Identification and Distribution of High-Abundance Proteins in the Octopus Spring Microbial Mat Community

Abstract

A shotgun metaproteomics approach was employed to identify proteins in a hot spring microbial mat community. We identified 202 proteins encompassing 19 known functions from 12 known phyla. Importantly, we identified two key enzymes involved in the 3-hydroxypropionate CO₂ fixation pathway in uncultivated Roseiflexus spp., which are known photoheterotrophs.

TEXT

The microbial mat communities of Octopus Spring (OS) and Mushroom Spring (MS), Yellowstone National Park, are well studied (20–22). Some of the predominant taxa have been cultivated and examined to elucidate their functional roles within the mat (2, 6, 15, 19). The two predominant phototrophs are Synechococcus and Roseiflexus spp. (11). Another anoxygenic phototroph, “Candidatus Chloracidobacterium thermophilum,” has also been recently described and may contribute significantly to mat function (6). Previous studies on this community have indicated that filamentous anoxygenic phototrophs (FAP; namely, Chloroflexus and Roseiflexus spp.) perform CO₂ fixation via the 3-hydroxypropionate (3-OHP) pathway (10, 17–19). To date, the linkage between taxa and function has been unclear, since Roseiflexus sp. RS-1, the major FAP in the mat, has not been shown to have photoautotrophic activity, even though it contains the genes for the 3-OHP pathway (10). Other extant taxa have not been brought into cultivation, but metagenomic and metatranscriptomic analyses have revealed their functional potential (11–13). In contrast to metatranscriptomics, which indicates the potential for protein expression and, thus, function in a complex sample, metaproteomics reveals the proteins present in the sample at any given time. Since gene expression does not always correlate with protein content (reviewed in reference 7) and, to date, analysis of specific proteins in the OS microbial mat community has been limited to a targeted study of nitrogen fixation (16), the current study was performed to help understand the functions of individual OS mat community members and to predict how they contribute to the processes occurring in the mat. This report is the first metaproteomics analysis of the OS microbial mat community.

Microbial mat cores were taken from OS at an average temperature of 60°C under midday high-light conditions. Proteins were extracted from mat samples, digested with trypsin, and then analyzed by offline Mud-PIT liquid chromatography-tandem mass spectrometry (LC-MS/MS) (see the supplemental material). The resulting spectra were used by the Mascot search engine (Matrix Science, London, United Kingdom) to query databases comprised of either protein sequences from NCBI associated with thermophiles or the translated metagenome from OS and MS (11). Unique proteins (based on spectra) from each database search were merged into a final data set (see the supplemental material). Functions of identified proteins were determined using the Kyoto Encyclopedia of Genes and Genomes (KEGG; see the supplemental material).

A total of 1,560 spectra, corresponding to 202 proteins (approximately 1% of the known metagenome [12]), were identified (see the supplemental data set). Proteins associated with 12 bacterial phyla were identified; however, 89% of the proteins were associated with three phyla, Chloroflexi, Cyanobacteria, and Chlorobi (Fig. 1A). The percentage of proteins associated with these taxa is in general agreement with metagenomic and metatranscriptomic data (11, 12).

Fig 1.

Protein distribution in the Octopus Spring cyanobacterial mat community. Proteins identified by Mascot database searching were grouped by phylum (A) or function (B) in descending order based on their relative percentage of the total population. (A) “Other” consists of Acidobacteria, Actinobacteria, Aquificae, Bacteroidetes, Cyanidophyta, Deinococcus-Thermus, Firmicutes, Proteobacteria, Thermotogae, and unknown (all less than 2.5% of the total). (B) “Other” consists of cell wall synthesis, DNA repair/recombination, nucleotide metabolism, oxidative phosphorylation, phospholipid synthesis, porphyrin and chlorophyll metabolism, proteolysis, survival, and transcription (all less than 2.5% of the total).

Identified proteins were mapped to function (Fig. 1B). The most populated five functions included protein folding, transport, carbohydrate metabolism, protein synthesis, and stress defense. The protein-folding category consists of six types of chaperones/cochaperones (36 proteins/439 spectra; after this point, only the numbers of proteins/spectra will be listed) distributed through most of the identified phyla. The most abundant protein in this category is the chaperonin GroEL (19/243). Analyses of the GroEL transcripts for many of the abundant mat inhabitants indicate that many of the GroEL homologs have differential transcription patterns throughout a diel cycle, and some exhibit midmorning or midday increases in mRNA levels while others are constitutively expressed (C. G. Klatt, Z. Liu, M. Ludwig, M. Kühl, S. I. Jensen, D. A. Bryant, and D. M. Ward, submitted for publication). Chaperonins, including GroEL variants, are highly expressed under conditions of stress that may lead to a loss of function and viability due to improper folding. These stressors include heat shock, hyperosmolarity, starvation, and oxidative stress (reviewed in reference 14). Since OS mat cyanobacteria exhibit evidence of oxidative stress (9), the high levels of chaperonins observed are not surprising. Abundant GroEL is also observed in the metaproteomes of other “stressed” communities, such as those from a chlorobenzene-contaminated site (4), an activated sludge wastewater treatment plant (23), and an acid mine drainage site (3).

The transport category (16/146) consisted primarily of ABC transporters (8/89) and extracellular binding proteins/receptors (9/57) observed in Chlorobi, Chloroflexi, Cyanobacteria, and Proteobacteria. The carbohydrate metabolism category (16/119) was observed mainly in Chloroflexi (13/108). The protein synthesis category (14/111) was observed in Chlorobi, Chloroflexi, Cyanobacteria, and Proteobacteria and consisted of elongation factors (7/62) and ribosomal proteins (7/49). The stress defense category (15/94) was observed in Acidobacteria, Bacteriodetes, Chloroflexi, and Cyanobacteria and consisted primarily of peroxiredoxin/AhpC/TSA family members (7/52).

The two most abundant phyla were further examined to correlate protein functions with taxa (Fig. 2). The majority of the proteins associated with Chloroflexi mapped to the genus Roseiflexus (85%). These results are in agreement with those of Liu et al. (12), who reported that 84% of Chloroflexi in the phototrophic layer of the mat were Roseiflexus spp., based on rRNA analysis. The proteins assigned to Chloroflexi were grouped into 17 functional categories (Fig. 2B). In general, central carbon metabolism (carbohydrate, amino acid, fatty acid, and nucleotide metabolism, carbon fixation, and protein synthesis) was the dominant category for this phylum (48/570 associated with metabolism from a total of 99/766). Proteins from Roseiflexus spp. that participate in glycolysis and gluconeogenesis were detected. Most notably, the two unidirectional enzymes 6-phosphofructokinase (RoseRS_1831/Rcas_2513) and phosphoenolpyruvate carboxykinase (RoseRS_2496) were observed. The former is considered to be a control point for the regulation of glycolysis (5). The simultaneous presence of these two proteins obfuscates the ability to determine the direction of carbon flux in these organisms; however, this proteomic analysis identifies only the presence of proteins, not their activities. Thus, phosphoenolpyruvate carboxykinase and 6-phosphofructokinase may operate as pathway regulators whose activities are controlled via allosteric regulation, depending upon metabolite concentrations. The metatranscriptomic sequencing of mat samples taken over a 24-h period reveals that genes for both of these key enzymes are most highly transcribed at night (Klatt et al., submitted), indicating that more in-depth metabolic studies must be undertaken to understand which of these pathways is active in Roseiflexus spp. at different times during the diel cycle.

Fig 2.

Protein distribution among the two major phyla in the Octopus Spring cyanobacterial mat community. Proteins associated with Chloroflexi and Cyanobacteria were examined with respect to their association with genera and species (A and C) and protein functions (B and D). For each panel, organisms or functions are grouped in descending order based on their relative percent abundance in the phylum. “Other” consists of DNA repair/recombination, oxidative phosphorylation, porphyrin and chlorophyll metabolism, proteolysis, survival, and transcription (all less than 2.5% of the total).

Interestingly, three proteins that can participate in CO₂ fixation were also identified for Roseiflexus spp.: phosphoenolpyruvate carboxylase (RoseRS_3110), the bifunctional malonyl coenzyme A (malonyl-CoA) dehydrogenase/reductase (RoseRS_3201), and the trifunctional acrylyl-CoA reductase (NADPH)/3-hydroxypropionyl-CoA dehydratase/3-hydroxypropionyl-CoA synthase (RoseRS_3202). The multifunctional enzymes RoseRS_3201 and RoseRS_3202 are directly involved in the 3-OHP CO₂ fixation pathway and perform five sequential steps converting malonyl-CoA to propionyl-CoA (1, 8). The genes encoding these 3-OHP enzymes are most highly transcribed during the day (Klatt et al., submitted), consistent with their presence at 2:00 p.m. in this study. The presence of these enzymes (i) strongly suggests that Roseiflexus spp. are not entirely dependent upon organic substrates from other bacteria for their carbon inputs and (ii) supports previous genetic and metabolic labeling studies suggesting the presence of the 3-OHP pathway in Roseiflexus sp. RS-1 (10, 17–19).

Cyanobacterial proteins were second in abundance (Fig. 2C). The majority of proteins (93%) were associated with the genus Synechococcus, particularly Synechococcus spp. strains JA-3-3Ab and JA-2-3B′a(2-13). Although most proteins were able to be associated with one or the other of these closely related strains, peptides from 8% of the proteins were identical for both strains and therefore were not able to be further categorized. These results are similar to those from Liu et al. (12), who showed by rRNA analysis that Synechococcus spp. comprised 99% of the cyanobacterial population. Functionally (Fig. 2D), the majority of cyanobacterial proteins are associated with protein folding, transport, photosynthesis, protein synthesis, and stress defense. Notably, superoxide dismutase (SOD) from Synechococcus sp. JA-3-3Ab, an enzyme involved in neutralization of reactive oxygen species (ROS), was observed. Its presence in this high-light sample supports previous OS studies on the diel expression of enzymes involved in cyanobacterial ROS defense (9). These studies determined that mRNA transcripts for SOD increased in high light, presumably to counteract the increase in ROS from high irradiance and increased oxygen levels.

Based on these data, only 6% (15/87) of mat community proteins were from Chlorobi (the taxonomic and functional classifications are shown in Fig. S1 in the supplemental material). The spectra in this phylum are predominantly from “Candidatus Thermochlorobacter aerophilum” (74% of the spectra), a novel community member that is an aerobic photoheterotroph (13). While chaperones were the most prevalent (3/16 for “Candidatus Thermochlorobacter aerophilum”), two proteins associated with phototrophy were identified. The non-“Candidatus Thermochlorobacter aerophilum” spectra exhibit homology to various Chlorobium species, and as members of this phylum contribute less to the gene and protein pools in the mat (11), it is currently difficult to assess functional contributions by metaproteomics, as the most abundant proteins observed are probably overrepresented (chaperones; 5/37).

In summary, we have shown that the metaproteome of the OS mat community is accessible and corresponds very well with data obtained from metagenomic and metatranscriptomic analyses. The metaproteomics data reported here provide insight into Roseiflexus species functions at midday, evidenced by the large percentage of peptides involved in central carbon metabolism. More importantly, the metaproteomics data revealed two of the three CO₂ fixation proteins which are multifunctional enzymes in Roseiflexus spp. that perform five sequential reactions in the 3-OHP CO₂ fixation pathway. The presence of these enzymes is supported by the metatranscriptome analysis (Klatt et al., submitted) and strongly supports the hypothesis that Roseiflexus spp. can perform CO₂ fixation even though it is primarily thought to grow photoheterotrophically. The large number of GroEL peptides at 2:00 p.m. is supported by metatranscriptomic data showing increases from morning to noon in GroEL gene expression by many of the mat community members (Klatt et al., submitted). This first investigation of the proteins produced in the OS microbial mat community, their functions, and their correspondence to individual taxa provides a framework for future investigation of the diel cycling of these proteins and will enable direct experimental linkage of function to taxa.