To identify the PRP genes in cucumber, the annotated cucumber genome (cv. 9930, Li et al., 2019) was downloaded from the website of National Centre for Biotechnology Information (http://www.ncbi.nlm.nih.gov, accessed on 22 December 2021).

The sequences of all known PRPs proteins in Arabidopsis (At1g54970 (AtPRP1), At2g21140 (AtPRP2), At3g62680 (AtPRP3) and At4g38770 (AtPRP4)) were used as queries to search the cucumber genome by BLASTp. Upon finding the candidate sequences, they were then tested against the SWISS-PROT database to determine if the identified PRP family was closely related to other plants.

To obtain the sequence length, molecular weight, and isoelectric point information of the identified PRP proteins, the ExPASy online tool (https://web.expasy.org/protparam, accessed on 3 January 2022) was utilized. Additionally, the PLANT-PLOC online tool, accessed on 6 January 2022) was employed to predict the protein subcellular localizations (Chou and Shen, 2007). The signal peptides and their cleavage sites were predicted by SignalP - 6.0 (https://services.healthtech.dtu.dk/services/SignalP-6.0/, accessed on 24 March 2023).

A phylogenetic tree was constructed using MEGA-X and the maximum likelihood method with Poisson model and 1000 bootstrap replications (Kumar et al., 2018). PRP protein sequences of Arabidopsis and rice were obtained from UniProt website (https://sparql.uniprot.org, accessed on 15 January 2022). The gene structures of CsPRPs were analyzed by gene structure display server (GSDS) online program (http://gsds.cbi.pku.edu.cn, accessed on 16 January 2022). The NCBI BATCH CD-search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi, accessed on 26 January 2022) was used to examine the conserved domain structure based on the corresponding protein sequence, using the Pfam database. The conserved motifs were predicted by the MEME Suite tools (http://meme-suite.org, accessed on 6 February 2022, Bailey et al., 2009), the number of motif parameters was limited to less than 10 manually. The structural characteristics were evaluated with SWISS-MODEL online tools (Guex et al., 2009; Bertoni et al., 2017; Bienert et al., 2017; Waterhouse et al., 2018; Studer et al., 2020).

The physical location information of CsPRP genes was obtained from the GCF_000004075.3_Cucumber_9930_V3_genomic database in NCBI, and all the identified genes were mapped to the cucumber chromosomes by TBtools V1.089 (Chen et al., 2020). Multiple Collinearity Scanning toolkit (MCScanX) with default parameters were used to analyze the gene duplication events (Wang et al., 2012). The expression of CsPRPs was analyzed using FPKM (fragments per kilobase of exon per million mapped fragments) values from RNA-seq datasets, including PRJNA307098, PRJNA321023, PRJNA80169, PRJEB7612, PRJNA214462, PRJNA263870, PRJNA271595, PRJNA285071, PRJNA292785, PRJNA312872, PRJNA319011, PRJNA376073, PRJNA258122, PRJNA345040, PRJNA380322, PRJNA382994, PRJNA388584, PRJNA419665, PRJNA431940, PRJNA437579, PRJNA438923, and PRJNA483118.

The cultivation of plant materials, sample collection, and processing of RNA-seq data were performed as described in the Cucurbit Expression Atlas (http://www.cucurbitgenomics.org/rnaseq/home).

To assess the binding activity of PRP, the coding sequence of PRP1 and PRP3 with or without the putative signal peptide was amplified from cucumber cDNA using primers containing NcoI and XhoI restriction site (as listed in Supplementary Table 1 ). The PCR products were then fused to a 6*His tag in the pET-28a expression vector and transformed into BL21(DE3) pLysS (CD701-02, Transgen, China). Expression of the fusion gene was induced with IPTG and the protein was purified by ProteinIso^® Ni-IDA Resin (DP111-01, Transgen, China).

The silicon deposition characteristics of CsPRP1 and CsPRP3 were then examined, with slight modifications to the method described by Kauss et al. (2003). This involved adding 10 μL of 0.5 mg/mL protein to 180 μL sodium phosphate/citrate buffer of different pH, followed by 10 μL 0.1M orthosilicic acid. After incubation at 25 °C for 30 min, the solution was centrifuged at 10000g for 10 min, the precipitation was collected and washed with distilled water. Silicon content in the precipitation was determined as previously described (Sun et al., 2020).

To investigate the subcellular localization of CsPRP1 and CsPRP3, the coding sequences of them was amplified with primers containing SalI and BamHI restriction sites. The PCR products were fused to the 5’ -end of enhanced green fluorescent protein (eGFP) under the CaMV 35S promoter in the pTF486 expression vector respectively (Primer sequences were listed in Supplementary Table 1 ). Expression of CsPRP1 and CsPRP3 in onion epidermal cells was performed by gene gun transduction as described previously (Duan et al., 2022). Onion epidermal cells transformed with empty vector (pTF486) was introduced as a positive control. The fluorescence signal was determined with a confocal laser scanning microscope (TCS SP8 SR Leica) at an excitation wavelength of 488 nm for GFP.

To investigate the expression character of CsPRP1 and CsPRP3, the promoter (2000bp upstream of the start codon) of CsPRP1 was isolated and inserted at the EcoRI and NcoI sites, and that (2000bp upstream of the start codon) of CsPRP3 was amplified and inserted at the KpnI and NcoI sites of pCambia 1305, of which the promoters may direct the expression of GUS (Primer sequences were listed in Supplementary Table 1 ). The constructs were transformed into Arabidopsis thaliana (Columbia) following a floral dip protocol. To analyze GUS expression patterns, Arabidopsis thaliana seedlings and plants were left to soak in GUS staining buffer (composed of 2 mM X-Gluc in 0.5 M sodium phosphate buffer, pH 7.2, with 2 mM potassium ferrocyanide, 2 mM potassium ferricyanide, and 0.2% Triton X-100) overnight.

Iterative protein BLAST analysis was employed to identify the PRP genes in cucumber, using previously identified Arabidopsis PRP sequences as queries. The candidate Cucumis sativus PRP (CsPRP) genes were further scrutinized using BLASTp in the Swiss-Prot database and Batch-CDD tests. As a result, seven CsPRP genes were identified and their basic characteristics, including CDS lengths, amino acid sequence lengths, molecular weights, and isoelectric points, were analyzed and presented in Table 1 . The localization of these proteins was predicted, with CsPRP2 being predicted to be located at the cell wall, CsPRP6 at the plasma membrane, CsPRP7 at the Cell membrane, Cytoplasm and Nucleus, and the other PRPs at the Nucleus. CsPRP6 and CsPRP7 were identified as hydrophilic proteins, while the other PRPs in cucumber were all hydrophobic. Additionally, a signal peptide with a length of 21–35 Aa was identified at the N terminal of the CsPRPs protein.

An unrooted maximum likelihood tree was constructed using the MEGA software to analyze the phylogenetic relationship between PRPs from cucumber, Arabidopsis, and rice. These plants are representative of dicotyledonous and monocotyledonous model organisms. Figure 1 illustrates that CsPRPs exhibit a closer relationship with AtPRPs than OsPRPs. Notably, CsaV3_2G012830(CsPRP3) is closely related to CsaV3_2G012840(CsPRP1).

Chromosome distribution of PRP gene families was analyzed based on the physical location of the GCF_000004075.3_Cucumber_9930_V3_genomic database from the NCBI website. Gene duplication events of these genes including segmental and tandem duplications were analyzed using the MCScanX software. Results demonstrated that the CsPRP genes were distributed on two chromosomes-PRP1, PRP2, PRP3 and PRP4 on chromosome 2, and PRP5, PRP6, PRP7 on chromosome 7. CsPRP1-CsPRP3 and CsPRP6-CsPRP7 were identified as tandem duplication, CsPRP7 and CsPRP5 were identified as segmental duplication ( Figure 2A ).

The Gene Structure Display Server (GSDS) online program was used to analyze the intron-exon structure. Analysis on the conserved domain structure was based on the online program of NCBI BATCH CD-search tool (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi, accessed on 12 April 2020). The MEME search tool was employed to predict the conserved protein motifs. The results showed that the CsPRP genes had no more than one intron ( Figure 2B ). Pollen_Ole_e_I like (pfam01190) is the only member of the superfamily cl03128. DNA polymerase III subunits gamma and tau (PRK14950) is the only member of the superfamily cl36446. Large tegument protein UL36 (PHa03247) is the only member of the superfamily cl33720. Predic_Ig_block (TIGR04526) is the only member of the superfamily cl37462. CsPRP4, CsPRP5, CsPRP6 and CsPRP7 had Pollen_Ole_e_I domain, PHA03247 superfamily domain was conserved in CsPRP2 and CsPRP4, and predic_Ig_block superfamily domain only existed in CsPRP2 ( Figure 2C ). Alignment of CsPRPs demonstrated that the sequences differed and conserved motifs analysis indicated that CsPRP1 and CsPRP3 share similar motif pattern, both have four motifs. CsPRP4 and CsPRP5 share similar motif composition but the order was distinctive ( Figure 2 ; Supplementary Figures 1 , 2 ), which may be associated with the segmental duplication event. The tertiary structure analysis indicates that CsPRP1 and CsPRP3 share a similar spatial conformation ( Figures 2D–J ), and it indicates that they may share similar biological functions.

This study achieved the prokaryotic expression of CsPRP1 and CsPRP3 in Escherichia coli BL21 by fusing their coding region sequences with a HIS tag protein on the pET28a vector, and then purified them with ProteinIso Ni IDA Resin. The binding experiment with silicon revealed that both CsPRP1 and CsPRP3, with or without a signal peptide, had a noticeable affinity for silicon, though their optimal pH values were different. CsPRP1 and CsPRP3 with signal peptide removed had the strongest binding capacity to silicon at pH 6.0, while CsPRP1 showed the most obvious binding at pH 7.0-7.5. The optimal pH of CsPRP3 binding to silicon was likely lower than 5.5 ( Figure 3 ).

In this research, the coding sequences of CsPRP1 and CsPRP3 were fused with green fluorescent protein (GFP) and introduced into onion epidermal cells using the gene gun method. The results showed that the empty vector control caused the interior of the onion cells to emit green fluorescence, except for the large vacuoles. However, the proteins of CsPRP1 and CsPRP3 were specifically localized on the cell walls and exhibited distinct polarity distribution patterns. The presence of green fluorescence on adjacent cell walls suggested that these proteins may have the ability to move between neighboring cells ( Figure 4 ).

Expression investigation of PRP family genes from high-throughput sequencing data showed that CsPRP6 and CsPRP7 had similar expression patterns, with expression only detected in the root differentiation and elongation regions, indicating their potential role in cucumber root development. CsPRP1 and CsPRP3 were found to be expressed in cucumber roots, leaves, and flower organs, but not in stems and fruits. The expression of CsPRP2, CsPRP5, and CsPRP4 varied across different tissue parts of cucumber, with CsPRP2 mainly expressed in roots and fruits, and CsPRP5 and CsPRP4 mainly expressed in leaves, fruits, and flower organs. Additionally, under GA treatment conditions, the expression of CsPRP1 and CsPRP4 was significantly reduced, suggesting their involvement in gibberellin-mediated signal transduction and growth and development in cucumber ( Figure 5 ).

The promoter regions of CsPRP1 and CsPRP3 were isolated and ProCsPRP1::GUS and ProCsPRP3::GUS expression vectors were constructed and used to generate transgenic Arabidopsis. GUS staining at different growth stages of the transgenic Arabidopsis revealed that blue was observed in the mature leaves and roots of ProCsPRP1::GUS transgenic Arabidopsis seedlings, while no GUS gene expression was detected in the young leaves, stems, and root tips ( Figure 6A ). The same pattern was observed in ProCsPRP3::GUS transgenic Arabidopsis seedlings ( Figure 6B ). These findings suggest that CsPRP1 and CsPRP3 are likely only expressed in mature leaves and roots during the seedling stage.

During the reproductive growth stage, CsPRP1 is primarily expressed in the leaves, roots, petals, and stamens of the plant. The expression of CsPRP1 in the leaves increases with age, although it is scarcely detected in young leaves. In the roots, expression is limited to the lowest-level lateral roots ( Figures 7A–D ). CsPRP1 is also expressed in the petals and stamens ( Figure 7I ). CsPRP3 exhibits a similar expression pattern to CsPRP1 in the leaves. In the roots, it is only expressed in the lowest lateral roots, but at significantly lower levels than CsPRP1 ( Figures 7E–H ). CsPRP3 is expressed solely in the petals of the flowers, and neither gene is expressed in the pods or seeds ( Figures 7J–L ).