3145
Journal of Cell Science 111, 3145-3156 (1998)
Printed in Great Britain © The Company of Biologists Limited 1998
JCS3826
ELAV protein HuA (HuR) can redistribute between nucleus and cytoplasm and
is upregulated during serum stimulation and T cell activation
Ulus Atasoy1,*, Janice Watson1,*, Dhavalkumar Patel2,3 and Jack D. Keene1,3,‡
Departments of 1Microbiology, 2Immunology and 3Medicine, Duke University Medical Center, Durham, NC 27710 USA
*These two authors contributed equally to this work
‡Author for correspondence
Accepted 28 August; published on WWW 14 October 1998
SUMMARY
ELAV proteins are implicated in regulating the stability
and translation of cytokine and growth regulatory mRNAs
such as GM-CSF, IL-2, c-myc, c-fos and GLUT1 by binding
to their AU-rich 3′UTRs. The tissue-specific ELAV protein
HuB (aka. Hel-N1) is predominantly cytoplasmic and has
been shown to stabilize GLUT1 and c-myc mRNAs and to
increase their translation following ectopic expression in
3T3-L1 cells. We report that the most widely expressed
mouse ELAV protein, mHuA, is predominately nuclear in
cultured NIH-3T3 cells, but is localized in the cytoplasm
during early G1 of the cell cycle. Therefore, much like the
primarily cytoplasmic HuB, HuA becomes temporally
localized in the cytoplasm where it can potentially regulate
the stability or translation of bound mRNAs. Moreover, we
report that stimulation of mouse spleen cells using either
mitogenic or sub-mitogenic levels of anti-CD3/CD28
resulted in a dramatic increase in the level of HuA.
Upregulation of HuA corresponds to previously
documented increases in cytokine expression which are due
to increased mRNA stability following T cell activation.
Consistent with these findings, HuA was down regulated in
quiescent cells and upregulated in 3T3 cells following
serum stimulation. The increase of murine HuA during the
cell cycle closely resembles that of cyclin B1 which peaks
in G2/M. Together with our earlier studies, these data
indicate that mammalian ELAV proteins function during
cell growth and differentiation due in part to their effects
on posttranscriptional stability and translation of multiple
growth regulatory mRNAs. This supports the hypothesis
that ELAV proteins can function as transacting factors
which affect a default pathway of mRNA degradation
involved in the expression of growth regulatory proteins.
INTRODUCTION
distribution among the four classes and their temporal
appearance during embryonic development (Good, 1995)
suggests that protein interaction signals outside of the RRMs
could endow each member with unique functional properties
(King et al., 1994; Gao and Keene, 1996; Antic and Keene,
1998).
Although none of the mammalian ELAV proteins localizes
to a single compartment, the tissue-specific forms HuB, HuC
and HuD, are predominantly cytoplasmic in cultured neurons
and medulloblastoma cells and are distributed in clusters along
the dendrites of neurons (Gao and Keene, 1996; Antic and
Keene, 1998; reviewed by Antic and Keene, 1997). While
biochemical data suggest that the more ubiquitously expressed
form, HuA (or HuR), is predominately nuclear (Vakaloupolu
et al., 1991; Myer et al., 1997; Ma et al., 1996), it is difficult
to imagine how HuA could be involved in mRNA stability and
translation unless it has access to the cytoplasmic machinery.
In this study, we demonstrate that the mouse ELAV protein,
mHuA, distributes throughout the cell during breakdown of the
nuclear membrane at mitosis and is delayed in its return to the
nuclei of daughter cells in G1. Additionally we show
redistribution of HuA to the cytoplasm following treatment
Embryonic lethal abnormal vision (ELAV) RNA binding
proteins are thought to be involved in cellular growth and
differentiation via posttranscriptional regulation (reviewed by
Antic and Keene, 1997). ELAV proteins are members of the
RNA recognition motif (RRM) superfamily (Query et al.,
1989) and each consists of three highly conserved RRMs
(Robinow et al., 1988; Szabo et al., 1991; King et al., 1994).
Although their exact functions are not known ELAV proteins
have been implicated in the stability and translation of early
response gene (ERG) messenger RNAs, such as those encoding
protooncoproteins and cytokines (Levine et al., 1993; Gao et
al., 1994; Jain et al., 1997; Myer et al., 1997). Direct binding
of ERG mRNA 3′ untranslated regions (UTR) by ELAV
proteins Hel-N1 and Hel-N2 (here termed HuB) was
demonstrated following in vitro selection from combinatorial
RNA libraries (Levine et al., 1993; King et al., 1994; Gao et
al., 1994). To date, four classes of ELAV proteins have been
described and all appear to have similar properties of binding
to AU-rich elements (ARE) in these 3′UTRs (reviewed by
Antic and Keene, 1997). However, differences in tissue
Key words: RNA binding protein, RRM, AU-rich elements, 3′
Untranslated region, RNA stability, Translation, Cytokine,
Paraneoplastic disease
3146 U. Atasoy and others
with transcription inhibitors actinomycin D (ActD), and 5,6dichloro-β-d-ribofuranosyl benzimidazole (DRB) and reimport
into the nucleus upon washout of DRB. Moreover, we find that
in untreated NIH-3T3 cells, endogenous HuA can be found
predominantly in the cytoplasm during early G1 of the cell
cycle.
Our previous demonstration that ectopic expression of the
human ELAV protein HuB (Hel-N1) can result in increased
stability and translation of the ERG mRNA encoding
endogenous GLUT1 protein (Jain et al., 1997), as well as,
endogenous c-myc mRNA (unpublished data) was unexpected
since ARE-binding proteins were assumed to help destabilize
target mRNAs (Vakaloupolu et al., 1991). These observations
were confirmed by the finding that vascular endothelial growth
factor (VEGF) mRNA is stabilized during hypoxia following
ectopic HuA (HuR) expression (Levy et al., 1998).
Accumulated data suggest that ELAV proteins may function as
transacting factors which modulate a default pathway of AREmediated mRNA degradation.
It is well established that cytokine mRNAs are stabilized
following proliferation and artificial stimulation of T cells
(Lindsten et al., 1989). Bohjanen et al. (1991) demonstrated
following T cell activation that an inducible protein, AU-B,
cross-links to the AREs of RNAs encoding cytokines such as
IL-2, TNF-alpha and GM-CSF. However, this protein has not
been identified. We have investigated the expression of ELAV
proteins, in particular HuA, during proliferation and T cell
activation to determine whether its levels increase during
periods following activation and subsequent cytokine
expression. We report that following costimulation of murine
T cells by CD3 and CD28 activation, levels of mHuA increased
significantly, as one might expect of a transacting factor(s)
involved in ARE-mediated stabilization of cytokine mRNAs.
Since activation of T cells results in their recruitment from
quiescence into the cell cycle, we examined HuA expression
in NIH-3T3 cells. We report that rapidly growing 3T3 cells
have increased levels of the mHuA protein, while quiescent
cells have lower levels. The kinetics of mHuA expression were
found to be similar to those of cyclin B1 expression, which is
maximally expressed during G2/M, but then diminishes.
In total, these results are compatible with our previous
findings that ELAV proteins, as exemplified by HuB,
participate in the stabilization of ERG mRNAs presumably via
binding to ARE 3′UTR sequences in the cytoplasm. These
findings suggest a role for the ELAV protein, HuA, in cellular
proliferation. Furthermore, by its increased levels in the
cytoplasm during and following cell division, mHuA may have
a temporal role in mediating expression of mRNAs via their
stabilization or translation. These findings have important
implications for the regulation of cell proliferation during
homeostatic growth, as well as during immunoregulation.
MATERIALS AND METHODS
Cloning of HuA forms of ELAV from mouse brain, spleen
and testis
The gene for mHuA (accession number U65735) was cloned by using
5′ and 3′ RACE (rapid amplification of cDNA ends) techniques using
the following Clontech products: Marathon-Ready murine spleen and
brain cDNA libraries, Advantage KlenTaq Polymerase Mix
(Clontech, Palo Alto, CA). Gene specific HuA primers (GSP1
5′AGAACCTGAATCTCTGTGCCTGG, position 307; GSP2
(5′CCTAGCAGGCGAGTGGTACAGCT, position 261) were
designed from the partial length murine HuA sequence (kind gift of
Peter Good, accession number U17595). For 5′ RACE, two nested
PCR were performed following the recommendations in the Clontech
manual. For the first reaction, forward primer AP1 and reverse primer
GSP1 were used. PCR was performed using ‘touchdown’ parameters:
95°C for 2 minutes, followed by 5 cycles of: 100°C for 2 seconds,
72°C for 1 minute then 5 cycles of 100°C for 2 seconds, 70°C for 1
minute; then 20 cycles of 100°C for 2 seconds, 68°C for 1 minute.
Two additional rounds of nested PCR were performed as per the
Clontech protocol using the following primers: forward primer AP2;
reverse primer GSP2. The final round of nested PCR used AP2 and
5′CCACTGAGCTCGGGCGAGCATA, position 422, human
sequence HuR, accession U38175). Nested 3′RACE was performed
using published Clontech touchdown parameters. For the first
reaction, the following primers were used: reverse primer AP1 and
forward primer GSP1 5′CGGTCAGAAGCAGAAGAG, position
113. For nested reaction: reverse primer AP2; forward primer GSP2
5′CTCTCCTCTCGCAGCTGTACCACTAGGTTCT, position 227.
PCR products were electrophoresed, transferred to Nytran
membranes and probed with 32P-end labeled HuA primers. Positive
bands were subcloned into the pNOTA vector (5Prime-3Prime) and
sequenced in both directions with universal M13 primers (−20 and −
48 mers) using ABI technology. Simultaneously, full-length clones
were obtained from the Marathon brain and spleen cDNA libraries
using the following primers (derived from human clone, HuR (HuA),
accession U38175) forward primer, 5′ ACAATGTCTAATGGTTATGA (position 116) and reverse primer, 5′ GAGCGAGTTATTTGTGGGA (position 1,106). The reactions were cycled using
the published Clontech protocol. A solitary band of the expected size
(approximately 1 kb), was confirmed by Southern analysis to be HuA
specific from both brain and spleen libraries. The full-length 1 kb
PCR products were subcloned into pNOTA (5Prime-3Prime,
Boulder, CO) and sequenced in both directions using the ABI method
and universal M13 and HuA specific primers. The sequences
obtained from the 5′, 3′ RACE reactions were identical to those
obtained from the full-length clones generated using human primers.
Additionally, putative full-length clones were obtained from mouse
testis cDNA (Clontech, Palo Alto, CA) using the published Clontech
protocol and the human specific primers described above. 5 µl of the
PCR reaction (representing 10% of the total volume) was
electrophoresed, blotted onto Nytran and probed with an internal HuA
primer. A prominent and single 1 kb band was identical to the fulllength clones obtained from mouse brain and spleen cDNA (data not
shown).
Derivation of mHuB probe
Gene specific primers (forward primer 5′ GCGATCAACACTCTGAATGG and reverse primer 5′ GTCACTGGACCAGCTGTTCT)
were designed from a partial length rat ELAV clone which had been
used to clone the human ELAV gene, Hel-N1 (King et al., 1994).
Reverse transcriptase-PCR was performed on mouse brain poly A+
(Clontech) using published Perkin Elmer protocols, the Gene Amp
RNA PCR kit (Norwalk, CT) and the rat primers above. Gel
electrophoresis of the final PCR product revealed a solitary band of
the expected size which was subcloned into pNOTA vector (5Prime3Prime) and confirmed by sequencing to be the murine homolog of
rat HuB cDNA. The plasmid was digested with BamHI and the ds
cDNA for mouse HuB (Mel-N1) cut out of the gel and used in random
prime labeling with [32P]dCTP using the random nonomer kit from
5Prime-3Prime and purified using Stratagene NucTrap columns
(LaJolla, CA).
Northern blot analysis
Northern blots of mouse and human tissues, as well as human tumor
Redistribution and upregulation of HuA (HuR) 3147
lines (Clontech) were hybridized with a PCR-generated 358 bp HuA
probe or with the mHuB probe described above. The PCR product
was subcloned into the pNOTA vector, sequenced to verify its identity
and then used as a template for further reactions. The double-stranded
DNA was labeled by the random nonomer method (Stratagene) and
the northern blots were hybridized with probes using ExpressHyb
hybridization solution (Clontech) and subjected to autoradiography
for various times. All northern blots contained 2 µg of poly(A)+ RNA
per lane and have been quality tested using actin DNA probe as
control to ensure equivalent amounts of RNA loaded per lane.
Cell lines
Cell lines were obtained from ATCC except where described and
cultured under ATCC recommended conditions unless indicated
otherwise. HFF (human foreskin fibroblasts) were obtained from Dr
Ross McKinney (Duke University Medical Center).
Cell cycle synchronization protocols
Serum starvation of NIH-3T3 cells
NIH-3T3 cells growing in DMEM+10% bovine calf serum (BCS)
were trypsinized, split 1:10 and seeded into P150 Petri dishes. After
overnight growth they were washed twice in DMEM and then starved
for 48 hours in DMEM+0.25% BCS. After serum starvation, the
medium was aspirated and fresh DMEM+10% BCS was added. Cells
were harvested at various time points (see Results) by trysinization,
washed with ice-cold PBS and then lysed using triple-detergent lysis
buffer (see below). At each time point, an aliquot of cells was stained
with propidium iodide (PI) and analyzed by FACS to confirm stages
of cell cycle. Typically, 86-95% of the cells were in G0 after 48 hours
of serum starvation.
Hydroxyurea block of NIH-3T3 cells
Exponentially growing cells in DMEM+10% BCS were serum starved
as above for 32-36 hours. These plates were aspirated, washed two
times with DMEM and treated with DMEM+10% BCS containing
hydroxyurea (final concentration, 2 mM; Sigma) for 21 hours. The
hydroxyurea was removed by washing the cells twice with DMEM.
DMEM+10% BCS was added back for varying times. Cells were
harvested and lysed as described above. Aliquots of unlysed cells were
assayed by FACS PI staining to confirm stages of cell cycle. The vast
majority of the cells were in G1/S after hydroxyurea treatment.
Activation of murine splenocytes
Single-cell suspensions of freshly teased and Ficoll-Hypaque (FicoLite, Atlanta Biologics, Atlanta, GA) isolated splenocytes from
C57B6 mice were prepared. These unstimulated murine splenocytes
were suspended at 2×106/ml with different concentrations of anti-CD3
(145-2C11 from Jeff Bluestone, Chicago, IL) and/or anti-CD28 (37.51
from J. P. Allison, Berkeley, CA) monoclonals. Anti-CD3 at 1:80 is
mitogenic and gives 100% response on [3H]thymidine uptake; at
1:6,400 it is sub-mitogenic and gives a thymidine response of less than
10% of maximal. Anti-CD28 at 1:800 does not stimulate by itself but
synergizes with anti-CD3 at 1:6,400 to give maximum response (Lee
et al., 1998). Aliquots of cells were removed at various points, washed
and lysed using triple detergent buffer and quantitated using Bradford
analysis. Equal amounts of protein were electrophoresed on 12% SDS
gel, transferred onto nitrocellulose and blotted with antisera to mHuA
or tubulin.
Antibody production, western blot analysis and
quantitation
cDNA for HuA was cloned into plasmid pGexKG for overexpression
of the recombinant GST-tagged protein. Thrombin was used in a
partial digest to remove the GST tag and the band corresponding to
full length HuA cut out of a SDS-PAGE gel and used for polyclonal
rabbit antibody production (Cocalico, Inc.). For immunofluorescence
staining the antibody was further purified over an affinity column
(ImmunoLink® Plus, Pierce) of GST-HuA and eluted with 0.1 M
glycine, pH 2.6, and neutralized with Tris-HCl, pH 9.5. Western blot
analysis of cell lysates demonstrated that a band of the expected size
for HuA was removed from the antibody flow-through and presents
as a solitary band in the predicted fractions eluted from the column.
By western blot analysis, HuA antisera did not cross-react with
endogenous forms of other neuronal specific ELAV family members.
However, prior to affinity purification we consistently observed a band
of approximately 28 kDa which may represent a degradation product,
but also could be an additional ELAV family member.
Whole cell lysates were prepared at 107 cells per ml in triple lysis
buffer of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 1%
Nonidet P40, 0.5% sodium deoxycholate and protease inhibitors (1
mM PMSF, 10 µg/ml aprotenin, 1 µg/ml leupeptin, 1 µg/ml pepstatin
from Sigma) and sonicated for 10 seconds three times. This method
solubilized all cellular proteins. Protein concentrations were
determined by Bradford assay and 20-40 µg loaded per lane in 1×
Laemmli buffer. Following transfer to nitrocellulose and blocking
with 5% non-fat milk in TBS-Tween, primary antibody incubations
were carried out in TBS-Tween for 2 hours at room temperature or
overnight at 4°C. The secondary antibodies conjugated to horseradishperoxidase (HRP) (Amersham) were incubated at room temperature
for 30 minutes and blots were developed using the ECL system
according to the manufacturer’s directions (Amersham). Anti-tubulin
murine monoclonal antibody was used at a dilution of 1:2,000
(Amersham).
Western blots were scanned densitometrically and HuA and tubulin
signals in the linear range were quantitated using ImageQuant
(Molecular Dynamics). To calculate relative HuA signal in
densitometric units, the values were normalized to the tubulin signals.
Limiting dilution western blots for HuA and tubulin were done using
ECL and 125I-labelled Protein A. The 125I blots were scanned by
phosphoroimager and quantitated using ImageQuant. Both methods
yielded results consistent with one another.
Immunofluorescence staining
3T3 cells were grown in chamber slides (Nunc). Cells were washed
in PBS and fixed in paraformaldehyde in poly-lysine and stained as
described previously (Gao and Keene, 1996). A further fixation in
50% v/v acetone/methanol for 1 minute was found to increase the
intensity of staining for HuA. Affinity purified antibody was used
diluted 1:1 in blocking buffer. For identification of 3T3 cells in S
phase combined with HuA staining, bromodeoxyuridine (BrdU) was
included in the culture medium for one hour as described by
DeGregori et al. (1995). DAPI was added at 2 mg/ml for 2-5 minutes
at room temperature, following immunostaining protocols where
described. There was no staining when anti-HuA primary antibody
was omitted. Secondary antibodies used included a donkey antirabbit-Texas Red (Jackson ImmunoResearch, Inc.) and an FITCconjugated mouse monoclonal against BrdU (Amersham).
Monoclonal antibody 4B10 to hnRNPA1 was provided by Gideon
Dreyfuss (University of Pennsylvania).
Cell fractionation
3T3 cells were separated into fractions based on the Weinberg and
Penman (1968) method. Cells were resuspended at 107/ml in a
hypotonic buffer of 10 mM NaCl, 10 mM Tris-HCl, pH 7.4, 1.5 mM
MgCl2 with protease inhibitors and allowed to swell for 5 minutes.
Lysis was achieved by addition of 0.5% Nonidet NP40 and nuclei
were pelleted at 1,000 g for 2 minutes. The supernatant was removed
and stored as cytoplasmic fraction C1. The outer nuclear membrane
(and attached material) was removed by washing nuclei in the above
buffer containing 1% NP40 and 0.5% sodium deoxycholate. The
solubilized supernatant after centrifugation at 1,000 g for 2 minutes
was designated C2. The nuclear pellet was sonicated in triple lysis
buffer and the fraction termed N. SDS-PAGE gels were loaded with
equal cell equivalents per lane. Transcription was inhibited by treating
3148 U. Atasoy and others
cells with 5 µg/ml actinomycin D for up to 3 hours. 5,6-dichloro-βd-ribofuranosyl benzimidazole (DRB) (Sigma) was also used to
inhibit transcription at 100 µM for 3 hours and was reversed by
washing out the drug following a further incubation in medium for 3
hours.
RESULTS
Cloning and expression of mHuA
We derived full length cDNAs (from brain and spleen)
encoding the mouse ELAV protein corresponding to Xenopus
elrA using the partial mouse clone derived by Dr Peter Good
at the National Institutes of Health (Good, 1995). Expression
of elrA and its human counterpart HuR (HuA) has been
reported at the mRNA level (Good, 1995; Ma et al., 1996;
Okano and Darnell, 1997). elrA has also been detected in
Xenopus oocytes by immunoblotting (Wu et al., 1997). We
investigated expression of HuA protein in a variety of
mammalian cell types using high titer polyclonal rabbit
antisera prepared against recombinant mHuA and found
relative differences between levels in transformed (Fig. 1A and
B, lanes 3 and 4) and nontransformed (Fig. 1B, lanes 1 and 2)
cell lines. Human foreskin fibroblasts and embryonic kidney
fibroblasts were chosen to illustrate extreme examples (Fig.
1B, lanes 1 and 4, respectively). In all cases examined, a
discrete protein of the expected size of 34 kDa was detected
using affinity purified mouse HuA antiserum. These and other
data not shown demonstrated that the antiserum recognized
mouse HuA (mHuA) and human HuA equally well. Since there
were differences in relative protein expression levels among
cell types, it was important to examine mHuA mRNA
expression in a variety of cell lines.
Fig. 1. Immunoblot of exponentially growing cells showing
examples of relative levels of HuA. Extracts from the following cell
lines are shown in A; lanes 1, HL60 (human lymphoblastoid line); 2,
Namalwa, Burkitt’s lymphoma; 3, 245MG, glioblastoma; 4, DT2H3,
neuroblastoma. B: 1, HFF, human foreskin fibroblasts; 2, NIH-3T3;
3, HeLa; 4, 293T, human embryonic kidney. 30 mg of protein extract
per cell line was separated on a 15% SDS-acrylamide gel, blotted
with anti-mHuA rabbit serum and visualized by ECL. In B, tubulin
levels on the same blot are shown for comparison.
Brain and testes express unique forms of HuA mRNA
By northern blot analysis we assessed the expression of both
HuA and HuB in a variety of mouse tissues. As shown in Fig.
2 quantitative and qualitative differences were also observed
among the tissues examined. In the majority of tissues HuA
was expressed as a single unique mRNA of approximately 2.6
kb (Fig. 2A, upper panel) that varied in relative intensity
among cell lines. However, in brain and testes the 2.6 kb band
was not expressed or was barely detectable, while new and
unique mRNA species of 6.6 kb and 1.8 kb appeared. For
comparison upon reprobing, the same northern blot
demonstrated the expression of HuB mRNAs of approximately
4.5 kb in brain and testes (Fig. 2A, lower panel). As expected,
other tissues and cell lines that expressed HuA did not have
high levels of mHuB mRNA under the same conditions of
blotting. These findings were confirmed further by examining
a variety of human tissues and tumors for expression of HuA
mRNAs (Fig. 2B, both upper and lower). The mRNAs detected
in these human tissues corresponded precisely with those
observed with the mouse tissues including brain and testes.
Thus, it is clear that HuA represents a generic or ubiquitous
member of the mammalian ELAV family which is detectable
in every tissue of mouse and human examined. Previously
Okano and Darnell (1997) reported the presence of a 1.8 kb
band of HuA mRNA in testes but did not report a higher sized
mRNA transcript in brain. Ma et al. (1996) reported only a
single sized band using RT-PCR, which may not have been able
to detect all potential transcripts.
Full length mHuA clones from both brain and spleen had
identical sequences in the open reading frame (data not shown).
Additionally, full-length clones obtained from mouse testis
cDNA libraries by PCR and Southern blotting also were
identical in size to those obtained from mouse brain and spleen
(data not shown). At the protein level, different sized bands
were not observed in whole mouse brain extracts using our
polyclonal antibody raised against recombinant HuA. Indeed,
we have consistently observed a band of approximately 34 kDa
which corresponds to mHuA. Hence, it seems plausible that
the differences in transcript size in mouse brain, spleen and
testis are due to differences in the untranslated regions of the
mRNAs. Okano and Darnell (1997) have reported
heterogeneity in the 3′UTRs of mHuA. However, we have not
ruled out differential splicing or the existence of isoforms of
the HuA mRNA in these tissues. We conclude that there are
HuA transcripts in both mouse brain and spleen encoding
identical full-length open reading frames of the HuA protein.
Expression of HuA and mHuB during mouse
embryogenesis
Given the potential role of ELAV proteins in differentiation
(Robinow and White, 1991; Wu et al., 1998; reviewed by Antic
and Keene, 1997), we decided to probe northern blots
representing different stages of murine development with the
mHuA probe. There was a single band of the expected size (2.6
kb) on days 11, 15 and 17 (Fig. 2C, upper). However, there
was no mHuB transcript detected at day 7 (Fig. 2C, lower
panel, lane 1) which is consistent with the fact that the murine
nervous system begins to develop after day eight. Hence mHuB
is expressed about the time of the earliest development of
neurons in the murine central nervous system and is consistent
with the time of appearance of ELAV in Drosophila (Robinow
Redistribution and upregulation of HuA (HuR) 3149
Fig. 2. Northern blots showing variations in HuA mRNA species expressed in mouse (A, upper panel) and human (B, upper panel) proliferating
cell lines and in developing mouse embryos (C, upper panel). Lower panels in A and C show the same blots probed for mHuB (Mel-N1). A:
Various mouse tissues were probed for mRNAs using mHuA (upper) and mHuB (lower) cDNA probes. Lanes: 1, heart, 2, brain, 3, spleen, 4,
lung, 5, liver, 6, skeletal muscle, 7, kidney, 8, testis. B: mRNA from human tissues (lanes 1-10) and cell lines (lanes 11-18) were probed with
only mouse HuA cDNA. Lanes: 1, spleen; 2, thymus; 3, prostate; 4, testis; 5, ovary; 6, small intestine; 7, colon; 8, peripheral blood leukocyte;
9, heart; 10, brain; 11, HL-60; 12, HeLa; 13, CML/K562; 14, T lymphoblastic leukemia, MOLT-4; 15, Burkitt’s/Raji; 16, colorectal
adenocarcinoma, SW480; 17, lung carcinoma, A549; 18, melanoma, G361. C: mRNA from developing mouse embryos probed with cDNA of
mHuA (upper panel) and mHuB (lower panel). Lanes: 1, embryos from day 7; 2, embryos from day 11; 3, embryos from day 15; 4, embryos
from day 17.
et al., 1988). In contrast to the developmental expression of
mHuB, however, mHuA appeared to be constitutively
expressed throughout development as determined at day 7 (lane
1), the earliest time examined. Furthermore, mHuA continued
to be expressed throughout intrauterine development as would
be expected of a constitutive generic cell protein. The fact that
the 6.6 kb and 1.8 kb transcripts were not detected in
developmental northern blots is probably due to the
vanishingly small amounts of the tissues from brain and testes
represented during these early stages of development. We
conclude that mHuA is expressed at early stages of murine
development, which is consistent with its being generic and not
tissue-specific.
Expression of mHuA during cell cycle progression
in NIH-3T3 cells
Given the apparent variation in expression of mHuA protein
(Fig. 1) and RNA (Fig. 2) among the cell lines examined, we
wondered whether there may be differences in protein levels at
various proliferative stages for any given cell type. Therefore,
cell cycle synchronization experiments were carried out using
3T3 with two established protocols: either serum starvation or
hydroxyurea block, followed in both cases by serum
restimulation. Levels of mHuA were monitored by western
blotting using affinity purified antiserum. As shown in Fig. 3A,
starved cells in G0 showed very low levels of protein (lane 2),
but the levels gradually increased after serum was added back
to cells, peaking at 18 hours post stimulation (arrow, lane 5).
HuA levels as measured in arbitrary densitometric units
increased approximately 2-fold during these experiments.
FACS staining of these same cells with propidium iodide (PI)
(not shown), confirmed that serum starvation placed the cells
in G0 and upon serum re-stimulation they progressed through
the cell cycle normally. The 18 hour time point (lane 5), during
which mHuA levels reached maximum, represented G2/M with
41% of the cells in G2/M by FACS PI staining (data not
shown). We reproducibly observed an inexplicable slight
decrease in mHuA at 24 hours post serum addition (lane 8).
Nonsynchronous 3T3 cells when harvested during logarithmic
growth showed higher levels of mHuA (lanes 1 and 9).
Consistent with these findings, nonsynchronous cells that
reached 100% confluence showed lower levels of mHuA (lane
10). Therefore, although the degree of synchrony was not
optimal when using the serum starvation method, a definitive
correlation between the levels of mHuA and certain stages of
proliferation was evident.
We further synchronized 3T3 cells using the more efficient
method of hydroxyurea block which arrests cells in G1/S (Jong
et al., 1995) followed by reentry into the cell cycle after
readdition of serum and removal of drug. Throughout these
experiments, FACS PI staining was used to monitor the stages
of cell cycle (data not shown). As shown in Fig. 3B, the levels
of mHuA decreased from time 0 (G1/S, lane 1) to 3 hours post
release (lane 2), increased progressively as proliferation ensued
reaching a peak at 8 hours (arrow, lane 4) and then
progressively decreased. The 8 hour time point shown in lane
3150 U. Atasoy and others
and exponentially growing cells (Fig. 3B, lane 9) showed levels
of mHuA that correlated with the proliferative state, while the
higher levels were always observed in actively growing cells.
The cells in lane 8 were untreated (negative control) and
completely confluent. Thus, the levels of HuA in these
untreated cells were very similar to those observed with starved
cells (lane 7). We conclude that levels of the ubiquitously
expressed ELAV protein, mHuA, are higher during cell
proliferation reaching a peak around G2/M and they appear to
approximate the expression of cyclin B1.
The higher level of mHuA found following hydroxyurea
blockage (Fig. 3B, lane 1) is consistent with the fact that
blockage occurs at the G1/S boundary. However, the
significance of the reproducible dip in HuA expression
following serum re-stimulation is presently unknown (Fig. 3B,
lane 2). Given these observations which indicate that levels of
mHuA vary with the state of proliferation, we decided to
examine the response of mHuA during activation of immune
cell proliferation.
Fig. 3. Western blot analysis showing expression of mouse HuA in
NIH-3T3 (A and B) or mouse splenocytes (C) following treatments
to activate proliferation. (A) Mouse 3T3 cells following serum
starvation (0.25%) and reactivation by serum addition (10%). Lanes:
1, exponentially growing, unstarved cells; 2, serum starved for 48
hours (G0); lanes 3-8: post serum addition; 3, 10 minutes; 4, 6 hours;
5, 18 hours; 6, 20 hours; 7, 22 hours; 8, 24 hours; 9, exponentially
growing 3T3 from another experiment; 10, confluent cells (G0). B:
mouse 3T3 cells following growth blockage with hydroxyurea and
subsequent release. Lanes: 1, t=0, no serum added; lanes 2-6: post
serum addition; 2, 3 hours; 3, 6 hours; 4, 8 hours; 5, 10 hours; 6, 12
hours; 7, starved cells with no hydroxyurea treatment; 8, untreated
cells; 9, positive control, exponentially growing 3T3 cells. G2/M is
approximately marked by arrows. C: mouse spleen cells treated with
anti-CD3 and/or anti-CD28 to activate T cells. Lanes: 1, cells at day
0; 2, cells unstimulated at day 1; 3, cells unstimulated at day 2; lanes
4-9, stimulated with either or both antibodies; 4, anti-CD3 (1:80) at
day 1; 5, anti-CD28 (1:800) at day 1; 6 anti-CD3 (1:6400)+antiCD28 (1:800) at day 1; 7 anti-CD3 (1:80) at day 2; 8 anti-CD28
(1:800) at day 2; 9 anti-CD3 (1:6400)+anti-CD28(1:800) at day 2;
10, control 3T3 cells in log phase of growth. In each blot, equivalent
amounts of protein were loaded in each lane and tubulin was probed
on the same blots as control. Results shown are representative of
three separate experiments.
4 represents G2/M as assayed by FACS PI staining and by
monitoring maximal cyclin B1 expression (Fig. 3B). Cyclin B1
is known to peak at G2/M in the mammalian cell cycle (Maity
et al., 1995). Interestingly, mHuA expression paralleled cyclin
B1 expression with the peak occurring around G2/M and the
lowest levels in G0/G1. In the hydroxyurea block-release
experiment shown in Fig. 3B, the levels of mHuA as measured
in arbitrary densitometric units varied up to 5-fold, whereas the
cyclin B1 levels varied by about 5-fold. Consistent with results
shown in Fig. 3A, serum starved 3T3 cells (Fig. 3B, lane 7)
mHuA expression is upregulated during T cell
activation in mouse splenocytes
One of the established examples in which post transcriptional
regulation at the level of mRNA stability is known to affect
cytokine protein expression is that of T cell activation via CD3
and CD28 signaling (June et al., 1987; Lindsten et al., 1989;
Thompson et al., 1989). Since the ELAV family members,
including mHuA, have been shown to bind to the 3′UTRs of
cytokine mRNAs and mHuA is the first known ELAV family
member to be expressed in cells of lymphoid lineage (Figs 1
and 2; spleen, thymus, MOLT-4, a T cell line, Burkitt’s
lymphoma and Namalwa, a type of Burkitt’s lymphoma) we
investigated whether mHuA levels were also altered during
activation of T cells. We examined mouse spleen cells for
levels of mHuA following standard methods of activation (see
Materials and Methods) with either mitogenic levels of antiCD3 or with sub-mitogenic, synergistic levels of anti-CD3
plus anti-CD28 (Lindsten et al., 1989). As shown in Fig. 3C,
lane 1, mHuA was barely detectable in untreated cells, which
is consistent with the quiescent (G0) state. Interestingly,
unstimulated human peripheral blood lymphocytes have
nearly undetectable levels of HuA mRNA (Fig. 2B, lane 8).
Following stimulation with mitogenic levels of anti-CD3, an
increase in mHuA was evident by as early as 24 hours (lane
4), but no change occurred upon addition of anti-CD28 alone
(lane 5) or in untreated controls (lane 2). An increase in mHuA
expression was also observed at day 1 when co-stimulating
with sub-mitogenic doses of anti-CD3 plus anti-CD28 (lane
6). By day 2 in the presence of anti-CD3 (lane 7) the level of
mHuA was higher as compared to unstimulated controls.
Therefore, consistent with data shown above using 3T3 cells
(Fig. 3A), mitogenic stimulation of T cells also upregulated
levels of HuA up to 12-fold as measured in arbitrary
densitometric units.
Moreover, when these lymphoid cells were stimulated with
sub-mitogenic levels of both anti-CD3 and anti-CD28, for two
days, there was approximately a 12-fold increase in the level
of mHuA (lane 9) as compared with either unstimulated or
anti-CD28 controls (lanes 2 and 8, respectively). Furthermore,
by day 2 following costimulation using both antibodies (lane
9) the levels of mHuA reached those consistently observed in
Redistribution and upregulation of HuA (HuR) 3151
all of our exponentially-growing 3T3 cells (lane 10).
Stimulation with only anti-CD3 at sub-mitogenic levels
(1:6,400) showed negligible HuA upregulation (data not
shown). These data demonstrate that stimulation of quiescent
T cells using antibodies to CD3 and CD28 surface receptors in
mitogenic or submitogenic levels resulted in dramatic increases
in the expression levels of the ELAV protein, mHuA. HuA
levels increase following stimulation of either NIH-3T3 cells
or murine splenocytes. The greatest increases are seen when T
cells are stimulated. This may be due to the fact that nearly
100% of unstimulated murine splenocytes are in G0, whereas
the previously employed methods used to block NIH-3T3 cells
never block 100% of the cells.
The sub-mitogenic response used in the experiment of Fig.
3C corresponds with well documented increases in cytokine
mRNA stability following costimulation of T cells (June et al.,
1987; Lindsten et al., 1989). Furthermore, it is consistent with
our findings that the ELAV protein, HuB (Hel-N1) results in
stabilization of target mRNAs following transfection of 3T3L1
cells (Jain et al., 1997; Antic and Keene, 1997; unpublished
data) and that ectopic HuA expression stabilizes VEGF mRNA
(Levy et al., 1998).
The correlation shown above between increased levels of
mHuA, previously documented increases in cytokine mRNA
stabilization and T cell activation may appear to be inconsistent
with the reported localization of HuR to the nucleus
(Vakalopoulou et al., 1991; Myer et al., 1997). However, as
shown below, mHuA can be detected also in the cytoplasm of
NIH-3T3 cells in low amounts and has a predominant
cytoplasmic presence at a particular stage of the cell cycle.
Cellular localization of mHuA
Given the potential regulation of mHuA during cell
proliferation, as shown above, it is believed that the stability
and translation of ERG polyadenylated mRNA most likely
takes place in the cytoplasm (see Richter, 1996, and articles
therein). The putative localization of mHuA to the nucleus was
based previously upon biochemical fractionation and binding
of cell extracts to ERG RNAs in vitro (Vakalopoulou et al.,
1991; Myer et al., 1997). Since antibodies to mHuA have not
been available to assess its expression in mammalian cells, we
used our antiserum against mHuA to examine its intracellular
location. Our affinity purified antiserum reacts only with the
ubiquitous protein HuA and does not recognize any other cell
proteins or other ELAV proteins based upon western blotting
of many cells and tissues. As expected, examination of 3T3
cells by indirect fluorescence using this antiserum revealed
predominantly nuclear staining (Fig. 4A). However, when a
large number of fields were examined, separating daughter
cells or recently separated cells were observed in which
cytoplasmic staining was dramatically increased (Fig. 4C and
D). The rarity of these cells appears to reflect the short time
period during which cells are in this stage of the cell cycle.
Interestingly, these cells had a strong pattern very similar to
that reported previously with medulloblastoma cells
expressing the HuB ELAV proteins Hel-N1 and Hel-N2 (Gao
and Keene, 1996). This suggests that at a distinct period
during the cell cycle, presumably during early G1, mHuA
protein is located in the cytoplasm and later returns to the
nucleus. Double staining with anti-BrdU-FITC to track the S
phase of the cell cycle clearly demonstrated mHuA was
predominantly nuclear during periods of DNA synthesis (Fig.
5A and B). The recently divided cells in Fig. 5A and B
(arrow), which displayed cytoplasmic mHuA staining did not
incorporate BrdU and hence were not in S phase. Fig. 5C and
D, as well as Fig. 5E and F, indicate cells which were involved
in various stages of mitosis as detected by DAPI staining. Fig.
5D (arrow) outlines the condensed chromosomes and displays
diffuse HuA staining (Fig. 5C) which may be a result of
nuclear membrane breakdown during mitosis. Fig. 5E and F
demonstrate the predominantly cytoplasmic staining of HuA
at times when the DAPI staining shows that new daughter
nuclei have formed.
Staining of endogenous HuA suggested that the protein
could redistribute to the cytoplasm in a cell cycle dependent
manner. We further investigated HuA localization after
inhibition of transcription which, originally shown for
hnRNPA1, can cause redistribution of some nuclear proteins
into the cytoplasm (Pinol-Roma and Dreyfuss, 1991). A further
rationale for this experiment comes from Katz et al. (1994)
who described a UV-crosslinked ARE binding protein similar
in size (34 kDa) to mHuA that accumulated in the cytoplasm
of T cells following actinomycin D (ActD) treatment. Upon
addition of ActD (5 µg/ml, 3 hours) to NIH-3T3 cells, mHuA
dramatically increased in the cytoplasm leaving in some cells
the nucleus almost devoid of staining (Fig. 4B). 5,6-dichloroβ-d-ribofuranosyl benzimidazole (DRB), which inhibits
transcription by a different mechanism than Act D also caused
redistribution of HuA to the cytoplasm (Fig. 6A). Following
washout of DRB, HuA returned to its predominantly nuclear
distribution (Fig. 6D). This experiment demonstrated both the
export of HuA into the cytoplasm and its reimport into the
nucleus. The effect of DRB (Fig. 6B) and DRB washout (Fig.
6E) on hnRNPA1 staining is also shown. In comparison to
HuA, the proportion of hnRNPA1 that leaves the nucleus
during treatment with inhibitor was small.
To further confirm these observations the expression levels
of HuA in the different compartments were investigated by
fractionation of Act D treated 3T3 cells and western blotting
(Fig. 7). Subcellular fractionation was performed using the
method outlined by Weinberg and Penman (1968) which
allows separation of the cytoplasm into soluble (C1) and outer
nuclear membrane-associated (C2) components. The integrity
of the fractionation before and after ActD treatment was
confirmed by monitoring of nuclear antigens, U1 70K and Sm
(data not shown). In untreated cells (control), HuA was
predominantly nuclear (N) but some soluble cytoplasmic (C1)
and detergent extractable protein (C2) was also evident. ActD
treatment resulted in a significant increase in the mHuA
cytoplasmic signal and a corresponding decrease in the nuclear
signal. These data demonstrate a dramatic transition of the
ubiquitously-expressed ELAV protein HuA as compared with
the other ELAV family members which appear to be
predominantly cytoplasmic.
Although small amounts of HuA can be detected in the
cytoplasm of cells at all phases of the cell cycle, it is highly
abundant during a specific phase of the cell cycle, and thus has
the opportunity to interact with ERG RNAs in the cytoplasm.
This observation does not preclude the possibility that a basal
level of movement in and out of the cytoplasm is occurring
continually, even though the predominant localization of
mHuA appears nuclear.
3152 U. Atasoy and others
Fig. 4. Indirect immunofluorescence
showing the localization of mHuA in
mouse 3T3 cells. (A) Predominantly
nuclear localization of mHuA in
exponentially growing, nonsynchronized
cells; B, cytoplasmic localization of
mHuA following treatment of
nonsynchronized 3T3 cells for 3 hours
with 5 µg/ml actinomycin D; (C and D)
same as A, but showing fields in which
mHuA is localized in the cytoplasm of
recently separated 3T3. HuA was
visualized using Texas Red.
Fig. 5. Double staining of 3T3 cells with mHuA
antisera and anti-BrdU reveals nuclear localization
of mHuA during S phase of cell cycle and
cytoplasmic distribution after mitosis. (A and B)
Cytoplasmic mHuA distribution in recently
dividing cells (arrow). Cells in B which stain with
anti-BrdU (FITC) are in S phase and co-localize in
the nucleus with mHuA (Texas Red). (C and D) A
cell with condensed chromosomes about to
undergo mitosis (arrow). (E and F) Cells which
have recently completed mitosis. Blue DAPI
staining clearly outlines nuclear material.
Redistribution and upregulation of HuA (HuR) 3153
Fig. 6. Staining of 3T3 cells for
HuA and hnRNPA1 (monoclonal
antibody 4B10) following
treatment with 100 µM DRB for 3
hours demonstrates the extensive
relocalization of HuA (A) and
detectable redistribution of
hnRNPA1 (B, arrows) to the
cytoplasm. Two examples of
representative fields are shown.
Wash-out of DRB and a further 3
hours of incubation in media
resulted in HuA and hnRNPA1
returning to their original nuclear
staining (D and E). Double staining
for HuA and hnRNPA1 is shown in
both cases in C (DRB treatment)
and F (DRB washout).
A
D
B
E
C
F
DISCUSSION
ELAV proteins have all been found to bind to AU-rich
sequences present in the 3′UTRs of mRNAs encoding protooncoproteins and cytokines. While HuA appears to have the
same in vitro RNA binding specificity as HuB, HuC and HuD,
it shows differences in expression and intracellular localization
which may indicate distinct functions. In this paper we have
investigated the level of expression, tissue specificity and
intracellular localization of HuA. We prepared rabbit antiserum
against recombinant HuA that does not recognize the neuronal
ELAV proteins, thus allowing specific detection of endogenous
HuA in all tissues examined. Good (1995) first cloned the
Xenopus counterpart, elrA, and detected its mRNA expression
by northern blot analysis. Previous reports concerning
mammalian HuA protein expression failed to detect
endogenous protein, possibly due to weak immunoreactivity of
the antisera used (Okano and Darnell, 1997; Wakamatsu and
Weston, 1997) but protein expression of the Xenopus protein
has been described (Wu et al., 1997). More recently, Levy et
al. (1998) reported detecting two bands of 34 kDa and 30 kDa
which are close to the predicted size of HuA (36 kDa) by
western blotting of 293T cells using a Hu patient antiserum
affinity purified against recombinant GST-HuR (HuA).
However, these authors did not rule out the presence of other
ELAV proteins in their 293T cells that may have reacted with
their paraneoplastic patient serum.
Intracellular localization of the HuA ELAV protein
Sequence comparisons of the mammalian ELAV proteins shows
that there is strong identity in the RRMs which is consistent
with their RNA binding activities being virtually
indistinguishable (Query et al., 1989; Levine et al., 1993; King
et al., 1994; Gao et al., 1994; Ma et al., 1996; Myer et al., 1997).
While all ELAV proteins can bind the same RNAs in vitro,
apparent differences in the intracellular localization of ELAV
3154 U. Atasoy and others
Fig. 7. Western blot showing redistribution of HuA from the nucleus
to the cytoplasm of mouse 3T3 cells. Immunoblot of 3T3
cytoplasmic extracts (C1); deoxycholate extracted from nuclei (C2)
and nuclear extracts (N). Proteins were separated on 15% SDSacrylamide gels and probed with HuA-specific rabbit antiserum.
proteins has raised the question of how they might find access
to the same mRNA targets in vivo. For example, the HuB
proteins, Hel-N1 and Hel-N2, are predominantly cytoplasmic
with some nuclear staining (Gao and Keene, 1996).
Endogenous HuA, on the other hand, appears to be mostly
nuclear but during certain times of the cell cycle is clearly
cytoplasmic (Fig. 4). Interestingly, Hel-N2 and HuA are the
most similar to one another among all human members of the
ELAV family in that they lack hinge segments and are expressed
in proliferating cells (Gao et al., 1994; Ma et al., 1996; Gao and
Keene, 1996; Antic and Keene, 1997). In this study, we report
that HuA is located predominantly in the cytoplasm following
mitosis and appears at that time to have an intracellular
distribution very similar to the HuB proteins (Gao and Keene,
1996), before it later returns to the nucleus during G1. Nuclearcytoplasmic movement of hnRNPA1 and other RRM proteins
has also been demonstrated previously by blocking
transcription with actinomycin D and DRB (Pinol-Roma and
Dreyfuss, 1991, 1992). However, by direct comparison of HuA
and hnRNPA1, we observed that HuA redistributed more
extensively to the cytoplasm following these same treatments.
It is logical to assume that HuA can interact with ERG mRNAs
in the cytoplasm and as shown previously for HuB (Gao and
Keene, 1996), would be available to affect mRNA stability and
translation (Antic and Keene, 1997; Jain et al., 1997).
During revision of this manuscript, Fan and Steitz (1998)
reported that the human homolog of HuA is a shuttling protein
based upon its ability to move between nuclei in heterokaryons
formed between human and mouse cells. Our data are
consistent with HuA being able to shuttle between the nucleus
and cytoplasm like other RNA binding proteins reported
previously (Katz et al., 1994; Pinol-Roma and Dreyfuss, 1991,
1992; Zinsnzer et al., 1997; Peng et al., 1998). A well
characterized type of nuclear export signal (NES) is a leucinerich sequence first described in HIV-1 Rev (Fischer et al., 1994)
and in PKI, a polypeptide inhibitor of the cAMP-dependent
protein kinase (Wen et al., 1995). Bogerd et al. (1996) have
shown the spacing between leucines (or isoleucines) is
important and can be either 3-3-1 (Rex), 2-2-1 (Rev) or 3-2-1
(PKI) while the identity of these surrounding residues can vary.
The sequence 78-I S T L N G L R L Q-87 in loop 5 of RRM
1 of mHuA is a potential NES of the Rev-type whose location
may allow access to interacting factors. Interestingly, the U170K snRNP protein was found to contain a basic amino acid
nuclear localization signal in the same position of the RRM
structure (Romac et al., 1994). The differences in intracellular
localization among ELAV family members cannot be solely
controlled by this putative NES since this sequence exists in
them all. The surrounding structure, including differences in
potential phosphorylation sites which are unique to HuA, may
control accessibility of the NES to interacting proteins.
Phosphorylation is known to play an important role in the
localization of many proteins (reviewed by Jans, 1995).
Additionally, the intracellular distribution of ELAV members
may result from a balance between their individual nuclear
export and import.
Upregulation of HuA expression and immune cell
activation
HuA levels increase dramatically following activation of
mouse spleen T cells suggesting a possible role in immune
function. T cell activation involving CD28 as the costimulatory molecule with CD3 has been shown to involve
stabilization of cytokine mRNAs (Lindsten et al., 1989). Posttranscriptional cytokine mRNA stabilization mediated via the
CD28 signaling pathway, allows activated T cells to escape
apoptosis and secrete large amounts of cytokines (Noel et al.,
1996). In contrast, cells activated solely via the TCR using only
anti-CD3 secrete small amounts of cytokines before
undergoing apoptosis. However, T cell stimulation does not
result in indiscriminate ARE mRNA stabilization as seen by
the stark increase in cytokine mRNAs in comparison with the
small increase in protooncogene mRNAs following
costimulation (Lindsten et al., 1989). This suggests that
additional regulatory factors discriminate between these
mRNAs. We report HuA upregulation during both mitogenic
anti-CD3, as well as sub-mitogenic anti-CD3 and CD28
stimulation. Further experiments will be needed to understand
the role(s) of HuA and other factors in regulating these two
modes of T cell activation.
Umlauf et al. (1995) confirmed the importance of mRNA
stabilization in T cell activation by detecting a 20-fold
upregulation in IL-2 mRNA levels and up to a 100-fold
increase in secreted IL-2 following T cell co-activation using
both anti-CD3 and anti-CD28 but not only anti-CD3.
Interestingly, nuclear pre-spliced IL-2 mRNA levels increased
by 8-fold and they suggested that a transacting RNA binding
protein capable of shuttling between the nucleus and cytoplasm
may interact with IL-2 mRNA. In fact, three proteins which
can bind to the AU-rich sequences of cytokine 3′UTRs as
detected by UV-crosslinking were described by Bohjanen et al.
(1991, 1992). Two of these, AU-B and AU-C, were upregulated
during T cell activation. The third protein, AU-A, is a 34 kDa
predominantly nuclear protein that is believed to shuttle
between the nucleus and cytoplasm and is constitutively
expressed. However, AU-A was reported not to be upregulated
following T cell activation by the methods they used (Katz et
al., 1994). Cloning of these factors has not been reported, nor
have they been biochemically defined and it remains possible
that AU-A and HuA are related proteins.
It is logical to expect that many proteins are upregulated
during cellular stimulatory processes which result in
Redistribution and upregulation of HuA (HuR) 3155
proliferation or metabolic activation. For example, other
investigators have reported that some hnRNP proteins are
elevated in PHA stimulated T cells (Biamonti et al., 1993).
Since ELAV proteins have been demonstrated to stabilize
ARE-containing mRNAs (Jain et al., 1997; Levy et al., 1998),
the consequences of HuA upregulation for cytokine production
in the immune system is an obvious possibility. Our results
suggest that the ELAV protein, HuA, is an ideal candidate for
a transacting factor which interacts with cytokine mRNAs and
regulates their expression. These results do not rule out a role
for HuA in also regulating proto-oncogene mRNA stabilities.
Additional experiments will be necessary to determine whether
these interactions directly influence cytokine production during
immunoregulation.
Implications for cell proliferation and differentiation
Little is known about post-transcriptional regulation of the
mammalian cell cycle. Protooncogenes, such as c-fos and cmyc are upregulated during the transition from G0 to G1,
although the levels of c-fos are transient and c-myc remains
high throughout the cell cycle. Among the known cyclins, only
cyclins B1 and A are known to be regulated posttranscriptionally. In recent papers, Maity et al. (1995 and 1997)
showed that the mRNA half-lives of both cyclins vary by as
much as 4- to 5-fold during the cell cycle, although they peak
at different times. The cyclin E mRNA half-life, however, did
not appreciably change. Hence, the cell posttranscriptionally
regulates cyclins B1 and A but not E. Although the 3′UTRs of
both cyclin B1 and A contain multiple AU repeats (HanleyHyde et al., 1992; Ravnik and Wolgemuth, 1996), RNA
binding proteins which interact with these mRNAs have not
been elucidated. The overlapping temporal upregulation of
HuA and cyclin B1 is intriguing.
We suggest that ERG products have evolved a default
pathway of mRNA degradation as marked by the ARE in the
3′UTRs and that transacting factors like the ELAV proteins
have the ability to reverse this effect by binding the mRNA
instability sequence. The functions of ERGs in the seemingly
distinct pathways of growth and differentiation may involve
differences in localization by ELAV proteins and bound
mRNA. Results with HuA reported here suggest that like HelN2, its expression correlates with increased proliferation
during cell growth in culture, as well as in isolated spleen cells
which are activated in vitro (Fig. 3).
Although our data are consistent with a role for HuA in the
cytoplasm, they do not preclude a role for HuA in RNA
stability, splicing or mRNA transport, while in the nucleus
(Antic and Keene, 1997, 1998). Data presented here are
compatible with earlier hypotheses implicating mammalian
ELAV proteins in the regulation of gene expression via their
interactions with growth regulatory mRNAs (King et al., 1994;
Gao and Keene, 1996; Antic and Keene, 1998). It is likely that
they participate in both nuclear and cytoplasmic functions by
binding to an mRNA subset and regulating their processing
and/or transport during growth and development. It will be
interesting to define other cellular components involved in
these regulatory pathways via their interactions with ELAV
proteins. These factors can be more readily investigated using
a recently devised cell-free deadenylation/degradation system
in which HuA and HuB were shown to stabilize AREcontaining transcripts (unpublished data).
We thank Scott Tenenbaum for assisting with the graphics, the
Nevins laboratory for assistance with cell cycle techniques, Carlos
Sune for providing cell lysates, Dawn Jones for assistance with
splenocytes and the Cell Culture Facility of the Duke Comprehensive
Cancer Center.
REFERENCES
Antic, D. and Keene, J. D. (1997). Embryonic lethal abnormal visual RNAbinding proteins involved in growth, differentiation, and posttranscriptional
gene expression. Am. J. Hum. Genet. 61, 273-278.
Antic, D. and Keene, J. D. (1998). Messenger ribonucleoprotein complexes
containing human ELAV proteins: interactions with cytoskeleton and
translational apparatus. J. Cell Sci. 111, 183-197.
Biamonti, G., Bassi, M. T., Cartegni, L., Mechta, F., Buvoli, M., Cobianchi,
F. and Riva, S. (1993). Human hnRNP protein A1 gene expression.
Structural and functional characterization of the promoter. J. Mol. Biol. 230,
77-89.
Bogerd, H. P., Fridell, R. A., Benson, R. E., Hua, J. and Cullen, B. R.
(1996). Protein sequence requirements for function of the human T-cell
leukemia virus type 1 Rex nuclear export signal delineated by a novel in
vivo randomization-selection assay. Mol. Cell. Biol. 16, 4207-4214.
Bohjanen, P. R., Petryniak, B., June, C. H., Thompson, C. B. and Lindsten,
T. (1991). An inducible cytoplasmic factor (AU-B) binds selectively to
AUUUA multimers in the 3′ untranslated region of lymphokine mRNA.
Mol. Cell. Biol. 11, 3288-3295.
Bohjanen, P. R., Petryniak, B., June, C. H., Thompson, C. B. and Lindsten,
T. (1992). AU RNA-binding factors differ in their binding specificities and
affinities. J. Biol. Chem. 267, 6302-6309.
DeGregori, J., Leone, G., Ohtani, K., Miron, A. and Nevins, J. R. (1995).
E2F-1 accumulation bypasses a G1 arrest resulting from the inhibition of G1
cyclin-dependent kinase activity. Genes Dev. 9, 2873-2887.
Fan, X. C. and Steitz, J. A. (1998). Overexpression of HuR, a nuclearcytoplasmic shuttling protein, increases the in vivo stability of AREcontaining mRNAs. EMBO J. 17, 3448-3460.
Fischer, U., Meyer, S., Teufel, M., Heckel, C., Luhrmann, R. and
Rautmann, G. (1994). Evidence that HIV-1 Rev directly promotes the
nuclear export of unspliced RNA. EMBO J. 13, 4105-4112.
Gao, F. B., Carson, C., Levine, T. and Keene, J. D. (1994). Selection of a
subset of mRNAs from combinatorial 3′ untranslated region libraries using
neuronal RNA-binding protein Hel-N1. Proc. Nat. Acad. Sci. USA 91,
11207-11211.
Gao, F. B. and Keene, J. D. (1996). Hel-N1/Hel-N2 proteins are bound to
poly(A)+ mRNA in granular RNP structures and are implicated in neuronal
differentiation. J. Cell Sci. 109, 579-589.
Good, P. J. (1995). A conserved family of elav-like genes in vertebrates. Proc.
Nat. Acad. Sci. USA 92, 4557-4561.
Hanley-Hyde, J., Mushinski, J. F., Sadofsky, M., Huppi, K., Krall, M.,
Kozak, C. A. and Mock, B. (1992). Expression of murine cyclin B1 mRNAs
and genetic mapping of related genomic sequences. Genomics 13, 1018-1030.
Jain, R. G., Andrews, L. G., McGowan, K. M., Pekala, P. H. and Keene,
J. D. (1997). Ectopic expression of Hel-N1, an RNA-binding protein,
increases glucose transporter (GLUT-1) expression in 3T3-L1 adipocytes.
Mol. Cell. Biol. 17, 954-962.
Jans, D. A. (1995). The regulation of protein transport to the nucleus by
phosphorylation. Biochem. J. 311, 705-716.
Jong, A. Y., Wang, B. and Zhang, S. Q. (1995). Pulsed field gel
electrophoresis labeling method to study the pattern of Saccharomyces
cerevisiae chromosomal DNA synthesis during the G1/S phase of the cell
cycle. Anal. Biochem. 227, 32-39.
June, C. H., Ledbetter, J. A., Gillespie, M., Lindsten, T. and Thompson,
C. B. (1987). T-cell proliferation involving the CD28 pathway is associated
with cyclosporine-resistant interleukin 2 gene expression. Mol. Cell. Biol.
7, 4472-4481.
Katz, D. A., Theodorakis, N. G., Cleveland, D. W., Lindsten, T. and
Thompson, C. B. (1994). AU-A, an RNA-binding activity distinct from
hnRNP A1, is selective for AUUUA repeats and shuttles between the
nucleus and the cytoplasm. Nucl. Acids Res. 22, 238-246.
King, P. H., Levine, T. D., Fremeau, R. T., Jr. and Keene, J. D. (1994).
Mammalian homologs of Drosophila ELAV localized to a neuronal subset
can bind in vitro to the 3′UTR of mRNA encoding the Id transcriptional
repressor. J. Neurosci. 14, 1943-1952.
3156 U. Atasoy and others
Lee, D., Staats, H. F., Sundy, J. D., Patel, D., Sempowski, G. D., Scearce,
R. M., Jones, D. M. and Haynes, B. F. (1998). Immunologic
characterization of CD7-deficient mice. J. Immunol., 160, 5749-5756.
Levine, T. D., Gao, F., King, P. H., Andrews, L. G. and Keene, J. D. (1993).
Hel-N1: an autoimmune RNA-binding protein with specificity for 3′
uridylate-rich untranslated regions of growth factor mRNAs. Mol. Cell. Biol.
13, 3494-3504.
Levy, N. S., Chung, S., Furneaux, H. and Levy, A. P. (1998). Hypoxic
stabilization of vascular endothelial growth factor mRNA by the RNAbinding protein HuR. J. Biol. Chem. 273, 6417-6423.
Lindsten, T., June, C. H., Ledbetter, J. A., Stella, G. and Thompson, C. B.
(1989). Regulation of lymphokine messenger RNA stability by a surfacemediated T cell activation pathway. Science 244, 339-343.
Ma, W. J., Cheng, S., Campbell, C., Wright, A. and Furneaux, H. (1996).
Cloning and characterization of HuR, a ubiquitously expressed Elav-like
protein. J. Biol. Chem. 271, 8144-8151.
Maity, A., McKenna, W. G. and Muschel, R. J. (1995). Evidence for posttranscriptional regulation of cyclin B1 mRNA in the cell cycle and following
irradiation in HeLa cells. EMBO J. 14, 603-609.
Maity, A., McKenna, W. G. and Muschel, R. J. (1997). Cyclin A message
stability varies with cell cycle. Cell Growth Differ. 8, 311-318.
Myer, V. E., Fan, X. C. and Steitz, J. A. (1997). Identification of HuR as a
protein implicated in AUUUA-mediated mRNA decay. EMBO J. 16, 21302139.
Noel, P. J., Boise, L. H., Green, J. M. and Thompson, C. B. (1996). CD28
costimulation prevents cell death during primary T cell activation. J.
Immunol. 157, 636-642.
Okano, H. J. and Darnell, R. B. (1997). A hierarchy of Hu RNA binding
proteins in developing and adult neurons. J. Neurosci. 17, 3024-3037.
Peng, S.-Y. S., Chen, C.-Y. A., Xu, N. and Shyu, A.-B. (1998). RNA
stabilization by the AU-rich element binding protein, HuR, an ELAV
protein. EMBO J. 17, 3461-3470.
Pinol-Roma, S. and Dreyfuss, G. (1991). Transcription-dependent and
transcription-independent nuclear transport of hnRNP proteins. Science 253,
312-314.
Pinol-Roma, S. and Dreyfuss, G. (1992). Shuttling of pre-mRNA binding
proteins between nucleus and cytoplasm. Nature 355, 730-732.
Query, C., Bentley, R. C. and Keene, J. D. (1989). A common RNA
recognition motif identified within a defined U1 RNA binding domain of
the 70K U1 snRNP protein. Cell 57, 89-101.
Ravnik, S. E. and Wolgemuth, D. J. (1996). The developmentally restricted
pattern of expression in the male germ line of a murine cyclin A, cyclin A2,
suggests roles in both mitotic and meiotic cell cycles. Dev. Biol. 173, 6978.
Richter, J. D. (1996). Dynamics of poly(A) addition and removal during
development. In Translational Control (ed. J. W. B. Hershey, M. Mathews
and N. Sonenberg), pp. 481-503. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY.
Robinow, S., Campos, A. R., Yao, K. M. and White, K. (1988). The elav
gene product of Drosophila, required in neurons, has three RNP consensus
motifs. Science 242, 1570-1572.
Robinow, S. and White, K. (1991). Characterization and spatial distribution
of the ELAV protein during Drosophila melanogaster development. J.
Neurobiol. 22, 443-461.
Romac, J. M., Graff, D. H. and Keene, J. D. (1994). The U1 small nuclear
ribonucleoprotein (snRNP) 70K protein is transported independently of U1
snRNP particles via a nuclear localization signal in the RNA-binding
domain. Mol. Cell. Biol. 14, 4662-4670.
Szabo, A., Dalmau, J., Manley, G., Rosenfeld, M., Wong, E., Henson, J.,
Posner, J. B. and Furneaux, H. M. (1991). HuD, a paraneoplastic
encephalomyelitis antigen, contains RNA-binding domains and is
homologous to elav and sex-lethal. Cell 67, 629-639.
Thompson, C. B., Lindsten, T., Ledbetter, J. A., Kunkel, S. L., Young, H.
A., Emerson, S. G., Leiden, J. M. and June, C. H. (1989). CD28 activation
pathway regulates the production of multiple T-cell-derived
lymphokines/cytokines. Proc. Nat. Acad. Sci. USA 86, 1333-1337.
Umlauf, S. W., Beverly, B., Lantz, O. and Schwartz, R. H. (1995).
Regulation of interleukin 2 gene expression by CD28 costimulation in
mouse T-cell clones: both nuclear and cytoplasmic RNAs are regulated with
complex kinetics. Mol. Cell. Biol. 15, 3197-3205.
Vakalopoulou, E., Schaack, J. and Shenk, T. (1991). A 32-kilodalton protein
binds to AU-rich domains in the 3′ untranslated regions of rapidly degraded
mRNAs. Mol. Cell. Biol. 11, 3355-3364.
Wakamatsu, Y. and Weston, J. A. (1997). Sequential expression and role of
Hu RNA-binding proteins during neurogenesis. Development 124, 34493460.
Weinberg, R. A. and Penman, S. (1968). Small molecular weight
monodisperse nuclear RNA. J. Mol. Biol. 38, 289-304.
Wen, W., Meinkoth, J. L., Tsien, R. Y. and Taylor, S. S. (1995).
Identification of a signal for rapid export of proteins from the nucleus. Cell
82, 463-473.
Wu, L., Good, P. J. and Richter, J. D. (1997). The 36-Kilodalton EmbryonicType Cytoplasmic Polyadenylation Element-Binding Protein in Xenopus
laevis is ElrA, a Member of the ELAV Family of RNA-Binding Proteins.
Mol. Cell. Biol. 17, 6402-6409.
Zinszner, H., Sok, J., Immanuel, D., Yin, Y. and Ron, D. (1997). TLS (FUS)
binds RNA in vivo and engages in nucleo-cytoplasmic shuttling. J. Cell Sci.
110, 1741-1750.