REPRODUCTION · Translational regulation is important in a wide variety of cell types but may be even more prevalent in germ cells, where periods of transcriptional quiescence necessitate

R
EPRODUCTIONREVIEW
The DAZL and PABP families: RNA-binding proteins withinterrelated roles in translational control in oocytes

Matthew Brook1,2,3, Joel W S Smith2 and Nicola K Gray1,2,3

1MRC Human Reproductive Sciences Unit, Queen’s Medical Research Institute, Centre for Reproductive Biology, 47Little France Crescent, Edinburgh EH16 4TJ, UK, 2MRC Human Genetics Unit, Institute for Genetics and MolecularMedicine, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK and 3School of Clinical Sciences andCommunity Health, University of Edinburgh, The Chancellor’s Building, 49 Little France Crescent,Edinburgh EH16 4SB, UK

Correspondence should be addressed to N K Gray; Email: [email protected]

JWS Smith is now at Edinburgh Research and Innovation Ltd, 1-7 Roxburgh Street, Edinburgh EH8 9TA, UK

Abstract

Gametogenesis is a highly complex process that requires the exquisite temporal, spatial and amplitudinal regulation of gene expression at

multiple levels. Translational regulation is important in a wide variety of cell types but may be even more prevalent in germ cells, where

periods of transcriptional quiescence necessitate the use of post-transcriptional mechanisms to effect changes in gene expression.

Consistent with this, studies in multiple animal models have revealed an essential role for mRNA translation in the establishment and

maintenance of reproductive competence. While studies in humans are less advanced, emerging evidence suggests that translational

regulation plays a similarly important role in human germ cells and fertility. This review highlights specific mechanisms of translational

regulation that play critical roles in oogenesis by activating subsets of mRNAs. These mRNAs are activated in a strictly determined

temporal manner via elements located within their 3 0UTR, which serve as binding sites for trans-acting factors. While we concentrate on

oogenesis, these regulatory events also play important roles during spermatogenesis. In particular, we focus on the deleted in

azoospermia-like (DAZL) family of proteins, recently implicated in the translational control of specific mRNAs in germ cells; their

relationship with the general translation initiation factor poly(A)-binding protein (PABP) and the process of cytoplasmic mRNA

polyadenylation.

Reproduction (2009) 137 595–617

Introduction

Impaired fertility affects w15% of couples worldwidewith a significant proportion of cases being due todefective gametogenesis (Matzuk & Lamb 2008). Inorder to define possible points of therapeutic interven-tion in infertility, it is necessary to have an in-depthunderstanding of the normal processes of gametogenesisand, in particular, to delineate the molecularmechanisms by which gametogenesis is regulated. Inrecent years, a great deal of research has concentrated onthe control of gene expression during gametogenesisand, more specifically, has investigated the importantrole of translational control during oogenesis (Piccioniet al. 2005; see Box 1 and Fig. 1 for an overview ofoogenesis). The limited availability of human ovariantissue and the tractability of organisms such as mice,Xenopus, sheep, cows, Drosophila and Caenorhabditiselegans means that most studies are performed usingthese models, in which many of the key oogenic and

q 2009 Society for Reproduction and Fertility

ISSN 1470–1626 (paper) 1741–7899 (online)

translational control processes are evolutionarily con-served, albeit to varying extents. We primarily discussstudies of 3 0UTR-mediated translational activationduring oogenesis in small laboratory vertebrates suchas Xenopus laevis and mouse.

X. laevis has proved an extremely powerful tool fordissecting the molecular mechanisms and signallingpathways that regulate gene expression during oogen-esis. Indeed many key control mechanisms that operatein mammalian oogenesis were first described usingX. laevis oocytes. One reason for the popularity of theX. laevis model is that the seasonality of X. laevisovulation can be uncoupled and hundreds of oocytes ofdifferent, clearly identifiable, stages can be recoveredfrom a single ovary. These oocytes can be readilymaintained and matured in vitro. Importantly, the largesize of these oocytes make them amenable to micro-injection with exogenous DNAs/RNAs/proteins/drugsand subsequent biochemical assays. Most studies havefocused on translational regulation during oocyte

DOI: 10.1530/REP-08-0524

Online version via www.reproduction-online.org

Downloaded from Bioscientifica.com at 02/22/2021 05:55:30PMvia free access

http://dx.doi.org/10.1530/REP-08-0524

Box 1 A comparative overview of oogenesis (see also Fig. 1).Primordial germ cells (PGCs) originate in the early embryo distinct from the location of the developing gonad. InXenopus laevis, PGCs derive from a cellular subset that inherited oocyte ‘germ-plasm’ during early embryonicasymmetric mitoses (Nieuwkoop & Faber 1994, Sive et al. 1999). Germ-plasm describes a region of mitochondrial-rich ooplasm containing numerous mRNA-rich germinal granules. Human and mouse oocytes have no recognisablegerm-plasm and PGCs do not appear to differentiate from a pre-determined cell population. Rather, mouse PGCsdifferentiate in response to extracellular signals derived from the proximal epiblast or primitive streak of the earlyembryo. PGCs of each species then proliferate and migrate to the developing gonad. PGC migration is completedprior to ovarian sexual differentiation but the germ cells continue to proliferate. In human and mouse ovaries theoogonia subsequently cease to proliferate and enter meiosis, whereas the X. laevis ovary retains a population ofmixed primary oogonia (self-renewing) and nests of secondary oogonia (committed to oocyte differentiation;Nieuwkoop & Faber 1994, Sive et al. 1999). The retention of self-renewing primary oogonia in X. laevis facilitatesseasonal production of large numbers of oocytes, but is fundamentally different from mammalian oogenesis.

The onset of meiosis I in oogonia occurs during embryogenesis in humans and mice and prior to metamorphosis inX. laevis. Human oogonia asynchronously begin to enter meiosis I in early second trimester and mice oogoniasynchronously enter meiosis I at E13.5. A large proportion of human and mouse oogonia die during meiosis I in aprocess called atresia, with most cell death occurring during late-pachytene or dichtyate. The dichtyate arrest(prophase I) of surviving human oogonia is complete by late third trimester and persists for up to 45 years. X. laevisoogonia begin toentermeiosis I from stage55 (wE32) onwardswith a marked increase inentry rate occurringat stage56 (wE38). By stage 62 (wE50) most oogonia have entered diplotene of meiosis I, termed stage I oocytes. Severalthousand X. laevis oocytes then progress through diplotene until many stage VI oocytes are present in the ovary(Nieuwkoop & Faber 1994, Sive et al. 1999).

During diplotene, oocytes synthesise huge amounts of rRNA, tRNAs, mitochondria, proteins and maternalmRNAs to prepare for their post-fertilization biosynthetic requirements until zygotic transcription resumes. Thisoccurs after 1, 2–3 or 12 post-fertilization mitotic divisions in mouse, human and X. laevis respectively (Sive et al.1999, Matova & Cooley 2001, Bownes & Pathirana 2003). Unlike mammals, X. laevis embryos do not receivenutrients placentally and instead oocytes accumulate a nutrient store during diplotene: a process termedvitellogenesis. Liver-secreted vitellogenin accumulates in oocytes until, by stage V, it comprises w80% of theooplasm and accounts for the w4000-fold greater volume of a X. laevis oocyte compared with a mouse oocyte(Nieuwkoop & Faber 1994, Sive et al. 1999). However, although transcription continues during diplotene onlyw2%of the ribosomes in late diplotene X. laevis oocytes are engaged in active translation and only w20% of the mRNAsare translated; indicating that the majority of the mRNA population is translationally silenced.

Folliculogenesis and ovulation differ between mammals and X. laevis, although there are some significantsimilarities. In humans w20 primordial follicles/day become activated, in a poorly understood process, and thendevelop in a hormone/steroid-independent manner for w65 days. During this time the granulosa cells (GCs)proliferate, thecal cells derived from the stroma surround the growing follicle and the enclosed granulosa cells beginto differentiate. After this time GC proliferation becomes FSH-dependent, an antral space develops and the oocyte issurrounded by specialized cumulus granulosa cells (termed a Graafian follicle). Ultimately, only a single follicle issufficiently developed to survive when FSH-levels decline mid-menstrually. It is for this reason that humans areusually mono-ovulatory. Mouse follicles can survive in lower FSH concentrations and, consequently, several folliclesprogress to ovulation.

In mammals, a surge in circulating LH levels stimulates Graafian follicle oocytes to reinitiate meiosis. In X. laevis,the oocyte is surrounded by a single layer of follicle cells which secrete progesterone in response to circulatinggonadotrophins and induce the oocyte to resume meiosis. The completion of meiosis I and progression through tometaphase of meiosis II is termed maturation and is highly conserved between mammals and X. laevis (see Box 2 andFig. 2 for a more detailed overview of oocyte maturation; reviewed in (Schmitt & Nebreda 2002, Vasudevan et al.2006)). Hence, X. laevis data have greatly informed studies of oocyte maturation in mammals.

596 M Brook and others

maturation (see Box 2 and Fig. 2 for an overview ofoocyte maturation) since their transcriptional quiescencemeans that changes in the pattern of protein synthesiscan only be achieved via the activation, repression ordestruction of pre-existing stored mRNAs. Many of thekey regulatory steps in maturation are conserved fromXenopus to human and oocyte maturation in X. laevis

Reproduction (2009) 137 595–617

can be easily scored by the appearance of a white spoton the animal pole.

Mouse models have greatly facilitated the study ofmammalian oogenesis. While the multiovulatory natureand lack of menstruation in mice differs from humans,the mouse ovary and the hormonal regulation ofoogenesis/folliculogenesis are comparable in many key

www.reproduction-online.org


Figure 1 Schematic to illustrate the comparative timing and duration of key events in vertebrate oogenesis. Timings are shown (where known) for thecompletion of the major primordial germ cell (PGC) pre-meiotic events (orange arrows). Due to the more asynchronous nature of the entry of thePGC/oogonia population into meiosis in human and X. laevis, the durations of individual stages of meiosis/folliculogenesis relate to the progress ofan individual oocyte/follicle (yellow arrows). See Box 1 (or Matova & Cooley 2001) for a brief comparative review of vertebrate oogenesis. Xenopusgerm-plasm can be detected throughout early embryogenesis, where it plays an important role in PGC specification, but the PGCs becomeembedded in the dorsal endoderm by stage 40 prior to migration to the site of genital ridge formation (Nieuwkoop & Faber 1994, Sive et al. 1999).A schematic of particular meiotic stages and pictorial representations of X. laevis oogenesis and mammalian folliculogenesis are also shown. GVBD,germinal vesicle breakdown.

DAZL and PABP mediated-translation in oocytes 597

aspects. Their short oestrus and reproductive cycle is alsoadvantageous and the seasonality of reproduction is bredout of most laboratory strains. Additionally, the geneticmanipulation of mice has proved extremely valuable tothe study of both male and female gametogenesis; atechnique that is not as widely applied in the tetraploidX. laevis. However, this advantage is countered by thefact that it is difficult to obtain sufficient numbers ofmouse oocytes to facilitate molecular and biochemicalstudies.

In this review, we will focus on the molecularmechanisms of 3 0UTR mediated translational activationthat play critical roles in oogenesis and highlight thecontributions of the X. laevis and mouse model systemsto the understanding of these regulatory events. Inparticular, we discuss the emerging importance of the


DAZL family of translational activator proteins and theirmolecular and functional interactions with the PABPfamily of translation initiation factors.

Translational regulation

Translational regulation is usually defined as a change inthe rate of protein synthesis without a correspondingchange in mRNA levels. However, this is an over-simplification as translation and mRNA turnover areoften tightly coupled and alterations in translation canhave downstream effects on mRNA stability (Newbury2006). Translation can also be spatially regulated toestablish protein gradients via localization of mRNAsand/or regulatory factors (Huang & Richter 2004).Translational control therefore facilitates the generation

Reproduction (2009) 137 595–617


Box 2 An overview of the major signalling and gene expression events that regulate vertebrate oocytematuration (see also Fig. 2).Maturation is triggered by progesterone in X. laevis, or LH in mammals, causing a rapid inhibition of adenylatecyclase activity (AC) and subsequent decline in cAMP levels, thereby reducing protein kinase A (PKA) activity andinitiating a number of intracellular signalling cascades. The earliest regulatory events in maturation are poorlyunderstood but, in X. laevis, rapid inducer of G2/M progression in oocytes (RINGO)/speedy (SPY) mRNAtranslation is initiated leading to activation of CDK2, aurora A and additional intracellular signalling pathways(the latter is not depicted) (Cheng et al. 2005, Padmanabhan & Richter 2006, Kim & Richter 2007). These events inturn lead to the translational upregulation of mRNAs encoding key meiotic proteins via modulation of theirpoly(A) tail length (including c-Mos, cyclin B1 and Wee1). One of the key events in vertebrate oocyte maturationis the translational upregulation of c-Mos, a serine/threonine kinase that promotes the activation of the MAPkinase (MAPK) pathway (Schmitt & Nebreda 2002). This is upstream of the polyadenylation of a large subset ofmRNAs. In X. laevis, but not mammals, c-Mos is required for complete activation of maturation promotingfactor (MPF), comprising a heterodimer of cyclin B and CDC2, which promotes germinal vesicle membranebreakdown (GVBD) and completion of meiosis I. c-mos null mouse oocytes progress normally through meiosis Ibut exhibit defects during meiosis II, indicating that c-Mos is required for the second meiotic arrest (Gebauer et al.1994). Indeed, in both mammals and X. laevis c-Mos is required for cytostatic factor (CF) activity which serves toinhibit the anaphase promoting complex (APC), thereby maintaining MPF activity and causing meiosis IImetaphase arrest.


of intricate and fine-tuned expression patterns, even inthe absence of ongoing transcription.

Translational regulation can be divided into two broadcategories: global and specific. Global control affectsmost cellular mRNAs in response to a given signal and isoften mediated by changes in the activities of basaltranslation factors. By contrast, specific regulation affectsonly a single or subset of mRNAs in a coordinate manner,normally via sequence elements located within thesemRNAs. Global and specific regulation can occursimultaneously to rapidly reprogram protein synthesis(reviewed in (Spriggs et al. 2008, Yamasaki & Anderson2008)).

Translation can be divided into three basic steps:i) initiation, during which many of the required factorsare recruited; ii) elongation, whereupon an mRNA isdecoded by the ribosome and associated factors andiii) termination, when the de novo synthesizedpolypeptide and the translational machinery arereleased. While each of these steps can be regulated,specific mRNA regulation occurs most frequently at thelevel of initiation.

Eukaryotic mRNAs exit the nucleus with a modified7-methyl guanosine at the 5 0end, termed the capstructure, and a poly(A) tail of w250 nucleotides at the3 0end. These features act as the primary determinants oftranslational initiation efficiency. Cap-dependentinitiation accounts for the majority of cellular translationand although other mechanisms of initiation are alsoutilized during gametogenesis they are beyond the scopeof this review (Spriggs et al. 2008). This pathway is acomplex process with several mRNA-dependent stepsthat are described in detail in Fig. 3. Briefly, initiationproceeds with 1) binding of a protein complex to the5 0 cap, 2) unwinding of mRNA secondary structure,

Reproduction (2009) 137 595–617

3) joining of the small (40S) ribosomal subunit, 4)scanning of the 5 0UTR by the 40S subunit and 5)recognition of the initiator codon, release of multipleinitiation factors and joining of the large (60S) ribosomalsubunit.

Intriguingly, although translation initiation occurs atthe 5 0end of the mRNA, the poly (A) tail at the 3 0end ofthe message is also important. Poly(A) tails are bound bya family of proteins called poly(A)-binding proteins(PABPs) which interact with proteins bound to the 5 0end,bringing the 5 0- and 3 0ends of an mRNA into proximity.This end-to-end configuration is termed the ‘closed-loop’(illustrated in Fig. 3B), is highly conserved throughoutevolution (Gallie 1998, Kahvejian et al. 2001) and formsthe basis for understanding regulation by factors boundto the 3 0UTR.

mRNA-specific regulation of translation initiation

Cap-dependent translation provides multiple points atwhich initiation can be regulated in an mRNA-specificmanner (Gray & Wickens 1998, Scheper et al. 2007).Such regulation requires cis-acting elements that arenormally located within the 5 0- and 3 0UTRs, but areoccasionally found within the main open reading frame.These elements can confer differential regulation tomRNAs transcribed from the same gene via theirinclusion/exclusion due to alternative promoter orpolyadenylation site use and/or splicing. To date themajority of examples of mRNA-specific regulationduring gametogenesis involve 3 0UTR-mediated control(Radford et al. 2008). 3 0UTRs act as a repository forelements that regulate mRNA translation, stability orlocalization; normally by acting as binding sites for avariety of trans-acting factors. These include miRNA



P

P

P

P

P

P

P

P

P

P

P

P

Progesteroneor LH

AC

PKAG2

arrest

Mos

Maternal mRNA translation(inc. via cytoplasmic polyadenylation)

CyclinB2/B5

CyclinB1/B4

Wee1

APC/C

CDC2

GVBD/Meiosis I

MPFPre-MPF

Metaphasearrest

Meiosis IIentry

CyclinB1/B4

Mos

ERK

MEK

MAPK

RSK1/2

MYT1

CyclinB2/B5

CDC2 CDC25

Eg2/AurA

CPEB

RINGO/SPY

CDC2 CDK2

CDC2

Figure 2 Schematic of the major signalling and gene expression events that regulate vertebrate oocyte maturation. Oocyte maturation is induced byexposure to hormones which initiate complex cascades of kinase/phosphatase activity and translationally regulated gene expression. These highlyregulated events facilitate the coordinated completion of meiosis I and entry into meiosis II. The process of oocyte maturation is described in greaterdetail in Box 2. AC, adenylate cyclase; APC, anaphase promoting complex; AurA, aurora A; CF, cytostatic factor; CPEB, CPE-binding protein; GVBD,germinal vesicle breakdown; MAPK, mitogen-activated protein kinase; MPF, maturation promoting factor; PKA, protein kinase A; RINGO/SPY, rapidinducer of G2/M progression in oocytes/speedy. Events in which an as yet unidentified intermediate may play a role are denoted by dotted arrows.Phosphorylation events are denoted by an encircled P.


binding sites or protein binding sites such as cyto-plasmic polyadenylation elements (CPEs). It shouldbe stressed, however, that individual mRNAs oftencontain multiple cis-acting elements that, rather thanacting in a binary manner, confer combinatorialregulatory control over the mRNA (Wilkie et al. 2003).A further level of complexity is added by the factthat many of the trans-acting factors are themselvespart of dynamic multiprotein complexes that canconfer differential regulation to a bound mRNA. Whilemany mRNAs are translationally regulated, relativelyfew of the trans-acting factors have been identifiedand the molecular mechanism of action has only


been delineated in a few notable cases. Althoughthese studies indicate that translation initiation canbe regulated at almost every step, (Wilkie et al.2003, Radford et al. 2008), most of the characterizedfactors are repressors; thus, the mechanism of action of3 0UTR bound activators remains largely enigmatic.

Translation control during oogenesis: the importance ofcytoplasmic polyadenylation

One important mechanism for the regulation of mRNAsduring gametogenesis is via the dynamic modulationof mRNA poly(A) tail length (Fig. 4). In general, an

Reproduction (2009) 137 595–617


1

1A

40S

Met

2

5

eIF-4G

4A

4B

5′ UTRm7GpppG

eIF-4E

1

BA

eIF-4G

4A

4Bm7GpppG

eIF-4E

2

eIF-4GeIF3

4A

4B

m7GpppG

eIF-4E

3

1

1A

40S4

Met

25

40S

60S

5

Met

AAAAAAA

3

4G

PABP1

AUG

UAG

PAIP1

4A

4B

4E

AUG

AUG

AUG

AUG

AUG

Figure 3 Overview of mRNA translation initiation. (A) The mRNA-dependent steps of cap-dependent translation initiation. (1) The eIF4F complex(eIF4A, eIF4E and eIF4G) binds the 5 0end of the mRNA via the cap-binding protein eIF4E. eIF4G coordinates the binding of eIF4E to the cap and theATP-dependent RNA helicase activity of eIF4A, which is stimulated by eIF4B. (2) eIF4A helicase activity removes secondary structure within the5 0UTR. (3) Removal of secondary structure facilitates small ribosomal subunit (40S) binding at or near the cap. The 40S also carries with it a numberof initiation factors; including the initiator tRNA complex (eIF2-GTP-tRNAi), eIF1, eIF1A, eIF5 and eIF3. Multiple interactions, including thatbetween eIF4G and eIF3, may facilitate this step. (4) 40S and its associated factors then ‘scan’ the 5 0UTR, in a poorly defined process, to locate anappropriate initiator codon (most frequently the first AUG encountered). (5) Initiator codon recognition results in GTP hydrolysis of eIF2.GTP,release of multiple initiation factors and large ribosomal subunit (60S) joining to form the translationally competent 80S ribosome. A number ofadditional eIFs drive this process. Initiation ends with the Met-tRNAi based paired to the AUG in the P-site of the 80S ribosome, with the A-site beingavailable for delivery of tRNA. For clarity, the function of some initiation factors that are not pertinent to this review have been omitted (for a moredetailed review see (Gray & Wickens 1998, Scheper et al. 2007)), and the spatial arrangement of factors does not depict all in vivo interactions.(B) The closed-loop model of translation initiation. PABP interacts with eIF4G, functionally linking the ends of the mRNA via the interaction of eIF4Gwith the cap-binding protein eIF4E. This interaction increases both cap (panel A, step 1) and poly(A)-binding by the end-to end complex. PABP alsointeracts with eIF4B and PAIP1, both of which interact with eIF4A; thus PABP may influence unwinding of secondary structure within the 5 0UTR viaeIF4A (panel A, step 2). PAIP1 may also influence eIF3 function, enhancing or stabilizing interaction of the 40S with the mRNA (panel A, step 3).While PABP-eIF4G interactions are the most studied, each of these interactions may play an incremental role in initiation complex stabilization,thereby enhancing various steps of translation initiation. A single molecule of PABP is shown for simplicity. Numbered factors, eIFs. Filled blackcircle, 5 0cap. Dotted lines denote lack of interaction between overlapping factors.


increase in poly(A) tail length correlates with trans-lational activation while a decrease correlates withtranslational silencing. Many X. laevis oocyte mRNAscontain 3 0UTR elements that target the mRNA for rapidpoly(A) tail shortening following export into thecytoplasm (e.g. cyclin B1 poly(A) length in thecytoplasm of immature oocytes is w30 nucleotides;Sheets et al. 1994). Such mRNAs remain stable but arestored translationally silent until, under permissiveconditions, the poly(A) length is increased again.Critically these regulated, and often dramatic, changesin poly(A) length occur within the cytoplasm by a

Reproduction (2009) 137 595–617

mechanism that is distinct from nuclear polyadenylation(Radford et al. 2008). Indeed, enucleation of X. laevisoocytes does not block this process (Fox et al. 1989).Pique et al. (2008) recently estimated that w45% ofstored mRNAs undergo cytoplasmic polyadenylation-dependent regulation during X. tropicalis oocytematuration and that O30% of mouse and humanoocyte mRNAs were similarly regulated. Cytoplasmicpolyadenylation is critical for gametogenesis in mouse,C. elegans, Drosophila and other model organisms andalso occurs in human oocytes (Gebauer et al. 1994,Prasad et al. 2008; reviewed in (Radford et al. 2008));



Maximal translational activation

B

Aa) b)

AAAAAADAZL

1 A

Translationallysilent

2 A

3 AAA

4 A

A

–

DAZL

AAAAAA

A

AAAAAA

DAZL

Figure 4 Translational activation of silent mRNAs during gametogen-esis. (1) Many mRNAs are activated by cytoplasmic polyadenylation,with some being bound by repressor proteins (3) to prevent low levelsof precocious translation prior to activation. Other mRNAs with shortpoly(A) tails (2 and 4) are activated without a requirement forcytoplasmic polyadenylation by mRNA-specific translational activatorssuch as DAZL family proteins. Many mRNAs contain multipleregulatory elements that facilitate complex patterns of activation, e.g.mRNAs are initially activated in the absence of cytoplasmicpolyadenylation (panel A) but their translational activity is thenmaintained or enhanced by subsequent polyadenylation (panel B).


but has been most intensively studied during X. laevisoocyte maturation.

In order for stage VI X. laevis oocytes to mature theyneed to exit prophase I-arrest and recommence meiosis,a process that can be induced by progesterone (Box 2and Fig. 2). This requires the translational activation ofmany mRNAs (e.g. c-Mos and cyclin B1) in a highlyorchestrated temporal sequence. This is achieved viaspecific cis-acting elements within their 3 0UTRs, whichpromote cytoplasmic polyadenylation (Belloc et al.2008, Radford et al. 2008).

The critical requirement for the cytoplasmic poly-adenylation of individual mRNAs for oocyte maturationwas demonstrated in seminal studies in X. laevis andmouse. Ablation of endogenous c-Mos mRNA poly-adenylation in X. laevis oocytes via selective removal ofits 3 0end, resulted in inhibition of c-Mos translation andprogesterone-induced maturation; both of which wererescued by annealing a prosthetic, polyadenylation-competent 3 0end or a prosthetic poly(A) tail to the mRNA(Sheets et al. 1995, Barkoff et al. 1998). Similarly,Gebauer et al. (1994) demonstrated the importance ofc-Mos polyadenylation for mouse oocyte maturation.

Mouse knock-out/down models have revealed thatother stages of mammalian gametogenesis are alsodirected by this process. By deleting a factor requiredfor the adenylation of many mRNAs (CPE-bindingprotein (CPEB)), a lack of cytoplasmic polyadenylationwas shown to result in the arrest of both male and femalegerm cells at the pachytene stage of meiotic prophase I(Tay & Richter 2001). This may be due, at least in part, tothe loss of synaptonemal complex protein (Sycp) 1 andSycp3 translation, which correlates with a failure ofhomologous chromosome pairing and resultant meioticfailure. When CPEB expression was reduced aftercompletion of the pachytene stage of prophase I, viaexpression of an siRNA under the control of the zona


pellucida 3 (ZP3) promoter, the loss of polyadenylationof a number of key mRNAs was demonstrated (Racki &Richter 2006). These include growth differentiationfactor 9 (GDF9), a critical folliculogenic growth factorexpressed by the oocyte; SMAD5, which functions in thetransforming growth factor (TGF) b signalling pathway;and H1Foo, an oocyte-specific histone 1 variant which isessential for oocyte maturation. A progressive decreasein oocyte and follicle number was observed in theseCPEB-knockdown mice. At the dichtyate stage, a numberof oocyte phenotypes were observed including apopto-sis, abnormal polar bodies and parthenogenetic celldivision. Additionally, oocyte-granulosa cell interactionsare disrupted leading to oocyte detachment from thecumulus granulosa cells and impaired folliculogenesisand these mice often developed ovarian cysts with age.Several of the phenotypes associated with these CPEB-knockdown mice have led to the proposal that they maymodel aspects of human premature ovarian failuresyndrome (POF; Racki & Richter 2006). Together, thesefindings indicate that cytoplasmic polyadenylation isrequired at multiple steps during the female meioticprogram (recently reviewed in (Belloc et al. 2008)).However, CPEB also functions as a translationalrepressor and loss of mRNA repression may contributeto these phenotypes, but this remains to be addressed.Interestingly, CPEB expression (and cytoplasmic poly-adenylation) is not restricted to germ cells but itsfunctions in somatic cells are beyond the scope of thisreview and are detailed elsewhere (Richter 2007).

Sequences required for cytoplasmic polyadenylation

All cytoplasmically polyadenylated mRNAs containvarying combinations of cis-acting sequences thatdetermine the precise timing and extent of theirpolyadenylation (reviewed in detail in (Belloc et al.2008)). CPEs are the best characterized and have beenpredominantly studied in X. laevis oocytes (McGrewet al. 1989). CPEs function in collaboration with thehexanucleotide (hex) signal AAUAAA (see (Mandel et al.2008) for review of nuclear polyadenylation), and bothelements are binding sites for multiprotein complexesthat can interact and are reviewed in detail elsewhere(Radford et al. 2008). Briefly, CPEB bound to the CPE(s)interacts with cytoplasmic polyadenylation and speci-ficity factor (CPSF) bound to the hex. CPEB then recruitsa cytoplasmic poly(A) polymerase GLD-2 and thepoly(A) RNase (PARN) which acts antagonistically.During many oogenic stages, PARN activity exceedsthat of GLD-2, resulting in deadenylation and trans-lational silencing of the mRNA. Thus, CPEs also play arole in rapid deadenylation and translational silencing ofregulated mRNAs as they exit the nucleus. Underpermissive conditions (e.g. during oocyte maturation)CPEB is phosphorylated and PARN dissociates fromthe complex, facilitating GLD-2-mediated mRNA

Reproduction (2009) 137 595–617



polyadenylation. Recent data suggest that PABPs mayinteract with polyadenylation complex components,raising the possibility that this may facilitate rapidPABP binding to newly synthesized poly(A) tails topromote the mRNA closed-loop conformation (Fig. 3B)thereby stimulating translation and protecting againstPARN-mediated deadenylation (Richter 2007, Radfordet al. 2008).

CPE-containing mRNAs exhibit varying patterns oftranslational activation during oogenesis; with thesequence, number and position of CPEs with respect tothe hex being important. Recent studies further definedrules which determine at what point during X. laevisoocyte maturation, and with how many adenosines, agiven CPE-containing mRNA will be polyadenylated(reviewed in (Belloc et al. 2008)). Regulation conferredby CPEs/CPEB can be further modulated by the presenceof additional cis-acting elements in the mRNA 3 0UTR,such as the pumilio 2 (PUM2)-binding element (PBE).PUM2 is a member of the PUM/Fem-3 mRNA-bindingfactor (FBF) (PUF) family of RNA-binding proteins whichhave well-defined roles in translational repression in thegerm-line of many species (Wickens et al. 2002, Kimble& Crittenden 2007). However, both CPEB-mediatedpolyadenylation, as well as repression, appears to beaugmented by PUM2 in X. laevis oocytes.

A number of X. laevis mRNAs are activated bypolyadenylation very early in oocyte maturation, in anapparently CPE-independent manner. To date, twofurther cis-acting cytoplasmic polyadenylation elementsare described; Musashi/polyadenylation responseelements (PREs) that are bound by Musashi (Charles-worth et al. 2000, 2004) and the translational controlsequence (TCS) whose trans-acting factor(s) is unknown(Wang et al. 2008). It is proposed that these elementsfunction during early oocyte maturation prior to CPE-dependent polyadenylation. The temporal regulation ofcytoplasmic polyadenylation is a contentious issuesince, in most of the cases described to date, PREs andTCSs are present in mRNAs that also contain CPEs.However, this may reflect a role of multiple cis-actingelements in the intricate regulation of translation viacombinatorial control. mRNAs that do not contain CPEsare normally silenced following maturation-inducedgerminal vesicle breakdown, via a default pathway inwhich their poly(A) tails are removed, further enhancingthe reprogramming of translation (Fox & Wickens 1990,Varnum & Wormington 1990).

PABPs

The primary function of PABPs is to transduce the signalprovided by the poly(A) tail and convey it to thetranslational machinery (see Fig. 3B). Indeed, tetheringPABPs to the 3 0UTR of non-adenylated reporter mRNAscan functionally replace poly(A) tails in X. laevis oocytes(Gray et al. 2000, Wilkie et al. 2005). Thus, cytoplasmic

Reproduction (2009) 137 595–617

polyadenylation is thought to exert its effects ontranslation largely by recruiting PABPs to the newlyextended poly(A) tails. Consistent with a role of PABP inpoly(A) tail mediated changes during oogenesis, over-expression of PABP1 in X. laevis oocytes blocks thematuration-induced deadenylation and translationalinactivation of mRNAs that lack CPEs, as well asinhibiting the recruitment of some CPE-containingmRNAs onto polysomes (Wormington et al. 1996).A role for PABP in oogenesis is also indicated by earlymeiotic defects in PAB-1-deficient C. elegans(Maciejowski et al. 2005).

Cytoplasmic PABPs are evolutionarily conserved fromyeast to humans and have multiple roles in translationand mRNA stability (Table 3; reviewed in (Gorgoni &Gray 2004)). Drosophila only encode a single cyto-plasmic PABP which, like PABP in budding yeast, isrequired for viability, whereas C. elegans encodes twocytoplasmic PABPs and vertebrate genomes encodemultiple cytoplasmic PABP genes (Fig. 5A; Gorgoni &Gray 2004). No vertebrate PABP-deficient phenotypeshave yet been described and only a handful of non-synonymous single nucleotide polymorphisms aredocumented within human PABP genes, possiblyreflecting an incompatibility of PABP mutation withviability.

In vertebrates, only the molecular functions of theprototypical PABP1 (also known as PABPC1), and to alesser extent ePABP, have been studied in any detail(Voeltz et al. 2001, Gorgoni & Gray 2004, Wilkie et al.2005). As few as 12 adenosines are required to bind asingle PABP molecule, but each PABP has a footprint ofw26 adenosines, binding cooperatively and contigu-ously such that long poly(A) tails may be associatedwith multiple PABPs (reviewed in Gorgoni & Gray2004). PABP1 comprises four highly conserved RNArecognition motifs (RRMs) that exhibit differing specifi-cities for RNA sequences and interacting proteins, suchas the translation initiation factor, eIF4G (Figs 3 and5B); a less well-conserved proline-rich region thatfunctions in cooperative PABP binding to poly(A)(Melo et al. 2003) and a C-terminal domain (termedthe PABC domain) which exhibits homology to HECTdomains of E3-ubiquitin ligases (Kozlov et al. 2001;Fig. 5B). Neither the proline-rich nor the PABC domainbinds the RNA but both interact with proteins; the latterbinding proteins which contain a PABC-binding domaincalled PAM2, such as the PABP-interacting protein 1(PAIP1; Figs 3 and 5B; Kozlov et al. 2001, Albrecht &Lengauer 2004, Kozlov et al. 2004, Lim et al. 2006).

PABP4, testis-specific PABP (tPABP -PABP2 in mouseand PABP3 in human) and ePABP exhibit a similardomain structure to PABP1 (Yang et al. 1995, Feral et al.2001, Guzeloglu-Kayisli et al. 2008), whereas PABP5has neither a proline-rich linker region nor a PABCdomain (Blanco et al. 2001). Vertebrates also encode anadditional protein, ePABP2, which, although apparently



Mm PABP5

Hs PABP5

Xt ePABP

Xl ePABP

Mm ePABP

Hs ePABP

Mm tPABP

Xt PABP1

Xl PABP1

Hs tPABP

Hs PABP1

Mm PABP1

Ce PAB-2

Cb PAB-2

Ce PAB-1

Cb PAB-1

Dm PABP

Xl PABP4

Mm PABP4

Hs PABP4

Hs hnRNPA1

Protein-proteininteractions

PAIP1

DAZLelF4G

eRF3

PABP1PAIP1

PABCRRM1 RRM2 RRM3 RRM4

A

B

Figure 5 PABP family evolution and key protein–protein interactions.(A) Phylogenetic tree based on the sequence of PABP family proteins.The tree is rooted to human hnRNPA1, an RNA-binding protein of theRRM family. Hs: Homo sapiens, Mm: Mus musculus, Xl: Xenopuslaevis, Xt: Xenopus tropicalis, Ce: Caenorhabditis elegans, Cb:Caenorhabditis briggsae, Dm: Drosophila melangaster. PABP4 andePABP are vertebrate-specific and PABP5 is mammalian-specific.tPABP is mammalian-specific and arose from a PABP1 retro-transposition event. (B) PABP1 domain structure and proteininteractions with translation factors and DAZL. PABP1 consists of fournon-identical RRM domains that have different RNA-binding specifi-cities. The majority of poly(A)-binding activity is conferred by RRMs 1and 2. The specificity of RRMs 3 and 4 is less clear but this region hasbeen linked to the ability of PABP1 to bind AU-rich sequences. Thestructure of RRMs 1 and 2 bound to poly(A) has been solved showingthat the RNP motifs, characteristic of RRMs, directly contact RNAleaving the opposite face free for protein–protein interactions includingeIF4G and PAIP1 (Gorgoni & Gray 2004). RRMs 3 and 4 also mediateprotein–protein interactions (not shown). The C-terminal region ofPABP is composed of a proline-rich region and the PABC domain;PABP1 homodimerization, and DAZL interaction map to this area.The PABC motif interacts with a variety of factors, including eRF3 andPAIP1, via their PAM2 domains. Bioinformatic analyses suggest thatPABP1 may associate many additional proteins via PABC–PAM2mediated interactions (Albrecht & Lengauer 2004). The region ofPABP1 that interacts with DAZL is also shown. The mapped region ofinteraction is depicted by a horizontal line. References: PAIP1 (Craiget al. 1998, Gray et al. 2000, Roy et al. 2002), eIF4G (Imataka et al.1998), PABP1 (Melo et al. 2003), DAZL (Collier et al. 2005), eRF3(Hoshino et al. 1999). Other experimentally determined proteininteractions are not shown including eIF4B.



predominantly cytoplasmic, is more closely related tothe nuclear PABPN1 (Cosson et al. 2004). ePABP2 may,like its nuclear counterpart PABPN1, regulate poly(A) taillength rather than translation (Good et al. 2004) and isbeyond the scope of this review.

The expression patterns of vertebrate PABPs aredocumented to varying degrees. PABP1 is expressed ina wide variety of tissues in humans, mice and X. laevisand available evidence suggests that PABP4 mRNA mayalso be widely expressed in these species (Kleene et al.1994, Gu et al. 1995, Yang et al. 1995, Cosson et al.2002; GS Wilkie & NK Gray, unpublished observations).Both PABP1 and PABP4 mRNAs are detected in the testisand ovary of humans, mice and X. laevis. Other PABPsappear to be more restricted in their expression. RT-PCRanalyses of human PABP5 suggest low-level expressionin a variety of tissues, with higher levels being detectedin the brain and gonads (Blanco et al. 2001). MousetPABP mRNA and protein and human tPABP mRNA areonly expressed in pachytene spermatocytes and roundspermatids (Kleene et al. 1994, 1998, Kimura et al.2009). ePABP mRNA is detected in both mouse andhuman oocytes and is downregulated during earlyembryogenesis (Seli et al. 2005, Guzeloglu-Kayisliet al. 2008, Sakugawa et al. 2008). ePABP mRNA isalso expressed at low levels in mouse testis (Wilkie et al.2005) but appears to be more widespread in humans.ePABP protein expression has been studied in X. laevisoogenesis where a developmental switch betweenPABP1 and ePABP appears to occur (Cosson et al.2002). Thus, ePABP is the predominant PABP presentduring most of oogenesis, and during oocyte maturationand early embryogenesis when regulated poly(A) tailchanges drive the translational selection of mRNAs(Voeltz et al. 2001). The physiological basis for thisswitch to ePABP remains to be determined, althoughePABP stimulates translation analogously to PABP1(Wilkie et al. 2005).

PABP1 regulates translation at multiple steps and bymultiple mechanisms (Gray et al. 2000, Kahvejian et al.2005). This is in keeping with the fact that PABP1contains several domains that can stimulate translationand can interact with numerous translation factors(Figs. 3B and 5B). The best-characterized interaction isbetween PABP1 and eIF4G, a component of the cap-associated complex eIF4F (Figs 3 and 5B), and thisinteraction is conserved from yeast to man (Derry et al.2006). Interaction of PABP1 with eIF4G is proposed toincrease both the poly(A)-binding affinity of PABP1 andthe cap-binding affinity of the eIF4F complex (Fig. 3A,step 1), establishing the ‘closed-loop’ mRNP confor-mation (Fig. 3B). This conformation is believed to beoptimal for translational initiation and ribosomalrecycling. Consistent with this, mutations that disruptthe PABP1–eIF4G interaction result in compromisedtranslational activity (Tarun et al. 1997). PABP1 alsobinds to eIF4B (Bushell et al. 2001) and PAIP1

Reproduction (2009) 137 595–617



(Craig et al. 1998; Figs 3 and 5B), with both interactionsproposed to promote eIF4A RNA helicase activity whichaids unwinding of secondary structure within the 5 0UTR(Fig. 3A, step 2). PAIP1 may also influence eIF3 function(Martineau et al. 2008) which is thought to facilitatethe 40S subunit joining step (Fig. 3A, step 3). PABP1 canalso enhance later steps in translation including 60Ssubunit joining (Sachs & Davis 1989, Kahvejian et al.2005) and translation termination. PABP1 regulatestermination via an interaction with the eukaryotictranslation termination factor 3 (eRF3; Hoshino et al.1999) and recent work in yeast described a potentialrole for the PAB-1–eRF3 interaction in 40S recruitment;but existence of this mechanism has not been describedin vertebrates (Amrani et al. 2008). The basis for thePABP1-mediated regulation of 60S joining remainsenigmatic (Kahvejian et al. 2005). By necessity mostPABP1 interactions have been studied in isolation and itis unclear to what extent either a single PABP1 moleculeor a cooperatively bound PABP1 oligomer can supportmultiple interactions with the same or multiple proteinpartners. Therefore, while significant progress has beenmade, many important questions remain pertaining tothe mechanisms of action of PABP1 and the functions ofother PABPs (reviewed in greater detail in (Gorgoni &Gray 2004, Derry et al. 2006)).

Although the amplitude of translational activationduring oogenesis is related to the length of the poly(A)tail added/number PABPs bound, this relationship iscomplex. Upon injection into X. laevis oocytes, reportermRNAs with the same length poly(A) tail but different5 0UTRs are translated to differing extents (Gallie et al.2000) and some endogenous oocyte mRNAs areactivated without changes in poly(A) tail length (e.g.Wang et al. 1999) (Fig. 4). Moreover, poly(A) tails ofsufficient length to support multiple PABP binding donot always confer translational activation. In immatureX. laevis oocytes, for instance, c-Mos mRNA has anaverage poly(A) tail length of 50 nucleotides but is nottranslated, although reporter mRNAs containing thesame length poly(A) tail are very efficiently translated(Sheets et al. 1994). This complexity is likely to bemediated, at least in part, via a poorly understoodinterplay between trans-acting repressors/activatorsbound to 3 0UTR cis-acting elements and the poly(A)tail/PABP (Fig. 4). For example, repressors may preventPABP binding to poly(A) tails or may modulate PABPactivity. In this regard, Kim & Richter (2007) recentlyshowed that ePABP associates with CPEB in a repressivecomplex in X. laevis oocytes prior to maturation. TheePABP–CPEB interaction is subsequently lost in maturingoocytes and ePABP binds the de novo synthesizedpoly(A) tails. However, our understanding has beenhampered by the relatively few trans-acting factors thathave been identified to date. Recently, we and othershave gained insights into the links between poly(A) tail

Reproduction (2009) 137 595–617

length, PABP function and the DAZL proteins, a proteinfamily with critical roles in gametogenesis.

DAZL family of proteins

DAZ, DAZL and BOULE comprise a family of proteins(Fig. 6A) that have important roles in gametogenesis.Deleted in azoospermia (DAZ) is located within theazoospermia factor locus (AZFc) on the humanY-chromosome and was first identified as a candidatefactor underlying diverse azoospermia and oligosper-mia defects in 5–10% of infertile men (Reijo et al.1995, Reynolds & Cooke 2005). Human DAZL (hDAZL)and BOULE (hBOULE) are both autosomal and theirroles in human fertility are less clear. hDAZL SNPs havebeen linked to Sertoli cell-only syndrome, oligospermiaand POF (Reynolds & Cooke 2005, Fassnacht et al.2006, Tung et al. 2006a, 2006b). An associationbetween a lack of hBOULE protein and adult malegerm cell meiotic arrest was observed (Luetjens et al.2004); although this does not provide any evidence thathBOULE deficiency is causative. Not all familymembers are conserved across species with DAZbeing found in humans and old world monkeys,while invertebrates only encode boule (Fig. 6A).

Germ cells are the main site of DAZL family proteinexpression. However, the mammalian germ cellexpression patterns of the DAZL family members arecomplex, although only DAZL has been detectedin oocytes (detailed in Table 1; reviewed in detail in(Reynolds & Cooke 2005)). In humans, mice and X. laevisDAZL expression is detectable in primordial germ cells(PGCs) and all oogenic stages examined so far (Table 1).DAZL has also been reported to be expressed in humanand mouse granulosa cells (Ruggiu et al. 1997, Dorfmanet al. 1999), human theca interna cells (Nishi et al. 1999)and in the granulosa-luteal cells of human corpus lutea(Pan et al. 2002, 2008), but this remains contraversial. Ininvertebrates, DAZL family protein expression is notrestricted to the germ cells/gonads with Boule beingexpressed in Drosophila larval neurons and adult brain(Joiner & Wu 2004, Hoopfer et al. 2008).

Knock-down of X. laevis DAZL (xDAZL) leads to adefect in PGC migration and proliferation and thus a lossof both female and male gametes (Houston & King2000). Deletion of mouse Dazl (mDazl) results in acomplete loss of female germ cells before birth and afailure to advance past meiotic prophase I in spermato-genesis (Ruggiu et al. 1997, Lin & Page 2005). However,it should be noted that the timing of male germ cell lossappears to vary depending on the genetic background;with the loss of male germ cells occurring earlier on aninbred background (reviewed in Reynolds & Cooke2005). Consistent with this, DAZL was recently demon-strated to be an intrinsic factor required for germ cellentry into meiosis (Lin et al. 2008). Only oogenesisis affected in worms; loss-of-function mutations in the




C. elegans Boule ortholog daz-1 and DAZ-1 knock-downin Caenorhabditis briggsae lead to pachytene arrest inthe female germline (Karashima et al. 2000) anddisruption of the hermaphroditic sperm/oocyte switch(Otori et al. 2006) respectively. Conversely, loss ofDrosophila boule results in defective G2/M transitionduring spermatogenesis only (Eberhart et al. 1996).Taken together, these observations strongly support a

Dr Dazl

Xt Dazl

Xl Dazl

Mm DAZL

Hs DAZL

Hs DAZ-2

Hs DAZ-3

Hs DAZ-1

Hs DAZ-4

Ce DAZ-1

Cb DAZ-1

Dm Boule

Mm Boule

Hs Boule

Hs hnRNPA1

DAZ

DAZDAZL

DAZ DAZ DAZ DAZ DAZ DAZ DAZ DAZRRM

DAZ

B

A

PUM2DZIP1DAZAP1 and DAZAP2

DAZRRM

DAZLPABP

DAZAP1 and DAZAP2

Dynein light chain (Dlc)

PUM2

DAZRRM

BOULE BOULE / DAZL

Figure 6 DAZL family evolution and key protein–protein interactions.(A) Phylogenetic tree based on the sequence of DAZL family proteins.The tree is rooted to human hnRNPA1, an RNA-binding protein of theRRM family. Hs: Homo sapiens, Mm: Mus musculus, Xl: Xenopus laevis,Xt: Xenopus tropicalis, Dr: Danio rerio, Ce: Caenorhabditis elegans,Cb: Caenorhabditis briggsae, Dm: Drosophila melangaster. DAZL isvertebrate-specific and is derived from the ancestral BOULE gene, whichis conserved from invertebrates to humans (Reynolds & Cooke 2005) andincludes the confusingly named nematode DAZ-1 genes. DAZ appears tohave arisen from a transposition of the Dazl gene, followed by subsequentgene amplification and modification, and is present in humans andold-world monkeys (Saxena et al. 1996). Although the human genomenormally contains four DAZ genes, copy number variation has beenreported (Saxena et al. 2000). (B) DAZL family protein domain structuresand interactions. Family members contain two recognisable motifs,a single RRM (except for some DAZ genes) and the characteristic DAZmotif, a repeat of a 24 amino acid sequence that is rich in glutamine,proline and tyrosine (which varies between 7 and 24 copies in DAZ).The figure depicts all interactions described in vertebrates. The describedregion of interaction is depicted bya horizontal line. See Table 2 for DAZLfamily protein interaction references.


critical and evolutionarily conserved role for DAZLfamily proteins in gametogenesis.

Phenotypic rescue experiments indicate that differentDAZL-family proteins may contribute to gametogenesisby related molecular mechanisms. For instance, ectopicexpression of xDAZL and hBOULE in Drosophila orhuman DAZ (hDAZ) expression in mice can partiallyrescue the boule and mDazl null male phenotypesrespectively (Houston et al. 1998, Slee et al. 1999, Xuet al. 2003). However, this redundancy is incompletesince many organisms, including mice, encode bothDazl and Boule, but mouse DAZL (mDAZL) deficiencyresults in male infertility even though mouse BOULE(mBOULE) expression partially overlaps with normalmDAZL expression during spermatogenic differentiation(Reynolds & Cooke 2005). While differential expressionmay contribute to this lack of redundancy, differentfamily members may have both overlapping and distincttargets and functions. A better understanding of theirindividual molecular functions and mRNA targets isrequired to clarify this issue.

DAZL family proteins are translational regulators

DAZL family proteins are identified by the presence oftwo functional domains; RRMs and DAZ domains, thelatter of which appears to be unique to this family.RRMs are found in many RNA-binding proteins (such asPABPs) which function in processes as diverse as mRNAsplicing, localization, translation and stability. Indeed,several studies have demonstrated that DAZL familyproteins are bona fide RNA-binding proteins (Reynolds& Cooke 2005), consistent with a role in nuclear and/orcytoplasmic post-transcriptional gene regulation. Thesubcellular localization of DAZL family proteinschanges during gametogenesis and, while DAZL familyproteins may have undefined nuclear functions, it is ofnote that the timing of germ cell loss in outbred malemDazl null mice corresponds to the meiotic stagesduring which mDAZL becomes localized to thecytoplasm (Reynolds & Cooke 2005, Anderson et al.2007). The importance of a cytoplasmic, rather thannuclear, function was elegantly demonstrated in Droso-phila, wherein nuclear exclusion of Boule did notdisrupt germ-cell entry into meiosis (Cheng et al. 1998).In humans and mice, DAZL is located in the cytoplasmthroughout oogenesis.

The first link to a role in translation came from geneticexperiments in Drosophila, where Twine/Cdc25A mRNAtranslation was found to be defective in boule null flies.However, it remains unclear whether Boule mediates itseffect directly on the translation of Twine/Cdc25A mRNAor indirectly via alterations in the expression of unknownmRNAs upstream of Twine/Cdc25A (Maines & Wasser-man 1999). In vertebrates, a translational or mRNAstability role for DAZL family proteins was intimated bythe detection of mDAZL, and subsequently exogenously

Reproduction (2009) 137 595–617


Table

1Im

munohis

toch

emic

alan

alys

isof

DA

ZL

fam

ily

pro

tein

sin

mam

mal

ian

germ

cells.

BOULE

DAZL

DAZ

Germ

cellstag

eH

sM

mH

sM

mH

s

PG

CNo

Xu

etal

.(2

001)

No

Xu

etal

.(2

001)

c/N

Rei

joet

al.

(2000),

Rugg

iuet

al.

(2000)

and

Ander

son

etal

.(2

007)

C/N

Xu

etal

.(2

001)

c/N

Rei

joet

al.

(2000)

and

Xu

etal

.(2

001)

C/N

Xu

etal

.(2

001)

Sper

mat

ogo

nia

No

Xu

etal

.(2

001)

No

Xu

etal

.(2

001)

and

Moore

etal

.(2

003)

c/N

Rei

joet

al.

(2000)

and

Xu

etal

.(2

001)

C/N

Rugg

iuet

al.

(1997)

C/n

Huan

get

al.

(2008)

c/N

Rei

joet

al.

(2000)

CR

ugg

iuet

al.

(2000)

and

Lin

etal

.(2

001)

c/N

Rei

joet

al.

(2000)

and

Xu

etal

.(2

001)

C/N

Xu

etal

.(2

001)

Lepto

tene

CX

uet

al.

(2001)

and

Kost

ova

etal

.(2

007)

CX

uet

al.

(2001)

and

Moore

etal

.(2

003)

CR

eijo

etal

.(2

000),

Rugg

iuet

al.

(2000),

Lin

etal

.(2

001)a

nd

Xu

etal

.(2

001)

CR

ugg

iuet

al.

(1997),

Rei

joet

al.(2

000)an

dX

uet

al.

(2001)

C/N

Xu

etal

.(2

001)

No

Huan

get

al.

(2008)

CR

eijo

etal

.(2

000)

Zyg

ote

ne

CX

uet

al.

(2001),

Luet

jens

etal

.(2004)a

nd

Kost

ova

etal

.(2007)

CX

uet

al.

(2001)

and

Moore

etal

.(2

003)

CR

eijo

etal

.(2000),

Rugg

iuet

al.(

2000),

Lin

etal

.(2

001)a

nd

Xu

etal

.(2

001)

CR

ugg

iuet

al.

(1997),

Rei

joet

al.

(2000)

and

Xu

etal

.(2

001)

No

Xu

etal

.(2

001)

and

Huan

get

al.

(2008)

CR

eijo

etal

.(2

000)

Pach

yten

eC

Xu

etal

.(2

001),

Luet

jens

etal

.(2

004),

Tung

etal

.(2

006a)

and

Kost

ova

etal

.(2

007)

CX

uet

al.

(2001),

Moore

etal

.(2

003)

and

Tung

etal

.(2

006a)

CR

eijo

etal

.(2

000),

Rugg

iuet

al.

(2000),

Lin

etal

.(2

001),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

CR

ugg

iuet

al.

(1997),

Rei

joet

al.

(2000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

No

Xu

etal

.(2

001)

and

Huan

get

al.

(2008)

CR

eijo

etal

.(2

000)

Seco

ndar

ysp

erm

atocy

teC

Xu

etal

.(2

001),

Tung

etal

.(2

006a)

and

Kost

ova

etal

.(2

007)

CX

uet

al.

(2001),

Moore

etal

.(2003)a

nd

Tung

etal

.(2006a)

CR

eijo

etal

.(2

000),

Rugg

iuet

al.

(2000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

CR

eijo

etal

.(2

000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

No

Xu

etal

.(2

001)

and

Huan

get

al.

(2008)

No

Luet

jens

etal

.(2

004)

CR

eijo

etal

.(2

000)


Reproduction (2009) 137 595–617 www.reproduction-online.org


Round

sper

mat

ids

No

Xu

etal

.(2

001),

Luet

jens

etal

.(2004),

Tung

etal

.(2

006a)

and

Kost

ova

etal

.(2007)

No

Xu

etal

.(2

001),

Moore

etal

.(2

003)

and

Tung

etal

.(2

006a)

CR

eijo

etal

.(2

000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

CX

uet

al.

(2001)

and

Tung

etal

.(2

006a)

No

Rei

joet

al.(2

000),

Xu

etal

.(2

001)

and

Huan

get

al.

(2008)

No

Lin

etal

.(2

001)

No

Rei

joet

al.

(2000)

CH

aber

man

net

al.

(1998)

Elonga

ting

sper

mat

ids

No

Xu

etal

.(2

001),

Luet

jens

etal

.(2

004),

Tung

etal

.(2

006a)

and

Kost

ova

etal

.(2

007)

No

Xu

etal

.(2

001),

Moore

etal

.(2

003)

and

Tung

etal

.(2

006a)

CR

eijo

etal

.(2

000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

No

Rei

joet

al.

(2000),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

No

Rei

joet

al.

(2000)

and

Xu

etal

.(2

001)

No

Lin

etal

.(2

001)

CH

aber

man

net

al.

(1998)

Mat

ure

sper

mNo

Xu

etal

.(2

001),

Luet

jens

etal

.(2

004)

and

Tung

etal

.(2

006a)

No

Xu

etal

.(2

001),

Moore

etal

.(2

003)

and

Tung

etal

.(2

006a)

CLi

net

al.

(2002)

No

Rugg

iuet

al.

(1997),

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

CH

aber

man

net

al.

(1998)

No

Xu

etal

.(2

001)

and

Tung

etal

.(2

006a)

No

Rei

joet

al.

(2000)

Foet

aloogo

nia

n.d

No

Moore

etal

.(2

003)

CR

ugg

iuet

al.

(2000)

and

Ander

son

etal

.(2

007)

CR

ugg

iuet

al.

(1997)

C/N

Dorf

man

etal

.(1

999)

Pri

mord

ial

follic

leoocy

tes

n.d

No

Moore

etal

.(2

003)

CD

orf

man

etal

.(1

999)

and

Nis

hi

etal

.(1

999)

CR

ugg

iuet

al.

(1997)

Pre

-antr

alfo

llic

leoocy

tes

n.d

No

Moore

etal

.(2

003)

CD

orf

man

etal

.(1

999),

Nis

hi

etal

.(1

999)

and

Moore

etal

.(2

003)

CR

ugg

iuet

al.

(1997)

Antr

alfo

llic

leoocy

tes

n.d

No

Moore

etal

.(2

003)

CD

orf

man

etal

.(1

999)

and

Nis

hi

etal

.(1

999)

CR

ugg

iuet

al.

(1997)

Bei

ng

Y-ch

rom

oso

mal

hD

AZ

isnot

expre

ssed

inova

ry.

n.d

.,not

det

erm

ined

;N

o,

not

obse

rved

inim

munohis

toch

emis

try;

C,

excl

usi

vely

cyto

pla

smic

loca

liza

tion;

N,

excl

usi

vely

nucl

ear

loca

liza

tion;C

/N,det

ecte

din

both

com

par

tmen

ts;C

/n,pre

dom

inan

tly

cyto

pla

smic

loca

liza

tion;c/

N,pre

dom

inan

tly

nucl

ear

loca

liza

tion.W

her

eco

nflic

ting

report

sex

istth

ese

are

pre

sente

don

indiv

idual

lines

;H

s,H

om

osa

pie

ns;

Mm

,M

us

musc

ulu

s.


www.reproduction-online.org Reproduction (2009) 137 595–617



expressed zebrafish DAZL (zDAZL), on actively translat-ing polyribosomes (polysomes; Tsui et al. 2000a, 2000b,Maegawa et al. 2002).

Direct evidence supporting a translational role forDAZL family proteins came from two sets of obser-vations. First, zDAZL upregulated reporter proteinsynthesis in an in vitro translation assay (Maegawaet al. 2002). Second, tethering of DAZL family membersto the 3 0UTR of reporter mRNAs microinjected intoX. laevis oocytes provided evidence that xDAZL,mDAZL, hDAZL, hDAZ and hBOULE exhibit a con-served and robust ability to stimulate the translation ofmRNAs to which they were bound. This suggested arole in mRNA-specific, rather than global, regulationof translation (Collier et al. 2005). It is of note thatthis functional conservation occurs despite the relativelylow sequence homology between some of theseproteins outside the RRM and DAZ domains. Finally,the 3 0UTRs from two potential mDAZL target mRNAsconferred DAZL-dependent translational regulation toreporter mRNAs, in microinjected X. laevis oocytes, andthe levels of the protein products of these endogenoustarget mRNAs were reduced in the remaining germcells in testis of mDazl null mice (see below; Reynoldset al. 2005, 2007).

mRNA targets of DAZL family protein-mediatedregulation of translation

The phenotypes associated with the loss of DAZL familyfunction suggest that the targets of DAZL-mediatedtranslational regulation are likely to be mRNAs thatencode proteins with key functions in oogenesis and/orspermatogenesis. Initial studies indicated that DAZLexhibited a general affinity for poly(U) RNA andsuggested that the affinity may be increased for poly(U)interspersed by G or C residues (Venables et al. 2001).These studies defined a loose consensus sequence,UUU[G/C]UUU, which was later refined to GUUCand U2–10[G/C]U2-10 for zDAZL (Maegawa et al. 2002)and mDAZL (Reynolds et al. 2005) respectively.Tethered-function and reporter mRNA studies in X. laevisoocytes (Collier et al. 2005) and zDAZL-transfected cellsrespectively (Maegawa et al. 2002) suggest thatfunctional DAZL-binding sites were likely to be locatedin 3 0UTRs of target mRNAs. Furthermore, the tethering ofmultiple molecules of DAZL stimulates translation to agreater extent than tethering of a single DAZL molecule;suggesting that multiple DAZL-binding sites may berequired for efficient translational activation of an mRNA(Collier et al. 2005). Consistent with this, several DAZLtarget mRNAs contain multiple DAZL-binding motifs(Reynolds et al. 2005, 2007) and identification of such3 0UTR motif clusters may prove important in identifyingbona fide target mRNAs. While the DAZL-bindingconsensus occurs relatively frequently throughout

Reproduction (2009) 137 595–617

vertebrate genomes; mRNAs with a 3 0UTR that containsmultiple binding sites occur much less frequently.

Efforts to identify in vivo mRNA targets of mammalianDAZL family proteins utilized varied approaches buthave been mainly restricted to mDAZL (reviewed in(Reynolds & Cooke 2005)) and have identified multipleputative targets (Jiao et al. 2002, Reynolds et al. 2005,2007). These studies focused on testis-expressed mRNAsbut it is likely that at least a subset of the identifiedmRNAs are regulated during oogenesis.

Mouse vasa homolog (Mvh) and Sycp3 mRNAs wereisolated by co-immunoprecipitation with mDAZL fromUV-crosslinked mouse testis extracts (Reynolds et al.2005, 2007). The male phenotypes of Mvh null andSycp3 null mice correspond to distinct blocks in meioticprophase I, reminiscent of the final block in the mDazlnull mice (Tanaka et al. 2000, Yuan et al. 2000).Combinations of in vitro RNA-binding assays andtranslation reporter assays in injected X. laevis oocytesindicated that mDAZL could regulate the translation ofreporters containing the 3 0UTR from these mRNAs.At least for SYCP3, stimulation was dependent on thepresence of intact mDAZL-binding sites. Importantly,MVH and SYCP3 protein levels were significantlyreduced in the surviving germ cells of mDazl null testis(5 or 7 days post-partum respectively), implicating thesemRNAs as the first physiological targets of a vertebrateDAZL protein (Reynolds et al. 2005, 2007).

However, MVH expression is dispensable for normalmouse oogenesis (Tanaka et al. 2000). Thus, while MvhmRNA may be regulated by mDAZL in oocytes, the lossof its translation does not underlie the defectiveoogenesis in mDazl null mice. In Sycp3 null femalemice, oogenesis and folliculogenesis progress tocompletion but the mature oocytes exhibit a highdegree of aneuploidy and a resultant failure ofembryogenesis (Yuan et al. 2002). The lack of acorrelation between the female phenotypes of themDazl null and Sycp3 null mice indicates thatdefective Sycp3 translation is not the primary lesionleading to the loss of germ cells in mDazl null females,but does not exclude DAZL-mediated Sycp3 translationcontributing to the production of viable femalegametes. It would be desirable to generate a con-ditional mDazl null mouse, to induce mDAZL defici-ency during later stages of oogenesis and therebyfacilitate the study of the roles of putative mDAZLtarget mRNAs; for instance during the diplotene arrestand maturation of oocytes.

Studies aimed at identifying human mRNA targets ofDAZL family proteins have proposed CDC25 andsevere depolymerization of actin 1 (SDAD1) asputative targets. Cdc25 was first suggested as a Bouletarget mRNA in Drosophila (Maines & Wasserman1999), subsequently as a mDAZL target mRNA inmouse (Venables et al. 2001, Jiao et al. 2002) andCDC25 protein was absent in testicular biopsies from




men with reduced hBOULE expression (Luetjens et al.2004). However, studies have disagreed regardingwhich CDC25 isoform is the putative target mRNAand regarding the location of the binding sequence forDAZL family members within the mRNA, althoughthis disparity may be, in part, the result of speciesdifferences (Maines & Wasserman 1999, Venableset al. 2001, Jiao et al. 2002, Maegawa et al. 2002).Nonetheless, the potential for DAZL-mediatedregulation of CDC25 translation warrants furtherstudy since Cdc25b null female mice are sterile(Lincoln et al. 2002). SDAD1 was identified in ayeast-based screen that recovered mRNAs associatedwith both human PUM2 (hPUM2) and hDAZL (Foxet al. 2005). hDAZL binds to the SDAD1 mRNA 3 0UTRin vitro but, to date, it is not known whether hDAZLregulates SDAD1 mRNA translation in vitro or in vivo.

To date, no vertebrate DAZL mRNA targets have beendefinitively identified in female germ cells and aconcerted effort is required to redress this lack ofknowledge, in order to better understand the role ofhDAZL in human fertility. However, a bona fide DAZLfamily target mRNA with a role in oogenesis has beenidentified in the hermaphroditic worm C. elegans.Translation of FBF mRNAs, which encode proteinsrequired for germline stem cell maintenance and thesperm-oocyte switch, is enhanced by DAZ-1 bindingduring the switch to oogenesis (Otori et al. 2006).

In vertebrates, a potential oocyte target mRNA wassuggested in X. laevis oocytes. RINGO/SPY activatesCDK2, initiating a phosphorylation cascade that iscritical for oocyte maturation (Box 2 and Fig. 2).In prophase I-arrested oocytes RINGO/SPY mRNA istranslationally repressed by a complex containingPUM2 (xPUM2), xDAZL and ePABP. RINGO/SPYtranslation is required prior to the onset of CPE-mediated cytoplasmic polyadenylation and its mRNAdoes not contain CPEs. When maturation is triggeredxPUM2 exits the complex and RINGO/SPY translationis activated, presumably in an ePABP and/or xDAZL-dependent manner (Padmanabhan & Richter 2006).A demonstration that RINGO/SPY translational activa-tion is mediated by xDAZL in X. laevis would raisethe possibility that it is similarly regulated duringmammalian oogenesis.

Mechanism of DAZL-mediated translationalstimulation

The observation that multiple DAZL family membersfrom a variety of species can stimulate the translation ofspecific mRNAs suggests that this conserved function iscritical to their roles in oogenesis (Collier et al. 2005,Reynolds et al. 2007). Therefore, it is important tounderstand the mechanism(s) by which they stimulatetranslation. Since the function of mRNA-specific trans-lational activators is poorly understood, few paradigms


exist for the function of DAZL family proteins. Sucrosegradient analysis of translation intermediates frommicroinjected X. laevis oocytes revealed that mDAZLregulates translation initiation (Collier et al. 2005),similar to characterized mRNA-specific repressors;thereby raising the possibility that DAZL influencesinitiation factor function.

A conserved interaction between DAZL familymembers and both PABP1 and ePABP was detectedin vivo (Collier et al. 2005; Fig. 6B). Importantly, severallines of evidence demonstrated protein-proteininteractions between DAZL family members andPABP1 or ePABP (Table 2), rather than simply detectingDAZL and PABP associated with the same mRNA,thereby implicating PABP1 and ePABP as attractivecandidates for DAZL-regulated initiation factors.

Two further observations support the hypothesis thata PABP–DAZL interaction is critical for DAZL familyproteins to stimulate translation. First, deletion of thePABP interaction site, but not other regions, withinmDAZL completely abrogated its ability to stimulatetranslation (Collier et al. 2005). This loss of activity wasindependent of mDAZL RNA-binding activity since theprotein was artificially tethered to the reporter mRNA.Second, the relative ability of mDAZL to activatereporter mRNA translation is reduced when the mRNAis polyadenylated; indicating that the amplitude ofmDAZL-mediated translational activation is alteredwhen the mRNA can recruit multiple PABPs via thepoly(A) tail (Collier et al. 2005). However, thedetermination of the extent to which PABPs arerequired for DAZL function in vertebrate oogenesiswill require the genetic confirmation of a role for theDAZL-PABP interaction. Interestingly, the C. elegansDAZ-1 loss-of-function phenotype is highly reminiscentof the C. elegans PAB-1-deficient phenotype, bothexhibiting early meiotic blocks in the female germline(Maciejowski et al. 2005, Maruyama et al. 2005).

The apparent requirement for an interaction withPABP, a known translation initiation factor, provides amodel for DAZL function. In this model, DAZL boundto a 3 0UTR recruits PABP which, in turn, interacts with5 0 end-associated factors to enhance the assembly ofthe closed-loop mRNP conformation (Fig. 7, see alsoFig. 3B). The interaction of DAZL with the C-terminalregion of PABP should permit PABP to simultaneouslyscaffold interactions with the key factors required topromote translation initiation, such as eIF4G (Fig. 5B).However, it remains to be examined whether PABPmaintains all its interactions with translation factorswhen complexed with DAZL. Nor is it known whichstep(s) of initiation DAZL regulates as PABP haspleiotropic effects on initiation. Moreover, this modeldoes not rule out the possibility that other DAZL-interacting factors may modulate or co-operate withPABP to promote DAZL-mediated stimulation.

Reproduction (2009) 137 595–617


Table 2 Available evidence for best-characterized DAZL-interacting proteins.

Protein Interacts withInteractionmapped? Methods used RNase used? References

Hs DAZL Hs DAZ Yes Y2H, GSTlys Tsui et al. (2000a) and Moore et al. (2003)Xl PABP1 No Y2H Collier et al. (2005)Hs DAZAP1 Yes Y2H, GSTlys RNase A Tsui et al. (2000a, 2000b)Hs DAZAP2 Yes Y2H, GSTlys Tsui et al. (2000a, 2000b)Hs BOULE Yes Y2H Urano et al. (2005)Hs PUM2 No Y2H* Moore et al. (2003)

Mm DAZL Xl PABP1 Yes Y2H Collier et al. (2005)Xl ePABP Yes oCoIP RNase 1 Collier et al. (2005)Mm Dlc Yes Y2H, GSTlys, oCoIP, eCoIP Lee et al. (2006)Mm DAZL Yes Y2H, GST Ruggiu & Cooke (2000)Hs DAZ No Y2H Ruggiu & Cooke (2000)

Hs DAZ Hs DAZ Yes GSTlys Tsui et al. (2000a, 2000b)Hs PUM2 Yes Y2H, oCoIP Moore et al. (2003)Hs DZIP1 Yes Y2H, oCoIP Moore et al. (2003, 2004)Hs DAZAP1 Yes Y2H, GSTlys RNase A Tsui et al. (2000a)Hs DAZAP2 Yes Y2H, GSTlys Tsui et al. (2000a)Hs DZIP2 No Y2H, oCoIP Moore et al. (2003, 2004)Xl PABP1 No Y2H Collier et al. (2005)Hs BOULE No Y2H Moore et al. (2003)

Hs BOULE Hs BOULE Yes Y2H, oCoIP Urano et al. (2005)Hs PUM2 Yes Y2H, oCoIP Urano et al. (2005)Xl PABP1 No Y2H Collier et al. (2005)

Xl DAZL Xl PABP1 No Y2H, oCoIP, eCoIP RNase 1 Collier et al. (2005)Xl ePABP No eCoIP, Y2H RNase 1 Collier et al. (2005)Xl PUM2 No oCoIP RNase A Padmanabhan & Richter (2006)

Ce DAZ-1 Ce CPB-3 No Y2H, oCoIP, IHC Hasegawa et al. (2006)

Yeast two-hybrid (Y2H) analysis is frequently provided as evidence of a direct protein interaction but interactions can, in some cases, be bridged byendogenous yeast proteins or RNA. * indicates yeast-based screen with multiple components. GST pull-downs (GST) utilizing recombinant proteinspurified to homogeneity indicate direct protein–protein interactions. When recombinant proteins are used to pull-down proteins from cell lysates orcell-free translation extracts (GSTlys) the use of RNase is required to distinguish between direct- and fortuitous RNA-mediated interactions. Likewise,other proteins can also bridge such interactions. Co-immunoprecipitation experiments are divided into those examining endogenous proteins(eCo-IP) or studies utilizing overexpressed proteins (oCo-IP). Co-immunoprecipitations performed in the absence of RNase cannot distinguishwhether a complex is mediated by protein–protein interactions or whether the proteins are merely bound to different parts of the same mRNA.Immunohistochemistry (IHC) can reveal the presence of potential partners within the same cell or cellular compartment, thus raising the possibilityof interaction, but does not provide any evidence of interaction. RNase 1 cuts between all nucleotides and so liberates PABPs from poly(A) RNA.RNase A cuts 3 0 to uridine and cytosine and will leave stretches of RNA including poly(A) tails intact. Interactions are only listed once.


DAZL interacting proteins and other roles of DAZL inpost-transcriptional regulation in the cytoplasm

The recently defined role of DAZL family proteins astranslational activators does not preclude them fromhaving additional translational/non-translational func-tions within the cytoplasm and/or nucleus. Indeed, manymRNA-binding proteins interact with different partnersin the same cell, as well as in different cell types ordevelopmental stages, to perform different functions(Abaza & Gebauer 2008). Thus, the composition ofDAZL-containing complexes may change throughoutgametogenesis in order to enable reprogramming ofprotein synthesis.

A number of DAZL family-interacting proteins havebeen identified (Fig. 6B and Table 2) including PUM2,dynein light chain (DLC) and DAZL-associated protein 1(DAZAP-1). The previously characterized molecular

Reproduction (2009) 137 595–617

functions of these proteins (Table 3) may provide insightinto their potential roles in DAZL-mediated regulation.However, little is actually known regarding their in vivofunctions with DAZL. Additional partners are alsodescribed for which no obvious functional role inDAZL-mediated regulation can be ascribed: DAZAP-2(Tsui et al. 2000a, 2000b) and DZIP1, 2 and 3 (Mooreet al. 2004). Table 2 summarises the evidence for thebest-characterized interactions with DAZL proteins.

The PUF family protein PUM2 is expressed in human,mouse and X. laevis oocytes and is also highly expressedin human and mouse PGCs, gonocytes and spermata-gonia and at lower levels in primary and secondaryspermatocytes (Moore et al. 2003, Xu et al. 2007).xPUM2 acts as a translational repressor via binding toPBEs (see sequences required for cytoplasmic polyade-nylation), and also contributes to regulated cytoplasmic



ADAZL

PABP1

AUG

UAG

3

4G

4E

PAIP1

4A

4B

Figure 7 Model for DAZL-mediated translational stimulation of targetmRNAs. DAZL family members bind to 3 0UTR sequence elements anddirectly recruit PABP molecules via protein-protein interactions. Likecytoplasmic polyadenylation, this is predicted to increase end-to-endcomplex formation (see Fig. 3B), leading to enhanced ribosomal subunitrecruitment (not shown). However, the molecular mechanism by whichthis is achieved remains to be determined and it is unclear whetherDAZL-bound PABP maintains all its translation factor interactions.While this activation does not require changes in polyadenylation, insome cases polyadenylation may contribute to increased activation of aDAZL-bound target mRNA or contribute to the maintenance of itsactivated state (Fig. 4). Target mRNAs bound by multiple DAZLmolecules may recruit multiple PABPs resulting in a higher degree oftranslational activation. PABPs may also be bound to the (short) poly(A)tail. Numbered factors, eIFs. Filled black circle, 5 0cap. Dotted linesdenote lack of interaction between overlapping factors.


polyadenylation (Table 3) (Nakahata et al. 2001, 2003).In immature X. laevis oocytes xPUM2 regulates RINGO/SPY translation (see mRNA targets), and interacts withxDAZL that is also bound to this mRNA (Padmanabhan &Richter 2006). The association of xPUM2, xDAZL andePABP with translationally repressed RINGO/SPY mRNAled to the proposal that xPUM2 may interfere with thetranslation activation function(s) of xDAZL/ePABP.However, the polyadenylation status of RINGO/SPYmRNA has not been determined and xPUM2 may act by

Table 3 Molecular functions of DAZL interacting proteins.

DAZL-interactingprotein Molecular function(s)

PABPs i) Translation initiation factorii) mRNA stability and deadenylationiii) Translational repressoriv) Nonsense-mediated mRNA decay

PUM2 i) Translational repressor of CPE-dependent translii) Translational activator of CPE-dependent transliii) Direct translational repressor

Dynein light chain i) Transport of mRNAs and proteins along microtu

DAZAP-1 ?

DAZAP-2, DZIP1-3 ?

The described molecular functions are from a variety of species, some of whan interaction with DAZL family members. Experimental evidence for DAZfunction of these proteins when participating in complexes with DAZL.


interfering with the poly(A) tail, rather than xDAZL, asdescribed for Drosophila pumilio (Wreden et al. 1997).It also remains possible that xDAZL is an activecomponent of the repression complex. Thus, it is unclearwhether xPUM2 is repressing the stimulatory action ofxDAZL, whether xDAZL is an active component of therepressive complex or even whether xDAZL is involvedin the subsequent activation of this mRNA.

It has also been suggested that PUM2/DAZLinteractions may affect RNA-binding. Human PUM2(hPUM2) interacts with hBOULE and hDAZL (Mooreet al. 2003, Urano et al. 2005) and it has been suggestedthat hPUM2 may aid the recruitment of these proteinsto mRNAs. Consistent with this, RINGO/SPY mRNA inX. laevis does not contain putative DAZL binding sitesin its 3 0UTR (Urano et al. 2005). Conversely, it hasbeen suggested that the mRNA targets of PUM2 may bealtered as a function of the interaction with specificDAZL family proteins (Urano et al. 2005).

While data supports a role for xPUM2 in X. laevisoogenesis (Padmanabhan & Richter 2006), it remainsunclear whether any of its functions are mediated viaxDAZL. Genetic evidence for a physiologically import-ant role in mammalian oogenesis is absent since Pum2null mice are fertile. However, male mice do display asignificant reduction in testis size (Xu et al. 2007) and thelack of ovarian phenotype could be due to a level ofredundancy with mouse PUM1. Thus, further work isrequired to clarify the functional and physiologicalconsequences of DAZL–PUM2 interactions.

mDAZL also interacts in vitro with dynein light chain,a component of the microtubule-associated dynein–dynactin motor complex which has an important role inmRNA localization (Lee et al. 2006). Indeed, whenmDAZL was ectopically expressed in cultured somaticcells, it was able to regulate the cytoplasmic distributionof reporter mRNAs containing either the CDC25C (but

Experimental evidence forDAZL-dependent function

i) Tethered function assays; Collier et al. (2005)

ation –ation

bules i) Intracellular mRNA transport studies; Lee et al. (2006)

–

–

ich have multiple family members. These functions are independent ofL-mediated function refers to experiments aimed at directly testing the

Reproduction (2009) 137 595–617



not CDC25A), TPX1 or MVH 3 0UTRs (see DAZL targets).While mRNA localization is heavily linked to trans-lational control in organisms whose oocytes containgerm-plasm, the role of this process in mammalianoocytes is less well explored. It would be interesting todetermine the extent of mDAZL involvement in thelocalization of mRNAs that it subsequently activatesin vivo. Moreover, localized mRNAs are normallyrepressed prior to their translation, which could involvepartners such as PUM2 or even DAZAP-1.

The mRNA-binding protein DAZAP-1 (also known asproline-rich protein (PRRP)) is widely expressed in humanand mouse, with high-level expression in testis (Tsui et al.2000a, 2000b, Dai et al. 2001, Kurihara et al. 2004, Horiet al. 2005, Pan et al. 2005). Where investigated, DAZAP-1 appears to interact with DAZ motifs, a region importantfor the polysomal localization of DAZL, suggesting that itcould contribute to DAZL-activated translation ormodulate this function by preventing the binding ofother proteins to this region. Interestingly, in this regard,human DAZAP-1-associated DAZ does not interact withPABP in vitro (Morton et al. 2006), although the minimalbinding site for PABP does not include the DAZ motif.Thus, while no role in translation has been establishedfor DAZAP-1, these studies raise the possibility thatDAZAP-1 may repress, activate or modulate DAZLfamily-mediated translation.

DAZAP-1 may also have other roles. Human DAZAP-1is a nucleocytoplasmic shuttling protein (Lin & Yen 2006)and its X. laevis ortholog, PRRP, associates with mRNAlocalization elements in Vg1 and VegT mRNAs (Zhaoet al. 2001). However, it is unclear whether DAZAP-1plays an active role in mRNA localization and its role onthese mRNAs appears to be independent of xDAZL. Thewide expression pattern of DAZAP-1 also suggests that itfrequently functions independently of DAZL proteins.DAZAP-1 is a substrate of the ERK2 kinase, raising thepossibility that its interaction with protein partners maybe subject to regulation by phosphorylation. Consistentwith this notion, phosphomimetic forms of humanDAZAP-1 exhibit reduced DAZ binding activity in vitro(Morton et al. 2006).

Recently, Hsu et al. (2008) showed that most Dazap-1null (or hypomorphic) mice die perinatally due togrowth defects, indicating roles for DAZAP-1 in normalgrowth and development. The few surviving DAZAP-1hypomorphic male progeny have no post-pachytenespermatocytes indicating a role for DAZAP-1 in sperma-togenesis. However, Dazap-1 null mice exhibit apparentlynormal oogenesis, indicating that an mDAZL–DAZAP-1interaction is not an absolute requirement for mouseoogenesis, although DAZAP-1 is normally expressed inthe ovary.

Finally, a role for DAZL family proteins in regulatingmRNA stability may be indicated by the decrease inabundance of a subset of mRNAs in mDazl null mousetestis (Maratou et al. 2004). However, these changes in

Reproduction (2009) 137 595–617

mRNA abundance may be a direct consequence of germcell loss or an indirect consequence of mDAZLdeficiency. Nonetheless, changes in mRNA stability areoften linked to changes in mRNA translation. Thus,DAZL-mediated changes in the translation of targetmRNAs may alter their stability, but no evidencecurrently supports a distinct role in mRNA stability.

A model for DAZL as a translational activator of mRNAsduring oogenesis

Observations in C. elegans and mouse suggest that DAZLmay function at various defined times during gameto-genesis (Collier et al. 2005, Reynolds & Cooke 2005,Reynolds et al. 2007). DAZL function may change duringthe complex, multistage oogenesis program in order toregulate different aspects of mRNA metabolism; changeswhich may depend on protein partner interactions.However, to date, only a role in activating mRNAtranslation has been demonstrated. Cytoplasmic poly-adenylation is an important mechanism for regulatingthe translation of a large number of oligoadenylatedmRNAs, but not all mRNAs appear to require poly-adenylation for translational activation (Radford et al.2008). While loss of repression could contribute to theactivation of these mRNAs, it appears likely thatactivating proteins may be required to compensate fortheir short poly(A) tail status (Wang et al. 1999). SinceDAZL family proteins appear to be mRNA-specifictranslational regulators that stimulate efficient translationof mRNAs with short poly(A) tails, it was suggested thatthe binding of DAZL proteins (and their associatedPABPs) may be one mechanism by which such mRNAscan be activated (Collier et al. 2005, Reynolds et al.2005, 2007; Figs 4 and 7). This would allow theactivation of small subsets of mRNAs, for instance attimes that do not coincide with waves of cytoplasmicpolyadenylation. This hypothesis does not preclude theDAZL-mediated translational activation of mRNAs withlonger poly(A) tails but, rather, predicts that themagnitude of their stimulation would be reduced.

It seems unlikely that there is no link between DAZL-mediated translational activation and cytoplasmic poly-adenylation. Indeed, the cytoplasmic polyadenylation ofsome DAZL family target mRNAs may be required fortheir maximal translational activation (Fig. 4). Forinstance, translation of Sycp3 is attenuated in bothmDazl null and Cpeb null mouse germ cells. Sycp3mRNA contains both mDAZL binding sites and a CPEand has a short poly(A) tail (w20 As) in Cpeb nulloocytes, indicating that CPEB-dependent cytoplasmicpolyadenylation is required for full translational acti-vation of SYCP3 expression. The mDAZL binding site(s)in the Sycp3 3 0UTR overlaps with the CPE and, althoughit has not been formally excluded that both sites can beoccupied simultaneously, the inability to co-immuno-precipitate CPEB and mDAZL makes this unlikely




(Reynolds et al. 2007). The translation of Sycp3 may beinitially activated by mDAZL which is then eitherreplaced or further activated by CPEB to maintain and/or promote Sycp3 translation. In contrast, in C. elegans,an interaction between the BOULE ortholog DAZ-1 andthe CPEB ortholog CPB-3 has been detected (Hasegawaet al. 2006) raising the possibility that DAZL familyproteins can activate translation both directly via DAZL/PABP interactions and indirectly by facilitating cyto-plasmic polyadenylation. However, no DAZL familymembers have yet been shown to promote polyadenyla-tion and, given the dual role of CPEB in repression aswell as polyadenylation, the functional consequences ofthis interaction remain unclear. Greater insight into themolecular mechanisms of DAZL action and its relation-ship with other protein complexes will be required tounderstand how regulatory elements and associatedfactors, both individually and in combination, functionto achieve the changes in gene expression required tocomplete oogenesis.

Perspectives

While other putative functions remain to be elucidated,it appears that DAZL proteins fulfil their roles inoogenesis, at least in part, by activating the translationof target mRNAs. However, it is appealing to speculatethat DAZL may also participate in complexes thatlocalise and repress target mRNAs prior to theiractivation at defined times during oogenesis. The extentto which other family members may share thesefunctional properties also remains to be determined,but while DAZ and BOULE mRNA targets may overlapwith those of DAZL in spermatogenesis this appearsunlikely in vertebrate oocytes.

A greater understanding of the roles of DAZL proteinsduring vertebrate oogenesis is hampered by the currentdearth of identified female DAZL target mRNAs,although further study of RINGO/SPY in X. laevisappears warranted. This mRNA is activated early inmaturation, lacks CPEs, is bound by complexescontaining xDAZL and is critical for oocyte maturation.The DAZL binding consensus sequence needs to befurther defined prior to its use in bioinformatic searchesfor putative target mRNAs. Moreover, these searches mayrequire a clearer understanding of how these elementsfunction together with, or independently of, othercontrol sequences such as CPEs and PBEs. Furthermore,it will be important to determine whether DAZL-interacting proteins alter the RNA-binding specificity ofDAZL proteins. Ultimately, the functional definition andin vivo validation of DAZL target mRNAs will require acombination of molecular and genetic studies in modelssuch as Xenopus and mouse.

In oogenesis, DAZL-mediated translation may pre-dominantly play a role in the activation of mRNAs withshort poly(A) tails. Consistent with this, Sycp3 has a


short poly(A) tail in CPEB-deficient mice. Whilemolecular studies support an important role of PABP inDAZL-mediated activation, the physiological role of thisinteraction remains to be clarified. The targeted mutationof the PABP-binding region of DAZL by geneticallydisrupting this interaction in a model organism, such asthe mouse, would determine the extent to which DAZLutilises PABP for translational activation. Such a modelwould also provide insight into some of the roles ofPABPs during gametogenesis. Ultimately, furtherdefinition of the molecular mechanism of DAZL-mediated stimulation may require a detailed under-standing of the mechanisms by which PABP activates theinitiation pathway and which of these activities ismodulated by DAZL.

However, PABPs are also likely to have other essentialroles in oogenesis, including the translational activationof cytoplasmically polyadenylated mRNAs and bothdefault and adenosine/uridine-rich element (ARE)-mediated deadenylation of mRNAs. Expression ofmultiple PABPs within certain cell types, includingX. laevis oocytes, raises the possibility that PABPs haveboth overlapping and distinct roles during oogenesis. It istherefore tempting to speculate that PABPs may sharesome of their poly(A)-mediated functions but differ intheir ability to control specific mRNAs, for instance viaproteins such as DAZL. Generation of PABP-deficientvertebrate models will be crucial to increase ourunderstanding of the molecular and physiological rolesof this multi-functional protein family in gametogenesis.Moreover, it is clear that the study of PABPs will berelevant to other areas of reproductive and non-reproductive biology.

The study of DAZL function may also have unforeseenimplications because DAZL-mediated recruitment ofPABP in a poly(A)-independent manner raises thequestion of how many other 3 0UTR-binding proteinsfunction analogously. The idea that such a mechanismmay be more widespread is supported by the recentobservation that proteins such as BRCA1 similarly utilisePABP (Dizin et al. 2006); thus, studies of DAZL mayprovide a paradigm for understanding the function ofother translational activators in oocytes as well as othercell types.

In conclusion, defining the roles, mechanisms andRNA targets of these interrelated DAZL and PABPfamilies in gametogenesis may ultimately identifypotential sites of therapeutic intervention in infertility.However, it will be important to ascertain to whatextent these animal studies accurately model humanreproductive disorders.

Declaration of interest

The authors declare that there is no conflict of interest thatcould be perceived as prejudicing the impartiality of theresearch reported.

Reproduction (2009) 137 595–617



Funding

N K G is the recipient of MRC GIA and a MRC Senior Non-Clinical Fellowship and Wellcome Trust project grant funding.

Acknowledgements

We would like to thank Brian Collier for his initial contributionsto this review, other members of the laboratory for their helpfulcomments and Gavin S Wilkie (G S W) for the use of hisunpublished data. We are grateful to Ian Adams, RichardAnderson, Andrew Childs, Howard Cooke, Henry Jabbour,Alan McNeilly, Philippa Saunders and Tom Van Agtmael fortheir critical reading of the manuscript. We would like toapologise to those whose work we have not been able toinclude, discuss in sufficient detail or cite as a primaryreference due to space constraints.

References

Abaza I & Gebauer F 2008 Trading translation with RNA-binding proteins.RNA 14 404–409.

Albrecht M & Lengauer T 2004 Survey on the PABC recognition motifPAM2. Biochemical and Biophysical Research Communications 316129–138.

Amrani N, Ghosh S, Mangus DA & Jacobson A 2008 Translation factorspromote the formation of two states of the closed-loop mRNP. Nature453 1276–1280.

Anderson RA, Fulton N, Cowan G, Coutts S & Saunders PT 2007 Conservedand divergent patterns of expression of DAZL, VASA and OCT4 in thegerm cells of the human fetal ovary and testis. BMC DevelopmentalBiology 7 136.

Barkoff A, Ballantyne S & Wickens M 1998 Meiotic maturation in Xenopusrequires polyadenylation of multiple mRNAs. EMBO Journal 173168–3175.

Belloc E, Pique M & Mendez R 2008 Sequential waves of polyadenylationand deadenylation define a translation circuit that drives meioticprogression. Biochemical Society Transactions 36 665–670.

Blanco P, Sargent CA, Boucher CA, Howell G, Ross M & Affara NA 2001 Anovel poly(A)-binding protein gene (PABPC5) maps to an X-specificsubinterval in the Xq21.3/Yp11.2 homology block of the human sexchromosomes. Genomics 74 1–11.

Bownes M & Pathirana S 2003 Comparative aspects of oogenesis. InBiology and Pathology of the Oocyte, pp 53–64. Eds AO Trounson & RGGosden. Cambridge: Cambridge University Press.

Bushell M, WoodW, Carpenter G, Pain VM, Morley SJ & Clemens MJ 2001Disruption of the interaction of mammalian protein synthesis eukaryoticinitiation factor 4B with the poly(A)-binding protein by caspase- and viralprotease-mediated cleavages. Journal of Biological Chemistry 27623922–23928.

Charlesworth A, Welk J & MacNicol AM 2000 The temporal control ofWee1 mRNA translation during Xenopus oocyte maturation is regulatedby cytoplasmic polyadenylation elements within the 3 0-untranslatedregion. Developmental Biology 227 706–719.

Charlesworth A, Cox LL & MacNicol AM 2004 Cytoplasmic polyadenyla-tion element (CPE)- and CPE-binding protein (CPEB)-independentmechanisms regulate early class maternal mRNA translational activationin Xenopus oocytes. Journal of Biological Chemistry 279 17650–17659.

Cheng MH, Maines JZ & Wasserman SA 1998 Biphasic subcellularlocalization of the DAZL-related protein boule in Drosophila spermato-genesis. Developmental Biology 204 567–576.

Cheng A, Xiong W, Ferrell JE Jr & Solomon MJ 2005 Identification andcomparative analysis of multiple mammalian Speedy/Ringo proteins.Cell Cycle 4 155–165.

Collier B, Gorgoni B, Loveridge C, Cooke HJ & Gray NK 2005 The DAZLfamily proteins are PABP-binding proteins that regulate translation ingerm cells. EMBO Journal 24 2656–2666.

Reproduction (2009) 137 595–617

Cosson B, Couturier A, Le Guellec R, Moreau J, Chabelskaya S,Zhouravleva G & Philippe M 2002 Characterization of the poly(A)binding proteins expressed during oogenesis and early development ofXenopus laevis. Biology of the Cell 94 217–231.

Cosson B, Braun F, Paillard L, Blackshear P & Beverley Osborne H 2004Identification of a novel Xenopus laevis poly (A) binding protein. Biologyof the Cell 96 519–527.

Craig AW, Haghighat A, Yu AT & Sonenberg N 1998 Interaction ofpolyadenylate-binding protein with the eIF4G homologue PAIPenhances translation. Nature 392 520–523.

Dai T, Vera Y, Salido EC & Yen PH 2001 Characterization of the mouseDazap1 gene encoding an RNA-binding protein that interacts withinfertility factors DAZ and DAZL. BMC Genomics 2 6.

Derry MC, Yanagiya A, Martineau Y & Sonenberg N 2006 Regulation ofpoly(A)-binding protein through PABP-interacting proteins. Cold SpringHarbor Symposia on Quantitative Biology 71 537–543.

Dizin E, Gressier C, Magnard C, Ray H, Decimo D, Ohlmann T & DallaVenezia N 2006 BRCA1 interacts with poly(A)-binding protein:implication of BRCA1 in translation regulation. Journal of BiologicalChemistry 281 24236–24246.

Dorfman DM, Genest DR & Reijo Pera RA 1999 Human DAZL1 encodes acandidate fertility factor in women that localizes to the prenatal andpostnatal germ cells. Human Reproduction 14 2531–2536.

Eberhart CG, Maines JZ & Wasserman SA 1996 Meiotic cell cyclerequirement for a fly homologue of human deleted in azoospermia.Nature 381 783–785.

Fassnacht W, Mempel A, Strowitzki T & Vogt PH 2006 Premature ovarianfailure (POF) syndrome: towards the molecular clinical analysis of itsgenetic complexity. Current Medicinal Chemistry 13 1397–1410.

Feral C, Guellaen G & Pawlak A 2001 Human testis expresses a specificpoly(A)-binding protein. Nucleic Acids Research 29 1872–1883.

Fox CA & Wickens M 1990 Poly(A) removal during oocyte maturation: adefault reaction selectively prevented by specific sequences in the 3 0

UTR of certain maternal mRNAs. Genes and Development 4 2287–2298.Fox CA, Sheets MD &Wickens MP 1989 Poly(A) addition during maturation

of frog oocytes: distinct nuclear and cytoplasmic activities and regulationby the sequence UUUUUAU. Genes and Development 3 2151–2162.

Fox M, Urano J & Reijo Pera RA 2005 Identification and characterization ofRNA sequences to which human PUMILIO-2 (PUM2) and deleted inazoospermia-like (DAZL) bind. Genomics 85 92–105.

Gallie DR 1998 A tale of two termini: a functional interaction between thetermini of an mRNA is a prerequisite for efficient translation initiation.Gene 216 1–11.

Gallie DR, Ling J, Niepel M, Morley SJ & Pain VM 2000 "The role of 5 0-leader length, secondary structure and PABP concentration on cap andpoly(A) tail function during translation in Xenopus oocytes. NucleicAcids Research 28 2943–2953.

Gebauer F, Xu W, Cooper GM & Richter JD 1994 Translational control bycytoplasmic polyadenylation of c-mos mRNA is necessary for oocytematuration in the mouse. EMBO Journal 13 5712–5720.

Good PJ, Abler L, Herring D & Sheets MD 2004 Xenopus embryonicpoly(A) binding protein 2 (ePABP2) defines a new family of cytoplasmicpoly(A) binding proteins expressed during the early stages of vertebratedevelopment. Genesis 38 166–175.

Gorgoni B & Gray NK 2004 The roles of cytoplasmic poly(A)-bindingproteins in regulating gene expression: a developmental perspective.Briefings in Functional Genomics & Proteomics 3 125–141.

Gray NK & Wickens M 1998 Control of translation initiation in animals.Annual Review of Cell and Developmental Biology 14 399–458.

Gray NK, Coller JM, Dickson KS & Wickens M 2000 Multiple portions ofpoly(A)-binding protein stimulate translation in vivo. EMBO Journal 194723–4733.

GuW, Kwon Y, Oko R, Hermo L & Hecht NB 1995 Poly (A) binding proteinis bound to both stored and polysomal mRNAs in the mammalian testis.Molecular Reproduction and Development 40 273–285.

Guzeloglu-Kayisli O, Pauli S, Demir H, Lalioti MD, Sakkas D & Seli E 2008Identification and characterization of human embryonic poly(A) bindingprotein (EPAB). Molecular Human Reproduction 14 581–588.

Habermann B, Mi HF, Edelmann A, Bohring C, Backert IT, Kiesewetter F,Aumuller G & Vogt PH 1998 DAZ (deleted in azoospermia) genesencode proteins located in human late spermatids and in sperm tails.Human Reproduction 13 363–369.




Hasegawa E, Karashima T, Sumiyoshi E & Yamamoto M 2006 C. elegansCPB-3 interacts with DAZ-1 and functions in multiple steps of germlinedevelopment. Developmental Biology 295 689–699.

Hoopfer ED, Penton A, Watts RJ & Luo L 2008 Genomic analysis ofDrosophila neuronal remodeling: a role for the RNA-binding proteinBoule as a negative regulator of axon pruning. Journal of Neuroscience28 6092–6103.

Hori T, Taguchi Y, Uesugi S & Kurihara Y 2005 The RNA ligands for mouseproline-rich RNA-binding protein (mouse Prrp) contain two consensussequences in separate loop structure. Nucleic Acids Research 33190–200.

Hoshino S, Imai M, Kobayashi T, Uchida N & Katada T 1999 The eukaryoticpolypeptide chain releasing factor (eRF3/GSPT) carrying the translationtermination signal to the 3 0-poly(A) tail of mRNA. Direct association oferf3/GSPT with polyadenylate-binding protein. Journal of BiologicalChemistry 274 16677–16680.

Houston DW & King ML 2000 A critical role for Xdazl, a germ plasm-localized RNA, in the differentiation of primordial germ cells inXenopus. Development 127 447–456.

Houston DW, Zhang J, Maines JZ, Wasserman SA & King ML 1998 AXenopus DAZ-like gene encodes an RNA component of germ plasm andis a functional homologue of Drosophila boule. Development 125171–180.

Hsu LC, Chen HY, Lin YW, Chu WC, Lin MJ, Yan YT & Yen PH 2008DAZAP1, an hnRNP protein, is required for normal growth andspermatogenesis in mice. RNA 14 1814–1822.

Huang YS & Richter JD 2004 Regulation of local mRNA translation. CurrentOpinion in Cell Biology 16 308–313.

Huang WJ, Lin YW, Hsiao KN, Eilber KS, Salido EC & Yen PH 2008Restricted expression of the human DAZ protein in premeiotic germcells. Human Reproduction 23 1280–1289.

Imataka H, Gradi A & Sonenberg N 1998 A newly identified N-terminalamino acid sequence of human eIF4G binds poly(A)-binding protein andfunctions in poly(A)-dependent translation. EMBO Journal 177480–7489.

Jiao X, Trifillis P & Kiledjian M 2002 Identification of target messenger RNAsubstrates for the murine deleted in azoospermia-like RNA-bindingprotein. Biology of Reproduction 66 475–485.

Joiner ML & Wu CF 2004 Nervous system function for the testis RNA-binding protein boule in Drosophila. Journal of Neurogenetics 18341–363.

Kahvejian A, Roy G & Sonenberg N 2001 The mRNA closed-loop model:the function of PABP and PABP-interacting proteins in mRNA translation.Cold Spring Harbor Symposia on Quantitative Biology 66 293–300.

Kahvejian A, Svitkin YV, Sukarieh R, M’BoutchouMN & Sonenberg N 2005Mammalian poly(A)-binding protein is a eukaryotic translation initiationfactor, which acts via multiple mechanisms. Genes and Development 19104–113.

Karashima T, Sugimoto A & Yamamoto M 2000 Caenorhabditis eleganshomologue of the human azoospermia factor DAZ is required foroogenesis but not for spermatogenesis. Development 127 1069–1079.

Kim JH & Richter JD 2007 RINGO/cdk1 CPEB mediate poly(A) tail stabili-zation and translational regulation by ePAB. Genes and Development 212571–2579.

Kimble J & Crittenden SL 2007 Controls of germline stem cells, entry intomeiosis, and the sperm/oocyte decision in Caenorhabditis elegans.Annual Review of Cell and Developmental Biology 23 405–433.

Kimura M, Ishida K, Kashiwabara S & Baba T 2009 Characterization of twocytoplasmic poly(A)-binding proteins, PABPC1 and PABPC2, in mousespermatogenic cells. Biology of Reproduction 80 545–554.

Kleene KC, Wang MY, Cutler M, Hall C & Shih D 1994 Developmentalexpression of poly(A) binding protein mRNAs during spermatogenesis inthe mouse. Molecular Reproduction and Development 39 355–364.

Kleene KC, Mulligan E, Steiger D, Donohue K & Mastrangelo MA 1998 Themouse gene encoding the testis-specific isoform of poly(A) bindingprotein (Pabp2) is an expressed retroposon: intimations that geneexpression in spermatogenic cells facilitates the creation of new genes.Journal of Molecular Evolution 47 275–281.

Kostova E, Yeung CH, Luetjens CM, Brune M, Nieschlag E & Gromoll J2007 Association of three isoforms of the meiotic BOULE gene withspermatogenic failure in infertile men. Molecular Human Reproduction13 85–93.


Kozlov G, Trempe JF, Khaleghpour K, Kahvejian A, Ekiel I & Gehring K2001 Structure and function of the C-terminal PABC domain of humanpoly(A)-binding protein. PNAS 98 4409–4413.

Kozlov G, De Crescenzo G, Lim NS, Siddiqui N, Fantus D, Kahvejian A,Trempe JF, Elias D, Ekiel I, Sonenberg N et al. 2004 Structural basis ofligand recognition by PABC, a highly specific peptide-binding domainfound in poly(A)-binding protein and a HECT ubiquitin ligase. EMBOJournal 23 272–281.

Kurihara Y, Watanabe H, Kawaguchi A, Hori T, Mishiro K, Ono M,Sawada H & Uesugi S 2004 Dynamic changes in intranuclear andsubcellular localizations of mouse Prrp/DAZAP1 during spermatogen-esis: the necessity of the C-terminal proline-rich region for nuclear importand localization. Archives of Histology and Cytology 67 325–333.

Lee KH, Lee S, Kim B, Chang S, Kim SW, Paick JS & Rhee K 2006 Dazl canbind to dynein motor complex and may play a role in transport of specificmRNAs. EMBO Journal 25 4263–4270.

Lim NS, Kozlov G, Chang TC, Groover O, Siddiqui N, Volpon L, DeCrescenzo G, Shyu AB & Gehring K 2006 Comparative peptide bindingstudies of the PABC domains from the ubiquitin-protein isopeptide ligaseHYD and poly(A)-binding protein. Implications for HYD function.Journal of Biological Chemistry 281 14376–14382.

Lin Y & Page DC 2005 Dazl deficiency leads to embryonic arrest of germcell development in XY C57BL/6 mice. Developmental Biology 288309–316.

Lin YT & Yen PH 2006 A novel nucleocytoplasmic shuttling sequence ofDAZAP1, a testis-abundant RNA-binding protein. RNA 12 1486–1493.

Lin YM, Chen CW, Sun HS, Tsai SJ, Hsu CC, Teng YN, Lin JS & Kuo PL 2001Expression patterns and transcript concentrations of the autosomal DAZLgene in testes of azoospermic men. Molecular Human Reproduction 71015–1022.

Lin YM, Chen CW, Sun HS, Tsai SJ, Lin JS & Kuo PL 2002 Presence of DAZLtranscript and protein in mature human spermatozoa. Fertility andSterility 77 626–629.

Lin Y, Gill ME, Koubova J & Page DC 2008 Germ cell-intrinsic and -extrinsicfactors govern meiotic initiation in mouse embryos. Science 3221685–1687.

Lincoln AJ, Wickramasinghe D, Stein P, Schultz RM, Palko ME, DeMiguel MP, Tessarollo L & Donovan PJ 2002 Cdc25b phosphatase isrequired for resumption of meiosis during oocyte maturation. NatureGenetics 30 446–449.

Luetjens CM, Xu EY, Rejo Pera RA, Kamischke A, Nieschlag E & Gromoll J2004 Association of meiotic arrest with lack of BOULE proteinexpression in infertile men. Journal of Clinical Endocrinology andMetabolism 89 1926–1933.

Maciejowski J, Ahn JH, Cipriani PG, Killian DJ, Chaudhary AL, Lee JI,Voutev R, Johnsen RC, Baillie DL, Gunsalus KC et al. 2005 Autosomalgenes of autosomal/X-linked duplicated gene pairs and germ-lineproliferation in Caenorhabditis elegans. Genetics 169 1997–2011.

Maegawa S, Yamashita M, Yasuda K & Inoue K 2002 Zebrafish DAZ-likeprotein controls translation via the sequence ‘GUUC’. Genes to Cells 7971–984.

Maines JZ & Wasserman SA 1999 Post-transcriptional regulation of themeiotic Cdc25 protein Twine by the Dazl orthologue Boule. Nature CellBiology 1 171–174.

Mandel CR, Bai Y & Tong L 2008 Protein factors in pre-mRNA 3 0-endprocessing. Cellular and Molecular Life Sciences 65 1099–1122.

Maratou K, Forster T, Costa Y, Taggart M, Speed RM, Ireland J, Teague P,Roy D & Cooke HJ 2004 Expression profiling of the developing testis inwild-type and Dazl knockout mice. Molecular Reproduction andDevelopment 67 26–54.

Martineau Y, Derry MC, Wang X, Yanagiya A, Berlanga JJ, Shyu AB,Imataka H, Gehring K & Sonenberg N 2008 Poly(A)-binding protein-interacting protein 1 binds to eukaryotic translation initiation factor 3 tostimulate translation. Molecular and Cellular Biology 28 6658–6667.

Maruyama R, Endo S, Sugimoto A & Yamamoto M 2005 Caenorhabditiselegans DAZ-1 is expressed in proliferating germ cells and directs propernuclear organization and cytoplasmic core formation during oogenesis.Developmental Biology 277 142–154.

Matova N & Cooley L 2001 Comparative aspects of animal oogenesis.Developmental Biology 231 291–320.

Matzuk MM & Lamb DJ 2008 The biology of infertility: research advancesand clinical challenges. Nature Medicine 14 1197–1213.

Reproduction (2009) 137 595–617



McGrew LL, Dworkin-Rastl E, Dworkin MB & Richter JD 1989 Poly(A)elongation during Xenopus oocyte maturation is required for trans-lational recruitment and is mediated by a short sequence element. Genesand Development 3 803–815.

Melo EO, Dhalia R, Martins de Sa C, Standart N & de Melo Neto OP 2003Identification of a C-terminal poly(A)-binding protein (PABP)-PABP inter-action domain: role in cooperative binding to poly(A) and efficient cap distaltranslational repression. Journal of Biological Chemistry278 46357–46368.

Moore FL, Jaruzelska J, Fox MS, Urano J, Firpo MT, Turek PJ, Dorfman DM& Pera RA 2003 Human Pumilio-2 is expressed in embryonic stem cellsand germ cells and interacts with DAZ (deleted in azoospermia) andDAZ-like proteins. PNAS 100 538–543.

Moore FL, Jaruzelska J, Dorfman DM & Reijo-Pera RA 2004 Identificationof a novel gene, DZIP (DAZ-interacting protein), that encodes a proteinthat interacts with DAZ (deleted in azoospermia) and is expressed inembryonic stem cells and germ cells. Genomics 83 834–843.

Morton S, Yang HT, Moleleki N, Campbell DG, Cohen P & Rousseau S 2006Phosphorylation of the ARE-binding protein DAZAP1 by ERK2 induces itsdissociation from DAZ. Biochemical Journal 399 265–273.

Nakahata S, Katsu Y, Mita K, Inoue K, Nagahama Y & Yamashita M 2001Biochemical identification of Xenopus Pumilio as a sequence-specificcyclin B1 mRNA-binding protein that physically interacts with a Nanoshomolog, Xcat-2, and a cytoplasmic polyadenylation element-bindingprotein. Journal of Biological Chemistry 276 20945–20953.

Nakahata S, Kotani T, Mita K, Kawasaki T, Katsu Y, Nagahama Y &Yamashita M 2003 Involvement of Xenopus Pumilio in the translationalregulation that is specific to cyclin B1 mRNA during oocyte maturation.Mechanisms of Development 120 865–880.

Newbury SF 2006 Control of mRNA stability in eukaryotes. BiochemicalSociety Transactions 34 30–34.

Nieuwkoop PD & Faber J 1994 Normal Table of Xenopus laevis (Daudin):A Systematical and Chronological Survey of the Development Fromthe Fertilized Egg Till the End of Metamorphosis. New York: GarlandPublishers.

Nishi S, Hoshi N, Kasahara M, Ishibashi T & Fujimoto S 1999 Existence ofhuman DAZLA protein in the cytoplasm of human oocytes. MolecularHuman Reproduction 5 495–497.

Otori M, Karashima T & Yamamoto M 2006 The Caenorhabditis eleganshomologue of deleted in azoospermia is involved in the sperm/oocyteswitch. Molecular Biology of the Cell 17 3147–3155.

Padmanabhan K & Richter JD 2006 Regulated Pumilio-2 binding controlsRINGO/Spy mRNA translation and CPEB activation. Genes andDevelopment 20 199–209.

Pan HA, Tsai SJ, Chen CW, Lee YC, Lin YM & Kuo PL 2002 Expression ofDAZL protein in the human corpus luteum. Molecular HumanReproduction 8 540–545.

Pan HA, Lin YS, Lee KH, Huang JR, Lin YH & Kuo PL 2005 Expressionpatterns of the DAZ-associated protein DAZAP1 in rat and humanovaries. Fertility and Sterility 84 (suppl 2) 1089–1094.

Pan HA, Lee YC, Teng YN, Tsai SJ, Lin YM & Kuo PL 2008 CDC25 proteinexpression and interaction with DAZL in human corpus luteum. Fertilityand Sterility [in press] DOI: 10.1016/j.fertnstert.2008.09.025.

Piccioni F, Zappavigna V & Verrotti AC 2005 Translational regulationduring oogenesis and early development: the cap-poly(A) tail relation-ship. Comptes Rendus Biologies 328 863–881.

Pique M, Lopez JM, Foissac S, Guigo R & Mendez R 2008 A combinatorialcode for CPE-mediated translational control. Cell 132 434–448.

Prasad CK, Mahadevan M, MacNicol MC & MacNicol AM 2008 Mos 3 0

UTR regulatory differences underlie species-specific temporal patterns ofMos mRNA cytoplasmic polyadenylation and translational recruitmentduring oocyte maturation. Molecular Reproduction and Development 751258–1268.

Racki WJ & Richter JD 2006 CPEB controls oocyte growth and follicledevelopment in the mouse. Development 133 4527–4537.

Radford HE, Meijer HA & de Moor CH 2008 Translational control bycytoplasmic polyadenylation in Xenopus oocytes. Biochimica etBiophysica Acta 1779 217–229.

Reijo R, Lee TY, Salo P, Alagappan R, Brown LG, Rosenberg M, Rozen S,Jaffe T, Straus D, Hovatta O et al. 1995 Diverse spermatogenic defects inhumans caused by Y chromosome deletions encompassing a novel RNA-binding protein gene. Nature Genetics 10 383–393.

Reproduction (2009) 137 595–617

Reijo RA, Dorfman DM, Slee R, Renshaw AA, Loughlin KR, Cooke H &Page DC 2000 DAZ family proteins exist throughout male germ celldevelopment and transit from nucleus to cytoplasm at meiosis in humansand mice. Biology of Reproduction 63 1490–1496.

Reynolds N & Cooke HJ 2005 Role of the DAZ genes in male fertility.Reproductive Biomedicine Online 10 72–80.

Reynolds N, Collier B, Maratou K, Bingham V, Speed RM, Taggart M,Semple CA, Gray NK & Cooke HJ 2005 Dazl binds in vivo to specifictranscripts and can regulate the pre-meiotic translation of Mvh in germcells. Human Molecular Genetics 14 3899–3909.

Reynolds N, Collier B, Bingham V, Gray NK &Cooke HJ 2007 Translation ofthe synaptonemal complex component Sycp3 is enhanced in vivo by thegerm cell specific regulator Dazl. RNA 13 974–981.

Richter JD 2007 CPEB: a life in translation. Trends in Biochemical Sciences32 279–285.

Roy G, De Crescenzo G, Khaleghpour K, Kahvejian A, O’Connor-McCourt M & Sonenberg N 2002 Paip1 interacts with poly(A) bindingprotein through two independent binding motifs. Molecular and CellularBiology 22 3769–3782.

Ruggiu M & Cooke HJ 2000 In Vivo and in vitro analysis ofhomodimerisation activity of the mouse Dazl1 protein. Gene 252119–126.

Ruggiu M, Speed R, Taggart M, McKay SJ, Kilanowski F, Saunders P, Dorin J& Cooke HJ 1997 The mouse Dazla gene encodes a cytoplasmic proteinessential for gametogenesis. Nature 389 73–77.

Ruggiu M, Saunders PT & Cooke HJ 2000 Dynamic subcellular distributionof the DAZL protein is confined to primate male germ cells. Journal ofAndrology 21 470–477.

Sachs AB & Davis RW 1989 The poly(A) binding protein is required forpoly(A) shortening and 60S ribosomal subunit-dependent translationinitiation. Cell 58 857–867.

Sakugawa N, Miyamoto T, Sato H, Ishikawa M, Horikawa M, Hayashi H &Sengoku K 2008 Isolation of the human ePAB and ePABP2 cDNAs andanalysis of the expression patterns. Journal of Assisted Reproduction andGenetics 25 215–221.

Saxena R, Brown LG, Hawkins T, Alagappan RK, Skaletsky H, Reeve MP,Reijo R, Rozen S, Dinulos MB, Disteche CM et al. 1996 The DAZ genecluster on the human Y chromosome arose from an autosomal gene thatwas transposed, repeatedly amplified and pruned. Nature Genetics 14292–299.

Saxena R, de Vries JW, Repping S, Alagappan RK, Skaletsky H, Brown LG,Ma P, Chen E, Hoovers JM & Page DC 2000 Four DAZ genes in twoclusters found in the AZFc region of the human Y chromosome.Genomics 67 256–267.

Scheper GC, van der Knaap MS & Proud CG 2007 Translation matters:protein synthesis defects in inherited disease. Nature Reviews. Genetics8 711–723.

Schmitt A & Nebreda AR 2002 Signalling pathways in oocyte meioticmaturation. Journal of Cell Science 115 2457–2459.

Seli E, Lalioti MD, Flaherty SM, Sakkas D, Terzi N & Steitz JA 2005 Anembryonic poly(A)-binding protein (ePAB) is expressed in mouse oocytesand early preimplantation embryos. PNAS 102 367–372.

Sheets MD, Fox CA, Hunt T, Vande Woude G & Wickens M 1994 The 3 0-untranslated regions of c-mos and cyclin mRNAs stimulate translation byregulating cytoplasmic polyadenylation. Genes and Development 8926–938.

Sheets MD,WuM&Wickens M 1995 Polyadenylation of c-mos mRNA as acontrol point in Xenopus meiotic maturation. Nature 374 511–516.

Sive HL, Grainger RM & Harland RM 1999 Early Development of Xenopuslaevis: A Laboratory Manual, Cold Spring Harbor: Cold Spring HarborLaboratory Press.

Slee R, Grimes B, Speed RM, Taggart M, Maguire SM, Ross A,McGill NI, Saunders PT & Cooke HJ 1999 A human DAZ transgeneconfers partial rescue of the mouse Dazl null phenotype. PNAS 968040–8045.

Spriggs KA, Stoneley M, Bushell M & Willis AE 2008 Re-programming oftranslation following cell stress allows IRES-mediated translation topredominate. Biology of the Cell 100 27–38.

Tanaka SS, Toyooka Y, Akasu R, Katoh-Fukui Y, Nakahara Y, Suzuki R,Yokoyama M & Noce T 2000 The mouse homolog of Drosophila vasa isrequired for the development of male germ cells. Genes andDevelopment 14 841–853.




Tarun SZ Jr, Wells SE, Deardorff JA & Sachs AB 1997 Translation initiationfactor eIF4G mediates in vitro poly(A) tail-dependent translation. PNAS94 9046–9051.

Tay J & Richter JD 2001 Germ cell differentiation and synaptonemalcomplex formation are disrupted in CPEB knockout mice.Developmental Cell 1 201–213.

Tsui S, Dai T, Roettger S, Schempp W, Salido EC & Yen PH 2000aIdentification of two novel proteins that interact with germ-cell-specific RNA-binding proteins DAZ and DAZL1. Genomics 65 266–273.

Tsui S, Dai T, Warren ST, Salido EC & Yen PH 2000b Association of themouse infertility factor DAZL1 with actively translating polyribosomes.Biology of Reproduction 62 1655–1660.

Tung JY, Luetjens CM, Wistuba J, Xu EY, Reijo Pera RA & Gromoll J 2006aEvolutionary comparison of the reproductive genes, DAZL and BOULE,in primates with and without DAZ. Development Genes and Evolution216 158–168.

Tung JY, Rosen MP, Nelson LM, Turek PJ, Witte JS, Cramer DW, Cedars MI& Reijo-Pera RA 2006b Novel missense mutations of the deleted-in-azoospermia-like (DAZL) gene in infertile women and men.Reproductive Biology and Endocrinology 4 40.

Urano J, Fox MS & Reijo Pera RA 2005 Interaction of the conservedmeiotic regulators, BOULE (BOL) and PUMILIO-2 (PUM2). MolecularReproduction and Development 71 290–298.

Varnum SM & Wormington WM 1990 Deadenylation of maternal mRNAsduring Xenopus oocyte maturation does not require specific cis-sequences: a default mechanism for translational control. Genes andDevelopment 4 2278–2286.

Vasudevan S, Seli E & Steitz JA 2006 Metazoan oocyte and early embryodevelopment program: a progression through translation regulatorycascades. Genes and Development 20 138–146.

Venables JP, Ruggiu M & Cooke HJ 2001 The RNA-binding specificity of themouse Dazl protein. Nucleic Acids Research 29 2479–2483.

Voeltz GK, Ongkasuwan J, Standart N & Steitz JA 2001 A novel embryonicpoly(A) binding protein, ePAB, regulates mRNA deadenylation inXenopus egg extracts. Genes and Development 15 774–788.

Wang ZF, Ingledue TC, Dominski Z, Sanchez R & Marzluff WF 1999 TwoXenopus proteins that bind the 3 0 end of histone mRNA: implications fortranslational control of histone synthesis during oogenesis. Molecularand Cellular Biology 19 835–845.

Wang YY, Charlesworth A, Byrd SM, Gregerson R, MacNicol MC &MacNicol AM 2008 A novel mRNA 3 0 untranslated region translationalcontrol sequence regulates Xenopus Wee1 mRNA translation.Developmental Biology 317 454–466.

Wickens M, Bernstein DS, Kimble J & Parker R 2002 A PUF familyportrait: 3 0UTR regulation as a way of life. Trends in Genetics 18150–157.


Wilkie GS, Dickson KS & Gray NK 2003 Regulation of mRNA translation by5 0- and 3 0-UTR-binding factors. Trends in Biochemical Sciences 28182–188.

Wilkie GS, Gautier P, Lawson D & Gray NK 2005 Embryonic poly(A)-binding protein stimulates translation in germ cells. Molecular andCellular Biology 25 2060–2071.

Wormington M, Searfoss AM & Hurney CA 1996 Overexpression ofpoly(A) binding protein prevents maturation-specific deadenylation andtranslational inactivation in Xenopus oocytes. EMBO Journal 15900–909.

Wreden C, Verrotti AC, Schisa JA, Lieberfarb ME & Strickland S 1997Nanos and pumilio establish embryonic polarity in Drosophila bypromoting posterior deadenylation of hunchback mRNA. Development124 3015–3023.

Xu EY, Moore FL & Pera RA 2001 A gene family required for human germcell development evolved from an ancient meiotic gene conserved inmetazoans. PNAS 98 7414–7419.

Xu EY, Lee DF, Klebes A, Turek PJ, Kornberg TB & Reijo Pera RA 2003Human BOULE gene rescues meiotic defects in infertile flies. HumanMolecular Genetics 12 169–175.

Xu EY, Chang R, Salmon NA& Reijo Pera RA 2007 A gene trap mutation of amurine homolog of the Drosophila stem cell factor Pumilio results insmaller testes but does not affect litter size or fertility. MolecularReproduction and Development 74 912–921.

Yamasaki S & Anderson P 2008 Reprogramming mRNA translation duringstress. Current Opinion in Cell Biology 20 222–226.

Yang H, Duckett CS & Lindsten T 1995 iPABP, an inducible poly(A)-bindingprotein detected in activated human T cells. Molecular and CellularBiology 15 6770–6776.

Yuan L, Liu JG, Zhao J, Brundell E, Daneholt B & Hoog C 2000 The murineSCP3 gene is required for synaptonemal complex assembly, chromo-some synapsis, and male fertility. Molecular Cell 5 73–83.

Yuan L, Liu JG, Hoja MR, Wilbertz J, Nordqvist K & Hoog C 2002 Femalegerm cell aneuploidy and embryo death in mice lacking the meiosis-specific protein SCP3. Science 296 1115–1118.

Zhao WM, Jiang C, Kroll TT & Huber PW 2001 A proline-rich protein bindsto the localization element of Xenopus Vg1 mRNA and to ligandsinvolved in actin polymerization. EMBO Journal 20 2315–2325.

Received 17 December 2008

First decision 16 January 2009

Accepted 9 February 2009

Reproduction (2009) 137 595–617


Documents

REPRODUCTION · Translational regulation is important in a wide variety of cell types but may be even more prevalent in germ cells, where periods of transcriptional quiescence necessitate