Fungi Outcompete Bacteria for Straw and Soil Organic Matter Mineralization

Background: Understanding the effects of straw return and nitrogen (N) fertilization on soil organic matter (SOM) transformations will help to mitigate climate change and maintain crop production and soil function. A 100-day soil incubation experiment was conducted using a two-factorial design with three fertilization levels and four 13 C-labeled maize straw and N addition treatments. The competition and contributions of the bacterial and fungal communities were assessed with relation to straw mineralization. Results: Mineral fertilizer alone and with straw increased straw decomposition by 59% and 55% and SOM mineralization by 27% and 37%, respectively, compared with the unfertilized soil, due to raised β-N-acetylglucosaminidase and cellobiohyrolase activities. Conversely, priming effect was decreased by 59% and 39%, respectively. Priming effect increased with higher N additions and decreased with lower N additions because an improved C:N ratio for microorganisms. Straw additions increased bacterial and fungal abundance by 1.4 and 4.9 times. Fungal diversity decreased with N fertilization because lower C:N ratios increased the bacterial competition. Bacterial abundance decreased but diversity increased with the duration of incubation as bacteria preferred to utilize labile organic compounds abundant in the initial stages. Along with labile organic compounds depletion, fungal abundance was increased. Firmicutes, Actinobacteria, and Proteobacteria bacterial as well as Ascomycota, Basidiomycota, and Mucoromycota fungi dominated straw and SOM decomposition. Firmicutes were mostly involved in straw and SOM mineralization on day one because of their capacity for labile compound decomposition. Integrated co-occurrence networks revealed that fungal taxa had a stronger correlation with straw decomposition than bacterial groups. Straw and N addition increased the number of negative edges among bacterial taxa but these decreased within fungal groups when compared to trials without straw and N. The ratio for pairwise correlations between abundant fungal taxa, straw, and SOM mineralization (29.9%) was greater than with bacteria (1.2%). Conclusions: increased competition within bacterial taxa. Fungi outcompete bacteria for straw and soil organic matter mineralization in long-term fertilized soils.

on the bacterial and fungal activity. High Ca levels in the karst soils can stabilize and protect SOC by inhibiting its decomposition [36,37]. Microbial activity and abundance are relatively higher in karst soils that contain abundant organic matter than in the same region of red soil with low SOC [38]. Consequently, straw additions to karst soils were more easily decomposed than in non-karst ecosystems, due to their neutral or alkaline pH and high microbial biomass [39][40][41][42]. However, the biological mechanisms that keystone species of bacteria and fungi utilize to regulate organic matter decomposition in karst soils have not yet been investigated.
This study aimed to evaluate how microorganisms respond to straw returns coupled with N additions in the long-term unfertilized and fertilized soils of the karst ecosystem. Microbial properties included microbial abundance, biomass; and composition: phospholipid fatty acids (PLFA), bacterial and fungal community compositions; and extracellular enzyme activities. We hypothesized that: (i) The initial SOM decomposition phase was dominated by an increasing bacterial abundance (e.g., Firmicutes) that consumes labile organic C at the initial stages, but increasing fungal biomass (e.g., Ascomycota) decomposes recalcitrant material in the later stages; (ii) Straw with low N levels decreased the priming effects (PE) because of a good C/N balance for the microorganisms, but increased PE with high N additions; (iii) straw alone and coupled with N increased enzyme activity (β-D-glucosidase, β-Nacetylglucosaminidase, and Cellobiohyrolase) as well as bacterial and fungal abundance, resulting in raising SOM decomposition; and (iv) Fungi played a more important role in straw and SOM decompositions than bacteria in the karst cropland because of their strong adaptability to soils with high levels of Ca. (iii) mineral fertilizers plus maize and soybean straws (NPK+Straw): 107 and 9 kg N ha -1 , 30 and 25 kg P ha -1 , as well as 40 and 52 K ha -1 , during the maize and soybean planting seasons, respectively. At the same time, soybean straw (5.4 Mg ha −1 , that is, about 93 kg N ha -1 , 9 kg P ha -1 , and 60 kg K ha −1 ) and maize straw (1.48 Mg ha −1 , that is, about 13.5 kg N ha -1 , 1 kg P ha -1 , and 4 kg K ha −1 ) were applied during the maize and soybean planting seasons, respectively.

Study site and experimental design
In the maize planting season, mineral and organic fertilizers were added three times: before planting the maize, during maize shoot growth, and during the reproductive growth stage. During the soybean planting season, mineral and organic fertilizers were applied two times, once before planting the soybeans and once at the pre-owering stage. Organic and mineral fertilizers were added next to the plant by hand and covered with soil during each fertilization event. The experimental site was previously described in great detail [33,43].
Soil samples (0-20 cm) were collected in November 2016 from plots with three fertilization treatments (Control, NPK, and NPK+Straw) using a soil corer with a diameter of 10 cm. The impurities, including animal (e.g., earthworm) and plant residues were removed using a 2-mm mesh sieve. The soil samples were stored at 4 °C in a refrigerator until the start of the incubation experiment. The soil properties for the three fertilization methods are presented in the Additional File: Table S9.
Preparation of 13 C-labeled maize straw The 13 C-straw preparation was conducted according to a previous study [44]. Brie y, at the seedling stage, maize was transferred to an automatically controlled gas-tight growth chamber system that was 1.1 × 2.5 × 1.8 m. 13 C-labeled CO 2 was introduced into the chamber, and was generated by acid base reactions with H 2 SO 4 (0.5 M) and Na 2 13 CO 3 (99 atom % 13 C; Cambridge Isotope Laboratories, Tewksbury, MA, USA). The CO 2 concentration was monitored using an infrared analyzer (Shsen-QZD, Qingdao, China). When the CO 2 concentration was higher than 450 μL L −1 , the excess CO 2 was absorbed automatically by pumping the chamber air through NaOH solution. There were temperature and humidity sensors (SNT-96S, Qingdao, China) inside and outside the chamber. The air-conditioning system controlled the internal air circulation and maintained the temperature between the inside chambers and ambient environment within 1 °C. Maize was grown for 60 days in the closed transparent chamber to label the maize with 13 CO 2 continuously. The preparation was completed when the maize was harvested.
The soil (200 g dry weight) samples for each fertilization condition (Control, NPK, and NPK+Straw) were placed in 1 L jars. An ammonium nitrate solution was prepared (low N: 0.07 g kg −1 ; high N: 0.2 g kg −1 ) and added to each jar together with or without the 13 C-maize straw. The jars were then incubated in the dark at 25 °C for 100 days. Soil samples were collected after 0, 1, 5, and 100 days. Subsamples were stored at 4 °C in a refrigerator for the microbial biomass and soil enzyme activity analyses. Subsamples were further freeze-dried and stored at -80 °C for PLFA and DNA extraction. The rest of the soil samples were dried for analysis of the soil physical and chemical properties. The gas samples for the CO 2 and 13 CO 2 were collected after 0, 2, 5, 7, 10, 15, and 20 days, and then collected every 10 days until 100 days.

Microbial biomass
The microbial biomasses of the samples were measured using a chloroform-fumigation extraction method. A 20 g sample of each soil, fumigated and non-fumigated, was extracted with 80 mL K 2 SO 4 (0.5 M). A conversion factor of k EC = 0.45 was used to determine the microbial biomass C (MBC) and microbial biomass N (MBN) concentrations. The contents of the dissolved organic C (DOC) and dissolved organic N (DON) were determined using the non-fumigated soils with a total C and N analyzer (Phoenix 8000, USA). Additionally, 6 mL of ltrate was absorbed and freeze-dried to measure the 13 C-MBC and 13 C-DOC using an isotope ratio mass spectrometer (Thermo Scienti c MAT 253, USA).

Soil enzyme activities
The activities of three enzymes with β-D-glucosidase (BG), β-N-acetylglucosaminidase (NAG), and cellobiohyrolase (CBH) were measured. Fresh soil (1 g) was shaken off and added to 125 mL acetate buffer (

Phospholipid fatty acid analyses
The microbial community compositions of the soils were characterized using PLFA analysis, as previously described [45]. Total soil lipids were extracted from approximately 8 g of the freeze-dried soil using a single-phase chloroform-methanol-citrate buffer (1:2:0. The quality of the raw sequences was assessed in FastQC v0.11.5. Adapters and sequences with low quality were trimmed with Trimomatic v0.36. Then, clean data for the 16S and 18S were processed using the QIIME2-2020.2 environment. The amplicon sequence variants (ASV) and ASV table were rst predicted using the DADA2 plugin with "-p-trunc-len" set as 250 bp or 310 bp for bacteria and fungi, respectively. Taxonomic analysis of the ASV was performed using the q2-feature-classi er, with the corresponding taxonomic classi ers, which were trained with the sequenced regions of the 16S/18S rRNA gene sequences, from the Silva rRNA reference database, release 132. The taxonomies of each ASV were further veri ed according to the taxdmp database (ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdmp.zip).
After discarding the ASV that were not annotated as fungi or bacteria, the total sequence number of each sample in the ASV table was rare ed to 10,000 in R.

Calculations and statistical analysis
Calculations of CO 2 uxes, organic carbon mineralization, and priming effects To calculate the fraction of the CO 2 derived from the straw and SOC, the end-member mixing model was used. We used the methodology described previously to calculate the CO 2 ux, C mineralization, and priming effect (PE) [49,50]. Added organic matter (e.g., straw) input induced microbial mineralization of the soil organic matter, generally causing a "priming effect" (PE), described as positive and negative PE [2]. C total = CO 2 ux × incubation time (1) C straw = ((δ 13 C total −δ 13 C control )/( δ 13 C straw − δ 13 C soil )) × C total (2) Where C total is the total CO 2 ux, derived from the straw-treated soil; C straw and C SOC are the proportions of CO 2 derived from the straw and SOC, respectively; δ 13 C total is the δ 13 C value (‰) of the total CO 2 (include the decomposition of the soil and straw); δ 13 C straw is the δ 13 C value (‰) of the straw-derived from CO 2 ; δ 13 C control is the δ 13 C value (‰) of CO 2 in the untreated soil; δ 13 C soil is the δ 13 C value (‰) of the initial soil.

Statistical analysis
Signi cant differences were estimated based on Duncan's multiple-range test (p < 0.05). The effects of soil fertilization, straw with increasing N addition levels, and incubation time on the straw decomposition, SOC mineralization, PE, and soil microbial pro les were performed using multivariate analysis of variance (MANOVA) in R. The Pearson correlation coe cients were calculated in R to investigate the relationships between the soil properties or microbial indexes and C turnover (straw decomposition, SOC mineralization, and PE). The response of the bacterial and fungal community compositions to the soil fertilization, straw with increasing N addition levels, and incubation time, were conducted by using nonmetric multidimensional scaling (NMDS) ordination and permutation multivariate analysis of variance (PERMANOVA), in R using the vegan package. We constructed an integrated network with all samples from each fertilization level (Control, NPK, and NPK+Straw), and evaluated the main ecological clusters with three modules, for both bacteria and fungi. For each module, nodes were highly inter-connected as a set and the rate of inter-module edges was lower than the intra-module edges, which were determined to be sub-units or communities [24]. The network analysis was constructed using a similar method to that described in previous studies [22,23]. First, all pairwise correlations between the bacterial or fungal taxa and environmental factors (including soil properties and SOC mineralization, straw decomposition, and PE) were calculated with an SAPRCC algorithm using Fastspar software. The signi cant level of SPARCC correlation was determined based on 1000-permutations. Only the top 10000 and 1800 robust (|r| > 0.3) and signi cant correlations (p < 0.05), were included in the network construction for bacteria and fungi, respectively. Second, the networks were visualized using the software Gephi v0.9.1, and further divided into clusters using a Louvain algorithm.
Similarly, sub-networks for the bacteria or fungi in each fertilization level, for each treatment (S0+N0, S+N0, S+N1, and S+N2), were also constructed. Furthermore, the ASVs were classi ed into abundant and rare taxa, dependent on the relative abundance in the network analysis. To accurately estimate the diversity and taxonomic compositions of the samples, the ASVs (i.e. 100 % of the operational taxonomic units (OTUs)) were identi ed from amplicon sequences in QIIME2 with the DADA2 pipeline. We de ned ASVs with relative abundances above 0.1 % for all sequences as "abundant" and those with a relative abundance below 0.1 % as "rare" [51,52]. In order to demonstrate the relationships between the bacterial or fungal ASVs and three indices (straw decomposition, SOC mineralization, and PE), subnetworks were also constructed and the top 1000 edges between the ASVs and these three indices were subset and illustrated using the R package igraph.

Results
Effects of straw and nitrogen additions on microbial biomass carbon and dissolved organic carbon Straw alone and with the N addition increased MBC and DOC. The contents of the MBC, 13 C-MBC, DOC, and 13 C-DOC increased rapidly and peaked after 3 to 5 days of incubation, followed by a gradual decrease with the straw and N additions (Additional File: Figure S1). For the three fertilization types, MBC and DOC increased in S+N0, S+N1, and S+N2 compared with S0+N0, with the highest MBC and DOC levels in S+N2 (Additional File: Figure S1). The 13 C-MBC and 13 C-DOC in the NPK+Straw soils were higher with the S+N2 than with the others throughout the rst month of incubation (Additional File: Figure S1). These results highlighted that the effect of the straw and N addition on MBC and DOC occurred in the initial decomposition stage and then disappeared until the end of the incubation. CO 2 emissions, straw decomposition, soil organic matter mineralization, and priming effects Long-term fertilization increased CO 2 and 13 CO 2 uxes but decreased PE. Regardless of the straw and N additions, the cumulative CO 2 ux and 13 CO 2 uxes increased (CO 2 : 50%; 13 CO 2 : 69%) with the NPK and NPK+Straw soils when compared with the unfertilized control soil. The CO 2 and 13 CO 2 uxes rapidly increased during the rst 3 days of incubation, followed by decreases, and then stabilization after 30 days. An increase in the PE of 48.8%, however, was observed for the unfertilized control soil, when compared with the NPK and NPK+Straw soils (Additional File: Figure S2). Consequently, priming intensity was found to decrease as mineral and organic fertilizers increased.
CO 2 uxes were increased in the treatments with the straw alone and with N additions for the three fertilization levels. Furthermore, the cumulative 13 CO 2 uxes as well as the straw decomposition were increased with high N levels only in the unfertilized control soil. The straw decomposition and cumulative 13 CO 2 showed no variation with the straw and N additions for the NPK and NPK+Straw soils ( Fig. 1a; Additional File: Figure S2). Following this, SOM mineralization was found to be increased by 124%, 41%, and 44% by the straw alone and with N additions when compared with the S0+N0 with the control, NPK, and NPK+Straw soils, respectively (Fig. 1b). Straw with high N addition levels increased the PE by 30% compared with the S+N0 and S+N1, for the three soil fertilization levels. The lowest PE value was found for S+N1, with the unfertilized control and NPK+Straw soils at the end of the 100-day incubation period (Fig. 1c). Overall, the CO 2 emissions derived from the SOM were higher (1.2-2.0 times) than those from the straw (Fig. 1d).

Enzyme activities
Straw additions increased enzyme activities compared with S0+N0, with similar levels between S+N0, S+N1, and S+N2. Straw alone and with N additions compared to trials without straw and N increased β-Dglucosidase (BG), Cellobiohyrolase (CBH), and β-N-acetylglucosaminidase (NAG) activities by 3.9, 9.6, and 4.9 times, respectively. Enzyme activities of BG, CBH, and NAG increased during the rst 5 days, but decreased after 100 days of incubation compared to day 5 (Additional File: Figure S3). Taken together, straw rather than the N additions was found to increase enzyme activities.

Soil microbial communities
Soil microbial communities based on phospholipid fatty acid composition Straw additions increased the relative abundance of the phospholipid fatty acid (PLFA) composition, N levels had not effect on the PLFAs. During the whole incubation period, the total and bacterial PLFA abundances were higher in the NPK+Straw soil than the unfertilized control and NPK soils. The total, bacterial, and fungal PLFA abundances were highest in the rst 1-5 days, and then decreased until the end of incubation. Straw alone and with N additions compared with S0+N0 increased the total, bacterial, and fungal PLFA abundances levels by 54%, 82%, and 107%, respectively. Similar levels for the total, bacterial, and fungal PLFA abundances between the straw alone and straw with N addition (Additional File: Figure S4). Overall, straw addition increased PLFA abundances, particularly in the rst 1-5 days.
Soil bacterial and fungal abundance, diversity, and community composition Soil bacterial and fungal abundance (based on the real time qPCR) as well as fungal diversity shifted with the straw and N additions (Additional File: Table S1). Both the straw alone and with the N additions increased the 16S and 18S rRNA gene copies by 1.4 and 4.9 times, respectively. Similar contents with 16S and 18S rRNA gene copies between straw alone and straw with N levels. Straw with increasing N levels decreased fungal richness by 4.8-49.8% and the Shanon index was 1.2-23.3%. The opposite temporal tendencies were found in 16S and 18S rRNA gene copies. 16S rRNA gene copy levels were higher in the early incubation period (1-5 days) than after100 days in the fertilized soils. Soil bacterial diversity and 18S rRNA gene copies increased with incubation duration (Fig. 2) Figure S5). The development of bacterial communities varied during the 100 days ( Fig.  3a; Additional File: Table S2). The relative abundance of the bacterial Firmicutes was increased on day 1 and then decreased with the incubation duration. Actinobacteria abundance was higher during the rst 1-5 days than on day 1, but the relative abundance of Proteobacteria was increased with the incubation duration (Additional File: Figure S6). Straw with N levels increased the relative abundance of Mucoromycota on day 5 but decreased the relative abundance of Blastocladiomycota and Zoopagomycota (Additional File: Figure S7). Thus, bacterial community compositions had dynamic development with the incubation duration and the N fertilization levels affected the fungal community compositions ( Fig. 3; Additional File: Table S2, S3).
Soil bacterial and fungal co-occurrence networks associated with environment factors The integrated co-occurring networks were built for three ecological clusters that closely co-occurred with each other (modules 1, 2, and 3) to detect the dominant bacterial and fungal phylotypes (Fig. 4). The relative abundance of modules 1 was strongly increased by 54% compared to those for modules 2 and 3. Bacteria phylum of Actinobacteria, Firmicutes, and Proteobacteria were highly clustered in module 1 (Additional File: Figure S8, Table S4). Ascomycota, Basidiomycota, and Mucoromycota were the most dominant phylum of fungi in modules 1, 2, and 3 (Additional File: Figure S8, Table S4). These multiple  Table S4). Thus, Actinobacteria, Firmicutes, Proteobacteria Ascomycota, Basidiomycota, and Mucoromycota were the main drivers for straw and SOM decomposition.
Fungal taxa compared with bacterial groups had strong correlations with environmental factors (including soil properties, e.g., DOC, MBC, and enzyme activities; straw and SOM decomposition) (Fig. 4,   5). Edge numbers and ratio for pair-wised correlations of the bacterial community compositions were highest among abundant taxa themselves as well as being interrelated with rare groups. For the fungal communities, abundant taxa were more widely associated with environmental factors and interrelated with themselves (Table 1). Fungal taxa more than bacterial groups were interrelated with SOM mineralization, straw decomposition, PE, and soil properties, particularly for abundant taxa (ratio for pairwised correlations: 21.9 vs. 1.2%) ( Table 1; Additional File: Figure S9). Overall, abundant fungal taxa outcompeted bacteria for straw and SOM mineralization. Environmental factors include soil properties and soils organic carbon (SOC) mineralization, straw decomposition, and priming effect (PE), for the no fertilization (Control), mineral fertilizers only (NPK), and mineral fertilizers plus maize and soybean straws (NPK+Straw) soils.
Four sub-networks in each unfertilized control and fertilization soils were used to explored the linkages between straw alone or with N level additions and bacterial as well as fungal taxa (Additional File: Table   S5, S6). Compared to trials without straw and N, straw alone and with N levels increased bacterial ratio of negative edges (correlation with each other among taxa) for 7.3% (Additional File: Table S5). The opposite was observed for the fungal taxa, as straw alone or with the added N levels decreased fungal ratio of negative edges by 12.6% (Additional File: Table S6). Similar results were found in the bacterial and fungal negative edges between straw alone and with addition N levels (Additional File: Table S5, S6).
The negative correlations between the bacterial taxa and organic C mineralization were higher after 100 days than after 1-5 days incubation. In contrast, negative correlations between the fungal taxa and organic C mineralization decreased over 100 days compared with the rst 1-5 days, except for those without straw and N fertilizer (Fig. 5). Taken together, straw addition decrease and increase competition among fungal and bacterial taxa with low and high negative edges, respectively.
Key factors for straw decomposition, soil organic matter mineralization, and priming effects Incubation time, soil fertilization, and straw with N additions, affected SOC mineralization, straw decomposition, and PEs (Additional File: Table S1). SOC mineralization, straw decomposition, and PEs  Table S8).

Discussion
Organic matter mineralization is linked to fungi taxa more than bacteria taxa The co-occurring network provided empirical evidence that Actinobacteria, Proteobacteria, and Firmicutes bacteria phyla were the keystone taxa in organic matter decomposition. Actinobacteria, Proteobacteria, and Firmicutes had highest interactions with other taxa as well as straw and SOM decomposition in module 1 (Fig. 5; Additional File: Table S4). This con rms that these bacterial groups are widely considered to assimilate maize C [53][54][55]. The relative abundance of Firmicutes was increased on day 1 and decreased from 5 to 100 days (Additional File: Figure S6). Additionally, the correlation coe cient for Firmicutes was linked to straw and SOM decomposition was higher than with other fungal groups on day 1 (Fig. 5). Consequently, copiotrophic lineages (e.g., Firmicutes) mainly mediated the decomposition of labile organic compounds in the initial stage [53,54]. As the labile compounds were gradually decomposed, the copiotrophic lineages of bacteria were substituted by oligotrophic lineages [19,56], increasing the relative abundance of Proteobacteria on day 100 (Additional File: Figure S6). Therefore, the bacteria decreased capacity to degrade recalcitrant compounds during the later incubation stages followed by increased bacterial diversity but decreasing 16S rRNA gene copies (Fig. 2).
Fungi, more than bacteria, were the key drivers regulating straw and SOM decomposition. Though nodes (represented as amplicon sequence variants) for fungi were lower than for bacteria. Soil properties, such as MBC and DOC, as well as organic matter decomposition with fungal taxa had remarkably higher connection numbers than those with bacterial groups within the integrated network (Fig. 4, 5). This was because: 1) The N content of the limestone (210-280 mg kg -1 ) in karst area is high [37] due to the abundance of legumes xing atmospheric N and N rich bedrock [57]. Microbial C limitation were aggravated more than N in the karst regions [38]. Fungi have high C utilization and a competitive advantage for low N nutrient acquisition over the bacteria by increasing the abundance of some abundant taxa (e.g., Ascomycota) [16,17,58]. Straw additions increased fungal abundance more than bacteria, as 16S and 18S rRNA gene copies increased by 1.4 and 4.9 times, respectively (Fig. 2). This con rm that fungi, when compared with bacteria, were more inspired by straw addition; 2) Soil Ca supressed microbial cell communications and cell motility [59]. Fungi compared with bacteria could tolerate more extreme soil environments (e.g., high Ca) [14,51]. Consequently, fungi had direct and closer correlations with organic matter decomposition compared with bacteria.
Abundant fungal groups dominated SOM decomposition. Particularly, the ratio for pairwise correlations between abundant fungal taxa and soil properties as well as organic matter decomposition increased by 20% more than those within rare groups (Table 1; Additional File: Figure S9). The Ascomycota prefer to grow in rich nutrients during organic matter and N additions [60][61][62]. The Basidiomycota phylum is generally most effective for lignin decomposition [63]. Straw with the increased N levels increased Mucoromycota abundance but declined fungal diversity by decreasing Blastocladiomycota and Zoopagomycota, which showed that raising N availability suppressed most fungal taxa growth (Additional File: Figure S7). Consequently, abundant fungal taxa of Ascomycota, Basidiomycota, and Mucoromycota had the highest degree numbers in the integrated networks (Additional File: Table S6). This was clearly as abundant fungal groups (e.g., Ascomycota, Basidiomycota, and Mucoromycota) had more important functions in straw and SOM decomposition than rare taxa, especially for soil organic fertilized with N.
The positive and negative correlations within the microbial groups re ect the cooperation and competition relationships, respectively [64,65]. The ratio of negative correlations among the fungal groups (45.7%) were greater than those within bacterial taxa (5.6%), showing that bacteria was dominated by cooperation relationships while fungi had erce competition for C and N (Additional File: Table S5, S6). Bacteria prefer nutrient rich environments [17,19], and consequently less labile organic compounds in the later stages caused strong competition among the bacterial taxa (Fig. 5) [53,66].
Consequently, bacterial taxa had higher negative correlations with the straw and SOM mineralization after 100 days than in the initial 1-5 days (Fig. 5). Conversely, straw alone or with N additions decreased the negative links between the fungal taxa and organic matter decomposition more after 100 days than during the rst 1-5 days (Fig. 5). Straw provides more cellulose and lignin for fungi in the later stages [7,13], increased 18S rRNA gene copies thereby decreased competition with low negative correlations among fungal taxa (Fig. 2, 5; Additional File: Table S6). Consequently, straw additions decreased fungal but increased bacterial competition relationships among taxa.

Long-term fertilization increases straw and soil organic matter mineralization
Long-term applications of mineral and organic fertilizers increased C and N availability for microbial growth [54,[67][68][69]. The SOC, available N, enzyme activity (CBH), and bacterial and fungal abundance were higher in the NPK and NPK+Straw soils than in unfertilized control soil before the straw and N additions ( Fig. 2; Additional File: Table S9, Figure S3). Consequently, higher CO 2 , 13 CO 2 ux, and SOM mineralization were occurred in the NPK and NPK+Straw than in unfertilized control soil ( Fig.1; Additional File: Figure S2). Conversely, increases in C and N limitations had consequence for microbial growth limited by these nutrients. As labile C is rapidly exhausted, increased microbes immediately and e ciently decomposed more SOM for C and N acquisition, leading to a stronger PE in unfertilized control soils than in NPK and NPK+Straw soils [4,66]. Therefore, straw return is necessary to decrease PE as well as increase SOM accumulation. Consequently, long-term fertilization increased straw and SOM decomposition but decreased PE.
Response of organic matter mineralization and the microbial characteristics to straw and nitrogen additions There are large CO 2 uxes in the initial stages followed by increasing bacterial and fungal abundance as well as enzyme activity. Microbes prefer to utilize soluble substrates from straw instead of SOM, and thus increased their CO 2 and 13 CO 2 emissions during the rst 0-30 days, but they peaked after the 3 rd day [66,70,71]. Microbial activity and biomass increased as the straw released su cient soluble organic C for microbial growth [66,72,73]. The enzyme activity (BG and NAG), bacterial and fungal PLFA abundances, MBC, and DOC rapidly increased after straw additions during the initial 1-5 days and then decreased, con rmed for the temporal dynamics of CO 2 emissions (Additional File: Figure S1, S2, S3, and S4) [72,74,75]. Consequently, straw and N additions with greater C and N content as well as larger microbial populations and higher activity caused higher CO 2 uxes in the karst calcareous soils during the initial 1-3 days [66,70].
An elemental stoichiometric balance of the microbial requirements for C:N regulated straw and SOM decomposition [66,70]. Increasing N availability causes relative de ciency of C [66]. The highest PE in the straw was with the high N (S+N2) (Fig. 1) and showed that decomposer activities increased by increasing SOM decomposition to meet the demands of the stoichiometric microbial nutrients [4,66]. In contrast, straw coupled with appropriate N additions (S+N1) would decrease the PE for the unfertilized control and NPKS soils [7]. Consequently, it was found that when N fertilization is split into small doses it can suitably decrease PE [70,76].
Straw alone or with additional N, increased the bacterial and fungal abundance by increasing substrate availability (MBC, DOC, DON, NH 4 + , and NO 3 − ) ( Fig. 2; Additional File: Figure S1) [14,54,77]. Karst soils have developed from limestone or dolomite and consequently have large Ca contents [30,32]. Increased soil Ca promoted soil organic carbon, resulting in the low availability of organic C (e.g., DOC) in the karst soils is low when compared to non-karst soils [78,79]. The straw additions increased the labile C compounds (MBC and DOC), causing raised enzymatic activities (BG CBH and NAG), bacterial and fungal PLFA abundances, as well as 16S and 18S rRNA gene copies ( Fig. 2; Additional File: Figure S1, S3, and S4). As described above, microorganisms in karst soils were more susceptible to C than N restrictions [38]. Consequently, there was little difference in enzyme activities, and bacteria and fungi populations (both PLFAs and gene copies) between the S+N0, S+N1, and S+N2. Enzyme activity and microbial abundance were sensitive to straw rather than N additions.

Microbial characteristics and soil properties regulate organic matter decomposition
Microbe changes accompany soil nutrient changes to drive SOM decomposition [4,6,14,77]. Close correlations between the microbial characteristics (bacterial and fungal PLFAs abundance as well as Shannon index) with straw and SOM decomposition, as well as PE show that microorganisms were the key drivers in organic matter mineralization (Additional File: Table S7). Straw alone and with N additions indirectly increased SOM mineralization by increasing the bacterial and fungal abundance as well as enzyme activity towards raised the labile substrates (e.g., DOC, 13 C-DOC, MBC, DON, NH 4 + , and NO 3 − ) [74,80]. Bacterial and fungal community compositions (especially with fungal abundant groups) are directly altered by increases in the DOC, DON, NH 4 + , NO 3 − , and enzyme activity (Fig. 6), and result in increased SOM decomposition [4,66,70].

Conclusions
Long-term mineral fertilizers increase straw decomposition and CO 2 emissions but decreased PE compared with the unfertilized control soil. Straw alone and with additional N induced SOM decomposition because of the increasing enzyme activities (BG, CBH, and NAG), as well as bacterial and fungal abundance. Straw with low N additions increase SOC accumulations by decreasing PE in the control as well as mineral fertilizers with straw soils. This provided empirical evidence that split N fertilizer additions mitigate straw decomposition with low PE. The 16S rRNA gene copies were decreased when easily available compounds were depleted with incubation time, thereby increasing the competition among bacterial taxa by raising diversity. In contrast, in the late decomposition stage, 18S rRNA gene copies were increased but competition among fungal taxa decreased due to their capacity for recalcitrant compounds. Bacterial phyla Firmicutes (1 day), Actinobacteria (5-100 days) and Proteobacteria (5-100 days) regulated SOM decomposition. Compared with bacteria, the fungi taxa had stronger connections with straw and SOM decomposition because of their capacity to utilize C and N in high Ca and pH environments. Additionally, the ratio for pairwise correlations between abundant fungal taxa (e.g., Ascomycota, Basidiomycota, and Mucoromycota) and organic matter decomposition was higher than those with bacteria did. Consequently, fungal abundant groups rather than rare taxa outcompeted bacteria for organic matter decomposition in karst soils. In conclusion, straw and N additions indirectly increased SOM decomposition, followed by directly increasing the available nutrients (e.g., DOC and N), Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.        Networks edges linked to bacterial or fungal taxa and soil carbon and straw decomposition. Networks containing the top 100 edges that directly link the bacterial (top) or fungal (bottom) taxa that directly connected to straw decomposition, soil organic carbon mineralization, and priming effects, after 1, 5, and 100 days depending on straw and nitrogen additions. S0+N0, neither straw nor nitrogen fertilizer addition; S+N0, addition of 13C-maize straw; S+N1, addition of 13C-maize straw and low nitrogen; S+N2, addition of 13C-maize straw and high nitrogen. Nodes ( lled with dark red) in the center refer to priming effects, straw decomposition, and soil organic carbon mineralization from left to right. The size of surrounding nodes which refer to ASVs denote the absolutely value of SparCC correlation (r). The red and bule lines between each pair of nodes indicate positive and negative correlations, respectively. The numbers right below each sub gure represent ratio of positive and negative correlations.

Figure 5
Page 35/36 Networks edges linked to bacterial or fungal taxa and soil carbon and straw decomposition. Networks containing the top 100 edges that directly link the bacterial (top) or fungal (bottom) taxa that directly connected to straw decomposition, soil organic carbon mineralization, and priming effects, after 1, 5, and 100 days depending on straw and nitrogen additions. S0+N0, neither straw nor nitrogen fertilizer addition; S+N0, addition of 13C-maize straw; S+N1, addition of 13C-maize straw and low nitrogen; S+N2, addition of 13C-maize straw and high nitrogen. Nodes ( lled with dark red) in the center refer to priming effects, straw decomposition, and soil organic carbon mineralization from left to right. The size of surrounding nodes which refer to ASVs denote the absolutely value of SparCC correlation (r). The red and bule lines between each pair of nodes indicate positive and negative correlations, respectively. The numbers right below each sub gure represent ratio of positive and negative correlations.