Download Krp1, a novel kelch related protein that is involved in pseudopod

ã
Oncogene (2000) 19, 1266 ± 1276
2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00
www.nature.com/onc
Krp1, a novel kelch related protein that is involved in pseudopod elongation
in transformed cells
Heather J Spence1, Imogen Johnston1, Karen Ewart1, Sarah J Buchanan1, Una Fitzgerald1 and
Bradford W Ozanne*,1
1
Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Bearsden, Glasgow, G61 1BD
We have previously shown that the transcription factor
AP-1 regulates the expression of genes which allow
neoplastically transformed rat ®broblasts to become
invasive. Searches for further AP-1 target genes led to
the identi®cation of a gene encoding a novel rat kelch
family member, named kelch related protein 1 (Krp1).
Kelch family members are characterized by a series of
repeats at their carboxyl terminus and a BTB/POZ
domain near their amino terminus. Rat Krp1 has a
primarily cytoplasmic localization, and a small fraction
appears to accumulate and co-localize with F-actin at
membrane ru‚e-like structures in the tips of pseudopodia. Overexpression of Krp1 in transformed rat ®broblasts led to the formation of dramatically elongated
pseudopodia, while expression of truncated Krp1 polypeptides resulted in a reduction in the length of
pseudopodia. We propose that the transformationspeci®c expression of Krp1 is required for pseudopod
elongation, which are structures that are required for cell
motility and invasion. Oncogene (2000) 19, 1266 ± 1276.
Keywords: Kelch; Krp1; v-fos; pseudopodia; cytoskeleton
Introduction
One of the major issues of cancer progression is the
complex process that results in a benign tumour
becoming malignant. The ®nal change in tumours
during this process is that they acquire the ability to
invade, which allows them to penetrate the basement
membrane and migrate into neighbouring tissues. Such
processes require alterations in the following: cell ± cell
interactions; cell ± extracellular matrix (ECM) adhesion;
cytoskeleton; and cell motility (reviewed in Liotta et
al., 1991a). We have recently proposed that the
transcription factor AP-1 in v-fos transformed 208F
rat ®broblasts (FBR cells) co-ordinates the changes in
gene expression that allow the cells to become invasive
(Hennigan et al., 1994; Lamb et al., 1997a,b).
V-fos encodes the transforming agent of the FBR
murine sarcoma virus, and forms part of the transcription factor AP-1 (Finkel et al., 1975; van Beveren et al.,
1984; Curran and Verma, 1984; Curran and Franza,
1988). AP-1 is a heterodimeric protein composed of
various combinations of the Jun and Fos family proteins,
and this ¯exibility permits transactivation or transrepres-
*Correspondence: BW Ozanne
Received 4 October 1999; revised 23 December 1999; accepted 23
December 1999
sion of gene expression (Bushel et al., 1995; De Cesare
et al., 1995; reviewed in Karin et al., 1997). Upon
transformation of 208F cells with v-fos, the cells show
extensive reorganization in their cytoskeleton and
become invasive (Hennigan et al., 1994; Lamb et al.,
1997a). These cytoskeletal rearrangments result in a
striking change in the morphology of the cells: they
change from being ¯at with F-actin stress ®bres to being
bipolar with long pseudopodia, which are lined with
cortical F-actin cables that have dynamic actin-rich
membrane ru‚e-like structures at their tips. Pseudopodia have previously been shown to be implicated in
motility and invasion of tumour cells, highlighting the
importance of cytoskeletal reorganization in invasion
(Guirguis et al., 1987; Hay et al., 1989; Lamb et al.,
1997b; Liotta et al., 1991b; Nabi et al., 1999).
We have previously shown that AP-1 is essential for
invasion of FBR cells and 208F cells treated with EGF
(Hennigan et al., 1994; Lamb et al., 1997a). This was
achieved using antisense oligonucleotides to c-jun, and
the expression of a c-Jun dominant negative mutant
(TAM-67) to inhibit invasion. Studies in c-fos null mice
showed that Fos was needed for premalignant
papillomas to progress to malignant skin carcinomas,
when subjected to chemical carcinogens or an activated
ras oncogene (Saez et al., 1995). Transformation of
®broblasts by oncogenes, such as raf and ras, have also
been shown to be dependent on AP-1 activity (Johnson
et al., 1996; Lloyd et al., 1991; Rapp et al., 1994;
Suzuki et al., 1994). The role of AP-1 in transformation is not limited to ®broblasts, since mouse and
human squamous cell carcinomas expressing the
dominant negative Jun mutant have reduced tumourigenicity and in vitro invasiveness (Domann et al., 1994;
Dong et al., 1997; Malliri et al., 1998).
Studies of AP-1 target genes in oncogenically
transformed cells have also highlighted the role of AP1 in invasion. Examples of proteins expressed are type 1
collagenase and stromelysin 1, both of which are
proteases that are involved in the degradation of the
ECM (Kerr et al., 1988; Schonthal et al., 1988). In order
to identify more AP-1 regulated genes that are needed for
FBR invasion, we have used di€erential screening and
subtractive suppression hybridization (Diatchenko et al.,
1996; Hennigan et al., 1994; Johnston et al., 1999,
unpublished). These searches have revealed genes which
encode proteins that have previously been implicated in
invasion, as well as numerous novel genes that may play
a role in transformation. An example of the genes whose
expression are up-regulated is that encoding the cell
surface hyaluronan receptor CD44, which is localized at
the tips of pseudopodia in v-fos and EGF transformed
208F cells (Lamb et al., 1997a). CD44 antisense
Krp1 contributes to pseudopod elongation
HJ Spence et al
oligonucleotides caused a reduction in invasiveness in
both of these cells. The expression of ezrin, which is a
member of the ezrin/radixin/moesin (ERM) family and is
thought to link CD44 to the cell's cytoskeleton, was also
shown to be upregulated (Lamb et al., 1997b; Jooss and
Muller, 1995; Tsukita et al., 1994). Ezrin is located at the
tips of pseudopodia in FBR cells and EGF-treated 208F
cells, and removal of it from pseudopodia using the
micro-CALI technique resulted in retraction of pseudopodia, suggesting that ezrin has a role in pseudopod
extension (Lamb et al., 1997b).
Here we present data on a novel protein, named
Kelch-related protein 1 (Krp1), which was isolated due
to its upregulation upon transformation of 208F cells
with v-fos and v-ras. We show that Krp1 is a member
of a family of proteins that is de®ned by a carboxyl
terminal series of approximately 50 amino acid repeats
(the kelch repeats), and a protein ± protein interaction
domain, named the BTB/POZ domain, near its amino
terminus (Xue and Cooley, 1993; Hernandez et al.,
1997; Kim et al., 1998). Members of this family have
been shown to be involved in cytoskeleton organization
and function (Xue and Cooley, 1993; Hernandez et al.,
1997; Kim et al., 1998; Robinson and Cooley, 1997).
We report that rat Krp1 is cytoplasmically localized in
FBR cells, and accumulates at the tips of pseudopodia,
where it co-localizes with F-actin. In addition, we show
through overexpression studies of the entire rat Krp1
protein in FBR and Ras cells, and the BTB/POZ
domain or the Kelch repeats in FBR cells, that Krp1 is
required for pseudopod elongation.
revealed that krp1 was also upregulated in the Ras
cells, although to a lesser extent than in the FBR cells
(Figure 1b). These data illustrate that increased krp1
expression correlates with the transformation of 208F
cells, and its expression is dependent on AP-1.
To examine whether or not krp1 is expressed in a
tissue-speci®c manner, we performed Northern blots of
poly A+ RNA from various rat tissues, and found that
krp1 is expressed primarily in muscle tissue (Figure 1d,e).
1267
Results
Isolation of a cDNA which is upregulated in 208F cells
transformed with v-fos and v-ras
A cDNA library representing the mRNAs which are
upregulated in FBR cells compared to 208F cells was
prepared using the suppression subtractive hybridization method (Diatchenko et al., 1996), and 500 random
clones were sequenced (Johnston et al., 1999, unpublished). The open reading frame (ORF) from one
400 bp clone, designated krp1, showed 31% amino acid
sequence identity to a Drosophila kelch protein, which
is involved in crosslinking actin ®laments in the ring
canals during development (Xue and Cooley, 1993;
Robinson and Cooley, 1997). As transformation by vfos results in reorganization of the cytoskeleton, we
predicted that such a protein may play a role in this
remodelling, and the cDNA clone was chosen for
further investigation.
To determine the expression pattern of krp1, Northern blots were prepared with total RNA isolated from
the following cell lines: 208F, FBR, 208F transformed
with ki-ras (Ras), and FBR expressing a dominant
negative c-Jun mutant (FBR-TAM67). Upon hybridization with krp1 an RNA species of approximately
2.4 kbp was detected in FBR cells but not in 208F
cells, indicating that it is upregulated upon v-fos
transformation (Figure 1a). In addition, its expression
was reduced in FBR-TAM67 cells. Since TAM67 is a
dominant negative Jun mutant and blocks AP-1
function, this again con®rms that krp1 expression is
AP-1 dependent. A longer exposure of the blot
Figure 1 krp1 upregulation in FBR and Ras cells and expression
in rat tissue. Northern blot analysis of total RNA isolated from
FBR-TAM67, Ras, FBR and 208F cell lines. The blot was probed
with the krp1 cDNA clone and left for a 3 h exposure (a) and a 3
day exposure (b). The blot was stripped and then probed with s7
as a loading control (c). A Rat poly A+ tissue blot from Clontech
was probed with krp1 (d) and then stripped and probed with actin
as a loading control (e)
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1268
Isolation and characterization of rat and human krp1
cDNAs
To isolate a full-length cDNA clone of krp1, a library
derived from FBR mRNA was screened. Six clones
were isolated, and two of these, designated pkrp14 and
pkrp16, contained an identical ORF of 1821 bp. The
ORF encoded a putative polypeptide with a molecular
weight of 68 212 Da, which we named Kelch-related
protein 1 (Krp1; Figure 2a). Five non-identical repeats
of approximately 50 amino acids are apparent at the
carboxyl terminus, which are characteristic of proteins
belonging to the kelch family (see Figure 2b). At the
amino terminus of the putative polypeptide (amino
acid residues 5 ± 122) there is a domain that shows 42%
sequence similarity to the BTB/POZ domain. This
domain is thought to be involved in protein ± protein
interactions and is found in many of the kelch family
Figure 2 Amino acid sequence of rat and human Krp1. (a) Shows the complete amino acid sequence of rat Krp1. Also shown are
the amino acids which di€er between rat and human Krp1. The BTB/POZ domain is underlined and the kelch repeats of Krp1 are
shown in bold. (b) Alignment of the rat Krp1 kelch repeats. The bottom line shows the consensus sequence. White letters on the
black background represent conserved residues, while white letters on a grey background represent similar amino acids. K1-5
represents the ®ve kelch repeats. (c) Schematic representation of some of the kelch family members
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1269
Figure 3 Krp1 localization in FBR, FBR-Tam67 and 208F cells with anti-Krp1 antiserum. (a) Western blot analysis of Ras, 208F
and FBR cell lysates probed with anti-Krp1 antiserum. The blot was stripped and reprobed with anti-ERK2 antisera as a loading
control. (b) Double immuno¯uorescence staining of FBR cells with Krp1 antiserum with a FITC-labelled anti-rabbit secondary
antibody and tubulin antiserum with TRITC-labelled anti-mouse secondary antibody. Double immuno¯uoresence of FBR cells (c ±
e), 208F cells (f) and FBR-TAM67 cells (g) with Krp1 antiserum with a FITC-labelled anti-rabbit secondary antibody and TRITCphalloidin for F-actin staining. (d) Shows tips of a FBR psuedopod with the same labelling, but at a higher magni®cation. (e,1)
Shows immuno¯uoresence staining without pre-permeablization of the cells and in (e,2 and 3) with pre-permeablization of the cells
in 0.02% Triton x100/PBS for 1 min. Actin is shown in red and Krp1 in green. The ®gure also shows the merged images where
yellow represents where both Krp1 and F-actin co-localize. Arrows highlight the localization of Krp1 at the tips of pseudopodia.
Bars=25 mm
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1270
of proteins, as well as in some DNA binding proteins
such as the Drosophila Broad Complex (reviewed in
Albagli et al., 1995). A schematic representation of
some members of the kelch family, highlighting the
positions of the kelch domains and BTB/POZ domains,
is shown in Figure 2c.
To isolate the human homologue of krp1, a human
skeletal muscle cDNA library was screened with the rat
krp1 full-length cDNA. This library was used because
rat krp1 is expressed primarily in muscle tissue (Figure
1d,e). Four clones were isolated, two of which were
predicted to contain the complete ORF. The human
gene encodes a putative polypeptide with a molecular
weight of 68 151 Da, which shows 95% amino acid
sequence identity to rat Krp1, con®rming we have
isolated a human gene which is homologous to rat
Krp1 (Figure 2a). The amino acid sequence identity of
95% between rat and human Krp1 is higher than the
41% identity observed between rat Krp1 and another
human kelch family member, NRP/B (Kim et al.,
1998). These data suggest that within the kelch family
there are sub-families which are more closely related.
Rat Krp1 co-localizes with F-actin at the tips of
pseudopodia in FBR cells
To analyse the subcellular localization of endogenous
Krp1 in FBR cells, polyclonal rabbit antiserum was
raised to bacterially-expressed full-length rat Krp1
fused to a 6 Histidine epitope. A Western blot was
performed with FBR, Ras and 208F cell lysates and
was probed with the anti-Krp1 antiserum. A band of
approximately 68 kDa was detected in the FBR and
Ras cell lysates, corresponding to the predicted size of
the Krp1 polypeptide, while nothing was detected in
the 208F cell lysate (Figure 3a). These data, along with
the Northern analysis (above), con®rm that Krp1
expression is speci®cally upregulated upon transformation of 208F cells.
Immuno¯uorescence experiments performed on FBR
cells using anti-Krp1 antiserum revealed that Krp1 has
a primarily cytoplasmic localization, and a small
fraction accumulates at the tips of transformationspeci®c pseudopodia (Figure 3b ± d). Cell fractionation
experiments con®rmed that Krp1 is primarily found in
the cytoplasm, since the majority of it was present in
the soluble fraction (data not shown). Examination of
Krp1 expression in pseudopod tips at a higher
magni®cation revealed that it was present in membrane
ru‚e-like structures that appeared to be forming a
looping conformation (Figure 3d). Because it is known
that some kelch family proteins interact with actin
(Xue and Cooley, 1993; Hernandez et al., 1997;
Soltysik-Espanola et al., 1999), and that the yeast
Tea1 kelch family member localizes at the ends of
microtubules (Mata and Nurse, 1997), we performed
two double immuno¯uorescence experiments on FBR
cells: anti-Krp1 antiserum and phalloidin to detect Factin; anti-Krp1 and tubulin antisera (Figure 3b ± d).
Both tubulin and Krp1 were present in the cytoplasm,
but while Krp1 extended into the loop-like structures
at the tips of pseudopodia, tubulin stopped just short
of these structures (Figure 3b). Krp1 and F-actin colocalized at the tips of pseudopodia, but not in the
cytoplasm, where F-actin was found as long cables
(Figure 3c,d). To con®rm that Krp1 does not co-
Oncogene
localize with F-actin cables in the cytoplasm, live cells
were treated with triton X-100 to remove di€usable
proteins before ®xing for immuno¯uorescence (Figure
3e). This revealed that Krp1 remained at the tips of
pseudopodia with F-actin, but did not associate with
the F-actin cables. Double immuno¯uorescence with
anti-Krp1 antiserum and phalloidin was also performed on 208F cells. Krp1 was not expressed in these
cells (Figure 3f), con®rming the Northern and Western
data. The same experiment performed on the FBRTAM67 cells, which have a reverent phenotype in that
they have a ¯attened morphology, and showed loss of
pseudopodia and a gain of stress ®bres, indicated that
Krp1 had a cytoplasmic localization (Figure 3g). It
should also be noted that in FBR-TAM67 cells Krp1
does not co-localize with F-actin in the stress ®bres or
in the membrane ru‚e structures (Figure 3g). These
data con®rm that Krp1 expression and localization is
transformation-speci®c, and that the protein only colocalizes with F-actin at the membrane ru‚e-like
structures at the tips of pseudopodia.
Rat Krp1-myc localizes at the tips of pseudopodia in
transformed 208F cells
To con®rm the localization of Krp1 at the tips of
pseudopodia, the full-length rat krp1 ORF was cloned
into the mammalian expression vector pcDNA3.1 MycHis, creating a translation fusion in which the 9 amino
acid myc epitope was placed at the carboxyl terminus
of Krp1 (Krp1-myc). The construct was then transfected into FBR cells and the distribution of Krp1-myc
assessed 24 h later using immuno¯uorescence with an
anti-myc monoclonal antibody. The subcellular localization of Krp1-myc appeared identical to that of
endogenous Krp1, con®rming that Krp1 is primarily
cytoplasmic and that a small fraction localizes at the
tips of pseudopodia in FBR cells (Figure 4a).
The Krp1-myc construct was next transfected into
208F cells to determine its e€ect and localization in
untransformed cells. Krp1-myc was distributed throughout the cytoplasm, and did not appear to associate with
the membrane (Figure 4b). Notably, Krp1-myc did not
co-localize with the cytoplasmic F-actin, which form
stress ®bres in these cells. In addition, expression of krp1myc did not appear to have any e€ect on the morphology
of 208F cells. To test whether this localization was
speci®c for v-fos transformed 208F cells, we next
expressed Krp1-myc in 208F cells which were subsequently treated with EGF and in Ras cells. Treatment of
208F cells with 100 ng/ml of EGF results in membrane
ru‚ing after 4 h, and removal of stress ®bres and the
formation of pseudopodia after 24 h (Kaplan and
Ozanne, 1983). We determined the localization of
Krp1-myc in 208F cells after 5 min, 2 h, 4 h and 24 h
of EGF treatment. At both 5 min and 2 h of EGF
treatment Krp1 remained cytoplasmic, while at 4 h Krp1
was found to localize at membrane ru‚es (Figure 4c).
After 24 h a small fraction was localized at the tips of
pseudopodia (Figure 4c). It should be noted that not all
the cells after 24 h had long pseudopodia, but were
instead only partially transformed, as they had small
pseudopodia with ru‚ing tips where Krp1 localized. Ras
cells have a distinct morphology to FBR cells in that they
have many short pseudopodia with actin-rich structures
at their tips, while FBR cells have only two long
Krp1 contributes to pseudopod elongation
HJ Spence et al
1271
Figure 4 Localization of Krp1-myc in FBR, 208F, EGF treated 208F and Ras cells. Transfection of Krp1-myc into FBR (a), 208F
(b), 208F treated with 100 ng/ml of EGF for 5 min, 2, 4 and 24 h (c) and Ras (d). Cells were viewed in (a) and (d) by
immuno¯uorescence with myc monoclonal antibody followed by a TRITC-labelled anti-mouse secondary antibody, and in (b) and
(c) by double immuno¯uorescence with monoclonal antibody to myc-epitope followed by a TRITC-labelled anti-mouse secondary
antibody and FITC-phalloidin for F-actin. Arrows point to the accumulation of Krp1-myc at the tips of pseudopodia and at the
leading edge of the cells. Size bars=25 mm
pseudopodia. Despite this, Krp1-myc was still localized
in the cytoplasm and accumulated at the tips of all
pseudopodia (Figure 4d). These data, together with the
localization of endogenous Krp1 in FBR and FBRTAM67 cells, suggest that the localization of Krp1 in the
tips of pseudopodia is transformation-speci®c.
Overexpression of rat Krp1-myc and dominant negative
Krp1 mutants suggest that Krp1 has a role in pseudopod
elongation
To determine whether overexpression of rat Krp1-myc
in FBR and Ras cells had any e€ect on the
morphology of the cells, transfections were performed
and the phenotypes of Krp1-myc positive cells
analysed. Thirty-six per cent of FBR cells and 25%
of RAS cells showed elongated pseudopodia (Figure
5b ± e). Elongated pseudopodia were particularly
apparent in Ras cells, as they normally have many
short pseudopodia. To verify that Krp1-myc overexpression was responsible for this phenotypic change,
two Krp1 variants were cloned into the pcDNA3.1
Myc-His vector. The ®rst construct contained amino
acids 1 ± 181, encompassing the BTB/POZ domain,
and the second contained the kelch repeats (amino
acids 335 ± 606). To con®rm that these constructs were
expressing products of the expected sizes, they were
transfected into COS-7 cells and Western blots were
performed on the cell lysates. Bands of 68, 20 and
30 kDa were detected by an anti-myc monoclonal
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1272
Figure 5 E€ects of overexpression of Krp1-myc and the variant Krp1 constructs in FBR and Ras cells. (a) The Krp1-myc,
Krp1repeats-myc and Krp1poz-myc constructs were transfected into COS-7 and the cell lysates were used in a Western blot analysis.
Blots were probed with anti-myc monoclonal antibody. (b) and (d) Illustrates elongated pseudopodia produced when Krp1-myc is
overexpressed in both FBR and Ras cells respectively. Cells were visualized using anti-myc monoclonal antibody and secondary
TRITC-labelled anti-mouse antibody. (c) and (e) Quantitative analysis of the average percentage of FBR and Ras cells respectively
that showed elongated pseudopodia. The average length of pseudopodia in FBR cells was approximately 30 mm and in RAS cells
10 mm, anything more than two times the length of the average was classi®ed as elongated. The experiment was performed on three
separate occasions on FBR cells and two times on Ras cells, on each occasion the experiments were performed in duplicate. (f) and
(g) Show the reduced pseudopodia size when FBR cells are transfected with either the Krp1poz-myc and Krp1repeats-myc. (h)
Quantitative analysis of the average percentage of FBR cells showing a reduction in pseudopodia length. Shortened pseudopodia
were classi®ed as anything smaller than 15 mm. The experiment was repeated on three separate occasions, and was performed in
duplicate each time. Bars=25 mm
antibody, which were the expected sizes for the Krp1myc, Krp1poz-myc and Krp1repeats-myc polypeptides
respectively (Figure 5a). Transfection of these constructs into FBR and Ras cells revealed that elongated
pseudopodia were seen more frequently in cells
expressing the entire Krp1 protein, indicating that
the whole protein is needed to exert this e€ect. In
FBR cells transfected with Krp1poz-myc or Krp1
repeats-myc, 54 and 55% of the cells respectively
displayed pseudopodia shortened by up to 80%
(Figure 5f ± h). A control construct expressing Betagalactosidase fused to the myc epitope did not cause
any signi®cant lengthening or shortening of pseudopodia, indicating that these e€ects are Krp1 speci®c. It
is likely that the truncated Krp1 polypeptides are
acting as dominant negative mutants, which are
interfering with wild type Krp1 function in pseudopod
elongation.
Oncogene
Discussion
Through our investigation of genes which are
upregulated as a consequence of v-fos transformation
of rat ®broblasts, we identi®ed a new member of the
kelch family of proteins, which are characterized by a
BTB/POZ domain at the amino terminus and a series
of repeats at their carboxyl terminus. We have shown
that the expression and co-localization of Krp1 with
F-actin at membrane ru‚e-like structures at the tips
of pseudopodia is dependent upon transformation of
rat ®broblasts. In addition, overexpression of fulllength Krp1 results in dramatic elongation of
extending pseudopodia, while expression of either
the BTB/POZ domain or the carboxyl terminal
repeats in isolation results in truncated pseudopodia.
These data suggest that Krp1 has a role in
pseudopod elongation.
Krp1 contributes to pseudopod elongation
HJ Spence et al
Krp1 is a member of the kelch family of proteins,
which have been found in many di€erent species. It
has become apparent that the kelch family members
have many diverse functions. Some of them are
thought to interact with actin through their kelch
repeats. An example is Drosophila kelch (Xue and
Cooley, 1993), which is found in the ring canals
(actin-rich structures that link the nurse cells to the
developing oocyte) where it is thought to stabilize
actin ®laments. Another protein, alpha-scruin of
Limulus, which has kelch repeats present at both its
amino and carboxyl termini but lacks a BTB/POZ
domain, cross-links actin ®laments within the acrosomal processes of sperm (Way et al., 1995a). Further
kelch proteins are also thought to interact with actin:
murine and human ENC1 (Hernandez et al., 1997,
1998); spe-26 of C. elegans (Varkey et al., 1995); and
human Mayven (Soltysik-Espanola et al., 1999). Many
other family members are thought not to interact with
actin; beta-scruin of Limulus, for instance, is found in
sperm acrosomal vesicles, which do not contain actin
(Way et al., 1995b). In addition, Schizosaccharomyces
pombe kelch family member Tea1 is found at the tips
of growing cells and is thought to interact with the
ends of microtubules rather than with actin (Mata and
Nurse, 1997). Several other family member have been
characterized, but it is not known what, if any,
cytoskeletal components they interact with: Kel1p and
Kel2p of Saccharamyces cerevisiae (Philips and
Herskowitz, 1998); human NRP/B (Kim et al.,
1998); and mouse Muskelin (Adams et al., 1998). It
is apparent from the above that we can not assign a
speci®c function to Krp1 based upon its sequence
similarity to the kelch family, as di€erent members
appear to have diverse functions. It is conceivable,
however, that within the whole kelch-family there are
sub-families with the same or related functions, and it
is therefore interesting to note the high level of
sequence identity (95%) between rat and human
Krp1.
The sub-cellular localization studies presented reveal
that Krp1 co-localizes with F-actin in a dynamic
membrane ru‚e-like structure, which forms looping
conformations at the tips of pseudopodia in FBR
cells. Although Krp1 co-localizes with F-actin within
this structure, it does not do so in the cytoplasm: here
F-actin appears as large cables that run the length of
the cell body, while Krp1 appears to be dispersed
throughout the cytoplasm. In Ras cells Krp1-myc also
only co-localized with F-actin at the tips of
pseudopodia. It should also be noted that when
Krp1-myc was expressed in 208F cells, and when we
examined the low level of endogenous Krp1 that is
present in FBR-TAM67 cells, no co-localization with
F-actin containing stress ®bres was detected. This
result contradicts work presented by Hernandez et al.,
(1997) and Soltysik-Espanola et al. (1999), who
reported that two other kelch family members,
ENC1 and Mayven, interact with stress ®bres containing F-actin. Together these data suggest that Krp1
only co-localizes with F-actin when it is in the looped
membrane ru‚e-like structure at the tips of pseudopodia. Krp1 may not directly interact with actin, but
instead with unidenti®ed proteins that are required for
actin to adopt this conformation. It is interesting to
note in this regard that we have been unable to show
any direct biochemical interactions between actin and
Krp1.
The presence of Krp1 in the cytoplasm raises the
question of what, if any, function it has here. One
possibility is that this simply represents a cytoplasmic
pool of Krp1 that is constantly being moved to the
dynamic processes at the tips of pseudopodia. Preliminary two-dimensional SDS ± PAGE data suggest
that Krp1 in FBR cells is phosphorylated (data not
shown). Perhaps Krp1 is phosphorylated when present
at the tips of pseudopodia, while cytosolic Krp1 is
inactive and not phosphorylated.
The localization of Krp1 at the tips of pseudopodia
in FBR cells, and the recruitment of Krp1-myc to this
structure in 208F cells treated with EGF for 24 h and
Ras cells, is reminiscent of the products of two other
AP-1 regulated genes: CD44 and ezrin. Both of these
proteins are up-regulated in FBR cells and EGFtreated 208F cells and appear to cluster at the tips of
pseudopodia (Lamb et al., 1997a,b). This raises the
question of whether these proteins might be interacting
at the looped structure. It has been reported that ezrin
links the cytoplasmic domain of CD44 to the actin
cytoskeleton in baby hamster kidney cells and mouse L
®broblasts (Tsukita et al., 1994; Yonemura et al., 1998,
1999; Legg and Isacke, 1998), but we do not know if
this interaction occurs in FBR cells. None of our data
allow us to conclude that Krp1 interacts with these
proteins, but the presence of all of them at pseudopod
tips upon transformation of 208F cells supports the
hypothesis that AP-1 regulates the expression of these
genes, and that they are recruited to a structure that is
likely to allow transformed cells to become invasive
(Guirguis et al., 1987; Hay, 1989; Lamb et al., 1997b;
Liotta et al., 1991b; Nabi et al., 1999). Further
evidence for this proposal is seen when Krp1-myc,
ezrin or CD44 were individually expressed in 208F
cells, since they did not result in the formation of large
pseudopodia (Lamb et al., 1997a,b). These data imply
that AP-1 must regulate the expression of numerous
genes which are all needed for the cells to become
invasive. Further analysis of genes isolated from the
subtractive suppression hybridization libraries representing genes up-regulated in v-fos transformed cells
may give us a better understanding of the range of
genes needed for invasion.
To gain further insight into the function of Krp1, we
presented data from overexpression studies in FBR and
Ras cells. Dramatically elongated pseudopodia were
seen in both FBR and Ras transformed cells when they
were subjected to overexpression of a Krp1-myc
construct, suggesting that Krp1 may have a direct role
in pseudopod elongation. This is supported by the
result of overexpressing the BTB/POZ domain or kelch
repeats in isolation, which resulted in shortening of the
pseudopodia, presumably because these Krp1 variants
are acting as dominant negative mutants which
interfere with endogenous Krp1 function. A comparison between Krp1 and ezrin function can again be
made here, since ablation of endogenous ezrin led to
the loss of pseudopodia. We are not the ®rst to draw a
comparison between ezrin and kelch proteins, since
Vega and Solomon (1997) reported the similarity
between ezrin and the yeast kelch protein, Tea1. They
reported that both proteins are localized at tips of cells,
Tea1 being present at the tips of growing yeast cells
1273
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1274
and ezrin and other ERM proteins in growth cones,
and that they both depend on the integrity of
microtubules for localization. We have also shown
that the positioning of Krp1 at the tips of pseudopodia
is dependent on the integrity of the microtubules (data
not shown). These data, together with the localization
data, suggest that Krp1 may function with the ERM
proteins at the tips of cells, but as stated above our
data does not allow us to conclude that these proteins
interact directly. What this does suggest, however, is
that Krp1, ezrin and Tea1 are related in de®ning the
structures and processes that occur at cell tips.
Overexpression of other kelch family members (Tea1,
Kel1p, Kel2p and NRP/B) in their respective cell types
led to various changes in cell morphology (Mata and
Nurse, 1997; Philips and Herskowitz, 1998; Kim et al.,
1998). This suggests that the broad function of the
kelch family proteins may be in creating and
modulating various aspects of cell shape and structure.
Materials and methods
Cell lines
208F is a subclone of the Rat-1 ®broblast cell line originally
obtained from K Quade. FBR cell lines were originally
obtained from Tom Curran and are transformed nonproducer 208Fs infected with FBR-MuSv. The Ras transformants in this study were generated by transfections of 208F
with a ki-ras expressing construct. All cell lines were routinely
passaged before con¯uence and maintained in Dulbecco's
modi®ed Eagle medium (DMEM; Life Technologies) supplemented with 10% foetal calf serum (Advanced Protein
Products Ltd) at 378C in 5% CO2.
Isolation of rat and human krp1 cDNAs
A lZap11 FBR cDNA library was produced using the
protocols described by Stratagene. Approximately 106 plaque
forming units from the library were screened with the 32Plabelled cDNA clone krp1. After three rounds of screening six
clones were isolated and excised from lambda phage in the
form of the plasmid pBluescript, following the protocols
described by Stratagene. The clones were sequenced on both
strands using the ABI automatic sequencer 373A and 377.
Human krp1 was isolated by screening approximately 106
plaque forming units from a lZap11 human skeletal muscle
library (Stratagene), with 32P-labelled rat krp1 cDNA. Four
positive clones were isolated and were characterized using the
methods described above.
While this work was in progress, data base searches
revealed that a number of muscle speci®c cDNAs have been
identi®ed (Sarcosin, AA192617, AA22114096, AA196024)
which show high levels of sequence similarity to human
krp1 (Hillier et al., 1996; Taylor et al., 1998). Analyses of
these sequences revealed that none of them contain as long
an open reading frame as the one predicted from clones
isolated in this work, suggesting that they represent partial
cDNAs.
Northern blot analysis
Total RNA was isolated using RNAzol B according to the
manufacturer's protocol (Biogenesis). For Northern blot
analysis, 10 mg of total RNA was separated electrophoretically as described by Sambrook et al. (1989). The RNA was
stained with ethidium bromide, photographed, and then
transferred on to Hybond N (Amersham) as described by
Oncogene
Sambrook et al. (1989). Membranes were UV cross-linked
with a Stratalinker 1800 (Stratagene). The krp1 cDNA probe
was 32P-labelled with a random priming kit from Pharmacia.
Hybridization and stringency washes were performed as
previously described (Sambrook et al., 1989).
Preparation of Anti-Krp1 antiserum
To prepare the His-Krp1 tagged recombinant protein, the
complete open reading frame of krp1 was ampli®ed by PCR
with pkrp14 as a template using the following primers: 5k, 5'GTC TCG AGG TAC CTA AGC AGC GG-3' and the T7
primer of pBluescript. The PCR product was cloned into
pCR-SCRIPT (Stratagene) then subcloned into the KpnI site
of pTrcHisC (Invitrogen). Sequencing was performed to
con®rm that no errors had been introduced by the Pfu
polymerase and to determine the correct orientation of the
insert. Recombinant His-Krp1 was overexpressed in E. coli
strain DH5a and anity puri®ed on a Ni column
(Invitrogen). The puri®ed protein was then used to raise
polyclonal rabbit antisera to Krp1. To determine the
speci®city of the anti-Krp1 antiserum, Western blot analysis
of the bacterially produced His-Krp1 was performed and was
probed with this antibody. A single band of 68 kDa was
detected which is the predicted size of the Krp1 polypeptide,
con®rming that the antibody recognized bacterially produced
His-Krp1.
Western blot analysis
Cells were washed twice in PBS and then lysed in EBC bu€er
(50 mM Tris-HCL pH 8, 120 mM NaCl, 0.5% Nonidet P-40,
1 mM EDTA) containing 10 mg/ml aprotinin, leupeptin and
1 mM PMSF (Hernandez et al., 1997). Protein concentrations
were measured with a commercial reagent (Bio-Rad) and
50 mg of proteins were electrophoresed on either a SDS-7.5%
or 10% polyacrylamide gel using standard methods. Proteins
were transferred to polyvinylidene di¯uoride membranes
(Immunobilon P:Millipore) with a transblot wet blotting
apparatus (Bio-Rad). Membranes were blocked in blocking
bu€er (PBS containing 0.1% Tween 20 and 5% nonfat milk)
for 1 h and probed with the following antibodies at the given
dilution in blocking bu€er: anti-Krp1 rabbit polyclonal
antiserum at a 1 in 2500; anti-myc (9E10; Invitrogen) 1 in
100 dilution; anti-Erk2 (Transduction Laboratories) 1 in
5000. Membranes were then rinsed three times in blocking
bu€er followed by a 1 h incubation with the secondary
antibodies, either peroxidase-labelled goat anti-rabbit or
peroxibase-labelled goat anti-mouse at a 1 in 5000 dilution
in blocking bu€er. Three ®nal washes were performed in PBS
containing 0.1% Tween 20 and the blots were visualized
using ECL (Amersham).
Construction of Krp1 fusions
The Krp1-myc construct was derived by PCR ampli®cation
with pkrp14 as a template using the following primers: 5k, 5'AGA AGC TTA TGG ATT CCC AGC GG-3' and 3k, 5'CCA AGC TTT AGT TTA GAC AG-3'. This PCR product
was cut with HindIII and ligated into the HindIII site of
expression vector pcDNA3.1/Myc-HisA (Invitrogen) creating
Krp1-myc. The Krp1 variants were also generated by PCR
using pkrp14 as a template. Krp1poz-myc was generated
using the following primers: p5, 5'-TAT CTA GAA TGG
ATT CCC AGC GGG AGC TTG-3' and p3, 5'-ATT CTA
GAA TGA CAG AGA TCA GCT CCT GTG GGG AC-3'.
The Krp1poz PCR product was cut with XbaI and ligated
into the XbaI site of pcDNA3.1/Myc-His B (Invitrogen). To
create Krp1repeat-myc the following primers were used: 5'
repeat, 5'-TAT CTA GAA TGA ATC ATT CCA GTA TTG
TTA CC-3' and 3k, 5'-CCA AGC TTT AGT TTA GAC
AG-3'. This PCR product was cut with XbaI and HindIII and
Krp1 contributes to pseudopod elongation
HJ Spence et al
was cloned into XbaI and HindIII site of pcDNA3.1/MycHisA. To check the orientation of all constructs, as well as to
determine whether errors had occurred during PCR, all the
constructs were completely sequenced on both strands using
ABI Automatic sequencer 373A or 377.
Transfectionos of Krp1 constructs
Cells were transfected with 1.5 mg of the relevant plasmid
using Superfect lipid transfection reagent (Qiagen). Immuno¯uorescence of transfected cells was performed 24 h later by
either pre-permeablizing the live cells in Triton X-100/PBS
for 1 min before ®xing them with 4% paraformaldehyde for
15 min at 378C, or they were ®xed without any prepermeablization. After ®xing, the cells were washed extensively in PBS and permeablized in PBS/20 mM glycine/0.05%
triton X-100 for 5 min if they had not been pre-permeablized.
Blocking of non-speci®c binding was performed for 20 min in
blocking bu€er (PBS/0.5% BSA/10% FCS), followed by a
1 h incubation at room temperature of the relevant primary
antibody diluted in blocking bu€er: a 1 in 400 dilution of
anti-Krp1; 1 in 30 dilution of anti-tubulin monoclonal
antibody (a gift from E Schieble); and 1 in 100 dilution of
anti-myc monoclonal antibody. After extensive washing in
blocking bu€er, the relevant secondary antibody was added
in blocking bu€er and, when needed, either TRITC- or
FITC-phalloidin was added and incubated at 45 min at room
temperature. Three ®nal washes in PBS were performed and
then the cells were mounted with Vectashield (Vector
Laboratories). Cells were visualized on a Bio-Rad MRC
600 confocal illumination unit attached to a Nikon Diaphot
inverted microscope with various magni®cations.
1275
Acknowledgments
We would like to thank Prof. JA Wyke and Drs MC Frame
and G Stapleton for critical reading of the manuscript. We
also thank P McHardy for assistance with microscopy.
Thanks to E Schieble of the Beatson Institute for the
tubulin monoclonal antibody. We gratefully acknowledge
the Cancer Research Campaign for their ®nancial support.
References
Adams JC, Seed B and Lawler J. (1998). EMBO J., 17,
4964 ± 4974.
Albagli O, Dhordain P, Deweindt C, Lecocq G and Leprince
D. (1995). Cell Growth Di€er., 6, 1193 ± 1198.
Bushel P, Kim JH, Chang W, Catino JJ, Ruley HE and
Kumar CC. (1995). Oncogene, 10, 1361 ± 1370.
Curran T and Franza Jr. BR. (1988). Cell, 55, 395 ± 397.
Curran T and Verma IM. (1984). Virology, 135, 218 ± 228.
De Cesare D, Vallone D, Caracciolo A, Sassone-Corsi P,
Nerlov C and Verde P. (1995). Oncogene, 11, 365 ± 376.
Diatchenko L, Lau YF, Campbell AP, Chenchik A,
Moqadam F, Huang B, Lukyanov S, Lukyanov K,
Gurskaya N, Sverdlov ED and Siebert PD. (1996). Proc.
Natl. Acad. Sci. USA, 93, 6025 ± 6030.
Domann FE, Levy JP, Birrer MJ and Bowden GT. (1994).
Cell Growth Di€er., 5, 9 ± 16.
Dong Z, Crawford HC, Lavrovsky V, Taub D, Watts R,
Matrisian LM and Colburn NH. (1997). Mol. Carcinog.,
19, 204 ± 212.
Finkel MP, Reilly CA and Biskis BO. (1975). Front. Radiat.
Ther. Oncol., 10, 28 ± 39.
Guirguis R, Margulies I, Taraboletti G, Schi€mann E and
Liotta L. (1987). Nature, 329, 261 ± 263.
Hay ED. (1989). Cell Motil. Cytoskel., 14, 455 ± 457.
Hennigan RF, Hawker KL and Ozanne BW. (1994).
Oncogene, 9, 3591 ± 3600.
Hernandez MC, Andres-Barquin PJ, Holt I and Israel MA.
(1998). Exp. Cell. Res., 242, 470 ± 477.
Hernandez MC, Andres-Barquin PJ, Martinez S, Bulfone A,
Rubenstein JL and Israel MA. (1997). J. Neurosci., 17,
3038 ± 3051.
Hillier LD, Lennon G, Becker M, Bonaldo MF, Chiapelli B,
Chissoe S, Dietrich N, DuBuque T, Favello A, Gish W,
Hawkins M, Hultman M, Kucaba T, Lacy M, Le M, Le N,
Mardis E, Moore B, Morris M, Parsons J, Prange C,
Rifkin L, Rohl®ng T, Schellenberg K, Marra M. (1996).
Genome Res., 6, 807 ± 828.
Johnson R, Spiegelman B, Hanahan D and Wisdom R.
(1996). Mol. Cell. Biol., 16, 4504 ± 4511.
Jooss KU and Muller R. (1995). Oncogene, 10, 603 ± 608.
Kaplan PL and Ozanne B. (1983). Cell, 33, 931 ± 938.
Karin M, Liu Z and Zandi E. (1997). Curr. Opin. Cell. Biol.,
9, 240 ± 246.
Kerr LD, Holt JT and Matrisian LM. (1988). Science, 242,
1424 ± 1427.
Kim TA, Lim J, Ota S, Raja S, Rogers R, Rivnay B,
Avraham H and Avraham S. (1998). J. Cell. Biol., 141,
553 ± 566.
Lamb RF, Hennigan RF, Turnbull K, Katsanakis KD,
MacKenzie ED, Birnie GD and Ozanne BW. (1997a).
Mol. Cell. Biol., 17, 963 ± 976.
Lamb RF, Ozanne BW, Roy C, McGarry L, Stipp C,
Mangeat P and Jay DG. (1997b). Curr. Biol., 7, 682 ± 688.
Legg JW and Isacke CM. (1998). Curr. Biol., 8, 705 ± 708.
Liotta LA, Steeg PS and Stetler-Stevenson WG. (1991a).
Cell, 64, 327 ± 336.
Liotta LA, Stracke ML, Aznavoorian SA, Beckner ME and
Schi€mann E. (1991b). Semin. Cancer Biol., 2, 111 ± 114.
Lloyd A, Yancheva N and Wasylyk B. (1991). Nature, 352,
635 ± 638.
Malliri A, Symons M, Hennigan RF, Hurlstone AF, Lamb
RF, Wheeler T and Ozanne BW. (1998). J. Cell. Biol., 143,
1087 ± 1099.
Mata J and Nurse P. (1997). Cell, 89, 939 ± 949.
Nabi IR. (1999). J. Cell. Sci., 112, 1803 ± 1811.
Philips J and Herskowitz I. (1998). J. Cell. Biol., 143, 375 ±
389.
Rapp UR, Troppmair J, Beck T and Birrer MJ. (1994).
Oncogene, 9, 3493 ± 3498.
Robinson DN and Cooley L. (1997). J. Cell. Biol., 138, 799 ±
810.
Saez E, Rutberg SE, Mueller E, Oppenheim H, Smoluk J,
Yuspa SH and Spiegelman BM. (1995). Cell, 82, 721 ±
732.
Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular
Cloning: a Laboratory Manual, Cold Spring Harbor
Laboratory, Cold Spring Harbor NY.
Schonthal A, Herrlich P, Rahmsdorf HJ and Ponta H.
(1988). Cell, 54, 325 ± 334.
Soltysik-Espanola M, Rogers RA, Jiang S, Kim TA,
Gaedigk R, White RA, Avraham H and Avraham S.
(1999). Mol. Biol. Cell., 10, 2361 ± 2375.
Suzuki T, Murakami M, Onai N, Fukuda E, Hashimoto Y,
Sonobe MH, Kameda T, Ichinose M, Miki K and Iba H.
(1994). J. Virol., 68, 3527 ± 3535.
Taylor A, Obholz K, Linden G, Sadiev S, Klaus S and
Carlson KD. (1998). Mol. Cell. Biochem., 183, 105 ± 112.
Tsukita S, Oishi K, Sato N, Sagara J and Kawai A. (1994). J.
Cell. Biol., 126, 391 ± 401.
Oncogene
Krp1 contributes to pseudopod elongation
HJ Spence et al
1276
Oncogene
Van Beveren C, Enami S, Curran T and Verma IM. (1984).
Virology, 135, 229 ± 243.
Vega LR and Solomon F. (1997). Cell, 89, 825 ± 828.
Varkey JP, Muhlrad PJ, Minniti AN, Do B and Ward S.
(1995). Genes Dev., 9, 1074 ± 1086.
Way M, Sanders M, Chafel M, Tu YH, Knight A and
Matsudaira P. (1995a). J. Cell. Sci., 108, 3155 ± 3162.
Way M, Sanders M, Garcia C, Sakai J and Matsudaira P.
(1995b). J. Cell. Biol., 128, 51 ± 60.
Xue F and Cooley L. (1993). Cell, 72, 681 ± 693.
Yonemura S, Hirao M, Doi Y, Takahashi N, Kondo T and
Tsukita S. (1998). J. Cell. Biol., 140, 885 ± 895.
Yonemura S and Tsukita S. (1999). J. Cell. Biol., 145, 1497 ±
1509.