Substrate promiscuity of polyketide synthase enables production of tsetse fly attractants 3-ethylphenol and 3-propylphenol by engineering precursor supply in yeast

Tsetse flies are the transmitting vector of trypanosomes causing human sleeping sickness and animal trypanosomiasis in sub-saharan Africa. 3-alkylphenols are used as attractants in tsetse fly traps to reduce the spread of the disease. Here we present an inexpensive production method for 3-ethylphenol (3-EP) and 3-propylphenol (3-PP) by microbial fermentation of sugars. Heterologous expression in the yeast Saccharomyces cerevisiae of phosphopantetheinyltransferase-activated 6-methylsalicylic acid (6-MSA) synthase (MSAS) and 6-MSA decarboxylase converted acetyl-CoA as a priming unit via 6-MSA into 3-methylphenol (3-MP). We exploited the substrate promiscuity of MSAS to utilize propionyl-CoA and butyryl-CoA as alternative priming units and the substrate promiscuity of 6-MSA decarboxylase to produce 3-EP and 3-PP in yeast fermentations. Increasing the formation of propionyl-CoA by expression of a bacterial propionyl-CoA synthetase, feeding of propionate and blocking propionyl-CoA degradation led to the production of up to 12.5 mg/L 3-EP. Introduction of a heterologous ‘reverse ß-oxidation’ pathway provided enough butyryl-CoA for the production of 3-PP, reaching titers of up to 2.6 mg/L. As the concentrations of 3-alkylphenols are close to the range of the concentrations deployed in tsetse fly traps, the yeast broths might become promising and inexpensive sources for attractants, producible on site by rural communities in Africa.


Material and Methods
Strains and plasmids. Yeast strains and plasmids used in this study are described in Hitschler and Boles 14 or are listed in Table 1. S. cerevisiae was cultivated in YPD medium (20 g/L peptone, 10 g/L yeast extract, 20 g/L glucose) from freshly streaked YPD agar plate cultures. For fermentations the medium was supplemented with 100 mM potassium phosphate (KP i ) buffer (pH 6.5). Appropriate antibiotics (200 mg/L hygromycin or 200 mg/L G418) were added to media for plasmid maintenance. Escherichia coli DH10ß (Gibco BRL, Gaithersburg, MD) was grown in lysogeny broth (LB)-medium (10 g/L trypton, 5 g/L yeast extract, 5 g/L sodium chloride, pH 7.5) supplemented with appropriate antibiotics (100 mg/L carbenicillin, 50 mg/L kanamycin or 25 mg/L chloramphenicol) for plasmid maintenance and cloning.
Plasmid and strain construction. The codon-optimized DNA sequences, opt patG (GeneBank accession number MK791645), Ppopt MSAS (MK791642), opt npgA (MK791644) and opt prpE (MT219994), were obtained with the JCat tool 29 and ordered as GeneArt Strings DNA fragments from Thermo Fischer Scientific. Genomic DNA of CEN.PK2-1C or plasmids were used as templates for PCR amplification of yeast open reading frames, promoters and terminators with 35 bp homologous overlaps. Primers and genes used in this study are described in Hitschler and Boles 14 , Schadeweg and Boles 21,22 or are listed in Supplementary Tables S1 and S2.
Plasmid assembly in yeast via homologous recombination or in E. coli via Gibson assembly 30 and plasmid propagation were conducted as described previously 14 . Genomic integrations into the ura3 locus of CEN.PK2-1C were performed with the CRISPR/Cas9 system 25 as described in Hitschler and Boles 14 . For deletions, CRISPR/ Cas9 plasmids carrying the guide RNA (gRNA) for the specific deletion were amplified via PCR and assembled via Gibson. The donor DNA consisting of 40 bp upstream and 40 bp downstream sequences of the open reading frame were ordered as primers and were annealed to double-stranded DNA as described in Reifenrath and Boles 31 . However, for deletion of ACS2 in the JHY185 strain a pUG6-based deletion cassette 23 was amplified, conferring resistance to hygromycin and carrying 40 bp overhangs to the ACS2 locus for integration via homologous recombination.
For genomic integration of the 'reverse ß-oxidation' pathway genes (Schadeweg and Boles) the Golden Gate system 27 was utilized for construction of an integration vector. Part plasmids were obtained from Lee et al. 27 or PCR fragments with part type specific overhangs and pYTK01 as backbone were assembled with Esp3I as described previously 27 incubating the reaction mixture for 10 min at 37 °C, using 15 cycles of digestion and ligation (37 °C 2 min, 16 °C 5 min) and heat inactivating the enzymes at 60 °C for 10 min and 80 °C for 10 min and transformed into E. coli. Next cassette plasmids were formed carrying BsmBI overhangs flanking the cassette for subsequent assembly of the integration plasmid. To build the cassette plasmid pJHV62, part plasmids pYTK3.41, pYTK05, pYTK09, pYTK053, pYTK72 and pYTK95 were assembled with BsaI-HF. For assembly of the integration vector pJHV65 with Esp3I, the reaction mixture containing the cassette plasmids pYTK_ ERG10, pYTK_hbd, pYTK_crt, pJHV62 and SiHV110 were incubated for 10 min at 37 °C, using 25 cycles of digestion and ligation (37 °C 1.5 min, 16 °C 3 min), 37 °C for 5 min and heat inactivating the enzymes at 50 °C for 5 min and 80 °C for10 min and transformed into E. coli. The integration vector pJHV65 was digested with NotI and 500 bp homologous sequences to the upstream and downstream region of LEU2 flanked the integration cassette and natMX cassette for homologous recombination and selection in yeast. After transformation of yeast with respective DNA fragments according to Gietz and Schiestl 32 , cells were grown on selective YPD agar plates.
Cell cultivation. Cells were cultivated in 150 mL YPD medium supplemented with corresponding antibiotics and buffered with 100 m potassium phosphate buffer (KP i ) at pH 6.5 to avoid unwanted effects of weak acids. Overnight cultures were harvested in exponential phase and utilized for inoculation of 25 mL KP i buffered YPD medium (pH 6.5) to an optical density (OD 600 nm ) of 4 or more. For consumption or biotransformation experiments, 10 mM butyrate or propionate were added, respectively. Cultures were shaken at 180 rpm at 30 °C for 144 h in a waterbath (Memmert, Germany) or in a 30 °C container to prevent inhalation of 3-alkylphenols.
Growth and metabolite analysis. The spectrophotometer Ultrospec 2100 pro (GE Healthcare, USA) was utilized to follow cell growth at an optical density of 600 nm. Culture supernatants for HPLC analysis of 3-alkylphenol formation were prepared as described previously 14 and analysis was performed via HPLC (Dionex) with an Agilent Zorbax SB-C8 column (4.6 ×150 mm, 3.5 µm) at 40 °C and at a flow rate of 1 mL/min. 3-methylphenol and 3-ethylphenol were separated by the same gradient of solvent A (0.1% (v/v) formic acid in ddH 2 O) and solvent B (0.1% (v/v) formic acid in acetonitrile) mentioned before 14 . The same gradient applied for 3-propylphenol analysis with the exception that the gradient stayed at 40% B for 5 min before it switched to 100% B to prolong the separation before the washing step. The 3-alkylphenols were detected at 270 nm in an UV detector (Dionex UltiMate 3000 Variable Wavelength Detector). For quantification and calibration, 3-alkylphenol standards were prepared in ddH 2 O from m-cresol purchased from Carl Roth (9269.1), 3-ethylphenol from Sigma-Aldrich (210-627-3) and 3-propylphenol from Alfa Aesar (621-27-2).
For propionate analysis 50 µL 50% (w/v) sulfosalicylic acid was added to 450 µL culture supernatant. Samples were analysed in the HPLC equipped with the ion exchange column HyperREZ XP Carbohydrate H + (7.7 × 300 mm, 8 µm) and a refractive index detector (Thermo Shodex RI-101). The metabolites were separated with 5 mM sulfuric acid as liquid phase at a flow rate of 0.6 mL/min and 65 °C. For quantification, propionate standards of different concentrations were prepared in ddH 2 O from propionic acid purchased from Carl Roth (6026.2). Data analysis and graphing were performed utilizing the software Prism 5 (Graphpad).

Results and Discussion
Production of 3-ethylphenol from propionyl-CoA. Propionate supplementation enables 3-ethylphenol formation. We aimed to synthesize 3-ethylphenol (3-EP) in vivo from glucose via MSAS and MSA decarboxylase by provision of propionyl-CoA as a priming unit for MSAS. This approach relies on high intracellular levels of propionyl-CoA to successfully compete with acetyl-CoA as the cognate priming unit of MSAS. In principle, enhanced propionyl-CoA concentrations can be achieved by either manipulating endogenous pathways leading to propionyl-CoA, feeding of propionate to the cells, blocking of propionyl-CoA degradation and/or enhancing direct propionyl-CoA synthesis. In yeast cells, endogenous pathways for generation of propionyl-CoA exist. For example, propionyl-CoA is an intermediate in yeast threonine catabolism which takes place in the mitochondria 18,19 . Threonine degradation to propionyl-CoA is initiated by threonine deaminase, catalyzing the conversion of threonine to 2-ketobutyrate. The 2-ketoacid dehydrogenase complex can then catalyze the oxidative decarboxylation of 2-ketobutyrate to propionyl-CoA. In addition, the acetyl-CoA synthases of S. cerevisiae are able to convert externally supplied propionate to propionyl-CoA 33 .
We first wanted to test whether it is possible to rely on endogenous pathways to provide enough propionyl-CoA for 3-EP production. For this, we utilized the 3-methylphenol (3-MP) production strain JHY162 from our previous work 14 . Strain JHY162 expresses Ppopt MSAS, opt npgA and opt patG under control of the strong constitutive pPGK1, pHXT7 -1-392 and pFBA1 promoters, respectively, which were stably integrated in the ura3 locus of the S. cerevisiae strain CEN.PK2-1C 14 . A high-OD fermentation (starting OD = 5) in KP i buffered YPD at pH 6.5 revealed that strain JHY162 only produced 3-MP (296 mg/L) but 3-EP could not be detected in the supernatants of the cultures (Fig. 2). This indicated that propionyl-CoA, either generated by threonine catabolism 18,19 or other endogenous pathways, is not available at sufficient concentrations to outcompete acetyl-CoA for conversion by MSAS. Moreover, endogenous propionyl-CoA might be directly degraded in the 2-methylcitrate cycle by 2-methylcitrate synthases Cit2 and Cit3 18,19,34 .
Nevertheless, when adding 10 mM propionate to the medium, the same strain (JHY162) produced up to 4.4 mg/L 3-EP ( Fig. 2A), indicating synthesis of propionyl-CoA and its conversion by MSAS and PatG to 3-EP. We assume that propionate is converted to propionyl-CoA by the endogenous acetyl-CoA synthetases of S. cerevisiae 33 . In spite of this success, the main product was still 3-MP which accumulated up to 372 mg/L (Fig. 2B) reflecting the preference of MSAS for acetyl-CoA as priming unit and of PatG for MSA [15][16][17] . Surprisingly, also 3-MP formation was stimulated by the addition of propionate.
To increase propionyl-CoA levels as the priming unit for 3-EP formation, we aimed at blocking propionyl-CoA degradation which is mediated by 2-methylcitrate synthases. It was shown that abolishment of 2-methylcitrate synthase activity in a Δcit2 Δcit3 deletion strain prevented propionate degradation 34 . To confirm this in our strains, we deleted CIT2 or CIT3 individually or both together in CEN.PK2-1C and performed fermentations with a starting OD of 4 in KP i buffered YPD medium at pH 6.5 supplemented with about 10 mM propionate. Indeed, the Δcit2 Δcit3 strain JHY174 did not consume any propionate over 144 hours. The single knock-out strains revealed that even a cit3 deletion alone (JHY175) is enough to abolish propionate degradation. Externally added propionate was completely consumed by the CEN.PK2-1C wildtype strain and the Δcit2 strain JHY164 (Fig. 2C).
Although deletion of only CIT3 already abolished propionate degradation, we used the Δcit2 Δcit3 double deletion strain to test an influence on 3-EP formation. We expressed Ppopt MSAS, opt npgA and opt patG together under control of the strong constitutive pPGK1, pHXT7 -1-392 and pFBA1 promoters, respectively, integrated into the ura3 locus of strain JHY174 (resulting in strain JHY197), and performed high-OD fermentations (starting OD = 5.0) in KP i buffered YPD at pH 6.5. Strain JHY197 showed a slightly better growth (not shown) and 3-MP production (296 mg/L) compared to JHY162 (247 mg/L) (Fig. 2B) was not sufficient to provide enhanced endogenous propionyl-CoA levels for 3-EP formation ( Fig. 2A). Even when supplemented with 10 mM propionate in the medium, blocked propionyl-CoA degradation did not further increase 3-EP formation (up to 3.6 mg/L in JHY197) ( Fig. 2A). Interestingly, also in the cit2/3 deletion strain 3-MP formation was somehow stimulated by the addition of propionate. The results indicate that degradation of propionyl-CoA is not limiting 3-EP formation in the yeast cells.
A heterologous propionyl-CoA synthetase increased 3-EP formation. In order to provide additional propionyl-CoA, we expressed the codon-optimized propionyl-CoA synthetase gene opt prpE from Salmonella typhimurium 35 under control of the strong promoter pTDH3, integrated into the sfa1 locus in both strains JHY162 and JHY197 (resulting in strains JHY218 and JHY185, respectively). Fermentations were performed as described above. 3-MP production with JHY218 and JHY185 was not significantly influenced compared to the strains without opt PrpE (Fig. 3B). When the medium was supplemented with 10 mM propionate, 3-EP production noticeably increased to titers of up to 12.5 mg/L with strain JHY218 and 11.6 mg/L with strain JHY185 (Fig. 3A). This result demonstrates that normally endogenous yeast propionyl-CoA synthetase activity is limiting 3-EP formation. Moreover, as both strains -with or without the 2-methylcitrate cycle -produced comparable amounts of 3-EP it confirms that degradation of propionyl-CoA is not limiting 3-EP production. Interestingly, expression of opt prpE led to the formation of 3-EP even without externally added propionate with both strains (Fig. 3A) (3.1 mg/L with JHY185 and 2.9 mg/L with JHY218 after 144 hours). This indicates an endogenous unknown source of propionate. We speculate that this propionate derives from endogenous propionyl-CoA which is hydrolyzed (unspecifically) by thioesterases. Obviously, hydrolyzation of propionyl-CoA  14 ) and with the Δcit2Δcit3 double deletion (JHY197) were utilized for high-OD fermentations (starting OD = 5.0) at 30 °C in KP i buffered YPD medium (pH 6.5) with or without supplementation of 10 mM propionate. Propionate consumption was followed in S. cerevisiae wild-type strain CEN.PK2-1C and deletion strains that either had peroxisomal (Δcit2), mitochondrial (Δcit3) or both methylcitrate synthases (Δcit2/Δcit3) deleted and were cultured (starting OD = 4) at 30 °C in KP i buffered YPD medium (pH 6.5) supplemented with 10 mM propionate. Culture supernatants were analysed via HPLC for 3-alkylphenol production and propionate. Error bars represent standard deviations of biological duplicates.

Production of 3-propylphenol from butyryl-CoA. Butyrate feeding is not sufficient for 3-propylphenol
formation. 3-Propylphenol (3-PP) is another promising attractant for tsetse flies 5 . In principle, it can be formed from butyryl-CoA as a priming unit of MSAS followed by decarboxylation of the formed 6-propylsalicylic acid  14 ) and additionally the propionyl-CoA synthase opt prpE, with or without the Δcit2Δcit3 double deletion (strains JHY185 and JHY218, respectively), were inoculated at an OD of 5 and cultivated for 144 h in KP i buffered YPD medium (pH 6.5) with or without supplementation of 10 mM propionate. Culture supernatants were analysed via HPLC for 3-alkylphenol production. Error bars represent standard deviations of biological duplicates.
(PSA). It has already been shown in vitro that MSAS can use butyryl-CoA as a substrate although conversion to PSA proceeds with 9% of the activity with acetyl-CoA 15,16 . The MSA decarboxylase PatG has not yet been tested for its activity with PSA.
In order to enable 3-PP production in S. cerevisiae, the priming unit butyryl-CoA is required in sufficient amounts. As shown above the endogenous acyl-CoA synthetases are able to convert propionate to propionyl-CoA. Moreover, Luo et al. 37 recently even demonstrated the conversion of exogenously supplied hexanoic acid to hexanoyl-CoA for the production of olivetolic acid. Therefore, we tested whether the supplementation of 10 mM butyrate in KPi buffered YPD medium can provide enough butyryl-CoA for 3-PP formation via MSAS and MSA decarboxylase. However, high-OD fermentations of strain JHY162 expressing Ppopt MSAS, opt npgA and opt patG under control of the strong constitutive pPGK1, pHXT7 -1-392 and pFBA1 promoters, respectively, did not result in any 3-PP formation (Supplementary Fig. S3A). Butyrate supplementation did not influence 3-MP production ( Supplementary Fig. S3B).
To exclude the possibility that the formed butyryl-CoA is rapidly degraded via ß-oxidation, we deleted the fatty acyl-CoA oxidase encoding gene POX1 in CEN.PK2-1C to abolish ß-oxidation 20 , and integrated the heterologous 3-MP production pathway as above, resulting in strain JHY211. For improved uptake of butyrate we transformed JHY211 with multi-copy plasmids overexpressing the endogenous monocarboxylic acid transporters Jen1 or Ady2 [38][39][40] under control of the strong pHXT7 -1-392 promoter or an empty vector as control. However, high-OD fermentations in KPi buffered YPD medium supplemented with 10 mM butyrate did not result in any 3-PP formation in any of the strains (Supplementary Fig. S3C). These results indicate that either butyryl-CoA concentrations cannot be enhanced to levels at which it can compete with acetyl-CoA for priming MSAS or that MSAS and/or MSA decarboxylase are not able to convert butyryl-CoA and/or PSA, respectively, in vivo.
Heterologous 'reverse ß-oxidation' pathway supplies enough butyryl-CoA as precursor for 3-propylphenol production. Recently, our group established a 'reverse ß-oxidation' pathway for efficient n-butanol production from acetyl-CoA in yeast 21,22 . This pathway forms butyryl-CoA as an intermediate. In the pathway, two acetyl-CoA are condensed by endogenous thiolase Erg10 to acetoacetyl-CoA, which is then converted to 3-hydroxybutyryl-CoA and crotonyl-CoA by hydroxybutyryl-CoA dehydrogenase Hbd and crotonase Crt from Clostridium acetobutylicum and finally reduced to butyryl-CoA by trans-2-enoyl-CoA reductase Ter from Treponema denticola. To use the pathway for 3-PP production, we integrated the genes required up to butyryl-CoA formation (ERG10 from S. cerevisiae, codon-optimized opt hbd and opt crt from C. acetobutylicum and opt ter from T. denticola) under control of the strong constitutive promoters pPGK1, pCCW12, pENO2, pTDH3, respectively, into the leu2 locus of yeast strains JHY162 and JHY211, resulting in strains JHY194 and JHY212, respectively.
High-OD fermentations of JHY194 and JHY212 and their parent strains as controls were performed in KP i buffered YPD (pH 6.5) at 30 °C. As observed before, strain JHY162, with the 3-methylphenol production pathway, and JHY211, with additional pox1 deletion, were unable to produce 3-PP (Fig. 4A). However, expression of the 'reverse ß-oxidation' pathway resulted in up to 2 mg/L 3-PP formation with strain JHY194 (Fig. 4A). The highest 3-PP titer (2.6 mg/L) was achieved when POX1 was additionally deleted (strain JHY212), indicating that butyryl-CoA degradation was partially limiting 3-PP production. As expected, 3-methylphenol was still produced in high amounts by all strains (Fig. 4B).
The 'reverse ß-oxidation'-based n-butanol production was limited by the trans-2-enoyl-reductase Ter, and additional Tdopt ter overexpression improved the final butanol titers 21 . To test whether Ter might also limit 3-PP formation, Tdopt ter was additionally overexpressed from a multi-copy plasmid under control of the strong constitutive pHXT7 -1-392 promoter in JHY194 and JHY212 and high-OD fermentations (starting OD = 5) were performed. However, additional Tdopt ter overexpression did not increase 3-PP titers (1.1 mg/L and 0.7 mg/L with or without Ter overexpression in JHY194 and JHY212, respectively) indicating that other factors limit 3-PP production.
To conclude, obviously the 'reverse ß-oxidation' pathway provides more butyryl-CoA than exogenous addition of butyrate, and these levels are high enough to compete at least partially with acetyl-CoA and to be transformed into 3-PP.

conclusions
In this work we show that yeast engineered to provide increased intracellular formation of propionyl-CoA or butyryl-CoA and expressing MSAS and MSA decarboxylase can be exploited to produce 3-EP and 3-PP from sugars. The approach is based on the broad substrate tolerance of MSAS and MSA decarboxylase, shown before in vitro. In spite of this success, 3-MP derived from acetyl-CoA as the preferred priming substrate of MSAS and decarboxylation of the intermediate 6-MSA by MSA decarboxylase remained the main product of the engineered strains. Acetyl-CoA cannot be eliminated as it is an essential metabolite and is required for the production of malonyl-CoA as the elongation substrate of MSAS. Therefore, further approaches to increase production of 3-EP and 3-PP will require the engineering of the substrate specificity of MSAS. Since all the enzymatic domains of MSAS can essentially account for reduced rates in the turnover of substrates with elongated alkyl moiety, an elaborate engineering strategy is necessary including mutations of binding sites and swaps of catalytic domains 41 . A corresponding engineering of the MSA decarboxylase will be necessary.
Concerning the use of the 3-alkylphenols (3-MP, 3-EP, 3-PP) as baits in tsetse fly traps the titers achieved in our work are close to the natural concentrations in cattle urine (50 mg/L 3-MP, 5.5 mg/L 3-EP and 12.5 mg/L 3-PP) and the concentrations deployed in tsetse fly traps 1,5 . However, as higher concentrations improved catch rates 5 and 3-EP and 3-PP are more effective than 3-MP, MSAS and MSA decarboxylase engineering might be useful to improve the effects. It remains to be tested whether it will be possible to simply use the whole yeast cultures, yeast extracts or supernatants to prepare the traps. Our work is a first step in facilitating the preparation of the traps by the simple and direct "brewing" of 3-alkylphenols. As they shall be produced locally by poor rural  14 ) with or without additional expression of the 'reverse ß-oxidation' pathway (ERG10, opt hbd, opt crt and opt ter 21,22 ) (strains 194 and 162, respectively), and with additional pox1 deletion (strains JHY212 and JHY211, respectively). High-OD fermentations (starting OD = 5) were performed in biological duplicates at 30 °C in KP i buffered YPD medium at pH 6.5. Culture supernatants were analysed via HPLC for 3-alkylphenol production. Error bars represent standard deviations.