LsNRL4 enhances photosynthesis and decreases leaf angles in lettuce.

Lettuce (Lactuca sativa) is one of the most important vegetables worldwide and an ideal plant for producing protein drugs. Both well-functioning chloroplasts that perform robust photosynthesis and small leaf angles that enable dense planting are essential for high yields. In this study, we used an F2 population derived from a cross between a lettuce cultivar with pale-green leaves and large leaf angles to a cultivar with dark-green leaves and small leaf angles to clone LsNRL4, which encodes an NPH3/RPT2-Like (NRL) protein. Unlike other NRL proteins in lettuce, the LsNRL4 lacks the BTB domain. Knockout mutants engineered using CRISPR/Cas9 and transgenic lines overexpressing LsNRL4 verified that LsNRL4 contributes to chloroplast development, photosynthesis and leaf angle. The LsNRL4 gene was not present in the parent with pale-green leaves and enlarged leaf angles. Loss of LsNRL4 results in the enlargement of chloroplasts, decreases in the amount of cellular space allocated to chloroplasts and defects in secondary cell wall biosynthesis in lamina joints. Overexpressing LsNRL4 significantly improved photosynthesis and decreased leaf angles. Indeed, the plant architecture of the overexpressing lines is ideal for dense planting. In summary, we identified a novel NRL gene that enhances photosynthesis and influences plant architecture. Our study provides new approaches for the breeding of lettuce that can be grown in dense planting in the open field or in modern plant factories. LsNRL4 homologues may also be used in other crops to increase photosynthesis and improve plant architecture.

Introduction

Due to the continued expansion of urban areas, loss of soil fertility and climate change, agricultural land is increasingly becoming a scarce resource (Foley et al.,  2011; Lambin et al.,  2013; Pandey and Seto, 2015). Thus, to use agricultural land efficiently, we have a critical need to improve photosynthesis in crop plants that can produce high yields when grown in high‐density conditions. Another approach to improve land‐use efficiency is the deployment of plant factories. To make plant factories profitable and sustainable, it is critical to grow crops such as lettuce (Lactuca sativa) with high photosynthesis efficiencies and small leaf angles (Touliatos et al.,  2016).

Chloroplast, converting solar energy into biologically usable forms of energy, is one of the most critical organelles to improve land‐use efficiency and crop yields (Russo et al.,  2019). Genes regulating chloroplast development and chlorophyll biosynthesis are part of a complex network that responds to light and hormones that can be exploited to improve photosynthesis and the production of biomass (Cackett et al.,  2022). For example, the GOLDEN2‐LIKE genes (GLKs) regulate chloroplast development to increase photosynthesis and consequently, enhance both the biomass and grain yields in crops (Li et al.,  2020b; Yeh et al.,  2022). Each plant cell has many chloroplasts, which lack gene silencing or position effects, providing a perfect system for uniform and high levels of expression of introduced foreign genes (Daniell et al.,  2021).

Leaf angle, as a major component of ideal plant architecture, is an important agronomic trait that affects the efficiency of photosynthesis and crop yields (Dong et al.,  2018). For example, AUXIN RESPONSE FACTORS and PUT ON WEIGHT1, regulate leaf angle in response to auxin and brassinosteroids, respectively, and the development of compact plant architecture to increase yield in rice (Huang et al.,  2021; Zhang et al.,  2021). ZmILI1 and Upright Plant Architecture2 regulate leaf angle in response to cytokinins and brassinosteroids, respectively, decrease leaf angle to facilitate high‐density planting and consequently, improve yields in maize (Ren et al.,  2020; Tian et al.,  2019). The secondary cell wall determines the mechanical strength of lamina joints and plays an important role in the regulation of leaf angle (Huang et al.,  2021). Although the mechanisms underlying leaf angle are intensively studied in monocots, factors that contribute to leaf angle in dicots are largely unknown (Jin et al.,  2021).

The NPH3/RPT2‐Like (NRL) protein family is prevalent in algae and land plants. The NRL family has three main domains, including an N‐terminal BTB (bric‐a‐brac, tram track and broad complex) domain, a central NPH3 domain and a C‐terminal coiled‐coil domain (Liscum et al.,  2014). The BTB domain interacts with Cullin3 in vivo to form an E3 ligase, with the BTB protein subunit and functions in substrate recognition (Figueroa et al.,  2005). The coiled‐coil domain is proposed to facilitate the localization of NRLs to the plasma membrane and to bind phototropin (Inoue et al.,  2008). The NRL family is involved in many important biological pathways, although the functions of the majority of NRL proteins remain unknown (Christie et al.,  2018). The NRL homologues studied so far are mainly involved in phototropism, petiole positioning, leaf expansion, chloroplast accumulation and the regulation of auxin transport (Cheng et al.,  2007; Christie et al.,  2018). For example, PHOTOTROPIC HYPOCOTYL 3 (NPH3) and ROOT PHOTOTROPISM 2 (RPT2), two members of the NRL family, are localized on the plasma membrane and interact with phototropin to modulate phototropism in Arabidopsis (Fankhauser and Christie, 2015; Liscum et al.,  2014). RPT2 and NRL PROTEIN FOR CHLOROPLAST MOVEMENT 1 (NCH1) mediate chloroplast accumulation in land plants (Suetsugu et al.,  2016). It remains unclear whether other NRLs contribute to chloroplast development.

The NRL4 and NRL12 proteins are two unique members of the NRL family in Arabidopsis because they lack the conserved BTB domain (Pedmale et al.,  2010). The lack of the BTB domain provides evidence that NRL4, NRL12 and their orthologs cannot interact with Cullin3 and therefore, do not contribute to the ubiquitin/proteasome pathway (Figueroa et al.,  2005). Hence, NRLs lacking the BTB domain and the NRLs containing the BTB domain may have different functions. Nonetheless, the impact of deficiencies in the BTB domain on the function of NRLs remains unknown. Therefore, to gain a better understanding of the NRL family, we have a critical need to study the functions of NRL proteins that lack the BTB domain.

Lettuce (Lactuca sativa) is one of the most popular vegetables worldwide. It has also been engineered to produce protein drugs, such as valuable pharmaceutical and edible vaccines (Daniell et al.,  2020; Daniell et al.,  2022; Power et al.,  2021; Singh et al.,  2021). For example, angiotensin‐converting enzyme 2 was expressed in lettuce chloroplasts as an oral drug to attenuate pulmonary arterial hypertension or to trap the SARS‐CoV‐2 and decrease infectivity (Daniell et al.,  2020; Daniell et al.,  2022). In addition, lettuce leaves were expressed with the SARS‐CoV‐2 receptor‐binding domain subunit of the surface‐exposed spike glycoprotein for the oral vaccine against SARS‐CoV‐2 infection (Power et al.,  2021). In this study, we used map‐based cloning to clone the gene LsNRL4, which plays important roles in the development and proliferation of chloroplasts. LsNRL4 enhances the accumulation of chlorophyll and photosynthesis and decreases leaf angles to increase the yield of lettuce and meet the ever‐increasing requirement of the production of lettuce leaf biomass. Our results provide new insights into the functions of the NRL gene family and contribute a new gene resource for the breeding of high‐yielding lettuce and other crops for modern agriculture.

Results

Phenotypic and physiological characterization of pale‐green cultivars of lettuce

The variations in chlorophyll levels in leaves were investigated in 240 lettuce accessions. Five accessions, all stem lettuce accessions, have pale‐green leaves (Figure 1a). Each of the five accessions has a wild‐type LsGLK. Moreover, their pale‐green leaves are distinct from those of the Lsglk mutants (Zhang et al.,  2022).

Figure 1

Phenotypic characterization of dark‐green lettuce cultivar S34 and pale‐green lettuce cultivar S23. (a‐b) Top views (a) and side views (b) of cultivars S34 and S23. Bar = 5 cm. (c‐f) Photosynthetic pigment contents (c), photosynthetic rates (Pn) (d), effective quantum yield of PSII (YII) (e) and photochemical quenching (qP) (f) in S34 and S23 (means ± SD; n = 3 to 9). Statistically significant differences were determined using a Student's t‐test (*P < 0.05, **P < 0.01). (g) The photosynthetic electron transport rate (ETR) for cultivars S34 and S23 in different intensities of photosynthetically active radiation (x‐axis; means ± SD; n = 3). [Colour figure can be viewed at wileyonlinelibrary.com]

To investigate the mechanism that contributes to the development of pale‐green phenotypes, we quantified the chlorophyll content of leaves from a pale‐green cultivar, S23, and a dark‐green cultivar, S34 (Figure 1a‐b). The chlorophyll and carotenoid contents of leaves from S23 were significantly lower than those from S34 (Figure 1c). Moreover, key photosynthetic parameters, including net photosynthetic rate (Pn), effective quantum yield of PSII (YII), photochemical quenching (qP) and photosynthetic electron transport rate (ETR) were significantly decreased in S23 relative to S34 (Figure 1d‐g). The decreases in the total chlorophyll content and other traits associated with photosynthesis are consistent with the pale‐green phenotype of S23.

Construction of a segregating population for the pale‐green trait

To dissect the genetics underlying the pale‐green leaves, we used an F2 segregating population derived from a cross between S23 and S34. Individuals in the F2 segregating population had either dark‐green leaves or pale‐green leaves, which is consistent with leaf color segregating as a qualitative trait in this population. Interestingly, amongst the 94 individuals that we initially screened, 73 individuals had dark‐green leaves and small leaf angles. The remaining 21 individuals had pale‐green leaves and large leaf angles (Figure S1a‐d). Thus, the pale‐green leaves and large leaf angles were segregated in a 3:1 ratio, as expected for a Mendelian trait (χ 2 = 0.355, P > 0.05). These data are consistent with a mutation in a single gene or with mutations in two tightly linked genes leading to the development of pale‐green leaves and large leaf angles.

We used bulked segregant analysis and RNA‐seq to genetically map the gene controlling leaf color and leaf angle. From the F2 population, 20 individuals with pale‐green leaves were mixed to construct a ‘pale‐green‐leaf’ pool, and 20 individuals with dark‐green leaves were mixed to construct a ‘dark‐green‐leaf’ pool. RNA was extracted from each pool and sequenced. The RNA‐seq data were analysed, and the Δ(SNP‐index) was calculated for each SNP. The average of Δ(SNP‐index) of the SNPs in a 3‐Mb sliding window with a 1‐Mb step was plotted along the nine chromosomes of lettuce, and a single significant peak was detected on Chromosome 4 (Figure 2a, Tables S1, S2, S3). A marker (AGH305) around this peak co‐segregated with leaf color and leaf angle in the F2 family (Table S4). Therefore, we successfully constructed a population in which chlorophyll deficiency and enlarged leaf angles segregating as one locus.

Figure 2

Map‐based cloning of Lsnrl4. (a) BSR‐seq showing a single locus on Chromosome 4. The y‐axis represents the average of Δ(SNP‐index) amongst the SNPs in a 3‐Mb sliding window with a 1‐Mb step. The x‐axis represents the nine chromosomes of lettuce. The red and green lines indicate 99% and 95% confidence intervals (i.e. P = 0.01 and P = 0.05), respectively. (b) Genetic mapping of the candidate gene. The numbers in the parentheses refer to the number of recombinants between the two neighbouring markers from 3168 progeny. (c) Nine open reading frames present in the candidate interval. The dashed lines and gene names show the microsynteny between cultivars S34 and S23. (d) Diagram for the protein structure of NRLs with or without the BTB domain. The grey boxes indicate the conserved regions I to IV. [Colour figure can be viewed at wileyonlinelibrary.com]

Map‐based cloning of the candidate gene controlling the development of pale‐green leaves and large leaf angles

To genetically fine map and clone the gene controlling pale‐green leaves and large leaf angles, we used two flanking markers, AGH318 (173.240 Mb) and AGH180 (183.607 Mb), to screen 3168 individuals from the F2 family and obtained 90 recombinants (Figure 2b). We mapped the candidate gene to a 731‐kb interval between markers AGH471 (177.262 Mb) and AGH424 (177.993 Mb) on Chromosome 4 by genotyping the recombinants with a series of markers from the candidate region (Figure 2b). Nine open reading frames (ORFs) were predicted in the candidate interval (Figure 2c; Table S5).

To identify additional polymorphisms between the two parents in the candidate interval, the DNA from both parents was re‐sequenced. The re‐sequencing data were mapped to the lettuce reference genome. The SNPs and the structural variations in the nine ORFs were analysed. No reads were mapped to LG4_398401 in the pale‐green parent (i.e. S23), in striking contrast to a large number of reads mapped to LG4_398401 in the dark‐green parent (i.e. S34; Figure S2a). The re‐sequencing data suggested that the deletion in parent S23 extended at least 467‐Kb, which was further supported by four markers showing presence/absence polymorphism between the two parents (Figure S2a). The RNA‐seq data from the two pools also showed that LG4_398401 was not expressed in the ‘pale‐green‐leaf’ pool and that it was highly expressed in the ‘dark‐green‐leaf’ pool (Figure S2b). We used LG4_398401‐specific primers to amply PCR products from the two parents. PCR products were amplified from S34 but not from S23 (Figure S2c). Therefore, LG4_398401 is deleted in parent S23 and exhibits presence/absence polymorphism between the two parents. We hypothesize that LG4_398401 may contribute to the development of dark‐green leaves with small leaf angles.

LG4_398401 encodes an NPH3/RPT2‐like (NRL) protein lacking a BTB domain

The LG4_398401 gene encodes an NPH3/RPT2‐like (NRL) protein. The lettuce reference genome encodes 34 NRL proteins. The Arabidopsis genome encodes 33 NRL proteins. A phylogenetic tree for the 67 NRL proteins encoded by the lettuce and Arabidopsis genomes was constructed (Figure S3). The phylogenetic tree indicates that LG4_398401 encodes a protein that is orthologous to NRL4 from Arabidopsis. For convenience, we hereafter refer to LG4_398401 as LsNRL4.

LsNRL4 and its ortholog NRL4 from Arabidopsis encode NRL proteins lacking the N‐terminal BTB domain (Figure 2d). To explore the evolutionary significance of the lack of the BTB domain in some NRLs, the amino acid sequences of the NRLs from 78 representative land plants were examined. Although the NRLs are present in all 78 land plants included in this study, the NRLs lacking the BTB domain can be found only in angiosperms (Figure S4a). Moreover, the NRLs lacking the BTB domain maintain obvious orthologous relationships, which is consistent with NRL4 proteins performing conserved functions in different plant species (Figure S4b).

We determined the proportion of cultivated and wild lettuce (Lactuca serriola) that contain null alleles of LsNRl4. Re‐sequencing and/or sequencing of PCR products showed that the 488 accessions of wild lettuce included in this study had intact alleles of LsNRL4 (Figure S5). Therefore, the loss of LsNRL4 most likely occurred after domestication. We also investigated the LsNRL4 genotypes in 198 lettuce cultivars (Zhang et al.,  2017). Amongst them, only five stem lettuce cultivars have the null allele of LsNRL4 (Figure S5). We conclude that the deletion of LsNRL4 is unique to stem lettuce and is likely associated with the breeding and development of stem lettuce.

Expression pattern and subcellular localization of LsNRL4

We then analysed the expression pattern and subcellular localization of LsNRL4. We quantified the relative expression levels of LsNRL4 using qRT‐PCR and found that LsNRL4 is expressed at the highest levels in leaves and lamina joints and at low levels in flowers, stems, roots and seeds (Figure 3a). In situ hybridization was carried out to investigate the expression patterns of LsNRL4 in the two parents. We found that LsNRL4 was expressed mainly in the mesophyll cells of leaves in S34 and that its expression was undetectable in the leaves of S23 due to the absence of the LsNRL4 gene (Figure 3b). Then, we transformed the fusion gene LsNRL4‐GFP driven by the CaMV 35S promoter into S23. Fluorescence emitted from the LsNRL4‐GFP fusion protein was detected only in the plasma membrane, but not in the nuclei or chloroplasts, which indicates that LsNRL4 is a cell membrane‐localized protein in lettuce (Figure 3c). The subcellular localization of LsNRL4 is consistent with that of NPH3, RPT2 and NCH1, which contain the BTB domain (Suetsugu et al.,  2016).

Figure 3

Expression pattern of LsNRL4 and subcellular localization of LsNRL4‐GFP. (a) Relative expression of LsNRL4 in different tissues (means ± SD; n = 4). (b) In situ hybridization of LsNRL4 transcripts in S34 and S23. Antisense probe was used in the left panel, and sense probe was used in the right panel. Bars = 50 μm. ‘ab’ and ‘ad’ represent the abaxial and adaxial domains of leaves, respectively. (c) Subcellular localization of LsNRL4‐GFP in lettuce leaf cells. The imaging was performed using a confocal laser scanning microscope. The plasma membrane was labelled with FM4‐64. [Colour figure can be viewed at wileyonlinelibrary.com]

Genetic demonstration of function

To test whether the deletion of LsNRL4 leads to the development of pale‐green leaves and large leaf angles in lettuce, we transformed S23 with a transgene that uses the CaMV 35S promoter to drive the expression of the LsNRL4 gene. We obtained three positive transformants (LsNRL4‐OX #1, #2, #3) that produced dark‐green leaves and developed smaller leaf angles relative to S23 (Figure 4a‐b). The total photosynthetic pigment content was significantly higher in the LsNRL4‐OX leaves than in S23 (Figure 4c). The expression level of the LsNRL4 gene was significantly higher in the LsNRL4‐OX plants than in S23 (Figure 4d). The expression of the N‐terminus Myc‐tagged LsNRL4 protein was detected in the LsNRL4‐OX plants (Figure 4e). Furthermore, the elevated levels of chlorophyll and small leaf angles co‐segregated with the transgene in a T1 population (Figure S6).

Figure 4

Functional analysis of the LsNRL4 gene. (a) Front view of S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants. Bar = 5 cm. (b‐d) Leaf angles (means ± SD; n = 7) (b), photosynthetic pigment contents (means ± SD; n = 3) (c) and expression of LsNRL4 (means ± SD; n = 4) (d) in S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants. Different letters indicate statistically significant differences based on a one‐way ANOVA analysis, followed by Tukey's multiple comparison test (P < 0.05). (e) Expression of an N‐terminus Myc‐tagged LsNRL4 protein in LsNRL4‐OX plants. (f) DNA sequences deleted from LsNRL4 in the Lsnrl4‐KO plants. [Colour figure can be viewed at wileyonlinelibrary.com]

We also used CRISPR/Cas9 technology to test whether LsNRL4 is required for the accumulation of chlorophyll in S34. We transformed S34 with a recombinant CRISPR/Cas9 vector that expresses a sgRNA that specifically binds the coding region of the LsNRL4 gene. Two knockout mutants with modified LsNRL4 sequences were obtained (Figure 4f). The homozygous knockout mutants Lsnrl4‐KO #1 and #2 produced pale‐green leaves that accumulated lower levels of chlorophyll than S34 and resembled the leaves of S23, and the leaf angles of the Lsnrl4‐KO plants were significantly larger than the leaf angles of S34 (Figure 4a‐c). We conclude that loss‐of‐function mutations in LsNRL4 lead to the development of pale‐green leaves and large leaf angles in lettuce.

enhances chloroplast coverage and photosynthesis

The chlorophyll content in LsNRL4 genotypes, such as S34 and the LsNRL4‐OX plants, was significantly higher relative to the Lsnrl4 genotypes, such as S23 and the Lsnrl4‐KO plants. We fixed leaf tissue with glutaraldehyde and used differential interference contrast (DIC) microscopy to quantify the numbers and sizes of chloroplast and the amount of space in the cell devoted to all chloroplasts (i.e. chloroplast coverage) in S34, S23, LsNRL4‐OX and Lsnrl4‐KO plants (Figure 5a). We found that chloroplasts were larger in the Lsnrl4 genotypes than in the LsNRL4 genotypes. However, the number of chloroplasts per cell plan area and chloroplast coverage in the Lsnrl4 genotypes were significantly reduced compared to the LsNRL4 genotypes (Figure 5 b‐d).

Figure 5

LsNRL4 promotes chloroplast coverage and photosynthesis. (a) Glutaraldehyde‐fixed mesophyll cells from S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants. Bars = 20 μm. (b‐d) Quantification of chloroplast number per cell plan area (means ± SD; n = 30) (b), chloroplast plan area (means ± SD; n = 150) (c) and chloroplast coverage (means ± SD; n = 10) (d) in S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants. (e) Transmission electron microscopy of mesophyll cells from S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants. Bars = 1 μm. (f‐i) Pn (f), YII (g), qP (h) and ETR (i) in S23, LsNRL4‐OX, S34 and Lsnrl4‐KO plants (means ± SD; n = 3 to 9). Different letters indicate statistically significant differences based on a one‐way ANOVA analysis, followed by Tukey's multiple comparison test (P < 0.05). [Colour figure can be viewed at wileyonlinelibrary.com]

We also investigated the ultrastructure of chloroplasts in S23, S34, LsNRL4‐OX and Lsnrl4‐KO plants using transmission electron microscopy. In S34 and LsNRL4‐OX plants, the chloroplasts have well‐developed thylakoid membrane systems with numerous grana. In contrast, the chloroplasts from S23 and the Lsnrl4‐KO plants are enlarged and have fewer thylakoid membranes and fewer grana thylakoids (Figure 5e). Thus, the loss of LsNRL4 gene function leads to abnormal chloroplast development.

To determine whether the reduced chlorophyll content and the impaired chloroplast development in the Lsnrl4 genotype affect photosynthesis, we measured key photosynthetic parameters in S23, S34, LsNRL4‐OX and Lsnrl4‐KO plants. The Pn, YII, qP and ETR in the Lsnrl4 genotype were significantly decreased relative to the LsNRL4 genotype (Figure 5f‐i). We conclude that LsNRL4 promotes chloroplast coverage and chlorophyll content and consequently, enhances photosynthetic capacity.

and independently promote the accumulation of chlorophyll in lettuce leaves

A recent study reported that the LsGLK gene is attenuated by the insertion of a CACTA transposon in some cultivars. This transposon insertion explains 29.2% of the variation in total photosynthetic pigments in a natural lettuce population (Zhang et al.,  2022). We investigated the photosynthetic pigments from 60 accessions that had a wild‐type LsGLK (Figure 6a). In these accessions, LsNRL4 explained 78.0% of the variation in chlorophyll a, 56.9% of the variation in chlorophyll a + b, 68.4% of the variation in carotenoids and 65.1% of the variation in total photosynthetic pigments. Therefore, the loss of LsNRL4 has a significant influence on chlorophyll content in lettuce, and similar to LsGLK, the LsNRL4 gene plays important roles in the variation in chlorophyll content in lettuce.

Figure 6

LsNRL4 and LsGLK independently promote the accumulation of chlorophyll and the proliferation of chloroplasts in lettuce leaves. (a) Photosynthetic pigment content of LsNRL4 genotypes and Lsnrl4 genotypes from 60 accessions in genetic backgrounds containing wild‐type LsGLK. Statistically significant differences were determined using a Student's t‐test (**P < 0.01). (b) Photosynthetic pigment contents from different LsNRL4 and LsGLK genotypes (means ± SD; n = 3). (c) Glutaraldehyde‐fixed mesophyll cells from different LsNRL4 and LsGLK genotypes. Bars = 20 μm. (d‐f) Quantification of Chloroplast number per cell plan area (means ± SD; n = 50) (d), chloroplast plan area (means ± SD; n = 100) (e) and chloroplast coverage (means ± SD; n = 10) (f) from different LsNRL4 and LsGLK genotypes. Different letters refer to statistically significant differences based on a one‐way ANOVA analysis followed by Tukey's multiple comparison test (P < 0.05). [Colour figure can be viewed at wileyonlinelibrary.com]

To determine the genetic relationships between LsNRL4 and LsGLK, the photosynthetic pigment content of four genotypes, LsNRL4/LsGLK, Lsnrl4/LsGLK, LsNRL4/Lsglk and Lsnrl4/Lsglk, from a natural population was measured in each genotype from three randomly chosen cultivars (Figure 6b). The results showed that compared to wild type, the photosynthetic pigment content of the LsNRL4/Lsglk and Lsnrl4/LsGLK mutants was reduced by 45.5% and 33.9%, respectively. And the photosynthetic pigment content of the Lsnrl4/Lsglk (i.e. the double mutant) was reduced by 51.9%, leading to more severe pale‐green phenotypes than the single mutant.

We also quantified chloroplast development using DIC microscopy in different LsNRL4 and LsGLK genotypes. Our results indicate that the loss of LsNRL4 decreased the number of chloroplasts per cell but increased the average size of individual chloroplasts (Figure 6c‐e). In contrast, LsGLK mainly regulates the size of individual chloroplasts but has no obvious effects on the number of chloroplasts per cell (Figure 6c‐e). Chloroplast coverage was reduced in Lsnrl4 and Lsglk single mutants and was further reduced in the Lsnrl4 Lsglk double mutant (Figure 6f). We conclude that LsNRL4 and LsGLK both promote the accumulation of chlorophyll and the proliferation of chloroplasts and that the loss of function of either gene leads to decreases in the chlorophyll content of lettuce leaves. Moreover, because the double mutants accumulate significantly less chlorophyll and allocate significantly less cellular space to chloroplasts relative to the single mutants, we conclude that LsNRL4 and LsGLK contribute to different mechanisms that promote the accumulation of chlorophyll and the proliferation of chloroplasts.

decreases leaf angle by regulating secondary cell wall biosynthesis

Leaf angle is determined by the mechanical properties of lamina joint tissues. The sclerenchyma tissue of lamina joints is largely responsible for providing the mechanical strength necessary to maintain leaf erectness (Zhou et al.,  2017). We examined the sclerenchyma tissues in the lamina joints of S23, S34 and LsNRL4‐OX. The cross sections revealed that the thickness of the secondary cell walls of the sclerenchyma cells of S23 was significantly reduced relative to S34 and the LsNRL4‐OX plants (Figure 7a‐b). These data indicate that the mechanical strength of the sclerenchyma tissue was attenuated in the lamina joints of S23. We conclude that the increased leaf angle in cultivar S23 results from defective sclerenchyma tissue.

Figure 7

Influence of LsNRL4 on secondary cell wall biosynthesis and leaf angles. (a) Cross sections showing the anatomical structures of sclerenchyma tissue in the lamina joints of S34, S23 and LsNRL4‐OX plants. Lignin was stained using sarranine. High magnification images (right panels) of sclerenchyma tissue indicated with red rectangles (left panels) are shown. Bars = 50 μm (left panel). Bars = 10 μm (right panel). (b) Cell wall thickness in sclerenchyma cells (means ± SD; n = 11). (c) Gene ontology enrichment analysis of genes co‐expressed with LsNRL4 and DEGs from BSR‐seq. (d) Relative expression of LsCOMT, LsCAD and LsF5H in S23, S34 and LsNRL4‐OX plants (means ± SD; n = 4). (e) Lignin content of the lamina joints from S23, S34 and LsNRL4‐OX plants (means ± SD; n = 3). Different letters indicate statistically significant differences based on a one‐way ANOVA analysis, followed by Tukey's multiple comparison test (P < 0.05). [Colour figure can be viewed at wileyonlinelibrary.com]

To determine how LsNRL4 influences the formation of sclerenchyma tissue, we analysed co‐expression networks in 240 accessions from natural populations of lettuce (Zhang et al.,  2017). We found that 665 genes were co‐expressed with the LsNRL4 gene (Table S6). Gene ontology (GO) analysis using the 665 co‐expressed genes and 1421 DEGs from BSR‐seq showed that the GO terms related to ‘cell wall organization,’ ‘lignin metabolic process,’ ‘cellulose metabolic process’ and ‘pectin metabolic process’ were significantly enriched (P < 0.00017 to 2.1 × 10−8; Figure 7c; Tables S7‐S9). We chose LsCOMT, LsCAD and LsF5H for further analysis, which encode key enzymes for the lignin biosynthesis (Wu et al.,  2019). The expression of these three genes was significantly reduced in S23 relative to S34 (Figure 6d). Consistently, the accumulation of lignin in the lamina joints of leaves from S23 was considerably reduced relative to S34 (Figure 6e). Overexpressing the LsNRL4 gene led to increases in the expression of LsCOMT, LsCAD and LsF5H, and consequently, led to the accumulation of more lignin in the lamina joints of the LsNRL4‐OX plants (Figure 7d‐e). Therefore, we conclude that the loss of LsNRL4 affects the expressions of a large number of genes associated with cell wall biosynthesis and leads to the development of defective sclerenchyma tissue and consequently, large leaf angles.

Discussion

The ( ) gene family

In this study, we cloned a gene that belongs to the NPH3/RPT2‐Like (NRL) gene family that controls chlorophyll content, chloroplast development and secondary cell wall development. Particular NRL proteins coordinate different aspects of the signalling activated by phototropins, which are blue‐light receptors (Christie et al.,  2018). Phototropins are present in green algae and flowering plants and their functions are highly conserved (Li et al.,  2015). The NRL proteins serve as effectors in phototropin signalling and are also conserved in algae and plants (Christie et al.,  2018). Compared to green algae, phototropin signalling in land plants is complicated in that it contributes to phototropism, chloroplast movement, leaf positioning and leaf flattening. Consistent with these observations, a large number of NRL proteins are present in land plants to facilitate the regulation of these biological processes by phototropins (Suetsugu et al.,  2016).

Most of the NRL proteins have BTB domains. The BTB domains of NPH3 and RPT2 can form a heterodimer in yeast. Moreover, the BTB domain of NPH3 can serve as a substrate adapter in a Cullin3‐based E3 ubiquitin ligase in Arabidopsis (Roberts et al.,  2011). However, NRL4 and NRL12 from Arabidopsis and LsNRL4 from lettuce lack the BTB domain (Pedmale et al.,  2010). The impact of these deficiencies in the BTB domain remains unclear because prior to this study, there were no reports on the biological functions of NRL proteins lacking the BTB domain. Prior to this study, the only information available on members of the NRL genes that do not encode proteins with BTB domains was that NRL4 and NRL12 are expressed in leaves and flowers in Arabidopsis, according to the TAIR database.

In this study, we showed that LsNRL4 is associated with the plasma membrane and promotes the development of chloroplast and secondary cell walls. The loss of LsNRL4 leads to the development of pale‐green leaves and large leaf angles. However, the biochemical functions of LsNRL4 remain to be studied. We hypothesize that LsNR4 may not interact with Cullin3 or target the phototropin light‐activated kinase for ubiquitination due to its lack of a BTB domain. Instead, the conserved C‐terminal consensus sequence of LsNRL4 may contribute to its biochemical function. Two serine residues, S744 and S746, located in the conserved C‐terminal domain of Arabidopsis NPH3 are required for NPH3 to promote phototropism and to influence the positioning of petioles (Sullivan et al.,  2021). Phosphorylation of serine residue S744 in Arabidopsis NPH3 leads to its interaction with 14–3‐3 proteins, which play important roles in phototropin signalling. (Sullivan et al.,  2021). Similar to NPH3, LsNRL4 has a conserved C‐terminal domain, including the third to last residue (i.e. S744 in NPH3), which, in turn, may enable the binding of 14–3‐3 proteins and the development of normal leaf angles in lettuce.

contributes to the development of chloroplasts and photosynthesis

The loss of LsNRL4 leads to abnormal chloroplast development, decreases in chlorophyll content, enlarged chloroplasts and less cellular space devoted to chloroplasts in leaves. A complex network that responds to light and hormones controls chloroplast development and the accumulation of photosynthetic pigments (Cackett et al.,  2022). A mechanism that drives increases in chloroplast volume determines how much space in the cell is devoted to chloroplasts. When chloroplasts reach a particular threshold size, chloroplast division is activated and breaks up the chloroplast compartment into many small chloroplasts. Thus, when the chloroplast division machinery is knocked out, mesophyll cells contain one enlarged chloroplast that occupies as much cellular space as all of the chloroplasts of wild‐type mesophyll cells combined (Cackett et al.,  2022; Larkin et al.,  2016). In Lsnrl4, chloroplasts are 70.3% larger than wild type and are fewer in number, which indicates that chloroplast division is attenuated. Also, in Lsnrl4, the amount of cellular space allocated to chloroplasts is reduced 44.0% relative to wild type. In contrast, when all of the REDUCED CHLOROPLAST COVERAGE (REC) from Arabidopsis are knocked out, the amount of cellular space allocated to chloroplasts is reduced 50.0% relative to wild type (Larkin et al.,  2016). The striking decrease in the accumulation of photosynthetic pigments, small increase in the size of the individual chloroplasts relative to chloroplast division mutants, smaller reduction in the amount of cellular space allocated to chloroplasts in Lsnrl4 relative to the rec mutants and the precedents for NRL proteins contributing to signalling are consistent with LsNRL4 influencing a network that controls chloroplast development. Consistent with numerous chloroplast defects in Lsnrl4, we found that the photosynthesis capacity decreased in Lsnrl4. A total of 26 genes that were differentially expressed in the LsNRL4 and Lsnrl4 genotypes were associated with chlorophyll metabolic and biosynthetic processes (Table S9). Our results showed that LsNRL4 plays important roles in the accumulation of chlorophyll and the development and proliferation of chloroplasts.

Previous studies have shown that in land plants, two members of the NRL gene family, RPT2 and NCH1, redundantly mediate the chloroplast accumulation response, which refers to the movement of chloroplasts towards low fluence rate light (Suetsugu et al.,  2016; Wang et al.,  2021). Although RPT2 and NCH1 contain four conserved regions including the BTB domain, the LsNRL4 protein in lettuce lacks the BTB domain. Therefore, LsNRL4 may use a different mechanism to influence chloroplasts. In this study, we discovered obvious decreases in the accumulation of chlorophyll and in the development and proliferation of chloroplasts in Lsnrl4, which lacks the entire LsNRL4 gene. It remains to be investigated how LsNRL4 regulates the development and proliferation of chloroplasts in lettuce.

contributes to the formation of secondary cell walls and leaf angles

The loss of LsNRL4 in lettuce attenuates secondary cell wall (SCW) biosynthesis in the sclerenchyma tissue of lamina joints. The lamina joint connects the leaf to the stem and is considered to be the most important tissue governing leaf angle (Zhou et al.,  2017). In rice and maize leaves, the leaf angle largely depends on cell division, cell expansion and cell wall composition in the joints of rice and maize leaves (Feng et al.,  2016). Notably, most of the genes that are known to influence the leaf angle are either transcriptional regulators or signalling components that respond to a variety of phytohormones (Cao et al.,  2022). Here, we showed that LsNRL4, an atypical NRL protein, promotes the development of small leaf angles by promoting SCW biosynthesis. The GO enrichment of genes that were co‐expressed with LsNRL4 and the genes that were differentially expressed in LsNRL4 relative to Lsnrl4 indicates that LsNRL4 promotes the development of small leaf angles by serving as a key regulator of SCW biosynthesis. This is the first report that members of the NRL family contribute to SCW biosynthesis. To gain a comprehensive understanding of the functions of the NRL family, the mechanisms used by LsNRL4 to regulate SCW biosynthesis, need to be investigated.

may regulate the defence system

We found many genes associated with biotic stress tolerance in the list of genes that were differentially expressed in LsNRL4 relative to Lsnrl4 (Table S9). These DEGs include genes that contribute to salicylic acid signalling, systemic resistance and defence responses. Consistent with these findings, recent studies reported that StNRL1 is a plant susceptibility gene that enables the late blight pathogen Phytophthora infestans to infect potatoes (Garcia‐Ruiz et al.,  2021; Yang et al.,  2016). The StNRL1 protein enhances pathogen infection by helping to degrade StSWAP70, an immune regulator (He et al.,  2018). Homodimerization of StNRL1 is required for StNRL1 to bind and help degrade StSWAP70. Mutations in the BTB domain of the StNRL1 protein attenuate its homodimerization activity and abolish its interaction with StSWAP70 and thus, prevent the degradation of StSWAP70 and reduce P. infestans infection (He et al.,  2018; Naqvi et al.,  2022). However, we suggest that because LsNRL4 lacks the BTB domain, LsNRL4 and StNRL1 may use different mechanisms to influence the defence response. It will be interesting to investigate how LsNRL4 contributes to the defence response and whether LsNRL4 helps lettuce tolerate biotic stress.

Furthermore, LsNRL4 induces increases in the thickness of the secondary cell walls, which is the first defensive barrier against plant pathogens (Miedes et al.,  2014). Indeed, secondary cell wall thickening limits the spread of pathogen infection (Miedes et al.,  2014; Yogendra et al.,  2015). For example, OsMYB30 induces cell wall thickening in sclerenchyma cells and inhibits the penetration of Magnaporthe oryzae in rice leaves at an early stage of infection by inducing increases in the expression of genes associated with lignin biosynthesis (Li et al.,  2020a). In this study, we showed that LsNRL4 up‐regulates the expression of genes associated with lignin biosynthesis. Lignin is considered to be one of the main components of the plant cell wall that can reduce the infiltration of pathogens (Liu et al.,  2018). The increased accumulation of lignin in the genotypes containing LsNRL4 may not only support the optimal development of leaf angles and chloroplasts but may also promote disease resistance in lettuce.

contributes to an ideal plant architecture for dense planting

We found that the leaf angles of the LsNRL4 genotype were smaller than the leaf angles of the Lsnrl4 mutant. Leaf angle is an important agronomic trait for plant architecture. Upright leaves can enhance photosynthetic capacity when plants are grown in high‐density conditions and thus, increase grain yields for cereal crops (Cao et al.,  2022). For leafy vegetables, small leaf angles contribute to a compact architecture that reduces the amount of space required to grow each plant, enables dense planting and promotes vegetable production. LsNRL4 could be exploited to produce new lettuce cultivars with upright leaf architectures. Orthologs of NRL4 may reduce leaf angle in other crops and consequently, help to increase yields when crops are densely planted.

Experimental procedures

Plant materials and growth conditions

Lactuca materials used in this study were described previously (Zhang et al.,  2017). Stem lettuce cultivars S23 and S34 were crossed to generate the F1 hybrids. The F1 hybrids were self‐crossed to generate an F2 population. All plants were planted in a field at Huazhong Agricultural University, Wuhan, China.

Bulked segregant analysis and RNA‐seq

A bulked segregant analysis in combination with RNA‐seq was used for the map‐based cloning. As described in the Results, total RNA was extracted from the two pools of tissue and sequenced using the Illumina Hiseq2500 platform (Personalbio, China). Approximately 4.05 and 4.69 Gb clean data were obtained for pale‐green and dark‐green pools, respectively, and the sequencing data were mapped to the lettuce reference genome using the Bowtie software (Langmead et al.,  2009; Reyes‐Chin‐Wo et al.,  2017). SNP calling was performed using SAMtools (Li et al.,  2009). The ∆(SNP‐index) was calculated by subtracting the SNP‐index value (allele frequency) of the pale‐green pool from the dark‐green pool. The average of Δ(SNP‐index) of the SNPs in a 3‐Mb sliding window with a 1‐Mb step was plotted along the nine chromosomes of lettuce. The primers used in the genetic mapping are shown in Table S10.

Construction of the overexpression and knockout lines

To construct lines that overexpress LsNRL4, the coding sequence of LsNRL4 was amplified from cDNA prepared from S34 and was inserted into pH7LI9 (with an N‐terminus Myc‐tag). Agrobacterium tumefaciens GV3101 was transformed with this vector using a thermal stimulation method and then, used to transform S23. For knockout assays, a CRISPR/Cas9 vector that expresses sgRNA specific to the coding region of LsNRL4 was constructed. GV3101 was transformed with this construct and then used to transform stem lettuce cultivar S34. All primers used for vector construction are shown in Table S10. Lettuce was transformed as described previously (Curtis et al.,  1994).

Quantitative RT‐PCR analysis

Total RNA was extracted from leaves using RNAiso plus (Takara, Japan). cDNA was synthesized using TransScript One‐Step SuperMix (TransScript, China). The qRT‐PCR analysis followed the guidelines and protocols described previously (Udvardi et al.,  2008). Ubiquitin (LG4_16296), a housekeeping gene, was used as an internal standard. The relative expression levels were quantified using the 2−ΔΔCq method. Statistically significant differences were calculated using a Student's t‐test. The primers used in the qRT‐PCR analysis are shown in Table S10.

Evaluation of photosynthetic pigment content

Photosynthetic pigments (i.e. chlorophylls and carotenoids) were extracted from 2‐month‐old lettuce leaves in triplicate. Approximately 100 mg of leaves were ground in liquid nitrogen, incubated in 10 mL of 95% ethanol in the dark for 24 h at room temperature and centrifuged for 10 min at 13 523 g. The absorbance of the supernatant was measured at 470 nm, 649 nm and 665 nm using spectrophotometry. The concentration of photosynthetic pigments was calculated as described previously (Lichtenthaler and Wellburn, 1983).

Total lignin assay

The tissues of lamina joints from 2‐month‐old S23, S34 and LsNRL4‐OX plants were used for lignin measurements. Total lignin, including acid‐soluble and acid‐insoluble lignin, was measured using the Laboratory Analytical Procedure of the National Renewable Energy Laboratory (Wu et al.,  2013). All analyses were performed from independent experiments performed in triplicate.

Measurement of photosynthetic parameters

For the chlorophyll fluorescence measurements, firstly, 2‐month‐old lettuce leaves from S23, S34, LsNRL4‐OX and Lsnrl4‐KO plants were kept in dark for 30 min. A chlorophyll fluorescence system (Imaging‐PAM, Walz, Germany) was used to measure the chlorophyll fluorescence parameters, namely YII, qP and ETR, in leaves. The Pn was measured between 9:00 and 11:00 on a sunny day. Nine individuals were randomly selected from each genotype and the third, fully unfolded lettuce leaves from the top of the plants were used for measurements. The data were measured and recorded using a portable photosynthesis system (LI‐6400XT, LI‐COR, USA).

Analysis of chloroplasts by microscopy

Leaf sections were fixed with glutaraldehyde. Fixed mesophyll cells were released from the leaf sections and were visualized using differential interference contrast microscopy as previously described (Pyke and Leech, 1991). The chloroplast number per cell plan area, chloroplast plan area and total chloroplast plan area/cell plan area (i.e. chloroplast coverage) were quantified as described previously (Pyke and Leech, 1991).

Tissue section and in situ hybridization

For tissue sections, the tissues from lamina joints were excised and immersed in 70% FAA buffer (70% ethanol, 5% acetic acid and 3.7% formaldehyde). After the application of a vacuum for 15 min, samples were dehydrated using a series of ethanol solutions (70%, 85%, 95% and 100%). The ethanol was gradually replaced with xylene and further immersed in an increasing concentration of paraffin and finally embedded in absolute paraffin. Paraffin sections were prepared using a rotary microtome (Leica). Images were taken using a bright‐field microscope (ZEISS).

The in situ hybridization and immunological signal detection were performed as described previously (Samach et al.,  1997). The probes were amplified using gene‐specific primers. The PCR fragment was inserted into the SpeI/ScaII linearized pGEM‐T and the probes were transcribed in vitro from either the T7 or SP6 promoter for sense and antisense probe synthesis using the Digoxigenin RNA labelling kit (Roche). The primers used in the in situ hybridization are shown in Table S10.

Accession numbers

The datasets and the gene sequence are available in the National Center for Biotechnology Information under the accession number PRJNA794278 and OM156463.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

G.A. performed the molecular experiments; Y.Q. assisted with the subcellular localization and vector construction; W.Z. provided assistance with bioinformatics; H.G. helped with lignin analysis; J.Q. helped with the chloroplast analysis; H.K. and J.C. designed the experiment; G.A. wrote the manuscript with the help from H.K., J.C. and R.M.L.

Figure S1 Phenotypes in the F2 segregating population.

Figure S2 LsNRL4 gene was lost in the pale‐green parent S23.

Figure S3 Phylogenetic analysis of the NRL proteins from lettuce and Arabidopsis.

Figure S4 Evolution of NRLs in land plants.

Figure S5 Frequency of LsNRL4 and Lsnrl4 genotypes in Lactuca populations.

Figure S6 Variations in leaf color and leaf angle in a T1 population co‐segregating with the vector insertion.

Click here for additional data file.

Table S1 The source data of the average of Δ(SNP‐index) in BSR‐seq.

Table S2 The 181 SNPs in the region with average of Δ(SNP‐index) > 0.6.

Table S3 The 268 genes in the region with average of Δ(SNP‐index) > 0.6.

Table S4 The phenotype and genotype of 94 individuals used for primary mapping.

Table S5 Nine open reading frames in the candidate interval.

Table S6 Genes co‐expressed with LsNRL4.

Table S7 GO terms enriched in the co‐expressed genes.

Table S8 The 1421 DEGs in LsNRL4 relative to Lsnrl4.

Table S9 GO terms enriched in the DEGs.

Table S10 Primers used in this research.

Click here for additional data file.