Schultheis, Christian et al. (2009) International Worm Meeting "Improving optogenetic methods in Caenorhabditis elegans."

Nagel, Georg, Gottschalk, Alexander, Schultheis, Christian

[International Worm Meeting, 2009]

The optogenetic approach uses exogenous, light-sensitive proteins for in-vivo light-dependent depolarization (Channelrhodopsin-2; ChR2) and hyperpolarization (Halorhodopsin; NpHR) of neurons or muscles, respectively. Optogenetic tools are becoming widely used for functional characterization of synaptic transmission and neuronal networks (Zhang et al.; Nature 2007; Liewald et al.; Nat. Meth. 2008). However, some issues prevent optogenetic tools to be used to their full potential: Two of those issues are challenged in the following experiments. To make use of intersecting promoters for a more cell-specific expression, Channelrhodopsin-2 was fragmented genetically in different loops between the TM domains. Reconstitution of two fragments after coexpression in body wall muscle cells was then monitored by fluorescence and by contraction effects resulting from the photoactivation of ChR2. Significant contraction effects confirm a functional reconstitution of two complementary fragments, though these effects were strongly reduced compared to the positive control. In a similar approach, Halorhodopsin was fragmented and analyzed for functional reconstitution. In a second approach for cell-specific expression of ChR2, we apply the FLP-recombinase (Davis et al., PloS Genetics, 2008). Expression of ChR2 from a first promoter is prevented by a transcriptional stop flanked by FLP recombination target (FRT) sites (FRT-block). Following FLP-recombinase expression with a second promoter, the FRT-block is excised and ChR2 is hence expressed in cells at the intersection of the two promoter expression patterns. In our experiments we adjusted this system for use in cholinergic motorneurons and the AVA-neurons. To allow a prolonged depolarization of excitable cells, e.g. to influence cellular events during development, various mutants of ChR2 were analyzed for use in C. elegans. Mutation of C128 results in accumulation of M, N and O intermediates during the photocycle of ChR2, and hence the channel stays much longer in the open state (Berndt et al., Nat Neurosci 2009). Upon expression of various ChR2 C128-mutants in cholinergic motorneurons we found prolonged contraction effects in comparison to full-length ChR2 after photoactivation with blue light. In addition to the directed inactivation of ChR2 C128 mutants with green light this allows a temporally very precise regulation of ChR2-dependent depolarization of excitable cells using only a short lightpulse of reduced intensity. The variants of ChR2 investigated here hence complement optogenetic tools for a precise light-driven in-vivo stimulation of neurons in C. elegans.

Steven J Husson et al. (2009) European Worm Neurobiology Meeting "Optogenetics tools to dissect a nociceptive neuronal network and neuropeptide signalling pathways in C. elegans"

Alexander Gottschalk, Jana Liewald, Steven J Husson, Liliane Schoofs, Martin Brauner, Wagner Steuer Costa

[European Worm Neurobiology Meeting, 2009]

Light-gated cation channels such as the blue light-activated depolarizing Channelrhodopsin-2 (ChR2), allow optical activation of individual neurons of live and behaving C. elegans at the millisecond time-scale in a non-invasive manner (Nagel et al., 2005, Zhang et al., 2007, Liewald et al., 2008). This optogenetics approach paves the way for further functional dissection of peptide signalling pathways or individual neuronal networks in a detail that is not possible in higher organisms. The huge advantage is that we can specifically stimulate the sensory input neurons, while other potentially contributing neurons are kept silent. The sensory PVD neurons that envelop the nematode with highly branched dendritic arbors are involved in harsh touch nociception. Expression and activation of ChR2 in PVD results in a forward escape movement and sometimes a reversal. These results are in line with the fact that the PVD neurons make synaptic contacts with the locomotory command interneurons PVC and AVA that regulate forward and backward movement, respectively. Using electrophysiology, we will assess the physiology of the PVD cells, as well as the downstream interneurons in response to photoactivation, while the involvement of different ion channels, receptors or neurotransmitters will be assessed by RNAi. We are also investigating the flp-15 and nlp-38 neuropeptidergic signalling pathways by optogenetics. Neuropeptide release can be triggered by photo-activating the respective neurons in an acute fashion while effects on behaviour can be observed at the same time. This way, we can correlate neuropeptide action with acute behavioural changes or effects, about which very limited knowledge is currently available in any system. References Liewald JF, Brauner M, Stephens GJ, Bouhours M, Schultheis C, Zhen M, and Gottschalk A (2008) Optogenetic analysis of synaptic function. Nat Methods, 5, 895-902. Nagel G, Brauner M, Liewald JF, Adeishvili N, Bamberg E, and Gottschalk A (2005) Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans triggers rapid behavioral responses. Curr Biol, 15, 2279-2284. Zhang F, Wang LP, Brauner M, Liewald JF, Kay K, Watzke N, Wood PG, Bamberg E, Nagel G, Gottschalk A, and Deisseroth K (2007) Multimodal fast optical interrogation of neural circuitry. Nature, 446, 633-639.

Nagpal, Jatin et al. (2015) International Worm Meeting "Optogenetic control of cGMP mediated signal transduction ."

Woldemariam, Sarah, Schneider, Martin, Gottschalk, Alexander, Nagel, Georg, Nagpal, Jatin, Gao, Shiqiang, L'Etoile, Noelle, Bethke, Mary, Brueggemann, Chantal, Steuer Costa, Wagner

[International Worm Meeting, 2015]

Cyclic guanosine monophosphate (cGMP) is a widely used 2nd messenger in cellular signaling, acting via protein kinase G (PKG), or cyclic nucleotide gated (CNG) channels. In sensory neurons, cGMP allows for signal modulation and amplification, before depolarization. Manipulating cGMP levels is required to access this signalling and provide insights into signal encoding. We achieve this by implementing two photo-activatable guanylyl cyclases - 1) guanylyl cyclase rhodopsin from Blastocladiella emersonii (BeCyclOp) and 2) a mutated version of Beggiatoa sp. bacterial light-activated adenylyl cyclase (BlaC), with specificity for GTP, termed BlgC or bPGC (Beggiatoa photoactivated guanylyl cyclase). BeCyclOp enabled rapid and precise light-triggered cGMP increases in heterologous cells (Xenopus oocytes) and in C. elegans. BeCyclOp exhibits an unusual 8 transmembrane topology and cytosolic N-terminus. Oocyte experiments revealed light/dark activity ratio of ~5,000 and no cAMP production. Via co-expression of the TAX-2/-4 CNG channel in C. elegans body wall muscle, BeCyclOp photoactivation induced rapid light-driven depolarization and contraction of muscle cells. In C. elegans O2/CO2 sensory BAG neurons, BeCyclOp activation rapidly triggered slowing of locomotion, consistent with the normal sensory function of BAG, and in agreement with previous BAG activation by channelrhodopsin (ChR2). Interestingly, a quick 'recovery' of the slowing response was observed both in ChR2 and BeCyclOp stimulation, despite ongoing photostimulation, arguing that this apparent desensitization is neither mediated at the level of cGMP nor the CNG channel, but at the output synapses or in downstream networks.Light activation of bPGC expressed in muscle cells along with TAX-2 and TAX-4 caused a relatively slower and less pronounced contraction as compared to BeCyclOp. We could validate these differences in cGMP production from the two cyclases by directly imaging the cGMP rise, using a genetically encoded cGMP sensor, WincG2. WincG2 (or worm indicator of cGMP) is based on FlincG3, a circularly permutated EGFP fused to cGMP binding domain of PKG. Currently, we are expressing the cyclases in a variety of C. elegans sensory neurons that use cGMP as the 2nd messenger and are performing behavioural experiments that recapitulate cGMP mediated signal transduction in these sensory neurons using optogenetic activation of the cyclases.

Henss, Thilo et al. (2021) International Worm Meeting "Optogenetic tools for manipulation of cyclic nucleotides, functionally coupled to CNG-channels"

Nagel, Georg, Henss, Thilo, Hirschhauser, Alexander, Scheib, Ulrike, Schneider-Warme, Franziska, Hegemann, Peter, Gao, Shiqiang, Gottschalk, Alexander, Pieragnolo, Alessia, Nagpal, Jatin

[International Worm Meeting, 2021]

The cyclic nucleotides cAMP and cGMP are ubiquitous second messengers that regulate numerous biological processes by activating e.g. protein kinases (PKA and PKG) or cyclic nucleotide gated channels (CNGCs). In eukaryotic GPCR signalling, cAMP is generated predominantly by membrane-bound adenylyl cyclases (mbACs), which are located in microdomains together with GPCRs, PK(A) and their targets. The existing optogenetic toolbox in C. elegans is restricted to soluble adenylyl cyclases (i.e. microbial photoactivatable adenylyl cyclases (PACs) from Euglena (euPAC) and Beggiatoa (bPAC), and the synthetic phytochrome-linked cyclases IlaC22 k27 and PaaC), the membrane-bound Blastocladiella emersonii guanylyl cyclase opsin (BeCyclOp) and hyperpolarising rhodopsins (e.g. Natronomonas pharaonis halorhodopsin - NpHR). Yet missing are membrane-bound photoactivatable adenylyl cyclases (mbPACs) and hyperpolarizers based on K+-currents. To obtain mbPACs, we mutated 2-3 key amino acids in the active site of Blastocladiella and Catenaria CyclOps, which are particular in combining a rhodopsin and a guanylyl cyclase domain. For characterization of photoactivatable nucleotidyl cyclases, we expressed the proteins alone or in combination with CNGCs ("two-component optogenetics") in muscle cells and cholinergic motor neurons. To investigate the extent of optogenetic cNMP production and the ability of the systems to de- or hyperpolarise cells, we performed behavioural analyses (locomotion, muscle contraction), measured cNMP content in vitro, and compared in vivo expression levels. We implemented Catenaria CyclOp as a new tool for cGMP production, allowing fine-control of cGMP levels. We established the mbPACs YFP-BeCyclOp(A-2x) and YFP-CaCyclOp(A-2x), based on mutated versions ("A-2x") of Be and Ca CyclOp, enabling more efficient and specific cAMP signalling compared to soluble bPAC, despite lower overall cAMP production. For hyperpolarization of excitable cells by two-component optogenetics, we introduced the cAMP-gated K+-channel SthK from Spirochaeta thermophila and combined it with bPAC, BeCyclOp(A-2x), or YFP-BeCyclOp(A-2x). As an alternative, we implemented the Blastocladiella emersonii cGMP-gated K+-channel BeCNG1 together with BeCyclOp. In summary, we established a comprehensive suite of optogenetic tools for cNMP manipulation, useful for applications in many cell types, including sensory neurons, and for potent hyperpolarization.

Husson, Steven J et al. (2009) International Worm Meeting "Optogenetics tools to dissect small neuronal networks and neuropeptide signalling systems."

Liewald, Jana, Husson, Steven J, Schultheis, Christian, Schoofs, Liliane, Gottschalk, Alexander, Brauner, Martin, Erbguth, Karen, Schedletzky, Thorsten

[International Worm Meeting, 2009]

Light-gated ion channels or pumps such as the blue light-activated depolarizing Channelrhodopsin (ChR2) and the yellow light-driven hyperpolarizing Halorhodopsin (HR) allow optical activation or inhibition in muscles and neurons of live and behaving C. elegans (Zhang et al., 2007). Furthermore, inward currents evoked by either ChR2 or HR, as well as muscle currents in response to activating ChR2 in motor neurons, can be directly measured by electrophysiology, while photo-evoked body contraction or elongation of the animal could be monitored at the behavioural level (Nagel et al., 2005; Liewald et al., 2008). These state-of-the-art technologies pave the way for further functional dissection of individual neuronal networks in a detail that is not possible in higher organisms. Doing so, we are investigating some defined neuropeptidergic signalling pathways. Neuropeptide release can be triggered by photo-activating the respective neurons in an acute fashion while effects on behaviour can be observed at the same time. This way, we can correlate neuropeptide action with acute behavioural changes or effects, about which very limited knowledge is currently available in any system. While higher organisms display millions of contributing neurons, only a handful of neurons take part in individual neuronal networks in C. elegans. This opens the possibility to study the contribution of each neuron to the function of a small network, for example involved in nociception. The huge advantage of our optogenetics tools is that we can specifically stimulate the sensory input neurons, while other potentially contributing neurons are kept silent. The involvement of different ion channels, receptors or neurotransmitters can be assessed by using different genetic backgrounds, while the physiological properties of each individual neuron will be monitored by electrophysiology. Liewald JF, Brauner M, Stephens GJ, Bouhours M, Schultheis C, Zhen M, and Gottschalk A (2008) Optogenetic analysis of synaptic function. Nat Methods, 5, 895-902. Nagel G, Brauner M, Liewald JF, Adeishvili N, Bamberg E, and Gottschalk A (2005) Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans triggers rapid behavioral responses. Curr Biol, 15, 2279-2284. Zhang F, Wang LP, Brauner M, Liewald JF, Kay K, Watzke N, Wood PG, Bamberg E, Nagel G, Gottschalk A, and Deisseroth K (2007) Multimodal fast optical interrogation of neural circuitry. Nature, 446, 633-639.

Anusha Narayan et al. (2007) International Worm Meeting "Transfer at C. elegans synapses."

Shawn Lockery, Paul Sternberg, Tod Thiele, Gilles Laurent, Anusha Narayan

[International Worm Meeting, 2007]

Neural circuits in C. elegans have been studied using light and electron microscopic techniques, focal laser ablations and, more recently, calcium imaging techniques. For a clearer functional understanding of these circuits, however, some knowledge of the rules of synaptic information transfer is required. How is the dynamic range of the post-synaptic neuron set? What are the mechanisms for synaptic integration and gain control? Questions such as these can best be answered by monitoring or controlling connected pre- and post-synaptic neurons simultaneously. We chose to focus on the synapses between the AFD/ASER and AIY neurons, since the functional relevance of these neurons has been established and there is anatomical evidence for synapses between them. Channelrhodopsin-2 (chR2) is a light activated cation channel with fast kinetics (order of milliseconds1). We express chR2 under a neuron-specific promoter2 in the presynaptic neuron, and use whole-cell patch-clamp recording techniques to monitor membrane voltage or currents in the postsynaptic neuron. We are first calibrating the response to light of chR2-expressing neurons. Currently, we are calibrating this light response in worms expressing chR2 in ASER. We have observed depolarizations of 10-30 mV in response to light (450-490 nm) in current clamp, and inward currents of 5-10 pA in voltage-clamp. We have also seen evidence of spontaneous synaptic activity, in the form of discrete synaptic events (potentials or currents) with different reversal potentials (some depolarizing, others Cl-dependent). We are beginning to characterize the ASER-AIY synapse, and will then move on to the AFD-AIY synapse. References 1. Boyden ES, Zhang F, Bamberg E, Nagel G, Deisseroth K. Millisecond-timescale, genetically targeted optical control of neural activity, Nat. Neurosci. 8 (2005), pp. 1263-1268. 2. Nagel G, Brauner M, Liewald JF, Adeishvili N, Bamberg E, Gottschalk A. Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans triggers rapid behavioral responses. Curr. Biol. 15 (2005), pp. 2279-2284.

Schultheis C et al. (2010) Neuronal Development, Synaptic Function and Behavior, Madison, WI "Using Channelrhodopsin-2 Slow Mutants To Complement The Optogenetics Toolbox"

Nagel G, Gottschalk A, Schultheis C

[Neuronal Development, Synaptic Function and Behavior, Madison, WI, 2010]

Channelrhodopsin-2 (ChR2) is a blue light-gated cation-channel from the green algae Chlamydomonas reinhardtii (Nagel et al. 2003). The optogenetic approach uses heterologous expression of ChR2 for in-vivo light-dependent depolarization of neurons or muscles, for functional characterization of synaptic transmission and of neuronal networks (Nagel et al. 2005; Liewald et al. 2008; Liu et al. 2009). However, the required high light intensities for photoactivation do not allow long-term applications, as permanent illumination can cause phototoxicity, thus hampering applications e.g. during development, where long-term altered neuronal activity may influence certain processes or pathways. Recently, it was shown that mutation of the Cys128 residue of ChR2 vastly delays the inactivation of the channel (Berndt et al. 2009). Hence a short light pulse of reduced intensity is sufficient to render these "slow mutants" open for several minutes. Furthermore, slow mutants can be inactivated with green light. To test these ChR2 slow mutants in C. elegans, we expressed them in body wall muscle cells. We found that a brief (1 sec) blue light pulse with reduced intensity (200x less than for WT-ChR2) triggers body contraction for several minutes. A subsequent green light pulse was able to terminate the contraction. Similar results were obtained for photoactivation of ChR2 slow mutants in cholinergic and GABAergic motorneurons to trigger long-term synaptic transmission. In C. elegans two pairs of command interneurons (PVC and AVB) are required to trigger forward locomotion (Zheng et al. 1999). Backward command interneurons (AVA, AVE and AVD) appear to be less active, depending on sensory neuron input, which explains the dominant forward movement. Through photoactivation of ChR2(C128S) in both types of command interneurons, we could permanently disrupt the dominance of forward command interneurons, which significantly increased the ratio of backward movement until a green light pulse terminated ChR2 activity. Our results show that the ChR2 slow mutants can be employed to investigate long-term behavioural effects with minimum light delivery and that this activity can be precisely terminated using green light. We are now investigating possibilities to alter neuronal function in the long-term during animal development.

Erbguth, Karen et al. (2011) International Worm Meeting "Red-shifted optical excitation with a Chlamydomonas / Volvox hybrid Channelrhodopsin."

Hegemann, Peter, Lu, Hang, Prigge, Matthias, Erbguth, Karen, Gottschalk, Alexander, Stirman, Jeffrey N., Schneider, Franziska

[International Worm Meeting, 2011]

How neural circuits evoke behavior is one of the most fascinating questions in the neurosciences. To aid the functional elucidation of neural networks, several optogenetic tools of microbial origin have been adopted for use in C. elegans. Chlamydomonas Channelrhodopsin-2 (ChR2) (1, 2) is a well established optical tool to trigger the depolarization of neurons with blue light, while yellow-green light-triggered hyperpolarization can be achieved using Cl- pumps like Natronomonas pharaonis Halorhodopsin (NpHR) (3) or proton pumps like Leptosphaeria maculans bacteriorhodopsin (MAC) (4). Along with the investigation of tissue specific expression systems, optogenetic tools pave the way for a straightforward and systematic approach to dissect neural networks in vivo. To expand possibilities for multimodal investigation of neural networks, activation of different cell types of one circuit in a temporally and spectrally independent manner is desired. Channelrhodopsin-1 from Volvox carteri (vChR1), another depolarizing cation channel, exhibits red-shifted absorption properties compared to ChR2 (5). However, vChR1 expresses only weakly in metazoan cells including, as we found, C. elegans. Thus, an improved version would help to expand the optogenetic toolbox. To assess alternatives to vChR1, we generated different chimeras of vChR1 and Chlamydomonas Channelrhodopsins (from now on termed cChRs), and screened for robust expression and beneficial properties, i.e. red-shifted action spectrum and light-sensitivity. Candidate proteins were compared to cChR2, and tested for optical compatibility and distinguishable functionality in different cell types. Finally, we identified C1V1duo, a hybrid with specific point mutations that showed peak activation at ~540nm (cChR2 peaks at ~460nm). We could activate both proteins with 400 or 570nm light, respectively, in mechanosensory neurons, evoking withdrawal behavior. cChR2 expressed in GABAergic neurons (activated at 400nm) could counteract activity of C1V1duo in muscles (570nm). C1V1duo appears to be more light-sensitive and efficient than cChR2. We currently assess the compatibility of C1V1duo with yellow Cameleons for Ca2+-imaging as a readout of network activity. (1) Nagel et al. (2005): Curr Biol 15, 2279-84; (2) Liewald et al. (2008): Nat Methods 5, 895-902; (3) Zhang et al. (2007): Nature 466, 633-9; (4) Stirman et al. (2011): Nat Methods 8, 153-8; (5) Zhang et al. (2008): Nat Neurosci 11, 631-3.

Martin Brauner et al. (2007) International Worm Meeting "Combination of the light-driven Cl--pump Halorhodopsin with Channelrhodopsin-2 for independent, bidirectional optical control of muscles and neurons."

Alexander Gottschalk, Jana Liewald, Martin Brauner

[International Worm Meeting, 2007]

Current methods to study the role of specific neurons in their circuitry (e.g. laser-ablation, sensory stimulation) are limited in both temporal precision and (fast) reversibility. To overcome these limitations, we recently established the use of Channelrhodopsin-2 (ChR2), an algal cation channel activated by blue light. ChR2 facilitates rapid photo-depolarization and thus activation of different types of neurons (and muscles), in live, non-disturbed animals (Nagel et al. (2005) Curr Biol 15: 2279). This allows us e.g. to trigger GABA or Acetylcholine release at the neuromuscular junction simply by expressing and photoactivating ChR2 in different classes of motoneurons. The fact that ChR2 permits stimulating specific neurons with millisecond precision motivated us to search for a complementary light-activated hyperpolarizing agent to acutely inhibit neural function in vivo. To this end, we chose the yellow light driven Cl--pump Halorhodopsin (NpHR) from Natronomonas pharaonis. Electrophysiological whole-cell recordings from transgenic body-wall muscle cells expressing NpHR showed rapid outward currents that persisted as long as the illumination. Consistent with the photocurrents observed, photoactivation of NpHR in muscles led to body-relaxation and immediate arrest of locomotion. Similar behavioral effects could be observed in animals expressing NpHR in cholinergic neurons, showing that NpHR can inhibit specific types of neurons in vivo (Zhang et al. (2007) Nature, in press). Intriguingly, the action spectrum of NpHR is red-shifted in comparison to ChR2, which offers the possibility of bidirectional optical modulation of neural activity. Indeed, when coexpressed in muscles or cholinergic neurons, ChR2 and NpHR could be activated independently and concurrently by blue and yellow light. Furthermore, photoactivation of NpHR rapidly and reversibly counteracted contractions observed during activation of ChR2 alone. We now want to establish the combined use of ChR2 and NpHR in different cells within specific neural circuits, such as the command interneuron circuit. In sum, ChR2 and NpHR enable for the first time the repeated and reversible stimulation or inhibition of muscle cells and neurons in live animals with high temporal precision. Since both proteins can be activated separately and in synchrony simply by light of different wavelengths, this method will be very useful in dissecting the neural circuits controlling the numerous behaviors of C. elegans.

Husson, Steven J. et al. (2011) International Worm Meeting "Microbial proton pumps as hyperpolarizers complement the optogenetics toolbox in Caenorhabditis elegans."

Stirman, Jeffrey N., Husson, Steven J., Gottschalk, Alexander, Liewald, Jana F., Lu, Hang

[International Worm Meeting, 2011]

Optogenetic technologies use light to gain endogenous control of defined cells or tissues in a non-invasive manner. Neuronal activity can be manipulated at the millisecond timescale by expressing the light-activated depolarizing cation channel channelrhodopsin-2 (ChR-2) and subsequent illumination with blue light (1). In contrast, yellow light-triggered inhibition of neuronal activity can be achieved by activation of the hyperpolarizing halorhodopsin (NpHR) (2). However, only few successful attempts were undertaken to inhibit C. elegans neurons using NpHR, probably due to the insufficient trafficking to the plasma membrane, and thus need for high expression levels, or the limited hyperpolarizing power of this Cl- channel. An extensive screen of type I microbial opsins from archaebacteria, bacteria, plants and fungi recently revealed powerful outward directed proton pumps as valuable alternative hyperpolarisers (3). As cells and extracellular fluid are strongly buffered, shuffling protons across the membrane is not expected to cause any appreciable pH changes. The yellow-green light-sensitive archaerhodopsin-3 (Arch) from Halorubrum sodomense appears to be significantly more powerful than NpHR. Another proton pump from the fungus Leptosphaeria maculans, Mac, enables neuronal silencing by green-blue light. This opens the possibility to inhibit different neuronal populations, depending on the illumination wavelengths used. Here we present the use of these outward-directed proton pumps as potent circuit breakers in C. elegans. Electrophysiological recordings on dissected muscle cells allowed us to quantify the outward current evoked by either NpHR, Arch and Mac. As Mac can be stimulated using blue light, we can activate a subset of neurons using ChR2 while Mac could be used to simultaneously inhibit downstream neurons, using the same colour of light. Alternatively, illumination of predefined body segments with different colours of light using an integrated LCD projector as light source (4) allows using the more potent Arch (maximal hyperpolarizing power with green light) for neuronal inhibition and simultaneous ChR2-induced activation of other cells with blue light. A few examples for circuit dissection with either bacteriorhodopsin will be presented at the meeting. (1) G. Nagel et al., Curr. Biol. 15, 2279 (2005); (2) F. Zhang et al., Nature 446, 633 (2007); (3) B. Y. Chow et al., Nature 463, 98 (2010); (4) J. N. Stirman et al., Nat. Meth. 8, 153 (2011).