High temperatures affect the hypersensitive reaction, disease resistance and gene expression induced by a novel harpin HpaG-Xcm

Harpin proteins are produced by plant-pathogenic Gram-negative bacteria and regulate bacterial pathogenicity by inducing plant growth and defence responses in non-hosts. HpaG-Xcm, a novel harpin protein, was identified from Xanthomonas citri pv. mangiferaeindicae, which causes bacterial black spot of mango. Here, we describe the predicted structure and functions of HpaG-Xcm and investigate the mechanism of heat resistance. The HpaG-Xcm amino acid sequence contains seven motifs and two α-helices, in the N- and C-terminals, respectively. The N-terminal α-helical region contains two heptads, which form the coiled-coil (CC) structure. The CC region, which is on the surface of HpaG-Xcm, forms oligomeric aggregates by forming hydrophobic interactions between hydrophobic amino acids. Like other harpins, HpaG-Xcm was heat stable, promoted root growth and induced a hypersensitive response (HR) and systemic acquired resistance in non-host plants. Subjecting HpaG-Xcm to high temperatures altered the gene expression induced by HpaG-Xcm in tobacco leaves, probably due to changes in the spatial structure of HpaG-Xcm. Phenotypic tests revealed that the high-temperature treatments reduced the HR and disease resistance induced by HpaG-Xcm but had little effect on growth promotion. These findings indicate that the stability of interactions between CC and plants may be associated with thermal stability of HpaG-Xcm.

A phylogenetic tree was constructed based on conservative sequence analysis of the complete amino acid sequences of HpaG-Xcm and the nine homologous harpin proteins from Xanthomonas spp. (Fig. 2a). Seven relatively conservative motifs that were arranged in the same order were identified in all 10 ten proteins r (Fig. 2b,c). The phylogenetic tree grouped the 10 proteins into three branches: HpaG-Xcm, HpaG-Xag, HpaG-Xam, Hpa1Xac and HpaXm; Hpa1Xoc, Hpa1Xoo-2 and Hpa1Xoo; and XopA-Xcv and HpaXcc (Fig. 2a). The shortest of the seven conserved motifs was composed of 11 amino acid residues, and the longest motif was composed of 29 amino acid residues (Fig. 2c). The motif characteristics of the homologous proteins were basically consistent with their position on the phylogenetic tree.
Secondary structure, tertiary structure and coiled-coil predictions of HpaG-Xcm. Structure prediction analyses indicated that the secondary structure of HpaG-Xcm consists of helixes, strands and coils. There is an α-helix at the N-terminal and C-terminal of HpaG-Xcm, which have the sequences 37-EKQLDQLLTQLIMALLQ-53 and 89-QYTQMLMNIVGDILQAQN-106, respectively (Fig. 3a). The solvent accessibility prediction of amino acid residues showed that the values of Leu40, Leu44, Ile48, Met93, Leu94 and Scientific RepoRts | (2019) 9:990 | https://doi.org/10.1038/s41598-018-37886-9 Val98 were 0 ( Fig. 3a), indicating that these amino acid residues were located within the structure. Furthermore, these amino acids were distributed in the α-helix regions. The coiled-coil (CC) prediction indicated that Gln33 to Gln60 of the HpaG-Xcm amino acid sequence had the possibility of forming a CC, and that there were two complete heptads at the N-terminal: 33-QGISEKQ-LDQLLTQ-46 (Fig. 3b). The 3D model of HpaG-Xcm showed that the structure of this protein was predominantly made up of helical regions connected by turns and loops (Fig. 3c). Leu40, Leu43, Leu47 and Ala50 are critical amino acids for the formation of the CC structure (Fig. 3b), and are located on the hydrophobic side of the α-helix (Fig. 3c).
Expression and purification of HpaXcm. In order to explore the optimum expression conditions for GST-HpaXcm (the fusion protein GST-HpaG-Xcm), we investigated the effect of different isopropyl-β-D-thiogalactoside (IPTG) concentrations (0.05 mM and 0.1 mM), different induction times (3 h and 5 h) and different induction temperatures (28 °C and 37 °C) on the expression of the fusion protein. SDS-PAGE analysis indicated that GST-HpaXcm was expressed as two forms: soluble proteins and inclusion bodies (Supplemental Fig. S1). Although the GST-HpaXcm fusion protein (39 kDa) could be expressed under different conditions, different induction conditions led to a different proportion of soluble protein in the total protein. When the expressed fusion soluble proteins were infiltrated into the intercellular space of tobacco leaves, all the proteins induced by the different conditions were able to induce HR production in tobacco leaves; however, the necrosis ratios (i.e., the ratio of necrotic area to injection area) were different. The maximum necrosis ratio was close to 1, which was induced by soluble protein C (Supplemental Fig. S1). Therefore, the optimum expression conditions for GST-HpaXcm is an IPTG concentration of 0.1 mM and 5 h of induction at 28 °C.  SDS-PAGE and a western blot were performed using a purified sample of the GST-HpaXcm fusion protein. A band of GST-HpaXcm appeared at 39 kDa, which was consistent with the theoretical molecular weight (the molecular masses of the GST-tag protein and HpaXcm are 26 kDa and 13.56 kDa, respectively) (Supplemental Fig. S2). GST-HpaXcm was digested by thrombin to obtain the GST-tag protein and the purified HpaXcm (Supplemental Fig. S2).

Induced HR, disease resistance and root growth promotion by HpaXcm. To determine whether
HpaXcm could induce an HR response in tobacco leaves, purified HpaXcm (10 μM), purified and boiled HpaXcm-B (10 μM) or purified Hpa1Xoo (10 μM) were injected into tobacco leaves; phosphate-buffered saline (PBS) was used as a negative control. Unlike the negative controls, HpaXcm, HpaXcm-B and Hpa1Xoo induced a visible HR in the tobacco leaves (Fig. 4 left). The ratio of the necrotic area to the injected area of HpaXcm was significantly greater than that of HpaXcm-B (P < 0.05) and Hpa1Xoo (P < 0.01) (Fig. 4

right)
To determine whether HpaXcm could induce disease resistance, equal volumes of HpaXcm, HpaXcm-B, Hpa1Xoo or PBS (which was used as a negative control) were sprayed on tobacco leaves 16 h before inoculating the leaves with TMV. After 6 d, all the tobacco leaves had developed necrotic spots (Fig. 5a); however, the proportion of the leaf area that was necrotic was significantly smaller (P < 0.01) for leaves that had been treated with HpaXcm, HpaXcm-B or Hpa1Xoo compared with that of the control group. The necrotic area of leaves treated with HpaXcm-B was greater (P < 0.01) than that of leaves treated with HpaXcm; however, the necrotic areas of leaves treated with HpaXcm or Hpa1Xoo were not significantly different (Fig. 5b).
To determine whether HpaXcm could promote root growth, Arabidopsis thaliana seeds were soaked in the same concentration of HpaXcm, HpaXcm-B, Hpa1Xoo or PBS (which was used as a negative control) for 6 h, before transferring the seeds to MS medium for culture. After 15 d, seeds soaked in HpaXcm, HpaXcm-B or Hpa1Xoo produced stronger roots and more branches than those in the control group (Fig. 6a). The mean root length and mean fresh weight were also significantly greater (P < 0.05) than those of the control group. The mean root length of A. thaliana plants that grew from seeds that had received the HpaXcm treatments were significantly greater (P < 0.05) than those that grew from seeds that had received the Hpa1Xoo treatment (Fig. 6b). However, the plants that grew from seeds that had received the HpaXcm or HpaXcm-B treatment were not significantly different (Fig. 6b). The phylogenetic tree was constructed using the Maximum Likelihood method, and the Bootstrap value was set to 1000. HpaG-Xcm from X. citri pv. mangiferaeindicae, HpaG-Xag from X. axonopodis pv. glycines, HpaG-Xam from X. phaseoli pv. manihotis, Hpa1Xac from X. axonopodis pv. citri, HpaXm from X. citri subsp. malvacearum, Hpa1Xoc from X. oryzae pv. oryzicola, Hpa1Xoo-2 from X. oryzae pv. oryzae, Hpa1Xoo from X. oryzae pv. oryzae, XopA-Xcv from X. campestris pv. vesicatoria, HpaXcc from X. campestris pv. campestris. (b,c) Conservative analysis for characteristic domains of HpaG-Xcm and nine harpins from Xanthomonas spp. and their positions in sequences determined using MEME (http://meme-suite.org/tools/meme). The stacking height of amino acids at different sites of the sequence are shown as the conserved degree of the site, and the height of the single amino acid in the stack shows the relative frequency of this amino acid at this position. qRT-PCR. In order to further understand whether HpaXcm induces HR, disease resistance and growth promotion because related genes were stimulated to increase their expression levels, and to explore the effects of temperature on protein functions, we compared the relative expression levels of the HR-related genes Hin1 16 and HSR203J 17 , the defence-related genes NPR1 and PR-1a 11 , and the growth promotion-related gene NtEXP6 18 in tobacco leaves at 1 h, 3 h and 6 h with the expression levels in tobacco leaves at 0 h by performing quantitative real-time PCR (qRT -PCR). The relative expression levels of genes were calculated using the 2 −ΔΔCT formula with EF-1a as a control for normalization. In general, the expression of NtEXP6, Hin1, HSR203J and NPR1 was clearly  detected in tobacco leaves within 6 h of being treated with unheated HpaXcm (28 °C HpaXcm). However, when leaves were injected with HpaXcm that had been heated at different temperatures, the relative expression levels of NtEXP6, Hin1, HSR203J, NPR1 and PR-1a changed (Fig. 7).
NtEXP6. Leaves treated with unheated (28 °C) HpaXcm did not express NtEXP6 1 h after treatment; however, significant expression (P < 0.01) of NtEXP6 was observed at 3 h and at 6 h. The highest levels of expression were observed at 6 h. Leaves treated with unheated (28 °C) HpaXcm induced higher levels of NtEXP6 expression than those treated with HpaXcm that had been heated to 100 °C, 150 °C or 200 °C. The relative expression induced by 28 °C HpaXcm and 100 °C HpaXcm peaked at 6 h; however, the relative expression induced by 150 °C HpaXcm  and 200 °C HpaXcm peaked at 1 h. The peak values of relative expression induced by HpaXcm treated at different temperatures ranging from high to low were: 28 °C HpaXcm, 100 °C HpaXcm, 150 °C HpaXcm and 200 °C HpaXcm (Fig. 7).
Hin1. Leaves treated with 28 °C HpaXcm showed significant (P < 0.05) relative expression of Hin1 at 1 h, expression peaked at 3 h, and then declined; however, the relative expression was still remarkable compared with that observed at 0 h. Leaves treated with 100 °C HpaXcm, 150 °C HpaXcm or 200 °C HpaXcm showed significant expression levels of Hin1 within 6 h of the treatment, with relative expression peaking at 6 h, unlike the leaves subjected to the 28 °C HpaXcm treatment. The peak values of relative expression induced by HpaXcm treated at different temperatures ranging from high to low were: 150 °C HpaXcm, 200 °C HpaXcm, 100 °C HpaXcm and 28 °C HpaXcm (Fig. 7).

Discussion
Phylogenetic analysis revealed that HpaXcm and the nine harpin proteins from Xanthomonas contained seven conservative motifs and could be divided into two classes: HpaG-Xcm, HpaG-Xag, HpaG-Xam, Hpa1Xac, HpaXm, Hpa1Xoc, Hpa1Xoo-2 and Hpa1Xoo; and XopA-Xcv and HpaXcc (Fig. 2). Conserved motifs are closely related to protein structure and function 19 . Seven motifs were found in all eight amino acid sequences of the first group, whereas only four and two motifs were found in XopA-Xcv and HpaXcc, respectively. Harpins can be classified as 'T' or 'C based on the Thr residue and Cys residue in the N-terminal α-helix region. Based on their 'T' and 'C' type, the first group of Xanthomonas spp. harpins can be divided into two branches: HpaG-Xcm, HpaG-Xag, HpaG-Xam, Hpa1Xac and HpaXm belong to the 'T' type and Hpa1Xoc, Hpa1Xoo-2 and Hpa1Xoo belong to the 'C' type 6 . HpaXcm contains seven motifs in a complete sequence and a Thr residue in the N-terminal α-helix region. On the basis of the amino acid composition and the arrangements of motifs of the 10 harpins we investigated, HpaXcm showed the closest relationship to HpaG-Xag. Previous studies have indicated that Xanthomonas spp. harpins generally have two α-helices located in the N-and C-terminal. The N-terminal α-helices are crucial for inducing the HR of tobacco leaves 14 , and the two heptads in the α-helical regions influence the HR activity of proteins 10 . The secondary structure of HpaXcm involved two α-helices in the N-and C-terminal, and the solvent accessibility values of Leu40, Leu44, Ile48, Met93, Leu94 and Val98 were 0 (Fig. 3a). In other words, these amino acid residues located in the α-helical regions are almost buried. The a and d positions of the CC structure are hydrophobic amino acids, which are responsible for the formation of the CC. The hydrophobic interactions between hydrophobic amino acids prompt the two α-helices to form supercoils 20 . The CC prediction involved two complete heptads in the N-terminal α-helical region: 33-QGISEKQ-LDQLLTQ-46, and Gln33, Ser36, Leu40, Leu43, Leu47 and Ala50 (residues at a and d positions) were key amino acids in the formation of the CC (Fig. 3b). The predicted HpaXcm tertiary structure model was more intuitive so that Leu40, Leu43, Leu47 and Ala50 were kept at the hydrophobic side of the α-helix (Fig. 3c). Hydrophobic interactions between these amino acids cause α-helices to aggregate and form oligomeric aggregates, which play an important role in protein function. Accordingly, we concluded that the CC structure of HpaXcm associated with functions is located at the protein surface.
Tobacco leaves injected with HpaXcm and HpaXcm-B produced obvious necrotic patches, which indicated the activity of the HR and the thermostability. The necrosis ratio produced by HpaXcm-B was significantly different from that of HpaXcm, suggesting that a temperature of 100 °C affects the HR activity but is not high enough to inhibit it (Fig. 4 right). The necrosis ratios of tobacco leaves treated with HpaXcm and HpaXcm-B before infection with TMV were significantly lower than that of the control leaves, which indicated that HpaXcm induces resistance to TMV in tobacco and has thermal stability (Fig. 5a). However, the necrosis rate of leaves treated with HpaXcm-B was higher than that of leaves treated with HpaXcm (Fig. 5b), suggesting that the ability of HpaXcm to induce disease resistance had been weakened by boiling HpaXcm at 100°C. The root lengths and fresh weights of plants that germinated from seeds soaked with HpaXcm and HpaXcm-B were significantly higher than that of the control group, and there was no significant difference between the HpaXcm and HpaXcm-B treatments, indicating that HpaXcm has the ability to promote root growth and that this was not affected by the 100 °C treatment (Fig. 6b). We observed that although the HR activity and disease resistance of HpaXcm was reduced at 100 °C, the 100 °C treatment had little impact on growth promotion. In previous studies, the growth-promoting of Hpa1Xoo has been reported 21,22 , and its application has been studied 23 . Thus, we chose Hpa1Xoo to provide a contrast when comparing the activity of Hpa1Xoo with that of HpaXcm. Compared with Hpa1Xoo at the same concentration, the necrosis rate induced by HpaXcm was apparently higher (Fig. 4 right), that is, the HR activity of HpaXcm is stronger than that of Hpa1Xoo. Compared with Hpa1Xoo at the same concentration, the root lengths and fresh weights of plants that germinated from seeds soaked with HpaXcm were apparently higher, that is, the growth promoting ability of HpaXcm was stronger than Hpa1Xoo (Fig. 6b), which means that HpaG-Xcm has potentially more value in terms of agricultural applications and development as a growth promoter.
The HR functional domains of HpaXcm and Hpa1Xoo are located in the N-terminal α-helical regions, and both of them have the possibility of a CC structure. The most important difference between them is that the N-terminal α-helical region of HpaXcm contains Thr and is classified as a T-type harpin, whereas Hpa1Xoo contains Cys and is classified as a C-type harpin. The hydrophobicity of Cys (which has a hydrophobic parameter of 2.5) is stronger than Thr (which has a hydrophobic parameter of −0.7) and, hence, Hpa1Xoo has stronger hydrophobic interactions. Previous studies have suggested that the HR activity of the Hpa1Xoo36-52 fragment was increased after heat treatment 22 , which is the opposite of our findings for HpaXcm. We speculate that the interactions of the CC structure probably affects the HR activity and the thermal stability of HpaXcm.
Unlike the PR-1a gene, NtEXP6, Hin1, HSR203J and NPR1 were significantly expressed in leaves 6 h after being subjected to the 28 °C HpaXcm, 100 °C HpaXcm or 200 °C HpaXcm treatments. All five genes were expressed in leaves after being subjected to the 150 °C HpaXcm treatment. The expression of PR-1 is regulated by many factors, such as the accumulation of salicylic acid and the expression of NPR1 24 . The lack of PR-1a expression in leaves 6 h after being subjected to the 28 °C HpaXcm, 100 °C HpaXcm or 200 °C HpaXcm treatments is probably due to an insufficient accumulation of salicylic acid. The peak times of NtEXP6 and PR-1a expression changed with heating temperature, whereas Hin1, HSR203J and NPR1 peak expression times remained unchanged (Fig. 8). The results showed that heating HpaXcm at different temperatures had different effects on the expression of the five genes. As the protein treatment temperature increased, the expression of NtEXP6 decreased, i.e., the higher the temperature, the lower the ability of HpaXcm to induce NtEXP6 expression. By contrast, the higher the temperature, the stronger the ability of HpaXcm to induce NPR1 expression. With the increase of protein treatment temperature, the expression of Hin1 and HSR203J increased and then decreased, with the highest expression levels induced by the 150 °C HpaXcm treatment (Fig. 8). Protein-specific functions are determined by their specific conformations. Changes in the spatial structure of protein are accompanied by changes in biological activity. We speculate that the spatial structure of HpaXcm changes with high temperature heating, resulting in an enhanced, weakened or even a loss of the ability to induce gene expression.
In this study, we showed not only that HpaXcm has the common characteristics of a harpin protein but also that the HR activity and growth promoting effect of HpaXcm were stronger than those of Hpa1Xoo at the same concentration. We speculate that the spatial structure of HpaXcm may change when subjected to high temperatures and that the stability of interactions between CC and plants is probably associated with the HR activity and thermal stability. These findings provide a basis for further exploring the functional mechanism and heat resistance mechanism of harpins.

Materials and Methods
Bacteria, virus and plant materials. Xanthomonas citri pv. mangiferaeindicae HNHK and BL21/pGEX-HpaXm were maintained in the laboratory at −80 °C. Seeds of tobacco (Nicotiana tabacum L. cv. NC89) and Arabidopsis thaliana, ecotype Columbia were stored in the laboratory at 4 °C. X. citri pv. mangiferaeindicae strains were cultured in nutrient broth (NB) and on nutrient agar (NA) medium 25 at 28 °C. Escherichia coli DH5α was cultured in Luria broth (LB) and on LB agar medium with a final concentration of 100 µg/ml ampicillin at 37 °C 6 . The leaves of Tobacco mosaic virus (TMV) were stored at −80 °C after rapid freezing in liquid nitrogen.
Cloning and construction of the hpaG-Xcm expression vector. The hpaG-Xcm fragment was cloned from X. citri pv. mangiferaeindicae using PCR technology, the primers (Supplemental Table S1) were designed based on the hpaXm sequence (GenBank Accession No. DQ643828) from X. citri subsp. malvacearum. X. citri pv. mangiferaeindicae genomic DNA was extracted using the OMEGA HP Plant DNA Kit (OMEGA, Norcross, GA, USA.). The following procedure was used for the PCR reaction: pre-denaturation at 95 °C for 5 min, followed by 35 cycles of amplification (95 °C for 30 s, 58 °C for 30 s, 72 °C for 1 min), and then a final extension of 7 min at 72 °C 26 . The PCR product was purified using the OMEGA Gel Extraction Kit (OMEGA, Norcross, GA, USA), connected with the pMD18-T vector (TransGen Biotech, Beijing, China) to become T-hpaG-Xcm, and then transformed into the E. coli strain DH5α and sequenced. The T-hpaG-Xcm and pGEX-HpaXm plasmids were extracted using a Plasmid Mini Kit (OMEGA, Norcross, GA, USA). hpaG-Xcm and pGEX-EF 27 , the fragments obtained by double digestion, were connected using the BamHI-SacI sites, and then transformed into E. coli strain BL21 (DE3) and sequenced.
HpaG-Xcm protein expression, purification and western blot. LB liquid medium containing ampicillin was inoculated with BL21/pGEX-HpaXcm (pGEX-HpaG-Xcm transformed into BL21 (DE3)). Isopropyl-β-D-thiogalactoside (IPTG) with final concentrations of 0.05 mM and 0.1 mM were added to induce protein expression, until the optical density at 600 nm increased to 0.6-0.8, and then cultured at 28 °C and 37 °C for 3 h and 5 h, respectively. The bacteria collected after centrifugation were resuspended in 1 × PBS (phosphate buffered saline) (Solarbio, Beijing,China) and broken using an ultrasonic wave. After centrifugation, the supernatants and precipitates were collected and identified by performing 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) 6,25 . Purified GST-HpaXcm (the fusion protein of a glutathione S-transferase (GST)-tag and HpaG-Xcm) was obtained from crude protein using a GST-tag Protein Purification Kit (Beyotime, Shanghai, China), and GST-HpaXcm was cleaved using thrombin (GE, Boston, MA, USA) at 22 °C for 16 h to obtain the purified HpaG-Xcm. HpaG-Xcm was diluted to 10 μM using PBS before determining the protein concentration by performing a bicinchoninic acid assay (BCA, Solarbio, Beijing, China). GST-HpaXcm was blotted using a polyclonal antibody against GST and a goat anti-rabbit IgG-HRP antibody.
Hypersensitive response (HR) assays. An HR experiment was conducted on the completely unfolded leaves of 7-8-week-old tobacco plants. Proteins (10 μM) were injected into tobacco leaves, with PBS used as a negative control, using methods that have been previously described 10,34 . Scabs were observed after 2 d. The ratio of necrotic area to injection area represents morbidity. The areas of injection and necrosis were calculated using ImageJ software. Experiments were repeated three times and 10 plants were treated with each protein.
Tobacco TMV resistance assays. The completely unfolded tobacco leaves were pretreated with proteins at concentrations of 10 μM for 16 h and then inoculated with TMV. PBS was used as a negative control 8 . The scabs were observed after 6 d. The percentage of necrotic area to leaf area represents morbidity. The areas of injection and necrosis were calculated using ImageJ software. Experiments were repeated three times and 10 plants were treated with each protein.
Root growth promotion assays. A. thaliana seeds were soaked in protein solutions (10 μM) at 28 °C for 6 h, and then transferred to MS agar medium. The plates were then placed vertically in a 24 °C incubator with a 10 h night cycle 21,22 . Seeds soaked in PBS were used as a negative control. The root length and the fresh weight of the germinated seeds were recorded after 15 d. Experiments were repeated three times and 30 seeds were treated with each protein.
qRT-PCR. HpaG-Xcm proteins (10 μM) were pretreated for 10 min at 28 °C, 100 °C, 150 °C, or 200 °C and then cooled to room temperature before injecting the proteins into whole tobacco leaves. Leaves were collected at 0, 1, 3 and 6 h; the leaves collected at 0 h were used as a control. The RNA of leaves subjected to the different treatments was extracted using an RNAprep Pure Plant Kit (TIANGEN, Beijing, China) and TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TRANS, Beijing, China) was used to reverse transcript cDNA. The relative expression of the HR-related genes Hin1 16 and HSR203J 17 , the defence-related genes NPR1 and PR-1a 11 , and the growth promotion-related gene NtEXP6 18 were detected by quantitative real-time PCR (qRT-PCR) using SuperReal PreMix Plus (SYBR Green) (TRANS, China). The relative expression of genes was calculated using the 2 −ΔΔCT formula with EF-1a as a control for normalization. The following procedure was used for the PCR reaction: pre-denaturation at 95 °C for 10 min, followed by 40 cycles of amplification (95 °C for 10 s, 60 °C for 35 s, 72 °C for 20 s) 22 . The primers are listed in Supplemental Table S1. Experiments were repeated three times and 10 plants were treated with each protein.
Data processing. All data were counted and analysed using SPSS software. The values shown in the results were the means and standard deviations (SD) of three repetitions. Significant differences were analysed using multiple comparisons (LSD method, *P < 0.05, **P < 0.01).