SlLAX1 is Required for Normal Leaf Development Mediated by Balanced Adaxial and Abaxial Pavement Cell Growth in Tomato

SlLAX1 is Required for Normal Leaf Development Mediated by Balanced Adaxial and Abaxial Pavement... Abstract Leaves are the major plant organs with a primary function for photosynthesis. Auxin controls various aspects of plant growth and development, including leaf initiation, expansion and differentiation. Unique and intriguing auxin features include its polar transport, which is mainly controlled by the AUX1/LAX and PIN gene families as influx and efflux carriers, respectively. The role of AUX1/LAX genes in root development is well documented, but the role of these genes in leaf morphogenesis remains unclear. Moreover, most studies have been conducted in the plant model Arabidopsis thaliana, while studies in tomato are still scarce. In this study, we isolated six lines of the allelic curly leaf phenotype ‘curl’ mutants from a γ-ray and EMS (ethyl methanesulfonate) mutagenized population. Using a map-based cloning strategy combined with exome sequencing, we observed that a mutation occurred in the SlLAX1 gene (Solyc09g014380), which is homologous to an Arabidopsis auxin influx carrier gene, AUX1 (AtAUX1). Characterization of six alleles of single curl mutants revealed the pivotal role of SlLAX1 in controlling tomato leaf flatness by balancing adaxial and abaxial pavement cell growth, which has not been reported in tomato. Using TILLING (Targeting Induced Local Lesions IN Genome) technology, we isolated an additional mutant allele of the SlLAX1 gene and this mutant showed a curled leaf phenotype similar to other curl mutants, suggesting that Solyc09g014380 is responsible for the curl phenotype. These results showed that SlLAX1 is required for normal leaf development mediated by balanced adaxial and abaxial pavement cell growth in tomato. Introduction Leaves are the major plant organs whose primary function involves photosynthesis. Leaves play a major role in sensing the quality, quantity and duration of light, all of which are crucial for complete plant growth and development. Understanding leaf initiation and development is important in plant biology. Most leaves are dorsoventrally (upper to bottom) flattened and develop distinct upper (adaxial) and lower (abaxial) surfaces. Balanced co-ordination of polarity, auxin response and cell division is essential for formation and development of normal and flat leaves. Any imbalance of this co-ordination results in altered leaf shapes such as curly, crinkly, twisted, rolled, radial or shrunken leaves (Serrano-Cartagena et al. 1999, Yu et al. 2005, Liu et al. 2010, Liu et al. 2011). The formation of flat leaves enables the optimum capture of sunlight during photosynthesis. An important factor controlling leaf morphogenesis is the phytohormone auxin. Indole-3-acetic acid (IAA) is the natural form of auxin that controls various aspects of plant growth and development, including cell division, expansion and differentiation, leaf initiation and morphogenesis. One of the unique and intriguing features of auxin is its transport (Paciorek et al. 2005, Tromas and Perrot-Rechenmann 2010). It is known that auxin is synthesized in young leaves and in the shoot apex, and is transported basipetally to all plant organs (reviewed in Bennett et al. 1998, Tromas and Perrot-Rechenmann 2010). Auxin transport involves two patterns: long-distance transport through phloem and short-distance or cell to cell transport called polar auxin transport (PAT). At the cellular level, IAA is distributed through a combination of membrane diffusion (passive uptake), carrier-mediated uptake and proton-driven distribution (Delbarre et al. 1996). PAT contributes to 85% of short-distance auxin transport. It is well established that polar auxin localization controls the direction of auxin movement in whole-plant organs. Several auxin carriers have been identified, including AUX1/LAX (LAX: like AUX1), PIN (PIN-FORMED) and PGP/MDR (P-glycoprotein/multidrug resistance)-like proteins. AUX1/LAX is reported to be an auxin influx carrier that facilitates auxin movement from outside to inside the cell, while PIN is an efflux carrier that pumps auxin from the cell into the intercellular space. PGP/MDR-like proteins are reported to have the ability to be either influx or efflux carriers (Yang and Murphy 2009), but the contribution of these proteins is considerably small compared with that of the AUX1/LAX and PIN families (Kramer and Bennett 2006, reviewed in Swarup and Péret 2012). There are numerous studies highlighting the effects of mutations in the AUX/LAX gene family in the model plant Arabidopsis. However, most studies have focused on root phenotypes. For instance, the AUX1/LAX family has been reported to promote lateral root emergence and formation (Marchant et al. 2002, Swarup et al. 2008, reviewed in Péret et al. 2009), root gravitropism (Bennett et al. 1996, Marchant et al. 1999) and root–pathogen interactions (Lee et al. 2011). Recently, AUX1 function in the aerial parts of plants has received interest, but studies are still considerably scarce. In Arabidopsis, AUX1 has been reported to control phyllotaxis patterning (Bainbridge et al. 2008), vascular patterning, xylem differentiation (Fàbregas et al. 2015) and leaf serration (Kasprzewska et al. 2015). Additionally, although PAT is governed and maintained by the co-ordinated action of AUX1/LAX and PIN carrier proteins, among auxin carriers PIN1 is the most studied. The role of the PIN protein family in leaf morphogenesis is well documented, yet the role of AUX1/LAX remains neglected or is underestimated. Furthermore, almost all studies have been carried out in the model plant Arabidopsis, while the role of auxin influx carriers in other model plants such as tomato is poorly understood. In this study, we isolated six lines of curly leaf (curl) mutants from the ‘Micro-Tom’ mutant population that had been previously established by γ-ray irradiation and EMS (ethyl methanesulfonate) treatment (Saito et al. 2011, Shikata et al. 2016). The curl mutants showed dorsoventrally impaired leaf flatness, which exhibited severe upward bending or hyponasty on the transverse axis. Through map-based cloning combined with exome sequencing (ES), we characterized six alleles of the curly leaf mutants, which have a nonsense mutation in the SlLAX1 gene. We reported that the SlLAX1 gene controls the curly leaf phenotype in the tomato curl mutants. This feature has never been characterized. The characterization of several alleles of single curl mutants in this study sheds light on the pivotal role of SlLAX1 in controlling leaf flatness mediated by normal adaxial–abaxial pavement cell growth. We also combined forward and reverse genetic approaches to validate the candidate gene. Using TILLING (Targeting Induced Local Lesions IN Genome) technology, we screened another nonsense mutant allele that consistently shows a similar curly leaf phenotype to that of the curl mutants obtained by a forward genetic approach. Results Isolation and phenotypic characterization of the curly leaf mutants We previously developed a large mutant population of ‘Micro-Tom’, a model tomato cultivar, using γ-ray irradiation and EMS mutagenesis (Saito et al. 2011, Shikata et al. 2016). Currently, we have 9,216 EMS mutant lines. From the M3 generation of this mutant population, we isolated six mutant lines exhibiting a severe upward curly leaf (hyponastic) phenotype; three lines were used for further analysis (Fig. 1A, B). The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the wild type (WT) (Fig. 1C, D), suggesting that the impairment of leaf curvature was not detectable at the early vegetative stage. The leaves became curly about 1 month after sowing and were continuously curly until the end of the growing period. The initiation of curly leaves was not related to the transition from the vegetative to the reproductive stage, and the leaf phenotype could not be restored at any stage once the curly leaves had formed. Growing the curl mutants in a high-humidity environment in in vitro culture could not rescue the curly phenotype (Fig. 1E). Additionally, curly leaves continuously appeared irrespective of water availability in the soil medium (Fig. 1F, G). The leaf water potential of the mutants and WT was also comparable (Supplementary Table S1). These data suggested that the curly leaf mutant phenotype is persistent, irrespective of relative humidity or water availability. Fig. 1 View largeDownload slide Leaf morphology of the WT ‘Micro-Tom’ and three alleles of the curl mutants. (A, B) Mature leaf morphology of mature curl mutants in the (A) adaxial and (B) abaxial view. The leaf images were captured from 2-month-old plants from the fifth leaflet. Scale bar = 2 cm. (C, D) Appearance of the young leaves of curl mutants. (C) Adaxial and (D) abaxial view. The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the WT. Scale bar = 1 cm. (E) Representative of the curl mutant (curl-1) when grown in in vitro culture. Scale bar = 2 cm. The curly phenotype was not restored. (F, G) Wild-type (F) and representative curl mutant (G, curl-1) grown under well-watered conditions in the greenhouse. Plant images were captured from 2-month-old plants. Scale bar = 1.5 cm. Fig. 1 View largeDownload slide Leaf morphology of the WT ‘Micro-Tom’ and three alleles of the curl mutants. (A, B) Mature leaf morphology of mature curl mutants in the (A) adaxial and (B) abaxial view. The leaf images were captured from 2-month-old plants from the fifth leaflet. Scale bar = 2 cm. (C, D) Appearance of the young leaves of curl mutants. (C) Adaxial and (D) abaxial view. The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the WT. Scale bar = 1 cm. (E) Representative of the curl mutant (curl-1) when grown in in vitro culture. Scale bar = 2 cm. The curly phenotype was not restored. (F, G) Wild-type (F) and representative curl mutant (G, curl-1) grown under well-watered conditions in the greenhouse. Plant images were captured from 2-month-old plants. Scale bar = 1.5 cm. We also analyzed the percentage of reduced leaf area and perimeter in both young and mature leaves of mutants by flattening the curl mutant leaves. In the young leaves, leaf area was markedly reduced (41.0–56.0%, Table 1). The leaf perimeters of the WT and mutants were comparable. Consistently, in the mature leaves, the reduction in leaf area was more evident (55.8–64.0%) (Table 1), indicating a progression of severity that was concomitant with leaf maturity. Then, to investigate how and when the curly leaf is formed and its progression at the organ level, we measured the curvature index (CI) in both young and mature leaves according to the method of Liu et al. (2010). Negative curvature represents upward bending of the leaf. At the early stage of leaf initiation and development, mutants developed and maintained flat leaves; after several (4–7) days following leaf initiation, the leaves gradually became curly, and the curly leaf severity increased concomitant with leaf maturity (Supplementary Fig. S1A–C). The leaf incurvature was initiated from young leaves firstly at the tip along the transversal axis to a low extent, while the longitudinal axis remained flat in all mutant lines (Table 2, Fig. 2A, B). To understand the curly leaf progression, the global curvature of mature leaves of all mutants was also measured (Table 2). Consistently, leaf incurvature was observed along the transversal axis to a high extent and the longitudinal axis remained normal at all stages of leaf development. In the mature leaves, the whole leaf had become curly (Fig. 2C, D). These data suggested that leaf incurvature was more severe as leaf maturity progressed. Table 1 Leaf area and leaf perimeter of young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf area and perimeter were observed when the curly leaf was being formed, about 6 d after leaf initiation. The mature leaf leaf area and perimeter were observed when the leaf had become completely curly, about 10 d after leaf initiation. Values are means ± SE (n = 15). The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01); WT mean values were used as controls. The reductions in leaf area and leaf perimeter (%) were measured by comparing the values before and after flattening (multiplied by 100). Table 1 Leaf area and leaf perimeter of young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf area and perimeter were observed when the curly leaf was being formed, about 6 d after leaf initiation. The mature leaf leaf area and perimeter were observed when the leaf had become completely curly, about 10 d after leaf initiation. Values are means ± SE (n = 15). The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01); WT mean values were used as controls. The reductions in leaf area and leaf perimeter (%) were measured by comparing the values before and after flattening (multiplied by 100). Table 2 The global curvature index of the young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Values are means ± SE (n = 15). The curvature index (CI) of mutants was measured by a method introduced by Liu et al. (2010). CI = (ab – a’b’)/ab. ab = the distance between points a and b on two margins of curvature before flattening of leaves. a’b’ = the distance between a and b on two margins after flattening. The CI was measured in the middle of the leaves. The flatness of either young or mature leaf curl mutants was impaired along the transverse axis, whereas the longitudinal axis was normal. A negative CI represents upward curvature. The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01). Table 2 The global curvature index of the young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Values are means ± SE (n = 15). The curvature index (CI) of mutants was measured by a method introduced by Liu et al. (2010). CI = (ab – a’b’)/ab. ab = the distance between points a and b on two margins of curvature before flattening of leaves. a’b’ = the distance between a and b on two margins after flattening. The CI was measured in the middle of the leaves. The flatness of either young or mature leaf curl mutants was impaired along the transverse axis, whereas the longitudinal axis was normal. A negative CI represents upward curvature. The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01). Fig. 2 View largeDownload slide Adaxial and abaxial surfaces of young (upper panel) and mature (bottom panel) tomato leaflets. (A) Adaxial (upper) surface of young tomato leaflets. (B) Abaxial (bottom) surface of young tomato leaflets. Young leaflets were detached from 1.5-month-old plants. (C) Adaxial (upper) surface of mature tomato leaflets. (D) Abaxial (bottom) surface of mature tomato leaflets. Mature leaflets were detached from the fifth leaflet of 2.5-month-old plants. Scale bar: upper panel = 3 cm; bottom panel = 2 cm. Fig. 2 View largeDownload slide Adaxial and abaxial surfaces of young (upper panel) and mature (bottom panel) tomato leaflets. (A) Adaxial (upper) surface of young tomato leaflets. (B) Abaxial (bottom) surface of young tomato leaflets. Young leaflets were detached from 1.5-month-old plants. (C) Adaxial (upper) surface of mature tomato leaflets. (D) Abaxial (bottom) surface of mature tomato leaflets. Mature leaflets were detached from the fifth leaflet of 2.5-month-old plants. Scale bar: upper panel = 3 cm; bottom panel = 2 cm. Genetic mapping of curl mutants To examine the inheritance pattern of the curl mutants, we crossed the mutants with the WT ‘Micro-Tom’ and another tomato cultivar ‘Ailsa Craig’, and observed the segregation ratio in the F2 population. Phenotypic observation was carried out visually according to the presence or absence of the curly leaf phenotype. The mutant phenotype appeared in the F2 generation only as a recessive genetic trait (Table 3). The ratio of WT and mutant phenotypes fit to the Mendelian segregation ratio for monogenic traits (3:1), indicating a monogenic recessive inheritance of all curl mutants. Similarly, in the ‘Ailsa Craig’ background, the inheritance of the curl mutants was also recessive (Supplementary Table S2). An allelism test was performed to observe complementation effects among mutant lines and to examine whether mutations occurred because of the same causal gene. The complementation effect was determined in the F1 generation. All crosses between each pair of mutant lines showed curly leaf phenotypes (Table 4), indicating that they are allelic, which means that causal mutation occurred in the same locus. The curl-6 mutant generated from EMS treatment (see ‘Plant material and growth conditions’) was also allelic with the other mutants which were generated from γ-ray irradiation. We confirmed that all curl mutant lines were allelic; therefore, for further analyses, we used only three mutant alleles, namely curl-1, curl-2 and curl-6. Table 3 Segregation analysis of the curl mutants back-crossed to the wild-type ‘Micro-Tom’ Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  a The curl mutants were crossed to the wild-type ‘Micro-Tom’. b The number of progeny exhibiting normal (WT) and curly leaf phenotype is shown. c χ2 value was calculated based on progeny segregation in the F2 population. d χ2 distribution in the table reference value, with probability >0.05 and degree of freedom 1. e Inheritance pattern of the curl mutants, estimated based on the χ2 value at the 95% (P < 0.05) significance level. Table 3 Segregation analysis of the curl mutants back-crossed to the wild-type ‘Micro-Tom’ Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  a The curl mutants were crossed to the wild-type ‘Micro-Tom’. b The number of progeny exhibiting normal (WT) and curly leaf phenotype is shown. c χ2 value was calculated based on progeny segregation in the F2 population. d χ2 distribution in the table reference value, with probability >0.05 and degree of freedom 1. e Inheritance pattern of the curl mutants, estimated based on the χ2 value at the 95% (P < 0.05) significance level. Table 4 The result of the allelism test among the curl mutants Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    The allelism test was carried out by crossing all possible pairs and observing the results in the F1 generation visually by the presence of a curly leaf phenotype. Normal represents the wild-type phenotype; Curly represents the curly leaf phenotype; ND, not determined; ♀, ♂, female recipient and male donor, respectively. Table 4 The result of the allelism test among the curl mutants Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    The allelism test was carried out by crossing all possible pairs and observing the results in the F1 generation visually by the presence of a curly leaf phenotype. Normal represents the wild-type phenotype; Curly represents the curly leaf phenotype; ND, not determined; ♀, ♂, female recipient and male donor, respectively. To identify the candidate gene controlling the curly leaf phenotype, we performed a map-based cloning approach using PCR-based DNA markers including CAPS (cleaved amplified polymorphic sequence) and SNPs (single nucleotide polymorphisms) (Shirasawa et al. 2010, Ariizumi et al. 2014, Chusreeaeom et al. 2014, Hao et al. 2017). We found that the mutation probably occurred in the short arm of chromosome 9 (Supplementary Table S3). The highest ‘Micro-Tom’ allele frequency was observed in this chromosome region between markers tomInf5375 and 14109_151, and ranged from 0.68 to 0.89, suggesting that the gene responsible could be localized in the short arm of chromosome 9 close to marker 14109_151 (physical position SL2.40ch09:2052389, Fig. 3). Fig. 3 View largeDownload slide Partial chromosome mapping result of the curl mutant locus. The curl locus was found to associate with the marker 14109-151 on chromosome 9 in the F2 mapping population derived from the cross between S. lycopersicum cv. ‘Ailsa Craig’×S. lycopersicum cv. ‘Micro-Tom’ curl-2. The marker information was obtained from the Kazusa DNA Research Institute AMF2 database (http://marker.kazusa.or.jp/). No such association was observed in other chromosomes (Supplementary Table S3). Fig. 3 View largeDownload slide Partial chromosome mapping result of the curl mutant locus. The curl locus was found to associate with the marker 14109-151 on chromosome 9 in the F2 mapping population derived from the cross between S. lycopersicum cv. ‘Ailsa Craig’×S. lycopersicum cv. ‘Micro-Tom’ curl-2. The marker information was obtained from the Kazusa DNA Research Institute AMF2 database (http://marker.kazusa.or.jp/). No such association was observed in other chromosomes (Supplementary Table S3). The SlLAX1 gene is commonly mutated in several curl mutant alleles To narrow down the candidate region obtained by rough mapping, we performed ES. Four lines of the curl mutants, curl-1, curl-2, curl-3 and curl-6, were used for the ES analysis. The F2 progeny derived from the cross between mutant and WT ‘Micro-Tom’ were divided into flat leaf and curly phenotype based on the presence or absence of the curly leaf phenotype, and then flat leaf and mutant bulked segregants were subjected to ES. By the bowtie2-GATK pipeline using the tomato genome reference version SL2.50 as a reference (see the Materials and Methods), we identified 5,430, 5,110, 5,050 and 4,829 genome-wide mutations for curl-1, curl-2, curl-3 and curl-6 mutant segregants, respectively. When allele frequencies were compared between these mutants, a strong association was found around the top region of chromosome 9 in all four mapping populations (Fig. 4). This result suggested that the causal gene for the curly phenotype is located in this chromosome region, in agreement with the result of rough mapping of chromosoms (Fig. 3; Supplementary Table S3). Furthermore, we then searched for the gene in which mutation commonly occurs in some of the curl mutants. We found that mutations commonly occur in Solyc09g014380.2.1, which is a homolog of Arabidopsis AtAUX1 (AT2G38120; BLASTx E-value = 0.0, protein amino acid similarity = 93%). The Solyc09g014380.2.1, tomato locus SlLAX1, gene spans an approximately 3.8 kb genomic region, while cDNA including the untranslated region (UTR) spans 1.8 kb. The SlLAX1 gene has seven exons, including a UTR in both the 5' and 3' ends (Fig. 4). The curl-2 and curl-6 mutants had a nucleotide substitution from G to A in exon 6, physical position SL2.50ch09: 6,010,739 bp (Table 5). This SNP produced a premature stop codon (W262*) in the deduced protein sequence of SlLAX1. According to the SL2.50 tomato genome reference, WT ‘Micro-Tom’ produced 411 amino acids of the SlLAX1 protein, whereas the curl-2 and curl-6 mutants produced only 261 amino acids of the protein, losing the last 150 amino acids (63.7% of the WT protein). curl-1 and curl-3 had an SNP from G to T in the splicing junction of intron 4, physical position SL2.50ch09: 600,292 bp. These mutations were also confirmed by dideoxy sequencing of cDNA (Fig. 5A, B). Table 5 Predicted mutation position, amino acid substitution and mutation type based on the whole-exome sequence result Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  a The location in the chromosome in the tomato genome. b The position of the nucleotide substitution according to the tomato genome sequence database, version SL2.50 (Sol Genomics Network). c Tomato genome sequence reference according to the position in the second column. d Alternative nucleotide sequence/nucleotide substitution according to the position in the second column. e Location of nucleotide substitution of the gene in the sixth column. f Gene mutated according to the Sol Genomic Network database. g *A stop codon. Table 5 Predicted mutation position, amino acid substitution and mutation type based on the whole-exome sequence result Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  a The location in the chromosome in the tomato genome. b The position of the nucleotide substitution according to the tomato genome sequence database, version SL2.50 (Sol Genomics Network). c Tomato genome sequence reference according to the position in the second column. d Alternative nucleotide sequence/nucleotide substitution according to the position in the second column. e Location of nucleotide substitution of the gene in the sixth column. f Gene mutated according to the Sol Genomic Network database. g *A stop codon. Fig. 4 View largeDownload slide Identification of SlLAX1 (Solyc09g01480.2) as the most plausible candidate gene responsible for the curl phenotype. Genome-wide allele frequency data were obtained by exome sequencing of BCF2 bulked segregants that show the curl mutant phenotype. To narrow down candidates efficiently, four mapping populations derived from independent curl alleles (curl-1, 2, 3 and 6) were constructed and subjected to exome sequencing. In all four mapping populations, a strong association was commonly observed for mutations within the SlLAX1 (Solyc09g01480.2) gene, which is a homolog of the Arabidopsis AUXIN RESISTANT1 (AUX1) transporter gene. Black boxes indicate exons, transparent boxes indicate UTRs, and lines between boxes indicate introns. Fig. 4 View largeDownload slide Identification of SlLAX1 (Solyc09g01480.2) as the most plausible candidate gene responsible for the curl phenotype. Genome-wide allele frequency data were obtained by exome sequencing of BCF2 bulked segregants that show the curl mutant phenotype. To narrow down candidates efficiently, four mapping populations derived from independent curl alleles (curl-1, 2, 3 and 6) were constructed and subjected to exome sequencing. In all four mapping populations, a strong association was commonly observed for mutations within the SlLAX1 (Solyc09g01480.2) gene, which is a homolog of the Arabidopsis AUXIN RESISTANT1 (AUX1) transporter gene. Black boxes indicate exons, transparent boxes indicate UTRs, and lines between boxes indicate introns. Fig. 5 View largeDownload slide Changes in protein amino acid sequence and SlLAX1 gene expression in curl mutants. (A, B) A partial alignment of the SlLAX1 cDNA sequence (A) or deduced protein amino acid sequence (B) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-2 and curl-6. The mutation in curl-2 and curl-6 causes a premature stop codon, as shown by the red box (W262*). (C, D) A partial alignment of SlLAX1 cDNA sequence (C) or deduced protein amino acid sequence (D) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-1 and curl-3. cDNA sequences were obtained by dideoxy sequencing (A, C). (E) Donor and acceptor splicing sites in intron 4 of the wild type, curl-1 and curl-3 mutants. Square brackets indicate splicing sites. Double square brackets indicate the alternative splicing site in the curl-1 and curl-3 mutants. The one-letter code indicates an amino acid. Upper case indicates an exon, whereas lower case indicates an intron sequence. The bold letter indicates a mutated sequence in intron 4 of the curl-1 and curl-3 mutants. The asterisk represents the stop codon in curl-1. (F) qRT-PCR analysis of SlLAX1 gene expression. qRT-PCR primers were designed to target downstream of the stop codon mutation in exon 6. The asterisks represent statistically significant differences in the mean with equal variants compared with the wild type (WT) based on Student’s t-test (**P < 0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Fig. 5 View largeDownload slide Changes in protein amino acid sequence and SlLAX1 gene expression in curl mutants. (A, B) A partial alignment of the SlLAX1 cDNA sequence (A) or deduced protein amino acid sequence (B) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-2 and curl-6. The mutation in curl-2 and curl-6 causes a premature stop codon, as shown by the red box (W262*). (C, D) A partial alignment of SlLAX1 cDNA sequence (C) or deduced protein amino acid sequence (D) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-1 and curl-3. cDNA sequences were obtained by dideoxy sequencing (A, C). (E) Donor and acceptor splicing sites in intron 4 of the wild type, curl-1 and curl-3 mutants. Square brackets indicate splicing sites. Double square brackets indicate the alternative splicing site in the curl-1 and curl-3 mutants. The one-letter code indicates an amino acid. Upper case indicates an exon, whereas lower case indicates an intron sequence. The bold letter indicates a mutated sequence in intron 4 of the curl-1 and curl-3 mutants. The asterisk represents the stop codon in curl-1. (F) qRT-PCR analysis of SlLAX1 gene expression. qRT-PCR primers were designed to target downstream of the stop codon mutation in exon 6. The asterisks represent statistically significant differences in the mean with equal variants compared with the wild type (WT) based on Student’s t-test (**P < 0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. As described above, the curl-1 and curl-3 mutants had a mutation in the first nucleotide or splicing junction of intron 4 (Fig 4; Table 5). Interestingly, sequencing of SlLAX1 cDNA in these alleles revealed that abnormal splicing occurred around intron 4, which led to deletion of five nucleotides within exon 4 (nucleotides 433–437, Fig. 5C). Given that mutation in curl-1 and curl-3 is a G to T substitution in the splicing junction of intron 4, presumably there was an alteration in the donor and recipient sites for intron splicing. Splicing of intron 4 occurred 435 bp from the start codon in the tomato genome of the WT, whereas intron splicing occurred 5 bp upstream of the end of exon 4 (430 bp from the start codon) in both the curl-1 and the curl-3 alleles. Then the next sequence following exon 5 is GGTTGA; this TGA may produce a premature stop codon, which is at 435 bp from the start codon (Fig. 5E). Thus, curl-1 and curl-3 alleles could produce a C-terminal truncated SlLAX1 protein that is only 145 amino acids in length (Fig. 5D). We also analyzed the transcript level of the SlLAX1 by quantitative real-time PCR (qRT-PCR) using mature curly leaf cDNA. The expression of the SlLAX1 gene in the three curly leaf mutants was significantly reduced to only 35–40% of WT expression (Fig. 5F), which indicates low abundance of this gene transcript in the mutants. Taken together, these results indicated that all of curl mutants carried a loss-of-function mutation in the SlLAX1 gene. Screening a new allele of the nonsense mutation of SlLAX1 by TILLING Because our research group had previously developed large mutant resources in the ‘Micro-Tom’ background and proved that TILLING is an efficient tool for isolating desired mutants from the ‘Micro-Tom’ mutant collection (Okabe et al. 2011), we utilized TILLING to search for other SlLAX1 mutant alleles. We screened 4,608 lines in the M2 and M3 generations to obtain new SlLAX1 mutant alleles. In addition, because we only had one EMS mutant screened by forward genetics (curl-6), we attempted to obtain other mutant alleles to confirm consistency of the phenotype. We designed a primer pair to amplify 865 bp along exon 6 of the SlLAX1 gene for the TILLING screening target, and found five new mutant alleles that carried intron, missense and nonsense mutations (Supplementary Fig. S2A, B; Supplementary Table S5). The curl-6/TOMJPE8506, which was previously isolated by forward genetics, was also confirmed by TILLING screening. Then, to validate the mutant phenotype, one line that carried a nonsense mutation, TOMJPW601-1, was renamed ‘curl-7’ and used for further analysis. This mutant line carried a 1 bp substitution from G to A in the 554th nucleotide from the start codon, which led to the conversion of tryptophan to a premature stop codon at the 185th amino acid (Fig. 6A). The curl-7 mutant exhibited the curly leaf phenotype like the other curl mutant alleles (Fig. 6B). Furthermore, by dideoxy sequencing, we confirmed the consistency of the TILLING result (Fig. 6C, D). This result supports the evidence that SlLAX1 is the gene responsible for the curly leaf phenotype in tomato. These results again indicated that mutation in SlLAX1 produces the curly leaf phenotype. Mutations in the same gene consistently resulted in the same phenotype, strongly suggesting that SlLAX1 functions in controlling the tomato curly leaf phenotype. Fig. 6 View largeDownload slide TILLING screening results and confirmation of the presence of the curly leaf phenotype, and cDNA and amino acid sequence alignment of the new mutant allele, TOMJPW601-1/curl-7. (A) A polyacrylamide gel image of TILLING screening. The mutation in TOMJPW601-1/curl-7 is shown as an intense spot on the lanes both in IRD-700 (red circle) and in IRD-800 (green circle). A single nucleotide change is shown on the sequence chromatogram (red arrowhead). (B) Whole-plant images of curl-6 (left); a representative of the curl allele obtained using forward genetics; (middle and right) confirmation of the presence of curly leaves in the new selected allele, curl-7, in the M3 generation. Plant images were captured from 2-month-old plants when the curly leaf phenotype progressed. Scale bar = 2 cm. (C) A partial alignment of the SlLAX1 cDNA sequence among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Nucleotide substitution in the curl-7 mutant is shown by gray highlighting. (D) Partial protein amino acid sequence alignment of SlLAX1 (Solyc09g01480.2) among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Mutation in curl-7 led to the conversion of tryptophan to a premature stop codon. The wild type (WT) produced a 411 amino acid product, whereas curl-7 produced only a 185 amino acid product. The premature stop codon is indicated by a red box. Fig. 6 View largeDownload slide TILLING screening results and confirmation of the presence of the curly leaf phenotype, and cDNA and amino acid sequence alignment of the new mutant allele, TOMJPW601-1/curl-7. (A) A polyacrylamide gel image of TILLING screening. The mutation in TOMJPW601-1/curl-7 is shown as an intense spot on the lanes both in IRD-700 (red circle) and in IRD-800 (green circle). A single nucleotide change is shown on the sequence chromatogram (red arrowhead). (B) Whole-plant images of curl-6 (left); a representative of the curl allele obtained using forward genetics; (middle and right) confirmation of the presence of curly leaves in the new selected allele, curl-7, in the M3 generation. Plant images were captured from 2-month-old plants when the curly leaf phenotype progressed. Scale bar = 2 cm. (C) A partial alignment of the SlLAX1 cDNA sequence among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Nucleotide substitution in the curl-7 mutant is shown by gray highlighting. (D) Partial protein amino acid sequence alignment of SlLAX1 (Solyc09g01480.2) among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Mutation in curl-7 led to the conversion of tryptophan to a premature stop codon. The wild type (WT) produced a 411 amino acid product, whereas curl-7 produced only a 185 amino acid product. The premature stop codon is indicated by a red box. Endogenous IAA levels and the expression of auxin-related genes in curl mutants As described above, all curl mutants commonly have a mutation in the SlLAX1 gene, which encodes an auxin influx carrier. To test the potential function of SlLAX1 as an auxin transporter in tomato, we measured the leaf auxin content at three stages: (i) in young leaves, before curly leaves formed; (ii) when leaves just became curly; and (iii) in mature leaves, after leaves were fully curly. The IAA content significantly decreased from young leaves to mature leaves in both the WT and three curl mutants (Supplementary Fig. S3A). However, the IAA content at each leaf stage was comparable between the WT and the curl mutants. Similarly, IAA conjugates and total IAA between the WT and the curl mutants were also comparable (Supplementary Fig. S3B, C). In Arabidopsis, numerous findings have indicated the role of the LAX1/AUX1 family in root gravitropism and lateral root formation (Bennett et al. 1996, Marchant et al. 2002, reviewed in Swarup and Péret 2012). Importantly, root agravitropism is the most prominent defect and well-characterized trait of the Arabidopsis aux1 mutant. In addition, the aux1 mutant also showed lateral root formation defects (Marchant et al. 2002). Thus, we further tested these traits in the curl mutants; as expected, the curl mutants showed agravitropism as well as reduced lateral root formation, in agreement with the Arabidopsis aux1 mutant phenotype (Supplementary Fig. S4), suggesting the possibility of involvement of SlLAX1 as an auxin influx carrier in tomato similar to AtAUX1. Abaxial pavement cell size of the curl mutants was significantly larger Because SlLAX1 gene function was commonly disabled in curl mutants and auxin has been known to affect pavement cells (Pérez-Pérez et al. 2010, reviewed in Sandalio et al. 2016), we hypothesized that the curly leaf formation may be related to differential cell growth on the adaxial and abaxial surfaces. To observe histological features of the curl mutants, we measured pavement cell size using scanning electron microscopy (SEM) of the adaxial and abaxial surfaces at the mature leaf stage in the curly part (Table 6; Fig. 7). We noted that cell enlargement in the curl mutants was more prominent on the abaxial side, while there was no significant difference in adaxial pavement cells. As a consequence, the ratio of abaxial and adaxial pavement cells was more prominent in the curl mutants. We also quantified the pavement cell number on both the adaxial and abaxial surfaces. The number of pavement cells on both surfaces was comparable (Table 6). These data revealed that impairment of leaf flatness in the curl mutants is likely to be due to the differential cell growth between the adaxial and abaxial epidermal layers. Most probably, the curly leaf phenotype is related to cell enlargement on the abaxial side. Table 6 Adaxial and abaxial pavement cell size of the curl mutants in the curly part measured by SEM Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Values are means ± SE (n = 9). The asterisks represent statistically significant differences in means with equal variants based on the Student’s t-test (**P < 0.01). The cell features were measured at the mature leaf stage when the leaves were completely curly, precisely in the same regions on the adaxial and abaxial surfaces. The curl mutants showed a significantly larger abaxial/adaxial pavement cell size ratio compared with the wild type (WT). Table 6 Adaxial and abaxial pavement cell size of the curl mutants in the curly part measured by SEM Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Values are means ± SE (n = 9). The asterisks represent statistically significant differences in means with equal variants based on the Student’s t-test (**P < 0.01). The cell features were measured at the mature leaf stage when the leaves were completely curly, precisely in the same regions on the adaxial and abaxial surfaces. The curl mutants showed a significantly larger abaxial/adaxial pavement cell size ratio compared with the wild type (WT). Fig. 7 View largeDownload slide Adaxial and abaxial pavement cells in the wild type (WT) and the curl mutants in the curly part. (A) The adaxial pavement cell size of the WT and mutants was comparable. Scale bar = 20 µm. (B) The pavement cell size of all curl mutants on the abaxial surface was significantly larger compared with that of the WT. Scale bar = 10 µm. (C) The adaxial and abaxial sides of the curly part of the leaf that were subjected to SEM. Images were captured using a scanning electron microscope with ×400 magnification in the curly part at precisely the same position in both the adaxial and abaxial surfaces. Fig. 7 View largeDownload slide Adaxial and abaxial pavement cells in the wild type (WT) and the curl mutants in the curly part. (A) The adaxial pavement cell size of the WT and mutants was comparable. Scale bar = 20 µm. (B) The pavement cell size of all curl mutants on the abaxial surface was significantly larger compared with that of the WT. Scale bar = 10 µm. (C) The adaxial and abaxial sides of the curly part of the leaf that were subjected to SEM. Images were captured using a scanning electron microscope with ×400 magnification in the curly part at precisely the same position in both the adaxial and abaxial surfaces. Relative expression of auxin-related genes in the curl mutants Recently, some studies have reported that impairment of auxin biosynthesis, signaling, degradation and conjugation results in leaf development defects such as wrinkled, curled leaf and rounded leaf phenotypes. We checked the relative expression of some putative tomato auxin-related genes which were reported to be involved in controlling the leaf flatness phenotype such as AtDof5.1 (Kim et al. 2010) which is homologous to SlDof25 and SlDof28 in tomato (Cai et al. 2013), LCR (LEAF CURLING RESPONSIVENESS) (Song et al. 2012), PNH (PINHEAD) (Newman et al. 2002) and YUC1 (Cheng et al. 2007). At the young leaf stage, the expression level of the LCR gene was slightly decreased in the curl mutants compared with the WT, but was increased in the mature leaf (Fig. 8C, I). YUC1 expression was also significantly decreased in both the young and mature leaves of the curl mutants (Fig. 8D, J). There was no significant difference in Sldof28 and PNH at either stage (Fig. 8B, H, E, K). The SlDof25 expression level was increased in the curl mutants at the mature leaf stage (Fig. 8G), while there was no significant change at the young leaf stage (Fig. 8A). It has been reported that the Arabidopsis activation tagging mutant Dof5.1-D exhibited an upward curling leaf phenotype by promoting Revoluta (Rev) transcription (Kim et al. 2010). Rev is an adaxial specification gene (Emery et al. 2003, Prigge et al. 2005) and, most importantly, in tomato it has also been reported that overexpression of a microRNA166-resistant version of SlREV (35S::REVRis) showed an upward curly leaf phenotype (Hu et al. 2014). The expression of SlDof 25 and SlRev was consistent with these findings (Fig. 8G, L). Fig. 8 View largeDownload slide Relative expression of auxin-related genes which were reported to control leaf flatness, observed by qRT-PCR at young and mature leaf stages. (A–E) Relative expression of genes at the young leaf stage: (A) SlDof25, (B) SlDof28, (C) SlLCR, (D) SlYUC1, (E) SlPNH and (F) the adaxial specification gene SlRev. (G–L) Relative expression of the genes at the mature leaf stage, when the leaves were completely curly: (G) SlDof25, (H) SlDof28, (I) SlLCR, (J) SlYUC1, (K) SlPNH and (L) the adaxial specification gene SlRev. Values are means ± SE (n = 3). The asterisks represent statistically significant differences in means with equal variants compared with the wild-type (WT) based on Student’s t-test (*P<0.05, **P<0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Fig. 8 View largeDownload slide Relative expression of auxin-related genes which were reported to control leaf flatness, observed by qRT-PCR at young and mature leaf stages. (A–E) Relative expression of genes at the young leaf stage: (A) SlDof25, (B) SlDof28, (C) SlLCR, (D) SlYUC1, (E) SlPNH and (F) the adaxial specification gene SlRev. (G–L) Relative expression of the genes at the mature leaf stage, when the leaves were completely curly: (G) SlDof25, (H) SlDof28, (I) SlLCR, (J) SlYUC1, (K) SlPNH and (L) the adaxial specification gene SlRev. Values are means ± SE (n = 3). The asterisks represent statistically significant differences in means with equal variants compared with the wild-type (WT) based on Student’s t-test (*P<0.05, **P<0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Discussion The SlLAX1 gene is responsible for the curly leaf phenotype in tomato We characterized several alleles of tomato mutants exhibiting severe upward curling leaf phenotypes at the mature leaf stage (Fig. 1A, B). This mutant phenotype occurred irrespective of water content or relative humidity (Fig. 1E, G;Supplementary Table S1). Six lines were isolated using a forward genetic approach by visually selecting curly leaf phenotypes in a previously generated tomato mutant population (Saito et al. 2011, Shikata et al. 2016). Map-based cloning combined with ES revealed that the mutation occurred in the SlLAX1 (Solyc09g014380) gene (Figs. 3, 4). Then, to validate the candidate gene, we utilized TILLING to obtain an additional allelic line with a nonsense mutation, curl-7, which was generated by EMS. The curl-7 mutant leaves displayed similar curly leaves to the other curl mutants (Fig. 6B). Furthermore, we confirmed the full-length coding sequence of SlLAX1 (Fig. 6C, D), which supported the evidence that SlLAX1 is the gene responsible for the curly leaf phenotype. Taken together, the characterization of multiple alleles in this study that consistently showed similar phenotypes is strong evidence for the role of SlLAX1 in controlling the curly leaf phenotype. To our knowledge, this study is the first example of the successful application of ES in tomato for the identification of a causal gene preceded by a forward genetic approach. SlLAX1 encodes a transmembrane amino acid transporter protein and belongs to the amino acid/auxin permease (AAAP) family. Homology searches indicated that the SlLAX1 protein sequence is homologous to Arabidopsis thaliana AtAUX1 (AT2G38120). In Arabidopsis, AUX1 is one of four auxin influx carriers belonging to the AUX1/LAX family that controls several developmental processes including gravitropism responses, venation patterns and lateral roots (Bennett et al. 1996, Vieten et al. 2007). Although recent findings have indicated that the AUX/LAX1 family also control aerial part development such as leaf serration (Kasprzewska et al. 2015), phyllotaxis patterning, vascular patterning and xylem differentiation (Bainbridge et al. 2008, Fàbregas et al. 2015), the role of the AUX1/LAX gene family in leaf curling is poorly understood. In contrast, mutations in many auxin-related genes produced an impaired leaf flatness phenotype (Zgurski et al. 2005, Esteve-Bruna et al. 2013). In tomato, a few studies have shown a relationship between auxin and leaf flatness; for instance, SlARF4-RNAi (RNA interference) lines produce hyponastic leaves (Sagar et al. 2013) and SlPIN4-RNAi lines show leaf flatness defects as well as altered plant architecture (Pattison and Catalá 2012). However, the role of SlLAX1 in controlling the leaf curly phenotype has not been reported in tomato or other major crops. In tomato, the AUX1/LAX family consists of five genes (SlLAX1–SlLAX5). They share high identity and similarity; the identity of SlLAX2, SlLAX3, SlLAX4 and SlLAX5 with SlLAX1 is 80.36, 79.70, 92.65 and 80.87%, respectively (Sol Genomics Network). All SlLAX genes are expressed in the mature leave and root of tomato (Pattison and Catalá 2012). The single mutants depleted in SlLAX1 used in this study, curl-1–curl-7, showed a severe phenotype effect in leaf flatness, suggesting the importance of SlLAX1 in controlling leaf flatness in mature leaves. Although the functional redundancy of the AUX1/LAX family, in addition to the function of SlLAX1 itself, is poorly characterized in tomato, their function in Arabidopsis is well characterized particularly in root development. Although four AUX1/LAX genes share high sequence identity and similarity, AtAUX1 has the strongest auxin influx activity (Péret et al. 2012, Rutschow et al. 2014). Péret et al. (2012) also reported that subfunctionalization of the AUX1/LAX family in root is based on their distinct pattern of spatial expression and their subcellular localization. In contrast, the AUX1/LAX genes have redundant roles in the context of phyllotaxy, vascular patterning and xylem differentiation (Bainbridge et al. 2008, Fàbregas et al. 2015). Therefore, the functional redundancy of the SlLAX gene family in the tomato leaf curling phenotype awaits further investigation. Loss of function of SlLAX1 protein is related to the curly leaf phenotype Based on experimental evidence, Swarup et al. (2004) reported that Arabidopsis AUX1 protein has 11 transmembrane (TM) helixes. Using a publicly available server, we checked the prediction of TM helices in the SlLAX1 protein. According to a prediction program in http://www.cbs.dtu.dk/services/TMHMM/, both AtAUX1 and SlLAX1 (Supplementary Fig. S5A) have 10 TM helixes. The curl-2 and curl-6 mutants (Fig. 5B) carry a nonsense mutation which is located in TM helix VII (Supplementary Fig. S5C) according to TMHMM, which is equivalent to the central region of AtAUX1 and has proven to be particularly important for protein function (Swarup et al. 2004). In addition, both curl-1 and curl-3 mutations (Fig. 5D) are located in TM helix IV (Supplementary Fig. S5B), which is in a similar part of the N-terminal half of AtAUX1 and is essential for its correct localization (Péret et al. 2012). The curl-7 mutant has only five TM helixes, losing the other five (Supplementary Fig. S5D). Furthermore, the curl-1/curl-3, curl-2/curl-6 and curl-7 mutations were nonsense mutations that can produce only 35, 63 and 45% of the WT protein, respectively (Figs. 5B, D, 6D). Additionally, the relative expression of the curl mutant alleles (curl-1, curl-2 and curl-6) was <40% compared with the WT (Fig. 5F). These reasons presumably account for the loss-of-function mutations of the SlLAX1 gene. To test the potential function of SlLAX1 as an auxin transporter, we first measured the leaf endogenous auxin content. However, IAA content was comparable between the WT and the curl mutants at all stages (Supplementary Fig. S3). Numerous findings have indicated that AtAUX1 plays an important role in root gravitropism and lateral root development (Bennett et al. 1996, Marchant et al. 1999). The root gravitropism response is also commonly used to check auxin response and distribution. Therefore, we next carried out these assays and found that the root gravitropism response of the curl mutants was affected by the SlLAX1 mutation. In addition, lateral root emergence was also disrupted (Supplementary Fig. S4). Although the functional characterization of SlLAX1 has not been conducted in tomato and we do not yet have direct evidence in this study, agravitropism and lateral root formation defects of the curl mutants indicated that SlLAX1 may have a potential function as an auxin transporter similar to AtAUX1, and SlLAX1 might participate in local auxin distribution without affecting the total endogenous auxin content of the whole leaf. Functional analysis of the SlLAX1 gene remains to be carried out. The curly leaf phenotype of the curl mutants is presumably caused by an imbalance of pavement cell enlargement between the adaxial and abaxial sides The curly leaf phenotype was not observed at the early stage of leaf development (Fig. 1C, D;Supplementary Fig. S1), and is not related to relative humidity and water availability (Fig. 1E, G). Thus, we hypothesized that the curly leaf phenotype was caused by alteration of the adaxial/abaxial cell ratio rather than an impairment in adaxial–abaxial polarity since adaxial–abaxial polarity is established at the very early stage of leaf development, i.e. at the primordium stage. As expected, pavement cell size in the abaxial side in the curl mutants was significantly larger compared with that of the WT, while there was no significant difference in the adaxial side. The number of pavement cells in adaxial and abaxial sides was comparable. The upward curling of the curl mutants might be explained by the differential growth of pavement cells in the adaxial and abaxial cell surfaces, which is supported by a similar observation of incurvata6 (icu6), the semi-dominant allele of AUXIN RESISTANT3 (AXR3), that showed an upward curly phenotype caused by a reduced adaxial/abaxial cell size ratio (Pérez-Pérez et al. 2010). The imbalance in epidermal adaxial–abaxial cell growth which led to either epinastic (downward curvature) or hyponastic leaves is not a new phenomenon. It was previously reported that an auxin hyperaccumulation plant produced leaf epinastic curvature, formed due to an increased growth of the leaf adaxial side (Klee et al. 1987, Romano et al. 1993, Kim et al. 2007), which was induced by reduced auxin export that may cause its hyperaccumulation on the adaxial side. Taken together, SlLAX1 might have a function not in the establishment of adaxial–abaxial polarity but rather in balancing the adaxial/abaxial cell size ratio in later stages of leaf development. The evaluation of auxin distribution and/or analysis of SlLAX1 gene expression on the adaxial and abaxial leaf surfaces should allow for a better understanding of the SlLAX1 function in this process. According to the relative expression of some tomato putative auxin-related genes controlling leaf flatness, SlYuc1 showed prominent changes in both young and mature leaves of the curl mutants. YUC is a family of genes that are orthologs to ToFZY (Expósito-Rodríguez et al. 2007), which has a function in local auxin biosynthesis (Zhao et al. 2001). In a previous study it was reported that aux1 and yuc mutants in Arabidopsis have a synergistic effect to enhance each other in order to control leaf development. Lower expression of SlYuc1 does not change leaf auxin content presumably because SlYuc is a family of genes (Expósito-Rodríguez et al. 2007). Other SlYuc genes may compensate total auxin biosynthesis, resulting in comparable auxin contents in the entire leaf. In Arabidopsis, activation tagging of AtDof5.1 resulted in an upward curly leaf phenotype (Kim et al. 2010). Dof5.1 was demonstrated to promote Rev gene expression by binding to its promoter. Similar to these finding (Kim et al. 2010), expression of SlDof25, an ortholog of Dof5.1, was increased in all the curl mutants (Fig. 8G). The SlRev expression level was also increased (Fig. 8L). Up-regulation of AtDof5.1 also repressed transcript levels of auxin biosynthesis genes, which is consistent with a low expression level of SlYuc1 in the curl mutants (Fig. 8D, J). In tomato, it has also been reported that overexpression of a microRNA166-resistant version of SlREV (35S::REVRis) showed an upward curly leaf phenotype (Hu et al. 2014). Collectively, our findings are similar to previous findings which reinforce the partial disturbance of auxin homeostasis in the SlLAX1 mutants. The fact that lower auxin content triggers cell expansion is well established (Ishida et al. 2010, reviewed in Velasquez et al. 2016). We hypothesize that the loss of function of SlLAX1 in the curl mutants results in imbalanced adaxial/abaxial pavement cell growth leading to the curly leaf phenotype. Depletion of SlLAX1 in the curl mutants disrupts auxin transport in either the adaxial or abaxial leaf surface. Given that there was no significant difference in the adaxial pavement cell size (Table 6; Fig. 7A), SlLAX1 action appears to be restricted to the abaxial side. SlLAX1 belongs to the SlLAX family, and other members are known to be expressed in leaves (Pattison and Catalá 2012). It is possible that other influx transporters compensate for the loss of function of SlLAX1 in the adaxial side. In contrast, our data suggest that SlLAX1 is a major determinant of auxin transporter which is dominant in the abaxial side. In the adaxial side, where SlLAX1 is not a major auxin influx carrier, presumably the auxin content of the curl mutants was similar to or higher than that of the WT, while a decreased auxin content in the abaxial side was due to low influx carrier activity. Auxin, which is not taken up by abaxial cells, may be accumulated in the adaxial side, or alternatively accumulated in the extracellular space (most probably the latter because the cell number in the adaxial and abaxial sides was comparable, meaning that there was no increase in cell division in the adaxial side). Therefore, the auxin content in the curl mutants could be maintained at a similar level to the WT (Supplementary Fig. S3). Imbalanced adaxial/abaxial cell growth due to differential auxin accumulation is also well established (Pérez-Pérez et al. 2010, reviewed in Sandalio et al. 2016). We speculate that the lower auxin content on the abaxial cell surface triggers cell expansion and imbalanced cell growth on both surfaces, leading to the emergence of curly leaves. This hypothesis awaits further investigation. In brief, this study contributes to the newly characterized role of SlLAX1 in controlling leaf development in tomato by balancing the adaxial–abaxial pavement cell enlargement potentially mediated by auxin. The evaluation of auxin distribution and/or analysis of SlLAX1 gene expression on the adaxial and abaxial leaf surfaces should allow for a better understanding of the SlLAX1 function in this process. Additionally, analysis of double mutants with other LAX or PIN family members and other adaxial–abaxial specification genes would be helpful to dissect the precise mechanism of SlLAX1 in normal leaf development in plants. Materials and Methods Plant material and growth conditions Tomato (Solanum lycopersicum cv. ‘Micro-Tom’) curly leaf (curl) mutants were generated by EMS and γ-ray irradiation. The mutants were obtained from the National BioResources Project (NBRP) Project at the University of Tsukuba (Saito et al. 2011, Shikata et al. 2016). From the M3 mutagenized population, we isolated six lines of the curly leaf phenotype mutants, herein referred to as ‘curl’ mutants. The mutant screening was carried out visually using mature plants showing severe curly leaf phenotypes. Five mutant alleles, curl-1–curl-5, were generated by γ-ray irradiation, and one mutant allele, curl-6, was generated by EMS mutagenesis. Furthermore, using TILLING screening, we screened another EMS mutant, curl-7. These mutants were registered in the TOMATOMA mutant database (Saito et al. 2011, http://tomatoma.nbrp.jp/). The NBRP accession numbers are listed in Supplementary Table S6. Unless otherwise stated, further analyses of the curl mutants were conducted after two backcrosses to the WT ‘Micro-Tom’ to remove any possible background mutation following the mutagenesis treatment. The plants were grown under standard cultivation conditions in the greenhouse facility at the University of Tsukuba. Genomic DNA extraction, construction of the mapping population, DNA markers and genetic analysis Genomic DNA was extracted from 2-month-old plants. A maximum of 100 mg of fresh leaf sample was extracted using a Maxwell® 16 Tissue DNA Purification Kit (Promega). To perform rough mapping using DNA markers, curl-2 was crossed to another tomato cultivar, ‘Ailsa Craig’, to obtain a mapping population. From approximately 100 plants of the F2 mapping population, 19 plants exhibiting the curly mutant phenotype were isolated, and genomic DNA was extracted from the leaves of the individual plants. These plants were subjected to rough mapping experiments. All SNP and CAPS DNA markers were designed according to the AMF2 (F2: Solanum lycopersicum ‘Ailsa Craig’×S. lycopersicum ‘Micro-Tom’) linkage map information that is publicly available from the Kazusa DNA Research Institute (KDRI) webpage (http://marker.kazusa.or.jp/Tomato/; Shirasawa et al. 2010). The primers and restriction enzyme used in rough mapping chromosomes are listed in Supplementary Table S7. Exome sequencing and variant identification ES was performed to narrow down the candidate genes. Four alleles, curl-1, curl-2, curl-3 and curl-6, of the curl mutants of the F2 mutant population backcrossed to the WT were used. The mutants and WT phenotypes were selected in the F2 population based on the presence or absence of curly leaves among approximately 100 F2 plants for each line, after which their DNA samples were bulked based on phenotype. ES analysis was then performed based on the Roche exome sequence SeqCap® EZ SR protocol (http://sequencing.roche.com/). Briefly, genomic DNA was treated with a Covaris® S220 Ultrasonicator (Covaris) to achieve an average length of 200 bp. Then, a multiplex next-generation sequencing (NGS) library was constructed using a KAPA® Library Preparation Kit and SeqCap® adaptor kit (Roche). After constructing the NGS library, exome capture was conducted using a custom probe set that was designed based on the tomato genome reference version SL2.50 (supporting dataset, Sol Genomics Network, https://solgenomics.net). This probe set was designed to capture 49.5 Mb of exonic DNA regions (Supplementary Data S1). The resultant exome library was amplified by 14 cycles of post-capture ligation-mediated PCR with KAPA HiFi HostStart ReadyMix (Roche) and then subjected to Illumina HiSeq-2000 sequencing set to the 100 bp paired-end mode. Paired-end short read data were subjected to quality filtering using the FASTXToolkit with the parameters of –Q 20 –P 90. Then, short reads were aligned to the tomato genome reference version SL2.50 using bowtie2 software with the following parameters: L, 0,-0.16 –mp 2, 2 –np 1 –rdg 1, 1 –rfg 1, 1. On average, 98.8 ± 0.03% of the target exonic regions were covered by short reads. The average read depth was 18 ± 1.5. Genome-wide DNA polymorphisms and mutations were identified based on the alignment results by the HaplotypeCaller function of the Genome Analysis Toolkit (GATK) with the following parameters: -mmq 5 -forceActive -stand_call_conf 10 -stand_emit_conf 10. The resultant DNA variant information was further combined into one genomic VCF data set with the GenotypeGVCFs function of the GATK. Three wild-type ES data sets [DNA Data Bank of Japan (DDBJ) accession Nos. DRR097500–DRR097502], two WT whole-genome NGS data sets (DDBJ accession Nos. DRR097503 and DRR097504) and one publicly available WT whole-genome NGS data set (Kobayashi et al. 2014) were used as controls to remove intracultivar variations that are present between WT ‘Micro-Tom’ lines. DNA variants were further removed if their allele frequencies was >90% in the WT F2 bulked segregants because they were also expected to be intracultivar variations. Those variants with <20% allele frequency or with a read depth <6 were also removed because they were likely to be false positives. ES data sets for curl mutants are available in the DDBJ (accession Nos DRR097492–DRR097502). RNA extraction and cDNA synthesis Total RNA was extracted from young and mature leaves (when the leaves were completely curly) using an RNeasy Mini Kit (QIAGEN) according to the manufacturer’s protocol. To remove genomic DNA contamination, two steps were applied: an on-column RNase-free DNase Set (QIAGEN) and an RNA Clean & Concentrator™-5 (Zymo Research). Subsequently, cDNA was synthesized from 2,000 ng of total RNA by a SuperScript III First Strand Synthesis Kit (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. Cloning and sequencing of the full-length coding sequence of the SlLAX1 gene The full-length coding sequence (1,236 bp) of the SlLAX1 gene from three independent plants was amplified by PCR. The primer sequences are listed in Supplementary Table S4. Subsequently, PCR products were loaded onto a 0.8–1.5% agarose gel, which was then electrophoresed for 45–60 min. Next, the band was visualized under 70% UV and then cut either with a gel cutter or a blade. Any visible desired product band was individually cut, removed and subsequently subjected to purification using a Wizard® SV Gel and PCR Clean-Up System (Promega). DNA purification by centrifugation was applied. The purified PCR product was then cloned into the entry vector pCR8/GW/TOPO (Invitrogen, http://www.lifetechnologies.com/) using an In-Fusion® HD Cloning Kit (TAKARA BIO INC.) according to the manufacturer’s protocol. Then, plasmids from clones were purified using a FastGene Plasmid Mini Kit (Nippon Genetics). The plasmid fragments were sequenced using M13 primers (Supplementary Table S4). qRT-PCR analysis The mRNA expression level was quantified using qRT-PCR. A 10 ng µl−1 cDNA template of three biological replicates was used for gene expression analyses. The SlActin gene was used as an internal control (Løvdal and Lillo 2009). qRT-PCR was carried out using a CFX96 Real-Time System (Bio-Rad) with SYBR Premix ExTaq II (Ili RNase H Plus; TAKARA BIO INC.). The primers used for qRT-PCR are listed in Supplementary Table S4. Relative gene expression was quantified using the ΔΔCT method (Pfaffl 2001). The qRT-PCR mixture and thermal cycle conditions were as described by Shinozaki et al. (2015). The primers for qRT-PCR were designed using the Primer3 Plus website (http://primer3plus.com/); two exons in the forward or reverse primer were joined to exclude any possibility of contamination of genomic DNA. Screening new SlLAX1 mutant alleles by TILLING The TILLING population was previously described by Okabe et al. (2013), and the TILLING experiments were performed as described by Okabe et al. (2011). We attempted to screen for mutations in the coding region of the SlLAX1 gene. The primer pair was designed to span exon 6. Given that exon 6 is the longest exon, we also identified an EMS mutant line, curl-6, that carries a nonsense mutation in exon 6 of SlLAX1. The primer pair used in the TILLING experiment was forward 5'-TGGTACATGGGAACTAGCTAAGCC-3' and reverse 5'-ACCTGACGAGCGGATGATTTTC-3', which amplified 865 bp of genomic DNA template; the 5' end of each primer was labeled with DY-681 or DY-781, which are equivalent to IRDye 700 or IRDye 800 (https://www.biomers.net/), respectively. Morphological analysis The CI of mutants was measured on the fifth leaflet in accordance with the method introduced by Liu et al. (2010). Leaf area and perimeter analyses were conducted at the young and mature leaf stages; 15 leaves harvested from the same position were used as samples. Leaf images were captured using a digital camera, and the leaf area and perimeter were measured using CellSensStandard imaging software (Olympus). The leaf perimeter and leaf area were measured by following the edge of the leaf using a closed polygon measurement tool within the CellSensStandard software. The reduction in leaf area and leaf perimeter (%) was measured by comparing the values before and after flattening (multiplied by 100). SEM The leaf epidermal surface was observed using a scanning electron microscope (Hitachi Tabletop Microscope TM3000). The cell features were measured at the mature leaf stage when the leaves were completely curly, and precisely in the same regions on the adaxial and abaxial surfaces. Mature fresh leaves were sampled and flattened before being subjected to microscopic observation. Approximately 0.5×0.5 cm2 of adaxial or abaxial surface was placed into a sample box, after which the epidermal pavement cell was imaged at ×400 magnification for at least three biological replications. The cell size was quantified separately using CellSensStandard software. All measurements were obtained for at least three independently captured SEM images for each replication and three fields of view for each image. For quantification of the number of pavement cells, leaf samples were cut from midway exactly between the midrib and the margin of fully curly leaves. We used precisely the same position on both the adaxial and abaxial sides; one side was used for adaxial pavement cell observation and the other was used for the abaxial pavement cells. An approximatley 2–4 mm leaf sample in the tip area of the transversal axis was cut irrespective of the size from the midrib to the margin, and it was subjected to SEM (Supplementary Fig. S6). The cell number was counted throughly in that region. Measurements were obtained from three biological replications. Measurement of the auxin content in leaves Leaves were sampled at three stages from the same positions in (i) young leaves, before curly leaves formed; (ii) when leaves just turned curly; and (iii) mature leaves, after leaves were fully curly. Three biological replications were included at each stage. At least 100 mg of fresh leaves was immediately frozen in liquid nitrogen and crushed into a fine powder using a TissueLyser (Qiagen). Endogenous auxin was measured using a UHPLC-Q-Exactive (Thermo Fisher Scientific) system. Measurements were conducted as described by Kojima et al. (2009) and Shinozaki et al. (2015). Leaf water potential measurements Leaf water potential was measured using a pressure chamber. A leaflet from the same position was cut and immediately placed into the chamber. Pressure was gradually increased until water was exuded from the petiole. Six biological samples were tested. Statistical analysis Unless otherwise stated, the data are presented as the mean ± SE. Student’s t-test (at the 95% and 99% significance levels) was used to analyze the significance level between two values with equal variance. χ2 tests were performed using MS Excel 2016 to examine the goodness of fit between the expected and observed Mendelian ratio in the segregating F2 population of mutants backcrossed to the WT ‘Micro-Tom’, and the degrees of freedom and expected Mendelian ratio used for monogenic traits were 1 and 3:1 (WT:mutant phenotype), respectively. Supplementary Data Supplementary data are available at PCP online. Funding This work was supported by the Japan Society for the Promotion of Science (JSPS) [KAKENHI Grant-in-Aid for Research Activity start-up (15H06071 to R.Y.)]; Program to Disseminate Tenure Tracking System [to T.A.]; and the Japan Advanced Plant Science Network [to H.E. and H.S.]. Acknowledgments Tomato ‘Micro-Tom’ and the curl mutant seeds were obtained from the National BioResource Project (NBRP), Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. We are grateful for helpful comments and discussion of the manuscript from Dr. Kentaro Ezura. We would like to express our sincere gratitude to all our laboratory members for their great support and helpful discussion throughout the work. Disclosures The authors have no conflicts of interest to declare. References Ariizumi T., Kishimoto S., Kakami R., Maoka T., Hirakawa H., Suzuki Y., et al.  . ( 2014) Identification of the carotenoid modifying gene PALE YELLOW PETAL 1 as an essential factor in xanthophyll esterification and yellow flower pigmentation in tomato (Solanum lycopersicum). Plant J.  79: 453– 465. Google Scholar CrossRef Search ADS PubMed  Bainbridge K., Guyomarc’h S., Bayer E., Swarup R., Bennett M., Mandel T., et al.  . ( 2008) Auxin influx carriers stabilize phyllotactic patterning. Genes Dev . 22: 810– 823. Google Scholar CrossRef Search ADS PubMed  Bennett M.J., Marchant A., Green H.G., May S.T., Ward S.P., Millner P.A., et al.  . ( 1996) Arabidopsis AUX1 gene: a permease-like regulator of root gravitropism. Science  273: 948– 950. Google Scholar CrossRef Search ADS PubMed  Bennett M.J., Marchant A., May S.T., Swarup R. ( 1998) Going the distance with auxin: unravelling the molecular basis of auxin transport. Philos. Trans. R. Soc. B: Biol. Sci  353: 1511– 1515. Google Scholar CrossRef Search ADS   Cai X., Zhang Y., Zhang C., Zhang T., Hu T., Ye J., et al.  . ( 2013) Genome-wide analysis of plant-specific Dof transcription factor family in tomato. J. Integr. Plant Biol.  55: 552– 566. Google Scholar CrossRef Search ADS PubMed  Cheng Y., Dai X., Zhao Y. ( 2007) Auxin synthesized by the YUCCA flavin monooxygenases is essential for embryogenesis and leaf formation in Arabidopsis. Plant Cell  19: 2430– 2439. Google Scholar CrossRef Search ADS PubMed  Chusreeaeom K., Ariizumi T., Asamizu E., Okabe Y., Shirasawa K., Ezura H. ( 2014) A novel tomato mutant, Solanum lycopersicum elongated fruit1 (Slelf1), exhibits an elongated fruit shape caused by increased cell layers in the proximal region of the ovary. Mol. Genet. Genomics  289: 399– 409. Google Scholar CrossRef Search ADS PubMed  Delbarre A., Muller P., Imhoff V., Guern J. ( 1996) Comparison of mechanisms controlling uptake and accumulation of 2,4-dichlorophenoxy acetic acid, naphthalene-1-acetic acid, and indole-3-acetic acid in suspension-cultured tobacco cells. Planta  198: 532– 541. Google Scholar CrossRef Search ADS PubMed  Emery J.F., Floyd S.K., Alvarez J., Eshed Y., Hawker N.P., Izhaki A., et al.  . ( 2003) Radial patterning of Arabidopsis shoots by class III HD-ZIP and KANADI genes. Curr. Biol . 13: 1768– 1774. Google Scholar CrossRef Search ADS PubMed  Esteve-Bruna D., Perez-Perez J.M., Ponce M.R., Micol J.L. ( 2013) incurvata13, a novel allele of AUXIN RESISTANT6, reveals a specific role for auxin and the SCF complex in Arabidopsis embryogenesis, vascular specification, and leaf flatness. Plant Physiol . 161: 1303– 1320. Google Scholar CrossRef Search ADS PubMed  Expósito-Rodríguez M., Borges A.A., Borges-Pérez A., Hernández M., Pérez J.A. ( 2007) Cloning and biochemical characterization of ToFZY, a tomato gene encoding a flavin monooxygenase involved in a tryptophan-dependent auxin biosynthesis pathway. J. Plant Growth Regul.  26: 329– 340. Google Scholar CrossRef Search ADS   Fàbregas N., Formosa-Jordan P., Confraria A., Siligato R., Alonso J.M., Swarup R., et al.  . ( 2015) Auxin influx carriers control vascular patterning and xylem differentiation in Arabidopsis thaliana. PLoS Genet.  11: e1005183. Google Scholar CrossRef Search ADS PubMed  Hao S., Ariizumi T., Ezura H. ( 2017) SEXUAL STERILITY is essential for both male and female gametogenesis in tomato. Plant Cell Physiol.  58: 22– 34. Google Scholar CrossRef Search ADS PubMed  Hu G., Fan J., Xian Z., Huang W., Lin D., Li Z. ( 2014) Overexpression of SlREV alters the development of the flower pedicel abscission zone and fruit formation in tomato. Plant Sci . 229: 86– 95. Google Scholar CrossRef Search ADS PubMed  Ishida T., Adachi S., Yoshimura M., Shimizu K., Umeda M., Sugimoto K. ( 2010) Auxin modulates the transition from the mitotic cycle to the endocycle in Arabidopsis. Development  137: 63– 71. Google Scholar CrossRef Search ADS PubMed  Kasprzewska A., Carter R., Swarup R., Bennett M., Monk N., Hobbs J.K., et al.  . ( 2015) Auxin influx importers modulate serration along the leaf margin. Plant J.  83: 705– 718. Google Scholar CrossRef Search ADS PubMed  Kim H.S., Kim S.J., Abbasi N., Bressan R.A., Yun D.J., Yoo S.D., et al.  . ( 2010) The DOF transcription factor Dof5.1 influences leaf axial patterning by promoting Revoluta transcription in Arabidopsis. Plant J . 64: 524– 535. Google Scholar CrossRef Search ADS PubMed  Kim J.I., Sharkhuu A., Jin J.B., Li P., Jeong J.C., Baek D., et al.  . ( 2007) yucca6, a dominant mutation in Arabidopsis, affects auxin accumulation and auxin-related phenotypes. Plant Physiol . 145: 722– 735. Google Scholar CrossRef Search ADS PubMed  Klee H.J., Horsch R.B., Hinchee M.A., Hein M.B., Hoffmann N.L. ( 1987) The effects of overproduction of two Agrobacterium tumefaciens T-DNA auxin biosynthetic gene products in transgenic petunia plants. Genes Dev . 1: 86– 96. Google Scholar CrossRef Search ADS   Kobayashi M., Nagasaki H., Garcia V., Just D., Bres C., Mauxion L.P., et al.  . ( 2014) Genome-wide analysis of intraspecific DNA polymorphism in ‘Micro-Tom’, a model cultivar of tomato (Solanum lycopersicum). Plant Cell Physiol . 55: 445– 454. Google Scholar CrossRef Search ADS PubMed  Kojima M., Kamada-Nobusada T., Komatsu H., Takei K., Kuroha T., Mizutani M., et al.  . ( 2009) Highly sensitive and high-throughput analysis of plant hormones using MS-probe modification and liquid chromatography–tandem mass spectrometry: an application for hormone profiling in Oryza sativa. Plant Cell Physiol . 50: 1201– 1214. Google Scholar CrossRef Search ADS PubMed  Kramer E.M., Bennett M.J. ( 2006) Auxin transport: a field in flux. Trends Plant Sci.  11: 382– 386. Google Scholar CrossRef Search ADS PubMed  Lee C., Chronis D., Kenning C., Peret B., Hewezi T., Davis E.L., et al.  . ( 2011) The novel cyst nematode effector protein 19C07 interacts with the Arabidopsis auxin influx transporter LAX3 to control feeding site development. Plant Physiol . 155: 866– 880. Google Scholar CrossRef Search ADS PubMed  Liu Z., Jia L., Mao Y., He Y. ( 2010) Classification and quantification of leaf curvature. J. Exp. Bot.  61: 2757– 2767. Google Scholar CrossRef Search ADS PubMed  Liu Z., Jia L., Wang H., He Y. ( 2011) HYL1 regulates the balance between adaxial and abaxial identity for leaf flattening via miRNA-mediated pathways. J. Exp. Bot . 62: 4367– 4381. Google Scholar CrossRef Search ADS PubMed  Løvdal T., Lillo C. ( 2009) Reference gene selection for quantitative real-time PCR normalization in tomato subjected to nitrogen, cold, and light stress. Anal. Biochem . 387: 238– 242. Google Scholar CrossRef Search ADS PubMed  Marchant A., Bhalerao B., Casimiro I., Eklöf J., Casero P.J., Bennett M., et al.  . ( 2002) AUX1 promotes lateral root formation by facilitating indole-3-acetic acid distribution between sink and source tissues in the Arabidopsis seedling. Plant Cell  14: 589– 597. Google Scholar CrossRef Search ADS PubMed  Marchant A., Kargul J., May S.T., Muller P., Delbarre A., Perrot-Rechenmann C., et al.  . ( 1999) AUX1 regulates root gravitropism in Arabidopsis by facilitating auxin uptake within root apical tissues. EMBO J  18: 2066– 2073. Google Scholar CrossRef Search ADS PubMed  Newman K.L., Fernandez A.G., Barton M.K. ( 2002) Regulation of axis determinacy by the Arabidopsis PINHEAD gene. Plant Cell  14: 3029– 3042. Google Scholar CrossRef Search ADS PubMed  Okabe Y., Ariizumi T., Ezura H. ( 2013) Updating the Micro-Tom TILLING platform. Breed. Sci.  63: 42– 48. Google Scholar CrossRef Search ADS PubMed  Okabe Y., Asamizu E., Saito T., Matsukura C., Ariizumi T., Bres C., et al.  . ( 2011) Tomato TILLING technology: development of a reverse genetics tool for the efficient isolation of mutants from Micro-Tom mutant libraries. Plant Cell Physiol . 52: 1994– 2005. Google Scholar CrossRef Search ADS PubMed  Paciorek T., Zažímalová E., Ruthardt N., Petrášek J., Stierhof Y.-D., Kleine-Vehn J., et al.  . ( 2005) Auxin inhibits endocytosis and promotes its own efflux from cells. Nature  435: 1251– 1256. Google Scholar CrossRef Search ADS PubMed  Pattison R.J., Catalá C. ( 2012) Evaluating auxin distribution in tomato (Solanum lycopersicum) through an analysis of the PIN and AUX/LAX gene families. Plant J . 70: 585– 598. Google Scholar CrossRef Search ADS PubMed  Péret B., Rybel B.D., Casimiro I., Benkova E., Swarup R., Laplaze L., et al.  . ( 2009) Arabidopsis lateral root development: an emerging story. Trends Plant Sci . 14: 399– 408. Google Scholar CrossRef Search ADS PubMed  Péret B., Swarup K., Ferguson A., Seth M., Yang Y., Dhondt S., et al.  . ( 2012) AUX/LAX genes encode a family of auxin influx transporters that perform distinct functions during Arabidopsis development. Plant Cell  24: 2874– 2885. Google Scholar CrossRef Search ADS PubMed  Pérez-Pérez J.M., Candela H., Robles P., López-Torrejón G., Del Pozo J.C., Micol J.L. ( 2010) A role for AUXIN RESISTANT3 in the coordination of leaf growth. Plant Cell Physiol . 51: 1661– 1673. Google Scholar CrossRef Search ADS PubMed  Pfaffl M.W. ( 2001) A new mathematical model for relative quantification in real-time RT–PCR. Nucleic Acids Res.  29: e45– e4521. Google Scholar CrossRef Search ADS PubMed  Prigge M.J., Otsuga D., Alonso J.M., Ecker J.R., Drews G.N., Clark S.E. ( 2005) Class III homeodomain-leucine zipper gene family members have overlapping, antagonistic, and distinct roles in Arabidopsis development. Plant Cell  17: 61– 76. Google Scholar CrossRef Search ADS PubMed  Romano C.P., Cooper M.L., Klee H.J. ( 1993) Uncoupling auxin and ethylene effects in transgenic tobacco and Arabidopsis plants. Plant Cell  5: 181– 189. Google Scholar CrossRef Search ADS PubMed  Rutschow H.L., Baskin T.I., Kramer E.M. ( 2014) The carrier AUXIN RESISTANT (AUX1) dominates auxin flux into Arabidopsis protoplasts. New Phytol.  204: 536– 544. Google Scholar CrossRef Search ADS PubMed  Sagar M., Chervin C., Mila I., Hao Y., Roustan J.-P., Benichou M., et al.  . ( 2013) SlARF4, an auxin response factor involved in the control of sugar metabolism during tomato fruit development. Plant Physiol . 161: 1362– 1374. Google Scholar CrossRef Search ADS PubMed  Saito T., Ariizumi T., Okabe Y., Asamizu E., Hiwasa-Tanase K., Fukuda N., et al.  . ( 2011) TOMATOMA: a novel tomato mutant database distributing Micro-Tom mutant collections. Plant Cell Physiol . 52: 283– 296. Google Scholar CrossRef Search ADS PubMed  Sandalio L.M., Godiguez-Serrano M., Romero-Puertas M.C. ( 2016) Leaf epinasty and auxin: a biochemical and molecular overview. Plant Sci.  253: 187– 193. Google Scholar CrossRef Search ADS PubMed  Serrano-Cartagena J., Robles P., Ponce M.R., Micol J.L. ( 1999) Genetic analysis of leaf form mutants from the Arabidopsis Information Service collection. Mol. Gen. Genet . 261: 725– 739. Google Scholar CrossRef Search ADS PubMed  Shikata M., Hoshikawa K., Ariizumi T., Fukuda N., Yamazaki Y., Ezura H. ( 2016) TOMATOMA update: phenotypic and metabolite information in the Micro-Tom mutant resource. Plant Cell Physiol.  57: e11. Google Scholar CrossRef Search ADS PubMed  Shinozaki Y., Hao S., Kojima M., Sakakibara H., Ozeki-Iida Y., Zheng Y., et al.  . ( 2015) Ethylene suppresses tomato (Solanum lycopersicum) fruit set through modification of gibberellin metabolism. Plant J.  83: 237– 251. Google Scholar CrossRef Search ADS PubMed  Shirasawa K., Isobe S., Hirakawa H., Asamizu E., Fukuoka H., Just D., et al.  . ( 2010) SNP discovery and linkage map construction in cultivated tomato. DNA Res.  17: 381– 391. Google Scholar CrossRef Search ADS PubMed  Song J.B., Huang S.Q., Dalmay T., Yang Z.M. ( 2012) Regulation of leaf morphology by microRNA394 and its target LEAF CURLING RESPONSIVENESS. Plant Cell Physiol . 53: 1283– 1294. Google Scholar CrossRef Search ADS PubMed  Swarup K., Benková E., Swarup R., Casimiro I., Péret B., Yang Y., et al.  . ( 2008) The auxin influx carrier LAX3 promotes lateral root emergence. Nat. Cell Biol.  10: 946– 954. Google Scholar CrossRef Search ADS PubMed  Swarup R., Kargul J., Marchant A., Zadik D., Rahman A., Mills R., et al.  . ( 2004) Structure–function analysis of the presumptive Arabidopsis auxin permease AUX1. Plant Cell  16: 3069– 3083. Google Scholar CrossRef Search ADS PubMed  Swarup R., Péret B. ( 2012) AUX/LAX family of auxin influx carriers—an overview. Front. Plant Sci.  3: 1– 11. Google Scholar CrossRef Search ADS PubMed  Tromas A., Perrot-Rechenmann C. ( 2010) Recent progress in auxin biology. C. R. Biol.  333: 297– 306. Google Scholar CrossRef Search ADS PubMed  Velasquez S.M., Elke B., Jürgen K., José M.E. ( 2016) Auxin and cellular elongation. Plant Physiol . 70: 1206– 1215. Vieten A., Sauer M., Brewer P.B., Friml J. ( 2007) Molecular and cellular aspects of auxin-transport-mediated development. Trends Plant Sci . 12: 160– 168. Google Scholar CrossRef Search ADS PubMed  Yang H., Murphy A.S. ( 2009) Functional expression and characterization of Arabidopsis ABCB, AUX 1 and PIN auxin transporters in Schizosaccharomyces pombe. Plant J . 59: 179– 191. Google Scholar CrossRef Search ADS PubMed  Yu L., Yu X., Shen R., He Y. ( 2005) HYL1 gene maintains venation and polarity of leaves. Planta  221: 231– 242. Google Scholar CrossRef Search ADS PubMed  Zgurski J.M., Sharma R., Bolokoski D.A., Schultz E.A. ( 2005) Asymmetric auxin response precedes asymmetric growth and differentiation of asymmetric leaf1 and asymmetric leaf2 Arabidopsis leaves. Plant Cell  17: 77– 91. Google Scholar CrossRef Search ADS PubMed  Zhao Y., Christensen S.K., Fankhauser C., Cashman J.R., Cohen J.D., Weigel D., et al.  . ( 2001) A role for flavin monooxygenase-like enzymes in auxin biosynthesis. Science  291: 306– 309. Google Scholar CrossRef Search ADS PubMed  Abbreviations Abbreviations CAPS cleaved amplified polymorphic sequence CI curvature index curl curly leaf EMS ethyl methanesulfonate ES exome sequencing LAX like AUX1 NGS next-generation sequencing PAT polar auxin transport qRT-PCR quantitative real-time PCR SEM scanning electron microscopy SNP single nucleotide polymorphism TILLING Targeting Induced Local Lesions IN Genome TM transmembrane UTR untranslated region WT wild type © The Author(s) 2018. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Plant and Cell Physiology Oxford University Press

SlLAX1 is Required for Normal Leaf Development Mediated by Balanced Adaxial and Abaxial Pavement Cell Growth in Tomato

Loading next page...
 
/lp/ou_press/sllax1-is-required-for-normal-leaf-development-mediated-by-balanced-HC4MbHIzyz
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: journals.permissions@oup.com
ISSN
0032-0781
eISSN
1471-9053
D.O.I.
10.1093/pcp/pcy052
Publisher site
See Article on Publisher Site

Abstract

Abstract Leaves are the major plant organs with a primary function for photosynthesis. Auxin controls various aspects of plant growth and development, including leaf initiation, expansion and differentiation. Unique and intriguing auxin features include its polar transport, which is mainly controlled by the AUX1/LAX and PIN gene families as influx and efflux carriers, respectively. The role of AUX1/LAX genes in root development is well documented, but the role of these genes in leaf morphogenesis remains unclear. Moreover, most studies have been conducted in the plant model Arabidopsis thaliana, while studies in tomato are still scarce. In this study, we isolated six lines of the allelic curly leaf phenotype ‘curl’ mutants from a γ-ray and EMS (ethyl methanesulfonate) mutagenized population. Using a map-based cloning strategy combined with exome sequencing, we observed that a mutation occurred in the SlLAX1 gene (Solyc09g014380), which is homologous to an Arabidopsis auxin influx carrier gene, AUX1 (AtAUX1). Characterization of six alleles of single curl mutants revealed the pivotal role of SlLAX1 in controlling tomato leaf flatness by balancing adaxial and abaxial pavement cell growth, which has not been reported in tomato. Using TILLING (Targeting Induced Local Lesions IN Genome) technology, we isolated an additional mutant allele of the SlLAX1 gene and this mutant showed a curled leaf phenotype similar to other curl mutants, suggesting that Solyc09g014380 is responsible for the curl phenotype. These results showed that SlLAX1 is required for normal leaf development mediated by balanced adaxial and abaxial pavement cell growth in tomato. Introduction Leaves are the major plant organs whose primary function involves photosynthesis. Leaves play a major role in sensing the quality, quantity and duration of light, all of which are crucial for complete plant growth and development. Understanding leaf initiation and development is important in plant biology. Most leaves are dorsoventrally (upper to bottom) flattened and develop distinct upper (adaxial) and lower (abaxial) surfaces. Balanced co-ordination of polarity, auxin response and cell division is essential for formation and development of normal and flat leaves. Any imbalance of this co-ordination results in altered leaf shapes such as curly, crinkly, twisted, rolled, radial or shrunken leaves (Serrano-Cartagena et al. 1999, Yu et al. 2005, Liu et al. 2010, Liu et al. 2011). The formation of flat leaves enables the optimum capture of sunlight during photosynthesis. An important factor controlling leaf morphogenesis is the phytohormone auxin. Indole-3-acetic acid (IAA) is the natural form of auxin that controls various aspects of plant growth and development, including cell division, expansion and differentiation, leaf initiation and morphogenesis. One of the unique and intriguing features of auxin is its transport (Paciorek et al. 2005, Tromas and Perrot-Rechenmann 2010). It is known that auxin is synthesized in young leaves and in the shoot apex, and is transported basipetally to all plant organs (reviewed in Bennett et al. 1998, Tromas and Perrot-Rechenmann 2010). Auxin transport involves two patterns: long-distance transport through phloem and short-distance or cell to cell transport called polar auxin transport (PAT). At the cellular level, IAA is distributed through a combination of membrane diffusion (passive uptake), carrier-mediated uptake and proton-driven distribution (Delbarre et al. 1996). PAT contributes to 85% of short-distance auxin transport. It is well established that polar auxin localization controls the direction of auxin movement in whole-plant organs. Several auxin carriers have been identified, including AUX1/LAX (LAX: like AUX1), PIN (PIN-FORMED) and PGP/MDR (P-glycoprotein/multidrug resistance)-like proteins. AUX1/LAX is reported to be an auxin influx carrier that facilitates auxin movement from outside to inside the cell, while PIN is an efflux carrier that pumps auxin from the cell into the intercellular space. PGP/MDR-like proteins are reported to have the ability to be either influx or efflux carriers (Yang and Murphy 2009), but the contribution of these proteins is considerably small compared with that of the AUX1/LAX and PIN families (Kramer and Bennett 2006, reviewed in Swarup and Péret 2012). There are numerous studies highlighting the effects of mutations in the AUX/LAX gene family in the model plant Arabidopsis. However, most studies have focused on root phenotypes. For instance, the AUX1/LAX family has been reported to promote lateral root emergence and formation (Marchant et al. 2002, Swarup et al. 2008, reviewed in Péret et al. 2009), root gravitropism (Bennett et al. 1996, Marchant et al. 1999) and root–pathogen interactions (Lee et al. 2011). Recently, AUX1 function in the aerial parts of plants has received interest, but studies are still considerably scarce. In Arabidopsis, AUX1 has been reported to control phyllotaxis patterning (Bainbridge et al. 2008), vascular patterning, xylem differentiation (Fàbregas et al. 2015) and leaf serration (Kasprzewska et al. 2015). Additionally, although PAT is governed and maintained by the co-ordinated action of AUX1/LAX and PIN carrier proteins, among auxin carriers PIN1 is the most studied. The role of the PIN protein family in leaf morphogenesis is well documented, yet the role of AUX1/LAX remains neglected or is underestimated. Furthermore, almost all studies have been carried out in the model plant Arabidopsis, while the role of auxin influx carriers in other model plants such as tomato is poorly understood. In this study, we isolated six lines of curly leaf (curl) mutants from the ‘Micro-Tom’ mutant population that had been previously established by γ-ray irradiation and EMS (ethyl methanesulfonate) treatment (Saito et al. 2011, Shikata et al. 2016). The curl mutants showed dorsoventrally impaired leaf flatness, which exhibited severe upward bending or hyponasty on the transverse axis. Through map-based cloning combined with exome sequencing (ES), we characterized six alleles of the curly leaf mutants, which have a nonsense mutation in the SlLAX1 gene. We reported that the SlLAX1 gene controls the curly leaf phenotype in the tomato curl mutants. This feature has never been characterized. The characterization of several alleles of single curl mutants in this study sheds light on the pivotal role of SlLAX1 in controlling leaf flatness mediated by normal adaxial–abaxial pavement cell growth. We also combined forward and reverse genetic approaches to validate the candidate gene. Using TILLING (Targeting Induced Local Lesions IN Genome) technology, we screened another nonsense mutant allele that consistently shows a similar curly leaf phenotype to that of the curl mutants obtained by a forward genetic approach. Results Isolation and phenotypic characterization of the curly leaf mutants We previously developed a large mutant population of ‘Micro-Tom’, a model tomato cultivar, using γ-ray irradiation and EMS mutagenesis (Saito et al. 2011, Shikata et al. 2016). Currently, we have 9,216 EMS mutant lines. From the M3 generation of this mutant population, we isolated six mutant lines exhibiting a severe upward curly leaf (hyponastic) phenotype; three lines were used for further analysis (Fig. 1A, B). The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the wild type (WT) (Fig. 1C, D), suggesting that the impairment of leaf curvature was not detectable at the early vegetative stage. The leaves became curly about 1 month after sowing and were continuously curly until the end of the growing period. The initiation of curly leaves was not related to the transition from the vegetative to the reproductive stage, and the leaf phenotype could not be restored at any stage once the curly leaves had formed. Growing the curl mutants in a high-humidity environment in in vitro culture could not rescue the curly phenotype (Fig. 1E). Additionally, curly leaves continuously appeared irrespective of water availability in the soil medium (Fig. 1F, G). The leaf water potential of the mutants and WT was also comparable (Supplementary Table S1). These data suggested that the curly leaf mutant phenotype is persistent, irrespective of relative humidity or water availability. Fig. 1 View largeDownload slide Leaf morphology of the WT ‘Micro-Tom’ and three alleles of the curl mutants. (A, B) Mature leaf morphology of mature curl mutants in the (A) adaxial and (B) abaxial view. The leaf images were captured from 2-month-old plants from the fifth leaflet. Scale bar = 2 cm. (C, D) Appearance of the young leaves of curl mutants. (C) Adaxial and (D) abaxial view. The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the WT. Scale bar = 1 cm. (E) Representative of the curl mutant (curl-1) when grown in in vitro culture. Scale bar = 2 cm. The curly phenotype was not restored. (F, G) Wild-type (F) and representative curl mutant (G, curl-1) grown under well-watered conditions in the greenhouse. Plant images were captured from 2-month-old plants. Scale bar = 1.5 cm. Fig. 1 View largeDownload slide Leaf morphology of the WT ‘Micro-Tom’ and three alleles of the curl mutants. (A, B) Mature leaf morphology of mature curl mutants in the (A) adaxial and (B) abaxial view. The leaf images were captured from 2-month-old plants from the fifth leaflet. Scale bar = 2 cm. (C, D) Appearance of the young leaves of curl mutants. (C) Adaxial and (D) abaxial view. The newly developed young leaves of the curl mutants were flat and indistinguishable from those of the WT. Scale bar = 1 cm. (E) Representative of the curl mutant (curl-1) when grown in in vitro culture. Scale bar = 2 cm. The curly phenotype was not restored. (F, G) Wild-type (F) and representative curl mutant (G, curl-1) grown under well-watered conditions in the greenhouse. Plant images were captured from 2-month-old plants. Scale bar = 1.5 cm. We also analyzed the percentage of reduced leaf area and perimeter in both young and mature leaves of mutants by flattening the curl mutant leaves. In the young leaves, leaf area was markedly reduced (41.0–56.0%, Table 1). The leaf perimeters of the WT and mutants were comparable. Consistently, in the mature leaves, the reduction in leaf area was more evident (55.8–64.0%) (Table 1), indicating a progression of severity that was concomitant with leaf maturity. Then, to investigate how and when the curly leaf is formed and its progression at the organ level, we measured the curvature index (CI) in both young and mature leaves according to the method of Liu et al. (2010). Negative curvature represents upward bending of the leaf. At the early stage of leaf initiation and development, mutants developed and maintained flat leaves; after several (4–7) days following leaf initiation, the leaves gradually became curly, and the curly leaf severity increased concomitant with leaf maturity (Supplementary Fig. S1A–C). The leaf incurvature was initiated from young leaves firstly at the tip along the transversal axis to a low extent, while the longitudinal axis remained flat in all mutant lines (Table 2, Fig. 2A, B). To understand the curly leaf progression, the global curvature of mature leaves of all mutants was also measured (Table 2). Consistently, leaf incurvature was observed along the transversal axis to a high extent and the longitudinal axis remained normal at all stages of leaf development. In the mature leaves, the whole leaf had become curly (Fig. 2C, D). These data suggested that leaf incurvature was more severe as leaf maturity progressed. Table 1 Leaf area and leaf perimeter of young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf area and perimeter were observed when the curly leaf was being formed, about 6 d after leaf initiation. The mature leaf leaf area and perimeter were observed when the leaf had become completely curly, about 10 d after leaf initiation. Values are means ± SE (n = 15). The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01); WT mean values were used as controls. The reductions in leaf area and leaf perimeter (%) were measured by comparing the values before and after flattening (multiplied by 100). Table 1 Leaf area and leaf perimeter of young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf  Mature leaf  Line  Leaf area  Leaf perimeter  Leaf area  Leaf perimeter  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  Before flattening (mm2)  After flattening (mm2)  Reduction (%)  Before flattening (mm)  After flattening (mm)  Reduction (%)  WT  685.2 ± 47.7  662.0 ± 40.4  3.5  125.7 ± 5.6  124.6 ± 5.7  0.9  1489.2 ± 63.2  1,440.4 ± 57.3  3.4  188.1 ± 4.3  190.0 ± 5.8  –1.0  curl-1  397.1 ± 54.7**  694.3 ± 45.4  –42.8  119.3 ± 3.7  118.9 ± 3.9  0.3  530.4 ± 72.9**  1,471.8 ± 77.1  –64.0  175.0 ± 7.9  191.3 ± 7.8  –8.6  curl-2  289.3 ± 54.3**  664.9 ± 24.1  –56.5  130.1 ± 3.1  129.4 ± 2.4  0.5  314.9 ± 70.4**  1,362.1 ± 98.0  –76.9  180.6 ± 5.9  190.7 ± 7.2  –5.3  curl-6  316.6 ± 27.7**  649.7 ± 36.1  –51.3  121.1 ± 5.3  118.4 ± 3.5  2.3  697.2 ± 81.6**  1,575.5 ± 122.7  –55.8  189.9 ± 9.2  214.7 ± 11.3  –11.6  Young leaf area and perimeter were observed when the curly leaf was being formed, about 6 d after leaf initiation. The mature leaf leaf area and perimeter were observed when the leaf had become completely curly, about 10 d after leaf initiation. Values are means ± SE (n = 15). The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01); WT mean values were used as controls. The reductions in leaf area and leaf perimeter (%) were measured by comparing the values before and after flattening (multiplied by 100). Table 2 The global curvature index of the young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Values are means ± SE (n = 15). The curvature index (CI) of mutants was measured by a method introduced by Liu et al. (2010). CI = (ab – a’b’)/ab. ab = the distance between points a and b on two margins of curvature before flattening of leaves. a’b’ = the distance between a and b on two margins after flattening. The CI was measured in the middle of the leaves. The flatness of either young or mature leaf curl mutants was impaired along the transverse axis, whereas the longitudinal axis was normal. A negative CI represents upward curvature. The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01). Table 2 The global curvature index of the young and mature leaves of the curl mutants Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Young leaf  Mature leaf  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  Line  Direction  Axis  Transverse CI  Longitudinal CI  Extent  WT  Flat  –  0.0 ± 0.0  0.0 ± 0.0  –  WT  Flat  –  0.0 ± 0.0  0.00 ± 0.0  –  curl-1  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-1  Upward  Transverse  –0.7 ± 0.2**  –0.02 ± 0.0  High  curl-2  Upward  Transverse  –0.2 ± 0.0**  0.0 ± 0.0  Low  curl-2  Upward  Transverse  –0.8 ± 0.2**  0.00 ± 0.0  High  curl-6  Upward  Transverse  –0.3 ± 0.0**  0.0 ± 0.0  Low  curl-6  Upward  Transverse  –0.8 ± 0.2**  –0.01 ± 0.0  High  Values are means ± SE (n = 15). The curvature index (CI) of mutants was measured by a method introduced by Liu et al. (2010). CI = (ab – a’b’)/ab. ab = the distance between points a and b on two margins of curvature before flattening of leaves. a’b’ = the distance between a and b on two margins after flattening. The CI was measured in the middle of the leaves. The flatness of either young or mature leaf curl mutants was impaired along the transverse axis, whereas the longitudinal axis was normal. A negative CI represents upward curvature. The asterisks represent statistically significant differences in means with equal variants based on Student’s t-test (**P < 0.01). Fig. 2 View largeDownload slide Adaxial and abaxial surfaces of young (upper panel) and mature (bottom panel) tomato leaflets. (A) Adaxial (upper) surface of young tomato leaflets. (B) Abaxial (bottom) surface of young tomato leaflets. Young leaflets were detached from 1.5-month-old plants. (C) Adaxial (upper) surface of mature tomato leaflets. (D) Abaxial (bottom) surface of mature tomato leaflets. Mature leaflets were detached from the fifth leaflet of 2.5-month-old plants. Scale bar: upper panel = 3 cm; bottom panel = 2 cm. Fig. 2 View largeDownload slide Adaxial and abaxial surfaces of young (upper panel) and mature (bottom panel) tomato leaflets. (A) Adaxial (upper) surface of young tomato leaflets. (B) Abaxial (bottom) surface of young tomato leaflets. Young leaflets were detached from 1.5-month-old plants. (C) Adaxial (upper) surface of mature tomato leaflets. (D) Abaxial (bottom) surface of mature tomato leaflets. Mature leaflets were detached from the fifth leaflet of 2.5-month-old plants. Scale bar: upper panel = 3 cm; bottom panel = 2 cm. Genetic mapping of curl mutants To examine the inheritance pattern of the curl mutants, we crossed the mutants with the WT ‘Micro-Tom’ and another tomato cultivar ‘Ailsa Craig’, and observed the segregation ratio in the F2 population. Phenotypic observation was carried out visually according to the presence or absence of the curly leaf phenotype. The mutant phenotype appeared in the F2 generation only as a recessive genetic trait (Table 3). The ratio of WT and mutant phenotypes fit to the Mendelian segregation ratio for monogenic traits (3:1), indicating a monogenic recessive inheritance of all curl mutants. Similarly, in the ‘Ailsa Craig’ background, the inheritance of the curl mutants was also recessive (Supplementary Table S2). An allelism test was performed to observe complementation effects among mutant lines and to examine whether mutations occurred because of the same causal gene. The complementation effect was determined in the F1 generation. All crosses between each pair of mutant lines showed curly leaf phenotypes (Table 4), indicating that they are allelic, which means that causal mutation occurred in the same locus. The curl-6 mutant generated from EMS treatment (see ‘Plant material and growth conditions’) was also allelic with the other mutants which were generated from γ-ray irradiation. We confirmed that all curl mutant lines were allelic; therefore, for further analyses, we used only three mutant alleles, namely curl-1, curl-2 and curl-6. Table 3 Segregation analysis of the curl mutants back-crossed to the wild-type ‘Micro-Tom’ Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  a The curl mutants were crossed to the wild-type ‘Micro-Tom’. b The number of progeny exhibiting normal (WT) and curly leaf phenotype is shown. c χ2 value was calculated based on progeny segregation in the F2 population. d χ2 distribution in the table reference value, with probability >0.05 and degree of freedom 1. e Inheritance pattern of the curl mutants, estimated based on the χ2 value at the 95% (P < 0.05) significance level. Table 3 Segregation analysis of the curl mutants back-crossed to the wild-type ‘Micro-Tom’ Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  Mutant linea  F1b WT:curly  F2b WT:curly  χ2 valuec  χ2 referenced  P-value  Inheritance patterne  curl-1  4:0  105:25  2.30  3.84  0.13  Monogenic recessive  curl-2  1:0  79:31  0.59  3.84  0.44  Monogenic recessive  curl-3  5:0  70:25  0.08  3.84  0.76  Monogenic recessive  curl-6  2:0  123:30  2.37  3.84  0.12  Monogenic recessive  a The curl mutants were crossed to the wild-type ‘Micro-Tom’. b The number of progeny exhibiting normal (WT) and curly leaf phenotype is shown. c χ2 value was calculated based on progeny segregation in the F2 population. d χ2 distribution in the table reference value, with probability >0.05 and degree of freedom 1. e Inheritance pattern of the curl mutants, estimated based on the χ2 value at the 95% (P < 0.05) significance level. Table 4 The result of the allelism test among the curl mutants Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    The allelism test was carried out by crossing all possible pairs and observing the results in the F1 generation visually by the presence of a curly leaf phenotype. Normal represents the wild-type phenotype; Curly represents the curly leaf phenotype; ND, not determined; ♀, ♂, female recipient and male donor, respectively. Table 4 The result of the allelism test among the curl mutants Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    Mutant line ♀  Mutant line ♂  WT  curl-1  curl-2  curl-3  curl-4  curl-6  WT    Normal  Normal  Normal  Normal  Normal  curl-1  Normal    ND  ND  ND  Curly  curl-2  Normal  Curly    ND  ND  Curly  curl-3  Normal  ND  Curly    Curly  Curly  curl-4  Normal  ND  Curly  Curly    Curly  curl-6  Normal  Curly  Curly  Curly  Curly    The allelism test was carried out by crossing all possible pairs and observing the results in the F1 generation visually by the presence of a curly leaf phenotype. Normal represents the wild-type phenotype; Curly represents the curly leaf phenotype; ND, not determined; ♀, ♂, female recipient and male donor, respectively. To identify the candidate gene controlling the curly leaf phenotype, we performed a map-based cloning approach using PCR-based DNA markers including CAPS (cleaved amplified polymorphic sequence) and SNPs (single nucleotide polymorphisms) (Shirasawa et al. 2010, Ariizumi et al. 2014, Chusreeaeom et al. 2014, Hao et al. 2017). We found that the mutation probably occurred in the short arm of chromosome 9 (Supplementary Table S3). The highest ‘Micro-Tom’ allele frequency was observed in this chromosome region between markers tomInf5375 and 14109_151, and ranged from 0.68 to 0.89, suggesting that the gene responsible could be localized in the short arm of chromosome 9 close to marker 14109_151 (physical position SL2.40ch09:2052389, Fig. 3). Fig. 3 View largeDownload slide Partial chromosome mapping result of the curl mutant locus. The curl locus was found to associate with the marker 14109-151 on chromosome 9 in the F2 mapping population derived from the cross between S. lycopersicum cv. ‘Ailsa Craig’×S. lycopersicum cv. ‘Micro-Tom’ curl-2. The marker information was obtained from the Kazusa DNA Research Institute AMF2 database (http://marker.kazusa.or.jp/). No such association was observed in other chromosomes (Supplementary Table S3). Fig. 3 View largeDownload slide Partial chromosome mapping result of the curl mutant locus. The curl locus was found to associate with the marker 14109-151 on chromosome 9 in the F2 mapping population derived from the cross between S. lycopersicum cv. ‘Ailsa Craig’×S. lycopersicum cv. ‘Micro-Tom’ curl-2. The marker information was obtained from the Kazusa DNA Research Institute AMF2 database (http://marker.kazusa.or.jp/). No such association was observed in other chromosomes (Supplementary Table S3). The SlLAX1 gene is commonly mutated in several curl mutant alleles To narrow down the candidate region obtained by rough mapping, we performed ES. Four lines of the curl mutants, curl-1, curl-2, curl-3 and curl-6, were used for the ES analysis. The F2 progeny derived from the cross between mutant and WT ‘Micro-Tom’ were divided into flat leaf and curly phenotype based on the presence or absence of the curly leaf phenotype, and then flat leaf and mutant bulked segregants were subjected to ES. By the bowtie2-GATK pipeline using the tomato genome reference version SL2.50 as a reference (see the Materials and Methods), we identified 5,430, 5,110, 5,050 and 4,829 genome-wide mutations for curl-1, curl-2, curl-3 and curl-6 mutant segregants, respectively. When allele frequencies were compared between these mutants, a strong association was found around the top region of chromosome 9 in all four mapping populations (Fig. 4). This result suggested that the causal gene for the curly phenotype is located in this chromosome region, in agreement with the result of rough mapping of chromosoms (Fig. 3; Supplementary Table S3). Furthermore, we then searched for the gene in which mutation commonly occurs in some of the curl mutants. We found that mutations commonly occur in Solyc09g014380.2.1, which is a homolog of Arabidopsis AtAUX1 (AT2G38120; BLASTx E-value = 0.0, protein amino acid similarity = 93%). The Solyc09g014380.2.1, tomato locus SlLAX1, gene spans an approximately 3.8 kb genomic region, while cDNA including the untranslated region (UTR) spans 1.8 kb. The SlLAX1 gene has seven exons, including a UTR in both the 5' and 3' ends (Fig. 4). The curl-2 and curl-6 mutants had a nucleotide substitution from G to A in exon 6, physical position SL2.50ch09: 6,010,739 bp (Table 5). This SNP produced a premature stop codon (W262*) in the deduced protein sequence of SlLAX1. According to the SL2.50 tomato genome reference, WT ‘Micro-Tom’ produced 411 amino acids of the SlLAX1 protein, whereas the curl-2 and curl-6 mutants produced only 261 amino acids of the protein, losing the last 150 amino acids (63.7% of the WT protein). curl-1 and curl-3 had an SNP from G to T in the splicing junction of intron 4, physical position SL2.50ch09: 600,292 bp. These mutations were also confirmed by dideoxy sequencing of cDNA (Fig. 5A, B). Table 5 Predicted mutation position, amino acid substitution and mutation type based on the whole-exome sequence result Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  a The location in the chromosome in the tomato genome. b The position of the nucleotide substitution according to the tomato genome sequence database, version SL2.50 (Sol Genomics Network). c Tomato genome sequence reference according to the position in the second column. d Alternative nucleotide sequence/nucleotide substitution according to the position in the second column. e Location of nucleotide substitution of the gene in the sixth column. f Gene mutated according to the Sol Genomic Network database. g *A stop codon. Table 5 Predicted mutation position, amino acid substitution and mutation type based on the whole-exome sequence result Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  Chromosomea  Positionb (bp)  REF nucc  ALT nucd  Withine  Genef  Strand  Amino acid substitution  Mutation type  Arabidopsis homolog  Arabidopsis homolog name  curl mutant allele  SL2.50ch09  6010739  G  A  Exon 6  Solyc09g014380.2.1  Plus  W262*g  Nonsense  AT2G38120.1  AtAUX1  curl-2, curl-6  SL2.50ch09  6009292  G  T  Intron4  Solyc09g014380.2.1  Plus  –  Intron  AT2G38120.1  AtAUX1  curl-1, curl-3  a The location in the chromosome in the tomato genome. b The position of the nucleotide substitution according to the tomato genome sequence database, version SL2.50 (Sol Genomics Network). c Tomato genome sequence reference according to the position in the second column. d Alternative nucleotide sequence/nucleotide substitution according to the position in the second column. e Location of nucleotide substitution of the gene in the sixth column. f Gene mutated according to the Sol Genomic Network database. g *A stop codon. Fig. 4 View largeDownload slide Identification of SlLAX1 (Solyc09g01480.2) as the most plausible candidate gene responsible for the curl phenotype. Genome-wide allele frequency data were obtained by exome sequencing of BCF2 bulked segregants that show the curl mutant phenotype. To narrow down candidates efficiently, four mapping populations derived from independent curl alleles (curl-1, 2, 3 and 6) were constructed and subjected to exome sequencing. In all four mapping populations, a strong association was commonly observed for mutations within the SlLAX1 (Solyc09g01480.2) gene, which is a homolog of the Arabidopsis AUXIN RESISTANT1 (AUX1) transporter gene. Black boxes indicate exons, transparent boxes indicate UTRs, and lines between boxes indicate introns. Fig. 4 View largeDownload slide Identification of SlLAX1 (Solyc09g01480.2) as the most plausible candidate gene responsible for the curl phenotype. Genome-wide allele frequency data were obtained by exome sequencing of BCF2 bulked segregants that show the curl mutant phenotype. To narrow down candidates efficiently, four mapping populations derived from independent curl alleles (curl-1, 2, 3 and 6) were constructed and subjected to exome sequencing. In all four mapping populations, a strong association was commonly observed for mutations within the SlLAX1 (Solyc09g01480.2) gene, which is a homolog of the Arabidopsis AUXIN RESISTANT1 (AUX1) transporter gene. Black boxes indicate exons, transparent boxes indicate UTRs, and lines between boxes indicate introns. Fig. 5 View largeDownload slide Changes in protein amino acid sequence and SlLAX1 gene expression in curl mutants. (A, B) A partial alignment of the SlLAX1 cDNA sequence (A) or deduced protein amino acid sequence (B) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-2 and curl-6. The mutation in curl-2 and curl-6 causes a premature stop codon, as shown by the red box (W262*). (C, D) A partial alignment of SlLAX1 cDNA sequence (C) or deduced protein amino acid sequence (D) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-1 and curl-3. cDNA sequences were obtained by dideoxy sequencing (A, C). (E) Donor and acceptor splicing sites in intron 4 of the wild type, curl-1 and curl-3 mutants. Square brackets indicate splicing sites. Double square brackets indicate the alternative splicing site in the curl-1 and curl-3 mutants. The one-letter code indicates an amino acid. Upper case indicates an exon, whereas lower case indicates an intron sequence. The bold letter indicates a mutated sequence in intron 4 of the curl-1 and curl-3 mutants. The asterisk represents the stop codon in curl-1. (F) qRT-PCR analysis of SlLAX1 gene expression. qRT-PCR primers were designed to target downstream of the stop codon mutation in exon 6. The asterisks represent statistically significant differences in the mean with equal variants compared with the wild type (WT) based on Student’s t-test (**P < 0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Fig. 5 View largeDownload slide Changes in protein amino acid sequence and SlLAX1 gene expression in curl mutants. (A, B) A partial alignment of the SlLAX1 cDNA sequence (A) or deduced protein amino acid sequence (B) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-2 and curl-6. The mutation in curl-2 and curl-6 causes a premature stop codon, as shown by the red box (W262*). (C, D) A partial alignment of SlLAX1 cDNA sequence (C) or deduced protein amino acid sequence (D) among the tomato reference (SL2.50), wild-type Micro-Tom, curl-1 and curl-3. cDNA sequences were obtained by dideoxy sequencing (A, C). (E) Donor and acceptor splicing sites in intron 4 of the wild type, curl-1 and curl-3 mutants. Square brackets indicate splicing sites. Double square brackets indicate the alternative splicing site in the curl-1 and curl-3 mutants. The one-letter code indicates an amino acid. Upper case indicates an exon, whereas lower case indicates an intron sequence. The bold letter indicates a mutated sequence in intron 4 of the curl-1 and curl-3 mutants. The asterisk represents the stop codon in curl-1. (F) qRT-PCR analysis of SlLAX1 gene expression. qRT-PCR primers were designed to target downstream of the stop codon mutation in exon 6. The asterisks represent statistically significant differences in the mean with equal variants compared with the wild type (WT) based on Student’s t-test (**P < 0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. As described above, the curl-1 and curl-3 mutants had a mutation in the first nucleotide or splicing junction of intron 4 (Fig 4; Table 5). Interestingly, sequencing of SlLAX1 cDNA in these alleles revealed that abnormal splicing occurred around intron 4, which led to deletion of five nucleotides within exon 4 (nucleotides 433–437, Fig. 5C). Given that mutation in curl-1 and curl-3 is a G to T substitution in the splicing junction of intron 4, presumably there was an alteration in the donor and recipient sites for intron splicing. Splicing of intron 4 occurred 435 bp from the start codon in the tomato genome of the WT, whereas intron splicing occurred 5 bp upstream of the end of exon 4 (430 bp from the start codon) in both the curl-1 and the curl-3 alleles. Then the next sequence following exon 5 is GGTTGA; this TGA may produce a premature stop codon, which is at 435 bp from the start codon (Fig. 5E). Thus, curl-1 and curl-3 alleles could produce a C-terminal truncated SlLAX1 protein that is only 145 amino acids in length (Fig. 5D). We also analyzed the transcript level of the SlLAX1 by quantitative real-time PCR (qRT-PCR) using mature curly leaf cDNA. The expression of the SlLAX1 gene in the three curly leaf mutants was significantly reduced to only 35–40% of WT expression (Fig. 5F), which indicates low abundance of this gene transcript in the mutants. Taken together, these results indicated that all of curl mutants carried a loss-of-function mutation in the SlLAX1 gene. Screening a new allele of the nonsense mutation of SlLAX1 by TILLING Because our research group had previously developed large mutant resources in the ‘Micro-Tom’ background and proved that TILLING is an efficient tool for isolating desired mutants from the ‘Micro-Tom’ mutant collection (Okabe et al. 2011), we utilized TILLING to search for other SlLAX1 mutant alleles. We screened 4,608 lines in the M2 and M3 generations to obtain new SlLAX1 mutant alleles. In addition, because we only had one EMS mutant screened by forward genetics (curl-6), we attempted to obtain other mutant alleles to confirm consistency of the phenotype. We designed a primer pair to amplify 865 bp along exon 6 of the SlLAX1 gene for the TILLING screening target, and found five new mutant alleles that carried intron, missense and nonsense mutations (Supplementary Fig. S2A, B; Supplementary Table S5). The curl-6/TOMJPE8506, which was previously isolated by forward genetics, was also confirmed by TILLING screening. Then, to validate the mutant phenotype, one line that carried a nonsense mutation, TOMJPW601-1, was renamed ‘curl-7’ and used for further analysis. This mutant line carried a 1 bp substitution from G to A in the 554th nucleotide from the start codon, which led to the conversion of tryptophan to a premature stop codon at the 185th amino acid (Fig. 6A). The curl-7 mutant exhibited the curly leaf phenotype like the other curl mutant alleles (Fig. 6B). Furthermore, by dideoxy sequencing, we confirmed the consistency of the TILLING result (Fig. 6C, D). This result supports the evidence that SlLAX1 is the gene responsible for the curly leaf phenotype in tomato. These results again indicated that mutation in SlLAX1 produces the curly leaf phenotype. Mutations in the same gene consistently resulted in the same phenotype, strongly suggesting that SlLAX1 functions in controlling the tomato curly leaf phenotype. Fig. 6 View largeDownload slide TILLING screening results and confirmation of the presence of the curly leaf phenotype, and cDNA and amino acid sequence alignment of the new mutant allele, TOMJPW601-1/curl-7. (A) A polyacrylamide gel image of TILLING screening. The mutation in TOMJPW601-1/curl-7 is shown as an intense spot on the lanes both in IRD-700 (red circle) and in IRD-800 (green circle). A single nucleotide change is shown on the sequence chromatogram (red arrowhead). (B) Whole-plant images of curl-6 (left); a representative of the curl allele obtained using forward genetics; (middle and right) confirmation of the presence of curly leaves in the new selected allele, curl-7, in the M3 generation. Plant images were captured from 2-month-old plants when the curly leaf phenotype progressed. Scale bar = 2 cm. (C) A partial alignment of the SlLAX1 cDNA sequence among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Nucleotide substitution in the curl-7 mutant is shown by gray highlighting. (D) Partial protein amino acid sequence alignment of SlLAX1 (Solyc09g01480.2) among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Mutation in curl-7 led to the conversion of tryptophan to a premature stop codon. The wild type (WT) produced a 411 amino acid product, whereas curl-7 produced only a 185 amino acid product. The premature stop codon is indicated by a red box. Fig. 6 View largeDownload slide TILLING screening results and confirmation of the presence of the curly leaf phenotype, and cDNA and amino acid sequence alignment of the new mutant allele, TOMJPW601-1/curl-7. (A) A polyacrylamide gel image of TILLING screening. The mutation in TOMJPW601-1/curl-7 is shown as an intense spot on the lanes both in IRD-700 (red circle) and in IRD-800 (green circle). A single nucleotide change is shown on the sequence chromatogram (red arrowhead). (B) Whole-plant images of curl-6 (left); a representative of the curl allele obtained using forward genetics; (middle and right) confirmation of the presence of curly leaves in the new selected allele, curl-7, in the M3 generation. Plant images were captured from 2-month-old plants when the curly leaf phenotype progressed. Scale bar = 2 cm. (C) A partial alignment of the SlLAX1 cDNA sequence among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Nucleotide substitution in the curl-7 mutant is shown by gray highlighting. (D) Partial protein amino acid sequence alignment of SlLAX1 (Solyc09g01480.2) among the tomato reference (SL2.50), wild-type ‘Micro-Tom’ and TOMJPW601-1/curl-7. Mutation in curl-7 led to the conversion of tryptophan to a premature stop codon. The wild type (WT) produced a 411 amino acid product, whereas curl-7 produced only a 185 amino acid product. The premature stop codon is indicated by a red box. Endogenous IAA levels and the expression of auxin-related genes in curl mutants As described above, all curl mutants commonly have a mutation in the SlLAX1 gene, which encodes an auxin influx carrier. To test the potential function of SlLAX1 as an auxin transporter in tomato, we measured the leaf auxin content at three stages: (i) in young leaves, before curly leaves formed; (ii) when leaves just became curly; and (iii) in mature leaves, after leaves were fully curly. The IAA content significantly decreased from young leaves to mature leaves in both the WT and three curl mutants (Supplementary Fig. S3A). However, the IAA content at each leaf stage was comparable between the WT and the curl mutants. Similarly, IAA conjugates and total IAA between the WT and the curl mutants were also comparable (Supplementary Fig. S3B, C). In Arabidopsis, numerous findings have indicated the role of the LAX1/AUX1 family in root gravitropism and lateral root formation (Bennett et al. 1996, Marchant et al. 2002, reviewed in Swarup and Péret 2012). Importantly, root agravitropism is the most prominent defect and well-characterized trait of the Arabidopsis aux1 mutant. In addition, the aux1 mutant also showed lateral root formation defects (Marchant et al. 2002). Thus, we further tested these traits in the curl mutants; as expected, the curl mutants showed agravitropism as well as reduced lateral root formation, in agreement with the Arabidopsis aux1 mutant phenotype (Supplementary Fig. S4), suggesting the possibility of involvement of SlLAX1 as an auxin influx carrier in tomato similar to AtAUX1. Abaxial pavement cell size of the curl mutants was significantly larger Because SlLAX1 gene function was commonly disabled in curl mutants and auxin has been known to affect pavement cells (Pérez-Pérez et al. 2010, reviewed in Sandalio et al. 2016), we hypothesized that the curly leaf formation may be related to differential cell growth on the adaxial and abaxial surfaces. To observe histological features of the curl mutants, we measured pavement cell size using scanning electron microscopy (SEM) of the adaxial and abaxial surfaces at the mature leaf stage in the curly part (Table 6; Fig. 7). We noted that cell enlargement in the curl mutants was more prominent on the abaxial side, while there was no significant difference in adaxial pavement cells. As a consequence, the ratio of abaxial and adaxial pavement cells was more prominent in the curl mutants. We also quantified the pavement cell number on both the adaxial and abaxial surfaces. The number of pavement cells on both surfaces was comparable (Table 6). These data revealed that impairment of leaf flatness in the curl mutants is likely to be due to the differential cell growth between the adaxial and abaxial epidermal layers. Most probably, the curly leaf phenotype is related to cell enlargement on the abaxial side. Table 6 Adaxial and abaxial pavement cell size of the curl mutants in the curly part measured by SEM Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Values are means ± SE (n = 9). The asterisks represent statistically significant differences in means with equal variants based on the Student’s t-test (**P < 0.01). The cell features were measured at the mature leaf stage when the leaves were completely curly, precisely in the same regions on the adaxial and abaxial surfaces. The curl mutants showed a significantly larger abaxial/adaxial pavement cell size ratio compared with the wild type (WT). Table 6 Adaxial and abaxial pavement cell size of the curl mutants in the curly part measured by SEM Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Line  Pavement cell size (µm)  Abaxial/adaxial pavement cell size ratio  Adaxial  Abaxial  WT  43.36 ± 2.1  42.11 ± 3.4  0.97  curl-1  36.04 ± 1.8  57.83 ± 6.4**  1.59**  curl-2  36.90 ± 1.2  58.69 ± 4.1**  1.61**  curl-6  38.07 ± 1.8  60.18 ± 1.3**  1.66**  Line  Pavement cell number (cell)  Abaxial/adaxial pavement cell number ratio  Adaxial  Abaxial  WT  1,317.3 ± 49.5  1,110.6 ± 70.8  0.84  curl-1  1,207.5 ± 80.6  1,073.5 ± 65.2  0.89  curl-2  1,389.2 ± 105.2  1,173.9 ± 26.9  0.85  curl-6  1,304.3 ± 73.6  1,156.8 ± 59.6  0.89  Values are means ± SE (n = 9). The asterisks represent statistically significant differences in means with equal variants based on the Student’s t-test (**P < 0.01). The cell features were measured at the mature leaf stage when the leaves were completely curly, precisely in the same regions on the adaxial and abaxial surfaces. The curl mutants showed a significantly larger abaxial/adaxial pavement cell size ratio compared with the wild type (WT). Fig. 7 View largeDownload slide Adaxial and abaxial pavement cells in the wild type (WT) and the curl mutants in the curly part. (A) The adaxial pavement cell size of the WT and mutants was comparable. Scale bar = 20 µm. (B) The pavement cell size of all curl mutants on the abaxial surface was significantly larger compared with that of the WT. Scale bar = 10 µm. (C) The adaxial and abaxial sides of the curly part of the leaf that were subjected to SEM. Images were captured using a scanning electron microscope with ×400 magnification in the curly part at precisely the same position in both the adaxial and abaxial surfaces. Fig. 7 View largeDownload slide Adaxial and abaxial pavement cells in the wild type (WT) and the curl mutants in the curly part. (A) The adaxial pavement cell size of the WT and mutants was comparable. Scale bar = 20 µm. (B) The pavement cell size of all curl mutants on the abaxial surface was significantly larger compared with that of the WT. Scale bar = 10 µm. (C) The adaxial and abaxial sides of the curly part of the leaf that were subjected to SEM. Images were captured using a scanning electron microscope with ×400 magnification in the curly part at precisely the same position in both the adaxial and abaxial surfaces. Relative expression of auxin-related genes in the curl mutants Recently, some studies have reported that impairment of auxin biosynthesis, signaling, degradation and conjugation results in leaf development defects such as wrinkled, curled leaf and rounded leaf phenotypes. We checked the relative expression of some putative tomato auxin-related genes which were reported to be involved in controlling the leaf flatness phenotype such as AtDof5.1 (Kim et al. 2010) which is homologous to SlDof25 and SlDof28 in tomato (Cai et al. 2013), LCR (LEAF CURLING RESPONSIVENESS) (Song et al. 2012), PNH (PINHEAD) (Newman et al. 2002) and YUC1 (Cheng et al. 2007). At the young leaf stage, the expression level of the LCR gene was slightly decreased in the curl mutants compared with the WT, but was increased in the mature leaf (Fig. 8C, I). YUC1 expression was also significantly decreased in both the young and mature leaves of the curl mutants (Fig. 8D, J). There was no significant difference in Sldof28 and PNH at either stage (Fig. 8B, H, E, K). The SlDof25 expression level was increased in the curl mutants at the mature leaf stage (Fig. 8G), while there was no significant change at the young leaf stage (Fig. 8A). It has been reported that the Arabidopsis activation tagging mutant Dof5.1-D exhibited an upward curling leaf phenotype by promoting Revoluta (Rev) transcription (Kim et al. 2010). Rev is an adaxial specification gene (Emery et al. 2003, Prigge et al. 2005) and, most importantly, in tomato it has also been reported that overexpression of a microRNA166-resistant version of SlREV (35S::REVRis) showed an upward curly leaf phenotype (Hu et al. 2014). The expression of SlDof 25 and SlRev was consistent with these findings (Fig. 8G, L). Fig. 8 View largeDownload slide Relative expression of auxin-related genes which were reported to control leaf flatness, observed by qRT-PCR at young and mature leaf stages. (A–E) Relative expression of genes at the young leaf stage: (A) SlDof25, (B) SlDof28, (C) SlLCR, (D) SlYUC1, (E) SlPNH and (F) the adaxial specification gene SlRev. (G–L) Relative expression of the genes at the mature leaf stage, when the leaves were completely curly: (G) SlDof25, (H) SlDof28, (I) SlLCR, (J) SlYUC1, (K) SlPNH and (L) the adaxial specification gene SlRev. Values are means ± SE (n = 3). The asterisks represent statistically significant differences in means with equal variants compared with the wild-type (WT) based on Student’s t-test (*P<0.05, **P<0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Fig. 8 View largeDownload slide Relative expression of auxin-related genes which were reported to control leaf flatness, observed by qRT-PCR at young and mature leaf stages. (A–E) Relative expression of genes at the young leaf stage: (A) SlDof25, (B) SlDof28, (C) SlLCR, (D) SlYUC1, (E) SlPNH and (F) the adaxial specification gene SlRev. (G–L) Relative expression of the genes at the mature leaf stage, when the leaves were completely curly: (G) SlDof25, (H) SlDof28, (I) SlLCR, (J) SlYUC1, (K) SlPNH and (L) the adaxial specification gene SlRev. Values are means ± SE (n = 3). The asterisks represent statistically significant differences in means with equal variants compared with the wild-type (WT) based on Student’s t-test (*P<0.05, **P<0.01). The SlActin gene was used as an internal control. The expression level of the curl-1, curl-2 and curl-6 mutants was relative to the WT expression. Discussion The SlLAX1 gene is responsible for the curly leaf phenotype in tomato We characterized several alleles of tomato mutants exhibiting severe upward curling leaf phenotypes at the mature leaf stage (Fig. 1A, B). This mutant phenotype occurred irrespective of water content or relative humidity (Fig. 1E, G;Supplementary Table S1). Six lines were isolated using a forward genetic approach by visually selecting curly leaf phenotypes in a previously generated tomato mutant population (Saito et al. 2011, Shikata et al. 2016). Map-based cloning combined with ES revealed that the mutation occurred in the SlLAX1 (Solyc09g014380) gene (Figs. 3, 4). Then, to validate the candidate gene, we utilized TILLING to obtain an additional allelic line with a nonsense mutation, curl-7, which was generated by EMS. The curl-7 mutant leaves displayed similar curly leaves to the other curl mutants (Fig. 6B). Furthermore, we confirmed the full-length coding sequence of SlLAX1 (Fig. 6C, D), which supported the evidence that SlLAX1 is the gene responsible for the curly leaf phenotype. Taken together, the characterization of multiple alleles in this study that consistently showed similar phenotypes is strong evidence for the role of SlLAX1 in controlling the curly leaf phenotype. To our knowledge, this study is the first example of the successful application of ES in tomato for the identification of a causal gene preceded by a forward genetic approach. SlLAX1 encodes a transmembrane amino acid transporter protein and belongs to the amino acid/auxin permease (AAAP) family. Homology searches indicated that the SlLAX1 protein sequence is homologous to Arabidopsis thaliana AtAUX1 (AT2G38120). In Arabidopsis, AUX1 is one of four auxin influx carriers belonging to the AUX1/LAX family that controls several developmental processes including gravitropism responses, venation patterns and lateral roots (Bennett et al. 1996, Vieten et al. 2007). Although recent findings have indicated that the AUX/LAX1 family also control aerial part development such as leaf serration (Kasprzewska et al. 2015), phyllotaxis patterning, vascular patterning and xylem differentiation (Bainbridge et al. 2008, Fàbregas et al. 2015), the role of the AUX1/LAX gene family in leaf curling is poorly understood. In contrast, mutations in many auxin-related genes produced an impaired leaf flatness phenotype (Zgurski et al. 2005, Esteve-Bruna et al. 2013). In tomato, a few studies have shown a relationship between auxin and leaf flatness; for instance, SlARF4-RNAi (RNA interference) lines produce hyponastic leaves (Sagar et al. 2013) and SlPIN4-RNAi lines show leaf flatness defects as well as altered plant architecture (Pattison and Catalá 2012). However, the role of SlLAX1 in controlling the leaf curly phenotype has not been reported in tomato or other major crops. In tomato, the AUX1/LAX family consists of five genes (SlLAX1–SlLAX5). They share high identity and similarity; the identity of SlLAX2, SlLAX3, SlLAX4 and SlLAX5 with SlLAX1 is 80.36, 79.70, 92.65 and 80.87%, respectively (Sol Genomics Network). All SlLAX genes are expressed in the mature leave and root of tomato (Pattison and Catalá 2012). The single mutants depleted in SlLAX1 used in this study, curl-1–curl-7, showed a severe phenotype effect in leaf flatness, suggesting the importance of SlLAX1 in controlling leaf flatness in mature leaves. Although the functional redundancy of the AUX1/LAX family, in addition to the function of SlLAX1 itself, is poorly characterized in tomato, their function in Arabidopsis is well characterized particularly in root development. Although four AUX1/LAX genes share high sequence identity and similarity, AtAUX1 has the strongest auxin influx activity (Péret et al. 2012, Rutschow et al. 2014). Péret et al. (2012) also reported that subfunctionalization of the AUX1/LAX family in root is based on their distinct pattern of spatial expression and their subcellular localization. In contrast, the AUX1/LAX genes have redundant roles in the context of phyllotaxy, vascular patterning and xylem differentiation (Bainbridge et al. 2008, Fàbregas et al. 2015). Therefore, the functional redundancy of the SlLAX gene family in the tomato leaf curling phenotype awaits further investigation. Loss of function of SlLAX1 protein is related to the curly leaf phenotype Based on experimental evidence, Swarup et al. (2004) reported that Arabidopsis AUX1 protein has 11 transmembrane (TM) helixes. Using a publicly available server, we checked the prediction of TM helices in the SlLAX1 protein. According to a prediction program in http://www.cbs.dtu.dk/services/TMHMM/, both AtAUX1 and SlLAX1 (Supplementary Fig. S5A) have 10 TM helixes. The curl-2 and curl-6 mutants (Fig. 5B) carry a nonsense mutation which is located in TM helix VII (Supplementary Fig. S5C) according to TMHMM, which is equivalent to the central region of AtAUX1 and has proven to be particularly important for protein function (Swarup et al. 2004). In addition, both curl-1 and curl-3 mutations (Fig. 5D) are located in TM helix IV (Supplementary Fig. S5B), which is in a similar part of the N-terminal half of AtAUX1 and is essential for its correct localization (Péret et al. 2012). The curl-7 mutant has only five TM helixes, losing the other five (Supplementary Fig. S5D). Furthermore, the curl-1/curl-3, curl-2/curl-6 and curl-7 mutations were nonsense mutations that can produce only 35, 63 and 45% of the WT protein, respectively (Figs. 5B, D, 6D). Additionally, the relative expression of the curl mutant alleles (curl-1, curl-2 and curl-6) was <40% compared with the WT (Fig. 5F). These reasons presumably account for the loss-of-function mutations of the SlLAX1 gene. To test the potential function of SlLAX1 as an auxin transporter, we first measured the leaf endogenous auxin content. However, IAA content was comparable between the WT and the curl mutants at all stages (Supplementary Fig. S3). Numerous findings have indicated that AtAUX1 plays an important role in root gravitropism and lateral root development (Bennett et al. 1996, Marchant et al. 1999). The root gravitropism response is also commonly used to check auxin response and distribution. Therefore, we next carried out these assays and found that the root gravitropism response of the curl mutants was affected by the SlLAX1 mutation. In addition, lateral root emergence was also disrupted (Supplementary Fig. S4). Although the functional characterization of SlLAX1 has not been conducted in tomato and we do not yet have direct evidence in this study, agravitropism and lateral root formation defects of the curl mutants indicated that SlLAX1 may have a potential function as an auxin transporter similar to AtAUX1, and SlLAX1 might participate in local auxin distribution without affecting the total endogenous auxin content of the whole leaf. Functional analysis of the SlLAX1 gene remains to be carried out. The curly leaf phenotype of the curl mutants is presumably caused by an imbalance of pavement cell enlargement between the adaxial and abaxial sides The curly leaf phenotype was not observed at the early stage of leaf development (Fig. 1C, D;Supplementary Fig. S1), and is not related to relative humidity and water availability (Fig. 1E, G). Thus, we hypothesized that the curly leaf phenotype was caused by alteration of the adaxial/abaxial cell ratio rather than an impairment in adaxial–abaxial polarity since adaxial–abaxial polarity is established at the very early stage of leaf development, i.e. at the primordium stage. As expected, pavement cell size in the abaxial side in the curl mutants was significantly larger compared with that of the WT, while there was no significant difference in the adaxial side. The number of pavement cells in adaxial and abaxial sides was comparable. The upward curling of the curl mutants might be explained by the differential growth of pavement cells in the adaxial and abaxial cell surfaces, which is supported by a similar observation of incurvata6 (icu6), the semi-dominant allele of AUXIN RESISTANT3 (AXR3), that showed an upward curly phenotype caused by a reduced adaxial/abaxial cell size ratio (Pérez-Pérez et al. 2010). The imbalance in epidermal adaxial–abaxial cell growth which led to either epinastic (downward curvature) or hyponastic leaves is not a new phenomenon. It was previously reported that an auxin hyperaccumulation plant produced leaf epinastic curvature, formed due to an increased growth of the leaf adaxial side (Klee et al. 1987, Romano et al. 1993, Kim et al. 2007), which was induced by reduced auxin export that may cause its hyperaccumulation on the adaxial side. Taken together, SlLAX1 might have a function not in the establishment of adaxial–abaxial polarity but rather in balancing the adaxial/abaxial cell size ratio in later stages of leaf development. The evaluation of auxin distribution and/or analysis of SlLAX1 gene expression on the adaxial and abaxial leaf surfaces should allow for a better understanding of the SlLAX1 function in this process. According to the relative expression of some tomato putative auxin-related genes controlling leaf flatness, SlYuc1 showed prominent changes in both young and mature leaves of the curl mutants. YUC is a family of genes that are orthologs to ToFZY (Expósito-Rodríguez et al. 2007), which has a function in local auxin biosynthesis (Zhao et al. 2001). In a previous study it was reported that aux1 and yuc mutants in Arabidopsis have a synergistic effect to enhance each other in order to control leaf development. Lower expression of SlYuc1 does not change leaf auxin content presumably because SlYuc is a family of genes (Expósito-Rodríguez et al. 2007). Other SlYuc genes may compensate total auxin biosynthesis, resulting in comparable auxin contents in the entire leaf. In Arabidopsis, activation tagging of AtDof5.1 resulted in an upward curly leaf phenotype (Kim et al. 2010). Dof5.1 was demonstrated to promote Rev gene expression by binding to its promoter. Similar to these finding (Kim et al. 2010), expression of SlDof25, an ortholog of Dof5.1, was increased in all the curl mutants (Fig. 8G). The SlRev expression level was also increased (Fig. 8L). Up-regulation of AtDof5.1 also repressed transcript levels of auxin biosynthesis genes, which is consistent with a low expression level of SlYuc1 in the curl mutants (Fig. 8D, J). In tomato, it has also been reported that overexpression of a microRNA166-resistant version of SlREV (35S::REVRis) showed an upward curly leaf phenotype (Hu et al. 2014). Collectively, our findings are similar to previous findings which reinforce the partial disturbance of auxin homeostasis in the SlLAX1 mutants. The fact that lower auxin content triggers cell expansion is well established (Ishida et al. 2010, reviewed in Velasquez et al. 2016). We hypothesize that the loss of function of SlLAX1 in the curl mutants results in imbalanced adaxial/abaxial pavement cell growth leading to the curly leaf phenotype. Depletion of SlLAX1 in the curl mutants disrupts auxin transport in either the adaxial or abaxial leaf surface. Given that there was no significant difference in the adaxial pavement cell size (Table 6; Fig. 7A), SlLAX1 action appears to be restricted to the abaxial side. SlLAX1 belongs to the SlLAX family, and other members are known to be expressed in leaves (Pattison and Catalá 2012). It is possible that other influx transporters compensate for the loss of function of SlLAX1 in the adaxial side. In contrast, our data suggest that SlLAX1 is a major determinant of auxin transporter which is dominant in the abaxial side. In the adaxial side, where SlLAX1 is not a major auxin influx carrier, presumably the auxin content of the curl mutants was similar to or higher than that of the WT, while a decreased auxin content in the abaxial side was due to low influx carrier activity. Auxin, which is not taken up by abaxial cells, may be accumulated in the adaxial side, or alternatively accumulated in the extracellular space (most probably the latter because the cell number in the adaxial and abaxial sides was comparable, meaning that there was no increase in cell division in the adaxial side). Therefore, the auxin content in the curl mutants could be maintained at a similar level to the WT (Supplementary Fig. S3). Imbalanced adaxial/abaxial cell growth due to differential auxin accumulation is also well established (Pérez-Pérez et al. 2010, reviewed in Sandalio et al. 2016). We speculate that the lower auxin content on the abaxial cell surface triggers cell expansion and imbalanced cell growth on both surfaces, leading to the emergence of curly leaves. This hypothesis awaits further investigation. In brief, this study contributes to the newly characterized role of SlLAX1 in controlling leaf development in tomato by balancing the adaxial–abaxial pavement cell enlargement potentially mediated by auxin. The evaluation of auxin distribution and/or analysis of SlLAX1 gene expression on the adaxial and abaxial leaf surfaces should allow for a better understanding of the SlLAX1 function in this process. Additionally, analysis of double mutants with other LAX or PIN family members and other adaxial–abaxial specification genes would be helpful to dissect the precise mechanism of SlLAX1 in normal leaf development in plants. Materials and Methods Plant material and growth conditions Tomato (Solanum lycopersicum cv. ‘Micro-Tom’) curly leaf (curl) mutants were generated by EMS and γ-ray irradiation. The mutants were obtained from the National BioResources Project (NBRP) Project at the University of Tsukuba (Saito et al. 2011, Shikata et al. 2016). From the M3 mutagenized population, we isolated six lines of the curly leaf phenotype mutants, herein referred to as ‘curl’ mutants. The mutant screening was carried out visually using mature plants showing severe curly leaf phenotypes. Five mutant alleles, curl-1–curl-5, were generated by γ-ray irradiation, and one mutant allele, curl-6, was generated by EMS mutagenesis. Furthermore, using TILLING screening, we screened another EMS mutant, curl-7. These mutants were registered in the TOMATOMA mutant database (Saito et al. 2011, http://tomatoma.nbrp.jp/). The NBRP accession numbers are listed in Supplementary Table S6. Unless otherwise stated, further analyses of the curl mutants were conducted after two backcrosses to the WT ‘Micro-Tom’ to remove any possible background mutation following the mutagenesis treatment. The plants were grown under standard cultivation conditions in the greenhouse facility at the University of Tsukuba. Genomic DNA extraction, construction of the mapping population, DNA markers and genetic analysis Genomic DNA was extracted from 2-month-old plants. A maximum of 100 mg of fresh leaf sample was extracted using a Maxwell® 16 Tissue DNA Purification Kit (Promega). To perform rough mapping using DNA markers, curl-2 was crossed to another tomato cultivar, ‘Ailsa Craig’, to obtain a mapping population. From approximately 100 plants of the F2 mapping population, 19 plants exhibiting the curly mutant phenotype were isolated, and genomic DNA was extracted from the leaves of the individual plants. These plants were subjected to rough mapping experiments. All SNP and CAPS DNA markers were designed according to the AMF2 (F2: Solanum lycopersicum ‘Ailsa Craig’×S. lycopersicum ‘Micro-Tom’) linkage map information that is publicly available from the Kazusa DNA Research Institute (KDRI) webpage (http://marker.kazusa.or.jp/Tomato/; Shirasawa et al. 2010). The primers and restriction enzyme used in rough mapping chromosomes are listed in Supplementary Table S7. Exome sequencing and variant identification ES was performed to narrow down the candidate genes. Four alleles, curl-1, curl-2, curl-3 and curl-6, of the curl mutants of the F2 mutant population backcrossed to the WT were used. The mutants and WT phenotypes were selected in the F2 population based on the presence or absence of curly leaves among approximately 100 F2 plants for each line, after which their DNA samples were bulked based on phenotype. ES analysis was then performed based on the Roche exome sequence SeqCap® EZ SR protocol (http://sequencing.roche.com/). Briefly, genomic DNA was treated with a Covaris® S220 Ultrasonicator (Covaris) to achieve an average length of 200 bp. Then, a multiplex next-generation sequencing (NGS) library was constructed using a KAPA® Library Preparation Kit and SeqCap® adaptor kit (Roche). After constructing the NGS library, exome capture was conducted using a custom probe set that was designed based on the tomato genome reference version SL2.50 (supporting dataset, Sol Genomics Network, https://solgenomics.net). This probe set was designed to capture 49.5 Mb of exonic DNA regions (Supplementary Data S1). The resultant exome library was amplified by 14 cycles of post-capture ligation-mediated PCR with KAPA HiFi HostStart ReadyMix (Roche) and then subjected to Illumina HiSeq-2000 sequencing set to the 100 bp paired-end mode. Paired-end short read data were subjected to quality filtering using the FASTXToolkit with the parameters of –Q 20 –P 90. Then, short reads were aligned to the tomato genome reference version SL2.50 using bowtie2 software with the following parameters: L, 0,-0.16 –mp 2, 2 –np 1 –rdg 1, 1 –rfg 1, 1. On average, 98.8 ± 0.03% of the target exonic regions were covered by short reads. The average read depth was 18 ± 1.5. Genome-wide DNA polymorphisms and mutations were identified based on the alignment results by the HaplotypeCaller function of the Genome Analysis Toolkit (GATK) with the following parameters: -mmq 5 -forceActive -stand_call_conf 10 -stand_emit_conf 10. The resultant DNA variant information was further combined into one genomic VCF data set with the GenotypeGVCFs function of the GATK. Three wild-type ES data sets [DNA Data Bank of Japan (DDBJ) accession Nos. DRR097500–DRR097502], two WT whole-genome NGS data sets (DDBJ accession Nos. DRR097503 and DRR097504) and one publicly available WT whole-genome NGS data set (Kobayashi et al. 2014) were used as controls to remove intracultivar variations that are present between WT ‘Micro-Tom’ lines. DNA variants were further removed if their allele frequencies was >90% in the WT F2 bulked segregants because they were also expected to be intracultivar variations. Those variants with <20% allele frequency or with a read depth <6 were also removed because they were likely to be false positives. ES data sets for curl mutants are available in the DDBJ (accession Nos DRR097492–DRR097502). RNA extraction and cDNA synthesis Total RNA was extracted from young and mature leaves (when the leaves were completely curly) using an RNeasy Mini Kit (QIAGEN) according to the manufacturer’s protocol. To remove genomic DNA contamination, two steps were applied: an on-column RNase-free DNase Set (QIAGEN) and an RNA Clean & Concentrator™-5 (Zymo Research). Subsequently, cDNA was synthesized from 2,000 ng of total RNA by a SuperScript III First Strand Synthesis Kit (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. Cloning and sequencing of the full-length coding sequence of the SlLAX1 gene The full-length coding sequence (1,236 bp) of the SlLAX1 gene from three independent plants was amplified by PCR. The primer sequences are listed in Supplementary Table S4. Subsequently, PCR products were loaded onto a 0.8–1.5% agarose gel, which was then electrophoresed for 45–60 min. Next, the band was visualized under 70% UV and then cut either with a gel cutter or a blade. Any visible desired product band was individually cut, removed and subsequently subjected to purification using a Wizard® SV Gel and PCR Clean-Up System (Promega). DNA purification by centrifugation was applied. The purified PCR product was then cloned into the entry vector pCR8/GW/TOPO (Invitrogen, http://www.lifetechnologies.com/) using an In-Fusion® HD Cloning Kit (TAKARA BIO INC.) according to the manufacturer’s protocol. Then, plasmids from clones were purified using a FastGene Plasmid Mini Kit (Nippon Genetics). The plasmid fragments were sequenced using M13 primers (Supplementary Table S4). qRT-PCR analysis The mRNA expression level was quantified using qRT-PCR. A 10 ng µl−1 cDNA template of three biological replicates was used for gene expression analyses. The SlActin gene was used as an internal control (Løvdal and Lillo 2009). qRT-PCR was carried out using a CFX96 Real-Time System (Bio-Rad) with SYBR Premix ExTaq II (Ili RNase H Plus; TAKARA BIO INC.). The primers used for qRT-PCR are listed in Supplementary Table S4. Relative gene expression was quantified using the ΔΔCT method (Pfaffl 2001). The qRT-PCR mixture and thermal cycle conditions were as described by Shinozaki et al. (2015). The primers for qRT-PCR were designed using the Primer3 Plus website (http://primer3plus.com/); two exons in the forward or reverse primer were joined to exclude any possibility of contamination of genomic DNA. Screening new SlLAX1 mutant alleles by TILLING The TILLING population was previously described by Okabe et al. (2013), and the TILLING experiments were performed as described by Okabe et al. (2011). We attempted to screen for mutations in the coding region of the SlLAX1 gene. The primer pair was designed to span exon 6. Given that exon 6 is the longest exon, we also identified an EMS mutant line, curl-6, that carries a nonsense mutation in exon 6 of SlLAX1. The primer pair used in the TILLING experiment was forward 5'-TGGTACATGGGAACTAGCTAAGCC-3' and reverse 5'-ACCTGACGAGCGGATGATTTTC-3', which amplified 865 bp of genomic DNA template; the 5' end of each primer was labeled with DY-681 or DY-781, which are equivalent to IRDye 700 or IRDye 800 (https://www.biomers.net/), respectively. Morphological analysis The CI of mutants was measured on the fifth leaflet in accordance with the method introduced by Liu et al. (2010). Leaf area and perimeter analyses were conducted at the young and mature leaf stages; 15 leaves harvested from the same position were used as samples. Leaf images were captured using a digital camera, and the leaf area and perimeter were measured using CellSensStandard imaging software (Olympus). The leaf perimeter and leaf area were measured by following the edge of the leaf using a closed polygon measurement tool within the CellSensStandard software. The reduction in leaf area and leaf perimeter (%) was measured by comparing the values before and after flattening (multiplied by 100). SEM The leaf epidermal surface was observed using a scanning electron microscope (Hitachi Tabletop Microscope TM3000). The cell features were measured at the mature leaf stage when the leaves were completely curly, and precisely in the same regions on the adaxial and abaxial surfaces. Mature fresh leaves were sampled and flattened before being subjected to microscopic observation. Approximately 0.5×0.5 cm2 of adaxial or abaxial surface was placed into a sample box, after which the epidermal pavement cell was imaged at ×400 magnification for at least three biological replications. The cell size was quantified separately using CellSensStandard software. All measurements were obtained for at least three independently captured SEM images for each replication and three fields of view for each image. For quantification of the number of pavement cells, leaf samples were cut from midway exactly between the midrib and the margin of fully curly leaves. We used precisely the same position on both the adaxial and abaxial sides; one side was used for adaxial pavement cell observation and the other was used for the abaxial pavement cells. An approximatley 2–4 mm leaf sample in the tip area of the transversal axis was cut irrespective of the size from the midrib to the margin, and it was subjected to SEM (Supplementary Fig. S6). The cell number was counted throughly in that region. Measurements were obtained from three biological replications. Measurement of the auxin content in leaves Leaves were sampled at three stages from the same positions in (i) young leaves, before curly leaves formed; (ii) when leaves just turned curly; and (iii) mature leaves, after leaves were fully curly. Three biological replications were included at each stage. At least 100 mg of fresh leaves was immediately frozen in liquid nitrogen and crushed into a fine powder using a TissueLyser (Qiagen). Endogenous auxin was measured using a UHPLC-Q-Exactive (Thermo Fisher Scientific) system. Measurements were conducted as described by Kojima et al. (2009) and Shinozaki et al. (2015). Leaf water potential measurements Leaf water potential was measured using a pressure chamber. A leaflet from the same position was cut and immediately placed into the chamber. Pressure was gradually increased until water was exuded from the petiole. Six biological samples were tested. Statistical analysis Unless otherwise stated, the data are presented as the mean ± SE. Student’s t-test (at the 95% and 99% significance levels) was used to analyze the significance level between two values with equal variance. χ2 tests were performed using MS Excel 2016 to examine the goodness of fit between the expected and observed Mendelian ratio in the segregating F2 population of mutants backcrossed to the WT ‘Micro-Tom’, and the degrees of freedom and expected Mendelian ratio used for monogenic traits were 1 and 3:1 (WT:mutant phenotype), respectively. Supplementary Data Supplementary data are available at PCP online. Funding This work was supported by the Japan Society for the Promotion of Science (JSPS) [KAKENHI Grant-in-Aid for Research Activity start-up (15H06071 to R.Y.)]; Program to Disseminate Tenure Tracking System [to T.A.]; and the Japan Advanced Plant Science Network [to H.E. and H.S.]. Acknowledgments Tomato ‘Micro-Tom’ and the curl mutant seeds were obtained from the National BioResource Project (NBRP), Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. We are grateful for helpful comments and discussion of the manuscript from Dr. Kentaro Ezura. We would like to express our sincere gratitude to all our laboratory members for their great support and helpful discussion throughout the work. Disclosures The authors have no conflicts of interest to declare. References Ariizumi T., Kishimoto S., Kakami R., Maoka T., Hirakawa H., Suzuki Y., et al.  . ( 2014) Identification of the carotenoid modifying gene PALE YELLOW PETAL 1 as an essential factor in xanthophyll esterification and yellow flower pigmentation in tomato (Solanum lycopersicum). Plant J.  79: 453– 465. Google Scholar CrossRef Search ADS PubMed  Bainbridge K., Guyomarc’h S., Bayer E., Swarup R., Bennett M., Mandel T., et al.  . ( 2008) Auxin influx carriers stabilize phyllotactic patterning. Genes Dev . 22: 810– 823. Google Scholar CrossRef Search ADS PubMed  Bennett M.J., Marchant A., Green H.G., May S.T., Ward S.P., Millner P.A., et al.  . ( 1996) Arabidopsis AUX1 gene: a permease-like regulator of root gravitropism. Science  273: 948– 950. Google Scholar CrossRef Search ADS PubMed  Bennett M.J., Marchant A., May S.T., Swarup R. ( 1998) Going the distance with auxin: unravelling the molecular basis of auxin transport. Philos. Trans. R. Soc. B: Biol. Sci  353: 1511– 1515. Google Scholar CrossRef Search ADS   Cai X., Zhang Y., Zhang C., Zhang T., Hu T., Ye J., et al.  . ( 2013) Genome-wide analysis of plant-specific Dof transcription factor family in tomato. J. Integr. Plant Biol.  55: 552– 566. Google Scholar CrossRef Search ADS PubMed  Cheng Y., Dai X., Zhao Y. ( 2007) Auxin synthesized by the YUCCA flavin monooxygenases is essential for embryogenesis and leaf formation in Arabidopsis. Plant Cell  19: 2430– 2439. Google Scholar CrossRef Search ADS PubMed  Chusreeaeom K., Ariizumi T., Asamizu E., Okabe Y., Shirasawa K., Ezura H. ( 2014) A novel tomato mutant, Solanum lycopersicum elongated fruit1 (Slelf1), exhibits an elongated fruit shape caused by increased cell layers in the proximal region of the ovary. Mol. Genet. Genomics  289: 399– 409. Google Scholar CrossRef Search ADS PubMed  Delbarre A., Muller P., Imhoff V., Guern J. ( 1996) Comparison of mechanisms controlling uptake and accumulation of 2,4-dichlorophenoxy acetic acid, naphthalene-1-acetic acid, and indole-3-acetic acid in suspension-cultured tobacco cells. Planta  198: 532– 541. Google Scholar CrossRef Search ADS PubMed  Emery J.F., Floyd S.K., Alvarez J., Eshed Y., Hawker N.P., Izhaki A., et al.  . ( 2003) Radial patterning of Arabidopsis shoots by class III HD-ZIP and KANADI genes. Curr. Biol . 13: 1768– 1774. Google Scholar CrossRef Search ADS PubMed  Esteve-Bruna D., Perez-Perez J.M., Ponce M.R., Micol J.L. ( 2013) incurvata13, a novel allele of AUXIN RESISTANT6, reveals a specific role for auxin and the SCF complex in Arabidopsis embryogenesis, vascular specification, and leaf flatness. Plant Physiol . 161: 1303– 1320. Google Scholar CrossRef Search ADS PubMed  Expósito-Rodríguez M., Borges A.A., Borges-Pérez A., Hernández M., Pérez J.A. ( 2007) Cloning and biochemical characterization of ToFZY, a tomato gene encoding a flavin monooxygenase involved in a tryptophan-dependent auxin biosynthesis pathway. J. Plant Growth Regul.  26: 329– 340. Google Scholar CrossRef Search ADS   Fàbregas N., Formosa-Jordan P., Confraria A., Siligato R., Alonso J.M., Swarup R., et al.  . ( 2015) Auxin influx carriers control vascular patterning and xylem differentiation in Arabidopsis thaliana. PLoS Genet.  11: e1005183. Google Scholar CrossRef Search ADS PubMed  Hao S., Ariizumi T., Ezura H. ( 2017) SEXUAL STERILITY is essential for both male and female gametogenesis in tomato. Plant Cell Physiol.  58: 22– 34. Google Scholar CrossRef Search ADS PubMed  Hu G., Fan J., Xian Z., Huang W., Lin D., Li Z. ( 2014) Overexpression of SlREV alters the development of the flower pedicel abscission zone and fruit formation in tomato. Plant Sci . 229: 86– 95. Google Scholar CrossRef Search ADS PubMed  Ishida T., Adachi S., Yoshimura M., Shimizu K., Umeda M., Sugimoto K. ( 2010) Auxin modulates the transition from the mitotic cycle to the endocycle in Arabidopsis. Development  137: 63– 71. Google Scholar CrossRef Search ADS PubMed  Kasprzewska A., Carter R., Swarup R., Bennett M., Monk N., Hobbs J.K., et al.  . ( 2015) Auxin influx importers modulate serration along the leaf margin. Plant J.  83: 705– 718. Google Scholar CrossRef Search ADS PubMed  Kim H.S., Kim S.J., Abbasi N., Bressan R.A., Yun D.J., Yoo S.D., et al.  . ( 2010) The DOF transcription factor Dof5.1 influences leaf axial patterning by promoting Revoluta transcription in Arabidopsis. Plant J . 64: 524– 535. Google Scholar CrossRef Search ADS PubMed  Kim J.I., Sharkhuu A., Jin J.B., Li P., Jeong J.C., Baek D., et al.  . ( 2007) yucca6, a dominant mutation in Arabidopsis, affects auxin accumulation and auxin-related phenotypes. Plant Physiol . 145: 722– 735. Google Scholar CrossRef Search ADS PubMed  Klee H.J., Horsch R.B., Hinchee M.A., Hein M.B., Hoffmann N.L. ( 1987) The effects of overproduction of two Agrobacterium tumefaciens T-DNA auxin biosynthetic gene products in transgenic petunia plants. Genes Dev . 1: 86– 96. Google Scholar CrossRef Search ADS   Kobayashi M., Nagasaki H., Garcia V., Just D., Bres C., Mauxion L.P., et al.  . ( 2014) Genome-wide analysis of intraspecific DNA polymorphism in ‘Micro-Tom’, a model cultivar of tomato (Solanum lycopersicum). Plant Cell Physiol . 55: 445– 454. Google Scholar CrossRef Search ADS PubMed  Kojima M., Kamada-Nobusada T., Komatsu H., Takei K., Kuroha T., Mizutani M., et al.  . ( 2009) Highly sensitive and high-throughput analysis of plant hormones using MS-probe modification and liquid chromatography–tandem mass spectrometry: an application for hormone profiling in Oryza sativa. Plant Cell Physiol . 50: 1201– 1214. Google Scholar CrossRef Search ADS PubMed  Kramer E.M., Bennett M.J. ( 2006) Auxin transport: a field in flux. Trends Plant Sci.  11: 382– 386. Google Scholar CrossRef Search ADS PubMed  Lee C., Chronis D., Kenning C., Peret B., Hewezi T., Davis E.L., et al.  . ( 2011) The novel cyst nematode effector protein 19C07 interacts with the Arabidopsis auxin influx transporter LAX3 to control feeding site development. Plant Physiol . 155: 866– 880. Google Scholar CrossRef Search ADS PubMed  Liu Z., Jia L., Mao Y., He Y. ( 2010) Classification and quantification of leaf curvature. J. Exp. Bot.  61: 2757– 2767. Google Scholar CrossRef Search ADS PubMed  Liu Z., Jia L., Wang H., He Y. ( 2011) HYL1 regulates the balance between adaxial and abaxial identity for leaf flattening via miRNA-mediated pathways. J. Exp. Bot . 62: 4367– 4381. Google Scholar CrossRef Search ADS PubMed  Løvdal T., Lillo C. ( 2009) Reference gene selection for quantitative real-time PCR normalization in tomato subjected to nitrogen, cold, and light stress. Anal. Biochem . 387: 238– 242. Google Scholar CrossRef Search ADS PubMed  Marchant A., Bhalerao B., Casimiro I., Eklöf J., Casero P.J., Bennett M., et al.  . ( 2002) AUX1 promotes lateral root formation by facilitating indole-3-acetic acid distribution between sink and source tissues in the Arabidopsis seedling. Plant Cell  14: 589– 597. Google Scholar CrossRef Search ADS PubMed  Marchant A., Kargul J., May S.T., Muller P., Delbarre A., Perrot-Rechenmann C., et al.  . ( 1999) AUX1 regulates root gravitropism in Arabidopsis by facilitating auxin uptake within root apical tissues. EMBO J  18: 2066– 2073. Google Scholar CrossRef Search ADS PubMed  Newman K.L., Fernandez A.G., Barton M.K. ( 2002) Regulation of axis determinacy by the Arabidopsis PINHEAD gene. Plant Cell  14: 3029– 3042. Google Scholar CrossRef Search ADS PubMed  Okabe Y., Ariizumi T., Ezura H. ( 2013) Updating the Micro-Tom TILLING platform. Breed. Sci.  63: 42– 48. Google Scholar CrossRef Search ADS PubMed  Okabe Y., Asamizu E., Saito T., Matsukura C., Ariizumi T., Bres C., et al.  . ( 2011) Tomato TILLING technology: development of a reverse genetics tool for the efficient isolation of mutants from Micro-Tom mutant libraries. Plant Cell Physiol . 52: 1994– 2005. Google Scholar CrossRef Search ADS PubMed  Paciorek T., Zažímalová E., Ruthardt N., Petrášek J., Stierhof Y.-D., Kleine-Vehn J., et al.  . ( 2005) Auxin inhibits endocytosis and promotes its own efflux from cells. Nature  435: 1251– 1256. Google Scholar CrossRef Search ADS PubMed  Pattison R.J., Catalá C. ( 2012) Evaluating auxin distribution in tomato (Solanum lycopersicum) through an analysis of the PIN and AUX/LAX gene families. Plant J . 70: 585– 598. Google Scholar CrossRef Search ADS PubMed  Péret B., Rybel B.D., Casimiro I., Benkova E., Swarup R., Laplaze L., et al.  . ( 2009) Arabidopsis lateral root development: an emerging story. Trends Plant Sci . 14: 399– 408. Google Scholar CrossRef Search ADS PubMed  Péret B., Swarup K., Ferguson A., Seth M., Yang Y., Dhondt S., et al.  . ( 2012) AUX/LAX genes encode a family of auxin influx transporters that perform distinct functions during Arabidopsis development. Plant Cell  24: 2874– 2885. Google Scholar CrossRef Search ADS PubMed  Pérez-Pérez J.M., Candela H., Robles P., López-Torrejón G., Del Pozo J.C., Micol J.L. ( 2010) A role for AUXIN RESISTANT3 in the coordination of leaf growth. Plant Cell Physiol . 51: 1661– 1673. Google Scholar CrossRef Search ADS PubMed  Pfaffl M.W. ( 2001) A new mathematical model for relative quantification in real-time RT–PCR. Nucleic Acids Res.  29: e45– e4521. Google Scholar CrossRef Search ADS PubMed  Prigge M.J., Otsuga D., Alonso J.M., Ecker J.R., Drews G.N., Clark S.E. ( 2005) Class III homeodomain-leucine zipper gene family members have overlapping, antagonistic, and distinct roles in Arabidopsis development. Plant Cell  17: 61– 76. Google Scholar CrossRef Search ADS PubMed  Romano C.P., Cooper M.L., Klee H.J. ( 1993) Uncoupling auxin and ethylene effects in transgenic tobacco and Arabidopsis plants. Plant Cell  5: 181– 189. Google Scholar CrossRef Search ADS PubMed  Rutschow H.L., Baskin T.I., Kramer E.M. ( 2014) The carrier AUXIN RESISTANT (AUX1) dominates auxin flux into Arabidopsis protoplasts. New Phytol.  204: 536– 544. Google Scholar CrossRef Search ADS PubMed  Sagar M., Chervin C., Mila I., Hao Y., Roustan J.-P., Benichou M., et al.  . ( 2013) SlARF4, an auxin response factor involved in the control of sugar metabolism during tomato fruit development. Plant Physiol . 161: 1362– 1374. Google Scholar CrossRef Search ADS PubMed  Saito T., Ariizumi T., Okabe Y., Asamizu E., Hiwasa-Tanase K., Fukuda N., et al.  . ( 2011) TOMATOMA: a novel tomato mutant database distributing Micro-Tom mutant collections. Plant Cell Physiol . 52: 283– 296. Google Scholar CrossRef Search ADS PubMed  Sandalio L.M., Godiguez-Serrano M., Romero-Puertas M.C. ( 2016) Leaf epinasty and auxin: a biochemical and molecular overview. Plant Sci.  253: 187– 193. Google Scholar CrossRef Search ADS PubMed  Serrano-Cartagena J., Robles P., Ponce M.R., Micol J.L. ( 1999) Genetic analysis of leaf form mutants from the Arabidopsis Information Service collection. Mol. Gen. Genet . 261: 725– 739. Google Scholar CrossRef Search ADS PubMed  Shikata M., Hoshikawa K., Ariizumi T., Fukuda N., Yamazaki Y., Ezura H. ( 2016) TOMATOMA update: phenotypic and metabolite information in the Micro-Tom mutant resource. Plant Cell Physiol.  57: e11. Google Scholar CrossRef Search ADS PubMed  Shinozaki Y., Hao S., Kojima M., Sakakibara H., Ozeki-Iida Y., Zheng Y., et al.  . ( 2015) Ethylene suppresses tomato (Solanum lycopersicum) fruit set through modification of gibberellin metabolism. Plant J.  83: 237– 251. Google Scholar CrossRef Search ADS PubMed  Shirasawa K., Isobe S., Hirakawa H., Asamizu E., Fukuoka H., Just D., et al.  . ( 2010) SNP discovery and linkage map construction in cultivated tomato. DNA Res.  17: 381– 391. Google Scholar CrossRef Search ADS PubMed  Song J.B., Huang S.Q., Dalmay T., Yang Z.M. ( 2012) Regulation of leaf morphology by microRNA394 and its target LEAF CURLING RESPONSIVENESS. Plant Cell Physiol . 53: 1283– 1294. Google Scholar CrossRef Search ADS PubMed  Swarup K., Benková E., Swarup R., Casimiro I., Péret B., Yang Y., et al.  . ( 2008) The auxin influx carrier LAX3 promotes lateral root emergence. Nat. Cell Biol.  10: 946– 954. Google Scholar CrossRef Search ADS PubMed  Swarup R., Kargul J., Marchant A., Zadik D., Rahman A., Mills R., et al.  . ( 2004) Structure–function analysis of the presumptive Arabidopsis auxin permease AUX1. Plant Cell  16: 3069– 3083. Google Scholar CrossRef Search ADS PubMed  Swarup R., Péret B. ( 2012) AUX/LAX family of auxin influx carriers—an overview. Front. Plant Sci.  3: 1– 11. Google Scholar CrossRef Search ADS PubMed  Tromas A., Perrot-Rechenmann C. ( 2010) Recent progress in auxin biology. C. R. Biol.  333: 297– 306. Google Scholar CrossRef Search ADS PubMed  Velasquez S.M., Elke B., Jürgen K., José M.E. ( 2016) Auxin and cellular elongation. Plant Physiol . 70: 1206– 1215. Vieten A., Sauer M., Brewer P.B., Friml J. ( 2007) Molecular and cellular aspects of auxin-transport-mediated development. Trends Plant Sci . 12: 160– 168. Google Scholar CrossRef Search ADS PubMed  Yang H., Murphy A.S. ( 2009) Functional expression and characterization of Arabidopsis ABCB, AUX 1 and PIN auxin transporters in Schizosaccharomyces pombe. Plant J . 59: 179– 191. Google Scholar CrossRef Search ADS PubMed  Yu L., Yu X., Shen R., He Y. ( 2005) HYL1 gene maintains venation and polarity of leaves. Planta  221: 231– 242. Google Scholar CrossRef Search ADS PubMed  Zgurski J.M., Sharma R., Bolokoski D.A., Schultz E.A. ( 2005) Asymmetric auxin response precedes asymmetric growth and differentiation of asymmetric leaf1 and asymmetric leaf2 Arabidopsis leaves. Plant Cell  17: 77– 91. Google Scholar CrossRef Search ADS PubMed  Zhao Y., Christensen S.K., Fankhauser C., Cashman J.R., Cohen J.D., Weigel D., et al.  . ( 2001) A role for flavin monooxygenase-like enzymes in auxin biosynthesis. Science  291: 306– 309. Google Scholar CrossRef Search ADS PubMed  Abbreviations Abbreviations CAPS cleaved amplified polymorphic sequence CI curvature index curl curly leaf EMS ethyl methanesulfonate ES exome sequencing LAX like AUX1 NGS next-generation sequencing PAT polar auxin transport qRT-PCR quantitative real-time PCR SEM scanning electron microscopy SNP single nucleotide polymorphism TILLING Targeting Induced Local Lesions IN Genome TM transmembrane UTR untranslated region WT wild type © The Author(s) 2018. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Plant and Cell PhysiologyOxford University Press

Published: Mar 8, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off