Biomedical Engineering Reference
In-Depth Information
MΩ to 1.2 GΩ, with 27% of the seals at more than 250 MΩ and 5% at higher than 1
GΩ. In whole-cell mode, the cells had a seal resistance of 100 to 250 MΩ at −80 mV.
hese giga-seal yields and seal resistance qualities were not competitive with the best
yields achievable by a human operator and a pipette, but the device ofered straight-
forward automation and multiplexing for pennies, because it was the irst device
that could integrate multiple units with microluidics using inexpensive polymer
micromolded materials (PDMS). he lateral aperture design has several advantages
over the planar aperture coniguration: (a) it does not interfere with the microscopic
observation of the cells; (b) aperture fabrication is simpler because it does not require
thinning down of a thick substrate and it can be molded; (c) luidic access to the res-
ervoirs is already designed in the photomask, does not require additional packaging,
and allows for integrated luidic automation such as microvalves and micropumps,
unlike in the planar patch clamp coniguration; and (d) last but not least, in the planar
geometry, the conducting bufers are separated by a relatively thin substrate, which
produces a nonnegligible capacitance, whereas in the lateral patch clamp chip design,
the two luids are separated by an insulator wall.
he author's group at the University of Washington in Seattle conducted an investi-
gation of the fundamental reasons why Luke Lee's PDMS design produced such low
yields and concluded that a key factor was its inability to keep the aperture clean from
cell debris and proteins. (In pipette-based recordings, to mitigate this problem, the
pipette is positively pressurized so as to continuously perfuse the aperture rim with
clean, nonproteic bufer until the very last moment, when the tip is very close to the
cell; only then, is the pressure inverted.) A lateral patch clamp PDMS design that
incorporated a dedicated “rinse line” to keep the aperture clean was demonstrated
(“D” for “drain” in
Figure 5.55
). Using RBL cells, the device delivered high-stability
giga-seals with success rates comparable with those of pipettes. he high stability
enabled exchanges of both the extracellular solution (delivered through nanochannel
nCh
EC
in
Figure 5.55
) and intracellular solution (delivered through nCh
IC
) during
whole-cell recordings. In a test of 103 diferent devices, 66 cells (64%) were success-
fully immobilized at the patch aperture; 38 cells (58% of immobilized cells, 37% of
all cells) were successfully giga-sealed; and 25 cells (65% of giga-sealed cells, 34% of
immobilized cells, 24% of all cells) were successfully perforated for whole-cell access.
In the last group of 27 experiments, 79% of the cells could be immobilized, of which
68% could be giga-sealed and 46% perforated for whole-cell access, indicating that
dexterity was still important.
How about if it were possible to create lateral apertures
in glass
? Ater all, glass seems
to be unmatched in terms of giga-seal quality as a material (cell membranes just stick
to clean glass more than any other material!). his is exactly what Levent Yobas' group
from the Institute of Microelectronics in Singapore did in 2007. In fact, Yobas' “lat-
eral” apertures are truly lateral, not cornered, as they are elevated from the bottom
of the channel and the cell is patched against a lat surface. he process relies on the
same physical principles as the preparation of conventional micropipette electrodes
(heat pulling and ire polishing) but using phosphosilicate glass (PSG, 8 wt.% phos-
phorus content) deposited through plasma-enhanced chemical vapor deposition in
microfabricated silicon trenches. A 2-μm-wide, 3.5-μm-deep Si trench was coated
with a 4-μm layer of PSG (which “closed” the trench, leaving a keyhole void inside)
and was heated to 1150°C for 30 minutes, resulting in relow (melting) of the keyhole
void into a cylindrical cavity. Apertures with a diameter of approximately 1.5 μm
(variation of <10%) were demonstrated. One hundred apertures were tested on RBL-1
cells: 61% formed giga-seals (>1 GΩ) and of those, approximately 48% (29% of all)
achieved whole-cell recordings (
Figure 5.56
).
