Biological production of hydroxylated aromatics: Optimization strategies for Pseudomonas putida S12

To replace environmentally unfriendly petrochemical production processes, the demand for bio-based production of organic chemicals is increasing. This thesis focuses on the biological production of hydroxylated aromatics from renewable substrates by engineered P. putida S12 including several cases of strain improvement. Chapter 2 describes the construction of a P. putida S12 strain that produces p-hydroxybenzoate via the aromatic amino acid tyrosine. Previous research on biosynthesis of aromatic compounds has culminated in the construction of P. putida S12_427. This strain, which has an optimized carbon flux towards tyrosine, was employed as platform host for aromatics production. By introducing the heterologous gene pal/tal (encoding phenylalanine/tyrosine ammonia lyase; Pal/Tal) the conversion of tyrosine into p-coumarate was established, which compound is further converted into p-hydroxybenzoate by endogenous enzymes. The degradation of p-hydroxybenzoate was prevented by inactivating the gene pobA, which encodes hydroxybenzoate hydroxylase. The resulting strain, P. putida S12palB1, accumulated p-hydroxybenzoate at a yield of 11 Cmol % on glucose or glycerol in shake flask cultivations. A glycerol-limited fed-batch cultivation was performed to increase the product titers, yielding a final p-hydroxybenzoate concentration of 12.9 mM (1.8 g l-1) with a product-to-substrate yield of 8.5 Cmol %. Tyrosine availability was identified as the main bottleneck for p-hydroxybenzoate production in glycerol-limited chemostat cultivations. Since the enhanced flux towards p-hydroxybenzoate in P. putida S12palB1 mainly originated from a random mutagenesis approach, multiple system-wide changes were expected. Therefore, a chemostat-based comparative transcriptomics and proteomics analysis was performed as described in Chapter 3, to gain insight into the genetic background of the enhanced strain performance. The overall expression differences between parent strain and P. putida S12palB1 confirmed the system-wide changes effectuated by the strain improvement procedure. The higher net metabolic flux towards p-hydroxybenzoate was reflected in the upregulation of genes involved in tyrosine biosynthesis and the conversion of p-coumarate into p-hydroxybenzoate. Notably, on glucose some of the p-hydroxybenzoate biosynthetic genes were upregulated to a higher extent than in the glycerol-grown chemostats, while the phydroxybenzoate accumulation was not affected by the carbon source applied. Furthermore, a multidrug efflux transporter (PP1271-PP1273) was identified that may have a major role in phydroxybenzoate export. The 2.8-fold upregulation of hpd (encoding 4-hydroxyphenylpyruvate dioxygenase; first enzyme of tyrosine degradation via the homogentisate pathway) on glucose suggested that part of the tyrosine was directed away from p-hydroxybenzoate biosynthesis. This was confirmed by a 22-% increase of the product-to-substrate yield upon eliminating hpd. Although the hpd gene was not significantly differentially expressed on glycerol (1.5-fold), hpd deletion resulted in a 21 % improved p-hydroxybenzoate yield also on glycerol. The key precursors for the production of aromatic compounds, phosphoenolpyruvate (PEP) and erythrose-4-phosphate (E4P), are withdrawn from the lower glycolysis, respectively, the pentose phosphate (PP) pathway. Since the flux through the PP pathway is typically low in P. putida S12, the availability of E4P may be a bottleneck for p-hydroxybenzoate production. In Chapter 4, it was attempted to increase the availability of E4P by co-feeding xylose directly into the PP pathway. To this end, the genes xylABFGH (encoding xylose isomerase, xylulokinase and a high affinity xylose transporter) were introduced in the phydroxybenzoate producing strain to establish growth on xylose via the PP pathway. Growth performance on xylose was improved by an evolutionary selection approach. Surprisingly, this also resulted in a 25-% improvement of p-hydroxybenzoate production on either glycerol or glucose, indicating that the resulting strain, P. putida S12pal_xylB7, had an intrinsically elevated PP pathway activity. Chemostat experiments demonstrated that co-feeding of xylose (replacing part of the glucose or glycerol in the feed) considerably increased the production efficiency. On glucose, co-feeding of xylose improved the product-to-substrate yield by a factor 1.5 to 7.9 Cmol %, while on glycerol the product-to-substrate yield doubled to 16.3 Cmol %. Interestingly, product formation was not further improved by replacing more than 25 % of the carbon feed with xylose. This suggested that the availability of E4P was no longer the limiting step in p-hydroxybenzoate biosynthesis. The ability of co-utilizing xylose and glucose and the resultant improved production parameters implicated that lignocellulosic feedstock, containing around 20 % xylose and 50 % glucose, is a very suitable substrate for the production of aromatic compounds by engineered P. putida S12. In Chapter 5 the utilization of the industrial grade renewable feedstock crude glycerol was evaluated. In high-cell density fed-batch fermentations, P. putida S12 strains performed consistently better on crude glycerol than on purified glycerol, as shown by the higher biomass-to-substrate yield, maximum biomass production rate and substrate uptake rate. Moreover, the production of p-hydroxybenzoate by an engineered P. putida S12 strain was more efficient on crude glycerol. On crude glycerol a maximum p-hydroxybenzoate concentration of 43.5 mM (6.0 g l-1) was obtained with a product-to-substrate yield of 6.6 Cmol %, compared to 38.4 mM (5.3 g l-1) and 5.9 Cmol % on purified glycerol. In contrast, E. coli DH5? showed a decreased biomass-to-substrate yield and growth rate on crude glycerol compared to purified glycerol, and increased acetate formation (11.5 and 16.2 g l-1 on purified and crude glycerol respectively). The majority of the optimization approaches described in this thesis for p-hydroxybenzoate production can be applied for efficient production of other tyrosine derived compounds. This is exemplified in Chapter 6, where the production of the value-added compound phydroxystyrene from glucose is described. Production of p-hydroxystyrene was established by introducing the genes pal/tal and pdc, encoding phenylalanine/tyrosine ammonia lyase and p-coumaric acid decarboxylase respectively, into the tyrosine overproducing strain P. putida S12_427. These enzymes allow the conversion of the central metabolite tyrosine into p-hydroxystyrene, via p-coumarate. Degradation of the p-coumarate intermediate was prevented by inactivating the fcs gene encoding feruloyl-CoA synthetase. In fed-batch cultivation on glucose a maximum p-hydroxystyrene concentration of 4.5 mM was obtained with a yield of 6.7 Cmol %. At this p-hydroxystyrene concentration, growth and production were completely halted due to the toxicity of p-hydroxystyrene. Product toxicity was overcome by the application of a second phase of 1-decanol to extract p-hydroxystyrene during fedbatch cultivation. This approach resulted in a final p-hydroxystyrene concentration of 147 mM (17.6 g l-1) in the decanol phase, with a twofold increase of the maximum volumetric productivity (0.75 mM h-1) and a fourfold higher total concentration (21 mM). The work described in this thesis illustrates how hydroxylated aromatics can be efficiently produced by engineered P. putida S12 strains. Several strategies for analysis and optimization were performed to improve the aromatics production and understand the background and impact of the improvements. The biological production of various hydroxylated aromatics was established and largely improved. Hence, this work presents an important contribution to our efforts to replace petrochemical production by bio-based processes.