Heavy grazing reduces soil bacterial diversity by increasing soil pH in a semi-arid steppe

Study site and experimental design

Our experiment was performed in the Xilamuren desert steppe (41°15′10″N, 111°13′26″E), Inner Mongolia, China. The mean annual precipitation and temperature are 280 mm and 2.5 °C, respectively. A dominant plant is Stipa krylovii, while S. breviflora, Cleistogenes songorica and Artemisia frigida are common species in this area. The soil type at this site is loamy sand texture.

A randomized complete block design was utilized in this experiment. Three blocks were established and each of them contained three treatments, with the total of 9 plots in the experimental area. The three treatments were no grazing (with a fenced enclosure undisturbed by livestock; CK), light grazing (with a stocking rate of 0.63 sheep/ha/year (3 horses); LG) and heavy grazing (with a stocking rate of 1.05 sheep/ha/year (5 horses); HG). The grazing platform has been implemented in each growing season (i.e., from June to September) since 2012. Adult Mongolian horses were selected to graze. The detailed information on study site and experimental design can be found in our previous study (Wang et al., 2022c).

Sampling and measurement

In August 2021, we assessed species composition, and the height, density, individual biomass and relative density of S. krylovii population based on ten 1 m × 1 m quadrats per plot. Then, we randomly selected 30 clumps of S. krylovii (30 cm soil layer each plant individual) in each plot, removed the loose soil on the surface of roots, and obtained the rhizosphere soil (RS) by the shaking method. Every ten samples were mixed into one composite sample. At the same time, ten non-rhizosphere soils (NRS) of equal depth were excavated along the Z-shape and mixed into one composite sample in each plot. In addition, we collected other soil cores for soil bulk density. The composite samples were passed through a two mm sieve and divided into two parts. One part was naturally dried at room temperature for the determination of soil chemical properties except for soil inorganic nitrogen content with fresh soil, and the other part was placed at −80 °C for the determination of microbial diversity. Data on inorganic nitrogen (${\mathrm{NH}}_4^{+}$-N and ${\mathrm{NO}}_3^{-}$-N), soil pH, total carbon (TC), total nitrogen (TN), soil organic carbon (SOC), total phosphorus (TP), available phosphorus (AP) and soil bulk density were collected as previously described in Zhang et al. (2020a).

Determination of soil microbial diversity

After the soil sample was frozen for two days, the DNA of microbial was extracted from soil using the PowerSoil DNA Isolation Kit (MO BIO laboratories, Carlsbad, CA, USA). The DNA extract was determined on 1% agarose gel, and the concentration and purity of DNA were examined using a NanoDrop 2000 UV-Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The V4 region of the 16S gene was sequenced by PCR: 515 F (GTGCCAGCMGCCGCGGTAA) and 806 R (GGACTACHVGGGTWTCTAAT), while PCR: ITS5-1737F (CTTGGTCAT TTAGAGGAAGTAA) and ITS2-2043R (GCTGCGTTCTTCATCGATGC) was sequenced for the ITS1 region gene. The PCR product was tested in a 2% agarose gel. The construction of library was established in a TruSeq® DNA PCR-Free Sample Preparation Kit. The sequencing work was completed by Beijing NuoHeZhiYuan Technology Co. Ltd. The 16S rRNA and ITS rRNA sequences are available at NCBI Sequence Read Archive (SRA) database (Accession Number: PRJNA1014180 and PRJNA1014553).

Statistical analysis

To test the interaction effects between grazing intensities (i.e., CK, LG and HG) and soil positions (both in rhizosphere and non-rhizosphere soils) treatments on the microbial diversity and soil chemical properties, we used a two-way ANOVA with grazing treatments as the main-plot effect, and soil position treatments as the sub-plot effect. Microbial diversity was calculated based on OTUs number, including Richness, Shannon-Wiener index and Chao1 index, calculated by the following formula:

Richness = S_obs

Shannon-Wiener index: $H_{\text{shannon}}=-{\sum }_{i=1}^{S_{\text{obs}}}\frac{n_i}{N}ln\frac{n_i}{N}$

Chao1 index: S_chao1 = S_obs + n₁(n₁ − 1)/2(n₂ + 1).

Where: S_obs is the OTUs number actually observed; S_chao1 is the estimated OTUs number; n₁ represents the number of OTUs containing only one sequence; n₂ represents the number of OTUs containing only two sequences; n_i is the number of sequences contained in the i th OTUs; N represents all the sequence numbers. Besides, spearman correlation analysis was used to assess the relationships between the microbial diversity (i.e., both bacterial and fungal) and soil chemical properties. Data analysis was conducted using SPSS 19.0 (IBM Corp., Armonk, NY, USA), and all figures were prepared using Origin 2021 (OriginLab Corporation, Northampton, MA, USA).

The interaction effects between grazing intensities and soil positions on both bacterial and fungal community diversity

Heavy grazing decreased diversity (i.e., richness, Shannon and Chao1) of bacterial community in the non-rhizosphere soil (F = 26.915, P = 0.001, F = 10.96, P = 0.01, F = 8.625, P = 0.017, for richness, Shannon and Chao1, respectively; Figs. 1A, 1C and 1E), however, there had no significant differences in diversity (i.e., richness, Shannon and Chao1) of bacterial community among grazing treatments in the rhizosphere soil (F = 0.514, P = 0.604, F = 1.505, P = 0.242, F = 0.336, P = 0.718, for richness, Shannon and Chao1, respectively; Figs. 1A, 1C and 1E). Overall, bacterial diversity (i.e., richness and Chao1) in the rhizosphere soil was higher than that of non-rhizosphere soil only in the heavy grazing treatment (P < 0.001 and 0.042 for richness and Chao1, respectively; Figs. 1A and 1E). There was significant variation in bacterial diversity (i.e., richness, Shannon and Chao1) among grazing intensity treatments, as well as soil position treatments, and a significant interaction between grazing intensity and soil position treatments (Table 1).

For fungal community, however, there had no significant differences in fungal diversity (i.e., richness, Shannon and Chao1) among grazing treatments both in the rhizosphere and non-rhizosphere soils (for the rhizosphere soil, F = 0.814, P = 0.455, F = 0.769, P = 0.474, F = 0.774, P = 0.472, for richness, Shannon and Chao1, respectively, for the non-rhizosphere soil, F = 3.163, P = 0.115, F = 1.199, P = 0.365, F = 3.93, P = 0.081, for richness, Shannon and Chao1, respectively; Figs. 1B, 1D and 1F). Likewise, there had no significant differences in fungal diversity (i.e., richness, Shannon and Chao1) between rhizosphere and non-rhizosphere soil in all grazing intensity treatments (all P > 0.05; Figs. 1B, 1D and 1F). There had no significant interaction in fungal diversity (i.e., Shannon and Chao1) between grazing intensity and soil position treatments, except for richness of fungal community (Table 1).

The interaction effects between grazing intensities and soil positions on soil chemical properties

For the rhizosphere soil, light grazing increased soil pH value but decreased ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N and inorganic N content (F = 14.849, P = 0.005, F = 74.575, P < 0.001, F = 184.481, P < 0.001, F = 91.036, P < 0.001, for soil pH value, ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N and inorganic N content, respectively; Figs. 2A, 2E, 2F and 2G), while heavy grazing increased soil pH value, total P, ${\mathrm{NO}}_3^{-}$-N, and inorganic N content (F = 14.849, P = 0.005, F = 11.679, P = 0.001, F = 184.481, P < 0.001, F = 91.036, P < 0.001, respectively; Figs. 2A, 2C, 2F and 2G), but decreased available P and ${\mathrm{NH}}_4^{+}$-N (F = 5.575, P = 0.043, F = 74.575, P < 0.001, respectively; Figs. 2D and 2E). For the non-rhizosphere soil, however, light grazing decreased soil pH value and available P (F = 42.304, P < 0.001, F = 13.561, P = 0.006, respectively; Figs. 2A and 2D), while heavy grazing increased soil pH value but decreased available P and ${\mathrm{NH}}_4^{+}$-N content (F = 42.304, P < 0.001, F = 13.561, P = 0.006, F = 5.184, P = 0.049, respectively; Figs. 2A, 2D and 2E).

Besides, soil pH value in the non-rhizosphere soil was higher than that of rhizosphere soil in all grazing treatments (P = 0.007, P = 0.003, P < 0.001, for control, LG and HG treatment, respectively; Fig. 2A). Conversely, soil ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N, inorganic N content in the non-rhizosphere soil was lower than that of rhizosphere soil in all grazing treatments (all P < 0.001 in the control, P = 0.023, 0.027 and 0.011 in the light grazing treatment, all P < 0.001 in the heavy grazing treatment for ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N, inorganic N content, respectively; Figs. 2E, 2F and 2G). Besides, soil organic carbon in the non-rhizosphere soil was lower than that of rhizosphere soil only in the control treatment (P = 0.017; Fig. 2B), while both soil total P and total N content in the non-rhizosphere soil was lower than that of rhizosphere soil only in the heavy grazing treatment (P = 0.049 and 0.004, respectively; Figs. 2C and 2I). Further, soil available P content in the non-rhizosphere soil was lower than that of rhizosphere soil both in the light and heavy grazing treatment (P < 0.001, P = 0.003, respectively; Fig. 2D), while total C content in the non-rhizosphere soil was lower than that of rhizosphere soil both in the control and heavy grazing treatment (P = 0.018 and 0.012, respectively; Fig. 2H). A significant interaction between grazing intensities and soil position treatments in soil pH value, ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N, inorganic N content; however, there had no significant interaction in soil organic carbon, available P, total C, N and P content (Table 1).

Relationships between diversity of microbial community and soil chemical properties

In the non-rhizosphere soil, the diversity of bacterial community (i.e., richness, Shannon and Chao1) was significantly negatively correlated with the soil pH value (all P < 0.05, r =  − 0.80, −0.67 and −0.78, respectively; Fig. 3). For fungal community, however, the Chao1 was only significantly positively correlated with available P content in the non-rhizosphere soil (r = 0.76, P > 0.05; Fig. 3). Also, soil available P content was significantly positively correlated with ${\mathrm{NH}}_4^{+}$-N content but negatively correlated with ${\mathrm{NO}}_3^{-}$-N content (all P > 0.05; r = 0.82 and −0.76, respectively; Fig. 3), while soil ${\mathrm{NH}}_4^{+}$-N content was significantly negatively correlated with ${\mathrm{NO}}_3^{-}$-N content (r =  − 0.87, P > 0.05; Fig. 3). In the rhizosphere soil, soil ${\mathrm{NH}}_4^{+}$-N content was significantly negatively correlated with the richness of bacteria and fungi, soil pH value (all P < 0.05, r =  − 0.80, −0.70 and −0.80, respectively; Fig. S1). Soil total P in the rhizosphere soil was significantly positively correlated with soil pH but negatively correlated with soil available P (all P < 0.05, r = 0.67 and −0.73, respectively; Fig. S1). Soil total N in the rhizosphere soil was significantly negatively correlated with soil available P but positively correlated with total C (all P < 0.05, r =  − 0.74 and 0.70, respectively; Fig. S1).