J. Agric. Food Chem.

1983, 31 I 545-548 545

Damodaran, S.; Kinsella, J. E. J . Agric. Food Chem. 1981b, 29, D., Ed.; Interstate Printers and Publishers, Inc.: Danville, IL,
1249. 1976; p 904.
Ehler, K. F.; Bernhard, R. A.; Nickerson, T. A. J . Agric. Food Milner, M.; Scrimshaw,N. S.; Wang, D. I. C. “Protein Resources
Chem. 1979,27,921. and Technology: Status and Research Needs”;Avi Publishing
Eldridge, A. C. “Soybeans: Chemistry and Technology”;Smith, Co., Inc.: Westport, CT, 1978; pp 288, 290.
A. K.; Circle, S. G., Eds.; Avi Publishing Co.: Westport, CT, Morowitz, H. J. “Entropy for Biologists”;Academic Press: New
1978. York, 1971.
Franzen, K. L.; Kmella, J. E. J. AgTic. Food Chem. 1974,22,675. National Agricultural Research Policy Advisory Committee
Gale, R. L.; Beebe, R. A. J. Phys. Chem. 1964,68, 555. “National Soybean Research Needs”; National Soybean Re-
Goosens, A. E. Food Process. I n d . 1975, 44, 20. search Coordinating Committee: Washington, DC, 1974.
Green, S. A.; Pust, H. J . Phys. Chem. 1958, 62, 55. Rackis, J. J.; Sessa, D. J.; Honig, D. H. J . Am. Oil Chem. SOC.
Gremli, H. A. J. Am. Oil Chem. SOC. 1974,51,95A. 1979, 56, 262.
Hammonds, T. M.; Call, D. L. Chem. Technol. 1972, 2, 156. Sawyer, D. T.; Brookman, D. J. Anal. Chem. 1968, 40, 1847.
Honig, D. H.; Warner, K. A.; Seike, E.; Rackis, J. J. J. Agric. Food Schutte, L.; Van Der Ouweland, G. A. M. J. Am. Oil Chem. SOC.
Chem. 1979,27, 1383. 1979, 56, 289.
Johnson, D. W. “World Soybean Research Proceedings of the Sessa, D. J. J. Agric. Food Chem. 1979, 27, 2340.
World Soybean Research Conference”; Hill, L. D., Ed.; In- Smith, A. K.; Circle, S. J. “Soybeans: Chemistry and Technology”;
terstate Printers and Publishers, Inc.: Danville, IL, 1976; p Avi Publishing Co.: Westport, CT, 1978.
1014. Twigg, G.; Kotula, A. W.; Young, E. P. J. Anim. Sci. 1977, 44,
Kalbrener, J. E.; A. C.; Moser, H.; Wolf, W. J. Cereal Chem. 1971, 218.
48,595. Wolf, W. J.; Baker, F. L. Cereal Chem. 1975, 52, 387.
Kiselev, A. V.; Yashin, Yy. I. “Gas-AdsorptionChromatography”; Wolf, W. J.; Cowan, J. D. ”Soybeansas a Food Source”;CRC Press:
Plenum Press: New York, 1969. Cleveland, OH, 1975.
Lyon, L. Food Prod. Deu. 1980, 14, 58.
MacKenzie,C. A. “Unified Organic Chemistry”;Harper and Row: Received for review July 14,1982. Revised manuscript received
New York, 1962. December 30,1982. Accepted January 10,1983. Portions of this
McMullin,S. L.; Bernhard, R. A.; Nickerson,T. A. J. Agric. Food paper were presented at the World Soybean Research Confer-
Chem. 1975,23,452. ence-11, March 26-29, 1979, Raleigh, NC. This work was sup-
Meyer, E. W. “Proteins as Human Food”; Lawrie, R. A., Ed.; Avi ported by the Iowa State University Research Foundation and
Publishing Co.: Westport, CT, 1970; p 346. the Iowa Agriculture and Home Economics Station (Journal Paper
Meyer, E. W.; Williams, L. D. “World Soybean Research Pro- No. 5-10721, Project 2164, a contributing project to the North
ceedings of the World Soybean Research Conference”;Hill, L. Central Regional Research Project NC-136).

Characterization of the Main Secondary Components of the Liquid Sugars from

Cane Molasses

Gerardo Palla

The main nonsugar components of liquid sugars from cane molasses have been identified as phenolic
and phenylpropanetriolic glucosides coming from the lignin material of cane during juice extraction
and processing. The glucosides have been recovered by column absorption using nonionic polymers,
have been fractionated through a silica gel column, and have been characterized by GC-MS, lH NMR,
and 13C NMR techniques.

In a previous work (Palla, 1982) we reported an efficient weight has also been detected (10-30%). The character-
method to recover minor nonsugar organic colored com- ization of the compounds has been made by MS, GC-MS,
ponents of sugar juices and syrups obtained by cane mo- lH NMR, and l3 C NMR and by reference to synthetic
lasses demineralization. We reported also the structure standards such as phenols, phenyl glucosides, and phe-
of some phenolic compounds that we considered the main nylpropanetriols.
components of the sugar color. We have continued the
EXPERIMENTAL SECTION
work on the characterization of the nonsugar components,
and we can now give an exhaustive description of the minor Recovery and Fractionation. The recovery of the
components of this type of sugar syrup. To this purpose colored nonsugar material has been made following the
we examined the silica gel chromatographic fractions of absorption method described for nonionic polymers (Palla,
the raw organic material recovered from a wide series of 1982); in this way 1.2 g of colored material (80% of the
liquid sugars; we found the greatest amount of nonsugars total nonsugars) has been recovered from 300 g of liquid
was composed by phenyl glucosides (10-20%), phenyl- sugar sample (70% of dry matter, 99.3% of total sugars).
propanetriol glucosides (40-60%) and minor amounts of The material recovered was dissolved in methanol and
other phenolic derivatives. A fraction with high molecular filtered. The amount dissolved (1.05 g) was chromato-
graphed through a silica gel column (1.5 X 70 cm, Kieselgel
60 Merk), eluted with ethyl acetate and methanol (713
Istituto di Chimica Organica dell’Univerdta’, 43100 v/v). A control on the column effluent showed eight main
Parma, Italy. chromatographic bands: A (9 mg, R, 0.78, pale yellow), B
0021-8561/83/1431-0545$01.50/0 0 1983 American Chemical Society
546 J. Agric. Food Chem., Vol. 31, No. 3, 1983 Palla

fraction, gave glucose and a series of aglucons that were

silanized and characterized by GC-MS as follows: the raw
hydrolysate (100 mg) was passed through a silica gel col-
umn (1X 20 cm) eluting with ethyl acetate and methanol
(8/2 v/v); the UV-absorbingfraction was collected, dried,
and treated with 1 mL of silanizing reagent. After being
ppm 150 100 so warmed, the mixture was analyzed by GC-MS (Figure 3A).
Figure 1. l3 C NMR spectrum of the 3,4-dimethoxyphenylp- The phenolic nature of the aglucons was confirmed by
D-glucopyranoside,with the probable assignment of the lines. GC-MS spectra of the peracetylated derivatives that were
obtained by treating 100 mg of the raw hydrolysate with
2 mL of acetic anhydride and 1 mL of acetic acid under
reflux for 2 h (Figure 3B).
Bands G H (93 mg) had UV characteristics similar to
the ones of F fraction and gave the same green reaction
with FeC13. Enzymatic and acid hydrolyses did not give
enough material for further analyses. The 'H NMR
spectrum showed signals in the aromatic region, as well
as the products of the F fraction.
For the brown final methanolic fraction (110 mg), both
Figure 2. Mass spectrum of the 3,4-dimethoxybenzyl /3-D- enzymatic and acid hydrolyses released UV-absorbing
glucopyranoside. products with slightly increased Rf compared with that of
the starting material. Silanized products, however, did not
(182 mg, R, 0.63, orange), C (48 mg, Rf 0.50, pale yellow), show volatility enough for GC-MS analysis.
D (32 mg, Rf 0.42, pale yellow), E (14 mg, R 0.35, pale Synthesis of the Phenolic Standards. Phenolic
yellow), F (512 mg, Rf 0.30, yellow), G (22 mg, 8,0.18, pale glucosides, benzylic glucosides, and phenylpropanetriols
yellow), H (71 mg, Rf 0.08, pale yellow). The strongest UV have been synthesized by known methods to facilitate the
absorption was given by the bands B and F, corresponding identification of the compounds isolated. Glucosides have
to the fractions I11 (B) and I1 (F) recovered with Sephadex been obtained by reaction of phenol, 3,4-dimethoxyphenol,
G-10 (Palla, 1982). Another brown fraction (110 mg) benzylic alcohol, and 3,4-dimethoxybenzylicalcohol with
corresponding to the fraction I from Sephadex has been acetobromoglucose,in presence of silver oxide (Helferich
recovered by washing the silica column with methanol after et al., 1935). The acetobromoglucose has been synthesized
completion of the chromatographic elution. This brown from peracetylated glucose and hydrobromic acid (Horn-
fraction was 68% retained in dialysis experiments by ing, 1955). Phenylpropanetriols have been synthesized
membranes with retention limits of 10000. from the corresponding cinnamic acids (Adler and Gus-
Characterization of the Chromatographic Bands. tafsson, 1963).
Band A (9 mg) gave a color reaction with resorcinol and
sulfuric acid but not with FeC1,; the UV spectrum showed RESULTS AND DISCUSSION
a large absorption band between 250 and 300 nm; the The analytical studies on the chromatographic fractions
identification was unsuccessful. Band B (182 mg), as A-H permitted us to recognize the nature of about 7040%
previously reported, is composed of several phenolic glu- of the nonsugar material recovered from the liquid sugars
cosides, the most abundant being the 3,4-dimethoxyphenyl examined. The @-glucosidesof polyphenols and phenyl-
P-D-glucopyranoside,of which we can now give the 13C propanetriols represented 18 and 5070,respectively, of the
NMR spectrum, recorded on a Varian XL-100 instrument total raw material extracted in the sample studied.
(FT 25.15 MHz, in CD,OD) (Figure 1). Phenyl glucosides, whose structure has been previously
Bands C-E (96 mg) have been considered together. described (Palla, 19821, greatly contribute to the color of
They seemed composed of small amounts of several com- the syrup, as well as the phenylpropanetriols, which gen-
pounds; free glucose and fructose were also detected in this erate yellow compounds under oxidative conditions. The
fraction. The mass spectra performed on a Varian Mat series of silylated and peracetylated phenylpropanetriols
CH-5 spectrometer, at 70 eV, suggested the presence of with GLC profiles and the main m / e peaks is reported in
glucosides of ketols such as 1-(3,5-dimethoxy-4-hydroxy- Figure 3.
phenyl)-3-hydroxy-2-propanone(M+ = 388) and 144- The structure of the most abundant aglucons of the F
hydroxy-3-methoxyphenyl)-3-hydroxy-2-propanone (M+ band has been confirmed by their 13CNMR spectra being
= 358) or their isomers. Traces of 3,4-dimethoxybenzyl in good agreement with the ones of similar compounds
glucoside have also been detected (M+ = 330, identical with coming from lignin degradation (Ludeman und Nimz,
the synthetic sample, see Figure 2). The enzymatic hy- 1974) and by comparison with the synthetic products ob-
drolysis of these fractions by a-and p-glucosidase failed; tained from cynnamic acids. The I3C patterns of the (4-
the treatment with methanolic acid solution (HC1,8%, for hydroxy-3-methoxyphenyl)-, the (3,5-dimethoxy-4-
2 h, at 60 OC) provided products whose GC-MS analyses, hydroxyphenyl)-, and the (3,4,5-trimethoxyphenyl)-
performed on a Varian gas chromatograph connected with propanetriols are reported in the Figure 4.
a single-focus Varian Mat CH-5 mass spectrometer, using The line assignments to the carbons have been made by
a 1.5% SE-30 glass column and programmed conditions considering similar products and models; in many cases,
(100-250 OC, 20 OC/min), gave a series of peaks apparently however, especially for substituted aromatic carbons, the
confirming the presence of the ketols (M' = 226 and M+ assignments are probable but not certain. A problem is
= 196) and simple polyphenols such as dimethoxyphenols also the assignments of the configuration to the a- and
(M+ = 154), dimethoxyhydroxyphenols (M+ = 170), and @-carbonsof the side chain (carbons a and b in Figure 4):
trimethoxyphenols (M' = 184). we think the more abundant isomers have the threo form,
Band F (512 mg, brown-green reaction with FeCI,) was as this one seems to be favored in the enzymatic synthesis
the most abundant fraction recovered; the enzymatic hy- (Higuchi et al., 1974). But in this way we assign the lower
drolysis, carried out as previously reported for the B GLC retention time to the threo form, in contrast with the
Main Nonsugar Components of Liquid Sugars J. Agric. Food Chem., Vol. 31, No. 3, 1983 547

A.GLC P r o f i l e s
- Peaks Structure mle ( r e l . i n t . )
'" c o
S t . S t e e 1 column: d=lmm,
3% SE-30
Progr. 140-220", 20"Imin
1=2m
cH3a$\$]i CHCHCH2 239(9) ,225(90) ,194(9) ,147(20) ,137(11) , 1 1 7 ( 4 ) ,
73(100).

B.GLC P r o f i l e s
P r o g r . 150-200". 15"Jmin

n EHgHEHZ
384 (M+, 11), 3 2 4 ( 1 4 1 , 2 3 9 ( 8 ) , 2 2 2 ( 4 8 ) , 1 9 7 (100).

J
A i 2 3 4 bmm
Figure 3. GLC Profiles and m / e values of the silylated aglycons (A) and of the acetylated aglycons (B) recovered from the F band
hydrolysate.

results of Nurok et al. (1968) that proposed lower GLC

retention times for a series of erythro derivatives of bu-
tanediols.
We studied also the bonding site of glucose in the
phenylpropanetriols. Examples are reported of bonding
to the a-,the 0-,and the y-carbons (meander, 1965). We
could reasonably state through a mass fragmentation study
made on natural and synthetic glucosides that the glucose,
in our phenylpropanetriols, is bound to the a-carbon. In
Figure 5 we report the typical fragmentation of a phenyl
glucoside (path A), a benzyl glucoside (path B), a phe-
nylpropanetriol glucose (path C),and a phenylpropanetriol
(path D).
We can see that if glucose is bound to the aromatic ring
or if the a-carbon brings a free hydroxy group, then the
main mass peak retains the glycosidic oxygen or the hy-
ppm bo 100 50
T-
droxy group (A,m l e = 154; D,m l e = 183); the glycosidic
Figure 4. '%!N M R spectra and probable assignmentsfor natural oxygen is instead missing when the glucose is bound to the
phenylglycerols. a-carbon (B, m l e = El),and this happens also in the

CHO-
3 n

CH3 co
, / 6i-i CHO

mle 151 I1001

167 (8)
m e ' 167 (1001
1 8 3 I101
m t e ' 183 (100)
167 1321

Figure 5. Typical mass fragmentation of phenolic glucosides, benzyl glucosides, and phenylpropanetriols.
548 J. Agric. Food Chem. 1983, 3 1 , 548-553

fragmentation of our phenylpropanetriol glucosides (C, molasses, due, perhaps, to the higher content of phenolic
m l e = 167). compounds.
The phenylpropanetriols contribute to the color of the
ACKNOWLEDGMENT
liquid sugars examined, probably as their oxidation or
dehydration products. We found these products to behave I thank W. V. Turner for the NMR spectra, E. Rosa for
like the phenyl glucosides, generating a bright yellow color the mass spectra, and G. Simonazzi for experimental as-
when exposed to air on silica plates. The glucosides iso- sistance.
lated most probably come from the lignin fraction of stalks Registry No. Lignin, 9005-53-2; 3,4-dimethoxyphenyl8-D-
of cane. It is known that acid media and high tempera- glucopyranoside, 84812-00-0; 3,4-dimethoxybenzyl 8-D-glUC0-
tures accelerate the depolymerization of the lignin with pyranoside, 81381-73-9; 1-(4-hydroxy-3-methoxyphenyl)-3-
formation of phenolic material (Lundquist and Lundgren, hydroxy-2-propanone, 4899-74-5;1-(3,5-dimethoxy-4-hydroxy-
1972). It is not surprising that sugar syrups from beet phenyl)-3-hydroxy-2-propanone,35263-53-7; glucose, 50-99-7;
molasses, produced without grinding and without acid fructose, 57-48-7;(3,4,5-trimethoxyphenyl)propanetriol, 76774-
03-3; (4-hydroxy-3-methoxyphenyl)propanetriol, 1208-42-0;
treatment with phosphoric and sulfurous acid in the (3,5-dimethoxy-4-hydroxyphenyl)propanetriol, 4204-29-9.
clarification steps, do not contain appreciable amounts of
these phenolic derivatives. In the syrups from beet mo- LITERATURE CITED
lasses we never detected more than 0.04% of phenolics Adler, E.; Gustafsson, B. Acta Chem. Scand. 1963, 17, 27.
(percent based on the total sugar content). Harborne, J. B. In “Biochemistryof Phenolic Compounds”;Ac-
The knowledge of the chemical structure of the colorants ademic Press: London, 1964; Chapters 4 and 13.
suggests that anionic macroporous resins would decolorize Helferich, B.; Scheiber, H. E.; Streeck, R.; Vorsatz, F. Justus
these syrups. The resins can interact with the free phenolic Liebigs Ann. Chem. 1935,518, 211.
groups and adsorb the organic polar compounds. Exper- Higuchi, T.; Nakataubo, F.; Ikeda, Y. Holzforschung 1974,28,189.
Homing, E. C. In “Organic Syntheses”;Wiley: New York, 1955;
iments until now gave satisfactory absorption only by Collect. Vol. 111, p 11.
passing the diluted and well-demineralizedsyrup through Ludemann, H. D.; Nimz, H. Makromol. Chem. 1974,175,2393.
the resins, after regeneration with ammonia. Work is in Lundquist, K.; Lundgren, R. Acta Chem. Scand. 1972,26,and
progress to improve the decolorizing process. references cited therein.
The chemical and the biological significance of the Nurok, D.; Taylor, G. L.; Stephen, A. M. J. Chem. SOC.B 1968,
presence of phenolic glucosides in liquid sugars has also 291.
been considered; trace amounts of these phenolics are very Palla, G. J. Agric. Food Chem. 1982, 30, 764.
widespread in nature (Harborne, 1964))and toxic effects Theander, 0. Acta Chem. Scand. 1965, 19, 1972.
for human nutrition have never been reported in literature.
The polyphenol derivatives can exert some antioxidant and Received for review August 24,1982. Accepted January 26,1983.
antimicrobial effect (Harborne, 1964); we found that di- This work has been partially supported by the National Research
luted liquid sugars from cane molasses were more micro- Council (Italy) and by the Sugar Research Department of the
biologically stable than the corresponding syrups from beet Reggiane O.M.I. of Reggio Emilia (Italy).

Gas-Liquid Chromatography-Chemical Ionization Selected Ion Monitoring Assay

for Glycerol Formal in Animal Tissues
William J. A. VandenHeuvel, III,* Francis P. Baylis, Joseph E. Brown, Dennis H. Wallace,
David H. Minsker, Richard T. Robertson, Avery Rosegay, and Robert W. Walker

Glycerol formal, a 60140 mixture of the two cyclic condensation products [5-hydroxy-1,3-dioxaneand
4-(hydroxymethyl)-l,3-dioxolane] of glycerol and formaldehyde, is used as a nonaqueous solvent in
parenterally administered animal health products. Recently this substance was reported to be a teratogen
when administered to rats in large doses. We have developed an assay for glycerol formal in the edible
tissues (fat, kidney, liver, muscle) and plasma of food-producing animals (cattle, swine, horses). The
assay utilizes packed column GLC-chemical ionization (isobutane) selected ion monitoring (glycerol
formal-d2 serving as the internal standard) and possesses an overall lower limit of sensitivity of 0.05
ppm. The maximum glycerol formal residue found in an edible steer tissue (injection-site muscle) at
5 days postdose (14.6 mg/kg of body weight) is -0.1 ppm; Our teratolgy studies demonstrate that doses
of glycerol formal less than 150 mg/ kg do not elicit a teratogenic response in rats. Thus, the pressure
of the negligible residue (-0.1 ppm) in steer injection-site tissue does not appear to represent a significant
hazard to the consumer.

Although numerous studies aimed at detecting submi- ducing animals are carried out routinely in the develop-
crogram quantities of drugs in the tissues of food-pro- ment of new animal health drugs, little attention is usually
given to the tissue concentration of the components of the
Merck Sharp & Dohme Research Laboratories, Rahway, dosing vehicle. Glycerol formal, a mixture of the cyclic
New Jersey 07065 (W.J.A.V., F.P.B., J.E.B., A.R., and condensation products of glycerol and formaldehyde, has
R.W.W.), Merck Sharp & Dohme Research Laboratories, been used as a nonaqueous solvent in parenterally ad-
Fulton, Missouri 65251 (D.H.W.), and Merck Sharp & ministered products (Spiegel and Noseworthy, 1963) such
Dohme Research Laboratories, West Point, Pennsylvania as trivetrin (“ABPI Compendium of Data Sheets for
19486 (D.H.M. and R.T.R.). Veterinary Products”, 1978) and oxytetracycline (Green-

0021-8561/83/1431-0548$01.50/0 0 1983 American Chemical Society