Ultrastructural correlates of vesicular docking in the rat dentate gyrus

Neuroscience Letters 378 (2005) 92–97

Ultrastructural correlates of vesicular docking in the rat dentate gyrus Diano F. Marronea,∗ , Janelle C. LeBoutilliera , Ted L. Petita,b a b

Department of Psychology, University of Toronto, 1265 Military Trail, Toronto, Ont., Canada M1C 1A4 Program in Neuroscience, University of Toronto, 1265 Military Trail, Toronto, Ont., Canada M1C 1A4 Received 5 October 2004; received in revised form 30 November 2004; accepted 6 December 2004

Abstract To determine the extent to which CA1 synapses are typical of those found in other regions of the hippocampal formation, we have carried out a quantitative analysis of synapses in the middle molecular layer of the rat dentate gyrus, reconstructed from serial electron microscopy, and have compared these data with previous observations from CA1. In general, the morphology of synapses in areas CA1 and the dentate agree, other than an increased density of multisynaptic boutons. Thus, it seems that either area may form an equally effective model for the function of individual synapses in the hippocampal formation. In addition, the current study examines presynaptic curvature, which recent mathematical models have suggested may have profound effects on synaptic transmission. When synapses of distinct curvature profiles (i.e., presynaptically concave, convex, and flat) are examined, the average characteristics of these three synapse populations are distinct. In general, concave synapses have a greater number of morphologically docked vesicles, and thus, likely a greater probability of release. This, however, seems to be accounted for by the fact that these synapses are larger—the spatial density of docked vesicles remains identical across these curvature profiles. This study provides crucial data for further modeling of individual synapse function. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Curvature; Synaptic vesicle; Active zone; Hippocampus; Dentate gyrus; Release probability

The hippocampal formation plays a critical role in understanding neural physiology due to its simple structure and its importance in learning and memory. Much of what is known about the structure of synapses in the hippocampal formation, however, is derived from extensive morphometric studies of Schaffer collateral synapses between regions CA3 and CA1 (e.g., [1,11–13,19,21]). Despite these data, several questions about synaptic physiology within the hippocampal formation remain unaddressed. To what extent are Schaffer collateral synapses structurally like those found in other regions of the hippocampal formation? To approach this question, quantitative analysis of excitatory synapses has been undertaken in the dentate gyrus middle molecular layer (MML). This region is suitable for the physiological analysis of non-Schaffer collateral synapses since (a) it has an equally simple structure and (b) it has been implicated in many electrophysiological (e.g., [14]), ∗

Corresponding author. Tel.: +1 416 287 7449; fax: +1 416 287 7642. E-mail address: [email protected] (D.F. Marrone).

0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.12.019

anatomical (e.g., [4]), and molecular (e.g., [23]) studies of hippocampal plasticity in vivo. Thus, one of the main goals of the current study is to determine the physiologically relevant morphological properties of synapses of the dentate gyrus and compare this to the currently available information on synapses from the hippocampus proper. A second issue to be dealt with is the relationship between several morphological parameters, such as terminal and postsynaptic density (PSD) size, and synaptic curvature. Mathematical models have suggested that compartmentalization (see [16]) may have profound effects on the probability of transmitter release (PTR). Transmitter release has long been known to depend on intracellular calcium concentrations ([Ca2+ ]i ), but the exact mechanism by which [Ca2+ ]i facilitates synaptic transmission is unclear [9,10,16,24]. Although it is not yet possible to image these phenomena in vivo, modeling local microgeometry has suggested that the smaller the compartment surrounding the release site, the higher the local [Ca2+ ]i (since it is less able to diffuse), increasing PTR [9,10]. Initial modeling research has parsed the

D.F. Marrone et al. / Neuroscience Letters 378 (2005) 92–97

critical features contributing to compartmentalization: (a) the width of the terminal; (b) the distance from the site of vesicular release to the terminal wall, and (c) the existence and size of terminal partitions. Many of these parameters, however, have yet to be explored through quantitative morphometry. One factor critical to compartmentalization is synaptic curvature. For example, concave synapses are generally thought to be more compartmentalized than convex synapses. In fact, in subsequent modeling research [24], synaptic curvature has been found to affect PTR in two distinct ways. Increasing a terminal’s concavity (i.e., the degree to which the presynaptic terminal protrudes into the postsynaptic element) enhances PTR. In addition, PTR is enhanced by increasing the proximity of a docked vesicle to the concave inflection point (i.e., the point of maximal concave curvature). Given that curvature is an established plastic feature of the synapse (see [16] for a review) and that it may have profound impact on synaptic function [9,10], exploratory spatial characterization of these synapse types are required. These data allow for the correlation of curvature with other morphological measures of PTR. Arguably, the best established of these correlates is the presence of morphologically docked vesicles [17], which are vesicles that appear in electron micrographs immediately adjacent to the active zone membrane. The readily releasable pool, which refers to quanta available for immediate release upon high frequency stimulation, determines PTR, and evidence suggests this pool coincides with the morphologically defined docked vesicle pool [17,19,22]. Thus, simultaneously quantification and correlation of the size, curvature, and the number of morphologically docked vesicles in individual synapses may address how they mediate PTR. Six male Long–Evans rats aged 45–60 days were anesthetized with 2-ml/kg sodium pentobarbital (Nembutal), and perfused with 10 ml physiological saline followed by fixative [2% paraformaldahyde, 2% glutaraldehyde in 0.1 M phosphate buffer (PB), pH 7.2]. Brains were post-fixed for 24 h and given three 1 h washes in PB. Four 0.5 mm mid-dorsal sections were then randomly dissected from the ventral dentate gyrus, placed in 1% OsO4 (4 ◦ C, 1 h), dehydrated in a graded series of ethanol, and embedded in Spurr’s (Ladd Research Industries, Burlington, Vermont) medium. Embedded sections included the granule cell layer and the entire extent of the dentate gyrus. The MML was identified as the middle third of this expanse, and trimmed using Toluidine Blue stained semi-thin (l ␮m thick) sections for orientation. Using an Ultracut microtome with a diamond knife, a series of 22–33 (mean = 30) ultra-thin serial sections per block were mounted with Formvar onto slotted copper grids and counterstained with uranyl acetate and lead citrate. AnalySIS 3.0 (Soft Imaging Systems) software was used to digitally photograph and analyze the MML on a Hitatchi H-7500 electron microscope at 30,000× magnification. Synapses were sampled if vesicles, as well as both pre- and postsynaptic densities were observed. All synapses within the reconstructed volume were included provided: (a) synapses

93

Fig. 1. Typical electron micrograph used in the current study. Depicted in the figure are two simple monosynaptic boutons (Bs ) forming synapses on dendritic spines (Sp ), as well as one multisynaptic bouton (Bm ) excluded from the study. Docked vesicles are shown with arrows. Scale bar = 500 nm.

were asymmetric, axospinuous, and macular; (b) boutons contained only a single synapse; (c) the PSD was cut tangentially (i.e., all sections had clear membrane outlines to permit accurate judgment of docked vesicles); and (d) the entire terminal and postsynaptic spine were within the volume reconstructed. Synaptic contacts were classified into several categories (presynaptically concave, convex, or flat) according to curvature (e.g., [15]). Vesicles were considered docked when the vesicle membrane was immediately adjacent to the active zone membrane (see Fig. 1). For each sampled synapse, a series of measurements were taken at every profile for a given synapse within the series. Section thickness was calibrated using the cylindrical diameters method [7]. Similar to previous methods [19,20], the area of the PSD equated to the length of the boundary of the PSD in each individual section times the section thickness. The sum of these products yielded the PSD area. Similarly, boutons were reconstructed by tracing their area in every section in which they appeared and multiplying this by section thickness. Curvature was measured by taking the shape of the PSD as a uniform arc (see Fig. 2). Through trigonometric identities, the angle of this arc from its origin (θ ARC ) can be deduced from the angle of the PSD (θ PSD ) by the following formula [3]: θARC = 360 − (2θPSD ) To maintain direction, all concave curvatures were made positive while convex curvatures were made negative.

94

D.F. Marrone et al. / Neuroscience Letters 378 (2005) 92–97

Fig. 2. Diagram illustrating the method for quantifying synaptic curvature in the current study. Depicted are (a) typical macular axospinuous (a) concave, (b) flat, and (c) convex synapses in the dentate gyrus MML. Below each micrograph is an enlarged picture of the same synapse depicting measurements for estimating curvature of the PSD (θ PSD ). The calculated θ PSD for the (d) concave synapse is 151.08◦ , while θ PSD for the (e) flat and (f) convex synapses are 178.83◦ and 159.71◦ , respectively.

Synaptic dimensions were analyzed using a one-way analysis of variance (p < 0.05) with curvature as the single factor, as well as Pearson product–moment correlations. Post hoc analyses were carried out using Tukey’s HSD. All tests used the Statistical Package for the Social Sciences 11.0. Over the entire synapse population, synapses in the MML exhibit similar characteristics to Schaffer collateral synapses (see Table 1). Dentate boutons most commonly make only a single synapse (82–87%), and show broad range from 0.009 ␮m3 to 0.134 ␮m3 (mean = 0.048 ␮m3 ). Similarly, docked vesicle number showed a wide range from 2 to 32 (mean = 8.78). The areas of PSDs are also broadly distributed (0.096–0.006 ␮m2 , mean = 0.029 ␮m2 ) with a marked skew. The structural characteristics of dentate synapses are (like CA1) highly correlated so that larger boutons generally have larger PSDs, more vesicles, and more docked vesicles. Furthermore, there is a small significant correlation between

docked vesicle density (i.e., ␮m2 /docked vesicle) and active zone size (r = 0.373), such that the vesicles of larger synapses are further apart. No morphological measure correlated with PSD curvature across the entire synapse population (see Table 2). Although there is a great deal of overlap between synapses of different curvature profiles, the average characteristics of these synapse populations are distinct (see Table 1). Foremost, dentate PSDs of different classifications exhibited a significant difference in their degree of curvature [F(2,66) = 52.054, p < 0.001; all contrasts p < 0.001]. Dentate PSDs of different curvatures also exhibited a significant size difference [F(2,66) = 4.371, p = 0.016]. Concave PSDs were significantly larger than flat (p = 0.017), but not convex (p = 0.706) synapses, while convex and flat synapses showed no significant difference (p = 0.104) post hoc. Dentate boutons of different synaptic curvatures significantly differed in their number of docked vesicles [F(2,66) = 4.610, p = 0.013].

Table 1 Summary of presynaptic morphology across synapse types Shape Concavea N Bouton volumeb Area/docked vesiclec PSD curvatured PSD areac Percent of vesicles docked Number of docked vesicles Total vesicles a b c d e f g

0.052 0.003 12.037e,f 0.034f 3.466 10.190f 348.190

Convexa ± ± ± ± ± ± ±

21 0.008 0.000 1.460 0.004 0.382 1.206 37.882

0.056 0.003 −26.321f,g 0.031 2.763 9.520 392.390

Values are provided ±standard error of the mean. Measures are provided per ␮m3 . Measures are provided per ␮m2 . Measures are provided in degrees. Significant difference from convex synapses (Tukey’s HSD, p < 0.05). Significant difference from flat synapses (Tukey’s HSD, p < 0.05). Significant difference from concave synapses (Tukey’s HSD, p < 0.05).

Flata ± ± ± ± ± ± ±

23 0.007 0.000 4.280 0.002 0.266 0.668 35.272

0.036 0.003 2.678e,g 0.023e 2.783 6.920e 270.960

Totala ± ± ± ± ± ± ±

25 0.004 0.000 1.559 0.002 0.215 0.535 23.998

0.048 0.003 −3.778 0.029 2.984 8.780 334.940

± ± ± ± ± ± ±

69 0.004 0.000 2.527 0.002 0.167 0.495 19.366

D.F. Marrone et al. / Neuroscience Letters 378 (2005) 92–97

95

Table 2 Summary of correlations in morphology across all synapses Synapse type

Parameter

PSD curvature (PC)

Area/docked vesicle (AV)

Bouton volume (BV)

Overall

PC AV BV PA DV TV

1

0.005 1

−0.081 0.165 1

Concave

PC AV BV PA DV TV

1

0.144 1

0.286 0.144 1

Convex

PC AV BV PA DV TV

1

−0.054 1

0.024 0.301 1

Flat

PC AV BV PA DV TV

1

−0.190 1

−0.027 0.077 1

∗ ∗∗

PSD area (PA)

Number of docked vesicles (DV)

Total vesicles (TV)

0.076 0.373** 0.441** 1

0.074 −0.173 0.391** 0.839** 1

−0.094 0.038 0.859** 0.364** 0.374** 1

0.816** 0.481* 0.472* 1

0.815** −0.104 0.460* 0.817** 1

0.096 0.018 0.806** 0.246 0.280 1

−0.084 −0.420* 0.181 0.813** 1

0.138 0.243 0.886** 0.384 0.223 1

0.435* −0.144 0.282 0.832** 1

0.135 −0.148 0.851** 0.431* 0.578** 1

−0.153 0.172 0.369 1

0.284 0.415* 0.279 1

Correlation is significant at the 0.05 level (two-tailed). Correlation is significant at the 0.01 level (two-tailed).

Concave synapses displayed a greater number of docked vesicles than flat (p = 0.017), but not convex (p = 0.838) synapses, while convex and flat synapses showed no significant difference (p = 0.062) post hoc. However, when PSD area is accounted for, there is no difference in volumetric (i.e., PSD area per docked vesicle) density [F(2,66) = 0.032, p = 0.968]. Boutons with different curvatures also showed a significant main effect in their total numbers of vesicles [F(2,66) = 3.805, p = 0.027], although none of these parameters reached significance post hoc. Bouton volume showed a non-significant trend towards the largest boutons being convex, and boutons with flat synapses being the smallest [F(2,66) = 3.039, p = 0.055]. No significant difference was observed in the percentage of docked vesicles [F(2,66) = 1.860, p = 0.164]. Across all curvature profiles, many structural characteristics of dentate synapses are highly correlated (see Table 2). Convex synapses, however, did not show correlations between bouton size and PSD area, or between curvature and docked vesicle number. Thus, while concave and flat synapses on larger boutons have larger PSDs and more docked vesicles, this is not true of convex synapses. Furthermore, convex synapses show a significant negative correlation between docked vesicle number and density (i.e., the greater the number of vesicles, the further apart they are). Flat synapses showed no correlation between the total number of vesicles and either PSD area or docked vesicle number.

Based on this data, two basic conclusions can be made. When analyzed collectively, synapses of CA1 and the dentate MML seem consistent in their general morphology. In addition, when synapses are categorized based on their curvature, they show several morphological distinctions that are of relevance to PTR. Before examining these conclusions, however, the consistency of the current data with previous research must be addressed. Although several notable reports have been published concerning synaptic morphology in the dentate gyrus (e.g., [5,6]), these results were based on single random sections. As this forms a major source of bias [2], these studies are not addressed here. Furthermore, most studies utilizing serial sections in the dentate gyrus (e.g., [8]) assess only synaptic density. Due to these factors, attention will be turned to morphological studies conducted in CA1. Across the entire synapse population, the data reported here are in agreement with previous observations on Schaffer Collateral synapses, although the values reported here for several statistics are somewhat smaller than previously reported. The number of docked vesicles observed here (8.78) are smaller than values obtained by both Harris and Sultan [13], who reported 15.6 vesicles (range = 2–36) per active zone, as well as Schikorski and Stevens [19], who reported 10.3 vesicles (range = 2–27). The average bouton volume reported here (0.048) is also smaller than both Harris and Sultan (0.11 ␮m3 ) and Schikorski and Stevens (0.086 ␮m3 ). This,

96

D.F. Marrone et al. / Neuroscience Letters 378 (2005) 92–97

however, is most likely accounted for by the sampling procedures used. Both multisynaptic and perforated synapses, which are among the largest synapses, have been excluded from the current study to more accurately assess synaptic curvature. In addition, the density of docked vesicles reported here (0.003 ␮m3 /vesicle) agree with those of Schikorski and Stevens (62 nm × 62 nm = 0.0036 ␮m3 /vesicle), and both studies show the same linear relationships between bouton size, PSD size, and the number of both docked and total vesicles. These observations also replicate many of the relationships between docked vesicles, bouton size, and PTR previously reported (e.g., [18,22,25]). Thus, it seems that synapses in these two areas are consistent in their gross morphology. When considering curvature, concave synapses have the most docked vesicles, while flat synapses tend to have the fewest. This, however, is accounted for entirely by size differences in these synapses. There was also a substantial correlation between the degree of curvature and the number of docked vesicles in both concave and flat synapses. If increased compartmentalization is acting to sequester Ca2+ and increase PTR, then one should expect to see this correlation between the morphological correlates of compartmentalization (PSD curvature), and PTR (docked vesicles) in concave synapses. By the same logic, if convex synapses inefficiently sequester Ca2+ , as mathematical models [9,10] suggest, these correlations should not be observed. In fact, no correlations were seen between bouton size and PSD area or between curvature and number of docked vesicles in convex synapses. These conclusions, however, should be made with caution. While not analyzing both CA1 and dentate synapses in the same animals is a limiting factor for this study, the impact of this is minimal considering that the current data agree with the observations of multiple labs using different methods. In addition, these results cannot legitimately be extrapolated to the entire dentate gyrus since the current study examined only the MML. This may mask laminar differences, as seen in other regions [20]. However, this same limitation is true of the CA1 observations cited, which have almost exclusively examined the stratum radiatum. In summary, while the morphological data suggest that concave synapses have a higher PTR than flat or convex synapses, this seems largely because concave synapses are larger. Their volumetric density of docked vesicles remains unchanged. Mathematical models, however, suggest that the location of release along the active zone may impact synaptic function. While in area CA1 the pattern of docking appears random across the entire synapse population [19], vesicles may dock in distinct locations in synapses of different curvatures. Conceivably, curvature may alter PTR because concave terminals (a) have a greater density of docked vesicles; or (b) tend to have vesicles docked in locations permitting greater local residual [Ca2+ ]i . Although the current data confirm that (a) is not the case, (b) is possible. This issue remains to be addressed.

Acknowledgement This Research was supported by the Natural Sciences and Engineering Research Council of Canada in the form of an operating grant to T.L.P. and a postgraduate scholarship to D.F.M.

References [1] C. Boyer, T. Schikorski, C.F. Stevens, Comparison of hippocampal dendritic spines in culture and in brain, J. Neurosci. 18 (1998) 5294–5300. [2] R. Coggeshall, H. Lekan, Methods for determining the number of cells and synapses: a case for more uniform standards of review, J. Comp. Neurol. 364 (1996) 6–15. [3] H.S.M. Coxeter, Introduction to Geometry, H.M. Wiley & Sons, New York, 1963. [4] T. Deller, C.A. Haas, M. Frotscher, Reorganization of the rat fascia dentata after a unilateral entorhinal cortex lesion: Role of the extracellular matrix, Ann. N.Y. Acad. Sci. 911 (2000) 207–220. [5] N.L. Desmond, W.B. Levy, Synaptic correlates of associative potentiation/depression: an ultrastructural study in the hippocampus, Brain Res. 265 (1983) 21–30. [6] N.L. Desmond, W.B. Levy, Changes in the postsynaptic density with long-term potentiation in the dentate gyrus, J. Comp. Neurol. 253 (1986) 476–482. [7] J.C. Fiala, K.M. Harris, Cylindrical diameters method for calibrating section thickness in serial electron microscopy, J. Microsc. 202 (2001) 468–472. [8] Y. Geinisman, L. de Toledo-Morrell, F. Morrell, R.E. Heller, M. Rossi, R.F. Parshall, Structural synaptic correlate of long-term potentiation: formation of axospinous synapses with multiple, completely partitioned transmission zones, Hippocampus 3 (1993) 435–445. [9] T. Ghaffari-Farazi, J. Liaw, T.W. Berger, Impact of synaptic morphology on presynaptic calcium dynamics and synaptic transmission, Soc. Neurosci. Abstr. 23 (1997) 2105. [10] T. Ghaffari-Farazi, J. Liaw, T.W. Berger, Consequence of morphological alterations on synaptic function, Neurocomputation 26/27 (1999) 17–27. [11] K.M. Harris, F.E. Jensen, B.H. Tsao, Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation, J. Neurosci. 12 (1992) 2685–2705. [12] K.M. Harris, J.K. Stevens, Dendritic spines of CA1 pyramidal cells in the rat hippocampus: serial electron microscopy with reference to their biophysical characteristics, J. Neurosci. 9 (1989) 2982– 2997. [13] K.M. Harris, P. Sultan, Variation in the number, location and size of synaptic vesicles provides an anatomical basis for the nonuniform probability of release at hippocampal CA1 synapses, Neuropharmacology 34 (1995) 1387–1395. [14] T.L. Ivanco, R.J. Racine, Long-term potentiation in the reciprocal corticohippocampal and corticocortical pathways in the chronically implanted, freely moving rat, Hippocampus 10 (2000) 143–152. [15] D.F. Marrone, J.C. LeBoutillier, T.L. Petit, Comparative analyses of synaptic densities during reactive synaptogenesis in the rat dentate gyrus, Brain Res. 996 (2004) 19–30. [16] D.F. Marrone, T.L. Petit, The role of synaptic morphology in neural plasticity: structural interactions underlying synaptic power., Brain Res. Rev. 38 (2002) 291–308. [17] V.N. Murthy, T.J. Sejnowski, C.F. Stevens, Heterogeneous release properties of visualized individual hippocampal synapses, Neuron 18 (1997) 599–612.

D.F. Marrone et al. / Neuroscience Letters 378 (2005) 92–97 [18] J.P. Pierce, L.M. Mendell, Quantitative ultrastructure of Ia boutons in the ventral horn: scaling and positional relationships, J. Neurosci. 13 (1993) 4748–4763. [19] T. Schikorski, C.F. Stevens, Quantitative ultrastructural analysis of hippocampal excitatory synapses, J. Neurosci. 17 (1997) 5858–5867. [20] T. Schikorski, C.F. Stevens, Quantitative fine-structural analysis of olfactory cortical synapses, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 4107–4112. [21] K.E. Sorra, K.M. Harris, Occurrence and three-dimensional structure of multiple synapses between individual radiatum axons and their target pyramidal cells in hippocampal area CA1, J. Neurosci. 13 (1993) 3736–3748.

97

[22] L.C. Streichert, P.B. Sargent, Bouton ultrastructure and synaptic growth in a frog autonomic ganglion, J. Comp. Neurol. 281 (1989) 159–168. [23] C.S. Wallace, G.L. Lyford, P.F. Worley, O. Steward, Differential intracellular sorting of immediate early gene mRNAs depends on signals in the mRNA sequence, J. Neurosci. 18 (1998) 26–35. [24] G.-Y. Wu, J. Kleim, D.F. Marrone, T.W. Berger, T.L. Petit, Morphological basis of neural plasticity, Winter Conference on Neural Plasticity, Guadeloupe, French Antilles, 2003. [25] M.B. Yeow, E.H. Peterson, Active zone organization and vesicle content scale with bouton size at a vertebrate central synapse, J. Comp. Neurol. 307 (1991) 475–486.