Evaluation of salt resistance in three Gossypium spp.
The seedlings of three Gossypium spp. (G. arboreum L. Shixiya1, G. raimondii, and G. hirsutum L. TM-1) were subjected to salt stress. Morphological changes such as severe dwarfing, leaf wilting, and chlorosis were observed in all three Gossypium spp. after irrigation with salt solution (300 mM NaCl) (Fig. 1a, b). The Na+ concentration in the shoot and roots of stressed seedlings was measured. Both shoot and roots accumulated high Na+ under salt stress (Fig. 1c). The results indicate that the seedlings of the three cotton species were all affected when grown in soils with high salt concentrations.
Identification of cotton NHXs
NHXs are evolutionarily and functionally conserved proteins. To identify and compare the NHX members of cotton, the NHX protein sequences of Arabidopsis, rice, tomato, soybean, and mulberry [30, 31, 32, 33, 34, 35] were downloaded from NCBI and used to build HMM to search against cotton genome database [23, 36, 37]. All putative NHXs were further analyzed by InterPro and SMART to reconfirm their signature sequences and conserved domains. The predicted protein sequences with the Na+/H+ exchanger domain were consider as cotton NHX. We identified 12, 12, and 23 NHXs from G. arboreum, G. raimondii, and G. hirsutum, respectively (Table 1). The details of the identified NHX proteins are listed in Table S1, including accession number, sequence length, predicted isoelectric point (pI), and molecular weight (Mw). The NHXs of the three Gossypium spp. encoded proteins that varied from 211 to 1,193 AA in size with predicted pI ranging from 5.05 to 9.14 and Mw ranging from 22.95 to 132.84 kD.
Phylogenetic tree of NHXs in cotton
To investigate the evolutionary relationships between NHXs of G. arboreum, G. raimondii, G. hirsutum and other plant species such as rice, mulberry, and Arabidopsis (Table S2), we built an unrooted phylogenetic tree using ClustalW and MEGA 6.0 with neighbor-joining (N-J) method. Similar to the previous studies in other plant species [30, 31, 32, 33, 34], the 47 NHX candidates of Gossypium were divided into three distinct classes based on the sequence conservation (Table 1, Fig. 2). The majority of NHXs (75%) of Gossypium spp. were categorized into class I (9 NHXs of G. arboreum, 9 of G. raimondii, and 17 of G. hirsutum). Class II had 2 GaNHXs, 2 GrNHXs, and 4 GhNHXs. According to previous research, class III is the smallest group, and only 1 GaNHX, 1 GrNHX, and 2 GhNHXs were categorized into this class in our study.
All NHXs of G. arboreum had orthologs in G. raimondii. For instance, Ga10585 and Gr10021256 of class I are high homologous. They shared 99.26% identity in the protein sequence with only a few single nucleotide polymorphisms observed (Figure S1) and had similar predicted pI and Mw values (Table S1). The phylogenetic tree and sequence alignment showed common gene duplication events associated with chromosome doubling in G. hirsutum compared to G. arboreum and G. raimondii, which were also found in the previous reports of gene family studies in G. hirsutum [38]. These gene duplication events contributed to the expansion of NHX family in G. hirsutum, which has twice as many NHXs in diploid cotton (Table 1). The only exception is Gr10038122 of class I. The NHX corresponding to Gr10038122 in the D sub-genome of G. hirsutum could be lost during the evolution process after G. hirsutum formation.
We further analyzed the genomic distribution of cotton NHXs. The genomic sequences of NHXs were used to query with BLAST to assess their chromosomal locations. We observed that the distribution pattern of NHXs in G. hirsutum is symmetrical and most members split evenly between A and D sub-genomes, while NHX genes of diploid cotton are distributed uniformly on the genome (Figure S2). The Circos cycle demonstrated the syntenic relationship between cotton and Arabidopsis (Figure S2), which revealed that most of the NHXs in Arabidopsis are homologous with the syntenic genes pairs in cotton. For example, AtNHX7 located on At2 chromosome of Arabidopsis is highly homologous to GhD02G1385 and GhA03G0996 in D02 and A03 of G. hirsutum.
Structural divergence of cotton NHX genes
The structural analysis of cotton NHX gene family revealed that all the identified NHXs possess several introns (Fig. 3). Generally, the class I NHX genes from G. arboreum and G. raimondii are conserved and contain 13 introns. Contrarily, one-third of the class I GhNHXs evolved different structure compared to the orthologous in G. arboreum or G. raimondii (eg., Ga36304/GhA09G0594/Gr10009549/GhD09G0593). Ga36304 and Gr10009549 with conserved protein sequence and gene structure contain 14 exons, and the longest intron is 1 kb and 0.85 kb, respectively. On the other hand, GhA09G0594 and GhD09G0593 evolved several extremely long introns. The longest intron of GhA09G0594 is 3.8 kb and GhD09G0593 is 5.2 kb. As a consequence, the genome sequences of these two GhNHX genes expanded to 18 kb and 16 kb, respectively, and the genome sequences of Ga36304 and Gr10009549 are both 7 kb. Besides, the predicted protein sequences of these four orthologs were different (Figure S4). Similar evolutionary events were observed in class III. Gh02G1385 has a 17.2 kb intron at the 3’ end and a total genome sequence of 31 kb. These findings revealed the complexity of the gene structure of cotton NHXs that may be related to the biological roles of NHX family.
Expression pattern of NHXs in cotton under salt stress
The expression pattern of NHX genes in G. hirsutum under salt stress was studied using previously published transcriptome data [23]. RPKM (reads per kilobase per million mapped reads) values in the leaves of G. hirsutum under NaCl treatment at indicated time points were used to create the salt-induced heat map of GhNHXs. As shown in Fig. 4, The 2 genes from class III, homologous to SOS1, were upregulated within 1 h of salt treatment. Four genes from class II showed no significant difference. NHX members of class I showed significantly different expression patterns. Seven genes were significantly induced and 3 genes were significantly suppressed under salt stress, while 6 NHXs were not expressed in all samples. These findings indicated differences in the potential roles of NHX classes under salt response. We cloned the gene pair GhA11G2132/GhD11G2440 from TM-1 transcripts for further study.
qPCR was performed to reconfirm the expression of GhA11G2132/GhD11G2440 in different tissues and in respond to salt treatment. Because of the high similarities of GhA11G2132/GhD11G2440 gene pair and high requirements for qPCR primers, it is very difficult to study the expression changes of the two genes individually with qPCR-based methods. The gene-specific primers were designed to amplify the GhA11G2132 and GhD11G2440 together. As shown in Fig.5, GhA11G2132/GhD11G2440 gene pair was expressed in root, stem, and leaf, and the expression level of GhA11G2132/GhD11G2440 in leaf and root was two times than that in stem. GhA11G2132/GhD11G2440 was significantly induced by 300 mM NaCl, which was consistent with the results of heat map analysis.
Analysis of GhNHX1A/D protein sequence
The sequences of GhA11G2132/GhD11G2440 were studied. Phylogenetic tree (Fig. 2) and sequence alignment (Fig. 6) revealed that GhA11G2132/GhD11G2440 pair was homologous to NHX1 and NHX2 from Arabidopsis. GhA11G2132 from sub-genome A was denoted as GhNHX1A, and the GhD11G2440 from sub-genome D was denoted as GhNHX1D. Prediction methods highlighted 12 transmembrane domains for GhNHX1A and GhNHX1D, and they showed high similarity (96.17%) with an extra five amino acid segment at 444~448 AA in GhNHX1D. The putative amiloride-binding site (LFFIYLLPPI) that inhibits eukaryotic NHX function was found in transmembrane domain III.
The biological roles of plant NHXs were closely related to their subcellular localization. To study the possible roles of the GhNHX1A/D in upland cotton, we fused GFP to the N-terminal of GhNHX1A/D. The vacuolar marker protein δ-TIP [39] fused with C-terminal RFP was cotransferred with GFP-NHX to the Arabidopsis protoplast. As shown in Fig. 7, GFP expressed throughout the protoplast, while red fluorescence expressed only on the vacuolar membrane in the protoplast transformed with GFP and δ-TIP-RFP vectors. The coexpression of GFP-GhNHX1A/D and δ-TIP-RFP in the protoplast, represented by overlapped red and green fluorescent signals, indicated that majority of GhNHX1A/D localized to the vacuolar system where the δ-TIP protein was located.
GhNHX1A/D regulate salt tolerance in cotton
The potential roles of GhNHX1A/D in cotton response to salt stress were studied using the VIGS method. A 376 bp fragment from 3’ end of GhNHX1A was cloned and inserted into the VIGS vector to silence both GhNHX1A/D. After 3 weeks, the silencing efficiency was detected in cotton leaves using qPCR. GhNHX1 expression in GhNHX1-silenced (TRV:NHX1) plants reduced to 10% compared with the empty vector transferred control (TRV:00) plants (Fig. 8a). The phenotypic appearance of TRV:00 and TRV:NHX1 was studied and no significant difference was observed under normal conditions.
Cotton seedlings were subjected to water or 300 mM NaCl treatment for 6 days. High salt concentrations restricted cotton growth and development; salt treatment resulted in shorter plants with wilting and yellowing symptoms on leaves. Compared to TRV:00, the TRV:NHX1 showed reduced tolerance to salt stress with more wilting of leaves and less biomass of both roots and leaves (Fig. 8b, c). The Na+ content was measured in the roots and shoots of TRV:00 and TRV:NHX1 plants after water or salt treatment (Fig. 8d). After 6 days of salt treatment, the Na+ concentration in TRV:NHX1 was increased compared with TRV:00. These findings indicated a positive role of GhNHX1 in cotton response to salt stress.