FY2014 Annual Report

Information Processing Biology Unit

Unit: Information Processing Biology Unit
Principal Investigator: Ichiro Maruyama
Research Theme: Information Processing by Life

Abstract

We are interested in understanding how cellular networks in the nervous system processes environmental information to regulate animal behaviors, including decision-making, learning and memory. In order to survive, animals must closely monitor environmental and body fluid pH. However, little is known about how animals monitor alkaline pH in the environment at molecular and cellular levels. In this study, we employ C. elegans as a model organism. While C. elegans is attracted to mildly alkaline pH, it is repelled by strongly alkaline pH. In previous years, we demonstrated that multimodal sensory neurons (ASH) sense strong environmental alkalinity through activation of transient receptor potential vanilloid type (TRPV) ion channels encoded by the osm-9 and ocr-2 genes. Furthermore, mildly alkaline pH is perceived by a gustatory sensory neuron, ASEL, in which GCY-14, a receptor-type guanylyl cyclase, serves as a cell-surface sensor molecule. We also found that the behavioral decision-making is the result of competition between the two sensory neurons, ASEL and ASH. In the present fiscal year, we have tried to neuronal circuits responsible for the decision-making by behavioral analysis of mutants and Ca2+ imaging analyss of neuronal activity. 

In previous years, we also tried to develop protocols to study learning and memory in C. elegans, and found that the animal can learn and form associative long-term memory. It can learn to associate an appetitive or aversive olfactory stimulus with an unconditioned stimulus, and can retain the associative long-term memory. In the present fiscal year, we have been trying to map memory traces or engrams of the associative memories.

We are also interested in understanding how neurons/cells detect extracellular information and transmit it into inside the cell. For the last three decades, ligand-induced dimerization has been widely thought to be a common property of transmembrane signaling mechanisms for all known growth factor and cytokine receptors, among others. In previous years, however, we found that receptors for epidermal growth factor (EGF), nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF) have a preformed, yet inactive, dimeric structure prior to ligand binding. We proposed an alternative ‘rotation model’, in which ligand binding to the extracellular domains induces a rotation of the transmembrane domains in parallel to the plane of the membrane. This activates the intracellular domains, which often encode or physically interact with enzymes such as kinases, by rearranging their dimeric structures. To examine the model, we are currently trying to determine three dimensional structures of the receptor dimers. We are also analyzing co-operative interaction between EGF and it's receptor EGFR, which is observed only in dimers.

Graduate students are also now producing results from their own projects.

These results provide insights into the molecular mechanism underlying information transfer from the outside of neurons/cells to the inside, as well as an understanding of neuronal networks that control animal behavior in response to external stimuli. These findings may also be invaluable in developing pharmaceuticals for human diseases such as cancers and mental diseases.

1. Staff

  • Rehab Fouad Ahmed Abdelhamid, Staff Scientist (since July 1, 2014)
  • Satomi Ito, Staff Scientist (since April 1, 2014)
  • Takeshi Mise, Staff Scientist (until March 31, 2015)
  • Takashi Murayama, Staff Scientist
  • Hideki Muto, Staff Scientist (until August 31, 2014)
  • Hitomi Ohtaki, Research Administrator
  • Endang Rinawati Purba, Postdoctoral Scholar (since February 1, 2015)
  • Eiichiro Saita, Technical Staff (since May 1, 2014)

2. Graduate and other students

  • Dina Mostafa Abdelazim, Rotation Student (January 13-April 18, 2015)
  • Mohamed Abdelhack, PhD Student
  • Tosif Ahamed, PhD Student
  • Filippo Birocchi, Research Intern (July 11-September 2, 2014)
  • Erin Boldt, Research Intern (January 8-May 14, 2014)
  • Leonidas Georgiou, Rotation Student (January 9- April 30, 2014)
  • Kazuto Kawamura, PhD Student
  • Christina Lee, Rotation Student (September 1-December 26, 2014)
  • Viktoras Lisicovas, PhD Student
  • Ziyuan Wang, Research Intern (February 28-March 26, 2015)

3. Collaborations

Theme: Functional analyses of small G protein and their downstream proteins

  • Type of collaboration: Joint research
  • Researchers: 
    • Ken-ichi Kariya, University of the Ryukyus
    • Tsuyoshi Asato, University of the Ryukyus
    • Kimiko Nonaka, University of the Ryukyus

Theme: Study of signal transduction pathways that regulate cellular functions

  • Type of collaboration: Joint research
  • Researchers: 
    • Ken-ichi Kariya, University of the Ryukyus
    • Tsuyoshi Asato, University of the Ryukyus
    • Kimiko Nonaka, University of the Ryukyus

Theme: Phenotype analysis of the Rap2 knockout mice toward an understanding of molecular mechanisms underlying the pathology

  • Type of collaboration: Joint research
  • Researchers: 
    • Ken-ichi Kariya, University of the Ryukyus
    • Tsuyoshi Asato, University of the Ryukyus
    • Kimiko Nonaka, University of the Ryukyus

Theme: Social interaction and anxiety of genetically modified mice

  • Type of collaboration: Joint research
  • Researchers: 
    • Ken-ichi Kariya, University of the Ryukyus
    • Tsuyoshi Asato, University of the Ryukyus
    • Kimiko Nonaka, University of the Ryukyus

4. Activities and Findings

4.1. Project Aims

All forms of life are separated from their environments by cell membranes, and all neurons/cells have cell-surface receptor proteins that span the membrane in order to transmit external information, such as environmental changes and cell-cell communications, into inside the cell. Such information flow is fundamental for all forms of life, from bacteria to humans. Dysregulation of cell surface receptor molecules causes a variety of impairments, including mental and developmental diseases and cancer in humans. (1) We wish to understand the information processing through cellular networks. Specifically, we want to know how the external information is sensed and transmitted through sensory neurons, how it is processed by the nervous system, and how it controls animal behaviors, including decision-making, learning and memory. (2) We wish also to understand at the molecular level how the external information is sensed and transmitted into the inside of neurons/cells by cell-surface receptors. We also seek to know how the information is processed, how it is transferred to other parts of the cells, and how it regulates other cellular activities.

4.2. Progress Report

4.2.1. Information processing by the nervous system

4.2.1.1. Alkalinity sensation in C. elegans

4.2.1.1.1. Background

Like other animals, C. elegans detects various environmental cues, such as tastes and odors, mainly through its amphid sensilla. The amphids are the largest chemosensory organs, and each amphid includes 12 sensory neurons (ADF, ADL, AFD, ASE, ASG, ASH, ASI, ASJ, ASK, AWA, AWB, and AWC) with ciliated dendrites, as well as one sheath and one socket cell. Apart from AFD, the sensory dendrites of 11 neurons penetrate the sheath-cell ending, and the cilia of eight of these neurons, except those of AWA, AWB, and AWC, extend into the doughnut-like pore created by the socket cell where they are directly exposed to the external medium. These amphid neurons have roles in chemotaxis, thermotaxis, mechanosensation, osmotaxis, and dauer pheromone sensation.

Animals can survive only in a narrow pH range, and continuously monitor the pH of environment and body fluids. However, little is known about how animals monitor alkaline pH in the environment. Mildly alkaline pH in the environment is one of attractive cues for C. elegans. From pH 2.8-10.4, wild-type animals prefer higher alkaline pH ranges, but they avoid pH ≥11.0 and ≤4.0. However, neurons and cell-surface molecules that sense mildly and strongly alkaline pH have not been elucidated in C. elegans. Neural networks responsible for this worm’s attraction to mildly alkaline pH and aversion from strongly alkaline pH can also be efficiently analyzed, since wiring diagrams of all neurons have been reconstructed from electron micrographs of serial thin sections of the whole C. elegans body. 

4.2.1.1.2. Sensation of mildly alkaline pH

Previously, we devised an agar-plate assay with a linear pH gradient in order to investigate cellular and molecular bases for C. elegans’ chemotaxis toward mildly alkaline pH. Along the pH gradient from pH 6.8-8.5, wild-type animals were attracted to higher pH regions, whereas che-1 mutants defective in chemosensory ASE neurons were not. To search for a molecular sensor for the mildly alkaline pH, we performed a series of RNAi knock down experiments of genes encoding various channels, cell-surface receptors, and guanylyl cyclases, in conjunction with analysis of chemotaxis mutants. Among the genes analyzed, tax-2, tax-4 and gcy-14 were found to be involved in C. elegans’ chemoattraction to mildly alkaline pH. By imaging Ca2+ concentration changes using a Ca2+-sensing fluorescent protein, we also found that ASE-left (ASEL) is activated by a pH up-shift. The transmembrane receptor-type guanylyl cyclase (RGC) GCY-14, which is tagged with green fluorescent protein (GFP), was localized to ASEL sensory cilia, and in the gcy-14 mutant, ASEL did not respond to a pH up-shift. While wild-type ASI sensory neurons do not respond to pH up-shift, ASI neurons ectopically expressing GCY-14 were activated by a pH up-shift from pH 7.0-10.0. This suggests that GCY-14 is sufficient for alkaline pH sensation. These results indicate that GCY-14 acts as a mildly alkaline pH sensor, and an increased concentration of cGMP opens the cGMP-gated cation channel TAX-2/TAX-4 for activation of ASEL. We have further dissected the GCY-14 molecule in order to determine which part of it recognizes environmental alkaline pH. GCY-14 consists of an extracellular domain, a membrane-spanning alpha-helix, and intracellular kinase-homology, and guanylyl cyclase domains. We swapped these domains one by one with those of other alkaline pH-insensitive RGCs, DAF-11 and ODR-1, to identify a domain that detects mildly alkaline pH. Results of this experiment demonstrate that the extracellular domain is responsible for the sensation. Furthermore, the results also indicate that GCY-14 acts as a homodimer on the cell surface of ASEL cilia. we have also tried to identify a histidine residue(s) in the extracellular domain that may play an essential role in alkaline-pH sensation. A mutated GCY-14 protein that substitutes Gln for His-174, was defective in rescue experiments of gcy-14 mutants, indicating that His-174 plays a critical role in alkaline pH sensation. Furthermore, when GCY-14 was ectopically expressed in alkaline pH-insensitive ASG chemosensory neurons, ASG was activated by pH up-shift from 6.8-10.0. This indicates that GCY-14 is a sensor molecule for mildly alkaline pH (Murayama et al., 2013).  

4.2.1.1.3. Sensation of strongly alkaline pH

During the work on mild alkalinity sensation described above, we found that C. elegans avoids higher pH than ~10.5 as a noxious stimulus in chemotaxis assays using agar plates divided into four quadrants. Using the plate assay, in previous years, we identified several genes that play pivotal roles in the avoidance behavior: che-2, dyf-3, osm-1, and osm-5, which are required for chemosensation, and osm-9 and ocr-2, which encode subunits of type-V transient-receptor potential (TRPV) channels. Rescue experiments that employed dyf-3 and osm-9 mutants by injecting the respective cDNA into ASH chemosensory neurons indicate that ASH is a major neuron responsible for strongly alkaline-pH sensation in C. elegans. This was confirmed by Ca2+ imaging in which a Ca2+ concentration increase was observed in ASH following stimulation with strongly alkaline pH. We further confirmed that ASH sensory neurons are responsible for nociception of strongly alkaline pH using genetic rescue experiments and laser microsurgery of the neurons. Furthermore, genetic rescue of osm-9 mutants by specifically expressing OSM-9 in ASH showed that TRPV channels play an essential role in sensing of high pH. Ca2+ imaging in vivo also revealed that ASH neurons were activated by high pH stimulation, but ASH of osm-9 or ocr-2 mutants were not. These results demonstrate that in C. elegans, high pH is sensed by ASH nociceptors through opening of OSM-9/OCR-2 TRPV channels (Sassa et al., 2013). We have also found that C. elegans mutants deficient in a G-protein a subunit, GOA-1, failed to avoid strongly alkaline pH with normal Ca2+ influx into ASH. These results suggest that GOA-1 regulates signal transmission downstream of Ca2+ influx through OSM-9/OCR-2 TRPV channels in ASH (Sassa & Maruyama, 2013).

4.2.1.1.3. Decision-making in C. elegans alkaline pH chemotaxis

As described above, monitoring of environmental and tissue pH is critical for animal survival, and C. elegans is attracted to mildly alkaline pH, but avoids strongly alkaline pH. However, little is known about how the behavioral switching or decision-making occurs. Genetic dissection and Ca2+ imaging have previously demonstrated that ASEL and ASH are the major sensory neurons responsible for attraction and repulsion, respectively. In this fiscal year, we have found that unlike C. elegans wild type, mutants deficient in ASEL or ASH were repelled by mildly alkaline pH, and were strongly attracted to alkaline pH, respectively. These results suggest that signals through ASEL and ASH compete to determine the animal’s alkaline-pH chemotaxis. Furthermore, mutants with two ASEL neurons were more efficiently attracted to mildly alkaline pH than the wild type with a single ASEL neuron, indicating that higher activity of ASEL induces stronger attraction to mildly alkaline pH. This stronger attraction was overridden by normal activity of ASH, suggesting that ASH-mediated avoidance dominates ASEL-mediated attraction. Thus, C. elegans chemotactic behaviors to alkaline pH seems to be determined by signal strengths from the sensory neurons ASEL and ASH, and the behavior decision-making seems to be the result of competition between the two sensory neurons (Murayama & Maruyama, 2013). To test this working model, this fiscal year we have analyzed downstream information flows from ASEL and ASH. As shown in Figure 1, ASEL and ASH are connected to the first layer of interneurons, AIA, AIB, AIZ and AIY. In addition, ASH is also connected to command interneurons, AVA, AVB, AVE and AVD. A transcription factor, TTX-3, is expressed in AIA and AIY, and is required for functional differentiation of these interneurons. The ttx-3 mutant animals were more efficiently attracted to pH 10 than the wild type, or failed to more efficiently avoid stronger alkaline pH than pH 10, while they avoided normally pH 11.4. The hen-1 gene, which encodes a secreted protein, is expressed in ASE and AIY. The hen-1 mutants avoided a wide range of alkaline pH higher than pH 8.8, indicating that they failed to be attracted to mildly alkaline pH. SCD-2 is a putative receptor for HEN-1, and functions in AIA. GLC-3, a glutamate-gated Cl- channel, also functions at least in AIA and AIY. The glc-3 mutants normally avoided strongly alkaline pH, but are poorly attracted to mildly alkaline pH. Taken together, AIA interneurons seem to send the ASEL attractive signal to downstream of the circuit.

                                                                                                       

Figure 1. Model neural networks that may be involved in alkaline-pH chemotaxis. Interneurons that regulate motor neurons for head movement are indicated by a yellow background, and interneurons that are involved in forward and backward locomotion are shown in blue and red, respectively. Arrows and I-shaped bars represent chemical synapses and gap junctions, respectively. Thickness of arrows and I-shaped bars indicate relative strengths of the connections.

4.2.1.2. Associative learning and memory in C. elegans

4.2.1.2.1. Background

C. elegans is a model organism suited for the study of learning and memory. The hermaphrodite nervous system consists of only 302 neurons, and neural circuits, chemical synapses, and electrical gap junctions of these invariant neurons have completely been reconstructed from serial thin sections of electron micrographs. The worm body is transparent throughout its life so that neural activities can directly be observed using Ca2+-sensitive and/or voltage-sensitive fluorescent proteins in living animals.

Associative learning in C. elegans was first suggested from the finding that worms return to their temperature of cultivation if they had food at that temperature. Most of learning paradigms in C. elegans are based on pairing chemical cues or cultivation temperature with food or starvation. Rather than pairing chemical cues with food or starvation, however, it would be preferable for subsequent analysis of neural circuits responsible for associative learning and memory to use two defined chemical cues for conditioning of animals in associative learning paradigms.

4.2.1.2.2. Appetitive olfactory learning and long-term associative memory

In previous years, we developed an associative learning paradigm in C. elegans by classical (Pavlovian) conditioning of the animals with 1-nonanol, as a conditioned stimulus (CS), and potassium chloride (KCl) as an unconditioned stimulus (US). Before conditioning, animals avoided nonanol, an aversive olfactory stimulus, and were attracted by KCl, an appetitive gustatory stimulus, in chemotaxis assays. In contrast, after eight-cycle massed (without intertrial intervals, ITI) or spaced (with 10-min ITI) training, animals were attracted to nonanol. Memory induced by the massed training was extinguished within three hours, while the spaced training induced memory that was retained for more than 12 hours. Animals treated with cycloheximide or actinomycin D failed to form long-term memory by the spaced training, whereas memory induced by the massed training was not significantly affected. These results indicated that the memory formation by the spaced training, but not by the massed training, required protein synthesis and mRNA transcription. Therefore, the memories induced by the massed and spaced training are classified as short-term/middle-term (STM/MTM) and long-term (LTM) memories, respectively. In support of this, C. elegans mutants defective in nmr-1, encoding an NMDA receptor subunit, failed to form both STM/MTM and LTM, while mutations in crh-1 encoding the CREB transcription factor affected only formation of LTM. Last fiscal year, we determined sensory neurons that detect nonanol in C. elegans as a step toward elucidating the neural networks responsible for associative learning and memory with 1-nonanol and KCl. Olfactory avoidance assays of animals whose amphid AWB or ASH neurons were separately ablated by laser microsurgery, demonstrate that AWB is responsible for the perception of a low concentration, 0.1%, of 1-nonanol (Nishijima & Maruyama, manuscript in preparation). We are currently trying to identify neurons required for learning and memory using rescue experiments of the mutants.

Figure 2. Diagrams showing steps for associative conditioning of C. elegans by using nonanol as a CS and KCl as a US. These steps are repeated several times with or without the rest interval for training.

4.2.1.2.3. Aversive olfactory learning and long-term assicuatuve memory

In previous years, we also developed protocols for classical conditioning of C. elegans with 1-propanol, as a CS and hydrochloride (HCl), pH 4.0, as a US. Before conditioning, worms were attracted to 1-propanol, and avoided HCl in chemotaxis assays. However, after spaced or massed training, animals were either not attracted at all or were repelled by propanol on the assay plate. The memory after the spaced training was retained for 24 hours, while the memory after the massed training did not even persist for 3 hours. Animals pretreated with transcription and translation inhibitors failed to form the memory by spaced training, whereas the memory after massed training was not significantly affected by inhibitors and was sensitive to cold-shock anesthesia. Furthermore, memory after spaced training was reasonably disrupted by extinction learning, in which the worm is repeatedly exposed to the CS in the absence of US. Therefore, memories after spaced and massed training can be classified as LTM and STM/MTM, respectively. Consistently, like other organisms including Aplysia, Drosophila, and mice, C. elegans mutants defective in nmr-1, encoding an NMDA receptor subunit, failed to form both LTM and STM/MTM, while mutations in crh-1 encoding the CREB transcription factor affected only LTM (Amano & Maruyama, 2011). We are currently trying to identify neurons required for learning and memory using rescue experiments of the mutants.

4.2.2. Information processing by neurons/cells

4.2.2.1. Background

The bacterial cell-surface receptor, Tar, recognizes aspartate molecules in the environment, and directs bacterial cells toward higher concentrations of the attractant as a nutrient, or toward lower concentration of repellents, such as nickel and cobalt ions. This transmembrane signaling is mediated by Tar in its homodimeric form on the cell surface. We have previously shown that Tar activity is regulated by its ligands, which bind to the extracellular domain of the receptor and lock/freeze the rotational/twist movement of the receptor’s transmembrane domains (Maruyama et al., 1995). This locking/freezing of the rotation/twist at one position by the attractant is likely to inhibit the associated histidine kinase Che A, while the locking/freezing at another position by the repellent seems to activate the kinase activity (the rotation/twist model).

We also analyzed the molecular mechanism underlying activation of the human epidermal growth factor receptor (EGFR) family of cell-surface receptor tyrosine kinases, also known as ErbB or HER (Moriki et al., 2001). EGF/ErbB receptors play a pivotal role in the development of organisms, and are frequently implicated in human cancers. Furthermore, these receptors also regulate neural activities, and mutations of these receptor genes are frequently associated with mental diseases. The receptor family consists of four members, EGFR/ErbB1, ErbB2/Neu/HER2, ErbB3/HER3 and ErbB4/HER4, and has a large (~620 amino acid) extracellular ligand-binding region, a single, transmembrane alpha-helix, and an intracellular region containing the tyrosine kinase and its regulatory domain. They form a network of homo- and heterodimers. ErbB2 can only be regulated indirectly, and is thought to be the preferred heterodimerization partner for other ErbB receptors. ErbB3, on the other hand, must associate with an ErbB family member that has an active tyrosine kinase in order to respond to its own ligand, neuregulin (NRG).

Ligand-induced dimerization has widely been thought to be a property common to the transmembrane signaling mechanism of all known growth factor receptors including the EGF/ErbB receptors (Ligand-induced Dimerization Model). According to the model, receptor dimerization is responsible for autophosphorylation of the intrinsic kinase activity, which is mediated by an intermolecular process. The model holds that ligand binds to the monomeric form of the receptor, inducing its dimerization and resulting activation. However, it remains controversial whether the receptor actually has a monomeric or dimeric structure prior to ligand binding

Using chemical cross-linking and sucrose density-gradient centrifugation, we recently discovered that in the absence of bound ligand, EGFR has the ability to form a dimer and that the majority (>80%) of receptors exist as preformed dimers on the cell surface. We also analyzed receptor dimerization by inserting cysteine residues at strategic positions along the longitudinal axis of the alpha-helical extracellular juxtamembrane region. Mutant receptors spontaneously formed disulfide bridges and transformed NIH3T3 cells in the absence of ligand, depending upon the positions of the cysteine residues inserted. Kinetic analysis of disulfide bonding indicates that ligand binding induces flexible rotation or twist of the juxtamembrane region of the receptor in a plane parallel with the lipid bilayer. The binding of an ATP competitor to the intracellular kinase domain also induced similar flexible rotation/twist of the juxtamembrane region. All disulfide-bonded dimers had flexible ligand-binding domains with the same biphasic affinities for ligand as the wild type. Based on these results, we have proposed an alternative ‘rotation/twist’ model (Fig. 3) for the molecular mechanism of EGF receptor activation, in which ligand binding to the flexible extracellular domains of the receptor dimer induces rotation/twist of the juxtamembrane regions, hence the transmembrane domains, rearrange the kinase domains for the receptor activation. Indeed, this rotation/twist model (Moriki et al., 2001; Tao and Maruyama, 2008) is consistent with recent results by others in which the receptor kinase, transmembrane and unactivated extracellular domains are shown to have homodimeric structures.                                    

Figure 3. ‘Rotation/twist’ model for molecular mechanism underlying activation of EGFR by EGF binding (Maruyama, 2014).

To support the ‘rotation/twist’ model, we have recently determined preformed, homo- and heterodimeric structures of EGFR and ErbB2 at physiological expression levels (~104 molecules per cell), using fluorescence microscopy, fluorescence resonance energy transfer (FRET) and fluorescence cross-correlation spectroscopy (FCCS) (Liu et al., 2007). When fluorescent protein (FP)-fused EGFR and ErbB2 were expressed on the cell surface of Chinese hamster ovary cells at physiological expression levels, FRET was detected between the donor and acceptor FPs in the absence of ligand. Furthermore, cross-correlation between FPs separately fused to EGFR or ErbB2 was also observed by FCCS, indicating that EGFR and ErbB2 molecules diffuse together as homo- or heterodimers in the cell membrane. These results demonstrate that prior to ligand binding, the cell-surface receptors can spontaneously form homo- and heterodimers, irrespective of their expression levels ranging from ~2 x 104 to ~5 x 106 molecules per cell.

Furthermore, we have been analyzing preformed homo- and heterodimeric structures between all the members, EGFR, ErbB2, ErbB3, and ErbB4, of the receptor family by employing bimolecular fluorescence complementation (BiFC) assay. We have found that all members display preformed, yet inactive, homo- and heterodimeric structures in the absence of bound ligand (Tao & Maruyama, 2008). Ligand-independent dimerization of EGF/ErbB receptors occurs in the endoplasmic reticulum (ER) before newly synthesized receptor molecules reach the cell surface. Furthermore, we have also found that ErbB3 was localized in the nucleus when expressed alone or together with ErbB4. When coexpressed with EGFR or ErbB2, however, ErbB3 was located in the plasma membrane. These results indicate that all the EGF/ErbB receptors exist as homo- and heterodimers before ligand binding, consistent with the ‘rotation/twist’ model. ErbB receptors exist as dimers on the cell surface, mainly through interaction between their transmembrane domains, intracellular kinase domains and C-terminal tails. Receptor dimers have flexible extracellular domains, and perhaps can take two major conformations, closed (tethered) and open (untethered) states. Ligand binding to the open form of the receptor dimer stabilizes the extracellular domains, resulting in approximately 140° rotation or twist of the transmembrane domains about its helix axis parallel to the plane of the cell membrane, dissociate the symmetric back-to-back kinase domains, and then rearrange the kinase domains to take head-to-tail asymmetric configuration for the receptor activation.

As described above, different cell-surface receptors, bacterial Tar and human EGF/ErbB receptors, seem to be similarly regulated by their ligands in order to transmit the extracellular information to the inside of the cell. Ligand binding regulates the rotation/twist of the receptor’s transmembrane domain parallel to the plane of the plasma membrane. Therefore, we have been continuing to test the “rotation/twist” model for the molecular activation mechanism of other cell-surface receptors, including Tar, EGF/ErbB receptors and neurotrophin receptors. Indeed, we have recently elucidated that the neurotrophin receptor TrkA is present as a preformed, yet inactive, dimer in living cells (Shen & Maruyama, 2011). We have also previously found that the intracellular domain of the EGF/ErbB receptors plays a crucial role in the spontaneous formation of receptor dimers (Tao & Maruyama, 2008). Thus, an increasing number of studies demonstrate that many transmembrane receptors, which include receptors previously thought to be activated by ligand-induced receptor dimerization, exist as preformed, yet inactive, homo- and heterodimers in living cells. These receptors include the aspartate receptor Tar, EGF/ErbB receptor family members, erythropoietin receptor (EpoR), growth hormone receptor (GHR), Toll-like receptor-9 (TLR9), natriuretic peptide receptor A (NPRA), nerve growth factor (NGF) receptor TrkA, and brain-derived neurotrophic factor (BDNF) receptor TrkB. In the cases of Tar, EGFR, GHR, and NPRA, it has been proposed that ligand binding induces or stabilizes the rotation/twist of transmembrane domains of preformed receptor dimers for the rearrangement of intracellular domains necessary for activation.

4.2.2.2.  Crystal structure of Tar with repellent

Three-dimensional (3D) structures of the extracellular ligand-binding domain of the bacterial aspartate receptor Tar, with and without bound aspartate were determined 20 years ago, and others have reported that the transmembrane domain shifts vertically 1-2 Å when aspartate binds. This suggests that the receptor structures without and with bound aspartate are very similar, consistent with our ‘rotation/twist’ model for the mechanism of Tar activity regulation by aspartate (Maruyama et al., 1995). Furthermore, the ‘rotation/twist’ model predicts that another cofactor for Tar, nickel, a repellent in E. coli chemotaxis, stabilizes the transmembrane domains in a distinct rotational orientation. That is, nickel induced rotation/twist of the transmembrane domains parallel with the plane of the cytoplasmic membrane. To test this model, we are currently trying to crystalize the extracellular domain of Tar in the presence of Ni2+ or Co2+ ions (Mise et al., 2014).

4.2.2.3. Kinetic analysis of EGF interaction with EGFR

To analyze kinetics of EGF interaction with EGFR, this fiscal year we have employed total internal reflection fluorescence microscopy (TIRFM) to observe single molecules of fluorescently labeled EGF bound to EGFR on the cell surface. Using the technique, we found that one or two EGF molecules bound to its receptor on the cell surface, and a majority of the events was very brief (less than 3 seconds). By employing multicolor TIRFM imaging, we are currently analyzing association and dissociation rates between EGF and EGFR at the single molecule level on the surface of living cells, and association of EGFR with downstream effector proteins in the cytosol.

4.2.3. Graduate student projects

4.2.3.1. Optogenetic approach to the study of learning and memory in C. elegans (Viktoras Lisicovas)

Project summary: The small nervous system of C. elegans coordinates a wealth of behaviors and exhibits learning and memory. We suggest that imposing a novel memory onto the nervous system alters activity patterns of implicated neurons. Experiences can be stored by altered molecular interactions in sensory cells or changes in strength of glutamatergic synapses between neurons. The location and extent of the created engram should correlate with the observed memory persistence. To assess these hypotheses, we are developing an optical associative conditioning paradigm using optogenetic stimulation of sensory neurons. We will further measure ensuing change in putative networks using genetically encoded calcium indicators. This will elucidate the formation and locations of memory engram in the neural tissue. Such knowledge may prove useful in finding better treatments for memory disorders such as Alzheimer’s disease, designing of brain-computer interfaces and more robust neural network architectures.

Progress report: Gustatory and olfactory cues have been previously shown to predictably alter behavior of animals upon training. A conditioning paradigm, using an acidic solution as a US and 1-propanol as a CS, was shown to induce odor associations lasting for as long as 24 hours. Based on report that optogenetic activation of ASH sensory neuron produces behavioral responses similar to those induced by osmotic stress and acidic pH, we set out to develop an optogenetic associative conditioning paradigm. Optogenetic stimulation targets single, identified neurons, offering a precise control over sites of information inputs to the neural network.  An auto-tracking confocal microscope was used to implement an aversive conditioning paradigm with optical stimulation of ASH as a US and exposure to vapor of 1-paropanol on a cotton swab as a CS (Figure 4A). We tested, whether associative memory can be formed through activation of ASH neurons with channelrhodopsin-2 (ChR2) photosensitive ion channel. By tracking individual animals, we found that conditioning by optogenetically activated ASH nociceptive neuron can change behavioral response (probability of a reversal) to 1-propanol. This induced plasticity persists for at least 6 hours depending on intertrial interval (ITI) used in conditioning procedure (Figure 4B).                       

Figure 4. Optogenetic conditioning of C. elegans.

4.2.3.2. Forward genetic screen for new regulators of adult-onset neuromuscular dysfunction (Kazuto Kawamura).

Background: The causes of adult-onset neuromuscular disorders such as amyotrophic lateral sclerosis (ALS) and myopathies remain largely unknown. C. elegans is a model organism that has been used to identify key genes involved in complex biological processes such as aging. For example, a C. elegans forward genetic screen led to the identification of a mutation in the age-1 gene that causes a 40% extension of mean lifespan (Friedman and Johnson, Genetics 1988). Surprisingly, an unbiased forward genetic screen has not been conducted for adult-onset disease phenotypes in C. elegans. Approximately 65% of human disease genes have a counterpart in C. elegans, making it a promising animal model to find new disease-relevant genes (Sonnhammer and Durbin, Genomics 1997). Taking a forward genetic approach can lead to findings that contribute to basic knowledge as well as clinical applications.

Progress report: I established a procedure called the “edge assay” to test the motor function of worms. The edge assay is conducted on a 9-cm circular agar plate with Escherechia coli OP50 strain bacterial feed located only on the outer edge. Up to several hundred adult worms are placed on the center of the plate. Motile worms reach the E.coli and remain close to the edge but slow or paralyzed worms can be found remaining in the center of the plate. Optimal time points for the worms to complete the task were determined by comparing the completion rates with known developmental motor deficit mutants. A two-step screening procedure using the edge assay was conducted to isolate worms that were able to complete the task on adult day 1, but unable to complete the task on adult day 5. The screening procedure on adult day 1 removes worms that likely have abnormal development and cannot move properly during larval stages or young adulthood. Only the worms that reached the edge on adult day 1 are selected and tested again on adult day 5. Worms that fail to reach the edge on adult day 5 are classified as adult-onset motor deficit mutants and kept for further analysis. Worms were chemically mutagenized using ethyl methanesulfonate (EMS). A total of 1000 haploid genomes were screened, and 22 viable mutants were obtained.

In order to confirm that the mutants were in fact adult-onset motor deficit mutants, we tested the offspring of several mutant lines. The A20-1 mutant is a promising candidate with no obvious morphological abnormalities (Figure 5A). The edge assay completion rate in one hour for the backcrossed A20-1 line was not significantly different on the first four days of adulthood but showed a significant decline on adult day 5 (Figure 5B). This decline was similar to that of a C. elegans model of ALS based on a mutation in the SOD1 gene (G127insTGGGstop).

                               

Figure 5. Confirmation of A20-1 adult-onset motor deficit. (A) Representative images of N2 and A20-1 animals. No obvious morphological abnormalities were seen in A20-1 animals. (B) Relative completion rates from adult day 1 to 5 in WT, SOD1-127X, and A20-1 animals (*p<0.05, **p<0.01).

4.2.3.3. Computational neuromodulation in C. elegans (Tosif Ahamed).

Background: Over the past two decades a computational theory called Reinforcement Learning (RL) has shown promise in integrating the activities of different monoamines (and other neuromodulators) under a common normative framework. The central hypothesis of RL is that neuromodulators are profoundly involved in the computations associated with decision making under uncertainty. Although RL has had considerable success in explaining the nature of phasic dopamine signaling in different animals, the theory has still not been able to account for the activity of other monoamines and the interactions between them. Most studies in RL have been carried out in mammalian systems where it's difficult to observe activities of different monoaminergic neurons in alive freely behaving animals. Moreover, decisions in mammals are influenced by many different factors most of which are out of experimenter's control.

The goal of this project is to study neuromodulation in C. elegans under the framework of RL using a combination of experimental and theoretical techniques. C. elegans offers many advantages, in addition to the genetic tractability and the fact that different monoaminergic neurons can be simultaneously observed in freely moving animals. Most of the behaviors displayed by C. elegans can be directly explained by either a drive to obtain food or stay away from danger suggesting that C. elegans might behave close to an ideal reinforcement learner. During the course of my PhD, I would like to address the following questions: (1) Are monoamines involved in the decision to leave an attractant? (2) What computations are the monoamines involved with and what are the interactions between them?

Progress report: Wild type C. elegans are attracted to certain chemicals such as sodium acetate (CH3COONa), and move towards the peak concentration of the chemical. Once at the peak it prefers to explore the area densely before deciding to go down the gradient (Figure 6A). The decision to leave the attractant is presumably driven by hunger and shows plasticity (Figure 6B) as the staying times decrease on subsequent encounters with the same attractant.

We study the role of monoamines in the attractant leaving behavior and the plasticity associated with this behavior. For the first set of experiments we tested all the monoamine mutants and their ability to leave the attractant compared to the wild type. Preliminary results suggest that dopamine and serotonin both modulate the decision of when to leave an attractant.

After establishing the involvement of monoamines in the attractant leaving behavior we will build a reinforcement learning model of the behavior. To test and build the model further we will observe calcium activity from all monoaminergic neurons of freely moving animals in the assay described above. Most of the tools required for these set of experiments are already in place. We will analyze the neural activity and the interactions between different monoamines and use the data to expand upon the model. To test the model and its predictions, we will do perturbations by optogenetics or exogenous application of monoamines.

(A)                            

(B)

                                       

Figure 6. (A) Wild type C. elegans when placed in the middle of a chemical attractant (sodium acetate in this case) prefer to stay around the attractant and sample the area densely however after about 45 minutes the animals leave the attractant. (B) The attractant leaving behavior is plastic as animals spend less time exploring the attractant on subsequent encounters. In preliminary experiments this decision of when to leave the attractant and the plasticity associated with decision is modulated by both dopamine and serotonin. 

 

5. Publications

5.1. Journals

  1. Maruyama, I. N. Mechanisms of activation of receptor tyrosine kinases: monomers or dimers. Cells (2014) 3(2): 304-330.  
  2. Sassa, T., Murayama, T. & Maruyama I.N. Decision making in C. elegans chemotaxis to alkaline pH. The FASEB (2014) 28. 
  3. Mise, T., Matsunami, H., Samatey, F.A., & Maruyama, I.N. Crystallization and preliminary X-ray diffraction analysis of the periplasmic domain of the Escherichia coli aspartate receptor Tar and its complex with aspartate. Acta Crystallographica Section F (2014) 1219-1223. 

5.2. Oral and Poster Presentations

  1. OIST International Workshop, Single Protein Dynamics in Cellulo (SPDC) 2014; OIST Seaside House, Okinawa, Japan (April 21-25, 2014). Title: Mechanisms of activation of receptor tyrosine kinases; I. Maruyama
  2. Sassa, T., T, Murayama and Maruyama, I.N. Decision making in C. elegans chemotaxis to alkaline, in ASBMB Annual Meeting, San Diego, CA, USA (April 26-30, 2014)
  3. 3rd Symposium on cancer chemotherapy: from academia to ant-cancer drug development; Bankoku Shinryokan, Okinawa, Japan (May 12, 2014). Title: Mechanism of activation of human epidermal growth factor receptor and its inhibitors. Maruyama, I. N.
  4. Shen, J. and Maruyama, I.N. TrkA and TrkB exist as pre-formed dimers on the surface of living cells, in Cold Spring Harbor Asia (CSHA)/NGF 2014 Joint Conference, Suzhou, China (June 23-27, 2014)
  5. Mise, T. and Maruyama, I.N. Structural Analysis of the ligand binding domain of the Aspartate Receptor Tar from Escherichia coli, in 14th Annual Meeting of the Protein Science Society of Japan, Yokohama, Japan (June 25-27, 2014)
  6. Murayama, T. and Maruyama, I.N. Behavioral changes in C. elegans chemotaxis to alkaline pH, in the C. elegans Development, Cell Biology, and Gene Expression Topic Meeting in association with the 6th Asia-Pacific C. elegans Meeting, Nara, Japan (July 15-19, 2014)
  7. Murayama, T. Behavioral changes in C. elegans chemotaxis to alkaline pH, in OIST internal seminar, Okinawa, Japan (August 22, 2014)
  8. Neuroscience 2014 (37th Annual Meeting of the Japan Neuroscience Society), Pacifico Yokohama, Kanagawa, Japan; September 11-13, 2014. Title: Behavioral choice in C. elegans chemotaxis to alkaline pH (September 11, 2014; Poster #: P1-230); Takashi Murayama, Toshihiro Sassa and Ichiro Maruyama.
  9. Murayama, T and Maruyama, I.N. Decision-making in C. elegans chemotaxis to alkaline pH, in Society for Neuroscience - Neruscience 2014, Washington, USA (November 15-19, 2014)
  10. Murayama, T. and Maruyama, I.N. Decision-making in C. elegans chemotaxis to alkaline pH: Competition between two sensory neurons, ASEL and ASH in The 2014 ASCB/IFCB Meeting, Philadelphia, USA (December 6-10, 2014)

6. Meetings and Events

6.1. International Workshop

Single Protein Dynamics in Cellulo (SPDC 2014) 

  • Date: April 21-25, 2014
  • Venue: OIST Seaside House
  • Organaizers: Ichiro Maruyama & Ulf Skoglund
  • Co-organizers: Masataka Kinjo & Thorsten Wohland
  • Participants: 49

Speakers:

  • Hiroko Bannai, Laboratory of Brain Function and Structure Nagoya University, Japan
  • Andrew Clayton, Engineering & Industrial Sciences Swinburne University of Technology, Australia
  • Duccio Fanelli, University of Florence, Italy
  • Charlotte Fournier, OIST, Japan
  • Mike Heilemann, Institute of Physical & Theoretical Chemistry, Johann Wolfgang Goethe University, Germany
  • Masataka Kinjo, Advanced Life Science, Hokkaido University, Japan
  • Akihiro Kusumi, Institute for Frontier Medical Sciences, Kyoto University, Japan
  • Jun Liu, Department of Pathology & Laboratory Medicine, University of Texas Medical School, USA
  • Tomoko Masaike, Tokyo University of Science, Japan
  • Shigetoshi Oiki, Department of Medicine, Fukui University, Japan
  • Yasushi  Sako, RIKEN, Japan
  • Takayuki Uchihashi, Institute of Science and Engineering, Kanazawa University, Japan
  • Vladana Vukojević, Karolinska Institutet Department of Clinical Neuroscience Center for Molecular Medicine, Sweden
  • Paul Wiseman, Depts. of Physics and Chemistry, McGill University, Canada
  • Thorsten Wohland, National University of Singapore, Singapore

6.2. Seminar

Title: Metabolism and functions of soluble epoxide hydrolase (sEH)

  • Speaker: Endang R. Purba, Kwansei Gakuin University
  • Date: October 3, 2014
  • Venue: C015, Level C, Lab1, OIST

Title: Crystal structures of two enzymes involved in degradation pathway I for pyridoxine, from Mesorhizobium loti

  • Speaker: Andrew Mugo, Kochi University
  • Date: November 25, 2014
  • Venue: B043, Level C, Lab1, OIST

Title: Complementary and conflicting forms of sensory adaptation during salt chemotaxis in Caenorhabditis elegans

  • Speaker: Netta Cohen, University of Leeds
  • Date: December 15, 2014
  • Venue: C016, Level C, Lab1, OIST