Seeds of the tomato (S. lycopersicum) trichome mutant odorless-2 (Kang et al., 2010a) were kindly provided by Dr. Gregg Howe (Michigan State University, MI, United States). Tomato plants used for Agrobacterium leaf infiltration and all subsequent analyses were grown from seeds in Sungro^® soil mixture (Sun Gro Horticulture, Agawam, MA, United States) in multi-trays (288 cells, 5 ml per cell), and seedlings were transplanted into 4-inch square pots. Plants were grown under a 16-h photoperiod in a climate-controlled growth room (23–25°C, 50–60% relative humidity) without pesticide application.

For the cloning of the two multicistronic gene constructs under the control of a tissue-specific promoter, the binary vector pMCS:GW was used (obtained from the Arabidopsis Biological Resource Center, stock # CD3-1933) that contains a multiple cloning site upstream of the Gateway cassette (Michniewicz et al., 2015). To obtain the promoter region of the AtCER5 gene (At1g51500; Pighin et al., 2004), a 2,614 base pair fragment was amplified by PCR from Arabidopsis thaliana genomic DNA using Taq polymerase (GenScript, Piscataway, NJ, United States) and a pair of oligonucleotides each carrying a specific restriction site (AtCER5-fwd-EcoRI: 5'-CGGAATTCTTTAGTTTGCTTG AGTTCTCATG-3'; AtCER5-rev-SalI: 5'-GCGTCGACTGTTTT TGTTTGATCTTGAAAAAGATC-3'). The resulting PCR product was cloned into the vector pMD20 (Takara Bio, United States) and sequenced to verify its correct amplification. The AtCER5 promoter fragment was subsequently excised from the pMD20 vector by EcoRI and SalI digestion, and then ligated between the EcoRI and XhoI sites of the pMCS:GW vector.

The two multicistronic gene constructs were obtained by gene synthesis (Twist Bioscience, San Francisco, CA, United States). These gene constructs were flanked by attL-sequences and inserted into the Gateway Entry vector pTwist ENTR (Twist Bioscience, San Francisco, CA, United States). Both of these multicistronic constructs contained three consecutive coding regions encoding a prenyltransferase, a terpene synthase, and enhanced green fluorescent protein, respectively. Stop codons of the prenyl transferase and terpene synthase coding sequences were removed and instead 60-bp wild type F2A sequences (5'- CAGCTGTTGAATTTTGACCTT CTTAAGCTTGCGGGAGACGTCGAGTCCAACCCTGGG CCC-3'; Ryan et al., 1991; Kenneth et al., 2012) inserted such that the three coding regions were linked into one ORF. In addition, the 5'-UTR sequence of the A. thaliana Farnesyl Diposphate Synthase 2 AtFPPS2 gene (At5g47770) was fused upstream of the start codon of the prenyl transferase coding sequence in both gene constructs.

Each of the two multicistronic gene constructs was transferred from the pTwist ENTR vector into the pMCS:GW vector carrying the AtCER5 promoter fragment by performing a standard LR recombination reaction. One μl of the destination vector (150 ng/μl) and 0.5 μl of the entry vector (30 ng/μl) were mixed with 6.5 μl TE buffer and 2 μl LR Clonase II Plus enzyme mix (Invitrogen, Thermo Fisher Scientific), and incubated at 25°C for 1 h. After the addition of 1 μl of proteinase K (Invitrogen, Thermo Fisher Scientific), reactions were incubated at 37°C for 10 min. Aliquots of each LR reaction were transformed into competent Escherichia coli DH5α cells (Thermo Fisher Scientific) and colonies with recombinant plasmids selected on LB agar plates with kanamycin (100 μg/ml).

The resulting binary pMCS:GW vectors carrying the multicistronic gene constructs under the control of the AtCER5 promoter were introduced into Agrobacterium tumefaciens (strain GV3101), and those were subsequently used for transient transformation via infiltration of tomato leaves as described previously (Norkunas et al., 2018). Selected Agrobacterium clones were grown in LB medium containing kanamycin (100 μg/ml), rifampicin (25 μg/ml), and gentamycin (50 μg/ml) to an OD₆₀₀ of 0.8–1.0. Bacterial cultures were harvested by centrifugation and resuspended in infiltration buffer [10 mm MES-KOH (pH 5.7), 10 mm MgCl₂ and 200 μm acetosyringone]. Agrobacterium suspensions were adjusted to an OD₆₀₀ of 0.6–0.8 and further incubated at room temperature for 2 h prior to leaf infiltration. Agrobacterium-mediated transient transformation was subsequently conducted on the second and older true leaves of 4-week-old odorless-2 tomato plants. The Agrobacterium suspension was injected into leaves from the abaxial side using a 5 ml plastic syringe (without hypodermic needle) thus infiltrating the intercellular spaces of the entire leaves. After the infiltration plants were covered by a humidity dome and continued to grow under a 16-h photoperiod at 23–25°C until further analysis.

Six days after the Agrobacterium infiltration total RNA was isolated from infiltrated tomato leaves as previously described (Eggermont et al., 1996). For the RT-PCR analyses, total RNA was pretreated with RNase-free DNase (New England Biolabs, Ipswich, MA, United States) and cDNA was synthesized using reverse transcriptase (Superscript II, Invitrogen, Carlsbad, CA, United States). To evaluate the expression of both multicistronic gene constructs, the cDNA was subsequently used for PCR utilizing three primer pairs specific for: AtFPPS2 (AtFPPS-fwd: 5'-CGGATCTGAAATCAACCTTCCTCGAC-3'; AtFPPS-rev: 5'- CAATGCCTTAACTACCAACCAGGAGC-3'), ShzFPPS (ShzFPPS-fwd: 5' -CAAATTCACCTCTGACAGTGTCTGC-3'; ShzFPPS-rev: 5'-GTGTGTCCACCAAAACGTCTATGCC-3'), and eGFP (eGFP-fwd: 5'- CGACGTAAACGGCCACAAGTT CA-3'; eGFP-rev: 5'- ACTTGTACAGCTCGTCCATGCC-3'). The PCR conditions were as follows: 94°C for 5 min for one cycle, followed by 35 cycles of 95°C for 60 s, 57°C for 60 s and 72°C for 60 s, and a final extension at 72°C for 10 min. The amplification products were separated by agarose gel electrophoresis, stained with GelRed^® (Biotium, United States), and analyzed using the ChemiDoc Gel Imaging System and Image Lab 5.1 software (Bio-Rad, Hercules, CA, United States).

Four to five days after Agrobacterium infiltration tomato leaves were collected, and cross- and surface sections were prepared. Leaf cross- and surface sections were analyzed with a Zeiss Axio Imager M1 compound microscope (Zeiss, Oberkochen, Germany) equipped with an X-Cite series 120Q fluorescent illuminator. Fluorescence microscopy was performed using a 485/20 excitation filter in combination with a green filter cube, and a 545/25 excitation filter in combination with a red filter cube for the analysis of GFP fluorescence and chlorophyll autofluorescence, respectively. All images were captured using a monochrome AxioCam MRm camera mounted on the Imager M1 microscope and further processed with the AxioVision software. Composite images from GFP and chlorophyll fluorescence microscopy were overlaid using the ImageJ software (National Institutes of Health).¹

Tomato leaves were collected at different time points (0, 3, 6, 9, 12, and 15 days) after the Agrobacterium infiltration. Leaves were photographed, and their surface areas were determined using the ImageJ software (National Institutes of Health, see footnote 1). Leaves were stirred in 50 ml of methyl tert-butyl ether (MTBE) for 20 min to extract terpenes as described previously (Gutensohn et al., 2013). After removing the leaves, extracts were concentrated under a gentle stream of nitrogen gas to a volume of 200 μl and centrifuged for further purification. For the analysis of terpene profiles, extracts were transferred into GC vials and supplemented with 3.33 μg of naphthalene as an internal standard.

To further investigate the accumulation of terpenes in different leaf tissues, glandular trichomes, epidermis, vasculature, and mesophyll were isolated from tomato leaves 15 days after Agrobacterium infiltration. Type VI glandular trichomes were collected from tomato leaves as described previously (Schilmiller et al., 2009) using a stretched Pasteur pipette under a stereomicroscope (SZ-ST, Olympus, Tokyo, Japan). A total of 200 trichomes from each leaf sample were accumulated into 10 ml of MTBE for the extraction of terpenes. For the isolation of the other leaf tissues, a protocol was adapted from previous studies (Endo et al., 2016; Svozil et al., 2016). Entire tomato leaflets were sandwiched between two layers of clear scotch tape, and the abaxial leaf epidermis was peeled by gently pulling off one tape. The tape with the attached abaxial leaf epidermis was immediately transferred into 10 ml MTBE for terpene extraction. The other layer of tape with the remainder of the leaf was transferred into a 50 ml tube containing 15 ml of enzyme solution [1.00% (w/v) cellulase R-10, 0.25% (w/v) macerozyme R-10, 0.4 M mannitol, 8 mm CaCl₂, and 5 mm MES-KOH, pH 5.7]. After a 10 min incubation, the leaf vasculature was isolated by using a dissecting needle and transferred into 10 ml MTBE. The remaining leaf tissue was further digested for an additional 15 min for removing mesophyll cells into the enzyme solution. Subsequently, the tape with the attached adaxial leaf epidermis was transferred into the vial which already contained the tape with the abaxial leaf epidermis. The mesophyll cells isolated by the enzyme treatment were further purified using a 30 μm cell strainer (pluriSelect, Leipzig, Germany), centrifuged at 1,000 g for 5 min, and resuspended into 10 ml MTBE. All isolated leaf tissue fractions were allowed to shake in MTBE at 100 rpm for 30 min for terpene extraction. All extracts were concentrated under a gentle stream of nitrogen gas to a volume of 200 μl and centrifuged for further purification. Concentrated and purified extracts were transferred into GC vials and supplemented with 3.33 μg of naphthalene as an internal standard.

All extracts from entire tomato leaves and leaf tissue fractions were analyzed by combined gas chromatography–mass spectrometry (GC–MS) using a TRACE 1310 gas chromatograph system linked to a TSQ 8000 Triple Quadrupole mass spectrometer (Thermo Fisher Scientific, Waltham, MA, United States). Two μl of each sample was injected under a spitless mode, volatilized at 220°C, and then separated on a TraceGOLD TG-5MS GC column (30 m length, 0.25 mm I.D., and 0.25 μm film; Thermo Fisher Scientific, Pittsburgh, PA, United States). The initial column temperature was held at 40°C for 3 min and then ramped at 5°C/min to 120°C, 10°C/min to 180°C, and 20°C/min to 300°C which was maintained for 2 min. The helium carrier gas flow was 1.3 ml/min. All samples were analyzed using the total ion chromatogram (TIC) mode. Individual terpene compounds were identified by comparing their mass spectra (15–300 m/z) with those deposited in the NIST/EPA/NIH Mass Spectral Library (NIST11; National Institute of Standards and Technology NIST, Scientific Instrument Services, Inc., NJ, United States),² as well as those reported previously (Sallaud et al., 2009). Terpenes identified in the leaf and leaf tissue extracts were quantified using a previously determined average response factor for sesquiterpenes (Wang et al., 2020) in combination with the internal naphthalene standard.

A potato aphid (M. euphorbiae) colony was established from apterae collected in the WVU Evansdale Greenhouse (Morgantown, WV, United States). To avoid experience on tomato plants prior to the non-choice assays on odorless-2 plants, aphids were allowed to reproduce parthenogenetically on potted potato plants in an insect rearing room under a 16-h photoperiod at 20–22°C. The aphid species was confirmed through barcode sequencing.

The performance (longevity and fecundity) of M. euphorbiae apterae on agroinfiltrated leaves of odorless-2 tomato plants was determined in a climate-controlled growth room (23–25°C, 50–60% relative humidity, and 16-h photoperiod) as described previously (Eichele-Nelson et al., 2018; Wang et al., 2020). Before the assay apterae (F₀) reared on potato plants were introduced on the leaf surface of odorless-2 plants. Four days after the introduction three neonate F₁ nymphs (considered as 1 day old) were carefully transferred to the surface of a young odorless-2 leaf (second or third fully expanded leaf) which had been infiltrated with Agrobacterium 2 days earlier, and subsequently enclosed in a clip cage (BioQuip Products, Rancho Dominguez, CA, United States) that was attached to the leaf. Four tomato plants each with three clip cages attached (in total 12 clip cages) were used for each of the treatments (two expression constructs and one control). Over the course of the experiment F₂ nymphs and exuviae were removed daily from the clip cages. The longevity of F₁ nymphs and their fecundity represented by the number of F₂ nymphs in each cage were recorded. The longevity and fecundity of aphids were analyzed by one-way ANOVA followed by multiple comparisons using Tukey’s HSD (α = 0.05).

To engineer formation of the two groups of sesquiterpenes with activity against M. euphorbiae that we have identified in S. habrochaites (Wang et al., 2020) into the epidermis of S. lycopersicum, we have designed two multicistronic expression constructs (Figure 1) in the binary vector pMCS:GW (Michniewicz et al., 2015). Both constructs were put under the control of the promoter of the AtCER5 gene (At1g51500; Figure 1), encoding a plasma membrane localized ABC transporter involved in the export of cuticular wax, that was shown to direct epidermis-specific expression (Pighin et al., 2004). To achieve high levels of terpene formation, we opted to co-express each of the selected S. habrochaites terpene synthases together with a respective prenyl transferase which assures sufficient availability of the required prenyl diphosphate substrate in the correct subcellular compartment. Despite the significant difference in β-caryophyllene/α-humulene formation between cultivated and wild tomato accessions (Wang et al., 2020), in both tomato species these sesquiterpenes are produced by Terpene Synthase 12 (TPS12) utilizing E,E-FPP as substrate (Schilmiller et al., 2010; Bleeker et al., 2011b). Thus, for the design of the respective expression construct (Figure 1), the coding sequence for the S. habrochaites Terpene Synthase 12 (ShTPS12, GenBank accession JN402389; Bleeker et al., 2011b) was paired with that for the A. thaliana Farnesyl Diphosphate Synthase 2 (AtFPPS2, At4g17190; Keim et al., 2012). The terpene synthase responsible for the formation of (−)-endo-α-bergamotene/(+)-α-santalene/(+)-endo-β-bergamotene in some S. habrochaites accessions, Santalene and Bergamotene Synthase (ShSBS), accepts Z,Z-FPP as substrate which is synthesized by cis-Farnesyl Diphosphate Synthase (ShzFPPS; Sallaud et al., 2009). In contrast to AtFPPS2 and ShTPS12, which are localized in the cytosol, ShzFPPS and ShSBS both carry N-terminal transit peptides that target them toward plastids. Thus, for the design of the second expression construct (Figure 1), the coding sequences of ShSBS (GenBank accession FJ194970) and ShzFPPS (GenBank accession FJ194969) (Sallaud et al., 2009) were paired. To obtain coordinated and stable expression of the multiple transgenes under the control of the AtCER5 promoter, the open reading frames encoding the terpene synthase and prenyl transferase in both expression constructs (Figure 1) were linked by a short 60 bp nucleotide sequence encoding the foot-and-mouth disease virus 2A oligopeptide (F2A; Ryan et al., 1991; Kenneth et al., 2012). This F2A sequence represents a self-processing peptide that via a ribosome skipping mechanism during the translation process leads to the separation between the upstream polypeptide ending with the C-terminal 2A sequence and the next translation product downstream (Ryan and Flint, 1997; Donnelly et al., 2001). As third part, the coding region of enhanced green fluorescent protein (eGFP) was added to both multicistronic expression constructs (Figure 1) and was likewise linked by an F2A sequence to the 3' end of the terpene synthases, ShTPS12 and ShSBS, respectively. In summary, both constructs (Figure 1), the pC5-FTG construct (AtCER5P-AtFPPS-ShTPS12-eGFP) and the pC5-zFSG construct (AtCER5P-ShzFPPS-ShSBS-eGFP), will result in the formation of three separate proteins upon expression in planta: a prenyl transferase and terpene synthase pair catalyzing the synthesis of the desired sesquiterpenes, as well as eGFP that will serve as a visual marker of the epidermis-specific expression.

We tested the function of the newly designed multicistronic expression constructs by infiltrating leaves of cultivated tomato with Agrobacterium carrying the pMCS binary vector with the inserted pC5-FTG and pC5-zFSG constructs (Figure 1), respectively, as well as the empty pMCS vector as a negative control. For these transient transformation assays, we used the tomato trichome mutant odorless-2 (Kang et al., 2010a) since it is deficient in the formation of TPS20-derived monoterpenes and TPS12-derived sesquiterpenes that are naturally found in S. lycopersicum trichomes and thus could interfere with the analysis of the engineered sesquiterpenes. To determine the expression of both multicistronic constructs upon transient transformation of the tomato leaves reverse transcription-PCR (RT-PCR) analyses were performed using primer pairs (Figure 1) specific for the first coding region (AtFPPS and ShzFPPS) and the third coding region (eGFP) in the pC5-FTG and pC5-zFSG constructs, respectively. An AtFPPS specific 788 base pair cDNA fragment could be amplified from leaves infiltrated with Agrobacterium carrying the pC5-FTG construct (Figure 2A). However, this AtFPPS fragment was not detected with untransformed odorless-2 control leaves or leaves infiltrated with Agrobacterium carrying the empty pMCS vector and the pC5-zFSG construct, respectively. In contrast, a ShzFPPS-specific 794 base pair cDNA fragment was only amplified from leaves infiltrated with Agrobacterium carrying the pC5-zFSG construct (Figure 2B) but was absent from the odorless-2 and empty vector controls as well as leaves infiltrated with Agrobacterium carrying the pC5-FTG construct. The eGFP-specific 656 base pair cDNA fragment could be amplified from leaves infiltrated with Agrobacterium carrying either the pC5-FTG or pC5-zFSG construct (Figure 2C) but was not found with the odorless-2 and empty vector controls. The fact that transcripts of the prenyl transferases and eGFP representing the first and last coding region in both constructs were detected by RT-PCR upon the transient transformation of tomato leaves (Figure 2) suggests that the entire multicistronic constructs were expressed under the control of the AtCER5 promoter.

While the RT-PCR analysis (Figure 2) in general demonstrated expression of the multicistronic constructs in tomato leaves upon transient transformation, the tissue specificity of their expression under the control of the AtCER5 promoter remained to be shown. Toward this goal, we performed confocal fluorescence microscopy of tomato leaves that had been infiltrated with Agrobacterium carrying the empty pMCS vector, pC5-FTG construct, and pC5-zFSG construct, respectively, to determine the tissue-specific accumulation of eGFP which is encoded in both multicistronic constructs. When cross-sections of tomato leaves were analyzed, GFP fluorescence was exclusively detected in both epidermal layers of leaves transiently transformed with the pC5-FTG and pC5-zFSG constructs (Figure 3A), while no respective fluorescence was observed with the empty vector control. Moreover, the GFP fluorescence did not overlap with the chlorophyll fluorescence detected in the chloroplast containing parenchyma cells of the leaf cross-sections (Figure 3A). In addition, we analyzed surface sections of the transiently transformed tomato leaves by fluorescence microscopy to further verify the expression of eGFP in epidermis cells. Upon the transient expression of the pC5-FTG and pC5-zFSG constructs in tomato leaves, GFP fluorescence could be observed in the cytosolic rim of the epidermal pavement cells (Figure 3B). The results of these fluorescence microscopy analyses not only provide further evidence that the entire pC5-FTG and pC5-zFSG constructs including eGFP are expressed, but also indicate that their expression under the AtCER5 promoter is indeed restricted to the epidermis of transformed tomato leaves.

To determine if the transient expression of the multicistronic constructs, both encoding pairs of prenyl transferases and terpene synthases, in the leaf epidermis, resulted in the formation of the expected sesquiterpenes, tomato leaves were extracted with methyl tert-butyl ether (MTBE) 15 days after Agrobacterium infiltration. The subsequent analysis of the leaf extracts by combined gas chromatography–mass spectrometry (GC–MS) demonstrated that leaves of the odorless-2 tomato mutant expressing the pC5-FTG construct had accumulated the expected ShTPS12 products β-caryophyllene and α-humulene (Figure 4A; Supplementary Figure 1). A similar analysis of leaves infiltrated with Agrobacterium carrying the empty pMCS vector found no β-caryophyllene and α-humulene accumulation (Figure 4C) which is in line with the previous characterization of the odorless-2 tomato mutant (Kang et al., 2010a; Wang et al., 2021) showing the absence of these two sesquiterpenes in this trichome mutant. In contrast, the analysis of tomato leaves expressing the pC5-zFSG construct revealed a different profile of accumulated terpenes (Figure 4B; Supplementary Figure 2) including (−)-endo-α-bergamotene, (+)-α-santalene, (−)-exo-α-bergamotene, (−)-epi-β-santalene, and (+)-endo-β-bergamotene. These five sesquiterpenes have been observed previously in in vitro enzyme assays as well as in a transgenic tobacco line as products of ShSBS when Z,Z-FPP was provided as substrate by ShzFPPS (Sallaud et al., 2009).

The quantitative analysis of the terpene accumulation in tomato leaves expressing the pC5-FTG and pC5-zFSG constructs showed that the ShTPS12- and ShSBS-derived sesquiterpene products, respectively, could be detected for the first time 6 days after the Agrobacterium infiltration (Figures 4D,E; Supplementary Table 1). Subsequently, the amounts of the sesquiterpene products in the tomato leaves continued to increase until 12 days after the Agrobacterium infiltration and appeared to remain constant afterward (Figures 4D,E; Supplementary Table 1). Remarkably, the total amount of sesquiterpenes produced after 15 days in leaves expressing the plastid localized ShzFPPS and ShSBS were 3.1-fold higher than in leaves expressing the cytosolic AtFPPS and ShTPS12 (Figures 4D,E; Supplementary Table 1).

To further verify the tissue specificity of the novel metabolic engineering approach described here, we studied the accumulation of sesquiterpenes in different tissues of tomato leaves transiently expressing the pC5-FTG and pC5-zFSG constructs. Fifteen days after Agrobacterium infiltration tomato leaves were separated into epidermis, mesophyll and vasculature fractions which were subsequently extracted with MTBE and analyzed for their terpene content by GC–MS. The ShTPS12-derived sesquiterpenes β-caryophyllene and α-humulene were found in the epidermis and mesophyll fractions of tomato leaves expressing the pC5-FTG construct (Table 1), while they were absent in the vasculature. Likewise, three of the ShSBS-derived sesquiterpenes, (−)-endo-α-bergamotene, (+)-α-santalene, and (+)-endo-β-bergamotene, were found in the epidermis fraction of leaves expressing the pC5-zFSG construct (Table 1), while only (+)-α-santalene and (+)-endo-β-bergamotene were detected in the mesophyll fraction and no ShSBS-derived sesquiterpenes were present in the vasculature of these leaves. In addition to the epidermis, mesophyll, and vasculature fractions, we also analyzed the terpene content of glandular trichomes collected from tomato leaves expressing the pC5-FTG and pC5-zFSG constructs, however, did not observe any accumulation of ShTPS12- and ShSBS-derived sesquiterpenes, respectively (Table 1). In summary, these analyses revealed that the vast majority of the ShTPS12- and ShSBS-derived sesquiterpenes, 92.99 and 92.24%, respectively, accumulate in the epidermis (Table 1) of the transiently transformed tomato leaves, thus indicating that expression of the pC5-FTG and pC5-zFSG constructs under the control of the AtCER5 promoter indeed results in the epidermis-specific production of the engineered sesquiterpenes.

As a first approach to characterize the potential of the sesquiterpene formation engineered into the leaf epidermis to affect the potato aphid (M. euphorbiae), we performed non-choice assays utilizing tomato leaves that transiently express the pC5-FTG and pC5-zFSG constructs. Newly emerged M. euphorbiae nymphs were reared in clip cages on the surface of tomato leaves that previously have been infiltrated with Agrobacterium carrying the pC5-FTG and pC5-zFSG constructs or the empty pMCS vector control, and their longevity and fecundity (represented by the number of offspring) were determined. Compared to the non-infiltrated odorless-2 control, infiltration of leaves with Agrobacterium carrying the empty pMCS vector did not significantly affect longevity (t = 0.201, p = 0.997) or fecundity (t = 0.368, p = 0.983) of M. euphorbiae. In contrast, the longevity of M. euphorbiae on tomato leaves expressing the pC5-FTG construct (20.39 ± 0.55 days), characterized by β-caryophyllene and α-humulene production in their epidermis (Figure 4; Table 1), was significantly decreased (t = 5.420, p = 0.001) compared to that on leaves infiltrated with Agrobacterium carrying the empty pMCS vector (22.78 ± 0.53 days; Figure 5A). The longevity of M. euphorbiae (Figure 5A) on tomato leaves expressing the pC5-zFSG construct (18.33 ± 0.65 days), which accumulated the ShSBS-derived sesquiterpenes in their epidermis (Figure 4; Table 1), was also significantly decreased (t = 2.678, p = 0.048) compared to that on leaves infiltrated with the empty pMCS vector control, and even further decreased compared to that on leaves expressing the pC5-FTG construct (t = 2.742, p = 0.035). Similar effects as observed for the longevity were also found for the fecundity of M. euphorbiae (Figure 5B) on tomato leaves expressing the pC5-FTG and pC5-zFSG constructs, while their fecundity on leaves infiltrated with Agrobacterium carrying the empty pMCS vector was not significantly affected (t = 0.368, p = 0.983). The number of M. euphorbiae offspring was significantly reduced (t = 3.160, p = 0.015) on leaves expressing the pC5-zFSG construct (17.83 ± 2.01 nymphs) compared to leaves infiltrated with Agrobacterium carrying the empty pMCS vector (25.33 ± 1.94 nymphs; Figure 5B). A similar trend toward a reduced number of offspring (20.83 ± 1.86 nymphs) was observed with M. euphorbiae on leaves expressing the pC5-FTG construct (Figure 5B), although their fecundity was not significantly different to that of aphids on leaves infiltrated with Agrobacterium carrying the empty pMCS vector (t = 1.679, p = 0.347) or the pC5-zFSG construct (t = 1.481, p = 0.457).