Reliable synaptic connections between pairs of excitatory layer 4 neurones within a single 'barrel' of developing rat somatosensory cortex

Abstract

1. Dual whole-cell recordings were made from pairs of synaptically coupled excitatory neurones in the 'barrel field' in layer (L) 4 in slices of young (postnatal day 12-15) rat somatosensory cortex. The majority of interconnected excitatory neurones were spiny stellate cells with an asymmetrical dendritic arborisation largely confined to a single barrel. The remainder were star pyramidal cells with a prominent apical dendrite terminating in L2/3 without forming a tuft. 2. Excitatory synaptic connections were examined between 131 pairs of spiny L4 neurones. Single presynaptic action potentials evoked unitary EPSPs with a peak amplitude of 1.59 +/- 1.51 mV (mean +/- s. d.), a latency of 0.92 +/- 0.35 ms, a rise time of 1.53 +/- 0.46 ms and a decay time constant of 17.8 +/- 6.3 ms. 3. At 34-36 C, the coefficient of variation (c.v.) of the unitary EPSP amplitude was 0. 37 +/- 0.16 and the percentage of failures to evoke an EPSP was 5.3 +/- 7.8 %. The c.v. and failure rate decreased with increasing amplitude of the unitary EPSP. 4. Postsynaptic glutamate receptors in spiny L4 neurones were of the AMPA and NMDA type. At -60 mV in the presence of 1 mM Mg2+, NMDA receptors contributed 39.3 +/- 12.5 % to the EPSP integral. In Mg2+-free solution, the NMDA receptor/AMPA receptor ratio of the EPSC was 0.86 +/- 0.64. 5. The number of putative synaptic contacts established by the projection neurone with the target neurone varied between two and five with a mean of 3.4 +/- 1.0 (n = 11). Synaptic contacts were exclusively found in the barrel in which the cell pair was located and were preferentially located on secondary to quarternary dendritic branches. Their mean geometric distance from the soma was 68.8 +/- 37.4 microm (range, 33.4-168.0 microm). The number of synaptic contacts and mean EPSP amplitude showed no significant correlation. 6. The results suggest that in L4 of the barrel cortex synaptic transmission between spiny neurones is largely restricted to a single barrel. The connections are very reliable, probably due to a high release probability, and have a high efficacy because of the compact structure of the dendrites and axons of spiny neurones. Intrabarrel connections thus function to amplify and distribute the afferent thalamic activity in the vertical directions of a cortical column.

Figure 1. Structure of rat barrel cortex

A, living, unstained thalamocortical slice. The barrel field in L4 of rat somatosensory cortex is shown under DIC. The dark borders are clusters of cell somata, while the bright hollows contain primarily dendrites and axon collaterals. B, cytochrome c oxidase stain of a fixed barrel cortex slice. The staining is shown in false colours (yellow/orange) to enhance the contrast. Scale bars in A and B are 500 μm. C, three spiny stellate neurones located at the border between two barrels. The two neurones on the left are from the same barrel and are synaptically coupled; the spiny stellate cell on the right is located in the adjacent barrel. Dendrites of spiny stellate neurones had an asymmetrical orientation and were confined to the barrel in which their somata were located. D, spiny stellate neurone in a cytochrome oxidase-stained barrel. Same section as shown in B. Scale bars in C and D are 100 μm.

Figure 2. Analysis of EPSP recordings

Analysis of amplitude, noise, rise time, decay time and latency of EPSPs. A single presynaptic AP (top) and postsynaptic response (bottom) recorded in a spiny stellate cell pair are shown. V_m, membrane potential. Upper and lower voltage calibrations refer to pre- and postsynaptic recordings, respectively. The two crosses mark 20 and 80% of the peak EPSP amplitude between which the rise time was calculated. The EPSP decay was fitted with an exponential function to the falling phase of the EPSP (thick line). The latency was defined as the interval between the peak of the presynaptic AP and the onset of the EPSP (dashed vertical lines). The onset of the EPSP was obtained from a parabolic fit of the EPSP rising phase to the baseline. The windows selected for the measurement of the noise-contaminated EPSP amplitude and the baseline noise are denoted A1, B1 and A2, B2, respectively (see text for details).

Figure 7. Reliability of synaptic connections between spiny stellate cells

A, examples of success and failure of a presynaptic AP (top trace) to elicit an EPSP. A single response and a failure are shown in the middle and bottom traces. The trace labelled Pre I shows the time course of the current injected via the presynaptic recording pipette to elicit an AP. Upper and lower voltage calibration refer to pre- and postsynaptic recordings, respectively. B, histogram of failures (▪) and EPSP peak amplitudes (□); the failure rate in this experiment was 8·6%. Same cell pair as in A. C, distribution of the c.v. of unitary EPSPs in 131 connections measured at near-physiological temperature (34–36 °C). D, histogram showing the failure rate (%) in morphologically identified spiny stellate cell pairs. The failure rate was determined from 50–400 trials for each connection.

Figure 8. Decrease of failure rate and c.v. with increasing EPSP amplitude

A, percentage of failures plotted as a function of the unitary EPSP peak amplitude in L4 spiny cell connections. B, c.v. plotted as a function of the unitary EPSP peak amplitude in L4 spiny cell connections. Data are from synaptic connections recorded at near-physiological temperature (34–36 °C, n = 131; •) and from those at room temperature (n = 20; ○). The two continuous lines in A and B are fits to the data obtained at 34–36 °C and represent the predictions of binomial release statistics for the percentage of failures as a function of EPSP amplitude with n = 4 and q_s = 0·15 mV (right) and q_s = 0·8 mV (left) as p_r increases from 0·08 to 0·6 (right-hand lines) and from 0·05 to 1·0 (left-hand lines). The p_r values refer to the two endpoints of each curve.

Figure 11. Uni-directionally connected pair of spiny stellate neurones

A, low magnification of a light microscopic image of a synaptically coupled cell pair filled with biocytin. The dendritic arbor is confined to L4 whereas the axonal arbors span the cortex from L1 to the white matter with extensive arborisation in L2/3 and L4. B–E, putative synaptic contacts established by the upper neurone with the bottom neurone in the encircled areas in A shown at higher magnification. Arrows mark the en passant axons. In B, an en passant synaptic contact established on a secondary dendrite close to the soma is shown while C–E represent synaptic contacts established with tertiary dendrites at different distances from the soma. Scale bar in A, 100 μm; B–E, 10 μm.

Figure 12. Uni-directionally coupled pair of star pyramidal neurones in L4

A, low magnification of a light microscopic image of a synaptically coupled cell pair filled with biocytin. In contrast to spiny stellate cells, these neurones possess a prominent, thick apical dendrite terminating in L2/3. The axons span the cortical L1 to L6. B–D, putative synaptic contacts established by the upper neurone with the bottom neurone in the encircled areas in A shown at higher magnification. Arrows mark the en passant axons. B, en passant synaptic contact established on an apical oblique dendrite; C and D, synaptic contacts on basal dendrites at different distances from the soma. E, autaptic contact on a primary dendrite close to the soma of the presynaptic neurone. Scale bar in A, 100 μm; B–E, 10 μm.

Figure 3. Morphology and AP firing pattern of spiny L4 neurones

A and B, AP firing pattern of a spiny stellate cell (A) and a star pyramidal neurone (B) in L4 of the barrel field. To elicit APs a 300 ms-long current pulse (I; 100–400 pA) was injected into the cell. Spiny stellate cells and star pyramidal neurones displayed a regular spiking pattern with a variable extent of after-hyperpolarisation and spike adaptation (cf. A and B). C and D, reconstructions of a spiny stellate cell (C) and a star pyramidal neurone (D) in L4 of the barrel field at low magnification; axons are not shown in their entirety. While spiny stellate cell dendrites were confined to L4, star pyramidal neurones had a short, untufted apical dendrite that terminated in L2/3. Axons are marked by arrows.

Figure 4. Uni- and bi-directional connection between pairs of spiny stellate neurones

Simultaneous voltage recordings from a pair of synaptically coupled spiny stellate cells. Traces were obtained at 34–36 °C and represent means of 50 records. A, uni-directional connection: an AP in cell 1 fails to elicit an EPSP in cell 2 while an AP in cell 2 results in an EPSP in cell 1. B, bi-directional connection: an AP evoked in either cell 1 or cell 2 elicits an EPSP in the other cell. Traces labelled I show the time course of the current injected via the recording pipette; V_m, membrane potential; AHP, after-hyperpolarisation.

Figure 5. Latency fluctuation of unitary EPSPs in a spiny stellate cell

A, presynaptic AP (top trace) and unitary EPSPs (bottom traces) recorded from a pair of spiny stellate cells. The EPSP recording shows small fluctuations in latency, measured between the peak of the presynaptic AP and the beginning of the EPSP. Upper and lower voltage calibration refer to pre- and postsynaptic recordings, respectively. B, distribution of latencies of unitary EPSPs. C, plot of latency against peak EPSP amplitude showing a weak inverse relationship, which was statistically not significant. Correlation coefficient is −0·17 and slope is −0·09 ms mV⁻¹ (at 34–36 °C).

Figure 6. Time course and amplitude of EPSPs in excitatory L4 neurones of the barrel cortex

A and B, time course of EPSP rise (A; arrows mark 20 and 80% of peak amplitude) and decay (B) at 35 °C (same recording). Traces represent means of 20 unitary EPSPs. C, histograms of EPSP latency, 20–80% rise time, unitary EPSP amplitude and decay time constant in excitatory L4 neurones. Data from spiny stellate and star pyramidal neurones were pooled. Decay times were obtained by fitting a single exponential to the decay of the mean unitary EPSP. □, recordings at 34–36 °C; ▪, recordings at room temperature (21–23 °C). The mean latency of unitary EPSPs was 2·55 ± 0·51 and 0·92 ± 0·35 ms, the mean rise time 3·01 ± 1·00 and 1·53 ± 0·46 ms, and the mean decay time 47·3 ± 26·3 and 17·8 ± 6·3 ms at room temperature and 34–36 °C, respectively.

Figure 9. Suprathreshold unitary EPSPs in spiny L4 neurones

A, spiny stellate cell pair in which the presynaptic AP evoked suprathreshold EPSPs in the postsynaptic cell (V_m =−61 mV). Five representative, consecutive responses are shown. The latency of the postsynaptic AP had a considerable jitter, depending on the amplitude of the unitary EPSP. B, AP generation in a postsynaptic neurone (V_m =−65 mV) by a high-frequency (100 Hz) burst of APs in the presynaptic cell. Calibration bars in B also apply to A.

Figure 10. NMDAR/AMPAR ratio of unitary EPSPs and EPSCs in spiny L4 neurones

A, unitary EPSPs measured in a spiny stellate cell before (Control) and after the addition of 10 μM CNQX to the bath solution to isolate the NMDAR-mediated component. Traces are means of 50 sweeps. The membrane potential was set to −60 mV during the recording. The thin line represents the EPSP component mediated by AMPARs obtained by subtraction of the EPSP in CNQX-containing saline from that in control saline. B, histogram of the fractional NMDAR-mediated EPSP integral (n = 17) at −60 mV recorded in the presence of CNQX. The mean integral of the NMDAR-dependent EPSP was 39·3 ± 12·5% of that in the absence of the AMPAR antagonist. C, voltage dependence of the NMDAR-mediated EPSP integral measured between −90 and −10 mV. D, unitary EPSCs measured in a spiny stellate cell before (Control) and after addition of 10 μM CNQX to the bath solution to isolate the NMDAR-mediated component. The postsynaptic neurone was held at a membrane potential of −60 mV. The thin line represents the AMPAR-mediated component of the EPSC and was obtained by subtracting the peak current in the presence of CNQX from that in its absence. E, ratio of peak AMPAR current vs. peak NMDAR current plotted for 12 cell pairs at −60 mV. The dashed line represents unity, i.e. when the peak AMPAR current is equal to the NMDAR current.

Figure 13. Reconstruction of a pair of reciprocally connected spiny stellate cells

Computer-assisted three-dimensional reconstruction of a pair of reciprocally coupled L4 spiny stellate neurones. The axon of the black spiny stellate cell is drawn in green (A), the axon of the red spiny stellate cell is drawn in blue (B). Putative synaptic contacts by the green axon with the dendrites of the red neurone are marked by green dots (C); contacts established by the blue axon with the dendrites of the black neurone are marked by blue dots (D). Scale bar, 50 μm. The centres of the somata of the two neurones were 19 μm apart in the horizontal direction and 7 μm apart in the vertical direction.

Figure 14. Geometric dendrogram of the pair of reciprocally connected neurones

A and B, geometric dendrogram of the cell pair shown in Fig. 13. Dendrites are pointing upwards in this figure, the axons and their collaterals downwards. A, dendritic and axonal arbor of the black cell in Fig. 13. B, dendritic and axonal arbor of the red cell in Fig. 13. The colour coding of dendritic and axonal branches and the labelling of potential synaptic contacts is the same as that in Fig. 13. Note that boutons on axonal arbors and contacts on dendritic arbors are close (within 260 μm) to the somata.

Figure 15. Location of synaptic contacts on dendrites

A, histogram of geometric distances from the soma of visually (i.e. on the light microscopic level) identified putative synaptic contacts in 10 pairs of spiny L4 neurones. Inset, distribution of the number of synaptic contacts per connection. B, relationship between unitary EPSP amplitude and the number of synaptic contacts per connection. The correlation coefficient, r, obtained for B was 0·59 and the slope was 0·44 mV per contact. The regression line was forced through zero. C, relationship between unitary EPSP amplitude and the mean geometric distance from the soma of synaptic contacts in a connection. The correlation coefficient was 0·37 and the slope was −0·013 mV μm⁻¹. For both graphs, the correlation was not statistically significant.