Beyond bottlenecks: expression of complementary enzymes and permeabilization of cell membranes to improve performance of CYP153A6 in Escherichia coli whole cell biocatalysis

White, Bronwyn Elizabeth

Beyond bottlenecks: expression of complementary enzymes and permeabilization of cell membranes to improve performance of CYP153A6 in Escherichia coli whole cell biocatalysis

Doctoral Thesis

2022

Abstract

Cytochrome P450s are a diverse and versatile class of enzymes, able to carry out oxyfunctionalisation of hydrocarbons with exceptional regio- and stereospecificity. They show great promise within the medical and fine chemical industries, and could potentially be applied for the activation of linear alkanes into platform chemicals. However, bulk chemicals require higher titres, rates and yields than fine chemicals to be economically viable. Application of these enzymes in vitro is hampered by low biocatalyst stability, inhibitory upstream processing costs, and the need for costly cofactor supplementation. In vivo operation can overcome these issues, but brings its own set of limitations. This thesis presents research on the in vivo biotransformation of n-octane to 1-octanol by CYP153A6 and its natural redox partners, ferredoxin (Fdx) and ferredoxin reductase (FdR), from Mycobacterium sp. HXN-1500. These three proteins were heterologously expressed in Escherichia coli BL21DE3. Experiments were carried out at the millilitre scale, using resting cells. Potential limitations for the whole-cell biocatalyst include insufficient expression of the CYP153A6 or its redox partners, leading to a lack of biocatalytic “active sites”; lack of oxygen for biotransformation and cell metabolism; poor transport of n-octane from the organic phase into the cytoplasm, leading to limitations in substrate; toxicity resulting from the accumulation of substrate or product; and insufficient availability of NADH cofactor. With regards to heterologous enzyme expression, the decision was made to focus on CYP153A6 expression levels (with FdR and Fdx expression levels forming part of a related project). To investigate the effects of intracellular concentration of active CYP153A6, it was necessary to find a reliable method of varying P450 expression. The haem precursors δ-aminolevulinic acid (δ-ALA) and ferric chloride were applied during the growth and expression stage, and it was found that P450 expression could be optimised by varying the concentration of these chemicals, with co-addition of the two chemicals having a synergistic effect on expression levels. However, higher intracellular levels of active CYP153A6 did not lead to increased biotransformation rates or product titres in biotransformations carried out with whole cells. This demonstrated that the biotransformation of n-octane was not limited by the availability of active CYP153A6, even at concentrations as low as 0.3 µmolP450.gDCW-1 . Furthermore, these results suggest that optimising the expression of CYP153A6 in whole-cell biocatalysts does not equate to maximising its expression. Aminolevulinic acid constitutes a substantial portion of media formulation costs in the proposed whole-cell process, and these findings allow for more efficient application of this and other precursor chemicals. Based on an analysis of the literature, it was hypothesised that oxygen was unlikely to be a limiting substrate, due to the relatively low biotransformation rates being recorded, as well as the reduced oxygen demand of resting cells. Nevertheless, two sets of experiments were carried out to assess the effects of oxygen supply to the system. In the first set, the quantity of oxygen in the headspace was increased by enlarging the vials, while keeping cross-sectional area constant. In the second set, the headspace was replenished at regular intervals to increase the average concentration of oxygen in the headspace over the course of the biotransformation. In both sets of experiments, the differences between lower-oxygen vials and higher-oxygen vials were neither substantial nor consistent. This demonstrated that biotransformations by resting cell cultures at a density of 4 – 5 gDCW.L-1 were not limited by the quantity of oxygen available. The transport of hydrocarbons across the cell envelope affects the availability of n-octane substrate in the cytoplasm, and impacts on the in situ extraction of 1-octanol product. Thus, hydrocarbon transport has the potential not only to be a rate-limiting step, but also to control levels of cytotoxicity. To explore the significance of cross-membrane transport, cell membranes were permeabilised via the application of chemicals or via mechanical breakage, and biotransformations were also carried out using cell free extract. To investigate cofactor limitations, the NADH regeneration rate of E. coli was enhanced through over-expression of glycerol dehydrogenase (GLD) and supply of glycerol as sacrificial substrate. These interventions were implemented both separately and concurrently, and the results compared to biotransformations by low-GLD, non-permeabilised cells. The interactions between substrate transport and cofactor regeneration were found to be complex In cultures of intact or chemically permeabilised cells, strains with enhanced NADH regeneration capacity produced equal, and sometimes lower, titres than control cells carrying only the empty vector. The co-expression of GLD tended to reduce product titres in high-density cultures, by exacerbating instability of the whole-cell biocatalyst, which was already substantially reduced relative to low-density cultures. In low-density cultures, co-expression of GLD was not harmful, but it did not lead to any notable improvements in product formation rates. However, across numerous cultures, enhanced cofactor regeneration capacity correlated with enhanced CYP153A6 stability. Furthermore, in many low-density cultures, the co-expression of GLD provided marginal improvements in overall stability of the whole-cell biocatalyst when biotransformations were extended beyond 48 h. Thus, under the conditions tested, n-octane biotransformation rates were not limited by the availability of NADH. At the same time, the results suggest the possibility of harnessing enhanced cofactor regeneration to extend biotransformation times, presumably by improving the overall metabolic state of the whole cell – but also show that such a strategy has a ‘tipping point', related to cell density, beyond which the presence of additional dehydrogenase becomes harmful. The application of Triton X-100 or Polymyxin B substantially increased initial product formation rates relative to untreated cells. In low-density cultures, this led to chemically permeabilised cells achieving significantly higher final titres than untreated cells. However, the use of these chemicals in high-density cultures reduced the stability of the whole-cell biocatalyst; biotransformations levelled off rapidly, limiting the maximum titres that could be obtained. The co-expression of GLD was not beneficial in these cases. Conversely, the physical breakage of cell envelopes (via passage through a high-pressure homogeniser) improved product titres and space time yields relative to unpermeabilised cells, provided strains were co-expressing GLD. This clearly demonstrated the importance of balancing the supply of alkane and cofactor, with the expression of additional GLD compensating for reductions in metabolic capacity associated with the physically broken membranes. Mechanically permeabilised cells and cell free extract remained stable over extended periods of biotransformation, even though culture densities were high in these cases. Due to their capacity to maintain initial reaction rates at high cell density, mechanically permeabilised cells produced the highest final titres of any system tested in this study. If GLD was not co-expressed (or if CYP153A6 and GLD were expressed in separate cells), the biotransformation rates of mechanically permeabilised cells were substantially reduced. These results identified cross-membrane transport of hydrocarbons as the key limitation in the biotransformation of n-octane by whole-cell biocatalysts. Supply of substrate was a rate-limiting step in the reaction (hence the increase in product formation rates in low-density cultures of permeabilised cells). The degree of cell permeabilizationalso affected how the cell responded to substrate and product toxicity (hence the stability of high-density cultures of mechanically permeabilised cells). However, cofactor regeneration rates quickly became a limiting factor if the cofactor pool was disrupted (as was the case when the cell envelope was physically broken). To improve cross-membrane transport without destroying the cell membrane, E. coli strains were engineered to express AlkL, a passive membrane transporter protein that evolved to facilitate the growth of Pseudomonas putida on alkanes. The expression strategy included manipulating the supply of haem precursors δ-ALA and FeCl3, in a drive to achieve comparable expression of CYP153A6 in strains with and without AlkL. Even though the AlkL expression strategy was not fully optimised, cells co-expressing the transport protein demonstrated 50 – 100 % improvements on their yield on biomass after 48 h. These results confirmed the importance of overcoming the cross-membrane transport limitation. Overall, the findings highlight the importance of balancing reaction components. Cross-membrane transport of hydrocarbons emerged as the key limitation in biocatalysis by CYP153A6 in whole cells. This limitation could be alleviated by permeabilising the cell. Mechanical breakage of the cell envelope was particularly effective, but this in turn impacted negatively on the cell's ability to maintain its cofactor pool, necessitating the co-expression of glycerol dehydrogenase to boost NADH regeneration rates. In this regard, the membrane channel protein AlkL showed great promise, allowing cells to increase their rates of product formation even while relying on their native cofactor pool. These performance enhancements were achieved despite low intracellular concentrations of active CYP153A6 in cells co-expressing AlkL. Furthermore, in biotransformations with intact cells, intracellular concentration of CYP153A6 could be substantially reduced (via a reduction in the concentration of haem precursor), without any negative impact on biotransformation being observed. Increasing the cell density of biotransformation cultures allowed for rapid product accumulation, but substantially reduced the stability of the whole-cell biocatalyst. This instability correlated closely with the volumetric rate of product formation, suggesting that the effect was linked to reaction or product toxicity. Isolated improvements in enzyme expression, substrate supply, cofactor regeneration, or biomass loading did not translate into improved biocatalytic performance overall. This demonstrates that the creation of a useful P450-based biocatalyst will depend on the simultaneous optimisation of multiple components. In the field of biocatalysis, the focus tends to be on the characterisation and modification of the enzymes themselves. While this enzymatic understanding is crucial, the equally important task of understanding host physiology is often not explored in the same exacting detail. Unlocking the full potential of biocatalysis within the bulk chemical space requires an understanding of the cell architectures within which the enzymes of interest must function. This study has served to shed light on the links between cells, enzymes, and reactants – and on how much there still is to understand when it comes to whole cell performance.

Keywords

Chemical Engineering

Reference:

Collections

PhD / Doctoral

Full item page