Introduction

Rehmannia glutinosa is a member of Scrophulariaceae family, and it is one of the 50 basic herbal medicines in traditional Chinese medicine (TCM). However, the replant problems of R. glutinosa extremely restrict its production. Fields used for R. glutinosa cultivation are typically replanted every 15–20 years (Yang et al. 2011b). Jiaozuo City, Henan Province, is the geo-authentic production area of R. glutinosa (34° 48′ N to 35° 30′ N, 112° 02′ E to 113° 38′ E) (Yang et al. 2011b). Because of the lack of availability of fields in Jiaozuo, R. glutinosa is replanted in the same location at intervals of less than 15 years or outside the Jiaozuo area to meet market demand. All of these factors result in a vicious circle that is characterized by a lower tuber yield, worse product quality and higher disease incidence. Therefore, understanding the underlying mechanism of R. glutinosa replant problems is urgent.

The inhibitory effect of allelochemicals is regarded as the primary factor inducing replant problems (Muller 1969; Singh et al. 2010; Zhu et al. 2014). According to the previous studies, the occurrence of phenolic acids were frequently reported and regarded as one class of important secondary metabolites possess strong allelopathic activity (Li et al. 2010; Muscolo and Sidari 2006; Olofsdotter et al. 2002). What is noteworthy is that the secondary metabolism of plants presents unique diversity and richness and results in more than 200,000 secondary products, and the phenolic acids are only part of the products of various secondary metabolic pathways (Dixon and Strack 2003; Seigler 1998; Weston et al. 2012). Thus, lots of potential allelochemicals should be identified to explore the plants’ allelopathy. In recent years, increasing compounds (like ketones, alkaloids, terpenes) were reported as potential allelochemicals of other plants (Guo et al. 2015; Kato-Noguchi et al. 2011; Ma et al. 2009). Interestingly, the growth inhibitory activity of some potential allelochemicals was much greater than that of phenolic acids (Kato-Noguchi et al. 2012), and some compounds presented a synergistic effect when exhibiting their allelopathic effect (Al Harun et al. 2015; Zhang et al. 2012; Zuo et al. 2015). In previous studies, the polar compounds in ethyl acetate extraction of R. glutinosa leaves and fibrous roots have been preliminary analysed for the identification of allelochemicals (Li et al. 2012; Wang et al. 2010). However, the potential autotoxins left in the soil after R. glutinosa cultivation are still unknown. Thus, a relative integrated identification of potential autotoxins of R. glutinosa is necessary and meaningful in understanding the mechanism of its autotoxicity.

Plant-microbe interactions can positively or negatively influence plant growth through a variety of mechanisms, and root exudates are one of the most important chemical signals in the determination of whether an interaction will be benign or harmful (Bais et al. 2006). Generally, allelochemicals are regarded as the distinctive components of root exudates and somewhat influence the microbial community (Bertin et al. 2003). According to a recent report of Pseudostellaria heterophylla replant problems, the accumulated autotoxins in rhizosphere soil increased the harmful microorganisms and decreased beneficial microorganisms, resulting in an imbalance of microbial community structure and the degradation of soil ecological function (Lin et al. 2015). The previous studies found that the structural and functional diversity of soil microbial communities in newly R. glutinosa planted soil were distinctly different from consecutively mono-cultivated soils, and the fungal community decreased with the increase of mono-cultivated years (Wu et al. 2011; Zhang et al. 2011). However, information on microbial community distribution within root-zone soil of R. glutinosa and the influence of R. glutinosa cultivation on microbes are still unclear. Thus, investigation is necessary to lay a foundation for the further understanding of the interactions among plant, autotoxins and microbes.

In this study, the ‘autotoxic circle’ of R. glutinosa was evaluated by seedling tests and verified by plantlet tests. We then quantified five central focal phenolic acids (4-hydroxybenzoic acid, vanillic acid, syringic acid, 3,4-dihydroxybenzoic acid and vanillin) in allelopathic research and provided a more precise identification and confirmation of potential autotoxins to understand R. glutinosa autotoxicity. Twenty new compounds emerged in the ‘autotoxic circle’, four of which were chosen based on the availability of the chemicals, the specific dimensionality distribution in the soil and previous reports to track their autotoxicity. Finally, the bacterial and fungal community distributions in the root-zone were investigated by PCR-DGGE.

Materials and methods

Container-culture experiment

Soil from Jiaozuo City that had never been used for R. glutinosa planting was selected for this container-culture experiment. The R. glutinosa cultivar ‘Wen 85-5’ is widely planted in the main production region. It was used as the seed and cultivated in the centre of the pots (diameter 100 cm, depth 40 cm) based on the production regulation. The following root-zone soil samples were subsequently collected: close soil (CS), including the soil collected after the harvest roots were shaken and the soil brushed from the surface of the roots; intermediate soil (IS), including the soil within 10 to 20 cm of the centre of the pots and distant soil (DS), including the soil within 20 to 30 cm of the centre of the pots. The soil from the pots that was not cultivated with R. glutinosa was collected as a control sample (CK). The soil cultivated with or without R. glutinosa was watered and managed equally, and the surface soils (3 cm deep) were abandoned before sampling to exclude environmental influences. All of the above soil samples were air-dried, blended and sieved through a 2-mm mesh for subsequent analysis.

Assessment of ‘autotoxic circle’

To define the R. glutinosa ‘autotoxic circle’, a seedling test was conducted to estimate the autotoxic effect of the following root-zone soils: CS, IS, DS and CK were weighed (20 g) and transferred into Petri dishes in triplicate, and 10 mL of distilled water was sprayed every 3 days after 20 R. glutinosa seeds were planted in the soil. In addition, cyclopentaneacetic acid, 3-oxo-2-(2-pentenyl)-, methyl ester, [1R-[1a,2b(Z)]]- (compound 2), trans-2-hexadecenoic acid (compound 10), tert-hexadecanethiol (compound 14), palmitoleic acid (compound 19) and their mixtures, which were identified and selected from the GC-MS results, were dissolved in a small volume of ethanol and transferred onto the double-layered filter paper in the Petri dishes. In a draft chamber, the solvents evaporated for 1 h. The filter paper that contained these compounds was moistened with 5 mL distilled water, and the final concentrations of the compounds in water were 2, 5 and 10 μg/mL. Distilled water was used as a control, and 20 R. glutinosa seeds were placed on the filter paper. All dishes were maintained at 26 °C under fluorescent lights for 11 h (0800 hours–2000 hours), and the fluorescent light intensity was 4.68 ± 0.14 × 103 lx. The radicle length was measured after incubation for 7 days. Additionally, a plantlet test was performed to verify the ‘autotoxic circle’ under aseptic conditions, in which the influence of polar and non-polar compounds within different root-zone soils on plantlets root growth was assessed as follows: Soil samples were weighed (200 g) and extracted using 400 mL HPLC-grade methanol and n-pentane for 24 h, separately. Then, the extracts were filtered and diluted with corresponding solvents to 400 mL. Eighty millilitres methanol extraction, as well as 80 mL n-pentane extraction were transferred into glass bottles and evaporated in a draft chamber for 8 h. Segment of stems with four leaves, including the apical meristem approximately 3 cm in length from the tissue culture plantlets of R. glutinosa, were used as explants and cultured in the bottles according to the same conditions of the seedling tests. The MS formulation was used as the nutrient medium, and six replications of each soil sample extract were performed. After 4 weeks, the roots of plantlets were washed and scanned using an Epson Expression 10000XL scanner (Epson Co., Ltd., Long Beach, CA, USA). Root length was determined by analysing the scanned images using WinRHIZO version Pro2007d software (Regent Instruments Inc., Quebec, Canada).

Quantitative analysis of phenolic acids in different root-zone soils

The five phenolic acids (4-hydroxybenzoic acid, vanillic acid, syringic acid, 3,4-dihydroxybenzoic acid and vanillin) were analysed by HPLC, and their concentrations were determined by the resulting peak areas combined with standard curves. The standard curves were plotted based on single and mixed phenolic acids. The methods were as follows: standardized phenolic acids (Sigma-Aldrich Co., MO, USA) were weighed (0.025 g) and dissolved in 25 mL Milli-Q water as single standards, and then each single standard was combined and diluted to different gradients (0.5, 1, 5 and 10 μg/mL) for the mixed standards. The injection samples for the soil samples were obtained by the following steps: soil samples were weighed (5.0 g) and extracted using 100 mL HPLC-grade ethyl acetate (Fisher, NJ, USA), dried via rotary evaporators and redissolved in 5 mL HPLC-grade methanol (Fisher, NJ, USA). The HPLC system was composed of an LC-20AD liquid chromatography system, a SPD-20A ultraviolet detector, a SIL-20A autosampler, a CTO-10AS column oven and a CBM-20A communication module (Shimadzu, Kyoto, Japan). The chromatographic separation was performed on an Agilent XDB-C18 column (5 μm, 250 mm × 4.6 mm; Santa Clara, CA, USA) using methanol (solvent A; 28:72, v/v) and 2 % acetic acid (solvent B) as a mobile phase at a flow rate of 0.8 mL/min. During the experiment, the injection volume was 20 μL and the chromatogram was monitored at a wavelength of 280 nm. The column temperature remained at 30 °C. The samples and standard solutions were filtered through a 0.22-μm membrane before injection. The stability was tested, and the percent of recovery rates was measured (Zhang et al. 2015).

Identification of compounds in different root-zone soils

The polar and non-polar chemicals in the soil samples were extracted as follows: soil samples were weighed (50 g) and extracted using 200 mL HPLC-grade methanol and n-pentane, separately. Then, extracts were dried in rotary evaporators before redissolving in a 5 mL homologous extraction. Extracted samples were filtered through a 0.22-μm membrane for injection. A GC-MS analysis was performed with Agilent 7890A gas chromatography, which interfaced with an Agilent 7000 mass selective detector equipped with an electron impact ion (ionization energy 70 ev). A 1 μL extracted sample was injected via a ‘splitless’ mode. The full-scan monitoring mode (m/z 50–1500) was employed as the detection mode. The capillary column was DB-5 (30 m × 0.25 mm × 0.25 μm, Agilent, USA). The oven temperature was set to 50 °C for 2 min, followed by a gradient of 4 °C/min up to 250 °C and then held for 2 min. Helium was used as the carrier gas at 1 mL/min. The interface was 200 °C, and the ion source was 250 °C.

Analysis of microbial community in different root-zone soils

PCR-DGGE technology was employed to analyse the microbial community of the soil samples. The implementation method was as follows: soil samples were weighed (500 mg) to extract total genomic DNA in triplicate using a Fast DNA Spin kit for soil (Q-Biogene, USA) following the manufacturer’s specifications. Bacterial 16S ribosomal DNA (rDNA) was amplified using the universal bacterial primer pair GC-338F and 518 R, and fungal 18S rDNA was amplified using GC-Fung and NS1. In a final volume of 50 μL, the PCR amplifications of the bacteria were performed, which included 100 ng soil DNA, 3.2 μL dNTP (2.5 mM), 0.4 μL rTaq (5 U/μL), 1 μL GC-338F (20 mM), 1 μL 518 R (20 mM), 5 μL 10× PCR buffer and an amount of ddH2O. The 50 μL PCR amplifications for fungi consisted of 100 ng soil DNA, 4 μL dNTP (2.5 mM), 0.5 μL rTaq (5 U/μL), 0.5 μL GC-Fung (20 μM), 0.5 μL NS1 (20 μM), 5 μL 10× PCR buffer and an amount of ddH2O. The thermal cycling for bacteria consisted of 30 cycles of pre-denaturation at 94 °C for 4 min, denaturation at 94 °C for 30 s, renaturation at 55 °C for 30 s, extension at 72 °C for 30 s and a final extension at 72 °C for 10 min. The thermal cycling for fungi proceeded as follows: 30 cycles of pre-denaturation at 94 °C (5 min), denaturation at 94 °C (1 min), renaturation at 50 °C (45 s), extension at 72 °C (1 min) and final extension at 72 °C for 10 min. The DGGE was performed for bacteria using 8 % (w/v) acrylamide gel with a 35–55 % denaturant gradient and run in 1× TAE (Tris-acetate-EDTA) buffer for 4 h at 60 °C and 150 V. In addition, a DGGE was performed for fungi using 7 % (w/v) acrylamide gel with 15–35 % denaturant gradient and run in 1× TAE buffer for 6 h at 60 °C and 150 V.

After electrophoresis, the gels were stained for 30 min with SyBR green I nucleic acid stain and photographed using a Bio-Rad trans-illuminator (Bio-Rad Laboratories, Segrate, Italy) under UV light. The visually prominent bands from the gels were excised using sterilized pipette tips and then recycled using Ploy-Gel DNA Extraction Kit. Two microlitres of buffer containing DNA were used as a template for PCR using the same primer pair but without the GC-clamp. The PCR products were purified and cloned into Escherichia coli DH5 using the PMD 18-T plasmid vector system for bacteria and pEASY-T for fungi to sequence. Three clones per band were sequenced. The 16S rDNA and 18S rDNA sequences were compared against those available in the NCBI database.

Calculations and statistical analyses

An index of the inhibition rate (IR) was used to compare the radicle length and the plantlet root length of treatments with the controls.

$$ \mathrm{I}\mathrm{R}=\frac{\mathrm{control}-\mathrm{treatment}}{\mathrm{control}}\times 100\ \% $$

where IR >0 indicates growth inhibition and IR <0 indicates growth promotion (Li et al. 2012).

The samples’ peaks from GC-MS results were identified by their mass spectra combined with a Kovats index (KI), which was calculated by the injection of a solution containing a homologous series of normal alkanes (C8-C40 purchased from ANPEL Scientific Instrument Co., Ltd., Shanghai, China).

$$ \mathrm{K}\mathrm{I}=100n+100\left({t}_x-{t}_n\right)/\left({t}_{n+1}-{t}_n\right) $$

where n is the carbon number of normal alkanes, t x is the retention time of the sample peak, t n is the retention time of the n-alkanes peak eluted immediately before the sample peak and t n+1 is the retention time of the n-alkanes peak eluted immediately after the sample peak. The chemicals in the results that showed the highest KI similarity with the KI* of compounds listed in the NIST library were considered reliable.

For analysis of the molecular community profiles, the Molecular Analyst Fingerprinting software (version 1.61, BioRad, Veenendaal, NL) was used. From the molecular (PCR-DGGE) community fingerprints analysed by the Molecular Analyst software, the Shannon-Weaver index (H′) and the evenness index (E) were calculated according to the aforementioned method (Garbeva et al. 2008). The most similar sequences in NCBI blasting results were selected and listed as the identification results. The data of seedlings’ radicle length, plantlets’ root length, Shannon-Wiener diversity index (H′) and evenness index were subject to an analysis of variance by SPSS software. Each value was expressed as the mean of three replicates ± SE.

Results

Assessment and verification of ‘autotoxic circle’ of R. glutinosa

The radicles’ length of the R. glutinosa seedlings grew in the close soil, the intermediate soil was significantly inhibited and the radicle length in the distant soil was not statistically different from that of the control soil (Figs. 1 and 2). The strongest inhibitory effect was observed firstly in the close soil (IR = 27.52 %) and secondly in the intermediate soil (IR = 12.30 %). These results indicated that the autotoxic effect of R. glutinosa was significant within 20 cm (close soil and intermediate soil) of the plant, and this effect lessened considerably at a distance beyond 20 cm (distant soil).

Fig. 1
figure 1

Different root-zone soils of R. glutinosa. CS the close soil shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa

Full size image
Fig. 2
figure 2

Inhibition rate of different root-zone soils on radicle length of R. glutinosa seedlings. Different letters indicate significant differences among different radicle length at P < 0.05 according to Duncan’s test. CS the close soil shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa, CK the control soil that was not cultivated with R. glutinosa

Full size image

To exclude the influence of microbes on the assessment of the autotoxic circle, the polar and non-polar compounds within different root-zone soils were extracted and their effects on root growth of plantlets were evaluated through tissue culture. The plantlet test results were similar to the seedling tests: The strongest inhibitory effect was observed in the close soil (IR = 61.92 %) and then in the intermediate soil (IR = 25.65 %). The plantlets’ root length which treated with extraction of distant soil showed no significant difference with that of the control soil (Fig. 3). This similar result indicated that the different compounds’ constitution within different root-zone soils resulted in the formation of R. glutinosa ‘autotoxic circle’. The diameter of this ‘autotoxic circle’ was 40 cm.

Fig. 3
figure 3

Inhibition rate of extracts of different root-zone soils on root length of R. glutinosa tissue culture plantlets. Different letters indicate significant differences among different radicle length at P < 0.05 according to Duncan’s test. CS the close soil shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa, CK the control soil that was never cultivated with R. glutinosa

Full size image

Concentration of five phenolic acids in the root-zone soils

HPLC results showed that 4-hydroxybenzoic acid, vanillic acid, syringic acid, 3,4-dihydroxybenzoic acid and vanillin were present in all of the treatment soils and the control soil. A comparison of the concentration of these five phenolic acids showed that they were relatively higher in the close soil and intermediate soil than in the distant soil and control soil, and the distant soil did not present significant differences from the control soil (Fig. 4). Furthermore, the concentrations of 4-hydroxybenzoic acid, vanillic acid and syringic acid in the intermediate soil and close soil were at the same level, whereas the concentrations of 3,4-dihydroxybenzoic acid and vanillin were significantly higher in the intermediate soil than in the close soil. This finding indicates that the concentration of phenolic acids in soil within 20 cm of R. glutinosa increased by the cultivation of this plant, and the soil beyond 20 cm was unaffected.

Fig. 4
figure 4

Concentration of five phenolic acids in different root-zone soils. Different letters indicate significant differences among different concentrations at P < 0.05 according to Duncan’s test. CS the close soil shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa, CK the control soil that was never cultivated with R. glutinosa

Full size image

Identification of autotoxins in the root-zone soils

The GC-MS spectrograms of the polar and non-polar fractions in the soil samples showed that the close soil containing the most peaks was followed by the intermediate soil. The spectrogram for the distant soil was highly similar to that of the control soil (Fig. 5). To identify the specific chemicals originating from R. glutinosa cultivation, the control soil was used as the background, and 20 specific peaks (10 in methanol extraction and 10 in n-pentane extraction) were selected for identification by NIST database searching combined with KI screening. The molecular formula and KI of these chemicals were listed in Table 1. Notably, eight chemicals (compounds 1, 2, 10, 11, 12, 14, 18 and 20) present in the close soil were also specific to the intermediate soil. However, they were not detected in the distant soil or control soil. Therefore, these eight compounds were derived from R. glutinosa cultivation.

Fig. 5
figure 5

GC-MS spectra of compounds in methanol (a) and n-pentane (b) extracts of different R. glutinosa root-zone soils. Same peak numbers indicate same compounds. CS the close soil that was shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa, CK the control soil that was never cultivated with R. glutinosa

Full size image
Table 1 Identification of 20 compounds that specific to the close soil and intermediated soil
Full size table

According to previous studies, some compounds detected in this work have been reported as allelochemicals in other plants (Cetin et al. 2010; Colvin and Gliessman 2011; Preston et al. 2004; Roy and Barik 2014; Tang and Gobler 2011). Based on these reports and the availability of the chemicals and the specific dimensionality distribution in the soil, four chemicals and their mixtures were selected and tested for their potential autotoxic effects on R. glutinosa radicle growth (Table 2). The results showed that these chemicals had different effects on seedling growth. Particularly, the seedling radicles were inhibited by 33.71–74.29 % when they were treated with the single compounds and mixtures at different concentrations. Additionally, hormesis, the phenomenon whereby a usually detrimental environmental agent (e.g., chemical substance, radiation) changes its role to provide beneficial effects when administered at low intensities or concentrations (Furst 1987), was indicated by the radicle growth of R. glutinosa when the seeds were treated with compound 10 and compound 14 at a concentration of 2 μg/mL (P < 0.05). This stimulatory effect changed to an inhibitory effect at concentrations of 5 and 10 μg/mL. Moreover, seed germination was prevented (the radicle length was less than 1 mm) under treatment with the compound mixture and compound 2 at relatively higher concentrations (5 and 10 μg/mL compound mixture and 10 μg/mL compound 2). A comparison of effects on the IR among different compounds indicated that the compound mixture induced the strongest allelopathic effect on the seedling growth of R. glutinosa. The results demonstrated that these specific compounds extracted from the close and intermediate soils of R. glutinosa were potent inhibitors of seedling growth. There is a strong probability that these compounds belong to the autotoxins of R. glutinosa.

Table 2 Effect of the selected autotoxins on the radicle growth of R. glutinosa
Full size table

PCR-DGGE analysis of the microbial community in different root-zone soils

The DGGE patterns showed that some microbes were only detected or showed relatively higher intensity in the corresponding soil range (Fig. 6).

Fig. 6
figure 6

DGGE profiles of the bacterial 16S rDNA (a) and fungal 18S rDNA (b) fragments. CS the close soil shaken and brushed from the R. glutinosa roots, IS the intermediate soil within 10 to 20 cm from the R. glutinosa, DS the distant soil within 20 to 30 cm from the R. glutinosa, CK the control soil that was never cultivated with R. glutinosa. Three replicates for each soil

Full size image

These 23 soil-specific and individual density bands were excised and used as a template for PCR amplification. The sequence results of these PCR products were compared against those available in the NCBI database, and the closest relative species were listed (Tables 3 and 4).

Table 3 Identification of specific microbial species in different R. glutinosa root-zone soils
Full size table
Table 4 Identification of specific fungal species in different R. glutinosa root-zone soils
Full size table

As can be seen from the Tables 3 and 4, the sequences from these bands showed 94–100 % similarity to the closely related database sequences in GenBank. The close soil presented the most specific species (nine bacterial species and four fungal species), followed by the intermediate soil (one bacterial species and three fungal species) and the distant soil (one bacterial species and one fungal species). Among these species, eight were uncultured species, and the others belong to different genera including Chitinophaga, Hydrogenophaga, Bacillus, Pseudomonas, Ochrobactrum, Micrococcus, Ramicandelaber, Orbilia, Acremonium, Humicola, Heyderia, Lasiosphaeria, Dothidotthia and Coniosporium. Additionally, the soil samples’ Shannon-Wiener diversity index and evenness index of total bacteria, as well as total fungi, were calculated according to the 16S/18S rDNA PCR-DGGE banding patterns (Table 5). Overall, the microbial diversity and evenness index of the three soil types exhibited a descending order: close soil > intermediate soil > distant soil. In the close soil, the bacterial diversity index, the fungal diversity index and the fungal evenness index were significantly higher than that in the intermediate and distant soils, with values of 3.29 ± 0.032, 2.97 ± 0.104 and 1.01 ± 0.040, respectively. Significant differences in the bacterial evenness were not observed in these three soil samples. The findings presented here indicate that the microbial community changed due to R. glutinosa cultivation. The microbial constitution in the different R. glutinosa root-zone soils shifted dramatically, particularly within 20 cm of the plant that characterized by the appearance of various exclusive species. However, the microbes in the soil beyond this range were slightly affected.

Table 5 Shannon-Wiener diversity and evenness indices based on the 16S rDNA/18S rDNA PCR-DGGE patterns
Full size table

Discussion

Various allelochemicals/autotoxins result in the plants’ allelopathy/autotoxicity, and the autotoxicity is regarded as one of the important factors responsible for the plants’ replant problems (Muller 1969; Singh et al. 2010; Zhu et al. 2014). Our results demonstrated that the soil cultivated with R. glutinosa developed an autotoxicity effect on the growth of R. glutinosa seeds, and this effect decreased as the distance from the plant increased. Specifically, the autotoxic effect faded away at a distance beyond 20 cm. Moreover, the assessment of autotoxic effect under aseptic conditions showed similar results with soil culture conditions, implying that an ‘autotoxic circle’ was formed after R. glutinosa cultivation. This autotoxic range was induced by the compounds that appeared after R. glutinosa cultivation.

To further understand the autotoxins that induced R. glutinosa autotoxicity, phenolic acids, which were reported as the allelochemicals of many plants (Bais et al. 2006; Li et al. 2010), within the ‘autotoxic circle’ and beyond were analysed. The HPLC results showed that the concentrations of five phenolic acids within the ‘autotoxic circle’ were significantly higher compared with their concentrations in the soil beyond, implying that R. glutinosa cultivation increased the concentration of phenolic acids. According to the previous reports of R. glutinosa, the allelopathy intensity of root exudates of different R. glutinosa growth stages showed positive correlation with its syringic acid concentration (Zhang et al. 2015). Seven phenolic acids were detected in the acetate-soluble extracts of R. glutinosa fibrous roots, and their bioassay showed a significant suppressive function on the seedling growth (Li et al. 2012). Similar autotoxic effects of phenolic acids were also reported in other plants. For instance, nine phenolic compounds (p-hydroxybenzoic acid, vanillin, syringic acid, vanillic acid, coumaric acid, ferulic acid, cinnamic acid, salicylic acid, benzoic acid) were detected in the soil of commercially cultivated American ginseng, and all these compounds inhibited the radicle and shoot growth of itself (Bi et al. 2010). The autotoxic effect of ferulic acid and p-coumaric acid on cucumber and soybean was also reported (Politycka and Mielcarz 2007; Suzuki et al. 2008). Based on our results and these researches of other plants, these five phenolic acids are components of the autotoxins found in R. glutinosa. Although this evidences was powerful, it raised more questions than answer.

A noticeable phenomenon is that some preceding crops of R. glutinosa (such as wheat, maize) have been reported to release various phenolic acids, but they did not induced replant problems in R. glutinosa (Jia et al. 2014; Machinet et al. 2011; Maksimovic et al. 2008; Pham et al. 2011). However, the replant problems occurred when R. glutinosa was cultivated as a preceding crop of R. glutinosa. Why is it that the phenolic acids released by Gramineae crops did not induce the R. glutinosa replant problems? Based on the studies of other plants, the synergistic effect of allelochemicals drew our attention. Research on canola suggested that the synergistic effect of sinapyl alcohol, p-hydroxybenzoic and 3,5,6,7,8-pentahydroxy flavones played a role for canola allelopathy against annual ryegrass (Asaduzzaman et al. 2015). Research on boneseed’s allelopathy indicated that the mixture of ferulic acid, catechin, phloridzin and p-coumaric acid showed a stronger inhibiting effect on Isotoma axillaris compared to individual phenolics (Al Harun et al. 2015). The combined allelopathic effect of three allelochemicals of Chara vulgaris also expressed synergistic inhibitory effects on the growth of Microcystis aeruginosa (Zhang et al. 2009). Thus, we hypothesized that there are some other R. glutinosa allelochemicals/autotoxins besides phenolic acids, and their synergistic effect exerts a stronger autotoxicity. Our GC-MS and bioassay results provided powerful evidence for this hypothesis. The GC-MS results showed that the variety of compounds decreased as the distance increased from the close soil, to the intermediate soil, to the distant soil; the compounds’ constitution in the distant soil was very similar to the control soil that was not planted with R. glutinosa. Moreover, some compounds were only detected in the close and intermediate soil, implying these compounds were specific to the soil within the ‘autotoxic circle’ and emerged after the R. glutinosa cultivation. To track their potential autotoxic effect, four of the compounds were chosen for bioassay. The bioassay illustrated the following results: the trans-2-hexadecenoic acid (compound 10) and the tert-hexadecanethiol (compound 14) promoted the radicle growth with relative lower concentration, while inhibiting the radicle growth with higher concentrations, and the methyl jasmonate (compound 2) and the palmitoleic acid (compound 19) inhibited the radicle growth with even low concentrations. We suggest that they are the unreported members of R. glutinosa autotoxins. Additionally, a synergistic effect, which characterized by the inhibitory effect of the mixtures of these four compounds were much stronger than each single compound, was found. Similar results indicated that methyl jasmonate released from sagebrush had an allelopathic effect on Nicotiana attenuata when combined with other compounds (Preston et al. 2002). Therefore, there is a high probability that R. glutinosa autotoxicity is expresses by the synergistic effect of various autotoxins, including phenolic acids and other autotoxins (such as methyl jasmonate and other compounds listed in our result). The phenomenon of why the problems associated with replanted R. glutinosa do not occur after the cultivation of Gramineae crops may be partly explained by the absence of some other important autotoxins. Further studies are necessary to determine the synergistic effect of various autotoxins and to gain a more in-depth understanding of R. glutinosa autotoxicity. The identification of more potential autotoxins by combining GC-MS with other instruments (like LC-TOF-MS, LC-NMR) would be useful.

When tracking the potential functions of the identified microbes, we found that most of them were still unknown. However, some species belong to the same genus as our identified species were confirmed to have positive effect on plant growth. For instance, some species of Bacillus, Pseudomonas, Acremonium and Humicola could promote plant growth due to their antagonistic activity, the production of IAA or specific substance to activate resistance for control of diseases (Dharni et al. 2014; Hassan et al. 2015; Khabbaz et al. 2015; Wenhsiung 2009; Yang et al. 2014). On the other side, some species of these genera showed negative effect on plant growth due to their broad-spectrum pathogenic activity (Elbanna et al. 2014; Furuya et al. 2005; Iakovleva et al. 2013; Racedo et al. 2013; Zhang et al. 2014). Additionally, the report of other plant suggested that the harmful microbes were increased with the decrease of beneficial microbes under mono-cultivated conditions, and the soil ecological function was affected (Lin et al. 2015). Our previous report confirmed that continuous cropping of R. glutinosa decreased the bacteria species and simplified the bacterial community structure (Zhang et al. 2010). Thus, the waxing and waning of specific microbes may be also one important reason for the R. glutinosa replanted problem. More importantly, two identified species (Arthrobotrys oligospora, Acremonium sclerotigenum) may play an important role in the micro-ecological of R. glutinosa rhizosphere according to previous reports. It was confirmed that A. oligospora not only increased the plant growth and enhanced nutritional value of tomato but also functioned in controlling nematodes (Niu and Zhang 2011; Singh et al. 2013; Yang et al. 2011a). A. sclerotigenum was confirmed to be the pathogen of brown spot disease in apple (Li et al. 2014). Further studies on whether and how these two species shift under consecutive mono-cultivated conditions would be meaningful.

Furthermore, the newly research results of our team indicated that phenolic acid mixture could promote mycelial growth, sporulation and toxin production of Fusarium oxysporum while inhibiting growth of the Pseudomonas sp. W12 (Wu et al. 2015). Research has confirmed that F. oxysporum is the pathogenic pathogens of R. glutinosa disease (Li et al. 2013) and that most strains of Pseudomonas showed antagonistic activity against F. oxysporum (Borah et al. 2015; Chin-A-Woeng et al. 2000; Srinivasan et al. 2009). In this work, we found that some congeneric species showed a close interaction with allelochemicals. For instance, Pseudomonas catalysed (-)-catechin conversion into 3,4-dihydroxybenzoic acid, and 3,4-dihydroxybenzoic acid showed a higher phytotoxicity than (-)-catechin (Masuda et al. 2007); Bacillus aquimaris involved in the degradation of allelopathic m-tyrosine (Khan et al. 2013); F. oxysporum, a replant disease pathogen in P. heterophylla rhizospheric soil, increased sharply under a consecutive-replant system, and some phenolic acids promoted the growth of F. oxysporum f.sp. (Zhao et al. 2015). In production of R. glutinosa, 15 to 20 years of crop rotation is taken for its replanting (Yang et al. 2011b). In other words, the auto-inhibitory effect of R. glutinosa could last for a long time, and the farmland spends at least 15 years on itself-rehabilitation. During the process of formation and rehabilitation of the replant problems, some specific microbes may play an important role to catalyse, to transform and to degrade the autotoxins. Thus, more studies should be conducted not only on the various soil microbes, the autotoxins and the plants’ abnormal metabolism but also on their interactions and functions in mediating plants’ replant problems.

Conclusions

To summarize, the ‘autotoxic circle’ of R. glutinosa was firstly assessed in this work. Moderate yield practices have surpassed high yield practices to guarantee the concentration of active ingredients in traditional herb production. Defining the range of R. glutinosa’s autotoxic effect provided a refined map by which planting density can be optimized and the corresponding standards can be set in GAP and GMP. Moreover, we performed a more precise identification of several potential autotoxins of R. glutinosa based on the ‘autotoxic circle’. Our results indicated that there are other R. glutinosa autotoxins in addition to phenolic acids. The synergistic effect of various potential autotoxins may contribute to the manifestation of their autotoxic activity. Furthermore, we found that many specific species emerged in the root-zone soils after R. glutinosa planting. One of these species is a potential pathogen, and another is a potential plant growth promoting fungus. This plant beneficial fungus may be utilized in the alleviation of R. glutinosa replanted problems. However, in addition to harness the power of this species, we need to understand how and why it matters with R. glutinosa. Future research should investigate the identification of other potential autotoxins through the means of multiple technologies and verify the function of specific microbes in the rhizosphere ecosystem.