SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Abstract  Over a century ago, the Spanish anatomist Ramón y Cajal described ‘mossy fibres’ in the hippocampus and the cerebellum, which contain several presynaptic boutons. Technical improvements in recent decades have allowed direct patch-clamp recordings from both hippocampal and cerebellar mossy fibre boutons (hMFBs and cMFBs, respectively), making them ideal models to study fundamental properties of synaptic transmission. hMFBs and cMFBs have similar size and shape, but each hMFB contacts one postsynaptic hippocampal CA3 pyramidal neuron, while each cMFB contacts ∼50 cerebellar granule cells. Furthermore, hMFBs and cMFBs differ in terms of their functional specialization. At hMFBs, a large number of release-ready vesicles and low release probability (<0.1) contribute to marked synaptic facilitation. At cMFBs, a small number of release-ready vesicles, high release probability (∼0.5) and rapid vesicle reloading result in moderate frequency-dependent synaptic depression. These presynaptic mechanisms, in combination with faster postsynaptic currents of cerebellar granule cells compared with hippocampal CA3 pyramidal neurons, enable much higher transmission frequencies at cMFB compared with hMFB synapses. Analysing the underling mechanisms of synaptic transmission and information processing represents a fascinating challenge and may reveal insights into the structure–function relationship of the human brain.

  • image

[ Igor Delvendahl (left) received his MD degree from University of Freiburg in 2011, where he started his scientific career in clinical neurophysiology using transcranial magnetic stimulation (TMS). He studied neuroplasticity of the human cortex and the underlying physiological mechanisms of TMS, whereupon he became greatly interested in cellular neuroscience. Since 2012 he works with electrophysiological and calcium-imaging techniques on cerebellar mossy fibre synapses in Stefan Hallermann's laboratory. Stefan Hallermann (right) is a neurophysiologist with a background in medicine and physics. After leading an independent research group at the European Neuroscience Institute in Göttingen, Germany, he became head of a Department for Neurophysiology at the University Leipzig, Germany. He analyzed the mechanisms of synaptic transmission in the cerebellum, the hippocampus, and the cortex of vertebrates, as well as the neuromuscular junction of Drosophila melanogaster. He uses cutting-edge electrophysiological, optical, and computational techniques, such as direct whole-cell patch-clamp recordings from presynaptic nerve terminals, capacitance measurements, and two-photon calcium imaging.]

Abbreviations 
AZ

active zone

CA3

cornu ammonis region 3

EPSC

excitatory postsynaptic current

LTD

long-term depression

LTP

long-term potentiation

MFB

mossy fibre bouton

In the late 19th century, innovations in tissue staining and microscopy led to a great leap forward in the morphological description of the vertebrate nervous system. Golgi, Sala and Schaffer all mentioned nerve fibres, which the Spanish anatomist Ramón y Cajal described in detail (Ramón y Cajal, 1894, 1911). These fibres with their numerous, characteristic varicosities and filopodial extrusions along their axons were reminiscent of moss to Ramón y Cajal, and he thus coined the term ‘fibres moussues’ (mossy fibres; Ramón y Cajal, 1894; for more information on the historical naming see also Palay & Chan-Palay, 1974). Mossy fibres with similar appearance can be found exclusively in two very different brain areas, the hippocampus and the cerebellum.

With respect to the investigation of neuronal function, both hippocampal and cerebellar mossy fibre boutons (hMFBs and cMFBs, respectively) have two important advantages. (1) Direct presynaptic recordings are feasible (Geiger & Jonas, 2000; Rancz et al. 2007), which facilitates analysis of the fundamental mechanisms of neurotransmission at these synapses. To our knowledge, these are the only presynaptic terminals in the cortex or the cerebellum that allow direct presynaptic recordings. (2) The beautiful lamination of the hippocampus and cerebellum was seductive to neuroscientists. Therefore, these two brain areas are among the most thoroughly studied of the mammalian brain. For example, detailed computational models of the function of these brain regions have been developed (Marr, 1969; Lisman, 1999; Medina & Mauk, 2000). This combination offers the unique chance to understand the mechanisms by which specific synapses perform information processing within higher brain circuits. Furthermore, the structural similarity of hMFBs and cMFBs (Rollenhagen & Lübke, 2006), which are embedded in neuronal circuits with very distinct functions, should provide insights into the structure–function relationship of the human brain. In this review, we systematically compare anatomical and functional properties of hMFBs and cMFBs, and aim to relate fundamental mechanisms of synaptic transmission to the function of both synapses within the neuronal network.

Anatomical structure

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Several structural similarities exist between mossy fibres in hippocampus and cerebellum. In the hippocampus, the unmyelinated mossy fibres arise from granule cells in the dentate gyrus and exhibit collateralization in the hilar region, reaching a terminal axon diameter of ∼0.2 μm (Claiborne et al. 1986). Hippocampal mossy fibres possess a large number of characteristic swellings, which vary in size from submicrometre range to giant presynaptic boutons (Table 1; Fig. 1), forming excitatory, mostly en passant synapses onto their target cells. Mossy fibres have 10–18 large hMFBs, which make synaptic contact with one CA3 pyramidal cell each (Acsády et al. 1998). The size of these MFBs is in the range of ∼5 μm in adult animals (see Table 1). In addition, each mossy fibre contains ∼140 small varicosities (diameter ∼1 μm), and each hMFB has filopodial extensions; both target GABAergic interneurons (Fig. 1F; Acsády et al. 1998). These filopodia provide feedforward inhibition and exhibit structural plasticity during physiological and pathophysiological conditions (Frotscher et al. 2006; Torborg et al. 2010; Ruediger et al. 2011).

Table 1.  Properties of hMFBs and cMFBs
PropertyhMFBscMFBsReferences
  1. Abbreviations: MFB, mossy fibre bouton; AZ, active zone; RRP, release-ready pool; GC, granule cell; RT, room temperature. *All anatomical data are from adult rats. †Due to active zone clustering, the nearest distance does not reflect overall active zone density, which is ∼10-fold smaller in hMFBs than in cMFBs. ‡Based on release probability <0.3 with elevated Ca2+ concentration (cf. fig. 6C in Lawrence et al. 2004).

Bouton diameter (μm)2–8*3–12*(Claiborne et al. 1986; Jakab & Hámori, 1988; Rollenhagen & Lübke, 2006; Rollenhagen et al. 2007)
Bouton surface area (μm2)32–10669–266(Jakab & Hámori, 1988; Xu-Friedman & Regehr, 2003; Rollenhagen et al. 2007)
Nearest distance between AZs (μm)0.40†0.46(Xu-Friedman & Regehr, 2003; Rollenhagen et al. 2007)
Number of AZs per MFB18–45190–440(Xu-Friedman & Regehr, 2003; Rollenhagen et al. 2007)
RRP per AZ∼401–2(Hallermann et al. 2003, 2010)
Release probability<0.1‡0.5(Lawrence et al. 2004; Hallermann et al. 2010)
Connectivity1:1 (CA3)1:50 (GC)(Jakab & Hámori, 1988; Acsády et al. 1998)
τEPSC (ms) at RT6.21.3(Jonas et al. 1993; Silver et al. 1992)
In vivo firing patternGamma bursts (30–100 Hz)Rate coding (1–1000 Hz)(Csicsvari et al. 2003; Arenz et al. 2008)
image

Figure 1. Anatomical structure  A and B, illustrations of a hippocampal (Ramón y Cajal, A) and a cerebellar (Gray, B) sagittal slice. C and D, drawings of a microscopic view showing mossy fibres and their postsynaptic target cells. In the hippocampus, the granule cells of the dentate gyrus give rise to mossy fibre axons, which travel into the CA region. In the cerebellum, mossy fibres are the primary input and originate from, e.g., pontine nuclei in the brainstem. E and G, schematic illustrations of hippocampal (E) and cerebellar (G) mossy fibre boutons (magenta) and their postsynaptic targets (green), the CA3 pyramidal neuron (CA3) and the granule cell (GC), respectively. Inhibitory interneurons (blue) in the hippocampus (Int) and Golgi cells in the cerebellum (GoC) are targeted by filopodia of the MFBs. F and H, artistic drawing (F) and fluorescent image (H) of MFBs (magenta) showing their filopodial extensions making synaptic contacts with inhibitory interneurons (blue). (Figures modified, with permission, from: A, Ramón y Cajal (1911); B, Gray (1918); C, Ramón y Cajal (1933); D, Ramón y Cajal (1894); F, Acsády et al. (1998) Society for Neuroscience; H, Ruediger et al. (2011) Macmillan Publishers Ltd.)

Download figure to PowerPoint

In the cerebellum, the myelinated mossy fibres represent one of the major inputs to the cerebellum and originate from, e.g., pontine nuclei or the spinal chord. Each cerebellar mossy fibre forms several collaterals and features a large number of boutons along its axon at intervals of 20–80 μm (Palay & Chan-Palay, 1974). Mossy fibres enter the granule cell layer of the cerebellar cortex, where their main postsynaptic targets are cerebellar granule cells. Compared with hippocampal mossy fibres, the axon diameter is larger (in the range of 1 μm; Palay & Chan-Palay, 1974), which might contribute to the high-frequency firing observed at cMFBs (Jörntell & Ekerot, 2006; Rancz et al. 2007). Each cMFB, having a rosette-like shape, makes synaptic connections with approximately 50 granule cells (Jakab & Hámori, 1988), thereby forming a glomerular complex with the dendritic digits of granule cells. Thus, mossy fibre synapses show a high degree of divergence in the cerebellum as opposed to the hippocampus. cMFBs have a highly complex shape and three types of mossy fibre rosettes have been distinguished by their morphological appearance (Palay & Chan-Palay, 1974). The size of cMFBs is also highly variable, but seems on average to be greater than that of hMFBs (Table 1). During maturation, the size of cMFBs increases and the shape becomes more irregular (Hámori & Somogyi, 1983). In addition to their principal targets, each cMFB has filopodia, which target inhibitory Golgi cells (Fig. 1H). Thus, the filopodia of cMFBs provide feedforward inhibition like the filipodia of hMFBs, and show similar structural plasticity (Ruediger et al. 2011). In summary, the structure of MFBs is similar in the hippocampus and the cerebellum, but with one obvious exception: hMFBs contact one whereas cMFBs contact 50 principal neurons. In the following, we focus on the function of the synapses formed by the large hMFBs and cMFBs onto their principal target cells.

Short-term plasticity of synaptic transmission

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

When an action potential enters a presynaptic terminal, voltage-gated Ca2+ channels open and Ca2+ triggers exocytosis of neurotransmitter-filled synaptic vesicles. Both hMFBs and cMFBs use glutamate as neurotransmitter, which elicits excitatory postsynaptic currents (EPSCs) in both CA3 pyramidal neurons in the hippocampus and granule cells in the cerebellum with a peak amplitude of ∼100 pA (Fig. 2A and B). However, the kinetics of EPSCs is faster in the cerebellum (Table 1; Silver et al. 1992; Jonas et al. 1993). During repeated synaptic transmission, short-term plasticity can cause changes in EPSC amplitude due to various mechanisms such as presynaptic facilitation and vesicle depletion as well as postsynaptic glutamate receptor saturation and desensitization (Neher & Sakaba, 2008). While the initial EPSC amplitude is similar in hippocampal and cerebellar mossy fibre synapses, their short-term plasticity is very different as discussed in the following.

image

Figure 2. Short-term plasticity of synaptic transmission  A and B, examples of EPSCs in hippocampal CA3 pyramidal neurons (A) and cerebellar granule cells (B) elicited by activation of a single mossy fibre input. C and D, example EPSC traces of paired-pulse experiments (top) and paired-pulse ratio (defined as 2nd amplitude/1st amplitude) versus inter-pulse interval (bottom) at hippocampal (C) and cerebellar (D) mossy fibre synapses. E and F, examples of stimulus train experiments (top) and peak EPSC current as a function of stimulus number (bottom), indicating short-term synaptic facilitation at hMFBs (E) and short-term synaptic depression at cMFBs (F), respectively. (Figures modified, with permission, from: A, Jonas et al. (1993); B, Silver et al. (1992) Macmillan Publishers Ltd; C, Salin et al. (1996) National Academy of Sciences, U.S.A.; D, Saviane & Silver (2006) Macmillan Publishers Ltd; E, Lawrence et al. (2004); F, data taken from Hallermann et al. (2010).)

Download figure to PowerPoint

Hippocampal mossy fibre synapses show facilitation in paired-pulse experiments (Fig. 2C) and during short trains of action potentials (Fig. 2E). Furthermore, hMFBs display remarkable frequency facilitation occurring at frequencies as low as once every 40 s (Salin et al. 1996). Presynaptic kainate receptors seem to contribute to this facilitation (Schmitz et al. 2001). The strong facilitation suggests a low release probability and a large pool of release-ready vesicles. Indeed, fluctuation analysis indicates a very low release probability (<0.1; Lawrence et al. 2004), and presynaptic capacitance measurements suggest a large pool of release-ready vesicles per active zone (∼40; see Table 1; Hallermann et al. 2003). Consistent with the low release probability, a long diffusional distance from Ca2+ channels to Ca2+ sensor has been suggested for hMFBs (Blatow et al. 2003).

On the other hand, cerebellar mossy fibre synapses exhibit paired-pulse depression (Fig. 2D) and short-term depression during high-frequency transmission (Fig. 2F), suggesting high release probabilities and a small pool of release-ready vesicles. Indeed, quantal analysis indicates a release probability of ∼0.5 and only 1–2 release-ready vesicles per active zone (Table 1). To sustain high-frequency transmission, this small pool of release-ready vesicles is rapidly reloaded from a large number of releasable vesicles (300 per active zone; Saviane & Silver, 2006). For this rapid reloading of vesicles, filamentous active zone proteins might play an important role (Hallermann & Silver, 2012). Thus, hMFBs and cMFBs have very different functional properties such as different release probability, resulting in distinct short-term plasticity.

Long-term plasticity

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

The fundamental property of exhibiting long-term changes in synaptic strength is expressed at both hMFBs and cMFBs (Fig. 3). Long-term potentiation (LTP) and long-term depression (LTD) were initially described in hippocampus and cerebellum, respectively. The induction and expression mechanisms of LTP and LTD at pre- and postsynaptic sites are a matter of ongoing debate and have been investigated at both hippocampal and cerebellar MFB synapses (Nicoll & Schmitz, 2005). Here, we focus on presynaptic expression mechanisms of LTP at these two synapses.

image

Figure 3. Long-term potentiation  A and B, example EPSC traces before (black) and after (blue) high-frequency stimulation in hippocampal (A) and cerebellar (B) mossy fibre synpases (top), and average normalized EPSP slope (A) and EPSC amplitude (B) before and after LTP induction in hippocampal (A) and cerebellar (B) mossy fibres (bottom). Blue arrows indicate the time of high-frequency stimulation (HFS). C and D, paired-pulse ratios decrease following LTP induction in hippocampal (C) as well as cerebellar (D) mossy fibre synapses, indicative of a presynaptic contribution to LTP. (Figures modified, with permission, from: A, Galvan et al. (2008) Society for Neuroscience; B and D, Sola et al. (2004); C, Goussakov et al. (2000) Society for Neuroscience.)

Download figure to PowerPoint

LTP in the mossy fibre pathway of the hippocampus displays the typical time course of increased postsynaptic responses (Fig. 3A). However, in contrast to other excitatory hippocampal synapses, LTP is expressed presynaptically in hMFBs. For example, the decrease in paired-pulse facilitation upon induction of LTP (Fig. 3C) suggests an increase in release probability (Goussakov et al. 2000). In the cerebellum, high-frequency stimulation of mossy fibres leads to LTP of the postsynaptic response recorded from granule cells (Fig. 3B). The expression of LTP at the cMFB–granule cell synapse includes several mechanisms with a strong presynaptic component. During LTP, the paired-pulse ratio decreases (Fig. 3D), suggesting increased release probability (Sola et al. 2004). Thus, while the postsynaptic contribution to LTP expression seems larger in cerebellar than in hippocampal MFB synapses (Nicoll & Schmitz, 2005; Andreescu et al. 2011), the dynamic increase in transmitter release appears to be an important mechanism for the expression of LTP at both hMFBs and cMFBs. However, the presynaptic mechanisms underlying LTP are still not well understood. Presynaptic recordings may offer the chance to analyse these mechanisms at hMFBs and cMFBs directly.

Presynaptic recordings

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Our current knowledge of the fundamental aspects of synaptic transmission is largely based on analysis of the few synapses that offer direct access to presynaptic terminals, such as, for example, the calyx of Held (Neher & Sakaba, 2008; Borst & Soria van Hoeve, 2012). Recently, technical improvements have made recordings from hMFBs and cMFBs possible (Geiger & Jonas, 2000; Bischofberger et al. 2006b; Rancz et al. 2007), which in the cerebellum can also be performed in vivo (Rancz et al. 2007). In both terminals, the capacitive currents in response to small voltage steps decay multi-exponentially, representing the charging of the bouton and the adjacent axonal compartments. Consistent with their morphological similarity, the membrane capacitance of the MFBs is similar (1.4 vs. 1.8 pF in hMFBs and cMFBs; Hallermann et al. 2003; Rancz et al. 2007). Long depolarizations in current-clamp mode elicit up to three action potentials in hMFBs, while only a single action potential occurs in cMFBs (Fig. 4C and D). Interestingly, action potentials recorded from hMFBs are metabolically very efficient (Alle et al. 2009). However, metabolic inefficiency supports conduction reliability of high-frequency action potentials (Hallermann et al. 2012). As cMFBs seem to operate in the kilohertz regime (Table 1), it will be interesting to analyse, for example, sodium channel properties (Khaliq et al. 2003) and their impact on action potential efficiency at this terminal. Finally, both hMFBs and cMFBs demonstrate a pronounced outward rectification, but only cMFBs display a ‘sag’ upon hyperpolarization (Fig. 4D, arrow), which is characteristic for Ih currents.

image

Figure 4. Presynaptic recordings  A, confocal image of an hMFB filled with biocytin. The position of the patch pipette is indicated (CA3, cornu ammonis region 3; strat. luc., stratum lucidum). B, projection of a two-photon image stack of a whole-cell patch-clamped cMFB in a Thy1 mouse with fluorescently labelled MFBs (green). The patch-clamped cMFB, its axon and the cMFBs along the axon are filled with a fluorescent dye (magenta) via the patch pipette (ML, molecular layer; PC, purkinje cell layer; GC, granule cell layer; WM, white matter). C and D, voltage response of a hippocampal (C) and a cerebellar (D) MFB recorded in current clamp mode to hyper- and depolarizing current pulses. The cMFB shows a characteristic ‘sag’ in response to hyperpolarizing currents (arrow). E and F, voltage commando with sinusoidal stimulation (± 50 mV; holding potential −100 mV; 5 kHz in E and 1 kHz in F) and a depolarization to 0 mV for 30 ms (top trace); resulting pharmacologically isolated Ca2+ current (middle trace); and capacitance increase elicited by 30 ms depolarization (bottom trace; note different time scale). (Figures modified, with permission, from: A, Bischofberger et al. (2002) Society for Neuroscience; C, Geiger & Jonas (2000) Elsevier; D, Rancz et al. (2007) Macmillan Publishers Ltd; E, Hallermann et al. (2003); B and F, our unpublished data.)

Download figure to PowerPoint

Presynaptic capacitance measurements allow a rather direct quantification of transmitter release. This technique is based on the increase in electrical capacitance of a presynaptic terminal upon vesicle fusion (Neher & Marty, 1982). At hMFBs, the size of the capacitance increase indicates a large pool of vesicles, resulting in an estimate of ∼40 vesicles per active zone that can fuse during 30 ms (Fig. 4E; Hallermann et al. 2003). However, hMFBs also contain a small proportion of large dense-core vesicles, which might lead to an overestimation of the pool size. On the other hand, simultaneous exo- and endocytosis will lead to an underestimation. Up to now, it was only possible at very few synapses to demonstrate that neurotransmitter-filled vesicles exclusively contribute to the capacitance increase (von Gersdorff et al. 1998; Sakaba, 2006). Nevertheless, a large pool of release-ready vesicles could well explain the pronounced short-term- and frequency-facilitation of hMFBs (Fig. 2; Salin et al. 1996; Lawrence et al. 2004).

At cMFBs, we recently established capacitance measurements (Fig. 4F; unpublished data). The results indicate that the capacitance increase is larger than at hMFBs. However, considering the ∼10-fold higher number of active zones at cMFBs compared with hMFBs (Table 1), the number of release-ready vesicles per active zone seems to be smaller in cMFBs, consistent with the 1–2 vesicles estimated by quantal analysis (Hallermann et al. 2010). Thus, presynaptic recordings now allow application of techniques such as capacitance measurements and photolytic Ca2+ uncaging at both hMFBs and cMFBs. This will facilitate investigation of fundamental mechanisms of transmitter release and its modulation during short- and long-term plasticity, with the ultimate aim to gain insight into the mechanisms of information processing within the hippocampal and cerebellar neuronal network.

Network function of mossy fibres

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Although the ultrastructure and principal function of MFBs are beginning to become clearer due to intensive investigation of these two synapses, elucidating the role of mossy fibre synapses for information processing is more difficult. In the hippocampus, gamma oscillations (30–90 Hz) are evoked under specific behavioural conditions, such as exploration, when they typically coexist with theta oscillations (3–8 Hz; Buzsáki & Draguhn, 2004; Bartos et al. 2007). This interplay allows encoding of independent information by the frequency of action potentials (rate coding) as well as by the phase relation of action potentials to the underlying theta rhythm (temporal coding). For example, the location of a rat running on a linear track seems to be encoded by the phase relationship of action potential bursts to the concurrent cycle of the hippocampal electroencephalogram (Fig. 5A), while the speed of the rat appears to be encoded by action potential frequency (Huxter et al. 2003).

image

Figure 5. Network function of mossy fibres  A, illustration of a rat running on a linear track (top); colour-coded firing frequency of a place cell depending on the position of the rat (middle); and EEG theta rhythm and place cell firing (red) on a single run (bottom). Grey lines through theta waves indicate corresponding points within the phase. The timing of place cell bursts relative to the phase of the theta cycle correlates with the position of the rat. B, illustration of horizontal rotation of a mouse (left); and rotational velocity (orange) and EPSCs recorded in vivo from a cerebellar granule cell (black trace) show that firing rates at the cMFB-granule cell synapse correlate with velocity of rotation in one direction (right). C, in vivo intracellularly evoked action potentials in a hippocampal granule cell (top) and simultaneous extracellular recordings in the hippocampal CA3 region (bottom). The time-locked response of the putative CA3 pyramidal cell to the granule cell action potentials indicates a ‘conditional detonator’ function of hMFB synapses. D, in vivo whole-cell recordings of action potentials in a cMFB (top) and EPSCs in a cerebellar granule cell (bottom) during air puff stimulation of the lip (indicted in blue). (Figures modified, with permission, from: A, Huxter et al. (2003) Macmillan Publishers Ltd; B, Arenz et al. (2008) AAAS; C, Henze et al. (2002) Macmillan Publishers Ltd; D, Rancz et al. (2007) Macmillan Publishers Ltd.)

Download figure to PowerPoint

The hMFB–CA3 synapse is an important part of a trisynaptic circuit, and modelling suggests that this synapse has a ‘detonator’ function (Lisman, 1999; Bischofberger et al. 2006a). In vivo studies indicate that this may indeed be the case, but only under specific conditions (‘conditional detonator’; Henze et al. 2002). The mossy fibre discharge can precisely and with high fidelity time the CA3 output (Fig. 5C). However, the chance of CA3 firing is increased during a high-frequency (gamma) burst of mossy fibre activity (Henze et al. 2002). The pronounced short-term facilitation of hMFBs in the gamma range (cf. Fig. 2E) should help to implement the CA3 firing upon the condition of granule cell bursting.

The cerebellum seems to operate primarily with rate coding, reaching frequencies as high as 1 kHz (Jörntell & Ekerot, 2006; Rancz et al. 2007). During horizontal rotations of mice, for example, the frequency of EPSCs of some granule cells linearly correlates with the speed of rotation in one direction (Fig. 5B; Arenz et al. 2008). In contrast to the hippocampus, sensory information in the cerebellum can be encoded by a constant frequency for sustained periods (van Kan et al. 1993). In addition, rhythmic neuronal oscillations seem less prominent in the granule cell layer of the cerebellum (but see D’Angelo et al. 2009).

At the cMFB–granule cell synapse, rapid vesicle reloading sustains high bandwidth synaptic transmission (Saviane & Silver, 2006; Hallermann et al. 2010), which might enable a linear computational algorithm in the cerebellum (Walter & Khodakhah, 2006). Furthermore, short-term plasticity of cMFBs contributes to neuronal gain control (Rothman et al. 2009) and supports coincidence detection by granule cell integration (Dean et al. 2010). However, a burst of spikes in a single cMFB can evoke granule cell spiking, indicating a ‘detonator’ function analogous to the hMFBs (Rancz et al. 2007). Therefore, temporal coding might also play a role for integration of cMFB inputs by granule cells (D’Angelo & De Zeeuw, 2009). Thus, the hippocampus and cerebellum use both temporal as well as rate coding, with a preference in the hippocampus for temporal and in the cerebellum for rate coding. Furthermore, cMFBs can use ∼10-fold higher transmission frequencies than hMFBs (Table 1).

Conclusion

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Despite a similar morphology, which justified the identical naming, hMFBs and cMFBs show marked differences in synaptic transmission, short-term plasticity and their function within the neuronal network. To our knowledge, hMFBs and cMFBs are the only presynaptic terminals in cortex or cerebellum that allow direct presynaptic recordings. The hippocampus and the cerebellum are furthermore among the most intensely studied brain areas, providing knowledge on possible functions of the mossy fibre synapses within the neuronal network. This offers the unique chance to analyse higher brain function from a biophysical up to a behavioural level. Here, we argue that the distinct biophysical properties of hMFBs and cMFBs are optimally suited to perform the required synaptic function within the neuronal network. In the hippocampus, a large number of release-ready vesicles of hMFBs (Hallermann et al. 2003) and a low release probability (Lawrence et al. 2004) contribute to pronounced short-term facilitation (Salin et al. 1996). This allows ‘conditional detonation’ of CA3 pyramidal neurons depending on the bursting activity of the hMFBs in the gamma frequency range during theta frequency cycles (Lisman, 1999; Henze et al. 2002). In the cerebellum, on the other hand, a small number of release-ready vesicles, high release probability and rapid vesicle reloading allow cMFBs to sustain high-frequency transmission over prolonged periods (Saviane & Silver, 2006; Hallermann et al. 2010). This enables rate-coded signalling (Arenz et al. 2008), with a fidelity that allows a single cMFB input to ‘detonate’ the postsynaptic granule cell (Rancz et al. 2007). But despite the progress in our understanding of synaptic, cellular and network function, several questions remain to be elucidated: What are the molecular mechanisms underlying the distinct function of hMFBs and cMFBs? How can cMFBs operate ∼10-fold faster than hMFBs? How is information processed by the corresponding synapses? Comparing the mechanisms of neurotransmission at hMFBs and cMFBs should provide interesting insights into the function of our brain.

References

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Appendix

  1. Top of page
  2. Abstract
  3. Anatomical structure
  4. Short-term plasticity of synaptic transmission
  5. Long-term plasticity
  6. Presynaptic recordings
  7. Network function of mossy fibres
  8. Conclusion
  9. References
  10. Appendix

Acknowledgements

S.H. received funding from the Heisenberg Program of the German Research Foundation (DFG HA 6386/1-1 and 2-1).