EcoFAB Technologies

EcoFAB Component Selection

A great diversity of EcoFABs can be generated in principle and the details depend on the questions they are designed to address. Hence modular EcoFAB design is very desirable such that there would be ‘core’ EcoFABs relevant for a particular native ecosystem that could be disseminated between labs and then extended as part of the efforts of individual labs. Hence, our initial goal is to define the base system(s) that could be refined or expanded by the individual scientist to their specific needs. We envision that groups of scientists would work together and develop EcoFABs to address the needs of their community.

Field Site: The native ecosystem that is serving as the reference system is perhaps one of the most critical aspects of the EcoFAB design. We envision a stable source of model soil and microbes and a long-term experimental site for validation of our findings. Importantly, the “natural” reference community not only provides the source for the soil microorganisms (and potential metazoans) but also allows direct observation of the patterns of community organization and potential interactions (plant-microbe-soil). For example, which microbes regularly co-occur with each other, what is the proportion of saprotrophs, mutualists, and pathogens in a natural soil community? What are the temporal and spatial pattern alterations of the rhizosphere microbiome?

Figure 1. Aerial photo of Russell Ranch.

This workshop identified that the UC Davis Russell Ranch Sustainable Agriculture Facility (Figure 1) is well-suited as a field site for EcoFAB development. This 300-acre research farm has an ongoing 100-year experiment measuring long-term impacts of crop rotation, farming management practices, and water and nutrient inputs in organic and conventional production systems (Kong et al., 2011). The Russell Ranch sustains continuous monitoring of environmental data through smart-sensor technology and remote sensing; maintains archived datasets of soil samples, farm operations, soil quality, crop yields, carbon sequestration, and greenhouse gases; and serves as training space for students and farmers. The Ranch houses 72 one-acre plots with rotations for different crop systems as well as native grasslands serving as control plots. Nested within each plot, there are 40 microplots, allowing researchers to perform experimental work on a short-term basis or measure soil and biological properties that are relevant at multiple scales: from small (soil aggregate, rhizosphere) to large (landscape, ecosystem) scales. The UC Davis Russell Ranch is an excellent reference system for studying plant-soil-microbe interactions important for the understanding of carbon cycling, plant growth promotion, and soil microbiomes that are ecologically relevant and sustainable.

EcoFab_Image_3
Figure 3. Illustration of using synthetic biology tools to construct reporter microbes (green triangle) that are chemiluminescent in the presence of a specific metabolite.
EcoFab_Image_4
Figure 2. Photo of Brachypodium distachyon.

Model Plant: The EcoFAB concept is predicated on having as much control as possible over all inputs and organisms. Thus, it is desirable to use plants that are well characterized, genetically transformable, experimentally tractable and small enough to complete their lifecycle in the confines of a compact EcoFAB. Thus a wide range of plants could be used, including plants of the genera Arabidopsis, Setaria, and Brachypodium, among many others. For our initial experiments we have chosen to use two related model grasses, one that is annual (Brachypodium distachyon, Figure 2) and one that is perennial (B. sylvaticum). Having both annual and perennial model plants will facilitate investigation of soil C and microbiome dynamics under conditions that mimic farming with annual grains as well as perennial forage/biomass grasses. Both Brachypodium species form mycorrhizal symbioses (unlike Arabidopsis) and possess a typical grass root system. The genus Brachypodium occupies an intermediate evolutionary position in the grass family that allows it to be used as a general model for the grasses including grains and the grasses grown as biomass crops (Draper et al., 2001; Steinwand et al., 2013). The extensive genomic, genetic and experimental resources established for these model grasses (e.g. high-quality genome sequences, extensive sequenced mutant collections, large germplasm collections, and highly efficient transformation methods) will greatly accelerate studies to understand the plant’s contribution to plant-microbiome interactions.


Model Bacteria and Genetic Tools:
Bacteria should originate from the field site, either as isolates or intact communities, depending on the experimental goals. Complex communities can be obtained by filtering soil pore water and enriching the community under a diversity of growth conditions (Breidenbach et al., 2015). As a parallel effort, a multi-condition isolation scheme (using both low and high-throughput approaches) should be conducted to get an isolate library. These isolates can be used to construct synthetic communities whose phenotypes can be compared to the native, enriched communities. Isolates and laboratory communities enable application of exometabolomics approaches to determine the metabolites uptaken and released by isolate organisms to inform possible crossfeeding and resource competition (Baran et al., 2015) and then test predictions by monitoring community metabolism within the EcoFAB. In addition, a great advantage of EcoFAB versus native ecosystems is the ability to use genetic tools and mutants to enable the discovery of important genetic determinants of ecological processes. For example, the genetic tools being developed at Berkeley Lab can be applied to these isolates to gain a mechanistic understanding of their roles in microbial and plant interactions (Deutschbauer et al., 2011; Firrincieli et al., 2015; Lundberg et al., 2012). These approaches include random barcode transposon site mutagenesis coupled to sequencing to examine the fitness of all nonessential genes under in situ conditions and a complementary approach, Dub-Seq that enables examination of the impact of gene overexpression on organismal fitness and therefore can be used to characterize essential genes. Early stage technologies will provide additional tools for manipulation and visualization of the community including as phage engineering to selectively manipulate individual members of a complex microbiome, and reporter constructs for in situ spatial visualization of microbes in their native environment (for example, in association with plant roots, Figure 3). Specifically, work is underway at the Joint Genome Institute to develop chemiluminescent plant growth promoting bacteria to enable their visualization within the root system. Taken together, these approaches will enable the transition from correlative studies of microbiome structure to predictive models based on a detailed molecular understanding of microbe-plant interactions.

EcoFab_Image_2
Figure 4. Illustration of plant microbe metabolite exchange.

Model Fungi: Arbuscular mycorrhizal fungi (AMF) represent one the most ancient growth promoting plant-microbe symbioses and are present in >72% of angiosperm species (Parniske, 2008). As such, they represent the natural condition under which most other plant-microbe interactions occur. For this reason, including AMF is an important component of the EcoFAB design. Despite this, AMF present many challenges for a model system. They are obligate biotrophs and cannot be cultured in the absence of a host, key details of their sexual cycle and nuclear condition remain unknown, and their large genomes (up to 150 Mb) have proved highly challenging to sequence (Boon et al., 2015; Tisserant et al., 2013). Thus, tractability is a key concern for selecting an AMF for inclusion in EcoFAB. For this reason, strong consideration should be given for selection of a model AMF species, such as Rhizophagus irregularis, for which a complete genome exists and spores can be easily obtained from existing AMF culture collections (such as INVAM – The International Vesicular Arbuscular Mycorrhizal Collection). However, due to the promiscuous nature of AMF associations and the nearly global distribution of some AMF taxon, it is likely that a model AMF taxon such as R. irregularis is already present at Russell Ranch or, if not, that the interaction of R. irregularis with Brachypodium sp. and soil bacteria will still be a useful representation of the natural AMF community.

Other Soil Organisms: Small fauna (e.g. arthropods, annelids, mollusks and nematodes), microbial eukaryotes and protists, RNA and DNA viruses associated with the rhizosphere of field Brachypodium are to be screened and isolated for their introduction to the artificial environments in different combinations to test for their influence in the development of the model plants as well as their influence in the overall microbial makeup of the ecosystem, i.e. changes in bacterial, fungal and phage populations. These fauna represent an important addition to EcoFABs in that they play critical roles in soil carbon cycling and other nutrient cycling in the rhizosphere. For example, maceration of plant debris by soil fauna greatly accelerates microbial decomposition, and the predation and subsequent digestion of soil microbes by nematodes can mobilize nitrogen and other nutrients enhancing plant growth (Neher, 1999).

Model Soil: Soil is one of the most difficult control aspects of an EcoFAB design. Field soils are generally unsuitable for controlled experiments for several reasons: they are undefined, they are not reproducible, they are difficult to sterilize without negatively changing soil chemistry, and they perform poorly in pots/containers largely due to poor drainage. While acid washed quartz sand is often used because it is inert, stable and reproducible, it is still undesirable because it has weak cation exchange capacity and lacks pores, both of which negatively impact plant health (Hendershot & Duquette, 1986; Jiang et al., 2009). A kiln-fired ceramic substrate manufactured as a soil amendment for golf courses offers an attractive alternative. This material is stable, reproducible, easily sterilized, and has a high percentage of pore space and high cation exchange capacity (Steinberg et al., 2005). In addition, this substrate contains no organic carbon which will facilitates carbon flux analyses. While this substrate does not possess all the properties of a field soil (e.g. aggregation of small particles and soil organic matter) it is a robust, reproducible reagent that will allow an EcoFAB to be standardized and reproduced in any laboratory. This said, an important research direction for EcoFABs is the development of synthetic soils. These can be a blend of acid washed quartz sand, defined organic amendments and synthetic clays to match the composition at Russell Ranch. While the preparation of such synthetic soils that accurately recapitulate native soils is an important long-term goal, in the short term combinations of simpler materials like the kiln fired ceramic substrate and gamma irradiated native soils provide viable starting points for the construction of EcoFABs.

Growth Chambers: There are many relevant examples of rhizotrons that provide starting points for the design of EcoFABs. 3-D printing technologies have the potential to greatly enable the construction of EcoFABs either through direct fabrication or through printing molds used to cast PDMS or other materials. For example, the root chip system which uses a PDMS chamber affixed to a glass slide space (Figure 5A) (Grossmann et al., 2011). Its plant reservoir is sealed with agarose gel, and the roots can penetrate the gel and grow along the observation chamber be imaged using microscopy. Inlets and outlets enable the contents of the chamber to be changed for example in the case of hydroponic growth.

EcoFab_Image_1
Figure 5. Rhizotrons designs that will be useful for EcoFAB construction. (A) Root-Chip (Grossmann et al., 2011). (B) GLO-Roots platform (Rellán-Álvarez et al., 2015). (C) Minirhizotron (Wang et al., 2004).

Another example is the GLO-Roots system recently reported that uses a 2 mm thin sheet of soil between two sheets of polycarbonate plastic enabling the cultivation of plants such that their root system can be visualized especially for engineered plants expressing luciferase (Figure 5B) (Rellán-Álvarez et al., 2015). Plastic racks hold the rhizotrons vertically and further protect the roots from light. These rhizotrons and rack are placed in a black tub and a small amount of water is added to the bottom to maintain moisture in the rhizotrons during plant growth. The volume of soil in the rhizotrons is similar to small pots commonly used for Arabidopsis growth and supports growth of Arabidopsis throughout its entire lifecycle. One final rhizotron system that provides additional capabilities for EcoFABs construction is the Minirhizotron (Figure 5C) (Wang et al., 2004). In this case, a portion of the rhizotrons plate contains small tubes positioned within the rhizotron connected to a manifold enabling spatially defined introduction and sampling of the rhizotrons contents. Integration of these three designs could provide a powerful EcoFAB design enabling both control of bulk flow through the chamber, spatially defined chemiluminescence imaging of the root system (and microbes, Figure 6), and the ability to sample and add microbes and materials to the root system in a spatially defined manner.

EcoFab_Image_6
Figure 6. Illustration of an EcoFAB containing chemiluminescent plant and reporter microbes signaling the presence of different metabolites from root exudates.

All of this could be done while simultaneously imaging the rhizotrons contents using spectroscopic techniques. We also envisioned a range of advanced sensors could be integrated into the EcoFABs. These could be spatially defined micro-electrodes, opcodes and even wireless microdevices which together can provide near real time information on the environmental conditions within the chamber. This can be extremely powerful when integrated with the rhizotrons technologies described above. For example, arrays of pH sensors within an EcoFAB can be used to indicate microbial growth relative to the growing chemiluminescent root system while Minirhizotron sampling and sequencing through mass spectrometry analysis can identify the organisms and metabolites within this same region. It is important to note that while these examples are focused on plant-soil-microbe systems, the plant can be omitted to focus on soil-microbe interactions. A great advantage of using 3-D printing is that it can provide a great deal of flexibility enabling EcoFABs to be scaled appropriately for the question being asked. We can envision designs for EcoFABs be disseminated such that scientists around the world would be able to easily leverage each other’s designs and local 3-D printing facilities to rapidly construct EcoFABs relevant to their question and may also be useful for educational purposes. For example: (http://rr-lab.github.io/GLO-Roots/).

EcoFAB Characterization and Validation: Another important component of reproducible experiments is reproducible and standardized measurement technologies. There is a vast array of technologies which could be deployed within the EcoFAB context. Several methods were proposed for characterization of microbial community membership, activity and metabolism and are summarized briefly below:

Measurements of community structure, gene expression, protein expression and metabolic potential:

  • Standard Illumina-based tag sequencing of marker genes including 16S rRNA and fungal Internal Transcribed Spacer (ITS) regions.
  • Shotgun sequencing with Illumina, Pacific Biosciences or Oxford Nanopore.
  • NanoStrings, RNA-seq to monitor gene expression.
  • Standard proteomic methods can be used to characterize protein expression from biomass collected from the EcoFAB.

Assessment of active populations, metabolic flux, metabolic exchange and nutrient cycling during EcoFAB operation:

  • 16S rRNA vs 16S rDNA sequencing to identify dormant taxa.
  • Stable isotope probing (SIP), in which a substrate containing an isotopic label such as 13C or 15N is added to the community such that members able to take up and metabolize the substrate will incorporate the label into their nucleic acids. Isolation and separation of the heavy and light nucleic acids followed by sequence-based characterization enables identification of those microbes.
  • BioOrthogonal NonCanonical Amino acid Tagging (BONCAT) coupled with Fluorescence Activated Cell Sorting (FACS), in which synthetic amino acids are taken up by metabolically active cells which can be rendered fluorescent by click chemistry and physically isolated for sequence-based characterization.
  • D2O coupled with Raman spectroscopy and microfluidic separation to isolate metabolically active organisms incorporating label into their lipids, followed by sequence-based metabolomics characterization.
  • Electrochemical sensors can be included in the EcoFAB to provide near real-time analysis of the wide array of chemical parameters including pH, oxygen and electronic transfer. One exciting possibility is to use immobilized microbes that transfer electrons to the surface in response to specific environmental conditions.
  • Optodes provide another approach for measuring chemical conditions within the EcoFAB and are well suited for measuring parameters such as oxygen.
  • Exometabolomics provides a powerful tool for characterization of the metabolic transformations happening within the EcoFAB. This can be performed on spatially sampled regions of the chamber as well as the inflow and outflow to determine which metabolites are being produced and consumed within the various regions. Methods for analysis can include standard methods such as liquid chromatography tandem mass spectrometry and gas chromatography mass spectrometry but could also include high throughput laser desorption ionization methods that are particularly suited for very small sample volumes.
  • Spectroscopic methods such as Raman and FTIR which enable analysis of biopolymer composition.
  • Enzyme activity assay either using surrogate substrates coupled with colorimetric or fluorescence analysis or native substrates with mass spectrometry analysis can provide important insights into the enzymatic activities occurring within the EcoFAB.

Synthetic biology and genetic tools:

  • High throughput genetics can be used to construct plant and microbial mutants for discovery with genes responsible for important phenotypes such as plant genes selecting for beneficial microbes. Mutant fitness analysis using Tn-SEQ and dub-SEQ can be used to discover microbial genes conferring fitness under specific environmental conditions or space responsible for ecological interactions.
  • Synthetic biology tools can be used to introduce biosynthetic pathways to test their ecological function, for example, the production of secondary metabolites associated with intraspecific interactions, plant growth promotion, nutrient mobilization. It is also possible to construct reporter systems that will emit chemiluminescence in response to abiotic or biotic conditions of interest.

Methods of characterizing EcoFAB endpoints:

  • X-ray techniques enable diverse analyses of the resulting EcoFAB contents. For example, x-ray microtomography can be used to create a three-dimensional reconstruction of the pore structure within the EcoFAB. X-ray absorptionspectroscopy can localize minerals and even measure at the nanoscale the absorption of metabolites and biopolymers onto mineral surfaces.
  • Mass spectrometry imaging can be used to localize biomolecules especially metabolites and elements at the conclusion of an EcoFAB study. For example nanoSIMS can locate specific elements with 50 nanometer resolution providing vital information on the localization of stable isotopes within individual bacteria. Soft desorption ionization mass spectrometry approaches can be used to localize metabolites within the soil. DESI, NanoDESI, LAESI, and MALDI (Fang & Dorrestein, 2014) are all examples of technologies that can be used to localize metabolites within EcoFAB contents. A related technology, nanostructure-initiator mass spectrometry (Deng et al., 2015; Northen et al., 2007) has been used to image both enzyme activity and metabolites from fungal growth on switchgrass and is shown in Figure 7. NIMS imaging has sufficient spatial resolution to image fungal hyphae and is effective at localizing isotope incorporation given that high resolution mass spectrometers are used in this application.
Figure 7. NIMS imaging of enzyme activity and metabolites from Aspergillus growing on switchgrass showing the rich mass spectra obtained.

Multiple aspects of EcoFABs will require different modeling approaches. Gene and functional distributions can be understood in terms of population genetics models, while organism abundances and interaction distributions require ecological and game theory models. Integration of abundance and spatial information will require integrating plant development models with the population genetics and ecological models mentioned above. Integration of metabolic and chemical data with genomic content information can be achieved with multi-species flux based models. Ultimately, EcoFAB will produce a multi-scale modeling approach that combines fine- and coarse-grained methods on data from a common experimental platform. Moreover, key plant and microbial traits and environmental constraints can become incorporated into large-scale modeling efforts (e.g. CENTURY model).

EcoFAB Modeling and Data Analysis: The only way to derive robust and generalizable design principles for soil communities is to underpin this effort by predictive theoretical models. We will first decide on the community properties that will be most beneficial to predict (e.g. abundance distribution, diversity as a function of distance, robustness to perturbations, resilience). Once these are decided on, and depending on the types of data which are available in a repeatable and abundant fashion, the right modeling framework will be chosen, e.g. flux-based vs trait-based, coarse-grained vs spatially explicit, statistical vs mechanistic.