1. Introduction
Nitrogen is a key mineral nutrient playing a crucial role in plant growth and development. The soil microbiome contributes to nitrogen acquisition, and among the best studied endosymbiotic interactions are those with N-fixing rhizobia and arbuscular mycorrhizal (AM) fungi. Legumes gain access to N through symbiotic association with rhizobia which convert N
2 gas into ammonia in nodules. Although several efforts have been made to incorporate biological N fixation capacity into non-legume plants [
1], agricultural crop production without N fertilization is currently not conceivable. AM fungi help plants in nutrient acquisition and much progress has been made in understanding the molecular basis of P and N transfer from the fungal partner to the host plant [cf. 2]. Less is known about endophytes, although they show relatively little host specificity and have therefore a great potential for agricultural applications [
3].
A well-studied endophytic fungus is
Piriformospora (
Serendipita)
indica, which interacts with numerous host plants and promotes their growth and resistance against biotic and abiotic stresses [4, 5]. Stimulation of growth of its hosts suggests that the fungus promotes nutrient acquisition, including nitrogen. An effect of
P. indica on nitrate uptake and the nitrogen metabolism in the hosts has been reported repeatedly. On full medium, the fungus promotes nitrogen accumulation and the expression of nitrate reductase in Arabidopsis [
6]. In sunflower,
P. indica increases the absorption of nitrogen by the root [
7]. Strehmel et al. [
8] showed that the concentration of nitrogen-rich amino acids decreased in inoculated Arabidopsis plants. Ghaffari et al. [
9] proposed that the nitrogen metabolism plays an important role in systemic salt-tolerance in leaves of
P. indica-colonized barley. Furthermore, Lahrmann et al. [
10] showed that the
P. indica ammonium transporter Amt1 functions as a nitrogen sensor mediating the signal that triggers the in planta activation of the saprotrophic program. In Chinese cabbage, especially the amino acid γ-amino butyrate is de novo synthesized in colonized roots [
11]. Bandyopadhyay et al. [
12] demonstrated that
P. indica together with
Azotobacter chroococcum facilitates higher acquisition of N and P in rice.
P. indica also improves chickpea productivity and N metabolism in a tripartite combination with Mesorhizobium [
13]. Finally,
Serendipita williamsii does not affect P status but C and N dynamics in AM tomato plants [
14]. These examples highlight the importance of the N metabolism on numerous beneficial effects of
P. indica for different plant species, however, how the fungus influences the host N metabolism is not clear. In this study, we use the model plant
Arabidopsis thaliana to investigate how
P. indica interferes with N uptake and metabolism under N limiting conditions.
3. Discussion
N limitation has severe consequences for plant performance [
16], and endophytes may help plants to better adapt to the shortage. We used the well-investigated symbiotic interaction between the model plant Arabidopsis and
P. indica to address this question. We demonstrate that under severe N limitation, the fungus does not stimulate the uptake of nitrate into the host plant, but rather N-label from fungal metabolites appears in the leaves of the host. Since our N limiting medium contains barely any nitrate, the absence of a detectable stimulatory effect of the fungus on nitrate uptake into the host is not surprising. The N metabolites which are translocated from the hyphae to the plants under N limitation did not results in fungus-induced growth promotion (
Figure 1), suggesting that the N supply to the host by the fungus might only compensate deficits. Furthermore, N-translocation from the fungus to the host occurs only under N limitation conditions suggesting the involvement of an N sensing system [cf. 10]. Successful transfer of
15N by arbuscular mycorrhizal fungi to host plants has been shown earlier [17-19]. More recently, Hoysted et al. [
20] investigated clover (
Trifolium repens) colonized by Mucoromycotina fungi and showed that the host gained both
15N and
33P tracers directly from the fungus in exchange for plant-fixed C. Whether the N supply to the host in our study system has comparable symbiotic features with profit for both partners or it is just a stress-related withdrawal of N from the fungus by the plant without any profit for the microbe remains to be investigated. However, since the fungus can grow and propagate on the host under our –N conditions, the N translocation to the host does not restrict hyphal growth. It appears that the conditions are not strong enough to induce changes in the symbiotic interaction [
10]. It is also not clear which metabolites and how they are transported from the microbe to the plant. In
Medicago truncatula, three AMT2 family ammonium transporters (AMT2;3, AMT2;4, and AMT2;5) are involved in the uptake of N in form of ammonium from the periarbuscular space between the fungal plasma membrane and the plant-derived periarbuscular membrane [
21]. In exchange, host plants transfer reduced carbon to the fungi [22-24]. Also Cope et al. [
25] showed that colonization of
M. truncatula with
R. irregularis led to an elevated expression of the mycorrhiza-induced
AMT2;3 and the nitrate transporter
NPF4.12 as well as the putative ammonium transporter
NIP1;5 in the roots. A dipeptide transporter from the arbuscular mycorrhizal fungus
Rhizophagus irregularis is upregulated in the intraradical phase [
26]. To investigate how the fungus manipulates the host N metabolism we performed a comprehensive metabolome and transcriptome analysis for N-related metabolites and genes (
Table 2 and Table 3).
No major impact of the colonization by
P. indica on the changes of shoot metabolite levels in response to N limitation has been observed in this study. Liu et al. [
27] demonstrated that raffinose positively regulates maize drought tolerance by reducing leaf transpiration. The raffinose family oligosaccharides are associated with various abiotic and biotic stress responses in different plant species [e.g., 28-32]. It is conceivable that the stimulatory effect of
P. indica on the raffinose level in N-limited leaves reduces stress.
Nitrate transporter genes are often upregulated under N starvation, however, the role of endophytic microorganisms in nitrate acquisition is not fully understood. In rice, the arbuscular mycorrhizal fungus
R. irregularis remarkably promoted growth and N acquisition, and about 42% of the overall N could be delivered via the symbiotic route under nitrate limiting condition [
33]. Nitrate uptake occurs via NITRATE TRANSPORTER1/PEPTIDE TRANSPORTER FAMILY (NPF)4.5, a member of the low affinity nitrate transporter family which is exclusively expressed in arbuscles of Gramineous species [
33]. A comparable mechanism does not exist in our endophyte/Arabidopsis model, and the putative Arabidopsis NPF4.5 homolog is not upregulated in colonized roots under N limitation. However, we observed a highly specific response of several NPF/NRT1 and NRT2 family members to
P. indica colonisation which allow conclusions how the fungus interferes with the plant N metabolism. The nitrate transporters NRT2.2 and NRT2.4 [
34] are only upregulated in the rosettes when the roots are colonized by
P. indica, while their expression in the roots is not responding to the fungus. This suggests that the fungus promotes nitrate scavenging that is released from the vacuole in response to N starvation. In fact, NRT2.4 has been shown to be expressed close to the phloem in rosettes and to contribute to nitrate homeostasis in the phloem under limiting nitrate supply, since in nitrate-starved
nrt2.4 mutants, nitrate content in shoot phloem exudates was decreased [
34]. Likewise,
NRT1.7 (
NPF2.13) and
NRT1.8 (
NPF5.3) are upregulated by
P. indica in leaves, but not in roots. NRT1.7 loads excess nitrate stored in source leaves into phloem and facilitates nitrate allocation to sink leaves. Under N starvation, the
nrt1.7 mutant exhibits growth retardation, indicating that NRT1.7-mediated source-to-sink remobilization of stored nitrate is important for sustaining growth in plants [
35].
NRT1.8 is expressed predominantly in xylem parenchyma cells within the vasculature and functional disruption of
NRT1.8 significantly increased the nitrate concentration in xylem sap [
36]. In contrast,
NRT2.3 and -2.6 are down-regulated under N limiting conditions and this is further promoted by the fungus. NRT2.6 has been linked to biotic and abiotic stress responses [
37], and it appears that downregulation of
NRT2.6 expression by
P. indica alleviates the stress responses in the roots. Finally,
NRT1.9 (NPF2.9) is strongly downregulated by
P. indica in the leaves.
NRT1.9 is expressed in the companion cells of phloem. In
nrt1.9 mutants, downward nitrate transport was reduced, suggesting that NRT1.9 facilitates loading of nitrate into the phloem and enhances downward nitrate transport to the roots [
38], apparently a process that is restricted by the fungus. Taken together, the analysis of the regulation of the Arabidopsis nitrate transporter genes by
P. indica suggests that the root-colonizing fungus supports nitrate transport to and availability in the aerial parts of the host under our nitrate limiting conditions. This is further supported by the up-regulation of
NRT1.15 (
NPF5.14) by
P. indica in the leaves. NRT1.15 is a tonoplast-localized low-affinity nitrate transport [
39] and overexpression of the gene significantly decreased vacuolar nitrate contents and nitrate accumulation in Arabidopsis shoots. NRT1.15 regulates vacuolar nitrate efflux, and the reallocation might also contribute to osmotic stress responses other than mineral nutrition [
39].
Since the medium does not contain NH
4+, the plant can only receive NH
4+ from the fungus via ammonium transporters (AMTs) [
35,
40,
41].
AMT1-4 expression is upregulated by
P. indica in roots under N limitation. Since AMT1-4 is root-specific [
42], this suggests that the plant tries to compensate its N limitation by stimulating NH
4+ uptake. NH
4+ might also originate from the fungus and it is conceivable that withdrawal of this ion from the fungus might ultimately result in a change of the symbiotic interaction towards saprophytism [cf. 10]. Furthermore, expression of
AMT1-3 and
AMT1-5 as well as
DUR3 coding for an urea transporter [
43,
44] is stimulated by
P. indica in the shoots. Root colonization might create a metabolite environment in the host that requires these transporters for proper distribution of the N metabolites in the aerial parts.
Seven amino acid transporters are regulated >log2-fold by
P. indica colonisation in nitrate-deprived Arabidopsis seedlings. In roots, the fungus prevents down-regulation of the gene for glutamine secreting GLUTAMINE DUMPER (GDU)1 [
45] suggesting that the microbe wants to become access to the plant glutamine. Furthermore, the broad-specificity high affinity amino acid transporter LYSINE HISTIDINE TRANSPORTER (LHT)1 [
46], AMINO ACID PERMEASE (AAP)4, γ
-AMINOBUTYRIC ACID TRANSPORTER (GAT)1 and CATIONIC AMINO ACID TRANSPORTER (CAT)5 are upregulated in the leaves of
P. indica-colonized seedlings. These transporters have been proposed to be involved in nitrogen recycling in plants [
47]. Apparently, a better or different N metabolism management is required for the plant when the roots are associated with the endophyte. LHT1 and -2 are also involved in the transport of 1-aminocyclopropane carboxylic acid, a biosynthetic precursor of ethylene [
48] which might indicate an increased stress by the interaction with the fungus under N limiting conditions. An involvement in nitrogen recycling has also been proposed for 5 of the 10 USUALLY MULTIPLE ACIDS MOVE IN AND OUT TRANSPORTERS (UMANIT13, -20, -40, -45 and -47) [
47]. which are regulated >log2-fold in either roots or shoots of
P. indica-colonized seedlings under N limitation.
4. Materials and Methods
4.1. Plant and fungus material, corresponding growth conditions
Arabidopsis thaliana seeds (
Col-0) were surface-sterilized and sown on N-free MGRL medium supplemented with 2.5 mM NH
4NO
3 and 3 g/L gelrite [
49]. The KNO
3 and Ca(NO
3)
2 in the MGRL medium were replaced by KCl and CaCl
2 to ensure ion equilibrium. After 48 h of stratification at 4 °C in the dark, the seeds were transferred to long-day conditions with 22 °C, 16 h light/8 h dark, 80 μmol m
−2 s
−1 for 10 days.
Piriformospora indica was cultured on Kaefers medium as described before [
50,
51]. As described previously, plugs of a 4-week-old fungal culture were used for co-cultures with the seedlings. The fungus was pre-grown for 7 days on PNM medium (PNM+N) with a nylon membrane in the dark at 22 °C. For N limiting conditions (0 mM total N, PNM-N), KNO
3 and Ca(NO
3)
2 were replaced by KCl and CaCl
2. For control plates without fungus, only empty KM plugs were placed on top of the membrane.
4.2. Plant-fungus co-cultures and determination of growth promotion
For plant-fungus co-cultures for 5 days, 4 plants (per petri dish) were placed on top of the pre-grown fungal lawn as described previously [
51], with some adaptations. Plates were sealed with 3M
TM Micropore tape to reduce the condensation and 10 d-old plants were used for co-cultivation, to reduce the amount of N which accumulates in the plants on MGRL medium before the co-culture. In pilot experiments, we showed that the reduced age did not affect the establishment of the symbiosis with the fungus. The co-cultures were incubated at 22 °C, 16 h light/8 h dark, 80 μmol m
−2 s
−1 with light from the top.
After 5 days, roots and shoots of the plants were harvested separately. For that, 5 plates (= 20 plants) were harvested as 1 sample. Both roots and shoots were washed in sterile distilled water and carefully dried before weighing and direct freezing in liquid nitrogen. Samples were stored in - 80 °C until further use. These experiments were repeated 3-4 times independently.
To determine growth promotion by the fungus, the weight of the sample with fungus was normalized (divided) to the weight without fungus. This was done for total weights sampled from full medium (PNM+N) as well as from N-limited medium (PNM-N). Final growth promotion values presented in the figures are averages of 3 replicates from independent cultures.
4.3. 15N Labeling experiments in the medium
To analyze the uptake of nitrogen by the plant, 2.5 % of the total KNO3 (which equals 0.125 mM KNO3) of the PNM medium was replaced by K15NO3 (Eurisotop, Saint-Aubin, France) dissolved in distilled water. For proper comparison, the 2.5 % of K15NO3 was also added to the N-free medium (PNM-N) resulting in a final concentration of 0.125 mM nitrate. Finally, PNM-N control plates without 15N were used and contained 0.125 mM unlabeled KNO3. Plants grown on these plates were compared to those grown on PNM+N plates to analyze the natural abundance of 15N in the plant tissue. As described before, the fungus or control plug was pre-incubated on the PNM with nylon membrane for 1 week before plants were placed on the plates. The co-cultures were incubated for 5 days to ensure that enough 15N was taken up by the plant.
4.4. 15N fungus labeling experiments
To analyze if the fungus can directly transfer N or N-containing metabolites to the plant, it was labeled with 15N before the co-culture. A modified KM medium without the N-containing components (20 g/L dextrose, 50 mL/L macronutrients, 10 mL/L micronutrients and 1 mL/L Fe-EDTA, 1 mL/L vitamin mix, pH 6.5) was prepared and supplemented with 10g/L ISOGRO®-15N (CortecNet, Les Ulis, France) according to the manufacturers protocol. P. indica plugs of 2 mm diameter were incubated (23 °C, 50 rpm, dark) in 2 mL of KMISOGRO in Greiner CELLSTAR® 12-well plates (Greiner Bio-One, Frickenhausen, Germany) sealed with 3MTM Micropore tape. After 14 days of growth, the fungal tissue was separated from the remaining medium and carefully washed 3 times with N-free liquid PNM to remove 15N bound to the hyphal surface. A 76.66 % enrichment in 15N was achieved using this protocol. The fungus was carefully cut in 5x5 mm pieces and placed on PNM-N and PNM+N plates to start the co-cultures. To minimize 15N uptake by the plant from dead fungal material due to the washing and handling procedure, the fungal plugs were placed in minimum 1 cm distance from the roots. Under these conditions, contact between the two symbionts requires active growth of the hyphae towards the roots. Co-cultivation was performed with 3 plants per plate for 14 days to ensure that enough 15N was taken up by the plant.
4.5. Isolation and clean-up of RNA
Samples of root or shoot material were stored in -80 °C. For homogenization, the samples were ground with mortar and pistil in liquid nitrogen. A maximum of 100 mg material was used for RNA extraction. RNA was extracted with TrizolTM (ThermoFisher Scientific, Waltham, USA) and chloroform according to the manufacturers protocol. Briefly, the plant material was mixed with 1 mL of TrizolTM and incubated on a shaker at room temperature for 15 min. After addition of 250 μl chloroform and a second incubation phase, the sample was centrifuged (30 min, 4 °C). The supernatant was mixed with isopropanol and incubated on ice, followed by centrifugation. The pellet was washed twice with 80 % ethanol, dried and resuspended in RNAse-free water. The RNA isolation was followed by an additional cleaning step to remove access salts originating from the fungus tissue. For that, the sample was mixed with 3 M sodium acetate [1/10 (v/v) in RNAse-free water, pH= 5.2] and 600 µL of ice-cold 100 % ethanol and incubated at -20 °C for at least 1 hr. After centrifugation and 2 cleaning steps with 80 % ethanol the sample was resuspended in RNAse-free water. The quality and concentration of the extracted RNA was tested by absorbance analysis using a NanoVue (GE Healthcare, Uppsala, Sweden).
4.6. RNAseq and data analysis
After transfer of samples to Novogene Genomics Service (Cambridge, UK), the RNA sample integrity was checked with a Bioanalyzer 2100 (Agilent). After samples passed the quality check, the service laboratory proceeded with the library construction and RNA sequencing (PE150) on Illumina NovaSeq™ 6000 platforms as described in a previous study [
52].
The RNAseq libraries were filtered and quality-trimmed with fastp (v0.23.2) [
53], i.e., read ends were truncated to achieve a Phred quality score of 30 or more. Reads below 15 nt length or those comprising at least 2 ambiguous N bases were removed from the dataset. Read qualities were monitored by FastQC (v0.11.3;
https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Hisat2 (v2.2.1) [
54] was used with default parameters to map the quality-trimmed RNAseq libraries to the
A. thaliana reference genome (TAIR10, Ensembl release 51). The mapping allowed spliced reads and single-read mapping to multiple best-fitting locations. FeatureCounts (v1.5.3) [
55] was applied to perform read counting based on the
A. thaliana reference annotation (TAIR10, Ensembl release 51). The Bioconductor DESeq2 (v1.10.0) package [
56] was utilized to identify DEGs in the different pairwise mutant and wild type comparisons. Benjamini and Hochberg’s False Discovery Rate (FDR) approach [
57] was employed to adjust the calculated
p-values for multiple testing.
To identify DEGs of transporters predicted to transport major N compounds, the obtained results were in a first step filtered according to their p-value (p<0.05). Next, DEGs were sorted according to their log2-fold change, here only changes with numbers >= +1.5 and <= -1.5 were further analyzed. This list was cross-checked with targets identified from a search in the UniProt database (
https://www.uniprot.org) using keywords like “NH
4 transport”.
4.7. Analysis of fungal colonization by qPCR
1 mg of RNA was used for the synthesis of cDNA. The Omniscript RT Kit (Qiagen, Hilden, Germany) was used according to the manufacturers protocol with the oligo(dT)
18 primer (ThermoFisher Scientific, Waltham, USA). qPCR was performed with fifty nanograms of the synthesized cDNA as template in a Bio-Rad CFX96 Real-Time PCR Detection System (Feldkirchen, Germany) by use of DreamTaq Polymerase (ThermoFisher Scientific, Waltham, USA) and Evagreen (Biotium, Fremont, USA). The data were normalized with respect to the Arabidopsis
RPS18B (
At1g34030) gene by the 2
-∆∆CT method [
58]. To quantify the
P. indica colonization level of Arabidopsis roots, the expression of
PiTEF1 [
59] was analysed in comparison to the plant´s housekeeping gene
RPS18B (At1g 34030). Following primers were used:
PiTEF1: CGCAGAATACAAGGAGGCC and CGTATCGTAGCTCGCCTGC;
RPS18B: GTCTCCAATGCCCTTGACAT and TCTTTCCTCTGCGACCAGTT [
60]. The colonization was compared between plants grown on PNM-N and PNM+N media (set as 1.0), by the 2
-∆∆CT method.
4.8. Determination of total nitrogen and 15N enrichment
Total N and 15N contents were quantified on 1–2 mg aliquots of dry tissue, after drying a ground tissue aliquot at 65-70 °C for at least 48 h. N elements were detected by gas chromatography on a FLASH 2000 Organic Elemental Analyzer (Thermo Fisher Scientific, Villebon, France). The 15N/14N isotopic ratio was subsequently quantified by a coupled mass spectroscope (Delta V advantage IRMS; Thermo Fisher Scientific, Villebon, France). The total N content was only determined in plant shoots, because a discrimination of N from plant or fungus was not possible in root material.
4.9. Metabolomic analysis
For GC-MS-based quantifications, 25 mg of finely ground plant material was resuspended in 1 mL of frozen (-20 °C) water: acetonitrile: isopropanol (2:3:3, v/v/v) containing Ribitol at 4 ug/mL and analysed as described in [
61].