Improved asymmetric syntheses of (R)-(-)-homocitrate and (2R,3S)-(-)-homoisocitrate, intermediates in the α-aminoadipate pathway of fungi

Article abstract of DOI:10.1016/S0040-4039(00)01707-X

Improved asymmetric syntheses of the title compounds are reported. Both products are produced through diastereoselective alkylation of malic acid; (R)-homocitrate synthesis uses the self-regeneration of stereocenters approach. Both procedures represent an improvement in yield over existing methods without loss of stereoselectivity. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01707-X

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 9209–9212
Improved asymmetric syntheses of (R)-(−)-homocitrate and
(2R,3S)-(−)-homoisocitrate, intermediates in the
a-aminoadipate pathway of fungi
Guoxiang Maa and David R. J. Palmera,b,
*
aDepartment of Chemistry, University of Saskatchewan, Saskatoon, SK, Canada S7N 5C9
bDepartment of Biochemistry, University of Saskatchewan, Saskatoon, SK, Canada S7N 5C9
Received 21 September 2000; accepted 28 September 2000
Abstract
Improved asymmetric syntheses of the title compounds are reported. Both products are produced
through diastereoselective alkylation of malic acid; (R)-homocitrate synthesis uses the self-regeneration of
stereocenters approach. Both procedures represent an improvement in yield over existing methods without
loss of stereoselectivity. © 2000 Elsevier Science Ltd. All rights reserved.
(R)-(−)-Homocitrate and (2R,3S)-homoisocitrate are natural products known to be intermedi-
ates in the a-aminoadipate pathway of lysine biosynthesis.1 This pathway is found primarily in
fungi, although it has recently been shown to be present in some archaea. Because this pathway
is absent from plants and animals, it is considered a promising target for anti-fungal therapy.2
Fungal infections are a severe medical problem, especially among immunocompromised patients
such as AIDS sufferers, organ transplant recipients, and patients undergoing chemotherapy. The
need for novel anti-fungal therapies has been expounded upon in recent reviews.3 (R)-Homo-
citrate is also a compound of interest because it is a required component of the active site of the
iron-molybdenum protein of nitrogenase.4
One aspect of the a-aminoadipate pathway is poorly understood.1,5 Homocitrate is converted
to homoisocitrate in a series of reactions apparently in simile with the conversion of citrate to
isocitrate by the enzyme aconitase (Fig. 1). While the intermediates homocitrate, homoaconitate
and homoisocitrate have been isolated from Saccharomyces cerevisiae mutants, which have been
impaired in their ability to produce lysine, the process by which homocitrate is converted to
homoisocitrate has been characterized so poorly that it is not clear whether one enzyme or two
are involved. Our laboratory is currently investigating this portion of the pathway to resolve this
issue. One of the factors which has confounded the study of this process is the lack of
* Corresponding author.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01707-X

9210
availability of the compounds shown in Fig. 1: only racemic homocitric lactone is available
commercially (at >US$1000/g). Nitrogenase research has been similarly affected. While models
of the FeMo protein active site continue to be investigated,6 (R)-homocitrate, a required
component of the enzyme’s reaction center, has not been incorporated into any such models.
Figure 1. Initial steps of the a-aminoadipate pathway. Lys20=homocitrate synthase, Lys4=‘homoaconitate
hydratase’. The enzyme catalysing the second step, perhaps Lys4, has not been identified experimentally.
Only one asymmetric synthesis of homocitric lactone has been published.7,8 The initial steps
of this synthesis were based on the work of Seebach,9 using trimethylacetaldehyde as a
protecting group to convert
L
-lactic acid to an asymmetrically hindered chiral alkene. This
compound was then converted to (R)-(−)-homocitric lactone in four steps with a reported yield
of 7.5% (3% from
results in 92% ee.
L
-lactic acid). The key step in this scheme is a Diels–Alder addition which
We felt that the Seebach strategy might be more effective using a diastereoselective alkylation
rather than cycloaddition. To this end, we pursued the procedure shown in Scheme 1.
Trimethylacetal-protected
D
-malic acid (1) was synthesized and alkylated with allyl bromide.9
The planar a-carbon of the dilithium enolate can be alkylated stereofacially (>95% diastereo-
selectivity) by virtue of the bulky protecting group. This is referred to as the self-regeneration of
stereocenters (SRS) approach.10 Esterification of the carboxylic acid, followed by hydroboration
of the alkene with oxidative workup gave 3 in 25% yield from 2. Deprotection gave the
trisodium salt of (R)-homocitrate.11 The overall yield from
D
-malic acid was 12%.
Scheme 1. (a) 2 equiv. of LiHMDS/THF, −78°C. (b) Allyl bromide. (c) AcCl, reflux 1 h; MeOH, reflux 4 h. (d)
BH3·SMe2/CH2Cl2, 25°C, 4 h. (e) Pyridinium dichromate/DMF, 25°C, 16 h. (f) 50% trifluoroacetic acid, 95°C, 2 h.
(g) 3% NaOH, 1 equiv., 25°C, 2 h.
This method, using the SRS approach, represents a substantial improvement over the
literature procedure. Considering the potential of SRS it is surprising how rarely this method
has been reported by researchers other than Seebach himself.12 Our results reinforce how
practical the SRS approach can be for the synthesis of carboxyl-containing compounds. If low

9211
cost is a dominant criterion to enantiomeric purity, the cost of starting material can be defrayed
by preparing
D
-malic acid from
D
-aspartic acid (about US$1.50/g).13 Reaction of
-aspartic acid
D
with HNO2 proceeds with retention of configuration, an ee of >95%, and is purified from the
side product 2-chlorosuccinic acid by separation of the dimethyl ester. We can isolate
acid in 80% yield by this method.
D
-malic
It should be noted that the alkylation procedure is only suitable for active electrophiles.
Seebach’s procedure reports alkylations with methyl iodide, allyl bromide and benzyl bromide,9
and these reactions have been successfully repeated in our laboratory. A more direct route to (R)-
homocitrate was sought using less activated electrophiles 3-bromopropionitrile and 2-(2-bromo-
ethyl)-1,3-dioxolane, but in our hands these electrophiles did not react with 1.
(2R,3S)-Homoisocitrate is no longer commercially available.14 The previously reported asym-
metric synthesis of this compound15 used a diastereoselective alkylation of dimethyl
D
-malate
using trimethylsilylpropargyl bromide, resulting in a 51% yield of a 9:1 mixture of diastereomers,
which were separated after acetylation by column chromatography. The alkyne was then treated
with benzenesulfenyl chloride to give a vinylic addition product from which chlorotrimethyl-
silane was eliminated using fluoride ion. The resulting alkynyl thioether was cleaved in the
presence of Hg2+ to give the carboxylic acid. Deprotection gave the desired product in an overall
yield from dimethyl -malate of 21%. In our hands, however, hydrolysis of the alkynyl thioether
D
resulted in a complex mixture.
We have adapted this procedure by substituting allyl bromide for trimethylsilylpropargyl
bromide in the alkylation step, as shown in Scheme 2. Note that allyl bromide is less than 1/50
the cost of the trimethylsilylpropargyl bromide, and reacts with the dianion of dimethyl malate
with improved diastereoselectivity (19:1) and comparable yield (80%). This 3-allylmalic dimethyl
ester was then acetylated in 91% yield to give 4. Hydroboration of the alkene 4, followed by
deprotection gave the desired compound in an overall yield of 30% from dimethyl
D
-malate.16
In conclusion, we have synthesized two important natural products with high stereoselectivity
and improved yields over literature methods. We have purposefully sought a unified approach
to minimize the method development time for laboratories wishing to repeat both these
preparations. We hope that improved syntheses of these compounds can have an immediate
impact on research into the a-aminoadipate pathway of fungi and on the investigation of
nitrogenase and nitrogenase models.
Scheme 2. (a) LDA/THF, −78°C; allyl bromide. (b) Ac2O, Et3N, DMAP. (c) BH3·SMe2/CH2Cl2, 25°C, 2 h. (d)
Pyridinium dichromate/DMF, 25°C, 16 h. (e) 2.5% KOH, 1 h; Dowex 50X8-100(H+).
Acknowledgements
This work has been supported by a grant from National Sciences and Engineering Research
Council and by the University of Saskatchewan. The authors gratefully acknowledge helpful
discussions with Professor Marek Majewski.

9212
References
1. Zabriskie, T. M.; Jackson, M. D. Nat. Prod. Rep. 2000, 85–97.
2. Suvarna, K.; Seah, L.; Bhattacharjee, V.; Bhattacharjee, J. K. Curr. Genet. 1998, 33, 268–275.
3. See, for example: Alexander, B. D.; Perfect, J. R. Drugs 1997, 54, 657–678.
4. Hoover, T. R.; Imperial, J.; Ludden, P. W.; Shah, V. K. Biochemistry 1989, 28, 2768–2771.
5. Gamonet, F.; Lauquin, G. J.-M. Eur. J. Biochem. 1998, 251, 716–723.
6. Couvouvanis, D.; Konstantinos, D. D.; Chang, G. K.; Dunham, R. W.; Kampf, J. W. J. Am. Chem. Soc. 1993,
115, 2244–3345.
7. Rodriguez, G. H. R.; Biellmann, J.-F. J. Org. Chem. 1996, 61, 1822–1824.
8. Preparation of (R)-homocitrate by resolution of diastereomeric salts has also been reported, Ancliff, R. A.;
Russell, A. T.; Sanderson, A. J. Tetrahedron: Asymmetry 1997, 8, 3379–3382.
9. Seebach, D.; Naef, R.; Calderari, G. Tetrahedron 1984, 40, 1313–1324.
10. Seebach, D.; Sting, A. R.; Hoffmann, M. Angew. Chem., Int. Ed. Engl. 1996, 35, 2708–2748.
11. Note that 1 and 2 are novel compounds only in the sense that they are enantiomers of compounds prepared in
Ref. 9. Physical properties matched those reported previously, except for the direction of optical rotation.
Compound 3 (oil): 1H NMR (CDCl3) l=10.91, 5.05, 3.61, 2.78, 2.42, 2.20, 2.02, 0.87. 13C NMR l=178.2, 173.6,
168.7, 108.2, 79.4, 52.3, 39.1, 34.3, 28.3, 27.8, 23.7. [h]25 (CHCl3, c=2.4)=−11.20°. High resolution mass [M+1]
calcd (found) for C13H20O7: 289.1287 (289.1295). Physical properties of trisodium (R)-homocitrate and its
corresponding lactone matched those previously reported.
12. Other applications reported: Juaristi, E.; Lopez-Ruiz, H.; Madrigal, D.; Ramirez-Quiros, Y.; Escalante, J. J. Org.
Chem. 1998, 63, 4706–4710. Einhorn, J.; Einhorn, C.; Ratajczak, F.; Gautier-Luneau, I.; Pierree, J.-L. J. Org.
Chem. 1997, 62, 9385–9386. O’Donnell, M. J.; Fang, Z.; Ma, X.; Huffman, J. C. Heterocycles 1997, 46, 617–630.
13. Winitz, M.; Bloch-Frankenthal, L.; Izumiya, N.; Birnbaum, S. M.; Baker, C. G.; Greenstein, J. P. J. Am. Chem.
Soc. 1956, 78, 2423–2430.
14. Homoisocitrate was until recently available from Research Plus, Inc.
15. Schmitz, C.; Rouanet-Dreyfuss, A.-C.; Tueni, M.; Biellmann, J.-F. J. Org. Chem. 1996, 61, 1817–1821.
1
16. Compound 4 (oil): H NMR (CDCl3) l=5.65 (m 1H), 5.15 (d, J=4.8, 1H), 4.94–5.00 (m, 2H), 3.65 (s, 3H), 3.59
(s, 3H), 3.02 (m, 1H), 2.44 (m, 1H) 2.18 (m, 1H), 2.03 (s, 3H). 13C NMR l=171.2, 170.0, 169.2, 134.2, 118.3,
71.4, 52.6, 52.2, 46.3, 31.9, 20.5. [h]25 (CHCl3, c=1.77)=+0.71°. High resolution mass [M+1] calcd (found) for
C11H16O6: 245.1025 (245.1020). Physical properties of (2R,3S)-homoisocitric acid matched those previously
reported.
.