Four-year-old C. chinensis samples were collected from Gaochuan Township, Anzhou District, Mianyang City, China (31°37′49.8″N, 104°12′5.7″E). Seedlings of C. chinensis were cultivated in a greenhouse maintained at 22°C under 16-h day/8-h night cycles with 70% relative humidity at the Key Laboratory of Plant Genetics and Molecular Breeding for 7d. Based on previous research, the seedlings were washed and then transferred into aerated containers filled with half-strength Hoagland’s solution supplemented with 100μM CdCl₂ for Cd treatment (Huang et al., 2019). After exposure to Cd for 0h, 6h, 12h, 24h, and 48h, the roots and leaves of C. chinensis seedlings were harvested for analysis of the expression patterns of the CcMTP gene family. Samples were immediately frozen in liquid nitrogen and stored at –80°C.

The genome sequence of C. chinensis was retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_015680905.1/). To identify all MTP genes in C. chinensis, the HMM profile of the cation efflux domain (PF01545) was downloaded from the Pfam protein family database (http://pfam.xfam.org) and used to screen for potential MTP genes within the C. chinensis genome using HMMER 3.2.1 (http://hmmer.janelia.org/) (Prakash et al., 2017). After removing redundant sequences, the putative CcMTP protein sequences were submitted to the NCBI Conserved Domains Database (https://www.ncbi.nlm.nih.gov/cdd/), SMART database (http://smart.embl.de/) and Pfam for additional confirmation of conserved domains (Marchler-Bauer et al., 2017; Letunic et al., 2021). To further validate the completeness of our identification, we conducted a BLASTP search against the C. chinensis proteome using experimentally characterised MTP protein sequences from A. thaliana and Oryza sativa as queries. This independent analysis confirmed that all candidates identified via the HMM-based approach were consistently recovered, with no additional CcMTP homologs detected.

The chromosomal location data of CcMTPs were retrieved from the assembled C. chinensis genomic repository. These spatial arrangements and gene duplication events were subsequently subjected to graphical representation through TBtools software (Chen et al., 2020). The rates of synonymous (Ks) and non-synonymous (Ka) substitutions per site for duplicated gene pairs were also calculated using TBtools (Chen et al., 2020).

Twenty-five CcMTP protein sequences were obtained from NCBI. Subsequently, protein sequences of the MTP gene family from Arabidopsis and rice were used to construct a phylogenetic tree using MEGA version 10.0, based on the maximum likelihood method with bootstrap values calculated from 1000 replicates (Tamura et al., 2011). The phylogenetic tree was annotated and visualised using EvolView 2.0 server (https://www.evolgenius.info/evolview-v2/).

The physicochemical properties of the confirmed CcMTP proteins, including molecular weight (MW) and theoretical isoelectric point (pI), were calculated using the ExPASy server (https://web.expasy.org/protparam/). TMHMM and WoLF PSORT were used to predict transmembrane domains and subcellular localisation, respectively (https://wolfpsort.hgc.jp/) (Krogh et al., 2001; Yu et al., 2006). Conserved motifs within CcMTP proteins were identified using the MEME server (http://meme-suite.org/tools/meme) with default parameters, specifying up to 10 motifs per gene and motif widths between 15 and 40 amino acids. Conserved domains of CcMTP sequences were further confirmed using InterPro (https://www.ebi.ac.uk/interpro/). Exon-intron structures of CcMTPs were analysed using the GSDS 2.0 online tool (http://gsds.gao-lab.org/). The 2000 bp upstream regions of the 25 identified CcMTPs were extracted as promoter sequences, and related cis-regulatory elements were identified using the PLACE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). Computational structural modelling of CcMTP proteins was performed using the intensive mode of the Phyre2 webserver (http://www.sbg.bio.ic.ac.uk/phyre2/) (Kelley et al., 2015).

Protein sequences of MTPs from A. thaliana, O. sativa, and C. chinensis were subjected to pairwise alignment using the NCBI BLASTP algorithm with default parameters (e-value ≤ 1e–5). Gene pairs with sequence similarity greater than 46% were retained. An SSN was constructed by EFI-EST (Oberg et al., 2023) and visualised using Cytoscape software (v3.10.3). In the network, nodes represent individual protein sequences, and edges represent significant BLASTP alignments meeting a significance threshold (e-value ≤1e-5), and topological organisation was optimised via Cytoscape force-directed layout (Shannon et al., 2003).

Gene-specific oligonucleotide primers were designed and optimised using Primer3 software ( Supplementary Table S2 ) for RT-qPCR analysis. Total RNA was extracted from the 12 samples described previously using the HiPure Total RNA Plus Kit (Magen, R411103). First-strand cDNA synthesis was performed using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, RR047A), following the manufacturer’s protocol. The synthesised cDNA served as a template for qPCR, which was performed using the CFX96™ Real-Time PCR Detection System (Bio-Rad, USA) with SYBR Premix Ex Taq™ (Tli RNaseH Plus, Takara Bio Inc.). The amplification protocol was as follows: initial denaturation at 95°C for 3min, followed by 40 cycles of 95°C for 10s and 60°C for 30s. All reactions were performed in triplicate under identical conditions. The fold changes in RNA transcripts were computed using the 2^−ΔΔCt method (Livak and Schmittgen, 2001), with the 18S rRNA-encoding gene (DQ406855) in C. chinensis used as the internal reference (Liu et al., 2022) ( Supplementary Table S2 ).

To validate the functional roles of CcMTPs, the full-length coding sequences (CDS) of CcMTPs were amplified using Phanta Max Super-Fidelity DNA Polymerase (Vazyme, China). Following sequencing confirmation, PCR amplicons of three target genes (CcMTP11, CcMTP16, and CcMTP24) were directionally cloned into Hind III/Sac I restriction sites of the pYES2 yeast expression vector, generating three recombinant plasmids (pYES2-CcMTP11, pYES2-CcMTP16, and pYES2-CcMTP24) along with an empty vector control. These plasmids were transformed into the Cd -hypersensitive Saccharomyces cerevisiae mutant strain Δycf1 via lithium acetate-mediated transformation for heterologous complementation analysis. Primary transformants were selected on uracil dropout (SD-Ura) agar medium. Cultures were grown at 28°C with orbital shaking (225 rpm) at 28°C until reaching mid-exponential phase (OD₆₀₀ = 1.0 ± 0.05) and serially diluted 10-fold (10^-1, 10^-2, 10^-3). A 50-μL aliquot from each dilution was spotted onto SGR–Ura plates supplemented with various heavy metals: Cd²⁺ (0, 10, 30, 55μM), Mn²⁺ (0, 20, 30, 40mM), Fe²⁺ (0, 40, 60, 65mM), and Zn²⁺ (0, 8, 10, 12mM). Plates were incubated at 30°C for 3–5 d before imaging.

To assess Cd -induced growth inhibition, yeast cells were cultured in 20mL SGR–Ura medium to an OD₆₀₀ = 0.2 and treated with 35μM CdCl₂. OD₆₀₀ was measured at 6-h intervals to generate a growth curve of the transformants. To analyse Cd accumulation within yeast cells, yeast cells were propagated in 20 mL of SGR-Ura medium until the OD₆₀₀ reached 0.2, followed by a 48-h exposure to 35 μM CdCl₂ under standard culture conditions. The yeast cells were then harvested and washed three times with distilled water. Subsequently, they were lyophilised for 48h at 85°C to constant weight, and digested in 5mL nitric acid. Intracellular metal content was determined using inductively coupled plasma-optical emission spectrometry (ICP-OES 7000, Thermo Fisher Scientific, New York, USA).

Statistical analyses of numerical data were performed using one-way analysis of variance (ANOVA) and independent-sample t-tests, followed by post-hoc Duncan tests (p < 0.01) and least significant difference (LSD) comparisons. All computations were executed via the SPSS, version 23.

Using 12 AtMTP and 10 OsMTP protein sequences as the queries, a total of 25 MTP proteins were identified in the C. chinensis genome. According to the sequence analysis, multiple distinct characteristics of the CcMTPs were elucidated ( Table 1 ; Supplementary Figure S1 ): (1) ORF lengths of CcMTPs were determined to span 300–1572 base pairs (CcMTP1 to CcMTP25), encoding 99–523 amino acid residues. Predicted molecular weights (MWs) ranged from 11.4 to 58.6 kDa, and isoelectric points (pI) ranged from 4.72 to 9.05 across the CcMTP family. (2) Most of CcMTPs were predicted to contain two to six transmembrane domains (TMDs), except for CcMTP1, CcMTP18, and CcMTP21, which harboured ten, seven, and eight TMDs, respectively. (3) Subcellular localisation by Wolfpsort predicted that 24 of the 25 proteins localised in both the plasma membrane and tonoplast, with only CcMTP11 predicted to be located in the Golgi apparatus. (4) Multiple sequence alignment revealed that three distinct subfamilies (Zn-, Zn/Fe-, Mn-MTPs) exhibited low sequence homology ( Supplementary Figure S1 ). The CcMTP members classified within the Zn-MTP and Zn/Fe-MTP clades exhibited a conserved HXXXD (X = variable residue) signature sequence spanning both TMD-I and TMD-V. Conversely, the Mn-MTP subgroup was characterised by a divergent DXXXD motif localised in these equivalent transmembrane regions, suggesting evolutionary adaptation for differential metal ion selectivity ( Supplementary Figure S1 ).

To characterise the chromosomal distribution of C. chinensis MTP genes, we mapped the genomic locations of the 25 CcMTP genes across the C. chinensis genome ( Figure 1A ). This study demonstrated a heterogeneous distribution pattern across the nine chromosomes. Chr2, Chr5, and Chr7 harboured the highest number of CcMTP genes (4) CcMTP genes, with Chr1 and Chr6 following closely behind, each containing three CcMTP genes. Chr4, Chr8, and Chr9 each harboured two CcMTP genes, whereas the lowest number (one CcMTP gene) was identified on Chr3.

In this study, we employed the TBtools software to investigate gene duplication events within the CcMTP gene family. Genomic analyses revealed that the CcMTP gene family generated a single segmental duplication event involving the CcMTP12 and CcMTP20 gene pair, as evidenced by phylogenetic and syntenic conservation patterns ( Figure 1B ). This isolated chromosomal duplication event likely represents an evolutionary mechanism driving the functional diversification of metal transport systems in C. chinensis. To investigate the evolutionary dynamics and selective pressures influencing the CcMTP gene family, the values of Ka, Ks, and the Ka/Ks ratio were calculated in the C. chinensis genome. The duplicated CcMTP gene pairs exhibited a Ka/Ks ratio of 0.69 ( Supplementary Table S3 ), suggesting that MTP family genes in C. chinensis may have undergone selective pressure during their evolutionary history.

To investigate the evolutionary relationships of the MTP gene family, phylogenetic analyses were conducted using the maximum likelihood (ML) method ( Figure 2 ). A phylogenetic tree was constructed using MEGA X following multiple sequence alignment of 25 C. chinensis MTP proteins (CcMTPs), 12 A. thaliana MTP proteins (AtMTPs), and 10 rice MTP proteins (OsMTPs). CcMTP protein sequences are listed in Supplementary Table S4 . Based on the topology of the ML phylogenetic tree, MTP genes from C. chinensis, A. thaliana, and rice were divided into three major clades: Zn-MTP, Zn/Fe-MTP, and Mn-MTP. The three subfamilies were resolved to comprise 19, 13, and 15 MTP proteins, respectively. Multiple sequence alignment analysis of CcMTPs and AtMTPs revealed significant sequence homology, highlighting conserved structural features among these proteins ( Supplementary Figure S1 ).

All twenty-five CcMTPs were modelled using Phyre2 with >98% confidence in normal mode to elucidate their molecular mechanisms in C. chinensis. Four known structural templates, c7kzxA and c8f6fA (Cd and zinc efflux pump FieF), c8j7tF and c8j7yA (zinc transporter 7), c7y5gA and c6xpdA (zinc transporter 8), and c8f6fD (another FieF structure), were used for modelling CcMTP1–9, CcMTP11–14, CcMTP16–22, and CcMTP24, respectively. Models of c3h90D (ferrous-iron efflux pump FieF), c7ttiA (solute carrier family 12 member 4), and c8xmaB and c8zsbA (proton-coupled zinc antiporter SLC30A1) were used for the remaining five CcMTPs ( Supplementary Figure S2 ; Supplementary Table S1 ).

The SSN was defined as a graph structure composed of nodes and edges, where each node represents a non-redundant protein sequence (90% identity threshold), and edges connect node pairs with >40% pairwise sequence identity (BLASTP, e-value <1e-10). The MTP family SSN comprised 47 nodes (non-redundant at 90% identity) and 520 edges (≥46% sequence identity). A total of five clusters were obtained from the SSN, with clusters 1, 2, and 3 accounting for 89% of the total nodes ( Supplementary Figure S3 ). The results showed that 11 CcMTP members within the Zn-MTP subfamily clustered in cluster 2, 9 transporters classified as Zn/Fe-MTPs in cluster 3, and 5 CcMTPs within the Mn-MTP clade in cluster 1. In cluster 1, six CcMTPs and AtMTP8 exhibited sequence homology ranging from 47% to 56%, whereas CcMTP1 shared as high as 67% sequence homology with AtMTP12 at the amino acid level. Seven CcMTP genes, CcMTP3, CcMTP6, CcMTP7, CcMTP8, CcMTP15, CcMTP17, and CcMTP19, showing significant homology with AtMTP1 and AtMTP3 were co-clustered in cluster 2 ( Supplementary Figure S3 ).

To acquire a more in-depth understanding of the structural properties of CcMTPs in C. chinensis, conserved motif analysis was performed using the MEME server to identify 10 evolutionarily conserved protein motifs ( Supplementary Figure S4 ). The 25 CcMTP genes exhibited motif counts ranging from 1 to 5, demonstrating quantitative heterogeneity in conserved domain architectures ( Figure 3A ). Motif composition analysis showed that most genes (n = 20) harboured 2–4 motifs, whereas two genes (CcMTP10 and CcMTP23) harboured only a single motif. Three genes (CcMTP14, CcMTP20, and CcMTP21) possessed five motifs. Notably, tandem duplication of motif 1 was detected in 64% of the analysed CcMTP proteins, whereas duplication was absent in CcMTP3, CcMTP12, CcMTP13, CcMTP14, CcMTP20, CcMTP21, CcMTP22, CcMTP24, and CcMTP25. Additionally, CcMTP10 and CcMTP23 exclusively contained motif 1, suggesting that motif 1 may serve as a core functional element, potentially mediating key substrate recognition activities involved in metal transport.

Notably, phylogenetic subgrouping revealed that members within the same clade demonstrated significantly similar conserved motif distribution patterns, whereas members from different clades displayed marked divergence in motif architecture ( Figure 3A ). This observation supports the hypothesis that phylogenetic clustering within the same clade correlates with functional homology, as reflected by the conserved motif arrangements among closely related members. A systematic investigation was conducted to examine the topological distribution of conserved structural motifs within transmembrane α-helices in six representative MTP family members (CcMTP1, CcMTP7, CcMTP8, CcMTP19, CcMTP21, and CcMTP22), with particular emphasis on their spatial arrangement across helical domains ( Figure 3C ).

To assess this further, domain architecture analysis of the CcMTP proteins was performed. The results showed that all CcMTP proteins contained the cation efflux domain. In contrast, the zinc transporter domain was exclusively identified in CcMTP5, CcMTP14, CcMTP20, and CcMTP21 ( Supplementary Figure S5 ).

To further investigate the genomic organisation of CcMTPs, a comprehensive analysis of exon-intron structure was systematically performed via comparison of genomic DNA and CDS sequences ( Figure 3B ). Genomic characterisation of the CcMTP gene family revealed a stable exon number distribution from 1 to 12 across all members. The CcMTP gene family exhibited a distinct distribution pattern, with 10 members (the majority) harbouring 3–4 exons, whereas five genes (CcMTP22, CcMTP9, CcMTP21, CcMTP18, and CcMTP5) represented the extremes with 2, 6, 9, 10, and 12 exons, respectively. In contrast to the conserved motif patterns, no significant congruence was observed between phylogenetic grouping and exon-intron architecture among CcMTP genes within the same clade.

To characterise the cis-regulatory elements within the promoters of CcMTP genes, 2000 bp upstream sequences from the translation start site were computationally analysed for all family members. A total of 16 major cis-acting elements were identified across the 25 CcMTP gene promoters ( Figure 4 ). These elements were functionally categorised into four major regulatory pathways: growth and development, biotic/abiotic stress response, light responsiveness, and phytohormone responsiveness. Promoter analysis revealed that CcMTP gene promoters were particularly enriched in elements associated with light and biotic/abiotic stress responses, including motifs for light responsiveness, drought inducibility, low temperature responsiveness, and general defence and stress response. In addition to the major categories, CcMTPs gene regulatory regions comprised a small subset of elements related to phytohormone response and growth and development responsiveness, such as those responsive to abscisic acid, MeJA, and salicylic acid, along with circadian control elements.

The relative expression patterns of 24 CcMTP genes under Cd²⁺ stress at 0, 6, 12, 24, and 48 h were assessed using qRT-PCR ( Figure 5 ). CcMTP3 was excluded because of unsuccessful primer design. The CcMTP genes exhibited diverse expression responses to Cd²⁺ exposure across roots, stems, and leaves.

In roots, Cd²⁺ increased the expression of 11 CcMTPs (CcMTP2, CcMTP5, CcMTP6, CcMTP9, CcMTP10, CcMTP11, CcMTP14, CcMTP16, CcMTP19, CcMTP24, and CcMTP25); however, it repressed the expression of five CcMTPs (CcMTP7, CcMTP8, CcMTP12, CcMTP13, and CcMTP17). In stems, Cd²⁺ markedly enhanced the expression of 62.5% of CcMTP genes, except for CcMTP6, CcMTP10, CcMTP20, CcMTP21, and CcMTP23. In contrast, it remarkably suppressed the expression of CcMTP12, CcMTP13, CcMTP14, and CcMTP22. Notably, the transcript level of most of CcMTPs reached the peak at 48 h after Cd treatment. For instance, the expression levels of CcMTP4 in stems reached their highest at 48 h, which were 40.5-fold higher than those in the control. In leaves, 58.3% of the CcMTP genes were remarkably upregulated under Cd treatment, except for CcMTP1, CcMTP6, CcMTP8, CcMTP18, CcMTP20, CcMTP23, and CcMTP25, whereas the expression levels of CcMTP12, CcMTP13, and CcMTP14 were downregulated. Excess Cd significantly upregulated the expression levels of CcMTP2, CcMTP5, CcMTP9, CcMTP11, CcMTP16, CcMTP19, and CcMTP24. In contrast, it substantially downregulated the expression levels of CcMTP12 and CcMTP13 in three tissues. Given that smaller gene size facilitates PCR amplification, efficient transformation into yeast for functional characterisation, as well as proper expression and folding, CcMTP11, CcMTP16, and CcMTP24 were selected for subsequent functional validation in this study. CcMTP11, CcMTP16, and CcMTP24 were selected for subsequent functional validation in this study.

Based on the expression patterns of CcMTP genes in Cd-treated C. chinensis plants, CcMTP11, CcMTP16, and CcMTP24, which exhibited high expression levels in roots, stems, and leaves, were selected as the candidate genes for functional analysis in yeast. The yeast mutant Δycf1, deficient in yeast Cd factor 1, was transformed with the yeast expression vector pYES2 carrying CcMTP11, CcMTP16, or CcMTP24, as well as with the empty vector as a control. When cultured on standard SGR-Ura medium without Cd, or with 10 µM and 30 µM Cd, the growth phenotypes of yeast strains overexpressing CcMTPs were indistinguishable from those harbouring the empty vector. However, on medium supplemented with 55 μM Cd, CcMTP24 significantly enhanced the Cd tolerance of the Δycf1 mutant, whereas CcMTP11 and CcMTP16 marginally improved Cd tolerance ( Figure 6A ).

Growth curve analysis further confirmed that yeast strains transformed with CcMTP11, CcMTP16, and CcMTP24 exhibited significantly higher growth rates under Cd stress compared with the Δycf1-pYES2 control ( Figure 6C ). In contrast, no significant growth differences were observed under normal conditions ( Figure 6B ). Notably, after 48 h of exposure to 35 µM Cd, all three transformants accumulated lower Cd levels than the Δycf1 control, with CcMTP16 and CcMTP24 showing especially reduced Cd accumulation ( Figure 6D ). These results suggest that CcMTP24 may mediate Cd detoxification by facilitating its efflux into the extracellular medium.

Additionally, we investigated whether CcMTP11, CcMTP16, and CcMTP24 conferred tolerance to other heavy metals, such as Mn, Zn, and Fe. When expressed in yeast, CcMTP11 and CcMTP24 conferred tolerance to Mn, Zn, and Fe, whereas CcMTP16 enhanced tolerance only to Mn and Zn ( Figure 7 ). These findings demonstrate that the heterologous expression of CcMTP11, CcMTP16, and CcMTP24 in yeast enhances cellular tolerance to various heavy metals.