Biosynthesis in Vitro of Homarine and Pyridine Carboxylic Acids in Marine Shrimp*

Minces and homogenates of muscle obtained from the marine shrimp Penaeus duornrum are capable of syn-thesizing homarine from [14C]glycine. Glycine carbon atoms are incorporated into homarine but not signifi-cantly into picolinate or quinolinate. [2-"C]Acetate is readily incorporated into quinolinate in the in vitro system but only slightly into homarine and not at all into picolinate. Quinolinic acid is rapidly methylated to N-methyl quinolinate which is not decarboxylated to form homarine. Procedures have been developed for the satisfactory separation of N-methyl quinolinate from homarine.

We have previously reported (1) that homarine (N-methyl picolinic acid) i s endogenously synthesized by the salt water shrimp Penaeus duorarum. After injection of a number of C-labeled substances and subsequent isolation of radioactive homarine, interpretation of the results was found to be sufflciently difficult to warrant efforts to develop an in vitro system capable of incorporating 14C-labeled precursors into homarine.
Homogenized shrimp tail muscle was found to be capable of converting [I4C]glycine into homarine, while labeled acetate was incorporated into quinolinic acid and not into homarine. When these I4C-labeled precursors were incubated in the presence of nonlabeled potential intermediates, it was found that two separate biosynthetic pathways were evident. 1) Picolinic and quinolinic acids are not intermediates in the conversion of glycine to homarine; 2) [2-I4C]acetate, on the contrary, is incorporated into quinolinic acid and not into homarine. The resulting quinolinic acid is readily methylated to form N-methyl quinolinic acid. 14

MATERIALS AND METHODS
In Vitro System for Homarine Biosynthesis-The shrimp, P. duorarum, were collected and maintained as previously described (1). After chilling on ice, the inert shrimp were shelled, and the muscle portion was finely minced and blended for 3 to 5 s in ice-cold citrate/ phosphate buffer (pH 7.4, 0.6 osmolar). Generally 20 ml of the buffer was used with 10 g of shrimp muscle. In typical experiments lo-'' mol each of ATP, DPN, FAD, and MgCly were added. Prior to incubation, i4C-labeled precursors were added, and, on occasion, nonradioactive potential intermediates (usually 5 mg of picolinic acid, quinolinic acid, or N-methyl quinolinic acid) which were subsequently isolated in addition to the homarine. At the end of the incubation (usually 6 h at 20 to 25"C), 5 mg more of nonradioactive carrier was added just prior to the work-up.
Fractionation and Purification-The isolation of purified hom-* This study was aided in part by a grant from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. arine has been previously described (1). Cell homogenates were deproteinized either 1) by addition of 10 or more volumes of methanol followed by overnight chilling, centrifugation, and evaporation of the aqueous methanolic supernatant fluid to a small volume or 2) by precipitation with 20% trichloroacetic acid. Following the methanol procedure, repeated vigorous shaking with a s m d volume of chloroform (0.1 to 0.2 volume) yielded a protein-free solution. All subsequent fractionations described here were done after deproteinization. Separation of Homarine and N-Methyl Quinolinic Acid-Following deproteinization by the methanol method (and shaking with chloroform) the concentrated aqueous preparation was evaporated in U~C U O to 2 to 3 ml, adjusted with a few drops of concentrated NH4OH to pH 10.9, and chromatographed on a column (1.5 X 20 cm) of Bio-Rad AG 1-X8 resin (OH-form, 100 to 200 mesh) previously equilibrated with 0.5% NH,OH. The column was eluted with the same solvent until homarine, N-methyl quinolinate, and other UV-absorbing quaternary nitrogen compounds were removed. After evaporation of NH3, the resulting syrup was adjusted to pH 4 to 5 with dilute HC1 and evaporated to a volume of 1 to 2 ml . The solution was then chromatographed on a column (1.5 X 20 cm) of Bio-Rad AG 50W-X8 resin (H+ form); the column was thoroughly washed with water and eluted with 0.01 N HCI (400 to 600 ml required). Both homarine and NMQ' are simultaneously eluted. Following evaporation of the eluate to 1 to 3 ml, both substances are co-precipitated by phosphotungstic acid in 1 N H2S04. The mixture, after removal of phosphotungstate, cannot be separated by thin layer chromatography with the solvent systems previously employed (1). Separation of the two substances can, however, be achieved by column chromatography on SP-Sephadex C-25 resin in 0.01 N HCI at pH 2. Homarine is readily eluted with 1 to 2 bed volumes of 0.01 N HCI, while NMQ can only be removed with 1 to 2 bed volumes of 0.1 M NaCl in 0.01 N HCI. Satisfactory separation of the two substances can also be obtained by thin layer chromatography (Analtech MN 300 microcrystalline cellulose) with isopropanol/water (85/15); RF for homarine is 0.41 and for NMQ is 0.23. It should be emphasized that trace contamination of the recovered homarine and NMQ by radioactive precursors was prevented by repeated additions of nonradioactive carrier precursors at suitable stages in the isolation procedure (prior to AG-50 column chromatography, phosphotungstic acid precipitation, and prior to final chromatography on SP-Sephadex columns). Highly radioactive products were checked several times by such washing-out procedures and rechromatographed for final 'C-labeled measurements.
Picolinic Acid-After incubation of [i4C]glycine or ['Tlacetate in the presence of carrier picolinic acid and deproteinization with trichloroacetic acid, the picolinic acid was precipitated along with homarine by phosphotungstic acid in 1 N H2S04 and chilled several hours; after Centrifugation, the precipitate was dissolved in dilute NaOH to pH 7 and decomposed with 10% Ba(0H);. solution. Barium ions were removed with dilute H2S04, and after evaporation of the aqueous fraction, the residue was extracted into a small volume of methanol to eliminate salts. The methanol solution was evaporated to dryness, and the residue was dissolved in 1 to 2 ml of 0.5% NH,OH (pH 10.9) and chromatographed on a column of Bio-Rad AG I-X8 resin as previously described. The column was eluted with 0.5% NH40H until homarine and other quaternary nitrogen compounds were completely removed. After washing with H20, picolinic acid was eluted with 1 to 4 bed volumes of 0.05 N HCI. Fractions were monitored by UV ' The abbreviation used is: NMQ, N-methyl quinolinic acid.
The picolinic acid fraction was evaporated to dryness and dissolved in minimal H20, the pH was adjusted to 2 to 3, and the solution was chromatographed on a column (1.5 X 20 cm) of Bio-Rad AG 50W-X8 resin (H+ form, 100 to 200 mesh). After subsequent washing with 250 ml of 0.01 N HCl, picolinic acid was eluted with 3 to 5 bed volumes of 1 N HCl. Following evaporation of the eluate, the residue was thoroughly dried and converted to the methyl ester by refluxing in methanolic HCl for 5 to 6 h. Solvent was removed by evaporation in vacuo, free base was liberated with NaHCOJ, and methyl picolinate was extracted into benzene. After drying over MgS04, the benzene was evaporated, and the residue was dissolved in 0.5 ml of methanol for chromatography on a Hewlett-Packard model 400 gas-liquid chromatograph equipped with a flame ionization detector (6-foot column of 3% JXR on 100/120 Gas-chrom Q; carrier flow rate, 60 ml/min; column temperature, llO°C; retention time, 162 s). The product was collected by means of an effluent splitter and trapped at -77"C, then dissolved in methanol for quantitative UV assay at 264 nm (c = 3200 M" cm" in CH30H). An aliquot was evaporated to dryness for radioassay in a Beckman LS 230 liquid scintillation counter.
Quinolinic Acid-Since quinolinic acid is not precipitated by phosphotungstic acid, this step was omitted. Chromatography on AG 1-X8 anion exchange resin in 0.5% NHIOH effectively removed homarine and other quaternary bases; the column was then successively After removal of the fonnic acid by evaporation in vacuo at 30-35°C. the residue was dissolved in 2 to 3 ml of dilute H2S04 at pH 3 and converted to the CU" salt by addition of powdered CuS04 with stirring; a precipitate of crystalline copper quinolinate formed within a few minutes. After centrifugation, the precipitate was washed once with a small volume of dilute H804 (pH 3) and decomposed with dilute NaOH. The resulting slightly soluble CU(OH)Z was removed by centrifugation, the aqueous phase was acidified to pH 3 with dilute HzSOI and evaporated in vacuo to a syrup, and Na2S04 was precipitated with methanol. The methanol-soluble fraction was evaporated, dried at 1 to 2 mm Hg, and converted to the dimethyl ester as described for picolinic acid. The dimethyl quinolinate was subjected to gas-liquid chromatography as above (2) and collected by means of an effluent splitter (carrier flow rate, 60 ml/min; column temperature, 145OC; retention time, 105 s). The dimethyl quinolinate was dissolved in methanol for quantitative UV assay at 264 nm (e = 2450 M" cm" in CHaOH), and a measured sample was evaporated for radioassay.
N-Methyl Quinolinic Acid-Synthetic NMQ was prepared by reaction of quinolinic acid with methyl iodide. Quinolinic acid in 50% aqueous methanol was neutralized to pH 7.5, treated with a 5-to 10fold excess of methyl iodide in a chilled pressure tube, and heated at 100-105°C for 24 h. After evaporation to dryness the residue was dissolved in minimal 0.5% N R O H and passed through a column of AG-1 resin at pH 10.9 to remove unreacted quinolinic acid (which is retained by the column). After evaporation of NHs and pH adjustment to 4 to 5 with dilute HCl, the solution was evaporated to dryness and recrystallized from ethanol/l-butanol (4/1). The UV spectrum showed a maximum at 264 nm (e = 4500 M" cm" for C1-salt in H20) and a minimum at 244 to 246 nm. On thin layer chromatography only a single spot was obtained.

RESULTS
A series of in uitro incubations with [2-I4C]glycine as well as [l-14C]glycine demonstrated that radioactive homarine is biosynthesized by such preparations. Although the radioactive yield is small, nevertheless, consistent and far greater radioactivity was obtained with glycine than with any of the other amino acids listed (Table I). It should additionally be pointed out that free glycine is present in large amounts (approximately 45 mg/lO g of shrimp tissue). The radioactive glycine is, therefore, diluted to an enormous extent by endogenous free glycine. When carrier picolinate or quinolinate was added prior to incubation with [14C]glycine, little or no radioactivity was recovered in the carrier substances while the isolated homarine remained radioactive. It is, therefore, clear that    glycine carbons are incorporated into homarine without intermediate formation of either picolinate or quinolinate. In a preliminary experiment we have observed that dipicolinic acid (2,6-pyridine dicarboxylic acid) is likewise not an intermediate.
The exclusion of picolinate as an intermediate suggested that glycine might be methylated early to form sarcosine which could then be incorporated into homarine.
[ l-14C]Sarcosine, upon incubation, was converted to homarine (Table I) but only half as efficiently as glycine. Inasmuch as free endogenous glycine is present in very high concentrations in shrimp (as noted above), labeled sarcosine was diluted to approximately the Same extent with 45 mg of carrier sarcosine prior to incubation. The low radioactivity obtained in the recovered homarine suggests that sarcosine is converted to glycine prior to incorporation into homarine.
In an earlier report (1) labeled acetate injected into live shrimp appeared to be incorporated into homarine; when incubated in uitro, however, little or no [2-'4C]acetate was converted to homarine (Table 11). Significant activity was recovered in carrier quinolinate (none in picolinate). In our earlier in uiuo experiments (1) labeled quinolinate also appeared to be a good precursor of homarine. Again, upon incubation of [6-14C]quinolinate with shrimp homogenate, relatively little radioactivity was recovered in the isolated homarine (Table 11). These results suggested that our earlier preparations of radioactive homarine derived from acetate or quinolinate were contaminated with traces of a highly radioactive substance. On the theory that quinolinic acid might undergo N-methylation to form N-methyl quinolinate during in vitro incubations, NMQ was chemically synthesized by methylation of quinolinate with methyl iodide. The resulting compound was found to be very similar in its properties to homarine and the two substances separable with diffkulty. Radioactive preparations of homarine derived from acetate or quinolinate lost most of their radioactivity upon additional chromatography and elution on SP-Sephadex in 0.01 N HC1, while the NMQ could be subsequently eluted only with 0.1 M NaCl in 0.01 N HC1. As a result of this additional final purification step, the data listed in Table I1 help to clarify the phenomena under study. 1) Acetate is converted to quinolinate which is readily methylated to NMQ. 2) Labeled tryptophan is converted to quinolinate which is in turn methylated. 3) Labeled glycerol, glutamate, and probably aspartate are incorporated into quinolinate. None of the substances listed in Table I1 appear to be converted to any significant extent into homarine.

DISCUSSION
It is apparent from this work that two separate pathways have been established. 1) Acetate carbon atoms are incorporated into quinolinic acid which is subsequently methylated to NMQ. 2) Conversion of glycine to homarine occurs by a pathway which does not involve picolinate or quinolinate as intermediates. Whether endogenous synthesis of quinolinate (from acetate) by shrimp provides a significant source of nicotinic acid has not been established; it is possible that low or moderate levels of nicotinate may be related to rapid methylation of quinolinate to form NMQ which is then decarboxylated to trigonelline.
Since picolinate is not an intermediate in the biosynthesis of homarine, it may be postulated that glycine condenses with a suitable 4-carbon compound to yield a di-or tetrahydropyridine carboxylic acid which is subsequently converted to homarine. We have unsuccessfully investigated the possible condensation of glycine with succinic monoaldehyde, hoping that Schiff base formation followed by an aldol type cyclization would yield an important precursor of homarine. Efforts will be continued to find more effective intermediates in this metabolic pathway.
It is recognized that quinolinate is formed by several pathways: 1) from tryptophan in mammals, yeast, and Neurospora (4); 2) from aspartate, acetate, and formate in Clostridium butylicum (5); and 3) from aspartate and glycerol (or a closely related intermediate) in higher plants (6) and various bacteria (7,8). It would appear that the latter mechanism is similar to what we have observed in shrimp.