An expeditious synthesis of DPD and boron binding studies

Article abstract of DOI:10.1021/ol047695j

(Chemical Equation Presented) A practical synthesis has been developed for DPD (4,5-dihydroxypentane-2,3-dione), an unstable small molecule that is proposed to be the source of universal signaling agents for quorum sensing in bacteria. The synthesis allows preparation of isotopically labeled DPD and ent-DPD as well as detailed studies of spontaneous binding to borate to give the unusual borate complex 6, the signal for marine bacteria such as Vibrio harveyi.

Full text of DOI:10.1021/ol047695j

ORGANIC
LETTERS
2005
Vol. 7, No. 4
569-572
An Expeditious Synthesis of DPD and
Boron Binding Studies
Martin F. Semmelhack,*,† Shawn R. Campagna,† Michael J. Federle,‡ and
Bonnie L. Bassler‡
Departments of Chemistry and Molecular Biology, Princeton UniVersity,
Princeton, New Jersey 08544
mfshack@princeton.edu
Received November 9, 2004
ABSTRACT
A practical synthesis has been developed for DPD (4,5-dihydroxypentane-2,3-dione), an unstable small molecule that is proposed to be the
source of universal signaling agents for quorum sensing in bacteria. The synthesis allows preparation of isotopically labeled DPD and ent-
DPD as well as detailed studies of spontaneous binding to borate to give the unusual borate complex 6, the signal for marine bacteria such
as Vibrio harveyi.
DPD (1) is an enigmatic molecule. It has been known since
1971 as the product of catabolism of S-adenosylhomocysteine
Scheme 1. Biological Forms of DPD
in many bacteria,1 and more recently, it is proposed to be
the core molecule from which all bacterial AI-2 signaling
molecules are derived.2 These molecules are widely used in
inter-species communication in the bacterial world.3 It is a
simple molecule but was reported to be quite unstable toward
rearrangement4 and oligomerization5,6 and has only recently
been synthesized and tentatively characterized in dilute
solution.6 DPD can exist as an equilibrium mixture of three
isomers (1, 2, and 3), hydrated versions (4 and 5), and borate
† Department of Chemistry.
‡ Department of Molecular Biology.
(1) Duerre, J. A.; Baker, D. J.; Salisbury, L. Fed. Proc. 1971, 30, 1067.
(2) Miller, S. T.; Xavier, K.; Campagna, S. R.; Taga, M. E.; Semmelhack,
M. F.; Bassler, B. L.; and Hughson, F. M. Mol. Cell 2004, 15, 677-687.
This paper also presents a nomenclature system for the structures in Scheme
1.
(3) Federle, M. J.; Bassler, B. L. J. Clin. InVest. 2003, 112, 1291-1299.
(4) (a) Winzer, K.; Hardie, K. R.; Burgess, N.; Doherty, N.; Kirke, D.;
Holden, M. T. G.; Linforth, R.; Cornell, K. A.; Taylor, A. J.; Hill, P. J.;
Williams, P. Microbiology 2002, 148, 909-922. (b) Slaughter, J. C. Biol.
ReV. 1999, 74, 259-276.
(5) Semmelhack, M. F.; Campagna, S. R.; Hwa, C.; Federle, M. J.;
Bassler, B. L. Org. Lett. 2004, 6, 2635-2637.
(6) Meijler, M. M.; Hom, L. G., Kaufmann, G. F.; McKenzie, K. M.;
Sun, C.; Moss, J. A.; Matsushita, M.; Janda, K. D. Angew. Chem., Int. Ed.
2004, 43, 2106-2108.
complexes (e.g., 6), collectively known as autoinducer 2 (AI-
2).2 V. harVeyi recognizes the 2,3-borate diester (6) of the
hydrated R-anomer (4) of DPD as AI-2.7 Salmonella Typh-
imurium recognizes the hydrated â-anomer (5) of DPD
without borate.2 The incorporation of borate into the AI-2
signal may be a function of the natural habitat of the bacteria.
V. harVeyi is a marine bacterium that lives in an environment
(7) Chen, X.; Schauder, S.; Potier, N.; Van Dorsselaer, A.; Pelczer, I.;
Bassler, B. L.; Hughson, F. M. Nature 2002, 415, 545-549.
10.1021/ol047695j CCC: $30.25
© 2005 American Chemical Society
Published on Web 01/19/2005

with high borate concentration compared to S. typhimurium,
an enteric bacterium. Here we report an efficient synthesis
of DPD that allows characterization of DPD isomers,
hydration processes, and boron complexation; and that
facilitates the preparation of isotopically labeled versions.
Our synthesis strategy (Scheme 2) parallels that recently
yield as a white solid.11 Oxidative cleavage with KIO4 led
to aldehyde 8 (76%). Following a literature procedure,12 the
Corey-Fuchs protocol gave rise to the dibromoalkene 9,
which was purified (67% yield) prior to treatment with
nBuLi; the intermediate alkyne anion was quenched with
methyl iodide to give alkyne 10 in 64% yield after careful
chromatography. Alternative quenching with water gave the
alkyne 12 (79%); then methylation proceeded in 98% yield
to 10. The critical oxidation process followed the protocol
of Seebach13 with RuO2/NaIO4 and produced the R-diketone
11 in 70% yield as a bright yellow crystalline solid. The
yield is ca. 20% overall from gulonic acid γ-lactone via 12.
Compound 11 showed a strong tendency to hydrate at the
C-3 carbonyl group to give 13 (Figure 1), which complicates
Scheme 2. Synthesis Strategy for DPD
reported6 but employs different tactics which significantly
improve the efficiency and make DPD readily available in
multigram quantities. The previous synthesis of DPD made
use of a labile methyl ortho ester, which could be removed
in dilute aqueous solution to produce DPD in situ for
biological evaluation. The delicacy of that protecting group
required that it be introduced at a late stage as a replacement
for the acetonide group. However, this protection strategy
led to complications with the oxidation to form the R-dike-
tone (analogue of 11; 10% yield). In addition, the acetonide
series involves two intermediates of inconveniently high
volatility. With access to DPD prepared enzymatically,8 we
showed that it is stable at pH 1.5-2 in dilute aqueous
solution for extended periods (no significant decrease in
bioactivity9 after 16 h at 20 °C). While enantio-pure
glyceraldehyde equivalents provide a convenient conceptual
starting point for the synthesis of DPD and other small
molecules, the handling of these molecules can be difficult.10
The cyclohexylidene group (Scheme 2) offers lower
volatility and efficient isolation, and can be removed rapidly
at pH 1.3-1.5 and 20 °C. Importantly, the cyclohexanone
byproduct from deprotection does not inhibit cell growth at
concentrations under 1 M. The monocyclohexylidene deriva-
tive (7) of L-gulonic acid γ-lactone was prepared in 75%
Figure 1. Derivatives of DPD.
the characterization.9 When 11 is placed in H2O or D2O at
pH 1.5 and 20 °C at concentrations up to 30 mM, depro-
tection was >95% complete after 2.0 h with no detectable
byproducts (1H NMR). After adjusting the buffer (0.5 M
potassium phosphate, pH 7.3), synthetic DPD was observed
to be equal in activity to enzymatically prepared DPD at
equal concentration in the V. harVeyi bioassay.9 Using this
route, ent-DPD (14) and DPD labeled with 13C at C-1 (15)
were synthesized.9 ent-DPD (14) had only 1% of the activity
of DPD (1) in the V. harVeyi bioassay.9
Synthetically produced DPD was stable under acidic
conditions; a 30 mM sample of the molecule was monitored
for decomposition via 1H NMR for 5 h at 20 °C and pH 1.5.
Under these conditions, no decomposition products were
observed. A further examination of a 100 mM sample stored
at 20 °C and pH 1.3 for 16 h showed no loss in activity as
monitored by the V. harVeyi bioassay.9 The purity of the
synthetic DPD was further established by reaction with
(8) Schauder, S.; Shokat, K.; Surette, M. G.; Bassler, B. L. Mol.
Microbiol. 2001, 41, 463-476.
(9) For full data, see Supporting Information.
(10) Isopropylidene glyceraldehyde can be generated as either enantiomer
in a few steps, but it is quite volatile and dimerizes readily. These drawbacks
have led to the development of other protecting groups for glyceraldehyde.
(a) For a general discussion, see: Schmid, C. R.; Bradley, D. A. Synthesis
1992, 587-590. (b) For discussion and elaborate solutions, see: Michel,
P.; Ley, S. V. Synthesis 2003, 10, 1598-1602. Aube, J.; Mossman, C. J.;
Dickey, S. Tetrahedron 1992, 48, 45, 9819-9826. (c) First reported by:
Grauert, M.; Schollkopf, U. Liebigs, Ann. Chem. 1985, 1817-1824
(11) First reported by: Vekemans, J. A. J. M.; Boerekamp, J.; Godefroi,
E. F.; Chittenden, G. J. F. Recl. TraV. Chim. Pays-Bas 1985, 104, 266-
272.
(12) Yoshida, J.; Nakagawa, M.; Seki, H.; Hino, T. J. Chem. Soc., Perkin
Trans. 1 1992, 343-350.
(13) Zibuck, R.; Seebach, D. HelV. Chim. Acta 1988, 71, 237-240.
570
Org. Lett., Vol. 7, No. 4, 2005

o-phenylenediamine to produce (S)-1-(3-methylquioxalin-2-
yl)-ethane-1,2-diol in >98% purity.9
Hydration of the carbonyl group at C-3 of the cyclic form
(2) of DPD is important for borate binding in the V. harVeyi
signal, 6. Initially, it was not known whether hydration and
subsequent borate addition were intrinsically favorable or
required the presence of the protein receptor.14 We showed
that laurencione, which is 4-deoxy-DPD, hydrates spontane-
ously and associates strongly with borate in aqueous media
at pH 7.8.5 It forms complexes with both 1:1 and 2:1
laurencione:borate stoichiometry. For synthetic DPD, the
appearance of five 13C NMR signals in the spectrum at pH
1.5 (30 mM DPD, aq) in the region from 90 to 110 ppm
assigned to the hemiacetal and hydrated carbons strongly
suggests that the majority of DPD, both open and closed
isomers, is hydrated at C3. It is less stable at pH >8 unless
bound to borate. Unlike laurencione, which is lacking the
C-4 hydroxyl group and therefore in the hydrated form
contains only one site for borate complexation, hydrated DPD
(4) has two sites for borate complexation. Since both 2,3-
and 3,4-DPD borates are possible for each of the two
anomers, a complex mixture of borate species is expected.
Borate binding was followed with 13C-labeled DPD, 15,
by both 11B and 13C NMR spectroscopy. As shown in Figure
2a, in H2O (5% D2O) and in the absence of borate, 15 mM
15 produces three main peaks from the incorporated label,
assigned to the hydrated open form 16 (δ 24.9 ppm), and
the two hydrated cyclic forms, 17/18 (δ 19.9 or 20.4 ppm).
As borate is added, the complexity of the 13C NMR spectrum
increases, consistent with the hypothesis that a number of
1:1 and 2:1 DPD-borate complexes are present in solution;
with 15 mM 15 and 45 mM B(OH)3 (Figure 2b) the signals
for 16-18 are gone, confirming a high affinity of these
isomers for boron. When the solution is saturated with boric
acid (∼0.9 M) and the concentration of 15 is kept at 15 mM,
the 13C NMR spectrum is dominated by a peak at δ 22.3
ppm to which we assign the 2:1 borate complex, 19 (Figure
2c).10 Although 19 appears crowded and possesses two
negatively charged groups, precedence for this structural type
exists in work showing that small polyhydroxylated mol-
ecules can complex multiple tetrahedral phenyboranate ions
in close proximity.15
Figure 2. Titrating borate into 15 mM DPD at pH 7.8. 13C and
11B NMR monitoring.10 13C NMR titration: (a) no added B(OH)3;
(b) 45 mM B(OH)3; (c) saturated B(OH)3. 11B NMR titration: (d)
no added B(OH)3; (e) 40 mM B(OH)3; (f) saturated B(OH)3.
The peak at δ 4.7 ppm is at a position typical of 1:1
sugar-borate complexes16 and is tentatively assigned to the
3,4-borate complex, 21. The family of peaks from δ 8-11
ppm, have positions consistent with a mixture of 2:1 DPD-
borate complexes (e.g., 20) with borate bound at either the
2,3 or 3,4 position.5 As the concentration of borate is
increased to saturation while keeping the concentration of 1
constant at 15 mM, the intensity of the peaks from δ 8-11
ppm decrease while the peaks at δ 4.7 and 5.8 ppm increase
and become almost equal in intensity (Figure 2f).
These data are consistent with the conversion of 2:1 DPD-
borate complexes to 1:1 DPD-borate complexes (e.g., 6),
followed by further conversion to 1:2 DPD-borate complex
19 as the borate concentration is increased. The inherently
large number of borated DPD species formed coupled with
the difficulties associated with accurately measuring the
equilibrium constant for sugar borate binding17 has not
allowed quantitation of the equilibrium constant for the
association of DPD with borate.
When the identical titration is followed by 11B NMR, no
signal was observed initially with 15 mM 1 and no B(OH)3
(Figure 2d), but as the B(OH)3 concentration was increased
to 40 mM while holding the concentration of 1 constant
(Figure 2e), new peaks appeared: two broad peaks at δ 4.7
and 5.8 ppm and a family of peaks at δ 8-11 ppm (excess
borate at δ 18.3 ppm not shown). The peak at δ 5.8 ppm is
assigned to the natural product, 6 (and its anomer), based
on analogy with 6 bound to the V. harVeyi receptor LuxP
(6.1 ppm),7 6 released from the LuxP receptor (δ 5.8 ppm),2
6 (δ 5.8 ppm) observed when 17 is released into borate from
its receptor, LsrB, from S. typhimurium,2 and the 1:1 borate
complex of laurencione (δ 5.9 ppm).5
To explore the binding of DPD with borate under
conditions that more closely resemble those in the natural
habitat of V. harVeyi and to correlate this study with previous
work using DPD produced in vivo and released from the V.
harVeyi receptor protein, LuxP, the 11B NMR spectrum of a
18
100 µM DPD solution in 400 µM B(OH)3 at pH 7.5 was
(14) Pei and co-workers proposed that the C-3 carbonyl of DPD
(enzymatically prepared) was hydrated in aqueous media: Zhu, J. G.; Hu,
X. B.; Dizin, E.; Pei, D. H. J. Am. Chem. Soc. 2003, 125, 13379-13381.
(15) Lorand, J. P.; Edwards, J. O. J. Org. Chem. 1959, 24, 769-774.
(16) Van Duin, M.; Peters, J. A.; Kieboom, A. P. G.; Van Bekkum, H.
Tetrahedron Lett. 1985, 41, 3411-3412.
(17) Springsteen, G.; Wang, B. Tetrahedron 2002, 58, 5291-5300.
Org. Lett., Vol. 7, No. 4, 2005
571

recorded. Under these conditions, one peak at δ 5.8 ppm
corresponding to 1:1 DPD-borate complexes with borate
bound at the 2,3-positions (6 and its anomer) was observed
significantly above noise level. Based on the ratio of integrals
for the B(OH)3 peak and the DPD-borate peak, the
concentration of DPD-borate complex was calculated to be
ca. 10 µM, consistent with ca. 10% DPD bound at this
dilution. We were unable to observe conclusively the
presence of complexes with a 2:1 DPD-borate stoichiometry
due to sensitivity limitations of the 11B NMR experiment,
although a peak with a S/N of 2 was recorded at δ 9.8 ppm.
We have provided an effective path to DPD that is
amenable to preparation of isotopically labeled versions and
allows a detailed study of the isomerization, hydration, and
boron complexation phenomena.
Acknowledgment. This work was supported by a grant
(PHS NIH AI054442), a graduate fellowship to SC (ND-
SEG), and a postdoctoral fellowship to MF (NIH). We thank
Dr. G. Tranmer for assistance in the early stages of this work.
Supporting Information Available: Procedures and
spectra related to Scheme 2, bioassay data, and borate
titration data. This material is available free of charge via
the Internet at http://pubs.acs.org.
(18) Dickson, A. G. Deep-Sea Res. 1990, 37, 755-766.
OL047695J
572
Org. Lett., Vol. 7, No. 4, 2005