|Home | About | Journals | Submit | Contact Us | Français|
Neuropeptides activate G-protein coupled receptors to acutely modulate cellular excitability and synaptic transmission. However, due to the lack of reagents for precise delivery of peptides within dense brain tissue, the spatiotemporal scale over which neuropeptides act is unknown. To achieve rapid and spatially delimited delivery of neuropeptides in mammalian brain tissue, we developed photoactivatable analogues of two opioids: [Leu5]-enkephalin (LE) and the 8 amino acid form of Dynorphin A (Dyn-8). These peptides are functionally inactive prior to photolysis and exposure to UV light causes clean release of LE and Dyn-8. Recordings from acute slices of rat locus coeruleus (LC) demonstrated that photorelease of LE activates mu opioid receptor-coupled K+ channels with kinetics that approach the limits imposed by G-protein mediated signaling. Temporally-precise and spatially-delimited photorelease revealed the kinetics and ionic nature of the mu opioid response and the mechanisms that determine the spatial profile of enkephalinergic volume transmission in LC.
Neuropeptides are expressed and secreted throughout the mammalian brain, typically in combination with a fast neurotransmitter such as glutamate or GABA (Hokfelt et al., 2000). Neuropeptides are packaged in vesicles and several are known to be released in an activity-dependent manner (Ludwig and Leng, 2006). Neuropeptide expression is often regulated by neuronal activity and many neurons are classified by their selective expression of different neuropeptides and neuropeptide receptors (Hokfelt et al., 2000). Such regulated and heterogeneous expression of neuropeptides suggests a precise function in neuron-to-neuron signaling. Indeed, many aspects of synapse and cell function are modulated by neuropeptide-dependent activation of G-protein coupled receptors (GPCRs) (Strand, 1999; Tallent, 2008). At the behavioral level, neuropeptides have profound and complex neuromodulatory effects on brain function: they regulate social bonding (Insel, 2010), feeding (Morton et al., 2006), sleep (Adamantidis et al., 2010), aversion (Knoll and Carlezon, 2010) and reward (Le Merrer et al., 2009).
Studies into neuropeptide systems have been limited by a paucity of experimental tools. The conditions that trigger neuropeptide release from neurons are largely unknown and currently available methods of activating neuropeptide receptors in brain tissue prevent quantitative studies of their function. Although small molecule agonists for many neuropeptide receptors are available, many GPCRs exhibit functional selectivity such that they are incompletely or unnaturally activated by synthetic ligands (Urban et al., 2007). Furthermore, neuropeptides can bind and activate multiple receptor subtypes present on the same cell with similar affinities (Lupica et al., 1992; Svoboda et al., 1999). Thus exogenous application of peptide ligands, rather than synthetic agonists, more accurately mimics endogenous peptidergic signaling. However, compared to traditional pharmacological agents, peptides are large, hydrophobic molecules and thus diffuse slowly within the brain. Direct peptide application in vivo and in brain slices by perfusion, pressure injection (Williams et al., 1982) or iontophoresis (Travagli et al., 1995) produces a slowly rising, prolonged and spatially imprecise presentation of the peptide. These methods offer poor control over the concentration of peptide delivered, largely limiting quantitative analysis to the effects of saturating doses for consistency (Duggan and North, 1983). However, such doses can rapidly trigger receptor desensitization and internalization, which limits robustness and experimental throughput. Therefore, typical peptide delivery methods can only reveal slow and spatially imprecise neuropeptide actions, leaving the possibility of short-lived, local neuropeptide signaling unexplored.
In dissociated neurons, peptide signaling reaches full activation within several seconds of agonist exposure and deactivates within seconds of washout (Ingram et al., 1997). However, in intact brain tissue, neuropeptide receptors are often found up to hundreds of microns from peptide release sites (Khachaturian et al., 1985) suggesting that neuropeptides are capable of volume transmission. Indeed, strong evidence for this phenomenon has been generated in the spinal cord (Duggan, 2000). The spatiotemporal extent of neuropeptide signaling will be determined by the poorly understood interactions of rapid GPCR signaling downstream of ligand binding, slow peptide diffusion and the action of extracellular peptidases, leaving the limits of neuropeptide signaling in the brain remain undefined.
In order to overcome these technical limitations and gain insight into the spatiotemporal dynamics of peptidergic signaling, we have developed a strategy to produce photoactivatable neuropeptides that can be applied to brain tissue at high concentrations in an inert form. These molecules can be rapidly photolyzed to trigger release of the endogenous neuropeptide with high temporal and spatial precision (Ellis-Davies, 2007). Our initial efforts focus on opioid neuropeptides, since these short peptides and their receptors are known to regulate pain sensation (Scherrer et al., 2009), behavioral reinforcement (Le Merrer et al., 2009) and addiction (Gerrits et al., 2003). Opioid peptides and their receptors are prominent in many brain regions including hippocampus, cerebellum, striatum, amygdala and the locus coeruleus (Khachaturian et al., 1985; Mansour et al., 1994). The opioid receptor family consists of three classically-recognized receptors: mu, delta and kappa. These are activated with differential affinity by the endogenous opioid peptides enkephalin and dynorphin and all couple to Gαi/o such that their activation typically inhibits electrical excitability and neurotransmitter release via the opening of K+ channels and inhibition of voltage-sensitive calcium channels (Wagner and Chavkin, 1995).
To enable rapid, spatially delimited delivery of opioid peptides in neural tissue, we have developed `caged' LE and Dyn-8 peptides that can be released by exposure to ultraviolet (UV) light. These peptide analogues contain a photolabile chemical moiety in a position that attenuates activity at opioid receptors. Exposure to light causes the blocking group to detach, thereby releasing the peptide agonists. As photolysis occurs with microsecond kinetics, release can be initiated on the timescale of neurotransmission. Importantly, the concentration and spatial extent of released peptide can be regulated by varying laser power and the area of illumination. We characterized these tools in brain slices and used them to define the spatiotemporal dynamics of opioid signaling with unprecedented resolution.
Enkephalins and dynorphins are the most prominent opioid peptides in the brain (Khachaturian et al., 1985). We chose to work with LE and Dyn-8 (Figure 1A) because they are the smallest and most chemically stable endogenous opioids from these peptide families. LE activates delta and mu opioid receptors with nanomolar affinity, but is inactive at kappa receptors (Toll et al., 1998). The three additional C-terminal amino acids found in Dyn-8 confer nanomolar potency at kappa receptors in addition to mu and delta receptors (Toll et al., 1998). To render these peptides inactive until exposed to light, we produced analogues modified at the N-terminal tyrosine side chain with the carboxynitrobenzyl (CNB) chromophore, which photo-releases tyrosine with high quantum yield (~0.3) (Sreekumar et al., 1998) on the microsecond time scale (Tatsu et al., 1996) (See Supporting Information for information on peptide production and handling). Extensive studies into the structure-activity relationships of enkephalins (Morley, 1980) and dynorphins (Chavkin and Goldstein, 1981) have revealed an essential role for their common N-terminal tyrosine (Y) in receptor activation. In particular, alkylation (Beddell et al., 1977) or removal (Terenius et al., 1976) of the phenolic OH group reduces the potency of enkephalin analogues, suggesting that modification of the tyrosine side chain of LE and Dyn-8 may be a viable caging strategy. Based on these considerations, we designed CNB-Y-[Leu5]-enkephalin (CYLE) and CNB-Y-Dyn-8 (CYD8) (Figure 1A) to release LE and Dyn-8, respectively, in response to illumination with ultraviolet (UV) light. The chemical structure of CYLE is shown in Figure 1B. Reverse-phase high pressure liquid chromatography experiments confirmed that both peptides cleanly photorelease their parent peptides in pH 7.4 phosphate buffered saline in response to 355 nm laser illumination (Figure S1) and that they are stable in the dark at room temperature for > 48 hrs (data not shown).
To determine if CYLE and CYD8 are inactive at opioid receptors prior to photolysis, we compared their activity on opioid receptors relative to that of LE and Dyn-8, respectively, using an in vitro functional cellular assay. To detect opioid receptor activation, we utilized HEK293 cells that stably express a Gαs-Gαi chimera (Liberles and Buck). This chimeric protein allows GPCRs that normally do not signal through Gαs to stimulate adenylate cyclase and control the transcription of a cAMP-dependent reporter construct. Cells were co-transfected with the opioid receptor of interest and the reporter construct such that receptor activation leads to production of secreted alkaline phosphatase (SEAP). The amount of SEAP that accumulates in the media is assayed on a fluorescence plate reader with the fluorogenic substrate 4-methylumbelliferone phosphate (Liberles and Buck, 2006).
Using this assay, we obtained dose-response curves for CYLE and CYD8 at mu, delta and kappa opioid receptors and compared those to the actions of the parent peptides (Figure 1C, D and Figure S2). The potency of CYLE was reduced 100–500 fold with respect to LE at both delta (LE EC50: 3.2 nM ± 0.8 nM, CYLE EC50: 1.7 μM ± 0.4 μM) and mu receptors (LE EC50: 90 nM ± 11 nM, CYLE EC50: 16 μM ± 1.7 μM, Figure 1D). Similar to LE, CYLE did not activate kappa receptors (Figure S2) and the presence 100 nM CYLE did not reduce the affinity of LE at mu and delta receptors (Figure 1C) or that of Dyn-17 at kappa receptors (Figure S2), indicating that CYLE does not act as an antagonist. CYD8 exhibited similar reductions in potency in comparison to Dyn-8 at kappa (Dyn-8 EC50: 7.1 nM ± 0.8 nM, CYD8 EC50: 16 μM ± 1.7 μM), mu (Dyn-8 EC50: 63 nM ± 4.4 nM, CYD8 EC50: 23 μM ± 2.6 μM, Figure 1D) and delta receptors (Dyn-8 EC50: 9.6 nM ± 2.3 nM, CYD8 EC50: 3.9 μM ± 0.6 μM, Figure S2), with no indications of antagonism. Summary dose-response data are tabulated in Table S1. These data reveal that CYLE and CYD8 possess no significant agonist or antagonist activity at concentrations that should strongly activate receptors following photolysis.
In order to characterize the ability of these molecules to activate neuropeptide receptors in brain slices with spatiotemporal precision, we took advantage of the well-described opioid receptor-mediated activation of K+ channels in neurons of the locus coeruleus (LC). The LC is heavily innervated by enkephalinergic afferents (Curtis et al., 2001; Drolet et al., 1992) and LC neurons express a high density of mu opioid receptors in their somata and dendrites (Van Bockstaele et al., 1996a; Van Bockstaele et al., 1996b), activation of which pauses spontaneous firing (Pepper and Henderson, 1980) by producing large outward currents (Travagli et al., 1995; Travagli et al., 1996; Williams et al., 1982), at least in part through G-protein coupled inward rectifier K+ (GIRK) channels.
We obtained whole-cell recordings from neurons in acute horizontal slices of rat LC (Figure 2A) and characterized the ability of CYLE to photoactivate mu receptors by releasing LE. In current clamp recordings, these cells spontaneously fired action potentials at 2–8 Hz, while local perfusion of 10 μM LE caused a strong hyperpolarization and transient pause in spiking (Figure 2B), consistent with previous studies (Williams et al., 1982).
In voltage-clamp recordings at a holding potential of −55 mV, application of 10 μM LE via a perfusion pipette placed near the cell of interest evoked average (n = 6) outward currents of 199 ± 23 pA in amplitude (Figure 2C). In contrast, when applied to the same cells 10 μM CYLE evoked an average outward current of 5.6 ± 4 pA, measured 3.5–4 min after addition to the circulating ACSF (Figure 2C). A subsequent uncaging stimulus consisting of a 5 ms flash from a 124 μm diameter beam of 355 nm light induced a large outward current that was blocked by the addition of 2 μM naloxone (Nal), a broad-spectrum opioid receptor antagonist (Figure 2D). Additionally, photolysis of an isomer of CYLE in which the amino acid sequence was scrambled to render it inactive at opioid receptors did not produce currents (Figure S3).
To evaluate the extent and kinetics of photoactivation, we compared responses (n=6) evoked by local application of 10 μM LE and a subsequent UV light flash in the presence of 10 μM CYLE (Figure 2E). This analysis revealed that photorelease of LE produces currents similar in amplitude to those evoked by the same concentration of locally applied LE (peak current = 207 ± 19 pA versus 179 ± 9 pA for local perfusion and photolysis, respectively) (Figure 2F). Nevertheless, consistent with rapid delivery of LE directly to the recorded cell, the onset kinetics of the light-evoked response were nearly two orders of magnitude faster than for local perfusion (τon = 349 ± 26 ms versus 11.64 ± 2.22 s for photolysis and local perfusion, respectively) (Figure 2F), such that the peak current was reached within 1–2 s after the light flash. In contrast, the kinetics of deactivation for the uncaging response were only ~2-fold faster (τoff = 24 ± 2 s versus 14 ± 1 s for local perfusion and photolysis, respectively) (Figure 2F). Additionally, the responses to 15 uncaging stimuli delivered to the same cell once every 3 minutes were stable (Figure 2G). Thus CYLE enables rapid and robust delivery of enkephalin in brain slices.
One advantage of caged compounds is the ability to photorelease molecules in a graded or analog fashion by varying the amount of photolysis light. This can be readily achieved by manipulating the light intensity or the area of illumination. To explore the former approach, we applied 5 ms flashes of a focused, 30 μm diameter spot of UV light to the soma during voltage clamp recordings and varied the light intensity. Under these conditions, the light-evoked currents in individual cells increased in amplitude with light power (Figure 3A) such that the average (n = 8) peak currents ranged from 31 ± 3 pA at 1 mW to 300 ± 32 pA at 91 mW.
Another approach to analog delivery is to vary the area of illumination. This was achieved by shaping a collimated beam of fixed power density with a field diaphragm placed in a conjugate image plane of the laser path. The area of the field of illumination was varied from 250 μm2, which is smaller than a typical LC cell body, to 12·103 μm2, which covers the soma and a large region of the proximal dendrites (Figure 3B). Responses were measured in both voltage- and current-clamp recording configurations (Figure 3C). In voltage-clamp, varying the area of illumination produced light-evoked outward currents of various magnitude (peak current = 178 ± 29 pA versus 18 ± 6 pA for 12·103 μm2 and 250 μm2 fields, respectively) and duration. In current clamp, the same uncaging stimuli produced pauses in spontaneous firing of graded duration (t = 29 ± 5 s versus 4 ± 0.3 s for 12·103 μm2 and 250 μm2 fields, respectively) and hyperpolarizations of graded amplitude. In individual cells and across cells, the responses to each photolysis condition in both recording configurations were tightly correlated (Figure 3D). Furthermore, the onset kinetics of the light-evoked currents did not vary across the different uncaging stimuli (τon = 349 ± 26 ms versus 400 ± 66 ms versus 417 ± 112 ms for 12·103 μm2, 4.2·103 μm2 and 1.2·103 μm2 fields, respectively; one-way ANOVA p = 0.81; kinetics could not be reliably derived for responses to the 250 μm2 uncaging stimulus). These results indicate that photolysis delivers LE directly to the site of action over a range of areas.
The ability to tightly regulate the area over which LE is applied provides an opportunity to study the ionic conductances that underlie the mu opioid response in LC with unprecedented accuracy. Although it has been clearly demonstrated that mu opioid receptor activation opens GIRK channels in LC neurons (Torrecilla et al., 2002), reversal potentials determined for the evoked currents in brain slices are frequently much more negative (−140 mV to −120 mV) than predicted for a pure K+ conductance according to the Nernst equation (~ −105 mV typically). This observation might be accounted for by the inability to voltage clamp currents generated in the large (Shipley et al., 1996), gap-junction-coupled dendrites (Ishimatsu and Williams, 1996; Travagli et al., 1995) of LC neurons. Several studies suggest that inhibition of a standing, voltage-insensitive Na+ current may contribute 50% of the observed outward current response to enkephalin (Alreja and Aghajanian, 1993; Alreja and Aghajanian, 1994). Thus, the complete ionic nature of the enkephalin-evoked outward currents has been a subject of debate (Alreja and Aghajanian, 1993; Alreja and Aghajanian, 1994; Osborne and Williams, 1996; Torrecilla et al., 2002; Travagli et al., 1995).
To address this issue, we measured the reversal potential of the LE evoked outward current while restricting the uncaging area to the soma and proximal dendrites where voltage clamp is expected to be optimal (Williams and Mitchell, 2008). Importantly, the responses to the uncaging stimuli shown in Figure 3B were not significantly attenuated by the gap junction inhibitor carbenoxolone (Figure S4), suggesting that gap junctions do not contribute to the LE mediated currents evoked by uncaging CYLE around the soma. To measure reversal potentials in the voltage range of a K+ conductance, we held cells at −55 mV and applied negative voltage ramps to −140 mV over 500 ms during the peak of the outward current (Figure 4A). A response to the ramp alone is presented with a response to the ramp after an uncaging stimulus corresponding to the 4.2·103 μm2 beam (the green ring in Figure 2B) in Figure 4B. The average (n=12) LE-evoked current (Figure 4C, black trace) was isolated by subtracting the baseline from the light-evoked response. After correction for the junction potential this current reversed at −107.2 ± 1.8 mV (Figure 4D), closely matching the calculated K+ reversal potential of −106 mV ([K+]out = 2.5 mM, ([K+]in = 145 mM). In the presence of 3.5 mM BaCl2, a GIRK channel blocker, the outward current was largely attenuated and only a small response remained (peak current = 142 ± 17 pA versus 29 ± 4 for control and 3.5 mM BaCl2, respectively). The average (n=13) LE-evoked, Ba2+-insensitive current (Figure 4C, D), appeared to reverse near −140 mV. We were unable to block or occlude this residual outward current with the voltage-sensitive Na+ channel blocker TTX, the voltage-gated K+ channel blocker 4-AP, the voltage gated Ca2+ channel blocker Cd2+, or the HCN channel blocker ZD7288 and it persisted in K+ free, Cs+ based internal (data not shown). Thus, at least 80% of the outward current evoked by LE at the soma and proximal dendrites is carried by a Ba2+-sensitive K+ current.
The finding that somatodendritically-evoked currents reverse in the predicted range for a K+ conductance suggests that dendritic currents may indeed account for the negatively-shifted reversal potentials previously obtained with bath-applied enkephalin. Therefore, using a focused, 30 μm uncaging spot, we compared the reversal potential of currents evoked by uncaging on the soma and ~150–200 μm away on a dendritic branch. Because the dendritic current was typically much smaller than the somatic current evoked by a given uncaging stimulus, we also measured the somatic current after reducing laser power to produce a response of similar amplitude. As shown for a representative cell in Figure 4E, whereas somatically-evoked currents of different amplitude reverse at the same membrane potential, the reversal potential of the dendritically-evoked current is shifted to more negative values. In 2 of 5 cells, the reversal potential was shifted by −12 and −16 mV, but evoked currents in the other 3 cells did not reverse even at −140 mV, the most negative potential reached in our ramps. For quantification purposes, these were assigned a reversal potential of −140 mV. By this measure, the average (n=5) large somatic currents (182.4 ± 8.5 pA), small somatic currents (63.4 ± 5.9 pA) and dendritic currents (59.6 ± 4.3 pA; p=0.71, paired two-tailed t-test between small somatic and dendritic currents) reversed, respectively, at −110.4 ± 1.9 mV, −109.2 ± 1.6 mV and −132.2 ± 4.8 mV* (p<0.05, paired two-tailed t-test between small somatic and dendritic currents). Thus, mu opioid-activated currents in the dendrites of LC neurons cannot be reversed by somatic voltage clamp in the membrane potential range expected for K+ currents.
Although neuropeptides are generally thought to be degraded by extracellular proteases, the extent to which this clearance mechanism limits the duration of peptide signaling after release is unknown. To determine the contribution of peptide degradation via proteolysis to the termination of enkephalin signaling, we repeated the voltage clamp experiment shown in Figure 3B and compared the currents evoked by light before and after the addition of the peptidase inhibitors bestatin and thiorphan (Figure 5A). This inhibitor combination blocks the degradation of enkephalin in brain tissue by >95% and increases the EC50 of bath applied [Met5]-enkephalin by 7-fold in LC (Williams et al., 1987). When the uncaging beam was restricted to the soma (250 um2 and 1.2 um2/103 beam areas), no significant difference in the peak current, charge transfer or moment (the time at which half of the total charge transfer occurs) of the uncaging response was observed (Figure 5B). This indicates that when opioids are released with spatial heterogeneity, the local response is determined by the local time course of peptide release and diffusion, without contribution of peptidase-mediated degradation. However, with larger uncaging areas, peptidase inhibition significantly enhanced all of these parameters. The total charge transfer at the largest area examined was particularly sensitive to this manipulation, as it was enhanced 1.8 fold, while the moment was enhanced 1.6 fold. The moment was used to quantify the deactivation time course because the decay kinetics in peptidase inhibitors were not well fit by a monoexponential, as expected from the complex kinetics of buffered-diffusion reactions. These results indicate that peptidases limit the spread of enkephalin signaling when released in large volumes but that diffusion plays a larger role in limiting the spread from spatially confined release sites.
To determine the spatial precision with which LE can signal in LC, we focused the photolysis beam down to a nominal spot of ~ 2 μm in diameter and measured the current responses as the uncaging stimulus was applied at various distances along a straight line from the cell body. Due to strong scattering of UV photons by brain tissue, the effective illumination spot will be larger (Sarkisov and Wang, 2007). Laser power was adjusted to elicit a response of ~100 pA upon photolysis at the soma. Although care was taken to choose a trajectory that avoided major dendritic branches, due to the somewhat radial nature of the dendrites in the x-y plane, dendritic processes were typically present near the uncaging stimulus and most likely contributed to the measured responses. Nonetheless, the flash-evoked current amplitudes decreased with distance from the soma (Figure 6A), yielding a half-maximal response at a distance of ~100 μm. The activation kinetics of the evoked currents similarly decreased with distance from the soma (τon = 0.31 ± 0.01 s versus 0.89 ± 0.14 s for photolysis at the soma and at a distance of 150 μm, respectively; Figure 6B). Kinetic parameters could not be derived from the small individual responses to the uncaging stimulus at a distance of 300 μm and were thus excluded from this analysis. To determine if the observed profile depends on the shape of the uncaging stimulus (in this case a cone of light focused to a 2 μm diameter spot), we repeated this experiment using a collimated beam of 10 μm in diameter and adjusted the light intensity to again produce a response of ~100 pA at the soma. As shown in Figure S4, the spatial profiles for the elicited currents are super-imposable, suggesting that the spatial extent of signaling observed reflects the spread of enkephalin signaling and is not a consequence of the optical configuration used for uncaging.
Neuropeptides are an important class of neurotransmitters that has received relatively little attention in comparison to other neuromodulators such as acetylcholine and the monoamines. Because it has been difficult to selectively stimulate neuropeptide release from distinct cell types (however see (Ludwig and Leng, 2006)), our understanding of neuropeptide signaling dynamics is limited. Photoactivatable molecules enable spatiotemporally precise delivery of endogenously occurring ligands in relatively intact brain tissue preparations. We were able to generate photoactivatable opioid neuropeptides that are sufficiently inert to allow large responses to be generated with a brief uncaging stimulus. The caged LE analogue CYLE provided robust, rapid and graded delivery of LE in acute brain slices. The ability to spatially restrict release allowed us to selectively evoke currents from regions of neurons that can be effectively voltage clamped in order to accurately measure the reversal potential of the mu opioid receptor mediated K+ current, which was not previously possible in brain slices of LC. These features further enabled us to quantitatively characterize the mechanisms governing peptide clearance and delineate the spatial profile of enkephalinergic volume transmission for the first time.
Based on extensive prior pharmacology, we identified the N-terminal Tyrosine side chain as a caging site where the relatively small CNB chromophore sufficiently attenuates potency on both LE and Dyn-8. Peptides may be inherently more difficult to `cage' than small molecules as the caging group will only interfere with one of multiple interaction sites with receptors. In particular, hydrophobic interactions contribute greatly to peptide-receptor binding and hydrophobic side chains lack functional handles for attaching caging groups. For these reasons the full length Dyn-17 or beta-endorphin may be more difficult to cage by the same approach.
CNB-tyrosine photolysis occurs with microseconds kinetics following a light flash (Sreekumar et al., 1998; Tatsu et al., 1996). Thus the time-course of activation we observed in slices likely reflects the time required for ligand binding and engagement of the G-protein mediated signaling pathway that activates GIRK channels. Indeed, we observed a 50–100 ms delay from the flash to the current onset and a peak response within 1–2 seconds, which closely matches the rates observed for GABABR-mediated GIRK activation in dissociated cells using rapid perfusion techniques (Ingram et al., 1997; Sodickson and Bean, 1996). However, the offset kinetics we observed are orders of magnitude slower than those measured in dissociated cells, where currents cease within 1–2 s of agonist washout. Instead, photorelease produced deactivation kinetics that were only two-fold faster than those obtained with local perfusion, likely reflecting slow diffusion of released peptide away from the recorded cell in neural tissue and concomitant proteolytic cleavage. Indeed, addition of a protease inhibitor cocktail slowed deactivation of the response to photolysis over large areas.
Prompted by the ability to spatially confine LE release, we revisited previous studies into opioid actions on rat LC neurons that were unable to unambiguously identify a K+ current using reversal potential measurements in brain slices (Osborne and Williams, 1996; Travagli et al., 1995). The reversal potential of the LE-dependent current that we measured is accounted for by a pure K+ current and 80% of the outward current was blocked by a high concentration of Ba2+, consistent with a dominant role of GIRKs. It has been also been proposed that down-regulation of a cAMP-dependent standing Na+ current contributes 50% of the opioid response in rat LC (Alreja and Aghajanian, 1993; Alreja and Aghajanian, 1994). Although we cannot rule out that this component mediates the remaining 20% of the current not sensitive to Ba2+ or that this Na+ permeable channel may be enriched in the dendritic regions not activated by our somatodendritically-restricted uncaging stimulus, our results clearly demonstrate that the majority of the somatodendritic current is carried by K+ channels and thus cannot reflect the closing of Na+ channels. Consistent with previous work that suggests that poor voltage clamp of K+ currents originating in LC dendrites may underlie the apparent negative shift in K+ reversal potential (Ishimatsu and Williams, 1996; Travagli et al., 1995), photorelease of LE in the distal dendrites evoked currents that reversed at membrane potentials much more negative than somatically-evoked currents. This shifted membrane potential can arise from poor space-clamp of the large LC dendrites, K+ currents activated in unclamped and gap junction-coupled neighboring cells, or from other dendritically localized and opioid modulated conductances.
By varying the laser power and uncaging area used to photorelease LE, we found a correlation between the amplitude and duration of the outward current in voltage clamp and the duration of the pause in spontaneous firing recorded in current clamp (Figure 3). The responses to the smallest stimuli demonstrate that small outward currents near the detection limit are sufficient to prevent just a few action potentials without causing significant hyperpolarization. At the opposite extreme, strong uncaging causes a large somatic hyperpolarization and pauses action potential firing for 30 s or longer. Thus, the effect of enkephalin on LC firing can be subtle or dramatic, highlighting that neuropeptides are capable of temporally precise actions in addition to volume transmission.
We found that LE could generate opioid-receptor mediated currents when released ~150 μm from the recorded cell. The slower onset kinetics observed when LE was released at locations distant from the soma suggest that the photolyzed peptide diffused from the release site to activate receptors on the soma and proximal dendrites. These distances are large compared to those over which fast-acting neurotransmitters such as glutamate (Carter et al., 2007) and GABA (Chalifoux and Carter, 2011) can spread, as clearance mechanisms for these neurotransmitters are present at high density in neural tissue. Under the conditions of our experiments, LE was nearly inactive when released 300 μm from the soma, which reflects the limit of detection by mu opioid receptors due to dilution of the peptide as it diffuses away from the release site. Assuming a diffusion-limited process, this absolute boundary depends not only on the initial quantity released, but also on the affinity of the receptor for the ligand. Our results may overestimate the mobility of LE in LC due to activation of receptors on dendrites that are closer to the release site than the soma and contributions from currents originating in gap junction-coupled neurons. Nonetheless, our results indicate that enkephalin can indeed function as a volume transmitter in LC and define the spatial profile of the spread of enkephalinergic signaling from a single release site. The spatial profile of signaling may be different in other brain regions due to variations in the densities and identities of proteases and possible differences in diffusional mobility.
Although we obtained similar results using two differently shaped photolysis beams, UV light scatters extensively in brain tissue. Studies in which similar spot sizes (10–25 μm) were employed for UV-uncaging of glutamate in brain slices report 25–50 μm lateral resolution (Katz and Dalva, 1994; Kim and Kandler, 2003). Below the surface of the brain slice, light scattering enlarges the photolysis spot by approximately two-fold in the x-y dimensions (Sarkisov and Wang, 2007), consistent with these observations. Because 1-photon uncaging provides poor spatial control in the z-dimension, it is most practical to consider our results in terms of area of photolysis in the plane of the recorded cell. Thus we estimate that the 10 μm diameter collimated uncaging stimulus illuminates an area of ~300 μm2 at the depth of our recordings. At the light intensity employed, LE evoked substantial currents when released 150 μm from the uncaging site within 1–2 s of the light flash, indicating that the enkephalin signal rapidly spreads through at least 70,000 μm2 of tissue, which is ~200-fold greater than the area of origin. To more accurately address peptide mobility in brain tissue, two-photon sensitive caging groups must be employed so that release will be restricted to μ3-scale volumes.
Here we have described novel photoactivatable tools for the study of opioid signaling within the mammalian brain. By caging both LE and Dyn-8, we provide reagents that can be used to study mu, delta and kappa receptors. Using UV-mediated photolysis of caged LE in brain slices, we demonstrated that somatic mu receptors in the LC generate an outward current mediated primarily by K+ channels. These reagents allowed us to probe the mechanisms that regulate the spread of opioid signaling in brain tissue and revealed that with graded, temporally precise and spatially confined release, neuropeptides are capable of subtle and relatively short-lasting modulation of neuronal function. This approach represents a general strategy for probing the spatiotemporal dynamics of neuropeptides and should be applicable to other peptide transmitters.
These reagents are expected to interface well with 2-photon Ca2+ and voltage imaging methods, as the CNB chromophore exhibits poor sensitivity to two-photon excitation. Similarly, one photon excitation with the visible wavelengths used to image fluorophores such as GFP and activate light-sensitive ion channels such as channelrhodopsin, is also compatible with our probes. However, the intense UV light used for uncaging can photobleach fluorophores and partially activate channelrhodopsin, so care must be taken to control the area of illumination and minimize the requisite UV light intensity in these contexts. Extension to in vivo studies, including amperometry to measure the effects of opioids on monoamine release, should be possible by equipping optrodes and fiber-optic coupled carbon fibers with perfusion lines for peptide delivery and may thus enable spatiotemporal studies into opioidergic modulation of behavior with unprecedented precision.
Custom chemical synthesis was carried out by Peptech Corporation (Burlington, MA) using standard Fmoc-based solid phase peptide synthesis. The carboxynitrobenzyl-modified tyrosine was prepared by modifying established protocols (Sreekumar et al., 1998). After arrival, CYLE and CYD8 were typically handled under lighting filtered using Rosculux # 312 Canary optical filter paper to remove any traces of UV light that could lead to unintentional photolysis. It was essential to further purify the synthetic material by semi-preparative reverse-phase high pressure liquid chromatography (RPHPLC, Agilent) to remove contaminating photolysis products, which typically included ~1% LE or Dyn-8. Crucially, the UV (and VIS) lamps on the detector were turned off during the purification to prevent photolysis of the purified material. Purified, solid material was dissolved in 50% DMSO/water and diluted to a concentration of 10 mM. Stock concentrations were determined by UV/VIS spectroscopy using an extinction coefficient of 1,000 M−1 cm−1 at 355 nm for the carboxynitrobenzyl-modified tyrosine (Sreekumar et al., 1998).
A 100 μl sample of either 1 mM CYLE or CYD8 in phosphate buffered saline, pH 7.2, contained in a quartz cuvette, was placed for 30 s in the path a 100 KHz pulsed q-switched UV laser (DPSS, Santa Clara, Ca) producing ~800 mW of 354.7 nm light. The samples were then analyzed by RP-HPLC (Agilent) on a 4.6 mm × 150 mm Zorbax SBC8 column (5 μm particle size) in H2O, MeCN, 0.1% TFA at 1 ml/min. LE and CYLE were resolved by a gradient from 30% to 70% MeCN over 10 min. Dyn-8 and CYD8 were resolved by a gradient from 25% to 70% MeCN over 10 min. Samples were monitored at 210 nm, 280 nm and 350 nm. Non-photolyzed CYLE, CYD8, LE and Dyn-8 samples were prepared and run under the same conditions. Although addition of the parent peptides to photolyzed samples increased the product peaks as expected, further analysis by mass spectrometry confirmed photolytic conversion of CYLE to LE and CYD8 to Dyn-8 (data not shown).
HEK293 cells stably transfected with a chimeric Gs-Gi protein (Liberles and Buck; Liberles and Buck, 2006) were grown in DMEM (Invitrogen) containing 5% FBS (Invitrogen) and 500 μg/ml G-418 (Invitrogen) and maintained at 37°C in an atmosphere of 5% CO2. Cells were plated in 96-well plates at 50,000 cells/well and co-transfected with the GPCR and reporter plasmid using Lipofectamine® (Invitrogen) and PLUS® reagent (Invitrogen). The transfection media was replaced with ligand-containing DMEM (200 μl/well) and cells were incubated for 36–48 hrs 37°C/5% CO2. Care was taken to avoid exposure to bright, non-filtered light. Each plate was then wrapped in plastic wrap and incubated at 68°C for 2 hrs to heat-inactivate native phosphatases. After transferring 100 μl aliquots from each well to a fresh 96-well plate, 100 μl of an aqueous buffer containing 2 M diethanolamine bicarbonate and 1.2 mM methylumbelliferone phosphate, pH 10.0, was added to each well. Plates were imaged using a Perkin Elmer Envision plate reader using optical settings for methylumbelliferone fluorescence, taking care to image all conditions to be compared on the same day for a single receptor at the same time-point after MUP addition. In every individual experiment, the non-caged parent ligand (LE or Dyn-8) was evaluated alongside the caged compound (CYLE or CYD8) at the same receptor and used to normalize data across trials. cDNAs encoding human delta (OPRD1) and human kappa opioid receptors (OPRK1) contained an N-terminal 3xHA epitope tag (Missouri S&T cDNA Resource center, www.cdna.org) and the murine mu opioid receptor (OPRM1) contained an N-terminal FLAG epitope tag. Data shown reflect the average of 6 replicates run in parallel for each condition.
Animals were handled according to protocols that were approved by the Harvard Standing Committee on Animal Care and are in accordance with Federal guidelines. Post-natal day 22–29 Long-Evans rats were anesthetized by inhalation of isoflurane and cardiac perfused with ice-cold ACSF containing (in mM): 127 NaCl, 2.5 KCl, 25 NaHCO3, 1.25 NaH2PO4, 2.0 CaCl2, 1.0 MgCl2, and 25 glucose, equilibrated with 95% O2/5% CO2, osmolarity 307. Horizontal slices of locus coeruleus were prepared in a cold choline-based artificial cerebrospinal fluid (choline-ACSF) containing (in mM): 25 NaHCO3, 1.25 NaH2PO4, 2.5 KCl, 7 MgCl2, 25 glucose, 1 CaCl2, 110 choline chloride, 11.6 ascorbic acid, and 3.1 pyruvic acid, and equilibrated with 95% O2/5% CO2. Slices of 260 μm thickness were cut with a Leica VT1000s (Leica Instruments, Nussloch, Germany) and transferred to a holding chamber containing ACSF. Slices were incubated at 32°C for 30–45 min and then left at room temperature (20–22°C) until recordings were performed.
All recordings were performed within 5 hours of slice cutting in a submerged slice chamber perfused with ACSF warmed to 32°C and equilibrated with 95% O2/5% CO2. Whole-cell recordings were made from LC neurons visualized using Dodt gradient contrast. The LC was identified in horizontal brainstem slices as a distinct, relatively translucent cluster of cells with exceptionally large somata, typically 20–30 μm in diameter. For current clamp recordings and voltage clamp recordings measuring K+ currents, patch pipettes (open pipette resistance 1.6–2.2 MΩ) were filled with an internal solution containing (in mM): 135 KMeSO4, 5 KCl, 5 HEPES, 1.1 EGTA, 4 MgATP, 0.3 Na2GTP and 10 Na2Creatine Phosphate (pH 7.25, osmolarity 286). For the experiments in Figures 4E and Figure 6, 20 μM Alexa 594 (Molecular Probes) was included in the internal solution.
Recordings were made with an Axoclamp 200B amplifier (Axon Instruments, Union City, CA). Data were filtered at 5 kHz and sampled at 10 kHz. Cells were held at −55 mV in voltage clamp mode, and no current was injected in current clamp mode. Cells were rejected if holding currents exceeded −200 pA. Series and input resistance were measured throughout the experiment and recordings were discarded if series resistance exceeded 15 MΩ. Series resistance was not compensated. Liquid junction potentials of ~ −8 mV were not corrected except when calculating the reversal potentials in Figure 4. In these experiments we only accepted recordings in which series resistance was between 5–8 MΩ and compensated for whole cell capacitance and series resistance by 60–80%.
In all experiments, the following pharmacological agents were used in the extracellular solution at the following final concentrations (in μM): 10 CPP (Tocris), 10 NBQX (Tocris), 25 picrotoxin (Tocris). In some experiments, additional agents were added, as indicated in the text: 2 μM naloxone (Tocris), 100 μM carbenoxolone (Tocris), 3 μM thiorphan (Sigma), 20 μM bestatin (Sigma), 3.5 mM BaCl2 (Sigma), 1 μM TTX (Tocris), 1 mM 4-AP (Tocris), 300 μM CdCl2 (Sigma), and 50 μM ZD7288 (Tocris).
For UV uncaging, we used a custom setup based on a BX51Wl microscope (Olympus). The output of a 100 KHz pulsed q-switched UV laser (Model 3501, DPSS, Santa Clara, Ca) producing ~800 mW of 354.7 nm light was launched into a multimode, 200 μm inner diameter optical fiber with a numerical aperture of 0.22 (OZ Optics, Ottawa, Ontario, Canada). The beam was shuttered at the laser head (OZ Optics, part number HPUC-2,A3HP-355-M-10BQ-1-SH) and collimated at the output of the fiber using either a factory (OZ Optics, part number HPUC0-2,A3HP-355-M-25BQ) or custom built collimator to produce a 10 mm diameter beam. Laser pulses were controlled by opening the shutter, waiting for mechanical vibrations in the fiber launch to dampen, and then q-switching the laser on and off. Light power levels were monitored with a PDA25K amplified photodiode (Thorlabs). Uncaging areas were measured by imaging laser-evoked fluorescence from a thin layer of an aqueous fluorescein solution that was sandwiched between two glass coverslips and placed in the sample chamber.
For the experiments in Figure 2, Figure 3B–D, Figure 4B–D and Figure 5, the 10 mm diameter beam was focused using a planoconvex lens onto the back focal plane of a 60X water-immersion, infinity-corrected objective with a numerical aperture of 0.90 (Olympus) to produce a collimated beam of ~124 μm in diameter. Light intensity was attenuated to ~25 mW, measured as a 10 mm diameter beam at the back aperture of the objective with the planoconvex lens removed from the light path. An iris placed in the light path in a conjugate image plane served as a field diaphragm. The iris was adjusted such that the diameter of the area in the tissue exposed to UV light was either ~124 μm, ~73 μm, ~39 μm or ~18 μm, corresponding to the beam areas of 12·103 μm2, 4.2×103 μm2, 1.2×103 μm2 or 250 μm2, respectively, as indicated in the text. For the experiments in Figure 3A and Figure 4E, the beam was launched directly into the objective to produce a focused UV spot of ~30 μm in diameter and power was modulated with neutral density filters to range from 1 mW to 91 mW, measured as a 10 mm diameter beam at the back aperture of the objective. In this optical configuration, photolysis at light intensities greater than 91 mW led to unstable recordings. For the experiments in Figure 6, the output from a multimode, 25 μm inner diameter optical fiber with an numerical aperture of 0.13 (OZ Optics, Ottawa, Ontario, Canada) was collimated to a 10 mm diameter beam and launched directly into the objective to produce a focused UV spot of ~2 μm in diameter at the sample. Power was modulated empirically to yield a ~100 pA response at the soma. For the experiments in Figure S4, the output from the 25 μm fiber was collimated to a 2.5 mm diameter beam that was focused using a planoconvex lens onto the back focal plane of the objective. The field diaphragm was adjusted to produce a collimated beam of 10 μm in diameter at the sample.
Neurons in acute slices were loaded with 100 μm Alexa Fluor 594 via a patch pipette, and images were taken using a custom-built 2-photon microscope and a Chameleon Ti-Sapphire laser tuned to 840 nm.
All data are expressed as the mean ± SEM. In the figures, average traces are shown as the mean (line) ± the SEM (shaded regions or bars). A two-tailed t test was used to determine significance of differences across conditions in Figure 2 and Figure 5. p<0.05 was considered significant.
The authors thank members of the Sabatini laboratory and G. Pekkurnaz for helpful comments during the preparation of the manuscript. We thank S.D. Liberles, L.B. Buck and J. Lemon for assistance with the SEAP assays and for the gift of reagents; J.L. Whistler for the mu opioid receptor plasmid DNA; J.B. Cohen and D. Chiara for HPLC access; J.T. Williams for insightful discussions; and R. Shah for technical support. The work was funded by a grant from the Helen Hay Whitney Foundation (postdoctoral fellowship) to M.R.B. and a grant from the National Institute of Mental Health (MH085498) to B.L.S.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AUTHOR CONTRIBUTIONS M.R.B. conducted the experiments and data analysis. M.R.B. and B.L.S. designed the experiments and wrote the manuscript.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.