Spartina alterniflora invasion alters soil microbial community composition and microbial respiration following invasion chronosequence in a coastal wetland of China

The role of exotic plants in regulating soil microbial community structure and activity following invasion chronosequence remains unclear. We investigated soil microbial community structure and microbial respiration following Spartina alterniflora invasion in a chronosequence of 6-, 10-, 17-, and 20-year-old by comparing with bare flat in a coastal wetland of China. S. alterniflora invasion significantly increased soil moisture and salinity, the concentrations of soil water-soluble organic carbon and microbial biomass carbon (MBC), the quantities of total and various types of phospholipid fatty acids (PLFAs), the fungal:bacterial PLFAs ratio and cumulative microbial respiration compared with bare flat. The highest MBC, gram-negative bacterial and saturated straight-chain PLFAs were found in 10-year-old S. alterniflora soil, while the greatest total PLFAs, bacterial and gram-positive bacterial PLFAs were found in 10- and 17-year-old S. alterniflora soils. The monounsaturated:branched PLFAs ratio declined, and cumulative microbial respiration on a per-unit-PLFAs increased following S. alterniflora invasion in the chronosequence. Our results suggest that S. alterniflora invasion significantly increased the biomass of soil various microbial groups and microbial respiration compared to bare flat soil by increasing soil available substrate, and modifying soil physiochemical properties. Soil microbial community reached the most enriched condition in the 10-year-old S. alterniflora community.


Results
Soil and plant properties. Soil moisture, salinity, WSOC, SOC, and SON in S. alterniflora soils were significantly higher than those in bare flat soil (Table 1). Soil moisture was the highest in 17-and 20-year-old S. alterniflora soils followed by 6-and 10-year-old S. alterniflora soils ( Table 1). The pH in S. alterniflora soils were significantly lower than that in bare flat soil with the lowest pH in 6-and 17-year-old S. alterniflora soils ( Table 1). The highest salinity and the lowest WSOC were found in 20-year-old S. alterniflora soil, while the greatest SOC concentration was found in 17-year-old S. alterniflora soil (Table 1). Aboveground biomass was the highest in the 17-year-old S. alterniflora community, followed by 20-, 10-, and 6-year-old S. alterniflora communities (Table 1).

Soil microbial biomass and structural diversity. The highest MBC concentration was found in
10-year-old S. alterniflora soil, followed by 17-, 20-, 6-year-old S. alterniflora and bare flat soils (Fig. 1a). In contrast, MBN did not significantly vary in the S. alterniflora invasion chronosequence (Fig. 1b). The MBC:MBN ratio in 10-year-old S. alterniflora soil was significantly higher than that in 6-year-old S. alterniflora and bare flat soils (Fig. 1c). The MBC:SOC ratio in bare flat soil was significantly higher than that in S. alterniflora soils, and the MBC:SOC ratio in 10-year-old S. alterniflora soil was significantly higher than that in 6-, 17-, 20-year-old S. alterniflora soils (Fig. 1d). MBC concentration was strongly associated with soil and plant properties except soil pH ( Table 2). The quantities of the total PLFAs, bacterial, fungal, gram + bacterial, gram − bacterial, arbuscular mycorrhizal fungi (AMF), actinomycete, monounsaturated, branched, and saturated straight-chain (SSC) PLFAs in S. alterniflora soils were significantly higher than those in bare flat soil (Figs 2 and 3). The quantities of soil total PLFAs, bacterial, and gram + bacterial PLFAs were the highest at 10 and 17 years, followed by 6 years and 20 years after S. alterniflora invasion (Fig. 2a,b,f). The lowest fungal PLFAs were found in 20-year-old S. alterniflora soil (Fig. 2c). The quantity of AMF PLFAs gradually declined following S. alterniflora invasion in the chronosequence (Fig. 2h). The quantities of actinomycete and branched PLFAs gradually increased from 6 to 17 years after S. alterniflora invasion, but declined in the 20-year-old S. alterniflora soil (Fig. 3a,d). The quantities of gram − bacterial and SSC PLFAs were the greatest in 10-year-old S. alterniflora soil and the lowest in 20-year-old S. alterniflora soil (Figs 2e and 3b). The quantity of monounsaturated PLFAs was the most enriched in 10-and 6-year S. alterniflora soils, and it declined in 17-and 20-year S. alterniflora soils (Fig. 3c).
The fungal:bacterial PLFAs ratio in all S. alterniflora soils were considerably higher than that in bare flat soil (Fig. 2d). The lowest gram − :gram + ratio was found in 17-year-old S. alterniflora soil, and the highest gram − :gram + ratio was found in 10-year-old S. alterniflora and bare flat soils (Fig. 2g). The monounsaturated:branched PLFAs ratio in 17-and 20-year-old S. alterniflora soils were considerably lower than that in 6-and 10-year-old S. alterniflora soils (Fig. 3e). The bacterial stress index in 17-and 20-year-old S. alterniflora soils were significantly higher than that in 6-and 10-year-old S. alterniflora and bare flat soils (Fig. 3f).
Soil microbial respiration and temperature sensitivity. Cumulative microbial respiration in the 0-30 cm soil layer after 30 days of incubation at 35 °C was significantly greater than soil from the bare flat and S. alterniflora invasion chronosequence incubated at 25 °C (Fig. 4a). Cumulative microbial respiration in over 10-year-old S. alterniflora soils were considerably higher than that in 6-year-old S. alterniflora soil, which was higher compared with bare flat soil (Fig. 4a). Cumulative microbial respiration was not only significantly related to soil moisture, salinity, SOC, WSOC, SON, and aboveground and root biomass ( Table 2) but also strongly associated with total and all types of PLFAs (Table 3).
Similarly, microbial respiration on a per-unit-PLFAs and qCO 2 after 30-days of incubation at 35 °C were significantly higher than in soils from all communities that were incubated at 25 °C, the exception being qCO 2 in 6-year-old S. alterniflora soil (Fig. 4b,c). Microbial respiration on a per-unit-PLFAs gradually increased following the S. alterniflora invasion chronosequence (Fig. 4b). The qCO 2 in 6-and 20-year-old S. alterniflora soils were significantly higher than that in 10-year-old S. alterniflora and bare flat soils (Fig. 4c). The Q 10 value of microbial respiration did not significantly change across bare flat soil and the S. alterniflora invasion chronosequence (Fig. 5). Pearson's correlation analysis showed that qCO 2 was significantly associated with total and all types of PLFAs (Table 3) and that the Q 10 value of microbial respiration was not significantly related to soil and plant properties ( Table 2), and it was also unrelated to total and all types of PLFAs (Table 3).
Controls on soil microbial community. Eight variables of soil and plant properties, including, soil moisture, pH, salinity, SOC, WSOC, SON, aboveground and root biomass, explained 84.1% of the total variability in the PLFAs (Fig. 6). The variations in the PLFAs were strongly correlated with soil moisture (F = 131.16, P = 0.0020), WSOC (F = 8.75, P = 0.0040), and salinity (F = 10.07, P = 0.0020) (Fig. 6). The biggest variation, at 82.1%, was explained by the total variations of the PLFAs in Axis 1, and Axis 2 explained 1.9% of the total variations of the PLFAs (Fig. 6). Meanwhile, Pearson's correlation analysis showed that the PLFAs were significantly positively correlated with soil moisture, salinity, SOC, WSOC, SON, aboveground and root biomass, but they were negatively associated with soil pH ( Table 2).

Discussion
Our findings not only added to various evidence that S. alterniflora invasion greatly accelerated soil organic C and N accumulation due to greater biomass input 14,29 (Table 1), but also found that S. alterniflora invasion significantly increased MBC concentration and the quantities of the total and all types of PLFAs compared with bare flat soil (Figs 1-3). Soil C sources are considered as crucial ecological driving factors for microbial community dynamics 32 . Increased biomass input and the soil substrate following S. alterniflora invasion 14,16 (Table 1) possibly enhanced MBC and all types of PLFAs. This speculation was supported by our Pearson's correlation analysis that MBC and all types of PLFAs were highly associated with SOC, WSOC, SON, and above-and below-ground biomass ( Table 2). The aboveground biomass and SOC content progressively increased in 6-to 17-year-old S. alterniflora soils and then fell in soils collected afterwards (Table 1). Interestingly, the highest MBC, gram − bacterial and SSC PLFAs were found in 10-year-old S. alterniflora soil (Figs 1-3), and the greatest total PLFAs, bacterial and gram + bacterial PLFAs were found in 10-and 17-year-old S. alterniflora soils (Fig. 2), implying that the soil microbial community reached the most enriched condition in 10-year-old S. alterniflora soil. The WSOC is the most important available substrate and directly provides available C and energy for soil microbial metabolism 33,34 . Although 17-year-old S. alterniflora soil had a bigger SOC stock compared with 10-year-old S. alterniflora soil (Table 1), there was no significant difference in the quantity of total PLFAs between 10-and 17-year-old S. alterniflora soils (Fig. 2) due to the same level of WSOC in both samples (Table 1). Meanwhile, the decrease in total PLFAs, bacterial, fungal, gram + bacterial, and branched PLFAs in 20-year-old S. alterniflora soil compared with 10-and 17-year-old S. alterniflora soils (Figs 2 and 3), may be caused by lower levels of readily   Fig. 6). The fungal:bacterial PLFAs ratio can be used to reflect the physiological state of the soil microbial community that is particularly involved in SOM accumulation and turnover 35 , and the ecosystem's buffering capacity 36 . In this study, S. alterniflora invasion significantly increased the fungal:bacterial PLFAs ratio compared with bare flat soil (Fig. 2d), indicating a higher C accumulation and self-buffering capacity in S. alterniflora soils. Previous studies reported that fungi have higher C assimilation efficiency relative to bacteria 37,38 , owing to their stronger ability to decompose plant compounds 39,40 . Higher C assimilation efficiency in fungi may result in more organic C being converted into more recalcitrant humic materials 37 . Hence, the increased fungal:bacterial PLFAs ratio in S. alterniflora soil can possibly enhance soil organic C sequestration following S. alterniflora invasion.
We found that the gram − :gram + PLFAs ratio ranged from 1.52 to 1.78 across S. alterniflora invasion chronosequence (Fig. 2g), suggesting that gram − bacteria dominated in bare flat and S. alterniflora salt marsh and that there were copiotrophic condition in this coastal wetland ecosystem 22 . Previous studies showed that higher soil pH would increase gram − bacteria and decrease gram + bacteria 25,26,27 . 10-year-old S. alterniflora and bare flat soils had a higher pH compared with 17-year-old S. alterniflora soil (Table 1), and this may be one of the reasons that the lowest gram − :gram + PLFAs ratio was found in 17-year-old S. alterniflora soil (Fig. 2g). This result was consistent with our finding that 17-and 20-year-old S. alterniflora soils had a greater bacterial stress index than 10-and 6-year-old S. alterniflora and bare flat soils (Fig. 3f). Generally, a high bacterial stress index represents a slow rate of growth and long turnover time for gram − bacteria 41 . Thus, higher bacterial stress index in 17-and 20-year-old S. alterniflora soils indicated that they had slower growth rates and lower turnover rates of the gram − bacteria community relative to 10-and 6-year-old S. alterniflora and bare flat soils (Fig. 3f). Additionally, previous studies have reported that gram − bacteria preferentially utilize fresh plant residual as an available C source, while gram + bacteria prefer to use older, humified and more microbially processed SOM 42,43 . Thus, the lowest gram − :gram + PLFAs ratio and the highest bacterial stress index were found in 17-year-old S. alterniflora soil (Figs 2g and 3f), indirect suggesting increased degree of SOM decomposition and humification in 17-year-old S. alterniflora soil compared with 10-year-old S. alterniflora and bare flat soils.
The monounsaturated and branched PLFAs were generally used to indicate aerobic and anaerobic microorganism biomass, respectively 41,44 . We found that the highest levels of monounsaturated PLFAs were found in 10and 6-year-old S. alterniflora soils (Fig. 3c), while the highest levels of branched PLFAs was found in 17-year-old S. alterniflora soil across the invasion chronosequence (Fig. 3d), suggesting that the quantity of aerobic microbes was the highest in the early stage of S. alterniflora invasion, whereas the quantity of anaerobic microbes reached maximum levels at the later stage of S. alterniflora invasion. The ratio of monounsaturated:branched PLFAs continually declined during the invasion chronosequence (Fig. 3e), implying that the percent of anaerobic microbes gradually increased and the percent of aerobic microbes gradually decreased during the invasion chronosequence. This may be highly associated with the gradual increase in soil moisture during the invasion chronosequence (Table 1). Higher soil moisture provides stronger soil anaerobic conditions, which might be more suitable for anaerobic microorganism growth and facilitate SOM accumulation 45,46 .
Soil microbial respiration was highly dependent on soil temperature, moisture, and C inputs 47 , and was strongly associated with the quantities of the soil microbes and WSOC concentration 34 . In this study, cumulative microbial respiration at 25 °C and 35 °C at different invasion times of S. alterniflora soils were significantly higher than that in the bare flat (Fig. 4a), which was highly correlated with total and various types of PLFAs, WSOC, aboveground and root biomass (Tables 2 and 3). Thus, the increased cumulative microbial respiration in S. alterniflora soils may be greatly attributed to higher C inputs, and increase in available substrate (e.g., WSOC) and the microbial biomass (Table 1; Figs 1-4). We found that cumulative microbial respiration and the microbial respiration:total PLFAs ratio at 35 °C in all communities were significantly higher than that at 25 °C (Fig. 4a,b), primarily because elevated temperature increases soil enzyme activities and further drives SOM decomposition 48 . Interestingly, the Q 10 value of microbial respiration showed no obvious changes during the S. alterniflora invasion chronosequence (Fig. 5), likely because the Q 10 of microbial respiration is not influenced by the differences in the microflora 34 (Table 3). Although cumulative microbial respiration at 25 °C and 35 °C showed no significant differences between 10-, 17-, and 20-year-old S. alterniflora soils (Fig. 4a), cumulative microbial respiration on a per-unit-PLFAs basis at 25 °C and 35 °C progressively increased following the increase of invasion time (Fig. 4b), suggesting that S. alterniflora invasion may decrease microbial C utilization efficiency and enhance respiration loss in this coastal wetland ecosystem 48 . Generally, an increase in qCO 2 may reflect a decrease in microbial C utilization efficiency and ecosystem stabilization 36,49 , and a higher MBC:SOC ratio could indicate an increase in microbes use C efficiency 16,50 . In this study, the lowest qCO 2 and the highest MBC:SOC ratio were found in 10-year-old S. alterniflora soil following invasion from 6 to 20 years (Figs 1d and 4c), indicating that 10-year-old S. alterniflora soil had the highest microbial C utilization efficiency and the greatest ecosystem stabilization following invasion from 6 to 20 years. In addition, bare flat soil had higher the MBC:SOC ratio and lower cumulative microbial respiration on a per-unit-PLFAs and qCO 2 relative to S. alterniflora soils (Figs 1d and 4b,c), implying that S. alterniflora invasion resulted in low microbial C utilization efficiency compared to bare flat 16,50 , which may be due to S. alterniflora with lower quality and more recalcitrant substance (e.g., lignin) is difficult to be utilized by microbes 14 .  Table 3. Pearson correlation coefficients between soil microbial respiration and microbial biomass across the communities. * P < 0.05; ** P < 0.01. MRP: microbial respiration: total PLFAs ratio; qCO 2 : respiration quotient; See Table 2 for abbreviations. In conclusion, this study highlighted the variations of soil microbial community structure and activity following bare flat was converted to S. alterniflora salt marsh in a invasion chronosequence in a coastal wetland of China. Specifically, S. alterniflora invasion greatly increased the total and various types of soil microbial biomass and cumulative microbial respiration but significantly decreased microbial C utilization efficiency compared to bare flat. 10-year-old S. alterniflora community had the most enriched soil microbial community and the highest microbial C utilization efficiency across 6 to 20 years S. alterniflora invasion. Soil microbial biomass decreased after 17 years S. alterniflora invasion. The variations in the microbial community structure and activity may in turn deeply affect SOM accumulation and ecosystem C and N cycling. This study represents a step forward in our understanding of microbial communities as affected by plant invasion, and it provides valuable insights regarding the better understand the influence mechanism of plant invasion on soil organic C pool.

Methods
Site description and sampling. This study was conducted at the core area of the Jiangsu Yancheng Wetland National Nature Reserve, Rare Birds, China (JYWNNRRB) (32°48′ 47″ -34°29′ 28″ N, and 119°53′ 45″ -121°18′ 12″E). This area is characterized as warm temperate with an average annual temperature of 13.8 °C, average annual precipitation of 1000 mm, and average annual sea water salinity of 3.09% 16 . JYWNNRRB was designated as an internationally important wetland site (Ramsar) in 2002. S. alterniflora was introduced to the bare flat of the JYWNNRRB in 1983, and it quickly expanded to form large areas of S. alterniflora salt marshes following mudflat aggrading 14 . The bare flat and S. alterniflora salt marshes are located on the low and middle areas of the intertidal zone with semidiurnal tidal periodicity 20 . The seaward invasion region of S. alterniflora is a bare flat that had no vegetation prior to S. alterniflora invasion 16 .
The sampling region, with its different S. alterniflora invasion times, was identified based on analyses of  marked from the bare flat to the invasive 6, 10, 17, and 20 years S. alterniflora communities. Three 2 m × 2 m plots were randomly established within each location. Three soil samples (5-cm diameter × 30 cm depth) were collected randomly in each plot. The soil samples from each plot were mixed evenly to form a composite sample. Three 50 cm × 50 cm quadrats were established to collect aboveground biomass (i.e., the sum of leaves, stems and litter), and three root sampling blocks (15-cm length × 15-cm width × 30 cm depth) were excavated to collect root biomass in each community of each transect. All soil and plant samples were stored at 4 °C in the field and then transported to the laboratory for subsequent analysis.
Laboratory analysis. Each root sampling block was put through a 100 mesh sieve and flushed with water, and the roots remaining in the sieve were collected at the final step 14 . All plant samples were carefully cleaned and oven-dried at 65 °C for determining aboveground and root biomass. The visible plant and fauna residues were removed from the soil samples, and soil samples were then divided into three subsamples after thorough mixing. One subsample was air-dried and passed through 1-mm sieves to measure soil pH, salinity, SOC and SON. A subsample of 2-mm sieved fresh soil was stored at 4 °C to determine WSOC, MBC, MBN and microbial respiration. Another subsample was passed through 2-mm sieves and stored at −80 °C as quickly as possible after freeze-drying and was used to analyze for PLFAs. The soil subsample was weighed and oven-dried at 105 °C to determine soil moisture. Soil pH was measured in a soil-water suspension (1:2.5 soil:water) with a glass electrode. Soil salinity was determined in a soil-water suspension (1:5 soil:water) with a conductivity meter. Before the SOC and SON analyses, approximately 10 g of dried soil subsamples were treated with 1 M HCl at room temperature for 24 h to eliminate total inorganic C and N 29 , and unhydrolyzed residues were analyzed with a CN elemental analyzer (Vario PYRO cube elemental analyzer, Germany) to obtain SOC and SON concentrations 14 . WSOC was determined using the method described by Yang et al. 16 . Briefly, WSOC was extracted from 10 g moist soil samples after addition of 20 mL distilled water. The extracted fluid was vacuum filtered through a 0.45 μ m filter, and C concentration of the filtrate was rapidly determined by a Liqui TOCII analyzer (Elementar Analysensystem GmbH, Germany).
Microbial respiration measurements. Microbial respiration was measured by alkali absorption of CO 2 evolved at 25 °C and 35 °C for 30 days in a laboratory aerobic incubation experiment with soil 37,48 . Briefly, the fresh soil sample (20 g dry weight equivalent) was evenly placed in a 50 mL glass beaker. Distilled water was added to the soil samples to maintain moisture at 60% of water-holding capacity. The glass beaker was placed in a 500 mL mason jar, and the glass tubes containing 10 mL 0.5 M NaOH solution was placed in each mason jar to capture CO 2 evolved by the soil in the mason jar. The mason jar was sealed and incubated at 25 °C and 35 °C in the dark for 30 days. After incubation for 6, 12, 18, 24, and 30 days, the glass tubes that were equipped with NaOH were removed, and the mason jar was opened for several minutes to maintain sufficient O 2 levels. The amount of CO 2 was determined by titration of the NaOH solution with 0.1 M HCl in two drops BaCl 2 .
Temperature sensitivity (Q 10 ) of microbial respiration was determined using equation (1)  where R 2 and R 1 are the mean microbial respiration rate at T 2 (35 °C) and T 1 (25 °C), which are the temperature levels within 30 days of incubation. The microbial respiration quotient (qCO 2 ) was calculated by dividing the microbial respiration (mg CO 2 30 day) per kg by the MBC 36 .
Statistical analyses. All of the statistical analyses were performed using SPSS Statistics 19 software. Data not meeting assumptions of normality and homogeneity of variance were log-or cube root-converted prior to statistical testing. One-way analysis of variance (ANOVA) was used to determine the statistical significance of the effect of S. alterniflora invasion time on soil and plant properties, microbial biomass and various types of PLFAs, cumulative microbial respiration, microbial respiration:total PLFAs ratio, qCO 2 and Q 10 . One-way ANOVA was also used to determine the statistical significance of the incubation temperatures on cumulative microbial respiration, the ratio of microbial respiration:total PLFAs and qCO 2 . Pearson's correlation analysis was performed to correlate soil microbial indexes with the soil and plant characteristics, and to correlate soil microbial respiration indexes with microbial biomass (i.e., each of the PLFAs). Soil and plant characteristics were tested for significant contributions to explain the variations in the PLFAs data with redundancy analysis (RDA) using CANOCO software for Windows 4.5. The statistical significance of the RDA was tested using the Monte Carlo permutation test (499 permutations; P < 0.05).