curve deviates slightly from a straight line. The above standards equal 0, 25, 50, 75, 100, 125) 150, 175 and 200 y of cycloserine, respectively. DETERMINATION
APPARATUS
SPECTROPHOTOMETER, Beckman Model DU, with 1-cm. Corex cells
The determination is the same as the calibration, except the sample is prepared so that a 1.0-ml. aliquot of 0.100N sodium hydroxide solution does not contain more than 1507 of cyclo-
ANALYTICAL CHEMISTRY
40 serine per milliliter. If the color is too dark t o read, the analysis must be repeated, using a smaller aliquot. The colored complex cannot be diluted. APPLlCATlOSS
Table I.
Determination of Crystalline Cycloserine (microgram per milligram)
B1 o 1o g i c a 1
Lot s o .
This method has been applied successfullj- t o the determination of crystalline cycloserine and ita salts in pharmaceutical preparations, fermentation, and process samples, and in blood, urine, cerebrospinal fluid, and other biological fluids.
Theory Color Assay 988.3 1000 1000 990 988 , 1000 1000 0 1000 a Plate a-say method using .llzcrococcus p u o g e n e s hfi538P.
Assaya
1031
2
1103
T-aI.
1033 1005 aureus .1TCC
DETERhIINATION O F CRYSTALLISE CYCLOSERINE
Weigh accurately 0.10 t o 0.12 gram of dry, crystalline cj-closerine, transfer quantitatively t o a 100-ml. volumetric flask, and dilute to volume with 0.100~Vsodium hydroxide. Transfer 10 ml. of this solution t o a second 100-ml. volumetric flask, and dilute t o volume with 0.100.\- sodium hydroxide. A4nalyzea 1.0-ml. aliquot of this solution for cycloserine, using for a blank 1.0 ml. of 0.10OS sodium hydroside solution treated in the same manner as the sample. Typical results on the determination of crystalline cycloserine by the color method as compared to a biological assay are given in Table 1. CYCLOSERINE IN BLOOD PLASJIA
Pipet 2.0 ml. of plasma t o a 15-ml. centrifuge tube, add 4.0 ml. of the tungstic acid reagent, mix well, and centrifuge for several minutes until a clear filtrate is obtained. Pipet 3.0 ml. of the clear filtrate (this is equivalent t o 1.0 ml. of plasma) t o a test tube. Add 1.0 ml. of 3.000S acetic acid, and 1.0 ml. of the freshly prepared color reagent and mix. Allow the reaction to st,and at room temperature for 10 minutes. Transfer t o a 1-em. Corex cell and read the absorbance a t 625 mp, using a blank solution prepared from 1.0 ml. of 0.1005 sodium hydroxide and reagents t o set the spectrophotometer. Data for the determination of cycloserine in blood plasma are: Found, Y/hIl. 10 49 99 201
10
50 100 200
CYCLOSERINE IN CEREBROSPINAL FLUID
Pipet 1.0 t o 3.0 ml. of clear cerebrospinal fluid t o a test tube.
If necessary, add water t o make 3.0 ml. Add 1.0 ml. 0.f 3.0005 acetic acid and 1.0 ml. of color reagent and proceed as in above procedures, using a blank prepared from 1.0 ml. of 0.100N sodium hydroxide and reagents t o set the spectrophotometer. Data for the determination of cycloserine in cerebrospinal fluid are: Found, y/Ml. 5.0 9.9 50.1
y/m.
5 10 50 100
CYCLOSERINE IX URINE
Data for the determination of cycloserine in urine are given below. The low recoveries of known amounts of cycloserine from urine indicate the presence of some interference. .y/hll.
Y/Xll.
100 100
94.0 97.0 93.0
100
100
Added, 100
100
Vol. 4.00.\Acetic .\cid, 111.
Absorbance IOOT of Cycloserine
mentation is high and the cycloserine titre is low, the broth should be decolorized n i t h char before analysis. Analyze a 1 to 3-ml. aliquot, using 1.0 ml. of 0.lOOS sodium hydroxide and the reagents as a blank. EXPERIMENTAL
The effects of several variables were investigated for their influence on the color format'ion and quant'itative application of the reaction. The factors included the absorbtion curve, stability and intensity of color, time and temperahre of reaction, stability and concentration of sodium nitrit,opentacyanoferroate, acid-base concentration, conformity to Beer's law, and interference from other compounds. A Beckman Model DU spectrophotomet,er was used. Khen sodium nitroprusside is treated at' room t'emperat.ure with a strong aqueous solution of sodium hydroxide, an iron(I1) compound is formed in which the nitroso group is oxidized by iron(II1) to the nitro group ( 6 ) , and sodium nitritopentacyanoferroate, Sa4[Fe"(CN)j(N02)],is formed. This reagent is easily osidized by light and oxygen, and the oxidized reagent does not give the charact'erist'ic color with ctcloserine. Fresh reagent must be prepared for each set of determinations. For convenience, two stock solutions of 4 . 0 5 sodium hydroxide and 4% aqueous sodium nitroprusside are prepared and equal portions are mised for each anal>-sisof cycloserine. The reagent must be used within 15 minutes after preparation. ABSORPTION SPECTRUM
Dilute the urine so that a milliliter aliquot contains between 50 and 150 y of cycloserine per milliliter and then analyze. I n cases where pigmentation is high and cycloserine content is low the urine should be neutralized and decolorized with char before analysis. A4nalyzea 1 t o 3-ml. aliquot, using 1.0 ml. of 0.1005 sodium hydroxide and the reagents as a blank.
Found,
hbsorbanre 100 y of Cycloserine
The absorbance curve of the blue comples of cycloserine and sodium nit,ritopentacyanoferroate shows a masimum absorbance a t 625 nip.
100
Added, -f/XIl.
111.
REAGENT PREPARATION
Normal cerebrospinal fluid can be analyzed for cycloserine without preliminary treatment.
Added,
Table 11. Acid-Rase Effect on Color Vo1. 4.00.\llcetic Acid,
Found y/hll. 96.0 87.5 97.0
CYCLOSERINE IN FERRIE\TATION MEDIA
Fermentation media and process recovery samples can be analyzed for cycloserine after filtration and dilution. If pig-
TRANSBIITTANCE AND CONCENTRATION
Calibration curves were determined for the cycloserinesodium nitritopentacyanoferroate by plott,ing absorbance against concentratmionon linear graph paper. The color deviates slightly from Beer's law in the range of 10 to 200 -1 of cycloserine, The curve is reproducible a t all point,s when a standardized procedure is followed. COLOR STABILITY
Cycloserine was treated to develop the colored complex. The absorbances of the samples were determined a t various time intervals. The color was found to he stable for a t least 4 hours. TIME AND TEMPERATURE OF REACTION
The color complex was prepared and spectrophotometer readings were made intermittently in order to determine the required
V O L U M E 2 8 , N O . 1, J A N U A R Y 1 9 5 6 amount of time for complete color development. -4development time of 10 minutes was chosen as optimum to ensure complete color intensity. The complex between cycloserine and sodium nitritopentacyanoferroate was found to be sensitive toward temperature changes. Although the color develops rapidly a t room temperature, the highest precision and reproducibility were obtained in a temperature-controlled laboratory where the temperature was 25' & 1' C. and the relative humidity \vas 50 f 2y0. Cooling samples below 15OC.before color development resulted in slow color formation, and heating samples above 50OC. completely destroyed the color.
41 acid, methionine, glucurono-lactone, rutin, urea, benzylpenicillin, streptomycin, dihydrostreptomycin, bacitracin, Aureomycin, Terramycin, neomycin, glucose, lactose, and sucrose. These compounds did not give a blue color. -4 positive reaction was given by derivatives of cycloserine which still retain the basic ring structure of cycloserine. The derivatives prepared in the laboratory and tested were: monoacetyl, monobenzoyl, mono-and diisocyanate, and desaminocycloserine. The test appears to be specific for the determination of the csycloserine molecule in a wide variety of samples. SUMMARY
REAGEF-T COIVCENTRATION
The concentration of reagent is not critical, as long as an excess amount is added. 4 2% solution of sodium nitritopentacyanoferroate in 2.000N sodium hydroxide gives a low reagent blank and maximum color. ACID-BASE RELATION TO COLOR
The complex develops in weakly acidic medium. Acetic or phosphoric acids can be used, but stronger acids such as sulfuric, hydrochloric, or nitric destroy the color complex. The optimum amount of acid required was investigated. Several aliquots containing 100 y of cycloserine were treated with 1.0 ml. of color reagent in 2.ON alkali. The amount of 4.0S acetic and the water content were varied. The total volume was held constant a t 5.0 ml. The results are given in Table 11. Three milliliters of a 1.00OS solution of acetic acid was chosen as optimum when 1.0 ml. of color reagent equivalent to 2.000.Y alkali was used. INTERFERENCE
In order to determine the specificity of the nitritopentacyanoferroate method toward cycloserine several compounds !$ere tested. These included degradation compounds, amino acids and other antibiotics. The compounds tested were: u-serine, serine amide, serine hydrosamic acid, hydroxylamine hydrochloride, D-threonine, L-tyrosine, D-valine, L-proline, L-hydroxyproline, D-lysine, u-glutamic acid, glycine, p-alanine, s-meth? 1glucamine, L-hiqtidine, L-cysteine, asparagine, nicotinamide, uric.
A colorimetric procedure for the specific determination of cycloserine, a new antibiotic, has been presented. The proposed test quantitatively determines the antibiotic in amounts as little as 2 to 3 y and has an accuracy within f 2% with a precision ivithin & 1%. The method has been adapted to the determination of crystalline cycloserine and its salts in pharmaceutical preparations, blood plasma, urine, cerebrospinal fluid, other biological fluids, fermentation broths, and all phases of proceseing. Interference has been encountered only in urine. A close correlation has been found between a biological assay and the color apsav. LITERATURE CITED
Cuckler, A. C., Frost, 13. XI., McClelland. L., and Solotororsky, AI., Antibiotics &- Chernofherapy5 , 191 (1955). Epstein, I. G., Sair, K. G . S.,and Boyd, L. J., Antibiotic M e d . 1, 80 (1955).
Harned, R. L., Hidy, P. H., and LaBaw, E. K., A~itibiotic & Chemotherapy 5, 204 (1955). Harris, D. A . , Ruger, AI., and others, Ibid.,5, 183 (1955). Hidy. P. H., Hodge, E. B., and others, J . Am. Chem. SOC.77, 2345 (1955).
Hofman, K. A . H., A m . Chem., Justus Liebigs 312, 1 (1900); Chem. Zentr.. 557-9 1900 I I .
Kuehl, F. d.,Jr., Wolf. F. J . , and others, J . A m . Chem. SOC. 77, 2344 (1955).
Stammer. C. H., Filson, .I,X., Holly, F. W..and Folkers, K., Ibid..77. 2346 (1955). ~, Welch, H.,'Putrnan, L. E., and Rantiall, IV. A , , Aiifibiotic M e d . 1 , i 2 (1955). RECEIVED for review July 18, 1955.
Accepted September 18 19%.
Color Reaction of Salicylic Acid and Nitrite with Cupric Ion A . L. UNDERWOOD D e p a r t m e n t of Chemistry, Emory University, Emory University, G a .
The variables which affect the determination of copper h>-the Jorissen reaction (the formation of a red color upon heating a solution containing salicylate, nitrite, and cupric ion) have been investigated. The pH of the solution, the concentration of nitrite and of salicylate, and the duration of the heating period must be controlled if this reaction is to yield reliable results. Under the conditions suggested as a result of this study, the recommended copper concentration range is about 2 to 9 p.p.m. A t the optimal concentration (about 5 p.p.m.) the standard deviation is about 0.04 p.p.m. Although a detailed study of interferences has not been made, the method appears to be fairly selective for copper. Information regarding the identity of the red reaction product is also presented.
T
HE red color fornied upon heating :t solution containing salicylic acid, inorganic nitrite, and cupric ion has been known for many years. I t \vas first reported by Jorissen ( 2 ) and later st'udied by Schott ( 4 ) and Sherman and Gross ( 5 ) . More recently, this color reaction has been mentioned by 1Iuller and Burtsell ( 3 ) and has fourid its way into several standard treatises ( 1 , 6, 7') where it, i? suggested as the basis of a colorimetric copper determination and also as a confirmatory test for salicylates. In spite of this attention, the reaction has never been studied criticslly to establish optimal conditions for analytical procedures, nor has the nature of the red reaction product h r e n elucidated. In the present investigation, it has tieen found that the variables affecting the development, of the red color must be controlled much more carefully than previous reports have in-