Browning Reaction Theories Integrated in Review

DEHYDRATED FOODS Chemistry of Browning Reactions in Model Systems

JOHN E. HODGE Northern Regional Research Laboratory, Peoria, III.

Many different types of organic reactions lead to the production of brown pigments at moderate temperatures. In spite of many reviews of the subject, there has been no comprehensive organization of the reactions. In this review some relationships are shown to exist among the carbonyl-amino, the nonamino, and the oxidative types of browning. Recent findings have provided the basis for an integration of the several isolated partial theories of browning (Maillard, sugar fission, ascorbic acid, furaldehyde) heretofore proposed. The significance of the occurrence of the Amadori rearrangement in the Maillard reaction is stressed, and a mechanism for browning in sugar-amine systems based upon the rearrangement is outlined. Attention is directed to the little-studied but important role of dehydrogenated reductones in both enzymatic and nonenzymatic browning reactions. Investigations of browning reactions in model systems during the past 3 years are reviewed, with the pertinent older studies, and the results of most of these are shown to fit into the proposed scheme of reactions. A classified directory to the major part of 201 references on browning in nitrogenous model systems (1 940 to March 1953) is included.

CHEMISTRY OF BROWNING RE- Model Systems Important oxidative browning reactions has been THEACTIONS in nitrogenous model sys- For Browning Research published by Joslyn and Ponting (82). tems has been adequately reviewed to the Browning, of whatever type, is caused Three broad of reac- year 1950 by Danehy and Pigman (33). types browning by the formation of unsaturated, colored tions are in food However, some important findings in recognized technology. polymers of varying composition (33, The most common model system studies since 1950 are inter- type, carbonyl- 39, 41, 45, 46, 82, 96, 180, 197). Com- amino includes the reactions related. reactions, pounds that engender browning usually of and More recent reviews of browning aldehydes, ketones, reducing contain a carbonyl or potential carbonyl with amines, amino acids, reactions (9, 30, 96, 97, 98, 107, 113, 131, sugars pep- grouping. [Pyrroles are an exception, tides, and Another called 136, 137, 138, 156, 158, 177, 183, 201) proteins. type, but their participation in food browning occurs when have, in general, emphasized the com- caramelization, polyhy- has not been adequately demonstrated plexity of the subject and the lack of droxycarbonyl compounds (sugars, polv- (cf. 79).] Polyhydroxy compounds (87. acids) are heated to specific knowledge on the chemical hydroxycarboxylic 168, 196) and sugars in which the car-

Downloaded via UNIV OF SAN DIEGO on March 14, 2021 at 21:27:00 (UTC). in the reactions and intermediates involved. relatively high temperatures bonyl function is blocked (116, 167, 170, absence of amino This The complexity of browning reactions compounds. 196) do not give rise to browning. of re- has recently been demonstrated (in a type browning characteristically Moreover, the extraction of a browning more to started than See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. superficial way), for chromatographic quires energy get system with a solvent that removes, the carbonyl-amino reactions, other investigations have shown the presence of inter alia, carbonyl compounds will retard conditions Both acids and at least 15 components in a glucose- being equal. or eliminate browning (69, 180). Brown- bases are known to catalyze carameliza- aqueous ammonia system (78) and 24 ing can also be inhibited by the addition tion reactions, but little more than this or more definite compounds in a glucose- of reagents that will combine with or is known. The of carameliza- glycine reaction mixture (28; cf. 70, 139). chemistry eliminate carbonyl groups. A discussion tion is However, along with such perplexing astonishingly underdeveloped of browning inhibitors comes later—the reviews 117. Neither car- disclosures have come revealing demon- (see 201). point to be made here is that browning nor caramelization reactions strations, through the characterization of bonyl-amino can be regarded as stemming from car- are the of oxy- a few isolated compounds, that certain dependent upon presence bonyl compounds. With this viewpoint, to A third definite of reactions are involved. gen produce browning. the relationships among the carbonyl- types broad of Because these scattered disclosures of type browning frequently amino, the nonamino, and the oxidative encountered the food is the reaction types are capable of orderly by processor types of browning reactions are more of oxidative reactions for arrangement, a perspective view of the group which, readily apparent. convert ascorbic and entire group of chemical reactions known example, acid poly- The naturally occurring compounds as browning can now be drawn. As phenols into di- or polycarbonyl com- from which browning originates usually many gaps in our present knowledge are pounds. These oxidations may or may contain, not only one, but a multiplicity indicated, the correlation should be of not be enzyme-catalyzed. A compre- of potential carbonyl groups. Examples value in programming future studies. hensive review of the enzyme-catalyzed of such compounds are catechols and

92g AGRICULTURAL AND FOOD CHEMISTRY the vicinal polyphenols, ascorbic acid, D. Sugar fragmentation (0.2 M) in an aqueous solution of sodium and the Catechols E. Amino acid degradation at the same Also, reducing sugars. III. Final hydroxide pH (76). are stage (highly colored) of easily oxidized to quiñones. Like- F. Aldol condensation the ultraviolet absorption spectra wise, ascorbic acid and other reductones G. Aldehyde-amine polymeriza- heated dilute aqueous solutions of d- [the term “reductone” is used in its tion; formation of hetero- glucosyl-n-butylamine correspond closely general sense (53) throughout this paper] cyclic nitrogen compounds to that of D-glucose in sodium hydroxide even more readily yield vicinal tri-, As reaction types B, C, D, E, F, and G (27). Nevertheless, one should not -dí-, or conjugated dicarbonyl com- can follow spontaneously from A. a conclude, at least in the case of the less pounds. Since these oxidations may be reaction scheme that unifies the in- strongly basic glycosylamines or of more auto- (<57) or enzyme-catalyzed (82) or dividual reaction types can be formu- concentrated solutions, that browning is reagent-induced, it is apparent that no lated. The interrelationships of these the result of only an alkaline sugar real difference need exist between reactions are outlined in Figure 1, which degradation (16, cf. 27). Generally, enzymatic and nonenzymatic browning shows how the partial mechanisms for .V-substituted glycosylamines remain un- after the initial oxidations. Sugars are browning—i.e., the furfural, sugar-fission, dissociated in aqueous solutions to also examples of naturally occurring and ascorbic acid (reductone) theories appreciable extents (27, 122, 148); compounds that can yield a multiplicity discussed by Stadtman (777)— can be hence, the irreversible transformations of carbonyl groups. Reducing sugars considered as part of the total sugar- that the undissociated molecules undergo are transformed (by dehydration and/or amine condensation (Maillard) theory. must be considered. The end results in to enediols to vs. fission) and osones, and All the reactions shown in Figure 1 the two cases (glycosylamine glycose reductones and dehydro reductones, are known to occur in model browning decomposition) should differ, but the which provide -dicarbonyl or con- systems, but the extent to which each exact nature of the alleged differences is jugated dicarbonyl groups and color. represents the browning processes of not known. Some comparatively simple organic natural products is unknown. Recent researches have established compounds will undergo polymerization In this review the types of reactions that sugars do condense with amino and browning at ordinary temperatures, which have been found to occur in model acids in aqueous solutions. As the solu- even when they are prepared in a state systems are discussed and correlated tions are concentrated, as in food de- of high purity. Especially important in according to the scheme of reactions hydration, the reversible sugar-amine this class are the ,ß-unsaturated alde- presented in Figure 1 (lettered para- condensation (Reaction A) goes toward hydes—e.g., furaldehydes—and the a- graphs of the text correspond to the completion (70, 101, 119). In the final dicarbonyl compounds—e.g., pyruval- lettered reactions in Figure 1). An concentrated solution (or dehydrated dehyde, diacetyl, and reductones in the attempt has been made to include all food) there are ¿V-glycosyl derivatives of dehydro form. Apparently the presence published work since the time of Danehy amines, amino acids, and proteins. of oxygen or oxidation products is re- and Pigman’s review (33) and through In a kinetic study of sugar-amino acid quired for the browning of these types of March 1953. Only such older work as is condensation, Haugaard, Tumerman, compounds in the pure state (34, 92, pertinent to the discussion is mentioned. and Silvestri (73) used a novel solubility 162). In the presence of amino com- technique. They measured the rate of pounds, the browning is even more rapid Initial Stage of Sugar-Amine Reaction solution of amino acids from the solid and intense 170, of a saturated solution (36, 144, 153, 168, 199). btadtman (777) and phase aqueous The interaction of and A. Sugar-Amine an aldose The initial reducing sugars Danehv and Pig- containing (pH 9). amines readily such Condensation reaction was and the amino produces spon- man (33) suggested reversible, reactive as those 1 1 taneously compounds that the term “Maillard reaction” be acid combined with the aldose in a to named above. ratio. irreversible reserved for browning reactions involving Later, decomposition It can be concluded that the most of the condensation the interaction of reducing sugars and sugar-amine product model for studies of was important systems amino compounds. It has not been indicated. reactions are those in which In another kinetic browning clear, however, to what extent sugar- study, Katchalsky -hydroxycarbonyl (or a-aminocar- amine condensation is involved in the and Sharon (85) added alkali contin- bonyl) are transformed into to the solutions of aldoses compounds Maillard reaction—for example, some uously aqueous unsaturated colored and amino acids to the of the polymers. authors have pointed to the catalytic keep pH reaction mixture constant as the effects of amino groups on reducing sugar- Mechanism of Browning in amine condensation reactions occurred. sugar transformations without apparent Sugar-Amine Systems mechanism involved sugar-amine condensation. This seem- Their proposed dual both and Nearly all of the research on browning ing divergence of views can now be catalysis by hydrogen reactions in model systems published explained. The evidence indicates that hydroxyl ions. The rate of the primary of since 1940 has involved reactions of sugar-amine condensation is the first reaction increased with the basicity the the of carbonyl with amino compounds. Classi- step in the so-called catalysis of amino amino component; velocity decreased with increase of fied references to these studies (through compounds. After condensation and the reaction and the extent of the reaction de- March 1953) are given in Table I. rearrangement, the sugar moiety is pH; creased with The rate The literature on sugar-amine brown- dehydrated and easily polymerizable decreasing pH. the reaction of ing reactions in model systems reveals and unsaturated compounds are formed in constants for primary seven different types of reactions which which the amine moiety is labile. Even- glycine with the various aldoses showed are that ease of of the known to occur during browning. tually, the amine is split off (in part) the opening pyranose These of form type reactions (lettered A through from the dehydrated sugar residue to give ring (percentage aldehyde pres- G) may be classified for discussion pur- the effect of a true catalysis. ent) was a contributing factor. In poses according to the three stages of It has been shown (16, 36, 174) that agreement with the latter conclusion are and Traitteur development: sugars and amines undergo browning in the findings of Speck (174) solution in tn the (183) that color formation in aqueous I. Initial no ab- aqueous proportion stage (colorless; acid mixtures sorption in near-ultraviolet) basic strength of the amines employed. reducing sugar-amino A. Augar-arpine condensation For example, the strongly basic D-glu- is directly proportional to the percentage B. Amadori rearrangement cosyl-piperidine (0.2 M) in aqueous solu- of aldehyde form of the reducing sugar. II. Intermediate (colorless or stage tion with at 25° C. used the same method yellow, with strong absorption glycine (2.0 M) Katchalsky in near-ultraviolet) (pH 8.4) undergoes browning at approxi- (keeping the pH constant) in an earlier C. Sugar dehydration mately the same rate as does D-glucose study with Frankel (57) and could find

VOL. 1, NO. 15, OCTOBER 14, 1953 929 no evidence for a reaction between fruc- Table.!. Studies Pertaining to Browning in Nitrogenous Model Systems tose and various amino acids in dilute aqueous solutions. However, fructose (Published since 1940) is known to Condense with amines under V-Substituted conditions. glycosyíamines , (4, 15, 16, 27, 36, 63-65, 74, 75, 76, 122, 125, 132, other 148, 149) Haugaard et al., Katchalsky et at., Amadori rearrangement products (4, 15, 63-65, 76, 122, 172) and Taufel and Iwainsky (179) were Ammonia and ammonium salts unable tó use the Van Slyke or formal Reactions with titration procedures to determine amino Hexoses (14, 36, 47, 78, 151, 169, 183, 186) in kinetic studies. The use of Aldehydes (141, 142,153) groups has con- Amines, amine salts, and ¡mines these methods in the past led to Reactions with flicting conclusions (33; cf. 21). Taufel Hexoses (16, 27, 36, 48, 122, 151, 174, 183, 186) and Iwainsky also showed the inade- Pentoses (174, 180) quacy of copper reduction methods Trioses ; (4, 16, 171, 172) for the determination of Biose ... . (16, 26) reducing sugars Aldehydes in sugar-amino acid reactions, since ' v'!· Fatty f35, 126, 176, 184) strongly reducing substances are gener- Unsaturated (55, 193) ated in the reactions. isolated Ketones and ketols t. They (8U26, 171) the Reductones (29. 53, 189) reducing sugar chromatographically Dehydro reductones (51, 72) before analysis and determined amino Amino acids acids by the method of Pope and Stevens Reactions with which involves the solubilization ; ! (150), Hexoses of from the Aldoses copper copper phosphate by (,1;:9, 21, 28, 31, 32, 36, 47, 58, 59, 63-67, 70, 71, amino acid. 73, 76, 77, 84, 85, 89, 91, 94, 96, 98, 107, 109, 110, 114, 115, 118, 119, 132, 133. 135, 140-143, Hannan and Lea (70, 71) demon- 745, 168-170, 174. 178. 181, 183, 185, 186, strated 1 to 1 reaction between glucose 195-197, 199) and the free amino groups of a-.V-acetyl- Ketoses (1, 32, 77, 89, 133. 145. 168, 185, 186) Pentoses lysine, polylysine, and casein. Con- (9, 21, 36, 77, 79, 96, 145, 174, 179, 183, 195, 197) Trioses (16, 36, 65, 67, 174) centrated solutions of the reactants at Biose (16) pH 8.5 were freeze-dried and stored at Hexose phosphates (36, 167) 37° C., pH 6.4, at different humidities. Deoxy sugars (77, 79, 80) In all three cases the Methylated sugars (174, 196) glucose (determined Disaccharides (36, 67, 83, 89. 127) by glucose oxidase) and the free amino Oligosaccharides (other than di- ) (36. 194) groups [determined by the Van Slyke and acids Glyconic glycaric (768) method and confirmed by Sanger’s Uronic acids (36, 86, 168. 183) fluorodinitrobenzene reaction (101, 166)} Keto acids (36, 67, 145, 164) Aldehydes disappeared in a 1 to 1 ratio in the initial Fatty (24, 40, 67, 126, 143) colorless stage of the reaction. Galac- Keto (1, 36, 39, 40. 44, 45, 164, 174) tose reacted with as did Unsaturated casein, glucose (36, 40, 67, 153, 179) Mohammad, Fraenkel-Conrat, Ketones and ketols (36. 79, 743. 164. 171) (105). and Olcott showed that Reductones (6, 29, 31, 32. 53, 76, 88, 140, 144, 145) (124) glucose Dehydro reductones (88, 140, 144, 164) and the free amino groups of bovine Peptides and proteins serum albumin and other proteins and Reactions with combined in a 1 to 1 ratio Hexoses peptides to Aldoses (13, 74, 84, preliminary browning. 87, 89, 100, 98-105, 111, 124, 134, 140, condensa- 765, 166) Crystalline glucose-glycine Glucosamine (98. 105, 106) tion products isolated from browning Ketoses (111, 765) reactions were reported by Patron (137) Pentoses 165) (141, and (115). Only small Deoxy sugars (98, 105) Mackinney were iso- Methylated sugars (13) amounts of these compounds Disaccharides (56, 99, 165) lated, and their structures have not been Uronic acids (111) determined; both compounds gave anal- Aldehydes (87, 126) yses a mole for mole reaction Reductones (20, 182) indicating Amides between glucose and glycine. reaction Formamide, dehydroascorbic acid (144) Three acidic glucose-glycine Acetamide (122, 126) products (neither crystalline nor homo- Pyrroles (79, 143) geneous) were isolated in low yields by No browning (under conditions used) elution of the reaction mixture (concen- Nonamino compounds trated aqueous solution heated at Hexitols (87, 168, 196) 45 °C., 5) from an anion-exchange Glucose-1 -phosphate pH (767) column Barnes and Methylated glucoses (13, 16, 73, 196) by Kennedy (10). O-Glycosides (170, 196) Determinations of carbon, hydrogen, and Sucrose (9, 27, 36, 76. 87, 89, 99) molecular (a) C8Hi708N — Lactic acid weight gave: (67) - Maleic acid 1.5H20, (b) [C8H1708N 2H20]2. (186) - Amino compounds and (c) [C8H17OsN 3H20]3 4- Triethylamine (783) The monomeric product, (a), was Triethanolamine (126, 174) nearly colorless, but it produced color />-Aminobenzoic acid cf. (79; 67) on heating in water. Products (b) and -V, ,V-Dimethylglycine (174, 196) were more three Urea (186) (c) highly colored. All Succinimide (186) condensation products were more stable Aminoanthraquinones (186) on standing in air than normal .V-sub- stituted giycosylamines.

930 AGRICULTURAL AND FOOD CHEMISTRY Ar-Glucosyl'derivatives of the sodium without an excess of aniline,' 2 moles: of gand (190), the reaction is acid an- Salts of glycine, serine, and lysine (2 (S-phenylglyceraldehyde will'react1 in' the alyzed: moles of with the two glucose reacted RNH r rnh RNH RNH amino -of were groups lysine) prepared II I 1 by Micheel and Kleiner (119) by evapo- HC----, ch CH ch2

* 1 I! I rating aqueous solutions of the compo- 1 +H 1 (HCOH),, O --· HCOH COH =r=± C.’O nents to dryness.: These compounds were Í contrast to the 1 1 analytically pure, in HC-j (HCOH),, (HCOH )n (HCOH)„ impure calcium and barium salts re- CH2OH ported previously by Kuzin and Polya- ch2oh _ ch2oh_ CH2OH kova (94). A-substituted Cation of Enol form A-Substicuted Several chromatographic studies which aldosylamine Schiff base 1 -amino-1 -deoxy- 2-ketose, keto form demonstrate sugar-amine condensation have been reported (28, 65, 70, 139, cold with one mole of aniline to form the It has never been demonstrated that From the reactions of 179). browning crystalline compound, CeHsNXCHOH.- the Amadori rearrangement product is and amino reducing sugars acids, spots CHOH.C6H5)2, which readily decom- in equilibrium with the Ar-substituted were obtained on in solutions. Whereas paper chromatograms poses to a brown tar. Diglycosylamines glycosylamine which reacted with both sugar spray this had been considered are readily formed from reducing sugars rearrangement reagents and ninhydrin, and also with to occur with and ammonia in aqueous alcohols, as previously only '-glycosyl p-dimethylaminobenzaldehyde in the derivatives of aromatic amines, has been shown on paper chromatograms primary manner of glycosylamine and amino Hodge and Rist (75, 76) demonstrated by Bayly, Bourne, and Stacey (14) and sugar derivatives. that the rearrangement occurs also with Raacke-Fels (151). Finally, the re- Chichester, Stadtman, and Mackinney derivatives of both primary searches of Mohammad, Fraenkel-Con- Ar-glycosyl (28) used C14-labeled glucose and C14- and aralkyl and alkylamines. and Olcott and of Lea and secondary labeled glycine with unlabeled glycine rat, (724) Furthermore, the rearrangement of N- Hannan revealed and glucose, and showed that the major (70, 101, 102) that, substituted glycosylamines occurred and the free reaction product was derived from both although reducing sugars spontaneously in the dry or nearly dry amino of combine glucose and glycine. Elution of this groups proteins state at 25° C. When the C-2 hydroxyl in a 1 to 1 the ratio in- major intermediate from a cellulose initially ratio, of a crystalline Ar-substituted glycosyl- creases and 1.5 to 1 column gave a crystalline compound approaches during amine was unsubstituted, the compound the of the reaction. It is which analyzed for carbon, hydrogen, latter stages was slowly transformed during storage to here that one mole of water nitrogen, and molecular weight in agree- significant a dark brovyn tar. From the tarry ment with the calculated values for a was liberated for each mole of sugar mixture, the Amadori rearrangement 1 to combined with the rnonohydrated glucose-glycine 1 protein (107, 104). product was isolated in 30 to 50% yields. condensation product (175). Another As yet, diglycosylamine formation re- On the other hand, if the C-2 hydroxyl 1 to 1 to be demonstrated in compound of a reducing sugar and mains sugar-amino was substituted (as in 2-O-methyl-D- an amino acid was isolated similarly by acid or sugar-protein browning systems. glucosylpiperidine), then the derivative Hannan and Lea (70) after paper chro- „ . . . Speculations that the remained white and stable after B. Amadori . , . storage matography of a freeze-dried glucose- „ ^ Amadori rearrange- for 2 at 25° C. Under - experi- Kearrangement° , years a-A-acetyllysme reaction mixture. ment ot -sub- mental conditions that caused the rapid It can stituted could be involved be concluded that reducing glycosylamines rearrangement of D-glucosylpiperidine, in nonenzvmatic reactions have sugars and amines condense in an equi- browning 2-O-methyl-D-glucosylpiperidine was re- molar in the in ratio in browning reactions. appeared literature, starting covered unchanged. Thus, it was shown The initial reaction is reversible, but 1945 (13, 15, 65, 71, 79, 103, 148), but that blocking the Amadori rearrange- soon its occurrence was not demon- irreversible reactions follow (70, actually ment of an Ar-substituted glycosylamine until Gottschalk and 73, 703). The accepted mechanism strated (63) Hodge blocked the browning which otherwise for sugar-amine condensation involves and Rist (75) reported their results in would have occurred. of the form of the 1952. Earlier, Barnes and opening ring sugar, Kennedy Several other groups of workers (16, addition of the amine to the (70), Gottschalk and Partridge (65), carbonyl 73, 76, 80, 196) have used different types group, and subsequent elimination of a and Lea and Hannan (103) had indi- of model systems composed of 2-deoxy molecule of water to form the -substi- cated how the rearrangement of first- or 2-substituted aldoses and amino tuted glycosylamine (cf. 85). formed glycosylamines would explain acids. Each group showed that the RNH C-2 hydroxyl of an aldose is essential RNH RN for the occurrence of a significant degree HC---, HC:0 CHOH CH of browning. Only one conflicting

. instance has been Lea and I +rnh2 | -h2o | (CHOH)„_, O reported: —± - (CHOH), ^-i (CHOH),, ,; (CHOH)„ I I Rhodes (105) found that 2-deoxy-n- HC--1 underwent a more CH2OH ÓH-OH CH2OH galactose rapid CH2OH browning reaction with casein than d- it reacted with Aldose in Addition Schiff base Ar-Substituted galactose, although the free amino more than aldehyde form compound (not isolated) glycosylamine groups slowly D-galactose. There was evidence in As the -glycosyl derivatives of am- some of their experimental findings. their study (cf. 706) that the browning of monia and of many primary aliphatic The available evidence that the Amadori 2-deoxy-D-galactose with casein, like amines crystallize initially with a mole of rearrangement is a key reaction for that of 2-deoxy-2-amino-D-glucose with “water of hydration” (68,. 122, 125, browning in aldose-amine, and ketose- casein, did not proceed mainly through 748), the final step of water elimination amine systems is presented below. carbonyl-amino condensation. may not actually occur in some cases. Aldoses. The Amadori rearrange- It is concluded that autoisomerization Diglycosylamines may be formed in ment (93, 790) is the isomerization of is a general reaction of Ar-substituted sugar-amine browning reactions. Smith A-substituted aldosylamines to 1-amino- aldosylamines on standing in moist and Anderson (777) have shown that, l-deoxy-2-ketoses. According to Wey- atmospheres, since most of them (un-

VOL 1, NO. 15, OCTOBER 1 4, 1953 931 substituted at the sugar hydroxyls) glucose-»-A"-acetyllysine compound also of hygroscopic, yellowish white crystals show decomposition with browning to gave no glucose on acid hydrolysis (70). were isolated. The compound in some extent on storage (75, 75, 76, 725, Chichester and Mackinney’s crystalline aqueous solution was colorless, acidic, 7J2). jV-Glycosyl derivatives pf pro- glucose-glycine reaction product (de- and strongly fluorescent, and it reduced teins should react similarly (cf. 103). scribed under A) gave no glucose upon alkaline silver nitrate in the cold to form Although the Amadori rearrangement acid hydrolysis and 50% of the glycine a silver mirror. It gave only a faintly products are more stable than the was recovered. Furthermore, on heat- positive Seliwanoff test; but, after original glycosylamines in moist, acidic ing with dilute acetic acid, 5-hydroxy- reaction with nitrous acid (fleeting pale atmospheres, they are still heat-labile, methylfurfural (HMF) was formed. blue color), the Seliwanoff test was even in the dry state. On heating, they It is apparent from Gottschalk’s study strongly positive, indicating the presence undergo dehydration and fission and (63) that hydroxymethylfurfural is much of hydroxymethylfurfural. Analysis yield colorless reductones as well as more readily formed from 1-amino-1- gave 5.3 ± 0.2% nitrogen, the same as brown fluorescent substances (76). deoxy-D-fructose derivatives than from for Chichester and Mackinney’s com- The Amadori rearrangement products D-fructose itself under equivalent condi- pound (calculated for CgHisCbN.HsO: undergo browning decomposition alone tions. 5.5% N). Because no further data were in aqueous solutions at alkaline pH, and Chichester and Mackinney’s crystal- obtained, it is not known whether this the rate of browning is considerably line compound was colorless, but it compound was the glycosylamine or the enhanced in the presence of amino acids darkened readily on warming. It gave rearrangement product; however, its (76). An alkaline pH is not necessary a relatively large levorotation, [»]D — stability to hydrochloric acid, its strong for browning, however; the browning 40°. [For comparison, the sodium reducing power, and its easy conversion reactions of 1-deoxy-l-piperidino-D-fruc- salt of D-glucosylglycine in water gives to hydroxymethylfurfural all indicate tose with glycine ethyl ester hydro- [»]n — 8 (10 minutes) +20° (779), the rearrangement product. Rigorous chloride at pH 6.3 (initial) were more and D-glucosylglycine ethyl ester in proof of structure for all of the above- rapid than those of o-glucosylpiperidine ethanol gives [»]D — 8 (5 minutes) -*· discussed glucose-amino acid compounds or of D-glucose with glycine at pH 8.4. + 2° (76, 799).] But, most revealing is still needed. The recent researches of Gottschalk was the fact that a solution of the com- Smith and Anderson (777, 772) (63) with D-glucose and phenylalanine pound in 0.1 N sodium hydroxide re- studied the rearrangement of trioses. and cf Hodge and Rist (76) with d- duced dichlorophenolindophenol solu- Their experiments showed that both a glucosylglycine ethyl ester showed that tion readily at room temperature (un- primary aromatic amine and acetic the Amadori rearrangement occurs also published experiment performed jointly acid were necessary to effect the re- with .V-glycosyl derivatives of amino by Chichester and the author, March arrangement of 3-phenylglyceraldehyde acids. More experiments of this type 12, 1953). The total physical and to ß-phenylpyruvaldehyde. Dimeric ß- are needed, however; for Gottschalk’s chemical behavior of this compound, phenylglyceraldehyde and A’-methyl- product was a hygroscopic sirup, and which was isolated from a browning aniline (772), when refluxed in absolute Hodge and Rist’s hygroscopic crystalline reaction conducted under Maillard’s alcohol containing a little acetic acid, product was not entirely homogeneous. conditions (28), indicates that it is prob- apparently underwent condensation and Nevertheless, both products analyzed ably a monohydrate of the Amadori a typical Amadori rearrangement to for an isomer of the 1 to 1 sugar-amine rearrangement product of D-glucosyl- yield (probably): C6HsN(CH3).CHs- condensation product and gave the glycine—namely, 1 - (A-carboxymcthyl) - CO.CHOH.CeH5. (The carbonyl and characteristic reducing behavior of 1- amino-1 -deoxy-D-fructose. hydroxyl groups could have formed on amino-1 -deoxy-2-ketoses. Patron (737) heated an aqueous solu- carbons 3 and 2, respectively.) The A comparative tabulation of the prop- tion of glucose (1 M) and glycine (1 M), white crystalline product reduced Fehl- erties of vV-substituted glycosylamines adjusted to pH 4.5 with acetic acid, at ing solution and alkaline methylene blue and the corresponding Amadori re- 100° C. for 30 minutes. The reaction at room temperature. The correspond- arrangement products is given in Table mixture exhibited strong blue fluores- ing compound from phenylalanine ethyl II. From the Amadori rearrangement cence in ultraviolet light, although there ester also reduced Fehling and alkaline product, none of the aldose is recoverable was only moderate darkening of the methylene blue solutions and decom- after acid hydrolysis, although about solution at this point. The fluorescent posed with browning after standing only 50% of the amine moiety is released. material was separated as a silver salt 1 hour at room temperature. This was the behavior of Gottschalk’s by a method described by Lewis (107, Ketoses. Although it is now estab- rearrangement product obtained from 108). After liberation of the compound lished that the Amadori rearrangement the reaction of glucose with phenyl- with hydrochloric acid (in which the occurs after aldose-amine condensation alanine (63). Hannan and Lea’s 1 to 1 compound was stable), about 100 mg. in model systems which undergo brown- ing, very little is known about the first reactions by which ketoses undergo Toble II. Properties of N-Substituted Glycosylamines and Corresponding browning. Maillard (776) reported that Amadori Rearrangement Products fructose browns with amino acids at a rate faster than I-Amino-1 ~deoxy-2- somewhat glucose. N-Substituted Glycosylamines keloses This finding has been corroborated Reducing power under various conditions by several In dilute NaOH at 25° C. others (36, 145, 168, 174, 183, 186). o-Dinitrobenzene after 1 hour after 1 minute Purple Purple fructose browns Methylene blue Very slow decolorization Rapid decolorization However, with casein, Dichlorophenolindophenol Very slow decolorization Rapid decolorization at a very slow rate, even more slowly Fehling solution than the disaccharides, maltose, and lac- At 25° C. Little or no reduction Reduction tose At 100° C. Reduction Rapid reduction (770). condense with amines Tollens reagent, 25° C. Little or no reduction Reduction Ketoses aliphatic Browning with amino acids at 25° C. Slow Rapid in aqueous alcoholic media more readily Ninhydrin reagent with heating Color Color11 than aldoses, according to the recent Products of acid hydrolysis (63) Glucose, amine, little Glucose, some amine, HMF much HMF communication of Erickson (48). Con- aldoses condense with aromatic « versely, Reported to be produced more rapidly with 1-deoxy-l-piperidinofructose than with more than do D-gtucosylpiperidine (76). amines readily ketoses, according to Barry and Honeyman (72).

932 AGRICULTURAL AND FOOD CHEMISTRY In Erickson’s experiments more than one by the spectra could develop either by Cantor (727); but whether or not this mole of octadecylamine apparently con- sugar dehydration alone or by sugar compound is formed in browning re- densed with fructose or sorbose, as fission and dehydration of the (recom- actions is still unknown. though multiple Amadori-type re- bined) fragments. A third reaction, Táufel and Iwainsky (181) heated a arrangements had occurred along the the Strecker degradation of a-amino neutral solution of glucose (20%) for ketose chain. Raacke-Fels (757) also acids, is included in the intermediate 37 hours at 96° C. without detectable had some indications from spot reactions stage, although it follows fission and/or change when developed on a paper on paper chromatograms that fructose dehydration or dehydrogenation of the chromatogram. After 42 hours of heat- combined with ethylamine and propyl- sugar residue. ing a spot appeared which corresponded amine in aqueous alcoholic solutions. Two types of in Rt and color reactions to hydroxy- C. Dehydration From the reactions of amines and sugar dehydra- methylfurfural, and the spot increased Of Sugar amino acids with model ketoses (such as Moiety tion reactions are in size and intensity on further heating. acetol, hydroxyacetophenone, ,ß-di- now known. In acidic systems, furfurals In the presence of amino acids, the spot hydroxypropiophenone, dihydroxyace- are formed. In the presence of amines hydroxymethylfurfural appeared after tone, phenyldihydroxyacetone, acetoin, in nearly anhydrous systems, 6-carbon only 14 hours of heating. Thereby, the benzoin, and propionoin), as reported and other reductones are formed. This accelerating effect of amino acids on by Hurd (79), Smith and Anderson reductone-producing reaction may pro- sugar dehydration was again demon- (777), and Carson (26), it appears that vide an important heretofore missing strated. an Amadori-type rearrangement is to link in the browning mechanism. The results of Gottschalk (63) explain be expected for A-substituted ketosyl- Furfural Formation. The accelerat- the “catalysis” of furfural formation by amines. The similarities of these re- ing effect of glycine, aspartic acid, and amines. When fructose was heated at actions with the Amadori rearrange- arginine (also ammonium chloride, so- 100° C. in 2 A acetic acid for 2 hours, ment are pointed out by Anderson (4). dium pyruvate, diacetyl, morpholine, no hydroxymethylfurfural was detectable The necessity for oxidation in the forma- and hydrogen peroxide) on the color- chromatographically. A considerable tion of a dicyclohexylimine derivative of producing reactions in aqueous solutions amount of hydroxymethylfurfural was diacetyl from acetoin and cyclohexyl- of furfural was demonstrated by Rice, obtained under these conditions, how- amine, as reported by Carson (26), Kertesz, and Stotz (153). Woifrom and ever, either from the D-glucose-phenyl- is an important point for consideration coworkers (779, 799) showed that glycine alanine Amadori rearrangement product in future studies of the browning re- accelerates the production of furfural or from 1 -deoxy-1 -/>-toluino-D-fructose. actions of ketoses. and color from o-xylose, and of hydroxy- D-Glucosyl-/>-toluidine was hydrolyzed According to Hurd (79), the model methylfurfural and color from n-glucose, under these conditions to D-glucose and ketoses, acetoin and acetol, undergo in dilute aqueous solutions. They fol- /i-toluidine. The explanation for the rapid browning with alanine; further- lowed the reactions by means of ultra- liberation of amine, but not sugar, by more, 2,4-dimethylpyrrole (which could violet absorption spectra and proposed a the hydrolysis of Amadori rearrange- be formed from acetol and alanine) mechanism for the dehydration based ment products is now evident. The undergoes rapid browning with acetol. upon the spectra. The mechanism amine is hydrolyzed from the Schiff Acetol and alanine formed 2,5-dimethyl- involved the initial loss of one mole of base of hydroxymethylfurfural after the pyrazine also, but the pyrazine deriva- water with production of the a-enolic- dehydration of the sugar moeity. The tive did not brown rapidly with acetol. ,/3-unsaturated aldehyde (I), which deoxyaminoketose per se is not hy- Benzoin reacted with alanine to yield would exist in equilibrium with the 3- drolyzed. The catalysis of furfural for- both tetraphenylpyrrole and tetraphenyl- deoxyosone (II). mation and resultant browning as first pyrazine. In view of the likelihood of outlined Gottschalk and HC=0 HC=0 by Partridge Amadori-type rearrangements with both (65; cf. 63) has been incorporated in aldoses and ketoses in browning reac- C—OH c=o Figure 1. After condensation of the tions, pyrrole and pyrazine derivatives sugar and the amine (Reaction A) and CH CH, should be sought in sugar-amine reac- rearrangement to the easily dehydrated tions and their role in the I or tautomeric browning (HCOH)„ (HCOH)„ deoxyaminoketose 1,2- established. Both fructosamine (1-am- enolic structure (Reaction B), the loss of ino- 1-deoxvfructose) and glucosamine CHoOH CH2OH 3 molecules of water (Reaction C) occurs have (2-amino-2-deoxyglucose) long I II readily (acid catalyzed) to form the been known to undergo self-condensation Schiff base of the furfural. The amine to substituted yield pyrazines. Because The argument for a mechanism in- is subsequently liberated from the Schiff of the that the forma- uncertainty exists, volving 1,2-enolization, rather than de- base (but only in part) by hydrolysis. tion of heterocyclic nitrogen compounds hydration, as the first step was given by The fraction of the amine not liberated in Ama- a sugar-amine browning via the Newth (128), and correlation of the is combined in the melanoidins (humin dori has not been in- enolization and mechanisms rearrangement dehydration substances) which accompany the re- cluded in 1. was more Figure presented recently by Petuely7 actions (63). (746). The role of furfurals in brow'ning Reductone Formation. Hodge and Intermediate has been discussed by Stadtman (777) Stage Rist (76) have show7n that when d- and Danehy and Pigman (33). The initial stage of browning (Reac- glucose and piperidine are heated to- The formation of hydroxymethyl- tions A and B) cannot be detected by gether at 70° to 80° C. in the presence of furfural in glucose-glycine browning measurements in the malonic acid and an excess of spectrophotometric reactions under Maillard’s conditions piperidine, ultraviolet; however, before visible the Amadori occurs first; was demonstrated chromatographically rearrangement has a on the chain browning begun, strong absorption by Chichester, Stadtman, and Mac- then, longer heating, sugar to announce the of is with the loss of 2 molecules appears beginning kinney (28). Compound 7 on their dehydrated the intermediate The initial water. stage. illustrated chromatogram has since been of The crystalline product, Cn- strong absorption occurs at 2770 to found by them to be isomeric with Hi;03N (isolated in 20% yield), had the 2850 A. (the furfural region) for the hydroxymethylfurfural (28). A crystal- properties of a true reductone (53, 147). reaction of reducing sugars with most line isomer of hydroxymethylfurfural The reductone (considered to be amino acids (28, 59, 70, 170, 199). identified as 2-(hydroxyacetyl)furan, has CeHgOj) retained the piperidyl radical, The conjugated unsaturation indicated been isolated recently by Miller and and the reductone-piperidine compound,

VOL 1, NO. 15, OCTOBER 1 4, 1953 933 C11H17O3N, was neutral in aqueous 193). Open-chain dienediolic inter- that carbonyl-amino reactions aré ex- solution (internal neutralization of the mediates obviously may form upon sugar cluded; on the contrary, the déhydro acidic enediol grouping by the amine dehydration and never undergo the reductones, like the furfurals; browri more radical). Glucosylpiperidine and galac- ring closure to furan derivatives but rapidly in the presence of amino com- tosylpiperidine gave the same optically polymerize instead to form melanoidins. pounds than alone. inactive reductone. Seaver and Kertesz that Galactosylpiperi- COOH COOH RC:0 (168) reported dine gave this compound (12%) after L-ascorbic acid undergoes browning with it had stood at 25° C. for 2 years in the C—OH C: O C—OH glycine (90° C., pH 4.3) at a faster rate II “dry” state. An aqueous solution of I! I! than the common sugars; however, no — + the CH 2H CH CH was to piperidyl-reductone turned brown x I attempt made exclude air in their of itself in air at 6. In the I Patron found that pH presence CH +2H + CH CH experiments. (137) of a more is the form of ascorbic acid amino acids, rapid browning | || it dehydro reaction occurred. C—OH C:0 C—OH that undergoes the browning reactions. The same was ob- | I When was excluded from solu- piperidyl-reductone COOH COOH R' oxygen tained from the brown residue formed by tions of L-ascorbic acid and glycine, heating glucosylpiperidine, galactosyl- III IV V there was no browning; dehydroascorbic or acid and piperidine, 1-deoxy-l-piperidino- The recent communication of Wol- glycine readily produced fructose in the dry state at 110° C. browning. Another reductone, di- from, Schlicht, Langer, and Rooney (cf. 125). When the heating was done hydroxymaleic acid (which is more (197) cites results indicating that the in a vacuum of 0.3 mm., a sirupy distil- easily oxidized than ascorbic acid), repeating units in the water-soluble, late (ca. 15%) which possessed reductone underwent browning very easily with nondialyzable aldose-glycine melanoi- properties was also formed. The dis- glycine or amines, and this reaction was dins of their study approach the sugar tillate was water-soluble initially, but accelerated in acetic acid-buffered solu- dehydration stage of a furfural-glycine underwent very browning on tions at pH 4.0. rapid condensation product. However, the standing in air and finally formed a dark Pecherer (144) showed that dehydro- results “do not require that the furan water-insoluble polymer. ascorbic acid (from the crystalline meth- rings are necessarily present in their The A-glucosyl derivative of a primary anol complex) gave a pink color with intact cyclized forms.” Under their amine (monoethanolamine) also was glycine or -alanine (and other amino conditions sugar dehydration (whether spontaneously dehydrated on standing to acids, some not a) when refluxed in to furfurals or to open-chain reductones) a reductone (76). However, in water or in alcohol (cf. 20, 88, 140); produce apparently is the dominant reaction in this case, the product was amorphous the pink color deepened to cherry red the formation of sugar-amine melanoi- and melanoidinlike, and may have been within 1 minute and became dark brown dins. However, sugar fission and re- slowly polymerized. The A-hydrogen after 50 minutes. Evaporation of the combination of the fragments to 6- atom in the prob- solvent gave an intractable brown, glucosylethanolamine carbon dehydrated residues are not ably allowed further condensations, amorphous material. ruled out by their study. which could not have occurred in the Proteins readily form complexes with Browning Reactions of Reductones. case of the tertiary glycosylpiperidines. copper, according to Thompson, Kocher, The development of reductonelike re- It is now apparent that the sugar in and Fritzsche (182), and the copper- ducing power is a w'ell-known charac- sugar-amine browning reactions can be proteins undergo a browning reaction teristic of sugar-amine browning dehydrated in two ways. In neutral or with reductones at 30° C., pH 6.5 to 7. reactions (22, 40, 42, 103, 179, 187). acidic systems, furfurals are The amorphous, fluorescent red-brown aqueous Furthermore, the bro .-n polymers which formed. In the dry state, or in non- product from the reaction of copper- form are known to possess strong reduc- aqueous solvents in the presence of casein with ascorbic acid contained no ing power and a redox system similar to amines, reductones are formed. It is reduced reductone groupings (indo- those of reductones (40, 42, 195). In possible that the reductones possess the phenol titration) but did contain the spite of these facts, the structures of the structure of furfurals without furan ring dehydro form of a reductone (dinitro- reductones formed during sugar deg- closure. Whether or not the 6-carbon phenylhydrazine reaction). There was radation are not known with certainty, reductone is formed without fission, or by no browning if any one of the three com- and in only a few model systems has the fission and recombination of the (copper, casein, or ascorbic acid) frag- browning of reductones with amines been ponents is not known. was omitted. Other reductones (re- ments, studied (Table I). and showed ductic catechol, Petuely Künssberg (147) Ascorbic acid is the best known reduc- acid, hydroquinone) that a,a'- reacted as rapidly as ascorbic acid with by polarographic analysis tone, and most of the browning studies acid is a true the copper-casein complex. Further- dihydroxymuconic (III) have involved this single enediol. It is reductone. It is a conjugated dienediol, more, furfural and furfuryl alcohol, probable that all reductones will be rather than an enediol-(a); neverthe- which (in the open-chain form) could found to undergo similar browning the two enolic are form dehydro reductones, also showed a less, hydroxyls readily reactions, as they all have the same type to give the dehydro positive but weaker browning reaction. dehydrogenated of conjugated unsaturation shown in the form Furfurals, by hydrolytic In a series of model systems represent- (IV). following general formula: scission of the ring, could yield the ing the constituents of orange juice, conjugated dienediolic structure (V; OH OH Curl (31, 32) measured the browning and R = R' = or 1 rH Hi i carbon dioxide evolution of various H; H, CH2OH, CH3). = —C=LC—C J„=C—C=* (n 0 or 1) 2-Hydroxyacetylfuran (121) also would synthetic combinations of ascorbic acid, yield V (R = CH2OH, R' = H) which, The unsaturation shown at the asterisk sugars (glucose, fructose, sucrose), and by analogy, probably is also a reductone is frequently, but not always, that of a amino acids (alanine, asparagine, arg- (cf. 147). As a matter of fact, furfurals carbonyl group (cf. 53); conjugated inine). The citrate-buffered solutions do produce reductonelike reducing power unsaturation is necessary for stability were sealed in tin cans and stored at upon acid or alkaline hydrolysis (147) of the reductone. It is not the reductone 49° C., pH 3.75. The color density for although the structures of the reductone, per se, but the dehydro form of the increased only slightly in 60 days formed are not known. A ring opening reductone, that is the prime source of ascorbic acid alone, for the mixture of of furfural, such as shown in V, is known browning. The browning of reductones sugars alone, and for ascorbic acid plus to occur with aromatic amine hydro- therefore belongs to the class of oxidative the mixture of amino acids. Appreci- chlorides in the Stenhouse reaction (55, browning reactions. This is not to say able browning was obtained, however,

934 AGRICULTURAL AND FOOD CHEMISTRY recovered were is in the solution containing sugars and lactone of glucuronic acid), when dis- fragments small, it. amino acids. The carbon dioxide pro- solved in 2 A sodium hydroxide at 90° conceivable that much larger amounts and one or could have been in the duction of the sugar-ascorbic acid C. , is completely converted, present initially were held back systems (37) was greatly increased by the more reductones containing the equiv- reaction mixture but by cf. addition of copper, iron, tin, or mercury alent of one enediol group per molecule melanoidin formation (38; 788). salts. Mixtures containing sugars, as- of glucurone are formed. With these Fructose gave much more pyruvaldehyde corbic acid, and amino acids yielded evidences of reductone formation from than did glucose with both glycine and about the same amount of carbon di- Uronic acids, the browning may occur ß-alanine in Speck’s experiments. oxide as concentrated orange juice. through the dehydro form of the reduc- The accepted mechanism for sugar Acceleration of the decomposition of tones as well as furfural formation. fragmentation is dealdolization—i.e., the L-ascorbic acid to carbon dioxide in the The fission prod- reverse of aldol condensation, As amines presence of amino acids was reported D. Fragmentation ucts of the sugars or amine salts are known to catalyze by Ashikaga (6). In the light of present Of Sugar Moiety vary considerably aldol condensation (24, 54, 92, 784), knowledge, the carbon dioxide could in their potential they will also catalyze the reverse re- have come from the carboxyl group of for browning. Fragments which retain action, if the reaction is truly reversible. the amino acid by way of the Strecker the ct-hydroxycarbonyl grouping will Speck and coworkers (774) investigated degradation (744, 763) as well as from undergo browning alone in aqueous the types of catalysts effective for the the carboxyl group of ,/3-diketo-L- solutions; and, in the presence of amino dealdolization . of diacetone alcohol. gulonic acid. The latter compound is compounds, the browning is greatly They found that the dealdolization is produced by opening of the lactone ring accelerated. The most highly reactive subject to amine catalysis, but not to of dehydroascorbic acid prior to decarbox- compounds are glycolaldehyde (76, 26), general base catalysis. The anions, not ylation (50-52). Both decarboxylation glyceraldehyde (36, 65, 774, 783), the zwitterions, of amino acids func- mechanisms would require preliminary pyruvaldehyde (36, 39, 44, 774, 783), tioned as catalysts. The magnitude of dehydrogenation of the reductone. acetol (79, 758, 788), dihydroxyacetone the catalytic constant for methylamine Most of the studies on the browning (774, 783), acetoin (26, 79, 758), and in this reaction was approximately 20 reactions of ascorbic acid have involved diacetyl (26, 36, 67, 774). Acetalde- times that for the /3-alaninate ion and 65 a natural product (cf. 777). According hyde (35, 39, 726) is only slightly less times that for the -alaninate ion. to both Krijt (90) and Miller (720), re- reactive, aldol is still a little slower (67, Solutions of xylose or glucose with ductones other than ascorbic acid are not 79, 726), and propionaldehyde is a very ethanolamine, diethanolamine, or tri- present in fresh fruits and vegetables, slow reactant (726). Even slower re- ethanolamine were adjusted to pH 6.0 but, after storage, all dehydrated and acting are the keto acids, pyruvic (36, with sulfuric acid and distilled. Mono- processed foods contain increasing 67) and levulinic (67, 770), while the ethanolamine sulfate produced severe amounts of the other reductones with saccharinic acids, lactic and acetic, do browning with the formation of black increasing storage time and tempera- not react with amino compounds (67). insoluble melanoidins, and pyruvalde- ture (cf. 40, 42). Baraud (7) recently Formic acid has not been shown to be a hyde was found in the distillate. Rela- reported that an unknown substance, source of browning, and formaldehyde is tively large amounts of furfural were which reacts with />-sulfophenylhydra- not only inactive (726), but is actually found in the distillate from the xylose zine in the manner of dehydrogenated an inhibitor for furfural (69), carameliza- solution. Diethanolamine sulfate pro- reductones, is formed from glucose in tion (83), and carbonyl-amino (cf. 33, duced a similar but less marked effect. various wines. The unknown dicar- 777) types of browning. No browning occurred with triethanol- bonyl compound was not derived from Triose-reductone, which is known to be amine sulfate, and only a small amount triose-reductone, dihydroxymaleic acid, produced in the alkaline degradation of a of pyruvaldehyde was found in the dis- or ascorbic acid. The same reductone- large number of reducing sugars (52, 86, tillate. Speck concluded that amines like substance was formed in borate- 789), forms crystalline condensation and amino acids which accelerate the buffered glucose solutions which were products with amines and amino acids degradation and browning of reducing heated on a boiling water bath. Greatly (29, 53, 789). A comparative rate of sugars do so through the participation of increasing amounts of the substance were browning has not been reported for the primary or secondary amino groups. obtained with increasing increments of triose-reductone. By analogy with other Figure 1 shows how the amines partici- pH in the range 5.6 to 7.8. Wendland reductones, it would be expected to un- pate in the reactions. After sugar-amine (787) showed that reductones were dergo rapid browning reactions in the condensation (A), the Amadori re- produced from buffered glucose solutions dehydro form—i.e., as mesoxalic di- arrangement (B) produces the 1,2-enol by heating on a boiling water bath (pH aldehyde, CHO.CO.CHO. or 2-ketose configuration in which the 6.9 and up) or on long standing at room The fragmentation of sugars in neutral C—C bond ,ß to the carbonyl group is temperature. The addition of glycine and slightly acidic solutions was first weakened. The amino compounds pres- or protein catalyzed the reaction. reported by Nodzu (729) and Enders ent then accelerate the dealdolization to It is apparent that more intensive and Marquardt (44). Nodzu identified trióse and other fission products. study of reductone formation from the acetol and/or pyruvaldehyde, pyruvic The isolation of 4(5)-methvlimidazole sugars and the manner in which de- acid, and diacetyl. Enders assumed and 2-methyl-5- and 2-methyl-6?-(D- hydrogenated reductones form brown that their product was pyruvaldehyde, craio-tetrahydroxybutyljpyrazine from pigments would throw much needed but Sattler and Zerban (758, 759) the browning reaction of D-glucose in light upon both enzymatic and non- showed that Enders could have had ace- aqueous ammonia at 37° C. (78), and enzymatic browning processes. tol instead of, or in addition to, pyruval- the detection of 4(5)-methylimidazole in As for the strong browning reactions dehyde, since their bishydrazone deriva- the browning reaction of glucose and gly- given by uronic acids and pectic sub- tives are identical. Pyruvaldehyde and cine (43) provide additional evidence for stances (36, 86, 777, 768), it should be diacetyl were positively identified in the the fragmentation of sugars during remembered that uronic acids are de- distillates from solutions of sugars and browning. composed on heating in alkaline solution amino acids by Speck (774). Other The Amadori rearrangement product, to yield indophenol-reducing substances volatile substances which reduce perio- 1 -deoxy-I-piperidino-D-fructose, was (reductones) (52, 86). In strong acids, date—e.g., acetol, glycolaldehyde—were shown by Hodge and Rist (76) to furfural and reductic acid are formed indicated (by difference) from periodate undergo fission on heating to yield, inter (86). Recently, Euler and Hasselquist oxidation analyses of the distillates. alia, piperidine acetate. Evidently, pi- (52) have shown that glucurone (a Although the amounts of the volatile peridine was split from the compound

VOL. 1, NO. 15, OCTOBER 14, 1953 and the sugar chain was rearranged to has not been properly recognized in pre- reaction of nonselectively labeled glucose yield acetic acid. The piperidine ace- vious reviews as an important partial with inactive glycine. Wolfrom et al. tate formed in this decomposition would mechanism for browning. Evidence for (797) have reported even higher per- catalyze the aldol condensation of sugar its existence is scattered throughout the centages (90 to 100%) of carbon dioxide fission products with accompanying literature back to the time of Maillard arising from carboxyl-labeled glycine. dehydration of the aldols (cf. 92, 784). (776, cf. 2, 3). Recently the degrada- Ambler (3) reported long ago that Hence, this is a mechanism to be con- tion has been reviewed experimentally by aldehyde production paralleled carbon sidered for the formation of the 6-carbon Schonberg, Moubacher, and Mostafa dioxide evolution in sugar-amino acid reductone. (764), who found that only carbonyl browning reactions. More recently, The volatile acid liberated by the compounds containing the structure: carbon dioxide evolution was found to treatment of with 0.035 N sodium color glucose O O parallel production (96, 709, 795) hydroxide at 35° C. in the absence of and show about the same dependence II I I II is acetic as was determined — Patron’s oxygen acid, —C— [ c=c—J c— upon pH (795); however, by Sowden and Schaffer (773). It experiments (737) under similar condi- would not be surprising to find other where n is zero or an integer, are capable tions did not confirm these reports. analogies between the decomposition of initiating the degradation. Com- Stadtman et al. (778) found that less reactions of Amadori rearrangement pounds easily converted to such di- carbon dioxide was produced per unit products, on the one hand, and of the carbonyl structures—e.g., reductones by of optical density at low temperatures Lobry de Bruyn-Alberda van Ekenstein dehydrogenation, imino analogs by hy- than at high temperatures, and the rearrangement products on the other. drolysis—are also reactive. The total melanoidins formed at 56.5° C. con- Barker and Cromwell (8) reported the reaction is illustrated by the following tained more labeled carbon from car- spontaneous decomposition in the solid equation: boxyl-labeled glycine than those formed state of a-hydroxy-d-piperidyl-/3-phenyl- under Maillard’s conditions at 100° C. RCOCOR + R'.CHNH2.COOH — to brown sub- In a propiophenone (VI) tarry R'-CHO + C02 + RCHNH2COR xylose-glycine system under stances which contained phenyl benzyl nitrogen at 95° C., pH 4 to 5, 2 moles diketone (VII). The transamination shown in the of xylose reacted to produce one mole of carbon and the reaction c5h10n C5H,oN dioxide, stopped at that point, according to Langer and HC0 HC0 HC0 CH20 Wolfrom (96, 795, 797). In the presence I of air the could an HCOH COH coh C:0 sugar decarboxylate excess of glycine. The osone structure 0C:O 0COH 0C:O 0C:O (II) or a sugar fragment, such as pyruv- VI VII aldehyde, were mentioned as likely -dicarbonyl intermediates which could Therefore, it is not unreasonable to above could be an equation important have caused the Strecker degradation. assume that reaction for of 1-deoxy-l-piperidinofruc- incorporation nitrogen The Amadori rearrangement product, tose lose piperidine simi- into the brown polymer. Akabori (2) (VIII) might 1 -deoxy-1 -piperidinofructose, produced to IX. found that little ammonia was larly yield very the Strecker degradation of a-amino- C5H10N CsHI0N phenylacetic acid in aqueous solution in the absence of air The trans- CH. (76). ch2 ch2 ch3 —* 11 formation VIII IX (shown under 1 I! Reaction would the C:0 COH _ c6H„N COH C.O D) give a-dicarbonyl | — structure necessary for the Strecker HOCH COH C:0 C:0 degradation. The fission products of such as and di- c3h7o3 C3H7O3 c3h:o3 C3H70; VIII, pyruvaldehyde as well as reductones VIII IX acetyl, dehydro- genated by dismutation, are other Fragments such as glycolaldehyde and liberated in the degradation of amino possible sources of the required di- acetylglycolyl could be formed from acids by glucose or furfural. The likeli- carbonyl compounds. IX by dealdolization. Glycolaldehyde hood of transamination should also be Pecherer (744) showed that dehydro- could yield acetic acid by a saccharinic remembered in using the Van Slyke ascorbic acid in methanol degraded a- acid rearrangement, and acetylglycolyl determination of -amino nitrogen in alanine to carbon dioxide and acet- could yield it by oxidative scission. browning reactions. If the a-dicar- aldehyde with concurrent browning. Piperidine acetate may have arisen by bonyl compound is an osone—e.g., Quinones, which are formed from poly- such transformations on heating VIII. Compound II or pyruvaldehyde—the phenols by oxidases, would also cause Weygand and Bergmann (797) oxi- -amino nitrogen is not lost in the re- the Strecker degradation of a-amino dized Amadori rearrangement products action, but merely transferred to the acids (763, 764). The low temperatures in an ammoniacal solution with oxygen -position of the osone moiety where it which prevail during enzymatic brown- and a platinum catalyst and obtained could continue to react with nitrous acid. ing may strongly limit the extent of the arabonic acid (ammonium salt), the Maillard (776) concluded that the Strecker degradation, but the tempera- amine, and carbon dioxide. This type of carbon dioxide liberated in his browning tures and other conditions necessary oxidative fission may be a possible source reactions must have come from the car- for significant reaction have not been of carbon dioxide from the sugar radical boxyl group of the -amino acid, rather satisfactorily fixed (cf. 2, 763). (carbon 1) for sugar-amine browning than from the sugar radical. Recent The aldehydes formed by the Strecker reactions that occur in the presence of tracer experiments with carboxyl-labeled degradation are a source of browning. oxygen. glycine have upheld Maillard’s hypothe- They could condense with themselves, _ „ , The Strecker degrada- sis. Stadtman, Chichester, and Mac- with sugar fragments, with furfurals E. Strecker . c _ . or , tion (763) of a-amino kinney (778) showed under Maillard’s and other dehydration products, with Degradation . , V . , , , . acids (to aldehydes con- conditions that over 80% of the carbon aldimines (734) and ketimines to form Ot Amino . . , , . . . one less and ^ taming carbon dioxide generated from glucose brown pigments. Mohammad, Olcott, Acid Moiety1 ”, . . , than the ammo acid, glycine came from the carboxyl group, and Fraenkel-Conrat (726) demon- with the liberation of carbon dioxide) and less than 10% came from the strated that both proteins and amino

936 AGRICULTURAL AND FOOD CHEMISTRY Figure I. Amador! rearrangement in integration of known reactions leading to browning in sugar-amine systems acids undergo very rapid browning Strecker degradation is a nonessential Final Sfage reactions with acetaldehyde, but no but possibly concomitant important In the final of the browning was obtained with form- reaction leading to melanoidin forma- stage browning, intermediates and un- aldehyde. Since glycine (which yields tion. polymerize saturated, fluorescent, colored formaldehyde by the Strecker degrada- The source and amounts of carbon polymers are formed. The chief reactions in- tion) browns in some cases more and in dioxide formed in brow'ning reactions volved are to be aldol con- other cases nearly as rapidly as alanine where the Strecker degradation could thought densation, aldehyde-amine (which yields acetaldehyde), it is ap- not occur—e.g., with ammonia, amines, polymeriza- tion, and the formation of parent that the production of aldehydes and amino acids other than a—have heterocyclic such as from amino acids by the Strecker deg- not been fully investigated. Lewis, nitrogen compounds, pyrroles, imidazoles, and radation could not be the major color- Esselen, and Fellers did show, however, pyridines, pyrazines. The of the melanoidins has producing reaction. Another argument that appreciable amounts of carbon chemistry been reviewed Enders is that some non-a-amino acids, which dioxide were liberated during the reac- by (41, 42), Enders and Theis Lüers supposedly are not active in Strecker- tion of glucose writh />-aminobenzoic acid (46), (113), type degradations, react more rapidly (108), as well as with some nitrogen-free Traitteur (183), Danehy and Pigman with reducing sugars to produce mel- organic acids (109). The participation (33), and Langer (96). anoidins than the corresponding a- of amines in the decarboxylation of The relationships between color pro- amino acids (65, 67, 71, 97, 174, 183). ascorbic acid and pyruvic acid has been duction and fluorescence were studied Therefore, it is concluded that the discussed by Euler and Hasselquist (51). recently in model systems (59, 133;

VOL. 1. NO. 15, OCTOBER 14, 1953 937 cf. 132). The results led to the con- have contained polymerized 3-carbon by sugar fission (D), and by the Strecker clusion that the fluorogens are pre- sugar fragments (which give no anthrone degradation (E), it is evident that aldol cursors of the brown pigments but are reaction); whereas those formed at condensation (F) is a highly probable not identical with them. In the presence higher temperatures contained poly- reaction for melanoidin formation. of bisulfite, color formation was in- merized 6-carbon residues from non- Browning reactions have been demon- hibited, yet the fluorogens accumulated. fragmented sugars. Wolfrom etal. (197) strated for pyruvaldehyde alone (36), The nonidentity of the fluorogens and obtained experimental data which they for furfurals alone (36, 153, 161, 170, the pigments was shown also by the interpreted as support for the latter case. 179, 198, 199), and between pyruvate isolation of colorless fluorescent inter- Inhibitors for browning have been and furfural (153). Sugars alone show mediates from a glucose-glycine system shown to be, chiefly, carbonyl reagents, less browning under equivalent condi- (737), from a glucose-a-X-acetyllysine such as cyanide (9, 25), Dimedon (42, tions (199). The aldol condensation of system (70, 77), and from potatoes 44, 89, 96, 155), hydroxylamine (9, 25), pyruvaldehyde has been postulated by (777). Although these compounds were hydrazines (9), mercaptans (67), and Enders and Sigurdsson (45) as the initial not identified, evidence was given that bisulfite (9, 47, 59, 76, 110, 130, 177, reaction in melanoidin formation. Since all three compounds were formed by 188). However, mercaptans and bi- aldol, CH3 · CHOH · CH2 · CHO, is sugar-amine condensation. The kinetics sulfites (which are the best of the above- known to undergo browning with of fluorescence production in glycose- listed inhibitors from a practical stand- glycine (67, 96), it is likely that nitrogen- glycine (141, 142) and in acetaldehyde- point) are also reducing agents; and, free aldols in general will react with ammonia (141) systems was reported by as such, they would keep reductones in amino compounds, aldimines, and ket- Pearce. the inactive reduced form, rather than imines to produce nitrogenous mela- Because of the similarity in composi- the active dehydro form. Other in- noidins, as indicated in Figure 1. tion and properties of the melanoidins hibiting factors, such as the lowering of The aldol condensation of diacetyl derived from the reaction of glycine pH, temperature, water content, and proceeds by first forming the ordinary with glucose, mannose, xylose, furfural, removal of reactants, have been re- aldol between two molecules, then or pyruvaldehyde, it has been assumed viewed by Olcott (130, 131). intramolecular condensation takes place that all were derived from similar inter- When a model system has browned to yield 2,5-dimethyl-/i-quinone. Car- mediates (39, 45, 179, 197). Traitteur only as far as the yellow-orange stage, son (26) recently investigated the reac- (183) has now reported significant the color can be appreciably diminished tion of diacetyl with cyclohexylamine at differences among furfural-glycine melan- by adding sulfhydryl or sulfite com- 0° C. He isolated crystalline diacetyl oidin (high content of ether-bound pounds; but, after further browning to dicyclohexylimine (X) and a highly oxygen), glucose-glycine melanoidin a dark brown color, very little diminu- reactive, colorless crystalline compound, (high content of alcoholic hydroxyl tion in color is obtained (67). Guss C20H32N2O, for which the structure groups), and pyruvaldehyde-glycine (67) also found that a few drops of ß- (XI) was proposed. The colorless melanoidin (high content of enolic mercaptoethanol (and several other intermediate (XI) hydroxyl groups and low content of mercaptans) would inhibit the browning CHS ether-bound oxygen). Glucose-glycine of glucose (3 grams) in aqueous solutions melanoidin was considered to occupy (5 ml.) containing glycine, lysine hydro- C:N C6Hn an intermediate position between the chloride, or sodium glutamate (0.5 two extremes represented by furfural gram), respectively, at 95° C. As mer- C:N.CeH„ and melanoidins. The did not inhibit the of pyruvaldehyde captans browning CH3 melanoidins derived from acetaldehyde- furfural-glycine, it appeared that the X glycine or pyruvic acid-glycine gave inhibition occurred before the dehydra- colors in incident ultraviolet which tion or closure to form furfurals. light ring H.,C are different from that of the first- The critical role of water in sugar- named group, indicating possible differ- amine browning reactions has been C=CH ences in structure (45). In unpublished investigated by Olcott and Dutton Z \ Lea and Hannan C(HnN :C C:N-C6Hu work, Carson and Olcott found that the (132), (70, 98, 100, \ / melanoidins which are produced from 101, 104, 105), Rooney (155), Mash- HC-C acetaldehyde and amino compounds kovtsev (118), Volgunov and Pokhno H Z\ HO are not acidic, contrary to the case with (185), and Pigman and Rosen (149). CH, sugar-amine melanoidins. As sugar- These studies confirmed a point well XI amine melanoidins contain reductone- known to the dehydrator—viz., brown- like reducing power (40, 42, 195), their ing of a thoroughly dehydrated food browns on exposure to moist air at room acidity may be caused chiefly by increases and passes through a maximum temperature and decomposes spon- enediolic hydroxyl groups. as the water content is increased. There taneously with color formation in many The naturally formed melanoidin is a striking resemblance in the shape of organic solvents. In methanol at 25° C., isolated from nonenzymatically browned the curves obtained by Rooney (155) on XI is dehydrated and yields 2,5-di- apricots by Weast and Mackinney (186) plotting browning against water content methyl-7?-quinone dicyclohexylimine, a was shown by Sattler and Zerban (160) for the model system, xylose-glycine, and yellow crystalline compound. Since not to contain 6-carbon sugar residues, the potato dehydration - browning -dicarbonyl compounds—e.g., diacetyl, as it gave a negative test for hydroxy- curves obtained by C. E. Hendel (re- osones, dehydro reductones—are gen- methyl furfural with the anthrone ported by Olcott, 130). erated in sugar-amine browning reactions, On the other the Several in- and ¡mines (which reagent. hand, F. Aldol independent quiñones quinone melanoidin Weast are so formed at low synthetic prepared by Condensation vestigations (24, 54, 94, readily tempera- and Mackinney (186) from the reaction 184) have shown that tures) could conceivably be involved in of fructose with asparagine at 100° C. amino compounds (particularly amine melanoidin formation. did contain a 6-carbon as it and According to skeleton, salts), including peptones egg G. Aldehyde-Amine a are for the yielded hydroxymethyl furfural and albumin, effective catalysts and Sprung (776), of and Polymerization alde- positive anthrone test. Sattler and aldol condensation acetaldehyde Formation of Hetero- simple Zerban that the melanoidins crotonaldehyde. Since amine catalysts hydes and suggested cyclic Nitrogen Com- which were formed at low are and aldehydes can be amines, even temperatures present, pounds (as in the sun-drying of apricots) may generated by sugar dehydration (C), acetaldehyde

938 AGRICULTURAL AND FOOD CHEMISTRY and ammonia, react readily at low found that the sodium salts of nitrogen- acids accelerated the discoloration of temperatures and form polymeric colored free organic acids cause a more or less reducing sugar solutions, but they did products the structures of which are typical browning reaction with sugars. so by acting synergistically rather than not known, a, ß-Unsaturated;aldehydes As in sugar-amine systems, the rate of separately. Malic acid gave more color react with amines at temperatures color formation is markedly increased than citric or tartaric acids in their above —10° C. to produce resins and by increased pH (addition of sodium systems. A similar study by Patron complex aldollike condensation products. hydroxide), carbon dioxide is liberated, (737) showed that the phosphate, iron, Patrick (734) showed that aliphatic the pH drops during the reaction, the and copper salts of a synthetic mixture aldehydes will condense with aldimines color development is inhibited by sulfur (representing the minerals in orange to form two different types of products: dioxide, and an oxygen atmosphere juice) also acted synergistically on the a dimeric, unsaturated, conjugated al- has only a slight enhancing effect on browning of glucose-organic acid and dimine—e.g., RCHaC :C(R')CH:NR"— the amount of color produced. The glucose-glycine-organic acid systems in and a substituted dihydropyridine. source of the carbon dioxide and the the presence of oxygen. The effect Both types of condensations could occur mechanism by which it is produced was much greater when glycine was also in melanoidin formation, as both remain obscure. present. It is apparent from these aldehydes and aldimines are present. It is apparent from Lewis’ data (707, two reports that the presence of small The formation of pyridines in this way 709) that sodium acetate, oxalate, amounts of minerals and organic acids might explain the fact that pyridines tartrate, lactate, and fumarate do not in natural systems may have important have been distilled from melanoidins produce browning as rapidly with effects on the browning reactions. (47, 46). Pyrroles, also, have been dis- glucose at 100° C., pH 7.2, as does The accelerated decomposition of tilled from glucose-glycine melanoidins glycine; on the other hand, sodium reducing sugars with color production in the presence of zinc dust (47, 46, 60, citrate causes even more browning by phosphates and other buffers (ace- 757). than glycine under these conditions. All tates, citrates, phthalates) is frequently Positive color reactions have been of these carboxylic acid salts accelerated mentioned in the literature (22, 38, obtained from model browning mixtures browning considerably over what is 47, 83, 96, 709, 729); but it is not with such nonspecific tests as the pine- known to occur with glucose alone at known with certainty which of these splinter (60, 757), Paulys (44), and Elson- 100° C., initial pH 7.2. However, in salts have specific action and which have Morgan (64, 65, 70, 77, 77), which the latter unbuffered case, the pH drops only the general accelerating action of indicated the presence of pyrroles, more and faster during the browning, bases. Phosphates have been reported imidazoles, or other nitrogenous hetero- thereby providing a self-inhibiting action to have a specific accelerating effect on cyclic compounds. Additional incon- on color production. The apparent sugar-amine condensation (7), sugar clusive evidence for the presence of accelerating effect of fatty acid salts fission (38, 67, 729), production of pyrroles in model browning systems was could possibly be caused by approximate reductones (22, 723), and browning (47, given by Pearce and Bryce (743). They maintenance of the initial pH. In such 745; cf. 83). Other reports (64, heated solutions of sugars (glucose, case the catalysis is only that of bases in 724; cf. 83) indicate that there is no mannose) and sugarlike substances general. The sugar-carboxylic acid specific acceleration of browning by (acetoin, diacetvl) with amino acids systems of Lewis were reinvestigated phosphate. Obviously, experiments ex- (glycine, glutamic acid) and showed for browning in acid solution by Patron pressly designed to settle the question that the fluorescence spectra of these (737). He concluded that temperatures are needed. Until the question of solutions were similar to the spectra of upwards of 100° C. were required, or specific vs. general base effect on the pyrroles. (Glycols, which do not un- very little browning would occur in the decomposition is settled for the various dergo browning with amino acids, also absence of amino compounds—for ex- buffer salts, stability studies on sugars, gave pyrrolelike fluorescence spectra ample, aqueous solutions of 0.5 M such as that of McDonald (774) (who when heated with amino acids.) On glucose which contained either 0.05 M used phosphate and glycinate buffers), the other hand, 4(5)-methylimidazole citric or malic acid gave virtually no cannot bear a general interpretation. (43, 78), tetraphenvlpyrrole (79), and browning after 10 days at 50° C. The main reactions that occur in several substituted pyrazines (64, 78, The browning of fructose in the pres- carbonyl-amino browning also occur 79), have been detected and actually ence of malic acid at 60° to 70° C. in in nonamino browning. For example, isolated from sugar-amine browning acid solutions was investigated by 1,2-enolization, sugar dehydration to reactions in model systems. Neverthe- Livingston (772). The browning reac- furfurals, and sugar fission are known less, the presence of such heterocyclics tion was more rapid than that of fructose caramelization reactions. Just as the in the melanoidins remains to be proved. with hydrochloric acid at the same pH. Amadori rearrangement (1,2-enolization Absorption spectra measurements and of glycosylamines) is now known to Browning in Nonamino Systems control experiments indicated that the provide a more labile configuration of the It is known that sugars, polysac- fructose was dehydrated to hydroxy- 1-amino sugar chain, so the Lobry de charides, polyhydroxycarboxylic acids, methyl furfural. [From the recently Bruyn-Alberda van Ekenstein rear- reductones, -dicarbonyl compounds, published results of Petuely (746), it rangement (1,2-enolization of re- and quiñones will undergo browning in can safely be said that the fructose was ducing sugars) provides the labile sugar the absence of amino compounds (cf. also 1,2-enolized in the acid solution.] chain for nonamino browning reactions. 68). Such decompositions, even in the Fructose, sucrose (cf. 745), hydroxy- Transformations of reducing sugars absence of catalysts, are of some im- methyl furfural, and ascorbic acid gave of the Lobry de Bruyn type that occur portance to the food chemist; but they browning in the presence of malic acid; in weak alkali are known to cause occur mainly at high temperatures not but glucose, furfural, pyruvic, and browning (62), and to occur in acid as often encountered in normal food proc- levulinic acids did not. By heating well as in alkaline media. Wolfrom essing. In most foods these nonamino various dicarboxylic acids with hydroxy- and Shilling (200) obtained 0.1% reactions would occur in the presence methyl furfural, it was shown that an glucose by column chromatography after of accelerators, such as carboxylic acids -hydroxyl group in the dicarboxylic refluxing an aqueous solution of fructose and their salts, phosphates, and metallic acid is required for the occurrence of (80%) at 113° for 16 hours (pH 6.9 — ions, and thereby they would approach browning. 2.7). The cooled solution was very (but not equal) the low energy require- The model system studies of Pederson, dark brown and had a strong burnt ments of sugar-amine reactions. Beattie, and Stotz (745) indicated that sugar odor. Petuely (746) and Berner Lewis, Esselen, and Fellers (709) organic acids, phosphates, and amino and Sandlie (79) also used chroma to-

VOL. 1, NO. 15, OCTOBER 1 4, 1 95 3 939 graphic methods to identify fructose in involves the dehydration of a poly- or other polyphenolic substrates to form heated, acidic glucose solutions. Petuely hydroxycarbonyl compound to produce quiñones. Catechols are enediols in detected very small amounts of mannose an -dicarbonyl compound and a a conjugated system and are classed as in the acidic glucose reaction mixture. saccharinic acid (acetic acid), thereby reductones (53). Although aromatic In addition, glucose was isolated and resembling sugar browning reactions. enediols are easily oxidized, nonaro- identified in the reaction mixture ob- It is concluded that organic acids and matic reductones are oxidized even tained by heating an aqueous solution their salts accelerate the caramelization more readily (5, 53). Some of the of fructose and tartaric acid in an auto- of sugars by promoting enolization of reactions of the quiñones after they are clave at 140° C. for 1.5 hours. Furfural the sugar. Enolization occurs in both formed may conceivably be enzyme- formation was a competing reaction, acid and alkaline media (746, 796), but catalyzed, but nonenzymatic browning strongly so below pH 3. Speck and far more easily with increasing alka- reactions are believed to predominate Forist (775) reported that both weak linity. The enolized sugar is more (82). Ascorbic acid is also a substrate organic acids and their salts are capable easily dehydrated and fragmented than for enzymatic oxidation. The dehydro- of catalyzing the Lobry de Bruyn- the initial ring form. The dehydration ascorbic acid formed would undergo Alberda van Ekenstein rearrangement and fragmentation products contain browning as described above. of glyceraldehyde and dihydroxyacetone. (as the most reactive intermediates) un- Little is known about the browning Simultaneous dehydration occurred to saturated osones and dehydro reductones mechanism for quiñones (82). It is yield pyruvaldehyde. The products of ( -dicarbonyl compounds) which so known, however, that quiñones (also this latter transformation, starting with readily form polymeric brown pigments. dehydroascorbic acid) will cause the the 3-C-phenyltrioses and acetic acid, Thus, the browning of nonamino and Strecker degradation (Reaction E) of were isolated and identified by Smith carbonyl-amino systems is similar. amino acids (763, 764). In addition, and Anderson (4, 777). There is one outstanding difference: quiñones (72), and probably also form- Following 1,2-enolization of the sugars, The acceleration of browning is appre- aldehyde (752), will oxidize ascorbic dehydration and fission occur. However, ciably greater through glycosylamine to dehydroascorbic acid, which will dehydration without fission is not known enolization (the Amadori rearrange- undergo browning of itself. Since p- in alkaline media. Furfural formation ment) than through enolization of the quinones can be formed in typical sugar- is strongest in strongly acid media unsubstituted sugar (Lobry de Bruyn- amine browning systems (see Reaction (below pH 3), but it occurs to a measur- Alberda van Ekenstein rearrangement). F), the reductones present could be the able extent at elevated temperatures .. . _ . The heating or dehydrogenated by quiñones; Nonammo Browning , ,. r even in the 6.0 to 6.7 (77, of thereby, browning of the reductones pH range Of Reductones long standing 758, 767, 770, 798-200). Sugar fission pure glucose could occur in the absence of oxygen. is known to occur to a limited extent in solutions produces reductones (7, 787). Conclusion weakly acidic media (37, 67, 729, 758), Furthermore, the reductones are un- and it increases rapidly with increasing stable and apparently undergo oxidative The several types of browning, the pH (37, 38, 788; cf. 7). If sulfite is degradation (787). Ascorbic acid alone many different reactants and reactions used to inhibit color development in in aqueous solution will undergo brown- involved in each type, and the large heated glucose solutions (alkaline pH), ing on heating at 98° C. or above (768, number of controlling variables make the yield of acetol obtainable from the 770), forming furfural and carbon the chemistry of browning reactions an distillates is significantly increased (788). dioxide (770, 753). The browning is intricate subject. It should be possible According to Berl and Feazel (78), increased with increasing pH, and above to analyze the browning complex by the ultraviolet absorption spectra of pH 7 autoxidation and color production identifying the individual reactions and sugars in alkali at pH 8 to 10 are simple occur readily at 25° C. (77, 737). studying each in model systems; yet, and correspond neither to pyruvalde- Neutral solutions of the reductones even with all the possible pathways to hyde, acetol, aldehydo sugars, nor derived from glycosylpiperidines (Sec- melanoidin formation outlined, it would exactly to triose-reductone. The spec- tion C) also gave color on standing in still remain to be determined which of trum is identical with that of triose- air at 25° C. Patron (735) found that these routes are operative and to what reductone, except the shift of the crude dihydroxymaleic acid turned extent one affects the other in a given maximum from 2950 to 2650 A. comes brown in the solid state and produced food (or other system) under a given at pH 7, rather than pH 5. [From an odor like maple sirup. Apparently set of conditions. Although we are recent polarographic measurements on all reductones, including the polyphenols, still in the first stages of solving the AEi/2 in alkaline methods and -— undergo strong darkening problem, chromatographic triose-reductone (23) a shift in solution in contact with air (707, 737, isotopic tracer techniques give a promis- was observed at pH 5.5 and also at 207). As in the case of the ascorbic ing outlook for an acceptable solution about pH 8.5. Moreover, the value acid-amino systems previously discussed in the future. As a step toward even- was the same above and below these (Section C), it is the dehydro form of tual solution, this review has discussed limits of pH.] The intensity and rate the enediol that is so unstable (50, 52, the reactions involved in browning, pro- of development of the spectrum are 737). According to the chromatographic posed a mechanism for browning in much greater for the trioses than for experiments of Euler and Hasselquist sugar-amine systems which correlates the the pentoses and hexoses, suggesting (52), dehydroascorbic acid is hydrolyzed known reactions, and pointed out im- that the chromogen is formed from the by acid and yields three different re- portant unexplored areas. latter through the condensation of 3- ductones. The promotion of color The mechanism for sugar-amine carbon or smaller intermediates. formation through the catalytic oxida- browning presented here (Figure 1) Like the sugars, polyhydroxycar- tion of reductones by copper (77, 745, contains several distinct routes to mela- boxylic acids will caramelize if heated 782) and the stoichiometric oxidation noidin formation, all of which stem to sufficiently high temperatures. with complex formation by trivalent from the Amadori rearrangement of Lamure (95) demonstrated that potas- iron and titanium compounds (5, 49, sugar-amine condensation products. sium tartrate undergoes browning and 50, 707, 738, 792) are also known In interpreting Figure 1 it should not be decomposes rapidly on heating the pure mechanisms for darkening in foods. concluded that all sugar-amine browning hemihydrate at 240° C. Water, carbon _ . Darkening of the surfaces passes through the Amadori rearrange- Enzymatic dioxide, potassium pyruvate, potassium of fresMy cut fruitg is ment (27, 75, 799); and the term rowning acetate, and carbon were identified as attributed to one or more Amadori rearrangement should not be products of the reaction. This pyrolysis oxidases which act on tyrosine, catechol, construed as meaning the complete

940 AGRICULTURAL AND FOOD CHEMISTRY transformation of an aldose to the 2- The value of this correlation of (8) Barker, N. G., and Cromwell, N. J. Am. Chem. ketose configuration. For example, 1,2- browning reactions will be directly H., Soc., 73, 1051-3 (1951). enolization of a gly cosy lamine, as postu- to the degree to which it proportional (9) Barnes, . M., and Kaufman, lated Gottschalk (63, comes within solutions to the of by 64), promotes problems C. W., Ind. Eng. Chem., 39, the as this two areas for the scope of term, enolization food technology. At least 1167-70 (1947). yields the tautomeric form of the ketose. practical applications are now apparent: (10) Barnes, . M., and Kennedy, A mechanism for C. Division of Chemis- browning providing For in which the Amadori A., Sugar for 1,2-enolization of the sugar radical systems try, Abstracts of Papers, 112th rearrangement of sugar-amine condensa- of a without Am. Chem. Soc., p. glycosylamine passing tion is the methods Meeting products key reaction, New York, N. Y., 1947. through the Amadori rearrangement— should be for the rear- IQ, sought blocking E. De Meto, R. H., i.e., by elimination of the amine radical No research directed (11) Barron, S., rangement. specifi- and Klemperer, F., J. Biol. to produce the unsubstituted sugar toward this end has yet been tried. cally Chem., 112, 625-40 (1936). enol—has been by Patron The multitudinous reactions which follow proffered (12) C. P., and Honeyman, J.. have the would then be elimi- Barry, (137). Others (174, 196) sug- rearrangement J. Chem. Soc., 1952, 4147-50. nated. gested 1,2-enolization of the sugar E. C., and Haw- The reaction of with (13) Bate-Smith, without sugar-amine condensation to glucose piperidine thorne, J. R., J. Soc. Chem. a new of stable the labile intermediate for both yields type crystalline T-297-302 (1945). provide reductone. No doubt antioxidants of the Ind., 64, (128, 146) and fission (14) Bayly, R. J., Bourne, E. J., and dehydration ascorbic acid type would find greater use of the It has been the Stacey, M., Nature, 169, 876-7 (174) sugar. in food preservation, if they could be made to show (1952). purpose of this article that in cheaply. In the piperidyl-reductone and (15) Bayne, S., and Holms, W. H., concentrated solutions of aldose sugars its there is the of a analogs possibility J. Chem. Soc., 1952, 3247-52. and amino and in the cheap, neutral antioxidant, which compounds, nearly relatively (16) Beacham, . H., and Dull, M. F., the could be tailored for various uses varia- dry state, experimental evidence by Food Research, 16, 439-45 tion of the amine or amino acid radical. favors the mechanism outlined in (1951). 1. condensation M. and Holmes, E., Figure Sugar-amine Dehydrators would think of browning (17) Bergdoll, S., and the Amadori are Ibid., 16, 50-6 (1951). rearrangement as an undesirable reaction; yet, in the key reactions to the 1,2-enol (18) Berl, W. G., and Feazel, C. E., leading many instances a particular food—e.g., and, from it, the accelerated production Division of Sugar Chemistry, coffee, fresh bread crust, maple sirup, Abstracts of 122nd of brown pigments occurs Papers, through roasted nuts—could not be identified Am. Chem. Soc., several different routes. Meeting p. were it not for its characteristic desirable Atlantic N. 1952. One feature of the 5-6R, City, J., striking proposed “browned” flavor and odor. The con- (19) Berner, E., and Sandlie, S., mechanism 1) is the extent to (Figure trol of browning reactions to produce Chemistry & Industry. 1952, 1221- which addition reactions, con- simple only wanted flavors and odors is an 2. densations, enolizations, 209-10 dehydrogena- intriguing possibility. Control of brown- (20) Block, A., Science, 102, tions, and rearrangements are (1945). repre- ing to do man’s will is the ultimate goal sented. All of these are considered to (21) Boretti, G., Chimica e industria of browning research, but progress be proton-transfer reactions. Such an (Milan), 26, 31-3 (1944). toward this goal can be made only as and H.. explanation is required by the known (22) Borsook, PL, Wasteneys, the reaction mechanisms are better Biochem. J., 19, 1128-37 (1925). independence of sugar-amine browning understood. R., and Zuman, P., from oxidation. This is (23) Brdicka, atmospheric Collection Czechoslov. Chem. Com- not to that oxidation say, however, by Acknowledgment muns., 15, 766-79 (1950). the atmosphere plays no part in sugar- The and (24) Budnitskaya, E. V., Biokhimiya. amine browning. In connection with guidance helpful suggestions 146-54 of C. E. Rist, head of the Starch and 6, (1941). the browning of reductones, some type W. G., J. Soc. Chem. Ind.. Dextrose Division of this laboratory, are (25) Burton, cf oxidation is essential, but need not 64, 215-18 (1945). gratefully acknowledged. Many of be provided by the atmosphere. Air- Carson, F., Division of the here were (26) J. Organic oxidation might also diminish the extent concepts presented gained Chemistry, Abstracts of Papers, as a result of the free of un- of color formation by exchange 123rd Meeting Am. Chem. Soc. converting very information the reactive compounds—e.g., pyruvalde- published among p. 4M, Los Angeles, Calif., author’s in the four J. Am. Chem. Soc.. hyde—into less active compounds— colleagues Regional 1953; 75, Research Laboratories of the U. S. 4337 e.g., pyruvic acid. Of the three main 4300, (1953). of and the L. and routes to melanoidins outlined in the Department Agriculture (27) Cavalieri, F., Wolfrom, research contractors to the M. L., Ibid., 68, 2022-5 (1946). intermediate stage, sugar dehydration Quarter- master Food and Container Institute (28) Chichester, C. O., Stadtman, F. and sugar fission are of prime im- of the Armed Forces. H., and Mackinney, G., Ibid.. amino acid is portance; degradation 74, 3418-20 (1952); private and secondary. Doubtless, dehydration Literature Cited communication. fission hand in hand. Whether one go (29) Cocker, W., O’Meara, R. A. Q., 321-8 precedes the other is not known. The (1) Agren, G., Enzymologia, 9, Schwartz, J. C. P., and Stuart. relative importance of each may vary (1941). E. R., J. Chem. Soc.. 1950, 143-50 with the controlling conditions. (2) Akabori, S., Ber., 66, 2052-8. (1933). The formation of reductones, a- (30) Coulter, S. T., Jenness, R., and (3) Ambler, J. A., Ind. Eng. Chem.. dicarbonyl, and conjugated dicarbonyl Geddes, W. F., in “Advances in 21, 47-50 (1929). Vol. compounds (dehydro reductones) has Food Research,” 3, pp. (4) Anderson, R. H., “Reactions of New Academic been encountered repeatedly in this Amines with 97-105, York, Polyhydroxycar- Press, 1951. correlation. Yet, very little study has bonyl Compounds,” Ph.D. A. L., Food Research, 13, been to the reactions thesis, University of Minne- (31) Curl, given browning 381-6 (1948). of these compounds in model systems. sota, 1950. and (32) Ibid., 14, 9-14 (1949). More research on the decomposition (5) Arndt, F., Loewe, L., Ayca, 333-42 (33) Danehy, J. P., and Pigman. W. and polymerization reactions of pertinent E., Ber., 84, (1951). (6) C., J. Chem. Soc. W., “Advances in Food Re- should estab- Ashikaga, Japan, dicarbonyl compounds help 64, 639-44 (1943). search,” Vol. 3, pp. 241-90, lish the mode of melanoidin formation (7) Baraud, J., Bull. soc. chim., 1953, New York, Academic Press, for many of the browning pathways. 521-5. 1951.

VOL. 1, NO. 15, OCTOBER 1 4, 1953 94] A. P. and Research and De- O. (34) Dunlop, P., Stout, R., cago, 111., L, Biokhimiya, 6, 113-21 ' Swadesh, S., Ind. Eng. Chem., velopment Associates, Food and (1941). 38, 705-8 (1946). Container Institute, Inc., 1952. (95) Lamure, J., Compt, rend., 230, (35) Edwards, B. G;, and Dutton, (68) Haas, V. A., and Stadtman, E. R., 302-3.(1950). H. J„ Ibid., 37, 1121-2 (1945). Ind. Eng. Chem., 41, 983—5 (96) Danger, A. W., Jr., “The Maillard (36) Enders, C.. Biochem. Z.. 312, 339— (1949). Reaction,” Ph.D. thesis, Ohio 53 (1942). (69) Haas, V. A., Stadtman, E. R., State University, 1951. (37) Ibid., pp. 349-53. Stadtman, F. H., and Mackin- (97) Lea, C. H., Chemistry & Industry, (38) Ibid., 313, 265-9 (1943). ney, G., J. Am. Chem, Soc., 70, 1950, 155-8. (39) Ibid., pp. 352-71. 3576-9 (1948). (98) Lea, C. H., in “Contributions of (40) Ibid., 314, 389-98 (1943). (70) Hannan, R. S., and Lea, C. H., Browning Research to Ration 141) Enders, C., Kolloid-Z., 85, 74-87 Biochim. et Biophys. Acta, 9, Item Stability,” pp. 7-19, Chi- (1938). ,-,r. 293-305 (1952). cago, 111,, Research and De- (42) Enders, C., Wochenschr.furBrauerei, (71) Hannan, R. S., and Lea, C. H., velopment Associates, Food and 60, 98-104 (1943). Nature, 168, 744-5 (1951). Container Institute, Inc., 1952.’ (43) Enders, C., Fischler, F., and (72) Hasselquist, H., Arkiv Kemi, 4, (99) Lea, C. H., J. Dairy Research, Zellweger, R., Brauwissenschaft, 369-76 (1952). 15, 369-76 (1948). 1948, 85-7. (73) Haugaard, G., Tumerman, L., (100) Lea, C. H., and Flannan, R. S., (44) Enders, G., and Marquardt, R., and Silvestri, H., J. Am. Chem. Biochim. et Biophys. Acta, 3; Naturwissenschafteri, 29, 46-7 Soc., 73, 4594-600 (1951). 313-25 (1949). (1941). (74) Henry, K. M., and Kon, S. K., (101) Ibid., 4, 518-31 (1949). (45) Enders, C., and Sigurdsson, S., Biochim. et Biophys. Acta, 5, (102) Ibid., 5, 433-54 (1950). C. and R. Ber., 76, 560-3 (1943). 455-6 (1950). (103) Lea,' H., Hannan, S., (46) Enders, C., and Theis, K., Brenn- E., and Rist, C. E., Nature, 165, 438 (1950). ' (75) Hodge, J. staff-Chem., 19, 402-49 (1938). J. Am. Chem. Soc.. 74, 1494-7 (104) Lea, C. H., Hannan, R. S., and (47) Englis, D. T., and Hanahan, D. J., (1952). Rhodes. D. N., Biochim. et J. Am. Chem. Soc.. 67, 51-4 (76) Ibid., 75, 316-22 (1953). Biophys. Acta, 7, 366-77 (1951). (1945). (77) Horowitz, . N., Ikawa, M., and (105) Lea, C. H.. and Rhodes, D. N.; (48) Erickson, J. G., Ibid., 75, 2784 Fling, M., Arch. Biochem.. 25, Ibid., 9, 56-60 (1952). (1953). 226-9 (1950). (106) Lea, C. H., Rhodes, D. N., and (49) Ettori, J., Compt. rend., 202, (78) Hough, L., Jones, J. K. N., and Borrell, S., Nature, 169, 1097-8 852-4 (1936). Richards, E. L.. J. Chem. Soc.. (1952).' (50) Euler, H. v., and Hasselquist, H., 1952, 3854-7. (107) Lewis, V. M., “Studies on the Arkiv Kemi, 4, 169-74 (1952). (79) Hurd, C. D., in “Contributions of Non-Enzymatic Browning of (51) Ibid., pp. 476-85. Browning Research to Ration Foodstuffs,” Ph.D. thesis, Uni- (52) Ibid., 5, 49-55 (1953). Item Stability,” pp. 21-3, Chi- versity of Massachusetts, 1948. (53) Euler, H. v., and Hasselquist, H., cago, 111., Research and De- (108) Lewis, V. M., Esselen, W. B., “Reduktone,” Vol. 50 in velopment Associates, Food and Jr., and Fellers, C. R.. Ind. “Sammlung chemischer und Container Institute, Inc., 1952. Eng. Chem., 41, 2587-91 (1949). chemisch - technischer Vor- (80) Hurd, C. D., and Kelso, C. D., (109) Ibid., pp. 2591-4. trage,” Stuttgart, Germany, J. Am. Chem. Soc.. 70, 1484-6 (110) Lewis, V. M., and Lea, C. H., Ferdinand Enke, 1950. (1948). Biochim. et Biophys. Acta, 4, (54) Fischer, F. G., and Marschall, (81) Ingraham. L. L.. and Corse, J., 532-4 (1950). A., Ber., 64, 2825-7 (1931). Ibid., 73, 5550-3 (1951). (111) Lewis, W. R., and Doty, D. M„ (55) Foley, W. M., Jr., Sanford, G. E., (82) Joslyn, M. A., and Pon ting, J. Am. Chem. Soc., 69, 521-3 and McKennis, H., Jr., J. Am. J. D., in “Advances in Food (1947). Chem. Soc., 74, 5489-91 (1952). Research,” Vol. 3, pp. 1-44, (112) Livingston, G. E., J. Am. Chem. (56) Fraenkel-Conrat, J., Cook, B. B., New York, Academic Press, Soc., 75, 1342-4 (1953). and Morgan, A. F., Arch. Bio- 1951. (113) Lüers, H., Brewers Digest, 24(10), chem. Biophys., 35, 157-70 (83) Kass, J. P., and Palmer, L. S., 41-4, 48 (1949). (1952). Ind. Eng. Chem., 32, 1360-6 (114) McDonald, E. J., J. Research (57) Frankel, M., and Katchalsky, A., (1940). Natl. Bur. Standards, 45, 200-3 Biochem. J., 31, 1595-1604 (84) Katchalsky. A.. Biochem. J., 35, (1950). (1937); 32, 1904-7 (1938). 1024-7 (1941). (115) Mackinney, G., in “Contributions (58) Ibid., 35, 1028-38 (1941). (85) Katchalsky, A., and Sharon, N., of Browning Research to Ration (59) Friedman, L., and Kline, O. L., Biochim. et Biophys. Acta, 10, Item Stability,” p. 21, including J. Biol. Chem., 184, 599-606 290-301 (1953). footnote, Chicago, 111., Research (1950) . (86) Kertesz, Z. L, “The Pectic Sub- and Development Associates, (60) Fuhlrott, R., Ph.D. dissertation, stances,” pp. 16, 151-3, New Food and Container Institute, Technische Hochschule, Mün- York. Interscience Publishers, Inc., 1952. chen, 1945; cited by Danger 1951. (116) Maillard, L. C., Compt. rend., 154, in (96). (87) Kline, R. W., and Stewart, G. F., 66-8 (1912). (61) Goto, R., J. Chem. Soc. Japan, Ind. Eng. Chem., 40, 919-22 (117) Martin. L. F., Division of Carbo- 64, 999-1007. 1054-63, 1183-5 (1948). hydrate Chemistry, Symposium (1943). (88) Koppanyi, T., Vivino, A. E., on Use of Sugars and Other (62) Gottfried, J. B., and Benjamin, and Veitch, F. P., Jr., Science. Carbohydrates in the Food D. G., Ind. Eng. Chem., 44, 101, 541-2 (1945). Industry, Abstracts of Papers, 141-5 (1952). (89) Kretovich, V. L., and Tokareva, 123rd Meeting Am. Chem. Soc., (63) Gottschalk, A., Biochem. J., 52, R., Biokhimiya, 13, 508—15 Los Angeles, Calif., p. 7D, 1953. 455-60 (1952). (1948). (118) Mashkovtsev, M. F., Biokhimiya, (64) Gottschalk, A., Nature, 167, 845-7 (90) Krijt, G., Arch, neerland. physiol., 16, 615-20 (1951). and (1951) . 26, 457-71 (1942). (119) Micheel, F., Klemer, A., (65) Gottschalk, A., and Partridge, (91) Kubota. T., J. Biochem. (Japan), Chem. Ber., 1083-6 (1952). S. M., Ibid., 165, 684 (1950). 34, 119-41 (1941). (120) Miller, M. C., Food Research, 12, (66) Graham, N. D., Hsu, P. Y., and (92) Kuhn, R., Badstüímer, W., and 343-59 (1947). McGinnis, J., Science, 110,217-18 Grundmann, C., Ber., 69, 98- (121) Miller, R. E., and Cantor, S. M., (1949). 107 (1936). J. Am. Chem. Soc., 74, 5236-7 (67) Guss, C. O., in “Contribution of (93) Kuhn, R., and Weygand, F., (1952). Browning Research to Ration Ibid., 70, 769-72 (1937). (122) Mitts, E., and Hixon, R. M., Ibid., Item Stability,” pp. 26-9, Chi- (94) Kuzin, A. M., and Polyakova, 66, 483-6 (1944).

942 AGRICULTURAL AND FOOD CHEMISTRY (123) Miyake, S., Hayashi, K., and (149) Pigman, W., and Rosen, L., Division of Sugar Chemistry, Kurasawa, F., J. Soc. Trop. Division of Sugar Chemistry, Abstracts of Papers, 122nd Agr., Taihoku Imp. Univ., 14, Abstracts of Papers, 122nd Meeting Am. Chem. Soc., Atlan- 298-303 (1942). Meeting Am. Chem. Soc., At- tic City, N. J., p. 9R, 1952. (124) Mohammad, A., Fraenkel-Conrat, lantic City, N. J., p. 9R, 1952; (176) Sprung, . M., Chem. Revs., 26, H., and Olcott, H. S., Arch. J. Am. Chem. Soc., 75, 3460-4 297-338 (1940). Biochem., 24, 157-78 (1949). (1953). (177) Stadtman, E. R., “Advances in (125) Mohammad, A., and Olcott, H. (150) Pope, C. G., and Stevens, M. F.. Food Research,” Vol. 1, pp. S., J. Am. Chem. Soc., 69, 969 Biochem. J., 33, 1070-7 (1939)’. 325-72, New York, Academic (1947) . (151) Raacke-Fels, I. D., Arch. Biochem. Press, Í948. (126) Mohammad, A., Olcott, H. S., Biophys., 43, 289-98 (1953). (178) Stadtman, F. H., Chichester, C. and Fraenkel-Conrat, H., Arch. (152) Reithel, F. J.. and Wither, R. P., O., and Mackinnev, G., J. Am. Biochem., 24,270-80 (1949). J. Am. Chem. Soc., 71, 1879-80 Chem. Soc.. 74, 3194-6 (1952). (127) Moster, J. B., and Chapman, (1949). (179) Taufel, K., and Iwainsky, H., R. A., Can. J. Research, 27F, (153) Rice, R. G., Kertesz. Z. L, and Biochem. Z., 323, 299-308 429-34 (1949). Stotz. E. H., Ibid'., 69, 1798- (1952) . (128) Newth, F. H., “Advances in 1800 (1947). (180) Tan, T.-L., Wolfrom, M. L., and Carbohydrate Chemistry,” Vol. (154) Roe, J. H., Science, 80, 561 (1934). Langer, A. W., Jr., Ibid., 72, 6, pp. 83-106, New York, (155) Rooney, C. S., in “Contributions 5090-5 (1950). N. Y., Academic Press, 1951. of Browning Research to Ration (181) Tarr, H. L. A., Nature, 171, 344-5 (129) Nodzu, R., Bull. Chem. Soc. Japan, Item Stability,” p. 41, Chicago, (1953) . 10, 122-30 (1935). 111., Research and Development (182) Thompson, J. B., Kocher, R. B., (130) Olcott, H. S., “Chemistry of Associates, Food and Container and Fritzsche, H. W., Arch. Dehydrated Vegetables,” Divi- Institute, Inc., August, 1952. Biochem., 18,41-9 (1948). sion of Agricultural and Food (156) Ross, A. F., in “Advances in Food (183) Traitteur, H., Brauwissenschaft, 9, Chemistry, Symposium on De- Research,” Vol. 1, pp. 257-290, 153-8 (1951); hydrated Foods, 123rd Meeting New York, Academic Press, (184) Triebs, W., and Krumbholz, K., Am. Chem. Soc., Los Angeles, 1948. Chem. Ber., 85, 1116-19 (1952). 1953. (157) Ruckdeschel, W., Z. ges. Brauw., (185) Volgunov, G. P., and Pokhno, (131) Olcott, H. S., in “Contributions of 37, 430-2, 437-40 (1914). . T., Biokhimiya, 15, 67-74 Browning Research to Ration (158) Sattler, L., and Zerban, F. W., (1950). Item Chi- Ind. 1401-5 C. and Stability,” pp. 51-6, Eng.’ Chem., 41, (186) Weast, A., Mackinney, cago, 111., Research and De- (1949). G., Ind. Eng. Chem., 33, 1408-12 velopment Associates, Food and (159) Sattler. L., and Zerban, F. W., (1941). Container Institute, Inc., 1952.: J. Am. Chem. Soc.. 70, 1975-6 (187) Wendland, G., Arch. Pharm.. 285, (132) Olcott, H. S., and Dutton, H. J., (1948). 71-9, 109-120 (1952). Ind. Eng. Chem., 37, 1119-21 (160) Sattler, L., and Zerban, F. W., (188) Werle, E., and Stiess, P., Biochem. (1945). Science, 108, 207 (1948). Z., 321, 485-91 (1951). (133) Overby, L. R., and Frost, D. V7., (161) Scallet, B. L., and Gardner, J. H., (189) Weygand, F., Arkiv Kemi. 3, J. Nutrition, 46, 539-50 (1952). J. Am. Chem. Soc., 67, 1934-5 11-16 (1951). (134) Patrick, T. M., Jr., J. Am. Chem. (1945). (190) Wevgand, F., Ber., 73, 1259-78 Soc., 74, 2984-6 (1952). (162) Schenk. G. O.. Ber., 77, 661-8 (1940). (135) Patron, A., Fruits

(143) Pearce, J. A., and Bryce.' W. A., Soc., 70, 517—22 (1948). and Cavalieri, L. F.. Ibid.. Ibid., 1, 210-19 (1947). (171) Smith. L. L, and Anderson, R. H., 70, 514-17 (1948). (144) Pecherer, B., J. Am. Chem. Soc., J. Org. Chem., 16, 963-71 (199) Ibid., 71, 3518-23 (1949). 73, 3827-30 (1951). (1951) . (200) Wolfrom. M. L., and Shilling, (145) Pederson, C. S., Beattie. H. G., (172) Ibid., pp. 972-80. W. L., Ibid., 73, 3557-8 (1951). and Stotz, E. ., . Y. (Geneva) (173) Sowden, J. C., and Schaffer, R., (201) Zerban, F. W., “Color Problem Agr. Expt. Sta. Bull.. 728, 3-32 J. Am. Chem. Soc., 74, 499-504 in Sucrose Manufacture,” Tech- (1947). (1952) . nological Report Series No. 2, (146) Petuely, F., Monatsh.. 24, 298-312 (174) Speck, J. C., Jr., in “Contribu- pp. 1-31, Sugar Research Foun- (1953). tions of Browning Research to dation, Inc., 1947. (147) Petuely, F., and Künssberg, U., Ration Item Stability,” pp. Received review 1953. Ibid., 83, 80-94 (1952). 29-36, Chicago, 111., Research for August 3, Accepted (148) Pigman, W., Cleveland, E. A., and September 17, 1953. Presented before the Development Associates, Division and Food Couch, D. H., and Cleveland, Food and Container of Agricultural Chemistry, Institute, Symposium on Dehydrated Foods, at the 123rd J. Am. Chem. Soc.. 1952. J. H., 73, Inc.. Meeting of the American Chemical Society, 1976-9 (1951). (175) Speck, J. C., Jr., and Forist. A. A., Los Angeles, Calif.

VOL. 1, NO. 15, OCTOBER 14, 1953