Kun Liu ,Xin LuoJiu Jimmy Jio*,Ji-ong Gu,Rmon Arven

a Department of Earth Sciences, Faculty of Science, The University of Hong Kong, Hong Kong 999077, China

b Department of Ecology and Biodiversity, School of Biological Sciences, The University of Hong Kong, Hong Kong 999077, China

c Department of Earth and Environmental Sciences, University of Waterloo, Waterloo N2L 3G1 , Canada

d China Institute of Geo-Environment Monitoring, China Geological Survey, Beijing 100081, China

Keywords:Anammox Ammonium oxidizing archaea (AOA)Ammonium oxidation bacteria (AOB)Aquitard Groundwater Hydrogeological survey engineering Pearl River Delta China

ABSTRACT A mmonia-oxidizing archaea (AOA),ammonia-oxidizing bacteria (AOB),and anaerobic ammoniaoxidation (anammox) bacteria are very important contributors to nitrogen cycling in natural environments.Functional gene abundances of these microbes were believed to be well relevant to N-cycling in groundwater systems,especially in the Pearl River Delta (PRD) groundwater with unique high intrinsic ammonia concentrations.In this research,20 sediment samples from two in the PRD were collected for porewater chemistry analysis and quantification of N-cycling related genes,including archaeal and bacterial amoA gene and anammox 16S ribosomal Ribonucleic Acid (rRNA) gene.Quantitative Polymerase Chain Reaction (qPCR) results showed that gene abundances of AOA,AOB,and anammox bacteria ranged from 3.13×105 to 3.21×107,1.83×104 to 2.74×106,and 9.27×104 to 8.96×106 copies/g in the sediment of the groundwater system,respectively.Anammox bacteria and AOA dominated in aquitards and aquifers,respectively,meanwhile,the aquitard-aquifer interfaces were demonstrated as ammonium-oxidizing hotspots in the aspect of gene numbers.Gene abundances of nitrifiers were analyzed with geochemistry profiles.Correlations between gene numbers and environmental variables indicated that the gene abundances were impacted by hydrogeological conditions,and microbial-derived ammonium loss was dominated by AOA in the northwest PRD and by anammox bacteria in the southeast PRD.

1.Introduction

Nitrification is a central biochemical process of the global nitrogen cycling,and its first and rate-limiting step,i.e.ammonia oxidation,was traditionally believed only attributed to aerobic ammonium oxidation bacteria (AOB) (Koops HP et al.,2006; Kowalchuk GA and Stephen JR,2001),functioned by theα-subunit of ammonium monooxygenase (amoA)(Rotthauwe JH et al.,1997).Then this view was changed by the discovery of ammonium oxidizing archaea (AOA)(Könneke M et al.,2005; Treusch AH et al.,2005; Venter JC et al.,2004),which is now identified as Thaumarchaeota,sister archaeal group to the Crenarchaeota (Brochier-Armanet C et al.,2008; Spang A et al.,2010).AOA also haveamoAgenes and can process ammonium oxidization under lowoxygen conditions in nitrogen and carbon cycles,oxidizing NH4+to NO2−(Beman JM et al.,2008; Francis CA et al.,2005,2007; Karner MB et al.,2001; Lam P et al.,2007;Molina V et al.,2010).Indicated by either cell numbers oramoAgene numbers,the abundance of AOA outnumbered AOB in marine and soil-like environments (Cao H et al.,2013; Karner MB et al.,2001; Leininger S et al.,2006;Schleper C,2010),with several exceptions in estuarine environments (Caffrey JM et al.,2007; Mosier AC and Francis CA,2008; Santoro AE et al.,2008).However,controversy continues with new evidence that though AOA has an advantage in numbers,15N-labeled NH4+amendment experiment in soils suggested the nitrification process was dominated by AOB (Di HJ et al.,2009; Jia Z and Conrad R,2009).

Meanwhile,a novel ammonium oxidation pathway,anaerobic ammonium oxidation (anammox),had beendiscovered in bioreactors in 1995 (Mulder A et al.,1995; Van de Graaf AA et al.,1995).This reaction combines nitrite and ammonium to form nitrogen gas under anoxic or suboxic conditions,and the N2gas finally leaks out of the reaction system (Kartal B et al.,2011; Kuenen JG,2008).Later,anammox bacteria have been discovered,identified,and quantified in various natural environments,including marine oxygen minimum zones (Füssel J et al.,2011; Pitcher A et al.,2011),deep-sea sediments (Kuypers MMM et al.,2003;Prokopenko MG et al.,2013),coastal estuaries (Dale OR et al.,2009; Trimmer M et al.,2003),freshwater lakes and rivers(Schubert CJ et al.,2006; Zhu GB et al.,2013),and contaminated groundwater (Clark L et al.,2008; Moore TA et al.,2011).N2production attributed to anammox in marine and soils may be up to 67% (Dalsgaard T et al.,2005) and 37%(Zhu GB et al.,2011) of the total nitrogen loss of those systems,respectively,indicating anammox is a nonignorable process in ether marine or terrestrial systems.The coexistence of AOA,AOB and anammox bacteria has been investigated in different kinds of environments (Erguder TH et al.,2009; Li M et al.,2011; Zhu GB et al.,2013),while there is an obvious lack of knowledge on links and feedback mechanism between nitrifiers abundances and hydrogeological conditions of natural coastal groundwater systems.

Coastal areas ecotones between terrestrial ecosystem and marine ecosystem,and the groundwater systems of coastal areas are always hydraulically and biochemically complicated(Alley WM et al.,2002).By definition,a coastal delta begins to form when rivers flow into the ocean,and this process is largely controlled by sea-level change (Blum MD and Roberts HH,2009; Blum MD and Törnqvist TE,2000; Lambeck K and Chappell J,2001).Historical sea-level changes and related sea transgression and regression cycles lead to sedimentary sequences of aquifer and aquitard.Gravels and coarse sands compose productive aquifers with sound hydraulic conductivity while silt and clay make good aquitards with poor permeability.Fresh recharging groundwater,paleo water,intrusive seawater interact mutually by advection and diffusion (Reilly TE and Goodman AS,1985),forming hydrogeological interfaces where nitrogen cycling is often elevated (McClain ME et al.,2003; Mcmahon PB and Chapelle FH,1991; Santoro AE,2010; Zhu GB et al.,2013; Zhan J et al.,2018; Zhou NQ et al.,2019).Interfaces between different hydrogeological units were found as hot spots for geomicrobial activities in other studies,for example,microbial nitrogen cycling at the saltwater-freshwater interface under “subterranean estuaries ” of coastal areas(Santoro AE,2010),and anammox at the land-freshwater interfaces under the riparian zone of Baiyangdian Lake (Zhu GB et al.,2013).In the Pearl River Delta (PRD) sediment,bacteria at the aquifer-aquitard interface showed higher diversity and richness in community structure (Liu K et al.,2011,2014).Chemical gradients among different hydrogeological layers,such as organic content,NOx−concentration,salinity,and pH,are key factors influencing abundances of AOA,AOB and anammox bacteria (Dalsgaard T and Thamdrup B,2002; Nicol GW et al.,2008; Rysgaard S et al.,1999; Schleper C,2010).

The Pearl River System is China’s third longest river after the Yangtze River and Yellow River,and the PRD,which is located in the subtropical region in Guangdong Province,is one of the most fertile areas in Southern China (Fig.1).Groundwater in the PRD has an abnormally high concentration (up to 390 mg/L) of naturally occurred ammonium (Jiao JJ et al.,2010),which was generated from the decomposition of organic matter in the organic-rich sediment in the adjacent aquitards,and released into aquifers by diffusion (Jiao JJ et al.,2010; Wang Y and Jiao JJ,2012a).Nitrifiers communities and N-cycling related genes abundances in this area have been comprehensively studied in river water,riparian soil,and surface sediment of the Pearl River (Jin T et al.,2011; Li M et al.,2013; Liu ZH et al.,2011b; Sun W et al.,2013,2014a,2014b; Wang S et al.,2012),however only a few reports had been conducted in the coastal groundwater system,which demonstrated that microbial communities in the deltaic sediment were largely controlled by the geological conditions,and AOB outnumbered AOA inamoAgene abundance in some aquitard sediment of the PRD (Lee KH,2010).

Hence,the authors chose the PRD,a coastal delta next to the South China Sea,as the study site to investigate abundances of AOA,AOB and anammox bacteria by Quantitative Polymerase Chain Reaction (qPCR)quantification of gene markers,and to explore the relationship between gene abundance and geological conditions in the groundwater environment of the PRD.

2.Materials and methods

2.1.Sediment sampling

The multilayer aquifer system in the PRD mainly has two aquifers and two aquitards interposed.Formation history and stratigraphy of the PRD had been abundantly described in previous geological research (Li PR and Qiao PN,1982; Wu C et al.,2007; Yim WWS et al.,2008; Zong Y et al.,2009b;Zong YQ et al.,2012).Based on lithological characteristics,the whole PRD aquifer-aquitard system can be distinguished into four units: M1 aquitard,T1 aquifer,M2 aquitard,and T2 aquifer (Fig.1c).The T2 layer,the basal aquifer overlying on bedrocks,is composed of sand and gravel.The M2 layer was formed in the Pleistocene mainly composed of silt and clay.The T1 layer,which could be either sandy fluvial deposits or clayey silt weathered from the M2 sediment,was formed during the last glacial period (Yim WWS et al.,2008),and could perform as a local intermediate aquifer when it is sandy.The major layer of the system,aquitard M1,was formed during the Holocene with fine silt and clayey materials.The thick aquitard M1 (over 10 m on average) is characterized by very low hydraulic conductivity of 2.2×10−10-8.2×10−9m/s(Yang L et al.,2014),which prevents underneath aquifers and aquitard M2 from most of the input of modern precipitation and oxygen,and the regional horizontal groundwater recharge is also very slow due to the gentle topographic slope of the PRD.

Two locations,Beijiao Town and Minzhong Town in the PRD area were selected to represent an inland near-mountain area and a coastal estuarine area,respectively.At these two locations,some residual seawater of the last transgression still existed in the aquifer and gradually mixed with freshwater recharged from piedmont areas.Beijiao is very close to the mountainous area in the northwest of the PRD,so the residual Holocene seawater in Beijiao strata was largely replaced by fresh water,while Minzhong groundwater still has very high salinity due to its location at the coast.However,groundwater chemistry in different hydrogeological layers distinctively varied from each other in both Beijiao and Minzhong.Two boreholes,BJP8 and MZ4P9,were drilled in Beijiao and Minzhong to take sediment samples,respectively.To avoid contamination,percussion drilling was adopted with a stainless steel sampler assembled at the lower end of the drilling rod.Core sediment was collected and sealed in sterile plastic bags after being pushed out of the sampler.Core samples were then transferred to the laboratory in an icebox immediately after sampling.In the laboratory,peripheral sediment of the core was peeled and discarded,while the inner core sediment was subsampled in a clean bench using an aseptic technique.Based on the lithology and chronology analysis,a total of 20 samples at different depths and layers were selected for deoxyribonucleic acid (DNA) extraction and subsequent qPCR (Table 1).The rest of the sediment was used for porewater extraction for porewater chemical analyses.

Fig.1.a-Location of Pearl River Delta plain and estuary; b-research section in the PRD; c-simplified geological profile along section HJ1-MZ4P9.Lithological descriptions of BJP8 and MZ4P9 were based on this study,while borehole HJ1 and DL1 were based on Zong Y et al.(2009a).

For sediment chronology,optically stimulated luminescence (OSL) dating method was adopted to estimate the time when sediment grains were last exposed to daylight(Murray AS and Olley JM,2002).Sediment samples for dating were wrapped with aluminum foil and transferred to Luminescence Dating Laboratory,the University of Hong Kong.Dating was performed with TL/OSL reader (Risø-TLDA-10) following procedures described by Bøtter-Jensen L and Duller G (1992).The dating results,as showing in Fig.2,were then used for stratigraphic subdivision.

2.2.DNA extraction and purification

Genomic DNA was extracted from sediment samples using fast DNA®SPIN Kit for Soil (MP Biomedicals,LLC)according to the manufacturer ’s protocol with adaptions specifically designed for the sediment of high salinity.Concentration and purity of extracted crude DNA were determined by Eppendorf®Biophotometer (Netheler & Hinz,Hamburg,Germany) by absorptions of lightwave at 230 nm,260 nm,and 280 nm (Yeates C et al.,1998).High salinity and organic-rich sediment or water often yield DNA of low concentration and poor quality,thus purification of DNA was an essential step for the PRD sediment samples.

Thawed sediment samples (approximately 0.5 g,wet weight) were transferred into a sterile 1.5 mL Lysing Matrix tube (MP Bio).Then,1 mL Sodium Phosphate Buffer and 200 μL MT Buffer (MP Bio) were added into each of the lysing tubes.For thorough homogenization,the tubes were secured horizontally on a flat-bed vortex pad with tape,and then vortexed at maximum speed for 20 mins.After being centrifuged at 14000×g for 15 mins at room temperature,the supernatant of lysing tubes was transferred to a clean 2 mL microcentrifuge tube.A volume of 250 μL Protein Precipitation Solution (MP Bio) was then added.Pellets precipitated after centrifuging at 14000×g for 10 mins,and the supernatant was transferred to a clean 15 mL tube.Then 1 mL resuspended binding matrix suspension was added into the supernatant.The tubes were inverted by hand for 5 mins to allow DNA binding,and placed in a rack for 30 mins to allow settling of silica matrix.Then approximately 600 μL of the binding matrix was transferred to a SPIN Filter (MP Bio) and centrifuged at 14000×g for 2 mins.Then the filtered liquid in the catch tube was transferred into the SPIN Filter and centrifuged again at 14000×g for 1 min.The waste in the catch tube was discarded and a new 600 μL of the binding matrix was added into the SPIN Filter.After all binding matrix from the 15 mL tube was double-filtered,500 μL prepared SEWS-M was added to the filter to wash the matrix by centrifuging at 14000×g for 2 mins.Empty the catch tube and repeat this centrifuge step,and air dry the filter for 20 mins at room temperature.At last,100 μL of DNase/Pyrogen-Free Water was added to the filter to resuspend the binding matrix.Water bath incubation of the filter at 55°C for 5 mins was applied to increase DNA yields.Eluted DNA was collected in a new clean catch tube by centrifuging at 14000×g for 1 min.

Crude DNA was purified by Wizard®Genomic DNA Purification Kit (Promega,Madison,WI) based on the manufacturer’s protocol.Basically,300 μL crude DNA of each sample was concentrated into 50 μL purified DNA.The purified DNA was then diluted 5 times and 25 times for amplification by regular PCR to test PCR inhibition (Jaeschke A et al.,2010).The archaealamoAgenes were amplified using the primer set Arch-amoA-1F and Arch-amoA-2R(Francis CA et al.,2005).The bacterialamoAgenes were amplified using the primer setamoA-1F andamoA-2R(Rotthauwe JH et al.,1997).The 16S rRNA gene fragments of anammox bacteria were amplified using the primer sets A438F and A684R (Humbert S et al.,2012).Optimized PCR thermal profiles for different genes amplification are listed in Table 2.Extracted DNA with no significant PCR inhibition was stored at −20°C and will be used in later qPCR amplification.

Table 1.Lithology description of sediment samples for qPCR experiments.

In a final volume of 25 μL PCR system,a reaction mixture contained: 1 μL of extracted template DNA (1-10 ng),2.5 μL of 10X Ex TaqBuffer (Mg2+free) (Takara,Hong Kong),and 2.5 μL of MgCl2(25 mM,Takara),2 μL of dNTP Mixture (2.5 mM,Takara),1 μL BSA (10 mg/mL,Promega,Hong Kong),0.5 μL of each forward and reverse primer (20 μM,synthesized at BGI,China),and 0.2 μL of TaKaRa Ex Taq(5 units/μL,Takara).PCR products were analyzed by agarose gel electrophoresis in 1% agarose gels in TAE (20 mM Tris-acetate pH 8.0; 0.5 mM EDTA) at 120 V for 30 mins(Amersham Biosciences,Electrophoresis Power Supply 301).The gels were stained by the addition of GelRed nucleic acid stain (Biotium) and photographed by using a Bio-Rad®GelDoc™station.

2.3.Quantitative PCR experiment

The abundances of archaeal and bacterialamoAgenes and anammox bacterial 16S rRNA genes in extracted DNA were determined in triplicate using an Applied Biosystems StepOnePlus™ Real-Time PCR System at the Centre Lab of School of Biological Sciences,The University of Hong Kong(HKU).The quantification was based on the fluorescent dye SYBR-Green using FastStart Universal SYBR Green Master(Rox) Kit (Roche,Germany).The primer sets composed of Arch-amoA-1F and Arch-amoA-2R,andamoA-1F andamoA-2R were used for the amplification of the AOA and AOBamoAgenes,respectively.The primer set composed of A438F and A684R was used for the amplification of the 16S rRNA genes of anammox bacteria.Optimized qPCR thermal profiles were also listed in Table 2.Each reaction was performed in a 20 μL volume system containing 10 μL of Power SYBR Green PCR Master Mix (Applied BioSystems),1 μL of DNA template (1-10 ng),0.2 μL of each primer (20 μM); 1 μL BSA (10 mg/mL),and 7.8 μL of autoclaved deionized water.A standard plasmid carrying archaeal or bacterialamoAgene or anammox bacteria 16S rRNA gene was generated by amplifying the gene from DNA extracted from marine surface sediment and cloning into the pMD 18-T Vector (Takara) as described by Han P et al.(2013a,2013b).The plasmid DNA concentrations were determined,and the gene abundance in terms of copy numbers was calculated directly from the concentration of the extracted plasmid DNA.Tenfold serial dilutions of a known copy number of the plasmid DNA were subjected to a quantitative PCR assay in triplicate to generate an external standard curve.Non-specific binding or primer dimers were ruled out by melt curve analysis with a single peak.The qPCR amplification efficiencies were 82.1%(AOA-amoA),89.9% (AOB-amoA),and 114.4% (anammox-16S),respectively.The correlation coefficient (R2) of the standard curve was >0.93 for all of the standard curves.It has been observed so far thatamoAand 16S rRNA genes are only found in one copy number per genome in nitrifiers and anammox bacteria,the determining gene copy numbers were used to represent the abundances of nitrifiers and anammox in samples for later discussion.

2.4.Porewater chemistry measurement

Porewater from bulk sediment was extracted by centrifuging at the speed of 5000 rounds/min,after a small part of the sediment was subsampled for DNA extraction.Porewater separated from the sediment was filtered using syringes with 0.45 μm pore-size filters.Concentrations of major ions (Na+,K+,Mg2+,Ca2+,NH4+,Cl−,NO2−,NO3−,SO42−) by Dionex ICS-1100 ion chromatography system in the Department of Earth Science,HKU.Porewater pH and TDS were measured with pH meter (HANNA HI98127 pHep®4 Tester) and TDS meter (HANNA HI 98312 DiST®6 Tester),respectively.Dissolved inorganic carbon (DIC),majorly dissolved carbonate and bicarbonate in porewater,was measured by HACH Digital Titrator.Total organic carbon (TOC) and total nitrogen (TN) contents were determined by Elementar VarioEL Cube (Elementar,Germany) at the State Key Laboratory of Marine Geology of Tongji University,Shanghai.Forδ15N-NH4isotopic analysis,10-15 mL filtered porewater was sealed in plastic bottles and sent to the Environmental Isotope Lab,University of Waterloo,Canada within icebox and measured by Micromass IsoChrom isotope-ratio mass spectrometer (Micromass,Manchester,U.K.) using a modified version of the diffusion method.

Geochemical results were shown in Table 3.Correlations between gene abundances and environment variables were analyzed by Pearson correlation method using SPSS 19.0.0 software (SPSS Inc.,IBM company,USA) and were shown in Table 4.

3.Results

3.1.Chemical profiles in PRD sediments

Geochemical and isotopic parameters were measured and shown in Table 3.Fig.2 showed the depth profiles of NH4+,NO2−,NO3−,TN,15N,pH,TDS (Total Dissolved Solids),DIC(Dissolved Inorganic Carbon),TOC (Total Organic Carbon),and TOC/TN.Significant variations existed among different geological layers in both Beijiao and Minzhong sites.Concentrations of major ions which are abundant in seawater(including Na+,K+,Mg2+,and Cl−) were mainly controlled by the sea level during the evolution of the PRD.At both sites,they were higher in the marine layers M1 and M2 and were lower in aquifers T1 and T2.Concentrations of Ca2+and SO42−in sediment porewater were largely influenced by mineralization and recrystallization processes in forming minerals such as gypsum and pyrite,demonstrated bycalculated saturation index and SEM observation of these minerals in sediment at similar locations of the PRD (Wang Y and Jiao JJ,2012a).The abundances of nitrogen and carbon species are important factors to the survival and growth of nitrifiers and can be significantly altered by the nitrification processes.Their results will be discussed with microbial information in the discussion part.

Table 2.PCR and qPCR primer sets and programs.

Ta bl e 3.C on ce nt ra ti on o f t es te d c he mi ca l o f t he s am pl es i n t he P ea rl R iv er D el ta.No.De pt h/m Na+K+ Mg2+ Ca2+ NH4+Cl−NO2− N O3−SO42− DI C TD S pH TO C/% T N/‰ T OC/T N δ15 NNH4/‰ AO A AO B AM X/( mg/L)Co pi es p er g ra m w et s ed im en t BJ-2 2.6 25.72 4.99 12.68 87.06 1.66 24.61 0.238 10.475 109.16 174.7 451.03 6.56 1.55 1.18 13.14 6.09 4.15 E+06 1.77E+05 6.02E+05 BJ-5 5.0 187.22 15.23 30.47 30.90 28.52 100.54 0.017 3.166 8.41 642.3 1043.64 7.03 1.53 1.15 13.30−1.45 3.13 E+05 7.17E+04 4.80E+05 BJ-8 8.2 442.61 23.95 46.69 28.61 46.29 249.60 0.119 3.576 6.34 1153.4 1999.99 7.27 2.89 1.12 25.80 0.21 1.94 E+07 1.19E+06 1.35E+06 BJ-10 10.0 698.00 32.67 62.90 26.31 64.06 398.67 0.279 6.534 4.27 1664.5 2956.36 7.27 1.58 0.49 32.24 1.29 3.21 E+07 2.74E+06 9.79E+05 BJ-14 13.9 705.80 27.88 51.96 28.15 48.71 457.10 0.368 13.347 0 1545.0 2871.77 6.89 1.40 0.91 15.38 4.93 1.69 E+06 1.83E+04 8.96E+06 BJ-17 17.6 887.82 35.02 30.71 44.72 48.81 660.27 0.351 22.242 127.08 1417.8 3252.24 7.87 1.59 0.94 16.91 5.57 1.22 E+06 1.75E+05 1.35E+06 BJ-18 18.6 774.49 30.07 29.93 45.39 40.45 628.16 0.226 19.025 100.74 1197.7 2846.98 6.58 1.79 1.07 16.73 10.89 6.77 E+06 1.47E+05 6.44E+05 BJ-20 20.7 661.16 25.13 29.15 46.06 32.10 596.05 0.407 24.448 74.40 977.6 2441.71 8.43 1.82 1.09 16.70 8.52 5.29 E+06 4.80E+04 3.53E+06 BJ-21 21.2 547.83 20.18 28.38 46.72 23.74 563.95 0.623 7.452 48.06 757.5 2036.46 6.56 1.91 1.01 18.91 2.57 1.29 E+06 5.47E+04 6.48E+06 BJ-24 24.3 434.50 15.25 27.60 47.39 15.39 531.84 0.485 18.313 21.72 537.5 1631.19 8.16 1.64 1.18 13.90 3.39 9.17 E+05 3.60E+04 1.29E+06 MZ-5 5.6 4742.71 143.79 455.95 76.92 109.45 7982.54 0.159 2.238 106.09 431.2 14175.21 7.90 1.56 1.42 11.14 6.0 9.75 E+06 1.09E+06 2.70E+06 MZ-8 8.1 6299.56 218.66 825.08 92.75 110.44 10548.13 0.013 1.231 369.02 302.2 19070.42 7.80 1.48 1.23 12.33 7.4 1.19 E+06 2.19E+05 4.54E+06 MZ-11 11.3 6975.60 241.43 765.41 89.05 123.49 11914.43 0.054 0.792 4.73 3682.3 23940.92 7.36 1.81 1.31 13.92 6.0 2.03 E+06 4.78E+05 4.43E+06 MZ-15 15.5 7983.04 271.62 845.02 60.89 148.60 13765.07 0.038 0.104 133.60 3080.1 26459.65 7.66 1.29 0.90 14.33 5.9 2.59 E+06 3.37E+05 5.86E+06 MZ-20 20.6 8744.95 310.33 980.21 82.10 187.26 15448.82 0.039 0.738 9.27 1188.0 27127.23 7.69 1.45 1.12 13.18 5.8 2.37 E+06 1.53E+05 6.51E+06 MZ-26 26.0 8776.44 309.35 775.58 64.30 233.18 15458.51 0.046 0.898 177.53 1798.6 27804.89 7.30 1.70 1.28 13.08 6.3 3.24 E+05 3.90E+05 7.87E+06 MZ-31 31.2 9010.68 306.23 787.84 81.97 260.31 16029.92 0.017 1.755 148.95 478.5 27320.99 7.62 1.68 1.08 15.27 6.7 3.26 E+06 3.68E+05 4.62E+06 MZ-34 34.0 4453.74 139.25 28.84 41.07 135.76 8027.21 0.678 6.115 12.28 254.9 13672.77 8.84 2.31 1.51 15.40 4.7 8.52 E+06 1.53E+06 9.27E+04 MZ-37 37.6 9965.19 397.55 1599.76 642.19 301.40 16270.12 0.003 3.323 697.27 334.4 30853.78 7.91 1.63 1.27 12.54 5.2 4.32 E+06 6.95E+04 4.45E+05 MZ-40 40.0 8398.52 318.68 735.66 292.95 227.41 15870.38 0.494 4.771 53.96 575.0 26544.49 6.32 1.13 1.04 11.30 NA 5.13 E+05 6.84E+04 3.69E+05 Notes: N A- no t a pp li ca bl e d ue t o n ot e no ug h p or ew at er y ie ld f or δ15 NN H4 a na ly si s; m aj or i on s, p H, D IC,T DS,a nd 15 N- NH4 i so to pe f ro m p or ew at er o f s ed im en t; T OC,T N c on te nt s, a nd T OC/T N f ro m b ul k s ed im en t; ab un da nc e o f A OA- a mo A, A OB-a mo A, a nd a na mm ox 16S r RN A g en es f ro m s ed im en t.

3.2.Gene abundances in PRD sediments

Gene abundances (in the unit of copies per gram wet soil)of archaealamoA,bacterialamoA,and anammox bacterial 16S rRNA were plotted in Fig.3,marked as AOA,AOB,and AMX,respectively.Detailed numbers are in Table 3.Fig.3a showed the distribution of gene copy numbers in Beijiao along with depth.AOA ranged from 3.13×105copies/g to 3.21×107copies/g in the Holocene M1 aquitard,which was highest among four layers in Beijiao.AOA then dropped in the underlying terrestrial T1 aquifer and old marine M2 aquitard,and the basal aquifer T2 had the lowest AOA number on average.AOB had a similar distribution pattern as that of AOA.The distribution of the anammox 16S rRNA gene was very different from those of AOA or AOB.Anammox bacterial number generally increased with depth in aquitards,and then dropped in aquifers.In general,abundances of anammox bacteria were comparable to those of AOA in each layer,while abundances of AOB were 1-2 orders in magnitude lower than those of anammox bacteria and AOA.

Table 4.Pearson correlations between genes abundances and environmental variables.

Fig.3b showed distributions of AOA,AOB,and anammox bacteria in Minzhong.amoAgene numbers of archaea and bacteria both increased in aquifer layer T1 and dropped to a very low level in old marine layer M2.For anammox bacteria,16S rRNA gene numbers slightly increased in aquitard layers with depth and decreased in aquifers.Generally speaking,anammox bacteria were 1-2 times higher than AOA and over 10 times higher than AOB in the aquitard M1,except for the top sediment sample at −5 m,where anammox bacteria were lower than AOA but still higher than AOB.For layers T1 and M2,AOA outnumbered anammox bacteria in gene numbers,while AOB was still the lowest in average compared with the other two nitrifiers.

4.Discussion

4.1.N-cycling in the PRD groundwater system

In Beijiao,NH4+,NO2+,and NO3+had a similar trend along with depth.They were higher in top sediment and aquifers T1 and T2,due to recharged freshwater with more dissolved oxygen.δ15N-NH4ranged from −1.45‰ in the aquitard M1 to 10.89‰ in the aquifer T1,indicating the ammonium in every layer was decomposed from natural materials such as plants and organic nitrogen (Amberger A and Schmidt H,1987; Böttcher J et al.,1990).δ15N-NH4values had a very similar distribution pattern to those of NOx−,implying that in Beijiao the enrichment of15N in ammonium of porewater was majorly contributed by the process of aerobic ammonium oxidation by AOA and AOB because microorganisms preferentially oxidize lighter isotopes of NH4+.In Minzhong,the C and N contents were quite different from those in Beijiao,due to the near-coast location and different deltaic formation processes.The M1 aquitard layer in Minzhong is over 25 m and much thicker than that in Beijiao.Thus,the reducing condition in this layer in Minzhong was much stronger.This highly reducing environment caused very low concentrations of NO2−(less than 0.2 mg/L) and NO3−(less than 2.5 mg/L),and a very high concentration of NH4+(up to 260 mg/L).Then in the relatively fresh layer aquifer T1,NO2−and NO3−dramatically increased to their highest points while NH4+decreased to its lowest concentration.For the same reason of reducing condition,the concentrations of NOx−decreased in aquitard M2 and increased in aquifer T2,while the concentration of NH4+had the opposite changing trend.δ15N-NH4ranged from 7.4‰ in aquitard M1 to 4.7‰ in aquifer T1.The depletion of15N in aquifer T1 might have two reasons.One reason was that aquifer T1 had been influenced by15N-depleted materials,such as precipitation water and synthetic fertilizer,which hadδ15N values around zero.The other possibility was that the15N-enriched ammonium in aquitards was attributed to the anammox process.

4.2.Gene abundances of AOA,AOB,and anammox in PRD sediments

The results of gene abundances of AOA,AOB,and anammox bacteria were comparable to previous studies in paddy soils and shallow riparian sediments of less than one meter deep (Hefting MM et al.,2006; Humbert S et al.,2012;Jin T et al.,2011; Shen JP et al.,2012; Wang S et al.,2012;Yang A et al.,2015).Similar to soil-like sediments,in the PRD aquitard-aquifer system,AOA,other than AOB,was the dominant aerobic nitrifiers inamoAgene numbers because the conditions of salinity,NH4+concentration,pH,and organic carbon content all favored the survival and growth of archaeal microbes (Francis CA et al.,2005; Liu S et al.,2013; Santoro AE et al.,2008; Wang S et al.,2011).amoAgenes of AOA and AOB in the investigated sediment were 2-3 orders higher than those in Pearl River water (Fig.1) (Liu ZH et al.,2011b),but 2-3 orders lower than those in PRD paddy soils (Zhu GB et al.,2011).The 16S rRNA gene abundances of anammox bacteria were about one order lower than that in river riparian sediment of the PRD (Wang S et al.,2012),and 1-3 orders lower than that in fertilized paddy soil (Wang et al.,2012b).In this research,abundances of anammox bacteria in Beijiao andamoAgenes in Minzhong reached their maximum points at the interface between aquifer T1 and aquitard M2,and obvious changes were also notified near the interface between aquitard M1 and aquifer T1 (Fig.3).Significant changes of NOx−concentrations,pH values,and TOC contents at those interfaces may be accounted for by the elevation of gene numbers.Thus,for the first time,the interfaces between aquifers and aquitards in deep deltaic sediments were demonstrated as hot zones for anammox andamoAgenes.

Fig.3.Gene abundances of archaeal amoA,bacterial amoA,and anammox bacterial 16S rRNA in sediment samples of Beijiao (a) and Minzhong (b).Red circles represented archaeal amoA (AOA); green triangles represented bacterial amoA (AOB); blue diamonds were for anammox bacterial 16S rRNA (Anammox).Mean values and standard deviations were calculated from triplicate assay within a single qPCR setup.OSL ages were obtained from the OSL dating experiment in this study,while one 14 C age was from a government report previously published(GHT,1981).

4.3.Relationships between gene abundances and environmental variables in PRD sediments

To investigate the relationship between nitrifiers and environment conditions in the PRD groundwater system,Pearson correlations between gene abundances (amoAgene of AOA and AOB,16S rRNA gene of anammox bacteria) and major environmental variables (NH4+,NO2−,NO3−,TN,δ15N,DIC,TOC and TOC/TN together with pH and TDS) were conducted and shown in Table 4.Those variables had been reported as controlling factors affecting gene abundances and community structures of AOA,AOB,and anammox bacteria(Francis CA et al.,2005; Liu S et al.,2013; Moore TA et al.,2011; Risgaard-Petersen N et al.,2004; Trimmer M et al.,2003).To relate the discussion more specifically to the hydrogeological conditions,the four strata were divided into two groups for each study site.Group one was the silty-clayey uniformed marine aquitard M1 recently formed in the Holocene with low permeability,and group two was the heterogeneous aquifer-aquitard-aquifer (T1-M2-T2) system with much higher groundwater conductivity and geochemical activities.Pearson correlations between gene abundances and environmental variables were separately produced in the M1 group,T1-M2-T2 group,and all-strata group for each study site.

For aquitard M1 in Beijiao (Table 4-A),abundances of AOA and AOB were both positively correlated with TOC/TN ratio (R2=0.861 and 0.869,respectively),showing thatamoAgenes favor for higher TOC/TN conditions,and AOA and AOB shared similar ecological niches,as they were positively correlated with each other (0.842).Anammox bacteria were significantly correlated with the depth (0.882) and NO2−(0.707).Nitrite was a controlling factor for the anammox process (Dalsgaard T and Thamdrup B,2002; Kuypers MMM et al.,2003),and in this oxygen-limited aquitard,NO3−and NO2−were supplied from the underlying T1 aquifer,which explained why the anammox process was intensified with depth.For the T1-M2-T2 group in Beijiao (Table 4-B),AOA and AOB were no longer dependent on TOC/TN.AOA was well linked with15N enrichment,indicating the elevated15N in ammonium was derived by the archaealamoAgene.AOB was positively related to the concentrations of NH4+(0.926),TDS (0.926),and HCO3−(0.926),but negatively related with depth (−0.928) and TN (−0.724),showing a different ecological niche from AOA.In this group,the abundance of AOB was one to two orders of magnitude lower than that of AOA,indicating the alternation of AOB’s ecological niche was forced by the overwhelming number of their competitor AOA.Anammox bacteria were still positively correlated with the concentration of NO2−,as in aquitard M1,demonstrating that NO2−was still the limiting factor for the anammox process in this system.For all strata in Beijiao (Table 4-C),the correlation result was very similar to that in aquitard M1,showing that aquitard M1 was the dominating hydrological layer in controlling the numbers of nitrifiers.

For aquitard M1 in Minzhong (Table 4-D),AOA and AOB were significantly correlated with NO2−(0.965 and 0.874,respectively),while anammox bacteria were negatively correlated with NO2−(−0.831) and positively with depth(0.899).This result suggested that in the reducing environment of aquitard M1,NO2−was generated by AOA and AOB from oxidizing NH4−with dissolved oxygen supplied from precipitation.At the bottom of the M1 aquitard,where dissolved oxygen was much lower,anammox bacteria increased within two orders in magnitude,rather than AOA and AOB.For the aquifer-aquitard-aquifer group in Minzhong(Table 4-E),AOA and AOB were no longer significantly correlated with NO2−or NO3−,while anammox bacteria were still negatively correlated with NOx−.Also,15N-NH4abundance in ammonium was negatively correlated with AOA and positively with anammox bacteria.These observations all indicated that in this kind of hydrological group,the anammox process,rather thanamoAoxidation,was the dominant NH4+utilizing pathway.Dissolved oxygen from recharged groundwater in aquifers inhibited anammox bacteria,and elevated the concentration of NO2−and NO3−.AOA and AOB were also positively linked with TOC content and TOC/TN ratio,implying that aerobic ammonium oxidation might demand more organic supplement than anammox,because decomposition of organisms could provide the NH3that AOA and AOB needed,while anammox could utilize NH4+which were dissolved in groundwater or absorbed on sediment particles.In this system,pH condition was positively linked with AOA and AOB,which had also been reported in mangrove sediment of the PRD (Li M et al.,2011).For all strata in Minzhong (Table 4-F),16S rRNA gene abundances of anammox bacteria were negatively correlated with depth (−0.642),NO2−(−0.824),NO3−(−0.967) and positively withδ15N (0.746),while no strong correlation was observed between AOA,AOB and other environmental conditions,except a positive correlation between AOA and pH (0.754) and a negative correlation between AOB and TDS(−0.757).This result showed that anammox was the predominant pathway for NH4+oxidation in the Minzhong groundwater system.

The two study sites,Beijiao in the northwest of the PRD and Minzhong in the southeast of the PRD,showed distinguished relationships between gene abundances and environmental variables,even in the similar hydrological unit.In the aquitard M1,amoAgenes in Beijiao were majorly affected by nutrition conditions,represented by an index of TOC/TN,while in Minzhong they were linked with NO2−concentration.In the aquifer-aquitard-aquifer system,AOB in Beijiao positively correlated with TDS and HCO3−,while in Minzhong the correlation became negative.For anammox,the gene copy numbers of 16S rRNA were always correlated with the concentration of NO2−,however,the correlation was positive in Beijiao but negative in Minzhong.The differences were believed to be caused by different geological and geochemical conditions at these two places.In Beijiao,the groundwater system was more active than in Minzhong,because Beijiao was near the mountains in the north part of Guangdong Province,while Minzhong was much flatter in topography and its aquitard M1 was 10 m thicker than that of Beijiao.Also,the permeability in Beijiao was greater than in Minzhong because of bigger grain particles and less clay content.Thus,in Beijiao,AOA and AOB were better supported with higher oxygen and organic inputs,and the elevated NO2−supported anammox bacteria.In Minzhong,however,oxygen was limited,only in aquifers and the top of aquitard M1 that NO2−could be conducted by AOA and AOB,and the generated NO2−was continuously consumed by anammox bacteria.That might explain why the correlation between anammox and NO2−was positive in Beijiao and negative in Minzhong.

5.Conclusion

In this study,the abundances of archaeal and bacterialamoA genes and anammox 16S rRNA gene in two representative locations in the PRD were determined and analyzed with geochemical data obtained from the same samples.qPCR results showed that gene abundances of AOA,AOB,and anammox bacteria ranged from 3.13×105copies/g to 3.21×107copies/g,1.83×104copies/g to 2.74×106copies/g,and 9.27×104copies/g to 8.96×106copies/g in the sediment of the groundwater system,respectively.In the Beijiao study site,AOA was the predominant nitrifiers in most of the strata,followed by anammox bacteria,while in Minzhong,anammox bacteria were overwhelming in aquitard samples,while AOA was dominant in aquifers.AOB was much lower than AOA and anammox bacteria in gene numbers in most of the samples.Geochemical profiles and Pearson correlations indicated that AOA was majorly responsible for ammonia oxidation in Beijiao,while anammox bacteria were attributed to microbial-derived ammonium loss in the Minzhong groundwater system.Hotspots of geomicrobial activities were observed at aquifer-aquitard interfaces,indicated by significant variation of copy numbers of bothamoAand 16S rRNA gene.It is the first time that AOA,AOB,and anammox bacteria were quantitatively investigated in subsurface deltaic sediment,showing their ubiquitous existence in paleo sediment.Significant connections between gene abundances and geochemistry were noted and discussed.This study provided first-hand geomicrobiological data of nitrifying microbes in the ammonium-rich groundwater system of the PRD and can help better understand their important roles in geochemistry in the groundwater environment.

CRediT authorship contribution statement

Kun Liu conceived the presented idea,processed experiments,analyzed data,drew all the figures and prepared the manuscript.All authors discussed the results and contributed to the final manuscript.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgment

This study was financially supported by the General Research Fund of the Research Grants Council,the Hong Kong Special Administrative Region,China (HKU 702612P and HKU 703010P).Xing-xing Kuang and Xin Luo are thanked for their help in field sampling; Ping Han,Jing Chen,and Jessie Lai are thanked for their help in lab facility arrangement.