ABSTRACT
Site-specific cleavage of human insulin-like growth factor II mRNAs requires two
cis
-acting elements, I and II, that are both located in the 3
'
untranslated region and separated by almost 2 kb. These elements can interact and form a stable RNA-RNA stem structure. In this study we have initiated the investigation of
trans
-acting factors involved in the cleavage of IGF-II mRNAs. The products of the cleavage reaction accumulate in the cytoplasm, suggesting that cleavage occurs in this cellular
compartment. By electrophoretic mobility shift assays, we have identified a
cytoplasmic protein with an apparent molecular weight of 48-50 kDa, IGF-II cleavage unit binding protein (ICU-BP), that binds to the stem structure formed by interaction
of parts of the
cis
-acting elements I and II. The binding is resistant to high K
+
concentrations and is dependent on Mg
2+
. In addition, ICU-BP binding is dependent on the cell density and correlates inversely with
the IGF-II mRNA levels.
In vivo
cross-linking data show that this protein is associated with IGF-II mRNAs
in vivo
.
Human insulin-like growth factor II (IGF-II) is encoded by a single gene that contains four promoters P1-P4 which give rise to a family of mRNAs. These promoters are
differentially active in a tissue- and development-specific manner and render mRNAs of 5.3 (P1), 5.0 (P2), 6.0 (P3) and
4.8 kb (P4) respectively (
1
). The transcripts share the IGF-II coding region, but have different 5' untranslated regions (UTRs), leading to differential
translatability (
2
,
3
). All IGF-II mRNAs are targets for site-specific endonucleolytic cleavage in the 4.2 kb 3' UTR, rendering an unstable capped 5' cleavage product containing the coding region and an
unusually stable polyadenylated 3' cleavage product of 1.8 kb consisting of the 3' terminal region of the 3" UTR (
4
).
The importance of RNA processing as a regulatory level in the control of gene
expression becomes increasingly appreciated as more insight into post-transcriptional events is generated (
5
-
7
). Analysis of nuclear RNA-processing events has revealed many details of both pre-mRNA splicing (
8
) and 3' end formation (
9
). In contrast, relatively little is known about the processing of mature mRNAs
in eukaryotes, not in the least because of the difficulties in establishing
in vitro
degradation assays. It is clear that in these processes,
cis
-acting elements in the mRNA interact with specific
trans
-acting factors. In yeast cells, some of the mRNA degradation pathways are being elucidated, and genetic screens have
allowed the identification of different
trans
-acting factors (reviewed in
6
). Mammals do not allow a genetic analysis and the identification of
trans
-acting factors has to start with biochemical rather than genetic
approaches. Many of the mammalian
trans
-acting factors have been identified based on their ability to bind to
cis
-acting elements in certain mRNAs (
10
-
16
). Due to the structural versatility of RNA, binding of RNA by RNA-binding proteins is often dependent on both sequence- and structure-specific recognition, e.g. iron-responsive element binding protein (IRP) binding to the iron-responsive element (IRE) (
17
), human immuno deficiency virus (HIV) Tat protein to the Tar region (
18
) and HIV Rev to the Rev response element (RRE) (
19
).
Previously, we have identified two
cis
-acting elements in the 3'-UTR that are required for cleavage of IGF-II mRNAs (
20
,
21
). These two elements are separated by approximately 2 kb that can be deleted
without decreasing the cleavage efficiency. Element I is located at positions -2116/-2013 and contains no obvious features in its primary sequence or
secondary structure. Element II at -173/+150 encompasses the cleavage site and contains two regions that are
highly conserved among human, rat and mouse (
20
): two stem-loops upstream of the cleavage site (-139/-3) and a G-rich region at positions -14/+60 (positions are relative to the cleavage
site which is set to +1; Fig.
1
A). Cleavage
in vivo
requires the presence of an intact cleavage unit comprising both elements (Fig.
1
B). Site-specific endonucleolytic cleavage of IGF-II mRNAs is unique since, firstly, the two
cis
-acting elements are so widely separated from each other, and, secondly,
the sizes of these elements are very large. We have shown that a region of
element I (-2108/-2029) and a region in element II downstream of the cleavage site (+18/+101) form a very stable stem structure (Fig.
1
B), that is conserved among human, rat and mouse IGF-II mRNAs. In addition, our data clearly indicate that both the secondary structure and the sequence of this stem structure are
important for cleavage (
21
). This suggests that elements I and II cooperate in the binding of
trans
-acting factors involved in cleavage of IGF-II mRNAs. The identification of proteins binding to the
cis
-acting elements required for cleavage may provide critical insight into
the mechanism underlying the specific endonucleolytic cleavage of IGF-II mRNAs. In the present study we have initiated the identification of
trans
-acting factors that interact with the IGF-II cleavage unit.
Plasmid pBluescript II (KS
+
) was obtained from Stratagene (La Jolla, CA). Enzymes were purchased from Boehringer Mannheim (Germany), with
the exception of Pfu DNA polymerase (Stratagene, La Jolla, CA), RNase-free DNase (Promega, Madison, WI) and RNase T1 (CalBiochem, La Jolla, CA).
Enzymes were used according to the manufacturer's instructions. NTPs and dNTPs
were obtained from Kabi-Pharmacia (Uppsala, Sweden). G418 was purchased from Sigma (St Louis, MO).
A random priming DNA labeling kit was purchased from Boehringer Mannheim
(Germany) and a DNA sequencing kit from Kabi-Pharmacia (Uppsala, Sweden). Guanidinium thiocyanate was obtained from
Fluka (Buchs, Switzerland) and GeneScreen membranes from Du Pont de Nemours
(Dreiech, Germany). [[alpha]-
32
P]dCTP (3000 Ci/mmol) and [[alpha]-
32
P]CTP (760 Ci/mmol) were purchased from Amersham (Buckinghamshire, UK).
The cell lines HeLa and Ltk
-
were grown in Dulbecco's Modified Eagle's Medium (DMEM), the stable Ltk
-
cell line EP7-9 (
22
) was grown in the DMEM in the presence of 300 [mu]g/ml G418 and Hep3B cells were grown in [alpha]-Minimal Essential Medium ([alpha]-MEM). All media were supplemented with 10% fetal
calf serum, 100 IU/ml penicillin, 100 [mu]g/[mu]l streptomycin and 300 [mu]g/ml glutamine.
Cytoplasmic and nuclear cell fractions were prepared employing a method adapted
from Vakalopoulou
et al
. (
12
) with slight modifications. Briefly, the cells were lysed in 10 mM Tris pH 8.0,
10 mM NaCl, 3 mM MgCl
2
, 1 mM DTT, 1 mM phenylmethylsulphonic acid, 0.5% NP-40 and 0.5% sodium deoxycholate on ice for 10 min. Subsequently, the lysate was spun in a microfuge for 5 min at 4oC and the supernatant was collected and used for cytoplasmic RNA
isolation or stored at -80oC in 10% glycerol (cytoplasmic extract). The pellet was used to
isolate nuclear RNA or to prepare nuclear extracts as described in (
23
). Protein concentrations were determined by the Bradford protein assay (BioRad,
München, Germany).
In vivo
cross-linking experiments were performed essentially as in (
24
). Briefly, cells were grown on 100 mm plates, washed twice with cold PBS and
exposed to UV light (254 nm) at 1.9 J/cm
2
(Stratalinker, Stratagene, La Jolla, CA) in 5 ml PBS on ice prior to
preparation of the extracts.
RNA was isolated from the cytoplasmic, nuclear and total cell fractions by the
single-step guanidinium thiocyanate method (
25
). RNA (10 [mu]g) was glyoxalated and size-separated on a 1% agarose 10 mM sodium phosphate gel and transferred
to a GeneScreen membrane. The RNA was fixed on the membrane by irradiation with
long-wavelength UV light for 2.5 min and baking at 80oC for 2 h. Northern blots were hybridized in the presence of 50%
formamide according to GeneScreen protocols in glass cylinders with continuous
rotation at 42oC. DNA fragments were labeled by random priming with [[alpha]-
32
P]dCTP and added after 3 h prehybridization at a final concentration of 10
6
c.p.m./ml. After overnight hybridization, blots were washed to a final
stringency of 0.5 or 0.1* SSC, 1% SDS at 65oC (1* SSC is 0.15 M NaCl, 0.015 M sodium citrate) and exposed on
Fuji RX X-ray film. Two human IGF-II exon 9 probes were used: a 532 bp
Eco
RV-
Ava
I fragment encompassing the region between positions -557 and -26 (5'-specific probe) and a 1.0 kb
Sma
I fragment (positions +81/+1094; 3'-specific probe). A 26 nt-long oligonucleotide was used as a 28S ribosomal probe: 5'-AACGATCAGAGTAGTGGTATTTCACC-3'. Hybridization conditions for this
probe were identical to the IGF-II probe except for the formamide concentration (25%). Blots were washed
for 30 min in 2* SSC, 1% SDS and 30 min in 1* SSC, 1% SDS at 30oC. The autoradiographs were analyzed by densitometric
scannning.
For the preparation of BS-S/S, BS-AS/S and BS-AS/AS (Fig.
6
A), oligonucleotides 5'
Bam
HI -2340/-2334 3' and 5'
Eco
RI +169/+151 3' (numbers indicate positions in exon 9) were used as primers in PCR
reactions on plasmids S/S, AS/S and AS/AS respectively. The latter constructs
are IGF-II expression plasmids containing elements I/II in the sense (S) or
antisense (AS) orientation as indicated and are described in detail in (
21
). The PCR products were subcloned in the
Bam
HI and
Eco
RI sites of pBluescript KS
+
, resulting in the plasmids BS-S/S, BS-AS/S and BS-AS/AS. BS-II was prepared by PCR with primers 5'
Bam
H1 -173/-155 3' and 5'
Eco
RI +169/+151 3' on S/S. BS-AS/ASn was prepared from BS-AS/AS by replacing the
Bgl
II (-110)-
Eco
RI (+170) fragment with a PCR product made with primers 5'
Bgl
II -14/+4 3' and 5'
Eco
RI +169/+151 3' on template BS-AS/AS, resulting in a deletion of nucleotides -104/-15 from BS-AS/AS. All constructs were checked by
restriction enzyme analysis or sequencing if necessary.
Radiolabeled RNA probes and unlabeled competitor RNAs were synthesized using T7 RNA polymerase on linearized DNA templates according to
instructions of the manufacturer in the presence of 1 mM ATP, GTP and UTP each
and 20 [mu]Ci [[alpha]-
32
P]CTP and 0.1 mM CTP (radiolabeled) or 1 mM CTP (unlabeled). Templates used for
synthesis of the RNAs: I: BS-S/S linearized with
Bgl
II (-110), II: BS-II linearized with
Eco
RI, I/II: BS-S/S linearized with
Eco
RI, AS/S: BS-AS/S linearized with
Eco
RI, AS/AS: BS-AS/AS linearized with
Eco
RI, and AS/ASn: BS-AS/ASn linearized with
Eco
RI. After synthesis (1 h) the template was removed by DNase I treatment (1 U in
20 [mu]l reaction mixture for 15 min at 37oC) and the RNA was separated from unincorporated nucleotides by
Sephadex G25 spin dialysis. Subsequently, the samples were phenol/chloroform
extracted, ethanol precipitated, washed in 70% ethanol and renatured in
renaturation buffer (20 mM Tris-HCl pH 7.5, 100 mM KCl, 2 mM MgCl
2
), 5 min at 90oC and 30 min on ice. The integrity of the RNAs was checked by gel
electrophoresis.
For electrophoretic mobility shift assays (EMSAs), radiolabeled RNA probes (10
5
c.p.m.) were incubated in cellular extract with a protein content of 10 [mu]g in a 10 [mu]l reaction mixture containing 20 mM Hepes-KOH pH 7.5, 100 mM KCl, 1 mM MgCl
2
, 1 mM DTT and 0.01% NP-40 for 45 min on ice. In competition experiments, the competitor RNAs were
preincubated with the extract for 5 min before addition of the probe. After the
binding reaction the samples were incubated at room temperature for 20 min with
a mixture of 12.5 [mu]g/ml RNase A and 2 U/ml RNase T1. Samples were analyzed on a non-denaturing 6% polyacrylamide gel containing 0.01% NP-40.
The molecular weight of IGF-II cleavage unit binding protein (ICU-BP) was determined as described in (
26
). First the IGF-II cleavage unit RNA-protein complex (ICU-RPC) was UV cross-linked in an EMSA gel (254 nm, 5 J/cm
2
) and a gel slice containing this complex (localized by wet gel autoradiography)
was excised. Subsequently the gel slice was inserted into a 10% SDS-PAGE gel and electrophoresed with
14
C-methylated protein markers (Amersham, Buckinghamshire, UK) loaded in
parallel.
To investigate whether cleavage of IGF-II mRNAs takes place in a particular subcellular compartment, 10 [mu]g cytoplasmic or nuclear RNA from Hep3B cells were analyzed by
Northern blotting (Fig.
2
). Hep3B is a human hepatoma cell line that endogenously expresses both the 6.0
and the 4.8 kb mRNAs derived from promoters P3 and P4, respectively. The full-length mRNAs are present in comparable amounts in both the cytoplasmic and nuclear fractions. Cleavage of these mRNAs renders a polyadenylated 3' cleavage product of 1.8 kb that hybridizes exclusively with the 3'-specific IGF-II exon 9 probe. The capped 5' cleavage product derived from the 6.0 kb
mRNA, with a size of 4.2 kb, is detected by the 5'-specific probe. The amount of 5' cleavage product of the 4.8 kb mRNA is below the detection
level, probably due to the low relative levels of this mRNA species. In
contrast with the full-length IGF-II mRNAs, the 3' and 5' cleavage products are readily detected in the
cytoplasmic RNA fraction, but not in the nuclear fraction (Fig.
2
). The cytoplasmic localization of the cleavage products becomes even more
pronounced if we take into consideration that equal amounts of RNA were loaded
for each fraction, but that 10 [mu]g nuclear RNA is derived from eight times more cells than 10 [mu]g cytoplasmic RNA. Obviously, the cleavage products accumulate in the
cytoplasm, suggesting that cleavage takes place in the cytoplasm, although
nuclear cleavage and rapid transport can not be fully excluded.
To examine further the specific mRNA-protein interaction and to characterize the protein components of ICU-RPC, we cross-linked ICU-RPC in a gel and subsequently inserted the gel slice containing the cross-linked complex into a 10% SDS-PAGE gel (see Materials and Methods for
details). This analysis revealed that the protein binding to the cleavage unit
has an apparent molecular weight of 48-50 kDa (Fig.
4
A). Since the formation of ICU-RPC is abolished by proteinase K pretreatment of the extract, but not by
Micrococcal Nuclease pretreatment (data not shown), ICU-RPC does not require a nucleic acid component from the extract. Obviously,
the 48-50 kDa protein is the main protein component in ICU-RPC. We will refer to this protein as ICU-BP (IGF-II cleavage unit binding protein).
Interactions of proteins with nucleic acids are highly dependent on the salt
concentrations. Therefore, we analyzed the binding capacity of ICU-BP in the presence of increasing K
+
concentrations. ICU-BP does not bind at very low K
+
concentrations (Fig.
4
B), but binds efficiently at physiological K
+
concentrations and still forms a stable RNA-protein complex in 600 mM KCl. Divalent cations play an important role in
many cellular processes. When divalent cations are chelated from the binding
buffer by EDTA, binding of ICU-BP is strongly decreased. This inhibition can be overcome by Mg
2+
, and high concentrations of these ions increase the binding. The physiological
binding conditions used throughout our binding studies (100 mM KCl, 1 mM Mg
2+
) are obviously favourable for ICU-BP binding.
We have shown previously that the
cis
-acting elements I and II are both required for cleavage. Parts of these
elements form a long stem structure that stabilizes two additional stem-loops upstream of the cleavage site in element II, as was shown by RNase
T1 digestion experiments (Fig.
1
B;
21
). In addition, we showed that the stem structure functions in a sequence-specific manner (
21
), suggesting that recognition of the stem structure by
trans
-acting factors is important for cleavage. To investigate whether ICU-BP is involved in recognition of the stem structure, the requirements
for ICU-BP binding were analyzed using different probes in an EMSA (Fig.
5
A). First we analyzed the binding of ICU-BP to probes consisting of elements I or II separately (probes I and II
respectively), and compared this with probe I/II that contains both elements
(Fig.
5
B). ICU-RPC is not formed when either probe I or probe II is used, but only when
both elements are present, indicating that the elements cooperate in binding of
ICU-BP. In competition experiments with unlabeled competitor RNAs, the ICU-BP binding is competed very efficiently by the cleavage unit RNA
(I/II) (Fig.
5
C). Competition is almost complete already at the lowest competitor
concentration (10 ng/binding mixture). The individual elements I and II can
only compete when present in higher excess, probably because at these higher
concentrations (50-100 ng/binding mixture) they interact with the probe. Yeast tRNA does not
compete with I/II for binding, but instead the binding increases at higher
concentrations of tRNA, most likely caused by sequestration of non-specific factors that interfere with the formation of specific RNA-protein complexes.
If binding of ICU-BP is involved in cleavage of IGF-II mRNAs, it may be expected that its binding is regulated in a
manner that correlates with the expression pattern of IGF-II. Cytoplasmic extracts were prepared from Hep3B cells grown at various
densities (20, 50, 80 and 100%) and tested in an EMSA with probe I/II (Fig.
7
A). In extract from exponentially growing cells (20%), ICU-RPC is prominently present. In contrast, in extract from the confluent
cells (100%), the complex formation is strongly decreased. Densitometric
scanning of the autoradiogram shows a gradual decrease of ICU-RPC upon increasing cell density. This results in a 5-fold reduction of ICU-RPC formation when 20 and 100% confluent cells are compared,
and the total amount of protein used for the EMSA is identical (Fig.
7
A). In parallel with the decrease of ICU-RPC formation with increasing cell density, a yet unidentified lower
mobility complex appears (indicated by an asterix). In parallel total RNA was
isolated from dishes with cells grown to the various cell densities and 10 [mu]g of each sample was analyzed on a Northern blot. A strong increase in the levels of IGF-II mRNA in confluent cells is observed (Fig.
7
B). Apparently this applies to both the 6.0 and 4.8 kb mRNAs as well as to the 1.8 kb cleavage product. The amount of the 6.0 kb major IGF-II transcript was quantified by densitometric scanning and normalized with a 28S ribosomal probe. The 6.0 kb mRNA level gradually rises with
increasing cell densities up to 10-fold in confluent cells (Fig.
7
B). These results indicate that the formation of ICU-RPC correlates inversely with the IGF-II mRNA level in Hep3B cells.
Various lines of evidence suggest a role for ICU-BP in cleavage of IGF-II mRNAs: (i) both the protein and the cleavage products are highly
enriched in the cytoplasm; (ii) the protein binds to the
cis
-acting elements involved in cleavage; and (iii) the binding of the protein
correlates inversely with the IGF-II mRNA levels. To demonstrate further that ICU-BP also specifically interacts
in vivo
with IGF-II mRNAs, the following experiment was performed. Two cell lines, untransfected Ltk
-
cells that do not express IGF-II mRNAs and Ltk
-
cells stably transfected with an IGF-II minigene EP7-9 (
22
), were irradiated with UV light. If ICU-BP is associated with IGF-II mRNAs
in vivo
, UV irradiation will cross-link the protein to the mRNAs, thereby preventing it from binding to
radiolabeled I/II RNA after preparation of a cytoplasmic extract. As shown in
Figure
8
, ICU-BP is depleted from the Ltk
-
EP7-9 cell extract after the UV treatment, whereas no difference is observed
in the untransfected Ltk
-
cells, although the protein yields from both lines are comparable. This
indicates that the depletion is not due to non-specific cross-linking or artefactual damage to the complex, but that it is caused
specifically by cross-linking the protein to the IGF-II mRNAs. These results demonstrate that ICU-BP can also form a specific RNA-protein complex with IGF-II mRNAs
in vivo
.
Previously, we have identified two
cis
-acting elements required for cleavage of human IGF-II mRNAs (
20
) and we have demonstrated that part of these elements interact to form a stable
stem structure and that in addition two stem-loops in element II are required for cleavage (
21
). In this report we have studied in which subcellular compartment the cleavage
reaction takes place and initiated investigation of the
trans
-acting factors involved in cleavage. Processing of mature mRNAs
predominantly occurs in the cytoplasm and since it was shown that the products
of the cleavage reaction accumulate in the cytoplasm, it is likely that the IGF-II mRNA cleavage is also a cytoplasmic event. Although we can not formally
exclude the possibility that both cleavage products are very rapidly exported
from the nucleus to the cytoplasm after cleavage has occurred, the results
strongly suggest that the cleavage reaction mainly, if not exclusively, takes place in the cytoplasm. Therefore, the
trans
-acting factors involved in cleavage are most likely localized in the
cytoplasm. We have identified a novel RNA-protein complex (ICU-RPC) that forms on the cleavage unit and is strongly enriched in
the cytoplasm. The major extract component of this complex is a protein with an apparent molecular
weight of 48-50 kDa, which we named ICU-BP. The ICU-BP protein is present in both endogenously IGF-II expressing (Hep3B) and non-expressing cell lines (HeLa, Ltk
-
) and shows identical binding characteristics in these cells (data not shown). Similarly, the cleavage activity is not restricted to IGF-II expressing cells, because non-IGF-II expressing cell lines transfected with IGF-II mini-gene constructs are also able to cleave IGF-II mRNAs. Also, the fact that the RNA-protein complex is formed in both
human (Hep3B, HeLa) and mouse (Ltk
-
) extracts is not surprising, because human mRNAs are cleaved efficiently in
human cells as well as in mouse Ltk
-
cells (
22
).
We used elements I and II of the cleavage unit separately in an EMSA and showed
that binding of ICU-BP requires the presence of both elements. This by itself would not
necessarily call for an interaction between the elements, but could be
explained if the binding site would be partly located in element I and partly
in element II. However, we used an element I probe that still contains the
first 60 nt of element II that are not involved in stem structure formation and
this did not bind. Furthermore, in the authentic IGF-II mRNAs elements I and II are separated by 2 kb. Additional evidence for
the requirement for a cooperation of both elements was obtained by EMSA
analysis with orientation mutants, showing that the AS/S RNA which can not form
the stem structure does not bind ICU-BP, whereas the AS/AS RNA that forms a stem similar to the S/S RNA can
bind ICU-BP. We have shown before that the presence of the stem structure stabilizes the two additional stem-loop structures in element II (
21
). Therefore, the results with the orientation mutants could also be explained
if ICU-BP binds these stem-loops. The AS/AS[Delta] mutant which lacks the stem-loops, binds ICU-BP, albeit with a somewhat reduced efficiency.
This shows that the binding site for ICU-BP is located in the stem structure and not in the two additional stem-loops of element II. Why is ICU-BP able to bind the AS/AS mutant stem structure, while this
RNA is not a substrate for the cleavage reaction
in vivo
? The sequence from this mutant stem is of course very similar to the wild-type stem (
21
) and apparently the information suffices for binding, but not for cleavage. It
is possible that this is merely an orientation effect, since the main
difference between the AS/AS RNA and the S/S RNA is the orientation of the stem
structure relative to the cleavage site. It has been reported before that mRNA
binding capacity and activity of a protein can be separated. Proteinase K
abolishes the c-
myc
mRNA
in vitro
degradation without affecting the binding to the AU-rich element in the c-
myc
mRNA (
27
).
In summary, these data indicate that ICU-BP recognizes the stem structure formed between the regions of elements I
and II located at -2108/-2029 and +18/+101. Previous results showed that this stem
structure functions in a sequence-specific manner in cleavage (as was shown for the HIV Rev protein
recognition;
19
), thereby suggesting that it is a target for a
trans
-acting factor and not merely a structural element. ICU-BP, therefore, is likely to be an important component in cleavage of
IGF-II mRNAs. However to confer endonucleolytic cleavage additional entities
possibly interacting with the two additional stem-loops are required.
In Hep3B cells, the binding of ICU-BP to the stem structure is regulated in a growth-dependent manner. In cells grown to confluence the ICU-BP binding is 5-fold decreased as compared with exponentially growing
cells. These differences in binding correlate inversely with the IGF-II mRNA levels (Fig.
7
). Although the levels of 3' cleavage product seem to increase along with the full length mRNA
levels, this can still be consistent with a role for ICU-BP binding in cleavage of IGF-II mRNAs. Because many processes each with its own kinetics are
occurring at the same time (promoter activities, differential stability of the
mRNAs and the cleavage products) it is difficult to relate the observed steady
state mRNA levels to one process in particular. With respect to the observed
growth-related regulation it is interesting to note that translation of the 6.0
kb mRNA also appears to be dependent on the growth-status of the cells (
28
).
The physiological importance of ICU-BP is further indicated by the
in vivo
UV cross-linking experiments. UV cross-linking of intact cells has proven to be a powerful tool in the
isolation of RNA-binding proteins (
29
-
31
). Only proteins that are very tightly associated with the RNA can be cross-linked. The efficiency of cross-linking is variable, depending on the protein and RNA involved (
30
). Recently, this technique was used in combination with an
in vitro
RNA-binding assay to show that AU-rich sequence binding proteins that bind
in vitro
, are also associated with the mRNA
in vivo
(
24
). UV-irradiation depletes ICU-BP from the transfected (e.g. IGF-II mRNA containing) cell line, but not from the non-transfected cell line. Because the presence or absence
of IGF-II mRNA is the only difference between these cell lines this result
suggests that also
in vivo
ICU-BP is associated with IGF-II mRNAs.
Some crude
in vitro
mRNA degradation systems have been reported (
32
-
34
), and components of degradation pathways have been identified, but neither of
the two mRNA binding proteins involved in degradation that have been cloned to
date, AUF1 (
35
) and IRP (
36
), are sufficient for the regulated degradation of their cognate mRNAs. For
investigation of the function of ICU-BP in cleavage, an obvious experiment would be to purify the protein and
test it in an
in vitro
cleavage assay. The resistance of the complex to high salt and its requirement
for Mg
2+
may aid in the purification of the protein. Cleavage of IGF-II mRNAs can be performed
in vitro
, but we have not yet succeeded in establishing an assay that is sufficiently
reproducible (W. Scheper, unpublished data).
Our current working hypothesis for the mechanism of cleavage involves a
functional relation between the formation of ICU-RPC and the level of IGF-II mRNA. In this model, formation of ICU-RPC facilitates the cleavage reaction and thereby the IGF-II mRNA degradation. However, formation of the ICU-RPC is not sufficient for cleavage and other
proteins, possibly binding to the two additional stem-loops, are also required.
The function of the cleavage reaction is still elusive. It might be a first step in the degradation of IGF-II mRNAs, as has been suggested for other endonucleolytic cleavage
reactions (
37
-
39
). Alternatively, the 3' cleavage product may have some function in the cell by itself, since it
is unusually stable for a cleavage product. Obviously, RNA can serve more
functions than being an intermediate between the genome and the protein. In mammalian cells, effects of
3' UTRs on differentiation and tumorigenesis were reported (
40
,
41
). If the 1.8 kb RNA exerts a function in the cell, its accumulation in the
cytoplasm indicates that this is where to look for its action. Of course, these
two possible functions of the cleavage reaction are not mutually exclusive.
In previous studies (
20
,
21
) we have identified the
cis
-acting elements involved in site-specific cleavage of human IGF-II mRNAs. In this study we initiated the characterization of
trans
-acting factors that play a role in this process, and have identified a
novel protein, ICU-BP, that binds to the stem structure of IGF-II mRNAs. Future experiments will focus on: (i) the function and
regulation of ICU-BP; to address this problem, an
in vitro
cleavage assay has to be developed; and (ii) the possible function of the 1.8 kb
RNA. For the latter it will be interesting to investigate the effect of over-expression of the 1.8 kb RNA in cultured cells and transgenic animals.
We thank Tamás Henics for communicating data prior to publication, and Linda Nolten, Luc Rietveld and Richard Rodenburg for
stimulating discussions.
REFERENCES
Return