![[BioMatNet Database - FAIR Program]](../images/Fair.gif) |
FAIR-CT95-1099
Design and Scale-up of a Bioprocess for the Production of Natural Vanillin from Agricultural By-Products |
 |
Proposal No: | FAIR-CT95-1099 |
| Date Prepared: | September1999 | |
| Source: | Final Report
Abstract and
Executive Summary
Second Annual Progress Report |
Final Report Abstract
Source: Final Report, February 2000
Consortium: The contract was co-ordinated by the Pernod-Ricard Research Centre, Creteil (France), in partnership with Institute of Food Research, Norwich (UK), Laboratoire de Biochimie et Technologie des Glucides, INRA Nantes (France), Laboratoire de Biotechnologie des Champignons Filamenteux, INRA Marseille (France), Centre for Systems Engineering and Applied Mechanics, Louvain La Neuve (Belgium), agro-Industries recherche Et Developpment, Pomacle (France), Department of Applied Chemistry and Microbiology, University of Helsinki (Finland), Amersham-Pharmacia Biotech Europe GmbH, Freiburg (Germany).
Abstract
Introduction
Vanillin is one of the most widely used aromatic molecules, particularly in food but also in the pharmaceutical and fragrance industries. Considering the increasing consumer preference for natural flavours, our work was focused on the production of a biotechnological vanillin from a natural raw material by use of micro-organisms under the requisite conditions. Ferulic acid is an extremely abundant naturally occurring phenolic compound, with a molecular structure close to that of vanillin. It occurs in common agricultural residues such as cereal brans and sugar-beet pulp, and was therefore chosen as precursor for transformation to vanillin by fungi.
First, enzymatic fungal pathways of ferulic acid transformation leading to vanillin were studied for a better understanding of the bioconversion. The hypothesis of ferulic acid P- oxidation producing vanillic acid could not be confirmed, either in Aspergillus niger 1-1472 nor in Pycnoporus cinnabarinus MUCL 39533. Many attempts were performed for the study of vanillate reductase, responsible for the transformation of vanillic acid to vanillin. Starting from vanillic acid, no activity was conveniently detected to conclude. Therefore labelled ferulic acid was used in order to establish an efficient activity test. The activity responsible for vanillin production from ferulic acid was found in the cytosolic fraction of the fungal cell. When this fraction was boiled, the metabolism was inhibited, proving that the metabolism of ferulic acid to vanillin was associated to enzymes. From these results, further experiments were conducted in order to isolate the enzymes, but no real pre-purification was obtained. Vanillate hydroxylase was also studied, which is a well-known enzyme in other fungi, decarboxylating vanillic acid into methoxyhydroquinone. Vanillic acid metabolism was studied by cell free extracts, and from fractions of chromatographic runs, prepared from the white-rot fungus Pycnoporus cinnabarinus. Several, up to 4-5 intracellular cell-free enzyme activities were described for the first time in this fungus, and one activity for the first time in any fungus. Firstly, vanillic acid was decarboxylated to methoxyhydroquinone (MHQ) oxidatively by NADPH dependent vanillate hydroxylase. Secondly, another compound was also frequently found during the enzyme activity assay procedure, namely hydroxyhydroquinone (1,2,4- trihydroxybenzene, THB). The third enzyme activity found in the cell-free system was the vanillic acid reducing enzyme (VAR) or enzyme complex. The enzyme catalysing the reduction of vanillin to vanillyl alcohol, vanillin reductase (VR), was also found in cell free fractions from P. cinnabarinus. Formation of P-benzoquinone and hydroquinone may indicate that P. cinnabarinnus produces 1,4-benzoquinone reductase. Despite many further experiments, the vanillate hydroxylase could not be purified.
The optimisation of vanillic and ferulic acids bioconversion by P. cinnabarinus was performed in bioreactors using synthetic pure molecules. First different reactor designs and incubation conditions were studied. The bench- top bioreactor with a marine impeller was chosen because it is a good compromise between fungal growth (low shear stress) and vanillin production (not too high oxidative conditions enhancing n - methoxyhydroquinone formation). With vanillic acid as precursor and in maltose medium, 1,2 g/l of vanillin were obtained in 7 days after optimisation of the inoculum, the precursor feeding and aeration. Various medium composition were tested for growth improvement compared to the reference medium with maltose. By increasing yeast extract concentration and adding other carbon sources such as glucose, fructose or phospholipids, high density cultures were obtained of more than 10 g/l dry weight after 3 days. Vanillin production was then tested in the various media and optimised for the incubation temperature, pH regulation, aeration conditions and precursor and cellobiose feedings. Finally, use of glucose-maltose or glucose-phospholipids media allowed to produce more than 600 mg/l of vanillin from ferulic acid in 10 days. The use of XAD2 resin with specific adsorption toward vanillin reduced the vanillyl alcohol formation. In this way 1.6 g/l of vanillin were obtained from vanillic acid and 800 mg/l from ferulic acid. Mathematical modelling of the bioconversion was done for vanillic and ferulic acids using experimental data in a bioreactor. The model was then cross-validated by new experiments, that were not used for its establishment. It was also modified considering the influence of essential parameters (like mycelial death or dissolved oxygen). The cross-validation was almost perfect with vanillic acid as precursor. As more reactions are involved in the bioconversion of ferulic acid, the predictions of the model were not so good in this case, but convenient enough to be used in simulation studies. These studies have been carried out for the determination of the optimal strategies to maximise the vanillin production. Extensive simulations of the vanillin obtained for a given feeding rate or a given addition time of the precursor allowed to determine the optimal feeding rates and an optimal switching time. Finally the optimal duration of the biotransformation was also determined with respect to the competitivity of the process.
A complete process was established for production of cellobiose enriched fractions from sugar-beet pulp. A preliminary heat treatment was applied before enzymatic treatment by Celluclast 1.5L. The scale-up was successful and allowed to obtain larger amounts of cellobiose-enriched fractions. The produced fractions were tested for improvement of vanillin production from vanillic and ferulic acids in P. cinnabarinus cultures. Same inhibition of methoxyhydroquinone formation was observed as with pure commercial cellobiose. The production of ferulic acid from various raw materials was optimised. In order to improve accessibility of the enzyme to the substrates, two physical pre-treatments of maize bran have been explored. Autoclaving was preferred to flash explosion as it solubilised arabinoxylans as well as ferulic acid in a very efficient way for maize bran. Treating the samples for one minute at 190·C or at 160 ·C for l hour gave an optimal yield of 63% in weight of solubilised material. In both case small oligomers are produced and ferulic acid is mainly esterified to sugars. Same treatments were applied to wheat bran and sugar-beet pulp with good success. However the best raw material for ferulic acid production is maize bran (better yield, higher amount). In order to get free ferulic acid, subsequent treatment with a feruloyl esterase is then necessary. The optimisation of the production of free ferulic acid has shown that a high concentration of enzymes was required but the process was very efficient. Extraction of esterified ferulic acid has been optimised at pilot scale, i.e. 20 kg of linked ferulic acid were obtained in-powder form after autoclaving at 140'C for 3 hours. Then the various fractions were tested in A. niger and P. cinnabarinus cultures, respectively for vanillic acid and vanillin production. A preliminary study has shown that A. niger I-1472 and P. cinnabarinus MUCL 39533 were able to use fractions containing free ferulic acid (i.e. after enzymatic hydrolysis by commercial enzymes) with so good yields as pure synthetic ferulic acid. In contrast, neither fungi was able to transform esterified ferulic acid in the standard conditions (i.e. without any inducer for esterase production).
An economic study showed that a production of free ferulic acid would cost ten fold more than a production of esterified ferulic acid. Therefore a large study was performed to study the cell-wall degrading ability of both fungi. This included the analysis of the cell-wall itself and the preparation of substrates (esterified ferulic acid and dimers) for enzyme activity tests. No esterase activity was measured in P. cinnabarinus MUCL 39533 cultures whatever the carbon source was. Giving A. niger I-1472 raw materials induces the production of a large spectrum of enzymes. Esterase activity was the highest on maize bran and then on sugar beet pulp. More complete study on FAE III activity of A. niger showed that this enzyme in particular was able to release free ferulic acid from most of the mono- or di- esterified ferulate dimers.
These results led to a search for alternative strategies for vanillin production:
Sugar beet pulp was chosen as the best inducer for enzyme production in A. niger cultures and autoclaved maize bran was shown to be the best source of ferulic acid. With these conditions and after optimisation in 2-litre fermenter, vanillic acid production by A. niger was successful : the best result was 2.65 g/L on day 13 of cultivation with a molar yield of 96 %. The produced vanillic acid was recovered and concentrated before its use as precursor in a P. cinnabarinus 2-litre fermenter. Vanillin production reached 530 mg/L after 13 days with a molar yield of 73% and 691 mg/L with a molar yield of 78% when adding XAD2 resin.
For the second strategy, enzyme production by A. niger was carried out in fermenter (20 and 200 litres) with success. The enzymes were concentrated by ultrafiltration with a ratio of 5 to 10 fold and a recovery yield of 42 to 67 % depending on the followed activity (glucosidase or esterase). The filtered culture broth was used in P. cinnabarinus cultures fed with autoclaved maize bran. Vanillin production was optimised in flask and allowed to produce 687 mg/l in eleven days with 32 % molar yield. The addition of resin reduced the quantity of vanillyl alcohol and 802 mg/l of vanillin were obtained with 41 % molar yield. In 2-litre fermenter, the same results were reproduced. Experiments with concentrated enzymes were successful in flask but vanillin concentrations were lower in fermenter (less than 200 mg/l in 15 days). The scale-up in 20 litre fermenter was done and allowed to produce natural vanillin from autoclaved maize bran. Experiments are still in progress for improvement of this final step especially with the use of concentrated enzymes.
The natural biotechnological vanillin produced from maize bran was purified using LH20 column. This column was chosen for its specificity toward vanillin allowing 95- 98% recovery yield and 90% purity directly from the fermentation broth. Organoleptic study was done on liquid extracts. Our biotechnological vanillin exhibited a strong vanilla flavour and taste, like ethylvanillin, but with a woody phenolic secondary note. Isotopic ratio 13C/12C was measured on the crystallised vanillin and the - 18.3 %o PDB obtained value is close to the ratio obtained with pure vanillin from vanilla bean.
Source: Progress Report for 1/11/97 to 30/10/98
Introduction
This project is focused on the design and scale-up of an industrial biotechnological process for the transformation of agricultural by-products into high added-value natural vanillin. This non-food valorisation (flavours are considered as additives but not as food) will be realised on the basis of the experience and results obtained at the laboratory level in the previous project ( AIR1-CT92-0026) New processes for the biological transformation of agricultural residues for the production of high added value flavours.
The FAIR project addresses all problems associated with the transfer of basic research to industrial scale and is divided into four phases
Activities
The overall aim is the production of natural vanillin by a two-step procedure:
Because of inhibition of the growth of the fungus by the final product, a continuous process for vanillin extraction during the bioconversion is required. The complexity of this transformation also requires a high level of control. In addition it is necessary to define an analytical method that can be used to prove the origin of the vanillin obtained (SNIF-NMR, isotopic analysis). It is expected that several kg of vanillin will have been produced by the end of the project. The economic viability of the process can then be evaluated, taking into account the price of vanilla beans and the yield of our process.
Progress
Phase A: Preparation of fermentation feedstock by fractionation of plant cell walls Amplification of the gene coding for the CinnAE activity is in progress. Production of CinnAE and FAE-III are currently performed at fermentor scale. Pilot-scale amounts of ferulic acid have been obtained from wheat bran using this enzymatic technology. Importance of the dimers of ferulic acid in the structure of the cell-wall polysaccharides has been shown. Activities able to attack these dimers were found in A. niger cultures and in A. oryzae pure tannase.
Some physical pre-treatments of the maize bran were compared to improve the efficiency of the enzymatic hydrolysis. Autoclaving gave the best results, solubilising up to 80% of the ferulic acid as esters of sugars. The previous heat treatment was also tested on wheat bran and sugar beet pulp. Extraction of linked ferulic acid has been optimised at large scale. The ferulic acid can be freed by the enzyme defined in the previous project. The complete process gives more than 1 kg of ferulic acid per 100 kg of maize bran.
Two strains, A. niger I-1472 and P. cinnabarinus MUCL-39533 were studied for their production of polysaccharide-degrading enzymes on various raw materials. The A. niger strain was the most interesting, producing a wide spectrum of enzymes. A crude extract of this strain was able to degrade sugar beet pulp or autoclaved maize bran and to produce free ferulic acid. The same crude extract is also able to transform the ferulic acid into vanillic acid.
Use of cellulose to obtain cellobiose (to be used as a carbon source in the subsequent bioconversion) has been optimised. Heat treatment was also tested. The production was performed at pilot scale and production yield is 50% of the cellulose content of the solid fraction, choosing heat treatment as enzymatic hydrolysis. This process will be developed at large scale. The cellobiose produced at pilot scale was used with success for P. cinnabarinus trials.
Phase B: Modification of fungal pathways Involvement of P-oxidation in the first step of the bioconversion (ferulic acid to vanillic) is still not demonstrated. But some hypothetical metabolic intermediates of the 0-oxidation were synthesised and tested in vivo. The first results seemed to confirm the hypothesis of beta-oxidation but they must be confirmed by in vitro experiments.
The VAR enzyme responsible for the last reaction (vanillic acid to vanillin) has been studied in vitro. A test of enzyme activity has been established but it was not possible to purify the enzyme. Using a more sensitive method to assay the VAH activity (responsible for the production of the by-product MHQ) it was possible to detect the activity. However, the VAH enzyme has not yet been purified.
Phase C: Industrial fermentation process scale-up Optimisation of the two-step and one-step processes in fermentors gave way to final concentration of vanillin up to 1.6 g/l using adsorbent in the medium. A model based on observation has been calculated to explain the mechanisms of formation of the phenolic compounds. Natural ferulic acid produced from maize bran and sugar beet pulp was tested with success at laboratory scale. Also, cellobiose from natural origin was tested with success. A continuous extraction device has been defined, using first a separation and then a trapping of the phenolics, with subsequent testing on 2-litre fermentor.
Phase D: Authentication and labelling A 20-litre fermentor was used to produce 15 litres of culture medium, containing more than 500 mg/l vanillin. This vanillin is being purified.
Activities
The following milestones were achieved:
Future activities
Activities in the third year will:
© Copyright 2006 Policy Statements
Updated
by CPL Press:
03/07/2007
- biomatnet@biomatnet.org
 |
 |
 |
News |
 |
Events |
|
|