The plasma membrane ABC transporter Sl-ABCB5 mediates acylsugar secretion from tomato trichomes to confer insect resistance

Abstract

Acylsugars are structurally diverse specialized metabolites secreted by glandular trichomes of Solanaceae plants and play a crucial role in defense against herbivores and pathogens. However, the mechanism underlying acylsugar secretion remains unclear. This study identifies and characterizes the plasma membrane ATP-binding cassette (ABC) transporter Sl-ABCB5 as a key component mediating acylsugar transport in trichome secretory cells of cultivated tomato (Solanum lycopersicum). Using CRISPR–Cas9-generated Sl-ABCB5 knockout mutants, we observed a dramatic reduction in trichome acylsugar accumulation, accompanied by increased susceptibility to western flower thrips (Frankliniella occidentalis). In vitro transport assays using membrane vesicles and protoplasts confirmed the ability of Sl-ABCB5 to transport acylsugars. Furthermore, virus-induced gene silencing and CRISPR-mediated knockout of ABCB5 orthologs in wild tomato (Solanumpennellii) and black nightshade (Solanum nigrum) demonstrated functional conservation of this transporter across Solanaceae species. Together, these findings reveal a conserved acylsugar transport mechanism and establish ABCB5 as a promising target for engineering enhanced insect resistance in Solanaceae crops.

Acylsugars secreted by glandular trichomes are key defensive compounds in tomato, yet the mechanism of their export from secretory cells has remained elusive. This study identifies the plasma membrane ABC transporter Sl-ABCB5 as the essential acylsugar exporter in cultivated tomato. Loss of Sl-ABCB5 function leads to markedly reduced surface acylsugars and increased susceptibility to thrips. Functional analyses in Solanum pennellii and Solanum nigrum further reveal that ABCB5-mediated acylsugar transport is evolutionarily conserved, highlighting its potential for engineering insect resistance in Solanaceae crops.

Introduction

Plants orchestrate a dynamic chemical dialogue with their environment by synthesizing a structurally diverse repertoire of specialized metabolites that mediate interactions with pathogens, herbivores, and pollinators (Huang and Dudareva, 2023; Xu and Gaquerel, 2025). Intermediates and final bioactive products of these biosynthetic pathways often require efficient transport among different organelles, cells, and tissues to reach their sites of action (Halkier and Xu, 2022; Huang and Dudareva, 2023). This intricate transport process is fundamental to plant ecological success, as specialized metabolites are typically synthesized in specific cell types (e.g., glandular trichomes and laticifers) or compartmentalized within distinct subcellular organelles (e.g., vacuoles and plastids), whereas their biological functions often occur at different sites (Onoyovwe et al., 2013; Jørgensen et al., 2015). The spatial separation of biosynthesis and bioactivity therefore necessitates sophisticated transport mechanisms, representing a mechanistically complex and underexplored area of research in plant biology (Xiao et al., 2025).

The plant epidermis serves as a protective interface with the environment and is frequently equipped with specialized structures known as trichomes. These epidermal outgrowths are generally classified into two categories (Callow et al., 2000; Fahn, 2000). Non-glandular trichomes, such as those found in the model plant Arabidopsis thaliana, primarily function as mechanical barriers against herbivores and help reduce water loss (Levin, 1973). In contrast, glandular trichomes contain specialized secretory cells that synthesize a structurally and physically diverse array of bioactive compounds (Werker, 2000; Schilmiller et al., 2008). Beyond mediating interactions with insects and microbes, metabolites secreted by glandular trichomes represent a valuable source of pharmaceuticals, flavorings, and insecticides (Schuurink and Tissier, 2020; Xu et al., 2024). Notable examples include aromatic compounds in hops and Mediterranean herbs (Wang et al., 2008), psychoactive cannabinoids in Cannabis sativa (Tanney et al., 2021), and the antimalarial drug artemisinin produced by Artemisia annua (Yin et al., 2025).

Within the Solanaceae (nightshade) family, glandular trichomes secrete acylsugars, a taxonomically distinctive suite of defensive specialized metabolites (Fan et al., 2019; Liu et al., 2019; Lou et al., 2021). Tomato, for example, possesses several distinct types of glandular trichomes, including types I, IV, VI, and VII, each associated with specialized metabolic outputs. Among these, type I/IV glandular trichomes have been identified as the primary sites of acylsugar biosynthesis and secretion. These structurally diverse compounds, composed of a sugar core acylated with various fatty acid chains, function in both direct and indirect plant defense against microbial pathogens and herbivorous insects (Weinhold and Baldwin, 2011; Luu et al., 2017). Acylsugar metabolism has been extensively characterized across multiple genera within the nightshade family, including Solanum, Physalis, Nicotiana, Petunia, and Salpiglossis (Moghe et al., 2017; Lybrand et al., 2020; Lou et al., 2021; Schenck et al., 2022; Fiesel et al., 2024). In cultivated and wild tomatoes, acylsugar biosynthesis primarily occurs in the tip cells of type I/IV glandular trichomes. This process involves several key enzymes with defined catalytic roles, including isopropylmalate synthase-like 3 (Sl-IPMS3), enoyl-coenzyme A hydratase (Sl-AECH1), acyl-coenzyme A synthetase (Sl-AACS1), four BAHD acyltransferases (Sl-ASAT1–Sl-ASAT4), and an invertase-like enzyme, Sl-ASFF1 (Fan et al., 2019; Leong et al., 2019; Mandal et al., 2020). Despite detailed elucidation of the acylsugar biosynthetic pathway, the mechanisms underlying the subsequent transport and secretion of acylsugars from trichome cells to the external plant surface remain largely unknown.

Secretion of specialized metabolites from glandular trichomes requires dedicated transport machinery and energy expenditure, particularly when these compounds accumulate at high concentrations in extracellular spaces against concentration gradients. ATP-binding cassette (ABC) transporters, which are ubiquitous across all domains of life, fulfill this role by harnessing ATP hydrolysis to translocate diverse substrates across biological membranes (Vasiliou et al., 2009; Ofori et al., 2018). In plants, these transporters mediate critical physiological processes, including detoxification, nutrient acquisition, hormone transport, and specialized metabolite secretion (Banasiak and Jasiński, 2022). Several notable examples illustrate the versatility of plant ABC transporters. In Arabidopsis, ABCB19 exports both auxin and brassinosteroids across the plasma membrane (PM), potentially integrating growth and developmental signaling pathways (Ying et al., 2024). Similarly, Arabidopsis ABCG36 mediates the export of both indole-3-butyric acid and the defense compound camalexin, thereby balancing growth and defense responses (Aryal et al., 2023). In specialized metabolism, Catharanthus roseus CrTPT2 transports anticancer alkaloid precursors into the vacuole (Yu and De Luca, 2013), whereas Petunia hybrida PhABCG1 facilitates the emission of volatile organic compounds across the plasma membrane (Adebesin et al., 2017).

ABC transporters facilitate the movement of trichome-derived specialized metabolites in several plant species. In Artemisia annua, AaPDR3 mediates PM export of the sesquiterpene β-caryophyllene from the cytosol of glandular trichomes to the extracellular space (Fu et al., 2017). Recent evidence further suggests that Cannabis sativa ABC transporters mediate the transport of cannabinoids and terpenes from the cytoplasm to the vacuole in glandular trichomes (Mi et al., 2023). Together, these examples highlight the capacity of ABC transporters to facilitate the transport of structurally diverse specialized metabolites and underscore their potential as targets for plant metabolic engineering.

In this study, we identified and characterized a trichome-specific ABC transporter, Sl-ABCB5, and demonstrated its role in acylsugar transport in cultivated tomato. Using an integrated approach combining co-expression analysis, gene knockout and silencing technologies, subcellular localization, and both in vitro and in vivo transport assays, we demonstrated that Sl-ABCB5 functions in the transport of tomato trichome acylsugars. We further established the functional conservation of this transporter across multiple Solanaceae species, suggesting an evolutionarily conserved mechanism for acylsugar transport within this family. Notably, Sl-ABCB5 knockout plants exhibited increased susceptibility to western flower thrips (WFT; Frankliniella occidentalis), demonstrating that efficient acylsugar transport to the plant surface is critical for effective insect resistance. Together, these findings advance our understanding of specialized metabolite transport mechanisms in glandular trichomes and identify Sl-ABCB5 and its homologs as promising molecular targets for enhancing insect resistance in cultivated tomato.

Results

Identification of the ABC transporter Sl-ABCB5 by co-expression analysis with acylsugar biosynthesis-related genes

Acylsugars in cultivated tomato (Solanum lycopersicum) are synthesized in the unicellular tip cells of type I/IV glandular trichomes (McDowell et al., 2011; Schilmiller et al., 2012). Consistent with this localization, previous studies have identified multiple enzymes involved in tomato acylsugar biosynthesis (Moghe et al., 2017; Fan et al., 2019), including Sl-AECH1, Sl-AACS1, Sl-IPMS3, Sl-ASFF1, and Sl-ASAT1–4, all of which are specifically expressed in type I/IV trichome tip cells (Schilmiller et al., 2012, 2015; Fan et al., 2016, 2019; Leong et al., 2019). We hypothesized that acylsugar biosynthesis and secretion are coordinated processes and that genes encoding acylsugar transporters would therefore exhibit expression patterns similar to those of biosynthesis-related genes. To test this hypothesis, we analyzed gene co-expression patterns using publicly available tomato transcriptomic datasets to identify candidate acylsugar transport genes expressed in trichomes.

Using RNA sequencing (RNA-seq) data from 25 tomato tissue samples with varying trichome densities (Moore et al., 2020), we performed hierarchical clustering analysis to identify genes co-expressed with known acylsugar biosynthesis-related genes. This analysis revealed a cluster of 153 co-expressed genes that included the established acylsugar biosynthesis enzymes Sl-ASAT1, Sl-ASAT2, Sl-ASAT3, Sl-AECH1, and Sl-AACS1 (Figure 1A). Within this cluster, we identified two ABC transporter genes: Solyc03g005860, an ABCB family transporter designated Sl-ABCB5 and Solyc04g010200, a half-size ABCG-class transporter. Notably, the expression of Sl-ABCB5 showed a strong correlation with that of core acylsugar assembly genes, particularly the BAHD acyltransferases Sl-ASAT1–Sl-ASAT3. Consistent with this observation, Sl-ABCB5 was also identified within the previously reported “gray60” co-expression module, which was derived from weighted gene co-expression network analysis and is enriched for trichome-expressed genes, including several characterized acylsugar biosynthesis genes (Kerwin et al., 2024). Based on this strong and consistent co-expression with known components of the acylsugar biosynthetic machinery, we hypothesized that Sl-ABCB5 functions as an acylsugar-associated ABC transporter in S. lycopersicum.

Figure 1.

Screening of candidate genes involved in tomato trichome acylsugar transport.

(A) Identification of putative acylsugar transporter genes by hierarchical clustering analysis using public transcriptome datasets from cultivated tomato. An ATP-binding cassette (ABC) transporter gene, Solyc03g005860 (Sl-ABCB5), was identified as being co-expressed with known acylsugar biosynthesis genes, including Sl-ASAT1, Sl-ASAT2, Sl-ASAT3, Sl-AACS1, Sl-AECH1, and Sl-IPMS3.

(B) RT–qPCR analysis of Sl-ABCB5 transcript levels in roots, stems, young leaves, old leaves, ripe fruits, green fruits, sepals, and type I trichomes. Expression levels were normalized to stem tissue using the 2^−ΔΔCt method. Sl-ABCB5 was predominantly expressed in trichome-rich tissues (stems, young leaves, and sepals), with the highest expression observed in type I trichomes. Data represent means ± SEM (n = 3).

(C) Phylogenetic analysis of 154 ABC transporters in S. lycopersicum. Trichome-enriched ABC transporter genes are highlighted in red. The complete phylogenetic tree is provided in Supplemental Figure 1.

To experimentally validate the trichome-specific expression of Sl-ABCB5 indicated by transcriptomic data (Figure 1A and Supplemental Figure 1), we performed RT–qPCR analysis across multiple tomato tissues. Sl-ABCB5 showed the highest expression in isolated type I trichomes and was also abundantly expressed in trichome-rich tissues, including stems, young leaves, and sepals (Figure 1B). This tissue-specific expression pattern is consistent with the hypothesis that Sl-ABCB5 functions in acylsugar secretion from type I/IV glandular trichomes.

Phylogenetic analysis placed Sl-ABCB5 within the ABCB subfamily, which comprises full-size transporters containing two nucleotide-binding domains and two transmembrane domains (Figure 1C and Supplemental Figure 2A). Among the 154 ABC transporters identified in the genome of cultivated tomato (Ofori et al., 2018), Sl-ABCB5 is one of ten that show significant trichome-enriched expression (Supplemental Figure 1 and Figure 1C). Although some ABCB transporters function in auxin transport in A. thaliana (Geisler et al., 2017), Sl-ABCB5 clusters within a phylogenetic clade distinct from known auxin transporters (Supplemental Figure 3), suggesting a functional role different from auxin transport. Collectively, these findings identify Sl-ABCB5 as a trichome-specific ABC transporter that is co-expressed with core acylsugar biosynthesis genes, positioning it as a strong candidate for mediating acylsugar transport in tomato trichomes.

Disruption of Sl-ABCB5 dramatically reduces acylsugar accumulation in tomato trichomes

To test the function of Sl-ABCB5 in trichome acylsugar transport, we generated sl-abcb5 knockout mutants in S. lycopersicum cv. M82 using CRISPR–Cas9-mediated genome editing. Two guide RNAs (gRNAs) targeting the eighth exon of Sl-ABCB5 were designed, resulting in two independent mutant lines: sl-abcb5_4del (carrying a 4-bp deletion) and sl-abcb5_6del (carrying a 6-bp deletion) (Figure 2A). The 4-bp deletion introduced a premature stop codon, truncating the protein in its central region, whereas the 6-bp in-frame deletion removed two amino acids (residues 171 and 172) located within a predicted transmembrane domain of Sl-ABCB5 (Figure 2A and Supplemental Figure 2).

Figure 2.

CRISPR–Cas9-mediated disruption of trichome-expressed Sl-ABCB5 reduces acylsugar accumulation in S. lycopersicum M82.

(A) Schematic representation of gRNA target sites in Sl-ABCB5. Target sites (colored) and protospacer adjacent motifs (PAMs; black underlined) are indicated. Mutations in sl-abcb5_4del (4-bp deletion) and sl-abcb5_6del (6-bp deletion) are highlighted by blue dashed lines.

(B) Representative ESI-mode LC–MS chromatograms showing acylsugar profiles in sl-abcb5 mutants. sl-abcb5_4del produced nearly undetectable acylsugar peaks compared with the M82 control, whereas sl-abcb5_6del retained some long-chain acylsugars but lacked detectable short-chain acylsugars (S3:15, S4:16, and S4:17).

(C) Quantification of total trichome acylsugars in sl-abcb5 mutants compared with M82. Data are presented as means ± SEM (n = 3). Significant differences were determined using an unpaired Student’s t-test (∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

(D) Quantification of seven major trichome acylsugars in sl-abcb5 mutants compared with M82. Data represent means ± SEM (n = 3). Different letters indicate significant differences (p < 0.05), as determined by one-way ANOVA followed by Tukey’s honestly significant difference test. LC–MS/MS spectra of the major acylsugars are provided in Supplemental Figure 4.

(E) Comparison of type Ⅰ/IV trichome density on the leaves of M82 and sl-abcb5 plants. Data are means ± SEM (n = 8). Statistical significance was assessed using an unpaired Student’s t-test, ns, not significant (p > 0..05).

To assess the impact of Sl-ABCB5 disruption on acylsugar production, we analyzed leaf trichome acylsugars from both mutant lines using liquid chromatography–mass spectrometry (LC–MS). We name acylsugars based on their sugar core (e.g., S for sucrose, G for glucose), the number of acyl chains, and the total acyl chain carbon number. For example, S4:17 (2,5,5,5) represents the most abundant S. lycopersicum trichome acylsugar and consists of a sucrose core bearing four acyl chains with a total of 17 carbons, distributed as individual chains of 2, 5, 5, and 5 carbons. In both sl-abcb5_4del and sl-abcb5_6del mutants, total acylsugar levels were significantly reduced compared with wild-type M82 plants (Figure 2B and 2C). Notably, the sl-abcb5_4del mutant exhibited a more severe phenotype, with nearly no acylsugars detected. Both short-chain acylsugars, including S3:15 (5,5,5), S4:16 (2,4,5,5), and S4:17 (2,5,5,5), and long-chain acylsugars, such as S3:20 (5,5,10), S3:22 (5,5,12), S4:22 (2,5,5,10), and S4:24 (2,5,5,12), were drastically reduced or became undetectable in the sl-abcb5_4del mutant compared with control plants (Figure 2D and Supplemental Figure 4). Importantly, knockout of Sl-ABCB5 did not affect the density of type I trichomes compared with M82 (Figure 2E), indicating that the reduction in acylsugars was not due to altered trichome development. By contrast, the sl-abcb5_6del mutant, which retains the reading frame but lacks two amino acids within a predicted transmembrane domain (Supplemental Figure 2), displayed a partial loss-of-function phenotype. In this line, short-chain acylsugars were nearly absent, whereas several long-chain acylsugars, including S3:22 (5,5,12), S4:22 (2,5,5,10), and S4:24 (2,5,5,12), remained detectable, albeit at reduced levels (Figure 2D).

To validate that the acylsugar reduction phenotype was specifically caused by Sl-ABCB5 disruption rather than off-target CRISPR mutations, we generated an additional independent set of CRISPR mutants using alternative gRNAs targeting the third and fourth exons of Sl-ABCB5 (Supplemental Figure 5A). Three additional mutant alleles, each resulting in premature translation termination, exhibited acylsugar profiles similar to that of sl-abcb5_4del, with the major acylsugars S3:15 (5,5,5), S4:17 (2,5,5,5), S3:22 (5,5,12), and S4:24 (2,5,5,12) nearly undetectable (Supplemental Figure 5B).

The substantial reduction in acylsugar content across five independent mutant alleles strongly supports the involvement of Sl-ABCB5 in acylsugar production in cultivated tomato trichomes. To further elucidate whether Sl-ABCB5 functions as an acylsugar transporter in tomato trichomes, we next analyzed its tissue-specific expression and subcellular localization.

Sl-ABCB5 fusion proteins localize to the plasma membrane of type I/IV trichome tip cells

To determine the tissue- and cell type-specific expression pattern of Sl-ABCB5, we generated stably transformed tomato lines expressing GFP under the control of a 1.8-kb promoter region upstream of the Sl-ABCB5 start codon (Figure 3A). Confocal microscopy revealed strong GFP signals specifically in the tip cells of type I/IV glandular trichomes (Figure 3A). This trichome-specific expression pattern is consistent with previous comparative transcriptomic and RT–qPCR analyses, which showed enriched Sl-ABCB5 expression in trichomes (Figure 1B and Supplemental Figure 1). Notably, the expression profile of Sl-ABCB5 closely mirrors that of known acylsugar biosynthesis-related genes (Fan et al., 2019), further supporting its involvement in acylsugar metabolism in tomato trichomes.

Figure 3.

Sl-ABCB5 is specifically expressed in type I/IV trichome tip cells and localizes to the plasma membrane in S. lycopersicum M82.

(A) Confocal images of leaf trichomes from S. lycopersicum M82 stably expressing a Sl-ABCB5 promoter-driven GFP reporter construct (proSl-ABCB5::GFP). GFP fluorescence (white arrows) is specifically detected in the tip cells of type I/IV trichomes.

(B) Subcellular localization of Sl-ABCB5-YFP in transiently transformed epidermal cells of N. benthamiana expressing a nucleus-localized RFP marker; white arrows indicate the nuclear RFP signal.

(C–E) Colocalization of Sl-ABCB5–YFP with a PM-localized RFP marker (PM-RFP) in transiently transformed N. benthamiana epidermal cells. From left to right: YFP channel (Sl-ABCB5-YFP), RFP channel (PM-RFP), and merged image.

(F–H) Confocal images of protoplasts isolated from N. benthamiana leaves transiently expressing Sl-ABCB5–YFP. From left to right: YFP channel (Sl-ABCB5-YFP), chlorophyll autofluorescence (chloroplasts), and merged image. Scale bars, 25 μm.

To experimentally determine the subcellular localization of Sl-ABCB5, we attempted to clone its coding sequence for fusion with a fluorescent protein followed by transient expression in Nicotiana benthamiana (N. benthamiana) leaves. However, repeated attempts to clone the Sl-ABCB5 coding sequence in Escherichia coli (E. coli) were unsuccessful. This difficulty was presumably due to leaky expression of the gene, which is toxic to bacterial cells—a phenomenon commonly encountered when cloning plant ABC transporters (Xu et al., 2016; Supplemental Figure 6). To circumvent this obstacle, we successfully cloned a sequence-confirmed full-length Sl-ABCB5 genomic DNA fragment in E. coli. The resulting construct, in which the genomic sequence was fused to a C-terminal yellow fluorescent protein (YFP) tag (gDNA-YFP), was transferred into Agrobacterium tumefaciens and infiltrated into N. benthamiana leaves, relying on the plant splicing machinery to correctly process the transcript (Xu et al., 2016; Supplemental Figure 6). Confocal microscopy revealed fluorescent signals that distinctly outlined cell boundaries, indicating PM localization. To distinguish between PM and vacuolar membrane localization, we performed co-localization analyses using a nucleus-localized red fluorescent protein (RFP) marker (Figure 3B) and a PM-localized RFP marker (PM-RFP; Figure 3C–3E). The observation that nuclei were surrounded by RFP fluorescent signals (Figure 3B) supported PM localization, as the vacuolar membrane does not encircle the nucleus. Consistently, merged signals of Sl-ABCB5-YFP and the PM-RFP marker further confirmed the PM localization of Sl-ABCB5. In addition, protoplasts isolated from N. benthamiana leaves expressing Sl-ABCB5 were examined for localization (Figure 3F–3H). In these cells, the chloroplast autofluorescence was surrounded by the Sl-ABCB5-YFP signal (Figure 3F–3H), providing additional evidence that Sl-ABCB5 localizes to the PM.

Collectively, these results demonstrate that Sl-ABCB5 is a PM-localized transporter specifically expressed in type I/IV trichome tip cells. This subcellular localization is consistent with previous findings for other members of the ABCB subfamily, as most functionally characterized full-size ABCB transporters in Arabidopsis localize to the PM (Blakeslee et al., 2007; Lee et al., 2008; Yu and De Luca, 2013; Ofori et al., 2018). Combined with our mutant analyses, this localization pattern strongly suggests that Sl-ABCB5 functions in the transport of acylsugars across the PM in tomato trichomes.

Functional characterization of Sl-ABCB5 transport activity

To explore the transport activity of Sl-ABCB5, we developed two complementary in vitro systems using N. benthamiana as a heterologous expression platform. Using the genomic Sl-ABCB5-YFP fusion construct described in the previous section, we transiently expressed the protein in N. benthamiana leaves. Leaves showing successful expression of Sl-ABCB5, as confirmed by YFP fluorescence, were used to prepare either membrane vesicles or isolated protoplasts for subsequent transport assays.

The membrane vesicle system was generated by isolating membrane fractions from homogenized N. benthamiana leaves. These vesicles contained Sl-ABCB5 proteins oriented in both inward- and outward-facing directions (Figure 4A). Because ABC transporters require ATP to drive the conformational changes necessary for substrate translocation (Petra et al., 2012; Ofori et al., 2018), and because the vesicular lumen lacks ATP, the addition of ATP to the external buffer allowed us to measure substrate transport from the external medium into the vesicular lumen. To assess ATP dependence and time-dependent substrate accumulation, we performed time-course assays in which vesicles were incubated with the short-chain acylsugar S4:17 (2,5,5,5) for defined time intervals. Vesicles expressing Sl-ABCB5 exhibited significant accumulation of S4:17 (2,5,5,5) over 5, 10, and 15 min compared with control vesicles expressing the fluorescent protein alone (Figure 4B). Importantly, this transport activity was abolished in the absence of ATP or in the presence of sodium orthovanadate, a classical inhibitor of ABC transporters (Figure 4B). Together, these results provide direct biochemical evidence that Sl-ABCB5 functions as an active, ATP-dependent transporter of acylsugars.

Figure 4.

Functional characterization of Sl-ABCB5 in vitro transport activity.

(A) Schematic representation of the membrane vesicle transport assay. Membrane vesicles containing ABC transporters in both inward- and outward-facing orientations were incubated with ATP and substrate. Addition of external ATP drives substrate transport from the buffer into the vesicular lumen, which is quantified by LC–MS analysis.

(B) LC–MS analysis of accumulation of the short-chain acylsugar S4:17 (2,5,5,5) in membrane vesicles expressing Sl-ABCB5 or control constructs. Membrane vesicles were prepared from N. benthamiana leaves transiently expressing Sl-ABCB5 or an empty vector (EV) and collected 5, 10, and 15 min after ATP addition and incubation at 28°C. Vesicles expressing Sl-ABCB5 without ATP or with both ATP and sodium orthovanadate served as additional controls. Data are presented as mean ± SEM (n = 4). Statistical significance between Sl-ABCB5 + ATP and other experimental groups was determined using an unpaired Student’s t-test (∗∗∗∗p < 0.0001).

(C) Schematic of the protoplast transport assay. Protoplasts maintain intact cellular structure with a native, outward-facing orientation of ABC transporters. Substrate is loaded via passive diffusion during incubation in a high-concentration substrate buffer. After transfer to fresh buffer, substrate export is monitored over time by LC–MS analysis.

(D) LC–MS analysis showing that protoplasts expressing Sl-ABCB5 export significantly higher levels of the short-chain acylsugar S4:17 (2,5,5,5) than controls expressing either EV or a transport-dead Sl-ABCB5 mutant lacking the Walker A motif (DM). Samples were collected 5, 10, and 15 min after transfer of substrate-loaded protoplasts to betaine buffer and incubation at 28°C. Data are presented as mean ± SEM (n = 4). Statistical significance between Sl-ABCB5 and other experimental groups was determined using an unpaired Student’s t-test (∗∗∗∗p < 0.0001).

As a complementary approach to characterize Sl-ABCB5 transport activity, we employed a protoplast-based transport system. In protoplasts, PM proteins maintain their physiological outward-facing configuration and have access to endogenous ATP, enabling direct measurement of substrate export. Protoplasts were loaded with S4:17 (2,5,5,5) by incubation in a high-concentration substrate buffer, allowing passive diffusion into the cells. After transferring the loaded protoplasts to a fresh buffer lacking S4:17 (2,5,5,5), exported substrates were quantified using LC–MS (Figure 4C). Our results showed that protoplasts expressing Sl-ABCB5 exported significantly higher levels of S4:17 (2,5,5,5) than controls in a time-dependent manner (Figure 4D). To rigorously distinguish whether Sl-ABCB5 functions as a transporter or acts as a regulator of endogenous N. benthamiana transporters, we generated a transport-dead mutant by disrupting the Walker A motif essential for ATP binding and hydrolysis (Figure 4D). Protoplasts expressing this Walker A mutant showed no significant difference in S4:17 export compared with controls (Figure 4D). This result supports a role for Sl-ABCB5 as a transporter rather than a regulator.

To investigate the functional impact of specific mutations observed in planta, we tested the Sl-ABCB5^del2AA variant (corresponding to the sl-abcb5_6del mutant, which lacks two amino acids in a predicted transmembrane domain; Supplemental Figure 2A) using the protoplast assay. Consistent with the partial loss-of-function phenotype observed in sl-abcb5_6del plants (Figure 2D), protoplasts expressing Sl-ABCB5^del2AA showed no significant difference in S4:17 export compared with YFP controls (Supplemental Figure 2B). This loss of transport activity indicates that the two-amino-acid deletion compromises Sl-ABCB5’s ability to transport short-chain acylsugars, reinforcing our observations of reduced acylsugar accumulation in the corresponding plant mutants.

The collective in vitro data therefore provide direct biochemical evidence that Sl-ABCB5 functions as an ATP-dependent acylsugar transporter. However, the in planta phenotypes of the sl-abcb5 knockout mutants (Figure 2B–2D and Supplemental Figure 5B) revealed a more profound impact than impaired export alone. Total surface-extractable acylsugar levels were markedly reduced or became undetectable, even though the organic extraction buffer (EB) used recovers both intra- and extracellular acylsugar pools. This outcome was unexpected, as disruption of a transporter would be predicted to primarily affect extracellular acylsugar levels, potentially leading to intracellular accumulation. The drastic reduction in total acylsugars suggests that disruption of Sl-ABCB5 affects not only secretion but also overall acylsugar biosynthesis within trichomes. We therefore hypothesized that loss of Sl-ABCB5 activity might downregulate acylsugar biosynthetic genes, creating a feedback loop between transport and biosynthesis. To test this hypothesis, we performed comparative transcriptomic analysis on approximately 1,000 handpicked type I trichomes from sl-abcb5_4del mutants, using wild-type M82 trichomes as controls (Supplemental Figure 7A). Strikingly, most acylsugar biosynthesis-related genes, including Sl-AACS1, Sl-AECH1, and four BAHD family acyltransferase genes (Sl-ASAT1–Sl-ASAT4), were significantly downregulated in the mutant (Supplemental Figure 7B and 7C). This reduction was further validated by RT–qPCR analysis (Supplemental Figure 6D). These findings suggest that functional Sl-ABCB5 not only mediates acylsugar transport but also directly or indirectly influences the expression of biosynthetic genes, revealing a feedback mechanism that coordinates acylsugar production and secretion in tomato trichomes.

Collectively, these results indicate that Sl-ABCB5 functions as a PM-localized transporter responsible for acylsugar secretion in tomato trichomes. The observation that disruption of Sl-ABCB5 not only blocks acylsugar export but also downregulates upstream biosynthetic genes demonstrates that Sl-ABCB5 serves as a key integrative node linking acylsugar production with secretion, thereby governing the overall metabolic flux of this defense pathway.

Conservation of ABCB5 sequence and function beyond cultivated tomato

Acylsugars are widely distributed across the Solanaceae family, with most extant species producing these specialized metabolites in their trichomes over more than 70 million years of evolution (Huang et al., 2023; Fiesel et al., 2024). This broad conservation prompted us to investigate whether the ABCB5 transporter is similarly conserved across Solanaceae species. We therefore analyzed ABCB5 homologs in several high-acylsugar-producing species, including Solanum pennellii, Solanum nigrum, Solanum quitoense, Hyoscyamus niger, Salpiglossis sinuata, and Petunia axillaris. These species possess well-characterized trichome transcriptomes, enabling us to explore the expression patterns and evolutionary relationships of ABCB5 homologs (Moghe et al., 2017; Hurney, 2018; Fan et al., 2019; Lou et al., 2021).

We constructed a phylogenetic tree of Sl-ABCB5 and its close homologs identified by protein sequence BLASTp analysis. Sl-ABCB5 clustered within a small, well-supported clade containing its closest homologs from each species: Sopen03g001870 (S. pennellii), SN-c70630 (S. nigrum), SQ-c35546 (S. quitoense), Hn-c70387 (H. niger), SS-c58904 (S. sinuata), and GBRU01024532 (P. axillaris) (Supplemental Figure 8). Reciprocal BLAST analyses further confirmed these genes as potential orthologs of Sl-ABCB5. Analysis of trichome transcriptomes revealed strong trichome-specific expression (enrichment ratios >10) exclusively within this clade, with particularly high enrichment observed for the Solanum orthologs (Supplemental Figure 8A). This conserved pattern of trichome-enriched expression suggests that ABCB5 and its orthologs share a common functional role in acylsugar production across Solanaceae species.

To functionally validate the role of ABCB5 orthologs in acylsugar production, we employed virus-induced gene silencing (VIGS) to knock down ABCB5 expression in two Solanum species with distinct acylsugar profiles. Specifically, we targeted Sopen03g001870 (hereafter Sp-ABCB5) in S. pennellii, which produces both acylsucroses and acylglucoses (Leong et al., 2019; Lybrand et al., 2020), and SN-c70630 (hereafter Sn-ABCB5) in S. nigrum, which produces acylglucoses and acylinositols but not acylsucroses (Lou et al., 2021). The effectiveness of the VIGS procedure was confirmed by parallel silencing of the phytoene desaturase (PDS) gene, which resulted in the characteristic albino phenotype (Figure 5A and 5D). Leaves from plants subjected to ABCB5 VIGS were then harvested for acylsugar extraction. RT–qPCR analysis confirmed reduced expression of Sp-ABCB5 and Sn-ABCB5 in VIGS-treated plants (Figure 5B and 5E, respectively). Consistently, LC–MS analysis revealed significantly lower acylsugar levels in silenced plants of both species compared with control plants (Figure 5C and 5F; Supplemental Figures S9 and S10). These results indicate that ABCB5 orthologs play a conserved functional role in acylsugar production across Solanum species that diverged from a common ancestor approximately 45 million years ago.

Figure 5.

VIGS of Sl-ABCB5 orthologs reduces acylsugar accumulation in S. pennellii and S. nigrum.

(A) Representative images of S. pennellii plants subjected to VIGS. Silencing of the phytoene desaturase (PDS) gene produced an albino phenotype that served as a visual indicator of VIGS efficiency. EV infiltration served as a negative control.

(B) RT–qPCR analysis confirming reduced expression of Sp-ABCB5 in VIGS-treated S. pennellii plants relative to controls.

(C) Total acylsugar levels were significantly reduced in Sp-ABCB5-silenced plants compared with controls. Data are presented as mean ± SEM (n = 12). Statistical significance was determined using an unpaired Student’s t-test (∗p < 0.05, ∗∗∗p < 0.001).

(D) Representative images of S. nigrum plants subjected to VIGS targeting Sn-PDS and Sn-ABCB5. EV infiltration served as a negative control.

(E) RT–qPCR analysis confirming reduced expression of Sn-ABCB5 in VIGS-treated S. nigrum plants relative to controls.

(F) Total acylsugar levels were significantly reduced in Sn-ABCB5-silenced plants compared with controls. Data are presented as mean ± SEM (n = 6). Statistical significance was determined using an unpaired Student’s t-test (∗p < 0.05, ∗∗∗p < 0.001). Scale bars, 1 cm.

To investigate the conserved function of ABCB5 and overcome the limitations of transient silencing, we generated CRISPR–Cas9 knockout mutants of its ortholog Sp-ABCB5 in the acylsugar-rich, insect-resistant species S. pennellii LA0716. Two gRNAs targeting different exons were designed (Figure 6A), yielding three independent mutant lines, each harboring 5–6 nt deletions in Sp-ABCB5 that resulted in either in-frame amino acid deletions or premature translational termination. Knock-out of Sp-ABCB5 did not reduce the density of type I trichomes compared with wild-type S. pennellii (Figure 6B). LC–MS analysis of trichome acylsugars in these mutants revealed significant reductions in total acylsugar levels, including both acylsucroses and acylglucoses (Figure 6C and 6D). Specifically, eight major acylglucoses—G3:13 (4,4,5), G3:14 (4,5,5), G3:15 (5,5,5), S3:18 (4,4,10), G3:19 (4,5,10), G3:20 (5,5,10), G3:21 (4,5,12), and G3:22 (5,5,12)—were significantly reduced in the mutants (Figure 6E and Supplemental Figure 9). Strikingly, the reduction in acylsugar content was also evident at the microscopic level, as sp-abcb5 mutants lacked the characteristic acylsugar droplets typically observed around trichome tip cells in wild-type S. pennellii LA0716 (Figure 6F).

Figure 6.

CRISPR–Cas9-mediated knockout of Sp-ABCB5 reduces acylsugar accumulation in S. pennellii.

(A) Schematic representation of CRISPR–Cas9 target sites in Sp-ABCB5. gRNA target sequences are shown in color, and PAMs are underlined in black. Multiple nucleotide deletions in the sp-abcb5#1, #3, and #4 mutant lines are highlighted in blue.

(B) Comparison of type I/IV trichome density on leaves of wild-type S. pennellii and sp-abcb5 mutant plants. Data are presented as mean ± SEM (n = 8). Statistical significance was assessed using an unpaired Student’s t-test, ns, not significant (p > 0.05).

(C) Representative ESI⁺ LC–MS chromatograms showing reduced acylsugar accumulation in sp-abcb5 mutants compared with wild-type S. pennellii. Notable reductions were observed in acylglucose peaks, including G3:13, G3:14, G3:15, G3:18, G3:19, G3:20, G3:21, and G3:22, across all three mutant lines.

(D) Significant reduction of total acylsugar levels, including acylsucroses and acylglucoses, in sp-abcb5 compared with wild-type controls. Peak areas were normalized to the internal standard and leaf dry weight. Each mutant line represents one data point. Data are presented as mean ± SEM (n = 3). Statistical significance was determined using an unpaired Student’s t-test (∗p < 0.05).

(E) Quantification of eight major acylglucoses (G3:13, G3:14, G3:15, G3:18, G3:19, G3:20, G3:21, and G3:22) in sp-abcb5 mutants and wild-type plants. Data are presented as means ± SEM (n = 3). Statistical significance was assessed using an unpaired Student’s t-test (∗p < 0.05 ). LC–MS/MS spectra of these acylglucoses are provided in Supplemental Figure 8.

(F) Microscopic analysis of trichome acylsugar accumulation. Insets (white rectangles) show enlarged views of leaf trichomes. White arrows indicate acylsugar droplets in wild-type trichome tip cells, which are absent in sp-abcb5 mutants. Scale bars, 100 μm.

Together, these findings establish ABCB5 as a conserved acylsugar transporter within the Solanum genus, capable of mediating the transport of acylsucroses, acylglucoses, and acylinositols. Moreover, the presence of trichome-enriched ABCB5 orthologs in other Solanaceae species, combined with their phylogenetic relationships, suggests that this transport mechanism may be broadly conserved across the family.

Disruption of ABCB5 compromises thrips resistance in both wild and cultivated tomato

Acylsugars are well documented for their role in conferring insect resistance in Solanaceae species. Previous studies have utilized the high-acylsugar-producing S. pennellii LA0716 to generate introgression lines with varying acylsugar levels, enabling the investigation of acylsugar-mediated insect resistance (Ben-Mahmoud et al., 2019; Mutschler et al., 2023). Our sp-abcb5 mutant, in which acylsugar secretion is disrupted, provided a unique opportunity to directly assess the contribution of acylsugars to insect resistance in S. pennellii.

To evaluate insect resistance, we selected WFT (F. occidentalis), a major tomato pest that damages leaves and fruits and acts as a vector for viral transmission (Navas et al., 1991). We conducted choice feeding assays in which detached leaves from mutant and control plants were placed side by side in Petri dishes and exposed to WFT feeding. Leaf damage was quantified by measuring the affected areas to assess relative plant resistance to WFT (Figure 7A). Wild-type S. pennellii LA0716 leaves showed no detectable damage after WFT exposure, demonstrating strong resistance to the pest (Figure 7B and 7C). In contrast, the sp-abcb5 mutant, which produces almost no detectable trichome acylsugars, suffered extensive damage following WFT feeding. Approximately 75% of the leaf area of sp-abcb5 was damaged (Figure 7B and 7C), indicating that blocking acylsugar transport converts S. pennellii from thrips resistant to thrips susceptible.

Figure 7.

Loss of ABCB5 function compromises western flower thrips (WFT) resistance in S. pennellii and S. lycopersicum.

(A) Schematic of the insect feeding assay. Detached leaves from mutant and control plants were placed in the same Petri dish to assess WFT feeding preference. Leaf damage areas were quantified to evaluate plant resistance to WFT.

(Band C) Increased WFT feeding damage in sp-abcb5 leaves. Representative images (B) and quantification of damaged leaf area (C) show that sp-abcb5 leaves exhibit significantly greater damage than wild-type S. pennellii controls following WFT feeding. Data are presented as mean ± SEM (n = 7). Statistical significance was determined using an unpaired Student’s t-test (∗∗∗p < 0.001). Scale bars, 1 cm.

(Dand E) Increased WFT feeding damage in sl-abcb5 mutant leaves. Representative images (D) and quantification of damaged leaf area (E) show that sl-abcb5 mutant leaves exhibit significantly greater damage than wild-type S. lycopersicum M82 controls following WFT feeding. Data are presented as mean ± SEM (n = 7). Statistical significance was determined using an unpaired Student’s t-test (∗∗∗p < 0.001). Scale bars, 1 cm.

(F) Proposed working model for ABCB5-mediated acylsugar secretion underlying tomato thrips resistance. In the tip cells of tomato glandular trichomes, acylsugars are synthesized through the coordinated action of acyltransferases (ASAT1–ASAT4). The PM-localized transporter Sl-ABCB5 exports these acylsugars across the cell membrane into the extracellular space. In WT plants, efficient secretion and surface accumulation of acylsugars confer resistance to WFT. In abcb5 mutants, acylsugar export is impaired, resulting in reduced surface accumulation and increased susceptibility to thrips feeding damage.

To evaluate the impact of acylsugars in an agricultural species, we assessed WFT resistance in cultivated tomato M82 and its sl-abcb5-4bp mutant. Unlike S. pennellii, M82 is a low-acylsugar producer and is naturally more susceptible to WFT feeding. WFT damage was further exacerbated in the sl-abcb5-4bp mutant, resulting in significantly greater leaf damage compared with M82 (Figure 7D and 7E). This finding highlights the importance of ABCB5 in maintaining basal levels of acylsugar-mediated insect resistance, even in cultivated tomato.

Taken together, these results establish ABCB5 as a key genetic determinant of acylsugar-mediated insect resistance in two tomato species (Figure 7F). The sp-abcb5 mutant, in particular, may serve as a valuable genetic resource for dissecting additional factors underlying tomato–insect interactions beyond acylsugars.

Discussion

Acylsugars serve as a crucial chemical defense mechanism across Solanaceae species, particularly within the Solanum genus, which exhibits remarkable diversity in acylsugar profiles (Fiesel et al., 2024). While extensive research has elucidated the biosynthetic pathways of acylsugars in cultivated tomato (Fan et al., 2019), wild tomato (S. pennellii) (Fan et al., 2017; Leong et al., 2019), and black nightshade (S. nigrum) (Lou et al., 2021), the mechanisms governing their secretion from glandular trichomes to the plant surface—where their ecological functions are executed—remain elusive. Our study addresses this knowledge gap by identifying and characterizing Sl-ABCB5, a PM-localized ABC transporter, as an essential component of the acylsugar secretion machinery in tomato trichomes. The discovery of Sl-ABCB5 provides molecular evidence for how these vital defensive metabolites are transported from their site of synthesis within glandular trichome tip cells to the external environment, thereby directly linking intracellular biochemistry with plant ecological interactions.

The identification of Sl-ABCB5 as a trichome acylsugar exporter expands the known functional repertoire of the plant ABCB subfamily. ABC transporters operate across all domains of life and participate in diverse biological processes, including detoxification, nutrient uptake, and the transport of structurally and functionally diverse substrates (Rees et al., 2009; Ofori et al., 2018). In plants, they play crucial roles in the transport of phytohormones, lipids, and a wide range of specialized metabolites (Banasiak and Jasiński, 2022). Although several members of the plant ABCB subfamily, typically full-size transporters, have been extensively characterized for their roles in auxin transport (Geisler et al., 2017), Sl-ABCB5 clusters phylogenetically apart from these known auxin transporters. This distinction, together with its specific expression in trichome tip cells, co-localization with acylsugar biosynthesis, and demonstrated transport activity, establishes a previously unrecognized role for an ABCB transporter in the secretion of glandular trichome-derived specialized metabolites.

The contrasting phenotypes of the frameshift sl-abcb5_4del mutant and the two-amino-acid-deletion sl-abcb5_6del mutant provide insights into the substrate specificity and transport mechanism of Sl-ABCB5. Whereas the frameshift mutation in sl-abcb5_4del resulted in a near-complete loss of acylsugars, the sl-abcb5_6del mutant retained detectable levels of long-chain acylsugars, while short-chain variants were almost undetectable. Transcript analysis revealed a modest reduction in Sl-ABCB5 mRNA levels in sl-abcb5_6del (Supplemental Figure 6D), suggesting that reduced transporter abundance likely contributes to the overall decrease in acylsugar accumulation observed in this line. However, reduced expression alone cannot fully explain the observed metabolic phenotype, as a simple reduction in protein abundance would be expected to proportionally decrease the transport of all acylsugar species. Instead, the pronounced disproportionate effect—where short-chain acylsugars are nearly abolished while long-chain variants remain partially preserved—indicates that the defect is also intrinsic to transporter function. Deletion of these two amino acids within the transmembrane domain likely alters the substrate-binding pocket or the conformational changes required for translocation, thereby specifically impairing the recognition or transport of shorter acyl chains. These observations suggest that, although Sl-ABCB5 exhibits substantial substrate promiscuity, specific residues within its transmembrane region are critical for distinguishing among chemically diverse acylsugars with different chain lengths. This resembles other plant transporters with broad yet selective substrate profiles, such as AtABCB19, which transports both auxins and brassinosteroids with different affinities (Ying et al., 2024).

The reduction in total acylsugar levels observed in the sl-abcb5 knockout mutants, together with the downregulation of key biosynthetic genes (Supplemental Figure 6D), suggests tight coordination between acylsugar synthesis and secretion. We interpret this transcriptional downregulation as evidence of a negative feedback regulatory mechanism. In plant specialized metabolism, intracellular accumulation of end products is known to repress upstream biosynthetic genes to prevent cellular toxicity and metabolic waste (Li et al., 2024). Therefore, the altered gene expression observed in the mutants is likely a secondary consequence of the transport defect rather than a direct role of Sl-ABCB5 in transcriptional regulation.

Furthermore, the near-complete loss of acylsugars in the mutants raised the question of whether Sl-ABCB5 might transport biosynthetic intermediates or precursors rather than the final defensive compounds. However, the specific localization of Sl-ABCB5 to the PM provides strong evidence for a role in exporting final products to the cell surface. Biosynthetic intermediates typically require transport between intracellular organelles (Linka and Weber, 2010). In contrast, PM transporters facilitate secretion into the apoplast or external environment, as exemplified by AaPDR3 in Artemisia annua (Fu et al., 2017) and PhABCG1 in Petunia hybrida (Adebesin et al., 2017). When considered together with our biochemical data demonstrating direct, ATP-dependent transport of fully assembled acylsugars, the most parsimonious conclusion is that Sl-ABCB5 primarily functions in the final export step. Although we cannot entirely exclude the possibility that Sl-ABCB5 may also accept certain biosynthetic intermediates, given the known substrate promiscuity of ABC transporters (Geisler, 2024), the collective evidence supports a model in which ABCB5 acts as the terminal gatekeeper delivering acylsugars to the trichome surface.

Our in vitro transport assays, using both membrane vesicles and protoplasts from cells heterologously expressing Sl-ABCB5, demonstrated its capacity to transport short-chain acylsugars. However, the inherently low aqueous solubility of long-chain acylsugars posed a technical challenge for their direct testing in these aqueous, buffer-based systems. This limitation suggests that, within trichome tip cells, additional mechanisms may facilitate the trafficking of these highly hydrophobic compounds from their sites of biosynthesis to the PM, where Sl-ABCB5 is localized. Such mechanisms could involve lipid carrier proteins, specialized lipid microdomains, or vesicular trafficking pathways that shield hydrophobic molecules during their transit through the aqueous cytoplasm. Future structural biology studies—such as cryo-electron microscopy analyses of Sl-ABCB5 in complex with different acylsugar substrates—together with investigations of its protein interactome, may elucidate how Sl-ABCB5 accommodates chemically diverse acylsugars and clarify the complete intracellular trafficking pathway.

The functional conservation of ABCB5 orthologs across Solanum species underscores the fundamental importance of this transport mechanism. Gene silencing of Sp-ABCB5 in S. pennellii and Sn-ABCB5 in S. nigrum, followed by CRISPR-mediated knockout of Sp-ABCB5, resulted in significant reductions in acylsugar accumulation. These results demonstrate that ABCB5 orthologs are required for the accumulation of diverse acylsugar classes—including acylsucroses, acylglucoses, and acylinositols—across Solanum species that diverged over 45 million years ago (Messeder et al., 2024). The conserved, trichome-specific expression of these orthologs suggests that ABCB5-mediated transport represents an ancient and broadly adopted mechanism, providing a crucial export step that enables diverse Solanaceae species to weaponize acylsugars for defense.

The reduced resistance to WFT observed in both sl-abcb5 (cultivated tomato) and sp-abcb5 (S. pennellii) mutants provides direct evidence for the critical role of ABCB5-mediated acylsugar secretion in plant defense. WFT is a formidable agricultural pest that causes extensive damage through feeding and virus transmission and is notoriously difficult to control (Mouden et al., 2017; Reitz et al., 2020). Knockout of Sp-ABCB5 in S. pennellii, a species known for its robust acylsugar-based defenses, dramatically reduced thrips resistance. Even in the relatively low acylsugar-producing cultivated tomato, loss of Sl-ABCB5 function exacerbated thrips damage, highlighting the contribution of this transporter to basal defense levels. The sp-abcb5 mutant, by specifically blocking acylsugar secretion, represents a valuable genetic tool for dissecting additional, potentially secondary layers of insect resistance in S. pennellii that are normally masked by the potent effects of surface-accumulated metabolites. Although these controlled laboratory experiments suggest that acylsugars are a major factor in thrips resistance in this context, plant–insect interactions in complex field environments are likely influenced by multiple interacting factors. Furthermore, because sl-abcb5 mutants affect both intracellular and secreted acylsugar pools, future studies will be required to dissect the specific ecological contributions of the surface-deposited acylsugar layer alone.

In conclusion, this study identifies Sl-ABCB5 as a PM transporter essential for efficient acylsugar secretion in tomato trichomes, elucidating a previously unknown step in the acylsugar defense pathway. Our findings demonstrate its conserved function across Solanum species and its direct importance in conferring insect resistance in tomato. Future research should focus on elucidating the molecular mechanisms underlying ABCB5 substrate recognition and transport, as well as testing a broader range of herbivorous insects affected by ABCB5-mediated defenses. Expanding investigations of ABCB5 orthologs across a wider range of Solanaceae species will further clarify how this transport system evolved and diversified within the family and how it has shaped ongoing evolutionary interactions between Solanaceae plants and their insect herbivores.

Methods

Plant materials and growth conditions

The seeds of cultivated tomato (S. lycopersicum L. M82) and wild tomato (S. pennellii LA0716) were obtained from the C.M. Rick Tomato Genetic Resource Center (https://tgrc.ucdavis.edu/). Plants were grown in a controlled growth chamber under a 16-h photoperiod at 28°C (day) and 22°C (night), with 50% relative humidity.

Acylsugar extraction and LC–MS analysis

Acylsugars were extracted from leaf surfaces as previously described (Schilmiller et al., 2010); the detailed protocol is available on Protocols.io (https://doi.org/10.17504/protocols.io.xj2fkqe). Fully expanded apical leaves from the uppermost branch of 3-week-old plants were used for trichome acylsugar extraction. Briefly, each leaf was dipped into a 1.5-ml centrifuge tube containing 1 ml of acylsugar extraction buffer (EB; acetonitrile:isopropanol:water, 3:3:2) supplemented with 0.1% formic acid and 1 μM telmisartan as an internal standard. The samples were gently agitated for 2 min, after which the extraction solvent was collected and stored in LC–MS vials at −20°C until analysis. Metabolite analysis was performed using an Agilent 1290 Infinity II ultra-high performance liquid chromatography system coupled to an Agilent 6545B quadrupole time-of-flight mass spectrometer equipped with an electrospray ionization (ESI) source operating in negative mode. The quadrupole time-of-flight parameters were as follows: gas temperature, 275°C; gas flow, 8 l/min; nebulizer pressure, 35 psig; sheath gas temperature, 300°C; sheath gas flow, 11 l/min; capillary voltage, 3500 V; nozzle voltage, 0 V; fragmentor voltage, 150 V; and skimmer voltage, 65 V. Chromatographic separation was carried out on a C18 column maintained at 40°C, with 5-μl sample injections separated using a 21-min binary gradient of water (A) and acetonitrile (B) at a flow rate of 0.3 ml/min. The gradient program was as follows: 5%–60% B (0–3 min), 60%–100% B (3–15 min), 100% B (15–18 min), 100%–5% B (18–18.1 min), and 5% B (18.1–21 min). Data were acquired in auto MS/MS mode over a mass range of m/z 50–1700 and processed using Agilent MassHunter Workstation software (v10.0).

CRISPR–Cas9 vector construction

CRISPR–Cas9 vector construction was performed as described previously (Brooks et al., 2014; Leong et al., 2019; Fan et al., 2020). Two gRNAs were designed to target one or two exons of each gene of interest to ensure efficient gene editing. The gRNAs were synthesized as gene blocks (Sunya, Hangzhou, China), and their sequences are provided in Supplemental File 1. For vector assembly, four plasmids—pICH47742::2×35S-5′ UTR-hCas9 (STOP)-NOST (Addgene, 49771), pICH41780 (Addgene, 48019), pAGM4723 (Addgene, 48015), and pICSL11024 (Addgene, 51144)—were combined with the synthesized gene blocks. Seamless DNA assembly was carried out using the Golden Gate Assembly Kit (New England Biolabs) according to the manufacturer’s instructions.

Vector construction for tissue-specific localization analysis

To investigate tissue-specific gene expression patterns, a 1.8-kb region upstream of the translational start site of Sl-ABCB5 was amplified and cloned. Primers used for amplification are listed in Supplemental File 1. The amplified fragments were fused to a GFP reporter gene and cloned into the entry vector pENTR/D-TOPO (Thermo Fisher Scientific). The resulting constructs were then recombined into the Gateway destination vector pKWGFS7 using LR Clonase II (Thermo Fisher Scientific). The final vectors were stably transformed into tomato plants to visualize tissue-specific expression patterns via GFP fluorescence.

Tomato transformation

Tomato transformation was performed using A. tumefaciens strain AGL0, following established protocols (McCormick, 1997; Leong et al., 2019). Briefly, cotyledon explants from S. lycopersicum M82 and S. pennellii were co-cultivated with A. tumefaciens harboring CRISPR–Cas9 or promoter–GFP constructs. Transformed plants were selected on kanamycin-containing medium, and putative transgenic lines were confirmed by PCR-based genotyping. Primers used for genotyping are listed in Supplemental File 1. PCR products were purified and subjected to Sanger sequencing to verify the presence of the desired mutations or insertions.

N. benthamiana transient gene expression

To analyze the subcellular localization of Sl-ABCB5, the full-length genomic DNA sequence of Sl-ABCB5 was amplified using primers listed in Supplemental File 1. The amplified fragment was cloned into the entry vector pENTR/D-TOPO (Thermo Fisher Scientific) and subsequently recombined into the destination vector pEarleyGate101, which carries a C-terminal YFP tag, using LR Clonase II (Thermo Fisher Scientific). A PM marker (PM-RFP) based on AtPIP2A was used for co-infiltration (Nelson et al., 2007). The resulting construct was transformed into A. tumefaciens strain GV3101 for transient expression in N. benthamiana.

Transient expression assays were performed as previously described (Batoko et al., 2000; Fan et al., 2020). Briefly, A. tumefaciens cultures carrying the recombinant vectors were grown overnight, harvested by centrifugation, and resuspended in infiltration buffer (20 mM acetosyringone, 50 mM 2-(N-morpholino)ethanesulfonic acid [pH 5.7], 0.5% [w/v] glucose, and 2 mM Na₃PO₄) to an OD₆₀₀ of 0.05. The bacterial suspensions were infiltrated into the abaxial side of leaves from approximately 4-week-old N. benthamiana plants using a needleless syringe. Plants were grown under controlled conditions at 21°C with an 8-h photoperiod (short-day conditions). After infiltration, plants were maintained for 3 days to allow sufficient protein expression prior to confocal microscopy analysis.

Confocal microscopy

Subcellular localization of YFP-tagged Sl-ABCB5 was visualized using a Nikon A1 laser scanning confocal microscope equipped with Nikon NIS-Elements Advanced Research software. GFP, YFP, RFP, and chlorophyll fluorescence signals were detected using excitation wavelengths of 488, 513, 561, and 488 nm, respectively, with emission windows of 505–525 nm (GFP), 521–554 nm (YFP), 600–650 nm (RFP), and 650–750 nm (chlorophyll). Microscope settings were optimized according to the manufacturer’s guidelines to ensure high-resolution imaging. Confocal images were acquired and processed using Nikon NIS-Elements Advanced Research software. Fluorescence channels were overlaid to determine the subcellular localization of Sl-ABCB5. At least three independent biological replicates were analyzed to ensure reproducibility.

Acylsugar purification for transporter assays

Acylsugars used for transporter activity assays were isolated and purified from S. lycopersicum M82 as follows. Leaves and stems from 4-week-old plants were harvested from 40 individual plants and extracted using a dipping buffer consisting of acetonitrile:isopropanol:water (3:3:2 [v/v/v]) containing 0.1% (v/v) formic acid. Plant tissues were submerged in 1000 ml of dipping buffer for 5 min with gentle agitation to ensure thorough extraction. The resulting extract was filtered through a Buchner funnel lined with filter paper to remove plant debris. The solvent was then evaporated to dryness under vacuum using a rotary evaporator, and the residue was re-dissolved in 10 ml of acetonitrile:isopropanol:water (3:3:2 [v/v/v]). The acylsugar extract was diluted 1:4 with water and loaded onto a pre-equilibrated C18 solid-phase extraction cartridge. Short-chain and long-chain acylsugars were sequentially eluted using 6 ml of 45% acetonitrile and 6 ml of 90% acetonitrile, respectively. The eluted fractions were evaporated to dryness using a SpeedVac vacuum concentrator. Purified acylsugar samples were stored at −20°C for subsequent use in transport assays. The quality and relative abundance of the purified acylsugars were verified by LC–MS. Because of the low solubility of long-chain acylsugars in aqueous solutions, only short-chain acylsugars were used for transporter assays.

Quantification of purified acylsugars used for transporter assays was performed using a saponification method adapted from a previous study with minor modifications (Schilmiller et al., 2016). Briefly, 100 μl of purified acylsugar dissolved in methanol was transferred to a 1.5-ml glass vial. An equal volume (100 μl) of freshly prepared 1 M sodium hydroxide was added, and the mixture was incubated at room temperature (25°C) for 48 h to ensure complete saponification. Following incubation, 100 μl of water and 700 μl of methanol were added, resulting in a final solvent composition of 4:1 (v/v) methanol:water. The samples were incubated for an additional 24 h at room temperature to ensure complete solubilization. Samples were then diluted either 2-fold or 100-fold with 4:1 (v/v) methanol:water, and stable isotope-labeled sucrose-¹³C₁₂ was added as an internal standard at a final concentration of 0.5 mM. Quantitative analysis of sugar concentrations was performed using an ExionLC high performance liquid chromatography system coupled to an AB Sciex QTRAP 6500 Plus mass spectrometer.

Plasma membrane vesicular transport assay

For the vesicular transport assay, the full-length genomic DNA sequence of Sl-ABCB5 was cloned into the pEAQ-HT vector using primers listed in Supplemental Table 1. Protein expression was carried out in 4-week-old N. benthamiana plants via Agrobacterium-mediated transient expression following established protocols (Batoko et al., 2000; Fan et al., 2020).

Plant vesicles were prepared according to modified protocols (Casado-Vela et al., 2010; Yepes-Molina et al., 2020). Approximately 200 mg of leaf tissue was ground to a fine powder in liquid nitrogen using a pre-chilled mortar and pestle. The powdered tissue was transferred to a 1.5-ml tube containing 200 μl of EB (100 mM HEPES–KOH [pH 6.8], 25% [w/v, 0.81 M] sucrose, 5% [v/v] glycerol, 10 mM EDTA [pH 8.0], 5 mM KCl, 1 mM DTT, and protease inhibitors [cOmplete Ultra Tablets, 1 tablet per 10 ml]). The homogenate was vortexed for 30 s to ensure thorough mixing. To remove phenolic compounds, the homogenate was mixed with pre-prepared polyvinylpolypyrrolidone (PVPP) pellets. For each 200 mg of plant material, 10 mg of PVPP was used. PVPP pellets were prepared by suspending 100 mg of PVPP in 2 ml of EB (without protease inhibitors), vortexing for 3 min, and centrifuging at 5,000 g for 2 min at 20–23°C. The supernatant was discarded, and the PVPP pellets were used immediately. The homogenate–PVPP mixture was incubated for 5 min to allow adsorption of phenolic compounds, followed by centrifugation at 2,000 g for 3 min. The supernatant was transferred to a new 1.5-ml tube, and the pellet was re-extracted three times with 100 μL of EB, vortexed vigorously, and centrifuged at 2,000 g for 2 min. All supernatants were combined and centrifuged at 2,000 g for 5 min at 4°C. The clarified supernatant was aliquoted into 100-μl portions in new 1.5-ml tubes, and an equal volume of water was added to each tube to reach a final volume of 200 μl. Samples were centrifuged at 21,000 g for 2 h at 4°C. The supernatant was carefully removed, and membrane pellets were washed with 150 μl of wash buffer (10 mM HEPES–KOH [pH 6.8], 1 mM EDTA, and 1 mM PMSF). The washed pellets were recentrifuged at 21,000 g for 1 h, after which the wash buffer was discarded. Membrane pellets were stored at 4°C for immediate use.

For the PM vesicular transport assay, membrane pellets were resuspended in 150 μl of transport buffer (50 mM HEPES–KOH [pH 6.8], 3% [v/v] glycerol, and 0.1 M KCl). Protein concentrations of the membrane vesicles were determined using the Bradford assay. For each reaction, 100 μl of membrane vesicles containing 50 μg of total protein was used. Purified short-chain acylsugar was added as a substrate to a final concentration of 4 μM. Transport reactions were initiated by the addition of Mg-ATP to a final concentration of 5 mM and incubated at 28°C. In parallel, reactions lacking ATP or containing 5 mM Mg-ATP supplemented with 1 mM sodium orthovanadate were included as controls. At 5, 10, and 15 min, 50 μl aliquots of the reaction mixture were immediately loaded onto prewetted nitrocellulose filters (0.45-μm pore size; Millipore) and rapidly washed three times with 5 ml of ice-cold HEPES buffer (50 mM HEPES–KOH [pH 6.8]). Filter-bound vesicles were dissolved by adding 200 μl of acylsugar EB and transferred to 1.5-ml tubes, followed by centrifugation at 12,000 g for 10 min. The resulting supernatants were transferred to LC vials with glass inserts for LC–MS analysis. Data are presented as means ± SEM from three technical replicates.

Protoplast isolation and transport assay

For transport assays using the protoplast system, the full-length genomic DNA sequence of Sl-ABCB5 was amplified using primers listed in Supplemental File 1. Fragments corresponding to a transport-dead Sl-ABCB5 mutant, in which the Walker A motif was disrupted, were amplified using primers listed in Supplemental Table 1 and assembled into the full-length gene by homologous recombination. The amplified Sl-ABCB5 or transport-dead Sl-ABCB5 was cloned into the entry vector pENTR/D-TOPO (Thermo Fisher Scientific) and subsequently recombined into the destination vector pEarleyGate101, which harbors a C-terminal YFP tag, using LR Clonase II (Thermo Fisher Scientific). Additionally, Sl-ABCB5_6del was cloned into pENTR/D-TOPO and recombined into pEarleyGate104, which contains an N-terminal YFP tag. Protoplasts were isolated from N. benthamiana leaves as previously described (Geisler et al., 2005; Elejalde-Palmett et al., 2021). Briefly, leaves from 4-week-old N. benthamiana plants were co-infiltrated with A. tumefaciens strain GV3101 harboring the p19 silencing suppressor and the pEarleyGate101 constructs. After 5 days, infiltrated leaves were excised into 1–2 mm strips and incubated in digestion buffer (MCP medium: 500 mM D-sorbitol, 10 mM 2-(N-morpholino) ethanesulfonic acid (MES), and 1 mM CaCl₂ [pH 5.6]) supplemented with 1% (w/v) cellulase YC, 1% (w/v) pectolyase Y23, and 1% (w/v) BSA for 90 min at 30°C. The digested tissue was filtered through a 320-mesh stainless steel filter, and protoplasts were purified using MCP medium and a Percoll gradient. Isolated protoplasts were maintained on ice for subsequent experiments. Protoplast concentration was determined using a hemocytometer, and cell density was adjusted to a uniform concentration with MCP buffer to ensure consistent experimental conditions across replicates.

The transport assay was performed as previously described (Geisler et al., 2005), with minor modifications. Protoplasts were loaded with equal substrate concentrations across all treatments. Briefly, purified acylsugars were added to the protoplast suspension to a final concentration of 35 μM, and the mixture was incubated on ice for 15 min to allow substrate uptake. External acylsugars were removed by separating protoplasts using a 50%/25%/5% Percoll gradient. Protoplasts were recovered from the 25%/5% Percoll interface and diluted 1:3 with betaine buffer (400 mM betaine, 30 mM KCl, and 20 mM HEPES [pH 7.2]). Protoplasts were then incubated at 28°C to facilitate acylsugar export. Aliquots (70 μl) were collected at 0, 5, 10, and 15 min, mixed with 200 μl of acylsugar EB, and centrifuged at 12,000 × g for 10 min. The supernatants were transferred to LC vials with glass inserts for LC–MS analysis. Data are presented as means ± SEM from three technical replicates.

VIGS and RT–qPCR analysis

To investigate the function of Sl-ABCB5 homologs in S. pennellii and S. nigrum using VIGS, gene-specific fragments were amplified using primers listed in Supplemental File 1. These fragments were cloned into the pTRV2-LIC vector via homologous recombination. The PDS gene was included as a positive control for VIGS. Recombinant pTRV2-LIC vectors were individually transformed into A. tumefaciens strain GV3101. VIGS was performed as previously described (Fan et al., 2020; Leong et al., 2020). Briefly, Agrobacterium cultures harboring the recombinant vectors were infiltrated into 4-week-old plants. Approximately 4 weeks post inoculation, the fourth true leaf from each infiltrated plant was harvested and divided into two halves: one for acylsugar quantification and the other for gene expression analysis.

Total RNA was extracted using the RNAsimple Total RNA Kit (Tiangen, China). First-strand cDNA was synthesized from total RNA using TOROBlue qRT PreMix with gDNA Eraser 2.0 (Toyobo). RT–qPCR was performed using TOROGreen qRT Master Mix (Toyobo) on a LightCycler 480 II real-time PCR system (Roche, Germany). Relative gene expression levels were calculated using the 2^−ΔΔCt method.

Trichome counting and trichome collection for comparative transcriptome analysis

Fully expanded apical leaves from the three uppermost branches of 4-week-old tomato plants were sampled for trichome counting and trichome transcriptome analysis. For type I/IV trichomes, at least three fields on the adaxial surface of each leaf were imaged, and the number of type I/IV trichomes within each image was counted. Counts from four leaves per seedling were averaged to obtain one biological replicate, and eight biological replicates were analyzed. For trichome collection, type I trichomes were manually collected from stems using fine-tipped tweezers under a dissecting microscope. Type I trichomes were identified based on their characteristically long stalks, which facilitated manual collection. Approximately 1,000 trichomes were collected per replicate. For total trichome isolation, stems were snap-frozen in liquid nitrogen and gently brushed with a pre-chilled brush to remove trichomes. Total RNA was extracted from collected trichomes using the RNeasy Plant Mini Kit (QIAGEN, Germany). RNA purity and concentration were assessed using a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA), and RNA integrity was evaluated with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA libraries were constructed using the TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. Sequencing was performed on an Illumina HiSeq 2500 platform, generating 125-bp paired-end reads. Approximately 45 million raw reads were obtained per sample. Raw reads in FASTQ format were processed using Trimmomatic (Kim et al., 2015) to remove low-quality sequences, resulting in approximately 43 million clean reads per sample. Clean reads were aligned to the S. lycopersicum SL3.0 reference genome for downstream analyses. Transcriptome sequencing and data analysis were conducted by OE Biotech (Shanghai, China).

Phylogenetic analysis, heatmap construction, and gene co-expression analysis

Genes belonging to the ABC transporter family were retrieved from a previous study (Ofori et al., 2018). Homologs of Sl-ABCB5 in Solanaceae species and their trichome expression profiles were obtained from publicly available transcriptome datasets, as described previously (Guo et al., 2015; Moghe et al., 2017; Fan et al., 2020). Phylogenetic analysis was performed using RAxML v8.0.6 (Stamatakis, 2014). Amino acid sequences were aligned via multiple alignment using fast fourier transform (Rozewicki et al., 2019) with default parameters. A maximum-likelihood phylogenetic tree was constructed with 1000 bootstrap replicates to assess branch support. The resulting phylogenetic tree was visualized using the web-based application tvBOT (Xie et al., 2023), and further annotation and integration with expression heatmaps were performed using Evolview v3 (Subramanian et al., 2019). RNA-seq datasets from tomato were used for gene co-expression analysis, as described previously (Moore et al., 2020). Hierarchical clustering analysis of gene expression was performed using the GenePattern platform (https://www.genepattern.org/) (Reich et al., 2006). Gene expression profiles were analyzed and visualized as heatmaps using TBtools-II (v2.095) (Chen et al., 2023). Cluster analysis was conducted to identify patterns of gene expression across different conditions.

Prediction of transmembrane domains and structural modeling of Sl-ABCB5

The tertiary structure of Sl-ABCB5 was predicted using the SWISS-MODEL online tool (https://swissmodel.expasy.org/) (Waterhouse et al., 2018). PyMOL (v4.60) was used to render and generate cartoon representations of the predicted structure. Transmembrane helices were predicted using TMHMM v2.0 (https://services.healthtech.dtu.dk/services/TMHMM-2.0/).

Insect choice feeding assays

WFT (F. occidentalis) were propagated using fresh chrysanthemum flowers under controlled conditions: 50% relative humidity, 26 °C, light intensity of 50,000 lux, and a 16:8 h (light:dark) photoperiod.

Approximately 10 ml of 0.75% agar solution was poured into tilted Petri dishes. After the agar solidified, a layer of plastic wrap was placed over the agar surface to prevent water evaporation. The prepared agar plates were stored at room temperature until use.

Two small slits were made in the plastic wrap, into which leaves of similar size and developmental stage from wild-type and mutant plants were inserted. Fifteen ice-anesthetized adult and 15 nymphal WFTs were placed equidistant between the two leaves in each Petri dish. The dishes were then covered and sealed, allowing WFTs to feed freely. After 3 days, each leaf was photographed, and the damaged leaf area was quantified.

Statistical analysis

Data were analyzed using an unpaired Student’s t-test or one-way ANOVA followed by Tukey’s honestly significant difference post hoc test, with at least three biological replicates analyzed in each case. For the unpaired Student’s t-test, statistical significance was defined as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. For one-way ANOVA followed by Tukey’s post hoc test, different letters indicate statistically significant differences (p < 0.05). All error bars represent SEM unless otherwise noted.

Data and code availability

The raw RNA-seq data used for Supplemental Figure 7 have been deposited in the NCBI Sequence Read Archive under accession number PRJNA1114878 and in the China National Genomics Data Center Genome Sequence Archive under accession number CRA037903 (CNCB-NGDC Members and Partners, 2025, Zhang et al., 2025). All other data are included in this article and the supplemental materials.

Funding

This study was supported by the National Key Research and Development Project of China (grant 2024YFD2001003), the National Natural Science Foundation of China (grant 32472718), the Leading Innovative and Entrepreneurial Team Program of Zhejiang Province (grant 2025R01007), the Natural Science Foundation of Zhejiang Province (grant LZ22C150005), the Starry Night Science Fund of Zhejiang University Shanghai Institute for Advanced Study (SN-ZJU-SIAS-0011), the Organized Scientific Research in the Horticulture Discipline of Zhejiang University (grant B231220.0005-26), and the Innovative Development of the Horticulture Discipline of Zhejiang University (grant B231220.0005-25).

Acknowledgments

We thank Dr. Wenwu Zhou (Zhejiang University) for providing western flower thrips, Dr. Fangbin Cao (Zhejiang University) for providing the PM-RFP marker, and Dr Han Guo (Kunming Institute of Botany,CAS) for assistance with the VIGS experiments. We also thank Fengjun Yang (Agricultural Experiment Station, Zhejiang University) for assistance with glasshouse tomato cultivation. No conflict of interest is declared.

Author contributions

T.L. and P.F. designed the experiments. T.L., L.H., J.J., J.W., and H.Z. performed most of the experiments. T.L., X.X., and S.F. conducted acylsugar analyses. R.L.L. and J.Y. provided guidance on experimental design. T.L. wrote the first draft of the manuscript. P.F. and R.L.L. revised the manuscript with input from all authors.

Published: February 17, 2026