OUP user menu

Comparison of diazotroph community structure in Lyngbya sp. and Microcoleus chthonoplastes dominated microbial mats from Guerrero Negro, Baja, Mexico

Enoma O. Omoregie, Lori L. Crumbliss, Brad M. Bebout, Jonathan P. Zehr
DOI: http://dx.doi.org/10.1016/S0168-6496(03)00301-5 305-318 First published online: 1 March 2004


The nitrogenase activity and phylogenetic diversity of nitrogen fixing microorganisms in several different cyanobacterial mat types from Guerrero Negro, Baja California, Mexico were investigated by acetylene reduction assay, and by amplification and sequencing of the nitrogenase nifH gene. Acetylene reduction assays performed on a Lyngbya sp. and two Microcoleus chthonoplastes dominated microbial mats showed a typical diel pattern of nitrogenase activity in these mats. The highest rates of activity were found at night, with 40 and 37 μmol C2H4 m−2 h−1 measured in the Microcoleus mats, and 9 μmol C2H4 m−2 h−1 in the Lyngbya mat. Nitrogenase sequences were obtained that clustered with sequences from cyanobacteria, γ-Proteobacteria, and cluster 3 of nifH. In addition, novel and divergent sequences were also recovered. The composition of nifH sequence types recovered differed between the Lyngbya and Microcoleus mats. Interestingly, nifH sequences belonging to filamentous cyanobacteria were absent in most mat samples even though both mats were dominated by filamentous cyanobacteria. nifH sequences clustering with those of unicellular cyanobacteria were found, some of which were virtually identical to the nifH sequence from Halothece sp. MPI96P605, which had previously been isolated from the mat. In manipulation experiments, the Lyngbya and Microcoleus mats were allowed to re-colonize a cleared surface. In these developing mats, nifH sequences not previously observed in the mats were discovered. Our results showed that organisms capable of N2 fixation were present in N2 fixing mats, that the composition of the N2 fixing communities differs between mats, and that filamentous cyanobacterial diazotrophs may not have a large role in the early stages of mat development.

  • Nitrogenase
  • nifH
  • Polymerase chain reaction
  • Cyanobacterial mat
  • Nitrogen fixation
  • Microcoleus
  • Lyngbya

1 Introduction

Microbial mats are composed of diverse assemblages of microorganisms that catalyze high rates of carbon, nitrogen, and sulfur transformations. Modern day (extant) microbial mats are analogous to early microbial assemblages that first populated the Earth billions of years ago [13]. The productivity of many ecosystems is limited by the availability of fixed inorganic or organic nitrogen [4,5] and biological nitrogen (N2) fixation can be a source of nitrogen in an otherwise nitrogen limited environment [68]. Cyanobacterial mats exhibit high rates of photosynthesis [9, 10], and since they are found in low nutrient environments, also often exhibit high rates of N2 fixation [11].

Mats are composed of diverse microorganisms of which many are potentially N2 fixers [1214], but it is difficult to ascertain which of the many taxa comprising a microbial mat are involved in this process [15]. It is also difficult to ascertain whether phototrophic or heterotrophic metabolisms support the energetically costly process of N2 fixation in the mat. Photosynthetic microbial mat communities often exhibit distinct vertical lamination, typically a top layer composed primarily of cyanobacteria and diatoms, a middle layer consisting mostly of purple bacteria and a bottom layer comprised mainly of anaerobic bacteria [10]. The penetration of oxygen and light in the different layers determines microbial community structure and composition [10]. Previous work has shown N2 fixation to be under the control of the activities of photoautotrophs, such as cyanobacteria and/or diatoms (the latter have not been shown to fix N2, but diatoms affect oxygen sensitive nitrogen fixation due to photosynthetic oxygen evolution) [15], but other studies have indicated that heterotrophs could be involved, at least in some mats [16].

Nitrogenase is composed of two metalloproteins encoded by the nifHDK genes [17]. The nitrogenase genes are found throughout the prokaryotes, and are highly conserved among closely related organisms indicating either an early evolution or early lateral gene transfer [1820]. The conserved nature of the nitrogenase genes makes it a useful marker for the genetic capability of N2 fixation [19,21] and previous studies of cyanobacterial mats have shown there to be a diverse array nifH genes [12,13].

The Exportadora de Sal (ESSA) salt works in Guerrero Negro, Mexico, has well-developed mat communities that have been the subject of several studies. Previous work on these mat communities has elucidated biogeochemical transformations of C, N, O, and S [9,10,22,23], as well as identified some of the organisms responsible for these transformations [14,2426]. Investigations into N2 fixation on the mats from ESSA have focused on two cyanobacteria dominated mats: one by Lyngbya sp., and the other by Microcoleus chthonoplastes. These mats have distinctly different biogeochemical characteristics, that typically include vastly different rates of N2 fixation [23,27]. Prior work on these mats has illustrated the pattern of N2 fixation as well as quantified their requirement for fixed N derived through N2 fixation [23]. However, to date, there has not been any work conducted to identify the nifH containing organisms in these mats.

The primary objective of this study was to attempt to elucidate the organisms most important in mat N2 fixation using comparisons of the diversity and community structure of diazotrophic organisms populating a Lyngbya sp. mat and a M. chthonoplastes mat from ESSA in Guerrero Negro, Mexico [23]. Furthermore, experiments were performed in which portions of the well-developed mats were excised and replaced with concrete blocks to determine which microorganisms were involved in the early stages of mat formation (when nitrogen demand would presumably be high, and rates of nitrogen recycling low).

2 Materials and methods

2.1 Field site

The samples used in this study were collected from the extensive mat communities found at the ESSA salt works in Guerrero Negro, Baja California Sur, Mexico. ESSA is located on the Pacific coast approximately 700 km south of the US–Mexico border. Both Lyngbya sp. and M. chthonoplastes are filamentous non-heterocystous cyanobacteria that help to form extensive, stable mat communities by excreting a thick adhesive polysaccharide sheath. The M. chthonoplastes dominated mats were 4–5 cm thick, cohesive, smooth surfaced mats that were collected from a subtidal hypersaline pond (Area 4, salinity=ca. 90‰) [14,21,28]. The Lyngbya sp. dominated mats were 2–3 cm fibrous, rough surfaced mats that were collected from a tidal flat that undergoes frequent alternating periods of desiccation/aeration and tidal flooding [14,21,28].

Three types of mats were studied: established mats, developing mats and greenhouse mats. Established mat samples were collected from intact natural mats, developing mat samples were collected from mats that developed on an artificial substrate placed in the mat, and the greenhouse mats were obtained from field collected Microcoleus mats maintained (for long term experiments under controlled conditions) at a greenhouse facility at NASA Ames [28]. Samples of established mats were collected in June 2001. Cores, approximately 25 mm in diameter, were taken directly from the field sites in Mexico. The top 3–5 mm of the cores were sectioned and placed into plastic Petri dishes that were promptly placed in liquid nitrogen until they could be transferred into a −80°C freezer. Developing mat samples were also collected in June 2001. Thirteen months earlier, 36×36×5-cm portions of both mats had been excised and replaced with concrete blocks, placed so that the surface of the block was flush with the surface of the mat; there was no gap between the sides of the mats and the sides of the block. Three blocks were placed into each of the field sites. In June 2001, a sample consisting of the entire thickness of these mats (ca. 2 mm) was removed from each of the concrete blocks, placed into a plastic Petri dish, promptly frozen in liquid nitrogen, and later transferred to a −80°C freezer.

The greenhouse mat was a section of established Microcoleus mat originally also collected in Guerrero Negro, in June 2001, but which had been subsequently transferred to a greenhouse at the NASA Ames Research Center in Mountain View, CA, USA. These mats had been maintained for approximately 1 month before the samples were collected for this study (in July 2001). Cores taken from the greenhouse mat were collected using a 1 cm inner diameter stainless steel cylinder; then were sectioned into 5-mm increments (top, middle, and bottom) using a clean scalpel. After sectioning, the core samples were put into cryovials and placed directly into liquid nitrogen until they could be transferred to a −80°C freezer.

2.2 Light microscopy

Microbial community composition was observed at the light microscope level using simple squash mounts of small portions of the microbial mats. A Nikon Microphot FX/A (Nikon USA) having optics for differential interference contrast microscopy was used. No attempt was made to employ quantitative microscopy techniques.

2.3 Oxygen microelectrode and photosynthetic rate measurements

We used Clark-type oxygen microelectrodes incorporating guard cathodes (Diamond General 737-GC, Diamond General Development, Ann Arbor, MI, USA) to measure both oxygen concentrations and rates of oxygenic photosynthesis in the mats. Oxygen microelectrodes were positioned using motorized micromanipulators. The micromanipulators were controlled by, and the oxygen electrode signal data acquired with, custom software written in the LabVIEW programming environment (National Instruments, Austin, TX, USA). Oxygen microelectrode signal output was calibrated using a two-point calibration, (1) the air-bubbled water overlying the mats, and (2) zero oxygen (taken deep within the mats). The exact value of the water column oxygen concentration can then be calculated using published equations [29]. The second point in the calibration, the output of the electrode at an oxygen concentration of zero, was provided by the asymptotic minimum of electrode current within the permanently anoxic lower parts of the mat. Oxygenic photosynthesis was quantified using the dark shift technique [30,31].

2.4 Porewater extraction and analysis

Cores of microbial mats for porewater analysis were obtained using cut-off 20-ml syringes and sectioned at 2-mm intervals using a metal blade. Porewater was extracted from the mat sections by gentle exchange (20 min on a rocker table) with ca. 15 volumes of an isotonic solution of sodium chloride. This method is much gentler than centrifugation or squeezing, which, in these mats, may recover some intracellular contents (data not shown). Porewater ammonium concentrations were subsequently quantified using the phenol hypochlorite method [32].

2.5 Acetylene reduction assay

Nitrogenase activity was measured by the acetylene reduction assay following the procedure of Bebout et al. [15]. Cores of ca. 20 mm2 by 5 mm were placed in 38-ml serum bottles containing 20 ml of water (overlying water from each mat). Three replicate cores were taken from each mat during each sampling period. Bottles were sealed with rubber stoppers and 2 ml of headspace removed, and replaced with 5 ml of acetylene. Bottles were returned for incubation to the flumes, trays or ponds from which they were sampled to maintain similar light and temperature conditions. After 3–4 h, incubations were terminated by shaking each bottle for 10 s. Two milliliters of head space were injected into a 9.1-ml serum bottle filled with a 3.4 M solution of NaCl, displacing the salt solution through a vent needle, for short term storage. Ethylene concentrations were measured (usually within hours) by gas chromatography. Shimadzu GC14A or GCMini2 gas chromatographs (Kyoto, Japan), having 2-m Porapak N columns held at 80°C and flame ionization detectors, were used to quantify 0.1-ml aliquots of the stored headspace samples.

2.6 Genomic DNA extraction from mats

One replicate core from each mat sample was used for nucleic acid extraction. Genomic DNA was isolated from microbial mats by modification of the MoBio UltraClean Soil DNA Kit (MoBio Laboratories, Carlsbad, CA, USA). Approximately 50 mg of mat sample was added to a 2-ml Bead Solution tube, then 60 μl of solution S1 was added and the tube was inverted once to mix. Next, 200 μl of inhibitor removal solution (IRS) was added to the tube, which was then vortexed for 10 s. The mixture in the bead beat tube was heated at 65°C for 10 min, vortexing every few minutes. Samples were then bead beaten in a Fast Prep machine (Qbiogene Carlsbad, CA, USA) for 1 min at speed 6. Samples were centrifuged at 10 000×g for 5 s, then 250 μl of lysate was transferred to a clean microcentrifuge tube. Two hundred microliters of extraction buffer (550 μl of the solution in a bead beat tube with 60 μl of solution S1, 200 μl of IRS) and 250 μl of solution S2 were added. Samples were then processed according to the manufacturer's protocol.

2.7 Genomic DNA extraction from cultures

A culture of Halothece sp. MPI96P605 [26] was provided by F. Garcia-Pichel (Arizona State University). A small volume of culture was pelleted and resuspended in 50 μl of 1×TE buffer. This solution was then added to a bead beat tube (MoBio UltraClean Soil DNA Kit, MoBio Laboratories). Sixty microliters of solution S1 was added and the tube was inverted once to mix. Two hundred microliters of IRS was then added to the tube. The bead beat tube was then processed as described above, with the exception that no heat step was performed during the extraction.

2.8 Nested polymerase chain reaction (PCR) amplification

Two sets of primers were used in a nested PCR. The outer primer sequences were 5′-TTYTAYGGNAARGGNGG-3′ and 5′-ATRTTRTTNGCNGCRTA-3′ for nifH4 and nifH3, respectively [33]. For the nested reaction, nifH1 (5′-TGYGAYCCNAARGCNGA-3′) and nifH2 (5′-ADNGCCATCATYTCNCC-3′) were used in the second stage [33]. The final concentrations in each 50 μl reaction were 4 mM MgCl2, 1×PCR buffer, 0.8 mM total dNTP, 1 μM each of forward and reverse primer, and 2.5 U of either Pfu (Stratagene, La Jolla, CA, USA) or Taq (Promega, Madison, WI, USA) DNA polymerase. This core mix was filtered through a 100 MW Biomax filter (Millipore, Billerica, MA, USA) and centrifuged at 3000 rpm for 7 min before enzyme was added to the mix, in order to reduce DNA contamination in reagents.

The first round of PCR contained 48 μl of core mix and 2 μl of DNA extract, while the second round reactions each contained 49 μl of core mix and 1 μl of first round PCR product. Cycling conditions for both rounds were as follows: 94°C for 1 min, followed by 30 cycles of 94°C for 1 min, 57°C for 1 min, and 72°C for 1 min. Final annealing at 72°C for 7 min was followed by a 4°C hold. PCR products were run in a 1.5% agarose gel with 1 μg ml−1 ethidium bromide in 1×TAE buffer, and analyzed under a UV transilluminator (Bio-Rad, Hercules, CA, USA).

2.9 Cloning and sequencing

The PCR products were extracted from the gel and purified with the QIAquick gel extraction kit (Qiagen, Hilden, Germany). Ligations and transformations were done using the Promega pGEM-T vector kit and JM109 High Efficiency Competent Cells. Standard Promega protocols were used for both the ligation and transformation reactions. After incubation, the transformation reactions were plated on LB agar plates pretreated with ampicillin (100 μg ml−1), 0.5 mM isopropyl-β-d-thiogalactose, and 80 μg ml−1 X-gal. Individual colonies were isolated into culture tubes containing 3 ml LB broth treated with ampicillin (100 μg ml−1), then incubated with agitation (200 rpm) at 37°C for 16–20 h. The cell cultures were purified using a QIAprep Spin Miniprep Kit (Qiagen). DNA was eluted with 50 μl EB buffer. Purified plasmids were sequenced in one direction with an ABI Prism 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) using the ABI Prism BigDye Terminators v3.0 Cycle Sequencing kit (Applied Biosystems). Sequences were edited and predicted amino acid sequences generated using the editor available within the software program GCG (Accelrys, San Diego, CA, USA). Probabilistic alignments were generated using the HMMER program in GCG. Percent identities were generated from genetic distances in GCG, with Kimura correction and the distances normalized to 100%. Representative sequences were chosen and then sequenced in the reverse direction. Representative sequences were sequences that shared less than 98% identity with any other sequence from the respective mat sample, or were one out of a group of two or more sequences from the same sample that had 98% or greater identity. Phylogenetic distances were calculated from predicted amino acid sequences using the algorithm of Tajima-Nei [34] in the software program TREECON for Windows [35]. Phylogenetic tree reconstructions from distance approximations were done using the neighbor-joining method [36] in TREECON. The trees were bootstrapped (100 replicates) using TREECON.

3 Results

3.1 Physical and chemical characterization of mats

The Lyngbya and Microcoleus mats used in this study were visibly laminated at a millimeter to sub-millimeter scale (Fig. 1), although the laminations were much more apparent in the Microcoleus mat. Light microscope observations of the Lyngbya mats revealed a number of cyanobacterial morphotypes, in addition to that of Lyngbya spp., including filamentous species as well as a few unicellular cyanobacteria. Purple and colorless sulfur bacteria were infrequently observed in the Lyngbya mats. Microcoleus mats revealed a diversity of microorganisms far exceeding that of the Lyngbya mats, including diatoms, filamentous and unicellular cyanobacteria, purple and colorless sulfur bacteria, as well as large numbers of other microbes which were more difficult to distinguish at the light microscope level. Microorganisms indistinguishable at this level included filamentous (colored and colorless) and coccoid morphologies. Nematodes were the most prominent metazoans, but a number of other grazing organisms were observed including flagellated and ciliated protozoans.


Photographs of Lyngbya (a–c) and Microcoleus (d–f) mats. Panels a and d show the overall appearance of the mats. Panels b and e show the vertical lamination and stratification in the mats. Panels c and f are typical light microscope views of samples taken from the mat surface.

Oxygenic photosynthetic activity, even at irradiances characteristic of those observed at high noon at the field sites, was confined to the top 2 mm in both mats (Fig. 2). Profiles of nitrogenase activity, performed at the much coarser resolution of 5 mm, showed that detectable nitrogenase was also confined to the surface interval in both mats (data not shown). In the case of the Lyngbya mat, the surface 5 mm frequently comprised the entire thickness of the community. Nitrogenase activity in general is strongly repressed by high concentrations of ammonium [37]. Profiles of porewater ammonium, presumably produced by the decomposition of buried mat organic matter, revealed high concentrations of ammonium at depth, and a pronounced depletion in the surface intervals of the Microcoleus mat (Fig. 2). Ammonium concentrations in the Lyngbya mat are frequently at or below the limit of detection of our protocol (ca. 5 μM) (data not shown). A reliable ‘deep source’ of ammonium, similar to that observed in the Microcoleus mat, does not exist due to the very loose attachment of the Lyngbya mat to its underlying (nutrient poor) sandy substrate.


Rates of oxygenic photosynthesis (circles) and concentrations of dissolved oxygen (squares) in the Lyngbya (a) and Microcoleus (b) mats at saturating irradiances near those typical of high noon at both field sites. Profiles of porewater ammonium in Microcoleus mats (c) shortly after collection (▪) and after 1 year of growth in the greenhouse (◻).

3.2 Acetylene reduction assay

Acetylene reduction activity was not detected below 5 mm depth in the mat in preliminary assays on the greenhouse Microcoleus mats, hence all subsequent assays were performed on only the upper 5 mm. All mats assayed (Fig. 3) had distinct diel variability in acetylene reduction rates with maximal nitrogenase activity occurring during the night and negligible nitrogenase activity during the daylight hours. This is consistent with the acetylene reduction results obtained during assays performed on these microbial mat communities in 1990 and 1991 [23], and in October 2001 [38]. The highest rates of activity, 40 μmol C2H4 m−2 h−1, were observed in the greenhouse mat. The maximum rates for the established Lyngbya and Microcoleus mats were 9 and 37 μmol C2H4 m−2 h−1, respectively. Developing mats were not assayed for nitrogenase activity.


Patterns of nitrogenase activity (●) and irradiance (▲) for the greenhouse (a), established Microcoleus (b), and established Lyngbya (c) mats.

3.3 Community composition of nifH containing organisms from established mats

nifH genes were amplified from the established Lyngbya and Microcoleus mats as well as the greenhouse Microcoleus mat. A total of 141 sequences were obtained. The recovered nifH sequences ranged from 68 to 100% identity to previously reported nifH protein sequences (Table 1). The phylogenetic positions of the representative nifH sequences are shown in Figs. 46. For ease of reference, sequences were divided into sequence groups defined by their closest isolate, and their position in the phylogenetic tree. The recovered nifH sequences fell into 25 different sequence groups. In general, there are four major clusters of nifH sequences [19,39]. Clusters 1 and 3 contain the conventional Mo-containing nitrogenases, cluster 2 the second alternative (Fe) nitrogenases, and cluster 4 the deeply divergent nifH-like proteins. Of the 25 sequence groups, five were in cluster 1, and 19 were within cluster 3. One group of six divergent nifH sequences (group 33, Fig. 6) was recovered from the established Lyngbya mat. This group of sequences did not fall into any of the previously described clusters of nitrogenase, and may not represent a functional nitrogenase gene. The majority of the 141 sequences in the mat clone libraries (Table 1) were in cluster 3 (68%), or in the cyanobacteria (19%) and γ-Proteobacteria (9%) groups of cluster 1. None of the recovered sequences represented nifH sequences from α-Proteobacteria or the nifH-like homologs (cluster 4) phylotypes.


Relationship of mat nifH sequence groups (Figs. 46) to nifH sequences from cultivated isolates


nifH protein tree of cluster 1 showing the phylogenetic position of the sequences recovered from each mat. Numbers in parentheses show the number of sequences recovered having 98% or greater identity to the indicated sequence.


Top half of nifH protein tree of cluster 3 showing the phylogenetic position of the sequences recovered from each mat. Numbers in parentheses show the number of sequences recovered having 98% or greater identity to the indicated sequence.


Lower section of nifH protein tree of cluster 3 showing the phylogenetic position of the sequences recovered from each mat. Numbers in parentheses show the number of sequences recovered having 98% or greater identity to the indicated sequence.

Almost half of the cyanobacterial sequences (48%) recovered from the mat samples originated from the greenhouse Microcoleus mat (mid-section). These sequences were most closely related to the unicellular halotolerant cyanobacterium Halothece. Several sequences (groups 11 and 12) from the established Lyngbya mat, the greenhouse Microcoleus mat (all sections) and the established Microcoleus mats were most closely related to protein sequences from unicellular cyanobacteria, such as Cyanothece, Myxosarcina, Xenococcus and Halothece. One sequence from the greenhouse Microcoleus mat (top section) was most closely related to the second copy nitrogenase protein from the heterocyst forming cyanobacterium Anabaena variabilis. Several sequences (group 1) from the greenhouse Microcoleus mat (top section) and the established Lyngbya and Microcoleus mats were most closely related to nitrogenase protein from the γ-Proteobacterium Azotobacter vinelandii. One sequence (group 13) recovered from the greenhouse mat was in cluster 1, but did not group within any of the major cluster 1 groups. This sequence was most closely related to the filamentous non-heterocyst forming cyanobacterium Phormidium.

Cluster III nifH protein sequences were recovered from the established Lyngbya and Microcoleus mats as well as the greenhouse Microcoleus mat (all sections). None of the cluster 3 sequences (Figs. 56) were closely related (Table 1) to previously recovered sequences from cultivated isolates. They ranged from 81 to 92% identity to previously recovered isolate sequences. The most similar sequences from cultivated isolates were those from the δ-Proteobacteria and Low GC gram positives. Most of these sequences were most similar to sequences recovered from other mats, stromatolites and/or ecosystems that contain anaerobic environments.

3.4 Community composition of nifH containing organisms from developing mats

Seventy nifH sequences were recovered by PCR amplification from the developing Lyngbya and Microcoleus mats. These sequences fell into 17 different sequence groups (Table 1). They ranged from 83 to 97% identity to previously recovered nifH protein sequences. The majority of the recovered sequences were in cluster 3 (51%), followed by the cyanobacteria (30%) and γ-Proteobacteria (17%) groups.

Several of the cluster 1 sequences from both mats were most closely related to the nifH sequence from A. vinelandii, and Klebsiella sp. Two sequences from the developing Microcoleus mat were most closely related to the nifH gene from Burkholderia. Most of the cluster 1 sequences (81%) recovered from these two mats originated from the Lyngbya mat, the majority (77%) of which were most closely related to the nifH sequences from cyanobacteria. These sequences (groups 6–9) were most similar to sequences from filamentous cyanobacteria such as Anabaena and Phormidium, as well as unicellular cyanobacteria, such as Dermocarpa and Synechocystis. Only one cyanobacterial sequence was recovered from the developing Microcoleus mat, which was most closely related to a second copy of nifH in A. variabilis. Cluster 3 sequences made up just over half of the sequences recovered from both mats. As with the sequences from the greenhouse and the established mats, these sequences were not closely related (81–92%) to any previously recovered sequences from cultivated isolates, but were most closely related to sequences recovered from environments that contain anaerobic regions.

4 Discussion

We observed high rates of nitrogenase activity (Fig. 3) and low concentrations of inorganic nitrogen (Fig. 2) in the surface layers of both the greenhouse and the established Microcoleus mats. This is consistent with the hypothesis that N2 fixation in the surface layers of the mats is important in supporting the nitrogen demands associated with primary production. Further support for this hypothesis is provided by the increase in nitrogenase activity (and concomitant decrease in porewater nitrogen) observed in mats incubated for long periods of time (up to 1 year) without nutrient additions [28]. The observed patterns of nitrogenase activity (Fig. 4) illustrate the typical temporal separation utilized by the organisms within the mats to avoid the inhibitory O2 levels present during photosynthetically active times of the day. Similar patterns of nitrogenase activity have been observed in other mats dominated by filamentous, non-heterocystous cyanobacteria [15,40], as well as mats dominated by unicellular cyanobacteria (Gloeothece sp. and Oscillatoria sp. mats) [41].

The established Microcoleus mats are subtidal mats that have been shown to be at or near steady state with respect to carbon exchange with their environment [9,23]. Essentially, there is very little net flow of carbon into or out of these mat systems, therefore the demand for an external source for N2 is low. The rates of 40 and 37 μmol C2H4 m−2 h−1 are almost three times higher than the rates typically observed in these mats [23]. The reason for the changes in rates of nitrogenase activity within these mats is unclear, but may reflect seasonal changes of N dynamics within the mat.

In contrast, the established Lyngbya mat is an intertidal mat that is periodically flooded, desiccated and physically displaced as a result of the tidal cycles. These cycles force the Lyngbya mat to be in pioneering mode where its demand for fixed N is high [23]. The maximum rate of 9 μmol C2H4 m−2 h−1 measured in this mat is low for a mat that has such a high N demand [23]. However, these mats were incubated in situ in Pond 4 in Guerrero Negro; the salinity of this pond is substantially higher (85‰) than the lagoon water (40‰) to which these mats are typically exposed. The effects of salinity on N2 fixation are not well understood, but salinity has been shown to influence nitrogenase activity in microbial mats [42].

The high prevalence (28–87% of individual mat clone libraries) of cluster 3 sequences (Table 1 and Figs. 56) in the clone libraries suggests that these organisms could potentially have a role in N2 fixation in these mats. It is difficult to identify the types of organisms responsible for these sequences as they are not very similar to sequences previously recovered from isolates (Table 1). Most of the sequences recovered in this cluster of the nifH phylogeny are derived from anaerobic or micro-aerobic environments, such as copepod and termite guts, rice roots, sea grass sediments and other microbial mats. Sequences from organisms such as the sulfate reducers, Clostridium, spirochaetes, green sulfur bacteria, and methanogens also cluster in this part of the nifH phylogeny. Therefore, it is very likely that the organisms that these sequences originate could be anaerobes. Diazotrophy in several anaerobe species, and in anaerobic environments, has been demonstrated [4346]. More recently, work by Steppe and Paerl in 2002 [16] has suggested that sulfate reducing bacteria (SRB) may be involved in N2 fixation in cyanobacterial mats. It is unclear whether some of the cluster 3 sequences recovered in this study are from SRB, as none of them show high similarity to SRB sequences. However, the SRB do not show a tight coherence within cluster 3, so the lack of clustering of these sequences to previously cultivated SRB sequences is not necessarily an indication that these sequences are not derived from SRBs.

The cluster 1 nifH protein sequences (Table 1 and Fig. 4) were the next most abundant group of sequences after the cluster 3 sequences. The sequences within this cluster were most similar to those of unicellular and filamentous cyanobacteria as well as members of the γ-Proteobacteria. It is not surprising that filamentous and unicellular cyanobacterial nifH sequences were recovered from these mats, as patterns of nitrogenase activity are consistent with the presence of these types of cyanobacteria. These types of cyanobacteria typically fix N2 during the night [47,48], which is consistent with the observed patterns of nitrogenase expression in Fig. 3. It is also interesting to note that unicellular cyanobacterial sequences have been recovered in this study, which is also consistent with the observed patterns of N2 fixation. Recently, these types of cyanobacteria have been implicated in N2 fixation in the Pacific and Atlantic Oceans [49,50], but their role in cyanobacterial mats had not been previously recognized.

The fact that cyanobacterial nifH sequences made up a small portion of the sequences recovered from the established Lyngbya and Microcoleus mats as well as the top 5 mm of the greenhouse mats suggests that cyanobacteria may have a smaller role in fixing N2 in mats than previously believed [15,40]. However, amplification of nifH DNA only indicates which organisms in the mats possess the ability to fix N2, not necessarily those that are active. Therefore, the relative abundance of cyanobacterial nifH sequences derived from genomic DNA may not reflect their role in N2 fixation. In fact, studies with these mats indicate that not all nifH phylotypes are expressed at all times. However, diverse nifH sequences from cyanobacteria and cluster 3 groups were found to be expressed at night in the Guerrero Negro mats [51]. Detailed studies of expression are needed to determine if many or all of the phylotypes detected in DNA are expressed at certain times or under certain conditions.

The identity of the organisms in groups 1–5 (Fig. 4) is unclear. Therefore, it is difficult to determine whether these sequences represent sequences derived from heterotrophic or autotrophic microorganisms. Both phototrophic and heterotrophic metabolisms have been shown to be involved in N2 fixation in microbial mats [52,53]. These sequences cluster deep within the γ/β-Proteobacterial cluster of nifH when larger sets of nifH sequences are used in the analysis (clustering is better supported using the entire database of over 2000 nifH sequences, data not shown).

All the developed mats in this study are characterized by an abundance of filamentous cyanobacteria using light microscopy (Fig. 1). However, it is interesting that sequences which were most closely related to nifH from filamentous cyanobacteria were only 0–3% of the sequences recovered from each of the established mats. To date, no isolates of Microcoleus have been shown to fix N2 or possess nifH genes [54]. At least one species of Lyngbya[55] has been shown to possess nifH, however, none of the sequences from the established mats are similar to this sequence. The lack of sequences grouping with filamentous cyanobacteria indicates that the dominant mat building organisms may be incapable of fixing N2. It may be energetically advantageous for mat building organisms to obtain their N through other means. The detection of green and purple sulfur bacteria, both of which have known N2 fixing members, occurring in intimate associations with Microcoleus, in fact within Microcoleus sheaths [56], lends credence to this idea.

We attempted to address the commonly held hypothesis [57] that filamentous nitrogen fixing cyanobacteria are important in the development of microbial mats by sampling mats in the process of formation on concrete blocks. The blocks had been in place for 13 months. Although concrete does not completely mimic the chemical composition of the bottoms of the ponds, or of the lagoon sediments, it does contain many of the same minerals (e.g., calcium carbonate, silicates, and gypsum) and is of sufficiently low porosity to have excluded potential sources of dissolved inorganic nitrogen from below the mats. In the case of the blocks placed into Area 4 (developing Microcoleus), the mats that developed on the surface of the blocks appeared very similar to the mats that grew on other surfaces in the ponds, including the natural pond bottom. Casual microscopic observations revealed that the community growing on the blocks was dominated by small filamentous cyanobacteria, similar in appearance to Phormidium sp., as well as unicellular cyanobacteria and diatoms. In the case of the blocks placed into the tidal flats inhabited by the Lyngbya mats, the blocks were colonized by overgrowth of the natural mats from the sides of the block. Microscopic observations of these developing mats revealed a community of M. chthonoplastes and Lyngbya sp. indistinguishable from the surrounding mat.

Although microscopic observations of both mats showed a high prevalence of cyanobacteria, especially filamentous types, filamentous cyanobacterial phylotypes appeared to be absent in the clone libraries of both developing mats. The sequences in group 8, which are from the developing Lyngbya mat, cluster with a sequence derived from Lyngbya sp. However, they are much more closely related (Table 1) to a sequence from a unicellular cyanobacterium, Dermocarpa. This result, coupled to the fact that sequences closely related to that of filamentous cyanobacteria are only a small part of the developing Microcoleus mat, indicates that filamentous diazotrophs may not have a large role in the early stages of mat development. Although there are differences in groups of cluster 3 sequences that were recovered from the various mat samples, cluster 3 organisms are present and likely to be important in all stages of mat development [22].

It is not surprising to find cluster 3 nifH sequences in the middle and bottom sections of the greenhouse Microcoleus mat, even in the absence of detectable nitrogenase activity. Many of the previously isolated organisms which have cluster 3 nifH genes are anaerobes [13,19]. It is interesting to note that many of the sequences recovered from those regions form groups different from those that were recovered from the top 5 mm in the mat, indicating that there is a distribution of different nifH-containing anaerobes throughout the mat.

It is curious to find unicellular cyanobacterial nifH sequences in the lower regions of the mat. These sequences presumably represent organisms that would not survive in a permanently dark environment. However, the organisms that contain these genes may be cells that have been buried but not yet fully degraded in the mat.

Novel protein sequences similar to the ones reported here from group 33 have been recovered in other studies [58]. It is not possible to say whether the novel protein sequences such as those that were recovered from the established Lyngbya mat are proteins that function in N2 fixation. These sequences do not seem to be part of any of the recognized clusters of nitrogenase, and therefore their exact function is unclear. However, the divergent nifH sequence (group 13) recovered from the top 5 mm of the greenhouse mat may belong to an as yet unidentified diazotroph. Although this sequence is not closely related to any nifH sequence, its grouping within cluster 1 suggests that it is in fact a nifH sequence.

Our work to determine the N2 fixing microorganisms in the mat communities located at ESSA complements several studies utilizing molecular tools to study the microorganisms that populate these mat communities. These studies have revealed a large diversity of organisms capable of photosynthesis [14,25,26,59], as well as organisms capable of sulfate reduction [24]. Work by Garcia-Pichel et al. and Nubel et al. [25,26] has elucidated the phylogeny of halotolerant cyanobacteria, some of which are likely to have been detected in this study. In addition, it is likely that some of the organisms with cluster 3 nifH sequences are closely related to the SRB detected by Risatti et al. [24] in their study of Guerrero Negro mats. The large diversity of unidentified nifH sequences could mean that the functional groups of bacteria responsible for these sequences are diverse. This study clearly adds to the breadth of emerging knowledge of the function and diversity of organisms that inhabit the mat systems in ESSA.

In summary, we report here that established Lyngbya and Microcoleus mats, as well the greenhouse incubated Microcoleus mats, were actively fixing nitrogen. Activity was largely in the surface 5 mm, a layer characterized by high rates of oxygenic photosynthesis, and low concentrations of inorganic nitrogen. The pattern of nitrogenase activity appeared to be cyclic, with highest rates occurring during the dark and lowest rates occurring during the day, consistent with N2 fixation driven by non-heterocystous cyanobacteria. nifH protein sequences recovered from those mats as well as from developing Lyngbya and Microcoleus microbial mats were similar to sequences recovered from cyanobacteria, γ-Proteobacteria, as well as sequences recovered from termite guts, estuaries, sea grasses and other microbial mats. The cluster 3 sequences were the overwhelming majority of sequences recovered in this study, and their abundance in both the established and developing mats suggests they have a large role in N2 fixation in all stages of mat growth. Finally, while the detection of nitrogenase genes identifies the types of microorganisms that are capable of fixing N2, it cannot define whether these microorganisms are involved in N2 fixation. Reverse transcriptase PCR and quantitative reverse transcriptase PCR are needed to determine which microorganisms are actively transcribing nifH and are thus responsible for nitrogenase activity observed at different times and under different conditions [51].


We would like to acknowledge Dr. Bethany Jenkins for her advice, as well as Dr. Dave Des Marais, Kendra Turk, and many others at the NASA Ames Research Center, for their logistical support. Steve Carpenter made the microelectrode measurements, and Mary Hogan helped perform acetylene reduction analyses, and Lee Prufert-Bebout took the photomicrographs. We would also like to thank Dr. Ferran Garcia-Pichel's lab for providing the culture of Halothece sp. This work was funded through NASA Cooperative Agreement NCC 2-5370 to the University of California, Santa Cruz. We are very grateful for continued access to the Guerrero Negro field site, as well as logistical support, from Exportadora de Sal S.A. de C.V.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
View Abstract