A Truncated TIR-NBS Protein TN10 Pairs with Two Clustered TIR-NBS-LRR Immune Receptors and Contributes to Plant Immunity in Arabidopsis

The encoding genes of plant intracellular nucleotide-binding site (NBS) and leucine-rich repeat (LRR) domain receptors (NLRs) often exist in the form of a gene cluster. Several recent studies demonstrated that the truncated Toll/interleukin-1 receptor-NBS (TIR-NBS) proteins play important roles in immunity. In this study, we identified a large TN gene cluster on Arabidopsis ecotype Col-0 chromosome 1, which included nine TN genes, TN4 to TN12. Interestingly, this cluster also contained two typical TIR-NBS-LRR genes: At1g72840 and At1g72860 (hereinafter referred to as TNL40 and TNL60, respectively), which formed head-to-head genomic arrangement with TN4 to TN12. However, the functions of these TN and TNL genes in this cluster are still unknown. Here, we showed that the TIR domains of both TNL40 and TNL60 associated with TN10 specifically. Furthermore, both TNL40TIR and TNL60TIR induced cell death in Nicotiana tabacum leaves. Subcellular localization showed that TNL40 mainly localized in the cytoplasm, whereas TNL60 and TN10 localized in both the cytoplasm and nucleus. Additionally, the expression of TNL40, TNL60, and TN10 were co-regulated after inoculated with bacterial pathogens. Taken together, our study indicates that the truncated TIR-NBS protein TN10 associates with two clustered TNL immune receptors, and may work together in plant disease resistance


Introduction
The growth environment of plants also contains a range of microorganisms, including viruses, bacteria, fungi, and oomycetes, some of which infect plants to exploit nutrition [1]. In addition to non-host resistance, including natural physical barriers and constitutive synthesis of various enzymes and some secondary metabolites, plants also rely on the innate immune system to defend against pathogens [2][3][4][5]. The innate immune system of plants is mainly divided into two layers. Pattern recognition receptors (PRRs) at the plasma membrane recognize the conserved components of pathogens, pathogen-associated molecular patterns (PAMPs), or microbe-associated molecular patterns (MAMPs), and activate the first layer of immune responses, termed pattern-triggered immunity (PTI) [2,6,7]. The plant PRRs can be classified into two types: receptor-like kinases (RLKs) and receptor-like proteins (RLPs) based on the presence or absence of a kinase domain, respectively [7,8]. The well-studied RLKs in Arabidopsis include FLAGELLIN SENSING 2 (FLS2) and EF-TU RECEPTOR (EFR), which recognize conserved 22-amino acids peptide of bacterial flagellin (flg22) and the bacterial elongation factor TU (EF-TU), respectively [9][10][11][12][13][14].
Some host-adapted pathogens can secrete effectors that contribute to pathogen virulence into plant cells to interfere with plant immunity. Plants have evolved corresponding intracellular nucleotide-binding site (NBS) and leucine-rich repeat (LRR) domain receptors (NLRs), which are also called resistance (R) proteins that specifically recognize some effectors directly or indirectly and activate the second layer of immune responses, termed effector-triggered immunity (ETI) [2,15,16]. Compared with PTI, the response of ETI is often more intense and rapid, usually followed by hypersensitive responses (HR), a type of programmed cell death at the infection site [17,18].
In plants, the typical NLR protein often contains a specific N-terminal domain, a conserved NBS domain, and a highly variable LRR domain at the C-terminus [19,20]. The specific N-terminal domain is typically a coiled-coil (CC) domain, a Toll/interleukin receptor (TIR) domain or a RESISTANCE TO POWDERY MILDEW 8 (RPW8)-like domain. According to the difference of their N-terminal domain, NLRs are mainly divided into three subfamilies: CC-NBS-LRR (CNL), TIR-NBS-LRR (TNL), and RPW8-NBS-LRR (RNL) [15,18,21]. In addition to the typical NLRs containing all three characteristic domains, there are also some truncated NLRs [22,23]. For example, 21 truncated TIR-NBS (TN) proteins lacking C-terminal LRR domains are identified in Arabidopsis ecotype Col-0 [18,[21][22][23]. Several recent studies demonstrated that TNs play an important role in plant immunity as well. For instance, the loss of function of TN8 and TN11 genes leads to enhanced susceptibility to Pseudomonas syringae pv. tomato (Pto) strain DC3000 [24]. Ectopic expression of the TIR domain of TN2 triggers cell death in tobacco. The autoimmune responses in exo70B1 mutants are caused by the activation of TN2 and co-expression EXO70B1 with TN2 can suppress the TN2-induced cell death in tobacco, indicating that TN2 may monitor EXO70B1 integrity [25,26]. Moreover, TN2 interacts with CALCIUM-DEPENDENT PROTEIN KI-NASE 5 (CPK5) and stabilizes the kinase activity of CPK5 [27]. TN13 is reported to be required for resistance against Pto DC3000 lacking the effectors AvrPto or AvrptoB [28]. The gain-of-function mutant of TN1, also known as CHILLING SENSITIVE 1 (CHS1), displays chilling sensitivity at low temperatures, and the chilling sensitive phenotypes of chs1-2 mutants at low temperatures are dependent on the full-length TNL protein SUPPRESSOR OF chs1-2, 3 (SOC3) [29][30][31]. SOC3 is found to associate with TN1 and TN2, respectively, to monitor the homeostasis of Arabidopsis E3 ubiquitin ligase SENESCENCE-ASSOCIATED E3 UBIQUITIN LIGASE1 (SAUL1). The over-accumulation of SAUL1 is monitored by the SOC3-TN2 pair, while SAUL1 disappearance is guarded by the SOC3-CHS1 pair [32,33].
Interestingly, SOC3 and CHS1/TN2 show a head-to-head orientation arrangement on the gene cluster [31,33]. The head-to-head clusters are frequently found for the plant NLR genes. For instance, RPS4-RRS1 from Arabidopsis is one of the best characterized head-to-head NLR pairs. RPS4 interacts with RRS1 through their TIR domains [34]. The WRKY domain of RRS1 was shown to be able to recognize different effectors to activate ETI responses. The ability of RPS4-RRS1 to activate the defense responses depends on both proteins being present and functional [35]. CHS3 and CSA1 is another head-to-head NLR pair in Arabidopsis [36]. Autoimmunity in the chs3-2D mutants requires the presence of a functional CSA1 protein [37], suggesting a cooperative activation mechanism requiring both CSA1 and CHS3. In rice, the head-to-head NLR pairs are also identified, such as RGA4-RGA5 pair [38] and Pik-1-Pik-2 pair [39], indicating that the head-to-head arrangement of NLRs appears to be a conserved mechanism across species. The possible reason for this arrangement is that the paired NLRs are beneficial in co-regulation and co-evolution. Therefore, the investigation of the interaction between NLRs in gene clusters may reveal the functional link of unobserved NLR protein pairs [40].
In this study, through an analysis of the Arabidopsis ecotype Col-0 genome sequences, we found a large TN-type gene cluster on chromosome 1. This cluster contains nine TN  genes named TN4, TN5, TN6, TN7, TN8, TN9, TN10, TN11, and TN12. Interestingly, two full-length TNL genes, TNL40 and TNL60, were located at this gene cluster. Here, we show that TN10 interacted with TNL40 and TNL60. Transient expression of both TIR domains of TNL40 and TNL60 triggered cell death in N. tabacum. TNL40 localized in the cytoplasm, while TN10 and TNL60 localized in both the cytoplasm and nucleus. The co-expression of TNL40, TNL60, and TN10 was observed after the treatment of Pto DC3000 and Pto DC3000 avrRpt2. Our study suggested that TN10 may contribute to plant immunity by associating with both TNL40 and TNL60.

Results
2.1. The TN4-TN12 Gene Cluster Contains Two Typical TNL Genes: TNL40 and TNL60 The genome of Arabidopsis ecotype Col-0 contains 21 TN genes that are encoded by 20 genes and 1 pseudogene [23]. We found that the largest TN cluster was located on chromosome 1. As shown in Figure 1, this gene cluster contained nine TN genes, TN4-TN12, and all of these TN genes were arranged in the same direction. It was reported that TN10 and TN11 shared the highest homology with TN8 and TN9, respectively [23]. Among these TN genes, the transcripts of TN7 and TN10 were reported to be higher than those of other TNs [22]. Interestingly, we observed that there were two full-length TNL genes, TNL40 and TNL60, in this TN gene cluster (Figure 1), suggesting that the function of TN4-TN12 genes in plant immunity may be related to TNL40 and TNL60. Furthermore, TNL40 and TNL60 formed a head-to-head genomic arrangement with TN4-TN12 (Figure 1), which was reminiscent of the head-to-head arrangement of TNL gene SOC3 and TN genes TN1 and TN2 in the genome [31]. As both the TN1 and TN2 paired with SOC3 to monitor the homeostasis of SAUL1 [32,33], we hypothesized that TNL40 and TNL60 may also employ these TN proteins to monitor different substrates and activate downstream immune responses. of TNL40 and TNL60 triggered cell death in N. tabacum. TNL40 localized in the cytoplasm, while TN10 and TNL60 localized in both the cytoplasm and nucleus. The co-expression of TNL40, TNL60, and TN10 was observed after the treatment of Pto DC3000 and Pto DC3000 avrRpt2. Our study suggested that TN10 may contribute to plant immunity by associating with both TNL40 and TNL60.

The TN4-TN12 Gene Cluster Contains Two Typical TNL Genes: TNL40 and TNL60
The genome of Arabidopsis ecotype Col-0 contains 21 TN genes that are encoded by 20 genes and 1 pseudogene [23]. We found that the largest TN cluster was located on chromosome 1. As shown in Figure 1, this gene cluster contained nine TN genes, TN4-TN12, and all of these TN genes were arranged in the same direction. It was reported that TN10 and TN11 shared the highest homology with TN8 and TN9, respectively [23]. Among these TN genes, the transcripts of TN7 and TN10 were reported to be higher than those of other TNs [22]. Interestingly, we observed that there were two full-length TNL genes, TNL40 and TNL60, in this TN gene cluster (Figure 1), suggesting that the function of TN4-TN12 genes in plant immunity may be related to TNL40 and TNL60. Furthermore, TNL40 and TNL60 formed a head-to-head genomic arrangement with TN4-TN12 ( Figure  1), which was reminiscent of the head-to-head arrangement of TNL gene SOC3 and TN genes TN1 and TN2 in the genome [31]. As both the TN1 and TN2 paired with SOC3 to monitor the homeostasis of SAUL1 [32,33], we hypothesized that TNL40 and TNL60 may also employ these TN proteins to monitor different substrates and activate downstream immune responses. Exons and introns are represented by black boxes and solid lines, respectively. Arrows indicate the orientation of genes. The long bar at the bottom represents Arabidopsis chromosome 1. The gray dot represents the position of the centromere and the red dot represents the location of the indicated TN gene cluster. The red bars and numbers represent the boundaries of these TN gene cluster.

TN10 Interacts with TNL40 and TNL60
To investigate the functional link between these TNs and two TNL proteins, we first tested their interaction by performing firefly split-luciferase complementation (LUC) assays. As the activation and over-accumulation of TNL proteins may be toxic to plants, we generated a series of truncations of TNL40 and TNL60 based on their protein domains, including TNL40TIR, TNL40NBS, TNL40LRR, TNL60TIR, TNL60NBS, and TNL60LRR, Exons and introns are represented by black boxes and solid lines, respectively. Arrows indicate the orientation of genes. The long bar at the bottom represents Arabidopsis chromosome 1. The gray dot represents the position of the centromere and the red dot represents the location of the indicated TN gene cluster. The red bars and numbers represent the boundaries of these TN gene cluster.

TN10 Interacts with TNL40 and TNL60
To investigate the functional link between these TNs and two TNL proteins, we first tested their interaction by performing firefly split-luciferase complementation (LUC) assays. As the activation and over-accumulation of TNL proteins may be toxic to plants, we generated a series of truncations of TNL40 and TNL60 based on their protein domains, including TNL40TIR, TNL40NBS, TNL40LRR, TNL60TIR, TNL60NBS, and TNL60LRR, respectively. In this LUC assay, the MYC-tagged N terminus of firefly luciferase (Nluc) was fused to the truncations of TNL40 and TNL60. Moreover, the C terminus (Cluc) was fused to each TN, except TN6 whose full-length coding sequence (CDS) was not cloned from Arabidopsis leaves, respectively. Different pairs of constructs were transiently expressed in 4-week-old N. benthamiana leaves and the luminescence signals were observed two days post inoculation (dpi). As shown in Supplementary Figure S1, we observed strong luminescence signals in both leaves co-expressing TN10 with TNL40TIR or TNL60TIR. We also confirmed the correct expression of these fusion proteins in the N. benthamiana leaves (Supplementary Figure S2), suggesting that TN10 may be the unique TN protein in this cluster that associated with both TNL40 and TNL60.
The TIR domain of the TNL protein is an intracellular signaling domain that triggers immunity, and the homo-or hetero-dimerization of TIR domains is crucial for immune signaling transduction [17,18,34]. To further confirm the interaction between TNs and TNL40 or TNL60, we next fused the TIR domain of each TN protein to Cluc and performed LUC assays with TNL40TIR-Nluc and TNL60TIR-Nluc, respectively. As shown in Supplementary Figure S3, we only observed strong luminescence signals in leaves coexpressing TNL40TIR-Nluc or TNL60TIR-Nluc with TN10TIR-Cluc, but not with other TNs Cluc-tagged TIR proteins, indicating that TNL40 and TNL60 mainly interacts with TN10. Therefore, we focused on TN10 mainly in the further investigation.
To confirm the interaction between TN10 and TNL40TIR or TNL60TIR, we next performed the LUC assay with appropriate controls. In this assay, we fused yellow fluorescence protein (YFP) to Cluc and Nluc, respectively, to act as negative controls. After co-expression of different pairs in N. benthamiana leaves, we found that TN10-Cluc associated with both TNL40TIR-Nluc and TNL60TIR-Nluc, but not with YFP-Nluc, and there is no interaction between YFP-Cluc and TNL40TIR-Nluc or TNL60TIR-Nluc ( Figure 2A). We also confirmed the correct expression of all these fusion proteins (Supplementary Figure S4). These data indicate that TN10-Cluc associates with TNL40TIR-NLuc and TNL60TIR-Nluc specifically.
To further confirm this result, we next performed a co-immunoprecipitation (Co-IP) assay in which we transiently expressed GFP, TNL40TIR-GFP, or TNL60TIR-GFP with TN10-MYC in 4-week-old Arabidopsis protoplasts ( Figure 2B). When we immunoprecipitated GFP, TNL40TIR-GFP, and TNL60TIR-GFP proteins using an anti-GFP antibody, TN10-MYC protein was detected with an anti-MYC antibody only in the precipitate from the leaves that co-expressed both TN10-MYC and TNL40TIR-GFP or TNL60TIR-GFP, but not from the negative control leaves that expressed TN10-MYC and GFP ( Figure 2B), indicating again that TN10 can form a complex with TNL40TIR or TNL60TIR in Arabidopsis.

TN10 Contains A Typical TIR Domain
To gain more insights into the TN10 protein, we performed the alignments of protein amino acid sequences with the well-characterized TNL protein L6 and RESISTANCE TO PSEUDOMONAS SYRINGAE 4 (RPS4), as well as TN2. We found that all TN10, TNL40, and TNL60 contained a typical TIR domain (Supplementary Figure S5), including four conserved TIR subdomains (TIR1-TIR4) [30,36]. Several recent studies demonstrated that the TIR domains of TNL proteins were able to cleave nicotinamide adenine dinucleotide in its oxidized form (NAD + ) and transmit signals downstream to trigger cell death [41][42][43]. Furthermore, a conserved glutamic acid (Glu, E) was found to be required for the activity of plant TIR domain to act as a NAD + -cleaving enzyme (NADase) [41]. As shown in Supplementary Figure S5, we also found that the conserved Glu was present in the TIR2 subdomain of all aligned TIR-type proteins, indicating again that TN10, TNL40, and TNL60 carry typical TIR domains.

TN10 Contains A Typical TIR Domain
To gain more insights into the TN10 protein, we performed the alignments of protein amino acid sequences with the well-characterized TNL protein L6 and RESISTANCE TO PSEUDOMONAS SYRINGAE 4 (RPS4), as well as TN2. We found that all TN10, TNL40, and TNL60 contained a typical TIR domain (Supplementary Figure S5), including four conserved TIR subdomains (TIR1-TIR4) [30,36]. Several recent studies demonstrated that the TIR domains of TNL proteins were able to cleave nicotinamide adenine dinucleotide in its oxidized form (NAD + ) and transmit signals downstream to trigger cell death [41][42][43]. Furthermore, a conserved glutamic acid (Glu, E) was found to be required for the activity of plant TIR domain to act as a NAD + -cleaving enzyme (NADase) [41]. As shown in Supplementary Figure S5, we also found that the conserved Glu was present in the TIR2 subdomain of all aligned TIR-type proteins, indicating again that TN10, TNL40, and TNL60 carry typical TIR domains.
Moreover, we found that the motifs defining the NBS, ARC1 (Apaf-1, R proteins, CED-4, 1), and ARC2 subdomains were conserved in TNL40, TNL60, L6, and RPS4 (Supplementary Figure S5) [44]. As typical TNL proteins, TNL40 and TNL60 also contained GKT-type Walker A/P-loop motifs, a conserved GNNNNGKT sequence (N represents any Moreover, we found that the motifs defining the NBS, ARC1 (Apaf-1, R proteins, CED-4, 1), and ARC2 subdomains were conserved in TNL40, TNL60, L6, and RPS4 (Supplementary Figure S5) [44]. As typical TNL proteins, TNL40 and TNL60 also contained GKT-type Walker A/P-loop motifs, a conserved GNNNNGKT sequence (N represents any amino acid residue) that is essential for the binding of ATP [45,46] in their NBS domain, while it is a GRS-type P-loop motif (GNNNNGRS) in TN2 (Supplementary Figure S5). In addition, the conserved MHD (methionine-histidine-aspartate) motif required for function/activity of TNL was also observed in the NBS domains of TNL40 and TNL60 [44,47], but not in the NBS domains of TN2 and TN10 (Supplementary Figure S5), indicating again that TNL40 and TNL60 are typical TNL proteins. Different from other TN proteins, we found that TN10 not only lacks LRR domains, but also misses most of the NBS domain sequences, including the P-loop motif in its NBS domain (Supplementary Figure S5), suggesting that the function of TN10 in plant immunity may be different from other TN proteins.

Subcellular Localization of TNL40, TNL60, and TN10
In order to explore the function of TNL40, TNL60, and TN10, we next examined TNL40-GFP, TNL60-GFP, and TN10-GFP subcellular localization in N. benthamiana leaves. The expression of GFP alone was used as a negative control. Consistent with the previous report [22], we observed that TN10-GFP localized in both the cytoplasm and nucleus ( Figure 3). Interestingly, TNL40 mainly localized in the cytoplasm (Figure 3 and Supplementary Figure S6), while TNL60 localized in the cytoplasm and nucleus ( Figure 3). All proteins of the correct size without free GFP were detected by an immunoblotting analysis with an anti-GFP antibody (Supplementary Figure S7).
while it is a GRS-type P-loop motif (GNNNNGRS) in TN2 (Supplementary Figure S5). In addition, the conserved MHD (methionine-histidine-aspartate) motif required for function/activity of TNL was also observed in the NBS domains of TNL40 and TNL60 [44,47], but not in the NBS domains of TN2 and TN10 (Supplementary Figure S5), indicating again that TNL40 and TNL60 are typical TNL proteins. Different from other TN proteins, we found that TN10 not only lacks LRR domains, but also misses most of the NBS domain sequences, including the P-loop motif in its NBS domain (Supplementary Figure S5), suggesting that the function of TN10 in plant immunity may be different from other TN proteins.

Subcellular Localization of TNL40, TNL60, and TN10
In order to explore the function of TNL40, TNL60, and TN10, we next examined TNL40-GFP, TNL60-GFP, and TN10-GFP subcellular localization in N. benthamiana leaves. The expression of GFP alone was used as a negative control. Consistent with the previous report [22], we observed that TN10-GFP localized in both the cytoplasm and nucleus (Figure 3). Interestingly, TNL40 mainly localized in the cytoplasm (Figure 3 and Supplementary Figure S6), while TNL60 localized in the cytoplasm and nucleus ( Figure 3). All proteins of the correct size without free GFP were detected by an immunoblotting analysis with an anti-GFP antibody (Supplementary Figure S7).

Overexpression of TNL40TIR and TNL60TIR Induced Cell Death in N. tabacum
Overexpression of TNL proteins or their TIR domains in N. tabacum leaves often leads to cell death [18,41]. However, whether these TN proteins can trigger cell death is unclear. Then, with TN2 as a positive control, we transiently expressed TN4, TN5, TN7, TN8, TN9, TN10, TN11, and TN12 protein in N. tabacum leaves. Four days after injection, we observed that the expression of TN2 induced significant cell death, which is consistent with

Overexpression of TNL40TIR and TNL60TIR Induced Cell Death in N. tabacum
Overexpression of TNL proteins or their TIR domains in N. tabacum leaves often leads to cell death [18,41]. However, whether these TN proteins can trigger cell death is unclear. Then, with TN2 as a positive control, we transiently expressed TN4, TN5, TN7, TN8, TN9, TN10, TN11, and TN12 protein in N. tabacum leaves. Four days after injection, we observed that the expression of TN2 induced significant cell death, which is consistent with a previous report [26]. Similarly, we found that TN11 and TN12, but not other TNs in this cluster, can also trigger obvious cell death ( Figure 4A), and the TN4, TN5, TN7, TN8, TN9, and TN10 proteins were accumulated at the correct size ( Figure 4B). In addition, we also observed that the expression of the TIR domains, but not the NBS or LRR domains, of TNL40 and TNL60 can trigger cell death in N. tabacum leaves ( Figure 4C,D).
Interestingly, we found that the cell death triggered by TNL40TIR was weaker than that by TNL60TIR. To further test this, we next expressed TNL40TIR-GFP, TNL60TIR-GFP, and TN10-GFP in N. tabacum leaves and observed the occurrence of cell death. As shown in Figure 5A, we found that the TNL60TIR triggered cell death was obvious 3 days after injection, while the TNL40TIR induced cell death was still not very obvious, indicating that the hypersensitive response activated by TNL40 was weaker than TNL60 ( Figure 5A). Additionally, we also tested the accumulation of protein levels of TNL40TIR, TNL60TIR, and TN10, and found all proteins were expressed and accumulated ( Figure 5B), indicating again that TNL40 and TNL60 carry typical and functional TIR domains.
observed that the expression of the TIR domains, but not the NBS or LRR domains, of TNL40 and TNL60 can trigger cell death in N. tabacum leaves ( Figure 4C,D).
Interestingly, we found that the cell death triggered by TNL40TIR was weaker than that by TNL60TIR. To further test this, we next expressed TNL40TIR-GFP, TNL60TIR-GFP, and TN10-GFP in N. tabacum leaves and observed the occurrence of cell death. As shown in Figure 5A, we found that the TNL60TIR triggered cell death was obvious 3 days after injection, while the TNL40TIR induced cell death was still not very obvious, indicating that the hypersensitive response activated by TNL40 was weaker than TNL60 ( Figure  5A). Additionally, we also tested the accumulation of protein levels of TNL40TIR, TNL60TIR, and TN10, and found all proteins were expressed and accumulated ( Figure  5B), indicating again that TNL40 and TNL60 carry typical and functional TIR domains.

Expression of TNL40, TNL60, and TN10 after Infected with Bacterial Pathogens
The pattern of expression of TNL40, TNL60, and TN10 under inoculation with pathogens may provide some clues to their cellular roles. In order to explore the role and functional link of TNL40, TNL60, and TN10 in disease resistance, we next inoculated the 4week-old Arabidopsis wild-type Col-0 leaves with Pto DC3000, Pto DC3000 avrRpt2, and Pto DC3000 avrRps4, respectively, and analyzed the mRNA expression levels of TNL40,

Expression of TNL40, TNL60, and TN10 after Infected with Bacterial Pathogens
The pattern of expression of TNL40, TNL60, and TN10 under inoculation with patho gens may provide some clues to their cellular roles. In order to explore the role and functional link of TNL40, TNL60, and TN10 in disease resistance, we next inoculated the 4-week-old Arabidopsis wild-type Col-0 leaves with Pto DC3000, Pto DC3000 avrRpt2, and Pto DC3000 avrRps4, respectively, and analyzed the mRNA expression levels of TNL40, TNL60, and TN10 by quantitative reverse transcription (qRT)-PCR. The initial transcription level of TN10 was very high (Figure 6), which is consistent with the previous reports [22,23]. We found that both TNL genes, TNL40 and TNL60, were up-regulated after being infected with Pto DC3000, Pto DC3000 avrRpt2, and Pto DC3000 avrRps4 (Figure 6), suggesting that TNL40 and TNL60 may play a role after pathogens invasion. Interestingly, we found that the transcripts of TN10 increased significantly six hours after treatment with Pto DC3000 and Pto DC3000 avrRpt2, but decreased significantly six hours after treatment with Pto DC3000 avrRps4 (Figure 6), indicating that TN10 plays different roles in the resistance response mediated by different effectors. However, different to the inoculation with Pto DC3000 avrRpt2, we observed that the up-regulated expression of TN10 was dropped to initial level 12 h after inoculation with Pto DC3000 ( Figure 6A,B), suggesting that TN10 may only play a role in the resistance to Pto DC3000 at the early time of infection.  DC3000 avrRpt2 (B), and Pto DC3000 avrRps4 (C). The y-axis represents the relative expression level that was normalized to those of ACTIN2, and the x-axis represents the indicated time point (hours) after inoculation. Lower-case letters indicated statistically significant differences (p < 0.05, ANOVA, n = 3).

Discussion
The head-to-head arranged NLR gene pairs have been reported in many studies [33,38,40,[48][49][50][51][52]. In this article, we found that there was a large TN-TNL gene cluster on Arabidopsis chromosome 1 and showed the association of TN10 with TNL40TIR and TNL60TIR. As far as we know, this is the first report that an atypical TN protein was shown to interact with the TIR domain of two full-length TNL proteins. Furthermore, we tested the subcellular localization of TNL40, TNL60, and TN10, and showed that TNL40 mainly localized in the cytoplasm, while TN10 and TNL60 localized in both the cytoplasm Figure 6. Expression of TN10, TNL40, and TNL60 genes in Arabidopsis Col-0 plants after infiltration with different Pto DC3000 strains. The expression levels of TN10, TNL40, and TNL60 genes were measured using qRT-PCR after inoculation with Pto DC3000 (A), Pto DC3000 avrRpt2 (B), and Pto DC3000 avrRps4 (C). The y-axis represents the relative expression level that was normalized to those of ACTIN2, and the x-axis represents the indicated time point (hours) after inoculation. Lower-case letters indicated statistically significant differences (p < 0.05, ANOVA, n = 3).

Discussion
The head-to-head arranged NLR gene pairs have been reported in many studies [33,38,40,[48][49][50][51][52]. In this article, we found that there was a large TN-TNL gene cluster on Arabidopsis chromosome 1 and showed the association of TN10 with TNL40TIR and TNL60TIR. As far as we know, this is the first report that an atypical TN protein was shown to interact with the TIR domain of two full-length TNL proteins. Furthermore, we tested the subcellular localization of TNL40, TNL60, and TN10, and showed that TNL40 mainly localized in the cytoplasm, while TN10 and TNL60 localized in both the cytoplasm and nucleus. In addition, the observation of co-regulation of transcripts of TN10 with TNL40 and TNL60 further indicated that TN10 may form complexes with TNL40 and TNL60 to regulate plant immunity.
The head-to-head genomic arrangements play important roles in regulating the cooperation of NLR proteins as almost all of the identified paired NLRs are encoded by physically linked genes [40]. For example, the SOC3-CHS1 pair and SOC3-TN2 pair are the two firstly identified unique NLR pairs formed by a full-length TNL and a truncated TN protein [33,49]. The over-accumulation and disappearance of Arabidopsis E3 ligase SAUL1 are monitored by SOC3-TN2 and SOC3-TN1, respectively [32,33]. SOC3, CHS1, and TN2 genes also show head-to-head genomic arrangement on chromosome 1 [31,33,49]. These reports suggested that the analysis of the interaction of NLR proteins in gene clusters may reveal new NLR protein pairs [35,[38][39][40]51]. We next analyzed the Arabidopsis genome and found that the largest TNs gene cluster was located on chromosome 1. Two typical TNL genes were found in this cluster, suggesting that they may function together in plant immunity. Interestingly, in our further investigation, we found that only TN10 in this cluster interacted strongly with both TIR domains of TNL40 and TNL60, indicating that TN10 may transmit signals to TNL40 and TNL60 to regulate plant immunity. Why other TNs in this cluster did not obviously interact with TNL40 or TNL60 and what targets were guarded by TNL40-TN10 and TNL60-TN10 pairs needs to be further investigated.
In plants, the TIR domain of TNL plays a crucial role in immune signaling since ectopic expression of certain TIR domain triggers strong cell death in tobacco [31,53]. Recent studies showed that, similar to the TIR domain of animals, the plant TIR domain also has NADase activity, which is required for cell death induction [41,42]. As shown in Supplementary Figure S5, we found that both the TIR domains of TNL40 and TNL60 contained the conserved glutamic acid, which is necessary for their potential NADase activity, in the putative catalytic residue. Consistent with this, both TNL40TIR and TNL60TIR could trigger cell death in N. tabacum leaves. Interestingly, although the alignments of protein sequences showed that TN10 contained the typical TIR domain and the conserved glutamic acid in its putative catalytic residue, we did not observe obvious cell death in N. tabacum leaves expressing TN10 in our experimental conditions. However, the TN10 protein was accumulated in N. tabacum leaves. Moreover, it is noteworthy that we found very high transcripts of TN10 in the untreated plants, which is consistent with a previous report [22,23]. Together with the fact that TN10 protein misses most of the NBS domain sequences, including P-loop and MHD motifs, we proposed that TN10 may mainly function in transmitting immune signals. Furthermore, it is worth noting that the well-characterized TIR-only functional immune receptor RESPONSE TO HOPBA 1 (RBA1), which is sufficient to trigger cell death in response to HopBA1, functions as a pathogen sensor [54]. It is possible that TN10, as a TIR-only receptor, may be functionally similar to RBA1 in effector recognition and immune activation.
Similar to the pentamerized resistosome formed by the CNL protein ZAR1 [19,55], two more recent studies showed that the TNL proteins RPP1 and ROQ1 can form the tetramerized resistosome required for their NADase activity [56,57]. The existence of TNL resistosome may serve as a platform to recruit TIR-only proteins or other oligomerized TNLs to form complexes and expand TNL-mediated immune signals [56]. Similarly, whether TNL40 and TNL60 can form tetramerized resistosome remains to be determined. It is possible that TN10 may also participate in the process of TNL tetramerization resistosome.
The prominent TIR domain in the resistosome enhances its NADase enzyme activity, so it will be interesting to determine whether TN10 affects the prominent TIR domain in the TNL resistosome. In addition, whether TN10 can form a tetramerization by its TIR domain is also an interesting question to be determined.
The mRNA expression levels of TN10, TNL40, and TNL60 were co-regulated after inoculation with Pto DC3000 and Pto DC3000 avrRpt2, but not with Pto DC3000 avrRps4, suggesting that TN10 may function together with TNL40 and TNL60 in response to different effectors. Interestingly, we found that the up-regulation of TN10 transcripts significantly occurred six hours after injection with Pto DC3000 and Pto DC3000 avrRpt2, but the expression of TNL40 and TNL60 was obviously increased 12 h after injection with Pto DC3000 and Pto DC3000 avrRpt2, indicating that the up-regulation of TNL40 and TNL60 is later than TN10, and TN10 may sense the attack of bacteria and transmit the signals to TNL40 and TNL60 to activate immunity. The expression trend of TN10 after inoculation with Pto DC3000 is similar to the treatment of salicylic acid (SA) [22], indicating that TN10 plays a positive role in the early stage of immune responses, which is consistent with the previous report that the TN10 overexpression plants display enhanced resistance to Pto DC3000 [22]. However, it will also be interesting to investigate why the transcripts of TN10 are induced by Pto DC3000 avrRpt2 as the effector AvrRpt2 is recognized by the well-characterized CNL RPS2.
Arabidopsis ecotype Col-0 has only about 82 TNLs and 51 CNLs, which are sufficient to deal with the invasion and colonization of most external pathogens [23]. How plants use limited NLR proteins to resist the complex environment has become a key issue. Previous studies have suggested that the proteins encoded by the clustered NLRs may act together in activating immune responses and may be likely to provide coregulatory benefits [40]. Our results here showed that the atypical TN protein TN10 can interact with TNL40TIR and TNL60TIR. Based on these results, we proposed a possible working model for the functional mechanism of TN10, TNL40, and TNL60 in plant immunity (Figure 7). In the resting cell, TNL40 and TNL60 are kept in inactive states through the interaction between their domains. When pathogens such as bacteria secrete effectors into plant cells, TN10 interacts with the TIR domain of TNL40 or TNL60, leading to the change of the conformation of the full-length TNL proteins, thereby releasing the full-length TNL to activate ETI responses. Additionally, how TN10 activates and helps the full-length TNL proteins perform functions are also interesting questions that need to be determined in the further investigation. TNL40 and TNL60 to activate immunity. The expression trend of TN10 after inoculation with Pto DC3000 is similar to the treatment of salicylic acid (SA) [22], indicating that TN10 plays a positive role in the early stage of immune responses, which is consistent with the previous report that the TN10 overexpression plants display enhanced resistance to Pto DC3000 [22]. However, it will also be interesting to investigate why the transcripts of TN10 are induced by Pto DC3000 avrRpt2 as the effector AvrRpt2 is recognized by the wellcharacterized CNL RPS2. Arabidopsis ecotype Col-0 has only about 82 TNLs and 51 CNLs, which are sufficient to deal with the invasion and colonization of most external pathogens [23]. How plants use limited NLR proteins to resist the complex environment has become a key issue. Previous studies have suggested that the proteins encoded by the clustered NLRs may act together in activating immune responses and may be likely to provide coregulatory benefits [40]. Our results here showed that the atypical TN protein TN10 can interact with TNL40TIR and TNL60TIR. Based on these results, we proposed a possible working model for the functional mechanism of TN10, TNL40, and TNL60 in plant immunity (Figure 7). In the resting cell, TNL40 and TNL60 are kept in inactive states through the interaction between their domains. When pathogens such as bacteria secrete effectors into plant cells, TN10 interacts with the TIR domain of TNL40 or TNL60, leading to the change of the conformation of the full-length TNL proteins, thereby releasing the full-length TNL to activate ETI responses. Additionally, how TN10 activates and helps the full-length TNL proteins perform functions are also interesting questions that need to be determined in the further investigation. In summary, our results here showed that TNL40, TNL60, and TN10 were located in the largest TN-TNL gene cluster in a head-to-head genomic arrangement. Transient expression of the TIR domains of TNL40 and TNL60 triggered cell death in N. tabacum Figure 7. A possible model summarizing the function of TNL40, TNL60, and TN10 proteins in plant immunity. In the resting cell, TNL40 and TNL60 are kept in inactive states through the interaction between their domains. When plants were attacked by pathogens, TN10 interacted with TNL40 or TNL60 and resulted in the changes of the conformation of TNL proteins followed by the activation of effector-triggered immunity (ETI) responses. TNL40, TNL60, and TN10 may also function as accessory components in signal transduction during defense responses.
In summary, our results here showed that TNL40, TNL60, and TN10 were located in the largest TN-TNL gene cluster in a head-to-head genomic arrangement. Transient expression of the TIR domains of TNL40 and TNL60 triggered cell death in N. tabacum leaves. Moreover, TN10 interacted with both TNL40TIR and TNL60TIR, and the observation of co-expression of TNL40, TNL60, and TN10 after infection with Pto DC3000 suggested that TN10 pairs with TNL40 and TNL60 may contribute to immune signaling.

Plant Materials and Growth Conditions
The wild-type Arabidopsis thaliana accession used in this study was Columbia-0 (Col-0), seeds were surface-sterilized by 70% ethanol and planted on plates containing a halfstrength Murashige and Skoog (1/2 MS) medium with 1% sucrose. Plates were kept in 4 • C for 3 days and moved to the growth room, and then transplanted into soil (nutrient soil/vermiculite 2: 1) 7 days after germination. Arabidopsis thaliana plants were grown in a growth room at 22 • C with approximately 60% relative humidity and 9-h-light/15-h-dark photoperiod, with a light intensity of 7000-8000 lux for protoplasts preparation and gene expression assay. Nicotiana benthamiana and Nicotiana tabacum plants were grown under the same growth conditions as Arabidopsis [58].

Plasmid Constructs
All CDS sequences containing TN4-TN12, TNL40, and TNL60 used in this research were derived from the Arabidopsis thaliana Col-0 accession. Plasmids used in this study were constructed by In-fusion technology's One-Step Cloning Kit (Vazyme, Nanjing, China. C112). All clones were verified with DNA sequencing. For the firefly split-luciferase complementation assay, the CDS of TN4, TN5, TN7, TN8, TN9, TN10, TN11, TN12, TN4TIR,  TN5TIR, TN7TIR, TN8TIR, TN9TIR, TN11TIR, and TN12TIR were amplified with specific primers and then cloned into pCAMBIA1300-Cluc, respectively. The CDS of TNL40TIR, TNL40NBS, TNL40LRR, TNL60TIR, TNL60NBS, and TNL60LRR were amplified with specific primers and then cloned into pCAMBIA1300-Nluc-MYC, respectively. For the Co-IP assay, TNL40TIR and TNL60TIR was cloned into pSuper1300 with a GFP tag, TN10 was cloned into pSuper1300 with a 6 × MYC tag. For the subcellular localization assay, the fulllength TNL40, TNL60 and TN10 CDS sequences were amplified and then cloned into pSuper1300 with a GFP tag. The primers used for the plasmid constructs are listed in Supplementary Table S1.

Cell Death Assay in N. tabacum Leaves and Tobacco Protein Extraction
Agrobacterium strain GV3101 carrying the targeted plasmid was resuspended to OD 600 = 0.4 in an injection buffer (10 mM MES [pH 5.6], 10 mM MgCl 2 , 150 µM Acetosyringone), then gently injected into the tobacco leaf using a 1-mL needle-free syringe. The tobacco was cultivated in the dark for 12 h, then transferred to the short-day growth conditions. For cell death leaf observation experiments, the pictures were taken at 3 or 4 days after injection. Then, the indicated N. tabacum leaves were stained with trypan blue and decolorized overnight. For protein extraction experiments, the samples were taken 36 h after injection (before cell death occurrence). The leaves were frozen quickly in liquid nitrogen and ground into powders in a mortar. Then, the sample was transferred to a new 2-mL centrifuge tube. Native extraction buffer (50 mM Tris-MES [pH 8.0], 0.5 M Sucrose, 1 mM MgCl 2 , 10 mM EDTA, 5 mM DTT, and 1% [w/v] protease inhibitor cocktail S8830 [Sigma, St. Louis, MO, USA]) was added in a ratio of 1 g to 2 mL. The tubes were placed on ice for 30 min and centrifuged at 12,000 g for 15 min, and then the supernatant was transferred to a new centrifuge tube to be used in the next protein immunoblotting assay performed as described previously [59].

Firefly Split-Luciferase Complementation Assay
Pairs of Agrobacterium GV3101 strains containing the targeted plasmid combinations were resuspended in an injection buffer (the final concentration of each stain injection was OD 600 = 0.4), and then co-transformed into 4-week-old N. benthamiana leaves and incubated in the short-day condition growth room for 48 h. The N. benthamiana leaves were sprayed with 1 mM precooled luciferin, and the samples were placed in darkness for 5 min. Subsequently, LUC images were captured with a low-light cooled CCD imaging apparatus. The details of this assay were described previously [60].

Transient Expression in Arabidopsis Protoplasts
Protoplasts were prepared from 4-week-old Arabidopsis Col-0 leaves and transfections were done as described previously [61]. After transfection, protoplasts were incubated at room temperature under weak light for 16 h, and protein extracts were prepared for immunoblotting and Co-IP assays.

Protoplast Protein Extraction, Immunoprecipitation, and Immunoblotting
Protoplast proteins were processed in extraction buffer ( Aliquots of supernatants were used as input samples. For immunoprecipitation assays, 20 µL GFP-Trap Magnetic Agarose (Chromotek, Munich, Germany) was added to 1 mL total protein extract for 2 h incubation at 4 • C. Beads were collected by centrifugation and washed four times with a PBS buffer containing 0.1% (v/v) Triton X-100. Then, beads were resuspended in 80 µL of PBS buffer with SDS loading buffer and incubated at 95 • C for 8 min. Proteins were separated via SDS-PAGE gels and analyzed by immunoblotting. The anti-bodies used were anti-GFP (Sigma Aldrich, Louis, MO, USA, 11814460001) and anti-MYC (Sigma Aldrich, Louis, MO, USA, 11867423001).

Subcellular Localization Assay
The constructs were transformed into the Agrobacterium GV3101 strain and injected into N. benthamiana leaves. Plants were placed in the dark for 12 h and then transferred to the short-day growth conditions for 36 h, then the leaves were taken for observation. Green fluorescent protein (GFP) fluorescence was visualized using a confocal laser scanning microscope (Zeiss 880) under a × 20 objective, zoom × 0.8, 2048 × 2048 pixels, and imaged TN10-GFP and TNL40/60-GFP with 2 and 10% of the 488-nm line, respectively, and detected in the range 494-598 nm.

Pathogen Infection Assays and Gene Expression Analysis
Pseudomonas syringae pv. tomato (Pto) DC3000 or Pto DC3000 AvrRpt2, Pto DC3000 AvrRps4 strains were cultured on the King' B medium containing appropriate antibiotics at 28 • C for two days. The 10 mM MgCl 2 containing buffer was used to resuspend bacteria to OD 600 = 0.1. Leaves from 4-week-old plants were infiltrated with suspensions of different Pto DC3000 strains. The samples were taken at 0, 6, and 12 h after inoculation and frozen in liquid nitrogen. Total RNA was extracted by the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions, and the first strand was synthesized using a RT reagent kit (code number is RR047A; Takara, Dalian, China). Accumulation of transcripts was examined by qRT-PCR using the RT reagent Kit (code number is RR047A; Takara, Dalian, China). All qRT-PCR assays were performed with the Premix Ex Taq Kit (code number is RR420A; TaKaRa, Dalian, China) in a CFX Connect Real-time PCR System (BIO-RAD, Hercules, CA, USA). The ACTIN2 gene was used as an internal control. All of the primers used for qRT-PCR are listed in Supplementary Table S1.

Conclusions
In this study, we identified the largest TN-TNL gene cluster in the Arabidopsis ecotype Col-0 genome. Two full-length TNL genes, TNL40 and TNL60, were arranged in a head-tohead orientation with TN4-TN12. TNL40 and TNL60 interacted strongly with TN10, which is almost a naked TIR domain. TNL40, TNL60, and TN10 localized in the cytoplasm, and TN10 and TNL60 also localized in the nucleus. The transcripts of TNL40, TNL60, and TN10 were co-induced when inoculated with Pto DC3000. Furthermore, both TNL40TIR and TNL60TIR induced cell death in N. tabacum leaves. Based on these data, we hypothesized that TNL40 and TNL60 were kept in inactive states in the resting cell. When bacterial pathogens attacked, TN10 interacted with TNL40 or TNL60 to change the conformation of the full-length TNLs, and then activated the ETI immune signaling.