Final Report Summary - HETEROGENIUS (Heterogeneous activity underlying the control of animal behaviour by a fungus)
One of the most dramatic examples of parasites controlling behaviour is the manipulation of ants infected by the fungus Ophiocordyceps unilateralis, causing them to bite into vegetation before dying. Most work so far has described the ant’s behaviour. This work has focussed on taking the first steps towards studying the processes underlying this manipulated behaviour from the fungal genome point of view.
To be able to move towards studying the molecular mechanisms, first this system had to be moved from the field into the lab. Protocols have been established to isolate the fungal parasite from infected ant cadavers collected in the field, to grow and to maintain fungal cultures. All cultures obtained have been genotyped and identified as Ophiocordyceps sp. Subsequently, infection protocols have been developed to reconstruct Ophiocordyceps infections and behavioural manipulation under controlled laboratory conditions. This allowed for the study of the survival probability of infected ants and species-specificity of the fungal parasite. Furthermore, it has allowed for the collection of specimens during the characteristic manipulated biting behaviour for mixed transcriptomics studies to discover the ant host and fungal parasite genes involved in establishing this. Finally, fungal isolates have been used to set up protocols to study the molecular aspects of the internal biological clock of the parasite.
To elucidate the fungal metabolites that are secreted in different parts within the host an ex vivo ant tissue culturing system was developed and combined with a metabolomics approach. This novel technique has been tested in a study in which the metabolites secreted by Beauveria and Metarhizium as a reaction to growth in the presence of dead and live ant brains and muscles were analysed. The study showed that both these entomopathogens react heterogeneously to different host tissues. Furthermore, the study led to the discovery of novel cyclic depsipeptides, a group of mycotoxins that has generally been well studied. Subsequently, this technique was used to study species-specificity and heterogeneous metabolite secretion in O. unilateralis. This study showed that besides this fungus displaying a different secretome on ant brain tissue versus muscle tissue, it also reacts heterogeneously to brains of different ant species. The study also led to the discovery of the first candidate compounds that are likely involved in brain manipulation. Despite metabolite discovery being a difficult task at this point in time, the function of two of these candidate compounds was determined: sphingosine and guanidinobutyric acid. Both are, according to the literature, known to be involved in neurological diseases.
To determine which genes are involved in the manipulated biting behaviour that O. unilateralis induces in Carpenter ants, infection studies were performed followed by RNAseq. From these infection studies ants were sampled at the moment of manipulated biting (while still alive) and after manipulation (death after biting). Mixed RNA from whole heads of these infected ants was compared to RNA from healthy ant heads and the fungal parasite grown in culture, which functioned as controls. In line with this part of the project the genome of O. unilateralis has been sequenced, assembled into contigs, and annotated to function as a reference for the RNAseq data. For the ant RNA the published genome of Camponotus floridanus was used. This mixed transcriptomics approach, together with a comparative genomics study, showed that the majority of the fungal genes that are up-regulated during manipulated biting behavior are unique to the O. unilateralis s.l. genome. This study furthermore revealed that the fungal parasite might be regulating immune- and neuronal stress responses in the host during manipulated biting, as well as impairing its chemosensory communication and causing apoptosis. Moreover, we found genes up-regulated during manipulation that putatively encode for proteins with reported effects on behavioral outputs, proteins involved in various neuropathologies and proteins involved in the biosynthesis of secondary metabolites such as alkaloids.
The return phase of the project concerned the fungal circadian clock. Field studies have shown that manipulated biting behaviour is synchronized at noon. This indicates that external synchronizers (i.e. Zeitgebers), such as light and temperature might have a big influence on the parasitic behavioural control observed. Furthermore, several studies concerning ant behaviour and related gene expression have shown that ant behaviour is highly depending on the circadian clock. In fact, the introduction of 24-hour cycles was key to the protocols developed to reconstruct the fungal infection and manipulation of ant behaviour in the lab. The fungal circadian clock might thus be an important element in establishing behavioural manipulation. Therefore, the return phase has focused on designing chronobiology experiments to characterize the molecular elements of the O. unilateralis clock. Candidate clock genes present in the fungus have been identified using the sequenced O. unilateralis genome mentioned above and preliminary experiments demonstrating the endogenous rhythmicity of one of the main players of the fungal molecular clock have been performed.
The results I obtained during this project, have paved the way towards unravelling the complex mechanisms underlying behavioural manipulation by a fungal parasite. Future studies will focus on performing functional studies to determine the importance of the candidate genes and metabolites identified during this project, as well as the further exploration of the importance of biological clocks. This research project will also be very informative for other studies concerning parasite-host interactions and specifically those that also focus on behavioural manipulation. Furthermore, the discovery of genes and metabolites involved in this phenomenon is opening up opportunities towards applied research within the fields of drug discovery and biological control of insect pests. Both the results from this project and the recent literature point towards these future socio-economical impacts. Moreover, this project has generated a great amount of interest from both researchers and the general public, creating various outreach opportunities.
Project Objectives for the Period
The Project has been based on 3 Research Objectives with the time line set to spend about 1 year on each of them. There were no recommendations from previous reviews. Objectives as proposed in the written application:
Outgoing phase – Pennsylvania State University, State College, United States
Host 1: Dr David Hughes - The Huck Institute of the Life Sciences, Center for Infectious Disease Dynamics (CIDD)
Research objective 1: Adaptive gene expression is a feature of brain manipulating fungi
Proposed technique: LCM of heterogeneous expression
Goals year 1:
In the first year of the project I intend to perform LCM on the different stages and body parts of lab-infected ants to look at heterogeneous expression (Objective 1). The techniques I will learn include ant keeping, artificial ant infection and behavioural observations. I will also further enhance my LCM and RNA isolation techniques. The analysis of the transcriptomics data in collaboration with the Cuomo lab from the Broad will further advance my understanding of bioinformatics and lead to the elucidation of genes important in the different stages of infection and behavioural manipulation. This will lead to one big, or several smaller papers as this approach is completely novel.
Research objective 2: Fungal heterogeneity is expressed via the metabolome inside infected ants
Proposed technique: LC/MS of metabolites in cells transferred within an artificial ant
Goals year 2:
In the second year, I would like to complete the transfer experiments in the artificial ant and analyse the dynamic metabolome of the fungus (Objective 2). This data set will reveal the interesting metabolites the fungus is able to produce as a reaction to the interaction with different ant parts. I will learn the state of the art LC/MS technique and gain ant-dissecting skills. These findings will definitely result in a high-impact paper. Furthermore, it will, together with the results from Objective 1, provide the first insight into the mechanisms underlying heterogeneous activity of the parasite and reveal compounds interesting for biological pest control and medicine development. Therefore, these results will be a nice starting point to apply for follow-up funding within the EU that ask for a more applied approach.
Next to completing Objective 2, I want to already start preparing for Objective 3, which will be done at the Return Host in Germany. I want to use my gained skills from Objective 1 to already prepare the chronobiology samples needed for the last part of this proposed project in which the candidate clock genes of fungi grown inside the ant will be analysed.
Return phase – University of Munich Medical School, Munich, Germany
Host 2: Professor Martha Merrow - Institute for Medical Psychology (IMP)
Research objective 3: Behaviour, heterogeneity and the circadian clock
Proposed technique: Chronobiology and RT-QPCR
Goals year 3:
In the third year I will relocate to Munich at the start of the year. I will elucidate Ophiocordyceps clock genes and hopefully provide the first proof that heterogeneous fungal activity, that underlies the manipulation of host behaviour, is a function of the fungal circadian clock. Working on this third objective, I will learn chronobiology techniques and how to interpret this type of data. The identification of clock genes in Ophiocordyceps will also result in a paper. If I am also able to link the parasite’s clock to the activities underlying infection and behavioural manipulation, this will lead to a paper with significant impact since these findings would be extremely novel and interesting for several research fields such as molecular fungal genetics, chronobiology and parasitology.
Training goals:
1) Advanced training in fungal pathogenicity
2) Training in fieldwork and ant keeping, behavioural observation and artificial infection of ants
3) Advanced training in LCM and transcriptomics
4) Advanced training in LC/MS
5) Training in chronobiology
Work Progress and Achievements During the Period
Work progress has been made as presented in Annex I of the Grant Agreement, however the time line has been restructured. The time line for year 1 and 2 of the Outgoing Phase have been swopped. The reason for restructuring the time line has been some unexpected set backs in terms of resources and equipment at the Outgoing Host Institute. At the start of the project the arrival of the Laser Capture Microscope, which was planned to be used for the transcriptomics experiments to achieve Objective 1, had been delayed by a couple of months. Furthermore, the grant that would support sequencing, assembly and annotation of the Ophiocordyceps genome in collaboration with the Broad Institute did not come through, leaving me without a genome to map the RNAseq data to. Therefore, the choice was made to start with Objective 2 first, which would allow for time to solve the problems that had arisen with Objective 1.
In order to reach Objective 2 Fungal heterogeneity is expressed via the metabolome inside infected ants novel protocols have been developed. These protocols portray a new method to study tissue specific parasite-host interactions. The interactions between parasite and host are very complex, especially concerning the heterogeneity within the host and how the parasite deals with this over time. The possibility to study the metabolites involved in each individual host-tissue interaction is therefore very informative. Hereto, an “artificial ant tissue culturing system” was developed in which ant tissues are kept alive outside the host’s body and the fungal parasite is cultured in the presence of these tissues to measure the metabolites secreted by the fungus as a reaction to this. First, this technique was tested with the rather well studied entomopathogens Beauveria and Metarhizium grown on dead and live Carpenter ant brains and muscles. This resulted in the confirmation that fungal pathogens indeed react heterogeneously to different live and dead tissues inside infected ant hosts by displaying a different metabolome. It also resulted in the discovery of novel cyclic depsipeptides that, though this group of mycotoxins is rather well studied, hadn’t been described before. This work has been published in PLoS One.
Subsequently, this protocol has been applied in a study with Ophiocordyceps unilateralis as proposed in Annex I of the Grant Agreement. This again resulted in the affirmation of the proposed hypothesis concerning heterogeneity. Furthermore, experiments with different ant species’ brains has shown that O. unilateralis secretes a significantly different array of metabolites as a reaction to the various species. This is in line with infection experiments in these ant species in the lab in which in some species manipulated behaviour was observed and in others it was not. This is also in line with what has been observed in the field: some species get infected and manipulated while other related species do not. These experiments thus show that this species-specificity might have a molecular basis. As part of these studies 2 metabolites, that were significantly enriched in O. unilateralis – ant brain interactions, were identified: sphingosine and guanidinobutyric acid. Both these metabolites are related to various neurological diseases and cancers making them the interesting first candidate compounds involved in brain manipulation. This interdisciplinary work (including, infection studies, behavioural studies and metabolomics profiling) has been published in BMC Evolutionary Biology and has reached “Highly Accessed” status within the first two weeks the pre-publication appeared online. This work also received an enormous amount of media attention with write-ups in Internet outlets such as Science Magazine, Wired, LiveScience.com IFLScience and Spektrum.de. Moreover, F1000 Prime recommended the article to its readers. This article made it to BMC Evol Biol’s Top 5 Most Downloaded Articles of 2014, even though the article was published late in the year (end of August). It also already made it into the Top 50 of most accessed articles of all time for this journal. At the moment of writing this report, this article has been accessed almost 18,000 times.
In order to reach Objective 1: Adaptive gene expression is a feature of brain manipulating fungi again novel protocols were developed. These protocols describe a method to reliably infect Carpenter ants with Ophiocordyceps and create the conditions in which manipulation of ant behaviour is successfully reconstructed under lab conditions. These methods have been published together with the results of Objective 2 as described above. From these infection set ups infected ants were sampled during the hallmark manipulated biting behaviour (while still alive) and after manipulation had taken place (dead). The initial approach was to use Laser Capture Microscopy (LCM) to precisely sample only the fungal cells from the heads of infected ants sampled at these conditions. When the Laser Capture Microscope arrived at the Huck Institutes of Life Sciences Core Facility, as well as a brand new cryostat, which is another important piece of equipment for this approach, several months were spent to try and optimize the conditions towards single cell excision from ant heads as planned. Successful cryosections of healthy ant heads were achieved, however the fungus has proven to damage the ant tissues inside the head too much, making useful cryosections impossible. Formalin fixation and sections were considered but rejected since the repeatability of RNA extractions from these type of samples is still too unreliable. To still be able to achieve Objective 1 and examine gene expression during brain manipulation, the approach of using LCM was abandoned and replaced by doing mixed transcriptomics on whole ant heads. This provides the opportunity to not only look at fungal gene expression but also to that of the ant host.
To overcome the second problem with respect to Objective 1, there not being an annotated genome of O. unilateralis generated by the Broad Institute as expected, a collaboration was started with Dr Andreas Brachmann from the LMU (the return phase institution) to perform genome sequencing as part of this Marie Curie project. Together with him, the genome of O. unilateralis was sequenced using the Illumina MiSeq platform and assembled into contigs. For the annotation of the genome, collaboration was started with fungal bioinformatics expert Dr Robin Ohm from Utrecht University. This has led to a high-quality genome, which can be used for future comparative studies as well as a reference for transcriptomics with 7831 predicted gene models and a CEGMA completeness of 98.69%. Through working with these two experts and completing a graduate course on Applied Bioinformatics taught by the director of the Bioinformatics Facility at Penn State, the training goal to obtain more experience in processing “omics” data has thus been fulfilled.
RNASeq using an Illumina HiSeq platform has been performed successfully, leading to high quality paired end reads with a 16 million reads per sample coverage. To look at differential gene expression between the various conditions (infection during manipulation, infection after manipulation and non-infection controls) the respective sample reads have been mapped to the recently obtained O. unilateralis genome and the published Camponotus floridanus genome using TopHat and Bowtie. This resulted in 86% of the fungal reads being mapped to the fungal genome and 36-40% of the ant reads being mapped to the ant genome, which is both within expectations according to literature. Subsequently, Cuffdiff was used to determine differential gene expression. Genes that were at least 2 fold up or down-regulated, with a significant p-value (p<0.05) and a FPKM of 4 have been included in further analyses. Further analyses included an enrichment study and comparative genomics studies, as well as homology searches against the NCBI database. These approaches showed that the majority of the fungal genes that are up-regulated during manipulated biting behavior are unique to the O. unilateralis s.l. genome. This study furthermore revealed that the fungal parasite might be regulating immune- and neuronal stress responses in the host during manipulated biting, as well as impairing its chemosensory communication and causing apoptosis. Moreover, we found genes up-regulated during manipulation that putatively encode for proteins with reported effects on behavioral outputs, proteins involved in various neuropathologies and proteins involved in the biosynthesis of secondary metabolites such as alkaloids. These results will soon be published together with the annotated genome for O. unilateralis in an extensive article, which was recently accepted by BMC Genomics. In this article we discuss in great detail the various candidate manipulators and possible host pathways through which manipulated biting behavior as seen in O. unilateralis s.l.-infected Carpenter ants could be established. While some of these candidates could be essential to the process, we expect these to work in concert since the manipulated behavior in this system is complex and precise. Future functional gene expression studies will be performed to determine this. Finally, this data set also demonstrated that 2 important clock genes in the ant host are significantly higher expressed during manipulation compared to the healthy ant controls. This is another indication that Objective 3 (below) will eventually result in a positive result.
For Objective 3: Behaviour, heterogeneity and the circadian clock I moved to the return host at the LMU in Munich. For this objective I started by identifying candidate clock genes in the O. unilateralis genome through searching for known signature PFAM domains such as FREQUENCY, TIMELESS and PAS. Among the candidates are homologs of the in the model organism Neurospora crassa found genes encoding for frequency, wc-1, wc-2, vvd, phy-1, ckb-1, ckb-2 and timeless which is a clock gene found in Drosophila melanogaster. QPCR primers for these genes have been designed and tested for their efficiency and specificity. Moreover, many reference genes have been tested for their stability throughout circadian time course samples making use of the recent literature regarding proper use of references in RT-QPCR experiments, which has proven to be a rather time consuming issue. As an important component of Objective 3 I have learned how to conduct circadian time course experiments to verify if the identified clock candidate genes are indeed behaving as reported for N. crassa. This would demonstrate that O. unilateralis indeed has a functional clock, which is essential basic knowledge needed to make further progress regarding Objective 3. Much time has been spent to find, from scratch, the right experimental conditions in which to perform the time course experiments for O. unilateralis. This has been a successful exercise, which already led to preliminary data showing that the gene frequency, a major player in the feed-back loop of the N. crassa clock, is an endogenous oscillator as hypothesized. These preliminary findings will be followed up in the near future to further tease apart the molecular clock of the fungal parasite O. unilateralis.
This project has been implemented with multiple training goals for the 3 Research Objectives. Having isolated and cultured several different species of fungal entomopathogens and setting up the infection and manipulation protocols has provided me with advanced training in fungal pathogenicity and artificial infection of ants. I have also led 3 fieldwork expeditions to South Carolina and participated in expeditions to the Brazilian Atlantic Rainforest and Amazonian Rainforest. Part of this work was collecting and observing manipulated ant behaviour. As part of Research Objective 2 I have performed metabolomics studies doing everything from the experimental set up to sample preparation to data analysis. This has provided me with an advanced training in LC/MS. As part of Research Objective 1 I have learned how to perform cryosections, and how to analyse next generation sequencing data for both genomics and transcriptomics. Towards Research Objective 3 I have learned how to design protocols for and conduct chronobiology experiments, as well as interpret their results. This work has also led to a better understanding of and advanced training in conducting and analysing RT-QPCR data.
Additional Information
In addition to the proposed Objectives an Opinion paper in the journal Comparative and Integrative Biology has been published with both the Outgoing and the Incoming host as co-authors. This has aided in international collaboration between the two labs I visited during the reporting period. I also co-authored 3 papers with my colleagues from Utrecht University related to my research done before the starting of this Marie Curie project. Furthermore, I supervised a visiting Master student from Utrecht University on a proteomics project for 6 months, exploring the proteins secreted by various Ophiocordyceps species. Additionally, during my time at the return host, I supervised a Biology Master student for 3 months. In the multidisciplinary lab at Penn State I have been mainly driving the molecular research. Therefore, I have been teaching molecular techniques to two of the graduate students in the lab and as such have been highly involved in their research projects. As a result, I will be co-authoring a behavioural paper and another transcriptomics paper. Moreover, I have supervised undergrads working in the lab and in the field and have been part of outreach activities such as The Great Insect Fair informing kids and parents about our research. My teaching activities at the LMU involved designing and teaching a short course on Parasite Biology and Ecology for medical students. In addition, I have started collaborations with Dr Brachmann from the LMU and Dr Robin Ohm from the University of Utrecht. To prepare for data analysis I have been taking a graduate course in Applied Bioinformatics taught by the Director of the Penn State Bioinformatics Facility. To prepare for the next step in my career I have taken part in a Leadership Workshop for postdocs organized at Penn State.
Furthermore, I have actively participated in the interdisciplinary seminars and meetings within the Centre for Infectious Disease Dynamics both as a speaker and a seminar organizer. I have been part of the organizing committee of the Penn State Postdoctoral Society and as such organized events such as the yearly research exhibition. I have been presenting my work at different Departments within Penn State and the LMU, other academic institutes (Senckenberg, Leiden University, Utrecht University, Erskine College, Harvard University, Nottingham University, Max Planck Institute, University of Cambridge) and at several international meetings (FGC, ESEB, SICB, YSF) in the form of a lecture. Lectures regarding the research performed the last 3 years have also been planned for upcoming October (Fungal Molecular Biology Meeting in Berlin and University of Copenhagen) and November (University of Düsseldorf).
Lastly, I have also applied for new funding to continue the research performed during this project and dive deeper into the molecular mechanisms that underlie parasitic behavioural manipulation. As a result I will continue my work at the LMU starting September 2015 as a Humboldt Research Fellow. During the project period I have been awarded a “The Huck Institutes of the Life Sciences Shared Technology Facilities Pilot Projects” grant, which funded part of the costs of my LC-MS experiments, a “Pennsylvania State University Post Doctoral Award” that aided towards the costs for travel to the ESEB conference, a “Genetics Society of America DeLill Nasser Award for Professional Development in Genetics”, which aided towards the costs for travel to the ECFG conference, and “Friedrich Bauer Stiftung Funding”, which allowed me to buy essential equipment for RNA extractions.
Project Management
The time line for year 1 and 2 of the Outgoing Phase have been swopped. The reason for restructuring the time line has been an unexpected set back in terms of resources and equipment at the Outgoing Host Institute. At the start of the project the arrival of the Laser Capture Microscope had been delayed by a couple of months. Furthermore, the grant that would support sequencing, assembly and annotation of the Ophiocordyceps genome in collaboration with the Broad Institute did not come through, leaving this project without a genome to map the RNAseq data to. Therefore, the choice was made to start with Objective 2 first, which would allow some time to solve the problems that had arisen with Objective 1. Subsequently, the Laser Capture Microscopy approach has proven to not be a viable approach resulting the change of the approach to mixed transcriptomics. This does not change the goal of Objective 1 and rather results in the positive outcome that this way not only the fungal transcriptome during manipulation can be assessed (as proposed), but also the ant transcriptome. This furthermore has led to a collaboration with Dr Andreas Brachmann from the LMU (the return phase institution) and Dr Robin Ohm from Utrecht University concerning the genome sequencing. The set back of having no genome when the project was started has thus resulted in an extra training opportunity, namely that of genomics.
Objective 2 has been achieved in the first year of this project. Objective 1 has taken a bit longer due to above mentioned set backs and extra work regarding the generation of a novel genome and has been achieved half-way into the third year. Objective 3 has been started while analyses for Objective 1 were still in progress. Setting up the protocols for Objective 3 have been very time consuming, as well as designing the right set up for proper RT-QPCR analysis. While this has been difficult, this has provided me with chronobiology training and advanced RT-QPCR training as envisioned. Objective 3 has however not been completely achieved in terms of final results. The important basis has none the less been made and the preliminary results are the first steps towards unravelling the molecular basis of the clock of the fungal parasite as was proposed. I will continue this work on new funding (a Humboldt Research Fellowship) for which I applied during the last year of this project.
To be able to move towards studying the molecular mechanisms, first this system had to be moved from the field into the lab. Protocols have been established to isolate the fungal parasite from infected ant cadavers collected in the field, to grow and to maintain fungal cultures. All cultures obtained have been genotyped and identified as Ophiocordyceps sp. Subsequently, infection protocols have been developed to reconstruct Ophiocordyceps infections and behavioural manipulation under controlled laboratory conditions. This allowed for the study of the survival probability of infected ants and species-specificity of the fungal parasite. Furthermore, it has allowed for the collection of specimens during the characteristic manipulated biting behaviour for mixed transcriptomics studies to discover the ant host and fungal parasite genes involved in establishing this. Finally, fungal isolates have been used to set up protocols to study the molecular aspects of the internal biological clock of the parasite.
To elucidate the fungal metabolites that are secreted in different parts within the host an ex vivo ant tissue culturing system was developed and combined with a metabolomics approach. This novel technique has been tested in a study in which the metabolites secreted by Beauveria and Metarhizium as a reaction to growth in the presence of dead and live ant brains and muscles were analysed. The study showed that both these entomopathogens react heterogeneously to different host tissues. Furthermore, the study led to the discovery of novel cyclic depsipeptides, a group of mycotoxins that has generally been well studied. Subsequently, this technique was used to study species-specificity and heterogeneous metabolite secretion in O. unilateralis. This study showed that besides this fungus displaying a different secretome on ant brain tissue versus muscle tissue, it also reacts heterogeneously to brains of different ant species. The study also led to the discovery of the first candidate compounds that are likely involved in brain manipulation. Despite metabolite discovery being a difficult task at this point in time, the function of two of these candidate compounds was determined: sphingosine and guanidinobutyric acid. Both are, according to the literature, known to be involved in neurological diseases.
To determine which genes are involved in the manipulated biting behaviour that O. unilateralis induces in Carpenter ants, infection studies were performed followed by RNAseq. From these infection studies ants were sampled at the moment of manipulated biting (while still alive) and after manipulation (death after biting). Mixed RNA from whole heads of these infected ants was compared to RNA from healthy ant heads and the fungal parasite grown in culture, which functioned as controls. In line with this part of the project the genome of O. unilateralis has been sequenced, assembled into contigs, and annotated to function as a reference for the RNAseq data. For the ant RNA the published genome of Camponotus floridanus was used. This mixed transcriptomics approach, together with a comparative genomics study, showed that the majority of the fungal genes that are up-regulated during manipulated biting behavior are unique to the O. unilateralis s.l. genome. This study furthermore revealed that the fungal parasite might be regulating immune- and neuronal stress responses in the host during manipulated biting, as well as impairing its chemosensory communication and causing apoptosis. Moreover, we found genes up-regulated during manipulation that putatively encode for proteins with reported effects on behavioral outputs, proteins involved in various neuropathologies and proteins involved in the biosynthesis of secondary metabolites such as alkaloids.
The return phase of the project concerned the fungal circadian clock. Field studies have shown that manipulated biting behaviour is synchronized at noon. This indicates that external synchronizers (i.e. Zeitgebers), such as light and temperature might have a big influence on the parasitic behavioural control observed. Furthermore, several studies concerning ant behaviour and related gene expression have shown that ant behaviour is highly depending on the circadian clock. In fact, the introduction of 24-hour cycles was key to the protocols developed to reconstruct the fungal infection and manipulation of ant behaviour in the lab. The fungal circadian clock might thus be an important element in establishing behavioural manipulation. Therefore, the return phase has focused on designing chronobiology experiments to characterize the molecular elements of the O. unilateralis clock. Candidate clock genes present in the fungus have been identified using the sequenced O. unilateralis genome mentioned above and preliminary experiments demonstrating the endogenous rhythmicity of one of the main players of the fungal molecular clock have been performed.
The results I obtained during this project, have paved the way towards unravelling the complex mechanisms underlying behavioural manipulation by a fungal parasite. Future studies will focus on performing functional studies to determine the importance of the candidate genes and metabolites identified during this project, as well as the further exploration of the importance of biological clocks. This research project will also be very informative for other studies concerning parasite-host interactions and specifically those that also focus on behavioural manipulation. Furthermore, the discovery of genes and metabolites involved in this phenomenon is opening up opportunities towards applied research within the fields of drug discovery and biological control of insect pests. Both the results from this project and the recent literature point towards these future socio-economical impacts. Moreover, this project has generated a great amount of interest from both researchers and the general public, creating various outreach opportunities.
Project Objectives for the Period
The Project has been based on 3 Research Objectives with the time line set to spend about 1 year on each of them. There were no recommendations from previous reviews. Objectives as proposed in the written application:
Outgoing phase – Pennsylvania State University, State College, United States
Host 1: Dr David Hughes - The Huck Institute of the Life Sciences, Center for Infectious Disease Dynamics (CIDD)
Research objective 1: Adaptive gene expression is a feature of brain manipulating fungi
Proposed technique: LCM of heterogeneous expression
Goals year 1:
In the first year of the project I intend to perform LCM on the different stages and body parts of lab-infected ants to look at heterogeneous expression (Objective 1). The techniques I will learn include ant keeping, artificial ant infection and behavioural observations. I will also further enhance my LCM and RNA isolation techniques. The analysis of the transcriptomics data in collaboration with the Cuomo lab from the Broad will further advance my understanding of bioinformatics and lead to the elucidation of genes important in the different stages of infection and behavioural manipulation. This will lead to one big, or several smaller papers as this approach is completely novel.
Research objective 2: Fungal heterogeneity is expressed via the metabolome inside infected ants
Proposed technique: LC/MS of metabolites in cells transferred within an artificial ant
Goals year 2:
In the second year, I would like to complete the transfer experiments in the artificial ant and analyse the dynamic metabolome of the fungus (Objective 2). This data set will reveal the interesting metabolites the fungus is able to produce as a reaction to the interaction with different ant parts. I will learn the state of the art LC/MS technique and gain ant-dissecting skills. These findings will definitely result in a high-impact paper. Furthermore, it will, together with the results from Objective 1, provide the first insight into the mechanisms underlying heterogeneous activity of the parasite and reveal compounds interesting for biological pest control and medicine development. Therefore, these results will be a nice starting point to apply for follow-up funding within the EU that ask for a more applied approach.
Next to completing Objective 2, I want to already start preparing for Objective 3, which will be done at the Return Host in Germany. I want to use my gained skills from Objective 1 to already prepare the chronobiology samples needed for the last part of this proposed project in which the candidate clock genes of fungi grown inside the ant will be analysed.
Return phase – University of Munich Medical School, Munich, Germany
Host 2: Professor Martha Merrow - Institute for Medical Psychology (IMP)
Research objective 3: Behaviour, heterogeneity and the circadian clock
Proposed technique: Chronobiology and RT-QPCR
Goals year 3:
In the third year I will relocate to Munich at the start of the year. I will elucidate Ophiocordyceps clock genes and hopefully provide the first proof that heterogeneous fungal activity, that underlies the manipulation of host behaviour, is a function of the fungal circadian clock. Working on this third objective, I will learn chronobiology techniques and how to interpret this type of data. The identification of clock genes in Ophiocordyceps will also result in a paper. If I am also able to link the parasite’s clock to the activities underlying infection and behavioural manipulation, this will lead to a paper with significant impact since these findings would be extremely novel and interesting for several research fields such as molecular fungal genetics, chronobiology and parasitology.
Training goals:
1) Advanced training in fungal pathogenicity
2) Training in fieldwork and ant keeping, behavioural observation and artificial infection of ants
3) Advanced training in LCM and transcriptomics
4) Advanced training in LC/MS
5) Training in chronobiology
Work Progress and Achievements During the Period
Work progress has been made as presented in Annex I of the Grant Agreement, however the time line has been restructured. The time line for year 1 and 2 of the Outgoing Phase have been swopped. The reason for restructuring the time line has been some unexpected set backs in terms of resources and equipment at the Outgoing Host Institute. At the start of the project the arrival of the Laser Capture Microscope, which was planned to be used for the transcriptomics experiments to achieve Objective 1, had been delayed by a couple of months. Furthermore, the grant that would support sequencing, assembly and annotation of the Ophiocordyceps genome in collaboration with the Broad Institute did not come through, leaving me without a genome to map the RNAseq data to. Therefore, the choice was made to start with Objective 2 first, which would allow for time to solve the problems that had arisen with Objective 1.
In order to reach Objective 2 Fungal heterogeneity is expressed via the metabolome inside infected ants novel protocols have been developed. These protocols portray a new method to study tissue specific parasite-host interactions. The interactions between parasite and host are very complex, especially concerning the heterogeneity within the host and how the parasite deals with this over time. The possibility to study the metabolites involved in each individual host-tissue interaction is therefore very informative. Hereto, an “artificial ant tissue culturing system” was developed in which ant tissues are kept alive outside the host’s body and the fungal parasite is cultured in the presence of these tissues to measure the metabolites secreted by the fungus as a reaction to this. First, this technique was tested with the rather well studied entomopathogens Beauveria and Metarhizium grown on dead and live Carpenter ant brains and muscles. This resulted in the confirmation that fungal pathogens indeed react heterogeneously to different live and dead tissues inside infected ant hosts by displaying a different metabolome. It also resulted in the discovery of novel cyclic depsipeptides that, though this group of mycotoxins is rather well studied, hadn’t been described before. This work has been published in PLoS One.
Subsequently, this protocol has been applied in a study with Ophiocordyceps unilateralis as proposed in Annex I of the Grant Agreement. This again resulted in the affirmation of the proposed hypothesis concerning heterogeneity. Furthermore, experiments with different ant species’ brains has shown that O. unilateralis secretes a significantly different array of metabolites as a reaction to the various species. This is in line with infection experiments in these ant species in the lab in which in some species manipulated behaviour was observed and in others it was not. This is also in line with what has been observed in the field: some species get infected and manipulated while other related species do not. These experiments thus show that this species-specificity might have a molecular basis. As part of these studies 2 metabolites, that were significantly enriched in O. unilateralis – ant brain interactions, were identified: sphingosine and guanidinobutyric acid. Both these metabolites are related to various neurological diseases and cancers making them the interesting first candidate compounds involved in brain manipulation. This interdisciplinary work (including, infection studies, behavioural studies and metabolomics profiling) has been published in BMC Evolutionary Biology and has reached “Highly Accessed” status within the first two weeks the pre-publication appeared online. This work also received an enormous amount of media attention with write-ups in Internet outlets such as Science Magazine, Wired, LiveScience.com IFLScience and Spektrum.de. Moreover, F1000 Prime recommended the article to its readers. This article made it to BMC Evol Biol’s Top 5 Most Downloaded Articles of 2014, even though the article was published late in the year (end of August). It also already made it into the Top 50 of most accessed articles of all time for this journal. At the moment of writing this report, this article has been accessed almost 18,000 times.
In order to reach Objective 1: Adaptive gene expression is a feature of brain manipulating fungi again novel protocols were developed. These protocols describe a method to reliably infect Carpenter ants with Ophiocordyceps and create the conditions in which manipulation of ant behaviour is successfully reconstructed under lab conditions. These methods have been published together with the results of Objective 2 as described above. From these infection set ups infected ants were sampled during the hallmark manipulated biting behaviour (while still alive) and after manipulation had taken place (dead). The initial approach was to use Laser Capture Microscopy (LCM) to precisely sample only the fungal cells from the heads of infected ants sampled at these conditions. When the Laser Capture Microscope arrived at the Huck Institutes of Life Sciences Core Facility, as well as a brand new cryostat, which is another important piece of equipment for this approach, several months were spent to try and optimize the conditions towards single cell excision from ant heads as planned. Successful cryosections of healthy ant heads were achieved, however the fungus has proven to damage the ant tissues inside the head too much, making useful cryosections impossible. Formalin fixation and sections were considered but rejected since the repeatability of RNA extractions from these type of samples is still too unreliable. To still be able to achieve Objective 1 and examine gene expression during brain manipulation, the approach of using LCM was abandoned and replaced by doing mixed transcriptomics on whole ant heads. This provides the opportunity to not only look at fungal gene expression but also to that of the ant host.
To overcome the second problem with respect to Objective 1, there not being an annotated genome of O. unilateralis generated by the Broad Institute as expected, a collaboration was started with Dr Andreas Brachmann from the LMU (the return phase institution) to perform genome sequencing as part of this Marie Curie project. Together with him, the genome of O. unilateralis was sequenced using the Illumina MiSeq platform and assembled into contigs. For the annotation of the genome, collaboration was started with fungal bioinformatics expert Dr Robin Ohm from Utrecht University. This has led to a high-quality genome, which can be used for future comparative studies as well as a reference for transcriptomics with 7831 predicted gene models and a CEGMA completeness of 98.69%. Through working with these two experts and completing a graduate course on Applied Bioinformatics taught by the director of the Bioinformatics Facility at Penn State, the training goal to obtain more experience in processing “omics” data has thus been fulfilled.
RNASeq using an Illumina HiSeq platform has been performed successfully, leading to high quality paired end reads with a 16 million reads per sample coverage. To look at differential gene expression between the various conditions (infection during manipulation, infection after manipulation and non-infection controls) the respective sample reads have been mapped to the recently obtained O. unilateralis genome and the published Camponotus floridanus genome using TopHat and Bowtie. This resulted in 86% of the fungal reads being mapped to the fungal genome and 36-40% of the ant reads being mapped to the ant genome, which is both within expectations according to literature. Subsequently, Cuffdiff was used to determine differential gene expression. Genes that were at least 2 fold up or down-regulated, with a significant p-value (p<0.05) and a FPKM of 4 have been included in further analyses. Further analyses included an enrichment study and comparative genomics studies, as well as homology searches against the NCBI database. These approaches showed that the majority of the fungal genes that are up-regulated during manipulated biting behavior are unique to the O. unilateralis s.l. genome. This study furthermore revealed that the fungal parasite might be regulating immune- and neuronal stress responses in the host during manipulated biting, as well as impairing its chemosensory communication and causing apoptosis. Moreover, we found genes up-regulated during manipulation that putatively encode for proteins with reported effects on behavioral outputs, proteins involved in various neuropathologies and proteins involved in the biosynthesis of secondary metabolites such as alkaloids. These results will soon be published together with the annotated genome for O. unilateralis in an extensive article, which was recently accepted by BMC Genomics. In this article we discuss in great detail the various candidate manipulators and possible host pathways through which manipulated biting behavior as seen in O. unilateralis s.l.-infected Carpenter ants could be established. While some of these candidates could be essential to the process, we expect these to work in concert since the manipulated behavior in this system is complex and precise. Future functional gene expression studies will be performed to determine this. Finally, this data set also demonstrated that 2 important clock genes in the ant host are significantly higher expressed during manipulation compared to the healthy ant controls. This is another indication that Objective 3 (below) will eventually result in a positive result.
For Objective 3: Behaviour, heterogeneity and the circadian clock I moved to the return host at the LMU in Munich. For this objective I started by identifying candidate clock genes in the O. unilateralis genome through searching for known signature PFAM domains such as FREQUENCY, TIMELESS and PAS. Among the candidates are homologs of the in the model organism Neurospora crassa found genes encoding for frequency, wc-1, wc-2, vvd, phy-1, ckb-1, ckb-2 and timeless which is a clock gene found in Drosophila melanogaster. QPCR primers for these genes have been designed and tested for their efficiency and specificity. Moreover, many reference genes have been tested for their stability throughout circadian time course samples making use of the recent literature regarding proper use of references in RT-QPCR experiments, which has proven to be a rather time consuming issue. As an important component of Objective 3 I have learned how to conduct circadian time course experiments to verify if the identified clock candidate genes are indeed behaving as reported for N. crassa. This would demonstrate that O. unilateralis indeed has a functional clock, which is essential basic knowledge needed to make further progress regarding Objective 3. Much time has been spent to find, from scratch, the right experimental conditions in which to perform the time course experiments for O. unilateralis. This has been a successful exercise, which already led to preliminary data showing that the gene frequency, a major player in the feed-back loop of the N. crassa clock, is an endogenous oscillator as hypothesized. These preliminary findings will be followed up in the near future to further tease apart the molecular clock of the fungal parasite O. unilateralis.
This project has been implemented with multiple training goals for the 3 Research Objectives. Having isolated and cultured several different species of fungal entomopathogens and setting up the infection and manipulation protocols has provided me with advanced training in fungal pathogenicity and artificial infection of ants. I have also led 3 fieldwork expeditions to South Carolina and participated in expeditions to the Brazilian Atlantic Rainforest and Amazonian Rainforest. Part of this work was collecting and observing manipulated ant behaviour. As part of Research Objective 2 I have performed metabolomics studies doing everything from the experimental set up to sample preparation to data analysis. This has provided me with an advanced training in LC/MS. As part of Research Objective 1 I have learned how to perform cryosections, and how to analyse next generation sequencing data for both genomics and transcriptomics. Towards Research Objective 3 I have learned how to design protocols for and conduct chronobiology experiments, as well as interpret their results. This work has also led to a better understanding of and advanced training in conducting and analysing RT-QPCR data.
Additional Information
In addition to the proposed Objectives an Opinion paper in the journal Comparative and Integrative Biology has been published with both the Outgoing and the Incoming host as co-authors. This has aided in international collaboration between the two labs I visited during the reporting period. I also co-authored 3 papers with my colleagues from Utrecht University related to my research done before the starting of this Marie Curie project. Furthermore, I supervised a visiting Master student from Utrecht University on a proteomics project for 6 months, exploring the proteins secreted by various Ophiocordyceps species. Additionally, during my time at the return host, I supervised a Biology Master student for 3 months. In the multidisciplinary lab at Penn State I have been mainly driving the molecular research. Therefore, I have been teaching molecular techniques to two of the graduate students in the lab and as such have been highly involved in their research projects. As a result, I will be co-authoring a behavioural paper and another transcriptomics paper. Moreover, I have supervised undergrads working in the lab and in the field and have been part of outreach activities such as The Great Insect Fair informing kids and parents about our research. My teaching activities at the LMU involved designing and teaching a short course on Parasite Biology and Ecology for medical students. In addition, I have started collaborations with Dr Brachmann from the LMU and Dr Robin Ohm from the University of Utrecht. To prepare for data analysis I have been taking a graduate course in Applied Bioinformatics taught by the Director of the Penn State Bioinformatics Facility. To prepare for the next step in my career I have taken part in a Leadership Workshop for postdocs organized at Penn State.
Furthermore, I have actively participated in the interdisciplinary seminars and meetings within the Centre for Infectious Disease Dynamics both as a speaker and a seminar organizer. I have been part of the organizing committee of the Penn State Postdoctoral Society and as such organized events such as the yearly research exhibition. I have been presenting my work at different Departments within Penn State and the LMU, other academic institutes (Senckenberg, Leiden University, Utrecht University, Erskine College, Harvard University, Nottingham University, Max Planck Institute, University of Cambridge) and at several international meetings (FGC, ESEB, SICB, YSF) in the form of a lecture. Lectures regarding the research performed the last 3 years have also been planned for upcoming October (Fungal Molecular Biology Meeting in Berlin and University of Copenhagen) and November (University of Düsseldorf).
Lastly, I have also applied for new funding to continue the research performed during this project and dive deeper into the molecular mechanisms that underlie parasitic behavioural manipulation. As a result I will continue my work at the LMU starting September 2015 as a Humboldt Research Fellow. During the project period I have been awarded a “The Huck Institutes of the Life Sciences Shared Technology Facilities Pilot Projects” grant, which funded part of the costs of my LC-MS experiments, a “Pennsylvania State University Post Doctoral Award” that aided towards the costs for travel to the ESEB conference, a “Genetics Society of America DeLill Nasser Award for Professional Development in Genetics”, which aided towards the costs for travel to the ECFG conference, and “Friedrich Bauer Stiftung Funding”, which allowed me to buy essential equipment for RNA extractions.
Project Management
The time line for year 1 and 2 of the Outgoing Phase have been swopped. The reason for restructuring the time line has been an unexpected set back in terms of resources and equipment at the Outgoing Host Institute. At the start of the project the arrival of the Laser Capture Microscope had been delayed by a couple of months. Furthermore, the grant that would support sequencing, assembly and annotation of the Ophiocordyceps genome in collaboration with the Broad Institute did not come through, leaving this project without a genome to map the RNAseq data to. Therefore, the choice was made to start with Objective 2 first, which would allow some time to solve the problems that had arisen with Objective 1. Subsequently, the Laser Capture Microscopy approach has proven to not be a viable approach resulting the change of the approach to mixed transcriptomics. This does not change the goal of Objective 1 and rather results in the positive outcome that this way not only the fungal transcriptome during manipulation can be assessed (as proposed), but also the ant transcriptome. This furthermore has led to a collaboration with Dr Andreas Brachmann from the LMU (the return phase institution) and Dr Robin Ohm from Utrecht University concerning the genome sequencing. The set back of having no genome when the project was started has thus resulted in an extra training opportunity, namely that of genomics.
Objective 2 has been achieved in the first year of this project. Objective 1 has taken a bit longer due to above mentioned set backs and extra work regarding the generation of a novel genome and has been achieved half-way into the third year. Objective 3 has been started while analyses for Objective 1 were still in progress. Setting up the protocols for Objective 3 have been very time consuming, as well as designing the right set up for proper RT-QPCR analysis. While this has been difficult, this has provided me with chronobiology training and advanced RT-QPCR training as envisioned. Objective 3 has however not been completely achieved in terms of final results. The important basis has none the less been made and the preliminary results are the first steps towards unravelling the molecular basis of the clock of the fungal parasite as was proposed. I will continue this work on new funding (a Humboldt Research Fellowship) for which I applied during the last year of this project.