JIA-2019-11

2611 CHEN Xu et al. Journal of Integrative Agriculture 2019, 18(11): 2605–2618 and BF (average of 18.8 and 33.4%), but the abundances of the individual PLFAs with the various additions also followed a pattern similar to that of CK. The most abundance of fungal, bacterial and actinomycetic PLFAs was from the glucose addition, which produced 21.8–36.0% more for BF and 12.2–28.8% for FL compared to the control. The addition of alanine increased fungal, bacterial and actinomycetic PLFAs 10.4–34.8% for BF and 11.8–21.4% for FL compared to the control. The addition of substrates had little effect on the fungi/bacteria and G + /G – ratios between IAD and RAD throughout the incubation, except the ratios were higher in BF-RAD (Appendices D and E). 3.3. PLFA compositions in grassland Early in the incubation, the fungal PLFA 18:1 ω9c was most abundant in GL-RAD soil with or without substrate addition (Appendix F). The other fungal PLFAs did not differ between IAD and RAD. RAD fungal PLFAs later in the incubation did not differ significantly, except for 18:1ω9c in CK. Similarly, the PLFA 10Me18:1ω7c (actinomycetic PLFA marker) was the main factor affecting the amount of actinomycetes after aggregate-size reduction. The PLFA10Me16:0 also became a determinant later in the incubations without substrate additions. The abundances of two bacterial PLFAs (i16:0 and a17:0) were significantly high for RAD at day 3 without additions (Appendix G), and the abundances of six PLFAs (16:0, i15:0, 18:1ω7c, a15:0, cy17:0 ω7c and cy19:0 ω7c) were high after 60 days of incubation. The inorganic-N additon increased the bacterial PLFA numbers from one (18:1ω5c) at early to two (i15:0 and 18:1ω7c) at the end of the incubation. The abundances of some bacterial PLFA (16:0, 16:1ω7c, 18:1ω7c, cy17:0ω7c and 18:1ω5c) were significantly higher in the RAD treatment under the glucose and alanine additions ( P <0.05), but decreased toward the end of the incubation (only 16:0 under glucose addition). The bacterial PLFA 18:1 ω5c was not detected in GL at the end of the incubation. 3.4. Enzymatic activities GL soil had higher enzymatic activities than FL and BF soils during the incubation (Fig. 3). With alanine additon, BG and NAG activities in BF were 13.0 and 89.2% higher for BF, 2.5 and 53.0% higher for FL compared to the control, respectively. All the treatments and substrate addition combinations peaked at nearly the same time (day 3 or 15), and then decreased. The activities of BG and NAG from GL was more than 1.5-fold larger than BF and FL at the end of incubation regardless of treatment. The RAD treatment increased the enzyme activities, which were higher in GL. BG and NAG activities in GLwithout substrate addition were 17.4 and 7.6% higher in RAD than that in IAD, respectively. The differences between RAD and IAD disappeared over time and often became insignificant at day 60 in GL, except for CK and inorganic N. Almost no difference was observed in RAD and IAD for BF and FL in any treatment. 3.5. Relationships between the microbial parameters and soil properties Soil management, substrate addition and aggregate distribution significantly influenced the microbial parameters, except for the fungi/bacteria and G + /G – ratios by aggregate- size distribution ( P <0.05, Table 2). The PCA indicated that the soil samples were grouped by SOC concentration (Fig. 4). The microbial parameters were highly correlated with GL, especially for RAD. PC1 and PC2 accounted for 97.6 and 1.5% of the total variation, respectively. These results confirmed that SOC content, PLFA abundance and microbial enzymatic activities were more sensitive to RAD in GL than in FL and BF. The GL-RAD and IAD samples were separated from the FL and BF samples along PC1. Soil management had the largest effect on PLFAs and enzymatic activities (adonis, R 2 =0.829, P <0.001). Importantly, the effect of aggregate size reduction was significant for GL (adonis, R 2 =0.803, P <0.001) but not for FL (adonis, R 2 =0.068, P =0.176) or BF (adonis, R 2 =0.020, P =0.584), whereas the substrate addition exerted a similar effect on FL (glucose:adonis, R 2 =0.862, P <0.001; alanine:adonis, R 2 =0.869, P <0.01) and BF (glucose:adonis, R 2 =0.864, P <0.01; alanine:adonis, R 2 =0.914, P <0.01), but had no significant impact in GL. The inorganic N addition showed the only significant effect on BF (adonis, R 2 =0.267, P <0.05). The PCA was performed for GL only to determine the effect of aggregate disruption under higher C and macroaggregate levels (Appendix H). The PC1 (76.4%) and PC2 (19.2%) ordinations also identified two clusters of macroaggregate reduction along PC1, regardless of substrate input. The microbial parameters were significantly correlated with the RAD treatment, further illustrating that SOC content, PLFA abundance and microbial enzymatic activities were sensitive to macroaggregate reduction in GL. 4. Discussion 4.1. Effect of aggregate-size distribution on the microbial communities A reduction in soil aggregate size can release the decomposable C that was physically protected by