Influence of vanillin on growth and metabolism of Amycolatopsis sp. ATCC 39116
When Amycolatopsis sp. was cultivated with vanillin initially added in the medium over the range 2–10 g/L, the cells were observed to have no growth (shown by OD600), suggesting that vanillin has an almost completely inhibitory effect on cell growth when present at the lag phase (data not shown). When vanillin was provided 12 h after inoculation, that is, at the exponential phase, cell growth was significantly repressed subsequently. However, the cell density of the control culture was observed to have increased during the initial 20 h (Figure 1A). These results indicate that the presence of vanillin may also have a significant inhibitory effect on cell growth during the exponential phase. When vanillin was added to the broth of Amycolatopsis sp. at the stationary phase, the cell density was observed not to decrease over several concentrations (Figure 1B). Interestingly, growth inhibition due to the presence of vanillin at the stationary phase was observed to not be dependent on the concentration of vanillin added and in the case of 6 g/L supplementation seems to show a higher increase in cell density compared with other cases. An analysis of the broth samples during stationary phase revealed that the ferulic acid in the control culture (without vanillin addition) was completely consumed by 24 h; however, the broth with different separate additions of vanillin showed a very low consumption of substrate (Figure 2A) by 24 h. By 30 h, both ferulic acid and vanillin added in the broth were observed to disappear and different concentrations of vanillic acid and guaiacol were detected (Figure 2B). These results suggest that the presence of vanillin in the stationary phase (i.e. added at this phase) would have unfavorable effects on vanillin production, as it can be utilized as a substrate for cell growth thereby reducing its concentration, which was also observed previously.
Figure 1. The effect of added vanillin in the medium on cell growth of Amycolatopsis sp. ATCC 39116.
A. The inhibition of vanillin added at the exponential phase. B. The influence of vanillin added at the stationary phase. The arrows represent the time point of vanillin addition to the broth.
Download figure to PowerPoint
Figure 2. Metabolic analysis of transformation in the broth with different concentrations of vanillin added at the stationary phase.
A. An analysis of results of transformation at 24th h. B. An analysis of results of transformation at 30th h. Arrow in (A) represents that no ferulic acid is supplemented in broth. Arrow in (B) shows that no vanillin is detected at 30th h.
Download figure to PowerPoint
Several studies have investigated the toxicity of vanillin on the cell growth of recombinant E. coli and Pycnoporus cinnabarinus, but there have been no reports related to Amycolatopsis sp. ATCC 39116 and the influence of the presence of vanillin on the metabolism of producers has not been reported to date. The present results suggest that vanillin may have a significant inhibitory effect on cell growth during the lag and exponential phases and less of an effect at the stationary phase, and therefore vanillin transformations are often performed during or after the stationary phase. Nevertheless, even here, the presence of or produced vanillin may result in vanillin utilization, which is a significant drawback for the establishment of vanillin bioproduction. This work has shown for the first time the consumption of vanillin during the biotransformation by Amycolatopsis sp. ATCC 39116, and therefore the potential advantage of producing vanillin and simultaneously extracting it from the bioconversion broth during the stationary phase to avoid its further degradation.
Determination of polymer bead partition coefficients
The partition coefficients of candidate polymers are summarized in Table 1, which shows that Hytrel®G4078W and Hytrel G3548L have a superior partitioning capacity for vanillin but have a low affinity for ferulic acid, a desirable result. Both these polymers are block copolymers of poly (butylene terephthalate) and polyether, with differing amounts of hard and soft segments. Hytrel polymers are block copolymers of poly(butylene terephthalate) as the hard segment and polyether as the soft segment, with harder grades containing a larger proportion of poly (butylene terephthalate). Importantly, Hytrel is also an Food and Drug Administration compliant polymer for use in the food and fragrance industry, is odorless, biocompatible, and nonbioavailable, and is thermally resistant to sterilization even at 121 °C. Hytrel C4078W was thus a good option for this biotransformation and was used in the TPPB transformation by Amycolatopsis sp.
The difference in the uptake performance of polymers such as Hytrel G4078 W in absorbing vanillin and ferulic acid support a different mechanism than that of hard, resin polymers. Previously, macroporous adsorbent resins such as DM11 and XAD-2 resin showed relatively good uptake results for vanillin, although there is some confusion in the article as to whether the uptake of vanillin by these polymers is via absorption or adsorption. Generally speaking, adsorption by these resins is based on the interaction between the outer surface of the polymers and the target molecules and thereby their adsorption is limited by the polymer's surface area, whereas absorption by soft polymers such as Hytrel G4078W is by dissolution of the molecule into the physical structure of the polymers themselves.
Transformation at bioreactor scale
As there are no reports concerning the bioproduction of vanillin by Amycolatopsis sp. ATCC 39116 at bioreactor scale, transformations by this strain were performed at 280 and 400 rpm in a 5-L reactor, to provide single phase benchmark performance and establish appropriate conditions for further transformations. The time course of substrate consumption, relevant metabolites, and product accumulation are illustrated in Figure 3A,B. A maximum vanillin concentration of 9.4 g/L was reached in the 280 rpm case with a molecular yield of 89.2% by the time the substrate was depleted (Figure 3A), while the transformation at 400 rpm accumulated a concentration of only 6.8 g/L vanillin with a yield of 62% (Figure 3B). During the transformation at 280 rpm, the percentage saturation of dissolved oxygen was shown to decline quickly to 36.5% and the rpm was then adjusted to 370 to avoid possible oxygen limitation. This suggests that low stirring rate may be inadequate for the transformation. However, with a higher stirring speed of 400 rpm, the molecular yield and maximum concentration of vanillin were lower than that of the transformation at low stirring speed, which may be in part ascribed to the deleterious effect of stirring on cell viability and transformation performance in transformation as Amycolatopsis sp. is a filamentous cell and excessive stirring could destroy its morphology and thereby may decrease cell viability. It was indeed observed that the filaments of this strain operating at 400 rpm were shorter and more dispersed than those in the transformation at 280 rpm. In light of these, and the earlier results, ISPR was implemented in a 5-L TPPB at a stirring speed not exceeding 400 rpm.
Figure 3. Transformation of vanillin in single aqueous phase.
A. The transformation at 400 rpm. B. The transformation at 280–370 rpm.
Download figure to PowerPoint
Fed-batch transformation of vanillin in a TPPB
Figure 4 shows the time course for substrate, vanillin, and the main by-products including vanillic acid and guaiacol throughout the fed-batch biotransformation. Compared with the previous single phase result (Figure 3B), it is clear that operating in TPPB mode increased the transformation time and the amount of substrate addition. A similar quantity of substrate was consumed more quickly and more substrate was added into the reaction system than in the transformation in single phase (Figure 3B). It is also clear that the TPPB operating strategy was highly successful in improving reactor performance as a total supplementation of 48 g/L ferulic acid to the reactor was achieved over 56 h before the reaction rate decreased significantly. This run obtained a vanillin volumetric productivity of 450 mg/L h with a vanillin concentration of 8.9 g/L in the aqueous phase and 106 g/kg in the polymer phase (Figure 4 and Table 2). As reported previously, the volumetric productivity was calculated by dividing the final system vanillin mass by the total system working volume (aqueous plus polymer) and the biotransformation time. The final aqueous concentrations of by-products including vanillic acid and guaiacol in this operating reached 0.5 and 1.1 g/L, respectively. The reactor performance criteria (volume of substrate added before transformation termination, obtainable biotransformation time, and vanillin volumetric productivity) were greatly improved compared with those achieved in the single aqueous phase reactors as is summarized in Table 2, along with the two other reported solid–liquid phase reactors, which used adsorptive resin as the immiscible solid phase (Table 2).[4, 12] Additionally the “online” separation of vanillin from the broth (i.e. by absorption into the polymer) decreased the possibility of vanillin being consumed as an energy source or carbon source and thereby increased the total production of vanillin in the process. This investigation showed that by selecting an appropriate polymer, a solid–liquid TPPB system can greatly enhance the biotransformation process and achieve high final vanillin concentration and high productivity, although further polymer screening may identify an even more effective sequestering phase than Hytrel. Overall, the solid–liquid TPPB system showed promising results in removing a toxic and easily consumed product via an ISPR technique and is also an efficient initial concentrating step for downstream processing.
Figure 4. TPPB transformation of vanillin using a fed-batch strategy.
The arrows represent the time of substrate addition to the broth.
Download figure to PowerPoint
Table 2. Summary and Comparison of TPPB Bioreactor Performance
|Mode||Scale||Aqueous Vanillin||Polymer Vanillin||Polymer (w/v)||Volumetric Productivity (mg/L. h)|
|Single phase||3 L working volume reactor||9.4 g/L||NA||NA||268|
|Fed batch TPPB||3 L working volume reactor||Total 19.5 g/La||Hytrel® (10%)||450|
|Fed-batch||500 mL flask||Total 19.2 g/L||DM11(8%)||349|
|Batch||100 mL flask||Total 2.9 g/L||XAD-2 (50%)||60.4|
A closer examination of the vanillin curve in Figure 4 suggests that the aqueous concentration of vanillin before 30 h had been relatively invariable although the concentration of ferulic acid decreased rapidly. Also, the concentration of vanillic acid, a main by-product of vanillin by Amycolatopsis sp., was observed to decline with time after 38 h and these results suggest that the substrate was transformed to vanillin but did not accumulate in the aqueous solution; rather it must be taken up by the polymer beads. In addition, with each substrate addition after 38 h, the aqueous concentration of vanillin was shown not to increase in accordance with the consumption of substrate. These results also suggest that the application of the TPPB had reduced the concentration of vanillin in the aqueous phase thereby reducing its toxicity to cell growth as well as preventing biodegradation of vanillin in the aqueous phase.
Compared with the instantaneous equilibrium that occurs in a liquid–liquid TPPB system, the equilibrium between a solid and liquid phase requires more time due to the diffusion of the solute into the solid polymer matrix. The diffusion rates of vanillin into the polymer was not the focus of this study; however, there are some reports involving the diffusivities of similarly sized-molecules (benzene and phenol) into Hytrel polymers and ethylene–vinyl acetate.
As reported before, aside from the improved performance in volumetric productivity, increase in substrate addition and biotransformation time, there are several operational advantages over the two liquid–liquid phase bioreactor. In the case of vanillin, it is interesting to note that as a typical filamentous bacterium, the filaments of Amycolatopsis sp. ATCC 39116 are fragile and susceptible to mixing shear forces and that too strong a shear force could destroy the filaments. However, the cells in the TPPB transformation at 400 rpm still displayed branched-rich filaments, a morphology of this strain itself in liquid culture, and there were no any ndications of the destructive effects of strong shear force on cells as had been seen in the single phase bioreactor configuration (data not shown). The incorporation of the polymer beads in the bioreactor may decrease or buffer the destructive effects of strong shear force on the cells, which is significant for the performance of such strains.
In previous reports of two phase biotransformations with resin beads as the sequestering phase such as XAD-2, HZ802, and DM11, no information was provided concerning the adsorption of the by-products of vanillin by the resins; however, it is known that adsorption by such resins is nonselective. Therefore, such adsorptive resins may inadvertently bind essential nutrients or adsorb metabolites in the system potentially affecting performance. In case of vanillin bioproduction by Amycolatopsis sp., the accumulation of vanillic acid to 200 mg/L was thought to be a prerequisite for the initialization of vanillin production. Since vanillic acid is dissociated in the medium (pH 8.0), Hytrel polymers would not absorb such molecules.
Using polymer fractions of 0.1 kg/L yielded vanillin production enhancements of 19.5 g/L relative to the single-phase control. However, polymer fractions above 0.1 kg/L were not examined here. In other studies involving the adsorption of butanol by hard resins, excessive binding of medium components was reported to be the cause of decreased performance at high polymer fraction of resin. Increasing the proportion of these polymer resins in the aqueous phase may possibly result in further production enhancement as was addressed previously in the bioproduction of 2-phenylethanol in a TPPB. From this limited data set combined with the limitation of reactor volume, there appears to be an optimum polymer fraction in a reactor whose specification is a critical design element and will require further investigation.
Vanillin recovery and polymer regeneration
The ability to recover and purify vanillin from the polymer phase, thereby regenerating the polymer for subsequent reuse, was investigated using different organic solvents. The results of three extraction–regeneration cycles are shown in Table 3. No significant difference in extraction performance was seen among extractions by the three solvents, achieving a total desorption of 64.6% from the recovered polymer in the bioreactor by methanol. The decreased desorption of vanillin may be ascribed to volatilization during the drying of polymer beads. In addition, some guaiacol was desorbed from the polymer, although it is not predominant in the desorbed mixture. However, there was no detectable ferulic acid, vanillic acid, and other by-products present in the desorbed mixture. These results demonstrated that the application of TPPB did not unfavorably affect the metabolic pathway of vanillin as the accumulation of vanillic acid with 200 mg/L has been confirmed to be a necessary condition for vanillin production by Amycolatopsis sp. The above results also demonstrated that it is possible to efficiently recover the products from the beads using a small volume of methanol or other organic solvent. In addition to product recovery, the results also suggest that this downstream step allows for the regeneration of the polymer beads for future reuse using organic solvents generally used in the food industry. Product extraction from polymer beads in this manner may be much easier than the removal of products from a liquid organic extracting phase where distillation of the solvent or back extraction using another immiscible solvent is required. The use of polymers in TPPBs is confirmed again to be advantageous in the microbial bioproduction of flavor/fragrance compounds, as the polymers in medium do not produce any fugitive off-aromas or chemicals that could affect product quality or other organoleptic properties.
Table 3. Vanillin Recovery Capabilities Using Organic Solvent Extraction and Performance of Polymer Regeneration
|Organic Solvent||Total Vanillin Extraction Ratio (%)*||Total Ferulic Acid Extraction|
In addition, after extraction of blank polymers (without solute absorbed) using methanol for 24 h, it was observed that there were no chemical species in the methanol as detected using gas chromatography (100–300 °C of oven scanning), suggesting no inherent contaminants within the polymer itself. Vanillin was easily and quantitatively recovered from the polymers mostly by single stage extraction into methanol or other organic solvents used in food industry, simultaneously regenerating polymer beads for reuse.