Shifts in Archaeaplankton Community Structure Along Ecological Gradients of Pearl Estuary Jiwen Liu1, Shaolan Yu1, Meixun Zhao2, Biyan He3,4 & Xiao-Hua Zhang1

Home , Geodia barretti

RESEARCH ARTICLE Shifts in archaeaplankton community structure along ecological gradients of Pearl Estuary Jiwen Liu1, Shaolan Yu1, Meixun Zhao2, Biyan He3,4 & Xiao-Hua Zhang1

1College of Marine Life Sciences, Ocean University of China, Qingdao, China; 2Key Laboratory of Marine Chemistry Theory and Technology, Ministry of Education, Ocean University of China, Qingdao, China; 3State Key Laboratory of Marine Environmental Science, Xiamen University, Xiamen, China; and 4School of Bioengineering, Jimei University, Xiamen, China Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021

Correspondence: Xiao-Hua Zhang, College Abstract of Marine Life Sciences, Ocean University of China, 5 Yushan Road, Qingdao 266003, The significance of archaea in regulating biogeochemical processes has led to China. Tel./fax: +86 532 82032767; an interest in their community compositions. Using 454 pyrosequencing, the e-mail: [email protected] present study examined the archaeal communities along a subtropical estuary, Pearl Estuary, China. Marine Group I Thaumarchaeota (MG-I) were predomi- Received 18 April 2014; revised 29 July nant in freshwater sites and one novel subgroup of MG-I, that is MG-Im, was 2014; accepted 31 July 2014. Final version proposed. In addition, the previously defined MG-Ia II was grouped into two published online 28 August 2014. clusters (MG-Ia II-1, II-2). MG-Ia II-1 and MG-Ik II were both freshwater- a k DOI: 10.1111/1574-6941.12404 specific, with MG-I II-1 being prevalent in the oxic water and MG-I II in the hypoxic water. Salinity, dissolved oxygen, nutrients and pH were the most Editor: Gary King important determinants that shaped the differential distribution of MG-I subgroups along Pearl Estuary. Marine Group II Euryarchaeota (MG-II) domi- Keywords nated the saltwater sites, but their abundance was higher in surface waters. The 16S rRNA gene; 454 pyrosequencing; habitat phylogenetic patterns of MG-I subgroups and their habitat preferences provide preference; hypoxia; Marine Group I; Pearl insight into their phylogeographic relationships. These results highlight the Estuary. diversification of various ecotypes of archaea, especially of MG-I, under distinct environmental factors in Pearl Estuary, which are of great value for further exploring their ecological functions.

ammonia-oxidizing archaea (AOA) (Mosier & Francis, Introduction 2008; Santoro et al., 2008; Bernhard et al., 2010) from Archaea were previously thought to be present only in freshwater to marine systems. Characterized by strong extreme environments, but since the ﬁrst discovery of mixing of freshwater and saltwater with sharp salinity and Crenarchaeota in a temperate habitat (DeLong, 1992; nutrient gradients, estuaries are ideal ecosystems to eluci- Fuhrman et al., 1992), distinct archaeal lineages have date the response of different microbial phylotypes to been found in non-extreme habitats such as ocean waters, physicochemical variations. freshwater sediments, and soils (Auguet et al., 2010). Marine Group I Thaumarchaeota (MG-I) and Marine Thus, studies on structures of archaeal community and Group II Euryarchaeota (MG-II) are the most abundant their distribution patterns against environmental charac- groups of archaea in marine systems (Takai et al., 2004; teristics are essential to understand their role in ecosys- Galand et al., 2006). Members of MG-I, a major compo- tems. nent capable of aerobic ammonia oxidation (Konneke€ The signiﬁcance of salinity in shaping bacterial com- et al., 2005), are widespread not only in oceans but also munity structure has been well documented (e.g. Kirch- in freshwater and soils (e.g. DeLong, 1992; Fuhrman man et al., 2005; Lozupone & Knight, 2007). et al., 1992; Schleper et al., 1997; Inagaki et al., 2003; Nevertheless, the evolutionary separation of archaeal Brown et al., 2009; Durbin & Teske, 2010; Pester et al.,

MICROBIOLOGY ECOLOGY MICROBIOLOGY assemblages dwelling in different salinity ranges is still 2012). Different phylogenetic ecotypes of MG-I have been elusive as few phylogenetic studies are available regarding deﬁned previously (Massana et al., 2000; Sørensen et al., the transition of community structure of total archaea 2004; Takai et al., 2004; Durbin & Teske, 2010; Jorgensen (Crump & Baross, 2000; Galand et al., 2006, 2008) and et al., 2012), of which MG-Ia and MG-Ic are phylotypes

ª 2014 Federation of European Microbiological Societies. FEMS Microbiol Ecol 90 (2014) 424–435 Published by John Wiley & Sons Ltd. All rights reserved Shifts in archaeal community structure 425 commonly retrieved from shallow and deep seawater, Corporation, Billerica, MA). The membranes were stored respectively (Massana et al., 2000; Takai et al., 2004). in liquid nitrogen onboard and at 80 °C in the lab until However, the habitat preference of these subgroups, espe- nucleic acid extraction. Samples for nutrients (NO2 , + 3 cially that of terrestrial ecosystems, remains largely NO3 ,NH4 and PO4 ) were analyzed with spectropho- unknown. tometric and colorimetric methods (Dai et al., 2006; Han The wide distribution of MG-I subgroups raises the et al., 2012). question of the driving forces of such distribution. Depth was considered a predominant factor in the structure of Nucleic acid extraction and high-throughput archaeal assemblages (Brown et al., 2009; Hu et al., sequencing 2011). Dissolved oxygen (DO) was also an important determinant, as a species of MG-I, Nitrosopumilus mariti- Community DNA was extracted using the method Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 mus (Konneke€ et al., 2005), was found to grow under described by Yin et al. (2013) with a modified step to fully aerobic conditions, whereas some specific ecotypes maximize the DNA output, in which a Fast Prep-24 of AOA could tolerate DO concentration near 1 lM Homogenization System (MP Biomedicals, Irvine, CA) (Erguder et al., 2009). Therefore, the adaptations and was used to intensify cell lysis at maximum speed for growth responses of different MG-I subgroups to low DO 40 s. DNA integrity was checked on a 1% (w/v) conditions should be investigated further. Compared with agarose gel. Amplification of archaeal 16S rRNA gene the widely studied MG-I, the distribution pattern of MG- was conducted using barcode and adaptor added pri- II remains largely unknown. mer 344F (50-ACGGGGYGCAGCAGGCGCGA-30) and The archaeal ecology of subtropical estuaries remains 915R (50-GTGCTCCCCCGCCAATTCCT-30) (Casamayor elusive. Pearl Estuary, a typical subtropical estuary in et al., 2002). A 20 lL PCR reaction included 4 lLof59 China, receives discharges from four eastern outlets, FastPfu Buffer, 2 lL of 2.5 mM deoxynucleoside triphos- Humen, Jiaomen, Hongqimen and Hengmen (Support- phate (dNTP) mix, 0.8 lL of each primer (5 lM), 0.4 lL ing Information, Fig. S1) (PRWRC/PRRCC, 1991), and of TransStart Fastpfu DNA Polymerase (TransGen), 1 lL is subject to hypoxia caused by excessive anthropogenic of template DNA and 11 lL of double-distilled H2O. A inputs in both the upstream and downstream areas triplicate PCR reaction was performed and PCR products (Yin et al., 2004; Dai et al., 2006). The archaeal com- were purified with an AxyPrepDNA Gel Extraction Kit munity distribution along gradients of various environ- (Axygen, Hangzhou, China) and quantified using a mental factors of Pearl Estuary has not been Quant-iT PicoGreen double-stranded DNA assay systematically investigated. To gain an insight into the (Invitrogen, Carlsbad, CA). The amplicons from each environmental and geochemical controls on archaeal reaction mixture were pooled in equimolar ratios based community composition and to test whether archaea on concentration and subject to emulsion PCR to are adapted to different environments such as salinity generate amplicon libraries, as recommended by 454 Life and DO gradients, 454 pyrosequencing of the archaeal Sciences. Sequencing was conducted on a Roche Genome 16S rRNA gene along Pearl Estuary was performed in Sequencer FLX Titanium platform at Majorbio this study. Our study highlights that different MG-I Bio-Pharm Technology Co. Ltd. (Shanghai, China). subgroups evolved to accommodate different environments. Sequence processing and operational taxonomic unit (OTU) assignment Materials and methods All the sequence analysis processes followed the pipeline of MOTHUR (Schloss et al., 2009). All reads completely Sampling matching the barcodes were retained as well as reads with a Surface and bottom water samples of six sites along the maximum single mismatch to the primers. Reads were then salinity gradient of Pearl Estuary (Fig. S1) were collected trimmed by removing the sequencing adaptor, barcodes during 16 July to 12 August, 2012. P01, P03 and P07 rep- and primer sequences. Reads shorter than 200 base pairs, resented the sites located in the upstream freshwater with an average quality score lower than 25 and with any region, and A08, C2 and F412 the sites situated in the ambiguous bases were further removed. Chimeric downstream saltwater region. Samples were collected with sequences were identified and removed using UCHIME a Sealogger CTD (SBE 25, Sea-Bird Co.) rosette water (Edgar et al., 2011). After quality control, the average sampler. One liter of water samples for nucleic acid was length of the reads was 490 bp. The selected reads were prefiltered through 3-mm-pore size filters before collec- compared with the SILVA v111 16S rRNA gene reference tion on 0.22-lm polycarbonate membranes (Millipore database (http://www.arb-silva.de) and taxonomically

FEMS Microbiol Ecol 90 (2014) 424–435 ª 2014 Federation of European Microbiological Societies. Published by John Wiley & Sons Ltd. All rights reserved 426 J. Liu et al. assigned using a 3% distance level at 80% confidence tively, in the following analysis) from six sites were exam- threshold. ined. Environmental factors changed vertically in the The representative read of each of the top 50 OTUs saltwater sites, but remained consistent in the freshwater obtained was blasted against the National Center for Bio- sites (Table 1). The average water depth, ranging from 3 technology Information (NCBI, http://www.ncbi.nlm.nih. to 27 m across all sites, was approximately 10 m. The gov) database and the top-hit sequence was downloaded salinity ranged from 0 to 31.4 PSU along the estuary. 1 for each OTU. The sequences of MG-I subgroups described Nutrient levels of NO3 (4.4106.4 lmol L ), NO2 1 + 1 by Durbin & Teske (2010) and Jorgensen et al. (2012) were (1.8–43.1 lmol L ), NH4 (0.8140.1 lmol L ) and 3 1 retrieved to define different clusters. In addition, several PO4 (0.4–4.8 lmol L ), and chlorophyll a (Chl a, sequences cloned from freshwater or coastal water systems 0.9117.36 lgL 1) in the freshwater sites were signifi- (Auguet & Casamayor, 2008; Galand et al., 2008) that affil- cantly higher than in the saltwater sites (Wilcoxon’s rank + Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 iated to MG-I were also retrieved to construct the phyloge- test, P < 0.05). NH4 levels of the saltwater sites were clo- netic trees. A neighbor-joining (NJ) phylogenetic tree was ser to or under the detection limit of 0.5 lmol L 1.In 1 constructed using MEGA5 (Tamura et al., 2011) with the contrast, the levels of DO (0.22–6.95 mg L ) and pH K2 + G model. A maximum-likelihood (ML) phylogenetic (6.98–8.13) were lower in the freshwater sites (P < 0.05). tree was also constructed to verify the topological structure Among the three freshwater sites, DO level was lower given by the neighbor-joining tree. The tree topology was (< 1mgL 1) in P01 and P07, and higher in P03 checked by 1000 bootstrap replicates. (> 3mgL 1). The DO level of the bottom water in the The 454 raw reads were deposited in the NCBI Short saltwater region was lower than that of the surface water, Read Archive (SRA) database under accession number but did not reach hypoxia. In addition, the surface water SRP028510, and sequences of the top 50 OTUs, which in the saltwater sites has a significantly lower level of were used to construct the phylogenetic trees, were also salinity and turbidity but a higher level of NO3 com- deposited in the NCBI under accession numbers pared with the bottom water (Student’s test, P < 0.05). KF240565–KF240614 (OTU1–OTU50). Diversity of archaeal community and Statistical analyses community comparison The Chao I richness (Chao & Bunge, 2002) and Shannon The 12 samples studied here yielded a total of 31 308 diversity index (Magurran, 1988) used as diversity proxies high quality reads ranging from 1563 to 4091 in all the and Good’s coverage (Good, 1953) were calculated in samples (Supporting Information, Table S1). In total, MOTHUR. For statistical comparisons, the distribution and 1890 OTUs were clustered after randomly resampling, abundance matrix of the OTUs was normalized to the ranging from 103 to 393 OTUs per sample, at a 3% dis- sample with the smallest number of reads by randomly similarity level. The normalized Shannon and Chao I resampling using the perl script daisychopper.pl (Gilbert index were used as proxies for diversity and richness, et al., 2009). Complete linkage hierarchical clustering respectively. No significant differences in both measures based on Bray–Curtis dissimilarity using the normalized were observed between the freshwater and saltwater sites and square root-transformed OTUs matrix was performed (P > 0.05, Wilcoxon’s rank test). In addition, there were in PRIMER 5 (Plymouth Marine Laboratory, West Hoe, no significant differences in the diversity measures Plymouth, UK). Redundancy analyses (RDA) with 9999 between surface and bottom waters in either the freshwa- Monte Carlo permutation tests were conducted with ter or saltwater sites. Good’s coverages ranged from CANOCO (Version 4.5, Microcomputer Power) using square 85.4% to 98.2% in all the samples, indicating that most root-transformed data to evaluate the relationship of the species in the study areas could be represented by between phylotypes and environmental factors. Wilco- the libraries generated by 454 pyrosequencing. xon’s rank test was performed to statistically compare the For statistical analysis, the samples were randomly environmental factors and diversity measures in R (RDC resampled to equalize the sampling effort according to Team, 2008). the sample with the lowest number of sequences (1563 sequences). A clear transition of archaeal community composition was observed from the freshwater to saltwa- Results ter sites as shown in the hierarchical clustering dendrogram (Fig. 1). In the freshwater sites, the archaeal Environmental characteristics assemblages were consistent in different water layers and The physicochemical parameters of 12 samples including the non-hypoxic site P03 was distinguished from the two surface and bottom waters (designated _0 and _2, respec- hypoxic sites P01 and P07. The archaeal communities

Table 1. Environmental parameters of the 12 water samples

3 + Depth Temperature Salinity Turbidity DO Chl a PO4 NO3 NO2 NH4 Samples (m) (°C) (PSU) pH (FTU) (mg L1) (lgL1) (lmol L1) (lmol L1) (lmol L1) (lmol L1) P01_0 1 30.14 0.00 7.01 24.10 0.22 16.45 4.8 78.6 14.8 139.7 P01_2 7 30.10 0.00 7.01 24.11 0.51 17.36 4.7 96.1 14.7 140.1 P03_0 1 29.61 0.00 7.24 24.10 3.47 5.16 2.0 102.5 24.6 21.4 P03_2 3 29.61 0.00 7.21 24.08 3.14 6.45 2.2 103.0 26.4 30.6 P07_0 1 29.35 0.00 6.99 22.14 0.72 11.54 2.0 105.0 43.1 34.4 P07_2 19 29.32 0.10 6.98 24.11 0.62 9.73 2.0 106.4 41.9 31.1 A08_0 1 28.34 16.50 8.04 5.26 6.45 4.81 1.0 61.5 8.5 0.8 A08_2 13 26.45 31.50 8.09 23.42 4.68 0.91 0.5 4.4 4.6 n.d.

C2_0 1 28.31 18.66 8.01 9.08 5.48 1.55 1.0 54.3 6.5 1.1 Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 C2_2 6 27.22 28.34 8.03 24.11 4.09 1.13 0.8 18.5 4.9 n.d. F412_0 1 27.50 23.50 8.11 3.90 6.95 4.29 0.9 36.9 4.2 n.d. F412_2 14 26.39 31.41 8.13 7.82 5.39 n.a. 0.4 4.4 1.8 n.d. n.a., no analysis; n.d., non detectable. appeared to vary vertically between the surface and bot- eal clades were observed in the freshwater sites. Further- tom waters of the saltwater sites, although sample C2_0 more, the abundance of MG-I was higher in sites P01 separated earlier than the other surface samples. and P07 than in site P03. More lineages, such as Deep Sea Euryarchaeotic Group (DSEG), South African Gold Mine Group 1 (SAGMCG-1) and Soil Crenarchaeotic Taxonomic assignment Group (SCG), were observed in the freshwater sites, espe- Taxonomic assignment using the SILVA v111 database cially P03 and P07, than in the saltwater sites, reflecting grouped the reads of all samples into mainly two phyla, the higher diversity measures in those sites (Table S1). Euryarchaeota and Thaumarchaeota, with a few belonging to Crenarchaeota and others (including unclassified Diversity of the MG-I and their correlations groups) (Fig. 2a). The two major phyla displayed an with environmental factors opposite trend along Pearl Estuary, with Thaumarchaeota and Euryarchaeota dominating the freshwater and saltwa- For reducing complexity and excluding rare organisms, ter sites, respectively. At a finer taxonomic resolution, the top 50 OTUs (with sequences numbers ranged from MG-I were prevalent across all sites, especially abundant 80 to 2607) were extracted. Thaumarchaeota occupied 27 in the freshwater sites, whereas MG-II were more pre- OTUs: 24 MG-I, 2 SAGMCG-1 and 1 SCG. Euryarchaeota dominant than MG-I in the saltwater sites and almost accounted for 23 OTUs: 19 affiliated to MG-II, 1 to Mar- absent in the freshwater sites (Fig. 2b). The abundance of ine Group III, 1 to DSEG, 1 to Deep Sea Hydrothermal MG-II in the saltwater sites was higher in the surface than Vent Group 6 (DHVEG-6) and 1 to genus Haloterrigena the bottom water, whereas no vertical variations of archa- in Halobacteria (Fig. S2).

Fig. 1. Hierarchical clustering based on the OTU composition and abundance showing the similarity of microbial communities. The OTU matrix was resampled with ‘daisychopper.pl’ to equalize sampling effort. The dendrogram was constructed based on Bray–Curtis dissimilarity using PRIMER 5. SS, saltwater samples; FS, freshwater samples; H, hypoxic sites; NH, non-hypoxic site.

(a) Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021

(b)

Fig. 2. Relative abundance of major archaeal clades of Pearl Estuary at different taxonomic levels: (a) at phylum level; (b) at deeper taxonomic levels. Reads were assigned in the SILVA database by using a minimum support threshold of 80%. The salinity range for each sample is provided and the hypoxic samples are indicated with asterisks.

The MG-I phylogenetic trees constructed identiﬁed of MG-Ik subgroup suggested it was an ancestor clade of three new clusters: a II-1, a II-2 and m (nu). All the MG- MG-I group, despite there presence of a ‘phylogenetic I OTUs (except OTU 33) obtained in this study fell into jump’ between MG-Ik I and MG-Ik II, as shown in trees four clades, with subgroups a I, a II-1, k II (lambda II) with different algorithms. and m occupying 12, 4, 6 and 1 OTUs, respectively. MG- The correlation of MG-I subgroups with various envi- Im was a monophyletic clade supported by both the topo- ronmental factors was analyzed. RDA1 and RDA2 logical structure and bootstrap value and was represented together explained 49.6% of the total variation (Fig. 4a). by 1 OTU (OTU 17) of this study and several sequences The relative abundance of subgroups was correlated with retrieved from the NCBI database (Fig. 3). MG-Im was salinity and pH (P < 0.05). Moreover, their distributions predominant in the saltwater sites of this study. MG-Ik II were also related to N-nutrients (P<0.05). Most OTUs was the most abundant clade in the freshwater sites, fol- of the MG-Ia I were inversely correlated to N-nutrients, lowed by MG-Ia II-1. Most of the MG-Ik II and MG-Ia whereas MG-Ik II was positively correlated. The other II-1 sequences were obtained from freshwater systems. freshwater subgroup, MG-Ia II-1, did not seem to be dri- Whereas MG-Ik II mainly derived from anoxic subsurface ven by the environmental factors analyzed in the overall waters, MG-Ia II-1 mainly derived from oxic river/lake RDA. Since many environmental factors in estuarine sys- systems. Among MG-Ia I, members spanned different tems, such as DO, pH and turbidity, showed similar or salinity ranges, with more present in the saltwater sites opposite trends to salinity, the overwhelming variation of (Fig. 4a). Additionally, the ancestral phylogenetic position salinity might obscure the impacts of other factors.

Fig. 3. Neighbor-joining phylogenetic tree showing the subgroups of MG-I with SAGMCG-1 and SCG (OTU28) serving as outgroups. The phylogenetic tree was constructed with sequences from this study, the top-hit sequences retrieved from NCBI and sequences cloned from some freshwater or coastal water systems that afﬁliated to MG-I. The outer color circle around the phylogenetic tree suggests the different habitats. The inner color circle suggests the sub-habitats of seawater and freshwater. The nomenclatures follows previous studies.

Therefore, the freshwater MG-I subgroups were analyzed found only in saltwater sites. The MG-II sequences of this alone with the factors that varied only in freshwater sites study accounted for nine subgroups with bootstrap values (Fig. 4b). The result showed that RDA1 and RDA2 could all greater than 95%, as shown in the phylogenetic tree explain 63.4% of the total variation and that DO (Fig. 5). Within the MG-II OTUs, sequences from OTU 2 (P<0.05) was the most signiﬁcant factor shaping the and 7 were mainly from the surface water, whereas those distribution of MG-I subgroups. Although both MG-Ia from OTU 4 were mainly from the bottom water. The II-1 and MG-Ik II were abundant in the freshwater sites, differences in salinity between the surface and bottom they varied inversely with DO; whereas MG-Ia II-1 was waters contributed the most to their heterogeneous distri- positively correlated to DO, MG-Ik II was negatively cor- bution (P<0.05). However, OTUs within a single sub- related. group showed different vertical distributions.

Diversity of the MG-II Discussion Nineteen OTUs were affiliated to MG-II and, unlike the A significant finding of the present study was that fresh- generally distributed MG-I archaea, MG-II archaea were water-specific MG-I subgroups inhabited the sites

(a) (b) Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021

Fig. 4. Redundancy analysis (RDA) diagram illustrating (a) the relationship between MG-I subgroups with environmental factors in all sites and (b) the relationship between three major MG-I subgroups with environmental factors in only freshwater sites. The environmental factors and MG- I subgroups are indicated by red arrows and different shapes, respectively. RDA analysis was performed with CANOCO software (version 4.5). upstream of Pearl Estuary, which were, as shown in the et al., 2012; Auguet & Casamayor, 2013; Cao et al., phylogenetic trees, distant from the subgroups associated 2013). However, since the environmental factors of salin- with sediment and deep water (Takai et al., 2004; Dur- ity, pH and nutrients are highly correlated with each bin & Teske, 2010). Some freshwater-specific Cre- other in estuarine areas, it is difficult to judge which is narchaeota/Thaumarchaeota groups have been delineated. the true influencing factor. Additionally, members of For example, the Freshwater Group I.1a clade from an MG-Ia I were found to inhabit a wide range of salinity. Arctic lake (Galand et al., 2008) was phylogenetically The aquatic MG-Ia I components were phylogenetically related to MG-Ia I of this study, whereas the Freshwater distant from those of sediment (Fig. 3), but these MG-Ia 1.1a clade from a high mountain lake (Auguet & Casa- I subclusters displayed very low bootstrap values and mayor, 2008) was more related to the SAGMCG-1 group therefore we did not name them. (Fig. 3). Results of the present study suggested that clus- MG-I related lineages have been demonstrated to prefer ters of MG-I could adjust to specific niches partitioned suboxic/microoxic habitat (Durbin & Teske, 2012) and by different environmental conditions, which raises ques- the shift of MG-I subgroups was observed accompanying tion of what determining factors shape the distribution the change of oxygen in oligotrophic sediments (Durbin of MG-I subgroups. & Teske, 2010). Results of the present study also high- Ammonia is an important factor that influences the lighted the importance of DO in governing the distribu- distribution of AOA (Herfort et al., 2007; Liu et al., 2011; tion of the two freshwater subgroups (MG-Ia II-1 and Sintes et al., 2013). In this study, MG-Ik II and MG-Ia MG-Ik II), which exhibited opposite correlation with II-1 were the two MG-I (major components of AOA) DO, indicating that these two lineages evolved to accom- subgroups abundant in the freshwater sites of Pearl Estu- modate different water oxygenation. In the upstream of ary, and they appeared to be inversely related to salinity. Pearl Estuary, nitrification has been demonstrated to con- On the other hand, MG-Ik II, but not MG-Ia II-1, tribute substantially to DO depletion (Dai et al., 2006, showed a significantly positive correlation with N-nutri- 2008). The predominance of MG-I clade in this area thus + ents (NO3 ,NO2 and NH4 ), suggesting that this sub- might play an important role in mediating the aerobic group of MG-I might favor environments with high ammonia oxidation, the first and rate-limiting step of nutrient levels. The freshwater and saltwater sites of Pearl nitrification, in the Pearl River, since few known ammo- Estuary were inhabited by different MG-I subgroups, nia-oxidization bacteria were obtained from our study which could be attributed to the significant differences in (data not shown). A finer resolution of which subgroup salinity and pH between the two regions (Fig. 4), consis- plays the most significant role in nitrification and the tent with previous studies regarding the significance of extent to which AOA could function under hypoxic con- salinity and pH in driving the distribution of AOA (Biller dition of Pearl Estuary remains to be explored.

Fig. 5. Neighbor-joining phylogenetic tree showing the subgroups of MG-II and other lineages (of the top 50 OTUs) detected in this study belonging to Euryarchaeota. Bootstrap values greater than 50 are shown. The OTU sequences of the present study are shown in bold. The sequence numbers for each OTU are given in parentheses.

There are several successful enrichment and isolated 2012). Jorgensen et al. (2012) grouped the MG-I cultures of AOA genera, Nitrosopumilus (Konneke€ et al., sequences retrieved from the SILVA database into several 2005), Nitrososphaera (Hatzenpichler et al., 2008), Nitros- subgroups and divided MG-Ia into four clusters, MG-Ia ocaldus (de la Torre et al., 2008), Nitrosotalea (Lehtovirta- I to IV. According to the topology of both the neighbor- Morley et al., 2011) and Nitrosoarchaeum (Mosier et al., joining and maximum-likelihood trees of this study,

MG-Ia II was further divided into two clusters, MG-Ia and freshwater flowing out of the estuary at the surface II-1 and II-2. The sequences of MG-Ia II-1 were phyloge- (Dong et al., 2004), to which the vertical variation prob- netically affiliated to Nitrosoarchaeum and were obtained ably could be attributed. The differences were mainly mainly from oxic river/lake systems (Mosier et al., 2012). mirrored by the differential distribution of OTUs In contrast, MG-Ia II-2 consisted of MG-I archaea associ- belonging to MG-II, with some OTUs dominating the ated with marine sponges (Fig. 3). MG-I are major archa- surface water and others the bottom water. Diversifica- eal components associated with marine sponges (Preston tion of the MG-II OTUs was also observed, and several et al., 1996; Holmes & Blanch, 2007; Kennedy et al., new clusters were designated, expanding the subgroups 2014) and may be involved in ammonia oxidation (Hoff- designated by Takai et al. (2004). However, OTUs mann et al., 2009). This suggests that the sponge-associ- within the same subgroups of MG-II showed inconsis- ated archaea were phylogenetically distinct from those tent vertical variations. Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 present in ambient seawater and marine sediment (Fig. 3; As the majority of microorganisms in the environ- Turque et al., 2010). ment are uncultivated at present, the significance of The phylogenetic depth of the MG-Ik subgroup was molecular ecological methods in elucidating the distribu- comparable to that of SAGMCG-1 and SCG (Fig. 3, Fig. tion patterns of these microbes was evident (Lozupone S3), indicating that MG-Ik might be an ancestor clade of & Knight, 2007; Auguet et al., 2010). Previous studies MG-I. MG-Ik I and MG-Ik II are two clusters of MG-Ik have reported that Euryarchaeota such as RC-V and subgroup defined by Jorgensen et al. (2012), and both are LDS (phylogenetic to DHVEG-6), methanogenic archa- found in freshwater systems. MG-Ik II was a freshwater eon and thermoplasmatales (Gonzalez-Toril et al., 2003; clade dominant in the hypoxic sites of Pearl Estuary, and Galand et al., 2006; Vieira et al., 2007) and Thaumarch- its representatives were retrieved mainly from under- aeota such as MG-I (Crump & Baross, 2000; Herfort ground water including aquifers and various deep mines et al., 2009) were abundant in various river systems. (Fig. 3), which are usually environments lacking oxygen. The different habitat types could be an important rea- Thus, it could be inferred that the groundwater discharge son for this discrepancy of the archaeal community in of the Pearl River has significant effects on structuring the different river systems (Auguet et al., 2010; Biller et al., downstream microbial assemblages, and that MG-Ik II 2012). As more and more archaeal clades and subgroups might represent a AOA clade functioning under low salin- are detected, there is clearly a need for more in-depth ity and low DO levels. It is noticeable that sequences of studies to determine the metabolic characteristics of MG-Ik I were retrieved mainly from high-temperature these abundant clades under distinct environmental con- waters such as a geothermal spring and Yellowstone lake. ditions. This, together with the ancestral position of MG-Ik Iat both the phylogenetic trees, led to the hypothesis that the Conclusions modern mesophilic MG-I might have originated from a thermophilic group, which needs further consolidation. Using high-throughput sequencing, an environmentally The different subgroups of MG-I found in different habi- shaped variability in archaeal community, especially MG-I, tats and under distinct environmental conditions highlight was observed in Pearl Estuary. MG-I were the most dom- the diversification of MG-I groups. Their evolutionary inant clade in the freshwater sites. Following the nomen- history and physiological characters thus need further clature of previous studies, we proposed one novel investigation. subgroup of MG-I, MG-Im, and divided the previously Evolutionary separation of archaeal communities has described MG-Ia II into two clusters: MG-Ia II-1 and been observed between marine and freshwater areas of MG-Ia II-2. MG-Ia II-1 and MG-Ik II were predominant estuarine ecosystems (Crump & Baross, 2000; Vieira in the freshwater sites, and DO was the determining fac- et al., 2007; Galand et al., 2008). Similar to the present tor driving their heterogeneous distributions. MG-Im,as study, clades of MG-I and MG-II were abundant in the well as MG-Ia I, was predominant in the saltwater sites. freshwater and marine areas, respectively, in a tropical MG-II were also prevalent in the saltwater sites, for which environment (Silveira et al., 2013). At the same time, phylotype diversification was also observed. Meanwhile, our study pointed to an obvious vertical variation of ar- the abundance of OTUs of MG-II differed according to chaeal communities in the saltwater sites of Pearl Estu- depth, but this difference was not observed within the ary. Although the water depth of Pearl Estuary was same subgroup. The present study showed that phyloge- shallow, the surface and bottom waters in the saltwater netically related MG-I ecotypes inhabited different envi- sites were dominated by two different water masses, ronmental conditions, and the functions of these ecotypes with marine water entering the estuary at the bottom need to be further elucidated.

DeLong EF (1992) Archaea in coastal marine environments. P Acknowledgements Natl Acad Sci USA 89: 5685–5689. We thank all of the scientists and crew members on the Dong L, Su J, Ah Wong L, Cao Z & Chen J-C (2004) Seasonal R/V Tian Long during the expedition for their great variation and dynamics of the Pearl River plume. Cont Shelf – efforts and help in sample collection. We also thank Min- Res 24: 1761 1777. han Dai of Xiamen University for organizing this expedi- Durbin AM & Teske A (2010) Sediment-associated microdiversity within the Marine Group I Crenarchaeota. tion and Jianping Gan of Hong Kong University of Environ Microbiol Rep 2: 693–703. Science and Technology for providing CTD records. This Durbin AM & Teske A (2012) Archaea in organic-lean and work was supported by the National Natural Science organic-rich marine subsurface sediments: an environmental Foundation of China through grants 41276141, 41221004 gradient reflected in distinct phylogenetic lineages. Front and 41130857. Microbiol 3: 168. Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 Edgar RC, Haas BJ, Clemente JC, Quince C & Knight R References (2011) UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27: 2194–2200. Auguet JC & Casamayor EO (2008) A hotspot for cold Erguder TH, Boon N, Wittebolle L, Marzorati M & Verstraete crenarchaeota in the neuston of high mountain lakes. W (2009) Environmental factors shaping the ecological Environ Microbiol 10: 1080–1086. niches of ammonia-oxidizing archaea. FEMS Microbiol Rev Auguet JC & Casamayor EO (2013) Partitioning of 33: 855–869. Thaumarchaeota populations along environmental gradients Fuhrman JA, McCallum K & Davis AA (1992) Novel major in high mountain lakes. FEMS Microbiol Ecol 84: 154–164. archaebacterial group from marine plankton. Nature 356: Auguet JC, Barberan A & Casamayor EO (2010) Global 148–149. ecological patterns in uncultured Archaea. ISME J 4: 182–190. Galand PE, Lovejoy C & Vincent WF (2006) Remarkably Bernhard AE, Landry ZC, Blevins A, de la Torre JR, Giblin AE diverse and contrasting archaeal communities in a large & Stahl DA (2010) Abundance of ammonia-oxidizing arctic river and the coastal Arctic Ocean. Aquat Microb Ecol archaea and bacteria along an estuarine salinity gradient in 44: 115–126. relation to potential nitrification rates. Appl Environ Galand PE, Lovejoy C, Pouliot J, Garneau ME & Vincent WF Microbiol 76: 1285–1289. (2008) Microbial community diversity and heterotrophic Biller SJ, Mosier AC, Wells GF & Francis CA (2012) Global production in a coastal Arctic ecosystem: Astamukhi lake biodiversity of aquatic ammonia-oxidizing archaea is and its source waters. Limnol Oceanogr 53: 813–823. partitioned by habitat. Front Microbiol 3: 252. Gilbert JA, Field D, Swift P, Newbold L, Oliver A, Smyth T, Brown MV, Philip GK, Bunge JA, Smith MC, Bissett A, Lauro Somerfield PJ, Huse S & Joint I (2009) The seasonal FM, Fuhrman JA & Donachie SP (2009) Microbial structure of microbial communities in the Western English community structure in the North Pacific Ocean. ISME J 3: Channel. Environ Microbiol 11: 3132–3139. 1374–1386. Gonzalez-Toril E, Llobet-Brossa E, Casamayor EO, Amann R Cao H, Auguet JC & Gu JD (2013) Global ecological pattern & Amils R (2003) Microbial ecology of an extreme acidic of ammonia-oxidizing archaea. PLoS ONE 8: e52853. environment, the Tinto River. Appl Environ Microbiol 69: Casamayor EO, Massana R, Benlloch S et al. (2002) Changes 4853–4865. in archaeal, bacterial and eukaryal assemblages along a Good IJ (1953) The population frequencies of species and the salinity gradient by comparison of genetic fingerprinting estimation of population parameters. Biometrika 40: 237– methods in a multipond solar saltern. Environ Microbiol 4: 264. 338–348. Han A, Dai M, Kao S-J, Gan J, Li Q, Wang L, Zhai W & Chao A & Bunge J (2002) Estimating the number of species in Wang L (2012) Nutrient dynamics and biological a stochastic abundance model. Biometrics 58: 531–539. consumption in a large continental shelf system under the Crump BC & Baross JA (2000) Archaeaplankton in the influence of both a river plume and coastal upwelling. Columbia River, its estuary and the adjacent coastal ocean, Limnol Oceanogr 57: 486–502. USA. FEMS Microbiol Ecol 31: 231–239. Hatzenpichler R, Lebedeva EV, Spieck E, Stoecker K, Richter Dai M, Guo X, Zhai W, Yuan L, Wang B, Wang L, Cai P, A, Daims H & Wagner M (2008) A moderately Tang T & Cai W-J (2006) Oxygen depletion in the upper thermophilic ammonia-oxidizing crenarchaeote from a hot reach of the Pearl River estuary during a winter drought. spring. P Natl Acad Sci USA 105: 2134–2139. Mar Chem 102: 159–169. Herfort L, Schouten S, Abbas B, Veldhuis MJW, Coolen MJL, Dai M, Wang L, Guo X, Zhai W, Li Q, He B & Kao S-J Wuchter C, Boon JP, Herndl GJ & Damste JSS (2007) (2008) Nitrification and inorganic nitrogen distribution in a Variations in spatial and temporal distribution of Archaea large perturbed river/estuarine system: the Pearl River in the North Sea in relation to environmental variables. Estuary, China. Biogeosciences 5: 1545–1585. FEMS Microbiol Ecol 62: 242–257.

Herfort L, Kim JH, Coolen MJL, Abbas B, Schouten S, Herndl Pester M, Rattei T, Flechl S, Grongroft A, Richter A, GJ & Damste JSS (2009) Diversity of Archaea and detection Overmann J, Reinhold-Hurek B, Loy A & Wagner M (2012) of crenarchaeotal amoA genes in the rivers Rhine and Tet. amoA-based consensus phylogeny of ammonia-oxidizing Aquat Microb Ecol 55: 189–201. archaea and deep sequencing of amoA genes from soils of Hoffmann F, Radax R, Woebken D, Holtappels M, Lavik G, four different geographic regions. Environ Microbiol 14: Rapp HT, Schlappy ML, Schleper C & Kuypers MM (2009) 525–539. Complex nitrogen cycling in the sponge Geodia barretti. Preston CM, Wu KY, Molinski TF & DeLong EF (1996) A Environ Microbiol 11: 2228–2243. psychrophilic crenarchaeon inhabits a marine sponge: Holmes B & Blanch H (2007) Genus-speciﬁc associations of Cenarchaeum symbiosum gen. nov., sp. nov. P Natl Acad Sci marine sponges with group I crenarchaeotes. Mar Biol 150: USA 93: 6241–6246. 759–772. PRWRC/PRRCC (1991) The Pearl River Records (Zhujiang

Hu A, Jiao N & Zhang CL (2011) Community structure and Zhi). Guangdong Science and Technology Press: Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 function of planktonic Crenarchaeota: changes with depth Guangzhou, pp. 162–165 (in Chinese). in the South China Sea. Microb Ecol 62: 549–563. RDC Team (2008) R: A Language and Environment for Inagaki F, Suzuki M, Takai K, Oida H, Sakamoto T, Aoki K, Statistical Computing. R Foundation for Statistical Nealson KH & Horikoshi K (2003) Microbial communities Computing, Vienna. associated with geological horizons in coastal subseafloor Santoro AE, Francis CA, de Sieyes NR & Boehm AB (2008) sediments from the Sea of Okhotsk. Appl Environ Microbiol Shifts in the relative abundance of ammonia-oxidizing 69: 7224–7235. bacteria and archaea across physicochemical gradients in a Jorgensen SL, Hannisdal B, Lanzen A et al. (2012) Correlating subterranean estuary. Environ Microbiol 10: 1068–1079. microbial community profiles with geochemical data in Schleper C, Holben W & Klenk HP (1997) Recovery of highly stratified sediments from the Arctic Mid-Ocean crenarchaeotal ribosomal DNA sequences from Ridge. P Natl Acad Sci USA 109: E2846–E2855. freshwater-lake sediments. Appl Environ Microbiol 63: Kennedy J, Flemer B, Jackson SA, Morrissey JP, O’Gara F & 321–323. Dobson AD (2014) Evidence of a putative deep sea specific Schloss PD, Westcott SL, Ryabin T et al. (2009) microbiome in marine sponges. PLoS ONE 9: e91092. Introducing mothur: open-source, platform-independent, Kirchman DL, Dittel AI, Malmstrom RR & Cottrell MT (2005) community-supported software for describing and Biogeography of major bacterial groups in the Delaware comparing microbial communities. Appl Environ Microbiol Estuary. Limnol Oceanogr 50: 1697–1706. 75: 7537–7541. Konneke€ M, Bernhard AE, de la Torre JR, Walker CB, Silveira CB, Cardoso AM, Coutinho FH, Lima JL, Pinto LH, Waterbury JB & Stahl DA (2005) Isolation of an Albano RM, Clementino MM, Martins OB & Vieira RP autotrophic ammonia-oxidizing marine archaeon. Nature (2013) Tropical aquatic Archaea show environment-specific 437: 543–546. community composition. PLoS ONE 8: e76321. Lehtovirta-Morley LE, Stoecker K, Vilcinskas A, Prosser JI & Sintes E, Bergauer K, De Corte D, Yokokawa T & Herndl GJ Nicol GW (2011) Cultivation of an obligate acidophilic (2013) Archaeal amoA gene diversity points to distinct ammonia oxidizer from a nitrifying acid soil. P Natl Acad biogeography of ammonia-oxidizing Crenarchaeota in the Sci USA 108: 15892–15897. ocean. Environ Microbiol 15: 1647–1658. Liu ZH, Huang SB, Sun GP, Xu ZC & Xu MY (2011) Sørensen KB, Lauer A & Teske A (2004) Archaeal phylotypes in Diversity and abundance of ammonia-oxidizing archaea in a metal-rich and low-activity deep subsurface sediment of the the Dongjiang River, China. Microbiol Res 166: 337–345. Peru Basin, ODP Leg 201, Site 1231. Geobiology 2 : 151–161. Lozupone CA & Knight R (2007) Global patterns in bacterial Takai K, Oida H, Suzuki Y, Hirayama H, Nakagawa S, diversity. P Natl Acad Sci USA 104: 11436–11440. Nunoura T, Inagaki F, Nealson KH & Horikoshi K (2004) Magurran AE (1988) Ecological Diversity and Its Measurement. Spatial distribution of marine crenarchaeota group I in the Princeton University Press, Princeton, NJ. vicinity of deep-sea hydrothermal systems. Appl Environ Massana R, DeLong EF & Pedros-Alio C (2000) A few Microbiol 70: 2404–2413. cosmopolitan phylotypes dominate planktonic archaeal Tamura K, Peterson D, Peterson N, Stecher G, Nei M & Kumar assemblages in widely different oceanic provinces. Appl S (2011) MEGA5: molecular evolutionary genetics analysis Environ Microbiol 66: 1777–1787. using maximum likelihood, evolutionary distance, and Mosier AC & Francis CA (2008) Relative abundance and maximum parsimony methods. Mol Biol Evol 28: 2731–2739. diversity of ammonia-oxidizing archaea and bacteria in the de la Torre JR, Walker CB, Ingalls AE, Konneke M & Stahl San Francisco Bay estuary. Environ Microbiol 10: 3002–3016. DA (2008) Cultivation of a thermophilic ammonia Mosier AC, Allen EE, Kim M, Ferriera S & Francis CA oxidizing archaeon synthesizing crenarchaeol. Environ (2012) Genome sequence of ‘Candidatus Nitrosoarchaeum Microbiol 10: 810–818. limnia’ BG20, a low-salinity ammonia-oxidizing archaeon Turque AS, Batista D, Silveira CB et al. (2010) Environmental from the San Francisco Bay estuary. J Bacteriol 194: shaping of sponge associated archaeal communities. PLoS 2119–2120. ONE 5: e15774.

Vieira RP, Clementino MM, Cardoso AM, Oliveira DN, Fig. S1. Location of sampling sites in Pearl Estuary. Albano RM, Gonzalez AM, Paranhos R & Martins OB Fig. S2. The heatmap generated in R with function ‘heat- (2007) Archaeal communities in a tropical estuarine map.2’ of the top 50 abundant OTUs of all samples with ecosystem: Guanabara Bay, Brazil. Microb Ecol 54: 460–468. their phylogenetic relationship shown on the right tree Yin K, Lin Z & Ke Z (2004) Temporal and spatial distribution (constructed by FastTree using maximum-likelihood of dissolved oxygen in the Pearl River Estuary and adjacent method). coastal waters. Cont Shelf Res 24: 1935–1948. Fig. S3. Maximum-likelihood phylogenetic tree showing Yin Q, Fu B, Li B, Shi X, Inagaki F & Zhang XH (2013) the subgroups of MG-I with SAGMCG-1 and SCG Spatial variations in microbial community composition in (OTU28) serving as the outgroups. surface seawater from the ultra-oligotrophic center to rim of the South paciﬁc gyre. PLoS ONE 8: e55148. Table S1. The original reads, normalized observed OTUs, Good’s coverage, species estimator Chao1 and Shannon Downloaded from https://academic.oup.com/femsec/article/90/2/424/2680468 by guest on 01 October 2021 diversity index of the 12 samples at 3% distance level Supporting Information

Additional Supporting Information may be found in the online version of this article: