Well, I'm not sure why your friend is using that internal solution, but recording at -65mV will not give him true EPSCs as the Erev for Cl in his case is around -85 mV. 

Here is the site where you can calculate your Erev for different ions

http://www.pages.drexel.edu/~ek89/nernst.html

Hi Bradley,

The Cs-gluconate solution you described should be fine for EPSC/IPSC recordings. Of course you should hold the cell at the reversal for EPSCs when recording IPSCs and vice versa. Remember also to adjust you holding potential for the junction potential given by your internal and external solutions (you can calculate it in Clampex).

Hello,

Maximiliano is right. Using a Cs-Gluconate solution gives you a junction potential around 10-15 mV meaning that when you hold your cell at -70 mv you're actually at -85 mV. This is due to the different mobility of Cs vs gluconate at the tip of your pipette.

https://en.wikipedia.org/wiki/Liquid_junction_potential

you can calculate it with :

http://web.med.unsw.edu.au/phbsoft/LJP_Calculator.htm

For reliable recording it is important to use K-gluconate. In addition to the problem of juction potential,  Cs blocks K channels and causes decrease of resting potential. So, parts of the cell, which you can't properly clamp (this is usual problem at whole-cell recording from neurons) will be artificially depolarised. 

Both internals are fine for patch clamp recording. K-based internals are best when needing to record action potential firing in current clamp mode.  Cs-based internal is best when trying to record small events like minis or post synaptic events in voltage clamp mode, because Cs blocks K channels and leads to what is called better space clamp. You can't record APs with Cs based internal. As far as the chloride reversal, there are multiple ways to deal with this to allow recording of mEPSCs and mIPSCs from single recorded neurons. You can use Cs internal and as recommended change the holding potential to -70mV for mEPSCs (where mIPSCs are negligible), versus 0 mV for mIPSCs (where mEPSCs are negligible). You can also use K-based internal and a low chloride formulation to make the chloride reversal very negative (more than your colleagues recipe), which allows you to measure mEPSCs as inward currents and mIPSCs as outward currents using a single holding potential, usually more like -60mV but depends on the exact chloride reversal. 

Two more quick points. 1) it is very challenging to record whole cell from CA3 beurons in acute slices, so congrats to you for being able to do this routinely.  2) a clarification that you can't record mIPSCs in voltage clamp at 0 mV holding potential using K-based internal solution. Depolarizing the cell beyond about -55mV leads to strong activation of K channels and prolonged holding at zero would destabilize and probably kill your cell. You should verify exactly what your colleague is doing to record mIPSCs because this can't be correct. Maybe you have a simple miscommunication on the details. 

Hello,

May I know what extracellular solution you are using? Is the Erev you reported based on calculation or direct measurement?

I am analyzing the reversal potential of hippocampal neurons from 198 papers. I have a similar problem. In the Cs and gluconate based intracellular solutions, the GABA-A reversal potentials that I calculate with GHK equation have a significant  difference with experimentally measured values in which authors claim they have corrected junction potential with agar bridge. HCO3 and Cl cotransporters might not work in these solutions. pH difference between intracellular and extracellular solution change the HCO3 concentration inside the cell. The O2/CO2 oxygenation might be another factor that changes the HCO3 concentration and ultimately the Erev of GABA-A.

For example look at this experiment in whole-cell mode in which authors report that the junction potential is corrected with agar.

from Hájos 1997 Synaptic communication among hippocampal interneurons: properties of spontaneous IPSCs in morphologically identified cells.

Extracellular solution: 126 NaCl, 2.5 KCl, 26 NaHCO3, 2 CaCl2, 2 MgCl2, 1.25 NaH2PO4

Intracellular solution: 135 Cs-gluconate, 5 CsCl, 20 HEPES, 2 MgCl2 , 2 MgATP

Calculations:

Cao= 2

Cli= 9

Clo= 136.5

Csi= 140

HCO3o= 26

HCO3i=0

Ko= 2.5

Nao= 153.25

Assuming PCl = 1 and and PHCO3 = 0.18

GABA-A Erev wit GHK = -72.86, but reported experimentally measured value = -44.6±2.6.

If we assume pH in the bath is 7.4 and pH inside the cell is 7.25 the HCO3 concentration before the start of whole-cell recording should be 

HCO3i = exp(pHo - pHi)  * HCO3o = 28.73

GABA-A Erev with GHK = -60.84, which is 12 mV closer the experimental value. I am not sure how we can explain the remaining difference.

Assuming the HCO3i is going to diluted with the intracellular solution to one half of its original value the Erev should be -66.17

I have found one study that suggests gluconate ions permeates the GABA-A. they have calculate the permeability of 0.11 relative to PCl. This might be a factor. However, in many experiments that I have analyzed calculations without considering gluconate permeability provides better results.

Fatima-Shad 1993 Anion permeation in GABA- and glycine-gated channels of mammalian cultured hippocampal neurons.