I V"
‘ 4 ,«m‘fimm

Mug. 4 -
é;- " ‘

 

1'9 44.. u :wm fr“
1:.“13‘m‘.

..

<.

. mum
4

“w

:‘I

an-

, ‘Kv

 

 

 

 

 

195?.

b
1‘

......

 

V mar

 

‘n....

wcnm

“ .V. g ml
4‘“
.nuu

 

.r
":2

”g,

n.
“q? ~
:3 if?

035
1.1

m.
lane-u

w.‘ ..
.

$3395
Mi???
~31, 1;-

h
éfii‘fig; a3! .
'4

 

.o
‘ I ‘~. I

v‘ 'r,
31.22.913‘4
‘ziv' : l
’5“: ‘ :
go: -"
t."

   

Z.

I ‘-
c z
15%}
"i1?
‘ P

S!
i p'

..._.I‘
59“

3!

I
l

 

1'74

8

"3:.
gr}

"Ci-‘7

l ‘ .‘ 51!: 'u ‘
53%,; -
kr‘ $38313“,

 

1

 

91;}; .
%‘."§. Ii
2.? .,g~
'I .fi't‘g

i
saw

5

 

THESE

IHIHHIIHHIIIHIHIHlHlJlllllllllHlHlllllllllllllll

31293 01572 2030

LIBRARY
Michigan State
University

 

 

 

 

This is to certify that the

thesis entitled

MACROMOLECULAR SIZE DETERMINATION BY
SCANNING FORCE MICROSCOPY

presented by

Martha Gilchrist

has been accepted towards fulfillment
of the requirements for

M. O . h
S degree In C emistry

Cit-wk

Major professor

 

 

 

Date H/Z§/?(

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE ll RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or before dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l ICE
II II]
l | l

MSU In An Affirmative Action/Equal Opportunity inetitmion
W

i
l
l
l
l

 

 

 

 

 

 

 

 

 

m1

_—..____.____.._. ____

MACROMOLECULAR SIZE DETERMINATION BY SCANNING FORCE
MICROSCOPY

By
Martha Gilchrist

A THESIS -

' Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE,

Department of Chemistry
1996

ABSTRACT

MACROMOLECULAR SIZE DETERMINATION BY SCANNING FORCE
MICROSCOPY

By
Martha Gilchrist

A fast and precise method of determining protein size has been developed using
Scanning Force Microscopy (SFM). The technique, termed Random Adsorbtion Molecular
Sizing (RAMS), has provided nanometer resolution for examinations of macromolecules
under conditions in which the sample does not require the extensive fixation protocols
normally associated with other high resolution techniques such as IBM or SEM. A model
is proposed which allows accurate determination of the three dimensional size of molecules
randomly oriented on a flat surface. This makes possible studies on changes in quaternary
structure due to point mutations and ligand-induced oligomerization. Using RAMS, the
molecular size for the enzyme ADP-glucose pyrophosphorylase (ADngp) is determined,
and the enzyme is shown to undergo a change from tetrainer to octamer caused by a single
point mutation. RAMS is also used to study ligand-induced oligomerization in two

molecules, Streptavidin and Wheat Germ Agglutinin.

ACKNOWLEDGEMENTS

‘For such an interdisciplinary work as this, I have many people to thank. First,
thank you to Professor Melvin Schindler, who has been my main link to Biochemistry and
has furnished my ideas for this project. He has been key in the editing of the paper, to
be submitted to the journal Biochemistry, on which this thesis is based. Professor Jack
Preiss has been a valuable source of advice on protein structure in general and his protein,
the ADP glucose pyrophosphorylase, in particular. The help of Dr. Min Wu in Preiss’s lab
was invaluable when I was doing the electrophoresis. He gave me both instruction and
space at his lab bench. I am very thankful to Professor John Wang, who pointed out some
important structural features of Con A. Thank you to Joseph Leykarn, whose generosity in
teaching me to use the HPLC and allowing me to use his equipment has made this project
run more smoothly. Finally, the enthusiasm and advice from Professor Marcos Danurs has
been the driving force in developing this project, and I truly appreciate his patience and
encouragement.

I am grateful to the funding agencies which provided support for me for several
semesters and money with which to purchase the Scanning Force Microscope. This
research was partially funded by the REF-Protein Structure, Function and Design Award
(MSU), the Center for Fundamental Materials Research (MSU) and a Camille and Henry
Dreyfus New Faculty Award to Professor Dantus.

iii

TABLE OF CONTENTS

LIST OF
TABLES .............................................................................................. iii

LIST OF
FIGURES ............................................................................................. iv

CHAPTER 1
INTRODUCTION .................................................................................... 1

CHAPTER 2
IMAGING THE MOLECULAR DIMENSIONS AND OLIGOMERIZATION
OF SINGLE PROTEIN MOLECULES BY RANDOM ABSORPTION MOLECULAR

SIZING (RAMS) MICROSCOPY ................................................................. 9
CHAPTER 3

LIGAN D BINDING STUDIES ................................................................... 48
CHAPTER 4

CONCLUSIONS .................................................................................... 66
APPENDIX A: SOLID FILMS OF PROTEIN .................................................. 74
APPENDIX B: VISUAL BASIC PROGRAM TO AUTOMATICALLY MEASURE

AND TABULATE THE HEIGHT OF MOLECULES IN AN SFM IMAGE. .............. 78
APPENDIX C: VISUAL BASIC PROGRAM TO AUTOMATICALLY OPEN

SEVERAL TEXT FILES AND PROCESS THE DATA THEN CLOSE THE FILES ..... 87
APPENDIX D: C-H- PROGRAM TO CONVERT SFM IMAGE INTO AN

X, Y, ARRAY SUITABLE TO READ INTO EXEL ........................................... 88
APPENDIX E: MATHEMATICA ROUTINE TO READ A FILE OF TABULATED
RAMS DATA AND FIT THE DATA TO GAUSSIAN CURVES ............................ 90

LIST OF TABLES

Table 1: Summary of sizes of colloidal gold particles by TEM and SFM 18

 

Table 2: Summary of the protein dimensions derived from RAMS microscopy and x-ray
crystallography ..................................................................................................................... 47

LIST OF FIGURES

Hgmel:SchematicrepresentationofthemainworkingpartsoftheSFM. ......................................... 7

Frgure2:Diagramoftipconvolutioneffectanderrorduetomisalignmentoftipandprotein. ................. 8
Figme3:ThreeorientationsofamoleculetheshapeofConAarrangedonamicalattice ...................... 30
Figure 4: 2.5 by 2.5 pm image of WGA on mica. ................................ . ........................................ 31

Figure 5: Correlation between elongation and the probability that the two tallest dimensions are vertical..32

Figure 6: RAMS on 4.9, 9.0, and 18.1 nm colloidal gold fitted to Gaussians. ..................................... 33
Figure 7: RAMS on Ferritin ............. . ......................................................................................... 34
Figure 8: RAMS on WGA fitted to (A) monomer and (B) dimer dimensions ....................................... 35
Figure 9: RAMS on gluteraldehydeueatedWGA fitted to (A) monomerand (B) 55% monomerand45%
dimer ..................................................................................................................................... 36
Figure 10: RAMSonConAfittedto(A)monomer(B)dimer .......................................................... 37
Figure 11: RAMS on Con A fitted to (A) tetramer (B) 13% monomer, 67% dimer, 20% tetramer ........... 38
Figure 12: HPLC on (A) 3 rig/ml Con A (B) 5 mg/ml FlTC-sCon A ................................................ 39
Figure 13: RAMSonProteinAfittedto(A)monomer(B)PaflAplus24%from24and40A .............. 40
Figure 14: RAMS on Streptavidin fitted to monomer ..................................................................... 41
Figure 15: RAMS on gluteraldehyde treated Streptavidin fitted to (A) monomer (B) 66% monomer 34%
dime ..................................................................................................................................... 42
Figure 16: Non-denaturing gel on ADngp wild type, G336D, and Anabaena. ..................................... 43
Figure 17: SFM data on ADngp wild type, G336D, and Anabaena ................................................. 43
Figure 18: RAMSdataonADngp wild type... ............................................................................ 44
Figure 19: RAMS data on ADngp GB36D mutant. ....................................................................... 45
Figure 20: RAMS data on ADngp from Anabnena fitted to (A) monomer (B) 68% monomer 31%

tetramer .................................................................................................................................. 46
Figure 21: Pre-dilution mono-GleNac treated WGA fitted to (A) monomer (B) dimer ............................ 56

VI

Figrne 22: Pie-dilution mono-GLcNac treated WGA fitted to 50% monomer, 50% dimer ...................... 58

Figure 23: Post-dilution mono-GLCNac treated WGA fitted to (A) monomer (B) dimer .......................... 59
Frgrn'e 24: Pie-dilution di—GLcNac treated WGA fitted to 68% monomer, 32% dimer ............................ 60
Figure 25: Post—dilution di-GLcNac treated WGA fitted to (A) monomer (B) dimer ............................... 61
Figure 26: Pre—dilution tri-GLcNac treated WGA fitted to 81% monomer 19% dimer ............................. 62
Figure 27: Post-dilution tri-GlcNac treated WGA fitted to (A) monomer (B) dimer ................................ 63
Figure 28: Pre-dilution biotin treated Streptavidin fitted to tetramer ................................................... 64
Figure 29: Post-dilution biotin treated Streptavidin fitted to (A) monomer (B) dimer ............................. 65

Figure 30. Post-dilution biotin treated Streptavidin fitted to (A) tetramer (B) 24% monomer 76% dimer...66

 

Figure 31: Two-dimensional array of Streptavidin ............................. - ......... 77

Figure 32: Two-dimensional array of ADngp wild type ................................................................. 77

CHAPTER 1: INTRODUCTION

Protein aggregation has been shown to play a key role in regulating enzyme activity
for many proteins including Ca2+ -ATP-ase1 3 and ADP-glucose pyrophosphorylase,3 and
ligand-induced oligomerization has been suggested to cause enhanced binding for a large
number of DNA binding proteins such as p53,4 Par A,5 Mu B,6 Rec A, 7 and many others.
To understand the function of such proteins it is critical to have a good understanding of
their molecular size and oligomerization state under physiologically relevant conditions
when the molecules are subjected to treatment with appropriate ligands. Unfortunately,
few techniques are presently available with the necessary sensitivity and resolution to
measure molecular dimensions and subunit organization under conditions that can maintain
biological activity. Here we present a new technique which ean quickly determine the size
and oligomerization state for virtually any water-soluble protein molecule.

This research has focused on the development of techniques aimed at obtaining
measurements of biological macromolecular structures by Scanning Probe Microscopy
(SPM). The primary experiment was to develop a method for determining the size of a
protein molecule using Scanning Force Microscopy (SFM). The method developed can
give a reliable three dimensional size for virtually any water soluble protein only a few
minutes after sample preparation. In addition to general structural studies, this method can
be used to study ligand-induced conformational changes in a variety of molecules. When
applied to study mutant varieties of a protein this method has also furnished information
about the relative stability of the quaternary structure of the mutants, because if quaternary
structure is intact this will affect the apparent size of the protein. Also, a model has been
devised which correctly predicts the dimensions and orientation of protein molecules when
they are randomly oriented on the miea surface.

SPM uses a sharp probe in close proximity to the sample to obtain high resolution

images at a given x, y, position. The signal may be electrical conductance, as in scanning

tunneling microscopy (STM),8 attractive or repulsive forces, as in SFM,9 or absorption or
transmission of light, as in near-field scanning optical microscopy (NSOM).lo SFM does
not require the sample to be conductive as in STM or frozen, dehydrated, and coated with
metal as in electron microscopy (EM). SPM is therefore an ideal tool for studying
biological macromolecules and structures,”12 which may be dismpted or decompose when
subjected to metal coating or vacuum drying. The resolution that may be obtained by SFM
I is similar to that obtained by EM, on the angstrom level. Using SFM we have been able to

image, with no special treatment, biological structures that still hold their solvation water.

SCANNING FORCE MICROSCOPY

Basic Operation. Scanning force microscopy works by measuring the
interaction between a surface and a sharp tip on a cantilever. The cantilevers with
integrated tips are commercially available from several companies; the ones used in these
experiments were either silicon nitride with pyramidal tips having an aspect ratio of about 1
andaradius of curvature as small as 50 nm, or silicon with higher aspect ratio tips and a
radiusofcurvatureassmallas lOnm. InFrgure 1 the mainworking parts ofthe SFM are
pictured. These parts are the piezoelectric scanner, a, which moves the sample, b,
underneath the tip, c, which is mormted on a flexible cantilever spring, (1. The cantilever is
deflected by the interaction forces between the tip and the sample. A diode laser, e, shines
a beam onto the top of the cantilever which reflects the beam into a position sensitive
photodiode, f. The proportion of the beam in each half of the photodiode is detemrined by
the angle of the cantilever deflected by the sample. The position sensitive photodiode
monitors changes in height by detecting changes in the direction of reflection of the diode
laser reflecting off the top of the cantilever. The signal from the position sensitive
photodiode is fed into a feedback loop to regulate the height of the sample, so that as the

sample is scanned constant force is maintained between the sample and the tip. The amount
thatthe sample has to bemovedtomaintainthetipatconstant heightand constantforce at a
given x, y, position is recorded as the 2 value of the topographic scan. If the feedback loop
is operating properly the deflection of the cantilever changes very little because the changes
in topography are compensated for by expansion or contraction of the scanner tube. This
technique is called constant-force contact SFM.

Alternatively, the cantilever is vibrated with an amplitude of about 30 A at its
resonant frequency (360 kHz for the 2 pm thick, 200 um long triangular silicon nitride
cantilevers) about 20 A from the surface being scanned. As the tip approaches the sample,
the resonant frequency of the cantilever changes because of the attractive Van der Waals
forces between the tip and surface and this change in frequency is detected by the
photodiode. The feedback system compensates for the change in distance between the tip
and sample, and the amormt of vertical movement necessary to return the resonant
frequency to its previous value is registered as the change in height of the sample at a given
x, y, position. The tip‘is typically maintained at a distance of 15-100A above the surface.
This mode of operation is called non-contact SFM, and is usually less disruptive to a
surface because the forces between tip and sample are smaller. If a small amount of debris
is picked up by the tip during scanning, this will disrupt the resonance frequency in non-
contact mode and prevent the acquiring of images until the debris is removed or the
resonance frequency is reset. Also, if the sample has a thin water layer on it, the capillary
forces may be strong enough to pull the tip into contact with the sample making non-contact
imaging impossible. Contact mode has the disadvantage of having the tip actually touching
the sample, introducing the possibility of damaging a soft sample. Lateral forces in contact
mode may be high, sweeping insecurely fastened objects in front of the tip as it is scanned.
However, most published research using SFM is done in contact mode because non-
contact mode tends to be unstable and more easily disrupted.

‘3'” which combines non-

Tapping mode is a recently introduced mode of operation
contact methodology with some additional measure of stability due to intemrittent contact
between the tip and the surface. This mode of operation eliminates most of the lateral force
induced by contact imaging, but still allows the possibility of sample damage due to vertical
compression. It is still less stable than normal contact imaging due to the resonance
frequency dependence of the imaging, but more stable than non-contact imaging because

the forces measured in acquiring the image are much larger.

Capabilities and Limitations of SFM. Using SFM, one can measure height
with a resolution of better than one angstrom, and on a flat crystalline surface, lateral
dimensions can be measured with a resolution of better than 3 A. The best resolution is
obtainedonflatcrystalline samplesbecauseinthesecasesthe image is acquired using only
the very endmost atom oratoms of the tip. On a surface with particles on the order of the
size of the probe, the image of the particle is convoluted with a reverse image of the probe.
Since the tip is measuring the surface by scanning in contact, the finite dimensions of the
tip become part of the imlrge.“"""‘7'18 A simple algorithm has been proposed to separate
out tip convolution effects when imaging spherical particles with known size. Dimensions
can be calculated as

d = wz/4h (1)

where d is the tip diameter and w is the apparent width measured for sample height It.18
Once the tip diameter and geometry are known, the image can be deconvoluted and probe
effects separated from the true inmge of the sample. In these kinds of images, the sharper
the tip compared to the sample size, the less the error caused by the probe convolution. At
this time, the sharpest tip commercially available has a radius of curvature of 10 nm, a

value bigger than the diameter of most protein molecules. The error introduced by the tip
convolution can widen the horizontal measurements of 50 A diameter spheres by as much
as 950 A using a tip with radius of 500 A. Fortunately, this tip convolution does not
distort measurements of vertical dimensions. We have designed our protein sizing
experiment based on vertical measurements only. Figure 2 illustrates how horizontal

measurements are broadened due to this tip convolution effect.

Vertical Probe Tip Corrections. Height measurements of small spherical
particles with SFM are slightly reduced when images are taken at low resolution. This
measurement efiect occurs because the end of the tip, which can be modeled as a sphere, is
not necessarily at the apex of the protein. For example, in an inmge 2.5 microns square
which is 256 by 256 pixels, the end of the tip could be as far as 48 A from the molecule if it
is located exactly between two scan lines. This effect can be reduced by using tips with a
larger radius or having a larger number of pixels in a scan. In Figure 2, the idealized
shapes oftheendofthetipandaspherical proteinmolecule are illustrated. Because ofthe
finite number of points which can be recorded, the apex of the protein will usually not be
directly over the end of the tip, therefore the measured height of the particle will be less
than the actml height of the particle. The discrepancy, e, can be calculated as follows:

6 = (R + r) - ((R + r)2 -X2) “2 (2)

where R is the radius of curvature of the tip, r is the radius of a spherical protein, and X is
the horizontal distance between the center of fire tip and the center of the protein when this
distance is smallest. For a very fine tip with 100 A radius of curvature, a protein with a
diameter of 40 A, and taking an image with 256 by 256 points and 2.5 micrometer square,

the height deviation can be as much as 10 A. However, since the molecules are randomly

distributed on the surface, the effect on a set of measurements will be significantly less.
For 500 random horizontal distances between 0 and 48 A from the apex of the protein to
the end of the tip, the average error is only 3.5 A. The typical radius of the tips used in
these experiments was 750 A, as calculated from width and height measurements for
spherical particles.‘ 8 The average height error expected for our measurements due to
tip/protein misalignment is therefore 0.5 A.

Because of the large tip convolution effect, SFM is most useful when used to image
the height of very small particles spread on a surface, the surface topography of regular
crystalline surfaces or the three dimensional structure of particles and surfaces with featm'es
much larger than the dimensions of the probe, or to measure forces between
surfacesmuuw’ SFM has developed into a very useful tool for investigating biological
samples, having been used to image regular 2-D crystalline arrays of protein at molecrrlar

2‘25 measure adhesion forces and elasticity of lysozyme on mica,26 and observe

resolution,
the changes in a cell upon infection by a virus.27 Also, there has been extensive work in
imaging isolated DNA moleculeszugaw and individual protein molecules.323“"34 Work in
our group has concentrated on imaging large molecular structures and has branched into
two areas: developing a technique for fast protein sizing and measuring ligand-induced

conformational changes in protein molecules.

Em 05 Co 88a wen—.53 FEE 2: ho 5:858..qu canfioaom 3 Beam

 

 

13:22:88 oEoofioaoE Am

 

 

 

 

All 355 BEE
332% common C

 

 

£20.:— 98 a
w ..o 325.».
...EE 2 one note 2.. «echo :0:
. 202.8 a:
. ..o 88920 .
. seam

 

 

 

W . I
"I. O .010OI‘IUl“ ‘
......o. ....... ......e.....eqem...e..q..t.....u.. .1 . N e 1
...... . .... ... ... ... e. . ...".r t
........ .0 . ......e......e.e . ...... . 00......0W om: I
.0... .. .. 00.... ...... .... . ...... .. .......e. .00. 0 0.. 000 .dl.‘ Iqlqi
.... .e......... .0 .ee.............. .. ..e....... .. ...ee .. .... ....e ...-.... ......e...e.... .... 0......Nloeoe1.d v: .iwiJ.II
00 I. .0 0... 0 .. 0. ...... . 00.. o. I. .0 0. 00.0000 .0 M 0 0.000 I .... 0. 0. I. 0. . 00.. .0 I ...... 0 ..0 0. 0. 0. .0 0‘00‘040 l I I
000 I 0 0.0. .0 0.0 0. 0.. 0 00 0.0 0 I... I. 0. . 00.00I I
.. ........ ...... ..... . ..... ... ....... ... .............. ........ .... ..................o............ ....f..d.....t .N . I
8...... e........ ......... .0 .............. ..... .. ....... .. ..... .. ....e.... .. .. .. .. ...... ......e. .e ......e...... o. .. .. ....... .... ... .. .. coco-...... .. .. e. .. .d..t.. .1 ...t . .1 . eidlt
e ... ......e... .....e... ......o....ee .. .... ... e. e .... ......ee... .. .e .e .....e... 0......00. ... 1 t
.. ......... .. . . ... .. .... .... . . . ..... .... ..... .... .. ...... . ....... . ........... .. .... .. ....... ......... .. .. .... .. Moo... .... ..... 0...... . .... .. .... .. ....... .. .... .......... .. .....d.. .... 5%.. fl du‘ .ul.
. ...... ....... .... ...... ......... s. .. ..... ......o....... ... .... ....... . .......... . ... ........ ............. .... ..... ................. ... . . t
................. ... .. .... ........ ......................... ....... ......Jfl ...... .. ........ ....... .. a............ ......e.. ....... .. .- .. .. ......o ...... ...... .. ....... .. ..... .. f... . . .... ..... .. ......e......... .. ...............N...q.......”...”.u mfl. N . 1 . t
...e. .0 e.... o ....e e. I. ......e ......o. o. o .....e...... ...-.... ..ee. ........... N .... ...... 5...... ......O... .. no... ”...-......" ......ceo ...... .. ...-... ......e e... I. ... ......o..... ......e. .....f...... ......e. ...-...... .... ... ”...-”...... ......o. e....e&... 01.. N . t . 1
...-...... .. ...... .eee...ee..... ......aoo .. .. .............e.o.e ee...ee...e.ee....e...e.. ...N...e.e. e. ......eeoeeoooweeeeeooeeeoefi .... o.......e.....ee ...-.....e. ... o ....e ...e. ... ....e... ......a....e.e ......e ......e........fl... ......n...e. ......e... ... ......e. ....... ......fl. .eoeeeeeeeeeee ......ch..em.H.q at" . t
000.0 ..... 0 . .. 0.0000 ...... ......I....%.. .. ......I.. .00.... ... ..... .0.0.~... 0.. .. .... .... .... ... ..’0.0..I .. a ... .... .........I.....0..... .... ....00.... ...... ........0...... 0......0.’ e........0......0. .... .....0.....0. 0.. ......cel..0.. .. .... .....0.......0. .. .00H...0..........0 .I... .......0e....0 ......0......00.00... I. 0. .00..00.1.0.m1.0 .1 .l.{.. 0
I 0. e. I . 0. ..0.. 0. I. e. ...... a 0. . 0... 0 . .....I. I .00. e. . ...... 0.... ...... 0. ... . .... 0.... .0... I. .0 ... 0.. .I .. ...... . .... .. ... ...... ... .. ...... ......0 . 0. ........0 . .- .. 0. ... 0. 0 . ... . ... 0.. 0. . .00.. ..0I. .0 .... I .. .... .0.0.... I... 0 ......00 I. 0.... e. 0 ............ I. f
coma-...... .... . . .... .... no... .....e. .. ...0. .....e.... ......eeeooeeoo...e.. .....eue ...... e ......e.. e.......... .. ......neeeo... ...... ......on........ ..eeeeoeeee...ee. e. ......o.. ... .....e....e... .. ....e.....foe.... e...”...ee.e.eee....e..e...e. ......efieoffioe. .- e...ee.e.....e.. no... ......u....e.... .. ...... ...... ...... . ...........o....”...”..
. on . ... e .. e o a. .. ... .. .... . ...... ... e ... ... .. ... . . . . . ... .- .... I . ...e. .. .... .... . e. . ...... .. .... a. ...e. e a .o o. ....e .. ... c. e. e. . ... .e . .... e. e e. .. e. .... .. co... ...... .... .. .... e. o. e e e. .. . ...... . e o. .. or... e . ea... . e e . e .
70.00.00.000... .0... . ......I000... 00...... .... 0.00... .- 000.0000. .... I 000.00....000 0. 0.0 00.0.0..000..00. 00.00. I00...00..0..0I .0 ’0 0.000 .0 00.000000 .00... 00.0 00.0.... 00 I .0 0.0 00.... .. 000.0000 .000. 0.0.000 00.0...... 00.0000... 0. 00I 000 I. 0000. .0 ......000...00 0.00. .....l... ...... 0 0.0.0 ......000. .0000 .0 I .0. 000.000....0...00.000.000.00.
...... ... ..... ...-.... ......o...... .. ..... ...... . .. ... ... . ....o .. ................. .. ..... ....... ..... . ....... ..... .. .. ... ............... .. .. .... ....... . ............ ..... .. .. ..... ..... .. .. ............. ... .... .... ..... . ... .. ............. .. ....- .... ...... .... ... .............. .. .... .......... . ...... ...... .
W 0.0.0 00.0.... .... .I .000...0 In... .. H .... ......e. ..e.. . le. . .I I... I. .0. ........ ....0 H ’.0 ..... I ..... .0 .0. . ...... .... .- .....I. I. ...I. e. ......0. 0 ... e. 0.... .... 0 ... .. 0. ..... ...0. 0. .... .0 ...... .. . 0. ...... ....I I. 0.0.. .. . ... .0 ...I. 0. .. .... I. .. ..... ...II... .. 0e... ..."... 0.. .I ......00
ee ...-......- ..e. e....e..e.oe...ee......u........ ..... .... ... .....eee ....... .....ee...eeee....e.e....o. ......eo ............ee..e.... .... ..........e. ...eee. .... .. .......... e ........e.. ... ......e...eee...
0 00 0.0 000000.0000’0 .00fOI’I0l0000 0.00.00.00.00... 0 000.00 00... 0 0000 000000 .00....0000...0 .00 .. 000.0......0.II0. 0 I00..."...0 .0 .00....0 0. ...... I 0...... 0.... 00.000I.Il0.000.00
. 0 ~0. 0. .0 00... .. 00.00.000.000. ...... 000.000... 000...... .0 .0000’000 0. ...... ...-0.0 .0... ..... . 00.000.00.00. .000. .0...0..000...00. 0. 00.0.0.0 .- .00....0’ .0...I.... 000.00..0 00.00.... I... 0... I. 00.0... . 0. .... 0...I.0..000..0 000... 000...... 00000000.. I... .0. I.I.00000....0 .I ...... .......00..000.. 000.... 0000...... .0000...I .000... I...
. . . .. 5 .....f .. .. .......... ............ ..... ...... .. .... . ................ .......................e.. a . . . . ............. ... ......... .......... ......... ............ ................. ......o.......... ... ... .................... .... ..... .. .. ....... ...........
. . e I .e . e .. ... .e o. .. ......o .. ..... .. . ... ... ......I. .... .... .. .... ... .....e. ....e . . ......a. .. ... 0...... ......I... e c .....e...... e e. . ... ....... e. e. ......0
0 0 00 “O 00 0 .0000 .0 ’00..”0...0.0f... 0......” 0.....0... o ME...000000.000 0.0.0.0...”000000.00. .0000”. I... 0... ......I000. .00. 000000 0 .0...000..00.I.0..... .00.. .0... 0......0.....00..0 I00. 00.......0 ....0000H.0000000. 00000.00H.0.0I0.. .000 0.0.0.... 0.0...
O 0 0 0. “.00 0.....00 .0 0.0 00 0 .0 . 0.000.. 000......00. 0.00.. .......I..00.0. ... 0.00 000.000.00.0I.00. 0....000.00.0 .0 .000.0.000.0.0.000...0I. 00.. .......0.00I.. 0000.000... 00...... 00.00.00. .I 000......000
. .. ~e’...o.ee... ..efi... “...... ..e. .........e.e.................. ......... .. .... 0......
. . . .. 9.. ...... . . .. .... .. . . .. ... .. ............ ........ .... ...... .. ... .... .. .. .. ... . . ...... ....
. . . ... . ... e ......o... ......o......... ...... ..... .... ................ ......... ............ ... ....... .e... ....
. . 0 .... 0 0.0.0... 0.0.. ......0..0.0.. . .....0....0...0. ...... . .0. e. 0. ... ......0.....
0 0 0. 000.00.00.00. 000”"...0..I.00000”.. 00000000....000 0 0.. I
. . e. . . .... .. .... e........
0 . . 00

 

£38m

r’ f Q
5 %£ 80g Omwfim

h+m

9

H

 

 

 

 

CHAPTER 2: IMAGING THE MOLECULAR DIMENSIONS AND
OLIGOMERIZATION OF SINGLE PROTEIN MOLECULES BY RANDOM
ABSORPTION MOLECULAR SIZING (RAMS)’5 MICROSCOPY‘

INTRODUCTION ,

A variety of analytical techniques have been developed that provide information on
the size and shape of proteins. Chromatography, electrophoresis, transmission and
scanning electron microscopy (TEM and SEM), light scattering, nuclear magnetic
resonance (NMR) and equilibrium ultra-centrifugation are all commonly employed methods
‘ to determine the nrolecular size and shape of proteins." Although of considerable
importance in the characterization of protein structure, these techniques do not have the
ability to provide molecular dimensions of large macromolecules at high resolution, with
the exception of TEM and SEM, which require harsh fixation and dehydration protocols,
and NMR, which is most effective with small proteins. Such measurements are essential
for the characterization of the changes in protein conformation or state of aggregation that
have been associated with ligand binding, protein-protein interactions, or activator induced
oligomerization. At present, the only technique that can provide angstrom resolution for
the direct measurement of molecular dimensions is electron diffraction microscopy.37 To
perform such measurements, however, crystals must be prepared from protein solutions.
A particular difficulty with this method is that in the absence of a generalized crystal growth
protocol for oligomeric and membrane proteins, the utility of this approach has been limited
to predominantly soluble low molecular weight proteins.38 In addition, co-crystallization of
proteins with ligands is a more challenging task.

To circumvent the need of crystals for diffraction based approaches while still
maintaining die capacity for dimensional analysis at high resolution, we have developed an

analytical technique that uses scanning force microscopy to measure the dimensions of

 

' The work presented in this chapter is based entirely on a publication to be submitted- see
reference number 35.

10

individual protein molecules with angstrom resolution. Since its recent introduction, SFM
has been extensively used for imaging biomolecules.11 The technique has provided
nanometer resolution for examinations of macromolecules under conditions in which the
sample does not require the extensive fixation protocols normally associated with other
high resolution techniques such as TEM or SEM.8

Here we extend the rrtility of SFM to measurements of molecular dimensions with
angstrom resolution. A loss of resolution is normally introduced into SFM measurements
as aresult of probe tip geometry’imms as shown in Figure 2, a schematic representation
of the tip/protein interaction. To avoid this resolution loss, we have devised a measrnement
strategy that uniquely measures the height of protein molecules that have randomly
adsorbed to the atornically flat surface of mica. Figure 3 exhibits a representative drawing
of Concanavalin A dimers arranged in three orientations on a mica surface with an SFM tip
positioned on top of one of the molecules. The Statistical Orientation Model (SOM),
described in Theory and Data Analysis, assumes that the proteins may be represented as
parallelepipeds, and predicts the orientation according to the relative dimensions of the
parallelepiped. In Figure 3 the size of the molecules, the tip and the mica lattice are drawn
to scale. The technique to determine molecular size presented here has been termed
Random Adsorption Molecular Sizing (RAMS).35 The use of the height of adsorbed
protein molecules as the defining parameter for molecular dimension makes the RAMS
technique minimally dependent on probe tip geometry and extends the range of the SFM
from nanometer to angstrom resolution The relative ease and speed with which these high
resolution measurements of molecular dimension may now be performed should provide
investigators with new opportunities to examine changes in protein eonforrmtion and
oligomerization that depend on protein concentration and ligand or allosteric activator
binding. A fruthcr advantage of RAMS for structme frmction studies of proteins is that the
method provides a rapid screening protocol in which the molecular dimensions of large

11.

numbers of recombinant and site-mutated proteins may now be determined in a minimum

amount of time (15 minutes) with a minimum amount of material (~100 molecules).
MATERIALS AND METHODS

Materials. Colloidal gold particles were obtained from Sigma (St. Louis, MO)
and used without further purification. Three sizes of gold particles were used for this
experiment: 49 :i: 6 A (3.8 x 1013 particles/ml), 90 :l: 12 A (6.0 x 1012 particles/ml), and 181
:i: 8 A (6.7 x 1011 particles/ml). Protein A, wheat germ agglutinin (WGA), and
Streptavidin were obtained from Sigma (St. Louis, MO). Ferritin was obtained from
Polysciences (Warrington, PA) as the cationized protein. Concanavalin A (Con A) was a
gift from Professor John Wang, Department of Biochemistry, Michigan State University.
ADP-glucose pyrophosphorylase (ADngp) enzymes were a gift from Professor Jack
Preiss, Michigan State University. HEPES was from Sigma (St. Louis, MO) and the
MgCl2 was ACS grade from Columbus Chemical Industries (Columbus, OH). The water
used in preparing solutions was purified with a MilliQ MilliPore water purification unit.

High Performance Liquid Chromatography. HPLC was performed on Con
A and fluorescein derivatized succinylated Con A (FlTC-sCon A). The Beckrnan SEC-
2000 size exclusion column was used with an elution buffer of 40 QM Tris, 300 .m_M
NaCl, and 300 m_M_ glucose pH 6.8. The flow rate was 0.8 ml/minute and the protein was
monitored spectrophotometrically at 214 nm as it eluted from the column. Molecular size
standards included Trypsin, (27 kDa), and bovine senun alburrrin (68 kDa), and FITC-
sCon A (51 kDa). The concentration of Con A and FITC-sCon A was 3 rig/ml and 5
mglrnl respectively dissolved in 6 mM NaCl, 10 mM HEPES pH 7.5.

12

Non-denaturing gel electrophoresis. Non-denaturing gel electrophoresis
was performed on the wild type and G336D mutant of the E. coli ADngp enzyme as well
as the enzyme from Anabaena. The electrophoresis was performed essentially as described
by Omstein and Davis,39 brrt without exposing the protein to SDS or DTI‘ in the buffers,
and without heating the proteins. The protein was run on 8% and 10% polyacrylarrride
gels. Resolving gels were rmde 400 mM in Tris pH 8.8 and polymerized with 0.1%
ammonium persulfate and 0.04% TEMED, and either 8% or 10% acrylamide. Stacking
gels were 5% acrylanride and 130 mM in Tris pH 6.8.

Preparation of Samples for Height Measurements with S FM.
Measurements of protein height required an atonrically flat substrate. In this study, we
used muscovite mica obtained from Ward’s Natural Science Establishment, Inc.
(Rochester, NY .). When freshly cleaved, mica has a net negative surface charge. Treating
a freshly cleaved surface of mica with a 5 mM MgCl2 solution results in the replacement of
endogenous K+ ions with Mg” making the surface more positively charged. The
positively charged surface has been demonstrated to enhance adsorption of
macromolecules.40 Ten ul of an approximately 3 pg per ml solution of the protein
dissolved in 10 mM I-IEPES, pH 7.5, is deposited onto the mica followed by the addition
of 10 pl of 0.5% gluteraldehyde. Fifteen minutes later, the surface is gently rinsed twice
with 200 rd portions of MilliQ water and allowed to dry for several hours. This
preparation yields a dispersed p0pulation of individual protein molecules when the rrrica
surfaceisimagedby SFM. Figure4isatypical RAMS scan used to collect protein height
measurements, a 2.5 pm by 2.5 pm image of WGA. All the images used in our study
showed a random distribution of individual particles. The protein concentration was
chosen to provide a large population of protein molecules that are adsorbed in difl’erent
orientations. In the presence of a relative humidity of about 30%, there is a monolayer of

13

water maintained on the mica surface to hydrate the adsorbed proteins, as described by Hu

et. a1.“1

Random Adsorption Molecular Sizing (RAMS) Microscopy.

The microscope used for this work was an Autoprobe CP scanning probe
microscope (Park Scientific Instruments), with a five um high resolution scanner. SFM
data were recorded in contact mode under controlled ambient conditions (30% humidity and
21° C). Ultralever cantilevers (Park Scientific Instruments) having a spring constant of
approximately 0.06 N/m and an integrated silicon tip with a radius of curvature of
approximately 10 nm, according to the manufacturer, were used as probes. The scanner
was calibrated for vertical measurements using Tobacco Mosaic Virus (diameter 18 am)
(American Type Culture Collection), and for horizontal measurements using a 1,000 x
1,000 x-y features per millimeter grating.

It was found that the protein molecules adhered to the mica well enough for
measurements only when the humidity was less than 40%, probably because high humidity
increases the attractive capillary forces between tip and sample.3“"“"‘2 In the summer, when
ambient humidity could reach 85%, the humidity was controlled by placing the microscope
in a loosely sealed glove bag with a hygrometer (Omega model RI-IDP-l) and flushing the
bag with a slow stream of dry nitrogen to maintain a humidity of about 30%. In the winter
when ambient humidity was typically less than 40%, measrnements were performed
without the glove bag and in ambient air. A humidity of less than 20% was found to
interferewiththeprocessoftipapproachtothesampleprobablybecauseofstatic charge
buildup between the tip and the charged mica surface.

Height measurements were performed using the minimum force required for stable
imaging in order to avoid compression of the molecule which would produce force induced

height artifacts in our measurements. Images were taken within the range of the attractive

l4

capillary forces. In order to accomplish this an approach was performed at a high force
setting, then the force was lowered to scan the image. Scan size for all irrrages was 2.5 pm
by 2.5 pm and 256 by 256 pixels, and scans were recorded at a rate of 2 lines per second.

The height of each molecule was determimd by taking the difference in height
between the top of me particle and the surrounding substrate. The SFM software permits
theusertodraw alineataspecific plaeeintlreimage and measure height differences along
theline. Forall ourmeasrnementsthelinewas drawn parallel tothefastscan direction and
placed so tint it intersected the highest point on the molecule, enabling a measurement of
the height of each molecule. The roughness of the mica surface in the prepared samples
was less than 2—3 A, a surface distortion that is probably the result of a small amount of
residual buffer salts. Occasionally, large clumps of protein were formed instead of the
desired dispersed field of individual molecules. Only the heights of isolated and individual
particles were measured.

THEORY AND DATA ANALYSIS

Statistical Orientation Model (SOM). Most of the proteins imaged for this
work were asymmetrical or oligomeric demonstrating a distribution of measured heights.
In order to interpret these single molecule measurements, we devised a statistical orientation
model (SOM) to represent the distribution of heights that could be obtained by molecules of
specific dimensions binding randomly to the mica. The SOM assumes each protein can be
enclosed by a rectangular box (see Figure 3). A favorable interaction between the
negatively charged protein molecules and the positively charged mica surface would favor
adsorption such that the largest axis (represented by largest surface area) of the protein is in
contact with the mica The surface display of protein orientations that are observed in the
RAMS measurements can be easily modeled macroscopically by performing an experiment

15

dropping a large number of small rectangular wood boxes. Each box represents an
individual protein molecule adsorbed with a particular orientation. The parallelepipeds
(with increasing dimensions a, b, and c) were allowed to fall under the influence of gravity,
and their orientation was tabulated Comparing the height distribution profile for a group
of proteins obtained by the RAMS technique, we found that protein orientation on the mica
surface (in the absence of specific chemical recognition between the molecule and the
surface) closely resembled the distribution of heights obtained after dropping scale model
parallelepipeds onto a surface. A phenomenological relation was found for the probability
of measuring a given side of the parallelepiped For example, the probability of measuring
a, the smallest dimension, was found to be proportional to the area of the base (b x c)
divided by the distance of that area to the center of mass (d2) such that:

P (a) = 2(b X C)/0 (1)

In order to test the SOM we found a relation between the product of the normalized
probability of measuring side b and side c such that,

C x B = (a/(b x c))2 (2)

This formula allows us to compare the experimentally measured probabilities B and C to
the predicted value calculated from the known values for a, b, and c. The conelation
between these measurements is shown in Figure 5, a comparison between the
experimentally measured values for C and B (the relative probability that c and b will be the
measured height of the block when it is allowed to fall under the influence of gravity) and
the calculated d(b x c)2 which is a measure of the elongation of the molecule. The
relationship predicted by the SOM agrees well with the actual experimental values observed

16

when rectangular blocks are allowed to fall under the influence of gravity, as shown by a
comparison between the experimental points and the straight line with slope of 1.

Using SOM we are able to predict the height distribution from a randomly adsorbed
protein given a set of approximate dimensions a, b, c. While there is no physical reason
why protein adsorption should closely mimic the distribution of measured heights observed
for the wood blocks we have found a very close correspondence. The SOM, therefore,
provides a mathematical tool to interpret the height distribution profiles and obtain the
relative amounts of each molecular species under analysis e. g. monomer, dimer, tetramer.
Deviations from this model by adsorbed proteins imply a preferred orientation that is
presumably related to recognition between a protein domain and the surface. The extent of
deviation from the SOM may be useful in defining such preferred binding to ligand coated
surfaces or ligand induced conformational changes leading to the preferred orientation.

Data analysis consisted of two diflerent procedures, one to determine molecular
dimensions and the other to assign tertiary and quaternary structure. The raw data was first
tabulated and smoothed with a 5A window. A combination of Gaussian crrrves was then
used to fit the data, using Peakfit (Jandal Corp.). The center of the Gaussian curves
determined the measured molecular dimensions while their area determined their
prevalence. In order to assign a tertiary and quaternary structure for the macromolecule the
obtained measurements were used to propose the three dimensions (a, b, c) of a
parallelepiped which most closely resembled that macromolecule. Given the dimensions of
a proposed monomer the sizes of the higher oligomers were predicted and those predictions
were checked against the observed dimensions. Once the sizes of monomers, dimers and
higher oligomers were determined we used the SOM to obtain the predicted height
distribution of such species. Finally, we used a Mathematica routine (see Appendix E) to fit
the percentage of each oligorner in the data keeping the SOM values fixed. In this way we

17

have been able to analyze mixtures of different oligomers of a protein and detemrine their

size and concentration.

RESULTS

Measurements of the Height of Colloidal Gold Standards with
RAMS. Monodisperscd spherical colloidal gold particles were initially employed to
demonstrate that the RAMS technique is capable of making sub-nanometer height
measurements that are minimally influenced by tip geometry. The particles were chosen
because their sizes were comparable to that of protein molecules. Fifty microliters of
sample were deposited onto a 1 cm2 piece of mica that was mounted on a stainless steel
disk for imaging as described in materials and methods, then the sample was rinsed twice
with 200 pl aliquots of MilliQ water and allowed to dry overnight. This preparation
resulted in a uniform distribution of gold particles on the mica surface. The results of
height measurements on these particles are presented in Figrne 6 and Table 1. There is
good agreement between the manufacturer's stated size (determined with TEM) and the size
measured by SFM. The similarity between standard deviations as calculated for TEM and
SFM measurements suggest that the dominant source of error in these measurement may
reside in the size distribution of the particles themselves rather than the measurement

technique or tip geometry.

1 8
Table 1: Summary of sizes of colloidal gold particles by TEM and SFM.

 

 

 

 

 

Size Size ' Size by Number of Particles
Determined Determined by Gaussian fit to Measured
by TEM (nm) SFM (um) SFM Data (nm)
4.9 :t: 0.6 5.2 :i: 1.6 4.8:l:1.9 353
9.0 :l: 1.2 9.9 d: 2.0 10.4:16 207
18.1 :i: 1.2 17.8 :1: 2.8 16911.8 241 total
22812.6

 

 

 

 

 

 

Ferritin. To validate the RAMS technique for the measurement of the
macromolecular dimensions of proteins, we performed measurements on a diverse group
of proteins for which high resolution structures had been previously determined by either
electron microscopy or x-ray crystallography. The first protein employed using the RAMS
technique was horse spleen Ferritin. Ferritin has been demonstrated to be a roughly
spherical molecule comprised of 24 polypeptides (ratios of homologous H and L chains)
and bound iron (Fe3*). Fenitin has a molecular weight of 445,000 and it has been shown
to have an outer diameter of 120 A with a hollow center of about 80 A diameter by electron
microscopy43 and x-ray crystallography.“ The hollow center is used for iron storage.
RAMS Ferritin isolated from horse spleen yielded a distribution of measured heights with a
dominant peak at 112 A (see Figure 7 and Table 2), in excellent agreement with the
previously reported results for ferritin. Figure 7 is the RAMS data for Ferritin fitted with
three Gaussians centered at 73 A, 112 A, and 130 A, representing the iron core, the intact
ferritin molecule, and a presumed Ferritin trimer, respectively. The relative areas under the
curves are 20, 41 and 39% representing the iron core, monomer and trimer. The additional
peaks in the height distribution indicate a diversity of ferritin populations that are related to
the presence of the 80 A iron oxide core and Ferritin trimers,45 all of which have been
observed with adsorbed Ferritin samples on mica grids using STEM.

19

Wheat Germ Agglutinin ( WGA). Wheat Germ Agglutinin (WGA) is a plant
lectin (i.e. a carbohydrate binding protein that can agglutinate erythrocytes and other types
of cells) with aflinity for N-Acetylglucosamine (GlcNac), and GlcNac oligomers.“ The
dimer has two highly specific,“7 independent binding sites per polypeptide chain.“8 WGA
is composed of two identical subunits with molecular dimensions of the monomer
approximately 22 x 38 x 48 A each, and the subunits can join to form a dimer under
physiological conditions with dimensions 41 x 46 x 61 A.49 Figure 8 shows RAMS data
for wheat germ agglutinin, prepared as described in Materials and Methods. Part A is
RAMS datafitted to dimensions corresponding to the monomer, 20 x 35 x 46 A. The area
under the curves is determined by the SOM ratios for these three measurements. As
observed in Figure 7A and Table 2, SOM analysis demonstrates that WGA molecules are
found almost exclusively as monomers. Part B is the data fitted to the dimensions
corresponding to the dimer, 45 x 45 x 64 A. The poor fit shown in B suggests that there is
not a significant proportion of dimers in the molecules measured in this experiment. These
observations suggest that under the conditions of the experiment, in which low
concentrations of protein (5-10 rig/ml) are visualized, WGA dissociates into monomers.
To further test the possibility of a concentration dependent dissociation of WGA dimer to
monomer, WGA (1 mglrnl) was reacted with 2% gluteraldehyde to stabilize the dimer prior
to dilution for SFM analysis. As observed in Figure 9, the RAMS distribution can no
longer be attributed to pure monomer but has acquired a significant amount of dimer. Part
A shows that the observed RAMS data is no longer fit by WGA monomers but is more
closely fitted to a 55% monomer and 45% dimer combination, shown in part B. These
observations are consistent with other reports of a concentration dependent dissociation of
WGA into the monomer"0 and many reports in the literature describing the dissociation of
oligomeric proteins under similar conditions of high protein dilution (1-10 pg protein
/ml).‘"5‘

20

Concanavalin A (Can A). Concanavalin A (Con A) is a lectin, obtained from
the jackbean (Canavalia ensiformis), that binds to mannose, glucose, and glycoconjugates
containing these saccharides. Con A has been extensively used as a probe for
glycoconjugates in both animal and plant cells, and ean serve as a histochemical
probe.5253"“55 The native protein is composed of four identical subunits each with
molecular weight 25,500 Daltons"“s and molecular dimensions of approximately 33 x 39 x
57 A.’7

Figure 10 shows the height distribution profile for Con A that has been fitted to (A)
monomer, 28x36x57A, and(B)dimer, 28x46x68 A, and Figure 11 shows the same
datafittedto(A)tetramer46x 59 x78 A and (B) a mixture of 13% monomer, 67% dimer
and 20% tetramer. Although ConAis normally described as atetramerat pH 7.0,“5 it was
previously shown by Gordon58 that Con A ean dissociate into subunits as a consequence of
dilution and ionic strength. To determine whether Con A was a dimer under conditions
similar to that employed for SFM measmements, we examined the elution profile of Con A
(3 rig/ml) on high performance liquid chromatography (FIPLC). As observed in Figm'e 12,
the dominant molecular species at 3 pg/ml of Con A was the dimer (Fig. 12A). Further
confirmation was obtained by a comparison with the elution profile for FlTC-sCon A
obtained from a 5 mg/ml solution of FITC-sCon A, known to be a dimer at pH 7 .055”60
(Figure 128). Note the peak in both chromatographs at 8.6 minutes corresponding to Con
A dimers. From these HPLC measurements we show that Con A is predominantly a dimer
at the concentrations used in the RAMS experiments. Our experiments confirm the
observations of Gordon58 in which gel chromatography was used to demonstrate that the
dimer/tetramer equilibrium of Con A was sensitive to changes in the concentration of Con

A within the same ranges of concentration employed in the RAMS measurements.

21

Protein A. Protein A, a cell wall protein from Staphylococcus aureus, is a
monomer with four neariy identical domains, A, B, C, and D , each consisting of 60 amino
acids, and a fifth domain, the C-temrinal, with approximately 150 residues.“”32 The ability
oftheproteintobindthch panofIngromvarious specieshasmotivated its useasatool
in immrmochemical and cell-surface structural studies.63 The crystal structure of fragment
B is an oval 13 x 24 x 24 A,“ and the other homologous domains are expected to have
similar dimensions. From hydrodynamic studies and sedimentation equilibrium analysis,
the molecule was found to have a very extended shape. ‘5 It may therefore be concluded
that the five domains are arranged in an extended array, giving the molecule dimensions of
approximately 13 x 24 x 120 A. Such an extended shape would be expected to result in a
molecular orientation of Protein A on the mica surface in which adsorption predominantly
occurs along the long molecular axis. The expected height measured by RAMS would
therefore be 13 A. As observed in the height distribution profile (Figure 13, Table 2), the
dominant peak at 12 A agrees very well with the 13 A dimension derived from the crystal
structure. Part A shows RAMS data fitted to dimensions corresponding to the monomer of
Protein A excluding the C-terminal, 12 x 20 x 120 A. Part B shows RAMS data fitted to
the above measurements together with a 24% contribution from heights at 24 A and 40 A.
These are attributed to measurements on the larger C-terminal domain of Protein A. The
SOM fit confirms that the molecule, under the conditions of the experiment, is a monomer.
The RAMS data reveals structures at24 and 40 A which amount to approximately 24% of
the total area under the curve. These heights may be attributed to the C-terminal domain
whichcontain32.5 times more amino acids and is therefore expected to be largerthan the
other four domains. The data in Figure 138 has been fit to a combination of two structures,
onewithdimensions12x20x 120Aandtheother24x40x 120A.

22

Streptavidin. Streptavidin, a tetrameric component of the cell wall in
Streptomyces avidinii,“ has four sites of high affinity binding to biotin.67 The
Streptavidin tetramer has dimensions from the crystal structure of approximately 44 x 45 x
60 A and is comprised of monomers with the molecular dimensions 18 x 24 x 47 A.68 The
biological function of Streptavidin is poorly understood, but may involve an antibiotic
role.” Our results again demonstrate the disaggregation of the tetramer to form
predominantly amonomer (measured dimensions of 15 x 24 x 51 A) with minor amounts
of other oligomeric species. Figure 14 shows RAMS data for Streptavidin. The data has
been fitted to dimensions corresponding to the monomer of Streptavidin, 16 x 27 x 49 A,
with the SOM ratios determining the areas under the curves. Notice that for Streptavidin
the smallest dimension is more prevalent compared to the largest dimension than that
predicted by SOM, probably because of selective protein/substrate interactions. It seems
likely that the dissociation into ‘ monomer results from the high dilution of the protein
solution analogously to the previously seen dissociation of subunits in Con A and WGA.

To test the possibility of a concentration dependent dissociation of Streptavidin
from tetramer to monomer, a solution was made 2 mglrnl streptavidin in 10 mM HEPES
pH 7 .5, then made 0.3% in gluteraldehyde. It was incubated at room temperature for 45
minutes, then diluted to 0.002 mglrnl streptavidin and 0.0003% gluteraldehyde with 10
mM HEPES pH 7.5. The dilute solution was deposited on MgCL, treated miea and
incubated at room temperatme for 20 minutes, then rinsed twice with 200 yd aliquots of
MilliQ water. The protein was analyzed with the SFM and sizes tabulated, and the results
are presented in Figure 15. This treatment produces a distribution of 66% monomer, 34%
dimer, and 0% tetramer when analyzed with the Gaussian fitting routine in Appendix 13,
indicating that by fixing the protein at a high concentration, some of the protein molecules

are held together in the dimer form, but the native tetramer structure is still not retained. It

23

is possible, but not tested, that a higher gluteraldehyde concentration would cause complete

retention of the tetramer structure.

ADP-glucose pyrophosphorylase (ADngp). ADP glucose
pyrophosphorylase (ADngp) functions as the key regulatory enzyme in the biosynthesis
of bacterial glycogen" by catalyzing the formation of ADP-glucose from glucose 1-
phosphate and ATP. In E. coli, the enzyme is composed of four identieal subunits,71 each
with a molecular weight of about 50,000 kDa. 72 A mutant ADngp (G336D) is expressed
in E. coli [(12 strain 618 in which a single amino acid replacement occurs at position 336,
an aspartic acid is substituted for glycine.71 This mutant enzyme: a) stores about 33%
more starch than the wild type, b) is more resistant to its inhibitor AMP, and c) is less
dependent on the allosteric activator fructose l, 6-bisphosphate (FbP). Characterization of
a mutant enzyme from bacterial strain 865-504 by sedimentation equilibrium
centrifugation, polyacrylamide gel electrophoresis, and gel filtration column
chromatography shows that in the presence of FbP, this enzyme aggregates, while native
enzyme isolated from a wild type strain (AC70R1) does not exhibit this activator-induced
aggregation.72 The data suggest that a relationship may exist between altered chemical
activity in the ADngp mutants and changes in enzyme structure resulting in aggregation.
Molecular sizing by non-denaturing gel electrophoresis and Scanning Force Microscopy
was performed on wild type ADngp, the mutant ADngp containing a single amino acid
substitution (G336D) and an ADngp from the cyanobacteria Anabaena. Like the E. coli
enzyme, the enzyme isolated from Anabaena is composed of four identical subunits73 with
approximately the same molecular weight as the E. coli enzyme. There is a 33% sequence
homology between the ADngp from the two species, so a similar size and gross structure
is expected between these two molecules.“

24

Here we demonstrate the use of RAMS35 microscopy to rapidly determine the
molecular size of a family of wild-type and mutant ADngp molecules. The data obtained
from these measurements demonstrate that: a) a single amino acid replacement in the E.
coli ADngp G336D mutant can induce aggregation of the protein, b) the ADngp enzyme
from Anabaara shows different results indicating two oligomeric forms of the enzyme, the
monomer and dimer, and c) RAMS provides a high resolution tool for performing
comparative studies of structure/function relationships between wild type and mutant
macromolecules. Such analyses can now be performed without recourse to harsh fixation
techniques, high vacuums, crystallization, gel filtration, or ultracentrifugation. This
analysis shows that the protein sizing technique presented here can be useful for a protein
when the structure of the protein is not well understood.

As shown in the electrophoresis gel in Figure 16, G336D has a larger molecular
weight compared to the wild type enzyme. The enzyme from Anabaena shows two bands,
one at the lower molecular weight end of the broad E. coli enzyme band, and one much
lower than the E. coli enzyme band. The larger molecular weight for the G336D enzyme
must result from aggregation of molecules to form a higher molecular weight form. These
findings are corroborated with the RAMS data.

For RAMS analysis the three enzymes were prepared on mica substrates as
descfibedindreMatefialsandMeflrodssecfiomand the results are presented in Figure 17.
To provide a more detailed analysis of the data and a more graphic means to compare size
data between different proteins, the data are shown in a format similar to an electrophoresis
gel, where a darker band indicates a larger quantity of protein at the size corresponding to
the position on the graph. The size is plotted on a log scale for accurate comparison to the
electrophoresis gel. This plot clearly shows the type and amount of each molecular species
for each type of ADngp. It is observed that the single amino acid substitution in the

ADngp G336D mutant results in a considerable increase in the measured size when

25

compared to the wild type enzyme. The large increase observed reflects aggregation of the
mutant enzyme caused by the mutation imparting greater stability of the aggregate form.
Since this protein does not have a crystal structure solved, there is more potential
for error in interpreting the RAMS data, however, we have obtained enough data to nuke
reasonable predictions for molecular size for the monomer, tetramer and octamer of this
enzyme. In Appendix A,‘ two dimensional arrays of Streptavidin and ADng protein
molecules are described, formed when the protein concentration is higher than that required
to produce isolated molecules on the surface. The array is 60 A tall in the case of
Streptavidin, and 100 A tall in the case of ADngp. If an analogy is made between
ADng and Streptavidin,- we could conclude that the largest dimension of the tetramer of
ADngp is about 90 A, because the two-dimensional arrays of ADngp are 100 A tall and
the Streptavidin two dimensional arrays suggest that these molecules align themselves with
the longest axis as the height, and a slightly larger height of the two-dimensional array than
the longest axis of the tetramer. The other two dimensions of the protein are left to be
determined by the single particle RAMS measurements and by SOM. The molecular
weight of the tetramer is about 200,000 Daltons,72 and the volume of the molecule can be
estimated at 332 nm3 if one assumes the protein has a density equal to that of water. This
would nuke dimensions of the tetramer of approximately 50 x 70 x 90 A a possible set of
three dimensions of the molecule that would be consistent with RAMS results seen for the
wild type enzyme, shown in Figure 18. When the tetramer has this set of dimensions, the
octamer could be formed by joining two tetrarners with the largest face together forming a
molecule with dimensions 75 by 90 by 95 A. These dimensions agree well with the RAMS
data for the G336D E. coli mutant enzyme (Figure 19), suggesting that the point mutation
converting glycine to aspartate at position 336 causes aggregation of the enzyme from the
normal teuamertoanoctamer. ForthisE. colienzymeandtheenzyme fiom Anabaena,

both of which in the native state are tetrarners composed of identical subunits, the monomer

26

could have dimensions of about 25 x 35 x 90 A. From the electrophoresis and the RAMS
data, onecan see thatthe enzymefromAnabaena is composed ofasignificant proportion of
monomers and an additional proportion of a higher order oligomer. The fit of the RAMS
data for the enzyme from Anabaena is shown in Figure 20. Figure 20A shows the data
fitted to dimensions of the monomer alone, 25 x 35 x 90 A, and figme 208 shows the
RAMS data fitted to a combination of 68% monomer and 31% tetramer (dimensions 50 x
70x90A). Thefitusingthesedimensions misses a peak in the (hta at about 63 A. This
could be caused by small structural differences between the E. coli enzyme, from which the
tetramer dimensions were deduced, and the Anabaena enzyme. The two enzymes should
be quite similar in structure since the sequences are highly homologous. Finally, it can be
reasonably concluded based on RAMS and on electrophoresis that the Anabaena enzyme

exists as a combination of monomer and tetramer.

DISCUSSION

Analytical Methods for Measuring Molecular Dimensions and
Oligomerization of Macromolecules. Investigations with receptor proteins in
membranes, allosteric enzymes, and DNA binding proteins have all demonstrated a
structural promiscuity for proteins in their ability to undergo ftmctionally important
conformational changes and oligomerization in response to changes in protein
concentration, protein-protein interactions, ligand/effector binding and post-translational
modification.‘ To understand and measure such phenomenon has proven difficult, since
the concentration range in which purified proteins are studied is usually far above that
found in the intracellular environment. In many instances substrate or ligand induced
changes in protein conformation that are ftmctionally significant are masked as a

consequence of the necessity for performing measurements at high concentrations of

27

protein. Recent measurements with a variety of oligomeric enzymes clearly show that their
activities can be regulated by controlling their ratio of oligomers e. g. dimer to tetrarrrer.“2
In many cases membrane receptor activation requires ligand induced aggregation. 75
Enhanced binding/avidity and activity for a large number of diverse DNA binding proteins
e.g. p53, Rec A, Par A, CAP has been suggested to occur through substrate induced
oligomerization of DNA bound protein. ‘5'7'76'77 Such observations suggest that new
methods are necessary to determine the molecular dimensions of proteins at low abundance
and under conditions of protein concentration that are normally encountered within the cell.
These methods should. more accurately reflect the changes in conformation and
oligomerization that are suggested to be responsible for regulating biological activity.
Unfortunately, a limited number of analytical techniques are presently available with
the requisite sensitivity and resolution to measure the molecular dimensions and
organization of individual protein molecules under conditions that can maintain biological
activity of proteins at low protein concentration. New approaches with NMR have
provided the opportunity to measure the conformation of proteins in solution and to
determine the arrangement of specific amino acids within the polypeptide chain.
Oligomeriration and ligand binding studies have also been performed using these
methods.78 Although these techniques are useful, it is presently too complex to interpret
NMR data for molecular investigations of proteins larger than 40 kDa. A further linritation
is the need to use relatively high concentrations of protein and expensive NMR
instrumentation for these investigations. Electron microscopy (scanning (SFM) and
transmission (TEM)) may be successfufly employed to measure the molecular organization
of large proteins and macromolecular structures.7”°' A significant disadvantage of such
procedures, however, is that for most studies the preparation techniques involve
dehydration and metal coating, both of which can damage protein structure and certainly

destroy protein activity. In addition, electron microscopy is not particulariy sensitive to

28

smaller proteins (<100 kDa). Polyacrylarnide gel electrophoresis (denaturing and non-
denattuing), gel and liquid chromatography have both been useful in providing
measurements of molecular weight and protein oligomerization, but do not have the
resolution to provide molecular dimensions and generally require significant concentrations
of protein. Although sensitive to protein aggregation and protein shape, equilibrium
sedimentation ultracentrifugation can not provide molecular dimensions and requires high

concentrations of protein.

RAMS- A New Method for Measurements of Molecular Dimension
and Protein Oligomerization with Angstrom Resolution. The RAMS technique
provides a means to measure the molecular dimensions for the complete size range of
monomeric and oligomeric proteins. This approach extends the nanometer resolution of
SFM to that of angstroms by avoiding excessive image broadening of biological samples
introduced by the lateral distortions introduced by geometry of the probe tip. 15.1mm Tip-
protein interaction, mainly lateral and vertical pressure, can cause the molecule to appear
shorter. 1’ With regard to protein-substrate interaction, there is an attractive interaction
between the polar surface of the protein molecule and the charged mica surface, possibly
deforming the molecule and causing it to flatten against the mica surface, however, study
on lysozyme has shown that it retains its activity when adsorbed onto mica and imaged
with SFM, so these deformations must not be very extensive. 8‘ From Our measurements
we have observed 10% or less compression for the sizes 50A or larger, and very little
deformation in the smaller molecules.

By using a simple mathematical model, the statistical orientation model (SOM), to
interpret the height distribution profiles, it is possible to convert the observed distribution
of heights measured in a large population of randomly oriented molecules into 3
dimensional coordinates. A particularly important aspect of this model is that it provides a

29

means to determine the contribution of each molecular species, e. g. monomer, dimer,
tetramer, to the height distribution obtained by RAMS. An examination of the preferred
orientation of the molecules also provides the opportunity to predict the organization of
individual subunits within the protein oligomer. Measurements with angstrom resolution
may be performed on fimctional complexes. The sensitivity and resolution of the RAMS
. technique make it a powerful and unique tool for the determination of protein structure,
protein oligomerization as a regulator of enzyme activity, DNA/RNA organization, gene
regulation, chaperonin function, and the effects of post-translational modification on

protein organization and activity.

 

30

 

53:2 5:: a :3 39:25 < :3. v be 2:23 2: 2.522: a 3 £5__S=u_._c 98:... ”m “=st

r

.2:

cm.

  

00H:

mm

om

mm.

 

 

 

 

 

Fr

6
e

31

0.5 l

ure 4: 2.5 by 2.5 pm image of WGA on mica.

 

32

 

 

 

 

 

 

 

 

l
0.75

f b
L) 0.5
0.25
0

0 0.25 0.5 - 0.75 1
2
(a/(b*C))

Figure 5: Correlation between elongation and the probability that the two tallest dimensions are vertical

 

33

 

 

...m
B
..u.
... ..
cocoa
e ea 00..
0‘.
on”
......
.0 O” O
a: .

P P D D D P D D D P I F D D P DDDDDDDD

00 6 4 2 0

335$ we eon—=52

150
Height (A)

100

50

 

250

200

Figure 6: RAMS on 4.9, 9.0, and 18.1 nm colloidal gold fitted to Gaussians.

 

34

 

 

 

 

 

6, ............................ n. -
85f
—- l
8 .
.24:
g .
t... L
o3»
:- : .
.82’ ’ .
E ; '.". r
l; -0 ;
‘. ."'. l
0 r

 

 

25 50 75 100 125 150 1'75 200
Height (A)

Figure 7: RAMS on Ferritin.

 

35

 

 

 

 

 

 

 

 

       

 

 

 

10 - - . f g g - - Ta
A

:38: -. J
= : ‘
o O

2 I :
¢6r fl

2 E :

“a : .
54f ‘
0 L ‘

'2 ’ .
=21 1

Z ; j
0' - - . - - a, .

Height (11)
10-+- -fi #
LB ,

g l

38: j
9 ’ . . ‘

.2 : :
e6. . .. ‘

5 : -

"5 : . .

h4: ' '

g ; - .. .

g b ° .. j

22: . 1
0’ - - . - f"'°':m.:'“f°'..A - °"2-.-~°--,,- .

Height (A)

Figure 8: RAMS on WGA fitted to (A) monomer and (B) dimer dimensions.

36

...-s
O
l
d
J
i
t

 

O\ 00

v V v v

 

Number of Molecules

N

    

 

 

40 60" , 80 100
Height (A)

 

v ' v v v v ‘ v v v v I 1

Number of Molecules
N

 

20 40 60 80 100
Height (A)

 

 

-
v v v

 

Figure 9 RAMS on WGA fitted to (A) monomer and (B) 55% monomer and 45% dimer.

37

 

r—s r—t
O Ur

Number of Molecules
Ur

 

 

 

 

 

 

 

 

Number of Molecules

 

 

 

 

Height (A)

Figure 10: RAMS on Con A fitted to (A) monomer (B) dimer

 

' U I I I I I I I ' ‘ U ' ' I ' ' f """

)— r—-
0 UI
>

Number of Molecules
Ur

 

 

 

20 40 6O 80
' Height (A) .

 

 

f f T U fi' *1 T r

r—- r—t
O LII

Number of Molecules
Ur

 

 

O . L 4 . . . . . .
20 40 - 60 80
Height (A)

a a l a n m a L

Figure 11: RAMS on Con A fitted to (A) tetramer (B) 13% monomer, 67% dimer, 20% tetramer.

 

39

 

+‘Ililllllll1rljill‘lf

 

 

 

 

 

 

:A B 4 1.5
0.01 '— I
r d
: - 1.0
0.005 :— j
r d
- - 0.5
0 — 1
0.1). - U . 0.1)
1— —__________j _ 0
r r r r I r J r r l l l l l l l l l J -
0 . s 0 5 10

Time (min.)

Figure 12: I-IPLC on (A) 3 ng/ml Con A (B) 5 mg/ml FlTC-sCon A.

 

 

‘5

‘v V T v f v 1 w v v v v v a a
v v v v ' v v v 1 v v '7 v v i ' r

or
O
:D

N
O

 

Number of Molecules

p—r

 

 

 

 

 

 

20 40 60 80 100 120
Height (A)

40 fl v —v v v V V 1 fi' " ' v ' v v v ' v ‘

 

Number (3‘ Molecules
o

 

 

 

 

20 4O .
. Height (A)

Figure 13: RAMS on Protein A fitted to (A) monomer (8) part A plus 24% contribution from 24 and 40 A.

60 80 100 120

41

 

M
o

i

i

i

A

1

1

A
O

 

W
O

Number of Molecules
N
o

...-r
O

 

 

 

 

 

20 40 . 60 80
Height (A)

Figure 14: RAMS on Streptavidin fitted to monomer.

 

p—n
O

42

 

CD

Number of Lgolecules

 

 

.p.

 

 

  

 

80

 

Number of Molecules

' V v v V v v v

' v v v

 

N

 

Figure 15: RAMS on gluteraldehyde treated Streptavidin fitted to (A) monomer (B) 66% monomer 34% dimer.

 

 

Oi:

 

80

 

 

43

   

Q

'9‘. I

..I",

h

<3

a o =
>‘ \O 0
H m (B
'o m '3
= O =
B <

Figure 16: Non-denaturing gel on ADngp wild type, G336D, and Anabaena.

 

 

Height(nm)

 

 

 

.—

 

 

 

G336D

 

Anabaena

Figure 17: SFM data on ADngp wild type, G336D, and Anabaena

44

N
O

 

v v 1 v V v f w v V ‘v v v v t v v v f ' v v v v V v v

r—I ....
O Ur

Number of Molecules
Ur

 

 

 

 

 

20 . 40 60 80 100
Height (A)

Figure 18: RAMS data on ADngp wild type.

45

OJ
C

t

J

J

1

1

i

J

 

N
U]

N
O

Number of Molecules
r— l-|
o u:

U!

 

 

O
I

 

 

 

20 40 60 80 100 120 140
Height (A)

Figure 19: RAMS data on ADngp G336D mutant

20

Number of Molecules

~— N
u: o c?

Number of; Molecules
o

46

 

p—a
UI

p—s
0

UI

 

 

 

 

 

 

Ur

 

 

 

 

 

A
20 40 6O 80 100 120 140
Height (A)
B
- a - - - a A A - - A n+4.£:..:fil 3.. . - -
20 40 6O 80 100 120 140

Height (A)

Figure 20: RAMS data on ADngp from Anabaena fitted to (A) monomer (B) 68% monomer 31% tetramer.

47

.50.. wit... 0... 0. a... 00.3....» 0... 50... 0000...... «a .0550. 0... .05... ..050005 .0 0...... .

...050...0. ....0 850005 .20.. 50.... 20.50.5500 00.. < on .0 .00.. 0.... 8...... so... 0... ...... ..05... ...... 350005

50.. 50... «03.3.3000 ...... < am .0 .00.. 0... .00.... 00.... 0... 0.... .8583. ...... .05... £0... .0 30.53.5000 ...... < 9. .0 .30.. 0... 8...... 00... 0... .< 5.0505000 .0...

60...... 20a 0... 0. 002.3500 8.80 ..0.. .053 05.3.3502? ..000 ...... 0.0.0.... ..000 .0. 08022.00 00..- .0.0. 0... ..0 09:08.2. 0 v... 005000.... ....2. 0... 8...... §< .

....0. 0... ... 000.58.. 0... 00300.00 0. A...... 6.30.05 0... .0 :32. 0... a. 0...? 9.23.0035". 0... ...... ......» 0050.00 3.08.05 0... ..0 09.5005... 00.0.00... .
.530... ... :30... .....s 000.33...
83 0.30.05 0.... ....0. 0... ... 00880.0. 8 0.5.90.5 5.0.0.... 5035.00... 0... ....3 00:800.. 80.0....» .250 0... 50... 00.0805 30.5.05... 0... 0... 82.... .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

on an.
.4 a: .
8 2. n.” 8. . 5......
a 0N. _ 8.
«a 8 «u ..N
.... N. 2. n. I .0523... < 50.8..
a on
..n 9 6
c on 3 .3533.
a a. 30.3.8.0
«N ....
c on ..N 6 as... 5.25585
N 9. .. ....
..N ..N mm on 005000.:
_ .5 2 an .. b 533.585
8 .0
n.” 0..
. .. .. 6
... 9. ... .... :5... <03
5 n.” a an
. am am no «N D 505000... <95
m. a. n. a
... a 8 on i
8... a. 0.. R .v
... a. a 2.
a 8 a ... i .055... < .30
S... 2 ..a 8 .3
a on 2 R as... < .30
3.. on an on
...... R an 00 nn 0 0050005 < :00
.034: War-500.00 2. 2.0.505... £05082. 5..
8.3.8.825 2% 82.30 02:32 .28 30...... a8.2.2.5 2:832

 

.....00..m0._...mb0 >8-.. ...... 30820.5 as}; 50.... ..0>..0.. 0.. 000.0005... 0.0.0.... 0... .0 b.5555 "N 0.5....

 

CHAPTER 3 : LIGAND BINDING STUDIES

INTRODUCTION

Recent measurements with a large number of diverse DNA binding proteins
including Rec A, Par A, and CAP, have suggested that enhanced binding and activity occur
through substrate-induced oligomerization of DNA bound proteins.”33'“ A mutant ADP-
glucose pyrophosphorylase has demonstrated activator-induced oligorrreriration.72 Thus it
is critical to the understanding of protein structure to be cognizant of substrate-induced
conformational changes. Wrth the tools and theory presented in the previous chapter,
studies can be done to determine ligand-induced oligomerization of proteins. RAMS” has
proven to be a technique readily applicable to this type of study as will be demonsu'ated in
experiments on Wheat Germ Agglutinin and Streptavidin.

Wheat Germ Agglutinin (WGA) binds N-acetylglucosamine (GLcNac) and
derivative sugars (i.e., the B-(l-4)-linked oligomers such as di—N-acetylchitobiose or tri-N-
acetylchitobiose)“'85'“ known to be present on membrane surfaces of muss-’38” The
binding of WGA to oligosaccharides of GLcNac displays even higher aflinities, with a
dissociation constant of 4.0 and 1.2 x 105 M for the di- and ui-saccharide, respectively,
compared with a dissociation constant of 7.6 x 10“ M for the mono-saccharide.47
Tryptophan residues at the binding site seem to be essential for binding activity, as shown
by examining binding following u’eaunent with a tryptophan modifying agent.90 RAMS
analysis indicated that WGA was a monomer under the high dilution and low ionic strength
conditions used in experiments described in chapter 2. It was suspected that the subunit
association may be strengthened by substrate binding, so a study was undertaken to
determine if WGA would be stabilized in the dimer form upon binding GLcNac, di-
GLcNac, and tri-GLcNac. Since the high dilution conditions were suspected to cause
dissociation of the subunits, the protein was treated with the sugar before dilution. Also,

48

49

experiments were performed treating the protein with the sugar after dilution to measure the
ability of the sugar to promote re-association of the subunits.

Streptavidin, a protein produced by the actinobacterium Streptomyces avidinii, is
remarkable for its extraordinarily strong aflinity for d-biotin. The dissociation constant for
the streptavidin-biotin link is 1015 M."1 This protein has a tetrameric structure, and each
identical subunit has a molecular weight of 13.5 ltDa91 and dimensions from the crystal
structure 18Aby 24Aby 47 A. The tetramerhas dimensions 42Aby 42Aby 56A.“
Despite the lack of sulfur containing residues and disulfide bonds,"2 the tetramer is
relatively stable with regard to dissociation, and this stability is increased upon binding
biotin.93 One notable structural feature is that contacts made by tryptophan (Trp)-120 of
one subunit with biotin bind by an adjacent subunit through the dimer-dimer interface,
where two stable symmetric dimers are associated to form a tetramer having dihedral D2
symmetry.92 In an experiment by Cantor’s group in Boston University the wild type
protein bound to biotin did not dissociate from tetramer to monomers in an SDS gel unless
the protein was heated prior to running the gel. A mutant was constructed which had a
biotin-Streptavidin dissociation constant of about 10‘8 M compared to the biotin-
Streptavidin dissociatibn constant for wild type of 1015 M.“ This mutant Streptavidin,
which bound biotin less strongly because of conversion of tryptophan 120 to phenylalanin,
dissociated into monomers in an SDS gel whether or not the protein was heated prior to
running the gel,94 indicating that biotin binding plays a pivotal role in increasing the
stability of the quaternary structure.

MATERIALS AND METHODS

Materials. Materials used were described in the previous chapter, with a few
additions. Biotin and N-Acetylglucosamine (GLcNac) were obtained from Sigma. Di-

50

GLcNac and tri-GLcNac were a gift from Professor Melvin Schindler, Department of
Biochemistry, MSU. RAMS data was collected and analyzed as described in the previous
chapter, and the Mathematiea routine in Appendix E was used to evaluate the proportion of

the different oligomers in the protein size distribution.

WGA-GLcNac sample preparation. WGA was treated with the sugar while
still in the concentrated form, in an effort to avoid dilution induced dissociation of the
subunits of the dimer WGA. The monosaccharide treatment was performed in the
following manner: A solution was made 4.8 mglrnl in wheat germ agglutinin (WGA) and
0.16 M N-acetylglucosamine (GLcNac) in 10 mM I-IEPBS pH 7.5. After about 30 minutes
of incubation at room temperature, the solution was diluted with 10 mM HEPES pH 7.5,
making a solution 2.4 pig/ml in WGA and 0.00008 M in GLcNac. The dilute solution was
deposited on MgCl2 treated mica and incubated at room temperature for 20 minutes, then
rinsed twice with 200 pl aliquots of MilliQ water. The disaccharide treatment was
performed as follows: A solution was made 1 M di-GLcNac and 6 mg/ml WGA. After
incubating at room temperature for several hours, the above solution was diluted to 0.01 M
GLcNac and 0.06 mg/ml WGA, and after refrigeration for 7 days diluted again to 0.0003
M GLcNac and 1.8 rig/ml WGA. The dilute solution was deposited on MgC]2 treated mica
and incubated at room temperature for 20 minutes, then rinsed twice with 200 pl aliquots of
MilliQ water. The tri—GLcNac treatment was as follows: A solution was made 0.2 M tri-
GLcNac and 6 mglrnl WGA. After incubating at room temperature for several hours, the
above solution was diluted to 0.02 M GLcNac and 0.06 mglrnl WGA, and after
refrigeration for 7 days diluted again to 0.0006 M GLcNac and 1.8 pig/ml WGA. The dilute
solution was deposited on MgCl2 treated mica and incubated at room temperatrue for 20
minutes, then rinsed twice with 200 pl aliquots of MilliQ water.

51

To asses the ability of GLcN ac to promote reassembling of the WGA monomers
into the native dimer form, the protein was diluted before treatment with the three sugars.
The mono—GLcNac preparation was done as follows: Twenty pl of 0.06 mg/ml WGA in 10
mM HEPES pH 7.5 was added to 1.9 mg solid mono-GLcNac, making a solution 0.43 M
in GLcNac. After 30 minutes the solution was diluted 10x with 10 mM HEPES pH 7 .5,
and deposited on freshly cleaved mica along with 20 pl 5 mM MgClz. The di-GLcNac
preparation was done as follows: Forty pl of 0.06 mg/ml WGA in 10 mM HEPES pH 7.5
was added to 4.9 mg of solid di—GLcNac, making a solution 0.28 M in di-GLcNac. After
30 minutes the solution was diluted 10x with 10 mM HEPES pH 7.5, then deposited on
MgCl2 treated mica. The tri-GLcNac preparation was done as follows: Ten pl of 0.06
mg/ml WGA was added to 0.9 mg of solid tri-GLcNac, making a solution 0.14 M in tri-
GLcNac. After 30 minutes the solution was diluted 10x with 10 mM HEPES pH 7.5 and
deposited on MgCl2 treated mica. All samples were incubated at room temperatme for 10
minutes, then rinsed twice with 200 pl aliquots of MilliQ water and allowed to dry
overnight in petri dishes under ambient temperature and humidity.

S treptavidiu-biotin sample preparation. The biotin treatment was performed
as follows: a solution was made 0.03 mg/ml in streptavidin and 0.008 M biotin in 10 mM
HEPES pH 7 .5. That was diluted to 0.003 mg/ml streptavidin and 0.0008 M biotin several
hours later. The dilute solution was deposited on MgCl2 treated mien and incubated at room
temperature for 20 minutes, then rinsed twice with 200 ml aliquots of MilliQ water.

The ability of biotin to promote the reassembling of the Streptavidin monomers was
tested by first diluting the Streptavidin to 2 pg/ml in 10 mM HEPES pH 7.5, then after
incubation at room temperatme for 10 minutes making this solution 0.0008 M in biotin.
The solution was deposited on MgCl2 treated mien and incubated at room temperature for
20 minutes, then rinsed twice with 200 pl aliquots of MilliQ water. i

52

RESULTS

WGA . In all cases, the pre-dilution GLcNac treatment produced both monomers
and dimers distributed on the mica, while the post-dilution treatment produced height
distributions indicative of monomer only. This further confums that a dissociation of
WGA from dimer to monomer” is caused by dilution of the protein. The separation of
subunits does not seem to be reversible by the addition of GLcNac after dilution.

The pre-dilution mono-GLcNac heated WGA RAMS data is presented in Figure
21. ThedatahasbeenfittedtomonomerdimensionsB x32x 48 A (Figure 21A) and
dimer dimensions 36 x 48 x 60 A (Figure 213). Neither monomer nor dimer fits the data
well, but as shown in Figure 22, a combination of 50% monomer 50% dimer improves the
fit considerably. The post-dilution mono-GLcNac treated RAMS data is presented in
Figure 23. Figure 23A shows the datafitted to monomer dimensions 21 x 32 x 47 A, and
figure 238 shows the data fitted to dimer dimensions 39.5 x 47 x 59 A. The combination
of the monomer and dimer curves indicate that these molecules are composed of 99%
monomer and 1% dimer.

The pre-dilution di-GLcNac WGA RAMS data was fitted to dimensions
corresponding to monomer, 25 x 33 x 46 A, and dimensions corresponding to dimer, 40 x
46 x 60A. Figure 24 shows that a mixtme of 68% monomer and 32% dimer fit the data
much better than either monomer or dimer alone. The RAMS data for the post-dilution di-
. GLcNac WGA treatment is presented in Figure 25. The fit to (A) the dimensions
corresponding to monomer, 19 x 29 x 43 A, and (B) the dimensions corresponding to

53

dimer, 42 x 42 x 59 A, clearly show that all molecules are accounted for by monomer
only.

The RAMS data for the pre-dilution tri-GLcNac WGA treatment is presented in
Figure 26. The data was fitted to dimensions corresponding to monomer, 25 x 38 x 46 A,
and dimensions corresponding to dimer, 38 x 46 x 63 A. The best fit to the data is
produced by a linear combination of 81% of the monomer curve and 19% of the dimer
curve, shown in Figure 26. The post-dilution treatment with tri-GLcNac produced the data
in Figure 27. fits to dimensions corresponding to (A) monomer, 23 x 34 x 46 A, and (B)
dimer, 40 x 46 x 56 A, reveal that this data is produced by monomers only of WGA.

S treptavidiu. When a concentrated solution of Streptavidin is treated with biotin
then diluted, the molecules retain their native tetramer conformation. However, when the
solution is first diluted then biotin is added, the molecules dissociate into a mixture of
dimers and monomers. This indicates that the dissociation of Streptavidin into subunits,
caused by the dilution, is not completely reversible, but leaves most of the molecules in the
dissociatedstate. ThiscanbecomparedtotheRAMS measurementsmadcwithoutthe
addition of biotin at all” (chapter 2), which show the Streptavidin molecules completely
dissociated into monomers. In the dilute protein solution biotin seems to play a role in
partially stabilizing the dimer, even making possible the rejoining of monomers to dimers,
but is not able to bring about complete return to the tetramer state. If biotin is already
bound to Streptavidin the dimer-dimer attraction is increased to the point that only tetramers
are detectable. Further dilution at this point does not dissociate the tetramer.

The pre-dilution biotin treatment produces a distribution of molecules with the main
peak in the height histogram at 60 A as shown in Figure 28. The curve is the SOM
prediction for Streptavidin in the tetramer oligomaization state. The data does not closely
fit the SOM predictions in this case, probably because there is preferential orientation of the

54

molecules in their upright orientation as opposed to the majority of the molecules oriented
having the largest surface area in contact with the mica surface, as is assumed for SOM.
When Sueptavidin forms two-dimensional crystals the molecules also orient standing
upright, as is shown in Appendix A. It is likely that the molecule has chemical aflinity for
the mica that preferentially turns it upright, both as individual molecules and when two-
dimensional crystals are formed.

The Streptavidin solution treated with biotin after dilution produced the RAMS data
presented in Figures 29 and 30. The data is fitted to dimensions corresponding to (A)
monorrier18 x24x47A, (B)dimer,28x44x45 A, in Figure 29 and (A) tetramer, 44 x
45 x 56, and (B) a linear combination of 24%monomer and 76% dimer, in Figure 30.

DISCUSSION

Here we have demonstrated that the RAMS technique can be readily applied to
study ligand-induced conformational changes. For both wheat germ agglutinin and
Streptavidin ligand binding causes a shift in the distribution of oligomers from lower to
higher oligomerization state if the protein is treated with the ligand before dilution. This
ligandeffectishighlyreducedoreliminatedifligandtreatrmntis performedafterthe
protein is diluted. ‘

The shift in oligomerization is readily observable for the mono-, di- and tri-
GLcNac treated WGA. The statistical model developed in chapter 2” holds well for these
molecules as judged by the reasonably good fit between the observed and predicted
distribution of oligomers. One may therefore conclude that there is no preferential
orientation beyond that predicted by SOM. It is not surprising that GLcNac treatment does
not cause the rejoining of subunits in the dilute solution, since the low protein concentration

makes the average intermolecular distances in the solution large.

55

For Streptavidin, biotin binding appears to definitely promote the retention of
tetramer conformation in the concentrated solution, but preferential orientation of the
molecules on the mica make SOM fail in this case. It can reasonably be concluded,
however, that the Streptavidin in the pre-dilution biotin treated solution does not dissociate
into an oligomer smaller than the tetramer, since the height observed is larger than any
dimension of monomer or dimer for this molecule. It is furthermore reasonable that the
oligomer observed in this case is tetramer, since the height of two-dimensional crystals of
Streptavidin is 60 A, indicating that preferential orientation of individual Streptavidin
molecules in the same manner is also highly possible. Again, it is not surprising that the
post-dilution biotin treatment does little to promote reassembly of the tetramer, considering
the high dilution of the protein at the time the biotin is added. It is interesting that a
significant proportion of dimers are found in the post-dilution biotin treatment of
Streptavidin, but if biotin treatment is omitted, the protein forms monomers.

The difference. in oligomerization effects between post-dilution and pre-dilution
ligand treatment for both WGA and Streptavidin show that the dissociation of subunits
frequently seen in chapter 2 for Con A, Streptavidin, and WGA is indeed caused by the
low dilution of the protein used in those studies. Since pre-dilution treatment causes an
increase in the measured size for both Streptavidin and WGA but post-dilution treatment
causes little or no increase in the measured size, this experiment confirms previous
experiments which conclude that high dilution causes dissociation of oligomeric

proteins?”8

56

O\
r
1
1
4
i
l
l

 

 

 

 

 

 

 

 

 

{A :
35‘. . :
_ ' .
=3 ' .
o ’ I
24- -
a I i
“3 1
° . :
r- ; i
.82: ... j
E : .0...” O. ..
£1 O O... :

. 1

0.’ -.---Mn --r.; ‘
20 40. 20 80
Herght()

:B :

35: . :
fl ’ ..
a : O. O.
243 :
g , i
«H3 1
e . :
t- I I
32: :
s: r
231% ':
0’

 

 

 

 

 

 

20 40 60
Height (A)

80

Frgure 21: Bra-dilution mono—GLcNac treated WGA fitted to (A) monomer (B) dimer.

57

ON
J
r
r
r

 

--.---
O

A

N

Number of Molecules
i—t DJ

 

 

c?
L

20 4O 60 80
Height (A)

Frgure 22: Pie-dilution mono-GLcNac treated WGA fittedto 50% monomer, 50% dimer.

 

58

 

p—a
O\ OO O

1

J

l

q

d

Number; of Molecules

 

 

20 4o vvvvv 60 80

AJALLA

 

 

' A
10 remefirljeelgahtg?+

# ON 00

Number of Molecules

 

O

 

 

 

 

20 80

Height (A)

Figure 23: Post-dilution mono-GLcNac treated WGA fitted to (A) monomer (B) dimer.

59

...—i
ON

 

w v w v T v w r v ' v f f r I v v fi v V v v v v

_
00 N

Number of Molecules
.h

 

 

_o

20 40 60 80
Height (A)

Frgure 24: Pro-dilution di-GLcNac treated WGA fitted to 68% monomer, 32% dimer.

 

60

j—a
O

4

1

1

4

J

r

l

l

l

4

 

# O\ 00

Number of Molecules
N

 

 

 

OO

 

0\

Number; of Molecules

N

 

O
l
i.
r
L

0
b

 

 

 

 

20 40 60 80
Height (A)

Frgure 25: Postdilution di-GLcNac heated WGA fitted to (A) monomer (B) dimer.

61

O\
l
d
r
J
1
J
J

 

..1..
O
O

4;

Number of Molecules
N w

p—d

 

 

 

C?
l

 

 

20 40 60 80
Height (A)

Figure 26: Pre-dilution h'i-GLcNac heated WGA fitted to 81% monomer 19% dimer.

62

 

p—h
O\ 00 O
J
1
1
l
1
1

A

 

Ntbmber of Molecules

 

 

20 40 60 m 80

 

 

A O\ OO

Ngmher of Molecules
' O

 

 

O

A A l J A A A

 

 

 

20 4o . §0 80
Height (A)

Frgure 27: Post-dilution hi-GlcNac treated WGA fitted to (A) monomer (B) dimer.

63

 

j—a
-F O\ OO O
1
a
a
a
o
a
a
0

Number of Molecules

N

 

 

 

 

O
8

 

 

 

20 4o 60
Height (A)

Frgure 28: Pre-dilution biotin heated Sh'eptavidin fitted to tehamer.

 

64

N
O

 

v w h v 1 v W 1—' ‘ v w v 1 t v v v T V v v v v

...-t
Ur

Number of Molecules
in 3

 

c?

 

 

 

N
O

 

 

p—h
U!

Number of Molecules
o. E‘:

 

 

 

 

 

Height (A)

Frgure 29: Poshdilution biotin heated Sheptavidin fitted to (A) monomer (B) dimer.

65

 

N
O
l

1

J

t—t
LII

Number of Molecules
o. S

     

 

 

 

20 4O 6O 80

 

 

20 -a-..-e--!Ie.ig.hUA.)_

p—t
Ur

t—a
O

U!

 

Number of Molecules

 

 

CP
1

20 40 po 80
Height (A)

 

 

 

Figure 30: Post-dilution biotin treated Sheptavidin fitted to (A) tehamer (B) 24% monomer 76% dimer.

CHAPTER 4: CONCLUSIONS

In this thesis we have presented numerous examples of how RAMS microscopy
can be used to detemrine molecular size and ligand-induced conformational changes.
Experiments with colloidal gold particles have demonshated that the height measurements
used in RAMS are not afl‘ected by the artifacts caused by finite tip size. Analysis of Ferritin
by RAMS microscopy has demonshated that protein molecules can be accurately sized
using SFM RAMS on WGA and Con A have shown that the technique can accurately
measure the three dimensions of an asymmehical molecule. The study of Protein A has
demonshated that even a molecule with a very irregular shape consisting of an elongated
shing of 4 small domains and one larger domain can give results consistent with the actual
shape of the molecule, identifying the smaller sized part and the larger sized part of the
molecule. The RAMS analysis of Sheptavidin shows how the quaternary sh'ucture of a
protein can be deduced using this technique. The application of RAMS to determine
ligand-induced changes in oligomerization of protein molecules has been shown for two
molecules, wheat germ agglutinin and Sheptavidin. This demonshates that RAMS35 is the
technique of choice for the of study ligand-induced conformational changes for most other
proteins.‘”'“"'7

Random Adsorption Molecular Sizing is the only technique which can furnish direct
measurements of molecular size without requiring high protein concenhation or harsh
fixation methods. Techniques exist which can be used to give an estinmte of molecular
size, but each has disadvantages compared to RAMS. Gel filhation is commonly used to
determine the size of oligomeric proteins,95 but the column must be calibrated by passing a
set of proteins of known dimensions though it, and the reliability of the measurement is
dependent on how closely the protein of interest resembles the proteins used to calibrate the
columns"5 Since the protein of interest is of unknown shape, this makes the reliability of

this technique uncertain. Sedimentation equilibrium measurements are dependent on the

66

67

amino acid composition, overall shape, hydration, and surface roughness of the
molecule.” SDS-PAGE will give an accurate molecular weight for a monomeric protein,
but generally does not give information about quaternary shucture because the technique
involves dissociation of subunits with detergents.98 Non-denaturing gel elechophoresis is
highly afl‘ected by molecular charge, which may be changed upon binding a highly charged
ligand, so this makes ligand binding studies using this method impractical.” The growth
of crystals with the quality required for X-ray crystallography can be a diflicult and time
consuming task. All of these techniques require high concenhations of protein, a possible
problem when one wants to study concenhation-dependent oligomeric changes.
Microscopy techniques such as SEM and TEM may be used for large molecules such as
Ferritin, but usually require the sample to be dehydrated and metal coated, eliminating the
possibility that the molecule is in an active conforrrration.79'8° RAMS can reveal the three-
dimensional size and oligomerization state of molecules that still hold their solvation
water,“1 and give direct measurements of size without relying on uncertain calibration with
other protein molecules, and can produce these size measurements with only a few minutes
of sample preparation time. This technique should find many applications in the study of
the interrelationship of protein shucture and function.

The technique to determine molecular size presented here, termed Random
Adsorption Molecular Sizing, has demonshated its usefulness in measurements on a
number of different molecules. The use of the projected height of adsorbed protein
molecules as the defining parameter for molecular dimensions makes the RAMS technique
minimally dependent on probe tip geomehy and extends the range of the SFM from
nanometer to angstrom resolution. The relative ease and speed with which these high
resolution measurements of molecular dimension may now be performed should provide

investigators with new opportunities to examine changes in protein conformation and

68

oligomerization that depend on protein concenhation and ligand or allosteric activator

binding.

REFERENCES

References

 

’ Kosk-Kosicka, D., Bzedga, T., Wawrzynow, A. (1989) J. Biol Chem. 264, 19495-9.
2 Kosk-Kosicka, D., Bzedga, T., (1988) J. Biol. Chem. 263, 18184—18189.

3 Cgllso7nésg. A., Parsons, T. F., & Preiss, J. (1976) The Journal of Biological Chemistry

‘ Stenger, J. E., Teghneyer, P, Mayr, G. A., Reed, M., Wang, Y. Wang, P., Hough, P. V. C.,
Masnangelo, I. A. (1994) EMBO J. 13 6011-6020.

5 Davey, M. J., Funnel], B. E. (1994) J. Biol Chem. 269 29908-29913.

6 Koblan, K, S., Ackers, G. K. (1991) Biochemistry 30, 7817-7821.

7 Takahashi, M. (1989) J. Biol. Chem. 264-(1), 288-95.

8 Bennig, G., Rohrer, H. (1982) Helv. Phys. Acta., 55, 726.

’ Bennig, G., Quatc, c. F., Gerber, C. (1986) Phys. Rev Lett., so, 930.

‘° Pohl, D. w., Denk, w., Lanz, M. (1984) Appl. Phys. Lett. 44, 651.

u Lal, R., John, S. A. (1994) Am. J. Physiol., 266, C1-C2].

‘2 Marti, O., Amrein, M. SLM_a_nd SFM in Biology, Academic Press, New York (1993).

‘3 Puhnan, C. A. J; Van der Werf, K. 0.; Grooth, B. G.; Van I-Iulst, N. F.; Greve, J. (1994)
Appl. Phys. Lett. 64 (18) 2454-2456.

“ Zhong, Q.; Inniss, D.; Kjoller, K; Elings, V. B. (1993) Surf. Sci. Lett. 290 L688.
‘5 Keller, D. J., & Franke, F. S. (1993) Surf. Sci. 294, 409.
“ Eppel, S. J., Zypman, F. R., & Marchant, R E. (1993) Langmuir 9, 2281.

‘7 Allen, M. J., Hud, N. V., Balooch, M., Tench, R. J., Seikhais, W. J., & Balhorn, R.
(1992) Ultramicroscopy 42-44, 1095.

‘3 Markiewics, P., & Gob, M. C. (1994) langmuir 10, 5-7.

‘9 Radmacher, M., Cleveland, J. P., Fritz, M., Hansma, H. G., Hansma, P. K. (1994) Biophysical
Journal 66, 2159-2165.

2° Ducker, W. A.; Xu, Z.; Israelachvili, J. N. (1994) Langmuir 10, 3279-3289.
2' Rabinovich, Y. 1.; Yoon, R.-H. (1994) Langmuir 10, 1903-1909.

2’ Weisenhorn, A. L.; Maivald, P.; Butt, H.-J.;Hansma, P. K. (1992) Physical Review B 45,
11226-11232.

69

70

 

1’ Tsao, Y.H.; Evans, D. F.; Wennershem, H. (1993) Science 262, 547-549.

3‘ F. Schabert; A. I-Iefti; K. Goldie; A. Stemmer; A. Engel; E. Meyer; R Overney; H.-J.
Giintherodt, Ultramicroscopy 42-44, 1118-1124 (1992).

’5 Ohnishi, S., Hara, M., Furuno, T., Sasabe, H. (1992) Biophys. J., 63, 1425-1431.

2‘ Radmacher, M., Fritz, M., Cleveland, J. P., Walters, D. A., Hansma, P. K Langmuir, 10,
3809-3814 (1994).

’7 Hiberle, W., Horber, J. H. K., Ohnesorge, F., Smith, D. P. E., Binnig, G. (1992)
Ultramicroscopy, 42-44, 1161-1167.

1' Bustamante, C., Vasenka, J., Tang, C. L., Reese, W., Guthold, M., & Keller, R.
(1992) Biochemistry 31, 22.

'9 Hansma, H. G., Sinsheimer, R. L., Li, M. Q., & Hansma, P. K. (1992a) Nucleic Acids
Res. 20, 3585. and Hansma, H. G., Vasenka, J., Seigerist, C., Kelderrnan, G., Morret
H., Sinsheimer, R. I... Elings, V., Bustamante, C., & Hansma, P. K. (1992b) Science,
242, 209.

3° Henderson, E. (1992) Nucleic Acids Res. 20, 445.

31Shaiu, W.-L., Vesenka, J ., Jondle, D., Henderson, E., & Larson, D. D. (1993) J. Vac.
Sci. Technol. A 11(4), 820-823.

3’ Roberts, C. J., Williams, P. M., Davies, J., Dawkes, A. C., Sefton, J., Edward, J. C.,
Haymes, A. G., Bestwick, C., Davies, M. C., & Tendler, S. J. B. (1995) Langmuir 11,
1822-1826.

3" III, C. R., Keivens, V. M., Hale, J. E., Nakamura, K K, Jue, R. A., Chang, S.,
Melcher, E. D., & Drake, B. (1993) Biophys. J. 64, 919.

3‘ Marchant, R. 13., Lea, A. S., Andrade, J. D., & Bockenstedt, P. (1992) Journal of
Colloid and Interface Science 148, 261.

”Gilchrist, M., PfcISS,J., Schindler, M., Dantus, M. ”117-534 : Ir ' 01in __ ' turn 10 1 Li '
0 Comri ...-inc. ' ’ mlanI-r _' n '11”: ‘ - ' .-. ' '
Ulnar... o no ..or . .iochemistry (in preparation).

3‘ Cantor, C. R. W W. H. Freeman, San Francisco 1980.

 

3" Sommerville, J., Scheer, U. P _ .- -
men, IRL Press, Washington, DC 1987.

38Dykstra, M. J. . ._ _ _-
Plenum, New York 1993.

’9 Omstein, L., & Davis, B. J. (1962) W Reprinted by Distillation
Products Industries, Rochester.

‘° Bezanilla, M., Manne, S., Laney, D. E., Lyubchenko, Y. L., & Hansma, H. G. (1995)
Langmuir 11, 655.

 

 

71

 

“ Hu, J., Xiao, X.-D., Ogletree, D. F., & Salmeron, M. (1995) Science 268, 267.

‘2 Thundat, T., Zheng, X.-Y., Chen, G. Y., & Warmack, R. J. (1993) Surface Science
Letters 294, L939—L943.

° Harrison, P. M. (1963) J. Mol. Biol. 6, 404.

“ Massover, W. H. ( 1993) Micron 24, 389-437.

‘5 Yau, S.-T., Zhou, Y. (1995) Modern Physics Letters B 9, 187-193.

“ Allen, A. K., Neuberger, A., Sharon, N. (1973) Biochem. J. 131, 155-162.

" Nagata, Y., Burger, M. M. (1974) J. Biol. Chem. 249, 3116-3122.

‘3 Privat, J. P., Delmotte, F., Monsignay, M. (1974) FEBS Lens. 46, 224—228.

‘9 Protein DataBank, 9WGA, C. 8. Wright

5° Steindler, D. A., Bradley, R. H. (1983) Neuroscience 10(1), 219—241.

5‘ Denman, R. B., Wedler, F. C. (1984) Arch. Biochem. Biophys. 232(2), 427-40.

5’ Bowles, D.J., Pappin, D.J. (1988)TIBS 13, 60-64.

’3 Sharon, N., Lis, H. (1989)Science 246, 227-235.

5‘ Kéry, V. (1991)Int. J. Biochem. 23, 631-640.

5’ Lin, S. S., Levitan,1. B. (1991) Trends Neurosci. 14, 273-278.

5‘ Reeke, G.N., Becker, J. W., Cunningham, B. A., Wang, J. L., Yahara, I., Edelman,
G. M. Structure and Function of Concanavalin A. From W. T. K.
Chodhury and A. K. Weiss Ed. Plenum Publishing Crop., New York

5" Protein DataBank, 1CN1, Shoham et. al.

5' Gordon, J. A., Young, R. K. (1978) The Journal of Biological Chemistry 254, 1932-
1937. c

’ Gunther, G. R., Wang, J. L., Yahara, 1., Cunningham, B. A., Edelman, G. M. (1973)
Proc. Nat. Acad. Sci. USA 70, 1012

‘° Dumas, M., Schwab, M. a, Thoenen, H. (1979) J. Neurobiol. 10, 179—197.

“ Gouda, H., Torigoe, H., Saito, A., Sato, M., Arata, Y., Shimada, I. (1992) Biochemistry 31,
9665-9672.

‘2 Sjodahl, J. (1977) Eur. J. Biochem. 73, 343-351. and Sjodahl, J. (1977) Eur. J. Biochem.
78, 471-490.

9 Ghetie, V., Nilsson, K., Sjoquist, J. (1974) Eur. J. Immunol. 4, 500-505.

72

 

“ Protein DataBank, 1FC2, Deisenhofer 1981.
‘5 Bj¢rk, I, Petersson, B.-A., Sj¢quist, J. (1972) Eur. J. Biochem. 29, 579-584.

“ geber, P. C., Ohlendorf, D. H., Wendoloski, J. J., Salemme, F. R. (1989) Science 243, 85-

" Green, N. M. (1975) Adv. Protein Chem. 29, 85-133.
6' Protein DataBank, 1PTS, Dupont Protein Crystallography Group

9 Hendrickson, W. A.; Pahler, A.; Smith, J. L.; Satow, Y.; Merritt, E. A.; Phizackerley, R. P.
(1989) Proc. Natl. Acad. Sci. 86 2190-2194.

7° Ghosh, P., Meyer, C., Remey, E., Peterson, D., & Preiss, J. (1992) Archives of
Biochemistry and Biophysics 296, 122.

7‘ Stark, D. M, Timmerman, K. T., Barry, G. F., Preiss, J., & Kishore, G. M. (1992) Science
258, 287-292.

7’ Carlson, C. A., Parsons, T. F., & Preiss, J. (1976) The Journal of Biological Chemistry
251, 7886.

"-’ Iglesius, A. A., Kakefuda, G., Preiss, J. (1991) Plant Physiol. 97, 1187-1195.

7‘ Chamg, Y.-y., Kakefuda, G., Iglesius, A. A., Buikema, W. J., & Preiss, J. (1992)
Plant Molecular Biology 20, 37.

7’ Fanger, B. 0., Stephens, J. 13., Staros, J. V. (1989) FASEB J. 3(1), 71-5.
7‘ Krais, S., Quaiser, A., Oren, M., Montenarh, M. (1988) J. Viral. 62(12), 4737-44.

7" Ghosaini, L. R., Brown, A. M., Sturtevant, J. M. (1988) Biochemistry 27( 14), 5257-
61.

7‘ Clore, G. M, Gronenborn, A. M. W, CRC Press, Boca Raton 1993.
7’ Revet, B., Brahms, S., Brahms, G. (1995) Proc. Natl. Acad. Sci. USA 92(16), 7535-9.

°° Smith, J. M. A., Ford, G. C., Harrison, P. M., Yariv, J., Kalb, A. J. (1989) J. Mol.
Biol. 205, 465-467.

3‘ Radmacher, M., Fritz, M., Hansma, H. G., & Hansma, P. K. (1994b) Science 265,
1577.

'3 Takahashi, M. (1989) J. Biol. Chem. 264(1), 288-95.
'3 Davey, M. J., Funnel], B. E. (1994) J. Biol Chem. 269 29908-29913.

“ Ghosaini, L. R., Brown, A. M., Sturtevant, J, M. ( 1988) Biochemistry 27(14), 5257-
61.

” Burger, M. M., Goldberg, A. R. (1967) Proc. Natl. Acad. Sci. USA 57, 359-366.

72

 

6‘ Protein DataBank, 1FC2, Deisenhofer 1981.
‘5 Bj¢rk, I, Petersson, B.-A., Sj¢quist, J. (1972) Eur. J. Biochem. 29, 579-584.

“ geber, P. C., Ohlendorf, D. H., Wendolosld, J. J., Salemme,’ F. R. (1989) Science 243, 85-

" Green, N. M. (1975) Adv. Protein Chem. 29, 85-133.
‘3 Protein DataBank, lPTS, Dupont Protein Crystallography Group

‘9 Hendrickson, W. A.; Pahler, A.; Smith, J. L.; Satow, Y.; Merritt. E. A.; Phizackerley, R. P.
(1989) Proc. Natl. Acad. Sci. 86 2190-2194.

7° Ghosh, P., Meyer, C., Remey, E., Peterson, D., & Preiss, J. (1992) Archives of
Biochemistry and Biophysics 296, 122.

7‘ Stark, D. M, Timmernmn, K. T., Barry, G. F., Preiss, J., & Kishore, G. M. (1992) Science
258, 287-292.

’2 Carlson, C. A., Parsons, T. F., & Preiss, J. (1976) TheJournal of Biological Chemistry
251, 7886.

7’ Iglesius, A. A., Kakefuda, G., Preiss, J. (1991) Plant Physiol. 97, 1187-1195.

7‘ Chamg, Y.-y., Kakefuda, G., Iglesius, A. A., Buikema, W. J., & Preiss, J. (1992)
Plant Molecular Biology 20, 37.

7‘ Fanger, B. 0., Stephens, J. E., Staros, J. V. (1989) FASEB J. 3(1), 71-5.
7‘ Krais, S., Quaiser, A., Oren, M., Montenarh, M. (1988) J. Virol. 6202), 4737-44.

7’ Ghosaini, L. R., Brown, A. M., Sturtevant, J. M. (1988) Biochemistry 27( 14), 5257-
61.

7‘ Clore, G. M, Gronenbom, A. M. W, CRC Press, Boca Raton 1993.
7’ Revet, B., Brahms, S., Brahms, G. (1995) Proc. Natl. Acad. Sci. USA 92(16), 7535-9.

'° Smith, J. M. A., Ford, G. C., Harrison, P. M., Yariv, J., Kalb, A. J. (1989) J. Mol.
Biol. 205, 465-467.

3‘ Radmacher, M., Fritz, M., Hansma, H. G., & Hansma, P. K. (1994b) Science 265,
1577.

'2 Takahashi, M. (1989) J. Biol. Chem. 264(1), 288-95.
” Davey, M. J., Funnel], B. E. (1994) J. Biol Chem. 269 29908-29913.

3‘ Ghosaini, L. R., Brown, A. M., Stunevant, J, M. (1988) Biochemistry 27(14), 5257-
61.

3’ Burger, M. M., Goldberg, A. R. (1967) Proc. Natl. Acad. Sci. USA 57, 359-366.

73

 

8‘ Privat, J. P., Charleir, M. (1978) Eur. J. Biochem. 84, 79-85.

'7 Gurd, J. W. (1977) Biochemistry 16, 369-374.

'3 Nicholson, G. L. (1974) Int. Rev. Cytol. 39, 89-190.

'9 Pinto de Silva, P., Torrisi, M. R., Kachar, B. (1981) J. Cell. Biol. 91, 361-372.

9° Jordan, F., Bassetl, 13., Redwood, W. R. (1977) Biochem. Biophys. Res. Comm. 75 1015-
1021.

’1 Weber, P; Ohlendorf, D. H.; Wendoloski, J. J.; Salemme, F. R (1989) Science 243, 85-88.

’2 Pahler, A.; Hendrickson, W. A.; Kolks, M. A. G.; Argaraiia, C. E.; Cantor, C. R.
(1987) The Journal OfBiological Chemistry 262, 13933-13937.

9" Sano, T., Pansdori, M. W., Smith, C. L., Cantor, C. R. (1994) Advances in Biomagnetic
Separation, eds. Uhlén, M., Homes, E., Olsvik, 0. (ann, Natick, MA)pp. 21-29.

9‘ Sano, T.; Cantor, C. R. (1995) Proc. Natl. Acad. Sci. 92 3180-3184.

’5 Chinami, M., Sasaki, S., Hachiya, N., Yuge, K., Ohsugi, T., Maeda, H., Shingu, M. (1994) J.
Gen. Virol. 75, 227-281.

9‘ LeMaire et. al. (1989) Anal. Biochem. 177, 50-56.

’7 Minton, A. P. (1989) Anal. Biochem. 176, 209-216.

9‘ Shapiro, A. L. (1967) Biochem. Biophys. Res. Comm. 28, 815-820.

9’ Hedrick, J. 1... Smith, A. J. (1968) Arch. Biochem. Biophys. 126, 155-164.

APPENDICES

APPENDIX A
SOLID FILMS OF PROTEIN

Introduction If the concentration of the protein solution is too high to allow the
molecules to isolate themselves on the mica surface (generally greater than 0.05 'mg/ml), a
solid film of protein is formed. For streptavidin and ADngp, this film has characteristics
I of . a two dimensional crystal, with uniform vertical dimensions, while in other cases the
film is rather amorphous. In both situations, some structural information can still be
gleaned from analyzing these images. In the case of an amorphous film that completely
covers the mica, it is necessary to scrape away a portion of the film to reveal the flat

background of the mice in order to be sure of measuring the correct thickness of the film.

S trepta vidin. Streptavidin has previously been known to form 2 dimensional
crystals by spreading on lipid layers,1 and by spreading on a film of poly(l-benzyl-L-
histadine).2 The crystals formed by spreading on the poly(l-benzyl-L-histadine) revealed
ordered 2-D arrays of Streptavidin molecules with 44 A spacing. Since the molecular
dimensions are approximately 42 x 42 x 56 A3 this lattice spacing would indicate that the
molecules are oriented with their long sides together. Therefore, the thickness of the layer
would be the larger dimension of the molecule, 56A. In six different SFM experiments
using protein concentrations of 05 to 5 mg/ml, Streptavidin forms islands or a lattice type
formation on the mice after deposition of the protein and drying the sample. This type of

formation was seen in all experiments when the protein concentration was greater than 0.05

 

‘ Ku, A. C.; Darst. s. A.; Robertson, c. R.; Gast, A. P. Kornberg, R. D. (1993) J. Phys. Chem. 97,
3013-3016.

2 Fumno, T.; Sasabe, H. (1993) Biophysical Journal 65, 1714-1717.

3 Weber, P. C.; Ohlendorf, D. H.; Wendoloski, J. J.; Salemme, F. R. (1989) Science 243, 85.

74

75

mg/ml. These formations always have a very uniform thickness of about 60 A throughout
the entire formation. This indicates a type of 2-D molecular ordering similar to that found
in the 2-D crystals of Streptavidin. See figure 31 for a typical two-dimensional array of
Streptavidin.

ADngp. When ADngp wild type formed a layer on the mica, it was in the form
of a uniform lattice or islands 100 A tall. The concentration necessary to form a layer
instead of isolated particles is about 0.08 mglrnl. In four of five experiments in which a
solid coating was formed this uniform layer was observed. The remaining sample
produced a layer solidly covering the mica, for which the height could not be determined
because the flat mica surface was not visible by either gaps in the layer or scraping a
portion of the layer with the tip. Since the crystal structure for this protein is unknown, the
interpretation of this data is somewhat uncertain, however, the molecular weight of the
intact tetrameric enzyme, 200kDa, would be amenable to the interpretation that the molecule
was not terribly elongated and one dimension of the molecule was 90 A. The single
molecule SFM data then can be used to predict the other two dimensions of the molecule to
be 50 A and 70 A. The uniformity of the formations within and between different
experiments suggests that these are a formation with at least short-range ordering of the
molecules. The one sample of the ADngp enzyme from Anabaena in which a height could
be determined also showed uniform flat islands 100 A tall, indicating a similar structure for
this analogous enzyme. See figme 32 for a typical image of a two-dimensional array of
ADngp wild type molecules.

76

Protein A. Protein A, as mentioned in chapter 2, is an elongated molecule about
18 A in diameter and 120 A long. When the concentration of protein is too high to form
individual particles on the mica, Protein A will form a layer, and in the six different
experiments in which a layer was seen, it was always less than 30 A tall. A concentration
greater than 0.05 mg/ml will form this layer instead of isolated particles on the surface.
The layer was not smooth or uniform in height, indicating the absence of long-range order
to the molecular packing, but the molecules apparently lay flat on the mice and form one or

two layers of protein under the conditions of these experiments.

Con A. When Con A forms a solid layer on the mica, the layer is uneven, and it
is 40 to 75 A tall. This would indicate that these formations are composed of Con A dimers
with some tetramer form also possibly present. The thicker layer could also be the result of

two layers of dimers.

77

 

Figure 31: Two-dimensional array of Streptavidin.

 

Figure 32: Two-dimensional array of ADng wild type.

APPENDIX B

VISUAL BASIC PROGRAM TO AUTOMATICALLY MEASURE AND
TABULATE THE HEIGHT OF MOLECULES IN AN AFM IMAGE

Sub DifferentZO
ApplicafionScreenUpdating = False
Dim B

DimCo

DimCV

DimACV
DimBCV
DimCCV
DimDCV
DimECV
DimFCV
DimGCV

B = 0

Co = 0
SHEEl‘sC'Sheetl ").Select
Range("F5").Select
SHEET s("Sheet2").Select
Range("B1").Select
ActiveCell.FormulaR1C1 = "value"
Range("C1").Select
ActiveCell.FormulaR1Cl = "background"
Range("Dl").Select
ActiveCell.FormulaR1Cl = "height"
Range("D2").Select
ActiveCell.FormulaR1C1 = "=RC[-2]-RC[-ll"

78

79

Range("D2").Select
Selection.Copy
Range("D3:D500").Select
ActiveSheetPaste
Application.CutCopyMode = False
Range("B2").Select
SHEETs("Sheet1").Select
For Lines = 1 To 244 Step 1
Do Until ActiveCell = ""

CV = Selection.Value

If ActiveCell < 50 Then

B = CV + B

Co = C0 + 1

End If

If ActiveCell < 50 Then GoTo Line2

ActiveCell.Offset(-l, -1).Range("A1").Select
ACV = Selection.Value

ActiveCell.Offset(0, 1).Range( "A 1 ").Select
BCV = Selection.Value

ActiveCell.Offset(0, 1).Range("A1").Select
CCV = Selection.Value

ActiveCell.Offset(1, -2).Range("A1").Select
DCV = Selection.Value

ActiveCell.Offset(0, 2).Range("A1").Select
BCV = Selection.Value

ActiveCell.Offset( 1, -2).Range("A1").Select
FCV = Selection.Value

ActiveCell.Offset(0, 1).Range("A1").Select
GCV = Selection.Value

ActiveCell.Offset(O, 1).Range("A1 ").Select
HCV = Selection.Value

ActiveCell.Offset(-3, -3).Range("Al ").Select
ICV = Selection.Value

ActiveCell.Offset(0, 1).Range("Al ").Select
JCV = Selection.Value

ActiveCell.Offset(0, 1).Range("A1 ").Select
KCV = Selection.Value

ActiveCell.Offset(0, l).Range("Al ").Select

80

LCV = Selection.Value

ActiveCell.Offset(O, 1).Range( "A 1 ").Select
MCV = Selection.Value

ActiveCell.Offset( l , -4).Range( "Al ").Select
NCV = Selection.Value

ActiveCell.Offset(O, 4).Range("A1 ").Select
OCV = Selection.Value

ActiveCell.Offset(l, -4).Range("Al ").Select
PCV = Selection.Value

ActiveCell.Offset(O, 4).Range("A1 ").Select
QCV = Selection.Value

ActiveCell.Offset( 1 , -4).Range("A1 ").Select
RCV = Selection.Value

ActiveCell.Offset(O, 4).Range("A1").Select
SCV = Selection.Value

ActiveCell.Offset( 1 , -4).Range("Al ").Select
TCV = Selection.Value

ActiveCell.Offset(O, 1).Range( "A1 ").Select
UCV = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
VCV = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
WCV = Selection.Value

ActiveCell.Offset(O, 1).Range( "Al ").Select
XCV = Selection.Value

ActiveCell.Offset(-2, -2).Range("A1").Select

If CV < ACV Then GoTo Line2
If CV < BCV Then GoTo Line2
If CV < CCV Then GoTo Line2
If CV < DCV Then GoTo Line2
If CV < ECV Then GoTo Line2
If CV < FCV Then GoTo Line2
If CV < GCV Then GoTo Line2
If CV < HCV Then GoTo Line2
If CV < ICV Then GoTo Line2

If CV < JCV Then GoTo Line2

81

If CV < KCV Then GoTo Line2
If CV < LCV Then GoTo Line2
If CV < MCV Then GoTo Line2
If CV < NCV Then GoTo Line2
If CV < OCV Then GoTo Line2
If CV < PCV Then GoTo Line2
If CV < QCV Then GoTo Line2
If CV < RCV Then GoTo Line2
If CV < SCV Then GoTo Line2
If CV < TCV Then GoTo Line2
If CV < UCV Then GoTo Line2
If CV < VCV Then GoTo Line2
If CV < WCV Then GoTo Line2
If CV < XCV Then GoTo Line2
Linel:
Selection.Font.ColorIndex = 3
If ActiveCell > 300 Then Selection.Font.ColorIndex = 4
ActiveCell.Copy
'Calculates the average background

Dim B32
Dim B33
Dim B34
Dim B35
Dim B36
Dim B37
Dim BB8
Dim B39
Dim B40
Dim B41
Dim BKl
Dim BK2
Dim BK3
Dim BK4

Dim BK

82

ActiveCell.Offset(-5, -5).Range("Al ").Select

B1 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
BZ = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B3 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B4 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
BS = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B6 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B7 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B8 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
B9 = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
BIO = Selection.Value

ActiveCell.Offset(O, 1).Range("A1").Select
Bll = Selection.Value

ActiveCell.Offset( 1 , 0).Range("A 1 ").Select
B12 = Selection.Value

ActiveCell.Offset( l , 0).Range("A1").Select
813 = Selection.Value

ActiveCell.Offset( l , 0).Range("A1 ").Select
B14 = Selection.Value

ActiveCell.Offset( 1 , 0).Range("A1").Select
B15 = Selection.Value

ActiveCell.Offset( l , 0).Range("A1").Select
816 = Selection.Value

ActiveCell.Offset(l, 0).Range("Al").Select
B17 = Selection.Value

83

ActiveCell.Offset(l, 0).Range("A1").Select
B18 = Selection.Value

ActiveCell.Offset(l , O).Range("A1").Select
B19 = Selection.Value

ActiveCell.Offset( 1 , 0).Range("A1").Select
820 = Selection.Value

ActiveCell.Offset(l , 0).Range("A1").Select
321 = Selection.Value
ActiveCell.Offset(-1, 0).Range("Al ").Select
822 = Selection.Value

ActiveCell.Offset(-1, 0).Range("A1").Select
323 = Selection.Value
ActiveCell.Offset(-l, 0).Range("A1").Select
824 = Selection.Value

ActiveCell.Offset(-1, 0).Range("A1").Select
B25 = Selection.Value

ActiveCell.Offset(- 1 , 0).Range("Al ").Select
B26 = Selection.Value

ActiveCell.Offset(-1, O).Range("A1").Select
827 = Selection.Value

ActiveCell.Offset(-1, O).Range("A1").Select
B28 = Selection.Value

ActiveCell.Offset(-l , O).Range("A1").Select
829 = Selection.Value

ActiveCell.Offset(-1, 0).Range("A1").Select
B30 = Selection.Value

ActiveCell.Offset(- l , 0).Range("A 1 ").Select
B31 = Selection.Value

ActiveCell.Offset(O, -l).Range("A1").Select
B32 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1").Select
B33 = Selection.Value

ActiveCell.Offset(O, -l).Range("A1").Select
B34 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1").Select
B35 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1 ").Select
B36 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1").Select
B37 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1 ").Select
B38 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1 ").Select
, B39 = Selection.Value

ActiveCell.Offset(O, -l).Range("A1 ").Select
B40 = Selection.Value

ActiveCell.Offset(O, -1).Range("A1 ").Select
B41 = Selection.Value

ActiveCell.Offset(6, 5).Range("A1 ").Select

821)
831)

8K1=(81+82+83+B4+85+B6+B7+88+89+810)
8K2=(Bll+812+813+814+815+816+817+818+819+820+

BK3=(BZI+822+823+824+825+826+827+828+B29+B30+

8K4: (832 + 833 + 834-1- 835 + 836 + 837 + 838 + 839 + 840 + 841)
BK=(8K1+8K2+8K3 +8K4)/41

'Excludes background points if < 80% or > 120% of average
If81>1.2 * BK Then 81: BK
If81<0.8 * BK Then 81: BK
IfB2>1.2 * BK Then 82 = BK
If82 <0.8 * BK Then 82 = BK
IfB3 >1.2 * BK Then B3 = BK
IfB3 <0.8 * BK Then B3 = BK
IfB4> 1.2* BKThenB4=BK
IfB4<0.8*8KThen84=8K
IfBS >1.2 * BK Then BS = BK
IfBS <0.8 * BK Then 85 = BK
lfB6>1.2 * BK Then B6 = BK
IfB6<0.8 * BK Then B6 = BK
If87> 1.2 * BK Then B7 = BK
IfB7 < 0.8 * BK Then B7 = BK
IfBS>1.2 * BK Then B8 = BK
If88 <0.8 * BK Then B8 = BK
IfB9> 1.2 * BKThen 811=8K
If811<0.8 * BK Then 811 = BK
If812>1.2 * BK Then 812 = BK
If812<0.8 * BK Then 812=8K
If813 >1.2 * BK Then 813 = BK
IfBl3 < 0.8 * BK Then 813 = BK
IfBl4>1.2 * BK Then 814 = BK
lfBl4<0.8 * BK Then Bl4= BK
IfBlS >1.2 * BK Then BIS = BK
IfBIS < 0.8 * BK Then BIS = BK
IfBl6> 1.2 * BK Then 816 = BK
lfBl6<0.8 * BK Then 816=8K
lfBl7 >1.2 * BK Then 817 2: BK
1f817 <0.8 * BK Then 817 = BK
If 818 >1.2 * BK Then 818 = BK
IfBlS < 0.8 * BK Then 818 = BK
IfBl9>1.2 * BK Then 819 = BK
IfBl9<0.8 * BK Then 819=8K
IfBZO > 1.2 * BK Then B20 = BK
If821>1.2 * BK Then B21 = BK
IfBZl <0.8 * BK Then 821 = BK
If 822 > 1.2 * BK Then 822 2 8K
If 822 < 0.8 * 8K Then 822 = 8K
IfB23 > 1.2 * 8K Then 823 = 8K
If 823 < 0.8 * 8K Then 823 = 8K

B31)

85

If 824 > 1.2 * 8K Then 824 = 8K
If 824 < 0.8 * 8K Then 824 2 8K
If 825 > 1.2 * 8K Then 825 = 8K
If 825 < 0.8 * 8K Then 825 2 8K
If 826 > 1.2 * 8K Then 826 = 8K
If 826 < 0.8 * 8K Then 826 = 8K
If 827 > 1.2 * 8K Then 827 = 8K
If 827 < 0.8 * 8K Then 827 = 8K
If 828 > 1.2 * 8K Then 828 = 8K
If 828 < 0.8 * 8K Then 828 2 8K
If 829 > 1.2 * 8K Then 829 = 8K
If 829 < 0.8 * 8K Then 829 = 8K
If 830 > 1.2 * 8K Then 830 = 8K
If 830 < 0.8 * 8K Then 830 2 8K
If831>1.2 * 8K Then B31: 8K
If831<0.8 * 8K Then 831: 8K
If 832 > 1.2 * 8K Then 832 = 8K
If 832 < 0.8 * 8K Then 832 2 8K
If833 > 1.2 * 8K Then 833 2 8K
If 833 < 0.8 * 8K Then 833 2 8K
If834>1.2 * BK Then 834 = 8K
If 834< 0.8 * 8K Then 834 = 8K
If 835 > 1.2 * 8K Then 835 = 8K
If 835 < 0.8 * 8K Then 835 2 8K
If 836 > 1.2 * 8K Then 836 = 8K
If 836 < 0.8 * 8K Then 836 = 8K
If 837 > 1.2 * 8K Then 837 = 8K
If 837 < 0.8 * 8K Then 837 = 8K
If 838 > 1.2 * 8K Then 838 = 8K
If 838 < 0.8 * BK Then 838 = 8K
If 839 > 1.2 * 8K Then 839 2 8K
If 839 < 0.8 * 8K Then 839 = 8K
IfB40>1.2 * BKThen 840: 8K
IfB40<0.8 * 8KThen840=8K
If 841 > 1.2 * 8K Then 841 = 8K
If 841 < 0.8 * 8K Then 841 2 8K

8K1=(81+82+83+84+85+86+87+88+89+810)
BK2=(811+812+813+814+815+816+817+818+819+820)
8K3=(821+822+823+824+825+826+827+828+829+830+

8K4 = (832 + 833 + 834 + 835 + 836 + 837 + 838 + 839 + 840 + 841)
8K = (8K1 + 8K2 + 8K3 + 8K4) ./ 41

SHEET s( "Sheet2").Select

ActiveSheet.Paste

ActiveCell.Offset(O, 1).Range( "A 1 ").Select

Selection.Value = 8K

ActiveCell.Offset(l , -I).Range(”A1 ").Select
SHEET s( "Sheetl ").Select

86

Line2:
ActiveCell.Offset(l, 0).Range("A1").Select
Loop .
ActiveCell.Offset(-250, 1).Range( "A1 ").Select
Next Lines
Application.ScreenUpdating = True
SHEETs("Sheet3").Select
ActiveCell.Offset(Z, O).Range("A1").Select
Selection.Value = C0
ActiveCell.Offset(Z, O).Range("A1").Select
Selection.Value = 8
End Sub

Sub Color()
Do Until ActiveCell = '"‘
Do Until ActiveCell = ""
Application.ScreenUpdating = False
If Selection.Font.ColorIndex = 2 Then Selection.FontColorIndex = 3
ActiveCell.Offset( 1, 0).Range("A1 ").Select
Loop
ActiveCell.Offset(-250, 1).Range("A1").Select
Loop
Application.ScreenUpdating = True
End Sub

87

APPENDIX C

VISUAL BASIC ROUTINE TO AUTOMATICALLY OPEN SEVERAL
TEXT FILES AND PROCESS THE DATA THEN CLOSE THE FILES.

Sub series()

Workbooks.0pen Filename:="D:\EXCEL\MARTHA\cona2.XLS"
SHEET s.Add

SHEETs.Add

Application.Run Macro:="PICKS.XLS ldifferent2"
ActiveWorkbook Save

AcfiveWorkbookClose

Workbooks.0pen Filename:="D:\EXCEL\MARTHA\cona3.XLS"
SHEET s.Add

SHEET s.Add

Application.Run Macro:="PICKS.XLS ldifferent2"
ActiveWorkbook. Save

AcfiveWorkbookClose

Workbooks.0pen Filename:="D:\EXCEL\1VIARTHA\cona4.XLS"
SHEETs.Add

SHEETs.Add

Application.Run Macro:="PICKS.XLS ldifferentZ"
ActiveWorkbook. Save

ActiveWorkbookClose

Workbooks.0pen Frlename:="D:\EXCEL\MARTHA\cona5.XLS"
SHEETs.Add

SHEET s.Add

Application.Run Macro:="PICKS.XLS ldifferent2"
ActiveWorkbook Save

AcfiveWorkbookClose

End Sub

88

APPENDIX D

C-H- PROGRAM TO CONVERT PARK SCIENTIFIC INSTRUMENTS
BINARY IMAGE TO X, Y TEXT ARRAY SUITABLE FOR READING
INTO EXCEL.

”primitive binary to text file converter for Park scientific
#include <stdio.h>
#include <stdlib.h>

#define FNAME "1220003a"
int main(void)

FILE *fp,*fpl;
int word;

/* place the word in a file */
fp = fopen(FNAME, "rb");
fpl = fopen("testtxt","wb");
if (fp == NULL)

{

printf("Error opening file %s\n", FNAME);
exit(1);

}
if (fpl == NULL)

printf("Error opening file %s\n", "testtxt");
exit(1);

for (int i=1;i<=256;i-H~) {
for (int j=l ;j<=256;i++) {

word=getw(fp);

if (ferr0f(fP)){
printf "Error writing to file\n");
exit(1);

}.

else
fprintf(fp1,"%d",word);
if(i=256) fprintf(fpl,"\n");
else fprintf(fp1,"\t");

}
// fprintf(fpl,"\n");

fclose(fp);

fclose(fp1);

return 0;

89

APPENDIX E

MATHEMATICA ROUTINE FOR READING A TEXT FILE OF
TABULATED AF M DATA AND FITTING THE DATA TO APPROPRIATE
GAUSSIAN CURVES.

(*Reads a text file with the data and fits the data to a specified gaussian curve. The
coefficient to each exponential is the percentage of molecules at a particular height
according to SOM theory. If the gaussians have different widths, it is necessary to change
the coefficient to nuke the area under the curve correspond to the expected distribution of
protein sizes according to SOM theory.*)

111-54 Bxp[-((x-28)“2)/50] + 33 Bxp[-((x-36)‘2)/50] + 13
Exp[—((x-57)‘2)/50]

data=ReadList["ConAteth",Number, RecordLists->True];
Piudata, {it}, {x}]

92=Plot[%, (x, 0, 100}, PlotRange->{O, 20].];
gl-ListPlotmata, PlotRange->{{o, 100}, {0, 20}}];
Sthslr 92]

‘ damn-r .. '

 

 

 

"llllllllllll’llll