JlAO Hui-jun,WANG Hong-wei,RAN Kun,DONG Xiao-chang,DONG Ran,WEl Shu-wei ,WANG Shao-min

Shandong Institute of Pomology,Tai’an 271000,P.R.China

Abstract Arabinogalactan proteins (AGPs) are widely distributed in the plant kingdom and play a vital role during the process of plant sexual reproduction. In this study,we performed a comprehensive identification of the PbrAGPs expressed in pear pollen and further explored their influences on pollen tube growth. Among the 187 PbrAGPs that were found to be expressed in pear pollen tubes,38 PbrAGPs were specifically expressed in pollen according to the RNA-seq data.The PbrAGPs were divided into two groups of highly expressed and specifically expressed in pear pollen. We further tested their expression patterns using RT-PCR and RT-qPCR. Most of the PbrAGPs were expressed in multiple tissues and their expression levels were consistent with reads per kilobase per million map reads (RPKM) values during pollen tube growth,implying that PbrAGPs might be involved in the regulation of pear pollen tube growth. We also constructed phylogenetic trees to identify the functional genes in pear pollen tube growth. Therefore,19 PbrAGPs (PbrAGP1 to PbrAGP19) were selected to test their influences on pollen tube growth. Recombinant proteins of the 19 PbrAGP-His were purified and used to treat pear pollen,and 11 of the PbrAGP-His recombinant proteins could promote pear pollen tube growth. Additionally,pollen tube growth was inhibited when the expression levels of PbrAGP1 and PbrAGP5 were knocked down using an antisense oligonucleotide assay. PbrAGP1 and PbrAGP5 were localized in the plasma membrane and might not alter the distribution of pectin in the pollen tube. In summary,this study identified the PbrAGPs expressed in pear pollen and lays the foundation for further exploring their functions in pollen tube growth.

Keywords: arabinogalactan proteins,pear,pollen tube,antisense oligodeoxynucleotides,sexual reproduction

1.lntroduction

Arabinogalactan proteins (AGPs) belong to the hydroxyproline-rich glycoprotein (HRGP) superfamily,and they are rich in hydroxyproline/proline (Hyp/Pro),alanine (Ala),serine (Ser) and threonine (Thr). Based on the composition of amino acids,AGPs are divided into five subfamilies: classical AGP,moderately glycosylated extensins-AGP (EXT-AGP),fasciclin-like AGPs (FLA),arabinogalactan (AG) peptides and lysine-rich AGP. The classic AGP protein contains a signal peptide,a Hyp/Pro rich region and a C-terminal hydrophobic domain. The amino acid sequences of the proteins in several other subfamilies are similar and contain diverse conserved domains. Nevertheless,the genetic structures of AGPs are complicated,as their backbones are attached by various glycans containing arabinose and galactose(Schultzet al.2002;Johnsonet al.2003). As a result,the mechanisms for the specific functions of AGPs remain poorly understood.

Based on a previous study,β-glucosyl Yariv reagent(β-GlcY) and monoclonal antibodies are often applied to test the AGPs (Yarivet al.1962). β-GlcY can interact with AGPs and reveal their cellular distribution (Nothnagel 1997;Seifert and Roberts 2007). InArabidopsis,β-GlcY can inhibit root growth and alter the morphology of epidermal cells (Willats and Knox 2010). They can influence the assembly of pollen tube cell walls and alter the distributions of the pectin,cellulose and callose during pollen tube growth (Royet al.1998). Anti-AGPs mAbs,including JIM8,JIM13 and MAC207,were also used to identify AGPs (Coimbra and Duarte 2003;Lora and Hormaza 2018). However,due to the common carbohydrate epitopes that anti-AGPs mAbs often have,they cannot specifically identify AGPs (Nothnagel 1997).In recent years,in situhybridization,microarray data analysis,real-time (RT)-PCR and promoter analysis have also been applied to the study of AGPs (Pereiraet al.2014),and many AGPs have been identified fromArabidopsis,tomato,pear and tobacco (Duet al.1996;Sunet al.2004;Showalteret al.2010;Jiaoet al.2018).Most AGPs are expressed throughout the plant,but half of them show specific expression in different tissues,such as pollen,roots,stems,leaves and siliques. It has been reported that AGPs participate in multiple developmental processes,including root morphology(Seifertet al.2002),pollen tube growth (Cheunget al.1995;Wuet al.1995;Levitinet al.2010),plant hormonal signaling pathways (Suzukiet al.2002),somatic embryogenesis (Hengel and Keith 2010b),programmed cell death (Motoseet al.2004) and wound responses(Guan and Nothnagel 2004).

AGPs have been identified from transmitting tissue(TT),embryo sac,functional megaspore (FM),synergids,filiform apparatus,extracellular matrix (ECM),style,stigma and ovary,and play a crucial role in the process of plant sexual reproduction. AGP18 was identified fromArabidopsisand located in ovules. In theAGP18RNAi line,female gametophytes could not grow and divide mitotically (Acosta-Garcia and Vielle-Calzada 2004). AGP22 and AGP24 were reported to be involved in megaspore mother cell (MMC) and FM.AGP19was shown to be expressed in the styles,TT,siliques and ovary wall,and the knockout mutant exhibited a significant reduction in flowers and had fewer and shorter siliques(Yanget al.2010). AGPs have also been detected in ovular tissues and are involved in pollen tube guidance.AGP1,AGP9,AGP12andAGP15have also been identified in stigmatic cells,styles,TT and septum (Pereiraet al.2014). A recent report indicated that AGP30 can participate in root regeneration and seed germination(Hengel and Keith 2010a). In tomato,LeAGP1 was found to be localized in the ECM of style transmitting tissue.Additionally,LeAGP1 presented a relatively high level of glycosylation in immature flowers,indicating that it could be involved in pollen tube germination and growth in the style (Gao and Showalter 2000). In olive,AGPs are synthesized and increased in concentration in conjunction with the process of olive pollen germination (Castroet al.2013). AGPs can be used as markers for reproductive development as their expression levels are affected by time and space (Coimbraet al.2007). In apple flowers,three putative AGP genes specifically expressed in anthers were isolated (Choiet al.2010),andOsAGPsthat are predominantly expressed in inflorescences were identified from rice (Anand and Tyagi 2010).

AGP6,AGP11,AGP23,AGP40andFLA3are specifically expressed in pollen;whileAGP15,AGP21,AGP22andAGP24are expressed in multiple tissues(Lalanneet al.2004;Pinaet al.2005;Pereiraet al.2006).As indicated by GFP fluorescence,AGP6andAGP11were detected in pollen grains and pollen tubes (Coimbraet al.2008,2009). It is notable thatAGP6is only present in the vegetative cell wall (Pereiraet al.2013). The double mutant ofagp6agp11presented remarkable phenotypic traits,including the developmental defects in the pollen grains,and the pollen germination and pollen tube growth rates were reduced (Coimbraet al.2009,2010). Interestingly,a triple mutant ofagp6agp11agp40exhibited an obvious phenotype and produced few seeds(Nguema-Onaet al.2012). Hence,AGP6 and AGP11 are crucial for pollen grain and pollen tube growth.Additionally,the AGPs are essential for pollen tube growth,but it is unclear how they regulate the growth of pollen tubes. The C-terminal of AGP proteins contains the glycosylphosphatidylinositol (GPI) lipid anchor. AGP can be bound to the external surface of the plasma membrane by GPI and released into the extracellular matrix under the cleavage action of specific phospholipases,and these changes can alter the functional properties of the proteins (Lalanneet al.2004). Moreover,AGPs can be regarded as nutrients for pollen tube growth due to their rich sugar contents and abundant distribution in the style. They can also act as signaling molecules due to their high sugar content,and the sugars can be cleaved by specific enzymes and released into the extracellular medium (Showalter 2001). NaTTS can form a top-down concentration gradient in the style to guide pollen tube growth (Wuet al.2000),and NaTTS de-glycosylation by the pollen tube cell wall was also predicted as a source of nutrients to support pollen tube growth (Cheunget al.1995). Additionally,AGPs can act as calcium chelators and contribute to the release of Ca2+in cells under specific conditions (Lamport and Péter 2012;Lamportet al.2014).

Based on previous reports,we identified 522PbrAGPsfrom pear (Jiaoet al.2018). Among them,187PbrAGPswere expressed in pollen,and 38PbrAGPswere specifically expressed in pollen by screening others in the RAN-seq database. Nineteen of the PbrAGPs might have the potential function of regulating pollen tube growth,as identified using RT-PCR,RT-qPCR and phylogenetic tree analysis. By treating pear pollen tubes with purified recombinant proteins,11 PbrAGP-His recombinant proteins were found to promote pollen tube growth. The antisense oligonucleotides (as-ODN)assay showed that inhibiting the expression ofPbrAGP1andPbrAGP5led to the repression of pear pollen tube growth. Moreover,the proteins PbrAGP1 and PbrAGP5 were localized in the plasma membrane. In this study,we identified thePbrAGPsthat are expressed in pear pollen and determined their effects on pollen tube growth.The results of this study lay the foundation for further functional exploration of the role of AGPs in pollen tube growth.

2.Materials and methods

2.1.Gene identification and heatmap analysis

A total of 522PbrAGPswere identified from pear based on previous research (Jiaoet al.2018). The reads per kilobase per million map reads (RPKM) values of thePbrAGPsin pollen developmental stages were acquired from the transcriptome database of pollen tube development (Zhouet al.2016). Mature pollen grains(MP,0 min post-culture),hydrated pollen grains (HP,40 min post-culture),growing pollen tubes (PT,6 h post-culture)and stopped-growth pollen tubes (SPT,15 h post-culture)were used to construct RNA-seq libraries and obtain RNAseq data. TheP.bretschnriderigenome was used as the reference genome,and the raw reads with low quality sequences removed were mapped to the pear genome.The remaining sequences were mapped to the pear genome using SOAPaligner/soap2 Software with default parameters,except that mismatches were set to 2 (Liet al.2008). A variation of the RPKM method was used to calculate gene expression,and differentially expressed genes were filtered using R package DEGseq (Zhouet al.2016). We screenedPbrAGPsRPKM values from the RNA-seq database,and used them as the reference for analyzing the expression levels ofPbrAGPgenes during pollen tube growth. The RPKM values were normalized into a range of -1 to 1.5 and displayed as heat maps using TBtools-master.

2.2.Expression analysis

The roots,stems,leaves,pollen tubes,fruit and styles were collected from ‘Dangshan Suli’ and stored at -80°C.Total RNA was extracted from the frozen tissue using a Plant Total RNA Isolation Kit based on the manufacturer’s instructions. We used the TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech,Beijing,China) to reverse transcribe the cDNA and 1 μg of total RNA was used for each sample. RTqPCR analysis was carried out using LightCycler 480 SYBR Green I Master Mix (Roche,Basel,Switzerland)according to the manufacturer’s protocol,and the relative expression levels were calculated by the 2-ΔΔCtmethod.All RT-qPCR were performed in biological triplicates.The primers used in the RT-PCR and RT-qPCR assays are shown in Appendix A.PyrusTubulin(accession no.AB239681) was used as the internal control gene.

2.3.Expression and purification of recombinant PbrAGPs-His proteins from Escherichia coli

We identified 19PbrAGPsthat are expressed in pollen tubes and further analyzed the signal peptides using Signal 4.0. The full-length genes of 19PbrAGPswithout termination codons and signal peptides were amplified and fused into the pCold-TF DNA vector. We used the prokaryotic expression method to acquire the recombinant proteins of the PbrAGPs. The fusion plasmids were transformed into theE.colihost strain BL21 separately.We selected the positive clones and cultured them in Luria-Bertani (LB) medium containing 100 μg mL-1of ampicillin. The bacteria were incubated with shaking at 37°C and 220 r min-1for 16 h,and then transferred to fresh medium with dilution. After the LB medium was added,culturing was continued with shaking at 37°C and 220 r min-1until the OD600reached approximately 0.6.Then,the medium was quickly cooled on ice and allowed to stand for 40 min,and 0.5 mmol L-1IPTG was added into the medium. The bacterial cells were incubated with shaking at 15°C and 220 r min-1for 20-24 h,and collected using the centrifuge at 4°C and 12 000×g for 5 min. The precipitate was re-suspended using PBS buffer. The bacterial cell walls were broken using an Ultrasonic Cell Disruption System (Atpio,Nanjing,China)in low temperature conditions until the solution became clear. After centrifuging for 20 min,the supernatant was filtered using a 0.22-μm filter membrane. The supernatant containing the soluble proteins was loaded onto a column loaded with 5 mL Ni-NTA His Bind Resin (Novagen,Madison,America). The proteins were eluted with the binding buffer (500 mmol L-1NaCl,20 mmol L-1Tris-HCl,10 mmol L-1imidazole,pH 7.9). The recombinant proteins of PbrAGP-His were tested using SDS-PAGE.

2.4.The treatment of ‘Dangshan Suli’ pollen tubes with the PbrAGP-His recombinant proteins

The PbrAGP-His recombinant proteins were dialyzed(Spectra/por?membrane,molecular cutoff 2 000-3 000)against 1 L pollen basal medium at 4°C for 24 h and then stored at -80°C.

The ‘Dangshan Suli’ pollen grains were pre-cultured for 40 min at 25°C with basal medium (which consists of 10%sucrose,30 mmol L-12-(4-morpholino) ethanesulfonic acid (MES),0.04% Ca (NO3)2·4H2O,0.01% H3BO3,pH 6.2 adjusted with Tris). The pre-cultured pollen was divided into 2 mL centrifuge tubes and a concentration gradient of PbrAGP-His recombinant proteins was used to detect the effect on pollen tube growth. After 2 h,images were obtained using a microscope. The pollen tube length was measured by the software IPWIN32.

2.5.Phylogenetic tree analysis of PbrAGPs from pear and Arabidopsis thaliana

The sequences ofArabidopsiswere downloaded from NCBI,and the sequences of pear were acquired from pear genome database (Wuet al.2013). The amino acid sequences of PbrAGPs were used to construct the phylogenetic tree,and the trees were constructed by MEGA7.0. The Jones-Taylor-Thornton and Gamma(JTT+G) model was the optimal model for constructing the phylogenetic tree,as tested by MEGA 7.0 (Appendix B).Amino acid sequence alignments were performed by Muscle with default parameters. The evolutionary history was inferred using the Maximum Likelihood (ML) and Neighbor-Joining (NJ) methods based on the JTT+G model. The bootstraps of the phylogenetic trees were set as 1 000. The NWK file of the phylogenetic tree was uploaded to iTOL,and then the tree was modified with a color scheme and displayed.

2.6.Gene inhibition of as-ODN

The construction ofPbrAGPmRNAs were predicted using the RNA fold Web Server (http://rna.tbi.univie.ac.at/cgibin/RNAfold.cgi). The as-ODN primer design standards were that the length of the primer was 18-22 bp and the GC contents were higher than 60%. SnapGene was used to evaluate candidate as-ODN sequences. The primer sequences were modified by phosphorothioated and purified by high-performance liquid chromatography. The primers for as-ODN analysis are shown in Appendix A.The primers were mixed with transfection reagent for 15 min,and they were then added into the pre-cultured pollen and culturing continued for 1 h. The images were taken using a Nikon Eclipse E100 microscope and IPWin32 Software was used to measure pollen tube lengths.

2.7.Sub-cellular localization of PbrAGP1 and PbrAGP5

The full-length cDNAs ofPbrAGP1andPbrAGP5were fused into the vector pCAMBIA1302 under the control of the CaMV 35S promoter to form a fusion vector. After confirmation of the correct fusion by DNA sequencing,the fusion construct and the control vector (35S:GFP) were transferred intoAgrobacteriumtumefaciensstrain GV3101 by the freeze/thaw method. The PbrAGP1:GFP and PbrAGP5:GFP constructs were introduced into tobacco epidermal cells byAgrobacterium-mediated transfection.After two days,the leaves were cut into 1 cm2pieces which were placed into centrifuge tubes with 5 mL ddH2O.Then,the solution of DAPI and FM4-64 was added into the centrifuge tubes,and the tubes were vacuumtreated for 10 min. The samples were then examined for green fluorescence signals using an LSM800 (Zeiss,Oberkochen,Germany). The primers of PbrAGP1-GFP and PbrAGP5-GFP are shown in Appendix A.

3.Results

3.1.ldentification and analysis of PbrAGPs expressed in pollen tubes of pear

Based on our previous report,522PbrAGPgenes were identified from the pear genome database (Jiaoet al.2018). To explore the roles of PbrAGPs during pollen tube growth,we collected 187PbrAGPsgenes that are expressed in pollen tube growth. We further analyzed thePbrAGPexpression patterns based on the RNA-seq database of dynamic pollen tube development,and found that 62PbrAGPsgenes were highly expressed in pollen(RPKM＞10) and 52 of them were expressed in the four stages of pollen tube growth (Appendix C). Additionally,we also explored the expression profiles ofPbrAGPsby referring to the RNA-seq data from other pear tissues,including root,petal,leaves,fruit,stem,ovary,sepal and bud (Appendix D). The results showed that most of thePbrAGPsgenes were expressed in multiple tissues,and only 38 genes were specifically expressed in pollen tubes (fragments per kilobase million,FPKM＞0.715).Additionally,we also annotated the functions of PbrAGPs using the Blast2Go Software,and the annotations were complicated. This implies that PbrAGPs may play a key role during pollen tube growth.

3.2.Expression pattern analysis of highly expressed PbrAGPs in pear

To further explore the influences of PbrAGPs on pollen tube growth,we performed RT-PCR analysis to determine their expression patterns. Initially,PbrAGPswere divided into two groups,for those highly expressed in pollen tubes and specificity expressed in pollen tubes,based on the RNA-seq data analysis. We then further examined their expression patterns in root,stem,leaf,fruit,pollen tube and style. We firstly tested the expression patterns of thePbrAGPswhich were highly expressed in pollen tubes,and the results showed that 23PbrAGPswere expressed in multiple tissues:Pbr032240.1,Pbr03847.1,Pbr032242.1,Pbr040022.1,Pbr000522.1,Pbr010124.1,Pbr032370.1,Pbr008237.1andPbr039425.1were mainly expressed in pollen and style;whilePbr014494.1,Pbr023217.1andPbr033446.1were specifically expressed in the pollen tube (Fig.1).

Fig.1 RT-PCR analysis of the pollen highly-expressed PbrAGPs in six tissues of pear.

Additionally,RT-qPCR was used to examine their expression trends during pollen tube growth. The results indicated that the expression levels of 12PbrAGPswere variable and consistent with transcriptome data during pollen tube growth.Pbr023217.1,Pbr040022.1,Pbr008237.1,Pbr020882.1,Pbr007577.1,Pbr042667.1,Pbr028004.1andPbr002660.1expression levels were up-regulated along with pollen tube development and then down-regulated at the late pollen tube development stage;whilePbr027542.1andPbr021053.1were downregulated along with pollen tube growth (Fig.2).

Fig.2 RT-qPCR analysis of the pollen highly-expressed PbrAGPs during pollen tube growth. MP,HP,PT and SPT correspond to the four different developmental stages of matured pollen,hydrated pollen,growing pollen tubes after six hours of hydration and stopped-growing pollen tubes,respectively.

A total of 31PbrAGPgenes expressed in pollen were identified using RT-PCR,however,the expression levels of 10PbrAGPgenes were not detected using RTqPCR. AlthoughPbr025277.1andPbr042667.1were expressed in six tissues,and their expression levels were consistent with the RPKM values,and the RT-PCR results showed that their expression levels were lower than those of the otherPbrAGPs. On the other hand,multiplePbrAGPs,such asPbr039425.1,Pbr033446.1,Pbr032270.1,Pbr003847.1,Pbr000552.1,Pbr032242.1,andPbr010124.2,were identified with RT-PCR and mainly expressed in pollen and style. However,their expression levels were not consistent with the RPKM values during pollen tube growth (Appendix E).Pbr023217.1showed low expression in pollen,and its expression trend was consistent with the RPKM value.Additionally,the RPKM value ofPbr023217.1was considerably higher than the band brightness in RTPCR. Therefore,based on the RT-PCR and RT-qPCR results,ninePbrAGPs(Pbr040022.1,Pbr027542.1,Pbr000392.1,Pbr021053.1,Pbr008237.1,Pbr020882.1,Pbr007577.1,Pbr042667.1andPbr025277.1)were selected and their functions in regulating pollen tube growth were analyzed further.

3.3.Expression pattern analysis of PbrAGPs expression specificity in pear

Considering that 38PbrAGPsisolated from the RNA-seq data were expressed only in pollen,we further analyzed their expression patterns. Although 38PbrAGPswere detected from the RNA-seq database,the expression patterns of only 12 genes were detected using RT-PCR.The results showed thatPbr034785.1,Pbr012328.1,Pbr019138.1andPbr036733.1were either not expressed or only minimally expressed in pollen;whilePbr013082.1andPbr019138.1were expressed in multiple tissues;andPbr012855.1,Pbr013718.1,Pbr012857.1,Pbr001817.1,Pbr027436.1,Pbr011956.1andPbr029836.1were mainly or specifically expressed in pollen tubes (Fig.3).

Fig.3 RT-PCR analysis of the pollen-specific expressed PbrAGPs in six tissues of pear.

We also used RT-qPCR to examine the expression trends of the 38PbrAGPsduring pollen tube growth.Unfortunately,most of thePbrAGPsspecifically expressed in pollen tubes were not detected and their expression levels were not consistent with the RPKM values during pollen tube growth (Appendix F). Therefore,the 12 genes detected using RT-qPCR were revealed,and their expression levels were consistent with RPKM values.The expression levels ofPbr013718.1,Pbr012857.1,Pbr001817.1,Pbr027436.1andPbr011956.1presented down-regulatory trends during pollen tube growth and were consistent with the RPKM-values of RNA-seq. In contrast,thePbr034785.1,Pbr012328.1,Pbr019138.1andPbr036733.1expression levels showed up-regulatory trends during pollen tube growth. The expression patterns of the remaining genes are relatively complex(Fig.4),implying that PbrAGPs may co-regulate pollen tube growth. Based on the above resultsPbr013082.1,Pbr012857.1,Pbr012855.1,andPbr011956.1were subjected to further analyses.

Fig.4 RT-qPCR analysis of the pollen-specific expressed PbrAGPs during pollen tube growth. MP,HP,PT and SPT correspond to the four different developmental stages of matured pollen,hydrated pollen,growing pollen tubes after six hours of hydration and stopped-growing pollen tubes,respectively.

3.4.ldentification and collection of PbrAGP6/11 in pear

AGP6andAGP11were identified inArabidopsisand regulate pollen tube growth. Although we have identified multiplePbrAGPsduring pollen tube growth,the PbrAGP family has many members,and their constructions are relatively complex. Hence,we generated phylogenetic trees to identify the PbrAGPs which could play a role similar to AGP6 and AGP11.

The phylogenetic trees built by the ML and NJ methods showed that AGPs fromArabidopsisthalianawere mainly clustered into three and five branches,respectively. In the ML phylogenetic tree,43 genes from pear were clustered into three branches. Pbr015994.2,Pbr030851.1,Pbr004165.1,Pbr009796.1,Pbr09752.1,Pbr032240.1,Pbr014995.1 and Pbr012520.1 were clustered with AGP6 and AGP11 (Fig.5). On the other hand,the PbrAGPs and AGP6/11 showed clustered scattering in the NJ phylogenetic tree;and Pbr030851.1,Pbr008237.1,Pbr039425.1 and Pbr10124.2 were clustered with AGP6 and AGP11 (Appendix G).Considering that the amino acid structures of PbrAGP are highly diverse and the NJ phylogenetic tree may have long-branch attraction,the clustering result of ML method was used as the main reference.Additionally,Pbr015994.2,Pbr030851.1,Pbr009796.1andPbr09752.1were not expressed in pollen;whilePbr032240.1,Pbr014995.1andPbr012520.1were highly expressed in pollen.Pbr004165.1was mainly expressed at the late stage of pollen tube growth,and the amino acid sequences of Pbr004165.1 and AGP6/11 also contained an N-terminal domain,Proline rich regions and a C-terminal conserved domain (Appendix H). Sequence similarity between Pbr014995.1 and Pbr012520.1 was up to 99%,and we regarded them as the same gene in this study. Based on above results,Pbr032240.1,Pbr014995.1andPbr004165.1were preliminarily regarded as candidate genes. Additionally,we also selected two genes from other two branches to study their functions,includingPbr035082.1,Pbr032913.1,Pbr041476.1andPbr021831.1.

Fig.5 Phylogenetic tree analysis of the PbrAGP family in pear and Arabidopsis thaliana. The three branches are marked by yellow,blue and red,which contain the AGPs from pear and Arabidopsis thaliana. The remaining branches are shown in gray.

3.5.PbrAGPs influence pollen tube growth

We identified 19PbrAGPsfrom the pear genome based on the RT-PCR,RT-qPCR and phylogenetic tree results.Before analyzing their effects on pollen tubes,we analyzed their genetic properties and named them fromPbrAGP1toPbrAGP19(Table 1). They range in length from 216 to 1 332 amino acids,and their MW range is from 6.83 to 47.34.

Table 1 The characteristics of 19 PbrAGP identified from pear1)

To further explore whether the PbrAGP proteins could influence pollen tube growth,we purified the expressed recombinant proteins by prokaryotic expression. Firstly,we used SDS-PAGE to test the quality of the PbrAGPHis recombinant proteins (Appendix I-a). Based on the SDS-PAGE results,10 and 20 μL of recombinant proteins were independently added into the medium for culturing pollen tubes,and both amounts could promote pollen tube growth for several of the proteins,such as PbrAGP1,PbrAGP5 and PbrAGP18 (Appendix I-b). Therefore,we added 20 μL of recombinant proteins into the pollen medium to test their influences on pollen tube growth.The recombinant protein of empty vector pCold-His was added into the CK treatment. The results showed that 11 of the proteins (PbrAGP1,PbrAGP4,PbrAGP5,PbrAGP8,PbrAGP9,PbrAGP10,PbrAGP12,PbrAGP14,PbrAGP16,PbrAGP18 and PbrAGP19) could significantly promote pollen tube growth (Fig.6).

Fig.6 The influences of 19 PbrAGP-His recombinant proteins on ‘Dangshan Suli’ pollen tube growth in vitro. A,the morphological changes of pollen tubes under treatment with the 19 PbrAGPs-His. Bar=100 μm. B,statistical analysis of pollen tube lengths under treatment with the 19 PbrAGP-His proteins. Bars are SD (n=3). Student’s t-test: ＊,P＜0.05;＊＊,P＜0.01.

We also knocked-down the expression levels of these 11PbrAGPusing the ODN method to test their influences on the pollen tube growth. The pollen tubes grew normally under ODN treatment,and the pollen tube lengths were measured (Fig.7-A). The lengths of pollen tubes treated with PbrAGP1-asODN and PbrAGP5-asODN were significantly shorter than CK,but the other PbrAGP-asODN treatments did not influence pollen tube growth (Fig.7-B). The pollen tube lengths under treatments with PbrAGP4-asODN,PbrAGP8-asODN and PbrAGP12-asODN were shorter than CK,and we also analyzed the significant differences between CKand the treatment groups by the Student’st-test (P＜0.05)(Appendix J). However,there were no significant inhibitory effects on pollen tube growth,except for PbrAGP1-asODN and PbrAGP5-asODN. The RT-qPCR results further indicated that pollen tube growth could be significantly inhibited by suppressing thePbrAGP1andPbrAGP5expression levels (Fig.7-C and D). Based on the above results,PbrAGP1 and PbrAGP5 were treated as candidate genes and their mechanism for regulating pollen tube growth was further explored.

Fig.7 PbrAGP can inhibit pollen tube growth under treatment with antisense oligonucleotides (as-ODN). A,the pollen tube images of 11 PbrAGP under treatment with as-ODN. Bar=100 μm. B,statistical analysis of pollen tube lengths. C and D,the expression levels of PbrAGP1 and PbrAGP5 were tested using RT-qPCR under treatments with as-ODN and sense oligonucleotide (s-ODN).Bars are SD (n=3). Student’s t-test: ＊,P＜0.05.

3.6.Sub-cellular location of PbrAGP1 and PbrAGP5

Determining the sub-cellular localization of PbrAGP1 and PbrAGP5 is critical to studying their functions. To find the sub-cellular localization of PbrAGP1 and PbrAGP5,we cloned them into the N-terminal of the GFP reporter protein driven by a CaMV 35S promoter,generating two fusion proteins,PbrAGP1:GFP and PbrAGP5:GFP. The fusion plasmids were transformed separately into tobacco leaf epidermis viaAgrobacterium-mediated transformation.The results showed that the control GFP was uniformly distributed throughout the whole cell (Fig.8-C),whereas the PbrAGP1:GFP and PbrAGP5:GFP fusion proteins were found in the plasma membrane (Fig.8-A and B). To explore the exact locations of PbrAGP1 and PbrAGP5,we also used FM4-64 to mark the plasma membrane,which indicated that PbrAGP1 and PbrAGP5 were both mainly observed in the plasma membrane (Fig.8).

Fig.8 Sub-cellular localization of PbrAGP1 (A),PbrAGP5 (B) and GFP (C). Bar=20 μm.

4.Discussion

A growing number ofAGPshave been identified from diverse plants,including poplar,Arabidopsis,maize,rice,cork oak and pear (Mashiguchiet al.2004;Schindleret al.2010;Showalteret al.2010,2016;Costaet al.2015;Jiaoet al.2018). They play key roles in various physiological processes. In addition,the functions of AGPs in the processes of pollen formation,germination and development have been investigated. A subset ofAGPgenes were identified from pollen grains and pollen tubes (Pereiraet al.2006). Researchers have identified a total of 85ArabidopsisAGP genes,11 of which were expressed in pollen grains and pollen tubes based on their microarray data (Yuanet al.2009;Costaet al.2013);andAGP6,AGP11,AGP23,AGP24andAGP40were highly expressed in pollen. By analyzing the phenotypes ofagp6andagp11mutants,AGP6andAGP11were found to participate in the processes of pollen grain development and pollen tube growth (Levitinet al.2010;Coimbraet al.2010),and are probably involved in the double fertilization process (Pereiraet al.2006). However,the potential functions of PbrAGPs in pear pollen tube growth are still unclear.

A total of 522PbrAGPgenes were identified from the pear genome,and 187 of them were expressed in pear pollen tubes,indicating that PbrAGPs might be involved in multiple processes of pear plant growth and perform different functions. The reason for the relatively larger AGP gene family in pear compared withArabidopsismay be that the genome of the pear is larger than the genome ofArabidopsis. In the phylogenetic tree analysis,the PbrAGP clustering was not consistent between the ML and NJ trees. On one hand,AGP is a class of proline-rich proteins that exhibits high diversity and contains five subfamilies. Different subfamilies of AGPs possess various Hyp-rich motifs. The contents and percentages of Hyp/Pro,Ala,Ser and Thr,as well as the presence of unique motifs,were the criteria used to search for AGPs (Showalteret al.2010;Jiaoet al.2018).On the other hand,the amino acid lengths of PbrAGP members varied from 158 to 6 582 (Appendix D). The sequence similarity was low,and the NJ phylogenetic tree may have long-branch attraction resulting in its inaccuracy. Therefore,these reasons may explain the inconsistency between the two phylogenetic trees. To further explore the biological functions of PbrAGPs,we extracted the recombinant proteins of 19 PbrAGPHis,and found that 11 of the PbrAGP-His recombinant proteins could promote pollen tube growth. PbrAGP1 and PbrAGP5 could significantly promote pollen tube growth,suggesting that they play significant roles in pear pollen tube growth. Additionally,since they are localized in the plasma membrane,PbrAGP1 and PbrAGP5 might be regarded as the receptor(s) that receive the external environmental signals which regulate pear pollen tube growth.

Based on phylogenetic tree analysis,PbrAGP1 and PbrAGP5 did not cluster together with the AGP6 and AGP11. Biased amino acid compositions and conserved motifs of the AGP amino acid sequences can explain this phenomenon. The results of RTqPCR showed that the expressions levels ofPbrAGP1andPbrAGP5increased in the stages of MP,HP and PT,and then decreased in the stage of SPT. Pollen tube elongation is a polar growth process that requires the continuous synthesis of cell wall and membrane materials. AGPs are extracellular glycoproteins that form the amorphous component of the cell wall as structural proteins. Therefore,we used ruthenium red to detect the distribution of pectin under the treatments with PbrAGP1-His and PbrAGP5-His recombinant proteins separately. S7-RNase,which significantly inhibits pear pollen tube growth,was used as the negative control. No significant differences were detected in the distribution of pectin between the control and treatments with PbrAGP1-His or PbrAGP5-His recombinant proteins (Appendix K).These results were also consistent with the plasma membrane localization,suggesting that PbrAGP1 and PbrAGP5 may act as the receptors that receive external environmental signal which regulate pollen tube growth.

5.Conclusion

In this study,we identified 187PbrAGPs that were expressed in pear pollen tube. Of these,38PbrAGPswere found to be specifically expressed in the pear pollen tube by screening from multiple RNA-seq data.ThePbrAGPgenes were classified into two groups,those highly expressed and specifically expressed during pollen tube growth,to further verify their expression patterns using RT-PCR and RT-qPCR.Ultimately,19PbrAGPswere identified from pear,which might influence pollen tube growth. Among them,11 PbrAGP-His recombinant proteins could promote pollen tube growthinvitro. Additionally,pollen tube growth was inhibited when the expression levels ofPbrAGP1andPbrAGP5were suppressed by as-ODN.Both PbrAGP1 and PbrAGP5 were localized in plasma membrane,and they seem to have no obvious effects on pollen tube assembly. These findings suggest that the regulatory mechanisms of PbrAGP1 and PbrAGP5,which might be regarded as receptors that receive the signals of the external environment which regulate pollen tube growth,are complicated and need further exploration.

Acknowledgements

This research was supported by the earmarked fund for China Agriculture Research System (CARS-28-37),the Agricultural Science and Technology Innovation Project of Shandong Academy of Agricultural Sciences,China(CXGC2022E21),the Youth Foundation of Shandong Institute of Pomology,China (GSS2022QN11) and the Natural Science Foundation of Shandong Province,China(ZR2019BC075,ZR2020MC141,and ZR2021MC177).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm