Determination of Neomycins B and C in Neomycin Sulfate A. A. BROOKS, ARLINGTON A. FORIST, and BARBARA F. LOEHR Research Laboratories, The Upjohn Co., Kalamazoo,
Mich.
Among the reactions which might serve to distinguish between the two molecules, the formation of furfural or substituted furfural ( 2 , 3) in sulfuric acid seemed the most promising. The neutral equivalent also seemed to be a reasonable method on which to base a set of simultaneous assays. Later it was found that the difference in the temperature coefficient of optical rotation could be used as an assay procedure. Three procedures based on the above observations are preseiited.
This procedure for the determination of neomycins B and C in neomycin sulfate involves measurement of the optical rotation and determination of furfural produced from the pentose portions of the molecules by acid degradation. This procedure is compared with two other methods: measurement of change of optical rotation with temperature and titration of free base combined with optical rotation. The furfural procedure is preferred. It has an accuracy on pure mixtures ~ neomjciri B, within about 1% on total neomycin, 3 q on and up to 10% on neomycin C (when small percentages of C are involved). The errors on finished neomycin are about 2% on the total and about 5q~ o n neomycin B.
EQUIPMENT
h standard model Beckman D U spectrophotometer equipped 7%-ithhydrogen lamp and 1-em. quartz cells was used for absorbance measurements. The optical rotations were determined on a Gaertner polarimeter using a 2-dm. tube, except when the temperature variation was measured in a 4-dm. Landolf-Clerget tube through which thermostated water was circulated. Titrations were carried out with a Beckman Model G p H meter with glass electrode, using semimicroequipment under a carbon diouide-free helium blanket.
T
H E development of a suitable chemical assay has been hampered by the lack of complete knowledge of the structure of the neomycins and by the presence of functionally similar impurities. Seomycins B and C have been shown to be hexaamino compounds consisting of neamine, a pentose, and a diaminohesose ( 1 , 2, 4, 6) and possessing t,he empirical formula C?3H4J6013. St,ructural differences appear t,o reside in the diaminohexose. Several procedures for the determination of total neomycin have been described. The colorimetric ninhydrin method of O'Keefe and Russo-Alesi ( 7 ) measures only total amino groups. Hamre and associates ( 5 ) determined total neomycins by an anthrone reaction, indicating equivalent responses by B and C. Dutcher, Hosansky, and Sherman (3) estimat,ed t'he furfural produced by acid degradation of the neomycins and found that R produced about twice as much furfural as did C. Two methods for the analysis of neomycin B-C mixtures have been published : the microbiological procedure of Sokolski and Carpenter (9) and the semiquantitative papergram method of Pan and Dutcher (8) for S-acetyl derivatives on neomycins B and C. Because the specific rotations of neomycins B and C are differcnt, a combination of a determination of optical rotation with a second measurement, in Lvhich the txyo compounds respond in a different ratio should permit an analysis of B-C mixtures.
REAGENTS
Neomycins B and C Sulfates. The materials used as standards are the same as described by Ford and coworkers ( 4 , l O ) and are believed to be the purest materials available. The specific optical rotation C CY]*^ in 0 . 1 s sulfuric acid) and the equivalent xeight of the free neomycin bases are defined as follou-s: neomycin B, +83.2", 102.95; neomycin C, +120.8", 102.95. A11 other chemicals are reagent grade. METHOD I.
Optical Rotation. The optical rotations are determined a t ea. 26' C. on solutions that contain about 5 grams of neomycin sulfate in 25 ml. of 0 . 1 S sulfuric acid. All rotations are taken a t the s a D line. The rotations are corrected for temperature variation h y the foiloffing formulas: Seomycin B
ff:,
I
Figure 1.
40
60
100 MINUTES
80
120
140
11 - 0.001il ( t - 2511 0'
a:,
ff25
ffh
f f 2D5
25"
< t < 80"
< t < 80"
C.
C.
+ 0.000112 ( t - 25'1'1 0" < t < 25'
(1) (2)
C.
13)
Furfural Reaction. The furfural reaction is run on a solution containing not more than 1 gram of base per liter. This solution is obtained by quantitative dilution with water of the solution used for optical rotation. Three milliliters of the diluted solution are placed in a glass-stoppered test tube and cooled in an ice bath for several minutes. T o this tube, still immersed in ice water, is added 2 ml. of concentrated (specific gravity 1 8 4 ) sulfuric acid. The solutions are mixed and cooled again in an ice hath. The stoppered tube is then placed in a boiling water bathwith the m-ater level above the liquid level in the tube-for 100 zk 2 minutes and then quenched in ice water. The absorbance of a 1 to 10 dilution of this solution with water is then measured at the furfural peak (-277) mw in a 1-cm. quartz cell against a 4'34 sulfuric acid blank which has been prepared from the concentrated acid in a manner identical to the sample.
4 -
20
ff25
Seomycin C
,500
0
OPTICAL ROT4TION-FURFURAL RESPONSE
160
I80
Development of furfural absorbance vs. time
The absorptivities of acid-degraded neomycins B and C are determined from known solutions of the standards using the same procedure. The calculations are made in terms of the diluted neomvcin solution.
1788
V O L U M E 2 8 , N O . 11, N O V E M B E R 1 9 5 6
1789
Calculations. Calculations based on the diluted neomycin solution are as follows: (4)
(5) \)-here B and C
= concentrations of neomycins B and C free bases,
grams per liter absorptivities (Beer's law constants) of aciddegraded pure neomycins B and C free bases a t -277 mp, liters per gram-centimeter a2gJand a': = optical rotations of pure neomycins B and C free bases t: 25' C., liter per gram-dm. ( 0 1 % ~ = 0.0832 ; aL5 = 0.1208") il = observed absorbance a t 277-mp furfural peak 01 = optical rotation for diluted neomycin solution calculated from observed rotation of concentrated solution b and c
=
V E T H O D 11.
VARIATION O F O P T I C A L ROTATION U I T H TEMPER 4TURE
Optical rotiition is measured on the solution described in Method I, using a 4-dm. jacketed tube maintained a t ca. 75' and 25" C. The concentration of neomycins is calculated from the following expressions:
H =
R l E T H O D 111.
"h
0.00171
-
all
[aB]'j
(t, - t2)
O P T I C A L R O T 4 T I O N - N E U T R A L EQUIVALENT
Optical rotation a t 25" C. is measured as in Method I. I n addition, a 5-ml. aliquot of the sample solution is mixed with an excess of saturated barium hydroxide solution and back-titrated with 0 2&Vsulfuric acid. Inflection points are determined by the method of equal increments and consist of a sharp end point a t approxim:itely p H 4.0 and a less sharp inflection a t about pH 10.5. Total neomjcin is calculated from the normality, while the % B is rend from a plot of ag1-Y LS. % B, n hich was shonn to be linear.
Table I.
assay applied t o a set of typival bulk neomycin preparations, are given in Table 11. Figure 1 presents the data required t o establish the optimum time for acid degradation. It was established from the data of Table I t h a t the optical rotations of neomycins B and C are additive. T h e ultraviolet absorption of the furfural a t the 277-mp peak obeys Beer's law t o a t least 1 gram of base per liter in the undiluted, acid-degraded solution.
.%sa>-of Neomycin B-C 3Iixtures by 3lethod I
ICnonn Concn,,
Grams BaseILiter So. B C Total 10 1 . 0 0 0 2a 1 000 3 0 : 7 7 5 0 150 0 : 9 6 5 4 0.613 0 374 0.987 5 0 530 0 521 1 . 0 5 1 60 1 . 0 0 0 ... 1 000 7a 8 0:900 0 100 1 000 5 0.800 0 200 1 . 0 0 0 10 0.700 0 300 1.000
:
.Ibsorbance zt,'Eev, (n) 543 =t3 . 5 265 f 4 . 0 484 i 2 . 0 446 f 1 0 $42 =t 2 7 042 f 0 7 267 =k 1 5 519 i 1 9 495 1 0 463 =t 1 7
*
Calcd.
Found Concn., Grams Base/Liter Dilution B C Total 0.0832 b ... 0.1208 0 . 0 8 7 9 0 806 0 : i 7 2 0 978 0 . 0 9 6 2 0 . 6 4 8 0 350 0 , 9 9 8 0.567 0.522 1,089
:
,..
0 0 0 0
...
...
1208 0870 0:9i3 o:OQi 1 :004 0907 0 , 8 2 2 0.185 1.007 0945 0.710 0 . 2 9 3 1.003
Reference samples. Optical rotations calculated from readings on concentrated solutions. Optical rotations calculated from known standards used t o prepare above mixtures. a b C
Table 11. Assay of Typical Bulk Yeomycin Sulfatesn (Gram of total neomycin bases per gram of sample and 70 neomycin B based on total neomycins) Method Method Method BioSample I I1 111 Average assayb 1 0.644 0.614 0 710 0.656 0.675 81.6 51.8 102.5 2 0.615 0.627 0.650 0.631 0.705 94.1 99.0 106 3 0.677 0.654 0 700 0.677 0 787 101 51.5 108 4c 0.786 0.772 0 781 0.779 0.768 53.9 89.9 92.0 5 0.651 0 665 0.683 0.680 0.714 96.3 88.6 97.0 a Neomycin sulfate used contains about 64% neomycin free base calculated from degree of neutralization. b K . pneumoniae turbidimetric assay dpscribed in ( 1 0 ) . C Neomycin hydrochloride.
The results on the k n o n n samples lie mithin the propagated experimental error and the scatter is reduced by use of a 100minute degradation time. T h e errors on total neomycin of about I %, on neomycin B of 3%, and on neomycin C of about 10% are satisfactory, considering the effect of a simultaneous assay on the propagation of errors. T h e results of the three methods as applied to typical finished neomycin preparations are in good agreement for total neomycin. T h e deviation from the average is the least for the optical rotation-furfural response method and is random in character. The deviations of the other two methods are not random. The agreement n i t h the biological assay is not so good as the agreement between the physical-chemical methods, but the difference is not unduly large. The values for the percentage of neomycin B show a considerable scatter among the three methods. It is difficult to determine the influence of the interfering impurities present in typical finished neomycin powders. I n most cases the impurities are unknov-n, although the method of preparation would lead one t o expect the presence of optically active or inactive weak bases. Both of these would contribute to high values of total neomycin and probably to high percentages of B by the optical rotationneutral equivalent method. I t is difficult to predict the effects of impurities upon the change of optical rotation with temperature. T h e presence of furfuralyielding impurities is less likely than the presence of optically active or inactive weak bases. The effect upon the optical rotation-furfural response method could be in either direction, although the effects tend to be self-compensating in the total neomycin value. On the basis of the present data the effects of impurities would appear to be less than about 2 % on total neomycin and less than about 5% on the percentage of neomycin B. T h e optical rotation-furfural response and optical rotation-neutral equivalent methods lend themselves t o large numbers of routine samples, whereas the measurement of optical rotation a t 75" C. is slow and awkward. The preferred method would appear t o be the optical rotation-furfural response procedure.
DISCUSSION
T h e results of the optical rotation-furfural response method applied t o several known samples are summarized in Table I. T h e results of the application of all three methods, as well as a bio-
ACKNOWLEDGMENT
The authors are indebted to the Department of Biological Control for the bioassays reported.
ANALYTICAL CHEMISTRY
1790 LITERATZTRE CITED
(1) Duteher, J. D., Donin, M. N., J. Am. Chem. SOC.74, 3420-2
(1952).
(2) Dutcher, J. D., Hosansky, N., Donin, M. N., Wintersteiner, 0..
Ibid., 73, 1384-5 (1951). (3) Duteher, J. D., Hosansky, N., Sherman, J. H., Antibiotics & Chemotherapy 3, 534-6 (1953). (4) Ford. J. H., Berm, & E., 'I Brooks, . A. A., Garrett. E. R., Alberti, J.. Dyer, J. R., Carter, H. E., J . Am. Cham. Soc. 77, 5311-14
(1955). (5) Hsmre, D. M., Pansy. F. E., Lepedes, D. N., Pedman, 0..
Bayan.A. P., Donoviek. R.. AnAbiotics & Chemotheram2, 13541 (1952). (6) Leach, B. E., Teeters, C. M., 3. Am. Chhem. Soc. 73, 2794-7 (1951). (7) O'Keefe, A. E.. Russo-Alesi. F. M., Division of Biological Chemistry, 116th Meeting, ACS, Atlantic City, N. J.. 1949. ( 8 ) Pan, S. C.. Dutcher, J. D., ANALCHEM. 28,836 (1956). (9) Sokolski, W. T.. Carpenter, 0. S., "Antibiotics Annual 19551956, P.383, Medical Encyelopedia. Ino., New York, 1956. (10) U. S. Pharmacopoeia, 15th Revision, p. 855, Mack Publishing Co., Easton, Pa.. 1956. R ~ c ~ ~ for v areview n ~ p r i 25, l 1956. Accepted jUly 20, 1956
CRYSTALLOGRAPHIC DATA
139. 2-Amino-2-methyl-l,3-propanediol HARRY A. ROSE end ANN VAN CAMP, Lilly Research Lsborstorior, Indianapolis 6, Ind.
HO-CHS
I I
CHa-C-NHs HO-CHs Structural formula for 2-amino-2-metYJ.-.,.-=~"~~"="~".
C
# =93O32' L a
2-h~1~o-Z-methyl-1,3-propsnediolcan be crystallized from acetone or from the lower alcohols. The crystals used in this study were obtained by allowing a warm isopropyl alcohol solution to cool. The resulting crystals are needles and rods elongated parallel to the c axis.
b-