THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. No. 30, in A. Two … · 2001-07-17 · THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 263, No. 30, Issue of October 25, pp. 15626-15633,1988

Printed in U. S. A.

Two Distinctive 0-Methyltransferases Catalyzing Penultimate and Terminal Reactions of Macrolide Antibiotic (Tylosin) Biosynthesis SUBSTRATE SPECIFICITY, ENZYME INHIBITION, AND KINETIC MECHANISM*

(Received for publication, April 7, 1988)

Adam J. KreuzmanS, Jan R. Turner#, and Wu-Kuang YehSIl From the $Biochemical Development Division and SLilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, Indiana 46285

S-Adenosyl-L-methionine:demethylmacrocin O-me- thyltransferase catalyzes the conversion of de- methylmacrocin to macrocin as the penultimate step of tylosin biosynthesis in Streptomyces fradiae. The 0- methyltransferase was purified to electrophoretic ho- mogeneity by a conventional chromatographic proce- dure. The purified enzyme appears to be trimeric with a molecular weight of 122,000-126,000 and a subunit size of 42,000. Its isoelectric point was 6.0. The en- zyme required Mg2+ for maximal activity and was catalytically optimal at pH 7.8-8.5 and 42 “C. The 0- methyltransferase catalyzed conversion of demethyl- macrocin to macrocin at a stoichiometric ratio of 1:l. The 0-methyltransferase also mediated conversion of demethyllactenocin + lactenocin. The corresponding V,,,/K,,, ratios for the two analogous conversions var- ied only slightly. Both enzymic conversions were sus- ceptible to an extensive and identical range of meta- bolic inhibitions. Steady-state kinetic studies for initial velocity, substrate analogue, and product inhibitions are consistent with Ordered Bi Bi as the reaction mech- anism of demethylmacrocin 0-methyltransferase. Ex- cept for an identical kinetic mechanism, demethylma- crocin 0-methyltransferase can be readily differen- tiated from macrocin 0-methyltransferase by its physical and catalytic properties as well as metabolic inhibitions.

Demethylmacrocin 0-methyltransferase and macrocin 0- methyltransferase catalyze the penultimate and the terminal methylation reactions, respectively, of tylosin biosynthesis in Streptomyces fradiae (Fig. 1; Refs. 1-3). Significant accumu- lations of macrolide substrates (i.e. demethylmacrocin and macrocin) in tylosin fermentations of S. fradiae suggest that both demethylmacrocin 0-methyltransferase- and macrocin 0-methyltransferase-catalyzed reactions are rate-limiting for the antibiotic biosynthesis (4). The consecutive enzymic methylations which control tylosin biosynthesis represent a unique type of molecular regulation in secondary metabolism. In the preceding paper, we derived from the enzyme studies that macrocin 0-methyltransferase-catalyzed conversion of macrocin to tylosin in vivo is likely to be affected by lactenocin as a competing macrolide substrate and also to be controlled by broad but specific metabolic inhibitions. To further our understanding of the metabolic regulation of enzymic meth-

* 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.

7 To whom correspondence should be addressed.

ylation in tylosin biosynthesis, we report here purification and characterization of demethylmacrocin O-methyltransfer- ase. In comparison to macrocin 0-methyltransferase, deme- thylmacrocin 0-methyltransferase is distinctively and more extensively susceptible to metabolic inhibitions. Also, the two 0-methyltransferases are readily distinguishable by their physical, catalytic, and kinetic properties.

EXPERIMENTAL PROCEDURES AND RESULTS’

Substrate Specificity-In addition to methylation of de- methylmacrocin to macrocin, the 0-methyltransferase also catalyzed effectively methylation of demethylactenocin to lactenocin (as described under “Initial Velocity Pattern” of the Miniprint). No other macrolide compound as listed in Table I11 was a substrate for demethylmacrocin O-methyl- transferase. In contrast, AdoMet’ was the only methyl group donor for the demethylmacrocin O-methyltransferase-cata- lyzed methylation of demethylmacrocin and demethyllacten- ocin; no enzymic methylation of either macrolide compound occurred when AdoMet was substituted with L-methionine, 5’-methylthioadenosine, N5-methyltetrahydrofolate, betain or choline.

Enzyme Inhibition-The demethylmacrocin O-methyl- transferase-catalyzed conversion of demethylmacrocin to ma- crocin was inhibited by macrolide compounds in the following order: lactenocin, 20-dihydrolactenocin > 20-dihydromacro- cin > mycaminosyltylactone, 20-dihydro-23-deoxymycami- nosyltylonolide, 23-deoxy-O-mycaminosyltylonolide, 23-de- mycinosyloxytylosin, 0-mycaminosyltylonolide (Table 111). No inhibition was observed with tylactone, desmycosin, tylo- sin, and relomycin. Similarly, the alternative enzymic conver- sion of demethyllactenocin to lactenocin was inhibited by 20- dihydromacrocin, macrocin, 20-dihydro-23-deoxymycamino- syltylonolide, 23-deoxy-0-mycaminosyltylonolide, 23-demy- cinosyloxytylosin, and demycinosyltylosin (Table 111). Also similarly, little or no inhibition of the alternative conversion was observed with tylactone, desmycosin, tylosin, and relo- mycin. The demethylmacrocin + macrocin and the deme-

Portions of this paper (including “Experimental Procedures,” part of “Results,” Figs. 2-10, and Tables I and 11) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

The abbreviations used are: AdoMet, S-adenosyl-L-methionine; DMOMT, demethylmacrocin 0-methyltransferase; MOMT, macro- cin 0-methyltransferase; AdoHcy, S-adenosyl-L-homocysteine; PMSF, phenylmethylsulfonyl fluoride; HPLC, high performance liq- uid chromatography; FPLC, fast protein liquid chromatography; and SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electropho- resis.

15626

Demethylmacrocin and Macrocin 0-Methyltransferases 15627

0

Mp’+h

FIG. 1. Demethylmacrocin 0-methyltransferase-catalyzed conversion of demethylmacrocin to macrocin.

thyllactenocin + lactenocin conversions were not susceptible to inhibition by their own macrolide substrates (Table 111). The two demethylmacrocin 0-methyltransferase-catalyzed conversions were similarly inhibited by AdoHcy, sinefungin and A9145C (Table IV); however, the degrees of inhibition of the two conversions by each compound varied slightly. No inhibition of either conversion was observed with D-isomer of AdoHcy, AdoMet (up to 900 p ~ ) , adenine, and 5”methyl- thioadenosine.

TABLE I11 Effect of macrolide compounds on demethylmacrocin 0-

methyltransferme-catalyzed conversions of demethylmacrocin + macrocin and demethyllactenocin + lactenocin

Inhibition(%)c

9+12 8+10 Macrolide Compounda Structural Representationb

Tylactone(1)

20-Dihydro-23-deoxy- Mycaminosyttylactone(2)

23-Deoxy-0-mycarninosyl- rnycarninosyltylonolide(3)

tylonolide(4)

OMycaminosyltylonolide(5)

23-Dernycinosyloxytylosin(6)

DernydnosyltylosinQ)

Demethyllactenodn(8)

Demethylmacrocin(9)

Lactenodn(l0)

20-Dihydrolactenodn(ll)

Macrocin(l2)

20-Dihydromacrodn(l3)

Desmycosin(l4)

Tylosin(l5)

Relornycin(l6)

20 - 32 30

- 15

- nd

nd-

81 - 76 - - 47

5 5 5 6

nd nd

nd nd

nd nd

DISCUSSION

Both demethylmacrocin and macrocin O-methyltransfer- ases, which catalyze the penultimate and terminal methyla- tions, respectively, of tylosin biosynthesis in S. fradiae, have been purified to electrophoretic homogeneity. The two 0- methyltransferases from cell-free extracts of S. fradiae are separable by an anion-exchange chromatography. Purified demethylmacrocin and macrocin 0-methyltransferases are readily distinguishable by their physical and catalytic prop- erties, substrate specificities, and metabolic inhibitions. The two 0-methyltransferases are sensitive differentially to inhi- bition by adenine-containing compounds and particularly to sinefungin; the profound difference in sensitivity to sinefun- gin is also evident from the study of demethylmacrocin + macrocin + tylosin conversions in uiuo. In addition, the amino acid compositions and primary structures of the two O-meth- yltransferases, the latter of which have been almost com- pletely determined (to be published elsewhere), share no sig- nificant similarity. However, demethylmacrocin and macrocin 0-methyltransferases appear to follow the same kinetic reac- tion mechanism. Those enzymological data lead us to specu- late that either the two 0-methyltransferases arise by conver- gent evolution or they represent gene products of divergent evolution that is very extensive and thus not detectable by their structure-function analyses.

Differential Physical and Catalytic Properties-Purified de- methylmacrocin and macrocin 0-methyltransferases are quite stable at 4 “C in the presence of AdoMet and ethanol. The optimal reaction pH values of the two 0-methyltransferases are similar, and Mg2C was required for maximal activity of both enzymes (Table V). Maximal activity of macrocin 0- methyltransferase but not demethylmacrocin O-methyltrans- ferase can also be obtained when M g + is substituted by Mn2+ or Co2+. The two 0-methyltransferases, however, differ sig-

Macrolide compounds which were tested for inhibition of de- methylmacrocin 0-methyltransferase-catalyzed conversions include all nine metabolites (1-5, 8,9, 12, 15) from the preferred pathway (I), two metabolites (10, 14) from the alternative methylation route, three shunt metabolites (6, 7, 16), and two C-20 reduced macrolide substrates (11, 13). Except for 6, the metabolites are related to each other as shown in Fig. 11.

* Each macrolide compound is represented in the following manner: (20)CH3

I S,-O-H~C-T”O”S~-S~ / \

(3”’)HO OH(2”’) (20)CHa

I where Hac-T-O=tylactone, Sl=mycaminose, Sz=mycarose and

2 (3”’)HO OH(2”’)=6-deoxy-~-allose.

e 9 + 12 and 8 + 10 designate demethylmacrocin O-methyltrans- ferase-catalyzed conversions of demethylmacrocin + macrocin and demethyllactenocin + lactenocin, respectively. The uninhibitory ac- tivity of demethylmacrocin 0-methyltransferase as determined with 30 pM demethylmacrocin and 600 @M AdoMet was 0.58 nmol/min and that as determined with 30 pM demethyllactenocin and 600 p~ AdoMet was 0.54 nmol/min. For inhibition analysis of the enzyme, each macrolide compound was added at 60 p~ (i.e. at a total concen- tration of 90 pM when used as a macrolide substrate).

nd, not detectable (i.e. +3% of either uninhibitory activity). e -, not analyzed.

nificantly by their oligomeric structures, isoelectric points, and optimal reaction temperatures (Table V).

Distinctive Substrate Specificities-Among those macrolide compounds examined (1-16 of Table 111) for substrate dis- appearance and product formation by HPLC, only demethyl- macrocin and demethyllactenocin are macrolide substrates of

15628 Demethylmacrocin and Macrocin 0-Methyltransferases

demethylmacrocin 0-methyltransferase. In contrast, the mac- rolide substrates for macrocin 0-methyltransferase are ma- crocin, 20-dihydromacrocin, lactenocin, and 20-dihydrolac- tenocin. The distinctive substrate specificities of the two 0- methyltransferase are consistent with previously described bioconversion data of the macrolide compounds by S. frudiae (1). The bioconversion data also suggest that 20-dihydrode- methylmacrocin and 20-dihydrodemethyllactenocin are mac- rolide substrates for demethylmacrocin 0-methyltransferase; however, this substrate specificity of the enzyme has not been examined in vitro. Thus, the two 0-methyltransferases share a single methyl group donor but no common macrolide sub- strate. The K,,, of demethylmacrocin 0-methyltransferase for demethylmacrocin or demethyllactenocin is similar to that of macrocin 0-methyltransferase for macrocin or lactenocin, and the K,,, of demethylmacrocin 0-methyltransferase for AdoMet with either macrolide substrate is 5-fold higher than that of macrocin 0-methyltransferase for AdoMet with either mac- rolide substrate (Table VI). The V,, for demethylmacrocin 0-methyltransferase-catalyzed conversion of demethylmacro- cin "+ macrocin or demethylactenocin - lactenocin is similar to that for macrocin 0-methyltransferase-catalyzed conver- sion of macrocin + tylosin or lactenocin "+ desmycosin. The substrate specificities and associated kinetic data of de- methylmacrocin and macrocin 0-methyltransferases substan- tiate the methylation route for conversions of demethylma- crocin + macrocin + tylosin (1, 3, 9, 10) and are also consistent with an alternative methylation route for conver- sions of demethyllactenocin + lactenocin - desmycosin in

TABLE IV Effect of adenine-containing compounds on demethylmacrocin 0- methyltransferase-catalyzed conversions of demethylmncrocin -+

macrocin and demethyllactenocin + lactenocin

tylosin biosynthesis of S. frudiae (as shown in Fig. 11). Ac- cording to the "preferred" and the alternative methylation routes, demethylmacrocin 0-methyltransferase-catalyzed 2"'- 0-methylation precedes macrocin 0-methyltransferase cata- lyzed 3"'-0-methylation, and both types of methylations occur within 6-deoxy-~-allose bound to tylactone mycaminose with or without mycarose. Thus, the consecutive methylations result in a common tylactone-bound mycinose. The preferred methylation route is also supported by an observation that cell-free extracts from tylosin-producing strains of S. fradiae can convert demethylmacrocin to both macrocin and tylosin, whereas the same extracts can convert macrocin to tylosin only (1,9, and Footnote 3). A different methylation route for conversions of 6-deoxy-~-allose to mycinose before attach- ment of the sugar to tylactone has also been proposed in tylosin biosynthesis of S. fradiae (11). This methylation route may be elucidated simply by analyzing whether free 6-deoxy- D-allose and a monomethylated (2-0 or 3-0) derivative can serve as substrates for demethylmacrocin and macrocin 0- methyltransferases, respectively, or for other different 0- methyltransferases. The substrate specificity of the putative 0-methyltransferases has yet to be examined.

Independent Metabolic Inhibitions-The inhibitions from regular and shunt macrolide metabolites of tylosin biosyn- thesis for demethylmacrocin 0-methyltransferase-catalyzed conversions of demethylmacrocin "-* macrocin and demethyl- lactenocin + lactenocin are totally different from those for macrocin 0-methyltransferase-catalyzed conversions of ma- crocin + tylosin and lactenocin + desmycosin (Fig. ll). For either 0-methyltransferase, the inhibitions for one conversion are not distinguishable from those for the other conversion. Demethyllactenocin and demethylmacrocin, the substrates of

A. J. Kreuzman, N. J. Bauer, and W.-K. Yeh, unpublished data.

Adenine-containing Concentration Inhibition" compound 9-12 8-10

P M

Adenine 450

5'-Methylthioadenosine 450

AdoMet 900

AdoHcy 112.5 450

D-Isomer of AdoHcy 450

Sinefungin 112.5 450

A9145C 112.5 450

% NDb ND

ND ND

ND ND

25 7 74 44

ND ND

15 ND 43 24

45 19 89 54

a The uninhibitory activity of demethylmacrocin O-methyltrans- ferase as determined with 25 pM demethylmacrocin and 450 p~ AdoMet was 0.36 nmol/min and that as determined with 25 p~ demethyllactenocin and 450 p~ AdoMet was 0.33 nmol/min. The designations for 9 + 12, 8 + 10 are the same as those in Table 111.

ND, not detectable.

TABLE VI Kinetic constants of demethylmacrocin and macrocin

0-methvltransferases

Kinetic constant

With a preferred" macrolide substrate K,,, (pM) for

Demethylmacrocin 6 Macrocin 5 AdoMet 111 23

VmaX (pmol/min/mg protein) 0.15 0.23

substrate K,,, ( p M ) for With an alternative macrolide

Demethyllactenocin 10 Lactenocin 7 AdoMet 111 22

Vmax (pmol/min/mg protein) 0.14 0.21 From the preferred pathway of tylosin biosynthesis (1).

TABLE V Physical and catalytic properties of demethylmacrocin and macrocin 0-methyltransferases

Parameter 0-methyltransferase Demethylmacrocin

0-methyltransferase Macrocin

Oligomeric structure Molecular weight Subunit size Isoelectric point Metal ion for maximal activity Optimal reaction pH Optimal reaction temperature ("C)

Trimer or dimer-tetramer Dimer

42,000 32,000 6.0 4.5

122,000-126,000 65,000

M e M&+, MnZ+, Co2+ 7.5-8.5 7.5-8.0 42 31

Demethylmacrocin and Macrocin 0-Methyltransferases 15629

Tylactone

OoMycarninosyltylactone 0 0

c I

0 0 2O-Dihydro-23-deoxyrnycarninosyltylonolide 00

NC c 0 023Deoxy-0-rnycaminosyltylonolide 0 0

NC 4 0 Dernycinosyltylosin 0”- 0 00-Mycarninosyltylonolideoo

C I Dernethvllactenocin.

C / Demethylmacrocin 0 \

nd Macrocin 00 K- MOMT AdoMet-4

AdoHcy 4 nd Tylosin tc- - NF Desrnycosin nd

i Relornycin

FIG. 11. Me 5tabolic inhibitions of demethylmacrocin 0- methyltransferase- and macrocin O-methyltransferase-cata- lyzed conversions. A metabolic step of the preferred pathway (1) is indicated by +, an alternative metabolic step (postulated) by -+, and a shunt metabolic step (postulated) by -+. For demethylmacrocin 0- methyltransferase-catalyzed conversions (in bold arrows); weak, mod- erate, and potent inhibitions are represented, respectively, by 0 ,O 0, and 0 0 0. For macrocin 0-methyltransferase-catalyzed conversions (in bold arrows); weak and moderate inhibitions are represented by 0 and 0 0, respectively. The extent of inhibition for demethylmacro- cin 0-methyltransferase-catalyzed demethylmacrocin + macrocin conversion or for macrocin 0-methyltransferase-catalyzed macrocin + tylosin conversion is marked to the left of the inhibitory metabolite, whereas the extent of inhibition for demethylmacrocin O-methyl- transferase-catalyzed demethyllactenocin + lactenocin conversion or for macrocin 0-methyltransferase-catalyzed lactenocin -+ desmy- cosin conversion is marked to the right of the inhibitory metabolite. Inhibition by either macrolide reaction product in its own formation as catalyzed by either 0-methyltransferase was not determined (in- dicated by n d ) . C, competitive inhibition; NC, noncompetitive inhi- bition.

demethylmacrocin 0-methyltransferase, are moderate inhib- itors of macrocin 0-methyltransferases. Demethylmacrocin 0-methyltransferase is susceptible to moderate inhibition by the four early sequential metabolites from mycaminosyltylac- tone to 0-mycaminosyltylonolide, whereas macrocin O-meth- yltransferase is subject to weak inhibition by tylosin and desmycosin, the reaction products as two late metabolites of the revised pathway (Fig. 11). Also, demethylmacrocin 0- methyltransferase is weakly inhibited by demycinosyltylosin, the early shunt metabolite, and macrocin 0-methyltransferase is weakly inhibited by relomycin, the late shunt metabolite. In contrast to the weak product inhibitions of macrocin 0- methyltransferase, demethylmacrocin 0-methyltransferase is moderately and strongly inhibited by macrocin and lactenocin (i.e. reaction products), respectively. Thus, although both 0- methyltransferases are susceptible to broad metabolic inhi- bition rather than specific product inhibition, the inhibition scope by macrolide metabolites for demethylmacrocin 0- methyltransferase is independent from that for macrocin 0- methyltransferase. On structural basis, the minimal require- ment for a metabolic inhibitor of demethylmacrocin O-meth- yltransferase is tylactone-bound mycaminose (Table 111) and that of macrocin 0-methyltransferase is tylactone-linked my- caminose plus 6-deoxy-~-allose or mycinose (1).

Distinguishable Inhibitions by Adenine-containing Com- pounds-Both demethylmacrocin and macrocin O-methyl- transferases are susceptible to inhibition by AdoHcy (the common nucleoside reaction product), sinefungin, and A9145C (both are side-chain analogues of AdoHcy). The inhibitions of demethylmacrocin 0-methyltransferase by the three adenine-containing compounds vary greatly in extent from those of macrocin 0-methyltransferase. Demethylma- crocin 0-methyltransferase is similarly sensitive to the three compounds (Tables IV and VII), whereas macrocin O-meth- yltransferase is sensitive to them in the following order: sinefungin, A9145C > AdoHcy (1). Thus, sinefungin is a potent inhibitor for macrocin 0-methyltransferase but a poor inhibitor for demethylmacrocin 0-methyltransferase; the Ki. of macrocin 0-methyltransferase for sinefungin (0.06 pM) is three orders of magnitude lower than that of macrocin 0- methyltransferase for this side-chain analogue (77 p ~ ) . The inhibition efficiency of each of the three nucleosides for macrocin 0-methyltransferase (but not for demethylmacrocin 0-methyltransferase) appears to be governed by the side chain of the compound.

When measured in whole cells, demethylmacrocin + ma- crocin conversion (i.e. demethylmacrocin O-methyltransfer- ase-catalyzed 2”’-O-methylation of bound 6-deoxy-~-allose) and macrocin + tylosin conversion (i.e. macrocin O-methyl- transferase-catalyzed 3” -0-methylation of bound 2”‘-0- methylated 6-deoxy-~-allose) are differentially susceptible to inhibition by sinefungin. The synthesis of macrocin from demethylmacrocin by the mutant strain (macrocin O-meth- yltransferase-deficient) of S. frudiae is essentially unaffected (i.e. no accumulated macrocin) by sinefungin, whereas the synthesis of tylosin from macrocin by the wild-type strain is completely inhibited by this compound (Fig. 10). Conversion of demethylmacrocin to macrocin by the mutant strain and accumulation ( i e . lack of metabolism) of macrocin by the wild-type strain when either strain was incubated in the presence of sinefungin can be attributed to low sensitivity of demethylmacrocin 0-methyltransferase and high sensitivity of macrocin 0-methyltransferase, respectively, to the nucleo- tide inhibitor. The distinguishable inhibitions of demethyl- macrocin and macrocin 0-methyltransferases by sinefungin both in uiuo and in vitro as well as the distinctive substrate specificities of the two 0-methyltransferases, as described in the preceding section, are consistent with the metabolic flow of demethylmacrocin + macrocin + tylosin and thus sub- stantiate the notion that the 2’”-OH group of 6-deoxy-~- allose is methylated first and the methylation occurs after attachment of the sugar to tylactone (1, 9, 10) rather than before glycosylation as independently proposed previously (11). It remains to be elucidated whether demethylmacrocin 0-methyltransferase and macrocin 0-methyltransferase or other different 0-methyltransferases can catalyze 2-0- and 3- 0-methylations of free 6-deoxy-~-allose by an additional methylation route prior to glycosylation for tylosin biosyn- thesis in S. fradiae.

Identical Kinetic Reaction Mechanism-The substrate in- teraction pattern for demethylmacrocin O-methyltransferase- catalyzed conversion of demethylmacrocin + macrocin is intercepting (Fig. 5), suggesting that the enzymic methylation follows a sequential mechanism (12). In contrast to the com- petitive inhibition of the macrocin O-methyltransferase-cat- alyzed conversion of macrocin --.* tylosin by demethylmacro- cin and demethyllactenocin (2), the demethylmacrocin 0- methyltransferase-catalyzed conversion of demethylmacrocin + macrocin is inhibited noncompetitively by 23-demycino- syloxytylosin, 23-deoxy-0-mycaminosyltylonolide and O-my- caminosyltylonolide (Fig. 6). Therefore, none of the three

15630 Demethylmacrocin and Macrocin O-Methyltramfermes

TABLE VI1 Inhibition oatterns and constants of demethvlmacrocin 0-methyltransferase

Variable substrate

Demethylmacrocin

AdoMet

Demethylmacrocin AdoMet Demethylmacrocin AdoMet Demethylmacrocin

AdoMet AdoMet Demethylmacrocin AdoMet

Demethyllactenocin AdoMet Demethylmacrocin

AdoMet

Inhibitor

23-Deoxy-0-mycaminosyl-

23-Deoxy-0-mycaminosyl-

0-mycaminosyltylonolide 0-mycaminosyltylonolide 23-Demycinosyloxytylosin 23-Demycinosyloxytylosin

A9145C

tylonolide

tylonolide

A9145C Sinefungin Lactenocin Lactenocin

Macrocin Macrocin AdoHcy

AdoHcy

Fixed substrate"

AdoMet

Demethylmacrocin

AdoMet Demethylmacrocin AdoMet Demethylmacrocin AdoMet

Demethylmacrocin Demethylmacrocin AdoMet Demethylmacrocin

AdoMet Demethyllactenocin AdoMet

Demethylmacrocin

Inhibition patternb K,' Kii'

WM

NC 73 f 16

uc d -

NC 65 f 13 uc - NC 186 f 89 uc - NC' 108 f 28 uc - C 57 f 7 C 77 f 16 NC 24 f 6 NC' 22 f 10 uc - NC 37 f 22 NC 67 f 18 NC' 226 f 49 uc - C 110 f 14

BM

153 f 28

40 f 3

125 f 16 41 f 4 82 f 21

102 29 89 f 19 39 f 8 - -

15 f 2 7 f 0.7 8 f 1

91 * 20 42 f 6 43 * 3

202 f 19 -

AdoMet, 250 p ~ ; demethylmacrocin or demethyllactenocin, 20 /*M except as specified. * C, competitive inhibition; NC, noncompetitive inhibition; UC, uncompetitive inhibition. e Ki,, apparent slope inhibition constant; Kii, apparent intercept inhibition constant.

e The concentration of AdoMet (fixed substrate) was at 62.5 pM. 'The concentration of demethylmacrocin (fixed substrate) was at 10 PM.

-, not applicable.

macrolide substrate analogues is useful in dissecting the sub- strate binding order of demethylmacrocin O-methyltransfer- ase. However, the inhibition of demethylmacrocin O-methyl- transferase-catalyzed conversion of demethylmacrocin + ma- crocin by A9145C is competitive with regard to AdoMet and noncompetitive with regard to demethylmacrocin (Fig. 7). The noncompetitive inhibition by the nucleoside substrate analogue with respect to demethylmacrocin is consistent with AdoMet as the leading substrate of demethylmacrocin 0- methyltransferase (13). Lactenocin, which is used as an alter- native reaction product because of its HPLC separation from macrocin (the normal reaction product), is noncompetitively inhibitory to demethylmacrocin O-methyltransferase-cata- lyzed conversion of demethylmacrocin + macrocin with re- spect to either demethylmacrocin (Fig. 8 A ) or AdoMet (data not shown). The noncompetitive inhibition with variable AdoMet is shifted to the uncompetitive inhibition (Fig. 8B) when the fixed concentration of demethylmacrocin is in- creased from 10 (presumably unsaturated) to 20 PM (presum- ably saturated). The inhibition of the demethylmacrocin 0- methyltransferase-catalyzed conversion of demethylmacrocin + macrocin by AdoHcy, the nucleoside reaction product, is competitive with regard to AdoMet and noncompetitive with regard to demethylmacrocin (Fig. 9). The five kinetic patterns of product inhibitions (Table VII) support AdoMet as the leading substrate and are consistent only with AdoHcy as the last released product in an Ordered Bi Bi mechanism (14), as shown in Scheme 1, for demethylmacrocin O-methyltransfer- ase-catalyzed conversion of demethylmacrocin + macrocin. The intercepting substrate interaction pattern (data not

SCHEME 1. E, demethylmacrocin 0-methyltransferase.

shown) and a noncompetitive inhibition by macrocin (as an alternative reaction product) with respect to either substrate (Table VII) for demethylmacrocin O-methyltransferase-cata- lyzed conversion of demethyllactenocin + lactenocin indi- cates that this alternative enzymic methylation also follows Ordered Bi Bi as its kinetic mechanism. Thus, both deme- thylmacrocin 0-methyltransferase-catalyzed conversions and both macrocin 0-methyltransferase-catalyzed conversions (2) share a common kinetic mechanism. Interestingly, the same kinetic mechanism has also been described recently for both soluble and membrane-bound catechol 0-methyltransferases (15).

Different Versus Identical Kinetic Metabolic Inhibitions- The demethylmacrocin 0-methyltransferase-catalyzed con- versions and the macrocin 0-methyltransferase-catalyzed conversions are susceptible to independent inhibitions by macrolide metabolites (Fig. 11). However, with regard to a macrolide substrate, the inhibition patterns of the two 0- methyltransferases from an early inhibitory metabolite are different and those from a late inhibitory metabolite are identical (Table VI1 and Ref. 2) . For example, the demethyl- macrocin 0-methyltransferase-catalyzed conversion of de- methylmacrocin + macrocin is inhibited noncompetitively by 23-deoxy-0-mycaminosyltylonolide, indicating that the early

Demethylmacrocin and Macrocin 0-Methyltransferases 15631

metabolite may exert its inhibitory effect at a regulatory site of the enzyme. In contrast, macrocin O-methyltransferase- catalyzed conversion of macrocin + tylosin is inhibited com- petitively by demethylmacrocin (2), suggesting that this early metabolite may affect the enzyme at the active site. However, the noncompetitive inhibition in demethylmacrocin O-meth- yltransferase-catalyzed conversion of demethylmacrocin + macrocin by lactenocin and the noncompetitive inhibition in macrocin 0-methyltransferase-catalyzed conversion of ma- crocin + tylosin by desmycosin indicate that either late metabolite (used as an alternative reaction product) may exert its inhibitory effect at a regulatory site on the respective 0- methyltransferase. Since the conversion of the alternative macrolide substrate to its product by each enzyme is kinetic- ally very similar to that of the preferred substrate to its product (Table VI), we conclude that macrocin and tylosin may also exert an inhibitory effect at a regulatory site on demethylmacrocin and macrocin 0-methyltransferases, re- spectively.

Potential Practical Applications-The inhibitory effect on both 0-methyltransferases by macrolide metabolites, regard- less of differential kinetic behaviors, might play a significant regulatory role for tylosin biosynthesis in S. fradiae. We speculate that such metabolic regulation of demethylmacrocin 0-methyltransferase- and macrocin O-methyltransferase-cat- alyzed conversions, if present as indicated from our kinetic inhibition studies in vitro as described in this and the preced- ing papers, could be alleviated by increasing activities of both enzymes possibly by using a “gene cloning” approach (16,17). Furthermore, the putative allosteric regulation of demethyl- macrocin and macrocin 0-methyltransferases, respectively, by macrocin and tylosin, which are present as two main macrolide metabolites from high tylosin-producing strains of S. fradiae, provides a desirable working model to analyze yield improvement of this industrially important antibiotic by site- directed modification of either or both of the two distinctive rate-limiting 0-methyltransferases.

Acknowledgments-We are grateful to Dr. Gene M. Wild for sup- plying highly purified macrolide compounds; Mark L. Slisz and Rich- ard M. Van Frank for estimation of the isoelectric points; Robert M. Ellis for determination of the amino acid compositions; Drs. Richard

H. Baltz, Charles L. Hershberger, and Eugene T. Sen0 for helpful comments on the two manuscripts; Burton L. Hall for artwork; Jerry L. Chapman and Betty Lou Ems for manuscript preparation; and the management of Eli Lilly and Company for continuous support on enzymological development of tylosin biosynthesis. Specially, we thank Dr. C. B. Grissom (University of Berkeley) for providing all computer-aided programs and associated instruction used in the kinetic studies.

REFERENCES 1. Baltz, R. H., Seno, E. T., Stonesifer, J., and Wild, G. M. (1983)

2. Bauer, N. J., Kreuzman, A. J., Dotzlaf, J. E., and Yeh, W. K.

3. Seno, E. T., and Baltz, R. H. (1981) Antimicrob. Agents Chemo-

4. Seno, E. T., and Baltz, R. H. (1982) Antimicrob. Agents Chemo-

5. Yeh, W.-K., Bauer, N. J., and Dotzlaf, J. E. (1984) J. Chromutogr.

6. Bradford, M. M. (1979) Anal. Biochem. 72, 248-254 7. Cleland, W. W. (1979) Methods Enzymol. 63, 103-138 8. Dotzlaf, J. E., and Yeh, W.-K. (1987) J. Bacteriol. 169, 1611-

1618 9. Baltz, R. H., and Seno, E. T. (1981) Antimicrob. Agents Chemo-

ther. 20,214-225 10. Baltz, R. H. (1982) in Genetic Engineering of Microorganisms for

Chemicals (Hollaender, A., ed) pp. 431-444, Plenum Publishing Corp., New York

11. omura, S., Sadakane, N., and Matsubara, H. (1982) Chem. Pharm. Bull. (Tokyo) 30,223-229

12. Cleland, W. W. (1970) in The Enzymes (Boyer, P. D., ed) Vol. 2, pp. 1-65, Academic Press, Orlando, FL

13. Fromm, H. J. (1983) in Contemporary Enzyme Kinetics and Mechanism (Purich, D. C., ed) pp. 233-251, Academic Press, Orlando, FL

14. Segel, I. H. (1975) Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems, pp. 560- 656, Wiley Interscience, New York

15. Jeffery, D. R., and Roth, J. A. (1987) Biochemistry 26, 2955- 2958

16. Cox, K. L., Fishman, S. E., Larson, J. L., Stanzak, R., Reynolds, P. A., Yeh, W. K., Van Frank, R. M., Birmingham, V. A., Hershberger, C. L., and Seno, E. T. (1986) J. Nat. Prod.

17. Fishman, S. E., Cox, K. L., Larson, J. L., Reynolds, P. A., Seno, E. T., Yeh, W. K., Van Frank, R. M., and Hershberger, C. L. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 8248-8252

J. Antibiot. (Tokyo) 36 , 131-141

(1988) J. Biol. Chem. 263 , 15619-15625

ther. 20,370-377

ther.21,756-763

288,157-165

(Lloydia) 49,971-980

SUPPLEMENTARY MATERIALS TO

Two DIStInnIVe 0-MethylIransterases Catalymg Penult8mate and Termlnal Reacmns 01 MacrOINde Anllbmtc (Tylosin) B1OSynthesis. Substrate Spec~l#c~ly. Enzyme lnh8bItlOn and Klnetlc Mechanism

Adam Kreurman, Jan R Turner. and Wu-Kuang Yeh BY

EXPERIMENTAL PROCEDURES was determlned by uslng demethyllactenocm (rather than demethylmacroc~n) as the Substrate and

and the preparation 01

”. Unless specitled otherwise, the acllvlty 01 demethylmacroc~n 0-methylIransterase was determined by usmg demethylmacrocin as the Substrate and monltorlng macroon format8on at 285 nm wlth HPLC as modltied from the macroctn 0-methyltransterase assay (2. 5) A typical reaction m<xture

demethylmacrocm 0 15 pmol 01 AdoMet. IO mmol 01 MgClz and 0 2-3 OxtO-‘ unit (as detlned below) 01 of 0.5 ml lorthe demethylmacrocm O-methynranslerase assay contained 0.01 5 pmolot

the enzyme In 100 mM KH2P0,. pH 7 5. The enzymatic reactton was carned out at 42°C tor 5 mln Macrocr lormatoon was hear wllh reacmn tlme One unlt 01 the acI1v8ty 8s detlned a5 the amount 01 the enzyme to cause lormat8on 01 one p o l of maCrOCm p e r m r trom demethylmacrocr Spec~l~c anlvlty is delined as units per mg 01 protein. The plotem content was determlned by the method 01 Bradford (6) usmg bovme Semm albumm as the standard E@zvme Purilicqtlpn. The same crude extranot S kxbe was used as the enzyme source for both demethylmacroc8n and macrmin 0-methyltransterases (2) Demethylmacroc8n 0-methylIransterase was separated from rnacrocm 0-methynransterase by DEAE-Trisacryl chromatography. and tunher purlfled by Sephacryl S-200 gel bnrat~an. hydroxylapatlte chromatography. and Mono (I FPLC under ldentlcai condilms as descrlbed tor the putil~camn 01 maCrOCln 0-methyltransteiase (2) A ~ ~ “ g l e adlwty peak 01 demethylmacroc8n 0-methyttransterase was observed aner each 01 the this1 three Chmmatograph~c steps, however. two ma,” adtvtty peaks appeared aner Mono 0 FPLC Only the Mono Q pool wlth the higher spec~tlc anlwly was puntled tunher by gel flltrat80n with a Bso-Gel A0 5m column (1 .Ex100 cm). the activity and protein proflies are Shown ~n Flg 2 K!!EIESU~&~ Steady-state hfnetbc studies and adlwty analyses 01 DMOMT-catalyzed convers~ons were mnduned simllarly to those 01 MOMT~catalyzed convers~ons (21. In the absence 01 an Inh8bltor the Chmethyltransterase anlvlty was determrned by usmg demethylmacrmn or demethyllactenoc8n 8s the substrate and monttomg macrocin or lactenwin lormatlon. respectively The acl8wty data were analyzed graphlcally and by the SEQUENO program 01 Cleland (7) In the presence 01 an #nh#b,lor the O~methyltransterase aalvity was determmed by manitomg macrmn foimal~on with demethylmacroan as the Substrate. except lor macrocln lnhlbmon stud8es I“ which the en:yme a c w l y

The sources anWorqualltles 01 chemlcal compounds, the prouders ot chromalographlc m a t e m s hd!ascells are the Same as those desmbed ~n the preceding

paper (2).

determined by using demethylmacrocin (ratherthan demethyllactenocln) as the Substrate and monitonng Ianenocln formallon, and lor lactenocln lnhiblt80n studies ~n Which the enzyme acI1wty was

monltonng macrocln Iormatlon. Reproduc8ble measurement 01 product tormallon by HPLC lor the enzyme anlvlty dlctated use 01 an alternat8ve product ~n ellher macrmn or lactenoc8n mhrb#t#On studles Macrwm ladenocln and AdoHCy were used as produn Inhfbltors: whereas

and Stnetungin were used as dead-end mhlbltors The inh8bIllOn data were analyzed by the COMPO 0-mycaminosynylonobde. 23deaxy-O-mycam~nosynylonollde. 23-demyclnosylaxytylasln ~ 9 1 4 5 C

NCOMP and UNCOMP programs of Cleland (7)

S. hd!as were used in this study. a moderate tylosln-producmg stialn and a mutant s l r m detlclent ~n the activity of macmcin 0-methyltransterase (3.4). Complex vegetatlve and termentatton medla were used to prepare cells tor examlnatlon (3). The inoculum lor each s t r m was stored ~n liqurd nltrogen qulchly thawed and used to inmulate the vegetatlve medlum Aner mubatton at 30% lor 68 hr. an aliquot 01 each cuiture was used to mocuiate the termentatlon medwm The termentat~on was Incubated at 30-C. Smetungin was dissolved 0.2 M potass~urn phosphate. pH 7 0. sterIl#zed through a Mlllex-SR tlner (0.5 micron, Mllllpore). and added to the termentallon at the tlme 01 lnOCUlation lo a final COnCentratlOn 01 0.2 mgiml Duplicate CuIlures were prepared lor the w8ld~type and the mutant

f m h s p l ~ C o n v e n l o n s h Y l V P b y S ~ i n e t u n o ~ n Twostra1n5oI

Samples (1 ml) were removed from the termentatlons at the tlme 01 moc~lat~on and every 24 hr thereaner Up 10 144 hr. Methanol 12 ml) was added to the sampler. lollowed by mlxing. cenInlugat8on and 1~ltrat~on 01 the supernatant tractton through a M~llex-SR t~lter HPLC analys~s was pedormed uslng a C-18 reversed phase column i t 5 crn x 4 6 mm. Waters ASSOC ) which was developed w8Ih a 50 Io 80% gradlent 01 methanol.0.3% ammonium formate butler p~ 4 0 The eluate was mon#tOred at 282 nm and macrolide compunds were quantitated from absorption peaks with a ~ewlet t Pachard

QLkIm. Procedures lor SDS-PAGE. 85Oelectr1c point estimation and amino acid analys~s are 3390A mtegrator.

those as described prev8ously (8 )

15632 Demethylmacrocin and Macrocin 0-Methyltransferases RESULTS 60 T

0.20

1 0.15

E 2 0.10

" * 45 ..

- 31 L1 21

14 1 2 3

1 .oo

0.75 1 P E

0.50 - v

.- 0,

C

0 L

0.25 &

"1 20 \ ,DMOMT

$ 10

s - 6 x 8

2

= 4 ! 0 0.3 , \ , 0.4 0 5 0.6 0.7 0.8

-1 Macrocin 1 P

Demethylmacrocin a

I I I I

2 4 6 8 10 Time (min)

Flq 4 Slo,ch~omelr~c analys6 01 DMOMT catalyzed demelhylmaCroCm - maCIOCln COnvelSlOn

Demethylmacrocin and Macrocin 0-Methyltransferases 15633

20 , I

7 15 1

0.00 0.04 0.08 0.12 0.000 0.003 0.006 0.009

lI[Demethylmacrocin), p M - 1 lIIAdoMetj, p M - 1

Fag 6 Dead-end inhlbttmn 01 DMOMT-catalyzed demethyimacrocln-macrocln conversion by 0-mycam,nosyilyionoilde. The inhlbltor concentrat~on~ 01 0 (0)

25 io!. 50 (XI. 75 (m], and 100 pM (A) were used wllh demethylmacracr (A! or AdoMet (81 as the varlable Substrate AdOMel Concentralton was Itxed at 250 !,M iA1, whereas demethylmacrocln cOncentral~on was f w d at 20 )rM (81

2o IA. A 1 I4]E. 12 A

0.000 0.005 0.010 0.015 0.020 0.00 0.04 0.08 0.12 I I [ A d o M e t ] , p M - 1 lIIDemethylrnacrocin],pM-l

FIQ 7 Dead end lnhlbrtlon 01 DMOMT-catalyzed demethylmactoc,n-macrocln

(81. and 100 g M (r! were used w l h AdoMet (A! or demelhylmacrocr iB1 as the COnverSlOn by A9145C The inhlbllor on cent rations 01 0 (b). 25 (0). 50 (x). 75

varlable Subslrale Dernethyimacracln COnCentral~On was Itxed at 20 gM (A! whereas AdoMet COnCentratlOn was fixed at 62 5 pM (Bl

;&ciIic (though less than stoichiometnc) aCC~mulal~Oti 01 microc$n (Fig lOA) 'Theretbre.

conversion m w was selectively and enlclently blocked by lhw lnhlbmr The mutant Straln 01 S MOMT-catalyzed methylation a1 3 ~ O H 01 2-0-methylated 6-deoxy-D-allose tn macioc~n4ylosr

was deliclent on the acttvlty 01 macrocln 0-methyltranslerase (data not shown) and. ~n the absence of sinetungin. accumulated macmcln wllh no tylosm as expected (4) In the presence of

was detectable. Thus, DMOMTcalalyzed methylation at ?"'-OH 01 6-deOxy-D~alloSe I"

sinetungin, macrocm synthesis was not affected (Flg 108) and no demethylmacroc8n accumulat~on

demethyimacrocln~macrocln converson was not select8vely blocked by th8s Inhibitor

E. Mutant s t ra in I i

50 100 Time (hr)