‘ iii ”i”,

 

iii

” ”U” ”iiiiii” ii

”” iii” ””i”ii”ii. ..:
ii”; ‘ iii“ “Gigi!” Jiiiiiiii””””i”i ' a: '
i ”iii:

I i
”ii” ii ii” i

i.“ in EU,
”NIH.” J
i
I i
i
”J
i

ii
i ”i ?
.ii‘
i i
ii

iii.
."iiiiiii “ ”i Jiiiii

”i” ii iii ,ii ii ii iii
‘ ‘iii’i Iii.” “W

”ii iii' “ii
”I” ””i ”ii ii ”' ”ii
'”ii 1”“ iii”i|”iii' [iii ”ii iii. ”H b“ Ji” (‘iii

iiii
:' ii W M” ii” M” i‘ ”i "ii ii” ”ii.” ”’i
i , jii ii i“ i li iiiwwt‘iiiiil‘. iiiifhii’iigii}

iii ii

 

.
-- J’
‘ a- v

' ’. ‘x .s..':~;’

1 .J—V- __’ A --¥.
.
.__.:
-w- -M:
‘9...—

i ‘ . I ,
i,i"' ‘1 | :L 1
i. -". ’ -‘-
,i‘. i_ ' [5." i
.il i i illii ‘ii‘ i 2:, g I, ’i'
i l I .. i ‘l‘l -.i 1,? hi} 1” _
J ” 'i 'i‘ i” i
' ” ‘ *x i
‘ ” .r. ‘ .. | ‘ ‘
. . ”I J f‘ul”: 3W "”6131; [)4 z. I. .
M “Li‘s '“l :11”? ‘_.‘i.i. 3'“. 4‘ '3
i” 4.‘ ‘ VI, Hi J'J‘ ;.l‘.- M V)" l ‘ .':
‘ J 1‘1“” all“ Ji :_’ ‘4_ ,( ‘1‘ ET\"A:I“‘ {'31)}, ,
I . ‘~' g I. > - - , ”L. ‘. I i, I , .
” | M! ”J” I, ”'3‘ "h.” J I .J 'LJ '4 i ‘ i” J '4,“ h J. ‘4 ‘
l. \ o . i \' ‘ ~ ' ,I u.
l» ' “if u i i ‘ i w "J '
I . J j 1‘ »’ I '
‘ J J 1 V i l b .
' v.
0 J ‘1

”W ”i 'i ” i '. ' - .' ~. ”iii
””iiiiiiiiii _, i ii
iiiii”i ”iii iii” “4 " " J

iiiii.I ' “ i
ii . i

i/‘flii : ‘iJi Ii ' imii . ” i i '2‘7 ” i”

"' ‘ i i

I .i‘l‘” ‘ I... ” v." Ji'i 'Ii”:- i'}

iii”: i i ‘ "Li "M .4 '..i i

”iii.” i. ”‘ .‘i i‘ " ._ . ”I

i i i i i -. a _
Hiil ’ i... i

A g T
A
‘5 4 Al ' ~ ; , . ; . .
i ‘ ‘ .'
. - 7 s
9- U.-.‘.mh_th J t
,1 3“? i
‘“ mwflx "y n... V

’4.‘ 4

This is to certify that the

thesis entitled

FORMATION CONSTANT AND PHOTOLABELING OF GRAMICIDIN
CHANNELS IN LIPID BILAYER MEMBRANES

presented by

Hsiao Yung Guh

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemistry

 

 

fl£fi

Major professor \

Date May 19, 198]

0-7 639

OVERDUE FINES:
25¢ per day per item

RETURNING LIBRARY MATERIALS:

______—__—-—-—-——-—-—-—

    
   

fl\\

“ ‘ ‘VA )(lfi‘l-II\\\\‘ 'ix

     

.” _‘ “w q Place in book return to remove
\‘x “‘3‘"?! . charge from circulation records

 

 

 

FORMATION CONSTANT AND PHOTOLABELING OF
GRAMICIDIN CHANNELS
IN LIPID BILAYER MEMBRANES

By

Hsiao Yung Guh

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Chemistry

1981

ABSTRACT

FORMATION CONSTANT AND PHOTOLABELING OF
GRAMICIDIN CHANNELS
IN LIPID BILAYER MEMBRANES
By

Hsiao Yung Guh

Gramicidin, a polypeptide of 15 alternating D and
L amino acids, has been shown to increase the cation
permeability of biological membranes. Substantial evidence
indicates that ion transport across the membrane is
mediated by gramicidin via the pore mechanism and that
the gramicidin channel consists of two molecules.

By use of the charge injection technique. we have
successfully obtained the formation constants of gramicidin
channels in lipid membranes as well as the partition
coefficients of gramicidin between the bilayer and torus
phases. It was found that both of the parameters depend
greatly on the membrane thickness. The dimerization
constant. for instance, increases from 1.9 x 10“ l-mole-1
for decane membranes (46.5A.in thickness) to 9.9 x 105
l-mole-1 for hexadecane membranes (32.0A in thickness). In

addition. a correlation was discovered between the dimeriza-

tion constant and partition coefficient between the membrane

and torus regions. It is of interest to note that although
relaxation experiments have been performed on the membranes.
the resulting physical parameters represent the equilibrium
values for the same parameters prior to the charge
perturbation.

In order to study the orientation of gramicidin
channels in lipid bilayers. 2-nitro-5-azidobenzoic acid
(NABA), a photolabeling reagent. was devised. In the
presence of this reagent, a remarkable blocking effect
on gramicidin-mediated ion transport was observed upon
illumination. The reduction in membrane conductance is
attributed to the steric hindrance imposed by photolabels
attached to the channel opening. Unfortunately. the chemical
analysis of photoaltered gramicidin failed to reveal the
exact location of the photolabeling. As a result, no
definite conclusion was made regarding the conformation
of gramicidin channels in lipid membranes.

Due to the limited capacity of the existing charge
injection instrument, a microprocessor-controlled charge
injection device was developed. It surpasses the old unit
in performance as well as in the system design. The new
device is able to inject the charge onto a capacitor in
less than 100 ns and monitor the resulting voltage
transient at a rate up to 15 MHz for 1K data points. The
signal-to—noise ratio is at least 256 for an input voltage

of 0.67 V registered at 15 MHz. The recording system

responds to a step voltage change of 0.5 V within #00

nanoseconds.

Acknowledgements

I would like to express my sincere thanks to
Professor Christie G. Enke for his guidance and encourage-
ment throughout this study. Thanks go to Professor Michael
Weaver for serving as the second reader.

Special thanks go to Dr. Tom Atkinson for his help.
Many thanks also go to Marty Raab and the membranes in the
Enke research group for their help and friendship.

Deep appreciation to my parents for their unfailing
support and encouragement throughout my education. Special

thanks to my wife. Min. for her love and patience.

ii

Chapter

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

I. INTRODUCTION ....

The Biological Membrane .....

Pore Transport System .........

Gramicidin Channels ............

II. CHARGE INJECTION INSTRUMENT

The Measurement Problem .....

Charge Injection Approach ...

Description of Microprocessor—

Controlled Charge Injection System .........

1.

Overview of the System

Electrochemical Cell

Pulse Generator and Cell

Amplifier ..

Digital Recorder ........

Microcomputer

Operation of the System

Performance

iii

Page

15

16
18

21
2h
26

28
3o
35
36
37

Chapter Page

III. FORMATION CONSTANT OF GRAMICIDIN

CHANNELS IN LIPID MEMBRANES .. ....... . ........ . #6
INTRODUCTION .......... ......... ............... 46
Current Relaxation Approach . ............... . 47
Analog Approach .. ..... . ................... .. 51
Voltage Relaxation Approach .. .............. . 52
EXPERIMENTAL .......... ............. ........... 55
Material and Apparatus .......... ...... ...... 55
Membrane Formation . ......................... 56
Membrane Data Collection . ............. . ..... 57
Data Treatment .. ............ .. ............. . 58
RESULTS AND DISCUSSION ............. ...... ..... 6O
Membrane Capacitance Measurement . ...... ..... 60

Voltage Relaxation across
Gramicidin—doped Membranes ............ ...... 65

Formation Constant of
GrmiCidin Chmels .00....IOOOOOIOOOOOOIOOOI 68

Partition Coefficient of Gramicidin
between the Membrane and Torus Phases ....... 82

Comments on the Assumptions . ............... . 85

Conclusion . ............. . ................... 86

IV. PHOTOLABELING OF GRAMICIDIN

CHANNELS IN LIPID MEMBRANES ............ . ...... 88
INTRODUCTION ' I o o ooooooo o oooooooooooooooooooo o o o 88
Photolabeling . ............................. . 89
Arylazides . .............................. ... 91

iv

Chapter Page

z-NitrO'B-aZidObel’lZOIC ACid o o I o o o o o o o o o o 0 o o o 93

EXPERIMENTAL .. .................... . ......... .. 94
Reagent ..... ................. ... ............ 9h
Apparatus .............. ....... . ....... . ..... 95
Procedure .. .......... . ..... . ............. ... 96

RESULTS AND DISCUSSION .. ...................... 99
Photolysis of NABA .... ...................... 99

Effect of Photolysis of NABA on
Gramicidin-mediated Ion Transport ........... 101

Chemical Characterization of Gramicidin
Isolated from Irradiated Liposomes . ......... 106

APPENDIX A Peripheral Schematic Diagrams ......... 118
APPENDIX B Program Listings . ........... ' ........ .. 121

REFERENCES ......................................... lhh

LIST OF TABLES

Table Page

3-1 The thickness and capacitance of
3% GMO-cholesterol/n-alkane
membranes (w/v) .. ..... .. .......... ........... 62

3-2 Relaxation time. membrane conductance
and concentration of gramicidin-doped
GMO-cholesterol/n-decane membranes
at 2A 1 1°C .................................. 70

3-3 Relaxation time. membrane conductance
and concentration of gramicidin-doped
GMO—cholesterol/nbtetradecane
membranes at 2h : 1°C ......... .......... ..... 71

3-4 Relaxation time. membrane conductance
and concentration of gramicidin-doped
GMO-cholesterol/n-hexadecane
membranes at 24‘: 1°C ........ ....... ......... 72

3-5 Dimerization constant. K. and partition
coeffieient. D. of gramicidin in
bilayer lipid membranes ...................... 76

3-6 Free energy of dimerization and partition
of gramicidin in alkane membranes .... ..... ... 84

4-1 Molecular weight and characteristic
mass peaks of identified eluents ............. 110

vi

LIST OF FIGURES

Figure

1-1

1-2

1-3

1-5
1-6

2-1

2-3 '

2.1.

Schematic representation of the Davson-
Danielli lipid membrane model. The polar

head groups of the lipids are included in

the aqueous phase. The shaded areas

represent membrane proteins. .. ....... ........

Simplified schemes for carrier and
channel transport .. ..........................

Stepwise changes in conductance for
gramicidin-doped membrane. The aqueous
phase is 1 M KCl (37) . .................... ...

Membrane treated as a series of n
activation energy barriers. The
energy of barrier n is fn ....................

Structure of gramicidin A ....................

Possible conformations of gramicidin
channels in lipid membranes (58)

Electrical representation of electrodes
placed in two aqueous compartments
separated by a membrane ......................

Membrane capacitance charging curve for
the voltage step and high voltage
pulse experiments ...... ......................

Block diagram of the old charge injection
instrument .... ........................ . ..... .

Block diagram of the microprocessor-
controlled charge injection device .. .........

vii

Page

10

13

17

20

23

Figure Page

2—5 Schematic of electrochemical cell ............ 27
2-6 Schematic of pulse generator and

cell amplifier ...... ....................... .. 29
2-7 Schematic of timing circuit .................. 32
2-8 Schematic of ADC section .. ................... 34

2-9 Response of the digital recorder to
a 0.5-V loo-KHz square wave. Data
collection rate: 15 MHz . ..................... 38

2-10 Signal-to-noise ratio of data
collected at 15 MHz .. ...................... .. 39

2-11 Response of the digital recorder
to a 0-1 v 2-KHz ramp wave .. ....... . ........ . 41

2-12 Voltage transient resulting from
a pulse application to a resistor
and capacitor in parallel . ................... 42

2-13 Voltage transient resulting from a pulse
application to resistor and capacitor in
parallel. Data collection rate: 5 MHz ...... .. 45

3-1 Voltage transients resulting from pulse
applications to 3% GMO-1% cholesterol/
n-tetradegane membranes in 1 M KCl
at 24 1 1 C .. .............................. .. 61

3-2 Voltage transients resulting from pulse
applications to gramicidin-do ed a
membranes in 0.5 M CaCl2 at 2 i 1 C ... ...... 64

3-3 Voltage relaxations resulting from pulse
applications to gramicidin-doped
membranes in 1 M KCl at 24‘: 1°C ..... ... ..... 66

viii

Figure . Page

3-4 The voltage relaxation time and membrane
capacitance of gramicidin-doped membranes
(3% GMO-1% cholesterol/n-hexadecane) as
a function of film age ....................... 69

3-5 Dimeric channel concentration, N , as a
function of the gramicidin concefitration.
B, in GMO-cholesterol/n-decane membranes. .... 73

3-6 Dimeric channel concentration. N . as a
function of the gramicidin concefitration.
B, in GMO-cholesterol/n-tetradecane
membranes .................................... 74

3-7 Dimeric channel concentration. N , as a
function of the gramicidin concefitration.
B. in GMO-cholesterol/n-hexadecane
membranes ......... ..... ........ ....... ....... 75

4-1 Reactions open to a nitrene after
its formation 0 ......... ......OOOIOOOOOOOOOIOO 92

4-2 The effect of photolysis at 320 nm on
the UV spectrum of NABA. NABA was dissolved
in 0.1 M phosphate buffer (pH 9.4) and
irradiated at 320 nm for 10 min. until no
further change in the spectrum ............... 100

4-3 The effect of photolysis on membrane
conductance. Gramicidin was added to the
membrane forming solution. Membrane otential:
50 mV; electrolyte: 0.1 M phosphate pr 7.6) . 102

4-4 The effect of photlysis on voltage
relaxation time resulting from pulse
applications to the membrane ..... ......... ... 104

4-5 Experiments that would result in substan-
tial amount of photolabeled gramicidin
for chemical characterization ........ ........ 107

4-6 Gas cheomatogram of derivatized hydrolysate
of gramicidin isolated from control liposome
using mass spectrometer as the detector ...... 108

ix

Figure

4—7

Gas chromatogram of derivatized hydroly-
sate of gramicidin isolated from
irradiated liposomes using mass

spectrometer as the detector .. ............

The mass spectrum of derivatized

tryptophan .......... ........ ...............

GC trace of the derivatives of tryptophan
which has been treated with trifluoro-
acetic acid-const. distilled HCl at

110 C for 1.5 hours ..........................
4K RAM board schematic diagram ..... .....
1K PROM board schematic diagram ..... ........
USART board schematic diagram . ........ ..

Page

109

111

113
118
119

120

CHAPTER I
INTRODUCTION

In the past twenty years, it has become plain that
the biological membrane plays a crucial role in almost all
cellular activities. Its ability to regulate the transport
of substances into and out of cellular organelles underlies
the basis for nerve conduction (1). muscle function (2).
energy assimilation (3) and conduction of metabolites
through cell walls (4, 5). Consequently, a thorough study of
transport system across the membrane will contribute to the
understanding of these physiological phenomena. In this
research. gramicidin channel. an excellent model for the
study of pore transport mechanism in lipid bilayers. was

photolabeled and its formation constant determined.

The Biological Membrane

The first clue that the biological membrane
contained lipids came in 1899 from Overton's observation
(6) that cellular membranes were permeable to lipids and
lipidlike substances. Since then. numerous efforts (7-9)
have been made to elucidate the structure of biological
membranes. It is now generally accepted (10) that the
biological membrane (Fig. 1-1) consists of two back-to-back

1

32-
g;—

1

//////////:/

 

 

 

 

 

g2???

 

\\\

 

 

\

///////

T

 

 

 

 

N\

 

“gags;

”23"?

Fig. 1-1 Schematic representation of the Davson-Danielli
lipid membrane model. The polar had groups of the lipids
are included in the aqueous phase. The shaded areas
represent membrane proteins.

3

layers of lipid molecules with the polar head groups
pointing to the aqueous environment and the hydrocarbon
tails buried inside the interior of the membrane phase.
Proteins and other membrane constituents either reside on
the surface of the lipid bilayers or penetrate deeply into
the membrane interior. The bilayer model successfully
accounts for the amphipathic nature of lipid molecules and
the ultrathin membrane structure (less than 100A) as well
as for the versatile biological functions manifested by the
, associated proteins.

Experimentally. there are two approaches in the
studies of biological membranes. One involves the utiliza-
tion of spherical lipid vesicles. the so called liposomes
(11), which can be prepared by sonicating lipids in aqueous
solutions. The liposomes resemble the cellular membranes in
that they have a curved closed form. a large surface to
volume ratio and solvent-free lipid bilayers. The small
physical size of liposomes (300-1000A in diameter). how-
ever. presents obstacles to investigations where probes are
required to be placed on both sides of the membrane.

This difficulty was circumvented in 1962 by Mueller
et al. (12). when they successfully formed a planar bimole-
cular lipid membrane (BLM) by brushing a complex mixture of
brain lipids. n-tetradecane. silicone fluid and mineral oil
across a narrow orifice between two aqueous compartments.
The planar geometry of BLM allows easy application of

perturbations to and direct measurement of responses of

L.

the membrane. As a result. a wealth of information regar-
ding membrane permeability (13), membrane photochemicsty
(14, 15). membrane transport mechanism (16-28) etc., has

been learned in the past two decades.

Pore Transport System

 

Due to the nonpolar nature of the hydrocarbon
chains of lipid molecules, plain lipid membranes are not
permeable to metal ions and exhibit an electrical resistance
as high as 1o“- 1010 ohm-cmz.

Upon the addition of some chemical substances. the
resistance of lipid bilayers. however. can be reduced by
several orders of magnitude. Depending on the way they
facilitate the transport of ions across the membrane.
these chemical modifiers can be classified into two
categories: carriers (29-31) and pore formers (32-37).

In the carrier mechanism. as depicted in Fig. 1-2.
the ions form complexes with carrier molecules and the
subsequent diffusion of the complexes across the membrane
translocates the ions from one side of the bilayer to the
other side of the bilayer.

In the pore mechanism (Fig. 1-2). as the name
implies. the chemical modifiers form transmembrane
channels through which the ions diffuse across the

membrane.

The ion translocation mediated by transmembrane

 

.puommcmuu
Hmscmno 02m smwhumo pom mmsmsom omflwfiHQSwm NaH .wwm

5:3 .832".
u
_ .
..III. r._ + U w...
2 GT T..Tu.0
v u

 

 

 

 

 

6

channels displays several interesting features not shared
by the carrier system. It is observed that membranes doped
with pore formers have very high ion transport rate
(4 x 10.11:?"1 for a single gramicidin channel in 1 M
potassium chloride at 25°C) (37) and the membrane conduc-
tance is essentially independent of membrane thickness (38).
Furthermore. the high conductance of a pore-doped bilayer
membrane persists when the lipid film is solidified by
lowering the temperature (39). The most striking aspect of
the channel transport system is. however. the finding that
when the concentration of pore formers in the bilayers is
very low. the membrane conductance fluctuates in a stepwise
fashion with definite heights (Fig. 1-3) (37). The
commencement and termination of a transition corresponds to
the opening and closing of a single conducting channel.

Over the years. a number of models have been
proposed for the mechanism of pore transport system (21-
28). Although they differ in sophistication. all the
postulations are based on Eyring rate theory (40). in which
the transmembrane channel has been considered as a sequence
of energy barriers over which the ion has to jump
(Fig. 1-4). The kinetics of pore mediated transport are
governed by

(1) the number and the heights of potential

barriers in the channel and
(2) the simulataneous occupancy of various ion

binding sites.

mg n o .Ammv aux 2H ma own a .
c on cflnwoflsmnh sow mosmposocoo CH mmmcmmwowmwmmwmw Mcmscems
. . -H .mfla

-ovaH v: mo. .

 

.-Nu .-

 

 

 

 

 

 

Membrane
Interior

Fig 1-4 Membrane treated as a series of n activation
energy barriers. The energy of barrier n is fn'

The potential difference between the two aqueous phase
*
is f .

9

Recently. Haegglund et al. (28) put forth a three-
site-four-barrier model. which successfully accounts for
the observed concentration-dependences of flux-ratio,
conductance, I-V characteristic and permeability displayed
by gramicidin-doped membranes. As the membrane conductance
induced by gramicidin is voltage-independent. it is obvious
that a more sophisticated and elaborate model is needed
before voltage-dependent conductance manifested by certain

pore formers could be satisfactorily rationalized.

Gramicidin Channels

Of the various pore formers known today. gramicidin
system is probably one of the most investigated because of
(1) its commercial availability.
(2) structural simplicity (both primary structure
and channel structure) and
(3) resemblance to physiological channels (Na+ and
K+ channels in nerve. for example).
Historically. gramicidin was first isolated from
Bacillus brevis by Hotchkiss and Dubos in 1941 (41). In
1965. Sarges and Witkop reported that gramicidin is a
linear polypeptide. consisting of 15 alternating D and L
amino acids (Fig. 1-5) with N-terminus blocked by a formyl
group and C-terminus derivatized to an amino alcohol (42).
It is noted that none of gramicidin's amino acid

constituents is polar or ionic. As a result, gramicidin is

10

.Aessv 6 one Aesv m .Aemmv a caeaessmsm so
mhsest a ma swcwoflsmnw manwawm>m hHHdfiohmasoo .zam>wpoommoh .Ahpnq
new ozmnq an umomammu ma «a sowpfimom CH Shaun easy Cw < swcwowamuw
Souk pmmmwu u can m sacfloflsmuo .¢ :chOHsmpw Ho opaposppm mud .mwm

235.255

=s~=s~=s=z-E-T ...r. a
.
.....sl..E......_§L-E.... 1.3-75
.
w
_sL-_s..._..sTTstauszi-5....TT8:

11

10 M

highly hydrophobic with a solubility of less than 10-
in aqueous solutions.

In 1972. through the measurement of biionic
potentials and transference numbers of various electrolytic
solutions partitioned by gramicidin-doped membranes, Haydon
and Mayers (43) found that the gramicidin channel is not
permeable to anions. is blocked by divalent cations and
is selective for univalent cations with the following
sequence. H+>NHu+ >cS"'> Rb+>K+> Na" >I.i+ >( CH3) 4N+.

Substantial evidence suggests that gramicidin
channels in lipid bilayers are dimeric in nature. Tosteson
(44) and Goodall (45) observed that the membrane conduc-
tance increases with the square of the gramicidin concentra-
tion in the aqueous phase. In 1973, by means of voltage-
jump experiments. Bamberg and Lauger (46) acquired results
which are consistent with the hypothesis of an equilibrium
in the membrane between a non-conducting monomer and a
conducting dimer of gramicidin A. In 1975, based on the
simultaneous measurements of membrane fluorescence and
conductance. Veatch et al. (47) reported that dancyl
gramicidin C. a highly active analog of gramicidin. forms
a dimeric channel in lipid membranes. Similar results have
also been obtained from an autocorrelation analysis of the
conductance fluctuations (48). Recently. X-ray diffraction
studies on a cation-containing gramicidin structure (49)
shows a 26A-length to the pore and a 3.4A-length to the

radius. which further supports the dimer concept of the

12

channel on the basis of the molecular size of monomeric
gramicidin.

In the past decade. several structures have been
proposed for the conformation of gramicidin channels in
lipid membranes. The salient features of these conforma-
tions are depicted in schematic form in Fig. 1-6. Model A
is the N-terminal-to-N-terminal single stranded helical
dimer proposed by Urry et al. (50). Model B. a C-terminal-
to-C-terminal helix. was considered by Bradley et al. (51),
when they discovered that N-succinyldeformyl gramicidin
methyl ester is able to induce membrane conductance.

Model C is the antiparallel double helix proposed by
Veatch et al. (52-54) as one of the dimer conformations
found for gramicidin in nonpolar organic solvents. Model D
is a parallel double helix with both N-termini at one end
of the channel.

By chemically modifying gramicidin and subsequently
examining their ability in inducing membrane conductance
(51, 55-57). it was concluded that the double-stranded
helics (model C and D) could not adequately represent the
conformation of gramicidin channels in lipid membranes.
More recently, Veatch et al. (58) incorported 13C and 19F
nuclei at both N and C termini of gramicidin and carried
out NMR experiments to determine the relative accessibility
of these 13C and 191" nuclei to paramegnetic probes in the
aqueous solution. They found that the chemical shifts and

spin lattice relaxation rates of C-terminal labels are

13

 

Fig. 1-6 Possible conformations of gramicidin channels
in lipid membranes (58). '
A. N-terminal-to-N-terminal single-stranded helix
B. C-terminal-to-C-terminal single-stranded helix
C. antiparallel double-stranded helix
D. parallel double-stranded helix

14

affected much more than those of N-terminal labels by the
aqueous paramagnetic ions. It is thus believed that the
most probable channel structure of gramicidin in membranes
is model A.

The experiments mentioned above. however. suffer
from one serious drawback. i.e.. they all utilized
chemically altered gramicidins. It is noted (51) that the
polarity and physical size of chemical groups derivatized
to gramicidin greatly affect the way the channel is formed.
Consequently. in order to obtain unambiguous information
regarding the channel conformation. some new techniques
have to be developed which would use native gramicidin as

the major investigating tool.

CHAPTER II

CHARGE INJECTION INSTRUMENT

An often-used technique to study ion transport
across a lipid membrane is to make steady-state conductance
measurements on a given membrane and transport system for
a family of ions. However. in order to evaluate many of
the kinetic parameters. it is necessary to perform measure-
ments on a short time scale after a certain perturbation.

In 1977, a computer-controlled data acquisition
system was developed in this laboratory (77). The device
is able to monitor the voltage transient resulting from a
pulse application to a membrane at a rate up to 10 voltage
readings per microsecond. It was demonstrated that rate
constants regarding monactin-and dinactin-mediated ion
transport could be obtained from charge injection experi-
ments by use of this instrument (78).

Recently. we have built a microprocessor-controlled
charge injection device. It will be shown in this chapter
that the new system surpasses the old device in performance
as well as in the system design.

This chapter is organized into two sections. The
first part includes a brief discussion of the measurement
problems encountered in membrane kinetics and the

advantages of the charge injection technique. The second

15

16

section is a description of the newly constructed system

and its performance.

The Measurement Problems

One approach to study the kinetics of fast reactions.
is to apply a sudden perturbation to the system of interest
and mointor the resulting relaxation. The most widely used
method to date in the study of membrane kinetics has been
the voltage step experiment. This technique has not been
totally successful in measuring very fast parameters
however. due to the membrane capacitance. The experimental
set up for an electrical measurement on a membrane can.be
represented by Figure 2-1; where :

CM = membrane capacitance.
G(V) = membrane conductance (function of voltage).

C = membrane-solution double layer capacitance.

I
RS = solution resistance.
CE = electrode-solution double layer capacitance. and
RF = Faradaic resistance at the electrode.

Several disadvantages of the voltage step technique
can be recognized from Figure 2-1. First, due to the fact
that the membrane must be charged. the voltage perturbation
is not a clean transition but rather a growing exponential
curve whose time constant depends on the resistances and -
capacitances mentioned above. This slow growth of the

forcing function completely obscures any information in

17

 

_L
A:
O
T—
a
w
.9."

 

 

G(V,T)

—----—--- a
an

?18- 2-1 Electrical representation of electrodes placed
in two aqueous compartments separated by a membrane.

18

the early part of the relaxation. Second. in a voltage
jump experiment the total current passing through the cell
is measured instead of that due only to G(V.t). As a
result. the current going to charge the membrane
capacitance further obsures the relaxation. Third. because
of the solution and Faradaic resistances the membrane
potential is always less than the externally applied
voltage.

Charge Injection Approach

The membrane capacitance problem has been avoided
in the charge injection technique (76). This method uses
the membrane capacitance to advantage by charging the
membrane very rapidly and monitoring the voltage decay as
the membrane's conductance (ion transport) depletes the
charge stored on the membrane capacitance.

The advantages that accrue to this method are the
following :

1. The entire conductance-voltage relationship can
be deduced from data obtained by a single pulse
on a single membrane and in a time short enough
to minimize changes in membrane area and
composition.

2. No current exists in the solution or external
circuitry during the measurement. Thus. the

solution resistance between the membrane and

19

the electrodes does not introduce an extraneous
voltage drop.

3. The time course of the voltage decay depends

on the ratio of the membrane capacitance to the
conductance and therefore is independent of
membrane area.

4. The voltage at any time after injection yields

the amount of charge which has crossed the
. membrane since application of the charge pulse.

5. The fast charge injection allows measurements

to be made on the system within a few hundred
nanoseconds.

The primary disadvantage of the technique is that
the relaxation occurs over a broad potential range and the
voltage dependencies as well as the time dependencies
must be determined. Furthermore. small amplitude relaxation
will be difficult to observe and analyze.

Traditionally, there are two ways to achieve a
rapid injection of charge onto a membrane. Both approaches
utilize a pulse generator with a high output voltage
compliance and rapid rise time. Since the pulse duration is
very short. a high voltage level that would normally break
the membrane can be successfully employed. When the pulse
generator is operated in a current mode. the time necessary
to charge the membrane to a given voltage would be given by
t = VC/i. where V is the desired voltage. C is the membrane

capacitance. and i is the amplitude of the current pulse.

20

Voltage Pulse

.. ..-—1— _______

Voltage Step

 

 

 

 

 

 

TIME

Figure 2-2. Membrane capacitance charging curves
for the voltage step and high voltage
pulse experiments. V is the voltage on
the membrane capacit ce.

21

As is evident from Figure 2-2. although the capacitance
must be charged through the solution resistance. a voltage
pulse of large amplitude can also bring the membrane
potential to a desired value in much less than RSCM seconds.
In practice this results in an improvement of the charging
time from tens of microseconds for the voltage step method.
to a hundred nanoseconds for the charge injection
technique using a voltage pulse.

Recently. due to the advance in electronic techno-
logy. a third approach has been developed. In this method.
a predetermined amount of charge is stored on a capacitor
and then dumped onto the membrane capacitance by means of an
ultrafast operational amplifier. The speed of the charging
action depends greatly on the settling time of the output of
the amplifier. With a state-of-the-art fast operational
amplifier. it is now feasible to accomplish a charge injec-
tion in a few hundred nanoseconds. The advantages of this
method over the use of a pulse generator are the low cost
and structural simplicity of the charging device and the

precision of the charge injection.

Description of Micr0pgoceeeor-Controlled Charge Injection
m

The first computer-controlled charge injection
instrument designed for membrane kinetics studies was

developed in this laboratory in 1977 (77, 78). The block

22

,diagram of the instrument is shown in Figure 2-3. It
consists of four major parts: a minicomputer. a pulse
generator. a cell amplifier and a Reticon transient recorder.
Two advantages of this system can be recognized. First. the
employment of a minicomputer minimizes the human errors and
the time to perform experiments. data acquisition and data
analysis. Second. the rapid injection of charge onto the
membrane and the ability of the Reticon recorder to collect
data at 10 MHz makes possible the study of fast kinetics

at microsecond level. However. a system that requires the
constant attention of a minicomputer is rather expensive.
In addition. the number of data points available at the
fast collection rate (0.1 to 10 MHz) is limited to 100.

In the past few years. due to the advance in micro-
electronics. microcomputers have become very cost-effective
and are capable of performing most of theitasks previously
assigned to a minicomputer. Furthermore. high speed
electronic components such as operational amplifiers.
analog-to-digital converters (ADC) and random access memory
(RAM) are also available at moderate prices. It thus
appears that a significant improvement in the design and
cost/performance ratio of the charge injection system could
be accomplished by use of these recently developed
electronic devices.

In this section, the details of the newly
constructed charge injection device, its operation and

performance will be discussed.

23

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

’ cru L
[true-11)) ' “”0
l7“
8/"
. t l '
1 Timing _ Reticon 5
- * J recorder

     
 

 

 

 

“gm“ Cell Hmplitierj

 

Figrue 2-3 Block diagram of the old charge injection

instrument.

24

1. Overview of the System

Figure 2-4 shows the block diagram of the newly
constructed system. It consists of a microcomputer. a pulse
generator, a cell amplifier. an electrochemical cell and a
fast digital recorder. Several features can be recognized
in this system. First, the control of the experiment and
the data acquisition is accomplished by a microcomputer.
This alleviates the working load of a minicomputer and
allows it to perform more significant tasks such as data
analysis. Second. a digital recorder capable of operation
at 15 MHz with 1K memory has been implemented. It surpasses
the Reticon analog recorder in speed. memory size. range
of data collection rate and circuit design. Third. the
charge pulse generator is cemposed of an ultrafast opera-
tional amplifier. a capacitor and a set of switches. This
simple device replaces the conventional pulse generator
which costs $1400. Fourth. the whole system (including the
microcomputer) is built on three circuit boards. The
simplicity of the system design makes the construction
and debugging relatively easy and reduces the overall cost
to about $1400. The only expensive component in the instru-
ment is the analog-to-digital converter used in the digital
recorder. It is an 8-bit device (0.39% resolution) and

cost $800.

25

 

 

.oow>on sowpoowsw omhmso

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

aria.

 

 

j :53 8. ha“.
:8
m, ........ s. new .......
.dflflwu lfluzaz M “nun“ uflfimfifl
_ z : a : s .
25 5.231 (V
l l c. w : : l
Es as... :3. T338 flu.

ooaaospsoouuommocosmohows one 90 amuwmwc xooam .sum opswflm

L--“------------------O-- J

26

2. Electrochemical Cell

The schematic of the electrochemical cell is shown

in Figure 2-5. It consists of two Teflon compartments.
Each has a volume of 45 ml separated by a Teflon sheet
(0.25 mm in thickness) with an orifice (1-2 mm in diameter)
in the center. Two Viton "0" rings. each situated in a
circular trough on the Teflon compartment. are used to
ensure tightness when the cell parts are clampped together.

On the face opposite to the O-ring trough side of
the compartment there is 1-inch circular hole. The mounting
of a glass window of the same size allows the visual
examination of the formed lipid membrane through a micro-
scope. To prevent solution leakage. a Teflon tape (0.5 inch
in width) of proper length is employed to achieve tight fit
of the glass window to the hole.

Two Ag/AgCl electrodes are shaped into coils and
mounted on a piece of Teflon plate. The length and separa-
tion of the electrodes is such that when the plate is
placed on top of the cell assembly. each electrode dips
into the solution in a different compartment. The
electrodes are made of silver wires coated with AgCl and
have a surface area of about 3.5 cm2. Since the AgCl on
the cathode is consumed in the experiments. the coating
must occasionally be reinforced by anodizing the electrodes
in NaCl with a Pt counter electrode. It should be noted
that each pulsing study uses approximately 5 nanocoulombs

and therefore the coating is not depleted rapidly.

 

 

 

 

 

 

 

 

 

 

 

 

 

28

3. Pulse Generator and Cell Amplifiep

The pulse generator and cell amplifier are shown
in schematic form in Figure 2-6. The pulse generator and
the amplifier are placed as close to the electrodes as
possible in order to minimize capacitance and stray signal
pickup in the connecting wires. The components are housed
in a small aluminum box which rides "piggy back" on the
electrochemical cell. The electrodes are attached through
a 2 prong Cinch-Jones connector on the bottom of the
housing.

The pulse generator is a digitally controlled
source of charge pulses. It consists of a set of switches.
a charging capacitor (capacitor D) and a charging amplifier
(amplifier 1). The input voltage in Figure 2—6 comes from
the microcomputer interface and is used to charge up
capacitor D. The switches A: B and C are directly controlled
by CPU through a buffer register. Normally. switch A and B
are open and switch C is closed to prevent the charging
amplifier from oscillating. When switch A is closed. capa-
citor D is charged up to Vin with a charge of CDVin C where
CD is the capacitance of capacitor D. Upon opening switches
A and C, and closing B. the charge stored on capacitor D is
injected by the action of charging amplifier onto the
membrane which is situated between the two electrodes.

The voltage transient produced is amplified by the LH0032
PET input amplifier (gain of X2) to produce a voltage

excusion in the range of 0-1 volt at the output. Theoutput

29

00¢.3

.Auo>flhvvoommnm

use Acoufizmvooomnm . .o.m.< xofimsq « m amnoomq « m ammofi . H
.powwflamem Haoo was nonspocow omasm ho owpmeonom mum .wam

2.! as?!

\

 

.P—_>/

 

 

 

8— 2p

 

)\ J\/\

 

30

signal is then buffered by a LM310 cable driver and sent via
coaxial cable (terminated with 75 ohms) to the digital
recorder. It should be noted that because electrode E is at
virtual ground. a.negative Vin will result in a positive
voltage at electrode F.

To satisfy the "high speed" requirement of the
pulse generator. a Teledyne Philbrick 1025 BET input
operational amplifier was used for the charging amplifier.
It has a high input impedance of 1011 ohms and a very short
settling time of 55 nanoseconds (to 1%). The switches are
Signetics SD5000 (switch) and SD5200 (driver) with an on

10ohms.

resistance of 30 ohms and an off resistance of 10
The on and off times of the switch are both 5 nanoseconds.
Power (115, +5 volts). common. and control signals
from the microcomputer are connected to the pulse generator
and cell amplifier by 6 and 4 prong Cinch-Jones connectors
respectively located on the side of the housing. The
power supply lines are decoupled to common with 6 uF
tantalum capacitors at the input. and with smaller ceramic

capacitors near the amplifiers. to allow good transient

response.

4. Digital Recorder

The digital recorder implemented in the present
system consists of a timing circuit and an analog-to-
digital converter with associated memory and address

counters. The timing circuit provides necessary signals

31

for the operation of the ADC section. The ADC section
converts the incoming analog signal (voltage) into digital
form and stores it in the RAM. The device is a general data
acquisition system and can be adapted to any instrument
provided proper buffer circuit is available between the

unit and the instrument.

a. Timing Circuit

The timing circuit provides the conversion signal
for the ADC, the "write enable" for the RAM and the clock
pulse for the address counter. The system is shown in
schematic form in Figure 2-7. The operation of the circuit
is as follows. The crystal controlled oscillator frequency
(30 MHz) is divided by a set of flip flops to generate
5 different clock frequencies that determine the data
collection rate (15, 10, 5, 2. 1 MHz). Due to the fact that
the digital output of the ADC is delayed by one and half
clock cycles. three flip flops (FBI. 2. 3) are employed to
ensure pr0per timing among the signals for ADC, RAM and
counter. Prior to the start of the experiment. the data
acquisition rate is loaded into a register and a "clear"
pulse issued by the microprocessor to reset the address
counter and flip flops. When a start signal is received
by the timing circuit. output Q of FF4 goes to "1" which
places the CPU on hold. On sensing "hold acknowledge"
from the microprocessor. the pulse generator is actuated

(output of gate 4 goes to 1), and 1K RAM is selected,

32

.pwsohwo media .Ho owpmsozom .mrm opsmwm

LL ...-E

 

 

 

 

 

.‘o 02.5
. 3<¢

 

7w.»

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

33

the ADC starts data conversion. and output Q of FF1 goes

to "1". On the next rising clock edge (out of gate 1),

gates 2 and 3 are enabled by FF2 and FF3. respectively.

As a result, the first "write enable" pulse for RAM occurs
one and halchlock cycles after the start of data conversion
and the address counter is updated after the storage of

the first data point. When the counter reaches 1024 (1K).

it issues a "stop" signal which disables the clock circuit
from sending further pulses to ADC section while the

microprocessor resumes the control of the system bus.

b. ADC Sectioh

The schematic of the ADC section is shown in
Figure 2-8. The system consists of an analog-to-digital
converter board. a 1K byte random access memory (RAM) and
an address counter.

The ADC board (TRW TDC1007PCB). obtained from TRW
LSI Products. has three basic units; an 8-bit ADC (TPC
1007J). a set of voltage regulators and a buffer circuit.
The board is mounted on a 8" x 10" board developed in this
laboratory which contains the rest of the components used
in this section. In its pesent configuration. it digitizes
a 1 volt peak-to-peak. 75 ohm signal with a maximum
conversion rate of 15 MHz. The RAM contains 8 Fairchild
93425 RAM IC's each organized as 1024 words by one bit with
a write time of 45 nanoseconds. The address lines of the RAM

are connected to the system bus and a counter through

34

was
621m

.COfipoom

00¢ mo owpmSonom .wum opsmwm

 

 

 

 

 

 

 

 

 

 

 

BDdLOOLOOJ.) 2°
oov g

 

l

:>

 

 

 

:40

..o
.230 ..E.

v30

 

 

 

(/ 2m

 

 

 

35

tristate buffers. Due to the tristate nature of the RAM
output, the data lines are directly coupled to the system
bus. Briefly. the system operates as follows. Prior to the
experiment. the counter is reset to zero. During the experi-
ment. the clock circuit provides the conversion pulse to the
ADC, the clock pulse to the address counter and the "write"
pulse to the RAM to record the digitized voltage. After 1K
data points have been recorded. the counter issues a flag to
the microprocessor to indicate the end of data collection.

The microprocessor then resumes control of the system.

5. Microcompute;

Of the various microcomputers. the Intel SDK-85
system was chosen for the control of experiments and data
collection because of its high performance (1.3 micro-
second instruction cycle). popular instruction set and low
cost. The system consists of an 8085A CPU, 2 kilobyte
read-only-memory (ROM). 256 byte RAM, 38 parallel. I/O
ports, a serial port and an interactive keyboard and LED
display. The system also has an excellent monitor program
residing in the 2K ROM. It allows the operator to examine
and load the registers and memory as well as to run and
single-step user's program.

However. the memory capacity of SDK-85 is rather
limited. To circumvent this problem, an additional 4K RAM
and a.1K PROM (programmable ROM) were interfaced to the

system. Also added to the system was a USART (Universal

36

Synchronous/Asynchrous Receiver Transmitter) to facilitate
the communication between a minicomputer and the micro-
processor. The schematic diagrams of these peripherals.
which were designed by Bruce Newcome in our group. are

shown in Appendix A.

6. The Operation of the Syetem

To operate the microprocessor-controlled charge
injection device. four programs are required. They are
DOWNLD.FTN. DATAIN.FTN, LOAD.MAC, and MEMB.MAC. A listing
of these programs is shown in Appendix B.

DOWNLD.FTN and DATAIN.FTN reside in the minicomputer
(PDP-11). The former is used to send object files from
the minicomputer to the microsystem through a serial link.
For the present system. MEMB.MAC in binary codes is
downloaded to the microcomputer when the charge injection
device is powered up. DATAIN.FTN receives data from the
microcomputer. It can then either plot the data on a
graphic terminal or store them in a file on any mass
storage device.

LOAD.MAC and MEMB.MAC reside in the microcomputer.
LOAD.MAC is situated in PROM and is available when power
is on. It receives files from the minicomputer and stores
them in RAM. MEMB.MAC controls experiments. data collection
and data transfer from the device to the minicomputer.

The data collection rate and charging voltage for the

charging capacitor in pulse generator are entered through

37

the keyboard to memory locations ZOOOH and 2001K
(hexidecimal). respectively. When the pulse experiment is
over. MEMB.MAC transfers data immediately to the minicom-
puter. This is done because at the moment the microprocessor
lacks graphic terminals and mass storage devices. The
examination of data can only be done on the minicomputer.
A program named EXPFIT.FTN can be used to
exponentially fit the acquired data. To use EXPFIT.FTN.
a data file name and two integers are given to the
program. The integers specify the first and the last data
points of a data block in the data file. EXPFIT.FTN will
linearize the exponential data and calls a linear square
fitting subroutine. The program will then return the

fitted parameters.

. P o anc

The performance of the system is illustrated in the
next five figures. Figures 2-9 to 2-11 demonstrate the rise
time. signal-to-noise ratio. linearity and accuracy of the
digital recorder. Figures 2-12 to'2-13 display the rise
time and accuracy of the pulse generator.

Figure 2-9 shows the response of the digital
recorder to a loo-KHz. 0.5-V square wave. It can be seen
from the figure that the system responds to voltage steps
within 3 data points and that it takes 3 to 5 data points
for the system to settle down to a constant voltage. Based

on the data collection rate (15 MHz). the rise and settling

38

ON

T

. . . .sxs ms
.mpmp sowpooaaoo meme .o>m3 msmsvm umxrooH

>rm.o m on soumooou Hmpwmwv on» mo oncoammm .mam onzmwm

nommomoiv 92:.

_ . _ _ _ _ . _ H 80.0
(1 d d a . . . H #

 

 

 

.r 83

(A) HOVI'I O/\

39

.szz ma pm umpomaaoo apmn mo owpma mmwocnOPercmwm .oHum opswflm

ammomoic mic.
on o

~0-
.1:—

000.0

4&-
~1-
«F
4:-
C1-
.1-

(A) BOVJJ O/\

 

 

Lfi 095.0

40

time of the recorder are approximately 0.2 and 0.3
microsecond. respectively.

The response of the digital recorder to a constant
voltage is shown in Figure 2-10. Apparently. only one value
for the voltage was recorded by the system. For an 8-bit
analog-to-digital converter. the resolution is 0.39% which
amounts to a quantization error of 3.9 mV for a 1-volt
signal. It is therefore concluded that the signal-to-noise
ratio is at least 256 and the actual fluctuations in
output voltage are less than 3.9 mV.

Figure 2-11 shows the response of the recording
system to a 0-1 V ramp wave at 2 KHz. The figure displays
’the rising half of the signal. Since a data collection
rate of 2 MHz was used. each data point corresponds to a
2 mV resolution in Vin' As a result. a stepwise curve for
V vs Vin is observed in Figure 2-11. It is noticed

out

that the slope of the curve is one and V equals Vin

out
within quantization error. These two facts demonstrate

that the response of the system is linear and that the
device has an accuracy of 3.9 mV. It is also observed in
Figure 2-11 that two small groups of data points do not fall
on the curve. These data points are attributed to instru-
mental errors which were not observed when a constant
voltage was applied to the recorder (Figure 2-10). As the
difference between the true value and the noise is much

greater than that between the two adjacent data points. the

noise can be eliminated from the data set by data treatment

41

.Nmz m .mpmp
cowpooaaoo mama .csonm ma o>m3 ecu mo mam: mcflmfia

mafimwp on» haso .uocnooou on» mo pampso one ma uzo>
ocm Hmcwwm on» mo oprHo> one ma CH> .o>m3 mews

u:xrm >firo m on povuoooa Hmpwwfic one 00 mwconmom .Hfirm mpsmflm

8 5>

all-
~1-
db
all.
«I-

I.
q-
d

d-

00.9 00.0
p1

00.0

 

w ‘”°A

42

 

 

 

 

UJ
2
S
O
> an-
0000 ‘r : r t r r I. : : ll
0 70
TIME (MICROSECJ
1.000
$ 8
UJ
2
5
O
> .
0.000 + r : . r : A 1 : 1'

 

70
TIME (MICROSECJ

Figure 2-12. Voltage transient resulting from a pulse
application to a resistor and capacitor
in parallel. Data were collected at 15 MHz.
A: transient, B: transient plus its
exponential fit.

43

. using software techniques.

To characterize the pulse generator. a predeter-
mined amount of charge is applied to a resistor (10.0 1
0.1% kn) and a capacitor (5000 1,1% pF) in parallel (dummy
cell) and the resulting voltage transient is monitored by
the system. Figure 2-12A shows the voltage transient
collected at 15 MHz. Its exponential fit plus the transient
are illustrated in Figure 2-12B. Two parameters were
obtained from fitting the curve: the initial voltage. Ve.
and the time constant. 7. of the transient. The initial
voltage can be used to calculate the amount of charge

injected onto the dummy cell according to equation (1)3
QO=CVO/2 00000000000000.0000. (1)

where: Q0

the amount of charge received by the
dummy cell. i

C : capacitance of the dummy cell.

Ve : initial voltage of the transient
obtained by exponential fit.

2 a voltage gain of the cell amplifier.

For this experiment. v. equals 0.809 V and Qe is
2.02 X 10"9 coul.. The amount of charge issued by the
pulse generator is equivalent to the product of the
capacitance of the charging capacitor (470 1 10% pF) and
the charging voltage (4.45 1 0.04 V). This is equal to

44

2.09 X 10'9 1_10% coul.. Apparently. the amount of charge
received by the dummy cell is equal to that released by
pulse generator within experimental error. The time
constant acquired from exponential fit of the data is
50.23 microseconds. It is comparable to the RC constant
(50.0 1 0.5 microseconds) of the dummy cell.

It is noted that it takes about 7 data points
(Figure 2-12A) to reach the first valid data point. This
amounts to about 0.5 microsecond. Since the recorder needs
approximately 0.4 microsecond to rise and settle. this
means the charging time of the dummy cell is probably
less than 100 nanoseconds. This amount of time is close
to the settling time of the charging amplifier specified
by the manufacturer.

Figure 2-13 shows the results of the same experiment
collected at 5 MHz. The Va and decay constant of the
transient obtained from curve fitting are 0.821 V and
50.02 microseconds. respectively. The variation in initial
voltages acquired at different collection rates is
attributed to the uncertainty in the validity of the
first data point. As the decay constant is irrelevant to
Ve. an excellent agreement between two experiments is

observed for this quantity.

45

 

 

 

 

 

 

1.000
S . A
m \
3:9 .
8
>
00000 [— r r I TJI Jr I r r I J]
0 210
TIME (MICROSECJ
. 1.000-
DJ
2
'8
>
0.000 I I t t ‘r } t Lr % 41
0 210

TIME (MICROSECJ

Figure 2-13. Voltage transient resulting from a pulse
application to a resistor and capacitor
in parallel. Data were collected at 5 MHz.
A: transient. B: transient plus its
exponential fit.

CHAPTER‘III

FORMATION CONSTANT OF GRAMICIDIN
CHANNELS IN LIPID MEMBRANES

INTRODUCTION

In the past few years. two studies have been
reported that concern the formation constant of
gramicidin channels in bilayer lipid membranes. In Bamberg
and Lauger's work (46). a voltage-jump relaxation method
was used in which the rate constants for the formation
and dissociation of the dimer were obtained from the
time course of the current after a sudden change of the
potential.

In 1975. based on the simultaneous measurements
of membrane fluorescence and conductance. Veatch et al.
(47) concluded that dancyl gramicidin C, a highly active
analog of gramicidin. forms a dimeric channel and has a
dimerization constant comparable to that of native
gramicidin channels.

'Recently. we have developed a new and unique
approach to the problem. It is based on the observation
that when lipid membrane is properly doped with

gramicidin. the voltage discharge through the membrane
46

47

after a charge pulse is purely exponential. As a conse-
quence. the number of dimeric channels in the membrane
could be derived from the Voltage relaxation time if the
membrane capacitance is known. With this new method. we
found that the formation constant of gramicidin channels
depends greatly upon the membrane thickness. Furthermore.
a constant partition coefficient of gramicidin between
the membrane and torus phases exists.

In this chapter. Lauger et al. and Veatch et al.'s
work will be discussed briefly. The major emphasis will
be on the newly developed approach and the results there-
from. We will show that our data are in accordance with
others and that the assumptions involved in the study

are legitimate and justified.

Curzent Relegation Apppoach

In all three approaches to the measurement of
dimerization constant of gramicidin. it is assumed that an
equilibrium between monomers (A) and dimers (A2) exists in
the membrane. This equilibrium may be characterized by an
association rate constant kR (sec’l-l-mole'i). a dissocia-
tion rate constant kD (sec-1) and an equilibrium constant

K (l-mole'l)

48

no
ll

k N

:11
ll
W
U
2

T - 2 cocoon-00000000090

where N1 : concentration of monomers A in mole-1"1

N2 : concentration of dimers A2 in mole-1"1

R : rate of channel formation in
mole-l’l-sec"1

D : rate of channel dissociation in

mole-1"1-sec"1

It is further assumed that the exchange of A and
A2 between membrane and aqueous phases is sufficiently low.
so that the total concentration N of gramicidin in the

membrane remains constant during the experiment 3
N = N1+2N2 ......OOOOOOOOOOOOOO (3)

In Lauger and Bamberg's approach. a differential
equation is set up which describes the rate of change in
channel concentration after the voltage jump and the

solution of this equation reads:

e-t/v
_t/, .. (4)

 

sz = N20») - (Nzow) - N2<o)) “we

49

(1+8NK)% (5)
'th =
m q 4K(N2(°°)-N2(O))

 

1
, = (6)

kD+4kRN1 (°°)

 

At a given voltage. the current density J(t) is
proportional to the concentration N2 of conducting
channels. Using equation (4). the time course of the

current may be expressed by:

J(t) - J(O) 1 e't/T
J(oo) -J(0) - "q‘ 1+q-e"“7T

 

 

....... (7)

Equation (7) reduces to an exponential form.

equation (8).

J(t) - J(o)

 

= i-e‘t/“' (IqI>>1) (s)
J(oo) - J(o)

for large values of the parameter. q. which implies either
small displacements from equilibrium (i.e. N20»)-N2(0)
----*e 0) or low gramicidin concentration (NK>>1).

Via the use of single channel conductance, G . and

Avogadro's number NAv' equation (6) assumes the form:

50

1

 

0° %
kD+Lr(1000kaRx( )/NAVG d)

where x(“0 is the membrane conductance ar t =¢n.
and d is the membrane thickness in cm and Go the single
channel conductance in 0'1.

It is thus evident that a plot of 1/+ vs. (10»))%
would result in the values of kR and kD. The dimerization
constant K could then be calculated according to
equation (2).

Within experimental error. Lauger and Bamberg
found that the time course of membrane current after the
voltage jump is indeed exponential. Accordingly. a value of
4.8 x 105 l-mole"1 was obtained for the formation constant
of the dimer in lecithin membranes at V=205 mV. It is note-
worthy that in this method no direct measurement of
gramicidin concentration in the membrane is necessary. as
is evident from equation (9). Furthermore. both kinetic
and thermodynamic constants are readily available from the
study. The disadvantage of the technique is that the
parameters are acquired under the influence of an applied
voltage. The rate and formation constants at AV=0 mV could
only be estimated by extrapolating to AV=0 the parameters

obtained at a series of nonzero AV's.

51

Analog Approach

Equation (2) and (3) suggests that the dimeriza-
tion cOnstant of gramicidin channels can be expressed in

terms of N and N2 according to the equation:

K = NL 2 (10)
(N-2N2)

 

In 1975, Veatch et al. devised dancyl gramicidin
C (DGC). a highly fluorescent and active analog of
gramicidin A. Assuming that monomeric and dimeric
gramicidin have same excitation characteristics and
quantam yield. N could be determined directly by measuring
the fluorescence intensity of the membrane. With the know-
ledge of N2. which is proportional to the membrane conduc-
tance. the dimerization constant. K. of DGC could be easily
calculated according to equation (10).

In spite of structural differences. Veatch et al.
obtained results comparable to those reported by Lauger
and Bamberg. They found that the dimerization constant is
greatly influenced by the membrane thickness. Furthermore.
a constant partition coefficient of gramicidin between the
torus and membrane phases exists when the concentration of
gramicidin in the membrane forming solution is sufficiently
low.

It is interesting to note that Veatch's approach

52

is thermodynamic in nature and not capable of yielding
any kinetic information. Because of the requirement of the
incorporation of a fluorescent probe into the channel
former. the analog approach is not as convenient nor as

universal as the voltage-jump experiment.

Voltege Relegatioh Approach

In 1976. in an attempt to clarify the cause of the
shift of the gramicidin monomer-dimer equilibrium in the
membrane after a voltage jump. Bamberg and Lauger (69)
discovered that the membrane capacitance and conductance
begin to relax several milliseconds after the voltage
perturbation. This finding is consistent with our observa-
tion that when the lipid membrane is properly doped with
gramicidin. the time course of the voltage relaxation after
a charge pulse is purely exponential.

0n the basis of these two facts. the following
conclusions were reached:

(1) For a fast voltage relaxation (less than a few
milliseconds). the membrane capacitance and
conductance are independent of the membrane
potential and can be deduced from the voltage
relaxation time.

(2) Assuming the interactions between gramicidin
channels are insignificant. the density of

gramicidin dimers can be estimated from the

53

membrane conductance according to the

equations:
G'.--C/T ....OOOOOOCOOIOOOOO. (11)

N2=1000G/G0NAvd oooooooooooooooooooo (12)

where G : membrane conductance in {fl-cm-2

C : membrane capacitance in F-cm"2

r : relaxation time in sec.

Ge : single channel conductance in 9-1

N : concentration of gramicidin channels
in mole-l"1

NAv‘ Avogadro's number

d : membrane thickness in cm.

(3) The dimer concentration. N2. obtained from
voltage relaxation experiments represents the
equilibrium concentration of gramicidin
channels in the membrane prior to the charge
perturbation.

To utilize equation (10) to obtain K. we assume

that the partition coefficient. D, of gramicidin between

the membrane and torus phases is a constant:

D = N/B .................... (13)

54

where B is the gramicidin concentration in the
membrane forming solution. Combining equation (10) and

equation (13) and solving for N equation (14) is

 

2’
obtained:
1.
DB 1 - (1+8KDB)2
N2=-——-——+ ooooooooooo (11+)
2 8K

Equation (14) consists of two experimentally
available variables. N2 and N. and two physical parameters.
K and D. By use of a weighted nonlinear least squares
program. K and D could be estimated by fitting experimental
data. N and N2. to equation (14).

Explicitly. the assumptions involved in our
approach are as follows:

(1) An equilibrium in the membrane between a
nonconducting monomeric gramicidin and a
conducting dimer exists.

(2) There is no interaction between gramicidin
channels so that single channel conductance
could be used to calculate channel density
from the membrane conductance.

(3) Conservation of membrane bound gramicidin
persists during the course of experiments.

(4) The partition coefficient of gramicidin
between the membrane and torus phases is a

constant over the concentration range used in

55

the experiments.

EXPERIMENTAL

Material as Apperetue

All chemicals and solvents used are reagent grade
and used without further purification. The glycerol
monooleate (GMO, technical grade) and cholesterol were
purchased from Matheson. Coleman & Bell and Fisher
Scientific Company. respectively. n-Decane was obtained
from Sigma. n-tetradecane from Phillips Petroleum and n-
hexadecane from Aldrich Chemical Company. Gramicidin
(isolated from Bacillus brevis). a mixture of gramicidin
A (85%). B (4%). and C (11%) (64). was purchased from
Sigma and used without further purification.

The electrolyte solution used for the charge
pulse experiments was 1M KCl. This was chosen because
potasium ion gives rise to higher membrane conductance
than lithium and sodium ion. The membrane solution was
prepared by adding GMO (3% by weight) and cholesterol (1%
by weight) to n-alkanes. A stock solution of gramicidin
in methanol was kept in the freezer. Appropriate amounts
of this solution were added to the membrane forming
solution. The resulting solution was sonicated for 30
minutes and used immediately. In this chapter. unless

specified. "the gramicidin concentration in the membrane"

56

is equivalent to "the gramicidin concentration in the
membrane forming solution".

The old instrument (see Chapter II and references
(77. 78)) was used to perform the charge pulse experiments
because membrane experiments and the construction of the
new instrument were carried out at the same time. A 200
nanosecond current pulse of appropriate amplitude was
applied to the membrane so that the initial membrane
potential is around 300 mV. Since no thermostate was used.
the temperature of the solution was checked before and
after each pulse experiment to ensure it is within 10 of
24°C. Usually. two to three measurements could be made
before the readjustment of solution temperature was

necessary.

W122;

The electrochemical cell used for the membrane ion
transport was described in Chapter II. The membranes were
formed on a 1.6 mm orifice in a thin Teflon barrier which
was sandwiched between the two electrolyte compartments.
The membrane formation was accomplished using a Pasteur
pipet to apply approximately 1-5 ul of the membrane forming
solution to the lip of the orifice. By using the Pasteur
pipet to gently create air bubbles along the barrier. just
below the orifice. the membrane solution can be repeatedly

swept across the orifice until a membrane forms. The

57

membrane area was measured with a microscope and graticule

2 cm2 (80% of the total area)

and was kept about 1.6 x 10'
for all the experiments. Depending on the membrane composi-
tion. it usually takes about 30 seconds for n-hexadecane
membranes to thin and 1.5 minutes for n-decane membranes.
Tetradecane membranes thin in a time intermediate between

these values.

Mehhrehe Date Cellection

The program used to control data collection is
named "MEMBRN.FTN:2". It is a modification of "MEMBRN.
FTNgl" (78) to suit our recently installed multiuser
operating system (RXS-il) on a Digital PDP-11/40 computer.
A listing of the program is shown in appendix B. This
program can be used to collect. plot, list. exponentially
fit and store data in a file on any mass storage device. To
collect data. the recording and readout rates of the
transient recorder. as well as the number of data points
collected at the fast recording rate are entered in a 16-
bit word through the terminal. The bit assignments are
shown in the program listing under "Data Collection-
Section". It should be noted that a base line scan (n
average) must be acquired before the program will honor a
store command. This is necessary to allow correction for
the transient recorder's background noise (see references

(77. 78)). After the data have been stored. a plot command

58

will plot the baseline corrected data. To exponentially

fit data. a data file name and two integers that specify
the first and last data points of a data block in the data
file must be provided to the program. The program will then
return the fitting results: the slope and intercept of

an vs. time and the standard deviation of fitting.

Data Treatment

The data collected from charge pulse experiments
were exponentially fit using Fit command in MEMBRN.FTN;2

to acquire the relaxation time according to the equation:
7: 1/-SR

where S : slope of an vs. t plot (V : membrane
potential: t : time)

R : data collection rate

The fit section of MEMBRN.FTN;2 is an exponential
curve fitting routine. The program linearizes the exponen-
tial data. calculates the proper weighting coefficients.
and then calls a weighted linear least squares fitting
subroutine.

The dimerization constant and partition
coefficient of gramicidin in the membrane were acquired by

fitting data (i.e. channel and gramicidin concentrations)

59

to equation (13) using KINFIT (74) on the MSU CDC-6500
computer.

For a given membrane forming solution of certain
gramicidin concentration. 7 is the average of at least 10
experiments. each performed on a new membrane. To calculate
the channel concentration. equation (12) is used assuming

4152-1 (37) and a

a single channel conductance of 4.1 x 10
bilayer thickness of 32.0A for hexadecane membranes. The
thickness of other membranes is estimated from that of
hexadecane membranes based on the ratio of their capaci-
tances.

KINFIT requires that initial estimates for the
parameters be given by the user. It is often a necessary
condition for convergence that these initial estimates be
fairly close to the actual values.

For n-tetradecane and n-hexadecane membranes. two
variables (channel and gramicidin concentrations) were
fit to equation (14) to acquire the dimerization constant
and partition coefficient by use of KINFIT. For n-decane
membranes. because of difficulties encountered in using
KINFIT (error mode of negative numbers in square root
operation despite various estimates for initial values).
partition coefficient, D. was given as a constant in
equation (14) and the standard deviation of fitting

evaluated for a number of Ds. It turned out that when D=1.

the fitting error is minimal.

60

RESULTS AND DISCUSSION

Memhrane Capacitance Measurehent

To obtain membrane conductance by means of the
charge injection method. a knowledge of the membrane
capacitance is required .

Figure 3-1.1 shows the result of a pulse application
to a.GMO-cholesterol membrane with no external bleeder
resistor. Two important facts can be recongnized from this
figure. First. it is clear that discharge through the
membrane or external circuitry is minimal. and therefore
the membrane capacitance can be determined accurately by
this method. Secondly. it shows that if any dielectric
relaxation occurs in the membrane. it must either have a
very small amplitude. or occur on a time scale which is
either very much shorter or very much longer than the time
scale of the charge pulse experiments. In either case.
Figure 3-1.1 shows that the results of the membrane experi-
ments will not be affected by a dielectric relaxation.

Figure 3-1.2 shows the transient resulting from the
application of a pulse to the parallel combination of the
membrane capacitance and a 10k 7: 0.1% resistor. The solid
curve is the result of a weighted linear least squares fit
on the linearized exponential transient decay curve. The

average of 7 such experiments. each performed on a new

61

_ . .AchH.o.H oHv

aeeusuea assesses as see; Heaasssasa :qm . m .sqm seaweeoesa . H .uzgem
op umSmm.a Scam owcmso opts sofipooaaoo mums on was ma hpflscflpsoomwm
.umc on pay moumsum unwed auspcocomxo mucomopmon o>pso caaom

.oeH H.3N pm Hog EH Cw mmcmunsos osmoocmapop:s\HopopmoHoco RH 1020 an
0p soflpmowammm omasa Scam msfipasmmu mpsmwmcmup ommpao> arm .mfim

“6322.... as:

 

   

.
c
on. . e1; 8 3 as e
.N . m I A I I I I I coed
O. u 00
.. n -1
.
.u
. 1‘
.-
u
.u 4, M
. I
1
n 1- D.
. 3
u .u
. Ll ‘II
n A
. ll‘
“ 1....
o
o
. 1'
o
o
a
n 1:
0.0000.OOOOOOOoooo—OOOOOOOOOOMOOOOOOOOOOOOOOOO00000000000.OOOOOOOOOCOPOOO 0000000000.. .0000. o
m . 90 0 .00... 0.0000. .000 1T
. .uloono

 

 

 

.: ventralllldi. d v .. ...
. 1'. .9 tr! J... .521! ¥C.b:nh§vua.

 

 

62

Table 3-1 The thickness and capacitance of 3% GMO-
1% cholesterol/n-alkane membranes (w/v).

 

 

 

Solvent Capacitgnze Thickness (A)b
(F-cm ) _
n-decane 4.31 1 0.08 x 10"7 46.5
n-tetradecane 5.45 1 0.10 x 10"7 36.8
n-hexadecane 6.26 5; 0.11 x 10’7 32.0

 

a : The membrane capacitance is the average of 5 measure-
ments. each performed on a new membrane.

b : The thickness of n-hexadecane membranes was taken as
O
32.0A according to reference (72-73). The thickness of

other membranes was calculated from that of n-
hexadecane membranes based on capacitance ratio.

63

membrane. resulted in a value of (5.45 1 0.10) x 10'7 F/cm2
(this precision is comparable to the estimated error in the
membrane area measurement).

In Table 3-1 are listed the capacitances and thick-
nesses of various n-alkane membranes. The bilayer thickness
was calculated from the capacitance ratio assuming a
32.0A-thickness for hexadecane membranes (72). The variation
of membrane capacitance and. in turn. the thickness as a
function of chain length of the hydrocarbon solvents is in
accordance with reported observations (72. 38). It simply
implies that as the chain length of solvent in the film gets
longer. the volume fraction of solvent in the membrane
becomes smaller and. as a result. the film approaches the
"true" bilayer membrane structure.

To measure the membrane capacitance of bilayers
doped with gramicidin. we utilized the fact that gramicidin
channels are not permeable to anions and divalent cations
(38). Figure 3-2.1 shows the voltage transient resulting
from a pulse application to a gramicidin-doped membrane in
0.5 M CaClz. It is evident that on the time scale of charge
pulse experiments. no appreciable voltage discharge through
the membrane occurs. Figure 3-2.2 shows the result of a
pulse application to a gramicidin-doped membrane in
parallel with an external resistor (10km). The relaxation
experiment results in value of 5.35 x 10.7 F/cm2 for the

membrane capacitance.

64

.COHpsaom msfisuom osmunsos one CH cacaoasmaw SoroH N no.“
.moCMHQSos osmoovmupoprc\aomopmoaono mauosufim . :ofipficsoo
.Nmmmn op uxsmm.a Scum owsmco mums sowpomaaoo mums op was
ma hpwsflpsoomwo .Mpmu op paw mmumsvm pmmoa Hmwpsosomxo mpcmmmhmon

o>uzo cfiaom .Oea H :N am Naumo 2m.o CH mosmunsos somocrsasHOHSmnw
op sowpmofiammm omazm Scum wcwpasmmh masofimsmpp mwmpao> Nun .mwm

33225 0...:

 

 

HAS. «SR: 00 0' ON 0
“00000000000ofi0000N " u w w “ b1 IlbfiI “ “ 00H00 080°
000 a .
00 . til
0 o
.
0 a r
. m I
.u -- M
o '
. 1.. 0‘-
. e
u 00
. 1.: a
u \I
u .i In
. (
o
a
o 1.!
000 0 0 0 0 0 0“0 0 0 0 0 0 0 0 0 0 0
0 0 0000000000000? 0 0 0 m 0 0 0 00000 0 0 0000000 000 000 00000000F000000000000 000 0000000000000 0 0 0000000
. 0 lrl
«
.

 

65

0n the basis that the capacitance of unmodified BLM
is equal to that of gramicidin-doped bilayers within
experimental error. it is concluded that the former can be
used to calculate membrane conductance from voltage relaxa-
tion time obtained from pulsing bilayers containing

gramicidin.

V t R t' a o. G am’c' ’n-d ed Me anes

Voltage transients resulting from pulse applications
to gramicidin-doped membranes in 1 M KCl are shown in
Figure 3-3. Both traces correspond to 8.32 x 10-7M
gramicidin in the membrane forming solution. The slight
vertical shift of curve 1 with respect to curve 2 and
discontinuity of the curves is due to a difference in
membrane area and a change in data acquisition rate.
respectively. The solid curves result from exponential
least squares fits. Two important conclusions can be drawn
from this figure. First. gramicidin-mediated voltage
relaxation is purely exponential in the submillisecond time
range. The exponential nature of the decay suggests that
the membrane conductance is independent of membrane
potential. As a result. the channel concentration derived
from the relaxation time represents the equilibrium dimer
concentration in the bilayer prior to the charge perturba-
tion. Second. the relaxation times for Figure 3-3.1 and

3-3.2 are 47.50 x 10"6 second and 42.6 x 10'6 second.

66

.cowpsaom mcwshom osmucsoa one a“ zauwowsmpw

25:0H x mm.m .osmoocmppoprfi\aouopmoaono *« rose fin . codewosoo

.umxmm ow nmsnm.a song owsmno mama :ofipooaaoo mama op mac

ma szSSApCoomfln .mpmc op paw moumsam unwed Hmvaocogxo pcomoumou
mo>pso cflaom .oea H am pm Hum EH :H mocmppsos somocrswuwowEmpw

op CowmeAHQQm omasm Sony mswpasmoh msofiumxmaop ommpao> mum .wwm

«.2.
p r

3823.5 3.:
00 . 00

h -

fi-

On

p
4

~-

 

a.uuuuu.:.qo:u~.uu..lwe

---... d ho-..

J-

‘F

1 a

(II) 9321101

 

67

respectively. which amounts to a precision of about 8%.
Although for some membranes a relative standard deviation
of 20% in membrane conductance,measurements was observed.
the precidion of the charge injection technique is compa-
rable to. if not better than. that of the voltage clamp
method. as is evident from data presented in references
(46). (47) and (70).

Figure 3-4 shows the capacitance and voltage
relaxation time for gramicidin-doped membranes as a
function of membrane age (t=0 at the instance the membrane
is fully thinned). As pointed out by White and Thompson
(71), to be a useful and valid model system for biological
membranes. planar bilayer lipid membrane must assume a
homogeneous equilibrium structure with a well defined
stoichiometry. A reliance of membrane capacitance on
membrane age implies a.non-equilibrium bilayer structure
which undergoes constant changes. Figure 3-4A shows a
zero-order dependence of membrane capacitance on time.
suggesting a stable configuration for gramicidin-doped
GMO-cholesterol membranes. However. as the film aged. a
slowdown in voltage relaxation obtained from pulsing the
membrane was observed (Figure 3-4B). The rise in relaxation
time is attributed to a decrease in membrane conductance
because the membrane capacitance is constant. This finding
is in accordance with the fact that gramicidin was added

to the membrane phase only. The very poor correlation of

68

Figure 3-4A with Figure 3-4B again shows that the membrane
capacitance is independent of gramicidin concentration in

the membrane.

Formatioh Cohetaht QI Grahicidin Channels

The voltage relaxation times resulting from charge
pulse experiments and the membrane conductance of various
GMO-cholesterol membranes at different gramicidin concentra-
tions are summarized in Tables 3-2. 3-3 and 3-4. The
membrane conductance and channel concentration were
calculated from relaxation time according to equations (10)
and (11). respectively. The corresponding N2 vs. B2 plots
and KINFIT fits to equation (13) are shown in Figures 3-5.
3-6 and 3-7. In Table 3-5 are listed the formation
constants and partition coefficients resulted from curve
fitting as well as literature values of these same
parameters.

Several interesting features can be recognized
from the figures and listed results. First. the relative
standard deviations of relaxation time measurements for
n-decane membranes are considerably larger than those for
n-tetradecane and n-hexadecane membranes. This is probably
due to the fact that for less permeable membranes. slight
flaws in membrane structure such as microlenses (71) or air

bubbles exert a greater effect on the membrane conductance

than for highly permeable membranes. Second. except for

69

AooQMpHomamOv .SNroa x 0.: . :HcHOfismuw .maomo En.o . ophdoupooao . m
AoSap Cowpmxmaouv .2010“ x N.H . cflcwowsmum .Hom EH . opaaoaaomao . d

.soccan SHHSH ma osanEos on» oocmpmsw 029 pm 0 mamsvo mafia
.Aocmoosmxonrs\aoumpmoaoco Rarozu gmv mocmhnsos somocrcwsflowsmpw
go mocmpwoQOO ocwunsos new eBay coapmxmdoh owmpao> one arm .wwm

3... V ...:

 

 

 

 

 

5

ON n— O— n O

«.2.
n
J
/
o .
mW.'°:

'6
o O. IN III. klllJlJI R .l I h is i
a ....
C -L‘“
O.“ ‘
\\\-‘0
Ila...
.-‘Q!
.--- 0"
\\‘O ‘
.lflx

 

(new) 1.

70

.Sofipzdom ms080oh ocmhnsos one :0 sacaowsmpw Ho soapmupCooCoo 0 m

 

 

 

 

N.-60 s 66.0 6-60 x 60.6 0.66.0 6.60 s 66.6 0.66.0 66 H 666 6-60 x 66.6
6-60 x 60.0 N.-60 x 66.6 H 66.0 6-60 x 66.6 6.66.6 00 H 666 6-60 x 66.6
6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 x 66.6 H 66.6 66 H 666 6-60 x 66.6
6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 x 66.6 H 66.6 66.H 666 6-60 x 66.0
6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 x 66.6 H 06.6 660 H 666 6-60 x 66.0
6-60 x 06.6 6-60 x 66.6.“ 66.6 6-60 x 66.6 H 06.6 666 H 6660 6-60 x 60.0
6-60 x 66.0 6-60 x 66.6 H 66.6 6-60 x 66.6 H 66.6 606 H 6660 6-60 x 66.6
6-60 x 66.6 6-60 x 60.6 H 66.0 6-60 x 66.6 H 66.0 606 H 0666 N.-60 s 66.6
6-60 x 66.0 6-60 x 66.6 0. 66.6 6-60 x 66.6 0. 06.6 666 0. 6060. 0-60 x 66.6
S E 6:80.000 mcsooomoHOHs 2
. :oHmeFCoocoo
006.60.66.60 gunman”... 0.6.666”? Emma. . 53.2.6

I

.AOHV :oHpmsoo op wC0ouooom vmpmasoamo ohm :owetppCoocoo amccmzo 6:6 occuposcsoo
02609802 .060 H.6N pm nonwhnSos osmooors\aouopmodo:orozu conoc::000oasmnw
Ho sowpmupcoocoo Hosanna 6:6 occupancCoo ocmhnsos .650» nowamxmaom Nun magma

71

.mrm 66969 mom . 6

 

 

 

 

6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 x 66.6 H 66.6 6.6 H 6.66 6-60 x 66.0
6.60 s 60.6 6-60 6 60.6_H 66.6 6-60 x 60.6 H 66.6 6.0 H 6.66 6-60 x 06.6
6-60 6 66.0 6-60 x 60.6 H 66.0 6-60 x 60.6 H 66.0 6.6 H 6.66 6..60 x 66.6
6-60 x 66.0 6-60 x 60.6_H 66.0 6-60 x 60.6 H 66.0 0.6 H 0.66 6-60 x 66.6
6-60 6 66.6 6-60 x 66.0 H 06.6 6-60 6 60.0 H 06.6 6.6 H 6.66 6-60 x 66.6
6.60 s 66.6 6-60 x 06.6 H.66.6 6-60 6 66.6 H 66.6 0.60 H 6.66 6-60 x 66.6
6-60 x 06.0 6-60 x 66.6 H 66.0 6-60 6 60.6 H 66.0 6.66 H 666 6-60 x 60.0
6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 x 66.6 H 66.6 666 H 6660 6-60 6 66.6
2 2 Nrso.0rc muCooomohoHs S
...—00 Hmcgr—D Cowfimhvfimonoo mo§m#036300 mew—H. NCOHPMHpfimocoo
00 #MHfioHdO Hwfifiwso 063660.802 0.5.3.0060me CHQwOHfimHO

 

 

.AOHV soHp6=do op w50600606 6696636666 006 :oHp600coosoo Hoss6co 6:6 mos6posusoo
6:609802 .060 H am 06 mocmhnSoa osmooU6Hpoprfi\Hono06666201020 concorchHOHsmuw
Mo :oHp609260coo Hossmso 6s6 mocmpososoo 62600565 .686» nowpmxmaom nun 6696a

72

 

 

 

 

.Nrm @6069 wow . 6
6-60 x 66.6 6-60 x 66.6 H 66.6 6-60 60 66.6 H 66.6 6.6 H 6.66 6-60 x 66.6
6-60 x 66.6 6-60 6 60.6 H 66.6 6-60 x 60.6 H 66.0 6.6 H 6.66 6-60 x 66.6
6-60 x 66.0 6-60 6 00.6 H 66.6 6-60 x 66.6 H 66.0 6.6 H 6.66 6-60 x 06.6
6-60 x 66.0 6-60 x 66.0 H 66.0 6-60 60 66.6 H 66.0 6.0 H 6.66 6-60 x 66.6
6-60 x 66.0 6-60 6 66.6 H 60.6 6-60 00 66.6 H 66.6 6.6 H 6.66 6-60 x 66.0
6-60 60 66.6 6-60 6 60.6 H 66.6 6-60 x 60.6 H 60.6 6.6 H 660 6-60 60 66.6
6-60 x 66.6 6-60 60 60.6 H 66.6 6-60 x 60.6 H 66.6 6.60 H 666 6-60 x 66.6
6.60 s 66.6 6-60 x 66.6 H 66.6 6..60 60 06.6 H 60.6 600 H 6660 6-60 x. 66.6

E S NIEo.0:0 masoommopows s

.soo Hossmso :oHp60psmoCoo oosmpo:UCo0 mafia 6Co6p60pSoocoo
cmpmasoamo Hocsmno mcmnnsoz soflpmxmaom CHUHoHE6pw

 

 

.Aoav Cowemsuo op wsfichooom copmasoa6o 606 cowpmhpsoocoo Hmcs6£o 0:6 mosmpozccoo

6:609862 .060 H 6N p6 66:600868 os66666xocrfi\aohopmmHone-020 sonouICHuHOHammw
Ho :oHp60psmoCoo 6622630 6:6 mocmpozccoo osmpnsos .oSHp :oHp6xmaom dun 0696a

73

.6000 comp6306.
op mapopo006 ”n.06 mo mmmcxownp mcmpnsme 6 wCHESmmm smp6H36666

663 N2 .6600 coppmsuo 6p p60 eHmsz Scum mpasmmp m>hso swaom
.662602868 osmoocrc\aouopmoaonoroso :0 .m .soHpmppcoocoo 206000860m

map Mo :60pocsh 6 66 .62 .060p60pcoocoo Hmc:6no 6606800 mum .mHm

 

that x

 

 

74

.6000 sowpmsvm op

mcacpooom mm.0m mo mmochan os6ppsos 6 wcfissmm6 Ump6azoa6o 663 N2

.ANHV so pmzdo op p00 eHmsz Scum mpasmmp o>hso cwaom .mosmanos
6:66666ppop::\aopmpmoao:orozu :0 .m .soHp60pcoocoo cHsHoas6pm

62p 06 sOHpoczm 6 mm .Nz .co0p60pcmocoo Hmccmno 0006800 cum .wpm

sssx

o. m o

W90|X

 

75

.Aofiv COHpmzvm op

wcflcnooom Mo.mm ho mmmchan mCMMQEmB m mcfiszmmm umpMHSOHmO mm: Nz

.ANHV co pmsvm o» paw eHmsz Scum mpasmmg m>u30 cwaom .mmcmunEws
mamowomxmnuc\H0hwpmmaosouozu a“ .m .coapMpHCmocoo CHUHOHEMQw

mzp mo cowpocsm m mm .mz .COHpmupcmocoo Hmccmco owhmEHQ mam .wfim

3::

n C n a — O

' x. b L
‘JI‘““{J.J“‘A“ “qu‘11‘1“111“41.““J‘I‘“q4‘4“‘IJ‘J

W¢0Ix

‘A‘LJA‘."+AAAAA‘L.LJALAAAA

 

76

 

 

'

xaoz wasp :.fi H m.mfi mofi x m.¢ H m.a Hopmummaogo-ozu manomcmxmnnc

 

 

 

2;: . age o: :3 x R 23
ago: wasp 3.0 H 0.0 no« a m.« H m.n Honmpmmaono-ozo mcmomcmgpmp-c
Auav - oA>e cHV no“ N fi.d ozu
ghoz wasp « 30H N N.o.H m.fl Houmpmmaononosu
Amav mfi.H .uHm oA>a oHv 5: 02m
Amdv I oA>a oHv no« N o.m om osmowwuc
Ammv - A>a mnfiv so” x o.e ow
Away - An>a nmflv mofi x m.H om
Am:v - ac“ x m.c mum
monzom Q HAHISV x caan pam>Hom

 

 

.mmcmunams vflmwa awkwawn Cw chfiowsmhw
mo .Q .mpCowoawmmoo soflpaphmm ccm .m .mp:Mpmcoo :oaaduwpman mum manna

77

.muospsa uopfio an vow: omonp 809% oqumno mum; 2 how mpfics may . H
.AHIE sod N m.Hv vow: mafimn M use pamoxm COHpMHSUHmo osmm . m

.H-2 m: ..m.H
.m pwmsp no owwmn Amdv mocmumMmp :H co>fiw «Hut 809% cmpmazoamo was a . c
.pCmsfihmgxo map ca com: mm: o :HoHOHEMpw Hamnmn . o
.umposu:oo ma: pcmawnmgxm msv scan: pm Hmwpzmpon mcmnnSmE . n
.mCHHozoWhqumnmmosm AhomHOHu u om . m

Acmscwpcoov mum manna

78

n-decane membranes. the fit of experimental data, B and N2.
to equation (13) is very good as is judged from random
deviations of channel concentrations from calculated values
as well as from relatively small standard deviations of
fitted results. Thirdly, the formation constant of
gramicidin channels depends markedly on the thickness of
the hydrocarbon core. Hladky and Haydon (38) found that the
mean duration of a single channel increases monotonically
with decreasing membrane thickness. They attributed this
effect to a variation in the extent of local thinning or
dimpling of the membrane in the vicinity of the conducting
channel. Our data are consistent with their proposal. The
nearly 52 fold decrease in the dimerization constant in
going from a gzfi to a 46.5K GMO-cholesterol membrane
corresponds to an increase in the free energy of dimeriza-
tion of 2.3 Keel/mole. This free energy difference might
represent the energy of deforming a small region of the
“6.53 membrane to match the length of the gramicidin
channel. Fourth. for a given dimerization constant, the
relative population of gramicidin channels rises as the
total amount of gramicidin in the membrane increases. For
n-decane membranes, based on our results, only 10% of
gramicidin is in the channel form for a membrane forming
solution nearly saturated with gramicidin. By contrast,
more than 75% of the gramicidin in hexadecane membranes is
in the active dimer form when the gramicidin concentration

is six times less than the saturation concentration. This

79

implies that, for solvent-free membranes (such as lipid‘
vesicles) which are even thinner than hexadecane membranes.‘
practically all the gramicidin molecules in the bilayer are
in the channel form. As will be discussed later in Chapter
IV, this is used to advantage in attempts to photolabel
gramicidin channels in lipid membranes.

Also listed in Table 3-5 are the literature values
of the dimerization constant. Similar to our results,
Veatch et al. (47) observed a dependence of channel formar
tion constant on membrane thickness despite of the fact
that dancyl gramicidin C, an active analog of gramicidin.
was used in the study. It is noticed, however. that the
present results are uniformly larger-than the values
reported by Veatch et al.. 7

A careful examination of data given in Veatch
et al.'s work results in the following findings:

1. In converting the observed fluorescence

intensity to the absolute surface density
of dancyl gramicidin a value of 1.0 was'
assigned to the relative efficiency of detection
of the dancyl fluorescence without citing any
reference. It was also assumed that both
monomeric and dimeric gramicidin have the same
excitation characteristics and quantum yield.

2. The dimerization constant of 9 x 1013 M"1

obtained from fluorescence measurements for

80

dancyl gramicidin C in dioleoyl phosphatidyl-
choline is a factor of five less than the value
obtained using the voltage-jump technique for
dancyl gramicidin C on the same membrane.
Veatch et al. attributed this disagreement to
the overestimate of the total gramicidin
concentration by fluorescence measurements.
They noticed that Raman scattering and
fluorescence arising from the solution can give
a much larger signal than the fluorescent
species in the membrane. It was also reported
that for a.263 GMO membrane an average value of
1.h was found for the ratio Cd'/Cd where Cd' is
the gramicidin concentration calculated from
membrane conductance (Cd'=2N2) and Cd the total
gramicidin concentration obtained from fluore-
scence measurement. As the error resulting from
conductance measurements is much smaller than
that from fluorescence measurement and in
principle the Cd'/Cd ratio should be equal to
or less than one (the ratio equals 1 when all
the gramicidin is in the dimer form), a value
of 1.4 for Cd'/Cd suggests an underestimate of
total gramicidin concentration in the membrane
by fluorescence measurements. However. no

explanation was given for these apparently

81

conflicting results on the fluorescence
measurements.

3. No statistical or mathematical means was

utilized in data treatment. In fact, except for
n-decane membranes. onlu ONE data point, i.e..
one simulataneous measurement of membrane
conductance and fluorescence, was reported for
each type of the membrane. Consequently. it is
rather difficult to estimate the reproducibility
of the measurements and. in turn. the relibility
of the inferred results.

Regardless of these findings, on the basis of
rather strong fluorescence background emitted by plain
membranes and solutions. it seems that the fluorescence
technique provides a lower boundary for high dimerization
constants.

Another interesting feature evident from Table 3-5
is the dependence of dimerization constant on membrane
composition. Veatch et al. discovered a ZOO-fold increase
in formation constant going from a GMO to a dioleoyl
phosphatidylcholine membrane. Parallel correlation also
exists between Lauger et al.'s and our results although the
difference is less drastic. Since all these membranes are
composed of neutral lipids with similar chain length and
use same alkane solvents. factors other than bilayer thick-

ness and charge distribution must be responsible for this

82

phenomenon. Consequently, the difference in bilayer
composition might be another reason for the discrepancy

between Veatch's and our results.

Partition Coefficient of Gramicidin between the Membrane
and Torus Phase§

Another important physical parameter obtained from.
our study is the partition coefficient of gramicidin
between the membrane and torus phases. As is evident from
Table 3-5. the partition coefficient also exhibits a
strong dependence on membrane thickness. It increases from
1 for n-decane membranes to 15.6 for n-hexadecane
membranes.

Recently, Sychev et al. (75) reexamined the
spectral properties of gramicidin A and its analogs in
dioxane (a nonpolar solvent) and obtained a value of
1x105 M'1 for dimerization constant. On the basis of
theoretical calculations in conjunction with observed
dissociation rate constants for the dimers, they concluded
that the predominant dimeric species in the nonpolar
solvent is different from that in the membrane and thus
suggested that the bilayer structure of a membrane has a
greater influence on the dimer conformation than on the
monomer's. If the torus region could be approximated by

the dioxane solution and the gramicidin monomers assume

similar structure in both membranes and nonpolar solvents.

83

the partition coefficient would then be a measure of
stability of membrane dimers with respect to dimers in
dioxane and the dimerization constant would represent the
stability of dimers relative to monomers. In Table 3-6

are listed G 's calculated from partition coefficients and
dimerization constants according to equation (15):

O

4AG = -RTan .................... (15)

where‘AG° : the change in standard free energy in
Kcal-mole'l
R : gas constant

T : temperature (24°C was used)

K a partition coefficient or dimerization

constant

Since‘AGgar equals zero for n-decane membranes. the

standard free energy of dimers in dioxane solution and

n-decane membrane should be the same. As a result. the

. . o , o . .
following relation between.AGdim. s andMAGPar. 13 derived :
o = O o
AGdim.,alkane ‘AGdim.,decane +‘AGpar.,alkane

................. (16)

where the alkane is either n-tetradecane or n-

O

dim. s for dioxane and

hexadecane. If the average of‘AG

8h

.ooa_H am pm oopooadoo mum; span

.pmm
modem- u cod
oSH .SH
.excaem- u .moo

o 8H6

.cowpzaom osmxoflo osm nonspnsoa osmooo now m. o

04 Ho ommpoam one

63.0013

 

 

 

I. I. I. . I cams Ems
a.o + w.HI :.H + w.na n.o + N.mn mofi N o.m + m.m osmooWMXm:

I. I _l I mama Ems
H.o + fi.HI :.o + 6.0 m.o + a.mu moa x m.H + m.m mcwomomopme

I I. I camp Ems

o a ca.o + n.01 .H.o + a.mn so“ N N.o + m.fi msmwmn
I I . I Ammv soapsaom

m o no“ N a ocmxown

o g .o
Amaoe\amosvo “wood. 6 amaoe\amoavn sawed. AHIEV a mound

 

 

.mmsmpnsos ocmxam
CH cfiowoflsmuw Ho cowpwpamn ego :oHpmenoaHo Mo hwuoco mosh wum manna

85

O
dim., decane'

evident that equation (16) holds within experimental error.

n-decane membrane is taken as theIAG it is
As shown in Table 3-5 only one set of data was
available wherefrom D could be calculated. One major reason
for the scarcity of literature values for this parameter is
that in voltage-clamp experiments. the total concentration
of gramicidin in the membrane is not needed for data
treatment. Based on Veatch et al.'s dimerization constant
(#7 M'l) for n-decane membranes. a value of 21 was obtained

1+ M-l) is USEdv

for D. If our dimerization constant (1.9 x 10
D then equals 1.1. This finding demonstrates that our

results are self-consistent even when their data is used.

Co nt n t e Ass t' ns

The agreement between the experimental data and the
proposed model observed in reported studies as well as in
our approach strongly supports the validity of the
assumptions involved in the investigation. However. one
additional condition was imposed in the present study.
i.e.. that the partition coefficient of gramicidin between
the membrane and torus phases is a constant over the
concentration range used in the experiments.

In analog experiments. Veatch et al. discovered
a linear relation between the total amount of gramicidin
in the membrane and the gramicidin concentration in the

membrane forming solution as long as the latter is less

86

than 1 x 10‘6

M. As shown in Tables 3-2 to 3-#. except for
some n—decane membranes. all the solutions used are equal
to or less than 1 x 10.6 M in gramicidin. It is also
obvious from Tables 3-3 and 3-4 that the partition
coefficient must be greater than one since the dimer
concentrations obtained from conductance measurements are
greater than the gramicidin concentration in the membrane
forming solution.

In 1971. based on the measurements of interfacial
tension. Fettiplace et al. (72) estimated the area per
molecule occupied by glyceryl monooleate (GMO) in the
alkane membranes. They found that the area covered by one
GMO is almost independent of the solvent and has a value
of 39.532. If it is assumed that the same value applies for
GMO-cholesterol membranes and the bilayer is solvent-free.
a value of #.5 x 105 for the lipid to dimer ratio is
obtained for our most permeable membrane. It means two
adjacent channels are separated by 670 lipid molecules. As
a result. the interaction between channels should be minimal
and single channel conductance could be employed to

calculate the dimer density from conductance measurements.

Conglusign

We have successfully obtained the formation

constants of gramicidin channels in lipid membranes as

well as the partition coefficients of gramicidin between

87

the bilayer and torus phases by use of the charge injection
technique. The results acquired are comparable to those
found in literature. In addition. a correlation was found
between the dimerization constant and partition coefficient.
It is of interest to note that although relaxation
experiments have been performed on the membranes. the
resulting physical parameters represent the equilibrium
values of the same parameters prior to the charge

perturbation.

CHAPTER IV

PHOTOLABELING OF GRAMICIDIN CHANNELS
IN LIPID MEMBRANES

INTRODUCTION

Since the discovery that ion transport across the
lipid membrane is mediated by gramicidin via the pore
mechanism (37). a number of studies (50. 58) have been
conducted to determine the molecular architecture of
gramicidin channels in lipid bilayers.

As described in Chapter I. four configurations
exist as the possible structure of gramicidin channels.
They are parallel double-stranded helix. antiparallel
double-stranded helix. C-terminal-to—C-terminal single-
stranded helix and N-terminal-to-N-terminal single-
stranded helix. Results obtained from various investiga-
tions indicated that N-terminal-to-N-terminal helix is the
major. if not sole. conformation of the transmembrane
gramicidin channel.

A careful examination of the reported studies.
however. shows one serious drawback. i.e.. altered
gramicidin or gramicidin analogs have been used in the

experiments. It has been noted that some of these analogs

88

89

are not capable of inducing membrane conductivity and. for
the "active analogues". their single channel conductance
greatly depends on the polarity of the attached chemical
groups. Consequently. in order to obtain unambiguous
information regarding the structure of transmembrane
gramicidin channels. a direct method based on the use of
native gramicidin is highly desired.

In this chapter. we describe the effect of
photolabeling on channel conductivity and its potential

use in unraveling channel conformations.

Phgtolabgligg

In the past ten years. photolabeling techniques
(59) have found wide application in the studies of bio-
chemical systems. Some typical examples are the mapping
and identification of ligand binding sites on biomacro-
molecules (60. 61). the location of membrane constituents
(62-66) and the topography of biopolymers (63).

In such methods. a photolabile probe which suits
the needs of the system in question is selected. Upon
illumination. the photogenerated reactive species '
covalently binds to the target. Subsequent analysis of the
photo-altered target thus reveals the immediate chemical
surroundings that are adjacent to the binding site.

Generally speaking. photolabeling has three major

advantages over the conventional labeling. First of all.

9O

photolabeling probes are inert until photolysis. which
permits some experiments to be dome before irradiation
without any irreversible interactions between the target
and probe. Secondly. many photolabeling reagents can
insert into carbon-hydrogen bonds. Thus. photolabeling
probes can mark any reaction site which contains carbon-
hydrogen bonds and which does not require the presence of
particular reactive functional groups at the binding
site. Thirdly. with an infarared filter. photolysis is
practically free from introducing any heat into reaction
vessel. which preserves the structural integrity of biolo-
gical systems. ‘

As such. photolabeling seemed to be an ideal method

for our study. The reasons are two-fold.

1. Native gramicidin can be used in the experiments
since. as memtioned above. photolabeling does not
require the assistance of any special chemical
groups at the reaction site.

2. Photolabeling probes could be tailored to mark
certain moieties of the channel so that the
analysis of photo-tagged gramicidin would yield
information regarding the channel conformation
in lipid membranes.

For a.transmembrane polypeptide. the amino acid

residues can be classified into two categories: those
located near or on the membrane surface and those in the

membrane interior. As gramicidin forms linear channels

91

across lipid membranes. it is expected that a few of its
15 amino acid constituents would be exposed to the aqueous
phase. It follows that labeling of gramicidin channels on
the membrane surface would produce relatively simple

photoproducts.

Aleaziggg

There are two common reactive species which meet
the carbon-hydrogen bond insertion requirement of photo-
labeling technique: carbenes and nitrenes.

Carbenes can be generated by the photolysis of
compounds such as diazoalkanes. diazirines. and

-ketodiazo compounds. which produce the carbene on loss .
of nitrogen. Although it is very reactive. carbene has
been ruled out as a good photolabeling reagent on the
° grounds that most of carbene precursors are not chemically
inert and stable. and that a-ketocarbene produced from
stable a-diazoketones. esters etc. readily undergoes the
intramolecular rearrangement to a.ketene. which is less
reactive and subject to attack by nucleophiles.

In contrast. nitrenes can be produced from
chemically inert azides by reactions similar to those
forming carbenes. As illustrated in Figure 4-1. a nitrene
could undergo a variety of reactions after its formation:
abstraction (normally of hydrogen from carbon). cyclo-

addition. direct insertion (usually into carbon-hydrogen

92

.cowpmsuoh mp“ hopes eschew: m on ammo wsoflpoMmm HI: .mwm

332.... 32:
..osougataom q

+ :2:

z
=
z
z
a?
\I/
oo

 

.... .... w. e .. =5.

93

bonds). attack by nucleophiles. and rearrangement. Direct
insertion. abstraction-coupling or addition reactions
will result in covalent attachment of label to the target.

Rearrangement reactions can. as with carbenes.
reduce the effectiveness of the reagent. Acylazides
rearrange smoothly on photolysis to isocyanates; whereas
arylnitrenes are much less susceptible to rearrangement.

In addition. alkyl azides have absorption maxima
around 200 nm and it is not possible to effect photolysis
without radiation damage to the polypeptides. With proper
substitution on the aromatic ring. arylazides can be
photolyzed to the arylnitrenes at wavelengths greater
than 300 nm. Thus it appears that arylazides are

appropriate reagents for photolabeling experiments.

z-Nitno-fi-Azidobgnzoic Acid

Of various arylazides. 2-nitro-5-azidobenzoic acid
(NABA) has been chosen as the photolabeling reagent in
our studies of gramicidin channels in lipid membranes.
The introduction of an electron withdrawing group (-N02)
into the aromatic ring shifts the absorption maximum and
pKa of the benzoic acid to 315 nm and 3.6 respectively.
It follows that. with proper buffering of the aqueous
solution. arylnitrenes photolytically generated from NABA
would essentially stay in the aqueous phase and covalently

bind to gramicidin channels on the membrane surface.

NO
' COOH

Q
fi_\

I
”3

2-nitro-5-azidobenzoic acid

NABA

EXPERIMENTAL

Regents

All chemicals and solvents used were reagent
grade and used without further purification. Freshly
deionized water was used to prepare all solutions. The
lipid used for liposome solutions was phosphatidylcholine
(type V-E. from egg yoke. Sigma Chemical Company).
Gramicidin. a mixture of gramicidin A (85%). gramicidin B
(ufi). and gramicidin C (11%) (64). was purchased from
Sigma and used as received. Sephadex G-50 (particle size
50-150 micro. pharmacia Fine Chemicals) was suspended in
0.1 M phorphate buffer. pH 10. for at least 6 hours before
column packing. 2-Nitro-5-amino—benzoic acid was purchased
from Pfaltz and Bauer: trifluoroacetic anhydride.
trifluoroacetic acid and n-decane from Sigma; glyserolmonoo-

leate (GMO) from Matheson, Coleman & Bell. and

95

N.N'-dinitroso-N.N'-dimethyl terephthalamide from Du Pont.

W

The UV and IR spectra were obtained by the use of
a Beckman Gray 17 spectrometer and a Perkin-Elmer #57
spectrometer respectively. The mixture analysis was
performed on a Finnigan 4000 automated gas chromatograph-
EI/CI mass spectrometer with INCOS data system. Thin-layer
chromatography was carried out on Analtech silica gel G
plates of 1-mm thickness. PH measurements were made on a
Heath EU-302A digital pH/volt meter with a Sargent
combined glass electrode. For photolysis experiments. a
Hanovia medium pressure #50 W mercury lamp with a Pyrex
cooling jacket was used as the radiation source. All the
reaction vessels for photolysis were of Pyrex glass. so
that UV light less than 300 nm was absent. The sonication
of lipid-gramicidin suspensions was achieved on a Mettler
Electronics ME h.6 bath sonicator. For voltage clamp
experiments. a Heath Voltage Reference Source was used to
provide membrane potential and a Kiethley 610A Electrometer
to monitor resulting membrane current. Electrodes.
electrolytic cells and charge injection instrument were
described in Chapter III. Gas chromatogram was obtained

by use of a Varian model 1h00 FID gas chromatograph.

96

Procednxe

1. S t i of -n°t - -azido-benzoic acid NABA

2-Nitro-5-amino-benzoic acid (9.93 g. 0.05 mole)
was suspended in aqueous H2804 (3.8 M. 95 ml) at -5°C. To
the maxture was added dropwise NaNO2 (7g. 0.1 mole. in
35 ml of water at -5°C) followed by additional urea

crystals (6 g. 0.1 mole) to destroy excess NaNO After

2.
stirring the solution for 15 minutes. NaNO3 (13.1 g, 0.2
mole. in 35 ml of water at -5°C) was added slowly over a
period of 30 minutes. After N2 evolution ceased. the
mixture was stirred for additional 20 minutes and the
solid product collected on a Buchner funnel. The product
was washed once with ice water and recrystallized from
MeOH:H20 (2.1 by volume) in a yield of 3.06 g (33%). The
pKa of NABA was determined by titrating 18 mg of the
product in 150 ml of water with 0.01 N NaOH. The course of
the titration was followed by a pH meter. The pH of the
solution at the midpoint of the titration was taken as the
pKa of NABA.

The UV spectrum (0.1 M NaZHPOu. pH 7.4) showed
max 315 nm (E9300); the IR spectrum (Nujol) showed: 2120 cm-1
(5, -N3), 3uoo cm"1 (s, broad. ~0H); m.p. 163-166C(lit.

165-166 C): PKa 3.6.

97

2. Pneparation of gnamicidin-doped liposomes

Phosphatidylcholin (100 mg. 0.13 mmole. in 1 ml
of CHClB) and gramicidin (24 mg. 0.013 mmole. in 0.5 ml of
methanol) were mixed and evaporated first on a rotatory
evaporator and then by a vacuum pump. The mixture was
under vacuum for at least one hour to ensure the complete
removal of organic solvents. NABA (31 mg. 0.15 mmole. in
h ml of 0.1 M phosphate buffer. pH 10) was then added to
the lipid-gramicidin mixture. vortexed for 10 minutes and
sonicated for one hour. The resulting liposome solution
was incubated in an oil bath for 8 hours at 55°C to
establish the incorporation of gramicidin into the lipid

membrane (65).

3. Photolysis

For liposome solutions. the irradiation lasted
about 36 hours until the IR peak of azide group (-N3.
2120 cm'l) of NABA completely disappeared.

b. Sennnation of gngnicidin and its analogs fnon
liposomes

The irradiated or control liposome solution
was diluted two—fold with 0.1 M phosphate buffer (pH 10)
and applied to a Sephadex G-50 column (1.3 cm x 27 cm)
in two portions. The liposomes were eluted with 0.1 M

phosphate buffer (pH 10) and the turbid fractions

98

collected. The recovered liposome solution was then
carefully evaporated on a rotatory evaporator to about

0.5 ml followed by extractions with two 25-ml portions of
methanol. After evaporation of the solvent. the extract

was dissolved in minimal methanol and separated by thin-
layer chromatography (solvent system: methanol). The fast
moving band near the solvent front was eluted with methanol.
The eluent showed a UV spectrum and imgration destance
similar to authentic gramicidin and. thus. was gramicidin

or its analog or their mixtures.

5. Pantigl nygnolygis and denivatizgtion of
am' ' ’ d 't o

Gramicidin or its photolabeled analog were
partially hydrolyzed with 2 ml of trifluoroacetic acid in
constant distilled HCl (1:1 (V/V)) at 110°C for 1.5 hour
in an evacuated and sealed thickwall glass tube. The
contents of the tube were then transfered into a 10 ml
flask and evaporated on a rotatory evaporator.

To the resulting mixture was added ethereal
diazomethane freshly prepared by reacting 2 g of N,N'-
dinitroso-N.N'—dimethyl terephthalamide with 10 ml of 10%
methanolic NaOH in 30 ml of dry ether at 0°C. After
evaporation of excess diazomethane and ether. the methyl
esters were transformed into their trifluoroacetyl-

derivatives by treatment with 1 ml of trifluoroacetic

99

anhydride in 0.5 ml of CHZCl2 for two hours. The resulting

mixture was then evaporated and stored in the refrigerator.

6. GC-MS annlygis

The derivatized bydrolysate of control or
irradiated gramicidin was dissolved in 0.2 ml of acetone
and 1 ul of this solution injected into gas chromatograph-
mass spectrometer. Glass columns (3') filled with 3% SE-30
on Gas Chrom Q (applied Science Lab.) were used in all
experiments. The initial temperature was 80°; a linear
temperature programming rate of 8°/min. was used to a
final temperature of 290°. Electron impact (70 eV) was

exploited to obtain the mass spectra.

RESULTS AND DISCUSSION

Photolysis of NABA

The UV-Visible spectrum of NABA (Figure 4-2) is
characterized by a single peak at 315 nm. Upon photolysis
with a medium pressure mercury lamp C>BZO nm). a stable
spectrum was obtained that displayed two shoulders near
265 nm and 298 nm. and the major absorption peak was
shifted (Amax = 345 nm) toward higher wavelengths. This
red spectral shift and creation of new spectral shoulders

is attributed to a solvolysis reaction of the nitrene

100

 

 
   

 

 

0..“ ‘
o ”(I
O
I
u 0
a
b
3
‘ 00‘“
‘
J.
0.24
Photolyzed NABA
\\/
200 ° :30 - 460 500

Wavelength, nm

Fig. h-2 The effect of photolysis at2820 nm on the UV
spectrum of NABA. NABA was dissolved in 0.1M phosphate
buffer (pH 9.9) and irradiated at;azo nm for 10 minutes
until no further change in the spectrum.

101

intermediate photogenerated from NABA.

Spectral changes of NABA in infrared regions was
also observed. The sharp absorption peak of azide group
(-N3) at 2120 cm”1 is destroyed upon illumination. In
photolabeling experiments. the disappearance of this
spectral peak was monitored to signal the end of the

photolysis.

Effect of Photolysis of NABA on Gramicidin-mediated ien
Tanager:

Two techniques have been employed to investigate
the effect of photolysis of NABA on the membrane
conductivity induced by gramicidin. They are the voltage
clamp experiment and the charge injection experiment.

In voltage clamp experiments. lipid bilayers
containing gramicidin were formed and the membrane current
resulting from an externally applied potential was
monitored. As the light was turned on and off. as shown in
Figure 9-3. a small decline in membrane current in the dark
is observed. Since the antibiotic was added to the membrane
forming solution. the decline is attributed to the less
lipid-bond gramicidin to the aqueous solution. It is
evident from Figure h-3 that in the absence of NABA
irradiation does not cause any change in membrane
conductance.

In the presence of NABA, however. a remarkable

102

.ommcm msoo3dm a“ ¢m<z . o l<m<z o: .o

.Ao.e may meandeoed sa.o

muzaouuooao a>som . HmeCmuom osmpASoE .Aosmoocnc cw Hohmummaono ea
I020 sav sowpsaom msflspom osmLQSos one on venom mm; sfioaowsmpo
.ooCMpososoo mamassms so mwmhaouonm mo pomeho one nu: .wwm

.5... .2:

 

 

 

o L 0
.3 a II
a /. ..
I6 HI
3 //
.3 ‘3 All
to a.— A--- .S

 

103

change in membrane conductance is observed upon illumina-
tion. The membrane current drops more than 90% from 9.2 x

10'7A to 9.0 x 10"8

A and reaches a steady state. As the
membrane current is a manifestation of active gramicidin
channels. its reduction reflects either a decrease in the
number of open channels or a partial loss of channel
activity. Upon discontinuing irradiation. a slow recovery
in membrane conductance was registered. The membrane
current enters another fall-and-rise cycle when one more
round of on-and-off of radiation source is effected.

Similar results were also obtained from charge
injection experiments.

As discussed in Chapter III. the relaxation time
of voltage transient resulting from pulse applications to
a membrane equals the ratio of membrane capacitance to
membrane conductance. As membrane capacitance is voltage-
independent and does not vary with time. the relaxation
time is inversely proportional to the concentration of
gramicidin channels in the membrane.

In the charge injection experiments. gramicidin
was added to the aqueous solution. As shown in Figure h-h,
the relaxation time decreases as gramicidin diffuses into
the lipid phase in the early stage of the experiment. It
was observed that during this period irradiation exerts no
effect on the membrane conductance. When the relaxation

time reached a steady state. NABA was added and photolysis

101:

..pCoeHuomxo one Ho massHMon

one pm Aw.m :Q .mpmnmmonm 2H.ov mmmnm maomsam one on cocoa mm:
swoflowampw .Aosmoocus Houopmmaono RHIOEU Ray escapees one op mmasm
owhmno washangm n oomamo mm; cofipmxmaop oprHo> osmuQEoEmsmpp

one .meu cowwmmeoh omegao> so mwmhaouonm Mo poowmo one :u: .wflm

A .3... v as:

an ON ca ON op Np c

V
O

4
db

 

 

105

effected. A rapid rise in relaxation was registered.
Unlike the voltage clamp experiments. explanations
of changes in voltage relaxation time are a little more
complicated. The attachment of negatively charged nitrene
intermediates to membrane surface would inevitably modify
the double layer structure and result in changes in
membrane capacitance. In view of the five-fold increase in
decay time. it is. however. unlikely that the entire change
in relaxation time would be due to a variation in membrane
capacitance. The depletion mechanism for membrane charge
must somehow have been altered by photolysis of NABA.
There are reasons to believe that the photoinduced
blocking of gramicidin channels arises from the binding of
nitrene intermediates to the channel openings. In both'
experiments. NABA was kept in aqueous phase and denied its
access to membrane interior by use of alkaline buffer
(pH 7.6). With a channel diameter of #A plus the fact that
gramicidin does not induce anion permeability. it is quite
impossible for NABA to be present inside the pores. _
Consequently. the only place left on gramicidin channels
that could be photolabeled by nitrene intermediates is
channel termini near membrane surface. It thus appears
that the reduction of membrane conductance upon photolysis
of NABA is due to the steric hindrance introduced by the
covalent attachments of nitrene intermediates to the

channel orifice.

106

cnemical Characterization of Gramicidin Isolated from
Innagiated Liposomes

If the blocking effect of photolysis of NABA on
gramicidin-mediated ion transport is indeed a result of
the covalent binding of nitrenes to channel mouth.
structural analysis of gramicidin photolabeled while in
its pore position would then reveal information regarding
the channel conformation in lipid membranes.

Unfortunately. BLM experiments do not yield enough
material for chemical characterization. One way to circum-
vent the difficulty is to employ gramicidin-doped lipid
vesicles. The solvent-free nature of liposome bilayers
provides an additional advantage over BLM. i.e.. essentially
all the gramicidin molecules in the membrane are in the
channel form.

The search for a technique that suits our need was
limited by four factors. First. even with the liposome
approach. the amount of gramicidin available is in the
milligram range. Second. the technique should be capable
of yielding some structural information. Third. the method
should allow mixture analysis. And last. it should be
practical with resources available.

In view of these requirements. gas chromatography-
mass spectrometry appeared to be the best choice among the
various techniques. The outline of the experiments is

shown in Figure 4-5. Briefly. gramicidin-doped liposomes

107'

.CoapmNHumPomsmco Hmowsmso pom sHoHoHEmaw omaonmaouonm
mo pcsosm Hmwpsmpmpzm CH cannon oasoz page mesmswpmmxm nu: .mwm

£1.23: A 3:25:20: IIv 02.99

a,

 

3: a... 32.3.2. .0» v 3322:.” II 322...... T: 3582.:

108

.popompoc
one we popmsoupommm mmms wcflms moSomoQHH Honpsoo song oopmaomfi
cfiowowempm Mo mummhaopoz: omuwpm>wpoo ho sapwoumsopco mac 0:: .wwm

 

 

 

 

 

 

 

 

2; 36a 33.“ 8.2 26
2.9 .3 3o 8.. as
F . L b h p — b h p
/\((/\\IIa.n\III.J 1 5 v
.. Nag-"l n: p
I.
m . new
P. 3n _ 4.w"
3 DUN I k
a To ._ I
.e e a
a».
I.
.h
. Ti
H .
_
‘
1|- _
3 fl _
u
a .
e? A u
w... T
Lax—ace
'
Mel
.3
......
5.3.

 

 

13

109

ut_m av.q~ onus“ o¢"0n ov.o. on.n_ 00.6. ov.m oa.n
xtuu «at «on cow 0&3 our aa—
. _

.popompoo
one mm popmeonpoomm mmms mean: moEomomfiH oopmwomapfi sopm oopmdomw
afloflowsmam mo mammzaopoz: omuwpm>whmo wo sapwopmsopno mmu mu: .mflm

r

 

  

. . r .
-.u)
«at .<1
.

 

 

 

 

 

/.u\I 9....
....n
1
a
....w ..u
mu fem"
.
n. m
n .
A
n.
:0
1
1
d
3
. m
n.
352...: n
2:

 

. e8 8... .
I— )? r If h b b h
_ 5 an III. a
3n 2. J
l l j 7;

 

 

 

 

‘Ia ‘m ';

 

u.“

 

Io.ao.

110

.m

 

 

 

ca: .emm .Smm .oom .omm .mmH mmm :eH-oee
ca: .Hmm .mdm .omm .mma ofie ace
3mm .som .mHm .moH mmm Hm>IHm>
mmm .on .mmH mam «HeIseq
mmH .mmH Ham seq
mma .mofi mmm He>
mmfl .smH .oeH med «Ha
mad .oma was aHu
m\S .322 mmHoQO

 

 

.moHom osHsm Ho masomw oon one mum mm .Hm when;

zoowmwmzmmwmzwn mo mH moHommm opp AoH MHseuoh amHSOmHos Hmumsow one .mpsozHo

omHHHHCmoH Ho mammg mmms oHHmempomumno cam pstmz HMHsomHoz HI : oHnme

111

.cmnmopmepp ooqum>Hmoo mo sanpoomm moms one wI: .mHm

 

 

 

 

 

.7... Q3 Gum 3.0 can or" OWN 93 a?“ W:
P>_bbPDDbth :tI—rthnbthIPhhh itnDbbb.)>bIrDDb-DDDDbDtbhIbbeihbfib‘bbbbbED-DDDD-{IT-FDII
....» .... mm .— 86 a? 1% km H
.....e n

”b“
A axle :Slx .

 

 

am tea.
a
1
r 3:.
a: 3 3 2. we
’D”b”’bI—r’bFi-FD.DI- 'L.‘.t’.->$fi
Id II< q «I «H ‘1
....m mm M mnn F ... mm
nu .

sin“ 3:. «3...... . .
z" .2 re ....»
._ u a

.5 .

 

0'"..- "‘“+"
a - o '-
3 .+_
«
a
‘I’
o

r 93...:

112

were prepared and photolyzed in the presence of NABA
followed by the separation and hydrolysis of gramicidin.
After transforming the hydrolysate into volatile species.
GC-MS analysis was carried out.

The gas chromatograms for control and irradiated
gramicidin. using mass spectrometer as detector. are
shown in Figure h-6 and h-7. respectively. The compounds
identified are listed in Table 0-1. A typical mass spectrum
of an identified GC eluent is shown in Figure 4-8. which
corresponds to the gragmentation pattern of derivatized
tryptophan. It is of interest to note that the elution
order of the compounds follow that of molecular weight.
However. glycine was not separated from-alanine partly
because of their light molecular weights and partly
because of their similar structures.

As is evident from GC-MS profile. the majority of
high mass GC effluents are not identified. The reasons
are two-fold.

1. Most of the mass spectra from the high mass

fractions lack high mass peaks.

2. For eluents of similar retention time. similar

mass fragmentation patterns were observed.

The situation is further complicated by the fact
that although the retention time of each peak is very
reproducible. the number of peaks after tryptophan is
not always a constant. 1

To alleviate the first problem. chemical ionization

113

 

 

 

 

 

,4

W/

x A. A A A g A

o 4.5 2.0 18.0 27.0
Time (min)

 

 

 

 

Fig. h-9 GC trace of the derivatives of tryptophane which
has been treated with trifluoro acetic acid-const.
distilled HCl (1:1 by volume) at 110 C for 1.5 hours.

Condition : column : 3% SE-30, temperature : 80-290°C at
8°C/min.. detector : FID.

 

110

mass spectrometry was employed to elimenate the excessive
fragmentation caused by energetic electron impact.
However. probably due to source fouling. the mass spectro-
meter consistently lost its sensitivity minutes after
sample injection. As a result. no useful information was
acquired.

Figure 0-9 shows the GC trace of the acid-treated
tryptophan derivative which displays multiple peaks. The
similarity of Figure u-9 to the region between trp and
trp-leu peaks in Figures h-h and 4-5 indicates that
eluents right after trp in GC-MS chromatogram are the
decomposed products of tryptophan during acid hydrolysis.

Although the determination of the molecular
structure of GC effluents is of importance. the essence
of the study. however. is to identify those eluents that
are unique to either irradiated or control gramicidin.

Unfortunately. the comparison of the GC-MS profile
of irradiated gramicidin to that of control gramicidin
does not yield any significant difference. To reconcile
this negative result with that obtained from BLM photolysis
experiments. we postulate that the photolabeled
constituents of gramicidin are such that they could not
survive the harsh acid hydrolysis.

There is evidence to support our hypothesis. In
1978. in an attempt to photolabel membrane interior with

lipophilic phenylazides. Knowles and Bayley discovered

 

115

that photogenerated phenylnitrene exhibits certain
electrophilicity and prefer electron-rich reaction sites

to C-H bonds (67). The current consensus on gramicidin
channels is that the amino alcohol at N terminus is

situated at the membrane surface. These two thoughts lead us
to the belief that upon photolysis the nitrene intermediate
reacts with the hydroxyl oxygen at the channel opening.
pforming a rather labile O-N bond (Scheme h-l) which later

breaks down during acid hydrolysis.

 

 

 

Channel , .
—-———I—CH2—Q-H ——1-CH2-O-H —1-CH2—o
I l
' N H
o N T I
C I . z, c
Q ,.\ I/\
5 I II '\j' 'e.' c
E \\‘I’ ‘coo \I ‘000’ ‘ 0°
" NO N02
2 N02 2
Scheme 4-1

To test the hypothesis. we could either adopt a
milder protein degradation process to retain the labile
bonds or find a carbene analog that would result in an
acid-resistant etherlike linkage (C-O) to the antibiotic.
Again. difficulties were encountered.

Because of its alternating D and L amino acid
configuration. gramicidin is not subject to enzymatic

cleavage by any presently known proteinase. A search for

116

chemical reagents that would break peptide bonds in less
harsh conditions also failed.

In 1975. Smith and Knowles (68) reported the
synthesis of p-(carboxymethoxyphenyl)-3H-diazirine. Upon

photolysis. diasirine generates a carbene intermediate which

 

N 71;!
CH Cf
4‘"‘ h” A ‘¢4\." + N2
. I: l. .
‘3/ ,
I I
ocnzcoon OCHchOH
diazirine

‘readily undergoes insertion into 0-H and C-H bonds and
forms highly stable C-0 and C—C linkages. Furthermore.
diazirine is fairly stable in aqueous solutions at room
temperature and has an absorption maxima at 362 nm.
However. the reported yield (1%) of diazirine synthesis
is too low to be of any practical value.

At this point. with reluctance. we abandoned the
attempted chemical proof of the gramicidin orientation.
The attempt was. however. not without value. We have
demonstrated that the photolysis of NABA exerts a remar-
kable effect on ion transport mediated by gramicidin. It
appears that the reduction in membrane conductance arises
from steric hindrance imposed by photolabels at the channel
opening. The failure to identify the photoproducts is

attributed to the acid—labile linkage between gramicidin

117

and photolabel. It is believed that as the availability of
various photolabeling reagents expands. the photolabeling
technique will become an indispensable tool in biochemical

studies.

 

APPENDICES

APPENDIX A
PERIPHERAL SCI-EEMATIC DIAGRAMS

118

ha

.emhono

             

vzm
90.19

    

a.nc.

            

oHHMEonom snoop 2<m m:

Dadom 24m v. v

n20

 

35
.o. «o. no.

no o§

   

n.

 

V2

   

.H-< deemed

 

nN—o
¢¢¢<¢ (44¢

 

 

 

 

 

 

 

 

 

 

 

 

 

fiz'é'elilisajiija]
Q
2

38: ~ J

 

 

 

" < S a 3 l2

tic diagram.

 

1120

.eeeoeee encaseeem semen exam:

 

 

 

 

 

 

 

 

 

 

.m-e eeeeee

 

 

 

 

 

 

 

taut
urn...
32.1305. humwz
oh no! a:
gzzflw o...
0.120 D
E IaU
a ......e «.mmo .
Mo. K..- a:
t . eflmmwmnnw
mm Mum _ 100.?
all) eh... I _. ....I.u.\|e
poll-III]. #- .
IIII _ 1.
N: no no

 

APPENDIX B
PROGRAM LI ST ING S

00

00 000

200

210

20
300

310

121

MEMBRN.FNTI2

MANAGER PROGRAM FOR MEMBRANE ION TRANSPORT STUDY
COMMON/VECTOR/XSCALE.YSCALE.IDEV

COMMON/VOLAY/ XOFF.YOFF.XLAST.YLAST

COMMON/VABS/ IX.IY

COMMON/OFILE/ LUNIT

BYTE FILNAM(32).DAT(9).TIM(S)

DIMENSION ACSOO).B(100).C(10)

ABDATA ARRAY. B=BASELINE ARRAY. C=LINEARITY
CORRECTION COEFF.

INTEGER POINTS(500).OPTION.AUTO

LUNTI 8 5

** DISPATCHER **

IOUE-2OO

IGUE82OO MEANS DATA COLLECTION ROUTINE
SHOULD BE INITIALIZED.

CALL COLECT(NPTS.IPUSH.POINTS.IOUEIISROPT)
HRITE(LUNTI.200)

FORMAT(1X.' OPTION:COLLECTION DATA. AVERAGE
IILIST DATA. PLOT DATA’I' STORE DATA.
2.3ASELINE PLOT. EXIT’I'

SIEXPONENTIAL FIT'I’S?’)
READ(LUNTI.210)OPTION

FORMAT(A1)

IF (OPTION.EO.1HC) GOTO 20

IF (OPTION.EG.1HL) GOTO 40

IF (OPTION.EG.1HB) GOTO 59

IF (OPTION.EG.1HP) GOTO 60

IF (OPTION.EG.1HS) GOTO 80

IF (OPTION.EG.1HA) GOTO 100

IF (OPTION.EQ.1HF) GOTO 600

IF (OPTION.NE.1HE) GOTO 10

STOP

** DATA COLLECTION SECTION ** .

NRITE(LUNTI.300)

FORMAT(’$ENTER # OF DATA POINTS(DEFAULT=128):’)
READ(LUNTI.310)NPTS

FORMAT(I4)

IF (NPTS.LE.O) NPTS=128

URITE(LUNTI.320)

FORMAT('$AUTO OR MANUAL?')

READ(LUNTI.210)AUTO

IPUSH=O

IF (AUTO.EQ.1HA) IPUSH=1

325

00000000 00000000

326

330

50
340

330

(.0000

60

122

NRITE(LUNTI.325)
FORMAT(1X.’ SET UP FAST DATA POINTS. FAST DATA
1.RATE AND SLOW DATA RATE ')

COLLECTION RATE OPTIONS:

BITS FUNCTION

0'6 SET 0 FAST DATA POINTS COLLECTED
AFTER TRIG..MAX.= 98. (PRE-TRIG.
RECORD I 98.- 0 SET)

8'10 SET FAST DATA RATE
RATE 8 10 MHZ/(2** QSET)

11-13 SET SLOW DATA RATE (AID CONVERT
RATE). RATE 8 39.0623 KHZ/(2**
“SET)

READ<LUNTI.326)JSROPT

FORMAT(06)

CALL COLECT(NPTS.IPUSH.POINTS.IGUE.JSROPT)
GO GET DATA

DO 30 I-l.NPTS

A(I)=FLOAT(POINTS(I))

CONTINUE

GOTO 10

** DATA LISTING SECTION *o

NRITE<LUNTI.330)

FORMAT(’$ENTER DESTINATION (LP.KB):’)
READ(LUNTI.210)AUTO

IVAR-S

IF (AUTO.EG.1HL) IVAR=6
NRITE(IVAR.340)(A(I).I-1.NPTS)
FORMAT(5(4X.F7.2))
NRITE<IVAR.350)IGUE

FORMAT(’ NOISE QUE = ’.I4/2X/2X/’1 ’)
GOTO 10

** DATA PLOTTING SECTION **

LB=1
KPTS=100
GOTO 61
LB=O
KPTS=NPTS

0‘000

000

000

123

INITIALIZE PLOT

RKPTSSKPTS

IDEV'I

CALL VECTOR<XS.YS.-3)
XSCALE=RKPTSlXS
YSCALE=204S./YS

CALL VECTOR(O.1.0.1.0)

DRAW AXES

CALL VECTOR(0.0.2048..1)
CALL VECTOR(0.0.0.0.2)
CALL VECTOR(RKPTS:0.0.2)

PLOT POINTS

DO 70 I=loKPTS

X8I

Y=A(I)

IFILB.EG.1) Y-B(I)
CALL VECTOR(X.Y.1)
CALL VECTOR(X;Y.2)
CONTINUE

CALL VECTOR(0.0.0.0.4)
GOTO 10

** DATA STORAGE SECTION **

IF (BASAVG.NE.0) GOTO 81
HRITE(LUNTI.35$)

FORMAT(’ NO BASELINE--NO ACTION TAKEN’)
GOTO 10

NRITE(LUNTI.360)
FORMAT('$ENTER FILNAME ’)
READ(LUNTI.370)FILNAM
FORMAT(32A1)

CALL DATE(DAT)

CALL TIHE(TIH)

CALL ASSIGN(4.FILNAM.32)
NRITE(4.380)NPTS
FORMAT(’NU’:IS)

BASELINE SUBTRACTION (ALLIGNS NOISE GUE FIRST)
K=IGUE

J=JBQUE
IOFF=MOD(NPTS.100)

DO 84 I=1.NPTS
ACK)=A(K)-B(J)+BASVAG
K=K+1

J=J+1

83
S4

85

88

90

390
392
395
400
405
410

420

12#

IF (K.GT.NPTS) GOTO 85
IF (J.GT.IOO) J=J-100
CONTINUE

GOTO 87

KfiK-NPTS

J'J+IOO-IOFF

GOTO 83

NTERM586

LINEARITY COEFFICIENTS: (FROM POLFIT)
C(1)--0.00035
C(2)=O.96329
C(3)=-0.01033
C(4)-0.01107
C(3)--0.00271
C(6)-0.00024

DO 90 I31.NPTS

YIO.O

TERM-1.0
YI-(5.0/2048)*A(I)
LINEARITY CORRECTION
DO 88 N=1aNTERMS

'Y-Y+C(N)*TERH

TERMITERHfiYI

CONTINUE
NRITE(4.390)I.Y
CONTINUE
FORMAT('RD’.I15.F15.5)
BASVAE=BASVAG*5.0/204S
HRITE(4.392)
FORMAT('ED’)
NRITE<41395)BASVAE.IGUE
NRITE(4.400)JSROPT
FORMAT('BA'.F10.6/’GU’.I4)
FORHAT(’SR’.06)
NRITE(4.4OS)FILNAM
FORMAT('FI'.1X.32A1)
NRITE(4.410)DAT
FORMAT(’DA’.1X.9A1)
NRITE(4.420)TIH
FORMAT<’TI’.1X.SA1)
CALL CLOSE(4)

GOTO 10

125

C
C »* BASE LINE AVERAGING SECTION **
c
100 NRITE<LUNTI.300)
500 FORNAT<'sHOU MANY SCANS?:’)
105 READ<LUNTI.505>NSCANS
so: FORHAT(16)
IF (NSCANS.LE.O) NSCANSalo
DO 110 I=I.Ioo
BCI)-0.0
IIo CONTINUE
DO 180 J=1.NSCANS
CALL COLECT(100.1.POINTS.JBGUE.98)
KaUBoUE
DO 120 I=I.Ioo
3(I)-B(I)+FLOAT(PDINTS(K))
K-K+1
IF (K.GT.IOO) KaK-Ioo
120 CONTINUE
190 CONTINUE
BASAVC=0.0
DO 185 I=1.Ioo
B(I)-B(I)/NSCANS
BASAVCaBASAVC+B<I)
185 CONTINUE -
BASVACaBASAVC/Ioo.o
UBOUEsI
GOTO 10
c
c ** EXPONENTIAL FIT SECTION **
C
600 NR1TE(5.6OS)
605 FORMAT(’$ENTER FILENAME:’/’$?’)
READ<5.610>FILNAN
610 FORMAT(32A1)
NRITE(5.615)
613 FORMAT(’$ENTER INITIAL AND FINAL POINTS’)
READ(3.620)J.K
620 FORMAT(215)
CALL EXPFIT<FILNAN.U.K)
GOTO 10

END

 

R0820
R1821
R2822
R3823
R4824
R5825
SP826
PC8Z7

3
POINTR:

NPTS:
IPUSH:
SROPT:
POL:

COLECT:

PASS2:

126

.TITLE COLECT.MAC
.IDENT /V0.04/
.GLOBL COLECT

.NLIST BEXITTM
.PSECT COLECT.RN.REL

.NORD
.NORD
.NORD
.UORD
.NORD

OOOOO

HOV Q2(R5).R3
NOV R3.NPTS

‘ HOV Q4(R5)aIPUSH

HOV 6(R3).R4
HOV R4.POINTR
CLR CHR

CHP QIO(R5).#100
BLOS PASSZ

HOV “GODDATO
NOV #4OOOTCMR
CLR 010(R5)
HOV “IOOOICHR
RTS PC

CLR POL

HOV QI2(R5).R0
HOV ROISROPT
DEC R0

HOV ROODATO
HOV #400.CMR
HOV “IOORI

NOV SRDPTIRO
SNAB R0

HOV RODRQ

BIC #177770aR0
ASH R0.R1

ASH #7STR2

BIC #17777OTR2
HOV #20'R0

ASH RQTRO

 

1‘:

CHECK:

STORE:

CUE:

OVER:

POS:

ADD
ADD
SOB
HOV
TST
BEG
HOV
BIT
BEG
HOV
HOV
SOB

CLR
HOV
INC
HOV
CON
CMP
BHI
BIT
BEG
HOV
BIT
BEG
HOV
BIC
HOV
CMP
BLO
RTS

.END

127

R0.R1
*2OTR1
R1.1$
02OOOTCMR
IPUSH
CHECK
“1000.CNR
GIOOOOOTFLAG
CHECK
DATI.(R4)+
“2°OOICHR
RSTCHECK

R2
POINTRCR4
R2

(R4).R1

R1

R2:#100
OVER
“400000R1
OVER
R2:GIO(R5)
#4OOOTR1
POS
#1:POL
#174OOOTR1
R1:(R4)+
R2oNPTS
STORE

PC

 

 

128

’..-.---.=-.-.-...-..‘333:888‘-=8-.-.¢8388..3..8-8‘8838.=

TITLE: MEMB.MAC - PROGRAM TO PERFORM MEMBRANE
EXPERIMENTS AND SEND DATA TO
POP-11 THROUGH A SERIAL PORT

AUTHOR: H. Y. GUH
DEPARTMENT OF CHEMISTRY
MICHIGAN STATE UNIVERSITY
EAST LANSING. MI 48824

DATE: 30-JUL-80

~0~o§o§n§o~o~o§o~o~05oQa

I..--...-I-ra‘-SS.=88388a888:88:88.8-8‘afi....:‘t-Sa..-.3
I

ITO USE THE PROGRAM. DATA COLLECTION RATE IS DEPOSITED
TAT ZOOOH AND CHARGING VOLTAGE AT ZOOIH.

IO788>13HHZ

30688>10MHZ

3058-) DMHZ

3048-) 2MHZ

308-8) IHHZ

.ADDRESS REGISTER FUNCTION

38000/100000 RATE DATA RATE

58040/100100 CMD BITO===>I:CLEAR FLIP FLOPS

; AND COUNTERS.

I BIT1-==)1:START EXP.
38080/100200 VOLT CHARGING VOLT. (O-1OV)
.80CO/100300 SNITCH BITOa==>1zOPEN FEEDBACK SN

I BIT288=>12CLOSE CHARG. SW

3F000/170000 UDATA USART88>DATA REG.
3F001/170001 UCMD USAR ==>CMD REG.
:C000/140000 DATADD FAST RAM(1K)
3E000/160000 RAM(4K)

I

5 —‘ .— - ————————————————————————————————————————————
RATE 8 100000

CMD 8 100100

VOLT 8 100200

SNITCH s 100300

UDATA 8 170000

UCMD 8 170001

DATADD = 140000

.8 160000 TSTARTING ADD. OF 4K RAM

 

 

 

START: LXI SP.20302 sINITIALIZE STACK POINTER

MVI A.116 iSET UP USART - ASYN MODE
STA UCMD
MVI A.S iSET UP USART - ENABLE TX

STA UCMD iAND RX

 

 

 

I INITIALIZATION OF REGISTERS
’ .......................................................

XRA A .CLEAR CMD

STA CMD

STA VOLT IAND VOLT

STA SWITCH iAND SWITCH

MVI A.01 ICLEAR COUNTERS AND FLIP
STA CMD sFLOPS

MVI B.05 :ALLOW COMPLETE ACTION
CALL DELA

 

 

 

LXI H.20000 :GET RATE FROM 2000(H)

MOV A.M

STA RATE

LXI H.20001 iGET CHARGING VOLTAGE FROM
MOV A.M 52001(H)

STA VOLT

MVI A.04 sCHARGE UP CAPCITOR

STA SWITCH

MVI 8.10 iDELAY SOMEWHAT TO ALLOW
CALL DELA iCOMPLETE CHARGING

MVI A.01 iOPEN CHARGING AND FEEDBACK
STA SWITCH TSWITCH

MVI 3.5 aALLOW COMPLETE ACTION
CALL DELA

 

‘ ———————————————————————————————————————————————— __v .-

3 START EXPERIMENT AND DATA COLLECTION
’ ----------------------------------------------------------

MVI A.02 iDUNP CHARGE AND COLLECT
STA CMD sDATA
’ .-.-Quasaaaaa=======ar:= ======a==a=====a==saaaaazaauaasaa

TRANSFER DATA TO 11

 

; ------

 

MVI 8.100 .GET TIME DELAY
MVI D.4 .SET BYTE COUNTER: 1K BYTES
MVI E.0

CALL RD1 ICHECK PROMPT FROM 11

 

TRANSFER DATA BLOCK

LXI H.DATADD iSET UP DATA BLOCK ADD.
MOV A.M iFETCH A DATA BYTE
CALL SDCO ;AND SEND IT OFF TO 11
CALL DELA IDELAY FOR A WHILE

130

INX H IUPDATE ADD. POINTER

DCX D IUPDATE BYTE COUNTER

MOV A.D TALL THE BYTES SENT ?

ORA E

JNZ MOBYT :IF NO. SEND ANOTHER BYTE
RST 1 ISOFT RESET

’ .........................................................
l SDCO - SEND OUT A BYTE AND CHECK ECHO

, ................................. __ === —- _—.__

'SDCO: STA UDATA iSEND BYTE OUT
RD1: LDA UCMD sCHECK ECHO
ANI 2 )RXRDY FLAG SET?
JZ RD1 INOI CHECK AGAIN
LDA UDATA IVES: RESET FLAG
RET -

DELA - KEEP CPU IN LOOP FOR SOME TIME
A.B. C. ARE USED

‘0‘”-

' DELAL MOV A.D
DEL A0 : MOV C I B
DELAI: OCR C

JNZ DELAI
OCR A

JNZ DELAO
RET

 

fiwmhofio

RDUS:

131

LOAD.MAC - READ MESSAGES FROM POP-11 AND STORE IT
INTO MICRO’S MEMORY WHICH STARTS AT E000(H)

/160000(O).
UCMD8170001
UDATA-170000
.8120000
LXI SP.20302
MVI A.116
STA UCMD
MVI A.5
STA UCMD
LXI H.160000
LDA UCMD
ANI 2
JZ RDUS
LDA UDATA
MOV M.A
STA UDATA
INX H
JMP RDUS

.END

LOAD.OBJ STARTS AT A000/120000(EPROM).

iINITIALIZE STACK POINTER
iDEFINE USART MODE:

iASYN MODE

IENABLE TX AND RX

sSET UP ADD. POINTER OF
iMESSAGE BLOCK. AND READ
sUSART STATUS. AND MASK OUT
:RXRDY FLAG. AND IF RX NOT
TREADY. BACK TO RDUS. AND
iREAD THE BYTE AND STORE IT
sECHO THE BYTE

iUPDATE ADD. POINTER

iGET NEXT BYTE

132

Cunt-Ilaszzaasucasaaa==a======ga=aazsszas=aaasaasasasuz
CV DO“NLD.FTN ' READ A BLOCK OF DATA AND SEND IT
.c TO A RENOTE TERMINAL

c.------.=3b-..88..3383? 2=8:38-88883:8-..‘.‘....33.....

BYTE IDATA(512).F(32).NAM(8).TT5(8)

INTEGER*2 NBYTE.ICNT.IERCN.ITABLE(96)

DATA LUNIN/1/.LUNTIN/2/.NAM/8*’ '/

DATA TT5/2*’T’.’5’.’:’.4*’ ’/
c ......................................... _ _ _
C GET FILENAME AND SET UP I/O PORT; DEFAULT PORT
C - TT5:
C

 

 

 

5 WRITE(5.150)
150 FORMAT(’$NAME OF THE FILE THAT WILL BE COPIED
1.INTO THE ARRAY ’)
READ(5.110) F
110 FORMAT(32A1)
WRITE(5.130)
130 FORMAT('$I/O PORT (DEFAULT - TT5: ) : ’)
READ(5.135) NAM
135 FORMAT(8A1)
IF (NAM(1).NE.32) GO TO 140
DO 200 I81.8
'200 NAM(I)8TT5(I)
140 CALL ASSIGN(2.NAM)
c .......................................................
C OPEN THE FILE - READ MODE
c ............................. _ _ ......—
CALL UNFORM(0.IERR. ITABLE. F. 0. 3.LUNIN)
IF (IERR NE.0) GO TO 9001
NERR80
MBYTE 80
NRD80
120 NBYTE 8 512 ‘
READ ONE RECORD OFF THE FILE INTO ARRAY ’IDATA'
.RETURN THE LENGTH OF THE RECORD IN NBYTE

 

 

 

 

0000

CALL UNFORM(1.IERR. ITABLE.IDATA.NBYTE)
IF (IERR.EG.-1) GO TO 1100
IF (IERR.GT.0) GO TO 9000
IF (NBYTE.NE.2) GO TO 60
GO TO 120

1100 WRITE(5.8500)

8500 FORMAT(’ END OF FILE’)
GO TO 8503

9000 WRITE(5.8501)

8501 FORMAT(’ ERROR IN READ’)
GO TO 8503 -

133

 

 

c ......................................................
C SEND DATA TO THE REMOTE TERMINAL
c _ ............................................
CALL SEND(IDATA.NBYTE.IERCN.LUNTIN)
NRD8NRD+1
WRITE(5.70) NRD
70 FORMAT(’ NUMBER OF RECORDS READ : ’.I3)
NERR8NERR+IERCN
GO TO 120
c .......................................................
C MORE FILE TO OPEN? IF YES. GO BACK AND DO IT
C ALL OVER AGAIN
c ....................................... _ =
15 WRITE(5.300) NERR.MBYTE

300 FORMAT(’ TOTAL NUMBER OF TRANSMISSION ERRORS
1.8 ’.I6/’TOTAL NUMBERS OF BYTES 8 ’.I6)
WRITE(5.90)

90 FORMAT(’$FILE IS CLOSED; MORE FILE TO OPEN
1.? Y OR N ’)

READ(5.95) B

95 FORMAT(1A1)

IF (B.EG.’Y’) GO TO 5
CALL CLOSE(2)

STOP

END

9001 WRITE(5.8502)
8502 FORMAT(’ ERROR OPENING FILE’)
8503 CALL UNFORM(2.IERR.ITABLE)

 

GO TO 15
c_= ........................................
C ADJUST BYTE COUNT AND
C DATA ARRAY
c ______________________________________________________
60 NBYTE8NBYTE‘2
MBYTE8MBYTE+NBYTE

DO 50 I81.NBYTE
50 IDATA(I) 8 IDATA(I+2)

13L.

SUBROUTINE SEND (IDATA.NBYTE.IERCT.LUNTIN)

c-.8...-.= ‘33--=8.88388=‘ :::===aa==a===zaataxasscs=====

 

C
C SEND.FTN - SEND INFORMATION TO A REMOTE
C TERMINAL
c --.-..I‘: 33-.-8==88.3.=:==g===-.=-‘a==8====.==‘=3382::
c
C IDATA - ARRAY OF DATA TO BE SENT
C NBYTE - NUMBER OF DATA BYTES
C IERCT - NUMBER OF TRANSMISSION ERRORS
c---——--—-——---——--——————--——————---—--———---_—-_--_-A
INTEGER*2 ISBTN(2).IPARTN(6).IERCT
BYTE LBUFF(2).IDATA(NBYTE).NBUFF.ISERR
EGUIVALENCE (ISERR.ISBTN(1))
DATA RRRR.WWWW/’TOII’.’TOBS’/
DATA IEFNTN/IO/
DATA IOATT.IODET.IORPR/OI400.02000.04432/
c —————————————————————————————————————————————————————
C ATTACH I/O PORT.SET UP ERROR COUNTER
C .....................................................
CALL,NTOIO<IOATT.LUNTIN.IEFNTN..ISDTN.IFARTN.
1.IDERTN)
IERCT-o
c .....................................................
C SEND OUT DATA AND GET ECHO
c ____________________________________________
IFARTN<5>=1
CALL GETADR (IPARTN(1).LBUFF(1))
IPARTN(2)=1
DO 10 I81.NBYTE
LBUFF<1)=IDATA(I)
NBUFF=IDATA(I)
CALL GETADR (IPARTN(4).IDATA(I))
CALL WTGIO(IORPR. LUNTIN.IEFNTN..ISBTN.IPARTN.
1.IDERTN)
IF (IDERTN.NE.1.0R.ISERR.NE.1)
WRITE(5.BSOO) uuwu.IDERTN.ISDTN
8500 FORMAT (’ TRANSMISSION ERROR ’.A4.308)
IF (NBUFF.EG.LBUFF(1)) GO TO 10
IERCT=IERCT+1
10 CONTINUE
C .....................................................
C DETACH AND CLOSE I/O PORT
C ....................................................

CALL WTGIO(IODET.!_UNTIN.IEFNTN..ISBTN.IPARTN.
1. IDERTN)

RETURN

END

135

c ....----8 83-38883..-8C2888038888 ‘....83.....-..--..'

 

 

 

 

 

 

 

C TITLE: DATAIN.FTN - PROGRAM TO FETCH DATA
C FROM A REMOTE INTELLIGENT TERMINAL
C THROUGH USE OF ’LISTEN’
C AUTHOR: H. Y. GUH
C DEPARTMENT OF CHEMISTRY
C MICHIGAN STATE UNIVERSITY
C EAST LANSING. MI 48824
C
C DATE: 31-JUL-80
C
c ......I.Sa..'.8=8...883 38383888333888383=..8888
REAL BUFF(1024).POINT(1024).PHYS(8)
INTEGER*2 NBYTE.IDATA(1024)
BYTE ITIME(8).IDATE(10).F(32).NAM(8).TT5(8)
DATA IDATE(10)/’ ’/.NAM/8§' '/
DATA TT5/2*’T’.’5’.':’.4*’ ’/
DATA IOATT.IODET.IORAL
l01400.02000.01010/
DATA LUNTIN/I/
c — ......................... —_ — ........
C SET UP I/O PORT; DEFAULT PORT - TT5:
c .................... ===__ _ — __ —_
WRITE(5.130)
130 FORMAT(’$I/O PORT(DEFAULT - TT5:) : ’)
READ(5.135) NAM
135 FORMAT(8A1)
IF (NAM(1).NE.32) GO TO 140
DO 200 181.8
200 NAM(I)8TT5(I)
140 CALL ASSIGN(1.NAM)
70 WRITE(5.100)
100 FORMAT(’$COLLECT. PLOT. STORE. EXIT .
BASELINE: ’)
READ(5.110) A
110 FORMAT(A1)
IF (A.EG.’B’) GO TO 500
IF (A.EG.’C’) GO TO 80
IF (A.EG ’P’) GO TO 120
IF (A.EG.’S’) GO TO 170
IF (A.EG.’E’) GO TO 450
GO TO 70
c = ..........................................
C FETCH DATA
c ................... _ __ _ _____ __
80 CALL LISTEN<LUNTIN.IDATA.NBYTE.IERR)

GO TO 70

136

 

000

500

505

510

000

CALL LISTEN(LUNTIN.IDATA.NBYTE.IERR)
SUM80

DO 505 I81.NBYTE
SUM8FLOAT(IDATA(I))+SUM
BASAVE8SUM/FLOAT(NBYTE)

WRITE(5.510) BASAVE
FORMAT(’BASELINE AVERAGE 8 ’.F10.6)
GO TO 70

PLOT DATA ARRAY

 

120

300

330

340

370
380

OD 300 I81.NBYTE

BUFF(I)8FLOAT(IDATA(I))

POINT(I)8FLOAT(I)

PHYS(1)80.0

PHYS(2)88.0

PHYSC3)8O.0

PHYS(4)86.0

CALL ASSIGN(3.’TTO:’)

IDDEV81

WRITE(5.330) NBYTE

FORMAT(’$NUMBER OF POINTS: MAX. ’.I4.’ ’)
READ(5.340) NPNT

FORMAT(I6)

IMODE=1

PHYS(5)8O.0

PHYS(6)=FLOAT(NPNT)

PHYS(7)80.0

PHYS(8)=256.0

CALL PLOT(POINT.BUFF.NPNT.PHYS.IDDEV.IMODE)
WRITE(5.370)

FORMAT(’$PLOT DATA ON LPt? N OR Y ’)
READ(5.380) A

FORMAT(AI)

IF (A.NE.'Y’) GO TO 70

IDDEV83

CALL PLOT(POINT.BUFF.NPNT.PHYS.IDDEV.IMODE)
GO TO 70

137

 

150 FORMAT(’$FILENAME: ’)
READ(5.160) F
160 FORMAT(32A1)
F(32)80
OPEN (UNIT82.NAME=F.TYPE8’NEW’)
CALL TIME (ITIME)
CALL DATE (IDATE)
WRITE (2.8501) IDATE.ITIME
8501 FORMAT (’3 ’.10A1.8A1)
WRITE(5.420)
420 FORMAT(’$DATA COLLECTION RATE AND CHARGING
1 VOLTAGE ’.I4)
READ(5.430) IRATE.IVOLT
430 FORMAT(2I4)
WRITE(2.440) IRATE.IVOLT
440 FORMAT(’I DATA RATE(MHZ) ’.I4/’.
1 CHARGING VOLTAGE ’.I4)
DO 40 I81.NBYTE
BUFF(I)8FLOAT(IDATA(I))

40 WRITE(2.50) FLOAT(I).BUFF(I)
50 FORMAT(’RD’.2E15.6)
WRITE(2.60)
60 FORMAT(’ED’)
WRITE(2.55) BASAVE
55 FORMAT(’BA’.F10.6)
CLOSE (UNIT82)
GO TO 70
450 CALL CLOSE (1)
STOP

END

138

SUBROUTINE LISTEN (LUNTN.IDATA.MBYTE.IERR)

c...‘8=..833'=8==:===8=======8=============8====8=8

00000000000000000000000000000000000

C——-_

C

C----

TITLE: LISTEN.FTN - SUBRUTINE TO INPUT AN
ARRAY OF BYTES FROM A REMOTE
INTELLIGENT TERMINAL

AUTHOR: T.V.ATKINSON
H.Y.GUH
DEPARTMENT OF CHEMISTRY
MICHIGAN STATE UNIVERSITY
EAST LANSING. MI 48824

DATE: 25-JUL-80

 

LOGICAL UNIT LUNTN MUST BE ATTACHED BEFORE
ENTERING

ARGUMENT LIST.

LUNTN iLUN FOR PORT FOR REMOTE
DATA INPUT

IDATA :INTERGER BUFFER TO
RECEIVE DATA

MBYTE iNUHDER OF BYTES OF DATA
TO BE RECEIVED

IERR sERROR FLAG (80 FOR NO
ERROE)

INTEGER*2 IPARRX(6).ISBRX(2).IDATA(1024)
INTEGER*2 NBYTE.KDATA(1024)

BYTE LBUFF(1024).ISERX.ICSOT.IBUFF(2048)

EGUIVALENCE (ISERX.ISBRX(1)).(LBUFF(1).NBYTE)

EGUIVALENCE (IBUFF(1).KDATA(1))

DATA ICSOT/’*’/.IEFNRX/10/

DATA IOATT.IODET.IORAL.IOWAL.IORPR
/Ol400.02000.0lOI0.00410.04410/

GET NUMBER OF BYTES

-----_-‘----—-----_-..O-O.“----‘---—----‘----—--

CALL WTGIO(IOATT.LUNTN.IEFNRX..ISBRX.IPARRX
1..IDERRX)
WRITE(5.200)

210

c ......

C

C ——————

139

FORMAT(’ WTGIO ATTACHED’)

CALL GETADR(IPARRX(1).LBUFF(1))
WRITE(5.220)

FORMAT(’$NUMBER OF BYTE --’)
READ(5.230) MBYTE

FORMAT(I4)

 

CALL GETADR(IPARRX.ICSOT)

IPARRX(5)81

IPARRX(2)8MBYTE

WRITE(5.210)

FORMAT(’ GOING TO CALL WTGIO-RPR’)

CALL GIO(IORPR.LUNTN.IEFNRX..ISBRX.IPARRX
1..IDERRX)

IF (ISERX.NE.1.0R.IDERRX.NE.1) GO TO 9000
WRITE(5.140)

FORMAT(’ WAITING TO KILL: ’)

READ (5.9878) III

FORMAT (IS)

CALL GIO("12.LUNTN.IEFNRX..ISBRX.IPARRX
1..IDERRX)

WRITE (5.9868) ISBRX

FORMAT (’ ISBRX: ’.208)

CONSTRUCT DATA ARRAY - CONVERT ONE-BYTE
DATUM INTO TWO BYTE INTEGER

WRITE(5.130)

FORMAT(’ DATA ARE IN’)

DO 110 I81.MBYTE

J82*I

K8J-1

IBUFF(K)8LBUFF(I)
IBUFF(J)8O
IDATA(I)=KDATA(I)

CONTINUE

WRITE(5.120) (IDATA(I). I81.MBYTE)
FORMAT(’ IDATA8’.5(1X.I4))
GO TO 9999

IERR81

WRITE(5-170)

FORMAT(’ ERROR FROM WTGIO’)

CALL WTGIO(IODET.LUNTN.IEFNRX..ISBRX.IPARRX
1..IDERRX)

RETURN

END

0000000000000000

000

605
610
615

620

1h0

EXPFIT.FTN
PROGRAM TO LINEARIZE DECAYING EXPONENTIAL DATA
AND CALL THE LINEAR LEAST SGUARES SUBROUTINE

AUTHOR: H.Y.GUH
DATE: AUG-80

READS A STANDARD ENKE FILE
RETURNS THE SLOPE. INTERCEPT. AND THEIR STD. DEV.
CREATES A FILE REPRESENTING THE FIT TO EXP(-X)

J : INITIAL POINT
K : FINAL POINT

DIMENSION X(1024).IBAD(50)

BYTE FILNAM(32). DAT(9). TIM(8).FILE(32)
WRITE(5.605)

FORMAT(’$ENTER FILENAME: ’)
READ(5.610)FILNAM

FORMAT(32A1)

WRITE(5.615)

FORMAT(’$ENTER INITIAL FINAL POINTS: ’)
READ(5.620)J.K

FORMAT(2I5)

 

300

 

DO 300 I8J+2.K
DEL8X(I-1)-X(I)
DELTA-(ABS(DEL))/X(I-1)

IF (DELTA.LT.0.15) GO TO 300
X(I)82*X(I-1)-X(I-2)
IBAD(N)8I

N8N+1

CONTINUE

1A1

 

C —————————— . ——————————————————————————————————————————————
C FIT EXPONENTIAL DATA
c _ _ E- .............................................
CALL TLLSG(J.Y.XDEV.YDEV.1)
N81
DO 20 I8J.K
IF (I.EG.IBAD(N)) GOTO 25
Y8ALOG(X(I)-BASE)
Z8FLOAT(I-J)
YDEV83.2/(350*X(I))
CALL TLLSG(Z.Y.XDEV.YDEV.2)
GO TO 20
25 N8N+1
20 CONTINUE

CALL TLLSG(Z.Y.XDEV.YDEV.3)
WRITE(D.104)Z.Y.XDEV.YDEV
104 FORMAT(' SLOPE 3’.F10.7.' INTERCEPT 8’.F10.7.
' Io'SDEV 8'.FI°.7.' IDEV".F10.7)
NRITE(5.200)
200 FURNAT(1X////' STORE EXPFIT DATA? - TYPE Y 0R N’)
READ(5.205)IOPT
205 FORMAT(AII
IF (IOPT.EG.IHY) GOTO 210
A GOTO 110
210 HRITE(5.215)
215 FORMAT(IX//’ ENTER DEV AND FILENANE'/' ')
READ(5.220)FILE
220 FORMAT(SZAI)
CALL ASSIGN(4.FILE.32)
A=EXP(Y)
DO 30 I‘J: K
V‘A*EXP(Z*(I“J))+BASE
URITE(4.IOD)FLOAT(I).V

105 FORMAT(’RD’.2F15.7)

30 CONTINUE
WRITE(4.106)Z.Y.XDEV.YDEV

106 FORMAT(’ED’/’i’.4F10.7)
WRITE(4.107)FILNAM

107 FORMAT(32A1)
CALL CLOSE(4)

110 STOP

END

0000000000

...
O

200
110

20
120

130

40
140

50
150

60
160

70
170

80
180

‘ 114.2

SUBROUTINE DSTRIP(FILNAM.D.NPTS.ICUE.ISROPT
.BASE.DAT.TIM)

BYTE LINE(80). FILNAM(32). DAT(9). TIM(8)
DIMENSION D(1024)

EGUIVALENCE (JFLAG.LINE(1))

FLAGS: NU 8 DATA POINTS

RD 8 REAL DATA. FIELD WIDTH 8 15
GU 8 NOISE INDEX PULSE

SR 8 SWITCH REGISTER OPTIONS

DA 8 DATE

TI 8 TIME

ED 8 END OF DATA BLOCK

K80

CALL ASSIGN(4.FILNAM.32)
READ(4.110.END8200)LINE

IF (JFLAG.EG.2HRD) GOTO 30
IF (JFLAG.EG.2HNU) GOTO 20
IF (JFLAG.EG.2HGU) GOTO 40
IF (JFLAG.EG.2HDA) GOTO 50
IF (JFLAG.EG.2HTI) GOTO 60
IF (JFLAG.EG.2HSR) GOTO 70
IF (JFLAG.EG.2HBA) GOTO 80
GOTO 10

CALL CLOSE(4)

RETURN

FORMAT(BOAI)
DECODE(7.120.LINE)NPTS
FORMAT(2X.I5)

GOTO 10

K8K+1
DECODE(32.130.LINE)D(K)
FORMAT(2X.15X.F15.5)

GOTO 10 ‘
DECODE(6.140.LINE)ICUE
FORMAT(ZX.I4)

GOTO 10
DECODE(12.150.LINE)DAT
FORMAT(3X.9A1)

GOTO 10
DECODE(11.160.LINE)TIM
FORMAT(SX.8A1)

GOTO 10
DECODE(8.170.LINE)ISROPT
FORMAT(2X.O6)

GOTO 10
DECODE(12.180.LINE)BASE
FORMAT(2X.F10.6)

GOTO 10

END

000000000000000000

...

11.3

SUBROUTINE TLLSG(X.Y.XDEV.YDEV.MODE)
LINEAR LEAST SGUARES FOR Y8MX+B

MODE:
IBZERO SUMS
28ACCUMULATE SUMS
38CALCULATE VALUES

VALUES GIVEN: X. Y. YDEV (8 ERROR ESTIMATE)
.MODE ‘ '

VALUES RETURNED:
X8SLOPE
Y8INTERCEPT
XDEV8SLOPE STANDARD DEVIATION
YDEV8INTERCEPT STANDARD DEVIATION

GO TO (1.2.3) .MODE
A180.
A280.
3180.
B280.
C180.
C280.
RETURN

SIG8YDEV

A18A1+X**2/SIG**2
A28A2+XISIG**2
B18B1+XISIG**2
B28B2+1/SIG**2
C18C1+(Y*X)/SIG**2
C28C2+Y/SIG**2

RETURN

D8A2*B1-B2*A1
X8(BI*C2-B2*C1)/D
XDEV8SGRT(-B2/(A2**2-A1*B2))
Y8(A2*C1-A1*C2)/D
YDEV8SGRT(-A1/(A2**2-A1*B2))
RETURN

END

000000

REFERENCES

13.

1h.

15.

16.
17.

REFERENCES

Hodgkin, "The Conduction of the Nervous Im ulse",
Liverpool University Press, Liverpool (196 ).

Ashley in "Membrane and Ion Transport", Vol. 2,
Ed. Bittar, pp. 1-23, Wiley, NY (1970).

Lehninger. "Biochemistry", Ch. 28, Worth, NY (1975).

F. Alvarado, and A. Mahmood, Biochem., V. 13 (1A),
2882 (1974).

Jain, "The BimOlecular Lipid Membrane". Ch. 7,
Van Nostrand, NY (1972).

E. Overton, Vjsehr. Naturf. Ges. Zurich, g3, 88 (1899).
I. Langmuir. J. Amer. Chem. Soc., 32, 18h8 (1971).

E. Gorter, F. Grendel, J. Exptl. Med., 31, 439 (1925).
J. Danielli, H. Davson, J. Cell Phys., 5, #95 (1935).
S. Singer, and G. Nicholson, Science, 175, 720 (1972).

A. D. Bangham. M. W. Hill,-N. G. Miller, Methods Memb.
Biol., 1, 1 (197A).

P. Mueller, D. Rudin. H. Tein, W. Wescott, Nature,
12%. 979 (1962).

H. T. Tien, "Bilayer Lipid Membranes (ELM): Theory
and Practice", Ch. 6-7, Marcel Dekker Inc., NY (1974).

H. T. Tien, "Photosynthesis in Relation to Model
Systems". Ed. J. Barber. pp. 116-173. Elsevier/North
Holland Biochemical Press. Amsterdam (1979).

H. T.)Tien, Separation Science and Tech., 15, 1035
1980 .

W. Knoll, G. Stark, J. Memb. Biol., 31, (1977).
R. Benz, P. Lauger, J. Memb. Biol., 21, 171 (1976).

1AA

18.

19.
20.

21.
22.
23.
2#.

25.

26.
27.

28..

29.

30.

31.

32.

33-

3a.
35.
36.
37.
38.

145

F. Gambale. A. Gliozzi. M. Robello. Biochim. Biophys.
Acta 32.. 325 (1973L

W. Knoll, G. Stark, J. Memb. Biol., 25, 2#9 (1975).

G. Stark, R. Benz. P. Lauger, Biophys. J. , 11, 981
(1971)

D. Levitt, Biophys. J., 22, 221 (1978).
B. Hille, J. Gen. Physiol., 66, 535 (1975).
P. Lauger, Biochim. Biophys. Acta.. 355. #93 (1976).

E. Neher, J. Sandblom, G. Eisenman, J. Memb. Biol..
59. 97 (1978).

E. Neher, J. Sandblom, G. Eisenman, J. Memb. Biol..

S. Hladky. J. Harris. Biophys. J., 1, 535 (1967).
P. Lauger. Biochim. Biophys. Acta., 311, #23 (1973).

J. Hagglund, J. Sandblom, B. Enos, G. Eisenman.
Biophys. J., 2_. 26a (1978).

A. Finkelstein, A. Cass, J. Gen. Physiol. 52. 1#5
(1968)

M. Shemyakin. V. I. Ivanov, V. K. Antonov, J. Memb.
B10109 19 “’02 (1969)'

G. Szabo, G. Eisenman, S. Ciani, J. Memb. Biol.,
1. 3#6 (1969).

E. Bamber , K. Janko, Biochim. Biophys. Acta., #26,
#47 (1976?. .

S. Schein, M. Colombim. A. Finkelstein, J. Memb. Biol.,
0. 99 (1976)-

K. F. Schnell, J. Memb. Biol...lZ.99 (1976).

M. Eisenberg, J. Kolb, J. Memb., 13, 1#3 (1973).
G. Boheim, H. Kolb, J. Memb. Biol.. 38. 99 (1978).
S. Hladky, D. A. Haydon, Nature, 225, #91 (1970).

B. Hladky, D. A. Haydon, Biochim. Biophys. Acta.,
22. 294 (1972)

39.

#0.

#1.

#2.

#3.

##.

1+5.
#6.
#7.

#8.

#9.

50.

51.

52.

5#.

55.

56.

1#6

S. Krasre, G. Eisenman, G. Szabo, Science, 12#, #12
1971 .

H. Eyring, B. I. Zwolinski, C. E. Reese, J. Phys.
Chem.. 53. 1#26 (19#9)

R. D. Hotchkiss, R. J. Dubos, J. Biol. Chem., #1, 155
(19#1).

R. Sarges, B. Witkep, J. Amer, Chem. Soc. , _Z, 2011
(1965L

V. B. Myers, D. A. Haydon, Biochim. Biophys. Acta.,
22. 313 (1972)

D. C. Tosteson, T. E. Andreoli, M. Tieffenberg. P. Cook,
J. Gen. Physiol., 5;, 3733 (1968).

M. C. Goodall. Biochem. Biophys. Acta.. 212. #71 (1970)-
E. Bamberg, P. Lauger. J. Memb. Biol.. 11. 177 (1973).

W. R. Veatch, M. Ersen erg, L. Stryer, R. Mathies,
J. Mol. Biol., 22, 75 1975).

A. Kolb, P. Lauger. E. Bamberg, J. Memb. Biol. , 20,
133 (1975)

R. Koeppe. J. M. Berg, K. 0. Hodgson, L. Stryer,
Nature. 2,22! 723 (1979)

P. W. Urry, Proc, Natl. Acad. Sci. U.S.A. Q2,
1610 (1972).

R. J. Bradley, D. W. Urry, K. Okamoto. R. Rapaka,
Science. 200 .435 (1978L

W. R. Veatch, E. T. Fossel. E. R. Blout, Biochemistry,
11. 5219 (1974).

W. R. Veatch, E. R. Blout, Biochemistry. _3. 5257
(197#)

E. T. Fossel. W. R. Veatch, A. Yu, Biochemistry, _3.
52#6 (1979)

H. J. Ap e1. E. Bamberg, H. Alpes, P. Lauger, J. Memb.
Biol.. 3_. 171 (1977L

E. Bamberg, H. J. Apel, H. Alpes, Proc. Natl. Acad.
Sci. U. SA .Z_. 2#02 (1977)-

57-

58.

59.

60.

61.

62.
63.

6#.
65.

66.
67.
68.

69.

70.

71.

72.

73-

7a.
75-

1#7

D. W. Urry, M. C. Goodal, J. D. Glickson, D. F. Meyers,
Proc. Natl. Acad. Sci. U.S.A., 68, 1907 (1971).

S. Weinstein, B. A.. Wallace, E. R. Blout, J. S. Morrow,
W. Veatch, Proc. Natl. Acad. Sci. U. S. A., Z_,
4230 (1979)

J. R. Knowles, Accounts of Chem. Res., 5, 155 (1972).

J. M. Stadel. D. B. Goodman, R. E. Galardy, H. Rasmussen,
Hiochemistry. _24 1#03 (1978L

R. E. Galardy, L. C. Craig, J. D. Jannieson, M. P.
Printz, J. Biol. Chem., 2# 2, 3510 (197#).

T. Bercovici, C. Gitler, Biochem., 31, 1#8# (1978).

P. R. Matheson, E. Vanwart, A. W. Burgess, L. I.
Weinstein, H. A. Scheraga, Biochemistry, 16, 396 (1977).

R. B. Pepinsky, G. W. Feigenson, Anal. Biochem.. §é.
512 (1978).

D. W. Urry, A. Spisni, A. Khaled, Biochem. Biophys. Res.
Comm., 88. 9#0 (1979L

J. V. Staros, F. M. Richards, Biochem., 33, 2720 (197#).
H. Bayley, J. R. Knowles, Biochemistry, 12, 2#1# (1978).

R. A. Smith, J. R. Knowles, J. Chem. Soc., Perkin
Trans.,-2, 686 (1975).

E. 6Bamberg, R. Benz, Biochem. Biophys. Acta.,
_2_ 570 1976L

H. J. Apeel, E. Bamberg, H. Alpes, P. Lauger, J. Memb.
Biol., 1;, 171 (1977).

S. M. White, T. E. Thompson, Biochem. Biophys. Acta.,
3 3. 7 (1973)

R. Fettiplace, D. M. Andrews, D. A. Haydon, J. Memb.
Biol.. 5. 277 (1971L

R. Benz, 0. Frolich, P. Lauger, M. Montal, Biochim.
Biophys. Acta.. 324. 323 (1975L

v. A. Nicely, J. L. Dye, J. Chem. Ed.. fig. #43 (1971).

S. V. Sychev, N. A. Nevskaya, S. T. Jordanov, E. N.
Shepel, Bioorg. Chem. ,42, 121 (1980).

148

76. S. Feldberg. G. Kissel. J. Memb. Biol., 29. 269 (1975).
77. T. A. Last, 0. G. Enke, Ana1.-Chem.. £2. 19 (1977).

78. ?. A.)Last, Ph. D. Thesis, Michigan State University
1977 -