Báo cáo khoa học: Characterization of eIF3k A newly discovered subunit of mammalian translation initiation factor eIF3 potx

Characterization of eIF3k
A newly discovered subunit of mammalian translation initiation factor eIF3
Greg L. Mayeur, Christopher S. Fraser, Franck Peiretti, Karen L. Block and John W. B. Hershey
Department of Biological Chemistry, School of Medicine, University of California, Davis, CA, USA
Mammalian translation initiation factor 3 (eIF3) is a
multisubunit complex containing at least 12 subunits with an
apparent aggregate mass of  700 kDa. eIF3 binds to the
40S ribosomal subunit, promotes the binding of methionyl-
tRNA
i
and mRNA, and interacts with several other initi-
ation factors to form the 40S initiation complex. Human
cDNAs encoding 11 of the 12 subunits have been isolated
previously; here we report the cloning and characterization
of a twelfth subunit, a 28-kDa protein named eIF3k. Evi-
dence that eIF3k is present in the eIF3 complex was
obtained. A monoclonal anti-eIF3a (p170) Ig coimmuno-
precipitates eIF3k with the eIF3 complex. Affinity purifica-
tion of histidine-tagged eIF3k from transiently transfected
COS cells copurifies other eIF3 subunits. eIF3k colocalizes
with eIF3 on 40S ribosomal subunits. eIF3k coexpressed
with five other ÔcoreÕ eIF3 subunits in baculovirus-infected
insect cells, forms a stable, immunoprecipitatable, complex
with the ÔcoreÕ. eIF3k interacts directly with eIF3c, eIF3g
and eIF3j by glutathione S-transferase pull-down assays.
Sequences homologous with eIF3k are found in the genomes
of Caenorhabitis elegans, Arabidopsis thaliana and Droso-
phila melanogaster, and a homologous protein has been
reported to be present in wheat eIF3. Its ubiquitous
expression in human tissues, yet its apparent absence in
Saccharomyces cerevisiae and Schizosaccharomyces pombe,

suggest a unique regulatory role for eIF3k in higher organ-
isms. The studies of eIF3k complete the characterization of
mammalian eIF3 subunits.
Keywords: protein synthesis; translation initiation; initiation
factor; eIF3; eIF3k.
Initiation of protein synthesis in eukaryotic cells involves
formation of an 80S ribosomal complex containing the
initiator methionyl-tRNA
i
bound to the initiation codon in
a mRNA. It is a multistep process promoted by a number
of proteins called eukaryotic translation initiation factors
(eIFs). Currently, at least 12 eIFs, minimally composed of
29 distinct subunits, have been identified [1]. eIF3, the
largest initiation factor, is a multisubunit complex with an
apparent molecular mass of  700 kDa. eIF3 plays an
essential role in translation by binding directly to the 40S
ribosomal subunit and promoting formation of the 40S pre-
initiation complex consisting of the Met-tRNA
i
eIF2GTP
ternary complex, eIF1, eIF1A and the 40S ribosomal
subunit [2]. eIF3 also promotes the binding of 5¢-m
7
G-
capped mRNA through its interaction with eIF4G, the
largest member of the eIF4F cap-binding complex [3,4]. The
40S preinitiation complex then scans the mRNA in a 5¢ to 3¢
direction, until the initiation codon AUG is selected. Upon
recognition of the initiation codon, eIF5 stimulates the

hydrolysis of the GTP bound to eIF2, and the eIFs are
ejected from the ribosome. The 60S ribosomal subunit then
joins the 40S initiation complex aided by eIF5B [5].
Mammalian eIF3 is known to interact with other initiation
factors including eIF1 [6], eIF4B [7] and eIF5 [2]. Clearly, it
plays a central role in the initiation pathway, possibly as an
organizer of other initiation factors on the surface of the
40S ribosomal subunit.
Eleven nonidentical subunits have been identified in
mammalian eIF3 and their cDNAs have been cloned and
characterized: a (p170) [8], b (p116) [9], c (p110) [10], d (p66)
[11], e (p48) [12], f (p47) [11], g (p44) [13], h (p40) [11], i (p36)
[10], j (p35) [13] and l (p69) [14]. This contrasts with eIF3
from Saccharomyces cerevisiae, which contains only six
subunits, all homologous with the mammalian subunits of
the same letter designation: a (TIF32) [15], b (PRT1) [16], c
(NIP1) [17], g (TIF35) [18], i (TIF34) [19] and j (HCR1) [20].
The presence of at least six additional subunits in mamma-
lian eIF3 suggests a potentially important regulatory role
for these subunits. Interactions of the various yeast subunits
with one another and with other eIFs have been extensively
characterized and mapped [21], but a high-resolution, three-
dimensional structure for neither yeast nor mammalian
eIF3 is yet available. The knowledge of eIF3 subunit protein
sequences provided by their cDNAs has proved useful in
our understanding of the structure and function of eIF3. We
therefore turned out attention to an uncharacterized protein
of 28 kDa that is found in preparations of eIF3 derived
from HeLa cells. In this work, we describe the isolation and
characterization of a cDNA encoding this 28 kDa protein,

which we have named eIF3k. Evidence is presented that
demonstrates that eIF3k is indeed the twelfth subunit of
mammalian eIF3.
Correspondence to J. Hershey, Department of Biological Chemistry,
School of Medicine, University of California, Davis, CA 95616, USA.
Fax: + 1 530 752 3516, Tel.: + 1 530 752 3235,
E-mail:
Abbreviations: eIF3, mammalian translation initiation factor 3;
GST, glutathione S-transferase.
(Received 23 July 2003, revised 20 August 2003,
accepted 26 August 2003)
Eur. J. Biochem. 270, 4133–4139 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03807.x
Materials and methods
Materials
eIF3 was prepared from human HeLa cells as described
previously [22]. Polyclonal antiserum against rabbit eIF3
was prepared in a goat and characterized as previously [23].
Monoclonal antibody specific for eIF3a was a generous gift
of J. T. Parsons (University of Virginia, USA). The
monoclonal antibody against c-myc was obtained from
Santa Cruz Biotechnology, Inc. Anti-FLAG agarose and
FLAG peptides were obtained from Sigma-Aldrich Co. All
cell culture materials were from Mediatech.
Cloning of cDNAs encoding eIF3k
The 28 kDa protein in highly purified eIF3 was separated
from the other subunits by SDS/PAGE. The protein was
digested in the gel with Lys-C protease and thefragments were
fractionated by high-performance liquid chromatography.
N-terminal sequences of internal peptides were obtained by
automated Edman degradation in the Protein Structure

Laboratory (University of California, Davis, USA). A
peptide sequence, YNPENLATLER, matched 11 of 11 pre-
dicted residues from the human EST clone zb98c06.r1. Clone
zb98c06.r1 (GenBank nucleotide accession number W38661)
was obtained from IMAGE (ATCC) and its 931-bp cDNA
insert was sequenced on both strands. The cDNA sequence
predicts a protein of 218 amino acid residues, consistent with
it encoding the 28 kDa eIF3k subunit. Additional BLAST
searches identified more than 20 overlapping human ESTs.
The cDNA insert in zb98c06.r1 was amplified by PCR with
the forward primer 5¢-CCCATATGGCC
ATGGCGA
TGTTTGAGCAG-3¢ and reverse primer 5¢-GCTCGAGA
AGC
TTACTGGGAGGAGGCCATG-3¢ (restriction sites
are in bold and eIF3k start and stop codons are underlined).
The NdeI- and XhoI-digested PCR product was ligated into
pET28c (Novagen) to generate a histidine-tagged bacterial
expression construct, pET-Hisp28. An untagged construct,
pETp28, was generated by digesting the PCR product with
NcoIandXhoI and ligating it into the equivalent sites in
pET28c. Similarly, the eIF3k cDNA was cloned into
pcDNA3.1/myc-His A (Invitrogen) for expression in
COS-1 cells. PCR amplification from zb98c06.r1 with the
forward primer 5¢-CCTCGAGC
ATGGCGATGTTTGA
GCAG-3¢ and reverse primer 5¢-GGAAGCTTGGGAG
GAGGCCATGAT-3¢ (restriction sites are in bold and eIF3k
startcodonisunderlined;thestopcodonhas beenremovedfor
C-terminal c-myc tag) followed by digestion with XhoIand

SfuI allowed construction of pcDNA/myc-His:p28.
Synthesis of recombinant eIF3k
in vitro
Radiolabeled eIF3k was synthesized in the TnT system
(Promega) programmed with pETp28 in the presence of
[
35
S]Met. The radiolabeled protein product was used
without further purification.
Transient transfection of COS-1 cells
COS-1 cells were routinely grown in Dulbecco’s medium
with 10% fetal bovine serum. Cells at approximately 60%
confluence were transiently transfected by the DEAE-
Dextran method [24] with 10 lg pcDNA/myc-His:p28. In
some cases, proteins were radiolabeled by incubating the
transfected cells in Met- and Cys-free DMEM (Mediatech)
containing 0.2 mCiÆmL
)135
S-Trans Label (NEN) for 18 h
prior to lysis at 48 h. Cells were lysed in Lysis Buffer
containing 10 m
M
Tris/HCl pH 7.4, 100 m
M
KCl, 5 m
M
MgCl
2,
0.5% NP-40, 0.5% deoxycholate, 0.1% Tri-
ton X-100, and Complete Protease Inhibitor (Promega).

Lysates were clarified by centrifugation at 10 000 g for
10 min.
eIF3k expression in baculovirus-infected insect cells
The FASTBAC vector (Invitrogen) was modified so that
expressed proteins would contain the FLAG epitope fused
at their N-termini (C. S. Fraser, J. Y. Lee, M. Bushell,
G. L. Mayeur, & J. W. B Hershey, unpublished observa-
tion). The NdeI-XhoI DNA fragment encoding untagged
eIF3k derived from pETp28 above was inserted into the
same restriction sites of the FLAG-FASTBAC vector to
generate FLAG-eIF3k. FASTBAC vectors for the expres-
sion of untagged eIF3a, eIF3b, eIF3c, eIF3g and eIF3i were
constructed similarly as described elsewhere (C. S. Fraser,
J. Y. Lee, M. Bushell, G. L. Mayeur, & J. W. B Hershey,
unpublished observation). The six recombinant FASTBAC
vectors were recombined individually with baculovirus
DNA using DH10BAC Escherichia coli (Invitrogen) and
the high molecular mass DNA (ÔbacmidÕ) was purified
according to the manufacturer’s guidelines. Sf9 cells were
transfected with bacmid DNA by using the calcium
phosphate method (Promega) and viral stocks were pre-
pared by three-step growth amplification according to the
manufacturer’s guidelines. Sf9 cells (1 · 10
7
) were coinfected
with a mixture of the six baculovirus strains and grown for
24 h according to procedures in the Life Technologies Bac-
to-Bac manual. Cells supplemented with 0.5 mCi [
35
S]Met

for an additional 36 h were harvested, after placing on ice,
by washing once with NaCl/P
i
(50 m
M
Na phosphate,
pH 7.0, 150 m
M
NaCl) and scraping into 1 mL of Buffer A
[20 m
M
Tris/HCl (pH 7.5), 120 m
M
KCl, 10 m
M
2-merca-
ptoethanol, 1% (v/v) Triton X-100, 10% glycerol]. Follow-
ing a 5-min incubation on ice with occasional vortexing,
extracts were centrifuged for 10 min at 12 000 g in a cooled
microcentrifuge. The supernatant was either used immedi-
ately or frozen in liquid nitrogen and stored at )70 °C.
Glutathione
S
-transferase (GST) pulldowns
For GST fusions, DNA encoding each of the subunits of
eIF3 was inserted into pGEX4T-1 (Amersham Pharmacia)
essentially as described [10, 13]. Expression of the con-
structs generated GST fused in-frame at the N-terminus of
the individual eIF3 subunits. GST-fusion proteins were
expressed in E.coli BL21 and purified according to

manufacturer’s instructions (Amersham Pharmacia). The
amount of cell lysate incubated with the beads was adjusted
so that a similar amount of each subunit was bound to the
beads. The beads were washed once and then resuspended
in 500 lL of binding buffer [75 m
M
KCl, 20 m
M
Hepes
(pH 7.5), 0.1 m
M
EDTA, 2.5 m
M
MgCl
2
,1m
M
dithiothre-
itol, 1% nonfat dry milk and 0.05% NP-40]. In vitro
4134 G. L. Mayeur et al. (Eur. J. Biochem. 270) Ó FEBS 2003
translated, [
35
S]Met-eIF3kwasincubatedwiththebeads
containing the GST fusion proteins for 2 h at 4 °C. Samples
were washed and bound proteins were eluted with 2· SDS
sample buffer and analyzed by SDS/PAGE.
Results
Preparations of eIF3 purified from HeLa cells routinely
contain a previously uncharacterized protein with a mobility
in SDS gels of 28 kDa. To clone the cDNA encoding the

28 kDa protein, a partial peptide sequence, YNPENLA
TLER, was obtained and used to identify human cDNA
sequences (ESTs) as described in Materials and methods.
A plasmid carrying a 921-bp insert was obtained from
IMAGE (ATCC). The insert contained an open reading
frame from nucleotides 1–637, with an 84 nucleotide
3¢-UTR containing a polyadenylation signal (AAUAAA)
at position 835. The first AUG was found at nucleotide
165–167, with two downstream, in-frame AUGs at 171–173
and 183–185. The first AUG probably serves as the initiator
codon, as it alone matches the Kozak consensus sequence
for initiation [25]. Furthermore, attempts to sequence eIF3k
indicated that its N-terminus is blocked, suggesting that the
amino acid following the initiator Met is small [26,27]. The
protein sequence starting at the first AUG is MAM
FEQMRA. Thus, the Met encoded by the first AUG is
followed by Ala, whereas those encoded by the downstream
AUGs are followed by Phe and Arg, residues that do not
normally follow acylated Mets. Finally, initiation at the first
AUG generates a protein with 218 residues and a mass of
25.1 kDa, in agreement with its migration in SDS gels.
Initiation at a possible AUG codon located upstream from
the DNA insert is unlikely, even though no in-frame
termination codon exists in the 164 nucleotide 5¢-UTR, as
the protein would contain over 270 residues, and thus would
be considerably larger than 28 kDa. We conclude that the
cloned cDNA (GenBank accession number AY245432)
encodes eIF3k, the 28 kDa subunit of eIF3.
The cloned cDNA appears to be nearly full length, as
Northern blot analysis of mRNAs from a number of cell

lines (Fig. 1) produces a single hybridization signal at
approximately 1.1 kb. While eIF3k mRNA expression is
ubiquitous, as determined by multiple tissue Northern blots,
the brain, testis and kidney express the highest levels of
eIF3k mRNA (data not shown). The eIF3k sequence
contains no obvious RNA binding motif, nor does the
protein bind RNA when analyzed by North-Western
blotting (results not shown). A number of putative phos-
phorylation sites for PKC and CK2 are present, but eIF3k
does not appear to be phosphorylated in vivo (data not
shown). The subunit is Leu-rich (10.6%), but the Leu are
distributed throughout the protein and do not show
characteristics of Leu zippers. The eIF3k gene [28] is
comprised of eight exons located on chromosome 19q13.2.
A putative TATA box is located 747 nucleotides upstream
from the identified start codon and 581 nucleotides
upstream from the beginning of the cloned insert. Three
retropseudogenes with sequence identities greater than 85%
are located on chromosomes 3, 4, and 5; however, the
longest translatable region only generates a peptide 60%
that of full-length eIF3k. A cDNA containing an ORF with
a sequence (GenBank acc. no. AF085358) identical to that
encoding eIF3k was cloned as the product of HSPC029,an
uncharacterized gene; however, the protein was neither
identified or characterized other than showing that its
mRNA is expressed ubiquitously [29].
A number of methods was used to demonstrate that
eIF3k is a true subunit of the eIF3 complex. First, eIF3k
cDNA was expressed in vitro and the product was compared
to the 28 kDa protein in eIF3. The eIF3k coding region was

subcloned into the E.coli expression vector pET28c to
create pETp28, as described in Materials and methods. The
vectorwasusedasatemplateforin vitro transcription and
translation in a rabbit reticulocyte lysate system. The
resulting
35
S-labeled protein comigrates with the 28 kDa
protein found in purified eIF3 (Fig. 2A).
Second, we asked if eIF3k is associated with eIF3
subunits partially purified by immunoprecipitation. Lysates
derived from COS-1 cells metabolically labeled with
35
S
were immunoprecipitated with an anti-eIF3a monoclonal Ig
in the absence or presence of RNase T1 and the precipi-
tates were fractionated by SDS/PAGE (Fig. 2B). A radio-
labeled band of 28 kDa in the immunoprecipitate
comigrates with recombinant eIF3k synthesized in vitro
from its cDNA. This result indicates that the 28 kDa
protein is a true component of eIF3, rather than a
copurifying contaminant. However, because we have been
unsuccessful in obtaining antibodies specific for eIF3k, we
cannot definitively prove that the immunoprecipitated
28 kDa protein actually corresponds to the eIF3k subunit
in this experiment. Therefore, to better demonstrate that the
28-kDa protein associated with eIF3 is eIF3k, the anti-
eIF3a immunoprecipitate was analyzed by two-dimensional
isoelectric focusing and SDS/PAGE. Other known eIF3
subunits are identified readily, as their migration positions
Fig. 1. Northern analysis of human eIF3k. Total RNA was prepared

from five cell lines by using the Qiashredder/RNeasy Kit (Qiagen) and
5 lg of RNA was loaded into each lane of a denaturing formaldehyde
agarose gel. Following electrophoresis, the RNA was transferred to a
nitrocellulose membrane and probed with a
32
P-labeled cDNA probe
constructed using the Random Primers (Life Technologies) system and
full-length eIF3k cDNA as template. Cell lines from which the RNA
was isolated are indicated at the top of the figure and the migration
positions of RNA size markers (kb) are shown on the left.
Ó FEBS 2003 Characterization of eIF3k (Eur. J. Biochem. 270) 4135
Fig. 2. Analysis of human eIF3 and eIF3k. (A) Purified HeLa eIF3 (lane 1) and in vitro synthesized recombinant eIF3k (lane 2) were subjected to
10% SDS/PAGE and Coomassie Blue staining (lane 1) or autoradiography (lane 2). Plasmid pETp28 carrying untagged eIF3k DNA behind a T7
promoter was expressed in the TnT-coupled transcription/translation system (Promega) containing [
35
S]Met. The migration positions of molecular
mass markers (kDa) are indicated on the left and eIF3 subunits are identified on the right with eIF3k shown in bold. (B) COS-1 cells were labeled
with [
35
S]Met and lysed as described in Materials and methods. RNase T1 (200 000 UÆmL
)1
) was included as indicated in the figure. Gamma-
Bind G Plus beads (Santa Cruz Biotechnology) were preloaded with nonimmune antibodies (control) or monoclonal anti-eIF3a Igs (anti eIF3a).
The beads were mixed with the radiolabeled COS-1 cell lysate for 30 min and washed three times in Lysis Buffer. Bound proteins were eluted in
SDS/PAGE loading buffer and fractionated by 10% SDS/PAGE. Radioactivity was detected by analysis on a phosphorimager (Molecular
Dynamics). The migration position of eIF3k was determined by examining the radiolabeled TnT product described in A (TNT p28). eIF3 subunits
are identified on the right. (C) Radiolabeled COS cells were lysed in Lysis Buffer and immunoprecipitated with anti-eIF3a monoclonal Igs as
described in B. The immunoprecipitate (right panel) or TnT-expressed radiolabeled eIF3k (left panel) was suspended in IPG strip rehydration buffer
and isoelectric focusing on 3–10 NL strips proceeded according to the IPGphor (Amersham Pharmacia) protocol. The second dimension was 10%
SDS/PAGE. Molecular mass markers are identified on the right, approximate pH is identified on the bottom and individual eIF3 subunits are

identified on the right panel. (D) COS-1 cells were transiently transfected with pcDNA3.1/myc-His:p28 or were mock-transfected as described
under Materials and methods. At 48 h post-transfection, the nonradiolabeled cells were lysed into Lysis Buffer and the lysate was diluted 1 : 10 into
Extraction Buffer (50 m
M
sodium phosphate pH 7.0, 300 m
M
NaCl) and loaded onto a Talon IMAC Column (Clontech). The column was washed
and eluted following the procedure recommended by the manufacturer. The load (10%) and entire TCA-precipitated eluate were analyzed by SDS/
PAGE and immunoblotted with anti-eIF3a and anti-c-myc monoclonal Igs. (E) Insect Sf9 cells were coinfected with recombinant baculovirus
strains expressing eIF3a, FLAG-tagged eIF3b, eIF3c, eIF3g and eIF3i with (lane 2) and without (lane 1) a virus strain expressing eIF3k,
35
S-labeled
and lysed as described in Materials and methods. Lysates were immunoprecipitated with anti-FLAG agarose, eluted with FLAG peptide and
analyzed by 10% SDS/PAGE and autoradiography. Radiolabeled eIF3 subunits are identified on the right. The asterisks mark degradation
products of higher molecular mass subunits.
4136 G. L. Mayeur et al. (Eur. J. Biochem. 270) Ó FEBS 2003
are known [30]. The 28 kDa protein exhibits an apparent pI
of 4.8 in the IEF dimension (Fig. 2C) and is located in the
same position in the gel as the in vitro synthesized,
radiolabeled eIF3k whose theoretical pI is 4.81. These
results provide strong support for the view that the 28 kDa
protein is eIF3k.
Additional evidence showing that eIF3k is present in eIF3
was generated by expressing recombinant eIF3k in mam-
malian cells. When DNA encoding eIF3k double-tagged
with (His)
6
and c-myc is transiently transfected into COS-1
cells and proteins are purified by Cobalt IMAC chroma-
tography, eIF3a is retained on the column along with the

(His)
6
-tagged eIF3k (Fig. 2D). A control where c-myc-
tagged eIF3k lacking the (His)
6
tagwasexpressedinCOS-1
cells does not result in retention of eIF3a. This supports the
view that eIF3k is truly associated with the eIF3 complex.
However, the apparent low yield of eIF3 associated with
tagged eIF3k is surprising, but could be explained by an
inefficient exchange of recombinant tagged eIF3k with the
endogenous eIF3k in eIF3. Similar results have been
observed in previous attempts to exchange other eIF3
subunits in our hands. We suggest that eIF3 is a highly
stable complex, a view consistent with results obtained by
pulse-chase labeling of cells followed by a time-course of
eIF3 immunoprecipitation [31].
Recent experiments in our laboratory have established
the baculovirus-infected insect cell expression system as a
suitable way to obtain large eIF3 subcomplexes (C. S.
Fraser, J. Y. Lee, M. Bushell, G. L. Mayeur, & J. W. B
Hershey, unpublished observation). Co-infection with up to
seven recombinant virus strains, each expressing a different
eIF3 subunit, has enabled us to generate and isolate a
variety of eIF3 subcomplexes. We asked if eIF3k might
associate with one or more of such subcomplexes, in
particular eIF3a,b,c,g,i, whose subunits correspond to
homologous yeast subunits found in the yeast eIF3 ÔcoreÕ
[32]. To this end, recombinant virus strains designed to
express untagged human eIF3a, eIF3c, eIF3g, and eIF3i,

along with FLAG-tagged eIF3b, were used to coinfect
insect cells in the presence or absence of recombinant eIF3k
virus as described in Materials and methods. Infected insect
cells were labeled with [
35
S]Met, and lysates were immuno-
precipitated with anti-FLAG agarose. The bound proteins
were subjected to SDS/PAGE and autoradiography.
Figure 2E shows the ÔcoreÕ complex of eIF3a, FLAG-
eIF3b, eIF3c, eIF3g and eIF3i has formed, and that eIF3k
copurifies with this complex. Proteins expressed in a control
experiment, in which none of the expressed proteins is
FLAG-tagged, do not bind to the affinity beads (data not
shown). The results indicate that eIF3k binds to one or
more of the subunits found in the eIF3 ÔcoreÕ complex.
GST-pulldown experiments (Fig. 3) were used to identify
the direct binding partners of eIF3k. Eleven subunits were
expressed as GST fusion proteins and used as bait for
[
35
S]Met-eIF3k synthesized in vitro. Direct interactions are
seen between eIF3k and eIF3c, eIF3g and eIF3j. We
expected that at least one member of the eIF3 ÔcoreÕ complex
(Fig. 2E) would interact with eIF3k in the GST pull-down
experiments, and indeed both eIF3c and eIF3g were
identified. It is likely, therefore, that eIF3k associates with
the eIF3c and eIF3g subunits, and possibly also with the
eIF3j subunit which is known to associate with eIF3b,g,i
(C. S. Fraser, J. Y. Lee, M. Bushell, G. L. Mayeur, & J. W.
B Hershey, unpublished observation).

If eIF3k is a subunit of eIF3, the portion of overexpressed
myc-tagged eIF3k that exchanges into endogenous eIF3 in
transiently transfected COS-1 cells is expected to bind to 40S
ribosomal subunits. We therefore examined the ribosome
binding of tagged eIF3k in relation to eIF3. COS-1 cells
transiently transfected with pcDNA3/myc-His:p28 were
Fig. 3. Identification of eIF3 subunits interacting with eIF3k. GST-
fusion proteins were expressed and purified as described in Materials
and methods. (Top panel) Coomassie Blue staining of purified GST-
fusion proteins. The eIF3 subunit present in the pulldown is identified
below, and the right-most lane contains GST alone. (Bottom panel)
Autoradiograph of in vitro [
35
S]Met-labeled eIF3k, produced using
pETp28 as a template, interacting with eIF3 subunits identified below.
Left-most lane indicates eIF3k input.
Fig. 4. eIF3k binds to 40S ribosomal subunits. COS-1 cells were
transiently transfected with pcDNA3.1/myc-His:p28 and lysed as
described under Materials and methods with the addition of
0.1mgmL
)1
cycloheximide, 0.2 mg mL
)1
heparin, 1 m
M
dithiothre-
itol and 24 UÆmL
)1
RNAguard (Promega) to the lysis buffer. Cleared
lysates were loaded onto a 10–50% sucrose gradient and centrifuged at

40 000 g in a Beckman SW40 rotor for 150 min. The gradient was
fractionated on an ISCO fractionator, 1 mL fractions were collected
and protein was precipitated with methanol. The proteins in the
fractions were analyzed by 10% SDS/PAGE and transferred onto a
poly(vinylidene difluoride) membrane. A portion of the membrane was
probed with anti-c-myc and anti-eIF3a monoclonal Igs and developed
with ECL (lower panels). The migration positions in the gradient of
40S, 60S and 80S ribosomes and polysomes, as identified by the A
260
scan (upper panel), are indicated above the immunoblot.
Ó FEBS 2003 Characterization of eIF3k (Eur. J. Biochem. 270) 4137
analyzed by sucrose gradient centrifugation, SDS/PAGE
and immunoblotting with anti-eIF3a and anti-c-myc Igs.
eIF3, as identified by anti-eIF3a Ig, was found mostly
associated with 40S ribosomal subunits (Fig. 4). A small
amount of eIF3k colocalizes with eIF3a in the 40S region of
the gradient, indicating an ability to bind to the ribosomal
subunit. However, most of the eIF3k is found at the top of
the gradient, as expected for an overproduced subunit of
eIF3. The low amount of 40S binding is consistent with the
coimmunoprecipitation experiments (Fig. 2B), again sug-
gesting that recombinant tagged eIF3k does not exchange
efficiently with endogenous eIF3k in the eIF3 complex.
Discussion
The cloning of a cDNA encoding eIF3k completes the
characterization of the known subunits of mammalian eIF3.
The calculated mass of the protein, 25.1 kDa, is consistent
with the apparent mass determined by SDS/PAGE,
28 kDa. The cDNA obtained from the partial peptide
sequence generates a polypeptide that migrates in one-

dimensional SDS/PAGE and two-dimensional IEF/SDS/
PAGE at exactly the same position as the human purified
eIF3k. However, the band corresponding to eIF3k does not
stain as heavily with Coomassie Blue as expected for a
subunit present in stoichiometric amounts. Similar to eIF3j,
the amount of eIF3k varies from preparation to prepar-
ation; perhaps due to the multistep purification process
required to obtain highly purified eIF3. Immunoprecipita-
tion from [
35
S]Met-labeled cells shows eIF3k to be closer to
stoichiometric than observed in the Coomassie stained gel.
Considerable additional data support the conclusion that
the cloned eIF3k cDNA codes for a true subunit of
mammalian eIF3. Recombinant eIF3k in transiently trans-
fected COS-1 cell lysates coimmunoprecipitates with other
eIF3 subunits by using a monoclonal antieIF3a antibody,
and this association is not RNA-dependent as shown by
insensitivity to RNase treatment. Similarly, other eIF3
subunits copurify with recombinant histidine-tagged eIF3k
following fractionation of the transiently transfected COS-1
cell lysates on a nickel affinity column. The protein also
associates in vivo with a ÔcoreÕ group of eIF3 subunits when
coexpressed in baculovirus-infected insect cells, and direct
interactions with eIF3c, eIF3g and eIF3j have been
observed. Recombinant eIF3k also binds to 40S ribosomal
subunits, presumably through eIF3. Finally, a protein
corresponding to eIF3k has been identified in preparations
of purified wheat and Arabidopsis thaliana eIF3 [33]. Its
identification as eIF3k was based on the similarity of its

sequence with that of the eIF3k protein described here; no
other characterization of the plant protein was reported. In
many of the experiments above, the exchange of recombin-
ant eIF3k into endogenous eIF3 appears to be inefficient,
making detection of the subunit’s association with eIF3
difficult. However, the sum of all of the results indicates that
eIF3k is a true subunit of eIF3.
The fact that eIF3k is not present in either S. cerevisiae
or S. pombe may indicate a specialized regulatory role in
the higher eukaryotic eIF3 complex. Further studies are
required to elucidate the precise location of eIF3k in the
structure of eIF3, and how this subunit contributes to the
activity of the factor.
Acknowledgements
We thank Susan MacMillan for preparations of purified human eIF3
and Chris Bradley for critically reading the manuscript. The work was
supported by National Institutes of Health. Grant GM22135 from the
U.S. Public Health Service.
References
1. Hershey, J.W.B. & Merrick, W.C. (2000) The Pathway and
Mechanism of Initiation of Protein Synthesis. In Translational
Control of Gene Expression (Sonenberg, N. Hershey, J.W.B. &
Mathews, M.B., eds), pp. 33–88, Cold Spring. Harbor Laboratory
Press, Cold Spring Harbor, NY.
2. Chaudhuri, J., Chowdhury, D. & Maitra, U. (1999) Distinct
functions of eukaryotic translation initiation factors eIF1A and
eIF3 in the formation of the 40S ribosomal preinitiation complex.
J. Biol. Chem. 274, 17975–17980.
3. Lamphear, B.J., Kirchweger, R., Skern, T. & Rhoads, R.E. (1995)
Mapping of functional domains in eukaryotic protein synthesis

initiation factor 4G (eIF4G) with picornaviral proteases. J. Biol.
Chem. 270, 21975–21983.
4. Korneeva, N.L., Lamphear, B.J., Hennigan, F.L. & Rhoads, R.E.
(2000) Mutually cooperative binding of eukaryotic translation
initiation factor (eIF) 3 and eIF4A to human eIF4G-1. J. Biol.
Chem. 275, 41369–41376.
5. Pestova,T.V.,Lomakin,I.B.,Lee,J.H.,Choi,S.K.,Dever,T.E.&
Hellen, C.U.T. (2000) The joining of ribosomal subunits in
eukaryotes requires eIF5B. Nature 403, 332–335.
6. Fletcher, C.M., Pestova, T.V., Hellen, C.U.T. & Wagner, G.
(1999) Structure and interactions of the translation initiation fac-
tor eIF1. EMBO J. 18, 2631–2639.
7. Me
´
thot, N., Song, M.S. & Sonenberg, N. (1996) A region rich in
aspartic acid, arginine, tyrosine and glycine (DRYG) mediates
eIF4B self association and interaction with eIF3. Mol. Cell. Biol.
16, 5328–5334.
8.Johnson,K.E.,Merrick,W.C.,Zoll,W.L.&Zhu,Y.
(1997) Identification of cDNA clones for the large subunit
of eukaryotic translation initiation factor 3. Comparison of
homologues from human, Nicotiana tabacum, Caenorhabditis
elegans and Saccharomyces cerevisiae. J. Biol. Chem. 272, 7106–
7113.
9. Me
´
thot, N., Rom, E., Olsen, H. & Sonenberg, N. (1997) The
human homologue of the yeast Prt1 protein is an integral part of
the eukaryotic initiation factor 3 complex and interacts with p170.
J. Biol. Chem. 272, 1110–1116.

10. Asano,K.,Kinzy,T.G.,Merrick,W.C.&Hershey,J.W.B.(1997)
Conservation and diversity of eukaryotic translation initiation
factor eIF3. J. Biol. Chem. 272, 1101–1109.
11. Asano, K., Vornlocher, H.P., Richter-Cook, N.J., Merrick, W.C.,
Hinnebusch, A.G. & Hershey, J.W.B. (1997) Structure of cDNAs
encoding human eukaryotic initiation factor 3 subunits. Possible
roles in RNA binding and macromolecular assembly. J. Biol.
Chem. 272, 27042–27052.
12. Asano, K., Merrick, W.C. & Hershey, J.W.B. (1997) The trans-
lation initiation factor eIF3-p48 subunit is encoded by int-6, a site
of frequent integration by the mouse mammary tumor virus
genome. J. Biol. Chem. 272, 23477–23480.
13. Block, K.L., Vornlocher, H.P. & Hershey, J.W.B. (1998) Char-
acterization of cDNAs encoding the p44 and p35 subunits of
human translation initiation factor eIF3. J. Biol. Chem. 273,
31901–31908.
14. Morris-Desbois,C.,Rety,S.,Ferro,M.,Garin,J.&Jalinot,P.
(2001) The human protein HSPC021 interacts with Int 6 and is
associated with eukaryotic translation initiation factor 3. J. Biol.
Chem. 276, 45988–45995.
4138 G. L. Mayeur et al. (Eur. J. Biochem. 270) Ó FEBS 2003
15. Vornlocher,H.P.,Hanachi,P.,Ribeiro,S.&Hershey,J.W.B.
(1999) A 110-kilodalton subunit of translation initiation factor
eIF3 and an associated 135-kilodalton protein are encoded by the
Saccharomyces cerevisiae TIF32 and TIF31 genes. J. Biol. Chem.
274, 16802–16812.
16. Keierleber, C., Wittekind, M., Qin, S.L. & McLaughlin, C.S.
(1986) Isolation and characterization of PRT1, a gene required
for the initiation of protein biosynthesis in Saccharomyces cerevi-
siae. Mol. Cell. Biol. 6, 4419–4424.

17. Gu, Z., Moerschell, R.P., Sherman, F. & Goldfarb, D.S. (1992)
NIP1, a gene required for nuclear transport in yeast. Proc. Natl
Acad. Sci. USA 89, 10355–10359.
18. Hanachi, P., Hershey, J.W.B. & Vornlocher, H.P. (1999) Char-
acterization of the p33 subunit of eukaryotic translation initiation
factor-3 from Saccharomyces cerevisiae. J. Biol. Chem. 274, 8546–
8553.
19. Naranda,T.,Kainuma,M.,MacMillan,S.E.&Hershey,J.W.B.
(1997) The 39-kilodalton subunit of eukaryotic translation
initiation factor 3 is essential for the complex’s integrity and for
cell viability in Saccharomyces cerevisiae. Mol. Cell. Biol. 17,
145–153.
20. Valasek,L.,Hasek,J.,Trachsel,H.,Imre,E.M.&Ruis,H.(1999)
The Saccharomyces cerevisiae HCR1 gene encoding a homologue
of the p35 subunit of human translation initiation factor 3 (eIF3)
is a high copy suppressor of a temperature-sensitive mutation in
the Rpg1p subunit of yeast eIF3. J. Biol. Chem. 274, 27567–27572.
21. Valasek,L.,Mathew,A.,Shing,B.S.,Nielsen,K.H.,Szamecz,B.
& Hinnebusch, A.G. (2003) Yeast eIF3 Subunits TIF32/a and
NIP1/c and eIF5 Make Critical Connections with the 40S Ribo-
some in vivo. Genes Dev. 17, 786–799.
22. Brown-Luedi,M.L.,Meyer,L.J.,Milburn,S.C.,Yau,M P.P.,
Corbett, S. & Hershey, J.W.B. (1982) Protein synthesis initiation
factors from human HeLa cells and rabbit reticulocytes are
similar: comparison of protein structure, activities and immuno-
chemical properties. Biochemistry 21, 4202–4206.
23. Meyer, L.J., Milburn, S.C. & Hershey, J.W.B. (1982) Immuno-
chemical characterization of mammalian protein synthesis initia-
tion factors. Biochemistry 21, 4206–4212.
24. Kaufman, R.J. (1997) DNA transfection to study translational

control in mammalian cells. Methods 11, 361–370.
25. Kozak, M. (1987) At least 6 nucleotides preceding the AUG ini-
tiator codon enhance translation in mammalian cells. J. Mol. Biol.
196, 947–950.
26. Boissel, J P., Kasper, T.J. & Bunn, H.F. (1988) Cotranslational
Amino-terminal Processing of Cytosolic Proteins. J. Biol. Chem.
263, 8443–8449.
27. Huang, S., Elliott, R.C., Liu, P.S., Koduri, R.K., Weickmann,
J.L.,Lee,J.H.,Blair,L.C.,Ghosh-Dastidar,P.,Bradshaw,R.A.,
Bryan, K.M., Einarson, B., Kendall, R.L., Kolacz, K.H. & Saito,
K. (1987) Specificity of cotranslational amino-terminal processing
of proteins in yeast. Biochemistry 26, 8242–8246.
28. UCSC Genome Broswer (2002), (genome,
Human; assembley, Nov.2002; position, AF085358).
29. Zhang, Q H. Ye, M., Wu, X Y., Ren, S X., Zhao, M., Zhao,
C J., Fu, G., Shen, Y., Fan, H Y., Lu, G., Zhong, M., Xu, X R.,
Han,Z G.,Zhang,J W.,Tao,J.,Huang,Q H.,Zhou,J.,
Hu, G X., Gu, J., Chen, S J. & Chen, Z. (2000) Cloning and
functional analysis of cDNAs with open reading frames for 300
previously undefined genes expressed in CD34+ hematgopoietic
stem/progenitor cells. Genome Res. 10, 1546–1560.
30. Duncan, R., Etchison, D. & Hershey, J.W.B. (1983) Protein
synthesis eukaryotic initiaiton factors 4A and 4B are not altered by
poliovirus infection of HeLa cells. J. Biol. Chem. 258, 7236–7239.
31. Scholler, J.K. & Kanner, S.B. (1997) The human p167 gene
encodes a unique structural protein that contains a centrosomin A
homology and associates with a multicomponent complex. DNA
Cell. Biol. 16, 515–531.
32. Phan, L., Zhang, X., Asano, K., Anderson, J., Vornlocher, H P.,
Greenberg, J.R., Qin, J. & Hinnebusch, A.G. (1998)

Identification of a translation initiation factor 3 (eIF3) core
complex, conserved in yeast and mammals, that interacts with
eIF5. Mol. Cell. Biol. 18, 4935–4946.
33. Burks, E.A., Bezerra, P.P., Le, H., Gallie, D.R. & Browning, K.S.
(2001) Plant initiation factor 3 subunit composition resembles
mammalian initiation factor 3 and has a novel subunit. J. Biol.
Chem. 276, 2122–2131.
Ó FEBS 2003 Characterization of eIF3k (Eur. J. Biochem. 270) 4139