515

Biochemical and genetic studies are revealing a network ofinteractions between eukaryotic translation initiation factors,further refining or redefining perceptions of their function. Thenotion of translated mRNA as a ‘closed-loop’ has gainedsupport from the identification of physical and functionalinteractions between the two mRNA ends and their associatedfactors. Translational control mechanisms are beginning tounravel in sufficient detail to pinpoint the affected step in theinitiation pathway.

AddressesGene Expression Programme, European Molecular Biology Laboratory,Meyerhofstrasse 1, D-69117 Heidelberg, Germany*e-mail: preiss@embl-heidelberg.de†e-mail: hentze@embl-heidelberg.de

Current Opinion in Genetics & Development 1999, 9:515–521

0959-437X/99/$ — see front matter © 1999 Elsevier Science Ltd. All rights reserved.

AbbreviationsACE adenylation control elementBRE Bruno response elementCPE cytoplasmic polyadenylation elementCPEB CPE-binding proteineIF4F eukaryotic initiation factor 4FIRE iron-responsive elementIRP iron regulatory proteinPABP poly(A)-binding proteinPAIP PABP-interacting proteinSLBP stem-loop binding proteinTIF translation initiation factortPA tissue-type plasminogen activatorUTR untranslated regionXFGFR1 Xenopus FGF receptor-1

IntroductionThe translation of eukaryotic mRNAs is a highly compet-itive and tightly regulated step in gene expression.Control is most commonly exerted on it at the rate-limit-ing initiation phase. Factors involved in translationinitiation have been known for some time and their bio-chemical activities were used to build the salient modelfor cap-dependent initiation of translation [1]. Accordingto this model, the 5′ cap structure (m7GpppN) attractsthe eukaryotic initiation factor 4F (eIF4F) complex to themRNA (Figure 1). eIF4F is a heteromultimeric complexcomposed of the cap-binding protein eIF4E, the RNA-dependent ATPase eIF4A, and the modular factoreIF4G. The small (40S) ribosomal subunit binds to the 5′end of an mRNA as a 43S complex including eIF3, a mul-tisubunit factor, and the ternary complex of eIF2 withGTP and Met-tRNAi. eIF4G can contact eIF3 whereaseIF4A, stimulated by eIF4B, is thought to unwind sec-ondary structure in the 5′ UTR. The resulting 48Scomplex then advances through the initiation cycle. A lat-eral movement of the 43S complex along the mRNA,

termed scanning, is the most plausible explanation for afaithful recognition of the (usually) first AUG triplet asthe start codon. Codon–anticodon base-pairing with Met-tRNAi triggers eIF2-bound GTP hydrolysis, catalysed byeIF5; it has been thought that this causes dissociation ofinitiation factors and large (60S) subunit joining to formthe 80S ribosome.

The scope of our review is to summarise the recent progressin establishing a network of interactions between initiationfactors and the mRNA substrate. We discuss physiologicalexamples of specific translational control with an emphasison the interface between regulatory mechanisms and theinitiation pathway. As space is limiting, the focus of thisreview could not be extended to include interesting devel-opments in the area of signal transduction to translationinitiation factors or progress in viral mRNA translation.

Initiation factor functionCircularisation of mRNAThe 3′ poly(A) tail of mRNAs also participates in transla-tion initiation [2] and acts synergistically with the capstructure [3]. In the mid nineties, a picture of the underly-ing mechanism began to emerge: an interaction betweenthe poly(A)-binding protein Pab1p and both eIF4G homo-logues was discovered in yeast [4,5]. It was shown that thisinteraction is required for poly(A)-dependent recruitmentof 43S complexes to the mRNA [5–7], leading to thehypothesis that simultaneous binding of eIF4E and Pab1pto eIF4G and concomitant circularisation of the mRNAtemplate may form the biochemical basis of translationalsynergy between cap structure and poly(A) tail (Figure 1a).

Functional and physical evidence for such a circular con-formation of the mRNA has now been obtained. In a yeastcell-free translation system, poly(A)-tail-promoted transla-tion lacks an inherent directionality towards the mRNA 5′end and can occur internally. The cap structure serves toanchor ribosome recruitment to the 5′ end, probably viaPab1p/eIF4G/eIF4E interactions. Translational synergybetween the mRNA end modifications originates from asuperior efficiency of this combined mode to attract initia-tion complexes, which becomes apparent when mRNAsare competing for limiting initiation factors as is usually thecase [8,9••]. Using recombinant yeast factors eIF4E,eIF4G and Pab1p and a synthetic capped and polyadeny-lated RNA, circular complexes were directly observed byatomic force microscopy [10••].

Poly(A)-binding protein (PABP) binds to eIF4G in plant andmammalian systems too. Wheat germ PABP interacts withboth cognate eIF4F isoforms and eIF4B, and a bindingdeterminant was found in the amino-terminal region of

From factors to mechanisms: translation and translationalcontrol in eukaryotesThomas Preiss* and Matthias W Hentze†

iso-eIF4G. In a purified system composed of cap analogue,(iso-)eIF4F, PABP and polyribo(A), all binding interactionswere found to positively reinforce each other [11,12]. Theinteraction between the mammalian proteins involves a pre-viously unrecognised amino-terminal region of eIF4G, asdemonstrated by two independent studies. First, eIF4G wasfound by the two hybrid assay to interact with the rotaviralNSP3 protein which itself is associated with the very 3′ endof the (nonadenylated) rotaviral mRNAs [13••]. NSP3 andPABP association with eIF4F were found to be mutually

exclusive in Rotavirus infected cells. Second, the presenceof additional sequence at the amino-terminus of the highlysimilar eIF4GII [14], prompted the cloning of additional 5′cDNA sequence of eIF4GI. Co-precipitation assays in vivoand in vitro demonstrated an interaction of PABP with theamino-terminal extension of both eIF4Gs [15•], which mapsto the region that binds NSP3. These findings uncover anew viral strategy to usurp the cellular translation machin-ery: NSP3 evicts PABP from the eIF4F complex and acts asits functional replacement specifically for viral mRNAs.

516 Differentiation and gene regulation

Figure 1

Initiation factor interactions on circularmRNAs. The mRNA substrate is shown withthe 5′ and 3′ ends in close proximity.Translation initiation is divided operationallyinto three phases: (a) binding of the 40Sribosomal subunit with associated initiationfactors near the 5′ end, (b) ‘scanning’ alongthe 5′ UTR, and (c) start codon recognitionand 60S subunit joining. A putative recyclingof translational components on thecircularised mRNA is also indicated. The'enlargements' highlight interactions ofinitiation factors on or near the 40S ribosomalsubunit. Factors shown in green contain RNA-binding motifs and/or have reported RNA-binding properties. Names of canonicalinitiation factors are given omitting the usual'eIF'-prefix. In (c), the core subunits of eIF3are shown in orange. The nomenclature usedhere for yeast eIF3 subunits is based on theirapparent molecular mass by SDS-PAGE (i.e.[25•,28]); p110 is synonymous to Tif32p, p93to Nip1p, p90 to Prt1p, p39 to Tif34p, andp33 to Tif35p. Lines with two arrowheadsindicate additional interactions for whichavailable information does not permit theassessment of evolutionary conservation. (Seetext for further details and references.)

RNA circularisation is thus emerging as a central feature ofeukaryotic translation initiation. The functional signifi-cance of circularisation may be to increase translationalfidelity by ensuring that only properly processed andexported mRNA molecules are translated [10••].Translational synergy could arise as a consequence of coop-erative binding events along the path to 43S recruitment.This could, for instance, affect the formation of theeIF4F/PABP bridging complex or the binding of the 40Ssubunit itself. Another interesting and not mutually exclu-sive possibility is that components of the translationmachinery (i.e. the ribosome) may be recycled on the samemRNA molecule after termination, aided by the proximityof the termini. This inherently attractive concept dividesthe translational life of an mRNA into the first and subse-quent rounds of translation which may differ in theirinitiation factor requirement. This issue was addressed byemploying poliovirus protease 2A as a tool [16•], whichcleaves eIF4GI and II carboxy-terminally from theeIF4E/PABP binding regions to block translation [17].Under conditions of near complete cleavage of eIF4GIand II, cellular mRNAs continued to be translated forhours whereas, for instance, a newly transcribed reportermRNA was very poorly translated [16•]. This points to adifferential requirement for physically intact eIF4G duringfirst and subsequent rounds of translation. The 2A-cleav-age of eIF4G does not separate the eIF4E and PABPbinding regions and additional interlocking interactionsmay exist which help to bridge between the mRNA ends[10••] and terminating ribosomes once the first translationcomplex has been assembled.

Interestingly, an eIF4G-related protein called PAIP, forPABP-interacting protein, has been discovered which bindseIF4A but not eIF4E [18•]. PAIP may provide an additionalcontact between the poly(A) tail and the initiation machinery.

Ribosomal scanningThe scanning model for initiation in eukaryotes canexplain the general adherence of eukaryotic mRNAs to thefirst start codon rule, and is compatible with pseudocircu-lar mRNA substrates. Its popularity notwithstanding,direct evidence for scanning or insight into its mechanismhas proven hard to obtain [19].

A reconstitution approach using highly pure or recombi-nant components and a toe-printing assay (i.e. the arrest ofprimer extension by reverse transcriptase at the leadingedge of mRNA-associated complexes), may now begin toshed light onto this question [20••]. Addition of 40S sub-units, ATP, eIF2, 3, 4A, 4B, and 4F led to formation of acap-proximal complex I (leading edge 21–24 nucleotidefrom the 5′ end), which could not be chased towards theinitiator codon. Inclusion of eIF1 and 1A, however, led toformation of an authentic 48S complex, centred over theAUG (complex II, leading edge 15–17 nucleotides 3′ of theAUG). eIF1 and 1A act synergistically in this assay and,when added to complex I formed in their absence, require

a cycle of dissociation/reassociation to assemble into com-plex II. Thus, although complex I is not a direct precursorof complex II, eIF1 and 1A are the first initiation factorsintimately linked to the positioning of the small ribosomalsubunit at the translation initiation codon (Figure 1b).They might positively affect the processivity of scanningand/or form part of a clamp that closes over the mRNA-binding cleft of the 40S subunit [21,22•].

Interaction networksThe yeast SUI (suppressor of initiator codon mutations)mutants which allow initiation at a UUG codon are affect-ed in either eIF1, 2, or 5 [23]. eIF1 itself was found tointeract with the Nip1p subunit of yeast eIF3 (known asp110 in human) [22•,24•], which also binds eIF5 [25•]. Inturn, eIF5 interacts with eIF2 [26], thus outlining anarrangement of factors involved in the final step of ‘scan-ning’: the recognition of the initiation codon (Figure 1).Completion of the cDNA cloning of all 10 subunits ofmammalian eIF3 [27] allowed an exhaustive search foridentifiable homologues in the yeast genome. Five yeastgenes — PRT1, TIF32, NIP1, TIF34, and TIF35 — werefound, and all five encoded proteins are part of a function-al eIF3 complex [25•]; thus, they are considered toconstitute a conserved core of eIF3 (Figure 1c). The com-position of yeast eIF3 preparations varies with thepurification procedure [28,29] and three additionalpolypeptides — encoded by the genes SUI1 (a yeasthomologue of eIF1), GCD10, and TIF31 — have beenimplicated as either more loosely bound eIF3 subunits oras factors interacting with the complex. A complex involv-ing eIF3 with its several RNA-binding subunits is anadditional candidate for the above mentioned clamp thatcould embed the mRNA in the 40S-binding cleft [19,22•].Information on which of the eIF3 subunits interact witheIF4G is eagerly awaited.

eIF4E is phosphorylated at a conserved Ser 209 in responseto external stimuli such as growth factors, hormones andmitogens, resulting in enhanced cap-binding [30,31]. On thebasis of crystal structure data, phosphorylated Ser 209 canform a salt bridge with Lys 159, which may serve as a anoth-er type of clamp over the bound mRNA [32]. A MAP kinaseactivated protein kinase, Mnk1 was found independently,by two-hybrid studies [33•] and by co-precipitation assays invivo and in vitro [34•], to bind the carboxyl-terminus ofmammalian eIF4GI&II as well as the eIF4G-related proteinp97 (or NAT-1/DAP-5). eIF4G can thus provide a dockingsite for efficient and specific phosphorylation of eIF4Eassembled into eIF4F, whereas p97, which cannot bindeIF4E, could play a regulatory role by sequestering Mnk1.

It came as a major surprise that the bacterial initiation fac-tor 2 (IF2) was found to be conserved throughoutevolution, with homologues identified in archae, yeasts,mammals, zebrafish, and maize [35••,36•]. S. cerevisiaestrains lacking the FUN12 gene, which encodes yeast(y)IF2, show a drastic shift from polysomes to inactive

Translation and translational control Preiss and Hentze 517

monosomes. Translation extracts prepared from suchmutant strains can be rescued by addition of recombinantyIF2 [35••]. Human (h) or archeal (a) IF2 can substitute foryIF2 [36•]. eIF2 and yIF2 can both stimulate first peptidebond formation in a heterologous model system, but theyare not functional isozymes: eIF2 and yIF2 cannot sup-press each other’s mutations. hIF2 and yIF2 depend onintact GTP-binding domains for full activity. Collectively,these findings force a rethinking of the previous modelthat eIF2 is the functional analogue of IF2. They also sug-gest the existence of a second GTP hydrolysis step ineukaryotic translation initiation. Further study of IF2 func-tion should reveal unexpected parallels betweenprokaryotic and eukaryotic protein synthesis.

Translational regulationImproved understanding of the translation initiation path-way provides a much stronger basis for understandingtranslational control mechanisms. Translation can be regu-lated globally by affecting initiation factor function (e.g.through phosphorylation of eIF4E, the 4E-binding pro-teins [31], and eIF2α, or by eIF4G cleavage [17]). Suchglobal control affects many cellular mRNAs and thus totalprotein synthesis. In addition, translational regulation ofspecific mRNAs has been studied and is generally exertedthrough cis-acting elements on the mRNA. Recentprogress in the latter area is discussed below.

Iron regulatory proteinsIron regulatory proteins (IRPs) inhibit ferritin translationby binding to specific 5′ UTR sites, the iron-responsiveelements (IREs). Using a novel procedure to purify initia-tion complexes, the regulatory mechanism was analysed inthe rabbit reticulocyte lysate system. The cap-proximalbinding of IRP-1 to the IRE was shown to permit assem-bly of eIF4F and eIF4B on the mRNA but to impede theassociation of eIF3 and the 40S ribosomal subunit [37••].As one might predict, moving the repressor complex fur-ther away from the cap allows the recruitment ofmammalian 43S complexes to the mRNA. After temporarystalling, they then progress linearly through the IRE to theinitiator codon [38]. The apparent scanning arrest is prob-ably overcome by active displacement of IRP-1, becausethe passive dissociation rate is slow. Regarding the mecha-nism of initiation, these results also indicate that eIF4Fcan bind to the cap without prior association with 40S-bound eIF3. The data also indicate that the RNA-helicaseactivity of eIF4A/4B alone does not suffice for IRP-1 dis-placement [37••], a property that the 48S complex seemsto possess [38]; it is thus uncertain whether eIF4A/4Bcould serve as the sole ‘motor’ of scanning.

Developmental control — masking and polyadenylationThe 3′ UTR of mRNAs is a common location of cis-actingelements that control their localisation and/or translation,particularly during oogenesis and early embryogenesis. Atfirst glance, involvement of 3′ UTR elements in transla-tional regulation seems counter-intuitive. However, thisappears in a different light when one considers the role ofthe poly(A) tail and the importance of 5′/3′ end interactions.

Two originally quite separate notions of translational con-trol mechanisms in development, mRNA masking andcytoplasmic polyadenylation, have begun to converge.Different possible explanations for an interdependencebetween masking and regulated polyadenylation havebeen proposed [39]: First, a 3′ UTR element repressestranslation by maintaining a short poly(A) tail; second,polyadenylation is required to inactivate a 3′ UTR repres-sor element/protein, or third, 3′ UTR element(s) controlpolyadenylation and translation independently — an elon-gated tail may then reinforce translational activity(Figure 2). The latter may act through counteracting adefault deadenylation pathway that sets in during matura-tion [40•] and/or to boost the translational efficiency of themRNA in the competitive cellular environment [8].Several recent studies described below highlight the inter-relation between masking and polyadenylation in differentsystems but do not (yet) allow one to propose a unifyingmodel of the underlying mechanism(s).

Tissue-type plasminogen activator (tPA) mRNA is dor-mant in primary mouse oocytes and has a short poly(A) tailof ~50 adenosines. Meiotic maturation leads to translation-al activation and cytoplasmic polyadenylation. Both, theinitial poly(A) shortening and the subsequent elongation

518 Differentiation and gene regulation

Figure 2

mRNA masking and cytoplasmic adenylation control. CPE/maskingelements in the 3′ UTR may affect translation by direct regulation oftranslation, and/or control of polyadenylation status. Arrows denoteprocesses which may lead to repression (white) and/or activation oftranslation (black). Cap ribose methylation stimulated by poly(A)elongation is indicated as ‘–CH3’, and ‘A(n+x)’ denotes a poly(A) tailundergoing dynamic changes in length.

–CH3

CPE

/ACE

60S 40S

40S

40S

AAUAAAA(n+x) cap

Current Opinion in Genetics & Development

require a UA-rich element, the ACE or CPE (adenylationcontrol or cytoplasmic polyadenylation element) in the tPA3′ UTR. Cytoplasmic polyadenylation additionallyrequires the ‘nuclear’ polyadenylation hexanucleotide[39]. Injection of ACE competitor transcripts into oocytesnow shows that the short poly(A) tail in primary oocytesalone is insufficient for dormancy and that a titrateable fac-tor needs to bind to the ACE. A candidate for this maskingfunction is an approx. 80 kDa protein which cross-links tothe ACE. Removal of the factor, however, only leads toawakening of the tPA mRNA in primary oocytes if at leasta short poly(A) tail is present. It also induces a more rapidpolyadenylation of the mRNA during maturation [40•].Thus, the findings with tPA mRNA are most consistentwith the third scenario listed above.

The mRNA for Xenopus FGF receptor-1 (XFGFR–1) isrepressed in immature oocytes as a result of the translationinhibitory motif in its 3′ UTR. Oocyte maturation inducespolyadenylation and activation of translation. Whenmaturation is induced directly, by injection of maturation-promoting factor, XFGFR translation occurs at normallevels concomitant with only minimal poly(A) tail elonga-tion on the mRNA. Treatment of c-mos-depleted oocyteswith progesterone, by contrast, does not induce progres-sion through meiosis or XFGFR translation but the mRNAgets efficiently polyadenylated [41•]. Here, masking aswell as unmasking can be uncoupled from polyadenylationand unmasking can occur in the presence of a short tail.

Translational activation of Xenopus cyclin B1 mRNA duringoocyte maturation also coincides with CPE-dependentpolyadenylation. Injection of excess CPE competitor RNAinto immature oocytes causes translational unmasking ofthe endogenous mRNA in the absence of poly(A) tail elon-gation [42•], consistent with the third hypothesis above. Bycontrast, reporter mRNAs bearing the cyclin B1 3′ UTRwith functional CPEs, a tail of 30 adenosines, but withoutthe hexanucleotide polyadenylation motif are not activat-ed during oocyte maturation. This latter result is moresuggestive of the second explanation. The CPE-bindingprotein (CPEB) is essential for cytoplasmic polyadenyla-tion. Additionally, its binding activity correlates stronglywith translational repression. These findings raise the pos-sibility that CPEB could function both positively andnegatively at different stages of development.Interestingly, an EMCV internal ribosome entry site-dri-ven reporter mRNA is much less sensitive toCPE-mediated repression. Internal ribosome entry site-driven translation, because of differential initiation factorrequirement, may not depend on interactions that are nec-essary for the cellular cyclin B1 transcript. It has also beensuggested that polyadenylation may function by stimulat-ing cap-ribose methylation [43].

Recent studies with maternal surf clam mRNAs also high-light the interplay between masking and polyadenylation.p82 was originally found to bind a U-rich masking element

in the 3′ UTR of ribonucleotide reductase mRNA and tomediate translational repression in vitro; recently it wasidentified as clam CPEB [44•,45•]. The ribonucleotidereductase mRNA 3′ UTR contains six CPE-like motifs,including two in the masking element, which together withthe hexanucleotide motif support p82-mediated polyadeny-lation in egg extract. Fertilisation leads to translationalactivation of masked mRNAs and p82 is phosphorylated bya cdc2-like kinase before being degraded.

The onset of spermatogenesis in the Caenorhabditis eleganshermaphrodite depends on translational repression of tra-2mRNA. Tra-2 repression is mediated via two 28 nucleotidedirect repeat elements (DREs or TGEs [tra-2 and GLI ele-ments]) in the 3′ UTR [46]. A yeast three-hybrid-screenidentified GLD-1, a germline-specific member of theSTAR-family of proteins, as a specific TGE-binding factor.GLD-1 is a component of the complex that forms onTGEs with worm extracts. Translational repression medi-ated through the TGEs by GLD-1 was demonstrated in ayeast in vitro system and depends on a functional KH-domain in the recombinant protein. Ectopic expression ofGLD-1 and reporter constructs in somatic cells causes spe-cific repression in vivo [47•]. Intriguingly, bothTGE-mediated control and the STAR protein family areevolutionarily conserved.

In contrast to most other cellular mRNAs, metazoan his-tone mRNAs lack a poly(A) tail. They end with aconserved stem-loop that fulfils many of the functions ofthe poly(A) tail. It is essential for translation [48], andinteracts with SLBP (stem-loop binding protein) in mam-malian somatic cells. SLBP participates in pre-mRNAprocessing [49] and is a component of polyribosomal his-tone mRNPs. Xenopus oocytes possess two SLBPs. SLBP-2is oocyte-specific, present at high levels in early oogenesisand degraded during early embryogenesis. SLBP-1 is thehomologue of mammalian SLBP and displays a reciprocaltemporal expression pattern [50•]. This suggests the possi-bility of a masking phenomenon also for thisnonadenylated mRNA, namely that exchange of SLBP-2for -1 on the 3′ stem-loop could activate histone mRNAtranslation at oocyte maturation.

Developmental control — mRNA localisationand translationNormal progression through oogenesis and embryogenesisin Drosophila requires coupling between translational con-trol and mRNA localisation to achieve proper temporal andspatial protein expression [51]. For instance, translation ofmaternal oskar mRNA is silenced during transport to theposterior pole of the oocyte and until Oskar protein isrequired. These processes require the 3′ UTR of oskarmRNA and do not involve noticeable changes in mRNApoly(A) tail length. The Bruno protein recognises repeatedconserved 3′ UTR sequences, the Bruno response ele-ments (BREs). Bruno prevents premature translation andcolocalises with oskar mRNA [52]. Oskar mRNA is

Translation and translational control Preiss and Hentze 519

translated as two functional isoforms by alternative startcodon usage on its mRNA and an element between the twoAUGs activates translation from both codons exclusively onlocalised mRNA [53••]. This region is a derepressor ratherthan an activator element, because it is only required whenthe 3′ BREs are active. Two proteins, p50 and p68 (distinctfrom Bruno), interact with the 5′ element and p50 alsobinds to the 3′ BRE, apparently simultaneously withBruno. These observations demonstrate a highly complexmechanism involving a growing list of factors to achievelocalisation-dependent derepression of oskar mRNA trans-lation. The functional interactions between 5′ UTRderepressor and 3′ UTR repressor elements again reflectsthe recurrent theme of end-to-end communication.

Conclusions and future directionsResearch on the translation initiation pathway is progressingtowards a more complete picture of molecular interactionsduring its different phases. In particular, the concept of 5′and 3′ end interactions opens up new possibilities to addresspoly(A)-tail-dependent and poly(A)-tail-independentmeans of translational regulation by specific mRNA-bindingproteins in a more informed context.

AcknowledgementsWe thank A Hinnebusch, N Gunkel and the members of the Hentze lab forcomments on the manuscript. P Riedinger is acknowledged for help withthe figures.

References and recommended readingPapers of particular interest, published within the annual period of review,have been highlighted as:

• of special interest••of outstanding interest

1. Merrick WC, Hershey JWB: The pathway and mechanism ofeukaryotic protein synthesis. In Translational Control. Edited byHershey JWB, Mathews MB, Sonenberg N. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory Press; 1996:31-69.

2. Jacobson A: Poly(A) metabolism and translation: the closed-loopmodel. In Translational Control. Edited by Hershey JWB, MathewsMB, Sonenberg N. Cold Spring Harbor, NY: Cold Spring HarborLaboratory Press; 1996:451-480.

3. Gallie DR: The cap and poly(A) tail function synergistically toregulate mRNA translational efficiency. Genes Dev 1991,5:2108-2116.

4. Tarun SZ Jr, Sachs AB: Association of the yeast poly(A) tail bindingprotein with translation initiation factor eIF-4G. EMBO J 1996,15:7168-7177.

5. Tarun SZ Jr, Wells SE, Deardorff JA, Sachs AB: Translation initiationfactor eIF4G mediates in vitro poly(A) tail-dependent translation.Proc Natl Acad Sci USA 1997, 94:9046-9051.

6. Tarun SZ Jr, Sachs AB: A common function for mRNA 5¢ and 3¢ endsin translation initiation in yeast. Genes Dev 1995, 9:2997-3007.

7. Kessler SH, Sachs AB: RNA recognition motif 2 of yeast Pab1p isrequired for its functional interaction with eukaryotic translationinitiation factor 4G. Mol Cell Biol 1998, 18:51-57.

8. Preiss T, Muckenthaler M, Hentze MW: Poly(A)-tail-promotedtranslation in yeast: implications for translational control. RNA1998, 4:1321-1331.

9. Preiss T, Hentze MW: Dual function of the messenger RNA cap•• structure in poly(A)-tail-promoted translation in yeast. Nature

1998, 392:516-520.We provide evidence for a functional cooperativity between the cap andpoly(A) tail affecting both quality and quantity of translation products.

Translational synergy is explained as a superior efficiency of jointly promotedinitiation by the cap structure and the poly(A) tail under competitive conditions.

10. Wells SE, Hillner PE, Vale RD, Sachs AB: Circularization of mRNA•• by eukaryotic translation initiation factors. Mol Cell 1998, 2:135-140.Direct demonstration by atomic force microscopy that the eIF4E/4G/Pab1pinteraction forms the core of a bridge between the mRNA cap structure andpoly(A) tail, and hence can mediate mRNA circularisation.

11. Le H, Tanguay RL, Balasta ML, Wei CC, Browning KS, Metz AM,Goss DJ, Gallie DR: Translation initiation factors eIF-iso4G andeIF-4B interact with the poly(A)-binding protein and increase itsRNA binding activity. J Biol Chem 1997, 272:16247-16255.

12. Wei CC, Balasta ML, Ren J, Goss DJ: Wheat germ poly(A) bindingprotein enhances the binding affinity of eukaryotic initiation factor4F and (iso)4F for cap analogues. Biochemistry 1998, 37:1910-1916.

13. Piron M, Vende P, Cohen J, Poncet D: Rotavirus RNA-binding•• protein NSP3 interacts with eIF4GI and evicts the poly(A) binding

protein from eIF4F. EMBO J 1998, 17:5811-5821.Discovery of a new viral strategy to reprogram cellular translation: NSP3 pro-tein binds specifically to rotaviral mRNA 3′ ends and displaces PABP fromeIF4F. Evidence is provided for a PABP/eIF4G interaction in mammalian cells.

14. Gradi A, Imataka H, Svitkin YV, Rom E, Raught B, Morino S,Sonenberg N: A novel functional human eukaryotic translationinitiation factor 4G. Mol Cell Biol 1998, 18:334-342.

15. Imataka H, Gradi A, Sonenberg N: A newly identified N-terminal• amino acid sequence of human eIF4G binds poly(A)-binding

protein and functions in poly(A)-dependent translation. EMBO J1998, 17:7480-7489.

Co-immunoprecipitation analysis identified a 29 amino acid stretch at theamino terminus of human eIF4G I and II as the PABP-binding site. Thus,together with [13••], this paper extends the concept of the eIF4E/4G/Pab1pbridging interactions to mammalian translation.

16. Novoa I, Carrasco L: Cleavage of eukaryotic translation initiation• factor 4G by exogenously added hybrid proteins containing

poliovirus 2Apro in HeLa cells: effects on gene expression.Mol Cell Biol 1999, 19:2445-2454.

Possible differences between the first and subsequent rounds of translationinitiation on the same mRNA molecule are so far largely unexplored. Thisstudy takes a first step in this direction by showing a differential requirementfor physically intact eIF4G.

17. Gradi A, Svitkin YV, Imataka H, Sonenberg N: Proteolysis of humaneukaryotic translation initiation factor eIF4GII, but not eIF4GI,coincides with the shutoff of host protein synthesis after poliovirusinfection. Proc Natl Acad Sci USA 1998, 95:11089-11094.

18. Craig AW, Haghighat A, Yu AT, Sonenberg N: Interaction of• polyadenylate-binding protein with the eIF4G homologue PAIP

enhances translation. Nature 1998, 392:520-523.This paper describes PAIP, a new human protein with homology to the cen-tral portion of eIF4G. This adds complexity to mammalian translation initia-tion models: PAIP interacts with PABP and eIF4A, but not with eIF4E.

19. Jackson RJ: A comparative view of initiation site selectionmechanisms. In Translational Control. Edited by Hershey JWB,Mathews MB, Sonenberg N. Cold Spring Harbor, NY: Cold SpringHarbor Laboratory Press; 1996:71-112.

20. Pestova TV, Borukhov SI, Hellen CU: Eukaryotic ribosomes require•• initiation factors 1 and 1A to locate initiation codons. Nature 1998,

394:854-859.Initiation reactions assembled from highly purified or recombinant componentsare combined elegantly with toe-printing analysis. This demonstrates that the'Cinderella' factors eIF1/1A are essential initiation factors that are absolutelyrequired for positioning of 43S complexes at the start codon in vitro

21. Pestova TV, Hellen CU: Ribosome recruitment and scanning:what’s new? Trends Biochem Sci 1999, 24:85-87.

22. Fletcher CM, Pestova TV, Hellen CU, Wagner G: Structure• and interactions of the translation initiation factor eIF1. EMBO J

1999, 18:2631-2637The solution structure of eIF1 was determined by NMR spectroscopy.Availability of structural information for this interesting factor (see [20••]) willhelp us understand its mode of action

23. Huang HK, Yoon H, Hannig EM, Donahue TF: GTP hydrolysiscontrols stringent selection of the AUG start codon duringtranslation initiation in Saccharomyces cerevisiae. Genes Dev1997, 11:2396-2413.

520 Differentiation and gene regulation

24. Asano K, Phan L, Anderson J, Hinnebusch AG: Complex formation• by all five homologues of mammalian translation initiation factor

3 subunits from yeast Saccharomyces cerevisiae. J Biol Chem1998, 273:18573-18585.

See annotation [25•].

25. Phan L, Zhang X, Asano K, Anderson J, Vornlocher HP, Greenberg JR,• Qin J, Hinnebusch AG: Identification of a translation initiation

factor 3 (eIF3) core complex, conserved in yeast and mammals,that interacts with eIF5. Mol Cell Biol 1998, 18:4935-4946.

Together with [24•], this study identifies an eIF3 core complex conservedbetween yeast and mammals and charts a map of intersubunit interactionsand contacts with other initiation factors.

26. Das S, Maiti T, Das K, Maitra U: Specific interaction of eukaryotictranslation initiation factor 5 (eIF5) with the beta-subunit of eIF2.J Biol Chem 1997, 272:31712-31718.

27. Block KL, Vornlocher HP, Hershey JW: Characterization of cDNAsencoding the p44 and p35 subunits of human translation initiationfactor eIF3. J Biol Chem 1998, 273:31901-31908.

28. Naranda T, MacMillan SE, Hershey JW: Purified yeast translationalinitiation factor eIF-3 is an RNA-binding protein complex thatcontains the PRT1 protein. J Biol Chem 1994, 269:32286-32292.

29. Danaie P, Wittmer B, Altmann M, Trachsel H: Isolation of a proteincomplex containing translation initiation factor Prt1 fromSaccharomyces cerevisiae. J Biol Chem 1995, 270:4288-4292.

30. Wang X, Flynn A, Waskiewicz AJ, Webb BL, Vries RG, Baines IA,Cooper JA, Proud CG: The phosphorylation of eukaryotic initiationfactor eIF4E in response to phorbol esters, cell stresses, andcytokines is mediated by distinct MAP kinase pathways. J BiolChem 1998, 273:9373-9377.

31. Sonenberg N, Gingras AC: The mRNA 5¢ cap-binding protein eIF4Eand control of cell growth. Curr Opin Cell Biol 1998, 10:268-275.

32. Marcotrigiano J, Gingras AC, Sonenberg N, Burley SK: Cocrystalstructure of the messenger RNA 5¢ cap-binding protein (eIF4E)bound to 7-methyl-GDP. Cell 1997, 89:951-961.

33. Waskiewicz AJ, Johnson JC, Penn B, Mahalingam M, Kimball SR,• Cooper JA: Phosphorylation of the cap-binding protein eukaryotic

translation initiation factor 4E by protein kinase Mnk1 in vivo. MolCell Biol 1999, 19:1871-1880.

See annotation [34•].

34. Pyronnet S, Imataka H, Gingras AC, Fukunaga R, Hunter T,• Sonenberg N: Human eukaryotic translation initiation factor 4G

(eIF4G) recruits Mnk1 to phosphorylate eIF4E. EMBO J 1999,18:270-279.

This paper and [33•] identify a physical link between mitogen- and stress-activated signalling and translation initiation. Mnk1 docks onto eIF4G tophosphorylate bound eIF4E.

35. Choi SK, Lee JH, Zoll WL, Merrick WC, Dever TE: Promotion of met•• Met-tRNAi

Met binding to ribosomes by yIF2, a bacterial IF2homolog in yeast. Science 1998, 280:1757-1760.

In addition to highlighting an unexpected universal conservation of a bacter-ial and a yeast initiation factor, this report and [36•]. indicate the existence ofan additional GTP-hydrolysis step in eukaryotic initiation.

36. Lee JH, Choi SK, Roll-Mecak A, Burley SK, Dever TE: Universal• conservation in translation initiation revealed by human and

archaeal homologs of bacterial translation initiation factor IF2.Proc Natl Acad Sci USA 1999, 96:4342-4347.

See annotation [35••].

37. Muckenthaler M, Gray NK, Hentze MW: IRP-1 binding to ferritin•• mRNA prevents the recruitment of the small ribosomal subunit by

the cap-binding complex eIF4F. Mol Cell 1998, 2:383-388.This paper describes a new method to specifically capture and purify trans-lation initiation complexes formed on a given mRNA. Analysis of the compo-sition of such complexes reveals the mechanism of translational control by a5′ UTR-binding protein at 'initiation factor resolution'.

38. Paraskeva E, Gray NK, Schlager B, Wehr K, Hentze MW: Ribosomalpausing and scanning arrest as mechanisms of translationalregulation from cap-distal iron-responsive elements. Mol Cell Biol1999, 19:807-816.

39. Wickens M, Kimble J, Strickland S: Translational control ofdevelopmental decisions. In Translational Control. Edited by

Hershey JWB, Mathews MB, Sonenberg N. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory Press; 1996:411-450.

40. Stutz A, Conne B, Huarte J, Gubler P, Volkel V, Flandin P, Vassalli JD:• Masking, unmasking, and regulated polyadenylation cooperate in

the translational control of a dormant mRNA in mouse oocytes.Genes Dev 1998, 12:2535-2548.

The complex interdependence between translational masking and control ofcytoplasmic polyadenylation is elegantly demonstrated for tissue-type plas-minogen activator mRNA.

41. Culp PA, Musci TJ: Translational activation and cytoplasmic• polyadenylation of FGF receptor-1 are independently regulated

during Xenopus oocyte maturation. Dev Biol 1998, 193:63-76.The role of a 3′ UTR motif in the appropriate timing of Xenopus FGF-recep-tor 1 translation during meiotic maturation is highlighted. The study providesan example of an apparent lack of a relationship between translational acti-vation and polyadenylation.

42. de Moor CH, Richter JD: Cytoplasmic polyadenylation elements• mediate masking and unmasking of cyclin B1 mRNA. EMBO J

1999, 18:2294-2303.Here, an involvement of 3′ UTR CPEs in both translational masking and unmask-ing of cyclin B1 mRNA during Xenopus oocyte maturation is demonstrated.

43. Kuge H, Richter JD: Cytoplasmic 3¢ poly(A) addition induces 5¢ capribose methylation: implications for translational control ofmaternal mRNA. EMBO J 1995, 14:6301-6310.

44. Walker J, Minshall N, Hake L, Richter J, Standart N: The clam 3¢ UTR• masking element-binding protein p82 is a member of the CPEB

family. RNA 1999, 5:14-26.See annotation [45•].

45. Minshall N, Walker J, Dale M, Standart N: Dual roles of p82, the• clam CPEB homolog, in cytoplasmic polyadenylation and

translational masking. RNA 1999, 5:27-38.This paper and [44•] illustrate a functional linkage between masking/unmask-ing and cytoplasmic polyadenylation.

46. Jan E, Yoon JW, Walterhouse D, Iannaccone P, Goodwin EB:Conservation of the C. elegans tra-2 3¢UTR translational control.EMBO J 1997, 16:6301-6313.

47. Jan E, Motzny CK, Graves LE, Goodwin EB: The STAR protein, • GLD 1, is a translational regulator of sexual identity in

Caenorhabditis elegans. EMBO J 1999, 18:258-269.Translational repression through 3′ UTR elements is mediated by the KH-domain protein GLD-1, a member of the STAR- family of proteins. A com-prehensive study providing both in vivo and biochemical evidences.

48. Gallie DR, Lewis NJ, Marzluff WF: The histone 3¢-terminal stem-loop is necessary for translation in Chinese hamster ovary cells.Nucleic Acids Res 1996, 24:1954-1962.

49. Martin F, Schaller A, Eglite S, Schumperli D, Muller B: The gene forhistone RNA hairpin binding protein is located on humanchromosome 4 and encodes a novel type of RNA binding protein.EMBO J 1997, 16:769-778.

50. Wang ZF, Ingledue TC, Dominski Z, Sanchez R, Marzluff WF: Two• Xenopus proteins that bind the 3¢ end of histone mRNA:

implications for translational control of histone synthesis duringoogenesis. Mol Cell Biol 1999, 19:835-845.

This study demonstrates that Xenopus oocytes have two SLBPs whose lev-els are inversely regulated during oogenesis. The authors suggest that anexchange of SLBP isoforms on the histone mRNA 3′ ends regulates theirtranslational activation.

51. Macdonald PM, Smibert CA: Translational regulation of maternalmRNAs. Curr Opin Genet Dev 1996, 6:403-407.

52. Kim-Ha J, Kerr K, Macdonald PM: Translational regulation of oskarmRNA by bruno, an ovarian RNA-binding protein, is essential. Cell1995, 81:403-412.

53. Gunkel N, Yano T, Markussen FH, Olsen LC, Ephrussi A:•• Localization-dependent translation requires a functional

interaction between the 5¢ and 3¢ ends of oskar mRNA. Genes Dev1998, 12:1652-1664.

5′ derepressor and 3′ UTR repressor elements cooperate to achieve spa-tially restricted oskar mRNA translation during Drosophila oogenesis. This isthe first detailed report that identifies an active role of a 5′ element in medi-ating derepression from 3′ UTR silencing.

Translation and translational control Preiss and Hentze 521