Introduction

Trichomes are protuberant structures originating from epidermal cells in several plant species and are involved in diverse biological functions. For example, as physical barriers, trichomes protect plants against herbivores and pathogens (Wagner et al. 2004) and external environmental stresses, such as heavy metals, ozone, and ultraviolet irradiation (Domínguez-Solís et al. 2004; Li et al. 2005; Karabourniotis et al. 2020). Additionally, trichome-enriched metabolites act as natural pesticides (Dayan and Duke 2003; Schilmiller et al. 2008) and have been used as flavor and fragrance ingredients, pharmaceuticals, and food additives (Singh et al. 2016; Shi et al. 2018; Schuurink and Tissier 2020).

Arabidopsis has a unicellular non-glandular trichome (Hülskamp et al. 1998). Because of its simple structure and easy access, the Arabidopsis trichome has been used as a model system for studying epidermal cell development (Larkin et al. 2003; Schellmann et al. 2007; Wang et al. 2019). Trichome development in Arabidopsis can be divided into formation and morphology. Studies have shown that trichome formation is initiated by plant hormones. For example, cytokinins and gibberellic acid induce the expression of several zinc-finger protein (ZFP) genes, which regulate a protein complex called MYB-bHLH-WD40 repeat (MBW) complex (Li et al. 2021). The MBW complex activates expression of the GLABRA 2 gene, which encodes a homeodomain leucine zipper (HD-ZIP) protein and induces trichome initiation (Wang and Chen 2008). Similarly, the formation of unicellular, non-glandular trichomes in cotton is regulated by R2R3-MYB transcription factors (GhMYB25) (Machado et al. 2009) and HD-ZIP IV transcription factors (GhHOX3, GhHD-1) (Walford et al. 2012; Shan et al. 2014; Pei 2015).

Presently, seven types of multicellular glandular (types I, IV, VI, and VII) and multicellular non-glandular (types II, III, and V) trichomes have been identified in tomato (Solanum lycopersicum L.) based on length, multicellular base cells, and glandular cells (Kang et al. 2010; Glas et al. 2012). Recent studies have identified several genes that control the development of multicellular trichomes. For example, the Woolly gene encoding an HD-ZIP IV protein and B-type cyclin gene (SlCycB2) regulate the initiation of type I trichomes, which are the longest in tomato (Yang et al. 2011). The Hair (H) gene and its closest homolog Hair2 (H2), which encodes a C2H2 ZFP, regulate type I trichome development (Chang et al. 2018; Chun et al. 2021; Hua et al. 2022; Zheng et al. 2022). SlMYC1, a bHLH transcription factor, regulates the formation of type VI trichomes and terpene synthesis (Xu et al. 2018; Hua et al. 2021). SlSCL3, a GRAS-family transcription factor, regulates the size of type VI trichomes and terpene synthesis (Yang et al. 2021). Additionally, the closest Arabidopsis homologs (GIS, ZFP8, and GIS2) of the two tomato ZFPs H and H2 regulate trichome formation in Arabidopsis (Gan et al. 2007; An et al. 2012). However, protodermal factor2, the closest Arabidopsis homolog of Woolly, does not regulate trichome formation in Arabidopsis (Abe et al. 2003). These findings suggest that the genetic pathway responsible for trichome formation is partially conserved between tomato and Arabidopsis.

After trichome formation, trichome cells are expanded to form shapes. The Arabidopsis distorted mutants exhibit impaired trichome morphology, displaying swollen and distorted trichome structures (Zhang et al. 2005). The Arabidopsis DISTORTED genes encode subunits of two protein complexes, Wiskott–Aldrich syndrome protein verprolin-homologous protein (WAVE) complex and the actin-related protein (ARP) 2/3 complex (Mathur 2005; Yanagisawa et al. 2013). The WAVE complex regulates the ARP2/3 complex, which regulates actin polymerization (Frank et al. 2004; Yanagisawa et al. 2013). Similarly, tomato hairless and distorted mutants display swollen and distorted trichome structures. The corresponding tomato Hairless (Hl), Distorted-1 (Dt-1), and Hairless-2 (Hl-2) genes encode SRA1, SCAR2, and NAP1 proteins, respectively, all of which are subunits of the WAVE complex (Kang et al. 2016; Chang et al. 2019; Xie et al. 2020). Arabidopsis NAP1 regulates autophagy, essential in nitrogen deficiency and salt stress (Wang et al. 2016). Tomato SRA1 regulates the mechanical strength of stems (Kang et al. 2016). Maize BRICK1, a WAVE component, regulates epidermal cell lobe formation in leaves (Frank and Smith 2002). Rice NAP1-like protein, a WAVE complex, is involved in drought sensitivity (Huang et al. 2019). These findings indicate the involvement of the actin cytoskeleton in diverse functions, including trichome morphology. Therefore, the study of genes regulating actin cytoskeleton formation is essential in understanding trichome development. Although the mechanism of trichome formation has been extensively characterized in Arabidopsis, it is poorly understood in tomato plants.

Therefore, this study aimed to identify the mechanism of trichome formation, morphology, and density using transcriptomic techniques. To achieve this, we characterized a mutant defective in trichome morphology, showing swollen and distorted trichomes. Previously identified mutants with similar distorted trichome phenotypes were reported as hl and hl-2 (Kang et al. 2010; Xie et al. 2020). Therefore, the mutant identified in this study was named hairless-3 (hl-3). Moreover, the gene responsible for the mutation was identified by a combined candidate gene search, gene expression, and sequencing analyses.

Materials and methods

Plant materials and growth conditions

Tomato (S. lycopersicum L.) seeds of hl-3 (accession number TOMJPW4375) and wild-type (WT) ‘Micro-Tom’ (accession number LA3911) were obtained from the Gene Research Center, University of Tsukuba, Ibaraki, Japan. The seeds were sterilized using 25% (v/v) sodium hypochlorite for 5 min, followed by rinsing four times with sterilized water. The seeds were germinated on 1/2 MS medium containing 2% (w/v) sucrose (Mbcell, mb-s-48425), 0.5 g/L MES hydrate (Sigma, M2933) (pH 5.8 with KOH), and 0.8% plant agar. Subsequently, 1-week-old seedlings were transferred to Jiffy peat pots (Hummert International) and grown in a growth chamber with 60% humidity under 16 h of light (200 μmol m−2 s−1) at 24 °C and 8 h of darkness at 18 °C.

Analysis of trichome morphology and density

Trichome morphology was examined by dissecting microscopy and scanning electron microscopy (SEM) according to previously described procedures (Jeong et al. 2017). The second leaflets of the third compound leaves and stems between the second and third compound leaves of 4-week-old plants were used.

RT-PCR and genomic DNA PCR

Total RNA was extracted from the second leaflets of the third compound leaves of 4-week-old plants using TRIzol reagent (#15,596,018, Invitrogen), and cDNA was synthesized using a cDNA synthesis kit (#K1622, Thermo Fisher Scientific). Subsequently, reverse transcription (RT)-PCR was performed using a T100 thermal cycler (BioRad) in a total volume of 20 µL as previously described (Jeong et al. 2017). All primer sets used for RT-PCR are listed in Supplemental Table S1. SlACTIN (Solyc03g078400) was amplified with the ACTIN-RT primer set (Supplemental Table S1) and used as the control. Genomic DNA was extracted as previously described (Hong et al. 2020). Fragments corresponding to ARPC1 from WT and hl-3 plants were PCR-amplified using the ARPC1-gDNA1 primer set (Supplemental Table S1). The amplified DNA fragments were gel-purified using the LaboPass kit (Cosmo Genetech) and were sequenced at Macrogen (Seoul, South Korea).

Quantitative RT-PCR

The second leaflets of the third compound leaves, stems between the second and third compound leaves, premature floral buds, and roots were collected from 4-week-old WT, hl-3, and two complemented transgenic plants (ARPC1-2 and ARPC1-5) for ARPC1 expression analysis. Total RNA isolation and cDNA synthesis were performed as described in the preceding subsection on RT-PCR. Quantitative RT-PCR (qRT-PCR) was performed as described previously (Chun et al. 2021). All primers used for qRT-PCR are listed in Supplemental Table S1. SlACT7 (Solyc03g078400) was used as the control.

Terpene measurement

The fresh weight of the second leaflet of 4-week-old plants was measured. Then, the leaflet was incubated in 1 mL of tert-butyl methyl ether containing 1 ng µL−1 of tetradecane at 20 °C and shaken at 100 rpm (NB-101S, N-Biotech, Bucheon-si, Korea) for 5 min. The resulting solution was used for gas chromatography–mass spectrometry (GC–MS) analysis as previously described, with minor modifications (Kang et al. 2010). Briefly, 1 µL of extract was injected into a DB-5 fused-silica column (30 m length, 0.25 mm i.d., 0.25 µm thick stationary phase; Agilent, Santa Clara, CA, USA). The injector temperature was 280 °C. The initial column temperature was held at 40 °C for 5 min and then ramped to 90 °C at 20 °C min−1, 110 °C at 10 °C min−1, 250 °C at 15 °C min−1, and finally 320 °C at 20 °C min−1, which was maintained for 3 min. The flow rate of the helium carrier gas was set to 0.6 mL min−1. All compounds were analyzed using an Agilent 7890B GC system interfaced to a 5977A quadrupole mass spectrometer (Santa Clara, CA, USA) operated using 70 eV electron ionization and mixed selected ion monitoring (m/z 85.1 and 93.1) per scan (m/z 50–150) mode. The terpene content was normalized to the fresh weight of the leaf samples used for each extraction. α-Humulene was used as a standard to determine response factors for terpenes.

Amino acid comparison and phylogenetic analysis

For amino acid comparisons, the ARPC1 amino acid sequences of S. lycopersicum (SlARPC1, Solyc05g006470), Arabidopsis thaliana (At) (AtARPC1A, NP_0180648.1; AtARPC1B, NP_180688.1), Solanum tuberosum (StARPC1A1B, XP_006357213.1), Nicotiana tabacum (NtARPC1A, XP_016461293.1), Oryza sativa (OsARPC1B, XP_015623093), Drosophila melanogaster (DmARPC1A1B, NP_476596.1), Bos taurus (BtARPC1A, NP_001068827.1; BtARPC1B, NP_001014844.1), Caenorhabditis elegans (CeARPC1A1B, NP_499570.1), and Saccharomyces cerevisiae (ScARPC1A1B, NP_009793.1) were obtained from The National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/). CLUSTALW was used for amino acid sequence alignment. Conserved amino acids were shaded using BoxShade (https://embnet.vital-it.ch/software/BOX_form.html). The phylogenetic tree was constructed based on the neighbor-joining method using MEGA7 with bootstrap tests using 1000 replicates.

Vector construction and tomato transformation

The entire coding sequence of SlARPC1 was PCR-amplified from cDNA synthesized from total RNA of leaves with the ARPC1-BX primer set (Supplemental Table S1). The resulting fragment was digested with XhoI and BamHI and cloned into the pBI121 binary vector, which was cut with the same restriction enzymes. An approximately 4.5-kb promoter fragment upstream of the start codon in ARPC1 was PCR-amplified from genomic DNA using the ARPC1-SB primer set (Supplemental Table S1). The resulting fragment was digested with SbfI and BamHI and cloned into the above vector, which was digested with the same restriction enzymes, to replace the 35S promoter. For transformation, we used BC1F3 plants by backcrossing hl-3 with WT (cv. Micro-Tom). The constructed pARPC1::ARPC1 binary vector was introduced into Agrobacterium tumefaciens strain LBA4404 and used to transform the hl-3 mutant as described previously (Kang et al. 2016). The presence of the T-DNA insert and expression of ARPC1 in transgenic plants were confirmed through genomic PCR using the ARPC1-gDNA2 primer set and though RT-PCR using the ARPC1-RT primer set, respectively (Supplemental Table S1).

Data analysis

All statistical data analyses were performed using GraphPad Prism 9 software (La Jolla, San Diego, CA, USA) and Microsoft Excel (Redmond, WA, USA). Results are expressed as mean ± SE (standard error). Unpaired t tests were used to compare the two experimental groups. P values were defined as follows: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Results

The hl-3 mutant possesses defective trichome morphology

The trichrome morphology of the hl-3 mutant was significantly different from that of WT plants (cv. Micro-Tom). Compared with the WT, in which type I trichomes look like needles, hl-3 exhibited swollen and distorted trichomes (Fig. 1a). SEM showed that all trichome types were irregular and distorted, with the base stalk cells of type I trichomes from the hl-3 mutant being highly swollen (Fig. 1b). The type VI trichome, which contains a stalk cell with a glandular head on the stalk's tip, showed a different cell division pattern in the hl-3 mutant. More than 95% of glandular heads of type VI trichomes on leaves, stems, hypocotyls, and sepals of the hl-3 mutant were slanted due to the bent stalk (Fig. 1b). The hl-3 mutant showed normal growth, including plant growth, fruit development, and root development, compared with that of WT plants, the exception being trichome morphology (Fig. S1).

Fig. 1
figure 1

Surface characteristics of the leaves, stems, hypocotyls, and sepals of WT and hl-3 plants. Photographs of leaves, stems, and hypocotyls of 4-week-old plants, and sepals of 5-week-old plants. a Dissecting micrograph. The arrows indicate type I trichomes. Scale bars of the leaves, stems, and hypocotyls are 1 mm, and those of the sepals are 500 µm. b Scanning electron microscopy (SEM) micrographs. The arrows indicate type I trichomes. The yellow and red arrowheads indicate normal and abnormal type VI trichomes, respectively. Representative type VI trichomes are shown in the insets. The scale bars of the leaves, stems, hypocotyls, and sepals are 500 µm, and those of the insets are 100 µm

Full size image

RT-PCR analysis of tomato genes encoding the ARP2/3-WAVE complex in the hl-3 mutant

A previous study demonstrated that Hl and Ini encode SRA1 and ARPC2A, respectively, the subunits of ARP2/3 and WAVE complexes (Kang et al. 2016; Jeong et al. 2017). Therefore, we hypothesized that the hl-3 mutant is defective in ARP2/3 or WAVE complex genes. To test this hypothesis, we compared their expression levels between WT and hl-3 plants. The results showed that the expression pattern of the putative genes except for ARPC1 was similar between the WT and hl-3 plants. Compared with the single PCR-amplified fragment of WT ARPC1, amplification of hl-3 ARPC1 produced multiple fragments. Moreover, the size of the main fragment of hl-3 ARPC1 was smaller than that of WT ARPC1 (Fig. S2).

The hl-3 mutant has a single nucleotide mutation that causes missplicing of ARPC1

RT-PCR analysis confirmed that the hl-3 ARPC1 mRNA was smaller than the WT ARPC1 mRNA (Fig. 2a), suggesting that ARPC1 is a strong candidate gene for Hl-3. Direct sequencing of the amplified fragments revealed that hl-3 ARPC1 had an 82 bp deletion, corresponding to the 586–667 bp region in WT ARPC1 (Fig. S3 and Fig. 2b). Amino-acid-sequence analysis showed that WT ARPC1 consists of 375 amino acids, whereas hl-3 ARPC1 consists of 199 amino acids due to the generation of a premature stop codon at position 200, resulting from the abovementioned deletion (Fig. S4). To examine whether the deletion identified in the hl-3 ARPC1 mRNA was retained in the hl-3 ARPC1 genomic DNA, genomic DNA PCR encompassing the deleted cDNA region was performed, and the result showed that the size of the PCR-amplified fragments was similar between WT and hl-3 plants (Fig. 2c). However, genomic DNA sequencing revealed that hl-3 ARPC1 had a single nucleotide substitution, i.e., G > A at position 7132 bp compared with WT ARPC1. This single nucleotide change produced a 5′ splice-donor site mutation in intron 9, which caused the 82 bp deletion in the hl-3 ARPC1 mRNA due to exon 9 skipping (Fig. 2d and Fig. S5). qRT-PCR was performed using RNA isolated from various tissues of WT plants to examine the expression pattern of ARPC1. The result showed that ARPC1 was expressed 3.6-, 2.5-, and 3.7-fold more in the leaves, stems, and floral buds than in the roots (Fig. 2e).

Fig. 2
figure 2

hl-3 mutant has a defect in the splicing of ARPC1 RNA transcripts and ARPC1 is highly expressed in leaves, stems, and floral buds. a RT-PCR amplification of full-length ARPC1 from WT and hl-3 leaves. ACTIN was used as the control. b Partial-length ARPC1 cDNA sequence of WT and hl-3 plants. c PCR amplification of partial-length ARPC1 genomic DNA encompassing the mutated region from WT and hl-3 plants using the ARPC1-gDNA primer set (Supplemental Table S1). d Genomic DNA structure of ARPC1 between WT and hl-3 plants and sequence variation of the exon 9-intron 9 junction region. The mutated single nucleotide sequence of hl-3 ARPC1 gene is highlighted in bold with a gray box. The predicted splice-donor site (GT) is underlined in the WT sequence. e ARPC1 mRNA levels of the leaves, stems, floral buds, and roots of four-week-old WT plants. Data are presented as mean ± SE of three biological replicates

Full size image

Transgenic hl-3 plants with WT ARPC1 gene restore normal trichome development and trichome density

To demonstrate whether the ARPC1 mutation was responsible for the swollen trichome phenotype of the hl-3 mutant, we constructed a plant binary vector containing full-length WT ARPC1 cDNA fused with a 4.5 kb ARPC1 native promoter (Fig. 3a) and introduced it into the hl-3 mutant by Agrobacterium-mediated transformation. Fifteen T0 transgenic plants were obtained, and all of them exhibited normal type I trichome phenotypes. Two representative T2 transgenic plants (ARPC1-2 and ARPC1-5), with the T-DNA insertion and expressing WT ARPC1, were used for further experiments (Fig. 3b and c). Dissecting microscopy and SEM analysis showed that transformation of the hl-3 mutants with the WT ARPC1 gene restored normal trichome structure on the leaves, stems, hypocotyls, and sepals (Fig. 3d and Fig. S6). Moreover, the number of the different types of trichomes on the leaves was counted to determine whether ARPC1 is involved in trichome formation. Compared with that on the WT plants, there was a 53–62% decrease in the number of type I and VI trichomes on the hl-3 mutant; in contrast, there was no significant difference in type I and VI trichome densities between the transgenic and WT plants. However, the WT, hl-3, and transgenic plants had similar type III, V, and VII trichomes densities (Fig. 4), indicating that tomato ARPC1 plays an essential role in trichome morphology and trichome formation.

Fig. 3
figure 3

Trichome morphology of pARPC1::ARPC1-complemented hl-3 transgenic plants. a Schematic presentation of a binary vector containing ARPC1 promoter fused with WT ARPC1. RB right border, LB left border, Pnos nopaline synthase promoter, Tnos nopaline synthase terminator, NPTII neomycin phosphotransferase II gene. b Genomic DNA PCR analysis to detect T-DNA in pARPC1: ARPC1-complemented hl-3 transgenic plants (ARPC1-2 and ARPC1-5). c RT-PCR analysis to determine ARPC1 expression levels in pARPC1:ARPC1-complemented hl-3 transgenic plants (ARPC1-2 and ARPC1-5). The ACTIN gene (Solyc03g078400) was used as the control. All gene amplifications were performed for 30 PCR cycles. d Dissecting micrographs of trichomes of WT, hl-3, ARPC1-2, and ARPC1-5. Four-week-old plants were used for leaves, stems, and hypocotyls, and 5-week-old plants were used for sepals. Arrows indicate type I trichomes. The scale bars of the leaves, stems, and sepals are 2 mm, and those of the hypocotyls are 1 mm

Full size image
Fig. 4
figure 4

Trichome density of pARPC1::ARPC1-complemented hl-3 transgenic plants. The number of trichomes on the leaves of WT, hl-3 mutant, ARPC1-2, and ARPC1-5 plants was determined. Data show means ± SE (standard errors) of eight biological replicates of 4-week-old plants. Asterisks represent significant differences between WT and hl-3 mutant or transgenic plants (unpaired t test: ***P < 0.001; ****P < 0.0001)

Full size image

Tomato ARPC1 regulates MYC1 and terpene synthase genes

Previous studies have shown that SlWo and SlCycB2 regulate type I trichome formation (Yang et al. 2011; Gao et al. 2017), and SlMYC1 and SlSCL3 regulate both type VI trichome development and terpene synthesis (Xu et al. 2018; Yang et al. 2021). To examine whether Hl-3 is involved in the expression of these genes, their relative transcript levels in the leaves of WT, hl-3 mutant, and transgenic plants were compared. There was no significant difference in SlWo, SlCycB2, and SlSCL3 expression levels between the WT, hl-3, and transgenic plants (Fig. 5a); however, there was a 38% decrease in SlMYC1 expression in the hl-3 plants compared with that in the WT plants. Moreover, there was no significant difference in SlMYC1 expression between the WT and transgenic plants (Fig. 5a), indicating that SlARPC1 was involved in regulating SlMYC1. In tomato, the type VI and I/IV glandular trichomes produced specialized metabolites such as volatile terpenes and non-volatile acylsugars, respectively (Kortbeek et al. 2021). Several terpene synthases (TPSs), including TPS5, TPS9, TPS12, and TPS20, have been implicated in the synthesis of monoterpenes and sesquiterpenes in glandular trichomes (Schilmiller et al. 2009; Falara et al. 2011; Spyropoulou et al. 2014; Kortbeek et al. 2016). Branched-chain keto-acid dehydrogenases (BCKDHs) and acylsugar acyltransferases (ASATs) are involved in acylsugar biosynthesis (Slocombe et al. 2008; Mandal et al. 2020). Therefore, we examined whether these terpene synthase genes and acylsugar biosynthetic genes were regulated by Hl-3. There was no significant difference in SlTPS5 (monoterpene synthase gene) expression between the WT, hl-3, and transgenic plants (Fig. 5b). In contrast, SlTPS20 (monoterpene synthase gene) and SlTPS9 and SlTPS12 (sesquiterpene synthase genes) expression levels were 74–77% lower in the hl-3 mutant than in the WT plants. However, there was no significant difference in the expression levels of the genes between the WT and transgenic plants (Fig. 5b). Regarding the acylsugar biosynthetic genes, there was no significant difference in BCKDH-E1-α, BCKDH-E1-β, and SlASAT1, 2, 3, 4 expressions between the WT, hl-3, and transgenic plants (Fig. 5c).

Fig. 5
figure 5

qRT-PCR analysis of trichome- (a), terpene synthesis- (b), and acylsugar biosynthesis-related genes (c) in leaves of WT, hl-3, and pARPC1::ARPC1-complemented hl-3 transgenic plants. Data are presented as means ± SE (standard errors) of four biological replicates of 4-week-old plants. SlACT7 (Solyc03g078400) was used as the internal standard. Asterisks represent significant differences between hl-3 mutant and ARPC1-2 and ARPC1-5, or WT plants (unpaired t test: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Full size image

Tomato ARPC1 modulates the production of mono- and sesquiterpenes

To examine whether the reduced transcript levels of the TPSs affect terpene production, the levels of mono- and sesquiterpenes in the leaves of WT, hl-3, and transgenic plants were analyzed. Five monoterpenes and three sesquiterpenes were detected in the leaves of ‘Micro-Tom’. Compared with the WT levels, there was a 78–94% decrease in the levels of four monoterpenes (α-pinene, 2-carene, α-terpinene, and β-phellandrene) in the hl-3 mutant, whereas the levels of these monoterpenes in the two transgenic plants were similar to or slightly higher than those in the WT plants. There was no significant difference in the α-phellandrene level between the WT, hl-3, and ARPC1-5 plants. However, the α-phellandrene level in the ARPC1-2 transgenic plant was higher than that in the other plants (Fig. 6a). Similarly, there was an 80–82% decrease in the levels of sesquiterpenes (δ-elemene, β-caryophyllene, and α-humulene) in the hl-3 mutant compared with the WT plants. Although the sesquiterpene levels in the transgenic plants were not fully restored to WT levels, the sesquiterpene levels were 135–204% higher in the transgenic plants than in the hl-3 mutants (Fig. 6b).

Fig. 6
figure 6

Terpene content of the leaves of WT, hl-3, and pARPC1::ARPC1-complemented hl-3 transgenic plants. Levels of monoterpenes (a) and sesquiterpenes (b) were analyzed by GC–MS. Terpene content was normalized to the fresh weight of the leaves. Bars represent the means ± SE (standard errors) of five biological replicates of 4-week-old plants. Asterisks represent significant differences between hl-3 mutant and ARPC1-2 and ARPC1-5, or WT plants (unpaired t test: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Full size image

Comparison of ARPC1 homologs between tomato and other species

The ARPC1 amino acid sequences from different species were obtained from the NCBI (http://www.ncbi.nlm.nih.gov/,Fig. S7). The InterProScan (http://www.ebi.ac.uk/InterProScan/) and SMART (http://smart.embl-heidelberg.de) programs predicted that SlARPC1 has six WD40 repeat domains (Fig. 7a and Fig. S8). These domains are known to form β-propellers (Robinson et al. 2001). The ARPC1 protein of the hl-3 mutant was predicted to lack two WD40 domains at the C-terminal region (Fig. 7a and Fig. S8). A homolog search using NCBI (http://www.ncbi.nlm.nih.gov/blast) and Sol Genomics Network (https://solgenomics.net/tools/blast/) indicated that SlARPC1 is a single copy gene. Phylogenetic analysis and amino acid sequence alignment were performed to examine the relationship between ARPC1 proteins of different species. The results showed that the ARPC1 of plants belonged to a distinct clade from that of yeast and animals. The ARPC1 of tomato had 72–99% sequence similarity to that of other plants and 34–42% similarity to that of yeast and animals (Fig. 7b and Fig. S8), indicating that ARPC1 is highly conserved. Additionally, the six WD40 repeat domains were highly conserved in the ARPC1 protein of plants, yeast, flies, and animals (Fig. S8), suggesting that the WD40 domains are important for the function of ARPC1.

Fig. 7
figure 7

Structure of tomato ARPC1 protein and similarity between tomato ARPC1 and its homologs in other species. a Schematic representation of tomato ARPC1 protein. ARPC1 has six WD repeats that are predicted to form a β-propeller. Numbers above the bar indicate the amino acid positions of the repeats. Numbers below the triangles indicate the position of the premature stop codon in the hl-3 mutant. b Phylogenetic tree of ARPC1 in various species including tomato. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown next to the branches. Scale bar represents 0.05 amino acid substitutions per site

Full size image

Discussion

The findings of this study demonstrated that ARPC1 is involved in trichome formation and morphology in tomato plants. Preliminary studies in Arabidopsis have identified several Distorted trichome mutant genes, which encode subunits of the WAVE and ARP2/3 complexes and regulate actin filament organization (Mathur 2005; Szymanski 2005). In a previous study, we reported that the tomato hl and ini mutants exhibited curled and swollen trichomes, similar to the trichomes of the distorted mutants. Map-based cloning and molecular studies indicated that Hl and Ini encode SRA1, a subunit of WAVE, and ARPC2A, a subunit of ARP2/3, respectively (Kang et al. 2010; Jeong et al. 2017). Given that the hl-3 mutant shows a similar trichome phenotype as the hl and ini mutants, it can be speculated that Hl-3 might be related to the actin cytoskeleton. Here, we provide molecular and genetic evidence to support the hypothesis.

The candidate gene approach and gene sequencing analysis revealed that the hl-3 mutant has a single nucleotide mutation in the ARPC1 gene. ARPC1 is one of the seven subunits of the ARP2/3 complex (ARP2-3 and ARPC1-ARPC5) (Szymanski 2005). This ARP2/3 complex is activated by the WAVE complex, consisting of five subunits (Yanagisawa et al. 2013). The WAVE protein family has a conserved VCA (verprolin homology, central, acidic) domain in the C-terminal region. This domain is critical for activating the ARP2/3 complex (Weaver et al. 2003; Padrick et al. 2011). Studies have shown that ARPC1 interacts with the VCA domain of the WAVE complex in Saccharomyces cerevisiae (Pan et al. 2004; Balcer et al. 2010), and ARPC1 mutation can suppress actin nucleation, resulting in lethality. Additionally, the interaction of ARPC1 with ARPC4 and ARPC5 is vital for regulating actin nucleation (Balcer et al. 2010). The findings of the present study showed that the ARPC1 gene of the hl-3 mutant has a single nucleotide mutation, which caused missplicing and generated a premature stop codon in the C-terminal region. Additionally, the transformation of hl-3 mutants with WT ARPC1 restored the normal trichome phenotype, indicating that ARPC1 is H1-3. Moreover, the findings of the present study showed that tomato ARPC1 has six WD40 domains, which play pivotal roles in several cellular processes. Particularly, the WD40 domain plays an important role in protein–protein interactions and protein assembly (Xu and Min 2011; Mishra et al. 2012). In contrast, the predicted hl-3 ARPC1 does not contain two WD40 domains at the C-terminus. Therefore, we speculated that these domains are crucial for interactions with other proteins, such as ARPC4, ARPC5, and WAVE proteins. Additionally, the hl-3 mutant showed several PCR bands of ARPC1 transcripts, including the main truncated transcript. Therefore, the various forms of mutated ARPC1 protein may possess residual functions. However, the effect of hl-3 mutation on ARPC1 function remains to be elucidated.

Several studies have shown that ARP2/3 or WAVE gene mutants exhibit defective trichome morphology, which has been identified in rice, soybean, wild legume, and barrelclover (Miyahara et al. 2010; Hossain et al. 2012; Campbell et al. 2016; Zhou et al. 2016). These results imply that the function of the actin cytoskeleton in trichome morphology is well conserved in plant species. Moreover, ARP2/3 and WAVE possess diverse functions in different plant species. For example, in Arabidopsis, ARPC5 controls pavement cell shape by regulating auxin transport (García-González et al. 2020), and SRA1 controls stomatal closure in the dark (Isner et al. 2017). Maize BRICK1 regulates the development of epidermal cell lobes and stomata (Frank and Smith 2002). Rice NAP1-like protein is involved in the development of leaf cuticles and stomata, thus regulating drought sensitivity (Huang et al. 2019). Wild legume ARPC1 is essential for Mesorhizobium loti infection for root nodulation (Hossain et al. 2012). ARPC3 of Solanum habrochaites, a wild tomato species, regulates plant immunity in response to powdery mildew fungus (Sun et al. 2019). However, the findings of the present study showed that tomato ARPC1 is required for trichome formation and morphology determination and terpene synthesis.

Several studies have identified critical genes for trichome development in tomato. For instance, Wo encoding an HD-ZIP IV protein, and H and its close homologs encoding ZFPs, regulate the formation of type I trichomes (Yang et al. 2011; Chang et al. 2018; Chun et al. 2021). However, these genes are not involved in the distorted trichome phenotype. Recently, tomato HDZIPIV8 was found to regulate the expression of Hl-2, which encodes NAP1, a subunit of the WAVE complex (Xie et al. 2020). Hl encoding SRA1, another subunit of the WAVE complex, regulates trichome density and morphology (Kang et al. 2016). The findings of the present study showed that Hl-3 regulates trichome density and morphology by encoding ARPC1. Overall, these findings suggest that tomato ARPC2/3 and WAVE genes play a critical role in trichome density and morphology. However, how these genes regulate trichome density remains to be elucidated.

Terpenes are synthesized from isopentenyl diphosphate and dimethylallyl diphosphate. In general, these substrates are produced from two pathways in plant cells: the mevalonate pathway in the cytosol and the 2-C-methyl-d-erythritol-4-phosphate pathway in plastids. The basic substrates are condensed and converted to various terpenes by several terpene synthases (Falara et al. 2011; Pazouki and Niinemets 2016). In tomato, glandular trichomes, primarily type VI, express diverse terpene synthase genes and produce several terpenes (Falara et al. 2011; Zhou and Pichersky 2020). A recent study demonstrated that MYC1, a bHLH transcription factor, regulates type VI trichome development and terpene biosynthesis in tomato. Particularly, MYC1 regulates expression of the monoterpene synthase genes TPS5 and TPS20 and sesquiterpene synthase genes TPS9 and TPS12 (Xu et al. 2018). The decrease in monoterpene and sesquiterpene production in hl-3 mutants in the present study can be attributed to the lower density of type VI trichomes and lower expression of MYC1-regulated terpene synthase genes (TPS9, TPS12, and TPS20) in the hl-3 mutant. The actin cytoskeleton is essential for the cellular trafficking of organelles and metabolites in plant cells (Sampathkumar et al. 2013; Breuer et al. 2017). Therefore, it is possible that the distorted stems of type VI trichomes in the hl-3 mutant hamper cellular trafficking, leading to the insufficient transport of metabolic resources to the gland cells and inhibition of terpene synthesis.

Wild species of tomato such as Solanum pennellii, Solanum galapagense, and S. habrochaites produce large amounts of acylsugars, which are mostly synthesized in type IV trichomes (Simmons and Gurr 2005; Schilmiller et al. 2010; Glas et al. 2012; Kortbeek et al. 2021). In general, acylsugar biosynthesis can be divided into two stages: the first for synthesizing branched-chain fatty acids and the second for esterification of these acyl molecules to glucose or sucrose (Mandal et al. 2020). For example, BCKDH-E1-α and BCKDH-E1-β are involved in the first stage of acylsugar biosynthesis (Slocombe et al. 2008; Mandal et al. 2020). Several ASATs are involved in the second stage of acylsugar biosynthesis (Schilmiller et al. 2012, 2015; Fan et al. 2016). Our results showed no significant difference in the expressions of these BCKDH-E1s and ASAT1-4 acylsugar biosynthetic genes between WT, hl-3, and transgenic plants. Given that the primary location of acylsugar biosynthesis is type IV trichomes, which are rarely found in cultivated tomatoes (Simmons and Gurr 2005; Kortbeek et al. 2021), and the reduced density of type I trichomes in the hl-3 mutant may not affect the synthesis of acylsugars. In conclusion, the findings of this study identified the role of ARPC1 in the context of trichome morphology and formation and terpene synthesis in tomato.

Author contribution statement

JIC and JHK conceived the research; JIC and SMK performed all the experiments under the supervision of SK, CJ, and JHK; JIC, SMK, SK, CJ, and JHK analyzed the data and wrote the manuscript. All authors read and approved the manuscript.