Professor Athanassios Giannis, medicinal chemist and head of the Institute of Organic Chemistry at the University of Leipzig, describes how (2R,3S)-isocitric acid, a forgotten member of the chiral pool, has been made accessible in kilogram amounts for organic syntheses
The citric acid cycle is a main metabolic process. Discovered in 1937 by H. A. Krebs, all its intermediates are commercially available in multigram amounts - with one forgotten exception: (2R,3S)-isocitric acid (DS-threo-isocitric acid). However, isocitric acid could be a new member of the chiral pool, and a valuable starting material for both organic synthesis and pharmaceutical research.
In the past, substantial effort has gone into the commercial fermentation at a 100,000 ton scale of the constitutional isomer citric acid, for use in consumer products. The principle of these procedures consists of prompting microbes to exude excess citrate as a metabolite of the citric acid cycle into a fermentation medium from which it can be isolated. Surprisingly, no attempts have previously been made to obtain (2R,3S)-isocitric acid by fermentation.
Of course, isocitric acid has been studied intensively, and there are some 15,000 papers written on this material. The majority deal with biochemical and analytical problems. Interestingly, none of them describes its use as a small chiral molecule in organic syntheses. In this respect, isocitric acid was regarded as an inaccessible intermediate of the Krebs cycle, and nothing more. But looking at this situation from a distance, there is really no reason why one should not try to prompt a microbe to exude iso-citric acid in an analogous manner.
possible routes
The key for to this problem is hidden in the essence of the cycle itself: any means that allows a shift in the ratio of citrate to isocitrate should be a successful tool to finally exude excess isocitrate into the fermentation medium.
Two routes seem possible: create a transgenic organism or force an unaltered microbe to make more isocitrate than normal. The second route has the advantage of not requiring a lot of special safety precautions. However, it is a specialist art to find a micro-organism that is flexible enough in its metabolic behaviour and resistant enough towards metabolic stress situations to do the job.
The route to an appropriable isocitrate building block has three steps.
Isocitrate by fermentation: A large number of yeasts have been known to produce citric- and (2R,3S)-isocitric acid in varying ratios when grown on long-chain n-alkanes or glucose.1 It is astonishing that these fermentations were optimised for high levels of citric acid only. Yeasts are preferable to bacteria because the latter can sometimes be pathogenic.
The thiamine-auxotrophic yeast Yarrowia lipolytica, for example, excretes large percentages of organic acids when grown on vegetable oils with an excess of thiamine under nitrogen-limited, aerobic conditions. Our aim was to achieve the highest possible isocitric to citric acid ratio and therefore a high isocitric acid concentration.
Eventually, we succeeded in manufacturing isocitric acid concentrations of 93g/L and ratios of up to 1.14:1 (2R,3S)-isocitric acid 1 to citric acid2 (see scheme 1) on the 100-litre pilot plant scale in the cultivation of wildtype Y. lipolytica EH59 on refined sunflower oil. This marks a hitherto un-rivalled achievement, particularly with regard to the use of cheap renewable vegetable raw materials.
This should not be underestimated from the viewpoint of sustainable chemistry. Work-up of the fermentation broth follows standard procedures: after filtration of the biomass, electrodialysis was performed to convert the obtained trisodium salts into the free acids; finally, water is removed under reduced pressure. Comparing this fermentative route to isocitric acid with any hypothetic synthetic route, it is obvious that the biotech method will be superior in terms of reducing environmental pollution and undesirable waste.
knowledge gap
Eventually, an adequate process to separate the two isomers had to be developed. This was made more problematic because the databases do not list complete physicochemical characteristics of pure (2R,3S)-isocitric acid and derivatives. A real knowledge gap still exists here. But the opportunity was to find an elegant and straightforward separation.
Separation of isocitric acid from citric acid: Citric acid and isocitric acid are constitutional isomers and have the same kind of functional groups. This means that a lot of physicochemical properties - polarity, acidity and solubility - are likely to be similar.
From the beginning, it was our aim to find out and use non-chromatographic separation methods that allow for an efficient scaling up to the kilogram scale. Whereas citric acid shows a good tendency for crystallisation, its mixture with isocitric acid obtained by water removal remains a syrup (later it was found that pure (2R,3S)-isocitric acid itself is syrupy). It is a basic property of carboxylic acids that among all possible esters, the methyl esters show by far the highest melting points if such esters crystallise at all.
The idea of forming the methyl esters of both acids and making use of their differing properties proved to be the key to the separation problem. Whereas trimethyl citrate is a solid of mp 75 °C , trimethyl isocitrate proved to be a liquid (hitherto unknown). This circumstance allows a principal separation of both esters by simple suction filtration. In this way enantiopure trimethyl isocitrate was obtained.
inherent symmetry
Reflecting upon this separation, it is obvious that a difference in one structural feature can be used to facilitate the separation. The key is in the compounds" inherent symmetry. Whereas citric acid is highly symmetrical, (2R,3S)-isocitric acid lacks any sym-metry but has, at the same time, a large number of degrees of freedom of rotation. This property is passed on to its trimethyl ester. Hence, the assembly of a crystal lattice is apparently an unfavourable process, and the compound remains in the liquid state.
(2R,3S)-isocitric acid as a chiral building block: Before any reactions are shown, the structure will be discussed. In this small chiral molecule a lot of functional groups are combined: three carboxylic acids and one hydroxyl group. Hence, it is necessary to establish a method to distinguish between the carboxylic acid groups in order to achieve regioselectivity. From the viewpoint of possible functional group transformations the potential of this compound is excellent. Isocitric acid has a high degree of oxidation. This should not be regarded as a problem; it is helpful to make certain compounds but does pose a challenge to the synthetic chemist.
However, the molecule offers various other possibilities besides reduction and functional group transformation, such as cyclisation and transformation into various heterocycles.
The molecule contains two chiral centres and can be used as a typical chiral compound in asymmetric synthesis. This offers the greatest opportunities for future applications. Both stereogenic centres contain C-H-acidic groups and can, therefore, be exploited for epimerisation (the changing of one stereo form of a compound into another by enzymatic action) under certain conditions. Regioselective epimerisation using mild basic conditions was successful and afforded isocitric lacton (3 in scheme 2) in enantiopure form.
In contrast to citric acid, isocitric acid forms a five-membered lactone that can be further converted into a cyclic anhydride (4 scheme 2). This highly appealing structure proved to be of paramount importance for further transformations. Its opening with various heteroatom nucleophiles proceeds with complete regioselectivity and gives rise to a range of lactone mono acids. By this stage, the differentiation of the three carboxylic acid moieties has already been accomplished, rendering this isocitric acid derivative accessible to further selective transformations.
By short sequences, non-naturally occurring lactone amino acids of type 5 in scheme 2 as well as isocitric acid derivatives of type 6 in scheme 2 can be obtained; all these synthetic manipulations take place with nearly quantitative yields and therefore again highlight the outstanding synthetic utility of this new member of the chiral pool.
promising applications
In summary, (2R,3S)-isocitric acid, a substance from the citric acid cycle that hitherto was of only analytical interest, is for the first time being produced on the kilogram scale. The new building block is made by fermentation of sunflower oil with the non-pathogenic yeast Yarrowia lipolytica EH59. Isocitric acid is then separated from citric acid by a convenient crystallisation step of the methyl esters that does not require chromatography and allows up-scaling.
First attempts at its transformation to new chiral building blocks point to its promising applications in natural product synthesis and its use as starting material in the pharma industry.