bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1 Article Title 2 An improved benthamiana strain for aphid and whitefly research 3 4 Running title 5 Acylsugars protect Nicotiana benthamiana 6 7 8 Author names 9 Honglin Feng1, Lucia Acosta-Gamboa2, Lars H. Kruse3, Alba Ruth Nava Fereira4, Sara Shakir1†, 10 Hongxing Xu1‡, Garry Sunter4, Michael A. Gore2, Gaurav D. Moghe3, Georg Jander1* 11 12 13 Author Affiliations 14 1Boyce Thompson Institute, Ithaca NY, USA 15 2Plant Breeding and Genetics Section, School of Integrative Science, Cornell University, Ithaca NY, 16 14853, USA 17 3Plant Biology Section, School of Integrative Plant Science, Cornell University, Ithaca NY, 14853, USA 18 4Department of Biology, University of Texas San Antonio, San Antonio TX, 78249, USA 19 †Present address: Gembloux Agro-Bio Tech Institute, the University of Liege, Gembloux, Belgium 20 ‡Present address: College of Life Science, the Shaanxi Normal University, Xi’an, China 21 22 *Correspondence: 23 Georg Jander 24 Boyce Thompson Institute 25 Ithaca, NY 14853 26 USA 27 Phone: 607-254-1365 28 Email: [email protected] 29 30 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

31 Abstract 32 Nicotiana benthamiana is used extensively as a platform for transient gene expression and as a model 33 system for studying plant-virus interactions. However, many -feeding insects, e.g. Myzus persicae 34 (green peach aphids) and Bemisia tabaci (whiteflies), grow poorly on N. benthamiana, limiting its utility 35 for research on plant-insect interactions. Using CRISPR/Cas9, we generated knockout mutations in two 36 N. benthamiana acylsugar acyltransferase genes, ASAT1 and ASAT2, which contribute to the biosynthesis 37 of insect-deterrent acylsucroses. Whereas ASAT1 mutations reduced the abundance of two predominant 38 acylsucroses, ASAT2 mutations caused almost complete depletion of foliar acylsucroses. Both M. persicae 39 and B. tabaci survived and reproduced significantly better on asat2 mutant than on wildtype N. 40 benthamiana. Furthermore, ASAT1 and ASAT2 mutations reduced the water content and increased the 41 temperature of , indicating that foliar acylsucroses can protect against desiccation. Improved aphid 42 and whitefly performance on ASAT2 mutants will make it possible to use the efficient transient 43 overexpression and gene expression silencing systems that are available for N. benthamiana to study 44 plant-insect interactions. Additionally, the absence of acylsugars in ASAT2 mutant lines will simplify 45 transient expression assays for the functional analysis of acylsugar biosynthesis genes from other 46 . 47 48

2 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

49 Introduction 50 Nicotiana benthamiana, a species of wild tobacco that is native to Australia, is commonly used by plant 51 molecular biologists as model system for laboratory research. Susceptibility to a wide variety of plant 52 viruses has made N. benthamiana a popular model for fundamental studies of plant-virus interactions 53 (Goodin et al., 2008; Bally et al., 2018). Scientists have developed N. benthamiana as a transgene 54 expression powerhouse by engineering viral vectors to express heterologous genes, including fluorescent 55 reporter genes to visualize cell structures (Bally et al., 2018). Antibodies, biofuel compounds, and other 56 protein and metabolite products have been produced in N. benthamiana (Arntzen, 2015; Powell, 2015). 57 Virus-induced gene silencing (VIGS), which is employed to study gene function in a variety of plant 58 species, was originally developed in N. benthamiana (Hayward et al., 2011). Recently, high-efficiency 59 CRISPR/Cas9 germline gene editing using virus-encoded guide RNA (gRNA) was demonstrated for the 60 first time in N. benthamiana (Ellison et al., 2020). Although N. benthamiana is hyper-susceptible to many 61 plant viruses, it is not a good host for two virus-transmitting Hemiptera, Myzus persicae (green peach 62 aphid) (Thurston, 1961; Hagimori et al., 1993) and Bemisia tabaci (silverleaf whitefly) (Simon et al., 63 2003), that grow well on cultivated tobacco (). 64 The poor growth of aphids and whiteflies on N. benthamiana may be attributed in part to 65 glandular . These epidermal secretory structures on the surface of ~30% of vascular plants 66 (Weinhold and Baldwin, 2011; Glas et al., 2012), have been found to play a crucial defensive role in 67 several ways: as a physical obstacle for insect movement on the plant surface (Cardoso, 2008), 68 entrapment (Simmons et al., 2004), production of volatiles and other defensive metabolites (Laue et al., 69 2000; Schilmiller et al., 2010; Glas et al., 2012), and production of proteins that repel 70 (phylloplane proteins, e.g. T-phylloplanin) (Shepherd and Wagner, 2007). In addition to their defensive 71 functions, glandular trichomes also protect plants from abiotic stresses such as transpiration water loss 72 and UV irradiation (Karabourniotis et al., 1995). 73 There are two main types of glandular trichomes on N. benthamiana leaves, large swollen-stalk 74 trichomes and small trichomes that are capped by a secretory head with one, two, or four cells (Slocombe 75 et al., 2008). The large trichomes have been shown to secrete phylloplane proteins in N. tabacum. The 76 small trichomes are the most abundant trichomes on tobacco leaf surfaces and secrete exudates, including 77 acylsugars (Wagner et al., 2004; Slocombe et al., 2008). Detached trichomes, a mixture of the large and 78 small trichomes, from N. benthamiana are able to synthesize acylsugars (Kroumova and Wagner, 2003), 79 and the secretory head cells alone are able to synthesize acylsugars in N. tabacum (Kandra and Wagner, 80 1988). 81 Acylsugars, generally sucrose or glucose esterified with aliphatic acids of different chain lengths, 82 are abundant insect-deterrent metabolites produced by Solanaceae glandular trichomes (Arrendale et al.,

3 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

83 1990; Slocombe et al., 2008; Moghe et al., 2017). Specific acylsugars are associated with aphid-resistant 84 Nicotiana species, while not being detected in more susceptible species in this genus (Hagimori et al., 85 1993). Furthermore, relative to cultivated tomatoes ( lycopersicum), acylsugars provide wild 86 tomatoes (Solanum pennellii) greater resistance against M. persicae (Rodriguez et al., 1993). 87 Diacylsucrose protects crops against tobacco aphids (Myzus persicae nicotianae), B. tabaci, and two- 88 spotted spider mites (Tetranychus urticae) (Chortyk et al., 1996; Alba et al., 2009). The synthetic sucrose 89 octanoate (an analog of Nicotiana gossei sugar esters) is effective in the field against Asian citrus psyllids 90 (Diaphorina citri), citrus leafminer (Phyllocnistis citrella), and a mite complex (including Texas citrus 91 mite, red spider mite, and rust mite) (McKenzie and Puterka, 2004). Interestingly, acylsucroses in 92 Nicotiana attenuata are metabolized to volatile fatty acids by neonate Manduca sexta (tobacco 93 hornworm) larvae, thereby tagging these larvae and attracting predatory ants, Pogonomyrmex rugosus 94 (Weinhold and Baldwin, 2011). 95 Acylsugars and leaf surface lipids more generally may contribute to plant drought tolerance. 96 Transcriptomic studies of drought tolerant wild (Solanum pennellii) populations showed that lipid 97 metabolism genes are among those that are most responsive to drought stress (Gong et al., 2010; Egea et 98 al., 2018). Additionally, acylsugars on the leaf surface in a native S. pennellii population contributed to 99 drought tolerance (Fobes et al., 1985). Acylsugars also are reported to provide protection against drought 100 stress conditions in Solanum chilense (O’ Connell et al., 2007). Similarly, abundant accumulation of 101 acylsugars with C7-C8 acyl groups the in the desert tobacco (Nicotiana obtusifolia) was suggested to 102 provide this species with high drought tolerance for its desert environment (Kroumova et al., 2016). 103 Although the mechanism is not completely understood, it has been proposed that the polar lipids reduce 104 the surface tension of adsorbed dew water, thereby allowing the leaves absorb more condensed water on 105 the surface (Fobes et al., 1985). 106 More recently, enzymes involved in the biosynthesis of acylsugars have been identified. Four 107 acylsugar acyltransferases (ASATs), SlycASAT1, SlycASAT2, SlycASAT3, and SlycASAT4, have been 108 biochemically characterized in cultivated tomato (Fan et al., 2016). SlycASAT1 catalyzes the first step of

109 sucrose acylation, using sucrose and acyl-CoA to generate monoacylsucroses via pyranose R4 acylation 110 (Fan et al., 2016). SlycASAT2 uses the product of SlycASAT1 (e.g. S1:5(iC5R4)) and acyl-CoA donor 111 substrates (e.g. iC4-CoA, aiC5-CoA, nC10-CoA, and nC12-CoA) to generate diacylsucroses (Fan et al., 112 2016). Further, SlycASAT3 uses the diacylsucroses generated by SlycASAT2 to make triacylsucroses by 113 acylating the diacylsucrose five-membered (furanose) ring (Fan et al., 2016). Then, SlycASAT4 makes 114 tetraacylsucroses by acetylating triacylsucroes using C2-CoA (Schilmiller et al., 2012). Notably, 115 SlycASAT4 (formerly SlycASAT2) is specifically expressed in the trichomes, where acylsugar 116 acetylation occurs (Schilmiller et al., 2012). The expression and activity of ASATs varies among different

4 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

117 plant species (including the order of the ASAT reactions in the pathway), which may contribute to the 118 observed chemical diversity (Kim et al., 2012). Although ASATs have been most intensively 119 studied in tomato, ASAT genes also have been annotated in the available Nicotiana genomes (Gaquerel et 120 al., 2013; Van et al., 2017; Egan et al., 2019). However, ASAT genes in N. benthamiana have not been 121 annotated and characterized previously, and their functions in defense against insect pests remained 122 unknown. 123 The goal of this study was to investigate the role of acylsugars in N. benthamiana resistance to 124 insect feeding, as well as to create an insect-susceptible ASAT mutant line to facilitate use of N. 125 benthamiana for laboratory research on plant-insect interactions. Here, we identified two ASAT genes in 126 N. benthamiana, NbenASAT1 and NbenASAT2. Using CRISPR/Cas9 to create mutant lines, we showed 127 that both NbenASAT1 and NbenASAT2 mutants had reduced acylsugar content. NbenASAT2 mutations 128 allowed increased M. persicae and B. tabaci survival, growth, and reproduction. Additionally, decreased 129 water content and elevated leaf temperature in NbenASAT1 and NbenASAT2 mutants indicated that N. 130 benthamiana acylsugars contribute to protection against desiccation. 131 132

5 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

133 Results 134 Identification of ASAT1 and ASAT2 in N. benthamiana 135 Using reciprocal comparisons to confirmed Solanaceae ASAT genes (Moghe et al., 2017), we identified 136 three highly homologous sequences in the N. benthamiana genome: Niben101Scf02239Ctg025, 137 Niben101Scf22800Ctg001, and Niben101Scf14179Ctg028 (Bombarely et al., 2012) (gene identifiers are 138 from annotations at solgenomics.net). While Niben101Scf02239Ctg025 and Niben101Scf22800Ctg001 139 were annotated as a full-length coding sequences with strong coverage in available RNAseq datasets, 140 Niben101Scf14179Ctg028 was annotated as a pseudogene because it appears to be a fragment of 141 Niben101Scf141790g02010.1, with no coverage in available RNAseq datasets. The 142 Niben101Scf02239Ctg025 and Niben101Scf22800Ctg001 sequences were confirmed in a more recently 143 assembled N. benthamiana genome (Schiavinato et al., 2019). In this assembly, the pseudogene 144 Niben101Scf14179Ctg028 was annotated as part of Niben101Scf02239Ctg025, and there were no 145 additional annotated ASAT candidates. 146 To infer ASAT evolution and function, we constructed a protein phylogenetic tree of previously 147 annotated Solanaceae ASATs (Figure 1, Figure S1; Table S1). In this tree, Niben101Scf02239Ctg025 148 formed a monophyletic group with other ASATs, including the biochemically characterized SsinASAT1, 149 PaxiASAT1 and HnigASAT1. Therefore, we named Niben101Scf02239Ctg025 as N. benthamiana 150 ASAT1 (NbenASAT1). Niben101Scf22800Ctg001 formed a monophyletic group with other ASATs 151 including the biochemically characterized NattASAT2, HnigASAT2, and PaxiASAT2. Therefore, we 152 named Niben101Scf22800Ctg001 as N. benthamiana ASAT2 (NbenASAT2). Notably, the ASAT2 153 monophyletic group also included the biochemically characterized SpenASAT1, SlycASAT1, and 154 SnigASAT1 (Figure 1). 155 156 Generation of ASAT mutants 157 Using CRISPR/Cas9 coupled with tissue culture, we obtained two independent homozygous mutants for 158 both NbenASAT1 and NbenASAT2. asat1-1 has a five-nucleotide deletion at the gRNA3 cutting site and a 159 single nucleotide insertion at the gRNA2 cutting site, leading to a frameshift between gRNA3 and 160 gRNA2. asat1-2 has a 318-nucleotide deletion between the gRNA3 and gRNA2 cutting sites (Figure 2A). 161 asat2-1 has a single-nucleotide deletion at the gRNA3 cutting site and single-nucleotide insertion at the 162 gRNA2 cutting site, leading to a translation frame shift between the two sites. asat2-2 has a 115- 163 nucleotides deletion at gRNA3 cutting site and a single-nucleotide insertion at the gRNA2 cutting site 164 (Figure 2B). 165 166 ASAT2 knockout depletes acylsugar biosynthesis

6 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

167 We quantified the acylsugar content in the ASAT mutants by LC/MS, comparing to that of wildtype 168 plants. In the LC/MS profile of N. benthamiana leaf surface washes, we characterized twelve mass 169 features as acylsucroses based on their characteristic peaks and neutral losses (Figure S2A). In a negative 170 electron spray ionization mode, the characteristic peak features included the mass of 341.11 for sucrose, 171 509.22 for sucrose + C2+C8, 467.21 for sucrose + C8, 495.21 for sucrose + C7, and 383.12 for sucrose + 172 C2; the neutral loss peaks included mass for 126.10 for C8 (acyl-chain with 8 carbons), 129.09 for C7 +

173 H2O, and 59.01 for C2 + H2O (Figure S2B). Those twelve m/z ratios included 383.12, 467. 21, 509.22, 174 555.23, 593.32, 621.31, M625.31, 635.32, 639.32, 667.32, 671.30, 681.34 (Figure S2A). Based on their 175 MS/MS peak features, retention times and relative abundances, we predicted that the identified mass 176 features are mainly derived from two acylsucroses as formate or chloride adducts, pathway intermediates 177 and/or have resulted from in-source fragmentation. We named the two acylsucroses S3:17(2,7,8) and 178 S3:18(2,8,8) (in the nomenclature, “S” refers to the sucrose backbone, “3:18” indicates three acyl chains 179 with total eighteen carbons, and the length of each acyl chain is shown in parentheses) (Figure S2A). In 180 wildtype plants, S3:18(2,8,8) is the dominant acylsucrose while S3:17(2,7,8) has relatively low 181 abundance (Figure 3). S2:16(8,8) and S2:15(7,8), which may be biosynthetic pathway intermediates for 182 S3:18(2,8,8) and S3:17(2,7,8), respectively, are present at lower levels (Figure 3). Compared to wildtype, 183 both asat2-1 and asat2-2 were almost completely depleted in both acylsucroses (Figure 3). For asat1-1 184 and asat1-2, the detected acylsucroses were less abundant, and significantly reduced in asat1-1 (Figure 185 3). Although acylsugar content was significantly reduced in ASAT mutants, the structure and abundance 186 of trichomes on the leaf surface were not changed (Figure S3). 187 188 Insect performance on ASAT2 mutant lines 189 We used ASAT1 and ASAT2 mutant lines to test the role of acylsugars in protecting N. benthamiana 190 against aphid pests. The ASAT2 mutations significantly improved M. persicae survival compared to both 191 wildtype N. benthamiana and plants with ASAT1 mutations (Figure 4A, S4A). Significant improvements 192 in aphid performance were observed as early as at 48 h post-feeding (p < 0.05) and increased until the end 193 of the 120-h monitoring period (p < 0.001) (Figure 4A, S4A). At the end of the survivorship assay, aphids 194 on both ASAT1 and ASAT1 mutants were larger than those on wildtype plants (Figure 4B, S4B). When we 195 measured progeny production by five adult aphids over a period of seven days, more than 200 nymphs 196 were produced on the ASAT2 mutants, significantly more than the number of nymphs produced on either 197 wildtype or ASAT1 mutants (p < 0.05, Figure 4C, S4C). In aphid choice assays with detached leaves, 198 aphids preferentially chose ASAT2 mutant leaves relative to wildtype and ASAT1 mutants (p < 0.05, 199 Figure 4D, S4D). This preference was consistently observed in pairwise choice assays in any combination 200 between ASAT2 mutants and either wildtype or ASAT1 mutants (p < 0.001, t-test, Figure S5). No

7 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

201 preference was observed when comparing wildtype N. benthamiana and ASAT1 mutants (p > 0.05, Figure 202 4D, S4D, S5). 203 When aphid colonies were allowed to grow long-term on asat2-1 mutant and wildtype N. 204 benthamiana in the same growth chamber, there were many more aphids on the mutant plants (Figure 205 S6A,B), likely resulting from a combination of host plant choice and increase growth on the asat2-1 206 mutant. It is noteworthy that, on the asat2-1 mutant plants, aphids were feeding on the more nutritious 207 younger leaves, but on the wildtype plants they only were able to feed on older leaves. Consistent with the 208 increased aphid presence, growth of the asat2-1 mutants was visibly reduced relative to wildtype N. 209 benthamiana (Figure S6B). 210 To determine whether depletion of acylsugars in ASAT2 mutants improves the performance of 211 other insects on N. benthamiana, we conducted experiments with B. tabaci, Heliothis virescens (tobacco 212 budworm) and Trichoplusia ni (cabbage looper). Survival of B. tabaci was greatly increased on the asat2- 213 1 mutant line relative to wildtype (Figure 5A and S6C). Moreover, whereas no eggs were observed on 214 wildtype N. benthamiana, B. tabaci on asat2-1 plants laid 0.42 +/- 0.15 eggs per adult in the course of the 215 three-day experiment. After 23 days of feeding, whiteflies of different life stages were observed on asat2- 216 1 mutant plants including eggs, nymphs, and adults (Figure S6D-F). 217 Survivorship of H. virescens and T. ni larvae on N. benthamiana was low and the mass of the 218 surviving larvae after ten days was not significantly increased on the mutant line relative to wildtype 219 (Figure 5B,C). The size of the surviving larvae was about ten-fold smaller than what we have observed 220 when these species were feeding on N. tabacum for a similar amount of time, and neither H. virescens nor 221 T. ni larvae survived until pupation. In a repeat of this experiment with 16 replicates, only four H. 222 virescens larvae survived for ten days on asat2-1 and none survived on wildtype N. benthamiana. No T. 223 ni survived on either asat2-1 or wildtype plants. 224 225 ASAT2 mutant plants lose water faster than wildtype plants 226 While investigating the ASAT mutants, we noticed that the mutant plants dried out faster than wildtype 227 plants. This effect was quantified using detached-leaf assays, in which leaves from ASAT2 mutants lost 228 significantly more water over 24 h than leaves from either wildtype or ASAT1 mutants (Figure 6A, S7A). 229 When plants are subjected to drought, high/low temperature, salinity, herbivores, or other stresses, they 230 absorb and reflect specific wavelengths of light, which have been used to determine vegetation indices. 231 For example, the water band index (WBI) (Penuelas et al., 1993) has been used to monitor changes in the 232 plant canopy/leaf water content. Using hyperspectral imaging, we determined that the leaf water content 233 of intact plants, as measured by the WBI, was significantly lower in ASAT2 mutants than in wildtype 234 (Figure 6B, S7B). Although the ASAT1 mutants did not lose water faster than wildtype in detached leaf

8 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

235 assays (Figure 6A), the leaf water content in ASAT1 mutants was significantly lower than wildtype 236 (Figure 6B, S7B). Measurement of leaf temperature by thermal imaging showed that, consistent with the 237 reduced leaf water content, the leaf temperature of the asat1-1, asat2-1, and asat2-2 mutants was 238 significantly higher than that of wildtype plants (Figure 6C, S7C). The temperature change in the asat1-2 239 mutant was not significant at the p < 0.05 level (Figure 6C, S7C). 240

9 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

241 Discussion 242 We were able to identify only two ASAT genes, NbenASAT1 and NbenASAT2, as well as a fragmented 243 pseudogene, in the N. benthamiana genome (Table S1). By contrast, in other Nicotiana species, there are 244 larger numbers of predicted ASATs, e.g. one ASAT1, one ASAT2, and 20 ASAT3-like genes in N. attenuata 245 (Gaquerel et al., 2013; Van et al., 2017), 35 ASAT3-like genes in N. tabacum, and 19 ASAT3-like genes in 246 N. tomentosiformis (Egan et al., 2019). Given the relatively small number of predicted ASAT genes in N. 247 benthamiana, other enzymes may also be involved in acylsugar biosynthesis. In Solanaceae plants, two 248 common pathways are known for aliphatic acid elongation via acetate: the fatty acid synthase (FAS) and 249 the alpha-ketoacid elongation (αKAE) pathways (Kroumova and Wagner, 2003). In the FAS pathways, 250 two carbons from an acetyl-acyl carrier are retained per elongation cycle, whereas in the αKAE pathway, 251 one carbon is retained per elongation cycle (Kroumova and Wagner, 2003). The knockdown of E1-β 252 branched-chain α-keto acid dehydrogenase (BCKD) significantly reduces acylsugars in N. benthamiana 253 (Slocombe et al., 2008). In addition, Isopropylmalate Synthase 3 (IPMS3) in cultivated and wild tomatoes 254 (Ning et al., 2015) and Acyl-Sucrose Fructo-Furanosidase 1 (ASFF1) in wild tomato (Leong et al., 2019) 255 are involved in determining acylsugar composition. Further studies will be needed to characterize other 256 genes involved in the N. benthamiana acylsugar biosynthesis pathway. 257 Acylsugars can be categorized as sucrose or glucose esters based on the sugar cores, which is 258 decorated with varying numbers or lengths of acyl chains (Kim et al., 2012). Whereas some wild 259 tomatoes produce a mixture of acylsucroses and acylglucoses, we only identified acylsucroses, consistent 260 with previous identification of these compounds in N. benthamiana (Matsuzaki et al., 1989; Matsuzaki et 261 al., 1992; Hagimori et al., 1993; Slocombe et al., 2008). Nevertheless, it has been reported that N. 262 benthamiana produces acylglucoses, although in a smaller proportion/abundance than acylsucroses 263 (Hagimori et al., 1993), and one glucose ester structure has been proposed (Matsuzaki et al., 1992). Our 264 failure to detect acylglucoses may be explained by the growth conditions, growth stage of plants, and/or 265 the LC/MS methods used. Whereas we used ~1-month old plants and LC/MS, Matsuzaki et al. used ~3- 266 month old plants and GC/MS to detect acylglucoses in N. benthamiana (Matsuzaki et al., 1992).

267 Acylsugars with C7-12 chains have been shown to be the most toxic sugar esters for small phloem- 268 feeding arthropods such as aphids, Asian citrus psyllids, and whiteflies (Chortyk et al., 1996; McKenzie 269 and Puterka, 2004; Song Z et al., 2006). Synthetic acylsucroses with di-heptanoic acid (C7), di-octanoic 270 acid (C8), and di-nonanoic acid (C9) acyl groups showed the highest mortality in bioassays with M. 271 persicae and B. tabaci (Chortyk et al., 1996; McKenzie and Puterka, 2004; Song Z et al., 2006). 272 Nicotiana gossei, a tobacco species that produces mainly C7-C8 acyl group acylsugars, has a high level of 273 insect resistance relative to close relatives with acylsugar profiles that not dominated by those with C7- 274 C12 acyl groups (Thurston, 1961; Kroumova and Wagner, 2003). In N. benthamiana, the two most

10 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

275 abundant acylsugars that we found contain C7 and predominantly C8 acyl groups, which is consistent to 276 previous findings of mainly 5- and 6-methyl heptanoate (C8) in N. benthamiana (Kroumova and Wagner, 277 2003; Slocombe et al., 2008) and in N. alata (Moghe et al., 2017). The almost complete depletion of 278 acylsugars in our ASAT2 mutants improved aphid and whitefly performance (Figure 4, 5A, S4, S6), 279 suggesting that the identified C8-chain acyl group acylsugars are providing insect resistance for N. 280 benthamiana. Increased water loss and leaf temperature in the asat2-1 and asat2-2 mutants (Figure 6, S7) 281 is consistent with prior studies that suggested a role for acylsugars in tomato drought tolerance (Fobes et 282 al., 1985; O’ Connell et al., 2007; Kroumova et al., 2016). 283 The characterized acylsugars were reduced to a greater extent in N. benthamiana ASAT2 than in 284 ASAT1 mutants (Figure 3). This observation has two implications. First, the reduction of acylsugars in 285 ASAT1 mutants indicates that NbenASAT2 functions upstream of NbenASAT1 in the acylsugar 286 biosynthesis pathway. Second, the incomplete depletion of the acylsugars suggests that NbenASAT1 287 functions may be partially compensated by other enzymes, perhaps including NbenASAT2. In the ASAT 288 phylogenetic tree (Figure 1), NbenASAT2 is closely related to some biochemically characterized ASAT1 289 proteins in other Solanaceae species, including the SpenASAT1, SlycASAT1, and SnigASAT1. Those 290 ASAT1s have some substrate overlap with the ASAT2s found in the corresponding species, indicating 291 that ASAT2 has moved toward utilizing the ASAT1 substrate in these species over time (Moghe et al., 292 2017). The final activity shift that has become fixed in the Solanum genus most likely occurred after the 293 divergence of the Solanum and Capsicum clades (Moghe et al., 2017). However, if our hypothesis of 294 partial complementation of NbenASAT1 by NbenASAT2 is correct, it may flag a transition stage or 295 suggest independent Nicotiana-specific evolution of the ASAT1/ASAT2 functions. Based on previous 296 knowledge of BAHD activities (Moghe et al., 2017), we postulate that S2:15 (7,8) and S2:16 (8,8) are 297 produced by NbenASAT1 and NbenASAT2, while the acetylation is carried out by another unrelated 298 BAHD enzyme – not unlike the distantly related SlycASAT4 and Salpiglossis sinuata ASAT5 enzymes. 299 Further characterization will be required to identify specific enzyme activities. 300 Although ASAT1 mutations cause only a partial reduction in the acylsugar profile, they 301 nevertheless lead to increased M. persicae size (Figure 4B, S4B). However, the absence of aphid 302 preference for ASAT1 mutants suggested that the insects were not able to detect the difference in the 303 acylsugar content relative to wildtype N. benthamiana. Despite the only partial decrease in the acylsugar 304 content of ASAT1 mutants, the decreases in water content and increases in leaf temperature were similar 305 to those of ASAT2 mutants (Figure 6B, C, S7B, C). The water loss phenotypes of the ASAT1 mutants 306 likely resulted from the reduced abundance of acylsugars, as asat1-1, with a relatively higher reduction of 307 acylsugars (Figure 3), has a lower leaf water content and a higher leaf temperature compared to those of

11 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

308 asat1-2 (Figure 6B, C, S7B, C). These results suggest that the two acylsucroses contribute to protecting 309 N. benthamiana against drought. 310 Nicotiana benthamiana has been used for virus transmission assays, VIGS of M. persicae genes, 311 and transient expression of genes from other species to determine their effects on aphid growth and 312 reproduction (Del Toro et al.; Ramsey et al., 2007; Bos et al., 2010; Pitino and Hogenhout, 2013; Casteel 313 et al., 2014; Elzinga et al., 2014; Rodriguez et al., 2014; Krenz et al., 2015; Tzin et al., 2015; Mulot et al., 314 2016; Mathers et al., 2017; Cui et al., 2019; Worrall et al., 2019). However, experiments of this kind are 315 hampered by the fact that many M. persicae isolates do not perform well on N. benthamiana. Similarly, 316 due to the very poor survival of B. tabaci on wildtype plants, N. benthamiana has not been useful as a 317 host plant for whitefly research. Therefore, the availability of N. benthamiana asat2 mutants will facilitate 318 future research with both aphids and whiteflies. 319 Although both H. virescens and T. ni larvae grow well on cultivated tobacco, they did not survive 320 or grew very slowly on both wildtype and ASAT mutant N. benthamiana (Figure 5B, C). This suggests 321 that other resistance mechanisms in N. benthamiana provide protection against these lepidopteran species. 322 Additional mutations that decrease insect resistance, perhaps regulatory genes such as COI1 or genes 323 affecting the production of specialized metabolites other than acylsugars, will be necessary to facilitate N. 324 benthamiana experiments with these lepidopteran species. 325 Our knockout of acylsugar biosynthesis is an important first step toward developing the excellent 326 N. benthamiana system as one that can also be used for studying plant interactions with M. persicae, B. 327 tabaci, and perhaps other agriculturally relevant hemipteran pests. Such experiments can include transient 328 expression assays to test the function of insect elicitors and insect-defensive genes from other plant 329 species in N. benthamiana, as well as VIGS to down-regulate insect gene expression. Furthermore, the 330 almost complete absence of acylsugars in the ASAT2 mutant lines, in combination with the facile transient 331 gene expression systems available for N. benthamiana, will make these mutants a suitable platform for 332 the functional analysis of ASATs from other Solanaceae. 333 334

12 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

335 Materials and Methods 336 Insect and plant cultures 337 A red strain of M. persicae (Ramsey et al., 2007; Ramsey et al., 2014) originally collected from N. 338 tabacum by Stewart Gray (Robert W. Holley Center for Agriculture & Health, Ithaca, NY) was 339 maintained on N. tabacum plants in a growth room at 23°C with a 16:8 h light:dark photoperiod. Insect 340 bioassays were conducted in the same growth room. Eggs of H. virescens and T. ni were purchased from 341 Benzon Research (www.benzonresearch.com). Nicotiana benthamiana wild type and mutant plants for 342 aphid and caterpillar experiments were maintained at 23°C and a 16:8 h light:dark photoperiod in a 343 Conviron (Winnipeg, Canada) growth chamber and, for seed production, in a greenhouse at 27/24°C 344 (day/night) with ambient light conditions. A virus-free B. tabaci MEAM1 colony was obtained from Jane 345 E. Polston (University of Florida). Nicotiana benthamiana wild type and mutant plants for whitefly 346 experiments were maintained at 26°C and a 16:8 h light:dark photoperiod in a growth room and, for seed 347 production, in a growth chamber (Percival Scientific, Perry, IA) at 27/24°C (day/night) with a 16:8 h 348 light:dark photoperiod. 349 350 Identification of ASAT1 and ASAT2 orthologs in N. benthamiana 351 To identify ASAT1 and ASAT2 orthologs in N. benthamiana, protein sequences of Salpiglossis sinuate 352 and Solanum lycopersicum ASAT1 and ASAT2 (Moghe et al., 2017) were compared to predicted proteins 353 encoded in the N. benthamiana genome. Sequences with >67% identity were selected as potential ASAT1 354 and ASAT2 candidates and nucleotide sequences were obtained from the Solanaceae Genomics Network 355 (www.solgenomics.net). The candidate ASAT sequences also were confirmed by comparing them to the 356 most recent published N. benthamiana genome assembly (Schiavinato et al., 2019). To confirm the 357 nucleotide sequences of N. benthamiana ASAT1 and ASAT2, genes were amplified with 358 ASAT1F/ASAT1R and ASAT2F/ASAT2R primers (Table S2) using genomic DNA as the template. 359 Amplified fragments were cloned in pDONOR™207 (ThermoFisher Scientific, US) and were sequenced 360 in their entirety using Sanger sequencing. This confirmatory sequencing showed no differences relative to 361 the published N. benthamiana genome. 362 363 Phylogenetic analysis of N. benthamiana ASATs 364 A protein phylogenetic tree of previously annotated Solanaceae ASATs (Figure 1, Figure S1; Table S1) 365 was constructed using maximum likelihood method. Briefly, the ASATs protein sequences were aligned 366 in program ClustalW (Thompson et al., 1994). Then the alignment was improved by removing the 367 spurious sequences and poorly aligned regions (gap threshold at 0.25) in program TrimAL v 1.4.rev22 368 (Capella-Gutierrez et al., 2009). Finally, an unrooted maximum likelihood tree was generated using the

13 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

369 improved alignment with a bootstrap of 1000 in RAxML v8.2.12 (Stamatakis, 2014). The tree was 370 visualized and presented using FigTree v1.4.4 (Computer program obtained from http://tree.bio.ed.ac.uk). 371 372 sgRNA design and plasmid cloning 373 Single-guide RNAs (sgRNA) targeting ASAT1 and ASAT2 were designed based on the coding regions 374 using online software, CRISPR-P v2.0 (Liu et al., 2017) and CRISPRdirect (https://crispr.dbcls.jp/), based 375 on two parameters, cleavage efficiency and potential off-targets. Additionally, only sgRNAs with >40% 376 GC content were selected. Three Cas9/gRNA constructs each were constructed for ASAT1 and ASAT2 377 following a previously developed CRISPR/Cas9 system (Jacobs et al., 2015). Four segments of DNA 378 were prepared with 20 bp overlaps on their ends: ssDNA gRNA oligo, linearized p201N:Cas9 plasmid 379 (Addgene 59175-59178), the Medicago truncatula (Mt) U6 promoter (377bp), and a scaffold DNA 380 (106bp). For ssDNA gRNA, oligos targeting either the sense or antisense of target genes were designed 381 as: sense oligo TCAAGCGAACCAGTAGGCTT--GN19--GTTTTAGAGCTAGAAATAGC, and 382 antisense oligo GCTATTTCTAGCTCTAAAAC--N19C—AAGCCTACTGGTTCGCTTGA. The gRNA 383 oligo sequences were shown in (Figure 2 and Table S2). The oligo sequences were synthesized by 384 Integrated DNA Technologies (www.idtdna.com), then 1 μl of each 100 μM oligo was added to 500 μl 1x 385 NEB buffer 2 (New England Biolabs, www.neb.com). The p201N:Cas9 plasmid was linearized by 386 digestion with Spe1 (www.neb.com) in 1x buffer 4 at 37°C for 2 h followed by a column purification and 387 a second digestion with Swal in 1 x buffer 3.1 at 25°C for 2 h. Complete digestions were confirmed on a 388 0.8% agarose gel. The MtU6 promoter and Scaffold DNAs were PCR-amplified from the pUC gRNA 389 Shuttle plasmid (Jacobs et al., 2015) using the primers Swal_MtU6F/MtU6R and 390 ScaffoldF/Spe_ScaffoldR, respectively (Table S2). The PCR reactions were performed with a high- 391 fidelity polymerase (2x Kapa master mix; www.sigmaaldrich.com) using the program: 95°C for 3 min 392 followed by 31 cycles of 98°C for 20 sec, 60°C for 30 sec, 72°C for 30 sec, and a final extension of 72°C 393 for 5 mins. Finally, cloning was done using the NEBuilder® HiFi DNA Assembly Cloning Kit. For each 394 reaction, the four pieces of DNA were mixed in a 20 μl reaction with the NEBbuilder assembly mix with 395 a final concentration of 0.011 pmol (~100 ng) of p201N:Cas9 plasmid, 0.2 pmol of MtU6 amplicon (~ 50 396 ng), scaffold amplicon (~ 12 ng) and ssDNA gRNA oligo (60-mer, 1 μl). The reactions were placed in 397 thermal cycler at 50°C for 1 h. Then, 2 μl of the reaction were transformed into 50 μl of the One Shot™ 398 Top10 chemically competent cells (Invitrogen, www.thermofisher.com) and plated with 50 μl/ml 399 kanamycin selection. Colonies with correct inserts were screened using the primers of Ubi3p218R and 400 IScelR (Table S2). Plasmids carrying the right gRNA constructs were then transformed into 401 Agrobacterium tumefaciens strain GV3101 for generating transgenic plants. All constructs were 402 confirmed by Sanger sequencing.

14 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

403 To avoid off-target effects, gRNAs were further checked by comparison against the reference N. 404 benthamiana genome v1.0.1 (www.solgenomics.net). Only two sites were found to have non-target 405 matches >17 nt (both with 1 internal mismatch), and with the NGG PAM sequence on the correct strand. 406 These two sites were checked by PCR amplification and Sanger sequencing and showed no unexpected 407 editing in our asat1-1, asat1-2, asat2-1, or asat2-2 mutant plants. Primers used for off-target Sanger 408 sequencing are listed in Table S2. 409 410 Transient mutagenesis of ASATs in N. benthamiana 411 For both ASAT1 and ASAT2, we Agrobacterium-infiltrated N. benthamiana plants at the four-leaf stage. 412 For ASAT1, we used Agrobacterium carrying the Cas9/gRNA constructs at OD of 1 and 3, whereas for 413 ASAT2, we used an OD 1.5. Each leaf was saturated with Agrobacterium solution. After infiltration, the 414 plants were cultured in a growth chamber for 2 days and then the infiltrated whole leaves were collected 415 for genomic DNA extraction and tested by PCR amplification for detection of insertion/deletion 416 polymorphisms in the target region (Figure S8). During the method optimization, a positive control 417 construct targeting the N. benthamiana Drm3 gene (gRNA: GCCACTATCTGGCCGGGGAC, provided 418 by the Greg Martin lab, Boyce Thompson Institute) was infiltrated in parallel. 419 420 Stable mutagenesis of ASATs using tissue culture 421 The stable ASAT mutant N. benthamiana plants were created in the Boyce Thompson Institute plant 422 transformation facility using CRISPR/Cas9 with gRNAs that had been confirmed to be functional as 423 described above and a previously described protocol (Van Eck et al., 2019), with minor modifications. To 424 prepare plants for transformation, we disinfected N. benthamiana seeds with 1.5 ml 1.25% sodium 425 hypochlorite (1:5 dilution of 5.25% sodium hypochlorite Clorox bleach), with 100 μl Tween-20 added, 426 for 20 min on a shaker platform. After sodium hypochlorite treatment, seeds were rinsed three times with 427 sterile water. Subsequently, we placed seeds on Tobacco Seed Germination Medium (composition per 428 liter: 1.08g Murashige and Skoog salt formulation, 30g sucrose, and 8 g agar, pH=5.7±0.1) and incubated 429 the seeds under yellow filtered light with a 16-hr photoperiod at 27°C. After ~2 weeks, the seedlings were 430 transferred to the Rooting Medium (composition per liter: 2.15 g Murashige and Skoog salt formulation, 431 30 g sucrose, 1 ml B5 vitamin/amino acid stock, and 8 g agar, pH=5.7±0.1). After ~6 weeks, the plants 432 were ready for Agrobacterium infection. First, Agrobacterium carrying the gRNA plasmids was amplified 433 in YEP medium (composition per liter: 10g yeast extract, 10g peptone, and 5g NaCl with antibiotics, 50 434 µg/µl gentamycin, 50 µg/µl kanamycin, and 50 µg/µl rifampicin), and Agrobacterium culture then was 435 pelleted and combined to a certain OD (OD 1.0 for ASAT1 gRNAs, and OD 1.0 or 1.5 for ASAT2 gRNAs)

436 with filtered cell suspension buffer (10 mM MES, 10 mM MgCl2 and 200 µM acetosyringone). Then,

15 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

437 fully expanded but immature N. benthamiana leaves were dissected into 5 mm segments by removing the 438 leaf margins. The leaf segments were incubated in the Agrobacterium culture for 30 min with shaking. 439 The incubated leaf segments were transferred to co-cultivation medium plates (composition per liter: 4.3g 440 Murashige and Skoog salt formulation, 30g sucrose, 1 ml B5 vitamin/amino acid stock, and 5g agar, 441 pH=5.7±0.1) and kept in the dark at room temperature. Leaf segments treated with YEP medium were 442 used as a negative control. After three days, the leaf segments were transferred from co-cultivation 443 medium to shoot bud initiation medium with selections (composition per liter: 4.3g Murashige and Skoog 444 salt formulation, 30g sucrose, 1 ml B5 vitamin/amino acid stock, 0.1 mg NAA, 1.0 mg BA, and 6g agar, 445 pH=5.7±0.1; with 200 μg/ml kanamycin and 250 μg/ml Timentin). Medium with Timentin but without 446 kanamycin was used as a positive control. Leaf segments on selection medium were incubated under a 447 16:8 h light-dark photoperiod at 24-25°C under yellow filtered light. The leaf segments were sub-cultured 448 every 2-weeks onto the same medium until shoot buds were visible. Finally, to generate roots, shoot buds 449 were transferred to rooting medium (composition per liter: 2.15g Murashige and Skoog salt formulation, 450 30g sucrose, 1 ml B5 vitamin/amino acid stock, 8g agar with 200 μg/ml kanamycin and 250 μg/ml 451 Timentin, pH=5.7±0.1). 452 Rooted plants from tissue culture were transferred to soil and checked for the intended mutations 453 by PCR amplification and Sanger sequencing. Confirmed homozygous ASAT1 and ASAT2 mutants were 454 kept for seed collection. The presence of gRNA and Cas9 in transgenic plants also was confirmed using 455 PCR (MtU6252F/p201R primer pair for gRNAs, and Cas9F7/nosT-rev2 primer pair for Cas9) (Table S2). 456 457 Acylsugar measurements by LC/MS 458 Liquid chromatography/mass spectrometry (LC/MS) was used to confirm the effect of the ASAT 459 mutations by measuring acylsugar content in leaf extracts from wildtype and ASAT mutant plants. New 460 leaflets were rinsed in acylsugar extraction solution (3:3:2 acetonitrile:isopropanol:water, 0.1% formic 461 acid, and 1 μM Telmisartan as internal standard) and gently agitated for 2 min. Then, the extraction 462 solutions were transferred to LC/MS glass vials, and the leaves were air dried for leaf weight 463 measurements. 464 Chromatography of leaf surface washes was performed on a ThermoScientific Ultimate 3000 465 HPLC with a glass vial autosampler and coupled with a Thermo Scientific Q Exactive™ Hybrid 466 Quadrupole-Orbitrap™ Mass Spectrometer (Mass Spectrometry Facility at Boyce Thompson Institute). 467 Acylsugar extracts were separated on an Ascentis Express C18 HPLC column (10 cm × 2.1 mm × 2.7 468 μm) (Sigma‐Aldrich, St. Louis, MO) with a flow rate of 0.3 ml/min, using a gradient flow of 0.1% formic 469 acid (Solvent A) and 100% acetonitrile (Solvent B). We used a 7-min LC method for metabolite profiling 470 involved a linear gradient from 95:5 A:B to 0:98 A:B. Full‐scan mass spectra were collected (mass range:

16 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

471 m/z 50–1000) in both positive and negative electron spray ionization (ESI) modes. Mass spectral 472 parameters were set as follows: capillary spray voltage 2.00 kV for negative ion-mode and 3.00 kV for 473 positive ion-mode, Source temperature: 100°C, desolvation temperature 350°C, desolvation nitrogen gas 474 flow rate: 600 liters/h, cone voltage 35 V. Acylsugars were identified and annotated using Thermo 475 Xcalibur Qual Browser (Thermo Fisher) and MS-DIAL v4.20 based on the MS/MS peak features and 476 neutral losses. The acylsugar abundances were estimated using peak areas at the respective m/z channel 477 under negative ESI mode. Acylsugar quantification was first normalized to the internal control 478 Telmisartan to account for technical variation between samples, and then normalized to the leaf dry 479 weight to allow comparisons between samples. 480 481 Insect bioassays 482 To measure aphid growth on ASAT mutant and wildtype plants, we placed two leaf cages on each plant 483 (Figure S9A). Twenty adult aphids from N. tabacum (naïve to N. benthamiana) were placed in each cage 484 and allowed to generate nymphs for ~12 hrs. Twenty-five nymphs were left in each cage and were 485 monitored for 120 hours to assess nymph survival. At the end of aphid survival monitoring period, five 486 aphids were left in each cage and reproduction was monitored for one week. Finally, the remaining aphids 487 were collected to measure aphid size by taking a picture and assessing the area of each aphid using 488 ImageJ (Schneider et al., 2012). 489 Aphids choice assays were performed with detached leaves. Briefly, similarly-sized leaves from 490 eight individual ASAT mutant and wildtype plants were cut and placed at each side of 15-cm Petri dishes, 491 with their petioles inserted in moistened cotton swabs (Figure S9B). Ten naïve adult aphids, which had 492 not previously encountered N. benthamiana, were released at the midpoint in between wildtype and asat1 493 or asat2 leaves, and the Petri dishes were placed in complete darkness. The aphids on each leaf were 494 counted at 24 h after their release in the Petri dishes. 495 To measure whitefly survival and fecundity on wildtype and ASAT2 mutant N. benthamiana 496 plants, we placed three plants at the seven-leaf stage in each of two cages. Ninety adult whiteflies reared 497 on Brassica oleracea (cabbage variety Earliana; Burpee catalog number 62729A) were introduced into 498 each cage with N. benthamiana (30 whiteflies/plant) and were allowed to feed for three days. The 499 numbers of whiteflies surviving on each host plant were counted, after which the remaining insects were 500 terminated using insecticidal soap. The following day, the number of whitefly eggs on each plant were 501 counted. 502 Eggs of H. virescens and T. ni were hatched on artificial diet (Southland Products, Lake Village, 503 Arkansas). Neonate larvae were confined onto individual N. benthamiana leaves, one larva per plant, 504 using organza mesh bags. After ten days, the surviving larvae were counted and weighed.

17 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

505 506 Leaf desiccation assay 507 To measure the leaf water loss, two leaves from each of eight plants were detached. The fresh weight of 508 each leaf was determined on a Sartorius Ultra Micro Balance. All leaves were placed at 23°C and a 16:8 h 509 light:dark photoperiod. Each leaf was weighted after 24 hours and the percentage of water loss was 510 calculated as [-(final weight - fresh weight)/fresh weight]*100%. 511 512 Thermal and hyperspectral imaging 513 Thermal images were acquired in the growth chamber environment using a thermal camera (A655sc, 514 FLIR Systems Inc., Boston, MA, USA) with a spectral range of 7.5–14.0 mm and a resolution of 640 x 515 480 pixels. The camera was placed approximately 1 m away from each plant and a white background was 516 used when the plant images were acquired. One region of interest, corresponding to the perimeter of each 517 leaf, was specified per leaf for each of 20 leaves per genotype. Using the FLIR ResearchIR Max software 518 (version 4.40.9.30) thermal images files were exported as CSV files. Images were segmented from the 519 background using Gaussian mixture models in MATLAB to determine the temperature of each leaf. After 520 segmentation, the temperature was averaged across the segmented leaf. 521 Hyperspectral images were acquired in a dark room using a hyperspectral imager (SOC710, 522 Series 70-V, Surface Optics Corporation, San Diego, CA, USA) that covered a spectral range from 400 to 523 1000 nm for 128 wavebands. Image acquisition and recording were performed using a Dell DELL XPS 524 15 9570 laptop computer that controls the camera. The camera was fixed using a stand with the lens 525 facing the plants and capturing top view images approximately 1 m away. A Spectralon tile (Labsphere, 526 Inc, North Sutton NH, USA) was placed next to the plant trays, covering one corner of the image to 527 facilitate subsequent image processing and calibration. The nominal reflectance value for the Spectralon 528 tile was 99% and it had a 30.5cm x 30.5cm reflective area. Lighting consisted of two halogen lamps 529 placed at ~ 45° angles on either side of the camera to create an even light distribution. All image analysis 530 was performed in HSIviewer, a MATLAB package (Stone et al., 2020). White reflectance calibration was 531 performed using the Spectralon tile. One region of interest ROI was specified for each of 20 leaves per N. 532 benthamiana genotype. This ROI corresponded to the perimeter of each leaf. From each hyperspectral 533 cube image, the vegetation pixels (green portion of the plant) were extracted using the Normalized 534 Difference Vegetation Index (NDVI). Mean reflectance (R) was calculated per band per 10 leaves in 535 order to obtained the WBI results 536 To calculate NDVI and WBI we used the following formulas, where 푅 corresponds to the 537 reflectance at a specific wavelength (nm): WBI = (R970/R900) (Penuelas et al., 1993) and NVDI = (R750 538 - R705)/(R750 + R705) (Gitelson and Merzlyak, 1994).

18 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

539 540 Statistical analysis 541 All statistical comparisons were conducted using SPSS v25, R and MATLAB R2019a (MathWorks, Inc., 542 Natick, MA, USA). ANOVA followed by a Dunnett’s post hoc test was used to determine differences in 543 leaf water loss, leaf temperature, and WBI across genotypes in each data set. ANOVA followed by a 544 Duncan post hoc test was used for aphid bioassay and LC/MS results. A two-tailed independent samples 545 t-test was used to test for differences in pairwise aphid-choice assays, whitefly, and lepidopteran assays.

19 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

546 Acknowledgements 547 We want to thank Patricia Keen and Joyce Van Eck for their help with N. benthamiana stable 548 transformation and access to the tissue culture facility, Ning Zhang and Greg Martin for sharing the 549 p201N-Cas9 construct and the Drm3 sgRNA control construct, and William Stone and Thomas Lawton 550 for providing custom image processing software. This research was supported by Cornell startup funds to 551 G.D.M., Deutsche Forschungsgemeinschaft award #411255989 to L.H.K., and United States Department 552 of Agriculture Biotechnology Risk Assessment Grant 2017-33522-27006, US National Science 553 Foundation award IOS-1645256, and Defense Advanced Research Projects Agency (DARPA) agreement 554 HR0011-17-2-0053 to G.J. G.S. is part of a team supporting DARPA's Insect Allies program under 555 agreement HR0011-17-2-0055. M.A.G. is part of a team supporting DARPA's Advanced Plant 556 Technologies program under agreement HR0011-18-C-0146. The views and conclusions contained in this 557 document are those of the authors and should not be interpreted as representing the official policies of the 558 U.S. Government. 559 560 Conflict of interest 561 The authors declare that there is no conflict of interest. 562 563 Author Contributions 564 G.J. and H.F. conceived the original research plans; H.F., S.S., L.A., H.X., L.K. and A. N. F. performed 565 the experiments; H.F., L.A., L.K., and G.D.M. analyzed the data; M.A.G., G.D.M., G.S. and G.J. 566 supervised the experiments; H.F. and G.J. wrote the article with contributions of all the authors; G.J. 567 agrees to serve as the contact author responsible for communication and distribution of samples.

20 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

568 References 569 Alba JM, Montserrat M, Fernandez-Munoz R (2009) Resistance to the two-spotted spider mite 570 (Tetranychus urticae) by acylsucroses of wild tomato (Solanum pimpinellifolium) trichomes 571 studied in a recombinant inbred line population. Experimental and Applied Acarology 47: 35-47 572 Arntzen C (2015) Plant-made pharmaceuticals: from 'Edible Vaccines' to Ebola therapeutics. Plant 573 Biotechnology Journal 13: 1013-1016 574 Arrendale RF, Severson RF, Sisson VA, Costello CE, Leary JA, Himmelsbach DS, Vanhalbeek H 575 (1990) Characterization of the sucrose ester fraction from Nicotiana glutinosa. Journal of 576 Agricultural and Food Chemistry 38: 75-85 577 Bally J, Jung H, Mortimer C, Naim F, Philips JG, Hellens R, Bombarely A, Goodin MM, 578 Waterhouse PM (2018) The rise and rise of Nicotiana benthamiana: a plant for all reasons. 579 Annual Review of Phytopathology, Vol 56 56: 405-426 580 Bombarely A, Rosli HG, Vrebalov J, Moffett P, Mueller LA, Martin GB (2012) A draft genome 581 sequence of Nicotiana benthamiana to enhance molecular plant-microbe biology research. Mol 582 Plant-Micr Int 25: 1523-1530 583 Bos JIB, Prince D, Pitino M, Maffei ME, Win J, Hogenhout SA (2010) A functional genomics 584 approach identifies candidate effectors from the aphid species Myzus persicae (green peach 585 aphid). Plos Genetics 6: e1001216. 586 Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T (2009) trimAl: a tool for automated alignment 587 trimming in large-scale phylogenetic analyses. Bioinformatics 25: 1972-1973 588 Cardoso MZ (2008) handling of a plant's trichome: The case of Heliconius charithonia (L.) 589 (Lepidoptera : Nymphalidae) and Passiflora Lobata (Killip) Hutch. (Passifloraceae). Neotropical 590 Entomology 37: 247-252 591 Casteel CL, Yang CL, Nanduri AC, De Jong HN, Whitham SA, Jander G (2014) The NIa-Pro 592 protein of Turnip mosaic virus improves growth and reproduction of the aphid vector, Myzus 593 persicae (green peach aphid). Plant Journal 77: 653-663 594 Chortyk OT, Pomonis JG, Johnson AW (1996) Syntheses and characterizations of insecticidal sucrose 595 esters. Journal of Agricultural and Food Chemistry 44: 1551-1557 596 Cui N, Lu H, Wang TZ, Zhang WH, Kang L, Cui F (2019) Armet, an aphid effector protein, induces 597 pathogen resistance in plants by promoting the accumulation of salicylic acid. Philosophical 598 Transactions of the Royal Society B-Biological Sciences 374 599 Del Toro FJ, Mencía E, Aguilar E, Tenllado F, Canto T HCPro-mediated transmission by aphids of 600 purified virions does not require its silencing suppression function and correlates with its ability 601 to coat cell microtubules in loss-of-function mutant studies. Virology 525: 10-18

21 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

602 Egan AN, Moore S, Stellari GM, Kang BC, Jahn MM (2019) Tandem gene duplication and 603 recombination at the AT3 locus in the Solanaceae, a gene essential for capsaicinoid biosynthesis 604 in Capsicum. Plos One 14: e0210510 605 Egea I, Albaladejo I, Meco V, Morales B, Sevilla A, Bolarin MC, Flores FB (2018) The drought- 606 tolerant Solanum pennellii regulates leaf water loss and induces genes involved in amino acid and 607 ethylene/jasmonate metabolism under dehydration. Sci Rep 8: 2791 608 Ellison EE, Nagalakshmi U, Gamo ME, Huang PJ, Dinesh-Kumar S, Voytas DF (2020) Multiplexed 609 heritable gene editing using RNA viruses and mobile single guide RNAs. Nat Plants 6: 620-624 610 Elzinga DA, De Vos M, Jander G (2014) Suppression of plant defenses by a Myzus persicae (green 611 peach aphid) salivary effector protein. Mol Plant Microbe Interact 27: 747-756 612 Fan P, Miller AM, Schilmiller AL, Liu X, Ofner I, Jones AD, Zamir D, Last RL (2016) In vitro 613 reconstruction and analysis of evolutionary variation of the tomato acylsucrose metabolic 614 network. Proc Natl Acad Sci U S A 113: E239-248 615 Fobes JF, Mudd JB, Marsden MP (1985) Epicuticular lipid accumulation on the leaves of Lycopersicon 616 pennellii (Corr.) D'Arcy and Lycopersicon esculentum Mill. Plant Physiol 77: 567-570 617 Gaquerel E, Kotkar H, Onkokesung N, Galis I, Baldwin IT (2013) Silencing an N-acyltransferase-like 618 involved in lignin biosynthesis in Nicotiana attenuata dramatically alters herbivory-induced 619 phenolamide metabolism. Plos One 8: e62336 620 Gitelson A, Merzlyak MN (1994) Spectral reflectance changes associated with autumn senescence of 621 Aesculus hippocastanum L and Acer platanoides L leaves - spectral features and relation to 622 chlorophyll estimation. J Plant Physiol 143: 286-292 623 Glas JJ, Schimmel BCJ, Alba JM, Escobar-Bravo R, Schuurink RC, Kant MR (2012) Plant 624 glandular trichomes as targets for breeding or engineering of resistance to herbivores. 625 International Journal of Molecular Sciences 13: 17077-17103 626 Gong P, Zhang J, Li H, Yang C, Zhang C, Zhang X, Khurram Z, Zhang Y, Wang T, Fei Z, Ye Z 627 (2010) Transcriptional profiles of drought-responsive genes in modulating transcription signal 628 transduction, and biochemical pathways in tomato. J Exp Bot 61: 3563-3575 629 Goodin MM, Zaitlin D, Naidu RA, Lommel SA (2008) Nicotiana benthamiana: its history and future 630 as a model for plant-pathogen interactions. Mol Plant Microbe Interact 21: 1015-1026 631 Hagimori M, Matsui M, Matsuzaki T, Shinozaki Y, Shinoda T, Harada H (1993) Production of 632 somatic hybrids between Nicotiana benthamiana and Nicotiana tabacum and their resistance to 633 aphids. Plant Science 91: 213-222 634 Hayward A, Padmanabhan M, Dinesh-Kumar SP (2011) Virus-induced gene silencing in Nicotiana 635 benthamiana and other plant species. Plant Reverse Genetics: Methods and Protocols 678: 55-63

22 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

636 Jacobs TB, LaFayette PR, Schmitz RJ, Parrott WA (2015) Targeted genome modifications in soybean 637 with CRISPR/Cas9. Bmc Biotechnology 15: 16 638 Kandra L, Wagner GJ (1988) Studies of the site and mode of biosynthesis of tobacco trichome exudate 639 components. Archives of Biochemistry and Biophysics 265: 425-432 640 Karabourniotis G, Kotsabassidis D, Manetas Y (1995) Trichome density and its protective potential 641 against ultraviolet-B radiation damage during leaf development. Canadian Journal of Botany- 642 Revue Canadienne De Botanique 73: 376-383 643 Kim J, Kang K, Gonzales-Vigil E, Shi F, Jones AD, Barry CS, Last RL (2012) Striking natural 644 diversity in glandular trichome acylsugar composition is shaped by variation at the 645 acyltransferase2 locus in the wild tomato Solanum habrochaites. Plant Physiol 160: 1854-1870 646 Krenz B, Bronikowski A, Lu XY, Ziebell H, Thompson JR, Perry KL (2015) Visual monitoring of 647 Cucumber mosaic virus infection in Nicotiana benthamiana following transmission by the aphid 648 vector Myzus persicae. Journal of General Virology 96: 2904-2912 649 Kroumova AB, Wagner GJ (2003) Different elongation pathways in the biosynthesis of acyl groups of 650 trichome exudate sugar esters from various solanaceous plants. Planta 216: 1013-1021 651 Kroumova ABM, Zaitlin D, Wagner GJ (2016) Natural variability in acyl moieties of sugar esters 652 produced by certain tobacco and other Solanaceae species. Phytochemistry 130: 218-227 653 Laue G, Preston CA, Baldwin IT (2000) Fast track to the trichome: induction of N-acyl nornicotines 654 precedes induction in Nicotiana repanda. Planta 210: 510-514 655 Leong BJ, Lybrand DB, Lou YR, Fan PX, Schilmiller AL, Last RL (2019) Evolution of metabolic 656 novelty: A trichome-expressed invertase creates specialized metabolic diversity in wild tomato. 657 Science Advances 5: eaaw3754 658 Liu H, Ding YD, Zhou YQ, Jin WQ, Xie KB, Chen LL (2017) CRISPR-P 2.0: an improved CRISPR- 659 Cas9 tool for genome editing in plants. Molecular Plant 10: 530-532 660 Mathers TC, Chen Y, Kaithakottil G, Legeai F, Mugford ST, Baa-Puyoulet P, Bretaudeau A, 661 Clavijo B, Colella S, Collin O, Dalmay T, Derrien T, Feng H, Gabaldon T, Jordan A, Julca 662 I, Kettles GJ, Kowitwanich K, Lavenier D, Lenzi P, Lopez-Gomollon S, Loska D, Mapleson 663 D, Maumus F, Moxon S, Price DR, Sugio A, van Munster M, Uzest M, Waite D, Jander G, 664 Tagu D, Wilson ACC, van Oosterhout C, Swarbreck D, Hogenhout SA (2017) Rapid 665 transcriptional plasticity of duplicated gene clusters enables a clonally reproducing aphid to 666 colonise diverse plant species. Genome Biology 18: 27 667 Matsuzaki T, Shinozaki Y, Hagimori M, Tobita T, Shigematsu H, Koiwai A (1992) Novel 668 glycerolipids and glycolipids from the surface-lipids of Nicotiana benthamiana. Bioscience 669 Biotechnology and Biochemistry 56: 1565-1569

23 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

670 Matsuzaki T, Shinozaki Y, Suhara S, Ninomiya M, Shigematsu H, Koiwai A (1989) Isolation of 671 glycolipids from the surface-lipids of Nicotiana bigelovii and their distribution in Nicotiana 672 species. Agricultural and Biological Chemistry 53: 3079-3082 673 McKenzie CL, Puterka GJ (2004) Effect of sucrose octanoate on survival of nymphal and adult 674 Diaphorina citri (Homoptera : Psyllidae). Journal of Economic Entomology 97: 970-975 675 Moghe GD, Leong BJ, Hurney SM, Jones AD, Last RL (2017) Evolutionary routes to biochemical 676 innovation revealed by integrative analysis of a plant-defense related specialized metabolic 677 pathway. Elife 6: e28468 678 Mulot M, Boissinot S, Monsion B, Rastegar M, Clavijo G, Halter D, Bochet N, Erdinger M, Brault 679 V (2016) Comparative analysis of RNAi-based methods to down-regulate expression of two 680 genes expressed at different levels in Myzus persicae. Viruses 8: 316 681 Ning J, Moghe GD, Leong B, Kim J, Ofner I, Wang Z, Adams C, Jones AD, Zamir D, Last RL 682 (2015) A feedback-insensitive isopropylmalate synthase affects acylsugar composition in 683 cultivated and wild tomato. Plant Physiol 169: 1821-1835 684 O’ Connell MA, Medina AL, Sanchez Pena Pand Trevino MB (2007) Molecular genetics of drought 685 resistance response in tomato and related species. In Razdan, M. K., & Mattoo, A. K. (eds). 686 Genetic Improvement of Solanaceous Crops 2: Enfield USA 261-283 687 Penuelas J, Filella I, Biel C, Serrano L, Save R (1993) The reflectance at the 950-970 nm region as an 688 indicator of plant water status. International Journal of Remote Sensing 14: 1887-1905 689 Pitino M, Hogenhout SA (2013) Aphid protein effectors promote aphid colonization in a plant species- 690 specific manner. Molecular Plant-Microbe Interactions 26: 130-139 691 Powell JD (2015) From pandemic preparedness to biofuel production: tobacco finds its biotechnology 692 niche in North America. Agriculture-Basel 5: 901-917 693 Ramsey JS, Elzinga DA, Sarkar P, Xin YR, Ghanim M, Jander G (2014) Adaptation to nicotine 694 feeding in Myzus persicae. J Chem Ecol 40: 869-877 695 Ramsey JS, Wilson AC, de Vos M, Sun Q, Tamborindeguy C, Winfield A, Malloch G, Smith DM, 696 Fenton B, Gray SM, Jander G (2007) Genomic resources for Myzus persicae: EST sequencing, 697 SNP identification, and microarray design. BMC Genomics 8: 423 698 Rodriguez AE, Tingey WM, Mutschler MA (1993) Acylsugars of Lycopersicon pennellii deter settling 699 and feeding of the green peach aphid (Homoptera, Aphididae). Journal of Economic Entomology 700 86: 34-39 701 Rodriguez PA, Stam R, Warbroek T, Bos JIB (2014) Mp10 and Mp42 from the aphid species Myzus 702 persicae trigger plant defenses in Nicotiana benthamiana through different activities. Molecular 703 Plant-Microbe Interactions 27: 30-39

24 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

704 Schiavinato M, Strasser R, Mach L, Dohm JC, Himmelbauer H (2019) Genome and transcriptome 705 characterization of the glycoengineered Nicotiana benthamiana line DeltaXT/FT. BMC 706 Genomics 20: 594 707 Schilmiller A, Shi F, Kim J, Charbonneau AL, Holmes D, Jones AD, Last RL (2010) Mass 708 spectrometry screening reveals widespread diversity in trichome specialized metabolites of 709 tomato chromosomal substitution lines. Plant Journal 62: 391-403 710 Schilmiller AL, Charbonneau AL, Last RL (2012) Identification of a BAHD acetyltransferase that 711 produces protective acyl sugars in tomato trichomes. Proc Natl Acad Sci U S A 109: 16377- 712 16382 713 Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. 714 Nature Methods 9: 671-675 715 Shepherd RW, Wagner GJ (2007) Phylloplane proteins: emerging defenses at the aerial frontline? 716 Trends in Plant Science 12: 51-56 717 Simmons AT, Gurr GM, McGrath D, Martin PM, Nicol HI (2004) Entrapment of Helicoverpa 718 armigera (Hubner) (Lepidoptera:Noctuidae) on glandular trichomes of Lycopersicon species. 719 Australian Journal of Entomology 43: 196-200 720 Simon B, Cenis JL, Demichelis S, Rapisarda C, Caciagli P, Bosco D (2003) Survey of Bemisia tabaci 721 (Hemiptera: Aleyrodidae) biotypes in Italy with the description of a new biotype (T) from 722 Euphorbia characias. Bull Entomol Res 93: 259-264 723 Slocombe SP, Schauvinhold I, McQuinn RP, Besser K, Welsby NA, Harper A, Aziz N, Li Y, Larson 724 TR, Giovannoni J, Dixon RA, Broun P (2008) Transcriptomic and reverse genetic analyses of 725 branched-chain fatty acid and acyl sugar production in Solanum pennellii and Nicotiana 726 benthamiana. Plant Physiol 148: 1830-1846 727 Song Z, Li S, Chen X, Liu L, Z S (2006) Synthesis of insecticidal sucrose esters. Forestry Studies in 728 China 8: 26-29 729 Stamatakis A (2014) RAxML version 8: a tool for phylogenetic analysis and post-analysis of large 730 phylogenies. Bioinformatics 30: 1312-1313 731 Stone WD, Pellicore MJ, Hagstrom S, Lawton TJ (2020) HSIviewer: Pushbutton hyperspectral image 732 analysis for rapid plant phenotyping. in review 733 Thompson JD, Higgins DG, Gibson TJ (1994) Clustal-W - Improving the sensitivity of progressive 734 multiple sequence alignment through sequence weighting, position-specific gap penalties and 735 weight matrix choice. Nucleic Acids Research 22: 4673-4680 736 Thurston R (1961) Resistance in Nicotiana to the green peach aphids and some other tobacco insect pest. 737 Journal of Economic Entomology 54: 946-949

25 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

738 Tzin V, Yang X, Jing X, Zhang K, Jander G, Douglas AE (2015) RNA interference against gut 739 osmoregulatory genes in phloem-feeding insects. J Insect Physiol 79: 105-112 740 Van Eck J, Keen P, Tjahjadi M (2019) Agrobacterium tumefaciens-mediated transformation of tomato. 741 Methods Mol Biol 1864: 225-234 742 Van T, Weinhold A, Ullah C, Dressel S, Schoettner M, Gase K, Gaquerel E, Xu SQ, Baldwin IT 743 (2017) O-acyl sugars protect a wild tobacco from both native fungal pathogens and a specialist 744 herbivore. Plant Physiology 174: 370-386 745 Wagner GJ, Wang E, Shepherd RW (2004) New approaches for studying and exploiting an old 746 protuberance, the plant trichome. Annals of Botany 93: 3-11 747 Weinhold A, Baldwin IT (2011) Trichome-derived O-acyl sugars are a first meal for caterpillars that 748 tags them for predation. Proceedings of the National Academy of Sciences of the United States of 749 America 108: 7855-7859 750 Worrall EA, Bravo-Cazar A, Nilon AT, Fletcher SJ, Robinson KE, Carr JP, Mitter N (2019) 751 Exogenous application of RNAii-inducing double-stranded rna inhibits aphid-mediated 752 transmission of a plant virus. Front Plant Sci 10: 265 753

26 NattASAT3_OIT27081.1

CanASAT3-1_RNaky_ACO48259.1

NtabASAT3_like_XP_016467639.1

NattASAT3_OIT29898.1

NtabASAT3_like_XP_016440226.1

NtabASAT3_like_XP_016489441.1

NtabASAT3_like_XP_016453912.1

NtomASAT3_like_XP_009603396.1|NtomASAT3_like_XP_033512582.1

NtabASAT3_like_XP_016464881.1

NtomASAT3_like_XP_033515849.1

NtabASAT3_like_XP_016474114.1

NtomASAT3_like_XP_009604796.1

NtabASAT3_like_XP_016494255.1

NtabASAT3_like_NP_001313112.1

NtomASAT3_like_XP_009631587.1

NattASAT3_OIT03546.1

NtabASAT3_like_XP_016462941.1

NattASAT3_OIT20175.1

NtabASAT3_like_XP_016490033.1|NtomASAT3_like_XP_009616124.1

NtabASAT3_like_XP_016453041.1

NattASAT3_OIS95899.1

NtabASAT3_like_XP_016489747.1

NtabASAT3_like_XP_016486234.1|NtomASAT3_like_xp_009630197.1

NattASAT3_OIS99435.1

NtabASAT3_like_XP_016447325.1

NtomASAT3_like_XP_009590647.1

NtabASAT3_like_XP_016447861.1

SsinASAT5_ART34014.1

HnigASAT5_c40105_g1_i1

HnigASAT4_c56915_g1_i1

SlycASAT4_M82_AFM77971.1

NattASAT3_OIT06250.1

NattASAT3_OIT33147.1

SneoASAT2_LA2133_ALU64008.1

SarcASAT2_LA2172_ALU64013.1

SlycASAT2_M82_ALU64015.1

SpimASAT2_LA1578_ALU64006.1

SgalASAT2_LA1401_ALU64010.1

SchiASAT2_LA1969_ALU64012.1

Scor_ASAT2_LA0107_ALU64011.1

SperASAT2_LA1278_ALU64007.1

SpenASAT2_LA1649_AUG68772.1|SpenASAT2_LA1911_AUG68773.1

ShabASAT2_1_LA2098_AUG68751.1|SpenASAT2_LA1367_AUG68770.1

HnigASAT3_c61998_g1_i2

SsinASAT3_ART34013.1

PaxiASAT3_AOR06333.1

NattASAT1_A4A49_42350(NattASAT3_OIT19390.1) bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 InternationalNtabASAT3_like_XP_016505083.1 license. NtomASAT3_like_XP_018631230.1

NattASAT3_OIT04893.1

NattASAT3_OIT39828.1

NtabASAT3_like_XP_016477109.1

SsinASAT1 SsinASAT1_ART34011.1 NbenASAT1 NbenASAT1 100 52 61 NtomASAT3-like NtomASAT3_like_XP_033514428.1 NattASAT3-like NattASAT3_OIT05194.1 HnigASAT1 HnigASAT1_c58659_g1_i1 63 PaxiASAT1 PaxiASAT1_AOR06331.1 PaxiASAT4 PaxiASAT4_AOR06334.1 SsinASAT2 SsinASAT2_ART34012.1 97 NbenASAT2 NbenASAT2 87 NattASAT2 NattASAT2_A4A49_04553(NattASAT3_OIT36748.1) 100 96 99 NattASAT3-like NattASAT3_OIT37557.1 NtomASAT3-like NtomASAT3_like_XP_009588593.1 99 NtabASAT3-like NtabASAT3_like_XP_016451255.1 53 HnigASAT2 HnigASAT2_ARR28781.1 58 PaxiASAT2 PaxiASAT2_AOR06332.1 SnigASAT1 SnigASAT1_ART34015.1 100 100 SlycASAT1 SlycASAT1_M82_NP_001316332.1 SpenASAT1 SpenASAT1_ALU64004.1 SgalASAT3_LA1401_AUG68801.1 Figure 1. ASAT phylogenetic tree SneoASAT3_LA2133_AUG68797.1 The evolutionary history of ASATs in the Solanaceae (Table S1) wasSarcASAT3_2_LA2172_AUG68799.1 inferred using the Maximum Likelihood method using RAxML. Presented is a subtree of a larger SpimASAT3_2_LA1578_AUG68796.1tree that includes all annotated ASATs (Figure S1). The branch labels indicate the percentage of trees in which the associated taxa clustered SpimASAT3_1_LA1578_AUG68795.1 together (bootstrap of 1000). Only values greater than 50 are presented. The two predicted N. benthamiana ASATs are highlighted in orange and ASATs that were previously chemicallySchiASAT3_LA1969_AUG68802.1 characterized are highlighted in blue. SperASAT3_2_LA1278_AUG68804.1 SperASAT3_1_LA1278_AUG68803.1

SarcASAT3_1_LA2172_AUG68798.1

ScorASAT3_2_LA0107_AUG68805.1

ScorASAT3_3_LA0107_AUG68807.1

ScorASAT3_1_LA0107_AUG68806.1

ShabASAT3_P_LA2650_AJF98595.1|ShabASAT3_P_LA2722_AJF98594.1

SpenASAT3_LA0716_AJF98583.1

ShabASAT3_F_LA2156_AJF98586.1

SlycASAT3_M82_AJF98582.1

NtabASAT3_like_XP_016500728.1

NtabASAT3_like_XP_016506125.1

NtabASAT3_like_XP_016514860.1

NattASAT3_OIT35573.1

NtomASAT3_like_XP_033507977.1

NattASAT3_OIS97679.1

NtabASAT3_like_XP_016495009.1

0.4 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A gRNA3 gRNA2 gRNA1 1 268 598 TGTTTACTTGGACAGTGACT 712 TGAGCCCATCTTAATGAGT 1299 l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l ASAT1 5’ GTTCATGTCGAGTGCAACGATATAGG CCCACAAATGAACCTGTCACTGAGCC CCAACTCGGGTAGAATTACTCACAA 3’ asat1-1 5’ GTTCATGTCGAGTGC ------T ATAGG CCCACAAAATGAACCTGTCACTGAGCC CCAACTCGGGTAGAATTACTCACAA 3’ 287 605 asat1-2 5’ GTTCATGTCGAGTGCAACGA ------AATGAACCTGTCACTGAGCC CCAACTCGGGTAGAATTACTCACAA 3’

B gRNA3 gRNA1 gRNA2

1 464 TGTTCCAACGACTACGAAT 601 GATCACCGACTAGGATCAG 700 TGTTGACACCTTCGACAGT 1257 l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l ASAT2 5’ CCCACAAGGTTGCTGATGCTTACAGT CCACTAGTGGCTGATCCTAGTCCAAA CCTACAACTGTGGAAGCTGTCACTGC 3’

3’ asat2-1 5’ CCCACA -GGTTGCTGATGCTTACAGT CCACTAGTGGCTGATCCTAGTCCAAA CCTACAAACTGTGGAAGCTGTCACTGC 414 529 asat2-2 5’ ------CCACTAGTGGCTGATCCTAGTCCAAA CCTACATACTGTGGAAGCTGTCACTGC 3’

Figure 2. N. benthamiana ASAT mutations produced with CRISPR/Cas9 Three gRNAs (sequences shown in red) were designed to edit either ASAT1 or ASAT2. For both ASAT1 and ASAT2, we obtained two distinct mutations resulting from the corresponding gRNA2 and gRNA3. Single-base mutations and deletions are shown in cyan. A) asat1-1 has a five-nucleotide deletion at gRNA3 and a single-nucleotide insertion at gRNA2. asat1-2 has a 318-nucleotide deletion between the gRNA3 and gRNA2 cutting sites. B) asat2-1 has a single nucleotide deletion at gRNA3 and single nucleotide insertion at gRNA2 leading to a translation frame shift between the two mutations. asat2-2 has a 115-nucleotide deletion at gRNA3 and a single-nucleotide insertion leading to a translation frame shift at gRNA2. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

2000 a Plant wt 1750 asat1-1 asat1-2 1500 asat2-1 asat2-2 abundance 1250 ab b

sugar 1000

acyl 750

500

250 c c Normalized a ab a ab b a b c c b b c c 0 c c S2:17(2,7,8) S2:17(2,8,8) S2:15(7,8) S2:16(8,8) Acylsugar

Figure 3. The abundances of two N. benthamiana acylsugars (S2:17(2,7,8) and S2:18(2,8,8)), and two likely pathway intermediates/fragmentations (S2:15(7,8) and S2:16(8,8)) Acyl sugar LC/MS peak areas were normalized relative to the peak area of Telmisartan, which was added as an internal control, and then to the leaf dry weight (per gram). Error bars represent standard errors from measurements of three plants of each genotype. Significant differences for each acylsugar between different genotypes were tested using one-way ANOVA followed by a Duncan’s post hoc test (p < 0.05). Differences between groups are denoted with letters. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 100 B 2.2 d bc ns a b cd 2.0 *** 1.8 80 ns )

2 1.6

Treatment m 1.4 m

60 ( wt 1.2 asat1-1 1.0 40 asat1-2 0.8 0.6 Aphid size Aphid

Percentage of aphids of Percentage asat2-1 20 0.4 asat2-2 0.2 0 0.0 0 24 48 72 96 120 wt asat1-1 asat1-2 asat2-1 asat2-2 Aphid feeding time (hour) Plant genotype C D 250 b b 70 b a a a 60 200 a a 50

150 40

30 100

20

50 aphids of Percentage 10 NO. of Nymphs/5 aphids/Week Nymphs/5 of NO. 0 0 wt asat1-1 asat1-2 asat2-1 asat2-2 wt asat1 asat2 Plant genotype Plant genotype

Figure 4. Myzus persicae bioassays with wildtype, asat1, and asat2 N. benthamiana (A) Aphid survivorship over five days, ns: not significant, *** p < 0.001, t-test. (B) Aphid growth, as measured by = body sizes after 120 hours feeding on N. benthamiana. (C) Aphid reproduction, as measured by the number of nymphs that were reproduced by five aphids in one week. (D) Aphid choice among detached leaves of each plant genotype. Significant differences between different groups (p < 0.05) were determined using ANOVA followed by a Duncan’s post hoc test and are indicated by lowercase letters above each group in panels B, C, and D. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A Bemisia tabaci B Heliothis virescens C Trichoplusia ni

100% 25 30 ) *** ns ) ns s.e. s.e. 25

s.e. 80% -

20 - - 20 60% 15 15 40% 10 10 20% 5 5 Larval Larval mass(mg+/ Larval Larval mass(mg+/ Percent survival Percent survival +/ 0% 0 0 WT asat2-1 WT asat2-1 WT asat2-1 percent percent 33% 83% 67% 33% survival survival

Figure 5. Bioassays of whiteflies and caterpillars on wildtype and asat2-1 mutant N. benthamiana (A) Whitefly survivorships after 3 days, *** p < 0.001, t-test. (B) Larval mass of surviving Heliothis virescens after 10 days of feeding. (C) Larval mass of surviving Trichoplusia ni after 10 days of feeding. ns: no significant difference (P > 0.05, t-test). bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 45 ** 40 * 35 ns ns 30 25 20 15 10 5 Percent of water loss water of Percent 0 wt asat1-1 asat1-2 asat2-1 asat2-2

B 0.98

*** *** 0.97 *** *

0.96 Water band index band Water

0.95 wt asat1-1 asat1-2 asat2-1 asat2-2

C 21.8 *** *** 21.6 * * 21.4

21.2

21

20.8 Leaf temperatureLeaf

20.6

20.4 wt asat1-1 asat1-2 asat2-1 asat2-2

Figure 6. Water loss and leaf temperature of wildtype, asat1, and asat2 N. benthamiana (A) Percent of water loss from detached leaves in 24 hours (n=15). (B) Leaf water content measured by the water band index from hyperspectral imaging (n=20). (C) Leaf temperatures from leaves of different plant genotypes (n=20). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, Dunnett’s test relative to wildtype control. Error bars = standard error. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

NattASAT3_OIT27081.1

CanASAT3-1_RNaky_ACO48259.1 100 NtabASAT3_like_XP_016467639.1 Figure S1. Phylogenetic analyses of annotated NattASAT3_OIT29898.1 97 NtabASAT3_like_XP_016440226.1 26 ASATs from Solanaceae species NtabASAT3_like_XP_016489441.1 6950 1490 NtabASAT3_like_XP_016453912.1 The ASAT proteins used in constructing the tree NtomASAT3_like_XP_009603396.1|NtomASAT3_like_XP_033512582.1 60 NtabASAT3_like_XP_016464881.1 are listed in Table S1. The evolutionary history of NtomASAT3_like_XP_033515849.1 55 100 88 NtabASAT3_like_XP_016474114.1 ASATs from Solanaceae species was inferred by NtomASAT3_like_XP_009604796.1

NtabASAT3_like_XP_016494255.1 74 using the Maximum likelihood method using 100 NtabASAT3_like_NP_001313112.1 100 NtomASAT3_like_XP_009631587.1 RAxML. The branch labels indicated the 100 NattASAT3_OIT03546.1 66 NtabASAT3_like_XP_016462941.1 percentage of trees in which the associated taxa 100 NattASAT3_OIT20175.1 95 NtabASAT3_like_XP_016490033.1|NtomASAT3_like_XP_009616124.1 clustered together (bootstrap of 1000). 100 NtabASAT3_like_XP_016453041.1

100 NattASAT3_OIS95899.1 NtabASAT3_like_XP_016489747.1 66 100 NtabASAT3_like_XP_016486234.1|NtomASAT3_like_xp_009630197.1 100 86 NattASAT3_OIS99435.1 NtabASAT3_like_XP_016447325.1

100 NtomASAT3_like_XP_009590647.1

NtabASAT3_like_XP_016447861.1

100 SsinASAT5_ART34014.1

98 HnigASAT5_c40105_g1_i1

100 HnigASAT4_c56915_g1_i1 73 SlycASAT4_M82_AFM77971.1

58 NattASAT3_OIT06250.1

NattASAT3_OIT33147.1

98 SneoASAT2_LA2133_ALU64008.1 75 99 SarcASAT2_LA2172_ALU64013.1 SlycASAT2_M82_ALU64015.1 99 8338 SpimASAT2_LA1578_ALU64006.1 SgalASAT2_LA1401_ALU64010.1

SchiASAT2_LA1969_ALU64012.1 78 100 100 Scor_ASAT2_LA0107_ALU64011.1

SperASAT2_LA1278_ALU64007.1

55 SpenASAT2_LA1649_AUG68772.1|SpenASAT2_LA1911_AUG68773.1 100 ShabASAT2_1_LA2098_AUG68751.1|SpenASAT2_LA1367_AUG68770.1 58 HnigASAT3_c61998_g1_i2

97 SsinASAT3_ART34013.1 49 PaxiASAT3_AOR06333.1

56 NattASAT1_A4A49_42350(NattASAT3_OIT19390.1) 99 100 NtabASAT3_like_XP_016505083.1

100 NtomASAT3_like_XP_018631230.1

100 NattASAT3_OIT04893.1

89 NattASAT3_OIT39828.1

NtabASAT3_like_XP_016477109.1

SsinASAT1_ART34011.1

100 52 NbenASAT1 61 NtomASAT3_like_XP_033514428.1

50 29 NattASAT3_OIT05194.1

63 HnigASAT1_c58659_g1_i1 100 86 PaxiASAT1_AOR06331.1

PaxiASAT4_AOR06334.1

SsinASAT2_ART34012.1 97 NbenASAT2 87 100 96 NattASAT2_A4A49_04553(NattASAT3_OIT36748.1) 99 NattASAT3_OIT37557.1

99 NtomASAT3_like_XP_009588593.1 53 NtabASAT3_like_XP_016451255.1

58 HnigASAT2_ARR28781.1

43 PaxiASAT2_AOR06332.1 SnigASAT1_ART34015.1 100 100 SlycASAT1_M82_NP_001316332.1 25 SpenASAT1_ALU64004.1

SgalASAT3_LA1401_AUG68801.1

96 SneoASAT3_LA2133_AUG68797.1 SarcASAT3_2_LA2172_AUG68799.1

SpimASAT3_2_LA1578_AUG68796.1 26 SpimASAT3_1_LA1578_AUG68795.1 41 100 3450 SchiASAT3_LA1969_AUG68802.1 11 SperASAT3_2_LA1278_AUG68804.1 70 SperASAT3_1_LA1278_AUG68803.1

SarcASAT3_1_LA2172_AUG68798.1 56 ScorASAT3_2_LA0107_AUG68805.1 100 34 2067 ScorASAT3_3_LA0107_AUG68807.1

100 ScorASAT3_1_LA0107_AUG68806.1

93 ShabASAT3_P_LA2650_AJF98595.1|ShabASAT3_P_LA2722_AJF98594.1

92 SpenASAT3_LA0716_AJF98583.1

ShabASAT3_F_LA2156_AJF98586.1

SlycASAT3_M82_AJF98582.1

NtabASAT3_like_XP_016500728.1

97 NtabASAT3_like_XP_016506125.1 75 82 NtabASAT3_like_XP_016514860.1 84 NattASAT3_OIT35573.1

NtomASAT3_like_XP_033507977.1

NattASAT3_OIS97679.1

NtabASAT3_like_XP_016495009.1

0.4 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A Parent ion m/z MS2 fragment/neutral loss peaks m/z rt* Annotation 681.34 635.33, 509.22, 491.21, 383.12, 179.06, 143.11, 126.10, 59.01, 18.01 1.76 S3:18(2,8,8) Formate Adduct 671.30 681.34, 671.30, 635.33, 509.22, 143.11, 126.10 1.76 S3:18(2,8,8) Chloride Adduct 635.32 509.22, 491.21, 467.21, 383.12, 365.11, 179.06, 143.11, 126.10, 18.01 1.76 S3:18(2,8,8) 639.32 593.32, 467.21, 449.20, 323.10, 143.11, 126.10 1.70 S2:16(8,8) Intermediate Formate Adduct 593.32 467.21, 323.10, 143.11, 126.10 1.70 S2:16(8,8) Intermediate 667.32 621.31, 509.22, 495.03, 383.12, 179.06, 143.11, 129.09, 126.10, 112.09 1.70 S3:17(2,7,8) Formate Adduct 621.31 495.03, 383.12, 365.11, 143.11, 129.09, 126.10, 59.01, 18.01 1.70 S3:17(2,7,8) 625.31 579.30, 467.21, 453.20, 323.10, 143.11, 129.09, 126.10 1.64 S2:15(7,8) Intermediate Formate Adduct 555.23 509.22, 383.12, 143.11, 126.10, 59.01 1.76 S2:10(2,8) In-Source Fragment Formate Adduct 509.22 449.20, 383.12, 143.11, 126.10, 59.01 1.76 S2:10(2,8) In-Source Fragment 467.21 341.11, 323.10, 144.11, 143.11 1.70 S1:8(8) In-Source Fragment 383.12 383.12, 341.11, 59.01 1.76 S1:2(2) In-Source Fragment * rt: retention time of the predominant peak for each mass.

B

(C7+H2O)

(C8+H2O) (S1:2(2)) S3:17(2,7,8)

(C2+H2O)

(S1:2(2)-H2O) (S2:9(2,7))

(C8+H O) 2 (S2:10(2,8)) (S1:2(2))

Relativeabundance S3:18(2,8,8)

(S1:2(2)-H O) 2 (S3:18(2,8,8))

Figure S2 MS/MS chromatography of identified acylsugars (A) Acyl sugars LC/MS peaks identification and annotation from N. benthamiana (B) MS2 chromatography of the two predominant acyl sugars, S2:17(2,7,8) and S2:18(2,8,8). The name of the acylsugars are highlighted in blue and the chromatography is aligned on the x axis based on m/z. Highlighted in orange boxes are some of the abundant characteristic mass features and their compositions in parenthesis. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

wt asat1-1 asat1-2 asat2-1 asat2-2 Front Back

asat2-1 wt asat1-1 asat2-1 asat1-2 wt

Figure S3. ASAT mutations do not affect trichome morphology and abundance The white arrow indicates the small trichomes and yellow arrow indicates the large swollen- stalk trichomes. Trichome pictures were taken from both the front and back sides of the leaf. Examples of leaves are put side-by-side from different genotypes to facilitate comparisons. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 2.4 c 100 B c d b ns 2.2 a 2.0 80 *** ) 1.8 2

Treatment ns m 1.6 m

60 ( wt 1.4 1.2 asat1-1 40 1.0 asat1-2 0.8

Aphid size Aphid 0.6

Percentage of aphids of Percentage asat2-1 20 0.4 asat2-2 0.2 0 0.0 0 24 48 72 96 120 wt asat1-1 asat1-2 asat2-1 asat2-2 Aphid feeding time (hour) Plant genotype C D 250 b b 70 b

a 60 200 a 50 a a 150 40 a

30 100

20

50 aphids of Percentage 10 NO. of Nymphs/5 aphids/Week Nymphs/5 of NO. 0 0 wt asat1-1 asat1-2 asat2-1 asat2-2 wt asat1 asat2 Plant genotype Plant genotype

Figure S4. Independent repeat of aphid bioassays shown Figure 4 (A) Aphid survivorship, ns: not significant, *** p<0.001. (B) Aphid growth, as measured by aphid body sizes after 120 hours feeding on N. benthamiana. (C) Aphid reproduction was measured by the number of nymphs that were reproduced by five aphids in a week. (D) Aphid choice among detached leaves of each plant genotype. Significant differences between different groups (p < 0.05) were determined using ANOVA followed by a Duncan post hoc test and are indicated by lowercase letters above each group in panels B, C, and D. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

100 ns 100 ns 100 *** 100 *** 90 90 90 90 80 80 80 80 70 70 70 70 60 60 60 60 50 50 50 50 40 40 40 40 30 30 30 30 20 20 20 20 10 10 10 10 0 0 0 0 wt asat1-1 wt asat1-2 wt asat2-1 wt asat2-2

100 100 100 100 90 *** 90 *** 90 *** 90 *** 80 80 80 80 Percentage of aphids of Percentage 70 70 70 70 60 60 60 60 50 50 50 50 40 40 40 40 30 30 30 30 20 20 20 20 10 10 10 10 0 0 0 0 asat1-1 asat2-1 asat1-1 asat2-2 asat1-2 asat2-1 asat1-2 asat2-2 Plant genotype (detached leaves)

Figure S5. Pairwise aphid choice assay with wildtype and ASAT mutant leaves Each experiment included detached leaves from 5-8 plants of each genotype and was repeated twice with similar results. ***p < 0.001, ns: not significant. Significance between different groups were tested using independent samples t-tests, error bars = standard deviation. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A Without Aphids B With Aphids Aphid bioassay Aphid

asat2-1 wildtype asat2-1 wildtype asat2-1 C 3d D 23d E 23d F 23d

Whitefly bioassay Whitefly wt asat2-1 asat2-1 asat2-1

Figure S6. Aphid and whitefly bioassays on wildtype and asat2-1 mutant N. benthamiana (A) without aphid challenge. (B) with aphid challenge. One wildtype (right) and one asat2-1 (left) plant were placed in the same tray in an insect cage. For aphid challenge, ten adult M. persicae (from an aphid colony on N. tabacum) were released on both the wildtype and the asat2-1 plants. The picture was taken about one month after the aphid release. Plants in three replications of this experiment looked similar. (C) Whiteflies died within three days after release on wildtype N. benthamiana. Arrow indicates dead whiteflies. (D-F) Whiteflies survived and completed their life cycle on asat2-1 mutant N. benthamiana. At 23 days after release, whiteflies of different life stages were discovered on asat2-1 mutant plants. Arrows indicate eggs (D), eggshells and nymphs (E), and adults (F). bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 35 *** 30 *** 25 ns ns 20 15 10 5 Percent of water loss water of Percent 0 wt asat1-1 asat1-2 asat2-1 asat2-2 B 1.01 1 *** * *** *** 0.99 0.98 0.97

Water band index band Water 0.96 0.95 wt asat1-1 asat1-2 asat2-1 asat2-2 C 20 *** ** *** 19.5 ns

19

18.5

Leaf temperatureLeaf 18

17.5 wt asat1-1 asat1-2 asat2-1 asat2-2 Plant genotype

Figure S7. Independent repeat of water loss and leaf temperature shown in Figure 6. (A) Percent of water loss from detached leaves in 24 hours (n=16). (B) Leaf water content measured by the water band index from hyperspectral imaging (n=10). (C) Leaf temperatures from leaves of different plant genotypes (n=10 for wildtype and n=5 for mutants). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, Dunnett’s test relative to wildtype control. Error bars = standard error. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A B O 2

M M ddH

bp ASAT1 500

200

C D E

M M M bp 1500 1000 ASAT2

500

Figure S8. ASAT mutagenesis by transiently expressing Cas9 and gRNAs in N. benthamiana leaves via agrobacterium infiltration Three gRNAs each were used for both ASAT1 and ASAT2 mutagenesis. (A) PCR of plants that infiltrated with combinations of ASAT1 gRNA constructs. (B) DNA sequence confirmation of a representative sample (i.e. g23OD3) that displayed a shorter band in panel A. (C) PCR of plants that were infiltrated with combinations of ASAT2 gRNA constructs. (D) Restriction enzyme digestion site loss in PCR of g23OD1.5 (from panel C), where, prior to the PCR reaction, genomic DNA was digested with KpnI to remove wild type ASAT2. The KpnI restriction site is only present in the expected deletion region. (E) Sequence confirmation of sample g23+KpnI from digestion site loss PCR in panel D. Arrow heads indicate the short versions of target genes after a deletion by the two working gRNA constructs. Arrows indicate the positions expected to be cut in gene sequences. The label for each lane in the gel: M, 100 bp DNA marker ladder; WT, samples from wild type plants; ddH2O, no genomic DNA PCR control. bioRxiv preprint doi: https://doi.org/10.1101/2020.08.04.237180; this version posted August 5, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

release point

Figure S9. Experimental setups for Myzus persicae bioassays (A) Aphid no-choice assays. Myzus persicae were caged on individual Nicotiana benthamiana leaves. Survival, reproduction, and size were assessed at time points described in the methods section. (B) Aphid choice assays. Two N. benthamiana leaves were placed in a 15 cm diameter Petri dish with their petioles inserted in moistened cotton swabs, ten M. persicae were released into the Petri dish at the indicated position, the dish was covered and placed in complete darkness, and after 24 hours the aphids residing on each leaf were counted