Estimates of quantal synaptic parameters in light of more complex vesicle pool models

Grit Bornschein Simone Brachtendorf Hartmut Schmidt ^*

• Leipzig University, Leipzig, Germany

The subdivision of synaptic vesicles (SVs) into discrete pools is a leading concept of synaptic physiology. To better explain specific properties of transmission and plasticity, it has been suggested initially that the readily releasable pool (RRP) of SVs is subdivided into two parallel pools differing in their release probability. More recently, evidence was provided that sequential pools with a single RRP and a series-connected finite-size replenishment pool (RP) inserted between the reserve pool (RSP) and RRP equally well or even better account for most aspects of transmission and plasticity. It was further suggest that a fraction of the presynaptic release sites (N) are initially unoccupied by SVs, with vesicle replenishment occurring rapidly during activity, and furthermore that the number of release sites itself changes with rapid dynamics during activity. Experimentally it is difficult to obtain insights into the organization of SV pools directly and the interpretation of the data typically requires complex modelling. Here we propose a framework that identifies specific signs of the presence of the series-connected RP without complex modelling, using a combination of two experimental electrophysiological standard methods, cumulative analysis (CumAna) and multiple probability fluctuation analysis (MPFA). CumAna uses cumulative EPSC amplitude plots recorded during trains of action potentials and estimates the pool of releasable SVs from the y-intercept (y(0)) of a linear fit to the amplitudes late in the train. MPFA estimates N (NMPFA) from a parabolic fit to a variancemean plot of EPCS amplitudes recorded under conditions of different release probabilities. We show here, in particular, that if y(0) > NMPFA this is a strong indication for a series-connected RP. This is due to the fact that y(0) reports the sum of RRP and RP. Our analysis further suggests that this result is not affected by unoccupied release sites, as such empty sites contribute to both estimates, y(0) and NMPFA. We discuss experimental findings and models in the recent literature in the light of our theoretical considerations.