Microbial analyses
At the end of the incubation, phospholipid fatty acids in the soils were measured following standard methods47-48 to characterize microbial community composition. Phospholipids were extracted from 3-g freeze-dried soil using 15 mL of a single-phase chloroform: methanol: citrate buffer (1: 2: 0.8) (0.15 M, pH 4.0), and measured using a gas chromatograph (Agilent 6890 Series, Agilent Technologies, Wilmington, DE) with a flame ionization detector. Nonadecanoic acid, 19:0, was used as an internal standard. Peaks were identified using the MIDI peak identification software (MIDI, Inc., Newark, DE) and a bacterial fatty acid standard49-50. A total of 34 phospholipid fatty acids were detected in these samples and assigned to different functional groups according to the nomenclature in the literature51, including G (+) bacteria (i14:0, i15:0, i16:0, i17:0, a13:0, a15:0, a16:0 and a17:0); G(-) bacteria (15:4ɷ3c, 14:0 3OH, 11Me 18:1ɷ7c, cy19:0 ɷ8c, 17:1ɷ8c, 18:1ɷ7c, 16:1ɷ7c, N 16:0, 15:3ɷ3c, 18:1ɷ9c, and 16:1ɷ6c and 15:1G); fungi (16:1ɷ5c, 18:2ɷ6,9c and 18:3ɷ6c); actinomycetes (10Me16:0, 10Me 17:0, and 10Me20:0) and general bacteria (14:0, 15:0, 16:0, 17:0 and 18:0).
After the incubation, soil amino sugars were also measured to determine the effects of clay mineralogy on fungal and bacterial residues52. A sufficient mass of each soil sample was ground to ensure extraction of approximately 0.3 mg N. Then the samples were placed in a closed hydrolysis flask (filled with N2gas) and hydrolyzed in 10 mL of 6 M HCl at 105 oC for 8 hours. The solutions were filtered through glass-fiber filters GF6 (Schleicher & Schuell, Germany), and dried at 40 oC using a rotary evaporator in vacuum. Afterwards, the samples were re-dissolved in deionized water, their pH adjusted to 6.6-6.8 using 0.4 M KOH and 0.01 M HCl and centrifuged (1000 g ) for 10 mins. Amino sugars were first recovered from post-incubation soils by freeze-drying and methanol washing and then extracted with dichloromethane from the aqueous solution as aldononitrile derivatives. Excess anhydride was removed with 1 M HCl and deionized water. The amino sugar derivatives were re-dissolved in a 300-μL mixed solvent of hexane and ethyl acetate in 1: 1 volume ratio for final analysis after the removal of dichloromethane by drying under nitrogen gas. The concentrations of amino sugars were quantified with the internal standard myo-inositol added before hydrolysis. N-methylglucamine was also added prior to derivatization to calculate the recovery of amino sugars and the amounts of amino sugars identified as glucosamine, galactosamine and muramic acid. The concentration of fungal-derived residues was calculated as the difference between total glucosamine and twice that of the muramic acid concentration, while the concentration of bacterial-derived residues was defined as the sum of bacterial glucosamine plus muramic acid and galactosamine53-54.
13C solid-state NMR spectroscopy
13C CP/TOSS experiments were conducted at a spinning speed of 5 kHz and a cross polarization CP time of 1 ms, with1H 90° pulse-length of 4 µs and a recycle delay of 0.8 s using a Bruker AVANCE400 spectrometer at 100 MHz for13C with 4-mm sample rotors. Four-pulse total sideband suppression was employed before detection, and two-pulse phase-modulated decoupling was applied for optimum resolution55. The spectra obtained with cross polarization/total sideband suppression were assigned to different C functional groups following previous literature3,56-57. The relative proportions of the functional groups in the total spectral area were obtained by integration58 and presented for all the SOM samples in comparison with those of the original litters in Supplementary Table 2. The average noise of three chemical shift regions sampled from beyond 0-220 ppm in each spectrum was used as an error to determine the differences in the functional group distributions between samples59.