Carbazole Estimation of Hexoses in’ Connective Tissue Extracts and Their Hydrolyzates R. E. GLEGG’ Department
of Anatomy, McGill University, M o n t r e a l , Canada
A modification of the carbazolcsulfuric acid meLhotl was used to show that fucose, galactose, glucose, ar d mannose are not significantly decomposed by heatirsg with the hydrogen form of a cation cxchange resin for. 96 hours. Application to the analysis of conncctiie tissue extracts which contained fucose, galactose, g l u cose, and mannose, and to their hydrolyzates, showed that hexoses could be accurately determined without knowledge of the exact ratios in which they occurrcd.
for 10 minutes. (If necessary, they could then be allowed to stand a t room temperature for 3 hours without change in the spectrophotonictric readings.) Absorbances were measured on these solutions using a Beckmari DU spectrophotometer. The instrument was adjusted to zero with the over-all reference blank and the absorbances of the other three solutions were measured against this blank. In order to obtain the corrected absorbance, the sum of the readings for the sugar (or unknown) blank and the carbazole blank was subtracted from the reading for thc standard sugar (or unknown) solution
-
(11).
Plots were made of absorbance against wave length over the range 400 to 580 mp to obtain absorption curves for the simple sugars glucose, galactose, mannose, glucuronic acid, and fucose, as me11 as mixtures of galactose-mannose (1 : l),galactose-glucosemmnose (1:1: I), and galactose-mannose-fucose (1: 1: 1). These monosaccharides and mixtures showed maximum absorption a t 518 mp, except that fucose absorbed maximally a t 528 mp and glucuronic acid a t 540 m9.
T
HE carbazole-sulfuric acid technique for the estimation of hexoses (1, 2, 7, 8) was modified by the introduction of , I ncw system of blanks (11) and used to determine the polysaccharides of serum (11) and influenza virus (IO). Holzman, hIacAllister, and Niemann (9) subsequently made a systematics study of the variables influencing the determination and estab. lished optimum conditions for the reaction by adjusting the con centrations of sulfuric acid and carbazole, the period of heating and the volumes of solutions to be used. I n the present investigation a modified procedure was developed on the basie of previous findings (9, 1 1 ) and used to estimate the carbohydrate content of connective tissue extracts (4, 6), the degree of decomposition caused by heating monosaccharides with a cation exchange resin ( 3 ), and carbohydrate-protein complexes with the resin, hydrochloric acid, and sulfuric acid.
3.
30.25:
EXPERIRiEh’TAL
Carbazole-Sulfuric Acid Technique. REAGENTS.Reagent grade carbazole was purified by dissolving i t in 84% sulfuric acid, precipitating with water, and recrystallizing the product from alcohol and then from toluene. The stock solution made up to contain 0.5% (w./v.) of the purified carbazole in absolute alcohol should remain colorless when mixed with 84% sulfuric acid. (A comparable product was obtained after four sublimations.) Sulfuric acid was of reagent grade, diluted to 84% by weight (30.3N). The eoncentration WIS checked by titration with sodium hydroxide. SUGARS.With the exception of glucose, which was a standard sample obtained from the National. Bureau of Standards, the sugars were commercial samples which had been recrystallized. Their purity was checked by polarimetric measurements. hIoDIFmD PROCEDURE. TITOreagents were prepared. For Reagent A 8.35 ml. of absolute alcohol was made up to 250 ml. with 84% sulfuric acid. For Reagent B 8.35 ml. of the 0.5% solution of carbazole in absolute alcohol was made up to 250 ml. with 84% sulfuric acid. Reagents A and B were stored at 4” C . for as long as 1 week, and allowed to warm up to room temperature just before being used. For any given determination, three blank solutions were used: the f‘over-sll reference blank” containing 9 ml. of Reagent A and 1 nil. of water; the “sugar (or unknown) blank” consisting of 0 ml. of reagent A and 1 ml. of the solution under investigation; and the “carbazole blank” containing 9 ml. of Reagent B and 1 ml. of water. The final solution for the measurement of color development between carbazole-sulfuric acid and carbohydrate was made up from 9 ml. of Reagent B and 1 ml. of the sugar (or unknown) solution. The 9-ml. aliquots of Reagents A and B were placed in &inch borosilicate glass test tubes, and 1-ml. portions of water and sugar (or unknown) solution were carefully layered above them. T h e contents of the tubes were then thoroughly mixed by shaking and the mixture was heated in a boiling water bath for 13 minutes. The tubes were removed and cooled in an ice-water mixture Present address, Department of Food Science a n d Technology, N. T.State Agricultural Experiment Station, Cornell University, Geneva, N. T. I
OV
dl
.02
.03
.04
CONCENTRATION
.05
.06
: TOTAL
.07
08
.09
.,b
Mg SUGAR/MI.
Figure 1. Relation of corrected absorbaiiccs (carbazole) to concentrations of galactose-mannose and galactose-glucose-mannose mixtures
Absorbances wcre measured a t 548 mp for galactosomannose (I: 1) over the concentration range of 0.02 to 0.1 mg. per ml. and for galactose-glucose-mannose (1: 1: 1) over the range of 0.03 to 0.075 mg. per ml. to obtain standard curves (Figure 1 ) . Recovery of Monosaccharides after Heating with Resin. RESIN. The resin, Permutit Q, a polystyrene sulfonic acid type of cation exchange resin, was acid-regenerated by shaking with 4.4N hydrochloric acid (900 ml. per liter of resin) in a separatory funnel, and then washed with large volumes of distilled mater until the washings mere free of chloride ions. The resin was refluxed for 7 days with several changes of distilled water until the washings, which were a t first dark brown became almost colorless, and mas then air-dried. The m a t e r h extracted from the resin with hot water absorbed strongly a t 548 mp and without this treatment there mould have been a high blank in the following work, which involved heating solutions of monosaccharides and other substances with the resin prior to quantitative determination by the carbazole-sulfuric acid technique. To determine whether the boiled resin mas still effective as a catalyst for hydrolysis, several glycoproteins were hydrolyzed under the usual conditions (4-6) and the hydrolyzates were
53.2
533
V O L U M E 28, NO, 4, A P R I L 1 9 5 6 Table I. Corrected Absorbances (Carbazole Technique) of Sugar Solutions before and after Treatment with Permutit Q at 100" C.
Substances Fucosc Glucose Cellobiose Galactose-mannose (1 : 1) Glucuronic acid
Sugar Solution Concn., hfg./hfl. 0.20 0.04 0.04 0.04 0.04
Corrected Absorbances Resin Treated Un48 96 treated hours hours 0.147 0.148 0,150 0.196 0.197 0.199 0.186 0,193 0.190 0.136 0.139 0.141 0 045 0.018 0.127
analyzed by paper chromatography. The results were the same as those obtained previously (4-6) , the monosaccharides being detected after spraying with aniline hydrogen oxalate. I n order to measure the decomposition caused by heating nionosaccharides with resin, solutions were made up containing 0.04 nig. per ml. of glucose, glucuronic acid, galactose-mannose (1: l ) , and cellobiose, and 0.2 mg. per ml. of fucose. Five-milliliter aliquots of these solutions were heated with 0.4 gram of resin in seded glass tubes a t 100" C. for 48 and 96 hours. I n order to correct for the small amount of colored material eluted from the resin under these conditions, 5 ml. of water and 0.4 gram of resin were also heated under the same conditions. The tubes were tied to a metal plate, which was rotated slowly by means of a motor and pulley system situated outside the oven, so that the contents of the tubes were continuously mixed throughout the period of heating. The treated solutions were filtered when cool arid aliquots verc taken from the filtrates for analysis by the carbazole-sulfuric acid method. The solutions obtained after heating with resin were slightly yellow. Ab,sorbances were measured a t 548 mp for the untreated and resin-treated solutions and the solutions obtained from resin and water. The readings for the resintreated sugar solutions were corrected for the "resin blank" (Table I). Quantitative Determination of Carbohydrate in Connective Tissue Extracts. The connective tissue extracts (fr'aetions 11) had been prepared in this laboratory from Achilles tendon, lung, skin, cartilage, and bone of cattle (4, 6). The tissue mas extracted with 0.5.47 sodium hydroxide in the cold for 4 days.
The extract was neutralized, concentrated, and fractionated with alcohol. Acid mucopolysaccharides (fraction I ) were precipitated by addition of alcohol to a concentration of 63%) and fraction I1 was obtained after the alcohol concentration had been increased to 84%. The main problem was t o determine whether the occurrence of protein in these samples (fractions 11) would interfere with the carbazole determination. The problem was examined in three ways. First, their carbazole absorption curves (400 to GOO mp) were compared with a similar curve for a solution containing 0.04 mg. each of galactose, mannose, and fucose per ml. of solution, to see whether the absorption maxima as well as the shapes of the curves were the same (Figure 2). Secondly, 2 ml. of a standard solution of galactose-mannose (1: 1) containing 0.025 mg. of each per ml. was added to 2 ml. of thc solutions of connective tissue extracts used above, and then tho absorbances were measured a t 548 mp (Table 11) to estimate recovery of the added sugar. Thirdly, because the monosaccharide residues in these extracts must be bound to one another or to amino acid residues, the carbazole determinations were also carried out on their hydrolyzates. The absorbances a t 548 nip for solutions of galactose-mannose in the ratios 2 to 1 and 1 t o 2, a t total sugar concentrations of 0.03 mg. per ml., fell on the straight line shown in Figure 1. Thus within the range of concentrations studied the total hexose may be determined in a mixture without exact knoivledge of the ratios of the different hexoses present.
Table 11. Corrected Absorbances (Carbazole) of SoluLioiis of Connective Tissue Extracts, Standard GalactoseMannose, and a Mixture off Both Tissue Extract Concn., M g . Dry Weight/AIl. Tendon Skin Cartilage Lung Bone
1.61'2
0.170 0.194 0.231
2.060
0.403 1.014 2.414
0,400
Figure 2.
489
520 560 609 WAVE LENGTH (mp)
Carbazole-sulfuric acid absorption curves
0.04 m g . per ml. of galactose, mannose; a n d fucose. 2.41, 1.01, 0.40, 2.06, and 1.61 mg. dry w e i g h t per ml. of extracts from lone, l u n g , cartilage, skin, and t e n d o n , respectively
Corrected. Absorbances ~ _ _ Gal.-man. Calcd. for Exptl. f o r soln. mixture mixture 0.101 0.103 0.180 0.172 0.180 0.175 0.180 0.157 0.179 0.160 0,207 0.204 0.180 0.290 0.293
Table 111. Corrected Absorbances (carbazole) of Solutions of Connective Tissue Extracts before and after Hydrolysis
Tissue Extract Concn., JIg. Dry TVeight/hII. 0.4 Cartilage Bone 1.5
0 0 1 " " ' 400 440
Tissue extract 0.142
Corrected Absorbances 0.5.V Resin HC1 UnhydroITydrolyzed, hydrolyzed Hours lyzed, solution 48 90 4 hr. 0 . 1 8 1 0.177 0.176 0.182 0.208 0.271 0 . 2 6 5 0.255
0.5N HzSOI hydrolyzcd, 4 hr. 0.179 0.239
Carbazole Estimations on Hydrolyzed Connective Tissue Extracts. Five-milliliter volumes of solution8 of the extracts from cartilage and bone containing 0.4 and 1.5 mg. per ml., respectively, were heated with 0.4 gram of resin a t 100' C. for 48 and 96 hours. Controls were also set up with water and resin alone to correct for the small amount of material leeched out of the resin. For comparison, 8 and 30 mg. of the extracts from cartilage and bone, respectively, were heated with 10 ml. of 0.5N hydrochloric acid for 4 hours a t 100' C.; similar samples mere heated under the same conditions with 10 ml. of 0.514' sulfuric acid. These conditions using mineral acid are an average of those used by several workers for hydrolyzing similar material. A11 these hydrolyses were carried out in sealed glass tubes, which were continuously I otated in the oven as previously described. The tubes were cooled, then the solutions filtered. I n the case of resin hydrolyzates, aliquots were taken directly from the filtrates for carbazole analysis. However, in the case of hydrolyzates from hydrochloric acid and sulfuric acid, 5-ml. nliquots of the filtrates were neutralized with 0.6N sodium hydroxide, then made up to 10 ml. with water. The results are summarized in Table 111. RESULTS AND DISCUSSION
Effect of Heating Sugars with Hydrogen Form of Cation Exchange Resin. The absorbances (Table I ) for solutions of fucose,
ANALYTICAL CHEMISTRY
534 Table IV. Absorbances a n d Hexose Contcnt of Connectivc Tissue Extracts
Tendon
Tissue Extract Conon., hlg. Dry Weight/Ml. 1.812
Carbarole blank 0.023
Absorbances Unknown Sugar blank readlng
-
zero 0.002 0.005
Skin
2.060
0.028 0.028
Cartilage
0.403
0.023
0.002
Lung
1.014
0.023 0.028
0.003 0.012 0.013
Bone
2.414
0.028
0.027
0.028 0.028
0.024
0.006
0,028
0.167
0.170 0.203 0.203 0.218 0.225
0.2G7 0.277 0.454 0.452
Corrected Absorbance 0.144 0.140 0.170 0.189 0.193 0.194 0.232 0.236 0.399 0.400
%
Hexose 2.4 2.2 13.0 6.3 4.5
glucose, cellobiose, and a mixture of galactose and mannose (1 :1: agree well before and after treatment with the resin for 48 or even 96 hours. Thus it is concluded that these monosaccharides are not significantly decomposed by the treatment with resin. On the other hand, glucuronic acid must have been almost completely destroyed. When the solution of glucuronic acid was heated without resin under similar conditions, the same absorbances were obtained as after resin treatment. Thus the effect on glucuronic acid is due to the heating and not to the presence of the resin. These findings are in agreement with those previously ieported on a qualitative basis from results obtained by visual estimation of the intensities of spots on paper chromatograms (S), except t h a t fucose is now shown to be stable under the conditions of hydrolysis. Determination of Carbohydrate in Connective Tissue Extracts. I t was kncnm from previous paper chromatographic analysis that these connective tissue extracts all contained galactose, mannose, and fucose (6). The absorption curves plotted for these glycoproteins were therefore compared with the curve for a mixture of galactose, mannose, and fucose (Figure 2). These curves have the same shape and the same absorption maxima a t 548 mp. This result is taken as partial evidence that the noncarbohydrate components of these substances do not interfere significantly with the carbazole determination. In order t o measure the recovery of sugar, a known amount of galactose-mannose (1:l) was added t o the solutions of connective tissue extracts and carbazole determinations were carried out. The experimental absorbances of the mixtures were compared with the calculated values, which mere the arithmetical means of the absorbances found for the extract {Table 11, column 3) and the pure galactose-mannose solution (Table 11, column 4). The excellent agreement between the calculated and experimental values indicates that the carbazole readings for these substances are not affected by the other constituents present in them or released during the procedure. While the above determinations mere done on untreated connective tissue extracts, further carbazole determinations mere also carried out on their hydrolyzates, to determine whether the same results would be obtained when the monosaccharide units were available in the free form instead of being bound in some unknown manner in the original material. Table I11 shoms excellent agreement between the absorbances of the unhydrolyzed solutions and those hydrolyzed with the resin €or 48 or even 96 hours. In the case of hydrochloric and sulfuric acid hydrolyzates, there is also good agreement with the unhydrolyzed solution of cartilage extract, but with both acids the values are lower for the bone extract. The loss of sugar may depend on the material being hydrolyzed with mineral acid and may be due to interaction between sugar and amino acids after their liberation. Such interaction would probably be reduced in the case of the resin, as a large part of the amino acids is held by the resin.
The amount of carbohydrate in these samples was calculated from the equation D = K C L. The extinction coefficient, K , was calculated from the values for D (corrected absorbance) and the corresponding values of C (the total sugar concentration in milligrams per milliliter) as shown in Figure 1 for galactosemannose (I : 1) and galactose-glucose-mannose (1: 1:1). L (the light path in centimeters) was unity. These standards were chosen because the connective tissue extracts had been shown by paper chromatographic analysis of their hydrolyzates to contain these sugars in different ratios (6); the values for these two standards fell on the same straight line when plotted. The mean value of K was 3.66, Duplicate determinations of absorbances for the “carbazole blank,” “unknown blank,” and the final readings on the solutions of connective tissue extracts are given in Table IV, together with the calculated “hexose” percentages. These hexose values are in error, because all the substances analyzed contain fucose, which also produces with carbazole a colored substance absorbing maximally a t 528 mp. However, a t 548 mp the absorption due to fucose (corrected absorbance of 0.044 for a solution containing 0.05 mg. per ml.) is small compared with that for a solution of galactose-mannose of the same concentration (corrected absorbance of 0.180). For this reason the absorption maximum for a mixture of galactose-mannose-fucose (1 :1:1) is the same as for a mixture of galactose-mannose (1 :l), and occurs a t 548 mp. If the fucose contents of these connective tissue extracts were determined, a correction could be made. Nevertheless, the intensities of spots observed on paper chromatograms (6) shouthat the ratio of hexoses to fucose in these materials is large, and thus the hexose values are probably close to the actual values. The modified carbazole-sulfuric acid procedure developed in this work makes use of the system of blanks developed by Seibert and Atno ( I I ) , but is simpler and faster, as only two solutions (instead of three) are added to the tubes. The procedure also incorporates the requirements for optical color development and minimal experimental error established subsequently by Holzman, RlacAllister, and Niemann (9). As the monosaccharide components of the connective tissue extracts had previously been identified (e), it mas not necessary, as in the case of previous i\-orBers (7, 8, IO), to use the method of E ratios as a means of qualitative identification; in this way, more suitable standards were set up for the particular substances under study. The correspondence between the shapes of the absorption curves for the extracts and the standards, the excellent recovery of added sugar, and the fact that the same carbazole readings were obtained for galactose, glucose, mannose, fucose, and connective tissue extracts before and after hydrolysis with the resin afford good evidence that the carbazole method is capable of giving an accurate value for the amount of bound hexose present in these particular substances. ACKNOWLEDGIIIENT
The work was supported by a grant from Eli Lilly 6: Co. t o C. P. Leblond. Excellent technical assistance by Imre Sarlos is acknowledged. LITERATURE CITED
(1) Dische, Z., Bioehem. 2. 189, 177 (1927). (2) Dische, Z., lllikrochemie 8, 4 (1930). (3) Glegg, R . E., Eidinger, D . , AXIL. CHEM.26, 1365 (1954). (4) Glegg, R. E., Eidinger, D . , A r c h . Biochem. 55, 19 (1955). (6) Glegg, R. E., Eidinger, D., Leblond, C. P., S c i e m e 118, 614 (1953). ( G ) I b i d . , 120, 839 (1954). (7) Gurin, S., Hood, D. B., J. Biol. Chem. 131, 211 (1039). (8) I b i d . , 139, 775 (1941). (9) Holzman, G., MacXllister, R., Niemann, C., Ibid., 171, 27 (1947). (10) Knight, C., J. E x p t l . N e d . 85, 99 (1947). (11) Seibert, F. B . , Atno, J., J . Diol. Chem. 163, 511 (1946).
RECEIVED for review June 24, 1955. Accepted October 31, 1955.
