1 | INTRODUCTION
According to data from the Global Forest Resources Assessment (IPCC, 2013), the total amount of carbon sequestered by global forests is as high as 1950-3150 Pg C each year; the carbon sequestration by vegetation reaches 450-650 Gt, accounting for approximately 43.5% of the total. On the other hand, Schlesinger (1990) noted that soil is the largest reservoir of carbon in terrestrial ecosystems, storing two-thirds of their organic carbon. The rhizosphere microecosystem is the link between plants, soil and microorganisms, the most active part of the global carbon biochemical cycle, and the focus of research on the systematic mechanisms of the global carbon cycle (Schweinsberg-Mickan, Jorgensen, & Muller, 2012; Carrillo, Dijkstra, Pendall, LeCain, & Tucker, 2014). Photosynthetic products are transported from leaves to various tissues, such as root tissues, for storage, and this transport occurs through material flow caused by pressure differences. The remaining photosynthetic products are released into the surrounding soil by the plant root system as various organic and inorganic compounds that form rhizodeposits. In addition, carbon that enters the soil through a series of biochemical processes is circulated and redistributed among roots, the soil and microorganisms to maintain the balance of rhizosphere carbon absorption and release (Jones, Nguyen, & Finlay, 2009).
Studies have shown that carbon sequestration by plants is particularly important for rhizosphere microorganisms. Over 40% of the complex carbon produced by photosynthesis enters rhizosphere soil through plant roots to nourish microorganisms and maintain their normal metabolic functions (Rodriguez, et al., 2019). Cheng et al. (1996) noted that root exudates, as carbon sources for microbial utilization, increase respiration by rhizosphere microorganisms. The differences in carbon utilization efficiency among microbial communities mainly depend on the differences in their microbial functional families related to carbon source utilization (Xu, 2012). Yin et al. (2018) found that increasing the atmospheric CO2 concentration to promote carbon metabolism in Kandelia candel did not significantly increase the abundance of the rhizosphere bacterial community. Moreover, Xiao et al. (2017) noted that carbon sequestration by Bothriochloa ischaemumsignificantly increased the contents of total PLFAs (phospholipid-derived fatty acids) and fungal PLFAs in rhizosphere soil. In addition, Feng et al. (2011) indicated that under increased carbon metabolism conditions in rice, the proportions of aerobic, anaerobic and phototrophic bacteria in the bulk soil increased (from 0.5% to 1.5%), while no significant effects were observed in rhizosphere soil. It has been noted that among rhizosphere microorganisms, the response of fungi to photosynthetic carbon sequestration by plants is clearer than that of bacteria (Xu, 2012) because fungal mycelia can accelerate the turnover cycle of the fungal carbon metabolism (which takes approximately one week), while bacteria generally need more than two weeks to turn over carbon (Ostle, et al., 2003; Staddon,Ramsey, Ostle, Ineson, & Fitter, 2003).
Masson pine (Pinus massoniana ) is a large perennial tree that is widely distributed in 17 provinces and autonomous regions in the southern Qinling Mountains in China (Wu, et al., 2020). Masson pine thrives in light, is shade intolerant and prefers a warm and humid climate. It can grow in red soil, gravel soil and sandy soil and is used as a pioneer tree species for forest restoration in barren mountainous areas (Wang, et al., 2019). Previous studies have shown that Masson pine has a high carbon sequestration ability. Elisa et al. (2003) showed that the carbon sequestration in Masson pine organs ranged from 533.93 to 568.08 g·kg-1, which is higher than the carbon contents of 32 common tropical tree species (444.0-494.5 g·kg-1). The carbon sequestration ability of plants directly affects the quantity of root exudates (Ainsworth, 2008). However, the response of soil microbial communities, particularly rhizosphere microorganisms, to plant carbon sequestration has rarely been studied, especially under Masson pine. In this study, based on Masson pine from different families, samples with significant differences in carbon sequestration ability were selected as experimental materials. The corresponding rhizosphere soil was obtained for 16S rRNA and ITS sequencing to analyze the differences in in the number and taxonomic diversity of bacteria and fungi and their patterns in response environmental factors. This research provides guidance toward further understanding the response of microorganisms to plant carbon sequestration, which will be helpful in predicting the effects of climate change on rhizosphere microbial communities.
2 | MATERIAL AND METHODS
2.1 | Study site
This study was conducted in the progeny test plantation of the Masson pine seed orchard at the Baisha State-Owned Forest Farm (25°15’N, 116°62’E), Shanghang County, Fujian Province. The samples from the forests were collected in 2001, and the experimental trees were planted in 2003. There were 68 families (Kang, 2012). Before the experiment was carried out, it was found that due to human activities, the number of samples in some families did not meet the requirements for statistical analysis. Therefore, given the situation, 24 families were selected as the experimental families (Table S2).
2.2 | Estimation of carbon storage in different Masson pine families
To avoid destroying trees, regression equations were used to estimate the average carbon sequestration by each family. Approximately 30 individuals in each family were randomly selected as experimental samples, and the height and DBH (diameter at breast height) of each sample were measured. The regression equations in professional standards released by China’s Forestry Administration (Cai, et al., 2014) were used to estimate the biomass of each organ (including the trunk, branches, leaves, bark, and roots) based on the tree height and DBH. Then, the biomass of each organ was multiplied by the corresponding carbon coefficient (trunk: 0.5186, branches: 0.5174, leaves: 0.5785, bark: 0.4994, and roots: 0.5082) to obtain the total carbon sequestration. The total carbon sequestered by a single tree was obtained by adding the carbon sequestered in each organ. The mean value of all samples from the same family was used as an index to evaluate the carbon sequestration level of the family. The relevant regression equations are provided in Table S1.
2.3 | Soil sampling
The families with high, low and intermediate carbon sequestration were selected for follow-up experiments. Three individuals with carbon sequestration values that were close to the mean value for each experimental family were selected as the samples. Five sampling points near each sample were chosen, and 5~10 cm bulk soil was dug up. The roots were carefully pulled out of the soil with a shovel, and the loosely attached soil was gently shaken off. The rhizosphere soil was closely attached to the roots. Litter and humus were removed from the soil surface before soil sampling, and the rhizosphere soil from the five sites near each sample tree was mixed together to form a composite soil sample. The soil samples were divided into two parts; one part was placed into a 5 ml freezing tube and immediately frozen in dry ice for sequencing, and the other part was loaded into a 50 ml centrifuge tube for analyses of soil physical and chemical properties.
2.4 | Detection of the physical and chemical properties of rhizosphere soil
The rhizosphere soil total organic carbon was determined by the combustion oxidation nondispersive infrared absorption method according to the Chinese Environmental Protection Standard (Jian, Zhai, Wang, & Cai, 2020). The experimental temperature was set at 900°C, and the oxygen pressure was 900 Mbar. The gas flow rate in the analysis module was 150-165 mL·min-1. The soil total nitrogen was determined based on the Kjeldahl method as provided in the Chinese Environmental Protection Standard (Zhang, et al., 2015). To determine the rhizosphere soil pH, 5 g soil samples were ground and passed through a 100 mesh sieve, and 12.5 mL ddH2O was added. The samples were mixed by vortexing and oscillation and centrifuged at 5000 r·min-1 for 5 min, and the supernatant was separated and directly tested with a pH meter. To determine the rhizosphere soil moisture, soil samples (20 g) were weighed and dried at 105°C for 6 h. After cooling to room temperature and weighing again, the difference between the two weights was divided by the fresh soil weight (20 g) to obtain the moisture content.
2.5 | DNA extraction, high-throughput sequencing and analysis
The FastDNATM Spin Kit for Soil (MP Biomedicals, California, USA) was used to extract DNA from 0.5 g soil samples. The operation process was performed in strict accordance with the instruction manual. The barcode primers 515F (5′-GTGCCAGCMGCCGCGG-3′) and 907R (5′-CCGTCAATTCMTTTRAGTTT-3′) were used to amplify the bacterial 16S rDNA. ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS2R (5′-GCTGCGTTCTTCATCGATGC-3′) were used to amplify the fungal ITS sequence. The PCR amplification procedure was carried out according to the instructions forTransStart ® FastPfu DNA Polymerase (TransGen Biotech, Beijing). Each sample was subjected to PCR three times. The PCR products from the same sample were mixed and detected with 2% agar-gel electrophoresis, recovered using the AxyPrepDNA gel recovery kit (Axygen Biosciences, CA, USA), and purified using the agar-gel DNA purification kit (TransGen Biotech, Beijing). The purified PCR products were sequenced according to the default parameters on the MiSeq PE300 platform. The original data were stored in the NCBI Sequence Read Archive database (accession number PRJNA662187 for bacteria and PRJNA662212 for fungi).
The raw sequence data were analyzed and quality-controlled using fastp (version 0.19.6, https://github.com/OpenGene/fastp). Bioinformatics statistical analysis was performed using Usearch (version 7.0, http://drive5.com/uparse/) for OTUs (operational taxonomic units) at 97% similarity. The OTUs were subsampled according to the minimum sample sequence number (35,000). The taxonomic analysis of the OTU representative sequences was carried out by the RDP classifier Bayesian algorithm (version 2.2, http://sourceforge.net/projects/rdp-classifier/, the default confidence threshold value was 0.7), and the community composition of each sample was counted at different taxonomic levels. The Silva bacterial 16S comparison database and the Unite fungal ITS comparison database were used.
2.6 | Statistical analyses
A Venn diagram of the microbial community diversity was constructed with the R Venn diagram package (Chen, & Boutros, 2011) based on the common and unique OTUs in the different samples. The Chao and Shannon index of α-diversity were calculated according to the corresponding formulas shown in Table S1. The relative abundance histogram at the phylum level was plotted using the R ggplot2 package (Kahle, & Wickham, 2013) based on the data sheet in the tax_summary_a folder. Welch T-tests were conducted to determine the significance of differences among families (Garcia-Lledo, Vilar-Sanz, Trias, Hallin, & Baneras, 2011). The Bonferroni method was used to conduct multiple test corrections to evaluate the significance level of taxonomic abundance differences and to identify the significantly different phyla among samples. For the core microbiome analyses at the genus level, common OTUs with relative richness values higher than 1% were extracted from different samples for graphical purposes. Pie and box charts were generated with the R ggplot2 package and SPSS 16.0, respectively (Perez-Jaramillo, et al., 2019). RDA (redundancy analysis) was used to clarify the relationships between soil physicochemical properties and the rhizosphere microbial community using the R vegan package (Ng, et al., 2014). Correlation heatmap analysis was performed by calculating the Pearson correlation coefficients between the environmental factors and the selected taxa and drawing the heatmap diagram using the R Pheatmap package. One-way ANOVA and Duncan multiple comparison tests were conducted using SPSS 16.0.
3 |