A latent oxazoline electrophile for N-O-C bond formation in pseudomonine biosynthesis

Article abstract of DOI:10.1021/ja804499r

Nitrogen-heteroatom bonds figure prominently in the structural, chemical, and functional diversity of natural products. In the case of Pseudomonas siderophore pseudomonine, an N-O hydroxamate linkage is found uncommonly configured in an isoxazolidinone ri

Full text of DOI:10.1021/ja804499r

Published on Web 08/19/2008
A Latent Oxazoline Electrophile for N-O-C Bond Formation in
Pseudomonine Biosynthesis
Elizabeth S. Sattely and Christopher T. Walsh*
Department of Biological Chemistry and Molecular Pharmacology, HarVard Medical School, Boston, Massachusetts
02115
Received June 18, 2008; E-mail: christopher_walsh@hms.harvard.edu
Nitrogen-heteroatom bonds figure prominently in the structural,
chemical, and functional diversity of natural products. For example,
homolytic cleavage of the weak N-O bond of mitomycin relative
FR900482 triggers formation of the active mitosene antitumor
agent,1 and protein tyrosine phosphatases generate a transient N-S
linkage by capture of the sulphenic acid present in the inactive form
of the enzyme.2 A different class of N-X bonds, the N-O-
containing hydroxamates, is featured in small molecule ligands
excreted by bacteria to scavenge ferric iron.3 In the case of
Pseudomonas metabolite pseudomonine,4 however, the hydroxam-
ate moiety is found uncommonly configured in an isoxazolidinone
ring.5 While less prevalent than uncyclized hydroxamates, these
cyclic N-O variants have surfaced in several naturally occurring
metabolites including well-known tuberculosis antibiotic D-cyclos-
erine6 and ꢀ-lactamase inhibitor lactivicin7 (Figure 1).
Figure 2. (a) The 21-kb pseudomonine biosynthetic gene cluster and
predicted function of encoded proteins. (b) Variations in NRPS domain
organization among related heterocyclic siderophores. Pseudomonine is a
Figure 1. Naturally occurring isoxazolidinones include antibiotics D-
cycloserine and lactivicin and Pseudomonas siderophore pseudomonine.
close structural relative of oxazoline acinetobactin. IPL ) isochorismate
pyruvate lyase; ICL ) isochorismate lyase; A ) adenylation, T ) thiolation,
Flavin monooxygenases are often implicated in the formation
of the ubiquitous N-O bond,8,9 yet the pathway of hydroxylamine
incorporation into the isoxazolidinone has not been delineated.
Through in vitro reconstitution of the pseudomonine synthetase,
we have uncovered a previously uncharacterized mode of hetero-
cyclization that involves an oxazoline electrophile for N-O-C
bond formation. Our analysis suggests an SN2 mechanism and
establishes the oxygen of a hydroxamate intermediate as the
nucleophile in this instance of isoxazolidinone biosynthesis.
Pseudomonine is a blue-fluorescent siderophore first isolated from
the surface of the Lake Victoria Nile perch.4 The biosynthetic genes
associated with production have been identified in the genome of
Pseudomonas fluorescens10 and have been putatively assigned by
analogy in Pseudomonas entomophila.11 As illustrated in Figure
2a, five core proteins are responsible for the functionalization of
L-His and subsequent ligation with Thr and salicylic acid (SA) to
generate the pseudomonine scaffold. Comparison of the pseudomo-
nine non-ribosomal peptide synthetase (NRPS) with those of related
siderophores vibriobactin12 and acinetobactin13 reveals similar
domain organization (Figure 2b) yet does not account for the
observed structural differences in the heterocyclic products. In
particular, acinetobactin and pseudomonine are nearly constitutional
isomers with variation consisting only of the second hydroxyl
substituent of the aromatic acid. In the well-studied vibriobactin
system, the cyclization domains, as condensation subtypes,14 have
been shown to act in tandem: Cy2 of VibF promotes formation of
Cy ) cyclization, and C ) condensation domain.
a salicyl amide, which then becomes the site of Cy1-catalyzed
cyclization of the Thr-hydroxyl side chain and dehydration to deliver
the oxazoline.12 The pseudomonine synthetase appears to be a
relative of VibF, albeit severed into two proteins. PmsD harbors
an analogous set of cyclization domains, while alignment with VibF
indicates remnants of a condensation (C1) domain at the C-terminus
of PmsD and N-terminus of PmsG.15 Despite this resemblance, we
anticipated that in vitro analysis of the pseudomonine assembly
line would elucidate a pathway for isoxazolidinone biosynthesis
distinct from that for the formation of oxazoline-containing sid-
erophores.
The five requisite genes for pseudomonine biosynthesis pm-
sADEGF were amplified from genomic DNA isolated from P.
entomophila and separately expressed in E. coli to provide the
derived C- or N-terminal His6-fusion constructs. Annotation of
PmsA and PmsF suggested these proteins are involved in provision
of N-hydroxyhistamine, which could serve as a nucleophile in the
release of a PmsG-tethered acyl product (eq 1). Although PmsA
was isolated as the apoprotein, PLP-bound enzyme was generated
upon titration with the cofactor as indicated by the characteristic
9
12282 J. AM. CHEM. SOC. 2008, 130, 12282–12284
10.1021/ja804499r CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1. Early Condensation Events in Pseudomonine
Biosynthesis and Evidence for Oxazoline Intermediate 1a
Our assessment of amide-bond formation catalyzed by PmsD
was carried out as indicated in Scheme 1; reaction analysis was
simplified by the fact that the thioester of the PmsG-tethered
condensation product undergoes adventitious hydrolysis, thus
allowing for catalytic turnover of the synthetase. Unexpectedly,
comparison of the released carboxylic acid with authentic synthetic
standards18 by HPLC cleanly identified oxazoline 2 as the hydroly-
sis product and implicated thioester 1 in PmsD catalysis (Scheme
1, path a). The distinct UV profiles for the salicylate-derived
oxazoline 2 and the uncyclized salicyl amide 4 helped to confirm
product assignment; this difference was a useful property for the
analysis of subsequent reconstitution experiments. The formation
of oxazoline 2 reveals that, in analogy to VibF, PmsD is capable
of promoting both amide coupling, and subsequent cyclodehydra-
tion. It is possible that this mode of heterocyclization is off-pathway;
however, our findings strongly suggest that the isoxazolidinone of
pseudomonine arises via an unusual mechanism, one which most
likely does not directly involve the action of PmsD.
In order to ascertain the intermediacy of oxazoline 1 in
pseudomonine biosynthesis, we next tested the condensation activity
of PmsG. The C domain of PmsG is proposed to catalyze acylation
of a soluble amine monomer with the protein-bound thioester; this
event constitutes the final step in chain extension and product
release. Condensation assays involved the in situ generation of 1
through the combination of PmsE, G, and D in the presence of
L-Thr and SA; the addition of either histamine or synthetic
N-hydroxyhistamine revealed that oxazoline 1 could be efficiently
captured by an amine nucleophile. The C domain of PmsG readily
accepts histamine as a substrate (Scheme 1, path b); the resulting
product is assigned as the intact oxazoline 3 based on LCMS data
and the characteristic UV absorption spectrum.18
In contrast, the presence of N-hydroxyhistamine in the condensa-
tion assay results in complete reconstitution of pseudomonine
biosynthesis19 (Scheme 1, path c). The observed assembly line
product is identical to pseudomonine isolated from cultures of P.
fluorescens as determined by HRMS, 1H NMR, and UV absorption
data. Pseudomonine is also produced in reconstitution experiments
that include the combination of histamine, the N-hydroxylase PmsF,
and requisite cofactors FAD and NADPH (data not shown). This
outcome further supports the assigned function of PmsF and
establishes N-hydroxyhistamine as the native substrate in pseudomo-
nine biosynthesis.
These data led us to propose that the isoxazolidinone of
pseudomonine is accessed through the rearrangement of an initial
oxazoline hydroxamate condensation product, named “pre-pseudomo-
nine”. As illustrated in Scheme 2, SN2 attack at the ꢀ-carbon of a
Thr-derived oxazoline results in the N-O-C bond connectivity
of the isoxazolidinone moiety. Close monitoring of the pseudomo-
nine reconstitution assay described in Scheme 1, path c reveals the
formation of a transient compound suspected to be pre-pseudomo-
nine (Scheme 2). Although this intermediate has the same mass as
pseudomonine (by LCMS analysis), the associated UV absorption
spectrum suggests it bears a salicyl-oxazoline.20
The proposed intramolecular SN2 reaction involves an inversion
of configuration at the ꢀ-carbon of the Thr side chain. This
stereochemical outcome accounts for the observation that, whereas
L-Thr is activated by the A domain of PmsD, the pseudomonine
isoxazolidinone bears a modified L-allo-Thr residue with substit-
uents positioned with trans relative stereochemistry about the
heterocycle.21
Additional evidence supporting the rearrangement mechanism
was garnered from the 18O-labeling study shown in eq 2. The
synthetic route developed for access to N-hydroxyhistamine could
a (a) Hydrolysis of thioester 1 and identification of oxazoline 2 by HPLC.
(b) Capture of 1 with histamine yields 3 (m/z ) 314.70). (c) In the presence
of synthetic N-hydroxyhistamine, PmsE, G, and D convert SA and L-Thr
to pseudomonine (m/z ) 331.14).
UV absorbance of the Schiff base at 416 nm.16 Decarboxylase
activity of PmsA in the presence of L-His was subsequently
confirmed by HPLC analysis. PmsF was characterized as a two-
component flavoprotein as indicated by NADPH oxidation in the
presence of FAD and histamine; catalytic turnover could also be
observed by the FAD and NADPH-dependent consumption of
histamine by HPLC. These studies corroborate the proposal of
N-hydroxyhistamine as a somewhat unstable yet soluble intermedi-
ate in pseudomonine biosynthesis. Elucidation of the pathway for
incorporation of this amine into the pseudomonine isoxazolidinone
necessitated the functional examination of NRPS components PmsE,
G, and D (Scheme 1).
On the basis of genetic analysis, we expected that N-acylation
of a Thr residue with an aromatic acid derivative is a common
starting point in the biosynthesis of the pseudomonine, acineto-
bactin, and vibriobactin. In vitro confirmation of this assumption
required the posttranslational modification of the carrier T domains
of PmsE and G with phosphopantetheinyl substrate attachment sites,
accomplished by incubation of each with CoA and Sfp, a promiscu-
ous phosphopantetheinyl transferase from B. subtillus.17 An
ATP-32PPi exchange assay established that the A domains of PmsE
and PmsD are each active for substrate adenylation. We propose
that, unlike the A domain of PmsE, that of PmsD functions in trans
to generate substrate-bound PmsG. Although the embedded Thr
moiety of pseudomonine is thought to be of the allo-configuration,4
we discovered that that PmsD activates proteinogenic L-Thr
preferentially over L-allo-Thr or D-Thr.
9
J. AM. CHEM. SOC. VOL. 130, NO. 37, 2008 12283

C O M M U N I C A T I O N S
Scheme 2. Proposed Rearrangement Mechanism for the
Formation of the Pseudomonine Isoxazolidinone Involving an
Oxazoline Electrophilea
amate oxygen as a nucleophile in isoxazolidinone biosynthesis and
may be indicative of a general strategy for the production of this
class of N-O-C bonds in nature.
Acknowledgment. We thank Jesus Mercado Blanco for sharing
the complete sequencing data for the pseudomonine gene cluster
in P. fluorescens WCS374 and Peter Bakker for the gift of this
strain. Frederic Boccard and Bruno Lemaitre are acknowedged for
the gift of P. entomophila L48. Michael Fishbach brought to our
attention the pseudomonine gene cluster in P. entomophila and is
thanked for helpful discussion. We are grateful to Carl Balibar for
experimental assistance. This work was supported by NIH Grant
AI 47238 (C.T.W.); Elizabeth Sattely is a Damon Runyon Fellow
supported by the Damon Runyon Cancer Research Foundation
(DRG-1980-08).
Supporting Information Available: Experimental details. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
a A time course of pseudomonine reconstitution reveals the intermediacy
of pre-pseudomonine. Compound 5, eluting at 13 min, is assigned as the
salicyl-Thr-N-hydroxyhistamine adduct, the relative stereochemistry of
which is unknown. Absorbance monitored at 247 nm.
(1) Beckerbauer, L.; Tepe, J. J.; Cullison, J.; Reeves, R.; Williams, R. M. Chem.
Biol. 2000, 7, 805–812.
(2) Salmeen, A.; Andersen, J. N.; Myers, M. P.; Ment, T-C.; Hinks, J. A.;
Tonks, N. K.; Barford, D. Nature 2003, 423, 769–773.
(3) (a) Crosa, J. H.; Walsh, C. T. Microbiol. Mol. Biol. R. 2002, 66, 223–249.
(b) Miethke, M.; Marahiel, M. A. Microbiol. Mol. Biol. ReV. 2007, 71,
413–451.
be modified for the incorporation of 18O at the hydroxylamine
oxygen. Use of the resulting labeled substrate in reconstitution
assays gave the expected pseudomonine product with a comparable
level of enrichment as measured by MS analysis. Although an
orthogonal set of experiments with SA labeled with 18O at the
carboxylate group generates unlabeled pseudomonine, this result
was complicated by the partial loss of the SA label under the
reaction conditions. Taken together, however, these data indicate
that the isoxazolidinone oxygen originates from the N-hydroxy-
histamine precursor.
(4) Anthoni, U.; Christophersen, C.; Nielsen, P. H.; Gram, L.; Petersen, B. O.
J. Nat. Prod. 1995, 58, 1786–1789.
(5) Rohboek L.; Christophersen, C. The Isoxazole Alkaloids. In The Alkaloids:
Chemistry and Biology; Cordell, G. A., Ed.; Academic Press: San Diego,
CA, 2001; Vol. 57, pp 185-233.
(6) Zhang, Y. Annu. ReV. Pharmacol. Toxicol. 2005, 45, 529–564.
(7) Macheboeuf, P.; Fischer, D. S.; Brown, T., Jr.; Zervosen, A.; Luxen, A.;
Joris, B.; Dessen, A.; Schofield, C. J. Nat. Chem. Biol. 2007, 3, 365–569.
(8) For a review of flavoproteins, see: (a) van Berkel, W. J. H.; Kamerbeek,
N. M.; Fraaije, M. W. J. Biotechnol. 2006, 124, 670–689.
(9) For reviews on the intermediacy of hydroxyl amines in the context of nitro-
group biosynthesis, see: (a) Winkler, R.; Hertweck, C. ChemBioChem 2007,
8, 973–977. (b) Timmons, S. C.; Thorson, J. S. Curr. Opin. Chem. Biol.
2008, 12, 297–305.
(10) (a) Mercado-Blanco, J.; van der Drift, K. M. G. M.; Olsson, P. E.; Thomas-
Oates, J. E.; van Loon, L. C.; Bakker, P. A. H. M. J. Bacteriol. 2001, 183,
1909–1920. (b) Djavaheri, M. Ph.D. Thesis, Utrecht University, Utecht,
The Netherlands, 2007. (c) Mercado-Blanco, J.; Bakker, P. A. H. M.
Personal communication, 2007.
(11) Boccard, F.; et al. Nat. Biotechnol. 2006, 24, 673–679.
(12) Marshall, C. G.; Hillson, N. J.; Walsh, C. T. Biochemistry 2002, 41, 244–
250.
(13) Mihara, K.; Tanabe, T.; Yamakawa, Y.; Funahashi, T.; Nakao, H.;
Narimatsu, S.; Yamamoto, S. Microbiology 2004, 150, 2587–2597.
(14) Rausch, C.; Hoof, I.; Weber, T.; Wohlleben, W.; Huson, D. H. BMC EVol.
Biol 2007, 7, 78.
The proposed initial product of the pseudomonine synthetase,
pre-pseudomonine, is nearly identical to acinetobactin (Figure 2),
yet pre-pseudomonine appears to undergo facile conversion to the
isoxazolidinone. In this regard, it remains to be seen if both pre-
pseudomonine and pseudomonine are physiologic siderophores. Our
results suggest that rearrangement is spontaneous; the rate of
conversion of isolated pre-pseudomonine, prepared enzymatically,
to pseudomonine at pH 7 is not altered by the presence of PmsG.
It is not clear how the additional hydroxyl substituent of acineto-
bactin precludes an analogous transformation, and how pH and
temperature influence the reaction. The pre-pseudomonine rear-
rangement is reminiscent of the postulated mechanism of retaining
hexosaminidases, which involves the hydrolysis of a substrate-
derived oxazolinium intermediate.22 However, to the best of our
knowledge, an oxazoline electrophile in natural products enzymol-
ogy has not been described.23 Our results characterize the hydrox-
(15) The C1 domain of VibF has been shown to be catalytically inactive and is
presumed responsible for dimerization of the synthetase, (a) Hillson, N. J.;
Balibar, C. J.; Walsh, C. T. Biochemistry 2004, 43, 11344–11351.
(16) Eliot, A. C.; Kirsch, J. F. Annu. ReV. Biochem. 2004, 73, 383–415.
(17) Quadri, L. E. N.; Weinreb, P. H.; Lei, M.; Nakano, M. M.; Zuber, P.; Walsh,
C. T. Biochemistry 1998, 37, 1585–1595.
(18) See Supporting Information for details.
(19) For a review on the reconstitution of NRPS and PKS assembly lines, see:
(a) Sattely, E. S.; Fishbach, M. A.; Walsh, C. T. Nat. Prod. Rep. 2008, 25,
757–793.
(20) An alternative pathway to pre-pseudomonine involving PmsF-catalyzed
N-hydroxlation of 3 was ruled out by the observation that PmsF promotes
NADPH oxidation in the presence of histamine but not 3.18
(21) Reduction of the pseudomonine N-O bond and comparison of the resulting
product with authentic synthetic standards by HPLC supports the assigned
relative stereochemistry.18
(22) Mark, B. L.; James, M. N. G. Can. J. Chem. 2002, 80, 1064–1074.
(23) For an example of an oxazoline electrophile utilized in the context of organic
synthesis, see: (a) Wilson, K. J.; Sabat, M.; McGarvery, G. J. J. Org. Chem.
1993, 58, 6180–6181.
JA804499R
9
12284 J. AM. CHEM. SOC. VOL. 130, NO. 37, 2008