Metabolic Sequences of Anaerobic Fermentation on Glucose-Based

Oct 31, 2012 - Visualizing microbial dechlorination processes in underground ecosystem by statistical correlation and network analysis approach. Akira...
0 downloads 6 Views 2MB Size
Download PDF
Article pubs.acs.org/jpr

Metabolic Sequences of Anaerobic Fermentation on Glucose-Based Feeding Substrates Based on Correlation Analyses of Microbial and Metabolite Profiling Yasuhiro Date,†,‡ Tomohiro Iikura,‡ Akira Yamazawa,‡,§ Shigeharu Moriya,‡,∥ and Jun Kikuchi*,†,‡,⊥ †

RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan Graduate School of Nanobioscience, Yokohama City University, 1-7-29 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan § Research Planning and Management Group, Kajima Technical Research Institute, KAJIMA Corporation, 2-19-1 Tobitakyu, Chofu, Tokyo 182-0036, Japan ∥ RIKEN Advanced Science Institute, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan ⊥ Graduate School of Bioagricultural Sciences, Nagoya University, 1 Furo-cho, Chikusa-ku, Nagoya, Aichi 464-0810, Japan ‡

S Supporting Information *

ABSTRACT: Degradation processes in various biomasses are managed by complex metabolic dynamics created by diverse and extensive interactions and competition in microbial communities and their environments. It is important to develop visualization methods to provide a bird’s-eye view when characterizing the entire sequential metabolic process in an environmental ecosystem. Here, we describe an approach for the visualization of the metabolic sequences in anaerobic fermentation ecosystems, characterizing the entire metabolic dynamics using a combination of microbial community profiles and metabolic profiles. By evaluating their time-dependent variation, we found that microbial community profiles and metabolite production processes were characteristically affected by the feeding of different glucose-based substrates (glucose, starch, cellulose), although the compositions of the major microbial community and the metabolites detected were likely to be similar in all experiments. This combinatorial approach to variation in microbial communities and metabolic profiles was used successfully to visualize metabolic sequences in anaerobic fermentation ecosystems, in addition to mining candidate microbiota for cellulose degradation. Thus, this approach provides a powerful tool for visualizing and evaluating metabolic sequences within the biomass degradation process in an environmental ecosystem. This is the first report to visualize the entire metabolic dynamic in an anaerobic fermentation ecosystem as metabolic sequences. KEYWORDS: metabolic sequences, NMR, multivariate statistical analysis, cellulose degradation, anaerobic fermentation

■

INTRODUCTION Anaerobic fermentation is considered the most ancient metabolic mechanism undertaken by complex microbial ecosystems.1 Primordial microbial communities were relegated to anaerobic environments when the emergence of photosynthetic organisms raised the oxygen concentrations on Earth. However, within contemporary anaerobic environments, such as soils, oceans, and symbiotic ecosystems, anaerobic consortia can degrade biomass made by photosynthetic organisms. Thus, anaerobic fermentation remains indispensable to the biogeochemical cycles of organic matter. Anaerobic fermentation, a kind of anaerobic digestion, enables microorganisms to degrade the carbon chains of highly polymerized biomass into C1 compounds. This metabolic process is one of the most significant global events, undertaken extensively in wet anaerobic environments such as wetlands, swamps, bogs, fens, tundra, rice fields, landfills, and the intestines of animals, including mammals.2 Additionally, microbial anaerobic © 2012 American Chemical Society

digestion of organic waste can be exploited for clean energy production from a renewable feedstock.3,4 The biological conversion of biomass to biogases, such as hydrogen and methane, has received increasing attention in recent years. Thus, anaerobic fermentation is an important process from both natural and industrial perspectives. In the anaerobic fermentation process, low-molecular-weight chemical compounds are produced by heterotrophic bacteria in microbial ecosystems, accompanied by the degradation of larger substrates, such as carbohydrates. The degradation processes of various carbohydrates by heterotrophic bacteria are complex and diverse. In particular, the degradation processes of polymeric macromolecules, such as cellulose, which are typically persistent substances, are managed by complex metabolic dynamics created by diverse and extensive interactions and competition in Received: August 8, 2011 Published: October 31, 2012 5602

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

Figure 1. Schematic overview of the visualization method for metabolic sequences in environmental ecosystems. (A) Schematic view of the degradation processes of glucose, starch, and cellulose in anaerobic fermentation ecosystems. (B) The experimental flow for evaluating metabolic sequences. Time-dependent variations in microbial community and metabolic profiles were evaluated using denaturing gradient gel electrophoresis (DGGE) fingerprinting and NMR spectroscopy. Data sets of time series variations were combined using multivariate correlation analysis. Correlated data sets were visualized and compared as metabolic sequences in each system.

but further modification and optimization is needed to visualize and characterize metabolic sequences in anaerobic fermentation ecosystems. In this study, we developed a means of visualizing metabolic sequences in an anaerobic fermentation ecosystem using community profiling and multivariate statistical analyses. We aspired to characterize completely the metabolic dynamics of the microbial community. To this end, we investigated the degradation processes of glucose, starch, and cellulose in an anaerobic fermentation ecosystem using a combinatorial approach of microbial community profiles and metabolic profiles (Figure 1).

the microbial community and its environment. Because of the complexity of sequential metabolism in anaerobic fermentation ecosystems, conventional bottom-up approaches to identify the degradation processes in an entire ecosystem may be inadequate to visualize the comprehensive metabolic dynamics and networks within a microbial community. Thus, it is important to develop visualization methods that provide a bird’s-eye view to characterize the entire sequential metabolic process in anaerobic fermentation ecosystems, as well as to evaluate environments in mammalian intestines5−10 and plants,11,12 by applying a metabonomics approach with multivariate statistical analyses.13 Within the field of metabonomics, measurement of system responses over time is usually important to obtain a complete picture of metabolic dynamics after or during the application of some input to an environment, such as a drug or food in human intestines, because the variations in each metabolite have intrinsic behaviors.1,14 Thus, time-dependent variations and timed responses should be considered and applied in the development of a visualization method when evaluating and characterizing metabolic dynamics in microbial ecosystems. Nuclear magnetic resonance (NMR)-based metabolomics or metabonomics have been used extensively to study metabolites in a wide range of biological systems in various environments.15−24 We have previously developed an approach that applies metabonomics to monitor metabolic dynamics in microbial ecosystems and to link relationships between microbial communities and their metabolic information.25−27 These approaches are powerful tools for the evaluation of metabolic dynamics in microbial ecosystems,

■

MATERIALS AND METHODS

Seed Material

A high-temperature methane fermentation sludge used for residual food processing was provided by Kajima Corporation, Japan. The microbial community responsible for methane production in the sludge has been characterized in previous reports.28−30 Since the performance and operating conditions of the methane fermentation reactor are also detailed in the previous reports, short-chain fatty acids as major carbon sources, such as acetic and butyric acids, are known to be successfully converted to methane by the methanogenic consortia. Thus, the conversions of short-chain fatty acids to methane and the key players in these processes (i.e., methanogenic archaea) are already well-characterized. Before the experiment started, the sludge was incubated with food waste as a substrate to activate biological 5603

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

Figure 2. Phylogenetic tree constructed using the neighbor-joining method, based on partial 16S rRNA gene sequences cloned from the detected DGGE bands. The sequences determined in this study and those retrieved from the databases were aligned using CLUSTAL W. A phylogenetic tree was then constructed with CLUSTAL W and the Tree View software using the neighbor-joining method. The 16S rRNA gene fragment was amplified using the Univ954f and Univ1369r primer sets. The clones obtained were expressed as bands 1−9. The 16S rRNA gene sequence of Escherichia coli str. K-12 [U00096] was used as the outgroup to root the tree. The bootstrap numbers indicate the value of 1000 replicate trees supporting the order; values below 700 were omitted.

processes in a stirred-tank reactor (STR) at a constant temperature of 55 °C.

The precipitate was collected by centrifugation, dissolved in pure water, and then stored at −20 °C.

Incubation Conditions

PCR-DGGE

Sludge was inoculated in each STR with 2.3 g of D-glucose (Wako Pure Chemical Industries, Ltd.: STRglu), 2.1 g of cornstarch (Wako Pure Chemical Industries, Ltd.: STRsta), or 2.1 g of α-cellulose (Nacalai Tesque Inc.; STRcel) as a substrate at a constant temperature of 55 °C under anaerobic conditions (Supplementary Figure 1). Substrate concentrations were chosen to satisfy a theoretical chemical oxygen demand of 5 kg/m3. Sludge samples were collected at 0, 1, 2, 3, 5, 11, and 19 days after incubation began.

For PCR-DGGE analyses, each DNA sample was amplified by PCR with universal bacterial primers 954f-gc (CGC CCG CCG CGC CCC GCG CCC GGC CCG CCG CCC CCG CCC CGC ACA AGC GGT GGA GCA TGT GG) and 1369r (GCC CGG GAA CGT ATT CAC CG) targeted to the V6 to V8 regions of the 16S rRNA gene.31 Reaction mixtures and PCR conditions were set as described previously.25 After confirmation of the PCR product by agarose gel electrophoresis, DGGE was performed using the Dcode universal mutation detection system (Bio-Rad Laboratories, Inc., Japan). Polyacrylamide gel conditions for the denaturing gradient, migration, and differentiation were set as described in the protocol reported by Yu et al.31 Electrophoresis was conducted at a constant voltage of 82 V at 60 °C for 15 h. Gels were stained with SYBR Green I (Lonza, Rockland, ME), and acquired using a Gel Doc XR (Bio-Rad Laboratories, Inc., Japan).

DNA Extraction

Pellets from collected samples (1 mL) were suspended in 500 μL of water and 500 μL of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0)-saturated phenol was added. The suspension was homogenized and disrupted with 10 mg of 0.1-mm zirconia/silica beads (BioSpec Products, Inc.) using a beadbeating instrument (TOMY MS-100R, Japan; 3000 rpm, 5 min, 4 °C). After centrifugation (15 000 rpm, 10 min, 4 °C), DNA was extracted using 500 μL of phenol/chloroform/isoamyl alcohol (25:24:1) and the aqueous phase with no debris was collected. Then, the phenol/chloroform/isoamyl alcohol was removed using a chloroform extraction. DNA was precipitated by adding 1/10 vol of 3 M sodium acetate and 2 vol of ethanol.

Phylogenetic Analysis

To identify the bacterial origin of DNA sequences in the gel, selected DGGE bands were excised from the original gels and their DNA fragments were reamplified with corresponding primers, as reported previously.25 5604

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

Figure 3. Principle components analysis (PCA) score plots and loading plots of microbial community (A and B) and metabolic (C and D) profiles in all STRs. Each number in the score plots (A and C) indicates the day after operation: STRglu (circles), STRsta (triangles), and STRcel (squares). The region from 4.6 to 4.8 ppm was omitted to eliminate the effects of imperfect water suppression. Blue bars, PC1; red bars, PC2 in loading plots (B and D). Abbreviations: B1 to B9, bands 1 to 9 detected in the DGGE analysis; A, acetic acid; B, butyric acid; F, formic acid; P, propionic acid.

determination (R2) above 0.999 were selected for each metabolite. Using the standard curves, the metabolite concentration estimates were calculated by averaging each signal (excluding any overlapping signals).

NMR Spectroscopy

Supernatants of collected samples were suspended with 10% (v/v) deuterium oxide (D2O) and 1 mM sodium 2,2-dimethyl2-silapentane-5-sulfonate (DSS) as an internal standard. All NMR spectra were recorded by a Bruker DRX-500 spectrometer equipped with a 1H inverse triple-resonance probe with triple-axis gradients, operating at 500.13 MHz for protons. Temperature of NMR samples was maintained at 25 °C. For 1 H NMR spectra, 32 768 data points with a spectral width of 10 000 Hz were collected into 128 transients and 1 dummy scans, and residual water signals were suppressed by Watergate pulse sequence with a 1.2-s cycle time. Prior to Fourier transformation, the free induction decays were multiplied by an exponential window function corresponding to a 2.0 Hz line broadening factor. The acquired spectra were manually phased and baseline-corrected. The methods for the NMR measurements of two-dimensional (2D) 1H−13C heteronuclear single quantum coherence (HSQC) and total correlation spectroscopy (TOCSY) were applied as described previously.25,32−38 NMR spectra were processed using NMRPipe software39,40 and were assigned using the SpinAssign program at the PRIMe Web site.41−43 The metabolites assigned as glucose, acetic acid, lactic acid, formic acid, butyric acid, propionic acid, and ethanol were quantified by the calibration curve method. Chemical standards (of five known concentrations) of each metabolite were precisely quantified under the same NMR measurement conditions. In constructing the standard curve, signals satisfying coefficient of

Principle Components Analysis (PCA) of DGGE and 1H NMR Spectra

DGGE images were read using the Quantity One software (Bio-Rad Laboratories, Inc.). Signal intensities and band positions in each lane were divided into a spectrum consisting of 100 variables. 1H NMR data were reduced by subdividing spectra into sequential 0.04 ppm-designated regions between 1H chemical shifts of 0.0 to 9.0 ppm. After exclusion of water resonance, each region was integrated and normalized to the total of DSS integral regions. The DGGE and 1H NMR data (normalized to a constant sum, or total intensities of the signals are constant between each datum) were statistically evaluated by PCA using the ‘R’ software. The PCs are calculated such that each component is a linear combination of the original variables, each PC is orthogonal to all other PCs, and each successive component describes progressively less variance in the data. Data were visualized as PC score plots and loading plots. The PC score plots provide visual realization of the variation and clustering within the data set, from which the sample distribution can be deduced. Each data point on the scores plot represents an individual sample. DGGE bands derived from bacteria and 1H NMR spectral data points related to metabolites are displayed on separate loading plots. Thus, the loading plots 5605

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

provide information about band positions or spectral regions corresponding to sample or sample cluster positions in the score plots.

Table 1. The List of Annotated Metabolites Found in Each STR

Correlation Analysis of DGGE and NMR Data

abbreviation

1

A 2D correlation map was calculated as a symmetric matrix using Pearson’s product-moment correlation coefficient in which an element at position (i, j) is defined as a correlation coefficient between the ith and jth positions in a set of 1H NMR spectra and DGGE gel bands of the identified major bacteria. A more positive (negative) coefficient indicates the existence of a positive (negative) correlation between the ith and jth peaks or bands.

■

metabolites

A Adi

acetic acid adipate

Ala

L-alanine

B

butyric acid

C

6-carboxyhexanoate

E

ethanol

F G

formic acid glucose

L

lactic acid

P

propionic acid

Ph

phenylacetic acid

RESULTS AND DISCUSSION

Microbial Community Profiles in Each STR

The volume of produced biogas was measured as indicator for evaluating substrate degradation activity by the microbial community. The biogas production profiles over a period of about 20 days were obtained by addition of each substrate into the STRs (Supplementary Figure 2). The biogas production rate peaked on the second day following substrate addition, and biogas productions ceased on the 11th day in all STRs. The profiling data of the biogas production tended to be similar in each STR. These data suggest that, in all STRs, maximum substrate degradation activity by the microbial community was achieved on the second day after substrate addition, and was lost by the 11th day. To identify and compare differences in microbial communities in the STRs, bacterial DNA was taken from samples at 0, 1, 2, 3, 5, and 11 days from all STRs and analyzed using DGGE. Nine predominant bands were observed as a result of DGGE analysis (Supplementary Figure 3A). In particular, bands 6 and 7 were specifically observed only in STRcel, whereas other major bands were detected in all STRs. These two bands were detected most strongly on the second day of the experiment, suggesting that the bacteria originating from bands 6 and 7 played an important role in cellulose degradation. It should be noted that Archaea, as well as bacteria, exist in the fermentation sludge. Although the universal bacterial primers we used were unable to detect Archaea, their presence in the original sludge has been confirmed in previous reports.28−30 The proportion of these Archaea is relatively small compared with Bacteria (comprising about 1% or less of the total sludge population).30 Since, in addition, the Archaea are not involved in the conversion of carbohydrates to short-chain fatty acids, universal bacterial primer is appropriate for evaluating the glucose/starch/cellulose-utilizing and short-chain fatty acidsproducing microbial community. To obtain more definitive information regarding the taxonomy of these major bands, a phylogenetic tree was constructed based on the 16S rRNA gene fragments derived from the DGGE gel bands (Figure 2). As a result of sequencing, DNA sequences from bands 1−5 were categorized in phylum Thermotogae, known as thermophilic bacteria, and those from bands 6−9 were categorized in phylum Firmicutes (Figure 2 and Supplementary Figure 3B). In particular, within this group [Firmicutes], the source organisms of bands 6 and 7 were related to Firmicutes bacterium VNS03 and Halocella cellulolytica, respectively. Some of these bacteria are known to possess the ability to degrade cellulose.44,45 Thus, known cellulose-degrading bacteria were detected in the phylogenetic tree as close relatives of the bacteria in band 7.

H chemical shift (ppm) 1.91 1.54 2.17 1.46 3.77 0.88 1.56 2.13 1.29 2.17 1.54 1.16 3.64 8.44 3.72 3.89 3.76 3.83 3.45 3.82 3.40 3.47 3.71 3.24 3.53 4.64 5.23 1.32 4.11 1.04 2.17 3.53 7.29 7.30 7.37

13

C chemical shift (ppm) 26.0 28.4 40.1 18.7 53.2 15.9 21.9 42.2 31.2 40.1 28.4 19.6 60.2 173.9 63.4 63.4 63.3 63.3 78.7 74.2 72.3 78.5 75.5 76.8 74.2 98.6 94.8 22.7 71.1 12.8 33.3 47.1 131.8 129.0 131.3

The intensities of the nine major bands detected in DGGE analysis were digitized and compared statistically to evaluate the variability of the microbial community in each STR (Supplementary Figure 4A). Plots of PCA scores for DGGE fingerprinting data revealed that the microbial community profiles were likely to cluster according to differences in the substrates added to the STRs (Figure 3A and Supplementary Figure 5A). In particular, two clusters were likely to separate along the PC2 direction. The score plots of the data in STRcel tended to be located in the positive direction in PC2 whereas data from the other STR were likely to be located in a negative direction in PC2. The variation in all of the major bands detected in the DGGE analysis, according to loading plot data (Figure 3B and Supplementary Figure 5B), contributed to this separation. Additionally, the daily variation of the microbial community in STRcel was relatively large compared with the other STRs. Moreover, the daily variation of the microbial community in the STRglu and STRsta tended to differ. These results suggest that variations in the microbial community in each STR were characteristically formed as a result of the impact of feeding different substrates although the compositions of the major microbial community detected by DGGE analysis were similar in all STRs. 5606

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

Figure 4. Two-dimensional correlation maps depicted by the relationships between the metabolic profiles detected by 1H NMR measurements and the major microbial community profiles observed using DGGE analysis for STRglu (A), STRsta (B), and STRcel (C). The 2D correlation maps were calculated for a time period ranging from 0 to 11 days for all STRs. Red and blue denote positive and negative correlations, respectively. A coloring cutoff of 0.7 was chosen to identify and omit positive (or negative) correlations due to noise. Bands 6 and 7 were specifically observed only in STRcel, and are highlighted in gray in STRglu and STRsta. Abbreviations: A, acetic acid; B, butyric acid; E, ethanol; F, formic acid; G, glucose; P, propionic acid; Ph, phenylacetic acid.

Metabolic Profiles in Each STR

These metabolites contributed to the metabolite production processes by the microbial community in each STR, as shown in the loading plots (Figure 3D and Supplementary Figure 5D). Moreover, some metabolites, such as phenylacetic acid detected by HSQC and TOCSY methods, were considered to be inherent ingredients originating from the seed sludge and remained as indigestible ingredients during the experiments. Next, variation in certain metabolites, such as acetic acid, butyric acid, formic acid, lactic acid, propionic acid, and ethanol, detected by 1H NMR, HSQC, and TOCSY analyses, were evaluated in terms of their peak heights of 1H NMR signals (Supplementary Figure 7). The glucose signal was completely erased after 2 days and some metabolites were produced by the microbial community in STRglu. In contrast, although starch and cellulose were expected to be hydrolyzed by enzymes to monosaccharides, no sugar signal was detected in STRsta or STRcel. This suggested that sugars produced from starch and cellulose were immediately consumed by microbial communities in the STRs when starch and cellulose were hydrolyzed. Moreover, variations in certain metabolites did exist between the STRs. For example, the concentrations of butyric acid were significantly increased at the second day in the STRglu and STRcel and at the third day in the STRsta. However, the concentration in the STRcel was markedly decreased on the third day, while the concentrations in the STRglu and STRsta continued to be high until the fifth day. Thus, differing metabolic dynamics of the microbial communities in each STR were revealed, although the major metabolites detected in the analyses were almost identical in all STRs.

To identify and compare the metabolic dynamics by microbial community in each STR, metabolic profiles in each STR were determined using a NMR-based metabonomics approach. All 1H NMR chemical shift data were digitized and statistically compared with evaluate metabolic variability in each STR (Supplementary Figure 4B). Subsequently, PCA of the 1 H NMR chemical shift data was performed excluding data from the first day in STRglu because the high intensity of glucose signals in the data made it extremely difficult to observe other data and markedly affected calculation (Figure 3C and Supplementary Figure 5C). Plots of PCA scores for NMR data revealed that metabolic profiles were not likely to cluster according to differences in substrates added to the STRs, while microbial community profiles were likely to cluster according to those differences (Figure 3 and Supplementary Figure 5). However, score plots of metabolic profiles on similar days in each STR tended to be relatively similar, and the plots on the 11th day gathered in one cluster in all STRs. This result suggests that the metabolites produced in the microbial community in each STR were all similar, although the metabolic profiles in each STR varied somewhat. Moreover, as time progressed, the metabolite production in each STR converged toward a common point on the PCA score plot, so that by day 11, differences in metabolite production processes had largely disappeared. To obtain more detailed information about the metabolites, a mixture of all samples from STRglu was assessed using HSQC and TOCSY methods to assign the metabolites in the STR (Supplementary Figure 6 and Table 1). These analyses revealed that some metabolites, such as acetic acid, butyric acid, formic acid, and propionic acid, were major metabolic products. 5607

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

Figure 5. Visualization of metabolic sequences processed by microbial community in STRglu (A), STRsta (B), and STRcel (C). Associations between specific metabolites and each band detected in the DGGE analysis are shown as the correlation lines indicated by red (positive) or blue (negative). Bands 6 and 7 were specifically observed only in STRcel, and are highlighted in gray in STRglu and STRsta. Solid lines in the right panels indicate the production of metabolites, and broken lines in the right panels indicate the utilization of metabolites by the microbial community. Each box in the right panels was defined in ranges from a top (or near top) peak after marked elevation of each metabolite to below half the amount of the top peak, calculated according to the quantitative data shown in Supplementary Figure 7. Abbreviations: B1 to B9, bands 1 to 9, detected in DGGE analysis; A, acetic acid; B, butyric acid; E, ethanol; F, formic acid; P, propionic acid.

Visualization of Metabolic Sequences Based on DGGE-NMR Correlation Analysis

that glucose was used by bacteria related to bands 2, 3, 4, 8, and 9. Additionally, band 8 was positively correlated with propionic acid and butyric acid in the STRglu. This indicates the possibility that propionic acid and butyric acid were produced by bacteria related to band 8. Taken together, the microbial variability and metabolic dynamics of the microbial community were found to be closely linked to each other. Interestingly, 2D correlation maps indicated differences in the roles and relationships between metabolites and bacteria in the substrate degradation processes in each STR. For example, the production of propionic acid was due to the bacteria related to band 8 in the STRglu and to bands 5, 8, and 9 in the STRsta, whereas it was due to bacteria related to band 1 in the STRcel. In this manner, the role played by individual bacteria in the metabolic dynamics of the microbial community in each STR differed, as summarized in Figure 5. Thus, using the approach described here, it was possible to visualize characteristic metabolic sequences in anaerobic fermentation ecosystems. In this study, glucose, starch, and cellulose were used as feeding substrates to enable visualization and comparison of metabolic sequences in anaerobic fermentation ecosystems. Starch and cellulose are composed of glucose as a monomeric component, but cellulose, in particular, is a persistent substance. Our results revealed that changing the feeding substrate (glucose, starch, or cellulose)

To link the variability in microbial community profiles and metabolic profiles in each STR, we performed microbial− metabolic correlation analyses based on DGGE fingerprints and 1 H NMR chemical shift data using Pearson’s product-moment correlation coefficient (Figure 4). The correlation coefficients in the DGGE and NMR matrices were calculated across all days (i.e., were determined from time course data). For example, in the STRcel, the correlation coefficient (value 0.88) between band 7 and acetate (1.91 ppm) was calculated from time series data (i.e., from the intensities of NMR and DGGE signals at 0, 1, 2, 3, 5, and 11 days) based on time course population changes of bacteria associated with band 7 and time course concentration changes of acetate. This correlation coefficient (positive correlation) implies that the acetate and band 7 exhibit similar behavior in the STRcel. In fact, the correlation coefficients of all NMR signals versus all DGGE bands were incorporated into a correlation coefficient matrix, from which the heat maps for each STR (shown in Figure 4) were derived. This DGGE-NMR correlation analysis identified the relationship between microbial community and metabolites. For example, bands 2, 3, 4, 8, and 9 were negatively correlated with glucose in the STRglu (Figures 4 and 5). This indicates the possibility 5608

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

+81455039439; Fax: +81455039489; E-mail: kikuchi@psc. riken.jp.

provided to the ecosystems caused perturbations in the microbial community; substrates were processed by similar reactions and alternate players, as summarized in Figure 5. In particular, the ecosystems processed cellulose using key players, such as those in bands 6 and 7, which were related to Firmicutes as detected by DGGE analyses, and which were capable of degrading persistent macromolecules. Eventually, the ecosystems performed similar tasks with all feeding substrates, although there were differences in the exometabolome, such as acetic acid and propionic acid being produced at different times in the substrate degradation processes. To our knowledge, this study is the first attempt to use a sequential metabolism approach to visualize the key players, reactions, and metabolic dynamics of a microbial ecosystem. This approach will enable researchers to evaluate and visualize the tasks of microbial ecosystems, including those in environmental ecosystems and host-microbial symbiotic ecosystems. Our approach is a powerful tool to visualize and evaluate metabolic sequences in specialized symbiotic ecosystems, such as those found in humans and animals.9,10,46,47 Humans can be considered superorganisms as a result of their close symbiotic associations with intestinal microbiota. These microbiota can process indigestible components, such as plant polysaccharides, and can be significantly affected by alterations in diet. Dietary variation, accompanied by compositional changes in the microbiota of the intestinal environment, is likely to affect human health and disease, so visualization of metabolic sequences using our approach may provide a foundation for evaluating systemic effects of drugs and diet that are relevant to personal and public health. This application of our approach is a step toward clarifying the metabolic dynamics in the complex intestinal microbial community. It will also allow the translation of microbial activity and function into host responses, and may even provide a means to engineer the metabolic activities of intestinal microorganisms to improve human health.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors wish to thank Eisuke Chikayama (RIKEN) for stimulate discussion and technical advice for NMR measurements and analysis. This research was supported in part by Grants-in-Aid for Scientific Research for challenging exploratory research (J.K.), and Scientific Research (A) (J.K.) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. This work was also supported, in part, by grants from the New Energy and Industrial Technology Development Organization (NEDO to J.K.).

■

■

CONCLUSIONS This study focused on the degradation processes of glucose, starch, and cellulose in anaerobic fermentation ecosystems, and visualized and compared metabolic dynamics in individual microbial communities using a sequential metabolism approach based on DGGE-NMR correlation analysis. This is the first report to visualize the entire metabolic dynamics in anaerobic fermentation ecosystems as metabolic sequences. This approach may be a useful tool for visualizing and evaluating the metabolic sequences of biomass degradation processes in environmental ecosystems. By applying our approach not only to anaerobic fermentation ecosystems but also to other complex and diverse microbial ecosystems, including various environments such as oceans, soils, and symbiotic ecosystems, we should be able to decipher and readily compare comprehensive metabolic sequences in a variety of microbial ecosystems.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional infotrmation as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Nicholson, J. K.; Holmes, E.; Lindon, J. C.; Wilson, I. D. The challenges of modeling mammalian biocomplexity. Nat. Biotechnol. 2004, 22, 1268−1274. (2) Wahlen, M. The global methane cycle. Annu. Rev. Earth Planet. Sci. 1993, 21, 407−426. (3) Chynoweth, D. P.; Owens, J. M.; Legrand, R. Renewable methane from anaerobic digestion of biomass. Renew. Energ. 2001, 22, 1−8. (4) Han, S. K.; Shin, H. S. Biohydrogen production by anaerobic fermentation of food waste. Int. J. Hydrogen Energy 2004, 29, 569−577. (5) Fukuda, S.; Toh, H.; Hase, K.; Oshima, K.; Nakanishi, Y.; Yoshimura, K.; Tobe, T.; Clarke, J. M.; Topping, D. L.; Suzuki, T.; Taylor, T. D.; Itoh, K.; Kikuchi, J.; Morita, H.; Hattori, M.; Ohno, H. Bifidobacteria can protect from enteropathogenic infection through production of acetate. Nature 2011, 469, 543−547. (6) Holmes, E.; Loo, R. L.; Stamler, J.; Bictash, M.; Yap, I. K. S.; Chan, Q.; Ebbels, T.; De lorio, M.; Brown, I. J.; Veselkov, K. A.; Daviglus, M. L.; Kesteloot, H.; Ueshima, H.; Zhao, L.; Nicholson, J. K.; Elliott, P. Human metabolic phenotype diversity and its association with diet and blood pressure. Nature 2008, 453, 396−400. (7) Jia, W.; Li, H.; Zhao, L.; Nicholson, J. K. Gut microbiota: a potential new territory for drug targeting. Nat. Rev. Drug Discovery 2008, 7, 123−129. (8) Li, M.; Wang, B.; Zhang, M.; Rantalainen, M.; Wang, S.; Zhou, H.; Zhang, Y.; Shen, J.; Pang, X.; Zhang, M.; Wei, H.; Chen, Y.; Lu, H.; Zuo, J.; Su, M.; Qiu, Y.; Jia, W.; Xiao, C.; Smith, L. M.; Yang, S.; Holmes, E.; Tang, H.; Zhao, G.; Nicholson, J. K.; Li, L.; Zhao, L. Symbiotic gut microbes modulate human metabolic phenotypes. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 2117−2122. (9) Waldram, A.; Holmes, E.; Wang, Y.; Rantalainen, M.; Wilson, I. D.; Tuohy, K. M.; McCartney, A. L.; Gibson, G. R.; Nicholson, J. K. Top-down systems biology modeling of host metabotype-microbiome associations in obese rodents. J. Proteome Res. 2009, 8, 2361−2375. (10) Yap, I. K. S.; Li, J. V.; Saric, J.; Martin, F. P.; Davies, H.; Wang, Y.; Wilson, I. D.; Nicholson, J. K.; Utzinger, J.; Marchesi, J. R.; Holmes, E. Metabonomic and microbiological analysis of the dynamic effect of vancomycin-induced gut microbiota modification in the mouse. J. Proteome Res. 2008, 7, 3718−3728. (11) Chen, F.; Zhang, J.; Song, X.; Yang, J.; Li, H.; Tang, H.; Liao, Y. C. Combined metabonomic quantitative real-time PCR analyses reveal systems metabolic changes of Fusarium graminearum induced by Tri5 gene deletion. J. Proteome Res. 2011, 10, 2273−2285. (12) Liu, C.; Hao, F.; Hu, J.; Zhang, W.; Wan, L.; Zhu, L.; Tang, H.; He, G. Revealing different systems responses to brown planthopper infestation for pest susceptible and resistant rice plants with the combined metabonomic and gene-expression analysis. J. Proteome Res. 2010, 9, 6774−6785. (13) Nicholson, J. K.; Lindon, J. C.; Holmes, E. ‘Metabonomics’: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica 1999, 29, 1181−1189.

AUTHOR INFORMATION

Corresponding Author

*Address: RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan. Phone: 5609

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610

Journal of Proteome Research

Article

flow anaerobic packed-bed reactor. Bioresour. Technol. 2008, 99, 4786− 4795. (31) Yu, Z.; Morrison, M. Comparisons of different hypervariable regions of rrs genes for use in fingerprinting of microbial communities by PCR-denaturing gradient gel electrophoresis. Appl. Environ. Microbiol. 2004, 70, 4800−4806. (32) Date, Y.; Sakata, K.; Kikuchi, J. Chemical profiling of complex biochemical mixtures from various seaweeds. Polymer J. 2012, 44, 888−894. (33) Fukuda, S.; Nakanishi, Y.; Chikayama, E.; Ohno, H.; Hino, T.; Kikuchi, J. Evaluation and characterization of bacterial metabolic dynamics with a novel profiling technique, real-time metabolotyping. PLoS One 2009, 4, e4893. (34) Kikuchi, J.; Hirayama, T. Practical aspects of uniform stable isotope labeling of higher plants for heteronuclear NMR-based metabolomics. Methods Mol. Biol. 2007, 358, 273−286. (35) Nakanishi, Y.; Fukuda, S.; Chikayama, E.; Kimura, Y.; Ohno, H.; Kikuchi, J. Dynamic omics approach identifies nutrition-mediated microbial interactions. J. Proteome Res. 2011, 10, 824−836. (36) Sekiyama, Y.; Chikayama, E.; Kikuchi, J. Profiling polar and semipolar plant metabolites throughout extraction processes using a combined solution-state and high-resolution magic angle spinning NMR approach. Anal. Chem. 2010, 82, 1643−1652. (37) Sekiyama, Y.; Chikayama, E.; Kikuchi, J. Evaluation of a semipolar solvent system as a step toward heteronuclear multidimensional NMR-based metabolomics for 13C-labeled bacteria, plants, and animals. Anal. Chem. 2011, 83, 719−726. (38) Sekiyama, Y.; Kikuchi, J. Towards dynamic metabolic network measurements by multi-dimensional NMR-based fluxomics. Phytochemistry 2007, 68, 2320−2329. (39) Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeifer, J.; Bax, A. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 1995, 6, 277−293. (40) Kikuchi, J.; Shinozaki, K.; Hirayama, T. Stable isotope labeling of Arabidopsis thaliana for an NMR-based metabolomics approach. Plant Cell Physiol. 2004, 45, 1099−1104. (41) Akiyama, K.; Chikayama, E.; Yuasa, H.; Shimada, Y.; Tohge, T.; Shinozaki, K.; Hirai, M. Y.; Sakurai, T.; Kikuchi, J.; Saito, K. PRIMe: a Web site that assembles tools for metabolomics and transcriptomics. In Silico Biol. 2008, 8, 339−45. (42) Chikayama, E.; Sekiyama, Y.; Okamoto, M.; Nakanishi, Y.; Tsuboi, Y.; Akiyama, K.; Saito, K.; Shinozaki, K.; Kikuchi, J. Statistical indices for simultaneous large-scale metabolite detections for a single NMR spectrum. Anal. Chem. 2010, 82, 1653−1658. (43) Chikayama, E.; Suto, M.; Nishihara, T.; Shinozaki, K.; Hirayama, T.; Kikuchi, J. Systematic NMR analysis of stable isotope labeled metabolite mixtures in plant and animal systems: coarse grained views of metabolic pathways. PLoS One 2008, 3, e3805. (44) Schwarz, W. H. The cellulosome and cellulose degradation by anaerobic bacteria. Appl. Microbiol. Biotechnol. 2001, 56, 634−649. (45) Simankova, M. V.; Chernych, N. A.; Osipov, G. A.; Zavarzin, G. A. Halocella cellulolytica gen. nov., sp. nov., a new obligately anaerobic, halophilic, cellulolytic bacterium. Syst. Appl. Microbiol. 1993, 16, 385−389. (46) Le Gall, G.; Noor, S. O.; Ridgway, K.; Scovell, L.; Jamieson, C.; Johnson, I. T.; Colquhoun, I. J.; Kemsley, E. K.; Narbad, A. Metabolomics of fecal extracts detects altered metabolic activity of gut microbiota in ulcerative colitis and irritable bowel syndrome. J. Proteome Res. 2011, 10, 4208−4218. (47) Swann, J.; Wang, Y.; Abecia, L.; Costabile, A.; Tuohy, K.; Gibson, G.; Roberts, D.; Sidaway, J.; Jones, H.; Wilson, I. D.; Nicholson, J.; Holmes, E. Gut microbiome modulates the toxicity of hydrazine: a metabonomic study. Mol. BioSyst. 2009, 5, 351−355.

(14) Nicholson, J. K.; Connelly, J.; Lindon, J. C.; Holmes, E. Metabonomics: a platform for studying drug toxicity and gene function. Nat. Rev. Drug Discovery 2002, 1, 153−161. (15) Bundy, J. G.; Davey, M. P.; Viant, M. R. Environmental metabolomics: a critical review and future perspectives. Metabolomics 2009, 5, 3−21. (16) Bundy, J. G.; Keun, H. C.; Sidhu, J. K.; Spurgeon, D. J.; Svendsen, C.; Kille, P.; Morgan, A. J. Metabolic profile biomarkers of metal contamination in a sentinel terrestrial species are applicable across multiple sites. Environ. Sci. Technol. 2007, 41, 4458−4464. (17) Bundy, J. G.; Lenz, E. M.; Osborn, J. M.; Weeks, J. M.; Lindon, J. C.; Nicholson, J. K. Metabolism of 4-fluoroaniline and 4-fluorobiphenyl in the earthworm Eisenia veneta characterized by highresolution NMR spectroscopy with directly coupled HPLC-NMR and HPLC-MS. Xenobiotica 2002, 32, 479−490. (18) Bundy, J. G.; Osborn, D.; Weeks, J. M.; Lindon, J. C.; Nicholson, J. K. An NMR-based metabonomic approach to the investigation of coelomic fluid biochemistry in earthworms under toxic stress. FEBS Lett. 2001, 500, 31−35. (19) Bundy, J. G.; Spurgeon, D. J.; Svendsen, C.; Hankard, P. K.; Weeks, J. M.; Osborn, D.; Lindon, J. C.; Nicholson, J. K. Environmental metabonomics: applying combination biomarker analysis in earthworms at a metal contaminated site. Ecotoxicology 2004, 13, 797−806. (20) Bundy, J. G.; Spurgeon, D. J.; Svendsen, C.; Hankard, P. K.; Osborn, D.; Lindon, J. C.; Nicholson, J. K. Earthworm species of the genus Eisenia can be phenotypically differentiated by metabolic profiling. FEBS Lett. 2002, 521, 115−120. (21) Frimmersdorf, E.; Horatzek, S.; Pelnikevich, A.; Wiehlmann, L.; Schomburg, D. How Pseudomonas aeruginosa adapts to various environments: a metabolomic approach. Environ. Microbiol. 2010, 12, 1734−1747. (22) Hughes, S. L.; Bundy, J. G.; Want, E. J.; Kille, P.; Sturzenbaum, S. R. The metabolomic responses of Caenorhabditis elegans to cadmium are largely independent of metallothionein status, but dominated by changes in cystathionine and phytochelatins. J. Proteome Res. 2009, 8, 3512−3519. (23) Viant, M. R.; Bearden, D. W.; Bundy, J. G.; Burton, I. W.; Collette, T. W.; Ekman, D. R.; Ezernieks, V.; Karakach, T. K.; Lin, C. Y.; Rochfort, S.; de Ropp, J. S.; Teng, Q.; Tjeerdema, R. S.; Walter, J. A.; Wu, H. International NMR-based environmental metabolomics intercomparison exercise. Environ. Sci. Technol. 2009, 43, 219−225. (24) Viant, M. R.; Rosenblum, E. S.; Tjeerdema, R. S. NMR-Based metabolomics: a powerful approach for characterizing the effects of environmental stressors on organism health. Environ. Sci. Technol. 2003, 37, 4982−4989. (25) Date, Y.; Nakanishi, Y.; Fukuda, S.; Kato, T.; Tsuneda, S.; Ohno, H.; Kikuchi, J. New monitoring approach for metabolic dynamics in microbial ecosystems using stable-isotope-labeling technologies. J. Biosci. Bioeng. 2010, 110, 87−93. (26) Everroad, R. C.; Yoshida, S.; Tsuboi, Y.; Date, Y.; Kikuchi, J.; Moriya, S. Metabolic profiling of microbial planktonic communities by nuclear magnetic resonance spectroscopy for environmental transomics approaches. J. Vis. Exp. 2012, 62, e3163. (27) Ogata, Y.; Chikayama, E.; Morioka, Y.; Everroad, R. C.; Shino, A.; Matsushima, A.; Haruna, H.; Moriya, S.; Toyoda, T.; Kikuchi, J. ECOMICS: A web-based toolkit for investigating the biomolecular web in ecosystems using a trans-omics approach. PLoS One 2012, 7, e30263. (28) Sasaki, K.; Haruta, S.; Tatara, M.; Yamazawa, A.; Ueno, Y.; Ishii, M.; Igarashi, Y. Microbial community in methanogenic packed-bed reactor successfully operating at short hydraulic retention time. J. Biosci. Bioeng. 2006, 101, 271−273. (29) Tatara, M.; Yamazawa, A.; Ueno, Y.; Fukui, H.; Goto, M.; Sode, K. High-rate thermophilic methane fermentation on short-chain fatty acids in a down-flow anaerobic packed-bed reactor. Bioprocess Biosyst. Eng. 2005, 27, 105−113. (30) Tatara, M.; Makiuchi, T.; Ueno, Y.; Goto, M.; Sode, K. Methanogenesis from acetate and propionate by thermophilic down5610

dx.doi.org/10.1021/pr3008682 | J. Proteome Res. 2012, 11, 5602−5610