The Green Manure ( Astragalus sinicus L.) Improved Rice Yield and Quality and Changed Soil Microbial Communities of Rice in the Karst Mountains Area

: The use of green manure plants for soil restoration is a viable agricultural practice that can mitigate soil degradation and biodiversity loss caused by the long-term application of inorganic fertilizers. However, the effects of green manure on soil microbial communities, rice yield, and quality in the karst mountains are largely unknown. The effects of no chemical fertilizer, chemical fertilizer, chemical fertilizer + different Astragalus sinicus L. (Chinese milk vetch, CMV) treatments on the microbial community, soil enzyme activities, soil nutrient content, and crop yield were investigated through ﬁeld experiments. A moderate application of chemical fertilizer with green manure can increase chlorophyll content, increase effective rice spikes, positive impact on rice yield, and increase crude protein, etc. Additional application of the moderate amount of CMV can increase alkali-hydrolyzable nitrogen and available phosphorus (a signiﬁcant increase of 54.87–72.65%), improve microbial biomass carbon (MBC) and microbial biomass nitrogen (MBN) (CFMV2 signiﬁcantly increased by 22.16%), improve soil urease and phosphatase activities, and the urease activity increased by 43.43–69.24% with CMV application compared to CK. Moreover, all bacterial communities were dominated by three major phyla ( Proteobacteria , Chloroﬂexi , and Acidobacteria ), where the application of chemical fertilizer with CMV increased the ratio of abundance of Proteobacteria and Acidobacteria in rice soils, and the effect of chemical fertilizer application on the dominant bacteria was regulated to some extent by additional green manure application, which may have a beneﬁcial effect on rice yield. Therefore, we conclude that the rational use of chemical fertilizers with CMV (22,500 kg ha − 1 ) in karst landscapes is one of the effective measures to achieve efﬁcient and sustainable use of rice ﬁelds.


Introduction
Rice is one of the most important food crops, with nearly half of the world's population relying on rice as a staple food. In the context of climate change and increasing competition for land, water, labor, and energy, the demand in the world for rice is likely to increase further [1]. However, the human and environmental costs of expanding agricultural lands are such expensive that most of the necessary production gains must be achieved on existing farmland.
The Southwest of China is famous for its karst area, and its soil is one of the major ecologically fragile areas in the world [2]. Due to the special geological and climatic con-was Zhongyou 5617 (three-line hybrid rice, Single-season rice), rice was planted according to local farmers' practice, and the transplanting row spacing is 20 cm × 30 cm. The following five experimental treatments were 3 replicates, with plots (5 m long and 3 m wide) and blocks divided by 0.5 m wide ridges (ridge height 30 cm), plastic film was used to cover the ridges to prevent water and fertilizer runoff. The five treatments of the experiment were given by the following: CK (no chemical fertilizers and no CMV), CF (chemical fertilizers), CFMV1 (chemical fertilizers + 11,250 kg ha −1 CMV), CFMV2 (chemical fertilizers + 22,500 kg ha −1 CMV), and CFMV3 (chemical fertilizers + 45,000 kg ha −1 CMV). CMV was directly sown at the recommended rate (45 kg ha −1 ) in mid-October 2018, the CK and CF treatment were not planted with green manure, and the amount of CMV was adjusted according to the experimental design. Aboveground biomass of CMV was incorporated into the soil at the blooming stage by plowing and puddling approximately 10 days (early April 2019) before rice planting. The nutrient content of CMV at the blooming period: N = 3.00%, p = 0.33%, K = 3.00%, C/N = 14.25. The amount of chemical fertilizer applied to rice was: 180 kg ha −1 urea (N 46%), 120 kg ha −1 superphosphate (P 2 O 5 12%), and 120 kg ha −1 potassium sulfate (K 2 O 50%). A total of 50% urea and all CMV, P, and K fertilizers were applied as base fertilizer at any time, whereas 30% were applied as top fertilizer at the tillering stage and the other 20% at the panicle initiation stage.
10 soil (0-20 cm) samples were taken from each plot and mixed homogeneously after the late rice harvest [5]. The soil collected from each plot was evenly mixed to form a composite sample. After removing visible plant and stone fragments, the soil sample passed through a 2 mm sieve. Each sample was divided into two parts. One part of the soil sample was air-dried for determination of chemical properties and the other portion was stored at 5-10 g in sterile centrifuge tubes at -80 • C for microbial analysis [5].

Rice Physicochemical and Basic Soil Analysis
Plant height, chlorophyll, and other plant traits were tested in the field; rice yield was tested at the harvest stage. Crude protein, reducing sugars, and nutrients were sampled and tested by laboratory analysis at the harvest stage. The specific test method is as follows: reducing sugar was analyzed with the Ferring reagent method (GB/T 5513-2019); crude protein was measured following the Semi-trace Kjeldahl method (NY/T 3-1982); chlorophyll was measured using a plant nutrient meter (Zhejiang Top Instruments Co., Hangzhou, China).
The soil for this experiment was yellow rice soil, and the specific soil properties are presented in Table 1. The Water leaching potential method was used to determine pH (NY/T1121. ; SOM was determined using the high-temperature external heat potassium dichromate oxidation capacity method (NY/T1121. ; TN was analyzed with the Kjeldahl method (NY/T 53-1987); TP was measured following the molybdenum blue colorimetric method (NY/T 88-1988); TK was determined using a flame photometer (NY/T 87-1988); AHN, AP, AK of soil were measured by the alkaline hydrolysis diffusion method (DB51/T 1875-2014), NaHCO 3 method (NY/T 1121.7-2014) and NH 4 OAc and determined by the flame photometry method (NY/T 889-2004), respectively; microbial biomass carbon (MBC) and microbial biomass nitrogen (MBN) was estimated using the determination by fumigation extraction (GB/T 39228-2020). Besides, soil urease activity was determined by the phenol-sodium hypochlorite colorimetric method; soil phosphatase activity was analyzed with the colorimetric method using sodium benzene phosphate [19]; protease activity was measured following the colorimetric method [20].
For rice, reducing sugar was analyzed with the Ferring reagent method (GB/T 5513-2019); crude protein was measured following Semi-trace Kjeldahl method (NY/T 3-1982); chlorophyll was measured using a plant nutrient meter (Zhejiang Top Instruments Co., Hangzhou, China).

Soil DNA Extraction, PCR Amplification, and Sequencing
To represent the microbial properties, soil samples were collected at the end of the rice maturity immediately and stored at −80 • C for microbial analysis. Each sample of DNA was isolated from 0.40 g of soil using the FastDNA ® Spin Kit for Soil (MP Biomedicals, Santa Ana, CA, USA) following the manufacturer's instructions. DNA quality and concentration were checked with a NanoDrop ® ND-2000c UV-Vis spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The hypervariable region V3−V4 of the bacterial 16S rRNA gene was amplified with 338F/806R (5 -ACTCCT ACGGGAGGCAGCAG-3 /5 -GGACTACHVGGG TWTCTAAT-3 ) primers to characterize the bacterial community [21]. The Illumina pair-end library preparation and Miseq PE300 sequencing were determined by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China) following standard protocols. The Quantitative Insight into Microbial Ecology (QIIME) pipeline was used to analyze the sequence data. Following quality control adopted, 740,143 high-quality sequences were received from the 15 samples for bacteria. The quality-filtered sequences were clustered into operational taxonomic units (OTUs) according to the 97% sequences similarity. The data were analyzed on the free online platform of Majorbio Cloud Platform (https://cloud.majorbio.com/, accessed on 7 September 2021). A resampling procedure was performed at the same reads before the calculation. The observed species, Chao, and Shannon indexes were used to assess the diversity of the bacteria.

Statistical Analysis
Microsoft Excel 2016 was used for data processing, and one-way analysis of variance (ANOVA) and Duncan's post hoc test was performed for data analysis using IBM Statistics SPSS version 24.0 [22]. To evaluate the change in the relative abundance of the bacterial community, sequences produced by bacterial 16S rRNA were analyzed using the free online platform of Majorbio Cloud Platform (www.majorbio.com, accessed on 1 June 2022). The representative sequences of OTUs (operational taxonomic unit) with 97% similarity level were analyzed by cluster command and the representative OTUs were compared with Silva (Release138 http://www.arb-silva.de, accessed on 1 June 2022) ribosomal database for bacterial sequence. The classified OTU tables were randomly sampled to the same sequencing depth; Shannon, Simpson, Chao1, and ACE were calculated by Qiime software using the weighted UniFrac method. Principal coordinates analysis (PCoA) and displacement multivariate analysis of variance (Adonis) based on sample Bray-Curtis distances (OTUs level) used RStudio (version 4.0.3) to draw dominant OTU heatmaps and correlation heatmaps; Canoco software was used to draw redundancy analysis (RDA); the correlation of OTUs (relative abundance ≥ 0.1%) was calculated by Spearman's correlation coefficient.

Changes in Soil Properties and Enzyme Activity
In this experiment, CMV had the effect of regulating soil pH, and the pH tends to decrease to some extent, but the change was not significant (p < 0.05; Table 1). Moreover, no significant differences in soil TN, TP, TK, SOM, AHN, AP, and AK were observed among treatments. Overall, TN, TP, AP, and AK showed an increasing trend in CF and CFMV and a decreasing trend in TK (except for CFMV2, where TP content was lower than CF; CFMV1 where AK content was lower than CF, Table 1). Chemical fertilizer application alone significantly reduced soil AHN at the rice harvesting stage, but AHN content recovered with the application of CMV. The application of chemical fertilizer increased AP in the soil, which was further increased by CMV application, with a significant increase of 72.65% and 54.87% for CFMV1 and CFMV2 compared with CK, and 44.74% for CFMV1 compared with CF.
In addition, chemical fertilizer application alone reduced soil MBC and MBN by 13% and 51.47%, respectively, compared to CK. While additional CMV application restored MBC and MBN (22.16% increase in MBN was evident for CFMV2), while excessive application of CMV reduced soil MBC (p < 0.05; Figure 1a) and MBN (Figure 1b), see Table S1 for raw data. The urease activity of the applied CMV was significantly different from CK and CF, and the urease activity increased by 43.43%-69.24% with CMV application compared to CK (p < 0.05, Figure 1c), the phosphatase activity of CMV treatment was not significant compared with CK (p < 0.05; Figure 1d). However, urease and phosphatase activities decreased in CF but recovered or even increased with the addition of CMV. There were no obvious differences in protease activities among the various treatments ( Figure 1e).

Microbial Composition, Diversity, and Community Structure in Response to CMV
To better explain the effect of CMV on microorganisms, α-diversity indices were calculated to describe the microbial species richness and diversity, as shown in Table 2. According to the indicators of OTUs, Chao, ACE, Shannon, and Simpson, there was no significant difference in the abundance and diversity of bacterial communities (p < 0.05; Table  2). However, there was a tendency for OTUs, Chao, ACE, and Shannon to be higher than CK for the application of CMV treatment and the opposite value for Simpson. When excessive CMV was applied, the relative abundance and diversity decreased instead (Table  2), which indicates that the moderate use of CMV increased soil bacterial abundance and diversity.
β-diversity was assessed using principal coordinates analysis (PCoA) to distinguish differences in the composition and structure of species communities. In this analysis, three

Microbial Composition, Diversity, and Community Structure in Response to CMV
To better explain the effect of CMV on microorganisms, α-diversity indices were calculated to describe the microbial species richness and diversity, as shown in Table 2. According to the indicators of OTUs, Chao, ACE, Shannon, and Simpson, there was no significant difference in the abundance and diversity of bacterial communities (p < 0.05; Table 2). However, there was a tendency for OTUs, Chao, ACE, and Shannon to be higher than CK for the application of CMV treatment and the opposite value for Simpson. When excessive CMV was applied, the relative abundance and diversity decreased instead (Table 2), which indicates that the moderate use of CMV increased soil bacterial abundance and diversity. β-diversity was assessed using principal coordinates analysis (PCoA) to distinguish differences in the composition and structure of species communities. In this analysis, three replicates of five samples were compared as 15 independent data sets, principal component axis 1 explained 41.58% of the variation in bacterial community structure, whereas the second axis explained 34.93% (Figure 2a), and the cumulative total explained 76.51% of the total variables, see Table S2 for raw data. The close distance between points CF and CFMV1 and CFMV2 indicates the high similarity of the bacterial flora between these treatment points. However, the CF, CFMV1, and CFMV2 sites were separated from the CK sites by some distance, indicating that the fertilization treatment affected the soil bacterial community structure. higher than CF and CFMV1 (Figure 3b). The average relative abundance of Proteobacteria showed a trend of CFMV3 > CK > CFMV2 > CFMV1 > CF. In contrast to the changes in Proteobacteria, the relative abundance of Chloroflexi (14.79%-22.94%) decreased after CMV application compared with CF ( Figure 2b). Although there was no significant variation between treatments, their mean relative abundance was higher than that of CK. Therefore, when CMV was added, Chloroflexi showed a declining trend.   According to the classifications of bacteria community at the phylum level, there were 16 phyla were identified (Figure 2b). The dominant phyla across all soil samples were Proteobacteria (28.10%-38.94%), Chloroflexi (14.79%-22.94%), Actinobacteria (10.45%-12.01%), and Acidobacteria (7.61%-12.22%), which accounted for more than 60.95%-96.11% of the bacterial sequences (Figure 2b). Proteobacteria were the dominant phylum with the highest relative abundance, and differences were observed in the mean relative abundance of Proteobacteria in each treatment (Figure 3a; Table S3), and CK was significantly higher than CF and CFMV1 (Figure 3b). The average relative abundance of Proteobacteria showed a trend of CFMV3 > CK > CFMV2 > CFMV1 > CF. In contrast to the changes in Proteobacteria, the relative abundance of Chloroflexi (14.79%-22.94%) decreased after CMV application compared with CF (Figure 2b). Although there was no significant variation between treatments, their mean relative abundance was higher than that of CK. Therefore, when CMV was added, Chloroflexi showed a declining trend.   Figure 4 represents the correlations between the distribution of bacterial communities and environmental factors in all treatments; see Table S4 for raw data. In this experiment, the distribution of bacterial communities was closely related to TP, TN, AP, MBC, MBN, and phosphatase (p < 0.05; Figure 4). Notably, different soil bacteria varied sensitivity to environmental factors (only the top 5 colonies in terms of relative abundance were discussed here). The abundance and diversity of Proteobacteria were a significantly negative correlation with MBN, and similar results were also found in Actinobacteria and Chloroflexi, which were significantly negatively correlated with MBC and significantly positively correlated with AP. However, the correlation of Bacteroidetes was not reflected in the heat map of this experiment. In addition, TN, TP, and phosphatase were also correlated with some bacteria in the soil, but the relative abundance of the affected flora was not significant and it is not discussed.   Table S4 for raw data. In this experiment, the distribution of bacterial communities was closely related to TP, TN, AP, MBC, MBN, and phosphatase (p < 0.05; Figure 4). Notably, different soil bacteria varied sensitivity to environmental factors (only the top 5 colonies in terms of relative abundance were discussed here). The abundance and diversity of Proteobacteria were a significantly negative correlation with MBN, and similar results were also found in Actinobacteria and Chloroflexi, which were significantly negatively correlated with MBC and significantly positively correlated with AP. However, the correlation of Bacteroidetes was not reflected in the heat map of this experiment. In addition, TN, TP, and phosphatase were also correlated with some bacteria in the soil, but the relative abundance of the affected flora was not significant and it is not discussed.

Changes in Rice Growth Index, Quality, and Yield
In this paper, the results indicated that various fertilizer treatments (including CF and CFMV) in rice plant height were significantly higher than that of CK, but no significant difference was observed in rice height between various CF and CFMV treatments (p < 0.05; Table 3). The number of rice effective panicles increased by 16.19%-29.08% in each fertilization compared with that of the CK, although CFMV and CF have no significant differences and the increase in CFMV was more pronounced (p < 0.05; Table 3). The application of chemical fertilizer and 22,500 kg ha −1 CMV contributed to rice yield, with CFMV2 yielding the highest, while the lowest rice yield was due to no application of chemical fertilizers. The results showed that turning a certain amount of CMV could promote the number of effective panicles and the plant height in rice, but over-application of CMV decreased rice yield.
In addition, in rice quality assessment, chlorophyll content did not differ significantly among the treatment groups at the jointing stage but decreased after the application of CMV; at the tillering stage, CFMV2 and CFMV3 increased significantly compared with the other treatments, it increased by 8.86% and 12.03% respectively to the CK; at the heading stage, the fertilizer treatment exhibited a noticeable improvement of 22.58%-34.68% than CK (p < 0.05; Table 3). Reducing sugars of rice under treatments has no significant difference compared with CK. While the crude protein content of rice showed a trend of CFMV3 > CFMV1 > CFMV2 > CF > CK, the results showed that the application of nitrogen fertilizer improved the crude protein content of rice, and the crude protein content of rice

Changes in Rice Growth Index, Quality, and Yield
In this paper, the results indicated that various fertilizer treatments (including CF and CFMV) in rice plant height were significantly higher than that of CK, but no significant difference was observed in rice height between various CF and CFMV treatments (p < 0.05; Table 3). The number of rice effective panicles increased by 16.19%-29.08% in each fertilization compared with that of the CK, although CFMV and CF have no significant differences and the increase in CFMV was more pronounced (p < 0.05; Table 3). The application of chemical fertilizer and 22,500 kg ha −1 CMV contributed to rice yield, with CFMV2 yielding the highest, while the lowest rice yield was due to no application of chemical fertilizers. The results showed that turning a certain amount of CMV could promote the number of effective panicles and the plant height in rice, but over-application of CMV decreased rice yield.
In addition, in rice quality assessment, chlorophyll content did not differ significantly among the treatment groups at the jointing stage but decreased after the application of CMV; at the tillering stage, CFMV2 and CFMV3 increased significantly compared with the other treatments, it increased by 8.86% and 12.03% respectively to the CK; at the heading stage, the fertilizer treatment exhibited a noticeable improvement of 22.58%-34.68% than CK (p < 0.05; Table 3). Reducing sugars of rice under treatments has no significant difference compared with CK. While the crude protein content of rice showed a trend of CFMV3 > CFMV1 > CFMV2 > CF > CK, the results showed that the application of nitrogen fertilizer improved the crude protein content of rice, and the crude protein content of rice was significantly higher than that in CK when CMV was applied, where CFMV3 was significantly higher than that in CF and CK (p < 0.05; Table 3).

Effects of Different Amounts of CMV on Soil Properties and Enzyme Activity
To maintain food security, it is necessary to study the effect of green manure cultivation on the drivers of rice yield [23,24]. Some trends appeared in soil pH, TN, TP, TK, SOM, AHN, AP, and AK among the treatments. High soil pH significantly affects the uptake of mineral nutrients by the aboveground parts of rice, which influences rice growth and thus yield [25,26]. The experiment showed a slight trend of decreasing pH, but the change in pH was not significant (Table 1), and there was only a certain mediation effect. Chemical fertilizer application alone significantly reduced soil AHN at the rice harvesting stage, and AHN content recovered with the application of CMV (Table 1). In fact, the content of soil AHN or TN was usually used to assess the indigenous nitrogen supply capacity [27]. Several studies showed that the application of green manure to crops continuously improved soil fertility as well as AHN content to reduce nitrogen fertilizer inputs to rice [28][29][30], and higher AHN of Pennisetum giganteum z.x.lin mixed with nitrogen-fixing biofertilizer application compared to the conventional application [31]. Moreover, when legume species are incorporated, they supply the soil with new N derived from the atmosphere (Ndfa), increasing the self-sufficiency of the system [32,33]. In the same trend as AHN, chemical fertilizer application increased AP in the soil, and CMV application further increased AP (Table 1). AP content increases with the use of chemical fertilizer and CMV, as well as the soil strongly adsorbs AP, leading to its release from the fertilizer; thus, different fertilization methods significantly alter soil properties [34].
MBC and MBN reflect the microbial scale and soil fertility status, and they are the living nutrient pool in the soil [35,36]. Our study showed that compared to chemical fertilizer alone, applying CMV with chemical fertilizers could further increase the soil MBC and MBN to higher levels. This outcome occurred because applying organic fertilizer along with chemical fertilizer improved the soil physicochemical properties, promoting the absorption and use of inorganic N by the rice crops, as well as the conversion of inorganic N into MBN and other organic forms of N [37]. In other words, organic inputs might positively affect MBC and MBN, but microbial colonization in soils [38,39], but the excessive application of CMV reduced MBC and MBN (Figure 1a,b). The study reported that the addition of biochar and straw has the effect of regulating the surface soil temperature, providing suitable temperature conditions for various enzymatic reactions and thus promoting the increase in soil microbial population [40].
In that experiment, urease and phosphatase activities decreased with chemical fertilizer alone and recovered or even increased with the addition of CMV (Figure 1c,d). Green manure can rearrange the fertilizer effect, thereby giving the soil a high nitrogen fixation capacity and high nitrogen use efficiency for that period [11,12], and the effective nitrogen content increased, thereby prompting the urease to rise, which was consistent with the enhancement of soil urease and phosphatase activities when straw and fertilizer were applied [41]. No significant difference in protease activity between different treatments ( Figure 1e). This result indicated that the application rate of CMV mainly affects urease and phosphatase activities. CMV application can eliminate the inhibitory effect of fertilizer application on urease and phosphatase, directly increase the fractions of labile Pi and NaHCO 3 -Po, and result in a significant direct effect on grain yields [42].

Effects of Different Amounts of CMV on the Bacterial Community Composition
Soil microorganisms play an essential role in the material and energy transformation of agroecosystems by participating in the energy flow and material cycling of soil ecosystems [43,44]. Bacterial abundance and diversity showed an increasing trend after CMV application [10]; when CMV was excessively applied, the relative abundance and diversity decreased instead (Table 2), which indicates that the moderate use of CMV increased soil bacterial abundance and diversity, the microbial community composition in the CMV treatment differed from the other treatments [45]. In fact, the relative abundance and diversity of soil microbial communities are affected by factors such as physicochemical properties and enzyme activity [21,45].
Taxonomic analysis of 16S rRNA sequences indicated that the phylum Proteobacteria was the dominant phylum with the highest relative abundance and CK was significantly higher than CF and CFMV1 (Figure 3a,b). Proteobacteria were considered a type of high nutrient availability that promotes copiotrophs [46,47]. Many studies had found that the relative abundance of some eutrophic microorganisms was lower when organic fertilizers were applied [48]. It is basically consistent with the results of CMV regulation of Proteobacteria in this experiment. In contrast to the changes in Proteobacteria, the relative abundance of Chloroflexi and Acidobacteria decreased after CMV application compared with CF (Figure 2b), but the relative abundance was higher than that of CK. Chloroflexi and Acidobacteria were slow-growing oligotrophic taxa with slower growth rates and the ability to metabolize nutrient-poor and recalcitrant C substrates in all likelihood. Members of this phylum were adapted to survive under the resource-limited conditions found in deeper horizons [47,49], and they decreased upon soil fertilization. However, Proteobacteria and Actinobacteriota were fast-growing copiotrophic taxa (those taxa that thrive in conditions of elevated C availability and exhibit relatively rapid growth rates), which facilitated the accumulation of soil organic carbon in the macroaggregates [47,50]. And the benefits to rice yield may be obtained by reducing the relative abundance of Nitrospirae and increasing the ratio of abundances of Proteobacteria and Acidobacteria in paddy soils [17].
The MBC in soil aggregates had a similar distribution as the soil organic carbon, and they were significantly positively related to each other [51], So MBC might affect the growth of Chloroflexi. In addition, chemical fertilizer or CMV application promoted an increase in AP, while AP was significantly and positively correlated with Chloroflexi, which could indicate that the increase in AP in soil was a synergistic increase in Chloroflexi abundance; the result is consistent with previous reports [52]. Some experiments reported that copiotrophic taxa have lower biomass C:N ratios and higher N demands than more oligotrophic taxa [46,47,53]. In addition, TN, TP, and phosphatase were also correlated with some bacteria in the soil, which is similar to the results [11,12]. However, the relative abundance of the affected flora was not significant and it is not discussed. Consistent with most of the experimental results showed that CMV planting in winter could improve the nutrient status of the soil because more CMV residues decompose in the soil, soil microbes produce C and N, thereby creating a positive cycle among fertilizer, crop, microbes, and soil components, and more C and N are fixed in the soil, thereby improving soil fertility and maintaining soil productivity [10,24].

Effects of Different Amounts of CMV on Single-Rice Cropping
Previous studies reported that green manure could be used as a substitute for chemical fertilizer without reducing rice yield [24]. The results showed that turning a certain amount of CMV could promote the number of effective panicles in rice, and CMV return to the field with N fertilizer promoted the plant height and chlorophyll content of rice and increased rice yield (Table 3). This phenomenon was mainly due to the fact that CMV promoted the colonization of soil microorganisms and their nutrient release during organic matter decomposition [11]. The process could improve crop yield by increasing the effective nutrient content in the soil, coordinating individual and group growth of rice, and promoting urea N and native soil N uptake by rice [10,24]. This finding was in line with previous studies that 22,500 kg ha −1 CMV promoted rice yield under the premise of chemical fertilizer application; CFMV2 showed the highest yield, while the lowest rice yield was due to no application of chemical fertilizer [9]. However, when the nitrogen fertilizer applied was higher than the theoretical nitrogen requirement of rice and CMV was over-applied, rice yield was decreased, which was mainly due to the excessive supply of nitrogen fertilizer that caused overgrowth of rice plants and increased uptake of nitrogen, leading to the greedy effect and late maturity [9]. In addition, excessive nitrogen would accumulate in rice stem sheaths and leaves, and fewer nutrients would be transferred to the seeds, thereby preventing the seeds from adequately filling [9]. Therefore, the application of the appropriate amount of CMV is conducive to improving soil fertility, maintaining soil productivity, and promoting the growth of soil microorganisms to increase the richness and diversity of the community, thus promoting the growth of rice and improving rice yield and quality.

Conclusions
This study provides important theoretical support to observe the effects of different amounts of CMV combined with chemical fertilizers on rice and soil and achieve more efficient resource utilization in karst landforms. Briefly, compositions and abundances of predominant bacteria in soils varied among the fertilizer treatments and all the bacterial communities were dominated by three major phyla (Proteobacteria, Chloroflexi, and Acidobacteria). Where the application of chemical fertilizer and CMV increased the ratio of abundance of Proteobacteria and Acidobacteria in rice soils, adjusting the effect of fertilizer on dominant bacteria, which may beneficially affect rice yield. A moderate application of chemical fertilizer with CMV (22,500 kg ha −1 ) can increase chlorophyll content, increase effective rice spikes, positive impact on rice yield, and increase crude protein, etc. Additional application of a moderate amount of CMV can increase AHN and AP, improve MBC and MBN, improve soil urease and phosphatase activities, and regulate the abundance of soil microbial dominant bacteria. The practice of mixing chemical fertilizers with CMV (22,500 kg ha −1 ) is encouraged in the karst mountains to support more sustainable agricultural practices in the rice-growing areas.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/agronomy12081851/s1, Table S1: Soil nutrient content and enzyme activity at rice harvesting stage, Table S2: Raw data from principal component analysis during the different treatment in Figure 2a, Table S3: Raw data of relative abundances of bacterial phyla in paddy soils in Figure 3, Table S4: Raw data for correlation heatmap analysis in  Data Availability Statement: Data can be shared through NCBI's BioProject database, and project information can be accessed at http://www.ncbi.nlm.nih.gov/bioproject/863100 (accessed on 1 June 2022).

Conflicts of Interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.