Revealing Community-Level Substrate Preferences of Soil Microbes with EcoPlates and Odin

Biolog’s EcoPlates™ provide a simple solution for probing the metabolic function of environmental communities. Odin™ instruments together with EcoPlates, hold the key to unlocking community analysis insights.

Soil microbes have important roles within their unique local ecosystems and contribute heavily to the growth of plants and animals through their innate processes, such as fixing nitrogen from the atmosphere.  However, phenotypically characterizing mixed samples from hard-to-work-with substrates such as soil and compost presents a unique challenge for microbiology researchers. Biolog’s EcoPlates™ provide a simple solution for probing the metabolic function of environmental communities through a set of substrates that have been selected to best characterize these types of samples. 

Each EcoPlate is a microplate that contains 31 substrates in triplicate, dried into the wells in a minimal media broth. A redox reporter dye is also included so that sample preparation only involves suspending a sample in buffer or water and transferring to the EcoPlate for automated incubation and kinetic optical density measurements in an Odin™ instrument (Figure 1). These plates were originally used for community analysis by Garland and Mills (1991) and have since been used in a variety of industries and research fields as an all-in-one tool for characterizing ecological microbiome communities. More recently, groups like Sofo and Riciutti have analyzed EcoPlate data using the average well color development over the whole plate as well as Shannon’s Diversity and Evenness indices which have classically been used in population studies. The advantage of these plate-wide parameters is that the researcher does not need to standardize, or even measure, the relative abundance of different species within the community being analyzed. Biolog has debuted a new Community Analysis feature within Odin software, which automates the calculation of these parameters and enables the comparison of communities from different sources or of unknown composition. Here we demonstrate the utility of the EcoPlate, and these Community Analysis features with Odin by screening soil samples from around the southeast San Francisco Bay region.

Methods

Soil samples were collected from around the southeast San Francisco Bay (including untreated, native soil from the South Bay and the East Bay, soil that had been recently treated with organic fertilizer, soil that had been treated with compost, and fresh compost) and stored at -80 °C prior to use. Samples were thawed at room temperature and 5 grams from each was suspended in 40 mL sterile deionized water. Samples were spun down at 2,000 x g for 5 minutes. For the geographical samples, 100 µL of each suspension was inoculated onto EcoPlates and transferred to Odin L for incubation and kinetic data collection at 590 and 740 nm every 20 minutes for 48 hours at 30 °C, following the standard Biolog protocol for EcoPlates. Samples used to compare soil treatment methods were inoculated onto 3 EcoPlates and data collected in the same manner. Data was analyzed using the Community Analysis feature in Odin software. Replicates were averaged, and samples were compared based on their average well color development (AWCD), Shannon’s Diversity Index (H`), Shannon’s Evenness Index (E), and maximum curve heights using a positive response cutoff of 0.25 and a pattern development threshold of 1.  The maximum curve height for each native soil sample was plotted against one another and plates were clustered using Pearson Correlation and average linkage clustering in MeV4 (Howe 2010). 

Results:

Carbon source preferences of native soil vary by locale
Determining the relative differences in metabolic flexibility is a crucial part of understanding microbiomes and how they can impact the macro-organisms we depend on. Of particular interest in this study was the variability in the functional abilities of unique microbial communities from around the San Francisco Bay area to digest different carbon substrates. Diverse substrate utilization is a marker of overall soil health and can inform decisions regarding fertilization composition and application rate, planting times and depths, waste disposal, and more (Zhang 2020). Soil samples were collected from seven locations around the southeast San Francisco Bay area including Palo Alto, Cupertino, Berryessa, Fremont, Union City, Blackhawk, and Dublin. The environmental samples were suspended in sterile water, the solids removed, and plated directly onto EcoPlates for assessment of carbon source utilization over the course of 24 hours
at 30 °C.

Each sample’s microbial community was able to grow on the EcoPlate in at least 12 of the 31 substrates and showed significant color development (metabolic activity) with a maximal OD₅₉₀ of 1.765 (Figure 2). The EcoPlate substrates can be divided up into functional guilds based on their carbon source type comprised of amines, amino acids, carbohydrates, carboxylic acids, phenols, and polymers. Interestingly, there did not appear to be a correlation between geography and which guild was most utilized. Several substrates were even utilized by samples from all locations including putrescine, arginine, serine, D-galactonic acid γ-lactone, mannitol, GlcNAc, pyruvic acid methyl ester, D-galacturonic acid, γ-amino butyric acid, and Tween 40 (Figure 2). α-keto butyric acid and 2-hyrdoxybenzoic acid were not utilized by any community samples, while threonine, erythritol, and α-cyclodextrin were barely used at all (Figure 2). Hierarchical clustering revealed that there was some relationship between substrates utilized and location around the Bay (Figure 2). The tightest grouping was between Fremont and Union City which are neighboring towns east of the Bay and these samples also clustered furthest from Palo Alto and Cupertino which are on the opposite side of the Bay. However, not all close sample locations showed similar metabolic profiles, such as Blackhawk and Dublin which are only 10 miles apart. The relationships between geography and metabolic profiles is very complex and often influenced by factors outside the scope of this study, however, this gives us a glimpse into the general soil microbial community health in each area (Zhao 2022).

Soil treatments can have a significant impact on metabolic flexibility
Once a general soil health profile is established, it can be determined whether to treat the soil to increase metabolic activity and therefore nutrient turnover and bioavailability for plants. Here we challenged native soil, pure leaf litter compost, compost-treated soil, and organic fertilizer- treated soil using the EcoPlate to determine which condition fostered the most metabolic activity. To begin this comparison, we plotted the kinetic metabolic curves for native soil vs compost (Figure 3A) and compost-treated vs fertilizer-treated soils (Figure 3B) and calculated the relative difference in maximum curve height for each substrate. Compost and compost-treated soil both showed significantly increased metabolic activity relative to that of the native soil or fertilizer-treated soil, as most of the 31 substrates showed a reproducible difference in OD of 0.4 between the two samples where the compost or compost-treatment showed higher levels of metabolic activity. For example, the native soil community did not outperform the compost in any substrates, was completely unable to utilize D-xylose (wells B2,6,10) or phenylethylamine (wells G4,8,12) and had a lag in utilization of other substrates indicating a lack of flexibility (Figure 3A). Compost treatment also increased metabolic flexibility compared to fertilization including the ability to metabolize glucose-1-phosphate and α-D-lactose. 

To make comparisons across the whole plate, we calculated three different metrics for each sample over time: average well color development (AWCD), Shannon’s Diversity Index (H`), and Shannon’s Evenness Index (E) as described by Sofo and Ricciuti (2020) all three of which are incorporated into Odin’s Community Analysis feature set. Data collected at 740 nm (biomass) is subtracted from the 590 nm (predominantly dye reduction) for each well to give a color value (c). The average of blank wells is subtracted from each well to give a corrected color value (c_i). AWCD for each sample was calculated as the average of all ci across all 93 substrate wells.

Shannon’s Diversity was calculated using p_i (c_i divided by the sum of all c_i for that sample) when the sum of all c_i is above the pattern development cutoff of 1.

Shannon’s Evenness was calculated using richness (S) which is the number of substrates whose c_i is above 0.25 set as the positive response cutoff.

Compost showed an earlier increase in AWCD relative to native soil which reflects the existing healthy community of microbes endemic to active, rich compost (Figure 4A). Native soil tends to be drier and devoid of many organic components which allow microbes to flourish. This may mean that many of the microbes have sporulated, and thus take more time to germinate than actively dividing cultures. The compost samples also showed higher overall AWCD which corresponds to more reduction in dye, and therefore more overall metabolic output over time. Shannon’s Diversity (H`) revealed that the compost community had an increased ability to rapidly metabolize a variety of carbon sources (Figure 4B). H`rose rapidly in the compost sample once the pattern development threshold was met and remained higher than the native soil sample, further demonstrating its metabolic flexibility (Figure 4B). Evenness (E) informs us of how evenly substrates are utilized, where a high E means that some substrates are preferred over others. Here we found that the native soil had a higher initial spike in E and took longer to level out at 1 where all substrates have relatively the same color formation (Figure 4C). Taken together, these data show that the microbial communities present in compost are significantly more flexible than those found in native soils sampled from the San Francisco Bay area. Furthermore, these communities can adapt rapidly to new conditions as evidenced by the early growth in nearly all positive wells.

Figure 3. Kinetic metabolic profile comparisons (OD590) of different treatments of garden soil.  EcoPlate internal replicates are automatically averaged by Odin software to generate the above composite overlay figures. These averages are then compared, and the differences displayed on each substrate for reference (pink) and test (cyan) samples. Panel A shows a comparison of pure leaf-litter compost to untreated native soil. Panel B shows a comparison of garden soils treated with organic fertilizer or leaf-litter compost. Overall compost by itself (panel A), or as a treatment to garden soil (panel B), showed a significantly higher rate of metabolic activity and earlier induction of activity indicating a more robust and flexible community.

Figure 4. Community analysis of leaf-litter compost vs native soil.  Community samples were compared using AWCD (Panel A), Shannon’s Diversity Index (Panel B), and Shannon’s Evenness Index (Panel C). The compost community (cyan) showed a significant increase in overall metabolic activity shown through the higher AWCD over the course of the assay. Additionally, the compost community had greater flexibility in carbon substrate utilization as indicated by the early rise in calculated diversity and evenness (H` and E respectively) as compared to native soil (pink).

Conclusion

Screening samples using EcoPlates can provide a profile of the overall health and variety present in the microbial community of any ecological niche. Fostering vibrant and diverse flora within the soil is critical to increasing nutrient turnover and therefore affects the health and production potential of the soil. Here we demonstrated the use of Odin’s Community Analysis features along with EcoPlates, and how they can help to reveal metabolic characteristics and trends. We were able to identify some shared and unique carbon source utilization phenotypes of communities from around the Southeast Bay area. Additionally, we confirmed that compost and compost-treated soils have much higher overall metabolic output and flexibility than the native soils treated with non-bioactive fertilizer.

References
Garland JL, Mills AL. Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level sole-carbon-source utilization. Appl Environ Microbiol. 1991 Aug;57(8):2351-9. doi: 10.1128/aem.57.8.2351-2359.1991. PMID: 16348543; PMCID: PMC183575.

Sofo, A.; Ricciuti, P. A Standardized Method for Estimating the Functional Diversity of Soil Bacterial Community by Biolog® EcoPlates™ Assay—The Case Study of a Sustainable Olive Orchard. Appl. Sci. 2019, 9, 4035. https://doi.org/10.3390/app9194035

Howe, E., Holton, K., Nair, S., Schlauch, D., Sinha, R., Quackenbush, J. (2010). MeV: MultiExperiment Viewer. In: Ochs, M., Casagrande, J., Davuluri, R. (eds) Biomedical Informatics for Cancer Research. Springer, Boston, MA. https://doi.org/10.1007/978-1-4419-5714-6_15

Zhang BH, Hong JP, Zhang Q, Jin DS, Gao CH. Contrast in soil microbial metabolic functional diversity to fertilization and crop rotation under rhizosphere and non-rhizosphere in the coal gangue landfill reclamation area of Loess Hills. PLoS One. 2020 Mar 9;15(3):e0229341. doi: 10.1371/journal.pone.0229341. PMID: 32150552; PMCID: PMC7062280.

Zhao L, Shao H, Zhang L, Panno SV, Kelly WR, Lin TY, Liu WT, Flynn TM, Berger P. Impact of salinity origin on microbial communities in saline springs within the Illinois Basin, USA. Environ Microbiol. 2022 Dec;24(12):6112-6127. doi: 10.1111/1462-2920.16241. Epub 2022 Nov 7. PMID: 36222141; PMCID: PMC10099389.