This is the accepted version of the article: Hu, Minjie; Peñuelas, Josep; Sardans i Galobart, Jordi; [et al.]. «Dynamics of phosphorus speciation and the phoD phosphatase gene community in the rhizosphere and bulk soil along an estuarine freshwater oligohaline gradient». Geo- derma, Vol. 365 (April 2020), art. 114236. DOI 10.1016/j.geoderma.2020.114236 This version is avaible at https://ddd.uab.cat/record/224211 under the terms of the license 1 Dynamics of phosphorus speciation and the phoD phosphatase gene community 2 in the rhizosphere and bulk soil along an estuarine freshwater-oligohaline 3 gradient 4 Minjie Hu 1, Josep Peñuelas 2, 3, Jordi Sardans 2, 3, Chuan Tong 4, Chang Tang Zhang 5*, Wenzhi Cao 1* 1 5 State Key Laboratory of Marine Environmental Science, College of the Environment and Ecology, Xiamen University, 6 Xiamen, Fujian 361102, China 2 7 CSIC, Global Ecology Unit CREAF-CSIC-UAB, Bellaterra, 08193 Barcelona, Catalonia, Spain 3 8 CREAF, Cerdanyola del Vallès, 08193 Barcelona, Catalonia, Spain 4 9 Key Laboratory of Humid Sub-tropical Eco-geographical Process of Ministry of Education, Fujian Normal University, 10 Fuzhou 350007, China 5 11 Department of Environmental Engineering, National I-Lan University, Sheen Lung Road, I-Lan 26047, Taiwan 12 Corresponding author information: 13 Prof. Wenzhi Cao 14 Institution: College of the Environment and Ecology, Xiamen University. 15 Address: Xiangan South Road #4226, Xiamen University, Xiamen 361102, Fujian, China. 16 E-mail: wzcao@xmu.edu.cn. 17 Tel: 086-0592-2185877. 18 Abstract: 19 Estuarine tidal marshes play a key role in phosphorus (P) retention and cycling; however, they are suffering 1 20 from small but significant increases in tidal saltwater intrusion. The likely impacts of these low-level saltwater 21 intrusions on P availability and microbial activity are unclear. Here, we investigated soil P speciation, alkaline 22 phosphatase (ALP) activity, and the phoD phosphatase gene community along a freshwater-oligohaline 23 gradient in the Min River estuary, southeast China. The results indicated that with the transition from 24 freshwater to oligohaline water, the levels of soil-water salinity, pH and sulfate (SO42-) content were greater, 25 and ALP activity was lower, which were associated with higher concentrations of organic P, available P, 26 aluminum-bound P, calcium-bound P, and occluded P and lower levels of iron-bound P. There was a strong 27 shift in the phoD phosphatase community composition in response to the freshwater-oligohaline gradient. Our 28 findings showed that with the transition from freshwater to an oligohaline environment, in addition to the 29 associated increases in salinity and soil pH and decreases in general microbial and biological activity and soil 30 organic carbon, there is a shift in soil P toward more recalcitrant and immediately available fractions with less 31 labile forms. 32 Keywords: Phosphorous; phoD phosphatase gene; Saltwater intrusion; Rhizosphere; Estuarine tidal marsh 2 33 1. Introduction 34 Tidal estuarine marshes, of which freshwater (salinity <0.5 ppt) and oligohaline (salinity = 0.5–5.0 ppt) 35 marshes may represent a significant area (Odum 1988; Weston et al. 2014), are globally distributed ecosystems 36 that play vital roles in ecological processes and nutrient cycling (Kirwan and Megonigal 2013; Tong et al. 37 2017). The complex biogeochemical cycles characteristic of these systems reflect various external stressors, 38 such as human activity, hydrodynamics, and varying salinity, leading to spatial heterogeneity and uncertainty 39 in the distribution of elements (Hu et al. 2018a). For example, estuarine studies of primary production have 40 shown shifts in phosphorus (P) limitation to nitrogen limitation due to the transition between freshwater and 41 seawater environments (Gireeshkumar et al. 2013; Hartzell et al. 2017). While nutrient cycling in coastal salt 42 marsh systems has been well studied, little is known about nutrient dynamics in a low-salinity gradient from 43 freshwater to oligohaline marshes, in which there are various microbial biogeochemical processes and plant 44 communities (Weston et al. 2014). This knowledge gap currently limits the understanding of the wetland 45 geochemical processes that drive nutrient cycling and the associated environmental responses. 46 Phosphorus is an essential element for living organisms and plays an important role in the regulation of 47 primary productivity and ecosystem function in wetlands (Lin and Guo 2016). The immobilization and release 48 processes of P in estuarine and coastal sediments are used in the quantification of the global P cycle due to the 49 considerable levels of P sequestration and the potential associated contributions to water eutrophication 50 (Gireeshkumar et al. 2013; Hartzell et al. 2010). Environmental conditions may drive the chemical speciation 51 of P that subsequently determines its environmental fate, cycling, and bioavailability in estuarine sediments 52 (Lin and Guo 2016). However, responses among P species to changes along a freshwater-saltwater gradient 53 are known to differ. For example, in the transition from freshwater to saltwater, Gireeshkumar et al. (2013) 54 reported decreased proportions of Fe-bound P (Fe-P) but increased proportions of calcium-bound P (Ca-P) and 3 55 total sulfur (TS) due to changes in sediment texture and redox conditions. In contrast, Paludan and Morris 56 (1999) showed that aluminum-bound P (Al-P) was an important inorganic P (IP) pool, regardless of salinity, 57 likely as a consequence of changes in ionic strength and aluminum availability. Previous studies of responses 58 of P speciation to freshwater-saltwater gradients have tended to focus on the effects of wide ranges in salinity 59 (Bai et al. 2017; Caraco et al. 1990; Gireeshkumar et al. 2013), while P responses to relatively narrow ranges 60 in salinity, such as from freshwater to oligohaline, are poorly understood. 61 Soil microbes are key drivers of P transformation and dominate the composition of P forms (Fraser et al. 62 2015; Stout et al. 2014). Alkaline phosphatase (ALP) describes a large group of enzymes that generally 63 originate from soil microbes and recycles organic P (OP) to orthophosphate via enzymatic hydrolysis (Ragot 64 et al. 2015). Three ALP-encoding gene families, comprising phoD, phoA, and phoX, have been identified 65 (Acuña et al. 2016). Among these genes, the distribution of the phoD phosphatase gene is widespread, and this 66 gene is considered the key ALP gene in marine sediment. The abundance of the phoD phosphatase gene has 67 been used as a measure of ALP bacterial diversity and distribution in a range of ecosystems (Fraser et al. 2015; 68 Lagos et al. 2016). The soil ALP activity and phoD phosphatase genes are affected by biotic and abiotic factors. 69 For example, Huang and Morris (2003) showed that ALP activity in tidal freshwater wetlands was positively 70 correlated with aboveground plant biomass and negatively associated with soluble reactive P concentration. 71 Acuña et al. (2016) found that ALP gene abundance in rhizosphere soils was positively correlated with ALP 72 activity but negatively correlated with P availability. Increased levels of salinity have been associated with 73 shifts in ALP activity, albeit with variable responses. Morrissey et al. (2014) reported a positive association 74 between increased salinity and ALP activity as a consequence of increases in the bioavailability of organic 75 substrates and changes in microbial community structure. However, Jackson and Vallaire (2009) observed that 76 an increase in salinity to 3.5 ppt decreased the phosphatase activity by almost 20%. Therefore, systematic 77 studies of P speciation, ALP activity, and phoD phosphatase genes would allow the evaluation of P dynamics 4 78 in tidal estuarine wetlands and the prediction of eutrophication risk due to P mobilization. 79 The Min River estuarine tidal marsh is the largest in southeastern China and is characterized by a 80 transition from freshwater to oligohaline water (Tong et al. 2017), providing an ideal model environment to 81 study responses of soil P availability and phoD gene community to variation along a freshwater-oligohaline 82 gradient. Previous studies in this estuary have found greater levels of porewater sulfate (SO42−), chloride 83 concentrations (Hu et al. 2019), and plant biomass, along with a larger pool of iron oxides and lower levels of 84 sulfide (Luo et al. 2019) in oligohaline marshes than in freshwater. However, variation in soil P dynamics and 85 phoD phosphatase genes between freshwater and oligohaline marshes remains unclear. Thus, the objectives of 86 this study were to (1) evaluate the P dynamics and differences in P speciation and the associated drivers in 87 estuarine marshes and (2) quantify the responses of ALP activity and phoD community composition to 88 freshwater and oligohaline transition and the associated interactions with P availability. We hypothesized that 89 soil P availability is greater at oligohaline sites due to shifts in P-related physicochemical properties (first 90 hypothesis) and that the transition from freshwater to oligohaline reshapes the bacterial phoD gene community 91 and modulates ALP activity due to the associated changes in nutrient levels (second hypothesis). 92 2. Materials and methods 93 2.1. Study sites 94 The study area at the Min River estuary (Fig. 1) is in a region with a humid, subtropical monsoon climate, 95 where the average annual temperature and precipitation are 19.85 °C and 1905 mm, respectively (Tong et al. 96 2017). The tides are semidiurnal over a 24-h cycle, and the soil surface is completely exposed at low tide (Tong 97 et al. 2014). Further details of the Min River estuary are described in our previous studies (Luo et al. 2019; 98 Tong et al. 2017). We selected three tidal marsh study sites (Fig. 1) that included a freshwater site in the 99 Tajiaozhou wetland (A; 25°56′59.9″N, 119°21′07.8″E) with an average salinity of 0.08 ±0.02 ppt and two 5 100 oligohaline sites at the mouth of the Min River estuary in the Bianfuzhou (B; 26°03′12.0″N, 119°33′25.1″E) 101 and Shanyutan (C; 26°0′50.7″N, 119°40′28.4″E) wetlands that are affected by tidal saltwater intrusion and 102 have average salinities of 1.27 ±0.09 and 3.31 ±0.14 ppt, respectively (Luo et al. 2019). At each marsh, we 103 selected an area dominated by the native sedge grass Cyperus malaccensis. 104 2.2. Soil and water sampling 105 In each marsh, two plots (each with 1 m × 1 m dimension) were established in 2018, one vegetated and the 106 other unvegetated. Three soil samples were randomly collected from each plot. For rhizospheres, entire plants 107 with roots were sampled from the vegetated plots, and rhizosphere soil was collected by scraping the soil that 108 was attached to the roots. For bulk soils, surface soil (0-20-cm deep) was collected from the unvegetated plots 109 and immediately placed in sterile, vacuum-sealed polyethylene bags that were transported to the laboratory in 110 a portable refrigerator containing ice within 12 h. The overlying water was simultaneously collected with the 111 soil samples from each plot during low tide, when the soil surface was exposed, and filtered using a 0.45-µm 112 cellulose membrane filter (Millipore Sigma, MA, USA). 113 At the laboratory, soil samples were sieved through a 2-mm mesh within an anaerobic glove box and then 114 divided into three subsamples. One subsample was immediately stored at -80 °C prior to DNA extraction and 115 bacterial phoD gene analysis, one subsample was stored at 4 °C for the measurement of ALP activity, and the 116 remaining subsample was immediately freeze-dried and stored in a glass vacuum desiccator for the 117 determination of soil physicochemical properties. 118 2.3. Analysis of soil and water physicochemical properties 119 Soil pH was measured in situ using an IQ150 meter (IQ Scientific Instruments, Carlsbad, USA), and electrical 120 conductivity (EC) was monitored using an EC meter (FieldScout 2265FS, Spectrum Technologies, Aurora, 121 USA). Soil total organic carbon (TOC) and TS concentrations were determined using a Vario MAX element 6 122 analyzer (Elementar, Frankfurt, Germany), and carbonate was removed using 10% HCl before TOC 123 determination. Soil particle size distribution was determined using a Malvern Mastersizer-2000 laser particle 124 size analyzer (Malvern Instruments, Malvern, UK). The overlying water SO42− concentration was measured 125 using ion chromatography (Dionex, Sunnyvale, USA), salinity was measured directly using a Salt 6+ salinity 126 meter (Oakton Instruments, IL, USA), and total P concentration (TPw) was determined using a continuous flow 127 analyzer (Auto Analyzer 3, Bran+Luebbe, Germany) following digestion with K2SO4. 128 2.4. Soil P speciation 129 Phosphorus speciation was characterized as total P (TP), inorganic P (IP), or organic P (OP) and was analyzed 130 according to the protocol method presented in Ruban et al. (1999) and Ruban et al. (2001). Further, the IP 131 fraction was classified as aluminum-bound P (Al-P), Fe-bound P (Fe-P), calcium-bound P (Ca-P), and occluded 132 P (O-P) and was determined following a sequential extraction procedure based on differential solubility in 133 different chemical extractants (Table S1), as described by Chang and Jackson (1957) and modified by 134 Hartikainen (1979). This method has previously been used for analysis of freshwater and saltwater sediments 135 (Laakso et al. 2016; Rahutomo et al. 2019; Ray et al. 2018; Zhang et al. 2015). Available P (AP) was measured 136 following a 0.5 M NaHCO3 extraction (Olsen and Sommers 1982). The P concentrations of extracts were 137 analyzed using the molybdenum blue spectrophotometry method (Paludan and Morris 1999). 138 2.5. Soil ALP activity assay 139 We estimated the ALP activity from the production of p-nitrophenol (pNP) from p-nitrophenyl phosphate (p- 140 NPP), as described by Tabatabai and Bremner (1969), where 1 g of soil (dry weight equivalent) was incubated 141 using para-nitrophenyl phosphate (Macklin Biotechnology, Shanghai, China) as the substrate in a modified 142 universal buffer at pH 11 (Tabatabai and Bremner 1969). Following incubation at 37 ºC for 1 h, the reactions 143 were terminated using 1 M NaOH and centrifuged at 4000 rpm for 15 min. The colorimetric determination of 7 144 pNP formation in the supernatant was analyzed using a spectrophotometer at 410 nm (Thermo Fisher Scientific, 145 MA, USA), and the ALP activity was expressed as micrograms of pNP released by 1 g of soil (dry weight 146 equivalent) per hour (µmol g−1 h−1). 147 2.6. Bacterial phoD gene analysis 148 Total genomic DNA was extracted from 0.5 g of fresh soil using Fast DNA SPIN extraction kits (MP 149 Biomedicals, Santa Ana, CA, USA) according to the manufacturer’s instructions and was stored at -20 ℃prior 150 to analysis. The quantity and quality of extracted DNAs were measured using a NanoDrop ND-1000 151 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and agarose gel electrophoresis, 152 respectively. 153 The primer set ALPS-F730 (5'-CAGTGGGACGACCACGAGGT-3') and ALPS-1101 (5’- 154 GAGGCCGATCGGCATGTCG-3′) was used to amplify the bacterial phoD gene (Sakurai et al. 2008), 155 where PCRs (30 µL) comprised 15 µL of Phusion High-Fidelity PCR Master Mix (New England Biolabs), 0.2 156 µM of forward and reverse primers, and approximately 10 ng of template DNA. Amplification was performed 157 in a thermal cycler with an initial denaturation at 98 ℃ for 2 min, followed by 25 cycles of denaturation at 158 98 °C for 15 s, annealing at 50 ℃ for 30 s, and elongation at 72 ℃ for 30 s, with a final extension for 5 min at 159 72 ℃(Huang et al. 2017). Purified amplicons were quantified on a microplate reader (BioTek, Vermont, USA) 160 using a Quant-iT PicoGreen dsDNA assay kit (Invitrogen, P7589) and were subsequently pooled in equal 161 amounts. Paired-end (2×300 bp) sequencing was performed using an Illumina MiSeq platform with a MiSeq 162 rReagent kit (v3) at Shanghai Personal Biotechnology (Personal Biotechnology Co., Ltd., Shanghai, China). 163 Although the ALPS primer have an amplification bias toward Alphaproteobacteria (Ragot et al. 2015; Tan et 164 al. 2013), this set of primers has been conducted by most of the previous studies (Acuña et al. 2016; Chen et 165 al. 2019a; b; Fraser et al. 2015; Fraser et al. 2017; Hu et al. 2018b; Huang et al. 2019; Luo et al. 2017; 166 Matsuoka et al. 2019; Sun et al. 2019; Tan et al. 2013; Valdespino-Castillo et al. 2014; Wan et al. 2019; Wei 8 167 et al. 2019), and making it possible to compare phoD-harboring bacterial community among studies. 168 2.7. Pyrosequence data processing 169 Raw sequence reads with exact matches to barcodes were assigned to respective samples and identified as 170 valid sequences. After chimera detection, the remaining high-quality sequences were classed into operational 171 taxonomic units (OTUs) using a sequence similarity threshold of 97%. To minimize differences in sequencing 172 depth across samples, an averaged, rounded rarefied OTU table was generated by averaging 100 evenly 173 resampled OTU subsets at <90% of the minimum sequencing depth for further analysis (QIIME, v1.8.0). 174 Abundance at the phylum, class, order, family, genus, and species levels was compared among samples or 175 groups using Metastats (White et al. 2009). Indices of OTU-level alpha diversity, such as the Chao1 richness 176 estimator, abundance-based coverage estimator (ACE), Shannon’s diversity index and Simpson’s evenness 177 index, were calculated using the OTU table in QIIME (Caporaso et al. 2010). We analyzed beta diversity to 178 investigate the structural variation in microbial communities using UniFrac distance metrics based on the 179 OTUs and visualized these results using principal coordinates analysis (PCoA) and pair-group method with 180 arithmetic mean (UPGMA) hierarchical clustering (Ramette 2007). 181 2.8. Statistical analysis 182 When necessary, data (i.e., P concentrations and environment variables) were log-transformed to meet the 183 ANOVA assumption of normality and homoscedasticity. Pearson's correlation coefficient was used to test the 184 potential correlation between soil P concentration, phoD gene diversity, ALP activity, and environmental 185 parameters, and the correlation matrix was visualized using the 'corrplot' package in R. Regression analysis 186 was used to explore the relationship between AP concentration, ALP activity, and phoD gene diversity. 187 Redundancy analysis (RDA) was performed to identify the main influencing factors of soil P dynamics using 188 Canoco 4.5 (Microcomputer Power, Ithaca, USA). Overall distributions and variations in soil P fraction, 189 environmental parameter, and phoD gene community among the study sites were summarized using a principal 9 190 components analysis (PCA) in Statistica 6.0 (StatSoft, Tulsa, USA). 191 3. Results 192 3.1. Soil and overlying water physicochemical properties 193 Physicochemical properties varied between the freshwater and oligohaline marshes (Table 1), where the levels 194 of soil pH, EC, and TS were greater, but that of TOC was lower, at oligohaline sites. Overall, surface soil 195 mainly comprised silt (55%), followed by sand (34%) and clay (11%). The proportions of silt and clay were 196 greater, while that of sand was lower, at oligohaline sites. The salinity and SO42- concentration of overlying 197 water at oligohaline site C were greater than at freshwater site A, and there was no difference in TPw 198 concentration among the study sites. In general, there were no within-study site differences between the 199 rhizosphere and bulk soil parameters; exceptions were for lower levels of pH, TS, and sand and higher levels 200 of TOC and clay in the rhizosphere at oligohaline site B and respectively higher and lower levels of silt and 201 sand in the rhizosphere at oligohaline site C (Table 1). 202 3.2. Soil P dynamics 203 There were some differences in soil P concentrations in the rhizosphere and bulk soils, where the rhizosphere 204 TP concentration was lower at freshwater site A than at oligohaline site B. The bulk soil IP concentration was 205 lower at the freshwater site than at oligohaline site B. The OP concentration was lower in rhizosphere and bulk 206 soils at the freshwater site than at the two oligohaline sites (B and C). The AP concentration in rhizosphere 207 soils was greater at oligohaline site C than at the freshwater site and oligohaline site B. The AP concentration 208 in bulk soils was greatest at oligohaline site B and lowest at the freshwater site (P < 0.05; Figs. 2a-d). Overall, 209 IP accounted for 66-89% of TP (Figs. 2a, b). Among the IP fractions, the concentrations of Fe-P in rhizosphere 210 and bulk soils were greater at the freshwater site than at the two oligohaline sites, whereas the concentration 10 211 of Ca-P in the two soil profiles was lower (Figs. 2e, g). The Al-P concentration in the rhizosphere was lower 212 at the freshwater site than at the two oligohaline sites and lower in the bulk soil at the freshwater site than at 213 oligohaline site C (Fig. 2f). The O-P concentrations in the rhizosphere and bulk soils were lower at the 214 freshwater site and oligohaline site B than at oligohaline site C (Fig. 2h). Al-P and O-P were the dominant 215 forms of IP (38 and 30%, respectively), followed by Fe-P (21%) and Ca-P (11%) at all sites (Fig. S1). There 216 were few within-site differences in the soil P concentrations between rhizosphere and bulk soils (Fig. 2). TP 217 and Fe-P at the freshwater site were respectively lower and greater in the rhizosphere than in the bulk soils, 218 while at oligohaline site B, the concentration of AP was lower in the rhizosphere than in bulk soils, but that of 219 Fe-P was greater. 220 3.3. Soil ALP activities 221 Soil ALP activity was 7-fold greater in the freshwater (0.7 µmol g-1·h-1) than in the oligohaline marshes (0.1 g- 1 222 ·h-1), but there were no within-study site differences in ALP activity between the rhizosphere and bulk soils 223 (Fig. 3). 224 3.4. Richness and alpha diversity of the phoD phosphatase gene 225 A total of 73,350 qualified sequences of the phoD phosphatase gene were recorded from the soil samples. The 226 richness and alpha diversity of the phoD phosphatase gene were greater in the rhizosphere than in the bulk 227 soils at site B, and overall, these values were greatest at site B and lowest at site A (Table 2). There was no 228 clear pattern in phoD gene diversity with the transition from freshwater to oligohaline environments. The PCoA 229 showed that the bacterial phoD gene communities at freshwater site A were more loosely clustered and distinct 230 from those at oligohaline sites B and C (Fig. 4). 231 3.5. phoD phosphatase gene community structure 232 Overall, the most abundant bacteria classes containing the phoD gene were the Alphaproteobacteria, 233 Betaproteobacteria and Gammaproteobacteria, accounting for 30-80% of total sequences (Fig. 5a). The 11 234 dominant genera in all samples were Pleomorphomonas, Streptomyces, Cupriavidus, Bradyrhizobium, and 235 Pseudomonas (relative abundance >1%) (Fig. 5b). Specifically, Streptomyces, Cupriavidus, and 236 Bradyrhizobium were the dominant genera at the freshwater marsh (relative abundance: 43, 19, and 9%, 237 respectively), and Pleomorphomonas, Streptomyces, and Bradyrhizobium were dominant at the oligohaline 238 sites (relative abundance: 52, 22, and 5%, respectively). 239 The relative abundance of the phoD-harboring bacterial shifted significantly along a freshwater- 240 oligohaline gradient, where the relative abundance of the Alphaproteobacteria class was greater at oligohaline 241 sites, but that of the Betaproteobacteria was greater at freshwater sites. The relative abundance of the 242 Pleomorphomonas genus was greater, whereas those of Streptomyces and Cupriavidus genus were lower, at 243 oligohaline sites. Hierarchical cluster analysis showed that bacterial communities formed three groups that 244 represented the three study sites (Fig. S2). However, there were no within-site differences between rhizosphere 245 and bulk soil communities. 246 3.6. Relationships between soil P and phoD gene communities with soil and overlying water variables 247 The concentrations of soil Al-P, Ca-P, O-P, and AP were positively correlated with soil pH, EC, TS and silt 248 content and negatively correlated with soil TOC and sand content. The soil Fe-P concentrations were 249 negatively correlated with pH, EC, and TS and positively correlated with soil TOC. The soil OP concentrations 250 were positively correlated with soil pH, EC and overlying water salinity (Fig. 6). The concentrations of soil 251 Ca-P and AP were positively associated with soil clay content. Soil Al-P, Ca-P, O-P, and AP were also found 252 to be positively related to overlying water salinity and SO42- concentration. 253 We found that the relative abundance of Pleomorphomonas was negatively associated with the contents 254 of TOC and sand and positively associated with other environmental variables; that of Streptomyces showed 255 the opposite trend (Fig. 6). The soil ALP activity was positively correlated with soil TOC and sand content and 256 negatively correlated with soil pH and EC, contents of TS, clay, and silt, and overlying water salinity and SO42-. 257 We examined the relationships between the AP concentration, phoD gene community, and ALP activity 258 (Fig. S3) and found that soil AP was positively correlated with the phoD gene diversity but negatively 12 259 correlated with the ALP activity. The soil AP was positively related to the relative abundance of 260 Pleomorphomonas and negatively related to the abundance of Streptomyces. The soil ALP activity was 261 negatively related to the bacterial phoD gene diversity and the relative abundance of Pleomorphomonas. 262 The first two axes of the RDA of the influence of biochemical variables (environmental variables, phoD 263 gene community, and ALP activity) accounted for 78.6% of the variation in P dynamics (P < 0.01) that 264 clustered into three groups (groups I, II, and III) representing the three study sites (A, B, and C, respectively) 265 (Fig. 7). The soil P dynamics in Group I (freshwater) were primarily influenced by the soil contents of TOC, 266 ALP activity, and the relative abundance of Streptomyces, while in Groups II and II (oligohaline), they were 267 primarily driven by overlying water salinity and SO42-, soil TS and pH, phoD gene diversity, and the relative 268 abundance of Pleomorphomonas. 269 4. Discussion 270 4.1. Soil P speciation responses to a freshwater-oligohaline gradient 271 Previous studies have indicated that P dynamics and associated speciation varied from freshwater to saltwater 272 due to shifts in soil physicochemical and microbial processes (Gireeshkumar et al. 2013; Paludan and Morris 273 1999). As expected, we observed higher of soil-water salinity, pH and SO42- content, lower soil TOC, and 274 greater associated soil AP concentration in the transition from freshwater to oligohaline sites (Table 1; Fig. 2), 275 supporting our first hypothesis that soil P availability is greater at oligohaline sites due to shifts in the P-related 276 physicochemical properties. It is possible that the salinity brought by seawater influences P dynamics through 277 changes to adsorption and desorption reactions triggered by an increase in ions in the sediment that compete 278 with phosphate ions (PO43-) for the sorption sites (Qu et al. 2018). This response was demonstrated by the 279 positive correlation between the AP concentration and overlying water salinity (Fig. 6). The abundance of 280 electron-accepting SO42- in the oligohaline marshes (Table 1) may enhance the release of P as a result of sulfate 13 281 reduction (SR), but it may also compete with PO43- for anion adsorption sites (Caraco et al. 1990), thereby 282 regulating P availability. However, we did not find a difference in soil IP concentration among the sites, even 283 though the OP levels were greater in the oligohaline sites than in freshwater (Figs. 2b, c), possibly as a result 284 of a slower capacity for P-mineralization in oligohaline soils than in freshwater, as indicated by the lower ALP 285 activity in the oligohaline soils (Fig. 3). 286 The concentrations of Al-P, Ca-P, and O-P in the IP fraction increased from freshwater to oligohaline sites 287 (Figs. 2f-h), while Fe-P markedly decreased (Fig. 2e), representing a shift from Fe-P to Ca-P and Al-P due to 288 the reduction of Fe along a freshwater-oligohaline gradient. These findings are consistent with the greater 289 availability of SO42- in oligohaline sites (Table 1). This greater availability promotes the formation of sulfide 290 and subsequent Fe(II) sulfides due to the high rates of SR, which increase the release of PO43- due to the 291 reduction of Fe(III)-bound-P, thus reducing Fe-P storage in the soil (Dierberg et al. 2011; Luo et al. 2019). It 292 is also likely that greater levels of alkalinity generated by SR in anoxic soils may inhibit P sorption onto iron 293 oxides in soils (Caraco et al. 1989). The shifts in soil physicochemical properties from freshwater to 294 oligohaline sites also affected the P fraction. We found that soil concentrations of Al-P, Ca-P, and O-P were 295 positively associated with silt content in the soil and negatively associated with sand content (Fig. 6), indicating 296 that soil texture is a driver of P dynamics in estuarine marshes. This finding may be explained by differences 297 in the substrate surface area and reactivity, as the greater surface area of silt particles provides more binding 298 sites for PO43- adsorption in the soil (Gireeshkumar et al. 2013). Our findings also indicate that the transition 299 from freshwater to an oligohaline environment produces more strongly occluded P in soil so that the long-term 300 P storage capacity in soil is enhanced with moderate increases in salinity. 301 We found that the IP fraction of soil P was primarily controlled by Al-P (38%), followed by O-P (30%), 302 Fe-P (21%), and Ca-P (11%) (Fig. S1). The Min River estuary is located in the subtropics, where the weathering 303 of parent rock is relatively strong under the warm, humid climate, resulting in high levels of organic matter 14 304 rich in Fe/Al oxide (Luo et al. 2014; Luo et al. 2019). The release of Fe and Al facilitates the enrichment and 305 migration of P as the adsorption carriers, thereby controlling the IP fraction. In this study, the average soil TP 306 concentration across study sites in the estuary (698 mg·kg-1) did not vary with the transition from freshwater 307 to oligohaline sites (Fig. 2a). In addition, the TP concentrations were similar to those recorded in estuaries in 308 Portugal (Coelho et al. (2004) and elsewhere in China (Jin et al. (2013) but lower than those recorded from 309 estuaries in the US (Jordan et al. 2008) and Sri Lanka (Gireeshkumar et al. 2013) (Table S2). Within-estuary 310 consistency in soil TP content may be the result of a combination of changing dominance of the actions of 311 various forms of P and their trade-offs because we found that while some P-fractions were more abundant at 312 oligohaline sites (AP and O-P), other, mostly moderately labile forms, were less abundant. Our finding that 313 soil TP, and its associated speciation, did not differ between rhizosphere and bulk soils was inconsistent with 314 previous studies (Hinsinger 2001; Luo et al. 2017). This inconsistency may be due to the periodic flooding 315 and associated variations in salinity that drive complex hydrological and nutrient cycling in estuarine 316 environments (Hu et al. 2018a), which partially offset possible underlying rhizosphere effects on soil P 317 dynamics. Moreover, the resuspension of sediment driven by tide and river runoff may have stimulated P 318 immobilization to change the transformation of soil P fractions (Labry et al. 2016). Thus, shifts in soil P 319 fractions from freshwater to oligohaline sites may be a consequence of the direct effects of salinity and the 320 indirect effects of altered soil/water physicochemical properties due to saltwater intrusion. 321 4.2. Effects of salinity on ALP activity and the phoD gene community 322 Greater rates of ALP activity were observed in the freshwater than the oligohaline sites, regardless of soil type 323 (rhizosphere/bulk) (Fig. 3). These findings corroborate a study by Jackson and Vallaire (2009) but are in 324 contrast to those by Morrissey et al. (2014) and Labry et al. (2016). It is known that pH plays a crucial role in 325 the regulation of phosphatase activity, and ALP activity is predominant in alkaline environments (Stout et al. 326 2014). However, and somewhat surprisingly, we did not observe this phenomenon in our study because the 15 327 soil ALP activity was negatively correlated with pH (Fig. 6). This absence of a positive effect of pH on ALP 328 activity may be due to the limited variation in pH value (6.4-7.4; Table 1) or the low concentrations of PO43- 329 that drive ALP activity in P-deficient environments (Fraser et al. 2015; Labry et al. 2016). In this study, soil 330 AP concentrations were greater in the oligohaline sites than the freshwater site, decreasing the requirement for 331 ALP synthesis to facilitate P uptake in microbes and plants, as indicated by the negative association between 332 ALP activity and soil AP (Fig. S3d). It should be noted that the ALP activity of this study refers to potential 333 phosphatase activity rather than actual soil phosphatase activity because the pH has been modified, which 334 allows it to be compared to other studies (Chen et al. 2019a; b; Fraser et al. 2015; Fraser et al. 2017). 335 Although phoD genes have been investigated extensively under various environmental conditions (Fraser 336 et al. 2015; Ragot et al. 2015), the influence of variations with the transition from freshwater to oligohaline 337 environments remains unclear. Here, the alpha-diversity of bacterial phoD genes did not vary with the 338 transition (Table 2), possibly due to the relatively narrow range of salinity (0.03–2.92 ppt) at the study sites, 339 which may not have been sufficient to result in differences in diversity. In contrast, the soil bacterial phoD 340 gene community compositions differed between the freshwater and oligohaline sites (Fig. 5). Overall, 341 Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria were the most abundant bacterial classes 342 in the study sites (>59% of total bacteria; Fig. 5), supporting other studies of various types of soil (Lagos et al. 343 2016; Luo et al. 2017; Tan et al. 2013) and indicating that the phoD-harboring bacteria community composition 344 is stable across different environments. The greater relative abundance of Pleomorphomonas and the lower 345 abundance of Streptomyces at oligohaline sites relative to freshwater (Fig. 5) may reflect their responses and 346 adaptations to nutrient availability and salinity fluctuations (Spohn et al. 2015). 347 Our analysis of the community composition of phoD-harboring bacterial genera showed that the relative 348 abundance levels of some genera, such as Pleomorphomonas, were much greater at oligohaline sites, while 349 that of Streptomyces was lower (Fig. 5). In addition to the direct influences of soil osmotic potential and water 16 350 stress caused by salinity (Chambers et al. 2013), changes in soil pH, TOC, and texture with the transition from 351 freshwater to oligohaline environments may have affected phoD gene community composition. For example, 352 we found that soil TOC was positively associated with the relative abundance of Acidobacteria (Fig. 6), likely 353 because substrates rich in C favor growth of some phoD-harboring species, leading to increases in the 354 abundance of the phoD genes and ALP activity (Luo et al. 2017). It is important to note that soil texture, which 355 is correlated with variations in bacterial community composition, may affect the adsorption of chemicals by 356 distinct microbial communities due to differences in surface properties and microenvironments (Hemkemeyer 357 et al. 2015). Overall, these results support our second hypothesis that the transition from freshwater to 358 oligohaline environments shapes bacterial phoD gene communities and influences ALP activity. 359 4.3. Linking ALP activity and phoD gene community structure with P availability 360 Soil ALP activity has been regarded as an indicator of changes in organic P mineralization and bacterial phoD 361 gene abundance in soils (Acuña et al. 2016; Huang and Morris 2003). The negative relation between the soil 362 ALP activity and AP concentration (Fig. S3d) provides additional evidence that ALP activity is only induced 363 at low P levels due to the inhibited synthesis of phosphatases at high AP concentrations (Acuña et al. 2016; 364 Fraser et al. 2015). The differences in relative abundance of the dominant phoD gene community may play a 365 key role in P availability because soil AP concentration was positively associated with the relative abundance 366 of Pleomorphomonas (Fig. S3b) but negatively associated with that of Streptomyces (Fig. S3c). Both of these 367 dominant genera are important for P solubilization and mineralization through the production of P-hydrolyzing 368 enzymes (Acuña et al. 2016; Ragot et al. 2015). 369 In general, potential ALP activity may have been directly regulated by the phoD phosphatase gene 370 community due to the greater production efficiency of extracellular alkaline phosphatases (Luo et al. 2017). 371 The high taxonomic diversity of the phoD gene community renders it better able to tolerate changes in salinity 372 and potentially affects ALP production (Fraser et al. 2015). Moreover, the responses of the phoD gene 17 373 community to the freshwater-oligohaline transition might be partially depend on P availability (Fraser et al. 374 2015), and high levels of P might inhibit phoD gene expression (Vershinina and Znamenskaya 2002). However, 375 this pattern only was observed in the phoD gene diversity, which showed a negative correlation with the ALP 376 activity. Measuring the gene and transcript levels with more universal phoD primers is thus warranted (Ragot 377 et al. 2015) and will provide a full understanding of ALP production and phoD gene diversity along an 378 estuarine freshwater-oligohaline gradient. 379 4.4. Implications and uncertainties 380 In summary, our data clearly showed the variations of soil P fractions, environmental parameters, and phoD 381 gene communities among the study sites, where the levels of salinity and pH were greater, the soil texture was 382 finer, and the contents of O-P/Al-P were greater at the oligohaline sites, which were associated with greater 383 levels of soluble and available P. In contrast, freshwater sites were characterized by coarser textured soils, 384 higher levels of TOC and Fe-P, lower levels of O-P and AP, and higher levels of ALP activity, despite the low 385 pH levels. These results indicate higher microbial and general biological activity, lower P retention capacity 386 and greater biological effort required for P uptake in freshwater conditions. Furthermore, the overall PCA 387 analysis (Fig. 8) indicated that the most diverse bacterial community and the higher, more evenly distributed 388 stocks of P among the rhizospheres and bulk soils occurred at the moderately saline site (site B). Thus, we 389 conclude that with the transition from freshwater to oligohaline sites, the associated increases in soil-water 390 salinity and SO42- and decreases in general microbial activity are key drivers of P fractions and availability, 391 which indicates that greater potential P losses may reduce nutrient availability for estuarine plants and microbes 392 under longer-term, climate change-mediated rises in sea levels. 393 Although these patterns have important implications for our understanding of wetland geochemical 394 processes that drive P cycling, some uncertainties and future works must be carefully considered. First, 395 hydrological conditions, such as periodic tidal processes and seawater intrusion across the estuarine tidal 18 396 marshes, might have significant effects on P speciation and the phoD gene community. Therefore, increasing 397 the consideration of tidal effects can contribute to more accurate estimates of soil P dynamics and microbial 398 community along a freshwater-oligohaline gradient. Second, our results support the conclusion of Tan et al. 399 (2013) and Ragot et al. (2015) that the ALPS primers seems to have an amplification bias caused by the primers 400 specificity, for example, numerous sequences were assigned to Alphaproteobacteria, which suggested this 401 primer set could be biased toward Alphaproteobacteria rather than the real distribution of phoD. Therefore, 402 although the present study provided some insights into the phoD bacterial communities along an estuarine 403 freshwater-oligohaline gradient, the interpretation and comparison of our results must be conducted with 404 some caution. Future research should consider newly designed primers based on metagenome databases 405 (Ragot et al. 2015), which probably provide better coverage of the phoD diversity. 406 5. Conclusions 407 Our results suggest that the transition from freshwater to an oligohaline environment drives the increases in 408 soil-water salinity, pH and SO42- content, a decrease in soil TOC, the associated increases in concentrations of 409 AP, Al-P, Ca-P, and O-P, and a decrease in Fe-P concentrations. These findings support our first hypothesis 410 that soil P availability is greater at oligohaline sites due to shifts in P-related physicochemical properties. These 411 findings also highlight the role of salinity as a substantial factor in P availability, where increases in salinity 412 and the associated changes in soil-water physicochemical properties (SO42-, pH, and TOC) as a consequence 413 of saltwater intrusion may exacerbate losses of P to water, eventually leading to losses in P-fertility for plants 414 and microbes. The lower levels of soil ALP activity and the altered composition of phoD gene communities at 415 oligohaline sites indicate that even small increases in the salinity levels and the associated shifts in 416 environmental factors can be caused by saltwater intrusion and may regulate microbial activity and reshape 417 the phoD gene community composition in estuarine tidal marshes, supporting our second hypothesis. Our 19 418 results showed that the moderately saline study site (site B) was characterized by the most diverse bacterial 419 community and higher and more evenly distributed stocks of P among the rhizosphere and bulk soils. Our 420 results increase our understanding of the main processes and mechanisms involved in the estuarine P dynamics 421 and phoD phosphatase gene communities. However, detailed studies and analyses of spatiotemporal coupling 422 and tidal action that control the available soil P pools are required. 423 Acknowledgments 424 This research was jointly funded by the National Science Foundation of China (41801062), the National Key 425 R&D Program of China (2016YFC0502901), the China Postdoctoral Science Foundation (2018M630731, 426 2019T120556), a European Research Council Synergy Grant (ERC-SyG-2013-610028 IMBALANCE-P), the 427 Spanish Government (CGL2016-79835-P), and the Catalan Government (SGR 2017-1005). The authors 428 especially thank Chao Jiang, Yinren Tao, Yuxue Zhang, and Tingyu Zhang for their help with field sampling 429 and laboratory analyses. 430 References 431 432 433 Acuña JJ, Durán P, Lagos LM, Ogram A, de la Luz Mora M and Jorquera MA (2016) Bacterial alkaline 434 phosphomonoesterase in the rhizospheres of plants grown in Chilean extreme environments. Biology and 435 Fertility of Soils 52: 763-773. 436 Bai JH, Ye XF, Jia J, Zhang GL, Zhao QQ, Cui BS and Liu XH (2017) Phosphorus sorption-desorption and effects of 437 temperature, pH and salinity on phosphorus sorption in marsh soils from coastal wetlands with different flooding 438 conditions. Chemosphere 188: 677-688. 439 Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman, F D and Costello EK (2010) QIIME allows analysis of 440 high-throughput community sequencing data. Nature Methods 7: 335-336. 441 Caraco N, Cole J and Likens GE (1990) A comparison of phosphorus immobilization in sediments of freshwater and 442 coastal marine systems. Biogeochemistry 9: 277-290. 443 Caraco NF, Cole JJ and Likens GE (1989) Evidence for sulphate-controlled phosphorus release from sediments of aquatic 444 systems. Nature 341: 316–318. 445 Chambers LG, Osborne TZ and Reddy RK (2013) Effect of salinity-altering pulsing events on soil organic carbon loss 446 along an intertidal wetland gradient: a laboratory experiment. Biogeochemistry 115: 363-383. 20 447 Chang SC and Jackson ML (1957) Fractionation of soil phosphorus. Soil Science 84: 133–144. 448 Chen X, Jiang N, Condron LM, Dunfield KE, Chen Z, Wang J and Chen L (2019a) Impact of long-term phosphorus 449 fertilizer inputs on bacterial phoD gene community in a maize field, Northeast China. Science of the Total 450 Environment 669: 1011-1018. 451 Chen X, Jiang N, Condron LM, Dunfield KE, Chen Z, Wang J and Chen L (2019b) Soil alkaline phosphatase activity and 452 bacterial phoD gene abundance and diversity under long-term nitrogen and manure inputs. Geoderma 349: 36- 453 44. 454 Coelho JP, Flindt MR, Jensen HS, Lillebø AI and Pardal MA (2004) Phosphorus speciation and availability in intertidal 455 sediments of a temperate estuary: relation to eutrophication and annual P-fluxes. Estuarine, Coastal and Shelf 456 Science 61: 583-590. 457 Dierberg FE, Debusk TA, Larson NR, Kharbanda MD, Chan N and Gabriel MC (2011) Effects of sulfate amendments on 458 mineralization and phosphorus release from South Florida (USA) wetland soils under anaerobic conditions. Soil 459 Biology and Biochemistry 43: 31-45. 460 Fraser T, Lynch DH, Entz MH and Dunfield KE (2015) Linking alkaline phosphatase activity with bacterial phoD gene 461 abundance in soil from a long-term management trial. Geoderma 257: 115-122. 462 Fraser TD, Lynch DH, Gaiero J, Khosla K and Dunfield KE (2017) Quantification of bacterial non-specific acid (phoC) 463 and alkaline (phoD) phosphatase genes in bulk and rhizosphere soil from organically managed soybean fields. 464 Applied Soil Ecology 111: 48-56. 465 Gireeshkumar TR, Deepulal PM and Chandramohanakumar N (2013) Phosphorous speciation in surface sediments of the 466 Cochin estuary. Environmental Monitoring and Assessment 185: 2535-2552. 467 Hartikainen H (1979) Phosphorus and its reactions in terrestrial soils and lake sediments. Journal of the Scientific 468 Agricultural Society of Finland 53: 537–624. 469 Hartzell JL, Jordan TE and Cornwell JC (2010) Phosphorus burial in sediments along the salinity gradient of the Patuxent 470 River, a subestuary of the Chesapeake Bay (USA). Estuaries and Coasts 33: 92-106. 471 Hartzell JL, Jordan TE and Cornwell JC (2017) Phosphorus sequestration in sediments along the salinity gradients of 472 Chesapeake Bay subestuaries. Estuaries and Coasts 40: 1607-1625. 473 Hemkemeyer M, Christensen BT, Martens R and Tebbe CC (2015) Soil particle size fractions harbour distinct microbial 474 communities and differ in potential for microbial mineralisation of organic pollutants. Soil Biology and 475 Biochemistry 90: 255-265. 476 Hinsinger P (2001) Bioavailability of soil inorganic P in the rhizosphere as affected by root-induced chemical changes: a 477 review. Plant and Soil 237: 173-195. 478 Hu MJ, Peñuelas J, Sardans J, Huang JF, Li DD and Tong C (2019) Effects of nitrogen loading on emission of carbon 479 gases from estuarine tidal marshes with varying salinity. Science of the Total Environment 667: 648-657. 480 Hu MJ, Peñuelas J, Sardans J, Sun ZG, Wilson BJ, Huang JF, Zhu QL and Tong C (2018a) Stoichiometry patterns of plant 481 organ N and P in coastal herbaceous wetlands along the East China Sea: implications for biogeochemical niche. 482 Plant and Soil 431: 273–288. 483 Hu YJ, Xia YH, Sun Q, Liu KP, Chen XB, Ge TD, Zhu BL, Zhu ZK, Zhang ZH and Su YR (2018b) Effects of long-term 484 fertilization on phoD-harboring bacterial community in Karst soils. Science of the Total Environment 628: 53- 485 63. 486 Huang B, Yan D, Ouyang C, Zhang D, Zhu J, Liu J, Li Y, Wang Q, Han Q and Cao A (2019) Chloropicrin fumigation 487 alters the soil phosphorus and the composition of the encoding alkaline phosphatase PhoD gene microbial 488 community. Science of the Total Environment: 135080. 489 Huang J, Cao C, Yan C, Liu J, Hu Q and Guan W (2017) Impacts of silver nanoparticles on the nutrient removal and 490 functional bacterial community in vertical subsurface flow constructed wetlands. Bioresour Technol 243: 1216- 491 1226. 21 492 Huang XQ and Morris JT (2003) Trends in phosphatase activity along a successional gradient of tidal freshwater marshes 493 on the Cooper River South Carolina. Estuaries 26: 1281-1290. 494 Jackson CR and Vallaire SC (2009) Effects of salinity and nutrients on microbial assemblages in Louisiana wetland 495 sediments. Wetlands 29: 277-287. 496 Jin XD, He YL, Kirumba G, Hassan Y and Li JB (2013) Phosphorus fractions and phosphate sorption-release 497 characteristics of the sediment in the Yangtze River estuary reservoir. Ecological Engineering 55: 62-66. 498 Jordan TE, Cornwell JC, Boynton WR and Anderson JT (2008) Changes in phosphorus biogeochemistry along an 499 estuarine salinity gradient: The iron conveyer belt. Limnology and Oceanography 53: 172-184. 500 Kirwan ML and Megonigal JP (2013) Tidal wetland stability in the face of human impacts and sea-level rise. Nature 504: 501 53-60. 502 Laakso J, Uusitalo R and Yli-Halla M (2016) Phosphorus speciation in agricultural catchment soils and in fresh and dried 503 sediments of five constructed wetlands. Geoderma 271: 18-26. 504 Labry C, Delmas D, Youenou A, Quere J, Leynaert A, Fraisse S, Raimonet M and Ragueneau O (2016) High alkaline 505 phosphatase activity in phosphate replete waters: The case of two macrotidal estuaries. Limnology and 506 Oceanography 61: 1513-1529. 507 Lagos LM, Acuña JJ, Maruyama F, Ogram A, de la Luz Mora M and Jorquera MA (2016) Effect of phosphorus addition 508 on total and alkaline phosphomonoesterase-harboring bacterial populations in ryegrass rhizosphere microsites. 509 Biology and Fertility of Soils 52: 1007-1019. 510 Lin P and Guo LD (2016) Dynamic changes in the abundance and chemical speciation of dissolved and particulate 511 phosphorus across the river‐lake interface in southwest Lake Michigan. Limnology and Oceanography 61: 512 771-789. 513 Luo GW, Ling N, Nannipieri P, Chen H, Raza W, Wang M, Guo SW and Shen QR (2017) Long-term fertilisation regimes 514 affect the composition of the alkaline phosphomonoesterase encoding microbial community of a vertisol and its 515 derivative soil fractions. Biology and Fertility of Soils 53: 375-388. 516 Luo M, Zeng CS, Tong C, Huang JF, Yu Q, Guo YB and Wang SH (2014) Abundance and speciation of iron across a 517 subtropical tidal marsh of the Min River Estuary in the East China Sea. Applied Geochemistry 45: 1-13. 518 Luo M, Zhu WF, Fang HJ, Liu YX, Duan X, Wu J and Tong C (2019) Anaerobic organic carbon mineralization in tidal 519 wetlands along a low-level salinity gradient of a subtropical estuary: Rates, pathways, and controls. Geoderma 520 337: 1245-1257. 521 Matsuoka S, Fujinaga S, Kobayashi Y, Hobara S and Osono T (2019) Bacterial 16S rDNA and alkaline phosphatase gene 522 diversity in soil applied with composted aquatic plants. Limnology: 1-8. 523 Morrissey EM, Gillespie JL, Morina JC and Franklin RB (2014) Salinity affects microbial activity and soil organic matter 524 content in tidal wetlands. Global Change Biology 20: 1351-1362. 525 Odum WE (1988) Comparative ecology of tidal freshwater and salt marshes. Annual Review of Ecology and Systematics 526 19: 147-176. 527 Olsen SR and Sommers LE (1982) Phosphorus. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil 528 Analysis, 2nd ed. Agronomy Series No. 9, Part 2 Soil Science Society of America, Inc., Madison, WI, pp. 403– 529 430. 530 Paludan C and Morris JT (1999) Distribution and speciation of phosphorus along a salinity gradient in intertidal marsh 531 sediments. Biogeochemistry 45: 197-221. 532 Qu FZ, Shao HB, Meng L, Yu JB, Xia JB, Sun JK and Li YZ (2018) Forms and vertical distributions of soil phosphorus 533 in newly formed coastal wetlands in the Yellow River Delta estuary. Land Degradation and Development 29: 534 4219-4226. 535 Ragot SA, Kertesz MA and Bünemann EK (2015) phoD alkaline phosphatase gene diversity in soil. Applied and 536 Environmental Microbiology 81: 7281-7289. 22 537 Rahutomo S, Kovar JL and Thompson ML (2019) Phosphorus transformations in stream bank sediments in Iowa, USA, 538 at varying redox potentials. Journal of Soils and Sediments 19: 1029-1039. 539 Ramette A (2007) Multivariate analyses in microbial ecology. FEMS Microbiology Ecology 62: 142-160. 540 Ray R, Majumder N, Chowdhury C, Das S and Jana TK (2018) Phosphorus budget of the Sundarban mangrove ecosystem: 541 Box model approach. Estuaries and Coasts 41: 1036-1049. 542 Ruban V, López-Sánchez JF, Pardo P, Rauret G, Muntau H and Quevauviller P (1999) Selection and evaluation of 543 sequential extraction procedures for the determination of phosphorus forms in lake sediment. Journal of 544 Environmental Monitoring 1: 51-56. 545 Ruban V, López-Sánchez JF, Pardo P, Rauret G, Muntau H and Quevauviller P (2001) Development of a harmonised 546 phosphorus extraction procedure and certification of a sediment reference material. Journal of Environmental 547 Monitoring 3: 121-125. 548 Sakurai M, Wasaki J, Tomizawa Y, Shinano T and Osaki M (2008) Analysis of bacterial communities on alkaline 549 phosphatase genes in soil supplied with organic matter. Soil Science and Plant Nutrition 54: 62-71. 550 Spohn M, Treichel NS, Cormann M, Schloter M and Fischer D (2015) Distribution of phosphatase activity and various 551 bacterial phyla in the rhizosphere of Hordeum vulgare L. depending on P availability. Soil Biology and 552 Biochemistry 89: 44-51. 553 Stout LM, Joshi SR, Kana TM and Jaisi DP (2014) Microbial activities and phosphorus cycling: An application of oxygen 554 isotope ratios in phosphate. Geochimica Et Cosmochimica Acta 138: 101-116. 555 Sun Q, Qiu H, Hu Y, Wei X, Chen X, Ge T, Wu J and Su Y (2019) Cellulose and lignin regulate partitioning of soil 556 phosphorus fractions and alkaline phosphomonoesterase encoding bacterial community in phosphorus-deficient 557 soils. Biology and Fertility of Soils 55: 31-42. 558 Tabatabai MA and Bremner JM (1969) Use of p-nitrophenyl phosphate for assay of soil phosphatase activity. Soil Biology 559 and Biochemistry 1: 301-307. 560 Tan H, Barret M, Mooij MJ, Rice O, Morrissey JP, Dobson A, Griffiths B and O’Gara F (2013) Long-term phosphorus 561 fertilisation increased the diversity of the total bacterial community and the phoD phosphorus mineraliser group 562 in pasture soils. Biology and Fertility of Soils 49: 661-672. 563 Tong C, Cadilloquiroz H, Zeng ZH, She CX, Yang P and Huang JF (2017) Changes of community structure and abundance 564 of methanogens in soils along a freshwater-brackish water gradient in subtropical estuarine marshes. Geoderma 565 299: 101-110. 566 Tong C, Wang C, Huang JF, Wang WQ, Yan E, Liao J and Yao C (2014) Ecosystem respiration does not differ before and 567 after tidal inundation in brackish marshes of the Min River Estuary, Southeast China. Wetlands 34: 225-233. 568 Valdespino-Castillo PM, Alcántara-Hernández RJ, Alcocer J, Merino-Ibarra M, Macek M and Falcón LI (2014) Alkaline 569 phosphatases in microbialites and bacterioplankton from Alchichica soda lake, Mexico. FEMS Microbiology 570 Ecology 90: 504-519. 571 Vershinina O and Znamenskaya L (2002) The Pho regulons of bacteria. Microbiology 71: 497-511. 572 Wan W, Wang Y, Tan J, Qin Y, Zuo W, Wu H, He H and He D (2019) Alkaline phosphatase-harboring bacterial community 573 and multiple enzyme activity contribute to phosphorus transformation during vegetable waste and chicken 574 manure composting. Bioresource Technology: 122406. 575 Wei X, Hu Y, Razavi BS, Zhou J, Shen J, Nannipieri P, Wu J and Ge T (2019) Rare taxa of alkaline phosphomonoesterase- 576 harboring microorganisms mediate soil phosphorus mineralization. Soil Biology and Biochemistry 131: 62-70. 577 Weston NB, Neubauer SC, Velinsky DJ and Vile MA (2014) Net ecosystem carbon exchange and the greenhouse gas 578 balance of tidal marshes along an estuarine salinity gradient. Biogeochemistry 120: 163-189. 579 White JR, Nagarajan N and Pop M (2009) Statistical methods for detecting differentially abundant features in clinical 580 metagenomic samples. PLoS Computational Biology 5: e1000352. 581 Zhang WL, Zeng CS, Tong C, Zhai SJ, Lin X and Gao DZ (2015) Spatial distribution of phosphorus speciation in marsh 23 582 sediments along a hydrologic gradient in a subtropical estuarine wetland, China. Estuarine Coastal and Shelf 583 Science 154: 30-38. 584 24