Imaging signalling and transport in complex systems
Biological Network analysis
We have pioneered network analysis of foraging basidiomycete fungi growing in unconstrained microcosms (Bebber et al. 2007a; Bebber et al. 2007b; Boddy et al. 2009; Boddy et al. 2010; Fricker et al. 2008a; Fricker et al. 2007; Fricker et al. 2008b; Heaton et al. 2010; Rotheray et al. 2008). We have shown using graph-theoretic analysis of digitised networks, that these indeterminate, de-centralized systems can yield adaptive networks with both high transport capacity and robustness to damage, but at a relatively low cost, through a 'Darwinian' process of selective reinforcement of key transport pathways and recycling of redundant routes (Bebber et al. 2007a). Furthermore, fungal networks are able to dynamically modify link strengths and local connectivity when subject to experimental attack to readjust the balance between transport capacity, robustness to damage and resource allocation, resulting in increased resilience as the environment becomes more challenging (Boddy et al. 2010; Rotheray et al. 2008). Furthermore, genetically disrupting network formation in Neurospora reduces long distance nutrient transport (Simonin et al., 2012).
Currently network extraction is a laborious manual exercise so we have developed high-throughput, automated imaging, visualisation and network analysis protocols to extract biological network organisation using Phase Congruency Tensors (PCTs) to specifically enhance curvi-linear features (Obara, 2012a-c). This approach can rapidly extract networks with 105 links in a few minutes with high fidelity.
The underlying mechanisms leading to the emergence of adaptive behaviour in macroscopic mycelial networks are unknown. However, models based on growth-induced mass flow through the experimentally determined macroscopic networks provide a high level of explanatory power for the transport of added radiolabel (Heaton et al. 2010; Heaton et al. 2012). We infer that bio-physical hydraulic coupling and internal flows observed in macroscopic networks may act as the central mechanism enabling coordinated growth across the complete range of scales in networked organisms. These models can be extended to capture the energetic constraints that determine fungal life history strategy (Heaton et al. 2015).
Slime mold networks
The work on fungal networks has also led to collaboration with Toshiyuki Nakagaki and his team in Hokkaido to understand network formation in the acellular slime mold Physarum polycephalum. We have shown that it can form networks with comparable efficiency, fault tolerance, and cost to those of real-world infrastructure networks such as the Tokyo rail system (Tero et al. 2010; Kunita et al., 2014), which was awarded an IgNobel prize in 2010. The core mechanisms needed for adaptive network formation can be captured in a biologically inspired mathematical model that may be useful to guide network construction in other domains. Network formation in fungi and slime molds can also be compared to network architecture in other domains using mesoscale community detection and clustering algorithms (Onnela et al., 2012; Lee et al., 2015).
Long distance nutrient transport in fungi
Analysis of the network architecture allows predictive modeling of the expected nutrient dynamics. The pattern of nutrient distribution can be mapped in vivo using novel scintillation imaging techniques we have developed to map the transport of 14C-labelled α-amino isobutyrate (14C-AIB) as a non-metabolised, radiolabelled amino-acid analogue in Phanerochaete velutina (Tlalka et al. 2002, Tlalka et al. 2003, Bebber et al. 2007, Bebber et al. 2007, Fricker et al. 2008, Tlalka et al. 2008). We have revealed novel N-transport phenomena, including rapid, preferential N-resource allocation to C-rich sinks, induction of simultaneous bi-directional transport, and abrupt switching between different pre-existing transport routes (Tlalka et al., 2002; Tlalka et al., 2003; Bebber et al., 2007; Fricker et al., 2007; Tlalka et al., 2007). Using a simulation model we have also shown that preferential N-allocation and growth in response to a new resource in a range of resource environments confers an ecological benefit (Darrah et al., 2014). There is also a pulsatile component to transport and colonies self-organise into well demarcated domains that are identifiable by differences in the phase relationship of the pulses (Tlalka et al., 2003; Bebber et al., 2007; Fricker et al., 2007; Tlalka et al., 2007).
Imaging signal transduction
We have a long track record in the development and application of imaging techniques to map redox homeostasis and signalling plant and fungal systems (Moore et al., 2006; Brandizzi et al,. 2002). We developed a system for quantitative imaging of total cytoplasmic glutathione (GSH) in vivo following GST-catalysed conjugation to monochlorobimane (MCB) to give a fluorescent glutathione-bimane (GSB) adduct. This has allowed measurement of concentrations of GSH in defined cell types in intact tissues, to dissect the control of GSH synthesis and to analyse activities of GSH-based xenobiotic detoxification pathways in intact tissues with sub-cellular resolution. The imaging assay is technically complex (Fricker et al., 2000; Meyer and Fricker, 2000; Meyer et al., 2001), but now successfully applied to GSH measurements in several different types of tissues including roots (Sanchez-Fernandez et al., 1997; Fricker et al., 2001), trichomes (Gutiérrez-Alcalá et al 2000), suspension culture cells (Meyer and Fricker, 2002), and mesophyll, epidermal and guard cells of wild type and transgenic poplar over-expressing g-ECS (Hartmann et al 2003), mutants in ER morphology (Au et al., 2012), and most recently pathogenic fungi (Samalova et al., 2014).
This work on total GSH measurements has now been combined with in vivo redox imaging using ratiometric roGFP1 and roGFP2 (Schwarzlander et al., 2008; Marty et al., 2009) and application of these probes to measure mitochondrial redox state (Wagner et al., 2015) using custom ratio imaging software (Fricker, 2015) under a range of stress conditions (Lehmann et al., 2008; Schwarzlander et al., 2009), in mutants deficient in MnSOD (Morgan et al 2008), the mitochondrial Ca2+–channel subunit MICU1 (Wagner et al., 2015) and during pathogen attack (Fuchs et al., 2015).
We have also shown that cpYFP does not act as a reporter for superoxide, but behaves as an effective pH indicator in vivo, and reveals the presence of transient alkalinisation events associated with flickering in mitochondrial membrane potential (Schwarzlander et al. 2011; 2012).
We have started to move these approaches across to the pathogenic rice blast fungus, Magnaporthe oryzae, coupling the redox and ROS probes with nitric oxide measurements (Samalova et al., 2013; 2014).