Proc. Nati. Acad. Sci. USA Vol. 80, pp. 901-905, Februarv 1983 Biochemistry Enzyme relationships in a sorbitol pathway that bypasses glycolysis and pentose phosphates in glucose metabolism (isozymes/diabetes/alcohol dehydrogenase/reductases/alcoholism) JONATHAN JEFFERY* AND HANS JORNVALLt *Department of Biochemistry, University ofAberdeen, Marischal College, Aberdeen, AB9 lAS Scotland, United Kingdom; and tDepartment of Chemistry I, Karolinska Institutet, S-104 01 Stockholm 60, Sweden Communicated by Sune Bergstr6m, September 17, 1982 ABSTRACT A pathway from glucose via sorbitol bypasses the fructose (8). Increased operation ofthe glycerol phosphate shut- control points of hexokinase and phosphofructokinase in glucose tle could also mediate the effect (9). metabolism. It also may produce glycerol, linking the bypass to The structural relationships, the diverse and unrelated met- lipid synthesis. Utilization of this bypass is favored by a plentiful abolic suggestions, the apparent associations of some interme- supply of glucose-hence, conditions under which glycolysis also diate compounds, and the possible convergent relationships is active. The bypass further involves oxidation of NADPH, so the motivate a search for unifying explanations, detailing the sor- pentose phosphate pathway and the bypass are mutually facilita- bitol pathway and its consequences. tive. Possible consequences in different organs under normal and pathological, especially diabetic, conditions are detailed. Enzymes with related structures (for example, sorbitol dehydrogenase and MATERIALS AND METHODS alcohol dehydrogenase, and possibly, aldehyde reductase and al- Glucose Metabolism via a Five-Step Sorbitol Pathway By- dose reductase, respectively) are linked functionally by this passing the scheme. Some enzymes of the bypass also feature in glycolysis Regulatory Steps of Glycolysis. In glucose metab- (aldolase and alcohol dehydrogenase), and these enzymes, with the olism, flux through the glycolytic and pentose phosphate path- reductases involved, are proteins known to occur in different ways is regulated at phosphofructokinase (10) and glucose-6- classes or multiple isozyme forms. Two of the enzymes (aldolase phosphate dehydrogenase (11). Hexokinase is generally also a and alcohol dehydrogenase) both involve classes with and without control point, though the effect in liver is overcome at high a catalytic metal (zinc). The existence ofparallel pathways and the glucose concentrations by glucokinase. However, as shown in occurrence of similar enzymic steps in one pathway may help to Fig. 1, a complete pathway similar to glycolysis but fully by- explain the abundance and multiplicity ofenzymes such as reduc- passing the hexokinase and phosphofructokinase steps also is tases, aldolases, and alcohol dehydrogenases. possible. It involves oxidation of one or two molecules of NADPH, thus favoring operation of the pentose phosphate Although alcohol dehydrogenase is widespread in nature, is pathway, and is linked to lipid synthesis via glycerol formation common in liver, and has been structurally characterized from (Fig. 1). several sources (1), its exact role in mammalian organs generally The first step of the pathway is conversion of glucose into has remained unclear. In addition to ethanol oxidation, func- sorbitol by aldose reductase and NADPH. Oxidation ofsorbitol tions in bile acid formation, fatty acid degradation, vitamin A by sorbitol dehydrogenase and NAD' then yields fructose. metabolism, a hydroxysteroid reaction, and other areas of me- Fructose is not a substrate for glucokinase, but in some circum- tabolism have been considered (2). Different and multiple func- stances hexokinase and ATP could convert it into fructose 6- tions of alcohol dehydrogenase would be consistent with the phosphate, leading into glycolysis at the phosphofructokinase considerable species variations and extensive evolutionary reaction. This loop via sorbitol then would serve as a transhy- changes found. drogenase function (reduction of NAD+ and oxidation of Recently, another common liver enzyme, sorbitol dehydro- NADPH). Alternatively, fructokinase and ATP could convert genase, was shown to be structurally, mechanistically, and an- fructose into fructose 1-phosphate. This compound can be cestrally related to alcohol dehydrogenase (3). The substrates of cleaved by aldolase B to give dihydroxyacetone phosphate and sorbitol dehydrogenase, fructose and sorbitol, are considered glyceraldehyde. Glyceraldehyde 3-phosphate results from the important in special organs, such as male sexual organs (4), or action oftriosephosphate isomerase on dihydroxyacetone phos- special disease states, such as hyperglycemic cataract formation phate or oftriokinase on glyceraldehyde. Entry at this stage into (5) and possibly diabetic neuropathy (6) and glomerulosclerosis glycolysis via glyceraldehyde 3-phosphate bypasses the hexo- (7), but a more general metabolic significance that is compatible kinase and phosphofructokinase reactions. Conversion of glyc- with the structural similarity of these enzymes has not been clear. eraldehyde to glycerol is an alternative to phosphorylation. The The structural link between alcohol and polyol dehydroge- reduction can be effected by alcohol dehydrogenase and NADH nases showed a type ofparallel or convergent evolution ofa sec- or by aldehyde reductase and NADPH. ond, different enzyme in each case with related specificity (3). Occurrence of the individual enzymes of the pathway is de- Enhancement of alcohol metabolism by fructose gave one as- scribed in relation to the organs mentioned below. The initial sociation between the substrates for sorbitol and alcohol dehy- steps of the pathway, with the glucose-fructose conversion drogenases. This effect may stem from avoidance of the rate- (sometimes also called the "sorbitol pathway"), have been dis- limiting NADH dissociation off alcohol dehydrogenase by an cussed further in relation to diabetes or insulin effects (12), but in situ coenzyme reoxidation with glyceraldehyde formed from usually not the associated, later combination into glycolysis or lipogenesis. The publication costs ofthis article were defrayed in part by page charge In summary, the essential steps providing the bypass are payment. This article must therefore be hereby marked "advertise- those leading via sorbitol and fructose 1-phosphate to glycer- ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. aldehyde 3-phosphate and linking, among other enzymes, sor- 901 Downloaded by guest on September 23, 2021 902 Biochemistry: Jeffery and Jbrnvall Proc. Natl. Acad. Sci. USA 80 (1983) bitol dehydrogenase and alcohol dehydrogenase with various ofthe aldose reductase type (that is, NADPH-linked aldehyde reductases and aldolases. reductases) have little (15) activity towards glucose, and much hepatic sorbitol dehydrogenase is present. These aspects sug- RESULTS AND DISCUSSION gest that liver may not be the main functional site ofthe entire Normal Occurrence. The scheme in Fig. 1 and the enzymes bypass for energy production but rather for permitting metab- shown demonstrate that the bypass or parts ofit may be offunc- olism of important compounds in any ratios. tional significance. In this respect, five tissues or organs appear Pancreas. Sorbitol is present in pancreas and glucose-induced of special interest in normal conditions. They are, on the one insulin release evidently requires both NADPH and aldose re- hand, the liver, pancreas, and placenta, where some com- ductase activity within the f3 cells (12). Reduction ofglucose to pounds ofthe scheme may have special roles, and, on the other sorbitol possibly primes or activates an element in the insulin hand, the brain and the reproductive system, where the bypass release mechanism (12). Therefore, it appears possible that in appears capable of having a more general metabolic role. the pancreas, the initial steps ofthe scheme in Fig. 1 may have Liver. As the main processorofingestedcompounds, the liver special functions. As in the liver, they may not be ofenergetic might be expected to encounter sorbitol. However, orally in- value in an entire bypass ofglycolysis but serve other functions, gested sorbitol is poorly absorbed in man and animals, and prod- which in the case of pancreas may be metabolic regulation via ucts ofits metabolism by intestinal microorganisms, rather than insulin release. sorbitol itself, are absorbed (13, 14). Nevertheless, sorbitol en- Placenta. Glucose is converted into fructose by the aldose tering the livercould be oxidized byhepatic sorbitol dehydroge- reductase and sorbitol dehydrogenase of normal placenta (16) nase in the same way that ethanol is oxidized by alcohol dehy- and umbilical cord (17) in species including man. High concen- drogenase. trations offructose in the fetal blood ofungulates (for example, Most important, liver deals with high postprandial concen- sheep and pig) apparently result from nonutilization offructose trations of glucose. Accumulation of sorbitol formed from glu- by their fetuses (18). Because fructose is not an important fetal cose within the liver then would be disadvantageous, raising the energy source in these cases, the loop from glucose to fructose osmotic pressure of the cells. Significantly, the liver enzymes may have an osmoregulatory role or serve a transhydrogenase Glucose aldose hexokinase (Iglucokinase) Glucose-6-P sorbi tol dehydrogenase Fructose I11 Fructose-6-P phospho- fructokinase fructokinase Fructose-l-P Fructose-1,6-bisP