Identifying the minimal enzymes required for anhydrotetracycline biosynthesis

Article abstract of DOI:10.1021/ja800951e

The cyclohexenone ring A of tetracyclines exhibits unique structural features not observed among other aromatic polyketides. These substitutions include the C2 primary amide, C4 dimethylamine, and the C12a tertiary alcohol. Here we report the identification and reconstitution of the minimum set of enzymes required for the biosynthesis of anhydrotetracycline (ATC, 5), the first intermediate in the tetracycline biosynthetic pathway that contains the fully functionalized ring A. Using a combination of in vivo and in vitro approaches, we confirmed OxyL, OxyQ, and OxyT to be the only enzymes required to convert 6-methylpretetramid 1 into 5. OxyL is a NADPH-dependent dioxygenase that introduces two oxygen atoms into 1 to yield the unstable intermediate 4-keto-ATC 2. The aminotransferase OxyQ catalyzes the reductive amination of C4-keto of 2, yielding 4-amino-ATC 3. Furthermore, the N,N-dimethyltransferase OxyT catalyzes the formation of 5 from 3 in a (S)-adenosylmethionine (SAM)-dependent manner. Finally, a non-natural anhydrotetracycline derivative was generated, demonstrating that our heterologous host/vector pair can be a useful platform toward the engineered biosynthesis of tetracycline analogues. Copyright

Full text of DOI:10.1021/ja800951e

Published on Web 04/19/2008
Identifying the Minimal Enzymes Required for
Anhydrotetracycline Biosynthesis
Wenjun Zhang,† Kenji Watanabe,‡ Xiaolu Cai,§ Michael E. Jung,§ Yi Tang,*,† and
Jixun Zhan*,†
Departments of Chemical and Biomolecular Engineering, and Chemistry and Biochemistry,
UniVersity of California, Los Angeles, California 90095, and Department of Pharmaceutical
Sciences, School of Pharmacy, UniVersity of Southern California, Los Angeles, California 90089
Received February 6, 2008; E-mail: yitang@ucla.edu; jixun@ucla.edu
The tetracycline family of broad-spectrum antibiotics are
synthesized by soil-borne streptomycetes using type II polyketide
synthases (PKSs).1 The heavily decorated cyclohexenone ring
A of tetracycline possesses unique structural features not
observed among other aromatic polyketides. These functional
groups, which are essential for the antibiotic properties of
tetracyclines, include the C2 primary amide, the C4 dimethyl-
amine, and the C12a tertiary alcohol. Although the biosynthesis
of tetracyclines has been well-studied using blocked mutants
of tetracycline producers,2 the enzymology of these distinctive
structural features has remained unresolved. Here we report the
identification and reconstitution of the minimum set of enzymes
required for the biosynthesis of anhydrotetracycline (ATC, 5),
the first intermediate in the tetracycline biosynthetic pathway
that contains the fully functionalized ring A.
Enzymatic conversion of the aromatic intermediate 6-methyl-
pretetramid 1 into 5 requires a cascade of tailoring reactions
catalyzed by unidentified enzymes. The key reaction is the double
hydroxylation of ring A, in which two oxygen atoms are introduced
into 1 to yield the proposed intermediate 4-keto-ATC 2. In addition
to the well-studied ATC-oxygenase oxyS (otcC),3 the oxytetracy-
cline (10) gene cluster encodes four uncharacterized oxygenase
genes (oxyL, oxyE, oxyG, and oxyR). We hypothesized that one or
a combination of these oxy oxygenases may be responsible for
forming the cyclohexendione ring A via insertion of oxygen at
the C12a and the C4 positions. The resulting C4 ketone moiety is
then converted to an amine group through reductive amination.
The oxy gene cluster contains two nitrogen-inserting enzymes
(OxyD and OxyQ), of which OxyD is an amidotransferase that
installs the C2 amide group.4 We putatively assigned the PLP-
dependent aminotransferase OxyQ to catalyze the conversion of 2
to 4-amino ATC 3. Last, a N,N-dimethylation reaction, presumably
catalyzed by the (S)-adenosylmethionine (SAM)-dependent me-
thyltransferase homologue OxyT, yields 5.
Figure 1. HPLC analysis (260 nm) of organic extracts from CH999 strains
transformed with different combinations of oxy genes.
product with a yield of ∼10 mg/L. The HPLC retention time, UV
spectrum, mass (m/z 427 [M + H]+), and NMR data all matched
precisely to those of the commercial 5 standard, confirming the
successful heterologous biosynthesis of 5.
To probe the necessity of individual enzymes toward forma-
tion of 5, we systematically removed each of the newly added
genes from pWJATC1 and compared the product profiles of
the resulting transformed strains. Deletion of either OxyQ or
OxyT completely abolished 5 biosynthesis from CH999,
demonstrating that both enzymes are indispensible. Interestingly,
removal of OxyL from the operon led to complete loss of 5
and recovery of 7, while removal of OxyG, OxyE, or OxyR
had no effect on either the product profile or the yield of 5.
These results suggest that OxyL may be the only oxygenase
required during the biosynthesis of 5. Indeed, coexpression of
OxyL, OxyQ, and OxyT in CH999/pWJATC4 afforded 5 as
the predominant polyketide product (Figure 1, CH999/pW-
JATC4). These three enzymes therefore constitute the minimum
set of enzymes required to convert 1 into 5.
We then characterized the functional roles of three individual
enzymes. OxyL is a FAD-binding protein that displays moderate
sequence homology to TcmG (27% identity), which catalyzes the
triple hydroxylation of Tcm A1 in a monooxygenase-dioxygenase
mechanism to yield the highly hydroxylated Tcm C.6 Coexpression
of OxyL alone with OxyABCDJKNF in CH999/pWJ135 led to
the disappearance of 7 and the emergence of WJ135 9 (m/z 227
[M - H]-) as the predominant product (Figure S2). NMR
characterization revealed 9 is a tricyclic ketone (Table S3) and is
a truncated compound that retained the D and C rings of 1. We
were not able to detect 2 in the extract using HPLC/MS. We reason
that 2 may be unstable and can undergo spontaneous intramolecular
rearrangement of the B ring to afford 8, which can then undergo
two tandem retro-Claisen cleavages to yield the observed degrada-
tion product 9 (Scheme 1 and Figure S6). Indeed, the facile in
vitro conversion of 2 to 9 was reported by Scott et al.7 To further
confirm the dioxygenation activity of OxyL in converting 1 to 2,
We have recently identified the enzymes required for the
biosynthesis of 1. Using the heterologous host/vector pair Strep-
tomyces coelicolor CH999/pWJ119, expression of the minimal oxy
PKS (OxyABC), OxyD, C-9 ketoreductase (OxyJ), cyclases
(OxyKN), and C-methyltransferase (OxyF) led to the biosynthesis
of 1 in high yield, which was completely oxidized to 7 (Figure
1).5 To examine the roles of the aforementioned set of enzymes
(OxyLEGRQT) in transforming 1 into 5, the six genes were
assembled into a single operon and inserted into pWJ119 to yield
pWJATC1. CH999 transformed with pWJATC1 developed a dark
brown color, and the organic extract contained a single predominant
† Department of Chemical and Biomolecular Engineering, UCLA.
‡ University of Southern California.
§ Department of Chemistry and Biochemistry, University of California,
Los Angeles.
9
6068 J. AM. CHEM. SOC. 2008, 130, 6068–6069
10.1021/ja800951e CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1
soluble, holo-OxyL was overexpressed from Escherichia coli and
purified (50 mg/L). The substrate 1 was prepared as described
previously.8 Due to the instability of both the substrate and product,
the reaction containing OxyL (20 µM), 1 (∼0.2 mM), and NADPH
(2 mM) was incubated at 25 °C for 1 h, after which the supernatant
was immediately injected into HPLC/MS. As expected, 1 was
converted to a polar compound with mass (m/z 396 [M - H]-)
consistent with that of 2 (Figure S8). The combination of in vivo
and in vitro results strongly indicates the role of OxyL as a
NADPH-dependent dioxygenase that hydroxylates 1 at both C12a
and C4 positions, possibly via a monooxygenase-monooxygenase
mechanism (Figure S5).
Coexpression of OxyQ and OxyL in CH999/pWJ209 yielded a
new compound with a nearly identical UV spectrum as 5 (Figures
S2 and S11). The compound was more stable than 2 and was
verified to be 39 using high-resolution mass spectrometry (m/z )
421.1006 [M + Na]+, C20H18N2O7Na, calcd: 421.1012). Biosyn-
thesis of 3 confirmed OxyQ is a reductive transaminase, which is
the first of its kind found among bacterial type II PKSs.
Purification of 3 from the above strain enabled us to verify the
role of the N,N-dimethyltransferase OxyT in vitro. To date, the
only other enzymes found in PKS pathways capable of N,N-
dimethylation are all associated with the biosynthesis of amino
deoxysugars such as desosamine.10 OxyT was overexpressed from
E. coli and purified to homogeneity (50 mg/L). OxyT (200 nM)
was then incubated with purified 3 (0.2 mM) and SAM (2 mM) at
25 °C. HPLC/MS analysis of the reaction mixture after short
reaction times revealed the presence of an intermediate (tR ) 11.0
min), which was verified by LC/MS to be the monomethylated
species 4.11 Prolonged incubation led to the conversion of both 3
and 4 into the expected 5 (Figure 2). Removal of SAM abolished
the synthesis of 4 and 5, establishing SAM as the essential methyl
donor in the reaction. Fitting the conversion data to the rate laws
for an irreversible unimolecular consecutive reaction (3 f 4 f 5)
led to the apparent kinetic constants of 0.009 and 0.08 min-1 for
the mono- and dimethylation steps, respectively.9 These results
confirmed the role of OxyT in the N,N-dimethylation of 3. OxyT
is therefore the first N,N-dimethyltransferase to use an aromatic
polyketide aglycon as a substrate.
Having identified and confirmed the roles of OxyL, OxyQ,
and OxyT in the biosynthesis of 5, we set out to rationally
biosynthesize the analogue 6-desmethyl-ATC 6 (Scheme 1).
The 6-methyl moiety is not necessary for the antibacterial
activity of tetracyclines and is absent in semisynthetic
tetracyclines such as minocycline. We constructed pJX67b
that contained all the enzymes required for the assembly of
5, except OxyF, which we previously identified as a C6
methyltransferase.5 Transformation of CH999 with pJX67b
resulted in the biosynthesis of a predominant new product
with a titer of 10 mg/L (Figure 1). The identity of the
compound was confirmed to be 6 using NMR spectroscopy
(Table S4). This result reveals that the three tailoring enzymes
studied here have relaxed substrate specificity toward sub-
stitutions at C6. The efficient biosynthesis of 6 demonstrates
that the heterologous host/vector pair can be a useful platform
toward the biosynthesis of tetracycline analogues.
Acknowledgment. This work is funded by NSF CBET #0545860.
Supporting Information Available: Experimental procedures,
and compound characterizations. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) (a) Kim, E. S.; Bibb, M. J.; Butler, M. J.; Hopwood, D. A.; Sherman, D. H.
Gene 1994, 141, 141–142. (b) Binnie, C.; Warren, M.; Butler, M. J. J.
Bacteriol. 1989, 171, 887–895. (c) Butler, M. J.; Binnie, C.; Hunter, I. S.;
Sugden, D. A.; Warren, M. DeV. Ind. Microbiol. 1990, 31, 41–50. (d)
Nakano, T.; Miyake, K.; Ikeda, M.; Mizukami, T.; Katsumata, R. Appl.
EnViron. Microbiol. 2000, 66, 1400–1404.
(2) (a) McCormic, J. R. D.; Johnson, S.; Sjolander, N. O. J. Am. Chem. Soc. 1963,
85, 1692–1694. (b) McCormic, J. R. D.; Joachim, U. H.; Jenson, E. R.; Johnson,
S.; Sjolander, N. O. J. Am. Chem. Soc. 1965, 87, 1793–1794. (c) McCormic,
J. R. D.; Jensen, E. R. J. Am. Chem. Soc. 1969, 91, 206–206.
(3) (a) Behal, V.; Hostalek, Z.; Vanek, Z. Biotechnol. Lett. 1979, 1, 177–
182. (b) Peric-Concha, N. A.; Borovicka, B.; Long, P. F.; Hrnaueli, D.;
Waterman, P. G.; Hunter, I. S. J. Biol. Chem. 2005, 280, 37455–37460.
(4) Zhang, W.; Ames, B. D.; Tsai, S. C.; Tang, Y. Appl. EnViron. Microbiol.
2006, 72, 2573–2580.
(5) (a) Zhang, W.; Watanabe, K.; Wang, C. C.; Tang, Y. J. Biol. Chem. 2007,
282, 25717–25725. (b) Zhang, W.; Wilke, B. I.; Zhan, J.; Watanabe, K.;
Boddy, C. N.; Tang, Y. J. Am. Chem. Soc. 2007, 129, 9304–9305.
(6) Rafanan, E. R., Jr.; Hutchinson, C. R.; Shen, B Org. Lett. 2000, 2, 3225–
3227.
(7) Scott, A. I.; Yamaguchi, E.; Chung, S. K. Tetrahedron Lett. 1975, 16, 1369–
1372.
(8) Hassall, C. H.; Thomas, G. J. J. Chem. Soc., Perkin Trans. 1 1978, 2, 145–
152.
(9) Blackwood, R. K.; Stephens, C. R. J. Am. Chem. Soc. 1964, 86, 2736–2737.
(10) (a) Chang, C.; Zhao, L.; Yamase, H.; Liu, H. Angew. Chem., Int. Ed. 2000,
39, 2160–2163. (b) Chen, H.; Yamase, H.; Murakami, K.; Chang, C. W.;
Zhao, L.; Zhao, Z.; Liu, H. W. Biochemistry 2002, 41, 9165–9183.
(11) Miller, P. A.; Saturnelli, A.; Martin, J. H.; Mitscher, L. A.; Bohonos, N.
Biochem. Biophys. Res. Commun. 1964, 16, 285–291.
Figure 2. HPLC analysis (270 nm) of the OxyT catalyzed N,N-
dimethylation of 3 to yield 5. (A) No OxyT; (B) 10 min after OxyT addition.
4 is the monomethylated compound; (C) 10 h; and (D) 5 standard. Traces
not drawn to scale.
JA800951E
9
J. AM. CHEM. SOC. VOL. 130, NO. 19, 2008 6069