Characterization of soybean chitinase genes induced by rhizobacteria involved in the defense against Fusarium oxysporum

Rhizobacteria are capable of inducing defense responses via the expression of pathogenesis-related proteins (PR-proteins) such as chitinases, and many studies have validated the functions of plant chitinases in defense responses. Soybean (Glycine max) is an economically important crop worldwide, but the functional validation of soybean chitinase in defense responses remains limited. In this study, genome-wide characterization of soybean chitinases was conducted, and the defense contribution of three chitinases (GmChi01, GmChi02, or GmChi16) was validated in Arabidopsis transgenic lines against the soil-borne pathogen Fusarium oxysporum. Compared to the Arabidopsis Col-0 and empty vector controls, the transgenic lines with GmChi02 or GmChi16 exhibited fewer chlorosis symptoms and wilting. While GmChi02 and GmChi16 enhanced defense to F. oxysporum, GmChi02 was the only one significantly induced by Burkholderia ambifaria. The observation indicated that plant chitinases may be induced by different rhizobacteria for defense responses. The survey of 37 soybean chitinase gene expressions in response to six rhizobacteria observed diverse inducibility, where only 10 genes were significantly upregulated by at least one rhizobacterium and 9 genes did not respond to any of the rhizobacteria. Motif analysis on soybean promoters further identified not only consensus but also rhizobacterium-specific transcription factor-binding sites for the inducible chitinase genes. Collectively, these results confirmed the involvement of GmChi02 and GmChi16 in defense enhancement and highlighted the diverse inducibility of 37 soybean chitinases encountering F. oxysporum and six rhizobacteria.


Introduction
The composition of soil microbial communities varies depending on numerous factors, including the plant species or variety present.Plants release root exudates, which can attract beneficial microbes moving from bulk soil into the rhizosphere (Lugtenberg and Kamilova, 2009;Martinez-Medina et al., 2016;Lundberg and Teixeira, 2018).Many rhizobacteria, such as plant growth-promoting rhizobacteria (PGPR), can stimulate plant growth, and these rhizobacteria may enhance plant health by interacting directly or indirectly with soil-borne pathogens.The interactions can be generalized into three types: antagonism (Elnahal et al., 2022), parasitism (Tian et al., 2007), and induced systemic resistance (ISR) (Zhu et al., 2022).In terms of antagonism, rhizobacteria may secrete antibiotics or siderophores to antagonize or compete for nutrients with pathogens.Rhizobacteria such as Bacillus amyloliquefaciens and Streptomyces sp.exhibit these capabilities, and some strains have been developed into commercial products (Boubekri et al., 2022;Luo et al., 2022).Regarding parasitism, rhizobacteria may secrete enzymes such as chitinases that degrade fungal cell walls.Rhizobacteria such as Burkholderia ambifaria and Enterobacter sp. also possess predatory behaviors on fungi, thereby reducing fungal pathogens in the rhizosphere (Mousa et al., 2016;Stopnisek et al., 2016;Chang et al., 2021).Furthermore, due to the presence of microbeassociated molecular patterns (MAMPs) in most rhizobacteria, such as Pseudomonas fluorescens (Orozco-Mosqueda et al., 2023), pattern-triggered immunity (PTI) and ISR can be activated in the absence of soil-borne pathogens, leading to a phenomenon known as defense priming (Mauch-Mani et al., 2017;Salwan et al., 2023).Accordingly, the diverse mechanisms and interactions between plants and the rhizosphere microbes together contribute to the overall plant health and support agricultural sustainability.
The expression of pathogenesis-related protein (PR-protein) genes is one of the important responses in defense responses to combat pathogens, and many PR-protein genes have been confirmed to enhance defense responses in various mechanisms (Huang et al., 2016;Su et al., 2016;Luo et al., 2023;Wang Y et al., 2021).Among 17 families of PR-proteins (van Loon et al., 2006), the PR-3, PR-4, PR-8, and PR-11 proteins all encode plant chitinases, which contain the glycosyl hydrolases (GH) domain capable of breaking the b-1,4glycosidic linkages of chitin, leading to disruption of fungal cell walls (Vaghela et al., 2022).These chitinases can be classified into GH18 and GH19 based on the similarity of their catalytic domains (CatD) (Kawase et al., 2004;Funkhouser and Aronson, 2007).The GH18 chitinases exhibit a barrel-like structure consisting of eight a-helices and eight b-sheets (Yang et al., 2010), and the GH19 chitinases possess a lysozyme-like domain composed of several a-helices (Kezuka et al., 2006).Plant chitinases have been further grouped into five classes (classes I-V) according to characteristics such as Nterminal sequences.While classes III and V belong to GH18, classes I, II, and IV belong to GH19 (Grover, 2012).Class I chitinases possess a chitin-binding domain (CBD) in their N-terminal region (Tang et al., 2004), and the C-terminal region of class I chitinases contains seven extended amino acids that facilitate their targeting to vacuolar and intracellular transport (Vaghela et al., 2022).Class II chitinases lack CBD in the N-terminal, and they are typically acidic proteins induced by pathogen infection and secreted to the extracellular space (Patil et al., 2000).Class III chitinases exhibit lysozyme activity without sequence similarity to classes I and II chitinases (Xu et al., 2016;Ma et al., 2017).Class IV chitinases possess both CBD and CatD similar to class I; however, due to deletions in the CBD and CatD domains, class IV chitinases are usually smaller than class I (Xu et al., 2016).Class V chitinases have a C-terminal extension for vacuolar targeting and may contain CBD (Taira et al., 2009;Grover, 2012).Accordingly, plant chitinases have evolved with diverse domains and variations.
The importance of plant chitinases in defense responses has been studied in several cases.For example, it was shown that 11 chitinase genes of rice were upregulated by R. solani infection.These rice chitinases were secreted into extracellular spaces, resulting in the degradation of fungal cell walls (Richa et al., 2016).Overexpression of the rice chitinase gene LOC_Os03g30470 enhanced defense against Botrytis cinerea and Diplocarpon rosae (Marchant et al., 1998;Nuñez de Caćeres Gonzaĺez et al., 2015).Overexpression of another rice chitinase gene, LOC_Os05g33130, increased defense responses to many diseases, and many studies have also demonstrated the defense contribution of various plant chitinases in different plant systems (Table 1).Other than the overexpression approach, gene silencing of the chili pepper chitinase gene CaChiIII7 resulted in larger foliar symptoms, less ROS accumulation in leaves, and reduced expression of defense-related genes (Ali et al., 2020).Collectively, the importance of plant chitinases in defense responses has been confirmed through overexpression and silencing approaches in different plant systems.
The advancement of high-throughput sequencing technology in the past decade has completed about 800 plant genomes (Marks et al., 2021;Sun et al., 2022), which speeded up the genome-wide characterization of plant chitinases in apple, Arabidopsis thaliana, Brassica rapa, cotton, cucumber, mulberry, rice, and tea (Grover, 2012;Xu et al., 2016;Chen et al., 2018;Bartholomew et al., 2019;Mir et al., 2020;Bordoloi et al., 2021;Haxim et al., 2022;Xin et al., 2022).For example, seven of the 24 chitinase genes discovered in the A. thaliana genome, such as AT1G19810, AT2G43570, AT2G43580, AT2G4359, and AT3G47540, were found to be upregulated upon infection by B. cinerea and Pseudomonas syringae.In the case of rice, 49 chitinase genes were characterized, and transcriptome analysis identified Os01g18400, Os01g19750, Os10g28050, and Os11g47510 being upregulated in response to Magnaporthe grisea infection (Grover, 2012).In the genomes of B. rapa and tea, 33 and 49 chitinase genes were discovered, respectively.Upregulation of several chitinase genes was also found during infections of the clubroot pathogen and three tea pathogens (Chen et al., 2018;Bordoloi et al., 2021).These findings demonstrate the power of high-throughput sequencing in genomewide identification of plant chitinases, which also enable investigations for their expressions to different microbes.
Soybean (Glycine max) is one of the most important crops worldwide, and soybean diseases have been one of the major yieldlimiting stresses for decades (Bandara et al., 2020;Bradley et al., 2021;Allen et al., 2022).However, there were limited studies on the functional validation of soybean chitinases (Lv et al., 2022) and their expressions induced by rhizobacteria.Therefore, this study

Results
Genome-wide identification of soybean chitinase genes A search of the GH18 (PF00704) and GH19 (PF00182) domains identified 37 chitinase genes in the soybean genome, and all of them were predicted with an N-terminal signal peptide.Following the classification system of Arabidopsis chitinase genes, the soybean chitinase genes can be further divided into five genes in class I, four genes in class II, nineteen genes in class III, three genes in class IV, and six genes in class V (Figure 1).Using MEME analyses for characterizing motifs, the conserved GH18 motifs were identified in classes III and V, and the GH19 motifs were identified in classes I, II, and IV.As reported in the previous literature (Ma et al., 2017), class III chitinases harbor both GH18 motifs and GH19 lysozyme domain (motifs 3 and 8) as a classification signature.These results together confirmed that the HMMER method in genome-wide identification of soybean chitinase genes is robust and precise.
One controversial classification appeared for Glyma.01G160100,Glyma.02G007400, and Glyma.19G221800,which should be classified as class II for the absence of CBD (motif 15) if they followed the conventional classification for Arabidopsis chitinase genes (Patil et al., 2000).Although the presence or absence of CBD has been used to identify class I or IV chitinases (Patil et al., 2000;Grover, 2012;Xu et al., 2016), the MEME analyses found exceptions not only in soybean but also in Arabidopsis.For example, AT3G47540 was recognized as class IV chitinase because CBD was absent.In addition, although AT1G02360 and AT4G01700 were grouped as class II chitinases based on the absence of CBD, both phylogenetic and MEME analyses suggested that their sequences and motif structures were closer to the class I chitinases.Therefore, this study suggests a phylogeny-based classification for soybean chitinase genes (Figure 1), where classes II, III, and V chitinase genes do not contain CBD motif, Genome-wide identification of soybean chitinase genes.The conserved HMM domains of the glycosyl hydrolase (GH) GH18 (PF00704) and GH19 (PF00182) chitinases were used as templates to identify soybean chitinases in the soybean reference genome.The 37 putative chitinases were further grouped using the neighbor-joining method into five classes according to the classification system built on Arabidopsis thaliana chitinases.There were 5, 4, 19, 3, and 6 chitinase genes in classes I, II, III, IV, and V, respectively, and there were 20 conserved motifs detected in the soybean chitinases.Bootstrap values above 60 were shown in the nodes.Chen et al. 10.3389/fpls.2024.1341181Frontiers in Plant Science frontiersin.orgwhile classes I and IV chitinase genes may contain CBD motif.

Gene expression of soybean chitinases in a tritrophic RNA-Seq experiment
Among the soybean chitinase genes, 12 genes exhibited upregulation in response to inoculation with F. oxysporum ( Figure 2A).The top 5 upregulated genes included Glyma.13G346700,Glyma.12G156600,Glyma.11G124500,Glyma.02G007400, and Glyma.02G024500,which displayed a log 2 fold change of 7. 27, 7.23, 6.44, 5.89, and 4.97, respectively.Meanwhile, three among these five genes (Glyma.13G346700,Glyma.11G124500, and Glyma.02G024500) were also upregulated by the inoculation of rhizobacterium B. ambifaria in the absence of F. oxysporum (Table 2).In addition, Glyma.16G173000 and Glyma.09G126200 were also induced by B. ambifaria, but the upregulation of these two genes by F. oxysporum was not as high as the others.These results highlighted that these five soybean chitinase genes (Glyma.02G024500,Glyma.09G126200,Glyma.11G124500,Glyma.13G346700, and Glyma.16G173000)participated in the defense responses induced by B. ambifaria, and three of the five genes were listed in the top 5 important chitinase genes in the defense responses to F. oxysporum infection (Figure 2B).Indeed, upon the co-inoculation of F. oxysporum and B. ambifaria, for which the biomass of F. oxysporum was reduced by the antagonism of B. ambifaria (Chang et al., 2021), the upregulation of Glyma.13G346700,Glyma.11G124500, and Glyma.02G024500 was about 20% to 33% reduced compared to inoculation with F. oxysporum alone (Table 2).Collectively, Glyma.13G346700,Glyma.11G124500, and Glyma.02G024500became the research focus not only for their inducibility but also for their expression trends reflecting the biotic stress created by the inoculation of F. oxysporum.

Phylogenetic analysis for soybean chitinases in defense responses
In order to assess the potentials of Glyma.13G346700,Glyma.11G124500, and Glyma.02G024500 in defense responses, a phylogenetic analysis for the 37 soybean chitinase genes was performed together with functionally validated plant chitinases from the literature (Table 1).For GH 18 chitinases, there were only three functionally validated plant chitinases, including cucumber, Chinese wild strawberry, and rice that can enhance Soybean chitinase genes in response to the inoculation of Burkholderia ambifaria, Fusarium oxysporum, and co-inoculation.(A) Venn diagram suggests 12 chitinases being upregulated in response to F. oxysporum inoculation, while 7 chitinase genes were significantly induced by B. ambifaria in the absence of F. oxysporum.There were 5 consensus responding to the inoculation of B. ambifaria, F. oxysporum, and co-inoculation.(B) Gene expression of the 37 soybean chitinases, and the asterisk indicated the statistical significance of differential expression at a q-value of 0.05.
defense responses to B. cinerea, Colletotrichum higginsianum, and R. solani, respectively (Kishimoto et al., 2004;Richa et al., 2017;Wen et al., 2020) (Figure 3A).There were more studies that confirmed the function of GH19 plant chitinases, and 27 functionally validated plant chitinases from apple, barley, cocoa, common bean, cucumber, maize, pepper, rice, and wheat were included in the phylogenetic analysis together with 12 GH19 soybean chitinases.These 12 soybean genes can be categorized into three groups (Figure 3B).The first group contains three soybean chitinase genes (Glyma.11G124500,Glyma.12G049200, and Glyma.13G346700), and the closest ortholog gene is the pepper CaChitIV, which enhances defense responses to Arabidopsis downy mildew (Kim et al., 2015).The second group includes four soybean chitinase genes, but only a strawberry gene, FvChi-14, which enhances Arabidopsis defense responses to C. higginsianum, is phylogenetically neighboring to these four genes (He et al., 2023).The last group with Glyma.01G160100,Glyma.02G024500, and Glyma.16G119200clustered with apple, barley, bitter melon, cucumber, wild tomato, and zoysiagrass that were previously shown to contribute to the defense responses against multiple fungal pathogens, including many soil-borne fungi such as F. oxysporum, R. solani, and Verticillium dahliae (Table 1).In addition, two chitinase genes (Glyma.02G007400and Glyma19G221800) were in the same clade but located distantly from the three soybean chitinase genes abovementioned.Among these genes, Glyma.02G024500 is the one that responded to the inoculation of F. oxysporum and B. ambifaria.On the other hand, t h e p h y l o g e n e t i c a l l y c l o s e d G l y m a .0 1 G 1 6 0 1 0 0 a n d Glyma.16G119200 did not seem to participate in the defense responses at least to F. oxysporum, nor be induced by B. ambifaria.The observation raised a question whether Glyma.01G160100,Glyma.02G024500, and Glyma.16G119200(hereafter referred to a GmChi01, GmChi02, and GmChi16) all contain antifungal capability or if only Glyma.02G024500 remains antifungal.Accordingly, GmChi01, GmChi02, and GmChi16 were selected for functional validation.

Functional validation for soybean chitinases in defense responses
The homozygous transgenic Arabidopsis lines overexpressing empty vector (EV), GmChi01, GmChi02, or GmChi16, respectively, were validated for the expression of soybean chitinases (Supplementary Figure S1) before using their T 4 generation for experiments.For seedling root length, rosette leaves, and plant height, the transgenic lines (EV_6-8, GmChi01_6-8, GmChi01_7-1, GmChi02_3-7, GmChi02_6-3, GmChi16_4-3, and GmChi16_7-1) exhibited no difference in phenotypes ( Figures 4A-C).However, the area under the disease progress curve (AUDPC) of these lines inoculated with F. oxysporum exhibited significant differences.The AUDPC of transgenic Arabidopsis overexpressing GmChi02 and GmChi16 were significantly lower than the controls (Figures 4D, E).Although transgenic Arabidopsis overexpressing GmChi01 exhibited reduced AUDPC in appearance, statistical analysis did not detect a significant difference.Meanwhile, identical results can be observed in soil inoculation with the conidial suspension of F. oxysporum, where transgenic Arabidopsis overexpressing GmChi02 and GmChi16 exhibited less seedling wilt.Transgenic Arabidopsis overexpressing GmChi01 again showed better survival in appearance, but the statistical analysis did not detect any significance (Figures 4F, G).These results indicate that GmChi02  and GmChi16 were indeed phylogenetically and functionally close, and these two soybean chitinases enhanced defense responses to F. oxysporum infection.However, the gene regulation of GmChi01, GmChi02, and GmChi16 appeared to be diversified, and only GmChi02 exhibited inducibility in response to B. ambifaria.
Specifically, GmChi02 can be significantly induced by B. diazoefficiens, B. ambifaria, and L. enzymogenes.On the other hand, GmChi01 or GmChi16 did not reach statistical significance for any rhizobacteria.As for other chitinase genes such as Glyma.13G346700,Glyma.12G156600,Glyma.11G124500, and Glyma.02G007400 that were induced by F. oxysporum infection (Figure 2B), Glyma.13G346700 and Glyma.02G007400can be significantly induced by B. ambifaria and P. fluorescens.On the other hand, Glyma.12G156600 was upregulated by B. ambifaria, while Glyma.11G124500 was upregulated by P. fluorescens.These results suggest that soybean chitinase genes upregulated in the defense responses to F. oxysporum infection all interacted with at least one of the six rhizobacteria.Therefore, transcription factorbinding sites may have emerged during the co-evolution between soybeans and these rhizobacteria.

Identification of transcription factor and transcription factor-binding sites for the rhizobacteria-inducible soybean chitinase genes
In order to identify the potential regulatory motifs, the 5′ UTR and 3′ UTR of soybean chitinase genes that responded to the six rhizobacteria were analyzed.There were 94, 62, 76, 59, 125, and 90 soybean transcription factor-binding sites (TFBSs) associated with transcription factors (TFs) for soybean chitinase genes induced by B. amyloliquefaciens, B. japonicum, B. ambifaria, L. enzymogenes, P. fluorescens, and R. rhizogenes, respectively (Figure 6A).Among Gene expression of soybean chitinases in response to six rhizobacteria.Asterisks highlight the significant upregulation induced by the inoculation of each rhizobacterium.GmChi02 responded to all rhizobacteria, and the inoculation of B. japonicum, B. ambifaria, and L. enzymogenes reached statistical significance.GmChi01 only responded to P. fluorescens without statistical significance, and GmChi16 was upregulated by L. enzymogenes and R. rhizogenes without statistical significance.
these genes, there were 55 TFBSs associated with TFs consensually identified for all rhizobacteria, while there were only 11, zero, three, zero, 66, and six TFs exhibiting a specificity to B. amyloliquefaciens, B. japonicum, B. ambifaria, L. enzymogenes, P. fluorescens, and R. rhizogenes, respectively.Focusing on the TF enriched for interacting with B. ambifaria, there were three unique TFs, including two h o m e o b o x d o m a i n T F s ( G l y m a .0 1 G 2 4 0 1 0 0 a n d Glyma.07G076800) and one SQUAMOSA promoter-binding protein (SBP)-box TF (Glyma03g29900) (Figures 6B-E).
As for TFBS without associated TFs, only one cis-regulatory element, NODCON1GM1, was found for B. amyloliquefaciens, B. japonicum, B. ambifaria, L. enzymogenes, and R. rhizogenes in contrast to the inducible and noninducible chitinase genes (Figure 6F).Based on the consensus and diversified inducibility of soybean chitinase genes, the results indicated that the regulatory mechanism of chitinase genes may have co-evolved with soybeanrhizobacteria interaction.

Discussion
The benefits of rhizobacteria for plant health have been greatly recognized in different aspects.Other than direct antagonism against soil-borne pathogens, rhizobacteria may stimulate defense responses to provide sustainable plant protection.However, some scholars have pointed out that crops grown in fields may already be in a constant defense priming and/or induced systemic resistance (ISR) state because they are persistently interacting with rhizobacteria, and simply applying or exposure to rhizobacteria may not be sufficient to combat pathogens (Pieterse et al., 2014).Other literature indicates the density of rhizobacteria is a crucial key for defense priming, or ISR.For example, the minimum bacterial density required for P. fluorescens alone to induce ISR in laboratory conditions required 10 5 (Raaijmakers, 1995), and it may need to be higher in field conditions.Therefore, it could be challenging to achieve a sufficient population density of rhizobacteria for defense priming or ISR throughout the entire growing season (Walters et al., 2013).An alternative strategy is molecular breeding for important defense genes to bypass the reliance on bacterial density.For example, Rushanaedy et al. (2012) found the koa tree chitinase genes AKchit1a and AKchit1b were significantly upregulated in the resistant cultivars against F. oxysporum compared to susceptible cultivars, providing a screening criterion for disease resistance.Another application of plant chitinase genes is the early detection of biotic stresses.For example, the chitinase activity of papaya fruits was significantly increased upon Colletotrichum gloeosporioides infection during both the preharvest and storage stages.Since C. gloeosporioides is a pathogen exhibiting a latent infection stage, the expression of chitinase genes may serve as diagnostic biomarkers for asymptomatic fruits (Lucas-Bautista et al., 2020).Regardless of being selected as breeding targets or diagnosis biomarkers, the characterization of plant chitinase genes can provide novel insights and a comprehensive understanding of defense responses for a plant species.
This study performed genome-and transcriptome-wide identifications of soybean chitinases and functionally validated three phylogenetically close-related genes (GmChi01, GmChi02, and GmChi16) for their involvement in defense responses.The results showed that GmChi02 and GmChi16 enhanced defense responses to F. oxysporum, but only GmChi02 can be induced by B. ambifaria.In the transcriptomic characterization of GmChi02 in different rhizobacteria, the results confirmed a significant upregulation by B. ambifaria and P. fluorescens, and the expression of GmChi02 also responded to the inoculation of B. amyloliquefaciens, B. japonicum, L. enzymogenes, as well as R. rhizogenes.The observation indicates that GmChi02 may have coevolved with multiple rhizobacteria to induce defense responses against soil-borne pathogens.On the other hand, although GmChi16 exhibited an equivalent defense effect as GmChi02, the expression of GmChi16 only responded to the inoculation of L. enzymogenes and R. rhizogenes.Collectively, these observations indicate that the regulatory mechanism of soybean chitinase genes may have diversity not only in the coding sequence level for functionality but also in the expression level in terms of inducibility.
Several TFs and TFBSs have been shown to regulate plant chitinase expression.For example, the homeodomain leucine zipper III TF CsHB15 of cucumber was found to bind the promoter of CsChi23 and induce gene expression in response to F. oxysporum (Bartholomew et al., 2022).The R2R3-MYB TF of brown mustard was shown to recognize the W-box-like-4 (Wbl-4) element to activate BjCHI1 in response to B. cinerea (Gao et al., 2016).Another example is the LrWRKY2 of lily, which induced LrCHI2 expression in response to F. oxysporum (Li et al., 2021).However, whether plant chitinase genes harbor conserved TF and TFBS in response to rhizobacteria has not been assessed.In this study, several TF genes and motifs were highlighted by contrasting the TFs and TFBS motifs between the inducible and non-inducible soybean chitinase genes in each rhizobacterium (Supplementary Table S2).One with particular interest would be the NODCON1GM, which has been known to be a nodule-specific regulatory element (Wang et al., 2022).Mutation and deletion of NODCON1GM (5′-AAAGAT) or another regulatory element, NODCON2GM (5′-CTCTT), were shown to decrease the number of nodule formations (Jørgensen et al., 1991).The presence of NODCON1GM in the promoters of rhizobacteria-inducible chitinase genes suggests a possibility that the regulatory mechanism to drive chitinase genes may rely on a similar manner as the regulatory element NODCON1GM.Additional studies on the emergence of NODCON1GM in some but not all soybean chitinase genes, along with studies on the presence of NODCON1GM in the promoters of other soybean PR-protein genes, may further illuminate the evolution of rhizobacteriainduced defense responses.
Research has shown that soybean has diverse rhizobacteria, including Bradyrhizobium, Bacillus, Burkholderia, and Rhizobium species (Biate et al., 2014;Zhong et al., 2019;Han et al., 2020;Yamazaki et al., 2021), that could affect soybean yield and disease incidence (Chang et al., 2017;Hussain et al., 2018).Future studies may focus on the selection pressure derived from soybean rhizobacteria on the expressions of PR-protein genes and the regulatory mechanisms of defense responses induced by different rhizobacteria.These research advances may provide a broad knowledge of the application of beneficial rhizobacteria to enhance plant health.

Plant and microbial materials
For routine cultivation of Arabidopsis thaliana, the seeds were surface-sterilized using 70% ethanol and 50% Clorox bleach (Oakland, CA, USA).After rinsing five times with sterile water, the seeds were placed in the dark at 4°C for 48 h for vernalization.Subsequently, the seeds were planted in a soil mixture (peat moss: vermiculite:perlite = 6:1:1) and cultured in a long-day condition (16 h light/8 h dark) at 22°C.

Genome-wide identification and phylogenetic analysis of soybean chitinases
To identify chitinase genes in the soybean genome, the HMMs of the GH18 (PF00704) and GH19 (PF00182) protein domains were downloaded from the Pfam database (Mistry et al., 2021).Subsequently, HMMER v3.3.2 was applied to search PF00704 and PF00182 in the 'Williams 82' (W82) (Gmax_508_Wm82.a4.v1.protein) at a threshold of 1 −10 E-value (Finn et al., 2011).The presence of GH18 or GH19 domain was double-checked using the NCBI Conserved Domain Database at a threshold of 1 −20 E-value.In addition, protein tertiary structure was assessed by predicting the folded structure of each soybean chitinase gene protein sequence using ColabFold (Mirdita et al., 2022).Furthermore, MEME v5.4.1 was utilized at a setting of a maximum motif length of 300 and a number of motifs of 20 to identify conserved motifs within the protein sequences (Bailey and Elkan, 1994).The Protparam (Gasteiger et al., 2005), SignalP5.0(Almagro Armenteros et al., 2019), andDeepLoc-1.0 (Almagro Armenteros et al., 2017) webtools were employed to investigate the amino acid composition, molecular weight, and isoelectric point of soybean chitinase proteins.
The protein sequences of soybean chitinases were aligned with 24 Arabidopsis thaliana chitinases sourced from the TAIR database (Lamesch et al., 2012).Alignment was performed using MAFFT v7 (Katoh et al., 2019), and the phylogenetic tree was constructed using the neighbor-joining (NJ) method in MEGA-X (Kumar et al., 2018).Additionally, the protein sequences of soybean chitinases were aligned with functionally validated chitinase sequences from 21 plant species (Table 1).The phylogenetic tree was constructed using the maximum likelihood (ML) method in IQ-TREE v2.2.0 (Nguyen et al., 2015).The visualization of the phylogenetic trees was generated using iTOL (Letunic and Bork, 2021).

Transcriptomic analysis of soybean chitinases
The tritrophic RNA-Seq data were obtained from a previous study on the gene expression of F. oxysporum in the roots of the soybean variety 'Jack' under the influence of the antagonistic bacterium B. ambifaria (Chang et al., 2021).The data can be categorized into four treatments: (1) soybean roots without B. ambifaria or F. oxysporum, (2) soybean roots inoculated with B. ambifaria, (3) soybean roots inoculated with F. oxysporum, and (4) soybean roots simultaneously inoculated with both B. ambifaria and F. oxysporum.Each treatment consisted of three biological replicates, totaling 12 samples.The RNA-Seq was performed using the Illumina HiSeq 4000 platform (Illumina, San Diego, CA, USA).The raw data underwent quality control to keep reads with a Phred score ≥ 30 using the FASTQC and FASTX-ToolKit v0.0.14.The soybean W82 transcriptome (Gmax_508_Wm82.a4.v1.cds.fa) was used as a template for Kallisto v0.46.1 (Bray et al., 2016).Subsequently, differential gene expression analysis was conducted using the R package Sleuth v0.30 (Pimentel et al., 2017) at a threshold of 0.05 q-value.Transcript per million (TPM) measurements of the 37 soybean chitinase genes were presented in a heatmap using the R package ComplexHeatmap v2.13.1 (Gu et al., 2016).
In the RNA-Seq experiment of soybean root inoculated by six rhizobacteria, the W82 soybean seeds were sterilized in 1% bleach for 15 min, followed by five rinses with sterile water.The sterilized seeds were vernalized in sterile water at 28°C without light to better synchronize the germination rate.The next day, the seed coats were removed, and the seeds were placed on 1.5% water agar plates in a growth chamber at 28°C without light for 3 days.After the seeds germinated and the hypocotyls elongated to approximately 3 cm to 5 cm, the seedlings were transferred to new water agar (WA) plates, where 100 µl (approximately 1 × 10 7 CFU/ml) of bacterial suspension was inoculated onto the soybean hypocotyls.The control group was inoculated with ddH 2 O.
The inoculated soybean seedlings were further incubated in a growth chamber at 28°C without light.After incubating for 2 days, the frozen taproot samples were homogenized in liquid nitrogen with Invitrogen ™ TRIzol ™ Reagent (Thermo Fisher Scientific, Waltham, MA, USA), followed by the extraction workflow using chloroform and isopropanol.With two biological replicates per rhizobacteria and control, a total of 14 samples were sent to RNA-Seq using the Illumina NovaSeq 6000 platform in a 150-bp pairended platform (Biotools, New Taipei City, Taiwan).

Generation of Arabidopsis transgenic lines using Agrobacterium floral dipping
The Agrobacterium tumefaciens GV3101 was cultured in the 523 liquid medium (8 g/L casein hydrolysate, 2 g/L K 2 HPO 4 , 0.3 g/L MgSO 4 •7H 2 O, 10 g/L sucrose, 4 g/L yeast extract; pH 6.9) supplemented with rifampicin (50 mg/L) and streptomycin (100 mg/L) at 125 rpm shaking for 24 h at 28°C.Upon the optical density (OD600) reaching 1.0 to 1.5, the Agrobacterium suspension was centrifuged at 4,500 rpm at 4°C.The bacterial pellet was resuspended in 20 mM CaCl 2 as competent cells.Three soybean chitinase constructs and an empty vector were individually transformed into A. tumefaciens GV3101 using the freeze-thaw method, including a 30-s liquid nitrogen immersion and a 37°C water bath for 5 min.The transformed bacterial cells were selected by kanamycin (50 mg/L).Colony PCR, using gene-specific primer pairs, was used for validation.The Agrobacterium strains were stored in 523/Kan+/Rif+/Strep+ medium with 50% glycerol at −80°C.
For Agrobacterium floral dipping, the desired Agrobacterium strains were freshly prepared in the 523/Kan+/Rif+/Strep+ medium, and the bacterial pellets were resuspended in a 5% sucrose solution containing 0.02% Silwet L-77 (PhytoTech Lab, Lenexa, KS, USA) to OD600 = 0.6 as inoculum.The floral dipping procedure followed the protocol by Zhang et al. (2006) with slight modifications; in brief, the siliques and pollinated flowers were removed from 6week-old A. thaliana ecotype Col-0, and the unopened Arabidopsis inflorescences were immersed in the Agrobacterium inoculum for 20 s.After immersion, the plants were kept in humid chambers before being routinely cultured at 22°C.The Arabidopsis seeds harvested after floral dipping represented the T 1 generation.The T 1 seeds were selected on the MS medium containing 40 ppm hygromycin.The T 1 plants with hygromycin resistance were further PCR-confirmed before generating the T 2 seeds.The T 2 seeds were selected on hygromycin to estimate the Mendelian segregation (3:1) for each T 1 lineage.T 1 lineages with a single T-DNA insertion were propagated into the T 3 generation.Approximately 100 T 3 seeds of each lineage were screened on hygromycin.If the T 3 germination rate was approximately 100%, the lineage was considered to be homozygous.On the other hand, if the germination rate was around 75%, the lineage was considered to be heterozygous at the T 2 generation.Phenotyping and pathogenicity assay were only performed using the progenies of homozygous T 2 lineages (Supplementary Figure S1).
For the Arabidopsis transgenic lines, the expressions of soybean chitinase (GmChi01, GmChi02, or GmChi16) were confirmed via RT-qPCR.In brief, foliar RNA of transgenic lines was extracted by the TRIzol procedure described above.The raw RNA was treated with the TURBO DNase (Thermo Fisher Scientific) before cDNA synthesis using the SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific) and oligo d(T)18 primer (Bioman).RT-qPCR was performed using the iQ ™ SYBR ® Green Supermix (Bio-RAD, Hercules, CA, USA) with the primers (GmChi01_qPCR; GmChi02_qPCR; GmChi16_qPCR; AtACT7) (Supplementary Table S1) on the CFX ConnectTM Real-Time PCR Detection System (Bio-RAD).Three-step thermocycling conditions were set: initial denaturation at 95°C for 5 min, 40 cycles of denaturation at 95°C for 15 s, annealing at 62°C for 10 s, and extension at 72°C for 10 s.The gene expression was presented using the formula DCt = Ct target gene − Ct AtACT7 .The melty curve of each RT-qPCR amplicon was assessed to confirm specificity, and the amplification efficiencies of primers were optimized to ensure the use of 2 −DCt (Livak and Schmittgen, 2001).

Phenotyping and pathogenicity assay on Arabidopsis transgenic lines
The hypocotyl length, radical length, rosette area, and stem length were measured for the wild-type A. thaliana Col-0 and the Arabidopsis transgenic lines.Hypocotyl and radical lengths were measured after 1 week of growth on MS medium, while rosette area was calculated using the software Easy Leaf Area (Easlon and Bloom, 2014) after another 3 weeks in pots.Stem length measurements were conducted at the 6-week growth stage.The experiments were repeated twice, and there were 15 biological replicates each time.These data were collected for statistical analyses.
The detached leaf assay was applied to evaluate the defense responses of Arabidopsis lines.A 5-mm-diameter PDA plug with the mycelial edge of F. oxysporum f.sp.rapae was inoculated onto Arabidopsis leaves with a needle wound on the leaf surface.The inoculated leaves were grown for 4 weeks.An ordinal disease index (DI) was measured daily for 1 week, for which the index at 0, 1, 2, 3, 4, and 5 indicates 0%, 1%-10%, 11%-25%, 25%-50%, 51%-75%, and 76%-100% of leaf yellowing, and index at 6 indicates a complete wilt and dead leaf (Supplementary Figure S2).The area under the disease progress curve (AUDPC) was calculated (Sparks et al., 2008).The experiments were repeated three times, and there were nine biological replicates each time.
In addition, soil inoculation was performed by spreading the conidial suspension of F. oxysporum f. sp.rapae onto the 1-weekold Arabidopsis lines.After inoculation, the pots were covered with plastic lids to maintain humidity and placed in the greenhouse at room temperature (25°C ± 2°C).The plastic lids were removed after 10 days postinoculation.The experiments were repeated three times, and there were four biological replicates each time.These data were collected for statistical analyses.

Statistical analysis for phenotypic data
The R v4.0.5 environment and RStudio v1.4.17 were used for statistical analyses.All data were analyzed using the nonparametric Kruskal-Wallis rank sum test, and Dunn's test was applied for mean separation at a threshold of a = 0.05.

Identification of TF and TFBS for the rhizobacteria-inducible soybean chitinase genes
Soybean chitinase genes were grouped into two categories, including rhizobacteria-inducible chitinase genes (regardless of upor downregulation) and nonrhizobacteria-inducible chitinase genes (Supplementary Table S2).The upstream 2,000 bp 5′UTR and downstream 500 bp 3′UTR of these genes were subjected to PlantPAN3.0analysis (Chow et al., 2019) using soybean as the model plant for searching TF and TFBS at 90% frequency of support.

FIGURE 3
FIGURE 3 Phylogenetic analysis of soybean chitinase genes with functionally validated plant chitinases.(A) Soybean GH18 chitinases were analyzed with 3 functionally validated chitinases from Chinese wild strawberries, cucumber, and rice.(B) Soybean GH19 chitinases were analyzed with 27 functionally validated chitinases from multiple plant species.Soybean chitinase GmChi02, GmChi01, and GmChi16 (Glyma.02G042500,Glyma.01G160100, and Glyma.16G119200) are phylogenetically close to each other and grouped with most functionally validated plant chitinases.Unlike GmChi02, GmChi01 and GmChi16 were not upregulated by F. oxysporum nor induced by B. ambifaria.Bootstrap values above 60 in the Maximum likelihood tree are shown.
FIGURE 4 Phenotypes of Arabidopsis transgenic lines with GmChi01, GmChi02, or GmChi16.(A) Seedling root length, (B) Rosette leaves, and (C) Plant height of Arabidopsis transgenic lines exhibited no significant difference to the wild type Col-0 or Arabidopsis transgenic line with empty vector.(D) Arabidopsis transgenic lines with GmChi02 or GmChi16 exhibited mild symptoms after the inoculation of F. oxysporum.(E) Quantification of foliar symptoms after the inoculation of F. oxysporum.(F) Arabidopsis transgenic lines with GmChi02 or GmChi16 exhibited better seedling survival rates after the inoculation of F. oxysporum.(G) Quantification of seedling survival rates after the inoculation of F. oxysporum.Different letters indicate significant difference among the Arabidopsis transgenic lines (P< 0.05).

FIGURE 5
FIGURE 5 FIGURE 6 Transcription factor (TF) and TF-binding site (TFBS) analysis for rhizobacteria-inducible soybean chitinase genes.(A) Venn diagram of inducible soybean chitinase genes by Bacillus amyloliquefaciens, Bradyrhizobium japonicum, Burkholderia ambifaria, Lysobacter enzymogenes, Pseudomonas fluorescens, and Rhizobium rhizogenes.Only three soybean TFs responded to B. ambifaria, specifically.(B) Two TFs containing the homeobox domain in response to the inoculation of B. ambifaria, specifically.(C) One TF containing the START domain in response to the inoculation of B. ambifaria, specifically.(D, E) One TF containing the SQUAMOSA promoter-binding (SPB) domain in response to the inoculation of B. ambifaria, specifically.(F) The nodule-specific cis-regulatory motif was found in response to B. amyloliquefaciens, B. japonicum, B. ambifaria, L. enzymogenes, and R. rhizogenes.

TABLE 1
Literature review of functionally characterized plant chitinases against fungal diseases.

TABLE 2
Soybean chitinase genes induced by the inoculation of Fusarium oxysporum, Burkholderia ambifaria, and co-inoculation.