It May, As Found for Hydroxylation of Proline

VOL. 55, 1966 BIOCHEMISTRY: POPENOE ET AL. 393

E. H. Lincoln, and J. A. White, J. Am. Chem. Soc., 79, 1682 (1957); Moffett, R. B., B. D. Tiffany, B. D. Aspergen, and R. V. Heinzelman, J. Am. Chem. Soc., 79, 1687 (1957); Bahner, C. T., L. R. Barclay, G. Biggerstaff, D. E. Bilanco, G. W. Blanc, M. Close, and M. M. Isenberg, J. Am. Chem. Soc., 75, 4838 (1953); Wright, T. B., E. H. Lincoln, and R. V. Heinzelman, J. Am. Chem. Soc., 79, 1690 (1957). 6French, F. A., and B. L. Freedlander, Cancer Res., 18, 172 (1958); French, F. A., B. L. Freed- lander, A. Hosking, and J. French, Acta, Unio Intern. Contra Cancrum, 15, 614 (1960); French, F. A., and B. L. Freedlander, Cancer Res., 18, 1290 (1958). 7Friedemann, T. E., J. Biochem., 73, 33 (1927). 8 Klotzsch, H., in Methods of Enzymatic Analysis, ed. H. U. Bergmeyer (New York: Academic Press, 1965), p. 283. 9 Dr. S. E. Luria's strain B/1, 5, phage T, T5-resistant (MIT, Boston). 10 Roberts, R. B., P. H. Abelson, D. B. Cowie, E. T. Bolton, and B. J. Britten, Studies of Bio- synthesis of E. coli (Carnegie Institution of Washington, Publ. No. 607, 1957), p. 5. " Similar results were mentioned by First, A., J. Miller, A. Gross, and W. Cutting, Acta, Unio Intern. Contra Cancrum, 16, 625 (1960), that Kethoxal inhibited cell division of E. coli (strain K-12), and that glutamic acid antagonized inhibition. Glutamic acid similarly antagonized the action of Kethoxal-bisulphite complex in tumor-bearing animal experiments. 12Egyftd, L. G., to be published elsewhere.

HYDROXYLYSINE FORMA TION FROM LYSINE DURING COLLAGEN BIOSYNTHESIS* BY EDWIN A. POPENOE, ROBERT B. ARONSON, AND DONALD D. VAN SLYKE

MEDICAL RESEARCH CENTER, BROOKHAVEN NATIONAL LABORATORY, UPTON, NEW YORK Communicated December 22, 1965 Hydroxylysine as a constituent amino acid of a protein has been found only in collagen.' The obligatory source of collagen hydroxylysine is lysine. 2-4 Hydroxy- lysine itself, either fed or injected intraperitoneally, is not incorporated into collagen by rats,5 or presumably by other animals. Hydroxylation of lysine occurs in the early stages of collagen synthesis4' 6; it may, as found for hydroxylation of proline, occur in the precollagen peptide chain on the ribosomes.7 The present paper reports a continuation of a study of the biochemical mechanism by which a hydroxyl group is attached to the number 5 carbon of lysine. Con- ceivably the hydroxylation could proceed by (a) direct uptake of an atom of oxygen on carbon 5; (b) oxidation of carbon 5 to a ketone -CO- group followed by reduction to -CH(OH)-; (c) dehydrogenation to form a 4-5 double bond followed by addition of the elements of water; (d) dehydrogenation to form a 5-6 double bond followed by addition of the elements of water. We have previously reported the results of studies using lysine-4-5-H3 as a precursor of collagen hydroxylysine.8 In these studies we found that only one of the four labeled H atoms of lysine-4-5-H3 was lost in the conversion to hydroxylysine. One hydrogen atom on carbon 5 was replaced by -OH; therefore none of the other three labeled H atoms could be involved in the hydroxylation in such a way that they became equilibrated with the H of the medium. The results eliminated 5- keto-lysine and 4,5-dehydrolysine (mechanisms b and c) as possible intermediates in the hydroxylation of lysine. Downloaded by guest on September 25, 2021 394 BIOCHEMISTRY: POPENOE ET AL. PROC. N. A. S.

Ill the present experiments we have used lysine-6-H3 as a precursor of collagen hydroxylysine, and have found that the conversion to hydroxylysine occurs without the loss of any of the tritium from carbon 6. Mechanism d thus appears to be eliminated. Materials and Methods.-Alpha-aminoadipic acid was purchased from California Corp. for Bio- chemical Research. D-Glucose-l-H3, uniformly labeled i-lysine-C14, and standards of C14- and H3-toluene were obtained from New England Nuclear Corp. 2,5-Diphenyloxazole (PPO) and 1,4-bis-2-(4-methyl-5-phenyloxazolyl) benzene (POPOP) scintillators were from Packard Instru- ment Co. New England Nuclear Corp. prepared Dilysine-6-H3 chemically by catalytic hydrogenation of 5-cyano-2-amino-pentanoic acid (personal communication, Dr. Thomas F. Sullivan, New England Nuclear Corp.). The radiochemical purity of the lysine was confirmed by paper chromatography. Biosynthetic preparation of L-lysine-6-H3: The 6-tritiated lysine was prepared by the action of baker's yeast (Saccharomyces cerevisiae) on alpha-aminoadipic acid in the presence of D-glucose-1-H3 by a procedure derived from the work of Broquist9 and his collaborators. A two-step reaction converts -COOH of the adipic acid to -CH2NH2. In both the initial and final steps, reduction is mediated by DPNH, although in crude systems TPNH can be substituted for DPNH. For our preparation, into each of two 50-ml beakers were placed 2.5 me (0.8 mg) of D-gluCose-l-H3, 3.4 mg of alpha-aminoadipic acid, 0.5 ml of phosphate buffer (1 M KH2PO4 adjusted to pH 3.5 with H3PO4), and a suspension of 125 mg (moist) of baker's yeast. The total volume in each beaker was 5 ml. The beakers were shaken vigorously in a water bath at 300 for 3 hr. The cells were then collected by centrifugation and heated at 1150 in sealed tubes with 0.2 N HCl (4 ml) for 10 min. Lysine was isolated from the 0.2 N HCl extract by chromatography on Amberlite IR-120 ion exchange resin as described below for isolation of the amino acids from the metabolic experiments. Paper chromatography (butanol, acetic acid, water, 5:1:4, upper layer) showed that the product was contaminated by a very small amount of slower-moving, ninhydrin-positive material. However, all detectable radioactivity, as measured on a paper strip scanner, migrated with the lysine. The yield of radioactive lysine was 10 liC. Determination of radioactivity: All counting was done in a Tri-Carb liquid scintillation spec- trometer, model 4322. The instrument permits simultaneous counting of C14 and H3 in the same sample with good accuracy if the ratio H3/C"4 does not deviate too much from the optimal range. Activities of the two isotopes, when they were determined in the same sample, were calculated by the discriminator-ratio method described by Kabara et al.10 Counting efficiencies were calculated from the count rates of standards of C'4 and H3 toluene. In determining the H3 activities of the products of degradation of the lysine-6-H3 preparations (Tables 1 and 2), weighed samples were dissolved in 0.1 ml of water to which was added 5 ml of a 2:1 (v/v) mixture of ethylene glycol monomethyl ether and ethanolamine, and 10 ml of toluene containing 15 gm of PPO and 50 mg of POPOP per liter. In determining the H3/C"4 ratios of lysine and hydroxylysine from the hydroxylation experiments (Tables 3 and 4), the amino acids, after ion-exchange separation and desalting, were brought to 1 ml vol, and aliquots of 0.1 ml were handled in the same way described above, except that the scintillation mixture contained 5 ml of ethylene glycol monomethyl ether instead of 5 ml of the 2: 1 mixture of ethylene glycol monomethyl ether and ethanolamine. Degradation of the lysine-6-H3 preparations to confirm the location of the tritium: A sample of the chemically prepared Dilysine-6-H3 was mixed with nonradioactive Di-lysine, and a sample of the biosynthetic zlysine-6-H3 was mixed with nonradioactive ilysine. Each mixture was crystallized to constant specific activity and degraded by the procedures of Strassman and Wein- house.'1 In each degradation, one sample of lysine was oxidized directly with hot acidic permanganate to glutaric acid. Another sample was converted by more gentle permanganate oxidation to delta-aminovaleric acid. This was then. converted by a Schmidt reaction to 1,4-diamino- butane followed by alkaline permanganate oxidation to succinic acid. The reactions were carried out exactly as described by Strassman and Weinhouse," except that the succinic acid was puri- fied by crystallization from water rather than by way of the silver salt. Melting points of the final products agreed satisfactorily with known values. The purity of the delta-aminovaleric acid was Downloaded by guest on September 25, 2021 VOL. 55, 1966 BIOCHEMISTRY: POPENOE ET AL. 395

further checked on the amino acid analyzer and the purity of the two dicarboxylic acids by thii layer chromatography.12 Incorporation of labeled lysine into the collagen of rat sponge biopsy tissues: The procedures followed were those of Kao and Botucek,13 and have been described in detail previously.8 In the present experiments, either the chemically prepared D-lysine-6-H' or the biosynthetic ilysine-6-H3 was mixed with an appropriate amount of uniformly labeled L-lysine-C4 and incu- bated with the sliced sponge biopsy tissue for 12 hr in Krebs-Henseleit solution under an atmos- phere of 95% 02 and 5% Co2. After the incubation, from four to six sponges were pooled for each experiment. Collagen was extracted from the tissues with hot trichloroacetic acid14 and hydro- lyzed. Lysine and hydroxylysine were isolated by chromatography on a 20 X 0.9-cm column of Amberlite IR-120 ion exchange resin with pH 4.26, 0.38 N buffer" at 300. The fractions containing lysine or hydroxylysine were pooled and desalted on a column of Dowex-50 of similar size, as in a previous report.8 Results.-Location of tritium in the lysine-6-H3 preparations: The results of the degradations are given in Tables 1 and 2. Table 1 shows that in the chemically prepared tritiated lysine approximately 76 per cent of the tritium was attached to carbon 6, 24 per cent either to carbon 4 or carbon 5, or divided between them. The degradation experiments cannot distinguish between attachment of tritium to carbon 4 and attachment to carbon 5, but carbon 5 seems more likely to become labeled during exposure of the 5-cyano-1-amino-pentanoic acid to tritium gas because of the strong electron-withdrawing effect of the adjacent nitrile group. The effect of the measured distribution of tritium in this lysine preparation upon the results of the biological hydroxylation experiments is considered below. Table 2 indicates that in the lysine labeled biosynthetically, all of the tritium was located on carbon 6. Incorporation of tritium of chemically labeled DL-lysine-6-H3 into the lysine and hydroxylysine of sponge biopsy collagen: Table 3 gives the results of the experiments with the chemically prepared DL-lysine-6-H'. They indicate that hydroxylation does not involve the loss of a hydrogen atom attached to carbon 6. Previous ex- periments8 showed that the hydroxylation involves the loss of one hydrogen atom from carbon 5, none from carbon 4. If all of the 24 per cent of the tritium found in the DL-lysine-6-H' preparation on carbons other than carbon 6 was on carbon 5, which, as noted, seems probable, loss of half of the H'I (one atom) from C-5, and none from C-6, would remove 12 per cent and leave 88 per cent of the total tritium label in the hydroxylysine. The observed 85 per cent agrees within the limit of experi- mental error with this estimate. The loss of one hydrogen atom from carbon 6 and one from carbon 5, which would occur if hydroxylation occurred by way of a 5-6 unsaturated derivative, would leave only 50 per cent of the lysine tritium in the hydroxylysine if the 24 per cent of tritium not attached to C-6 were attached to TABLE 1 TABLE 2 SPECIFIC ACTIVITIES OF COMPOUNDS SPECIFIc ACTIVITIES OF COMPOUNDS PRODUCED IN DEGRADATION OF CHEMICALLY PRODUCED IN DEGRADATION OF BIOSYNTHETIC PREPARED D-LYSINE-6-H3 PI-LYSINE-6-H I Specific Per Per activity cent Specific Positions of H' of Positions activityof H' centof in lysine (dpm/ lysine in lysine (dpm/ lysine Compound represented jmole) H' Compound representednsmole) H' Lysine 2 3,4, 5, 6 16,364 (100) Lysine 2,3,r4,5,6 1,374 (100) Aminovaleric Aioaei acid 3, 4, 5, 6 16,462 101 aminovaleric 1388 10 Glutaric acid 3, 4, 5 3,796 23 a acid 3, 4, 5,6 1,388 103 Succinic acid 4,5 4,199 26 Glutaric acid 3, 4, 5 13 1 Downloaded by guest on September 25, 2021 396 BIOCHEMISTRY: POPENOE ET AL. PROC. N. A. S.

TABLE 3 TABLE 4 RETENTION OF HI IN THE FORMATION OF RETENTION OF HI IN THE FORMATION OF HYDROXYLYSINE FROM CHEMICALLY HYDROXYLYSINE FROM BIOSYNTHETIC PREPARED DL-LYSINE-6-H3 L-LYSINE-6-H3 Experi- ,-Ratio H3/C14- Per cent Experi- - Ratio H'/C'4 -_ Per cent ment Hydroxy- Hs ment Hydroxy- Hs no. lysine Lysine retained no. lysine Lysine retained 1 19.4 i 0.1 22.8±0.1 85.4 1 8.94 ± 0.20 8.44±0.03 106 2 19.4 i 0.1 22.9 ± 0.2 84.8 2 8.38 ± 0.20 7.69 ± 0.36 109 3 27.6±zt0.3 32.0±zt0.3 86.2 i i The per cent H' retained is lOOX the ratio H3/C'4in 427.427.44-0.1332.4As-32.4 4- 0.0.2 84*8864. hydroxylysine divided by the ratio H3/C"4 in lysine. Average 85.3 H3/CI4 ratios are given with the standard errors of replicate determinations. The per cent H' retained is lOOX the ratio H'/C'4 in hydroxylysine divided by the ratio H3/C14 in lysine. H3/C'4 ratios are given with the standard errors of replicate determinations.

C-5, or 62 per cent if it were attached to C-4. The results appear to exclude hy- droxylatioii of lysine by way of a 5-6 unsaturated derivative. Incorporation of tritium from biosynthetically labeled L-lysine-6-H3 into the lysine and hydroxylysine of sponge biopsy collagen: Table 4 shows the results. In these experiments it is clear that none of the tritium attached to lysine carbon 6 was lost in the hydroxylation. Discussion.-Fujimoto and Tamiyal6 have reported that atmospheric 02 is not incorporated into hydroxylysine in chick embryos, whereas either 018 or H3 from labeled water is incorporated. This observation indicates that hydroxylation of lysine proceeds by a mechanism different from that which Fujimoto and Tamiyal7 and Prockop, Kaplan, and Udenfriend'8 had previously identified for hydroxylation of proline, which was shown to occur by incorporation of atmospheric 02 into the proline. Our earlier experiments with lysine4-5-H3 and our present experiments with lysine-6-H3 seem to eliminate mechanisms b, c, and d postulated in the in- troduction, and to leave direct uptake of oxygen, such as occurs with proline, as the most plausible mechanism. No mechanism for the hydroxylation of lysine that is consistent both with our results and with those of Fujimoto and Tamiya is readily apparent. Summary.-Experiments in which lysine preparations labeled with tritium on carbons 4 and 5 and on carbon 6 served as sources for biosynthesis of hydroxylysine in collagen indicated that during the hydroxylation of lysine no hydrogen atom other than the one which is replaced by a hydroxyl group becomes exchangeable. The most plausible interpretation is that the hydroxylation occurs with direct addition of an oxygen atom to lysine carbon 5. The authors wish to express their appreciation to Dr. Harry Broquist for valuable advice con- cerning the preparation of lysine-6-H3 by means of yeast, and to Mrs. Agnes Chen for expert assistance. * This research was supported by the U.S. Atomic Energy Commission. 1 Hamilton, P. B., and R. A. Anderson, J. Am. Chem. Soc., 77, 2892 (1955). 2Sinex, F. M., and D. D. Van Slyke, J. Biol. Chem., 216, 245 (1955). 3 Piez, K. A., and R. C. Likins, J. Biol. Chem., 229, 101 (1957). 4Van Slyke, D. D., and F. M. Sinex, J. Biol. Chem., 232, 797 (1958). 6 Sinex, F. M., D. D. Van Slyke, and D. R. Christman, J. Biol. Chem., 234, 918 (1959). 6Popenoe, E. A., and D. D. Van Slyke, J. Biol. Chem., 237, 3491 (1962). 7Peterkofsky, B., and S. Udenfriend, these PROCEEDINGS, 53, 335 (1965). 8 Popenoe, E. A., R. B. Aronson, and D. D. Van Slyke, J. Biol. Chem., 240, 3089 (1965). Downloaded by guest on September 25, 2021 VOL. 55, 1966 BIOCHEMISTRY: BRINTZINGER ET AL. 397

9 Jones, E. E., and H. P. Broquist, J. Biol. Chem., 240, 2531 (1965). 10 Kabara, J. J., N. R. Spafford, M. A. McKendry, and N. L. Freeman, in Advances in Tracer Methodology, ed. S. Rothchild (New York: Plenum Press, 1963), vol. 1, p. 76. 1 Strassman, M., and S. Weinhouse, J. Am. Chem. Soc., 75, 1680 (1953). 12 Petrowitz, H. J., and G. Pastuska, J. Chromatog., 7, 128 (1962). 13Kao, K.-Y. T., and R. J. Boucek, Proc. Soc. Exptl. Biol. Med., 98, 530 (1958). 14 Fitch, S. M., M. L. R. Harkness, and R. D. Harkness, Nature, 176, 163 (1955). 15 Spackman, D. H., W. H. Stein, and S. Moore, Anal. Chem., 30, 1190 (1958). 16 Fljimoto, D., and N. Tamiya, Biochem. Biophys. Res. Commun., 10, 498 (1963). 17 Fuljimoto, D., and N. Tamiya, Biochem. J., 84, 333 (1962). 18 Prockop, D., A. Kaplan, and S. Udenfriend, Biochem. Biophys. Res. Commun., 9, 162 (1962)

ON THE LIGAND FIELD OF IRON IN FERREDOXIN FROM SPINACH CHLOROPLASTS AND RELATED NONHEME IRON ENZYMES* BY HANS BRINTZINGER, GRAHAM PALMER, AND RICHARD H. SANDS

BIOPHYSICS RESEARCH DIVISION, INSTITUTE OF SCIENCE AND TECHNOLOGY, UNIVERSITY OF MICHIGAN, ANN ARBOR Communicated by J. L. Oncley, December 27, 1965 The suspected structural relationship between the ferredoxin group of enzymes, recently found to participate in a variety of biological electron transfer reactions,1 and other nonheme iron enzymes has acquired substantial support by our recent finding2 that the ferredoxin isolated from spinach chloroplasts exhibits under re- ducing conditions an EPR spectrum of the type now well known for other nonheme iron enzymes by the work of Beinert and co-workers.3 More recently, such an EPR spectrum has also been obtained with ferredoxin from Clostridium pasteurianum.4 In the EPR spectra of these ferredoxins, as in those of the other nonheme iron enzymes that have been reported so far, two of the absorption extrema are observed at g-values lower than 2 and one is found just above the free electron value, g, (see Table 1). EPR spectra of this kind are very atypical for iron complexes usually encountered and must reflect a highly specific structural arrangement of ligand groups around the iron atoms in these proteins. The purpose of this communication is to substantiate a particular model for the environment of iron in these protein complexes which is in accord with their spectral properties. Construction of an Orbital Scheme.-The view is taken that the deviations of observed g-values from the free electron value, g, indicate that an unpaired electron is in an orbital to which another state, related to it by rotation around a coordinate axis, can be admixed by spin orbit coupling, X. (It has also been pointed out by Blumberg and Peisach (in ref. 1, p. 101) that exchange interaction of diamagnetic Fe(II) complexes with radicals can be expected to give rise to g-values below ge via interaction with an excited triplet state of the metal ion.) With this in mind, the observation of two absorption extrema below g,, as in the EPR spectra of the ferredoxins, can only arise from a situation where either one or two electrons or holes are centered around the iron atom in a near-degenerate set of three molecular orbitals involving the metal d_, d2z, and d1z, or p., py, and pz orbitals. In either case the unpaired electron(s) would have a partial freedom for precession around two of the Downloaded by guest on September 25, 2021