ABSTRACT
Eukaryotic cellular mRNAs contain a cap at their 5
'
-ends, but some viral and cellular mRNAs bypass the cap-dependent mechanism of translation initiation in favor of internal
entry of ribosomes at specific RNA sequences. Cap-dependent initiation requires intact initiation factor eIF4G (formerly eIF-4
[gamma]
, eIF-4F
[gamma]
or p220), whereas internal initiation can proceed with eIF4G cleaved by
picornaviral 2A or L proteases. Injection of recombinant coxsackievirus B4
protease 2A into
Xenopus
oocytes led to complete cleavage of endogenous eIF4G, but protein synthesis
decreased by only 35%. Co-injection of edeine reduced synthesis by >90%, indicating that eIF4G-independent synthesis involved ongoing initiation. The spectrum of
endogenous proteins synthesized was very similar in the presence or absence of
intact eIF4G. Translation of exogenous rabbit globin mRNA, by contrast, was
drastically inhibited by eIF4G cleavage. The N-terminal cleavage product of eIF4G (cp
N
), which binds eIF4E, was completely degraded within 6-12 h, while the C-terminal cleavage product (cp
C
), which binds to eIF3 and eIF4A, was more stable over the same period. Thus,
translation initiation of most endogenous mRNAs in
Xenopus
oocytes requires no eIF4G, or perhaps only cp
C
, suggesting a cap-independent mechanism.
The translation of most cellular mRNAs is thought to be initiated by a 5'-end-dependent mechanism involving protein synthesis initiation
factors of the eIF4 group. These factors collectively bind the 7-methylguanosine-containing cap, unwind secondary structure in the mRNA and catalyze
its binding to the 43S initiation complex to form the 48S initiation complex (
1
,
2
). Members of this group include eIF4E, a cap binding protein, eIF4A, an RNA
helicase, eIF4B, which stimulates the helicase and has RNA annealing activity,
and eIF4G (formerly eIF-4[gamma], eIF-4F[gamma] or p220), which acts as a linker in this process,
specifically complexing with eIF4E, eIF4A and eIF3 to bring together the mRNA 5'-end, the RNA helicase activity and the 40S subunit (
3
). The isolated complex of eIF4E, eIF4A and eIF4G is referred to as eIF4F.
In some situations, however, cap-independent initiation becomes predominant. The best understood instance
of a switch from cap-dependent to cap-independent initiation involves picornavirus infection of mammalian
cells. Picornaviral RNA contains an internal ribosome entry site (IRES) which is capable of productively binding ribosomal subunits and initiating translation independent of the 5'-end, even directing translation initiation on circular RNA
(reviewed in
4
). Viral mRNAs are not alone in their use of IRES elements as a means to utilize
a cap-independent mechanism for initiation. mRNAs encoding the cellular proteins
immunoglobulin heavy chain binding protein, fibroblast growth factor 2,
Drosophila
Antennapedia and yeast TFIID and HAP4 also contain IRESes and can initiate
translation internally (
4
,
5
). Even the mRNA for initiation factor eIF4G itself has been found to contain a
potent IRES (
6
). Picornaviral infection results in a dramatic shutdown of host cell protein
synthesis (
7
). Within a few hours following infection by poliovirus, synthesis of cellular
proteins is nearly undetectable while synthesis of viral polyproteins rapidly
becomes predominant. Entero- and rhinoviruses contain a protease (protease 2A) which cleaves eIF4G at
a specific site (amino acid 486 of rabbit eIF4G;
8
,
9
), physically separating the domains of eIF4G which bind eIF4E from those which
bind eIF4A and eIF3 (
3
,
10
) and resulting in the separation of the cap binding function from the RNA-unwinding and ribosome-binding functions of eIF4G. Under these conditions, cap-dependent initiation is inhibited and cap-independent initiation prevails in the cell. Cleavage of
eIF4G by protease 2A correlates with the shutdown of host protein synthesis and
is thought to be an important step in viral take-over of the protein synthetic apparatus of the host cell. Translation of
cellular mRNAs and mRNA cap-binding activity can be restored to virus-infected cell extracts by adding the intact eIF4F complex (
11
,
12
). eIF4 factors, therefore, are at the center of the mechanistic decision to switch from cap-dependent to cap-independent initiation.
The relative contributions of cap-dependent and cap-independent initiation
in vivo
can be conveniently addressed in
Xenopus
oocytes. Fully grown oocytes are arrested both in cell cycle and development,
yet synthesize protein actively (
13
,
14
). The translational efficiency of
Xenopus
stage VI oocytes is comparable with that of mammalian reticulocytes at similar
temperatures (22oC;
15
). They have amassed stores of maternally inherited mRNAs which provide genetic
information for subsequent embryonic development. Following transcription and
splicing, many of these mRNAs are transported from the nucleus in a
ribonucleoprotein complex (mRNP), masked from the translational machinery in
the cytoplasm by RNA-binding proteins which include FRGY2, and do not enter the pool of mRNA
available for translation (
16
-
18
). Despite this fact, the pool of available mRNA remains in excess of the
translation initiation potential of the non-stimulated oocyte. Studies utilizing microinjection of mRNAs have
demonstrated that they compete with endogenous mRNAs and with each other for
the existing initiation machinery (
19
-
21
). The limiting activity is thought to be that of a member of the eIF4 group of
initiation factors (
22
). The stored maternal mRNAs remain untranslated until recruited for translation
in response to a developmental stimulus such as progesterone or fertilization
(reviewed in
23
,
24
). Their utilization in a cell cycle-dependent manner is likely mediated both by unmasking of these mRNAs and
regulation of the translation initiation apparatus (
25
).
In this study we address cap-dependent and cap-independent initiation in
Xenopus
oocytes by microinjecting purified recombinant protease 2A from coxsackievirus
B4 (CVB4) to bring about proteolytic cleavage of eIF4G. Since oocytes are both
synthetically active and non-proliferating, we have been able to focus on the isolated effects of
disrupting eIF4G function on overall translation in a single, living cell. We
find that eIF4G cleavage severely inhibits cap-dependent initiation but causes only minor inhibition of total protein
synthesis. The results suggest that much of the actively translating mRNA in
the oocyte may initiate translation via a cap-independent pathway, although there is no evidence at present to suggest
this is internal initiation.
Recombinant coxsackievirus B4 protease 2A (CV2A) was a gift of Dr Barry Lamphear
(Louisiana State University Medical Center). It was produced in an
overexpressing
Escherichia coli
strain (a gift of Dr Tim Skern, University of Vienna, Vienna, Austria) and
purified to homogeneity as previously described (
26
,
27
). Protease was diluted at least 10-fold in modified Barth's saline solution (MBS;
28
) just prior to microinjection and kept on ice.
Frogs were purchased from Xenopus I (Madison, WI). Ovaries were surgically
removed from non-hormone-stimulated females and rinsed with MBS. Oocytes were isolated
manually with watchmakers forceps and a toothpick, sorted to select stage VI
oocytes and remove those with blemishes, and cultured at room temperature (22-24oC) in MBS containing 10 [mu]g/ml penicillin G and 10 [mu]g/ml streptomycin sulfate. Solution volumes of 20, 25 or 30
nl were microinjected equatorially into the oocyte cytoplasm using a Nanoject
(Drummond, King of Prussia, PA) delivery system and beveled tip (30 [mu]m) microcapillary needles (
28
). The follicle cells surrounding individual oocytes were
not removed
by collagenase/pronase treatment due to the dramatic adverse effects on protein
synthesis in the 24 h period following treatment (
29
). Rabbit globin mRNA was isolated as described (
30
) and then further purified by a second round of oligo(dT)-cellulose chromatography (
31
).
Metabolic labeling was used in preference to microinjection of radioactivity to
prevent distortion of the endogenous amino acid pool, to allow more precise
dosage to all of the experimental oocytes and to provide a continuous supply of
radiolabel over the course of several hours which was not subject to depletion
or leakage. After culturing, groups of three oocytes (in either duplicate or
triplicate) were transferred to 1.5 ml microcentrifuge tubes and excess buffer
removed with a drawn Pasteur pipette. Labeling was begun by the addition of 20 [mu]l 0.5 mCi/ml [
35
S]methionine (>1000 Ci/mmol; ICN Radiochemicals) in MBS at room temperature (22-24oC). Labeling was ended by rinsing oocytes with 0.5-1 ml MBS twice and freezing in dry ice. Oocytes were
homogenized at 4oC in extraction buffer (EB; 50 mM Tris-HCl, pH 7.5, 0.5 M urea, 2% Nonidet P-40, 5% 2-mercaptoethanol, 1 mM phenylmethylsulfonylfluoride) with
microtube pestles (Sarstedt), centrifuged at 25 000
g
for 5 min in a microfuge and aliquots from the supernatant spotted to duplicate
Whatmann 540 filters. One filter from each sample was washed in 5% (w/v)
trichloroacetic acid (TCA), 1 mM methionine for >= 30 min, boiled in 10% (w/v) TCA for 5 min, rinsed in methanol, rinsed in acetone, air-dried and the radioactivity measured by liquid scintillation
spectrometry. The second filter was air-dried directly for assay of radioactivity by liquid scintillation
spectrometry and served as a measure of total radioactive uptake by the
oocytes. The initial rate of [
35
S]methionine uptake by oocytes was always much greater than the rate of
incorporation into TCA-precipitable material and was essentially unaffected by the various
treatments, indicating that the rate of incorporation represented the rate of
protein synthesis. Nevertheless, equilibration with the oocyte methionine pool
caused an ~10 min lag in incorporation into protein (data not shown) and may have
caused a slight underestimation of protein synthesis during the initial 20 min
of labeling. After 20 min, complete equilibration had occurred, as demonstrated
by the parity of radioactivity between extract and labeling medium. Under these
conditions, oocytes generally demonstrated linear incorporation of
radioactivity into protein following the lag period for at least 3 h. Estimates
of the endogenous methionine pool (43 pmol/oocyte;
32
) and nominal oocyte volume (1 [mu]l;
28
) were used to calculate synthetic rates.
Oocytes were homogenized at 4oC in buffer EB, centrifuged at 25 000
g
for 5 min and the supernatants resolved by SDS-PAGE. Immunoblotting for eIF4G was performed using N-terminal antiserum which recognizes amino acids 327-342 (
33
) or C-terminal antiserum which recognizes amino acids 653-666 (
3
). Radiolabeled extracts were resolved by SDS-PAGE on 10 or 15% gels, which were subsequently fixed in 45% methanol, 5%
acetic acid then soaked in 1 M sodium salicylate prior to drying and
autoradiography. Image analysis was by PhosphorImager detection with
quantitation using ImageQuant (Molecular Dynamics) or by scanning on a Hewlett
Packard ScanJet with quantitation by NIH Image v1.5 software.
Total RNA was prepared from oocytes by the SDS/proteinase K method (
34
). RNA from two oocytes was resolved by formaldehyde-1.5% agarose gel electrophoresis and capillary blotted to NitroPlus 2000
membranes (MSI Separations) (
35
). An antisense RNA probe to endogenous B9 mRNA (
36
) was transcribed with T7 RNA polymerase (Promega) and [[alpha]-
32
P]UTP (Dupont/NEN) from linearized plasmid. An antisense DNA globin probe was
primer extended on plasmid pOG9 (
37
), encoding a rabbit [beta] globin cDNA, using [[alpha]-
32
P]dCTP (Dupont/NEN) (
35
).
To verify that antibodies against human eIF4G would specifically recognize eIF4G
from
Xenopus
oocytes, we utilized anti-peptide antibodies directed at N- and C-terminal portions of human eIF4G protein (
3
,
33
) for Western blotting of extracts of whole oocytes. A closely spaced triplet of
polypeptides migrating on SDS-PAGE with an apparent molecular mass of ~230 kDa reacted strongly with the N-terminal antibody (Fig.
1
A). Mammalian eIF4G is also heterogeneous on SDS-PAGE, possibly due to post-translational modification (
3
). The other, fainter bands at ~150 kDa on the Western blot may represent an N-terminal breakdown product of eIF4G or a full-length eIF4G polypeptide which has not undergone the putative
post-translational modification. The C-terminal antibody also recognized the ~230 kDa proteins, though with reduced sensitivity (Fig.
1
B).
Most picornaviruses target eIF4G for inactivation as a means of disrupting cap-dependent translation initiation in order to favor the translation of
their genomic RNAs, which do not contain caps (reviewed in
4
). These viruses encode a specific protease, 2A, which cleaves the eIF4G
molecule (
8
,
38
), effectively separating the domain which binds to the mRNA cap-binding protein, eIF4E, from that which interacts with the rest of the 48S
initiation complex (
3
,
10
). Cleavage of eIF4G in mammalian cells infected with poliovirus, for example,
correlates with a dramatic shutdown of host cell translation within 2-3 h of infection (
9
). Intact eIF4F complex purified from non-infected cells restores active translation and mRNA cap-binding function to extracts derived from poliovirus-infected cells, indicating that the only lesion to protein
synthesis is the inactivation of this cap-binding complex by 2A cleavage (
11
,
12
). Mutations in the viral 2A gene which inactivate the protease alleviate the
shutdown of host cell synthesis during infection (
42
,
43
), again linking eIF4G cleavage to the loss of cellular mRNA translation. Thus,
there is much correlative evidence from studies of picornaviral infection that
the cleavage of eIF4G by protease 2A leads to host synthesis shutdown.
However, there is also evidence to the contrary, i.e., that the cleavage of
eIF4G cannot be solely responsible for the dramatic shutdown of host protein
synthesis. Temporally, host protein synthesis shutdown actually lags behind
eIF4G cleavage by ~1 h in poliovirus-infected cells; this lag can be extended to at least 5 h when
infected cells are shifted to 28oC (
44
). Prior to the shutdown of cellular synthesis, IRES-directed translation has already been stimulated, indicating that separate
mechanisms mediate these events (
45
). In the presence of certain ionophores and agents which inhibit replication of
the viral genome, complete eIF4G cleavage still occurs, but cellular protein
synthesis continues at 50-100% of the rate of uninfected cells, leading to simultaneous translation
of cellular and viral RNAs (
46
,
47
). These seemingly contradictory observations may possibly be reconciled if one
considers that virus infection produces changes in intracellular cation
concentrations, membrane permeability and cytoskeletal structures, which could
also inhibit protein synthesis (
48
-
50
). Picornavirus infection also leads to increased phosphorylation of eIF2[alpha] by PKR in reponse to dsRNA and interferon, causing inhibition at another
step of translation initiation (
7
).
Simplified systems in which protease 2A alone is added may provide more insight
into the effects of eIF4G cleavage on protein synthesis. Translation of globin
mRNA is dramatically inhibited by eIF4G cleavage
in vitro
using reticulocyte lysate treated with purified picornavirus protease 2A (
26
,
51
), similar to our observations in oocytes. Likewise, the action of FMDV leader
protease inhibits the translation of capped mRNAs to various extents (
52
,
53
). On the other hand, eIF4G cleavage can lead to stimulation of cap-independent initiation. Translation of RNAs containing some picornaviral
IRESs is stimulated 6- to 10-fold in lysates treated with protease 2A (
51
,
54
). Translation of several cellular mRNAs introduced as uncapped RNAs is also
stimulated 2- to 5-fold by eIF4G cleavage (
53
). The simple interpretation from the
in vitro
studies is that cleavage of eIF4G alone is sufficient both to disrupt cap-dependent translation and facilitate cap-independent translation, be it IRES-mediated or non-IRES-mediated.
Because the stable expression of protease 2A in cells appears to be quite toxic
(
55
), analyzing the specific effects of eIF4G cleavage on translation
in vivo
has been more complicated. To date, transient transfection of mammalian cells
with 2A-expressing vectors has been the only means to approach these questions
in vivo
, but it is clear from these studies that eIF4G cleavage leads to profound
translational inhibition of specific cap-dependent mRNAs (
56
,
57
). The stimulation of IRES-mediated translation initiation as well as that of uncapped cellular mRNAs
has also been borne out by these transfection studies (
45
,
57
), again indicating that eIF4G cleavage inhibits the cap-dependent pathway and stimulates the cap-independent pathway. However, the contribution of eIF4G cleavage to
host protein synthesis shutdown has not been addressed in these studies, as
only a subset of cells express protease 2A. As for the toxic effects of 2A
expression, they are difficult to assess in growing somatic cells, where the
secondary effects of inhibited translation of a specific mRNA, for example, may
impair the cell's subsequent ability to carry out DNA replication,
transcription or cell cycle events. In the present study, we have employed a
new
in vivo
system to assess the role of eIF4G cleavage in cap-dependent and cap-independent translation by injecting CV2A into quiescent
Xenopus
oocytes.
It is clear that the cap-dependent initiation pathway is active in oocytes. Indeed, we observe that
several species of
35
S-labeled proteins disappear upon cleavage of eIF4G, though they are
surprisingly few. A clearer demonstration, though, is that translation of
globin mRNA is completely abolished by cleavage of eIF4G. Globin mRNA is
efficiently utilized in stage VI oocytes, forming polyribosomes within 3-4 h of injection and completing a round of translation every 5-10 min; its utilization is 10- to 30-fold more efficient than in reticulocyte lysate (
13
,
21
). Early studies found that globin mRNA competes in a dose-dependent manner in the range 1-100 ng with endogenous mRNAs (
19
-
21
). The authors of these studies concluded that protein synthesis in the stage VI
oocyte is limited by a saturated translational capacity (e.g. factors,
subunits), not by limiting mRNA availability. We carefully chose the 5 ng dose
of globin mRNA in order to effect moderate competition such that globin
translation would accurately reflect the cap-dependent utilization of the available mRNA pool. The complete inhibition
of globin mRNA translation is evidence that the cap-dependent pathway was fully disrupted. Preloading of globin mRNA into
polyribosomes before eIF4G cleavage did not prevent inhibition, suggesting that
each new round of initiation involves a cap-dependent event. In general, the presence of a cap on an injected mRNA
enhances its translation ~25-fold in oocytes, whereas it increases stability only 5-fold (
58
). Translation of exogenous globin mRNA provides a very sensitive assay of cap-dependent translation initiation activity and is highly dependent on the
integrity of eIF4G.
In contrast to globin mRNA, most endogenous mRNAs in the stage VI oocyte can be
translated in the absence of intact eIF4G, suggesting a cap-independent pathway. Unlike virus-infected mammalian cells,
Xenopus
oocytes do not undergo a dramatic shutdown of host cell protein synthesis when
eIF4G is cleaved. Instead, only ~35% of cellular protein synthesis is inhibited. Given that translational
efficiency and ribosome transit times measured in
Xenopus
stage VI oocytes were found to be comparable with mammalian systems (
15
), these results cannot be reconciled by slow elongation rates in oocytes. Nor
is the modest inhibition by protease 2A due to low initiation activity in
oocytes, since edeine injection reveals that at least 90% of synthesis measured
within the 2 h labeling period is due to new initiation events. Edeine binds to
the small ribosomal subunit (
59
,
60
) and prevents joining of the large subunit to a 48S initiation complex,
probably by blocking initiation codon recognition by the scanning 40S subunit (
39
,
40
). Thus, edeine acts at an event subsequent to cap recognition and should
inhibit both cap-dependent and cap-independent initiation events, but not elongation (
39
). By using edeine in conjunction with eIF4G cleavage, we have demonstrated that
eIF4G-independent synthesis represents primarily
de novo
initiation as well. Insofar as cleavage of eIF4G completely disrupts cap-dependent translation initiation,
Xenopus
oocytes have the capacity to initiate translation on their resident mRNAs in a
primarily cap-independent manner. Our findings are consistent with earlier observations
that general protein synthesis in stage VI oocytes is resistant to inhibition
by even millimolar concentrations of cap analogs (
61
). However, our data do not suggest a particular mechanism for the cap-independent initiation of endogenous oocyte mRNAs; one cannot assume that
they are internally initiated in the fashion of IRES-containing RNAs (
4
).
We suggest that eIF4 factors preferentially stimulate the synthesis of a small
subset of proteins. These are likely to be encoded by mRNAs with a strong cap
dependence (
62
). For example, rapamycin inhibits insulin-stimulated phosphorylation of both PHAS-I and eIF4E in 32D cells containing the insulin receptor and IRS-1, reducing both eIF4E availability and affinity for mRNA, and
this results in inhibition of insulin-stimulated Myc synthesis (
63
). The stimulation of general protein synthesis, however, is inhibited only 10%
and actin synthesis, not at all. Overexpression of eIF4E, on the other hand,
preferentially enhances the translation of mRNAs encoding growth- and cell cycle-related proteins such as Myc, ornithine decarboxylase, ornithine
aminotransferase, bFGF, cyclin D1 and vascular proliferation factor (
64
-
67
). The idea that a small subset of mRNAs, which are specifically involved in
cell proliferation, is more dependent on the cap recognition and unwinding
machinery may explain the effects of eIF4E overexpression on growth phenotype.
These cells undergo oncogenic tranformation and display rapid proliferation,
both of which are reversible upon expression of antisense eIF4E RNA (
68
). Therefore, modulating the activity of eIF4 factors may regulate the
translation of growth-related mRNAs without greatly altering the general protein synthetic
activity of a cell. The 2A-induced cleavage of eIF4G causes similar inhibition of eIF4 factor
activity which is functionally equivalent to eIF4E inhibition or depletion.
Our results suggest that mRNAs encoding `housekeeping' proteins in meiotically
arrested oocytes may be less cap dependent than those encoding growth-related proteins and that cap-dependent initiation is limited to favor the expression of
housekeeping mRNAs. Stage VI oocytes remain arrested in the meiotic G
2
phase of the cell cycle for up to 3 months in the ovary without dividing,
differentiating or proliferating (
14
). As a consequence of progesterone stimulation, the oocyte may selectively
enhance cap-dependent initiation by increasing the activity or availability of the
eIF4 factors. Increased phosphorylation and complex formation of eIF4 factors
is observed upon hormone-induced meiotic maturation of
Xenopus
oocytes (
69
). Maturation also promotes the dissociation of FRGY2 from previously masked
mRNAs (
17
) and the oocyte experiences a 2- to 3-fold stimulation in protein synthesis which involves the recruitment
of maternal mRNAs, including those encoding c-Mos, cyclins A and B, and Cdk2, which are essential for cell cycle
progression (
24
,
70
). Coincident with the recruitment of these mRNAs, they undergo cytoplasmic 3' poly(A) elongation, which is also required for their mobilization into
polyribosomes. Translation of an mRNA in
Xenopus
oocytes is enhanced by a poly(A) tail and there is evidence that the extended
poly(A) tail facilitates reinitiation events and affects initiation events at
the mRNA cap (
71
,
72
). The prediction which follows from these studies is that maternal mRNAs
recruited during meiotic maturation in response to polyadenylation will be
strongly cap dependent for that recruitment.
It is tempting to suggest that stage VI oocytes naturally utilize cap-independent translation for most of their mRNAs, but we cannot discount
the possibilities that: (i) cap-independent synthesis is specifically induced by the cleavage of eIF4G, or
(ii) another protein stably bound to endogenous mRNAs functionally replaces
eIF4G. Possibility (i) is a quite attractive hypothesis, as it may ascribe a
function to the cleavage products of eIF4G which result from the action of the
protease. cp
N
accumulates only transiently and is rapidly further degraded by cellular
proteases, perhaps because it contains all of the significant PEST signals of
the eIF4G molecule. Since it also contains the eIF4E binding site, the function
served by eIF4E in 48S complex formation disappears with the degradation of cp
N
. cp
C
, on the other hand, is relatively stable in the oocyte after intact eIF4G has
disappeared. cp
C
contains the binding sites for eIF4A and eIF3 (and hence the 40S subunit), and
it seems reasonable to suggest that such a complex might be competent to
initiate translation on an mRNA in a cap-independent manner. Indeed, cp
C
-containing ribosomes were recently shown to support cap-independent initiation in reticulocyte lysate (
73
). We observe that the synthesis of a few proteins (e.g. actin) from endogenous
mRNA may even be enhanced by eIF4G cleavage. It has been suggested that highly
efficient mRNAs like actin become less dependent on cap-mediated mechanisms once they are assembled into active polyribosomes (
74
). The enhanced synthesis of actin appears to correlate with stable accumulation
of cp
C
, consistent with the notion that stimulation of the cap-independent pathway is advantageous for actin mRNA.
As a system to assay the effects of eIF4G cleavage on cap-dependent and cap-independent initiation
in vivo
,
Xenopus
oocytes offer advantages over mammalian cells either infected with picornavirus
or expressing protease 2A. The injection of recombinant protease 2A does not
elicit the deleterious side-effects on transcription, eIF2 activity and cellular physiology which are
observed during picornaviral infection of cells, allowing simpler
interpretation of experiments. The warfare waged by both virus and host
involving interferon response, dsRNA response, membrane disintegration,
additional viral proteases (e.g. 3C), RNA virus replication, IRES-driven translation, etc., does not complicate studies in oocytes. Because
they are arrested in the cell cycle and transcriptionally silent, secondary
effects on gene expression and cell proliferation can also be avoided. In such
a system, where isolated effects on translation may be assayed
in vivo
, we find that eIF4G cleavage causes only minor inhibition of total protein
synthesis, but severely inhibits cap-dependent initiation.
We thank Dr Tim Skern for the expression plasmid to produce recombinant CV2A and
Dr Barry Lamphear and Ai-Li Cai for purified protease. Dr Raul Mendez, Dr Barry Lamphear, Weiniu
Gan, Ming-Xing Gao, Vaishali Kereketta and Boyd Butler improved the manuscript with
their comments and Dr Wolfgang Sommergruber provided helpful discussion. This
study was supported by grant no. 3078 from the Council for Tobacco Research-USA, Inc. and grant no. GM20818 from the National Institute of General
Medical Sciences.
REFERENCES
Return