1. Introduction
Phosphorus (P) is an essential plant macronutrient. As a key component of biomolecules, such as nucleic acids, proteins, and phospholipids, P is involved in multiple biosynthetic and metabolic processes throughout plant growth and development [
1,
2]. Phosphate (Pi), the major form of phosphorus acquired by plants, is not only unevenly distributed in soils, but is also readily fixed onto soil particles into unavailable forms (e.g., aluminum-P, iron-P, and calcium-P) [
3,
4,
5]. Low P availability significantly decreases crop yields, and, thus, becomes a major constraint on crop growth and production [
6]. At the other end of the spectrum, excessive application of P fertilizer is inadvisable, due to depletion of limited P rock resources and eutrophication of marine systems by Pi in runoff that is not utilized by plants [
6,
7]. Intelligent use of moderate amounts of P fertilizer can be beneficial if crops are developed for such conditions. To meet these goals of developing smart crop cultivars with high P utilization efficiency requires further understanding of genetic and molecular mechanisms underlying plant adaptions to P deficiency [
8,
9,
10,
11,
12].
To date, a range of morphological, physiological and molecular processes have been associated with plant in adaptation to P deficiency. These processes include the remodeling of root morphology and architecture, increased exudation of organic acids and acid phosphatases, enhanced expression of Pi transporters, and formation of symbiotic interactions with mycorrhizal fungi or rhizobia [
13,
14,
15,
16,
17,
18,
19]. In recent years, many Pi starvation responsive genes and proteins have been identified and functionally characterized, which has filled in large gaps in our sketch of plant Pi signaling and regulatory networks [
5,
9,
11]. In the center of the Pi signaling network lie several important regulators, such as
phosphate starvation response 1 (
PHR1) and WRKY transcription factors, proteins containing the SYG1/PHO81/XPR1 domain (SPX), and the ubiquitin-like modifier E3 ligase [
20,
21,
22,
23,
24,
25]. Downstream responses include a set of genes directly involved in morphological and physiological responses to Pi starvation, such as
purple acid phosphatase (
PAP) genes functioning in extracellular organic P remobilization,
phosphate transporter (
Pht) genes involved in Pi acquisition, and
expansin (
EXP) genes that participate in alteration of root morphology and architecture [
26,
27,
28,
29,
30].
Among all of the genes,
Pht genes are widely characterized in plants, especially in rice (
Oryza sativa) and
Arabidopsis thaliana. It has been documented that transcripts of a set of
Pht genes were increased by Pi starvation, such as 4 of 9
Pht genes in Arabidopsis and 4 of 13 in rice [
31,
32]. Furthermore,
AtPht1;1 and
AtPht1;4 are suggested to be responsible for about 50% of Pi uptake under Pi starvation conditions in Arabidopsis [
33,
34]. Similarly,
OsPht1;1,
OsPHT1;9 and
OsPHT1;10 were found to modulates phosphate uptake and translocation [
35,
36,
37]. Recently, several
Pht genes have been suggested to regulate root growth, such as
AtPht1;5 for root hair formation and primary root growth in Arabidopsis,
OsPht1;8 for adventitious root elongation and lateral roots number [
38,
39]. Another critical gene family in regulating Pi homeostasis,
SPX family, has also been well characterized in plants. In Arabidopsis, transcripts of 3
AtSPX members were enhanced by P deficiency except to
AtSPX4 [
22]. Similar to Arabidopsis, 5
OsSPX members were up-regulated by P deficiency except to
OsSPX4 [
40]. Recently, a highly conserved mechanism has been suggested that SPX proteins might act as an intracellular Pi sensor mainly through interactions with PHR1/PHR2 in both rice and Arabidopsis [
41,
42,
43,
44,
45].
Additionally, plant phytohormones are suggested to regulate plant responses to Pi starvation, such as auxin, abscisic acid, ethylene, cytokinin [
9,
46,
47]. Furthermore, cross-regulation also occurs between Pi and nitrogen (N) starvation in both legume and non-legume plants [
48,
49]. For example, a critical regulator for Arabidopsis adaptation to nitrogen availability, N limitation adaptation (NLA) was suggested to regulate Pi homeostasis by recruiting PHOSPHAT2 (PHO2) to degrade Pht1;4 in Arabidopsis [
48]. Furthermore, transcription levels of
NLA were found to be regulated by a Pi starvation responsive
miR827 [
49], strongly suggesting there is a crosstalk between N and P deficiency. For legume plants, a crosstalk between N and P deficiency could be directly reflected by significant decreases of both N
2 fixation capability and growth in legume nodules by Pi starvation [
28,
50,
51]. However, one outstanding issue is that a large fraction of our knowledge of Pi signaling networks has been attained in model plants, such as
Arabidopsis thaliana and rice (
Oryza sativa). Verification and application of this knowledge remain fragmentary for most crops, particularly legume crops.
Soybean (
Glycine max L.) is an important legume crop that is a source of high-quality protein and oil [
52]. Similar to other legumes, soybean participates in symbiosis with rhizobia in the formation of nodules [
53]. It has been well documented that rhizobium establishment is a complex process, which is mainly regulated by phytohormones, such as auxin, cytokinin, ethylene, gibberellic acid, strigolactones, jasmonic acid, abscisic acid, and salicylic acid [
54,
55]. For example, it has been suggested that ethylene negatively regulates rhizobia infection and nodule organogenesis because suppression of both
LjEIN2a and
LjEIN2b led to a hypernodulation phenotype in
Lotus japonicus [
56]. Recently, gibberellic acids have been suggested to negatively regulate root nodule symbiosis in
Lotus japonicus and
Medicago truncatula [
57,
58,
59]. Furthermore, it has been documented that P availability adversely affects soybean nodule development and growth [
60,
61,
62]. In addition, it has been suggested that responses to P deficiency are similar between roots and nodules, because of significant increases of proton exudation, and, thus, decreases of rhizosphere pH were observed in soybean grown in low P conditions [
18,
63]. However, few studies were conducted to investigate gene expression patterns between nodules and roots in legume crops. For example, it has been documented that 11 of 14
GmPT members exhibited Pi-starvation responsive expression patterns in soybean roots [
28,
29], but information about transcripts of all
GmPT members responsive to Pi-starvation remains largely unknown in soybean nodules. Recently, increased transcription of
GmPT5 has been shown to play a critical role in maintaining Pi homeostasis in soybean nodules [
62]. Plus, a Pi starvation responsive gene,
GmEXPB2, plays vital roles in adaptive responses of both soybean roots and nodules to P deficiency, possibly through cell wall modifications [
26,
51]. The results strongly suggest that identification and functional analysis of Pi starvation responsive gene is critical for elucidating adaptive strategies to low P stress in soybean nodules. Yet, genome-wide transcriptome analysis has not been conducted to identify Pi starvation responsive genes in soybean nodules.
Although genome-wide transcriptome analysis has been successfully used to elucidate molecular mechanisms underlying complex adaptations of plants to P deficiency using RNA-seq technique, most of these studies focus on roots or leaves of plants grown under non-symbiotic conditions. Little transcriptome information is available for legume nodules. As far as the authors are aware, only three studies have been conducted to investigate global gene expression responses to Pi starvation in legume nodules, including with bean (
Phaseolus vulgaris),
Medicago truncatula and chickpea (
Cicer arietinum) [
64,
65,
66]. However, there is little information about genome-wide analysis of gene transcripts responsive to Pi starvation in soybean nodule. Furthermore, it is well known that the formation of symbiotic nodules and their responses to Pi starvation varies considerably among legume species and rhizobium strains [
66,
67]. Thus, it is important to investigate molecular mechanisms underlying nodule development and physiology for each commercially important legume crop under Pi starvation.
In this study, genome-wide transcriptomic analysis of soybean nodules in response to P deficiency was conducted via RNA-seq. Thousands of differentially expressed genes were identified in soybean nodules under P deficiency, with many involved in nutrient/ion transport, transcriptional regulation, key metabolic pathways, Pi remobilization, and signaling. These results will enable future researchers to further elucidate molecular processes within nodules adapted to P deficiency, which will ultimately lead to the development of P-efficient soybean varieties that can maintain symbiotic nitrogen fixation (SNF) in low or moderate P availability systems.
3. Discussion
Leguminous plants form nodules through symbiotic interactions with rhizobium species. These organs are the sites of SNF, which provide nitrogen for host plants. However, it is well documented that P deficiency significantly influences nodule growth and development in legume plants, such as in soybean, common bean,
Medicago truncatula, and chickpea [
61,
62,
64,
65,
66]. Consistently, in this study, P deficiency also led to significant inhibition of nodule growth and development, as reflected by decreases in nitrogenase activity, nodule fresh weight and nodule size with Pi starvation (
Figure 1). However, relative to effects on leaves and roots, decreases of total P content, soluble Pi concentration, and total P concentration in nodules were less affected by Pi starvation (
Figure 1), strongly suggesting that nodules are P sinks with a high capability of maintaining Pi homeostasis to reduce adverse effects of P deficiency on nodule growth and development [
60,
62,
65,
66].
With the aid of genome-wide analysis of gene expression through microarray or RNA-seq approaches, a group of Pi starvation responsive genes have been identified in plant leaves and roots, which has facilitated the elucidation of adaptive strategies employed by plants to minimize the detrimental effects of P deficiency through functional characterization of these DEGs [
11,
69,
70,
71,
72,
73]. Recently, with the aid of RNA-seq approaches, genome wide analysis of Pi starvation responsive genes in legume nodules has been studied, with the host plants being common bean,
Medicago truncatula and
chickpea [
64,
65,
66]. Common responses of legume nodules to Pi starvation could be demonstrated by identifying a set of DEGs with high homology among three legume species, such as
WRKY,
MYB and
NAC [
64,
65,
66]. However, it seems that more complex responses of soybean nodules to Pi starvation were elucidated as reflected by identification of 2055 Pi-starvation responsive genes, which was more than 495 in bean, 1140 in
Medicago truncatula and 540 in chickpea [
64,
65,
66]. For example, 8
GmPT members and
GmSPX members were found to be responsive to Pi starvation in soybean nodules, but only 1
SPX member in
Medicago truncatula, 1
PT member and
SPX member in chickpea have been identified [
64,
65,
66]. Furthermore, it seems that a set of genes preferring to increase transcripts in soybean nodules at low P levels were identified in the current study, such as
GmPT5,
GmSPX1, and
GmPAP11/30. For example, among eight Pi-starvation up-regulated
GmSPX members in soybean nodules,
GmSPX1 has been documented to exhibit no response to Pi starvation in soybean roots [
74]. Meanwhile, transcription of
GmPAP11/30 was found to have no response to Pi starvation in soybean roots [
27], suggesting complex responses of soybean nodules to Pi starvation.
Enhanced Pi mobilization and acquisition through increased exudation of organic acids and purple acid phosphatase, along with up-regulation of Pi transporters are well-documented strategies employed by plant roots in response to P deficiency [
75,
76,
77,
78,
79,
80,
81]. Nodules exhibit similar responses to roots in response to P deprivation, with up-regulation of genes related to Pi mobilization and acquisition, such as Pi transporters, and purple acid phosphatases, which allows for the maintenance of Pi homeostasis in nodules (
Table 3 and
Figure 4). In this study, eight out of 14
GmPT members were significantly enhanced in nodules as a result of P deficiency (
Table 3). Among them, Pi starvation up-regulated
GmPT5 might mediate Pi homeostasis in soybean nodules through control of Pi translocation from roots to nodules [
35]. In the present study, other three
GmPT members (i.e.,
GmPT2/
6/
14) were found to be up-regulated by Pi starvation, strongly suggesting other
GmPT members could mediate Pi acquisition and translocation in soybean nodules at low P level except to
GmPT5 [
62], which merits further analysis.
Accompanying increases in the abundance of nine
GmPT transcripts, 16
PAP transcripts were also observed as differentially expressed in nodules subjected to Pi starvation, which is consistent with observations of significantly increased APase activity in P deprived nodules (
Table 3 and
Figure 2). Increased
PAP transcription and APase activity are well known to play vital roles in the regulation of internal P metabolism and extracellular organic P mobilization in plants [
82,
83]. Although functions of several
GmPAP have been documented, including the involvement of
GmPhy and
GmPAP4 in phytate-P mobilization, and the participation of
GmPAP3 in ROS metabolism in plants under salt stress, functions of most Pi starvation up-regulated
GmPAPs, except
GmPAP21, remain largely unknown [
84,
85,
86,
87].
GmPAP21 overexpression leads to nodule growth inhibition in soybean, suggesting that it participates in internal P metabolism within soybean nodules [
87]. Furthermore, it was observed that organic-P utilization was enhanced in rhizobia inoculated in soybean, it is reasonable to hypothesize that Pi starvation responsive
GmPAPs might also be involved in extracellular organic-P utilization in soybean [
18]. Among Pi starvation up-regulated
GmPAPs, GmPAP11/20/23 exhibited high homology with SgPAP10 in stylo functions as mediating extracellular organic-P utilization [
81], suggesting that
GmPAP11/20/23 might contribute to extracellular organic-P utilization in soybean nodules.
In addition to
GmPT and
GmPAP, two
GmSPX genes,
GmSPX1 and
GmSPX3, are also potentially vital regulators of Pi signaling pathways in soybean [
74,
88]. Interestingly,
GmSPX1 and
GmSPX3, together with six other
GmSPX members were found to be significantly up-regulated in soybean nodules upon Pi starvation (
Table 3). This indicates that
GmSPX members are good candidates for genes involved in maintaining Pi homeostasis in soybean nodules.
In addition to differential expression associated with Pi acquisition and mobilization, many Pi-starvation responsive DEGs in soybean nodules were associated with nitrate/nitrite absorption and assimilation (
Figure S2). Similarly, Pi starvation can lead to significant increases in the concentrations of total amino acids and asparagine in common bean and chickpea [
64,
89]. Furthermore, consistent with increased asparagine accumulation, three
asparagine synthetase genes were found to be up-regulated by Pi starvation in soybean nodules (
Table 1 and
Table S3), strongly suggesting that Pi starvation significantly influences amino acid accumulations in nodules. Increased asparagine accumulation is known to inhibit the capacity for SNF in nodules, suggesting that asparagine plays a role in N feedback regulation of SNF [
90,
91,
92]. Plus, nitrogenase activity has been severely curtailed through phloem-feeding of asparagine, which further implicates asparagine as a phloem-mobile shoot-born factor that functions in systemic feedback regulation of SNF [
91]. However, these previous investigations did not include experiments of P effects. Therefore, further investigation of regulatory mechanisms underlying amino acid synthesis and transport involving nodules in responses to Pi starvation remains as a relevant subject for future researchers.
In this study, 38 plant hormone-related genes were identified as DEGs in response to P deprivation (
Table 4). This indicates that a variety of signaling pathways within nodules participate in responses to Pi starvation. For example, two genes (Glyma.10G056200 and Glyma.16G020800) coding auxin-responsive proteins were down-regulated, while one auxin responsive factor (Glyma.19G206100) and 5 AUX/IAA family members were up-regulated by Pi starvation in soybean nodules (
Table 4). This suggests that auxin is involved in nodule adaptation to Pi starvation. However, specific roles for auxin signaling in adaptive strategies of nodules to Pi starvation remain unknown. Although miR160 can negatively regulate
AUXIN RESPONSE FACTOR10 (
ARF10), and, thus, increase auxin sensitivity and inhibit soybean nodule development [
93], none of these genes were found to be significantly regulated by Pi starvation in the present study (
Table 4). Therefore, other auxin pathways might also regulate nodule responses to Pi starvation, which requires further investigation for more conclusive evidence.
Calcium signaling was also found to be important in the current work, as 24 Ca
2+ signaling related genes were found to be regulated by Pi starvation in soybean nodules, including two
calmodulin-like and four
calcium-dependent protein kinase. This suggests that low P availability affects Ca
2+ signaling, and thereby regulates nodule development (
Table 5). Consistent with this result, sustained oscillation of calcium concentrations is known to activate the expression of symbiosis-related genes after perception of rhizobia-derived nodulation factors [
94,
95]. Meanwhile, CCaMK, a nuclear calcium- and calmodulin-dependent kinase has been suggested as the central regulator in symbiotic development in plants [
96]. All these results strongly suggest that Ca
2+ signaling is also involved in regulating soybean nodule adaptations to Pi starvation.
Finally, significant alterations of transcriptional regulation are implied by the presence of 76 transcription factors among the DEGs responsive to P deficiency in soybean nodules. These numbers include 12 MYB and five GRAS transcription factors (
Figure 5). Although functions of MYB and GRAS transcription factors remains largely unknown in soybean nodule development and responses to Pi starvation, one MYB transcription factor,
LjIPN2, has been documented as capable of binding directly to the
NIN gene promoter and, thus, play an important role in the Nod signaling pathway in
Lotus japonicas [
97]. Meanwhile, it has been reported that the GRAS family transcription factors,
MtNSP1 and
MtNSP2 form a protein complex that is essential for root nodule symbiosis in
Medicago truncatula [
98]. The results herein are consistent with these previous reports and further suggest that complex transcriptional regulatory networks participate in soybean nodule adaption to Pi starvation.