Background

In plants, the innate immune system can be categorized into two layers: pattern-triggered immunity (PTI) and effector-triggered immunity (ETI) [1]. PTI is mediated by surface-localized pattern recognition receptors (PRRs) that can recognize pathogen-associated molecular patterns (PAMPs) of the pathogen. ETI is mediated by intracellular immune receptors, which evolve resistance (R) genes to recognize effectors of pathogens. R genes can be divided into at least five classes [2, 3], and the biggest category is nucleotide binding–leucine-rich repeats (NBS–LRRs) [4]. NBS–LRRs are distributed in various plant species. Many NBS–LRRs have been identified at the genome-wide level such as in Arabidopsis thaliana [5], Arachis duranensis [6], Arachis ipaensis [6], Glycine max [7], Medicago truncatula [8], Oryza sativa [9], and Triticum aestivum [10]. NBS–LRRs are classified into two types based on the N-terminal domain, coiled-coil (CC)–NBS–LRR (CNL) and toll/mammalian interleukin-1 receptor (TIR)–NBS–LRR (TNL) [5]. Generally, the NBS domain hydrolyzes ATP or GTP to obtain energy [2]. Overexpression of CC or TIR domains can reduce hypersensitive response in plants [11, 12]. The LRR domain undergoes more relaxed selection or positive selection because this domain interacts with pathogenic effectors [13,14,15], indicating that LRR domains are more diverse compared to NBS, TIR, and CC domains [13, 14, 16].

To date, a few studies have focused on the phylogenetic relationship of NBS–LRRs between polyploids and their donors. T. aestivum (AABBDD) is a hybrid of Aegilops tauschii (DD) and T. dicoccoides (AABB) which originated from a hybridization process between T. urartu (AA) and A. speltoides (BB) [17]. Many NBS–LRRs are extinct in T. aestivum compared to the NBS–LRRs in its donors; the evolutionary rate of NBS–LRRs of T. aestivum is also slower than that of its donors [10], causing disease resistance in T. aestivum to be lower than its donors. Similarly, Gossypium hrisutum (AADD) is a hybrid between G. raimondii (DD) and G. arboretum (AA) [18]. New NBS–LRRs are produced in G. hrisutum because of polyploidy, natural and artificial selection, gene duplication, and chromosomal recombination [19]. However, gene number and gene structure of NBS–LRRs are similar for Citrus sinensis and its donor, C. clementina [16]. Therefore, it is important to study the evolution and function between polyploids and parental donors.

NBS–LRRs involved in response to pathogens have been well documented. RFO1, WRR4, and RPW8 genes are NBS–LRRs that have been isolated from A. thaliana [20,21,22]. Functional analyses have shown that RFO1 genes provide resistance to a broad spectrum of Fusarium races [20], and RPW8 controls resistance to a broad spectrum of powdery mildew pathogens [21]. Overexpression of WRR4 in Brassica species can confer broad-spectrum white rust resistance [22]. In addition, a total of 15 NBS–LRRs from five rice cultivars have been introduced into a transgenic rice cultivar, increasing its broad-spectrum resistance to Magnaporthe oryzae [15]. In legumes, RCT1 from M. truncatula, which is classified as a TNL gene, confers broad-spectrum anthracnose resistance in transgenic susceptible alfalfa plants [23]. In Arachis, NBS–LRRs are involved in response to Aspergillus flavus and Meloidogyne arenaria infection [6, 24, 25].

Cultivated peanut (Arachis hypogaea L., AABB) is an allotetraploid hybrid between two wild peanuts, A. duranensis (AA) and A. ipaensis (BB) [26,27,28]. The complete genome sequences of A. hypogaea cv. Tifrunner and related diploids, A. duranensis and A. ipaensis, have been published [26, 29,30,31,32]. In addition, NBS–LRRs of A. duranensis and A. ipaensis have been identified and subjected to phylogenetic analyses [6]. These studies provided a powerful basis for the understanding of evolution and function of NBS–LRRs in A. hypogaea cv. Tifrunner. In this study, we identified 713 full-length NBS–LRRs in A. hypogaea cv. Tifrunner. We analyzed the sequence structure, evolution and function of NBS–LRRs in A. hypogaea cv. Tifrunner. We proposed that the low disease resistance of A. hypogaea cv. Tifrunner may be partially caused by the loss of LRR domains.

Results and discussion

NBS–LRR gene family in A. hypogaea cv. Tifrunner

We identified 1105 NBS-containing sequences using HMMER in A. hypogaea cv. Tifrunner. Among the NBS-containing sequences, 713 NBS-containing genes contained complete NBS domains and had full-length coding sequences (Additional file 1: Table S1). Previously, results were more difficult to interpret when the evolution of NBS–LRR proteins was analyzed using the incomplete NBS domain of Lotus japonicus [33]. Therefore, in our study, only 713 regular NBS–LRRs encoding intact NBS domains were used for further analyses. There are a total of 278 and 303 full-length NBS–LRRs in A. duranensis and A. ipaensis, respectively [6].

Among the 713 NBS–LRR proteins, 229 sequences contained TIR domains, and 118 sequences included CC domains (Additional file 1: Table S1). Interestingly, we found that 26 sequences contained both TIR and CC domains in A. hypogaea cv. Tifrunner (Additional file 1: Table S1). However, none of the sequences contained both TIR and CC domains in A. duranensis and A. ipaensis [6]. Previous studies have demonstrated that TNL and CNL have different origins [34,35,36]. We speculated that genetic exchange or gene rearrangement likely resulted in the fusion of the TIR and CC domains after tetraploidization. Bertioli et al. [30] found many crossovers between A and B subgenomes, and chromosome inversions were detected in A. hypogaea cv. Tifrunner. The chromosome translacations could change gene direction. In addition, we found three sequences that simultaneously contained an NBS domain and WRKY domain in A. hypogaea cv. Tifrunner. In other legumes, NBS–WRKY fusion proteins have only been identified in G. max, A. duranensis, and A. ipaensis [37]. The bacterial effectors AvrRps4 or PopP2 can trigger WRKY transcription factors that are involved in active NBS–LRR gene responses to pathogens [38]. We speculated that NBS–WRKY fusion proteins can play a crucial role in response to biotic stress in A. hypogaea cv. Tifrunner.

LRR domains play important roles in protein–ligand and protein–protein interactions; these LRR domains are involved in plant immune responses [39, 40]. In this study, we found that 348 NBS–LRR proteins contained four types of LRR domains in A. hypogaea cv. Tifrunner, namely, LRR1, LRR3, LRR4, and LRR8 (Additional file 1: Table S1). Among these sequences, the greatest number of LRR domains were classified as LRR8-type (308), followed by LRR3 (133), LRR4 (88), and LRR1 (7). A. duranensis and A. ipaensis had five types of LRR domains: LRR1, LRR3, LRR4, LRR5, and LRR8 [6]. Moreover, the greatest number of LRR domains in A. duranensis were classified as LRR8-type, followed by LRR4, LRR3, and LRR5 [6]. In A. ipaensis, the greatest number of LRR domains were classified as LRR8-type, followed by LRR4, LRR3, LRR5, and LRR1 [6]. The LRR5 domain only appeared in CNL proteins in A. duranensis and A. ipaensis [6]. We proposed that A. hypogaea cv. Tifrunner lost the LRR5 domain possibly due to genetic exchange or gene loss after tetraploidization or whole genome duplication (WGD).

Genetic exchange of NBS–LRRs in A. hypogaea cv. Tifrunner

A. hypogaea cv. Tifrunner has 20 chromosomes, Arahy.01–Arahy.20 [30]. The chromosomal location results showed that the greatest number of NBS–LRRs was located on Arahy.12, while the lowest number of NBS–LRRs were located on Arahy.17 (Fig. 1). The chromosomal location of NBS–LRRs was reported in A. duranensis (chromosome: A01–A10) and A. ipaensis (chromosome: B01–B10) by Song et al. [6]. A02 and B02 contained the highest number of NBS–LRRs in A. duranensis and A. ipaensis, respectively, and A06 and B07 had the lowest NBS–LRR number in A. duranensis and A. ipaensis, respectively [6]. In this study, the A subgenome was represented as Arahy.01–Arahy.10, and B subgenome was represented as Arahy.11–Arahy.20 in A. hypogaea cv. Tifrunner based on the number of NBS–LRRs on each chromosome (Fig. 2). This result was consistent with a previous description of chromosome assembly in A. hypogaea cv. Tifrunner by Bertioli et al. [30].

Fig. 1
figure 1

The number of NBS–LRRs distributed on each chromosome in Arachis hypogaea cv. Tifrunner

Full size image
Fig. 2
figure 2

Comparison of the location of representative NBS–LRRs on each chromosome among Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner

Full size image

A polyploidization event (or WGD) can cause gene duplication and loss [41, 42]. A. hypogaea had at least three WGDs [32]; therefore, the number of NBS–LRRs on each chromosome of A. hypogaea cv. Tifrunner changed and was different from the number of NBS–LRRs on each chromosome of A. duranensis and A. ipaensis. We found that although some NBS–LRRs were lost, the total number of NBS–LRRs was higher in A. hypogaea cv. Tifrunner. For example, the number of NBS–LRRs on Arahy.10, 17, and 20 decreased, and the number of NBS–LRRs on other chromosomes increased compared with A. duranensis and A. ipaensis (Fig. 2).

To further reveal the relationship of NBS–LRRs between wild and cultivated peanuts, we constructed one-to-one orthologs. A total of 99 one-to-one orthologous gene pairs were identified between A. hypogaea cv. Tifrunner and A. duranensis, and 142 one-to-one orthologous gene pairs were identified between A. hypogaea cv. Tifrunner and A. ipaensis (Fig. 3). Most one-to-one orthologs corresponded to a similar location on the chromosome between wild and cultivated peanut species. However, some NBS–LRRs from A. duranensis (A genome) corresponded to NBS–LRRs in the B subgenome of A. hypogaea cv. Tifrunner and vice versa (Fig. 3). These results indicated that there was genetic exchange in the A. hypogaea cv. Tifrunner genome, which is consistent with previous findings by Leal-Bertioli et al. [43], who demonstrated that A. ipaensis B genome segments were replaced by the A. hypogaea cv. Tifrunner A subgenome segments, and A. duranensis A genome segments were replaced by A. hypogaea cv. Tifrunner B subgenome segments. The genome structure was not the expected AABB, but was AAAA or BBBB in A. hypogaea cv. Tifrunner [30]. Specifically, approximately 14.8 Mb of the A subgenome sequences were transferred into the B subgenome, and 3.1 Mb of the B subgenome sequences migrated into the A subgenome based on genetic exchange or homoeologous exchange [30].

Fig. 3
figure 3

One-to-one orthologous NBS–LRR gene pairs among Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner. The orange line indicates orthologous NBS–LRR gene pairs in a similar chromosomal location between wild and cultivated peanuts. The blue line indicates orthologous NBS–LRR gene pairs in a different chromosomal location between wild and cultivated peanuts

Full size image

Relaxed selection acting on paralogous NBS–LRR gene pairs in A. hypogaea cv. Tifrunner

A total of 43, 87, and 756 paralogous gene pairs were found in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner, respectively (Additional file 2: Table S2 and Additional file 3: Table S3). A. hypogaea cv. Tifrunner had a greater number of paralogous gene pairs than A. duranensis and A. ipaensis. This could be explained by tetraploidization or WGD. Specifically, a polyploidization event may have retained many duplicated genes [41, 42]. The average Ka/Ks of paralogous NBS–LRRs in A. hypogaea cv. Tifrunner (0.60) was greater than the Ka/Ks of A. ipaensis (0.59) and A. duranensis (0.55, Fig. 4a). Nevertheless, the average Ka/Ks value of paralogous NBS–LRRs was greater than 0.5 in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner, indicating that the paralogous NBS-LRRs were under relaxed selection.

Fig. 4
figure 4

Comparison of selective pressure (Ka/Ks) of paralogous NBS–LRR proteins among Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner. A. Ka/Ks of paralogous NBS–LRR proteins; B. Ka/Ks of paralogous LRR domains. Ka/Ks: nonsynonymous to synonymous per site substitution rates. P < 0.05 indicates a statistically significant difference

Full size image

Compared to other domains of NBS–LRR proteins, the LRR domain underwent more relaxed selection or positive selection because this domain was implicated in pathogenic effector sensing [13,14,15]. Our results showed that the average Ka/Ks value of the LRR domain in A. hypogaea cv. Tifrunner (0.80) was greater the average Ka/Ks value of A. duranensis (0.33) and A. ipaensis (0.41, Fig. 4b), suggesting that LRR domains were under relaxed selection in A. hypogaea cv. Tifrunner, but under purifying selection in A. duranensis and A. ipaensis.

Young NBS–LRR paralogs in A. hypogaea cv. Tifrunner

In this study, the paralogs produced by gene duplication events that occurred before tetraploidization were considered old paralogs. Young paralogs were generated by gene duplication events after tetraploidization. We detected 29 old and 727 young paralogous NBS–LRR gene pairs in A. hypogaea cv. Tifrunner (Additional file 3: Table S3), indicating that many young NBS–LRR paralogs were generated as a result of gene duplication events after tetraploidization. In addition, some old paralogous NBS–LRR gene pairs were lost after tetraploidization, where A subgenome lost 35 paralogous NBS–LRR gene pairs, and B subgenome lost 66 paralogous NBS–LRR gene pairs compared with A. duranensis and A. ipaensis. Previous studies have reported that the properties of old and young genes have different features [44,45,46,47,48,49,50]. For example, young genes have faster evolutionary rates, relaxed selection, lower gene expression levels, shorter gene length, and higher intrinsic structural disorder (ISD) than old genes [46, 47, 49,50,51,52,53]. We found that the average Ka/Ks values of young paralogous NBS–LRRs (0.60) were higher than old NBS–LRRs (0.54, Fig. 5a), indicating that young paralogous NBS–LRRs were under relaxed selection. The average polypeptide length of young paralogous NBS–LRRs (1110 amino acids) was longer than old paralogous NBS–LRRs (1080 amino acids; Fig. 5b). The average ISD value of young paralogous NBS–LRRs (0.14) was lower than the old paralogous NBS–LRRs (0.15, Fig. 5c), indicating that the protein structure of young paralogous NBS–LRRs was stable compared to old paralogous NBS–LRRs. In contrast to these findings, previous studies have found that young genes often have shorter gene length and higher ISD compared to old genes [46, 49]. Young gene has essential function at least underwent 100 MYA [52]. However, the A. hypogaea origination is relatively late [26, 31]. Therefore, we speculated that young NBS–LRRs played the essential functions need more time, it was just rapidly fixed in A. hypogaea cv. Tifrunner.

Fig. 5
figure 5

Comparison of sequence features and substitution rates between old and young paralogous NBS–LRR proteins in Arachis hypogaea cv. Tifrunner. A. Selective pressure (Ka/Ks) between old and young paralogous NBS–LRR proteins in A. hypogaea cv. Tifrunner; B. Polypeptide length between old and young paralogous NBS–LRR proteins in A. hypogaea cv. Tifrunner; C. The intrinsic structural disorder (ISD) of old and young paralogous NBS–LRR proteins in A. hypogaea cv. Tifrunner. Ka/Ks: nonsynonymous to synonymous per site substitution rates. P < 0.05 and < 0.01 indicate significant differences

Full size image

NBS–LRR proteins lost LRR domains in A. hypogaea cv. Tifrunner

NBS–LRR orthologs in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner were under relaxed selection (Fig. 6a), indicating that the biological functions of NBS–LRRs diversified after the divergence of these three Arachis species. Relaxed selection acted on LRR domains of NBS–LRR orthologs between A. duranensis and A. ipaensis (0.53) and between A. duranensis and A. hypogaea cv. Tifrunner (0.71) and purifying selection acted on LRR domains from NBS–LRR orthologs between A. ipaensis and A. hypogaea cv. Tifrunner (0.39; Fig. 6b). These results indicated that the LRR domains between A. ipaensis and A. hypogaea cv. Tifrunner were conserved, and LRR domains between A. duranensis and A. hypogaea cv. Tifrunner were divergent. Moreover, we found that the average Ka/Ks value of homoeologous NBS–LRR proteins (0.57) and LRR domains (0.75) in A. hypogaea cv. Tifrunner was greater than the average Ka/Ks value of orthologs between A. duranensis and A. ipaensis (NBS–LRR: 0.55; LRR domain: 0.53; Fig.7). Taken together, the LRR domains were under more relaxed selection after tetraploidization.

Fig. 6
figure 6

Comparison of selective pressure (Ka/Ks) between orthologous NBS–LRR proteins among Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner. A. Ka/Ks of orthologous NBS–LRR proteins; B. Ka/Ks of orthologous LRR domains. DI. A. duranensis VS A. ipaensis; DH. A. duranensis VS A. hypogaea cv. Tifrunner; IH. A. ipaensis VS A. hypogaea cv. Tifrunner. Ka/Ks: nonsynonymous to synonymous per site substitution rates

Full size image
Fig. 7
figure 7

Comparison of selective pressure (Ka/Ks) between homoeologous NBS–LRR proteins and orthologous NBS–LRR proteins among Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner. A. Ka/Ks of NBS–LRR proteins; B. Ka/Ks of LRR domains. Ka/Ks: nonsynonymous to synonymous per site substitution rates

Full size image

The number of LRR domains in A. duranensis and A. ipaensis were greater than that in A. hypogaea cv. Tifrunner (average number: 2.35 vs 0.72; Fig. 8a). There were fewer types of LRR domains in A. hypogaea cv. Tifrunner NBS–LRRs compared to A. duranensis and A. ipaensis (average number of type: 1.45 vs 0.64; Fig. 8b). Similarly, the number of LRR domains in orthologs of A. duranensis and A. ipaensis was greater than the homoeologs of A. hypogaea cv. Tifrunner (average number: 2.48 vs 0.56, average number of type: 1.73 vs 0.48; Fig. 8c and d).

Fig. 8
figure 8

Comparison of number and type of LRR domains between wild and cultivated peanuts. A. Number of LRR domains between wild and cultivated peanuts; B. Type of LRR domains between wild and cultivated peanuts. C. Number of LRR domains between homoeologous NBS–LRRs and its orthologs; D. Type of LRR domains between homoeologous NBS–LRRs and its orthologs. E. Number of LRR domains from NBS–LRRs that respond to late leaf spot, tomato spotted wilt virus, and bacterial wilt between wild and cultivated peanuts. F. Type of LRR domains from NBS–LRRs that respond to late leaf spot, tomato spotted wilt virus, and bacterial wilt between wild and cultivated peanuts. P < 0.05 and < 0.01 indicate statistical significant differences

Full size image

Although relaxed selection had a greater effect on the NBS–LRRs of A. hypogaea cv. Tifrunner compared to A. duranensis and A. ipaensis, A. hypogaea cv. Tifrunner lost a greater number of LRR domains. These results indicated that the resistance of A. hypogaea cv. Tifrunner to biotic effectors was weaker than that of A. duranensis and A. ipaensis, likely because A. hypogaea cv. Tifrunner lost LRR domains. Similarly, Peele et al. [54] found that A. thaliana was sensitive to biotic stress due to the loss of LRR domains compared to Arabidopsis lyrata, Capsella rubella, Brassica rapa, and Eutrema salsugineum.

It is unclear whether A. duranensis donated the A subgenome to A. hypogaea [26]. A recent study showed that the genome of A. duranensis from Rio Seco, Argentina, was the most similar to the A subgenome of A. hypogaea using chloroplast and ribosomal DNA haplotypes from 50 accessions [30]. In this study, we used A. duranensis (no. V14167) from Argentina [26]. Although there may be differences in the species used in this study, our data suggests that these potential population-level differences did not influence our results. The A subgenome from A. hypogaea had an average DNA similarity of 99.76% to the A. duranensis Rio Seco accessions and 99.61% similarity to A. duranensis V14167 using whole-genome sequencing [30].

NBS–LRRs involved in biotic resistance based on QTLs in A. hypogaea cv. Tifrunner

The QTLs of resistance to late leaf spot, tomato spotted wilt virus, and bacterial wilt were identified in cultivated peanut using A. duranensis and A. ipaensis as reference genomes [55, 56]. Three QTLs with 27 NBS–LRRs, four QTLs with six NBS–LRRs, and one QTL with eight NBS–LRRs were involved in response to late leaf spot, tomato spotted wilt virus, and bacterial wilt, respectively (Table 1 and Additional file 4: Table S4). All of these QTLs were mapped onto the genome of A. hypogaea cv. Tifrunner. One QTL (qTSW_T10_B03_1) contained two NBS–LRRs in A. ipaensis, but its collinear region was absent in NBS–LRRs in A. hypogaea cv. Tifrunner (Table 1), indicating that some NBS–LRRs were lost in A. hypogaea cv. Tifrunner.

Table 1 The number of NBS–LRRs in QTLs that respond to late leaf spot, tomato spotted wilt virus, and bacterial wilt in Arachis duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner
Full size table

In the collinear region, A. duranensis and A. ipaensis had greater number of LRR domains than A. hypogaea cv. Tifrunner (average number: 2.56 vs 0.60, average number of type: 1.58 vs 0.56; Fig. 8e and f). These results indicated that the loss of LRR domains may have decreased ability of NBS-LRR to recognize effectors of bacterial wilt, late leaf spot, and tomato spotted wilt virus in A. hypogaea cv. Tifrunner. Many studies have demonstrated that A. duranensis and A. ipaensis have greater resistant to biotic stressors than cultivated peanut [57,58,59,60]. Thus, we proposed that we may have overestimated the disease resistance of cultivated peanut using A. duranensis and A. ipaensis as reference genomes.

In this study, we identified 31, 11, and 71 NBS–LRRs that responded to late leaf spot, tomato spotted wilt virus, and bacterial wilt in A. hypogaea cv. Tifrunner, respectively. Among these NBS–LRRs, we found 75 young NBS–LRRs and 38 old NBS–LRRs based on gene duplication events after tetraploidization. There were more young NBS–LRRs compared to old NBS–LRRs in A. hypogaea cv. Tifrunner, indicating that young NBS–LRRs were involved in the plant’s response against pathogens. Similarly, Song et al. [61] found that compared to old duplicated genes, young duplicated genes were more likely to be involved in response to biotic stressors in A. duranensis. Although no studies have demonstrated that young genes confer resistance to biotic stress in A. hypogaea cv. Tifrunner, our results indicated that young NBS–LRRs may be involved in response to late leaf spot, tomato spotted wilt virus, and bacterial wilt compared to old NBS–LRRs in A. hypogaea cv. Tifrunner.

Conclusions

We identified NBS–LRRs in A. hypogaea cv. Tifrunner. Genetic exchange events occurred in NBS-LRRs in A. hypogaea cv. Tifrunner compared to A. duranensis and A. ipaensis. Although the LRR domains were under relaxed selection, more LRR domains were lost in A. hypogaea cv. Tifrunner compared to A. duranensis and A. ipaensis. Based on the QTL data, we found that NBS–LRRs were involved in response to late leaf spot, tomato spotted wilt virus, and bacterial wilt in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner. Interestingly, the results suggested that young NBS–LRRs were more likely to be involved in disease resistance compared to old NBS-LRRs in A. hypogaea cv. Tifrunner.

Methods

Identification of the NBS–LRR gene family in A. hypogaea cv. Tifrunner

The complete genome sequence of A. hypogaea cv. Tifrunner has been published [30] and is available on PeanutBase (https://www.peanutbase.org/data/public/Arachis_hypogaea/) [29]. The hidden Markov models (HMM) of NBS (PF00931) and TIR (PF01582) domains were downloaded from the Pfam database [62]. We identified the NBS-containing sequences using NBS domain by HMMER [63] in A. hypogaea cv. Tifrunner. We extracted NBS-containing sequences using an in-house Perl script based on the sequencing ID. Subsequently, we uploaded the NBS-containing sequences to the Pfam database [62] and re-examined these sequences. Among the NBS-containing sequences, we used the same method to identify the TIR-containing sequences. In A. duranensis and A. ipaensis, we found the following five types of LRR domains: LRR1, LRR3, LRR4, LRR5, and LRR8 [6]. We downloaded these five HMMs of the LRR domain from the Pfam database [62] and identified the LRR domains in NBS-containing sequences using HMMER [63] in A. hypogaea cv. Tifrunner. The CC domains of NBS-containing sequences were surveyed using Paircoil2 (http://groups.csail.mit.edu/cb/paircoil2/). The P-score cutoff was 0.03.

Chromosomal location

The gff3 file of the A. hypogaea cv. Tifrunner genome has been released on PeanutBase (https://www.peanutbase.org/data/public/Arachis_hypogaea/) [29]. We used the TBtools program [64] to extract the chromosomal location of NBS–LRRs based on the sequencing ID. The chromosomal location of NBS–LRRs was reported in A. duranensis and A. ipaensis [6]. We used Circos v0.69 [65] to compare the chromosomal location of NBS–LRRs in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner.

Homology in Arachis species

Genes that are paralogs and orthologs in A. duranensis and A. ipaensis have been reported in previous studies [66, 67]. We identified NBS–LRR paralogs and homoeologs in A. hypogaea cv. Tifrunner, and NBS–LRR orthologs between wild and cultivated peanut species. The following evaluation criteria were used as thresholds to determine paralogs and homoeologs in local BLAST analyses [26]: (1) alignment coverage exceeding 80% of the two sequences, (2) identity > 80%, and (3) E-value ≤10− 10.

The paralogous, orthologous, and homoeologous NBS–LRR gene pairs were extracted using an in-house Perl script. MAFFT [68] was used to align pairs of amino acid sequences. PAL2NAL [69] was used to convert amino acid sequences into their corresponding nucleotide sequences. PAML 4.0 [70] was used to calculate the nonsynonymous substitution per nonsynonymous site (Ka), synonymous substitution per synonymous site (Ks), and nonsynonymous to synonymous per site substitution rates (Ka/Ks). Ka/Ks = 1, Ka/Ks > 1, and Ka/Ks < 1 indicated neutral, positive, and purifying selection, respectively. We estimated the Ks, Ka, and Ka/Ks of LRR domains using the same methods.

Polypeptide length and intrinsic structural disorder

The polypeptide length of each NBS–LRR sequence was estimated using codon W (version 1.4, http://codonw.sourceforge.net) with default parameters. The intrinsic structural disorder (ISD) was estimated using IUPred2A with default parameters [71]. The ISD value ranged from 0 to 1, where 0 indicated a stable protein structure, and 1 indicated an unstable protein structure.

Identification of the potential function of NBS–LRRs using quantitative trait loci analysis

To date, many recombinant inbred peanut lines have been constructed to improve biotic resistance, including resistance to bacterial, fungal, insect, and viral stressors. A number of major quantitative trait loci (QTL) were obtained using various molecular markers and genome sequencing methods [55, 56, 72,73,74,75]. Agarwal et al. [55] identified major QTLs related to response to early leaf spot, late leaf spot, and tomato spotted wilt virus using a recombinant inbred population (Tifrunner × GT-C20). Luo et al. [56] identified two QTLs that act in response to bacterial wilt using a recombinant inbred population (Yuanza 9102 × Xuzhou 68–4). The abovementioned QTLs were obtained using genome sequencing of A. duranensis and A. ipaensis as the reference genomes [55, 56]. We obtained these QTLs, and mapped them onto the genome sequences of A. hypogaea cv. Tifrunner using a local BLAST program [76]. The parameters were set as follows: (1) alignment coverage exceeding 80% of QTL sequences, (2) identity > 80%, and (3) E-value ≤10− 10. The NBS–LRRs were identified using the gene location information across the collinear areas in A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner.