Soybean biomass, nodulation and P nutrition response to treatments
Comparison of the shoot dry weight using two-way ANOVA indicated significant (P ≤ 0.05) effect of phytate (M0P1) and combined treatment (M1P1) compared to control (M0P0), whereas inoculum (M1P0) was insignificant. The root dry weight was unaffected by inoculum (M1P0), phytate (M0P1) and their interactions (M1P1) (Fig. 1A). Although phytate addition significantly (P = 0.017) increased shoot dry weight (P = 0.037) and the number of nodules per plant (Table 1), the microbial inoculum had no significant influence on nodulation. Total P levels in inoculated plants (M1P1) were higher (22 mg/plant) than in control plants (M0P0) (16.78 mg/plant), but the difference was statistically insignificant (Fig. 1B).
Table 1
Two-way ANOVA from the effects of treatment on fresh and dry weight of shoot and root, nodule formation and total phosphorus.
Source of variation | df | Shoot dry weight | Root dry weight | Nodule | Total P |
F | Pr(> F) | F | Pr(> F) | F | Pr(> F) | F | Pr(> F) |
Inoculum | 1 | 3.060 | 0.088 | 3.899 | 0.058 | 0.171 | 0.683 | 3.391 | 0.098 |
Phytate | 1 | 8.018 | 0.007 | 0.037 | 0.847 | 8.525 | 0.017 | 2.901 | 0.122 |
Inoculum: Phytate | 1 | 0.983 | 0.327 | 0.009 | 0.921 | 0.773 | 0.390 | 1.465 | 0.256 |
Bacterial And Fungal Community Structure In Different Biotopes
Raw data of Illumina MiSeq produced a total of 13,148,932 reads where 7,763,410 reads for the bacterial 16S rRNA and 5,385,522 reads for the fungal ITS region. DADA2’s filtering, trimming and quality controlling, resulted on a total of 2,896 335 reads for 16S rRNA (Fig. S2A-C) and 4,234,021 for ITS (Fig. S2D-F). Finally, we assembled forward and reverse filtered reads into 15,613 ASVs for bacteria and 2171 ASVs for fungi. We then analyzed the effects of treatments on the diversity and structure of the bacterial and fungal communities in the roots and rhizosphere biotopes separately.
Alpha diversity indices of bacteria were insignificant for microbial inoculum (Shannon P = 0.1503 and Simpson P = 0.3776), phytate (Shannon P = 0.8643 and Simpson P = 0.908) and their interactions (Shannon P = 0.7244 and Simpson P = 0.5756) (Fig. 1C). Similarly, fungal alpha diversity did not significantly differ for different treatments (Fig. 1D). Bacterial communities clustered by niche along the first axis of the PCoA ordination. Root bacteria formed distinct clusters under M0P0, M1P0, and M0P1, whereas rhizosphere bacteria were much more scattered (Fig. 1E). The clustering pattern of fungal communities showed an opposite pattern to those of bacteria. M1P1 caused fungal communities in the rhizosphere to cluster less closely than other treatments (Fig. 1F). According to the PERMANOVA test, microbial inoculum had significant effect on bacteria in the root (P = 0.001) and rhizosphere (P = 0.004). Phytate significantly influenced the structure of the bacterial communities in the root biotope (P = 0.024) (Table 2A). PERMANOVA test was insignificant for the root fungi, but inoculum had a significant (P = 0.007) impact on the structure of the rhizosphere fungi (Table 2B).
Table 2. Effect of microbial inoculation and phytate on the structure of the bacterial and fungal communities in root and rhizosphere according to PERMANOVA.
A. Bacteria
Variable
|
Source
|
DF
|
SumOfSqs
|
R2
|
F
|
Pr(>F)
|
Roots
|
Inoculation
|
1
|
0.2347
|
0.04366
|
1.7477
|
0.001 ***
|
Phytate
|
1
|
0.1686
|
0.03136
|
1.2555
|
0.024 *
|
Inoculation: Phytate
|
1
|
0.1384
|
0.02574
|
1.0306
|
0.354
|
Residual
|
36
|
4.8353
|
0.89924
|
|
|
Total
|
39
|
5.3771
|
1.00000
|
|
|
Rhizosphere
|
Inoculation
|
1
|
0.2676
|
0.05702
|
2.2687
|
0.004 **
|
Phytate
|
1
|
0.0878
|
0.01871
|
0.7446
|
0.929
|
Inoculation: Phytate
|
1
|
0.0908
|
0.01934
|
0.7696
|
0.849
|
Residual
|
36
|
4.2474
|
0.90490
|
|
|
Total
|
39
|
4.6937
|
1.00000
|
|
|
B. Fungi
Variable
|
Source
|
DF
|
SumOfSqs
|
R2
|
F
|
Pr(>F)
|
Roots
|
Inoculation
|
1
|
0.1592
|
0.01993
|
0.7662
|
0.751
|
Phytate
|
1
|
0.1368
|
0.01713
|
0.6551
|
0.878
|
Inoculation: Phytate
|
1
|
0.1730
|
0.02166
|
0.8283
|
0.653
|
Residual
|
36
|
7.5181
|
0.94128
|
|
|
Total
|
39
|
7.9870
|
1.00000
|
|
|
Rhizosphere
|
Inoculation
|
1
|
0.1502
|
0.04113
|
1.6240
|
0.007 **
|
Phytate
|
1
|
0.0789
|
0.02161
|
0.8532
|
0.791
|
Inoculation: Phytate
|
1
|
0.0928
|
0.02542
|
1.0035
|
0.428
|
Residual
|
36
|
3.3296
|
0.91184
|
|
|
Total
|
39
|
3.6515
|
1.00000
|
|
|
Planctobacteria And Ascomycota Dominated Soybean Microbiota
The 15,613 bacterial ASVs were assigned to 39 phyla (Table S1) and 196 orders (Table S2), with Planctobacteria being the most abundant phyla in both root (Fig. 2A) and rhizosphere (Fig. 2C) biotopes. We chose the top 10 orders based on their high relative abundance and eight (Tepidisphaerales, Gemmatales, Isophaerales, Pirellulales, Planctomycetales, Chthoniobacteriales, Phycisphaerales and Burkholderiales) of the 10 most abundant orders were dominant in both biotopes (Fig. 2B, D). The root biotope was dominated by Tepidisphaerales (Fig. 2B), while the rhizosphere biotope was dominated by Gemmatales (Fig. 2D). Bacterial indicator species analysis revealed 35 ASVs under inoculation treatment, with 19 ASVs enriched in the root (Fig. S3A) and 16 in the rhizosphere (Fig. S3B). Thirteen ASVs were enriched under phytate treatment, with 7 ASVs enriched in root (Fig. S3C) and 6 ASVs in the rhizosphere (Fig. S3D) (Table S3); however, indicator species analysis under combined inoculum and phytate (M1P1) treatment significantly (P ≤ 0.05) revealed BASV738 (Tepidisphaera mucosa) and BASV766 (Candidatus Anammoximicrobium moscowii) in the root; and BASV1092 (Pirellula sp.) in the rhizosphere biotope (Table S4).
In the fungal dataset, we identified six phyla: Ascomycota, Basidiomycota, Mucoromycota, Chytridiomycota and Blastocladiomycota with one not assigned (NA) to any phylum (Table S5) and 92 orders (Table S6). Ascomycota was the most abundant phylum in both root and rhizosphere biotopes (Fig. S4A, C). The Sordariales order dominated fungal communities both in the root and rhizosphere biotopes (Fig. S4B, D). Both biotopes shared six (Hypocreales, Sordariales, Pleosporales, Orbiliales, Glomererellales and Pezizales) of the top 10 orders (Fig. S4B, D). Fungal indicator species in the rhizosphere revealed 48 ASVs under inoculation treatment and nine ASVs under phytate addition. Just one fungal ASV, FASV113 (Arthrobotrys conoides) was found to be enriched in the root biotope under phytate treatment, while no ASV was found in the root biotope under inoculum treatment (Table S3). In the rhizosphere, four arbuscular mycorrhizal fungi (AMF)- Glomus sp., Claroideoglomus etunicatum, Funneliformis mosseae and Glomeromycotina sp.; one ericoid mycorrhiza Sebacina sp. and three Trichoderma species - Trichoderma aerugineum, T. Americanum and T. simmonsii were significantly identified as indicator species under inoculum treatment (Fig. S3E); and only Glomeromycotina sp. was revealed under phytate treatment (Fig. S3F). Indicator species analysis under M1P1 revealed FASV241 (Sebacina sp.) and FASV46 (Chaetomium grande) in the root biotope, and twelve fungal ASVs in the rhizosphere, including Funneliformis mosseae (Table S7).
Determining Eco- And Core-microbiota
One hundred ASVs were ubiquitous in all roots and 115 ASVs in each rhizosphere and they were attributed as the bacterial eco-microbiota of soybean roots and rhizosphere, respectively (Table S8). The bacterial eco-microbiota in the root belonged to 19 genera, with 91 of them being Planctobacteria (Table S8), whereas the rhizosphere eco-microbiota belongs to 21 genera, with107 Planctobacteria ASVs (Table S9). Tepidisphaera mucosa and Gemmata sp.were detected in 32 ASVs in the root and rhizosphere eco-microbiota, respectively (Fig. S5B and Table S9). A Venn diagram identified 63 unique ASVs in the root and 78 distinct ASVs in the rhizosphere, with 37 ASVs shared by the two biotopes (Fig. S5C). Thirty-three of the 37 shared ASVs belonged to nine bacterial genera (Tepidisphaera mucosa, Gemmata sp., Planctomyces maris, Isosphaera sp., Pirellula sp., Planctomicrobium piriforme, Lacipirellula parvula, Calycomorphotria hydatis, Algisphaera agarilytica), while four were not assigned to any taxon (NA) (Fig. S5D). The shared bacterial taxa were the bacterial core-microbiota in soybean. In the fungal community, only ASV2 (Humicola fuscoatra) was discovered as eco-mycobiota in the root (Table S10A). Fifteen ASVs were identified as eco-mycobiota in the rhizosphere and assigned to 14 genera (Fig. S5E, Table S10B). ASV2 (Humicola fuscoatra) was found in both biotopes and has been considered as the core-mycobiota (Fig. S5F).
Bacteria Regulates The Connectivity Of Soybean Microbiota
The interkingdom co-occurrence network in the rhizosphere was more complex (452 nodes and 2159 edges) than in the roots (285 nodes and 553 edges) (Fig. 3A, B). Four bacterial ASVs (BASV6, BASV87, BASV16, and BASV58) were classified as hub taxa in the root based on their node degree and betweenness centrality and these hub taxas were Planctobacteria (Fig. S6A and Fig. S7A-H). Based on mutual putative interactions in the subnetwork of the hub taxa in the root, we found: (i) BASV6 (Tepidisphaera mucosa) had positive putative interactions with 13 different bacterial ASVs but negative interactions with two different bacterial ASVs (Fig. S7A, E); (ii) BASV16 (Tepidisphaera mucosa) interacted negatively with BASV476 but positively with nine different bacterial ASVs and a fungal ASV, FASV2 (Fig. S7B, F); (iii) BASV58 (Gemmata sp.) had positive putative interactions with seven different bacterial ASVs but negative interactions with four BASVs (Fig. S7C, G) and (iv) positive putative interactions with seven different bacterial ASVs and negative putative interactions with five bacterial ASVs were found in ASV87 (Tepidisphaera mucosa) (Fig. S7D, H). The interkingdom network in the rhizosphere identified four hub taxa- BASV175, BASV148, BASV200 and ASV311 as Thermostilla marina, Chloroflexus aurantiacus, Zavarzinella formosa and Gemmata sp., respectively (Fig. S6B and Fig. S7I-P) and their interaction pattern revealed: (i) BASV175 (Planctomyces maris) had positive putative interactions with 10 bacterial ASVs but negative interactions with six bacterial ASVs and three fungal ASVS (Fig. S7I, M); (ii) BASV200 (Zavarzinella formosa) had positive interactions with 10 bacterial ASVs and negative interactions with seven bacterial ASVs and a fungal ASV (FASV7) (Fig. S7J, N); (iii) BASV311 (Gemmata sp.) had positive interactions with 11 bacterial ASVs and negative interactions with five bacterial and a fugal ASV (FASV74) (Fig. S7K, O); (iv) BASV148 (Chloroflexus aurantiacus) had positive interactions with 16 bacterial and a fungal ASV (FASV16), and negative interactions with a bacterial ASV, BASV268 (Fig. S7L, P).
Meta co-occurrence patterns of hub taxa revealed a network of eight modules (Fig. 3C; Table S11): (i) Module I centered on BASV200 found in the rhizosphere was connected to Module II centered on the hub taxa BASV148 via BASV44 and BASV91; (ii) BASV35 connected Module II to Module III; (iii) Module IV centered on the root hub taxa BASV311 was linked to Module III through BASV308, and BASV577 connected Module IV and Module V; (iv) BASV14 connected Module VI and VII which are respectively centered on the hub taxa BASV87 and BASV16; (v) BASV6, the root interkingdom network’s hub taxa is linked to Module VII via BASV196, and Module VI via BASV57; (vi) BASV30 connected the Module VIII and the Module II; (vii) BASV268 and BASV132, respectively, connected Module V to Module II and the Module VIII (Fig. 3C). BASV91 connecting Module I, II and VIII; and BASV308 connecting Module III, IV and VIII (Fig. 3C). Overall, we identified 11 different bacterial ASVs that established interactions among eight different hub taxa to broaden interactions in soybean microbiota. As a result, these 11 ASVs and eight hub taxa, for a total of 19 ASVs have been designated as global hub taxa (Table S11). These 19 global hub taxa were assigned to seven genera (Tepidisphaera mucosa, Gemmata sp., Chloroflexus aurantiacus, Pirellula sp., Ralstonia solanacearum, Thermostilla marina and Zavarzinella formosa). Only BASV148 was a Chloroflexi member, while 18 of the 19 ASVs were Planctobacteria (Table S11).