Salinity Tolerance of ‘Rootpac 40’ and ‘Nemaguard’. To evaluate the salinity tolerance of ‘Rootpac 40’ and ‘Nemaguard’, rootstocks were irrigated with control irrigation water (1.36 dS m− 1) (C) and treatment irrigation water (3.0 dS m− 1) (T) for ten months (Table 1). The survival rate analysis revealed a higher survival rate for ‘Rootpac 40’ (90.6%) compared to ‘Nemaguard’ (38.9%) (Fig. 1a). The relative percent change of ‘Rootpac 40’ trunk diameter (58%) was slightly greater (p-value = 0.06) than that of ‘Nemaguard’ (45.5%) (Fig. 1b).
Table 1
Control and treatment irrigation water composition
Treatment
|
Salt composition
|
Control (C)
|
Non-saline control [Na+ 1.65 mmolc L− 1, K+ 6.5 mmolc L− 1, PO43− 1.5 mmolc L− 1, Mg2+ 1.3 mmolc L− 1, SO42−1.5 mmolc L− 1, Cl− 1.5 mmolc L− 1, NO3− 5 mmolc L− 1 and micronutrients]
|
Salinity Treatment (T)
|
Mixed cations (Ca2+ = 1.25 Mg2+ = 0.25 Na+) with predominantly chloride (SO42− = 0.2 Cl−) [Na+ 15.5 mmolc L− 1, Ca2+3.8 mmolc L− 1, K+ 6.5 mmolcL−1, PO43−1.5 mmolc L− 1, Mg2+3.1 mmolc L− 1, SO42− 3.8 mmolc L− 1, Cl− 19 mmolc L− 1, NO3−5 mmolc L− 1 and micronutrients]
|
Effect of salinity on biochemical responses. Proline content, antioxidant capacity (oxygen radical absorbance capacity, ORAC), and total phenolics in leaves of ‘Rootpac 40’ and ‘Nemaguard’ were evaluated under control and saline treatments (Fig. 1c and Supplemental Fig. S1). Our data indicated a comparable proline accumulation in the control and salinity treatments for ‘Rootpac 40’, but ‘Nemaguard’ showed a significant increase in proline concentration in response to salinity (Fig. 1c). There were no significant changes in antioxidant capacity in either ‘Nemaguard’ or ‘Rootpac 40’ in response to salinity treatment (Supplementary Fig. S1a). Similarly, neither ‘Nemaguard’ nor ‘Rootpac 40’, showed significant differences for total phenolics in response to salinity treatment (Supplementary Fig. 1b).
Effect of salinity on gas exchange parameters. To study how salt stress affects gas exchange parameters in ‘Rootpac 40’ and Nemaguad rootstocks, we evaluated chlorophyll content by Soil–Plant Analysis Development (SPAD) analysis, net photosynthetic rate (Pn), and leaf stomatal conductance (gs). Although SPAD analysis resulted in no significant effect of salinity on chlorophyll content on either rootstock, the average SPAD value was slightly smaller in ‘Nemaguard’ than in ‘Rootpac 40’ in both control and salt-treated rootstocks (Supplementary Fig. S2a). The photosynthetic efficiency (Pn) value analysis of ‘Nemaguard’ and ‘Rootpac 40’ indicated that salinity significantly inhibited photosynthesis in leaves of both rootstocks (Supplementary Fig. S2b). The stomatal conductance (gs) data revealed that salinity significantly affected stomatal conductance in ‘Rootpac 40’, but no significant effect was observed in ‘Nemaguard’, which had a high variation in gs values compared to ‘Rootpac 40’ (Supplementary Fig. S2c). However, it should be noted that, under salinity, ‘Rootpac 40’ and ‘Nemaguard’ had similar stomatal conductance values (Supplementary Fig. S2c).
Leaf ion accumulation in Rootstock 40 and ‘Nemaguard’ in response to salinity. To study ion accumulation characteristics of ‘Rootpac 40’ and ‘Nemaguard’ in response to salinity stress, ion analysis was performed on leaf samples for Na, Cl, K, Ca, Mg, P, S, B, Cu, Fe, Mn, Mo, and Zn (Fig. 1d,e,f, and Supplementary Fig. S3). In response to salinity stress, both rootstocks showed a higher accumulation of Na compared to their corresponding control. However, in response to salinity, ‘Nemaguard’ leaves accumulated over eight times the concentration of Na than that found in ‘Rootpac 40’ leaves (Fig. 1d). Both ‘Rootpac 40’ and ‘Nemaguard’ showed a significantly higher accumulation of leaf Cl in response to salinity but, similarly to what was observed for Na, the average accumulation of Cl in ‘Nemaguard’ leaves was 1.75-fold higher than that of ‘Rootpac 40’ (Fig. 1e). It is worth noting that under the control condition, ‘Rootpac 40’ showed 12-fold less Na accumulation and three-fold less Cl accumulation than ‘Nemaguard’. Additionally, a significant increase in K accumulation was observed in ‘Rootpac 40’ in response to salinity, which also showed a treatment/control (T/C) K accumulation ratio of 1.14. However, a significant decrease in K accumulation in response to salt treatment was observed in leaves of ‘Nemaguard’ with a T/C ratio of 0.75 (Fig. 1f). There was no significant accumulation of Ca, Mg, S, B, Cu, Fe, Mn, Mo, or Zn between ‘Nemaguard’ and ‘Rootpac 40’ in response to salinity (Supplementary Fig. S3). However, ‘Rootpac 40’ had a small but significant decrease in leaf P accumulation. ‘Nemaguard’ maintained a similar leaf P accumulation under both control and salinity conditions (Supplementary Fig. S3c).
Transcript sequencing and gene expression. To understand the salinity tolerance mechanism at the transcriptome level of salt-tolerant ‘Rootpac 40’ and salt-sensitive ‘Nemaguard’ rootstocks, we performed a three-factor RNA-Seq experiment using two levels per factor to identify differential gene expression between the following variables: 1) treatment type (control vs. salt treatment); 2) rootstock (salt-sensitive, ‘Nemaguard’ vs. salt-tolerant, ‘Rootpac 40’); and 3) tissue type (leaf vs. root). We named our experimental samples as CNL for Control ‘Nemaguard’ Leaf, TNL for Treatment ‘Nemaguard’ Leaf, CNR for Control ‘Nemaguard’ Root, TNR for Treatment ‘Nemaguard’ Root, CRL for Control ‘Rootpac 40’ Leaf, TRL for Treatment ‘Rootpac 40’ Leaf, CRR for Control ‘Rootpac 40’ Root, and TRR for Treatment ‘Rootpac 40’ Root (Supplementary Table S1). Accordingly, we performed RNA sequencing for 24 samples that included three biological replicates of leaf and root tissues harvested from ‘Rootpac 40’ and ‘Nemaguard’ rootstocks. We observed a total of 1,586,189,238 raw reads for 24 libraries with an average of 66,091,218 raw reads/library (Supplementary Table S2). After removing adapter sequences, low-quality reads, and ambiguous nucleotides, we obtained 1,551,884,576 clean reads, consisting of 233 gigabases (Gb) with an average of 9.7 Gb per library (Supplementary Table S2).
To analyze differential gene expression, RNA-Seq reads from each sample were aligned to the annotated Prunus persica genome, which produced an average mapping of 92.27% for individual samples (Supplementary Table S2). The genome sequencing of P. persica predicted 27,852 protein-coding genes26. Our analyses identified 14,985 DEGs in at least one of the comparisons: Treatment vs. control, ‘Nemaguard’ vs. ‘Rootpac 40’, or Leaf vs. Root (Fig. 2 and Table 2). Gene expression-based cluster analysis identified two main groups based on tissue types, one for root tissues and the other for leaf tissues (Fig. 2a). Genes from CRR and TRR formed one subgroup, and genes from CNR and TNR formed the other subgroup within the root and leaf groups (Fig. 2a).
Table 2
Differentially Expressed Genes (DEGs) identified in different comparisons.
Comparison
|
Groups
|
DEGs
|
Upregulated
|
Downregulated
|
Salt vs. Control
|
TNL vs. CNL
TNR vs. CNR
TRL vs. CRL
TRR vs. CRR
|
61
48
9
7
|
4
30
1
1
|
57
18
8
6
|
CNL vs. CRL
CNR vs. CRR
TNL vs. TRL
TNR vs. TRR
|
3403
1347
2064
1559
|
1338
542
799
706
|
2065
805
1265
853
|
|
Leaf vs. Root
|
CNL vs. CNR
CRL vs. CRR
TNL vs. TNR
TRL vs. TRR
|
7853
6725
7913
7220
|
3456
3211
3231
3267
|
4377
3514
4682
3953
|
CNL, control ‘Nemaguard’ leaf; TNL, treated ‘Nemaguard’ leaf; CNR, control ‘Nemaguard’ root, TNR, treated ‘Nemaguard’ root; CRL, control ‘Rootpac 40’ leaf, TRL, treated ‘Rootpac 40’ leaf; CRR, control ‘Rootpac 40’ root; TRR, treated ‘Rootpac 40’ root |
Gene expression analyses were performed to identify DEGs in three different comparisons (treatment vs. control; salt-sensitive rootstock vs. salt-tolerant rootstock; and leaf vs. root) (Fig. 2, Table 2 and Supplementary Table S3). For the treatment vs. control comparisons, 122 DEGs were identified, including 61 for TNL vs. CNL (4 upregulated and 57 downregulated), 48 for TNR vs. CNR (30 upregulated and 18 downregulated), 9 for TRL vs. CRL (1 upregulated and 8 downregulated), and 7 for TRR vs. CRR (1 upregulated and 6 downregulated) (Fig. 2b, Table 2 and Supplementary Table S3). For the comparisons between salt-sensitive (‘Nemaguard’) vs. salt-tolerant (‘Rootpac 40’) rootstocks, 4,765 DEGs were identified, including 3,403 for CNL vs. CRL (1,338 upregulated and 2,065 downregulated), 1,347 for CNR vs. CRR (542 upregulated and 805 downregulated), 2,064 for TNL vs. TRL (799 upregulated and 1265 downregulated), and 1,559 for TNR vs. TRR (706 upregulated and 853 downregulated) (Fig. 2c, Table 2 and Supplementary Table S3). For the comparisons between leaf vs. root, 10,098 DEGs were identified, including 7,833 in CNL vs. CNR (3,456 upregulated and 4,377 downregulated), 6725 in CRL vs. CRR (3,211 upregulated and 3,514 downregulated), 7, 913 in TNL vs. TNR (3231 upregulated and 4682 downregulated) 7,220 in TRL vs. TRR (3,267 upregulated and 3,953 downregulated) (Fig. 2d, Table 2 and Supplementary Table S3).
Verification of DEGs using qRT-PCR. To validate RNA-Seq data, we randomly selected a total of 41 DEGs (18 upregulated and 23 downregulated genes) from different comparison groups to perform qRT-PCR (Supplementary Tables S4 & S5). Among the 41 DEGs, 33 genes were evaluated for single comparisons, and four genes were evaluated for two different comparisons. Relative normalized expressions data were compared for three genes for CNL vs. TNL; four genes for CNR vs. TNR; four genes for CRL vs. TRL; one gene for CRR vs. TRR; seven genes for CNL vs. CRL; eight genes for CNR vs. CRR; six genes for TNL vs. TRL; and eight genes for TNR vs. TRR (Fig. 3). For most genes, qRT-PCR assay results showed a general trend of expression profiles observed in the RNA-Seq experiment, confirming the validity of the RNA-Seq results (Fig. 3 & Supplementary Fig. S4). For example, Prupe.8G148400, which encodes a serine carboxypeptidase showed 8.9- and 10.3-fold upregulation in CRL compared to CNL, in RNA-Seq and qRT-PCR, respectively. Still, a few genes did not exhibit a similar fold change in qRT-PCR compared to RNA-Seq results. For example, Prupe.1G476500, which encodes a BURP protein, showed 95.3- and 42.3-fold upregulation in CNL compared to CRL, in RNA-seq and qRT-PCR, respectively. Although expression levels differed between RNA-Seq and qRT-PCR results for a few genes, overall trends (downregulation or upregulation) were the same.
Gene ontology (GO) enrichment analysis of DEGs. To study functional enrichment analysis of DEGs in treatment vs. control and ‘Nemaguard’ vs. ‘Rootpac 40’, we performed GO enrichment analysis primarily for three major categories: molecular function, MF; cellular component, CC; and biological processes, BP (Supplementary Table S6). In treatment vs. control comparisons, 111, 86, 20, and 31 GO terms were enriched in TNL vs. CNL, TNR vs. CNR, TRL vs. CRL, and TRR vs. CRR, respectively (Supplementary Table S6). In ‘Nemaguard’ vs. ‘Rootpac 40’ comparison, 1552, 901, 1145, and 931 GO terms were enriched in CNL vs. CRL, CNR vs. CRR, TNL vs. TRL, and TNR vs. TRR, respectively (Supplementary Table S6).
KEGG Enrichment Analysis of DEGs. To find out which biological pathways were enriched in treatment vs. control and ‘Nemaguard’ (salt-sensitive) vs. ‘Rootpac 40’ (salt-tolerant), KEGG enrichment analysis of DEGs was performed for each pairwise comparison. In salt treatment vs. control comparisons, 1, 4, 0, and 1 number of pathway(s) were enriched in TNL vs. CNL, TNR vs. CNR, TRL vs. CRL, and TRR vs. CRR, respectively (Supplementary Table S7). In salt-sensitive (‘Nemaguard’) vs. salt-tolerant (‘Rootpac 40’) comparisons, 7, 5, 3, and 4 pathways were enriched in CNL vs. CRL, CNR vs. CRR, TNL vs. TRL, and TNR vs. TRR, respectively (Supplementary Table S7).
DEGs associated with stress pathways. Multiple pathways regulate salt stress signaling in plants. Therefore, all identified DEGs were analyzed to determine their association with primary stress pathways such as phytohormone signaling, redox signaling, and calcium signaling (Fig. 4 and Supplementary Tables S8, S9, S10).
Hormonal Signaling. In treatment vs. control comparisons, two DEGs were upregulated, 1 for jasmonic acid (JA) and 1 for salicylic acid (SA) in TNL vs. CNL (Fig. 4 and Supplementary Table S8). In TNR vs. CNR, 5 DEGs were associated with hormonal signaling, 2 DEGs (one upregulated and one downregulated) for indole acetic acid (IAA), 1 downregulated DEG for JA, and 2 downregulated DEGs were for SA. No DEGs were identified for hormonal signaling in two comparisons: TRL vs. CRL and TRR vs. CRR (Fig. 4 and Supplementary Table S8). In salt-sensitive vs. salt-tolerant comparisons, 117, 26, 59, and 37 DEGs associated with hormone signaling were identified for the CNL vs. CRL, CNR vs. CRR, TNL vs. TRL, and TNR vs. TRR comparisons, respectively (Fig. 4a and Supplementary Table S8). The highest number of DEGs involved in hormonal signaling were identified for IAA, followed by SA and ABA.
Redox signaling. In treatment vs. control comparisons, 4 DEGs associated with redox signaling were downregulated in TNL vs. CNL; of these, 3 DEGs were for heme, and 1 DEG was for glutathione (GSH) (Fig. 4b and Supplementary Table S9). Three DEGs (1 upregulated and two downregulated) were involved in redox signaling, specifically associated with heme in TNR vs. CNR. No DEGs associated with redox signaling were identified for TRL vs. CRL and TRR vs. CRR comparisons (Fig. 4b and Supplementary Table S9). In salt-sensitive vs. salt-tolerant comparisons, 123, 64, 52, and 76 DEGs regulating redox pathways were identified for the CNL vs. CRL, TNL vs. TRL, CNR vs. CRR, TNR vs. TRR comparisons, respectively (Fig. 4b and Supplementary Table S9). The highest number of DEGs associated with redox signaling were identified for heme, followed by GSH.
Calcium signaling. In the salt treatment vs. control comparisons, one DEG was downregulated in TNL vs. CNL, and one DEG was upregulated in TNR vs. CNR (Fig. 4c and Supplementary Table S10). No DEGs were involved in calcium signaling in the TRL vs. CRL and TRR vs. CRR comparisons (Fig. 4c and Supplementary Table S10). In salt-sensitive vs. salt-tolerant comparisons, 40, 27, 10, and 14 DEGs involved in calcium signaling were identified for the CNL vs. CRL, TNL vs. TRL, CNR vs. CRR, TNR vs. TRR comparisons, respectively (Fig. 4c and Supplementary Table S10).
DEGs associated with transporters. Transporters play critical roles in ion distribution and homeostasis in plants. Our observations indicated that the salinity tolerance differences between ‘Rootpac 40’ and ‘Nemaguard’ were primarily due to contrasting accumulation of Na, Cl, and K. This prompted us to analyze DEGs that encode transporters. Our analysis identified a total of 1,194 transporter DEGs in treatment vs. control comparisons and salt-sensitive vs. salt-tolerant comparisons (Fig. 5 and Supplementary Table S11).
In TNL vs. CNL comparison, 8 DEGs were found encoding 7 families of transporters, including the ATP-binding cassette (ABC) superfamily, the divalent anion Na+ symporter (DASS) family, and the intracellular chloride channel (CLIC) family. In TNR vs. CNR, 6 DEGs were identified, encoding 6 families of transporters, including the ATP-binding cassette (ABC) superfamily and the autotransporter-1 (AT-1) family (Fig. 5 and Supplementary Table S11). No DEGs were identified encoding transporter families in TRL vs. CRL and TRR vs. CRR comparisons. In salt-sensitive vs. salt-tolerant comparisons, 488 DEGs were found encoding 76 transporter families in CNL vs. CRL, including the intracellular chloride channel (CLIC) family, the voltage-gated ion channel (VIC) superfamily, and the monovalent cation:proton antiporter-1 (CPA1) family. For TNL vs. TRL, 319 DEGs were found encoding 63 families of transporters, including the polycystin cation channel (PCC) family, the cation channel-forming heat shock protein-70 (Hsp70) family, the Glycoside-Pentoside-Hexuronide (GPH): cation symporter family, the calcium-dependent chloride channel (Ca-ClC) family and the K+ uptake permease (KUP) family. In CNR vs. CRR, 170 DEGs encoded 54 transporter families, including the voltage-gated K+ channel β-subunit (Kvβ) family, the anion exchanger (AE) family, and the proton-dependent oligopeptide transporter (POT/PTR) family. In TNR vs. TRR, 203 DEGs encoded 58 transporter families, including the DASS family and the autotransporter-1 (AT-1) family (Fig. 5 and Supplementary Table S11). Our analyses also revealed that DEGs encoding similar transporters across comparisons (Fig. 5 and Supplementary Table S11).