The experiments were carried out in Quanxi village, Shiyan city, Hubei province, China (B: 31˚58.734'N, 109˚40.213'E; 1103 m). This region has a typical subtropical monsoon climate with an average annual precipitation of 1000 mm and an average temperature of 14˚C. The soils are loam. The farmland was planted with corn (Zea mays L.) and the grove was planted with sumac (Toxicodendron vernicifluum) until it was three years old. The field contained four plots (5.0 m × 10.0 m) each in the farmland and grove, and each plot was repeated 3 times. Starting in 2013, konjac (Amorphophallus konjac K. Koch ex N.E. Br.) was cultivated each year in the field, and the study was completed in 2015.

Each growth cycle of the present study was initiated on or about April 5 with a few days of variation among years (Figure 1). Konjac corms (about 200 g) were buried on the top of raised paths (25 cm in height, 120 cm in width) and plants were spaced 30 cm apart along the row. Two rows were planted on each raised path with 40 cm between the two rows. The blended fertilizer with a ratio of 15:15:15 (NH₄)₂SO₄:P₂O₅:K₂O was used for 750 kg·ha⁻¹ in basal-dressing and 75 kg·ha⁻¹ in top-dressing. All other field management activities were performed manually. Konjac corms were harvested in late October. The operation in the second year was repeated in the first year, and so on.

Rhizosphere samples were numbered as follows: F0 (cultivated corn for 3 years in farmland); F1 (cultivated corn for 2 years and konjac for 1 year in farmland); F2 (cultivated corn for 1 years and konjac for 2 years in farmland); F3 (cultivated konjac for 3 years in farmland); G0 (cultivated sumac grown for 6 years) ; G1 (cultivated konjac grown for 1 year under 4-year-old sumac); G2 (cultivated konjac for 2 years under 5-year-old sumac); G3 (cultivated konjac for 3 years under 6-year-old sumac).

Soil samples were collected on July 20, 2015. In each plot, soil was collected at four diagonal points using a sterile plastic bag and combined into a single sample. Every point is randomly selected 8 healthy konjacs in addition to the blank control (F0 and G0), rhizosphere samples were collected by the root-shaking method [13] . Soil samples of F0 and G0 were obtained at a depth of 5 - 25 cm. The fresh soil samples were sieved through 2 mm mesh. The bulk soil samples were stored at 4˚C for physical and chemical characterization [12] , while the rhizosphere samples were stored at −70˚C prior to DNA extraction [14] .

Konjac plants were collected and assessed for the presence and severity of plant disease. Disease incidence statistics included 100 plants from each soil sample. Soft rot symptoms of the leaves and petioles were identified in infected plants [15] .

Disease Incidence (DI) was calculated using the following formula:

DI = The number of infected plants Total number of plants investigated × 1 00 %

Bacterial community DNA sequencing

Soil microbial DNA was isolated using the E.Z.N.A.^® Soil DNA kit (Omega Bio-Tec, Inc., USA) according to the manufacturer’s instructions. The DNA extracts were stored at −20˚C prior to their use in PCR amplifications. The V3-V4 hypervariable regions of 16S rRNA were amplified from the microbial genomic DNA using barcoded fusion primers (forward primers: 341 FCCTACACGACG CTCTTCCGATCTNCCTACGGGNGGCWGCAG, reverse primers: 805 R GACTGGAGTTCCTTGGCACCCGAGAATTCCAGACTACHVGGGTATCT AATCC). For each sample, PCR products were combined from three 50 μl reactions, quantified, and the PCR products from each soil sample were combined in equimolar concentrations for paired-end 250 cycle sequencing on the Illumina MiSeq instrument as described by Ding et al. [12] . All sequences were subjected to systematic checks to reduce sequencing and PCR errors. High quality sequence reads (i.e., >200 bp in length, quality score of >25, exact match to barcode and primer, and containing no ambiguous bases) were retained. Sequences were clustered and assigned to operational taxonomic units (OTUs) at a 3% dissimilarity level using the Uclust software (uclust v1.1.579). The bacterial diversity is shown by the number of unique OTUs [12] . The raw reads were deposited into the NCBI Sequence Read Archive (SRA) database under accession number SRP076136.

The results were analyzed using SPSS software (version 14.0 for Windows, Chicago, IL, USA). The relative abundance of bacterial genera in the rhizosphere soil between moncropped and intercropped konjac grown for three years were analyzed using one-way analysis of variance (ANOVA; Levene’s test was used to assess the equality of variances before performing ANOVA), and significant differences between the means were determined by the LSD test. The differences were considered statistically significant when P < 0.05. T-tests were performed with Microsoft Excel, and all p-values were adjusted by False Discovery Rate (FDR) using the BH method with the mt.rawp2adjp function in R. The substrate utilization pattern, which was based on the Ecoplate data, was analyzed by principal component analysis. A probability level of P < 0.05 was used to indicate statistical significance, unless otherwise stated. Additionally, the alpha diversity statistics including Chao1 and ACE richness estimator, Shannon diversity index and Simpson diversity index were calculated for each sample.

Konjac usually grows in slightly acidic or neutral soil. The most appropriate pH for normal growth of konjac is 6.0 - 7.5. Soil pH decreased dramatically in the monocrop and remained basically neutral in the intercropped fields over the course of the experiment (three years) (Table 1). The soil pH was maintained within an acceptable range in the intercropped fields, but the soil pH of farmland-grown konjac became acidic in years 2 and 3. The accumulation of phenolic acids ultimately results in a significant decrease in soil pH. The organic matter, available nitrogen, available phosphorus, and available potassium contents in the soil were higher in the third year than in the first two years for the monocrop, while the soil nutrients were generally unchanged under intercropping conditions, which indicating continuous cropping soil nutrient was eutrophication under monocrop and relative balance under intercropping under.

We observed significant differences in disease among the different tillage modes. As shown Figure 2, the disease incidences for soft rot under monocrop and intercropping conditions were 73.1% and 3.1%, respectively. Compared with the

Letters indicate the Shortest Significant ranges (SSR) at P = 0.05 for the different treatments. Different letters denote a significant difference at p < 0.05.

konjac plants grown under monocrop conditions, the intercropped plants had almost no detectable soft rot disease. These results show very clearly that the incidence of soft rot disease can be effectively reduced when konjac is grown under economic forest intercropping conditions.

Initially, a total of 1,277,396 reads were generated by sequencing the 16S rDNA gene fragments for all samples. After sequence filtering, 1,235,270 high quality sequences (average length of 273 bp) with a 97 % valid sequence ratio were used for this project. Based on 97% species similarity, a total of 2617 operational taxonomic units (OTUs) were separately obtained from all samples. Among the OTUs, 400 OTUs (~15%) were returned as unclassified. The average length of the sequence clusters was 440 bp, and they were classified into different taxonomic groups using uclust [16] .

A total of 1,277,396 paired-end 250-bp reads were acquired from all 24 samples. After initial quality control, 1,235,270 high quality sequences were obtained. With an average of 97,202.6875 sequence reads per sample (the minimum number from one sample was 117,323 and the maximum was 191,931) and used for this project. The average length of the sequence reads was 440 bp, and they were classified into different taxonomies using UCLUST software.

Moreover, the alpha diversity indices including richness (Chao 1 and ACE) and diversity (Shannon and Simpson) are shown in Table 2. For the monocrop rhizosphere soil, the Chao 1, ACE, Shannon, and Simpson indices showed significantly lower for G3. Compared with the field rhizosphere soil, the intercropped rhizosphere soil F2 is very low.

The β-diversity was estimated using the unweighted UniFrac distance metrics. The principal coordinate analysis (PCoA), which is based on the weighted

UniFrac distances of the 16S rRNA sequences, was performed to cluster bacteria populations from the soil samples from 0, 1, 2, and 3 years of continuous cropping in the different farming modes (Figure 3). The first two factors accounted for 48.4% and 11.7% of the variation. The bacteria community of each soil sample was separated by number of cropping years in the monocrop, but the bacteria communities were not well separated in the intercropped soil samples (Figure 3). These results demonstrate that the bacteria communities of the soil monocrop samples changed significantly with increasing years of continuous cropping, while the intercropped samples did not appear to show the same changes, except possibly for the soil sample from year 3.

Comparisons of the relative abundance of the top 43 classified bacterial genera (relative abundance > 0.3%) showed significant variations among the rhizosphere soils (Figure 4). The predominant genera were Gemmata (16.25% < relative abundance < 24.55%), Aquicella (4.36% < relative abundance < 12.11%), and Planctomyces (4.65% < relative abundance < 7.73%) across all samples. With increasing years of continuous cropping, Candidatus, Koribacter, Lactobacillus, Pirellula, Achromobacter, Chitinophaga, Spirochaeta, Sphingomonas, Akkermansia, Fimbriimonas, Rothia, and Luteolibacter increased under monocropping conditions, but decreased for the intercropped planting, while Paenibacillus, Afifella, and Sphingobium decreased for the monocrop but increased for the intercrop.

There were 11 out of the 43 classified bacterial genera that were related to the duration of continuous cropping in the two different planting modes (Table 3). The relative abundance of Pirellula, Achromobacter, Chitinophaga, Sphingomonas, Rothia, Nocardia, Asticcacaulis, Deinococcus, and Catellatospora were significantly affected by planting age and mode; Pseudomonas was significantly

Note: M: monocrop and intercrop; Y: Years; *Same letters indicates no significant difference, different letters mean significant difference.

affected by the planting mode; and Myroides was significantly affected by planting age.

As shown in Table 4, the abundance of Staphylococcus, Anaeromyxobacter, and Alicyclobacillus in the monocrop, and Luteolibacter, Oscillospira, Opitutus, Candidatus, Rhabdochlamydia, Devosia, Asteroleplasma, Corynebacterium, Bdellovibrio, and Streptococcus in the intercroppped soils were significantly different (P < 0.05) with respect to the duration of continuous cropping. Luteolibacter,

Note: *Same letters indicates no significant difference, different letters mean significant difference.

Opitutus, Lentzea, Allobaculum, Corynebacterium, Patulibacter, Asteroleplasma, Leptolyngbya, Spirosoma, Aquicella, Paenibacillus, and Nitrosopumilus were significantly different (P < 0.05) between the two planting modes. The differences in Luteolibacter were highly significant (P < 0.01), and Opitutus, Asteroleplasma and Corynebacterium differed significantly (P < 0.05) in the G3/G0 and F3/G3 soil sample comparisons.