The complete coding region of TaLTP3 (GenBank accession AY226580.1) was amplified from complementary DNA (cDNA) synthesized from RNA extracted from the seedlings of wheat line “Chinese Spring.” The complete coding region of TaLTP3 was constructed to binary T-DNA vector pLGY-02 (Ubi:gene, T-DNA). Transgenic wheat line overexpressing TaLTP3 gene was generated in the genetic background of spring common wheat cultivar “JW1” (selected from a segregating population of the cross of spring common wheat cultivars “Fielder” and “NB1”) with technique support from Bangdi Ltd. Company, Shandong, China. Transgenic wheat lines overexpressing TaPR1a in the same genetic background were derived from a previous study (Bi et al., 2020).

Seedlings of the wheat isogenic line carrying leaf rust resistance gene Lr39/41 in background of common wheat cultivar “Thatcher” were inoculated with Puccinia triticina (Pt), a virulent pathotype THTT following previously described procedure (Li et al., 2018). A high-resistant phenotype with the hypersensitive response (HR) was observed in Lr39/41 isogenic line after inoculation with Pt pathotype THTT. RNA samples for quantitative PCR (qPCR) were collected from wheat leaves inoculated with leaf rust or water control at 0, 1, 2, 5, and 8 days postinoculation (dpi). Four independent biological replicates were included for each time point and treatment. Seedlings of transgenic wheat line TaLTP3-OE and wild-type plants were inoculated with Pt virulent pathotype FHJQ following a similar procedure. The wild-type plant “JW1” showed a moderate susceptible phenotype (rust sporulation with necrotic spots on wheat leaf) toward most of the collected Pt pathotypes including FHJQ and THTT. The disease symptoms were photographed at 14 dpi. An image analysis software ASSESS version 2.0 (American Phytopathological Society, St. Paul, Minisota, MN, United States) was employed to evaluate the percentage of Pt sporulation area for each inoculated leaf (Lamari, 2008; Bi et al., 2020). Two independent transgenic lines of TaLTP3-OE with 9–13 biological replicates were included for the Pt inoculation. The Dunnett’s test was conducted by using statistical analysis software SAS version 9.4.

Model bacterial pathogen P. syringae DC3000 was inoculated in second leaves of wheat seedlings according to the previously described method (Bi et al., 2020). P. syringae DC3000 was cultivated in low-salt Luria-Bertani (LS-LB) medium with 50 μg/mL rifampicin (Rif) antibiotics for 48 h (Jacob et al., 2017) and then diluted to optical density (OD) 600 = 0.6 in distilled water. Second leaves of transgenic wheat line TaLTP3-OE were infiltrated with bacterial suspensions by using a 1-ml needless syringe through the abaxial surface. Each transgenic line consisted of 6–8 independent biological replicates. Seedlings of wild-type plants and water-infiltration served as controls. The infiltration region was marked by using a marker pen. RNA samples for qPCR and RNA-sequencing (RNA-seq) assays were collected from the infiltration region at 6 h post-inoculation (hpi).

Ribonucleic acid samples were sent to Novogene Corporation Ltd. for KAPA library construction and transcriptome sequencing following the default procedure. The Illumina HiSeq 1000 System was employed for the 12-Gb sequencing for each RNA sample. Transcriptome was assembled with reference genome of Triticum aestivum (common wheat “Chinese Spring” TGACv1 version) (Appels et al., 2018) by using the HISAT2 FDR and StringTie software (Kim et al., 2019). Assembled transcripts that could not be found in the gene models of the reference genome were annotated as “novel” transcripts. Gene expressions were profiled by using the HTSeq version 0.9.1 software (Trapnell et al., 2010). The expected number of fragments per kilobase of transcript sequence per million base pairs (FPKM) for each gene was used to evaluate the expression of assembled genes. FPKM-based heatmaps were generated by the MeV software. Differentially expressed genes (DEGs) were filtered the expression levels of genes between different groups for an false discover rate (FDR)-adjusted p-value < 0.05 and |log2-fold change| > 1 by using the DESeq2 software (Love et al., 2014). The Gene Ontology (GO) annotations were assigned to each DEGs with the GO sequencing (Young et al., 2010). The KEGG (Kyoto Encyclopedia of Genes and Genomes) annotations were employed to profile the regulatory pathways enriched with DEGs (Kanehisa and Goto, 2000).

Ribonucleic acids were prepared by using the RNA Extraction Kit (QIAGEN, Hilden, Germany, United Kingdom) and the first-strand cDNA was synthesized by using the Reverse Transcription Kit (Takara, Dalian, China). Primers for qRT-PCR assay were designed (Supplementary Table 1) and a preliminary test was performed on the Roche LightCycler 96 qRT-PCR machine by using a series of 2-fold diluted cDNA templates (1:1, 1:2, 1:4, 1:8, and 1:32) to evaluate the amplification efficiency for each pair of primers. Melting curves were generated to ensure the specificity of the amplification. The wheat reference gene TaActin (GenBank accession AB181991) was used as the endogenous control (Paolacci et al., 2009). The transcript level was expressed relative to that of TaActin based on the 2^–ΔCt method as described in a previous study (Wu et al., 2019). Mean and SE were calculated by using the Microsoft Excel software. The two-sample t-test and the generalized linear model (GLM) ANOVA were performed by using the SAS software version 9.4.

Full-length open reading frame (ORF) of TaLTP3 was cloned into transient expression vectors of pGWB5 (35S:gene-GFP, T-DNA) and pGWB454 (35S:gene-RFP, T-DNA), respectively. The recombinant vectors were transformed into Agrobacterium strain GV3101 and then used for infiltration of Nicotiana benthamiana (N. benthamiana). Corresponding empty vectors expressing only green fluorescent protein (GFP) and red fluorescent protein (RFP) served as controls. Green or red fluorescence was checked 48 hpi by using a Nikon Ti-2 microscope with a GFP or RFP filter, respectively. Epidermal peels of tobacco leaves were prepared from the infiltration region and soaked in 800 mM mannitol for 6 min to induce the plasmolysis (Ma et al., 2012).

Primers that were used to amplify the full-length ORF of TaLTP3 and five other plant defense-related LTP genes from cDNA synthesized from RNA isolated from the common wheat line “Chinese Spring” are listed in Supplementary Table 1. A neighbor-joint tree including all the LTP proteins in this study was generated by using the MEGA software version 7.0. The signal peptides of wheat LTPs were predicted by SignalP 5.0 server.¹ The amplified DNA fragment lacking the sequence encoding the signal peptide was cloned into a Gateway yeast two-hybrid (Y2H) vector pLAW10 bait domain (BD). Recombined Y2H activation domain (AD) vectors carrying different regions of the TaPR1a gene were derived from a previous study (Bi et al., 2020). The coding regions of TaPR1a_(25–93aa) and TaPR1a_(25–128aa) were cloned into a Gateway Y2H vector pLAW11 (AD). Yeast transformation was performed by using lithium acetate and polyethylene glycol 3350 as previously described (Gietz and Schiestl, 2007). Cotransformed yeast strains were spotted on SD-Leu-Trp-His-Ade selective medium.

Full-length ORF of TaLTP3 with signal peptide was constructed to pDEST-^GWVYNE (35S:gene-YFP^N, T-DNA). The recombinant construct of TaPR1a-pDEST-^GWVYCE carrying complete coding region of TaPR1a was derived from our previous study (Bi et al., 2020). The fusion genes were coexpressed in tobacco leaves by Agrobacterium infiltration. Yellow fluorescence was checked 48 hpi by using a Nikon Ti-2 microscope with a yellow fluorescent protein (YFP) filter. Coexpression of AvrLm1-pDEST-^GWVYNE and BnMPK9-pDEST-^GWVYCE in the same vectors was used as a positive control (Ma et al., 2018). Signal peptide-fused empty vectors of SP_(TaPR1a)-pDEST-^GWVYNE and SP_(TaPR1a)-pDEST-^GWVYCE were derived from our previous study (Bi et al., 2020) and coexpressed with TaPR1a-pDEST-^GWVYCE and TaLTP3-pDEST-^GWVYNE, respectively, as negative controls.

Deoxyribonucleic acid fragment encoding the signal peptide-truncated TaLTP3 protein was cloned into pGWB5 as 35S:TaLTP3_(30–122)-GFP. A specific reverse primer of TaPR1a with stop codon and sequence encoding cMYC tag was designed and used to amplify the DNA fragment encoding the signal peptide-truncated TaPR1a protein. The resulting DNA segment was constructed into pGWB5 as 35S:TaPR1a_(25–164)-cMYC. Recombinant pGWB5:TaLTP3_(30–122)-GFP and pGWB5: TaPR1a_(25–164)-cMYC were coexpressed in tobacco leaves by using Agrobacterium. Coexpression of GFP and TaPR1a_(25–164)-cMYC served as a negative control. Total proteins were extracted from Agrobacterium-infiltrated tobacco leaves at 3 dpi and a portion of the total protein was reserved as an input sample. α-GFP IP samples were prepared by using GFP-trap beads (LABLEAD, Beijing, China). Input and IP samples were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto membranes made of polyvinylidene difluoride (PVDF). Immunoblots were performed by using α-GFP or α-cMYC antibody (Solarbio, Beijing, China).

To visualize the accumulation of ROS in wheat leaves infiltrated with P. syringae DC3000, two chemical reagents, nitroblue tetrazolium (NBT) and 3,3-diaminobenzidine (DAB), were applied to stain the superoxide anion (O₂^–) and hydrogen peroxide (H₂O₂), respectively. The staining protocols were modified from a previous study (Wang et al., 2007). Leaf regions infiltrated with P. syringae DC3000 were collected at 2, 4, and 8 hpi. For O₂^– visualization, samples were soaked in 10 mM sodium azide (NaN₃) and 10 mM potassium phosphate buffer (pH 7.8) with 0.1% NBT (w/v) for 24 h. For H₂O₂ staining, samples were incubated in hydrochloric acid (HCl)-acidified solution (pH 3.8) containing 1 mg/mL of DAB for 24 h. Thereafter, samples were decolored in boiling 95% ethanol for 10 min. The percentage of stained area in each leaf was determined by using the ASSESS software (Lamari, 2008). The whole experiment was repeated twice with two independent transgenic lines and each repeat consisted of 6–10 biological replicates. The Dunnett’s test was conducted by using the SAS software version 9.4.

To determine the role of TaLTP3 in plant resistance, we first quantified the expression change of this gene in leaf rust resistance mediated by Lr39/41 gene. Interestingly, TaLTP3 was significantly upregulated at the wheat early-stage resistance toward leaf rust (Figure 1A). To further explore the function of TaLTP3, transgenic wheat lines overexpressing TaLTP3 (TaLTP3-OE) were generated in the genetic background of commonspring wheat line “JW1.” The expression levels of TaLTP3 in TaLTP3-OE reached 1.7- to 2.8-fold of those in the wild-type (Figure 1B). The third leaves of TaLTP3-OE and wild-type plants were inoculated with urediniospores of the Pt pathotype FHJQ and FHJQ exhibited high virulence on wild-type plants (Figure 1C). The sporulation of leaf rust pathogen was observed in both the TaLTP3-OE and wild-type plants. The percentage of Pt sporulation area for each inoculated leaf was evaluated by using the ASSESS software. Compared with wild-type plants, significantly enhanced resistance to Pt was observed in TaLTP3-OE as evidenced by decreased sporulation area (^**p < 0.01, ^***p < 0.0001, 9–30 biological replicates) (Figure 1C and Supplementary Table 2). We further investigated several agronomic traits of transgenic wheat line TaLTP3-OE including plant height, spike length, spikelet number per spike, and seed number per spike under greenhouse conditions. We did not observe any significant difference in these agronomic traits between TaLTP3-OE and wild-type plants (Figure 1D). However, when we checked the seed size and thousand grain weight, we found smaller and lighter seeds harvested from TaLTP3-OE than wild-type plants (Figure 1E), indicating overexpression of TaLTP3 might have influence on wheat production.

To reveal the mechanism of TaLTP3 during plant resistance, we have used model bacterial pathogen P. syringae pv. tomato DC3000 to trigger the immune response in wheat leaves. Although P. syringae DC3000 is not a wheat pathogen under a natural environment, a typical cell death was observed in wheat leaf infiltrated with P. syringae DC3000 at 48 hpi (Figure 2A), indicating a successful induction of non-host resistance. We applied RNA-seq on samples collected from P. syringae DC3000-infiltrated third leaves of wheat transgenic line TaLTP3-OE and wild-type plants at 6 hpi. Samples from the water-infiltrated leaves served as a control. Each of the treatments included three biological replicates and, in total, 12 samples were subjected to a 12-Gb Illumina sequencing (Supplementary Table 3). The reference genome of Triticum astivum from the Ensembl Genomes (TGACv1 version) was used to assemble the transcriptome. A total of 122,130 transcripts were mapped to the genome sequence (Supplementary Table 4). Significant (R² > 0.92) correlations of the overall gene expressions among biological replicates were detected (Supplementary Table 5). We used the FPKM value to predict the expression levels of transcripts. Raw reads from the RNA-seq were deposited at the National Center for Biotechnology Information (NCBI) BioProject PRJNA746113.

The expression levels of all 18 PR gene families were displayed in the heat map based on their FPKM values derived from the transcriptome data (Figure 2B). Most of the PR genes were highly induced by the infection of P. syringae DC3000, indicating an activation of a broad-range plant defense response by this model pathogen. Compared with the wild-type plants, TaPR1a, TaPR5, and TaPR15 showed higher levels of expression in TaLTP3-OE. DEGs in comparison of “WT_DC3000 vs. WT_WATER” were identified (|log₂-fold change| > 1 and q-value < 0.01) by using DESeq2. In the wild-type plants, we found 9,178 genes were significantly upregulated upon P. syringae DC3000 infection. The KEGG annotations of these P. syringae DC3000-sensitive genes indicated the activation of a broad-range plant defense responses including Ca²⁺-dependent ROS induction, mitogen-activated protein kinase (MAPK) signaling, and effector-triggered immunity (Supplementary Figure 1 and Supplementary Table 6). The GO annotations for these DEGs enriched in “plant–pathogen interaction” and “protein processing in endoplasmic reticulum” (Figure 2C).

From the comparison between “TaLTP3-OE_DC3000” and “WT_DC3000,” only 214 genes were significantly higher expressed (log₂-fold change > 1 and q-value < 0.01) in TaLTP3-OE transgenic upon P. syringae DC3000 infection. The GO annotations of these upregulated DEGs enriched in “Various types of N-glycan biosynthesis” and “Glycerophospholipid metabolism” (Figure 2D). Among all these TaLTP3-responsive upregulated DEGs, TaPR1a was the only gene annotated into the KEGG pathway of “plant–pathogen interaction” (Table 1). Furthermore, we found 173 genes with significantly lower (log2-fold change < -1 and q-value < 0.01) expressions in TaLTP3-OE transgenic line upon P. syringae DC3000 infection. Several key regulators in the hormone pathway of auxin and JA, including auxin response factor 9 (ARF9), A-type Arabidopsis response regulator (A-ARR), and protein TIFY 6a [JAZ (jasmonate ZIM-domain)], were downregulated in the KEGG pathway of “plant hormone signal transduction” (Table 1 and Supplementary Figure 2).

We further evaluated the expression levels of two SA synthesis genes, TaPAL and TaPAD4, during P. syringae DC3000 infection in TaLTP3-OE transgenic line by using qRT-PCR assay (Figure 3). qPCR showed that the expression levels of both the TaPAL and TaPAD4 were elevated upon P. syringae DC3000 infection in TaLTP3-OE transgenic line, which was consistent with those in RNA-seq (Figure 2B), suggesting that our RNA-seq data are reliable.

As overexpression of TaLTP3 affected TaPR1a expression during plant defense response (Figure 2B, Supplementary Figure 2, and Table 1) and both TaLTP3 and TaPR1a are predicted as secreted proteins by signal P, we hypothesized that TaLTP3 might function through physical interaction with TaPR1a in the apoplastic space. The apoplast localization of TaPR1a has been validated in our previous study (Bi et al., 2020). GFP- and RFP-tagged TaLTP3 were generated, respectively. The recombinant constructs TaLTP3-GFP and TaLTP3-RFP were transientlyexpressed in N. benthamiana by agroinfiltration. Plasmolysis was induced by treatment with 800 mM mannitol. Pronounced green and red fluorescent signals for TaLTP3-GFP and TaLTP3-RFP were visualized in the apoplastic space, respectively (Figure 4), but not for the empty vector (EV-GFP and EV-RFP) controls (Figure 4), suggesting that TaLTP3 localizes in the apoplastic space.

To examine the interaction between TaLTP3 and TaPR1a, a Y2H assay was performed by using a signal peptide-truncated version of these two proteins. Strong interaction was observed between TaPR1a_(25–164) and TaLTP3_(30–122) as yeast transformed with both the constructs was able to grow on the stringent synthetic dropout (SD) selective media lacking leucine, tryptophan, histidine, and adenine (SD-Leu-Trp-His-Ade) (Figure 5A). Because the C-terminal region of TaPR1a contains the conserved CAPE1 peptide in most PR1 proteins that might act as a DAMP in triggering plant defense response, we generated several truncated versions of TaPR1a protein, in which different C-terminal regions were deleted (Figure 5B). As shown in Figure 5A, an N-terminal region of TaPR1a ranging from 25 to 93 aa was sufficient to mediate the interaction with TaLTP3. However, we did not observe any interaction between TaPR1a (94–164 aa) and TaLTP3. As many wheat LTP genes, including TaLTP2, TaLTP3F1, TaLTP4.3, TaLTP4.1, and TaPR14, have been previously reported in wheat resistance to various pathogens, we also checked the interaction of TaPR1a with these LTP proteins and found that TaPR1a interacts with all these LTP proteins (Supplementary Figure 3). Collectively, these initial Y2H results indicated a conserved physical interaction between TaPR1a and LTP proteins.

To validate the in-planta interaction between TaLTP3 and TaPR1a, a bimolecular fluorescence complementation (BiFC) assay was performed in N. benthamiana. Strong yellow fluorescence signals in the apoplast were observed in N. benthamiana leaves coexpressing TaLTP3-YFP^N and TaPR1a-YFP^C (Figure 5C), whereas fluorescence in cytoplasm and nucleus was found in positive control coexpressing AvrLm1-YFP^N and BnMPK9-YFP^C (Ma et al., 2018). No YFP signals were observed in the N. benthamiana leaves coexpressing the negative controls: TaLTP3-YFP^N and SP_(TaPR1a)-YFP^C nor SP_(TaPR1a)-YFP^N and TaPR1a-YFP^C.

To further confirm the interaction in vivo, a co-immunoprecipitation (co-IP) assay was performed following transient expression of TaPR1a_(25–164)-cMYC and TaLTP3_(30–122)-GFP constructs in N. benthamiana. GFP was used as a negative control. cMYC-tagged TaPR1a co-immunoprecipitated with GFP-tagged TaLTP3, but not with GFP (Figure 5D and Supplementary Figure 4), indicating that TaLTP3 physically interacts with TaPR1a.

To explore the coeffects of TaLTP3 and TaPR1a during plant defense response, we evaluate the accumulation of H₂O₂ and O₂^– in the transgenic wheat lines TaLTP3-OE and TaPR1a-OE during P. syringae DC3000 infection. Leaf samples were collected at 2, 4, and 8 h after P. syringae DC3000 inoculation and stained with DAB and NBT to visualize H₂O₂ and O₂^–, respectively. Percentage of the stained area in each infiltrated leaf was calculated by using the ASSESS software. Compared with the wild-type plants, we observed significantly higher (*p < 0.05, ^**p < 0.01, 6–10 biological replicates) accumulation of H₂O₂ in the transgenic wheat lines TaPR1a-OE and TaLTP3-OE at 4 and 8 hpi (Figure 6A and Supplementary Table 7). For the NBT staining, the transgenic wheat line TaPR1a-OE showed a fast and higher accumulation of O₂^– at 2 and 4 hpi, whereas the transgenic wheat line TaLTP3-OE accumulated more O₂^– only at 4 hpi (Figure 6B and Supplementary Table 8). These results indicated that both the TaLTP3 and TaPR1a might function through a ROS-dependent plant defense response.