The fate of [2,3,3-2H3, 1,2-13C2]-D,L-glycerate in clavulanic acid biosynthesis

Article abstract of DOI:10.1039/a607078g

The hydrogen at C-2 of glycerate is lost in the biosynthesis of clavulanic acid.

Full text of DOI:10.1039/a607078g

View Article Online / Journal Homepage / Table of Contents for this issue
The fate of [2,3,3-2H3, 1,2-13C2]-d,l -glycerate in clavulanic acid biosynthesis
Janos Pitlik and Craig A. Townsend*
Department of Chemistry, The Johns Hopkins University, Baltimore, Maryland 21218, USA
The hydrogen at C-2 of glycerate is lost in the biosynthesis of
clavulanic acid.
desired quintuply labelled sodium glycerate 12 in nearly 60%
overall yield. 13C{1H} NMR analysis of the product [Fig. 1(a)]
displayed the expected carboxyl resonance as a clean doublet
(1JCC = 54.2 Hz) while the corresponding doublet for C-2 was
further split into a pair of 1:1:1 triplets owing to the directly
bound deuterium (1JCD = 22.1 Hz) and broadened slightly by
the adjacent C-3 deuteria.
Cultures (2.0 l) of Streptomyces clavuligerus (ATCC 27064)
were grown as previously described (500 ml/4 l flask).12 After
48 h, 12 (2.0 mmol/l) was administered in equal portions to the
fermentations under sterile conditions. After an additional 96 h,
the clavulanic acid produced was isolated by adsorption onto
carbon, conversion to its p-bromobenzyl ester and purification
by silica gel chromatography (40 mg).2,13 Careful 13C{1H}
NMR analysis of this product showed the natural abundance
singlet for the b-lactam carbonyl, C-7, flanked by a doublet
(1JCC = 38.9 Hz) indicating 0.80–0.85% intact incorporation of
the multiply labeled glycerate carbon skeleton [Fig. 1(b)].
Elusive to experiment and mechanistic understanding has been
the biosynthesis of the 4-membered ring of deoxyguani-
dinoproclavaminic acid 1, the first b-lactam-containing inter-
mediate in the anabolic pathway to clavulanic acid 2.1 These
three carbons are known to be derived efficiently from glycerol
3, accompanied by the loss of HC.2,3 Subsequent radiochemical
experiments with samples of this C3-carbohydrate bearing
tritium stereospecifically at the pro-(R) hydroxymethylene
established that chiral information was delivered through the
entire biosynthetic pathway to clavulanic acid 2 such that the
(1R,2S)-hydrogen (HA) of 3 is lost while HB is incorporated in
2 at C-5.4 Similarly, in the clavaminate synthase-catalysed
cyclization/dehydration of proclavaminic acid 5 to clavaminic
acid 6 it was determined that formation of the oxazolidine ring
involved specific replacement of the 4AS-label (HA) with
substrate oxygen (retention of configuration).5 Finally, the
demonstration that both HA and HB are retained from glycerol 3
to proclavaminic acid 5 allowed important deductions to be
made about the cryptic formation of the b-lactam ring.
Generation of the azetidinone N–C-4A bond takes place with (a)
loss of the glycerol oxygen, (b) no net change in oxidation state
at this methylene carbon and, (c) overall retention of configura-
tion,6 unlike monocyclic b-lactam formation in nocardicin7 and,
presumably, the monobactams. Lastly, the stereochemical
inversion that takes place between clavaminic acid 6 and
clavulanic acid 2 must occur with retention of HB.
The complete loss of HC from glycerol 3, but the retention of
both HA and HB into proclavaminic acid 5 followed by their
stereospecific loss and retention, respectively, into clavulanic
acid 2 point to the intervention of conventional glycolytic
metabolism, e.g. to glycerate 4, prior to uptake into the b-lactam
biosynthetic pathway. Earlier whole-cell studies revealed that
[2-3H, 1-14C]-d-glycerate gave efficient incorporation of carbon
label into the b-lactam ring (95–100%) of clavulanic acid, but
only 4–11% of the tritium could be accounted for in the
molecule by chemical degradation, principally at C-6 and C-8.3
Given the inherent inaccuracies of low radioactivity in this
technically difficult double label experiment, we have re-
examined the fate of the glycerate C-2 hydrogen by a
methodologically distinct means.
C3 Precursor
+
NH
NH2
NH
NH2
H2N
N
N
H
N
H
O
CO2H
CO2H
1
HB
O
HB
5
9
O
6
NH2
OH
8
N
N
O
O
CO2H
CO2H
2
6
HA
N
HB
OH
HA
OH
OH
HD
HC
HO
HO
HO
HB
HB
NH2
O
HA
CO2H
CO2H
5
4
3
Scheme 1
The experiment to test the retention or loss of H-2 was
designed to use two 13C-labels as internal measures of intact and
absolute carbon utilization and then multiple deuterium labels to
monitor changes at H-2 knowing that one of the glycerate H-3
hydrogens would be cleanly retained in clavulanic acid 2. Thus,
(±)-[2,3,3-2H3, 1,2-13C2]glycerate 12 (Scheme 2) was prepared
as follows. [1,2-13C2]Bromoacetate 7 was converted to its
crystalline p-nitrobenzyl (PNB) ester and treated with PPh3 to
form the phosphonium salt 8. Facile ylide formation in D2O
allowed rapid exchange at the C-2 methylene followed by
Wittig reaction with deuterioformaldehyde in D2O8,9 in the
presence of anhydrous K2CO3 yielded p-nitrobenzyl [2,3,3-2H3,
1,2-13C2]acrylate 10 in excellent yield. Sharpless dihydroxyla-
tion10,11 gave PNB glycerate 11. Hydrogenolysis in the
presence of NaHCO3 and reverse-phase HPLC purification
(Partisil 10 ODS 3, 25 3 250 mm, H2O elution) provided the
Scheme 2 Reagents and conditions: i, PNB–OH, EDC, DMAP, CH2Cl2,
room temp., 3 h, 95%; ii, PPh3, THF–CH2Cl2, room temp., 24 h; iii, D2O,
room temp., 24 h; iv, D2CO in D2O, K2CO3, 24 h, 80% for 3 steps; v,
ButOOH, cat. OsO4, Et4NOAc, acetone, 0 °C to room temp., 6 h, 85%; vi,
H2, 10% Pd–C, NaHCO3, H2O–THF, room temp., 24 h, then RP–HPLC,
92%
Chem. Commun., 1997
225

View Article Online
Centred upfield at d 46.39, the natural abundance peak for C-6
could be found flanked by upfield a- and a/b-shifted doub-
lets.14 That marked by (:) corresponds to 13C at C-6 paired to
13C at C-7 and 2H-label at C-5 (a- and b-heavy isotopes;
Dd = 0.144 ppm). This species represents 80% fully intact
incorporation of 12. The second, smaller doublet (5) is due to
doubly 13C-labelled molecules, but having 1H at C-5 (a-heavy
isotope; Dd = 0.019 ppm). While this 20% of incorporated
precursor seems high, it is an artifact4 of cummulative isotope
effects15 that discriminate against 2H-labelled molecules, but let
the very small proportion of [1H,13C]-labelled precursor and
intermediate molecules pass more swiftly along the pathway.
These observations notwithstanding, notably absent is a 1:1:1
triplet corresponding to deuterium directly bound to C-6 and a
concommitant b-isotope shift detectable in the doublet for C-7.
In an attempt to establish an upper bound for incorporation of
deuterium at C-6, a 2H{1H} NMR spectrum was recorded (data
not shown). Deuterium label at C-5 was easily observed and,
while some minor secondary incorporation at C-9 could be
detected, none was seen at C-6. Within the limits of the NMR
methods used, we place the possible level of deuterium
incorporation at C-6 from each of these NMR experiments at
< 0.05%.
We are led to conclude from these data that 2H-2 is lost from
12 upon incorporation into clavulanic acid. There is a caveat,
however remote, that H-2, unlike H-3, is labile among the
intermediates of glycolysis despite the use of a triglyceride-
based medium to favour gluconeogenesis.3,12 The possibility
exists that the metabolic flux among these intermediates is so
great relative to the rate of entry into the biosynthetic pathway
that the C-2 deuterium has exchanged for hydrogen before
detectable incorporation at C-6 can occur. This proviso aside,
the loss of H-2 from glycerate imposes new limitations on the
mechanism of b-lactam formation in clavulanic acid biosyn-
thesis, a process whose full understanding has thus far resisted
experimental enquiry.
References
1 S. W. Elson, K. H. Baggaley, M. Davidson, M. Fulston, N. H.
Nicholson, G. D. Risbridger and J. W. Tyler, J. Chem. Soc., Chem.
Commun., 1993, 1212.
(a)
2 I. Stirling and S. W. Elson, J. Antibiot., 1979, 32, 1125.
3 C. A. Townsend and M.-F. Ho, J. Am. Chem. Soc., 1985, 107, 1066.
4 C. A. Townsend and S.-S. Mao, J. Chem. Soc., Chem. Commun., 1987,
86.
5 A. Basak, S. P. Salowe and C. A. Townsend, J. Am. Chem. Soc., 1990,
112, 1654.
6 D. B. McIlwaine and C. A. Townsend, J. Chem. Soc., Chem. Commun.,
1993, 1346.
75.5
181.5
δ / ppm
(b)
7 C. A. Townsend, A. M. Brown and L. T. Nguyen, J. Am. Chem. Soc.,
1983, 105, 919.
8 M. Grayson and P. T. Keough, J. Am. Chem. Soc., 1960, 82, 3919.
9 C. A. Townsend, T. Scholl and D. Arigoni, J. Chem. Soc., Chem.
Commun., 1975, 921.
10 K. Akashi, R. E. Palermo and K. B. Sharpless, J. Org. Chem., 1978, 43,
2063.
11 H. C. Kolb, M. S. VanNieuwenhze and K. B. Sharpless, Chem. Rev.,
1994, 94, 2483.
12 S. W. Elson and R. S. Oliver, J. Antibiot., 1978, 31, 586.
13 C. A. Townsend and M.-F. Ho, J. Am. Chem. Soc., 1985, 107, 1065.
14 J. C. Vederas, Nat. Prod. Rep., 1987, 277.
15 S. P. Salowe, W. J. Krol, D. Iwata-Reuyl and C. A. Townsend,
Biochemistry, 1991, 30, 2281.
46.4
174.5
δ / ppm
Fig. 1 Partial 13C{1H} NMR spectra at 100 MHz of (a) [2,3,3-2H3,
1,2-13C2]-d,l -glycerate 12 in D2O, referenced to DSS; (b) biosynthetically
enriched p-bromobenzyl clavulanate in CDCl3, referenced to SiMe4, 21000
transients
Received, 17th October 1996; Com. 6/07078G
226
Chem. Commun., 1997