Interaction Networks Are Driven by Community-Responsive Phenotypes in a Chitin-Degrading Consortium of Soil Microbes
ABSTRACT
Soil microorganisms provide key ecological functions that often rely on metabolic interactions between individual populations of the soil microbiome. To better understand these interactions and community processes, we used chitin, a major carbon and nitrogen source in soil, as a test substrate to investigate microbial interactions during its decomposition. Chitin was applied to a model soil consortium that we developed, “model soil consortium-2” (MSC-2), consisting of eight members of diverse phyla and including both chitin degraders and nondegraders. A multiomics approach revealed how MSC-2 community-level processes during chitin decomposition differ from monocultures of the constituent species. Emergent properties of both species and the community were found, including changes in the chitin degradation potential of Streptomyces species and organization of all species into distinct roles in the chitin degradation process. The members of MSC-2 were further evaluated via metatranscriptomics and community metabolomics. Intriguingly, the most abundant members of MSC-2 were not those that were able to metabolize chitin itself, but rather those that were able to take full advantage of interspecies interactions to grow on chitin decomposition products. Using a model soil consortium greatly increased our knowledge of how carbon is decomposed and metabolized in a community setting, showing that niche size, rather than species metabolic capacity, can drive success and that certain species become active carbon degraders only in the context of their surrounding community. These conclusions fill important knowledge gaps that are key to our understanding of community interactions that support carbon and nitrogen cycling in soil.
IMPORTANCE The soil microbiome performs many functions that are key to ecology, agriculture, and nutrient cycling. However, because of the complexity of this ecosystem we do not know the molecular details of the interactions between microbial species that lead to these important functions. Here, we use a representative but simplified model community of bacteria to understand the details of these interactions. We show that certain species act as primary degraders of carbon sources and that the most successful species are likely those that can take the most advantage of breakdown products, not necessarily the primary degraders. We also show that a species phenotype, including whether it is a primary degrader or not, is driven in large part by the membership of the community it resides in. These conclusions are critical to a better understanding of the soil microbial interaction network and how these interactions drive central soil microbiome functions.
INTRODUCTION
The soil microbiome carries out several important ecological functions, including carbon (C) and nitrogen (N) cycling and plant growth promotion (1–3). Central to these functions are interactions between the species that comprise soil microbial communities (4–6). While the combined genomic and metabolic potential of the individual species of the soil microbiome is vast, novel functions can emerge at the community level through metabolic interactions (7). A better understanding of these interactions will lead to a more complete view of the constituent organism’s and community’s functional capacity and will greatly expand our knowledge of how interaction networks can be affected by nutrient or environmental shifts (8, 9).
One of the most enigmatic microbial functions that depends on species interactions is the decomposition of soil organic matter. Identification of metabolic interactions involved in decomposition is particularly challenging due to the complexity of organic substrates in soil, the biodiversity of organisms involved, and difficulty in extracting samples at microbiological scales in soil (10). As a result, few studies have interrogated the taxon-specific gene expression and community metabolism that occur during C decomposition in soil. Microbially driven breakdown of plant-derived matter, such as cellulose (11, 12), is of great interest due to its ubiquity in soil environments. While the genomic potential for cellulase enzyme production has been detected in almost 40% of bacterial genomes in the Carbohydrate-Active Enzyme database, only a small number of organisms have been shown to digest cellulose independently in pure culture (13). The challenges described above in soil have made interactions centered on breakdown of cellulose difficult to ascertain. Even less is known about the decomposition of additional abundant molecules such as chitin that contribute to both carbon and nitrogen cycling in soil (14, 15). While chitinases, like cellulases, are widespread in bacteria (16), their expression is not universal and is differentially controlled by different species or even within a population (17). Previous studies using fluorescent reporter assays focused on both transcripts and proteins showed that only a subpopulation of cells in a pure culture of a chitin-degrading strain actually produce chitinases, with the remaining cells feeding off breakdown products (18, 19). In addition, chitinases are often not cell associated, meaning breakdown products are available not only to the species expressing the chitinase but also to other species of local community as well. As such, there is a strong element of inter/intraspecies interactions centered on community metabolism of chitin and its decomposition products.
The molecular details of community interactions during chitin decomposition and the generation of breakdown products have been previously studied in some detail in non-soil environments, especially in aquatic systems (16, 20). Chitin is the polymer of (1→4)-β-linked N-acetyl-d-glucosamine (NAG) monomers, and the decomposition of chitin into its NAG monomers is driven either by membrane-bound chitinases or through excreted non-cell-associated enzymes. Once converted into NAG oligomers, import and intracellular metabolism are possible and dimers are converted to monomers via β-N-acetylglucosaminidases. Numerous species have been found that contain transporters for uptake of NAG molecules or metabolic genes that act on NAG without a corresponding set of chitin-degrading genes, suggesting that some bacteria rely on NAG produced through chitin breakdown carried out by other species in their neighborhood (21). In supporting chitin degraders, it is possible that NAG consumers contribute to more effective chitin breakdown by removing downstream metabolites to increase enzyme efficiency (22) or providing additional metabolic benefits such as vitamins to alleviate the energetic costs expended by chitinase producers (23). As a result, growth promotion is driven by chitin degraders and secondary consumers through metabolic cross talk, where growth of constituents is maximized through cooperation.
Ecological theory suggests that the ability of a species to take advantage of exometabolites for growth is driven in part by the size of a species’ fundamental niche (C sources they can metabolize themselves) compared to the realized niche (C sources that can be metabolized by the complete community) (24). A large fundamental niche indicates that species may be able to take advantage of the presence of many exometabolites independently of other species. Yet it is not clear who has an advantage during growth with chitin as the sole carbon source—chitin degraders or consumers—nor is it clear how this advantage may be related to a species’ fundamental niche size. Laboratory growth experiments have suggested that primary consumers of complex carbon sources do not have a growth advantage (25), but whether this is a consistent rule and how this extends to field environments is not known.
Finally, the identity of chitin degraders or consumers may shift in response to community dynamics because species can express emergent properties with other species that they do not express during growth alone. For example, previous work from our group explored interactions centered on chitin breakdown by investigating a naturally evolved community of soil microbes developed using chitin as the major C and N source (26). Analyses of this community, MSC-1 (model soil consortium-1), identified several species, primarily a species of Rhodococcus, that occupied central positions in a 16S rRNA gene amplicon coabundance network, suggesting that they may be dominant chitin degraders and in turn provide assimilable substrates to other community members. Following isolation of specific constituents of MSC-1, subsequent coculture work revealed that several genera of this community (including Ensifer, Dyadobacter, and Rhizobium grown with Rhodococcus and Streptomyces grown with Ensifer) showed higher biomass when cocultured versus in monoculture. These initial studies suggest a network of interactions centered on chitin breakdown. However, there is a knowledge gap regarding the molecular details of how a community of chitin degraders and nondegraders organize themselves to break down chitin and share metabolic products.
Here, we aimed to fill these knowledge gaps by constructing and examining a new model community assembled from MSC-1 isolates, model soil consortium-2 (MSC-2). To delineate how MSC-2 degrades chitin, we used a multiomics approach combining species abundance analyses and expressed functions (metatranscriptomics) and extracellular nutrient pools (metabolomics) applied to chitin growth assays of the MSC-2 consortium as a community as well as its constituents in monoculture. We set out to answer three questions. (i) Which species have a growth advantage in a community—chitin degraders or species with large fundamental niches that can take the most advantage of chitin breakdown products? (ii) To what degree do MSC-2 members organize themselves into a chitin-degrading community, with each member contributing certain aspects of the breakdown process (cleaving of chitin polymer bonds, breakdown of NAG trimers or dimers, and processing of NAG monomers into further C and N pathways)? Finally, with an eye on native soil ecology, (iii) to what extent is the chitin decomposition phenotype of a species defined by the composition of the community in which this species grows? Answering these questions will shed light on how diverse bacterial species contribute to community decomposition and to what degree the ability to degrade organic C and/or assimilate breakdown products translates into a growth advantage in a mixed community.
RESULTS
Development of MSC-2 to investigate community and constituent roles in chitin degradation.
Our previous work described the development and analysis of a naturally evolved model microbial consortium of ~35 species (model soil consortium-1 [MSC-1]) (26). Here, we first obtained new isolates from this community by plating onto agar media and validated isolate identities by sequencing. As a result, members of the Sphingopyxis and Sinorhizobium genera were isolated, as well as novel strains of the Ensifer, Dyadobacter, and Streptomyces genera (see Table S1 in the supplemental material). As MSC-1 was originally isolated by enrichment under high-chitin conditions, we aimed to determine if any of the MSC-1 isolates could grow using chitin as the sole source of C (Fig. 1) under well-mixed liquid cultivation conditions. This analysis showed that Dyadobacter, one Ensifer strain (strain PNNL_MSC-1_Str_011), Variovorax, Sinorhizobium, and Rhodococcus grew with chitin, while the remaining strains of Streptomyces, Ensifer, Neorhizobium, and Sphingopyxis did not (Fig. 1B). We also tested for growth of Streptomyces, Neorhizobium, and Sphingopyxis on larger amounts of chitin (500 and 1,000 ppm) and observed similar lack of growth (data not shown). Comparisons of the Ensifer strains that did and did not grow on chitin showed some minor sequence differences but no apparent differences in annotated pathways. These results lay out the monoculture phenotypes of these species and allow for future comparisons of isolate phenotypes to emergent phenotypes expressed during community growth. From this set of 12 species, we selected eight that would comprise microbial soil consortium-2 (MSC-2) (Table 1). Selection was carried out with the goals of (i) including both chitin degraders and nondegraders and (ii) maximizing taxonomic diversity in MSC-2. Subsequent discussion of MSC-2 is defined as these eight species cultured together.
FIG 1
TABLE 1
| Isolate designation | GTDB taxonomya | No. of genes | No. of unique genes | Chitin growth | Max OD600b | Max OD SD | Time to reach Max OD (h) | Chitinase genesc | SOX system |
|---|---|---|---|---|---|---|---|---|---|
| PNNL_MSC-1_Str_001 | Streptomyces sp001905665 | 3,029 | 14 | No | 0.197 | 0.007 | NAd | (7/8) mannan endo-1 (GH5) (K01218), β-glucosidase (GH3) (EC 3.2.1.21), hexosaminidase (GH20) (EC 3.2.1.52), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), chitin-binding protein (K03933), chitinase (GH18) (EC 3.2.1.14), β-glucosidase (GH3) (EC 3.2.1.21), peptidoglycan-N-acetylglucosamine deacetylase (CE4) (EC 3.5.1.104), endoglucanase (GH5) (EC 3.2.1.4), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25), N-acetylglucosamine PTS system EIICBA or EIICB component (EC 2.7.1.193) | No |
| PNNL_MSC-1_Str_005 | Neorhizobium tomejilense | 3,619 | 48 | No | 0.183 | 0.011 | NA | (6/8) soluble lytic murein transglycosylase (GH23) (EC 4.2.2.−), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), endoglucanase (GH5) (EC 3.2.1.4), putative chitinase (GH19) (K03791), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25), glucosamine kinase (EC 2.7.1.8) | Yes |
| PNNL_MSC-1_Str_007 | Dyadobacter fermentans | 2,270 | 30 | Yes | 0.351 | 0.008 | 99 | (5/8) membrane-bound lytic murein transglycosylase D (GH23) (EC 4.2.2.−), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), hexosaminidase (GH20) (EC 3.2.1.52), chitinase (GH18) (EC 3.2.1.14), endoglucanase (GH5) (EC 3.2.1.4), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25), chitosanase (EC 3.2.1.132) | Yes |
| PNNL_MSC-1_Str_008 | Sphingopyxis fribergensis | 2,257 | 62 | No | 0.188 | 0.008 | NA | (4/8) soluble lytic murein transglycosylase (GH23) (EC 4.2.2.−), chitinase (GH18) (EC 3.2.1.14), hexosaminidase (GH20) (EC 3.2.1.52), β-glucosidase (GH3) (EC 3.2.1.21), N-acetylglucosamine PTS system EIICBA or EIICB component (EC 2.7.1.193), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25) | No |
| PNNL_MSC-1_Str_011 | Ensifer adhaerens | 3,978 | 26 | Yes | 0.289 | 0.022 | 35 | (6/8) membrane-bound lytic murein transglycosylase F (GH23) (EC 4.2.2.−), soluble lytic murein transglycosylase (GH23) (EC 4.2.2.−), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), putative chitinase (GH19) (K03791), hexosaminidase (GH20) (EC 3.2.1.52), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25), glucosamine kinase (EC 2.7.1.8) | No |
| PNNL_MSC-1_Str_012 | Variovorax beijingensis | 3,996 | 0 | Yes | 0.322 | 0.015 | 109 | (3/8) peptidoglycan-N-acetylglucosamine deacetylase (CE4) (EC 3.5.1.104), β-glucosidase (GH3) (EC 3.2.1.21), membrane-bound lytic murein transglycosylase D (GH23) (EC 4.2.2.−), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), soluble lytic murein transglycosylase (GH23) (EC 4.2.2.−), poly-β-1,6-N-acetyl-d-glucosamine N-deacetylase (CE4) (EC 3.5.1.−) | No |
| PNNL_MSC-1_Str_014 | Sinorhizobium meliloti | 3,535 | 50 | Yes | 0.259 | 0.011 | 42 | (7/8) soluble lytic murein transglycosylase (GH23) (EC 4.2.2.−), putative chitinase (GH19) (K03791), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52), hexosaminidase (GH20) (EC 3.2.1.52), glucosamine kinase (EC 2.7.1.8), N-acetylglucosamine-6-phosphate deacetylase (EC 3.5.1.25) | Yes |
| PNNL_MSC-1_Str_016 | Rhodococcus sp003130705 | 3,602 | 210 | Yes | 0.697 | 0.101 | 78 | (3/8) mannan endo-1,4-β-mannosidase (GH5) (EC 3.2.1.78), peptidoglycan-N-acetylglucosamine deacetylase (CE4) (EC 3.5.1.104), β-N-acetylhexosaminidase (GH3) (EC 3.2.1.52) | No |
a
GTDB, Genome Taxonomy Database.
b
Maximum (Max) optical density at 600 nm for each species when grown on chitin. Results are the average from three replicates.
c
The numbers in parentheses indicate how many chitin metabolism genes this species contained out of eight included.
d
NA, not applicable (no growth apparent).
Measuring fundamental niche size of MSC-2 members.
As a major goal of this study was to demonstrate the power of model communities to test ecological theory, we next carried out experiments using MSC-2 to test how fundamental niche size affects success, as measured by relative abundance, in a community. We assessed niche size by measuring the growth of each species under a variety of C sources as described in Materials and Methods (Fig. 2). While many of the tested substrates could be potential C and/or N sources, all experiments were carried out using the substrates as the sole C source and with ammonium added as a source of N. In certain cases, this led to a change in C/N ratios for some substrates. Because the ability of MSC-2 species to use these substrates as a source of C was the focus of these experiments, all substrates were added at the same concentration (10 mM).
FIG 2
All MSC-2 species were able to use glucose, NAG, glutamate, or arabinose as the sole C source, and all species except Neorhizobium, Streptomyces, and Rhodococcus were able to use alanine, xylose, and fructose, respectively, as well (Fig. 2A). Only Ensifer was able to use glycine as a substrate. Growth on other C and N sources was mixed, but generally most strains were able to use most nutrient sources.
With this information, we next generated metabolic models for each of the eight MSC-2 species using the Department of Energy Systems Biology Knowledgebase (KBase) (27) and simulated the biomass production of each isolate grown axenically on the suite of carbon sources using flux balance analysis (FBA) (Fig. 2B). With this analysis, the biomass yields on NAG for all species were larger than those on glucose, fructose, arabinose, and several amino acids. Similarly, for those that grew on disaccharides, maltose, and sucrose, the biomass yield per millimole of C source was almost double that predicted when grown on glucose and was also higher than that on NAG. Amino acids in general led to lower predicted biomass yields for MSC-2 species than sugars. This may be an effect of these amino acids also acting as an N source in experiments where an abundant N source (ammonia) was also present, possibly leading to altered C/N ratios that may affect growth as has been seen previously (28, 29). This was one of main reasons we chose chitin in subsequent experiments, to investigate a coupled C and N source. All of the members of MSC-2 had high yields on the main C breakdown product of chitin (NAG) but also responded well to other C sources, showing that while they are well suited to a high-chitin environment, they have pathways to metabolize several other C sources that are prevalent in soil.
The predicted growth patterns of the MSC-2 species were used to infer their fundamental and realized niches (24) (Fig. 2C). Fundamental niches are defined as the C sources that an MSC-2 species is predicted to metabolize itself (i.e., when grown in monoculture). The realized niche is the set of carbon sources that can be metabolized by the community as a whole, the sum of each specie’s fundamental niche. This analysis revealed that Ensifer had the largest fundamental niche followed by Streptomyces, Variovorax, and Rhodococcus. In contrast, Neorhizobium and Dyadobacter had the smallest fundamental niches. This suggests that Ensifer can take the most advantage of interactions resulting from residing in the MSC-2 community, compared to monoculture growth on a single C source, due to its larger fundamental niche; species with smaller niches may not be able to benefit as much from metabolic products resulting from community growth.
Genomic analysis of MSC-2 members.
Whole-genome sequencing was used to identify MSC-2 members to the greatest taxonomic depth possible and to gather information on their metabolic potential. Sequencing results and taxonomic and strain information are shown in Table 1. All MSC-2 members were predicted to contain at least some genes involved in chitin degradation, but the specific genes that each member contained varied (Fig. 3 and Table 1). Coverage of the chitin breakdown pathway was defined by having a certain number of the eight genes we assigned to the chitin/NAG breakdown and assimilation pathway. In addition, several other central processes (sulfur [S] and N assimilation, denitrification, mobility from flagellar genes, and secretion systems) varied among MSC-2 members (Fig. 3). Specifically, we observed substantial variability in the presence of sulfur oxidation genes (i.e., the SOX system) which oxidize thiosulfate to sulfate and represent the potential for sulfur acquisition. We also observed variation in the pentose phosphate pathway (PPP), which is involved in the creation of aromatic amino acids, the production of nucleic acid, the generation of NADPH, interconnections with glycolysis, and C fixation. A more complete PPP may allow for species to take advantage of multiple C pathways as opposed to pathways specific to chitin breakdown, leading changes in fundamental niche size described above. Finally, numerous MSC-2 members contained variable secretion systems which, besides their understood role in virulence, export numerous proteins or other substrates outside the cell. These secretion systems are involved in the production of extracellular chitin-degrading enzymes produced by certain members of MSC-2 and are thus critical for the chitin-related interspecies interactions driving this community (30).
FIG 3
Streptomyces (Fig. 3A) contains genes for all stages of chitin metabolism (despite this species showing low growth on chitin under liquid conditions) (Fig. 1B). For example, Streptomyces contains chitinase genes encoding enzymes to convert chitobiose to NAG and additional genes encoding enzymes to move NAG into the central glycolysis pathway via fructose-6-phosphate. Streptomyces also contains NAG importers along with several other importers. N and S assimilation pathways are also present in this microorganism.
Neorhizobium (Fig. 3A) contains a chitin breakdown pathway, although it is not as complete as that in Streptomyces, and it lacks a NAG importer. Despite this, it contains a larger number of other importers and exporters as well as complete N and S assimilation pathways. Neorhizobium also contains a denitrification pathway as well as several genes of the SOX system.
Dyadobacter (Fig. 3A), a species showing growth with chitin under our conditions, contains a reduced chitin breakdown pathway that is similar to that found in Neorhizobium (a species for which we found little chitin growth). The main differences are the existence of GH18 (a chitinase) and GH20 (a β-hexosaminidase) in Dyadobacter (Fig. 3C). Dyadobacter also has N and S assimilation and SOX genes. This species also has a more complete PPP system than Neorhizobium and Streptomyces. However, Dyadobacter has very few importers and none for NAG. Sphingopyxis (Fig. 3A), another species showing no growth on chitin, contains a chitin breakdown pathway like Streptomyces along with a NAG importer (though few other importers). This species lacks a SOX system, but it does contain N and S assimilation pathways as well as a nitrate reduction set of genes. Sphingopyxis also contains both type I and type II sets of secretion systems.
Ensifer (Fig. 3B) contains perhaps the widest range of processes among the species of MSC-2 and shows good growth on chitin. Many importers are present in this Ensifer species, though, interestingly, not an importer for NAG. Ensifer contains a chitin breakdown pathway that is focused on early stages of chitin breakdown as well as N and S assimilation pathways and several genes involved in denitrification. Ensifer also contains more a more complete PPP system, similar to Dyadobacter. A type IV secretion system is also present in Ensifer.
Variovorax (Fig. 3B) has a similar genomic profile to Ensifer, but with an additional type II secretion system and less complete chitin decomposition and denitrification pathways. Variovorax has many importers and exporters, although fewer than Sinorhizobium (Fig. 3B), and contains a chitin breakdown pathway like Streptomyces, with genes critical early in the chitin breakdown process, as well as a SOX system and many uptake genes—nearly as many as Ensifer. Sinorhizobium contains an N and S assimilation pathway, but it has a much less complete SOX pathway and no denitrification genomic potential.
Finally, Rhodococcus (Fig. 3B) has a somewhat reduced genomic potential compared to many of the other MSC-2 isolates. The chitin breakdown genes of Rhodococcus are minimal, and it does not contain an enhanced PPP system as Variovorax, Sinorhizobium, Ensifer, and Dyadobacter do. Rhodococcus does contain N and S assimilation pathways and some nitrate reduction genes, but it has no SOX genes. Rhodococcus does contain both type I and type II secretion systems and some importers and exporters: more than Dyadobacter and Sphingopyxis but fewer than other MSC-2 members.
The large variation in predicted genomic potential of the MSC-2 species reflects their various phylogenies and suggests that interactions among these species may be prevalent. This may be especially true in a chitin-rich environment considering the variability in the growth of these species under conditions where chitin is the sole C source (Fig. 1B) and their genomic potential for chitin/NAG metabolism (Fig. 3C). Detailed information on genes coding for other metabolic pathways aside from chitin are shown in Fig. S1.
Transcriptomic analysis of MSC-2 members during community growth on chitin.
A major goal of our study was to understand how the phenotypes of species, especially decomposers such as chitin degraders, are defined by their surrounding community. To that end, we next explored how the complete MSC-2 community responded to growth on chitin as the sole C source and how a member’s role in the community shifted compared to monoculture growth. While optical density (OD) measurements were difficult to obtain due to the large amount of nonsoluble chitin used in these coculture experiments, increases in the amount of both DNA and RNA were observed across time points from 4 to 118 h (approximately a 5-day growth assay) (Fig. S2). These increases in DNA and RNA yields indicated that MSC-2 grew using chitin as the sole C source.
Metatranscriptomic analysis of MSC-2 was carried out after 70 and 118 h of growth on chitin. Ensifer comprised the largest proportion of aligned reads at 70 h, representing 53% ± 14% of all reads (Fig. 4 and Table 2). This was followed by Rhodococcus (30% ± 19%), Streptomyces (12% ± 6%), and Variovorax (1.8% ± 0.66%), with the remaining species each present at less than 1% of total reads. Results at 118 h were similar to those at 70 h (Fig. 4), with the exception of Sphingopyxis, which increased from 0.17% ± 0.069% at 70 h to 1.61% ± 0.45% at 118 h. This 9.4-fold increase was significant with a P value of <0.02. Species abundance measures made using RNA alignment matched strongly with 16S amplicon abundance measures taken at 94 h. (Fig. S3). Slight differences, for example, in the case of Rhodococcus and Streptomyces, likely result from the differences in the measurement types (16S DNA versus global RNA) and not shifts in the absolute abundance of organisms at 94 h compared to 70 or 118 h. Both measures of abundance revealed that MSC-2 is a community that contains multiple species at biologically significant abundance levels. We also found that there was a strong association between species that were abundant within MSC-2 when grown together on chitin and those that had large fundamental niches. Ensifer, Rhodococcus, and Streptomyces were the three most abundant species and had the largest fundamental niches. (The Rhodococcus niche size was the same as that of Variovorax.) Neorhizobium and Dyadobacter were among the least abundant and had the smallest fundamental niches. This is despite the fact that Dyadobacter is able to grow using chitin as the sole carbon source under these growth conditions and Streptomyces is not.
FIG 4
TABLE 2
| Species | % of abundance (RNA at 70 h) | Total no. of: | % of DEGsc | |
|---|---|---|---|---|
| DEGsa | Genesb | |||
| Streptomcyes | 12.111 | 48 | 7,143 | 0.67 |
| Neorhizobium | 0.866 | 128 | 7,622 | 1.68 |
| Dyadobacter | 0.053 | 0 | 6,281 | 0.00 |
| Sphingopyxis | 0.179 | 411 | 4,407 | 9.33 |
| Ensifer | 53.185 | 1937 | 7,069 | 27.40 |
| Variovorax | 1.793 | 1897 | 6,258 | 30.31 |
| Sinorhizobium | 0.841 | 895 | 5,833 | 15.34 |
| Rhodococcus | 30.972 | 3909 | 7,526 | 51.94 |
a
Total number of differentially expressed genes (see Materials and Methods).
b
Total number of genes in the genome.
c
Percentage of the genome comprised of DEGs: (no. of DEGs/no. of genes in genome) × 100.
To determine which species were most transcriptionally active, differentially expressed genes (DEGs) were identified for each species of MSC-2 (Table 2 and Fig. S4). This analysis revealed that Rhodococcus had the highest degree of differential gene expression (as measured by the percentage of DEGs in the genome, 51%). This was followed by Variovorax (~30%), Ensifer (~27%), and Sinorhizobium (~15%). Streptomyces and Neorhizobium showed the smallest changes in DEGs. This indicates that species with low relative abundances in MSC-2 (e.g., Variovorax at ~2% and Sinorhizobium at ~0.7%) could still contain a large number of DEGs and potentially contribute to community-level processes. Conversely, abundant species such as Streptomyces were shown to have minimal differential gene expression, suggesting that while they were highly abundant, they may contribute only a small number of processes to the system during growth on chitin at the time points measured. This apparent lack of contribution may be specific to the later stages of chitin growth examined here; their contributions may be more important early in the growth timeline when direct breakdown of chitin is more critical to providing C to the system.
We next examined which functional processes related to chitin metabolism were expressed by each of the species in MSC-2 (Fig. 5A), by comparing 70 h to 118 h of growth on chitin. Growth at this later time point caused an upregulation of chitin binding genes in Streptomyces, Neorhizobium, and to a more mixed degree, Rhodococcus, and an upregulation of genes encoding enzymes that are involved in conversion of shorter chitin or chitobiose polymers to NAG monomers in Sinorhizobium and Ensifer. At the same time, genes involved with the processing of NAG showed an increase in expression in Sinorhizobium, Rhodococcus, and Ensifer (Fig. 5A). Overall, when DEGs of chitinase decomposition pathways were compared over time (70 h compared to 118 h), Streptomyces and Neorhizobium contributed to carrying out processes early in the metabolism of chitin (expressing proteins that bind chitin directly or expressing known chitinases), Sinorhizobium and Ensifer contributed processing midway between chitin breakdown and NAG metabolism, and Sinorhizobium, Rhodococcus, and Ensifer contributed to later stages of conversion of NAG to even simpler C and N sources (e.g., NAG-6-P to glucosamine-6-P which leads to fructose-6-P and the glycolysis/gluconeogenesis pathway) (Fig. 5B). Transcriptomic expression of other pathways for MSC-2 members is shown in Fig. S5. These observations address the hypothesis that species in a chitin-degrading community tend to organize themselves into specific roles. Not all species express all chitin breakdown pathways, even if they possess them: rather, certain species express only certain pathways, likely to maximize both community and constituent growth.
FIG 5
Overall, these analyses revealed that a species abundance is tightly linked to fundamental niche size, that multiple species may be transcriptionally active, despite low abundances, and that certain other species with high abundances can have low transcriptional activities. The most active species in MSC-2 were Rhodococcus, Variovorax, and Ensifer, and the major species likely contributing to initial, energy-intensive processes in chitin degradation were Neorhizobium and Streptomyces (Fig. 5). Despite the specific regulation of chitin processes in Streptomyces, other pathways in this organism showed very little transcriptional regulation at the time points sampled. It is also of note that we did not find Streptomyces to be able to grow on chitin in monoculture under the conditions of our experiment. Interactions with other species of MSC-2 in a community may thus contribute to growth of Streptomyces in coculture compared to axenically. This observation helps to answer another of our major hypotheses: that the phenotype of species (i.e., whether it acts as a degrader of chitin and can grow using chitin as the sole carbon source) is strongly influenced by surrounding community members. The chitin growth phenotype is not expressed by Streptomyces in monoculture but is expressed in coculture with other MSC-2 species.
Metabolomic analysis of MSC-2 under chitin conditions.
To gain a better understanding of downstream products of chitin breakdown in this community intracellular metabolomic analysis was carried out at the same time points as the metatranscriptomic analysis (70 and 118 h after chitin addition). Metabolic profiles of the communities shifted over time (Fig. S6 and Tables S3 and S4). In total, 14 metabolites were significantly different between the two time points, with most showing increased abundances at 118 h compared to 70 h (Fig. 6). Trehalose, l-pyroglutamic acid, d-ribose, and NAG were among the identified metabolites showing increased abundances. Several other metabolites also showed increased abundance, but their identity could not be confirmed due to the presence of only a mass spectral match and not an accompanying retention time or index match. However, provisional annotation of these unknowns was possible, and they included N-acetyl-d-galactosaminitol (unknown 105), N-acetyl-d-galactosamine (unknown 151), and a disaccharide (unknown 122). NAG, and related annotated molecules, were of particular interest as they are some of the main breakdown products of chitin, and in fact NAG and N-acetyl-d-galactosaminitol showed the largest fold changes (6.55- and 6.64-fold higher at 118 h versus 70 h, respectively). Among metabolites showing decreased abundances at 118 h were an unknown metabolite (likely hydroxylamine) and putrescine. Decreases in putrescine may be related to this molecule’s link to chitin breakdown (31). At higher chitin levels (70 h), chitin is being degraded and putrescine levels are high. Later in the growth curve (118 h), chitin levels are lower, chitin breakdown slows, and putrescine levels fall. Fold changes for all significantly shifted metabolites are shown in Table S3. Raw data for all detected metabolites (both known and unknown) are shown in Table S4.
FIG 6
Integration of metatranscriptomic and metabolomic data to infer community processes.
The collection of multiple omics data sets during growth of MSC-2 on chitin allowed us to take an integrated approach for downstream analysis using Metabolite-Expression-Metabolic Network Integration for Pathway Identification and Selection (MEMPIS), a tool that combines metabolic and metatranscriptomic data to determine which functions a community is expressing (32, 33) (Fig. 7A and B). The integration of metatranscriptomic data and metabolomic data also provided a more robust way to identify expressed processes. Previous work has pointed out the sometimes spurious link between expressed genes and the functions they encode (34); our incorporation of metabolomic data will help alleviate that issue and improve accuracy. This analysis revealed that fatty acid metabolism and amino acid metabolism had the most coverage, indicating that a wide range of processes in these categories had the potential to be activated in the MSC-2 system. Regarding processes likely related to chitin breakdown, the amino sugar metabolism pathway also had many expressed metabolites and genes, but the coverage was small, suggesting that only certain parts of this broad pathway were activated. In contrast, N metabolism had few genes and metabolites being expressed, but the overall coverage of this pathway was relatively high, showing that much of the known pathway was being expressed.
FIG 7
The metatranscriptome was then screened to determine which species were expressing which genes (Fig. 7C and D). Variovorax expressed almost all the detected genes involved in fatty acid metabolism, with smaller contributions from Rhodococcus. Variovorax was also the major driver of pyruvate metabolism. Rhodococcus carried out purine metabolism, tryptophan metabolism, and phosphate processes. Ensifer participated in fewer processes (glyoxylate, pyrimidine, and alanine metabolism) than Variovorax and Rhodococcus, despite Ensifer being the major constituent of MSC-2 when measured in both DNA and RNA data sets. Looking specifically at processes centering on chitin and NAG metabolism, Variovorax was the main contributor to N metabolism and Ensifer to amino sugar metabolism.
In some cases, MSC-2 species had outsized roles in certain processes compared to their genomic potential for such processes (Fig. 7C and D). The large role of Ensifer in glyoxylate metabolism contrasts with the fact that Variovorax had much more genomic potential to participate in these processes but did not in this experiment, at least as measured at the transcriptional level. Similarly, Variovorax had the greatest gene expression related to starch metabolism, despite having less genomic potential for that process than Rhodococcus and Streptomyces. Likewise, Rhodococcus’ expression of genes associated with alanine metabolism occurred despite it having less genomic potential for that process than Neorhizobium, Ensifer, Variovorax, and Sinorhizobium. All of this suggests that species occupy unique niches within MSC-2 based not on each member’s genomic potential but on their expression of specific gene sets designed to maximize their growth and/or survival as they interact with other species either cooperatively or competitively.
DISCUSSION
The native soil microbiome is a complex community consisting of thousands of species with potentially millions of interactions between them. Because no species exists in isolation, it is these interactions that lead to the emergent properties of the soil microbiome and its role in cycling C and N substrates. Here, we sought to better understand organismal phenotypes, the roles of individual species, and how this is affected by fellow community members. By focusing on a well-defined, simplified consortium (MSC-2) that was derived from a naturally evolved soil consortium (MSC-1) (26), we were able to investigate species-specific contributions to community metabolism, using the decomposition of chitin as our study system.
One of our major findings is that within MSC-2, chitin degradation potential (as measured by the ability to grow with chitin as the sole C source in monoculture) is not the sole or even the major determinant of how abundant (as measured by metatranscriptomic analysis and evaluating how many reads align to a given species) a member will be in the community when grown on a complex carbon source like chitin. Analysis of the fundamental and realized niches of MSC-2 members as well as the growth advantages and disadvantages of member species in a community versus axenic growth showed that ranking the species from largest to smallest fundamental niche (Fig. 2C) aligned very well with ranking by the abundances of each species (Fig. 4 and see Fig. S3 in the supplemental material). For some species, we noticed a lack of abundance in MSC-2, even though the ability of these species to grow on chitin had been shown. The trade-off between the ability to grow on chitin in monoculture and abundance in the MSC-2 community is especially striking for Sinorhizobium. Although this species can grow on chitin and expresses chitin binding genes, its relative abundance in MSC-2 was low. This contrast between chitin growth ability and community abundance (during growth on chitin) may be because the fundamental niche of Sinorhizobium is small compared to those of other species such as Ensifer. As a result of Ensifer’s larger niche, while it may not drive chitin degradation, it may be able to grow on breakdown products provided by initial chitin degraders such as Sinorhizobium. In contrast, Sinorhizobium, with its smaller fundamental niche, is not able to take great advantage of the range of chitin breakdown products and is limited to focusing on chitin itself, which is constantly falling in concentration as the experiment proceeds and is difficult to metabolize, all of which may lead to less growth of Sinorhizobium.
The observation that microbes with large fundamental niches may survive off chitin breakdown products opens a new question: what do these secondary consumers provide to the community in exchange for receiving breakdown products? Several possibilities emerge. First, secondary consumers may survive off C and N sources that degraders are generating but cannot metabolize themselves as has been seen with phenylalanine and Escherichia coli (35, 36). Alternatively, consumption of metabolites by secondary consumers may relieve stoichiometric pressure on critical enzymatic reactions of degraders, allowing for more continued degradation of chitin (22). This may especially be the case for Streptomyces as high levels of NAG have been shown to repress chitin breakdown in this organism (17). Another possibility is that secondary consumers may produce important vitamin cofactors such as cobalamin. While this vitamin is critical to the soil microbiome, only a small number of bacteria and archaea produce B12 (23), suggesting that these B12 producers are important to community growth even if they do not provide C or N. Our experiments here included excess B12, but this may be the case for other vitamins or secondary metabolites that are limiting. For example, the major chitin degraders Neorhizobium and Streptomyces have importers for both branched-chain amino acids and xylose molecules (Fig. 3). Gaining these from nondegraders would be energetically favorable for these species. It is important to note that these fundamental niche conclusions are possible only after defining this soil-derived consortium and building metabolic models for each individual species. Such a detailed understanding of a soil microbial consortium will help to answer several questions beyond what we show here and will allow us to begin to bridge the gap between the molecular details possible with laboratory experiments and the more complex, but translational, results from direct field analysis.
Our investigations of MSC-2 provide a deeper understanding of how species that cannot grow on complex organic carbon axenically may alter their phenotypes and contribute to decomposition in the context of other community members. For example, under the growth conditions of this experiment Streptomyces was not able to grow in monoculture on chitin as the sole C source. This result was consistent for each of the three Streptomyces strains that we isolated (Fig. 1B). This lack of growth in monoculture combined with the observed high abundance of Streptomyces within MSC-2 and our metatranscriptomic analysis suggests that Streptomyces contributes directly toward chitin decomposition in MSC-2, but the energy limitations presented by monoculture growth on chitin prohibit enzyme production. Moving a species from an environment with labile C sources (our initial growth on R2A medium) to an environment with chitin as the only C substrate may not allow Streptomyces, when grown in monoculture, to generate chitin-degrading enzymes. Streptomyces may then be “stuck” with abundant chitin but not enough nutrients to allow for the initial synthesis of chitinases. We designed the experiment to account for this possibility by including a 3-day period of growth in chitin and NAG to allow for species to acclimate from richer medium (R2A) to chitin-only minimal medium. While this may have been useful for some organisms, it may have had the opposite effect on Streptomyces as chitinase production in Streptomyces species has been shown to be repressed by the presence of NAG (17). However, in contrast to monoculture, in the context of the complete MSC-2 community, other species can initially degrade chitin, provide breakdown products, allow Streptomyces to grow and make its own chitin-degrading enzymes, and contribute to chitin degradation later in the experimental timeline.
These conclusions emphasize the differences between inherent phenotypes of microorganisms in isolation (e.g., no growth of Streptomyces on chitin in monoculture) versus the phenotypes expressed when grown as a community (e.g., growth of Streptomyces on chitin in a community). There is an extension of this observation that is more closely related to soil ecology: that phenotypes of a species are dependent on the nature of the surrounding community. As communities are constantly shifting in soil as a function of environmental changes, testing this hypothesis further is critical for a deeper understanding of the community interactions that are at the foundation of soil microbial ecology. Emergent properties are often thought of as functions emerging from the community as a whole (37). However, species within a community also show phenotypic properties that emerge only when they are grown with certain other species. In fact, emergent properties of individual species are likely what leads directly to emergent properties of the community as a whole. These results also emphasize the importance of examining species in both the context of other microorganisms and axenically (where possible) to fully understand their contributions to community metabolism and growth. Incorporation of species’ actions during monoculture versus community growth is also critical for applying what is learned with model communities back to the field, where no species exist in isolation and community growth is the rule. While our development of MSC-2 will help bridge the gap between laboratory and field, it should be noted that this is a community of relatively few isolates that are combined from individual strains, so the discoveries learned here may not represent the phenotypes expressed in the field. However, two aspects of our studies suggest that MSC-2 is a valuable tool for understanding soil microbial ecology. First, while MSC-2 is comprised of isolates, these isolates were drawn from a naturally evolved community derived from the complete soil microbiome (MSC-1) (26). Members of MSC-1 (and therefore MSC-2) were chosen not by us by rather by existing interactions of soil microbiology that drove community assembly. Therefore, constituents of MSC-2 are likely to participate in interactions that are, in part, representative of processes of the native soil microbiome. Second, our use of MEMPIS in these studies allows us to apply the same methods and the same omics measurements between the lab (MSC-2) and the field (native soil microbiome). This will allow us to map processes happening in both systems and highlight where MSC-2 can be used to better understand and interpret native soil microbiome processes and where it cannot.
In summary, we describe a model chitin-degrading consortium and use multiomics analysis as well as axenic versus coculture studies to delineate the role of individual species within this consortium. Chitin-degrading members likely supported non-chitin degraders through sharing of chitin breakdown products. Because of this sharing, the main driver of a species growth and success was not the ability to carry out specific metabolic processes related to abundant C sources (chitin breakdown), but rather which species could take the most advantage of shared breakdown products. Also related to community growth, we show that the phenotype of a species, regarding its role as a primary decomposer or not, is driven by the surrounding community. The data and conclusions gathered here will be instrumental in guiding future experiments focused on combining individual constituents of MSC-2 in pairwise combinations and performing experiments in which members of MSC-2 are left out. These conclusions will also be critical to our understanding of how native soil microbiomes process soil organic C, especially substrates such as chitin that drive interactions and metabolite sharing, and how these processes may shift as community membership changes as a function of both biotic and abiotic pressures. Application of these conclusions to the native soil microbiome will greatly expand our ability to identify what the keystone species are, which may have the greatest advantage for growth, and how these communities are organized to promote C cycling in natural settings.
MATERIALS AND METHODS
Isolation and analysis of new constituents of MSC-2.
These experiments build on a previously described field-derived, reproducible consortium, MSC-1. MSC-1 is a chitin-enriched consortium containing ~35 species for which we have identified keystone species and a preliminary interaction network. To test how individual members of MSC-1 are contributing to community metabolism and growth, we developed a more-reduced-complexity consortium, MSC-2, as a subset of MSC-1. Isolates of MSC-1 for use in MSC-2 were collected as described previously (26) through plating on R2A agar medium followed by selection of colonies with morphological differences and 16S rRNA amplicon sequencing to determine identity down to the genus level. For genome assembly of isolates, DNA was extracted using the ZymoBIOMICS DNA microprep kit (Zymo, Irvine, CA). DNA was then sequenced using Illumina technology, collecting ~630 Mbp per genome (see Table S1 in the supplemental material). Reads were quality trimmed using bbduk (part of the bbtools package (https://sourceforge.net/projects/bbmap/) and the following parameters: int ow k = 27 hdist = 1 qtrim=f ref=sequencing_artifacts.fa.gz, phix174_ill.ref.fa.gz minlen = 35. The SPAdes assembler (v.3.13.1) (38) was then used to assemble reads with the following parameters: -k 21,33,55,77,99,127 –careful. Assembled genomes were analyzed using the DRAM (Distilled and Refined Annotation of Metabolism) pipeline in the reduced memory mode (–skip_uniref) (39). As part of the DRAM pipeline, KEGG Orthology (KO) numbers were assigned using the KOfam database, which were subsequently used to visualize pathways using the KEGG mapper tool (40, 41). The DRAM pipeline additionally provides broad metabolic categories for each annotated gene. Using R and the associated ggplot2 R package (https://ggplot2.tidyverse.org/), these categories were parsed across each genome and a heat map of functional potential was generated. Taxonomy was assigned using the standard classification workflow (classify_wf) of the Genome Taxonomy Database toolkit (GTDBtk) v.1.4.1, and GTDB reference data v.r95 (42). Because the final species-level taxonomy may shift as the GTDB reference data are updated, the members of MSC-2 are referred to here primarily by their genera. The eight isolates comprising MSC-2 are shown in Table 1, were chosen to maximize genomic diversity, and include both chitin degraders and nondegraders.
Cultivation of MSC-2 species individually under variable C and N sources.
Each MSC-2 strain was cultivated on R2A agar plates for 3 days prior to starting a growth assay for a particular C source. Multiple C sources were tested to generate fundamental niche profiles for each species. The resulting colonies on R2A were collected and washed three times in M9 (containing no C source and 18 mM NH4Cl) with centrifugation (5,000 rpm for 5 min) between each wash. Each strain was then resuspended in M9 with 18 mM NH4Cl and the C source under analysis to a final OD600 of 0.1. For growth with chitin experiments, colloidal chitin from shrimp shells (Megazyme, Bray, Ireland; product code P-CHITN) was used as the sole C source at a final concentration of 100 ppm (0.1 mg/mL). This amount was chosen so that insoluble chitin interfered as little as possible with OD readings. For all other C sources (glucose, C6H12O6; NAG, C8H15NO6; serine, C3H7NO3; glycine, C2H5NO2; alanine, C3H7NO2; maltose, C12H22O11; xylose, C5H10O5; glutamate, C5H9NO4; sucrose, C12H22O11; fructose, C6H12O6; and arabinose, C5H10O5), the final concentration used was 10 mM. We standardized by molarity of substrates, rather than the mass of C, because our primary goal was comparing growth among species on the same C source, not necessarily growth among C sources for each species; the latter was addressed via metabolic modeling (described below). Following inoculation with the C source, each strain was grown in biological triplicates in a plate reader with OD measurements collected every 20 min for 5 days. Each experiment contained a no-carbon control, with the OD of these wells being used to correct OD values in growth wells. Plates were shaken at 180 rpm during growth, and cultivation was carried out at 20°C.
Metabolic modeling of MSC-2 species.
We developed metabolic models of the eight isolates using the Department of Energy Systems Biology Knowledgebase (KBase) modeling pipeline (27). We first constructed drafted models using the “Build Metabolic Model” app, which were iteratively gapfilled using the “Gapfill Metabolic Model” app to match predicted biomass production with experimentally observed growth phenotypes in M9 minimal media. The resulting isolate models were highly consistent with experimental data, correctly predicting 95% of growth and nongrowth observations. With these metabolic models, we performed flux balance analysis (FBA) (43) to predict biomass yields (grams dry weight per millimole of C source) from various carbon sources.
A mixed-bag metabolic network was constructed by merging the eight isolate models without accounting for species boundaries (44). Using comparative growth simulations with isolates and mixed-bag models, we determined fundamental and realized niches of MSC-2 species. To do so, we first identified C sources that are included in the uptake reactions of the mixed-bag model. For each of those C sources, we subsequently performed FBA using isolate and mixed-bag models to determine whether individual species and the community utilize them for growth. All C sources that support biomass production in individual isolate models were recorded as part of an isolate’s fundamental niche; C sources that the mixed-bag model can utilize for growth were recorded as part of the realized niche, which is common for all isolates.
Cultivation of MSC-2 community with chitin.
Each MSC-2 strain was cultivated on R2A agar plates for 3 days. Each strain was resuspended in M9 liquid medium (all M9 media in these experiments contained 18 mM NH4Cl) containing 100 ppm chitin and 10 mM NAG to a final OD of 0.1. After 3 days of growth, cells were collected by centrifugation and washed three times in M9 with centrifugation between washes. Strains were then resuspended to a final OD of 1.0 in M9. Three hundred fifty microliters of each strain was added to a single 100-mL flask containing 30 mL of M9 and 500 ppm (0.5 mg/mL) chitin. Five biological replicate flasks, labeled A to E, were started and incubated at 20°C for 5 days. A larger amount of chitin was used for coculture growth as optical density data did not need to be gathered.
For DNA samples, three biological replicates (from flasks A to C) of 1-mL samples were collected at 0, 8, 16, 34, 46, 70, 94, and 118 h after the start of the experiment. Sample time points were chosen based on visual confirmation of growth of MSC-2 community and growth kinetics of MSC-2 constituents (Fig. 1B). Samples were centrifuged as described above, the supernatant was discarded, and the pellet was flash frozen in liquid N2 and stored at −80°C. For RNA, three biological replicates (from flasks A to C) of 2-mL samples were collected at 0, 34, 70, and 118 h after the start of the experiment. Samples were stored as for the DNA samples. For metabolites, five biological replicates (from flasks A to E) were collected at 70 and 118 h after the start of the experiment. Samples were stored as for the DNA samples. These later time points were chosen for omics analysis as we predicted that chitin breakdown would need several days to reach full efficiency and because biomass levels were higher at these later time points, which led to better metatranscriptomic and metabolic analyses. DNA was extracted using the ZymoBIOMICS DNA microprep kit. RNA was extracted using the ZymoBIOMICS RNA microprep kit with on-column DNase treatment. Metabolites from the cell pellets were extracted using the MPLEx extraction protocol as described previously (45).
Multiomic analysis of MSC-2 during growth on chitin.
Metatranscriptomic analysis was carried out on the three replicates of extracted RNA at 70 and 118 h using Illumina technology and paired-end reads of 300 bp in length. This number of replicates was chosen to maximize our statistical analysis within the resources available for these experiments. As there were differences in the growth rates of MSC-2 species and certain members did not reach their highest densities until up to 109 h after the start of the experiment (based on our monoculture growth data) (Table 1), these time points were chosen so that all MSC-2 members would have a chance to grow to optimal levels prior to multiomic analysis. Total reads and alignment values from each sample are shown in Table S2. Reads were aligned to a concatenated metagenome of each of the eight species individual genomes using the Burrows Wheeler Aligner (46). The resulting SAM files were then converted to raw count files using HTSeq (47). Transcripts were normalized using the DESeq2 package within the statistics program R (48), which was also used to identify differentially expressed genes (DEGs), which were defined as genes with an adjusted P value of ≤0.05 and absolute fold change value of >2 comparing 70 h to 118 h. Normalization was applied to each species separately so that abundance changes would not be misinterpreted as gene expression changes. We also analyzed 16S rRNA gene amplicons on two replicates of extracted DNA at 94 h using Illumina technology and proprietary primers that cover the V3 and V4 hypervariable regions with a total read length of 500 bp. A single replicate was used here since we sought mainly to gain a confirmation of abundance levels as determined by our metatranscriptomic levels above. This was also why we chose a midpoint of 94 h between our metatranscriptomic time points of 70 and 118 h. Total reads from each sample are shown in Table S2. 16S rRNA amplicon sequences were analyzed using QIIME2 (v.2021.4) (49), and the SILVA database (v.138) (50) was used during taxonomy assignment (q2-feature-classifer).
Metabolomic analysis of MSC-2 was carried out by gas chromatography-mass spectrometry (GC-MS) analysis as described previously (51), with the samples blocked and analyzed in random order for each experiment. GC-MS raw data file processing was done using Metabolite Detector software, and metabolites were identified by matching experimental spectra and retention indices to an augmented version of FiehnLib (52). The NIST 14 GC–MS library was also used to cross-validate the spectral matching scores obtained using the Agilent library and to provide identifications of unmatched metabolites. As metabolomic analysis is sometimes imperfect due to noise in the signal (53), we collected five biological replicates for these experiments to increase our statistical power and confidence.
Metabolite-Expression-Metabolic Network Integration for Pathway Identification and Selection (MEMPIS) (32) was used to identify specific reaction pathways and associated taxon-specific genes during chitin decomposition. We integrated transcriptomic and metabolomic data measured from MSC-2 using MEMPIS with the mixed-bag metabolic model reconstructed in KBase. The MEMPIS algorithm was implemented in Python using Pyomo (54), a Python-based open-source software package to model optimization problems and pass the problem to solvers on the NEOS server (55). COBRApy (56) was used to read a systems biology markup language (SBML) file for loading the mixed-bag model and to determine the reactions associated with these overexpressed genes using gene-protein-reaction (GPR) rules defined in the mixed-bag model. ModelSEED identifiers were used to map metabolites and these reactions to the mixed-bag model. We used CPLEX v20.1.0.0 on the NEOS server to solve the minimization problem in MEMPIS.
Data availability.
All sequencing data has been uploaded onto DataHub. Genome accession numbers for MSC-2 species in GenBank are uploaded at: https://data.pnnl.gov/group/nodes/dataset/33234. 16s amplicon data is uploaded at: https://data.pnnl.gov/group/nodes/dataset/33231. Metatranscriptomic data is uploaded at: https://data.pnnl.gov/group/nodes/dataset/33232.
ACKNOWLEDGMENTS
This research was supported by the Department of Energy’s Office of Biological and Environmental Research (BER) and is a contribution of the Scientific Focus Area “Phenotypic response of the soil microbiome to environmental perturbations” (FWP 70880). PNNL is operated for the DOE by Battelle Memorial Institute under contract DE-AC05-76RLO1830. A portion of this work was performed in the William R. Wiley Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by Office of Biological and Environmental Research and located at PNNL.
We declare no conflict of interest.
Supplemental Material
File (msystems.00372-22-s0001.pdf)
- Download
- 271.13 KB
File (msystems.00372-22-s0002.pdf)
- Download
- 58.37 KB
File (msystems.00372-22-s0003.pdf)
- Download
- 33.09 KB
File (msystems.00372-22-s0004.pdf)
- Download
- 135.86 KB
File (msystems.00372-22-s0005.pdf)
- Download
- 66.76 KB
File (msystems.00372-22-s0006.pdf)
- Download
- 30.67 KB
File (msystems.00372-22-s0007.pdf)
- Download
- 128.46 KB
File (msystems.00372-22-s0008.pdf)
- Download
- 86.57 KB
File (msystems.00372-22-s0009.pdf)
- Download
- 118.25 KB
File (msystems.00372-22-s0010.pdf)
- Download
- 404.91 KB
REFERENCES
1.
Deng S, Wipf HM, Pierroz G, Raab TK, Khanna R, Coleman-Derr D. 2019. A plant growth-promoting microbial soil amendment dynamically alters the strawberry root bacterial microbiome. Sci Rep 9:17677.
2.
Tian L, Lin X, Tian J, Ji L, Chen Y, Tran LP, Tian C. 2020. Research advances of beneficial microbiota associated with crop plants. Int J Mol Sci 21:1792.
3.
Scarlett K, Denman S, Clark DR, Forster J, Vanguelova E, Brown N, Whitby C. 2021. Relationships between nitrogen cycling microbial community abundance and composition reveal the indirect effect of soil pH on oak decline. ISME J 15:623–635.
4.
Wilpiszeski RL, Aufrecht JA, Retterer ST, Sullivan MB, Graham DE, Pierce EM, Zablocki OD, Palumbo AV, Elias DA. 2019. Soil aggregate microbial communities: towards understanding microbiome interactions at biologically relevant scales. Appl Environ Microbiol 85:e00324-19.
5.
Yuan MM, Guo X, Wu L, Zhang Y, Xiao N, Ning D, Shi Z, Zhou X, Wu L, Yang Y, Tiedje JM, Zhou J. 2021. Climate warming enhances microbial network complexity and stability. Nat Clim Chang 11:343–348.
6.
Michalska-Smith M, Song Z, Spawn-Lee SA, Hansen ZA, Johnson M, May G, Borer ET, Seabloom EW, Kinkel LL. 2022. Network structure of resource use and niche overlap within the endophytic microbiome. ISME J 16:435–446.
7.
Park H, Patel A, Hunt KA, Henson MA, Carlson RP. 2020. Artificial consortium demonstrates emergent properties of enhanced cellulosic-sugar degradation and biofuel synthesis. NPJ Biofilms Microbiomes 6:59.
8.
Diffenbaugh NS, Swain DL, Touma D. 2015. Anthropogenic warming has increased drought risk in California. Proc Natl Acad Sci USA 112:3931–3936.
9.
Ganguly AR, Steinhaeuser K, Erickson DJ, III, Branstetter M, Parish ES, Singh N, Drake JB, Buja L. 2009. Higher trends but larger uncertainty and geographic variability in 21st century temperature and heat waves. Proc Natl Acad Sci USA 106:15555–15559.
10.
Smercina DN, Bailey VL, Hofmockel KS. 2021. Micro on a macroscale: relating microbial-scale soil processes to global ecosystem function. FEMS Microbiol Ecol 97:fiab091.
11.
Beaton D, Pelletier P, Goulet RR. 2019. Microbial degradation of cellulosic material and gas generation: implications for the management of low-and intermediate-level radioactive waste. Front Microbiol 10:204.
12.
López-Mondéjar R, Zühlke D, Becher D, Riedel K, Baldrian P. 2016. Cellulose and hemicellulose decomposition by forest soil bacteria proceeds by the action of structurally variable enzymatic systems. Sci Rep 6:25279.
13.
Medie FM, Davies GJ, Drancourt M, Henrissat B. 2012. Genome analyses highlight the different biological roles of cellulases. Nat Rev Microbiol 10:227–234.
14.
Hui C, Jiang H, Liu B, Wei R, Zhang Y, Zhang Q, Liang Y, Zhao Y. 2020. Chitin degradation and the temporary response of bacterial chitinolytic communities to chitin amendment in soil under different fertilization regimes. Sci Total Environ 705:136003.
15.
Starke R, Morais D, Větrovský T, López Mondéjar R, Baldrian P, Brabcová V. 2020. Feeding on fungi: genomic and proteomic analysis of the enzymatic machinery of bacteria decomposing fungal biomass. Environ Microbiol 22:4604–4619.
16.
Beier S, Bertilsson S. 2013. Bacterial chitin degradation-mechanisms and ecophysiological strategies. Front Microbiol 4:149.
17.
Miyashita K, Fujii T, Saito A. 2000. Induction and repression of a Streptomyces lividans chitinase gene promoter in response to various carbon sources. Biosci Biotechnol Biochem 64:39–43.
18.
Baty AM, III, Eastburn CC, Diwu Z, Techkarnjanaruk S, Goodman AE, Geesey GG. 2000. Differentiation of chitinase-active and non-chitinase-active subpopulations of a marine bacterium during chitin degradation. Appl Environ Microbiol 66:3566–3573.
19.
Baty AM, III, Eastburn CC, Techkarnjanaruk S, Goodman AE, Geesey GG. 2000. Spatial and temporal variations in chitinolytic gene expression and bacterial biomass production during chitin degradation. Appl Environ Microbiol 66:3574–3585.
20.
Gooday GW. 1990. The ecology of chitin degradation, p 387–430. In Marshall KC (ed), Advances in microbial ecology, vol 11. Plenum Press, New York, NY.
21.
Keyhani NO, Roseman S. 1997. Wild-type Escherichia coli grows on the chitin disaccharide, N,N′-diacetylchitobiose, by expressing the cel operon. Proc Natl Acad Sci USA 94:14367–14371.
22.
Pacheco AR, Moel M, Segrè D. 2019. Costless metabolic secretions as drivers of interspecies interactions in microbial ecosystems. Nat Commun 10:103–112.
23.
Lu X, Heal KR, Ingalls AE, Doxey AC, Neufeld JD. 2020. Metagenomic and chemical characterization of soil cobalamin production. ISME J 14:53–66.
24.
Oña L, Giri S, Avermann N, Kreienbaum M, Thormann KM, Kost C. 2021. Obligate cross-feeding expands the metabolic niche of bacteria. Nat Ecol Evol 5:1224–1232.
25.
Poulsen JS, de Jonge N, Macêdo WV, Dalby FR, Feilberg A, Nielsen JL. 2022. Characterisation of cellulose-degrading organisms in an anaerobic digester. Bioresour Technol 351:126933.
26.
McClure R, Naylor D, Farris Y, Davison M, Fansler SJ, Hofmockel KS, Jansson JK. 2020. Development and analysis of a stable, reduced complexity model soil microbiome. Front Microbiol 11:1987.
27.
Arkin AP, Cottingham RW, Henry CS, Harris NL, Stevens RL, Maslov S, Dehal P, Ware D, Perez F, Canon S, Sneddon MW, Henderson ML, Riehl WJ, Murphy-Olson D, Chan SY, Kamimura RT, Kumari S, Drake MM, Brettin TS, Glass EM, Chivian D, Gunter D, Weston DJ, Allen BH, Baumohl J, Best AA, Bowen B, Brenner SE, Bun CC, Chandonia J-M, Chia J-M, Colasanti R, Conrad N, Davis JJ, Davison BH, DeJongh M, Devoid S, Dietrich E, Dubchak I, Edirisinghe JN, Fang G, Faria JP, Frybarger PM, Gerlach W, Gerstein M, Greiner A, Gurtowski J, Haun HL, He F, Jain R, et al. 2018. KBase: the United States Department of Energy Systems Biology Knowledgebase. Nat Biotechnol 36:566–569.
28.
Goldman J, Dennett M. 1991. Ammonium regeneration and carbon utilization by marine bacteria grown on mixed substrates. Mar Biol 109:369–378.
29.
Wang J, Yan D, Dixon R, Wang Y-P. 2016. Deciphering the principles of bacterial nitrogen dietary preferences: a strategy for nutrient containment. mBio 7:e00792-16.
30.
Connell TD, Metzger DJ, Lynch J, Folster JP. 1998. Endochitinase is transported to the extracellular milieu by the eps-encoded general secretory pathway of Vibrio cholerae. J Bacteriol 180:5591–5600.
31.
Dai Y, Yan Z, Jia L, Zhang S, Gao L, Wei X, Mei Z, Liu X. 2016. The composition, localization and function of low-temperature-adapted microbial communities involved in methanogenic degradations of cellulose and chitin from Qinghai-Tibetan Plateau wetland soils. J Appl Microbiol 121:163–176.
32.
Roy Chowdhury T, Lee J-Y, Bottos EM, Brislawn CJ, White RA, Bramer LM, Brown J, Zucker JD, Kim Y-M, Jumpponen A, Rice CW, Fansler SJ, Metz TO, McCue LA, Callister SJ, Song H-S, Jansson JK. 2019. Metaphenomic responses of a native prairie soil microbiome to moisture perturbations. mSystems 4:e00061-19.
33.
McClure RS, Lee J-Y, Chowdhury TR, Bottos EM, White RA, Kim Y-M, Nicora CD, Metz TO, Hofmockel KS, Jansson JK, Song H-S. 2020. Integrated network modeling approach defines key metabolic responses of soil microbiomes to perturbations. Sci Rep 10:10882.
34.
Rocca JD, Hall EK, Lennon JT, Evans SE, Waldrop MP, Cotner JB, Nemergut DR, Graham EB, Wallenstein MD. 2015. Relationships between protein-encoding gene abundance and corresponding process are commonly assumed yet rarely observed. ISME J 9:1693–1699.
35.
Zampieri M, Hörl M, Hotz F, Müller NF, Sauer U. 2019. Regulatory mechanisms underlying coordination of amino acid and glucose catabolism in Escherichia coli. Nat Commun 10:3354.
36.
Gralka M, Szabo R, Stocker R, Cordero OX. 2020. Trophic interactions and the drivers of microbial community assembly. Curr Biol 30:R1176–R1188.
37.
Osburn ED, Badgley BD, Strahm BD, Aylward FO, Barrett J. 2021. Emergent properties of microbial communities drive accelerated biogeochemical cycling in disturbed temperate forests. Ecology 102:e03553.
38.
Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol 19:455–477.
39.
Shaffer M, Borton MA, McGivern BB, Zayed AA, La Rosa SL, Solden LM, Liu P, Narrowe AB, Rodríguez-Ramos J, Bolduc B, Gazitúa MC, Daly RA, Smith GJ, Vik DR, Pope PB, Sullivan MB, Roux S, Wrighton KC. 2020. DRAM for distilling microbial metabolism to automate the curation of microbiome function. Nucleic Acids Res 48:8883–8900.
40.
Aramaki T, Blanc-Mathieu R, Endo H, Ohkubo K, Kanehisa M, Goto S, Ogata H. 2020. KofamKOALA: KEGG ortholog assignment based on profile HMM and adaptive score threshold. Bioinformatics 36:2251–2252.
41.
Kanehisa M, Goto S. 2000. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res 28:27–30.
42.
Chaumeil P-A, Mussig AJ, Hugenholtz P, Parks DH. 2020. GTDB-Tk: a toolkit to classify genomes with the Genome Taxonomy Database. Oxford University Press, Oxford, United Kingdom.
43.
Orth JD, Thiele I, Palsson BØ. 2010. What is flux balance analysis? Nat Biotechnol 28:245–248.
44.
Henry CS, Bernstein HC, Weisenhorn P, Taylor RC, Lee JY, Zucker J, Song HS. 2016. Microbial community metabolic modeling: a community data-driven network reconstruction. J Cell Physiol 231:2339–2345.
45.
Nakayasu ES, Nicora CD, Sims AC, Burnum-Johnson KE, Kim Y-M, Kyle JE, Matzke MM, Shukla AK, Chu RK, Schepmoes AA, Jacobs JM, Baric RS, Webb-Robertson B-J, Smith RD, Metz TO. 2016. MPLEx: a robust and universal protocol for single-sample integrative proteomic, metabolomic, and lipidomic analyses. mSystems 1:e00043-16.
46.
Li H, Durbin R. 2010. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26:589–595.
47.
Anders S, Pyl PT, Huber W. 2015. HTSeq—a Python framework to work with high-throughput sequencing data. Bioinformatics 31:166–169.
48.
Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15:550.
49.
Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, Alexander H, Alm EJ, Arumugam M, Asnicar F, Bai Y, Bisanz JE, Bittinger K, Brejnrod A, Brislawn CJ, Brown CT, Callahan BJ, Caraballo-Rodríguez AM, Chase J, Cope EK, Da Silva R, Diener C, Dorrestein PC, Douglas GM, Durall DM, Duvallet C, Edwardson CF, Ernst M, Estaki M, Fouquier J, Gauglitz JM, Gibbons SM, Gibson DL, Gonzalez A, Gorlick K, Guo J, Hillmann B, Holmes S, Holste H, Huttenhower C, Huttley GA, Janssen S, Jarmusch AK, Jiang L, Kaehler BD, Kang KB, Keefe CR, Keim P, Kelley ST, Knights D, et al. 2019. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 37:852–857.
50.
Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO. 2013. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41:D590–D596.
51.
Cao X, Pan X, Couvillion SP, Zhang T, Tamez C, Bramer LM, White JC, Qian W-J, Thrall BD, Ng KW, Hu X, Demokritou P. 2021. Fate, cytotoxicity and cellular metabolomic impact of ingested nanoscale carbon dots using simulated digestion and a triculture small intestinal epithelial model. NanoImpact 23:100349.
52.
Kind T, Wohlgemuth G, Lee DY, Lu Y, Palazoglu M, Shahbaz S, Fiehn O. 2009. FiehnLib: mass spectral and retention index libraries for metabolomics based on quadrupole and time-of-flight gas chromatography/mass spectrometry. Anal Chem 81:10038–10048.
53.
Ghosh T, Philtron D, Zhang W, Kechris K, Ghosh D. 2021. Reproducibility of mass spectrometry based metabolomics data. BMC Bioinformatics 22:1–25.
54.
Hart WE, Watson J-P, Woodruff DL. 2011. Pyomo: modeling and solving mathematical programs in Python. Math Prog Comp 3:219–260.
55.
Gropp W, Moré JJ. 1997. Optimization environments and the NEOS server. Argonne National Laboratory, Lemont, IL.
56.
Ebrahim A, Lerman JA, Palsson BO, Hyduke DR. 2013. COBRApy: COnstraints-Based Reconstruction and Analysis for Python. BMC Syst Biol 7:74.
Information & Contributors
Information
Published In
mSystems
Volume 7 • Number 5 • 26 October 2022
eLocator: e00372-22
Editor: Jeffrey Lawrence Blanchard, University of Massachusetts Amherst
PubMed: 36154140
Copyright
© 2022 McClure et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.
History
Received: 4 May 2022
Accepted: 24 August 2022
Published online: 26 September 2022
Keywords
Contributors
Editor
Jeffrey Lawrence Blanchard
Editor
University of Massachusetts Amherst
Metrics & Citations
Metrics
Note:
- For recently published articles, the TOTAL download count will appear as zero until a new month starts.
- There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
- Citation counts come from the Crossref Cited by service.
Citations
Citation
2022.
Interaction Networks Are Driven by Community-Responsive Phenotypes in a Chitin-Degrading Consortium of Soil Microbes. mSystems
7:e00372-22.
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.