Role of Lysine and ε-N-Trimethyllysine in Carnitine Biosynthesis

A lysine-carnitine precursor-product relationship was envisaged. To test this hypothesis a series of isotopic labeling experiments was conducted with Neurospora crassa lysine auxotrophs. Following growth of such mutants on a simple defined medium lacking carnitine but supplemented with variously labeled test carnitine precursors, biosynthesized carnitine was subsequently isolated from the mold mycelium by ion exchange chromatography and examined for radioactivity. Radioactivity from DL-[6-14C]lysineandDL-[4,5-3H]lysine was incorporated into carnitine without dilution of specific activity, whereas radioactivity from DL-[lJ4C]lysine and DL-[2-WIlysine was not found in biosynthesized carnitine in significant amounts. In an experiment employing [+lSN]lysine, it was demonstrated that the c-N atom of lysine becomes the nitrogen atom of carnitine; there was some exchange of the [Ei5N]amino group of lysine with 14N in the medium during carnitine synthesis. e-N-[methyl-3H]Trimethyl-L-lysine was incorporated in high yield (16 and 19%) into carnitine; and, as expected [mefhyZ-3H]methionine also labeled carnitine, but with markedly less efficiency (0.23 %) consistent with dilution by endogenous methionine. [ l-14C] y -Butyrobetaine (4 Ntrimethylamino [lr4C] butyric acid), but not (4-amino[l-i4C]butyric acid), was readily utilized (9.4%) for carnitine formation, in agreement with its established role as a carnitine precursor in animals. From these and other considerations it is postulated that in Neurospora the biogenesis of carnitine derives from the amino acids lysine and methionine in which either free or bound lysine is successively methylated to give e-N-trimethyllysine which is then cleaved in a series of as yet unknown transformations to lose carbon atoms 1 and 2 yielding y-butyrobetaine. The latter compound is then hydroxylated at the /3 position to give carnitine as previously shown by other workers in the field.

and, as expected [mefhyZ-3H]methionine also labeled carnitine, but with markedly less efficiency (0.23 %) consistent with dilution by endogenous methionine. [ l-14C] y -Butyrobetaine (4 -N-trimethylamino [l-r4C] butyric acid), but not (4-amino[l-i4C]butyric acid), was readily utilized (9.4%) for carnitine formation, in agreement with its established role as a carnitine precursor in animals. From these and other considerations it is postulated that in Neurospora the biogenesis of carnitine derives from the amino acids lysine and methionine in which either free or bound lysine is successively methylated to give e-N-trimethyllysine which is then cleaved in a series of as yet unknown transformations to lose carbon atoms 1 and 2 yielding y-butyrobetaine.
The latter compound is then hydroxylated at the /3 position to give carnitine as previously shown by other workers in the field. * This work was supported by Health Science Advancement Award NIH 5 SO4 RR 06067, National Institutes of Health Grants 5 TO1 AM05441 and 5 ROl AM 14338, and was taken from a portion of the dissertation of I). W. H. submitted to the Graduate College of Vanderbilt University in partial fulfillment of the requirements for t,he degree of Doctor of Philosophy.
Although it has been well established that y-butyrobetaine is a precursor of carnitine in animals (l-3) and that methionine is the penultimate source of the N-methyl groups (4, 5), the biosynthetic origin of the butyrate carbon chain and y-nitrogen atom of y-butyrobetaine has been obscure. e-N-Trimethyllysine and d-hydroxy-e-N-trimethyllysine have recently been discovered in diverse sources in nature (cf. Ref. 6 for references). Our attention was drawn to the structural features these compounds have in common with y-butyrobetaine and carnitine, respectively, and led us to consider the possibility that y-butyrobetaine, and consequently carnitine, might derive from lysine metabolism.
To test the possibility of a lysine-carnitine precursor-product relationship, Neurospora crassa lysine auxotrophs were selected for study as such mutants could be grown in a carnitine-free synthetic medium with appropriately labeled test lysines. If the lysine-carnitine precursor-product relationship were true, the biosynthesized carnitine should have the same specific activity as the initial proffered isotopic lysine.
From experiments employing isotope incorporation and isotope dilution techniques, evidence will be presented that lysine, e-Ntrimethyllysine, and y-butyrobetaine are precursors of carnitine in Neurospora.
The data support the conclusion that the entire lysine molecule, save C-l, C-2, and the cr-N atom, is involved in the lysine ---) carnitine transformations.
Brief accounts of certain of these studies have appeared (7-9).  Mazzetti and Lemmon (14). Both compounds were radiochemically pure as judged by thin layer chromatography as described previously (7) (15) using an LKB combined gas chromatograph, mass spectrometer employing a g-foot column of 3% OV-17 on Gas-chrom Q at 230", and a helium flow rate of about 10 ml per min. The ratio of the peaks at m/e 31 and m/e 30 (CHzNHz)+, corresponding to the e-methylene and e-amino moieties of lysine, was used to calculate the 15N enrichment of the synthetic [e-15N]lysine. This analysis showed that the product contained 89.6% atom excess 15N.
In an experiment to be described (Table IV) in which N. crassa 15069 was grown in medium containing [e-15N]lysine, the mycelial proteins were subsequently hydrolyzed to release bound lysine which was isolated, converted to the ethyl ester, and analyzed for 15N enrichment as described above for synthetic [c-15N]lysine. In this instance the ratio of the peaks at m/e 31 and 30 was first used to calculate the enrichment of the E-N atom of the bound lysine. The ratio of peaks at m/e 102 and 101 (corresponding to loss of the carbethoxy moiety of lysine ethyl ester) and m/e 175 and 174 (the molecular ion) were then used to calculate the total 15N content of the lysine and by subtraction the 15N enrichment of the cr-N atom of the lysine could be determined. The isolation and purification of bound [15N]lysine is given in the legend of Table IV. Neurospora Cultures, Source, and Maintenance-Cultures of N. crassa lysine auxotrophs strains 33933 and 15069 (Fungal Genetics Stock Center, Dartmouth College) were maintained on 2% agar slants (Difco Neurospora culture medium) and transferred monthly.
Mutant 33933 is blocked at an early step of the homocitrate-aminoadipate pathway of lysine biosynthesis (16) ; mutant 15069 lacks saccharopine dehydrogenase (17) the final enzyme of this pathway. Isolation of Carnitine from Neurospora Cultures, Assay Procedures-The experimental plan and specific procedures followed in much of this study are given in the legend of Fig. 1 in Reference 7, and in summary encompass the following steps: (a) growth of N. crassa cells in carnitine-free medium supplemented with lysine and variously labeled test precursors, (b) subsequent harvesting of the mycelium, (c) extraction of biosynthesized carnitine and hydrolysis of the 0-acyl carnitine esters, (d) isolation of carnitine by cation exchange chromatography (18), and (e) determination of the radioactivity and carnitine content of appropriate column eluates. Radioactivity of these fractions, Step e, was measured by removing 0.1.ml aliquots for liquid scintillation counting in a Packard Tri-Carb liquid scintillation spectrometer with the mixture employed previously (7). The efficiency of individual samples was determined by the external standard channels ratio method using secondary standards of toluene (New England Nuclear) quenched with chloroform. Carnitine was determined by carnitine acetyltransferase assay (19). From these data, the per cent incorporation of a given test isotope into biosynthesized carnitine could be made, and the specific activity of such carnitine could be determined.  Table I summarizes a series of experiments in which the degree of utilization of various isotopically labeled lysines by N. crassa lysine auxotrophs for carnitine biosynthesis was studied.
We have previously reported (7, 8) that nL-[6-14C]lysine, but not nL-[I-14C]lysine, was incorporated into carnitine by N. crassa strain 33933. The coincidence or absence of radioactivity in carnitine fractions following isolation of carnitine by ion exchange chromatography in these instances were given in Figs. 1 and 3 of Ref. 7 and are summarized as Experiments 1 and 4 of Table I herein.
Two additional trials with nL-[6-14C]lysine were conducted (Experiments 2 and 3, Table I). The specific activity data in all three experiments illustrate that a moiety of lysine which includes C-6 is incorporated into carnitine without dilution. Experiment 3, Table I, which involved N. crassa mutant 15069 is worthy of note as this mutant lacks saccharopine dehydrogenase.
Thus, intermediates of the homocitrate-aminoadipate pathway of lysine biosynthesis that might possibly arise  Table I). It is possible that, in the lysine catabolism of N. crassa, C-l and C-2 of the lysine molecule may ultimately contribute appreciably to the l-carbon pool which in turn would be expected to label the N-methyl groups of carnitine. Fig. 1A illustrates the coincidence of radioactivity with carnitine fractions following ion exchange chromatography of mycelial extracts of N. crassa 33933 grown with DL-[4,5-3H]lysine. This experiment was repeated and the data from these trials, together with an experiment with L-[G-3H]lysine, are all summarized in Table I (Experiments 6 to 8). As indicated in Footnote b of Table I, the "anticipated specific activity" of biosynthesized carnitine was calculated on the postulation that C-3, C-4, C-5, and C-6 of lysine indeed contribute the butyrate carbon chain of carnitine.
Such a postulation demands the loss of 2 hydrogen atoms from C-3 of lysine and loss of 1 hydrogen atom from C-5 of lysine.
The observed specific activities of the biosynthesized carnitine in these instances (Experiments 6 to 8, Table I) are consistent with these assumptions, indeed the values are even slightly higher than anticipated.
These differences may nob be significant or, if so, could imply an isotope effect. In any event the utilization of UL-[4, 5-3H]lysine in carnitine biosynthesis provides important evidence for the incorporation of C-4 and C-5 of lysine into carnitine, since the pattern of retention of the tritium label implies the participation of the respective carbon atoms in carnitine formation.
Unfortunately lysine isotopically labeled at C-3 was unavailable for testing as a carnitine precursor in these studies. c-N-Trimelhyllysine and y-Butyrobetaine as Precursors of Car-r&&e-Evidence that these compounds are involved in carnitine biosynthesis in Neurospora was first obtained from isotope dilution experiments summarized in Table II. Thus when N. crassa 33933 was grown with no-[4,5-3H]lysine or L-[G-3H]lysine but in the presence of substrate quantities of e-N-trimethyllysine or y-butyrobetaine, respectively, the incorporation of radioactivity from lysine into biosynthesized carnitine in these instances was markedly repressed (93 and 83%, respectively, Table II).
Neither t-N-trimethyllysine nor y-butyrobetaine interfered with the lysine metabolism of the mold, as these compounds had no effect on growth. Indeed a point of interest is that the yield of biosynthesized carnitine in the presence of either of these conpounds was enhanced, Table II. Direct evidence for the participation of e-N-trimethyllysine and y-butyrobetaine in carnitine synthesis was obtained in a series of labeling experiments summarized in Table III. there was an exact coincidence between column effluents containing carnitine with radioactivity derived from t-Ntrimethyllysine.
Sixteen per cent of the radioactivity of trimethyllysine was found in the isolated carnitine (Experiment 1, Table III).
The specific activity of the biosynthesized carnitine was reduced by only a third from the specific activity of the profferred ~-N-[methyZ-31-I]trimethyl-L-lysine, such dilution likely being due to competition from unlabeled lysine in the medium as a substrate for carnitine synthesis.
This experiment was repeated with very similar results as shown in Experiment 2, 2173 a unique lysine derivative with more limited functions in metabolism, one of which is to serve directly in carnitine biosynthesis.
It seemed likely that the origin of the N-methyl groups of e-N-trimethyllysine for carnitine biosynthesis would be methionine, particularly since it has been shown that the penultimate source of the N-methyl groups of carnitine in the rat is methionine (4, 5). The data of Fig. 2A show that radioactivity derived from the methyl group of methionine is associated with the carnitine fraction, but the extent of incorporation of radioactivity in this instance was only 0.23% (Experiment 3, Table III). Furthermore, the specific activity of the [methylPH]methionine was reduced over a 100-fold in the biosynthesized carnitine (Tam ble III).
This finding undoubtedly reflects the demand for the methyl group of methionine for diverse reactions of l-carbon metabolism and that the methyl group of methionine is constantly being resynthesized de novo from nonradioactive l-carbon precursors.
In any event, it is clear that the incorporation of E-N-[methyL3H]trimethyllysine into carnitine (Experiments 1 and 2, Table III), is direct and does not involve demethylation and subsequent reutilization of the methyl groups via methionine by some other biosynthetic mechanism for carnitine synthesis.
The data of Experiments 1 to 3 of Table III taken together suggest that methionine is the penultimate source of the methyl groups of carnitine in Neurospora, but that e-N-trimethyllysine is the direct metabolite concerned in carnitine biosynthesis. Fig. 3  In this instance, the reduction of specific activity of [PC]ybutyrobetaine was only about 50'$&, and the incorporation of radioactivity was 9.4% (Experiment 4, Table III). These data thus illustrate that y-butyrobetaine is an excellent precursor of carnitiue in Neurospora, in accord with the findings in animals that y-butyrobetaine (2, 3), but not y-aminobutyrate (I), is a precursor of carnitine, and make untenable the early postulation that y-butyrobetaine rnight arise via successive methylation of y-aminobutyrate (20). These data are consistent rather with the view that y-butyrobetaine derives from c-N-trimethyllysine metabolism.

Origin of Nitrogen
Atom of Carnitine-Based on the efficient incorporation of e-N-trimethyllysine into carnitine (Experiments 1 and 2, Table III) coupled with the pattern of lysine carbon atoms utilized for carnitine biosynthesis (Table I), it was attractive to speculate that the e-N-atom of either free or bound lysine is successively methylated yielding c-N-trimethyllysine and that the nitrogen atom of carnitine thus initially derives from the E-N atom of lysine.
To test this hypothesis [@N]lysine containing 89.6y0 '"S atorn excess JTas prepared (cj+. "Experimental Procedure") diluted to 44.8oj, IjS atom excess with [14N]lysinc, and N. crassa strain 15069 grown on carnitine-free medium containing such lysine. Mutant 15069 lacking saccharopine dehydrogenase was selected for this experiment to eliminate the possibility of lysine precursors figuring in the interpretation of the results. Carnitine was subsequently isolated from the mold mycelium and analyzed for 15N content as described in the legend of Table IV. The cxpcrirncnt depends on the alkaline degradation of carnitine to trimethylamine and analysis of the latter by combined gasliquid chromatography-mass spcctromctry. The TO-e.v. spectrum of trimethylamine is published (21). The base peak at nz/e 58 (CJI&)+, corresponds to loss of one hydrogen from the molecular ion at nr/e 59. Since the large peak at d1-1 would interfere with the calculation of the '"N enrichment, the spectra were taken at 13 e.v. to minimize the 11/-l base peak. The ratio of the peaks 60:59 T\-as used to calculate a 20.3% enrichment in W in the trirnethylamine and thus in the biosynthesizcd carnitine (Table IV).
Since [6-W]lysinc was incorporated into carrritine without dilution (Table I), it was surprisin g that the E-N atom of lysine should be diluted in contributing to carnitine biosynthesis. It seemed important to establish the 16N content of the lysine of the cell proteins to serve as a control for the experiment.
Accordingly, the N. crassa mycelium was hydrolyzed to release bound lysine which was isolated and analyzed for 15x enrichment as described ill Table IV and under "Experimental Procedure." As is showr~ in Table IV, no significant dilution of 16X into lysine of the mycelial proteins occurred by exogenous 14N of the medium.
Hence it is concluded that in the lysine carnitine transformations a unique lysine intermediate exists which is in an environment such that the 6-K atom can readily cschange with nitrogen sources in the medium.
This latter point will be more fully discussed below. DISCUSSlOX Hy using techniques of isotope incorporation and isotope dilution, it has been shown that a moiety of lysine is involved in the biosynthesis of carnitine in A'. crassa lysine ausotrophs. About 0.2% of the lysine requirement of these mutants is demanded for the synthesis of carnitine which is presurnably only needed in catalytic quantities by the cell. From the labeling patterns presented it seems reasonable that C-3, C-4, C-5, and C-6 together with S-6 of lysine constitute the butyric acid carbon chain b Analyzed following degradation to trimethylarnine. and y-Ii atorn of carnitine, respectively, as sho~~-vn in Fig. 4. Fig. 4 visualizes that either free or protein-bound lysine is SUCcessively methylated, yielding free or protein bound e-S-trimethyllysine.
If e-N-trimethyllysine is indeed forrned from protein-bound lyaine, as for example by an appropriate S-adenosyl-L-methionine : proteirl-lysine methyltransferase (cf. Ref. 6 for refercnces), such bound trimethyllysine may then be released to give free c-N-trimethyllysine.
The latter is then subject to a series of reactions referred to as cleavage (Fig. 4), to yield y-butyrobetainc which is then hydrosylated, finally giving carnitine. The possibility that stelx of carnitine biosynthesis involve bound lysine intermediates may figure in an explanation for the data of Table IV in which in the course of the lysine carnitine transformations the e-r\; atom of lysine exchanged with nitrogen sources in the rnediurn.
It is known that in the forrnation of desmosine a polyfunctional amino acid involved in cross-linking in elastin (22) It is clear that the apparent exchange of the e-N atom of lysine with nitrogen sources in the cell is limited in Neurospora growth and metabolism, as [GS]lysine of the cell proteins in general was not significantly exchanged with 14N sources in the cell (Line 3, Table IV).
It should be pointed out that in these studies the Neurospora cells were harvested 1 to 2 days after maximum growth was obtained to assure maximal carnitine synthesis.
Hence, in this latter period, an opportunity is still being provided for exchange of W into carnitine but not into the mycelial proteins which could account in part for the findings of Table IV. The nature of the postulated "cleavage reaction" (Fig. 4), wherein e-N-trimethyllysine is metabolized to y-butyrobetaine, forms the basis of on-going research in our laboratory.
No evidence for a strict precursor-product relationship between E-Ntrimethyllysine and y-butyrobetaine was presented herein, although, from the circumstances presented, this inference seems very likely.
In this respect, radioactive y-butyrobetaine has been isolated from rat liver and rat urine of lysine-deficient rats administered e-N-[methyZ-3H]trimethyl-lr,-lysine as described in an accompanying paper (23). An attractive speculation for the postulated e-N-trimethyllysine --j y-butyrobetaine transformations (Fig. 4) is drawn from studies of lysine catabolism in Clostridium sticklandii, wherein C-3, C-4, C-5, and C-6 of lysine give rise to butyrate, and C-l and C-2 yield acetate (24). By analogy to certain of these catabolic reactions, if the a-amino group of e-N-trimethyllysine migrated to the p position and then is oxidatively deaminated to a P-keto acid, cleavage of the latter acid would yield y-butyrobetaine and acetate. In preliminary studies, the hydroxylation of y-butyrobetaine to carnitine in a cell-free extract of N. crassa 33933 was shown (25). The system had the same requirements as in the rat liver hydroxylase system as recently described by Lindstedt et al. (26), namely oxygen, oc-ketoglutarate, Fe++, and ascorbate. Further evidence for a commonality in steps of carnitine biogenesis between the fungal system described herein and the rat n-ill be apparent from an accompanying paper (23) which establishes that lysine and e-A-trimethyllysine arc precursors of carnitine in the rat as well.