Digestion is the process by which ingested materials are reduced to molecules of small enough size or other appropriate characteristics for absorption, i.e., passage through the gut wall into the blood stream. This generally means that proteins are hydrolyzed to amino acids or to polypeptide chains of a few amino acids, digestible carbohydrates to simple sugars, and lipids to fatty acids and glycerol. Materials not absorbed are by definition indigestible and are eventually voided as faeces. Digestibility ranges from 100 percent for glucose to as little as 5 percent for raw starch or 5-15 percent for plant material containing mostly cellulose (plant fibre). Digestibility of most natural proteins and lipids ranges from 80 to 90 percent.

Digestion is a progressive process, beginning in the stomach and possibly not ending until food leaves the rectum as faeces. Most studies of digestion simply compare the protein, lipid and carbohydrate content of the faeces with that of the feed. A study on digestion in channel catfish by Smith and Lovell (1973) showed continuing digestion (and absorption) of protein during passage through each part of the gut (Table 1). The methods employed in this study are discussed in Section 4 below. The comparison of faeces collected from the rectum and from the water also points out the hazard of incomplete recovery of faecal matter being likely when collection is done from outside the gut. Most of the protein digestion occurred in the stomach, but also continued in the intestine.

Table 1 - Apparent Digestibility of Protein by Channel Catfish ^1/

^1/ from Smith and Lovell, 1973

Temperature and pH play major roles in determining the effectiveness of digestive enzymes as a whole (details for specific enzymes are given in Section 4 below). Although most enzyme production decreases at temperatures above or below acclimation temperature, most enzyme activity (for a given amount of enzyme) increases in proportion to the temperature over a wide range of temperatures.

In general, enzyme reaction rates continue to increase at higher temperatures, even though the temperatures increase beyond the lethal temperatures for the species, until the enzymes begin to denature around 50-60°C. On the other hand, enzymes have limited ranges of pH over which they function, often as little as 2 pH units. Data for channel catfish are probably representative of many teleosts. Acid concentrations (pH) in the stomach ranged from 2 to 4, then became alkaline (pH = 7-9) immediately below the pylorus, decreased slightly to a maximum of 8.6 in the upper intestine, and finally neared neutrality in the hindgut (Page et al., 1976). Fish having no stomach have no acid phase in digestion.

The site of secretion in teleost stomachs appears to be a single kind of cell which produces both HCl and enzyme(s). This contrasts with mammals where two types of cells occur, one for acid and one for enzymes. The production of acid in teleosts is presumably the same as in mammals - NaCl and H₂CO₃ react to produce NaHCO₃ and HCl, with the blood being the source of both input materials, which are later mostly reabsorbed in the intestine. One possible explanation for the loss of stomachs in some species of fish is that they live in a chloride-poor environment and that providing large amounts of chloride ion for operating a stomach is bioenergetically disadvantageous. In addition to acid and enzymes, the stomach wall also secretes mucus to protect the stomach from being digested. As long as the rate of mucus production exceeds the rate at which it is washed and digested away, the gut wall is protected from being digested. When mucus production slows or fails, e.g. during gut stasis, during stressful conditions, or post mortem, the gut wall can be eroded or even perforated by the gut's own digestive enzymes.

Two sites produce enzymes in the midgut - the pancreas and the intestinal wall. The intestinal wall is folded or ridged in simple patterns which can be species specific. Secretory cells for both mucus and all three classes of enzymes develop in the depths of the folds, migrate to tops of the ridges (closest to the gut lumen), and then discharge their products. The pancreatic cells produce enzymes and an alkaline solution which are delivered to the upper midgut through the common bile duct. The control of pancreatic secretions (and the pyloric sphincter) in fish is probably the same as in mammals, but there is no information on teleosts yet.

The physical state of food passing through the gut varies with species and type of food. Fish, such as salmonids, which eat relatively large prey, reduce the prey in size layer by layer. Gastric digestion proceeds in a layer of mucus, acid, and enzyme wherever the stomach wall contacts the food. Food appears liquified only in the midgut and solidifies somewhat again during formation of faeces. Pellets of commercial feed seem to be treated similarly, i.e., pellets get smaller and smaller in size with time, although stomachs of some recently-fed salmonids have been found to contain moderate amounts of liquified pellets. Stomachs of juvenile Pacific salmon captured in the open sea contained a thick slurry of pieces of amphi-pods in various stages of solubilization. Fish whose food contains high levels of indigestible ballast, e.g., common carp feeding on a mixture of mud and plants, probably show minimal change in the appearance or volume of their food while it passes through the gut. Microphagous fish, such as the milkfish (Chanos) whose food starts out as a suspension of fine particles, probably also keep it in much the same form all the way through the gut. In general: there seems not to be the same degree of liquifaction of food in fish as is commonly described for mammals.

Absorption of soluble food could begin in the stomach - it occurs in mammals, but has not been investigated in fish - but takes places predominantly in the midgut and probably to some degree in the hindgut. The sites and mechanisms of absorption are largely unstudied, except histologically. Several histologists have identified fat droplets in intestinal epithelial cells following a lipid-rich meal. Increased numbers of leucocytes in general circulation following a meal by the sea bream and increased number of fat droplets in them have been described (Smirnova, 1966). It was hypothesized that leucocytes entered the gut lumen, absorbed lipid droplets, and then returned to the blood stream. It is clear that the mammalian type of villi with their lymph duct (lacteal) inside are absent in fish, although there is some folding and ridging of the gut wall to increase surface area. Lacteals serve as a primary uptake route in mammals for uptake of droplets of emulsified lipids (chylomicra). Teleost fish have a lymphatic system which includes extensions into the gut wall, but its role in lipid uptake is unknown. Absorption of amino acids, peptides, and simple carbohydrates have been little studied, but presumably they diffuse through or are transported across the gut epithelium into the blood stream. What light microscopists identified as a brush border on the surface of the epithelial cells facing the gut lumen, has now been clarified with electron-microscopy as microvilli; i.e., subcellular, finger like projections of the cell membrane whose greatly increased surface area is probably involved in absorption.

Digested food, particularly proteins, is not fully available to a fish even after it has been absorbed into the blood stream. Amino acids, if used for building new tissue, could be used as absorbed. If amino acids are to be oxidized for energy, however, deamination (removal of the amino group) must occur first - a reaction which requires input of energy. This process, known as specific dynamic action (SDA), can be measured externally in fish as an increase in oxygen consumption beginning soon after ingestion of food followed by an increase in ammonia excretion.

The proportion of amino acids which get deaminated varies with the food and the fish's circumstances. Fish which are not growing because of low temperature or have their ration at maintenance level or below, would deaminate most or all of their amino acids. Fish kept at high rearing temperatures or at high activity levels and therefore having very high metabolic rates would do likewise. On the other hand, fish having rapid growth and high protein intake would deaminate a relatively small proportion of their digested protein, although the absolute quantity of amino acids deaminated could still be large enough to produce a relatively large SDA. The energy for deamination need not necessarily come from amino acids, but will be preferentially taken from carbohydrate or lipid, if available. Thus, salmonid aquaculturists long ago discovered this "protein-sparing" action of limited amounts of inexpensive carbohydrate in the diet as a way of reducing the cost of feed and still achieving a desired level of growth. The protein-sparing action of lipids appears to have been minimally investigated. One can thus minimize SDA costs, but not avoid them completely.

Pepsin is the predominant gastric enzyme of all vertebrates, including fish. Optimal pH for maximal proteolytic activity has been reported for several species, as follows:

(a) pH 2 - pike, plaice
(b) pH 3-4 - Ictalurus
(c) pH 1.3, pH 2.5-3.5 - salmon, probably similar for tuna (Kapoor et al., 1975)

Peptic activity has been shown in a number of cultures and commercial species including Anguilla japonica, Tilapia mossambica, Pleuronecthys, both Salmo and Oncorhynchus species, Ictalurus, Micropterus, Lepomis and Perca. The presence of pepsin is so universal in vertebrates having stomachs that its presence can be presumed in fish for which no data is available.

The histochemistry of gastric secretion has been little studied in fish, although there is agreement on the presence of only one type of secretory cell in fish which stains positively for indicators of pepsinogen (pepsin precursor) cells. There is some question whether there may be more than one pepsin present in some fish, but no chromatographic or other tests have been done to investigate this. Several attempts have been made to identify acid-secreting cells, but results were either negative or confusing.

Other gastric enzymes have been proposed, but not firmly identified. Chitinolytic activity with an optimum at pH 4.5 was claimed for the stomach of Salmo irideus, but in most cases is probably from exogenous sources. If fish are like higher vertebrates, then the stomach wall also produces the hormone gastrin which stimulates gastric secretion. A lipase may also be present.

There are two sources of enzymes for the midgut - the pancreas and the secretory cells in the gut wall - with the pancreas perhaps secreting the greater variety and quantities of enzymes in fish. Because of the variety of enzymes present in different species, there have been some attempts to correlate enzyme activities with diet. However, these enzyme studies are fragmentary and histochemical tests are too general. Much remains to be learned about intestinal digestion in fish.

Trypsin appears to be the predominant protease in the midgut. Since the enzyme appears not to have been isolated, most authors have just tested for proteolytic activity over the pH range of 7 to 11 and reported their results as tryptic activity. The diffuse nature of the pancreas in most cases has limited many researchers to making relatively crude extracts from mixed tissues, hampering localization of the enzyme. Tryptic activity has been found in four stomachless species in Japan: Seriola, two basses and a puffer. Since these fish lack pepsin, some such kind of protease in the intestine would be the primary means of protein digestion. Tryptic activity was found in extracts of both the pancreas of perch and Tilapia and in intestinal extracts of Tilapia, all having a pH optimum of 8.0-8.2. Proteolytic activity has been identified in the pyloric caecae and intestine of rainbow trout. In grass carp, tryptic activity was stronger in the intestine than in the pancreas. In a mixture of pancreatic and pyloric caecae tissue from chinook salmon, casein was digested maximally at pH 9. Tryptic activity has also been demonstrated in extracts of liver of Several species, probably because in fish having a diffuse pancreas, pancreatic tissue extends into the liver, around the portal veins, and around the gall bladder. In several of the cases above, when extracts of pancreas were mixed with extracts of intestine, the tryptic activity increased ten-fold or more, suggesting the presence in fish of the enzyme enterokinase in the intestinal wall which activates in mammals the pancreatic trypsin as it reaches the intestine.

Additional pancreatic enzymes are involved in midgut digestion, many of them yet to be discovered. For example, Japanese workers are studying the occurrence and characteristics of a pancreatic collagenase in several Japanese fishes (Yoshinaka et al., 1973). There have also been several reports of chitinolytic activity in some fish which eat crustaceans predominantly. This could also have resulted from bacterial activity.

The occurrence of at least one lipase may be assumed in all fishes and has been demonstrated for a number of species. In carp and killifish extracts of intestine showed lipolytic activity. In goldfish, lipase activity occurred in extracts of a mixture of liver and pancreas and in the intestinal contents. Esterase (another lipase) activity has been found in the liver, spleen, bile, intestine, pyloric caecae and stomach of rainbow trout. Use of radioisotope-labeled lipids in cod suggested that the cod's lipase acted in the same manner as mammalian pancreatic lipase, although it was not considered more than a suggestion that fish lipase is of pancreatic origin. Regardless of origin, some kind of lipase is essential to fish because fatty acids are essential dietary components for fish.

Carbohydrases have perhaps excited the most interest of all the enzymes, particularly because salmonids do not handle the large carbohydrate molecules very well, and many workers wanted to determine the reason. Further, because there are several carbohydrases, the possibility that different enzyme combinations might show adaptations to different diets also intrigued some investigators. Also, herbivorous fish might be expected to have more carbohydrase activity and less tryptic activity than carnivores or omnivores.

Amylase is a widespread starch-digesting enzyme which occurs in human saliva and in pancreatic secretions into the small intestine. Amylase activity has been found in goldfish and bluegill sunfish in extracts of mixed liver and pancreas, oesophagus (contamination from regurgitated food suggested) and intestine, but not in large-mouth bass. Similar activity has been seen as well in rainbow trout, perch, Tilapia, Pacific salmon, cod, common carp, eel, and flounder. In fish with a diffuse pancreas there may be no pancreatic duct and so amylase activity appears in the bile. In mackerel. Scomber spp., which have a compact pancreas, the bile had no amylase activity.

Other carbohydrases identified included glucosidases (rainbow trout, chum salmon, common carp), maltase (common carp, red sea bream, Archosargus, marine ayu, Plecoglossidae), and sucrase, lactase, melibiase, and cellobiase, all of the latter in common carp. The hypothesis that carnivores might be deficient in one or more carbohydrases is largely disproved by the widespread presence of amylase in salmonids and other predators and by the presence of maltase in sea bream and ayu. The apparently larger diversity of carbohydrases in common carp than in other fish seems mostly a lack of information about fish other than carp. The question of whether dietary differences influence the kind of enzymes present must remain open but the evidence so far remains largely negative. However, there seems to be some evidence to show that the amounts of various enzymes may relate to the diet. Data in Table 2 suggest that herbivores have de-emphasized the production of proteases compared to the carnivores and the reverse for carbohydrases.

Table 2 - Relative Activity Levels of Amylase and Trypsin in Selected Cyprinids (Kapoor et al., 1975)

Similarly, in studies of Trachurus, Scomber, Mullus, Mugil, and Pleuronectes, the predatory species, Trachurus and Scomber had the highest proteolytic and lipolytic activities, while the planktivore, Mugil, had the lowest proteolytic and the highest amylolytic activities. Also, stomachless fish (which lack pepsin) are usually herbivores or omnivores, while carnivorous fish have true stomachs with peptic digestion. On the other hand, differences in proteolytic activity between Tilapia and Perca were small, and some other investigations of a variety of species failed to find any species differences. Apparently, where fish are somewhat specialized in their diets, differences in their enzyme activities are apparent. Many fish, however, remain non-specialized and have diversified diets and enzymes.

The functions of bile have scarcely been studied in fish, but presumably resemble those in higher vertebrates. In mammals bile is composed mainly of bilirubin and biliverdin, which are breakdown products of haemoglobin, and is produced continuously. These salts act like detergents and serve to emulsify lipids, thus making lipids more accessible to enzymes because of the increased surface area, allowing some lipids to be absorbed undigested as micro-droplets. In mammals, about 80 percent of the bile is recycled through the liver and gall bladder.

There are a few studies in fish which suggest that bile serves similar functions in fish. Several histologists have histochemically identified micro-droplets of lipid in midgut epithelium of fishes. That the gall bladder in fish reabsorbs water as in mammals has been confirmed. That bile is produced continuously in fish is suggested by the presence of green mucus in the lumen of the atrophied gut of spawning salmon. There appear to be no studies in fish of gall bladder contraction or other mechanisms controlling the release of bile during digestion. An observation of salmon having impacted gall bladders seemed related to diet because the gall bladders returned to normal when their dry pellet diet was changed to a moist pellet. Fish having impacted (and presumably non-contractile) gall bladders were normal otherwise and were indistinguishable in appearance and growth rates from fish in the same population with normal gall bladders.

Anatomists have tried for many years to correlate the shape of the liver and the position of the gall bladder in the liver with some of its functions. The basic functions of the liver in processing the foods which have been digested and absorbed are entirely cellular and molecular in scope. Thus, there is no functional requirement for shape at any level above the cellular level; i.e., livers basically could be of any shape. On the other hand, some restrictions are created by its position in the circulatory system between the gut and the heart, and the necessary interdigitation of the portal and hepatic veins, hepatic arteries, and bile ducts, all of which must serve essentially every cell of the liver. In common carp, the liver seems to have no shape of its own and simply fills every available space between the loops of the intestine. On the other hand, many fish (e.g., salmonids) have distinctive shape and colour to their livers. Changes in normal size and shape can indicate dietary or other problems. For example, a large, yellowish liver, often with white blotches suggests fatty degeneration of the liver caused by too much starch or by using saturated (mammalian) fats in the diet.

Common carp are representative of many cyprinids, including goldfish, squawfish, minnows, dace, chubs, and tench in North America. Most of these fish, including common carp, are omnivores, similar in several respects to catfish, but also differing significantly in several respects. Carp have maxillary barbels (most cyprinids do not) and forage in mud like catfish. However, carp ingest a considerably greater amount of plants than catfish and then chew the plants using a set of interdigitating pharyngeal teeth placed just anterior to the oesophagus. Carp lack a stomach, but have a long intestine which winds extensively throughout the visceral cavity. The gall bladder rests on the dorsal surface of the anterior midgut and the bile duet opens into the intestine just anterior to the gall bladder. In addition, the liver has no specific shape, but seems to serve as packing material around the intestine. Food seems to be ingested in small particles in a relatively steady stream instead of intermittently in large units, so the storage function of a stomach probably is not missed. With the liver filling all the available visceral space, there would be no room for accommodating the stomach expansion of a large meal anyway.

REFERENCES

Harder, W. 1975, Anatomy of fishes. Part I. Text. Part 2. Figures and plates. Stuttgart. E. Schweizerbart'sche Verlagsbuchhandlung, Pt.1:612 p., Pt.2:132 p. 13 pl.

Kapoor, B.B. 1975, H. Smit and I.A. Verighina, The alimentary canal and digestion in teleosts. Adv.Mar.Biol. 13:109-239

Magnuson, J.J. 1969, Digestion and food consumption by Skipjack tuna. Trans.Am.Fish.Soc., 98(3): 379-92

Page, J.W. 1976 et al., Hydrogen ion concentration in the gastrointestinal tract of channel Catfish. J.Fish Biol., 8:225-8

Phillips, A.M. Jr., 1969 Nutrition, digestion and energy utilization. In Fish physiology, edited by W.S. Hoar and D.G. Randall. New York, Academic Press, vol. 1:391-432

Post, G., W.E. Shanks and R.R. Smith, 1965 A method for collecting metabolic excretions from fish. Prog.Fish-Cult. 27:108-88

Smirnova, L.I., 1966 Digestive leukocytosis of bream (Abramis brama). In Biology of fishes of the Volga reservoirs. Tr.Inst.Biol.Vnutr.Vod./Trans.Inst.Biol.Inland Waters, 10(13) -.143-7

Smith, B.W. and R.T. Lovell, 1973 Determination of apparent protein digestibility in feeds for channel catfish. Trans.Am.Fish.Soc., 102(4):831-5

Yoshinaka, R., M. Sato and S. Ideka, 1973 Studies on collagenase of fish. l. Existence of colla-genolytic enzyme in pyloric caecae of Seriola quinqueradiata. Bull. Japan. Soc. Sci.Fish., 39(3):275-81

BACK