Membrane technologies for channel proteinbased sensing

1
Membrane technologies forchannel protein-based
sensing

• Schmidt Group
• UCLA Department of Bioengineering
• schmidtlab.seas.ucla.edu

2
Channel proteins are natural sensors

• Channel proteins are typically 5-15nm in size and
  inhabit lipid bilayer membranes
• There is a water-filled channel which runs down
  the center of the protein
• Channel proteins can exhibit charge or size
  selectivity due to the presence of charged or
  steric constrictions within the channels
• Natural sensing Applied voltage or binding of
  ligands to the channel can induce conformational
  changes which gate its conductance

3
Probing channel proteins experimentally

• The conductance state of the channel can be
  probed electrically by measuring ionic currents
  flowing through the channel in response to an
  applied voltage
• The lipid bilayer membrane is electrically
  insulating, with resistance typically greater
  than 10GW (conductance lt 100pS) for membranes
  hundreds of mm in size
• The conductance of a typical channel protein is
  5-1000pS, which gives rise to a .5-100pA ionic
  current in response to a 100 mV applied potential
• Binding foreign material to the channel interior
  can significantly block current

4
High-throughput drug screening of channel proteins
Source Meyer et al. Assay and drug development
technologies. 2507-514 (2004)
5
Sensing using engineered proteins

• The Bayley group has engineered a large number of
  mutants of the bacterial pore a-hemolysin to
  contain different binding sites within the
  channel
• Example cation binding sites using His4 bind to
  a number of different cations, each
  distinguishable through examination of the
  magnitude and temporal signature of the ionic
  current it blocks
• Stochastic Sensing
• The occurrence and duration of each binding event
  is random, but statistically show the
  concentration of analyte in solution, as well as
  its affinity for the binding site
• Measurements occur on time scales on the order of
  minutes or less

6
Fast single molecule nanopore DNA sequencing

• Initial work by Kasianowicz (PNAS 1996) looked
  the current through aHL modulated by the passage
  of polymers of RNA and DNA through it
• Since the membrane is highly insulating and the
  rest of the solution highly conductive, there is
  a huge electric field in the pore which drives
  the charged polymer through very rapidly
• All 100 of these bases traverse the pore in lt2ms,
  about 10-20us/base (Akeson Biophys J 1999)
• We need to measure pA currents in high bandwidths

7
Obstacles toward the technological exploitation
of channel proteins

• We can direct the self-assembly of lipids to
  create membranes with a planar or spherical
  geometry
• Although vesicles are generally more robust than
  planar membranes, the planar geometry ensures
  that we have access to both sides of the membrane
  for full control of the electrical and chemical
  environment of the protein
• The primary hurdle in the creation of practical
  devices using channel proteins is the short life
  and fragility of planar membranes

8
Freestanding planar lipid bilayer membrane
fabrication
Figure from Mayer, M., et al., Biophys. J.
85(4)2684-2695. (2003).
9
Painted membranes (Black Lipid Membranes)
Figures from White in Ion Channel Reconstitution
10
Painted membranes (Mueller-Rudin) (Black Lipid
Membranes)
Membranes are short-lived, 12 hours
11
Solvent-free membranes(Montal-Mueller method)
Figures from White in Ion Channel Reconstitution

• Langmuir films of lipid form at the air-water
  interface and form a membrane when the water
  level is raised beyond a hole that has had a
  suitable pretreatment with a lipid/organic
  solution
• Not really solvent-free. Membranes are
  short-lived, 12 hours

12
Addressing these shortcomings

• Freestanding planar membranes are meta-stable and
  have intrinsic lifetime limits
• Fixes
• (Get rid of the membrane and protein channel?)
• Substitution of lipid with biomimetic polymers
• Supported membranes
• Membranes in contact with solid surfaces
• Membranes in contact with porous (gel) surfaces
• Automated microfluidic formation

13
Lipid substitutes

• Amphiphilic polymers
• E.g. pluronics
• There is a lot of interest in manipulating
  amphiphilic polymers to self-assemble into a
  range of macromolecular structures for drug
  delivery and other applications
• Di-block copolymers (Bates, Discher)
• Di-block copolypeptides (Deming)
• Tri-block copolymers (Meier)
• A number of experiments creating biomimetic
  membranes (9nm thick!) formed from these polymers
  containing protein
• The hydrophilic PMOXA groups also have a
  methacrylate group on the end, enabling them to
  be crosslinked
• Increases vesicle lifetime and robustness

Discher, Science 284, 1143 (1999)
Nardin et al., Langmuir 16 1035 (2000)
14
Channel proteins can be functionally incorporated
into polymer vesicles

• Meier incorporated a number of channel and
  pore-forming proteins (OmpF, LamB, Alamethicin,
  etc.) and demonstrated that these proteins retain
  their ability to form channels as well as their
  native properties
• Lambda phage docking with LamB incorporated into
  polymer vesicles
• OmpF gating in the presence of a Donnan potential
  
• Creation of asymmetric ABC triblock copolymers
  with controlled A and C blocks can control the
  orientation of inserted protein (Stoenescu,
  Macromol. Biosci. 2004, 4, 930)

Graff PNAS 99, 5065 (2002)
15
Planar polymer membranes

• All of the work above was done with protein
  incorporation into polymer vesicle solutions and
  the results measured with bulk fluorescence or
  spectroscopy
• Although we can see that the protein can insert
  and function in the membranes, we still dont
  know if the membrane environment is having some
  effect on the protein
• Measurement at the single molecule level sheds
  some light on this
• Electrical transport measurements of OmpF and
  maltoporin inserted into planar polymer membranes
  show protein activity at the few molecule level
  (27 trimeric pores)
• Following protein insertion, membrane showed
  conductance decrease upon polymerization (B),
  then further decreases upon the addition of sugar
  (arrows)

Nardin Langmuir 2000, 16, 7708
16
Polymer membrane lifetime and single molecule
transport measurements

• Using a shorter version (5-6 nm) of Meiers
  PMOXA-PDMS-PMOXA polymer (9-31-9, previous
  15-68-15) we created freestanding membranes on
  conventional Teflon substrates as well as
  micromachined orifices in Si to measure membrane
  lifetime
• Average lifetime of polymer membranes is gt50
  greater than that of lipid
• Commonly exceeds 24 hours
• Obtained a 4 day polymer membrane on a 150um Si
  hole
• Resistance typically exceeds 100GW, and is over
  30x that of lipid membranes on average.
• Also probed the effects of the polymer
  environment on protein insertion and function

17
Single molecule measurements of a-hemolysin in
DPhPC polymer
Conductance 0.72 nS
Conductance 0.79 nS
18
Summary of our single molecule measurements

• Other proteins incorporated and measured at the
  single molecule level (for thin polymer- for
  thicker polymer, OmpG inserted, but not aHL!)
• OmpG (80 mV applied)
• MscL (16 mmHg applied)
• Alamethicin

19
Stabilizing membranes with a solid
surfaceTethered lipid bilayer membranes

• Can create these structures in two ways
• 1) Must covalently attach lipid to solid surface
  (silane or thiol SAMs)
• 2) Non-specifically absorb lipid onto surface
  through vesicle fusion
• These membranes generally show outstanding
  robustness and can withstand dehydration and
  rehydration, although it is unknown whether small
  defects develop (e.g., nS in conductance)
• Any protein incorporated into the tethered
  membranes must be spaced from the surface to
  avoid any deleterious interactions with it

20
Sensors using tethered BLMs on gold

• We cannot perform any DC measurements because the
  bottom surface, if conductive, is usually gold
  and therefore can only function capacitively
• First experiment of this kind was Cornell et al.
  Nature 387 580 (1997)
• Used gramicidin
• Dimeric ion channel, whose conductance would be
  disrupted when one half of it would be pulled
  away to bind to an analyte
• Looked at complex conductance as a function of
  time as analytes were introduced

21
Tethered BLMs on gold

• Using impedance spectra for capacitively probed
  membranes
• Complicated to interpret
• Need to model capacitance of electrode, double
  layer, and membrane as well as the resistance of
  the membrane, incorporated ion channels, and the
  surrounding solution
• Look at real and imaginary components of
  impedance as a function of frequency

22
Advances in tBLMs

• If the resistance of the tBLM is sufficiently
  large, there can be a large RC time constant for
  the ions in the double layer (between the
  membrane and the electrode) to deplete
• When this happens, pseudo-DC (.01 Hz or slower)
  measurements of ion channels in the membrane are
  possible
• As of yet, none of these resistances are high
  enough to show single channels, but patterning
  the surface to limit the membrane area can cut
  down on membrane resistance and there is a path
  to single channel current measurements
• This would be a significant advance as these
  membranes are typically stable, long-lived and
  the substrates are easily integrated into a
  device configuration
• Duran group reported these results at recent ACS
  meeting this week

23
Porous membrane supports using gels

• By surrounding a BLM with an agarose gel on one
  or both sides, mechanical or other interactions
  with the gel may alleviate various membrane
  failure modes
• Early attempts at gel supported membranes used
  standard techniques to paint membranes on a
  Teflon partition and then bring gels in contact
  with membrane on either side
• Gel allows mechanical support while allowing ions
  and other analytes to diffuse to and from the
  membrane

24
Gel supported membranes

• Ide and Yanagida formed bilayer membranes on
  agarose gels using applied pressure, but instead
  used the relaxation of a compressed material to
  apply negative pressure to the bottom of the
  membrane, causing the membrane to immediately
  thin out
• Membrane formed in lt 10s
• Measured a number of proteins at the single
  channel level

Ide and Ichikawa, Biosensors and Bioelectronics
21 (2005) 672
25
In situ gel-encapsulated membranes

• In recent work, we have created Mueller-Rudin
  DPhPC lipid membranes in the presence of a
  hydrogel precursor solution
• Polymerization of the gel solution encapsulates
  the membrane within it, forming a mold of the
  membrane in almost continuous contact with it

PEG-DMA (1 kDa)
26
In situ gel-encapsulated membranes

• Initial observations
• Gel polymerization also accelerated membrane
  thinning and resulted in a stable solvent annulus
  at the membrane periphery
• Encapsulated membranes have longer lifetimes, and
  enabled measurements of single channels for days

Jeon, Malmstadt, Schmidt, JACS, 128, 42 (2006)
27
In situ gel-encapsulated membranes

• Mechanical perturbation- shaking/hitting the air
  table

16
28
In situ gel-encapsulated membranes

• Mechanical perturbation- poking the gel

15
29
(No Transcript)
30
Mechanical perturbation

• Facilitating membrane formation by manipulating
  the gel

17
31
Susceptibility of membrane to pressure (1)

• Experiments
• 500 um hole, 200x microscope
• 7.5 (w/v) PEG-DMA hydrogels with 1 Irgacure,
  400W (5 min. polymerization)

water
oil
oil
exp1
exp2
exp3
gel
gel
gel gel
water
control
32
Susceptibility of membrane to pressure (2)

• Experiment 1
• 500 um hole, 200x microscope
• 1ml added at once ? membrane fails

3
33
Susceptibility of membrane to pressure (3)

• Contd
• 1ml added at once and then removed ? membrane
  recovers

4
34
Susceptibility of membrane to pressure (4)

• Contd
• 50 ul added at each point

Membrane failed at higher pressure
35
Susceptibility of membrane to pressure (5)

• Experiment 1
• 500 um hole, 200x microscope
• 7.5 (w/v) PEG-DMA hydrogels with 1 Irgacure,
  400W (5 min. polymerization)

oil
exp1
120 ul
gel
120 ul
120 ul
36
Susceptibility of membrane to pressure (6)

• Experiment 2
• 500 um hole, 200x microscope
• 7.5 (w/v) PEG-DMA hydrogels with 1 Irgacure,
  400W (5 min. polymerization)

water
exp2
gel
37
Susceptibility of membrane to pressure (7)

• Experiment 3
• 500 um hole, 200x microscope
• 7.5 (w/v) PEG-DMA hydrogels with 1 Irgacure,
  400W (5 min. polymerization)

120 ul
oil
exp3
120 ul
gel gel
120 ul
120 ul
120 ul
38
The gel traps solvent within the membrane (1)
exp

• 1 sec time lapse ? 30 frames x 13 sec ? 390 sec

gel
control
1sec time lapse
Real time
5
6
39
The gel traps solvent within the membrane (2)
7
40
Robustness to applied voltage

• Experiments
• 500 um hole, 200x microscope
• 1 sec step function (with 5 mV increments)

exp1
exp2
exp3

-

-
gel

-
gel
gel gel

-

-

-

-

-

-

-

-

-
control

-

-

-

-
41
Robustness to applied voltage(2)
Bigger annulus, broke at 245mV
Smaller annulus, broke at 215mV
9
8
42
Robustness to applied voltage(3)
0 500 mV (with 5mV increments, 1sec each)
10
43
Robustness to applied voltage(4)
0 500 mV (with 5mV increments)
Poking the gel after electro-compression
11
12
44
Robustness to applied voltage(5)
0 500 mV (with 5mV increments), broke at 215mV
13
45
Robustness to applied voltage(6)
0 500 mV (with 5mV increments), broke at 375mV
14
46
Possible slowing of DNA translocation by the
encapsulating gel

• 150 base pair single-stranded DNA was added atop
  the hydrogel.
• The hydrogel appears to significantly slow the
  DNA diffusion through the mesh to the nanopore.
• Blockades as slow as 1 ms/base were detected.

47
Planar lipid bilayer fabrication by solvent
extraction in a microfluidic channel

• Design criteria for an automated lipid bilayer
  fabrication device
• Simple no need for operator intervention or
  human monitoring
• Fast new membranes can be formed in a matter of
  minutes
• High-quality membranes gigaohm seals for ion
  channel research and applications
• Ability to measure single-molecules

48
PDMS solvent incompatibility
Cross-linked poly(dimethylsiloxane) (PDMS)
elastomer
After Lee et al., Anal. Chem. 75(23)6544-6554
(2003).
49
Membrane formation by solvent extraction
Principle of operation
50
Device design
51
Experimental apparatus
Applied voltage in
V
Amplifier
Measured current out
CCD
I
52
Fluid compositions

• Aqueous phase
• 1 M KCl
• 5 mM Hepes
• pH 7.0

• Organic phase
• Solvent composed of 11 n-decane squalene
• Lipid 0.025 (w/v) diphytanoylphosphatidylcholine
  (DPhPC)
• 50 ppm perfluorooctane

53
Lipid solution droplet formation
Lipid solution stream
100 µm
54
Solvent extraction
Lipid solution droplet
100 µm
5x replay speed
55
Membrane capacitance during solvent extraction
56
Observed membrane resistances of 50-100 G?
This membrane has a resistance of 91 G?
57
Insertion of a-hemolysin into a microfluidic
membrane
58
Design criteria for an automated lipid bilayer
fabrication device

• Simple no need for operator intervention or
  human monitoring
• Valves are computer controlled
• Fast new membranes can be formed in a matter of
  minutes
• True, but lifetime is limited 15 minutes for
  full integrated device, 45 minutes for PDMS
  solvent extraction only
• High-quality membranes gigaohm seals for single
  molecule ion channel research and applications
• Unique geometry results in minimal background
  capacitance, resulting in very low noise
  measurements

59
Future work Hydrogel encapsulation

• Optimize Organic phase
• Solvent composed of 11 n-decane squalene
• Lipid 0.025 (w/v) diphytanoylphosphatidylcholine
  (DPhPC)
• 50 ppm perfluorooctane
• Lipid concentration
• Solvent choice and concentrations
• Fluorocarbon

60
Future work Ion channel assay platform
61
Acknowledgements

• Schmidt Group
• Tae-Joon Jeon
• Noah Malmstadt
• Jason Poulos
• Robert Purnell
• Denise Wong
• Funding provided by DARPA and ACS-PRF