Skip to main content

Bacterial Abiotic Cellular Stress and Survival Improvement Network

Final Report Summary - BACSIN (Bacterial Abiotic Cellular Stress and Survival Improvement Network)

Executive Summary:
The central idea behind the BACSIN project was to improve the rational use of bacterial catalytic activities, in particular for targeted pollution treatment, removal and prevention. Whereas in many cases specifically isolated bacterial strains for targeted biodegradation have been used, their success has been largely unpredictable and annectodal. The BACSIN project thus attempts to gather the basic knowledge necessary to understand environmental survival programs in a number of predefined bacsin isolates, by using frontline systems biology, single cell microbiology, classical genetics and modeling.
The project developed around four poles of activity that operated independently but in an iterative fashion as well. One pole was centered around obtaining basic knowledge on the behaviour of a number of bacterial strains with good catabolic potential (so-called bacsins). The strains were chosen on the basis of their target pollutants (e.g. alkanes, BTEX, PAHs, PCE/TCE, chlorophenols), and their (presumed) prefered ecological niche (aerobic/anaerobic, groundwater, soil, rhizosphere, marine systems). Methods were developed to study the strain's behaviour in a standardized manner under a variety of exposures and environments, so that strain-to-strain comparisons would become evident. This included transcriptomics under drought exposure, in soils, on plant leaves, during formulations, or after pollutant exposure. In addition, genome-wide techniques such as transposon mutagenesis were used to discover mutants impaired in survival functions. Results highlight how the strains react to changing conditions, how almost every different condition demands a different gene expression 'program', how specific protection is initiated through membrane changes, synthesis of compatible solutes or operation of parallel and redundant metabolic pathways. These reactions thus allow much better prediction of the behaviour of introduced bacsins in particular sites and the possible measures to promote their proper colonization and active pollutant degradation.
These generalized results were then complemented by a very wide range of microcosm studies to study the actual survival, general activity and pollution degradation capacity by bacsins. This included plant leaves, plant roots, soils, sediments, marine beaches, biofilms and reactors. Surprisingly simple tests provided good estimations of general survival behaviour. For example, a simple plant leaf spray test and dry-humid alteration showed clearly which of the bacsins had superiour survival capacity (i.e. Arthrobacter) and for which others more sensitive procedures would have to be developed. In essence, such studies also showed that after a number of easy microcosm tests almost any bacsin can be made to survive and be 'active' in pollutant degradation. The effect of other factors that compete for good biodegradation were examined and modeled, some of which are more difficult to solve satisfactorily (e.g. competition by endogenous microbes). Large emphasis was also placed on specific formulation procedures and a better understanding of formulation effects on bacsin physiology and rewetting success. Some bacsins resisted remarkably simple formulation procedures, such as mixing with peat or vermiculite and drying at ambient air conditions, with shelf lives of more than 6 months.
At the full scale, the project advanced in two directions in particular. The first one consisted of the development of good techniques that enable an unabridged characterization of the catabolic capacity of microorganisms present at contaminated sites. In particular a catabolic gene chip and the corresponding protocol were developed and extensively tested on different sites and by different people. It could also be shown how communities in different contaminated sites develop over time with the same contamination type, which is an important requisit to predict which functions may be necessary to complement by bioaugmentation. Specific 'second generation' bacsins, those to be particularly abundant at a contaminated site and therefore presumed to have superior survival capacities, were isolated. For some of those it could be established why they would be more successful (e.g. denitrification capacity), but (ironically), most of them were not so superior after all. This suggests that it is only important to find the right ecological niche characterization by pretesting, in order to ensure survival and activity of an 'appropriate' bacsin. Indeed, most of the initially selected bacsins could be used in controlled randomized field trials and were shown to survive and aid in contaminant degradation. In particular, treatment of a petroleum-oil contaminated area by combined planting/bacsin co-inoculation, and treatment of chlorophenol-contaminated soils with vermiculite-formulated bacsins, were successful examples of the concepts developed in the project.
The project hosted a regularly visited web-site and made extensive efforts for disseminating results and concepts. Particular highights for dissemination included an international student summer course in July 2011 in Lausanne with hands-on methods, exercises and seminars, and a final international symposium on rational use of bacteria for bioaugmentation in April 2012 in Amsterdam.

Project Context and Objectives:
The central idea behind the BACSIN project is to improve the rational use of bacterial catalytic activities, in particular for targeted pollution treatment, removal and prevention. Whereas in many cases specifically isolated bacterial strains for targeted biodegradation have been used, their success is largely unpredictable and annectodal. The BACSIN project thus attempts to gather the basic knowledge necessary to understand environmental survival programs in a number of predefined bacsin isolates, by using frontline systems biology, single cell microbiology, classical genetics and modeling. Starting from basic knowledge, a transition will be made to describe and predict where possible the ecological behaviour of such strains in real complex environments and contaminated sites. In addition, the project focuses on contaminated sites directly, trying to identify the key catabolic players at such sites, in order to compare them to the same level as the set of predefined Bacsins from the beginning of the project. By doing so, we hope to obtain generalized knowledge on the types of survival programs that are most appropriate for strains in particular environments (soil, water, plant surface).
At the same time, the BACSIN project would like to improve the rational use of bacteria or bacterial consortia with useful catalytic properties in environmental settings. Rational use could mean inoculation of single strains in particular formulations, development of specific consortia, utilization of bacteria in combination with plants, or rational stimulation of specific pre-existing in situ bacterial activities. The application part, i.e. how to preserve strains with useful catabolic properties, how to apply them in the field for optimal success, how to trace their activity, and how to monitor the activity of the whole community they will be introduced into, is an important focus of the project as a whole, and subject to specific work packages.
BACSIN focuses on combining advances in in situ detection technologies, of the availability of hundreds of complete microbial genome sequences, of new ‘-omics’ approaches, with new single-cell experimental methods and computational simulation models, to unravel and rationally exploit the catalytic properties and survival modes held by the various natural bacterial isolates and communities for remediating and mitigating environmental pollution. The goal of this is to derive more general principles and concepts of microbial management in complex systems (‘ecological engineering’), which the consortium hopes will transcend the temporary focus of pollution control.
Detailed understanding and rational exploitation of bacterial catalytic properties for the environment
The vision of the BACSIN project is to acquire a detailed understanding of the effect of the environmental setting on bacterial behaviour and to rationally use bacterial activities in remediating and mitigating environmental pollution. Rational use means knowing how and when to use certain bacteria for catalytic purposes and when to avoid using others. It also implies knowing how to ‘diagnose’ environments and strains for catabolic and stress survival properties, how to predict and stimulate possible pollutant transformations and metabolic fluxes within a microbial community, and suggesting where to complement in situ activities via release of specific bacterial strains. Finally, rationality means that we should be able to predict how bacterial communities may adapt and evolve under influence of pollutant stress and when appropriate even stimulate this process.
Scientifically speaking, the challenge for BACSIN is to understand why certain bacteria survive well and remain active in a particular environment, whereas others do not. How are catabolic activities embedded within the complex metabolic and signalling network of even a single bacterial cell and within the changing boundary conditions of the outside environment and competing microbial activities in the community? Technically, BACSIN strives to be able to exploit bacterial catabolic activity, propose solutions for waste and wastewater treatment, pollution control and prevention, which in some cases means introducing catalytically active bacteria ‘off-the-shelf’ in order to achieve targeted pollutant removal. In other cases it means evaluating the on-site microbial degradation properties, identifying the responsible community members followed by targeted in situ stimulation or isolation of strains active in pollutant degradation, their cultivation and subsequent re-introduction into the site. Experimentally, we will face the problem of tracing the activity of single bacterial cells in complex environments and communities.
Objectives of BACSIN
The main objectives of BACSIN in the context of the above are:
1. To generate a knowledge base of bacterial gene expression programs and of subsequent physiological activity with particular relevance to biodegradation under real stress conditions to such a detailed level as to allow rational and predictable application in the environment
2. To categorize the effect of environmental boundary conditions on bacterial behaviour of both individual cells and populations and within the context of in situ mixed microbial communities to a level that allows rational engineering optimization
3. To develop bacterial formulations in physical forms from low value waste materials that allow large-scale environmental inoculation procedures and to demonstrate the directed application of such formulations to achieve targeted removal, treatment and prevention of recurring and notoriously difficult pollutants.
In order to achieve these goals, the consortium has chosen an iterative approach of testing so-called BACSINs (i.e. bacterial strains that we think have potential useful properties for environmental release and biodegradation) in four poles of activity: basic knowledge acquisition - lab conditions - field conditions - formulations (Figure 1). BACSINs not only have to be potentially useful and promising strains for biodegradation, but need to have their genome fully characterized. As the project evolves and more knowledge on general BACSIN properties become available, the initial strains will become replaced by newer environmental isolates that have been specifically isolated on the basis of general survival and catabolic activity characteristics.
(Legend Figure 1. Schematic representation of the poles of activity in the BACSIN project.)
Key scientific objectives and translation into work packages
WP1. General methods development and standardization
The main idea in this WP is to develop and standardize a number of methods for the whole consortium that will enable to obtain the generalized basic knowledge on the BACSINs. For example, appropriate methods for inducing catabolic and environmental stresses in laboratory systems are developed; physiological growth modes for the different BACSIN strains are being standardized among partners (shaking flask, continuous culture, or biofilm flow). Appropriate assays to measure catabolic activity, general physiology and stress parameters in BACSIN strains will be standardized and instructed among the different partners, and finally, various microbial community composition methods will be evaluated, that are to be used in field studies.
WP2. Genome wide BACSIN behaviour in response to environmental stresses with particular relevance to catabolic networks
The second WP - one of the basic knowledge packages - will focus on characterizing the genome-wide catabolic and stress response pathways in the selected BACSIN strains as a function of target pollutant exposure and stresses mimicking environmental survival. First this will be done on the preselected five BACSIN strains - later on newer environmental isolates. Main methods here are transcriptomics and proteomics, additionally complemented by membrane composition studies. Key stress and survival factors common to all BACSIN strains and those unique to each will have to be identified and their role understood. The potential cellular factors determining optimal environmental activity of BACSIN strains and those limiting survival and pollutant transformation capacity are to be generalized.
WP3. Traits governing environmental survival and adaptation.
WP3 will operate in extension to WP2 as to understand which gene functions in BACSIN bacteria determine environmental survival and stress resistance. Describing as in WP2 is not enough: the physiological changes taking place between regular rapid laboratory growth and phases of slow growth, dormancy or cell death have to be understood. Specific regulatory and signalling networks governing physiological adaptability at the transcriptional, translational and post-translational level have to be dissected here, to provide the basis for understanding and better selection of useful BACSINs.
WP4. Interpretive and predictive modelling of catabolic activity under environmental stress
The fourth WP is set-up to derive regulatory models that can help to interpret the observed micro-array and proteomic stress and catabolic datasets, and that can aid to predict environmental stress induction and catabolic activity of BACSIN strains. This WP will also develop models to predict the population development (biomass) of BACSIN strains in the environment as a function of pollutant availability, and will try to model catabolic networks in natural communities under conditions of stress.
WP5. Construction and standardization of afp reporters for in situ stress and catabolic measurements
To make the transition to observing and understanding BACSIN behaviour in more complex settings than the lab, this WP is focused on constructing inducible autofluorescent (afp) reporter fusions in BACSIN strains. By means of reporting genes it is possible to monitor stress response, cellular survival or catabolic activity. Tagged BACSIN strains with constitutively produced afp proteins allow to follow environmental survival and microscale location.
WP6. Environmental behaviour of pure BACSINs with particular relevance to catabolic activity
This WP is one of the core activities focused on studying which factors in the environment determine catabolic activity of introduced BACSIN strains. Here we will try to understand whether survival and stress resistance expression of BACSIN strains in the environment are different than under laboratory conditions. A number of near-field experiments is proposed here: to introduce BACSIN strains in microcosms, follow their survival and in situ activity for degrading the target compound.
WP7. BACSIN survival and activity in the phytosphere
A specific role is reserved for BACSINs in the plant environment. This WP will study catabolic gene expression as a function of being on a plant leaf or plant root systems, will study the unique stress conditions influencing BACSIN behaviour on plant leaves and the specific bacterial factors favouring adaptation to these conditions. We will also try to take advantage of plants by determining whether pre-exposition of BACSIN strains to the phyllosphere (‘priming’) can improve their stress resistance and survival, and possibly their catabolic activity, and to select agricultural and forestry plants particularly prone to host BACSIN strains on their root systems. Finally, we will study here whether specific BACSIN plant root combinations can lead to improved stress resistance, survival and catabolic activity.
WP8. Molecular diagnostics for natural catabolic, stress and survival functions
Here the real field scale will be addressed: the diversity of stress and survival determinants occurring in contaminated sites will be identified, we will try to understand the genetic basis for what makes a key successful catabolic player at contaminated sites. Hereto molecular diagnostics for establishing the catabolic potential (‘catabolome’) at contaminated sites or samples must be developed and tested. The presence and activity of BACSIN related strains in the phyllosphere (i.e. on plant leaves) will be examined, and studied whether and how they have superior survival and stress resistance.
WP9. Rational control of native communities or consortia
WP9 will focus on the idea whether not only single species can be managed in the environment, but perhaps whole bacterial ecosystems, with useful properties for environmental restoration. Hereto, we will focus on understanding which environmental and geographical conditions determine prevalence of different natural isolates of the same BACSIN genus at contaminated sites. We will try to identify how native communities at contaminated sites cope with environmental and pollution stresses, will study, model and predict hierarchical metabolic fluxes in a community at a contaminated site as a function of catabolic capacity, pollutant types and concentrations, and prevalent stress conditions. Finally, we will test whether potentially missing catabolic functions at a contaminated site can be identified and complemented via BACSIN strains or via evolutionary adaptation.
WP10. Improved bacteria survival in formulations
This final scientific WP aims to improve the survival of BACSINs in formulations, for application purposes. Here we will try to understand the physiological stresses influencing survival and catabolic activity during BACSIN formulation and storage procedures, we will select for mutants with improved survival characteristics due to membrane composition changes, and will attempt to influence strain survival via induced heterologous Usp complex formation. We will test very directly the effects of different formulation conditions on improved BACSIN activity and survival in the final product, and will explore the effects during cellular resuscitation from BACSIN formulations. Product application and dispersal methods will be improved and validated on a field-scale bioremediation effort.
WP11 Dissemination
The dissemination WP serves to transfer knowledge and technology at different levels and throughout the duration of the project. Most importantly, the consortium strives at producing excellent science, which will be disseminated in form of peer-reviewed publications that must be able to withstand scientific scrutiny and criticisms from the external community of scientists in the field. Secondly, results from the project will be disseminated in form of deliverables (reports and prototypes), most of which are accessible to the public, but which have not undergone peer-review. Third, the project will attempt to promote its basic ideas and findings in form of popular press articles, radio and television interviews, or web-sites. As a fourth aspect, the project aims to transfer its ideas, techniques, protocols and knowledge in forms of hands-on courses to young scientists, both internally and externally to the project. Finally, the project plans an open international symposium at the end of the project phase, to discuss the field of bioremediation and bioaugmentation in the context of the main results obtained by the project.

Project Results:
Selection of first and second generation BACSINs (WP1)
From the onset the project agreed to work on a number of bacterial strains with potentially interesting catabolic potential, of which the genomes had been sequences and were available, and which potentially differed in their ecological niche. We felt this was important in order to increase the chance of observing patterns in catabolic activity related to environmental stress response that would allow generalized and not just strain-specific knowledge. On the basis of these considerations, and others - such as, absence of known pathogenicity, the project selected from the beginning five bacteria to work with.
The first was Pseudomonas putida mt-2/KT2440, a soil-dwelling microbe that is considered an environmentally safe organism, that attaches readily to plant roots, has been widely studied, that can be used as host (KT2440) to accept other catabolic plasmids, and that naturally carries the TOL plasmid (strain mt-2) which enables degradation of toluene, m- and p-xylene.
The second strain was Arthrobacter chlorophenolicus, an organism isolated from soil that degrades methylsubstituted phenols and chlorophenols, and was expected to survive well. The third strain was Sphingomonas wittichii RW1, a bacterium isolated from the river Elbe on its capacity to degrade unsubstituted and chloro-substituted dibenzofurans and dibenzodioxins. The organism had been applied to microcosms for bioaugmentation purposes but with varying success. The fourth strain was Desulfitobacterium hafniense, an anaerobic strain with dechlorination activity. Finally, the consortium started off with Alcanivorax borkumensis, a marine bacterium that specializes on degradation of short and long-chain alkanes. With these five organisms, the project aimed to cover both freshwater and marine environments, aerobic and anaerobic conditions, soils and the phyllosphere, and different target chemicals (BTEX, PAHs, chlorinated compounds and petroleum oil).
As will become apparent in the following, over the course of the project a number of 'second generation' bacsins were added to the list of strains above. These strains were isolated from contaminated sites on the basis of their presumed superior survival characteristics and/or catabolic properties. A number of other potentially interesting bacsins was added as well, for example, to complement the series of anaerobic organisms. Wherever possible, the genomes of such second generation strains was determined during the project phase, and the strains analyzed in the same 'pipeline' of stress conditions. The second generations bacsins thus consisted of: (i) Sphingomonas sp. LH127, a polycyclic aromatic hydrocarbon degrader with superior survival in soil (see further below) and complementing the range of target compounds (phenanthrene, naphthalenes) compared to RW1 (dibenzofurans); (ii) P. putida BIRD1 and DOT-T1E, organisms with superior root attachment and solvent resistance, respectively; (iii) Thauera aromatica, an anaerobic BTEX degrader; (iv) Pseudomonas veronii YdB and YdBTEX, to isolates from the Hradcany contaminated site (see further below in WP8), and (v) Arthrobacter sp., an isolate from plant leafs with superior survival degrading 4-chlorophenol (see further below). The genome of the last strain, however, did not become available during the project period.
Standardization of stress exposure conditions (WP1)
One of the main first tasks for the BACSIN project, to which mainly WP1 was devoted, was the standardization of experimental conditions under which to grow the BACSIN strains, the conditions under which to expose strains to the stress conditions, to analye their responses and stress-related effects on their physiology. Without such standardization, we felt we would never be able to compare and draw conclusions on specific environmental behaviour, for example, by genome-wide transcriptomics studies. Furthermore, in order to reduce environmental complexity we designed standardized laboratory stress exposure conditions, to which (ideally) all strains could be subjected, so as to obtain a very standardized dataset. In first instance, we thus decided to focus on (i) drought stress (e.g. in soils or on plant leafs), (ii) toxicant stress (e.g. the contact to high concentrations of the target compound or co-contaminants), and (iii) nutrient stress (e.g. the absence of N, P in the environment).
Drought-related stress was translated to classical aqueous batch cultures in which the water potential (water activity) was reduced by the addition of salt or swelling agents (PEG8000). Individual strains would thus be grown with decreasing water potential and the point where no more than 20% reduction in growth rate was observed, would be taken as the maximum level to study the stress response. Going beyond this would kill too many cells and would result in nonsense data from e.g. transcriptomics (in WP2). In essence, all bacsins were indeed subjected to this standardized drought regime. The second form of standardized drought stress was to filter cells onto membrane filters, expose them to different humidity conditions for a short period, which would allow to observe the reactions in their cellular metabolism. This second form did not work well for e.g. Sphingomonas wittichii and, therefore, was abandoned for others. The third standardized drought stress consisted of mixing cells with dry sandy soil and recovering them after a short while (1 h), to detect any responses. This worked well but was a bit more elaborate protocol and could not (yet) be carried out with more than three bacsins. As fourth method, we used exposure to drought and UV on plant leaf surfaces, which turned out to be very interesting, but - due to the more complex nature of the protocol, could not be easily performed on all bacsins. Finally, we used drying processes in protective materials (vermiculite) as a standard protocol for drought stress (see WP10).
As a standard toxicant stress, we used exposure to chlorinated phenols with increasing concentrations. Similar as for the water activity stress, the point where no more than 20% growth rate decrease occurred was taken as the maximum allowable stress level to study the genome-wide transcription response. In addition, a flow cytometry protocol was developed to study toxicant stress on exposed microbial populations using live-dead staining and ethidium-bromide efflux. Toxicant or stress-induced membrane damage was standard measured in specific fatty-acid methyl ester-derived protocols. Due to different sensitivities of strains, the toxicant stress effect could not be tested for all bacsins (or tested at growth rate level, but not for transcriptomic response). Finally, in particular for A. borkumensis, a protocol was developed to test nutrient stresses - assuming that this is the major stress occurring in marine systems during oil spills (far too little N and P to compensate balanced growth given the extraordinary large quantities of C available).
Apart from the standardization of stress exposure protocols, the consortium also standardized various cultivation techniques, such as batch culture mode, biofilm growth mode, plant leaf and plant root inoculation, and an inoculation-recovery protocol for soil. Most of the stress protocols were tested with one or more bacsins of the first and second generation in conjunction with determination of the genome-wide transcriptive response to the stress conditions (WP2 and next paragraph). All partners were instructed in the use of these methods and cultivation techniques. Standardized protocols were provided on a central project-internal web-server so that everyone would use the same methods. A number of partners also provided 'services' for the standardized exposure testing, for example, water activity test, toxicant-exposure text, the plant-leaf exposure test, the plant root survival test, or controlled drying.
Standardizing molecular techniques (WP1)
Further development in this work package consisted of standardizing a number of central techniques that were of use to all. This included the following: (i) A service for analysis of bacterial membrane composition as an biomarker for various stress effects, (ii) A custom “catabolic array” for detection of the presence and expression of key catabolic genes in the bacsin strains; (iii) Quantitative PCR detection of marker gene presence or expression in bacsins; (iv) microbial community analysis methods based on 16S rRNA gene amplification and diversity.
Of particular importance for further work in WP2 was the standardization of genome-wide gene transcription analysis by microarray. After some discussion, the consortium settled with the Agilent platform that enables small quantities of custom-made oligomer chips, such as the 8X45K or 8X60K formats that include eightfold repeated arrays on a single glass slide. Designs were made for all first generation and a number of second generation bacsins, which were ordered and processed. Centralized protocols for RNA isolation, cDNA labeling and hybridization were produced. Several partners acted as service platforms for others to perform the hybridizations and extract the raw data. Primary data analysis was developed (e.g. COG or GO-analysis), which was later complemented with web-based tools such as DAVID, by which cross-species transcriptomes can be interpreted and commonalities in responses of bacsins to the same stress types can be extracted. By contrast, proteomic methods, which were planned to complement transcriptome analysis, could not be properly developed and tested during the project phase. Only in the final period was a first attempt made on D. hafniensis.
A second set of tools of particular importance for the project as a whole consisted of catabolic gene detection. For catabolic gene detection and expression (both of individual bacsins or communities) we settled on using quantitative PCR and use of a catabolic microarray chip. In a later stage of the project newer development such as RNA sequencing were introduced. Selective primers for key catabolic and housekeeping genes (e.g. for 16S rRNA) in almost any of the first generation bacsins were developed and tested, both on DNA/RNA from pure cultures and on environmental DNA/RNA. This also required development and testing of RNA isolation procedures from complex environments (soil, plant leafs, clay minerals).
Development and standardization of the catabolic gene array was a major technical tour de force for the project. Although similar catabolic gene arrays exist, they suffer from lack of proper internal controls for hybridization efficiency to non-perfect probes. After a large number of trials finally a proper protocol could be established, which has been extremely helpful and important for analysis of bacsins and for screening the presence of (known) catabolic genes in contaminated sites. The catabolic chip was extensively tested, also during the US-EC student course in Environmental Biotechnology in July 2011. Further probes were designed and included in the catabolic chip. The BACSIN project believes that this catabolic array is a very powerful tool to pre-screen contaminated sites for potential catabolic activity, and to possibly detect the key members of the community (see further WP8).
Finally, this work package standardized techniques for microbial community analysis. Along the complete project techniques such as T-RFLP on amplified 16S rRNA genes were applied successfully. In a later phase, protocols were developed for fast and cheap sequencing of the V5-V6 variable region of amplified 16S rRNA genes by using Illumina HiSeq technology. Most of these techniques were validated multiple times and by different users, were introduced and tested in such prominent courses as the US-EC Environmental Biotechnology course in July 2011.
The main tasks of this activity were 'finished' after the second period, but regularly updated and instructed to all BACSIN participants. This activity was very fruitful and important for the project, and most of the planned tasks could be carried out or even improved upon. In short: appropriate methods for inducing catabolic and environmental stresses in laboratory systems were developed, and physiological growth modes for the different BACSIN strains have been standardized among partners. Appropriate assays to measure catabolic activity, general physiology and stress parameters in BACSIN strains were standardized and instructed among the different partners. Various microbial community composition methods were evaluated and used in field studies.
Genome-wide responses of BACSINs to stress (WP2)
The major effort in the second work package was to obtain a global understanding of catabolic and stress programs in the selected bacsins. What we attempted to understand was not only how one strain would behave as a function of target pollutant exposure or stress in the environment, but to compare such behavior between different bacsins. From such a strain-strain comparison and in relation with survival data we could possibly learn how strains operate different survival programs, couple stress to catabolic activity, and which markers could be informative to characterize 'good' environmental survivors from 'bad' ones. First we focused on five preselected bacsins, later this was complemented with a number of second generation bacsins. In short, it is fair to say that we generated an enormous amount of data very rich in all sorts of details, which will take much longer to analyze than we could possibly do within the lifetime of the project. We have observed important strain-to-strain differences but still lack global understanding of which factors would be key to its general survival capacities. In all cases, a large proportion of the differentially expressed genes do not have a known function, which limits our understanding of their molecular function.
The most successful activity in this part of the work consisted of the genome-wide analysis of gene expression in bacsins cultured under different conditions using microarray technology. In fact, we realized that the behavior of even wild-type strains could be analyzed in short time using the pipeline 'genome-sequencing, design of microarrays, exposure assays and microarray hybridizations'. Because of the ease for experimental repetitions and the repeated array format (8 repeated arrays on a single slide), excellent data quality could be obtained that permitted in depth statistical analysis. All but one bacsin (A. borkumensis, for which water activity decrease made no sense) were thus tested in the standardized liquid suspended culture protocol using salt and PEG-decreased water potential. This activity implicated basically all partner laboratories and several labs performed service for microarray analysis with strains from others. Current results suggest conserved, but also strain-specific regulons controlling the response to solute and matric stress. For example, results show that all bacsins induce gene clusters that have previously been known to be implicated in drought stress (e.g. synthesis of osmoprotectants), although the actual osmosprotectant may be different from one strain to the next. We found important differences between all bacsins with respect to their response to water activity reduction. All strains reduced growth rates in response to solute or matric stress, but not at the same water activity threshold. We also detected important influences of water activity reduction on catabolic gene expression (different than just growth rate reduction). Solute and matric stress affected membrane composition differently, leading e.g. in the case of S. wittichii to a higher degree of saturation after perturbation with sodium chloride but to a lower degree after perturbation with PEG8000. Finally, it appeared that cellular reactions to solute and matric stress themselves were very different, and on their turn very different from the response of cells to e.g. dry sand (Figure 2). Therefore, even though the matric and solute stress are the most easy standardized experiments between different bacsins, this response is not an exact representation of the behaviour in a dry soil.
(Legend Figure 2. Representation in Gene Ontology terms of the common and different responses of Sphingomonas wittichii RW1 cells pregrown in salicylate (Sal) or on dibenzofuran medium (DBF), and then incubated in sandy soil with or without DBF. Note, for example, how cells turn off their energy metabolism - intense red boxes, when being introduced into soil without contaminant that they can utilize ('Sand vs Control Sal').)
Interestingly, some bacsins like A. chlorophenolicus almost did not react to matric stress at all, but dramatically decreased its cell size in response, which may constitute a specific program for this bacterium to enter a protective state. Biofilm growth experiments performed with strain LH128 showed that matric and salt stress do not affect survival of the cells but changes biofilm structure. Perception of a stress, therefore, is not necessarily the same in suspended culture or biofilm. Results further showed that the overall transcriptional response to the stress is specific for the type and duration of stress. Gene expression profiles change drastically immediately after a short stress exposure, and acute stress leads to immediate induction and repression of a large number of genes. In contrast, most strains ‘adapt’ to the same stress under prolonged conditions and control gene expression to that under a non-stress situation. Notable exceptions to this were also found; strain LH128 adapts to continued matric stress but not to long-term solute stress. In contrast, P. putida mt-2 was better able to adapt to salt than to matric stress. Likely salt stress specific adaptations were up-regulations of genes coding for potassium and glycine betaine transport systems. Also intriguing is the preliminary finding that second generation bacsins do not necessarily display better resistance characteristics in the solute and matric stress test. Their environmental abundance at particular contaminated sites may thus have other specific causes than resistance to drought.
In separate set of experiments, we evaluated the strain-specific responses to toxic chemicals, including the target pollutant itself. Interestingly, both S. wittichii and Sphingomonas sp. LH128 showed what can be called 'parallelism' in their catabolic behaviour, implicating induction of several seemingly redundant metabolic branches simultaneously. In contrast, P. putida, A. borkumensis, and A. chlorophenolicus show more 'linear' singular pathway inductions. Acute exposure to dbenzofuran in S. wittichii also caused temporarily increased expression of several stress pathways, suggesting that the compound itself is stressful as well, and that the bacteria have to adapt to it. P. putida KT2440 exposure to a variety of toxicants such as TNT or formaldehyde caused multiple different stress mechanisms, with the MexEFOprN efflux pump as a shared resistance determinant. A new ABC efflux system, encoded by (PP2669/PP2668/PP2667) was also identified, which is required for chloramphenicol and naphthalene resistance. Later analysis of toxicant stress on the second generation bacsin P. putida DOT-T1E indicated a number of extremely powerful efflux pumps that enable this bacterium to withstand contact to pure solvents. In addition, several other genes involved in chemical stress, or in resistance to ultraviolet light exposure, were identified.
Nitrogen and phosphate, salinity, heat stress (40°C) and phenol were identified as major stress factors relevant for efficient oil degradation by A. borkumensis, especially with regard to the magnitude/speed of oil degradation at excessive/insufficient nitrogen and phosphate levels. EC25 and EC50 values were determined for a broad range of chemical and physical stress factors, including addition of 1-octanol, toluene, m-xylene and benzene, temperature, salinity, or depletion of phosphate and nitrogen. For all these conditions microarray analysis was performed and this extensive data set was then used as input for regulatory model predictions. Oxygen availability largely controls catabolic gene expression of toluene degrading aerobic P. putida and anaerobic T. aromatica. In systems with fluctuating aerobic/anaerobic conditions, both bacteria must modulate the expression of catabolic genes depending on the oxygen availability. Results on such cycling regimes showed fast adaptability, and after two oxic cycles T. aromatica increased catabolic gene expression once oxygen was depleted by P. putida. The genome of T. aromatica K172 was sequenced and annotated. This revealed the unexpected presence of a broad host plasmid of the IncP-1 family that confers resistance to several classes of widely used antibiotics. Therefore, it was concluded that T. aromatica is not an appropriate organism for further bioaugmentation studies.
Catabolic gene expression analysis in the anaerobic bacterium D. hafniense DCB-2 showed that multiple catabolic gene operons are differentially regulated in response to different chlorinated pollutants, indicating catabolic versatility and flexibility of these dedicated degraders. Exposure to meta- and ortho-chlorophenols induces distinct cprA genes. Increasing concentrations of the o-chlorophenolic substrate Cl-OHPA act as stressor that affects population growth as well as membrane fatty acid saturation, and overall cellular activity as measured by transcriptome and proteome analyses.
In order to make a transition between pure lab studies and more applied conditions in WP6, two sets of experiments were initiated. One consisted of trying to determine the transcriptomes of cells introduced to dry sand. What so far has become clear is that complex environments induce a further large variety of responses, very different of standardized drought stress response, indicating that standardized lab conditions not necessarily mimick environmental conditions very well. Most of the additional functions for which differential gene regulation is observed, concerns nutrient scavenging, oxygen regulation and carbon availability (i.e. energy generation). The other experiment analyzed the transcriptome of cells placed under low air humidity (27%) - as a transition to plant leaves. Finally, a transcriptome comparison was made for A. chlorophenolicus cells that are prepared for drying. Interestingly, one hour drying at ambient temperature produced no measurable effect on transcriptome, suggesting that the good desiccation resistance of A. chlorophenolicus is due to its incapablity to react to this as a change (see WP10).
Molecular basis of water and toxic compound stress response in BACSINs (WP3)
To be able to go more in the molecular and biochemical details of the genome-wide gene expression 'programs' that were observed in WP2, WP3 outlined a number of specific tasks to try and create mutant libraries or define biochemical interactions. First of all this involved a genetic and biochemical research to understand the possible role of Usp proteins in controlling the physiological changes taking place between regular rapid laboratory growth and phases of slow growth, dormancy or cell death. Then for the toluene/xylene metabolic pathways of P. putida TOL plasmid we set-up a detailed study to better understand the specific regulatory and signalling networks governing physiological adaptability at the transcriptional, translational and post-translational level. Finally, for the less well-understood bacsins we intended to create large transposon mutant libraries and screen for loss of function in drought or toxicant stress pathways, which would help to understand the results of the previous transcriptome studies, and provide a basis for better identification of key survival functions. In summary, here, the Usp investigation brought useful results but did not confirm the hypothesis of their importance in controlling survival in P. putida. The TOL pathway could be dissected in further great detail and was described in logic terms, which helped to clarify its basic architecture. A number of newly evolved pathways was studied that confirmed that pollutant metabolism itself actually create (mostly oxidative) stress on the bacterial cells, which, interestingly, seems to result in higher mutations rates. Pollutant exposure, therefore, may lead to more rapid evolution of functions. Transposon libraries were generated for several of the first generation bacsins. In particular a newly designed library sequencing approach generated useful results complementing transcriptome data.
More specifically for P. putida, we focused on proteins with so-called USP (universal stress) domains, in order to study their role in the stress defense mechanism. P. putida carries the genes for six different Usps. A site directed mutagenesis strategy was developed for the generation of seamless mutants in such proteins in P. putida, so they lose the USP module but maintain the rest of the expressable sequence. All mutants and several double mutants were created, which were screened for effects on adaptation of cells to different stresses, specifically matric stress and gross chemical toxicity. All results point at USP domains playing a role in the maintenance of envelope-related functions. Importantly, some USPs play a role in the P. putida survival. Various interactions were found between P. putida and E. coli USPs, indicating that they can form heterodimers.
Tilling array technology was developed for fine analysis of the transcripts of the toluene/xylene biodegradative genes in P. putida, and their alteration under diverse physiological scenarios. The tiling array showed an important contribution of read-through expression in the xyl genes. In a genetic approach the connection between chemotaxis and stress resistance was investigated. Interestingly, results on P. putida indicate that a non-flagellated strain resists better to toxic compound exposure than its natural flagellated counterpart. The third issue has been the dissection of the connection between endogenous oxidative stress associated to biodegradation of 2,4-dinitrotoluene and the evolution of the corresponding catabolic activities. This showed, interestingly, that metabolism of toxic pollutants can generate oxidative stress, which then increases mutagenesis frequencies, suggesting that an increased pool of variants is formed under stress from which appropriate favorable mutants may become selected.
Transposon mutagenesis was successfully established for the bacsins P. putida, S. wittichii, Sphingomonas PH128, and A. borkumensis. Transposon mutagenesis revealed a number of unique new functions in P. putida that contribute to resistance to bacteriostatic compounds. In contrast, none were detected for PAH or alkane resistance. Others are implicated in tolerance mechanisms to membrane damaging agents. The gene osmY was identified as the main descriptor of the stress brought about by a sudden exposure of P. putida cells to desiccation. Other important functions were discovered that govern rhizosphere survival and plant root adhesion for P. putida, which laid the basis for understanding the root-control behavior of P. putida strains.
Transposon libraries of the other BACSIN strains were screened by regular replica plating for loss of function to salt-induced stress and water unavailability response. Here as well, a number of interesting mutants were obtained. Mutants were also recovered for A. borkumensis that determine osmosensitivity, UV sensitivity and low temperature sensitivity. For example, an ompA mutant from Sphingomonas sp. LH128 was recovered which is important for survival in soil. We showed that the ompA mutant displayed impaired growth and pollutant degradation activity in both sterile and non-sterile soil. Washing the sand improved the survival of the mutant compared to non-washed sand. The results show that ompA is an important gene for strain LH128 for soil colonization.
Although a number of mutants in hypothetical genes were recovered from library screening for loss of function to salt-induced water unavailability resistance in S. wittichii, none of them had clear links to known stress or catabolic pathways. To increase the capacity of screening for functions that are difficult to reveal on agar plates (such as desiccation stress), we designed a procedure to screen transposon mutant libraries by ultrahigh throughput sequencing. This appeared to be an extremely rewarding method. Transposon libraries as a whole could be incubated under the test conditions, and after an estimated 10-50 generations of population growth, DNA is isolated and both the insertion position and abundance of each individual transposon insertion in the library is quantified by sequencing and compared to the control condition. As proof of principle, we found more than 600 absolutely essential genes for growth in S. wittichii and an additional 400 unique genes per individual growth condition, which when interrupted by a transposon resulted in dramatic fitness loss. As an example of gene groups important for survival, under conditions of salt stress and related water unavailability most essential functions concerned cell wall synthesis or sugar polymer synthesis. In contrast, good survival in dry sand with salicylate as carbon source implicated motility, whereas long-term survival in sand implicated cell-cell communication and oxidative stress responses. All three conditions demanded nitrogen scavenging functions to be present, and pointed to essential regulatory functions. Interestingly, also several gain of function mutants were detected, some of those increasing fitness (as growth rate) by 30%. One of those was again in a regulatory gene, which when inactivated, resulted in much higher growth rates on salicylate than before. A comprehensive combined analysis was performed to link gene expression of the presumed dibenzofuran and salicylate metabolic pathways to mutant analysis in S. wittichii RW1. This showed the importance of several new unexpected metabolic branches but also, as mentioned before, other stress-related gene functions with importance for fitness maintenance during growth on aromatic compounds.
Network regulation analysis (WP4)
Apart from a descriptory understanding of possible stress mechanisms and their implications on catabolic activity in BACSIN strains, the consortium also aimed to model such networks in a more predictive way. This basis was developed in WP4, by the use of regulation analysis and microeconomic theory. Regulation analysis is a mathematical tool to establish how either microbial communities or individual species regulate changes in flux through metabolic pathways. Previously, there were two separate tools for regulation analysis; hierarchical regulation analysis and ecological regulation analysis. Hierarchical regulation analysis allowed to establish how individual species regulate flux, at the metabolic or hierarchical level. Metabolic flux analysis ideally uses enzyme activities and mass balance reactions. If regulation occurs at the hierarchical level, it allowed to determine the contribution of transcription, translation, modification of RNA and protein, and RNA and protein degradation. Ecological regulation allowed to establish how microbial communities regulate changes in biogeochemical fluxes, by changes in cell numbers or by changes in cellular activities. These two regulation analysis tools have now been integrated, allowing to use full data sets, to detail how microorganisms, and the individual enzymatic activities in them, respond to stress. It allows for the integration for information on cell numbers, proteomics, transcriptomics and metabolomics data, to pinpoint quantitatively where microorganisms make changes to bring about changes in biogeochemical fluxes (e.g. pollutant biodegradation rates). By its application to large sets of data, collected for several stress conditions and several species, it is possible to establish if there are general rules on how microorganisms respond to stress (e.g. does for example always the first enzyme in a metabolic pathway change its activity in response to stress, while all other remain constant, etc.).
Hierarchical regulation analysis mainly focused on the bacsins A. borkumensis and Desulfitobacterium and a number of literature studies. The WP faced an important delay, because of delay in producing the appropriate enzyme activity data. Technical difficulties were encountered, in particular, to grow Desulfitobacterium in continuous cultures and determine enzyme activities. On the other hand, the most successful data produced by WP2 were transcriptomic data, which are good for gene identification, but were not immediately very valuable for metabolic models. Here we investigated the use of programs that base on gene ontology to compare responses across all strains (DAVID). Attempts were thus made in a later stage to work with all transcriptomic data of one bacsin. A large number of different environmental conditions were tested for A. borkumensis that affected its alkane utilization. Interestingly, mainly nutrients directly reduce alkane gene expression but not just via reduced growth rate. The large experimental data set on Alcanivorax was used in the hierarchical regulation analysis framework, in particular to relate changes in gene expression to changes in metabolic fluxes. Changes in gene expression only for a small part regulate changes in growth rates. Also the flux through the hexadecane degradation pathway upon a change in environmental conditions is only to a limited extent determined by changes in gene expression. No impact of the type of environmental change became obvious. Large standard errors, which may relate to the large experimental variation in microarray signals per gene, made it difficult to draw unambiguous conclusions on how hexadecane degradation flux is regulated by changes in expression of individual genes, and whether different regulatory responses are observed in response to changes in exposure conditions. Overall, the data suggest that there is no simple linear relationship between gene expression and growth rate.
Attempts were also initiated to build a metabolic-scale model for dibenzofuran degradation of S. wittichii based on the B-NICE computational approach. This model assumes strict thermodynamic considerations for metabolic intermediate reactions, which can be minimised and linked to genome data. This model could so far not be finished.
As an alternative, the toluene/xylene pathway of P. putida was used for microeconomic theory-based analysis of carbon catabolite repression. This theory can explain the costs and benefits of mixed carbon source degradation. Interestingly, the model suggests that cells focus on using the substrate that provides the most utility for their limited energy budget.
In addition, the model for ecological regulation analysis was extended to apply to dynamic systems (e.g. oscillatory or ‘chaotic’ systems), besides systems in steady state. THe ecological regulation analysis was in first instance verified on literature data. The time-dependent ecological regulation analysis enables quantification of changing fluxes in dynamic systems, and to determine whether such changes are due to cell numbers or cellular activity, or a combination of both. The model application is somewhat limited to the relative flux occurring in the experimental systems and the accuracy at which these changes can be quantified. In the final period the ecological regulation analysis was carried out for mixed sediment cultures with Desulfitobacterium and Sphingomonas. Modeling suggested that cells after reintroduction may enter into an active but non-culturable state, from which they slowly recover. We discovered that the performance of the PCE-respiring Desulfitobacterium Y51 was affected by the composition of the fermenting community. PCE degradation was most rapid with enriched communities under denitrifying or iron-reducing conditions. Higher PCE degradation activity was associated with higher Desulfitobacterium numbers.
in conclusion, we can say that several useful modeling approaches were initiated that helped to understand the behavior of single strains in complex situations (e.g. multiple carbon sources) or conditions, involving many other members of a microbial community. The modeling could not yet be used in a predictive way, which could then be tested experimentally. An important issue to maintain for future activities is that experimentalists try to obtain exactly those parameters that modelers require.
Autofluorescent gene reporters (WP5)
Activities in WP5 were concentrated on making a transition to BACSIN behavior under environmentally relevant situations. As we initially assumed that it would be quite difficult to apply genome-wide assays to complex environments, the strategy was to reduce the complexity of the 'complete' strain behavior to one or two single key 'markers' that would be representative for (part of) the complex behavior. In order to do so we envisioned the use of autofluorescent reporter proteins that could be coupled to inducible stress or catabolic pathways in the selected BACSIN strains. Cells with the reporter proteins could then be introduced in more complex systems, after which they could be recovered and the produced signals compared to those under standard conditions. In another scenario, we envisioned reporters that would show reproductive success of an introduced bacsin.
As many of the bacsins were wild-type strains without previous genetic tools available, a major portion of the activities in WP5 concentrated on (i) providing the genetic tools for labeling of the ‘wild-type’ bacsins, and (ii) producing relevant labeled strains. Not unsurprisingly, some wild-type strains (e.g. S. wittichii) turned out to be extremely refractory to genetic manipulation, or turned out to be quite unstable. Also the selection of appropriate 'key' promoters for the behavior of the strains was not completely without difficulties. On the other hand, new genetic tools came available by activity of notably partner CSIC, and most BACSIN strains of interest were finally tagged with fluorescent markers, either constitutively expressed, coupled to catabolic promoters or, in a few cases, coupled to stress promoters.
For S. wittichii, Sphingomonas LH128 and A. borkumensis, appropriate plasmid and transposon tools were successfully tested for afp delivery. In terms of promoter selection, again P. putida was the most advanced, and both catabolic and stress promoters could be identified. On the other hand, P. putida KT2440 was not the most informative bacsin in terms of environmental catabolic capacity, and many of the afp reporters were not tested in environmental settings.
A tetradecane inducible promoter-egfp fusion was produced for A. borkumensis. The strain was extensively calibrated with short and long-chain alkanes in liquid cultures, gel gradients and in situ microcosms (Figure 3). Interestingly, the response to crude oil was much higher than expected from tetra- or hexadecane alone, suggesting that unknown compounds stimulate this degradation pathway further. Microcosm studies further showed an extremely strong effect of nutrients shortage on expression of the catabolic fusion. In contrast, addition of typical 'slow-release nutrient solutions' caused good induction of the reporter, but still the strain did not degrade crude oil faster. This suggested that such slow-release solutions may itself contain compounds that induce one particular alkane degradation pathway, but which is not the most effective for crude oil degradation.
(Legend Figure 3. Strong fluorescence (right image) of an Alcanivorax borkumensis bioreporter strain after contact to crude oil. Quantification of signal intensities on images like these is an easy method to measure bioavailable alkanes for oil-degrading bacteria.)
Two potentially inducible promoters for dibenzofuran could be selected for S. wittichii, and were, together with one further 'constitutive' dibenzofuran promoter, successfully cloned in plasmids that could be relatively stably maintained by strain RW1. One of those was quite well inducible by growth on dibenzofuran, but less so with salicylate or phenylalanine. One presumed stress promoter was also cloned, but which did not appear to be particularly induced during any of the tested conditions so far. The RW1 reporter strains were then tested in sandy soil microcosms with dibenzofuran and/or PAH contaminated material. We could show that the reporter strains survive and grow when dibenzofuran is present. Interestingly, the reporter activity indicated that native microbial communities compete with the introduced RW1 strain not for dibenzofuran degradation, but for some of the intermediates that the strain produces (and apparently releases) during dibenzofuran degradation. Since these intermediates are lost for RW1, it produces less biomass and finally disappears.
The promoter region of the PAH catabolic gene phnA1f of the second generation bacsin Sphingomonas LH128 could be identified and was fused to a promotorless gfp to construct a catabolic reporter variant of Sphingomonas LH128. Interestingly, induction of the gfp gene was observed in response to the presence of bioavailable phenanthrene at concentrations up to 1 mg/l and an increase was observed with increasing nominal concentrations of phenanthrene. The reporter system was successfully used to identify historically PAH contaminated soil. Also, the promoter of the previously mentioned ompA gene was cloned and turned into an LH128 bioreporter, but this strain could not yet be further applied.
As an alternative strategy we focused on reporters for 'reproductive success'. The principle of such reporters is not induction of expression of an autofluorescent reporter protein under specific conditions, but dilution of a pre-established maximum amount of autofluorescent reporter in the cell. Reproductive success reporters thus have a genetic control that allows (i) maximum induction by an artificial compound, after which (ii) the promoter is further silence and no new production of afp occurs. In such a way the cells are maximally loaded with reporter protein but once they divide will dilute the existing amount of afp protein over the two daughter cells. By following the signal intensity in single cells, an extrapolation can be made as to how many times such as cell has divided in the particular environment where it is in. The system was successfully established and tested for a strain surviving on plant leaves, which was shown to be a remarkably heterogenous environment for cells. Inoculating bacteria on plant leaves (typically by spraying a fine mist) will result in some cells landing in hostile places where they cannot multiply. Future efforts will have to be directed to try and apply the reproductive success bioreporter to strains targeted for soil survival and activity.
A number of constitutively labeled reporters were made for second generation BACSIN strains, such as Pseudomonas veronii, which are helpful to follow and quantify the strains in microcosm studies. The principle of this was tested extensively in the 2011 US-EC course in Environmental Biotechnology, where students bioaugmented microcosms with contaminated material from Jonction, Geneva, with a variety of bacsins (S. wittichii RW, P. veronii, B. sartisoli) that were gfp or mCherry-labeled. Population development of the strains in the soils was then followed by re-extraction of cells at different time points and flow cytometry to detect fluorescent cells.
In conclusion, we can thus say that WP5 was quite successful, even though not all bacsins could be completely equipped with both catabolic and stress-promoter constructs. The reporter gene concepts and constructs were used both for a general measurement of pollutant bioavailability in contaminated samples and in microcosms to follow the fate of the introduced cells.
Survival and catabolic activity of BACSINS in microscosms (WP6)
The main idea of the work in this work package was to make a transition from the laboratory experiments to microcosm and near-field experiments, and to test whether the performance of bacsins as seen in the more fundamental work packages would hold true for more complex conditions. Extensive studies were performed hereto with A. borkumensis in seawater microcosms, with S. wittichii RW1 and Sphingomonas sp. LH128 in soil and wastewater microcosms, with A. chlorophenolicus in soils and pots, with P. putida KT2440 in contaminated soils and wastewater microcosms, and with Desulfitobacterium in a variety of anaerobic microcosms. A large amount of new information became available that will increase the chances of predicting successful survival and activity of bioaugmented strains in particular environments. Results also pointed to less expected but important effects, such as competition for released metabolites formed by the primary introduced bacsin by endogenous bacteria, which reduces the population development of the bacsin. Most importantly are good test systems by which survival and activity can be pre-assessed.
The behavior and performance of A. borkumensis was studied when inoculated in seawater microcosms containing either unpolluted seawater and seawater polluted with heavy oil (Bunker C). It was shown that the addition of fertilizers resulted in accelerated oil emulsification and degradation. Bioaugmentation produced the double amount of biomass and accelerated the degradation of Bunker C up to 66%. The Gfp-labeled oil-inducible A. borkumensis reporter strain developed in WP5 was also tested in such microcosms, which confirmed the behaviour of the wild-type strain in degrading the bunker C crude oil, while producing fluorescent light that can be measured simultaneously. Also marine beach microcosms were performed with or without oil pollution, and both with gfp labeled and unlabeled A. borkumensis. Augmented microcosms degraded oil faster, but with higher rates in the seawater than in beach material. We can thus conclude that it is possible to apply A. borkumensis for specific accelerated clean-up of crude oil contamination in the seawater when the appropriate culture measures are taken into account (e.g. nutrients), but less so in beaches where the oil seeps into the underground easily.
We tested extensively the effects of soil moisture content, soil types and soil nutrients on the performance and behavior of the PAH-degrading BACSIN strain Sphingomonas sp. LH128, both under sterile (pretreated soils) and non-sterile conditions. First, two different sterile soils (a loamy and a sandy soil) were studied, which were artificially contaminated with phenanthrene. In the loamy soil, numbers of colony forming units directly declined drastically after inoculation and the size of this decrease depended on the moisture content (interestingly, it was later discovered that LH128 cells go through a period of forming viable but non-culturable cells, which explains this drop in CFU but not viable cells). Under all tested moisture conditions, except for the driest one (0 % water), LH128 recovered its population size and grew at the expense of phenanthrene. The extent of phenanthrene degradation was significantly different between soils with different moisture content and could be linked to CFU numbers. Most phenanthrene was degraded in microcosms with 16 % moisture content where also the highest CFU numbers were recorded. In soil microcosms containing a higher moisture content, less phenanthrene was degraded possibly due to diffusion limitation of oxygen resulting in lower cellular activity and/or less growth. In the sandy soil, as in the loamy soil, after 2 days, CFU numbers were below the detection limit in all conditions tested. However, after 5 days, LH128 started to grow in the soils with the highest moisture content (19 and 38 % water). Up to now, only in the wettest conditions, significant phenanthrene degradation was observed (~30 %). The data showed that the initial behaviour of strain Sphingomonas sp. LH128 upon inoculation in soil strongly depends on the soil moisture content. Although initial 'survival' seems low the strain recovers in most cases and degrades PAHs except when absolutely no water is available. Preculturing and plating conditions to recover cells from the inoculated soils were important, but never completely recovered the apparent 'lack' of survival in the first stages. By using live/dead staining techniques we discovered that the dramatic drop in CFU numbers of strain LH128 upon inoculation in soil is probably due to a change into a viable but not cultivable state (VBNC) rather than to cell death. In fact, both experimental data and modeling as in WP4 suggested that the VBNC LH128 cells appear to actively degrade phenanthrene, even though they cannot be directly recovered and form colonies on agar plates. This showed that one has to be careful with survival data that are solely based on CFU determinations, because introduced cells may be in a dormant state and recover from there.
(Legend to Figure 4. Dynamics of Sphingomonas sp. LH128 populations (CFU/g soil) (left column) and phenanthrene concentration (right column) (mg/kg soil) in microcosms containing either soil 151, 152 and 283 at different moisture contents (expressed relative to the soil’s water holding capacity (WHC)) and at different sampling times (i.e. day 0.125 (3 hours), 2, 5, 9, 15, 22 & 35). Non-spiked controls (left column) and non-inoculated controls (right column) were sampled at day 0.125 15 & 35. Data points and error bars represent the mean values and standard deviations of triplicate measurements. Where error bars are not visible, they are hidden behind the symbols.)
In a second stage we focused on soil type and soil nutrients. A large set of microcosm experiments was hereto performed (Figure 4). In addition to soil moisture content, also pH and other soil-specific physico-chemical characteristics determined the Sphingomonas sp. LH128 inoculum survival rates and catabolic activity towards phenanthrene in sterile microcosms. Multivariate regression analysis against more than 25 determined soil parameters indicated that the most important abiotic soil parameters influencing LH128 survival and catabolic activity in soil are pH, exchangeable Al concentrations, oxalate extractable Mn concentrations, organic carbon content, nitrogen content and also the soil particle size distribution. Both Principal Component Regression (PCR) analysis and Partial Least Square Regression (PLSR) analysis were used, but PLSR resulted in a more accurate predictive model than PCR for both CFU numbers and phenanthrene degradation activity. Based on the developed models, it would be possible to predict the performance of strain LH128 in a particular soil based on the analysis of a limited number of parameters (such as pH and carbon and nitrogen content). This would help to assess the feasibility of bioaugmentation at contaminated sites, although biotic factors and climatic conditions must also be considered. Survival and PAH degrading activity of strain LH128 was then also tested in non-sterile and historically PAH contaminated soil, supplemented with or without extra phenanthrene. Data showed that LH128 survived and degraded PAHs in the soil, however, the presence of an endogenous community clearly negatively affected its population development in non-sterile soil. Replicated experiments with reporter strains from the other bacsin S. wittichii RW1 suggested that this 'negative' effect may be primarily due to endogenous bacteria scavenging for PAH metabolites that the primary introduced degrader is producing and releasing into its surroundings. The main conclusion of this part of the work was thus that survival of an introduced strain can be reasonably well controlled if a number of (trivial) physicochemical and soil parameters are taken into consideration. Importantly, we also observed that typically introducing large amount of cells is resulting in a fast decline whereas small populations of cells will survive better and increase in population size. This is because there is usually a direct relation between the amount of 'carbon' available from the pollutant, which can sustain or not a particular population size of the introduced organisms. The relatively good behaviour of LH128 was reason to use it in a controlled field trial (see WP9).
The good Sphingomonas survival was also used as basis to test the genome-wide changes after initial and prolonged survival in control and contaminated soils. Although much of this is still ongoing a number of interesting observations were made. Transcriptome data of both Sphingomonas sp. LH128 and S. wittichii indicated a very rapid change within one hour after addition to soil. Different effects were tested, e.g. bare sandy soil, effect of targeted carbon source, effect of preculturing conditions and effect of protozoan presence. As an example, survival in non-amended, non-sterile sandy soil resulted in a cellular reaction very different than seen with salt or PEG8000-induced stress, namely oxygen stress, nutrient scavenging and general loss of energy generation. introduction into non-sterile soil but amended with target carbon source (e.g. phenanthrene or dibenzofuran) completely alleviated these stress signatures, but only when cells were pregrown on the same carbon source. In case of RW1, preculturing on e.g salicylate, caused an immediate delay in response to dibenzofuran in soils, with accompanied reprogramming of the metabolism. Preculturing therefore can make an important difference for assessing the survival and activity of the strain in its new environment. Introduced cells simply search for their preferred carbon source and if it is present, adapt immediate their metabolism to continue to grow. If not, the cells just stop to grow and enter in a stationary phase.
Environmental stress response of P. putida KT2440 was studied taking into account two applications, i.e. degradation of formaldehyde in formaldehyde contaminated wastewater and degradation of PAHs in PAH contaminated soil. In agreement with its genome annotation analysis, it was found that strain KT2440 cannot use formaldehyde or formate as the sole carbon source but that it is able to convert 14C-formaldehyde into 14CO2, indicating mineralization of the chemical without incorporation of C1 into biomass. Experiments with lacZ based transcriptional fusions with promoter regions of putative formaldehyde detoxification genes showed that those genes were expressed during the stationary phase, regardless of the addition of formate or formaldehyde, suggesting that the expression of the detoxification system takes place from promoters whose expression are dependent on the starvation and stationary phase sigma factor sigmaS. KT2440 was further shown to tolerate naphthalene in naphthalene-polluted rhizospheric soil in lab microcosms. However, mutants affected in gene functions previously identified as to be associated with in vitro naphthalene tolerance were not affected by naphthalene in rhizosphere soil. Interestingly, soil-native catabolic plasmids could be introduced into KT2440, by which the strain effectively degraded naphthalene. This is an interesting practise because resulting KT2440 transconjugants are not considered genetically modified bacteria. The advantage of this is thus that a non-pathogenic host background can be used as chassis for a new degradation pathway, whereas natural Pseudomonas isolates can still frequently carry non-desired genes. Both P. veronii strains were also tested extensively in microcosms with contaminated material from different origins, with or without addition of benzene. This was done using labeled strains. In contrast to Sphingomonas, P. putida was more sensitive to moisture content, requiring at least 25%, although unique strains were isolated from soils after forest fires that could thrive at 2% moisture. Also P. putida mineralized the target pollutant naphthalene in clean contaminated soils.
Microcosms experiments with the TCE dehalogenating BACSIN-organism Desulfitobacterium in contaminated sediments first focused on the role of redox conditions and on the importance of interactions with other endogenous micro-organisms. From a well-described contaminated river sediment in the Netherlands, sediment-free and sediment-containing glucose-degrading enrichment cultures were obtained, which was used as experimental ecosystem together with inoculated with Desulfitobacterium. Interestingly, Desulfitobacterium augmented microcosms with sediment material strongly profited from native electron-donating bacteria to increase rates of PCE dechlorination. Experimental results with Desulfitobacterium sp. strain Y51 alone in microcosms suggested that the strain is capable of growth and activity under in situ conditions, but (again) the presence of an already established community can prevent substantial growth. Currently, at some polluted locations in situ sterilisation approaches are utilised in order to vaporize chlorinated compounds. Our results suggest that Desulfitobacterium could be added to remove residual PCE as under conditions where it is basically the only community member it grows and degrades PCE rapidly.
Not unexpectedly (in the line with previous literature), A. chlorophenolicus was found to have good survival in soils, and could completely degrade its target compound 4-chlorophenol within a few days. By using two different inoculation densities (105 or 108 per gram soil) we could easily demonstrate the strong effect that low or high inoculum has on degradation rates (faster with higher inoculum) and population growth (better with low inoculum). As an example, at inoculum density of (either as live dry-formulated bacteria or freshly grown bacteria) 2·108 cfu per g of dry soil, the degradation rates from either treatment were nearly identical. The initial concentration of 4-CP was 130 ?g per g dry soil and the detection limit for the HPLC assay for 4-CP was 2 ?g per ml corresponding to 4 ?g of 4-CP per g of dry soil. Already after three days more than 95 % of the initial amount of 4-CP was degraded and after four days no 4-CP could longer be detected. The bacteria native to the soil used in this mi- crocosm experiment did not significantly contribute to degradation of 4-CP. Therefore, both freshly grown cells or dry-formulated cells in vermiculite make active A. chlorophenolicus ingredients. Excellent degradation of 4-chlorophenol was also detected in percolating sand systems artificially contaminated with 4-chlorophenol. The latter system mimicks a percolating sandfilter for treatment of chlorophenol contaminated wastewater.
BACSIN rhizo- and phyllosphere applications
WP7 specifically focused on the rhizosphere and phyllosphere applications of bacsins. This work package was generally very successful and delivered very useful information concerning simple survival tests implicating plant leaves or plant roots, unexpected pollutant degrading microbes in the phyllosphere, or important rood adhesion behaviour. Most importantly, this work package demonstrated the success of combined plant/bacsin inoculation for on-site surface remediations.
In first instance, the tasks in this part of the work determined survival of the various selected bacsins in root systems and plant leaves, by different labs, different plants and at different geographic locations. The survival of P. putida was investigated on a large number of tree roots, from which a first selection of trees under which rhizosphere P.putida KT2440 can survive in field conditions was made. Second important observation in this WP was that P. putida KT2440 can survive very well in the rhizosphere of PAH contaminated soils. This was extremely important because it showed that trees, scrubs and bushes can be replanted more easily after devastating forest fires that produce large quantities of PAHs in the soil, when they are co-inoculated with P. putida. Hereto, the root balls of small plantlings are soaked in bacsin solution, before being planted at the site. The seasonal and geographical influence in rhizosphere colonization by P. putida KT2440 was determined. The bacterium did not exhibit any seasonal or geographical limitations in its ability to colonize the rhizosphere. Interestingly, P. putida KT2440 was found suitable to accept catabolic megaplasmids via natural conjugation. It expresses their degradative pathways and we showed on the example of naphthalene that such natural genetic augmentation works successfully. The presence of catabolic megaplasmids did not alter the ability of KT2440 to survive and to colonize the rhizosphere.
All bacsins were tested in rhizospheres of trees or bushes, which indicated that not only P. putida but also A. chlorophenolicus was an excellent root survivor. Sphingomonads were less so, but still survived considerable periods. New P. putida strains were isolated from natural environments with even better rhizosphere survival than the initial KT2440 strain. These were returned to WP1/2 for better characterization (i.e. P. putida BIRD1, TOD-D1E). 12 genes of P. putida KT2440 involved in bacterial adhesion to biotic surfaces were identified, which permitted the determination of the role of adhesion genes of P. putida KT2440 in bacterial rhizosphere colonization.
Survival and possible priming of bacsin cells was also tested in the phyllosphere. Results are very much in line with those of the more controlled model and drying conditions in the other work packages, which makes it interesting to compare with. Experiments applied bacsin cells in suspension on bean leaves, after which the plant was cycled through a two-day regime of full and partial humidity. Subsequently, cells were recovered and plated to test survival (Figure 5). Quite interestingly, A. chlorophenolicus A6 survived very well in the phyllosphere and withstood low air humidity stress. On the contrary, P. putida and S. wittichii RW1 survived much less on the plant leaf, even after previous 'priming' on the plant leaf. As this idea of improvement of phyllosphere performance of bacins by pre-exposition did not work successfully, we abandoned It.
(Legend Figure 5. Plant leaf survival of different bacsins Pseudomonas putida KT2440, Arthrobacter chlorophenolicus A6, Sphingomonas wittichii RW1, Sphingomonas sp. LH128 and the model leaf coloniser Erwinia herbicola 299R after 1 day at high air humidity, 1 day at low humidity and another day at high humidity. Error bars represent standard deviations (n = 4 leaves)).
The superior survival in the phyllosphere by A. chlorophenolicus urged us to make the first plant leaf transcriptome studies because of its drought stress resistant phyllosphere colonizing traits. Transcriptome profiles were established for the plant leaf, showing an interplay between phyllosphere stress and catabolic gene expression. Interestingly, the behaviour of strain A6 on a plant leaf is very different from that on just a support surface (e.g. agar). Transcriptome data also suggested that phenolic compounds leaking from plant leaves may be an important carbon source for establishment of Arthrobacter.
Interestingly, several other Arthrobacter strains were isolated from plants in orchards that are regularly treated with pesticides that can release 4-chlorophenol. Some of those survived even better on plant leaves than strain A6, but not by much. One of such isolates was investigated in further detail to understand its difference to A6. Also a wide range of toluene degrading bacteria that belonged to the genus Rhodococcus was isolated from plant leaves near highways. These strains were also excellent survivors on leaves, and this showed that there is an important and considerable potential for targeted biodegradation of airborne pollutants using the phyllosphere.
Two bacsins were specifically tested in planted fixed bed reactors (PFBR), a system optimized to treat contaminated shallow groundwaters by optimizing root stimulation and microbial communities adhering to those. Unfortunately, these plant systems go through diurnal cycles with oxic/anoxic fluctuations, which makes it difficult to predict bacsin behavior. The project therefore chose a combination of P. putida mt-2 and T. aromatica as BTEX degraders, which could possibly survive this anoix/oxic cycling and achieve coordinated degradation. Although both strains survived in the rhizosphere of Juncus effesus in a PFBR fed with toluene and benzoate for short periods of time, long-term colonization was not successful and the strains were outcompeted by natural toluene degraders. The short periods were sufficient to study specific gene expression from P. putida and Thauera pathways, but the lack of survival could not be exactly pinpointed. Instead several other anaerobic toluene degraders belonging to Magnetospirillum genus appeared, which were isolated. Further catabolic genes in the microbial community in plant bed reactors treating BTEX were identified using degenerate primer sets, suggesting that similar pathways as are carried by mt-2 are present.
A controlled replicate field rhizoremediation assay with a number of bacsins was carried out at a dump site of a petrochemical company in Spain. First, we tested survival in microcosms using bulk soil from an average contaminated sector on P. putida KT2440R, P. putida BIRD1, P. putida DOT-T1E and Sphingomonas sp. LH128. All of them were able to survive and maintain their populations over 105 CFU/g. Secondly, we examined what kind of bacteria are adapted to live themselves at the site, Re- garding the microbial biodiversity analysis we included different sectors differing in the time when they had been contaminated: 3 samples from sectors contaminated for 20 years, 3 from sectors polluted for 6 years, 3 from sector polluted for 2 years, and two pristine controls. All microbial biodiversity studies were done by metagenomic analysis of total DNA directly extracted from soil. After the filtering processes we obtained a minimum of 4000 sequences per sample, which were analyzed by QIIME. Rarefaction curves, which are indicative of the degree of biodiversity, showed that even though pristine soils (controls) had higher microbial diversity than the polluted site, the biodiversity decrease in polluted soils seemed to be independent of the time of contamination. Regarding the composition of microbial populations we found that in polluted samples, Acidobacteria were reduced and cyanobacteria were not detected, but no clear differences were found in biodiversity according with the time of pollution.
In the first design of the rhizoremediation assay, five bacteria were used: 1) P. putida BIRD1, which is a plant growth promoting bacterium, 2) P. putida DOT-T1E, which is a BTEX degrader, 3) P. putida RNM2, as a naphthalene degrader, 4) P. putida EF11-2, as a phenanthrene degrader, and 5) Sphingomonas sp. LH128, also a phenanthrene degrader. We chose two types of plants: avex III and clover, both are Mediterranean seasonal fast-growing plants. In all we tested 5 different treatments; three controls: (i) an untreated control, (ii) a control without microorganisms, and (iii) a control without plants, and two rhizoremediation treatments: (iv) with plants plus all the Pseudomonas strains, and (v) with plants plus all the Pseudomonas strains plus Sphingomonas sp. LH128.
The following parameters were monitored in each plot: a) soil recuperation indicators: plant growth, soil pH, dehydrogenase activity, b) bacterial survival, c) evolution of the microbial populations by biodiversity metagenomic analysis, d) evolution of total petroleum hydrocarbons (TPH), and e) catabolic routes encoded by microbial populations by functional metagenomic analysis. Results over the first 12 months showed that bacsin strains can survive in the rhizosphere in a heavily polluted soil under field conditions during several months. A reduction of total petrol hydrocarbons in the soil of a refinery´s dump was detected after 9 months of rhizoremediation field assay. Plants displayed much more healthy characteristics after addition with microbes degrading the pollutants (Figure 6). Since the beginning of the rhizoremediation field assay, parcels with plants and microorganisms had more vegetal cover and plants were more robust than those without microorganisms, and the effect was even more evident when Sphingomonas sp. LH128 was also included. Rhizoremediation treatments had a low impact on indigenous microbial communities. The main effect was an increase on Acidobacteria, which was found also in pristine soil. TPH concentrations in the soil displayed a decreasing tendency in the plots where the rhizoremediation treatment had been applied, even though the bioremediation process had not finished.
(Legend to Figure 6. Visual aspects of the different plots in the randomized field rhizoremediation trial at the petroleum oil contaminated site. Pictures were taken one month after the beginning of the trials. Legend below each picture indicates the treatment applied on the plot. Control: plot without any treatment. )
After evaluating all these parameters after the first period we concluded that a new application of rhizoremediation treatments was necessary with some improvements derived from the experience of the experience obtained this first year. The continuation of the rhizoremediation assay included a new inoculation and seeding of the same plots. In Spring 2012 we applied the same plant-bacteria combinations with two differences. First, P. putida RNM2 was not included since it did not survive properly. Secondly, the new bacterial consortium was reinforced by inclusion of alkane-degrading bacteria. An exhaustive screening was performed to choose the alkane-degrading strains. We isolated strains with positive growth on minimal media with alkanes and identified them using Gram and API tests, and amplification of the 16S rRNA fragment. Once identified, potential pathogenic strains were discarded. Different combinations of the selected strains were assayed with mixtures of several alkanes, to obtain an optimized alkane-degrading consortium. Then, degradation assays of several linear chain alkanes (from 8 to 60 carbon atoms) were performed to estimate the degradative capability and the growth rate of the consortium. Results from these tests can not be reported yet, because the treatments are still running as this report is being finished.
Selection of contaminated environments for BACSIN application (WP8)
To make the further transition of BACSIN activities to real contaminated sites, WP8 selected a number of appropriate contaminated sites and characterized them in chemical and soil physical terms. All in all four contaminated areas were available for the project, including the former military airport Hradcany, the Jonction former gasification site, the Dunafer coke plants, as well as the Givaudan wastewater treatment plant. Samples from Hradcany and Dunafer were distributed, and the Hradcany site, which has been subject to scientific investigations during a previous EC project, was selected as the first site for a further extensive analysis.
Particularly Hradcany, where active bioremediation had been ongoing for several years, was characterized in great detail. Contaminated sites were mapped in depth and across landscape, including in groundwater boreholes. Several of the highest contaminated spots were selected for further treatment involving chemical oxidation, but also for extensive community catabolic analysis. The results from the chemical oxidation process indicate drastic changes in soil characteristics during short periods of time (pH, O2), which, however seem to level off after arrest of the treatment. Strong chemical oxidation has a good temporary effect on removal of recalcitrant hydrocarbons, but because the site at Hradcany still contains spots of pure hydrocarbon matrix (solvents, NAPLs), these dissolve and recontaminate the clean groundwaters. This procedure would thus have to be repeated during longer times. Microbial community analysis after treatment end suggested that microorganisms reenter and recolonize the site. This could also be good moments to inoculate specific bacsins, because they will find an empty ecological niche. To test this was unfortunately, not allowed on the site.
Community analysis at the site included use of the catabolic chip, metagenomic sequence analysis of new catabolic isolates and isolation procedures for key on-site bacsins. Catabolome array analysis was optimized and validated by introducing quantitative standards (Deliverable 1.3). An RNA extraction method on environmental material was optimized, and a labeling procedure designed by which catabolic gene expression at a site can be analyzed using the catabolome array (Deliverable 8.2). Material from the site was re-exposed to naphthalene vapour in order to see the specific reaction of endogenous microorganisms. This allowed to us to achieve a list of genes upregulated under stress conditions. New sequences were retrieved which can be included in the catabolic gene array or be used for further applications such as design of probes for PCR analysis.
Two isolates (P. veronii YdB and YdBTEX), were recovered, which thrive in high benzene contaminated environments. we started identifying specific stress and survival determinants and compare those to the first generation bacsins. To better understand the genetic basis making those strains related to sequenced bacsins the key players at contaminated sites, we sequenced and annotated their genomes. Interestingly, this showed that two strains at a contaminated site possessed genes for nitrate respiration. This may be key for their superior survival, since it may help them to remain active in low oxygen environments for longer.
Fifty new pesticide and aromatic compound degraders were isolated from the phyllosphere. Bacteria degrading 4-chlorophenol typically belonged to Arthrobacter sp., whereas phyllosphere toluene degraders were Rhodococcus sp. One of the Arthrobacter isolates proved to be a better leaf colonizer than the bacsin A. chlorophenolicus (see WP7). Analysis of its catabolic properties indicate very similar reactivity as the initial bacsin strain A6.
Characterization of petroleum-hydrocarbon contaminated sites themselves by RNA-seq and catabolome arrays also indicated considerable increase in the abundance of genes for PAH metabolism, but also for known stress related genes. The main sites we focused on were Hradcany Airport, the Cokes wastewater plant, and the Jonction cokes plant contaminated site. As far as possible with limited access, the development of the contamination was followed over time and on-site reactive oxygen procedures were installed to improve further biodegradation rates. Five soils from different geographically locations (Australia, Brazil, Switzerland and Czech Republic) were incubated without contaminants and contaminated with benzene or BTEX. The contamination was kept constant during three months and bacterial communities were characterized on eight time points by catabolic microarray experiments and 16S RNA gene sequencing using Illumina. Results are shown in Task 3 and the details are described in Deliverable 8.4. Deliverable 8.2 describes a detailed list of catabolic genes selected in polluted and non-polluted scenarios. These genes can be taken as references for design of new primers. These primers can be afterwards applied for the selection of potential isolates which harbour relevant catabolic genes.
Natural BACSINs in contaminated environments (WP9)
With the major tools for community analysis and catabolic gene screening in hand, we concentrated in this part of the work on understanding the possible factors that govern the prevalence, distribution and development of pollutant-degrading bacteria under pollution stress. Hereto we focused first on the Hradcany site, which had also been used for remediation and from where a number of new bacsins had been isolated (e.g. P. veronii). When looking at their distribution on the site, we found that a large number of bacteria carrying genes for benzene degradation such as in P. veronii were present at an area in Hradcany predominantly contaminated with benzene. Also strains harboring extradiol dioxygenase genes belonging to the so-called K2 subfamily were abundant on-site and could be isolated from Hradcany. Genome libraries prepared with DNA from the site showed a close linkage between both catabolic pathway segments (benzene to catechol, catechol meta cleavage). These data suggested that strains such as P. veronii or with similar catabolic gene organization are wide-spread in Hradcany but in particular associated to benzene contamination.
As Hradcany is also highly contaminated with aliphatic compounds, we looked for the presence of alkane monooxygenase genes in the communities on site. Genes for alkane monooxygenase are common in the core genome of different Pseudomonas species, therefore, we investigated whether many strains might actually be alkane degraders as well. Apart from Pseudomonas, also Burkholderia and Bacilli were found to be of importance in Hradcany for organic solvent degradation, in particular BTEX.
A representative set of contaminated soils from diverse and distant geographical origins, which not only included samples from Hradcany or Jonction, but also from contaminated sites originating from Brazil, Mexico, Colombia and Australia, were then tested in microcosms to study whether communities reproducibly and predictably lead to selection of microbial degrader 'types'. These microcosms were supplemented with continuous supply of benzene or BTEX, and were followed during 3 months for community changes and catabolic gene enrichment. Interstingly, the incubation with benzene only selected for organisms harboring EXDO A genes as in P. veronii, and confirms the type of selection seen in Hradcany. In contrast, selection with BTEX resulted in a wide variety of catabolic genes becoming enriched over time, similar to some of the patterns seen at those contaminated sites already. This suggests that by a prescreening analysis, one can predict to a certain extent which organisms will become enriched with which particular contaminant.
To identify how native communities at contaminated sites cope with environmental and pollution stresses we further studied soils historically and artificially contaminated with phenanthrene. Exposition of the soils to cycles of drying and wetting showed that the artificial contaminated sample underwent rapid change in community structure. However, repeated wetting and drying did not have a significant influence on general microbial community or on Sphingomonas community composition in both soils. Both soils degraded phenanthrene even after various cycles of drying and wetting, with native Sphingomonas likely playing a major role in degradation. Surprisingly, phenanthrene degradation rates in the artificially contaminated samples were less influenced by drought stress than those of the historically (‘aged’) contaminated sample. Dry-wet cycles also had only a small effect on the bacterial and Sphingomonas community structure in both soils except for the most severe dry/wet regime, i.e. at 10% of the soil’s WHC, where two (in soil TM) or one (in soil CUF) (extra) dominant bands were observed in the Sphingomonas community profile. The dynamics of the Mycobacterium community structure was different. For soil CUF, drought/wetting stress did not inflict dramatic changes in the Mycobacterium community structure. This indicates that native Mycobacteria are as robust as previously found for Sphingomonas in the same soil. On the other hand, no Mycobacteria could be detected in soil TM after the first drying and rewetting period although they were present at the start and after the 10-days pre-adaptation period, i.e. before the dry-wet cycle. These experiments thus indicated that alternating dry-wet periods may influence degradation rates but are not so much a stress for endogenous degraders that prevail at the site.
To possibly better model the predicted behaviour of introduced strains within an existing community we used both literature data, pot experiments performed with A. chlorophenolicus and microcosms with the anaerobic PCE degrading Desulfitobacterium hafniense Y51. The modeling was again based on hierarchical regulation analysis. Parameters obtained from simple growth experiments in batch cultures provided a good fit of the A. chlorophenolicus population development and the observed biodegradation rates, in particular when cells were inoculated at very high density. Under the assumption that an introduced strain does not interact with other members of the community, the modeling suggested that the best strategy would be to select for bacsins with high pollutant uptake rates, low growth yields, that store pollutants and/or convert them to storage materials. Furthermore, ideally such strains have high maintenance requirements and die easily under the stress conditions in the field. Modeling for D. hafniense, however, indicated that the assumption of no-interaction is not always valid and that neighboring microorganisms may control the actual rate of biodegradation. D. hafniense depends for its growth and dechlorinating activity on the supply of carbon and energy by fermenting microorganisms. Indeed, when D. hafniense was cocultured with a variety of fermenting communities in batch very strong enhancement of PCE dechlorination rates were observed. However, when coinoculated in microcosms with native anaerobic contaminated sediments no positive effect of the fermenting communities was observed. In fact, in some cases delay of dechlorination occurred, which might be caused by competition of e.g. methanogens, iron reducers for the products of the fermentors of which D. hafniense should benefit.
Finally, we tested whether a cycle of site diagnosis, pre-enrichment of strains, stress prediction and survival, and inoculation could be achieved. For this we chose pesticide-degradation in on-farm biopurification systems (BPS), which are being increasingly used to treat wastewater of farms contaminated with a variety of different pesticides through a biofilter. Pesticide-primed soils were thus used to seed such reactors, which were then followed through cycles of drought, non-use or different pesticides. On the example of linuron we could show that linuron-mineralizing populations in BPS can withstand various sequential stresses and recover rapidly after restoration of conventional operation conditions. Consequently, in a second experiment, we tested whether BPS could be bioaugmented with pure cultures of axenic pesticide degrading bacteria and whether this was more successful than inoculation with primed materials. In this case, a multiple pesticide feed was applied consisting of linuron, metamitron, atrazine and isoproturon. Bacterial inoculants that can degrade the respective compound included Variovorax sp. SRS16, Sphingomonas sp. SRS2, Chelatobacter sp. SR38 and Rhodococcus sp. KSF. In contrast with the strategy of bioaugmentation with primed materials, inoculation with pure cultures at high densities resulted in an immediate high pesticide degradation capacity. In contrast, the axenic mixed coculture was less stable against drought stresses, interruptions and re-establishment. Complete mixed culture engineering is, therefore, a potential option but will require more research for to be operated successfully.
Preservation of BACSINs (WP10)
The final WP aimed to improve the survival of bacsins in formulations for application purposes. The activities were cornered both around product applications and more basic understanding of the potential physiological stresses during formulation and storage procedures. Finally, we attempted to obtain mutants with improved survival characteristics due to membrane composition changes, and will attempt to influence strain survival via induced heterologous Usp complex formation. Fair to say that those last aspects were not very promising, despite quite intensive trials.
All first line bacsins were tested in standard drying protocols and in protocols involving carrier materials (vermiculite, corncob powder). Optimal survival was found in particular for A. chlorophenolicus, which appeared very resistant to drying and survived very well simple mixing with vermiculite and regular ambient air-drying. In contrast, strains such as P. putida and Sphingomonas were rather sensitive. Ratios of P. putida and vermiculite were optimized with regard to initial survival rates and relevant matric stress levels and also powder quality in the resulting product. Differences could be seen with regard to intracellular compatible solutes in the dried cells depending on the drying regime applied. Large differences were also observed between the survival rates of stored P. putida when dried in fluidized bed and when slowly air dried and we speculate that this can be due to the differences seen in accumulated intracellular trehalose. However, P. putida was very sensitive to oxygen during drying as compared to drying under nitrogen gas. P. putida could be reasonably successfully formulated in water-soluble but otherwise hard dry gelatine capsules with long shelf life.
Storage of Alcanivorax was particularly successful at 4°C and -20°C for periods of up to 6 months on polypropylene-based oil sorbent matrix (SuperM®) that can be coated with hydrocarbons (Alcanivorax cells grow in a biofilm on this sorbent). Storage was tested with the oil sorbent material in combination with two different coatings of the balls: Bunker C heavy oil and hexadecane. Resuscitation/survival was tested after 3, 7, 14, 28, 56 and 168 days. Eventually, biofilms of the mixed culture improve survival throughout the conservation process better than biofilms from pure A. borkumensis cultures. In parallel, freeze drying was done, however, no resuscitation after freeze drying was obtained.
Arthrobacter chlorophenolicus could be formulated using Vermiculite and subsequently stored in a dry form at 4° C for at least 6 months and still maintain up to 20 % surviving cells. An improved protocol for the formulation of A. chlorophenolicus has yielded survival rates around 60 %, for periods of one year.
Spray-drying procedures were successful in stabilizing cultures of Alcanivorax borkumensis and Sphingobium sp. Freeze-drying methodology was successful to accomplish drying and maintain vaiability of the anaerobic dehalogenator D. hafniense.
In order to understand the physiological changes during drying and rewetting, we focused on membrane changes, isothermal calorimetry, synthesis of compatible solutes and transcriptome analysis. For this we selected in particular A. chlorophenolicus and Alcanivorax, because at least ~70% of viable cells would withstand the whole drying process. Interestingly, A. chlorophenolicus formulated in vermiculite shows very little expression differences compared to stationary phase. This was repeated multiple times with similar results and this suggests that A. chlorophenolicus is so resistant because it (either) is not realizing that it is being dried, or it has a mechanism to prevent specific programs to be initiated. The rewetting stage is still under investigation. Microarray data from rewetted Alcanivorax borkumensis after freeze-drying shows interesting patterns of reactivation of several important functions.
The success of the work in this WP was evident from the outdoor pot trials with vermiculite formulated and 3 months-stored A. chlorophenolicus preparation, which rapidly degraded 4-chlorophenol in soils with and without plants (Figure 7). In other experiments that were described under WP7 the ability of P. putida was tested to survive on seeds, bentonite, sepioloite and peat, as carrier for inoculation. Besides parameters such us survival rates, time of storage or proliferation also other issues such as the ease to transport and application in the field were evaluated. Eventually peat was chosen as a carrier in the field experiment (the petroleum hydrocarbon contaminated site). These results thus showed that formulation may have to be optimized for the particular bacsin one is working with, but workable solutions that result in active strains for storage and/or transport can be achieved.
(Legend to Figure 7. Targeted degradation of 4-chlorophenol contamination in soil by vermiculite-formulated A. chlorophenolicus A6. Left panel: Development of Arthrobacter counts in soil over time. Right panel: corresponding 4-chlorophenol concentrations over time.)

Potential Impact:
Main dissemination activities
For the dissemination of BACSIN results and concepts the project targeted different audiences and forms, depending on the type of information that needed to be transmitted. Audiences that we aimed to reach consisted of (i) the scientific community in general, (ii) a wider audience of lay people, politicians, interested third parties, (iii) students at schools and universities, and (iv) project internal formation. Different forms for dissemination that the project took, included (i) scientific manuscripts, (ii) other scientific documents, (iii) poster and oral presentations at conferences, (iv) newspaper articles, interviews and viewpoints, (v) glossy magazine summaries, (vi) website with downloadable material, (vii) webcast presentations, (viii) practical and bioinformatic courses, (ix) open days for lab visits, (x) course material, (xi) a scientific conference, and (xii) formal deliverables. Some of those activities will be shortly summarized in the following. Further details can be found in the periodic reports and online lists.
The BACSIN consortium convened in total for 7 General Meetings. In addition, a specific mid-term strategic meeting with project leaders was held half-way in September 2010, in order to reset the initial goals of the program with the newest results and progress, and in order to plan the remaining two years as efficient as possible. Two webinar series were organized with webcast seminars by all partners. A literature club series complemented this discussion and pointed all partners to important new developments or results. All project related material, official documents, reports, results, databases, protocols, presentations, deliverables, and an online discussion forum were accessible to the BACSIN partners (and EC officers) on an internal file server. This was a very fruitful activity and enhanced mutual understanding of the science and promoting further partnerships between different groups. Common protocols could be accessed by all, which helped standardization of techniques and developments of new experimental protocols and methods. Two specific project internal workshops were organized for all BACSIN partners and students. The first such workshop was organized in Leipzig and consisted of a common technical course for the students around the catabolome array method and the analysis of communities and strains by FAME (fatty acid methyl ester analysis). A second specific course was organized in 2010 in Lausanne, and was focused on bioinformatic analysis and modeling. Bioinformatics consisted of micro-array probe design and analysis, including statistical methods, T-RFLP community analysis, catabolic chip design, analysis and statistics, and hierarchical regulation modeling and flux-balance analysis. These courses were also attended by local students outside the consortium.
Dissemination to the scientific community in general occurred mainly through regular channels. Some 30-40 publications with contributions from BACSIN have already appeared in peer-reviewed journals. Many more will follow, since much of the BACSIN research was carried out by PhD students who will now be in the final phase of their thesis work, which is usually the time to finish manuscript submissions. Peer-reviewed publications are still one of the most important dissemination channels, because it ensures independent review and rigorous examination of BACSIN results by experts outside to the consortium. Quite a good number of scientific publications from the BACSIN project consist of collaborative publications between different partner groups, emphasizing the importance that collaborations had on the development of the project. A multitude of aspects from BACSIN were also presented in form of conference contributions (posters and oral presentation, or conference proceedings) by individual or groups of collaborating BACSIN researchers. In addition, BACSIN also developed a 'core' project presentation summarizing the complete project, its goals and its results, which was presented three times at international conferences. BACSIN has been invited to be presented as ‘European collaborative project’ at the ISB International Soceity for BIotechnology meeting from 15-17 September 2010 in Italy, and at an accompanying OECD Workshop. BACSIN has been further presented as ‘European collaborative project’ at the EMB2012 conference in Bologna (April 9-12, 2012), and at the 5th European Bioremediation Conference in Chania, Crete (July 4-7, 2011).
The public BACSIN website was put in place right before the official project start and attracted quite some outside consortium visitors, as specified by web statistics (see, for example, the first and second periodic reports). The website has a number of separate pages with partner and project information, which direct readers to specific aspects of the project, and with a variety of other links. In addition, the website directs to an internal project server for participants only. A specific BACSIN flyer was created and distributed to all partners (100 per partner), and was displayed at conferences.
The major events covered in first instance by the press related to its kick-off meeting. This remai- ned restricted to local newspapers, tv and radio in Switzerland, as listed below. Links to press PDFs are found on the website. • 20 minutes–Lausanne:‘Bactéries mangeuses depolluants étudiées’ • l’AGEFI–Quotidien économique Suisse: ‘L’UNIL pilotera un projet de l’UE de traitement de la pollution’ • Le Temps: ‘Les bactéries utiles qui dépolluent • Le Temps: ‘Nos amies ces bactéries qui depolluent’ • Entreprise Romande: ‘Les bactéries pur nettoyer les sols’ • Radio-Suisse Romande: ‘BACSIN – les bactéries pour dépolluer les sols’ • July 2008: Télévision Suisse-Romande: Histoire des savoirs ‘Nos amies les bactéries’
The BACSIN project was presented on 1 page in the PanEuropean Networks Dec 2011 Science and Technology issue, p. 177, a magazine that is distributed to 10,000 and more agencies, persons, politicians and institutes throughout Europe. A web-portrait and a web-movie of the BACSIN project and coordinator was prepared for Euresearch-Switzerland/UNIL and is available on the website: http://www. unil.ch/euresearch/page87603.html. Other web publicities, included http://ec.europa.eu/research/bioeconomy/pdf/120305_bacsin_de.pdf; http://www.prlog.org/11819883-small-creatures-big-jobs-using-bacteria-to-clean-up-toxic-waste.html;
http://lib.bioinfo.pl/projects/view/2038; http://www.science.apa.at/site/politik_und_wirtschaft/detail?key=SCI_20120306_SCI37350350971283561. Less stress, less mess (NIOO website: http://www. nioo.knaw.nl/content/less-stress-less-mess).
A biology student was hired part-time to develop a proposal for ‘BACSIN at school’. The idea was to find the most challenging and rewarding experiment for kids at the level of early secondary education that could demonstrate the ideas of BACSIN. The student chose to work out an idea of rescue of plant growth in contaminated areas via bacterial rhizosphere inoculation. The concept would be that kids can germinate mung beans to small plants with a single root and two leafs. These plants are then planted in regular garden soil (in a pot), contaminated or not with 5 ml gasoline or 5 g of PAH contamination (grinded coal), and treated or not with a culture of P. putida with the TOL or the NAH7 plasmid. In addition, a specific popular website was launched: 'Less stress, less mess' (NIOO website: http://www. nioo.knaw.nl/content/less-stress-less-mess). The BACSIN project was presented at the Journées Portes Ouvertes (Open Door) of the University of Lausanne on June 4, 5 and 6 2010. Demonstrations were given about the oil spill in the Mexican Gulf and the possible use of Alcanivorax borkumensis in cleaning up marine spills.
A specific highlight in terms of international coursework was the organization of the Joint US-EU Short Course on Environmental Biotechnology, under the topic of 'Microbial Catalysts for the Environment', at the University of Lausanne from July 9-21, 2011. This work shop was organized under the auspices of the EU-US Joint Task Force on Biotechnology Research, in collaboration with the BACSIN project. Local organizer was Jan Roelof van der Meer, BACSIN coordinator. Co-Organizer from the US-side was Gerben Zylstra, Rutgers University.
The Joint US-EU Short Course on Environmental Biotechnology is designed for several purposes. One of the central tenets is to bring together young scientists (at the late Ph.D. or early postdoctoral stages of their careers) in a forum that will set the groundwork for future overseas collaborative interactions. The course is also designed to give the scientists hands-on experience in modern, up-to-date biotechnological methods for the analy- sis of microbes and their activities pertinent to the remediation of pollutants in the environment. The University of Lausanne (UNIL), Switzerland has been the fourth organizer to host this short course. It has a strong microbiology faculty with expertise in many areas of environmental biotechnology. The 2011 course covered multiple theoretical and practical topics in environmental biotechnology. The practical part was centered around a full concise experiment to demonstrate the possibility for targeted remediation of contaminated soil. Experiments included chemical, microbiological, and molecular analyses of sediments and/or waters, contaminant bioavailability assessment, seeded bioremediation, gene probing, PCR amplification, microbial community analysis based on 16S rRNA gene diversity, and microarray analyses. Each of these topics is explained in more detail in the next section. The practical part of the course was complemented with two lectures per day, given by distinguished scientists from the US and from Europe, covering a research area related to what the students are doing in the course. Seminar speakers spent considerable time with the students to promote interaction and discussions.
The 2011 Theoretical and Practical Course on Microbial Catalysts for the Environment, offered jointly by the US and the EU, focused on oil-derived compounds (alkanes, BTEX and polycyclic aromatic hydrocarbons) as model substrates for investigating and illustrating biodegradative processes and the potential for remediation intervention. This group of hydrocarbons was chosen as a model substrate for the following reasons: (1) there is a considerable knowledge base related to their biodegradation both in natural and in engineered systems; (2) contaminants typically occur as mixtures rather than single pure compounds; (3) oil-related contamination is one of the most commonly encountered pollutions; (4) the course instructors had extensive experience investigating oil biodegradation and bioavailability, both in the laboratory as well as in the field; (5) existing pure cultures were available to ensure the success of the course exercises and (6) the objectives and tasks outlined for this two week course could be accomplished using the more readily degradable and relative water soluble contaminants (BTEX and short chain alkanes) as models.
In the laboratory, the students learned through a combination of well-defined experimental designs and practical lectures, the most relevant issues at stake to intervene at a contaminated site with microbial catalysts in order to achieve targeted biodegradation. The course started with a discussion of and field trip to a contaminated shallow aquifer site underlying a former gasification site in the city of Geneva. This site is contaminated with a number of organic solvents, including benzene, toluene, and xylenes (BTEX). This will be the location where students were instructed on various sampling procedures, real-life hydrological difficulties and treatment options for contaminated sediment and ground water. The students were introduced to a variety of site-related issues such as safety, site characterization, contaminant containment, sample transport, storage conditions, past history as well as the goals and objectives of the treatment. Samples from this site were then further used in the laboratory portion of the course.
The main experimental design for the complete course consisted of laboratory-scale soil microcosms derived from the material of the contaminated site. The main focus of the practical part was to study the potential of preselected organic-compound degrading strains to accelerate contaminant biodegradation in the material. Hereto the students applied several strains alone or in combination, notable a Pseudomonas veronii (degrading BTEX), a Sphingomonas wittichii and a Burkholderia sartisoli (both PAH degraders). Microcosms with or without inoculated strains were operated for the entire two-weeks of the course and were analyzed for (1) microbial community development: observing changes in the microbial community as a consequence of pollution, (2) catabolic gene profiling: determining the catabolic potential of the community in the beginning and its evolution over time of the contamination, (3) Catalyst survival: introduction of a specific microbial catalyst degrading one or more target compounds from the contamination, of which the survival was monitored over time in the microcosms, (4) catabolic gene activity: the expression of the catabolic genes in the microbial catalyst over time, (5) pollution bioavailability: the changes in pollution availability over time in the microcosms measured using bioreporter assays.
DNA was successfully extracted from the all microcosms on two occasions (wk 1 and wk 2), and then PCR amplified with standard «T-RFLP» (e.g. terminal restriction fragment length polymorphism) primers designed to amplify partial 16S rRNA gene sequences of bacteria. T-RFLP results clearly demonstrated the presence of the inoculated strains and their different development over time as compared to non-inoculated controls. The same purified DNA samples were then labeled in a specific procedure and analyzed on a catabolic gene chip. This chip contains a large number of oli- gonucleotides representing twelve important catabolic enzyme families, such as Rieske type non-heme iron oxygenases, ferredoxins, extradiol dioxygenases, benzyl succinate synthases or muconate cycloisomerases, and, as a control gene fragments representing a wide diversity of 16S rRNAs. Un- fortunately, not all of the slides were of excellent quality, but the student groups managed to interpret a number of them, which confirmed the T-RFLP results for the presence of the inoculated strains. In addition, they could see the large repertoire of genes that is present in the contaminated material itself, from bacteria which have slowly accumulated over time.
The survival of the introduced catalyst bacteria was followed by two independent methods (in addition to the indirect methods mentioned above): regular selective plating and through the constitutive gfp label that all strains had been marked with. Despite antibiotic resistance markers and selective media, not all students interpreted the types of colonies appearing on plates in the same manner. Therefore, the variability in plating results was high. Washed cell dilutions from the samples were also analyzed by flow cytometry for the presence of the gfp label. Both techniques showed that the cells survived during the first week but then declines in population size in the second week. Perhaps this was due to the toxic nature of the sample material. Further parallel inoculated microcosms served to isolate community RNA. This purified RNA preparation was then subjected to reverse transcriptase reaction and subsequently to quantitative PCR, in order to quantify the mRNA copy number for key catabolic genes of the inoculated microbial catalysts. As a control the students used the constitutively expressed gfp gene. Because of the complexity of this method, only two time points could be analyzed. Results were not dramatically good (high variability), but confirmed the trend seen with the plating and flow cytometry. Catabolic gene expression was thus detected in the first week but strongly decreased in the second week.
In this final experimental technique, the students used bacterial (luciferase) bioreporter assays to quantify the bioavailability of alkanes, PAHs, BTEX and heavy metals in an aqueous extract of the contaminated material. All students could produce excellent calibration curves of the reporter signal as a function of known concentrations of target compound. In contrast, they could not detect changes in available compound fractions over time of the incubation, which suggests that the inoculated strains did not function well. Because the microcosm experiments all concentrated around one experimental design, the students spent considerable time to combine all their results together. They presented the different parts of the experiments in oral form on the last day.
Speakers at the conference included Graeme Paton, University of Aberdeen (UK); Dietmar Pieper, Helmholtz Institute for Infectious Research, Braunschweig (Germany, Teresa Lettieri, European Joint Research Centre, Ispra (Italy); Carrie Harwood, University of Washington (US); Jerry Schnoor, University of Iowa (US); Vladimir Sentchilo, University of Lausanne (Switzerland); Larry Halverson, Iowa State University (US); Hermann Heipieper, Helmholtz Institute for Environmental Research, Leipzig (Germany); Diana Northup, University of New Mexico (US); Joe Suflita, University of Oklahoma (US); Wilfred Roling, University of Amsterdam (NL); Becky Parales, University of California at Davis (US); Terry McGenity, University of Essex (UK); Jennifer Pett-Ridge, Oak Ridge National Labs (US); Willy Verstraete, University of Ghent (Belgium); Hauke Smidt, Wageningen University Research (NL); David Johnson, Swiss Federal Institute for Aquatic Research (CH); Victor de Lorenzo, Centro Nacional de Biotecnologia, Madrid (E); Hauke Harms, Helmholtz Institute for Environmental Research, Leipzig (Germany).
The course involved 24 students, with 12 from each side of the Atlantic. Participants were from a number of different backgrounds, including microbial ecology, molecular microbiology, and environmental science, and both at PhD and postdoc level. The participants were selected on basis of an application, motivation and recom- mendation letters. Course advertizing was done in Science and Nature and through a mass electronic mailing to appropriate people in the field; and through the internet with an interactive web site and postings to the appropriate newsgroups. Students worked in groups of two, and care was taken to obtain a mixture of backgrounds in each group. Students also brought posters on their own work which were on display during the whole period. Participants were housed in University Apartments specifically constructed (in 2007) for this purpose.
All students were unanimously extremely enthusiastic about the course, despite its long daily working hours and the intensive program. Some citations from the overall evaluation:
• «Beyond my expectations regarding the content and organization»
• «I was very happy with the course. It was a very busy two weeks but I wouldn’t change that because I wouldn’t have wanted to cut anything out.»
• «I would highly recommend the course. It was my exact research interest for the theme. I really liked the packed and intense atmosphere so we could learn so much. I was humbled to be in association with most of the people whom I met.
• «Very good! Very intense, very interesting and very well-prepared and very flexible organizers!»
• «Very intense, ambitious, complete, well organized, a good experience, networking.»
• «Over the expectations: I think that many factors influenced the success of this course. One, the possibility to interact with many people from all the work, exchanging ideas with them and comparing all the personal ex- periences. Second, the opportunity to have really interesting scientific discussions with the lectures and at least the opportunity to learn and know other techniques.»
«This course was extremely valuable to my future as a scientific researcher. Not only did I learn many new practical techniques, but I also learned so much from the breadth of speakers. Perhaps most important, it has made me think more outside the US for research opportunities.»
The final scientific BACSIN symposium was organized under the title 'Pollutant biodegradation under environmental stress: towards rational bioaugmentation', from March 29 – 30, 2012, in the 'Trippenhuis', Amsterdam, The Netherlands. This symposium aimed and succeeded in that, and was also organised in relation to the finalisation of a large EU 7th framework project, BACSIN (Bacterial Abiotic Stress and Survival Improvement Network).
The symposium was attended by 60 scientists from all over Europe and from North- and South-America, half of the scientists was from outside the BACSIN research consortium. The symposium consisted of four (connected) sessions, and in each session two invited international renowned speakers (one from outside the BACSIN project, one from within the BACSIN projected) presented their recent work. In total there were eight invited presenta- tions and 16 contributing oral presentations. In addition a poster session was organised, and the posters were available for viewing and discussing throughout the symposium.
The first day was opened with a session concerning “Ecological principles for biodegradation and bioreme- diation”. It addressed the survival and activity of degraders, as depending on their interactions with other microorganisms and the abiotic environment. Enhanced understanding on how the polluted environment selects for certain degraders and how these species adapt to and deal with their environment, provides clues on how to remediate other polluted environments. Combined theoretical and experimental approaches are important in order to understand how microorganisms can establish themselves and function in the environment. Knowledge from well-designed laboratory-based experiments provides the baseline for subsequent experiments in real, polluted environments. Invited presentations were by Hauke Harms (UFZ, Leipzig-Germany) and Steve Lindow (UC Berkeley, US).
Next the symposium continued with the session “Stress response networks in relation to catabolic perfor- mance”. In this session researchers presented how understanding the key factors determining and/or influencing the interplay of abiotic stress signals, survival mechanisms and catabolic activity in metabolic networks of in- dividual degraders, will contribute to better prediction of environmental performance of dedicated degraders. This session in particular addressed laboratory-based experiments to establish the response of microorganisms under environmental-relevant stress conditions (e.g. drought, salt stress, matrix stress) and how this knowledge provides insight on the metabolism of degraders. Key note presentations were provided by Juan-Luis Ramos (CSIC Granada, Spain) and Pablo Nikel (CSIC Madrid, Spain).
The third session was also started on the first day, and continued on the second day as many participants indicated their interest to present in this session. This session concerned the “selection and behaviour of degraders in the polluted environment”. Enhanced understanding on how the polluted environment selects for certain degraders and how these species adapt to and deal with their environment, provides clues on how to remediate other polluted environments. Modern molecular techniques allow for large-scale assessment of community composition and activities. The results of this (molecular) microbial ecology research on polluted environments were presen- ted in this session. Invited presentations were by Tillmann Lueders (Helmholtz Zentrum München, Germany) and Ramiro-Vilchez-Vargas (Helmholtz centre for infection research, Germany).
The symposium ended with a session on “Bioaugmentation experiments: formulations and monitoring”, which focussed on the practical application of bioaugmentation. Adequate formulation of bacterial inocula is a key fac- tor in bioaugmentation experiments, and was described in this session. Also, novel methods to track introduced strains and their survival and activities were presented. Key note presentations were by Frank Löffler (University of Tennessee, US) and Jan Roelof van der Meer (University of Lausanne, Switzerland) who provided a general overview of the outcome of the BACSIN project.
By organising regular and long coffee and tea breaks, as well as organising a conference diner, we attempted to stimulate discussions between participants. We felt the symposium was very successful, with a nice mixture of scientists and presentations. The symposium programme and venue, as well as the social activities (conference dinner, and a walking tour in the evening of the first symposium day) were very much appreciated by the participants. An abstract book, with information on the programme and abstracts of all oral and poster presentations, was made available to all participants and can also still be downloaded and viewed at the symposium website, which will be maintained till the end of this year: www.falw.vu/~bacsin. We thank KNAW for providing funding the symposium, as well as for the opportunity to organise the meeting in the Trippenhuis.
Potential impact and exploitation of results
Trial-and-error (or spray-and-pray) approaches to release specific bacterial strains into the environment or to stimulate specific groups of bacteria for pollutant removal are often unsuccessful. The design of bioremediation strategies would benefit from a more rational approach to utilizing bacteria or bacterial consortia with useful catalytic properties in environmental settings. The boundary conditions for successful application or stimulation of bacteria are increasingly receiving attention. In recent years new technological developments and conceptual frameworks (e.g. ‘omics’ technologies, Systems Biology) have emerged which provide fresh approaches to explore complex biological settings. At the same time, the analytical tools to study microbial behaviour in populations and communities have advanced rapidly and allow cultivation-independent identification of in situ key players. It was therefore of relevance to gather key researchers in these areas and to present recent progress with respect to the understanding of the bacterial catalytic and survival capacities under conditions of stress and the environmental factors governing those responses, and how this knowledge may aid in the design of bioremediation strategies, especially bioaugmentation.
The results of the BACSIN project have helped to provide a better basis for designing bioremediation and bioaugmentation strategies. In particular, the catabolic gene screening tools developed by the project (catabolic microarray, qPCR tools, and metagenome analysis) permit now for the first time to get an unabridged overview of the potential of microbial communities at contaminated sites to carry out biodegradation of contaminants found at those sites. This is extremely important, because it allows to make a pre-assessment of what might potentially be needed to accelerate spontaneous bioremediation processes. Secondly, a number of tools elaborated and tested in the project enables to examine the physiological and physicochemical status of contaminated material, in order to assess how likely the existing or the bioaugmented microorganisms will function. Such tools include in particular gene reporters, which can be used to examine pollutant availability for cells, their growth potential at the expense of the pollutant and their reproductive success in the material, which is key to success of pollutant degradation. Morevoer, such tools included a variety of microcosm assays that can be used to test reactions of the endogenous communities to the proposed treatment, to examine the survival of introduced strains and the potential duration or completion for the treatment. Thirdly, the BACSIN project delivered detailed insight information on the cellular reactions and behaviour of a number of potentially interesting 'off-the-shelf' strains for catabolism of common pollutants (e.g. oil, BTEX, PAHs, TCE/PCE) in a number of environmental compartments (marine, groundwater, soil, plant leafs/roots). This detailed information permitted to draw conclusions on the potential usefulness and general behavior of such strains after introduction, and/or their specific needs. We believe that this has an important impact on the practise of bioaugmentation, because it can avoid critical mistakes and increase the chance of success of treatment. Finally, the project showed on a number of concrete cases how bacteria with catabolic potential are best formulated, preserved and used for site treatment. In some cases this involves simple techniques such as 'root-ball soaking', mixing with vermiculite for air-drying, or growing in presence of oil on polypropylene fibers, whereas in other cases more advanced techniques such as high-vacuum freeze-drying in nitrogen-atmosphere are required.
Exploitation of results has been organized at different levels. Apart from direct scientific publications, which are at the responsablity of the respective individual BACSIN researchers or collaborating groups, the consortium strived to have concrete concepts, prototypes and experimental evidence, which can be exploited by the commercial partners of the project (or by individual BACSIN research groups). This includes for example, the root-ball-soaking procedure for bacsin inoculation of plants that help to remediate contaminated soil sites or areas after forest fires. Also peat-mixing and vermiculite-mixing are techniques that relevant companies can directly exploit for their daily practise. Unfortunately, exploitation of foreground in environmental management or bioremediation by university / scientific institute partners in Europe is hardly commercially (financially) viable, since registration fees and fees for examination of applications are too elevated compared to the potential outcome. Prototypes developed by individual BACSIN partners, whether these are genetic tools (vectors, catabolic chips, qPCR methods), bacterial strains (e.g. reporter strains, specific bacsins for cleanup), or test methods (e.g. plant leaf survival assay, plant root survival assay, microcosm survival assay, biotreatment pile setup), will therefore remain 'public good'; accessible to anyone with a (non-commercial) interest or available as repeatable concept.

List of Websites:
http://www.bacsin.org
which is automatically redirected to http://www.unil.ch/bacsin