Biosynthesis of phenazine antibiotics in Streptomyces antibioticus: Stereochemistry of methyl transfer from carbon-2 of acetate

Article abstract of DOI:10.1021/ja991159i

Stable isotope labeling experiments have shown that the biosynthesis of the monomeric phenazines, the saphenyl esters, and their dimerization products, the esmeraldins, in Streptomyces antibioticus Tu 2706 proceeds from phenazine-1,6-dicarboxylic acid by

Full text of DOI:10.1021/ja991159i

VOLUME 121, NUMBER 24
J UNE 23, 1999
© Copyright 1999 by the
American Chemical Society
Biosynthesis of Phenazine Antibiotics in Streptomyces antibioticus:
Stereochemistry of Methyl Transfer from Carbon-2 of Acetate
Matthew McDonald, Barrie Wilkinson,† Clinton W. Van’t Land,‡ Ulla Mocek,§
Sungsook Lee,| and Heinz G. Floss*
Contribution from the Department of Chemistry, Box 351700, UniVersity of Washington,
Seattle, Washington 98195-1700
ReceiVed April 12, 1999
Abstract: Stable isotope labeling experiments have shown that the biosynthesis of the monomeric phenazines,
the saphenyl esters, and their dimerization products, the esmeraldins, in Streptomyces antibioticus Tu¨ 2706
proceeds from phenazine-1,6-dicarboxylic acid by chain extension with C-2 of acetate to 6-acetylphenazine-
1-carboxylic acid, which is reduced to saphenic acid. The latter is incorporated into both halves of the
esmeraldins, albeit differentially. By feeding of chiral acetate, degradation of the resulting saphenyl esters and
esmeraldins, and configurational analysis of the acetic acid formed, the chain extension process was found to
proceed with overall inversion of configuration at the methyl group. This suggests that the decarboxylation of
a hypothetical intermediate â-keto acid proceeds in an inversion mode. This result is discussed with reference
to analogous C-methylations of polyketide backbones by addition of C-2 of acetate.
Phenazines are a group of natural products which are quite
common in many bacteria; most of its about 75 members are
produced by Pseudomonas species, but a few are also found in
Streptomycetes.1,5 Streptomyces antibioticus, strain Tu¨ 2706,
produces a series of dimeric phenazine derivatives (Figure 1),
the intensely green esmeraldins A (1b) and B (1c),2 as well as
simpler phenazines such as saphenic acid (2a) and its esters
(2b and 2c).3 Several of the compounds showed antibacterial
activity and anticancer activity against murine tumors.2 Deg-
radation and derivatization of the esmeraldins allowed their
separation into a pair of diastereomers, esmeraldic acid dimethyl
esters I and II (3 and 4), which are epimeric at N-15. NMR
analysis of their Mosher’s esters further revealed that each
compound in turn was a mixture of epimers at C-22, but with
a single configuration at C-25.4
As reported previously,4 the biosynthesis of the esmeraldins
involves the same head-to-tail dimerization of an unknown
metabolite of the shikimic acid pathway that has been demon-
strated for the phenazines elaborated by Pseudomonas sp.5-10
* To whom inquiries should be directed. Phone: (206)543-0310. Fax:
(206)543-8318. E-mail: floss@chem.washington.edu.
† Present address: Glaxo Wellcome Medicines Research Centre, Bio-
processing Unit, Gunnells Wood Road, Stevenage, Hertfordshire, SG1 2NY,
UK.
‡ Present address: Alcon Laboratories, Inc., Ophthalmic Products-
Biodisposition, 6201 South Freeway (MS: R3-2B), Fort Worth, TX 76134-
2099.
§ Present address: Thetagen, 11804 North Creek Parkway, South, Bothell,
WA 98011-8805.
(3) Geiger, A.; Keller-Schierlein, W.; Brandl, M.; Za¨hner, H. J. Antibiot.
1988, 41, 1542.
(4) Van’t Land, C.; Mocek, U.; Floss, H. G. J. Org. Chem. 1993, 58,
6576.
(5) Turner, J. M.; Messenger, A. J. AdV. Microb. Physiol. 1986, 27, 211.
(6) Herbert, R. B.; Holliman, F. G.; Ibberson, P. N.; Sheridan, J. B. J.
Chem. Soc., Perkin Trans. 1 1979, 2411 and references therein.
| Present address: CEREP, Inc., 15318 NE 95th Street, Redmond, WA
98052.
(1) Ingram, J. M.; Blackwood, A. C. AdV. Appl. Microbiol. 1970, 13,
267.
(2) Keller-Schierlein, W.; Geiger, A.; Za¨hner, H.; Brandl, M. HelV. Chim.
Acta 1988, 71, 2058.
10.1021/ja991159i CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/05/1999

5620 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
McDonald et al.
Figure 1. Structures of esmeraldins and saphenic acid derivatives from Streptomyces antibioticus Tu¨ 2706.
and by some other Streptomycetes.11,12 The product, presumably
phenazine-1,6-dicarboxylic acid (6),13 is then further converted
into saphenic acid 2a or a derivative thereof by addition of a
one-carbon unit, a methyl group derived from C-2 of acetate.
Such one-carbon extensions of various structural backbones,
particularly polyketides, have been encountered in the biosyn-
thesis of a number of compounds, such as myxovirescin A1,14
virginiamycin,15 myxopyronin A,16 aurantinin,17 pseudomonic
acid,18,19 oncorhyncolide,20 cylindrophane,1 and pulvomycin.22
Preliminary evidence supports the notion that 2a in turn is
converted into 1b and 1c by dimerization and esterification.4
In the present paper we report experiments which provide
further support for the proposed pathway by which the esmer-
aldin structures are assembled and which shed some light on
the one-carbon extension process by determining its steric course
at the acetate-derived methyl group.
Scheme 1. Proposed Pathway for the Biosynthesis of the
Dimeric Phenazines, the Esmeraldins, in Streptomyces
antibioticus Tu¨ 2706
Results
A plausible pathway by which C-2 of acetate can be converted
into the methyl groups of the esmeraldin (1) and saphenamycin
(2) type compounds is outlined in Scheme 1. Acetate as its
coenzyme A thioester derivative is converted by acetyl-CoA
carboxylase to malonyl-CoA, which then undergoes a decar-
boxylative Claisen condensation with the mono-CoA thioester
of phenazine-1,6-dicarboxylic acid 6. Hydrolysis of the resulting
thioester and decarboxylation of the â-keto acid produces
6-acetylphenazine-1-carboxylic acid (7), which is then reduced
to saphenic acid 2a. Support for this pathway was ascertained
by feeding stable isotope-labeled phenazine-1,6-dicarboxylic
acid 6, 6-acetylphenazine-1-carboxylic acid 7, and saphenic acid
2a to Streptomyces antibioticus Tu¨ 2706 and demonstrating their
specific incorporation into esmeraldins 1b/1c and/or saphenyl
esters 2b/2c. Labeled 6 carrying 99% 13C in one carboxyl group
was synthesized from [7-13C]benzoic acid and 2-chloro-3-
nitrobenzoic acid by the method of Holliman and co-workers.23
Feeding of this material (50 mg to ten 100 mL shake cultures)
produced 1, which was degraded4 to esmeraldic acid dimethyl
esters 3 and 4 (16.2% excess M + 1 and 2.2% excess M + 2
species by MS). These showed enhanced 13C NMR signals for
C-21, C-22, C-24, and C-25. A second feeding experiment with
[carboxy-13C1]-6 (20 mg to five 100 mL cultures) produced 2
isolated as 5 (15 mg) after base hydrolysis and esterification
with diazomethane, which displayed enhanced 13C NMR signals
at C-11 and C-12 (0.5 and 0.7% enrichment), whereas MS
analysis indicated the presence of 2.1% excess M + 1 species.
The esmeraldins from the same experiment, isolated similarly
as 3 (1.0 mg) and 4 (1.2 mg), showed 1.8% excess M + 1
species by MS. The amounts of sample and enrichments were
too small to obtain meaningful 13C NMR data.
(7) Etherington, T.; Herbert, R. B.; Holliman, F. G.; Sheridan, J. B. J.
Chem. Soc., Perkin Trans. 1 1979, 622.
(8) Messenger, A. J.; Turner, J. M. Biochem. Soc. Trans. 1978, 6, 1326.
(9) Hollstein, U.; Mock, D. L.; Sibbitt, R. R.; Roitch, U.; Lingens, F.
Tetrahedron Lett. 1978, 2987.
(10) Hollstein, U.; McCamey, D. A. J. Org. Chem. 1973, 38, 3415.
(11) Gulliford, S. P.; Herbert, R. B.; Holliman, F. G. Tetrahedron Lett.
1978, 195.
(12) Buckland, P. R.; Herbert, R. B.; Holliman, F. G. J. Chem. Res. (M)
1981, 4219 and references therein.
(13) Podojil, M.; Gerber, N. N. Biochemistry 1970, 9, 4616.
(14) Trowitzsch, W.; Gerth, K.; Wray, V.; Ho¨fle, G. J. J. Chem. Soc.,
Chem. Commun. 1983, 1174.
(15) Kingston, D. G. I.; Kolpak, M. X.; LeFevre, J. W.; Borup-
Grochtmann, I. J. Am. Chem. Soc. 1983, 105, 5106.
(16) Kohl, W.; Irschik, H.; Reichenbach, H.; Ho¨fle, G. Liebigs Ann.
Chem. 1984, 1088.
(17) Nakagawa, A.; Konda, Y.; Hatano, A.; Harigaya, Y.; Onda, M.;
Omura, S. J. Org. Chem. 1988, 53, 2660.
(18) Martin, F. M.; Simpson, T. J. J. Chem. Soc., Chem. Commun. 1989,
207.
(19) Mantle, P. G.; Macgeorge, K. M. J. Chem. Soc., Perkin Trans. 1
1991, 255.
(20) Needham, J.; Anderson, R. T.; Kelly, M. T. J. Chem. Soc., Chem.
Commun. 1992, 1367.
(21) Bobzin, S. C.; Moore, R. E. Tetrahedron 1993, 49, 7615.
(22) Gro¨ger, S.; Priestley, N. J. Org. Chem. 1995, 60, 4951.
(23) (a) Challand, R. S.; Herbert, R. B.; Holliman, F. G. Chem. Commun.
1970, 1423. (b) Flood, M. E.; Herbert, R. B.; Holliman, F. G. J. Chem.
Soc., Perkin Trans. 1 1972, 622.

Biosynthesis of Phenazine Antibiotics
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5621
6-Acetylphenazine-1-carboxylic acid 7 bearing three deute-
rium atoms (98% H) in its methyl group was synthesized by
synthase/fumarase assay,26 which indicated a 45% enantiomeric
excess of S configuration. To confirm this result, sodium (S)-
[2-2H1,3H]acetate (3.84 mCi) was fed to two 100 mL cultures
of S. antibioticus in the same way and both 2, isolated as 5
(1.36 µCi), and 1, isolated as 3 (0.23 µCi) and 4 (0.16 µCi),
were obtained. Degradation of 5 produced sodium acetate in
12.5% yield, which assayed for 18% ee R isomer in the
configurational analysis.27 Degradation of 3 from the same
experiment gave sodium acetate (22% yield), which analyzed
for 42% ee R configuration. Thus, the conversion of the methyl
group of acetate into the C-methyl groups of 1 and 2 proceeds
with predominant overall retention of configuration accompanied
by a slightly more than 50% decrease in configurational purity.
2
exchange of the unlabeled compound with D2O and CD3OD in
the presence of NaOD; part of this material was further
converted to saphenic acid 2a bearing four deuterium atoms in
the hydroxyethyl group by reduction with NaB2H4. 6-[13-2H3]-
Acetylphenazine-1-carboxylic acid (17 mg) was fed to five 100
mL cultures and gave 1 isolated as esmeraldin dimethyl esters
3 and 4 (8 and 10 mg, respectively) carrying deuterium in both
C-methyl groups, C-23 and C-26, as shown by deuterium NMR
and ES-MS. Mass spectrometry indicated about 6% enrichment
in each C-methyl group, and careful analysis of the deuterium
NMR spectrum by line-fitting revealed, within the limits of
accuracy, equal enrichment in the two methyl groups (3,
intensity ratio C-26:C-23 1.0:0.988; 4, ratio C-26:C-23 ) 1.0:
0.9). Similarly, [12,13-2H4]saphenic acid (50 mg) was fed to
ten 100 mL cultures and gave esmeraldic acid dimethyl esters
3 and 4 (8 and 10 mg, respectively) showing a deuterium
distribution by MS of 0.8% D2, 4.2% D3, 14.7% D4, 0% D5,
0.4% D6, 1.4% D7, and 2.0% D8 species. Deuterium NMR
showed an intensity ratio of the methyl to the methine signal
of 3.94:1. Line-fitting analysis of the deuterium NMR spectrum
of the labeled 4 from this experiment indicated that [12,13-
2H4]-2a labels the two halves of 1 unequally (intensity ratio
C-26:C-23 ) 1.0:1.53 and C-25:C-22 ) 1.0:1.85). Since the
precursor was incorporated predominantly (70-80%) with
retention of all four deuterium atoms, it is evident that (a) 2a is
incorporated directly, not by way of 7, (b) it is incorporated
preferentially into the “Eastern” half of 1 and (c) both methyl
groups remain intact during the dimerization process.
Discussion
The results of the feeding experiments with the stable isotope-
labeled compounds 6, 7, and 2a provide considerable support
for the pathway of esmeraldin assembly proposed in Scheme
1. The fact that deuterated 2a labels the two halves of 1
differentially suggests that the dimerization reaction to produce
the phenazine ring system must involve two different phenazines
rather than dimerization of two identical molecules, such as 2a.
It is peculiar that such unsymmetrical labeling of 1 is not seen
with deuterated 7 as precursor, but this may be a fortuitous result
of the kinetics of the fermentation, i.e., the time courses of
incorporation of isotope into the two halves may just accidentally
cross over at the time of harvest. One could speculate that the
two molecules undergoing dimerization to the esmeraldin ring
system are saphenic acid, giving rise to the “Northwestern” half,
and a saphenyl ester, giving rise to the “Eastern” half. However,
the feeding experiment with saphenic acid myristoyl ester 2d
provides no support for such a process, although its negative
outcome could be due to lack of cellular uptake of the precursor.
Thus, although we cannot rule out the possibility that one of
them is a saphenyl ester, the exact nature of the two molecules
undergoing dimerization to 1 is not known at this time.
The experiments with chiral acetate shed some light on the
process by which one of the carboxyl groups of 6 is extended
to a two-carbon unit by addition of C-2 of acetate to give the
C-methyl groups of 7, 2, and 1. Scheme 2 elaborates on a
plausible mechanism for this process, involving carboxylation
of acetyl-CoA (8) to malonyl-CoA (9), decarboxylative Claisen
condensation with the coenzyme A thioester of 6, followed by
thioester hydrolysis and decarboxylation. The reaction stereo-
chemistry of several members of the three classes of enzymes
presumed to act upon the methyl group in this sequence has
been studied in other systems using the chiral methyl group
methodology. Sedgwick and Cornforth28 have shown that acetyl-
CoA carboxylase replaces a hydrogen in the methyl group of
acetyl-CoA by COOH with retention of configuration and a
negligible deuterium isotope effect. Numerous enzymes cata-
lyzing Claisen condensations have been shown without excep-
tion to operate with inversion of configuration at the methyl
group.26c However, â-decarboxylases do not show such stereo-
chemical consistency, as examples of both inversion and
retention of configuration have been found.29 In the acetyl-CoA
The facts that esmeraldic acid 1a has not been isolated from
the esmeraldin fermentation and that esmeraldin A and B, 1b
and 1c, carry the same acyl substituents as the saphenyl esters
2b and 2c suggested the possibility that the Eastern half of 1
may be derived more directly from the saphenyl esters than from
2a. This could explain the unequal incorporation of [12,13-2H4]-
2a into the two halves of 1, although it would be more consistent
with lower enrichment in the “Eastern” half than with the
observed higher enrichment. To examine this point, we prepared
the myristoyl ester of [12,13-2H4]-2a, [12,13-2H4]-2d, by
acylation of [12,13-2H4]-2a with myristoyl chloride according
to the procedure of Bahnmu¨ller et al.24 [12,13-2H4]-2d (8.5 mg)
was fed to five 100 mL cultures and the resulting mixture of
esmeraldins was analyzed by mass spectrometry. The analysis
gave no evidence for the presence of myristoyl-esmeraldin A
1d in the mixture. Deuterium NMR of the derived esmeraldic
acid dimethyl esters, 3 and 4, showed no detectable signal for
deuterium at either C-22 or C-23. This experiment thus provided
no indication for the direct incorporation of saphenyl esters into
the esmeraldins.
The finding that the methyl groups of [12,13-2H4]-2a were
incorporated intact into 1 made it possible to examine the steric
course of the one-carbon chain extension from C-2 of acetate.
To this end, sodium (R)-[2-2H1,3H]acetate (8.58 mCi)25 was
administered to a 100 mL culture of S. antibioticus Tu¨ 2706.
The resulting 2, isolated as 5 (0.24 µCi), was purified to constant
specific radioactivity and degraded by ozonolysis and oxidation
with KMnO4/K2CO3 to carve out the hydroxyethyl group as
sodium acetate. The purified sodium acetate (25% radiochemical
yield) was subjected to configurational analysis by the malate
(26) (a) Cornforth, J. W.; Redmond, J. W.; Eggerer, H.; Buckel, W.;
Gutschow, C. Eur. J. Biochem. 1970, 14, 1. (b) Lu¨thy, J.; Retey, J.; Arigoni,
D. Nature 1969, 221, 1213. (c) Cf.: Floss, H. G.; Tsai, M.-D. AdV. Enzymol.
Relat. Areas Mol. Biol. 1979, 50, 243.
(27) This sodium acetate sample was inadvertently exposed to base for
a prolonged period of time, which probably caused substantial racemization.
(28) Sedgwick, B.; Cornforth, J. W. Eur. J. Biochem. 1977, 75, 465.
(29) (a) Rose, I. A.; Hanson, K. R. In Applications of Biochemical
Systems in Organic Chemistry; Jones, J. B., Sih, C. J., Perlman, D., Eds.;
Wiley: New York, 1976; Part II, p 507 and references therein.
(24) Bahnmu¨ller, U.; Keller-Schierlein, W.; Brandl, W.; Za¨hner, H.;
Diddens, H. J. Antibiot. 1988, 41, 1552.
(25) Kobayashi, K.; Jadhav, P. K.; Zydowsky, T. M.; Floss, H. G. J.
Org. Chem. 1983, 48, 3510.

5622 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
McDonald et al.
Scheme 2. Proposed Mechanism and Stereochemical Course
of the Conversion of C-2 of Chiral (S)-Acetate into the
Methyl Groups of the Saphenic Acid Derivatives and
Esmeraldins in Streptomyces antibioticus Tu¨ 2706a
Scheme 3. Hypothetical Mechanisms for the One-Carbon
Extension of Polyketides from C-2 of Acetate
enzymatic aldol condensations proceed in a retention mode.26c,29,30
Unlike in the phenazine case, the nature of the subsequent steps
is uncertain and may in fact differ in different systems. Two
mechanisms have been proposed in the literature (Scheme 3).
For the introduction of the acetate-derived methyl groups
attached to C-2 and C-14 of cylindrocyclophane D, Bobzin and
Moore21 propose a decarboxylation/dehydration of the initial
aldol product to generate an exo-methylene group from C-2 of
the acetate, which is then reduced to the methyl group (path
a).31 Experiments with deuterated acetate support this mecha-
nism The overall stereochemical outcome in this case cannot
be predicted and may not be mechanistically informative. An
alternative mechanism has been proposed for the methylation,
from C-2 of acetate, at C-7 and C-11 of the pulvomycin
polyketide backbone.22 Aldol condensation was postulated to
be followed by dehydration to an R,â-unsaturated polyketide
in which the stage is set for a vinylogous â-decarboxylation to
give the methyl group (path b). On the basis of stereochemical
precedence, such a mechanism would probably involve aldol
condensation with retention and decarboxylation with inversion,
resulting in overall retention of the methyl group configuration.
To the best of our knowledge these reactions have not yet been
examined for their stereochemical outcome.
aHydrogen isotopes in parentheses denote the expected results if a
deuterium is abstracted by the acetyl-CoA carboxylase rather than a
hydrogen.
carboxylase reaction the removal of H or D from the chiral
methyl group without an isotope effect will produce two
isotopomers of 9, one containing H and T and the other D and
T, in a 1:1 ratio. In the final decarboxylation reaction the latter
will give a chiral methyl group and the former an achiral one;
thus, the configurational purity of the final methyl group should
be decreased to 50% of that of the starting material. The
experimental values, 45 and 42% ee, are very close to this
prediction. On the basis of the stereochemical precedence, a
sample of (S)-acetate will give malonyl-CoA 9 of predominantly
S configuration, which upon decarboxylative Claisen condensa-
tion will produce 10 of predominantly R configuration. To
generate a chiral methyl group of predominantly R configuration
in the final products, the observed overall stereochemical
outcome from (S)-acetate as substrate, the â-decarboxylation
of 10 must proceed with inversion of configuration (Scheme
2). The complementary result with (R)-acetate confirms this
conclusion.
The mechanism of the one-carbon extension from C-2 of
acetate in the formation of 1 and 2 may be unique and somewhat
different from those of the analogous processes encountered in
various polyketide biosyntheses.14-22 Whereas the process
described here must involve a decarboxylative Claisen conden-
sation between two acyl derivatives, the one-carbon chain
extensions of polyketides occur by connecting C-2 of acetate
to a carbonyl group of the polyketide chain. They should
therefore involve a mixed aldol reaction, which would be
predicted to proceed with retention of configuration, since
Experimental Section
1
2
General Materials and Methods. H, H, and proton-decoupled
13C NMR spectra were recorded on Brucker AF-300 (1H 300 MHz) or
WM-500 (1H 500 MHz) spectrometers at 298 K. Chemical shifts in
parts per million (ppm), using residual solvent signal as internal standard
(1H, 2H: 7.24 (singlet) for CDCl3, 3.30 (quintet) for 2H4-methanol, 11.50
(singlet) for CF3COOD; 13C: 77.0 (t) for CDCl3, 49.0 (septet) for
methanol, 116.4 (quartet) and 164.6 (quartet) for CF3COOD), are given
on the δ scale. Coupling constants (J) are given in hertz (Hz). The
proton-decoupled, inverse-gated 13C NMR spectra were obtained by
setting the delays at either 4 or 5 s. Mass spectral (MS) data were
determined on a Micromass Ltd. Quattro II Tandem Quadrupole
electrospray ionization mass spectrometer or on a Micromass 80SEQ
Tandem Hybrid mass spectrometer for FAB-MS. Column chromatog-
raphy was performed with use of Merck 60 µm silica gel. Preparative
TLC was performed with use of Merck silica gel 60 F-254 plates (2
mm precoated). Α Hewlett-Packard HP-5890A series II gas chromato-
graph (GC) with a capillary column coupled to a Hewlett-Packard
5971A mass selective detector was used for GC-MS. HPLC was carried
(30) Meloche, H. P.; Monti, C. T. J. Biol. Chem. 1975, 250, 6875.
(31) Cf.: O’Hagan, D. Nat. Prod. Rep. 1995, 12, 1.

Biosynthesis of Phenazine Antibiotics
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5623
Table 1
precursor
amount fed
culture vol
metabolite produced
[11-13C]-6 (1st feeding)
[11-13C]-6 (2nd feeding)
[13-2H3]-7
50 mg
20 mg
17 mg
50 mg
8.5 mg
8.58 mCi
3.84 mCi
1000 mL
500 mL
500 mL
1000 mL
500 mL
100 mL
200 mL
1.2 mg of 3; 1.1 mg of 4
1.0 mg of 3; 1.2 mg of 4; 15 mg of 5
8.0 mg of 3; 10 mg of 4
8.0 mg of 3; 10 mg of 4
6.8 mg of 3, 5.7 mg of 4
0.24 µCi of 5 (97.5 µCi/mmol)
1.36 µCi of 5 (35 µCi/mmol)
0.23 µCi of 3 (52.3 µCi/mmol)
0.16 µCi of 4 (46.0 µCi/mmol)
[12,13-2H4]-2a
[12,13-2H4]-2d
(R)-[2-2H1,3H]acetate
(S)-[2-2H1,3H]acetate
out on a Beckman dual pump System Gold HPLC equipped with a
Beckman UV detector module monitoring at 254 and 365 nm.
Radioactivity was measured in Atomlight (DuPont) or Biosafe II
(Research Products International Corp.) liquid scintillation fluid with
a Beckman LS-1801 liquid scintillation counter with [3H]- or [14C]-n-
hexadecane (Amersham) as internal standards. Melting points were
determined on a Mel-Temp apparatus and are uncorrected. Streptomyces
antibioticus Tu¨ 2706 was a gift from Prof. Hans Za¨hner, University of
Tu¨bingen, Germany.
10 PAC column (10 µm, 9.6 × 250 mm) with CH2Cl2/EtOAc 9:1 as
the mobile phase at a flow rate of 2.5 mL/min to give 2.4 mg of 3
(52.3 µCi/mmol) and 1.9 mg of 4 (46.0 µCi/mmol).
Degradation of Saphenic Acid Derived from (R)-[2-2H1,3H]-
Acetate. The biosynthetically labeled saphenic acid methyl ester 5 (1.4
× 105 dpm)4 was diluted with unlabeled material (6.8 mg, 0.02 mmol),
dissolved in 10 mL of methanol, and cooled to 0 °C. Ozone was bubbled
through the cold solution for 45 min, after which it was warmed to
room temperature and purged with argon for 5 min. Solvent was
removed in vacuo and the residue dissolved in 2 mL of water and stirred
at 0 °C. A solution of potassium carbonate (69 mg, 0.5 mmol) and
potassium permanganate (79 mg, 0.5 mmol) in 5 mL of water was
added, and the resulting mixture warmed to room temperature with
stirring. After 20 min, another 5 mL of the basic permanganate solution
was added to the brown suspension and stirring was continued for an
additional 40 min. This mixture was adjusted to a pH of 2 with 2 N
sulfuric acid and volatiles were isolated by bulb-to-bulb lyophylization.
To the collected volatiles, 0.3 mL of 50% sulfuric acid and 110 mg of
mercuric sulfate were added, and the solution was heated to reflux for
1 h to oxidize any formic acid present. After the solution was cooled
to room temperature, the pH was adjusted to 10 with 5 N NaOH and
the solution was freeze-dried. The residue was taken up in 5 mL of
water and adjusted to pH 2 with 50% sulfuric acid, and the volatiles
were again isolated by bulb-to-bulb lyophilization. The collected
volatiles were basified to pH 10 with 5 N NaOH and the solution was
lyophilized to yield sodium acetate (3.5 × 104 dpm, 25% overall
radiochemical yield). This procedure was repeated on the methyl
saphenate (2.2 × 106 dpm) and the combined esmeraldic acid dimethyl
esters (8.9 × 105 dpm) isolated from the (S)-[2-2H1,3H]-acetate feeding
experiment and yielded 3.8 × 105 (12.5%) and 1.95 × 105 dpm (22%)
of sodium acetate, respectively. The chirality of the methyl groups in
these acetate samples was analyzed by using the coupled malate
synthase/fumarase assay developed by Cornforth, Arigoni, and their
co-workers.26
Feeding Experiments. Production cultures of Streptomyces antibi-
oticus Tu¨ 2706 were grown and their phenazine metabolites harvested
as previously described.4 After 4 days of growth of the production
culture, appearance of a faint green color indicates the beginning of
esmeraldin production. At that point (approximately 100 h after
inoculation), half of the precursor (dissolved in water or dilute sodium
bicarbonate) was administered by sterile filtration to the production
flasks in equal quantities. Twelve to 16 h later, the second half of the
precursor was administered to the production flasks in the same manner
as the first. The cultures were harvested 6 days (approximately 150 h)
after inoculation and were worked up as described,4 producing the
results shown in Table 1.
Purification of Phenazines from (S)-[2-2H1,3H]Acetate Feeding.
(S)-[2-2H1,3H]Acetate (3.84 mCi) was pulse-fed to 200 mL of S.
antibioticus fermentation medium as described above. The fermentation
broth was harvested by centrifugation 6 days after inoculation. The
supernatant was acidified to pH 2 with 2 N HCl and extracted with
methylene chloride (3 × 100 mL) to give a yellow-brown oil after
concentration. The cellular pellet was broken up in 100 mL of a mixture
of CH2Cl2/MeOH (1:2) by stirring overnight. This mixture was filtered,
and the extraction was repeated twice. The residue obtained after
removing solvent from the combined filtrates was dissolved in 5%
aqueous sodium bicarbonate (200 mL) and washed with 100 mL of
n-hexane. After acidification to pH 1-2 with 2 N HCl, the aqueous
solution was extracted with methylene chloride (3 × 50 mL). Solvent
was removed in vacuo to give a green oily residue that was chromato-
graphed on silica gel (n-hexane/acetone 1:1) to give yellow and green
fractions. All yellow fractions (saphenyl metabolites 2) were concen-
trated, combined with the residue from the supernatant extraction, and
converted to methyl saphenate 5 by hydrolysis and esterification with
diazomethane.4 The crude 5 was chromatographed on silica gel (n-
hexane/acetone 1:1) and the major yellow fraction was collected and
evaporated to dryness (Rf ) 0.35; 85 mg, 78 µCi/mmol). This residue
was chromatographed on a preparativeTLC plate (CH2Cl2/EtOAc:
MeOH 85:10.5) and the major yellow band at Rf ) 0.45 was scraped
off and eluted with CH2Cl2/acetone 1:1 (30 mg, 65 µCi/mmol). This
sample was recrystallized from hexane/acetone yielding 16.5 mg of 5
(42 µCi/mmol), which was dissolved in CH2Cl2 and purified further
by preparative TLC (hexane/acetone 1:1) to give, after elution with
CH2Cl2/acetone, 14 mg of 5 (32 µCi/mmol). A final preparative TLC
(CH2Cl2/EtOAc/MeOH 85:10:5) was carried out to demonstrate radio-
chemical purity. The yellow band at Rf ) 0.4 yielded 12 mg of 5 (35
µCi/mmol); its 1Η and 13C NMR data were in agreement with published
values.2,4 The green fractions from the initial column chromatography,
containing the esmeraldins 1, were hydrolyzed and esterified, using
the same procedure as with the saphenyl metabolites. The resulting
blue oil was chromatographed on silica gel (CH2Cl2/EtOAc 9:1) to give
6.2 mg of the diastereomers 3 and 4 (56 µCi/mmol). The two
diastereomers were separated by HPLC on a semipreparative Partisil
[11-13C]Phenazine-1,6-dicarboxylic Acid ([11-13C]-6). [11-13C]-6
was prepared from [7-13C]benzoic acid according to the procedure of
Holliman and co-workers23 for unlabeled 6: mp >360 °C [lit. mp >290
1
°C]; H NMR (CF3COOD, 500 MHz) δ 9.54 (broad d, J ) 6.84 Hz,
2H), 9.23 (d, J ) 7.81 Hz, 2H), 8.76 (dd, J ) 6.84, 7.81 Hz, 2H); 13
C
NMR (CF3COOD, 125 MHz, inverse gated) δ 170.53 (s, enriched),
144.39 (s, 2C), 140.06 (s, 4C), 138.47 (s, 2C), 134.01 (s, 2C), 124.30
(d, J ) 69.82 Hz, 1C), 124.32 (s, 1C); ESI-MS (dimethyl ester) m/z
268 (0.3), 269 (M+ + H for unlabeled 6, 2.7), 270 (100), 271 (24.9),
272 (3.3), 273 (1.1).
2-[3-(1-Hydroxyethyl)anilino]-3-nitrobenzoic Acid (11). 2-Chloro-
3-nitrobenzoic acid (3 g), 3.2 g of 3-(1-hydroxyethyl)aniline, 2.3 g of
potassium carbonate, and 185 mg of copper powder in 30 mL of
anhydrous ethanol were refluxed for 16 h under an argon atmosphere.
After the solution was cooled to room temperature, ethanol was removed
under reduced pressure and the residue taken up in 50 mL of water.
Insoluble material was removed by filtration and washed thoroughly
with water. The filtrate was acidified with 2 N HCl to afford a sticky
brown-orange precipitate, from which the aqueous layer was decanted.
The residue was taken up in chloroform, dried over MgSO4, and
concentrated to give 3.0 g of orange-red crystals of 11: mp 145-147
°C; 1H NMR (CD3OD, 300 MHz) δ 8.28 (broad d, J ) 8.31 Hz, 1 H),
8.00 (d, J ) 7.81, 1H), 7.20 (t, J ) 7.81, 1H), 7.02 (broad d, J ) 7.81,
1H), 6.99 (t, J ) 7.82, 1H), 6.90 (broad s, 1H), 6.80 (broad d, J )

5624 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
McDonald et al.
7.81, 1H); 13C NMR (CD3OD, 75 MHz) δ 165.1, 148.9, 142.7, 141.8,
138.9, 136.3, 131.9, 130.2, 128.4, 121.9, 119.2, 118.6, 116.3, 70.5,
24.75; GC/MS (methyl ester) m/z 316 (M+ 100), 317 (18), 223 (49),
195 (61).
(0.26 mmol) of 7 in 2 mL of [2H4]methanol and 4 mL of deuterium
oxide. After 5 h of stirring at room temperature, a half molar equivalent
2
(5.3 mg, 0.13 mmol) of sodium borodeuteride (98% H) was added
and the reaction mixture was left overnight. Saturated ammonium
chloride solution was used to destroy the excess reducing agent, and
methanol was evaporated from the mixture. The remaining aqueous
solution was acidified with 1 N HCl and extracted with chloroform (3
× 25 mL). The combined organic phase was dried (MgSO4) and
Saphenic Acid (2a). Sodium ethoxide (8.5 g) in 20 mL of anhydrous
ethanol was added to a solution of 1.5 g of 11 in 15 mL of ethanol.
After the addition, 390 mg of sodium borohydride was stirred in and
the reaction mixture was heated to reflux for 7 h under argon. Upon
cooling the solution to room temperature, the resulting precipitate was
dissolved by adding water, and the ethanol was removed in vacuo.
Careful acidification of the aqueous residue with 5 N HCl and
subsequent extraction with CHCl3 (3 × 100 mL) produced, after drying
and solvent removal, 1.36 g of crude product. Flash chromatography
(CH2Cl2/EtOAc 9:1) yielded 903 mg of 2a: mp 203-204 °C [lit. 202-
204 °C]; 1H and 13C NMR data were in agreement with literature
values.3
1
concentrated to give 50 mg of [12,13-2H4]-2a: mp 198-200 °C; H
NMR (CDCl3, 300 MHz) δ 8.97 (dd, J ) 7.0, 1.4, 1H), 8.50 (dd, J )
8.8, 1.4, 1H), 8.17 (dd, J ) 7.9, 2.1, 1H), 8.03 (dd, J ) 8.8, 7.0, 1H),
8.00-7.91 (m, 2H); 2H NMR (CHCl3, 46.1 MHz) δ 5.76 (s, 1D), 1.78
(s, 3D); 13C NMR (CDCl3, 75 MHz) δ 165.8, 143.7, 142.2, 141.5, 140.2,
139.7, 137.4, 134.9, 133.2, 130.5, 127.8, 127.0, 124.7, 68.5-66.9 (m);
ESI-MS m/z 269 (M+ + H for unlabeled 2a, 1.7), 270 (40.0), 271 (15.0),
272 (6.2), 273 (100), 274 (21.0).
6-Acetylphenazine-1-carboxylic Acid (7). Saphenic acid (100 mg)
and activated MnO2 (700 mg) in 50 mL of dioxane were heated at
reflux for 5 h. Upon cooling the solution, the insoluble material was
filtered off and solvent was removed under reduced pressure. The
residue was purified by chromatography (CHCl3) to give 70 mg of 7:
[12,13-2H4]Saphenic Acid Myristoyl Ester (2d). 2d was prepared
according to the procedure of Keller-Schierlein and co-workers for
saphenic acid acylation:24 1H NMR (CDCl3, 300 MHz) δ 8.97 (dd, J
) 7.1, 1.2, 1H), 8.55 (dd, J ) 8.7, 1.2, 1H), 8.19 (m, 1H), 8.05-7.93
(m, 3H), 1.75-1.55 (m, 2H), 1.40-1.10 (m, 20H), 0.78 (t, J ) 6.5
1
mp 218-219 °C [lit. 219-221 °C]; H and 13C NMR data were in
2
agreement with literature values.3
Hz, 3H); H NMR (CHCl3, 46.1 MHz) δ 7.05 (s, 1D), 1.56 (s, 3D);
13C NMR (CDCl3, 75 MHz) δ 172.9, 165.9, 142.5, 142.1, 141.3, 139.9,
139.6, 137.5, 135.5, 132.9, 130.2, 127.2, 126.9, 124.8, 34.6, 33.6, 31.9,
31.6, 30.9, 29.6-29.1 (several overlapping peaks), 25.0, 24.7, 22.7,
22.6, 14.1.
[13-2H3]Acetylphenazine-1-carboxylic Acid ([13-2H3]-7). Sodium
deuterioxide (0.3 mL, 30% in deuterium oxide) was added to a solution
of 30 mg of 7 in 2 mL of [2H4]methanol and 4 mL of deuterium oxide.
The reaction mixture was stirred at room temperature overnight before
acidifying with 1 N DCl in D2O to pH ∼2. Methanol was removed by
rotary evaporation and the precipitate recovered and dried, giving 17
Acknowledgment. We thank Prof. H. Za¨hner, Tu¨bingen,
Germany for providing the initial culture of S. antibioticus Tu¨
2706 and we acknowledge with gratitude the services of the
National Tritium Labelling Facility, Berkeley, supported by NIH
Grant RR01237, in the preparation of the chiral acetate samples,
the services of the University of Washington Mass Spectrometry
Center, and financial support by the NIH through research
Grants GM 32333 and AI 20264.
1
mg of [13-2H3]-7: mp 216-219 °C; H NMR (CDCl3, 300 MHz) δ
9.01 (dd, J ) 7.4, 1.9, 1H), 8.55 (dd, J ) 8.7, 1.9, 1H), 8.41 (dd, J )
8.7, 1.2, 1H), 8.29 (dd, J ) 6.8, 1.2, 1H), 8.07 (dd, J ) 8.7, 7.4, 1H),
2
8.05 (dd, J ) 8.7, 6.8, 1H), 3.10-3.00 (m, 0.3H); H NMR (CHCl3,
46.1 MHz) δ 3.05 (s); 13C NMR (CDCl3, 75 MHz) δ 200.6, 165.5,
143.6, 142.2, 140.5 140.2, 140.0, 138.6, 135.8, 133.0, 132.9, 132.1,
131.5, 125.9; ESI-MS (methyl ester) m/z 280 (2.2), 281 (M+ + H for
unlabeled 7, 11.2), 282 (11.5), 283 (44.7), 284 (100), 285 (19.0).
[12,13-2H4]Saphenic Acid ([12,13-2H4]-2a). Sodium deuterioxide
(0.3 mL, 30% in deuterium oxide) was added to a solution of 70 mg
JA991159I