I. INTRODUCTION

Top Previous Next References

Ever since the discovery of phosphorylcreatine (PCr) in 1927 and of the creatine kinase (CK; EC 2.7.3.2) reaction in 1934 (see Refs. 140, 833), research efforts focused mainly on biochemical, physiological, and pathological aspects of the CK reaction itself and on its involvement in "high-energy phosphate" metabolism of cells and tissues with high-energy demands. In contrast, Cr (from greek kreas, flesh) metabolism in general has attracted considerably less attention. In recent years, however, a series of fascinating new discoveries have been made. For instance, Cr analogs have proven to be potent anticancer agents that act synergistically with currently used chemotherapeutics. Cyclocreatine, one of the Cr analogs, as well as PCr protect tissues from ischemic damage and may therefore have an impact on organ transplantation. Circumstantial evidence suggests a link between disturbances in Cr metabolism and muscle diseases as well as neurological disorders, and beneficial effects of oral Cr supplementation in such diseases have in fact been reported. Oral Cr ingestion has also been shown to increase athletic performance, and it therefore comes as no surprise that Cr is currently used by many athletes as a performance-boosting supplement. Some data suggest that Cr and creatinine (Crn) may act as precursors of food mutagens and uremic toxins. Finally, the recent identification, purification, and cloning of many of the enzymes involved in Cr metabolism have just opened the door to a wide variety of biochemical, physiological, as well as clinical investigations and applications.

The goal of this article is to provide a comprehensive overview on the physiology and pathology of Cr and Crn metabolism. Because some of these aspects have already been covered by earlier reviews (e.g., Refs. 55, 669, 1056, 1077), preference will be given to more recent developments in the field. The text is written in a modular fashion, i.e., despite the obvious fact that complex interrelationships exist between different parts of the text, every section should, by and large, be self-explanatory. It is our hope that this review will stimulate future multidisciplinary research on the physiological functions of the CK system, on the pathways and regulation of Cr metabolism, and on the relationships between disturbances in Cr metabolism and human disease.

	II. ABBREVIATIONS

Top Previous Next References

Cr	Creatine
Crn	Creatinine
PCr	Phosphorylcreatine
CK	Creatine kinase
Mi-CK	Mitochondrial CK isoenzyme
B-CK	Cytosolic brain-type CK isoenzyme
M-CK	Cytosolic muscle-type CK isoenzyme
AGAT	L-Arginine:glycine amidinotransferase
GAMT	S-adenosyl-L-methionine:N-guanidinoacetate methyltransferase
GPA	Guanidinopropionate, if not otherwise mentioned, the 3-guanidinopropionate or -guanidinopropionate isomer
GBA	Guanidinobutyrate
cCr	Cyclocreatine = 1-carboxymethyl-2-iminoimidazolidine
hcCr	Homocyclocreatine = 1-carboxyethyl-2-iminoimidazolidine
Gc	Glycocyamine = guanidinoacetate
Tc	Taurocyamine
L	Lombricine
PCrn, PGPA, PcCr, PhcCr, PArg, PGc, PTc, PL	N-phosphorylated forms of the respective guanidino compounds
ArgK	Arginine kinase
DNFB	2,4-Dinitrofluorobenzene
AdoMet	S-adenosyl-L-methionine
GSH	Reduced glutathione
GSSG	Oxidized glutathione
OAT	L-Ornithine:2-oxoacid aminotransferase

	III. THE PHYSIOLOGICAL RELEVANCE OF CREATINE: THE CREATINE KINASE REACTION

Top Previous Next References

To understand why nature has "developed" reaction pathways for the biosynthesis of PCr and of other phosphagens, one must briefly explain the main functions proposed for the CK/PCr/Cr system (for a detailed discussion and for references, see Refs. 837, 838, 1084, 1124). In textbooks of biochemistry, the participation of the CK/PCr/Cr system in energy metabolism is often neglected, and it is tacitly assumed that high-energy phosphate transport between sites of ATP production (mitochondria, glycolysis) and ATP consumption (all sorts of cellular ATPases) relies on diffusion of ATP and ADP alone. This concept may reflect the situation in tissues devoid of CK and PCr, like liver, but is clearly inadequate for CK-containing tissues with high and fluctuating energy demands like skeletal or cardiac muscle, brain, retina, and spermatozoa. In these latter tissues of mammals and birds, four distinct types of CK subunits are expressed species specifically, developmental stage specifically, and tissue specifically. The cytosolic M-CK (M for muscle) and B-CK (B for brain) subunits form dimeric molecules and thus give rise to the MM-, MB-, and BB-CK isoenzymes. The two mitochondrial CK isoforms, ubiquitous Mi-CK and sarcomeric Mi-CK, are located in the mitochondrial intermembrane space and form both homodimeric and homooctameric molecules that are readily interconvertible. All CK isoenzymes catalyze the reversible transfer of the -phosphate group of ATP to the guanidino group of Cr to yield ADP and PCr (Fig. 1).

View larger version (16K): [in this window] [in a new window]   Fig. 1. The creatine kinase (CK) reaction. PCr, phosphorylcreatine; Cr, creatine.

In fast-twitch skeletal muscles, a large pool of PCr is available for immediate regeneration of ATP hydrolyzed during short periods of intense work. Because of the high cytosolic CK activity in these muscles, the CK reaction remains in a near-equilibrium state, keeps [ADP] and [ATP] almost constant (over several seconds), and thus "buffers" the cytosolic phosphorylation potential that seems to be crucial for the proper functioning of a variety of cellular ATPases.

Heart, slow-twitch skeletal muscles, or spermatozoa, on the other hand, depend on a more continuous delivery of high-energy phosphates to the sites of ATP utilization. According to the "transport" ("shuttle") hypothesis for the CK system, distinct CK isoenzymes are associated with sites of ATP production (e.g., Mi-CK in the mitochondrial intermembrane space) and ATP consumption [e.g., cytosolic CK bound to the myofibrillar M line, the sarcoplasmic reticulum (SR), or the plasma membrane] and fulfill the function of a "transport device" for high-energy phosphates. The -phosphate group of ATP, synthesized within the mitochondrial matrix, is transferred by Mi-CK in the mitochondrial intermembrane space to Cr to yield ADP plus PCr. ADP liberated by the Mi-CK reaction may directly be transported back to the matrix where it is rephosphorylated to ATP. PCr leaves the mitochondria and diffuses through the cytosol to the sites of ATP consumption. There cytosolic CK isoenzymes locally regenerate ATP and thus warrant a high phosphorylation potential in the intimate vicinity of the respective ATPases. Cr thus liberated diffuses back to the mitochondria, thereby closing the cycle. According to this hypothesis, transport of high-energy phosphates between sites of ATP production and ATP consumption is achieved mainly (but not exclusively) by PCr and Cr. Whereas for the buffer function, no Mi-CK isoenzyme is required, Mi-CK may be a prerequisite for efficient transport of high-energy phosphates, especially if diffusion of adenine nucleotides across the outer mitochondrial membrane were limited (see sect. IVB). In accordance with these ideas, the proportion of Mi-CK seems to correlate with the oxidative capacity of striated muscles. It is by far higher in heart (up to 35% of total CK activity) than in fast-twitch skeletal muscles (0.5-2%).

Although the shuttle hypothesis seems logical and intelligible on first sight, there is an ongoing debate on whether it accurately describes the function of the CK system in endurance-type tissues (524, 837, 964). Therefore, the buffer and transport models for CK function should be regarded neither as strictly true nor as static views that can be applied directly to any one tissue; rather, the CK system displays a high degree of flexibility and is able to adapt to the peculiar physiological requirements of a given tissue. In skeletal muscle, for example, an adaptation of the CK system from a more buffer to a more transport type can be induced by endurance training or by chronic electrical stimulation (26, 861).

PCr and Cr, relative to ATP and ADP, are smaller and less negatively charged molecules and can build up to much higher concentrations in most CK-containing cells and tissues, thereby allowing for a higher intracellular flux of high-energy phosphates. Furthermore, the change in free energy (G°') (pH 7.0) for the hydrolysis of PCr is 45.0 kJ/mol compared with 31.8 kJ/mol for ATP, implying that in tissues with an active CK system, the cytosolic phosphorylation potential can be buffered at a higher level than in tissues devoid of the CK system. This factor may, again, be essential for the proper functioning of at least some cellular ATPases, e.g., the Ca²⁺-ATPase of the SR (see Ref. 646). Finally, by keeping [ADP] low, the CK/PCr/Cr system may also protect the cell from a net loss of adenine nucleotides via adenylate kinase, AMP deaminase, and 5'-nucleotidase.

	IV. CREATINE METABOLISM IN VERTEBRATES

Top Previous Next References

Although the pathways of Cr metabolism in vertebrates seem simple (Fig. 2), the situation is complicated by the fact that most tissues lack several of the enzymes required, thus necessitating transport of intermediates between the tissues through the blood to allow the whole cascade of reactions to proceed. In mammals, for instance, a complete urea cycle operates actively only in liver. The main site of Arg biosynthesis for other bodily tissues is, however, the kidney. Citrulline, synthesized in the liver or small intestine and transported through the blood, is taken up by the kidney and converted into Arg mainly by the proximal tubule of the nephron (273, 553). Arg formed within the kidney is then either released into the blood and consumed by other tissues or used within the kidney itself for guanidinoacetate synthesis.

View larger version (21K): [in this window] [in a new window]   Fig. 2. Schematic representation of the reactions and enzymes involved in vertebrate creatine and creatinine metabolism. The respective enzymes are denoted by numbers: 1) L-arginine:glycine amidinotransferase (AGAT; EC 2.1.4.1); 2) S-adenosyl-L-methionine:N-guanidinoacetate methyltransferase (GAMT; EC 2.1.1.2); 3) creatine kinase (CK; EC 2.7.3.2); 4) arginase (L-arginine amidinohydrolase; EC 3.5.3.1); 5) ornithine carbamoyltransferase (EC 2.1.3.3); 6) argininosuccinate synthase (EC 6.3.4.5); 7) argininosuccinate lyase (EC 4.3.2.1); 8) L-ornithine:2-oxo-acid aminotransferase (OAT; EC 2.6.1.13); N) nonenzymatic reaction.

The transfer of the amidino group of Arg to Gly to yield L-ornithine and guanidinoacetic acid (GAA) represents the first of two steps in the biosynthesis of Cr (Fig. 3) and is catalyzed by L-arginine:glycine amidinotransferase (AGAT; EC 2.1.4.1). GAA, by the action of S-adenosyl-L-methionine:N-guanidinoacetate methyltransferase (GAMT; EC 2.1.1.2), is then methylated at the amidino group to give Cr (M_r 131.1). In the course of evolution, both AGAT and GAMT seem to have evolved with the appearance of the lampreys (1056). These enzyme activities were not detected in invertebrates, whereas they are found in most but not all vertebrates examined. Some invertebrate species (e.g., some annelids, echinoderms, hemichordates, and urochordates) nevertheless contain significant amounts of Cr, PCr, and CK, particularly in spermatozoa (226, 811, 1056, 1092). This implies that these species either accumulate Cr from the environment or from the feed, or that the enzymes for Cr biosynthesis in these animals escaped detection so far.

View larger version (11K): [in this window] [in a new window]   Fig. 3. The AGAT reaction. The asterisk indicates the nitrogen atom to which a methyl group from S-adenosyl-L-methionine is transferred by GAMT to yield Cr.

Many of the lower vertebrates (fish, frogs, and birds) have both AGAT and GAMT in their livers and often also in the kidneys (see Refs. 634, 1056, 1077). In mammals, pancreas contains high levels of both enzymes, whereas kidneys express fairly high amounts of AGAT but relatively lower levels of GAMT. On the contrary, livers of all mammalian species tested so far contain high amounts of GAMT but display only low levels of Cr and almost completely lack CK activity. Although livers of cow, pig, monkey, and human also have high amounts of AGAT, livers of common laboratory mammals such as the rat, mouse, dog, cat, and rabbit were reported to lack AGAT activity. On the basis mostly of these latter findings and of the fact that the rate of Cr biosynthesis is considerably reduced in nephrectomized animals (248, 291, 554), it was postulated, and is still largely accepted, that the main route of Cr biosynthesis in mammals involves formation of guanidinoacetate in the kidney, its transport through the blood, and its methylation to Cr in the liver (Fig. 4). Cr exported from the liver and transported through the blood may then be taken up by the Cr-requiring tissues.

View larger version (36K): [in this window] [in a new window]   Fig. 4. Major routes of Cr metabolism in the mammalian body. The most part (up to 94%) of Cr is found in muscular tissues. Because muscle has virtually no Cr-synthesizing capacity, Cr has to be taken up from the blood against a large concentration gradient by a saturable, Na+- and Cl-dependent Cr transporter that spans the plasma membrane (). The daily demand for Cr is met either by intestinal absorption of dietary Cr or by de novo Cr biosynthesis. The first step of Cr biosynthesis probably occurs mainly in the kidney, whereas the liver is likely to be the principal organ accomplishing the subsequent methylation of guanidinoacetic acid (GAA) to Cr. It must be stressed that the detailed contribution of different bodily tissues (pancreas, kidney, liver, testis) to total Cr synthesis is still rather unclear and may vary between species (see text). The muscular Cr and PCr are nonenzymatically converted at an almost steady rate (~2% of total Cr per day) to creatinine (Crn), which diffuses out of the cells and is excreted by the kidneys into the urine.

Several lines of experimental evidence, however, demonstrate that this concept of the organization of Cr metabolism is too simplistic. Pyridoxine-deficient rats, despite a 65% decrease in kidney AGAT activity relative to controls, displayed increased liver and skeletal muscle concentrations of Cr (572). In another study, extrarenal synthesis was suggested to account for 40-60% of total Cr (290). Similarly, comparison of the hepatic and renal venous levels with the arterial levels of Arg, GAA, and Cr suggested that in humans, the liver is the most important organ contributing to biosynthesis of both GAA and Cr, whereas the kidney plays only a secondary role (842). In accordance with these observations, immunofluorescence microscopy revealed significant amounts of AGAT not only in rat kidney and pancreas, but also in liver (623). The apparent discrepancy from earlier investigations may be explained by the high levels of liver arginase interfering with AGAT activity assays. Furthermore, AGAT activity was detected in heart, lung, spleen, muscle, brain, testis, and thymus, and it has been estimated that the total amount of AGAT in these tissues approaches that found in kidney and pancreas (769, 1055). Although AGAT is absent from human placenta, the decidua of pregnant females displayed the highest specific AGAT activity of all rat tissues examined (1077), implying a major involvement of this tissue in Cr biosynthesis during early stages of development. In line with this conclusion, maternofetal transport of Cr was demonstrated in the rat (157).

On the other hand, GAMT mRNA and protein levels higher than those observed in male liver were detected in mouse testis, caput epididymis, and ovary (see Ref. 543). Likewise, Sertoli cells of rat seminiferous tubules, in contrast to germ cells and interstitial cells, were shown to synthesize guanidinoacetate and Cr from Arg and Gly (664). It may therefore be hypothesized that the Cr-synthesizing machinery in reproductive organs plays a role in the development or function of germ cells. GAMT activity was also detected in rat spleen, heart, and skeletal muscle, in sheep muscle, as well as in human fetal lung fibroblasts and mouse neuroblastoma cells (149, 1130, 1135, 1136). Although the specific activities in these tissues are rather low, the GAMT activity in skeletal muscle was calculated to have the potential to synthesize all Cr needed in this tissue (149). Finally, feeding of rats and mice with 3-guanidinopropionic acid (GPA), a competitive inhibitor of Cr entry into cells, progressively decreased the concentrations of Cr and PCr in heart and skeletal muscle but had only little influence on the Cr and PCr contents of brain (372). One possible explanation is that the brain contains its own Cr-synthesizing machinery (171). To conclude, the detailed contribution of the various tissues of the body to total Cr biosynthesis as well as the relevance of guanidinoacetate and Cr transport between the tissues are still rather unclear; this is due both to a lack of thorough investigations and to the pronounced species differences observed so far.

A specific, saturable, Na⁺- and Cl-dependent Cr transporter responsible for Cr uptake across the plasma membrane has been described for skeletal muscle, heart, smooth muscle, fibroblasts, neuroblastoma and astroglia cells, as well as for red blood cells and macrophages (149, 150, 250, 570, 659, 711, 876, 965). These findings have recently been corroborated by cDNA cloning and Northern blot analysis of the rabbit, rat, mouse, and human Cr transporters (295, 319, 415, 543, 691, 697, 840, 860, 927). Although the quantitative results of these latter studies differ to some extent, the highest amounts of Cr transporter mRNA seem to be expressed in kidney, heart, and skeletal muscle; somewhat lower amounts in brain, small and large intestine, vas deferens, seminal vesicles, epididymis, testis, ovary, oviduct, uterus, prostate, and adrenal gland; and only very low amounts or no Cr transporter mRNA at all in placenta, liver, lung, spleen, pancreas, and thymus.

An important aspect of Cr biosynthesis to add is that in humans, the daily utilization of methyl groups in the GAMT reaction approximately equals the daily intake of "labile" methyl groups (Met + choline) on a normal, equilibrated diet (671). Even if de novo Met biosynthesis also is taken into account, Cr biosynthesis still accounts for ~70% of the total utilization of labile methyl groups in the body. Upon lowering of the Met and choline levels in the diet, the deficit in labile methyl groups is compensated for by increased de novo Met biosynthesis, indicating that the delivery of labile methyl groups, in the form of S-adenosyl-L-methionine, should normally not become limiting for Cr biosynthesis. It may do so, however, in folic acid and/or vitamin B₁₂ deficiency (231, 945) as well as in other physiological and pathological conditions that are characterized by an impairment of S-adenosyl-L-methionine synthesis (e.g., Refs. 118, 122, 188, 243).

The highest levels of Cr and PCr are found in skeletal muscle, heart, spermatozoa, and photoreceptor cells of the retina. Intermediate levels are found in brain, brown adipose tissue, intestine, seminal vesicles, seminal vesicle fluid, endothelial cells, and macrophages, and only low levels are found in lung, spleen, kidney, liver, white adipose tissue, blood cells, and serum (61, 127, 175, 525, 547, 568, 570, 693, 759, 1080, 1082, 1083, 1108, 1136). A fairly good correlation seems to exist between the Cr transporter mRNA level and total CK activity which, in turn, also correlates with the tissue concentration of total Cr (Cr + PCr; Fig. 5). There may be only two exceptions. 1) Kidney displays a much higher Cr transporter content than expected from its CK activity (Fig. 5A, ), which might be due to an involvement of the Cr transporter in the resorption of Cr from the primary urine. 2) Liver has a considerably lower CK activity than expected from its Cr content (Fig. 5B, ), which may be an expression of a strict separation between Cr-synthesizing and CK-expressing tissues in the body. Such a separation may be a crucial prerequisite for independent regulation of Cr biosynthesis on one hand and CK function/energy metabolism on the other hand.

View larger version (17K): [in this window] [in a new window]   Fig. 5. Correlations between Cr transporter level, CK activity, and total Cr content in different mammalian (rat, human, cat, dog, rabbit, mouse, and guinea pig) tissues. The respective tissues are, from left to right, as follows: A: pancreas, spleen, ovary, lung, small intestine, prostate, brain, colon, heart, kidney (), and skeletal muscle. B: spleen, kidney, liver (), smooth muscle (carotid artery), macrophages, brown adipose tissue, uterus, brain, heart, and skeletal muscle. (Data were taken from Refs. 56, 61, 129, 137, 172, 227, 529, 569, 570, 691, 1020, 1030.)

Resting type 2a and 2b skeletal muscle fibers of rodents contain ~32 mM PCr and 7 mM Cr, whereas type 1 fibers comprise ~16 mM PCr and 7 mM Cr (525). The difference in PCr concentration between type 1 and type 2 muscle fibers is less pronounced in humans (337, 844, 1042); nevertheless, the concentration of total Cr seems to parallel the muscle glycolytic capacity in both rodents and humans. In serum and erythrocytes, as opposite extremes, [Cr] amounts to only 25-100 µM and 270-400 µM, respectively (175, 776, 1137), implying that Cr has to be accumulated by most Cr-containing tissues against a large concentration gradient from the blood. This accumulation via the Cr transporter is driven by the electrochemical potential differences of extracellular versus intracellular [Na⁺] and [Cl].

Because both seminal vesicles and seminal vesicle fluid of the rat and mouse contain considerable quantities of Cr and PCr, it was hypothesized "that both compounds are actively secreted by the seminal vesicle epithelium" (542). This hypothesis later turned out to be incorrect, in as far as seminal vesicles were shown to lack AGAT and GAMT but to contain moderate amounts of Cr transporter mRNA (543). Therefore, seminal vesicles most likely accumulate Cr from the blood.

For PCr and Cr, a single cytosolic pool is assumed by most researchers, especially by those who postulate near-equilibrium conditions for the CK reaction throughout the cell. However, tracer studies with [¹⁴C]Cr suggested distinct cytosolic pools of Cr in rat heart (850) and fast-twitch (white) muscle of the rainbow trout (369). In addition, quantitative X-ray microanalysis revealed that phosphorus compounds (presumably represented mostly by PCr and ATP) are highly compartmentalized in sarcomeric muscle, with a preferential occupancy of the I band as well as the H zone (549). Surprisingly, some researchers detected Cr and PCr in the matrix of heart mitochondria and provided evidence that PCr uptake into the mitochondrial matrix is mediated by the adenine nucleotide translocase (see Refs. 391, 796, 921). In the light of 1) the lower phosphorylation potential within the mitochondrial matrix compared with the cytosol and 2) the lack of Cr-utilizing processes in the mitochondrial matrix, it is questionable whether and to what extent Cr accumulation in the matrix is physiologically relevant or is due to postmortem or other artifacts. Clearly, further studies are needed to get a deeper insight into the subcellular compartmentation of Cr and PCr.

The degradation of Cr and PCr in vertebrates is, for the most part, a spontaneous, nonenzymatic process, as indicated in the top part of Figure 2. In vitro, the equilibrium of the reversible and nonenzymatic cyclization of Cr to creatinine (Cr  Crn) is both pH dependent and temperature dependent. Cr is favored at high pH and low temperature, whereas Crn (M_r 113.1) is favored at elevated temperatures and in acidic solutions (see Ref. 551). In both directions, the reaction is monomolecular. Starting with pure Cr solutions, 1.0-1.3% of the Cr per day is converted into Crn at pH 7.0-7.2 and 38°C. In vitro studies on the stability of PCr revealed that this high-energy phosphate compound is acid labile, yielding P_i and either Cr or Crn upon hydrolysis. Both the rate of PCr hydrolysis and the ratio of Cr to Crn formed depend on temperature and pH and can additionally be influenced in a concentration-dependent manner by molybdate (for reviews, see Refs. 226, 669).

In contrast to these in vitro studies, experiments with ¹⁵N-labeled compounds clearly showed that the conversion of Cr into Crn in vivo is an irreversible process (72). Upon feeding of rats with [¹⁵N]Cr, the isotopically labeled Cr distributed homogeneously over the total Cr pool in the body as well as over the urinary Crn. Even after 5 days, the specific labeling of the urinary Crn and the body Cr were still the same, suggesting that Cr is the only precursor of Crn. Upon feeding with [¹⁵N]Crn, however, most of the label was directly excreted into the urine, and no significant exchange of the label with the body Cr was observed. In accordance with in vitro studies, an almost constant fraction of the body Cr (1.1%/day) and PCr (2.6%/day) is converted nonenzymatically into Crn in vivo, giving an overall conversion rate for the total Cr pool (Cr + PCr) of ~1.7%/day (for a review, see Ref. 1077). Consequently, in a 70-kg man containing ~120 g of total Cr, roughly 2 g/day are converted into Crn and have to be replaced by Cr from the diet or from de novo biosynthesis (Fig. 4) (1050, 1077, 1085). With the assumption of an average content in muscle of 30 mM of total Cr (see above) and a quantitative uptake of the compound by the digestive tract, this loss could be compensated by ingestion of 500 g of raw meat per day. Because Crn is a very poor substrate of the Cr transporter (318, 319, 691), because no other specific saturable uptake mechanism exists for Crn (515), and because Crn, most likely due to its nonionic nature, is membrane permeable, Crn constantly diffuses out of the tissues into the blood and is excreted by the kidneys into the urine (Fig. 4) (759). Because the rate of nonenzymatic formation of Crn from Cr is nearly constant, and because >90% of the total bodily Cr is to be found in muscle tissue, 24-h urinary Crn excretion is frequently used as a rough measure of total muscle mass (768, 1067). However, this approach suffers various limitations.

Twenty to twenty-five percent of the in vivo conversion of PCr into Crn may proceed via phosphorylcreatinine (PCrn) as an intermediate (414). Accordingly, [PCrn] in rabbit white skeletal muscle was found to be 0.4% of [PCr], and commercial preparations of PCr (at least several years ago) contained 0.3-0.7% of PCrn.

Crn in mammals, and especially in humans, is still widely believed to be an inert substance that is excreted as such into the urine. Several lines of evidence, however, contradict this view. Using radiolabeled Crn, Boroujerdi and Mattocks (83) showed that in rabbits, some Crn is converted into Cr, Arg, guanidinobutyrate, or guanidinopropionate. Additional routes of Crn degradation become favored in states of renal insufficiency and seem to be relevant for human pathology. They are therefore discussed in detail in section IXH.

	V. REGULATION OF CREATINE METABOLISM IN VERTEBRATES

Top Previous Next References

In keeping with the rather complex organization of Cr biosynthesis and degradation in vertebrates, a variety of potential regulatory mechanisms have to be considered, for instance, allosteric regulation, covalent modification, or changes in expression levels of the enzymes involved in Cr metabolism. In addition, changes in the transport capacity and/or permeability of biological membranes for the intermediary metabolites, i.e., Cr, Crn, and guanidinoacetate, are also expected to have an impact on Cr metabolism as a whole (for an extensive review, see Ref. 1077).

The formation of guanidinoacetate is normally the rate-limiting step of Cr biosynthesis (see Ref. 1077). Consequently, the AGAT reaction is the most likely control step in the pathway, a hypothesis that is supported by a great deal of experimental work. Most important in this respect is the feedback repression of AGAT by Cr, the end-product of the pathway, which most probably serves to conserve the dietary essential amino acids Arg and Met. Circumstantial evidence indicates that in folic acid deficiency, where Cr biosynthesis is curtailed and the serum concentration of Cr is likely to be decreased, AGAT expression is upregulated (187). In contrast, an increase in the serum concentration of Cr, due either to an endogenous source or to dietary Cr supplementation, results in concomitant decreases in the mRNA content, the enzyme level, and the enzymatic activity of AGAT, thus suggesting regulation of AGAT expression at a pretranslational level (322, 1053; for a review, see Ref. 1077). Feedback repression of AGAT by Cr is most pronounced in kidney and pancreas, the main tissues of guanidinoacetate formation, but is also observed in the decidua of pregnant rats (see Ref. 1077). Immunological studies suggest the presence of multiple forms (or isoenzymes) of AGAT in rat kidney, of which only some are repressible by Cr, whereas others are not (314). Because the half-life of AGAT in rat kidney is 2-3 days (624), the changes in the AGAT levels described here are rather slow processes, thus only allowing for long-term adaptations.

Cyclocreatine, N-acetimidoylsarcosine, and N-ethylguanidinoacetate display repressor activity like Cr, whereas Crn, PCr, N-propylguanidinoacetate, N-methyl-3-guanidinopropionate, N-acetimidoylglycine, and a variety of other compounds are ineffective (809, 1077). L-Arg and guanidinoacetate have only "apparent" repressor activity. They exert no effect on AGAT expression by themselves but are readily converted to Cr, which then acts as the true repressor.

In addition to Cr, the expression of AGAT may be modulated by dietary and hormonal factors (for reviews, see Refs. 634, 1053, 1054, 1077). Thyroidectomy or hypophysectomy of rats decreases AGAT activity in the kidney. The original AGAT activity can be restored by injection of thyroxine or growth hormone, respectively. In contrast, injections of growth hormone into thyroidectomized rats and of thyroxine into hypophysectomized rats are without effect, indicating that both hormones are necessary for maintaining proper levels of AGAT in rat kidney. Because enzyme activity, protein, and mRNA contents are always affected to the same extent, regulation of AGAT expression by thyroid hormones and growth hormone occurs at a pretranslational level, very similar to the feedback repression by Cr (322, 625, 1053). Growth hormone and Cr have an antagonistic action on AGAT expression, as evidenced by identical mRNA levels and enzymatic activities of kidney AGAT in hypophysectomized rats simultaneously fed Cr and injected with growth hormone compared with hypophysectomized rats receiving neither of these compounds (322, 1053).

AGAT levels in liver, pancreas, and kidney are also decreased in conditions of dietary deficiency and disease (fasting, protein-free diets, vitamin E deficiency, or streptozotocin-induced diabetes) (273, 1057). These findings seem, however, not to rely directly on the dietary or hormonal imbalance that is imposed. For example, insulin administration to streptozotocin-diabetic rats does not restore the original AGAT activity in the kidney (273). On the contrary, fasting and vitamin E deficiency are characterized by an increased blood level of Cr (248, 480; see also Ref. 1077) which, in all likelihood, represents the true signal for the downregulation of AGAT expression.

Finally, AGAT levels in rat kidney, testis, and decidua may also be under the control of sex hormones, with estrogens and diethylstilbestrol decreasing and testosterone increasing AGAT levels (449; see also Ref. 1077). Oral administration of methyltestosterone to healthy humans not only stimulates AGAT expression and, thus, Cr biosynthesis, but also results in a 70% increase in the urinary excretion of guanidinoacetate (367). This finding might be taken to indicate that at increased levels of AGAT activity, GAMT becomes progressively rate limiting for Cr biosynthesis, thereby leading to an accumulation of guanidinoacetate in the blood. In conflict with this interpretation, dietary Cr supplementation, which is known to decrease AGAT levels in kidney and pancreas, also results in increased urinary guanidinoacetate excretion. Furthermore, guanidinoacetate excretion is much higher when Cr and guanidinoacetate are administered simultaneously than when Cr or guanidinoacetate is given alone (368). Therefore, it is more likely that in situations of elevated Cr concentrations in the blood, the increased levels of Cr in the primary filtrate compete with guanidinoacetate for reabsorption by the kidney tubules (see Ref. 1077).

Based on the findings that GAMT expression in the mouse is highest in testis, caput epididymis, ovary, and liver, and that GAMT expression is higher in female than in male liver, it has been hypothesized that GAMT expression might also be under the control of sex hormones (545). However, removal of either adrenals, pituitaries, gonads, or thyroids and parathyroids or administration of large doses of insulin, estradiol, testosterone, cortisol, thyroxine, or growth hormone had, if any, only minor effects on GAMT activity in rat liver (109). There is some indication that GAMT activity in the liver may be influenced by dietary factors (1019).

In contrast to the described repression by Cr of AGAT in kidney and pancreas, Cr does not interfere with the expression of GAMT or arginase in liver. Cr, Crn, and PCr also do not regulate allosterically the enzymatic activities of AGAT or GAMT in vitro (1077). In contrast, AGAT is potently inhibited by ornithine, which may be pathologically relevant, for instance, in gyrate atrophy of the choroid and retina (see sect. IXA) (897, 1077). A striking parallelism between the enzymes involved in vertebrate Cr metabolism (AGAT, GAMT, CK) is that they all are sensitive to modification and inactivation by sulfhydryl reagents (for reviews, see Refs. 270, 474, 1077). On the basis of current knowledge (e.g., Ref. 496), however, there is no reason to believe that modification by sulfhydryl reagents [e.g., oxidized glutathione (GSSG)] represents a unifying mechanism for the in vivo regulation of AGAT, GAMT, and CK.

In chicken kidney and liver, where AGAT is localized in the mitochondrial matrix, penetration of L-Arg through the inner membrane was found to occur only in respiring mitochondria and in the presence of anions such as acetate or phosphate (301). Consequently, the rate of Arg transport across the mitochondrial membranes might influence Cr biosynthesis.

Cr uptake into CK-containing tissues, e.g., skeletal muscle, heart, brain, or kidney, is effected by a specific, saturable, Na⁺- and Cl-dependent Cr transporter (see sect. VIIC). Even though the evidence is not as strong as in the case of AGAT, the expression and/or specific activity of the Cr transporter seems to be influenced by dietary and hormonal factors. A 24-h fast slightly increases [Cr] in the plasma but decreases Cr uptake into tibialis anterior and cardiac muscle of the mouse by ~50% (480). In rats, Cr supplementation of the diet decreases Cr transporter expression (317). Similarly, in rat and human myoblasts and myotubes in cell culture, extracellular Cr downregulates Cr transport in a concentration- and time-dependent manner (571). Na⁺-dependent Cr uptake is decreased by extracellular [Cr] >1 µM, with 50% inhibition being observed at 20-30 µM, i.e., in the range of the physiological plasma concentration of Cr. In media containing 5 mM Cr, transport of Cr is decreased by 50% within 3-6 h, and maximal inhibition (70-80%) is observed within 24 h. Upregulation of Cr transport upon withdrawal of extracellular Cr seems to occur more slowly. Excessive concentrations (5 mM) of guanidinoacetate and GPA also reduce Cr transport significantly, whereas D- and L-ornithine, Crn, Gly, and PCr are ineffective. Because the downregulation of the Cr transporter activity by extracellular Cr is slowed by cycloheximide, an inhibitor of protein synthesis, it has been hypothesized that Cr transport, like Na⁺-dependent system A amino acid transport (331), is controlled by regulatory proteins. However, no conclusive evidence for or against this hypothesis is currently available. It also remains to be clarified how extracellular [Cr] is transformed into an intracellular signal. Loike et al. (571) have presented weak evidence suggesting that Cr has to be taken up into the cells to exert its effect on Cr transporter activity. On the other hand, dietary Cr supplementation in humans and animals, despite an at least 3- to 20-fold increase in the serum concentration of Cr, results in only a 10-20% increase in the muscle levels of Cr (see sect. XI). Because, in addition, this latter increase in muscle [Cr] is much lower than the ones observed during physical exercise, it is difficult to envisage that intracellular [Cr] should be a key regulator of Cr uptake.

In a thorough investigation of the Cr transporter activity in cultured mouse G8 myoblasts, Odoom et al. (711) showed that Cr uptake is stimulated by isoproterenol, norepinephrine, the cAMP analog N⁶,2'-O-dibutyryladenosine 3',5'-cyclic monophosphate, and the ₂-agonist clenbuterol, but not by the ₁-adrenergic receptor agonist methoxamine. Likewise, the stimulatory action of norepinephrine is not affected by -adrenergic receptor antagonists but is inhibited by -antagonists, with the ₂-antagonist butoxamine being more effective than the ₁-antagonist atenolol. Thus the Cr transporter activity may be controlled predominantly by ₂-adrenergic receptors that have cAMP as their intracellular signal. In fact, analysis of the Cr transporter cDNA sequence revealed consensus phosphorylation sites for cAMP-dependent protein kinase (PKA) and for protein kinase C (PKC) (691, 927). However, in transiently transfected cells expressing the human Cr transporter, phorbol 12-myristate 13-acetate, an activator of PKC, displayed a small inhibitory effect on Cr uptake, whereas forskolin (an activator of adenylyl cyclase), okadaic acid (a phosphatase inhibitor), A23187 (a calcium ionophore), and insulin were ineffective. The last finding, in turn, contrasts with experiments on rat skeletal muscle where insulin significantly increased Cr uptake, whereas alloxan-induced diabetes had no effect on Cr accumulation (see Ref. 349). Insulin and insulin-like growth factor I also stimulated Cr uptake in mouse G8 myoblasts (711), and insulin at physiologically high or supraphysiological concentrations enhanced muscle Cr accumulation in humans (943). Insulin increases Na⁺-K⁺-ATPase activity which, indirectly, may stimulate Cr transporter activity (see Ref. 943). In this context, it seems noteworthy that guanidinoacetate, and to a lower extent Arg and Cr, were seen to stimulate insulin secretion in the isolated perfused rat pancreas (15). Despite using G8 myoblasts and myotubes as Odoom et al. (711; see above), and despite other indications that clenbuterol may exert some of its anabolic effects on muscle by stimulating Cr uptake, Thorpe et al. (1003) failed to detect an effect of clenbuterol on Cr transport.

The contents of Cr, PCr, and total Cr are decreased in hyperthyroid rat cardiac muscle by 13, 62, and 42%, respectively, with these changes being paralleled by an increased sensitivity of the heart to ischemic damage (874). Although this finding might be explained by a direct action of thyroid hormones on the Cr transporter, experiments with colloidal lanthanum suggest that it is due instead to an increased (reversible) leakiness of the sarcolemma. Kurahashi and Kuroshima (519) suggested that the 3,3',5-triiodothyronine-induced creatinuria and decrease in muscle Cr contents is due both to decreased uptake and increased release of Cr by the muscles. On the other hand, Cr uptake into mouse G8 myoblasts was shown to be stimulated by 3,3',5-triiodothyronine and by amylin which, in muscle, is known to bind to the calcitonin gene-related peptide receptor (711).

As to be expected from the Na⁺ dependence of the Cr transporter (see sect. VIIC), Cr uptake is diminished in deenergized cells and is also depressed by the Na⁺-K⁺-ATPase inhibitors ouabain and digoxin (58, 293, 515, 570, 711). When, however, L6 rat myoblasts are preincubated with ouabain or digoxin, and Cr uptake subsequently is analyzed in the absence of these inhibitors, it is even higher than in untreated control cells (58). Finally, in erythrocytes from uremic patients, the Na⁺-dependent component of Cr influx is 3.3 times higher than in normal human erythrocytes. This finding may be due, by analogy, to the known occurrence of inhibitors of Na⁺-K⁺-ATPase in uremic plasma (950, 984). Obviously, cells may respond to decreased Na⁺-K⁺-ATPase activity, which in turn likely decreases Cr transporter activity, by compensatory upregulation of Na⁺-K⁺-ATPase (382) and/or Cr transporter expression.

After incubation of L6 rat myoblasts for 20 h under control conditions, replacement of the conditioned medium by fresh control medium decreases Cr uptake by 32-45% (58). This may indicate that conditioned medium from L6 myoblasts contains a modulator of Cr transport.

Despite all these investigations on the regulation of Cr uptake, it cannot be decided yet whether regulation of Cr uptake is effected directly by modulating the expression and/or activity of the Cr transporter or indirectly via alterations of the transmembrane electrochemical gradient of Na⁺ which depends primarily on the Na⁺-K⁺-ATPase activity. Accordingly, it is still unclear whether Cr uptake via the Cr transporter is under kinetic or thermodynamic control. The findings that Cr uptake is inhibited by ouabain and digoxin and that 3,3',5-triiodothyronine, isoproterenol, and amylin not only stimulate Cr uptake but also increase the Na⁺-K⁺-ATPase activity and, thus, the membrane potential would favor indirect regulation of the Cr transporter by the electrochemical gradient of Na⁺. However, with the assumption of a Na⁺ to Cr stoichiometry of the Cr transporter of 1 or 2, the theoretical concentration ratio of intracellular versus extracellular Cr should be between 900 and 3,000 (286, 711). If the chloride dependence of the Cr transporter were also taken into account, this theoretical ratio would be even higher. In sharp contrast to these values, the actual concentration ratio in resting muscle is around 80. Because, in addition, dietary Cr supplementation over several days or weeks considerably increases [Cr] in human and animal serum, but only slightly enhances the Cr levels in muscle (see sect. XI), and because in rats fed GPA and cyclocreatine, these Cr analogs compete efficiently with Cr uptake into muscle and thereby largely deplete the intracellular pools of Cr and PCr, the hypothesis that the Cr transporter is kinetically controlled seems at present more plausible. Clearly, the question of how Cr uptake is regulated in detail is of importance for a deeper understanding of Cr metabolism in health and disease. In particular, it will be crucial to determine the exact Na⁺ and Cl stoichiometries of the Cr transporter.

Because part of the Cr that is accumulated in CK-containing tissues is converted to PCr, it might be anticipated that Cr uptake and phosphate uptake influence each other. In fact, in mouse myoblasts that are exposed to extracellular Cr, P_i uptake is transiently stimulated (773). This finding is probably not due to concerted regulation of the Cr and P_i transporters but may rely on a local decrease in P_i concentration due to phosphorylation of intracellularly accumulated Cr. In Langendorff-perfused rabbit hearts, the intracellular concentrations of Cr and of Cr plus PCr remain significantly higher when the perfusion medium is devoid of phosphate than when it contains 1 mM P_i (286). This effect was attributed to decreased Cr efflux during phosphate-free perfusion.

Only few and inconclusive data are available on Cr efflux from cells. Although in L6 rat myoblasts at 37°C, Cr efflux amounted to 2.8-3.6% of intracellular Cr per hour (571), the respective value for G8 mouse myoblasts at 37°C was 5%/day (711). The latter value is comparable to the fractional conversion rate of Cr to Crn and may indicate that the plasma membrane is largely impermeable for Cr once it is intracellularly trapped. Because the liver is the main site of Cr biosynthesis in the body, the plasma membrane of hepatocytes is expected to be more permeable for Cr than that of muscle cells. This finding agrees with the fact that upon administration of Cr, liver, kidney, and viscera constitute a rapidly expansible pool for Cr, whereas muscle and nervous tissues constitute a slowly expansible pool of Cr plus PCr (480; see also Ref. 1077). On the other hand, when transgenic mice expressing CK in liver were fed 10% Cr in the diet for 5 days, Cr efflux from the liver proved to be insignificant (606). Because high dietary intake of Cr makes de novo biosynthesis of Cr superfluous, a putative transport protein responsible for Cr export from the liver may simply have been downregulated in this experimental set-up. In any case, this finding should not be taken as evidence against a significant contribution of the liver to de novo biosynthesis of Cr in vertebrates. Finally, cultured Sertoli cells from the seminiferous epithelium of rats were shown to secrete Cr into the medium (665). Cr secretion was stimulated by physiological and toxicological modulators of Sertoli cell function like follicle-stimulating hormone, dibutyryl cAMP, mono-(2-ethylhexyl)phthalate, or cadmium.

The permeability itself as well as changes in permeability of the outer mitochondrial membrane may be critical for the stimulation of mitochondrial respiration and high-energy phosphate synthesis, as well as for the transport of these high-energy phosphates between sites of ATP production and ATP utilization within the cell (for reviews, see Refs. 94, 280, 838, 1124). Changes in permeability of the outer mitochondrial membrane pore protein (voltage-dependent anion-selective channel; VDAC) may be accomplished 1) by "capacitive coupling" to the membrane potential of the inner membrane, leading to a voltage-dependent "closure" of the pore, or 2) by a VDAC modulator protein which increases the rate of voltage-dependent channel closure by ~10-fold. To what extent these mechanisms operate in vivo and retard the diffusion of ADP, ATP, P_i, Cr, and PCr remains to be established.

To conclude, the most critical determinant for the regulation of Cr metabolism seems to be the serum concentration of Cr. An elevation of serum [Cr] over an extended period of time would point to excess de novo biosynthesis or dietary intake of Cr and, in addition, would indicate that the tissue pools of Cr and PCr are replenished. The observed or suspected effects of an elevated serum [Cr], namely, to downregulate the expression and/or activity of AGAT and possibly also the Cr transporter, would therefore help to spare precursors of Cr (Arg, Gly, Met) and to maintain normal, steady levels of Cr and PCr in CK-containing tissues. As a consequence, the rate of Cr biosynthesis is highest in young, healthy, fast-growing vertebrates under anabolic conditions on a balanced, Cr-free diet (1077).

	VI. PHOSPHOCREATINE AND CREATINE AS PURPORTED ALLOSTERIC EFFECTORS

Top Previous Next References

In some recent articles (e.g., Refs. 92, 641, 1068), the opinion has still been expressed that Cr and PCr may act as allosteric regulators of cellular processes. As a matter of fact, a number of studies, mainly performed in the 1970s, seemed to demonstrate that physiological concentrations of PCr inhibit glycogen phosphorylase a, phosphofructokinase, glyceraldehyde-3-phosphate dehydrogenase, pyruvate kinase, lactate dehydrogenase, AMP deaminase, and 5'-nucleotidase from a variety of species. Similarly, PCr was claimed to activate fructose-1,6-diphosphatase from rabbit skeletal muscle, while phosphorylarginine was suggested to inhibit phosphofructokinase from oyster adductor muscle (for references see Refs. 207, 247, 580, 666, 834, 951, 1012, 1099).

Subsequent studies, however, proved that inhibition of at least phosphofructokinase, pyruvate kinase, lactate dehydrogenase, AMP deaminase, and 5'-nucleotidase, but probably also of all other enzymes listed above, is not afforded by PCr itself; rather, these effects were due to contaminants present in the commercial preparations of PCr which, at that time, were no more than 62-75% pure (1099). The contaminating inhibitors were identified as inorganic pyrophosphate for AMP deaminase (1099) and oxalate for lactate dehydrogenase and pyruvate kinase (1012).

When added to the bathing medium of differentiating skeletal and heart muscle cells in tissue culture, Cr increased rather specifically the rate of synthesis as well as the specific activity of myosin heavy chain (406, 1153). In slices of the rat neostriatum, Cr inhibited the GABA-synthesizing enzyme glutamate decarboxylase as well as the veratridine-induced release of GABA, but significant effects were only observed at an unphysiologically high Cr concentration of 25 mM (864). In rat basophilic leukemia cells, PCr was seen to stimulate phospholipase C activity (196). Finally, in cell-free extracts of white gastrocnemius, soleus, heart muscle, and liver of the rat, PCr affected the extent of phosphorylation of various proteins, in particular of phosphoglycerate mutase and of a 18-kDa protein (742). Again, these findings are unlikely to be due to direct allosteric effects of Cr and PCr but are probably mediated indirectly, e.g., via alterations of the energy status of particular microcompartments or whole cells (195, 563, 742, 1153).

In the unicellular alga Gonyaulax polyedra, several functions show circadian rhythmicity, for example, cell division, photosynthesis, bioluminescence, motility, and pattern formation. If cultures of Gonyaulax are first grown under a 12:12-h light-dark cycle, and if the conditions are then changed to constant dim light, the circadian rhythmicity persists for several weeks. This condition is called free-running circadian rhythmicity, with its period  depending on the color and intensity of the constant dim light.

Extracts of several eukaryotic organisms, including bovine and rat brain and muscle, shorten the period of the free-running circadian rhythms in Gonyaulax. The substance responsible for this effect was identified as Cr. In the micromolar range (2-20 µM), Cr accelerates the circadian clock by as much as 4 h/day (Fig. 6A) (816). The Cr effect on  is very pronounced in constant dim blue light, whereas it is virtually absent in constant dim red light (Fig. 6B) (817). This finding, together with other lines of evidence, suggests that Cr interferes with light transduction pathways and in particular with the pathway(s) coupled to blue-sensitive photoreceptors. In addition to its effect on , Cr also affects light-induced phase changes of the circadian rhythmicity (817).

View larger version (38K): [in this window] [in a new window]   Fig. 6. Effects of creatine and gonyauline on the period of free-running circadian rhythms in the unicellular marine alga Gonyaulax polyedra. A: effects of authentic natural gonyauline, synthetic gonyauline, and Cr on the bioluminescent glow rhythm of Gonyaulax. [Modified from Roenneberg et al. (815).] B: relationship between Cr concentration and period length , under conditions of constant dim red or constant dim blue light. [Modified from Roenneberg and Taylor (817).] C: chemical structures of Cr and gonyauline. For further information see text.

A period-shortening substance with properties similar to Cr is present in extracts of Gonyaulax itself (816) and has been identified as gonyauline (S-methyl-cis-2-[methylthio]cyclopropanecarboxylic acid) (815). Its rather close structural similarity to Cr (Fig. 6C), the complete lack of Cr in extracts of Gonyaulax (815), as well as the indication that Cr is not active as such but has to be metabolized to exert its effects on the circadian rhythmicity (see Ref. 817) all suggest that, at least in algae, Cr itself is not a physiological component or modulator of the circadian clock.

To conclude, the data suggesting that Cr and PCr act as direct (allosteric) regulators of cellular processes (other than the repression by Cr of AGAT and possibly also of the Cr transporter) must be treated with skepticism, at least until new, convincing data are presented. There may be four notable exceptions that deserve further attention: 1) PCr stimulates glutamate uptake into synaptic vesicles (1127; see also sect. IXG). A series of control experiments showed that the effect of PCr is not mediated indirectly via CK and ATP. Remarkably, PCr-stimulated glutamate uptake was even higher than that stimulated maximally by ATP. 2) PCr at relatively high concentrations of 10-60 mM was found to promote efficient endonucleolytic cleavage of mammalian precursor RNA in vitro (364), which is a prerequisite for subsequent poly(A) addition. Neither CK nor ATP was required for this effect, and ATP could in fact inhibit 3'-end cleavage. PCr was not hydrolyzed, suggesting that it may act as an allosteric regulator. Phosphorylarginine (PArg) had a similar effect, whereas Cr was ineffective. 3) AMP-activated protein kinase (AMPK) from rabbit skeletal muscle is inhibited by PCr, whereby it is not yet completely clear whether this effect is CK independent or not (774). In turn, AMPK inhibits CK by phosphorylation in vitro and in differentiated muscle cells, and it also activates fatty acid oxidation. These findings suggest that CK, AMPK, and fatty acid oxidation form an intricate regulatory network for meeting energy supply with energy demands. In transgenic mice lacking both M-CK and sarcomeric Mi-CK, due to permanently high levels of PCr even during exercise, AMPK most likely remains inactive and, thus, cannot switch on fatty acid oxidation (774). In fact, these mice are defective in lipid metabolism and show signs of impaired capacity to utilize fatty acids (938). 4) Cr has been identified as an essential cofactor of thiamine-diphosphate (TDP) kinase from pig skeletal muscle, with half-maximal stimulation of enzymatic activity being observed at a [Cr] of 0.2 mM (881). In contrast, PCr, Crn, Arg, guanidinoacetic acid, and GPA had no effect on TDP kinase activity.

	VII. MICROBIAL CREATINE AND CREATININE DEGRADATION PATHWAYS

Top Previous Next References

In contrast to the nonenzymatic conversion of Cr and PCr to Crn in vertebrates, a growing number of microorganisms are being discovered to express specific enzymes for the degradation of Cr and Crn. Several lines of evidence suggest an involvement of microbial Cr and Crn degradation in vertebrate physiology and pathology. Bacteria and fungi capable of degrading Cr and Crn have been identified in chicken and pigeon droppings (278, 772), human urine (499) and feces (204, 992, 1029), as well as the bacterial flora of the human colon (204, 439). The latter bacteria may be particularly relevant to renal disease (see sect. IXH). In uremic patients in whom [Crn] in the serum is highly increased (163), Crn was suggested to diffuse into the intestinal tract where it induces bacterial creatininase, creatinase, and Crn deaminase activity, resulting ultimately in the breakdown of part of the body's Crn pool (439, 438) as well as in partial recycling of Cr (652).

In accordance with experiments on a variety of bacterial strains, 1-methylhydantoin produced by Crn deaminase is not further metabolized by the gut flora (439), but may, instead, be retaken up into the body and degraded there to 5-hydroxy-1-methylhydantoin, methylparabanic acid, N⁵-methyloxaluric acid, and oxalic acid plus methylurea (395). Because 1-methylhydantoin and 5-hydroxy-1-methylhydantoin were also detected in rabbit skin after vaccinia virus inoculation, a similar reaction cascade may proceed in inflamed tissue. Further microbial degradation products of Crn (e.g., methylguanidine) may act as uremic toxins (see sect. IXH), carcinogens, or carcinogen precursors (see sect. IXF). Finally, knowledge of the reactions and enzymes involved in Crn degradation may have an impact on routine clinical diagnosis where the Crn-degrading microbial enzymes may be used for specific enzymatic assays of [Crn] and [Cr] in serum and urine (see sect. X).

At least four alternative microbial Crn degradation pathways have to be considered (Fig. 7). 1) In some bacteria (Bacillus, Clostridium, Corynebacterium, Flavobacterium, Escherichia, Proteus, and Pseudomonas strains) and fungi (Cryptococcus neoformans and C. bacillisporus), Crn seems to be degraded solely to 1-methylhydantoin and ammonia (see Refs. 278, 484, 660, 772, 884, 895, 992). Crn can therefore be used by these microorganisms as a nitrogen source, but not as a carbon or energy source. In all microorganisms of this group that have been analyzed so far (Flavobacterium filamentosum, E. coli, Proteus mirabilis, and Pseudomonas chlororaphis), a single enzyme displays both cytosine deaminase and Crn deaminase activity (229, 484). The wide distribution of cytosine deaminases in microorganisms and the close structural similarity between cytosine and Crn may thus be the actual reasons why Crn deaminase activity, quasi as a side reaction, is also widely distributed. Although some of the Crn/cytosine deaminases are induced when the bacteria or fungi are grown on media containing Crn or cytosine (484, 772, 992), others are expressed in a constitutive manner or are even repressed by cytosine (484).

View larger version (22K): [in this window] [in a new window]   Fig. 7. Schematic representation of the reactions and enzymes involved in microbial Cr and Crn degradation pathways. The respective enzymes are denoted by numbers: 1) creatinine iminohydrolase (creatinine deaminase; EC 3.5.4.21); 2) cytosine aminohydrolase (cytosine deaminase; EC 3.5.4.1); 3) 1-methylhydantoin amidohydrolase [ATP dependent (EC 3.5.2.14) or non-ATP dependent]; 4) N-carbamoylsarcosine amidohydrolase (EC 3.5.1.59); 5) creatinine amidohydrolase (creatininase; EC 3.5.2.10); 6) creatine amidinohydrolase (creatinase; EC 3.5.3.3); 7) sarcosine reductase (EC 1.4.4.-); 8) not characterized so far; 9) methylguanidine amidinohydrolase (EC 3.5.3.16); 10) sarcosine oxidase (EC 1.5.3.1); 11) sarcosine dehydrogenase (EC 1.5.99.1) or dimethylglycine dehydrogenase (EC 1.5.99.2).

2) In several Pseudomonas, Brevibacterium, Moraxella, Micrococcus, and Arthrobacter strains, as well as in anaerobic Clostridium and Tissierella strains, 1-methylhydantoin is degraded further to N-carbamoylsarcosine and sarcosine. The enzymes involved in this degradation pathway, i.e., Crn deaminase, 1-methylhydantoin amidohydrolase, and N-carbamoylsarcosine amidohydrolase, are all highly induced when the bacteria are grown on Crn or 1-methylhydantoin as main source of nitrogen and, in some cases, carbon (see Refs. 170, 335, 357, 484, 714, 883, 884, 892). A comparison of the specific enzymatic activities revealed that the 1-methylhydantoin amidohydrolase reaction is the rate-limiting step of the pathway. Consequently, N-carbamoylsarcosine is in most instances either undetectable in these bacteria or is present in much lower concentration than the other intermediates (278, 884). Hydrolysis of 1-methylhydantoin, as catalyzed by the 1-methylhydantoin amidohydrolases of Pseudomonas, Brevibacterium, Moraxella, Micrococcus, and Arthrobacter strains, is stoichiometrically coupled with ATP hydrolysis and is stimulated by Mg²⁺ and NH₄⁺ or K⁺. In addition, hydantoin is hydrolyzed by these enzymes at a much lower rate than 1-methylhydantoin. In contrast, the 1-methylhydantoin amidohydrolases of anaerobic bacteria (357) are not affected by ATP and Mg²⁺, and hydantoin is hydrolyzed at a similar rate as 1-methylhydantoin.

3) In various Alcaligenes, Arthrobacter, Flavobacterium, Micrococcus, Pseudomonas, and Tissierella strains, still another set of enzymes is induced when they are grown on Cr or Crn as sole source of nitrogen and/or carbon. Creatininase (Crn amidohydrolase) converts Crn to Cr which is then further metabolized by creatinase (Cr amidinohydrolase) to urea and sarcosine (see Refs. 115, 257, 335, 487, 708, 884). Even though creatinase has also been detected in human skeletal muscle (655), this finding awaits confirmation and demonstration of its physiological relevance.

Sarcosine formed in degradation pathways 2 and 3 may be degraded further to Gly by a sarcosine oxidase or sarcosine dehydrogenase (487, 708, 884), or possibly to methylamine by the action of a sarcosine reductase (see Refs. 334, 335, 439). It also seems worth mentioning that glycocyamidine and glycocyamine (guanidinoacetate) can be degraded by microorganisms almost exactly as shown in pathway 3 for Crn degradation. Glycocyamidinase converts glycocyamidine to glycocyamine, which is then split by glycocyaminase (guanidinoacetate amidinohydrolase; EC 3.5.3.2) into Gly and urea (1150, 1151).

4) Finally, Pseudomonas stutzeri seems to convert Crn quantitatively to methylguanidine and acetic acid (1049). Methylguanidine was shown to be split in an Alcaligenes species by a highly specific methylguanidine amidinohydrolase into methylamine and urea (685).

The distinction between four alternative degradation pathways for Crn represents an oversimplification. For example, two of the degradation pathways may occur in the same organism, with the relative expression levels of the individual enzymes depending primarily on the nitrogen source used (884). When the Pseudomonas sp. 0114 is grown on Crn as main nitrogen source, Crn is degraded chiefly via Cr. When the same species is grown on 1-methylhydantoin, the 1-methylhydantoin amidohydrolase and N-carbamoylsarcosine amidohydrolase activities are induced so that in this case, Crn degradation via 1-methylhydantoin and N-carbamoylsarcosine prevails. The different Crn degradation pathways may also overlap. In the Pseudomonas sp. H21 grown on 1-methylhydantoin as main nitrogen source, creatinase activity is undetectable. However, Cr can still be degraded, but only indirectly via Crn, 1-methylhydantoin, and N-carbamoylsarcosine (884). When the same species is grown on Crn, creatinase is induced, and Cr can be degraded directly to sarcosine. Finally, the distinction between pathways 1 and 2 may seem arbitrary, even more so if it is taken into account that Clostridium putrefaciens and C. sordellii grown on basal medium degrade Cr and Crn solely to 1-methylhydantoin, while the same strains grown on a minced meat medium further degrade 1-methylhydantoin to sarcosine (278).

Clearly, microbial Crn degradation is at present only incompletely understood. To get a deeper insight into this topic, a wide evolutionary screening and detailed characterization of all enzymes involved are essential prerequisites. Relevant questions to be addressed are whether and how the expression of Cr- and Crn-degrading enzymes is regulated, and in which microorganisms Crn and cytosine deamination are catalyzed by a single or by separate enzymes.

	VIII. PROTEINS INVOLVED IN CREATINE METABOLISM

Top Previous Next References