In Viva and Enzymatic
162, 614-619 (1974)
by Streptomyces TAKAYOSHI
UEMATSU AND ROBERT J. SUHADOLNIK2
Showa University, School of Pharmaceutical Sciences, Hatanodai, i-Chome, Shinagawa-ku, Japan and Department of &o-organic Chemistry, Albert Einstein Medical Center, Philadelphia,
Received November 14, 1973 The pyrrolopyrimidine nucleosides, toyocamycin, sa,ngivamycin, and tubercidin are isolated from the culture filtrates of 14 species of the Streptomyces. Although earlier experiments showed that the biosynthesis of the pyrrolopyrimidine nucleosides require GTP as the common precursor, there was no experimental evidence to demonstrate the interconversion of these naturally occurring nucleoside analogs. The data presented here describe two types of experiments to prove that toyocamycin is the precursor for sangivamycin. First, in vivo experiments show that radioactive toyocamycin is converted to sangivamycin. Second, the enzyme, toyocamycin nitrile hydrolase, that catalyzes the conversion of toyocamycin to sangivamycin has been isolated and partially purified from the soluble fraction of 8. rimosus. The nitrile hydrolase is not present in cell-free extracts of the Streptomyces that synthesize tubercidin or toyocamycin. Activity can be assayed by measuring the formation of radioactive sangivamyein from toyocamycin. The enzyme has been purified 24-fold with an over-all yield of 5%. The pH optimum is 6.5 and the K, is 0.5 mM. Most nitriles tested are competitive inhibitors but they are not substrates. The activity of the hydrolase is limited to the conversion of the nitrile group to the carboxamide group. Hydrolase activity is observed in cell-frre estracts of S. rimosus before toyocamycin production begins. The in vivo and in vitro studies demonstrate that toyocamycin is not a precursor for tubercidin. The experimental evidence strongly suggests that there must be a branch point in the biosynthesis of the pyrrolopyrimidine nucleoside antibiotics.
Naturally occurring nitriles are widely distributed among higher plants, insects, and fungi. These nitriles are found as cyanogenic glycosides, aromatic nitriles, aliphatic nitriles, glycosinolates, and a nucleoside antibiotic, toyocamycin. The biosynthesis of these five groups of nitriles has been studied. The biosynthesis of the cyanogenic glycosides, aromatic nitriles, and ‘This investigation was supported by a research grant from the U. S. Public Health Service (NIH AI 08932-12) and a grant from the National Science Foundation (GB 32288X). Inquiries regarding this work should be addressed to Dr. R. J. Suhadolnik. 2 Research Career Development Awardee of the U.S. Public Health Service (5-K3-GM-7100-71).
glycosinolates proceeds via the loss of the carboxyl group of a related amino acid and conversion of the amino group to a nitrile. These biological transformations proceed via an aldoxime (l-5) and support the earlier proposal of Ahmad and Spenser (6) on the biosynthesis of nitriles. The biosynthesis of the aliphatic nitrile P-cyanoalanine in Bacillus megaterium, Escherichia coli, and Lupinus an,gustijolia is catalyzed by /3cyanoalanine synthase (7-l 1). Either serine or cysteine and HCN is the substrate. In contrast to the uptake and assimilation of cyanide by plants and bacteria to form organic nitriles, Streplomyces rimosus is unable to add [14C]eyanide as a C-l unit to carbon 5 of tubercidin to form either the
614 Copyright A11 rights
Q 1974 by Academic Press, of reproduction in any form
(il.5 nitriltb or carhoxamid(> group (12). Koukol et al. (13) also showrd t!hat cyanidfl was not the prrcursor for the nitrilc group of the c-yanog(9ic glvcc lsidcs. Sitrilo m&abolism in plants, bacteria, and fungi invoIvw wwral diff Pront pathLva1.s. Thtx products arc 4thcr HCN, an am.kic, or an acid. In higher plants, Pc~~anoalaninc~ h~.drolaso converts @-cyanoalaniw to asparagiw. l’urthrr hydrolysis to :tspartic acid tlc~>s not occur (7). With bac+ria, P-c~anoalaninc is convcrtcd to asparaginrk and finally to aspartic acid (S). The format,ion of an amidr from a nitrile> hl a f’err&‘lliu~t~ has bwn reported by Thcriault et al. (14). Although in aivn and in vitlo st)udies with thcb Streptomyces that product> th(l 1)Yrrolopyrirllidine nucleosides wtablishcd that guanosiw scrvw as thr carbonnitrogtrn pwcursor (12, 1.5, lci), no information is availabl(b concerning the interconvcbrsion of th(w thrw nat,urally occurring noc1wsidcs. This rqort describes the ill viva conwrsion of toyocamycin to sangivamycin and the partial purification of the cnqrne, to?-ocamycin nitrilc hydrolasc, that does this cotiwrsion. MATERIALS
Radiochemical measurements were made on a Packard scintillation Spectrometer Model 314E; counting solution was according to Bray (17). Toyocamycin was generally labeled with tritium by New England Nuclear Corporation by the gasexchange method of Wilzbaeh. Toyocamycin was p\lrified to constant specific activity by paper chromatography using water and crystallized to I,onstant specific a.ctivity from water and finally t:t,hanoI. Ultraviolet measurements were made on :t Beckman DB spectrophotometer. Ire vivo conversion of loyocamvcin lo sangivnmycirt anti trrbercit-lin. These cultures of Sfreptomyces were used: S. rirnosus (ATCC No. 14,G73) (toyocamycin and sangivamycin producer), S. rirnosus (.4TCC No. 14,500) (sangivamycin producer), and S. (t~bercitlic~cls (tubercidin producerj. The procedure for the production and in vivo conversion cxxprriments for toyocamgcin, sangivamycin, and tubercidin were as follows: Cultures were maintained on agar straws at -2O’C. One straw was added to a P-liter baffled flask containing 300 ml of t.he appropriate medium (18). incubations were done at 27°C for 30 hr on a rotary shaker. Two milliliters of inoc*lllum were added to Z-liter fasks cont.:tining 200 ml of the same medilun and
200 180 G;
FIG. 1. Biosynthesis of toyocamycin (e---O) and sangivamycin (a---e) by S. G/,/<,sicz (ATCC No. lJ,K73). were shaken at 27°C for 50 hr. Toyocanlycin pro.duction reached a maximum 3(i hr after inoc*ul:ltion. Hangivamyciu prodllction reached a lnaxi mum 16 hr later (Fig. 1). Tubercidin production reached a maximum 48 hr after inovulation. The in viva biosynthesis of sangivamycin was den~or~strated by adding [“H]t,oyocamycin (3 &i, 0.2 rmoles) to flasks containing inoculations trf Streptomyces rirnosus (ATCC No. 14,673) 5%hr inoculation. One hour later 2 ml of t,he culturr medium was removed and filtered. The filtrate was concentrated to 0.1 ml, added to a Whatman No. 3i14M paper chromatogram, and developed ~JI n-butanol-water (%:lJ, v/v); X, of toyocamycitr, tubercidin, and sangivamycin: 0.~0,0.:33, and 0.2’2, respectively. The sangivamycin area was cut, out, eluted with wat,er, and the specific uc*tivity tletcxrmined. The sangivamycin from the rnlt,ure rn+ dium was isolat,ed and crystallized :ts desrribetl (18). Similar experiments were tlonc~ I)y adding [U-“Hltoyocamycin to c[lltures of S. rirtrosrts (ilTCC No. 14,500) and 8. /uOercid;crts. Enzymaiic assays. l>eionized water was (used for all reagents. The general pro&ure used 1’o1 estimation of the rritrile hydrolase was as follows: [U-3Hlt.oyocamvci11 (‘2.2 X 10F4M; -1lii,‘mole) and extracts from the S(repiomuces strains tiesc*rit)td above were incubated in 0.1 M potassirun phosphate buffer (pH (;.5), at 30°C in R final volumta of 0.1 ml. Control samples were performed with boiled enzyme or without enzynle. Iteactions were terminated by the addition of 0.1 1n1 of ethanol a(*tftir acid (!): 1. v v). (‘ari-ier s:ingivamycirI 130
rg) was added, the mixture was spotted on Whatman No. 3111~paper, and the paper was developed with water. The area corresponding to sangivamytin (Rf 0.80) was cut out and eluted with 0.1 N HCl. Preparation of cell-free extracts. Extracts were made of the three Streptomyces listed above. All purification steps were carried out at 04”. The cells were harvested by centrifugation at 4”C, 48 hr after inoculation, washed twice with 0.15 M KCI, lyophilized, and stored at -20°C. The dried cells (5 g), obtained from 600 ml of medium, were suspended in 100 ml of 0.10 M potassium phosphate buffer (pH 7.4) and disrupted with a French press (twice at 16,000psi). The homogenate was suspended in 50 ml of buffer, stirred for 10 The supernatant and min, and centrifuged. washings were combined. Purijkation procedure. The purification of the hydrolase described here is for 8. rimosus (ATCC No. 14,673). The crude extract was brought to 30% saturation by the gradual addition of solid ammonium sulfate over a 30-min period with stirring. The precipitate was removed by centrifugation (35,OOOg,10 min). Ammonium sulfate was added to the supernatant to 60% saturation, centrifuged for 10 min (15,OOOg).The precipitate was dissolved in 10 ml of water, and dialyzed overnight with 2 liters of 1 mM EDTA (pH 7.4). Protamine sulfate treatment. A 1% protamine sulfate solution (pH 6.8) was added until the OD2&ODz60 of the supernatant became 0.90. The solution was centrifuged at 20,OOOgfor 10 min. The pellet was suspended in 5 ml of 0.5 M phosphate buffer (pH 6.5). After stirring for 2 hr, the solution was centrifuged at 2O,OOOg,10 min. The pellet was discarded and the supernatant was dialyzed against 3 liters of 1 rnM EDTA, 3.5 hr. Hydroxylapatite adsorption. The supernatant (13.5 ml) was passed through a hydroxylapatite column (10 ml). The elution of protein was carried out stepwise with 10 ml of 0.05, 0.08, 0.10, 0.12, 0.16, and 0.20 M phosphate buffer (pH 6.8). The hydrolase was eluted with 0.12 M buffer. RESULTS
Two experiments are described that show that toyocamycin is the precursor for sangivamycin. The first experiment is an in vivo study in which tritium-labeled toyocamycin is added to cultures of S. rimosus that produce sangivamycin. The second experiment describes the properties of the enzyme, toyocamycin nitrile hydrolase, that accomplishes this conversion. In vivo conversion of toyocamycin and sangivamycin by S. rimosus (ATCC No.
14,673). Figure 1 shows that S. rimosus (ATCC No. 14,673) produces two nucleosides. Toyocamycin production reaches a maximum 36 hr after inoculation. Sangivamycin production does not reach a maximum until 52 hr after inoculation. Although there was a 20 % conversion of [U-3H]toyocamycin to [U-3H]sangivamycin by S. rimosus (ATCC No. 14,673), there was no apparent uptake of toyocamycin. Therefore, the rate of uptake of toyoeamycin and the rate of synthesis of sangivamycin and excretion into the medium was the same. The conversion of toyocamycin to sangivamycin did not occur in the culture medium because the medium, when freed of the Streptomyces, did not make this conversion. S. tubercidicus did not take up [U-3H]toyocamycin, nor did it produce sangivamycin. In vitro studies: Purijcation and properties of toyocamycin nitrile hydrolase. To find the optimum conditions for purification of toyocamycin nitrile hydrolase, extracts were made at various times after inoculation. The specific activity of the enzyme was the same in extracts 16-70 hr after inoculation indicating that the hydrolase is active before toyocamycin production begins. This is in contrast to the report of Elstner and Suhadolnik (19) who showed that GTP-8formylhydrolase, the first enzyme in toyocamycin biosynthesis, is present in cellfree extracts only after the excretion of toyocamycin into the culture medium. The purification employed for the nitrilase is summarized in Table I. The purified hydrolase was stabilized by storage at -20°C in 50% glycerol with about 25 % activity remaining in 6 mo. The pH optimum of the enzyme is 6.5 (Fig. 2). There is no conversion of toyocamycin to sangivamycin at pH 6.5. Nonenzymatic conversion of the nitrile to the carboxamide occurs when the pH is greater than 10. The substrate velocity curve is shown in Fig. 3. The K, for toyocamycin is 0.5 mM (Fig. 4). The curves were linear and showed no curvature. Enzyme activity as a function of time and protein concentration was also studied. At 17 X 1O-4 M toyoeamycin and 5.3 units of protein, product formation was linear for 60 min at 3O’C; when the protein was varied between 43 and 172 pg, there \vas a linear increase in
ti17 TABLE PURIFICATIOK
--__ _----~ Crude extract (dialyzed) 3&55% Ammoninm sulfate (dialyzed) Extract of pratamine Hydroxylapat,ite
OF TOYOCAMYCIX .___.-
” A llnit, of enzyme is defined 0.1 ml incubation mixtllre.
as the formation
HYDRC)I..IGE .~ ---.----
‘l‘otal activity (units)‘”
____~~ 1014 201
28 2 2 of 1 nanomole
KWOVer> i(‘; j
1 .o 5.‘)
:?I 0 47.4
Specific Yurifiactivity cation lunitlmg protein) ._ ~__ __ ..-. ~.
per 30 min at :SO”(’ ill
FIG. -1. Lineweaver-Uurk plot For toyoeamyrin in the togocamycin nitriie hydrolasp reactions. Partially purified enzyme and radioartiv? assays arr described under Materials and Met hods.
FIG. 2. Effect of pH on the enzymatic and lionenzymatic conversion of toyocamycin to sangivamycin. Fnzymatic conversion was done with the following buffers (0.1 M): (O----O) acsetatr; (@---a) phosphate; (@---a) nonolzymatic conversion,
10e4 M TOYOCAMYCIN
FIG. 3. Substrate velocity curve for toyocamytin in the toyocamycin nitrile hydrolase reaction. Enzyme and assays as described under Materials and Methods were used.
product’ formation. The cf’fwt of heat on enzyme stability showd a 30’S hJsS of activity, when the fwzyme (0.4 mg protein; (i X lo-” M [U-“HltoyocamyciI1) LVBSheatcad for 10 min at 40°C; heating for 10 min at WC resulted in complctc~ loss of c~nzymo activity. Eflecf oj’ metals. Cupric ion, -\I$+, ZrP, Kc++, and Hg*+ art’ inhibitors of the bydrolasc (Table II); CaZ+ and AIriL+ had no c+fcct’ on the activity of the c~nzynw. Efiecf oj inhibifors of foyocamyr.it~ 11itrile kyrlrolase. Tabk III lists the rwult;; I II>taiwd with various cnzymc inhibitors on toyocamycin nitrile hydrolaw. In f,hcw c~xpc~rinwnts, the> inhibitors LVWVincubatSc:d ;i min with the enzymr br>forc substrate was addtad. Of the inhibitors studkd, p-hydroxybc>nzonitrile and tubrrcidin \\w(’ S~IOWI~ to be competitive inhibitors wit.h 1%6 i of 5.7 X IO-” >I and S.0 X lo-” 11, rcspwtively (i’ig. 5, b) Skthcr tuhcrcidin r,‘-rnonc)I)hosp~lnt,c~
TABLE II KFFICCTOE' METAL IONS ON TO~OCWYCIN NITRILE HYDROL.WP Addition (M) (2.5 x 10-Z)
nor the nitriles tested were substrates for the hydrolase. DISCUSSION
y0 of Control
None CaClz MnCl? ZnSOl MgCl, cuso4 HgCl: FeClz
In vivo and in vitro studies are described to show the conversion of toyocamycin to sangivamycin. Earlier in vivo studies showed that guanosine is the common precursor for the biosynthesis of the pyrrolopyrimidine nucleosides, toyocamycin, sangivamycin, and tubercidin (Fig. 6) (12). Subsequent studies with cell-free preparations showed that the first step in the biosynthesis of the pyrrolopyrimidine nucleosides involved the loss of C-8 of GTP by GTP-S-formylhydrolase (19). It had also been established that a C-l group (as a cyano or carboxamide group) could not be added to carbon-5 of tubercidin to form toyocamycin or sangivamycin (12). The studies described here have demonstrated (1) that toyocamycin is converted to sangivamycin both in vivo and in vitro and (2) that the enzyme, toyocamycin nitrile hydrolase, does this conversion. The enzyme does not hydrolyze the amide group of sangivamycin to the corresponding acid. Although GTP is the common precursor for the three pyrrolopyrimidine nucleoside antibiotics and toyocamycin is converted to sangivamycin, the biosynthesis of tubercidin is not dependent on either toyocamycin or sangivamycin (Fig. 6). Also, the enzymatic conversion of toyocamycin to sangivamycin occurs as the nucleoside and not the nucleotidc. The toyocamycin
100 112 100 53 34 0 0 0
o Substrate was 3.2 X 10e4 M; 5.71 units of enzyme, 0.4 &I phosphate buffer (pH 6.5), 3O”C, 30 min. TABLE III EFFECT OF INHIEITORS ON TOYOCXMYCIN NITRILE HYDROL~WY Inhibitor
Concentration (X ‘?I-~,
None p-Hydroxybenzonitrile Tubercidin Ricinine Tubercidin 5’-monophosphate Nicotinonitrile Hydroxytoyocamycin Demethyl rieinine O,N-Didemethyl ricinine n Substrate concentration 5.7 units of enzyme.
Inhibition (%I 0 81 64 32 25 20 16 15 2
2.5 5.8 3.2 1.9 3.6 2.7 1.4 1.2
was 3.2 X 10e4 M;
FIG. 5 a, b. Inhibition of toyocamycin tubercidin (b). Substrate concentrations: (O-0) 12.8 x lo-” M.
nitrile hydrolase by p-nitrobenzonitrile (a) and (A--A) 3.2 X lo+ M; (O---O) 6.4 X 10e4hl:
TOYOCAMYCIN NITRILE HYDROLASE
R’= RIBOSE \
& N Cl TUBERCIDIN
FIG. 6. Conversion
of GTP to toyoramycin,
nit&t hydrolasa from 8. rhesus is similar to p-cyanoalaninc hydrolasr isolated from Lupirrus anyusfi~olius and Sorghum vulyare (7) in that the nitrile is hydrolyzed to the amide but not to the free acid. However, txtracts of snow mold and Bacillus meqatwium hydrolyzed asparagine but not @cyanoalanine. The above enzymes differ from the E. roli hydrolase which hydrolyzw cyanoalaninc and asparagine (‘20). Tovocamycin nitrile hydrolase does not req&e a metal ion. Of the nitriles studied as illhibitors, tubercidin and p-hydroxybcnzonitrilc wrc the best competitive inhibitors. Thr physiological function of the hydrolase appnars to br specific for the conversion of t’hc cyano group of toyocamycin to sangivamycin since the Streptomyces that synthesizes toyocamycin and tubercidin does not haw an active toyocamycin nitrilc hydrolasc. The liz ~duo and in uitm evidence strongly support’s the idea that a branch point exists in the biosynthesis of the pyrrolopyriInidin(: nwl~wid~~ st&biotics (Fig. 0).
8. !I. 30. 11. 12. 13. 14.
15. 16. 17.
Hioch~~n. Biophys. 126, 873-883. 3. TAPPICR, R. A., Goss, 17. E,, .WD BUTLIX, (i. W. (lSCi7) dxh. Kitxhevr. f%ophys. 119, 59:s595.