Bioinformatics analysis of the two target proteins, MatK1 and MatK2, was performed using data retrieved from the UniProt database (MatK1, entry ID A0A2R2V059; MatK2, entry ID A0A125QY04) (Figure 1, Supplementary Figure S1–S5). The molecular weights of MatK1 and MatK2 were approximately 60.31 kDa and 60.89 kDa, respectively (Table 1). Their molecular formulas were C₂₇₈₄H₄₃₁₇N₇₅₁O₇₂₂S₁₄ for MatK1 and C₂₈₀₁H₄₃₅₀N₇₆₂O₇₃₆S₁₃ for MatK2. The amino acid composition of MatK1 included 36 negatively charged residues and 70 positively charged residues, while MatK2 contained 35 negatively charged residues and 70 positively charged residues. The grand average of hydropathy (GRAVY) values for MatK1 and MatK2 were −0.10 and −0.12, respectively (Table 1). The aliphatic indexes for MatK1 and MatK2 were 103.02 and 101.24, respectively (Table 1). Both MatK1 and MatK2 had an estimated half-life of 30 h (Table 1). The isoelectric points (pI) for MatK1 and MatK2 were 10.01 and 10.04, respectively, with protein instability indices of 51.32 and 50.08 (Table 1).

PSIPRED analysis (Buchan et al., 2024; Jones, 1999) revealed that MatK1 contains 227 alpha helices (44.86%) in its secondary structure, along with a significant number of extended strands and random coils (Figure 2a; Table 2). Similarly, MatK2 consists of 216 alpha helices (41.86%) and numerous strands and coils in its predicted secondary structure (Figure 2b; Table 2).

The three-dimensional (3D) structures of MatK1 and MatK2 were predicted using AlphaFold2 (Wayment-Steele et al., 2023; Jumper et al., 2021), which employs deep learning algorithms for more accurate and reliable protein structure predictions compared to traditional homology modeling methods.

To assess the quality of the predicted models (Figures 3a,d), we used the Ramachandran plot to evaluate the dihedral angles of the protein backbone, ensuring they fell within acceptable regions indicative of a stable conformation. For MatK1, 86.5% of the residues were in the most favored region, 11.9% in the additionally allowed region, 0.8% in the generously allowed region, and 0.8% in the disallowed region (Figure 3b; Table 3). For MatK2, 84.6% of residues were in the most favored region, 13.7% in the additionally allowed region, 1.4% in the generously allowed region, and 0.2% in the disallowed region (Figure 3e; Table 3).

ProSA analysis revealed Z-scores of −5.39 for MatK1 and -5.68 for MatK2 (Figures 3c,f), further supporting the high quality of the predicted models.

While the overall fold of MatK1 closely resembles that of MatK2 (Figure 4), the root mean square deviation (RMSD) for all atoms was 1.05 Å, with a sequence identity of 85.30% (Figure 4).

Using the predicted models (Figures 3–5), we utilized the GalaxyWEB program (Seok et al., 2021; Heo et al., 2013; Heo et al., 2016; Ko et al., 2012) to identify the active sites of MatK1 and MatK2. The analysis revealed that the active site residues of MatK1 are H33, N34, K51, S52, S53, and L54 (Figure 5a). For MatK2, the active site residues include K58, R59, T62, R63, and Q66 (Figure 5b). These residues are likely involved in substrate interactions, potentially forming bonds with the substrate’s side chain atoms.

To investigate the spatiotemporal expression profiles of MatK1 and MatK2, we performed real-time quantitative polymerase chain reaction (RT-qPCR) using gene-specific primers (Supplementary Table S1). The results showed that the highest expression of both MatK1 and MatK2 occurred in the leaves at 14:00 (Figure 6). Specifically, MatK1 expression was upregulated by 553.8-fold in leaves, 4.2-fold in flowers, 1.7-fold in stems, and 1.1-fold in roots at 14:00 (Figure 6). Similarly, MatK2 expression was upregulated by 267.5-fold in leaves, 4.2-fold in flowers, 1.3-fold in stems, and 1.0-fold in roots at 14:00 (Figure 6). These results suggest that MatK1 and MatK2 are primarily involved in chloroplast photosynthesis, aligning with previous studies (Muino et al., 2024; Hertel et al., 2013; Barthet and Hilu, 2007; Qu et al., 2018).

We conducted RT-qPCR analysis to examine the expression levels of the MatK1 gene under heat and salt treatments in leaves, as MatK1 exhibited higher expression in leaves compared to MatK2 (Figure 6). The results showed that MatK1 expression in leaves increased with the duration of heat treatment, peaking at 40°C for 3 h and 30°C for 6 h, before declining (Figure 7a). Similarly, MatK1 expression in leaves also increased with the duration of salt treatment, peaking at 300 mM NaCl for 3 h and 200 mM for 12 h, before decreasing (Figure 7a). These findings suggested that temperature and salt concentration influence the photosynthetic rate of lavender, supporting the link between MatK1 and photosynthesis.

The amino acid sequences of MatK1 (UniProt code A0A2R2V059) and MatK2 (UniProt code A0A125QY04) (Figure 1, Supplementary Figure S1–S5) were analyzed using the ProtParam (Gasteiger, 2003; Duvaud et al., 2021) to predict their chemical properties and physicochemical parameters.

Structural predictions of the target proteins (MatK1 and MatK2) were performed using the AlphaFold2 program (Wayment-Steele et al., 2023; Jumper et al., 2021). Secondary structures were predicted with the PSIPRED program (Jones, 1999; Buchan et al., 2024), and active site residues were identified using the GalaxyWEB program (Ko et al., 2012; Heo et al., 2013; Heo et al., 2016; Seok et al., 2021). Multiple sequence alignment data were obtained from the LSQKAB program within the CCP4 suite (Collaborative Computational Project, Number, 1994), and the root mean square deviation (RMSD) for Cα atoms was calculated. Structural images were generated using PyMOL 2.3.4 (https://www.pymol.org/2/).

To validate the tertiary structures, we used the PDBsum database (Laskowski, 2022; de Beer et al., 2014; Laskowski, 2004; 2009; Laskowski et al., 2017) to generate Ramachandran plots for MatK1 and MatK2. This tool helps assess and validate protein structure quality by identifying geometric errors and improving accuracy. The Ramachandran plot specifically evaluates the stereochemical properties of the structures, displaying the dihedral angles of amino acid residues, highlighting allowed conformational regions, and identifying disallowed orientations.

Additionally, ProSA (Protein Structure Analysis) is a widely used tool for analyzing and validating predicted protein models (Wiederstein and Sippl, 2007). It aids in the analysis of protein structures derived from X-ray crystallography and NMR spectroscopy, identifying structural errors and pinpointing problematic regions, thereby improving the interpretation of the protein structures.

To quantify the expression levels of MatK1 and MatK2 under different light conditions, real-time quantitative PCR (RT-qPCR) was performed using PowerUp SYBR Green Master Mix (Applied Biosystems). Total RNA was extracted with the Universal Plant Total RNA Extraction Kit (Bioteke, Beijing, China) according to the manufacturer’s instructions. cDNA was synthesized from RNA using the PrimeScript 1st Strand cDNA Synthesis Kit (Takara, Kyoto, Japan). The primers used are listed in Supplementary Table S1. RT-qPCR was conducted with the Applied Biosystems QuantStudio 5 instrument, and data were analyzed using the 2^−ΔΔCT method (Green and Sambrook, 2018; Schmittgen and Livak, 2008; Livak and Schmittgen, 2001; Liu et al., 2025d; Liu et al. 2025a; Liu et al. 2025b; Liu et al. 2025c; Liu et al. 2024a; Liu et al. 2024b). Relative expression ratios are presented as log₂ values in histograms. Beta-actin served as the housekeeping gene for normalization, with a positive control using the beta-actin gene. A ratio greater than zero indicated up-regulation, while a ratio less than zero indicated downregulation.

All experiments were conducted at least in triplicate. The data were expressed as mean ± SD. Statistical analysis was conducted using Origin 8.5, Microsoft Excel 2013 and SPSS 19.0. In the all statistical evaluations, p < 0.05 was considered statistically significant, and p < 0.01 was considered high statistically significant.