(Received for publication, October 23, 1996, and in revised form, January 6, 1997)
From the Program in Molecular Pharmacology and
Therapeutics, Memorial Sloan Kettering Cancer Center, New York, New
York 10021 and the § Department of Biochemistry, University
of New Mexico School of Medicine and Cancer Center, University of New
Mexico Health Sciences Center, Albuquerque, New Mexico 87131
Previous studies have identified three brain
proteins (40, 65 and 95 kDa, respectively) that specifically bind to
the 3-untranslated region of GAP-43 mRNA. In this study, using a
specific monoclonal antibody, we now show that the 40-kDa proteins are
members of the Elav-like protein family. This family of specific
RNA-binding proteins comprise three neural specific members called HuD,
HuC, and Hel-N1. We have shown that purified recombinant HuD can bind with high affinity to GAP-43 mRNA. In addition, we have mapped the
binding site to a highly conserved 26-nucleotide sequence within the
regulatory element. The binding of HuD to this site is readily
displaced by RNA oligonucleotides encoding other HuD binding sites. We
also show that only the first and second RNA binding domains of HuD are
required for selective binding to GAP-43 mRNA.
The Elav-like genes encode a family of RNA-binding proteins that were first described in Drosophila (1-3). Elav, the founder member of this family, is expressed immediately after neuroblastic differentiation into neurons (4). The Elav gene product is necessary for neuronal differentiation and maintenance. In mutant flies, neuroblasts fail to differentiate and continue to proliferate (1). Interest in these proteins was stimulated by the discovery that their human homologues are tumor antigens (5). Subsequent cloning studies have shown that there are four members of the human Elav-like family, namely, HuD, HuC, Hel-N1, and HuR (5-8). HuD, HuC, and Hel-N1 are expressed in postmitotic neurons and in neuroendocrine tumors (6, 9-12). Recent studies indicate that they are also required for neuronal differentiation (13). HuR, the fourth and most recently discovered member, is expressed in all cells and is overexpressed in many human tumors (7). Its normal cellular function has yet to be clarified.
All four Elav-like proteins contain three RNA recognition motifs
(RRM)1 of the RNP2/RNP1 type (14). The
first and second of these RRMs are in tandem and are separated from the
third by a segment rich in basic amino acids. A significant insight
into the mechanism of action of these proteins was provided by the
observation that they specifically bind to U-rich elements in the
3-UTR of mRNAs that regulate cell growth and differentiation (7,
12, 15-17).2 These elements were first
characterized by Shaw and Kamen (19), who showed that the U-rich
element in the 3
-UTR of granulocyte macrophage-colony-stimulating
factor mRNA regulates expression at the post-transcriptional level.
Recent data have shown that transfection of the Elav-like genes into
cells results in the increase in expression of mRNAs that contain
such U-rich elements (20). Thus, through binding to a common element in
many mRNAs, the Elav-like proteins are important components of
coordinate gene expression mechanisms.
One important question is the relationship between the HuR, which is ubiquitously expressed, and the neural specific members HuD, HuC, and Hel-N1. It is reasonable to postulate that HuD, HuC, and Hel-N1 interact with brain-specific mRNAs involved in pathways critical to neuronal differentiation. In our previous studies we have investigated the interaction between the Elav-like proteins and mRNAs that are expressed in all cells (16-17, 21).2 We have now investigated whether there are neuronal specific mRNAs that may be specifically regulated by the Elav-like proteins. Recent studies have indicated that post-transcriptional mechanisms significantly contribute to the regulation of GAP-43 gene expression (22-24). GAP-43 is a neuron-specific phosphoprotein that is required for the regeneration and remodeling of neuronal connections (25-34). Precise control of GAP-43 expression is thus of critical importance during nervous system development.
Two independent cis-acting elements that regulate expression have been
mapped to the 3-UTR of GAP-43 mRNA. One maps to the region
proximal to the termination codon (35). Insertion of this element into
a reporter mRNA results in destabilization of the reporter message
(36). This pathway is regulated by nerve growth factor (36). The other
element is U-rich in sequence and found within a highly conserved
region of the 3
-UTR (37). Insertion of this element also confers
instability to reporter constructs, which is reversed by treatment with
TPA (38). Thus GAP-43 expression is regulated by at least two different
pathways utilizing at least two different cis-acting elements.
Cross-linking studies have identified three brain proteins of 40, 65, and 95 kDa that specifically interact with the 3
UTR of GAP-43 mRNA
(37). It was noted that the 40-kDa protein was similar in size to the neuronal specific Elav-like proteins. Thus, in the current study, we
have investigated whether this 40-kDa protein corresponds to a neuronal
specific member of the Elav-like protein family and have investigated
its interaction with GAP-43 mRNA in a purified system.
Plasmid DNAs were
digested with the appropriate restriction enzymes and transcribed with
T3 RNA polymerase for GAP-43B, and GAP-43C or T7 RNA polymerase for
GAP-43A in the presence of [32P]uridine triphosphate
(Amersham Corp.) as described previously (39). pGAP43A was linearized
with HindIII, yielding a transcript of 718 nucleotides
containing coding region and part of 3-untranslated region. pGAP43B
and pGAP43C were linearized with EcoRI, yielding 221 nucleotides and 114 nucleotides of the 3
-untranslated region, respectively.
S100 extracts were prepared from freshly dissected brains from postnatal day 4 rats as described by Dignam et al. (40). Extracts containing 50 µg of protein were incubated with 0.5 ng of 32P-labeled RNA (1.5 × 105 cpm) and 10 units of RNasin (Promega) in a buffer containing 10 mM HEPES (pH 7.6), 3 mM MgCl2, 40 mM KCl, 5% (v/v) glycerol, and 1 mM dithiothreitol for 10 min at 4 °C (41). Following digestion with RNase T1 (1 unit/µl, Calbiochem), RNA-protein complexes were exposed to UV irradiation for 30 min at 4 °C using a germicidal lamp (Sylvania G30T8), and samples were then subjected to an additional digestion with RNase A (1 mg/ml) for 15 min at 37 °C. Immunoprecipitation assays were performed using 16A11 mAb (10) or an anti-tubulin mAb as a negative control. We used the protocol developed by De Graan et al. (42) with the following modifications. UV-cross-linked complexes were incubated in 200 µl of Nonidet P-40 buffer (10 mM Tris-HCl (pH 7.5), 1% Nonidet P-40, 1% bovine serum albumin, 150 mM NaCl, and 2 mM EDTA) containing the indicated mAb (at 1:100 dilution) for 1 h at 25 °C. Next, 20 µl of a 50% slurry of protein G-Sepharose (Sigma) in Nonidet P-40 buffer was added. Samples were further incubated for 4 h at 4 °C; beads containing the immunoprecipitates were then separated by centrifugation and proteins analyzed by 10% SDS-polyacrylamide gel electrophoresis. Gels were dried and the radioactivity associated with each band analyzed using a PhosphorImager (Molecular Dynamics).
Purification of GST-HuD ProteinsAn overnight culture of
Escherichia coli BL 21, transformed with each HuD construct
(17), was diluted in 1:50 LB media. At an A600
of 0.4, the culture was induced with
isopropyl--D-thiogalactopyranoside (0.1 mM).
After 4 h of further growth, cells were spun down and resuspended
in 10 ml of buffer A (50 mM Tris (pH 8.0), 200 mM NaCl, 1 mM EDTA). The cells were lysed by
adding lysozyme (0.2 mg/ml) and Triton (1%). The lysate was
centrifuged at 12,000 × g for 30 min. The resultant
supernatant was loaded onto a glutathione-agarose affinity column (13 mg of protein/ml of resin). After washing the column with buffer B (50 mM Tris (pH 8.0), 200 mM NaCl, 1 mM
EDTA, 1% Triton), GST-HuD was eluted with (50 mM Tris (pH
8.0), 5 mM glutathione). Active protein was determined by
RNA-complex formation, pooled, and stored at
70 °C.
Reaction mixtures
(0.02 ml) contained 50 mM Tris (pH 7.0), 150 mM
NaCl, 0.25 mg/ml tRNA, 0.25 mg/ml bovine serum albumin, 10 fmol of
labeled RNA and protein as indicated. Mixtures were incubated at
37 °C for 10 min. Following incubation, 4 µl of a dye mixture
(50% glycerol, 0.1% bromphenol blue, 0.1% xylene cyanol) was added,
and 20% of the reaction mixture was immediately loaded on a 1%
agarose gel in TAE buffer (40 mM Tris acetate, 1 mM EDTA). The gel was then electrophoresed at 40 V for
2.5 h. The gel was dried on DE81 (Whatman) with a backing of gel
drying paper (Hudson City Paper, West Caldwell, NJ) and exposed to XAR5
film (Eastman Kodak Co.) for 4-5 h at 70 °C.
Reaction mixtures (0.02 ml)
contained 50 mM Tris (pH 7.0), 150 mM NaCl,
0.25 mg/ml bovine serum albumin, 0.25 mg/ml tRNA, 20 fmol of
radiolabeled mRNA, purified HuD as indicated. After 10 min of
incubation at 37 °C, 0.5 unit of RNase T1 was added to each reaction
and incubated at 37 °C for 10 min further. The mixtures were diluted
1:6 with buffer F (20 mM Tris (pH 7.0), 150 mM
NaCl, 0.05 mg/ml tRNA) and filtered through nitrocellulose (BA 85, Schleicher & Schüll). After washing the nitrocellulose twice with
buffer F, bound HuD-RNA complex was extracted with phenol/chloroform
and concentrated by ethanol precipitation. The resultant RNA was
dissolved in formamide buffer and denatured at 65 °C for 2 min.
Samples were analyzed by 12% polyacrylamide/urea gel electrophoresis.
The gel was fixed with 1:1:8 acetic acid:methanol:water, dried, and
exposed to the XAR5 film at 70 °C overnight.
Reaction mixtures (0.02 ml) contained 50 mM Tris (pH 7.0), 150 mM NaCl, 0.25 mg/ml bovine serum albumin, 0.25 mg/ml tRNA, 10 fmol of radiolabeled mRNA, and purified HuD as indicated. After 10 min of incubation at 37 °C, the mixtures were diluted 1:6 with buffer F and filtered using nitrocellulose. After washing the filter twice with buffer F, bound radioactivity was determined by Cerenkov counting.
Fig.
1 shows the structure of the GAP-43 mRNA and the
transcripts (encoded by plasmids GAP-43H, GAP-43A, GAP-43C, and
GAP-43B) used in these studies. Kohn et al. (37) have
previously described 40, 65, and 90-kDa brain proteins that cross-link
to the transcript encoded by plasmid GAP-43H. To investigate whether
the 40-kDa protein was related to the Elav-like proteins, the labeled
cross-linked material was precipitated with a monoclonal antibody
specific for the neuronal members of the Elav-like protein family (10). The anti-Elav-like monoclonal (16A11) specifically precipitated the
40-kDa species (Fig. 2, lanes 1-3). The
40-kDa protein was not detected without added mAb nor precipitated on
addition of an irrelevant monoclonal antibody. Thus we concluded that
the previously characterized 40-kDa RNA binding factor is indeed a member of the neuronal Elav-like protein family. The 65- and 95-kDa proteins were not co-precipitated. This result indicates that there is
little protein-protein interaction between the 40-kDa protein and the
65- and 95-kDa proteins. It is possible, however, that they interact
when bound to RNA. In this experiment the complex was digested with
RNase before immunoprecipitation.
There are three neuronal members of the Elav-like protein family, namely HuD, HuC, and Hel-N1. (5, 6, 8) The present analysis cannot establish whether a particular family member is associated with GAP-43 mRNA. mAb 16A11 recognizes all three. Since, however, the RNA binding properties of all three are virtually identical (12, 17), and GAP-43 is expressed in virtually all neurons, it is likely that all three family members associate with GAP-43 mRNA.
Characterization of HuD-GAP-43 mRNA Complex FormationAlthough it is clear that the Elav-like proteins are
bound to GAP-43 mRNA, it is possible that other factors in the
extract are required to stabilize binding. Thus we have investigated
whether purified Elav-like proteins can bind to GAP-43 mRNA. For
these studies, we have chosen HuD, since its RNA binding properties have been documented previously (17). Purified recombinant HuD was
incubated with labeled transcripts and complex formation was assayed by
gel retardation analysis. HuD binds with high affinity to the
transcripts encoded by GAP-43B and GAP-43H (Gap-43H data are not
shown), but not to transcripts encoded by GAP-43A or GAP-43C (Fig.
3). GAP-43A contains the entire coding region and the
cis-acting element defined by Nishizawa et al. (35) GAP-43B
contains the cis-acting element defined by Tsai et al. (38).
Thus, purified HuD is sufficient to reconstitute binding and the
specificity of that binding correlates well with the known properties
of the 40-kDa factor (38).
The interaction between HuD and GAP-43B RNA was further investigated
using a quantitative RNA binding assay. We employed the same method as
originally used for the R17 coat protein (43). A low concentration of
labeled RNA (100 pM) was incubated with increasing
concentrations of HuD protein as indicated. The reactions were filtered
through nitrocellulose and the bound radioactivity determined. Fig.
4A shows that the formation of the GAP-43B
RNA-HuD complex is detectable at 7.2 nM, has a midpoint at
about 130 nM, and reaches a plateau above 720 nM with about 60% of the input RNA bound. Complex
formation with GAP-43A or GAP-43C RNA was not detectable under these
conditions. A plot of the log of complex/free RNA versus the
log of HuD concentration reveals a straight line with an intersect on
the x axis at 129 nM (Fig. 4B). Thus
the binding of HuD to GAP-43 mRNA is a simple molecular reaction
with an apparent Kd of 129 nM.
HuD Binds to a Conserved U-rich Segment of GAP-43 mRNA
We
have further mapped the binding site using an RNase T1 digestion
technique. In this technique the HuD-RNA complex is digested with RNase
T1, and the protein bound RNA fragments are isolated by absorption to
nitrocellulose (17). Fig. 5A shows the
HuD-dependent retention of a single fragment on incubation
with GAP-43B RNA. No fragments were selected from GAP-43A or GAP-43C
RNAs. This fragment is identical in size to a unique 26-nucleotide T1
fragment present in the total digest of GAP-43B RNA. This suggested
that this fragment is the HuD binding site. It was also conceivable, however, that it arose from the protection of a G residue between two
smaller T1 oligonucleotides, yielding a fragment of similar size. The
26-nucleotide fragment was isolated and redigested with RNase T1 (Fig.
5B). No digestion products were observed, therefore this
fragment indeed corresponds to the unique 26-nucleotide sequence that
is shown in Fig. 1. It is interesting to note that this binding site is
highly conserved between GAP-43 mRNAs from different species (37).
We have compared the sequence of the GAP-43 mRNA binding site with
those we have found found in c-Fos, c-Myc, Tau, interleukin-3, and
AdIVA2 mRNAs (Fig. 6). The GAP-43 binding site is
similar to the others in that it contains U-rich tracts. We have shown that these U-rich tracts are essential for the binding of HuD to the
element in the 3-UTR of c-Fos mRNA (17). Next, we tested whether
the HuD-GAP-43 complex would be displaced by the 27-nucleotide c-Fos
binding site. Complex formation was indeed readily displaced by the
27-nucleotide c-Fos oligonucleotide but not by a mutant oligonucleotide
(27-8) (17) which contains substitutions of the U-rich tracts (Fig.
7). Thus the GAP-43 binding site is functionally similar
to those we have described in other mRNAs.
The First and Second RNA Binding Domains Are Essential for Binding to GAP-43 mRNA
HuD contains three RNA binding domains (5).
The first and second are in tandem and are separated from the third by
a basic segment, which we call the basic domain. Previous studies have shown that the first and second RNA binding domains are both essential for binding to the AU-rich elements, whereas the third RNA binding domain binds to the poly(A) tail (17).2 To determine the
domains of HuD required for binding to GAP-43 mRNA, we utilized the
deletion constructs shown in Fig. 8. Gel retardation
analysis showed that the first and second RRMs were essential for
binding to the GAP-43 site (Fig. 8). The third RRM did not display any
detectable binding (Fig. 8, lanes 8 and 9). As
noted previously in our studies on the c-Fos element, the first and
second RRMs are required in tandem for complete binding activity (Fig.
8, lanes 4-6). The basic domain alone did not exhibit any binding activity.
The Elav-like RNA-binding proteins regulate gene expression via an
interaction with U-rich elements in the 3-UTR of specific mRNAs.
The mechanism underlying the specificity of this regulation is poorly
understood. There are four members of this family (HuD, HuC, Hel-N1,
and HuR) in human cells. HuR is ubiquitously expressed, whereas HuD,
HuC and Hel-N1 are selectively expressed in neurons (5-8). The
neuronal members are the homologues of Elav, a Drosophila gene required for neuronal differentiation. (2) The existence of three
family members in human neurons suggests that they specifically modulate mRNAs that regulate neuronal differentiation. The RNA binding specificity, however, of all four members is very similar (7,
12, 17). Thus, it is likely that HuD, HuC, and Hel-N1 exert their
function, not only through binding to U-rich elements, but also through
interactions with other proteins that selectively bind to the 3
-UTR of
neuronal mRNAs. The first step in investigating this is to identify
and study neuronal specific mRNAs that may be regulated by the
Elav-like proteins. Previous studies have shown that neuronal factors
of 40, 65, and 95 kDa bind specifically to the 3
-UTR of GAP-43
mRNA. (37, 38) GAP-43 is an important neuronal specific protein
that regulates synaptic remodeling (18, 30, 44). In this study we have
shown that the 40-kDa factor is a neuronal specific member of the
Elav-like protein family. We have shown that purified HuD binds to
GAP-43 mRNA with appropriate specificity and affinity. Thus it is
likely that the Elav-like proteins play a significant role in the
post-transcriptional control of GAP-43 expression. If this was so, then
one would expect that alteration of Elav-like gene expression would
affect GAP-43 expression. In addition one might suspect that deletion
of the Elav-like binding site from GAP-43 mRNA would significantly
affect its expression. These are important future experiments to
confirm the role of the Elav-like proteins in the regulation of GAP-43
expression.
The regulation of GAP-43 expression in PC12 cells illuminates the
specificity problem that we discussed above. Addition of TPA to PC12
cells results in the marked stabilization of GAP-43 mRNA (38). This
stabilization is effected through interactions with the segment of
GAP-43 mRNA encoded by plasmid GAP-43B. However, TPA treatment also
results in the increased degradation of many other mRNAs
(e.g. c-Fos) that contain U-rich elements and are selectively bound by the Elav-like proteins. The simplest hypothesis is
that the regulation of a particular mRNA may be governed by the
association of the Elav-like proteins with other RNA-binding proteins
that recognize disparate specificity elements in the mRNA. In this
paper we have defined the 26-nucleotide Elav-like binding site. This
site is well conserved between GAP-43 mRNAs from different species.
There are, however, additional elements within the sequence encoded by
GAP-43B (Fig. 1) that are also well conserved. It is reasonable to
assume that these elements and their associated binding factors also
play a role in the expression of GAP-43 mRNA. The 95-kDa protein
described by Tsai et al. (38) is a good candidate for such a
trans-acting factor. Thus, the TPA-induced stabilization of GAP-43
mRNA may depend on the binding of the 95-kDa protein, the binding
of the Elav-like proteins to the 26-nucleotide element, and appropriate
interactions between these two proteins. Further investigation of this
possibility will require the purification of the 95-kDa factor. It is
also important to point out that GAP-43 mRNA contains another
element at the 3 end of the open reading frame. This element is bound by the 65-kDa factor (38). It remains to be seen whether there is any
interaction between the 65-kDa factor and the Elav-like proteins. In
conclusion, the GAP-43-Elav-like protein complex described here will
provide an important reagent to answer these questions and may
facilitate the development of an in vitro system that
faithfully recapitulates regulation of GAP-43 mRNA turnover.
We thank Barry Nevins for his patience in preparing the manuscript.