Heterogeneity in synaptic transmission along a Drosophila larval motor axon - PubMed (original) (raw)

Comparative Study

doi: 10.1038/nn1526. Epub 2005 Aug 14.

Affiliations

Comparative Study

Heterogeneity in synaptic transmission along a Drosophila larval motor axon

Giovanna Guerrero et al. Nat Neurosci. 2005 Sep.

Erratum in

Abstract

At the Drosophila melanogaster larval neuromuscular junction (NMJ), a motor neuron releases glutamate from 30-100 boutons onto the muscle it innervates. How transmission strength is distributed among the boutons of the NMJ is unknown. To address this, we created synapcam, a version of the Ca2+ reporter Cameleon. Synapcam localizes to the postsynaptic terminal and selectively reports Ca2+ influx through glutamate receptors (GluRs) with single-impulse and single-bouton resolution. GluR-based Ca2+ signals were uniform within a given connection (that is, a given bouton/postsynaptic terminal pair) but differed considerably among connections of an NMJ. A steep gradient of transmission strength was observed along axonal branches, from weak proximal connections to strong distal ones. Presynaptic imaging showed a matching axonal gradient, with higher Ca2+ influx and exocytosis at distal boutons. The results suggest that transmission strength is mainly determined presynaptically at the level of individual boutons, possibly by one or more factors existing in a gradient.

PubMed Disclaimer

Figures

Figure 1

Figure 1

SynapCam expression does no affect on NMJ development or physiology. (a) SynapCams are cameleons with CD8 at the N-terminal and the PDZ-interaction domain of the Shaker K+ channel at the C-terminal. Three versions of cameleon, differing in their sensitivity to Ca2+ ions (red circles), were used: SynapCam2.1, with all four Ca2+ binding sites intact; SynapCam3.1, with one site mutated to reduce Ca2+ affinity, and SynapCamNull, with all four sites mutated to serve as a Ca2+-insensitive control. (b) When expressed under the control of the MHC promoter, SynapCam3.1 (red, YFP fluorescence) localized to muscle sites underlying presynaptic terminals of type Ib boutons (shown in green, stained with anti-Nc82, an active zone marker). (c-e) Flies expressing SynapCam3.1 (sc3.1) showed no observable differences in the levels or localization of pre- or postsynaptic markers when compared to control larvae (w1118). (c) Glutamate receptor subunits DGluRIIA or (d) DGluRIIB, (e) Dlg, or (f) Syt, were not perturbed by expression of SynapCam3.1. All images are muscles 6/7, except for anti-Syt which are muscle 8. Scale bar 10 μm. (g-h) Physiological parameters were also unaffected. Neither EJCs (g, n = 10 NMJs), nor the amplitude distribution of spontaneous miniature quantal events (mEJCs, h) were affected by expression of SynapCam3.1 or the experimental conditions (2 μM thapsigargin and 500 μM ryanodine) that prevented muscle contraction. The histogram in (h) is from six NMJs and 1953 events for w1118 (black) and nine NMJs and 2191 events from synapcam3.1 (gray). The holding potential for (g) and (h) was -80mV.

Figure 2

Figure 2

SynapCams report Ca 2+ flux through GluRs as an increase in FRET. (a) An NMJ at muscle 6 during nerve stimulation. Synaptic and non-synaptic (boxes) areas expressing SynapCam3.1 were selected for analysis (in blue for CFP and yellow for YFP). Scale bar 10 μm. (b) The total fluorescence of synaptic regions showed reciprocal CFP (lower black) and YFP (gray) intensity changes following stimulation of the motor axon (EJC, upper black trace). Non-synaptic areas (dashed lines) did not show fluorescent changes. Fluorescence levels for non-synaptic areas were adjusted to synaptic levels for display purposes. (c) Fluorescence changes were indicative of an increase in FRET between CFP and YFP upon synaptic transmission. Single stimuli to the motor axon evoked large FRET (YFP/CFP) changes measured from the entire synaptic area for both synapcam3.1 (lower black trace, 16 NMJs) and synapcam2.1 (gray, 13 NMJs). The synapcam2.1 ΔFRET was on average 18% greater than the synapcam3.1 response. FRET changes had a rapid onset (peak intensity reached after < 200 ms) and a gradual offset, which was fit with a first-order exponential, with SynapCam2.1 exhibiting a slower decay lasting up to 2 s. (d) FRET increases (black, single response for the synapse in a) were induced by Ca2+ transients, dependent on GluR activity and SynapCams did not report voltage-dependent Ca2+ influx. The Ca2+-insensitive SynapCamNull exhibited no change in FRET (green, six NMJs). FRET changes were abolished upon application of desensitizing concentrations of glutamate (1.5mM) (red, three NMJs). A voltage step from−80 to 0 mV (gray bar) did not elicit a change in FRET (gray trace, three NMJs).

Figure 3

Figure 3

SynapCam reveals transmission heterogeneity at the Drosophila NMJ. (a) Muscle sites postsynaptic to individual boutons were analyzed for Ca2+-dependent FRET changes in response to a single motor axon stimulus. Each color represents a different postsynaptic region highlighted in the CCD image. The FRET trace is displayed to the right in corresponding color as well as the ΔFRET values for each region. A 4.6-fold difference of ΔFRET magnitudes was observed for the NMJ displayed, with a mean difference of 5.3-fold for all experiments. Scale bar 10 μm. (b) In addition to a small decrease in ΔFRET as a result of depression, small fluctuations (arrows) in the magnitude of ΔFRET were observed within single postsynapses upon repeated stimulation (0.25 Hz). These fluctuations however were not observed in the average response for the entire postysnapse (mean FRET trace, black). Fluctuations in ΔFRET were independent of the performance of other postsynapses despite physical proximity. For example at two pairs of postsynapses (green and blue, or pink and yellow), FRET changes showed different fluctuation behavior, regardless of proximity. Therefore, the ΔFRET for a single postsynapse is not influenced by FRET changes at other postsynapses. (c) Comparison of FRET response for postsynapse pairs against the distance between pairs shows no correlation (156 trials, 61 boutons, six NMJs, r = 0.009). Distance is the pythagorean distance between postsynapses centers. The correlation between the FRET traces of a pair of postsynapses was calculated from the FRET time-course vectors in Matlab 7.0 (Mathworks, Natick MA) using the “corrcoef” function.

Figure 4

Figure 4

SynapCam3.1 is not saturated by single stimuli to the motor axon. NMJs like the one shown in (a) were subjected to trials of single stimulation (FRET responses are black traces) and trials where two stimuli were separated by 10 ms (gray traces). Although currents were depressed following the second pulse, the magnitude of ΔFRET was higher for dual stimulation recordings in all boutons imaged as a result of Ca2+ summation in the SynapCam response. Scale bar 10 μm. (b) This increase was observed in all boutons of five NMJs tested. When compared to the response after a single pulse, the pooled responses for two stimuli exhibited on average an 83.1 ± 5.3% increase, and the responses were fit by linear regression (r = 0.897, P < 0.0001), indicating a lack of saturation, even for boutons that responded strongly to a single stimulus. ΔFRET numbers in (b) are the mean ± s.e.m. of three single stimulation and three double stimulation trials for each NMJ. (c) No correlation was found between the level of reporter expression, as assayed by the average resting levels of sensitized YFP (rYFP), and the FRET response of the postsynapse (red line is linear fit, r = −0.047, P = 0.25). (d) No correlation was observed between the size of a bouton and the FRET response (red line is linear fit, r = 0.032, P = 0.33). For (c) and (d), n = 625 boutons, and 45 NMJs.

Figure 5

Figure 5

Prolonged imaging reveals the distribution of transmission strength of an NMJ. (a) Imaging conditions were optimized to allow extended imaging of the NMJ. One frame was acquired 200 ms before, and another 100 ms after, nerve stimulation (0.125 Hz), each of 50 ms exposure. The protocol was repeated for at least 30 trials, but more typically for 100-200 trials. (b) Image analysis produced mean ΔFRET scans of the entire synaptic region (left), and scans where the NMJ was partitioned into postsynaptic regions of interest with mean ΔFRET values for each postsynapse (right). (c) Array of 100 ΔFRET responses for the eight postsynapses numbered in (b). The Y-axis reflects postsynapse identity. The last two rows (separated by a blue bar) are mean ΔFRET for all postsynapses and EJC responses for each stimulus. (d-e) YFP, mean rFRET, and mean ΔFRET images of two different. (d) FRET changes were sometimes observed for type Is postsynapses (short arrows), even though reporter localization at these sites was low (see YFP image). (e) Adjacent postsynapses of similar YFP and rFRET values frequently produced different FRET changes (two examples, black and white arrows) indicating that the mean ΔFRET of a postsynapse was not determined by reporter expression or rFRET values. The scale bar for all images is 8 μm.

Figure 6

Figure 6

A proximal-distal gradient in transmission strength (a) HRP staining (green), overlayed on a SynapCam3.1 YFP image (red), confirms the origin and termini of two branches (marked with asterisks). Mean rFRET (b), mean ΔFRET (c), and ΔFRET partitioned and averaged within each synapse (d) for the NMJ in (a). (e) Scatter plots of the mean rFRET (top), or ΔFRET (bottom) of each postsynapse against its distance from the end of a branch for the NMJ depicted in (a-d). Color corresponds to asterisks in (a), lines are fits to depict trend. (f-i) Additional examples of the transmission strength gradient along the length of axonal branches. Branch ends are marked with asterisks, branch origins are marked with arrows, and muscle border is depicted by a dashed white line. Color bars represent rFRET for (b) and ΔFRET for all other images. Scale bar for all images is 8 μm. (j) Pooled data of all postsynapses' ΔFRET (gray) and rFRET (black) values against the distance from the branch's end show a stronger correlation for ΔFRET values. Linear fits, ΔFRET r = −0.649, P < 0.0001, rFRET r = −0.461, P < 0.0001. (k) Postsynapses at the branch end (first column) give on average greater ΔFRET than the postsynapses that follow them. Numbers are number of postsynapses averaged, mean ± s.e.m., ** P < 0.001, * P < 0.005 independent t-test. For (j) and (k), FRET values were normalized to the highest value within each branch for n = 440 postsynapses, 90 branches, 34 NMJs.

Figure 7

Figure 7

Presynaptic contribution to the gradient of transmission strength (a) Axons expressing cytoplasmic Cam2.3 were subjected to 2.2 s of 40 Hz stimuli (dashed lines in b and c). Before stimulation no axonal gradient was detected either in the amount of Cam2.3 at the synapse (rYFP) or the resting FRET (rFRET). However, during stimulation, boutons at the ends of axonal branches exhibited higher ΔFRET responses than more proximal boutons. (b) Average ΔFRET traces show higher presynaptic Ca2+ increase for end boutons than for ones 2-3 boutons away (ΔFRET/FRET = 18.95 ± 0.88 distal, 15.69 ± 0.73 proximal, P <0.005 independent t-test). Data from 19 distal and 32 proximal boutons from19 axonal branches of 7 NMJs. (c) Vesicle fusion was examined in animals expressing SpH and mDsRed (to aid in visualization of the axonal arbor) presynaptically. Distal boutons showed bigger fluorescent changes upon 40 Hz stimulation indicative of higher exocytosis (ΔF/F = 17.97 ± 1.49 distal, 13.20 ± 0.91 proximal, P <0.005 independent t-test). Data from 52 axonal branches in 14 NMJs, including 52 distal and 67 proximal boutons.

References

    1. Hua JY, Smith SJ. Neural activity and the dynamics of central nervous system development. Nat. Neurosci. 2004;7:327–332. - PubMed
    1. Constantine-Paton M, Cline HT. LTP and activity-dependent synaptogenesis: the more alike they are, the more different they become. Curr. Opin. Neurobiol. 1998;8:139–48. - PubMed
    1. Turrigiano GG, Nelson SB. Hebb and homeostasis in neuronal plasticity. Curr. Opin. Neurobiol. 2000;10:358–64. - PubMed
    1. Liu G. Presynaptic control of quantal size: kinetic mechanisms and implications for synaptic transmission and plasticity. Curr. Opin. Neurobiol. 2003;13:324–31. - PubMed
    1. Oertner TG. Functional imaging of single synapses in brain slices. Exp. Physiol. 2002;87:733–6. - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources