Structural and Functional Changes with Depth in Microbial Communities in a Tropical Malaysian Peat Swamp Forest
- First Online:
- Cite this article as:
- Jackson, C.R., Liew, K.C. & Yule, C.M. Microb Ecol (2009) 57: 402. doi:10.1007/s00248-008-9409-4
- 575 Views
Tropical peat swamp forests are important and endangered ecosystems, although little is known of their microbial diversity and ecology. We used molecular and enzymatic techniques to examine patterns in prokaryotic community structure and overall microbial activity at 0-, 10-, 20-, and 50-cm depths in sediments in a peat swamp forest in Malaysia. Denaturing gradient gel electrophoresis profiles of amplified 16S ribosomal ribonucleic acid (rRNA) gene fragments showed that different depths harbored different bacterial assemblages and that Archaea appeared to be limited to the deeper samples. Cloning and sequencing of longer 16S rRNA gene fragments suggested reduced microbial diversity in the deeper samples compared to the surface. Bacterial clone libraries were largely dominated by ribotypes affiliated with the Acidobacteria, which accounted for at least 27–54% of the sequences obtained. All of the sequenced representatives from the archaeal clone libraries were Crenarchaeota. Activities of microbial extracellular enzymes involved in carbon, nitrogen, and phosphorus cycling declined appreciably with depth, the only exception being peroxidase. These results show that tropical peat swamp forests are unusual systems with microbial assemblages dominated by members of the Acidobacteria and Crenarchaeota. Microbial communities show clear changes with depth, and most microbial activity is likely confined to populations in the upper few centimeters, the site of new leaf litter fall, rather than the deeper, older, peat layers.
Tropical peat swamp forests are globally important ecosystems that are poorly understood in regards to their ecology and biodiversity . These environments are restricted to Southeast Asia (Malaysia, Indonesia, Papua New Guinea, and southern Thailand) and are being lost to logging, fire, and drainage and conversion to agricultural land so that few pristine forests remain [2, 21]. The importance of these tropical peat swamp forests in the global carbon cycle is becoming increasingly apparent [29, 35] as they are one of the largest terrestrial carbon stores . Carbon can be sequestered in layers of peat up to 20 m deep and released following disturbance . For example, surges in global atmospheric CO2 levels have been linked to burning of peat in Borneo and Sumatra [24, 34].
It is commonly assumed that peat accumulates because the environmental conditions (acidic waterlogged soils, low nutrients, high tannins) inhibit microbial activity [14, 37, 45, 46]. In contrast to the dipterocarp rain forests that dominate Southeast Asia, peat swamp forests are thought to show low rates of nutrient cycling and organic matter decomposition. Autochthonous organic matter production by algae is also suppressed by a thick forest canopy and tannin-darkened waters. However, these systems harbor a high diversity and abundance of fish and aquatic invertebrate species [11, 44], which do not appear to be dependent upon particulate detritus for energy. Rather, stable carbon isotope and dietary analyses suggest that bacteria assimilate dissolved organic carbon (DOC) leached from leaf litter and peat, and the bacteria are ingested by invertebrates, which are in turn ingested by fish (Yule, unpublished data). Thus, bacteria may be the vital link between photosynthetic inputs of carbon and the higher order consumers and the key to carbon and nutrient cycling within tropical peat swamp forests. Indeed, litter breakdown studies have shown that while sclerophyllous leaves of native peat forest plants decompose slowly and accumulate as peat, nonendemic leaves are decomposed rapidly in these systems suggesting the presence of active microbial communities .
Little is known of the composition of microbial assemblages in tropical peat swamp forests, although a few studies have begun to describe the microbial populations inhabiting temperate or boreal peatlands [10, 39]. Novel representatives of the Actinobacteria have been isolated from peat swamp forests in Thailand , and an aeroaquatic fungus isolated from submerged organic material in a Malaysian peat swamp forest has been recently described as a new genus and species . However, there has been no systematic survey of the in situ microbial populations in these ecosystems. In this study, we sought to determine general patterns in bacterial and archaeal community composition in sediments in a Malaysian peat swamp forest using 16S ribosomal ribonucleic acid (rRNA) techniques. In particular, we focused on examining how these communities change structurally and functionally with depth into the peat sediment.
Materials and Methods
North Selangor Peat Swamp Forest (N 3°39′55.1″, E 101°19′51.2″, elevation 16 m above sea level) is a remnant of a larger peat swamp forest that has been reduced by drainage, conversion to agriculture, and logging. The climate is tropical with mean annual rainfall of more than 200 cm per year and average temperature of 28°C. The peat substrate is several meters deep, lying on a bed of marine alluvial clay, and is perpetually waterlogged with the forest floor becoming submerged during rainy periods. The water is acidic (pH 3–4), has a characteristic “blackwater” color due to high concentrations of tannins and humic acids (DOC is typically 80 mg L−1), and is low in dissolved oxygen because of lack of water flow. Inorganic nutrient concentrations are low (N and P < 0.1 mg L−1). The study site was selectively logged until 1985, but secondary growth forest has recovered since that time. The vegetation is dominated by Macaranga pruinosa, a common species in disturbed peat swamp forests. Other abundant vegetation includes ferns (Stenoclaena palustris, Nephrolepis biserrata, Asplenium longissimum, Dicranopteris sp.), palms (Pinanga sp., Ptychoraphis sp., Korthalsia sp.), Pandanus (Pandanus helicopus), trees (Macaranga hypoleuca, Campnosperma coriaceum, Ixora grandiflora, Cryptocarya impressa, Parartocarpus venosus, Aglaia odorata), and climbers (Pternandra galeata and Shorea platycarpa). Much of the vegetation produces tough, sclerophyllous leaves with toxic secondary compounds that are resistant to microbial decomposition. Many of the trees have pneumatophores to obtain oxygen, and the plants typically have a thick layer of roots in the top 50 cm of the peat to absorb the limited nutrients.
Sample Collection and Processing
Upper peat sediments were sampled from the forest on 16 January 2007 (during a rainy period). Peat samples were collected in sterile 50-mL tubes from the peat surface (0 cm), and at 10-, 20-, and 50-cm depths. The pH of the water at each depth ranged from 3.8 to 4.1, with the slightly lower pH values being found in the deeper samples. The 0- and 10-cm samples were collected by scooping sediment directly into the tubes, while sterile syringes and tubing were used to obtain the 20- and 50-cm samples. Five replicate samples of peat at each depth were collected within a 1-m2 area, to give a total of 20 samples (four depths × five replicates at each depth). Collection tubes were filled to minimize the amount of headspace, sealed, and shipped to the USA. Shipping time was under 72 h, and samples were processed immediately upon arrival. A subsample (0.5 g) from each sample was frozen for subsequent molecular analyses, and 5 g of the remaining material was used in assays of microbial extracellular enzyme activity.
Assays of Microbial Extracellular Enzyme Activity
Each sample was assayed for the activity of seven microbial extracellular enzymes using colorimetric procedures previously applied to wetland sediments [17, 18]. Specific enzymes assayed are involved in the aerobic degradation of cellulose (β-glucosidase, cellobiohydrolase [CBH]), hemicellulose (β-xylosidase), chitin (N-acetylglucosaminidase [NAGase]), lignin and polyphenolic compounds (phenol oxidase, peroxidase), or phosphorus cycling (acid phosphatase). Sediment material was homogenized in 50 mM acetate buffer (as per Jackson et al.  but adjusted to pH 4.0 to reflect environmental conditions) to yield 10-mL slurries. Four 150-μl replicates of each sample slurry were mixed with 150 μl of substrate solution for each enzyme assay and incubated for 2–4 h. Substrate solutions for β-glucosidase, CBH, β-xylosidase, NAGase, and acid phosphatase consisted of p-nitrophenyl-linked substrates, and the activity for these enzymes was determined from absorbance at 410 nm in the presence of 0.067 M NaOH, as described previously . l-3,4-Dihydroxyphenylalanine served as the substrate for the phenol oxidase and peroxidase assays, and peroxidase assays also received hydrogen peroxide to a final concentration of 0.015%. The activity for these enzymes was determined from a final absorbance at 460 nm . Duplicate analytical controls consisted of 150 μl of sample slurry with 150 μl of acetate buffer. An additional 10 g of each sediment sample was weighed, dried (65°C, 48 h), and ashed (500°C, 2 h) so that enzyme activity could be expressed per gram organic matter (i.e., μmole substrate consumed h−1 g organic matter−1).
DNA Amplification, DGGE, and Sequencing Analysis
Deoxyribonucleic acid (DNA) was extracted from each sample (0.5 g) using a PowerSoil DNA Kit (MoBio, Carlsbad, CA, USA) with an additional initial incubation step at 70°C for 10 min. Extracted DNA was used as the template in four polymerase chain reaction (PCR) amplifications of portions of the 16S rRNA gene. The first amplification used the bacteria-specific primers Bac1070f and Univ1392rGC [12, 19] to amplify a 323-bp section of the bacterial 16S rRNA. The second amplification used the Archaea-specific primer Arc931f  in conjunction with the Univ1392GC primer to amplify a 461-bp section of the archaeal 16S rRNA gene. Each reaction contained approximately 25 ng of template DNA (as determined from visualization in agarose gels). Amplification products from both of these sets of reactions were analyzed by denaturing gradient gel electrophoresis (DGGE). The third amplification used primers Bac8f and Univ1492r [19, 20] to amplify a larger (1,480+ bp) portion of the 16S rRNA gene for cloning–sequencing, while the fourth amplification used the archaeal-specific primer pair Arc2f and Univ1492r for the same purpose. Amplification conditions for all sets of primers were as described previously , and reactions also used 25 ng of template DNA.
DGGE of both bacterial and archaeal PCR products was performed over a denaturing gradient of 40–70% (where 100% denaturant is 7 M urea and 40% deionized formamide). Electrophoresis conditions were 75 V for 18 h at 60°C, and approximately 700 ng PCR product was loaded for each sample. Because of the number of samples of each type (20 total), samples were analyzed on three separate DGGE gels, each with standards (PCR products amplified from Pseudomonas aeruginosa and Staphylococcus aureus genomic DNA) to allow comparisons between gels. Following electrophoresis, gels were stained with SYBR Green I and visualized by UV transillumination. Composite images comparing samples from different gels were generated using a Kodak Gel Logic 200 running Molecular Imaging Software 4.0 (Eastman Kodak, Rochester, NY, USA). The bands present in each lane were detected digitally, and the data were converted into binary form (presence–absence of a band at a particular position). Cluster analysis was performed using Euclidean distances and an average linking method . All multivariate analyses were performed using Gingko, a subset of the VegAna software package (Department of Vegetal Biology, University of Barcelona).
Amplification products from selected Bac8–Univ1492 and Arc2–Univ1492 reactions were cloned in to artificial plasmid vectors (TA TOPO Cloning, Invitrogen, Carlsbad, CA, USA), and clone libraries were generated for bacterial and archaeal 16S rRNA genes at different depths. The insert from the first 96 clones in each library was amplified using standard M13 primers (Invitrogen), and clones within each library grouped into ribotypes based upon restriction patterns after digests with the restriction enzymes EcoRI, RsaI, and HaeIII [19, 20]. The frequency of each ribotype was used to estimate overall diversity in each library as SChao1  using a web-based interface .
The first 600–700 bp of the more abundant ribotypes in each library (those represented by at least two clones or 80 ribotypes in total) was sequenced using the M13 reverse primer. Sequences were compared to those in GenBank (BLAST search during September 2007) to identify closest relatives. Both the sequence and its closest GenBank match were imported into ARB  running an established 16S rRNA database . Sequences were aligned and incorporated into existing phylogenetic trees using procedures described previously  to reduce distortions that can arise from analysis of short, divergent gene fragments.
Nucleotide Sequence Accession Numbers
The partial 16S rRNA gene sequences from this study have been deposited in GenBank and have accession numbers EU402971–403051.
Patterns in Extracellular Enzyme Activity
DGGE Profiles with Depth
Clone Library and Phylogenetic Analysis
Properties of clone libraries generated from 16S rDNA amplified from different depths (0, 10, 20, 50 cm) of tropical peat swamp forest sediment using bacterial (Bac)- or archaeal (Arc)-specific primers
Sequences not affiliated with the Acidobacteria that were numerous (greater than one clone out of 96 sampled) in clone libraries of PCR-amplified bacterial 16S rRNA genes generated from different depths (0, 10, 20, 50 cm) of tropical peat swamp forest sediment
Sequence (number of clones)
Dyella koreensis (Gamma-Proteobacteria)
EF019370 (Alpha-Proteobacteria, Hyphomicrobiaceae)
Bradyrhizobium canariense (Alpha-Proteobacteria)
AY080915 (Beta-Proteobacteria, Oxalobacteraceae)
Syntrophobacter wolinii (Delta-Proteobacteria)
AB232813 (Alpha-Proteobacteria, Rhodospirillales)
AY913382 (Alpha-Proteobacteria, Bradyrhizobiaceae)
High acidity coupled with low concentrations of nutrients and frequent anoxic conditions make temperate and tropical peat wetlands extreme habitats where microbial activity has generally been thought to be inhibited [10, 45]. Despite these extreme conditions, tropical peat swamp forests have a high biodiversity of flora and fauna, with more than 900 species of flowering plants described from Borneo peat swamp forests  and at least 300 species of fish in peat swamp forests throughout Southeast Asia . Novel species of fungi and actinomycetes have been isolated from these systems [40, 42], but there have been no prior attempts to describe prokaryotic community structure. Based on 16S rRNA sequence data, archaeal assemblages in the peat swamp forest sediments sampled were dominated by previously uncultured members of the Crenarchaeota, while bacterial communities were largely composed of members of the Acidobacteria.
Archaeal 16S rRNA could only be amplified from samples taken from 20 or 50 cm below the surface, suggesting the absence of Archaea in surface sediments. The numbers of archaeal cells have been shown to increase with depth in other peat sediments [10, 26], and prior to sequence analysis, we suspected that this would be because anoxic conditions in deeper sediments facilitated the growth of methanogenic Euryarchaeota. Methanogenesis is recognized as an important biogeochemical process in peatlands worldwide, and molecular studies of methanogen communities have been carried out in a number of peat wetlands [5, 26]. It is surprising to note that all of the sequenced ribotypes from the Archaea clone libraries corresponded to members of the Crenarchaeota. Crenarchaeotal sequences typically account for 0.5% to 3% of the 16S rRNA ribotypes obtained from soil , although these organisms have generally not been studied in peat wetlands. That none of the archaeal ribotypes sequenced corresponded to methanogens could be because they were not numerically dominant (we only sequenced ribotypes corresponding to at least two clones) or that our PCR primers did not detect them. However, the same primers and reaction conditions have been used to characterize both Crenarchaeota and Euryarchaeota in other environments . It is also possible that the 3-day lag (because of shipping time) between sampling and analysis may have resulted in the loss of certain microbial populations, particularly anaerobic organisms such as methanogens. However, the amount of air in sample tubes was kept to a minimum, and stable carbon isotope studies of peat samples from the same site also indicate an absence of methanogenic and methanotrophic organisms (Yule, unpublished data).
Other than our inability to amplify archaeal DNA from the 0- and 10-cm samples, there were no clear patterns in the distribution of Archaea with depth. Ribotypes from both 20- and 50-cm-deep samples were affiliated with the same groups of Archaea, and certain ribotypes were similar enough to suggest the same populations at each depth. Estimates of diversity were similar for the two clone libraries, again suggesting that the 20- and 50-cm samples harbored similar archaeal communities. The Crenarchaeota sequences detected fell into two broad groups: Three sequences were affiliated with a cluster of Crenarchaeota previously identified in rice paddies [9, 28], whereas the majority of sequences fell within a larger lineage that includes the Finnish forest soil type B (FFSB) group [6, 33]. However, a number of ribotypes were only loosely affiliated with the FFSB group and were more similar to a sequence recently described from acid mine drainage . The Arc50-13 ribotype and a similar sequence (Arc50-08) were related to this sequence, and together these accounted for 17.5% of the 50 cm clone library, suggesting that this population is numerically important in deeper sediments.
While archaeal assemblages were somewhat similar in the 20-cm and 50-cm samples, bacterial communities showed broad depth related patterns, including a steady decrease in diversity in clone libraries as depth increased. Bacterial clone libraries were dominated by ribotypes affiliated with the Acidobacteria, and representatives of this lineage have been found to be common members of 16S rRNA clone libraries generated from other soils and sediments, typically accounting for 10–50% of the ribotypes observed [3, 22]. The Acidobacteria accounted for 27–54% of ribotypes in the four bacterial clone libraries generated in this study, and these values are likely underestimates as we only sequenced ribotypes present in at least two clones (i.e., there may be additional Acidobacteria sequences represented by single clones in the 96 clones screened for each library). While this suggests that the Acidobacteria are a major component of the prokaryotic community found in tropical peat swamp forests, others have found inconsistencies between the abundance of Acidobacteria-like sequences in clone libraries and the apparent abundance of Acidobacteria as determined microscopically . Thus, it is possible that this lineage is proportionally less abundant than its representation in our clone libraries might imply.
The particular 16S rRNA gene sequences recovered fell within three of the original eight subgroups of Acidobacteria outlined by Hugenholtz et al.  and recently expanded . The greatest diversity of ribotypes fell within the Acidobacteria-1 subgroup, which includes a cluster of sequences closely related to A. capsulatum, although representatives of this cluster (designated 1C) were not detected in the 50-cm samples. In contrast, the proportion of ribotypes affiliated with subgroup 1A was appreciably higher in the 50-cm clone library compared to the shallower depths. Inferring patterns based upon comparisons to existing environmental sequences in databases is not always reliable, as database entries may contain insufficient information about the environment sampled. However, the reference sequences in Acidobacteria-1A tend to have been obtained from organic, deeper soils (subsurface, peat, forest, iron–manganese nodules) that would be more similar to the deeper sediments sampled in this study. Many of the Acidobacteria ribotypes showed the greatest similarity to 16S rRNA gene sequences detected in evergreen forest soils in China. These forest soils are likely similar to the peat swamp forest examined in this study (organic rich, low nutrients, pH around 4), and sequences affiliated with the Acidobacteria were also the most abundant in those 16S rRNA clone libraries . That study suggested that members of Acidobacteria subdivisions 1–3 may favor lower nutrient environments, a suggestion supported by ribotypes within these groups being the dominant 16S rRNA gene sequences obtained in the nutrient-poor peat swamp forest sediments that we examined.
Functional changes (as detected by extracellular enzyme activity) with depth were more pronounced than structural changes. Enzyme activities in surface sediments (the site of leaf litter fall) were similar to those associated with decaying organic matter in a Louisiana cypress swamp , which were slightly lower than those previously reported for other wetland systems. Activities of microbial extracellular enzymes have been shown to decrease with depth in temperate peatlands [13, 36] and other wetland systems  and showed the same pattern in tropical peat swamp forest sediments. Activities of β-glucosidase, CBH, β-xylosidase, and NAGase have been directly linked to rates of organic matter decomposition in various wetlands [1, 17, 18] so elevated activities in surface layers likely correspond to more rapid carbon mineralization. Similarly, activities of NAGase and phosphatase influence rates of N and P mineralization, respectively [30, 32], suggesting that nutrient cycling is also more rapid in the top 10 cm or so of tropical peat swamp forest sediments. Declines in enzyme activity with depth can represent a decrease in substrate quality with depth , suggesting that the surface layers of tropical peat swamp forests contain more readily utilizable substrates for microbial activity. Increased depth corresponds to increased age and hence greater prior degradation of the peat. Surface and upper sediments are also likely to be more aerobic, which would also facilitate greater microbial activity. We were unable to detect phenol oxidase activity at any depth, which may be a function of low pH and low oxygen levels, both of which create suboptimal conditions for phenol oxidase activity in peatlands . Rather, peroxidase, which showed activity at all depths, may be more important for the degradation of phenolic materials in this system.
Tropical peat swamp forests are endangered ecosystems that are important globally. As well as supporting diverse animal and plant communities, these systems also harbor unusual bacterial assemblages that based on 16S rRNA gene sequencing appear to be dominated by Acidobacteria, along with diverse archaeal communities in deeper sediments. These microorganisms potentially serve as a link between dissolved organic material released from decaying peat and higher consumers. As such, these bacteria may be interesting to culture and novel cultivation approaches  could be adapted to this system. From an applied aspect, these microorganisms may be capable of processing recalcitrant organic material under relatively harsh (low pH, potentially anoxic) conditions. Extracellular enzyme activities suggest much cellulolytic activity may be confined to the surface, but oxidative processing of phenolic material (such as by peroxidase) could occur deeper in the sediments. Future studies in tropical peat swamp forests could emphasize an enrichment approach or examine functional genes to more closely tie patterns in prokaryotic community structure to the functioning of these unique ecosystems.
M.D. Barrows and S.E. Young assisted with enzyme assays and molecular work. Funding for this work was provided to C.R.J. by through an O.R.S.P. faculty research award from The University of Mississippi and to C.M.Y. through a Monash University research grant.