Response of soil arbuscular mycorrhizal fungal community to chemical fertilization in sugarcane

Study site and experimental design

The experiment was conducted in 2020 at the experimental field of the Li-Jian Scientific Research Station of Guangxi Academy of Agricultural Sciences (23°14′N, 108°33′E), where average annual temperature is 21.7 °C, average annual humidity is 78.6%, average annual precipitation is 85.3 mm, and average annual sunshine length is 118.0 h. The field experiment was designed with randomized block with six duplicates. Each experimental plot consisted of five rows, with a row length of 7 m, a row spacing of 1.2 m and planting depth of 30 cm. Three rows of the same sugarcane variety were planted between different plots to serve as isolation areas. The basic physicochemical properties of the soil used were as follows: pH 5.1, organic matter 32.3 g/kg, total nitrogen 2.3 g/kg, total phosphorus 1.8 g/kg, total potassium 10.9 g/kg, water soluble nitrogen 51.97 mg/kg, available phosphorus 73.2 mg/kg, and available potassium 110 mg/kg. Two sugarcane varieties were chosen: Gui-Tang 58 (GT58), known for its extensive cultivation, and Gui-Tang 29 (GT29), characterized by strong ratoon and tillering ability. The specific characteristics of the sugarcane varieties are as follows:

GT58: This variety was selected from the cross of YT85-177 × CP81-1254. It is currently a widely cultivated variety. Its distinctive characteristics include medium plant height, large stalk diameter, more millable canes, slightly purplish-red leaf sheaths, underdeveloped hair groups, and easy leaf shedding. It has strong ratoon ability, strong adaptability, upright stalk, and lodging resistance, and lacks aerial roots. GT58 is known for high and stable cane yields, medium maturity, high sugar content, high resistance to smut disease, but average drought resistance.

GT29: The parentage of this variety is YC94-46 × ROC22. It has strong ratoon ability and can have more than nine ratoons in production. Its characteristics include erect and compact plant structure, easily detached leaves, intermediate stalk, uniform canes, early ripening, and high sugar. It has excellent sprouting ability and strong tillering ability, strong cold tolerance, and is highly resistant to smut.

The sugarcane planting date was January 8, 2020. Due to variations in tillering in different sugarcane varieties, especially the strong tillering ability of GT29, 120,000 buds/ha for GT58 and 75,000 buds/ha for GT29 were used, respectively. Chemical fertilizer was applied for two times. At planting, 40% of the base chemical fertilizer was applied, and 60% of the additional chemical fertilizer was applied just before earthing-up at early elongation stage. Considering the common chemical fertilization about 1,500 kg/ha used in commercial sugarcane production in Guangxi and referring to recent field research results related to sugarcane mycorrhizal studies, the experiment employed four different chemical fertilization treatments: T1 (0.0 kg/ha), T2 (562.5 kg/ha), T3 (1,125.0 kg/ha), and T4 (2,250.0 kg/ha). The chemical fertilizer used was controlled-release compound chemical fertilizer with an N: P: K ratio of 17:7:17 (product of Hebei Tian-Ren Chemical Industry Co., Ltd., Qinhuangdao, China). The crop was managed the same as for common commercial production, and harvest was done in December 2020.

Sugarcane agronomic trait measurements

After emergence, the sprouting rate of sugarcane was investigated on April 10, 2020. On May 26, 2020, the number of tillers per plant was surveyed. On November 20, 2020, the number of millable stalks, that is, the plant height was 1.3 m or more, was investigated. In December 2020, the crop was harvested, and 20 plants with similar growth were selected from the middle three rows. Plant height was measured using a tape measure, stalk diameter was measured using a vernier caliper, and hardness was measured using a hardness tester (MYT-1). After harvested, all the canes within the same plot were collected and weighed, and the data were used for estimation of the cane yield per hectare.

Soil and root sample collection

On December 10, 2020, samples of sugarcane roots and rhizosphere soil from 0–30 cm soil layer were collected using a five-point sampling method, that is, the five points of the middle and four corners in a plot. The soil around the roots of five sugarcane plants were collected each plot and thoroughly mixed as a soil sample. There were five replicates.

Soil sample collection: At each sampling point, 0.5–1.0 kg of rhizosphere soil from 0–30 cm depth was collected and thoroughly mixed. Using the quartering method, 0.5–1.0 kg of soil was taken from each plot as the soil sample. The soil samples were sieved through a 20–mesh sieve, and 800–1,000 mg portions were put into 2 mL centrifuge tubes. Each soil sample was divided into three tubes and stored at −80 °C for subsequent total DNA extraction.

Root sample collection: Concurrently with soil sampling, root samples were collected. At each sampling point, after rough separation of the collected roots and rhizosphere soil, the roots were mixed as one sample that contained roots of varied sizes. The root sample treatment was the same as described in the previous report (Zhang et al., 2024).

Root AMF colonization assessment

Root staining was conducted to observe AMF colonization structures. The root staining method was referred to Liao et al. (2016). The cleaned roots were cut into about 1 cm long segments, placed in embedding boxes, and placed in 20% KOH solution for 90–120 min in a 90 °C water bath. The root AMF colonization was determined according to Nong et al. (2023) using the Nikon microscope Eclipse Ci-L and DS-Ri2 to observe the root colonization, and the MYCOCALC software to calculate the colonization rate based on the results.

Soil AMF high-throughput sequencing

FastDNA^® Spin Kit for soil DNA extraction was used to extract DNA from soil samples, and the specific operation was conducted according to the instructions. Nested PCR was performed to amplify the extracted AMF DNA. Small subunit rRNA (SSU) was selected for amplification. The high-throughput sequencing followed the procedure described in our previous report (Nong et al., 2023).

Data analyses

Trimmomatic 0.38 was used for quality control of raw sequencing data, followed by paired-end assembly using Flash 1.2.11. UPARSE 11 was employed to cluster sequences into distinct Operational Taxonomic Units (OTUs) based on a 97% similarity threshold. The resulted OTUs were aligned to the MaarjAM database for sequence matching and annotated using RDP Classifier with a confidence threshold of 0.7. Unclassified sequences at the phylum level were filtered out. The OTU table was normalized to minimum sample sequence counts, and subsequent analyses were performed. Mothur was used for α-diversity analysis, displaying species diversity with the Shannon index and richness with the Chao1 index. The Bray-Curtis dissimilarity matrix was built to estimate the β-diversity distances, Principal Coordinates Analysis (PCoA) was done to depict the influence of diverse fertilization treatments on the AMF community around sugarcane roots. The Adonis test was done to assess the significance in AMF communities in different fertilization treatments. The agronomic traits were compared using one-way ANOVA and Duncan’s test. Correlations between sugarcane agronomic traits and diverse AMF genera were examined through Mantel tests, while Pearson correlation analysis was done to explore the relationships between various sugarcane agronomic traits.

Effects of chemical fertilization on agronomic traits of sugarcane

The agronomic traits in sugarcane were influenced by chemical fertilization. For GT58, the sprouting rate was significantly affected by different chemical fertilization rates. T4 showed no significant difference from T2 (P = 0.067 > 0.05) but exhibited significant increases than T1 and T3 (P = 0.012 < 0.05). The tiller number was significantly impacted by different fertilization treatments. The number of tillers in T1, T2, T3 and T4 were 138.23, 146.40, 152.28 and 153.83, respectively. No significant differences were observed between T4, T3, and T2, however, both T4 and T3 showed significant increases than T1 by 11% and 10%, respectively. Moreover, the different fertilization rates significantly influenced the millable stalks in both GT58 and GT29. For GT58, there was no significant difference between T4 and T3, yet that in T4 was significantly higher than T1 and T2 by 13% and 8%, respectively. For GT29, there were no significant differences in T4, T3, and T2, however, T4 and T3 both showed a significant increase than T1 by 6%. Conversely, the plant height and stalk diameter in both GT58 and GT29 were not significantly affected by different fertilization rates (Table 1).

The impact of different chemical fertilization rates on cane yield and brix is illustrated in Fig. 1. Fertilization significantly affected the cane yield in GT58, while did not in GT29. For GT58, there was no significant difference in cane yield between T4, T3, and T2, however, both T4 and T3 show significant higher cane yield by 13% than T1. Different fertilization rates significantly influenced the brix in GT29, but not in GT58. For GT29, T3 exhibited the highest brix (22.31°Bx), which was 6% higher than that in T2 (22.13°Bx).

Mycorrhizal colonization in sugarcane roots with different chemical fertilization rates

It showed that different chemical fertilization treatments for both sugarcane varieties established favorable symbiotic relationships with AMF, with colonization rates exceeding 90%. However, there was no significant difference in colonization rates among different fertilization treatments (P = 0.037 < 0.05). For GT58, varying fertilization rates significantly influenced the colonization intensity, arbuscular abundance, and relative arbuscular abundance in sugarcane roots. Notably, T1 showed significant differences from T3 while no significant differences from T2 and T4. For GT29, colonization intensity and relative colonization intensity exhibited the pattern T3 > T2 > T1 > T4, showing a 47% increase in colonization intensity in T3 compared to that in T4. Arbuscular abundance and relative arbuscular abundance followed the pattern T2 > T1 > T3 > T4 in GT29, and T2 and T4 exhibited significant differences, demonstrating a 164% increase in arbuscular abundance and a 141% increase in relative arbuscular abundance in T2 compared to those in T4 (Table 2).

AMF species composition under different chemical fertilization rates

A total of 258 OTUs were identified in the four different chemical fertilization levels in GT58. The number of OTUs for the fertilization level T1, T2, T3, and T4 was 70, 65, 68, and 55, respectively. Among them, there were three unique OTUs in T1, none in T2, 5 in T3, and none in T4. Additionally, there were 44 common OTUs shared for all the treatments, representing 17.05% of the total OTUs. For sugarcane variety GT29, a total of 273 OTUs were detected in the four chemical fertilization treatments. The OTU counts for T1, T2, T3, and T4 were 70, 71, 73, and 59, respectively. The number of unique OTUs in each level was 3, 1, 2, and 0, respectively, while there were 51 common OTUs for all the treatments, accounting for 18.68% of the total OTUs (Fig. 2).

Diversity of AMF in sugarcane rhizosphere soil under different chemical fertilization rates

The impact of various fertilizations on the richness and diversity of AMF communities in the rhizosphere soils of different sugarcane varieties was observed. In the rhizosphere soil of sugarcane variety GT58, the Chao1 and Shannon indices were higher in T1, T2, and T3 compared to T4, and significant differences were found between T3 and T4 (P = 0.036 < 0.05). The diversity and richness exhibited a dynamic trend of increase, followed by decrease with increasing fertilization, reaching the highest at the T3 level. Conversely, the diversity and richness of AMF communities in the rhizosphere soil of sugarcane variety GT29 were not significantly influenced by different fertilization levels. This suggests that fertilization has a significant impact on the AMF community in the rhizosphere of GT58, but not in that of GT29 (Fig. 3).

AMF communities in sugarcane rhizosphere soil under different chemical fertilization rates

The composition of AMF communities in sugarcane rhizosphere soil under different chemical fertilization rates was illustrated in Fig. 4. For sugarcane variety GT58, the dominant genus of AMF in all the fertilization treatments was Glomus, with relative abundances exceeding 94%, and the highest abundance was observed in T3 fertilization treatment, reaching 99.74%. For different fertilization levels, T1, T2, T3, and T4 exhibited 4, 4, 5, and 3 fungal genera, respectively. The genera Glomus, Acaulospora, Paraglomus, and Scutellospora were present in T1, T2, and T3 treatments, while only Glomus, Acaulospora, and Paraglomus were found in T4 treatment. The genus Diversispora was exclusively present in T3 treatment. Similarly, for sugarcane variety GT29, the dominant genus of AMF was also Glomus, with relative abundances exceeding 93%, and the highest abundance was observed in T3 treatment, reaching 98.48%. For different fertilization levels, T1, T2, T3, and T4 exhibited 5, 4, 4, and 3 fungal genera, respectively. The genera Glomus, Acaulospora, and Paraglomus were present in all the fertilization treatments, while Scutellospora was present only in T1 and T2 treatments, with the relative abundances of 0.63% and 2.49%, respectively. The genus Diversispora was exclusively present in T1 and T3 treatments, with the relative abundances of 0.17% and 0.02%, respectively.

Discrepancy of AMF community in sugarcane rhizosphere soil under different chemical fertilization rates

Principal coordinates analysis (PCoA) was performed at OTU level to explore the similarity and dissimilarity in AMF community in sugarcane rhizosphere soils with different chemical fertilization treatments. As shown in Fig. 5, for sugarcane variety GT58, the PCoA 1 and 2 axes explained 54.71% and 20.97% of the variation, respectively. For sugarcane variety GT29, the PCoA 1 and 2 axes explained 39.52% and 28.61% of the variation, respectively. The samples of rhizosphere soil in GT58 with the same fertilization rate were clustered closely, indicating minor differences in the AMF community structure. Conversely, those in GT29 with the same fertilization level showed more dispersion, suggesting greater variability in the AMF community structure. The Adonis test results revealed a significant effect of chemical fertilization on the AMF community in the rhizosphere soil in GT58, but not in that in GT29 at P = 0.753 > 0.05 level.

Distinctive OTUs in sugarcane rhizosphere soil under different chemical fertilization rates

Because the result that different chemical fertilization rates only significantly affected the AMF community with GT58, the subsequent analyses focused on identifying the distinctive OTUs in the rhizosphere soil of GT58 under different fertilization levels. The results revealed a total of 5 distinctive OTUs, all belonging to the genus Glomus. These OTUs were identified as OTU 41, OTU 74, OTU 85, OTU 122, and OTU 140. Among these distinctive species, the highest abundance was consistently observed in T3 treatment. Specifically, for OTUs 41 and 122, T3 exhibited significantly higher relative abundances compared to the other fertilization treatments, reaching 0.30% and 0.03%, respectively. For OTU 74, T3 (3.33%) showed significant differences from both T1 (0.02%) and T4 (0.31%). Significant differences were also detected between T3 and T1 for OTU 85, as well as between T3 and both T1 (0.00%) and T2 (0.01%) for OTU 140 (Fig. 6).

Relationship between different AMF genera communities and agronomic traits

Pearson correlation analysis revealed significant positive correlations between cane yield and sprouting rate, plant height, and stalk diameter, while brix was significantly and negatively correlated with tiller number and millable stalks. Through Mantel tests between sugarcane agronomic traits and AMF genera community, it was found that the genus Acaulospora showed a highly significant positive correlation (P = 0.000 < 0.001) with millable stalks and a significant positive correlation (P = 0.026 < 0.05) with tiller number. The genus Scutellospora exhibited a significant positive correlation with sprouting rate. However, there were no significant correlations between the other AMF genera and sugarcane agronomic traits (Fig. 7).