In this study, we have begun to address some of the knowledge gaps by characterizing archaeal diversity in a salt marsh in Long Island Sound. We used the distinctive vegetation patterns of New England salt marshes, which are dominated by Spartina alterniflora and Spartina patens, as proxies for different edaphic conditions. S. alterniflora is found in two phenotypically distinct but genetically identical forms (2). Tall S. alterniflora (SAT) grows to heights of 1 to 2 m and is typically found at the edges of the marsh and along creek banks, while short S. alterniflora (SAS) may reach heights of only 30 cm and is found slightly higher on the marsh, where soil drainage is limited and sediments are more reduced (6, 10, 17). S. patens typically grows on the high marsh in areas that receive less flooding, where salinity is lower and redox is higher than on the low marsh (6).

Surface sediment samples (0 to 2 cm) were collected on six dates from October 2005 to October 2006 from the Barn Island tidal marshes in southeastern Connecticut from sites representing SAT, SAS, and S. patens (see references 20 and 21 for more-complete site descriptions). Triplicate cores (6.5 cm) were collected on all dates except in April, when only duplicate cores were collected. SAS and S. patens samples contained dense roots and consisted of dark, sulfidic-smelling mud with high clay content. SAT samples consisted of coarse sandy sediment in which roots were easily removed during processing. Pore water was extracted from sediments by centrifugation. Salinity was measured using a hand-held refractometer. pH was measured with a YSI model 100 meter (Yellow Springs Instruments, Yellow Springs, OH). Pore water ammonium (NH₄⁺) was measured using the phenol-hypochlorite method for seawater samples (19).

We extracted DNA from sediments by using a PowerSoil DNA isolation kit (MoBio, Carlsbad, CA). DNA concentrations were measured using a Quant-iT double-stranded-DNA HS assay kit and a Qubit fluorometer (Invitrogen, Carlsbad, CA). Archaeal 16S rRNA genes were amplified using the primers 21F (8) and 1492R (13) to produce nearly full-length (1,500-bp) sequences. Each 20-μl reaction mixture contained 10 μl iQ Supermix (Bio-Rad), 0.5 μM each primer, and 1 μl of a 1:10 dilution of DNA (approximately 2 to 10 ng). The cycle conditions were as follows: 95°C for 10 min, followed by 35 cycles at 95°C for 20 s, 53°C for 20 s, and 72°C for 2 min. PCR products were cloned into pSC vectors (Stratagene, Agilent Technologies, Santa Clara, CA). Inserts were amplified from selected clones by using T3 and T7 primers, and clones containing the correct-size insert were sequenced bidirectionally by High-Throughput Sequencing Solutions (Seattle, WA).

Analysis of 150 archaeal 16S rRNA gene sequences revealed distinct archaeal communities at each site (Fig. ​(Fig.1).1). Although sequences belonging to Euryarchaeota and Crenarchaeota were recovered from all three sites, Crenarchaeota sequences constituted at least 70% of the archaeal 16S rRNA sequences recovered from each site. Most sequences (60%) at the S. patens site were affiliated with group I.1a Crenarchaeota (CGI.1a) and shared ≥97% sequence identity with that of the aerobic ammonia-oxidizing “Candidatus Nitrosopumilus maritimus” isolate (12). At the SAT site, more than 25% of the clones clustered with CGI.1a, but no CGI.1a sequences were recovered at the SAS site. Sequences affiliated with Crenarchaeota group I.3b constituted the majority of clones at the SAS site.

Nonthermophilic Crenarchaeota have been reported as common constituents of marine communities in open ocean and estuarine habitats (e.g., references 7 and 11), but their presence in salt marshes has not been reported. Although Munson et al. (16) found archaeal communities in a United Kingdom salt marsh consisting exclusively of Euryarchaeota, the primers used in their study lacked the specificity to detect CGI.1a sequences. We also did not recover any sequences related to known methanogens and halophiles, which were quite prevalent in the United Kingdom study. Although community differences between the two marshes are at least partially due to differences in primer specificity, differences may also reflect sediment conditions. For example, the salinities reported to occur in samples from the United Kingdom study ranged from 49 to 53 practical salinity units (psu), which are considerably higher than the salinity values recorded at our sites (12 to 32 psu) (Table ​(Table1)1) and could explain the differences in archaeal communities.

Most Euryarchaeota clones in our study were closely related to sequences recovered from other marine environments, particularly sediments from Skan Bay, AK (Fig. ​(Fig.1).1). Euryarchaeota sequences were more frequently recovered from the SAT and S. patens sites (30% and 26% of all sequences, respectively) than from the SAS site (15%). Unfortunately, since there are no cultured representatives related to our Euryarchaeota clones, their physiology remains unknown.

Archaeal 16S rRNA gene richness was more than 3 times greater at the SAT site than at the SAS or S. patens site (Fig. ​(Fig.2).2). The SAT site is a coarse, sandy sediment along a creek bank that is well flushed and has ammonium concentrations that are typically 2 to 3 times higher than those at the SAS or S. patens site (Table ​(Table1).1). The increased diversity at the SAT site may be related to increased drainage and tidal flushing, due to its position along a creek bank, compared to the levels for the SAS and S. patens sites, which are higher on the marsh. SAT also experiences a greater range of salinity than SAS or S. patens. Other studies have reported lower levels of diversity of bacterial 16S rRNA genes (3), denitrifying bacteria (22), and ammonia oxidizing bacteria (4) associated with higher salinity, but the relationship between salinity and archaeal diversity has not been explored adequately.

Based on the unexpected prevalence of CGI.1a sequences in two of three clone libraries, we quantified the abundance of 16S rRNA genes related to “Ca. Nitrosopumilus maritimus” (designated the Nm cluster) at all three sites over 1 year by using quantitative PCR. Nm cluster 16S rRNA genes (ca. 415 bp) were amplified using the method of Könneke et al. (12), but we modified the forward primer (5′-CTGGAATGGTTTRTGTTYC) to target additional sequences recovered in our study. The specificity of Nm cluster primers was confirmed by routine melting curve analysis (95°C for 1 min, 57°C for 1 min, and then a 0.5°C increase every 10 s) and sequencing of over 50 clones generated from the three sites in our study. All sequences were related to the Nm cluster (data not shown). Standard curves were generated in each run by using standards ranging from 18 to 1.8 × 10⁵ gene copies/μl of DNA and used to estimate gene abundance in the sediment samples. Standards were plasmid DNAs purified from archaeal 16S rRNA clones recovered from the same marsh samples. We also tested for inhibitory effects by calculating the slopes of each sample at different dilutions (ranging from 1:5 to 1:15). Dilutions ranging from 1:8 to 1:12 gave similar slopes (coefficient of variation = 11.6%), so we used 1:10 dilutions for final analysis. Additionally, slopes were not significantly different among samples from different sites (analysis of covariance; P = 0.46).

Nm cluster 16S rRNA gene abundance ranged from below the detection level in some samples at SAS to 1.4 × 10⁹ copies g (dry weight) sediment⁻¹ at the S. patens site and were significantly lower at the SAS site than at the SAT or S. patens site (Kruskal-Wallis test; P < 0.0001) (Fig. ​(Fig.3).3). Although Nm cluster 16S rRNA gene abundance was slightly lower at all sites in October, abundance remained relatively high from March to July at the SAT and S. patens sites, and we did not detect any significant seasonal variation during the months that we sampled. Finding numbers as high as 10⁹ was surprising, particularly given that previous reports of total cell counts in marine sediments typically range from 10⁸ to 10¹⁰ cells per gram (9, 18). Although we measured only the gene copy number, there is only one 16S rRNA gene copy according to the genome of “Ca. Nitrosopumilus maritimus.” If this is true for all Nm cluster archaea, then gene copy number should approximate cell numbers, suggesting that Nm cluster archaea could account for a significant component of the microbial community in some samples. The lack of Nm cluster sequences in the clone library from SAS is likely due to the low abundance of these archaea in the samples. However, using Nm cluster primers to generate clones, we were subsequently able to confirm Nm cluster 16S rRNA sequences at the SAS site (data not shown).

The largest differences in Nm cluster 16S rRNA gene abundance were found between samples at the SAS and S. patens sites, yet there were few differences in salinity, pH, or NH₄⁺ concentration between these sites (Table ​(Table1).1). Differences in Nm cluster 16S rRNA gene abundance may be related to differences in alternative edaphic conditions. For example, significant differences in sulfide and redox levels have been reported previously for these two sites, with SAS sites showing higher sulfide levels and lower redox levels than S. patens sites (21). Additionally, factors that influence the oxygenation of the sediments, such as bioturbation (5) and oxygen from the roots (15), were not measured in this study but may be of particular importance when the Nm cluster archaea are also aerobic ammonia oxidizers, like “Ca. Nitrosopumilus maritimus.”

Our results indicate that nonthermophilic Crenarchaeota are prevalent members of microbial communities in salt marsh sediments in Long Island Sound. To our knowledge, this is the first report of group I Crenarchaeota in salt marshes, and our data indicate that these crenarchaea constitute a stable component of archaeal communities in some areas of the marsh. The differential distribution of Nm cluster 16S rRNA genes also suggests that alternative environmental variables (e.g., sulfide, redox, or plant root exudates from different grasses) may be important in regulating their distribution, but future studies are needed to address these hypotheses.