Environmental Microbiology 8, 684-696.Pdf

Home , Nitrifying bacteria, Nitrosomonas, Nitrosomonas aestuarii, Nitrosomonas europaea, Nitrosomonas eutropha, Nitrosomonas marina

Blackwell Science, LtdOxford, UKEMIEnvironmental Microbiology 1462-2912© 2005 The Authors; Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd ?? 200584684696ArticleOriginalA-mo nia oxidizers in a freshwater–marine estuarine gradientT. E. Freitag, L. Chang and J. I. Prosser

Environmental Microbiology (2006) 8(4), 684–696 doi:10.1111/j.1462-2920.2005.00947.x

Changes in the community structure and activity of betaproteobacterial ammonia-oxidizing sediment bacteria along a freshwater–marine gradient

Thomas E. Freitag, Lisa Chang and James I. Prosser* nas cryotolerans and to the nitrite oxidizer Nitrospira School of Biological Sciences, University of Aberdeen, marina. No sequence with similarity to the Nitrosos- Cruickshank Building, St Machar Drive, Aberdeen AB24 pira cluster 1 clade was recovered during SIP analy- 3UU, UK. sis. The potential role of Nitrosospira cluster 1 in autotrophic ammonia oxidation therefore remains uncertain. Summary To determine whether the distribution of estuarine Introduction ammonia-oxidizing bacteria (AOB) was influenced by salinity, the community structure of betaproteo- The intertidal areas in coastal zones include some of the bacterial ammonia oxidizers (AOB) was characterized most productive marine ecosystems and are thus vital as along a salinity gradient in sediments of the Ythan breeding and feeding grounds for many species of birds, estuary, on the east coast of Scotland, UK, by dena- fish and crustaceans. These areas are under threat from turant gradient gel electrophoresis (DGGE), cloning anthropogenic activities, through land reclamation, fisher- and sequencing of 16S rRNA gene fragments. Ammo- ies and in particular through eutrophication caused by nia-oxidizing bacteria communities at sampling sites nutrient input through terrestrial runoff of nitrogen. Estua- with strongest marine influence were dominated by rine intertidal systems are especially vulnerable because Nitrosospira cluster 1-like sequences and those with the river catchments in areas of intensive farming can strongest freshwater influence were dominated by result in the transport of highly enriched nutrients (Raf- Nitrosomonas oligotropha-like sequences. Nitroso- faelli, 1999). The low species diversity of brackish-water monas sp. Nm143 was the prevailing sequence type estuarine ecosystems (Remane and Schlieper, 1971; in communities at intermediate brackish sites. Diver- Attrill, 2002) may additionally enhance effects of elevated sity indices of AOB communities were similar at nutrient levels and promote growth of antagonistic phy- marine- and freshwater-influenced sites and did not toplankton or macroalgal species, resulting in low- indicate lower species diversity at intermediate brack- diversity, high-density ecosystems. Nitrification, the con- ish sites. The presence of sequences highly similar to version of ammonia to nitrate via nitrite, is central to the the halophilic Nitrosomonas marina and the freshwa- global cycling of nitrogen and, when linked to anoxic envi- – ter strain Nitrosomonas oligotropha at identical sam- ronments, where NO3 can serve as alternative electron pling sites indicates that AOB communities in the acceptor for denitrification, is a major regulating factor estuary are adapted to a range of salinities, while alleviating eutrophication processes (Kemp et al., 1990; individual strains may be active at different salinities. Rysgaard et al., 1994). However, nitrification processes Ammonia-oxidizing bacteria communities that were depend on the availability of free dissolved oxygen and dominated by Nitrosospira cluster 1 sequence types, may be totally eliminated if hypoxia occurs through for which no cultured representative exists, were sub- eutrophication (Kemp et al., 1990). 13 – jected to stable isotope probing (SIP) with C-HCO3 , The first, usually rate-limiting oxidation step in nitrifica- to label the nucleic acids of active autotrophic nitrifi- tion, is carried out by ammonia-oxidizing bacteria (AOB), ers. Analysis of 13C-associated 16S rRNA gene which can also reduce nitrite to nitric oxide (NO) and fragments, following CsCl density centrifugation, by nitrous oxide (N2O) at low oxygen tension. Ammonia- cloning and DGGE indicated sequences highly similar oxidizing bacteria consist of three evolutionary distant to the AOB Nitrosomonas sp. Nm143 and Nitrosomo- groups, two of which belong to the class Proteobacteria (Head et al., 1993; Koops et al., 2003). One group forms a deep branch within the Gammaproteobacteria and com- Received 20 June, 2005; accepted 27 September, 2005. *For corre- spondence. E-mail [email protected]; Tel. (+44) 1224 273254; prises only three recognized, closely related marine Fax (+44) 1224 272703. Nitrosococcus species. The second group, which includes

Ammonia oxidizers in a freshwater–marine estuarine gradient 685 the majority of cultured strains, forms a monophyletic estuarine environment, the Ythan estuary, for which data group within the Betaproteobacteria and consists of two on seasonal and spatial distributions of salinity, hydrogra- genera, Nitrosomonas and Nitrosospira. The third group, phy, nutrients and flushing times are available (Balls et al., belonging to the order Planctomycetales, is associated 1995; Gillibrand and Balls, 1998; Raffaelli et al., 1999). with the Anammox process and its importance in marine Ammonia-oxidizing bacteria communities were analysed systems has been suggested, particularly for oxic–anoxic by amplification of 16S rRNA gene fragments from envi- interfaces (Jetten et al., 2003). Molecular studies of AOB ronmental DNA and subsequent analysis by denaturing communities in freshwater and estuarine systems suggest gradient gel electrophoresis (DGGE), cloning, sequencing the dominance of betaproteobacterial AOB (Caffrey et al., and phylogenetic analysis. These techniques were com- 2003). Laboratory isolates of freshwater and marine bined with stable isotope probing (SIP; Radajewski et al., betaproteobacterial AOB laboratory isolates (Koops and 2000; Lueders et al., 2004) to characterize the active AOB Pommerening-Röser, 2001), and AOB 16S rRNA gene community. Application of SIP to determine AOB activity sequences from freshwater and estuarine environments involves incubation of environmental samples with 13C- 12 13 (Speksnijder et al., 1998; de Bie et al., 2001; Caffrey labelled CO2 and fractionation of extracted C- and C- et al., 2003) fall predominantly within Nitrosomonas lin- labelled nucleic acids. Molecular analysis of labelled and eages. In marine systems, prevalent sequence types are unlabelled nucleic acids characterizes active and inactive associated with the Nitrosomonas eutropha lineage (Phil- community members, respectively, and thereby provides lips et al., 1999) or belong to the Nitrosomonas cluster 5 a more reliable measure of links between community or Nitrosospira cluster 1 clone groups, for which no cul- structure and ecosystem function. tured representative has yet been isolated (McCaig et al., 1999; Hollibaugh et al., 2002; Freitag and Prosser, 2003; 2004). Results Salt requirement is a major ecophysiological parameter 16S rRNA gene analysis of environmental AOB in determining the niche specialization of many bacteria. communities Characterization of both laboratory isolates (Koops and Pommerening-Röser, 2001; Koops et al., 2003) and clone Denaturing gradient gel electrophoresis analysis of the libraries (Stephen et al., 1996) suggests that the influence CTO189f-CTO654r 16S rRNA gene polymerase chain of salt concentration is a distinguishing feature within both reaction (PCR) products nested with 357f-GC-518r, span- Nitrosomonas and Nitrosospira lineages. This suggests ning the hypervariable V3 region only, produced almost that patterns of diversity, community composition and identical patterns for replicate samples from each sam- activity of different sequence types of AOB will differ in pling site, demonstrating more than 30 clearly distinguish- estuaries in response to salinity gradients. The aim of this able sequence types (Fig. 1A). Denaturing gradient gel study was to test this hypothesis in a well-characterized electrophoresis migration patterns varied significantly

A Ma ri ne Freshwater B Salinity 28 16 11 10 0 1 Eb 2 freshwater ig o. ig o. 3 4 9 Ea ol ol 5 Nm ma r. / r. Nm ma Db 5 5 Da Cl Cl 10 Cb 6 11 brackish 143 Ca 7 B

Nm b Ba

1 12 A b marine Cl 8 13 Aa p

Ns 14 15

CL 1 CL 2 A B C D E Sa mpli ng lo cati on s

Fig. 1. Analysis of betaproteobacterial ammonia oxidizer communities in Ythan estuary sediments at sampling locations indicated in Fig. 5, with salinities ranging from close to seawater (A) to freshwater (E). Betaproteobacterial AOB-like 16S rRNA gene sequences were amplified from extracted DNA using CTO189f and CTO654r PCR primers, nested with 357f-GC and 518r. Closest BLAST matches of clone sequences shown in lanes CL1 and CL2 (marked by arrows) are shown in Table 2. A. Denaturant gradient gel electrophoresis analysis of amplification products. B. UPGMA dendrogram assessing the similarity of DGGE profiles illustrated in (A). Salinities (‰) according to Gillibrand and Balls (1998), calculated from conductivity measurements.

686 T. E. Freitag, L. Chang and J. I. Prosser

Table 1. Diversity indices obtained for DGGE migration patterns of AOB 16S rRNA gene fragments from duplicate Ythan estuary sediment samples.

Sampling site

Diversity index AB CDE

Shannon 2.67 (± 0.15) 2.85 (± 0.36) 2.95 (± 0.10) 2.84 (± 0.08) 2.91 (± 0.007) Evenness 0.86 (± 0.14) 0.95 (± 0.040) 0.929 (± 0.03) 0.95 (± 0.03) 0.97 (± 0.002)

between the sampling sites. Dominant bands of different characteristics to dominant environmental DGGE bands migration behaviour were present at each site and indi- were sequenced and closest relatives determined by cated a general shift from 16S rRNA gene fragments with comparison with GenBank database sequences, using relatively low migration at freshwater sites to higher migra- BLAST (Table 2). Putative identifications of individual tion at marine sites. These differences are reflected in the bands are indicated in Fig. 1A and show a shift from unweighted pair group method with arithmetic mean Nitrosospira cluster 1-associated sequences (clones 8 (UPGMA) dendrogram, based on a similarity matrix calcu- and 12–15; Fig. 1A), in samples with highest marine influ- lated from the presence/absence of DGGE bands. Pro- ence, to Nitrosomonas marina- and Nitrosomonas olig- files from sites A and E were least similar, while those for otropha-associated sequences (clones 1–5 and 9; Fig. 1), sites B, C and D clustered together with higher similarity with increasing freshwater influence. The AOB sediment (Fig. 1B). Ammonia-oxidizing bacteria richness, in terms community of marine station A was dominated by four of the number of different DGGE bands, was similar (8– sequence types associated with Nitrosospira cluster 1 9) for all sampling sites. Evenness and Shannon diversity with several faint bands matching sequences from index values were similar (0.929–0.97 and 2.84–2.94 N. marina/oligotropha lineages also detectable. The AOB respectively; Table 1) for sampling sites B–E, but even- community in station B was also strongly influenced by ness at marine station A (0.864) was outside the standard Nitrosospira cluster 1 sequence types but Nitrosomonas deviation range for all other samples and Shannon diver- cluster 5-, Nitrosomonas sp. Nm143- and N. marina/olig- sity index (2.666) was outside the standard deviation otropha-associated sequences types were present at range for all other samples except station B. similar relative abundances. Intermediate brackish station To determine the identity of DGGE bands, clone librar- C profiles were dominated by a Nitrosomonas sp. ies were obtained using CTO189f-CTO654r primers with Nm143-associated sequence type and three of the Nitro- secondary amplification using 357f-GC-518r primers sospira cluster 1-associated bands were no longer (see above). CTO clones showing identical migration detectable. The same sequence types were present at

Table 2. CTO189f-CTO654r 16S rRNA gene sequence fragments (0.45 kb) ampliﬁed from Ythan estuary sediments identiﬁed by BLAST closest matches.

Representative Closest relative BLAST closest match clone/library Name/putative associated lineage Reference accession number % identity

CL1-1/E Clone VM4/Nitrosomonas oligotropha cluster 6a-A 5, 7, 1 AJ003756 98

CL1-2/E Nitrosomonas aestuarii/Nitrosomonas marina cluster 6b 1 AF272420 98

CL1-3/C Nitrosomonas sp. C-113/Nitrosomonas marina cluster 6b 1 AF338200 96

CL1-4/E Nitrosomonas sp. 86/Nitrosomonas oligotropha cluster 6a 1 AJ621026 99

CL1-5/E Clone VM4/Nitrosomonas oligotropha cluster 6a 5, 7, 1 AJ003756 98

CL1-6/C, D Clone LD1-B6/Nitrosomonas sp. 143 2 AY114347 98

CL1-7/C Clone LD1-B6/Nitrosomonas sp. 143 2 AY114347 98

CL1-8/E Clone LD1-B28/Nitrosospira cluster 1 2 AY114345 98

CL2-9/B Nitrosomonas sp. R7c140/Nitrosomonas marina 6 AF386753 96

CL2-10/B Clone EnvA1-21/Nitrosomonas cluster 5 3 Z69091 98

CL2-11/B Clone LD1-B6/Nitrosomonas sp. 143 3 AY114347 98

CL2-12/B Clone DT-1.3/Nitrosospira cluster 1 4 U62867 99

CL2-13/A, B Clone LD1-A1/Nitrosospira cluster 1 2 AY114337 99

CL2-14/A Clone LD1-A10/Nitrosospira cluster 1 2 AY114348 99

CL2-15/A Clone LD1-A10/Nitrosospira cluster 1 2 AY114348 98

Clone sequences were assembled from libraries generated from each sampling station (A–E) as DGGE marker lanes (CL1 and CL2; Fig. 1) on the basis of co-migration with DGGE bands in environmental samples. References: 1, Purkhold and colleagues (2003); 2, Freitag and Prosser (2003); 3, McCaig and colleagues (1999); 4, Kowalchuk and colleagues (1997); 5, Speksnijder and colleagues (1998); 6, Bollmann and Laanbroek (2001); 7, Burrell and colleagues (2001).

© 2005 The Authors Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 8, 684–696 Ammonia oxidizers in a freshwater–marine estuarine gradient 687 comparable relative abundances at station D but with the cells. Use of ethidium bromide at a concentration lower additional presence of a band associated with the than those previously reported (Radajewski et al., 2000) N. marina lineage. The AOB communities in freshwater increased sensitivity of UV visibility and sharpness of 12C station E were characterized by five dominant sequence and 13C bands, and buoyant densities of 200 µl fractions types associated with the N. marina/oligotropha lineages. were established by measurement of refractive index. In The Nitrosospira cluster 1-associated sequence type most cases linear gradients were established (Fig. 2) but present as one distinct band in all samples (clones 8 and frequently the first and last two to three fractions did not 13; Fig. 1A) was associated with two different Nitrosos- show linearity, and were discarded. In addition, in some pira cluster 1 sequences (LD1-A1 and LD1-B28; Table 2) cases, the gradient of refractive index was not linear, that were not resolved by DGGE analysis and were indicating disruption through disturbance, and these were observed predominantly in marine (A and B) or freshwa- also discarded. 12C- and 13C-DNA were separated well ter clone libraries (E) respectively. The two additional with highest concentrations in buoyancies of 1.68 and bands common to marine stations A and B were also 1.73 respectively (Fig. 2). However, both were also detect- associated with Nitrosospira cluster 1, while bands com- able throughout several fractions, exhibiting Gaussian-like mon to DGGE migration patterns from sites B, C and D distributions. were associated with the N. marina and Nitrosomonas sp. Following extraction, centrifugation and fractionation of Nm143 lineages. The additional band common in SIP microcosms, distinct bands of 12C- and 13C-DNA could freshwater-influenced sites D and E was also associated not be visualized in density centrifugation tubes by stan- with the N. marina lineage. dard ethidium bromide staining and fluorimetry. Stable isotope probing analysis was performed on samples from sites A and B amplifying 16S rRNA genes using bacterial Stable isotope probing analysis of AOB sediment primers to enable inclusion of sequence types not microcosm communities detected by AOB-specific primers. Denaturant gradient Calibration studies were performed with density gradient gel electrophoresis profiles (Fig. 3A and B) therefore con- centrifugation of mixtures of DNA from fully 13C-labelled tained considerably more bands than those in Fig. 1A, Nitrosomonas europaea cells and 12C-Escherichia coli particularly for low buoyant density fractions, containing

40 5

B Di 4.5

35 (ml) volume tube splaced

) 12

-1 C 4 l 30 µ 13C 3.5 25

.(ng 3 20 2.5 15 2

DNA conc 10 1.5 5 1 13C 12C 0 0.5 1.74 1.72 1.7 1.68 1.66 1.64 Buoyant density (g ml-1)

Fig. 2. Fractionation of unlabelled and 100% 13C-labelled nucleic acids within a buoyant density gradient in 5.1 ml centrifugation tubes with an initial CsCl buoyant density of 1.7019 g ml−1 subjected to centrifugation at 227 640 g for 24 h. A. Nucleic acids of individual buoyant density fractions resolved and quantiﬁed by standard agarose gel ethidium bromide ﬂuorescence. B. Distribution of unlabelled and 13C-labelled nucleic acids () and buoyant density gradient (). Inset: CsCl density gradient tube containing approximately 2 µg of unlabelled E. coli nucleic acids (top) and 100% 13C-labelled nucleic acids extracted from N. europaea respectively (bottom).

12 13 A Station A the tightly clustered C fractions from the C fractions. 13C fractions were also grouped together but to a lesser 12 11 degree than the C fractions. 1 5 1 8 2 12 In analysing changes in DGGE banding patterns along

4 3 13 the buoyant density gradient, emphasis was placed on 15 4 14 13 5 bands that appeared with high relative intensity in high 3 6 15 buoyant density fractions. This was achieved by ranking 2 7 16 the ratios of the relative intensities of each band averaged 12 8 6 9 17 over the buoyant density ranges established by calibration trials for 13C and 12C incorporation (Fig. 2). Bands that appeared in only one fraction were ignored. Proﬁles from 47 10 18 fractions obtained from both sites show the presence of R CLA 1.74 1.67 CLB Buoyant density [g ml-1] several bands in high buoyant density 13C fractions alone, the appearance of some bands in all fractions and the B Station B presence of others in 12C fractions only. The clone library generated from high buoyant density fractions from the 7 11 most marine station A demonstrated 10 clones with differ- 4 1 6 12 2 ent DGGE migration behaviour and corresponding bands 3 13 3 13 14 4 in C fractions (Table 3, Fig. 3A). These clones were 5 associated with bands that showed 13C/12C banding inten- 1 15 6 sity ratios ranked between 2 and 47, with six clones 2 16 7 8 ranked between 2 and 8 and three further clones ranked 5 17 9 12, 13 and 15. With the exception of clone 10 (ranked 47) that corresponded with a band that was present in all 13 12 31 18 10 fractions, all clones showed C/ C banding intensity 13 12 R CLB 1.74 1.67 CLA ratios > 1. The band with the highest C/ C banding Buoyant density [g ml-1] intensity ratio (i.e. ranked 1) was not retrieved in the clone Fig. 3. Stable isotope probing-denaturant gradient gel electrophore- library (Fig. 3A). BLAST searches of 0.7 kb clone sis (SIP-DGGE) analysis of ammonia oxidizer communities in sediments from Ythan estuary station A (A) and station B (B). Analysis was performed on 160 bp eubacterial 16S rRNA gene sequences B-PC1 µ −1 + ampliﬁed from sediment microcosms amended with 30 g ml NH4 - 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 13 N and Na2 CO3 and incubated for 20 days. Nucleic extracts were -0.8 -0.8 subjected to isopycnic CsCl density centrifugation and fractionated according to their buoyant density. Lanes represent amplicons gen- -0.6 erated with the general bacterial 27f-1492r PCR primers nested with -0.6 357f-GC and 518r primers from individual buoyant density fractions. -0.4 Clone libraries (CLA for microcosm sediment station A and CLB for -0.4 microcosm sediment station B) were assembled from nested 27f- A-PC2 1492r and 357f-pf1053r PCR products of the three fractions showing 2 -0.2

PC -0.2 - -

the highest buoyant densities. Individual clones were reampliﬁed with B 357f-GC plus 518r for DGGE analysis and clone library assembly. 0 0 Closest BLAST matches of clone sequences shown in lanes CLA and 13 CLB (marked by arrows) are given in Table 3. C enrichment of clone- 0.2 13 12 associated bands is indicated by C/ C proportionate ranking posi- 0.2 tions (R; see text); bold typeface indicates 13C/12C ratios > 1. No clone 0 4 was isolated with association with the bands ranked at position 1, for . station A, and position 4, for station B. 0.4 0.20.30.40.50.60. 70.80.91 A-PC1

12C-DNA, where complex banding patterns were Fig. 4. Principal component scores of correlation matrices produced from DGGE band intensities of bacterial 16S rRNA gene fragments observed. These proﬁles are typical of those in traditional from 13C-amended marine microcosms (station A, squares, and sta- DGGE analysis generated by bacterial 16S rRNA gene tion B, circles) after isopycnic CsCl density centrifugation and frac- 12 primers and, as expected, were similar throughout low tionation. Open symbols represent C buoyant density fractions (≤ 1.7; see Fig. 3B); solid symbols represent 13C buoyant density buoyant density fractions. Principal components analysis fractions (≥ 1.7; see Fig. 3B). PC1 represents 59.3% and 60% of the of SIP-DGGE banding intensities (59 and 53 bands for total variance for 13C microcosms A and B, respectively, and PC2 sites A and B respectively) (Fig. 4) clearly discriminated represents 17.6% and 17.2% of the total variance respectively. Ellipses (A, dotted lines; B, closed lines) demonstrate the 70% con- buoyant density fractions that were associated with the ﬁdence region of the principle component scores of the 12C and 13C 12C and 13C distribution (Fig. 2), resulting in separation of buoyant density fractions.

Table 3. BLAST closest matches of Ythan estuary sediment microcosm 0.7 kb 357f-pf1053r 16S rRNA clone sequences generated from high buoyancy nucleic acid fractions.

Representative Closest relative BLAST closest match % clone/library Name/putative lineage Reference accession number identity Rank

CLA-1 Arctic sea ice clone ARK9971/Pseudomonas 4 AF468404 98 5

CLA-2 Cold seep sediment clone/Bacteroidetes u.p. AB189346 92 8 Nitrifying bioﬁlm clone Koll-22/Sphingomonas 1 AJ224942 92

CLA-3 Deep sea mud clone Napoli-1B-52/n.a. u.p. AY592607 96 4

CLA-4 Nitrosomonas cryotolerans Nm55 5 AJ298738 99 15

CLA-5 Temperate estuarine mud clone KM23/n.a. u.p. AY216447 96 13

CLA-6 Nitrospira marina 6 X82559 98 3

CLA-7 Nullarbor cave bioﬁlm clone wb1_P06/Acidobacteria 3 AF317769 92 2

CLA-8 Antarctic sediment clone/cand. Division OP11 MERTZ 21CM 151 2 AF424451 92 12

CLA-9 Antarctic sediment clone BS1-0-111/cand. Division OP11 7 AY255001 91 6

CLA-10 Deep sea sed. clone D101/n.a. u.p. AY375134 97 47

CLB-11 Arctic sea ice clone ARK9971/Pseudomonas sp. 4 AF468404 99 7

CLB-12 Deep sea clone JTB247/Gammaproteobacterium 8 AB015251 99 6

CLB-13 Deep sea sediment Isolate/Janthinobacterium sp. u.p. AJ551147 99 3 An8/Oxalobacteriaceae

CLB-14 Gas hydrate sediment clone Hyd89-87/Gammaproteobacteria 9 AJ535246 93 10

CLB-15 Nitrospira marina 6 X82559 98 1

CLB-16 Comamonas aquatica LMG 5937/Comamonas 10 AJ430346 99 2

CLB-17 Nitrosomonas sp. C-17/Nitrosomonas sp. 143 11 AF338202 98 5

CLB-18 Deep sea sediment. clone D101/n.a. u.p. AY375134 97 31

Clone sequences were assembled from libraries generated from each buoyancy fraction as DGGE marker lanes (CLA and CLB; Fig. 3A and B). References: 1, Egli and colleagues (2003); 2, Bowman and McCuaig (2003); 3, Holmes and colleagues (2001); 4, Brinkmeyer and colleagues (2003); 5, Aakra and colleagues (2001); 6, Ehrich and colleagues (1995); 7, Kim and colleagues (2004); 8, Li and colleagues (1999); 9, Knittel and colleagues (2003); 10, Wauters and colleagues (2003); 11, Purkhold and colleagues (2003). n.a., no lineage association information available; u.p., unpublished. Rank = ranking according to the 13C/12C proportionate band intensities; bold typeface marks 13C/12C proportions > 1. sequences produced closest match associations between Discussion 92% and 99% similarity with database sequences, most of which were associated with clones or pure cultures The aim of this study was to test the hypothesis that the from marine environments. Two clones were closely asso- distribution patterns and activities of AOB communities in ciated with pure culture sequences of the marine ammo- estuarine environments could be explained in terms of nia oxidizer, Nitrosomonas cryotolerans (clone 4, ranked salinity gradients. At the study site, the Ythan estuary, the 15), and the marine nitrite oxidizer, Nitrospira marina high tide salinity gradient (Gillibrand and Balls, 1998) (clone 6, ranked 3) (Table 3, Fig. 3A). Two further clones ranges from marine values at the river mouth (site A) to (clones 7 and 2, ranked 2 and 8) were associated with a predominant freshwater influence at site E. At low tides, clone sequences isolated in studies on nitrifying biofilms. salinities in the entire estuary up to site A decrease con- For station B, eight clones were sequenced corre- siderably, especially when the tidal range is small. Addi- sponding to DGGE migration behaviour of bands from 13C tional fluctuations result from seasonal rainfall, the buoyancy fractions. These were associated with bands influence of river flow and high water levels at neap and with 13C/12C banding intensity ratios ranked between 1 spring tides, when the marine water front can reach site and 31, with seven ranked between 1 and 10 and 13C/12C E (Gillibrand and Balls, 1998). Changes in community banding intensity ratios > 1 (Table 3, Fig. 3B). Again, structure along the estuary may therefore result from BLAST searches showed two clones with sequences selection of organisms adapted to particular salinities or closely associated with pure cultures of marine nitrify- to changes in salinity, even at sites with an otherwise ing strains, the betaproteobacterial ammonia oxidizer overriding marine or freshwater influence. Related studies Nitrosomonas sp. Nm143 (clone 17, ranked 5) and the on the diversity of bacterial communities along salinity nitrite oxidizer Nitrospira marina (clone 15, ranked 1) gradients often focus on the distribution of planktonic (Table 3, Fig. 3B). All other clones showed closest match assemblages (Cebron et al., 2004; Crump et al., 2004) associations with clones and isolates from marine envi- and the distinction of true planktonic riverine, estuarine ronments. As for station A, no sequence was generated and marine communities can be obscured by mixing of with similarity to the Nitrosospira cluster 1 clade that dom- water bodies and temporal changes. However, these influ- inated environmental DGGE migration patterns from ences will be less for sediment communities, where these stations. organisms are attached to particles or occupy interstitial

© 2005 The Authors Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 8, 684–696 690 T. E. Freitag, L. Chang and J. I. Prosser pore space, where pore water salinities may be buffered orous characterization of laboratory enrichment or pure against the peak changes over a tidal cycle. cultures. De Bie and colleagues (2001) also detected N. oligotropha- and N. marina-like sequences in the Schelde estuary and Bollmann and Laanbroek (2002) Changes in community structure along the estuarine concluded from culture experiments that salinity was salinity gradient responsible for differences in AOB communities. However, Communities of betaproteobacterial AOBs, comprising the maximum salinity of their study sites was only 19‰ communities of active and inactive AOB, were character- and no Nitrosospira cluster 1-, Nitrosomonas cluster 5- or ized by DGGE profiling and sequencing of 16S rRNA Nitrosomonas 143-like sequences were retrieved. Caffrey genes amplified with the CTO primer set. This primer set and colleagues (2003) reported similar sequences in the is specific for betaproteobacterial AOBs, but with potential Elkhorn Slough estuary in California, but with no evidence bias through mismatches in the CTO189f sequence of a relationship with salinity or sampling. Recently Ber- against some Nitrosomonas lineages, and there is nard et al. (2001) analysed the distribution of the func- also the possibility of amplification from naked DNA tional gene responsible for the first step in the oxidation (Demaneche et al., 2001). Stable isotope probing employ- of ammonia, the ammonia monooxygenase gene (amoA) ing eubacterial primers was used to characterize active along a salinity gradient in an estuarine system. Their organisms. The first of these approaches demonstrated results confirm the dominance of Nitrosospira-associated significant differences within AOB communities at sam- sequences at marine sites and the prevalence of pling sites along the estuary, with Nitrosospira cluster N. oligotropha- and Nitrosomonas sp. Nm143-associated 1-like sequences dominating AOB communities close to sequences at freshwater and intermediate sites. Evidence the river mouth and N. oligotropha/marina-associated from distributions of benthic invertebrates along a salinity sequences dominating freshwater sites. Denaturant gra- gradient in the Baltic (Remane and Schlieper, 1971) has dient gel electrophoresis profiles from intermediate sites led to the hypothesis that diversity will be lower in brack- were dominated by Nitrosomonas sp. Nm143 sequences ish-water ecosystems than in marine or freshwater sys- and, to a lesser extent, by Nitrosomonas cluster 5-like tems, but this was not supported here. The number and sequences. These trends in community structure are con- relative intensity of individual 16S rRNA gene sequences sistent with previous studies of marine and freshwater resolved on DGGE suggest similar richness of AOB com- environments and with physiological characteristics of lab- munities at all sampling sites but significantly reduced oratory culture representatives (where available), and evenness and, to some extent, reduced Shannon diversity suggest selection for organisms with different responses indices at the marine site. Similarities in richness along to salinity. With the exception of the freshwater/soil strain the gradient may result from temporal changes in salinity N. oligotropha, all BLAST closest match sequences from and diversity indices at sites A and B are strongly influ- cultured members of the N. oligotropha/marina group are enced by high relative abundance of Nitrosospira cluster obligate halophiles (with salt optima of 300–400 mM 1 and Nitrosomonas cluster 5 sequences, whose roles in NaCl) or, for Nitrosomonas 143, have been isolated ammonia oxidation are not yet confirmed. only from coastal environments (Jones et al., 1988; Koops et al., 2003). The presence of N. oligotropha-like Application of SIP to estuarine AOB community analysis sequences in sediments strongly influenced by marine water influx (station B), and of halophilic Nitrosomonas Stable isotope probing analysis was carried out to deter- aestuarii and N. marina sequences from sediments at mine which AOB were active at sites A and B, which had freshwater sites (E), may therefore reflect adaptation to a contrasting communities, and to assess whether Nitrosos- range of salinities, while individual strains may be active pira cluster 1 sequences in marine sites contributed to at different salinities. The uncultured Nitrosospira cluster activity. 13C-DNA-derived sequences will identify organ- 13 1- and Nitrosomonas cluster 5-like sequences dominate isms that have incorporated C-CO2 or secondary AOB sequences amplified from estuarine and marine sed- utilizers, incorporating organic compounds released by iments (McCaig et al., 1999; Bano and Hollibaugh, 2000; autotrophs or biomass following death. Molecular analysis Freitag and Prosser, 2003; 2004). Pure culture represen- was carried out using Bacteria-specific primers to ensure tatives of these groups have not been obtained, and their that all active autotrophs were detected, in addition to role in ammonia oxidation is based on clustering of secondary utilizers. Principle components analysis clearly their 16S rRNA genes within a monophyletic group for discriminated bacterial DGGE profiles derived from 12C- which all cultivated representatives are autotrophic AOB, and 13C-associated buoyant densities, indicating success- and on reports of non-persisting enrichment cultures of ful density centrifugation and fractionation, distinct Nitrosomonas cluster 5 (Stephen et al., 1996). Unequivo- microbial communities and minimal cross-contamination. cal evidence of their assumed role as AOB requires rig- The tight clustering of 12C fractions reflects microbial com-

© 2005 The Authors Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 8, 684–696 Ammonia oxidizers in a freshwater–marine estuarine gradient 691 munities unaffected by 13C incorporation spanning the sequence similarity (98%) to the Nitrosomonas sp. Nm143 buoyant density gradient according to the Gaussian-like lineage. Detection of Nitrospira marina sequence type distribution observed during calibration spins. The cluster- indicates high rates of autotrophic nitrite oxidation after ing of 13C buoyant density fractions was less pronounced, 20 days of microcosm incubation. The concurrent high reflecting more heterogeneity in DGGE profiles, due to the ranking of sequence types generally associated with presence of naturally abundant 12C-carbonate in sedi- heterotrophs suggests significant turnover of biomass ments and turnover of 13C substrates. This will dilute 13C generated from nitrifiers, the existence of previously incorporation, leading to accumulation of nucleic acids uncharacterized autotrophs, or may be indicative of het- that are not fully labelled in fractions with a relative lower erotrophic CO2 assimilation (Hesselsoe et al., 2005). How- buoyant density. In addition, low copy numbers in 13C ever, Hesselsoe and colleagues (2005) estimated that buoyant density fractions, due to low label incorporation, heterotrophic CO2 assimilation accounted for 3% of the may lead to stochastic amplification of individual tem- total cell carbon in Pseudomonas putida, whereas plates, generating more random DGGE profiles (Taberlet Manefield and colleagues (2002) suggested that 20% et al., 1996). nucleic acid isotope incorporation is the minimum for Stable isotope probing calibration trials demonstrated successful fractionation during SIP-based approaches. shallow gradient curves close to the minimum possible Furthermore, the appearance of members of slope, resulting in maximum interband distance and Burkholderiales-like sequence types (Commamonas and sharply defined 12C and 13C bands. However, the distribu- Oxalobacteriaceae) at higher ranking positions than the tion of nucleic acids in density fractions implies contami- AOB Nitrosomonas sp. Nm143 suggests their degradative nation of 13C-associated fractions by 12C nucleic acids, rather then syntrophic role within autotrophic communities. with potential for generation of false-positive results from Clone sequences related to Nitrosomonas cryotolerans 13C-associated fractions when the 13C/12C nucleic acid and Nitrospira marina were observed in microcosm A. The ratio is low and highly sensitive PCR strategies are used. obligate halophilic N. cryotolerans grows at temperatures In this case, the concentration of nucleic acids loaded onto as low as −5°C (Jones et al., 1988) and survives for pro- individual density centrifugation tubes is of considerable longed periods at low ammonia concentrations (Jones importance, as higher loads will increase the likelihood of and Morita, 1985; Johnstone and Jones, 1988). Few 16S contamination. However, confidence in interpretation of rRNA gene clones with high similarity to N. cryotolerans SIP data was increased by ranking DGGE bands accord- have been recovered from marine systems (Tal et al., ing to the ratio of the averaged relative intensities in 13C 2003) and nothing is known of its importance in marine and 12C buoyant density fractions and facilitated identifi- nitrification. Other clone sequences were associated cation of sequence types indicative of 13C incorporation, with an Acidobacter-like sequence from a lithotrophic which could otherwise only be achieved by extensive com- biofilm, with high abundance of Nitrospira-related clone parison of clone libraries. Ranking of 13C/12C relative inten- sequences (Holmes et al., 2001) and with sequences sim- sity ratios may bias against active organisms present at ilar to Pseudomonas, Actinobacter, the uncharacterized high relative abundance. Nevertheless, the study provides candidate division OP11 and a Sphingomonas sequence confidence in the use of SIP for analysis of active AOB in from a nitrifying biofilm with high relative abundance of natural environments and omission of DGGE bands Anammox and Nitrospira-like sequence types (Egli et al., present in one fraction only enables further discrimination 2003). The physiological characteristics of members of of sequence types representing label incorporation from these groups are largely unknown and a potential role in those resulting from stochastic PCR effects. autotrophic metabolism cannot be excluded. The respective low and high rankings of AOB and Nitrospira marina sequence types suggest that nitrite oxidation was the pre- Stable isotope probing analysis of AOB dominant autotrophic process after 20 days of incubation Screening of a relatively low number of clones generated and that AOB biomass had been degraded. The closest with bacterial primers from nucleic acids in SIP buoyant database relative of the AOB sequence type was density fractions indicating 13C incorporation yielded some Nitrosomonas sp. Nm143, which was also dominant in the sequences with closest BLAST matches to ammonia and environmental DGGE and clone survey. Only four cultured nitrite oxidizers. This demonstrates the sensitivity of the strains of Nitrosomonas sp. Nm143 have been isolated approach, as nitrifiers will constitute a small proportion of from marine environments (Purkhold et al., 2003) but 16S the total bacterial community. A 16S rRNA gene fragment rRNA gene sequences associated with this lineage have highly similar (98%) to the nitrite oxidizer Nitrospira marina been isolated from various marine environments, includ- was isolated from the station B sediment microcosm that ing anoxic sediments (Freitag and Prosser, 2003), co-migrated with the highest ranking DGGE band and the suggesting its abundance in marine and estuarine eco- clone sequence of the fifth highest ranking band had high systems. Nothing is known about its importance within the

© 2005 The Authors Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 8, 684–696 692 T. E. Freitag, L. Chang and J. I. Prosser marine nitrogen cycle and this is the first study confirming the activity of Nitrosomonas sp. Nm143 within an estuarine sediment ecosystem. The BLAST closest match AOB E sequence type identified by SIP in the microcosm from the most marine station A was N. cryotolerans, which was Ythan not identified in the environmental DGGE and clone survey. However, this may have been due to the high relative abundance of Nitrosospira cluster 1 sequence type. No clone with similarity to the Nitrosospira cluster 1 or Nitrosomonas cluster 5 sequence types was recovered from SIP microcosms, which may be due to the relatively high ammonia concentrations and temperature during incubation. However, their role in ammonia oxidation D therefore remains uncertain. The high proportion of non- nitrifier sequences in 13C fractions suggests that shorter incubation periods are required for better discrimination of active nitrifiers. The study therefore provides evidence for the selection C of different ammonia oxidizer ecotypes along an estuarine salinity gradient which correlates with the physiological characteristics of representative cultured strains, where these are available. The study also demonstrates the abil- B ity of stable isotope probing to distinguish active and inactive ammonia and nitrite oxidizers at different sites and the activity of other organisms, presumed to be secondary A utilizers. 1km North Sea Experimental procedures Site description and sampling Fig. 5. Map of the Ythan estuary in the northeast of Scotland, UK, Sediment samples were collected from the Ythan estuary on showing the sampling sites, covering a 6 km section along the estu- the east coast of Scotland, UK (Fig. 5), approximately 8 km ary. Salinities ranged from close to marine values at station A to long and with average and maximum widths of approximately freshwater values at station E. Dotted lines indicate the low tide water line. Inset: outline of Scotland and location of sampling area. 300 m and 620 m, respectively, containing a total intertidal area of approximately 1.85 km2 (Leach, 1971). Flushing times of the estuary range from one tidal cycle (see above) at each sampling station close to the water line. Sampling to 5–12 days (Balls, 1994) and approximately 95% of the stations were (Fig. 5): A: 57°18′32″N, 1°59′30″W; B: catchment area (approximately 680 km2) is agricultural 57°18′42″N, 1°59′41″W; C: 57°19′20″N, 1°59′43″W; D: (Lyons et al., 1993). Depending on seasonal rainfall and river 57°20′04″N, 2°00′10″W; and E: 57°21′37″N, 2°01′00″W. flow, salinity over a 6 km stretch ranges from close to marine Sediment samples were freed from larger particles and pro- values (25–30‰; averaged over the tidal cycle) at the river cessed on the same day. Subsamples for molecular analysis mouth (Fig. 5, station A) to 0–5‰ before the Logie Buchan were frozen at −80°C until required. bridge, and approaches freshwater values shortly beyond the bridge (Fig. 5; station E; Gillibrand and Balls, 1998). Ammo- nia concentrations show a negative correlation with salinity Stable isotope probing microcosm studies and median values range between 8 µM at freshwater sites to approximately 1 µM at marine salinity (Balls et al., 1995). Microcosms consisted of 25 g of sediment, bulked from two Average temperatures in Ythan river water and coastal samples from stations with the strongest marine influence surface water at the sampling time were 5°C and 6°C (A and B), in 250 ml Erlenmeyer flasks and amended with respectively (Scottish Environmental Protection Agency, 5 ml of artificial sea water (g l−1 distilled water: NaCl, 25;

Water Data, Harmonized Monitoring Scheme, http://www. MgCl2·7H2O, 5; CaCl2, 1; KCl, 1) supplemented with ammo- −1 + sepa.org.uk/data/hm/hm.asp; Coastal Long-term Monitoring nium sulfate at 30 µg ml NH4 -N, giving a ﬁnal concentration −1 + Programme, FRS Marine Laboratory, Aberdeen, http://www. of approximately 10 µg ml NH4 -N. Flasks were sealed with frs-scotland.gov.uk/Delivery/standalone.aspx?contentid=543). butyl rubber stoppers and the headspace was purged with 13 Sediment samples were taken at low tide in March 2004 at CO2-free air (Whitby et al., 2001). A C-labelled carbon ﬁve locations covering a 6 km section along the estuary. Two source was provided by dropwise addition of 1 ml of 5% 13 replicate surface samples were collected within a 5 m radius (w/v) Na2 CO3 (99% atom enriched; CK-Gas, Cambridge,

UK) to the sediment slurry, while simultaneously titrating with using a Gilson MiniPlus3 peristaltic pump (Anachem, Luton, an equimolar solution of HCl to give a final pH of 7.5. Flasks UK) at a flow rate of 0.5 ml min−1. Density gradients were were incubated at 20°C for 3 weeks with weekly flushing of aliquoted into 200 µl fractions with a LKB 8Romma 13 the headspace and replenishment of Na2 CO3. Flasks were 2112Redirac fraction collector (Pharmacia LKB, Cambridge, sampled destructively for extraction of nucleic acids. UK). Buoyant density was determined by measuring the refractive index in a 10 µl aliquot of each fraction and centrifugation tubes showing a disturbed linear density gradient DNA extraction were discarded. Nucleic acids were separated from CsCl by overnight precipitation in 2 vols of polyethylene glycol at 4°C Nucleic acids were extracted from sediments and SIP micro- (Griffiths et al., 2000) followed by centrifugation at 20 000 g cosms according to Griffiths and colleagues (2000) with pre- for 30 min. Residual salts were removed by washing with viously described modifications (Freitag and Prosser, 2003). 70% ethanol, pellets were air-dried and precipitated nucleic Nucleic acid extracts from microcosms were visualized by acids were re-eluted in 30 µl of sterile H2O. For verification ethidium bromide staining in standard agarose gel electro- of the buoyant density range of maximum concentrations of phoresis with Hyperladder DNA fragment size markers 13C-labelled and unlabelled DNA, precipitated nucleic acids (Bioline, London, UK) and quantified by densitometry from calibration tubes containing labelled N. europaea and analysis (TN-Image, freeware by Thomas J. Nelson, http:// unlabelled E. coli DNA were visualized and concentrations entropy.brneurosci.org/tnimage.html). estimated by ethidium bromide fluorescence in standard agarose gel electrophoresis with Hyperladder DNA fragment size markers (Bioline). Isopycnic density centrifugation and gradient fractionation

13C-labelled and unlabelled nucleic acids extracted from SIP Analysis of sediment AOB communities microcosms were fractionated by density centrifugation in CsCl gradients (Schildkraut et al., 1962; Radajewski et al., Extracted DNA was amplified by nested PCR using 16S 2000) in 5.1 ml Quick-Seal polyallomer tubes in a Vti65.2 rRNA gene primer sets specific for the majority of betapro- vertical rotor (Beckman Coulter, Palo Alto, CA, USA). CsCl teobacterial AOB (CTO189f and CTO654r) (Kowalchuk et al., gradients and ultracentrifugation conditions were initially cal- 1997). Polymerase chain reaction products were reamplified ibrated with 4 µg of DNA per tube, extracted from pure cul- for DGGE analysis using the general bacterial 357f-GC and tures of N. europaea grown in Skinner and Walker medium 518r primers (Muyzer et al., 1993). Before secondary ampli- −1 + fication, PCR products were diluted 1:20 to prevent amplifi- (Skinner and Walker, 1961) at 500 µg ml NH4 -N (Powell 13 cation of non-target sequences without affecting DGGE and Prosser, 1985) supplemented with Na2 CO3, as described above, and E. coli, cultivated in standard Luria– profiles. Cloning inserts of 0.45 kb, specific for betaproteo- Bertani nutrient broth (g l−1 distilled water: tryptone 10 g; bacterial AOB, were created by PCR amplification of DNA β β yeast extract 5 g; NaCl 10 g). Assuming a buoyant density of extracts with the -AMO161f and -AMO1301r primer set 1.7 g cm−3 for double-stranded supercoiled DNA (Schildkraut (McCaig et al., 1994), followed by PCR amplification with the et al., 1962), the initial CsCl concentration and rotor speed CTO189f and CTO654r primer set. Primer pairs and PCR were estimated for maximum interband distance according to conditions for all amplifications were used as described pre- the equilibrium CsCl gradient curves specific for the Vti65.2 viously (Freitag and Prosser, 2003). Denaturant gradient gel rotor (Beckman Coulter; Rotor supplements) and adjusted electrophoresis was carried out as described previously empirically for maximum interband distance with labelled (Kowalchuk et al., 1997) using the DCode Universal Mutation genomic DNA extracted from N. europaea and unlabelled Detection System (Bio-Rad Laboratories, Hemel Hempstead, E. coli DNA. Pre-centrifugation density of CsCl solutions and UK) and 1.5 mm, 8% polyacrylamide gels containing dena- density of fractionated gradients (see below) were verified by turant gradients of 35–62%, for analysis of 357f-GC-518r refractive index readings using a high-contrast, hand-held PCR products. Polyacrylamide gels were silver-stained refractometer ATAGO-R-5000 (UNI-IT, Tokyo, Japan). Refrac- (McCaig et al., 2001) for 15 min and destained for 10 min in tive index readings were converted into their proportional dH2O before digital image analysis (Phoretix 1-D gel analysis buoyant densities according to conversion tables (Beckman software; Phoretix International, Newcastle-Upon-Tyne, UK). Coulter; Ultracentrifuge supplements). Following optimiza- To compensate for variations in DNA loading between DGGE tion, 0.5 µg of nucleic acids extracted from microcosms was lanes, the total band intensity for each lane was normalized suspended in a total volume of 200 µl of CsCl in TE buffer to that of the lane with the lowest DNA loading (McCaig et al., (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) containing 1 µg 2001; Nicol et al., 2003). Similarity matrices, based on band of ethidium bromide, adjusted to a buoyant density of presence, were produced using the Dice coefficient, from 1.7019 g ml−1, thoroughly mixed, and then suspended in which similarity dendrograms were constructed by UPGMA. 4.8 ml of CsCl in TE buffer at a buoyant density of Additionally, the Shannon diversity index, which describes 1.7019 g ml−1. Residual volumes in the tubes were filled with both species richness and evenness, was calculated by mineral oil (molecular grade, Bio-Rad, Hemel Hempstead, counting the numbers of different band positions and their UK) and tubes were subjected to centrifugation at 227 640 g respective relative intensity per sample (McCaig et al., 1999). for 24 h at 20°C. Density gradient tubes were fractionated into equal parts Community analysis of 13C microcosms by displacement from the top using a Beckman fraction recovery system, supplying sterile water to the top of the tube Nucleic extracts from individual fractions obtained from CsCl

© 2005 The Authors Journal compilation © 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 8, 684–696 694 T. E. Freitag, L. Chang and J. I. Prosser density fractionation were analysed by nested PCR amplifi- subjected to nested PCR with the 16S rRNA gene primers cation using general bacterial primers 27f and 1492r (Lane, 357f-GC-518r (Muyzer et al., 1993) and screened by DGGE. 1991) for primary amplification, followed by secondary ampli- The 0.7 kb 357f-pf1053r PCR amplification products of fication with primers 357f-GC and 518r (Muyzer et al., 1993) clones showing DGGE migration patterns identical to bands for DGGE analysis. Polymerase chain reaction amplicons in amplicons generated from density fractions were selected (0.7 kb) of sufficient quality for use as cloning inserts were for sequencing. Representatives of environmental and micro- created by nested PCR amplification of the 27f-1492r PCR cosm clone libraries were amplified with vector primers M13f- product with the 357f primer in conjunction with the general M13r and M13 PCR products were purified by standard pre- bacterial primer pf1053r (Edwards et al., 1989). Denaturant parative agarose gel electrophoresis as described above. To gradient gel electrophoresis analysis of general bacterial 16S ensure sequence read without ambiguities, M13 vector prod- rRNA gene fragments was carried out as described above, ucts were sequenced with sufficient overlap of sequencing but with polyacrylamide gels containing denaturant gradients reads using the SP6 and T7 vector primers (Promega). of 35–70%. Denaturant gradient gel electrophoresis banding Sequencing reactions were performed using the BigDye Ter- profiles derived from DNA with buoyant densities correspond- minator Cycle Sequencing Kit (PE Biosystems, Warrington, ing to 13C and 12C fractions were analysed in three ways to UK), and the cycle sequencing products were analysed with assess active AOB: (i) qualitative assessment of increases in a Model ABI377 automated sequencer (PE Biosystems). 16S relative intensity, or appearance of bands, in 13C fractions in rRNA gene clone sequences were compared with closely comparison with 12C fractions, (ii) calculation, for each band, related sequences retrieved from the GenBank database of the ratio of relative intensity averaged across 13C and 12C using the BLAST algorithm (Altschul et al., 1990). Sequences fractions, with 13C:12C ratios > 1 indicating sequences derived have been deposited at GenBank under Accession Nos 13 from organisms incorporating C-CO2 and (iii) principal com- DQ068702–DQ068740. ponents analysis of normalized relative intensities of bands from 13C and 12C fractions (Systat 8.0; Systat Software UK Limited, Hounslow, UK). References Aakra, A., Utaker, J.B., Pommerening-Röser, A., Koops, Cloning and sequence analysis H.P., and Nes, I.F. (2001) Detailed phylogeny of ammonia- oxidizing bacteria determined by rDNA sequences and For analysis of betaproteobacterial AOB, one clone library DNA homology values. Int J Syst Evol Microbiol 51: 2021– per sampling point was created using PCR amplicons gen- 2030. erated with the CTO primer set, as described above. Two Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, PCR amplicons from each sediment sample were pooled to D.J. (1990) Basic local alignment search tool. J Mol Biol compensate for PCR drift and to ensure that dominant 215: 403–410. sequences were represented in the clone library. Polymerase Attrill, M.J. (2002) A testable linear model for diversity trends chain reaction fragments were purified by standard agarose in estuaries. J Anim Ecol 71: 262–269. electrophoresis and excised DNA bands were cleaned as Balls, P.W. (1994) Nutrient inputs to estuaries from 9 Scottish described by Boyle and Lew (1995), ligated into the pGEM east-coast rivers – influence of estuarine processes on T- vector system (Promega, Southampton, UK) and trans- inputs to the North Sea. Estuar Coast Shelf Sci 39: 329– formed into XL1-Blue MRF Kan supercompetent E. coli cells 352. (Stratagene, Cambridge, UK) according to the manufactur- Balls, P.W., Macdonald, A., Pugh, K., and Edwards, A.C. ers’ instructions. Transformed colonies were screened for (1995) Long term nutrient enrichment of an estuarine sys- inserts of the correct size by PCR amplification with the CTO tem – Ythan, Scotland (1958–1993). Environ Pollut 90: PCR assay, as described above, and 15 clones per library 311–321. with the correct fragment size insert were subjected to nested Bano, N., and Hollibaugh, J.T. (2000) Diversity and distribu- PCR with the 357f-GC-518r (Muyzer et al., 1993) primers. tion of DNA sequences with affinity to ammonia-oxidizing Polymerase chain reaction products were screened by bacteria of the beta subdivision of the class Proteobacteria DGGE for sequence differences within the hypervariable V3 in the Arctic Ocean. Appl Environ Microbiol 66: 1960– region. Clones selected for sequencing represented at least 1969. two replicate sequence types with DGGE migration patterns Bernard, L., Courties, C., Duperray, C., Schafer, H., Muyzer, identical to individual DGGE bands in amplicons of environ- G., and Lebaron, P. (2001) A new approach to determine mental samples. the genetic diversity of viable and active bacteria in aquatic Clone libraries were also generated from three gradient ecosystems. Cytometry 43: 314–321. 13 fractions from DNA extracted from each Na2 CO3-amended de Bie, M.J.M., Speksnijder, A.G.C.L., Kowalchuk, G.A., microcosm, within the buoyant density range of 13C-labelled Schuurman, T., Zwart, G., Stephen, J.R., et al. (2001) N. europaea DNA. Bacterial 16S rRNA gene fragments were Shifts in the dominant populations of ammonia-oxidizing generated, as described above, from at least three pooled beta-subclass Proteobacteria along the eutrophic Schelde amplicons of each density fraction, to ensure recovery of estuary. Aquat Microb Ecol 23: 225–236. dominant sequence types. Ligations and transformations Bollmann, A., and Laanbroek, H.J. (2001) Continuous culture were performed as described above and transformed colo- enrichments of ammonia-oxidizing bacteria at low ammo- nies screened for correct size inserts by PCR amplification nium concentrations. FEMS Microbiol Ecol 37: 211–221. with the M13f-M13r vector primers (Promega). Thirty clones Bollmann, A., and Laanbroek, H.J. (2002) Influence of oxy- with the correct fragment size insert of each library were gen partial pressure and salinity on the community com-

position of ammonia-oxidizing bacteria in the Schelde Head, I.M., Hiorns, W.D., Embley, T.M., McCarthy, A.J., and estuary. Aquat Microb Ecol 28: 239–247. Saunders, J.R. (1993) The phylogeny of autotrophic Bowman, J.P., and McCuaig, R.D. (2003) Biodiversity, com- ammonia-oxidizing bacteria as determined by analysis of munity structural shifts, and biogeography of prokaryotes 16S ribosomal RNA gene sequences. J Gen Microbiol 139: within Antarctic continental shelf sediment. Appl Environ 1147–1153. Microbiol 69: 2463–2483. Hesselsoe, M., Nielsen, J.L., Roslev, P., and Nielsen, P.H. Boyle, J.S., and Lew, A.M. (1995) An inexpensive alternative (2005) Isotope labeling and microautoradiography of active to glassmilk for DNA puriﬁcation. Trends Genet 11: 8. heterotrophic bacteria on the basis of assimilation of

Brinkmeyer, R., Knittel, K., Jurgens, J., Weyland, H., Amann, (CO2)-C-14. Appl Environ Microbiol 71: 646–655. R., and Helmke, E. (2003) Diversity and structure of bac- Hollibaugh, J.T., Bano, N., and Ducklow, H.W. (2002) Wide- terial communities in arctic versus antarctic pack ice. Appl spread distribution in polar oceans of a 16S rRNA gene Environ Microbiol 69: 6610–6619. sequence with affinity to Nitrosospira-like ammonia- Burrell, P.C., Phalen, C.M., and Hovanec, T.A. (2001) Iden- oxidizing bacteria. Appl Environ Microbiol 68: 1478–1484. tification of bacteria responsible for ammonia oxidation in Holmes, A.J., Tujula, N.A., Holley, M., Contos, A., James, freshwater aquaria. Appl Environ Microbiol 67: 5791–5800. J.M., Rogers, P., and Gillings, M.R. (2001) Phylogenetic Caffrey, J.M., Harrington, N., Solem, I., and Ward, B.B. structure of unusual aquatic microbial formations in Nullar- (2003) Biogeochemical processes in a small California bor caves, Australia. Environ Microbiol 3: 256–264. estuary. 2. Nitrification activity, community structure and Jetten, M.S., Sliekers, O., Kuypers, M., Dalsgaard, T., van role in nitrogen budgets. Mar Ecol Prog Ser 248: 27–40. Niftrik, L., Cirpus, I., et al. (2003) Anaerobic ammonium Cebron, A., Coci, M., Garnier, J., and Laanbroek, H.J. (2004) oxidation by marine and freshwater planctomycete-like Denaturing gradient gel electrophoretic analysis of ammo- bacteria. Appl Microbiol Biotechnol 63: 107–114. nia-oxidizing bacterial community structure in the lower Johnstone, B.H., and Jones, R.D. (1988) Physiological Seine River: impact of Paris wastewater effluents. Appl effects of long term energy source deprivation on the sur- Environ Microbiol 70: 6726–6737. vival of a marine chemolithotrophic ammonium oxidizing Crump, B.C., Hopkinson, C.S., Sogin, M.L., and Hobbie, J.E. bacterium. Mar Ecol Prog Ser 49: 295–303. (2004) Microbial biogeography along an estuarine salinity Jones, R.D., and Morita, R.Y. (1985) Survival of a marine gradient: combined influences of bacterial growth and res- ammonium oxidizer under energy source deprivation. Mar idence time. Appl Environ Microbiol 70: 1494–1505. Ecol Prog Ser 26: 175–179. Demaneche, S., Jocteur-Monrozier, L., Quiquampoix, H., Jones, R.D., Morita, R.Y., Koops, H.P., and Watson, S.W. and Simonet, P. (2001) Evaluation of biological and phys- (1988) A new marine ammonium oxidizing bacterium, ical protection against nuclease degradation of clay-bound Nitrosomonas cryotolerans sp nov. Can J Microbiol 34: plasmid DNA. Appl Environ Microbiol 67: 293–299. 1122–1128. Edwards, U., Rogall, T., Blocker, H., Emde, M., and Bottger, Kemp, W.M., Sampou, P., Caffrey, J., Mayer, M., Henriksen, E.C. (1989) Isolation and direct complete nucleotide deter- K., and Boynton, W.R. (1990) Ammonium recycling versus mination of entire genes – characterization of a gene cod- denitrification in Chesapeake Bay sediments. Limnol ing for 16S-ribosomal RNA. Nucleic Acids Res 17: 7843– Oceanogr 35: 1545–1563. 7853. Kim, B.S., Oh, H.M., Kang, H., Park, S.S., and Chun, J. Egli, K., Bosshard, F., Werlen, C., Lais, P., Siegrist, H., (2004) Remarkable bacterial diversity in the tidal flat sedi- Zehnder, A.J.B., and van der Meer, J.R. (2003) Microbial ment as revealed by 16S rDNA analysis. J Microbiol Bio- composition and structure of a rotating biological contactor technol 14: 205–211. biofilm treating ammonium-rich wastewater without organic Knittel, K., Boetius, A., Lemke, A., Eilers, H., Lochte, K., carbon. Microb Ecol 45: 419–432. Pfannkuche, O., et al. (2003) Activity, distribution, and Ehrich, S., Behrens, D., Lebedeva, E., Ludwig, W., and Bock, diversity of sulfate reducers and other bacteria in sedi- E. (1995) A new obligately chemolithoautotrophic, nitrite- ments above gas hydrate (Cascadia margin, Oregon). oxidizing bacterium, Nitrospira moscoviensis sp nov and Geomicrobiol J 20: 269–294. its phylogenetic relationship. Arch Microbiol 164: 16–23. Koops, H.P., and Pommerening-Röser, A. (2001) Distribution Freitag, T.E., and Prosser, J.I. (2003) Community structure and ecophysiology of the nitrifying bacteria emphasizing of ammonia-oxidizing bacteria within anoxic marine sedi- cultured species. FEMS Microbiol Ecol 37: 1–9. ments. Appl Environ Microbiol 69: 1359–1371. Koops, H.P., Purkhold, U., Pommerening-Röser, A., Timmer- Freitag, T.E., and Prosser, J.I. (2004) Differences between mann, G., and Wagner, M. (2003) The lithoautotrophic betaproteobacterial ammonia-oxidizing communities in ammonia-oxidizing bacteria. In The Prokaryotes: An Evolv- marine sediments and those in overlying water. Appl Envi- ing Electronic Resource for the Microbiological Commu- ron Microbiol 70: 3789–3793. nity. Dworkin, M. (ed.). New York, USA: Springer-Verlag, Gillibrand, P.A., and Balls, P.W. (1998) Modelling salt intru- URL http://link.springer-ny.com/link/service/books/10125. sion and nitrate concentrations in the Ythan estuary. Estuar Kowalchuk, G.A., Stephen, J.R., DeBoer, W., Prosser, J.I., Coast Shelf Sci 47: 695–706. Embley, T.M., and Woldendorp, J.W. (1997) Analysis of Griffiths, R.I., Whiteley, A.S., O’Donnell, A.G., and Bailey, ammonia-oxidizing bacteria of the beta subdivision of the M.J. (2000) Rapid method for coextraction of DNA and class Proteobacteria in coastal sand dunes by denaturing RNA from natural environments for analysis of ribosomal gradient gel electrophoresis and sequencing of PCR- DNA- and rRNA-based microbial community composition. amplified 16S ribosomal DNA fragments. Appl Environ Appl Environ Microbiol 66: 5488–5491. Microbiol 63: 1489–1497.

Lane, D.J. (1991) 16S/23S rRNA sequencing. In Nucleic Acid based phylogeny of 12 novel betaproteobacterial ammo- Techniques in Bacterial Systematics. Stackebrandt, E., and nia-oxidizing isolates: extension of the dataset and pro- Goodfellow, M. (eds). Chichester, UK: Academic Press, posal of a new lineage within the nitrosomonads. Int J Syst pp. 115–175. Evol Microbiol 53: 1485–1494. Leach, J.H. (1971) Hydrology of the Ythan estuary with ref- Radajewski, S., Ineson, P., Parekh, N.R., and Murrell, J.C. erence to the distribution of major nutrients and detritus. (2000) Stable-isotope probing as a tool in microbial ecol- J Mar Biol Assoc 51: 157. ogy. Nature 403: 646–649. Li, L.N., Kato, C., and Horikoshi, K. (1999) Bacterial diversity Raffaelli, D. (1999) Nutrient enrichment and trophic organi- in deep-sea sediments from different depths. Biodivers sation in an estuarine food web. Acta Oecol Int J Ecol 20: Conserv 8: 659–677. 449–461. Lueders, T., Wagner, B., Claus, P., and Friedrich, M.W. Raffaelli, D., Balls, P., Way, S., Patterson, I.J., Hohmann, S., (2004) Stable isotope probing of rRNA and DNA reveals a and Corp, N. (1999) Major long-term changes in the ecol- dynamic methylotroph community and trophic interactions ogy of the Ythan estuary, Aberdeenshire, Scotland; how with fungi and protozoa in oxic rice field soil. Environ Micro- important are physical factors? Aquat Conserv Mar Freshw biol 6: 60–72. Ecosyst 9: 219–236. Lyons, M.G., Balls, P.W., and Turrell, W.R. (1993) A prelim- Remane, A., and Schlieper, C. (1971) Biology of Brackish inary study of the relative importance of riverine nutrient Water. New York, USA: John Wiley and Sons. inputs to the Scottish North Sea coastal zone. Mar Pollut Rysgaard, S., Risgaardpetersen, N., Sloth, N.P., Jensen, K., Bull 26: 620–628. and Nielsen, L.P. (1994) Oxygen regulation of nitrification McCaig, A.E., Embley, T.M., and Prosser, J.I. (1994) Molec- and denitrification in sediments. Limnol Oceanogr 39: ular analysis of enrichment cultures of marine ammonia 1643–1652. oxidisers. FEMS Microbiol Lett 120: 363–367. Schildkraut, C.L., Marmur, J., and Doty, P.J. (1962) Determi- McCaig, A.E., Phillips, C.J., Stephen, J.R., Kowalchuk, G.A., nation of the base composition of deoxyribonucleic acid Harvey, S.M., Herbert, R.A., et al. (1999) Nitrogen cycling from its buoyant density in CsCl. Mol Biol 4: 430–443. and community structure of proteobacterial beta-subgroup Skinner, F.A., and Walker, N. (1961) Growth of Nitrosomonas ammonia-oxidizing bacteria within polluted marine fish europaea in batch and continuous culture. Arch Mikrobiol farm sediments. Appl Environ Microbiol 65: 213–220. 38: 339–349. McCaig, A.E., Glover, L.A., and Prosser, J.I. (2001) Numer- Speksnijder, A.G., Kowalchuk, G.A., Roest, K., and Laan- ical analysis of grassland bacterial community structure broek, H.J. (1998) Recovery of a Nitrosomonas-like 16S under different land management regimens by using 16S rDNA sequence group from freshwater habitats. Syst Appl ribosomal DNA sequence data and denaturing gradient gel Microbiol 21: 321–330. electrophoresis banding patterns. Appl Environ Microbiol Stephen, J.R., McCaig, A.E., Smith, Z., Prosser, J.I., and 67: 4554–4559. Embley, T.M. (1996) Molecular diversity of soil and marine Manefield, M., Whiteley, A.S., Ostle, N., Ineson, P., and 16S rRNA gene sequences related to beta-subgroup Bailey, M.J. (2002) Technical considerations for RNA- ammonia-oxidizing bacteria. Appl Environ Microbiol 62: based stable isotope probing: an approach to associating 4147–4154. microbial diversity with microbial community function. Taberlet, P., Griffin, S., Goossens, B., Questiau, S., Rapid Commun Mass Spectrom 16: 2179–2183. Manceau, V., Escaravage, N., et al. (1996) Reliable geno- Muyzer, G., Dewaal, E.C., and Uitterlinden, A.G. (1993) Pro- typing of samples with very low DNA quantities using PCR. filing of complex microbial populations by denaturing gra- Nucleic Acids Res 24: 3189–3194. dient gel electrophoresis analysis of polymerase chain Tal, Y., Watts, J.E.M., Schreier, S.B., Sowers, K.R., and reaction-amplified genes coding for 16S ribosomal RNA. Schreier, H.J. (2003) Characterization of the microbial Appl Environ Microbiol 59: 695–700. community and nitrogen transformation processes associ- Nicol, G.W., Glover, L.A., and Prosser, J.I. (2003) Spatial ated with moving bed bioreactors in a closed recirculated analysis of archaeal community structure in grassland soil. mariculture system. Aquaculture 215: 187–202. Appl Environ Microbiol 69: 7420–7429. Wauters, G., De Baere, T., Willems, A., Falsen, E., and Phillips, C.J., Smith, Z., Embley, T.M., and Prosser, J.I. (1999) Vaneechoutte, M. (2003) Description of Comamonas Phylogenetic differences between particle-associated and aquatica comb. nov and Comamonas kerstersii sp nov for planktonic ammonia-oxidizing bacteria of the beta subdivi- two subgroups of Comamonas terrigena and emended sion of the class Proteobacteria in the northwestern Med- description of Comamonas terrigena. Int J Syst Evol Micro- iterranean Sea. Appl Environ Microbiol 65: 779–786. biol 53: 859–862. Powell, S.J., and Prosser, J.I. (1985) The effect of nitrapyrin Whitby, C.B., Hall, G., Pickup, R., Saunders, J.R., Ineson, P., and chloropicolinic acid on ammonium oxidation by Parekh, N.R., and McCarthy, A. (2001) 13C-incorporation Nitrosomonas europaea. FEMS Microbiol Lett 28: 51–54. into DNA as a means of identifying the active components Purkhold, U., Wagner, M., Timmermann, G., Pommerening- of ammonia-oxidizer populations. Lett Appl Microbiol 32: Röser, A., and Koops, H.P. (2003) 16S rRNA and amoA- 398–401.