Purification and functional characterization of human polymorphonuclear leukocyte actin polymerization inhibitor


Material Information

Purification and functional characterization of human polymorphonuclear leukocyte actin polymerization inhibitor
Physical Description:
ix, 160 leaves : ill. ; 29 cm.
Maun, Noel Anthony, 1968-
Publication Date:


Subjects / Keywords:
Actins -- isolation & purification   ( mesh )
Actins -- biosynthesis   ( mesh )
Microfilaments -- physiology   ( mesh )
Microfilaments -- metabolism   ( mesh )
Microfilaments -- ultrastructure   ( mesh )
Cytoskeleton -- physiology   ( mesh )
Cytoskeleton -- metabolism   ( mesh )
Cytoskeleton -- ultrastructure   ( mesh )
Neutrophils   ( mesh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1995.
Includes bibliographical references (leaves 144-159).
Statement of Responsibility:
by Noel Anthony Maun.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 50281502
System ID:

This item is only available as the following downloads:

Full Text







This dissertation is dedicated to all the teachers who have guided
my education at the University of Florida.


I would like to thank the members of my graduate committee for
their assistance and willingness to be a part of my education. Drs. Purich,

Baker, Driscoll, and Nick's constant support throughout my graduate work
enabled me to carry the project forward. I am especially indebted to my
mentor, Dr. Frederick Southwick, for his continued encouragement and
constant scientific guidance. His traits of perpetual enthusiasm, sincerity,

willingness to learn and dedication to basic science and clinical medicine
have provided a model I hope to equal someday. I would also like to

thank Dr. Purich for always being available when advice was needed. Our
discussions were always enjoyable, and I hope someday to be able to
explain complicated topics as simply and elegantly as he always does.

Special thanks also goes out to Dr. Mary Jo Koroly for a wonderful

introduction to the basic sciences. My tenure in the lab would not have

been as enjoyable if it were not for the many friends that I have made
during graduate school.

Finally, I would like to thank my family for their love, patience, and

support throughout my education.


ACKNOWLEDGMENTS ........................................... iii

LIST OF FIGURES ............................................... vi

ABSTRACT.................................................. viii


1 INTRODUCTION ...................................... 1

Overview ............................................. 1
The Human Neutrophil ................................. 2
Actin ................................................ 4
Actin Binding Proteins ................... ..............14
Actin Dynamics in PMN ................................ 21
Calcium Independent Model For Polymerization ........... 31
PMN Actin Polymerization Inhibitor ...................... 35

2 MATERIALS AND METHODS ............................. 37

Isolation Of Human Polymorphonuclear Leukocytes
(PM N ).................... ... .................... 37
Purification Of CapZ From Human PMN ................... 37
Purification of Muscle Actin ............................ 40
Amino Acid Sequence Analysis .......................... 40
Polyacrylamide Gel Electrophoresis And Western Blot
Analysis .......................................... 41
Actin Binding Studies ................................. 43
Lipid-Binding And Capping-Inhibition Studies ............. 45
Promyelocyte Differentiation Studies ..................... 47
Fluorescence Microscopy .............................. 50


Introduction ......................................... 52
R results ................... ........................... 52
D discussion ........................................... 75


Introduction............... ... ....... ................ 88
Results .............................................. 89
Discussion..................... ..................... 103

LOCALIZATION OF ANNEXIN VI IN PMN .................. 113

Introduction ........................................ 113
Results ................... ... ....................... 115
Discussion.......................................... 131


Conclusions ......................................... 136
Future Directions .................................... 139

REFERENCES ................................................. 144

BIOGRAPHICAL SKETCH ....................................... 160


Figure page

1-1 Actin and its spontaneous polymerization in salt ............ 8

1-2 Rate constants for the association and dissociation
of ATP-actin and ADP-actin at filament ends .............. 11

1-3 fMet-Leu-Phe induced phospholipase C signaling
pathway ........................................... 26

1-4 Regulation of actin assembly in PMN ...................... 34

3-1 PMN actin polymerization inhibitor ....................... 54

3-2 Amino-terminal sequence analysis of PMN actin
polymerization inhibitor ...............................57

3-3 Nondenaturing polyacrylamide gel electrophoresis .......... 61

3-4 SDS-PAGE of phospholipid affinity chromatography,
and anti-annexin VI western analysis ................... .64

3-5 Western blot analysis of PMN inhibitory fractions
probed with capZ antisera ............................. 67

3-6 DEAE-anion exchange chromatography of PMN extract ........ 70

3-7 High S-cation exchange chromatography of PMN
actin polymerization inhibitor ......................... 72

3-8 Mono Q-anion exchange chromatography of PMN
actin polymerization inhibitor ......................... 74

3-9 Western blot analysis of PMN actin polymerization
inhibitor purified to Mono Q chromatography ............ 77

3-10 Hydroxylapatite column chromatography ................. 79

3-11 Silver stained peak fractions from Mono Q
and HA chromatography .............................. 81

4-1 Effects of purified capZ on actin filament
depolymerization ......................................91

4-2 Effects of purified capZ on actin filament polymerization
from spectrin/band 4.1/actin nuclei ..................... 94

4-3 Effects of neutrophil capZ on the extent of
actin polymerization .................................. 97

4-4 Effects of neutrophil capZ on G-actin nucleation ...........100

4-5 Actin filament severing assay ........................... 102

4-6 Effects of capZ on polymerization from
gelsolin:actin nuclei ................................ 105

4-7 Effects of PIP2 on capZ barbed-end
capping activity ............... .................. .. 107

5-1 Northern analysis of PMN and U937 ...................... 117

5-2 Northern analysis of HL-60 differentiated to
neutrophil-like or macrophage-like cells ................. 120

5-3 Western analysis of HL-60 differentiated to
neutrophil-like or macrophage-like cells ................. 123

5-4 Indirect immunofluorescence microscopy of PMN. 126

5-5 Confocal microscopy of PMN stained with
anti-annexin VI antibodies ............................ 128

5-6 Confocal microscopy of peripheral blood
monocytes stained with anti-annexin VI
antibodies ......................................... 130

5-7 Quantitation of annexin VI in human
neutrophil cytoplasmic extracts ....................... 133

Abstract of Dissertation Presented to the Graduate School of the University
of Florida in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy




May, 1995

Chairperson: Daniel L Purich, Ph.D.
Cochairperson: Frederick S. Southwick, M.D.
Major Department: Biochemistry and Molecular Biology
Actin polymerization inhibitor, an activity from human
polymorphonuclear leukocytes (PMN) which lowers the viscosity of actin
filament solutions, was previously purified by our laboratory. The calcium

independent nature of the activity suggests it may be involved in the
motile behavior of PMN based on the current theories regarding the

regulation of actin filament assembly. This activity has been found by
additional column chromatography steps and Western blot analysis to be

the heterodimeric capping protein capZ. The actin regulatory activity of
this non-muscle capZ was assessed utilizing pyrenylactin. Similar to
skeletal muscle capZ and other members of the capping protein family, I
report that non-muscle capZ purified from PMN blocks monomer exchange
at the barbed ends of actin filaments under both polymerizing and
depolymerizing conditions with an apparent disassociation constant for

capping of 3 nM. Similar to the activity reported for actin polymerization

inhibitor, PMN capZ's capping activity is independent of Ca2+ and is
inhibited by increasing the KCI concentration from 0.1 M to 0.6 M. As
observed in all previously studied barbed-end capping proteins, PMN
capZ's capping function is inhibited by phosphatidylinositol 4,5-
bisphosphate (PIP2) micelles, 1/2 maximal inhibition being observed at

5.5 pg ml-1. Phosphatidylcholine, phosphatidylserine, or
phosphatidylinositol (11 pg ml-1) fail to inhibit capping function. The
PMN capZ's effects on actin assembly are confined to an interaction with
the barbed ends of actin filaments. This protein does not cap actin
filament pointed ends, does not sever preformed filaments, and fails to
interact significantly with actin monomers. Kinetic studies reveal no
enhancement of the nucleation step in actin assembly. This first report of
a capZ-related protein isolated from mammalian phagocytes suggests that
capZ is likely to play an important role in the regulation of actin filament
assembly in phagocytic cells.



Cellular motility is a complex process which mandates the ability of
a cell to change its shape. In response to an external stimulus, outer
membrane receptors receive motivational signals and transmit this
information intracellularly. Subsequently, a cascade of events is initiated
which allow the cell to change its morphology and generate force for
motion towards the extracellular signal. A likely candidate which appears

central to this behavior is the self-assembling protein actin. Actin is the

most highly conserved and abundant protein in eukaryotic cells; and, in
addition, is the predominant filament of the cytoskeletal network. Unlike
in the sarcomere of muscle cells, non-muscle actin filaments are randomly
arranged in a fashion which enables the generation of force in three

dimensions rather than two dimensions. It is becoming increasingly

apparent that actin filament assembly/disassembly is vital to the process

of cell motility and shape change, and that the randomness of this
assembly is prevented by a group of regulatory proteins known
collectively as actinn binding proteins." This is a very exciting time for the
field of cellular motility and cell shape change. Amoebae, tumor cells,
leukocytes, fibroblasts, epithelial cells, and many other different cell types

share the ability to move by crawling across solid substrates utilizing a

similar process (Stossel, 1993). Our deeper understanding of the
molecular events involved may lead to a better understanding of the
biological processes which rely on cellular motility such as embryonic
development, inflammation, wound healing, tumor invasion, and host
defense. The work entailed in this dissertation centers around an actin
binding activity purified from human polymorphonuclear leukocytes

(PMN), one of the most highly motile animal cell types (10 20 pm/min)
(Howard & Meyer, 1984).

The Human Neutrophil

The neutrophil is one of the most motile cell types in the human
body; its primary purpose is to engulf and kill invading pathogens.
Histologically, polymorphonuclear granulocytes can be subdivided into
three distinct cell types named according to their staining characteristics;
the neutrophil (> 90 %), basophil, and eosinophil. Despite originating
from the same progenitor stem cell, each has a functionally distinct
immunological role (Williams et al., 1990). For the purposes of this
dissertation, the terms neutrophil, polymorphonuclear leukocyte (PMN),
and granulocyte will be used synonymously to denote neutrophil
leukocytes. In the normal human adult, the life span of the PMN is spent
in three environments: marrow, blood, and tissues (Bainton, 1992).
Proliferation and terminal differentiation occur primarily in the marrow
compartment over a period of approximately 10 days (Gallin, 1988). They
are subsequently released into the blood. This intravascular pool of
granulocytes representing only 5 % of the total number in the body is
equally divided between a freely circulating pool and a marginated pool

that is adherent to, or closely associated with, the lining of the blood
vessels (Golde, 1990). Individual cells move back and forth behaving
kinetically as a single pool. The neutrophils' stay in the blood is short,
with an approximate half-life of 6 7 hours, from where they subsequently
enter the extravascular space (Dancey et al., 1976). It is estimated that
approximately 100 billion PMN enter and leave the circulation daily
(Walker & Willemze, 1980). Despite this number, their fate after
migrating to tissues is unknown. Based on in vitro survivability of
peripheral PMN, it is thought that they probably remain functional for 1-2
days, and are then cleared (Golde, 1990).
Neutrophils are first-line defenders against invading
microorganisms. The frequent and severe infections that occur in persons
whose neutrophils are deficient quantitatively (neutropenia) or
qualitatively (e.g. chronic granulomatous disease or neutrophil actin
dysfunction) attest to the central role of this cell in host defense
mechanisms. They are the predominant leukocyte in peripheral blood
comprising two-thirds to three-fourths (4,000 8,000/mm3) of the
peripheral white blood cell count.

It is estimated that the bone marrow compartment houses PMN
reserves in excess of 100-fold the quantity found in peripheral blood.
These stores are readily discharged in response to bodily invasion.
Neutrophils are commissioned to seek, attack, and destroy pathogens and
so are endowed with the facility to rapidly advance and engage the enemy.
The major functions through which neutrophils carry out this mission
include chemotaxis, adherence, aggregation, phagocytosis, degranulation,
and generation of toxic metabolites. In part, these properties are
dependent upon an extensive rearrangement of the actin filament system

in certain parts of or throughout the cell (Stossel, 1992). The properties
of PMN chemotaxis, phagocytosis, adherence, and degranulation (the
principle movements) have gained the most attention in regards to the
relationship to cytoplasmic actin assembly.
In attempts to familiarize itself to the environment, the neutrophil
(and all animal cells to a certain extent) constantly forms and dissolves
diverse protrusions from its surface. Dependent upon external signals,
these protrusions take on particular shapes known as veils, ruffles, pleats,
filopodia, microvilli, and pseudopodia. Evidence suggests actin is
responsible for the elasticity of this organelle-deplete, actin-enriched
cortical region (Bretscher, 1991). The predominance of an agonistic signal
in certain regions allows for the formation of peripheral protrusions which
eventually lead to the motile properties of the neutrophil.


Actin Isoforms
In 1942 Bruno Straub discovered and isolated actin from rabbit
skeletal muscle (Straub, 1942). Nonmuscle cell actin was first purified
from Physarium by Hatano and Oosawa (1966). It has since been realized
that actin is ubiquitous prokaryoticc, plant, and animal cells), and often
the most abundant protein component in cells. Initially, it was thought
there were three major isoforms of actin in mammalian cells (Garrels &
Gibson, 1976). One muscle (a) and two non-muscle isoforms (3,
predominant non-muscle form, and y, second non-muscle form) could be
distinguished by isoelectric focusing; each with apparently distinct pIs
between 5.40 and 5.44. It has since been demonstrated in mammals and

birds that at least six actin isoforms exist (three a, one 3, and two y) and
are expressed in a tissue specific manner (Vandekerckhove & Weber, 1978;
Vandekerckhove & Weber, 1984). They differ from one another by only a
few amino acid residues, mainly in the first 20 N-terminal amino acids.
Utilizing near physiologic conditions, no major differences in the
polymerization characteristics of muscle vs. non-muscle isoforms of actin
are noted (Korn, 1982). As reviewed by Herman (1993), recent
compelling evidence suggests there are functional differences amongst the
muscle and non-muscle isoforms. Based on this evidence which includes
isoform-specific antibody localization and gene replacement studies, he
proposes a model cell wherein the nonmuscle isoforms are found in
regions of moving cytoplasm/cell cortex, and the muscle isoforms compose
the static microfilament bundles (stress fibers or myofibrils). Curiously, it
was recently noted that the cytoplasmic mRNAs for the p and y actin
isoforms are differentially segregated in myoblasts (Hill & Gunning, 1993),
and overexpression of either differentially impacts the cytoarchitecture of
these cells (Schevzov et al., 1992). This suggests subtle functional
differences exist even amongst the nonmuscle isoforms.

Actin Structure
Actin consists of a single polypeptide chain of 375 amino acid
residues, with a molecular weight of about 42 kDa (G-actin). It has long
been known that G-actin contains a divalent cation and an adenine
nucleotide binding site. Both sites are occupied in all forms of purified
actin. Due to its tendency to polymerize, the formation of actin crystals
and thus its atomic structure had eluded researchers for many years.
Utilizing the ability of deoxyribonuclease I (DNase I) to inhibit

polymerization, Kabash et al. (1990) were finally able to solve the
structure of the actin (muscle isoform) molecule to atomic (2.8 and 3.0 A)
resolution. Using this atomic structure along with lower resolution
analyses, the G-actin molecule can be described as an oblate sphere
(overall dimensions of 5.5 X 5.5 X 3.5 nm) composed of two major
domains (small and large) which are separated by a pronounced cleft. The
small and large domains are further divided into subdomains 1 & 2, and
subdomains 3 & 4 respectively. The four subdomains are stabilized
mainly by interactions with the nucleotide and divalent cation bound
within the cleft formed between the major domains. The crystalline
structure of beta-actin (non-muscle isoform) completed with profilin was
recently solved to 2.55 A. When compared to the c-actin structure of
Kabash et al., they appeared structurally similar with a notable 5' rotation
between the major domains.
The polar actin monomers can polymerize to form a polar actin
filament (F-actin). Geometrically, the filament can be described as a two-
start, double-stranded, right-handed helix with approximately 13 subunits
per turn. The precise orientation of the actin subunits is uncertain, but it

is thought that subdomain 2 of the monomer represents the "-" end (see
"Polymerization" below), and "+" end (see below) is within subdomain 3.

Multiple inter- and intrastrand contacts between the subunits exist. The
structure of the filament is very complicated and has yet to be directly
characterized at the atomic level. Holmes et al. (1990) have proposed an
atomic model of the actin filament which they derived from the atomic
model of the G-actin molecule and low resolution (8 A) F-actin structural
data. The intermolecular contact points in this model were recently
reviewed by Mannherz (1992).

Figure 1-1. Actin and its spontaneous polymerization in salt. The filament
shown here is decorated with heavy meromyosin which binds at a 45"
angle distinguishing the 2 ends of the filament. The barbed-end (+) is the
fast-growing end, and the pointed end (-) is the slow-growing end of the


SSalt ./


The 42,000 dalton actin monomer (G-actin) can non-covalently self-
associate and polymerize to form filaments (F-actin) in the presence of
monovalent (KC1) or divalent (MgCl2, CaC12) salts, and hence increase the

solution viscosity. The theory of actin polymerization has been reviewed
extensively by several authors (Korn, 1982; Stossel et al., 1985; Pollard &
Cooper 1986; Pollard, 1990), and will be briefly discussed. Actin
polymerization is at least a two-step process: nucleation followed by
elongation. Nucleation, the rate-limiting, thermodynamically-unfavorable
step, occurs when 3 molecules of actin associate to form the nucleus for
further assembly. Polymerization/elongation cannot proceed unless the G-
actin concentration is above a critical concentration (Cc), which is defined
as the minimum concentration of G-actin required to form a polymer
(Pollard & Cooper, 1986). The Cc value is sensitive to the ionic conditions,
pH, temperature, and nucleotide content of the reaction solution
(Zimmerle & Frieden, 1986; Zimmerle & Frieden, 1988). Despite the
minimal requirements, many groups have simulated physiologic
conditions (0.1 M KCI, 1mM MgCl2, EGTA-to chelate Ca2+, 1mM ATP, and

pH 7.4) during their analysis of polymerization kinetics. As alluded to
earlier, at near physiologic conditions the critical concentrations for
assembly of muscle and non-muscle actin isoforms are virtually
indistinguishable. Under defined conditions, actin filament assembly
reaches a steady-state as a result of actin molecules continually being
exchanged between actin filaments and the critical concentration of
monomers without affecting the overall F-actin content.
The actin filament has two kinetically different ends (barbed "+"
and pointed "-") defined by the arrowhead pattern produced when an

Figure 1-2. Rate constants for the association and dissociation of ATP-
actin and ADP-actin at filament ends. Adapted from Pollard (Curr. Opin.
Cell Biol. 1990. 2: 33-40)


1 0.7
1.2 \


0.13 .3
S 0.3

\ \8

Unit K+ = pmol-l s-1
K. = s-1

actin filament is labeled with heavy meromyosin ("rigor" conformation)
and visualized by electron microscopy (Huxley, 1963). The barbed (+)
ends have a greater exchange rate and a lower critical concentration (Cc+
= 0.1 pM) for monomer binding than the pointed (-) ends (Cc- = 0.6 pM)

(Bonder et al., 1983). Thus, the barbed ends are the more kinetically
active and preferred ends for filament growth. At steady-state, i. e. when
the association rate equals the dissociation rate, the apparent critical
concentration for the entire filament (Cc = 0.15 pM) is closer to that of the
barbed end (Korn, 1982). As a result, a steady-state condition termed
"treadmilling" occurs in which monomers slowly dissociate from the
pointed ends allowing new monomers to rapidly add on to the barbed
ends (Neuhaus et al., 1983; Wanger et al., 1985).
Actin polymerization is more rapid in the presence of ATP than in
the presence of ADP (Carlier et al., 1984; Pollard, 1984). This is not due to
energy made available through ATP-hydrolysis since polymerization can
occur faster than ATP-hydrolysis (Carlier et al., 1984). The differential
polymerization rates could be explained by the fact that ATP-G-actin has a
higher affinity and a lower dissociation rate than ADP-G-actin (Pollard &
Cooper, 1986). The hydrolysis of the ATP within the actin molecule
consists of two temporally distinct steps: the chemical cleavage of ATP
resulting in an ADP-Pi-actin followed by the slower release of Pi into the

medium (Korn et al., 1987). Additionally, it has been shown that the
cleavage of ATP occurs vectorally (Carlier et al., 1987). This suggests that
if one examines a filament during the polymerization of ATP-actin, the
newly added actin molecules are ATP-actin followed by ADP-Pi-actin, and
most internally, ADP-actin (Carlier, 1991). Since the dissociation rate of
ATP-actin is slower than ADP-actin, the delayed ATP hydrolysis promotes

actin filament assembly. Under these conditions, the extent of this "ATP-
cap" is dependent upon the polymerization rate. In contrast,
depolymerization is facilitated by the exposure of these internal ADP-actin
molecules. In this model, the presence of ATP and its hydrolysis may
partially regulate the rapid reorganization of actin structures.

Using electron microscopy to monitor the elongation rates of
preformed filaments as a function of added G-actin has enabled the
determination of the rate constants for the "elongation" reactions (Bonder
et al., 1983; Pollard, 1986). The differences between ATP-actin and ADP-
actin result in eight main reactions (Figure 1-2). Several labs have
determined various rate constants, and at least two labs have generated all
eight (Pollard, 1986; Korn et al., 1987). There is general agreement (at
least within an order of magnitude) regarding the values of these rate

The different affinities of the various forms of nucleotide-bound
actin monomers for filaments may be due to conformational changes. In
Kabsch's atomic models comparing ATP- and ADP-actin, there are direct

interactions between the calcium ion and the p- and y- phosphates of ATP
(3 alone with ADP) within the cleft, and subtle differences in the hydrogen
bonds formed between the phosphate groups with neighboring amino
acids (Kabsch et al., 1990). The atomic structure of Mg2+-ATP-actin,
although, may be quite different, as suggested by the fluorescence
difference seen when AEDANS (N-acetyl-Nl-[sulfo-1-
napthyl]ethylenediamine) labeled actin is in a Mg2+ verses a Ca2+
environment (Selden et al., 1986). Despite the literature available on the
effects of calcium or magnesium binding to actin (Estes, 1992), this

complicated topic will not be covered. Additionally, it is generally
accepted that Mg2+ is the divalent cation bound to actin in the cell.

Actin Binding Proteins

It was recognized early on that at least 50 % of the actin in extracts
from various non-muscle cells is nonpolymerized, in stark contrast to
muscle cells in which nearly 100 % of the actin is polymerized (Korn,
1982). Based on the earlier findings that nonmuscle isoforms of actin
share the same polymerization kinetics and critical concentrations as
muscle isoforms, it was hypothesized that the polymerization of actin in
nonmuscle cells was regulated through the interaction of actin with other
cellular components. This hypothesis was amply supported by the
identification of nonmuscle cell proteins that have specific interactions
with G-actin and/or F-actin.
On the basis of their ability to interact with actin in vitro, these
proteins were collectively termed "actin-binding proteins" (Stossel et al.,
1985; Pollard & Cooper, 1986; Hartwig & Kwiatkowski, 1991). Today, over
100 such nonmuscle proteins have been described. They are felt to be
responsible for the coordinated regulation of nonmuscle cell actin

polymerization considered necessary for motile activities as well as the
maintenance of structural integrity. Actin-binding proteins can be
subdivided into groups that either control the three-dimensional
arrangement of actin filaments (cross-linking and bundling proteins),

apply force to the actin-filament network (myosins), or regulate actin-
filament number and length (capping, severing, nucleating, and monomer-

binding proteins). Although the groups that control the three-dimensional
arrangement of actin filaments (gelation to solution or gel-sol reaction)
and those that apply force (contraction of actin gels) to the network are
almost certainly involved with the motile behavior of PMN (Condeelis,
1993), this discussion on actin-binding proteins will be limited to the
subset controlling actin-filament number and length.
The regulation of actin filament number and length is achieved by
four principle mechanisms. First, capping describes the ability of a
regulatory protein to bind to either the barbed ("+") or pointed ("-") end
of an actin filament preventing further addition or loss of monomers at
the particular end "capped." Severing is the mechanism whereby an actin-
binding protein is able to interact with the side of a filament and disrupt
the intermolecular interactions at that site, thus breaking the filament in
two. This is usually achieved by binding a "barbed" region within the
filament, and thus one of the newly severed fragments is capped at their
barbed end (Weeds & Maciver, 1993). The third mechanism, nucleation,
describes the ability of a protein to initiate elongation of actin filaments
without a lag phase. This may be achieved by either the rapid induction
of a polymerization-favorable conformation or stabilization of
intermediates in the formation of the trimeric nucleus. The forth
mechanism describes the ability of an actin binding protein to sequester
monomers in a 1:1 complex and inhibit their incorporation into filaments.
Therefore, if monomer-binding proteins are added to a solution of actin
monomers in stoichiometric amounts, the final extent of polymerization
upon the addition of salts will be decreased. The decrease will be
dependent upon the critical concentration of actin and the dissociation
constant of the interaction with actin monomers. It must be noted that the

in vitro analysis and in vivo interpretation of function of actin binding
proteins is often complicated by their ability to utilize more than one
mechanism to regulate actin filaments.

Monomer Sequestering Proteins

A high concentration of actin, well above the Cc necessary for
assembly at both ends of the filament, is stored in an unpolymerized form
in nonmuscle cells (100 200 UM or greater in PMN, assuming a uniform
distribution) and becomes available for assembly into filaments upon cell
activation. Monomer sequestering proteins are felt to account for a
majority of the unpolymerized actin found in these cells. The two major
families of proteins in this class are the profilins and the thymosins.
Mammalian profilins are low-molecular-weight proteins with Mr

around 15,000. Initially purified from spleen (Carlsson et al., 1977), it
has since been found in almost all mammalian cell types including PMN,
macrophages, and platelets (Southwick & Young, 1990; DiNubile &
Southwick, 1985; Markey et al., 1978). Purified profilin is able to
reversibly bind actin monomers with an apparent dissociation constant
(Kd) in the range of 1 to 10 pM (Larsson & Lindberg, 1988; Goldschmidt-
Clermont et al., 1991). For many years, profilin was thought to be the
major sequestering protein in nonmuscle cells, and simple sequestration
was its primary function. Estimations of profilin concentrations made
initially in platelets and subsequently in PMN suggested the concentration
of profilin (~40 pM) present within these cells is insufficient, when using
the simple model of sequestration, to account for the amount of
nonpolymerized actin (-100 200 pM) in resting states (Lind et al., 1987;
Southwick & Young, 1990).

In search for additional monomer sequestering proteins, a 5 kD
peptide was discovered in high concentrations in human platelets (Safer et
al., 1990). Sequence analysis revealed identity to a peptide believed to be
a thymic hormone, thymosin f4 (T94) (Safer et al., 1991). Based on its

wide distribution pattern and lack of secretary activity, this protein is no
longer believed to be a hormone. The dissociation constant (Kd) for the
TI4-actin complex has been reported to be in the 0.4 2.0 pM range

(Nachmias, 1993). It is calculated that in human PMN the cytoplasmic
concentration of Tf4 is ~150 pM, and together with profilin can account

for the sequestration of a majority of the G-actin in resting PMN
(Cassimeris et al., 1992). The amount of actin sequestered can be
estimated using the following equation for the dissociation constant (Kd)

of the sequestered monomer (Stossel et al., 1985):
Kd = [sequestering protein]free [Cc actin]/[sequestered G-actin]

Note that the concentration of sequestered actin is quite sensitive to the
critical concentration of actin.

With the recent suggestions that thymosin R4 alone is capable of
sequestering a majority of the G-actin in resting cells (platelets ~560 pM;
Weber et al., 1992), the question arose as to the role of the ubiquitous
protein profilin. Still controversial, several different actin regulatory
functions for profilin besides its ability to sequester monomers have been
proposed (Theriot & Mitchison, 1993). In the writer's opinion, the most
fascinating proposed mechanism is this protein's ability to interact
catalytically with actin in a fashion promoting exchange of the bound
nucleotide and divalent cation. It was noted early on that Acanthamoeba
profilin interacts with G-actin to increase the rate of exchange of the
bound nucleotide (Mockrin & Korn, 1980). This is consistent with

structural data obtained by chemical cross-linking and X-ray
crystallography which suggests profilin binds to actin in subdomain 3, a
region opposite the cleft wherein the adenosine nucleotide and divalent
cation bind (Vandekerckhove et al., 1989). The binding of DNase I to
subdomain 1 inhibits the nucleotide exchange rate, most likely by blocking
its exit from the cleft (Hitchcock, 1980; Mannherz et al., 1980). It is
foreseeable that conformational changes induced upon the binding of
profilin may alter the nucleotide and divalent cation binding sites within
the cleft. It was recently demonstrated that profilin, when bound to actin,
increases the off rate constant (k-) of the bound nucleotide by 1,000-fold

(Goldschmidt-Clermont et al., 1991). More importantly, kinetic modeling
of the nucleotide exchange rates revealed that at substoichiometric
amounts profilin is still able to accelerate the nucleotide exchange of the
whole actin population. In this model, the catalytic mechanism is possible
because the rate of profilin's exchange between actin monomers is
relatively rapid (sub-second), but slower than the rapid dissociation of
actin ligands (nucleotide, divalent cation) which occurs during the
transient binding of profilin to each actin molecule (Goldschmidt-
Clermont et al., 1991). As discussed earlier, the kinetics of actin
polymerization are highly dependent upon the type of nucleotide bound.
Therefore, this mechanism may be central to the cell's ability to rapidly
"recharge" actin molecules which are continuously being cycled through

polymerization in highly motile regions of the cell.

Capping Proteins

Experimental evidence suggests the assembly of actin in nonmuscle
cells occurs by addition of monomers to the barbed ends of actin filaments

(Wallace et al., 1984; Carson et al., 1986). All capping proteins purified
from non-muscle cells thus far cap (in vitro) the barbed ends of F-actin
filaments. Inhibiting exchange at the high affinity, barbed end effectively
raises the Cc to that of the pointed end. The steady-state result of the
interaction of these capping proteins with actin is increasing the
concentration of the G-actin pool and creating a large number of relatively
short filaments. Two proteins, accumentin and l-actinin, were originally
reported to cap the pointed end of F-actin filaments, but their activities
have since been shown to actually be barbed end capping of filaments
(acumentin) (Young, Maun, and Southwick unpublished data; Maruyama
et al., 1990). Barbed-end capping proteins are grouped based on their
functional and structural characteristics (Hartwig & Kwiatkowski, 1991;
Weeds & Maciver, 1993). The two major families of capping proteins are
the gelsolin family and the capZ family. The major functional difference
between these families is that proteins in the gelsolin family require
calcium to initiate the interaction with actin whereas the capping activity
of the capZ family occurs independently of calcium.
The gelsolin family of capping proteins is composed of the
structurally related mammalian (gelsolin, villin, capG, and scinderin) and
invertebrate proteins (fragmin and severin) (Mishra et al., 1994). Each
member is able to nucleate filament assembly and sever preformed
filaments (except capG; Southwick & DiNubile, 1986) in addition to their
collective ability to cap the barbed end. The founding and best
characterized member is gelsolin. Originally purified from rabbit alveolar
macrophages as a 91 kDa protein, gelsolin is now known to be widely
distributed (including muscle cells) (Yin & Stossel, 1979; Yin et al., 1981).

As reviewed by Stossel et al., (1985), the three in vitro effects of
gelsolin on actin are differentially dependent upon the Ca2+
concentration. Severing has the most stringent calcium requirement, and
does not occur in the presence of submicromolar Ca2+. Barbed-end
capping occurs to a small extent in the presence of submicromolar Ca2+,
but this activity is markedly enhanced at higher calcium concentrations.
Gelsolin is able to nucleate filament assembly in the presence of calcium

by forming a complex (1 gelsolin: 2 actin) with two molecules of actin
(Janmey et al., 1985). Removal of calcium by EGTA dissociates one of the
actin monomers leaving behind a stable 1:1 gelsolin-actin complex. This
EGTA-resistant complex is unable to sever filaments, but can now bind the
barbed ends of filaments with high affinity even at submicromolar calcium
concentrations. Gelsolin has also been isolated as an alternatively spliced,
93 kDa, plasma form differing from the cytoplasmic form by the presence
of an additional 25 N-terminal amino acids (Kwiatkowski et al., 1986).
The secreted gelsolin is hypothesized to maintain and buffer a low plasma
actin filament content even when the cytoplasmic components of cells are
spilled into the vasculature during processes such as injury.

The capZ family includes heterodimeric proteins with subunits of
30-35 kDa which are able to bind the barbed end of actin filaments
independently of calcium. Most members described can nucleate filament
assembly, but all lack the ability to sever filaments. CapZ, a heterodimeric
protein with subunits of Mr = 36,000 (a-subunit) and 32,000 (P-subunit)
daltons, is the best characterized member of this family (Casella et al.,
1986). Purified skeletal-muscle capZ caps the barbed end of filaments
with an apparent dissociation constant of approximately 0.5 1 nM
(Caldwell et al., 1989a). It was originally purified from skeletal muscle

and localized to the sarcomeric Z-line (Casella et al., 1987). CapZ has
subsequently been recognized in nonmuscle cells as well (Caldwell et al.,
1989b; Schafer et al., 1992). It has recently been recognized that barbed-
end capping proteins purified from Acanthamoeba (Isenberg et al., 1980),
Dictyostelium (Schleicher et al., 1984), Xenopus (Ankenbauer et al., 1989),
Saccharomyces (Amatruda & Cooper, 1992) and bovine brain (Kilimann &
Isenberg, 1982) are indeed analogues of capZ. The presence of
mammalian capZ in highly motile cell types such as PMN and macrophages
has yet to be noted. It was recently reported to be present in human
platelets (Barkalow & Hartwig, 1994; Nachmias et al., 1994).

Actin Dynamics In PMN

PMN Chemoattractants

Chemoattractants are substances which have the ability to induce
directed migration. The three major groups of neutrophil
chemoattractants are described as complement-derived (C5a), leukocyte-
derived (LTB4, PAF, 11-8), and bacterial-derived (formylated peptides, e.g.
fMet-Leu-Phe). Neutrophils can sense a chemotactic concentration
gradient as small as 1 % across their dimension (Zigmond, 1977). As
discussed below, the rapid, chemoattractant-induced polymerization of
actin is probably the best understood transduction mechanism involved in
regulating the actin microfilament network. Overall, the transduction
mechanisms which regulate cellular actin assembly, although, remain
poorly understood.
Curiously, nanomolar concentrations of chemoattractants are
enough to induce shape change and locomotion in neutrophils, whereas it

takes much higher (micromolar) concentrations to induce the microbicidal
mechanisms of PMN (Gallin, 1988). Chemoattractants transmit their signal
through specific surface receptors, and many appear to utilize a common
signal transduction pathway. The bimodall" response of PMN to
chemoattractants (differential triggering of motile verses microbicidal
functions) is currently thought to result from the activation of additional
transduction pathways dependent upon the extent of the chemotactic
stimulus (Snyderman & Uhing, 1992). The remaining discussion in this
chapter will focus on the rapid actin assembly occurring during the initial
phase as the transductional events of the second phase are poorly
understood. The fMet-Leu-Phe peptide has been termed the neutrophil
"pan-activator" for its ability to induce almost all major PMN functions in

addition to chemotaxis (Becker, 1987), and consequently its signal
transduction mechanisms have been extensively studied (for review see
Snyderman & Uhing, 1992).

Rapid Polymerization In Response To fMet-Leu-Phe

The overall morphology of PMN dramatically changes in response to
chemotactic stimulation (Fechheimer & Zigmond, 1983; Howard & Oresajo,
1985). At rest the cells are spherical, and upon stimulation form
peripheral ruffles or lamellipodia at their surface which have been shown
microscopically to be highly concentrated in F-actin (Fechheimer &
Zigmond, 1983; Sheterline, et al., 1984a). Neutrophils contain
approximately 5 pg actin/cell, which correlates to roughly 10 % of the
total cellular protein in these cells (Sheterline et al., 1984b). Others have
calculated that actin represents as high as 20 % of PMN extracts
(Southwick & Young, 1990). Probably the best evidence suggesting actin

dynamics is important in PMN cell motility are the changes in the ratios of
monomeric to filamentous actin seen when cells are stimulated to undergo
motility. The rapid, chemoattractant-induced actin polymerization
response occurs in many eukaryotic cells, in addition to PMN, which are
able to undergo "amoeboid-like" motility (Stossel, 1992; Caterina &
Devreotes, 1991).

In resting PMN approximately 30 40 % of the total actin
concentration is F-actin (Feichheimer & Zigmond, 1983; Sheterline et al.,
1984b; Bengtsson et al., 1986). Assuming a uniform distribution, this
corresponds to roughly 100 pM F-actin and 200 pM G-actin (Southwick &
Young, 1990). Remarkably, several laboratories have demonstrated that
within 30 seconds or less of fMet-Leu-Phe stimulation, the actin filament
content doubles (F-actin: -100 -> 200 pM) (Shalit et al., 1987; Howard &
Meyer, 1984; Fechheimer & Zigmond, 1983; Omann et al., 1987; Lofgren et
al., 1993). Cytochalasins, in vitro inhibitors of actin polymerization,
inhibit the morphological changes associated with neutrophil motility
(Zigmond & Hirsch, 1972) as well as the induced rise in polymerized actin
(Southwick et al., 1989; Cassimeris et al., 1990; White et al., 1983).
Furthermore, the dose dependency of the fMet-Leu-Phe-induced
chemotaxis and the rapid, actin-polymerization response are similar
(Howard & Meyer, 1984).

Dynamic changes in actin filament content occur during other PMN
motile events such as phagocytosis, degranulation, and adherence. A
nearly twofold rise in actin filament content is also observed during these
neutrophil functions (Boyles & Bainton, 1981; Fechheimer & Zigmond,
1983; Southwick et al., 1989), but the dissection of the transduction
cascade is complicated by the activation of multiple pathways. Although

these functions can be initiated by fMLP stimulation, their predominant
signal transduction pathways appear to be different (Southwick et al.,
1989; Rosales et al., 1994).

fMet-Leu-Phe Signal Transduction Pathway
The N-formylated methionyl peptides such as fMet-Leu-Phe activate
neutrophils by binding to a specific cell surface receptor. The primary
structure of the fMet-Leu-Phe receptor has been deduced from its cDNA
sequence, revealing it belongs to the G-protein-coupled receptor family
(Boulay et al., 1990). Nearly all G-protein-coupled receptors share a
similar sequence motif which allows for seven-transmembrane-segments
(7-TMS motif) that span the lipid bilayer (Dohlman et al., 1991). The
recent sequence data only confirmed what had been suspected about the
fMet-Leu-Phe signal transduction cascade based on biochemical studies.
Pertussis toxin, a well known inhibitor of some G-proteins, inhibits
many chemotactic responses in a concentration dependent fashion (Brandt
et al., 1985; Becker et al., 1985; Bengtsson et al., 1986). Guanine
nucleotides, and not adenine compounds, regulate the affinity state of the

plasma purified f-Met-Leu-Phe receptors (Sklar et al., 1987). Chemotactic
factors stimulate membrane associated GTPase-activity and guanine
nucleotides can potentiate fMet-Leu-Phe induced activation (Feltner et al.,
1986; Verghese et al., 1986). Additionally, a GTP-binding, 40 kDa
pertussis-toxin-substrate complex copurifies with the fMet-Leu-Phe
receptor after several chromatographic steps. Immunochemical
identification and cDNA sequence analysis reveal the PMN G-protein
coupled to the fMet-Leu-Phe receptor is of the Gi2 (pertussis toxin

Figure 1-3. fMet-Leu-Phe induced phospholipase C signaling pathway. Stimulation of PMN results in the
doubling of intracellular F-actin in < 30 sec.




Calciosome > TCa2+

sensitive, phospholipase C activator) subtype (Snyderman & Uhing, 1992;
Kaziro et al., 1991).
The kinetics and dose-dependency of the fMet-Leu-Phe-induced
actin polymerization burst coincides with the rapid, initial phase of the
fMet-Leu-Phe response (discussed under "PMN chemoattractants"). The
following sequence of major intracellular events comprise this initial
phase. In short, binding of fMet-Leu-Phe to its receptor results in the
dissociation of a membrane-bound, heterotrimeric, GTP-binding protein
(G-protein). This activated G-protein (Gi2-alpha subunit) transiently

activates phospholipase C to produce two second messengers, 1,2-
diacylglycerol (DAG) and inositol (1,4,5) triphosphate (IP3) from
phosphatidylinositol(4,5) bisphosphate (PIP2) hydrolysis. Inositol

triphosphate releases intracellular stores of calcium, while DAG activates
protein kinase C. The rapid cleavage of PIP2 by PKC results in the
conversion of other phosphoinositides such as PIP and PIP3 to PIP2. The
conversion of phosphoinositides is regulated by a large number of kinases
and phosphatases that add or remove phosphates on the inositol ring
(Janmey, 1994). Temporal studies show rapid (< 5 sec) elevations in
phosphoinositide metabolism and cytosolic calcium levels following fMet-
Leu-Phe stimulation. The molecular mechanisms responsible for the
cytochalasin-sensitive, rapid (< 15-30 sec) and dramatic increase in actin
polymerization in response to fMet-Leu-Phe remain unknown, but are
thought to involve these molecules.

Calcium And Protein Kinase C-Indeoendent. Phosohoinositide-Dependent
fMLP-Induced Actin Polymerization Response
Modulation of the transductional cascade using traditional
pharmacological agents demonstrates that neither calcium or protein

kinase C (PKC) are primarily associated with the rapid polymerization of
actin in PMN stimulated with fMet-Leu-Phe. Since this fMet-Leu-Phe-
induced increase in F-actin content in PMN is totally abolished if cells are
pretreated with pertussis toxin (Bengtsson et al., 1986; Omann et al.,
1991), investigators have actively pursued potential signals distal to the G-
protein, but proximal to calcium release and protein kinase C activation.
In addition to calcium, the in vitro actin regulatory activities of various
actin binding proteins are modulated by phosphoinositides. These in vitro
findings in combination with the observation that phosphoinositides are
rapidly turned over in PMN stimulated with fMet-Leu-Phe has led to the
hypothesis that phosphoinositides are primarily involved in the regulation
of the rapid, actin-polymerization response (Stossel, 1989; Janmey, 1994).
Phorbol esters, exemplified by phorbol myristate acetate (PMA),
have been shown to bind, translocate, and activate PKC in human
neutrophils (Castagna et al., 1982; Nishihira et al., 1986), and are
consequently utilized as specific PKC activators. Only minor increases in
F-actin content are noted when resting human neutrophils are stimulated
with levels of PMA that phosphorylate much more effectively than fMLP

(Bengtsson et al., 1986; Howard & Wang, 1987). Conversely, several
protein kinase inhibitors with relative specificity for PKC (i.e.,
staurosporine, CGP 41251, and H7), when used at functional
concentrations, do not significantly alter the rapid, 2-fold increase in F-
actin seen within 30 seconds of fMet-Leu-Phe stimulation of neutrophils
(Niggli & Keller, 1991; Keller & Niggli, 1993; Sham et al., 1993).
Additionally, neutrophil chemotaxis induced by fMet-Leu-Phe is only
slightly inhibited by staurosporine (Boonen et al., 1993).

Chemotactic stimulation increases the intracellular free calcium
concentration, [Ca2+], from the 100 nM level at rest to micromolar levels
(Janmey, 1994). Despite buffering resting neutrophil [Ca2+]i with calcium

chelators (e.g., EGTA, quin-2, BAPTA), the cells are still capable of
migrating at normal rates (Elferink & Deierkauf, 1985; Meshulam et al.,
1986; Zigmond et al., 1988). The fMLP-induced rise in [Ca2+] involves
both the release of calcium from intracellular stores (via IP3) and influx

across the membrane. The use of fluorescent calcium indicators (quin2 or
fura2) reveals the "biphasic" nature of this fMLP-induced increase (Lew et
al., 1984; Lew et al., 1986; Anderson et al., 1986). The prolonged second
phase is eliminated by removal of extracellular calcium, while the rapid
first phase remains unaffected. The rapid doubling of F-actin in response
to fMet-Leu-Phe stimulation is temporally associated with the initial rise in
[Ca2+]i caused by IP3 (Snyderman & Uhing, 1992). Preventing the fMet-
Leu-Phe-induced rise in calcium with calcium chelators has no effect on
the neutrophil's ability to double its F-actin content within 15 30 seconds
(Sha'afi et al., 1986; Downey et al., 1990; Sham et al., 1993). Neutrophil
migration is a complex process that involves the continuous cycling of
polymerization and depolymerization. It is possible that the repetitive
increases in intracellular calcium in migrating PMN (Marks & Maxfield,
1990; Jaconi et al., 1990) may actually enhance depolymerization of actin
filaments by activating the severing activity of proteins like gelsolin
(Downey et al., 1990).
With increasing evidence suggesting neither calcium nor protein
kinase C activation were necessary for the chemoattractant-induced actin
polymerization seen in PMN, the signaling role of phosphoinositides was
actively investigated. It has been known for over 30 years that

phosphoinositide phospholipids such as PIP2 can form complexes with

certain proteins (Janmey, 1994). Since phosphoinositide turnover via
phospholipase C hydrolysis of PIP2 is one of the major intracellular events

occurring proximal to either the increase in calcium or activation of
protein kinase C in activated PMN, the proposal that phosphoinositides
might regulate actin assembly seems plausible (Lassing & Lindberg, 1985;
Stossel, 1993). With the finding that phosphoinositides modulate the in
vitro interaction of profilin with actin, Lassing and Lindberg proposed
phosphoinositides may regulate actin assembly in cells. They were able to
demonstrate that actin sequestered by profilin can be dissociated by PIP2,

thus leading to actin polymerization (Lassing & Lindberg, 1985). Since
that time, numerous other actin-binding proteins have been shown to
interact with phosphoinositides (Isenberg, 1991; Janmey, 1994).
Interestingly, the interaction of barbed-end capping proteins with actin is
also specifically inhibited in the presence of polyphosphoinositide
phospholipids (Janmey, 1994).

Stossel has recently proposed a model that integrates motile signal
transduction events with the modulation of actin binding proteins in
attempts to explain how animal cells are able to crawl in response to a
stimulus (Stossel, 1993). The motile behavior of certain animal cells is
partly explained by the rapid polymerization of actin. In this model, a
motile stimulus results in the rapid turnover of intracellular
phosphoinositides. This in turn causes release of sequestered monomers
of actin, and prevention of barbed-end cap formation. The net result
would be a rapid polymerization of monomeric actin onto the barbed-ends
of actin filaments.

Utilizing fluorescent actin probes, it can be demonstrated that
fibroblasts undergo actin filament assembly and disassembly. The
lamellipodium of a cultured fibroblasts includes a dense network of actin
filaments (Wang, 1985). Fluorescence photobleaching of a discrete region
within the lamellipodium demonstrates centripetal movement and
eventual disappearance at the proximal edge of the actin band, a process
not unlike treadmilling. Their finding suggests actin filaments within the
lamellipodium are undergoing net assembly at their distal, membrane-
directed barbed ends and net disassembly at their proximal pointed ends
(Wang, 1985). Electron micrographs of neutrophils whose actin filaments
are decorated with heavy meromyosin reveal the barbed ends of actin
filaments are directed peripherally towards the membrane (Pryzwansky et
al., 1983). In addition, when PMN are stimulated with fMet-Leu-Phe, the
number of nucleation sites for actin polymerization is increased
transiently (Carson et al., 1986). These nucleation sites are sensitive to
cytochalasin, a drug whose actin inhibitory activity is partly explained by
its ability to cap the barbed-end of filaments, and they sediment with the
detergent insoluble actin cytoskeleton (Carson et al., 1986). These
findings suggests that neutrophils rapidly increase their F-actin content by
increasing the number of free barbed-ends upon chemotactic stimulation.

Calcium Independent Model For Polymerization

The two mammalian cell types which undergo a rapid (5 30 sec.)
doubling of actin in response to stimulation are phagocytes neutrophilss
and macrophages) and platelets. Profilin and thymosin 94 have both been

purified in these cell types. The only capping proteins that have been

purified from phagocytes are members of the calcium-dependent/gelsolin
capping-protein family. CapZ was recently noted in platelets (see
"Capping Proteins")
A simple model integrating what is currently known about
neutrophils and other motile cells is suggested by Stossel (1993). At rest,
a majority of the actin filaments are capped at their membrane-proximal
barbed-ends. This leads to a critical concentration of free actin monomers
close to the dissociation constant of the freely-exchanging minus-end (~1.0
pM). As discussed earlier, the concentrations of monomer sequestering
proteins (profilin and thymosin R4) and their affinities for actin
monomers (Kd ~0.5 5 pM) can explain the high amounts of

unpolymerized actin in resting neutrophils (> 60 %). Stimulation of
neutrophils with chemotactic peptide causes an increased
phosphoinositide turnover (Snyderman & Uhing, 1992; Janmey, 1994).
Based on the findings that profilin and the barbed-end capping proteins
bind the phosphoinositides PIP and PIP2 thereby inhibiting their ability to

bind to actin, it is thought that phosphoinositides mediate the signal
transduction between the surface receptors and actin assembly (Stossel,
1993). In this model it is explained that phosphoinositides cause the
uncapping of membrane-proximal barbed-ends, thus lowering the critical
concentration of actin monomers in that region closer to that of the
barbed-end (0.1 0.3 pM). Since the critical concentration of free actin
monomers is lowered, a majority of the sequestered actin can be released
into the polymerizable pool (see "Monomer Sequestering Proteins").
Stimulation of neutrophils may activate profilin's ability to increase the
nucleotide exchange rate of actin molecules, thus ensuring a continuous

supply of ATP-actin available for rapid polymerization (Goldschmidt-

Figure 1-4. Regulation of actin assembly in PMN. This simple model attempts to explain the rapid
assembly of actin in fMet-Leu-Phe stimulated PMN.





fMLP < 30 sec
(tPhosphoinositide Turnover)

+ CC _

[] OOO 0 ADP


r n ATP

Clermont et al., 1991). The role of profilin in activated PMN remains
controversial since the PIP2 effect would theoretically prevent ATP/ADP

There have been no reports of calcium-independent capping
proteins in neutrophils or macrophages. Two abstracts were recently
presented at the ASCB (American Society for Cell Biology) meeting
December, 1994 noting the presence of capZ in platelets (Barkalow &
Hartwig, 1994; Nachmias et al., 1994). These three cells are the
mammalian model cells for understanding the signal transduction
mechanisms leading to rapid actin assembly. As discussed above, the
rapid polymerization response in neutrophils can occur independently of
calcium. It is upon these premises that we have further investigated the
calcium insensitive activity of PMN Actin Polymerization Inhibitor, an
activity originally purified by my mentor Dr. Southwick.

PMN Actin Polymerization Inhibitor

A Maior Calcium-Independent Activity In PMN Extracts

Approximately 15 years ago, Southwick and Stossel isolated an actin
binding activity from human neutrophils thought to account for much of
the unpolymerized actin in granulocyte extracts (Southwick & Stossel,
1981). This activity, PMN actin polymerization inhibitor, was shown to
decrease the viscosity of purified skeletal-muscle actin under polymerizing
conditions in the presence of 1 mM EGTA and absence of added calcium.
Substoichiometric concentrations of the inhibitor were still able to
decrease the viscosity of purified actin allowed to polymerize in the
presence of 0.1 M KC1. PMN Actin Polymerization Inhibitor was purified

from granulocyte extracts by DEAE-ion exchange chromatography and gel-
filtration chromatography. The viscosity lowering activity of the column
fractions was monitored using a Cannon-Manning semi-microviscometer.
The purified inhibitor was noted to contain polypeptides of 65,000 and
62,000 daltons by SDS-PAGE. The Stokes radius of the inhibitor was
reportedly 32 A and the s20,w was 4.8. These data were felt to be

compatible with the inhibitor being a globular monomer with a native
molecular weight similar to that of the 65,000 and 62,000-dalton peptides
resolved by SDS-Page. Increasing KCI concentrations from 0.1 0.6 M KCI
reversed the inhibition.

All activity studies were done in the absence of added calcium and
presence of 1 mM EGTA to avoid the potential contaminating activity of
the calcium dependent protein gelsolin. The PMN actin polymerization
inhibitor is hence a major calcium-independent actin filament regulatory
protein. The recent findings that the PMN actin polymerization-burst in
response to fMet-Leu-Phe occurs independently of calcium has prompted
my further examination of this inhibitor's actin regulatory role.


Isolation Of Human Polvmorphonuclear Leukocvtes (PMN)

Fresh leukocyte enriched fractions from the whole blood of healthy
donors were obtained from the community blood bank. For each
preparation, 15 45 buffy coat units were used. Further isolation of
leukocytes was accomplished using a variation of the technique described
by Southwick and Stossel (1981). Briefly, leukocyte enriched fractions
were pooled and sedimented through Dextran T-500 (Pharmacia Biotech
Inc., Piscataway, NJ). This procedure yielded on average 85 % PMN the
remainder of the cells representing lymphocytes (12 %) and monocytes (3
%). In some cases this procedure was followed by Ficoll-sodium diatrizoate
sedimentation. This procedure increased the purity of PMN to 95 %. CapZ
obtained from granulocytes isolated with or without a Ficoll-sodium
diatrizoate sedimentation resulted in purified protein with identical actin
binding activity. Higher yields of PMN capZ were achieved using dextran

Purification Of CapZ From Human PMN

All steps were done at 4 'C whenever feasible. The initial
procedures used were nearly identical to those described previously

(Southwick & Stossel, 1981). Briefly, cells were cooled to 4 "C and washed
twice in normal saline, treated with 5 mM diisopropylfluorophosphate
(DFP) for 15 min and then resuspended in 3 volumes of a solution
containing 5 mM dithiothreitol (DTT), 2 mM ATP, 20 mM EGTA, 20 mM
imidazole-HCL, pH 7.5, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1
g/100 ml leupeptin, 4 g/100 ml aprotonin, and 7.8 g/100 ml benzamidine
(homogenization solution). Cells were broken by nitrogen cavitation at
900 psi. This procedure ruptured greater than 95 % of all cells, as
monitored by phase microscopy. The homogenates were centrifuged at
12,100 X g for 60 min. The supernatant was dialyzed against 500
volumes of S2 buffer (10 mM Imidazole pH 7.8, 5 mM DTT, 1 mM MgCl2,

1 mM ATP, 1 mM EGTA) for 3 hrs, then diluted 1:1 with S2 buffer and
applied to a 1.5 cm X 16.5 cm DEAE-Sepharose CL-6B anion exchange
column (Pharmacia Biotech Inc. Piscataway, NJ) equilibrated with 120-200
ml of S2 buffer. After application of the supernatant, the column was
washed with 60 ml of 0.08 M KCI in S2 buffer, followed by 250 ml linear,
0.08-0.4 M KCI, gradient in S2 buffer. Eluted fractions were analyzed by
coomassie blue stained SDS-PAGE and by falling ball microviscometry
(MacLean-Fletcher & Pollard, 1980). Fractions that lowered the viscosity
of F-actin under low Ca2+ conditions were pooled and concentrated to a
volume of 0.5-3.0 ml (< 10 mg/ml) in a nitrogen pressure concentrator
using an Amicon PM-30 ultrafiltration membrane (Amicon Corp.,
Lexington, Mass.). The concentrated sample was then subjected to gel
filtration using a protein-pak 125 (Waters Associates, Milford, MA) column
equilibrated, and eluted with 0.1 M KCI in S2 buffer. Several precautions
were taken to limit activation of the abundant proteases present in PMN.
Isolation of PMN was performed using plastic to avoid generalized

activation by adherence to glass, Purified PMN were treated with the
potent neutral serine protease inhibitor DFP, cells were homogenized in
the presence of EGTA and a protease inhibitor cocktail, and efforts were
made to maintain the purification at 4 *C. As originally reported
(Southwick & Stossel, 1981), the PMN viscosity lowering activity at this
stage copurified with two polypeptides in the 60 68 kDa range.
In efforts to further purify this activity, three additional columns
were utilized in the following order; High S (Bio-Rad Laboratories, Inc.,
Hercules, CA), Mono Q(Pharmacia Biotech Inc., Piscataway, NJ), and
HTP/hydroxylapatite (Bio-Rad Laboratories, Inc., Hercules, CA). To
minimize protein loss during screening activity of these additional
columns, the Band 4.1 capping assay was substituted for the falling-ball
microviscometry assay (see "Actin Binding Studies" below). For the High
S cation exchange chromatography, active fractions were pooled and
dialyzed against a buffer containing 8 mM MES pH 6.0, 1mM sodium azide,
and 0.1 mM PMSF (buffer S). The dialyzed fractions were then applied to
a 5 ml Econo-Pac High S cation exchange column which had been
equilibrated with buffer S. The column was washed with 5 ml of buffer S,
followed by a 30 ml (1 ml/min) linear salt gradient to 0.150 M KCI in
buffer S. Active fractions from the High S column were pooled then
dialyzed against S2 buffer. After preclearing the fractions by
centrifugation at 10,000 X g for 5 minutes, they were applied to a Mono Q
HR 5/5 anion exchange column equilibrated with S2 buffer. The column
was then washed with 5 ml (1 ml/min) of 0.1 M KCI in S2 buffer, and
followed by a 12 ml linear gradient, 0.1 0.3 M KC1, gradient in S2 buffer.
Similar to the purification step described by Casella et al. (1986), our final
step entailed pooling and dialyzing Mono Qactive fractions against a 10

mM potassium phosphate buffer pH 7.0 containing 0.1 mM DTT. This
sample was then applied to a 1 ml hydroxylapatite (Econo-Pac HTP)
column equilibrated with the same buffer, and eluted with a linear
gradient to 75 mM potassium phosphate pH 7.0, and 0.1 mM DTT. Active
fractions were quantified using the Quantigold assay (Diversified Biotech)
and then stored at -20 *C in 30 % ethylene glycol (under these storage
conditions the specific activity of the protein remained stable for >3

Purification Of Muscle Actin

Actin was purified from rabbit skeletal-muscle by the method of
Spudich and Watt (1971). Monomeric actin was gel filtered through a
Superdex 200 16/60 (Pharmacia Biotech Inc., Piscataway, NJ) column for
polymerization kinetic studies. Pyrenylactin was prepared according to
Kouyama and Mihashi (1981) with the modifications described previously
(Young et al., 1990). Actin concentrations were calculated using the
extinction coefficient of 24.9 mM-1 cm-1 at 290 nm with the correction for

pyrenyl absorption according to Selden and colleagues (1983).

Amino Acid Sequence Analysis

Peptides were electroblotted from SDS-polyacrylamide gels onto
polyvinylidene difluoride (PVDF) membranes using a transfer buffer
containing 12.5 mM Tris, 96 mM glycine, 10 % ethanol, pH 8.3 as
previously described (Mozdzanowski et al., 1992). The samples were
blotted, stained, and sent to our collaborator Dr. David W. Speicher at The

Wistar Institute Philadelphila, PA for amino acid sequence analysis.
Coomassie blue stained bands were excised and sequenced on an Applied
Biosystems model 475A sequencer using gas phase TFA delivery and an
on-line model 120A PTH analyzer with modifications as previously
described (Reim et al., 1992). Acetonitrile and tetrahydrofuran were from
J. T. Baker. All other sequencer solvents and reagents were from Applied

Polvacrvlamide Gel Electrophoresis And Western Blot Analysis


Protein samples were subjected to electrophoresis on 10 %
discontinuous pH mini (Bio-Rad Laboratories, Inc., Hercules,
CA) or mid-sized (Hoefer Scientific Instruments, San Francisco, CA) SDS-
PAGE. Molecular weights (Mr) of proteins of interest were determined by

simultaneous electrophoresis of other polypeptides of known molecular
weights, and plotting their relative mobility verses log Mr. Proteins were

visualized by staining the gels with Coomassie brilliant blue R-250 or
Silver (Silver stain plus kit, Bio-Rad Laboratories, Inc., Hercules, CA).


The electrophoresed proteins were then transferred to nitrocellulose
paper (graphite electroblotter, Millipore) using standard protocols
(Towbin et al., 1979). The transferred blots were subjected to a blocking
step in 3 5 % non-fat dry milk in phosphate-buffered saline (PBS) and 0.3
% tween-20 for at least three hours. The primary antibodies were diluted
appropriately in the blocking solution (anti human annexin VI at 1:4000

(monoclonal) & 1:7500 polyclonall), anti human L-plastin at 1:500, anti
chicken capZ alpha and beta 1:300) and incubated with the blot for 1 hour
at room temperature. After subsequent washes with PBS-tween, the blots
were incubated with the appropriate alkaline phosphatase-conjugated
secondary antibody for 40 60 minutes. The detection was in an alkaline
phosphatase buffer (0.1 M Tris, 5 mM MgCI2, 0.1 M NaCI) containing 30

pg/ml nitro blue tetrazolium (Sigma Chemical Company, Saint Louis, MO)
and 20 pg/ml BCIP (5-bromo-4-chloro-3-indolyl phosphate, Sigma).
The annexin VI concentration in neutrophil extracts was determined
using a method similar to that described by Dabiri et al. (1992). The
immuno-reactivities of the polypeptides in neutrophil extracts were
compared to known concentrations of purified annexin VI using a
scanning laser densitometer (Zeineh Model SLR-2D/1D, Biomed
Instruments, Inc.). A linear standard curve could be generated relating
densitometry integration units to protein concentration.

Nondenaturing Polvacrvlamide Gel Electrophoresis
Neutrophil actin polymerization inhibitor containing predominantly
annexin VI and L-plastin were subjected to 7.5 % nondenaturing
polyacrylamide gel electrophoresis (mid-sized gel, Hoefer Scientific
Instruments, San Francisco, CA) according to the method of Safer (1989).
Samples were added 1:1 (volume:volume) with 10 % glycerol in the
running buffer with <1 mg of bromophenol blue (tracking dye). A
heterogeneous population of monomers, dimers, and trimers of bovine
serum albumin (BSA) are formed when placed in solution, and therefore
15 pg of BSA (fraction V, Sigma Chemical Company, Saint Louis, MO; 1

mg/ml in water) was run simultaneously to the proteins of interest as
evidence of nondenaturing resolution of proteins.

Actin Binding Studies

Pyrene actin was used for all kinetic studies. Fluorescence intensity
was monitored using a Perkin-Elmer LS-5 fluorescence spectrophotometer
with excitation and emission wavelengths of 364 nm and 407 nm

Depolymerization Assay

Pyrene labeled actin (2 pM) was allowed to polymerize to steady
state in the presence of 1 mM EGTA or 1 mM calcium. The F-actin was
then diluted to 50 nM into a buffer containing 10 mM Imidazole pH 7.5,
0.5 mM ATP, 0.1 M KC1, 1 mM MgCl2, 1 mM DTT (Buffer P) in the presence

of varying concentrations of capZ (Southwick & DiNubile, 1986). The
reaction was observed for 30 min at 25 'C.

Actin-Spectrin Nuclei Elongation Assay

Red blood cell derived spectrin/band 4.1/actin nuclei (rbc nuclei)
were isolated according to the methods of Casella et al. (1986). Pyrene-

labeled G-actin (0.55 pM) was added to a mixture of 1.25 pg/ml of rbc
nuclei, varying concentrations of capZ and buffer P containing 1 mM
EGTA. In the absence of rbc nuclei, this concentration of pyrene actin
(0.55 pM) failed to spontaneously polymerize during the time course of
our experiments (15 min). The capZ and rbc nuclei were allowed to
incubate for 2 min prior to the addition of the pyrene actin.

Steady-State Assay

Pyrene-labeled actin in buffer P with 1 mM EGTA was allowed to
polymerize at room temperature for 18 hours (steady-state), and the
critical concentration was determined by plotting the total actin
concentration verses the steady-state fluorescence (correcting for
fluorescence values at t = 0). The effects of various concentrations
(substoichiometric relative to actin) of capZ on the steady-state
fluorescence of pyrene actin (0.9 pM or 3 pM) was measured. The steady-
state G-actin concentration (apparent critical concentration) of the
solutions was calculated from the decrease in steady-state fluorescence
relative to pyrene actin controls. The fluorescent signal of pyrene actin is
proportional to polymer weight concentration (Cooper et al., 1983).

Nucleation Assay

For this assay, 1.5 pM pyrene labeled G-actin was allowed to
polymerize in buffer P in the presence of varying concentrations of capZ.
The effects of capZ on this polymerization rate were assessed by
fluorimetry (Southwick & DiNubile, 1986).

Severing Assay

Gelsolin (0.01 pM) and pyrene actin (2 pM) were copolymerized
(molar ratio 1:200) to steady-state, forming barbed-end capped filaments
(Casella et al., 1986). These filaments were then diluted to 100 nM in
buffer P containing 1 mM calcium and varying concentrations of capZ.

Monomer Sequestration Assay

Gelsolin and non-pyrene labeled actin were copolymerized at high
molar ratios (1:16) in buffer P forming nuclei for pointed end
polymerization. Pyrene labeled G-actin (final of 0.8 pM) was added to
buffer P containing gelsolin/actin nuclei (15 nM/0.24 pM) and varying
concentrations of capZ (Young et al., 1990). This assay was repeated in
the presence of G-actin that was 50 % labeled to examine the possibility
that capZ binds pyrenylactin with lower affinity. These reactions were
performed in the presence of either 1 mM EGTA or 0.5 mM calcium.

Falling Ball Microviscometry

The falling ball assay of MacLean-Fletcher and Pollard (1980) was
used to follow the relative activities of individual elution fractions at
various stages during the purification of neutrophil capZ. Rabbit skeletal
muscle actin (final concentration of 10 pM) was polymerized in the
presence of sample. After 2 hours at room temperature, the solutions
were drawn into glass capillary tubes. Relative viscosities are reflected by

the speed of steel balls traversing the tube (angled at 15).

Lipid-Binding And Capping-Inhibition Studies

Phosphatidylcholine (PC), phosphatidylinositol (PI),
Phosphatidylserine (PS) were purchased from Sigma (St. Louis, MO) and
used without further purification. Phosphatidylinositol 4,5-bisphosphate
(PIP2) was obtained from Calbiochem.

Annexin VI Binding To Multilamellar Vesicles

For the annexin VI lipid binding studies PC and PS were treated as
previously described by Blackwood, et al. (1990). Partially purified
neutrophil annexin VI was incubated at 25 "C for 15 minutes with either PS
or PC in buffer P containing 0.2 mM EGTA and varying concentrations of
calcium. Each mixture was then centrifuged at 12,000 X g in a table top
microfuge for 15 min at 25 C. Supernatants and pellets were subjected to
SDS-PAGE and stained with Coomassie brilliant blue. The relative amounts
of annexin VI in the pellets and supernatants were determined by laser

Phospholipid Column
A 3.5 ml sepharose-4B phospholipid column composed of PS/PC/PE
(2:2:1) was built in attempts to separate annexin VI and L-plastin, the two
predominant proteins copurified with the PMN actin polymerization
inhibitor after gel-filtration chromatography. The phospholipid column,
which exploits the ability of annexin proteins to bind acidic phospholipids,
was constructed according to the methods of Meers and colleagues (1987).

Modulation Of Capping Activity By Phosphoinositides
The ability of various lipids to block capZ filament end capping was
also examined. The phospholipids PIP2, PI, PC and PS were treated as

previously described (Janmey & Stossel, 1989). The modulation of capZ
actin binding function was assessed using the depolymerization assay (see
above). CapZ (36 nM final concentration) was first added to
polymerization buffer, followed by the addition of phospholipids from
previously sonicated stock solutions and resonicated in a water bath

sonicator at room temperature for 15 sec. After 1 min to allow the lipid-

capZ mixture to equilibrate, pyrene labeled F-actin was added (25 pl of 2
pM F-actin to a final volume of 1 ml). The inhibition of capZ capping was
assessed by measuring the initial slopes of depolymerization in the
presence of varying amounts of PIP2.

Promyelocyte Differentiation Studies

RNA Isolation And Northern Blots

Total RNA was extracted by a guanidinium thiocyanate (GITC)
method (Chomczynski & Sacchi, 1987). The cells of interest were lysed in
a GITC buffer (4 M guanidinium isothiocyanate, 25 mM sodium acetate, pH
7.0, 0.5 % sarcosyl, and 0.1 M R-mercaptoethanol) with vigorous pipetting
and vortexing. Two milliliters (2 ml) of GITC buffer was added to 0.5- 1.0
X 107 cells. The following was sequentially added to the extract with

thorough mixing by inversion in between: 0.07 volume 3 M sodium
acetate pH 5.2, 1 volume cold water-saturated phenol, and 0.2 volume of
chloroform/ alcohol (49:1). The mixture was shaken vigorously for 10 sec
and kept on ice for 15 in and centrifuged in a Beckman JA-20 rotor at
10,000 X g for 20 min. The aqueous phase was carefully removed and
precipitated with 1:1 volume of cold for 1 hour to overnight at -20 C.
After another 10,000 X g spin, the RNA pellet was resuspended in 300 pI of
GITC buffer and precipitated another time with 2.5 volumes of cold 95 %
ethanol at -20 C. The final pellet was washed with 70 % ethanol, dried
and resuspended in 0.5 % SDS in diethylpyrocarbonate (DEPC) treated
water. To eliminate any contaminating proteins and to remove the SDS,
the samples were heated at 65 C for 15 min, kept on ice for 15 min (to

precipitate the SDS), and microfuged for 5 min. The clear supernatant was
transferred into a clean tube and the concentration and purity of the
sample was analyzed by a spectrophotometer.
The RNA samples were prepared for electrophoresis as follows. The
RNA (10 pg/sample) was completely lyophilized and resuspended in 10 pl
of a sample mix containing 250 pl deionized formamide, 90 p1 37 9
formaldehyde, 26 pl 10 X MOPS (3-[N-morpholino]propanesulfonic acid)
buffer (1 X MOPS buffer = 0.2 M MOPS, 0.05 M sodium acetate, 0.01 M
Na2EDTA, pH 7.0), 10 p1 of ethidium bromide (10 mg/ml stock solution)
and 30 p1 of loading dye. The samples were heated at 67 'C for 15 min,
cooled on ice and were electrophoresed in 1.2 % agarose gels with 2.2 M
formaldehyde in 1 X MOPS buffer. The gel was transferred to
nitrocellulose membrane in 20 X SSC (IX SSC = 0.15 M NaCI and 0.015 M
sodium citrate) by standard capillary blotting. The blots were hybridized
with random-primed probes (1 X 106 cpm/ml) at 42 C in 5 X Denhardt's
solution, 5 X SSPE (IX SSPE = 0.15 M NaCI, 0.01 M NaH2PO4-H20, and

0.001 M EDTA), 50 % formamide (Fluka), 0.1 % SDS, and 150 pg/ml
salmon sperm DNA. After a room temperature rinse in 2 X SSC, 0.1 % SDS,
the filter was washed at 65 'C in 2 X SSC, 0.1 % SDS for 30 min followed by
a brief wash in 0.2 X SSC, 0.1 % SDS until the background was low by a
Geiger counter survey.

Generation Of Annexin VI cDNA Probe
Total RNA was isolated from human neutrophils as described above.
Using the published cDNA sequence for human annexin VI, upstream and
downstream oligonucleotide (25 base-pairs each) primers were
synthesized (I.C.B.R. DNA synthesis laboratory, University of Florida). The

Perkin-Elmer Cetus R.T.-P.C.R. kit was utilized to generate a 1057 base-pair

(bp) annexin VI DNA fragment from neutrophil total RNA via a reverse
transcription reaction immediately followed by initiation of the
polymerase chain reaction (Perkin-Elmer Cetus DNA Thermal Cycler model
480). The ends of the 1057 bp DNA fragment were blunt-ended with the
Klenow fragment of DNA polymerase, and ligated into the E. coli vector
pBluescript II SK + (Stratagene Cloning Systems, Lajolla, CA) at the Sma I
polylinker site. The 1057 bp, subcloned, annexin VI-cDNA-fragment (bp #
32 1088) was verified by sequencing (I.C.B.R. DNA sequencing facility,
University of Florida).

Cell Culture And Induction Of Differentiation

The human promyeloblast leukemic cell line HL-60 (Gallagher et al.,
1979) was obtained from ATCC (American Type Culture Collection,
Rockville, MD) and maintained in RPMI 1640 medium supplemented with
10% iron-supplemented calf serum (Gibco BRL, Gaithersburg, MD) and 2
mM glutamine. This cell line can be induced to differentiate towards the
neutrophil lineage or macrophage lineage depending upon the
pharmacologic exposure (Collins, 1987). Differentiation of HL-60 cells was
induced according to the methods previously described by Dabiri et al.
(1992) Briefly, exponentially growing HL-60 cells (> 1 X 106 cells/ml)
were plated at 0.5 X 106 cells/ml and treated at time zero with the

appropriate pharmacologic agents. For macrophage differentiation, HL-60
cells were treated with phorbal 12-myristate-13 acetate (PMA, Sigma) at 10
nM (6.2 ng/ml), and for granulocyte differentiation the HL-60 cells were
treated with 1.5 % dimethyl sulfoxide (DMSO, Sigma Chemical Company,
Saint Louis, MO). At appropriate time points after differentiation, RNA or

cellular protein was isolated and analyzed by Northern or Western blots as

described earlier. For Western studies, the total protein concentrations
were calculated on the last day of the experiment using the same standard
curve. Samples from earlier time points were rapidly collected (saving an
aliquot for determination of protein concentration), placed in gel sample
buffer, and stored at -70' C until the final sample was processed.

Fluorescence Microscopy

Human neutrophils and monocytes were isolated from peripheral
blood by sedimentation through PolymorphprepTM, a sodium-metrizoate,
dextran 500 solution (Nycomed Pharma As, distributed by Gibco BRL,
Gaithersburg, MD). The cells were resuspended in Hank's balanced salt
solution (Gibco BRL, Gaithersburg, MD), and allowed to adhere to glass
coverslips for 15 20 min in a 37 "C C02 incubator. The adherent cells

were treated for immunofluorescence by fixation for 20 min at room
temperature in 3 % formaldehyde (prepared from paraformaldehyde)
made in standard salt (0.1 M KC1, 0.01 M KPO4 buffer, 0.001 M MgCl2, pH

7.0), permeabilization in 0.1 % Triton-X 100 in standard salt, 5 min
overlay with 50 mM ammonium chloride in standard salt, and rinses with
standard salt between steps (Sanger et al., 1980). The following reactions
were carried out in a moist chamber in a 37 "C C02 incubator, the cells

were then blocked with 10 % goat serum in standard salt for 30 min,
followed by 45 min incubation with monoclonal antibodies (0.01 Jg/pl
anti-human annexin VI, control antibodies were a kind gift from Maurice
Swanson, Univ. of Fla., College of Medicine: anti-human hnRNP-M protein
& anti-yeast polyadenylated RNA-binding protein, Nab-1) (Datar et al.,

1993; Wilson et al., 1994) and used at concentrations greater than the

anti-annexin VI as determined by immunoblots of the antibodies run on
SDS-PAGE, rinses in standard salt, incubation with fluorescein
isothiocyanate (FITC)-conjugated goat anti-mouse IgG1 (heavy chain
specific; Fisher Scientific, Pittsburgh, PA) for 30 min, followed with rinses
in standard salt. The coverslips were mounted in glycerol containing an
antioxidant to prevent fluorescence quenching (90 % glycerol, 10 % 0.25
M Tris pH 8.0, 2.3 % 1,4 1,4-diazabicyclo-[22.2.2] octane obtained from
Sigma). The cells were observed and photographed using a Nikon Diaphot
inverted microscope (Nikon, Tokyo, Japan) set up for epifluorescence.
Photographs were taken through the camera port using Tri-X Pan film
(Kodak). The film exposure lengths and print development times were
kept identical to photographs of PMN stained with anti-annexin VI.
Confocal microscopy was conducted with the assistance of Michael Paddy
(1.C.B.R. Confocal microscopy facility, University of Florida).
Approximately 30 images 1 pm thick were taken with a z-increment of 0.3
pm. Digitized images were obtained using Dr. Paddy's microscope and
processed with an Image-1 video image analyzer (Universal Imaging Corp.,
West Chester, PA), and finally stored on a Panasonic laser disk recorder.



The viscosity lowering activity of PMN actin polymerization

inhibitor, as originally published, was thought to be composed of two
related polypeptides of 62,000 and 65,000 daltons (Southwick & Stossel,

1981). The polypeptides purified with this activity, although, were not
identified. This chapter outlines the efforts to identify the PMN inhibitor,
and my eventual discovery that the ~65 kDa, heterodimeric-protein capZ

is solely responsible for the actin viscosity-lowering activity.


Identification Of The Two Maior Polypeptides (~ 66 And 68 kDa)
Copurified With The Neutrophil Actin Polymerization Inhibitor

The calcium independent activity called human neutrophil actin
polymerization inhibitor was purified using ion exchange chromatography
and gel filtration (see methods) (Figure 3-1, lane 2). Yields and purity
were identical to the previously reported purifications (Southwick &
Stossel, 1981). The two polypeptides copurified with the inhibitor were
noted to have Mr of 66,000 and 68,000 when fractionated through a

discontinuous pH, 10 % SDS-PAGE. Similar to previous reports, highly

Figure 3-1. PMN actin polymerization inhibitor. The PMN inhibitor
purified by DEAE-anion exchange and gel-filtration chromatography. Lane
2 represents the PMN inhibitor subjected to 10 % SDS-PAGE with standards
(lane 1; Mr = X 103) and stained with Coomassie blue. Fractions were
subjected to Western analysis and probed with antibodies to annexin VI
(lane 3) or L-plastin (lane 4).






3 4

1 2

purified fractions (> 90 % composed of the 66 and 68 kDa polypeptides as
assessed by densitometry scanning of Coomassie blue stained SDS-PAGE)

lowered the viscosity of actin filament solutions as measured by falling-
ball microviscometry. The minor proteins, if present, (usually < 10 %)
were composed of a 42 kDa polypeptide believed to be actin (rabbit
skeletal muscle actin comigrates with this band) and a 55 kDa polypeptide
later found to be an L-plastin degradation product (see below). Active
fractions, composed predominantly of the 66 and 68 kDa polypeptides,
were separated on SDS-PAGE, transferred onto polyvinylidene difluoride
(PVDF), stained and sent to our collaborator Dr. David W. Speicher (The
Wistar Institute, Philadelphia, PA) for amino-acid sequence analysis. The
original analysis performed after BNPS-Skatole digestion (the amino-
termini of the polypeptides were blocked) resulted in sequence
information solely from the 68 kDa polypeptide. As shown in Figure 3-2A,
two fragments from the 68 kDa polypeptide yielded sequences of 18 and
23 amino acids which were identical to amino acids 193 210 and 344 -
366 respectively of human annexin VI (Crompton et al., 1988; Sudhof et
al., 1988). Immunoblots probed with a monoclonal antibody to human
placental annexin VI (Zymed Laboratories, Inc., South San Francisco, CA)
only revealed the 68 kDa polypeptide (Figure 3-1, lane 3). Polyclonal
antibodies to human annexin VI (rabbit antisera was a kind gift from Dr.
Joel D. Ernst, University of California, San Francisco) detected the 68 kDa
annexin VI, but also failed to recognize the 66 kDa polypeptide (data not

Amino acid sequence analysis of the 66 and 68 kDa polypeptides
containing neutrophil activity was repeated. Samples were prepared as
described above, and sent to The Wistar Institute (Protein

Figure 3-2. Amino-terminal sequence analysis of PMN actin
polymerization inhibitor. Since the amino termini were blocked,
microsequence analysis was performed on the predominant cleavage
products of the upper peptide (Skatole) or lower peptide (Trypsin)
component. The 68 kDa upper band is identical to annexin VI and the 66
kDa peptide is identical to L-plastin.


188 213


339 369


62 = 79


512 534


Microsequencing Facility). For this analysis, the two polypeptide bands
were digested with trypsin. Comparative HPLC peptide maps (post-trypsin
digestion) with a trypsin control demonstrated that almost all major peaks
in the 68 kDa sample were different from those generated from the 66 kDa
sample. The masses of the major trypsin cleavage peptides were
determined by MALDI mass spectrometry. Most major peak masses from
the 68 kDa sample agreed with expected masses of limit tryptic peptides of
annexin VI. In contrast, only a few major peaks in the 66 kDa sample
could possibly fit with annexin peptides. Two of the major peptides
generated by trypsin digestion of the 66 kDa band were sequenced, and
revealed a perfect match with the protein L-plastin. As shown in Figure 3-
2B, N-terminal sequence analysis of these two major peptides yielded
sequences of 9 and 15 amino acids which were identical to amino acids 67
- 75 and 516 530 respectively of human L-plastin (Lin et al., 1988). To
further confirm the identity of the 66 kDa polypeptide western blot
analysis was performed using a mouse monoclonal anti-human L-plastin (a
kind gift from Dr. Yuziro Namba, Kyoto University, Japan). As shown in
Figure 3-1 (lane 4), the antibody specifically cross-reacts with the lower,
66 kDa polypeptide. The antibody also cross-reacts with a polypeptide
with approximate Mr of 55,000. This band is felt to be a degradation

product of the neutrophil L-plastin, as it is absent in fresh PMN extract and

increases with storage of the PMN inhibitor fractions (data not shown). A
similar molecular weight L-plastin degradation product has been
previously reported (Pacaud & Derancourt, 1993).

Separation Of Annexin VI From L-Plastin

The following experiments were conducted in attempts to correlate
the neutrophil actin polymerization inhibitory activity with one of the two
proteins or a complex formed between them. Additional chromatography
steps were added in attempts to separate annexin VI from L-plastin.
Neither phenylsepharose (Pharmacia Biotech, Inc., Piscataway, NJ) or Affi-
Gel blue (Bio-Rad Laboratories, Inc., Hercules, CA) chromatography
separated the two proteins. The viscosity lowering activity was noted to
coelute with the annexin VI and L-plastin proteins during Affi-Gel blue
chromatography (the viscosity lowering activity was not monitored in the
fractions obtained from phenylsepharose chromatography).
The difficulty in separating these two proteins raised the possibility
that annexin VI and L-plastin were forming a complex. Nondenaturing
polyacrylamide gel electrophoresis was used to test this possibility (Figure
3-3). Active fractions containing both annexin VI and L-plastin (Figure 3-
3, lane 2), purified annexin VI (lane 1) (via liposome purification, see
below), recombinant human L-plastin (lane 3) (purified from E. coll by Dr.
Clarence Young for unrelated studies in our laboratory), and bovine
albumin (lane *) (indicator of the nondenaturing conditions of the assay)
were each subjected to nondenaturing electrophoresis. If a complex was
formed by annexin VI and L-plastin in active fractions, one might expect a
migration pattern different from lanes containing the purified annexin VI
(Figure 3-3, lane 1) or purified L-plastin (Figure 3-3, lane 3). This
difference would most likely appear as a slower migrating band in
fractions where a complex is formed (analogous to the decreasing
mobilities of monomer, dimer, and trimer species of native bovine
albumin seen in Figure 3-3A lane *). The samples were run in triplicate

Figure 3-3. Nondenaturing polyacrylamide gel electrophoresis. PMN inhibitor (lane 2; 12pg) was
subjected to native gel electrophoresis in search of evidence suggesting complex formation between L-
plastin & annexin VI. Purified annexin VI (lane 1; 6 pg) and purified L-plastin (lane 2; 6 pg) were run to
enable the determination of a mobility shift when the two proteins are present. Panel A is a Coomassie
stained gel of the various fractions (starred lane contains 15 pg of albumin/BSA). Gels were run in
triplicate for western analysis with antibodies to annexin VI (Panel B) or L-plastin (Panel C).


* 1 2 3 1 2 3 1 2 3

for Coomassie blue staining and immunoblot analysis. As shown in Figure

3-3B, antibodies to annexin VI (Zymed Laboratories, Inc., South San
Francisco, CA) reveal one band in the active fraction. The similar
migration pattern shared by annexin VI in purified and active fractions

(Figure 3-3B, lane 1 and lane 2), suggests annexin VI does not take part in
complex formation. Although the L-plastin antibody reveals two closely
spaced reactive bands in the active fraction (Figure 3-3C, lane 2), they
both migrate separately from annexin VI. The detection of two L-plastin
bands in active fractions (Figure 3-3C, lane 2) and only one in the purified
(recombinant L-plastin) fractions (lane 3) most likely results from post-
translational modification of the PMN isolated L-plastin. Furthermore, L-
plastin is a known phosphoprotein (Zu et al., 1990). It is well established
that bacteria are incapable of many of the post-translational modifications
seen in eukaryotes, and thus could explain the appearance of one band
with L-plastin recombinantly expressed and purified from E. coli. It is
notable that recombinant L-plastin does not decrease the viscosity of actin
solutions (Dr. C. Young, unpublished data). These data derived from
nondenaturing polyacrylamide gel-electrophoresis suggest no complex is
formed between the annexin VI and L-plastin present in the neutrophil
actin polymerization inhibitory activity.

To further clarify which protein was responsible for the actin
viscosity lowering activity, attempts were made to separate annexin VI by
methods previously developed for studying the membrane binding
properties of annexins (Blackwood & Ernst, 1990). In this assay,
phosphatidylserine liposomes were prepared as described, and added to
active fractions. After incubation in the presence of calcium, the
liposomes were separated from the fractions by centrifugation. This

Figure 3-4. SDS-PAGE of phospholipid affinity chromatography, and anti-
annexin VI western analysis. PMN inhibitor was exposed to a mixed-lipid-
vesicle column in the presence of calcium. The activity remained in the
fall through (lane 2). Annexin VI (> 95 % pure) was eluted with 1mM
excess EGTA. Fractions were transferred to nitrocellulose and probed with
antibodies to annexin VI. The same molecular weight standards in Figure
3-1 were used (HMW Std).

- OWi

* 1 2

1 2

w -

method was able to remove > 60 % of annexin VI from those fractions

purified by DEAE-ion exchange and gel-filtration chromatography.
Phospholipid affinity chromatography (Chapter 2), as well, incompletely
separated annexin VI from the neutrophil activity (Figure 3-4, lane 2).
The neutrophil actin polymerization inhibitory activity remained in
fractions partially depleted of annexin VI and predominantly composed of
L-plastin, and was not associated with the highly purified annexin VI (> 90
%) (Figure 3-4, lane 1).

Purification And Identification Of PMN Actin Polymerization Inhibitor As
In efforts to avoid phospholipid contamination and the addition of
calcium, other chromatographic methods were tested. After DEAE-anion
exchange chromatography and gel-filtration, the neutrophil inhibitory
activity was applied to a Mono Qanion exchange column and eluted with a
KC1 gradient (Chapter 2). Surprisingly, the activity coeluted with annexin
VI and not L-plastin (data not shown). The peak activity, although,
correlated with a fraction in which no protein was seen by Coomassie
staining of the fractions run on SDS-PAGE (no predominant band was seen
by silver staining as well). However, Western blots of this and three other
active fractions from previous purifications probed with antisera from a
goat immunized with chicken, skeletal-muscle capZ (kindly provided by
Dr. J. Cooper, Washington University, St. Louis, MO) revealed two
immunoreactive bands (Figure 3-5) in the appropriate molecular weight
range of the heterodimeric, skeletal-muscle protein capZ (Mr between

30,000 40,000). The DEAE-anion exchange fractions from the
subsequent purification were subjected to immunoblot analysis and
probed with the capZ antisera as well (Figure 3-6). Peak capZ

Figure 3-5. Western blot analysis of PMN inhibitory fractions probed with
capZ antisera. Fractions from the four most recently purified PMN actin
polymerization inhibitor purified by DEAE-ion exchange and gel-filtration
chromatography each had detectable levels of both subunits (a and p) of


1 2 34

immunoreactive peptides (Figure 3-6C fraction # 36 42) were found to
correlate with the neutrophil actin polymerization inhibitory activity. As
seen in Figure 3-6, the viscosity lowering activity was maximal in fractions
36 42 (panel C). These fractions also represented the peak capZ, annexin
VI (68 kDa), and L-plastin (66 kDa) fractions.
A scheme utilizing three additional columns was empirically devised
to separate annexin VI and L-plastin and further purify the neutrophil
actin polymerization inhibitor (High S, Mono Q, and hydroxylapatite).
Active fractions composed predominantly of annexin VI and L-plastin were
further separated by High S-cation exchange chromatography (Figure 3-7)
under similar conditions utilized during the purification of skeletal muscle
capZ. Figure 3-7A represents the typical separation achieved by this
method. The polymerization inhibitory activity correlates with L-plastin
(fraction # 28 42). Fractions # 62 86 are composed of annexin VI
(Figure 3-7B) as detected by Western analysis. Similar to the
phospholipid-purified annexin VI, no activity was detected for fractions of
annexin VI separated by High S-ion exchange chromatography (Figure 3-
7C, fraction # 62 70).
Active fractions after High S chromatography were applied to a
Mono Qanion exchange column and eluted with KC1 (Figure 3-8A) as
described earlier (Chapter 2). Despite the presence of PMSF in dialysis
solutions and maintenance of all procedures at 4 C, the 55 kDa L-plastin
degradation product was formed in significant amounts by this stage of
the purification (Figure 3-8C, fraction # 28 34: lower band). Mono Q
chromatography separated the peak L-plastin fractions (Figure 3-8C
fraction # 29 32), but not its degradation product (see fraction # 26 28)
from the neutrophil actin polymerization inhibitor. The purification was

Figure 3-6. DEAE-anion exchange chromatography of PMN extract.
Extracts were eluted with a linear KCI gradient as described (Chapter 2).
Eluted proteins were subjected to 10 % SDS-PAGE (with HMW stds) and
stained with Coomassie (Panel A) or transferred to nitrocellulose and
probed with antisera to chicken skeletal-muscle capZ (Panel B). Relative
activities (Panel C) were reported as the ability of fractions to decrease the
viscosity of actin solutions (10 pM) as measured by falling ball



24 32 36 40
n 30 34 38 42

-. :4'

Fraction #


44 52

-n ,3 3i
11 i::;

Figure 3-7. High S-cation exchange chromatography of PMN actin polymerization inhibitor. PMN
inhibitor consisting predominantly of annexin VI and L-plastin was bound to a High S column (Chapter 2)
and eluted with a 0 0.15 M linear KC1 gradient. Fractions were subjected to 10 % SDS-PAGE (with HMW
stds) and stained with Coomassie (Panel A). Fractions were transferred to nitrocellulose and probed with
antibodies to annexin VI (Panel B). Relative activities are reported as the ability to prevent actin
polymerization from actin filament nuclei isolated from red blood cells (Panel C).


Anti-Annexin VI

62 66 70

Std 16 24 30 34 38 42 50 58 66 74 86
12 20 28 32 37 40 46 54 62 70 78 82 I


Fraction i#

Figure 3-8. Mono Q-anion exchange chromatography of PMN actin
polymerization inhibitor. Peak activities from High S chromatography
were pooled and allowed to adhere to a Mono Q column. The proteins
were eluted with a 0.1 0.3 M linear KC1 gradient, and subjected to 10 %
SDS-PAGE (with HMW stds). The 10 % SDS-PAGE gel shown was first
stained with Coomassie blue (Panel A), and then silver stained (Panel A').
The relative activities (Panel B) were measured as in Figure 3-7. Fractions
were also transferred to nitrocellulose and probed with antibodies to L-
plastin (Panel C).

14 18 22 24 26 28 30 34 38 Std 14 18 22 24 26 28 30 34 38 Std
12 16 20 23 25 27 29 32 36 40 12 16 20 23 25 27 29 32 36 40

C: 16 1 8 20 22 23 24 25 26 27 28 29 30 32 34 36 8


26 27 28 29 30 31 32

carried out to this extent on three separate occasions, each resulting in the
isolation of polypeptides with approximate Mr of 55,000 (L-plastin

degradation product), 36,000, and 32,000 seen upon silver staining (for
examples see Figure 3-8A' lanes 26 and 27 or Figure 3-11, lane 3).
Western analysis identified the 36 and 32 kDa bands as capZ (Figure
3-9A & B, lane 2). As shown in Figure 3-8 (panel A verses panel A', lanes
26 & 27), the presence of these peptides after Mono Qchromatography is
revealed by silver staining. Hydroxylapatite chromatography of these
fractions successfully separated the L-plastin degradation product from
the actin-filament-shortening activity (Figure 3-10, fraction 21 verses
fraction 22). The activity was strictly associated with capZ (Figure 3-10).
Although several minor bands were seen below the 36 and 32 kDa
polypeptides in Figure 3-10, they were not seen when fraction 21 was
separated on other occasions (Figure 3-11, lane 4). Several possibilities
including degradation, resolution of isomers or an artifact of the
procedure can explain this finding.


Proteins that bind to actin and modulate actin filament assembly are
also hypothesized to regulate the finely coordinated assembly and
disassembly of the microfilament network in motile cells. Based on
viscometric studies, actin polymerization inhibitor is thought to play an
important role in the regulation of actin filament length in PMN.
Originally, my purifications focused on the 62 and 65 kDa proteins, and so
the identity of these polypeptides was pursued.

Figure 3-9. Western blot analysis of PMN actin polymerization inhibitor
purified to Mono Q chromatography. Peak fractions from Mono Q
chromatography (lane 2) (see Figure 3-8) were separated by 10 % SDS-
PAGE (with HMW stds), and either stained (Panel A) with Coomassie or
subjected to Western analysis (Panel B) probing with antisera to chicken
skeletal-muscle capZ. Lane 1 is 30 pg PMN extract, while a total of 3 pg of
protein was loaded in lane 2.

1 2

1 2



Figure 3-10. Hydroxylapatite column chromatography. As a final
purification step, peak active fractions from Mono Qchromatography were
subjected to hydroxylapatite (HA) column chromatography as described
(Chapter 2). Fractions were subjected to 10 % SDS-PAGE and silver stained
(Panel A). Relative activities were measured as in Figure 3-7.

17 18 19 20 21 22 23 24 25





Fraction #

.r", ..; ..p .

Figure 3-11. Silver stained peak fractions from Mono Q and HA
chromatography. PMN extract (14 pg; lane 2), Mono Qpeak activity (2pg;
lane 3), HA peak activity (0.25 pg; lane 4), and HMW std (lane 1) were
subjected to 10 % SDS-PAGE and silver stained. Arrowheads point to the
36 kDa and 32 kDa subunits of capZ.

1 2 3

The molecular weights of the polypeptides copurified with the
viscosity lowering activity were currently found to be 66,000 and 68,000
when separated by 10 % SDS-PAGE (originally determined with 5 15 %
SDS-PAGE). Initial sequence analysis of the polypeptides resulted solely in
the identification of the 68 kDa upper band as annexin VI. Their shared
physical properties (Mr and pl) suggested the 66 kDa polypeptide was

related. It was speculated that the 68 kDa protein was susceptible to
known neutral proteolytic activity in human granulocytes (Southwick &
Stossel, 1981).
Despite the recognition of only the upper band by anti-annexin VI
immunoblot analysis, it was thought that the epitope recognized by this
monoclonal antibody (Zymed Laboratories, Inc., South San Francisco, CA)
was not present in the 66 kDa polypeptide. The annexins are a large
family of related proteins (Chapter 5). The structural similarity amongst
the annexin proteins limits the number of epitopes specific to each
member. The annexins have divergent sequences at their amino termini
(Creutz, 1992), and therefore it is likely that the monoclonal anti-annexin
VI-specific antibody recognizes an epitope in this region. It seemed
plausible that the 66 kDa polypeptide was a partially degraded form of
annexin VI that had lost the epitope recognized by the monoclonal
antibody. Alternative explanations included the possibility the 66 kDa
peptide was an alternatively spliced annexin VI lacking the epitope
(Crompton et al., 1988), the 66 kDa peptide was annexin VI post-
translationally modified in a manner preventing recognition of the
epitope, or the 66 kDa peptide was an unrelated peptide. To test these
possible explanations, the peptides were probed with a polyclonal antisera
specific to human annexin VI (Dr. Joel Ernst). Similar to the monoclonal

antibody to annexin VI, the polyclonal antisera only recognized the 68 kDa

polypeptide. These findings suggested that the 66 kDa protein was not a
degradation product of annexin VI.
To determine the identity of this protein, active fractions were again
prepared for amino acid sequence determination as described earlier (see
Results), and sent to our collaborator D. W. Speicher at The Wistar
Institute. Utilizing improved techniques, the 68 kDa band was once more
identified as annexin VI while the 66 kDa band was identified as L-plastin.
Based on previously published cDNA sequences, the predicted molecular
weights of annexin VI and L-plastin are 75,901 and 70,306 respectively
(Sudhof et al., 1988; Lin et al., 1988).
There have been no prior reports of annexin VI and L-plastin
copurified, probably a reflection of the abundance of both proteins in
neutrophils and the tissue specific expression of L-plastin (limited to
transformed and hematopoetic cells) (Lin et al. 1988). I was unable to
detect a complex formation between these proteins as evidenced by the
inability of annexin VI and L-plastin antibodies to recognize a common
band in active fractions resolved by native acrylamide gel electrophoresis.
Biochemical analysis reported by several laboratories repeatedly
demonstrates Mr between 64,000 70,000 and pi between 5.3 and 5.6 for

both proteins. The shared physical properties explains why we
consistently copurified these two proteins.

Earlier reports have implicated both annexin VI and L-plastin as
cytoskeletal regulatory proteins. They therefore remained likely
candidates responsible for the activity of neutrophil actin polymerization
inhibitor. Annexin VI isolated from bovine hepatic tissue was recently
observed to cosediment with actin filaments in a calcium dependent (>

100 pM) fashion (Hosoya et al., 1992). L-plastin has been shown to bundle
actin filaments in a calcium dependent manner (Namba et al., 1992). This

group was able to demonstrate by electron microscopy that L-plastin
induces actin bundle formation in vitro in the presence of low free calcium
concentrations (10 100 nM), and this property is lost as the calcium
concentration is elevated (1 10 pM). A separate group (Pacaud et al.,
1993) was able to demonstrate that L-plastin increases the viscosity of
actin solutions at low free calcium concentrations (< 0.1 pM), but the
viscosity of L-plastin/actin solutions was progressively lowered to control
levels actinn polymerized alone) as the free calcium concentration was
elevated (> 10 pM). The reported interactions of annexin VI or L-plastin
with actin, although, could not readily explain the calcium-independent
viscosity-lowering effect of the neutrophil actin polymerization inhibitor.
Annexin VI purified by either the phospholipid affinity column or
the High S cation exchange chromatography did not possess the
neutrophil inhibitory activity. Additionally, L-plastin could also be
separated from the neutrophil actin polymerization inhibitor (see Results)

by Mono Q-anion exchange chromatography. The ability to separate both
proteins from the activity proves the activity of PMN actin polymerization
inhibitor is not due to annexin VI or L-plastin.
It was subsequently discovered that capZ was responsible for the
inhibitory activity. Utilizing three additional chromatography columns,
the neutrophil actin polymerization inhibitor was successfully separated
from annexin VI and L-plastin (Figures 3-7, 3-8, and 3-9). Silver stained
polyacrylamide gels of the various purification steps reveal that active
fractions consistently correlate with the presence of the 36 and 32 kDa
capZ bands. My purification of the activity to these bands, and their

recognition by antibodies to capZ proves this protein is responsible for the
PMN actin polymerization inhibitor.
CapZ is a heterodimeric protein with subunits of Mr 36,000 (a-

subunit) and 32,000 (P-subunit) (Casella et al., 1986). Members of this
family are capable of binding the barbed end of actin filaments with high
affinity (~ 0.5 10 nM) independently of calcium. Binding to the barbed
end of actin filaments effectively decreases the apparent viscosity of actin
solutions by two mechanisms. First, this interaction results in the
shortening of the average filament length, and secondly it raises the
critical concentration to that of the pointed end. The elucidation that
PMN actin polymerization inhibitor is identical to capZ was complicated by
several factors. The limited number of human PMN available for
purification, capZ's high affinity interaction with actin, its decreased
sensitivity to Coomassie staining, the predominance of annexin VI and L-
plastin in active fractions, and most notably the physical properties shared
by these three proteins each added to my difficulties in identifying capZ as
the PMN inhibitor.

The use of the human neutrophil system greatly limits the starting
material. These immune cells have developed an effective antimicrobial
repertoire which includes a diverse collection of proteolytic molecules. A
packed cell volume of ~ 20 25 ml (derived from > 20 liters of peripheral
blood) was isolated during the largest preparations. With this amount of
cells, less than 1 gram of protein remains after clarification of the
homogenized cells. This is in stark contrast to the two previous published
purifications of capZ from animal cells (chicken skeletal muscle and
bovine brain). With the purification from chicken skeletal muscle, 1000 g
of chicken breast muscle are typically utilized to purify ~1 mg capZ

(Casella et al., 1986, Caldwell et al., 1989). Purification of capZ from
bovine brain similarly requires ~ 1000 g starting material (4 brains)
(Kilimann and Isenberg, 1982). Our final yield of PMN capZ after
hydroxylapatite chromatography (-10 pg) was about one-tenth of those in
chicken muscle capZ preparations. This finding is not surprising
considering the abundant proteolytic activity commonly seen in extracts
made from these phagocytes. Secondly, the ability of capZ to interact with
actin with such high affinity (> 250 ng/ml, based on Kd ~3 nM) (see

Chapter 4) allowed for detection of activity with minimal protein as seen
by stained polyacrylamide gels. The problem was further complicated by
the identification of the two major polypeptides copurified with the
activity as previously studied actin binding proteins.
Lastly, the elusiveness of capZ resulted primarily from its physical
properties and staining characteristics. The consistent purification of
annexin VI and L-plastin with capZ is currently explained by the similar
native molecular weights and isoelectric points shared by these three
proteins (annexin VI/L-plastin: Mr 64,000 70,000 and pi 5.3 5.6 for

both proteins). The molecular weights of bovine brain and chicken
skeletal muscle capZ were reported as 63,000 (pl 5.75) and 61,000
respectively (Kilimann and Isenberg, 1982; Casella et al., 1986). A Stokes
radius of 32 A was determined for the PMN actin polymerization inhibitor
by analytical gel filtration following the viscosity lowering activity
(Southwick & Stossel, 1981). This closely resembles the Stokes radii
determined for bovine brain (35.5 A) and chicken skeletal muscle (37 A)
(Kilimann and Isenberg, 1982; Casella et al., 1986). Many
chromatographic techniques rely on variations of MW or pi to separate
proteins. The physical properties shared by these three proteins

necessitates the use of several columns to purify the neutrophil inhibitory

In summary, the exhaustive purification protocols have
demonstrated human neutrophil actin polymerization inhibitor is capZ, a
heterodimeric protein consisting of two subunits with Mr of 36,000 and

32,000 as detected by silver staining active fractions separated by SDS-

PAGE. Western blot analysis of peak inhibitory activity from Mono Q
chromatography fractions reveals immunoreactivity of the 36 and 32 kDa

polypeptides with antisera made to chicken skeletal-muscle capZ,
supporting my conclusion that the 36 and 32 kDa polypeptides further
purified by hydroxylapatite were identical to the capZ immunoreactive

peptides. Together these findings presented in Chapter 3 indicate that the

previously published neutrophil actin polymerization inhibitor is indeed
related, if not identical, to the skeletal muscle form of capZ. In addition,

neither of the polypeptides originally identified as the neutrophil inhibitor
(currently identified in this work as annexin VI and L-plastin) contributes

to the activity of PMN actin polymerization inhibitor.



The PMN actin polymerization inhibitor was originally isolated in
the presence of EGTA primarily to distinguish it from the calcium-
dependent viscosity-lowering activity of gelsolin (Yin & Stossel, 1979).
The importance of this characteristic is suggested by the recent findings of

calcium-independent actin assembly in PMN (reviewed in Chapter 1).
Based on the current theories of actin regulatory proteins (Pollard &
Cooper, 1986), several interactions with actin could explain the actin
viscosity lowering effect originally reported for PMN actin polymerization

inhibitor (Southwick & Stossel, 1981). Proteins that sequester actin
molecules away from the polymerizable pool would lower the viscosity of

actin solutions. Additionally, interactions which shorten the average

length of actin filaments such as severing or capping (barbed or pointed
end) would also lead to a relative decrease in the final viscosity.

Conversely, actin binding proteins that cross-link or bundle actin

filaments would be expected to raise the viscosity of actin solutions. The
identification of the neutrophil activity as capZ suggests the viscosity

lowering effect results from the ability of this protein to cap the barbed-
end of actin filaments.

The characterization of the specific interactions of proteins with
actin was at first limited to those expert in the techniques of electron
microscopy. By this method the lengths of filaments can be directly
visualized, the kinetics of assembly at each end of the actin filament can

be simultaneously analyzed (using morphologically identifiable nuclei
such as heavy-meromyosin-labeled actin oligomers or Limulus sperm
acrosomal actin bundles), and the three-dimensional configuration of
actin solutions can be monitored. With the development of fluorescent
labels covalently attached to actin, most notably pyrene-labeled actin,
spectrophometric analysis of actin polymerization has become an
alternative, more accessible technique to study specific interactions and
kinetics of actin filament assembly (Kouyama & Mihashi, 1981; Cooper et
al., 1983). In efforts to further characterize the viscosity lowering activity
of PMN capZ, its effects on actin polymerization were studied utilizing
pyrene actin.


Barbed-End Capping Activity

As shown in Figure 4-1A, capZ slowed the rate of actin filament
disassembly in a concentration dependent fashion. The marked slowing in
the depolymerization rate was most consistent with blocking of monomer
release from the barbed end of actin filaments. The apparent dissociation
constant for the interaction of capZ with the barbed end (Kd app) was

approximately 3 nM. CapZ inhibited actin depolymerization to a similar
extent in the presence as well as in the absence of Ca2+ (Figure 4-1B).
When filaments are diluted, they rapidly depolymerize from their ends in

Figure 4-1. Effects of purified capZ on actin filament depolymerization. Pyrene actin (2pM) was allowed
to polymerize to steady state in the presence of 1 mM EGTA (Panel A) or 1 mM CaC12 (Panel B). At time
zero, aliquots of the F-actin were diluted 1/40 (final concentration 50 nM) into varying concentrations of
purified neutrophil capZ in buffer P containing EGTA (Panel A) or 1 mM CaC12 (Panel B). Fluorescence
intensity was monitored over time. Numbers next to symbols represent the final concentrations of capZ.





A 27nM
A 13.lnM
* 0Control

0 E

0 10 20 30

Time (min)


O aM
A 18iM
0 9UMn
S* 3.6nM
D Control


0 10 20 30

Time (min)

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EL7XQ7RKC_D4W9PY INGEST_TIME 2012-12-07T21:04:49Z PACKAGE AA00012900_00001