BIOCHEMISTRY of TRYPTOPHAN in HEALTH and DISEASE Contents

BIOCHEMISTRY OF TRYPTOPHAN IN HEALTH AND DISEASE

David A. Bender

Courtauld Institute of Biochemistry, The Middlesex Hospital Medical School, London WIP 7PN, U.K.

Contents

Chapter 1 THE DISCOVERY OF TRYPTOPHAN, ITS PHYSIOLOGICAL SIGNIFICANCE AND METABOLIC FATES 103

Tryptophan and glucose metabolism 105 Xanthurenic acid and insulin 105 The glucose tolerance factor 106 Inhibition of gluconeogenesis by tryptophan metabolites i07

Metabolic fates of tryptophan 108 Protein synthesis 108 Oxidative metabolism Ii0 5-Hydroxyindole synthesis 111 Intestinal bacterial metabolism iii

Chapter 2 THE 5-HYDROXYINDOLE PATHWAY OF TRYPTOPHAN METABOLISM; SEROTONIN AND OTHER CENTRALLY ACTIVE TRYPTOPHAN METABOLITES 112

Tryptophan 5-hydroxylase 112 Inhibition of tryptophan hydroxylase and the carcinoid syndrome 116

Aromatic amino acid decarboxylase 118 The specificity of aromatic amino acid decarboxylase 120

Tryptophan metabolism in the pineal gland 121

Monoamine oxidase 124

The uptake of tryptophan into the brain 124 The binding of tryptophan to serum albumin 127 Competition for uptake by other neutral amino acids 129 Changes in tryptophan metabolism in response to food intake 129 Tryptophan uptake into the brain in liver failure 131

Sleep and tryptophan metabolism 134

101 102 D.A. Bender

Tryptophan and serotonin in psychiatric disorders 135 Affective disorders 136 Evidence for a deficit of serotonin or tryptophan in depression 138 The use of tryptophan as an anti-depressant drug 140 Schizophrenia 141

Chapter 3 THE OXIDATIVE PATHWAY OF TRYPTOPHAN METABOLISM 145

Tryptophan oxygenase 145 Induction by hormones 145 Stabilisation and activation by tryptophan 149 Product inhibition 150 The availability of haem 151

Indoleamine dioxygenase (D-tryptophan oxygenase) 152 The metabolism of D-tryptophan 152

Formylkynurenine formamidase 153

The metabolism of kynurenine 154 The tryptophan load test for vitamin B 6 status 156 The tryptophan load test in women receiving oestrogens 158

The synthesis of nicotinamide nucleotides 160 The control of tissue concentrations of nicotinamide nucleotides 163

Pellagra 165 Non-nutritional pellagra 168 Hartnup disease 169 Carcinoid syndrome 169 Isoniazid therapy for tuberculosis 170 The pellagragenic effect of excess dietary leucine 171

Picolinic acid and the absorption of zinc 173

REFERENCES 177 Chapter I

The Discovery of Tryptophan, Its Physiological Significance and Metabolic Fates

The discovery of the amino acid tryptophan by Hopkins and Cole at the beginning of this century was an example of scientific serendipity - the facility for making happy and unexpected discoveries by accident and sagacity. One of the class practical exercises that medical students were expected to perform was the identification of proteins by means of the Adamkiewicz reaction - the formation of a coloured derivative when proteins were reacted with glacial acetic acid. The reaction worked well for some students, and failed for others. Perhaps today we would dismiss this as reflecting the inability of medical students to perform simple biochemical exercises. However, Hopkins was more persistent, and set out to investigate the reasons for the inconsistent performance of the reaction. He and Cole [I] showed that the active reagent was not acetic acid, but rather glyoxylic acid that was present as an impurity in some samples of acetic acid, but not in others. Today the chromogenic reaction of proteins with glyoxylic acid is generally known as the Hopkins-Cole reaction.

They went on to identify the component of proteins that was responsible for the reaction with glyoxylic acid, and found it was the same compound as was responsible for another colour reaction of proteins, the so-called 'tryptophane' reaction. 'The substance also gives the tryptophane reaction. If bromine water be cautiously added to aqueous solutions of the compound, excess being avoided, a fine rose-red colour is produced; and if the mixture be shaken with amyl alcohol the coloured product of the reaction is character- istically taken up by the latter solvent, showing in this the exact spectro- scopic absorption which is seen where the 'tryptophane' (proteinchrome) reaction has been obtained from the original mixture of tryptic digestion products. The tryptophane reaction, no less than the glyoxylic reaction, is seen to be associated with the compound that we are describing at each stage of the separation of the latter from the original digestion mixture .... there can be no doubt that our substance is the hitherto unknown precursor of the red colour of this familiar reaction ..... there is no doubt that our substance is the much sought tryptophane' [2].

The name tryptophan (the final '-e' was gradually dropped, for obscure reasons, during the decades following its discovery) seems to be derived from the fact that it is that which is 'revealed by tryptic digestion of a protein' The Greek ~cpoo means visible or evident. Acid hydrolysis of proteins destroys tryptophan, so that it can only be detected in enzymic or alkaline hydrolysates of proteins.

Tryptophan was the first of the amino acids to be shown to be a dietary essential, despite the fact that six of the other amino acids that we now know to be essential or indispensible had already been identified and isolated by the turn of the century. This reflects both the interest of Hopkins in essential nutrients, and also the fortunate chance that there were two specific reactions available for the detection and quantification of tryptophan - the glyoxylic (Hopkins-Cole) reaction and the 'tryptophane'

103 104 D. A, Bender

reaction. Furthermore, the amino acid could be destroyed by heating in strong acid, so that diets could be prepared for experimental animals that were essentially tryptophan-free. In fact, the early work establishing the indispensibility of tryptophan used diets based on zein, the principal protein of maize, as the sole protein source, since it had been established that zein was a poor source of tryptophan. Willcock and Hopkins [3] showed that adding tryptophan to such a diet led to an increase in the rate of growth of their animals. In view of this early work showing that zein, and hence maize, was deficient in tryptophan, it is perhaps somewhat surprising that the connection between tryptophan and pellagra arising in areas where maize was the dietary staple was not realised sooner than it was (see Chapter 3). The discovery and early investigation of tryptophan also marked another milestone in the development of the newly emerging science of biochemistry - the first use of bacterial degradation of a naturally-occurring compound as a means of determining its chemical structure [4].

Tryptophan probably has more entries in Index Medicus than any other single amino acid, and interest in this amino acid covers a wide range of scientific and medical specialties. A meeting on tryptophan may well attract biochemists, chemists and nutritionists on the one hand, and such diverse clinical special- ists as dermatologists, ophthalmologists, neurologists, psychiatrists, oncologists and endocrinologists on the other. The metabolism of tryptophan is affected by a number of medical conditions, and conversely, tryptophan and its metabolites have a number of profound effects on other metabolic and endocrine systems.

The naturally occurring form of tryptophan is the L-enantiomer, as with other amino acids. L-Tryptophan has a strong bitter flavour, which has caused some problems in the administration of doses of the amino acid to volunteers in laboratory studies, and, more seriously, to patients in various clinical trials. There are those who prefer to take the amino acid mixed with a strongly fruit-flavoured yoghourt, or even raspberry jam, to mask the flavour, and one preparation made available in Britain was mixed with cocoa powder (Optimax, Cambrian).

By contrast, the synthetic D-tryptophan has a pleasant, if somewhat lingering, sweet flavour. It is about 40-times as sweet as sucrose on a weight basis, and therefore might be of interest as a 'non-metabolisable' sweetener for use in energy or carbohydrate restricted diets. Of greater potential importance for such a use are two derivatives of D-tryptophan, 7-chloro- and 7-methyl-D-tryptophan. Both of these are several hundred times sweeter than sucrose. However, 7-chloro-D-tryptophan gives rise to metabolites that are potent inhibitors of the normal pathway of tryptophan oxidative metabolism [5], so that it is unlikely that this compound will be developed for use as a food additive.

The interest of oncologists in tryptophan is principally that when the amino acid is fed to animals at relatively high levels, over a prolonged period, it has some slight co-carcinogenic or promoter action in association with known chemical carcinogens [6]. It is difficult, if not impossible, to interpret these data in terms of human cancer and tryptophan metabolism. However, there is also evidence that one or more of the tryptophan metabolites that are excreted in increased amounts in vitamin B 6 deficiency (kynurenine, hydroxykynurenine, xanthurenic and kynurenic acids) may be associated with the development of bladder tumours. It has been suggested that vitamin B 6 supplementation may be appropriate for patients from whom such tumours have been removed surgically, in view of the normal high rate of recurrence [7]. Excretion of abnormally large amounts of these tryptophan metabolites may Biochemistry of Tryptophan 105 also underlie the high incidence of bladder cancer that is associated with schistosomal infection. The activities of both kynureninase and kynurenine aminotransferase are reduced in patients with schistosomiasis, and one of the main drugs used for therapy, oxamniquine, also inhibits both enzymes.

This means that these patients excrete very large amounts of kynurenine and hydroxykynurenine [8]. Several of the pyrolysis products of tryptophan that are formed in moderate amounts when meat and fish are grilled or broiled may also be carcinogenic [9]. Again it is difficult to interpret data obtained from experiments in animals, using relatively large amounts of the tryptophan pyrolysis products, with the human condition. Certainly there are epidemio- logical correlations between the intake of (grilled and broiled) meat and various forms of gastrointestinal cancer, but there are a great many other dietary factors involved, including differences in fat and fibre intake, and it would be premature to implicate the degradation products of tryptophan in the aetiology of human cancer.

Ophthalmologists also have reason to be interested in tryptophan. It has been reported that the concentration of the amino acid in the plasma of patients with senile or diabetic cataracts is very much higher than in matched control subjects with clear lenses. It has been suggested that one or more of the metabolites of tryptophan may bind to lens proteins and hence be responsible for the formation of cataracts [10]. It would be of interest to know whether there is any increase in cataracts above the expected prevalence in patients who are treated with relatively large amounts of tryptophan over prolonged periods for the control of depressive or manic-depressive illness.

The interest of dermatologists in tryptophan is obvious. Pellagra, the tryptophan and niacin deficiency disease, is most likely to present as a dermatological condition, and, as discussed in Chapter 3, there are a number of drug-induced and other non-nutritional forms of pellagra that also present with the characteristic skin lesions. Furthermore, acrodermatitis enteropathica may also be related to tryptophan metabolism, since it has been demonstrated that the zinc binding ligand that is required for the intestinal absorption of zinc is picolinic acid, a tryptophan metabolite. This is discussed in Chapter 3.

Tryptophan and Glucose Metabolism

Those interested in diabetes and the control of glucose metabolism also have reason to be interested in tryptophan and its metabolism, for at least three reasons: xanthurenic acid interacts with insulin; the glucose tolerance factor is believed to be a chromium-nicotinic acid chelate; and tryptophan and a number of its metabolites have been shown to inhibit gluconeogenesis.

Xanthurenic acid and insulin

Xanthurenic acid is normally a minor side-product of the oxidative pathway of tryptophan metabolism; however, in dietary or drug-induced vitamin B6 deficiency, and also as a result of the inhibition of kynureninase by oestrogen metaboLites, as discussed in Chapter 3, the formation of this compound is greatly increased. Xanthurenic acid forms a complex with insulin that has no biological activity [11] and it is likely that a high circulating concentration of xanthurenic acid has a similar effect in vivo. 106 D.A. Bender Some impairment of glucose metabolism has been observed in women treated with high-oestrogen oral contraceptives, and also during pregnancy (so-called gestational diabetes). The administration of supplementary vitamin B6 to women receiving oral contraceptives Leads to normalisation of their glucose metabolism, at the same time as it corrects their tryptophan metabolism [123. Gestational diabetes normally resolves at parturition, and to date there do not appear to have been any trials of vitamin B6 supplementation in this condition.

In the context of the inter-relationships between tryptophan metabolism and insulin, it is of interest that zinc, which is required for the normal storage of pro-insulin in the pancreas, is absorbed as a chelation complex with the tryptophan metabolite picolinic acid (see Chapter 3), and it is tempting to speculate that abnormalities of tryptophan metabolism may be involved in abnormalities of insulin metabolism and storage as a result of zinc deficiency. Impaired glucose tolerance has been reported in at Least some subjects in one report of nutritional zinc deficiency in the Middle East [133, although it is one of the Less important aspects of the deficiency.

The glucose tolerance factor

There have been reports of impaired glucose tolerance associated with chromium deficiency among under-nourished Palestinian refugees in Jordan [14] and also in patients maintained for prolonged periods on total parenteral nutrition [15]. In both cases, there was improvement in glucose tolerance following the administration of chromium supplements.

The study on the refugee children [14] was an interesting example of a well- controlled situation arising by accident. Two groups of under-fed children were investigated, both receiving the same United Nations emergency rations. Those in a hill-top refugee camp had impaired glucose tolerance, while those in a nearby camp on the valley floor had normal glucose tolerance. The difference appeared to be one of chromium intake, and it was shown that the drinking water supply at the top of the hill (largely rain-water collected in tanks) contained very little chromium, while the river water available at the valley floor contained three times as much of the mineral, leached out from the rocks and soil over which the river flowed.

The role of chromium in glucose tolerance has been elucidated by Schwartz and Mertz and their coworkers. They showed [16] that rats fed on a diet that was designed to produce liver necrosis developed impaired glucose tolerance which could be prevented or cured by the administration of a factor that they isolated from kidney powder. This 'glucose tolerance factor' was shown to contain chromium, and crude extracts from either pig kidney or brewers' yeast would correct the impaired glucose tolerance that developed in rats that were fed on diets based on Torula yeast. The administration of chromium supplements to chromium deficient animals led to increased responsiveness of isolated epididymal fat pad tissue to insulin, albeit with a moderate time lag before there was any effect. The administration of glucose tolerance factor led to a more or less immediate response; it was also effective in vitro when added to tissue preparations, whereas inorganic chromium salts were not [17, 18].

The glucose tolerance factor has been identified as a chromium-nicotinic acid chelate; it enhances the action of insulin in vitro [19, 20]. Whether chromium nutrition, and hence presumably also nicotinic acid nutrition and Biochemistry of Tryptophan 107 tryptophan metabolism, has any relevance to human diabetes is difficult to determine. It is obviously possible that in maturity onset diabetes, where there is apparently normal or greater than normal secretion of insulin, but an impaired tissue response to the hormone, there may be a deficiency of chromium nicotinate. Chromium is the only one of the essential trace minerals whose body content falls with age, and especially markedly so after repeated pregnancies; the demands of the fetus seem to be greater th~n the body's capacity to absorb chromium and form glucose tolerance factor. One small short-term trial showed a beneficial effect of chromium administration in three out of six maturity onset diabetics. There was no effect on glucose tolerance for some days after the initiation of chromium administration, presumably reflecting the time required for glucose tolerance factor synthesis. The authors stated that their evidence was not adequate to recommend chromium supplementation for the treatment of maturity onset diabetes [21].

Inhibition of gluconeogenesis by tryptophan metabolites

The administration of loading doses of tryptophan to rats leads to inhibition of gluconeogenesis, and hence to hypoglycaemia [22, 23]. With isolated perfused rat liver, the addition of relatively large amounts of tryptophan to the perfusion medium also inhibits gluconeogenesis [24], and the addition of large amounts of tryptophan to the culture medium has a similar effect on isolated rat hepatocytes [25]. The effect has been attributed to accumulation of quinolinic acid in response to the increased flux of metabolites through the oxidative pathway of tryptophan metabolism (see Figure I) [24]. In vitro, quinolinic acid will inhibit phospho-enol pyruvate carboxykinase, one of the key enzymes in the synthesis of glucose from amino acids and other metabolic intermediates [26, 27], although it has been suggested that the amounts of quinolinic acid that accumulate in the liver in response to the administration of even large amounts of tryptophan are unlikely to cause significant inhibition of the enzyme [28]. However, Smith et al. [29] have shown that there can indeed be a great enough accumulation of quinolinic acid in isolated rat hepatocytes in culture to account for considerable inhibition of glucose synthesis. The addition of tryptophan to the medium in which hepatocytes are cultured, in concentrations that are of the same order of magnitude as those that are found under physiological conditions, will inhibit gluconeogenesis significantly [30].

It has been suggested that the rat is atypical, and the results of these studies may not be applicable to man. There is no evidence of any inhibition of gluconeogenesis following tryptophan administration to human beings, and in isolated guinea pig hepatocytes the addition of tryptophan to the culture medium has no effect on the rate of gluconeogenesis, although quinolinic acid does inhibit phospho-enol pyruvate carboxykinase from this species [25, 31]. Studies on hepatocytes from sheep, guinea pigs and gerbils show that, in all of these species, tryptophan has little or no effect on gluconeogenesis, and a greater proportion of tryptophan is metabolised by way of total oxidation to acetyl CoA than by way of quinolinic acid, unlike the situation in rat hepatocytes [31]. Thus it seems likely that the activity of picolinic carboxylase, the enzyme that regulates the entry of tryptophan into the total oxidation pathway, rather than permitting an accumulation of acroleyl aminofumarate, and hence formation of quinolinic acid (see Figure 1) may be important in determining whether, and to what extent, tryptophan administration will inhibit gluconeogenesis. The same enzyme also affects the synthesis of nicotinamide nucleotides from tryptophan, as discussed in Chapter 3. 108 D.A. Bender Metabolic Fates of Tryptophan

There are three principal metabolic fates of dietary tryptophan: it may be incorporated into tissue proteins, or it may be metabolised, either by way of the 5-hydroxyindole pathway, leading to the neurotransmitter serotonin (5-hydroxytryptamine) (see Chapter 2), or by way of the oxidative pathway (see Chapter 3) which leads to complete oxidation through acetyl CoA or the net new synthesis of the nicotinamide nucleotides NAD and NADP.

Protein synthesis

It is a common misconception that the oxidative metabolism of tryptophan is limited to that part of the dietary intake that is 'surplus' to requirements for protein synthesis. This is obviously incorrect, since in an adult who is in nitrogen equilibrium there is no net new protein synthesis. The same amount of tissue protein is catabolised as is resynthesized and therefore an amount of tryptophan equivalent to the whole of the dietary intake will be available for metabolism. Even in a child who is growing, and is therefore in positive nitrogen balance, with a net retention of protein in the body, new protein synthesis in only a minute fraction of the total daily synthesis of proteins, and only a small proportion of the total dietary intake of protein (and hence of tryptophan) is retained. Again the majority is available for metabolism.

Figure I Pathways of tryptophan metabolism

(a) Tryptophan hydroxylase (L-tryptophan, tetrahydropteridine: oxygen oxidoreductase (5-hydroxylating) EC 1.14.16.4)

(b) Tryptophan oxygenase (L-tryptophan: oxygen oxidoreductase (decyclising) EC 1.13.11.11)

(d) Kynureninase (L-kynurenine hydrolase, EC 3.7.1.3)

(e) Picolinic carboxylase (amino-carboxy-muconic semialdehyde carboxy-lyase, EC 4.1.1.45)

(f) Non-enzymic cyclisation

(g) Kynurenine aminotransferase (L-kynurenine: 2-oxo-glutarate aminotransferase (cyclising) EC 2.6.1.7) Biochemistry of Tryptophan 109

CH 5" CH~" NH 2 PROTEINS

H serotonin J COOH

tryptophan °I " O COOH e I C- CH2" CH-NH 2 kynurenic acid

NH 2 urenine 0 COOH Ii I g C- CH~ CH-NH 2 xanthurenic acid : ~ _NH 2 OH hydr oxykynurenine

~ COOH O--CH ";~'-NH e / COOH 2

I II aminofumarate CO0 H O-CH COON2-NH2 ~" N ~COOH / /aminomuconic quinolinic / semialdehyde acid / NAD CO2

Figure 1

Pathways of tryptophan metabolism 110 D.A. Bender

Although the role of essential dietary amino acids was originally elucidated by Willcock and Hopkins [3] using the tryptophan deficient protein zein, tryptophan is not known to be the limiting amino acid in any dietary protein. That is to say, although the amount of tryptophan in a protein may be very low, there is always another essential amino acid that is present in lower amount relative to the requirement for protein synthesis. It is this amino acid which limits the utility of a dietary protein for tissue protein synthesis. In the case of zein the content of lysine is lower, relative to the requirement for lysine for protein synthesis, than that of tryptophan, and lysine is therefore the limiting amino acid of zein. It must be assumed that the improvement in growth that Willcock and Hopkins observed when they added tryptophan to their zein-based diets [3] was due to an increase in the availability of tryptophan for nicotinamide nucleotide synthesis; it was only when they also added lysine to the diets that they obtained a more or less normal rate of growth in their animals.

Nevertheless, the availability of tryptophan for protein synthesis may be an important factor in the regulation of protein turnover under normal physiological conditions. The physiological response to the administration of glucocorticoids, or an increase in their secretion, is net catabolism of tissue proteins and an increase in the rate of gluconeogenesis. This is one of the major regulatory events in the switch from the fed to fasting metabolic state - a change from protein deposition to protein catabolism and from glucose utilization to glucose synthesis in the liver and kidney. At a molecular level there are two principal responses to glucocorticoids - induction of tryptophan oxygenase, the first and rate-limiting enzyme of the tryptophan oxidative pathway, and induction of tyrosine aminotransferase, the rate-limiting enzyme of tyrosine catabolism. Thus, the effect of glucocorticoid stimulation is to increase the catabolism of two essential amino acids. This leads to a reduction in the tissue pools of these two amino acids, and most markedly in tissue levels of tryptophan. Under these conditions there is a reduction in the rate of new protein synthesis, since the amount of tryptophan that is available is now the limiting factor of protein synthesis. However, the rate of protein catabolism continues unchanged, so that there is now a net excess of catabolism over synthesis of tissue proteins. This means that there is an excess of the remaining amino acids that cannot be used once the available tryptophan has been exhausted. These are substrates for gluconeogenesis or ketogenesis. Hence, the overall result of the induction of tryptophan oxygenase by glucocorticoids is an increase in the net rate of protein catabolism and an increase in the synthesis of glucose from those amino acids that cannot be used otherwise. Tryptophan oxygenase has a short half-life (of the order of 2 hours), so that on cessation of glucocorticoid stimulation there is a rapid return to the normal rate of tryptophan catabolism, and a restoration of the normal balance between protein synthesis and catabolism. Sourkes [32] has observed similar effects, but lasting for very much longer, when tryptophan oxygenase is activated and stabilised against breakdown by the administration of ~- methyl tryptophan.

Oxidative metabolism

The oxidative pathway is the major route of tryptophan metabolism; under normal conditions some 99% of the body's tryptophan metabolism is by way of tryptophan oxygenase. Biochemistry of Tryptophan 111 It is often thought that the major product of this pathway is nicotinamide adenine dinucleotide (NAD), i.e., that the tryptophan oxidative pathway is primarily a pathway for the net new synthesis of NAD and NADP. However, as discussed in Chapter 3, this is not so. Under normal conditions, most of the tryptophan that enters the oxidative pathway is metabolised by way of acetyl CoA, and hence is oxidised completely, to carbon dioxide and water. It is only when this branch of the pathway is saturated, because of a high flux of metabolites through the pathway, that nicotinamide nucleotides become a major product of tryptophan metabolism. Nevertheless, there is evidence from a number of clinical conditions, discussed in Chapter 3, that tryptophan is a very important source of these coenzymes in the body, and not merely a 'substitute' for nicotinamide and nicotinic acid when the diet is deficient in these two compounds.

5-Hydroxyindole synthesis

The other important pathway of tryptophan metabolism in mammalian tissues is the 5-hydroxyindole pathway, leading to the synthesis of the neurotrans - mitter amine serotonin. Normally only about I% of the body's tryptophan metabolism is by way of 5-hydroxylation, yet the 5-hydroxyindoles are of considerable physiological importance, as discussed in Chapter 2. Indeed, very much the larger part of the medical importance of deviations from normal tryptophan metabolism depends on this quantitatively minor pathway.

Intestinal bacterial metabolism

There is some evidence that direct decarboxylation of tryptophan to the amine tryptamine may occur in mammalian tissues; certainly there is a relatively large amount of tryptamine in the central nervous system, and it may have a physiological function. However, it is also clear that a considerable proportion of the body's tryptamine is the result of bacterial decarboxylation in the intestinal lumen (see Chapter 2).

A great many tryptophan metabolites are found in faeces, including indole, skatole and other indole derivatives. These arise by bacterial metabolism of tryptophan in the intestinal lumen. Under abnormal conditions, such as the defective absorption of tryptophan that occurs in Hartnup disease, these metabolites are formed in larger than normal amounts, and may be absorbed from the large intestine, and be excreted in the urine [33].

While the resultant abnormal urine biochemistry may be of some interest (for example in the 'blue diaper' syndrome, where indigotonin, an oxidation product of the tryptophan metabolite indican, is excreted in relatively large amounts [34]), most of these 'exotic' bacterial metabolites of tryptophan do not seem to have any physiological function in man. They are the result of passive absorption from the large intestine, and are excreted in the urine either unchanged or after the usual processes of conjugation in the liver. Chapter 2

The 5-Hydroxyindole Pathway of Tryptophan Metabolism; Serotonin and other Centrally Active Tryptophan Metabolites

The pathway of serotonin synthesis from tryptophan is shown in Figure 2. The role of this 5-hydroxyindoleamine as a neurotransmitter was established during the 1960's, with the demonstration of its distribution in specific tracts of neurons in the central nervous system, especially those originating in the raphe nucleus, by the formaldehyde fluorescence technique [35]. Serotonin also fulfils other criteria for consideration as a neurotransmitter: it is synthesized in the neurons in which it is found; it is secreted on electrical stimulation of the appropriate regions of the nervous system; it produces electrical and other effects when it is applied iontophoretically to post-synaptic neurons; it is catabolised rapidly after release from pre- synaptic neurons; there is good evidence that there are specific post- synaptic serotonin receptors, possibly associated with the formation of cyclic nucleotides [36].

Tryptophan 5-Hydroxylase

The first step in the synthesis of serotonin is 5-hydroxylation of tryptophan to yield 5-hydroxytryptophan, 5-HTP. Friedman and coworkers [37] demonstrated that tryptophan hydroxylase is a biopterin-dependent enzyme. They noted that the stimulation by Fe(II) ions that had been reported by earlier workers was not the result of iron-dependence of the enzyme, but seemed to be the result of removal of hydrogen peroxide, which accumulates as the result of side reactions of the enzyme. In the presence of catalase to remove this peroxide, there is no stimulation by Fe(II) ions.

The hydroxylation of tryptophan requires molecular oxygen, and in the reaction the cofactor, tetrahydrobiopterin, is oxidised to dihydrobiopterin. This is reduced back to the active cofactor for further hydroxylation by a second enzyme, dihydrobiopterin reductase, which uses NADPH as the reductant. This means that tryptophan hydroxylase is the same type of enzyme as phenylalanine and tyrosine hydroxylases, in that a specific amino acid hydroxylase requires the presence of dihydrobiopterin reductase for continuing activity. In vitro it is possible to measure the activity of the hydroxylase without the presence of the reductase only if there is an adequate supply of reduced cofactor, either tetrahydrobiopterin itself, or one of a number of synthetic analogues that are also active.

A number of cases of 'atypical' phenylketonuria have been reported, in which phenylalanine hydroxylase is normal (in the 'classical' form of the disease this is the enzyme that is defective), but there is an impairment of biopterin metabolism. In most such cases, it is dihydrobiopterin reductase that is

112 Biochemistry of Tryptophan 113

COOH

CH,- &H-NH,

tryptophan

tetrahydro biopterin N AOP

b x dihydrobiopterin NADPH

t COOH

CHT LH-NH,

5-hydroxytryptophan

CHh CH,-NH,

5-hydroxytryptamine (serotonin) H

(a) Tryptophan hydroxylase CL-tryptophan, tetrahydropteridine: oxygen oxidoreductase (5-hydroxylating) EC 1.14.16.4)

(b) Dihydrobiopterin reductase (NADPH: 6,7_dihydropteridine oxidoreductase, EC 1.6.99.7)

Cc) Aromatic amino acid decarboxylase (L-aromatic amino acid carboxy-Lyase, EC 4.1 .I .28)

Figure 2

The synthesis of serotonin 114 D A Bender defective, and therefore as well as the defective metabolism of phenylalanine that leads to the formation and excretion of the phenylketones, there is also impairment of the hydroxylation of tyrosine, leading to defective synthesis of the catecholamines, and of tryptophan, leading to defective synthesis of serotonin. The few patients who have been identified with this variant of phenylketonuria show a cluster of neurological abnormalities, in addition to the defects of classical phenylketonuria, that would be compatible with defective synthesis of catecholamines and serotonin [38, 39]. In two patients a deficit of serotonin and dopamine in biopsy samples from the central nervous system and Low Levels of catecholamine metabolites and 5-hydroxy-indoleacetic acid, the principal metaboLite of serotonin, in the cerebrospinal fluid, have been demonstrated [403°

Two further patients with 'atypical' phenylketonuria of a different kind have been reported. Thes~ patients also failed to respond to dietary treatment but they had normal activities of both phenylalanine hydroxylase and dihydrobiopterin reductase. They had defects in the pathway of biopterin synthesis. They were fortunate, since suitable precursors of biopterin, distal to the defective step in the pathway, can be administered orally. For these patients biopterin or a suitable precursor is a vitamin, whereas normal people are capable of synthesizing the cofactor [39, 413. For other patients with 'atypical' phenylketonuria due to a deficit of dihydrobiopterin reductase, no such therapy is possible, since it is not feasible to administer tetrahydrobiopterin in adequate amounts to obviate the need for the reductase. Tetrahydrobiopterin is unstable to both light and air, so that it would not remain in an active form for Long enough to be taken up by tissues. It is possible that one of the synthetic tetrahydrobiopterin analogues that have been developed and used for in vitro studies of tryptophan and related hydroxylases may be useful in these cases.

There is excellent evidence that the activity of tryptophan hydroxylase is the rate Limiting step in the synthesis of serotonin in vivo. Changes in the availability of tryptophan to the brain Lead to predictable changes in the rate of serotonin synthesis and turnover. Furthermore, the kinetics of the enzyme are such that it normally functions below saturation, so that any increase in the concentration of tryptophan available for hydroxylation would be expected to lead to an increase in the rate of reaction. The steady state concentration of tryptophan in the brain under physiological conditions is in the range of 5 - 25 ~mol/kg; this is an average concentration, and ignores the possibility that there may be localised areas of very much higher concentration as a result of metabolic compartmentation in the central nervous system. When the physiological cofactor, tetrahydrobiopterin, is used the K m of the enzyme is of the order of 50 pmol/l [42]. When synthetic biopterin analogues are used to measure the activity of the hydroxylase, the apparent K m for tryptophan is very much higher than this, and early reports, based on such studies, suggested that tryptophan hydroxylase normally functions at a substrate concentration very considerably below its Km, so that any change in the concentration of substrate available would be expected to lead to a very considerable increase in the rate of hydroxylation. Even when the lower value of K m obtained using tetrahydrobiopterin as cofactor is considered, it is obvious that tryptophan hydroxylase normally operates at considerably less than its maximum rate of reaction (Vmax) , and therefore any increase in the concentration of tryptophan will lead to a corresponding increase in the rate of hydroxylation, and hence in the rate of serotonin synthesis.

Friedman and coworkers [37] noted that tryptophan hydroxylase was inhibited by an excess of tryptophan; they quoted a value of 1 mmol/l for the K~ Biochemistry of Tryptophan 115

Vmax

o t

cd i I

i' Km tryptophan concentration i ! ! , normal range of ~ in tryptophan 5 - 25 ~Imol/l

Figure 3

Tryptophan hydroxylase and the control of serotonin synthesis in the brain

(the concentration of tryptophan at which there is 50% inhibition of the enzyme). In their hands the enzyme had a K m of 0.29 mmol/l, and therefore they predicted that it would be inhibited before the enzyme was saturated with substrate - i.e. the predicted value of Vma x would never be achieved. This means that the increase in serotonin synthesis in response to increased availability of tryptophan in the brain would be less than might be predicted from consideration of only the Km and Vma x of the enzyme. Such an effect was observed by Grahame-Smith and coworkers [43] in a series of studies of the effects of tryptophan loading in animals. They showed no further increase in serotonin synthesis after the concentration of tryptophan in the brain had risen to about 0.33 mmol/l. However, Bensinger and coworkers [44] showed that in pineal gland cultures there was proportionality between the rate of tryptophan hydroxylation and the concentration of tryptophan in the culture medium up to as high as 0.5 mmol/1. It is not clear whether this reflects the different experimental conditions used by different workers, or the presence of different isoenzymes of tryptophan hydroxylase in different tissues. Certainly, the enzyme in the intestinal mucosa (mainly in the enterochromaffin cells) seems to be different from that in the central nervous system, and will hydroxylate D-tryptophan at about one third of the rate that is observed for L-tryptophan [45]. The brain enzyme shows no activity towards D-tryptophan [37].

Hamon and coworkers [46] showed that treatment of rats with methiothepin led to an increase in the affinity of tryptophan hydroxylase for its substrate, and an increase in the rate of synthesis of serotonin and 5-hydroxyindoleacetic acid. This effect was also observed when methiothepin was added to preparations of the enzyme in vitro, and therefore presumably it is an activation of the enzyme, perhaps by an effect on its quaternary 116 D.A. Bender

or subunit structure. Treatment of the enzyme with a variety of detergents, or changes in its phospholipid environment, also activate it reversibly [48]. When the enzyme is incubated in the presence of a moderately high concentration of calcium ions, there is an irreversible activation, accompanied by loss of the reversible detergent and phospholipid effects, and a reduction in molecular weight [48]. This suggests the loss of a regulatory subunit of the enzyme. The addition of calcium to the incubation medium increases the affinity of tryptophan hydroxylase for all three of its substrates (oxygen, tryptophan and tetrahydrobiopterin) [49] and the addition of a chelating agent to reduce the availability of calcium causes a change to sigmoid kinetics for all three substrates [49]. Brain, but apparently not other tissues, has a specific calmodulin (and hence indirectly calcium) dependent protein kinase which is active towards tryptophan and tyrosine hydroxylases [50]. This suggests that the effects of changes in calcium availability on tryptophan hydroxylase activity are the result of phosphorylation and perhaps dephosphorylation of the enzyme protein, a common mechanism of regulation of enzyme activity. Presumably the effects of detergent treatment and changes in the phospholipid environment of the enzyme [47, 48] which are reversible, are due to conformational changes similar to those caused by phosphorylation.

Inhibition of tryptophan hydroxylase and the carcinoid syndrome

A number of compounds have been shown to inhibit tryptophan hydroxylase in vitro; these include ring-substituted derivatives of phenylalanine and tryptophan, which compete with the substrate, a variety of catechols which are non-competitive with respect to both the substrate and the cofactor, and compounds that chelate iron [51]. This last group of compounds would not be expected to inhibit the enzyme in vivo, since, as discussed above, the enzyme is not iron-dependent, and the activation by Fe(II) ions is an artefact of incubation procedures [37].

The only inhibitor of tryptophan hydroxylase that has been used at all widely is p-chlorophenylalanine. Koe and Weismann [52] demonstrated that administration of this compound to experimental animals depleted brain serotonin gradually over several days. There was also a reduction in the concentrations of 5-hydroxy-indoleacetic acid both in the brain and peripherally, and a reduction in peripheral serotonin, p-Chlorophenylalanine also inhibits phenylalanine hydroxylase, but in these studies the effects on brain catecholamines were only marginal, suggesting that synthesis from pre-formed tyrosine continues unimpaired in the presence of p-chlorophenylalanine. In vitro, p-chlorophenylalanine is a competitive inhibitor of tryptophan hydroxylase, but after administration in vivo it leads to an irreversible non-competitive inhibition of the enzyme. It has been suggested that either some metabolite of the drug is a more potent inhibitor of tryptophan hydroxylase than is the parent compound, and that the delay in the development of maximum inhibition is the result of the gradual accumulation of this compound over a number of days, or possibly that p-chlorophenylalanine is incorporated into proteins in place of either tyrosine or phenylalanine. However, if this latter were the case, it would be difficult to account for the relatively specific effects of p-chlorophenylalanine; inhibition of a great many enzymes might be expected as the result of impaired protein synthesis. p-Chlorophenylalanine has been widely used both in animal studies as a means of depleting serotonin and in clinical medicine as a means of reducing serotonin synthesis in patients with carcinoid syndrome. A further, unrelated, Biochemistry of Tryptophan 117 use of p-chlorophenylalanine in experimental animals has been as an inhibitor of phenylalanine hydroxylase, in conjunction with loading doses of phenylalanine, as a model system for the study of phenylketonuria [53].

Carcinoid tumours are most commonly tumours of the enterochromaffin cells of the gastro-intestinal mucosa, but they can also occur as primary tumours of other endodermal tissues [54]. The carcinoid syndrome is a complex of flushing, diarrhoea with excessive intestinal motility and right-sided cardiac failure. The syndrome is rare in patients with only a primary intestinal tumour, but common when there are hepatic or other metastases of the intestinal tumour, or when the primary carcinoid tumour is at some site other than the intestine. Carcinoid tumours produce large amounts of serotonin, and this is responsible for the increased intestinal motility, and partially responsible for the flushing. It is probable that there are fewer such effects when the patient only has an intestinal primary tumour because the serotonin produced is largely metabolised to 5-hydroxyindole acetic acid in the liver. It is only when the serotonin enters the general circulation that the carcinoid syndrome is manifest. As well as serotonin, a number of other vaso-active compounds are produced by the tumour, and these have a considerable part to play in the symptomatology of the disease [55].

The administration of p-chlorophenylalanine to patients with carcinoid syndrome leads to a considerable degree of symptomatic relief, and reduces the excretion of 5-hydroxyindoleacetic acid to a marked extent. There is a gradual improvement over several days after the initiation of treatment in the nausea, diarrhoea and vomiting associated with the syndrome, which is correlated with the gradual reduction in 5-hydroxyindoleacetic acid excretion. The effect on flushing is equivocal. On withdrawal of p-chlorophenylalanine there is a gradual increase, over several days, in the excretion of 5-hydroxyindoleacetic acid, and a parallel exacerbation of the gastrointestinal symptoms. Low doses of p-chlorophenylalanine cause the patients to complain of tiredness, dizziness and headaches, while at doses of the order of I g per day, which have the maximum effect on the excretion of 5-hydroxyindoleacetic acid, there may be psychiatric effects as well, with either hallucinations or depression [56, 57].

In patients with severe carcinoid syndrome and widespread metastases of the tumour, as much as 60% of the body's tryptophan metabolism may be by way of the 5-hydroxyindole pathway, whereas in normal subjects only about I% is metabolised this way. This leads to a considerable reduction in the circulating concentration of tryptophan, as shown in Table 1; at the same time there is a considerable increase in the circulating concentration of serotonin, and the urinary excretion of 5-hydroxyindole acetic acid is very much higher than normal [58].

This severe depletion of the body's reserves of tryptophan, and the reduction in the amount of tryptophan being metabolised by way of the oxidative pathway, may lead to a sufficient reduction in the synthesis of nicotinamide nucleotides from tryptophan to cause the development of clinical pellagra, as discussed in Chapter 3. Lehman [59] cited only three cases of unequivocal pellagra in 140 carcinoid patients whom he discussed, and suggested that this low incidence may be because the syndrome is usually detected at a relatively early stage, so that although there is a considerable reduction in the new synthesis of nicotinamide nucleotides, the body's reserves have not been seriously depleted by the time that the patient presents for treatment. 118 D, A. Bender

TABLE 1

Plasma concentration of tryptophan and serotonin in carcinoid syndrome

normal range carcinoid patients tryptophan (~mol/l) 50-100 10-50 serotonin (pmol/1) 0.4-1.6 1.9-20

(data adapted from Warner [58])

Although the principal site of action of p-chlorophenylalanine is on tryptophan hydroxylase, it may also inhibit the second enzyme of serotonin synthesis, aromatic amino acid decarboxylase, as well. There is some evidence that it is a substrate for decarboxylation; after the administration of inhibitors of monoamine oxidase, p-chlorophenylethylamine can be detected [60]. This may explain some of the pharmacological actions of the drug that have been observed to occur before there is any significant depletion of serotonin. If monoamine oxidase is not inhibited, p-chlorophenylethylamine is rapidly oxidised, and p-chlorophenylpyruvate can be detected in the urine [61]. However, it is not clear whether the decarboxylation of p-chlorophenylalanine is the result of the action of mammalian aromatic amino acid decarboxylase, or of intestinal bacterial action. In rats that had been treated with p-chlorophenylalanine and phenylalanine as a means of inducing a phenylketonuria-like condition [53] there was no effect of the treatment on aromatic amino acid decarboxylase activity [Bender, D.A, Coulson, W.F. and Antonas, K.N., unpublished observations].

Aromatic Amino Acid Decarboxylase

The role of decarboxylation in the synthesis of serotonin was demonstrated by Udenfriend and coworkers in 1953 [62]. They suggested initially that the decarboxylase involved was distinct from that required for the decarboxylation of dihydroxyphenylalanine (dopa) in the synthesis of catecholamines, and showed that the two activities had very different pH optima (dopa decarboxylation is optimal at pH 6.8, and 5-hydroxytryptophan decarboxylation at pH 7.8), appeared to have different interactions with pyridoxal phosphate and could apparently be separated to some extent by ammonium sulphate frac- tionation and chromatography on columns of alumina [62, 63].

Later workers showed that the two activities are associated with the same protein; both activities are blocked to a similar extent by a number of inhibitors, and the two substrates are mutually inhibitory [64]. Throughout purification the two activities are recovered in a constant ratio [65], and both activities show the same pattern of four peaks on polyacrylamide gel electrophoresis [66]. This pattern of electrophoretically separable peaks seems to be an artefact, and the three slower running peaks of the activity seem to be metastable degradation products of the fastest running form of the enzyme [67]. The most convincing demonstration of the identity of dopa and 5-hydroxytryptophan decarboxylases came with the purification to electrophoretic homogeneity of a single protein with both activities [68] and the demonstration that titration of the enzyme with anti-serum prepared against the purified preparation led to loss of both activities at the same rate [69]. Biochemistry of Tryptophan 119 Despite the overwhelming evidence that a single enzyme has activity towards both dopa and 5-hydroxytryptophan, there remain a number of observations that suggest that the two activities might be due to separate enzymes. Sims and coworkers [70, 71] demonstrated different ratios of dopa: 5-hydroxytryptophan decarboxylase activity in different regions of the brain, and showed that partial thermal denaturation led to loss of the two activities at different rates. They also showed that intra-cisternal injection of 6-hydroxydopamine to destroy dopaminergic neurons led to a fall in dopa decarboxylase activity, with no effect on the decarboxylation of 5-hydroxytryptophan. Other workers have been unable to reproduce this finding, and have shown that the result of administering either 6-hydroxydopamine to destroy dopaminergic neurons or 5,7-dihydroxytryptamine to destroy serotoninergic neurons was a loss of the two decarboxylase activities at the same rate [72, 73, Bender, D.A and Coulson, W.F., unpublished observations].

Bender and Coulson [67] showed that a number of treatments that lead to partial denaturation of the enzyme (detergent treatment and the use of high concentrations of urea) also lead to changes in the ratio of dopa: 5-hydroxytryptophan decarboxylase activity. They suggested that this might reflect the presence of two separate substrate binding sites, associated with a single catalytic site. This would account for the ability of various treatments to alter the activity of the enzyme towards one substrate to a greater extent than towards the other. It would also explain the two very different pH optima of the enzyme for its two substrates, a difference which does not seem to be associated with the ionisation of the substrates, but rather to be a property of the enzyme itself. Such a model of the enzyme would recon- cile the evidence of changes in the ratio of the two activities of the enzyme with the irrefutable evidence that a single protein has both activities [68, 693.

Robins and coworkers [74] suggested that the activity of aromatic amino acid decarboxylase was the rate-limiting step in serotonin synthesis, because the activity of this enzyme in human brain is very low. However, as discussed above, there is abundant evidence that it is hydroxylation of tryptophan that is the rate-limiting step; aromatic amino acid decarboxylase has a very low Km for 5-hydroxytryptophan, and a maximum rate of reaction that suggests that it would be capable of decarboxylating all of the 5-hydroxytryptophan that would be formed by the hydroxylase either under normal conditions, or following a loading dose of tryptophan. Indeed, it is not possible to detect 5-hydroxytryptophan in the central nervous system at all unless an inhibitor of the decarboxylase has been given first [75].

Although the decarboxylation of 5-hydroxytryptophan is not normally the rate-limiting step of serotonin synthesis, changes in the activity of the decarboxylase may be physiologically important. Bowen and coworkers [76] have shown that among other changes in the brains of patients with presenile dementia, there is a significant loss of aromatic amino acid decarboxylase activity. In view of the importance of both serotonin and the catecholamines in affective disorders, and possibly also in psychosis, as discussed below, this may be an important finding, although it may not be specific to dementia, but rather reflect terminal anoxia.

A number of compounds are known that will inhibit aromatic amino acid decarboxylase. Two are of interest clinically because they do not cross the blood-brain barrier, and can therefore be used to inhibit the enzyme in peripheral tissues without affecting its activity in the central nervous system; they are Benserazide (Roche) and Carbidopa (Merck, Sharp and Dohme). 120 D.A. Bender

Both of these compounds are used in conjunction with dopa in the treatment of Parkinsonism, to minimise the extra-cerebral decarboxylation of dopa, and hence permit a smaller dose of the amino acid to be used, with the same central effect, but a lower incidence of side-effects associated with peripheral decarboxylation to dopamine. To date, neither has found any application clinically for the inhibition of 5-hydroxytryptophan decarboxylation, although it is conceivable that they would be useful in alleviating the symptoms of carcinoid syndrome, as alternatives to the use of p-chlorophenylalanine. However, as discussed in Chapter 3, both drugs are also potentially pellagragenic, since they inhibit kynureninase and tryptophan oxygenase, and it would therefore be undesirable to use them in a condition where there is already a risk of pellagra, especially since effective pallia- tive treatment is available already. Inhibition of tryptophan hydroxylation by the use of p-chlorophenylalanine, with sparing of tryptophan for oxidative metabolism and hence nicotinamide nucleotide synthesis, is obviously preferable to inhibition of the decarboxylation of 5-hydroxytryptophan, which cannot be used for any other purpose.

The specificity of aromatic amino acid decarboxylase

The extent to which aromatic amino acid decarboxylase will decarboxylate amino acids other than dopa and 5-hydroxytryptophan remains an area of some controversy. There is no dispute that the enzyme will also decarboxylate a number of unphysiological amino acids, such as dihydroxyphenylserine (a direct precursor of noradrenaline), and the o- and m-isomers of tyrosine. However, there are two schools of thought concerning the ability of the enzyme to catalyse the decarboxylation of phenylalanine to phenylethylamine, p-tyrosine (the physiological isomer of the amino acid) to tyramine and tryptophan to tryptamine. Udenfriend and coworkers [68, 77] have claimed that the enzyme has activity towards all of these amino acids, while others [67, 78, 79] have failed to demonstrate any significant activity of the enzyme towards phenylalanine, tyrosine or tryptophan. Furthermore, even at relatively high concentrations, these three amino acids do not compete with dopa and 5-hydroxytryptophan for decarboxylation, although these two amino acid substrates are mutually competitive [64].

Nevertheless, tryptamine and the other aromatic amines do occur in the central nervous system [80]. The concentration of tryptamine in the brain is increased following inhibition of tryptophan hydroxylase by p-chlorophenylalanine [81] or the inhibition of peripheral aromatic amino acid decarboxylase, although there is no increase in brain tryptamine if the inhibitor crosses the blood-brain barrier [82]. After inhibition of monoamine oxidase, the concentration of tryptamine in the brain follows changes in the circulating concentration of tryptophan [83]. All of this suggests that it is likely that the tryptamine that is found in the brain is the result of cerebral metabolism. Weil-Malherbe [84] claimed to have demonstrated the formation of tryptamine from tryptophan by brain slices. However, his study suffered from the problem, which he admitted, that his method was unable to distinguish between tryptamine and 5-methoxytryptamine, a possible onward metabolite of serotonin. Snodgrass and Iversen [84] demonstrated the formation of [14C] tryptamine from [14C]tryptophan by slices of rat spinal cord.

It remains possible that some of the tryptamine in the central nervous system is not the result of endogenous metabolism, but is the result of intestinal bacterial metabolism, since tryptamine crosses the blood-brain barrier readily. Young and coworkers [86] showed that part of the cerebral tryptamine is Biochemistry of Tryptophan 121 bacterial in origin, as is the indolepyruvic acid found in the brain, but they suggested that the indoleacetic acid in the brain, and part of the tryptamine is endogenous in origin - it persists after sterilisation of the gastro-intestinal tract with neomycin and sulfasuccidine. To date, no-one seems to have performed any studies on germ-free animals, which would demonstrate conclusively whether the tryptamine, phenylethyLamine and tyramine found in the brain are of mammalian (endogenous brain) origin or the result of intestinal bacterial and fungal action. The evidence from Young's studies [863 is strongly suggestive that, although activity of the decarboxylase towards tryptophan cannot be demonstrated in vitro, nevertheless in vivo there is a significant amount of decarboxylation of tryptophan in the brain. It remains unclear whether this tryptamine has any physiological function or whether it is simply an 'accidental' component of the central nervous system.

TryptophanMetabolismin the Pineal Gland For many years it was thought that the calcification of the human pineal gland that occurs progressively from puberty onwards meant that the gland had Little or no function in the adult. However, there has been a considerable advance in our knowledge and understanding of the metabolism of the pineal over the last decade. This has been the result, in the main, of the development of highly sensitive and specific methods of analysis, mostly depending on gas chromatography and mass spectrometry, and thus combining a high degree of sensitivity with considerable precision in the identification of metabolites.

Melatonin is formed from serotonin by a two step reaction, shown in Figure 4. It is generally assumed that the first step, the N-acetylation of serotonin, is rate-limiting. The activity of serotonine N-acetyltransferase shows a diurnal variation in the pineal that corresponds to the variation in concentrations of melatonin in the central nervous system and the circulation through the light-dark cycle [87]. These effects are prevented by the administration of cycloheximide, suggesting that they are the result of new synthesis of enzyme protein [88].

The synthesis of melatonin is greatest during the hours of darkness, but light does not seem to be the sole controlling factor. If adult rats are maintained in constant darkness, the cyclic variations in serotonin N-acetyltransferase activity persist, although there is a gradual increase in the period, of the order of one hour per week. Rats kept in constant light show a similar persistence of the diurnal rhythm, with an increase in the period of the order of three hours per week [87]. The normal dark-induced increase in the synthesis and activity of serotonin N-acetyltransferase can be prevented by removal of the superior cervical ganglia, or by treatment of the animals with propranolol (an adrenergic blocking agent) or reserpine (which depletes serotonin and catecholamines), while administration of the adrenergic agonist isoproterenol enhances the dark effect. This suggests that there is catecholaminergic control of pineal metabolism [88].

The amplitude of the circadian variation in serotonin N-acetyltransferase activity differs considerably in different species. In the rat there is as much as a 60-fold increase during the hours of darkness, while in the hamster and gerbil (which are also nocturnal animals) it is only 3-fold, and in the (diurnal) guinea-pig it is only 1.5-fold [89]. ]22 D.A. Bender

.. CH~-CH£NH2 N Jj serotonin

H acetyl CoA.~ a CoAi$ 0 I! CHTCH2- NH- C'CHj N-ace tyl serotonin

H S-adenosylmethionin~ b

S -adenosyl homocysteine O H3CO~ CH7 CH~ NH-C-CH3 melatonin

(a) Serotonin N-acetyltransferase (acetyl CoA: arylamine N-acetyltransferase, EC 2.3.1.5)

(b) Hydroxyindole 0-methyltransferase (S-adenosyl-L-methionine: N-acetylserotonin O-methyltransferase, EC 2.1.1.4)

Figure 4

The synthesis of melatonin from serotonin in the pineal Biochemistry of Tryptophan 123

The second enzyme of melatonin synthesis, hydroxyindole O-methyltransferase, is also subject to changes in activity as a result of changes in exposure to light. In rats that have been kept in constant darkness there is a 2 to 3-fold increase in the activity of the enzyme, and also in the amount of enzyme protein present, as measured by immuno-titration [90]. Interestingly, the enzyme is inhibited by scotophobin, the peptide that is produced in the brains of animals that have been trained to avoid darkness [91].

In man, as in laboratory animals, there is a circadian variation in the concentration of circulating melatorrin. There is a 3-fold increase from the minimum during daytime to the peak at night. The concentration begins to rise before the onset of darkness and sleep, again evidence for an endogenous 'zeit-geber', and not simply a response to the level of illumination [92].

In a number of animals, pineal activity and the release of melatonin into the circulation seems to be associated with seasonality and the sensitivity of breeding to day-length. In the white-foot mouse, exposure to a regime of constant short days leads to a delay in sexual maturation; ovaries in female mice are smaller than normal, and the onset of spermatogenesis in the male is delayed. Similar effects are seen if the animals are given subcutaneous implants of melatonin in a suitable carrier, so as to ensure constant slow release into the circulation [93]. In the Syrian golden hamster, which does not breed during the winter in captivity in northern Europe, there seems to be no effect of short day-length on the development of sexual maturity, but exposure to less than about 12.5 hours of light per day leads to a fall in testicular weight and a failure of spermatogenesis in mature males. In females there is a failure of ovulation. The inhibition of breeding by short day-length is abolished by pinealectomy [94].

It has been suggested that measurement of the circulating concentrations of melatonin, and of excretion of the principal melatonin metabolites (and perhaps especially 6-hydroxymelatonin sulphate) may reflect pineal metabolism [95]. However, there is a relatively large amount of melatonin production in tissues other than the pineal. The nocturnal peak of melatonin excretion is not wholly abolished by pinealectomy [96] and although pinealectomy smooths out the diurnal variation in plasma concentrations of melatonin, the mean concentration falls by only about 80% [97].

A number of other tryptophan metabolites are also found in the pineal, including acetyl and other derivatives of 5-hydroxy- and 5-methoxytryptophol. There is some evidence that these also show circadian variation. 5-Methoxy- tryptophol is detectable in 90% of rat pineals collected at night, but only in 30% of those collected during daylight [98]. By contrast, the circulating concentration of 5-methoxytryptophol is highest during the hours of daylight. To date, the possible physiological functions of these methoxyindoles remain obscure, but it is an area of active research. They are found in areas of the brain as well as in the pineal, and, at least in relatively high concentrations, different methoxyindoles have been demonstrated to cause regression of the gonads, antagonise the action of acetylcholine on muscle and inhibit lymphocyte transformation induced by lectins (phytohaemagglutinins).

It has been suggested [99] that pineal metabolism may be important in the aetiology and pathology of schizophrenia. Schizophrenics show low plasma concentrations of melatonin, and a lower than normal amplitude of diurnal variation. Furthermore, the formation of small amounts of 5-methoxy-N,N- dimethyl tryptamine has been demonstrated in the pineal; this is one of the 124 D.A. Bender 'abnormal' compounds that have been identified in the urine of schizophrenic patients, and is known to have hallucinogenic and psychotomimetic properties [1003.

Monoamine Oxidase

The principal metabolism of serotonin in the central nervous system is oxidation, catalysed by monoamine oxidase and aldehyde dehydrogenase, to yield 5-hydroxyindole acetic acid. Unlike the catecholamines, methylation of the hydroxyl group does not seem to be an important part of the inacti- vation of serotonin. Hydroxyindole O-methyltransferase seems to be important mainly in the pineal in the synthesis of melatonin, although, as noted above, there are a number of other methoxyindoles in other regions of the brain.

There are two isoenzymes of monoamine oxidase in the central nervous system. MAO-A acts preferentially on serotonin, with a K m of 178 pmol/l and a Vma x of 0.73 nmol/min/mg protein. MAO-B acts preferentially on phenylethylamine and the catecholamines; it has a K m for serotonin of 1170 pmol/l and a Vma x of only 0.09 nmol/min/mg protein [101]. This means that it is unlikely that under normal conditions MAO-B will act on serotonin to any significant extent. Fowler and Tipton [101] have shown that serotonin will inhibit the oxidation of phenylethylamine by MAO-B with a K i of the order of 1400 ~mol/1, which is very close to the observed K m of this isoenzyme for serotonin. This concentration of serotonin is of the same order of magnitude as might arise in the brain after inhibition of MAO-A with the specific inhibitor clorgyline, and when this MAO inhibitor is used therapeuti- cally, MAO-B will be the only enzyme available for the oxidation of serotonin.

As discussed below, the increase in central nervous sytem amines after the inhibition of MAO has been central to the development of the amine hypothesis of affective disorders. The development of inhibitors that are specific for one or the other of the isoenzymes (clorgyline for MAO-A and deprenyl for MAO-B) has aided the dissection in vivo of serotoninergic and catecholaminergic mechanisms, and the parts that they play both in the affective disorders and in normal patterns of behaviour.

The product of monoamine oxidase action on serotonin is 5-hydroxyindole- acetaldehyde. Normally this is oxidised further to 5-hydroxyindoleacetic acid by aldehyde dehydrogenase. Measurement of the excretion of 5-hydroxyindoleacetic acid is a useful index of serotoninergic activity. Under some metabolic conditions, when the redox potential of the brain is shifted to a more reducing state than normal, as, for example after the ingestion of alcohol, a significant proportion of the aldehyde may be reduced to 5-hydroxytryptophol, which is also excreted in the urine.

The Uptake of Tryptophan into the Brain

The observation, discussed above, that the hydroxylation of tryptophan is the rate-limiting step in the synthesis of serotonin in the central nervous system, and that the rate of hydroxylation depends on the availability of tryptophan in the brain, has led to a considerable interest in the mechanism of transport of tryptophan across the blood-brain barrier, and investigation of factors that may affect this uptake. Biochemistryof Tryptophan 125

CH2 CH2 NH2

N ~ serotonin al H

~. CH2 CHO

N /5-hydroxyindole acetaldehyde / H

HO~,~"'~ -CH 2 CH2 OH NO~ N,~-CH2 COOH

H H 5-hydroxytryptophol 5-hydroxyindoleacetic acid

(a) Monoamine oxidase (amine: oxygen oxidoreductase (deaminating) EC 1.4.3.4.)

(b) Aldehyde dehydrogenase (Aldehyde: NAD + oxidoreductase, EC 1.2.1.3)

Figure 5 The catabolism of serotonin 126 D, A. Bender A number of early observations showed that peripheral tryptophan metabolism may affect serotonin synthesis in the brain. Curzon and Green [1023 showed that following induction of tryptophan oxygenase in the Liver by the administration of glucocorticoids, there was a significant reduction in the rate of serotonin turnover, and in the steady state concentration of serotonin in the central nervous system. Conversely, a number of studies have shown that the administration of loading doses of tryptophan leads to an increase in the rate of serotonin synthesis, with an increase in the steady state concentration of the amine in the brain, and a greater increase in the concentration of 5-hydroxyindoleacetic acid in the brain, cerebrospinal fluid and urine, indicating a considerable increase in the rate of serotonin turnover [103]. Tagliamonte and coworkers [1043 showed that a number of drugs that are known to increase the rate of serotonin synthesis aLso increase the steady state concentration of tryptophan in the brain. ALL of this is compatible with the evidence discussed above that the Km of tryptophan hydroxylase is higher than the steady state concentration of tryptophan in the brain, so that the increase in the availability of subs,rate will Lead to an increase in the rate of hydroxylation [423, and hence in the rate of serotonin turnover.

Little is known of the enzymology of amino acid transport across membrane barriers. Meister and coworkers [1053 have proposed a cyclic formation of x-glutamyl pep, ides from glutathione, followed by hydrolysis and resynthesis of glutathione. This x-glutamyl cycle is an attractive hypothesis, and fits much of the available evidence; however, it is not universally accepted [1063.

More is known of the specificity of amino acid transport systems, regardless of the enzymology of the process. The mechanisms in the intestinal mucosa, renal tubule and at the blood-brain barrier seem to have a similar specificity. A common carrier transports not only tryptophan, but also phenylalanine, tyrosine, threonine, leucine, isoleucine and valine - the so-called large neutral amino acids. In the brain this common carrier system was first demonstrated in brain slices [107], and has since been demonstrated at the blood-brain barrier by single pass perfusion studies [108].

The rate of amino acid uptake into the brain from the bloodstream is extremely high. As much as 30-50% of a labelled amino acid present in an infusion bolus may be extracted during a single pass through the brain capillary bed [108]. This occurs without any change in the net pool size of the amino acid in the brain, so there must be considerable counter-transport as well. There appears to be a rapid exchange of amino acids across the blood- brain barrier, and much of this may be carrier-mediated rather than active transport [109]. However, there are also active transport systems for the net efflux of amino acids from the brain into the circulation. These were first demonstrated by Lajtha and Toth [110, 111], who showed that after direct injection into the brain, different amino acids were lost to the circulation at different rates. They also showed that this efflux could occur against a concentration gradient, and was therefore by a process of active transport.

Altogether there seem to be at least ten different transport systems for different groups of amino acids in brain tissue, concerned with uptake from the bloodstream, efflux from the central nervous system, exchange across the blood-brain barrier (counter-transport rather than active transport) and uptake into specific cells and regions of the brain. These systems can be differentiated by their differing sensitivities to inhibition by substrates Biochemistry of Tryptophan 127 and analogues [112]. The transport of amino acids into tissues has been reviewed briefly by Bender [106], and the kinetics of transport and cross- substrate inhibition between amino acids by Pardridge [113]. The uptake of tryptophan into the central nervous system was discussed in considerable detail at a meeting in Lausanne in 1978 [114]. Two major factors seem to affect the rate of tryptophan accumulation in th~ central nervous system: the concentrations in the bloodstream of amino acids that compete with tryptophan for transport, and the binding of tryptophan to serum albumin, which reduces the availability of tryptophan for transport.

The binding of tryptophan to serum albumin

Tryptophan is unique among the amino acids in that it is bound non-covalently to serum albumin, a phenomenon that was first observed and studied by McMenamy and Oncley [115]. The binding is specific, and only a limited number of analogues and metabolites of tryptophan will compete. The equilibrium is a rapid one and such that under physiological conditions only about 10% of tryptophan is freely diffusible; the remainder is associated with and bound to albumin. Albumin binding is specific for the L-isomer; D-tryptophan is not bound. This has been exploited for the resolution of racemic mixtures of tryptophan by chromatography on columns of immobilised albumin [116]. McMenamy and Oncley [115] originally studied the binding of tryptophan to albumin by equilibrium dialysis. Other workers have also used dialysis, either using specially constructed dialysis chambers to permit very small volumes of serum to be used (as little as 60 ~l of mouse serum has been studied in this way [117]), or, when relatively large amounts of serum have been available, by reverse phase dialysis, whereby a small volume of dialysate is suspended in a dialysis sac in a relatively large volume of serum [118]. Other workers have used gel filtration or ultrafiltration, either with centrifugation or under pressure. In general, these methods give closely comparable results for the extent of protein binding of tryptophan under different conditions, and indeed it seems to make little difference whether dialysis studies are carried out at 4"C, 37"C or at uncontrolled 'room temperature', or whether the atmosphere is modified so as to maintain constant pH or not [119]. Nevertheless, it has been agreed by workers in the field that for future studies it would be preferable to adopt a standard method, ultrafiltration under centrifugation, so as to permit more ready comparison of results from different laboratories [114].

Non-esterified fatty acids displace tryptophan from albumin binding. The binding sites for tryptophan and fatty acids are separate, but there seems to be some degree of overlap or interaction between them [120]. This may be important under physiological conditions. The addition of palmitate over the physiological range of 0.3 - 1.2 mmol/l leads to an increase in the percentage of tryptophan that is free rather than albumin-bound, and changes in the dissociation constant of the tryptophan-albumin complex [117]. Further- more, the percentage of tryptophan that is free increases with the concentration of non-esterified fatty acids in response to fasting and exercise, and falls following the intake of food [Bender, D.A., unpublished observations].

A number of drugs will also displace tryptophan from albumin binding. Such compounds as salicylates [118], indomethacin, benzoate and clofibrate [121, 122] displace tryptophan both in vivo and when added to albumin or serum in vitro. Other drugs, including a number of phenothiazines and other drugs used in the treatment of schizophrenia lead to displacement of tryptophan 128 D.A. Bender from albumin binding in vivo, but have no effect when added in vitro, suggesting that they act by an indirect mechanism, or that it is a metabolite rather than the parent drug that is responsible for the effect [123].

Curzon and coworkers have shown the importance of the albumin binding of tryptophan in controlling the rate of uptake of tryptophan into the brain, and hence the rate of serotonin synthesis. A number of experimental manipulations that increase the plasma concentration of non-esterified fatty acids, and therefore displace tryptophan from binding, also lead to an increase in the rate of uptake into the brain. Similarly, insulin and nicotinic acid, both of which cause a lowering of plasma non-esterified fatty acid concentrations, also increase the binding of tryptophan to albumin and reduce the rate of uptake into the brain [124, 125, 126]. Salicylates displace tryptophan from albumin directly [118] and also lead to an increase in brain uptake of tryptophan and serotonin synthesis [127]. After the administration of clofibrate to rats there is a reduction in the total concentration of tryptophan in serum, but an increase in the concentration that is freely diffusible, and an increase in the rate of uptake into the brain [122].

The equilibrium of tryptophan binding to albumin is very rapid, and it has been suggested that because of this binding will not be an important factor in controlling the uptake of tryptophan into the brain. Much of the bound tryptophan would be stripped off albumin during a single pass through the brain capillary bed [128]. This suggests that the factor that will be important in determining the rate of uptake into the brain will be the total concentration of tryptophan in the bloodstream, rather than that which is freely diffusible, and not bound to albumin.

However, this may not be so. Although the albumin-bound tryptophan represents a considerable buffer pool that can replenish the free pool of the amino acid and is considerably greater than the free pool, what is immediately available to the transport system, and is competing with the other large neutral amino acids for transport, is the 10% or so of serum tryptophan that is free. The role of albumin-bound tryptophan is to maintain this pool of free tryptophan at a more or less constant level despite uptake. Pardridge [113] has shown that the dissociation constant of the tryptophan-albumin complex is of the same order of magnitude as the K m for tryptophan transport, so that these two would compete for free tryptophan. This suggests that although the albumin-bound pool of tryptophan would be available to replenish the pool of free tryptophan, it would not be 'apparent' to the transport system.

Further evidence of the importance of albumin binding in controlling the uptake of tryptophan into the nervous system has come from the perfusion studies of Etienne and coworkers [129]. Inclusion of albumin in the infusion bolus used in single pass perfusion studies of amino acid uptake leads to a considerable decrease in the rate of tryptophan uptake, although it does not affect the uptake of tyrosine, which is not bound to protein. Similarly, Smith and Pogson [130] have shown that the binding of tryptophan to albumin is important in the uptake of the amino acid into isolated rat hepatocytes. They showed a considerable inhibition of tryptophan metabolism by isolated hepatocytes when (bovine) serum albumin was added to the culture medium in the presence of physiological concentrations of tryptophan (of the order of 100 ~mol/l), but not when there was a considerable excess of tryptophan (500 ~mol/1), when albumin would be saturated with tryptophan, and the concentration of free amino acid would be greater than the K m of the transport Biochemistry of Tryptophan 129 system. In the presence of 2% albumin and 100 ~mol/l tryptophan, the addition of palmitate to displace tryptophan from albumin increased the rate of oxidation of tryptophan by isolated hepatocytes.

Competition for uptake by other neutral amino acids

The affinities of the blood-brain barrier amino acid transport system for the various large neutral amino acids, their potencies as inhibitors of transport of other amino acids and their concentrations in the bloodstream are all closely similar. This suggests that competition between amino acids will be an important factor in determining the rate of uptake into the brain, and therefore not only will the concentration of (free) tryptophan be important in regulating serotonin synthesis, but also the concentrations of those amino acids that compete with tryptophan for transport [113].

This competition is seen in such conditions as phenylketonuria and maple syrup urine disease, where, if the disease is not controlled by dietary restriction, the concentrations of some amino acids in the bloodstream are grossly abnormal. There is some evidence of defective serotonin synthesis as a result of increased competition from phenylalanine in uncontrolled phenylketonuria and the excess of leucine, isoleucine and valine in uncontrolled maple syrup urine disease. This competition may also be important under normal conditions. Fernstrom and Wurtman [131, 132] have shown that the rate of uptake of tryptophan into the brain under various conditions of fasting and refeeding depends not only on the concentration of tryptophan, but also on the ratio of tryptophan:total neutral amino acids.

Fernstrom and Wurtman [131, 132] have not measured the binding of tryptophan to serum albumin in their studies, and indeed, very few workers have measured both albumin binding and the tryptophan : competing amino acid ratio under the various metabolic and other conditions that affect the uptake of tryptophan into the brain. Where such studies have been performed, it seems that both factors are important as predictors of brain uptake and serotonin synthesis, and perhaps the most useful index is the ratio of freely diffusible tryptophan: competing amino acids [114].

Changes in tryptophan metabolism in response to food intake - cues to appetite regulation

Fernstrom and Wurtman [131] showed that in response to refeeding fasted animals there was an increase in the rate of uptake of tryptophan into the brain, and an increase in the rate of serotonin synthesis. They showed that there was an increase in the concentration of tryptophan relative to competing amino acids, at least partly as a result of insulin secretion, which increases the tissue uptake of the branched chain amino acids (leucine, isoleucine and valine), which are important energy-yielding substrates in muscle and other tissues. At the same time, there is a high concentration of non-esterified fatty acids in the bloodstream (the normal effect of fasting), which will lead to increased displacement of tryptophan from albumin binding [117, 133]. Later, fatty acid levels fall, and so does the concentration of free tryptophan, although the total concentration remains high. The uptake of tryptophan into the brain also falls [G. Curzon, personal communication].

At least a part of the increased tryptophan in the bloodstream under these conditions may be released from the pancreas together with insulin. Bender and coworkers [134] have shown that in fasting rats there is a considerable accumulation of tryptophan (but not other amino acids) in the pancreas - up to 2-3 times the concentration that is found under non-fasting conditions. 130 D.A. Bender

In other tissues, the increase in free tryptophan in fasting is only some 20% above the concentration that is found in the fed state. This tryptophan seems to be accumulated specifically in the ~-cells of the pancreas, and is released into the bloodstream when glucose is administered, apparently at the same time as insulin is released. The effect is abolished in animals in which the ~-cells have been destroyed by streptozotocin or alloxan. These observations have led to investigations of whether changes in serotonin synthesis as a result of food intake may be connected with the hypothalamic satiety response, since serotonin is the only transmitter for which there is uncontrovertible evidence of changes in response to food intake. There is also some pharmacological evidence that serotonin may be involved in the mechanism of satiety. Noble [135] showed that the serotonin (and histamine) antagonist cyproheptadine caused a significant increase in food intake and a gain in body weight when it was given to healthy underweight adults. In rats, the direct intra-cerebral injection of p-chlorophenylalanine, in such a way as to cause a considerable reduction in serotonin synthesis with little effect on catecholamines, causes hyperphagia and the development of obesity [136]. Destruction of serotoninergic neurons by the intra-cerebral injection of 5,7-dihydroxytryptamine does not cause hyperphagia unless catecholamines are protected by the administration of desmethylimipramine [137].

Depression of appetite and frank anorexia can be caused in experimental animals by the administration of amphetamine or fenfluramine. Indeed, fenfluramine has been developed as an anorectic agent for the treatment of obesity in man, and at one time amphetamine was used for the same purpose. The mechanism of action of amphetamine is complex, and it is not possible to determine on which system it exerts its primary anorectic actions; it acts on both serotoninergic and catecholaminergic neurons. The actions of fenfluramine seem to be more specifically on serotoninergic systems; lesions of the serotoninergic raphe nucleus antagonise the anorectic effect of fenfluramine [138] while the administration of 5-hydroxytryptophan potentiates it [139].

The administration of tryptophan alone has no effect on food intake [140, 141], but in combination with the monoamine oxidase inhibitor iproniazid, tryptophan causes a significant reduction of food intake, correlated with the increase in brain serotonin. This effect is antagonised by the serotonin antagonist methysergide, as is the anorexia produced by fenfluramine or mazindol [141].

The administration of carbon tetrachloride to rats leads to an increase in the proportion of serum tryptophan that is freely diffusible, without altering the total serum concentration of the amino acid, and this is accompanied by an increase in the brain content of serotonin and 5-hydroxyindoleacetic acid, and a reduction in food intake [142]. It is difficult to determine whether this anorexia is the result of changes in brain serotonin metabolism, or a more general effect of carbon tetrachloride intoxication. In long-term studies, Ashley and Curzon [143] have shown that rats fed from weaning on diets deficient in tryptophan, which have relatively impaired serotoninergic function in the central nervous system, have a greater food intake than do animals fed on tryptophan-supplemented diets.

Ashley and Anderson [144, 145] have reported a series of studies in which rats were permitted free choice of foods. Manipulation of the plasma concentrations of amino acids so as to alter the ratio of tryptophan : neutral amino acids led to changes in the voluntary intake of protein-rich Biochemistry of Tryptophan 131 foods, with reduced intake of protein when the synthesis of serotonin would be expected to be higher. There was no effect on total food intake in response to changes in tryptophan. However, they also refer to preliminary studies in which the voluntary intake of dietary energy was affected by manipulations of the ratio of phenylalanine : neutral amino acids (which may have an effect on the rate of catecholamine synthesis). It seems likely, from the evidence cited above, that serotonin is concerned in the hypothalamic satiety response, and in the regulation of feeding behaviour. It is possible that the future development of appetite suppressant drugs will concentrate on more specific serotonin agonists, and, conversely serotonin antagonists may be useful in the treatment of anorexia nervosa, to overcome early satiety, and hence stimulate appetite.

Tryptophan uptake into the brain in liver failure

No satisfactory explanation has been advanced to account for the coma that accompanies ammonia intoxication in liver failure. It has been suggested that the increased cerebral concentration of ammonium ions leads to a depletion of 2-oxo-glutarate, by formation of glutamate and glutamine, as a means of 'detoxifying' the ammonia. This in turn leads to reduction of the capacity of the citric acid cycle, and hence a reduction in the energy- yielding metabolism of the brain [146].

More recently, an alternative explanation has been proposed, based on the observation of altered uptake of tryptophan into the brain both in clinical liver failure and in various animal models. Patients dying in fulminant hepatic failure have a very much higher than normal proportion of plasma tryptophan freely diffusible rather than bound to serum albumin, and at the same time there are higher than normal concentrations of tryptophan, serotonin and 5-hydroxyindoleacetic acid in the brain, and a very high concentration of 5-hydroxyindoleacetic acid in the cerebrospinal fluid [147]. These findings are compatible with a reduction in the rate of removal of tryptophan from the circulation by the liver, and hence enhanced cerebral uptake of the amino acid. In cirrhotic patients, the presence and severity of encephalo pathy correlates well with the elevation of diffusible plasma tryptophan [148]. Baldessarini and Fischer [149] have suggested that the deleterious effect of increased serotonin synthesis might be not so much an increase in the activity of what are normally serotoninergic neurons, but also some degree of adventitious formation of serotonin in other brain regions, where it might have a 'false transmitter' role.

Animal studies have used three models of liver failure, all of which show effects on the uptake of tryptophan into the brain that appear to be similar to those seen in patients with hepatic encephalopathy:

(i) Curzon and coworkers [150] used hepatic devascularisation of pigs as a model of acute liver failure. In the 5-8 hours that the animals survived after the operation there was a significant increase in the plasma concentration of non-esterified fatty acids, accompanied by an increase in the proportion of tryptophan that was freely diffusible, although there was no change in the total concentration of tryptophan in the bloodstream. This was accompanied by an increase in the uptake of tryptophan into the brain and other tissues.

(ii) Acute intoxication of rats with carbon tetrachloride has a similar effect on the plasma concentration of non-esterified fatty acids, and 132 D.A. Bender

again leads to an increase in the uptake of tryptophan into the brain, and an increase in the rate of serotonin turnover [142].

(iii) The third animal model that has been used is portacaval anastomosis, which seems to be a good model for chronic liver failure, since the animals can be kept alive for a considerable time. Curzon and c~workers [151] have shown that three weeks after operation in rats there is a considerable increase in the brain content of tryptophan and 5-hydroxyindoleacetic acid, with a smaller, but significant, increase in the concentration of serotonin. There was also some evidence of an increase in the activity of the transport mechanism for neutral amino acids; the uptake of tyrosine into the brain was also increased, although to a lesser extent than the uptake of tryptophan, and both tryptophan and tyrosine were increased in muscle. Further evidence for an increase in the rate of neutral amino acid uptake has come from the studies of Mans and coworkers [152], who studied the uptake of labelled amino acids into the brain in rats that had undergone portacaval anastomosis 7-8 weeks previously. They showed a 200% increase in tryptophan clearance from the bloodstream, an 80% increase in phenylalanine clearance, and a 30% increase in leucine clearance. This last occurred despite a decrease in the circulating concentration of leucine, and therefore must be attributed to an increase in the activity of the transport system; the effects on tryptophan and phenylalanine uptake could be accounted for in part by the reduction in leucine, isoleucine and valine concentrations after anastomosis. The administration of leucine has been used successfully to speed the recovery of consciousness in patients in heptatic coma [153].

Munro and coworkers [153] suggested that the reduction in circulating branched chain amino acids in liver failure was the result of increased circulating concentrations of insulin, resulting from reduced liver clearance of the hormone. This would enhance the uptake of branched chain amino acids into tissues for both protein synthesis and energy-yielding metabolism, and therefore would lead to lower circulating concentrations, and hence reduced competition for the uptake into the brain of tryptophan and other large neutral amino acids. Insulin also increases the activity of the blood-brain barrier transport system for large neutral amino acids [154].

Jejuno-ileal anastomosis in man leads to changes in plasma amino acids similar to those seen in liver failure and after portacaval anastomosis in animals, and has a similar effect on brain serotonin synthesis. A condition that has been called 'hepatic coma' has been reported as one of the side- effects of gut bypass surgery for obesity in man, with a 2.3% death rate in one series of patients studied [155].

James and coworkers [156] have proposed a unifying mechanism that may explain much of the observed biochemistry of both experimental and clinical liver failure. They suggest that as the capacity of the liver to synthesize urea is progressively impaired, episodes of hyperammonaemia become more frequent. The increase in blood ammonium ion concentration leads to an increase in the secretion of glucagon, which promotes gluconeogenesis from amino acids. This in turn leads to an increase in the amount of ammonia that must be metabolised, so that once it is established hyperammonaemia is self-sustain- ing to a great extent. The increased synthesis of glucose leads to hyper- glycaemia, which stimulates the secretion of insulin, at a time when the capacity of the liver to clear the hormone from the portal circulation is impaired. This means that relatively high circulating concentrations of insulin develop. One of the actions of insulin is to increase the uptake of branched chain amino acids into tissues for energy-yielding metabolism, a Biochemistry of Tryptophan 133 impaired urea /' g~ synthesis ]~ pancreas • NH a gluca NH3"~~~ /

ose kidn insulin

m/scle /

t tryptophanealbumin adipose

~ fatty acids J (-)~ ~ ~free tryptophan

brain ~

Figure 6 Metabolicchanges leading to increaseduptake of tryptophaninto the brain in liverfailure 134 D.A. Bender process that will increase yet further the burden of ammonia to be metabolised. This clearance of leucine, isoleucine and valine leads to an imbalance of the tryptophan : neutral amino acids ratio in the bloodstream, and therefore enhances the uptake of tryptophan into the brain and other tissues. At the same time the circulating concentrations of non-esterified fatty acids are higher than normal, as a result of both glucagon stimulation of lipolysis and reduced hepatic metabolism of fatty acids; this will result in decreased binding of tryptophan to albumin, and hence will also enhance the uptake of tryptophan into the brain and other tissues.

The increased formation of glutamine in the brain may have an effect on tryptophan uptake, as well as leading to depletion of 2-oxo-glutarate, and hence reducing the brain's capacity ~or energy-yielding metabolism [146]. The efflux of glutamine from the brain in largely by means of a counter- transport system that transports tryptophan in at the same time as glutamine is transported out of the central nervous system [156].

Sleep and Tryptophan Metabolism

A series of elegant studies from Jouvet's laboratory (reviewed in [157]) showed that serotonin turnover is important in the normal processes involved in sleep. Either electrical destruction of the serotoninergic raphe system of the brain or the administration of p-chlorophenylalanine as a relatively specific means of inhibiting serotonin synthesis leads to a decrease in the total amount of slow-wave sleep and a decrease in the frequency of episodes of rapid eye movement sleep (REM or paradoxical sleep) in experimental animals. The severity of these effects is correlated with the extent of destruction of the raphe system, and the degree of depletion of brain serotonin, as determined at post mortem examination. The administration of 5-hydroxytryptophan as a precursor of serotonin leads to restoration of normal amounts of slow-wave sleep after the administration of p-chlorophenylalanine although it has no effect in animals in which the raphe system has been destroyed [158].

If rats are deliberately deprived of REM sleep, there is an increase in the rate of serotonin synthesis and catabolism in the central nervous system. This appears to be the result of both an increase in the uptake of tryptophan into the brain and an increase in the rate of hydroxylation. There is no change in the rate of serotonin turnover when 5-hydroxytryptophan in injected directly into the brain, although there is following intracerebral injection of tryptophan, showing that there is an increase in the activity of tryptophan hydroxylase in the brain [159].

There is also evidence of the importance of tryptophan and serotonin metabolism in sleep mechanisms in man. Wyatt and coworkers [160] showed that the administration of p-chlorophenylalanine to patients with carcinoid syndrome, as discussed above, caused a decrease in the proportion of REM sleep. However, there was no change in the total duration of non-REM (slow-wave) sleep, a difference from the effects of such treatment in animals that cannot be explained. Similarly, Chert and coworkers [161] showed that there was a significant correlation between the concentration of diffusible tryptophan in plasma and the amount of REM sleep, although they also showed a negative correlation between diffusible tryptophan and non-REM sleep. Other workers have shown that the administration of tryptophan to volunteers leads to an increase in the total duration of sleep, mainly slow-wave sleep. The effects of tryptophan on REM sleep in man are inconsistent [160, 161, 162, 163]. Biochemistry of Tryptophan 135

A number of studies of the effects of administration of tryptophan to human volunteers have shown that one of the effects of a relatively high dose of the amino acid is drowsiness. This is frequently accompanied by an elevation of mood and feelings of euphoria, listlessness and increasing talkativeness as the dose increases [164].

Hartmann [165] has suggested that tryptophan may be a valuable addition to the pharmacopoeia as a hypnotic drug. This is based on its involvement in the normal processes of sleep and the observation that, unlike most conventional hypnotic drugs, it does not distort the normal electroencephalographic patterns of sleep. Furthermore, it seems to have no long-term effects and there are no adverse effects following its withdrawal. However, relatively large amounts are required (of the order of several grams), and this presents problems of administration - either several large pills must be taken, or some means must be found of masking the flavour of tryptophan to be given in solution or suspension; as noted above, tryptophan has an intensely bitter flavour.

Tryptophan and Serotonin in Psychiatric Disorders

As the neurotransmitter roles of the biogenic amines (and especially serotonin and the catecholamines) were elucidated, and as sensitive analytical tech- niques for the amines and their metabolites became available, it was obvious to investigate changes in amine metabolism as neurochemical correlates of psychiatric disease. Considerable problems arise in the interpretation of results that show abnormalities of amine metabolism in such conditions:

(i) There have been very great problems of classification of diseases and diagnosis. Although these continue to bedevil attempts at inter-laboratory collaboration and the comparison of results from different centres, use of the internationally agreed diagnostic criteria and systems of classification has been helpful. Nevertheless, there are considerable problems in the diagnosis of psychiatric illness because of the subjective nature of the diseases and the lack of truly objective criteria. Indeed, one hope is that biochemical deviations from normal will permit an alternative diagnostic system to be developed. It is possible that understanding of the neurochemistry of psychiatric illness, and of the pharmacology of the drugs available for use, may permit more precise prediction of drug response, and the development of new pharmacological agents.

(ii) Where there are abnormalities of amine metabolism in psychiatric patients, it is frequently difficult or impossible to determine to what extent these are the results of present or previous drug therapy, abnormal dietary habits or other confounding factors, rather than an integral part of the disease process. Even when these abnormalities do indeed seem to be a part of the pathology of mental illness, it is difficult to know whether the changes involved are primary, secondary to some other effect, or even epiphenomenal.

(iii) There are no acceptable animal models for psychiatric illness, although a number of drug-induced conditions in man are sufficiently similar to naturally occurring diseases for it to be possible to use these drugs in animals, at least as moderately satisfactory models.

(iv) For obvious ethical and practical reasons, human studies are limited largely to analysis of blood and urine; cerebrospinal fluid is also sometimes 136 D.A. Bender

available, although here it is difficult to obtain samples from normal control subjects for comparison. All of these body fluids are more or less remote from the area of interest, the central nervous system. Furthermore, what is apparent from changes in amines, their precursors and metabolites in body fluids is, and can only be, a reflection of the sum of whole brain metabolism, even if metabolism in other tissues can be accounted for, while it must be assumed that behavioural changes and psychiatric illness involve discrete and specific regions of the brain. It is possible that changes in amine metabolism in a limited area of the brain will not be apparent if there is no change in the rest of the brain.

(v) Where samples of post-mortem brain tissue are available for analysis, the method of storage since death, the time and method of removal, and a number of other factors will all have profound effects on the concentrations of transmitters, precursors and metabolites, and on the activities of the enzymes involved. The time of death may also be important, since there are circadian variations in neurotransmitter metabolism.

(vi) Studies of the neurochemical effects of drugs that are known to be useful in the treatment of psychiatric disease can give a considerable amount of information about the possible metabolic basis of the disease. This has been especially true in the case of affective disorders, as discussed below. However, in conditions such as schizophrenia this approach has not been particularly useful; the effective anti-psychotic drugs all have a very wide range of effects on a number of different transmitter systems, and it is not generally possible to determine which of these may be important, and which may simply represent side effects of the drugs.

Despite these problems, we have gained a considerable understanding of the neurochemistry of affective disorders, and to a lesser extent also of the psychotic disorders, during the last two decades.

Affective disorders

A major advance in our understanding of the biochemical basis of depressive illness, and later also of bipolar manic-depressive illness, came with the observation that the treatment of hypertension with reserpine led to a severe depressive reaction in some patients, very similar to endogenous depression [166, 167]. Reserpine is the principal alkaloid of the root of the Indian plant Rauwolfia serpentina, which has long been used in traditional Indian medicine as a sedative. Studies of the actions of reserpine showed that it causes a marked and prolonged depletion of catecholamines and serotonin in the central nervous system. In experimental animals the behavioural depression and sedation caused by reserpine are correlated with the degree of depletion of brain amines, and deliberate repletion by the administration of precursors leads to recovery more rapidly than would otherwise be the case [168].

The biochemical action of reserpine is on the storage of amines in vesicles both in the central nervous system and in peripheral tissues. The amines are released from storage and are made available for catabolism immediately, without release into the synaptic cleft, so that the effect of reserpine is to deplete amines without any preliminary 'spurt' of activity due to enhanced release into the synapse [168, 169]. BiochemistryofTryptophan 137 I Increased availability of tryptopha~ tryptophan increases the synthesis synthesis of serotonin. serotonin storage in vesicles

Inhibition of monoamine oxidase reduces catabolism, release permitting the reutilisation of serotonin. ~ serotonin Tricyclics reduce uptake receptor from the stimulatio synaptic cleft and potentiate serotonin action.

Figure 7 The effects of anti-depressant drugs at the serotoninergic synapse 138 D.A. Bender

Two further pieces of evidence were instrumental in the development of the 'amine hypothesis' of depressive illness:

(i) The anti-depressant drug iproniazid was shown to be an inhibitor of monoamine oxidase, and thus caused an increase in the brain content of biogenic amines [170];

(ii) At the same time the tricyclic anti-depressant imipramime was being used successfully, although it was not until some ten years after its introduction into psychiatry that it was shown to act by blocking the uptake of amines from the synaptic cleft, so that it results in an increased concentration of amines at the post-synaptic receptor, and hence potentiation of amine action [171].

Schildkraut [172] summarised much of the evidence of the involvement of the catecholamines in depressive illness, and formulated a clear hypothesis that the basis of depression was reduced synthesis of amine transmitters. Although he considered primarily the catecholamines, much of the evidence he cited applies equally well to serotonin. Glassman [173] considered that it was possible to make as strong a case for the involvement of reduced synthesis of indoleamines (principally serotonin) in the aetiology of depression as had been made for the involvement of the catecholamines. Over the years since the amine hypothesis of depression was first proposed [172, 173] thinking has changed to a great extent, and a 'modified amine hypothesis' that most workers in the field now use as the basis of their research emphasises a relative functional deficit of neurotransmitter amines rather than an absolute deficit. Thus Ashcroft and coworkers [174] suggested that in some unipolar depressive patients the problem might be one of a low rate of synthesis of amines while in others it might be one of reduced sensitivity of the amine receptors to normal levels of the transmitters. In either case an increase in the availability of amines in the synaptic cleft, achieved by inhibition of monoamine oxidase, inhibition of uptake from the synapse with tricyclic and other drugs or the administration of amine precursors, would have a beneficial effect, either increasing the concentration above normal, and so permitting adequate receptor response, or enhancing synthesis to a more or less normal level. Long-term recovery from depression could then be attributed to either an increase in the sensitivity or number of post-synaptic receptors, or to an increase in the rate of amine synthesis.

Evidence for a deficit of serotonin or tryptophan in depression

A key finding in the development of the indoleamine hypothesis of depression was that of Ashcroft and coworkers [175] that the concentration of 5-hydroxyindoleacetic acid in the cerebrospinal fluid of depressed patients was significantly lower than in normal controls or recovered depressives. Bourne and coworkers [176] demonstrated that the concentration of 5-hydroxyindoleacetic acid was significantly lower in the hind-brain of depressed suicide victims than in other (control) post-mortem samples. There was no change in the content of catecholamines or serotonin, suggesting that the effect was on the rate of serotonin turnover rather than on the absolute amount in the brain. Other studies of the amine and metabolite content of post-mortem suicide brain tissue have yielded conflicting results, reflecting both the problems of selecting suitable control samples and possible difference in the post-mortem handling of tissue, as well as analytical problems associated with the relatively low specificity and poor sensitivity of fluorimetric methods of analysis. Biochemistry of Tryptophan 139

Asberg and coworkers [177, 178, 179] have confirmed the relationship between low levels of 5-hydroxyindoleacetic acid in cerebrospinal fluid and depression and attempted suicide, using sensitive and specific analysis with gas chromatography and mass spectrometry. They showed that depressed patients who had attempted suicide had significantly lower concentrations of 5-hydroxyindoleacetic acid in cerebrospinal fluid than did a group of healthy control subjects. This was more marked in those patients who had attempted suicide by active or violent means rather than by passive means. In a prospective study of a total of 120 depressive and/or suicidal patients they showed that in those whose cerebrospinal fluid concentration of 5-hydroxyindoleacetic acid was below the median there was a 20% probability of death by suicide within 12 months. They have suggested [179] that about 30% of the depressive patients they have studied can be classified into a sub-group, on the basis of their apparently lower rate of serotonin turnover. Within this sub-group of patients, suicide and attempted suicide are considerably more common than among depressed patients as a whole, and the attempts are frequently of a more positive and determinedly active nature. They suggest that the measurement of cerebrospinal fluid concentrations of 5-hydroxyindoleacetic acid may be useful in deciding on patient management, and ultimately, as more specific serotonin agonists become available, in the prescription of drug therapy [177, 178, 179].

A number of studies have demonstrated an apparent decrease in the availability of tryptophan for uptake into the brain in depression, with lower than normal concentrations of either total or freely diffusible tryptophan in plasma, relative to the concentrations of amino acids that compete with tryptophan for transport. During prolonged haemodialysis there is a fall in plasma tryptophan that is correlated with depression [180]. Bipolar manic-depressive patients with a short cycle show a cyclical variation in plasma tryptophan, with coincidences between high plasma tryptophan and manic behaviour, and between low plasma tryptophan and depressive phases. This is less apparent in patients with relatively long cycles of depression and mania [181]. Similarly, in unipolar recurrent depressive patients Coppen and coworkers [182] showed that diffusible plasma tryptophan was lower than normal during depression (although the total concentration of tryptophan was more or less normal) and rose to normal on recovery, regardless of the type of therapy used. In women during the first 2-5 days after giving birth there is a significant correlation between lower than normal concentrations of plasma diffusible tryptophan and the incidence of post-natal depression [183].

In peri-menopausal women, the concentration of plasma diffusible tryptophan is considerably lower than in older or younger women, or men of any age. There is a significant inverse correlation between the concentration of tryptophan and the degree of depression as assessed by the Hamilton and other rating scales. During periods of remission of depression the concentration of diffusible tryptophan increases, apparently the result of endogenous secretion of oestrogens, and there is a similar increase in diffusible tryptophan, together with alleviation of depression, during hormone replacement therapy with piperazine oestrone sulphate [184]. In unpublished studies, Bender and coworkers have demonstrated a significant concordance between the alleviation of depression and an increase in the total concentration of tryptophan in plasma during menopausal hormone replacement therapy with either piperazine oestrone sulphate or ethinyl oestradiol.

In Handley's study of post-natal depression [183], those patients who were least subject to depression, and had the highest circulating concentrations of tryptophan, also had the lowest circulating concentrations of cortisol. 140 D.A. Bender

This may be an important factor. Cortisol will induce tryptophan oxygenase, the first and rate-limiting step of tryptophan oxidative metabolism (see Chapter 3) and will therefore reduce the amount of tryptophan that is available for serotonin synthesis. The administration of cortisol to rats leads to a depletion of brain serotonin, apparently as a result of this induction of tryptophan oxygenase [102]. Lapin and Oxenkrug [185] suggested that the secretion of cortisol and its effects on tryptophan oxygenase may be the key to the biochemistry of depression. Cortisol secretion is increased in depression, and this will lead to increased tryptophan oxygenase activity, and hence reduced serotonin synthesis, which will lead to deeper depression and increased secretion of cortisol - a biochemical spiral to match the behavioural spiral of endogenous depression. Tricyclic anti-depressants reduce the activity of tryptophan oxygenase in the liver of rats after administration in vivo, although there is no effect in the activity of the enzyme when the drugs are added to incubations in vitro [186]. Badawy and Evans [187] suggested that this was due to complexing of the haem cofactor of the oxygenase by the drugs, and noted that a number of compounds that elevate brain tryptophan inhibit tryptophan oxygenase in an apparently similar way. The use of allopurinol as an inhibitor of tryptophan oxygenase has been suggested as potentially useful in the treatment of depression [188]. However, it has been shown that allopurinol, although an effective inhibitor of tryptophan oxygenase in vitro, has little effect on tryptophan metabolism in vivo [189].

If the induction of tryptophan oxygenase by corticosteroids is important in the aetiology of depression then it should be possible to correlate the depression, cortisol secretion and tryptophan metabolism in patients with Cushing's syndrome, where there is greatly elevated secretion of cortisol. Although the depression associated with this condition does appear to be related to cortisol secretion, and is certainly relieved by surgical treatment, there are no clear changes in the serum concentrations of total or diffusible tryptophan, nor in the urinary excretion of tryptophan metabolites, that can be correlated with this [190, 191]. Other workers have not demonstrated any clear differences of serum total or diffusible tryptophan in depression [192, 193]. Indeed, Peet and coworkers [192] showed changes in tryptophan in mania that were in the opposite direction to those reported by others, with an increase in the concentration of freely diffusible tryptophan on recovery from mania. Riley and coworkers [193] were unable to demonstrate any differences in the dissociation constant of the tryptophan-albumin complex in serum samples from depressed patients compared with normal controls.

The use of tryptophan as an anti-depressant drug

The evidence discussed above that in at least some depressed patients there is a relative deficit of serotonin in the central nervous system, and the observations of lower than normal concentrations of tryptophan in the bloodstream in depression, together with the realisation that serotonin synthesis is at least partly controlled by availability of t ryptophan for uptake into the brain, have led to a number of trials of the amino acid as an anti-depressant drug.

Pare and coworkers [194] studied a group of depressed patients who had initially responded to treatment with monoamine oxidase inhibitors, but who relapsed when the dose was reduced. Six of their 14 patients showed a significant improvement when they were treated with tryptophan in addition to a low dose of the monoamine oxidase inhibitor deprenil. However, in 5 patients the incidence of side effects such as nausea and drowsiness was severe enough Biochemistry of Tryptophan 141 to force withdrawal from the trial. Coppen and coworkers [195] also demonstrated the superiority of tryptophan together with an inhibitor of monoamine oxidase over treatment with the inhibitor alone, and noted that when schizophrenic patients were treated with this combination of drugs there was an elevation of mood almost to the point of euphoria. Tryptophan may also be useful in conjunction with tricyclic anti-depressants such as amitriptyline, although the enhanced anti-depressant action of the combination may not be great enough to outweigh the increased incidence of unwanted side effects, so that there is little therapeutic advantage in the combination [196]. Shaw and coworkers found that the addition of tryptophan to treatment with tricyclic anti-depressants did not give any additional therapeutic benefit in unipolar depressive patients [197].

In an open study, Coppen and coworkers [198] showed that tryptophan alone was approximately as effective as electroconvulsive therapy in the treatment of depression. Hertz and Salman [199] reported a single patient who suffered from an annually recurring depression in whom prophylactic tryptophan led to an improvement of mood, greater emotional stability and prevented the development of the expected attack of depression. A number of other studies, generally open rather than double-blind trials, also suggested that tryptophan had a considerable potential as a very much safer and more acceptable anti-depressant than either tricyclics or inhibitors of monoamine oxidase. However, later controlled double-blind studies have in general been disapp- ointing. Herrington and coworkers [200, 201] showed that the initial response to tryptophan was as good as that to electroconvulsive therapy or amitriptyline, but the effect was not sustained. After 2-4 weeks the tryptophan- treated patients relapsed, while those receiving the conventional forms of therapy continued to improve. Murphy and coworkers [202] showed no beneficial effect of tryptophan in unipolar depressive patients, although they did claim benefits in bipolar manic-depressives. Cooper and Datta [203] showed that tryptophan was no more effective than placebo in a group of elderly depressed patients. In a review of trials of both tryptophan and 5-hydroxytryptophan as precursors of serotonin and hence as potential anti-depressant drugs, D'Elia and coworkers [204] concluded that there was no evidence of any beneficial effect of 5-hydroxy tryptophan, and that tryptophan alone was similarly ineffective. 'The only convincing evidence for tryptophan as an anti-depressant is that tryptophan enhances the effects of monoamine oxidase inhibitors', which has the advantage of permitting the use of lower doses of these potentially dangerous drugs. Carroll [205] concluded a similar review by saying that 'tryptophan has been shown to possess no significant anti- depressant properties in severely ill patients. There is some acceptable evidence that it potentiates monoamine oxidase inhibitors, but the routine use of this combination presents problems because of side effects'.

Schizophrenia

The biochemical study of schizophrenia suffers considerably from both the problem of defining the groups of patients to be studied and the problem of finding suitable groups of control subjects with whom to compare them. The majority of studies have been conducted on long-term chronically hospitalised patients, many of whom have been receiving potent anti-psychotic medication for many years, and it is almost impossible to find suitable groups of 'normal' subjects for comparison. The history of the field is a depressing catalogue of initially euphoric reports of quantitative or quali- tative differences in metabolism or the excretion of metabolites, followed by the discovery that the apparently 'unique' compounds are unrelated to schizophrenia, but may be due to such diverse causes as a high consumption 142 D. A, Bender

of instant coffee, abnormal intestinal bacterial action as the result of chronic constipation or even in one case the presence of a new plasticiser material in the polyethylene sheeting used to collect sweat from patients.

There is an underlying thread of evidence that abnormalities of tryptophan metabolism may be related to some forms of schizophrenic disease - much of the evidence has been reviewed by Gilka [206]. The depressive psychosis of pellagra, the tryptophan-niacin deficiency disease, resembles schizophrenia in many respects (see Chapter 3). A psychosis can be precipitated by the hallucinogenic drug dimethyltryptamine that is similar to some forms of schizophrenia, and lysergic acid diethylamide-induced psychosis also has some close similarities to schizophrenia. Lysergic acid diethylamide has been shown to bind to serotonin receptors in the central nervous system; both substances are displaced from binding by the same compounds, and they are mutually competitive [207]. A number of other hallucinogens have also been assumed to act by way of a serotoninergic mechanism because of a structural and stereochemical resemblance to the neurotransmitter.

Two studies have shown interesting effects of the administration of tryptophan together with inhibitors of monoamine oxidase in schizophrenic patients [208, 209]. In both cases the administration of tryptophan provoked an initial exacerbation of the condition, with considerable behavioural excitation, euphoria and garrulosity to a point described in one paper [208] as 'word salad'. This is similar to the hyperactivity syndrome that is produced in rats by a similar combination of tryptophan and a monoamine oxidase inhibitor [43]. In the patients, the initial period of excitation was followed by a rebound period during which they were somewhat improved compared with their initial state [208, 209].

Among other abnormal proteins and enzymes that have been reported as being 'unique' to schizophrenia, and therefore possibly related to the apparent genetic basis of the condition, one is of interest in relation to tryptophan metabolism. This is an ~2-globulin that differs in lipid content and secondary structure from the equivalent normal plasma protein. Among other effects, this protein has been shown to enhance the uptake of tryptophan into chick erythrocytes in an in vitro test, and it has been suggested that if it has a similar effect at the blood-brain barrier this might be important in the pathology of schizophrenia [210, 211].

In an open study, Bender and Bamji [212] showed that the total serum concentration of tryptophan was lower than normal in a group of chronic schizophrenics treated with chloropromazine than in a control group of unhospitalised normal subjects, with a considerably greater than normal proportion of tryptophan freely diffusible rather than bound to serum albumin. They showed that this effect could not be attributed to increased plasma concentrations of non-esterified fatty acids, and that the dissociation constant of the tryptophan-albumin complex was the same in the patients as in normal subjects. There appeared to be some non-competitive inhibition of binding, with less than half the normal concentration of tryptophan binding sites per mol of albumin. A number of other studies have also shown abnormalities of serum total or diffusible tryptophan in schizophrenia. Manowitz and coworkers [213, 214] showed lower than normal total serum tryptophan in newly hospitalised schizophrenic patients, and in those who responded to drug therapy there was an increase to the normal level. Lovett-Doust and coworkers [215] also reported lower than normal concentrations of total serum tryptophan in a group of schizophrenics receiving a variety of different types of medication, and Deniker and coworkers [216] showed elevated concen- Biochemistry of Tryptophan trations of diffusible tryptophan (with a more or less normal total concentration) in schizophrenic patients with or without drug therapy. Freedman and coworkers [217] demonstrated elevated plasma serotonin and a very low activity of monoamine oxidase in platelets, associated with low plasma tryptophan but an increase in the fraction that was freely diffusible in some chronic schizophrenics. However, Domino and Krause [218] were unable to show any abnormality of either total or diffusible tryptophan in schizophrenic patients who had been drug-free for 6 months.

In a longitudinal study in which chronic schizophrenics who had been receiving chlorpromazine for a number of years were transferred to placebo for four weeks, Bender and coworkers [219] showed that there was an increase in the ratio of diffusible tryptophan : total amino acids in plasma on drug withdrawal. This coincided with deterioration in the patients' condition. This would suggest that during the period of drug wash-out there was a progressive increase in the rate of uptake of tryptophan into the brain, and therefore presumably an increase in the rate of serotonin turnover. Immediately following the reintrodution of chlorpromazine there was no change in the ratio of diffusible tryptophan : total amino acids, and at this stage there was no therapeutic effect of the drug. This study suggests that relative overactivity of serotoninergic systems may be a factor in the pathogenesis of schizophrenia, and therefore it is possible that serotonin antagonists may be useful in treatment. The trials that have been reported have been dis- appointing; methysergide has no beneficial effect [220] and the results in a trial of cinanserin were inconclusive, although there was some indication of a beneficial effect [221]. To date there have been no reports of trials with the more recently developed and more specific serotonin antagonists.

There is some evidence that the clinically useful anti-psychotic phenothiazines block the serotonin receptor. Chlorpromazine will inhibit serotonin- sensitive adenyl cyclase in rat brain preparations, as will known serotonin antagonists such as cyproheptadine [222]. Similarly, chlorpromazine and fluphenazine will displace serotonin from specific membrane binding, again an effect shared by known serotonin antagonists, although promethazine, a phenothiazine without anti-psychotic activity, only does so at very high concentrations [207]. The phenothiazines with clinically useful anti-psychotic activity have effects on tryptophan metabolism in experimental animals, increasing the fraction of tryptophan that is freely diffusible, and enhancing the rate of serotonin synthesis in the brain, that are not shown by related drugs such as promethazine that do not have anti-psychotic properties [223].

Further support for the possible involvement of serotoninergic overactivity or super-sensitivity of serotonin receptors in schizophrenia comes from the studies of Yorkston and coworkers [224] showing a beneficial effect of high doses of propranolol in some schizophrenics. Although propranolol is a ~- adrenergic receptor blocking agent in the periphery, in the central nervous system it seems to have anti-serotoninergic rather than anti-adrenergic activity, and it blocks the syndrome of hyperactivity induced in rats by the administration of tryptophan and a monoamine oxidase inhibitor [225].

Despite this evidence that there are abnormalities of serotoninergic activity in schizophrenia, it is likely that the effects are secondary to a dopaminergic system which is superior in the organisational hierarchy of the central nervous system to the serotoninergic systems involved. A variety of drugs that affect the firing of serotoninergic neurons in the raphe nucleus do so by means of a catecholaminergic mechanism, and have no effect when they are 144 D.A. Bender applied directly to the raphe by iontophoresis [226, 227]. Indeed, some of the effects are inhibited by picrotoxin, which suggests that GABA mediation may also be important. Bird and coworkers [228] have suggested that there is a functionally superior GABA system that may provide the clue to the basic biochemical lesion of schizophrenia. Chapter 3

The Oxidative Pathway of Tryptophan Metabolism

The oxidative pathway is quantitatively the most important pathway of tryptophan metabolism; under normal conditions as much as 99% of the daily intake of tryptophan is metabolised in this way.

As was noted in Chapter I, the pathway can lead to either the total oxidation of tryptophan to acetyl CoA, and hence to carbon dioxide and water, or to the formation of nicotinamide nucleotides, the de novo synthesis of NAD and NADP. It is not clear how important this synthesis of nicotinamide nucleotides from tryptophan is, compared with the use of preformed nicotinamide and nicotinic acid from the diet. However, there are a number of conditions in which there is a disturbance of tryptophan metabolism which leads to the development of the niacin deficiency disease pellagra. This suggests that under normal conditions synthesis from tryptophan may be an important source of nicotinamide nucleotides.

Tryptophan Oxygenase

The first enzyme of the oxidative pathway is tryptophan oxygenase, sometimes also called tryptophan pyrrolase, and at one time known as tryptophan peroxidase. It is the rate-limiting step of tryptophan oxidation, and its activity controls the overall rate of tryptophan metabolism in the body. As might be expected for a rate-limiting enzyme, tryptophan oxygenase has a relatively short half-life, of the order of 2 hours, so that its activity can be controlled with considerable precision by control of the rate of synthesis and degradation of the enzyme.

At least four distinct mechanisms are involved in the regulation of tryptophan oxygenase activity in the liver: induction by hormones; stabilisation by tryptophan against degradation; inhibition by nicotinamide nucleotides and other intermediates of the oxidative pathway; activation of the apo- enzyme by its haem cofactor.

Induction by hormones a) Glucocorticoids

The induction of tryptophan oxygenase by glucocorticoids was first reported by Mehler and coworkers [229], and was subsequently shown to be true induction, in that there is an increase in the rate of synthesis of the enzyme protein [230] and of messenger-RNA for the enzyme [231]. This latter paper is of interest because it was apparently the first report of translation of mammalian mesenger-RNA in a heterologous system in vitro. Schutz and coworkers [231] showed that the product of their in vitro protein

145 146 D. A Bender synthesis was immunologicaLly the same as the 'normal' enzyme isolated from liver, and they demonstrated an increase in the content of messenger-RNA in the liver after administration of glucocorticoids. Nakamura and coworkers [2323 have also demonstrated induction of the enzyme by glucocorticoids in isolated rat hepatocytes in culture. The physiological importance of this induction of tryptophan oxygenase by glucocorticoids was discussed above in Chapter 1. b) Glucagon

Nakamura and coworkers [232] also studied the effects of glucagon on tryptophan oxygenase in isolated rat hepatocytes. Alone, this hormone has no effect, but together with tryptophan, which stabilises the enzyme against degradation, as discussed below, the increase in oxygenase activity is as great as is observed when glucocorticoids and tryptophan are added to the cell culture. In their system, glucagon could be replaced with the long- acting dibutyryl derivative of cyclic AMP. They also noted that the effect of tryptophan, glucocorticoids and glucagon on the activity of tryptophan oxygenase are all additive, suggesting that they act by different mechanisms. They showed that while cycloheximide would inhibit the effects of both glucagon and glucocorticoids, actinomycin prevented induction only by glucocorticoids. This suggests that while glucocorticoids act at the level of transcription of DNA to yield messenger-RNA (a conclusion that is also obvious from the earlier studies of Schutz and coworkers [2313), glucagon acts at the level of translation, increasing the rate of protein synthesis from existing messenger-RNA, but not affecting the transcription of DNA. c) Oestrogens

A number of observations of the effects of oestrogens used as oral contraceptive agents and menopausal hormone replacement therapy have suggested that they may act like glucocorticoids to induce synthesis of tryptophan oxygenase. Braidman and Rose [233] showed that the administration of oestradiol benzoate to intact female rats led to an increase in the activity of tryptophan oxygenase, tyrosine aminotransferase and alanine aminotransferase in the liver. After adrenalectomy, this response was diminished, suggesting that at least a part of the action of oestrogens was indirect. Further

Figure 8

The oxidative metabolism of tryptophan

(a) Tryptophan oxygenase (L-tryptophan: oxygen oxidoreductase (decyclising) EC 1.13.11.11)

(b) Kynurenine formamidase (arylformylamine amidohydrolase, EC 3.5.1.9)

(d) Kynureninase (L-kynurenine hydrolase, EC 3.7.1.3)

(e) 3-Hydroxyanthranilic acid oxidase (3-hydroxyanthranilate: oxygen 3,4-oxidoreductase (decyclising) EC 1.13.11.6) Biochemistry of Tryptophan 147

COOH CH2- CH -NH 2 tryptophan o2.1 HN O COOH ~ [~ - CH2oCH-NH2 NH -CHO H 2Oh-~ formylkynurenine 7M HCOOH'f ~ O COOH ~ C-CH2-CH-NHii ! 2

NH 2 kynurenine NADPHc~ 02

N ADPt ~ O COOH ~:" C H2-" CH-NH 2

H 2 3-hydr oxykynurenine OH

H20

COOH CH3"CH -NH 2 COOH

~ NH2 3-hydroxyanthranilic acid OH

NAD~ ~ COOH (see Fig. iO)

CO 2 ~ O=CH ~"NH2 acroleyl aminofumarate (see Fig. Ii) COOH Fig. 8 (see facing page) 148 D.A. Bender

evidence that this may be an indirect effect came from the observation that although ovariectomy or the administration of testosterone led to a lower response to oestrogen administration in female rats, and male rats generally have a lower activity of tryptophan oxygenase than do females, the administration of oestrogens to males also results in the induction of tryptophan oxygenase. Although this may reflect the presence of oestrogen receptors in the male liver, Braidman and Rose [233] suggested that it was more likely that the effect was due to displacement of corticosteroids from their plasma binding proteins by oestrogens, and hence enhanced uptake of the glucocorticoids into the liver.

Patnaik and Sarangai [234] also demonstrated a fall in tryptophan oxygenase activity in liver from female rats following ovariectomy, and interpreted their results as indicating that oestrogens have at least a permissive effect in maintaining the normal level of tryptophan oxygenase in the female rat. They demonstrated induction of tryptophan oxygenase by oestradiol in mature female rats; this effect was less marked in immature or older females, and they suggested that this reflected induction of the enzyme by oestrogens, mediated by liver oestrogen receptors, which are present in greatest amount at sexual maturity. We (D.A. Bender, A.E. Laing and J.A. Vale, unpublished observations) have been unable to repeat these findings. Our results show no induction of tryptophan oxygenase after the acute or chronic administration of oestradiol, ethinyl oestradiol or oestrone sulphate to intact or ovariectomised rats, and indeed there appears to be a slight reduction in activity over a few hours after the acute administration of the oestrogens. Furthermore, administration of oestradiol together with a loading dose of tryptophan leads to a reduction in the urinary excretion of tryptophan metabolites that might be expected to be increased following induction of tryptophan oxygenase. The difference may be due to the method used to measure the activity of tryptophan oxygenase; our method involves the use of an internal standard to correct for inhibition of kynureninase by oestrogen metabolites [289, 319], while that used by Paitnaik and Sarangai [234] does not. This means that in their system inhibition of kynureninase would appear to be increased activity of tryptophan oxygenase.

Kanke and coworkers [235] reported that the administration of oral contraceptives containing both oestrogens and progestogens to rats that had been rendered anaemic by feeding an iron-deficient diet led, over a period of 4 weeks, to an increase in the activity of tryptophan oxygenase in the liver. Although this might be interpreted as evidence of the induction of tryptophan oxygenase by oestrogens, it might also reflect the induction of haem synthesis by the hormones [236] and hence increased activation of the apoenzyme by its cofactor, as discussed below. Leonard and Hamburger [237] showed that repeated administration of oestradiol benzoate over 15 days led to an increase in the liver content of holo-tryptophan oxygenase, with no effect on the total amount of enzyme (apo-enzyme + holo-enzyme) present. Again this would appear to reflect the greater stability of the holoenzyme against proteolysis, and the effects of hormones on either the synthesis of haem or the activation of the apo-enzyme by haem.

Green and coworkers [238] suggested that oestrogens have no effect on the activity of tryptophan oxygenase. They showed that the plasma concentration of tryptophan after the administration of a dose of 50 mg/kg body weight to women was the same whether they were taking oestrogenic oral contraceptives or not. Although they reported an increase in the plasma concen- trationof kynurenine under these conditions, this was associated with reduced urinary excretion of kynurenine, and increased excretion of hydroxykynurenine and xanthurenic acid, which would be compatible with inhibition Biochemistry of Tryptophan 149 of kynureninase by oestrogens or their metabolites, but not with induction of tryptophan oxygenase. Similarly, studies with rats treated with oestrone sulphate showed that there was an increase in the concentration of tryptophan in plasma in response to the administration of the oestrogens - again this would not be compatible with induction of tryptophan oxygenase by the hormone. There was no detectable effect of oestrone sulphate administraton on the activity of tryptophan oxygenase in this study [239]. Bender and coworkers (unpublished studies) have observed a similar increase in the concentration of tryptophan in plasma in women receiving either piperazine oestrone sulphate or ethinyl oestradiol as menopausal hormone replacement therapy.

Stabilisation and activation by tryptophan

Greengard and Feigelson [240] demonstrated the apparent induction of tryptophan oxygenase by its substrate. The effect was a peculiar one, in that it was additive with induction by glucocorticoids, and in some of the early papers on the subject it is referred to as 'super-induction' of the enzyme. Schimk~ and coworkers [230, 241] showed that the effect of tryptophan on the oxygenase was not induction of new protein synthesis, but stabilisation of the enzyme against proteolysis. This means that as long as synthesis of the enzyme protein occurs at the normal rate, there will be an increase in the liver content of enzyme protein in response to trytophan administration as a result of the increased half-life of the stabilised enzyme. If the enzyme were induced by a glucocorticoid in addition to the administration of tryptophan this would result in a greater rate of enzyme protein synthesis, which, coupled with the reduced catabolism as a result of the effect of tryptophan, would mean a considerable increase in the liver content of the enzyme.

Schimke and coworkers [241] also showed that tryptophan oxygenase in liver slices in vitro was stabilised against degradation by the addition of either its substrate, tryptophan, or s-methyl tryptophan, which is not a substrate. There was a good agreement between the ability of various tryptophan analogues to stabilise the enzyme in this way and the activity of these analogues as 'inducers' of tryptophan oxygenase in vivo. They suggested that there are two tryptophan binding sites on the enzyme; one is catalytic and specific for L-tryptophan, while the other has a broader specificity. It is binding at this second site which stabilises the enzyme against catabolism.

Sourkes [32] showed that the administration of s-methyl trytophan to intact or adrenalectomised rats led to an increase in the rate of tryptophan catabolism that persisted for several days after a single dose. This resulted in an increase in the urinary excretion of xanthurenic and kynurenic acids, and a decrease in the plasma concentration of tryptophan, to about 30% of the pre-treatment level. This led to a reduction in the concentration of tryptophan in the liver and brain, and a reduction in the concentration of 5-hydroxyindoles in the central nervous system. He also noted an increase in the overall rate of amino acid catabolism, urea synthesis and gluconeogenesis, suggesting that the increase in tryptophan oxidation caused by the administration of s-methyl tryptophan can have the same effects on general amino acid metabolism as does the administration or secretion of glucocorticoids, discussed in Chapter 1. s-Methyl tryptophan also enhances the activity of tryptophan oxygenase when the concentration of tryptophan is sub-optimal. Under normal conditions the graph of substrate concentration versus rate of tryptophan oxidation is sigmoid, suggesting an allosteric effect. However, in the presence of ~- 150 D.A. Bender methyl tryptophan the graph is hyperbolic, suggesting that the enzyme is also activated by those analogues that bind at the 'stabilisation' site [242].

Product inhibition

Wagner [243] showed that a number of intermediates of the tryptophan oxidative pathway would inhibit tryptophan oxygenase. From the data shown in Table 2 it seems probable that inhibition by kynurenine, hydroxykynurenine and hydroxyanthranilic acid might be significant under physiological conditions, but it is unlikely that the concentration of nicotinic acid mononucleotide, nicotinamide mononucleotide or desamido-NAD would rise to a high enough level to cause any significant inhibition.

TABLE 2

Inhibition of tryptophan oxygenase by intermediates of tryptophan oxidation

inhibitor concentration (mol/1)

Inhibitor 5x10 -3 5x10 -4 5x10 -5 5x10-6

% inhibition of tryptophan oxygenase

Kynurenine 58 47 36 - Hydroxykynurenine 100 100 95 67 Hydroxyanthranilic acid 100 100 85 32 Quinolinic acid 12 10 - - Nicotinic acid mononucleotide 83 33 0 - Nicotinamide mononucleotide 50 0 - - Desamido-NAD 64 43 22 - NAD + 42 0 - - NADP + 12 - - -

NADH - 30 - 0

NADPH - 50 - -

(Data from Wagner [243] and Cho-Chung and Pitot [244])

Wagner's data [243] show inhibition of tryptophan oxygenase by NAD + and NADP + only at the relatively high concentrations of 5 x 10 -5 mol/l; the total concentration of nicotinamide nucleotides (NAD +, NADP +, NADH and NADPH) in the liver is of the order of 0.5 - 1.0 x 10 -3 mol/l [245], so it is unlikely that there would be significant inhibition in vivo. However, Cho-Chung and Pitot [244] showed inhibition of the enzyme by the reduced cofactors, and especially by NADPH, at concentrations that are nearer to those likely to be encountered in vivo.

The inhibition of tryptophan oxygenase by reduced nicotinamide nucleotides may be important in vivo. Badawy and Evans [246, 247] showed that following the chronic administration of alcohol to rats for 2 weeks, which leads to an increase in the proportion of liver nucleotides present in the reduced form, there is a fall in the activity of tryptophan oxygenase in the liver. This effect persists for several days after the withdrawal of alcohol, a period during which the redox state of the liver continues to be more than normally reduced. This lower rate of tryptophan oxidation as a result of chronic Biochemistry of Tryptophan 151 alcohol administration is accompanied by an increase in the rate of serotonin synthesis in the central nervous system, which gradually returns to normal at the same rate as the liver nucleotides return to their normal redox state, and the oxidation of tryptophan increases.

The availability of haem

Tryptophan oxygenase is a haem-dependent enzyme, and in human liver, as in a number of species, including the rat, there is a considerable proportion of the enzyme that is present as the apo-enzyme, and is therefore inactive until additional haem is made available, either by the administration of haem or haem precursors in vivo or by the addition of haematin during incubation in vitro.

Badawy and Evans [248] showed that the administration of haem to rats led to a considerable increase in the activity of tryptophan oxygenase. The administration of 5-aminolaevulinic acid as a haem precursor had a similar effect, although in this case the effect could be abolished by the administration of protein synthesis inhibitors, suggesting that 5-aminolaevulinic acid must induce one or more of the enzymes of haem synthesis. This led them to suggest that the availability of haem may be a major factor in controlling the activity of tryptophan oxygenase. This may be relevant to the psychiatric problems of porphyria; an increase in the activity of tryptophan oxygenase as a result of increased availability of haem might result in depletion of the body pools of tryptophan, and hence result in reduced synthesis of serotonin in the central nervous system, as discussed in Chapter 2. Goldberg and coworkers [249] have shown that a number of 17-oxo-steroids which are excreted in abnormally large amounts in intermittent acute porphyria induce the synthesis of 5-aminolaevulinate synthetase, the first and rate limiting enzyme of haem synthesis. Several progesterone analogues have a similar effect, and the administration of these hormones is associated with increased activity of tryptophan oxygenase [236].

The activation of apo-tryptophan oxygenase by haem is not a simple process; it seems to require the action of xanthine oxidase. Removal of xanthine oxidase from in vitro preparations of liver by immuno-precipitation prevents the activation of the apo-enzyme by added haem [250]. This may explain the action of allopurinol, which is an inhibitor of apo-tryptophan oxygenase, but not of the holo-enzyme, and seems to prevent the binding of haem to the apo-enzyme [251]. The administration of allopurinol to rats prevents the increase in tryptophan oxygenase activity that is normally seen in response to corticosteroid administration, because the drug prevents the activation of newly synthesized apo-enzyme by conjugation with haem. However, in perfusion studies this effect of allopurinol is overcome by high concentrations of tryptophan in the perfusion medium [252] so that it is unlikely that the use of allopurinol together with tryptophan would potentiate the central nervous sytem effects of tryptophan, as has been suggested [188].

The administration of cortisol to adrenalectomised rats permits considerably greater induction of 5-aminolaevulinate synthetase by the porphyrogenic drug allyl isopropacetamide than is normally the case. Morgan and Badawy [253] suggested that this was because of the induction of apo-tryptophan oxygenase by the corticosteroid. The apo-enzyme provides a 'sink' for the newly synthesized haem, and thus reduces the feed-back inhibition of 5-aminolaevulinate synthetase by haem. Additional evidence that the availability of apo-tryptophan oxygenase may be an important factor in the control of haem 152 D.A. Bender

metabolism is provided by the observation that the adminstration of haematin leads to both a decrease in the activity of 5-aminolaevulinate synthetase and an increase in the degree of saturation of tryptophan oxygenase by its cofactor; if allopurinol is also given to the animals, to inhibit the activation of apo-tryptophan oxygenase, and so increase the pool of free haem in the liver, there is a more marked inhibition of 5-aminolaevulinate synthetase [254].

Indoleamine Dioxygenase (D-Tryptophan Oxygenase)

Hayaishi and coworkers [255] have described an enzyme that appears, superfi- cially, to catalyse the same reaction as tryptophan oxygenase, but with a broader specificity. Tryptophan oxygenase is specific for L-tryptophan, and is found only in the liver in mammals; indoleamine dioxygenase catalyses the same cleavage of the pyrrole ring of the indole nucleus, but will act on D- or L-tryptophan, D- or L-5-hydroxytryptophan, tryptamine, serotonin and melatonin. This enzyme is widely distributed in a number of tissues, including brain, lung, gastric and intestinal mucosa, kidney, heart and adrenal gland. Although, like tryptophan oxygenase, it is a haem enzyme, indoleamine dioxygenase uses the superoxide anion (02 ) as the source of oxygen, rather than molecular oxygen. The activity of the enzyme from rabbit intestinal mucosa is of the same order of magnitude as that of tryptophan oxygenase in the liver.

The physiological significance of indoleamine dioxygenase is not clear, although it obviously provides a pathway for the conversion of any ingested D-tryptophan to D-kynurenine. One product, 5-hydroxykynuramine, formed by its action on serotonin, may be physiologically important; it is a potent inhibitor of the action of serotonin in promoting the aggregation of blood platelets. Hayaishi [255] has suggested that this may provide a measure of regulation in cases of over-synthesis of serotonin; not only an alternative pathway for the catabolism of the amine, but also the production of an inhibitor of one of its biological actions.

Indoleamine dioxygenase is distributed in a variety of tissues, unlike tryptophan oxygenase, which is found only in the liver and therefore it may provide a pathway for the metabolism of tryptophan in extra-hepatic tissues. Since tissues such as kidney have the other enzymes of the tryptophan oxidative pathway, this suggests that indoleamine dioxygenase may be important in the overall metabolism of tryptophan in the body.

The metabolism of D-tryptophan

Indoleamine dioxygenase provides a pathway for the oxidative catabolism of any ingested D-tryptophan (either administered as a drug when the racemic mixture is used or arising from intestinal bacterial metabolism). There is also evidence that D-tryptophan is usable by mammals as a source of tryptophan for the synthesis of proteins and serotonin. Yuwiler and coworkers [256] showed that the administration of D-tryptophan to rats led to an increase in the brain content of serotonin. Since the enzymes involved in serotonin synthesis are known to be specific for L-tryptophan and L-5-hydroxytryptophan, and there is no evidence of the presence of tryptophan (or any other amino acid) racemase in mammalian tissues, they suggested that there must be racemisation by an indirect mechanism - oxidative deamination to indole- Biochemistry of Tryptophan 153 pyruvate, catalysed by D-amino acid oxidase, and reamination of the symmetrical oxo-acid, catalysed by tryptophan aminotransferase. Two further explanations of Yuwiler's results are possible:

(i) Racemisation of D-tryptophan by intestinal bacterial action; there are known to be active amino acid racemases in a number of these organisms. However, the administration of D-tryptophan to germ-free rats leads to considerable excretion of indole carboxaldehyde in the urine, evidence that there is deamination of D-tryptophan in mammalian tissues [257].

(ii) The increase in brain serotonin may be due to an increase in the uptake of tryptophan into the brain as the result of displacement of L-tryptophan from albumin binding by the D-enantiomer. However, there is no evidence that D-tryptophan will bind to albumin, or otherwise displace L-tryptophan from binding [115, 116].

Further evidence that D-tryptophan can be used by mammals comes from the studies of McEwan and Carpenter [258] who showed that as long as at least half of the tryptophan requirement of weanling mice is met by the provision of the L-enantiomer, the addition of D-tryptophan to the diet will permit normal growth. There is an increase in the urinary excretion of N1-methyl nicotinamide and methyl pyridone carboxamide under these conditions, which is evidence that D-tryptophan can enter the main pathway of oxidative metabolism, as well as being a substrate for protein synthesis. Although D-tryptophan could enter the oxidative pathway by way of indoleamine dioxygenase, with racemisation at the level of D-kynurenine, the ability of the mouse to use D-tryptophan for protein synthesis means that there must be racemisation to L-tryptophan, either directly or indirectly by way of indole- pyruvate.

Formylkynurenine Formamidase

The immediate product of the action of tryptophan oxygenase is formylkynurenine. However, the activity of the hydrolase which removes the formyl group, formyl kynurenine formamidase, is extremely high. This means that under all conditions except when highly purified enzyme preparations are used, the apparent product of tryptophan oxygenase (or of indoleamine dioxygenase) is kynurenine rather than formylkynurenine. Although the activity of this enzyme is not normally a limiting factor, Seifert and Cassida [259] have shown that a number of pesticides that have a teratogenic effect in chick embryos also lead to reduced tissue concentrations of nicotinamide nucleotides, apparently by inhibiting formylkynurenine formamidase.

A means of measuring the activity of this enzyme has been developed as the result of an interesting accidental observation. Jacobson [260] reported that the commercially available formylkynurenine was not as described, but in• fact was N 1 ,N~ -dlformylkynurenlne;• the product of formamidase action on this compound is N ~-formylkynurenine, which is not a substrate for kynureninase or kynurenine hydroxylase activity, but accumulates in the incubation medium, and can readily be measured. When Nl-formylkynurenine is the substrate, as is the case in vivo, the product is kynurenine, which is a substrate for onward metabolism, so that measurement of formamidase activity is difficult. It was noted above that when tryptophan oxygenase activity is measured by the rate of formation of kynurenine, misleading 154 D.A. Bender results may be obtained unless there is correction for this onward metabolism of kynurenine by the use of an internal standard [289, 319].

The Metabolism of Kynurenine

As shown in Figure 9, there are three possible metabolic fates for kynurenine: cleavage to anthranilic acid and pyruvate; hydroxylation to 3-hydroxykynurenine followed by cleavage to 3-hydroxyanthranilic acid and pyruvate; transamination and ring closure to xanthurenic acid. A further possible route of metabolism, not shown in the Figure, is N-acetylation to yield N-acetylkynurenine, which is excreted in the urine in small amounts.

Under normal conditions the major pathway of kynurenine metabolism is by way of hydroxylation, followed by cleavage catalysed by kynureninase. There is normally little transamination, except after the administration of a loading dose of tryptophan. Ueda and coworkers [261] suggested, on the basis of studies with 14C-labelled substrates, that an alternative to the accepted pathway of kynurenine metabolism would be cleavage to anthranilic acid, catalysed by kynureninase, followed by hydroxylation to 3-hydoxyanthranilic acid, catalysed by the nonspecific hydroxylases of the liver microsomes. Bender and McCreanor [262] have shown that the more usually accepted pathway, of mitochondrial hydroxylation followed by kynureninase action, is more likely to predominate in the rat. The kinetics of the enzymes involved, relative to the steady state concentration of kynurenine in the liver, are such that the initial rate of kynurenine hydroxylation will be at least three times as great as the initial rate of cleavage. Furthermore, although the values of Km of kynureninase for its two substrates, kynurenine and 3-hydroxykynurenine, are approximately the same, the Vma x towards hydroxykynurenine is very much greater. Tanizawa and Soda [263] reported a lower K m of the enzyme for hydroxykynurenine than for kynurenine, and also concluded that the preferred route of kynurenine metabolism would be by way of hydroxylation rather than direct cleavage.

Bender and McCreanor [262] also showed that anthranilic acid is not a precursor of nicotinamide nucleotides in the rat, even in animals in which maximum activity of microsomal enzymes has been induced by the prior administration of hexobarbitone. However, there is some formation of anthranilic acid from kynurenine in vivo; measurable amounts of anthranilic acid, its glucuronide and the glycine conjugate o-aminohippuric acid are excreted in the urine after the administration of loading doses of tryptophan [267, 270].

Under normal conditions, little kynurenine or hydroxykynurenine undergoes transamination. The Km of the aminotransferase is considerably higher than those of either kynureninase or kynurenine hydroxylase [264], so until these two enzymes are saturated with their substrates, transamination will be a relatively unfavoured reaction. In response to an increased flux of metabolites through the pathway, as the result of induction of tryptophan oxygenase or the administration of a tryptophan load, there is an accumulation of kynurenine and hydroxykynurenine to a sufficient extent to permit significant transamination to occur. Inhibition of kynureninase has a similar effect. Kynurenic acid (arising from kynurenine) and xanthurenic acid (arising from hydroxykynurenine) are normally excreted unchanged, although there can be some further metabolism to quinaldic and hydroxyquinaldic acids respectively. Biochemistryof Tryptophan 155

O COOH OH II ~ ,C- CH~"CH'NH 2 COOH kynurenine ~ NH 2 d ~

kynurenic acid coo

cOOH OH II ,, ~COOH r~COOH ~C'CHE'CH'NH2

~""~.'~ N H 2 -,,~F/.N H2 d anthranilic acid OH hydr oxykynurenine OH ~ ~ coo. xanthurenic acid b ICH3-CH'NH2 ~ .COOH ~ N H 2 OH hydroxyanthranilic acid

(a) Kynurenine hydroxylase (L-kynurenine, NADPH: oxygen oxidoreductase (3-hydroxylating) EC 1.14.13.9)

(b) Kynureninase (L-kynurenine hydrolase, EC 3.7.1.3)

(c) Non-specific microsomal hydroxylation (d) Kynurenine aminotransferase (L-kynurenine: 2-oxo-glutarate aminotransferase (cyclising) EC 2.6.1.7)

Figure 9 Pathways of kynurenine metabolism 156 D.A. Bender

The tryptophan load test for vitamin B 6 nutritional status

Studies in Padua during the 1940s on rats fed high protein diets that were deficient in the B vitamins, led to the discovery of increased excretion of xanthurenic acid in vitamin B 6 deficiency. This was celebrated at the time by a ditty in a student magazine 'Grita festante il Dottore Perini, pisciani verdi il mei ratti albini' (Dr. Perini cried with glee 'my albino rats are pissing green'). Lepkovsky and Nielsen [265] showed that the green colour was due to the presence of xanthurenic acid, which was known to be a metabolite of tryptophan, and its reaction with rust from the grids of the cages to form ferric xanthurenate. Xanthurenic acid itself is yellow, but the colour would not be very noticeable against the normal yellow colour of rat urine; the ferric salt has a distinctive colour, and until the development of fluorimetric methods of analysis the reaction with ferric chloride was the usual method of measuring xanthurenic acid. They showed that in vitamin B 6 deficiency there was a considerably greater increase in the urinary excretion of xanthurenic acid after the administration of a loading dose of tryptophan than occurred in animals receiving an adequate intake of the vitamin.

This excretion of xanthurenic acid (and also of kynurenic acid) in vitamin B 6 deficiency suggests that the activity of kynureninase is reduced. Kynure- ninase is a pyridoxal phosphate (vitamin B 6) dependent enzyme, and in deficiency the activity of the holo-enzyme in rat liver falls to about half the normal level [266]. This suggests that the measurement of these two metabolites of tryptophan, and possibly also of kynurenine and hydroxykynurenine, may be a useful means of assessing vitamin B 6 nutritional state; measurement is relatively easy [267, 268, 269]. However, the excretion of these compounds varies considerably from day to day and from person to person, even when the intake of tryptophan is controlled relatively well [269]. Therefore it has become accepted that a more suitable test of vitamin B 6 nutritional status is the ability to metabolise a relatively large test dose of tryptophan - the tryptophan load test. The excretion of xanthurenic and kynurenic acids, and possibly also of kynurenine and hydroxy kynurenine, is measured before and after the administration of the dose of tryptophan, preferably using 24 hour collections of urine. The mean observed excretion of some tryptophan metabolites and the fractional increases after a tryptophan load, are shown in Table 3.

Early investigations of vitamin B 6 nutritional status used relatively large amounts of tryptophan as the loading dose, of the order of 5g of L-tryptophan and up to 10g of DL-tryptophan. This is several times greater than the normal daily intake of tryptophan, and would certainly provide a considerable metabolic stress. Coursin [271] noted that the use of different doses of tryptophan made comparison between different workers difficult, and also that the use of DL-tryptophan further confused the interpretation of results, because of the formation of D-kynurenine, which was not generally metabolised further (presumably this is the result of indoleamine dioxygenase activity [255]). There was also a considerable variation in the time over which urine was collected, varying between 4-24 hours. He therefore recommended standardisation of the method: 2g of L-tryptophan should be given as a single dose before breakfast in the morning, suspended in milk or orange juice to mask the flavour. He noted that although most of the tryptophan metabolites would be recovered within 8 hours, it would be preferable to make 24 hour collections of urine when this was possible. Allegri and coworkers [272] suggested that rather than a flat dose of 2g of tryptophan Biochemistry of Tryptophan 157

TABLE 3

The excretion of tryptophan metabolites by normal subjects and the increase after oral administration of 2g of L-tryptophan.

(Data from Price et al. [267] and Brown et al. [270]) metabolite mean excretion % increase after 2g ( ~moI/24 h) tryptophan ( ± I sd range) kynurenic acid 11.4 - 15 329 xanthurenic acid 7.7 - 12.5 191 kynurenine 9.3 - 17.7 163 hydroxykynurenine 10.2 - 28.4 95 acetylkynurenine 9.5 - 15.9 35 o-aminohippuric aciid 21 .I - 31.9 87 anthranilic acid glucuronide 3.8 - 7.0 69 quinolinic acid 24.0 - 68.4 68 Nl-methyl nicotinamide 21.5 - 37.1, 98 methyl pyridone carboxamide 50.7 - 117.9, 61

(* the excretion of these two metabolites of nicotinamide will depend on the dietary intake of nicotinamide and nicotinic acid, as well as the synthesis of nicotinamide nucleotides from tryptophan). to all subjects, it would be preferable to match the tryptophan load to body weight, and suggested a dose of 50 mg/kg body weight.

Wolf and coworkers [273] have suggested that a kynurenine load test might be a useful adjunct to, or replacement for, the tryptophan load test. They showed that following the administration of 150 mg of kynurenine the excretion of xanthurenic and kynurenic acids was the same as after 2g of tryptophan. However, there was no increase in the excretion of the metabolites of the nicotinamide nucleotides (Nl-methyl nicotinamide and methyl pyridone carboxamide), as there is after a tryptophan load; they suggested that this is because kynurenine is metabolised in a variety of tissues, whereas tryptophan is only oxidised in the liver, the only tissue to contain tryptophan oxygenase. However, this ignores the possible contribution of indoleamine dioxygenase to tryptophan metabolism in extra-hepatic tissues [255]. Bender and McCreanor [262] showed that the administration of a relatively large amount of kynurenine to rats did lead to an increase in the liver content of nicotinamide nucleotides, although not in the urinary excretion of Nl-methyl nicotinamide over the first 24 hours.

Knox [266] showed that the average rate of kynureninase activity in the rat was 11.1 pmol of substrate metabolised/h/g dry weight of liver, which was very close to the 'normal' rate of tryptophan oxygenase activity, 9.3 pmol/h/g. In response to a tryptophan load there was up to a 10-fold increase in tryptophan oxygenase activity. This would explain the accumulation of kynurenine and hydroxykynurenine after a tryptophan load, and hence the formation of increased amounts of various kynurenine metabolites even in people and animals receiving adequate amounts of vitamin B 6. This close matching of the basal rates of tryptophan oxygenase and kynureninase activities may also explain the considerable day to day variation in excretion of tryptophan and kynurenine metabolites in normal subjects under basal conditions, referred 158 D.A. Bender

to above [269]. Minor stresses that alter the secretion of cortisol will affect the activity of tryptophan oxygenase and a slight increase in this activity may well lead to a considerable accumulation of kynurenine. Wolf and coworkers [273] noted that one of the major advantages of their kynurenine load test over the tryptophan load test, apart from the very much smaller amount of test material that must be swallowed by the subject, was that there would be no increase in the activity of tryptophan oxygenase, and therefore the results would be easier to interpret in terms of vitamin B 6 nutritional status and kynureninase activity. However, a possible complication may arise from the inhibition of tryptophan oxygenase by kynurenine [243].

The tryptophan load test depends on the reduction in kynureninase activity in response to a deficiency of its cofactor, pyridoxal phosphate. However, the enzyme that is responsible for the formation of kynurenic acid from kynurenine and xanthurenic acid from hydroxykynurenine, kynurenine aminotransferase, is also pyridoxal phosphate dependent. The aminotransferase is found both in the cytosol and in the mitochondria; in vitamin B 6 deficiency the activity of the cytosol aminotransferase is markedly reduced, as is that of kynureninase which is found only in the cytosol, whereas the activity of the aminotransferase in the mitochondria is not affected. Physical compartmentation by the mitochondrial membrane seems to protect the aminotransferase from cofactor depletion [274].

The tryptophan load test in women receivin 9 oestrogens

Rose and coworkers [275] showed that 26 out of 31 women who had been receiving high oestrogen oral contraceptives for between 6-36 months showed an abnormal response to a tryptophan load test, indicative of vitamin B 6 deficiency. They also showed that this abnormal tryptophan metabolism could be corrected by the administration of relatively large amounts of vitamin B6, of the order of 20mg/day compared with the average recommended daily intake of 2mg [276]. Luhby and coworkers [277] also showed that supplements of vitamin B 6 of the same order of magnitude as the normal recommended intake were effective in correcting tryptophan metabolism in only a few women receiving oestrogenic contraceptives, and up to 25mg was required daily for complete normalisation of tryptophan metabolism in all of their subjects. Ethinyl oestradiol used as menopausal hormone replacement therapy has a similar effect [278].

There have been many reports of abnormalities of tryptophan metabolism in women receiving oestrogens as contraceptives or as menopausal hormone replacement therapy, and these have generally been interpreted as indicating oestrogen- induced vitamin B 6 deficiency or depletion [184, 279, 280]. As noted above, the administration of relatively large amounts of vitamin B 6 will restore tryptophan metabolism to normal in these women, and will also overcome some of the other metabolic and psychiatric side effects of oestrogen administration [12]. However, other indices of vitamin B 6 nutritional status are unaffected by oestrogen administration: Rose and coworkers [275] were able to demonstrate reduced excretion of 4-pyridoxic acid (the principal metabolite of vitamin B 6) in only 7 of the 26 subjects who showed abnormal tryptophan metabolism; the plasma concentration of pyridoxal phosphate and the ability to metabolise a test dose of methionine are unaffected [281, 282]; in most studies the activation of erythrocyte aspartate and alanine aminotransferases by pyridoxal phosphate added to the incubation in vitro has also been unaffected by oestrogen administration [281, 282, 283], although there have been results indicative of some degree of depletion of vitamin B 6 by this criterion in some studies [284, 285]. Biochemistry of Tryptophan 159

Wolf and coworkers [286] showed an increase in the urinary excretion of kynurenine metabolites after a test dose of tryptophan or kynurenine in women receiving oestradiol or oestrone as menopausal hormone replacement therapy, and stated that although their results would be compatible with oestrogen- induced vitamin B 6 deficiency, it was not possible from their study to exclude a direct effect of oestrogens on kynureninase. Such a mechanism was proposed by Mason and Gullekson [287], who showed that relatively high concentrations of synthetic oestrogen disulphates (which do not occur in vivo) compete with pyridoxal phosphate for the cofactor binding site of apo- kynurenine aminotransferase. They therefore suggested that the mechanism of oestrogen-induced vitamin B 6 depletion was competition for cofactor binding, which resulted in displacement of pyridoxal phosphate from protective protein binding, and hence its increased catabolism and excretion. However, it was noted above that oestrogen administration has no effect on the urinary excretion of the principal metabolite of vitamin B6, 4-pyridoxic acid [281, 282, 288]. Bender and Wynick [189] investigated this proposed mechanism using a partially purified preparation of kynureninase from rat liver. They showed that there was uncompetitive inhibition of cofactor binding by oestrone sulphate (which is widely used clinically, and is one of the human oestrogen metabolites). More importantly, their studies showed that there was competitive inhibition of kynureninase with respect to its substrate by oestrone sulphate and glucuronide. Since the inhibition is competitive, the effect of adminstration of oestrogens would be an increase in the liver content of kynurenine, and hence an increase in the formation and excretion of such kynurenine metabolites as xanthurenic and kynurenic acids. However, when the concentration of kynurenine rises sufficiently, there will be competition with the oestrogen metabolite, and hence restoration of a more or less normal rate of kynureninase activity, albeit with an increased liver pool of kynurenine and hydroxykynurenine. This means that the amounts of metabolites distal to kynureninase will be little affected by oestrogen administration; this explains the observation of earlier workers [279] that despite the apparently reduced activity of kynureninase, and increased excretion of kynurenine metabolites in women receiving oestrogens, there was no decrease, and sometimes even an increase, in the excretion of Nl-methyl nicotinamide, an observation that would be difficult to account for in terms of inhibition of kynureninase by any mechanism other than competition with the substrate.

Bender and Wynick [289] also showed that in rat liver there is a considerable amount of apo-kynureninase, so that on addition of pyridoxal phosphate to incubations in vitro there is a 4 - 5 fold increase in the activity of the enzyme. Knox [266] also showed the presence of a relatively large amount of the apo-enzyme in rat liver, and Coon and Nagler [290] have provided circumstantial evidence that there is a similar amount of apo-kynureninase which is normally inactive in human liver. They noted that in nutritionally normal subjects in whom tryptophan oxygenase activity is increased as a result of increased secretion of corticosteroids, or following the administration of corticosteroids, there is a considerable excretion of kynurenine and kynurenine metabolites, which is indicative of vitamin B 6 deficiency, and is correctable by the administration of supplements of the vitamin.

These findings suggest that it is possible to explain the effects of oestrogen administration in terms of a direct action of oestrogen metabolites on kynureninase. The correction of tryptophan metabolism by the administration of relatively large supplements of vitamin B 6 [276, 277] is the result of activation of the normally inactive pool of apo-kynureninase in the liver. 160 D.A. Bender

It therefore seems likely that the tryptophan load test is not a reliable indicator of vitamin B 6 nutritional status in subjects receiving oestrogens. Coon and Nagler [290] have suggested that it may similarly be unreliable in a variety of patients with serious disease, as a result of increased tryptophan oxygenase activity resulting from increased secretion of glucocorticoids (either a direct effect of the disease itself, or as a general response to the stress of serious illness) or changes in nitrogen balance and hence an additional load of tryptophan for catabolism, as occurs in wasting diseases and advanced cancer. A further complication is the possibility of direct effects of the drugs used in therapy on the activity of kynureninase or another of the enzymes of tryptophan metabolism, as is the case with oestrogens and their metabolites [289], and perhaps direct interference in analysis by drugs and their metabolites. They therefore suggest that abnormal tryptophan metabolism should not be interpreted as indicating vitamin B 6 deficiency without supporting evidence from other indicators of vitamin B 6 nutritional status.

The Synthesis of Nicotinamide Nucleotides

The immediate product of kynureninase action on hydroxykynurenine is hydroxy anthranilic acid; this is oxidised rapidly to acroleyl aminofumarate, which is at a metabolic branch point. Acroleyl aminofumarate can either undergo enzymic decarboxylation to aminomuconic semialdehyde, and thus enter the pathway that leads to complete oxidation to carbon dioxide and water, or it may undergo non-enzymic cyclisation to quinolinic acid, which is the precursor of nicotinamide nucleotides.

As shown in Figure 10, there are two possible sources of nicotinamide nucleotides; quinolinic acid, a metabolite of tryptophan, and dietary niacin (nicotinic acid and nicotinamide). This led to considerable confusion in the early studies of the nutritional deficiency disease, pellagra, since either sources of niacin or additional protein (now known to be a source of additional tryptophan) would cure or prevent the disease. Thus, even once the disease had been demonstrated not to be due to either an infection or a toxin in foods, its aetiology was unclear until the work of Goldberger and coworkers during the 1940s [291, 292]. Although we now understand the importance of tryptophan and niacin, and the separate roles that deficiency of either can play in the aetiology of pellagra, we still do not know the relative importance of these two sources of nicotinamide nucleotides in the body. It is commonly assumed that niacin is the 'major' precursor, and that tryptophan can act as an acceptable 'substitute' when the intake of niacin is inadequate, but, as discussed below, in a number of conditions where tryptophan metabolism is disturbed, pellagra results despite an apparently normal intake of niacin.

The nutritional equivalence of tryptophan and niacin is important for the purposes of estimating the effective niacin content of foods, and hence determining the extent to which diets provide, from both tryptophan and niacin, an adequate supply of precursors of nicotinamide nucleotides, and are protective against pellagra. Horwitt and coworkers [293] studied a relatively large number of subjects, over a period of 88 weeks of experimental diets, after an initial 3 month baseline period, and showed that there was a considerable variation between individuals in the equivalence of tryptophan and niacin. They suggested that on the basis of their results it would be reasonable to assume that a dietary intake of 60 mg of tryptophan Biochemistry of Tryptophan 161 was approximately equivalent to an intake of 1 mg of niacin. This was an under-estimate of the observed average conversion ratio, and thus provided a margin of safety for the preparation of tables of food composition and recommended daily dietary allowances. Obviously, this ratio does not reflect the simple chemical stoichiometry of the pathway, but rather the extent to which acroleyl aminofumarate arising from tryptophan is converted enzymically to aminomuconic semialdehyde or is permitted to undergo non-enzymic cyclisation to quinolinic acid, and hence form nicotinamide nucleotides.

To a certain extent, the results of Horwitt's study [293] have been misinter- preted, and it has become common to refer to the 'efficiency' of conversion of tryptophan to nicotinamide nucleotides, and to assume that the ratio is more or less precisely 60:1. A number of studies have shown that as well as the considerable range of individual variation reported by Horwitt and coworkers [293] there are also changes with hormonal status (for example ranging from 7:1 to 30:1 in late pregnancy [294]) and, perhaps more importantly, with the dietary intake of tryptophan. Thus, Nakagawa and coworkers [295] showed that adding gradually increasing amounts of tryptophan to niacin-free diets led to a change in the tryptophan : niacin equivalence from 122:1 when the intake of tryptophan was low Go 75:1 when the intake was higher. Horwitt and coworkers [296] have reviewed a number of such studies, and have suggested that in order to meet the niacin requirements of 97.5% of the population (the normal criterion adopted in defining dietary allowances) it would be most correct to assume an equivalence of tryptophan : niacin of 89:1. They noted that the equivalence changes with the intake of tryptophan, and that when the intake of tryptophan is low, there is relatively little formation of nicotinamide nucleotides, and hence the ratio of tryptophan : niacin is high.

The explanation of this change in the equivalence of tryptophan and niacin with the intake of tryptophan, and also with changes in hormonal status, seems to lie in the non-enzymic cyclisation of acroleyl aminofumarate to quinolinic acid, in competition with the decarboxylation of aminomuconic semialdehyde, catalysed by picolinic carboxylase. Members of the cat family seem to be unique among animal species that have been investigated in that they are unable to synthesize any significant amounts of nicotinamide nucleotides from tryptophan, but are wholly reliant on a source of preformed niacin in the diet. Ikeda and coworkers [297] showed that cats have a 30 - 50-fold greater activity of picolinic carboxylase than do species that are capable of de novo nicotinamide nucleotide synthesis from tryptophan. It thus seems likely that quinolinic acid is only formed to any significant extent when the flux of metabolites through the pathway is so great that picolinic carboxylase is saturated with its substrate, which therefore accumulates, permitting non-enzymic cyclisation to quinolinic acid. Further evidence for this suggestion comes from the observation that the administration of corticosteroids will prevent or reverse the fall in liver nicotinamide nucleotides that normally ocurs when animals are fed on niacin-free diets without the provision of additional tryptophan [298]. In studies with perfused liver, the addition of hydrocortisone to the perfusion medium leads to a several- fold increase in the activity of tryptophan oxygenase, but only a 20% increase in the oxidation of tryptophan to carbon dioxide [299], evidence that under these conditions of increased flux of metabolites through the pathway picolinic carboxylase is saturated, so that no additional tryptophan can be oxidised to acetyl CoA; the increased formation of acroleyl aminofumarate leads to increased formation of quinolinic acid, and hence increased nicotinamide nucleotide synthesis. Similarly, Smith and coworkers [29] have shown that an increase in the concentration of tryptophan in the incubation medium leads to a 5 - 6-fold increase in the activity of tryptophan oxygenase 162 D. A. Bender / Tryptophan / COOH CH~ C02 ~ " I ~L.NH2 acroleyl aminofumarate NH2= \ O=CH _/__. O= COOH a L;UUrl aminomuconic semialdehyde

bl COOH quinolinic acid ~ COOH ~f. phosphoryl ribose PP c k_ -- PPi COOH ~ COOH ~ . ~ nicotinic acid mononucleotide ~N ~ I nicotinic acid ribose-P

/NH3 Pi nicotinic acid adenine d.inucleotid.e

e (NAAD, desamido-NAD)

~ glutamine

glutamate CONH2 g

JI NAD P (nicotinamide adenine dinucleotide) f nicotinamide 1 Ni-methyl nicotinamide

methyl pyridone carboxamide Fig. 10 (see facing page) Biochemistry of Tryptophan 163

Figure 10

The synthesis of nicotinamide nucleotides

(a) Picolinic carboxylase (amino-carboxy-muconic semialdehyde carboxy-lyase, EC 4.1.1.45)

(b) Non-enzymic cyclisation

(c) Quinolinate phosphoribosyltransferase (nicotinate nucleotide: pyrophosphate phosphoribosyltransferase (carboxylating) EC 2.4.2.19)

(d) Nicotinate phosphoribosyltransferase (nicotinate nucleotide: pyrophosphate phosphoribosyltranferase, EC 2.4.2.11)

(e) Nicotinamide deamidase (nicotinamide amido-hydrolase, EC 3.5.1.19)

(f) Nicotinamide phosphoribosyltransferase (nicotinamide nucleotide: pyrophosphate phosphoribosyltransferase, EC 2.4.2.12)

(g) NADase (NAD glycohydrolase, EC 3.2.2.5)

and kynureninase in isolated rat hepatocytes, but only a 3-fold increase in the rate of oxidation of tryptophan to carbon dioxide - evidence that changes in the availability of tryptophan within the range of physiological variation can lead to saturation of picolinic carboxylase.

Shibata and coworkers [300] have reported that in rats fed on synthetic amino acid mixtures as the sole source of nitrogen the activity of picolinic carboxylase falls as the intake of tryptophan falls, so that a greater proportion of the acroleyl aminofumarate that is formed is available for non-enzymic cyclisation to quinolinic acid. They suggested that this was a major factor in the regulation of nicotinamide nucleotide synthesis. Sanada and coworkers [301, 302] showed that in alloxan-diabetic rats the fall in urinary excretion of Nl-methyl nicotinamide was proportional to the increase in activity of picolinic carboxylase. In studies with somato- tropin and predonine acetate [302] they showed that the activity of picolinic carboxylase was proportional to the ratio of tryptophan intake : urinary Nl-methyl nicotinamide, and inversely proportional to the liver content of nicotinamide nucleotides, when animals received no dietary intake of nicotinamide or nicotinic acid.

The control of tissue concentrations of nicotinamide nucleotides

Bender and coworkers [245] have shown that when rats are fed on diets providing relatively large amounts of tryptophan, with no preformed niacin, there is a slightly higher concentration of nicotinamide nucleotides in the liver, and a very much higher excretion of Nl-methyl nicotinamide in the urine than when they receive minimally adequate tryptophan together 164 D.A. Bender with a relatively large amount of nicotinic acid or nicotinamide. They showed a significant correlation between the activities of both nicotinamide deamidase and nicotinamide phosphoribosyl transferase and the liver content of nicotinamide nucleotides, suggesting that the activities of these two enzymes may be limiting in the incorporation of nicotinamide into nucleotides. The kinetics of the enzymes are such that both nicotinamide deamidase and the phosphoribosyltransferase are saturated with substrate at the normal steady state concentration of nicotinamide in the liver. By contrast, the activity of nicotinate phosphoribosyltransferase was not correlated with liver nicotinamide nucleotide content, and kinetic analysis suggested that this enzyme normally functions somewhat below its maximum velocity, so that in response to an increase in the concentration of nicotinic acid there would be a small increase in the rate of incorporation into nucleotides. Quinolinate phosphoribosyltransferase activity was also not correlated with liver nicotinamide nucleotide content; this enzyme appears to operate at a substrate concentration very much lower than its Km, and therefore would be capable of utilising as much substrate as might be available under any physiological conditions. Nevertheless, as shown in Table 3, the administration of a loading dose of tryptophan does lead to some increase in the excretion of quinolinic acid, suggesting that when the flux of metabolites is high, the capacity of the enzyme is overwhelmed, despite its very high K m relative to liver concentrations of quinolinic acid. This may be because the Vma x of the enzyme is relatively low [245].

These results suggest that the utilisation of nicotinamide, and possibly also of nicotinic acid, arising from the diet or resulting from the hydrolysis of tissue nicotinamide nucleotides is strictly limited, while the utilisation of quinolinic acid arising from tryptophan is not limited in the same way. Since the effect on tissue nicotinamide nucleotide concentrations of feeding experimental animals on a relatively high intake of tryptophan is less dramatic than might be expected, and there is a more marked increase in the urinary excretion of N1-methyl nicotinamide, this suggests that a great measure of control over the tissue content of nucleotides is maintained by hydrolysis to nicotinamide. The utilisation of this nicotinamide for further nucleotide synthesis is then strictly regulated by the limiting activity of both nicotinamide phosphoribosyltransferase and the deamidase. Most of the nicotinamide released by the hydrolysis of nucleotides will be methylated to Nl-methyl nicotinamide; this may either be excreted unchanged, or may undergo oxidation to methyl pyridone carboxamide. When the dietary intake of nicotinamide is very high, some may be excreted as nicotinamide N-oxide; similarly, a high intake of nicotinic acid leads to excretion of nicotinuric acid and other conjugates.

The hydrolysis of nicotinamide nucleotides is catalysed by NAD glycohydrolase (NADase). This is a microsomal enzyme with a very high activity, such that (at least in theory) it would be capable of hydrolysing the entire liver content of nucleotides within a few minutes; obviously, under physiological conditions, there is metabolic compartmentation to limit the activity of the enzyme. It may be that there are permeability barriers between the cytosol and the microsomes, or it may be that the relatively high K m of the enzyme (of the order of 0.1 mmol/1, G.M. McCreanor and D.A. Bender, unpublished observations) means that it only competes relatively feebly with the numerous enzymes in the cell that bind the nicotinamide nucleotides. This means that there will only be appreciable hydrolysis of NAD when all of the NAD requiring enzymes are more or less saturated with their cofactor

- i.e. when there is a relative excess of the nucleotides. Biochemistry of Tryptophan 165

There is another enzyme in Liver and other tissues that also hydroLyses nicotinamide nucleotides to yield free nicotinamide - this is the nuclear enzyme, poly-ADP-ribose synthetase. Its product is not ADP-ribose, as is the case with the gtycohydrolase, but rather a poLy-nucleotide. There is some evidence that this enzyme is involved in the nuclear DNA repair mechanisms that are important in response to carcinogen attack. A number of chemical carcinogens that act by alkylation of DNA Lead to increased activity of poly-ADP-ribose synthetase, and increased excretion of Nl-methyl nicotinamide [303]. It has been suggested that the depletion of tissue nicotinamide nucleotides that is caused by this activity might be severe enough to preci- pitate pellagra in people who are anyway only marginally adequately nourished with respect to tryptophan and niacin. The presence of carcinogenic myco- toxins in badly stored maize and millet may be a predisposing factor to the development of pellagra in people eating them [304]. However, studies with one of these compounds, the trichothecene T-2 toxin from Fusarium tricinctum, showed no effect on the excretion of Nlmethyl nicotinamide or on the Liver content of nicotinamide nucleotides in animals fed on a diet that provided only minimally adequate amounts of tryptophan and niacin (b.A. Bender and R. Schoental, unpublished observations).

Streptozotocin is a carcinogenic toxin which is widely used in experimental work to induce insulin-dependent diabetes; in small doses it Leads to a marked and specific destruction of the B-cells of the pancreas. Pretreat- ment of animals with picolinamide, which is an inhibitor of poLy-ADP-ribose synthetase, protects rats against both the streptozotocin-induced depletion of pancreatic NAD and also against the depression of pro-insulin synthesis that normally follows the administration of streptozotocin [3053.

Pellagra

The history of pellagra can be traced back to the 18th century, and the spread throughout southern Europe of the new cereal, maize, introduced from the Americas by the Conquistadores. The first description of pellagra was in 1735 by the Spanish physician Casal, and the name was coined in 1771 by Frapolli from the Italian 'pelle' = skin and 'agra' = rough. Thus the name of the disease is a description of its most prominent feature, a roughened skin. By the 19th century pellagra was common throughout Spain, Italy, the Balkans and the Ukraine, as well as north Africa and especially Egypt.

The arrival of pellagra in South Africa, where it is still a major problem of public health, coincided with an outbreak of rinderpest in 1897, which killed most of the cattle. This loss of animal foods led to a change in the life-stype of the Bantu, and from being a largely meat-eating community, they have become a maize-eating group, with little meat or milk. In some parts of South Africa today up to half the patients attending clinics for any reason at all show some signs of pellagra. Pellagra has also been found in other parts of Africa, and is a potential problem in much of sub- Saharan Africa, and especially the semi-arid Sahel region, although in north Africa it is very much less common than formerly.

The other part of the world where pellagra is an important problem today is India, both among the small groups eating maize as the dietary staple, and more importantly among the large numbers of people whose dietary staple is jowar, a variety of Sorghum vulgare. At certain times of the year it is reported that as many as 10% of the mental hospital admissions in Hyderabad are attributable to the mental disturbances of pellagra. 166 D.A. Bender

In the southern United States of America, pellagra became a serious problem following the American Civil War - the social and economic upheaval of the war led to large numbers of people subsisting on a very poor maize-based diet. Indeed, it was not until the entry of the United States into the second world war that increasing employment led to a sufficient degree of economic recovery to eradicate the disease. Pellagra continued to be a serious problem in Romania until the late 1950s.

It was in the southern United States that most of the work on the aetiology, prevention and cure of pellagra was performed; the history of this work, a classic of the development of the science of nutrition, has been told by Roe [291], Still [292] and others. In 1913 the Federal Government sent a team of investigators into the region to attempt to discover the cause and cure of pellagra. Although Funk, who coined the name 'vitamin', maintained that pellagra was a nutritional disease, there was, in 1913, a considerable school of thought that believed it to be an infectious disease. It was Goldberger's team who showed that the disease could not be transmitted to healthy people by any of the conventional means, but could be contracted by otherwise healthy individuals living away from the region in which it was endemic if they ate a maize-based diet. Zmprovement of the diet of pellagrins cured the disease, even in areas of high incidence. Having established that pellagra was due to a nutritional deficiency, it remained to discover which nutrient was lacking in the maize diet. Feeding additional protein was shown to be beneficial, and therefore it was assumed for some time that it was a general protein deficiency disease. Later it was held to be a tryptophan deficiency, and it was not until 1938 that it was shown that the vitamin niacin (either nicotinic acid or nicotinamide) would cure or prevent pellagra. It was not until 1947 that the relationship between tryptophan and niacin, and the formation of nicotinamide nucleotides from tryptophan, was elucidated, and the nature of pellagra as a disease due to dietary deficiency of both tryptophan and niacin was established.

It is interesting to note that nicotinic acid had been known to organic chemists since 1867, and was isolated from yeast extract and rice polishings by Funk in 1913. When Funk failed to show any activity of the compound against experimental polyneuritis (beri-beri) he lost interest in it, and little was done with the vitamin for many years. It was because when it was discovered in what was then known as 'vitamin B' nicotinic acid had already been identified that it was not assigned an alpha-numeric code in the list of B vitamins, but is always known by either its specific names (nicotinic acid and nicotinamide for the two compounds with vitamin activity) or by its generic descriptor, niacin. Somewhat confusingly, in the United States the term 'niacin' is also used specifically for the acid, and niacinamide for the amide form of the vitamin.

One of the problems in establishing the role of niacin in pellagra was that chemical analysis of maize shows the presence of a relatively large amount of the vitamin, and pellagra is extremely rare in Mexico, where again maize is the staple. The majority of the niacin in maize and sorghum is present combined with polysaccharides as niacytin, a complex which is not hydrolysed by mammalian digestive enzymes. This means that the niacin is not biologically available, although it is released by acid or alkaline hydrolysis, as is used in chemical analysis, so that analysis gives an unrealistic picture of the nutritional value of the cereals. The traditional Mexican method of preparing maize for making tortillas is to soak the grain overnight in an alkaline solution (lime water, a solution of calcium hydroxide). This Biochemistry of Tryptophan 167 causes hydrolysis of much, if not all, of the niacytin, and hence liberates nicotinic acid. Other maize-eating peoples do not have a similar practice, and so are unable to utilise the bound niacin of maize or sorghum.

Clinically, pellagra is characterised by the dermatitis for which it was named by Frapolli, a mental disturbance, which, although it is often called dementia is more akin to a depressive psychosis, diarrhoea (although this is not a constant feature, and constipation is sometimes encountered) and finally death. For this reason it is sometimes called the disease of the four D's - dermatitis, dementia, diarrhoea and death.

The dermatitis of pellagra is highly characteristic. It is erythematous and resembles severe sunburn. Typically, it develops only in areas of skin exposed to sunlight, and indeed it was the seasonal nature of pellagra, with a peak of onset at the beginning of summer, when people were living on the remains of the last harvest that had been stored throughout the winter, that led to the early belief that it was an infectious disease. The distribution of dermatitis around the face is especially characteristic, with clear, unaffected areas in the orbits of the eyes and other parts that are relatively protected from the sun, and overall a symmetrical distribution frequently called 'butterfly face distribution'. There is also typically an area of dermatitis around the neck where the skin is exposed to the sun - this is generally called 'Casall's collar' or 'necklace', in honour of the man who first described the disease. The exposed parts of the hands and fore-arms also develop dermatitis. When there is constant pressure or irri- tation, as at the elbows, knees, and perhaps at the wrists and ankles where clothes may rub, pellagrous skin lesions can develop in the absence of sunlight.

The biochemistry of the dermatitis of pellagra has not been satisfactorily explained in terms of the known biochemical abnormalities of the condition. There is known to be a lower than normal concentration of the histidine metabolite urocanic acid in the skin of pellagrins [306]. Urocanic acid is one of the major ultra-violet absorbing compounds in normal skin, and it is possible that this reduced skin content may predispose to increased burning on exposure to sunlight. It seems that this reduced skin content of urocanic acid may result from zinc deficiency, secondary to disturbed tryptophan metabolism, as discussed below. There are clinical similarities between the dermatitis of pellagra and that of the hereditary hypozincaemia disease, acrodermatitis enteropathica, as well as the dermatitis reported in the few cases of nutritional zinc deficiency [307]. A link between zinc depletion or deficiency and reduced skin content of urocanic acid has come from the studies of Hsu and Rubenstein [308], who have shown that in zinc deficiency there is an increase in the rate of histidine oxidation, with reduced tissue pools of both histidine and its immediate product, urocanic acid, due to an increase in the activities of both histidine ammonia lyase and urocanase. However, although this mechanism would fit the few observations of the biochemistry of skin in pellagra [306], it remains to be demonstrated that the disturbed tryptophan metabolism of pellagra is sufficiently severe to lead to a sufficient degree of zinc depletion to disturb histidine metabolism to any significant extent.

It is generally assumed that the depressive psychosis of pellagra is associated with impaired synthesis of serotonin in the central nervous system, as a result of reduced circulating concentrations of tryptophan, as discussed in Chapter 2. There is evidence of impaired synthesis of serotonin in blood 168 D.A. Bender platelets in pellagra [309] and the degree of this impairment seems to be related to the severity of the mental disturbance [310].

Mental disturbances are generally found in relatively advanced pellagra, and can range from mild delusions and confusion to full-blown florid psychosis. The stages in the development of the psychosis can be classified, in order of increasing severity, as follows [59]:

(i) In mild cases the intelligence is unimpaired, as are memory and orient- ation with respect to time and place. The patient may complain of headache, vertigo and sleep disturbances, and may show anxiety, depression, apathy, lack of volition, some degree of psychomotor retardation and thought disorder. Flight of ideas, hallucinations, religious speculation, and thoughts of suicide are common.

(ii) There is a general aggravation of the signs described in (i), and the psychomotor retardation may be intensified to the point of stupor. There may be some slowness in understanding and answering questions, but there is still no di sturbance of memory. There may be catatonic muscle movements.

(iii) The hallucinations increase, especially those involving fire, and there may be auditory and olfactory hallucinations as well. There is increasing confusion.

(iv) The hallucinations, confusion and motor agitation are increased. In some cases acute delirium and even sudden death have been reported.

(v) The symptoms are predominantly those of catatonia with stereotyped movements. In some cases there may be dementia, while in others intellec- tual activity is unimpaired.

(vi) The syndrome becomes predominantly one of anxiety psychosis with melan- cholia and intermittent periods of stupor.

(vii) The patient is severely demented and may show epileptiform convulsions.

These symptoms show obvious similarities to schizophrenia, but can normally be distinguished because of the intermittent periods of lucidity that occur in pellagra, even in the most severely demented patients. The involvement of serotoninergic systems, as discussed in Chapter 2, is obvious in the psychiatric picture of pellagra; this is due to both the low availability of tryptophan for serotonin synthesis and also possibly to an excessive concentration of leucine (and perhaps other branched chain amino acids) competing with such tryptophan as is available for uptake into the brain. Sorghum protein, and to a lesser extent also that of maize, contains a relative excess of leucine; the importance of this in the aetiology of pellagra is discussed below.

Non-nutritional pellagra

There are a number of conditions in which disturbances of tryptophan metabolism lead to the development of clinical pellagra despite an apparently adequate intake of niacin. These conditions suggest that under normal conditions the synthesis of nicotinamide nucleotides from tryptophan makes a major contribution to the body's niacin metabolism. The conditions include Biochemistry of Tryptophan 169 two reports of inborn errors of tryptophan oxidative metabolism [311, 312], Hartnup disease, carcinoid syndrome and the use of isoniazid as antituber- culosis therapy. a) Hartnup disease

Hartnup disease is characterised by massive aminoaciduria involving all of the large neutral amino acids, together with cerebellar ataxia and the development of pellagra. Indeed, in the first reported patient [313] and in a number of others, it was the skin lesions that were the presenting symptom on admission to hospital. The pellagra of Hartnup disease responds to the administration of niacin, but not to tryptophan.

The lesion of this rare genetic disease (some 40 -50 cases have been reported since the original description in 1956 [33]) seems to be one of the transport mechanism for the large neutral amino acids. In the intestinal mucosa the uptake of tryptophan and other neutral amino acids is very severely impaired, so that virtually the only source of these amino acids is by the uptake of di- and tripeptides which are hydrolysed intracellularly. Indeed, it was the study of Hartnup disease and the similar transport defect involving basic amino acids, cystinuria, that gave the first evidence that these small peptides are absorbed from the lumen of the intestinal tract intact, and are hydrolysed intracellularly [106]. In the kidney the same defect is apparent, so that the neutral amino acids that are filtered by the glomerulus cannot be reabsorbed, and are therefore excreted in the urine.

The net result of this failure to absorb tryptophan and other neutral amino acids and the massive losses in the urine is that the body's pool of tryptophan is very severely depleted, and little is available for oxidative metabolism. This means that the flux of metabolites through the oxidative pathway is very low, and hence, as discussed above, there will be very little synthesis of nicotinamide nucleotides from tryptophan. The reduced availability of tryptophan for uptake into the brain, together perhaps with a defect of the blood-brain barrier transport system like that in the intestinal mucosa and kidney will result in much reduced synthesis of serotonin, which presumably explains the neurological signs of the condition.

Patients with Hartnup disease excrete a number of abnormal tryptophan metabolites in the urine, including indican, indole, skatole and a number of conjugates. These seem to be due to the presence of large amounts of tryptophan in the large intestine, due to the failure of absorption in the small intestine, and therefore abnormal bacterial metabolism. These metabolites are absorbed passively across the large intestinal mucosa, and are excreted either unchanged or after conjugation in the liver. Sterilisation of the intestinal tract with such antibiotics as neomycin results in the loss of these 'exotic' tryptophan metabolites from the urine, and the appearance of large amounts of tryptophan in the faeces [33]. b) Carcinoid syndrome

Carcinoid is a tumour of the enterochromaffin cells of the intestinal mucosa, cells which normally synthesize serotonin. As the size of the tumour increases, so a considerable proportion of the body's tryptophan is diverted into the 5-hydroxyindole pathway, reducing the amount that is available for oxidative metabolism. This is especially true once metastases have become established in the liver and other tissues. Under normal conditions no more 170 D.A. Bender than about I% of the body's total tryptophan metabolism is by way of 5- hydroxylation, while in advanced carcinoid syndrome this may rise to as much as 60%; obviously this will lead to very much reduced synthesis of nicotinamide nucleotides from tryptophan. At the same time, the diarrhoea of carcinoid syndrome (possibly the result of much increased intestinal secretion of serotonin) will mean that there is some degree of malabsorption of such niacin as is available from the diet, thus exacerbating the condition, and increasing the likelihood of pellagra. Lehman [59] and others [314] have noted that the incidence of pellagra in carcinoid sypdrome is very much lower than might be expected in view of the disturbances of tryptophan metabolism, and have suggested that this may be because the patients generally seek medical attention at a relatively early stage, before there has been too serious a depletion of their tissue reserves of nicotinamide nucleotides. c) Isoniazid therapy for tuberculosis

Isoniazid (iso-nicotinic acid hydrazide) can form a biologically inactive adduct with pyridoxal phosphate, the metabolically active form of vitamin B 6. When the drug was first introduced into use for the treatment of tuberculosis, during the early 1950s, both peripheral neuritis and pellagra were common side effects. Both can be overcome by the provision of supplementary vitamin B6, whereas additional niacin, although it will cure or prevent the pellagra, does not affect the development of isoniazid-induced peripheral neuritis; this is presumably a direct effect of acute vitamin B 6 deficiency [315, 316, 317]. With the use of lower doses of isoniazid (especially in patients who are genetically slow acetylators of the drug) and the use of vitamin B 6 supplements together with the drug, isoniazid- induced pellagra has become somewhat of a rarity.

The mechanism of pellagragenesis by isoniazid is presumably inhibition of kynureninase by cofactor blockage; patients with isoniazid-induced pellagra show very much higher than normal excretion of xanthurenic and kynurenic acids and kynurenine after a tryptophan load than do normal subjects [318]. Isoniazid also inhibits tryptophan oxygenase [319, 320] and thus reduces the rate of entry of tryptophan into the oxidative pathway; this will again result in a reduced flux of metabolites through the pathway, and therefore reduced synthesis of nicotinamide nucleotides from tryptophan. There is no evidence that tryptophan oxygenase is a pyridoxal phosphate dependent enzyme and this inhibition by isoniazid and other hydrazine derivatives [319, 320] cannot be explained.

There are reports of the development of pellagra in patients being treated with isoniazid despite the provision of apparently adequate amounts of vitamin B 6. In one case [321] the patient was stated to be receiving 'a poor diet', and in the other two cases that have been reported [318, 322] the patients were Indian and strict vegetarians. It is likely that their diets were limiting in tryptophan and niacin, so that their status with regard to these two nutrients was anyway marginal before the initiation of isoniazid therapy. At the same time it is possible that their diets were marginal with respect to zinc, or at least available zinc, since the high phytate and dietary fibre content of the typical Indian vegetarian diet will lead to considerably reduced availability of zinc. As discussed above, zinc may be important in the development of the skin lesions of pellagra. In a preliminary study (D.A. Bender and E. Larsen, unpublished observations) three out of nine Indian vegetarians who had been treated with isoniazid and Biochemist~ of Tryptophan 171 vitamin B 6 for 6 - 9 months were excreting so little N1-methyl nicotinamide and methyl pyridone carboxamide that they would be classified as 'at risk' of developing pellagra by the generally used criteria for field studies. At this time they showed no clinical signs of pellagra. In all cases, the patients were excreting considerably larger amounts of 4-pyridoxic acid than normal, indicating that they were indeed taking their supplements of vitamin B6, and this was apparently being absorbed and metabolised in a normal way.

Two other hydrazine derivatives are also potentially capable of causing pellagra, although to date there have been no reports of any clinical signs. These drugs are Benserazide (Roche) and Carbidopa (Merck, Sharp & Dohme). They are used in the treatment of Parkinsonism; both drugs inhibit extra- cerebral aromatic amino acid decarboxylase, but do not cross the blood- brain barrier, and are therefore used in conjunction with dopa, reducing the extra-cerebral metabolism of the amino acid, and thus ensuring that a greater proportion is available for uptake into the brain, so that a smaller dose can be used. This means that there is a considerable reduction in the unwanted side effects of dopa therapy for Parkinsonism. Benserazide and Carbidopa form adducts with pyridoxal phosphate in the same way as does isoniazid, and thus cause inhibition of kynureninase; both drugs are also inhibitors of tryptophan oxygenase, both in vivo in experimental animals and in vitro with liver preparations [319, 320, 323, 325]. Although clinical pellagra has not been reported in any patient receiving either of these drugs, a group studied by Bender and coworkers [326] excreted only very small amounts of N1-methyl nicotinamide, and indeed a considerable number of those studied would be classified as either 'at risk' or even frankly niacin deficient by the generally accepted criteria for excretion of niacin metabolites.

The Pellagragenic Effect of Excess Dietary Leucine

Gopalan and Srikantia [327] noted that pellagra is common in India among people living on a diet based on jowar (Sorghum vulgare), but not among those eating a rice-based diet, although they claimed that the total tryptophan and niacin intake of these two groups was essentially similar. The tryptophan content of sorghum proteins is higher than that of maize, the traditionally pellagragenic cereal. They also suggested that the niacin of sorghum is available, unlike that of maize, which is chemically bound, and therefore nutritionally unavailable. However, this has been challenged [328] and it has been suggested that their method caused hydrolysis of niacytin before analysis. Less than 10% of the total niacin content of sorghum is chemically free, a figure that agrees with both chemical analysis and biological assay of the available niacin of maize [328].

Gopalan and Srikantia [327] suggested that the relative excess of leucine in the proteins of jowar might be responsible for the development of pellagra, and showed that feeding a relatively large amount of leucine (5 g/day) to volunteers led to an increase in the excretion of Nl-methyl nicotinamide, equivalent to the loss of 3.5 mg of niacin per day. Their initial suggestion was that the effect of excess leucine in the diet was to release nicotinamide nucleotides from the tissues, and increase catabolism.

In later studies [329] they showed that feeding dogs on a maize-based diet that was rich in leucine Led to the development of black tongue disease (the canine equivalent of pellagra), although feeding the animals on a diet based on 0paque-II hybrid maize, which contains less leucine did not. They 172 D.A. Bender

adjusted the tryptophan and lysine contents of the two diets, so as to overcome the other differences between Opaque-II and the traditional Deccan Hybrid varieties of maize, so that content of leucine was the major difference between the two diets. The workers in Hyderabad also showed [330] that there was a good correlation between the leucine content of different varieties of sorghum and the induction of black tongue disease in dogs.

Further studies from the Indian Institute for Medical Research in Hyderabad have shown that the addition of between 1.5 - 3% additional leucine to the diets of monkeys and rats led to impaired synthesis of nicotinamide nucleotides [331, 332] as well as the increase in catabolism that was originally reported [327]. Feeding supplements of leucine to human volunteers led to reduced synthesis of nicotinamide nucleotides in erythrocytes [333] and changes in the electroencephalographic pattern of pellagrins, together with the development of the characteristic signs of mental disturbance of advanced pellagra [334]. Feeding supplements of isoleucine together with leucine to monkeys [331], human beings [333, 334] and rats [332] reversed the effects of leucine administration, suggesting that the effect is at least partially one of amino acid imbalance [335].

Feeding rats on diets providing 3% excess leucine leads to an increase in the activity of tryptophan oxygenase, and an increase in the total oxidation of tryptophan to carbon dioxide (presumably therefore an increase in the activity of picolinic carboxylase) together with a reduction in the activity of kynureninase and quinolinate phosphoribosyltransferase [332, 336, 337]. Thus, the synthesis of nicotinamide nucleotides from tryptophan seems to be impaired by the dietary excess of leucine, although there does not seem to be any effect on the utilisation of such free niacin as is available from the diet.

A number of workers in other centres have challenged the Hyderabad findings, and have suggested that there is no pellagragenic effect of excess leucine in the diet. Nakagawa and coworkers [338] were unable to show any effect of the administration of leucine on the excretion of Nl-methyl nicotinamide, methyl pyridone carboxamide, nicotinic acid, quinolinic acid or 5-hydroxy indoleacetic acid in a group of volunteers who were fed on a diet in which the sole nitrogen source was a mixture of amino acids. Manson and Carpenter [339, 340] showed no effect of excess leucine in the diet on the growth of dogs, chicks or rats, or on the urinary excretion of Nl-methyl nicotinamide, when 1.5% additional leucine was incorporated in diets that were relatively poor in tryptophan and niacin.

Yamada and coworkers [341] did show a significant decrease in the liver content of nicotinamide nucleotides when rats were fed on diets providing 5 or 10% casein following the addition of 5% additional leucine. This effect was only present when the diet was niacin-free; when there was adequate niacin in the diet they showed no deleterious effect of leucine supplementation. Again they showed no effect of leucine on the excretion of nicotinic acid or Nl-methyl nicotinamide.

Magboul and Bender [342] showed that when rats were fed on maize-based diets that were only marginally adequate with respect to tryptophan and niacin the addition of 1.5% leucine led to a significant depletion of liver and blood nicotinamide nucleotides, although there was no significant effect on the urinary excretion of Nl-methyl nicotinamide. They also showed that this effect was only apparent when the niacin content of the diet was such that it provided for less than half of the requirement; the effect of adding Biochemistry of Tryptophan 173 leucine was most marked when the diets provided virtually no pre-formed niacin, so that the animals were more or less totally reliant on synthesis of nicotinamide nucleotides from tryptophan. This suggests that the effect of leucine is on the metabolism of tryptophan, and not on the utilisation of nicotinamide or nicotinic acid.

Magboul [343] showed that feeding leucine to animals, or the addition of leucine or its oxo-acid (2-oxo-isocaproic acid) to incubations in vitro, had no effect on the activities of the enzymes involved in the incorporation of niacin into nucleotides (nicotinate and nicotinamide phosphoribosyl- transferases and nicotinamide deamidase) or on quinolinate phosphoribosyl transferase. He showed an increase in the activity of tryptophan oxygenase after the administration of leucine, or when it was added in vitro, suggesting that leucine may activate the enzyme, possibly by an interaction with the non-catalytic (activator) substrate binding site as well as a reduction in the activity of kynureninase. There was a reduction in the activity of kynurenine hydroxylase in the presence of 2-oxo-isocaproate, but only at concentrations very much higher than are likely to be encountered in vivo, while the effects of leucine were apparent at concentrations of the same order as those that occur under normal physiological conditions.

Rats fed on diets containing 1.5% additional leucine excreted significantly more kynurenine than did control animals, reflecting the increased entry of tryptophan into the oxidative pathway as the result of increased activity of tryptophan oxygenase and the reduced activity of kynureninase. When the animals were fed on diets providing 1.5% additional leucine there was no effect on the activity of picolinic carboxylase.

It thus seems likely that the hypothesis proposed by Gopalan and Srikantia [328] that excess leucine in some varieties of sorghum may be a precipitating factor in pellagra is correct, especially when the intake of tryptophan and niacin is anyway only marginally adequate. In view of the continuing high incidence of pellagra in India, especially in areas where jowar is the dietary staple, and in parts of Africa where maize is the staple, as well as the resurgence of pellagra in countries from which it was previously absent, as a result of declining living standards [344] this is potentially important to plant geneticists, who are working to develop improved strains of food crops. As well as higher yield per hectare and a higher protein content of the grain, it may be desirable that new strains of maize and sorghum should have a lower content of leucine (or possibly a higher content of isoleucine) than do existing varieties.

Picolinic Acid and the Absorption of Zinc

The immediate product of picolinic carboxylase in the pathway that leads to the total oxidation of tryptophan is aminomuconic semialdehyde. This is normally oxidised to aminomuconic acid, and thence onwards to acetyl CoA, as shown in Figure 11. However, when the dehydrogenase is saturated with its substrate, as occurs when the flux of metabolites through the pathway is high, aminomuconic semialdehyde can undergo a non-enzymic cyclisation akin to that of acroleyl aminofumarate to quinolinic acid; in this case the produce is picolinic acid.

The physiological significance of picolinic acid has only recently been recognised. It has been known since about 1973 that there is at least one low molecular weight compound in the intestinal lumen, apparently from the 74 D.A. Bender

,~COOH b =-- COOH "-~ "-~ NAD

O:CH ,,,~NH2 acroleyl aminofumarate ~ COOH COOH quinolinic acid

a ~' CO2 b ID NH 2 O=CH / aminomuconic semialdehyde COOH COOH picolinic acid A°

T -~NADH

HOOC~NH2 aminomuconic acid COOH

NADP

NH 3 t ~NADPH

= 2 x CH s CO CoA HO0 C~ 0 oxo-adipic~ acid COOH

(a) Picolinic carboxylase (amino-carboxy-muconate semialdehyde carboxy-lyase, EC 4.1.1.45)

(b) Non-enzymic cyclisation

Figure 11

The oxidation of acroleylaminofumarate the pathway for total oxidation of tryptophan Biochemistry of Tryptophan 175 exocrine pancreatic secretion, that is essential for the normal intestinal absorption of zinc - the zinc binding ligand [345]. The role of this ligand was elucidated by studies of the rare genetic disease acrodermatitis enteropathica. For many years the only treatment available for infants with this condition was to feed them on human milk, and to continue to provide human milk after weaning, since otherwise the skin lesions recurred.

The protective effect of diiodoquin was discovered empirically, and was later shown to be due to its ability to chelate zinc, and thus promote the absorption of the mineral from the intestinal lumen. Provision of relatively large supplements of inorganic zinc (many times the estimated normal daily requirement) is also beneficial [346]. This suggests that the lesion of acrodermatitis enteropathica is one of the absorption of zinc, and in view of the role of chelation of the mineral by an endogenous zinc binding ligand in normal subjects [345], it is probable that it is synthesis of this ligand that is defective.

The beneficial effect of human milk in treating acrodermatitis enteropathica, and the ineffectiveness of cow's milk, despite a similar or even higher total content of zinc, suggests that human milk may contain a zinc binding ligand which is absent from cow's milk, so that the zinc of human milk can be absorbed even in the absence of intestinal zinc binding ligand. Evans and Johnson [347] demonstrated the presence of such a zinc binding ligand in human milk; it is indeed absent from cow's milk [348]. In the rat, zinc binding ligand is present in the milk during the early stages of lactation, when the gastro-intestinal tract of the pups appears to contain no endogenous ligand, and the amount of ligand in the milk falls from about the tenth day of lactation, at the same time as the synthesis of zinc binding ligand increases in the pups [349]. This is further evidence that the provision of exogenous zinc binding ligand can enhance the absorption of zinc in the absence of endogenous ligand.

Evans and Johnson [350] isolated the zinc binding ligand from human milk and showed that it was picolinic acid. The picolinic acid content of a pooled sample of human milk was 308 umol/1, whereas in two separate samples of cow's milk they found only 20 umol/l and less than 1 umol/l. They were unable to detect any picolinic acid or zinc binding ligand activity in four samples of infant formula preparations based on cows' milk [351].

They subsequently showed that in rats fed on low protein diets the absorption of zinc was enhanced by the addition of either picolinic acid or tryptophan to the diet [352, 353]. Bender (unpublished observations) has shown that animals fed on a low tryptophan diet excrete significantly less zinc in the urine than do those receiving an otherwise identical diet but providing an adequate amount of tryptophan. There is thus a considerable amount of evidence that picolinic acid is involved in the normal intestinal absorption of zinc, and that the availability of tryptophan for the synthesis of picolinic acid may be a limiting factor under some conditions.

Further evidence of the involvement of picolinic acid in the intestinal absorption of zinc has come from study of a patient with an atypical form of acrodermatitis enteropathica. Despite the presence of more or less normal plasma concentrations of zinc, this patient showed the typical skin lesions of the condition, and responded to the administration of 60 mg of inorganic zinc daily. Because the diagnosis of acrodermatitis enteropathica was in doubt, the patient was treated for a time with a pancreatic extract ('Viokase'), which also gave good control of the symptoms, despite a daily 176 D.A. Bender

intake of only 5 mg of zinc. Further investigation showed that the pancreatic extract contained a relatively large amount of picolinic acid, and the patient showed a good response to the administration of small amounts of zinc picolinate [354].

Picolinic acid is formed non-enzymically from aminomuconic semialdehyde, and therefore it is likely that significant amounts will only be formed when the flux of metabolites is so great that the enzyme that competes with the non- enzymic cyclisation for aminomuconic semialdehyde, the dehydrogenase, is saturated with its substrate. Thus, as in the case of quinolinic acid synthesis from acroleyl aminofumarate, the synthesis of picolinic acid will increase dramatically with the rate of tryptophan oxidation, and under conditions of low tryptophan oxidation there will be relatively little synthesis of picolinic acid. This will presumably lead to reduced absorption of zinc.

Krieger [307] has noted that there are similarities between the skin lesions of pellagra and acrodermatitis enteropathica. The most obvious difference is in the distribution; pellagrous dermatitis affects chiefly those areas of the skin that are exposed to sunlight, as discussed above, while in acrodermatitis enteropathica there is no apparent photosensitivity; this is probably because acrodermatitis enteropathica is diagnosed early in life, when the infant has not been much exposed to sunlight, whereas pellagra develops later in life. She has suggested that the underlying pathology of the dermatitis in both conditions may be the same. It was noted above that in pellagra the skin content of urocanic acid is very low [306], and that zinc deficiency leads to increased metabolism of histidine and urocanic acid [308] and so could account for this. Vitamin B 6 deficiency would also lead to reduced flux of tryptophan metabolites beyond kynureninase, and Krieger [307] has noted that in cases of vitamin B 6 deficiency there are again skin lesions that resemble those of both pellagra and acrodermatitis enteropathica.

It is perhaps fitting to conclude a review of some aspects of the biochemistry of an amino acid that was discovered by 'accident' by noting that what may well be its two most important metabolites, quinolinic acid as the precursor of nicotinamide nucleotides and picolinic acid as the ligand required for the absorption of zinc, are both 'accidental' metabolites. Both are formed by non-enzymic reactions that must compete with enzyme-catalysed reactions, so that it is only when the flux of metabolites through the pathway of tryptophan oxidation is so great that these enzymes are saturated that there is any significant formation of these products. References

I Hopkins F.G. & Cole S.W. (1901) On the proteid reaction of Adamkiewicz, with contributions to the chemistry of glyoxylic acid. Proc. Roy. Soc. 68 21 2 Hopkins F.G. & Cole S.W. (1901) A contribution to the chemistry of the proteids: Part I: a preliminary study of a hitherto undescribed product of tryptic digestion. J. Physiol. 27 428 3 Willcock E.G. & Hopkins F.G. (1905) The importance of individual amino acids in metabolism: observations on the effect of adding tryptophane to a dietary in which zein is the sole nitrogenous constituent. J. Physiol. 35 88 4 Hopkins F.G. & Cole S.W. (1903) A contribution to the chemistry of the proteids: Part II: the constitution of tryptophane and the action of bacteria upon it. J. Physiol. 29 451 5 Parli C.J., Krieter P. & Schmidt B. (1980) Metabolism of 6-chloro- 3-hydroxyanthranilic acid, a potent inhibitor of 3-hydroxyanthranilic acid oxidase. Archs. Biochem. Biophys. 203 161 - 166 6 Cohen S.M., Arai M., Jacobs J.B. & Friedell G.H. (1979) Promoting effect of saccharin and DL-tryptophan in urinary bladder carcinogenesis. Cancer Res. 39 1207 - 1217 7 Gailani S., Murphy G., Kenny G., Nussbaum A. & Silvernail P. (1973) Studies on tryptophan metabolism in patients with bladder cancer. Cancer Res. 33 1071 - 1077 8 E1-Toukhy M.A., Ebied S.A., E1-Zoghby S.M., Abdel-Rabbo H., Hamouda N. & E1-Gohary Y.A. (1980) Effect of oxamniquine therapy on kynurenine metabolism of normal and S. mansoni infected mice. Biochem. Pharmacol. 29 2513 - 2515 9 Nagao M., Yahagi T., Kawachi T., Sugimura T., Kosugu T., Tsuji K., Wakabayaishi K., Mizuzuki S. & Matsumtot T. (1977) Comutagenic action of norharman and harman. Proc. Jap. Acad. Sci. 53 95 - 98 10 Cotlier E., Sharma Y.R., Zuckerman J., Pucklin J., Teasley B. & Irvine J. (1981) Plasma tryptophan in humans with diabetic and senile cataracts. Exp. Eye Res. 33 247 - 252 11 Kotake Y. & Murakami E. (1971) A possible diabetogenic role for tryptophan metabolites and effects of xanthurenic acid on insulin. Amer. J. Clin. Nutr. 24 812 - 829 12 Adams P.W., Folkard J., Wynn V. & Seed M. (1976) Influence of oral contraceptives, pyridoxine and tryptophan on carbohydrate metabolism. Lancet 1 759 - 764 13 Sandstead H.H., Prasad A.S., Schulbert A.R., Farid Z., Miale A., Bassilly S. & Darby W.J. (1967) Human zinc deficiency: endocrine manifestations and response to treatment. Amer. J. Clin. Nutr. 20 422 - 442 14 Hopkins L.L., Ransome-Kuti O. & Najaj A.S. (1968) Improvement of impaired carbohydrate metabolism by chromium (III) in malnourished infants. Amer. J. Clin. Nutr. 21 203 - 211 15 Jeebjeebhoy K.N., Chu R.C., Marliss E.B., Greenberg G.R. & Bruce-Robertson A. (1977) Chromium deficiency, glucose intolerance and neuropathy reversed by chromium supplementation in a patient receiving long-term total parenteral nutrition. Amer. J. Clin. Nutr. 30 531 - 538

177 178 D.A. Bender

16 Schwartz K. & Mertz W. (1957) A glucose tolerance factor and its differentiation from factor 3. Archs. Biochem. Biophys. 72 515 - 518 17 Schwartz K. & Mertz W. (1959) Chromium (III) and the glucose tolerance factor. Archs. Biochem. Biophys. 85 292 - 295 18 Mertz W. & Schwartz K. (1959) Relation of glucose tolerance factor to impaired intravenous glucose tolerance of rats on stock diet. Amer. J. Physiol. 196 318 - 322 19 Mertz W. (1974) Biological function of nicotinic acid - chromium complexes. Fed. Proc. 33 659 20 Toepfer E.W., Mertz W., Polansky M.M., Roginski E.E. & Wolf W.R. (1977) Preparation of.chromium-containing material of glucose tolerance factor activity from brewer's yeast extracts and by synthesis. J. Agric. Food Chem. 25 162 - 166 21 Glinsmann W.H. & Mertz W. (1966) Effect of trivalent chromium on glucose tolerance. Metabolism 15 510 - 520 22 Ray P.D., Foster D.O. & Lardy H.A. (1966) Paths of carbon in gluconeogenesis and lipogenesis (iv) inhibition by L-tryptophan of hepatic gluconeogenesis at the level of phosphoenolpyruvate formation. J. Biol. Chem 241 3904 - 3908 23 Smith, S.A. & Pogson, C.I. (1977) Tryptophan and the control of plasma glucose concentrations in the rat. Biochem. J. 168 495 - 506 24 Veneziale C.M., Walter P., Kneer N. & Lardy H.A. (1967) Influence of L-tryptophan and its metabolites on gluconeogenesis in the isolated perfused liver. Biochemistry 6 2129 - 2138 25 Smith S.A., Elliot K.R.F. & Pogson, C.I. (1978) Differential effects of tryptophan on glucose synthesis in rats and guinea pigs. Biochem. J. 176 817 - 825 26 Snoke R.E., Johnston J.B. & Lardy H.A. (1971) Response of phosphopyruvate carboxylase to tryptophan metabolites and metal ions. Eur. J. Biochem. 24 342 - 346 27 McDaniel H.G., Reddy W.J. & Boshell B.R. (1972) The mechanism of inhibition of phosphoenolpyruvate carboxylase by quinolinic acid. Biochim. Biophys. Acta 276 543 - 550 28 McDaniel H.G., Boshell B.R. & Reddy W.J. (1973) Hypoglycaemic action of tryptophan. Diabetes 22 713 - 718 29 Smith S.A., Carr P.A. & Pogson C.I. (1980) The metabolism of L-tryptophan by isolated rat liver cells: quantification of the relative importance of, and the effect of nutritional status on, the individual pathways of tryptophan metabolism. Biochem. J. 192 673 - 686 30 Smith S.A. & Pogson C.I. (1977) Tryptophan and the control of plasma glucose concentrations in the rat. Biochem. J. 168 495 - 506 31 Munoz-Clares R.A., Lloyd P., Lomax M.A., Smith S.A. & Pogson C.I. (1981) Tryptophan metabolism and its interaction with gluconeogenesis in mammals: studies with the guinea pig, Mongolian gerbil and sheep. Archs. Biochem. Biophys. 209 713 - 717 32 Sourkes T.L. (1971) Alpha-methyltryptophan and its actions on tryptophan metabolism. Fed. Proc. 30 897 - 903 33 Jepson J.B. (1978) Hartnup Disease. Chapter 66, pp 1563 - 1578 in The Metabolic Basis of Inherited Disease. Stanbury J.B., Wyngaarden J.B. & Fredrickson D.S., Eds. McGraw-Hill, New York. Biochemistry of Tryptophan 179

34 Drummond K., Michael A., Ulstrom A. & Good R. (1964) Blue diaper syndrome: familial hypercalcaemia with nephrocalcinosis and indicanuria. Amer. J. Med. 37 928 - 948 35 Falck B. (1962) Histofluorescence method for brain monoamines. Acta Physiol. Scand. 56 Suppl. 197 36 Johnson D.A., Cho T.M. & Loh H.H. (1977) Identification of 5-hydroxytryptamine binding substances from the rat brain stem. J. Neurochem. 29 1105 - 1109 37 Friedman P.A., Kappelman A.H. & Kaufman S. (1972) Partial purification and characterization of tryptophan hydroxylase from rabbit hindbrain. J. Biol. Chem. 247 4165 - 4173 38 Kaufman S., Holtzman N.A., Milstein S., Butler I.J. & Krumholz A. (1975) Phenylketonuria due to a deficiency of dihydrobiopterin reductase. New Engl. J. Med. 293 785 - 790 39 Kaufman S., Berlow S., Summer G.K., Milstein S., Schulman J.D., Orloff S., Spielberg S. & Pueschel S. (1978) Hyperphenylalaninaemia due to a deficiency of biopterin. New Engl. J. Med. 299 673 - 679 40 Koslow S.H. & Butler I.J. (9177) Biogenic amine synthesis defect in dihydrobiopterin reductase deficiency. Science 198 522 - 523 41 Niederwieser A., Curtius H-C., Bieri J., Schircks B., Viscontini M. & Schaub J. (1979) Atypical phenylketonuria caused by 7,8-dihydrobiopterin synthetase deficiency. Lancet 1 131 - 133 42 Gal E.M. (1974) Tryptophan 5-hydroxylase: function and control. Adv. Biochem. Psychopharmacol. 11 1 - 11 43 Grahame-Smith D.G. (1971) Studies in vivo on the relationships between brain tryptophan, brain 5-hydroxytryptamine synthesis and hyperactivity in rats treated with monoamine oxidase inhibitors and L-tryptophan. J. Neurochem. 18 1053 - 1066 44 Bensinger R.E., Klein D.C., Weller J.L. & Lovenberg W. (1974) Radiometric assay of total tryptophan hydroxylation by intact cultured pineal glands. J. Neurochem. 23 111 - 117 45 Noguchi T., Nishino M. & Kido R. (1973) Tryptophan 5-hydroxylase in rat intestine. Biochem. J. 131 375 - 380 46 Hamon M., Bourgoin S., Hery F., Ternaux J.P. & Glowinski J. (1976) In vivo and in vitro activation of soluble tryptophan hydroxylase from rat brain stem. Nature 260 61 - 63 47 Hamon M., Bourgoin S., Hery F. & Simmonet G. (1978) Phospholipid- induced activation of tryptophan hydroxylase from the rat brainstem. Biochem. Pharmacol. 27 915 - 922 48 Hamon M., Bourgoin S., Artaud F. & Glowinski J. (1979) The role of intra-neuronal 5-hydroxytryptamine and of tryptophan hydroxylase activation in the control of 5-hydroxytryptamine synthesis in rat brain slices incubated in K + enriched medium. J. Neurochem. 33 1031 - 1042 49 Knapp S. & Mandell A.J. (I 979) Conformational influences on brain tryptophan hydroxylase by sub-micromolar calcium: opposite effects of equimolar lithium. J. Neural Transm. 45 1 - 15 50 Yamauchi T. & Fujisawa H. (1981) A calmodulin-dependent protein kinase that is involved in the activation of tryptophan 5-mono- oxygenase is specifically distributed in brain tissues, FEBS Lett. 129 117 - 119 51 Koe B.K. (1971) Tryptophan hydroxylase inhibitors. Fed. Proc. 30 886 - 896 80 D.A. Bender

52 Koe K.B. & Weissman A. (1966) p-Chlorophenylalanine: a specific depletor of brain serotonin. J. Pharm. Exp. Ther. 154 499 - 516 53 Antonas K.N., Coulson W.F. & Jepson J.B. (1974) Simulation of phenylketonuria in rats by extended p-chlorophenylalanine treatment. Biochem. Soc. Transact. 2 105 - 107 54 Anon. (1975) Diagnosis of malignant carcinoid syndrome. Brit. Med. J. 3 122 - 123 55 Sandler M. (1968) The role of 5-hydroxy-indoles in the carcinoid syndrome. Adv. Pharmacol 6B 127 - 142 56 Engelman K., Lovenberg W & Sjoedsma A. (1967) Inhibition of 5-hydroxytryptamine synthesis in patients with carcinoid syndrome. New Engl. J. Med. 277 1103 - 1108 57 Engelman K., Jequier E., Lovenberg W. & Sjoedsma A. (1967) Inhibition of serotonin synthesis by p-chlorophenylalanine in animals and in patients with the carcinoid syndrome. Clin. Res. 15 316 58 Warner R.R.P. (1972) Plasma tryptophan levels in carcinoid syndrome. Gastroenterol. 62 825 59 Lehman J. (1972) Mental and neuromuscular symptoms in tryptophan deficiency. Acta Psych. Scand. Suppl. 237 60 Edwards D.J. & Blau K. (1972) The in vivo formation of p-chlorophenyl ethylamine in young rats treated with p-chlorophenylalanine. J. Neurochem 19 1829 - 1831 61 Cremata V.Y. & Koe B.K. (1966) Clinical-pharmacological evaluation of p-chlorophenylalanine: a new 5-hydroxytryptamine-depleting agent. Clin. Pharm. Ther. 7 768 - 776 62 Udenfriend S., Clark C.T. & Titus E. (1953) 5-Hydroxytryptophan decarboxylase: a new route of metabolism of tryptophan. J. Amer. Chem. Soc. 75 501 - 502 63 Clark C.T., Weissbach H. & Udenfriend S. (1954) 5-Hydroxytryptophan decarboxylase: preparation and properties. J. Biol. Chem. 210 139 - 148 64 Yuwiler A., Geller E. & Eiduson S. (1959) Studies on 5-hydroxytryptophan decarboxylase (i) in vitro inhibition and substrate interaction. Archs. Biochem. 80 162 - 173 65 Rosengren E. (1960) Are dopa decarboxylase and 5-hydroxytryptophan decarboxylase individual enzymes ? Acta Physiol. Scand. 49 364 - 369 66 Coulson W.F., Bender D.A. & Jepson J.B. (1969) Multiple electrophoretic peaks of rat liver decarboxylase for 3,4-dihydroxyphenylalanine and 5-hydroxytryptophan. Biochem. J. 115 63 - 64p 67 Bender D.A. & Coulson W.F. (1972) Variations in aromatic amino acid decarboxylase activity towards dopa and 5-hydroxytryptophan caused by pH changes and denaturation. J. Neurochem. 19 2801 - 2810 68 Christenson J.G., Dairman W. & Udenfriend S. (1970) Preparation and properties of a homogeneous aromatic amino acid decarboxylase from hog kidney. Archs. Biochem. Biophys. 141 356 - 367 69 Christenson J.G., Dairman W. & Udenfriend S. (1972) On the identity of dopa decarboxylase and 5-hydroxytryptophan decarboxylase. Proc. Nat. Acad. Sci. USA 69 343 - 347 70 Sims K.L. & Bloom F.E. (1973) Rat brain dopa and 5-hydroxytryptophan decarboxylase activities: differential effects of 6-hydroxy-dopamine. Brain Res. 49 165 - 175 71 Sims K.L., Davis G.A. & Bloom F.E. (1973) Activities of dopa and 5-hydroxytryptophan decarboxylases in rat brain: assay characteristics and distribution. J. Neurochem. 20 449 - 464 Biochemistry of Tryptophan 181

72 Dairman W., Horst W.D., Marchelle E. & Bautz G. (1975) The proportionate loss of 3,4-dihydroxyphenylalanine and 5-hydroxytryptophan decarboxylating activity in rat central nervous system following intracisternal administration of 5,6-dihydroxytryptamine or 6-hydroxydopamine. J. Neurochem. 24 619 - 623 73 Melamed E., Hefti F. & Wurtman R.J. (1980) L-3,4-dihydroxyphenylalanine and L-5-hydroxytryptophan decarboxylase activities in rat striatum: effect of selective destruction of dopaminergic or serotoninergic input. J. Neurochem. 34 1753 - 1756 74 Robins E., Robins J.M., Croninger A.B., Moses S.G., Spencer S.J. & Hudgens R.W. (1967) The low level of 5-hydroxytryptophan decarboxylase in human brain. Biochem. Med. 1 240 - 251 75 Lindqvist M., Carlsson A. & Kehr W. (1972) Determination of dopa and 5-hydroxytryptophan in the brain after decarboxylase inhibition. Acta Pharm. Tox. 31 Suppl. 1 30 76 Bowen D.M., White P., Flack R.H.A, Smith C.B. & Davison A.N. (1974) Brain decarboxylase activities as indices of pathological change in senile dementia. Lancet 1 1247 - 1249 77 Lovenberg W., Weissbach H. & Udenfriend S. (1962) Aromatic L-amino acid decarboxylase. J. Biol. Chem 237 89 - 93 78 Awapara J., Perry T.L., Hanley C. & Peck E. (1964) Substrate specificity of dopa decarboxylase. Clin. Chim. Acta 10 286 - 289 79 Murali D.K. & Radhakrishnan A.N. (1970) Purification and properties of 5-hydroxytryptophan decarboxylase from monkey small intestine. Ind. J. Biochem. 7 13 - 18 80 Saavedra J.M. & Axelrod J. (1972) A specific and sensitive enzymatic assay for tryptamine in tissues. J. Pharm. Exp. Ther. 182 363 - 369 81 Marsden C.A. & Curzon G. (1974) Effects of lesions and drugs on tryptamine andserotoninin the rat brain. Biochem. Soc. Transact. 2 264 - 265 82 Saavedra J.M. & Axelrod J. (1974) Brain tryptamine and the effects of drugs. Adv. Biochem. Psychopharmacol. 10 135 - 139 83 Warsh J.J., Coscina D.V., Godse D.D. &Chan P.W. (1979) Dependence of brain tryptamine formation on tryptophan availability. J. Neurochem. 32 1191 - 1196 84 Weil-Malherbe H. (1976) Amine formation from L-tryptophan in brain slices. J. Neurochem. 27 829 - 834 85 Snodgrass S.R. & Iversen L.L. (1974) Formation and release of [3H]tryptamine from [3H]tryptophan in rat spinal cord slices. Adv. Biochem. Psychopharmacol 10 141 - 150 86 Young S.N., Anderson G.M., Gauthier S. & Purdy W.C. (1980) The origin of indoleacetic acid and indolepropionic acid in rat and human cerebrospinal fluid. J. Neurochem. 34 1087 - 1092 87 Deguchi T. (1975) Ontogenesis of a biological clock for serotonin: acetyl coenzyme A N-acetyl transferase in pineal gland of rat. Proc. Nat. Acad. Sci. USA 72 2814 - 2818 88 Deguchi T. (1982) Sympathetic regulation of circadian rhythm of serotonin-N-acetyltransferase in pineal gland of intact rat. J. Neurochem. 38 797 - 802 89 Rudeen R.K., Reito R.J. & Vaughan M.L. (1978) Pineal serotonin N-acetyltransferase activity in four mammalian species. Neuroscience Letts. 1 225 - 229 182 D.A. Bender

90 Yang H.Y. & Neff N.H. (1976) Brain N-acetyltransferase: substrate specificity, distribution and comparison with enzyme activity from other tissues. Neuropharmacol. 15 561 - 564 91 Satake N. & Morton B. (1979) Pineal hydroxyindole-O-methyl transferase: mechanism and inhibition by scotophobin A. Pharmacol. Biochem. Behav. 10 457 - 462 92 Wetterberg L., Arendt J., Paunier L., Sizonenko P.C., van Donselaar W. & Heyden T. (1976) Human serum melatonin changes during the menstrual cycle. J. Clin. End. Metab. 42 185 - 188 93 Petterburg L.J. & Reiter R.J. (1980) Effect of photoperiod and melatonin on testicular development in the white-footed mouse, Peromyscus leucopus. J. Reprod. Fertil. 60 209 - 212 94 Darrow J.J., Davis F.C., Elliott J.A., Stetson M.H., Turek F.W. & Menaker M. (1980) Influence of photoperiod on reproductive development in the golden hamster. Biol. Reprod. 22 443 - 450 95 Fellenberg A.J., Phillipou G. & Seamark R.F. (1980) Measurement of urinary production rates of melatonin as an index of human pineal function. Endocr. Res. Commun. 7 165 - 175 96 Lynch H.J., Ozaki Y., Shakal D. & Wurtman R.J. (1975) Melatonin excretion of man and rats: effect of time of day, sleep, pinealectomy and food consumption. Int. J. Biometeor. 19 267 - 279 97 Ozaki Y. & Lynch H.J. (1976) Presence of melatonin in plasma and urine of pinealectomized rats. Endocrinol. 99 641 - 644 98 Carter S.J., Land C.A., Smith I., Leone R.M., Silman R.E., Hooper R.J.L., Larsen-Carter D.L., Finnie M.D.A. & Mullen P.E. (1979) 5-Methoxytryptophol in rat pineal glands and other tissues. Progr. Brain Res. 52 267 - 269 99 Smith J.A. (1978) The pineal gland: its possible significance in schizophrenia, pp 105 - 110 in The Biological Basis of Schizophrenia, Hemmings G. & Hemmings W.A., Eds. MTP Press. Lancaster. 100 Guchait R.B. (1976) Biogenesis of 5-methoxy-N,N-dimethyl tryptamine in human pineal gland. J. Neurochem. 26 187 - 190 101 Fowler C.J. & Tipton K.F. (1982) Deamination of 5-hydroxytryptamine by both forms of monoamine oxidase in the rat brain. J. Neurochem. 38 733 - 736 102 Green A.R. & Curzon G. (1968) Decrease of 5HT in the brain provoked by hydrocortisone and its prevention by allopurinol. Nature 220 1095 - 1097 103 Eccleston D., Ashcroft G.W. & Crawford T.B.B. (1965) 5-Hydroxyindole metabolism in rat brain: a study of intermediate metabolism using the technique of tryptophan loading: (ii) applications and drug studies. J. Neurochem. 12 493 - 503 104 Tagliamonte A., Tagliamonte P., Perez-Cruet J. & Gessa G.L. (1971) Increase of brain tryptophan caused by drugs which stimulate 5-hydroxytryptamine synthesis. Nature-New Biology 229 125 - 126 105 Meister A. (1973) On the enzymology of amino acid transport. Science 180 33 - 39 106 Bender D.A.(1983) Chapter 5, Metabolic and Pharmacological Studies in The Chemistry and Biochemistry of the Amino Acids, Barrett G.C., Ed. Chapman and Hall, London. Biocnemistry of Tryptophan 183

107 Blasberg R. & Lajtha A. (1966) Heterogeneity of the mediated transport systems of amino acid uptake in incubated slices of brain. Brain Res. 1 86 - 104 108 Oldendorf W.H. (1970) Measurement of brain uptake of radio-labelled substances using tritiated water internal standard. Brain Res 24 372 - 376 109 Oldendorf W.H. (1971) Brain uptake of radio-labelled amino acids, amines and hexoses after arterial injection. Amer. J. Physiol. 221 1629 - 1639 110 Lajtha A. & Toth J. (1961) The brain barrier system: (ii) uptake and transport of amino acids by the brain. J. Neurochem. 8 216 - 225 111 Lajtha A. & Toth J. (1963) The brain barrier system: (v) stereospecificity of amino acid uptake, exchange and efflux. J. Neurochem. 10 909 - 920 112 Sershen H. & Lajtha A. (1979) Inhibition pattern by analogs indicates the presence of ten or more transport systems for amino acids in brain cells. J. Neurochem. 32 719 - 726 113 Pardridge W.M. (1977) Kinetics of competitive inhibition of neutral amino acid transport across the blood-brain barrier. J. Neurochem. 28 103 - 108 114 Baumann P. & Wurtman R.J. (1979) Transport mechanisms of tryptophan. Proceedings of an International Symposium, Prilly/Lausanne, July 1978. J. Neural Trans. 15 (supplement) 115 McMenamy R.H. & Oncley J.L. (1958) The specific binding of L-tryptophan to serum albumin. J. Biol. Chem. 233 1436 - 1447 116 Stewart K.K. & Doherty R.F. (1973) Resolution of DL-tryptophan by affinity chromatography on bovine serum albumin-agarose. Proc. Nat. Acad. Sci. USA 70 2850 - 2852 117 Bender D.A., Boulton A.P. & Coulson W.F. (1975) A simple method for the study of tryptophan binding to serum albumin by small-scale equilibrium dialysis: application to animal and human studies. Biochem. Soc. Transact. 3 193 - 194 118 Smith H.G. & Lakatos C. (1971) Effects of acetylsalicylic acid on serum protein binding and metabolism of tryptophan in man. J. Pharm. Pharmacol. 23 180 - 189 119 Baumann P. & Perry M. (1977) The analysis of free tryptophan in human blood with the ultrafiltrator: a comparison with other methods. Clin. Chim. Acta 76 223 - 231 120 McMenamy R.H. (1965) Binding of indole analogues to human serum albumin. J. Biol. Chem. 240 4235 - 4243 121 Iwata H., Okamoto H. & Koh S. (1975) Effects of various drugs in serum free and total tryptophan levels and brain tryptophan metabolism in rats. Jap. J. Pharmacol. 25 303 - 310 122 Spano P.F., Szyszka K., Galli C.L. & Ricci A. (1974) Effect of clofibrate on free and total tryptophan in serum and brain tryptophan metabolism. Pharmacol. Res. Commun. 6 163 - 173 123 Bender D.A. & Cockcroft P.M. (1977) Increase in brain tryptophan and 5-hydroxytryptamine on administration of phenothiazines to rats. Biochem. Soc. Transact. 5 155 - 157 124 Knott P.J. & Curzon G. (1972) Free tryptophan in plasma and brain tryptophan metabolism. Nature 239 452 - 453 184 D. A. Bender

125 Curzon G. & Knott P.J. (1974) Effects on plasma and brain tryptophan in the rat of drugs and hormones that influence the concentration of unesterified fatty acids. Brit. J. Pharmacol. 50 197 - 204 126 Curzon G. (1974) Availability of tryptophan to the brain and some hormonal and drug influences on it. Adv. Biochem. Psychopharmacol. 10 263 - 271 127 Tagliamonte A., Biggio G., Vargiu L. & Gessa G.L. (1973) Increase in brain tryptophan and stimulation of serotonin synthesis by salicylate. J. Neurochem. 20 909 - 912 128 Yuwiler A., Oldendorf W.H., Geller E. & Braun L. (1977) Effect of albumin binding and amino acid competition on tryptophan uptake into brain. J. Neurochem. 28 1015 - 1023 129 Etienne P., Young S.N. & Sourkes T.L. (1976) Inhibition by albumin of tryptophan uptake by rat brain. Nature 262 144 - 145 130 Smith S.A. & Pogson C.I. (1980) The metabolism of tryptophan by isolated rat liver cells: effect of albumin binding and amino acid competition on oxidation of tryptophan by tryptophan 2,3-dioxygenase. Biochem. J. 186 977 - 986 131 Fernstrom J.D. & Wurtman R.J.(1971) Brain serotonin content: physiological dependence on plasma tryptophan levels. Science 173 149 - 152 132 Fernstrom J.D. & Wurtman R.J. (1975) Control of brain monoamine synthesis by diet and plasma amino acids. Amer. J. Clin. Nutr. 28 638 - 647 133 Curzon G., Friedl J. & Knott P.J. (1973) The effect of fatty acids on the binding of tryptophan to plasma protein. Nature 242 198 - 200 134 Bender D.A., Armstrong A.J.McN., Monkhouse C.R. & Richardson J.P. (1975) Changes in pancreatic tryptophan in the rat in response to fasting: the effects of B-cytotoxic agents and variation through the oestrous cycle. Pflugers Archiv. 356 245 - 251 135 Noble R.E. (1969) Effect of cyproheptadine on appetite and weight gain in adults. J. Amer. Med. Ass. 209 2054 - 2055 136 Breisch S.T., Zemlan F.P. & Hoebel B.G. (1976) Hyperphagia and obesity following serotonin depletion by intraventricular p-chlorophenylalanine. Science 192 382 - 385 137 Sailer C.F. & Stricker E.M. (1976) Hyperphagia and increased growth in rats after intraventricular injection of 5,7-dihydroxytryptamine. Science 192 385 - 387 138 Samanin R., Ghezzi D., Valzelli L. & Garattini S. (1972) The effects of selective lesioning of brain serotonin or catecholamine containing neurons on the anorectic effect of fenfluramine and amphetamine. Eur. J. Pharmacol. 19 318 - 322 139 Blundell J.E. & Leshem M.B. (1975) The effect of 5-hydroxytryptophan on food intake and on the anorexic action of amphetamine and fenfluramine. J. Pharm. Pharmacol. 27 31 - 37 140 Weinberger S.B., Knapp S. & Mandell A.J. (1978) Failure of tryptophan load induced increases in brain serotonin to alter food intake in the rat. Life Sci. 22 1595 - 1602 141 Barrett A.M. & McSharry L. (1975) Inhibition of drug-induced anorexia in rats by methysergide. J. Pharm. Pharmacol. 27 889 - 895 Biochemistry of Tryptophan 185

142 Knott P.J. & Curzon G. (1975) Tryptophan and tyrosine disposition and brain tryptophan metabolism in acute carbon tetrachloride poisoning. Biochem. Pharmacol. 24 963 - 966 143 Ashley D.V.M. & Curzon G. (1981) Effects of long-term low dietary tryptophan intake on determinants of 5-hydroxytryptamine metabolism in the brains of young rats. J. Neurochem. 37 1385 - 1393 144 Anderson G.H. & Ashley D.V.M. (1978) Plasma amino acids, brain mechanisms and the control of protein intake, pp 237 - 246 in Nutrition in Transition, Proceedings of the Western Hemisphere Nutrition Congress V, American Medical Association. 145 Anderson G.H. (1977) Regulation of protein intake by plasma amino acids. Adv. Nutr. Res. 1 145 - 165 146 Visek W.J. (1972) Effects of urea hydrolysis on cell life-span and metabolism. Fed. Proc. 31 1178 - 1193 147 Knell A.J., Davidson A.R., Williams R., Kantamaneni B.D. & Curzon G. (1974) Dopamine and serotonin metabolism in hepatic encephalopathy. Brit. Med. J. 1 549 - 551 148 Cangiano C., Calcaterra V., Cascino A & Capocaccio L. (1976) Bound and free tryptophan plasma levels in hepatic encephalopathy. Rendic. Gastroenterol. 8 186 - 189 149 Baldessarini R.J. & Fischer J.E. (1973) Serotonin metabolism in rat brain after surgical division of the portal venous circulation. Nature-New Biology 245 25 - 27 150 Curzon G., Kantamaneni B.D., Winch J., Rojas-Bueno A., Murray-Lyon I.M. & Williams R. (1973) Plasma and brain tryptophan changes in experimental acute hepatic failure. J. Neurochem. 21 137 - 145 151 Curzon G., Kantamaneni B.D., Fernando J.C., Woods M.S. & Cavanagh J.B. (1975) Effects of chronic portacaval anastomosis on brain tryptophan, tyrosine and 5-hydroxytryptamine. J. Neurochem. 24 1065 - 1070 152 Mans A.M., Biebuyck J,F., Shelly K. & Hawkins R.A. (1982) Regional blood-brain barrier permeability to amino acids after porta-caval anastomosis. J. Neurochem. 38 705 - 717 153 Munro H.N., Fernstrom J.D. & Wurtman R.J. (1975) Insulin, plasma amino acid imbalance and hepatic coma. Lancet 1 722 - 724 154 Daniel P.M., Love E.R., Moorhouse S.R. & Pratt O.E. (1975) Amino acids, insulin and hepatic coma. Lancet 2 179 - 180 155 Herlin P.M., James J.H., Joffe S.N., Kulneff-Herlin A.E.A. & Fischer J.E. (1982) Effect of jejuno-ileal bypass on plasma and brain amino acids in the rat. J. Neurochem. 38 1170 - 1173 156 James J.H., Ziparo V., Jeppson B. & Fischer J.E. (1979) Hyperammonaemia, plasma amino acid imbalance and blood-brain amino acid transport: a unified theory of portal system encephalopathy. Lancet 2 772 - 775 157 Jouvet M. (1969) Biogenic amines and the states of sleep. Science 169 32 - 41 158 Pujol J-F., Buguet A., Froment J-L., Jones B. & Jouvet M. (1971) The central metabolism of 5-hydroxytryptamine in the cat during insomnia: neurophysiological and biochemical studies after administration of p-chlorophenylalanine or destruction of the raphe system. Brain Res. 29 195 - 212 186 D.A. Bender

159 Hery F., Pujol J-F., Lopez M., Macon J & Glowinski J. (1970) Increased synthesis and utilisation of 5-hydroxytryptamine in the central nervous system of the rat during paradoxical sleep deprivation. Brain Res. 21 391 - 403 160 Wyatt R.W., Engelmann K., Kupfer D.J., Scott J., Sjoedsma A. & Snyder F. (1969) Effects of p-chlorophenylalanine on sleep in man. Electroencephal. Clin. Neurophysiol. 27 529 - 532 161 Chen C.N., Kalney R.S., Hartman M.K., Lacey J.H., Crisp A.H., Bailey J.E., Eccleston E.G. & Coppen A. (1974) Plasma tryptophan and sleep. Brit. Med. J. 4 564 - 566 162 Williams H.L., Lester B.K. & Coulter J.D. (1969) Monoamines and the EEG states of sleep. Activitas Nervosa Superior 11 188 - 192 163 Wyatt R.J., Engelman K., Kupfer D.J., Fram D.H., Sjoedsma A. & Snyder F. (1970) Effects of L-tryptophan ( a natural sedative) on human sleep. Lancet 2 842 - 846 164 Smith B. & Prockop D.J. (1962) Central nervous system effects of ingestion of L-tryptophan by normal subjects. New Engl. J. Med. 267 1338 - 1341 165 Hartmann E. (1977) L-Tryptophan: a rational hypnotic with clinical potential. Amer. J. Psychiatr. 134 366 - 370 166 Muller J.C., Pryer W.W., Gibbons J.E. & Orgain E.S. (1955) Depression and anxiety occurring during Rauwolfia therapy. J. Amer. Med. Ass. 159 836 - 839 167 Harris T.H. (1957) Depression induced by Rauwolfia compounds. Amer. J. Psychiatr. 113 950 - 962 168 Iversen L.L. (1973) Monoamines in the mammalian central nervous system and the actions of anti-depressant drugs, pp 81 - 96 in Biochemistry and Mental Illness, Iversen L.L. & Rose S.P.R., Eds., Biochemical Society, London. 169 Shore P.A. (1962) The release of 5-hydroxytryptamine and catecholamines by drugs. Pharmacol. Rev. 14 531 - 550 170 Shore R.A., Mead J.A.B., Kuntzman R.G., Spector S. & Brodie B.B. (1957) On the physiologic significance of monoamine oxidase in brain. Science 126 1063 - 1065 171 Ross S.B. & Renyi A.L. (1967) Accumulation of tritiated 5-hydroxytryptamine in brain slices. Life Sci. 6 1407 - 1415 172 Schildkraut J.J. (1965) The catecholamine hypothesis of affective disorders: a review of supporting evidence. Amer. J. Psychiatr. 122 509 - 522 173 Glassman A. (1969) Indoleamines and affective disorders. Psychosom. Med. 31 107 - 115 174 Ashcroft G.W., Eccleston D., Murray L.G., Glen A.I.M., Crawford T.B.B., Pullar I.A. Shields P.J., Walter D.S., Blackburn I.M., Connechan J. & Lonergan M. (1972) Modified amine hypothesis for the aetiology of depressive illness. Lancet 2 573 - 577 175 Ashcroft G.W., Crawford T.B.B., Eccleston D., Sharman D.F., MacDougal E.J., Stanton J.B. & Binar J.K. (1966) 5-Hydroxyindole compounds in csf of patients with psychiatric or neurological disease. Lancet 2 1049 - 1052 176 Bourne H.R., Bunney W.E., Colburn R.W., Davis J.M., Shaw D.M. & Coppen A.J. (1968) Noradrenaline, 5-hydroxytryptamine and 5-hydroxyindoleacetic acid in hindbrains of suicidal patients. Lancet 2 805 - 808 Biochemistry of Tryptophan 187

177 Asberg M., Thoren P. & Traskman L. (1976) 'Serotonin depression': a biochemical subgroup within the affective disorders. Science 191 478 - 480 178 Asberg M., Traskman L. & Thoren P. (1976) 5HIAA in the cerebrospinal fluid: a biochemical suicide predictor? Archs. Gen. Psychiatr. 33 1193 - 1197 179 Traskman L., Asberg M., Bertilsson L. & Sjostrand L. (1981) Monoamine metabolites in csf and suicidal behavior. Archs. Gen. Psychiatr. 38 631 - 636 180 Unge G., Lins L-E. & Haltman E. (1977) Tryptophan in patients on chronic haemodialysis. Lancet 2 937 181 Rees J.R., Allsopp M.N.E. & Hullin R.P. (1974) Plasma concentrations of tryptophan and other amino acids in manic-depressive patients. Psychol. Med. 4 334 - 337 182 Coppen A., Eccleston E.G. & Peet M. (1973) Total and free tryptophan concentration in the plasma of depressive patients. Lancet 2 60 - 63 183 Handley S.L., Dunn T.L., Baker J.M., Cockshott C. & Gould S. (1977) Mood changes in puerperium and plasma tryptophan and cortisol concentrations. Brit. Med. J. 3 18 - 20 184 Aylward M. (1976) Estrogens, plasma tryptophan levels in peri- menopausal patients, pp 135 - 147 in The Management of the Menopause and Post-Menopause Years. Campbell S., Ed., MTP Press, Lancaster 185 Lapin I.P. & Oxenkrug G.F. (1969) Intensification of central serotoninergic processes as a possible determinant of the thymoleptic effect. Lancet 1 132 - 136 186 Samsonova M.L. & Lapin I.P. (1973) Anti-depressants and liver tryptophan pyrrolase activity. Biochem. Pharmacol. 22 1499 - 1507 187 Badawy A.A-B. & Evans M. (1981) Inhibition of rat liver tryptophan pyrrolase activity and elevation of brain tryptophan concentration by administration of anti-depressants. Biochem. Pharmacol. 30 1211 - 1216 188 Badawy A.A-B. & Evans M. (1974) Tryptophan plus a pyrrolase inhibitor for depression. Lancet 2 1209 - 1210 189 Joseph M.H., Young S.N. & Curzon G. (1976) The metabolism of a tryptophan load in rat brain and liver. The influence of hydrocortisone and allopurinol. Biochem. Pharmacol. 25 2599 - 2604 190 Kelly W.F., Checkley S.A. & Bender D.A. (1980) Cushing's syndrome, tryptophan and depression. Brit. J. Psychiatr. 136 125 - 132 191 Kelly W.F., Checkley S.A., Bender D.A. & Mashiter K. (1982) Cushing's syndrome and depression: a prospective study of 26 patients. Brit. J. Psychiatr, in press 192 Peet M., Moody J.P., Worrall E.P., Walker P. & Naylor G.J. (1976) Plasma tryptophan concentration in depressive illness and mania. Brit. J. Psychiatr. 128 255 - 258 193 Riley G.J. & Shaw D.M. (1981) Plasma tryptophan binding to albumin in unipolar depressives. Acta Psychiatr. Scand. 63 165 - 167 194 Pare C.M.B. (1963) Potentiation of monoamine oxidase inhibitors by tryptophan. Lancet 2 527 - 528 188 D. A. Bender

195 Coppen A., Shaw D.M. & Farrell J.P. (1963) Potentiation of anti-depressive effect of monoamine oxidase inhibitors by tryptophan. Lancet 1 79 - 81 196 Alino J.J.L-I., Gutierez J.L.A. & Iglesias M.L.M. (1973) Tryptophan and amitriptyline in the treatment of depression: a double-blind study. Int. Pharmacopsychiatr. 8 145 - 151 197 Shaw D.M., MacSweeney D.A., Hewland R. & Johnson A.L. (1975) Tricyclic anti-depressants and tryptophan in depression. Psychol. Med. 5 276 - 278 198 Coppen A., Shaw D.M., Herzeberg B. & Maggs R. (1967) Tryptophan in the treatment of depression. Lancet 2 1178 199 Hertz D. & Salman F.G. (1968) Preventing depression with tryptophan. Lancet 1 531 - 532 200 Herrington R.N., Bruce A., Johnstone E.C. & Lader M.H. (1974) Comparative trial of L-tryptophan and ECT in severe depressive illness. Lancet 2 731 - 736 201 Herrington, R.N., Bruce A., Johnstone E.C. & Lader M.H. (1976) Comparative trial of L-tryptophan and amitriptyline in depressive illness. Psychol. Med 6 673 - 678 202 Murphy D.L., Baker M., Goodwin F.K., Miller H., Kotin J. & Bunney W.J. (1974) L-Tryptophan in affective disorders: indoleamine changes and differential clinical effects. Psychopharmacologia 34

11 - 20 203 Cooper A.J. & Datta S.R. (1980) A placebo-controlled trial of L-tryptophan in depression in the elderly. Can. J. Psychiatr. 25 386 - 390 204 D'Elia G., Hanson L. & Raotma H. (1978) L-Tryptophan and 5-hydroxytryptamine in the treatment of depression: a review. Acta Psychiatr. Scand. 57 239 - 253 205 Carroll B.J. (9171) Monoamine precursors in the treatment of depression. Clin. Pharm. Ther. 12 743 - 761 206 Gilka L. (1975) Schizophrenia - a disorder of tryptophan metabolism. Acta Psychiatr. Scand. Suppl. 258 10 - 83 207 Bennett J.P. & Snyder S.H. (1976) Serotonin and lysergic acid diethylamide binding in rat brain membranes: relationship to post-synaptic receptors. Molec. Pharmacol. 12 373 - 381 208 Pollin W., Cardon P.V. & Kety S.S. (1961) Effects of amino acid feedings on schizophrenic patients treated with iproniazid. Science 133 104 - 105 209 Alexander F., Curtis C., Sprince H. & Crosby A.P. (1963) L-Methionine and L-tryptophan feedings in non-psychotic and schizophrenic patients with and without tranylcypromime. J. Nerv. Ment. Dis. 137 135 - 142 210 Frohman C.E., Warner K.A., Yoon H.S., Arthur R.E. & Gottlieb J.S. (1969) The plasma factor and transport of indoleamino acids. Biol. Psychiatr. 1 377 - 385 211 Frohman C.E., Harmison C.E., Arthur R.E. & Gottlieb J.S. (1971) Conformation of a unique plasma protein in schizophrenia. Biol. Psychiatr. 3 113 - 121 212 Bender D.A. & Bamji A.N. (1974) Serum tryptophan binding in chlorpromazine-treated chronic schizophrenics. J. Neurochem. 22 805 - 809 Biochemistry of Tryptophan 189

213 Manowitz P., Gilmour D.G. & Racevskis J. (1973) Low plasma tryptophan in newly hospitalized schizophrenics. Biol. Psychiatr. 6 109 - 118 214 Gilmour D.G., Manowitz P., Frosch W.A. & Shopsin B. (1973) Association of plasma tryptophan levels with clinical changes in female schizophrenic patients. Biol. Psychiatr. 6 119 - 128 215 Lovett-Doust J.W., Huszka L. & Lovett-Doust J.N. (1975) Psychotropic drugs and gender as modifiers of the role of plasma tryptophan and serotonin in schizophrenia. Compreh. Psychiatr. 16 349 - 355 216 Deniker P., Loo H., Zarifian E., Bousquet B., Dreux C. & Escande C. (1976) Etude du tryptophane libre et total dans le plasma: son interet en psychiatrie. L'Encephale 2 123 - 131 217 Freedman D.X., Belendiuk K., Belendiuk G.W. & Crayton J.W. (1981) Blood tryptophan metabolism in chronic schizophrenics. Archs. Gen. Psychiatr. 38 655 - 659 218 Domino E.F. & Krause R.R. (1974) Free and bound serum tryptophan in drug-free normal controls and chronic schizophrenic patients. Biol. Psychiatr. 8 265 - 279 219 Bender D.A., Pigache R.M., Gruzelier J. & Hammond N. (1975) Changes in serum tryptophan and albumin binding of tryptophan in chlorpromazine-treated chronic schizophrenics on withdrawal and restoration of drug therapy. Psychol. Med. 5 397 - 403 220 Gallant D.M., Bishop M.P., Steele C.A. & Noblin C.D. (1964) Anti- serotonin activity and clinical tranquilization. J. Neuropsychiatr. 5 273 - 278 221 Holden J.M.C., Hill T., Keskiner A. & Gannon P. (1971) A clinical trial of an anti-serotonin compound, cinanserin, in chronic schizophrenia. J. Clin. Pharmacol. 11 220 - 226 222 von Hungen K., Roberts S. & Hill D.F. (1975) Serotonin-sensitive adenylate cyclase activity in immature rat brain. Brain Res. 84 257 - 267 223 Bender D.A. & Cockcroft P.M. (1977) Increase in brain tryptophan and 5-hydroxytryptamine on administration of phenothiazines to rats. Biochem. Soc. Transact. 5 155 - 158 224 Yorkston N.J., Zaki S.A., Malik M.K.U., Morrison R.C. & Havard C.W.H. (1974) Propranolol in the control of schizophrenic symptoms. Brit. Med. J. 4 633 - 635 225 Green A.R. & Grahame-Smith D.G. (1974) TRH potentiates behavioural changes following increased brain 5HT accumulation in rats. Nature 251 524 - 526 226 Gallager D.W. & Aghajanian G.K. (1976) Effects of anti-psychotic drugs on the firing of dorsal raphe cells: (i) role of adrenergic systems. Eur. J. Pharmacol. 39 341 - 355 227 Gallager D.W. & Aghajanian G.K. (1976) Effects of anti-psychotic drugs on the firing of dorsal raphe cells: (ii) reversal by picrotoxin. Eur. J. Pharmacol. 39 357 - 364 228 Bird E.D., Spokes E.G., Barnes J°, MacKay A.V., Iversen L.L, & Shepherd M. (1977) Increased brain dopamine and reduced glutamic acid decarboxylase and choline acetyltransferase activity in schizophrenia and related psychoses. Lancet 2 1157 - 1160 190 D.A. Bender

229 Mehler A.H., McDaniel E.G. & Hundley J.M. (1958) Changes in the enzymatic composition of liver: (ii) influence of hormones on picolinic carboxylase and tryptophan peroxidase. J. Biol. Chem. 232 331 - 335 230 Schimke R.T., Sweeney E.W. & Berlin C.M. (1965) The roles of synthesis and degradation in the control of rat liver tryptophan pyrrolase. J. Biol. Chem. 240 322 - 331 231 Schutz G., Beato M. & Feigelson P. (1973) Messenger-RNA for hepatic tryptophan oxygenase; its partial purification, its translation in a heterologous cell-free system and its control by glucocorticoid hormones. Proc. Nat. Acad. Sci. USA 70 1218 - 1221 232 Nakamura T., Shinno H & Ichihara A. (1980) Insulin and glucagon as a new regulator system for tryptophan oxygenase activity demonstrated in primary cultured rat hepatocytes. J. Biol. Chem. 255 7533 - 7535 233 Braidman I.P. & Rose D.P. (1971) Effects of sex hormones on three glucocorticoid-inducible enzymes concerned with amino acid metabolism in rat liver. Endocrinol. 89 1250 - 1255 234 Patnaik S.K. & Sarangai S. (1980) Age-related response of tryptophan pyrrolase to 17-beta-oestradiol in the liver of female rats. J. Biochem. (Tokyo) 87 1249 - 1252 235 Kanke Y., Suzuki K., Hirawaka S. & Goto S. (1980) Oral contraceptive steroids: effects on iron and zinc levels and on tryptophan pyrrolase and alkaline phosphatase activities in tissues of iron- deficient anaemic rats. Amer. J. Clin. Nutr. 33 1244 - 1250 236 Rifkind A.B., Gillette P.N. & Song C.S. (1970) Induction of hepatic delta-aminolaevulinic acid synthetase by oral contraceptive steroids. J. Clin. Endocrinol. 30 330 - 335 237 Leonard B.E. & Hamburger A.D. (1974) Sex hormones, tryptophan oxygenase activity and cerebral monoamine metabolism in the rat. Biochem. Soc. Transact. 2 1351 - 1355 238 Green A.R., Bloomfield M.R., Woods H.F. & Seed M. (1978) Metabolism of an oral tryptophan load by women and evidence against the induction of tryptophan pyrrolase by oral contraceptives. Brit. J. Clin. Pharmacol. 5 233 - 241 239 Bender D.A., Tagoe C.E. & Vale J.A. (1982) Effects of oestrogen administration on vitamin B 6 and tryptophan metabolism in the rat. Brit. J. Nutr. in press 240 Greengard O. & Feigelson P. (1961) The activation and induction of rat liver tryptophan pyrrolase in vivo by its substrate. J. Biol. Chem. 236 158 - 161 241 Schimke R.T., Sweeney E.W. & Berlin C.M. (1965) Studies on the stability in vivo and in vitro of rat liver tryptophan pyrrolase. J. Biol. Chem. 240 4609 - 4620 242 Feigelson P. & Maeno H. (1967) Studies on the enzyme-substrate interactions in the regulation of tryptophan oxygenase activity. Biochem. Biophys. Res. Commun. 28 289 - 293 243 Wagner C. (1964) Regulation of the tryptophan - nicotinic acid - DPN pathway in the rat. Biochem. Biophys. Res. Commun. 17 668 - 673 244 Cho-Chung Y.S. & Pitot H.C. (1967) Feedback control of rat liver tryptophan pyrrolase. J. Biol. Chem. 242 1192 - 1198

192 D.A. Bender

260 Jacobson K.B. (1978) A new substrate for formylkynurenine formamidase, NI,N -diformylkynurenine. Archs. Biochem. Biophys. 186 84 - 88 261 Ueda T., Otsuka H., Goda K., Ishigori I., Naitro J. & Kotake Y. (1978) The metabolism of [carboxyl-14C]anthranilic acid: (i) the incorporation of radioactivity into NAD + and NADP +. J. Biochem. (Tokyo) 84 687 - 696 262 Bender D.A. & McCreanor G.M. (1982) The preferred route of kynurenine metabolism in the rat. Biochim. Biophys. Acta in press 263 Tanizawa K. & Soda K. (1979) Purification and properties of pig liver kynureninase. J. Biochem. (Tokyo) 85 901 - 906 264 Ueno Y., Hayashi K. & Shukuya R. (1963) Kynurenine transaminase from horse kidney. J. Biochem. (Tokyo) 54 75 - 80 265 Lepkovsky S. & Nielsen E. (1942) A green pigment-producing compound in urine of pyridoxine deficient rats. J. Biol. Chem. 144 135 - 138 266 Knox W.E. (1953) The relation of liver kynureninase to tryptophan metabolism in pyridoxine deficiency. Biochem. J. 53 379 - 385 267 Price J.M., Brown R.R. & Yess N. (1965) Testing the functional capacity of the tryptophan-niacin pathway in man by analysis of urinary metabolites. Adv. Metab. Disorders 2 159 - 225 268 Musajo L. & Benassi C.A. (1964) Aspects of disorders of the kynureninase pathway of tryptophan metabolism in man. Adv. Clin. Chem. 7 63 - 133 269 Watanabe M., Takahashi T., Yoshida M., Suzuki M. & Muramatsu S. (1979) Relationship between tryptophan intake and urinary excretion of 3-hydroxykynurenine, 3-hydroxyanthranilic acid, xanthurenic acid and kynurenic acid. J. Nutr. Sci. Vitaminol. 25 115 - 122 270 Brown R.R., Yess N., Price J.M., Linkswiler H., Swan P. & Hankes L.V. (1965) Vitamin B 6 depletion in man: urinary excretion of quinolinic acid and niacin metabolites. J. Nutr 87 419 - 423 271 Coursin D.B. (1964) Recommendations for standardization of the tryptophan load test. Amer. J. Clin. Nutr. 14 56 - 61 272 Allegri G., Costa C. & de Antoni A. (1978) A further contribution to the choice of the dose for tryptophan load test. Acta Vitaminol. Enzymol. (Milan) 32 163 - 166 273 Wolf H., Brown R.R. & Arendt R.A. (1980) The kynurenine load test, an adjunct to the tryptophan load test. Scand. J. Clin. Lab. Invest. 40 9 - 14 274 Ogasawara N., Hagino Y. & Kotake Y. (1962) Kynurenine transaminase and the increase of xanthurenic acid excretion. J. Biochem. (Tokyo) 52 162 - 166 275 Rose D.P., Strong R., Adams P.W. & Harding P.E. (1972) Experimental vitamin B 6 deficiency and the effect of oestrogen-containing oral contraceptives on tryptophan metabolism and vitamin B 6 requirements. Clin. Sci. 42 465 - 477 276 Rose D.P. & Adams P.W. (1972) Oral contraceptives and tryptophan metabolism: effects of oestrogen in low dose combined with a progestagen and of a low dose progestagen (megestrol acetate) given alone. J. Clin. Pathol. 25 252 - 258 277 Luhby A.L., Brin M., Gordon M., David P., Murphy M. & Spiegel H. (1971) Vitamin B 6 metabolism in users of oral contraceptive agents: (i) abnormal urinary excretion of xanthurenic acid and its correction by pyridoxine. Amer. J. Clin. Nutr. 24 684 - 693 Biochemistry of Tryptophan 193

278 Haspels A.A., Coelingh-eennink H.J.T. & Schreurs W.H.P. (1978) Disturbance of tryptophan metabolism and its correction during oestrogen treatment in post-menopausal women. Maturitas 1 15 - 20 279 Rose D.P. & Braidman I.P. (1971) Excretion of tryptophan metabolites as affected by pregnancy, contraceptive steroids and steroid hormones. Amer. J. Clin. Nutr. 24 673 - 683 280 Rose D.P. (1978) The interactions between vitamin B 6 and hormones. Vitamins & Hormones 36 53 - 99 281 Brown R.R., Rose D.P., Leklem J.E., Linkswiler H. & Arendt R.A. (1975) Urinary 4-pyridoxic acid, plasma pyridoxal phosphate and erythrocyte transaminase levels in oral contraceptive users receiving controlled intakes of vitamin B 6. Amer. J. Clin. Nutr. 28 10 - 19 282 Leklem J.E., Brown R.R., Rose D.P., Linkswiler H. & Arendt R.A. (1975) Metabolism of tryptophan and niacin in oral contraceptive users receiving controlled intakes of vitamin B 6. Amer. J. Clin. Nutr. 28 146 - 156 283 Wien E.M. (1978) Vitamin B 6 status of Nigerian women using various methods of oral contraception. Amer. J. Clin. Nutr. 31 1392 - 1396 284 Aly H.E., Donald E.A. & Simpson M.H.W. (1971) Oral contraceptives and vitamin B 6 metabolism. Amer. J. Clin. Nutr. 24 297 - 301 285 Salkeld R.L., Knorr K. & Korrer W.F. (1973) The effect of oral contraceptives on vitamin B 6 status. Clin. Chim. Acta 49 195 - 199 286 Wolf H., Walter S., Brown R.R. & Arendt R.A. (1980) Effect of natural oestrogens on tryptophan metabolism: evidence for interference of oestrogens with kynureninase. Scand. J. Clin. Lab. Invest. 40 15 - 22 287 Mason M. & Gullekson E.H. (1960) Estrogen-enzyme interactions: protection of kynurenine transaminase by the sulfate esters of diethylstilbestrol, estradiol and estrone. J. Biol. Chem. 235 1312 - 1316 288 Bender D.A., Coulson W.F., Papadaki L. & Pugh M. (1981) Effects of oestrone on apparent vitamin B 6 status in peri- and post- menopausal women. Proc. Nutr. Soc. 40 20a 289 Bender D.A. & Wynick D. (1981) Inhibition of kynureninase by oestrone sulphate: an alternative explanation for abnormal results of tryptophan load tests in women receiving oestrogenic steroids. Brit. J. Nutr. 45 269 - 275 290 Coon W.W. & Nagler E. (1969) The tryptophan load test as a test for pyridoxine deficiency in hospitalized patients. Ann. N.Y. Acad. Sci. 166 30 - 43 291 Roe D.A. (1973) A plague of corn: the social history of pellagra. Cornell University Press, Ithaca. 292 Still C.N. (1976) Nicotinic acid and nicotinamide deficiency: pellagra and related disorders of the nervous system, pp 59 - 104 in Handbook of Clinical Neurology 28 (part 2), Vinken P.J. & Bruyn G.W., Eds., Elsevier-North Holland, Amsterdam. 293 Horwitt M.K., Harvey C.C., Rothwell W.S., Cutler J.L. & Haffron D. (1956) Tryptophan : niacin relationships in man. J. Nutr 60 (suppl. 1) 1 - 43 194 D.A. Bender

294 Wertz A.W., Lojkin M.I., Bouchard B.S. & Derby M.B. (1958) Tryptophan - niacin relationships in pregnancy. J. Nutr. 64 339 - 353 295 Nakagawa I., Takahashi T., Suzuki T. & Masana Y. (1969) Effect in man of the addition of tryptophan or niacin to the diet on the excretion of their metabolites. J. Nutr. 99 325 - 330 296 Horwitt M.K., Harper A.E. & Henderson L.M. (1981) Niacin - tryptophan relationships for evaluating niacin equivalents. Amer. J. Clin. Nutr. 34 423 - 427 297 Ikeda M., Tsuji H., Nakamura S., Ichiyama A., Nishizuka Y. & Hayaishi O. (1965) Studies on the biosynthesis of nicotinamide adenine dinucleotide: (ii) a role of picolinic carboxylase in the biosynthesis of nicotinamide adenine dinucleotide from tryptophan in mammals. J. Biol. Chem. 240 1395 - 1401 298 Greengard P., Kalinsky H., Manning T.J. & Zak S.B. (1968) Prevention and remission by adrenocortical steroids of nicotinamide deficiency disease: (ii) a study of the mechanism. J. Biol. Chem. 243 4216 - 4221 299 Kim J.H. & Miller L.L. (1969) The functional significance of changes in the activity of the enzymes tryptophan pyrrolase and tyrosine aminotransferase after induction in intact rats and in the isolated perfused rat liver. J. Biol. Chem. 244 1410 - 1416 300 Shibata K., Motokawa K., Murata K. & Iwai K. (1980) Metabolic control for maintaining NAD in rats fed on tryptophan-limiting amino acid diets. J. Nutr. Sci. Vitaminol. 26 571 - 584 301 Sanada H., Miyazaki M. & Takahashi T. (1980) Regulation of tryptophan-niacin metabolism in diabetic rats. J. Nutr. Sci. Vitaminol. 26 449 - 459 302 Sanada H. & Miyazaki M. (1980) Regulation of tryptophan-niacin metabolism by hormones. J. Nutr. Sci. Vitaminol. 26 617 - 627 303 Green S. & Dobrzansky A. (1967) Relationships of the nicotinamide adenine dinucleotide glycohydrolase activity to nicotinamide adenine dinucleotide content and rate of proliferation of Ehrlich ascites tumor cells. Cancer Res. 27 2261 - 2266 304 Schoental R. (1980) Mouldy grain and the aetiology of pellagra: the role of toxic metabolites of Fusarium. Biochem. Soc. Transact. 8 147 - 150 305 Yamamoto H. & Okamoto H. (1980) Protection by picolinamide, a novel inhibitor of poly-(ADP-ribose) synthetase, against both streptozotocin-induced depression of pro-insulin synthesis and reduction of NAD in pancreatic islets. Biochem. Biophys. Res. Commun. 95 474 - 481 306 National Institute of Nutrition (Hyderabad, India) Annual Report 1968/9 pp 56 - 58 307 Krieger I.E. (1981) Acrodermatitis enteropathica and the relation to pellagra. Med. Hypoth. 7 539 - 547 308 Hsu J.M. & Rubenstein B.J. (1981) Role of zinc in histidine metabolism. Fed. Proc. 40 839 309 Raghuram T.C. & Krishnaswamy K. (1975) Serotonin metabolism in pellagra. Archs. Neurol. 32 708 - 710 310 Krishnaswamy K. & Ramanamurthy P.S.V. (1970) Mental changes and platelet serotonin in pellagrins. Clin. Chim. Acta 27 301 - 304 Biochemistry of Tryptophan 195

311 Tada K., Ito H., Wada Y. & Arakawa T. (1963) Congenital tryptophanuria with dwarfism. Tokohu J. Exp. Med. 80 118 - 134 312 Wang P.W.K., Forman P., Tabahoff B. & Justice P. (1976) A defect in tryptophan metabolism. Ped. Res. 10 725 - 730 313 Baron D.N., Dent C.E., Harris H., Hart E.W. & Jepson J.B. (1956) Hereditary pellagra-like skin rash with temporary cerebellar ataxia, constant renal aminoaciduria and other bizarre biochemical features. Lancet 2 421 - 428 314 Castiello R.J. & Lynch P.J. (1972) Pellagra and the carcinoid syndrome. Archs. Dermatol. 105 574 - 577 315 Knapp A., Gassmann B. & Zimmermann W. (1958) Uber vitamin B6-mangel und die Harn-ausscheidung von Xanthurensaure und anderen Tryptophan-metaboliten bei Kranken. Klin. Wochschr. 36 819 - 823 316 Standal B.R., Kao-Chen S.M., Yang G-Y. & Char D.F.B. (1974) Early changes in pyridoxine status of patients receiving isoniazid therapy. Amer. J. Clin. Nutr. 27 479 - 484 317 Weiner W.J. & Klawans H.L. (1976) Vitamin B6, pp 105 - 139 in Handbook of Clinical Neurology 28 (part 2) Vinken P.J. & Bruyn G.W., Eds., Elsevier-North Holland, Amsterdam 318 Bender D.A. & Russell-Jones R. (1979) Isoniazid-induced pellagra despite vitamin B 6 supplementation. Lancet 2 1125 - 1126 319 Bender D.A. (1980) Inhibition in vitro of the enzymes of the oxidative pathway of tryptophan metabolism and of nicotinamide nucleotide synthesis by Benserazide, Carbidopa and isoniazid. Biochem. Pharmacol. 29 707 - 712 320 Bender D.A. (1980) Effects of Benserazide, Carbidopa and isoniazid administration on tryptophan - nicotinamide nucleotide metabolism in the rat. Biochem. Pharmacol. 29 2099 - 2104 321 Harrison R.J. & Feiwel M. (1956) Pellagra caused by isoniazid. Brit. Med. J. 2 852 - 854 322 Meyrick-Thomas R.H., Rowland-Payne C.M.E. & Black M.M. (1981) Isoniazid-induced pellagra. Brit. Med. J. 283 287 - 288 323 Madras B.K. & Sourkes T.L. (1968) Effects of drugs on the metabolism of tryptophan: alpha-hydrazino-tryptophan and other amino acid analogues. Biochem. Pharmacol. 17 1037 - 1047 324 Young S.N., St. Arnaud-McKenzie D. & Sourkes T.L. (9178) Importance of tryptophan pyrrolase and aromatic amino acid decarboxylase in the catabolism of tryptophan. Biochem. Pharmacol. 27 763 - 767 325 Smith S.A. & Pogson C.I. (1981) Effects of Benserazide and Carbidopa on the metabolism of tryptophan by isolated rat liver cells. Biochem. Pharmacol. 30 623 - 628 326 Bender D.A., Earl C.J. & Lees A.J. (1979) Niacin depletion in Parkinsonian patients treated with L-dopa, Benserazide and Carbidopa. Clin. Sci. 56 89 - 93 327 Gopalan C. & Srikantia S.G. (1960) Leucine and pellagra. Lancet 1 954 - 957 328 Magboul B.I. & Bender, D.A. (1982) The nature of niacin in sorghum. Proc. Nutr. Soc. in press 329 Gopalan C., Belavady B. & Krishnamurthy D. (1969) The role of leucine in the pathogenesis of canine black tongue and pellagra. Lancet 2 956 - 957 330 Belavady B. & Rao P.U. (1979) Leucine and isoleucine content of jowar and its pellagragenicity. Ind. J. Exp. Biol. 17 659 - 661 196 D.A. Bender

331 Belavady B. & Udayasekhara-Rao P. (1973) Production of nicotinic acid deficiency in monkeys fed on leucine-supplemented diets. Int. J. Vit. Nutr. Res. 43 454 - 460 332 Bapu-Rao S., Raghuram T.C. & Krishnaswamy K. (1975) Role of vitamin B 6 on leucine-induced metabolic changes. Nutr. Metab. 18 318 - 325 333 Belavady B., Udayasekhara-Rao P. & Khan L. (1973) Effects of leucine and isoleucine on nicotinamide nucleotides of erythrocytes. Int. J. Vit. Nutr. Res. 43 442 - 453 334 Srikantia S.G., Reddy M.V. & Krishnaswamy K. (1968) Electroencepha- lographic patterns in pellagra. Electroencephal. Clin. Neurophysiol. 25 386 - 388 335 Gopalan C. & Rao K.S.J. (1975) Pellagra and amino acid imbalance. Vitamins & Hormones 33 505 - 528 336 Ghafoorunissa & Narasinga-Rao B.S. (1973) Effect of leucine on the tryptophan-niacin metabolic pathway in rat liver and kidney. Biochem. J. 134 425 - 430 337 Krishnaswamy K. & Bapu-Rao S. (1978) Effect of leucine at different levels of pyridoxine on hepatic quinolinate phosphoribosyltransferase and leucine aminotransferase in rats. Brit. J. Nutr. 39 61 - 64 338 Nakagawa I., Ohguri S., Sasaki A., Kujimoto M, Sasaki M. & Takahashi T. (1975) Effects of excess intake of leucine and valine deficiency on tryptophan and niacin metabolites in humans. J. Nutr. 105 1241 - 1252 339 Manson J.A. & Carpenter K.J. (1978) The effect of a high level of dietary leucine on the niacin status of chicks and rats. J. Nutr. 108 1883 - 1888 340 Manson J.A. & Carpenter K.J. (1978) The effect of a high level in dietary leucine on the niacin status of dogs. J. Nutr. 108 1889 - 1898 341 Yamada 0., Shin M., Sano K. & Umezawa C. (1979) Effects of dietary excess leucine on nicotinamide nucleotide level in rat liver. Int. J. Vitr. Nutr. Res. 49 376 - 385 342 Magboul B.I. & Bender D.A. (1981) The effect of dietary leucine excess on nicotinamide nucleotides in the rat. Proc. Nutr. Soc. 40 16a 343 Magboul B.I. (1982) PhD Thesis, University of London 344 Ermolieff S. & Grosshaus E (1979) Le pellagre: maladie en resurgence au Zaire. Ann. Dermatol. Venerol. 106 591 - 595 345 Hahn C.J. & Evans G.W. (1973) Identification of a low molecular weight 65Zn complex in rat intestine. Proc. Soc. Exp. Biol. Med. 144 793 - 795 346 Moynahan E.J. & Barnes P.M. (1973) Zinc deficiency and a synthetic diet for lactose intolerance. Lancet 1 676 - 677 347 Evans G.W. & Johnson P.E. (1976) Zinc-binding factor in acrodermatitis enteropathica. Lancet 2 1310 348 Eckhert C.D., Sloan M.V., Duncan J.R. & Hurley L.S. (1977) Zinc binding: a difference between human and bovine milk. Science 195 789 - 790 349 Hurley L.S., Duncan J.R., Sloan M.V. & Eckhert C.D. (1977) Zinc binding ligands in milk and intestine: a role in neonatal nutrition. Proc. Nat. Acad. Sci. USA 74 3547 - 3549 Biochemistry of Tryptophan 197

350 Evans G.W. & Johnson P.E. (1979) Purification and characterization of a zinc binding ligand in human milk. Fed. Proc. 38 703 351 Evans G.W. & Johnson P.E. (1980) Characterization and quantitation of a zinc-binding ligand in human milk. Ped. Res. 14 876 - 880 352 Evans G.W. & Johnson P.E. (1980) Growth stimulating effect of picolinic acid added to rat diets. Proc. Soc. Exp. Biol. Med. 165 457 - 461 353 Evans G.W. & Johnson P.E. (1980) Zinc absorption in rats fed a low protein diet and a low protein diet supplemented with tryptophan or picolinic acid. J. Nutr. 110 1076 - 1080 354 Krieger I. & Evans G.W. (1980) Acrodermatitis enteropathica without hypozincemia: therapeutic effect of a pancreatic enzyme preparation due to zinc-binding ligand. J. Pediatr. 96 32 - 35