This item is only available as the following downloads:
TISSUE-SPECIFIC REGULATION AND FURTHER CHARACTERIZATION OF
CYSTEINE-RICH INTESTINAL PROTEIN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
VEIRSy OF FLORIDA LIBRARIES
Dedicated to my husband, Mun-Yew Chee, with love.
I would like to acknowledge the support and advice of my advisor, Dr. R.J. Cousins,
whom I have worked with for both my master's and PhD programs. His knowledge and
experience have been very beneficial to me over the years. I would also like to extend my
gratitude to my committee members, Drs. Denslow, Percival, Shireman and Sitren, for their
willingness to contribute both their expertise and their moral support throughout my program.
I would like to thank Dr. Nora A. Hallquist who contributed greatly to a large part
of the studies (Chapters 2 and 3) carried out in this project. Her knowledge of immunology
added enormously to the success of the projects that we carried out together. I would like
to express my appreciation to Dr. Michael A. King and Dr. Don Samuelson for contributing
their time and expertise in the areas of brain anatomy and liver histology, respectively. I
would also like to acknowledge Dr. Raymond K. Blanchard's advice and assistance in the
studies described in Chapter 5.
Undying gratitude is extended to Virginia Mauldin, for her help and assistance at
every step of the way, and for listening to my quite frequent hysterical rampages.
My colleagues and fellow graduate students, past and present, Warren Clark, Pearl
Fernandes, Vicki Sullivan, Barbara Davis, Steve Davis, also deserve a real pat on the back
for tolerating my presence for so long. I would also like to thank them for always being
ready to offer assistance and advice as well as to pick up the occasional slack and to lend the
occasional shoulder. My deep gratitude to Ruth Davenport for tolerating my frequent phone
calls to inquire about the status of my sequences and for still answering my calls. Thanks
also to Karol Smith, another fellow student, for a lot of good times and many more to come,
I would like to thank my family in Malaysia, Singapore and Australia, and my aunt
and uncle in Maryland, for their love and continued encouragement. Again, last but not
least, I would like to thank, with all my heart, Mun-Yew Chee, now my husband of two
years, for being so incredibly tolerant, wonderful and supportive.
TABLE OF CONTENTS
ACKNOWLEDGMENTS ...................................... iii
ABBREVIATIONS ........................................ vii
ABSTRACT ............................................. viii
1 INTRODUCTION AND LITERATURE REVIEW ................. 1
Introduction ........................................... 1
O objectives ............................................ 6
Hypotheses ......................................... 6
Literature Review ....................................... 6
2 STRESS CHALLENGE (I) ............................... 28
Carbon-Tetrachloride Challenge ............................. 28
M materials and M ethods .................................. 28
Results ............................................. 31
D discussion ........................................... 44
3 STRESS CHALLENGE (II) ............................... 50
Lipopolysaccharide Challenge .............................. 50
M materials and M ethods .................................. 51
R results . . . . . 54
D discussion ........................................... 66
4 PROTEIN-PROTEIN INTERACTION ........................ 72
Introduction .......................................... 72
Strategy . . . . . 73
M materials and M ethods .................................. 75
Results ............................................. 80
D discussion ........................................... 95
5 HUMAN CYSTEINE-RICH INTESTINAL PROTEIN .............. 103
Background ......................................... 103
M materials and M ethods .................................. 104
Results ........................................... 107
D discussion ............... .... ..... ..... ..... ... ..... 112
6 SUMMARY, SPECULATIONS AND CONCLUSIONS ............. 114
Summary/Speculations .................................. 114
Conclusion ......................................... 128
A PROTEIN-PROTEIN INTERACTION FLOW CHART .............. 129
B [35S]CYSTEINE-rrCRIP PREPARATION ..................... 130
C PROPOSED MODELS FOR CRIP FUNCTION ........ .......... 131
LITERATURE CITED ......................................... 132
BIOGRAPHICAL SKETCH ................................... 154
antioxidant response element
bovine serum albumin
cell adhesion molecule
(human) cysteine-rich heart protein
(rat, human) cysteine-rich intestinal protein
recombinant rat cysteine-rich intestinal
(chicken, human) cysteine-rich protein
gamma butyric acid
immunoglobulin A or G
interleukin 1 or 6
manganese superoxide dismutase
metal response element
nerve growth factor
peripheral blood mononuclear cells
polymerase chain reaction
sodium dodecyl sulfate polyacrylamide gel
tumor necrosis factor
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
TISSUE-SPECIFIC REGULATION AND FURTHER CHARACTERIZATION OF
CYSTEINE-RICH INTESTINAL PROTEIN
Chairperson: Robert J. Cousins
Major Department: Food Science and Human Nutrition
Cysteine-rich intestinal protein (CRIP) is a developmentally-regulated LIM protein.
There are high levels of CRIP mRNA in intestine and immune cells, and lesser amounts in
other tissues. To examine CRIP regulation under altered physiologic status we used
lipopolysaccharide (LPS) as an immune challenge and carbon tetrachloride (CCl4)-induced
injury as a model of hepatic injury and inflammation.
The effect of CC14 administration was examined after 24 h. Since CRIP is a zinc-
finger protein, the effect of feeding a zinc-supplemented diet was also examined. The results
showed that the level of zinc used (300 mg/kg diet) did not affect plasma alanine
aminotransferase, an index of hepatic injury. CRIP mRNA and immuno-positive cells
increased in liver, but not in immune cells, following CCl4 treatment. Intestinal CRIP
mRNA was suppressed but supplemental dietary zinc prevented this decrease.
CRIP mRNA increased over time after LPS injection in immune tissues, intestine and
brain but not in liver. CRIP mRNA in peripheral blood mononuclear cells decreased and
then increased later in the response. Intestinal and plasma CRIP protein decreased by 6 h.
Intestinal, but not plasma CRIP protein, was normalized by 48 h. These results showed that
CRIP expression was altered by stress challenges and these changes were dependent on the
type of challenge and the site of response.
In vitro protein-protein interaction studies showed CRIP binding to serum albumin
and trypsinogen. The former supports identification of CRIP in the plasma. Another putative
protein partner for CRIP was an N-terminally blocked, 116 kDa protein.
These studies suggested that CRIP was secreted into the plasma and was bound to
serum albumin. Increased CRIP expression later in the immune and inflammatory response
showed that it was not involved in the acute phase response but could function in
development/repair after injury.
Human CRIP cDNA was cloned from human small intestine poly (A)' RNA by
polymerase chain reaction. An expression system was developed as a corollary to these
studies. Human CRIP protein was found to differ in two amino acids compared to rat CRIP,
and one amino acid compared to human CRIP cloned from the heart, suggesting the existence
of different alleles of the gene.
INTRODUCTION AND LITERATURE REVIEW
Cysteine-rich intestinal protein (CRIP) was isolated from rat intestinal cytosol by
Hempe and Cousins (1991) and found to bind zinc during transmucosal transport. Its zinc
binding characteristics and predominance in the intestine argued for a role as an intracellular
zinc trafficking protein (Hempe and Cousins, 1992). CRIP mRNA was first identified in the
mouse intestine. It was found to be developmentally regulated, with low abundance in the
intestine at birth, which increased to adult levels post-weaning (Birkenmeier and Gordon,
1986; Levenson et al., 1993). Increased binding of 65Zn to CRIP occurred during the post-
weaning period, corresponding to the increased level of CRIP mRNA and protein at that
time, but the mRNA levels did not appear to be regulated by zinc (Levenson et al., 1994b).
However, CRIP mRNA was increased by glucocorticoids (Needleman et al., 1993), and
analysis of the promoter region showed the presence of putative glucocorticoid response
elements (Levenson et al., 1994a). The primary sequence of CRIP was found to contain a
putative signal sequence that was implicated in protein translocation (Tsonis et al., 1988).
CRIP contains a single copy of a cysteine-rich domain called the LIM motif. When
two atoms of zinc are bound, the motif comprises a double zinc finger (Liebhaber et al.,
1990). This motif has been found in products of several genes involved in developmental
regulation such as the lin-11 (C. elegans cell lineage gene), isl-1 (rat insulin-1 enhancer
binding protein) and mec-3 (C. elegans gene required for differentiation of mechanosensory
neurons) genes (Freyd et al., 1990). The acronym LIM was coined from the first letters of
the names of those three genes. Many other genes have been reported to have this motif,
usually in tandem repeats of two or more, including rhombotin, an oncogene (Boehm et al,
1991) and the human cysteine-rich protein (Wang et al., 1992). Many of these LIM proteins
contained a homeodomain and were postulated to function as transcription factors (Li et al.,
1991). The LIM motif has been proposed to be involved in protein-DNA interaction, as in
classically defined transcription factors, or protein-RNA interactions. The LIM motif has
also been found to interact with other proteins or to form homodimers (Crawford et al.,
1992; Sadler et al., 1992; Schmeichel and Beckerle, 1994).
CRIP belongs to the class of LIM proteins that does not have the homeodomain
(LIM-only proteins). This argues against a DNA binding function. Consequently, CRIP is
thought to function through protein-protein interactions to mediate changes in cells. CRIP
contains only one copy of the LIM motif which has been shown to be sufficient for binding
to another LIM protein (Schmeichel and Beckerle, 1994).
CRIP expression was originally thought to be specific to a few tissues such as the
intestine. However, more recent evidence showed that CRIP expression was widespread, with
particularly high levels in immune tissues even under normal physiological conditions. Based
on these and other new discoveries about the different LIM-containing proteins (Dawid et al.,
1995), it is now postulated that CRIP is involved in a different function not related to
intestinal zinc trafficking. One of the suggested roles is in maintenance of cellular
The high abundance of CRIP in the intestine has focused on that organ as a site to
evaluate a function. A study by Fleet et al. (1993) found that treatment of Caco-2 cells with
la,25-dihydroxyvitamin D3 decreased the level of CRIP mRNA to 70% while zinc transport
was increased. Therefore, they questioned the role of CRIP in zinc absorption. Needleman
et al. (1993) proposed that CRIP was a mediator of glucocorticoid action rather than a
component of an intracellular zinc trafficking system because of the pattern of hormonal
regulation by dexamethasone and T4. However, recently, another group of researchers still
stated that the model of zinc absorption proposed by Hempe and Cousins (1992) was
supported by the study they carried out in diabetic animals. They found that metallothionein
was induced in these animals, resulting in decreased zinc binding to CRIP, and lowered zinc
transport (Escobar et al., 1995). Therefore, the functions) of CRIP still remains to be
Levenson et al. (1994a) showed that the promoter region of the CRIP gene did not
have a TATA box but had a GC-rich region. The general transcription factor, Spl, was
found to bind to the GC-rich region of housekeeping genes. This pointed to a housekeeping
function for CRIP such as in cell repair or maintenance of differentiation. CRIP may interact
with other LIM and non-LIM proteins while carrying out its function.
A human CRIP partial sequence was cloned as an expressed sequence tag from a
human fetal brain (Adams et al., 1992), and this sequence enabled the full cDNA to be
cloned from the human heart (Tsui et al., 1994). The human cysteine-rich heart protein
(hCRHP) was cloned using a procedure called 3' rapid amplification of cDNA ends (Tsui
et al., 1994). The similarity and identity of the human and rat genes were very high,
showing remarkable conservation of the gene over the span of evolution. Another LIM
protein, the human cysteine-rich protein (hCRP), was also found to be highly conserved when
the DNA of eukaryotic species from yeast to mammals (mouse and human) were analyzed
(Liebhaber et al., 1990).
This dissertation project is a prospective study involving several different sections,
each providing new information about CRIP, applicable to defining a physiological function
for the protein. The regulation of CRIP was studied in an animal model, Sprague-Dawley
rats, and focused on the intestine, nervous system, liver and immune tissues. Two different
stress challenges were used to study CRIP regulation since CRIP expression was different
in various tissues under normal physiologic states. A toxic challenge with carbon
tetrachloride and an immune challenge with lipopolysaccharide were used.
Carbon tetrachloride (CCl4), an organic toxin that affects the liver, was used to cause
a toxic challenge which leads to hepatotoxicity and neuropathy (Rechnagel and Glende, Jr.,
1973; Dfaz-Mufioz and Tapia, 1989). The liver expressed one of the lowest CRIP mRNA
levels of all the tissues examined and perturbation of hepatic function could cause a change
in CRIP expression. Since the brain has also been shown to be affected by this toxic insult,
CRIP levels in the brain were also examined. In addition, the effect of dietary zinc on CRIP
expression in this model of inflammation was analyzed as zinc has been indirectly implicated
as being cytoprotective against CCI4, possibly by inducing metallothionein, a zinc
metalloprotein associated with cellular defense (Chvapil et al., 1973; Clarke and Lui, 1986).
Recently, very high levels of CRIP mRNA were observed in residential peritoneal
macrophages and lipopolysaccharide (LPS)-stimulated peritoneal macrophages.
Consequently, CRIP mRNA and protein expression were investigated at different time points
in cells of the immune and nervous systems, intestine, liver and plasma in response to
immune challenge with lipopolysaccharide (LPS), a bacterial cell wall antigen that was able
to elicit an acute-phase immune response (Watson et al., 1994).
As a potential avenue for further determining the function of CRIP, the search for a
potential protein partner of CRIP was carried out with the idea that the protein partners)
would have a similar, complementary or opposing role. If the protein partner of CRIP was
a protein with a known function and subcellular localization, the information would suggest
a related function and localization for CRIP.
An additional part of this project was to further characterize CRIP by comparing the
recombinant rat and human protein. Using the published sequences for heart CRIP, primers
were designed to clone the cDNA from human small intestine using the polymerase chain
reaction (PCR). The cDNA was then used in an expression system to make sufficient
quantities of the human protein for comparisons with the rat recombinant protein and for
future studies with antibodies.
There are a number of areas where the biology of CRIP would be nutritionally
significant. CRIP is a zinc-finger protein that is likely to be important in the developing
intestine. CRIP has also been found in immune cells and could be an important factor in
determining or explaining the role of zinc in immune function. Zinc deficiency has been
found to suppress immune functions, suggesting specific zinc-dependent components are
involved (Keen and Gershwin, 1990). The function of CRIP during immune processes could
be used for diagnostic purposes and the antibody produced from the human protein would
have clinical implications. Overall, the role of zinc and its interactions with zinc-finger
proteins would help to define the importance of regulating this trace mineral in our diet.
1. To develop information needed to define a function for CRIP:
a. by studying the expression of the gene in the rat intestine, liver, brain and
immune tissues to stress challenges, including a toxic challenge by carbon-
tetrachloride and an immune challenge with lipopolysaccharide.
b. by identifying a potential protein partner that may have a known function that
could define a related function for CRIP using intestinal cytosolic and membrane
2. To further characterize CRIP by comparing the recombinant human protein produced
by a bacterial expression system using the CRIP cDNA generated from human
intestinal mRNA with the rat protein and the recombinant human protein produced
from heart mRNA.
1. The regulation of CRIP is tissue-specific and is dependent on the site of challenge and
response during an insult.
2. The human CRIP cDNA is present in the intestine and has a high homology with rat
CRIP but is not identical.
Cysteine-Rich Intestinal Protein
Cysteine-rich intestinal protein (CRIP) is a 77 amino acid, 8.55 kDa protein, with 7
cysteine residues. As reviewed earlier, cysteine and histidine residues, arranged in a
consensus sequence of CX2CXI7.19HX2CX2CX2CX7.11(C)X8C, form a double zinc finger
structure called the LIM motif (Freyd et al., 1990). Although the LIM motif resembles the
iron-sulfur centers of ferredoxins, in vivo and in vitro studies have shown that the LIM motif
does not bind iron or copper with high affinity (Archer et al., 1994; Khoo and Cousins,
1994). Structural studies of the two fingers showed that one tetrathiolate site (S4) and one
S3NA site each bound one zinc(II) ion tetrahedrally (Kosa et al, 1994). The authors also
found that slow metal exchange can occur with Co(II) and Cd(II) with the S3N1 site being
more kinetically facile. Metal binding appeared to stabilize the tertiary structure of the
protein. Mutational analysis studies (Michelsen, 1994) provided further evidence that the two
sites were arranged in a double zinc finger motif as had been proposed earlier by Wang et
The amount of CRIP purified from the soluble fraction of rat intestinal tissue was
found to be between 15-20 Mpg/g tissue (Khoo and Cousins, 1994). It was found to bind 65Zn
more avidly than metallothionein during transmucosal zinc transport in rats fed a normal zinc
diet (Hempe and Cousins, 1992). However, when metallothionein was induced by feeding
a high zinc diet, it was able to bind more "Zn than CRIP, supporting the postulated role of
metallothionein as a controller of zinc homeostasis. Levenson et al. (1993) showed that the
CRIP mRNA was expressed most abundantly in the mid-villar region of the intestine. They
also showed that dietary zinc had little influence on expression (Levenson et al., 1994b).
Fernandes et al. (1995) recently showed by immunohistochemistry that CRIP was localized
predominantly in the Paneth cells with lower levels at the tip of the villus. The Paneth cells
and the brush border enterocytes both differentiate from the same stem cells (Gordon et al.,
1992) and thus, high expression of CRIP message in the mid-villar region could support the
high protein abundance and protein turnover in the Paneth cells in the rapidly proliferating
intestinal epithelium lining the villi. This suggested that CRIP message was more abundant
in cells that were not as differentiated while protein expression was highest in more
differentiated cell types such as brush-border enterocytes and Paneth cells.
The Paneth cells, which lie in the crypt region of the small intestine, are
morphologically well-characterized but their function is still relatively unknown. It is the
only cell lineage that migrates downward from the crypt-villus junction and it is derived from
the same multipotent stem cells that produce villus enterocytes (Bry et al., 1994). They have
been implicated in the production and secretion of antimicrobial peptides (Selsted et al.,
1992) for intestinal defense against invading organisms, small amounts of pancreatic enzymes
such as trypsinogen as a supplement to the pancreas (Senegas-Balas et al., 1991), growth
factors (Bry et al., 1994), immunoglobulins (Erlandsen et al., 1976; Rodning et al., 1976),
and tumor necrosis factor (Keshav et al., 1990). Paneth cells also secrete lysozyme which
acts as another line of intestinal defense (Erlandsen et al., 1974). Therefore, it has been
proposed that the Paneth cells represent a first line of defense as an alternative to recruitment
of inflammatory cells as a method to preserve the integrity of the epithelial cells (Boman,
The zinc content of Paneth cells was shown to be very high in numerous studies
(Dinsdale, 1984; Dinsdale and Biles, 1986; Sawada et al., 1993). Consequently, there is
considerable interest in the zinc-binding proteins in the Paneth cells. CRIP is one of the few
zinc binding proteins identified in Paneth cells. Cryptdins (or defensins), are members of
a family of cysteine-rich antimicrobial peptides found in Paneth cells (Selsted et al., 1992).
The cysteine residues of defensins are believed to form disulfide bonds, which would
preclude zinc-binding. A recently discovered 90 kDa zinc-binding protein, called ZBPP-1,
was found to be localized in cytoplasmic granules of the Paneth cells by
immunohistochemistry (Sawada et al., 1993; 1994). Treatment with dithizonate, a zinc
chelator, caused ZBPP -containing granules to disappear. They reappeared again with
regeneration of Paneth cells. Normally, these cells turn over fairly rapidly, differentiating
during downward migration and reside in the crypt area for about 20 days before being
removed by phagocytosis (Bry et al., 1994; Gordon et al., 1992). Localization of CRIP in
these cells suggest that it could participate in cell proliferation/differentiation function or
mucosal defense. The zinc content of the Paneth cells and the destruction of these cells by
dithizonate suggest that zinc-binding proteins such as CRIP could be important in the
functions of the Paneth cells in the small intestine.
CRIP mRNA levels were also found to be high in the female reproductive organs
which undergo active cell proliferation and differentiation (Cousins and Fernandes, 1995).
LIM Motif/Other LIM Proteins
Recently, the LIM family of proteins were loosely grouped into four groups (Dawid
et al., 1995) based on the differences in structural domains. The first group of LIM proteins
contain a homeodomain implicated in DNA binding while the other three groups are LIM-
only proteins without the homeodomain. The second group consists of LIM-only proteins
with 1-5 LIM domains near to the N-terminus and little else; Proteins of the third group
contain 3-4 LIM domains at the C-terminus. The last group of LIM proteins has more varied
structures which include functional domains such as a kinase domain and a G-activating
protein (GAP) domain. The LIM motif has been added to the list of zinc finger motifs that
play roles in gene regulation and, with the LIM-only proteins, possibly signal transduction,
cell adhesion and oncogenesis (Boehm et al., 1991; Dawid et al., 1995; Klug and Schwabe,
Tissue specificity of some of the LIM family of proteins based upon detectable
mRNA levels is as follows: rat CRIP (rCRIP)-intestine, lung, colon, spleen, thymus,
placenta, heart, ovary, skin, adrenal, muscle and bone marrow (Birkenmeier and Gordon,
1986; Levenson et al., 1993), hCRP-lung, colon, brain, ovary, thymus, placenta, spleen,
skin, heart, adrenal, kidney and pancreas (Wang et al., 1992), zyxin-adhesion plaque isolated
from avian smooth muscle (Sadler et al., 1992), and chicken CRP-cytoskeleton, gizzard,
stomach and intestine (Crawford et al., 1992).
Several different types of LIM-homeodomain proteins have been discovered in the
brain, as discussed in the next section, which implicates this motif in regulating functions in
the nervous system as well as in other tissues. In addition to CRIP, two other highly related
LIM proteins, CRP2 and ESP1 (estrogen-stimulated protein 1), have been found in the brain
although these three proteins appear to be ubiquitously expressed (Okano et al., 1993; Nalik
et al., 1989). ESPI has been found to be stimulated by 176-estradiol in both male and
female rats. A novel protein kinase found to be highly expressed in the rat brain has also
turned out to be a LIM protein with two repeats of the LIM domain (Mizuno et al., 1994).
This protein, LIMK, is proposed to function in developmental and oncogenic processes
through interactions with other LIM proteins. This finding suggests that some LIM proteins
may form a signaling complex through protein-protein interactions with other proteins and
may utilize LIM kinases for activation.
A LIM-homeodomain transcription factor was found to function in the activation of
the gene for a mouse glycoprotein hormone ao subunit in the pituitary by binding to the
promoter region (Roberson et al., 1994). This hormone is a very early marker of pituitary
differentiation and tied the function of yet another LIM-homeodomain protein, recently
named P-LIM (Bach et al., 1995), to cell lineage determination. The three original genes
found to contain the LIM motif, i.e. lin-11, isl-1 and mec-3, were found to have that function
(Freyd et al., 1990).
Several of the LIM-homeodomain proteins are expressed in the central nervous
system during developmental stages such as Lim-1, Isl-1, and GSH-4 (Barnes et al., 1994;
Lumsden, 1995). Murine GSH-4 has been found to be important for the development of
respiratory control mechanisms and is expressed transiently in the central nervous system
during late development (Li et al., 1994). Primary neurons in zebrafish expressed Isl-1 at
the end of gastrulation (Korzh et al., 1993).
The LIM domain of Isl-1 inhibited the binding of the homeodomain to its target DNA
(Sdnchez-Garcfa et al., 1993). This could represent a method of gene regulation by LIM
proteins. The inhibition was lifted by truncation of the LIM domain. This could mean that
under physiological conditions either an activator, such as another LIM protein, is required
to bind to the LIM domain of 1sl-1 and remove the repression of the homeodomain or that
the LIM domain is cleaved by proteolytic activity. This represents a possible mechanism of
negative regulation of LIM-homeodomain proteins in vivo and suggests that the LIM motif
is important in protein-protein interactions in both the LIM-homeodomain proteins as well
as the LIM-only proteins.
It has been proposed that the LIM-only proteins could bind to the equivalent domain
in LIM-homeodomain proteins and function to activate or depress their transcriptional
activities. LIM-only proteins have received some attention directed at determining the nature
of their interactions with other proteins (Feuerstein et al., 1994; Sinchez-Garcfa and Rabbitts,
1994; Schmeichel and Beckerle, 1994).
Proteins are known to interact with other proteins in virtually every type of cellular
process. Some of these form multi-subunit complexes such as G-protein-coupled receptors,
proteosome complexes for ubiquitination and degradation of cellular proteins and
transcriptional machinery (Austin et al., 1994), while some of the interactions are more
transient. For example, protein modifications such as phosphorylation, proteolysis, and
transamination involve protein-protein interactions. Proteins also transiently interact during
cell growth, signaling, metabolism, transport as well as breakdown and reformation of
subcellular structures during cell cycles, immune reactions and cell differentiation/activation
(Austin et al., 1994; Calakos et al., 1994; Phizicky and Fields, 1995). These events are
usually mediated by ligands such as peptides to promote dimerization and present a method
of controlling protein activation through complex formation (Austin et al., 1994).
A LIM-only protein that has been studied extensively is zyxin, a cytoskeletal protein
with three tandem copies of the LIM motif, which is able to interact with another LIM
protein called the chicken CRP (Sadler et al., 1992). Both proteins are associated with the
cytoskeleton and are postulated to be involved in transmembrane signaling. Zyxin is also
able to bind a-actinin which is another cytoskeletal protein that does not have the LIM
domain (Crawford et al., 1992). It was proposed that the LIM domain is a modular protein-
binding interface with the LIM domain of zyxin binding to the LIM domain of chicken CRP.
The proline-rich region of zyxin appears to be involved in another interaction with ot-actinin
(Schmeichel and Beckerle, 1994). Other LIM proteins have been found to have these distinct
types of binding where LIM/non-LIM interactions appear to require some specificity in the
non-LIM proteins as well in the LIM protein partner. For example, a LIM protein called
ENIGMA was found to recognize tyrosine tight turns in its partner (Wu and Gill, 1994).
These tyrosine tight turns are found in proteins such as the human insulin receptor and are
"codes" required for endocytosis of cell receptors. Mutation in these sequences prevents the
binding of ENIGMA, showing that there is specificity in the recognition of these target
Paxillin is another important LIM-only protein. It is a focal adhesion protein in
fibroblasts and binds to vinculin, c-src and the focal adhesion tyrosine kinase ppl25'
(Turner and Miller, 1994). It was proposed to interact with growth factor receptors for
signal transduction through phosphorylation of its tyrosine residues (Salgia et al., 1995).
Paxillin and its partners represent examples of protein-protein interactions in the pathway for
microfilament assembly (Melamed et al., 1995).
LIM proteins and immune function
Both LIM-only and LIM-homeodomain proteins have been found to play roles in the
immune function. A homeobox gene, LH-2, was found while screening early markers of B-
lymphocyte differentiation (Xu et al., 1993). It was found in both B- and T-lymphoid cell
lines while the highest expression was in discrete areas of the developing central nervous
system. A LIM-only protein called PINCH, containing five LIM domains, was found to
contain an autoepitope in the third LIM domain homologous to senescent red blood cell
antigen (Rearden, 1994). This antigen is important in immunoglobulin-mediated removal of
aged cells. Paxillin, a focal adhesion protein found to be important in protein-protein
interactions, was shown to be present in adherent and non-adherent neutrophils (Fuortes et
al., 1994) and human B lymphocytes (Melamed et al., 1995). It was discovered that the
tyrosine phosphorylation of paxillin in these different cell types required not only activation
by growth factors and cytokines, but also the formation of aggregates and microfilament
assembly (Melamed et al., 1995). The aggregation of IgE receptors also appeared to induce
phosphorylation of paxillin (Hamawy et al., 1994). These data suggest that paxillin is
important in signal transduction either activated or mediated by the complex formation of
other proteins involved in the adhesion or aggregation functions of the immune response.
LIM proteins have also been implicated in human T cell leukemia via specific
association with basic helix loop helix proteins (Wadman et al., 1994). Rhombotin, an
oncogene with two LIM motifs, was found to be associated with T-cell leukemia
chromosomal translocations (Boehm et al., 1991) implicating LIM proteins in control of
proliferation of immune cells.
Zinc in Biology
With the presence of so many zinc binding proteins carrying out roles in regulating
gene activity and other functions, zinc as a micronutrient is obviously playing an important
biological role. The recommended dietary allowance for zinc is now set at 15 mg/day for
men and 12 mg/day for women (National Research Council, 1992). It has also been
established that there is an increased need for zinc during different physiologic states
including pregnancy and lactation (15 mg and 19 mg/day respectively) and immune stress
(Prasad, 1988). Under these conditions, there is increased cell growth and differentiation
which require zinc. Sandstead (1991) reported that mild zinc deficiency may be more
prevalent in more societies than we think.
Zinc status in the general population has been estimated by measuring zinc in plasma
and serum, hair, saliva and erythrocytes (King, 1990; Solomons, 1981), with methods such
as atomic absorption spectroscopy and inductively-coupled plasma emission. Reticulocyte
metallothionein has also proved to be a useful indicator of zinc status when measured using
enzyme-linked immunoassay (ELISA) (Grider, 1990). King (1990) suggested that both
plasma zinc and liver metallothionein should be measured since during a stress response,
when plasma zinc is low, metallothionein increases. Plasma zinc has been found to be
carried by albumin in the plasma (Etzel et al., 1982; Falchuk, 1977; Smith et al., 1979).
Plasma albumin levels are depressed under stress conditions.
Physiological functions and absorption
The nutritional essentiality of zinc was not recognized in humans until 1963 when
Prasad et al. (1979) discovered deficiency symptoms in Egypt. The primary symptoms
included dwarfism, hypogonadism and delayed sexual maturation but many other symptoms
were observed for mild to severe zinc deficiency (Prasad, 1984). These included mental
retardation, compromised cell-mediated immune functions, skin lesions and alopecia (Prasad,
1984). An inherited rare autosomal recessive gene cause a congenital zinc malabsorption
syndrome (Acrodermatitis enteropathica) characterized by those symptoms as well as
diarrhea, but this syndrome was reversed by oral zinc treatment (Hambidge and Walravens,
Due to the nutritional importance of zinc, the study of zinc absorption and transport
in the intestine has been intensively pursued as reviewed by Cousins (1982; 1985; 1989).
The knowledge as it stands now is that zinc uptake at the brush border involves both passive
diffusion and a saturable, carrier-mediated uptake (Cousins, 1989) through activity of a
membrane transporter that could require small zinc binding ligands (DiSilvestro and Cousins,
1983). At the basolateral membrane, the uptake is postulated to be energy-dependent
(Kowarski et al., 1974; Oestreicher and Cousins, 1989) while the efflux of intestinal zinc
back into the lumen is thought to be a mechanism for zinc homeostasis (Hoadley et al.,
1987). In 1991, Hempe and Cousins suggested CRIP as candidate for a zinc transporter, due
to its abundance in the intestine and its saturable affinity for zinc, but the later discovery of
the ubiquitous distribution of the protein argued against a transport function.
Zinc binding proteins
Zinc binding proteins have been found to play many roles in biology including as zinc
finger proteins (Klug and Schwabe, 1995), zinc storage proteins and enzymes (Coleman,
1992), antioxidants (Mulder et al., 1991), signal transducers (Hubbard et al., 1991), and
many others. These proteins bind zinc in a variety of motifs especially the transcription
factors which have either zinc fingers or zinc twists (Vallee et al., 1991) and metallothionein
which binds zinc in clusters. These different structures also bind zinc with different affinities
which can determine their ability to exchange zinc, and define a structural or catalytic role
for zinc in these proteins. The chelating ability of different proteins can be used to regulate
the function of another protein. For example, metallothionein (zinc cluster) chelates zinc
from transcription factors (zinc finger) such as Spl in vitro (Zeng et al., 1991a), and TFIII
or transcription factor III (Zeng et al., 1991b). CRIP belongs to the group of metalloproteins
which has the zinc finger motif (Klug and Schwabe, 1995) and binds zinc with the same
affinity as these proteins (Kosa et al., 1994). Therefore, metallothionein has the capability
of chelating zinc from CRIP in vitro. However, zinc absorption studies indicate that with
normal physiological zinc concentrations, CRIP was able to bind zinc in the presence of
excess metallothionein (Hempe and Cousins, 1991; Khoo, unpublished observations).
Metallothionein is induced during the acute-phase immune response as well as during
other stress responses such as those involving oxygen radical formation following irradiation
or exposure to chemicals (e.g. ozone, alcohol). During an immune response liver
metallothionein increases. This results in a redistribution of zinc from plasma into liver and
some other tissues such as bone marrow. This redistribution may decrease the availability
of zinc to invading organisms. Metallothionein is also thought to mediate protection against
free radical-induced damage, perhaps by interacting with the damaging agent or the products
e.g. hydroxyl radical (Clarke and Lui, 1986). The promoter region of the metallothionein
gene contains, among other regulatory sequences, metal response elements that allow it to
be induced by heavy metals such as zinc and cadmium (Hamer, 1986). The promoter also
contains a response element called antioxidant response elements, or ARE, for activation by
oxygen radicals (Dalton et al., 1994). Metallothionein is also inducible by glucocorticoids,
endotoxin and non-metallic compounds such as carbon tetrachloride and paraquat (De et al.,
1990; Fleet et al., 1990; Min et al., 1991). Therefore, metallothionein is a good indicator
of not only heavy metal toxicity but also the acute phase response and inflammation.
Other functions of zinc
Other than being just part of the structure of zinc fingers or zinc clusters, zinc is an
important trace element in itself. It has special chemistry that makes it useful in biological
systems such as the lack of redox chemistry which could cause free radical damage, the rapid
change of coordination numbers when binding, and the rapid exchange of ligand-complex
(Williams, 1984). One of the most important direct functions of zinc is regulating gene
expression by binding to proteins which in turn bind metal response elements (MREs) in gene
promoters as seen for the metallothionein gene (Cousins, 1994). Zinc has also been shown
to induce metal response binding factors in the nucleus of human HepG2 cells and to
facilitate their binding to the DNA. Therefore, zinc acts at two different levels to stimulate
genes containing MREs (Czupryn et al., 1992). The acute phase proteins, a1-acid
glycoprotein and C-reactive protein, have also been shown to be induced by zinc through
zinc-facilitated binding of transcription factors to MREs (Yiangou et al., 1991).
In the immune system, zinc deficiency affects T and B cell response as well as
macrophage functions (Keen and Gershwin, 1990; Vallee and Falchuk, 1993). Zinc
deficiency was found to interfere with the function of macrophages in taking up and killing
parasites (Wirth et al., 1989). This function was restored by zinc supplementation (Wirth
et al., 1989). Extreme thymic atrophy and loss of splenocytes had been reported in zinc
deficient mice as had alteration in cytokine production (Fraker et al., 1995; Serushago and
Chandra, 1995). Zinc was found influence apoptosis of immune cells as well (Fraker et al.,
Zinc has also been postulated to act as an antioxidant through two possible
mechanisms, by protecting sulfhydryl groups against oxidation and by inhibiting other
transition metals from producing reactive oxygen species (Bray and Bettger, 1990).
Numerous studies with administration of toxic agents such as carbon tetrachloride with either
an injected dose or excess dietary zinc, showed a protective effect of zinc against the toxic
agent (Chvapil et. al, 1973; Clarke and Lui, 1986). The mechanism of protection, as
discussed earlier, may involve induction of metallothionein to act as a free radical scavenger.
Zinc in the brain
The physiological role of zinc in the brain has gathered a lot of interest lately due to
findings on the possible pathological effects of zinc in neurological disorders such as
Alzheimer's disease (Bush et al., 1994; Price et al., 1991). Physiological roles have been
proposed for zinc in the central nervous system. It has been proposed that zinc may function
as a neuromodulator and as a developmental signal for synaptogenesis (Smart et al., 1994).
It is also thought that zinc may facilitate packaging and storage of neurotransmitters in
synaptic vesicles (Smart et al., 1994).
Zinc is found in differing concentrations in various parts of the brain and is postulated
to be made up of three separate pools, vesicular, free and protein-bound zinc (Frederickson,
1989). The concentration of zinc in the cerebrospinal fluid is the same as that of
microligand-bound zinc in the serum but studies with brain cells have shown that exchange
between serum and brain cells does not occur freely but requires an energy-dependent process
(Howell et al., 1984). Concentrations of zinc within brain zinc pools range from 200 pM
in normal conditions to 300 jpM during brief surges of zinc being released from terminal
boutons (Frederickson, 1989).
Neither dietary deficiency nor excess of zinc appeared to cause any immediate or
pronounced effect in zinc concentrations in most areas of the brain, showing that zinc is
remarkably conserved in this organ (Kasarkis, 1984; Wallwork, 1987). Although changes
in zinc concentrations are not detectable during zinc deficiency, it has been well established
that dietary zinc deficiency led to clinical signs of neurological damage (Prasad, 1979; O'Dell
et al., 1990). Therefore, subtle changes in zinc concentrations within specific regions may
be enough to cause this damage. Acting as a neurotransmitter, synaptic release of zinc may
be directly responsible for regulating the cyclic feeding patterns observed during dietary zinc
The cellular basis for the anomalous staining of zinc after the destruction of the
hippocampus and numerous other injury-related changes in zinc distribution in the brain are
still unknown (Frederickson, 1989; Lees et al., 1990; Yokoyama et al., 1987). Zinc is found
in high concentrations in the hippocampus where it is mostly localized in the mossy fiber
boutons and is postulated to act as a neurotransmitter in this region (Frederickson, 1989;
Slomianka, 1992). In hippocampal slices, zinc has been shown to be both taken up as well
as released upon different types of stimulation (Assaf and Chung, 1984; Howell et al., 1984).
These stimulations appear to release zinc bound to soluble proteins (Baba et al. 1989, Assaf
and Chung, 1984; Xie and Smart, 1991). This may be deleterious or beneficial, depending
on different physiological states. Several studies have implicated zinc in neuropathological
states such as Alzheimer's and epilepsy (Bush et al., 1994; Frederickson, 1989). A zinc-
binding protein that was found to be deficient in Alzheimer's disease brain is a
metallothionein-like protein (Uchida et al., 1991). This was subsequently shown to be
metallothionein-3, an isoform of metallothionein that is unique to the brain (Palmiter et al.,
Most of the data relating diseases of the nervous system to zinc are very conflicting.
For example, both abnormally high and abnormally low serum zinc concentrations have been
found to be associated with epileptic seizures while both dietary zinc deficiency and systemic
zinc injections have been found to facilitate electroconvulsive seizures (Frederickson, 1989).
It is clear that zinc can act as a proconvulsant as well as an anticonvulsant and the effect is
dependent on dose, route of administration, type of seizure, species, clinical history, status
and etiology of the disease. There is an obvious need for a delicate balance in the
distribution of zinc in the central nervous system. Smart et al. (1994) in their review
emphasized that zinc can play two roles, a neuroprotective role at low concentrations by
antagonizing the effect of calcium, and a neurotoxic role at higher concentrations through
direct toxicity or enhancement of the effect of glutamate-activated non-NMDA receptors.
Rapid uptake and release of zinc is very important to preserve the integrity of
neurotransmission not unlike the regulation of calcium levels. This appeared to be the case
as following systemic administration, "6Zn was detected immediately throughout the brain
(Kasarkis, 1984). Zinc was found to be involved with both inhibitory and excitatory amino
acid receptors such as GABA (inhibitory) and NMDA (excitatory) receptors by associating
with one or two binding sites on these receptors and causing different effects with different
types of neurons (Koh and Choi, 1988; Peters et al., 1987).
A phenomenon that occurs in denervated tissue is the appearance of reactive
macrophages that are stained intensely by a zinc-specific histologic stain. The appearance
of these macrophages in the areas of high zinc accumulation suggests that either these cells
are secreting a reactive pool of zinc or that they are sequestering free zinc liberated as a
result of the neuronal injury (Frederickson, 1989). CRIP mRNA has been found to be
abundant in macrophages and as a zinc-binding protein, it may play a role in regulating zinc.
During neuronal damage there is an increase in zinc-specific staining in degenerating axons
(Pdrez-Clausell et al., 1989). Again, the appearance of zinc in these dying neurons could
be an indication of the role of zinc in the response to cellular damage. One suggestion was
that zinc observed histologically was associated with a nerve growth factor (NGF)-like
trophic factor, a zinc-binding protein, for the purpose of cell repair or regeneration (Kesslak
et al., 1987).
As a result of the studies described above, there is considerable interest in zinc
binding components of various cells of the nervous system. CRIP mRNA is found in the
brain and as a cytosolic zinc-binding protein whose function is still unknown, it would be
interesting to see the effect of a toxic challenge, such as with CCl4, on brain CRIP
expression. Part of the toxic effect of zinc was postulated to occur inside neuronal cells
following entry of zinc and calcium after injury to the neurons (Weiss et al., 1993). A
systemic challenge was carried out with CCI4, as described in the following chapter, which
may or may not, affect the brain.
Two different types of stimuli that can cause changes in organs such as intestine,
liver, immune cells, and the central nervous system will be discussed in the following
sections. Since tissues are regulated by different mediators under physiologic stress, it is
possible that the expression of CRIP could be altered selectively. This would vary depending
on the need for the presence of CRIP in a given tissue at a particular stage in the stress
Carbon tetrachloride (CC!4) is a highly toxic inflammatory agent that can cause liver
damage as evidenced through the increase in the activity of liver enzymes such as alanine
aminotransferase (ALT) in the plasma. The hepatotoxicity results from cell damage caused
by the trichloromethyl free radical produced through metabolism of CCl4 by the cytochrome
P-450-dependent monooxygenase (Rechnagel and Glende, 1973). Cirrhosis occurs with
various routes of chronic CCi4 administration and can lead to loss of hepatic function but
absence of significant extrahepatic complications or abnormal nitrogen metabolism (Blondd-
Cynober et al., 1994). Increased nitric oxide production may also be another effect of CC!4
treatment (Chamulitrat et al. 1994).
A mechanism was proposed for the cellular events that determine the nature and
extent of the hepatotoxic injury (Mehendale et al., 1994). According to these authors,
stimulation of tissue repair is the key determinant of the extent of liver injury and the
protection against cell death. A small dose of CCl4 administered 24 h prior to a large lethal
dose confers maximal autoprotection by downregulating the cytochrome P-450 system and
also by stimulating hepatic cell division to facilitate tissue repair and recovery. The
downregulation of the cytochrome P-450 system is important as uncontrolled activation of
this system has been found to lead to massive liver injury.
Antioxidants have also been implicated in both augmentation and inhibition of the
CCl4-induced effects. Vitamin A was found to potentiate the effects of CCi4 by activating
Kupffer cells to release increased amounts of oxygen free radicals (ElSisi et al., 1993).
Kupffer cells were thought to participate in CC14 toxicity by producing toxic secretary
products to attract neutrophils to the site of injury, thus amplifying the inflammatory response
(Edwards et al., 1993). On the other hand, vitamin E was shown to inhibit CCl4-induced
liver damage in a dose-dependent fashion when delivered in liposomes (Yao et al., 1994).
Biodistribution studies showed that these liposome-associated antioxidants were localized in
the Kupffer cell populations, further emphasizing the role of the macrophage-like activity in
the etiology of the liver injury. Interestingly enough, Kupffer cell stimulation by C.parvwn
did not potentiate the CCl4-induced injury but instead appeared to lessen the effect (Raiford
and Thigpen, 1994). It was postulated that this immune stimulation led to a different pattern
of secretion of soluble mediators by the Kupffer cells and thus, altered the susceptibility to
Dietary as well as injected zinc have been shown to be protective against CCl4-
induced injury when administered in pharmacological doses. Chvapil et al. (1973) observed
that a high zinc diet of 1000 mg/kg was able to protect the liver against the toxic insult.
They proposed that zinc acted by interfering with lipid-peroxidation and led to tissue damage.
They later found that this occurred only in the liver and red blood cells, but not in the
kidney, lungs, brain and testes, suggesting that the liver and red blood cells were able to
accumulate the excess zinc in proportion to the diet (Chvapil et al., 1974). Other
investigators have also shown that dietary zinc affected the damage caused by CCl4.
DiSilvestro and Carlson (1994) reported that a mild dietary deficiency caused poor resistance
to the injury and negated the protective effect that an acute phase response was able to elicit,
possibly by inducing metallothionein. It was suggested that the protective mechanism
involved the oxidation of the sulfhydryl groups of metallothionein by CC!4 (Suntres and Lui,
1990). However, Hanna et al. (1993) presented evidence against significant effect of
metallothionein but proposed that Zn itself may mediate this protection by antagonizing the
negative effects of calcium or by stabilizing cell membranes.
Direct and indirect effects of CCl4 have also been implicated in neurotoxicity
(Clemedson et al., 1990; Yamamoto, 1990). The distribution of cytochrome P-450 system
in different regions of the brain suggested that it could function in controlling the metabolic
and toxic consequences of cytotoxic and carcinogenic metabolites such as alcohol and carbon
tetrachloride (Hansson et al., 1990). These authors found that the most intensely cytochrome
P-450-immunoreactive regions of the brain were the hippocampus, the nigrostriatal system
and the brain stem. Hepatic encephalopathy has been shown to result from hyperammonemia
associated with chronic (Dlaz-Mufioz and Tapia, 1989; Yamamoto, 1990) or acute (Bates et
al., 1989) treatment with CC4. CC4 has also been demonstrated to directly affect cell
membranes of neurons and astrocytes through increased lipid peroxidation (Clemedson et al.,
1990). The effects of CCl4-related injury to the brain have not been fully elucidated.
Lipopolysaccharide is a component of bacterial cell walls and is a potent antigen and
stimulates an acute-phase immune response (Watson et al., 1994). It has been implicated in
a variety of septic states. Multiple organ failure is the most common cause of death in
intensive care patients (Yoshikawa et al., 1994).
Intravenous administration of LPS, which has been used as a model of multiple organ
failure, cause functional disorders in the intestine, liver, lung and stomach due to
inflammatory mediators. The intestine and the lung appear to be severely affected. There
are extensive hemorrhagic lesions in the intestine and the lungs. However, although
extensive infiltration of neutrophils occurs in the lungs 45 mins after LPS challenge, there
does not appear to be any neutrophil infiltration in the intestine as reflected in the lack of an
increase in myeloperoxidase activity in the intestine in contrast to the lungs (Yoshikawa et
al., 1994). However, Valentine and Nick (1992) showed that exposure of intestinal epithelial
cell lines to LPS caused an induction of Mn-SOD mRNA which may be followed by depress
oxygen radical formation through increased Mn-SOD activity.
LPS appeared to mediate its effects through ligand-receptor interactions with many
immune cells including mononuclear phagocytes (Watson et al., 1994). A recent study
showed that LPS binds to an LPS-binding protein which interacts with the LPS membrane
receptor, CD14, to either internalize LPS or cause a signal transduction cascade (Gegner et
al., 1995). In macrophage cell lines, LPS stimulated the release of oxygen radicals which
aid in the intracellular killing of invading organisms (Inoue et al., 1995). As discussed
earlier, one of the postulated mechanisms that protected cells from free-radical injury is the
concomitant activation of metallothionein. LPS is able to stimulate metallothionein in the
liver and other tissues, probably mediated by cytokines such as IL-6 and TNF-a (De at el.,
1990; Etzel et al., 1982; Fleet et al., 1990; Min et al., 1991; Schroeder and Cousins, 1990).
It was proposed that metallothionein may inhibit LPS toxicity for macrophages by altering
intracellular zinc concentrations or facilitating transfer of zinc among the metal-requiring
proteins (Patierno et al., 1983). Since CRIP is found in high abundance in macrophages, it
could be involved in the regulation of zinc during the inflammatory process.
Up to 8 h after LPS administration, there is no increase in metallothionein mRNA in
the brain, suggesting that cytokines have a minimal role in regulating brain metallothionein
(Gasull et al., 1994). Inflammatory response to LPS in the brain was observed to be minimal
by Montero-Monei and colleagues (1994). They found that intracerebral LPS injections did
not cause a massive recruitment of immune cells to the brain, which could have led to
neuronal damage, except for an early recruitment of peripheral macrophages/monocytes.
However, Xu and Ling (1994) showed that intraperitoneal injections of LPS can affect the
brain and allow expression of major histocompatibility complexes. Nerve growth factor and
brain-derived neurotrophic factor have also been shown to be regulated by cytokines
(LeithAuser et al., 1993; Montero-Monei et al., 1994). Therefore, although the brain is
considered an immunologically privileged organ with anti-inflammatory properties, there
appears to be a balance between neurotoxicity and neurotropism to achieve wound repair.
CRIP is expressed abundantly in immune tissues and intestine but is present in much
lower levels in brain and liver. The change in the expression of CRIP when perturbed by
the immune (LPS) and toxic (CCl4) challenges, may help to define a system to study CRIP
function. The toxic challenge with CC14 is specifically geared towards stress on liver while
the LPS treatment is to stimulate immune cells. By studying these different systems, we may
be able to define a function for CRIP. Protein-protein interaction studies will also help to
characterize potential interactions for CRIP with proteins in specific cellular localizations or
proteins with known functions as had been done for other LIM proteins such as paxillin and
zyxin. Cloning of a human CRIP gene from the intestine is a corollary to these experiments
and will be useful in studies with human cells.
STRESS CHALLENGE (I)
The purpose of this study was to investigate CRIP expression in liver, a tissue with
low CRIP expression, after hepatic injury associated with CCI4 treatment. CRIP expression
was also studied in intestine and immune tissues for comparison because these tissues had
highly abundant CRIP message, and because in the next chapter, an immune challenge was
used to perturb the normal physiologic state of those immune tissues. Since CCI4 treatment
mainly affected liver, and liver CRIP mRNA level is normally low, the main focus of this
study was to see if a perturbation of the hepatic system would affect CRIP expression. CC14
has also been found to affect the central nervous system and hence, brain CRIP mRNA was
examined as well. As CRIP is a zinc finger protein, CCl4-induced changes in liver CRIP
expression of rats fed diets with zinc supplemented above the requirement were compared
to control animals fed normal zinc as previous studies using pharmacological levels of dietary
or injected zinc showed that it was able to prevent the CCl4-induced hepatotoxicity.
Materials and Methods
Male Sprague-Dawley rats (175-250 g) were fed diet and water ad libitum for two
weeks. The diet was the semi-purified AIN 76 pelleted diet (Research Diets, New
Brunswick, NJ), sufficient in all nutrients except zinc (Reeves et al., 1993). Zinc was added
in adequate (30 mg/kg diet) or excess (300 mg/kg diet) amounts as described previously
(Blalock et al., 1988). All the animals were fed ad libitum throughout the study.Animals
from both groups were injected intraperitoneally with I ml CC4/kg body wt 24 h prior to
killing while control animals were injected with saline. Northern analysis and
immunohistochemistry shown were representative of data from at least four individual
animals from 3 replicate experiments. Animal experiments were approved by the University
of Florida Animal Care and Use Committee and carried out according to guidelines from the
Committee on Care and Use of Laboratory Animals (National Research Council, 1985).
Blood was obtained from rats under Metofane@ anesthesia (Pitman-Moore, Inc.,
Mundelein, IL) and the rats were killed by exsanguination. Heparinized needles were used
to obtain the blood by open cardiac puncture. The heparinized blood was layered over a 1:1
Histopaque 1077 (Sigma Diagnostics, St. Louis, MO) gradient and centrifuged at 600 x g
for 30 min at room temperature. The top layer was removed and centrifuged at 2 500 x g
for 15 min to obtain cell free plasma. The interface layer from the first centrifugation step
contained PBMCs which were collected by aspiration and washed 3 x with phosphate-
buffered saline (PBS) by centrifugation at 250 x g for 10 min at 4C. These procedures
were carried out according to the manufacturer's protocol to obtain cell-free plasma and
peripheral blood mononuclear cells (PBMC). Plasma alanine aminotransferase (ALT) activity
was used as a marker for hepatic injury and was measured using the ALT diagnostic kit,
Procedure no. 59-UV (Sigma Diagnostics, St. Louis, MO), which was based on the method
of Wr6blewski and LaDue (1956). The data were transformed logarithmically and reported
as such because the values for this assay increased exponentially. Plasma ALT activity was
also used to assess the efficacy of the CCl4 injections. Plasma was also used for
measurements of plasma zinc using flame atomic absorption spectroscopy (Blalock et al.,
Total tissue RNA was obtained from liver, thymus, spleen, peripheral blood
mononuclear cells (PBMC), brain and intestine by the method of Chomczynski and Sacchi
(1987). Total RNA was electrophoresed in 1 % wt/vol agarose gels with 1 x MOPS buffer,
pH 7.0 (20 mM 3-(N-morpholino)-propanesulfonic acid, 5 mM sodium acetate, 10 mM
EDTA) containing formaldehyde and 1 jig ethidium bromide/lane (Maniatis et al. (eds.),
1982). RNA was transferred by capillary action onto nylon membranes (GeneScreen,
DuPont NEN, Boston, MA) and immobilized by UV cross-linking.
Northern hybridization analysis was carried out using a random-primed [o2P]-labeled
rat CRIP cDNA probe (Levenson et al., 1993) and a 60-mer oligonucleotide rat MT-1 probe
(Cousins and Lee-Ambrose, 1992). Equal loading (usually 20 jig RNA) and RNA integrity
were confirmed by ethidium bromide staining. Data were normalized with [&2P]-labeled rat
8-actin cDNA. For quantitation, Northern blots were scanned with a densitomqter. All
radiolabeling procedures used [a"P]-dCTP (Du Pont/NEN Research Products, Boston, MA).
Liver was fixed in 10% neutral buffered formalin and embedded in paraffin. Fixed
sections of 5 Am were used for immunohistochemistry with rabbit antibody raised against a
synthetic peptide representing amino acids 31-56 of rat CRIP and coupled to keyhole limpet
hemocyanin (Fernandes et al., 1995). The IgG fraction obtained by Protein A
chromatography (Pharmacia Biotech Inc., Piscataway, NJ) was used with the Histostain-Spr
kit (Zymed Laboratories, Inc., San Francisco, CA). After being deparaffinized, sections
were placed in methanol containing 0.3% hydrogen peroxide to inhibit endogenous
peroxidase activity. The sections were treated with goat blocking antibody by the protocol
included in the kit before incubation with anti-CRIP IgG. The antibody had been shown by
Western immunoblot analysis with intestinal cytosol to be specific for CRIP (Chapter 4,
Figure 4-1). The presence of peroxidase was detected by adding a substrate-chromogen
solution which gives a red color. These sections were counterstained with hematoxylin.
Intestine was fixed as above and stained with hematoxylin and eosin.
Two-way analysis of variance was carried out using SAS (General Linear Models,
SAS Institute, Inc., Carey, NC). Differences between means were evaluated using the least
square means procedure. Probability values of 0.05 or less were considered statistically
The effectiveness of the CCI4 injections was confirmed by the significant increase in
plasma alanine aminotransferase activity (Table 2-1). Supplemental dietary zinc significantly
increased plasma zinc concentrations compared to controls but did not affect the level of
activity of plasma alanine aminotransferase (Table 2-1). Both the supplemental zinc diet and
CCl4 administration caused an increase in liver metallothionein mRNA but the combined
treatment did not cause an additive increase in the metallothionein expression (Figure 2-1).
Metallothionein mRNA was used as a positive control for the response to supplemental zinc
intake as well.
Table 2-1. Plasma zinc or alanine aminotransferase of rats fed excess zinc and/or
injected with carbon tetrachloride (CC4).
Treatment Plasma alanine
Plasma zinc, aminotransferase
Diet CCl4 (Imol/L) log (Units/L)
Control (n= 10) 14.6 0.9*,t 2.2 0.3t
Control (n= 13) + 9.6 0.4*, 5.1 0.2*
Supplemental Zinc 17.7 0.9,t 2.4 0.31
Supplemental Zinc + 11.1 0.4" 5.3 0.2t
Data are means SEM of n= 10 rats for saline-injected animals and n= 13 for CC14-
injected rats. Values for plasma ALT activity have been transformed logarithmically
as described in Materials and Methods. Rats were fed diets containing 30 mg Zn/kg
diet (control) or 300 mg Zn/kg diet (supplemental) ad libitum for two weeks. Carbon
tetrachloride (CCl4) or saline was administered i.p. and the rats killed 24 h later. *,
indicate dietary differences; t,t relate to CC14 effects. Within treatments, unlike
superscripts within a column are significantly different (P<0.05). There were no
significant interactions between diet and injection.
C CCl4 C CCI4
Normal Zn Supplemented Zn
Effect of CCl4 and supplemental dietary zinc on liver metallothionein mRNA.
a) Densitometry of hybridization intensity of metallothionein mRNA
normalized to the hybridization intensity of B-actin mRNA to correct for
loading. Values are arbitrary units normalized to control animals fed normal
zinc (means + SD, n=4). Means with different letters are significantly
different (p < 0.05). b) Northern analysis of total liver mRNA obtained 24
h after CC14 administration. Animals were in four treatment groups as
described in Table 1. For each treatment group with at least four animals,
RNA from two representative animals per group are shown.
CRIP expression in the liver is normally low. CRIP mRNA levels in the liver of
animals injected with CC14 were increased two-fold in 24 h after the CCl4 injection (Figure 2-
2). This level was maintained up to 72 h (data not shown). Hepatic CRIP expression was
not affected by the dietary zinc intake as shown by Figure 2-2. As liver was the tissue most
affected by CC4, it was postulated that the induction in CRIP expression is a direct result
of this stress challenge.
Hematoxylin staining of the liver showed the typical pattern of cellular damage caused
by CCI4 administration including fatty acid infiltration (degeneration), e.g., cell edema and
sinusoid occlusion, and early stage necrosis (Figure 2-3b). Immunohistochemical staining
showed that, 24 h after the CC14 injection, there was an increase in the abundance of discrete
cells that were immunoreactive to the anti-CRIP antibody as well as a diffuse reaction in all
liver cells, when compared to cells from the control animals (Figure 2-3). Figure 2-3a and
c showed that mild staining occurred in cells adjacent to the central vein of each lobule in
control animals but the intensity of the stain was very much increased in discrete cells of the
same area in the carbon-tetrachloride treated animals (Figure 2-3b and d).
At higher magnification (Figure 2-3c and d), the results also revealed that the cells
with immunoreactive protein were hepatocytes with characteristic polyhedral morphology and
rounded nucleus. Some nuclear staining was also observed, but this was decreased in cells
that were further away from the central vein. Although CRIP was thought to be localized
mainly in the cytosol, under certain conditions, it could be present transiently in the nucleus.
A background level of non-specific immune reactivity was also possible.
CRIP expression was very high in immune cells such as peripheral blood
mononuclear cells. Expression in these cells did not appear to change due to either dietary
C CCI 4 C
Effect of CCl4 and supplemental dietary zinc on liver CRIP mRNA. a)
Densitometry of hybridization intensity of CRIP mRNA normalized to B-actin
mRNA to correct for loading (n=4). Values are arbitrary units normalized
to controls animals fed normal zinc (mean +_ SD). Means with different
letters are significantly different (p < 0.05). b) Northern analysis of total
RNA obtained 24 h after CCl4 administration. Animals were in four
treatment groups as described in Table 1. For each treatment group, RNA
from two representative animals per group are shown.
- -- ----
' -' ) "* :, _. t .'.. r '"',2,? :-',* .. "
, i ,,, i.. ... -ile : .
-- c .. '.
Light micrograph of rat liver. Immunohistochemical staining of CRIP in liver
from a control animal (a) and CCl4-treated animal (b) at 100x magnification,
and a control animal (c) and CCl4-treated animal (d) at 400x magnification.
Staining was carried out with rabbit anti-CRIP IgG and biotinylated secondary
antibody and streptavidin-conjugated peroxidase. Positive immunoreactivity
is indicated by darker color. Hematoxylin was used as a nuclear stain for
J 1-~ ~ V
Figure 2-3: ---continued.
CRIP mRNA in peripheral blood mononuclear cells (PBMC). The PBMC
were isolated from whole blood as described in the Materials and Methods.
Northern analysis of total PBMC RNA obtained 24 h after CC4
administration. Animals were in four treatment groups as described in Table
1. For each treatment group of at least 4 animals per group, RNA from two
representative animals per group are shown. CRIP mRNA data were
normalized by comparison to B-actin mRNA to correct for loading.
zinc or CC4 treatment (Figure 2-4). No changes in CRIP mRNA levels were seen in spleen
or thymus with CCl4 administration or with excess dietary zinc (data not shown). Similarly,
in the thymus, PBMC and spleen, there were no changes in the level of metallothionein due
to either treatment. Therefore, the inflammatory effects of carbon tetrachloride
hepatotoxicity did not appear to greatly affect the level of metallothionein or CRIP mRNA
in immune cells within a 24 h period of acute exposure.
In contrast to the liver, intestinal CRIP mRNA was decreased approximately 50% by
24 h after CCl4 treatment (Figure 2-5). Although the animals were given free access to food,
CCl4-treated rats did not consume any food during this 24 h period in contrast to control
animals (24 h). An effect of the lack of food could be that intestinal cells were being
sloughed off. This could affect CRIP expression. Therefore, the different effect observed
between the liver and the intestine could be due to the additional effects of diet on the
intestine. On the other hand, preliminary studies with showed that CRIP mRNA could be
induced at an earlier time point before the suppression at 24 h (data not shown). Two
different effects could be occurring in the intestine, one from CCl4 toxicity and one from
In animals that were fed a high zinc diet, supplemental zinc appeared to prevent the
suppression in intestinal CRIP mRNA levels after the CCI4 administration (Figure 2-5).
Since the CRIP promoter was not shown to respond to zinc (Levenson et al., 1994), it was
unlikely that zinc increased CRIP mRNA via enhanced transcription. This observation also
supported the fact that the suppression in intestinal CRIP mRNA was attributed to CCl4
toxicity as the intestinal CRIP mRNA levels in the CCl4-treated animals from the two dietary
groups responded differently even though both groups did not eat over the 24 h period.
, -.... .---
C CCI 4
Effect of CCI4 and supplemental dietary zinc on intestinal CRIP mRNA. a)
Densitometry of the hybridization intensity of CRIP mRNA normalized to the
hybridization intensity of 1-actin mRNA to correct for loading (n=4). Values
are arbitrary units normalized to control animals fed normal zinc (means _+
SD). Means with different letters are significantly different (p < 0.05). b)
Northern analysis of total intestinal RNA 24 h after CCI4 administration.
Animals were in four treatment groups as described in Table 2-1. For each
treatment group of at least 4 animals, RNA from two representative animals
per group are shown.
- -- -- -
Figure 2-6: Light micrograph of rat jejunum. Hematoxylin and eosin staining of
intestinal villi from (a) control and (b) CCl4-treated rats. Magnification at
H and E staining showed that 24 h after CCl4 administration, the intestine appeared
edematous with shortened and dilated villi (Figure 2-6). Therefore, it appeared that damage
had occurred to the intestine which could result in breakdown of the gut immune barrier.
Dietary zinc did not have a discernible effect on the morphological changes caused by CC14
(data not shown).
Densitometry on CRIP message in different brain regions; septal area, hippocampus,
diencephalon which included the hypothalamus and thalamus, brain stem and cerebral cortex,
was also carried out. CRIP mRNA appeared to be fairly abundant in the diencephalon
(hypothalamus and thalamus) compared to the other regions. The results showed that excess
dietary zinc did not have a significant effect on CRIP message in these areas although there
was a trend towards an increased expression in the cerebral cortex (Table 2-2).
CC!4 treatment significantly suppressed CRIP levels in the hippocampus and the
diencephalon. However, this effect was not altered by excess dietary zinc. There was a
trend towards increased expression of the CRIP mRNA in the septal area and the cerebral
cortex with the CCI4 treatment but while this effect was suppressed by excess dietary zinc
in the septal area, it appeared to be maintained by the zinc diet in the cerebral cortex. There
were no significant effects of the treatments on CRIP message in the brain stem.
Immunohistochemical staining of the brain with the anti-CRIP IgG showed that CRIP
was present in most neurons but not astrocytes (personal communication from Dr M.A.
King). Although CRIP was present in most parts of the brain, there were some densely-
stained areas in the medial/supraoptic nucleus of the hypothalamus, the arcuate and ventral-
medial hypothalamus, hippocampus and the lateral septal area. No change in CRIP
expression was detected by immunohistochemical staining after zinc or CC4 treatment (Dr.
Table 2-2: Densitometric analysis of the effect of excess dietary zinc and CC14 treatment
on the relative abundance of CRIP mRNA to 8-actin mRNA in different
regions of the brain.
Hippocampus 0.19 + 0.13 0.06 + 0.03a 0.11 + 0.06 0.05 + 0.04a
Septal area 0.65 + 0.16 1.05 + 0.52 0.59 + 0.24 0.58 + 0.05
Diencephalon 1.75 + 0.50 0.89 + 0.13a 1.97 + 1.32 0.78 + 0.16a
Cerebral 0.22 + 0.07 0.31 + 0.31 0.36 + 0.31 0.37 + 0.14
0.67 + 0.17 0.45 + 0.12
0.71 + 0.41
0.48 + 0.10
Hippocampus 1.0 0.3 0.6 0.3
Septal area 1.0 1.6 0.9 0.9
Diencephalon 1.0 0.5 1.1 0.5
Cerebral 1.0 1.4 1.6 1.7
Northern analysis of mRNA levels was carried out on animals fed control (Control) diet (30
mg Zn/kg diet) and supplemental (Supplemental) zinc diets (300 mg Zn/kg). Animals from
each dietary group were injected with saline or 1 ml/kg CC14 i.p. (Control/CC14 or
Suppl./CCl4) as described in Table 2-1. a) Data are means of arbitrary densitometric units
+ SD, from at least n=3 animals. Superscript a denotes significant difference from control
in each region (p < 0.05). b) Data from a) were normalized to control animals without
CC14 administration (Control) in each brain region.
M.A. King, personal communication). In addition, no obvious morphological damage was
observed in the brain slices.
In previous studies, high doses of injected zinc or dietary zinc (> 1000 mg/kg) were
found to prevent the increase in plasma ALT, an index of hepatic injury (Chvapil et al.,
1973; Clarke and Lui, 1986). In this study, we showed that a supplemental zinc diet of 300
mg/kg diet was not able to prevent liver injury as assessed by an increase in plasma ALT
The dietary requirement for zinc in the rat is 12 mg/kg and most purified diets
contain 30 mg/kg (Reeves et al., 1993). These data suggest that zinc, supplemented at the
level of 300 mg/kg diet for two weeks, while sustaining metallothionein expression at
elevated levels, was not sufficient for cytoprotection against the toxic insult. It was proposed
that high metallothionein expression was a primary factor in the protection against
CC4-induced injury (Clarke and Lui, 1986). Other mechanisms could also be responsible
for the protection. For example, the pharmacological levels of zinc used in other studies
(Chvapil et al., 1973; Clarke and Lui, 1986) could have induced other protective responses
that attenuated the CCl4-induced insult. DiSilvestro and Carlson (1994) showed that an
acute-phase response protected animals against CCl4 hepatotoxicity but this protection was
diminished with a mild dietary zinc deficiency. This suggests that the protective effect is
zinc-dependent, which supports a protective effect of zinc observed in CCl4 damage to
isolated hepatocytes (Schroeder and Cousins, 1990).
Metabolic activation of Kupffer cells may contribute to CC4-induced hepatotoxicity.
Vitamin A potentiation of the CCl4 injury was attributed to the release of active oxygen
species by Kupffer cells (Edwards et al., 1993; EISisi et al., 1993). However, release of
soluble mediators by C. parvum-stimulated Kupffer cells was found to suppress the
cytochrome P450IIE1 system and to diminish the CC4-induced liver injury (Raiford and
Current views on carbon tetrachloride-induced hepatotoxicity suggest that a sublethal
acute dose of CC4, such as that used in this study, initially causes liver damage, but
subsequently allows tissue regeneration to occur. This regeneration is characterized by
increased cell proliferation and repair (Mehendale et al., 1994). Stimulation of tissue repair
appear to occur within 24 h after the CC14 administration as a protective effect is observed
when a lethal dose is administered 24 h after the protective sublethal dose (Mehendale et al.,
1994). Maldonado et al. (1994) reported that hepatic apo B mRNA, which was important
for normalization of lipid metabolism during liver regeneration, was increased 24 h after
CCL4 administration. Intestinal apo B mRNA was also increased but the increase was much
CRIP mRNA, as well as CRIP-immunopositive cells, were increased in the liver 24
h after CC4 administration (Figure 2-2 and 2-3) which was the time course of tissue repair
reported by Mehendale et al. (1994). This information suggests that CRIP could be
important for maintenance of the function of the differentiated cells after they are
regenerated. Increased CRIP-immunopositive cells near the central lobular veins of the liver
were fully differentiated hepatocytes and could have been in the process of being shedded.
Alternatively, it is also possible that, due to the CC14 treatment, these hepatocytes have
differentiated into cells involved in secretary processes (Dr. D. Samuelson, personal
communications). This could function to suppress further mediators of inflammation. CRIP
has recently been found in the plasma (as will be discussed in Chapter 3). Consequently,
increased blood flow to the site of injury and/or increased capillary permeability may allow
CRIP to infiltrate into the tissue and cause the diffuse immunoreactivity seen in the
The intestine has been shown by several studies to be directly injured by the CCl4-
induced insult as well as to contribute to the pathology observed in the liver (Miura et al.,
1989). In this study, histological data confirmed those findings (Figure 2-6). CRIP has been
found to developmentally regulated in a way that correlated to changes in Paneth cell activity
and antimicrobial protection in the intestine which occurs just before weaning (Levenson et
al., 1993). CRIP was found by immunohistochemistry to be localized in Paneth cells
(Fernandes et al., 1995). These cells were found to have secretary and immune functions
(Erlandsen et al., 1976). The marked CCl4-produced reduction of intestinal CRIP expression
could represent a loss of differentiated intestinal cells and, in turn, could represent a loss of
an important functional barrier (Figure 2-5). Dietary zinc could play an important role in
the maintenance/recovery of the intestine from a stress challenge as it was able to normalize
CRIP expression after the insult (Figure 2-5).
The expression of CRIP mRNA in different regions of the brain after CC14
administration showed that the gene was differentially regulated within the brain as well
(Table 2-2). It was interesting to note that in the hippocampus, CRIP expression was
suppressed with CC14 treatment while in the septal area, it was unchanged or slightly
increased. It is not possible to explain the reason for the differential expression of CRIP
since the function is still unknown and also because of the lack of more information about
the effect of CC14 on the different brain regions. The brain is affected by toxicants such as
ethanol and CCI4 based on studies that described changes in brain morphology, microglial
activation, and changes in the levels of various neurotrophic and neurotoxic factors (Dfaz-
Mufioz and Tapia, 1989; Hansson et al., 1990).
In studies using primary cell culture, the septohippocampal, cerebral cortex and
cerebellum regions were found to be most susceptible to damage by ethanol as neurons
cultivated from these regions did not proliferate after ethanol treatment (Heaton, 1994) while
the hippocampus was not as severely affected (Walker et al., 1993). On the other hand,
Watanabe et al. (1986) found that CCl4 did not affect the brain directly when administered
systemically. Therefore, the suppression of CRIP mRNA in the hippocampus (Table 2-2)
could be due to a secondary effect of hepatic injury. This was supported by the failure to
observe tissue damage histologically.
As far as the effect of zinc is concerned, zinc uptake has been shown to vary among
the different regions of the brain, with the hippocampus having the highest rate of uptake and
the cerebellum, the lowest (Franklin et al., 1992; Pullen et al., 1991). The highest
concentration of zinc is in the hippocampus (Pullen et al., 1991). Therefore, it is not
surprising that supplemental dietary zinc had different effects on CRIP expression in the
various regions. Zinc concentration in the brain was found to be very tightly regulated
(Frederickson, 1989), suggesting zinc is important for specialized functions. Consequently,
the regulation of zinc-binding proteins, such as CRIP, in the brain could be important as
From the pattern of the immunohistochemical staining, it is also obvious that the
abundance of CRIP mRNA obtained from the different brain regions could be different
depending on the purity of the tissue sections taken (Dr. M.A. King, personal
communication). For example, in the septal area, immunohistochemistry showed that CRIP
was highly expressed in the lateral septum, which receives zinc-containing fibers from the
hippocampus (Heaton et al., 1994), in contrast to the low expression in the medial septal
area, which projects cholinergic fibers to the hippocampus (Saporito et al., 1993).
Therefore, it is possible that CRIP expression was induced in one area and suppressed in
another. This could mask a real change in the expression of CRIP in a particular area. For
example, it was not possible to separate the medial septal area from the lateral septal area for
Northern analysis due to proximity of their location and their size. However,
immunohistochemical staining showed that the distribution of CRIP was not uniform in these
two areas (Dr. M.A. King, personal communication).
There were no changes in the expression of CRIP in immune tissues and cells,
including PBMCs, spleen and thymus. The data suggest that these immune tissues are not
affected in the same way as the liver when challenged with CC14. This was further
confirmed by a lack of induction of metallothionein in these tissues (data not shown). On
the other hand, when lipopolysaccharide was administered, as will be discussed in Chapter
3, CRIP was induced in the immune cells and tissues but not in the liver. Therefore, it
appears that the toxic challenge by CC14 at this sublethal dose affects the liver, brain and
intestine but not the immune cells. CRIP expression was affected accordingly. These results
suggest some tissue specificity since immune cells, with highly abundant levels of CRIP
mRNA, showed very little change in expression following CCL4 treatment.
This study was the first to show tissue-specific regulation of CRIP due to toxic stress.
It also shows that CRIP is not involved in the acute phase response but could be important
in the maintenance of cellular function at the later stages of the inflammatory response when
tissue regeneration and repair is occurring. Although this study did not provide answers to
the possible function of CRIP, it provided a time frame and a model to study CRIP regulation
in liver cells.
The following study with LPS enabled further development of a system to study CRIP
expression. The LPS challenge differs from that of the CCI4 insult in that it is an immune
challenge that targets immune cells first before causing other tissue injury. The immune cells
have abundant CRIP mRNA compared to the liver, and are activated when challenged.
STRESS CHALLENGE (11)
CRIP expression is very high in intestine, compared to other tissues investigated,
suggesting a role for CRIP in that organ. Recently, very high levels of CRIP mRNA were
found in peritoneal macrophages. The CRIP promoter has been found to have putative
response elements with high homology to consensus sequences for cytokines (Levenson et
al., 1994a). As shown in the previous chapter, CCi4 was able to induce CRIP expression in
the liver, a tissue with low CRIP abundance. However, CRIP expression in immune tissues
was not affected by this challenge. Consequently, CRIP mRNA expression and protein levels
were investigated in the immune system, intestine and liver in response to challenge with
lipopolysaccharide (LPS) to stimulate an acute-phase immune response. The purpose of this
study was to examine changes in CRIP expression due to perturbation of the normal
physiologic state by an immune challenge. Since the intestine has high CRIP expression and
is the largest immune organ in the body while the brain has low CRIP expression and is
considered immunologicallyy privileged", CRIP expression in both these organs was studied
as well. As discussed in the literature review, there have been many studies on the
contributions of the intestine and central nervous system to the immune response (Castro and
Arntzen, 1993; Quan et al., 1994; Xu and Ling, 1994).
The possibility that CRIP is present in the plasma was also investigated due to the
presence of a putative signal sequence for secretion (Tsonis et al., 1988) and some
experiments, as described in the next chapter, showing the binding of CRIP to serum
Materials and Methods
Sprague-Dawley rats (Harlan-Sprague Dawley, Indianapolis, IN) were injected i.p.
with LPS (E. coli 0127:B8; Difco Lab., Detroit, MI) at a concentration of 1 /pg/g body
weight at various time points (6, 12, 24, 48 and 72 h) prior to killing. Control animals were
injected with saline at 24 h. Plasma zinc levels in plasma were determined by atomic
absorption spectrophotometry (Blalock et al., 1988). Decreased plasma zinc and induced
liver metallothionein MT mRNA were used to confirm the efficacy of the LPS injection.
These responses to LPS have been observed previously in a variety of species (Etzel et al.,
1982; Klasing, 1984; De et al., 1990). Expression of CRIP protein in intestine and plasma
was also examined with control, 6 and 48 h LPS-treated rats.
Plasma, PBMC and intestine were collected from rats at indicated time points after
LPS injection. Plasma and PBMC were collected by Histopaque-1077T gradient
centrifugation (Sigma Chemical Company, St. Louis, MO) as described in Chapter 2,
Materials and Methods. To collect peritoneal macrophages, 30 ml PBS in a 50 ml syringe
was injected into the rat peritoneum with a 16-gauge needle, after MetofaneT anesthesia.
The peritoneal solution was then drawn back into the syringe and the solution centrifuged at
300 x g for 30 min at 4C to collect the cells.
Cytosolic protein from the intestine and PBMC were prepared as described previously
(Hempe and Cousins, 199.1). The proteins were resolved by high resolution gel filtration
chromatography, Superdex 75 Hiload 16/60 for the plasma proteins, and two smaller
Superdex 75 columns in tandem for the PBMC cytosolic proteins (Pharmacia Biotech,
Piscataway, NJ) as described previously (Khoo and Cousins, 1994). Fractions containing
CRIP were pooled and concentrated with Centriprep-3 and Centricon-3 concentrators
(Amicon, Inc., Beverly, MA). Total protein was determined by Lowry (1951) and cell
protein (75-300 jg/lane as indicated) was electrophoretically separated by SDS-PAGE (15%
gel) as described previously (Khoo and Cousins, 1994). Protein was transferred onto 0.2 pim
nitrocellulose membranes (Schleicher & Schuell, Keene, NH) for zinc blot and Western
analyses (Hempe and Cousins, 1991).
Plasma proteins for microsequencing were semi-purified by gel filtration (Superdex
75 Hiload 16/60) and ion exchange chromatography (CM-Sepharose Fast Flow, Pharmacia).
These chromatography steps will be described in more detail in Chapter 4 (Materials and
Methods). CRIP-containing fractions were pooled, electrophoresed by SDS-PAGE (15%
discontinuous tris-tricine gel) and transferred onto 0.4 Am PVDF membranes (Khoo and
Cousins, 1994). The proteins were stained with 0.1% Coomasie Brilliant Blue R250 (Bio-
Rad Laboratories, Richmond, CA) and the appropriate band sent for microsequencing at the
University of Florida ICBR Protein Core, as described previously (Hempe and Cousins,
Total RNA was extracted from peritoneal macrophages, PBMC, spleen, thymus,
intestine, brain and liver by the guanidinium isothiocyanate method of Chomczynski and
Sacchi (1987). Northern analysis was carried out as described in Chapter 2 (Materials and
Methods) except that cDNA from human glyceraldehyde-3-phosphate dehydrogenase
(G3PDH) (pHcGAP, ATCC #57090) and/or mouse 18S rRNA (pN29III, ATCC #63178)
were used when B-actin cDNA was not suitable as a control. Data were normalized to actin
and/or G3PDH cDNA.
Zinc Blot Analysis
The procedure was carried out as previously described (Hempe and Cousins, 1991).
Briefly, nitrocellulose membranes were equilibrated for 2 h with blocking buffer containing
10 mM Tris-HCl (pH 7.5), 1 mM MnCl2 and 10 mM B-mercaptoethanol, a reducing agent.
The high level of 8-mercaptoethanol increased specificity of binding of the radiolabeled zinc
to CRIP, as shown previously (Hempe and Cousins, 1991). Membranes were submerged for
15 min in blocking buffer with 1 ACi/ml of 6Zn (Specific activity = 3.2 /Ci/Utg, Du Pont
NEN, Boston, MA) and washed three times before autoradiography.
Western Analysis and Immunocytochemistry with CRIP Antibody
Nitrocellulose membranes were rinsed with TBS-T (10 mM Tris, 154 mM saline,
0.1% Tween, pH 7.5). The membranes were incubated at room temperature for 2 h with
10% non-fat dry milk in TBS-T to prevent nonspecific binding. The membranes were then
incubated for 1 h at room temperature with rabbit IgG against a CRIP peptide (amino acid
residues 31-56) conjugated to keyhole limpet hemocyanin and incubated with secondary
antibody (1:1000), donkey anti-rabbit IgG linked to horseradish peroxidase (Sigma Chemical
Co., St. Louis, MO) for 1 h. The membranes were washed repeatedly with TBS-T after
each incubation step. Detection reagents from ECL Western Blotting Detection Kit
(Amersham Life Science, Arlington Heights, IL) were mixed and incubated at room
temperature with membranes for 1 min. This method is a non-radioactive method for
detection of immobilized antigen using a specific primary antibody and a horseradish
peroxidase-conjugated secondary antibody. Oxidation of a chemical compound, luminol,
catalyzed by the horseradish peroxidase/hydrogen peroxide reaction, causes an emission of
light, which is detected by the autoradiographic film. Membranes were immediately exposed
to film for 15 sec 10 min.
Peripheral blood mononuclear cells were plated on gelatin-coated coverslips for 24
h in medium with 10% serum. Non-adherent cells were washed with PBS and aspirated off.
The cells were fixed with 2% paraformaldehyde and quenched with methanol containing
0.3% hydrogen peroxide for 30 min to inhibit endogenous peroxidase activity. The adherent
cells were then analyzed by immunocytochemistry with rabbit antibody raised against
synthetic peptide representing amino acids 31-56 of rat CRIP, using the Histostain-SPT kit,
as described in Chapter 2 (Materials and Methods). The cells were incubated with anti-CRIP
IgG in 10% non-immune goat serum according to the protocol included with the kit, as
described earlier. These sections were counterstained with hematoxylin.
Data from plasma zinc concentration measurements were analyzed by analysis of
variance over time after LPS injection (General Linear Models, SAS Institute, Inc., Carey,
NC). Values with p < 0.05 were considered statistically significant.
Changes in the level of plasma zinc and liver metallothionein mRNA showed that the
LPS injections were effective in stimulating an acute-phase immune response. Plasma zinc
concentration dropped to 26% of control levels at 6 h then returned to control levels by 48 h
(Figure 3-la). Liver metallothionein mRNA was increased at 6-12 h then returned to normal
by 48 h, as characteristic of an acute phase response (Figure 3-1b). In PBMC,
metallothionein mRNA expression was barely detectable in control rats or in rats 6 h after
LPS injection. Thereafter (12-72 h), the level of metallothionein mRNA in PBMC was
markedly increased (Figure 3-1c). By flow cytometric analysis (University of Florida core
facility), granularity and cell size of PBMC samples showed that they contained three
different populations of white blood cells which were likely to be lymphocytes, monocytes
and granulocytes (data not shown). During the acute phase response between 6-48 h, the
proportion of granulocytes appeared to increase in proportion to the population of white
blood cells (data not shown). The pattern and distribution of PBMC returned to control
levels by 72 h. Spleen and thymus metallothionein mRNA (data not shown) was increased
in a similar pattern as PBMC.
In a survey of CRIP mRNA levels in cells and tissues, expression of CRIP mRNA
in peritoneal macrophages was found to be exceptionally high and comparable to the level
seen in intestine under normal conditions (Figure 3-2). PBMC also contained a substantial
amount of CRIP mRNA while thymus, spleen and lung expressed CRIP mRNA at a moderate
level. In tissues that expressed the lowest level of CRIP mRNA, brain, kidney and liver, the
message could only be demonstrated with a longer exposure of film to the blot (Figure 3-2).
The level of CRIP mRNA in peritoneal macrophages increased over time after LPS
injection (Figure 3-3). In contrast, CRIP mRNA in PBMC was decreased between 6-12 h,
and then increased markedly with a maximum obtained by 48 h after LPS injection, before
returning to control levels by 72 h (Figure 3-4). This was different from the pattern of
LPS injection altered plasma zinc concentration and MT mRNA level in liver
and peripheral blood mononuclear cells (PBMC). Rats were saline-injected
(0 h) or injected with 1 pg LPS/g body weight (i.p.) for 6, 12, 24, 48 and 72
h. (a) Plasma zinc, determined by atomic absorption spectrophotometry, at
various time points. Plasma zinc data (mean pooled SEM [n = 10-13])
were analyzed by analysis of variance. Means with a different superscript are
significantly different (p < 0.05). Northern blots of (b) liver, and (c)
PBMC. Total RNA was extracted and used for Northern analysis (20
IAg/lane) with a 32P-rat metallothionein-1 probe. Northern data from two
individual animals are representative of 6 animals/time point. MT mRNA
data were compared to B-actin mRNA.
0 6 12 24
Actin mRNA *amboM
S ac ,
: ". 4 ^
^^Hl ~ ~ U 4-. 9 '
.P 0 S 0
0 6 12 24 48 72
Effect of LPS injection on CRIP mRNA in peritoneal macrophages. Rats
were treated as described in Fig. 3-1. Northern analysis (10 /Ag/lane) was
carried out as described in Fig. 3-2. Northern data from each individual
animal/timepoint are representative of 1-4 animals/time point. CRIP mRNA
data were normalized by comparison to B-actin mRNA.
0 6 12 24
Effect of LPS administration on CRIP mRNA in PBMC. Northern data from
two individual animals/timepoint are representative of 6 animals/time point.
CRIP mRNA data were normalized by comparison to G3PDH mRNA.
PBMC metallothionein mRNA expression which had a sustained increase up to 72 h after
LPS administration (Figure 3-1). CRIP protein was not detectable by Western analysis of
PBMC cytosolic protein as the total amount of protein isolated from these cells was low (data
not shown). Therefore, immunocytochemistry of adherent cells from PBMC, containing
primarily monocytes, was carried out. A control reaction with non-immune serum was not
immunoreactive (Figure 3-5a) but incubation with anti-CRIP IgG showed that CRIP protein,
as indicated by the immunoreactivity (dark color), was present in the cytoplasm of these cells
In contrast to PBMC, CRIP protein was detected in plasma of normal rats by both
zinc blot (Fig 3-6a) and Western analysis (Figure 3-6b). Results of the Western immunoblot
analysis showed markedly lower CRIP protein in the plasma of rats injected with LPS for
6 and 48 h than in plasma of control rats. Analysis by zinc blotting, a high affinity detection
method for CRIP, was not able to detect the changes in the protein (Figure 3-6a). Use of
this method depends on the availability of metal-binding sites on the immobilized protein.
The presence of CRIP protein in the plasma was confirmed by protein purification and amino
acid sequencing. It was clear from the sequencing data, however, that CRIP was not the
most abundant protein in the purified protein band after gel electrophoresis (data not shown).
Nevertheless, this is the first evidence that showed the presence of CRIP in plasma, and the
possibility that cellular CRIP is secreted under these conditions. This possibility, however,
needs to be investigated further.
In the intestine, like in the PBMC, CRIP mRNA was most abundant at 48 h after
LPS (Figure 3-7a). There was little change in the CRIP mRNA level at 6 h. CRIP protein
level in intestinal cytosol when analyzed by the Western blot was observed to be dramatically
Immunocytochemistry showed cytosolic CRIP in PBMC. PBMC were
obtained by gradient centrifugation and immunocytochemistry was carried out
as described in the methods using a rabbit IgG prepared against a rat CRIP
peptide. A dark color indicates CRIP protein; (a) nonimmune rabbit serum
as a control and (b) nonimmune rabbit serum plus rabbit IgG against rat
u .- w
3, .u c
11 a C
oo (JoS -.e .
J ^ T I-^I
-1 ^ o LL
3 li a
decreased at 6 h, similar to the decrease seen for plasma CRIP. By 48 h after LPS, the
intestinal CRIP protein level had returned to normal in contrast to the lack of a return to the
control level seen in plasma CRIP (Figure 3-7b).
LPS did not appear to change the level of CRIP mRNA in liver or thymus while in
the spleen, CRIP mRNA showed a slight, but consistent, increase between 12-72 h (data not
shown). CRIP mRNA levels also increased at 48 h in the brain and more specifically, the
hippocampus (Figure 3-8a and b), similar to the pattern found in the other tissues.
Liver metallothionein mRNA level increased upon LPS stimulation as expected.
Increased metallothionein expression in animals exposed to various types of physiological
stresses has been well described (Dunn et al., 1987; Sato and Bremner, 1993). The increase
in metallothionein during immunostimulation is thought to either protect cells against
oxidative damage due to the oxygen radicals produced by PBMC effector cells attracted to
the site of infection or is needed for another zinc-requiring process. Actin mRNA levels in
the PBMC changed in response to LPS administration (Figure 3-1) since actin synthesis was
altered in these cells for cell remodelling (Goldblum et al., 1993).
The constitutively high basal level of CRIP mRNA in peritoneal macrophages
strongly suggests that CRIP is important in immune cell function. Since expression of CRIP
increased in response to antigenic challenge by LPS, CRIP could have a role related to
activation of macrophages since residential peritoneal macrophages differentiate into activated
macrophages upon immunostimulation (Johnston, 1988). Zinc has also been shown to be
involved in immunomodulation and thus, may be regulated by CRIP during this response
(Driessen et al., 1995a). CRIP is probably not the only zinc finger protein that is increased
The changing level of CRIP expression in PBMC over time after LPS injection with
decreased expression during the early acute-phase immune response and increased expression
at the later phase suggests that CRIP plays a role in immune cells during the response. Since
the increased level of CRIP mRNA in PBMC occurs late in the immune response coincident
with the repair/regeneration stage of the acute-phase response (Kasama et al., 1993), CRIP
could be involved in maintenance of healthy, differentiated cells and tissues or in
proliferation of new cells.
From flow cytometry, it was found that the subpopulation of cells collected in the
mononuclear fraction of the blood was altered during the response to LPS (data not shown).
This observation was based on the change in the number of cells in different populations
defined by their granularity and size. Change in the subpopulations of PBMC or transient
expression of different genes within the same population could account for the changes in
PBMC CRIP mRNA level seen during the immune response. The CRIP promoter has
response elements for glucocorticoid hormone and consensus sequences for cytokines
(Levenson et al., 1994a). Therefore, the production of CRIP by PBMC could also be
suppressed by cytokines or other factors that are associated with the immune response.
Intestinal CRIP mRNA was induced over control levels after 48 h. The intestine is
considered an important immune tissue since 60-80% of the body's lymphoid tissue is
gut-associated. It is able to respond to acute-phase mediators during an immune response
(Castro and Arntzen, 1993). Ogle and colleagues (1994) showed that isolated enterocytes
are capable of participating in an immune response by producing tumor necrosis factor,
interleukin-1 and interleukin-6 at levels comparable to that of LPS-activated intestinal
macrophages. Other researchers have also shown that the epithelial cells are stimulated by
LPS to secrete cytokines (Santos et al., 1990). Intestinal epithelial cells have also been found
to elicit an acute phase immune response by expressing acute phase plasma proteins such as
ai-antitrypsin, transferring and complement factors (Molmenti et al., 1993). Increased
intestinal CRIP expression upon immune challenge suggests that intestinal cells respond to
the antigenic stimulus and that CRIP could be involved in the intestinal response to injury.
Challenge with CCd4 caused damage to the intestine. Histological studies showed that tissue
damage had occurred after CC14 while damage was not clearly observed with LPS challenge.
CRIP mRNA in intestine appears to have responded differently to the toxic and immune
challenge. Therefore, the time frame of intestinal injury and the type of injury, as well as
the response to injury, caused by the two challenges appear to be different.
The abundance of CRIP protein in the intestine and the dramatic decrease in CRIP
6 h after LPS treatment, as shown by Western analysis (Figure 3-6), also suggests that
intestinal CRIP expression was not due to immune cells. The expression of CRIP in
macrophages was observed to be steadily increasing from 6 h, with no decrease after LPS
administration. The later increase in intestinal CRIP mRNA could be important to the
cellular response for repair or restoration of normal gut tissue.
CRIP has been purified to homogeneity from the intestine (Khoo and Cousins, 1994),
and the protein has been localized to the Paneth cells (Fernandes et al., 1995). The Paneth
cell is also the site where intestinal defensins, lysozyme and cytokines are produced to
provide an antimicrobial and antiparasitic barrier to the mucosa (Evans et al., 1992; Selsted
et al., 1992). Localization of CRIP in these cells as well as its presence in immune cells
suggest that CRIP could play a role in cellular maintenance. Paneth cells are terminally
differentiated cells, again suggesting that CRIP has a role in maintenance of the differentiated
state of these and also immune cells. No studies have been done to show the effect of
endotoxin treatment on Paneth cell physiology. The Paneth cells could be a major
contributor to intestinal CRIP protein expression after the LPS challenge. Previous studies
have determined that changes in mRNA levels occur in the midvillar region during intestinal
cell development (Levenson et al., 1993). Protein levels, however, are higher in the Paneth
cells and brush-border enterocytes at the tip of the villi.
The presence of CRIP protein in the plasma of control rats raises some questions as
to the source and target tissues of CRIP. Although most plasma proteins are produced in the
liver (Schreiber et al., 1989), CRIP mRNA is expressed at very low levels in the liver
suggesting that plasma CRIP protein is not of hepatic origin. PBMC expressed high levels
of CRIP mRNA and CRIP protein has been detected in PBMC cytoplasm. Consequently,
PBMC could secrete CRIP protein into the blood since CRIP was found to have a unique
signal sequence common to some exported proteins (Tsonis et al., 1988). This putative
consensus sequence was found to be 19 amino acid long and included the metal-binding
domain of CRIP, beginning from the Cys residue at position 28. The presence of plasma
CRIP in normal animals and the suppressed levels shortly after LPS challenge, corresponded
to the pattern of many negative acute-phase proteins, such as albumin (Schreiber et al.,
1989). Decreased plasma CRIP could be due to passive diffusion out of the plasma by
increased transendothelial flux across capillaries (Goldblum et al., 1993). This has been
shown to occur for albumin after LPS treatment. This pattern of acute-phase response is
well-documented for plasma zinc, a characteristic response, believed to be part of the host
defense mechanism (Dunn et al., 1987).
In the brain, CRIP expression was induced after 48 h which was similar to that found
in the other tissues. In the previous chapter, it was shown that CRIP mRNA responded
differently in different brain regions due to CCI4 challenge. As discussed previously,
different regions have varying susceptibilities to the challenge. Therefore, the response of
CRIP mRNA could reflect the extent of the injury or the lack of it, especially when
challenged with these two different stresses. The difference between the effects on the
hippocampus observed for CCl4 insult and the LPS challenge could be due to the ability of
these compounds to cross the blood-brain barrier, the severity of the trauma and the cell type
that responded to each challenge.
LPS has been shown by some studies to affect the brain. Researchers found that LPS
administration caused minimal inflammatory reaction suggesting that the brain had regulatory
mechanisms to reduce damaging inflammatory reaction (Montero-Monei et al., 1994). LPS
stimulated expression of MHC class I and II antigens on microglia cells (brain macrophages)
in the brain to present antigens for the immune response (Xu and Ling, 1994). However,
peripheral macrophages did not appear to be able to reach the site of injury unless the blood-
brain barrier was breached such as in ischemia (Giulian and Vaca, 1993; Montero-Monei et
al., 1994). LPS also downregulated interleukin-I receptors in the hippocampus but not in
the pituitary (Ban et al., 1993) showing that different regions of the brain were differentially
Acute phase reactants such as IL-1 have been found in the brain as early as 5-6 h
after LPS administration (Ban et al., 1993; Quan et al., 1994). At the later time points,
events that occur after injury include secretion of glucocorticoids to inhibit inflammatory
cytokines (Chrousos, 1995), secretion of neurotoxic substances to cause delayed neuronal
damage, activation of brain microglia cells to scavenge cellular debris (Giulian and Vaca,
1993; Xu and Ling, 1994), and the production of neurotrophic factors to initiate the healing
process (Giulian and Vaca, 1993; Montero-Monei et al., 1994). The late increase in the
expression of brain CRIP mRNA, like that found in the intestine and the PBMC, suggests
that it could function in the late stages of cell repair/regeneration. One of the processes
mentioned above could be important in regulating or contributing to the expression of CRIP
in the brain.
In summary, CRIP was found in abundance in leukocytes and in the Paneth cells of
the intestine. These are cells with many similar characteristics, such as having a high
turnover rate, and having a highly differentiated state. In cells that have low turnover rates,
such as the liver cells, CRIP mRNA expression was low until the system was perturbed by
toxic stress. Tissue injury occur as a result of this challenge, followed by tissue
regeneration. Under these conditions, CRIP mRNA and protein was induced. In this study,
LPS caused a decrease in CRIP mRNA in PBMCs in the early stages of the immune
response. However, at 48 h, when the immune stress was being downregulated, and the
tissues were undergoing, or had undergone repair, CRIP mRNA was induced.- CRIP could
be involved in maintenance of a healthy cellular state after the immune and toxic challenge
in those tissues. Since the level of CRIP in the plasma was altered upon immunostimulation
and plasma zinc was altered in a similar temporal pattern as plasma CRIP, the function of
CRIP could also be related to the function of zinc during an acute-phase immune response.
The cysteine-rich domain of CRIP is arranged in a sequence that is described as the
LIM motif; Cys-Xaa2-Cys-Xaal7.l9-His-Xaa2-Cys-Xaa2-Cys-Xaa2-Cys-Xaa7T.-(Cys)-Xaas-Cys
(Freyd et al., 1990). The LIM motif may be involved in protein-DNA, protein-protein or
protein-RNA interactions. The LIM proteins that have homeodomains are transcriptional
regulators while the LIM proteins without homeodomains (LIM-only) are postulated to be
involved in cell signalling/aggregation through LIM-LIM interactions (Crawford et al., 1994;
Dawid et al., 1995; Feuerstein et al., 1994; Melamed et al., 1995; Sanchez-Garcia et al.,
1993; Schmeichel and Beckerle, 1994). Since CRIP does not have a homeodomain, it is
hypothesized that CRIP forms protein-protein interactions in carrying out its function.
In this chapter, a study was carried out to look for a putative partner for CRIP that
could define a possible area to study the function of CRIP and that could support the
information obtained in the previous studies. Since immunohistochemistry with the CRIP
antipeptide IgG showed a predominantly cytosolic localization for CRIP (Fernandes et al.,
1995), rat intestinal cytosolic proteins were used as the primary source to isolate a protein
partners) for CRIP. A crude membrane fraction and a nuclear fraction were also prepared
for preliminary studies to identify a potential protein partnerss.
In this study, we decided to look for potential partners for CRIP in cytosolic,
membrane and nuclear fractions using several, different methods. It was hypothesized that
the protein partner of CRIP may have similar, complementary or opposing roles in the cell.
Identification of the partner and its cellular location would help to further elucidate the
function of CRIP.
The methods used in this study included: i) overlay assay with ["S]Cys-rrCRIP as
a probe or alternatively, the sandwich overlay using both CRIP, followed by the anti-CRIP
peptide IgG antibody, as probes, and ii) cyanogen bromide affinity chromatography with
CRIP coupled as a ligand to the column. The overlay method was successfully used to study
interactions between LIM proteins (Crawford et al., 1992), and cAMP-dependent protein
kinase anchoring (Carr and Scott, 1992), while affinity chromatography has been used in
many different immunologically-based purification schemes for antigen-antibody complexes.
Other strategies that were used to identify protein-protein interactions included an in vivo
dihybrid assay utilizing the catalytic and transactivational domain of GAL4A (Guarente, 1993;
Le Douarin et al., 1995), library screening methods, immunoprecipitation and surface
plasmon resonance. These methods were extensively reviewed by Phizicky and Fields
(1995). Wadman et al. (1994) used the dihybrid method to study an specific association
between rhombotin (a LIM protein) and basic helix-loop helix proteins. Crosslinking
followed by gel filtration was another method used successfully to study binding of
neuropeptide Y to intestinal membrane proteins (Nguyen et al., 1990). Variations of these
methods were utilized successfully to study other types of protein binding. Some of the
advantages and disadvantages of the methods used in this study are as follows (Phizicky and
Table 4-1. Advantages and disadvantages of the overlay assay and affinity chromatography.
1. Cell lysate can be used
without prior purification.
2. Direct interaction can be
determined between two
1. Sensitive method-can detect
2. Able to bind its partner from
a complex mixture of proteins-
the partner successfully
competes with other proteins in
the extract to bind to the ligand
on the column.
3. Multi subunit interactions
can be detected.
1. Proteins may be inactivated
by SDS PAGE and reducing
2. Complexes are not detected,
only direct interactions.
1. Due to its sensitivity, false
positives can be obtained.
2. Interactions detected may not
be direct and may be through a
second protein present in the
___________________ I _________________________________ __________________________________
The putative CRIP partners that were identified using these two initial methods were
characterized by sequencing and CRIP binding activity was confirmed by
immunoprecipitation or crosslinking (Flow chart, see Appendix A). These latter methods are
not as sensitive as the two initial methods but have been found. to be useful in confirming
interactions between proteins as they are specific when used with appropriate controls.
Materials and Methods
Intestinal Cytosol and Membrane Protein Preparations
Intestines were excised from rats anaesthetized with methoxyflurane (Metofane;
Pitman-Moore, Mundelein, IL) and killed by exsanguination. The intestines were flushed
with ice-cold saline, placed on an ice-cold plate and the mucosa scraped off with two glass
slides. The mucosa was homogenized in a homogenization buffer (10 mM Tris, 154 mM
NaCI, 10 jtM ZnSO4, 1 mM MnCl2, 2mM phenylmethanesulfonyl fluoride, 0.9 pg/ml
pepstatin, 0.6 4g/ml leupeptin and 10 mM 2-mercaptoethanol, pH 8.0), with a Potter-
Elvejhem glass-teflon homogenizer. The homogenate was centrifuged at 12 500 rpm (12 000
x g) for 30 min at 4C using the Sorvall SS-34 rotor (Kracht, et al., 1993). This supernatant
fraction was centrifuged at 40 000 rpm (100 000 x g) for 1 h in a Beckman 50 Ti rotor to
obtain the cytosolic fraction. Membrane proteins were obtained as a pellet. The pellet was
mixed with 2 mg CHAPS/mg protein in buffer and shaken for 30 min at 4C. The
solubilized protein was obtained in the supernatant fraction by centrifugation at 100 000 x
g for 1 h. These protein fractions were then aliquoted and stored at -700C until used.
Preparation of Nuclear Extract
Nuclear extract was obtained from a mouse macrophage-monocyte cell line RAW
264.7 which was found to produce abundant amounts of CRIP mRNA. The cells were
grown to confluency in four 150 cm2 flasks in DMEM medium, 10% FBS, supplemented
with glutamine. The cells were harvested by centrifugation for.10 min at 2 500 rpm (1 000
x g) at 4C using Sorvall H1000B swing bucket rotors and cell nuclei extracted according
to a modified procedure described below (Cousins and Lee-Ambrose, 1992; Gorski et al.,
1986; Shapiro et al., 1988). The pelleted cells were resuspended in 10 ml homogenization
buffer (10 mM HEPES, 0.5 mM spermidine, 0.15 mM spermine, 25 mM potassium
chloride, 2 M sucrose, 10 % glycerol, pH 7.6) with 0.1 mM PMSF and 1 mM DTT and the
cells broken with 5-8 strokes of Dounce glass homogenizer (pestle A). The homogenate was
diluted to 85 ml with homogenization buffer. Aliquots of 27 ml each were layered on 10 ml
homogenization buffer and centrifuged for 30 min at 18 000 rpm (40 000 x g) using the
Sorvall SS-34 rotor at 4C. The pellets were combined in 50 ml of buffer consisting of 9:1
homogenization buffer:glycerol and rehomogenized. Two aliquots of 25 ml each were
layered on 10 ml of homogenization buffer and centrifuged at 18 000 rpm for 30 min at 4
C as above. The pellets were resuspended in 1 ml resuspension buffer (20 mM HEPES,
0.75 mM spermidine, 0.15 mM spermine, 25% glycerol, 2 mM DTT, pH 7.9).
The cell nuclei were counted with a hemacytometer using Trypan Blue and diluted
to 1 X 10' nuclei/3 ml of resuspension buffer. The extract was rocked for 30 min on ice
followed by ultracentrifugation with the Beckman 50Ti rotor for 90 min at 49 000 rpm (150
000 x g) at 40C. Ammonium sulfate (0.33g/ml) was added to the supernatant and the
mixture rocked for 15-20 min on ice to dissolve the ammonium sulfate. The protein
precipitate was collected by ultracentrifugation at 40 000 rpm (100 000 x g, 50Ti rotor) for
20 min at 4C and dissolved in 1 ml dialysis buffer (20 mM HEPES, 100 mM KCI, 20%
glycerol, 2 mM DTT, pH 7.9). The sample was dialyzed against two changes (1.5 h each)
of 200 x volume dialysis buffer. Aliquots were stored at -700C until use.
The proteins were resolved by SDS PAGE as described previously (Khoo and
Cousins, 1994). Gels of 12% and 15% were used to detect high and low molecular weight
proteins, respectively. Proteins were transferred onto Immobilon pTM membranes (Millipore,
Bedford, MA), as described previously, for both overlay analysis and microsequencing (Khoo
and Cousins, 1994). Microsequencing was carried out at the University of Florida ICBR
Protein Core, as described in Chapter 3 (Materials and Methods).
I3S1Cys-rrCRIP Preparation (Appendix B)
Recombinant CRIP expressed in the pET pLysS vector with the E.coli host BL21
(DE3) was labeled with [135S-Cysteine (Spec. activity 22.2 TBq/mmol, NEN, Boston, MA)
to obtain the radiolabeled protein. The bacteria was streaked on minimal medium plates
made with M9 medium (M9 salts, 1.0 mM MgSO4.7H20, 0.2% glucose, 0.4 mM
nonessential and essential amino acids without Cys, ampicillin and chloramphenicol before
inoculation of M9 liquid medium (Ausubel et al., 1994a). The overnight liquid culture was
used to inoculate fresh medium (1:100) and grown until the AbSsoom = 0.6-0.7. The culture
was then induced with 0.4 mM IPTG for 15 min after which 200 MM Zn2" and 200-300 MCi
["S]Cys were added. The culture was incubated for another 3-4 h before being harvested.
The protein was then prepared as described by Kosa and colleagues (1994). The
supernatant was filtered with a 0.2 /AM disc filter and lower molecular weight proteins were
isolated using the Superdex S75 Hiload column (Pharmacia) with an isocratic elution using
10 mM potassium phosphate, 100 mM potassium chloride buffer, pH 8.0 (Khoo and Cousins,
1994). The CRIP-containing fractions were pooled based on "3S radioactivity and molecular
mass before dialysis against 200 x volume 10 mM potassium phosphate buffer, pH 8.0
overnight. Final purification was carried out with a weak cation exchange column (CM
Sepharose Fast Flow gel; XK 14 column; Pharmacia). The sample was loaded in 10 mM
potassium phosphate buffer, pH 8.0 and eluted with a linear gradient of 0 to 500 mM
potassium chloride (pH 8.0), for 100 min at a flow rate of 1.2 ml/min. Four ml fractions
were collected. CRIP eluted at approximately 150 mM potassium chloride.
Intestinal cytosol, membrane proteins and nuclear extract were prepared as described
above. Proteins from each sample (100-150 /g) were resolved by discontinuous denaturing
SDS-PAGE and transferred onto PVDF membranes. Non-radiolabeled recombinant CRIP
was also loaded on the gel in increasing amounts from 1-10 tg per lane. The assay was
carried out as described by Crawford et al. (1992). Briefly, the membrane was blocked with
5% non-fat dry milk in PBS overnight. The membrane was then rinsed briefly and incubated
with the radiolabeled CRIP in overlay buffer (50 mM HEPES, 1 % non-fat dry milk, 154 mM
NaCI, 1 mM EDTA, 1 % NP-40, 10 MM ZnSO4, 0.1% 2-13 mercaptoethanol, pH 7.5) for 4
h (Crawford et al., 1992). The radioactivity of the incubation buffer was between 200,000-
250,000 cpm/ml. After incubation, the membrane was washed once in PBS with 0.1 % SDS,
and three times in PBS with 0.1% Tween-20, for 5-10 min each time. The membrane was
dried and sprayed with Enhance' (Du Pont NEN, Boston, MA) to reduce the exposure time
for autoradiography (14-21 d).
As an alternative, a sandwich assay was used with the anti-CRIP IgG as a probe for
nonradiolabeled CRIP bound to its partner on the membrane after the initial overlay assay.
The antibody was then detected using the streptavidin-biotin amplification system (ECL
detection kit, Amersham Life Science, Arlington Heights, 11.). The detailed method for the
detection kit has been described in Material and Methods, Chapter 3.
Cyanogen-Bromide Activated Sepharose 4B Affinity Chromatography
The cyanogen-bromide activated Sepharose 4B gel (Pharmacia) was prepared as in
the manufacturer's protocol. Between 8-10 mg of recombinant rat CRIP (rrCRIP) was
dissolved in 2 ml of the coupling buffer (0.1 M NaHCO3, 0.5 M NaCI, pH 8.3), and mixed
with the gel at 40C overnight on a rotator plate. Excess ligand was washed off with at least
five gel volumes of coupling buffer and the column blocked for 2 h with 0.1 M Tris-base
buffer, pH 8.0. The gel was washed further with three (2 ml) alternating cycles of an acidic
buffer (0.1 M glycine, 0.5 M NaCI, pH 2.7), and a basic buffer (0.1 M NaHCO3, 0.5 M
NaCI, pH 8.0) and then packed into a column (0.5 cm x 3 cm) at 133% of the flow rate
using three column volumes of loading buffer (0.1 M NaH2PO4, pH 8.0).
The sample (5 mg in 300 Al) was diluted in 5 ml loading buffer, applied to the
column, and left overnight at 4C. The column was then washed extensively with loading
buffer until A280nm =0. Three different elution conditions (buffers) were used: 0.1 M
glycine, 0.5 M NaCI, pH 2.7; 0.1 M Ethanolamine, pH 11; and 0.1 M NaH2PO4, 1 M
NaCI, pH 8.0 or 0.1 M glycine, 1.0 M NaCI, pH 2.7. The fractions collected were pooled
from several preparations and concentrated with Centriprep-3 for the larger volumes and
Centricon-3 for smaller volumes (Amicon, Beverly, MA). The samples were then analyzed
by gel electrophoresis and stained for protein with Coomasie Blue stain (Biorad, Richmond,
Ca) followed by microsequencing of appropriate bands. Duplicate gels of the fractions from
the CNBr column were transferred onto PVDF membrane for the overlay assay.
CRIP was incubated with the putative partner overnight in overlay buffer and
immunoprecipitated with the anti-peptide CRIP IgG described in Chapter 2, Materials and
Methods. Protein A beads were added to facilitate precipitation. After three washes with the
overlay buffer, the beads were solubilized by boiling the samples in SDS PAGE buffer for
5 min, followed by centrifugation to precipitate the beads. The samples were then analyzed
For crosslinking, DSP (dithiobis(succinimidylpropionate)) (Pierce, Rockford, II.) was
used at a concentration of 0.2 mM. The putative partners were incubated with "S-rrCRIP
overnight at 4C in the overlay buffer. DSP was added for 15 min at room temperature
followed by quenching with 50 mM Tris, pH 7.5 for 30 min. The samples were then
analyzed by tandem Superdex 75 columns (Khoo and Cousins, 1994), and "S in the fractions
was measured by liquid scintillation counting.
Initial studies were carried out on intestinal cytosolic and membrane proteins and
RAW cell nuclear extract by Western analysis using the rabbit anti-CRIP IgG made against
the CRIP peptide consisting of amino acids 31-55 (Fernandes et al., 1995). The results
showed that the CRIP antibody was specific for CRIP in intestinal cytosol (Figure 4-1).
Western analysis of the intestinal membrane fractions and RAW nuclear protein extract
showed that there were no detectable CRIP protein in these fractions as prepared with the
methods above (data not shown).
The radiolabeled recombinant CRIP made with the above procedure was analyzed by
chromatography and SDS-PAGE to assess purity. The various preparations had a specific
activity of 5-20 jyg/pCi or approximately 220 MBq/mmol protein (Figure 4-2). Using this
probe for the overlay assay, we found that CRIP was able to bind a protein larger than 43
kDa when intestinal proteins were separated on a 15% SDS-PAGE gel (Figure 4-3). It was
also discovered that under these conditions, CRIP did not appear to form dimers as
increasing amounts of CRIP protein on the membrane did not appear to bind to the
radiolabelled CRIP probe (Figure 4-3). This was consistent with data from gel filtration
6 I- CRIP
Western analysis of rat intestinal cytosol separated by 15% SDS-PAGE and
transferred onto nitrocellulose. Rabbit anti-rat CRIP peptide (residues 31-56)
IgG at 1:100 dilution and secondary antibody, donkey anti-rabbit IgG
(conjugated to horseradish peroxidase), at 1:5000 dilution were used. The
procedure was carried out as described in Materials and Methods, Chapter 3,
with the ECL Western detection kit (Amersham Life Science). (a) shows the
autoradiograph of the Western analysis of the cytosol and (b) the Coomasie-
stained duplicate gel.
|'. & 82
r e .5
(M) Q%0 00 0 \O
1 I I S
analysis of CRIP with a crosslinker which showed that, under this non-reducing condition,
dimerization also did not occur (data not shown).
Using a 10% gel to resolve rat intestinal proteins, it was found that CRIP was able
to bind a 116 kDa protein in the intestinal cytosol fraction as calculated by electrophoretic
mobility (Rf) (Figure 4-4a and b). Mass measurements by gel electrophoresis are +5% of
actual mass. The 116 kDa protein was found to be N-terminally blocked and due to its low
abundance could not be sequenced without further processing. There was also light binding
to a group of bands between 49-68 kDa which, when sequenced were identified as serum
albumin (Figure 4-4). Control reactions using nonradiolabeled rrCRIP to compete with
radiolabelled rrCRIP were able to eliminate binding to the putative protein partners. Serum
albumin and the 116 kDa protein were also found to bind to rrCRIP using the sandwich
overlay assay with the rabbit anti-rat CRIP IgG (Figure 4-4c).
There were no distinct bands in the membrane fraction that bound to rrCRIP except
for a band at around 70 kDa that was serum albumin, a contaminant of the membrane
preparation (Figure 4-4a, lane 3). When intestinal cytosol was used for the CRIP affinity
column, elution with 1 M NaCI from the affinity column yielded small peaks that when
analyzed were found to be serum albumin (data not shown). Further analysis of serum
albumin binding was carried out by crosslinking and gel filtration analysis. It was found that
the peak for the unbound radiolabeled rrCRIP was around 29-30 ml (Figure 4-2a). When
serum albumin was incubated with rrCRIP and crosslinker, two peaks with significant
radioactivity were observed at elution volumes 18.5 and 20 ml, in the larger molecular
weight region (Figure 4-5). These corresponded to the absorbance peaks at 280 nm (data not
S ( .4 4 ^ -, +
C/ 0 1 9~ 8 ,; 4 c q M
Q .5 g -? G
ti- g g 2
4- >T, 4>
4 54 && 2
Sia s o S 9 Sa
g 0cc 0 Cg 0
00 C' CCU4'7
42 of 'mo
CSs 0 Q^s
"*i-2 ga3 s
0 ci s bA 0 4 bo
g C,, (= !I4
0 Q K 44 2 og 0
| 4- ^I-
0f SS ^5
*g & ^i': 11 g
-,.j a 0- a j
S 3 -a -a ^ o
C' s w- '^ 0 S
-^bO u cA :nf af -
t- 00 M
C o t
O- 00 C
ON o 11,
20 25 30
Gel filtration profile of DSP-crosslinked serum albumin and [35S]Cys-rrCRIP
analyzed using the tandem Superdex 75 columns. (U) BSA, [35S]Cys-rrCRIP
and DSP; (A) [35S]Cys-rrCRIP and DSP. Control runs (A) were carried out
with [35SjCys-rrCRIP and crosslinker without serum albumin. Fractions of
0.5 ml were collected and 35S was measured using liquid scintillation
In the membrane fractions of both rat intestine and RAW cells, preliminary studies
with crosslinking and gel filtration produced a radioactive peak in the void volume of the
Superdex 75 columns, suggesting that either rrCRIP was bound to a large protein or was
forming a complex with several proteins excluded from the Superdex matrix (data not
shown). Therefore, membrane proteins were resolved by SDS-PAGE and assessed for CRIP
binding by the Western (sandwich) overlay assay using ECL immunodetection. Two protein
bands were detected to bind to rrCRIP with this method, around 48 kDa and 70 kDa (data
not shown). A control reaction where CRIP was omitted from the sandwich assay, indicated
that one of the antibodies reacted with the band at 48 kDa. Therefore, this band was not
considered a partner for CRIP but rather, appeared to bind non-specifically to one of the
antibodies. This showed that CRIP did not bind to any proteins in the intestinal membrane
fraction as prepared. This method of preparation was used to produce a crude membrane
product just to distinguish it from the cytosolic fraction. As shown by the contamination with
albumin, this was not a pure membrane fraction. It was possible that this method caused the
loss of some cytoskeletal proteins in the initial centrifugation steps and some membrane
proteins along with those. In addition, membrane proteins could have been incompletely
solubilized due to inadequate amounts of detergent used. Therefore, the membrane partner
of CRIP, if it existed, may not have been released during the preparation.
A cyanogen bromide affinity column with rrCRIP bound as a ligand was used as a
non-denaturing method to obtain the protein partners) of CRIP. This method was able to
detect interactions with binding constants as weak as 105 M. This was within a valid range
as binding constants stronger than 103 M were considered physiologically relevant (Phizicky
and Fields, 1995). The profile of elution is shown in Figure 4-6. In repeated experiments,
Glycine I M NaCi
U i U i I I U I I U U i U U I | | U I I |
10 20 30 40 50 60 70 80
90 100 110
Elution profile of rat intestinal cytosol from the cyanogen bromide-rrCRIP
affinity column. The sample was washed extensively with 0.1 M sodium
phosphate, pH 8.0, before sequential elution using three different buffers, as
shown in the figure: 0.1 M glycine, 0.5 M NaCI, pH 2.7; 0.1 M
ethanolamine, pH 11, and 0.1 M sodium phosphate, 1 M NaCl, pH 8.0.
the same peak was obtained with elution using the ethanolamine buffer. Since earlier
evidence had been obtained from the studies described above that CRIP interacted with
proteins in the cytosolic fraction and since CRIP was present in the cytosol, the cytosolic
fraction was used with this method. Under these conditions, it was possible to detect a 25
kDa band which bound to rrCRIP on the column (Figure 4-7a). The 25 kDa band when
sequenced was found to be rat trypsinogen I. A smaller band of 3-6 kDa was also found to
bind to the column (Figure 4-7a), but was found later to be a mixture of peptides that bound
non-specifically to the column. In a control reaction, this band bound to the column prepared
without the rrCRIP ligand, while trypsinogen did not do so (Figure 4-7b).
When the eluted proteins were transferred onto membranes and tested with the
overlay assay using radiolabeled rrCRIP, the band at 25 kDa was able to bind rrCRIP as
shown by the autoradiograph (Figure 4-7c). Using the overlay assay on the non-purified
intestinal cytosol fraction, but with the antibody detection system, the antibody was able to
detect CRIP in the cytosol fraction as well as another band with a molecular weight of 25
kDa (Figure 4-8d). Control experiments with a Western analysis on the cytosol fraction,
without carrying out the overlay assay initially, did not yield this 25 kDa band (Figure 4-1).
The antibody was able to detect the immobilized rrCRIP bound to this band.
Binding of rrCRIP to trypsinogen was further analyzed by immunoprecipitation and
SDS-PAGE, as well as crosslinking. As shown in Figure 4-8, when trypsinogen was
incubated with the anti-peptide CRIP IgG, there was no precipitation product. However,
when rrCRIP was added, both rrCRIP and trypsinogen were precipitated. CRIP was
precipitated by the antibody without the presence of trypsinogen (data not shown). By
crosslinking and gel filtration analysis, a radioactive peak was observed at the elution volume
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 EJ5O0WHUS_Z82CL2 INGEST_TIME 2012-02-20T21:40:19Z PACKAGE AA00009031_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC