Study site and experimental platform
Our study was carried out in a temperate cropland characterized by a subcontinental climate (mean temperature 8.9 °C, and mean annual rainfall of 498 mm for the period 1896–2013; mean temperature 9.8 °C, and mean annual rainfall of 516 mm for the period 1995–2014) in the Global Change Experimental Facility (GCEF), a field research station of the Helmholtz Centre for Environmental Research in Bad Lauchstädt, Saxony-Anhalt, Germany (51°22′60 N, 11°50′60 E, 118 m a.s.l.). During the study period (2018), the mean temperature was 10.8 °C with an annual rainfall of 254 mm. The GCEF (Fig. S2a) was designed to comparatively investigate the consequences of future climate and ambient climate conditions on ecosystem processes in a 50 field plots (400 m2 each), with half of them subjected to ambient and future climatic conditions, respectively . The future climate regime is a consensus scenario across three models (COSMO-CLM , REMO  and RCAO ) of climate change in Central Germany for the years 2070–2100 that manipulate both precipitation and temperature. For this, future climate plots (Fig. S2b) are equipped with mobile shelters and side panels, as well as an irrigation system, the roofs are controlled by a rain sensor. As result of continuous adjustment of irrigation or roof closing, precipitation is reduced by ~20% in summer months and increased by ~10% in spring and autumn. To simulate the increase in temperature with asymmetry between daytime and nighttime warming, we used the standard method passive nighttime warming to maintain the higher daytime temperature for increasing night temperature . The shelters and panels automatically close from sundown to sunrise to increase the mean daily temperature by ~0.55 °C. The resulting changes in climate conditions before and during the study period were shown in our preliminary work . Ambient climate plots are equipped with the same steel constructions (but without shelters, panels, and irrigation system) to mimic possible microclimatic effects of the experimental setup . The wheat straw decomposition experiment was performed on the conventional farming plots under both ambient (five replicate plots) and future climate (five replicate plots) conditions. The conventional farming plots are characterized by a typical regional crop rotation (including winter rape, winter wheat, and winter barley) with tillage and application of mineral fertilizers and pesticides. Management details are given elsewhere .
Litterbag preparation, experimental design, and sampling
After harvest of winter wheat (Triticum aestivum L.), the straw left over (10 cm aboveground) was sampled from each GCEF field plot and placed in sterile plastic bags before transferred to the laboratory on ice. The wheat straw from each plot was oven-dried at 25 °C for 3 days to normalize the moisture content, and then 10 g was enclosed in a litterbag (20 × 15 cm, 5 mm mesh size) . Three litterbags were returned back to each field plot (five ambient conventional farming plots and five future conventional farming plots) in mid-August 2018. Our experiment was therefore established at the end of the drought period under future conditions in summer. We evaluated the effects of future climate four years after the start of climate manipulation (the manipulation of temperature and precipitation started in April and July 2014, respectively). To simulate the natural field conditions, the litterbags were placed on the soil surface at the beginning of the experiment and following the agricultural practices (tillage) they were buried at 5 cm depth after 20 days. First sampling occurred at the onset of the experiment (0 days). The litterbags were buried at the same original location in the plots after the process of tillage in September, second sampling was done in first 30 days and the third sampling at 60 days after field incorporation. For sampling, one litterbag per plot was placed in sterile plastic bag before transferred to the laboratory on ice.
Microbial DNA Extraction, PCR, and Illumina Miseq Sequencing
Wheat straw from each bag was homogenized with the aid of liquid N2 and were used for further analyses. DNA was extracted from 150 mg of homogenized wheat straw sample using a DNeasy PowerSoil kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions, then subjected to polymerase chain reaction (PCR). The V5–V7 region of the bacterial 16S rRNA gene was amplified using the following primers: BAC799F forward (5′-AACMGGATTAGATACCCKG-3′)  and BAC1193R reverse (5′-ACGTCATCCCCACCTTCC-3′) . These bacterial primer pairs were chosen because they do not amplify the chloroplast DNA in pyrosequencing . The ITS2 region of fungi was amplified using the following primers: fITS7 (5′-GTGARTCATCGAATCTTTG-3′)  and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) . The amplification was performed in a two-step process. First amplifications were performed in 25 µL reactions with a Qiagen HotStar Hi Fidelity Polymerase Kit (Qiagen Inc., Valencia, CA, USA), 1 µL of each 5 µM primer, and 1 µL of template. Reactions were performed on ABI Veriti thermocyclers (Applied Biosytems, Carlsbad, CA, USA). The first PCR conditions were 95 °C for 5 min, then 35 cycles of 94 °C for 15 s, 54 °C for 60 s, 72 °C for 1 min, followed by one cycle of 72 °C for 10 min and 4 °C hold. Amplicons from the first stage amplification were diluted 1:10 and then used as a template in the second PCR. During the second PCR, dual indexes were attached using the Nextera XT Index Kit with the same amplification conditions as the first stage, except for 10 cycles. Amplification products were visualized with eGels (Life Technologies, Grand Island, New York) as explained by the manufacturer. Products were then pooled equimolar and each pool was size selected in two rounds using Agencourt AMPure XP (BeckmanCoulter, Indianapolis, Indiana) in a 0.75 ratio for both rounds. Size selected pools were then quantified using the Quibit 2.0 fluorometer (Life Technologies). Sequencing was performed using MiSeq (Illumina, Inc. San Diego, California) 2 × 300 bp paired-end strategy according to manufacturer`s manual.
The primer sequences were trimmed from the demultiplexed raw reads using cutadapt . The pair-end raw reads of bacterial and fungal datasets were merged using the simple Bayesian algorithm with a threshold of 0.6 and a minimum overlap of 20 nucleotides as implemented in PANDAseq . Reads fulfilling the following criteria were remained for further analyses: a minimum length of 350 (bacteria) and 120 (fungi) nt; a minimum average quality of 29 (bacteria) or 25 (fungi) Phred score; containing homopolymers with a maximum length of 20 nt; without ambiguous nucleotides. We detected chimeric sequences using the UCHIME algorithm  as implemented in MOTHUR and removed from the datasets. The obtained reads were then clustered into operational taxonomic units (OTUs) using the CD-HIT-EST algorithm  at a threshold of 97% sequence similarity. The OTU representative sequences (defined as the most abundant sequence in each OTU) were taxonomically assigned against the reference sequences from the SILVA database v132  for prokaryote 16S rRNA gene and the unite database (version unite.v7)  using the naive Bayesian classifier as implemented in MOTHUR  using the default parameters. Rare OTUs (singletons and doubletons) which potentially might originate from artificial sequences  were removed. The read counts were normalized to the smallest read number per sample. Therefore, the final normalized dataset without rare OTUs was used for further statistical analysis, unless otherwise stated. The ecological and metabolic functions of bacterial OTUs were predicted using FAPROTAX  and the functional annotation tool of prokaryotic taxa v.1.1, whereas those of fungal OTUs were predicted using FUNGuild .
Mass loss and physicochemical analysis of wheat straw
The dry mass of wheat straw samples from the GCEF plots was determined after oven drying at 105 °C to constant weight (mostly after 24 h), and was used for determination of mass loss at three time points (0, 30, and 60 days) under ambient and future climate regimes. Physicochemical properties were determined for total of 30 samples (10 samples for each time point where half of these are incubated under ambient and future climate conditions, respectively). Total C and N concentrations were determined by dry combustion at 1000 °C with a CHNS-Elemental Analyzer (Elementar Analysensysteme GmbH, Hanau, Germany) according to manufacturer’s protocol. Available soil phosphorus was extracted and measured according to Bray 1 method . Concentration of cations (K+, Mg2+, Ca2+, and Na+) were determined by atomic absorption spectrophotometry, using a Z 5300 instrument (Hitachi—Science & Technology, Tokyo, Japan) following recommendations of the manufacturer. The pH of the wheat straw samples was measured using a WTW Multi 3510 IDS portable meter (Weilheim, Germany).
Assay of microbial enzyme activity in wheat straw
Activities of five microbial extracellular enzymes were measured in the same 30 samples of homogenized wheat straw. Of those, three are hydrolytic enzymes important for the acquisition of polymeric carbon (b-glucosidase), nitrogen (N-acetylglucosaminidase) and phosphorus (phosphatase); and two are oxidative enzymes related to the chemical modification of lignin (phenol oxidase and peroxidase) . Hydrolytic and oxidative enzymes were measured based on 4-methylumbelliferone (MUB) derivatives and 3, 3′, 5, 5′-tetramethylbenzidine (TMB), respectively, as described previously .
Statistical analyses were performed using the PAST program v.2.17c  and IBM SPSS Statistics (Version 24) software. All the analyses were conducted based on five independent replicate plots of the field experiment (n = 5) with time as within plot factor. Bacterial and fungal OTU richness were calculated for each sample using the ‘diversity’ function in the PAST program. Samples rarefaction curves of fungi and bacteria are shown in Fig. S3. As the rarefaction curves indicated sufficient OTU coverage, we used the observed OTU richness directly as a proxy for bacterial and fungal diversity. Permutational multivariate analysis of variance (NPMANOVA)  based on Jaccard distance (permutations = 999) was performed to test the impact of sampling times and climate conditions on bacterial and fungal (including plant pathogen and saprotrophs) communities over time. Non-metric multidimensional scaling (NMDS) was used to visualize the variations of bacterial and fungal community compositions among the three sampling time points (0, 30, and 60 days), under ambient and future climate conditions, respectively. We used the presence/absence data of bacterial and fungal communities and Jaccard dissimilarity distances (permutations = 999) to perform the NMDS ordination plot as they are more reliable than relative abundance data. All physicochemical properties that significantly affected bacterial and fungal community compositions (p < 0.05) were fitted in the respective NMDS ordination plots using PAST. T-test was applied to evaluate the effect of climate on mass loss of wheat straw at 30 and 60 days under ambient and future climate conditions, respectively. Effect of climate, time and their interaction on physicochemical properties of wheat straw and on microbial enzymes activity were assessed by time-series analysis using SPSS as the data sets vary over time. With this test, climate was used as a between‐subject factor and time was used as within plot factor. We tested the correlation between bacteria and fungi (community composition and richness) and wheat straw physicochemical properties. Additionally, the correlation between microbial communities and richness and enzyme activities was investigated. For correlation analyses, Jarque-Bera test was performed to evaluate normal distribution of the datasets . Pearson’s correlation and Spearman’s rank correlation were applied with normally distributed and not-normally distributed datasets, respectively. Due to the significance of fungal plant pathogens in field-incorporated wheat straw, we aimed to characterize the interactions between these pathogens and other microorganisms colonizing wheat residues. We performed ecological network analysis (ENA) using Spearman’s rank correlations (P < 0.05) for ambient and future climate conditions, separately. The ecological networks of potential interacting taxa were visualized using cytoscape 3.0.2. . Network properties were calculated using Network Analyser as implemented in cytoscape. These correlation networks included nodes that consisted of plant pathogenic fungi and fungal and bacterial OTUs as a proxy for ‘species’, while the edges represented the correlative relations among the OTUs . Microbial hubs are defined as strongly interconnected taxa, which can have a severe effect on microbial community compositions and networks if they were removed .