To identify the CHS family members in Z. marina, CHS profiles (PF00195 and PF02797) downloaded from the Pfam database¹ were used as queries to perform BLASTP searches against Z. marina database (Olsen et al., 2016) on Phytozome website². The gene models of ZosmaCHSs were constructed using the Gene Structure Display Server (GSDS,³) (Hu et al., 2015). The DNAMAN 9 software (LynnonBiosoft, United States) was used to predict the isoelectric points and molecular weights of CHS proteins. Plant Duplicate Gene Database (PlantDGD,⁴) was used for gene duplication analysis.

The Pfam database and Simple Modular Architecture Research Tool (SMART,⁵) were used to predict the conserved domains of the ZosmaCHS proteins. The online MEME program⁶ was used to identify the conserved motifs of the CHS sequences. The parameters set as following: any number of repetitions, maximum of 10 motifs, and optimum motif width of 6–50 amino acid residues.

The neighbor-joining method and maximum likelihood method with a 1,000 times bootstrap value were used to construct phylogenetic tree by multiple sequence alignment of the protein sequences of 11 ZosmaCHSs in MEGA X software. The accession numbers and resources of all CHSs used for the tree were also provided in Supplementary Table 6. All CHSs can be found in Uniprot database⁷. Default settings were used for the other parameters.

The PlantCare website⁸ was used to analyze the putative promoter sequences of ZosmaCHSs. All the predicted cis-elements except the TATA-box and CAAT-box were visualized using the TBtools software following the author’s instructions (Chen et al., 2020).

Zostera marina with intact rhizome-systems was collected during the growing season from 3 m deep sub-tidal seagrass beds in Rongcheng (37° 91′N, 120° 73′E), Shandong Province, China. Samples were then cultured in seawater-aquarium that was continuously aerated and renewed daily. The plants were pre-cultivated with a 10/14 h (light/dark) photoperiod under minimum saturation light intensity (100 μmol photons m^–2s^–1) at 15°C for 3 days before experimentation. The roots, leaves, flowers, stems, and rhizomes were collected during the plant’s flowering stage for tissue-specific expression analysis using quantitative real time-PCR (qRT-PCR). The leaves were further exposed to blue, red and white light at the intensity of 300 μmol photons m^–2s^–1, after dark-adaptation overnight. Sampling was done at 0, 1, and 3 h for the qRT-PCR. Blue light was the dominant spectral components in the ecological niche of Z. marina, while the red light was at low intensity (Olsen et al., 2016).

Total RNA was extracted from the five tissues as well as the leaves exposed to different light treatments using RNAiso Plus (Takara, Japan). RNA quality was determined through agarose gel electrophoresis and NanoQuant (TECAN, Switzerland). Subsequently, 1 μg of the total RNA was reverse transcribed to cDNA using HiScript^® II 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China) following the manufacturer’s instructions. The qRT-PCR assays were performed on a Bio-Rad CFX96 Real-Time PCR System using AceQ Universal SYBR qPCR Master Mix (Vazyme). Primer sequences used in qRT-PCR are listed in Supplementary Table 1. The housekeeping gene gapdh of Z. marina was used as the internal control. The qRT-PCR program was set as follows: 95°C for 10 s, followed by 40 cycles of 56°C for 10 s and 72°C for 30 s. Each reaction was performed in three biological replicates. The relative expression level of each gene was calculated using the 2^{–Δ Δ} CT method. Heatmaps were constructed based on the transformed log2 (2^–ΔΔCT + 1) values using the TBtools software.

The open reading frames of the five CHS genes (ZosmaCHS01, ZosmaCHS02, ZosmaCHS07, ZosmaCHS08, and ZosmaCHS11) were amplified using KOD -Plus- Neo DNA Polymerase (TOYOBO, Japan). Primer sequences with restriction enzymes (EcoRI and BamHI) cutting sites are listed in Supplementary Table 2. The amplified products were purified using a FastPure Gel DNA Extraction Mini Kit (Vazyme, Nanjing, China), cloned into pEASY^®-Blunt Simple Cloning Vector (TransGen Biotech, Beijing, China), and then transformed into T1 competent cell for sequencing to select sequences without the base mutations and deletions. Plastids with correct CHS sequences, which was extracted by TIANprep Mini Plasmid Kit (TIANGEN, Beijing, China), and pET-28a (+) vector were cut using EcoRI and BamHI restriction enzymes (Invitrogen, Carlsbad, CA, United States), purified using a FastPure Gel DNA Extraction Mini Kit, and then linked by T4 DNA ligase (Invitrogen, Carlsbad, CA, United States). The pET-28a (+)-CHS was transformed into Transetta (DE3) Chemically Competent Cell. Until their OD600 to 0.4–0.5, the transformant cultures were added with isopropyl β-D-thiogalactopyranoside (IPTG) to a final concentration of 1 mmol L^–1, and incubated at 16°C overnight. The heterologous expression of the CHS was then examined using SDS-PAGE. The following buffers were used for purification of the ZosmaCHSs: lysis buffer (50 mM Na₂HPO₄, 0.3 M NaCl, pH = 8.0), washing buffer (50 mM NaH2PO4, 0.3 M NaCl, 10 mM imidazole pH = 8.0), and elution buffer (50 mM NaH2PO4, 0.3 M NaCl, 250 mM imidazole pH = 8.0). The bacterial cells were centrifugation, resuspended in lysis buffer, and broken by ultrasonication for 30 min on ice. The proteins were loaded on High-Affinity Ni-NTA Resin (GenScript Biotech, Nanjing, China), washed by washing buffer, eluted by elution buffer and further examined using SDS-PAGE.

The enzyme activities of the five purified CHS proteins were examined using a CHS enzyme activity kit (GENMED Scientifics Inc., United States) following the manufacturer’s instructions. Their absorbance was measured at 412 nm using a multifunctional enzyme-labeled instrument (Tecan, Switzerland). Each reaction was performed in three technical replicates. The enzyme activity of ZosmaCHS represented the test enzyme activity subtracted the activity of the empty buffer.

Data were analyzed using one-way ANOVA and Tukey’s tests on SPSS 22.0. P < 0.05 indicated significant differences between groups.

Eleven full-length CHS homologs (ZosmaCHS01-11) corresponding to the Pfam CHS family were identified and described (Table 1). Among them, ZosmaCHS09 was the smallest protein with 23 kDa, while the remaining were about 43 kDa. The isoelectric point (pI) of the proteins ranged between 6.25 and 7.58. Most CHS genes had multiple exons, of which ZosmaCHS02, 04, 07, 08, and 09 had two exons, ZosmaCHS03 had three exons, ZosmaCHS10 had four exons, while ZosmaCHS01, 05, 06, and 11 had one exon (Figure 1A).

Ten of the 11 CHS homologs harbored integral chalcone/stilbene synthases N-terminal (Chal_sti_synt_N) and chalcone/stilbene synthases C-terminal (Chal_sti_synt_C) domain represented by motif 1, 2, 7, and 9, and motif 3, 4 and 6, respectively (Figure 1B). ZosmaCHS09, had only part of the two domains with the absence of motif 1, 2, and 7. Moreover, ZosmaCHSs within clades I and III shared similar motif composition. The duplicated pairs ZosmaCHS05/06 and ZosmaCHS 07/08 showed highly similar motif distribution, indicating that their protein architecture is conserved in duplicated homologs.

Five widely accepted duplication modes (whole genome, tandem, proximal, transposed, and dispersed duplication) were systematically scanned among the ZosmaCHS genes to explore the evolution clues of the CHS gene family. The gene pairs ZosmaCHS05/06, ZosmaCHS07/08, and ZosmaCHS01/11 originated from tandem, proximal, and transposed duplication, respectively. Both ZosmaCHS09 and 10 formed dispersed duplication pairs with ZosmaCHS03. To further elucidate the evolutionary trend of duplicated ZosmaCHS genes, the ratios of the number of non-synonymous substitutions per non-synonymous site (Ka) to the number of synonymous substitutions per synonymous site (Ks) (Ka/Ks) were calculated between the duplication pairs. Most duplicated pairs had a Ka/Ks ratio of less than one, indicating that the ZosmaCHS gene family could have undergone strong purifying selective pressure during evolution (Supplementary Table 3).

All the ZosmaCHS proteins contained conserved catalytic triad of Cys-His-Asn and the CHS superfamily-specific Pro386 residues (Supplementary Figure 1). All the CHS members retained the residue Phe 220 responsible for CoA binding, except ZosmaCHS09. Phe 271, which is also essential for CoA binding, varied amongst the ZosmaCHS proteins. It was not retained in ZosmaCHS10, converted to alanine in ZosmaCHS04-06 and leucine in ZosmaCHS03 and 09. However, it was conserved in the five ZosmaCHSs (ZosmaCHS01, ZosmaCHS02, ZosmaCHS07, ZosmaCHS08, and ZosmaCHS11).

The phylogenetic relationships between Z. marina and other CHS genes using the maximum likelihood (Figure 2) and neighbor-joining method (Supplementary Figure 2) exhibited similar phylogenetic topology. It revealed that 11 ZosmaCHSs clustered into three clades. Notably, ZosmaCHS01, 02, 07, 08, and 11 clustered with the validated chalcone synthases in land plants (clade I), indicating that these CHS genes could have CHS activity to generate naringenin chalcone. ZosmaCHS03, 09, and 10 clustered with Hydrangea macrophylla coumaroyl triacetic acid lactone synthase (CTAS) and Rheum palmatum benzalacetone synthase (BAS) (clade II). ZosmaCHS04-06 formed clade III, with no CHS homologs in other plants.

The putative cis-elements located 2.0 kb upstream of the start codon (ATG) were analyzed to investigate the potential regulatory mechanisms of ZosmaCHSs during different stress responses. Four types of cis-elements were enriched in the promoter region of the ZosmaCHS genes (Figure 3). Numerous light-related elements, including Box 4, G-box, and the GATA-, GT1-, and TCT- motifs, were detected. Box 4 was the most abundant light-responsive element. Methyl jasmonate (MeJA) responsive elements, including CGTCA- and TGACG-motif, were the common cis-acting elements indicating their involvement in regulating the MeJA-pathway. Besides the ubiquitous elements, the MYB, MYC and MBS signature sequences were in the promoter region of ZosmaCHSs, contributing to the plant tolerance to various stresses.

Cluster analysis of the upstream sequences of ZosmaCHSs (Figure 3 and Supplementary Table 5) further revealed that the promoter regions of the 11 ZosmaCHS genes clustered into three groups. TCT-motif, GA-motif, and MRE related to light response and TCA-element related to salicylic acid responsiveness were in group I (ZosmaCHS05-07 and 09) and II (ZosmaCHS02, 04 and 10), indicating that the genes could be regulated by light stress and the salicylic acid pathway. Nevertheless, circadian elements were in group III (ZosmaCHS01, 03, 08, and 11), suggesting the photoperiod had effects on regulating these CHS genes. The functional diversification of ZosmaCHSs was attribute to differences in the promoter region.

Blue light was the dominant spectral components in the ecological niche of Z. marina, while the red light was at low intensity (Olsen et al., 2016). The expression patterns of all the 11 ZosmaCHS genes were investigated to confirm whether the expression of ZosmaCHS genes was influenced by different light qualities. The ZosmaCHSs were highly expressed in red light than in other lights. Moreover, ZosmaCHS02 was predominantly induced at a relatively early stage (after 1 h treatment), while ZosmaCHS01, 07, 08, and 11, were significantly and continuously up-regulated and peaked at 3 h after treatment (Figure 4A). These ZosmaCHS genes were predicted to produce naringenin chalcone. However, most ZosmaCHSs clustering with non-CHS genes were slightly up-regulated in red light or hardly induced. The duplication pairs ZosmaCHS07/08, ZosmaCHS03/09, and ZosmaCHS03/10 exhibited different expression patterns, implying the subfunctionalization of the CHS family.

The 11 CHS homologs could be detected in all the five tissues, though at varying expression levels (Figure 4B). ZosmaCHS04-06 was highly expressed in the leaves. In contrast, most ZosmaCHSs, including ZosmaCHS01-03, 07-11, showed higher expression levels in rhizomes, stems, and roots than in leaves and flowers, possibly because the CHS was involved in light protection and other biological functions. The diverse tissue-specific expression patterns of ZosmaCHSs suggested the subfunctionalization of the CHS family.

As a crucial step to elucidate biological function of ZosmaCHSs, the proteins which clustered with the validated chalcone synthases were recombined into prokaryotic expression systems to further confirm the ZosmaCHSs producing naringenin chalcone. Five ZosmaCHSs (ZosmaCHS01, 02, 07, 08, and 11) from clade I were heterologously expressed and purified to perform in vitro enzyme activity assays. SDS-PAGE analysis showed that the corresponding protein bands (around 43 kDa) were induced in the IPTG treated sample (Figure 5A and Supplementary Figure 3B), implying that the pET-28a-ZosmaCHSs recombinant plasmid could be successfully induced and expressed in Escherichia coli (E. coli). The purification experiments indicated that His-tag affinity column could effectively enrich the recombinant proteins, while the third time elution buffer, which contained the relatively pure target proteins, could be used in enzyme activity assays. The loss of target band in empty vector control avoided the false positive result (Supplementary Figure 3A). The five proteins showed CHS activity with varying efficiencies (Figure 5B). ZosmaCHS07 had the highest CHS activity and significantly higher enzyme activity than ZosmaCHS08. However, there were no significant differences in the catalytic efficiency in the ZosmaCHS01/11 pair.