Production of moth sex pheromones for pest control by yeast fermentation

The use of insect sex pheromones is an alternative technology for pest control in agriculture and forestry, which, in contrast to insecticides, does not have adverse effects on human health or environment and is efficient also against insecticide-resistant insect populations.1,2 Due to the high cost of chemically synthesized pheromones, mating disruption applications are currently primarily targeting higher value crops, such as fruits.3 Here we demonstrate a biotechnological method for the production of pheromones of economically important moth pests using engineered yeast cell factories. Biosynthetic pathways towards several pheromones or their precursors were reconstructed in the oleaginous yeast Yarrowia lipolytica, which was further metabolically engineered for improved pheromone biosynthesis by decreasing fatty alcohol degradation and downregulating storage lipid accumulation. The sex pheromone of the cotton bollworm Helicoverpa armigera was produced by oxidation of fermented fatty alcohols into corresponding aldehydes. The resulting pheromone was just as efficient and specific for trapping of H. armigera male moths in cotton fields in Greece as a synthetic pheromone mixture. We further demonstrated the production of the main pheromone component of the fall armyworm Spodoptera frugiperda. Our work describes a biotech platform for the production of commercially relevant titres of moth pheromones for pest control by yeast fermentation. Significance statement Agriculture largely relies on insecticides and genetically modified crops for pest control, however alternative solutions are required due to emerging resistance, toxicity and regulatory issues, and consumer preferences. Mating disruption with sex pheromones that act by preventing insect reproduction is considered the most promising and scalable alternative to insecticides. This method is highly efficient and safe for human health and environment. The likelihood of insect resistance development is very low and can be handled by adjusting the pheromone composition. The high cost of chemically synthesized pheromones is the major barrier for the wider adoption of pheromones. A novel method based on yeast fermentation enables the production of insect sex pheromones as a lower cost from renewable feedstocks.

2 reconstructed in the oleaginous yeast Yarrowia lipolytica, which was further metabolically engineered for improved pheromone biosynthesis by decreasing fatty alcohol degradation and downregulating storage lipid accumulation. The sex pheromone of the cotton bollworm Helicoverpa armigera was produced by oxidation of fermented fatty alcohols into corresponding aldehydes. The resulting pheromone was just as efficient and specific for trapping of H. armigera male moths in cotton fields in Greece as a synthetic pheromone mixture. We further demonstrated the production of the main pheromone component of the fall armyworm Spodoptera frugiperda.
Our work describes a biotech platform for the production of commercially relevant titres of moth pheromones for pest control by yeast fermentation.

Significance statement
Agriculture largely relies on insecticides and genetically modified crops for pest control, however alternative solutions are required due to emerging resistance, toxicity and regulatory issues, and consumer preferences. Mating disruption with sex pheromones that act by preventing insect reproduction is considered the most promising and scalable alternative to insecticides. This method is highly efficient and safe for human health and environment. The likelihood of insect resistance development is very low and can be handled by adjusting the pheromone composition. The high cost of chemically synthesized pheromones is the major barrier for the wider adoption of pheromones. A novel method based on yeast fermentation enables the production of insect sex pheromones as a lower cost from renewable feedstocks.

Establishing pathways towards moth pheromones in yeast
The majority of identified moth sex pheromone components are unsaturated acetates, alcohols or aldehydes. These compounds are derived from the fatty acid metabolism and possess 10-18 carbon 3 atoms in the carbon skeleton. They are called Type I moth pheromone components and constitute approximately 75% of all known moth sex pheromone components. 4,5 The chemical diversity is to a large extent produced by the combined action of specific desaturases and limited chainshortening 6 . To establish pathways towards moth pheromone compounds in yeast, we first investigated a range of fatty acyl-CoA desaturases and reductases for the production of (Z)hexadec-11-en-1-ol (Z11-16OH) (  (Fig. 1B), which was an order of magnitude enhancement in comparison to the previous study. 8 The improvement was likely due to the utilization of a desaturase variant with a higher activity in yeast and due to expression of the genes from constitutive promoters using constructs stably integrated into the yeast genome. 9 Next, we wanted to achieve the biosynthesis of (Z)-tetradec-9-en-1-yl acetate (Z9-14Ac), which is the main sex pheromone component of the fall armyworm Spodoptera frugiperda, a rising pest with a high occurrence of insecticide resistance. 10,11 For this, we searched for a Δ9-desaturase with a higher activity and specificity towards tetradecanoyl-CoA than to hexadecanoyl-CoA (Fig. 1C). The activities of six heterologous desaturase candidates were tested in a S. cerevisiae ole1Δelo1Δ strain devoid of the native desaturation and elongation activities. The cells were cultivated with supplementation of methyl tetradecanoate (14Me) and the total lipids were analysed to determine the desaturated fatty acids (Fig. 1D). The strain expressing the desaturase from Drosophila melanogaster resulted in the highest concentration of 3.67±0.99 mg/L methyl (Z)-tetradec-9enoate (Z9-14Me) and in the highest Z9-14Me to methyl (Z)-hexadec-9-enoate (Z9-16Me) ratio, indicating a higher specificity towards the tetradecanoyl-CoA substrate. To establish the complete pathway towards Z9-14Ac in the yeast S. cerevisiae, we expressed the D. melanogaster Δ9 desaturase together with the H. armigera reductase and S. cerevisiae ATF1 known to catalyse acetylation of fatty alcohols 12 . The resulting strain produced 7.3±0.2 mg/L of Z9-14Ac in comparison to 1.4±0.4 mg/L in an analogous strain lacking the heterologous Δ9 desaturase (Fig.   1E).

Optimization of the oleaginous yeast Yarrowia lipolytica for moth pheromone production
We rationalized that an oleaginous yeast should be a more suitable cell factory for production of fatty alcohol-based moth pheromones than the baker's yeast that has a low content of the fatty acid precursor acetyl-CoA in the cytosol and can only accumulate small amounts of intracellular lipids.
In contrast, the oleaginous yeast Yarrowia lipolytica has a naturally high fatty acid metabolism and has been engineered for commercial production of polyunsaturated omega-3 fatty acids 13 and for production of lipids 14 . Robust genetic tools, including the CRISPR/Cas9 method, have recently been developed for Y. lipolytica, and allow the rapid engineering of this yeast species. 15,16 The first hurdle we encountered when co-opting Y. lipolytica for the production of pheromones, was the prevention of endogenous degradation of the target fatty alcohols Z11-16OH and (Z)tetradec-9-en-1-ol (Z9-14OH). We deleted one by one and in combination the genes encoding the enzymes potentially implicated in fatty alcohol degradation: fatty aldehyde dehydrogenases Hfd1p and Hfd4p 17 , as well as fatty alcohol oxidase Fao1p 18 ( Fig. 2A). Moreover, we deleted peroxisomal biogenesis factor Pex10p, thus interrupting the correct assembly of peroxisomes and preventing acyl-CoA degradation. Single deletions of HFD1/HFD4/FAO1/PEX10 genes increased the titre of 5 Z11-16OH two-to three-fold, while the combination of four deletions resulted in a 19-fold titre increase (Fig. 2B). The quadruple deletion strain (ST5789) produced 14.9±3.6 mg/L of Z11-16OH in comparison to 0.8±0.1 mg/L produced by a reference strain only expressing the biosynthetic pathway towards Z11-16OH (ST3844). When strains ST3844 and ST5789 were incubated with externally supplied 1 g/L Z11-16OH and 1 g/L Z9-14OH each, strain ST3844 largely degraded the supplied alcohols, with only 6.2±3.8 mg/L Z11-16OH left at the end of the cultivation. In contrast to that, strain ST5789 showed a remaining concentration of 630.9±137.1 mg/L Z9-14OH and 620.3±73.9 mg/L Z11-16OH. Less than 1 g/L of fatty alcohols were recovered probably due to evaporation and some losses during the recovery procedure. A control, which contained only cultivation medium and externally supplied fatty alcohols, showed a remaining concentration of 500.7±135.7 mg/L Z9-14OH and 536.9±166.1 mg/L Z11-16OH (Fig. S1). The experiment confirmed that the degradation rate of fatty alcohols was much lower in the strain with deletion of HFD1, HFD4, PEX10, and FAO1 genes.
Another challenge with Y. lipolytica as a host was to reduce the channelling of fatty acyl-CoAs, the fatty alcohol precursors, into storage lipids. We hence downregulated the expression of the gene encoding glycerol-3-phosphate acyltransferase (GPAT), which catalyses the first committing step towards glycerolipid-and glycerophospholipid biosynthesis. The downregulation was achieved by truncating the GPAT promoter to 100 base pairs and confirmed by qRT-PCR (Fig.   S2.A). The downregulation of GPAT improved the titre of Z11-16OH from 14.9±3.6 mg/L to 20.6±5.4 mg/L (Fig. 2B). At the same time, the total fatty acid content of the cells was reduced by 53% ( Fig. S2.B, C). The combination of Y. lipolytica genome edits that reduce the fatty alcohol degradation and lipid accumulation thus resulted in a basic platform chassis, where various moth pheromone pathways can be inserted. 6 The strain, however, predominantly produced fatty alcohols of 16-carbon chain length. In order to enable the production of 14-carbon pheromones, we introduced a mutation into fatty acid synthase subunit Fas2p I1220F , which was previously reported to benefit the biosynthesis of tetradecanoyl-CoA. 19 We expressed the pathway towards Z9-14OH in the engineered Y. lipolytica strains and obtained 4.9±1.4 mg/L titre in the basic platform chassis and 73.6±16 mg/L Z9-14OH in the platform chassis with additional Fas2p mutation (Fig. 3B). The mutation thus resulted in a 15-fold improvement of a 14-carbon product and should be beneficial for producing also other pheromones derived from tetradecanoyl-CoA.
To further improve the production of Z11-16OH, we integrated additional copies of desaturase and reductase genes to pull the flux towards pheromone biosynthesis. Integration of the second copy of the pathway increased the titre 4.6-fold. The strain with three copies of the pathway produced 169±14 mg/L Z11-16OH, a 9.7-fold increase in comparison to the single copy strain (Fig 3A).
When the optimized yeast strain was fermented in a 10L-bioreactor, 2.57 g/L of the target product Z11-16OH was obtained. The fatty alcohols were extracted from the yeast biomass using organic solvents and purified on a silica column. The eluted fractions with a high content of the product were pooled and oxidized into corresponding aldehyde using tetrakisacetonitrile copper(I) triflate/TEMPO catalyst system. 20 The composition of the aldehyde preparation was Z11-16Ald, hexadecanal (16Ald), and (Z)-hexadec-9-enal (Z9-16Ald) in ratio 82:13:5 (Fig. S3, S4). The Z11-16Ald is the major and Z9-16Ald is the minor pheromone component in H. armigera and C. suppressalis, where the reported ratios between the two pheromone components in H. armigera were from 99:1 to 93:7 21-23 , in C. suppressalis the reported ratio is 90:10 24 . 16Ald is also present in the pheromone glands of both insect species, but it does not elicit a behavioural response. The biologically produced pheromone composition may thus be close enough and well suited for trapping and mating disruption of these insect species with Z11-16Ald as a major and Z9-16Ald as a minor pheromone component. The biologically produced pheromone mix was subsequently subjected to activity tests on H. armigera in the laboratory and field.

Electrophysiological responses of male H. armigera
We measured the electroantennographic responses of male H. armigera adults to the yeastproduced pheromone blend (Bio-Ald), standard compounds, and mixtures of the standards (Fig.   4). Ald mix #1 contained Z11-16Ald, Z9-16Ald, tetradecanal (14Ald), and pentadecanal (15Ald) (80:5:5:5, respectively). Ald mix #2 contained equal volumes of each of the same components as Ald mix #1 (25:25:25:25 ratio). Bio-Ald elicited the same magnitude of response on the male antenna as Ald mix #1 and significantly higher to that of the equivolume Ald mix #2 and to Z9-16Ald, the secondary compound of the H. armigera pheromone. The major sex pheromone compound, Z11-16Ald, yielded a high antennal response, whereas the minor sex pheromone, Z9-16Ald, induced a considerably lower response. The significantly lower response to Ald mix #2 is a clear indication that the antennal response is mainly attributed to Z11-16Ald and when its quantity in the mixture is lowered, the antennal response also drops. These results indicate that biologically produced Z11-16Ald induces the same magnitude of sensory stimulation as the chemically synthesized Z11-16Ald, the major compound of the moth's native pheromone.

Monitoring of H. armigera flight in the field
Mean weekly male catches in traps baited with yeast-produced pheromone (Bio-Ald) and synthetic pheromone (Z11-16Ald : Z9-16Ald at 97:3 ratio, Control) dispensers are shown in Fig. 5

Conclusions
In summary, we have demonstrated biological production of practically and commercially relevant titres of several lepidopteran sex pheromones in yeast cell factories. A biocatalytic production is particularly advantageous for the production of chemicals for which isomeric composition is critical, such as moth pheromones. 5 The enzymes can deliver the required stereoisomers, while in chemical synthesis, a mix of isomers is often obtained and may be difficult to separate especially in large quantities. Furthermore, the fermentation is carried out in a cheap medium with glycerol as the sole carbon source, using yeast cells as the only catalyst. This is in contrast to chemical synthesis that will typically require special starting material, expensive catalysts, and several synthesis steps. Reduced production costs and lower environmental impact of the biotech route in comparison to the chemical synthesis have been established for multiple chemicals, particularly for natural products. 25,26 As an additional advantage, major and minor pheromone components can be produced in a single process in a ratio that is suitable for the target insect. The work creates the foundation for the production of pheromones at a lower cost enabling pheromone-based pest control in row crops, such as rice, cotton, and maize.