Gluconeogenic pathway in liver and muscle glycogen synthesis after exercise

To determine whether prior exercise affects the pathways of liver and muscle glycogen synthesis, rested and postexercised rats fasted for 24 h were infused with glucose (200 mumol.min-1.kg-1 iv) containing [6-3H]glucose. Hyperglycemia was exaggerated in postexercised rats, but blood lactate levels were lower than in nonexercised rats. The percent of hepatic glycogen synthesized from the indirect pathway (via gluconeogenesis) did not differ between exercised (39%) and nonexercised (36%) rats. In red muscle, glycogen was synthesized entirely by the direct pathway (uptake and phosphorylation of plasma glucose) in both groups. However, only approximately 50% of glycogen was formed via the direct pathway in white muscle of exercised and nonexercised rats. Therefore prior exercise did not alter the pathways of tissue glycogen synthesis. To further study the incorporation of gluconeogenic precursors into muscle glycogen, exercised rats were infused with either saline, lactate (100 mumol.min-1.kg-1), or glucose (200 mumol.min-1.kg-1), containing [6-3H]glucose and [14C(U)]lactate. Plasma glucose was elevated one- to twofold and three- to fourfold by lactate and glucose infusion, respectively. Plasma lactate levels were elevated by about threefold during both glucose and lactate infusion. Glycogen was partially synthesized via an indirect pathway in white muscle and liver of glucose- or lactate-infused rats but not in saline-infused animals. Thus participation of an indirect pathway in white skeletal muscle glycogen synthesis required prolonged elevation of plasma lactate levels produced by nutritive support.     

Quantitative estimation of the pathways followed in the conversion to glycogen of glucose administered to the fasted rat

When [6-3H,6-14C]glucose was given in glucose loads to fasted rats, the average 3H/14C ratios in the glycogens deposited in their livers, relative to that in the glucoses administered, were 0.85 and 0.88. When [3-3H,3-14C]lactate was given in trace quantity along with unlabeled glucose loads, the average 3H/14C ratio in the glycogens deposited was 0.08. This indicates that a major fraction of the carbons of the glucose loads was converted to liver glycogen without first being converted to lactate. When [3-3H,6-14C]glucose was given in glucose loads, the 3H/14C ratios in the glycogens deposited averaged 0.44. This indicates that a significant amount of H bound to carbon 3, but not carbon 6, of glucose is removed within liver in the conversion of the carbons of the glucose to glycogen. This can occur in the pentose cycle and by cycling of glucose-6-P via triose phosphates: glucose----glucose-6-P----triose phosphates----glucose-6-P----glycogen. The contributions of these pathways were estimated by giving glucose loads labeled with [1-14C]glucose, [2-14C]glucose, [5-14C]glucose, and [6-14C]glucose and degrading the glucoses obtained by hydrolyzing the glycogens that deposited. Only a few per cent of the glucose carbons deposited in glycogen were deposited in liver via glucose-6-P conversion to triose phosphates. Between 4 and 9% of the glucose utilized by the liver was utilized in the pentose cycle. While these are relatively small percentages, since three NADP3H molecules are formed from each molecule of [3-3H]glucose-6-P utilized in the cycle, a major portion of the difference between the ratios obtained with [3-3H]glucose and with [6-3H]glucose is attributable to metabolism in the pentose cycle. Because 3H of [3-3H]glucose is extensively removed during the conversion of the glucose to glycogen within liver the extent of incorporation of the 3H into liver glycogen is not the measure of glucose's metabolism in other tissues before its carbons are deposited in liver glycogen. The distributions of 14C from the 14C-labeled glucoses into the carbons of the liver glycogens mean that at a minimum about 30% of the carbons of the glucose deposited in the glycogen were first converted to lactate or its metabolic equivalent.     

SERGE, the subcellular site of initial hepatic glycogen deposition in the rat: a radioautographic and cytochemical study

Hormonal control of hepatic glycogen and blood glucose levels is one of the major homeostatic mechanisms in mammals: glycogen is synthesized when portal glucose concentration is sufficiently elevated and degraded when glucose levels are low. We have studied initial events of hepatic glycogen synthesis by injecting the synthetic glucocorticoid dexamethasone (DEX) into adrenalectomized rats fasted overnight. Hepatic glycogen levels are very low in adrenalectomized rats, and DEX causes rapid deposition of the complex carbohydrate. Investigation of the process of glycogen deposition was performed by light and electron microscopic (EM) radioautography using [3H]galactose as a glycogen precursor. Rats injected with DEX for 2-3 h and [3H]galactose one hour before being killed displayed an increasing number of intensely labeled hepatocytes. EM radioautography revealed silver grains over small (+/- 1 micron) ovoid or round areas of the cytosome that were rich in smooth endoplasmic reticulum (SER) and contained a high concentration of small dense particles. These distinct areas or foci of SER and presumptive glycogen (SERGE) were most numerous during initial periods of glycogen synthesis. After longer exposure to DEX (4-5 h) more typical deposits of cytoplasmic glycogen were evident in the SERGE regions. Several criteria indicated that the SERGE foci contained glycogen or presumptive glycogen: resemblance of the largest dense particles to beta-glycogen particles in EM; association of 3H-carbohydrate with the foci; removal of particles and label with alpha-amylase; and positive reaction with periodic acid-chromic acid-silver methenamine. The concentration of SER in the small foci and the association of newly formed glycogen particles with elements of SER suggest a role for this organelle in the initial synthesis of glycogen.     

Glycogen and glycogen enzymes in the liver and striated muscle of rats under altered thyroid states

Changes induced in liver and striated muscle glycogen and glycogen enzymes (glycogen synthetase, glycogen phosphorylase and alpha-amylase) by hypothyroidism and hyperthyroidism in rats have been determined. There were no changes in liver glycogen synthetase, phosphorylase and amylase activities in the hypothyroid group. Hyperthyroid rats showed lower liver glycogen synthetase, phosphorylase a and amylase activities. In muscle, hypothyroid rats had lower phosphorylase activity. In the hyperthyroid group glycogen synthetase was increased.--The results presented do not completely agree with the glycogen levels found in both tissues studied, and they are obviously more related to other factors such as glucose availability. It can be concluded that under the conditions studied, the glycogen enzyme levels could not alone explain the variations of glycogen levels.     

Synthesis of fatty acids in the perfused mouse liver

1. Fatty acid synthesis de novo was measured in the perfused liver of fed mice. 2. The total rate, measured by the incorporation into fatty acid of (3)H from (3)H(2)O (1-7mumol of fatty acid/h per g of fresh liver), resembled the rate found in the liver of intact mice. 3. Perfusions with l-[U-(14)C]lactic acid and [U-(14)C]glucose showed that circulating glucose at concentrations less than about 17mm was not a major carbon source for newly synthesized fatty acid, whereas lactate (10mm) markedly stimulated fatty acid synthesis, and contributed extensive carbon to lipogenesis. 4. The identification of 50% of the carbon converted into newly synthesized fatty acid lends further credibility to the use of (3)H(2)O to measure hepatic fatty acid synthesis. 5. The total rate of fatty acid synthesis, and the contribution of glucose carbon to lipogenesis, were directly proportional to the initial hepatic glycogen concentration. 6. The proportion of total newly synthesized lipid that was released into the perfusion medium was 12-16%. 7. The major products of lipogenesis were saturated fatty acids in triglyceride and phospholipid. 8. The rate of cholesterol synthesis, also measured with (3)H(2)O, expressed as acetyl residues consumed, was about one-fourth of the basal rate of fatty acid synthesis. 9. These results are discussed in terms of the carbon sources of hepatic newly synthesized fatty acids, and the effect of glucose, glycogen and lactate in stimulating lipogenesis, independently of their role as precursors.     

A (1----4)-alpha-D-glucan-protein involved in liver glycogen biosynthesis

A macromolecular (1----4)-alpha-D-[14C]glucan-protein complex was synthesized with a rat liver preparation and uridine diphosphate D-[14C]glucose. The size of the complex is contributed by both the protein and the (1----4)-alpha-D-glucosyl-oligomer components. Iodoacetamide treatment did not change the migration properties on Bio-Gel A-50m. Therefore, disulfide bonds linking glucan-protein subunits seem not to be involved. The [14C]glucan-protein, precipitated by diluted trichloroacetic acid, was digested by alpha-amylase, phosphorylase a, and proteases. The extent of proteolysis was greater for a complex having fewer D-glucose units incorporated. After proteolytic digestion of that complex, the labeled fragments behaved on electrophoresis, and ion-exchange and gel chromatography as [14C]glucosylated peptides. These findings support previous conclusions that the primer for liver glycogen synthesis is a protein on which glycogen is built up by covalent attachment.     

Lactate metabolism in the perfused rat hindlimb.

A preparation of isolated rat hindleg was perfused with a medium consisting of bicarbonate buffer containing Ficoll and fluorocarbon, containing glucose and/or lactate. The leg was electrically prestimulated to deplete partially muscle glycogen. The glucose was labelled uniformly with 14C and with 3H in positions 2, 5 or 6, and lactate uniformly with 14C and with 3H in positions 2 or 3. Glucose carbon was predominantly recovered in glycogen, and to a lesser extent in lactate. The 3H/14C ration in glycogen from [5-3H,U-14C]- and [6-3H,U-14C]-glucose was the same as in glucose. Nearly all the utilized 3H from [2-3H]glucose was recovered as water. Insulin increased glucose uptake and glycogen synthesis 3-fold. When the muscle was perfused with a medium containing 10 mM-glucose and 2 mM-lactate, there was little change in lactate concentration. 14C from lactate was incorporated into glycogen. There was a marked exponential decrease in lactate specific radioactivity, much greater with [3H]- than with [14C]-lactate. The 'apparent turnover' of [U-14C]lactate was 0.28 mumol/min per g of muscle, and those of [2-3H]- and [3-3H]-lactate were both about 0.7 mumol/min per g. With 10 mM-lactate as sole substrate, there was a net uptake of lactate, at a rate of about 0.15 mumol/min per g, and the apparent turnover of [U-14C]lactate was 0.3 mumol/min per g. The apparent turnover of [3H]lactate was 3-5 times greater. When glycogen synthesis was low (no prestimulation, no insulin), the incorporation of lactate carbon into glycogen exceeded that from glucose, but at high rates of glycogen deposition the incorporation of lactate carbon was much less than that of glucose. Lactate incorporation into glycogen was similar in fast-twitch white and fast-twitch red muscle, but was very low in slow-twitch red fibres. We find that (a) pyruvate in muscle is incorporated into glycogen without randomization of carbon, and synthesis is not inhibited by mercaptopicolinate or cycloserine; (b) there is extensive lactate turnover in the absence of net lactate uptake, and there is a large dilution of 14C-labelled lactate from endogenous supply; (c) there is extensive detritiation of [2-3H]- and [3-3H]-lactate in excess of 14C utilization.     

Carbohydrate metabolism in fructose-fed and food-restricted running rats

In chronically catheterized rats hepatic glycogen was increased by fructose (approximately 10 g/kg) gavage (FF rats) or lowered by overnight food restriction (FR rats). [3-3H]- and [U-14C]glucose were infused before, during, and after treadmill running. During exercise the increase in glucose production (Ra) was always directly related to work intensity and faster than the increase in glucose disappearance, resulting in increased plasma glucose levels. At identical work-loads the increase in Ra and plasma glucose as well as liver glycogen breakdown were higher in FF and control (C) rats than in FR rats. Breakdown of muscle glycogen was less in FF than in C rats. Incorporation of [14C]glucose in glycogen at rest and mobilization of label during exercise partly explained that 14C estimates of carbohydrate metabolism disagreed with chemical measurements. In some muscles glycogen depletion was not accompanied by loss of 14C and 3H, indicating futile cycling of glucose. In FR rats a postexercise increase in liver glycogen was seen with 14C/3H similar to that of plasma glucose, indicating direct synthesis from glucose. In conclusion, in exercising rats the increase in glucose production is subjected to feedforward regulation and depends on the liver glycogen concentration. Endogenous glucose may be incorporated in glycogen in working muscle and may be used directly for liver glycogen synthesis rather than after conversion to trioses. Fructose ingestion may diminish muscular glycogen breakdown. The [14C]glucose infusion technique for determination of muscular glycogenolysis is of doubtful value in rats.     

Inhibition of fatty acid synthesis and stimulation of glycogen breakdown by vasopressin in the perfused mouse liver

1. Vasopressin (anti-diuretic hormone, [8-arginine]vasopressin) inhibited the synthesis de novo of fatty acids (measured with (3)H(2)O and U-(14)C-labelled lactate or U-(14)C-labelled glucose) and stimulated glycogen breakdown in the perfused liver of fed mice. 2. The concentration dependence of these effects (range 200-1000muunits/ml, i.e. 0.5-2.5ng/ml) resembled that for the action on glycogen breakdown which was previously reported for rat liver. 3. The appearance of newly synthesized fatty acids in both phospholipids and triglycerides was inhibited by vasopressin, whereas synthesis of cholesterol was less affected. 4. Inhibition of hepatic lipogenesis by vasopressin is the most potent short-term hormonal action on this process yet reported. Aspects of the effect are discussed, including the lack of a role for cyclic AMP, and a possible link with vasopressin action on glycogen metabolism.     

Carbohydrate content and regulation following injection of different glycogenolytic enzymes.

The effect of injection of glycogenolytic enzymes on tissue glycogen, blood glucose and plasma insulin was studied in mice. No effects were observed following phosphorylase, whereas the hydrolytic enzymes, alpha-amylase and acid amyloglucosidase depressed liver glycogen. In addition acid amyloglucosidase induced a decrease in blood glucose, a slight elevation of plasma insulin and a marked increase in tolbutamide-stimulated insulin release. At the doses given none of the enzymes affected muscle glycogen. Amyloglucosidase pretreatment markedly enhanced insulin release induced by glibenclamide, leucine, isoleucine, lysine and glucose whereas insulin release stimulated by IPNA, ACTH, glucagon and "CCK-PZ" was unaffected. Injection of acid amyloglucosidase has a profound influence on carbohydrate content and regulation in mice. It is suggested that the dependence or independence of amyloglucosidase activity among the insulin secretagogues tested might reflect different or partially different mechanisms in the process of insulin secretion.     

Characterization of a glucose polymer from PC12 cells and neuronal cells of rat embryo

A large-sized glucose polymer was isolated by pronase digestion from line PC12 pheochromocytoma cells metabolically labeled with [1-3H]galactose. The polymer was included on a column of concanavalin A-Sepharose and could be eluted with 10 mM methyl-alpha-mannoside. Its slight retention in a column of Bio-Gel A-5m suggested that its molecular weight was in the several millions. Glucose was the component monosaccharide and there were two minor lipophilic components present. The polymer was digested with alpha-amylase into a series of oligosaccharides and was cleaved by glucoamylase into glucose residues. The disaccharide obtained by digestion with alpha-amylase was identified as maltose in several HPLC systems and by NMR spectroscopy. NMR measurement revealed the trisaccharide to be maltotriose. Susceptibility of the polymer molecule to alpha-amylase, and the digestion products obtained, indicated a resemblance to glycogen. An analysis for saccharide compositions before and after reduction of the polymer suggested the presence of an aglycon part. Contrary to expectations based on the presence of this moiety, the polymer displayed good solubility in neutral organic solvents. Two-thirds of the glucose polymer was also soluble in 10% TCA. A similar glucose polymer was isolated from neuronal cells of rat embryos metabolically labeled with [1-3H]galactose. Mouse neuroblastoma cells did not synthesize the polymer.     

The use of tritiated water to measure absolute rates of hepatic glycogen synthesis.

Glucogen synthesis in rat liver in vivo was measured by the incorporation of 3H from 3H2O into glycogen. In meal-fed rats incorporation and the incorporation of 3H into glycogen was linear up to 100 min. Before feeding glycogen concentration and the incorporation of 3H were both low; and both rose on feeding to give maximal values after 2-3h. The glycogen concentration was maintained for a further 5h but the incorporation of 3H rapidly declined to pre-feeding values. This shows that glycogen turnover was low in the post-prandial rat. Streptozotocin diabetes decreased the rise in glycogen concentration on feeding and had a similar effect on 3H2O incorporation. Both effects were reversed by insulin administration. The number of 3H atoms incorporated per glycogen glucose moiety formed in biosynthetic experiments (2.84 +/- 0.47) was relatively constant and allowed absolute biosynthetic rates to be calculated. Degradation of glucose from glycogen labelled by 3H2O showed that most of the 3H was located at C-2 and C-5. The incorporation would arise by rapid equilibration of hexose phosphates through phosphoglucose isomerase, transaldolase and triose phosphate isomerase.     

Region of Longitudinal Growth in Striated Muscle Fibres

THE use of radioactively labelled amino-acids to locate the newly formed contractile proteins of muscle is not specific, for they are incorporated into all the proteins being synthesized by the muscle which contain that amino-acid in their sequence. We have attempted to label more specifically the newly formed actin subunits in growing muscle using radioactive adenosine. (We thank Dr M. R. Iyengar of the University of Pennsylvania for first making this suggestion.) The principle of the method is that the adenosine is incorporated into the structural ADP^1,2 of the actin monomers and the free adenosine and most other adenosine-containing compounds are removed from the muscle fibres by subsequent glycerol extraction³.     

Some biochemical effects of triamcinolone acetonide on rat liver and muscle

1. The powerful anti-inflammatory glucocorticoid triamcinolone acetonide, administered to rats at 20 and 2.5mg/kg, leads to a decrease in the incorporation in vivo of [(3)H]uridine and [(32)P]orthophosphate into hind-limb skeletal muscle. 2. At the higher dose, this decrease in the rate of incorporation of precursors into RNA precedes a decrease in the incorporating ability of muscle ribosomes, which commences about 4-5h after drug administration, but is unaccompanied by any changes in the concentration of tissue ATP or free amino acids. 3. The ribosomal dysfunction extends to polyribosomes, which can only be successfully isolated from the muscle of triamcinolone-treated animals after the addition of alpha-amylase to the tissue homogenate to remove glycogen. 4. The specific radioactivity of muscle protein labelled in vivo with (14)C-labelled amino acids does not decrease progressively after triamcinolone administration. After 2h there is an apparent stimulation of incorporation which leads to an overall discrepancy between measurements of protein-synthetic activity made in vivo and in vitro. 5. There is a significant increase in muscle-glycogen concentration between 8 and 12h after the administration of triamcinolone acetonide (20mg/kg), although a significant decrease occurs after 4h. The fall in glycogen concentration may be due to a decrease in the rate of synthesis of protein essential for glucose uptake into the tissues. 6. As judged by (a) incorporation of (14)C-labelled amino acids into protein, (b) [(3)H]uridine and [(32)P]-orthophosphate incorporation into RNA, (c) the rate of induction of tryptophan pyrrolase and (d) changes in the pool sizes of taurine and tryptophan, the responses in liver followed the same time-course as those in muscle after administration of the drug.     

Escherichia coli produces a cytoplasmic alpha-amylase, AmyA.

In the gap between two closely linked flagellar gene clusters on the Escherichia coli and Salmonella typhimurium chromosomes (at about 42 to 43 min on the E. coli map), we found an open reading frame whose sequence suggested that it encoded an alpha-amylase; the deduced amino acid sequences in the two species were 87% identical. The strongest similarities to other alpha-amylases were to the excreted liquefying alpha-amylases of bacilli, with > 40% amino acid identity; the N-terminal sequence of the mature bacillar protein (after signal peptide cleavage) aligned with the N-terminal sequence of the E. coli or S. typhimurium protein (without assuming signal peptide cleavage). Minicell experiments identified the product of the E. coli gene as a 56-kDa protein, in agreement with the size predicted from the sequence. The protein was retained by spheroplasts rather than being released with the periplasmic fraction; cells transformed with plasmids containing the gene did not digest extracellular starch unless they were lysed; and the protein, when overproduced, was found in the soluble fraction. We conclude that the protein is cytoplasmic, as predicted by its sequence. The purified protein rapidly digested amylose, starch, amylopectin, and maltodextrins of size G6 or larger; it also digested glycogen, but much more slowly. It was specific for the alpha-anomeric linkage, being unable to digest cellulose. The principal products of starch digestion included maltotriose and maltotetraose as well as maltose, verifying that the protein was an alpha-amylase rather than a beta-amylase. The newly discovered gene has been named amyA. The natural physiological role of the AmyA protein is not yet evident.     

Histochemical detection of glycogen using Griffonia simplicifolia agglutinin II

The utility of a lectin from Griffonia simplicifolia (GSA II) for demonstrating glycogen in situ was tested on fixed paraffin-embedded sections of a variety of tissues from rodents and man. The histochemical specificity of GSA II conjugated to horseradish peroxidase (GSA II-HRP) for glycogen was documented by the lability of such staining on adjacent sections treated with either malt diastase or alpha-amylase. In some tissue sites, the lectin-HRP conjugate imparted cytoplasmic staining that was diastase- and amylase-labile but resisted digestion with N-acetylglucosaminidase. Reactivity in the latter sites was attributed to glycogen and occurred in cell types having well-documented glycogen content, including liver hepatocytes, skeletal muscle fibres and polymorphonuclear leukocytes. In other tissue loci, GSA II binding was confined to cell surfaces or intracellular compartments, was not affected by prior diastase or amylase digestion, but was abolished with N-acetylglucosaminidase. Staining in these sites was attributed not to glycogen, but instead to glycoconjugate having sugar chains terminated with N-acetylglucosamine. These findings document the affinity of GSA II for glycogen in situ but do not conflict with the biochemically demonstrated affinity of GSA II for terminal N-acetylglucosamine. The results reported here show that the GSA II-HRP method may be useful for detecting physiological or pathological changes in the glycogen content of cells.     

Effects of substrate branching characteristics on kinetics of enzymatic depolymerization of mixed linear and branched polysaccharides: II. Amylose/glycogen alpha-amylolysis

Enzymatic depolymerization of polysaccharides with alpha-amylase has been studied in mixed aqueous dimethylsulfoxide (DMSO)/water solvents. Polysaccharide substrate chemical compositions, configurational structures, and bonding pattersn are known to affect observed enzymatic reaction kinetics. The branching structures of polysaccharides and their effects on the kinetic mechanisms of depolymerization reactions via endo-acting hydrolyzing enzyme was studied via size exclusion chromatography coupled to low angle laser light scattering (SEC/LALLS). The glycogen branching structure is a heterogeneously distributed "cluster" structure rather than a homogeneously distributed "treelike" structure. The action pattern of alpha-amylase on glycogen, which is composed of highly branched clusters, as end-products, has a "pseudo-exo-attack" in contrast to an expected "endoattack" as seen in the hydrolysis of amylose or amylopectin substrates. These effects of branched substrates for mixed amylose/glycogen alpha-amylolysis have been predicted and demonstrated by both experimental and theoretical analysis using the kinetic model presented in this report. The "lumped" kinetic model employed, assumes that the enzyme simultaneously attacks both linear and branched substrates. In general, excellent agreement between the model predictions and the experimental observations, both qualitatively and quantitatively, was obtained. (c) 1995 John Wiley & Sons, Inc.     

Regulation of lysosomal glycogen metabolism: studies of the actions of mammalian acid alpha-glucosidases

1. Acid alpha-glucosidases were purified to homogeneity from rat liver, rat skeletal muscle and human placenta. The properties of these enzymes were investigated. 2. Their pH optima for activity toward various substrates were in the range 4-5. 3. Time course and pH dependence experiments revealed that all glycogen substrates were not hydrolysed at the same rate; the rate of hydrolysis was inversely related to the molecular size of the substrate. The most rapidly hydrolysed glycogen substrate was the smallest (commercial oyster) while the least rapidly hydrolysed was the largest (native rat or rabbit liver). Intermediate sized glycogens were hydrolysed at intermediate rates. 4. Glycogen hydrolysis was stimulated by added sodium ions; this stimulation was pH dependent. 5. It is suggested that lysosomal glycogen metabolism may be controlled by pH, salt concentration and the size of the glycogen substrate. 6. Since the high molecular weight glycogen associated with lysosomes is formed by disulphide bridges between lower molecular weight material it is proposed that an important step of lysosomal glycogen degradation is disulphide bond reduction.     

Quantification of the pathways followed in hepatic glycogen formation from glucose

Quantitative contributions of the direct and indirect pathways to liver glycogen formation from a glucose load have been estimated from 1) the distribution of label in glycogen formed from specifically carbon-labeled loads of glucose, 2) the specific activity of the glycogen compared with that of the circulating glucose, 3) the 3H:14C ratios in glycogen formed from loads specifically labeled with 3H and 14C, 4) the incorporation of 3H from 3H2O into the glycogen, and 5) the balance of glucogenic substrates across the splanchnic bed. A number of assumptions are made in the use of each of these methods. Estimates have been made for animals and humans fasted overnight or longer. Results obtained with the different methods are compared. Under these conditions, the contribution of the pathways appears to be determined by the size of the load, with larger contributions of the indirect pathway occurring with smaller loads.