Lipidome Plasticity Enables Unusual Photosynthetic Flexibility in Arctic vs. Temperate Diatoms.

The diatom lipidome actively regulates photosynthesis and displays a high degree of plasticity in response to a light environment, either directly as structural modifications of thylakoid membranes and protein-pigment complexes, or indirectly via photoprotection mechanisms that dissipate excess light energy. This acclimation is crucial to maintaining primary production in marine systems, particularly in polar environments, due to the large temporal variations in both the intensity and wavelength distributions of downwelling solar irradiance. This study investigated the hypothesis that Arctic marine diatoms uniquely modify their lipidome, including their concentration and type of pigments, in response to wavelength-specific light quality in their environment. We postulate that Arctic-adapted diatoms can adapt to regulate their lipidome to maintain growth in response to the extreme variability in photosynthetically active radiation. This was tested by comparing the untargeted lipidomic profiles, pigmentation, specific growth rates and carbon assimilation of the Arctic diatom Porosira glacialis vs. the temperate species Coscinodiscus radiatus during exponential growth under red, blue and white light. Here, we found that the chromatic wavelength influenced lipidome remodeling and growth in each strain, with P. glacialis showing effective utilization of red light coupled with increased inclusion of primary light-harvesting pigments and polar lipid classes. These results indicate a unique photoadaptation strategy that enables Arctic diatoms like P. glacialis to capitalize on a wide chromatic growth range and demonstrates the importance of active lipid regulation in the Arctic light environment.

Revealing the Host-Dependent Nature of an Engineered Genetic Inverter in Concordance with Physiology.

Broad-host-range synthetic biology is an emerging frontier that aims to expand our current engineerable domain of microbial hosts for biodesign applications. As more novel species are brought to "model status," synthetic biologists are discovering that identically engineered genetic circuits can exhibit different performances depending on the organism it operates within, an observation referred to as the "chassis effect." It remains a major challenge to uncover which genome-encoded and biological determinants will underpin chassis effects that govern the performance of engineered genetic devices. In this study, we compared model and novel bacterial hosts to ask whether phylogenomic relatedness or similarity in host physiology is a better predictor of genetic circuit performance. This was accomplished using a comparative framework based on multivariate statistical approaches to systematically demonstrate the chassis effect and characterize the performance dynamics of a genetic inverter circuit operating within 6 Gammaproteobacteria. Our results solidify the notion that genetic devices are strongly impacted by the host context. Furthermore, we formally determined that hosts exhibiting more similar metrics of growth and molecular physiology also exhibit more similar performance of the genetic inverter, indicating that specific bacterial physiology underpins measurable chassis effects. The result of this study contributes to the field of broad-host-range synthetic biology by lending increased predictive power to the implementation of genetic devices in less-established microbial hosts.

Daylight-driven carbon exchange through a vertically structured microbial community.

Interactions between autotrophs and heterotrophs are central to carbon (C) exchange across trophic levels in essentially all ecosystems and metabolite exchange is a frequent mechanism for distributing C within spatially structured ecosystems. Yet, despite the importance of C exchange, the timescales at which fixed C is transferred in microbial communities is poorly understood. We employed a stable isotope tracer combined with spatially resolved isotope analysis to quantify photoautotrophic uptake of bicarbonate and track subsequent exchanges across a vertical depth gradient in a stratified microbial mat over a light-driven diel cycle. We observed that C mobility, both across the vertical strata and between taxa, was highest during periods of active photoautotrophy. Parallel experiments with 13C-labeled organic substrates (acetate and glucose) showed comparably less exchange of C within the mat. Metabolite analysis showed rapid incorporation of 13C into molecules that can both comprise a portion of the extracellular polymeric substances in the system and serve to transport C between photoautotrophs and heterotrophs. Stable isotope proteomic analysis revealed rapid C exchange between cyanobacterial and associated heterotrophic community members during the day with decreased exchange at night. We observed strong diel control on the spatial exchange of freshly fixed C within tightly interacting mat communities suggesting a rapid redistribution, both spatially and taxonomically, primarily during daylight periods.

Microbial Community Dynamics during a Harmful Chrysochromulina leadbeateri Bloom in Northern Norway.

A harmful algal bloom occurred in late spring 2019 across multiple, interconnected fjords and bays in northern Norway. The event was caused by the haptophyte Chrysochromulina leadbeateri and led to severe fish mortality at several salmon aquaculture facilities. This study reports on the spatial and temporal succession dynamics of the holistic marine microbiome associated with this bloom by relating all detectable 18S and 16S rRNA gene amplicon sequence variants to the relative abundance of the C. leadbeateri focal taxon. A k-medoid clustering enabled inferences on how the causative focal taxon cobloomed with diverse groups of bacteria and microeukaryotes. These coblooming patterns showed high temporal variability and were distinct between two geographically separated time series stations during the regional harmful algal bloom. The distinct blooming patterns observed with respect to each station were poorly connected to environmental conditions, suggesting that other factors, such as biological interactions, may be at least as important in shaping the dynamics of this type of harmful algal bloom. A deeper understanding of microbiome succession patterns during these rare but destructive events will help guide future efforts to forecast deviations from the natural bloom cycles of the northern Norwegian coastal marine ecosystems that are home to intensive aquaculture activities. IMPORTANCE The 2019 Chrysochromulina leadbeateri bloom in northern Norway had a major impact on the local economy and society through its devastating effect on the aquaculture industry. However, many fail to remember that C. leadbeateri is, in fact, a common member of the seasonal marine microbiome and the same spring phytoplankton blooms that support the marine ecosystem. It is challenging to draw any conclusions about exact causation behind the harmful bloom of 2019, especially since the natural bloom cycles of C. leadbeateri are not well understood. This study begins to fill major knowledge gaps that may lead to future forecasting abilities, by providing a molecular-based investigation of the destructive 2019 bloom that presents new insights into a seasonal marine microbial ecosystem during one of these sporadically reoccurring events.

Flashing lights affect the photophysiology and expression of carotenoid and lipid synthesis genes in Nannochloropsis gaditana.

Nannochloropsis gaditana is a promising microalga for biotechnology. One of the strategies to stimulate its full potential in metabolite production is exposure to flashing lights. Here, we report how N. gaditana adapts to different flashing light regimes (5, 50, and 500 Hz) by changing its cellular physiology and the relative expression of genes related to critical cellular functions. We analyzed the differential mRNA abundance of genes related to photosynthesis, nitrogen assimilation and biosynthesis of chlorophyll, carotenoids, lipids, fatty acids and starch. Analysis of photosynthetic efficiency and high mRNA abundance of photoprotection genes supported the inference that excess excitation energy provided by light absorbance during photosynthesis was produced under low frequency flashing lights and was dissipated by photopigments via the xanthophyll-cycle. Increased relative expression levels of genes related to the synthesis of carotenoids and chlorophyll confirmed the accumulation of photopigments previously observed at low frequency flashing lights. Higher differential mRNA abundance of genes related to the triacylglycerol biosynthesis were observed at lower frequency flashing lights, possibly triggered by a poor nitrogen assimilation caused by low mRNA abundance of a nitrate reductase gene. This study advances a new understanding of algal physiology and metabolism leading to improved cellular performance and metabolite production.

Diversity and Selection of Surface Marine Microbiomes in the Atlantic-Influenced Arctic.

Arctic marine environments are experiencing rapid changes due to the polar amplification of global warming. These changes impact the habitat of the cold-adapted microbial communities, which underpin biogeochemical cycles and marine food webs. We comparatively investigated the differences in prokaryotic and microeukaryotic taxa between summer surface water microbiomes sampled along a latitudinal transect from the ice-free southern Barents Sea and into the sea-ice-covered Nansen Basin to disentangle the dominating community (ecological) selection processes driving phylogenetic diversity. The community structure and richness of each site-specific microbiome were assessed in relation to the physical and biogeochemical conditions of the environment. A strong homogeneous deterministic selection process was inferred across the entire sampling transect via a phylogenetic null modeling approach. The microbial species richness and diversity were not negatively influenced by northward decreasing temperature and salinity. The results also suggest that regional phytoplankton blooms are a major prevalent factor in governing the bacterial community structure. This study supports the consideration that strong homogeneous selection is imposed across these cold-water marine environments uniformly, regardless of geographic assignments within either the Nansen Basin or the Barents Sea.

Environment Constrains Fitness Advantages of Division of Labor in Microbial Consortia Engineered for Metabolite Push or Pull Interactions.

Fitness benefits from division of labor are well documented in microbial consortia, but the dependency of the benefits on environmental context is poorly understood. Two synthetic Escherichia coli consortia were built to test the relationships between exchanged organic acid, local environment, and opportunity costs of different metabolic strategies. Opportunity costs quantify benefits not realized due to selecting one phenotype over another. The consortia catabolized glucose and exchanged either acetic or lactic acid to create producer-consumer food webs. The organic acids had different inhibitory properties and different opportunity costs associated with their positions in central metabolism. The exchanged metabolites modulated different consortial dynamics. The acetic acid-exchanging (AAE) consortium had a "push" interaction motif where acetic acid was secreted faster by the producer than the consumer imported it, while the lactic acid-exchanging (LAE) consortium had a "pull" interaction motif where the consumer imported lactic acid at a comparable rate to its production. The LAE consortium outperformed wild-type (WT) batch cultures under the environmental context of weakly buffered conditions, achieving a 55% increase in biomass titer, a 51% increase in biomass per proton yield, an 86% increase in substrate conversion, and the complete elimination of by-product accumulation all relative to the WT. However, the LAE consortium had the trade-off of a 42% lower specific growth rate. The AAE consortium did not outperform the WT in any considered performance metric. Performance advantages of the LAE consortium were sensitive to environment; increasing the medium buffering capacity negated the performance advantages compared to WT. IMPORTANCE Most naturally occurring microorganisms persist in consortia where metabolic interactions are common and often essential to ecosystem function. This study uses synthetic ecology to test how different cellular interaction motifs influence performance properties of consortia. Environmental context ultimately controlled the division of labor performance as shifts from weakly buffered to highly buffered conditions negated the benefits of the strategy. Understanding the limits of division of labor advances our understanding of natural community functioning, which is central to nutrient cycling and provides design rules for assembling consortia used in applied bioprocessing.

Systems biology: current status and challenges.

We put together a special issue on current approaches in systems biology with a focus on mathematical modeling of metabolic networks. Mathematical models have increasingly been used to unravel molecular mechanisms of complex dynamic biological processes. We here provide a short introduction into the topics covered in this special issue, highlighting current developments and challenges.

Structure Dependent Determination of Organophosphate Targets in Mammalian Tissues Using Activity-Based Protein Profiling.

Acute and chronic exposures to organophosphates (OPs), including agricultural pesticides, industrial chemicals, and chemical warfare agents, remain a significant worldwide health risk. The mechanisms by which OPs alter development and cognition in exposed individuals remain poorly understood, in part due to the large number of structurally diverse OPs and the wide range of affected proteins and signaling pathways. To investigate the influence of structure on OP targets in mammalian systems, we have developed a series of probes for activity-based protein profiling (ABPP) featuring two distinct reactive groups that mimic OP chemical reactivity. FOP features a fluorophosphonate moiety, and PODA and CODA utilize a dialkynyl phosphate ester; both reactive group types target serine hydrolase activity. As the oxon represents the highly reactive and toxic functional group of many OPs, the new probes described herein enhance our understanding of tissue-specific reactivity of OPs. Chemoproteomic analysis of mouse tissues treated with the probes revealed divergent protein profiles, demonstrating the influence of probe structure on protein targeting. These targets also vary in sensitivity toward different OPs. The simultaneous use of multiple probes in ABPP experiments may therefore offer more comprehensive coverage of OP targets; FOP consistently labeled more targets in both brain and liver than PODA or CODA, suggesting the dialkyne warhead is more selective for enzymes in major signaling pathways than the more reactive fluorophosphonate warhead. Additionally, the probes can be used to assess reactivation of OP-inhibited enzymes by N-oximes and may serve as diagnostic tools for screening of therapeutic candidates in a panel of protein targets. These applications will help clarify the short- and long-term effects of OP toxicity beyond acetylcholinesterase inhibition, investigate potential points of convergence for broad spectrum therapeutic development, and support future efforts to screen candidate molecules for efficacy in various model systems.

Minimal Interspecies Interaction Adjustment (MIIA): Inference of Neighbor-Dependent Interactions in Microbial Communities.

An intriguing aspect in microbial communities is that pairwise interactions can be influenced by neighboring species. This creates context dependencies for microbial interactions that are based on the functional composition of the community. Context dependent interactions are ecologically important and clearly present in nature, yet firmly established theoretical methods are lacking from many modern computational investigations. Here, we propose a novel network inference method that enables predictions for interspecies interactions affected by shifts in community composition and species populations. Our approach first identifies interspecies interactions in binary communities, which is subsequently used as a basis to infer modulation in more complex multi-species communities based on the assumption that microbes minimize adjustments of pairwise interactions in response to neighbor species. We termed this rule-based inference minimal interspecies interaction adjustment (MIIA). Our critical assessment of MIIA has produced reliable predictions of shifting interspecies interactions that are dependent on the functional role of neighbor organisms. We also show how MIIA has been applied to a microbial community composed of competing soil bacteria to elucidate a new finding that - in many cases - adding fewer competitors could impose more significant impact on binary interactions. The ability to predict membership-dependent community behavior is expected to help deepen our understanding of how microbiomes are organized in nature and how they may be designed and/or controlled in the future.

Selection, Succession, and Stabilization of Soil Microbial Consortia.

Soil microorganisms play fundamental roles in cycling of soil carbon, nitrogen, and other nutrients, yet we have a poor understanding of how soil microbiomes are shaped by their nutritional and physical environment. In this study, we investigated the successional dynamics of a soil microbiome during 21 weeks of enrichment on chitin and its monomer, N-acetylglucosamine. We examined succession of the soil communities in a physically heterogeneous soil matrix as well as a homogeneous liquid medium. The guiding hypothesis was that the initial species richness would influence the tendency for the selected consortia to stabilize and maintain a relatively constant community structure over time. We also hypothesized that long-term, substrate-driven growth would result in consortia with reduced species richness compared to the parent microbiome and that this process would be deterministic with relatively little variation between replicates. We found that the initial species richness does influence the long-term community stability in both liquid media and soil and that lower initial richness results in a more rapid convergence to stability. Despite use of the same soil inoculum and access to the same major substrate, the resulting community composition differed greatly in soil from that in liquid medium. Hence, distinct selective pressures in soils relative to homogenous liquid media exist and can control community succession dynamics. This difference is likely related to the fact that soil microbiomes are more likely to thrive, with fewer compositional changes, in a soil matrix than in liquid environments. IMPORTANCE The soil microbiome carries out important ecosystem functions, but interactions between soil microbial communities have been difficult to study due to the high microbial diversity and complexity of the soil habitat. In this study, we successfully obtained stable consortia with reduced complexity that contained species found in the original source soil. These consortia and the methods used to obtain them can be a valuable resource for exploration of specific mechanisms underlying soil microbial community ecology. The results of this study also provide new experimental context to better inform how soil microbial communities are shaped by new environments and how a combination of initial taxonomic structure and physical environment influences stability.

Reconciling Ecological and Engineering Design Principles for Building Microbiomes.

Simplified microbial communities, or "benchtop microbiomes," enable us to manage the profound complexity of microbial ecosystems. Widespread activities aiming to design and control communities result in novel resources for testing ecological theories and also for realizing new biotechnologies. There is much to be gained by reconciling engineering design principles with ecological processes that shape microbiomes in nature. In this short Perspective, I will address how natural processes such as environmental filtering, the establishment of priority effects, and community "blending" (coalescence) can be harnessed for engineering microbiomes from complex starting materials. I will also discuss how future microbiome architects may draw inspiration from modern practices in synthetic biology. This topic is based on an important overarching research goal, which is to understand how natural forces shape microbial communities and interspecies interactions such that new engineering design principles can be extracted to promote human health or energy and environmental sustainability.

Forfeiting the priority effect: turnover defines biofilm community succession.

Microbial community succession is a fundamental process that affects underlying functions of almost all ecosystems; yet the roles and fates of the most abundant colonizers are often poorly understood. Does early abundance spur long term persistence? How do deterministic and stochastic processes influence the ecological contribution of colonizers? We performed a succession experiment within a hypersaline ecosystem to investigate how different processes contributed to the turnover of founder species. Bacterial and eukaryotic colonizers were identified during primary succession and tracked through a defined, 79-day biofilm maturation period using 16S and 18S rRNA gene sequencing in combination with high resolution imaging that utilized stable isotope tracers to evaluate successional patterns of primary producers and nitrogen fixers. The majority of the founder species did not maintain high abundance throughout succession. Species replacement (versus loss) was the dominant process shaping community succession. We also asked if different ecological processes acted on bacteria versus Eukaryotes during succession and found deterministic and stochastic forces corresponded more with microeukaryote and bacterial colonization, respectively. Our results show that taxa and functions belonging to different kingdoms, which share habitat in the tight spatial confines of a biofilm, were influenced by different ecological processes and time scales of succession.

Prediction of Neighbor-Dependent Microbial Interactions From Limited Population Data.

Modulation of interspecies interactions by the presence of neighbor species is a key ecological factor that governs dynamics and function of microbial communities, yet the development of theoretical frameworks explicit for understanding context-dependent interactions are still nascent. In a recent study, we proposed a novel rule-based inference method termed the Minimal Interspecies Interaction Adjustment (MIIA) that predicts the reorganization of interaction networks in response to the addition of new species such that the modulation in interaction coefficients caused by additional members is minimal. While the theoretical basis of MIIA was established through the previous work by assuming the full availability of species abundance data in axenic, binary, and complex communities, its extension to actual microbial ecology can be highly constrained in cases that species have not been cultured axenically (e.g., due to their inability to grow in the absence of specific partnerships) because binary interaction coefficients - basic parameters required for implementing the MIIA - are inestimable without axenic and binary population data. Thus, here we present an alternative formulation based on the following two central ideas. First, in the case where only data from axenic cultures are unavailable, we remove axenic populations from governing equations through appropriate scaling. This allows us to predict neighbor-dependent interactions in a relative sense (i.e., fractional change of interactions between with versus without neighbors). Second, in the case where both axenic and binary populations are missing, we parameterize binary interaction coefficients to determine their values through a sensitivity analysis. Through the case study of two microbial communities with distinct characteristics and complexity (i.e., a three-member community where all members can grow independently, and a four-member community that contains member species whose growth is dependent on other species), we demonstrated that despite data limitation, the proposed new formulation was able to successfully predict interspecies interactions that are consistent with experimentally derived results. Therefore, this technical advancement enhances our ability to predict context-dependent interspecies interactions in a broad range of microbial systems without being limited to specific growth conditions as a pre-requisite.

Species-specific transcriptomic network inference of interspecies interactions.

The advent of high-throughput 'omics approaches coupled with computational analyses to reconstruct individual genomes from metagenomes provides a basis for species-resolved functional studies. Here, a mutual information approach was applied to build a gene association network of a commensal consortium, in which a unicellular cyanobacterium Thermosynechococcus elongatus BP1 supported the heterotrophic growth of Meiothermus ruber strain A. Specifically, we used the context likelihood of relatedness (CLR) algorithm to generate a gene association network from 25 transcriptomic datasets representing distinct growth conditions. The resulting interspecies network revealed a number of linkages between genes in each species. While many of the linkages were supported by the existing knowledge of phototroph-heterotroph interactions and the metabolism of these two species several new interactions were inferred as well. These include linkages between amino acid synthesis and uptake genes, as well as carbohydrate and vitamin metabolism, terpenoid metabolism and cell adhesion genes. Further topological examination and functional analysis of specific gene associations suggested that the interactions are likely to center around the exchange of energetically costly metabolites between T. elongatus and M. ruber. Both the approach and conclusions derived from this work are widely applicable to microbial communities for identification of the interactions between species and characterization of community functioning as a whole.

Phenotypic responses to interspecies competition and commensalism in a naturally-derived microbial co-culture.

The fundamental question of whether different microbial species will co-exist or compete in a given environment depends on context, composition and environmental constraints. Model microbial systems can yield some general principles related to this question. In this study we employed a naturally occurring co-culture composed of heterotrophic bacteria, Halomonas sp. HL-48 and Marinobacter sp. HL-58, to ask two fundamental scientific questions: 1) how do the phenotypes of two naturally co-existing species respond to partnership as compared to axenic growth? and 2) how do growth and molecular phenotypes of these species change with respect to competitive and commensal interactions? We hypothesized - and confirmed - that co-cultivation under glucose as the sole carbon source would result in competitive interactions. Similarly, when glucose was swapped with xylose, the interactions became commensal because Marinobacter HL-58 was supported by metabolites derived from Halomonas HL-48. Each species responded to partnership by changing both its growth and molecular phenotype as assayed via batch growth kinetics and global transcriptomics. These phenotypic responses depended on nutrient availability and so the environment ultimately controlled how they responded to each other. This simplified model community revealed that microbial interactions are context-specific and different environmental conditions dictate how interspecies partnerships will unfold.

Primary and heterotrophic productivity relate to multikingdom diversity in a hypersaline mat.

Benthic microbial ecosystems are widespread yet knowledge gaps still remain on the relationships between the diversity of species across kingdoms and productivity. Here, we ask two fundamental questions: (i) How does species diversity relate to the rates of primary and heterotrophic productivity? (ii) How do diel variations in light-energy inputs influence productivity and microbiome diversity? To answer these questions, microbial mats from a magnesium sulfate hypersaline lake were used to establish microcosms. Both the number and relatedness between bacterial and eukaryotic taxa in the microbiome were assayed via amplicon-based sequencing of 16S and 18S rRNA genes over two diel cycles. These results correlated with biomass productivity obtained from substrate-specific 13C stable isotope tracers that enabled comparisons between primary and heterotrophic productivity. Both bacterial and eukaryotic species richness and evenness were related only to the rates of 13C-labeled glucose and acetate biomass incorporation. Interestingly, measures of these heterotrophic relationships changed from positive and negative correlations depending on carbon derived from glucose or acetate, respectively. The bacterial and eukaryotic diversity of this ecosystem is also controlled, in part, from energy constraints imposed by changing irradiance over a diel cycle.

Organismal and spatial partitioning of energy and macronutrient transformations within a hypersaline mat.

Phototrophic mat communities are model ecosystems for studying energy cycling and elemental transformations because complete biogeochemical cycles occur over millimeter-to-centimeter scales. Characterization of energy and nutrient capture within hypersaline phototrophic mats has focused on specific processes and organisms; however, little is known about community-wide distribution of and linkages between these processes. To investigate energy and macronutrient capture and flow through a structured community, the spatial and organismal distribution of metabolic functions within a compact hypersaline mat community from Hot Lake have been broadly elucidated through species-resolved metagenomics and geochemical, microbial diversity and metabolic gradient measurements. Draft reconstructed genomes of 34 abundant organisms revealed three dominant cyanobacterial populations differentially distributed across the top layers of the mat suggesting niche separation along light and oxygen gradients. Many organisms contained diverse functional profiles, allowing for metabolic response to changing conditions within the mat. Organisms with partial nitrogen and sulfur metabolisms were widespread indicating dependence on metabolite exchange. In addition, changes in community spatial structure were observed over the diel. These results indicate that organisms within the mat community have adapted to the temporally dynamic environmental gradients in this hypersaline mat through metabolic flexibility and fluid syntrophic interactions, including shifts in spatial arrangements.

Indirect Interspecies Regulation: Transcriptional and Physiological Responses of a Cyanobacterium to Heterotrophic Partnership.

The mechanisms by which microbes interact in communities remain poorly understood. Here, we interrogated specific interactions between photoautotrophic and heterotrophic members of a model consortium to infer mechanisms that mediate metabolic coupling and acclimation to partnership. This binary consortium was composed of a cyanobacterium, Thermosynechococcus elongatus BP-1, which supported growth of an obligate aerobic heterotroph, Meiothermus ruber strain A, by providing organic carbon, O2, and reduced nitrogen. Species-resolved transcriptomic analyses were used in combination with growth and photosynthesis kinetics to infer interactions and the environmental context under which they occur. We found that the efficiency of biomass production and resistance to stress induced by high levels of dissolved O2 increased, beyond axenic performance, as a result of heterotrophic partnership. Coordinated transcriptional responses transcending both species were observed and used to infer specific interactions resulting from the synthesis and exchange of resources. The cyanobacterium responded to heterotrophic partnership by altering expression of core genes involved with photosynthesis, carbon uptake/fixation, vitamin synthesis, and scavenging of reactive oxygen species (ROS). IMPORTANCE This study elucidates how a cyanobacterial primary producer acclimates to heterotrophic partnership by modulating the expression levels of key metabolic genes. Heterotrophic bacteria can indirectly regulate the physiology of the photoautotrophic primary producers, resulting in physiological changes identified here, such as increased intracellular ROS. Some of the interactions inferred from this model system represent putative principles of metabolic coupling in phototrophic-heterotrophic partnerships.