A total of 19 genes that encode HcSWEET proteins were identified after homology alignment and conservative domain verification. These genes were named HcSWEET1 to HcSWEET16 according to their homology with dicotyledons such as Arabidopsis AtSWEETs and monocotyledons such as rice OsSWEETs. HcSWEET gene characteristics of open reading frame (ORF) size, number of amino acid (AA/aa), protein molecular weight (MW), theoretical isoelectric point (pI), instability index (II), aliphatic index (AI), grand average of hydropathicity (GRAVY), number of predicted TMHs and conserved domain position are analyzed in Table 1. As a consequence, the deduced number of amino acids in the night lily SWEET proteins was from 229 to 299, with the MW varying widely from 25.617 kDa to 32.938 kDa, and the PI index is between 5.49 to 9.66. All proteins ranged from 27.08 to 42.19 on the instability index, from 105.08 to 133.65 on the aliphatic index, and from 0.419 to 0.934 on GRAVY. The majority of predicted TMHs number was seven other than 6 TMHs of HcSWEET5 and HcSWEET13a/d. Coupled with subcellular localization, the secondary structure including α-helix, extended strand, β-turn and random coil are summarized in Supplementary Table S1. Subcellular localization predictions suggested that the majority of HcSWEET genes located in the plasma membrane, are also found in the chloroplast, vacuole or cytoplasm (Supplementary Table S1). Considering the transmembrane characteristic, we predicted 3D structural models of 19 HcSWEET proteins using a homology modeling approach, shown in Supplementary Figure S1. The results demonstrated that various properties existed in different HcSWEET members.
Phylogenetic analysis of HcSWEET proteins
To better understand the evolutionary interrelatedness among these HcSWEET members, we constructed a phylogenetic tree containing SWEETs from six species (Hemerocallis citrina, Hemerocallis fulva, Oryza sativa, Zea mays, Arabidopsis thaliana, Vitis vinifera) based on amino acid sequences using the TBtools software (Fig. 1). SWEET protein sequences of six species were listed in Supplementary Information 1. The results revealed that 19 HcSWEET members were classified into four major clades, namely Clade I, II, III, and IV respectively. Multiple HcSWEETs are tightly clustered with other SWEETs. Among these clades, the member number varied significantly. Clade III contained the most HcSWEET members (9), followed by Clade II (5), I (4), and IV (1), indicating a similar distribution pattern and gene function to these of other species.
Conserved motif composition, domain, and gene structure analysis of HcSWEET genes
According to the results from phylogenetic analysis, we divided the identified 19 HcSWEET proteins into 4 clades (Fig. 2A). Ten different conserved motifs were identified by analyzing amino acid sequences of HcSWEET proteins using the MEME online tool (Fig. 2B). The number and distribution of these motifs showed variations in different proteins. Each protein had 5–7 motifs, in which Motif 1, 3, and 4 were found in all proteins, moreover, different clades had distinct motifs. Motif analysis suggested that the HcSWEET gene family showed both conservation and differences, which might result from the diversity of protein functions within clades. Conserved domain analysis revealed that HcSWEET proteins possessed domains consisting of MtN3_slv and MtN3_slv superfamily (Fig. 2C), implying high conservation during evolution. In addition, to understand the structural compositions of HcSWEET genes, a structural map including the untranslated region (UTR) and coding sequence (CDS) was constructed based on the genome sequence (Fig. 2D). The number of CDSs ranged from 5 to 6. Coding sequences of 19 HcSWEETs were displayed in Supplementary Information 2. Gene structure analysis suggested that members within the same evolutionary clade exhibited similar gene structures.
Chromosomal localization, gene duplication, and collinearity analysis of HcSWEETs in Hemerocallis citrina, Arabidopsis thaliana and Oryza sativa
To investigate the chromosomal localization of HcSWEET genes, we analyzed the distribution of identified 19 genes on night lily chromosomes (Fig. 3). The distributions of these genes were determined via the night lily genome reported in 2021 [36]. The results showed that 19 genes were evenly dispersed across 9 of 11 chromosomes (LGs) with gene counts ranging from 1 to 4 in total. Notably, the largest number of genes were located in LG2, in contrast, LG6 and LG8 hosted only one gene and no gene located in LG7 and LG9.
To set forth whether gene duplication events forced the evolution of the HcSWEET gene family, an intra-species collinear analysis was performed (Fig. 4A). Among all gene pairs of the night lily genome, four pairs of duplicated segments were identified in the HcSWEETs, implying that segmental duplication events were likely to contribute to genetic diversity. Subsequently, we constructed inter-species syntenic maps of Hemerocallis citrina with other two representative species including monocotyledon Oryza sativa and dicotyledon Arabidopsis thaliana to identify orthologous genes through utilizing TBtools (Fig. 4B). It was observed that seven genes exhibited collinearity relationships with two genes in Arabidopsis thaliana and ten genes in O. sativa. These results suggest that HcSWEET genes likely share more similarities in function with OsSWEET genes in Oryza sativa.
Identification and distributions of cis‑regulatory elements in HcSWEETs promoters
We specifically extracted 2000 base pairs upstream of the HcSWEET genes transcription start codon to explore their potential biological functions through the database named PlantCARE online. As shown in Fig. 5, a total of three types of main responsive elements were identified, hormone responses, plant growth and development, and responses to abiotic stress respectively. It was found that 19 genes contained 6 to 26 light-responsive elements, indicating that the majority of HcSWEET genes may be regulated by light signaling. Moreover, all 19 genes contained hormone-related and development-related cis-acting elements. Hormone-related elements were referred to as MeJA, salicylic acid, abscisic acid, auxin, and gibberellin response elements. In contrast, development-related elements contained zein metabolism regulation, anaerobic induction, cell cycle regulation, seed-specific regulation, circadian control, flavonoid biosynthesis, endosperm, and meristem expression, anoxic specific inducibility and phytochrome down-regulation response elements. Additionally, HcSWEET genes other than HcSWEET4c, HcSWEET13d, and HcSWEET14b had cis-acting elements linked to abiotic stress, involving defense and stress, drought, and low-temperature responsive elements. HcSWEET13d also contained the wound-responsive element. Overall, various types of cis-elements were contained in the promoter regions of 19 HcSWEETs, indicating their possible involvement in diverse biological processes, environmental stresses, and regulatory pathways.
Expression patterns of HcWEET genes across different tissues and phenotype observations under drought and salt stress
To further analyze the possible roles of HcSWEET genes in abiotic stress, 19 genes were analyzed for relative expression profiles in different tissues and developmental stages (tender root, mature root, bud, tender leaf, mature leaf, tender scape, and mature scape) (Fig. 6A). As listed in Supplementary Table S3, RNA-seq results of tissue-specific expression revealed that most HcSWEETs in Clade II (HcSWEET4a, HcSWEET4b, HcSWEET5) and III (HcSWEET13c, HcSWEET13d, HcSWEET14a, HcSWEET14b, HcSWEET14c, HcSWEET15) exhibited the higher transcript levels in roots than other organs. Among them, HcSWEET4a and HcSWEET13c were simultaneously in a related highly expression in the scape. Except these genes, HcSWEET2 in Clade I was expressed especially and abundantly in roots throughout the developmental process.
The transporters of the SWEET gene family in Arabidopsis thaliana have been reported to increase root proliferation by enhancing sucrose transport from shoot to root during abiotic stress [37]. Therefore, we carried out stress experiments to determine whether the case is suitable for HcSWEET genes in night lily. We treated 30 d seedling roots with drought stress for 0 h, 24 h, 48 h, 72 h, and 108 h, and identically salt stress for 0 h, 24 h, 48 h, 72 h, and 96 h. As was displayed in Fig. 6B and 6C, night lily plants primarily showed a slight symptom of loss of water when treated for 24 h, after treatment for 48 h, leaves presented severe yellow and rotten roots.
Expression pattern analysis of HcSWWTs under drought and salt stress
Based on the phenotypes observed in night lily plants treated by drought and salinity, roots were collected, followed by RNA-Seq analysis (Supplementary Table S4/5). Depicted in Fig. 7A, expressions of genes including HcSWEET1b, HcSWEET4a, HcSWEET4b, HcSWEET5, HcSWEET13b, HcSWEET13c, and HcSWEET13d were induced compared with the control group (0 h) under drought stress. On the contrary, HcSWEET2, HcSWEET13a and HcSWEET14a expressed in a down-regulation trend. Relative expression levels of two genes (HcSWEET1a, HcSWEET14b) first increased and then decreased, while HcSWEET14c exhibited a contrary expression pattern.
We further investigated expression levels of HcSWEETs under salt treatments (Fig. 7B). Five HcSWEET genes (HcSWEET4a, HcSWEET5, HcSWEET11, HcSWEET13c, HcSWEET13d) showed significantly elevated expression levels. By contrast, the expression of HcSWEET2, HcSWEET6, HcSWEET13b, and HcSWEET14b increased during different salt stresses. After salt treatments, there were 4 genes (HcSWEET1b, HcSWEET14a, HcSWEET14c, and HcSWEET16) initially up-regulating, subsequently decreasing again in expression levels. HcSWEET4c, HcSWEET13a, HcSWEET15 went down and then up. The various expression types showed that functional differences were likely to exist in HcSWEET genes in response to abiotic stresses.
To unveil the potential function of HcSWEET genes in responding to different stresses including drought and salinity, nine HcSWEET genes showing significant responses were selected and RNA-Seq results were further confirmed using RT-qPCR experiments (Fig. 8). HcSWEET4a, HcSWEET4b, and HcSWEET5 showed an up-regulation in expressions under two stresses, notably, the expressions of HcSWEET4a and HcSWEET5 significant increased after salt stresses, suggesting the potential possibilities of HcSWEET4a and HcSWEET5 involved in salt stress. HcSWEET2 exhibited a pattern of decreased expression compared with 0 h. Given whole wavelike patterns, the expressions of HcSWEET1a, HcSWEET4a, HcSWEET4b, HcSWEET5, HcSWEET13c had a sharp increase at the time point of 24–72 h, but decreased in 96–108 h throughout stresses. Overall, our analysis revealed that expression trends of 19 genes were consistent with those of the RNA-Seq and the majority of HcSWEETs presented a dynamic response pattern.