Changes in Ribosomal Binding Activity of eIF3 Correlate withIncreased Translation Rates during Activation of T Lymphocytes*

Received for publication, December 15, 2004, and in revised form, May 24, 2005Published, JBC Papers in Press, June 9, 2005, DOI 10.1074/jbc.M414129200

Suzanne Miyamoto‡§¶, Purvi Patel‡, and John W. B. Hershey‡

From the ‡Department of Biochemistry and Molecular Medicine, School of Medicine, University of California-Davis,Davis, California 95616 and the §Division of Hematology/Oncology, Cancer Center, University of California-Davis,Sacramento, California 95817

The rate of protein synthesis in quiescent peripheralblood T lymphocytes increases dramatically following mi-togenic activation. The stimulation of translation is due toan increase in the rate of initiation caused by the regula-tion of initiation factor activities. Here, we focus on eIF3,a large multiprotein complex that plays a central role inthe formation of the 40 S initiation complex. Using su-crose density gradient centrifugation to analyze ribosomecomplexes, we find that most eIF3 is not bound to 40 Sribosomal subunits in unactivated T lymphocytes but be-comes ribosome-bound following activation. Immunoblotanalyses of sucrose gradient fractions for individual eIF3subunits show that the small eIF3j subunit is unassoci-ated with the eIF3 complex in quiescent T lymphocytes,but upon activation joins the other eIF3 subunits andbinds 40 S ribosomal subunits. Because eIF3j has beenshown to be required for eIF3 binding to 40 S ribosomes invitro, the results suggest that mitogenic stimulation of Tlymphocytes leads to an activation of eIF3j, thereby ena-bling eIF3 to bind to the larger ribosome-free eIF3 sub-unit complex, and then to the 40 S ribosomes. The associ-ation of eIF3j with the other eIF3 subunits appears to beinhibited by rapamycin, suggesting a mechanism that liesdownstream from the mammalian target of rapamycinkinase. This association requires ionomycin togetherwith a phorbol ester, which also suggests that calciumsignaling is involved. We conclude that the complex for-mation of eIF3 and its association with the ribosomesmight contribute to increased translation rates during Tlymphocyte activation.

Primary, mature peripheral blood T lymphocytes are in anatural, quiescent (G0) resting state until challenged by anappropriate mitogenic stimulus that prompts a growth andproliferative immune response. This growth and proliferativeresponse is highly dependent on increased synthesis of proteinsfrom pre-existing and newly generated mRNAs and occursbefore new ribosomal RNA synthesis (1, 2). The resting Tlymphocyte in culture maintains its quiescent state, exhibitinga low protein synthesis rate even in the presence of nutrientsand serum growth factors in the culture media. When primaryT lymphocytes are treated in culture with mitogenic stimu-

lants, proteins synthesis rates dramatically increase. Thisoverall increase in protein synthesis is due to an increase in therate of translation (1–3) caused by an increase in the initiationrate (2, 4–7); the rate of polypeptide elongation on the ribosomedoes not appear to change with stimulation (2).

It is suggested that the mechanism responsible for increasedtranslation is complex and involves a number of regulatedinitiation steps (7). These include the phosphorylation state ofeIF2�1 (7, 8) and increased eIF2B exchange activity (7, 9), bothof which affect the level of the ternary complex (eIF2�GTP�Met-tRNAi

Met), an intermediate in the binding of the initiator me-thionyl-tRNA to the 40 S ribosomal subunit (10–12). Also ex-amined as potential regulatory mechanisms in lymphocytes areeIF4E and 4E-BP phosphorylation (6, 8, 13). Phosphorylationof eIF4E was not found to be regulatory, but phosphorylation of4E-BP1 and the release of eIF4E appear to play a regulatoryrole in increased initiation of protein synthesis in lymphocytes.However, as suggested above, regulation at several initiationsteps likely is involved. We now examine eIF3, the large mul-tisubunit complex involved in the formation of the 40 S preini-tiation complex, as a possible regulatory target during theactivation of lymphocytes.

eIF3 is the largest of the mammalian initiation factors and isimplicated in a number of steps in the initiation pathway. Itaffects ribosome dissociation/re-association, promotes or stabi-lizes ternary complex binding to 40 S subunits, helps positionmRNA on the 40 S ribosome through its interaction witheIF4G, and may contribute to AUG recognition by affectingeIF5 stimulation of eIF2 GTPase activity (reviewed in Refs. 12and 14). eIF3 consists of 12 non-identical subunits that rangein size from 28 to 170 kDa and are named eIF3a (p170), eIF3b(p116), eIF3c (p110), eIF3d (p66), eIF3e (p48), eIF3f (p47),eIF3g (p44), eIF3h (p40), eIF3i (p36), eIF3j (p35), eIF3k (p28),and eIF3l (p69) (15). Characterization of eIF3 in mammaliancells has been hampered by its large size and complexity, andonly recently have the cDNAs been cloned and sequenced for allof the subunits (16). The functions of the individual subunitsare not yet well established. In the yeast Saccharomyces cer-evisiae, eIF3 appears to consist of a core complex of only fivesubunits (eIF3a, -b, -c, -g, and -i), plus a non-stoichiometricsubunit, eIF3j. In mammalian cells, the corresponding ho-mologs also may constitute a “core” complex to which the othermammalian subunits bind and regulate eIF3 activity. YeasteIF3 interacts with eIF1, eIF2�, and eIF5 to form a multifactor* This work was supported by National Institutes of Health Grant

GM22135 (to J. W. B. H.) and by the Children Miracle Network, UCDavis Health Systems (to S. M.). The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

¶ To whom correspondence should be addressed: Division of Hema-tology/Oncology, UC Davis Cancer Center, 4501 X St., Rm. 3016, Sac-ramento, CA 95817. Tel.: 916-734-3769; Fax: 916-734-6415; E-mail:smiyamot@ucdavis.edu.

1 The abbreviations used are: eIF, eukaryotic initiation factor; PMA,phorbol myristate acetate (or ‘P’); I, ionomycin; PI3K, phosphatidyl-inositol 3-kinase; MAPK, mitogen-activated protein kinase; 4E-BP1,eIF4E-binding protein 1; I�P, ionomycin plus PMA; mTOR, mamma-lian target of rapamycin; MOPS, 4-morpholinepropanesulfonic acid;PHA, phytohemagglutinin.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 31, Issue of August 5, pp. 28251–28264, 2005© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 28251

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

complex (17, 18). Mammalian eIF3 also binds eIF1 and eIF5(19–21) and, in addition, binds eIF4B (22) and eIF4G (23).Together with its RNA-binding activity and its ability to bindstably to the 40 S ribosome, mammalian eIF3 may play anorganizing role on the surface of the 40 S ribosomal subunit.

Several lines of evidence suggest that eIF3 activity is regu-lated during formation of the 40 S preinitiation complex. Asmall subunit, eIF3j, is cleaved by caspases in apoptosis (24,25), reducing its ability to promote eIF3 binding to the 40 Ssubunit (26). Apoptosis also leads to eIF3f phosphorylation byan activated, truncated form of cyclin-dependent kinase 11,thereby inhibiting protein synthesis (27). eIF3, along witheIF4E, eIF4G, and small, but not large ribosomal subunits, issequestered in stress granules as part of the stress response incells, and these stress granules disassemble, and eIF3 returnsto its original subcellular localization during recovery from thestress situation (28, 29). The interferon-induced protein P56binds to eIF3e and inhibits translation in vitro by blocking theinteraction of eIF3 with the eIF2�GTP�Met-tRNAi

Met ternarycomplex (30). The cellular level of eIF3a is reported to vary,with low levels negatively affecting the translation of ribonu-cleotide reductase M2 mRNA (31). Finally, numerous eIF3subunits are implicated in cancer (32), suggesting that regula-tion of eIF3 activity may be important in cell growth control.

Because peripheral blood lymphocytes are in a natural qui-escent state, with low protein synthesis (translation) rates,they are an excellent system to study the regulation of trans-lation. It is sometimes difficult to perform growth and prolifer-ation studies in already proliferating, often transformed celllines because: 1) the normal regulatory pathways may be mu-tated; 2) the cells need to be synchronized with drugs; and/or 3)the cells need to be serum-starved to reduce translation ratesand then re-stimulated with serum or growth factors, leadingto low translation rate (2- to 3-fold) increases. Frequent prob-lems are high basal translation rates of serum-deprived cellsand/or failure to recover when re-fed with serum.

Two signals are required for optimum activation of T lym-phocytes in nature. The interaction of the T cell receptor witha major histocompatibility complex-peptide presented by anaccessory cell such as the macrophage provides the specificantigen-restricted signal. The second signal is providedthrough interaction of the CD28 molecule with its cognateligand, the B7 family of proteins. This second antigen-unre-stricted signal provides a co-stimulatory signal necessary forfull T cell activation. TCR engagement with the major histo-compatibility complex-peptide without B7 co-stimulation re-sults in T cell clonal anergy (33). Optimal activation of Tlymphocytes can be mimicked in cell culture with anti-CD3 andanti-CD28 and results in a sustained increase in protein syn-thesis rate (34). The combination of ionomycin (I) and thephorbol ester, phorbol myristate acetate (PMA or P), alsocauses a rapid rise in protein synthesis rates and the onset ofcell proliferation (4, 9, 35), comparable to anti-CD3/28. In con-trast, PMA by itself (PMA alone) results in less induction ofprotein synthesis and no proliferation (9), suggesting that inionomycin plus PMA (I�P)-activated T lymphocytes the secondsignal, provided by ionomycin, likely replaces the need for theanti-CD28 co-stimulation.

Prior studies in lymphocytes identified signaling pathwaysthat promote the overall increase in global translation ratesafter activation of T lymphocytes with I�P (9). Translationrates induced by I�P are greater than those induced by PMAalone, suggesting that, additional signals, likely Ca2�-acti-vated, are responsible for the continued rate increase. Trans-lation rates are more sensitive to treatment with PI3K andmTOR inhibitors than to MAPK inhibitors, suggesting that the

PI3K and mTOR pathways are more crucial for the translationrate increase (9). Currently, the only translation initiationfactors affected by the PI3K or mTOR inhibitors rapamycin andwortmannin are eIF4B (36) and eIF4E, the latter through thephosphorylation of 4E-BP1 (6, 9, 13). However, translationrates are sensitive to the combination of rapamycin and wort-mannin, suggesting that these agents not only inhibit 4E-BP1phosphorylation but also might independently target othertranslation initiation events. We already know that eIF4Ephosphorylation and eIF2B exchange activity are not inhibitedby these agents (9), and we cannot rule out a currently “unde-scribed” activity downstream of mTOR that might affect trans-lation in T lymphocytes.

The objectives of this study are to investigate if eIF3 subunitexpression and eIF3 activity change in I�P- or PMA-activatedT lymphocytes, elucidate how eIF3 interacts with other trans-lational components to form the 40 S preinitiation complex, anddetermine if eIF3-associated translation initiation events areaffected by rapamycin or wortmannin. We report the novelobservation that most eIF3 in the unactivated lymphocyte isnot associated with 40 S ribosomal subunits, but becomes al-most completely associated after 24 h with activation condi-tions that promote proliferation (I�P) and not with conditionsthat do not (PMA alone). We also find that eIF3j is not associ-ated with the eIF3 complex but joins the complex followingactivation. Because eIF3j is required for eIF3 binding to 40 Sribosomes in vitro (26), the results suggest that eIF3j associa-tion with eIF3 is regulated early in the activation response.Thus generation of a complete eIF3 complex, after activation,leads to eIF3–40 S ribosome binding and 40 S preinitiationcomplex formation, which are events inhibited by rapamycinand wortmannin.

MATERIALS AND METHODS

Cell culture reagents were obtained from Fisher; fetal bovine serumfrom Gemini; ionomycin and PMA from Sigma; nylon wool was fromRobbins Scientific; goat anti-human eIF3 antibodies and cDNAs to eIF3subunits have been described previously (16, 37–43).

Preparation and Activation of T Lymphocytes—Human T lympho-cytes were obtained from normal human peripheral blood from wholeblood buffy coats, discarded tonsils, or leukocyte filters and furtherenriched for T lymphocytes by nylon wool purification (44). Pig T lym-phocytes were isolated from pig whole blood, obtained from the UCDavis Meat Laboratory. Pig blood was mixed with 10% (final concen-tration) citrate-acetate to prevent coagulation and 1% dextran to helpsettle red blood cell. The upper layer above the settled red blood cells,containing the white blood cells, was removed and subjected to Ficoll-Hypaque separation (lymphocyte separation media from Cellgro). Thebuffy coat layer, containing a mixture of T and B lymphocytes, granu-locytes, and monocytes, was removed, and the cell mixture was passedthrough sterile nylon-wool columns to enrich for the T lymphocytepopulation as previously described (9). Cells were grown in RPMI 1640containing 10% heat-inactivated fetal bovine serum, 100 units/ml pen-icillin, 100 units/ml penicillin/streptomycin, and 100 �g/ml glutamine,at 37 °C, 5% CO2. T lymphocytes were activated with either 0.25 �M

ionomycin plus 10 ng/ml PMA (I�P), or 10 ng/ml PMA (PMA alone). Forsome experiments, T lymphocytes were isolated from human tonsilsusing these same methods and were found to respond in a similarmanner as peripheral blood lymphocytes to I�P or PMA (45).

Immunoblot or Immunoprecipitation of eIF3 and eIF3 Subunits—Whole cell extracts were prepared from T lymphocytes, activated withI�P or PMA (see above), and lysed in SDS-PAGE loading buffer or lysedin 0.4% Nonidet P-40 in Tris buffer containing protease inhibitors.Proteins were resolved by 10% SDS-PAGE and transferred to an Im-mobilon (Millipore) membrane. The blots were blocked with 1% gelatin(fish, Sigma) in 10 mM Tris-HCl, 0.15 M NaCl with 0.05% Tween 20,probed with anti-human eIF3 goat antibodies and rabbit anti-goatantibodies conjugated with horseradish peroxidase, and developed withchemiluminescent substrate (Lumiglo, Cell Signaling or Western Light-ening, PerkinElmer Life Sciences). Immunoblots were exposed toKodak X-Omat AR or BioMax film.

Immunoprecipitations were performed by first absorbing the whole

eIF3 Biology during Lymphocyte Activation28252

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

cell extracts with Protein G-Sepharose (Amersham Biosciences), thenthe lymphocyte whole cell extracts were incubated with 0.5 �l of eIF3polyclonal antibody (goat) or 1 �l of eIF3a monoclonal antibody, incu-bated for 1 h, 4 °C in lysis buffer (see above), then 20 �l of washedProtein G-Sepharose was added, incubated 1 h, 4 °C with rotation. Theimmunoprecipitates were washed (wash, 0.05 M Tris, pH 8, 0.5% Non-idet P-40, 0.14 M NaCl), analyzed by 10% SDS-PAGE, followed byimmunoblotting with affinity-purified eIF3j (goat) or eIF3 polyclonalantibody (goat), with ECL (PerkinElmer Life Sciences). RecombinantFLAG-tagged eIF3j was prepared and purified from Escherichia coli.

Northern Blot Analysis—Total RNA was isolated from pig lympho-cytes (1.54 � 108 cells), unactivated or activated with I�P or PMA alone(0–24 h, TriReagent, Molecular Research Center). RNA concentrationwas quantitated (A260/280 nm, Beckman DU640), and 10 �g of RNA ofeach time point was sample separated on a 0.8% formaldehyde agarosegel in MOPS buffer (44), transferred to Nytran membrane (S&S), andcross-linked using UV irradiation (Stratagene). Labeled cDNA probeswere prepared from each eIF3 subunit cDNA (provided by the Hersheylaboratory) by random prime labeling (Invitrogen) and [32P]dUTP (Am-ersham Biosciences) and hybridized to the blots (44). After washing, theblots were exposed to Kodak X-Omat AR film and developed or quan-titated using a PhosphorImager (Amersham Biosciences).

Sucrose Gradient Fractionation of T Lymphocyte Cell Extracts—Tlymphocytes were activated with 0.25 �M ionomycin plus 10 ng/ml PMAor 10 ng/ml PMA alone. After activation, cells were harvested, washed,and lysed with cell lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM MgCl2,0.25 mM KCl, 1 mM dithiothreitol, and 1% Triton X-100) by incubationon ice for 10 min, followed by centrifugation for 10 min (14,000 � g) topellet nuclei, microsomes, and unlysed cells. Supernatants were layeredon top of sucrose gradients (5–45% sucrose in 50 mM HEPES, pH 7.9,100 mM NaCl, 10 mM Mg(OAc)2, and 10 units/ml RNasin, Promega) andcentrifuged for 4 h at 38,000 rpm, 4 °C (Beckman Ti-41 rotor). Gradi-ents were analyzed by using an ISCO fractionator with UV254 nm detec-tor, and 0.5-ml fractions were collected and precipitated with 10%trichloroacetic acid. Trichloroacetic acid precipitates were centrifuged(30 min, 4 °C, 14,000 rpm, Eppendorf centrifuge), washed twice withacetone, and dissolved in SDS loading buffer. The precipitated proteinswere separated by 10% SDS-PAGE, transferred to Immobilon (Milli-pore) membranes, and subjected to Western blot analysis.

RESULTS

The rate of protein synthesis begins to increase in primaryperipheral blood lymphocytes grown in culture about 2–3 hafter the addition of a combination of ionomycin and PMA(I�P) (9, 46). Translation rates continue to increase in thepresence of I�P (8) in a manner similar to stimulation oflymphocytes with phytohemagglutinin (PHA) (47, 48) or a com-bination of anti-CD3 and anti-CD28 (34). In contrast, T lym-phocytes in the presence of PMA alone (PMA) increase theprotein synthesis rate for only the first 8 h, but cannot sustainthis rate increase without the presence of the calcium iono-phore ionomycin (9). The combination of I�P promotes prolif-eration in T lymphocytes, whereas the addition of PMA alonedoes not (49, 50). DNA synthesis generally does not occur until24 h after stimulation of lymphocytes, and cell division takesplace after 48 h of stimulation (51); therefore increased proteinsynthesis occurs before DNA synthesis and is necessary forlymphocyte proliferation activation (1, 52, 53).

Biochemical studies of translational control in lymphocyteshave been conducted using human or porcine (pig) lymphocytes(2, 6, 54–56). In early studies, pig lymphocytes were commonlyused as an alternate source of primary lymphocytes (2, 54–56).Fresh pig blood can be obtained directly following slaughtering,and T lymphocytes are rapidly prepared in large enoughamounts for biochemical studies. Pig T lymphocytes can beactivated with PHA, PMA, or the combination of I�P (2, 8, 9,54–57) in a manner similar to human T lymphocytes (34). Anexample of the increase in translation rate induced by I�P inpig lymphocytes is shown in Fig. 1. Similar to results fromhuman T lymphocytes, protein synthesis rates in pig T lym-phocytes increase 3-fold in the first 8 h of activation by I�P andcontinue to increase thereafter (data not shown) (2).

The cDNAs of all the mammalian eIF3 subunits have beencloned (16, 39–41, 58), and a number of antibodies are nowavailable, which makes it possible to study eIF3 expression andactivity in lymphocytes undergoing activation. A goat anti-serum raised against mammalian eIF3 has been characterizedas recognizing some of the subunit proteins in human (fromHeLa cells) and rabbit eIF3 (59). To determine if this goatpolyclonal antiserum also recognizes pig lymphocyte eIF3, im-munoblot analysis was performed on cell extracts preparedfrom unactivated and I�P-activated pig lymphocytes, and twohuman leukemic cell lines, Jurkat (60) and Kit225 (61). Thepatterns were compared with that of purified human eIF3prepared from HeLa cells (Fig. 2A). The immunoblot showsthat the anti-eIF3 polyclonal goat antibody strongly recognizesthe eIF3a, eIFb/c, and eIF3j subunits in the cell lysates ofunactivated and activated pig and human lymphocytes. Thesame size for each of the recognized pig and human eIF3subunits suggests that pig eIF3 is very similar to human eIF3,and supports the subunit identifications assigned to the pigproteins. Furthermore, a monoclonal antibody to eIF3a reactswith the pig eIF3a protein, as does a polyclonal antiserumspecific for eIF3b (results not shown). An affinity-purified an-tibody was prepared from the goat polyclonal anti-eIF3 anti-serum and used to specifically identify eIF3j in human and pigcell lymphocytes (Fig. 2B). Cell lysates from Jurkat cells andactivated pig lymphocytes, along with purified human eIF3,were immunoblotted with the affinity-purified eIF3j antibody.Only one protein band at �35 kDa is detected in the cell lysatesfrom human and pig lymphocytes, showing the specificity ofthis antibody for eIF3j.

To provide further proof that antibodies in the anti-eIF3serum recognize pig eIF3j, a human recombinant eIF3j wasused to compete with the pig protein for the antibodies. Puri-fied recombinant His-tagged eIF3j migrates slightly moreslowly than native eIF3j during SDS-PAGE (Fig. 2C, left, Coo-massie stain) and is recognized by anti-eIF3 (Fig. 2C, right,Immunoblot). Jurkat, unactivated, and activated pig and hu-man lymphocyte cell lysates were subjected to immunoprecipi-tation with anti-eIF3 antiserum. Immunoprecipitates were an-alyzed by SDS-PAGE and Western blotting with the same eIF3polyclonal antiserum (Fig. 2D, top panel) and with the affinity-purified eIF3j antibody (Fig. 2D, bottom panel). eIF3a andeIF3b is detected in all samples, although these subunits ap-pear to migrate slightly slower in unactivated pig lymphocytes,suggesting the possibility of post-translation modification ofthese eIF3 subunits. The eIF3j affinity antibody detected eIF3j

FIG. 1. Translation rates increase with I�P activation of pig Tlymphocytes. Pig lymphocytes (2 � 106 cells) were activated with I�P,incubated with [35S]methionine for 1 h before harvest, lysed with SDS,and analyzed by 10% trichloroacetic acid precipitation as describedunder “Materials and Methods.” Data points at the times indicated areaverages of duplicate samples with error bars� S.D.

eIF3 Biology during Lymphocyte Activation 28253

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

FIG. 2. Western blot analysis of eIF3 in human lymphocyte cell lines and pig lymphocytes. A, Western blot analysis with the goatanti-human eIF3 antiserum. Jurkat (lane 1), Kit225 (lane 2), unactivated (lane 3) or activated (lane 4) pig lymphocytes, and purified human eIF3(lane 5) were analyzed as described under “Materials and Methods.” eIF3 subunit bands are identified on the right, and molecular mass markersare shown in kilodaltons on the left; B, 10% SDS-PAGE of purified human eIF3 (eIF3), cell extracts from Jurkat cells (hum), or pig lymphocytesactivated with I�P for 24 h (pig), immunoblotted with affinity-purified anti-eIF3j antibody; C, Coomassie blue stain and immunoblot, with theaffinity-purified eIF3j antibody, of purified human eIF3 (HeLa) and recombinant eIF3j; D, immunoprecipitation of Jurkat, unactivated (UA),activated (Act) pig and human lymphocytes with eIF3 antibody and immunoblotting with eIF3 (top) and eIF3j affinity-purified antibody (bottom).Purified eIF3 is included in lane 6; E, blocking of eIF3j recognition by anti-eIF3 antiserum. eIF3 was immunoprecipitated from cell extracts ofJurkat leukemic cells (J) (lanes 1 and 4), unactivated (Un) (lanes 2 and 5) or activated (Act) (lanes 3 and 6) pig T lymphocytes with the anti-eIF3antiserum and extracted with Protein G-Sepharose. The bound protein complexes were eluted in SDS loading buffer and subjected to 10%SDS-PAGE. One set of samples was incubated with the anti-eIF3 antiserum (lanes 1–3 and left part of lane 4), and the other set of samples wasincubated with the anti-eIF3 antiserum plus recombinant eIF3f (left panel, right part of lane 4 and lanes 5 and 6). Arrows identify the eIF3j bandon the right of the figure. The blot was stripped and reprobed with the affinity-purified eIF3 antibody (right panel).

eIF3 Biology during Lymphocyte Activation28254

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

in unactivated and activated pig lymphocytes, and in activatedhuman lymphocytes, but was unable to detect it in unactivatedhuman lymphocytes, and it appears to be below the detectionlimit in the human unactivated lymphocytes, likely due to thesmaller number of human lymphocytes (0.5–1 � 108 cells)compared with pig (2 � 108 cells). The difference in cell numberis because of the greater availability of pig lymphocytes. Fi-nally, duplicate blots of eIF3 immunoprecipitates from Jurkat,unactivated, and activated pig lymphocyte cell lysates wereprepared and incubated with the same eIF3 polyclonal anti-body in the absence or presence of purified recombinant eIF3j(Fig. 2E). eIF3j is detected in the Jurkat and activated piglymphocytes lysates (Fig. 2E, lanes 1 and 3 and part of lane 4,still attached to the first blot), but not detected in the blotincubated with recombinant eIF3j (Fig. 2E, lanes 4–6), norwith purified eIF3 (Fig. 2E, lane 7). The failure to detect eIF3jis not due to a loss of eIF3j protein on the membrane, becausestripping and re-probing with antibodies to eIF3j readily detectthe protein (Fig. 2E, right side of figure). eIF3j in the immuno-precipitates from unactivated pig lymphocytes is not easilydetected by the polyclonal eIF3 antibody, but is weakly de-tected by the affinity-purified eIF3j antibody (Fig. 2E, right,lane 2, also observed in Fig. 2D). In summary, because anti-eIF3j antibodies in the eIF3 antiserum recognize a pig proteinof the same size as human eIF3j, and adding purified recom-binant eIF3j reduces this recognition, we conclude that the

eIF3j band in pig lymphocytes is truly the j-subunit of pig eIF3.eIF3 Protein Levels Remain Constant and Then Increase

16–24 h after Stimulation with I�P—To determine if geneexpression (protein and mRNA) of eIF3 subunits increases withactivation of T lymphocytes and correlates with the increase intranslation rate (Fig. 1), we performed Western blot and North-ern blot analyses of eIF3 subunits in unactivated and activatedlymphocytes. We suspected that eIF3 protein levels might below and limiting in unactivated lymphocytes, and increasedexpression might be needed for the increase in translation rate.Therefore we investigated whether eIF3 protein levels changeafter activation of human or pig T lymphocytes with I�P orPMA alone. Western blot analysis of eIF3 during the first 8 hafter I�P activation shows no change in eIF3 protein level inhuman (Fig. 3A) or pig T lymphocytes (Fig. 3B). A longer I�Pactivation period (24 h) with pig T lymphocytes shows largeincreases in eIF3 at 16 and 24 h (Fig. 3C). In contrast, activa-tion of pig T lymphocytes with PMA alone results in no increasein eIF3 protein during the first 4 h of activation and shows onlya slight increase at 16 and 24 h (Fig. 3D), suggesting that partof the longer, sustained increase in translation rate induced byI�P activation in lymphocytes may be due to increased eIF3protein levels promoted by the presence of ionomycin duringactivation. Because eIF3 levels do not change in the first 8 h,any change in translation due to eIF3 function must be as-cribed to increased specific activity and not to increased pro-

FIG. 3. eIF3 expression does not increase during the early period of T lymphocyte cell activation by I�P, but does increase by16–24 h. T lymphocytes were activated with I�P or PMA alone for 0–24 h as indicated, and cell extracts were prepared and analyzed by 10%SDS-PAGE followed by Western blotting with anti-eIF3 antiserum as described under “Materials and Methods.” Human (A) and pig (B) lymphocytewere activated with I�P; C, a longer time course with pig lymphocytes activated with I�P up to 24 h; D, pig lymphocytes activated with PMA alone,immunoblotted with anti-eIF3; E, time course (0–24 h) of pig lymphocytes activated with I�P, immunoblotted with the affinity eIF3j antibody(middle panel), eIF3 antibody (only eIF3a shown) (top panel) and with anti-�-actin (bottom panel). Molecular mass markers (left, in kilodaltons)and eIF3 subunits (right, with arrows) are indicated.

eIF3 Biology during Lymphocyte Activation 28255

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

tein. In these experiments, two detergent methods of cell lysiswere used, Nonidet P-40 and SDS lysis, with similar results,although cell extracts from unactivated pig T lymphocytes (Fig.3C, lane 1) when prepared with Nonidet P-40 detergent appearto show slightly less eIF3 than when prepared with SDS lysisbuffer (Fig. 3B, lane 1). eIF3j protein expression during lym-phocyte activation was separately examined using the affinityantibody (Fig. 3D, middle panel), and little change in eIF3j oreIF3a protein levels was observed in the 8 h after activation butdid increase after 16 and 24 h of activation with I�P. Incontrast, �-actin protein is highly expressed in unactivatedlymphocytes and appears to slightly decrease during the earlyactivation period (2 h). As was observed earlier (57), levels arehigh in unactivated lymphocytes, may slightly decrease, butactin protein levels appears to remain relatively constant until24 h of activation.

Levels of Most eIF3 Subunit mRNAs Increase following Stim-ulation by I�P, but Are Less Induced by PMA—Because some

eIF3 subunits show large increases in protein level only after16 h of activation with I�P, we wanted to determine if themRNA levels for eIF3 subunits also vary throughout the first24 h of activation. Therefore mRNA levels for eleven eIF3subunits (eIF3a, eIF3b, eIF3c, eIF3d, eIF3e, eIF3f, eIF3g,eIF3h, eIF3i, eIF3j, and eIF3k) were measured by Northernblot analysis of RNA extracted from pig T lymphocytes thatwere unactivated or activated with I�P or PMA alone. TotalRNA was obtained from the lymphocytes, harvested at varioustimes over a 24-h period (see “Materials and Methods” forexperimental details), and separately probed with 32P-labeledDNA that was specific for each eIF3 subunit (Fig. 4A). mRNAsencoding subunits eIF3a, eIF3b, eIF3c, eIF3d, eIF3e, and eIF3ishow low levels in unactivated lymphocytes, which appear toincrease upon stimulation by I�P, but much less so with PMAalone. In contrast, eIF3g and eIF3h mRNA levels are equallyinduced with I�P or PMA, whereas mRNAs for eIF3f and eIF3jare already present at significant levels in unactivated pig T

FIG. 4. Northern blot analyses of eIF3 subunit mRNAs show increased expression for most subunits with I�P activation. Pig Tlymphocytes were activated with I�P or PMA for the times indicated. Total RNA was isolated and 10 �g/sample run on a 0.8% formaldehydeagarose gel, transferred to nylon membrane, and hybridized with specific cDNA probes to each of the eIF3 subunits (see “Materials and Methods”).The blots were analyzed using the Storm PhosphorImager (Amersham Biosciences) or autoradiography. A, Northern blot analysis of individualeIF3 subunits with number of mRNA forms and estimated size (kb) indicated on the right; B, ethidium bromide stain of RNA after transfer to nylonmembrane; C, quantitation (�g/ml) of total RNA per cell measured by absorbance at 260nm and 280nm.

eIF3 Biology during Lymphocyte Activation28256

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

lymphocytes and do not appear to change with activation byeither I�P or PMA. Equal loading of total RNA for each sampleis shown in the ethidium bromide staining for RNA after trans-fer to a nylon membrane and before hybridization with thecDNA probe (Fig. 4B). The same RNA and Northern blot wasused for each eIF3 subunit, through sequential stripping andreprobing of the same blots.

Subunits eIF3b, eIF3d, and eIF3i appear to have multiplemRNA forms. eIF3b has an unusual gene expression pattern,with at least four mRNAs; two mRNAs (3.0 and 5.4 kb) appearin unactivated pig T lymphocytes, whereas the other twomRNAs (2.0 and 8.4 kb) appear to be expressed after activa-tion. Interestingly, the largest 8.4-kb mRNA was expressed at16 and 24 h in T lymphocytes activated by both the I�P andPMA alone, although the transcript was less well expressed incells activated with PMA alone. Subunit eIF3d exhibited amajor 2-kb and a minor 3-kb mRNA, both of which increasedduring longer activation times. The major eIF3i band was onlyabout 0.24 kb, and therefore was too small to be the mRNAencoding eIF3i and is likely an artifact. The minor eIF3i bandat 1.35 kb was therefore the only form of the eIF3i mRNA andwas induced by both I�P and PMA alone.

Overall, gene expression of most of the “core” eIF3 subunits(eIF3a, eIF3b, eIF3c, and eIF3i) appeared to be strongly in-duced by I�P, less so by PMA alone, whereas two (subunits gand j) were barely induced at all, suggesting that they might becoordinately expressed. Importantly, the increases in mRNAlevels for most eIF3 subunits occurred before increases in pro-tein levels. This “lag” in synthesis of eIF3 subunit proteinscorrelates with similar lags already reported for eIF2�(34),eIF4E (8), and eIF2B (62). The lag in new eIF synthesis is likelydue in large part to the low translation rates in early activatedlymphocytes. However, further work is needed to determinewhether or not eIF mRNAs are specifically regulated at thetranslational level in activating T lymphocytes. These observa-tions suggest that transcriptional up-regulation of eIF genesmay be important for the later increase in translation rates(after 8 h) in lymphocytes, but not for the initial 8-h increase.Increased eIF3 subunit gene expression also may be importantfor T lymphocyte proliferation, because I�P and not PMA alonecauses the greatest mRNA increase for most eIF3 subunits.

We also measured the amount of total RNA in our T lympho-cyte samples, activated with I�P or PMA alone, to determine ifan increase in total RNA/cell correlates with the translationrate increases induced by each condition. The results show thattotal RNA (per cell) did not increase in lymphocytes activatedwith I�P or with PMA alone for the first 8 h (Fig. 4B). How-ever, after 24-h activation, there was an almost 4-fold increasewith I�P, but not with PMA alone. Because ribosomal RNA isthe major contributor to total RNA (�80%), the constant totalRNA levels through 8 h of activation suggest that new ribo-somes are not generated, and therefore are not responsible forthe increase in protein synthesis activity. However, new ribo-some biogenesis at later times, together with new eIF synthe-sis, certainly contributes to enhancement of translation ratesat later times. Apparently, ribosome biogenesis following Tlymphocyte activation is dependent on calcium-activated path-ways induced through ionomycin. It is important to note thatthe increases in eIF3 subunit mRNAs shown in Fig. 4A under-measure the mRNA levels in late activated T lymphocytes,because equal amounts of RNA were examined, not RNA froman equal number of cells.

eIF3 Is Not Bound to 40 S Ribosome Subunit in Quiescent TLymphocytes—Previous analyses of eIF3 binding to ribosomesin rabbit reticulocyte lysates (63–65) and lysates of culturedtransformed cells (26) show that eIF3 binds stably to 40 S

ribosomal subunits and is found in the 80 S and polysomeregions of the sucrose gradients used to fractionate the celllysates (63). A minor portion of the eIF3 also is found in the 15S region of the gradient, corresponding to free, unbound eIF3(63). The finding that purified eIF3 binds in vitro to 40 Sribosomal subunits in the absence of all other translationalcomponents (66) reinforces the view that native 40 S ribosomalsubunits contain eIF3. Given that resting primary T lympho-cytes are in a “natural” quiescent Go state without the need tomanipulate them into a “pseudo” quiescent state, we felt thatthe true state of eIF3 in Go and its association with ribosomeswould be best studied in resting lymphocytes. To address thisissue, sucrose gradient centrifugation was performed on ly-sates isolated from quiescent pig T lymphocytes, T lymphocytesactivated with I�P or PMA alone, and exponentially growinghuman Jurkat T leukemic cells (Fig. 5). Fractions from eachsucrose gradient were analyzed by Western immunoblotting asdescribed under “Materials and Methods.” Surprisingly, nearlyall of the eIF3 in unactivated pig lymphocytes is located atabout 15 S between the top of the gradient and the 40 Sribosomal subunit (Fig. 5A, fractions 3–5). The sedimentationposition suggests that in unactivated G0 lymphocytes eIF3exists as a free factor not associated with the 40 S ribosomalsubunit. Only a very small amount of eIF3 was detected infractions corresponding to the 40 S ribosome in the unactivatedpig T lymphocytes (Fig. 5A, fractions 9–11). The absence ofribosome-bound eIF3 correlates with the low rate of translationin these cells.

eIF3 Shifts to the 40 S Subunit upon Stimulation with I�P—When pig T lymphocytes are activated for 24 h with I�P andanalyzed by sucrose gradient centrifugation, nearly all of theeIF3 shifts from the 15 S to the 40 S region of the gradientwhere the small ribosomal subunit sediments (Fig. 5B). A sim-ilar location for eIF3 was seen when Jurkat cells were analyzed(Fig. 5D). However, an extensive shift was not seen following24-h activation with PMA (Fig. 5C). Instead, only very smallquantities of eIF3 were 40 S ribosome-bound, the vast amountremained as free, unbound eIF3. Thus, the shifting of eIF3from a free state to one bound to 40 S ribosomal subunits,correlates with the increased translation rate.

The results obtained with the Jurkat T cell line (Fig. 5D) aresimilar to those obtained with many other cultured cell lines(41). Jurkat T cells are rapidly proliferating lymphocytes withhigh rates of translation (67). They contain two mutations inthe PTEN (phosphatase and tensin homolog) phosphatasegene and are interleukin-2-independent. Results from theWestern blot analysis for eIF3 in sucrose gradient fractions ofJurkat T cell extracts show that all of the eIF3 is associatedwith the 40 S subunit fraction, and none is in the free state(Fig. 5D). A similar result was seen when Jurkat T cells wereserum-starved for 24 or 48 h (results not shown). Clearly,eIF3 is fully capable of binding to ribosomes in proliferatingJurkat T cells.

The kinetics of eIF3 association with the 40 S ribosomalsubunit was investigated next. A time course of I�P stimula-tion of pig T lymphocytes was performed followed by sucrosegradient fractionation of cell extracts and SDS-PAGE andWestern blot analysis of sucrose gradient fractions. Results areshown only for times 0, 2, and 4 h of stimulation (Fig. 6). eIF3did not associate with the 40 S subunit (identified by verticalarrows in Fig. 6 near fraction 10) after 2 h of stimulation.However, a distinct shift of eIF3 proteins toward the 40 Sregion was seen at 4 h, where some of the eIF3 was bound to the40 S subunit. This increased association correlates with theactivation of protein synthesis, where a 2-fold increase wasseen at 4 h (Fig. 1).

eIF3 Biology during Lymphocyte Activation 28257

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

FIG. 5. I�P activation strongly increases the association of eIF3 with the 40 S subunit. Cytoplasmic extracts of pig or humanlymphocytes were fractionated on 5–45% sucrose gradients and analyzed as described under “Materials and Methods.” The left panels are theabsorbance scans (A260 nm) of the gradient, and the right panels are the Western blots with anti-eIF3 antiserum; sedimentation is from left to right.Only fractions 1–18, which include the 80 S monosome fraction were analyzed by Western blot. A, unactivated pig lymphocytes; B, pig lymphocytesactivated with ionomycin (0.25 �M) plus PMA (10 ng/ml) (I�P) for 24 h; C, pig lymphocytes activated with PMA alone (10 ng/ml) (PMA) for 24 h;D, human Jurkat T lymphocytes. Whole cell extract is whole cell extract. eIF3 subunits are identified on the right; molecular mass markers areshown in kilodaltons on the left.

eIF3 Biology during Lymphocyte Activation28258

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

eIF3 Complex Composition Changes following I�P Activa-tion of T Lymphocytes—A careful evaluation of eIF3 subunits inthe Western blots of quiescent pig T lymphocytes shown inFigs. 5 and 6 indicates that eIF3j is not present in, or issubstantially dissociated from, the eIF3 complex located atabout 15 S in the sucrose gradients. eIF3j is predominantly infractions 1 and 2 (Figs. 5 and 6) in quiescent cells, but becomesassociated with the eIF3 complex and ribosomes upon activa-tion. Two other high molecular weight bands also appear to beunassociated with eIF3, occurring in fractions 1 and 2 (Figs. 5and 6). However, these bands do not correspond to eIF3a andeIF3b/c; their appearance in these fractions is not reproducibleand appears to be an artifact of this experiment.

To better evaluate the distribution of eIF3 and eIF3j in thesegradients, additional experiments were performed. By usingantibodies specific for eIF3a, eIF3b, eIF3f, and eIF3j, sucrosegradient fractions from unactivated and activated human Tlymphocytes were immunoblotted (Fig. 7, A and B). eIF3a fromquiescent cells clearly is associated with eIF3, with essentiallyno cross-reacting material in gradient fractions 1 and 2 wherean unassociated subunit might sediment. Confirmation of thelocation of the eIF3 complex is provided by the immunoblotwith anti-eIF3b, a commercial polyclonal antiserum of proven

specificity (Fig. 7A). As shown previously, activation by I�Presults in a shift of eIF3a and eIF3b (representing the eIF3complex) into the 40 S region of the gradient, reflecting eIF3binding to the ribosomal subunit. The presence of smalleramounts of eIF3a and eIF3b in the 40 S region with unacti-vated T lymphocytes probably indicates partial activation ofthese cells. We conclude that eIF3a is an integral part of theeIF3 complex throughout T lymphocyte activation.

The presence of another eIF3 subunit, eIF3f, was also im-munoblotted with a specific antibody (27) (Fig. 7B). eIF3f alsoappears to shift into larger eIF3 complexes with activation oflymphocytes. It appears to form an intermediate complex withthe other eIF3 subunits and then a complex that associateswith the ribosomes in the activated lymphocytes. It is barelydetectable in unactivated lymphocytes and appears to be foundin slightly less dense fractions than eIF3a and eIF3b. Thereason for this difference is under investigation.

The situation with eIF3j is more intriguing. Our recent invitro experiments performed with purified eIF3 complexes andeIF3j indicate that eIF3j is required for eIF3 binding to 40 Sribosomal subunits (26). eIF3 complexes lacking only eIF3jbind with low affinity, whereas addition of eIF3j results in tightbinding. The data from Figs. 5 and 6 suggest that eIF3j is not

FIG. 6. eIF3 binds to 40 S ribosomalsubunits soon after I�P activation.Pig T lymphocytes stimulated with I�Pfor 0 (top panels), 2 (middle panels), and 4(bottom panels) h were lysed with 1% Tri-ton X-100 and subjected to sucrose gradi-ent centrifugation (5–45%) for 4 h at38,000 rpm (SW40). Gradients were frac-tionated and analyzed by immunoblottingwith anti-eIF3 antiserum as describedunder “Materials and Methods.” In theimmunoblots (left panels), proteins bandscorresponding to the eIF3 subunits areidentified on the right. In the A254 nmscans (right panels), the 40 S and 60 Ssubunits and 80 S monosomes are identi-fied by arrows and labeled.

eIF3 Biology during Lymphocyte Activation 28259

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

part of the eIF3 complex in unactivated pig and human lym-phocyte lysates, but rather barely sediments into the gradient.However, after I�P activation, eIF3j is no longer unassociatedfrom eIF3, but appears to become part of the larger eIF3

ribosome-free complex, and is found in gradient regions rich inribosomes. To confirm these observations in human T lympho-cytes, we employed the affinity-purified anti-eIF3j antibodythat was characterized in Fig. 2. As shown in Fig. 7B, eIF3j

FIG. 7. Western blot analysis of ini-tiation factors in sucrose gradientfractions from unactivated and acti-vated human lymphocytes. T lympho-cytes isolated from human tonsils or hu-man peripheral blood were activated for24 h, lysed in 1% Triton X-100 and wholecell extracts (WCEs) were analyzed on a5–45% sucrose gradient as described inthe legend to Fig. 6. A, immunoblots fromunactivated and activated (I�P, 24 h) hu-man tonsil T lymphocytes, probed with amonoclonal antibody to eIF3a, and thensequentially, after stripping (see “Materi-als and Methods”), with a specific anti-body to eIF3b; B, blot of material fromhuman peripheral blood T lymphocytesprobed with the affinity-purified antibodyto eIF3j; C, blots from A were strippedand re-probed with specific antibodies toeIF4G, eIF4E, eIF4B, and eIF2� (see“Material and Methods”). At the top of thefigure are representative scans (A254 nm)from whole cell extracts from unactivatedand I�P-activated human T lymphocytes.Arrows indicate the position of the eIF3subunits. Molecular mass markers are in-dicated on the left, and fraction numbersare below each blot.

eIF3 Biology during Lymphocyte Activation28260

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

from quiescent human T lymphocytes again is found in gradi-ent fractions 1 and 2, well separated from the eIF3 complex(fractions 5 and 6). This experiment was conducted with hu-man lymphocytes after 10 h of activation, so the data mayappear different from the other experiments that were con-ducted with 24 h of activation. Also, we were unable to detecteIF3j in the intermediate ribosome-free complex, likely due tothe low sensitivity of the affinity antibody and the low amountof eIF3j present in these human lymphocyte samples. Similarexperiments conducted with larger numbers of pig lymphocytesdid show the intermediate form of eIF3 with eIF3j (data notshown). We suspect that, upon activation, eIF3j leaves its un-associated free state, becomes associated with the larger eIF3complex, and then this complex binds to the small ribosomalsubunit.

An alternate approach to demonstrating that eIF3j is unas-sociated with the eIF3 complex in quiescent cells is to ask ifeIF3j co-immunoprecipitates with the eIF3 complex. Unacti-vated and I�P-activated human and pig lymphocyte lysateswere co-immunoprecipitated with anti-eIF3a monoclonal anti-bodies, and the adsorbed proteins were analyzed by SDS-PAGEand immunoblotting with the affinity-purified anti-eIF3j anti-body (Fig. 8A). eIF3j is detected in the eIF3 immunoprecipi-tates from control Jurkat T cells and from purified humaneIF3, as expected. Activated pig lymphocytes, and moreweakly, activated human lymphocytes also generate eIF3jbands. However, eIF3j is not detected in the unactivated hu-man lymphocyte immunoprecipitates, although eIF3j can bedetected in cell lysates from unactivated pig lymphocytes (Fig.2A, lane 3); this is likely due to the lower number of humanlymphocytes used in this experiment. The presence of eIF3aand eIF3b/c in these samples was confirmed with further prob-ing of the same blot with the eIF3 antibody (Fig. 8B). The resultsupports the view that eIF3j is unassociated with the eIF3complex but joins it upon lymphocyte activation.

The behavior of eIF3j is consistent with the hypothesis thateIF3j is in an inactive state in quiescent cells and therebycontributes to the low level of initiation of protein synthesis. Itsability to form a complex with other eIF3 subunits and bind to40 S ribosomal subunits appears to be contingent on stimula-tion of T lymphocytes and may therefore play a key role in theinduction of optimal translation rates needed for proliferation.We clearly demonstrate that eIF3j associates with eIF3a inJurkat T leukemic cells and in activated T lymphocytes, butthis association is absent in unactivated T lymphocytes.

The same blots in Fig. 7 (A and B) used for detecting thepresence of individual eIF3 subunits in sucrose gradients werestripped and re-probed with specific antibodies to eIF4G,eIF4E, eIF4B, and eIF2�. eIF4G, eIF4E, eIF4B, and eIF2� inunactivated lymphocytes are found near the top of the gradi-

ent, indicative of their not being a part of either eIF3 or ribo-somal complexes (Fig. 7C). eIF4E in unactivated lymphocytesmostly appears in the least dense fractions and seems to be ina different fraction from eIF4G and eIF4B, thereby suggestingthat it is not associated with eIF4G or eIF4B in unactivatedlymphocytes. Upon activation of lymphocytes (human) withI�P, a portion of each of the three factors sediments somewhatfurther into the gradient, around fractions 4 and 5, but noneappears to bind to 40 S ribosomes. The increase in density/sizeof eIF4E supports the findings of Morley et al. (13), in lympho-cytes. Interestingly, eIF4B shifts to a different fraction fromeIF4E and eIF4G. The failure of these initiation factors to bindto ribosomes contrasts with what is seen with eIF3. The ab-sence of a shift of eIF2� to the 40 S region of the gradient maybe due to the low level of I�P activation at 10 h. These resultsare consistent with earlier experiments of Boal et al. (8), whodemonstrated that there did not appear to be a change in theassociation of the ternary complex with ribosomes during theactivation of human T lymphocytes.

Treatment of T Cells with Rapamycin Inhibits eIF3 Associa-tion with 40 S Ribosomal Subunits—We previously reportedthat inhibitors of important signaling pathways in lymphocytesalso prevent increased translation rates in early I�P-activatedlymphocytes (9). Particularly effective at inhibiting the stimu-lation of protein synthesis are rapamycin, an inhibitor of themTOR pathway, and wortmannin, an inhibitor of the phos-phatidylinositol 3-kinase (PI3K) and Akt pathway. Less effec-tive at preventing increased translation rates are the MAPKinhibitors (9). Therefore, if eIF3 complex formation and asso-ciation with the 40 S ribosomal subunit are necessary for in-creased protein synthesis after activation of lymphocytes,agents that inhibit increased translation rates might also in-hibit this process. We asked whether or not these proteinkinase inhibitors affect eIF3 recruitment to the 40 S ribosomalsubunit. Quiescent human T lymphocytes were pre-treatedwith rapamycin or wortmannin and then activated with I�Pfor 10 h. Cell extracts were fractionated by sucrose gradientcentrifugation and analyzed by Western blotting. As alreadyshown, eIF3 in unactivated cells is mostly in the 15 S region ofthe gradient, with very little bound to 40 S ribosomal subunits(Fig. 9A), whereas following I�P stimulation, a substantialamount of eIF3 becomes bound to the 40 S subunits (Fig. 9B).When rapamycin (Fig. 9C) or wortmannin (Fig. 9D) are pres-ent, the shift of eIF3 onto the 40 S ribosomal subunits isreduced. Therefore, the PI3 kinase pathway operating throughmTOR appears to contribute to the stimulation of eIF3 bindingto 40 S subunits. It is known that rapamycin or wortmannintreatment prevents the phosphorylation of 4E-BP1 in T lym-phocytes activated by I�P, and that this would be expected toprevent translational activation (9). The novel finding reported

FIG. 8. eIF3j co-immunoprecipi-tates with eIF3 in activated, but notunactivated lymphocytes. eIF3 wasimmunoprecipitated from cell extracts ofJurkat T lymphocytes (hum), unactivated(UA pig), and activated (Act pig) (24 hwith I�P) pig T lymphocytes, unactivated(UA hum) and activated (24 h with I�P)(Act hum) human T lymphocytes with amonoclonal antibody to eIF3a. The boundproteins were analyzed by 10% SDS-PAGE and immunoblotting with the affin-ity-purified antibody to eIF3j. A, immuno-blot with affinity antibody to eIF3j, and B,the blot stripped and reprobed with anti-eIF3 for the other eIF3 subunits. eIF3 sub-units are indicated by arrows on the rightside of each figure; molecular mass mark-ers (kDa) are on the left.

eIF3 Biology during Lymphocyte Activation 28261

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

here indicates that these signal transduction pathways alsooperate on eIF3–40 S binding, which are events not expected tobe affected by the phosphorylation levels of 4E-BPs.

Rapamycin and wortmannin do not affect eIF3j identically,however. In wortmannin-treated cells, eIF3j association withthe eIF3 complex caused by I�P occurs normally, because verylittle of the subunit is present in fractions 1 and 2 (Fig. 9D). Incontrast, rapamycin appears to prevent eIF3j from joining theeIF3 complex, resulting in a substantial amount remaining atthe top of the gradient (Fig. 9C). The results indicate that theregulation of eIF3j lies downstream of mTOR signaling. How-ever, it is not yet known if eIF3j itself is the target, or if othereIF3 subunit proteins, other eIFs, or the eIF4E-binding pro-tein, 4EBP1, whose phosphorylation and release of eIF4E isknown to be inhibited by rapamycin, are also involved.

DISCUSSION

Activation of quiescent human or pig T lymphocytes withionomycin plus PMA (I�P) results in a dramatic increase in therate of protein synthesis. A modest increase occurs over the

first 8–12 h that is dependent on pre-existing translationalcomponents. Further increases in translation rate occur up to24 h post-activation, and such stimulation accompanies thesynthesis of new ribosomes and initiation factors. In the workreported here, we have studied primarily the early steps of theactivation process and have focused on the largest of the initi-ation factors, eIF3, which is a critical component of the preini-tiation complex (12).

The finding that numerous subunits of eIF3 are expressed atsignificantly higher levels during the latter half of the 24 h Tcell activation regimen with I�P is not surprising, because anumber of other initiation factors have been shown to increasein amounts during this period (7, 8, 34, 62). It is surprising,however, that the eIF3 subunits are not all regulated (transcip-tionally) identically, because each is present in the eIF3 com-plex at near-stoichiometric amounts. Only two subunits,namely eIF3f and eIF3j, fail to be induced, but instead appearin quiescent cells at cellular levels similar to those following24-h activation with I�P. It may be pertinent that eIF3f is

FIG. 9. Treatment of T lymphocytes with rapamycin or wortmannin reduces eIF3 binding to 40 S ribosomal subunits. Human Tlymphocyte lysates were subjected to sucrose gradient centrifugation and Western blot analysis with anti-eIF3 antiserum as described under“Materials and Methods” and the legend to Fig. 6. A, unactivated; B, activated with I�P for 24 h; C, cells treated with 10 nM rapamycin and thenactivated with I�P; D, cells treated with 10 �M wortmannin and then activated with I�P. eIF3 subunit bands are identified on the right; molecularmass markers are identified in kilodaltons on the left. WCE, indicates whole cell extract.

eIF3 Biology during Lymphocyte Activation28262

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

regulated in the A375 human melanoma cell line during apo-ptosis, when a constitutive active form of cyclin-dependentkinase 11 phosphorylates the subunit and converts it into aninhibitor of protein synthesis (27). eIF3j also may be a regu-lated subunit, as discussed below.

Examination of eIF3 in T lymphocyte whole cell extractsfractionated by sucrose gradient centrifugation led to the un-expected finding that eIF3 is not bound to 40 S ribosomes inquiescent cell lysates, but rather occurs as a free eIF3 complex.eIF3 in proliferating cells is mainly bound to ribosomes, asshown here in Jurkat cells (Fig. 5D) and in numerous earlierstudies (63, 68). Furthermore, stable eIF3 binding to 40 Ssubunits in vitro does not require any other component ofinitiation (66). It appears that either something is missing ordefective in the eIF3 or 40 S ribosomal subunit that is requiredfor binding, or that one or more components in these complexesis modified to prevent a stable association. Interestingly, Kay etal. (69) previously described the presence of an “inhibitor” inthe post-ribosomal supernatant of unactivated lymphocytesthat could repress translation in rabbit reticulocyte lysates, butthis putative inhibitor has not been characterized. The appar-ent regulation of eIF3j association with eIF3 and the subse-quent binding of eIF3 to the 40 S ribosomal subunit may beaffected by changes in the phosphorylation status of theseproteins. Little is known about eIF3 subunit phosphorylationand the effects on eIF3 activity.

The in vivo binding of eIF3 to 40 S subunits has been exam-ined in detail in the yeast Saccharomyces cerevisiae (70). Dele-tion analysis implicates portions of eIF3a, eIF3c, and eIF3j instabilizing yeast eIF3 binding to 40 S subunits and improvedtranslation (71). Interestingly, deletions that block the forma-tion of the multifactor complex and binding of the ternarycomplex to the 40 S subunit nevertheless allow stable eIF3�40S complexes to form. The results in yeast indicate that eIF3binding to 40 S ribosomal subunits does not require stabiliza-tion by components of the multifactor complex but can occurwith eIF3 alone (although roles for eIF1A and eIF1 are notruled out). In contrast to the yeast system, human eIF3 pos-sessing all of its subunits binds with high affinity to 40 Ssubunits in vitro, but eIF3 lacking only eIF3j fails to bindstably (26). Furthermore, the subcomplex, eIF3bgi, which binds40 S subunits poorly, is stabilized on the 40 S subunit wheneIF3j is added. Because intact eIF3a and eIF3c are present inthe non-binding eIF3 lacking eIF3j, and are absent in thebinding subcomplex, eIF3bgij, it appears that eIF3a and eIF3cdo not confer essential stabilizing elements for mammalianeIF3. Instead, eIF3j may play an essential role in eIF3 bindingto 40 S subunits, as well as a stabilizing role in forming theeIF3 complex.

A role for eIF3j in promoting eIF3 binding to 40 S subunits inactivated T lymphocytes is suggested by the analyses reportedhere. In quiescent cells, eIF3j is not part of eIF3, but is unas-sociated with any large complexes, barely sedimenting into thesucrose gradient (Figs. 5 and 6). Upon activation of T cells,eIF3j now is found associated with eIF3 or 40 S complexes.eIF3j association with eIF3 or 40 S subunits correlates with theability of eIF3 to bind to 40 S ribosomes and also correlateswith the increase in translation rates. The in vitro experimentswith purified eIF3 and eIF3j strongly support the view thateIF3j promotion of eIF3–40 S binding is defective in quiescentT lymphocytes. That the phenomenon is seen upon stimulationwith I�P, but not with PMA alone, suggests that calciumregulation may be involved in “activating” eIF3j.

A role for calcium in promoting eIF3–40 S binding is sup-ported by earlier work in Ehrlich ascites cells (64). Ehrlichascites cells treated with a calcium chelator exhibit an inhibi-

tion of protein synthesis, which is relieved by addition of cal-cium. In lysates of the inhibited cells, eIF3 is no longer boundto 40 S ribosomes, but binds again when calcium is replenished.Thus, calcium appears to regulate eIF3–40 S binding in thesecells. Because ionomycin (a calcium ionophore) is needed toefficiently stimulate eIF3j association with eIF3 and higher 40S complexes in T lymphocytes, we propose that calcium-regu-lated signaling may affect these steps. We postulate that cal-cium signaling alters either eIF3j, eIF3, or 40 S subunits toenable eIF3j to associate with eIF3 or 40 S subunits andthereby promote eIF3 binding to ribosomes. Such regulation inmammalian cells has not been studied, yet it provides an in-triguing area for further experimentation.

REFERENCES

1. Kay, J. E., Ahern, T., and Atkins, M. (1971) Biochim. Biophys. Acta 247,322–334

2. Kay, J. E., Ahern, T., Lindsay, V. J., and Sampson, J. (1975) Biochim. Biophys.Acta 378, 241–250

3. Ahern, T., and Kay, J. E. (1973) Biochim. Biophys. Acta 331, 91–1014. Ahern, T., Sampson, J., and Kay, J. E. (1974) Nature 248, 519–5215. Cooper, H. L., and Braverman, R. (1977) J. Cell Physiol. 93, 213–2256. Morley, S. J., Rau, M., Kay, J. E., and Pain, V. M. (1993) Biochem. Soc. Trans.

21, 397S7. Jedlicka, P., and Panniers, R. (1991) J. Biol. Chem. 266, 15663–156698. Boal, T. R., Chiorini, J. A., Cohen, R. B., Miyamoto, S., Frederickson, R. M.,

Sonenberg, N., and Safer, B. (1993) Biochim. Biophys. Acta 1176, 257–2649. Miyamoto, S., Kimball, S. R., and Safer, B. (2000) Biochim. Biophys. Acta

1494, 28–4210. Pain, V. M. (1996) Eur. J. Biochem. 236, 747–77111. Moldave, K. (1985) Annu. Rev. Biochem. 54, 1109–114912. Hershey, J. W. B., and Merrick, W. C. (2000) in Translational Control of Gene

Expression (Sonenberg, N., Hershey, J. W. B., and Mathews, M. B., eds) 2ndEd., pp. 33–88, Cold Spring Harbor Laboratory Press, Cold Spring Harbor,NY

13. Morley, S. J., and Pain, V. M. (1995) Biochem. J. 312, 627–63514. Chaudhuri, J., Chowdhury, D., and Maitra, U. (1999) J. Biol. Chem. 274,

17975–1798015. Browning, K. S., Gallie, D. R., Hershey, J. W., Hinnebusch, A. G., Maitra, U.,

Merrick, W. C., and Norbury, C. (2001) Trends Biochem. Sci. 26, 28416. Asano, K., Vornlocher, H. P., Richter-Cook, N. J., Merrick, W. C., Hinnebusch,

A. G., and Hershey, J. W. (1997) J. Biol. Chem. 272, 27042–2705217. Phan, L., Schoenfeld, L. W., Valasek, L., Nielsen, K. H., and Hinnebusch, A. G.

(2001) EMBO J. 20, 2954–296518. Asano, K., Phan, L., Anderson, J., and Hinnebusch, A. G. (1998) J. Biol. Chem.

273, 18573–1858519. Das, S., Maiti, T., Das, K., and Maitra, U. (1997) J. Biol. Chem. 272,

31712–3171820. Bandyopadhyay, A., and Maitra, U. (1999) Nucleic Acids Res. 27, 1331–133721. Majumdar, R., Bandyopadhyay, A., and Maitra, U. (2003) J. Biol. Chem. 278,

6580–658722. Methot, N., Song, M. S., and Sonenberg, N. (1996) Mol. Cell Biol. 16,

5328–533423. Imataka, H., and Sonenberg, N. (1997) Mol. Cell Biol. 17, 6940–694724. Bushell, M., Wood, W., Clemens, M. J., and Morley, S. J. (2000) Eur. J. Bio-

chem. 267, 1083–109125. Clemens, M. J., Bushell, M., Jeffrey, I. W., Pain, V. M., and Morley, S. J. (2000)

Cell Death Differ. 7, 603–61526. Fraser, C. S., Lee, J. Y., Mayeur, G. L., Bushell, M., Doudna, J. A., and

Hershey, J. W. (2004) J. Biol. Chem. 279, 8946–895627. Shi, J., Feng, Y., Goulet, A. C., Vaillancourt, R. R., Sachs, N. A., Hershey, J. W.,

and Nelson, M. A. (2003) J. Biol. Chem. 278, 5062–507128. Kedersha, N., Chen, S., Gilks, N., Li, W., Miller, I. J., Stahl, J., and Anderson,

P. (2002) Mol. Biol. Cell 13, 195–21029. Dunand-Sauthier, I., Walker, C., Wilkinson, C., Gordon, C., Crane, R., Nor-

bury, C., and Humphrey, T. (2002) Mol. Biol. Cell 13, 1626–164030. Hui, D. J., Bhasker, C. R., Merrick, W. C., and Sen, G. C. (2003) J. Biol. Chem.

278, 39477–3948231. Dong, Z., Liu, L. H., Han, B., Pincheira, R., and Zhang, J. T. (2004) Oncogene

23, 3790–380132. Hershey, J. W. B., and Miyamoto, S. (2000) in Translational Control of Gene

Expression (Sonenberg, N., Hershey, J. W. B., and Mathews, M. B., eds) pp.637–654, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

33. Lenschow, D. J., Walunas, T. L., and Bluestone, J. A. (1996) Annu. Rev.Immunol. 14, 233–258

34. Mao, X., Green, J. M., Safer, B., Lindsten, T., Frederickson, R. M., Miyamoto,S., Sonenberg, N., and Thompson, C. B. (1992) J. Biol. Chem. 267,20444–20450

35. Jagus, R., and Kay, J. E. (1979) Eur. J. Biochem. 100, 503–51036. Raught, B., Peiretti, F., Gingras, A. C., Livingstone, M., Shahbazian, D.,

Mayeur, G. L., Polakiewicz, R. D., Sonenberg, N., and Hershey, J. W. (2004)EMBO J. 23, 1761–1769

37. Meyer, L. J., Brown-Luedi, M. L., Corbett, S., Tolan, D. R., and Hershey, J. W.(1981) J. Biol. Chem. 256, 351–356

38. Chaudhuri, J., Chakrabarti, A., and Maitra, U. (1997) J. Biol. Chem. 272,30975–30983

39. Block, K. L., Vornlocher, H. P., and Hershey, J. W. (1998) J. Biol. Chem. 273,31901–31908

eIF3 Biology during Lymphocyte Activation 28263

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

40. Methot, N., Rom, E., Olsen, H., and Sonenberg, N. (1997) J. Biol. Chem. 272,1110–1116

41. Asano, K., Kinzy, T. G., Merrick, W. C., and Hershey, J. W. (1997) J. Biol.Chem. 272, 1101–1109

42. Bandyopadhyay, A., and Maitra, U. (1999) Nucleic Acids Res. 27, 1331–133743. Asano, K., Merrick, W. C., and Hershey, J. W. (1997) J. Biol. Chem. 272,

23477–2348044. Miyamoto, S., Chiorini, J. A., Urcelay, E., and Safer, B. (1996) Biochem. J. 315,

791–79845. Carini, C., Pini, C., DiFelice, G., Fattorossi, A., and Fratazzi, C. (1993) Int.

Arch. Allergy Immunol. 101, 31–3846. Grolleau, A., Kaplan, M. J., Hanash, S. M., Beretta, L., and Richardson, B.

(2000) J. Clin. Invest. 106, 1561–156847. Kay, J. E. (1968) Eur. J. Biochem. 4, 225–23248. Kay, J. E. (1969) Exp. Cell Res. 58, 185–18849. Kumagai, N., Benedict, S. H., Mills, G. B., and Gelfand, E. W. (1988) J. Cell

Physiol. 137, 329–33650. Kumagai, N., Benedict, S. H., Mills, G. B., and Gelfand, E. W. (1987) J. Im-

munol. 139, 1393–139951. Ahern, T., and Kay, J. E. (1971) Biochem. J. 125, 73P–74P52. Ahern, T., and Kay, J. E. (1975) Exp. Cell Res. 92, 513–51553. Kay, J. E. (1967) Nature 215, 737–73854. Wallace, D. M., and Kay, J. E. (1981) Biosci. Rep. 1, 539–54555. Wallace, D. M., Jagus, R., Benzie, C. R., and Kay, J. E. (1979) Biochem. J. 184,

277–28256. Kay, J. E., Meehan, R. T., and Benzie, C. R. (1983) Immunol. Lett. 7, 151–15657. Miyamoto, S., and Safer, B. (1999) Biochem. J. 344, 803–81258. Hershey, J. W., Asano, K., Naranda, T., Vornlocher, H. P., Hanachi, P., and

Merrick, W. C. (1996) Biochimie (Paris) 78, 903–90759. Meyer, L. J., Milburn, S. C., and Hershey, J. W. (1982) Biochemistry 21,

4206–421260. Weiss, A., Wiskocil, R. L., and Stobo, J. D. (1984) J. Immunol. 133, 123–12861. Tamori, S., Uchiyama, T., and Uchino, H. (1989) Nippon Ketsueki Gakkai

Zasshi 52, 996–100362. Welsh, G. I., Miyamoto, S., Price, N. T., Safer, B., and Proud, C. G. (1996)

J. Biol. Chem. 271, 11410–1141363. Peterson, D. T., Merrick, W. C., and Safer, B. (1979) J. Biol. Chem. 254,

2509–251664. Kumar, R. V., Wolfman, A., Panniers, R., and Henshaw, E. C. (1989) J. Cell

Biol. 108, 2107–211565. Tolan, D. R., Hershey, J. W., and Traut, R. T. (1983) Biochimie (Paris) 65,

427–43666. Benne, R., and Hershey, J. W. (1976) Proc. Natl. Acad. Sci. U. S. A. 73,

3005–300967. Nishida, J., Shiratsuchi, A., Nadano, D., Sato, T. A., and Nakanishi, Y. (2002)

J. Biochem. (Tokyo) 131, 485–49368. Ayuso-Parilla, M., Henshaw, E. C., and Hirsch, C. A. (1973) J. Biol. Chem. 248,

4386–439369. Kay, J. E., Wallace, D. M., Benzie, C. R., and Jagus, R. (1979) in Cell Biology

and Immunology of Leukocyte Function (Quastel, M. R., ed) pp. 107–114,Academic Press, NY

70. Valasek, L., Nielsen, K. H., Zhang, F., Fekete, C. A., and Hinnebusch, A. G.(2004) Mol. Cell Biol. 24, 9437–9455

71. Valasek, L., Nielsen, K. H., and Hinnebusch, A. G. (2002) EMBO J. 21,5886–5898

eIF3 Biology during Lymphocyte Activation28264

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from

Suzanne Miyamoto, Purvi Patel and John W. B. HersheyTranslation Rates during Activation of T Lymphocytes

Changes in Ribosomal Binding Activity of eIF3 Correlate with Increased

doi: 10.1074/jbc.M414129200 originally published online June 9, 20052005, 280:28251-28264.J. Biol. Chem. 

  10.1074/jbc.M414129200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/280/31/28251.full.html#ref-list-1

This article cites 68 references, 41 of which can be accessed free at

by guest on June 29, 2017http://w

ww

.jbc.org/D

ownloaded from