PACα– an optogenetic tool for in vivo manipulation of cellular cAMP levels, neurotransmitter release, and behavior in Caenorhabditis elegans

Authors

  • Simone Weissenberger,

    1. Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Frankfurt, Germany
    Search for more papers by this author
    • The present address of Simone Weissenberger is the Institute for Toxicology, University Wuerzburg, Versbacher Str. 9, D-97078 Wuerzburg, Germany.

    • These authors contributed equally to this study.

  • Christian Schultheis,

    1. Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Frankfurt, Germany
    Search for more papers by this author
    • These authors contributed equally to this study.

  • Jana Fiona Liewald,

    1. Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Frankfurt, Germany
    Search for more papers by this author
    • These authors contributed equally to this study.

  • Karen Erbguth,

    1. Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Frankfurt, Germany
    Search for more papers by this author
  • Georg Nagel,

    1. University Wuerzburg, Botanik I, Wuerzburg, Germany
    Search for more papers by this author
  • Alexander Gottschalk

    1. Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Frankfurt, Germany
    2. Frankfurt Institute for Molecular Life Sciences (FMLS), Goethe-University, Frankfurt, Germany
    Search for more papers by this author

Address correspondence and reprint requests to Alexander Gottschalk, Department of Biochemistry, Chemistry, and Pharmacy, Institute of Biochemistry, Goethe-University, Max-von-Laue-Straße 9, D-60438 Frankfurt, Germany. E-mail: a.gottschalk@em.uni-frankfurt.de

Abstract

J. Neurochem. (2011) 116, 616–625.

Abstract

Photoactivated adenylyl cyclase α (PACα) was originally isolated from the flagellate Euglena gracilis. Following stimulation by blue light it causes a rapid increase in cAMP levels. In the present study, we expressed PACα in cholinergic neurons of Caenorhabditis elegans. Photoactivation led to a rise in swimming frequency, speed of locomotion, and a decrease in the number of backward locomotion episodes. The extent of the light-induced behavioral effects was dependent on the amount of PACα that was expressed. Furthermore, electrophysiological recordings from body wall muscle cells revealed an increase in miniature post-synaptic currents during light stimulation. We conclude that the observed effects were caused by cAMP synthesis because of photoactivation of pre-synaptic PACα which subsequently triggered acetylcholine release at the neuromuscular junction. Our results demonstrate that PACα can be used as an optogenetic tool in C. elegans for straightforward in vivo manipulation of intracellular cAMP levels by light, with good temporal control and high cell specificity. Thus, using PACα allows manipulation of neurotransmitter release and behavior by directly affecting intracellular signaling.

Abbreviations used:
[cAMP]i

intracellular cAMP concentration

ACh

acetylcholine

ChR2

Channelrhodopsin-2

g.o.f.

gain-of-function

GFP

green fluorescent protein

lite-1

light-unresponsive-1

mEJP

miniature excitatory junction potential

mPSC

miniature post-synaptic current

NGM

nematode growth medium

NMJ

neuromuscular junction

PAC

photoactivated adenylyl cyclase

SV

synaptic vesicle

cAMP is a ubiquitous second messenger in intracellular signal transduction and involved in many cellular events and complex biological processes, including hormone signaling (Beavo and Brunton 2002), immune function (Torgersen et al. 2002), modulation of synaptic transmission (Kidokoro et al. 2004), and memory consolidation (Kandel 2001; Morozov et al. 2003). Synthesis of cAMP is accomplished by adenylyl cyclases, soluble or integral membrane proteins that convert ATP to cAMP. Because of the various functions of cAMP and its use in many different organisms, an extensive family of adenylyl cyclases has evolved. Usually, the enzyme activity is regulated by G-proteins. However, a photoactivated adenylyl cyclase (PAC) was isolated from the photosensory organelle of the freshwater flagellate Euglena gracilis (Iseki et al. 2002; Ntefidou et al. 2003). In this unicellular organism, PAC serves as an important photoreceptor that mediates cAMP-dependent phototaxis.

Photoactivated adenylyl cyclase is a blue-light receptor that is composed of four flavoprotein subunits: two PACα and two PACβ subunits. Each subunit consists of two BLUF (sensors of blue-light using FAD) domains binding FAD for photoreception (Anderson et al. 2005; Gauden et al. 2006) and two cyclase domains to catalyze the conversion of ATP to cAMP (Iseki et al. 2002). PACα and PACβ are independent in their catalytic function, and their ability to form cAMP can strongly and repeatedly be increased by blue light (Yoshikawa et al. 2005; Looser et al. 2009). However, the specific activity of PACα is about 100-fold higher than that of PACβ (Schroder-Lang et al. 2007).

To date, it has been shown that photoactivation of PACα can increase cAMP levels in HEK293 cells (Schroder-Lang et al. 2007) and led to a large increase in plasma membrane conductance of Xenopus oocytes when the blue-light receptor was coexpressed together with a PKA (cAMP-dependent protein kinase)-activated Cl channel (cystic fibrosis transmembrane conductance regulator – CFTR). In the sea slug Aplysia, injection of PACα into sensory neurons furthermore facilitated light-activated changes in spike width and amplitude (Nagahama et al. 2007). In adult Drosophila, stimulation of PACα in neurons modulates behavior by inducing bouts of hyperactivity, unusual freezing as well as a decline in grooming activity (Schroder-Lang et al. 2007).

The nematode Caenorhabditis elegans has proven to be a powerful in vivo model for neurobiological studies. Its elementary, yet versatile nervous system of exactly 302 neurons has been mapped down to the individual synapse, and numerous studies showed that basic protein machineries and neurotransmitters involved in mammalian neurotransmission are conserved (White et al. 1986; Miller et al. 1996; Bargmann 1998; Richmond 2007). However, compared to rodent models C. elegans is easily genetically tractable and most mutants with severe neurotransmission defects are viable.

In C. elegans, pre-synaptic cAMP plays a critical role in the regulation of locomotion. At least three Gα-signaling pathways (Gαo/i, Gαq, Gαs) appear to be involved in regulating different aspects of synaptic vesicle (SV) release and synaptic signalling. While Gαo/i acts inhibitory, the Gαq and Gαs pathways are involved in controlling the release of neurotransmitters and in driving locomotion (Reynolds et al. 2005; Schade et al. 2005; Charlie et al. 2006). The Gαq pathway exerts its effects on locomotion by activating phospholipase Cβ and producing diacyl-glycerol as a pre-synaptic second messenger. Ultimately, this increases release of the excitatory neurotransmitter acetylcholine (ACh) at neuromuscular junctions (Lackner et al. 1999; Miller et al. 2000). Another major effector important for the Gαq pathway is the RhoGEF (Guanine nucleotide exchange factor) Trio which activates RhoA and also drives locomotion (Williams et al. 2007). Neuronal Gαs and Gαq pathways converge to regulate synaptic activity, and the Gαs pathway depends on the Gαq pathway to exert its effects on locomotion. It is believed that in the cholinergic system of C. elegans an increase in the intracellular cAMP concentration ([cAMP]i) causes increased ACh release and can subsequently stimulate downstream muscle cells and neurons. Accordingly, gain-of-function mutations in the Gαs pathway, as well as gain-of-function mutations in the adenylyl cyclase ACY-1, result in increased neurotransmitter release and hyperactive, though highly coordinated locomotion (Schade et al. 2005). Yet, many questions regarding pre-synaptic functions of cAMP and the Gα pathway remain unresolved.

Manipulation of cellular signaling in live animals with the help of genetically encoded light-sensitive proteins, such as Channelrhodopsin-2 (ChR2) and Halorhodopsin, has become a highly studied and applied topic in recent years (Nagel et al. 2005; Schroll et al. 2006; Zhang et al. 2006, 2007; Liewald et al. 2008; Cardin et al. 2010). Among these ‘optogenetic’ tools, PAC is rather new. Previous attempts to influence intracellular cAMP levels in C. elegans by using membrane-permeable cAMP analogs were unsuccessful in inducing hyperactive locomotion, suggesting that the location and/or timing of cAMP increase are critical (Schade et al. 2005). In the present study, we used PACα in C. elegans for manipulation of intracellular cAMP levels of cholinergic neurons simply by illumination with blue light. Photoactivation led to behavioral changes, i.e. elevated swimming frequency and speed of locomotion, as well as a reduced number of backward locomotion episodes. Concomitantly, miniature post-synaptic current (mPSC) frequency at the neuromuscular junction was increased. Compared to the use of cAMP analogs, this optogenetic tool allows the in vivo manipulation of [cAMP]i in selected cells only and with high spatial and temporal control.

Materials and methods

Genetics

Caenorhabditis elegans strains were cultivated using standard methods on nematode growth medium (NGM) and fed E. coli strain OP50-1 (Brenner 1974). Transgenic strains were generated following standard procedures (Fire 1986). Strains used or generated: N2 (wild type), KG1180: lite-1(ce314), ZX784: lite-1(ce314);zxEx512[punc-17::GFP::PACα; pelt-2::mCherry]–‘Line 1’, ZX785: lite-1(ce314);zxEx513[punc-17::GFP::PACa; pelt-2::mCherry]–‘Line 2’, KG524: gsa-1(ce94), KG518: acy-1(ce2).

Molecular biology

Photoactivated adenylyl cyclase α cDNA (http://www.ncbi.nlm.nih.gov/protein/BAB85619.1) was kindly provided by Masakatsu Watanabe (Iseki et al. 2002). The punc-17::GFP::PACα construct was generated as follows: A DNA-fragment of green fluorescent protein (GFP) was PCR amplified from pAG48 (pacr-13::acr-13::GFP (Gottschalk et al. 2005); primers oSW24 (5′-GCCAGTGCTAGCATGAGTAAAGGAGAAGAACTTTTC-3′) and oSW28 (5′-GCCAGTGGTACCAGTCTGATCATTTGTATAGTTCATCCATGCCATGTGT-3′)) and subcloned into the punc-17-vector RM#348p (a gift from J. Rand) using NheI and KpnI to yield the punc-17::GFP intermediate plasmid. PACα cDNA was synthesized in vitro using the T7 cap scribe kit (Ambion, Austin, TX, USA) from the plasmid AB031225 and subcloned into pGEMR2 (a modified version of pGEM3z; Promega, Madison, WI, USA) to generate pGEMHE::PACα (Schroll et al. 2006). Thereafter, PACα cDNA was cloned from pGEMHE::PACα [primers oSW29 (5′-GACGTGATCATACATCCTTGTTTGGAAAGAAGG-3′) and oSW27 (5′-GCCAGTGGTACCTTAATGTTCATATTTGTGCGAACC-3′)] into the punc-17::GFP intermediate plasmid using BclI and KpnI to generate punc-17::GFP::PACα. A resulting in-frame stop codon between GFP and PACα from the BclI restriction site (5′-TGATCA-3′) was then mutated to a NotI restriction site (5′-GCGGCCGC-3′) by assembled PCR with primers A (5′-CCGGGCAATTGGCGATGGCCCTGTCC-3′), B (5′-GCGGCCCACACTCATGGGTTCAGGGGC-3′), C (5′-ACAAAGCGGCCGCGTACATCCTTGTTTGGAAAGAAGGCC-3′), and D (5′-CTTTCCAAACAAGGATGTACGCGGCCGCTTTGTATAGTTCATCCATGCC-3′), adding a nucleotide to restore the reading frame.

Generation of transgenic animals

Transgenic C. elegans were obtained by microinjection of 20 ng/μL of the punc-17::GFP::PACα plasmid and 20 ng/μL of the co-transformation marker pelt-2::mCherry, as well as 60 ng/μL pUC19, into the gonads of lite-1(ce314) nematodes by standard procedures. Extrachromosomal arrays were generated to yield the following strains: ZX784: lite-1(ce314); zxEx512[punc-17::GFP::PACα; pelt-2::mCherry]–‘Line 1’, and ZX785: lite-1(ce314); zxEx513[punc-17::GFP::PACα; pelt-2::mCherry]–‘Line 2’.

Fluorescence analysis

Several animals of the same strain were immobilized on an agar pad containing 20 mM NaN3 in M9 buffer. Images were recorded under 100× magnification on an Axiovert 200 inverse fluorescence microscope (Zeiss, Göttingen, Germany), equipped with an HBO 100 lamp and a GFP filter set. Afterwards fluorescence intensity was analyzed using ImageJ software (Wayne Rasband, National Institutes of Health, USA, http://rsb.info.nih.gov/ij/). A line was drawn for each image spanning the ventral nerve cord from the nerve ring in the head to the worm`s tail and fluorescence intensity (8-bit gray-scale) along this line was determined. After background subtraction, individual line scans were averaged, and also the mean intensity for each line was calculated for the whole worm length.

Behavioral assays

To determine locomotion parameters on solid substrate, young adults were filmed on plain NGM plates on a worm tracker platform (Zaber Technologies, Vancouver, BC, Canada) under 10× magnification for 15 s in darkness, 25 s under DPSS laser illumination (Pusch OptoTech, Baden-Baden, Germany; 473 nm, 25.6 mW/mm2), and 20-s post-illumination, using a DinoLite digital microscope. Their position at each time point was determined using the coordinates of the stage, as obtained from single-worm tracking software (Wormtracker v2.0.3.1, kindly provided by the Schafer lab, MRC-LMB, Cambridge, UK), which detects the animal in a digital image, follows it by steering an x,y-translational stage, and records its coordinates. From this data, the animals’ velocity was deduced, and from the resulting videos, the body length was calculated.

As an alternative approach to measure animal velocity, we used a custom developed tracking software, written in LabView, that also controls an x,y-translational stage, films the animal and projects light of a chosen color via an LCD projector onto the animal (Stirman et al., accepted, Nature Methods). This system was utilized in control experiments using lower light power (Figure S2), and thus only 2 mW/mm² (450–490 nm) was applied.

For analysis of reversals and bending angles animals were recorded with a Canon Powershot G9 digital camera for 2 min. After 30 s, a 25-s blue light pulse (2 mW/mm²; 450–490 nm) was applied. Subsequently, single frames were extracted from the videos, and a custom written script for ImageJ software was used to find worm medians (details available upon request). Medians were further divided into nine segments of equal length to calculate angles between the latter ones. Long and short reversals were counted by eye.

For analyzing behavior in liquid, thrashing assays of young adult hermaphrodites were performed in 96-well microtiter plates, containing 80 μL of NGM and 80 μL of M9 saline per well. To stimulate PACα activity, animals were illuminated with an HBO 50 lamp (Zeiss; 450–490 nm, 0.2 mW/mm2) under 2.5× magnification. Duration of illumination was defined by a computer-controlled shutter (Sutter Instruments, Novato, CA, USA). Assays were recorded with a Powershot G9 camera (Canon, Krefeld, Germany) and swimming cycles (the worm’s body bends forth and back per each cycle) were counted for defined time bouts before, during, and after blue light illumination.

In all experiments, light power was measured by placing the detector of a powermeter (Thorlabs, Newton, NY, USA) at the focal plane and position in which animals would be present during experiments.

Electrophysiology

Recordings from dissected body wall muscle cells were conducted as described previously (Liewald et al. 2008). Light activation was performed using an LED lamp (KSL-70, Rapp OptoElectronic, Hamburg, Germany; 470 nm, 8 mW/mm2) and controlled by the HEKA amplifier software. mPSC analysis was done by Mini Analysis software (Synaptosoft, Decatur, GA, USA, version 6.0.7).

Statistics

Data are given as means ± SEM. Significance between data sets after two-tailed Student’s t-test or after anova is given as p-value.

ARRIVE guidelines

The ARRIVE guidelines have been followed.

Results

PACα can be expressed in cholinergic neurons of C. elegans

We heterologously expressed PACα, N-terminally fused with GFP, from extrachromosomal arrays. The fusion protein was expressed from the unc-17 promoter to facilitate expression in cholinergic neurons (Alfonso et al. 1993). Since C. elegans generally avoids intense light (particularly blue – UV) in a photophobic response mediated by the putative light-unresponsive-1 (LITE-1) photosensor (Edwards et al. 2008; Liu et al. 2010), we used a lite-1(ce314) mutant background for generation of transgenic PACα strains. This null mutation in the lite-1 gene strongly reduces photophobic reactions to blue light.

Fluorescence microscopy confirmed expression of the GFP::PACα fusion protein in cholinergic motor neurons and their processes along the ventral nerve cord (Fig. 1a–c; arrows) as well as in cholinergic neurons in the head (Fig. 1a and d). Furthermore, it was detected in commissures connecting the nerve cords (Fig. 1b, arrowhead). Within the neuronal cell bodies the protein was evenly distributed in cytosol and nucleus.

Figure 1.

 Expression of PACα in cholinergic neurons of C. elegans. (a–c) Expression of GFP::PACα in cholinergic neurons is shown by fluorescence. Arrows point to neuronal cell bodies within the nerve cord. Scale bars: 50 (a) and 10 μm (b, c). Expression was also found in commissures (b; see arrowhead) which connect the nerve cords. (d) Fluorescence intensity was higher in PACα line 1 than in line 2. Scale bar: 150 μm. (e) Comparison of fluorescence intensity between transgenic lines 1 and 2. Relative fluorescence intensity is shown along the length of the animal (1 pixel corresponds to 1.25 μm). Values are displayed as means (n = 10).

In order to examine whether there is any dependency between potential PACα mediated effects and the amount of PACα expressed, we selected and subsequently worked with two different transgenic lines exhibiting different expression levels (Figs 1d,e and S1): Line 2 (ZX785) had only 64 ± 11% relative fluorescence intensity (and thus PACα expression) of line 1 (ZX784).

Photoactivation of PACα causes an increase in swimming frequency

Next, we analyzed whether photoactivation of PACα in cholinergic neurons has any effects on the behavior of the animals. Studies on C. elegans motility frequently employ the thrashing assay which is performed in liquid medium by counting lateral swimming cycles (body thrashes). Before illumination the thrashing frequency was highest in control animals (0.69 Hz; Fig. 2a) and PACα line 2 (low expression; 0.61 Hz), while line 1 (high expression) had a reduced frequency (0.32 Hz). Upon photoactivation with blue light (λ = 450–490 nm) the thrashing frequency immediately increased up to 255% and 164% for lines 1 and 2, respectively, and remained elevated for a few seconds after the end of the light stimulus (Fig. 2a and b; Video S1). In contrast, thrashing frequency in control animals stayed unaltered during the entire period of photostimulation.

Figure 2.

 Photoactivation increases swimming frequency in C. elegans expressing PACα in cholinergic neurons. (a) The thrashing frequency is displayed for lite-1(ce314) controls, PACα-expressing line 1 (high expression), and PACα-expressing line 2 (low expression). Swimming cycles were determined for 20- and 10-s intervals. Statistically significant differences are related towards the initial velocity of each line before illumination. (b) Relative thrashing frequency with statistical values being relative towards the control animals (lite-1(ce314)). In both graphs, illumination over a period of 20 s is indicated by a black bar. Values are displayed as means ± SEM (n = 10). *p < 0.05, **p < 0.01, and ***p < 0.001.

The results demonstrate that PACα was functionally expressed and that the elevation of intracellular [cAMP]i in cholinergic neurons by photoactivation of PACα was sufficient to activate targets of cAMP. According to the current state of knowledge an increase in [cAMP]i in cholinergic neurons promotes release of the excitatory neurotransmitter ACh (Reynolds et al. 2005; Schade et al. 2005; Charlie et al. 2006). Thus, the effects on motility we observed were presumably triggered by a cAMP-dependent release of ACh at neuromuscular junctions (NMJs) which in turn can stimulate downstream muscle cells and neurons. The relative increase in body thrashes was significantly larger in line 1 than in line 2 suggesting a direct correlation between the amount of cAMP synthesized and increase in the thrashing frequency (Fig. 2b).

Photoactivation of PACα causes an increase in crawling velocity

The locomotion on solid substrate was analyzed quantitatively with a computerized single-worm tracking system (Wormtracker v2.0.3.1). In absence of blue light, both PACα-expressing lines moved more slowly than control animals. However, when transgenic animals were illuminated for 25 s (473 nm; 25.6 mW/mm²), the light stimulus clearly affected their locomotion (Fig. 3a and b). Within a few seconds of illumination they showed a significant increase in velocity which persisted throughout the light stimulus, likely because of increased cAMP levels. A maximal acceleration to 145 ± 11% (line 1) and 150 ± 8% (line 2) of the initial velocity was reached about 18 s after light onset. After the stimulus ended, the velocity decreased, most likely reflecting cellular phosphodiesterase activity. The light-induced increases in velocities of both transgenic lines were similar despite different PACα expression levels, and were both significantly higher than the lite-1(ce314) controls, which did not alter their velocity in response to light. We also used a lower stimulation light intensity (2 mW/mm²), obtaining qualitatively similar results (Figure S2), although the extent of the velocity increase was lower (20–30% increase vs. 50% obtained at 25.6 mw/mm²). This indicates that 2 mW/mm² does not fully saturate PACα effects.

Figure 3.

 Effects of photoactivation of PACα on the velocity of C. elegans on solid substrate. Displayed are lite-1(ce314) controls as well as PACα-expressing line 1 (high expression) and line 2 (low expression). (a) Velocity during photostimulation is given in arbitrary units. Statistically significant differences are related towards the initial velocity of each line before illumination. (b) Relative velocity with values of each line being normalized towards the velocity at the start of the experiment. Statistically significant differences of PACα-expressing line 1 and line 2 are related towards controls. The period of the light stimulus (25 s) is highlighted by a black bar. Values are displayed as means ± SEM (n = 26–35). *p < 0.05, **p < 0.01, and ***p < 0.001.

The locomotion of PACα-expressing animals was normal and highly coordinated. For example, we analyzed mean bending angles of the animals during locomotion and did not observe any light-dependent alterations (Figure S3). This behavior is in accordance with studies on gain-of-function mutants in the Gαs pathway which are expected to have constitutively elevated cAMP levels and show hyperactive but highly coordinated locomotion (Schade et al. 2005). Interestingly, gain-of-function (g.o.f.) mutants in the adenylyl cyclase ACY-1 (allele ce2) moved with increased velocity when compared to the wild type: the fractional increase resembled the velocity increase of both PACα lines during photostimulation (Figure S2). Similarly elevated velocity was measured for gsa-1(ce94)s g.o.f. mutants (data not shown).

Stimulation of PACα influences mPSCs in body wall muscle cells

To directly examine the effects of PACα stimulation on synaptic transmission and neurotransmitter release we carried out electrophysiological recordings on whole-cell patch-clamped body wall muscle cells. We analyzed mPSCs (excitatory) that represent neurotransmitter release at the NMJ by spontaneous fusion of one or few SVs. mPSCs were compared between lite-1(ce314) controls and PACα-expressing line 1 (high expression).

In animals expressing PACα, blue light stimulation for 45 s changed the dynamics of spontaneous transmitter release. Illumination led to a slight increase in the amplitude of mPSCs from 26.6 ± 3.2 pA to 32.2 ± 3.3 pA (Fig. 4a and c). Importantly, we also observed a significant increase of the event frequency from 29.5 ± 5.4 Hz to 39.8 ± 5.6 Hz (Fig. 4a–c). The increase in mPSC frequency was observed within 750 ms after the onset of irradiation and this fast effect demonstrates the feasibility of influencing the cAMP level with good temporal control. The increase persisted for a few seconds after the light stimulus was turned off.

Figure 4.

 Photoactivation of PACα increases the frequency of miniature post-synaptic currents (mPSCs) in patch-clamped body wall muscle cells of C. elegans. (a) Original trace showing photo-stimulated currents in a transgenic animal expressing PACα in cholinergic neurons (line 1; high expression). Illumination over a period of 45 s is indicated by a black bar. (b) Changes in mPSC frequency in control animals (lite-1(ce314)) and PACα line 1 (high expression). Illumination over a period of 45 s is indicated by a black bar. Recorded values were analyzed in 1000 ms intervals. (c) mPSC amplitude in PACα line 1 (high expression) is increased during photoactivation compared to lite-1(ce314) controls. (d) mPSC frequency in PACα line 1 (high expression) is increased during photoactivation compared to control animals (lite-1(ce314)). Values are displayed as means ± SEM (n = 8 for lite-1(ce314); n = 10 for PACα line 1). *p < 0.05, **p < 0.01, and ***p < 0.001.

Based on the fact that photoactivation of PACα leads to an increased cAMP synthesis in cholinergic motor neurons this should subsequently stimulate neurotransmitter (ACh) release at the pre-synapse and lead to the detected increase in the number of mPSCs (Reynolds et al. 2005; Schade et al. 2005). This verifies that the behavioral effects we observed previously were caused by an increase in the ACh release rate. The altered mPSC frequency suggests that the release probability of SVs was elevated. The increase in mPSC frequency also tended to cause more simultaneous SV fusions which may have resulted in mPSCs of increased amplitude, especially just after the stimulus onset.

A significant increase in ACh release at the NMJ might cause muscle contractions and marked changes in body length – similar to what is observed in light-stimulated animals expressing ChR2 in cholinergic neurons (Liewald et al. 2008). Thus, we monitored whether light-stimulation affects the body length of PACα transgenic animals. However, we could not observe any changes in length when blue light was applied (Figure S4).

Photoactivation of PACα causes a decrease in the frequency of long reversals

The locomotion of C. elegans is a useful measure in studies of neurobiological signaling pathways as it is a complex behavior. It mainly consists of a sinusoidal forward movement, however, the crawling is regularly interrupted by discrete motor activities such as reversals (temporary backward crawling) and omega turns (a reversal followed by an almost 180° turn; the worm’s body resembles the Greek letter Ω). Directional changes allow the animal to explore its environment. The frequency of reversals is influenced by environmental conditions such as availability of food or mechanical stimulation (Zhao et al. 2003) – factors which are detected by sensory neurons, and communicated via command interneurons (Fig. 5b) to motor neurons (Gray et al. 2005).

Figure 5.

 Photoactivation leads to a decrease in the number of long reversals in PACα-expressing animals. (a) Displayed are values for lite-1(ce314) controls, PACα-expressing line 1 (high expression), and PACα-expressing line 2 (low expression). During illumination there was a significant decrease in the number of long reversals in both PACα-expressing lines compared to control animals. Values are displayed as means ± SEM (n = 16–17). (b) Schematic display of connections between cholinergic neurons and command interneurons that could be relevant for influencing long reversals (modified from Goodman 2006). *p < 0.05 and **p < 0.01.

Before illumination, the number of long reversals (= pullback length more than half the length of the head) was larger in PACα line 1 than in control animals. However, during photostimulation animals expressing PACα showed a strong decrease in the number of long reversals (Fig. 5a). For PACα line 1 (high expression) the number of long reversals decreased from 1.6 ± 0.4 to 0.1 ± 0.1 reversals/min, while for PACα line 2 (low expression) the number decreased from 1.0 ± 0.3 to 0.2 ± 0.2 reversals/min. Likewise, g.o.f. mutants in the Gαs pathway, i.e. gsa-1(ce94) and acy-1(ce2) (Schade et al. 2005), showed similar characteristics. These animals performed almost no long reversals at all and like photoactivated PACα animals, they moved mainly forward (data not shown). In contrast, control animals did show an increase in the number of long reversals from 0.8 ± 0.3 to 1.5 ± 0.4 reversals/min. In the PACα line with high expression level, we also found a light-dependent rise in body bends with particularly strong bending angle, such that head and tail directly contacted each other (Figure S5; not equivalent to the previously described omega turns, as no reversal was preceding the body bend.) Head/tail contacts increased from 0.6 ± 0.2 to 1.6 ± 0.5 per minute, however, this increase was not significant.

Discussion

In the present study, we showed that PACα can be applied as a useful optogenetic tool in C. elegans, in vivo, to rapidly and transiently manipulate intracellular cAMP levels in selected cells by simple photoactivation. Following PACα photoactivation in cholinergic neurons, increased cAMP production resulted that was seemingly sufficient to mimic an activation of the cellular Gαs signaling pathway, which together with the Gαq pathway regulates synaptic activity (Reynolds et al. 2005). We observed behavioral changes, namely increased swimming frequency and locomotion velocity, presumably triggered by cAMP-dependent release of the excitatory neurotransmitter ACh (Reynolds et al. 2005; Schade et al. 2005). Subsequently, ACh stimulated downstream neurons and body wall muscle cells.

An increased release of ACh at the NMJ was confirmed by electrophysiological recordings from body wall muscle cells where we found a reversible increase in the frequency and amplitude of mPSCs during photostimulation. The frequency returned to baseline levels within a few seconds after the end of the light stimulus, probably because of the fast activity of endogenous phosphodiesterases degrading cAMP. Similar results were obtained in Drosophila larvae by Bucher and Buchner (2009): Following expression in motor neurons, photoactivation of PACα caused an increased frequency of miniature excitatory junction potentials (mEJPs) at the NMJ. While in Drosophila changes in mEJP frequency occurred with a delay of about 1 min after the start of the photoactivation, effects in C. elegans occurred within 750 ms. This time frame very much resembles results in Euglena where the intracellular cAMP level significantly increased within 1 s after photoactivation (Yoshikawa et al. 2005). Our results suggest that the time frame for PACα-dependent stimulation of ACh release in C. elegans is rather short.

The naturally occurring PACα provides researchers with a means of increasing [cAMP]i in C. elegans only in genetically defined cells of interest by using cell type-specific promoters. We observed effects within a few seconds of photoactivation; in contrast pharmacological agonists increase intracellular cAMP levels less rapidly and without spatial specificity. Usually, pharmacological substances, such as the adenylyl cyclase agonist forskolin, or the membrane-permeable cAMP analog dibutyryl cAMP, are used to influence [cAMP]i. However, in Drosophila these drugs took about 15–30 min to cause increases in the mEJP frequency that were of the same dimension as those observed just shortly after a light pulse in animals expressing PACα in motor neurons (Yoshihara et al. 2000; Bucher and Buchner 2009). These factors are of special importance since a previous attempt to use membrane-permeable cAMP analogs to influence locomotion in C. elegans was unsuccessful and showed that timing and location of cAMP are crucial (Schade et al. 2005).

In our study, we also observed a reduction in the number of long reversals during photostimulation. Some of the cholinergic neurons in which PACα was expressed have established synapses and gap junctions to command interneurons (AVB, PVC, AVA, AVD; Fig. 5b; White et al. 1986) which in turn regulate aspects of locomotion. While AVB and PVC can trigger forward movement via B-type motor neurons, AVA and AVD can trigger backward movement via A-type motor neurons (Bhatla 2009). Importantly, each pair of command interneurons also inhibits the opposite type, forming a bi-stable switch that alternates between both directions of movement. A cAMP-dependent release of ACh by cholinergic motor neurons may thus cause a disparity in this network of command interneurons and alter the balance of forward and reverse locomotion.

Importantly, PACα photoactivation appeared to trigger cellular activity not in an ‘uncoordinated’ manner. Rather, neurons were still able to evoke coordinated network activity and locomotion but simply showed exaggerated output while no alteration in body length was seen. This is in contrast to the strong depolarizing stimulation of the cholinergic motor circuit using ChR2, which induces a simultaneous massive release of neurotransmitter from cholinergic cells, overrides any intrinsic locomotory program and causes strong body contractions and paralysis (Liewald et al. 2008). Compared to other methods and optogenetic tools PACα also is advantageous, as it does not require any additional chromophore, it is not harmful for the cell, and its substrate ATP, as well as its chromophore FAD, are readily available.

However, even before any exposure to light the swimming frequency in PACα-expressing line 1 was reduced. Similarly, the velocity on solid substrate in both transgenic lines only reached about 72% of those of control animals. This can be explained by the basal (dark) activity of PACα (Iseki et al. 2002). For Xenopus oocytes expressing PACα it was shown that even in the dark [cAMP]i was increased (Schroder-Lang et al. 2007). A chronically elevated cAMP level in cholinergic cells of C. elegans might evoke compensatory effects on neurotransmitter release pathways resulting in a down-regulation. In an unstimulated state, they generally may release less ACh and therefore the motility of the transgenic animals may be reduced. In accordance with this assumption, line 1, expressing high PACα levels, showed a stronger down-regulation of the swimming frequency than line 2 (low expression). This is again in line with results obtained on Xenopus oocytes where a correlation between PACα expression level and extent of dark activity was observed (Schroder-Lang et al. 2007). These effects could hint at mechanisms which may regulate transmitter release in cholinergic neurons and compensate for a chronic increase of [cAMP]i. In contrast, acy-1 g.o.f. mutants showed a constitutive increase in velocity. However, the acy-1 and unc-17 expression patterns are only partially overlapping (i.e. acy-1 is expressed in many more, if not all neurons, including the locomotion command interneurons; http://www.wormbase.org), which may explain conflicting results. The undesired basal activity exhibited by PACα potentially restricts its application but might be mitigated by using lower expression levels. Thus, one should find a balance between an expression which is low enough to minimize dark activity while being high enough to cause significant effects.

One point to consider is that naturally occurring cAMP signaling is restricted to small domains close to the plasma membrane (Beavo and Brunton 2002) which may well influence the way that this signaling molecule affects downstream pathways. PACα, however, is neither localized to the membrane nor restricted to small domains. Thus, cAMP produced by PACα may have more diverse and possibly unwanted effects. However, future modifications of the protein might help to restrict its subcellular localization.

In the future, PACα can be employed in C. elegans to examine in more detail the function of various (neuronal) cells, to better understand the signal transduction within neurons as well as to clarify the specific role of various neurons in complex neural circuits. Furthermore, it can be used to explore the pre-synaptic function of cAMP and the neuronal Gαs pathway in triggering the release of neurotransmitter and in modulating synaptic transmission. It would also be interesting to further investigate the adaptation of neurotransmitter release to a chronic [cAMP]i increase. Such studies will be helpful in further dissecting the large network of proteins involved in regulating neurotransmitter release and behavior. Finally, Ryu et al. (2010) and Stierl et al. (2010) just recently discovered a bacterial photoactivated adenylyl cyclase, called bPAC or BlaC, in Beggiatoa sp. This enzyme is smaller than PACα and has alternative properties. Furthermore, mutations of BlaC were engineered to generate a photoactivated guanylyl cyclase (BlgC) which now allows to manipulate cGMP levels (Ryu et al. 2010).

Acknowledgements

We thank W. Schafer and E. Yemini for kindly providing the Wormtracker v2.0 software, as well as J. Stirman and H. Lu for providing the combined tracking/projection system. Furthermore, we are grateful to K. Miller for providing the lite-1 strain. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center (CGC) which is funded by the NIH National Center for Research Resources (NCRR). This work was funded by grants from the Deutsche Forschungsgemeinschaft (SFB807-P11, GO1011/2-1, GO1011/4-1 and Cluster of Excellence Frankfurt—Macromolecular Complexes) to AG. The authors declare no competing financial interest.

Ancillary