Modulating the strigolactone pathway to optimize tomato shoot branching for vertical farming

ABSTRACT

Optimizing plant architecture for specific cultivation methods is essential for enhancing fruit productivity. Unlike indeterminate growth plants, the total productivity of determinate growth plants relies on cumulative fruit production and synchronized fruit ripening from both main and axillary shoots. Here, we focused on SlD14 and SlMAX1, two key genes involved in the regulation of strigolactone (SL) signaling and biosynthesis, with the goal of maximizing yield and synchronizing fruit ripening by fine‐tuning axillary shoot growth. Using clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR‐associated protein 9 (Cas9) technology, we found that the sld14, slmax1, and sld14 slmax1 mutant plants exhibited reduced plant height and increased axillary shoot proliferation compared to wild‐type plants. However, these mutants showed reduced yield and delayed ripening, likely due to a source‐sink imbalance caused by excessive axillary shoot development. A weak sld14 allele displayed a milder phenotype, maintaining total fruit yield and harvest index despite smaller individual fruit size. These findings indicate that allelic variation in SL‐related genes can influence plant architecture and yield components. Our results suggest that weak or partial alleles may serve as promising targets for tailoring tomato architecture to space‐limited cultivation systems.

Optimizing tomato architecture via CRISPR–Cas9 editing of strigolactone–related genes revealed that weak alleles can balance axillary shoot growth, yield, and ripening. Strong mutants reduced yield due to source—sink imbalance, but a weak allele maintained productivity, highlighting its potential for space–efficient cultivation systems.

INTRODUCTION

Over thousands of years of domestication and breeding, most crops have been selectively optimized for yield and quality based on their specific growing environments. Many key agricultural traits, collectively known as domestication syndrome traits, have been continuously modified through human selection (Alam and Purugganan, 2024). Examples of these traits include plant architecture, early flowering, non‐shattering seeds, reduced fruit drop, increased fruit size, and diminished shoot branching (Alam and Purugganan, 2024). In particular, reduced axillary shoot development concentrates nutrients within the main shoot according to the source‐sink principle, thereby increasing economic yield (Rameau et al., 2015). In tomato (Solanum lycopersicum) cultivars have been bred to suppress axillary shoot growth by eliminating shoot branching to maximize fruit production from the main shoot, especially in greenhouse cultivation (Navarrete and Jeannequin, 2000; Xu et al., 2020). However, this strategy is primarily suited for indeterminate tomato varieties. In open‐field conditions, maintaining axillary shoot development can actually enhance fruit productivity per plant (Pnueli et al., 1998). Additionally, various mechanized farming techniques have been developed for large‐scale open‐field cultivation and harvesting. In tomato, for example, mechanical harvesters enable simultaneous collection of fruits over a given area (Arazuri et al., 2007). Recently, advancements in greenhouse construction and the development of alternative cultivation techniques have highlighted the growing need for research aimed at customizing crop traits to suit diverse agricultural environments (Kwon, 2023).

Vertical farms, first proposed in the 1990s, are a type of smart farming system that utilizes information and communication technology (ICT) to remotely and automatically regulate crop‐growing environments (Benke and Tomkins, 2017). Because vertical farms operate indoors with controlled artificial lighting, temperature, humidity, and CO₂ levels, they can be utilized in regions where conventional crop cultivation is challenging (Van Delden et al., 2021). A key feature of vertical farms is high‐density planting within a limited space, making them particularly suitable for crops with a small planting footprint and a short growth cycle from sowing to harvest (SharathKumar et al., 2020; Teo and Yu, 2024). As a result, leafy greens are the primary crops cultivated in these systems. Among fruiting vegetables, strawberry is one of the few widely cultivated species in vertical farms due to their compact size and small planting area per plant (Kouloumprouka Zacharaki et al., 2024). However, for taller crops with longer cultivation cycles, such as tomato and pepper, targeted trait improvements are necessary to optimize their suitability for vertical farming (Kwon, 2023; Yu et al., 2025). A previous study introduced the “triple‐determinate” tomato plant as a promising candidate for urban agriculture and vertical farming due to its determinate stem growth, rapid flowering, and short stem internodes (Kwon et al., 2020). However, axillary shoot development in this cultivar remains suboptimal, and further breeding efforts are required to synchronize fruit ripening, ensuring maximum per‐plant productivity and efficient simultaneous harvesting in vertical farm systems (Kwon, 2023).

Strigolactones (SLs) are terpenoid‐derived plant hormones that regulate key developmental processes such as shoot branching, root architecture, and adaptation to environmental stresses (Yoneyama and Brewer, 2021). Their structure consists of a tricyclic lactone (ABC rings) linked to a conserved D‐ring, forming an enol ether bridge (Barbier et al., 2023). Strigolactones are classified into strigol and orobanchol types based on stereochemical differences in the BC ring junction (Zwanenburg and Pospíšil, 2013). Strigolactone biosynthesis begins with all‐trans β‐carotene conversion to carlactone (CL) via the action of DWARF27 (D27), carotenoid cleavage dioxygenase 7 (CCD7), and CCD8, followed by species‐specific modifications by cytochrome P450 (CYP) enzymes such as more axillary growth 1 (MAX1) in Arabidopsis and CYP722C/SlMAX1 in tomato (Abe et al., 2014; Zhang et al., 2018; Mashiguchi et al., 2021). Strigolactone production is primarily root‐based and is induced by environmental factors like phosphate deficiency (López‐Ráez et al., 2008). The α/β‐hydrolase receptor D14 plays a central role in SL perception by hydrolyzing SLs and enabling their recognition through interaction with MAX2 (Sánchez Martín‐Fontecha et al., 2024). Importantly, D14 hydrolysis generates an intermediate that drives the conformational changes necessary for signal transduction (Yao et al., 2016). More recently, structural and biophysical studies have clarified the dynamic behavior of the D14‐D3/MAX2‐D53 signaling complex, providing mechanistic insight into how SL perception promotes the degradation of D53, a key repressor of SL signaling, thereby activating downstream responses (Liu et al., 2023). Mutations in D53 that prevent its degradation lead to increased shoot branching and reduced sensitivity to SLs (Jiang et al., 2013; Zhou et al., 2013). While SL binding to D14 is known to recruit MAX2 for protein degradation, the precise molecular events following D53 breakdown remain understood (Mashiguchi et al., 2021; Barbier et al., 2023; Hu et al., 2024).

Among the various genes involved in the SL pathway, the genetic and molecular association of MAX1 and D14 with shoot branching has been extensively studied through genetic analyses in the model plants Arabidopsis and rice (Abe et al., 2014; Mashiguchi et al., 2021; Hu et al., 2024). The MAX1 protein catalyzes the conversion of CL to carlactonoic acid (CLA) in the SL biosynthetic pathway. Carlactonoic acid can either be methylated to produce methyl carlactonoate (MeCLA), a precursor of noncanonical SLs, or be converted into canonical SLs via a distinct catalytic reaction (Mashiguchi et al., 2022). Noncanonical SLs such as MeCLA are critical for the control of plant architecture, strongly modulating shoot branching and stem elongation, whereas canonical SLs contribute minimally to these developmental processes and serve mainly as rhizospheric signaling molecules (Wakabayashi et al., 2019; Ito et al., 2022). Methyl carlactonoate then binds directly to AtD14, the SL receptor, to promote signal transduction (Abe et al., 2014). These molecular mechanisms play a crucial role in inhibiting axillary shoot growth (Barbier et al., 2023). In Arabidopsis and rice, knockout mutants of MAX1 exhibit notably reduced plant height and increased shoot branching compared to wild‐type plants (Abe et al., 2014; Zhang et al., 2014). Similarly, in the tomato cultivars “MP‐1,” clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR‐associated protein 9 (Cas9)‐induced slmax1 mutants display reduced plant height, shorter root length, and increased shoot branching relative to the wild‐type plants (Bari et al., 2021). Loss‐of‐function mutations of D14 also induce shorter plant height and enhanced branching in Arabidopsis and rice (Seto et al., 2019; Hu et al., 2024). In the tomato cultivar “Micro‐Tom,” sld14 mutants exhibit a shorter stature and decreased chlorophyll content (Li et al., 2022). Despite these findings, most previous studies have focused on the molecular functions of MAX1 and D14, with limited exploration of their potential agricultural applications and the broader implications of SL pathway modulation for crop improvement.

In this study, SlD14 and SlMAX1, key genes involved in the SL pathway, were edited using the CRISPR–Cas9 system in cherry tomato cultivar “Sweet100” triple‐determinate, which carries mutations in SELF PRUNING (SP), SP5G, and SlERECTA (SlER) (Kwon et al., 2020). For phenotypic characterization, plants were grown and evaluated under three independent cultivation systems, which included ebb‐and‐flow hydroponics, the nutrient film technique (NFT), and cocopeat media. The sld14, slmax1, and sld14 slmax1 mutant plants exhibited rapid shoot branching, reduced shoot length, and delayed inflorescence emergence, ultimately leading to delayed fruit ripening. Additionally, excessive shoot branching in these loss‐of‐function mutants negatively impacted agricultural productivity, resulting in lower fruit yield per plant. However, we found that plants carrying a weak allele exhibited a milder phenotype, allowing agricultural productivity to be maintained. These findings suggest that precise modifications of SL biosynthetic and signaling pathways can be leveraged to optimize plant architecture and enhance productivity, particularly for vertical farming and urban agriculture.

RESULTS

A hypothetical model for optimizing shoot branching and identification of SlD14 and SlMAX1 orthologs

To increase per‐plant productivity by optimizing axillary shoot development in tomato, we focused on two crucial genes, SlD14 and SlMAX1, in the SL pathway (Figure 1A). Initially, we confirmed the presence of these genes in the tomato genome. Phylogenetic analysis with Arabidopsis, rice, eggplant, groundcherry, potato, petunia, pepper, and tobacco revealed that tomato contains a single SlD14 and a single SlMAX1 (Figure S1A, B). Moreover, protein sequence comparisons and motif analyses demonstrated that the location and composition of these motifs are highly conserved across various species, suggesting that the functions of SlD14 and SlMAX1 have remained largely unchanged throughout plant evolution (Figures S1C–F, S2, S3).

Figure 1.

A model for shoot branching optimization and expression patterns of SlD14 and SlMAX1

(A) Hypothetical model illustrating shoot branching optimization through modulation of the SL pathway. (B, C) Normalized read counts of SlD14 (B) and SlMAX1 (C) across different tissues and meristems, based on public gene expression databases for M82. RPKM, reads per kilobase of transcript per million mapped reads. (D) Normalized expression levels of SlD14 and SlMAX1 in root, stem, leaf, flower, fruit, and shoot apex of triple‐determinate plants using quantitative real‐time polymerase chain reaction (qRT‐PCR). Expression levels were normalized to SlUBQ3. SL, strigolactone.

Next, we analyzed the expression patterns of SlD14 and SlMAX1 across various tissues using global gene expression profiles based on the M82 genome (Park et al., 2012; Koenig et al., 2013). Both genes were expressed in most major tissues and meristems (Figure 1B, C). SlD14 showed predominant expression in the root and leaf, with relatively lower levels in the stem, whereas SlMAX1 exhibited high expression in the stem (Figure 1B, C). Additionally, both genes were consistently expressed across different stages of meristem development, with SlD14 generally showing higher expression levels than SlMAX1 (Figure 1B, C). We further evaluated the expression of these genes in Sweet100 triple‐determinate plants and observed similar patterns to those in the public database, although SlD14 expression in the root was relatively lower than in the leaf (Figure 1D). Given the genetic conservation and functional roles of SlD14 and SlMAX1 in regulating shoot architecture, we selected these genes for targeted modification to influence axillary shoot development in tomato.

Mutations in SlD14 and SlMAX1 led to accelerated axillary shoot development and reduced plant height

To investigate the role of SlD14 and SlMAX1 in regulating shoot architecture, we utilized the CRISPR–Cas9 system to generate targeted genetic modifications in the Sweet100 triple‐determinate variety. For SlD14 disruption, three guide RNAs (gRNAs) were designed, resulting in two knockout alleles and one weak allele that retained partial functionality due to an in‐frame mutation (Figure 2A, B). Structural modeling of the SlD14 protein confirmed that the frameshift mutations introduced premature stop codons, leading to truncated proteins (Figure S4A). In contrast, the six‐base pair in‐frame mutation resulted in the loss of two amino acids, suggesting partial conservation of protein function (Figure S4A). Similarly, SlMAX1 was targeted using three gRNAs, generating two independent loss‐of‐function mutations, both of which resulted in premature stop codons and truncated proteins (Figures 2C, D, S4B).

Figure 2.

CRISPR‐generated mutations and phenotypic characterization of sld14 , slmax1 , and sld14 slmax1 plants grown under ebb‐and‐flow hydroponics

(A–D) Gene structures and guide RNA (gRNA) sequences of SlD14 and SlMAX1. Black boxes indicate exons (A, C), red arrows mark the gRNAs (A, C), and blue dashes and letters denote deletions and insertions (B, D). (E) Shoots of triple‐determinate, sld14, slmax1, and sld14 slmax1 plants. (F) Time‐course measurement of axillary shoot number in triple‐determinate, sld14, slmax1, and sld14 slmax1 plants. Asterisks on the lines mark statistically significant difference (*P < 0.05, one‐way analysis of variance (ANOVA) and Tukey test). (G–L) Quantitative phenotypic traits of triple‐determinate, sld14, slmax1, and sld14 slmax1 plants, including axillary shoot number at two time points (G), plant height (H), stem internode length (I), first sympodial shoot length (distance between the first and second inflorescences) (J), number of leaves to the first inflorescence (K), and total inflorescence number at two time points (L). n, number of plants. DAS, d after sowing. (E–L) All data were collected from trials between February and July 2024. (G–L) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. The letters on the box plots indicate significance at P < 0.01 (one‐way ANOVA and Tukey test). Different letters between genotypes represent statistical significance.

The gene‐edited plants were grown under ebb‐and‐flow hydroponics, and phenotypic analyses revealed significant differences in shoot branching among sp sp5g sler (triple‐determinate), sp sp5g sler sld14 (hereafter sld14), sp sp5g sler slmax1 (hereafter slmax1), and sp sp5g sler sld14 slmax1 (hereafter sld14 slmax1) plants (Figures 2E, S5A). We first confirmed that the CRISPR‐generated mutants of each allele showed identical phenotypes. Further phenotypic characterization was performed using sld14 ^CR‐1 single, slmax1 ^CR‐1 single, and sld14 ^CR‐1 slmax1 ^CR‐1 double mutants. We found that sld14, slmax1, and sld14 slmax1 plants exhibited increased shoot branching, characterized by earlier axillary shoot initiation and a greater number of axillary shoots (Figure 2F, G). Specifically, these mutants initiated axillary shoots earlier and maintained a significantly higher shoot number than triple‐determinate plants, with statistically significant differences observed by 61 d after sowing (DAS) (Figure 2F). Additionally, at the middle growth stage (56 DAS), sld14, slmax1, and sld14 slmax1 plants displayed longer axillary shoots, further supporting their early and rapid shoot branching phenotype (Figure S5B–D). These mutants also exhibited reduced plant height and distinct stem morphology, including significantly shortened stem internode and sympodial shoot lengths compared to triple‐determinate plants (Figure 2H–J).

We next hypothesized that alterations in shoot architecture could impact other phenotypic traits, such as inflorescence formation and development. To evaluate this, we measured the number of leaves to the first inflorescence and observed that slmax1 and sld14 slmax1 plants produced one or two more leaves before transitioning to flowering compared to triple‐determinate plants, suggesting a potential delay in floral transition associated with SlMAX1 function (Figure 2K). Furthermore, we tracked total inflorescence production and found that triple‐determinate plants produced a greater number of inflorescences than slmax1 and sld14 slmax1 plants at 54 DAS, indicating that delayed inflorescence formation may reduce total inflorescence number at the middle growth stage (Figures 2L, S5E, F). However, by the late growth stage (89 DAS), the total inflorescence numbers of sld14, slmax1, and sld14 slmax1 plants surpassed that of triple‐determinate plants, consistent with their increased axillary shoot number (Figures 2L, S5F). These findings demonstrate that modifications to SlD14 and SlMAX1 influence shoot architecture by promoting early axillary shoot development and altering internode elongation, with potential implications for floral development and overall plant productivity.

Yield trials revealed negative effects on the agronomic and physiological traits of sld14 and slmax1 plants

To evaluate how synchronous shoot branching and changes in inflorescence number affect agronomic traits, we examined yield and yield‐related traits in sld14, slmax1, and sld14 slmax1 plants. Although these mutant plants exhibited increased shoot branching and a greater total number of inflorescences, their total yield was generally lower than that of triple‐determinate plants (Figures 2L, 3A, B). Furthermore, the plant weight excluding fruits was higher in sld14, slmax1, and sld14 slmax1 plants than in triple‐determinate plants, leading to a lower harvest index in these mutant plants (Figure 3C, D). At harvest, sld14, slmax1, and sld14 slmax1 plants produced fewer red fruits and a greater number of green fruits compared to triple‐determinate plants (Figure 3E–G). These results suggest that excessive axillary shoot development caused by mutations in SlD14 and SlMAX1 contributed to delayed fruit ripening.

Figure 3.

Yield performance of sld14 , slmax1 , and sld14 slmax1 plants under ebb‐and‐flow hydroponics

(A) Shoots of triple‐determinate, sld14, slmax1, and sld14 slmax1 plants at 111 d after sowing (DAS). (B–G) Yield and yield‐related traits of triple‐determinate, sld14, slmax1, and sld14 slmax1 plants, including total fruit yield (B), plant weight (C), harvest index (total yield/plant weight) (D), red fruit weight per plant (E), green fruit weight per plant (F), and proportion of red fruit in total fruit yield (G). (H–K) Photosynthetic pigment analysis of triple‐determinate, sld14, slmax1, and sld14 slmax1 plants, including Soil Plant Analysis Development (SPAD) value (H), total chlorophyll content (I), total carotenoid content (J), chlorophyll a/b ratio (K). (A–K) All data were collected from trials between February and July 2024. (B–K) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. The letters on the box plots indicate the significant groups at P < 0.01 (one‐way analysis of variance and Tukey test). Different letters between genotypes represent statistical significance. n, number of plants (B–G) and samples (H–K).

To further examine the effects of delayed fruit ripening on seed quality, we analyzed seed yield and germination. We found that seeds from triple‐determinate and sld14 plants exhibited similar germination rates, with only a slight difference observed at 4 DAS (Figure S6A, B). In contrast, seeds from slmax1 and sld14 slmax1 plants showed lower germination rates compared to those from triple‐determinate and sld14 plants (Figure S6A, B). We also measured seed yield per fruit and found that the total number of seeds per fruit was reduced in sld14, slmax1, and sld14 slmax1 plants relative to triple‐determinate plants (Figure S6C, D). These findings suggest that disruption of SL biosynthesis and signaling negatively impacts seed viability and production.

To further investigate factors beyond excessive shoot branching that may contribute to reduced fruit productivity, we examined phenotypic traits related to photosynthesis. Given the pronounced expression of SlD14 and SlMAX1 in leaves (Figure 1B–D), we aimed to determine whether mutations in these genes affect leaf biochemical properties. We first performed a Soil Plant Analysis Development (SPAD) assay to assess chlorophyll content non‐destructively and found that sld14, slmax1, and sld14 slmax1 plants showed lower SPAD values than triple‐determinate plants (Figure 3H). Next, we conducted destructive biochemical analyses to quantify total chlorophyll content, total carotenoid content, and the chlorophyll a/b ratio (Figure 3I–K). The results showed that triple‐determinate plants had the highest total chlorophyll and carotenoid content among the all the genotypes (Figure 3I, J). The chlorophyll a/b ratio remained unchanged between triple‐determinate and sld14 plants but was reduced in slmax1 and sld14 slmax1 mutants (Figure 3K). These findings suggest that the reduced yield observed in sld14, slmax1, and sld14 slmax1 plants is likely attributed to both excessive shoot branching and decreased leaf chlorophyll content, collectively impairing overall plant productivity.

Cocopeat cultivation mitigates the agronomic and physiological defects of sld14

We verified that loss‐of‐function mutations in two key SL pathway factors severely compromise fruit productivity and aimed to find a strategy to mitigate this negative effect. Excessive shoot branching was presumed to cause a nutrient imbalance in the plant, so we altered the growing conditions to address this issue. Specifically, we compared the phenotypes of triple‐determinate and sld14 plants grown on cocopeat slabs with drip irrigation, which is considered to be better suited for tomato cultivation than ebb‐and‐flow hydroponics (2, 3, 4, S5, S7). We found that sld14 mutants developed axillary shoots more rapidly and maintained a consistently higher number of these shoots into later growth stages (Figure 4A, B). This increased branching led to more total inflorescences in sld14 plants (Figures 4C, S7B–C). Additionally, sld14 plants had shorter main shoots, shorter first sympodial shoots, and reduced internode lengths compared with triple‐determinate plants, although there was no difference in the number of leaves to the first inflorescence (Figure 4D–G). Overall, the shoot architecture of sld14 grown on cocopeat closely resembled that of sld14 grown in ebb‐and‐flow hydroponics (2, 4).

Figure 4.

Phenotypic characterization of triple‐determinate and sld14 plants under cocopeat cultivation

(A) Shoots of triple‐determinate and sld14 plants grown on a cocopeat slab at 115 d after sowing (DAS). (B, C) Time‐course measurements of axillary shoot number (B) and total inflorescence number (C) in triple‐determinate and sld14 plants. (D–G) Comparison of shoot architecture traits between triple‐determinate and sld14 plants, including primary shoot length (D), first sympodial shoot length (distance between the first and second inflorescences) (E), stem internode length (F), and number of leaves to the first inflorescence (G). (H) Harvested fruits from triple‐determinate and sld14 plants. (I–L) Yield and yield‐related traits of triple‐determinate and sld14 plants, including total fruit yield (I), plant weight (J), harvest index (total yield/plant weight) (K), and proportion of red fruit in total fruit yield (L). (A–L) All data were collected from trials between September and December 2023. (B, C) P‐values were calculated using a two‐sample t‐test (*P < 0.05, **P < 0.01, ***P < 0.001). (D–G, I–L) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. The numbers indicate P‐values (two‐tailed, two‐sample t‐test). Different letters between genotypes denote statistically significant differences. n.s., not significant; n, number of plants.

Next, we assessed fruit productivity in triple‐determinate and sld14 plants grown on cocopeat (Figures 4H–K, S7D–I). The total yield of sld14 plants did not significantly differ from triple‐determinate plants, although sld14 plants had a higher plant weight and thus a lower harvest index (Figure 4H–K). Notably, when the triple‐determinate plants had reached about 90% fruit ripeness, sld14 plants were only at around 40% (Figures 4L, S7D, E). These findings indicate that while improved growing conditions can partially enhance the yield of sld14, delayed fruit ripening remains a significant challenge. Further analysis revealed that sld14 plants produced smaller and lighter fruits than triple‐determinate plants, although sugar content remained unchanged between genotypes (Figure S7F–I). Photosynthesis‐related traits revealed that sld14 plants had lower SPAD values, but similar total chlorophyll, total carotenoid, and chlorophyll a/b ratios (Figure S7J–M). Collectively, these findings indicate that sld14 plants cultivated in cocopeat media with drip irrigation showed improved yield performance compared to those grown under ebb‐and‐flow hydroponics (3, 4, S7). However, persistent issues such as delayed fruit ripening suggest that genetic fine‐tuning, in addition to optimizing growing conditions, is required to fully enhance crop traits.

The sld14 in‐frame mutation exhibits a milder phenotype compared to the sld14 frameshift mutation

We hypothesized that screening plants with moderated phenotypes via genetic modification, together with improved growing conditions, could mitigate the negative effects by fine‐tuning the SL pathway. To test this, we focused on the in‐frame CRISPR alleles of SlD14 (hereafter sld14 ^inf ) (Figures 2B, S4A). To validate the stability of this phenotype across generations, we performed three independent cultivation trials using ebb‐and‐flow hydroponics (Figures 5, S8). Overall, sld14 ^inf plants exhibited a less severe shoot architecture phenotype than sld14 and slmax1 null mutants, with moderated shoot branching and stem elongation, although some variation was observed (Figures 5, S8).

Figure 5.

Phenotypic characterization of sld14 ^inf plants grown under ebb‐and‐flow hydroponics

(A) Shoots of triple‐determinate, sld14, sld14 ^inf , and slmax1 plants. (B–H) Quantitative analysis of phenotypic traits in triple‐determinate, sld14, sld14 ^inf , and slmax1 plants, including axillary shoot number (B), stem internode length (C), first sympodial shoot length (distance between the first and second inflorescences) (D), number of leaves to the first inflorescence (E), total inflorescence number (F), inflorescence number on the main shoot (G), and inflorescence number on axillary shoots (H). (A–H) All data were collected from trials between September and December 2023. (B–H) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. Different letters indicate statistical significance among genotypes (P < 0.01, one‐way analysis of variance and Tukey test). DAS, d after sowing. n, number of plants.

A detailed phenotypic analysis revealed that sld14 ^inf plants had a higher number of axillary shoots than triple‐determinate plants, but this increase was comparable to or slightly lower than that observed in sld14 and slmax1 null mutants (Figures 5B, S8B, G). Similarly, sld14 ^inf plants displayed shorter stem internode and sympodial shoot lengths than triple‐determinate plants, although these traits were slightly elongated compared to sld14 and slmax1 null mutants (Figures 5C, D, S8C, D, H, I). No significant differences were detected in the number of leaves preceding the first inflorescence among triple‐determinate, sld14, and sld14 ^inf plants, while slmax1 mutants exhibited a slightly higher leaf number than the other genotypes (2, 5, S8E, J). Notably, the moderated shoot branching observed in sld14 ^inf plants also influenced total inflorescence production at late growth stages (85 DAS), resulting in sld14 ^inf plants exhibiting an intermediate inflorescence number among sld14, slmax1, and triple‐determinate plants (Figure 5F–H). These findings suggest that precise modulation of the SL pathway through the development of weak alleles can effectively fine‐tune shoot architecture, offering a promising strategy for improving plant productivity while mitigating the detrimental effects of excessive branching.

We next hypothesized that the attenuated shoot architecture phenotype of the sld14 ^inf plants could mitigate the severe effects observed in loss‐of‐function mutants. To test this, we conducted two independent yield trials using ebb‐and‐flow hydroponics (Figures 6, S9). At harvest, sld14 and sld14 ^inf plants exhibited significantly greater shoot branching than triple‐determinate plants (Figure 6A). Notably, while sld14 plants still had a substantial number of unfertilized flowers, sld14 ^inf plants produced a higher proportion of developed fruits (Figure 6A). We found that sld14 ^inf plants also exhibited higher total fruit yield compared to sld14 plants, while plant weight remained relatively similar, resulting in an improved harvest index (Figure 6B–D). Although some variation was observed between the spring and fall trials due to differences in cultivation environments, sld14 ^inf plants consistently exhibited a higher harvest index and a greater proportion of red fruit relative to total yield compared to sld14 and slmax1 null mutants (Figure S9A–D). These findings suggest that the weak sld14 allele may alleviate the severe phenotypic effects of complete gene disruption and ultimately enhance agronomic traits.

Figure 6.

Yield performance of sld14 ^inf plants grown under ebb‐and‐flow hydroponics

(A) Shoots of triple‐determinate, sld14, and sld14 ^inf plants at 102 d after sowing (DAS). (B–D) Yield and yield‐related traits, including total fruit yield (B), plant weight (C), and harvest index (total yield /plant weight) (D) in triple‐determinate, sld14, and sld14 ^inf plants. (E) Harvested fruits from triple‐determinate, sld14, and sld14 ^inf plants. (F–H) Comparison of red fruit weight per plants (F), green fruit weight per plant (G), and proportion of red fruit in total fruit yield (H) among triple‐determinate, sld14, and sld14 ^inf plants. (I–L) Photosynthetic pigment analysis of triple‐determinate, sld14, and sld14 ^inf plants, including Soil Plant Analysis Development (SPAD) value (I), total chlorophyll content (J), total carotenoid content (K), and chlorophyll a/b ratio (L). (B–L) All data were collected from trials between February and July 2024. (B–D, F–L) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. Different letters between genotypes represent statistically significant differences (P < 0.01, one‐way analysis of variance and Tukey test). n, number of plants (B–D, F–H) and samples (I–L).

To further examine the positive impact of the weak sld14 allele on fruit yield, we evaluated the proportion of red and green fruits at harvest. We observed that sld14 ^inf plants produced a significantly higher number of red fruits than sld14 null mutants, with the proportion of red fruits increasing to approximately 80% of the total fruit yield (Figure 6E–H). This trend was consistently observed across both yield trials (Figure S9E–H), suggesting that the weak sld14 allele may contribute to moderate axillary shoot formation, which could facilitate synchronized fruit ripening and contribute to an increase in total yield. Next, we investigated how fruit weight and size varied among genotypes. While sld14 ^inf plants exhibited increased fruit weight and size compared to sld14 plants, these variations were likely influenced by environmental factors (Figure S10A–C, E–G). Furthermore, no significant differences in fruit sugar content were observed among any of the genotypes (Figure S10D, H).

To determine whether the improved agronomic traits of sld14 ^inf plants were associated with biochemical changes, we measured SPAD values, total chlorophyll content, total carotenoid content, and the chlorophyll a/b ratio (Figures 6I–L, S11A, B). Interestingly, sld14 ^inf plants exhibited lower SPAD values, total chlorophyll, and total carotenoid content than triple‐determinate plants but higher levels than sld14 null mutants (Figure 6I–K). This pattern is consistent with other phenotypic observations of sld14 ^inf plants and suggests that the strong phenotypic effects of loss‐of‐function mutations can be mitigated through in‐frame mutations. Finally, we analyzed photosynthetic efficiency by measuring variable chlorophyll fluorescence/maximum chlorophyll fluorescence (F _v /F _m ). F _v /F _m ratios showed a similar trend to SPAD values, although variation among genotypes was relatively minor (Figure S11C, D), suggesting that differences in chlorophyll content had a moderate impact on photosynthetic efficiency.

Application of sld14 ^inf plants in a vertical farming system

The moderated phenotype of sld14 ^inf plants highlights their potential for application in vertical farming systems. Specifically, sld14 ^inf plants exhibit a more compact stature than triple‐determinate plants, which is expected to enhance their adaptability to vertical farms characterized by multi‐layered, space‐restrictive cultivation environments (5, 6, S8A). To further evaluate whether rapid shoot branching in vertical farming could contribute to increased fruit productivity, we established a vertical farm system incorporating NFT hydroponics for tomato cultivation. We confirmed that sld14 ^inf plants were more suitable in size for restrictive cultivation spaces than triple‐determinate plants (Figure 7A–E). Notably, while trends in shoot and stem internode length remained consistent with those observed in greenhouse conditions using ebb‐and‐flow hydroponics, the timing of first inflorescence emergence was slightly delayed in sld14 ^inf plants compared to triple‐determinate plants (Figure 7B–F). Additionally, axillary shoot development was accelerated in sld14 ^inf plants, leading to a higher total number of inflorescences at the late growth stage (94 DAS) than in triple‐determinate plants (Figures 7A, G, H, S12A, B).

Figure 7.

Phenotypic characterization and yield performance of sld14 ^inf plants in a vertical farm

(A) Triple‐determinate and sld14 ^inf plants grown in the vertical farm. (B–H) Quantitative phenotypic traits of triple‐determinate and sld14 ^inf plants, including plant height (B), primary shoot length (C), first and second sympodial shoot length (distance between consecutive inflorescences) (D), stem internode length (E), number of leaves to the first inflorescence (F), axillary shoot number (G), total inflorescence number (H). (I) Harvested fruits from triple‐determinate and sld14 ^inf plants. (J–M) Yield and yield‐related traits of triple‐determinate and sld14 ^inf plants, including total fruit yield (J), plant weight (K), harvest index (total yield /plant weight) (L), proportion of red fruit per total fruit yield (M). (N) Mature red fruits from triple‐determinate and sld14 ^inf plants. (O, P) Individual fruit weight (O) and sugar content (P) comparisons between triple‐determinate and sld14 ^inf plants. (A–P) All data were collected from trials between July and October 2024. (B–H, J–M, O, P) Box plots, 25^th–75^th percentile; center line, median; whiskers, full data range. The numbers on the boxplots indicate P‐values (two‐tailed, two‐sample t‐test). n, number of plants (B–H, J–M) and fruits (O, P). DAS, days after sowing.

We next compared the yield and yield‐related traits of triple‐determinate and sld14 ^inf plants grown in the vertical farm system. The results showed no significant differences between the two genotypes in total fruit yield, plant weight, or harvest index (Figure 7I–L). Similarly, red fruit weight, green fruit weight, and the proportion of red fruit per total yield did not significantly differ between sld14 ^inf and triple‐determinate plants (Figures 7M, S12C, D). Despite the lack of differences in total fruit yield, individual fruit weight and size were reduced in sld14 ^inf plants compared to triple‐determinate plants (Figures 7N, O, S12E). This suggests that sld14 ^inf plants produce a greater number of fruits due to an increased number of inflorescences (Figures 7H, I, S12A, B). We also found no significant differences in fruit sugar content between the two genotypes (Figure 7P). In summary, sld14 ^inf plants grown in a vertical farming system exhibited a significantly smaller plant size compared to triple‐determinate plants, while maintaining comparable total fruit yield and ripening rates. These findings suggest that precise modulation of SL pathway genes by generating weak alleles could serve as a viable strategy for developing crops optimized for vertical farming systems.

DISCUSSION

Trait improvement can be achieved via precision modification of the SL pathway

In this study, we targeted SlD14 and SlMAX1, two key genes involved in SL signaling and biosynthesis that regulate shoot branching. Using a gene and trait stacking strategy with CRISPR–Cas9 technology, we aimed to optimize axillary shoot development in triple‐determinate plants to maximize productivity per individual. Additionally, we sought to synchronize fruit production by coordinating shoot and inflorescence development. Simultaneous fruit maturation is a critical trait for fruit crops intended for vertical farming applications (Kwon, 2023), aligning with the concept of temporary harvesting in open‐field tomato cultivation using mechanical harvesters. Harvesting all fruits within a unit cultivation area at the same time can reduce labor demands and limit worker entry, thereby minimizing the risk of pest introduction and spread (Kwon, 2023).

The sld14, slmax1, and sld14 slmax1 plants exhibited shorter stem length and a higher total inflorescence number than triple‐determinate plants (Figure 2). However, these three genotypes also showed reduced fruit fertilization rates, leading to lower total yield and significantly delayed fruit maturation (Figure 3). The observed reductions in total yield, slow fruit ripening, and smaller fruit size and weight in sld14, slmax1, and sld14 slmax1 plants are likely attributed to nutrient imbalances resulting from excessive axillary shoot development (Rameau et al., 2015). Interestingly, the sld14 ^inf plants, carrying a weak allele, exhibited moderated axillary shoot development and showed significant improvements in yield and yield‐related traits compared to sld14 and slmax1 null mutants (5, 6, S8, S9, S10). These findings suggest that fine‐tuning shoot branching regulation could enable the development of high‐yielding varieties with superior productivity per plant compared to existing cultivars (Wang et al., 2018; Fang et al., 2025). Therefore, targeting cis‐regulatory elements, introducing single or multiple amino acid substitutions, or generating weak alleles via epigenetic regulation represent promising approaches facilitated by advancements in gene editing technologies and genome analysis techniques (Hendelman et al., 2021; Jogam et al., 2022; Zhou et al., 2023; Vu et al., 2024). Furthermore, optimizing multiple genes within the noncanonical SL biosynthesis and signaling pathways could further enhance plant trait improvements (Butt et al., 2018; Wang et al., 2020; Ito et al., 2022; Barbier et al., 2023; Kelly et al., 2023; Ban et al., 2025).

Integrating genetic and environmental factors for optimized crop cultivation

Crop traits result from complex interactions between genetic and environmental factors, necessitating an integrated approach to improve productivity and adaptability (Peng et al., 2020). Achieving high‐quality crop production requires both precision breeding and the establishment of optimal growing conditions to sustain maximum agricultural productivity. With rapid advancements in automated and precision agriculture, cultivation systems integrated with ICT‐based smart farming technologies are becoming increasingly sophisticated (Van Delden et al., 2021).

Our findings highlight the critical role of growing conditions in shaping agronomic traits. Notably, the yield of sld14 plants improved significantly when cultivated in cocopeat media with drip irrigation compared to ebb‐and‐flow hydroponics, although their yield and fruit ripening remained lower and later than those of triple‐determinate plants (Figures 4, S7). Furthermore, no significant differences in total yield, harvest index, and red fruit proportion were observed between triple‐determinate and sld14 ^inf plants when grown in a vertical farm system (Figures 7J–M, S12C, D). Although sld14 ^inf plants produced smaller individual fruits compared to triple‐determinate plants, their total yield remained comparable, likely due to an increased fruit number per plant (Figures 7N, O, S12E). Collectively, these results suggest that sld14 ^inf plants are better suited for indoor vertical farming systems than for conventional greenhouse cultivation (6, 7, S10, S12). These findings emphasize the importance of customized cultivation strategies tailored to specific crop species and cultivars.

Adaptive cultivation protocols should be designed to flexibly accommodate different geographic regions and farming techniques. This approach can be further enhanced through the integration of genetic and agronomic technologies with artificial intelligence‐driven big data analytics (Aijaz et al., 2025). Furthermore, because yield and yield‐related traits are highly sensitive to environmental factors, they exhibit substantial variation across individual plants and growing seasons (6, 7, S9, S10). This underscores the critical need to refine cultivation conditions to achieve phenotypic uniformity and maximize crop productivity.

Epistatic interactions between SlD14 and SlMAX1

In this study, we analyzed the sld14 slmax1 double mutant, generated through genetic crossing of sld14 and slmax1 plants, to investigate the epistatic relationship between these two genes (2, 3, S5). Phenotypic comparisons revealed no significant differences among sld14, slmax1, and sld14 slmax1 plants in plant height‐related traits, including primary shoot length, sympodial shoot length, and stem internode length (Figures 2H–J, S5B–D). Similarly, axillary shoot development and total inflorescence number showed no significant differences among the three genotypes, despite minor variations (Figure 2G, L). However, leaf number to the first inflorescence was higher in slmax1 and sld14 slmax1 plants compared to sld14 plants (Figures 2K, 5E), suggesting a flowering‐related phenotype specific to slmax1. The similarity in flowering phenotypes between slmax1 and sld14 slmax1 plants indicates that slmax1 is epistatic to sld14.

Strigolactones are known to inhibit flowering in Arabidopsis (Bai et al., 2024). In Arabidopsis, SL‐deficient mutants such as max2, max3, and Atd14 show accelerated flowering, supporting the role of SLs as negative regulators of floral induction (Bai et al., 2024). However, in tomato, SLs appear to function oppositely, promoting flowering by activating the miR319‐LANCEOLATE module, which induces the expression of the florigen SINGLE FLOWER TRUSS (SFT) (Visentin et al., 2024). This aligns with our findings that slmax1 mutants exhibited delayed flowering (Figures 2K, 5E), likely due to drastically reduced endogenous SL levels (Zhang et al., 2018), which in turn may affect SFT expression. Since SlD14 acts as an SL receptor rather than being directly involved in SL biosynthesis, flowering times were not significantly altered in sld14 mutants. Notably, in rice, d14 mutants exhibit increased SL content (Arite et al., 2009), further supporting that loss of SlD14 in tomato does not promote flowering. The contrasting roles of SLs in flowering regulation across species may stem from divergent genetic and physiological mechanisms in evolutionarily distinct plant lineages (Lozano et al., 2009; Jarillo and Piñeiro, 2011). A comparative analysis of genetic and physiological variations across species will be crucial to elucidating these differences.

Pleiotropic effects of SL pathway modifications and their applications for crop improvement

Modifications in the SL pathway exert pleiotropic effects beyond plant architecture, influencing multiple physiological traits. For instance, our findings revealed that sld14, sld14 ^inf , slmax1, and sld14 slmax1 plants showed slightly lower SPAD values, total chlorophyll content, and carotenoid content compared to triple‐determinate plants (3, 6, S7, S11). These results align with previous studies demonstrating that SlD14 plays a critical role in photosynthetic pigment accumulation and photosynthetic capacity (Li et al., 2022). A prior study on sld14 mutants in Micro‐Tom reported that these mutants displayed reduced chlorophyll content and F _v /F _m ratios compared to wild‐type plants (Li et al., 2022). This reduction in photosynthetic performance was attributed to downregulation of chlorophyll a/b‐binding protein genes, key components of the light‐harvesting complex in Micro‐Tom sld14 mutants (Li et al., 2022). Furthermore, the SL pathway regulates chlorophyll content and photosynthetic efficiency (Khalid et al., 2024), which may explain the relatively low chlorophyll content and F _v /F _m ratios observed in sld14 and slmax1 mutants (Figures 3, S7, S11). However, sld14 ^inf plants maintained F _v /F _m ratios comparable to those of triple‐determinate plants, suggesting that the weak allele effectively mitigated these negative effects, consistent with trends observed in yield and yield‐related traits (6, 7, S11C, D).

Previous studies have reported that disruption of the SL pathway can impair seed viability and quality (Yamada et al., 2019). In our study, sld14 mutants showed germination rates comparable to the triple‐determinate control, whereas slmax1 mutants exhibited a clear reduction in germination (Figure S6A, B). Furthermore, sld14, slmax1, and sld14 slmax1 plants retained ∼75%–80% of the seed count per fruit relative to the control (Figure S6C, D). These findings indicate that while SL pathway disruption can affect seed performance, the extent depends on the specific gene involved. For precision crop improvement, it is therefore essential to identify SL‐related genes with minimal pleiotropic effects and to fine‐tune gene expression through targeted modifications of cis‐regulatory regions (Hendelman et al., 2021; Zhou et al., 2023).

Given the broad regulatory role of SLs, a comprehensive understanding of SL pathway components is essential for bioengineering strategies that minimize unintended phenotypic trade‐offs (Wang et al., 2025; Zhou et al., 2025). With climate change intensifying environmental stresses, research has increasingly focused on understanding plant resilience mechanisms. Emerging studies confirm that SLs contribute significantly to stress tolerance by modulating key physiological and developmental pathways (Wang and Xi, 2022; Khalid et al., 2024; Ban et al., 2025). These insights underscore the need to elucidate the genetic factors within the SL pathway and their association with agronomic traits to enhance crop resilience and productivity. Finally, our study demonstrates that precise modifications of SL biosynthesis and signaling can be harnessed to develop crops optimized for diverse cultivation environments, extending beyond conventional field agriculture to vertical farming systems.

MATERIALS AND METHODS

Plant materials and growth conditions

Plants were grown in a greenhouse at Kyung Hee University, Yongin, Korea. Seedlings were managed using 40‐cell seedling trays following established protocols (Lim et al., 2024), and transplanted after 28–40 DAS into ebb‐and‐flow beds, cocopeat slabs with drip irrigation, or a hydroponic vertical farm. A hydroponic fertilizer (S‐FEED 15‐30‐15, Farm Hannong, Korea) was applied before transplanting and adjusted at different growth stages to optimize nutrient availability (Lim et al., 2024). Electrical conductivity (EC) levels were adjusted from 1.0 to 2.2 dS/m. Plants were grown under long‐d conditions (16‐h photoperiod, 24°C–26°C day, 20°C–22°C night, relative humidity of 40%–60%) with artificial lighting provided by high‐pressure sodium bulbs. The vertical farm consisted of four layers, each measuring 25 cm in height, 90 cm in width, and 30 cm in length, with a planting density of ∼150 cm² per plant. Nutrient solution was provided through the NFT, at 15‐min intervals and a 165‐min rest period. EC was gradually increased from 0.7 to 1.3 dS/m during the experiment. Plants were cultivated under long‐d conditions and were illuminated with red/blue LED lighting at 400–450 μmol/m²/s. Plants exhibiting signs of disease or shoot damage were excluded from the data analysis.

CRISPR–Cas9 genome editing and plant transformation

CRISPR–Cas9‐based tomato transformation was performed by following established protocols (Brooks et al., 2014; Van Eck et al., 2019). All gRNAs targeting SlD14 and SlMAX1 were designed using CRISPRdirect software (https://crispr.dbcls.jp/), and binary vectors were constructed using the Golden Gate cloning technique, then introduced into tomato seedlings by Agrobacterium tumefaciens‐mediated transformation (Van Eck et al., 2019). Transgenic mutants were selected with kanamycin (400 mg/L) and genotyped using specific primers (Table S1). Genomic DNA was extracted from at least three individual leaf samples from each T₀ individual. All primers and gRNA sequences are described in Table S1.

RNA extraction, complementary DNA synthesis and quantitative real‐time polymerase chain reaction

Sweet100 triple‐determinate plants were used to extract the total RNA. Established protocols with minor modifications were used, and normalized expression levels was quantified by quantitative real‐time polymerase chain reaction (qPCR) (Kwon et al., 2022). Briefly, total RNA was extracted from different tissues using the RNeasy Plant Mini Kit (Qiagen, Venlo, The Netherlands). After the extraction, 1 μg of total RNA was synthesized into complementary DNA (cDNA) using the iScript cDNA Synthesis Kit (Bio‐Rad, Hercules, USA). The iQ SYBR Green Supermix (Bio‐Rad, Hercules, USA) on a CFX96 Real‐Time PCR Detection System (Bio‐Rad, Hercules, USA) was used to perform the qPCR with gene‐specific primers (Table S1). For each biological replicate, five independent samples were pooled for stem, leaf, flower, fruit, and shoot apex tissues, while seven samples were pooled for root tissues. Tomato Ubiquitin3 (SlUBQ3) was used as internal control.

Plant phenotyping and imaging

Phenotypic characterization was performed on non‐transgenic homozygous tomato lines obtained through backcrosses or self‐pollination. Three transgene‐free homozygous sld14 alleles were obtained. Among them, sld14 ^CR‐1 and sld14 ^CR‐2 were null mutants, while the other was an in‐frame mutant (sld14 ^inf ). Since sld14 ^CR‐1 and sld14 ^CR‐2 showed identical phenotypes, sld14 ^CR‐1 was selected for further phenotypic analysis. Similarly, non‐transgenic homozygous slmax1 plants were generated, yielding two null mutants, slmax1 ^CR‐1 and slmax1 ^CR‐2 . As both showed the same phenotype, slmax1 ^CR‐1 was selected for further analysis. Homozygous sld14 slmax1 double mutants were generated through artificial crossing of sld14 ^CR‐1 with slmax1 ^CR‐1 . Shoot length and height were measured using a standard ruler (50 cm and 3 m, respectively), and the number of axillary shoots was counted when at least four leaves were fully developed. Inflorescence flowering time was recorded when the first flower fully bloomed, and the number of inflorescences on both the main and axillary shoots was counted. Data on inflorescence and axillary shoot numbers were collected at the same growth stage. Red fruit size and weight were measured using an electronic digital caliper (Mitutoyo, Kawasaki, Japan) and a digital scale (HANSUNG INSTRUMENT, Gwangmyeong, Korea), respectively. Sugar content was quantified by measuring the Brix value with a digital Brix refractometer (ATAGO, Tokyo, Japan). The number of samples (n) are indicated in all figures and details are provided in Tables S2–S6.

Seed germination and yield analysis

Seed germination was examined by placing seeds on moistened filter paper in Petri dishes and incubating them in the dark at 28°C. Each replicate consisted of 30 seeds, with three biological replicates per treatment. Seed yield was evaluated by counting the total number of seeds per fruit. The raw data are provided in Table S7.

Plant physiological analysis

Soil Plant Analysis Development values were measured using the SPAD‐502Plus device (Konica Minolta), and F _v /F _m ratios were evaluated using the LI‐600 fluorometer (LI‐COR, Lincoln, USA). Total chlorophyll, carotenoid content, and the chlorophyll a/b ratio were determined by sampling the first leaflet of the sixth and seventh leaves. Leaf extracts were prepared using an 80% acetone, and chlorophyll and carotenoid concentrations were measured using a UV/VIS spectrophotometer (Agilent, Santa Clara, USA) according to established protocols (Lichtenthaler and Buschmann, 2001). Sample numbers are given in the figures and in Table S8.

Gene annotation and accession numbers

Proteins identified with over 80% of SlD14 and SlMAX1 orthologs were selected from various crops (Table S9). The ortholog protein sequences were collected through the Sol Genomics Network (https://solgenomics.net/), The Arabidopsis Information Resource (https://www.arabidopsis.org/), Rice Genome Annotation Project (https://rice.uga.edu/), and the groundcherry genome assembly database (https://github.com/pan-sol/pan-sol-data/tree/main/Physalis).

Protein domain information is from STRING (https://string-db.org/cgi/input?sessionId=bnuM2I5Mv9I0&input_page_show_search=on), and protein structure data were from the AlphaFold Protein Structure Database (https://alphafold.ebi.ac.uk/) and generated using ANACONDA (https://www.anaconda.com/).

Phylogenetic analysis and expression analysis

A comparative phylogenetic tree was constructed by the maximum likelihood estimation (MLE) method using MEGA‐X software (https://www.megasoftware.net/). Bootstrap values from 1,000 replicates are shown at each node. Motif analysis of protein sequences was performed using Multiple Expectation Maximizations for Motif Elicitation (MEME; https://memesuite.org/meme/index.html). Multiple alignments were performed using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/-clustalo/). RNA sequencing data were obtained from Tomato eFP Browser (https://bar.utoronto.ca/efp_tomato/cgi-bin/efpWeb.cgi) and our previous RNA sequencing data (Park et al., 2012). All raw data for expression analysis are in Table S10.

Statistical analyses

Statistical analyses were performed using R (version 2022.12.0 + 353) in RStudio, Microsoft Excel, and the analysis of variance (ANOVA) calculator (https://www.statskingdom.com/180Anova1way.html#R). Data were analyzed by one‐way ANOVA with Tukey's post hoc test (Table S11).

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

J.L. performed the experiments, prepared the figures, and wrote the manuscript. M.‐G.S., Y.L., S.H., J.‐T.A., and H.‐Y.J. performed the tomato experiments and phenotypic characterization. C.L. helped perform the experiments. S.J.P. generated transgenic tomato plants and performed the tomato experiments. G.S. performed the experiments, prepared the figures, and wrote the manuscript. C.‐T.K. conceived the research, supervised and performed the experiments, prepared the figures, and wrote the manuscript. All authors read and approved the manuscript.

Supporting information

Additional Supporting Information may be found online in the supporting information tab for this article: http://onlinelibrary.wiley.com/doi/10.1111/jipb.70059/suppinfo

Figure S1. Phylogenetic analysis and multiple sequence alignment of SlD14 and SlMAX1

Figure S2. Multiple sequence alignment of SlD14 orthologs

Figure S3. Multiple sequence alignment of SlMAX1 orthologs

Figure S4. CRISPR‐Cas9‐induced alleles of SlD14 and SlMAX1

Figure S5. Axillary shoot and inflorescence phenotypes of sld14, slmax1, and sld14 slmax1 plants grown under ebb‐and‐flow hydroponics

Figure S6. Germination rates and seed yields of sld14, slmax1, and sld14 slmax1 plants

Figure S7. Phenotypic characterization and agricultural traits of sld14 plants grown on cocopeat slabs

Figure S8. Phenotypic comparison of triple‐determinate, sld14, and sld14 ^inf plants

Figure S9. Yield performance of sld14, sld14 ^inf , and slmax1 plants grown under ebb‐and‐flow hydroponics

Figure S10. Fruit weight, size, and sugar content of sld14, sld14 ^inf , slmax1, and sld14 slmax1 plants

Figure S11. Physiological analysis of sld14, sld14 ^inf , and slmax1 plants

Figure S12. Yield‐related traits of sld14 ^inf plants in a vertical farm

Table S1. Primers used in this study

Table S2. Length quantification of shoots and inflorescences

Table S3. Quantification of axillary shoot and inflorescence numbers

Table S4. Time‐course quantification of axillary shoot and inflorescence numbers

Table S5. Quantification of fruit weight, size, and brix

Table S6. Quantification of yield‐related traits

Table S7. Quantification of seed‐related traits

Table S8. Physiological analysis

Table S9. Protein sequences of SlD14 and SlMAX1 orthologs

Table S10. Expression levels of SlD14 and SlMAX1

Table S11. Exact P‐values in this study

Supplemental_file_01‐SlD14_CR‐1_guide_1‐2.ab1

Supplemental_file_02‐SlD14_CR‐1_guide_3.ab1

Supplemental_file_03‐SlD14_CR‐2_guide_1‐2.ab1

Supplemental_file_04‐SlD14_CR‐2_guide_3.ab1

Supplemental_file_05‐SlD14_inf_guide_1‐2.ab1

Supplemental_file_06‐SlD14_inf_guide_3.ab1

Supplemental_file_07‐SlMAX1_CR‐1_guide_1‐2.ab1

Supplemental_file_08‐SlMAX1_CR‐1_guide_3.ab1

Supplemental_file_09‐SlMAX1_CR‐2_guide_1‐2.ab1

Supplemental_file_10‐SlMAX1_CR‐2_guide_3.ab1

Biographies

Lee, J. , Seo, M. G. , Lim, Y. , Hong, S. , An, J. T. , Jeong, H. Y. , Lee, C. , Park, S. J. , Song, G. , and Kwon, C. T. (2026). Modulating the strigolactone pathway to optimize tomato shoot branching for vertical farming. J. Integr. Plant Biol. 68: 113–129.

Data availability statement

Raw data and information for this study are in Tables S1–S11. The raw Sanger sequence traces for edited sequences are in files.