Developmental Biology 355 (2011) 227–238

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

Lineage specific trimethylation of H3 on lysine 4 during C. elegansearly embryogenesis

Siyao Wang, Kate Fisher, Gino B. Poulin ⁎Faculty of Life Sciences, Michael Smith Building, The University of Manchester, Oxford Road, Manchester, M13 9PT, UK

⁎ Corresponding author.E-mail address: Gino.Poulin@manchester.ac.uk (G.B.

0012-1606/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.ydbio.2011.04.010

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 28 February 2011Accepted 17 April 2011Available online 22 April 2011

Keywords:MethylationMLL complexChromatinH3K4C. elegans

In many organisms early embryogenesis is characterised by a period refractory to transcription. InCaenorhabditis elegans, the one-cell embryo is transcriptionally inactive, but at around eight-cell stagetranscription is activated in the somatic lineage. This model suggests that histone tail modifications associatedwith activation of transcription, such as di- or trimethylation of histone 3 on lysine 4 (H3K4me2/me3) shouldbe enriched in the somatic lineage. Here, we have investigated the deposition of H3K4me3 duringembryogenesis and found that it is more dynamic than anticipated. In the eight-cell stage embryo, H3K4me3deposition is poor in the germline blastomere, as expected, but surprisingly three somatic blastomeres alsoremain poor in H3K4me3. All the other somatic blastomeres show robust deposition of H3K4me3.Interestingly, the three somatic blastomeres poor in H3K4me3 are descendants of the first germlineblastomere, implying an activity that impedes on H3K4me3 deposition in these cells. In contrast, thedeposition of H3K4me2 and H3K27me2/3 is not lineage restricted. Taken together, our data reveal thatH3K4me3 deposition is highly regulated according to the cell lineage involved.

Poulin).

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

One of themost fascinating processes of early embryogenesis is theproduction of cells with two fundamentally different programs: thesomatic program, which produces cells that are destined to die, andthe germline program responsible for immortality of the species.Hence, the germline blastomeres face a dilemma at each cell division:to retain totipotency or to engage the somatic program.

Transcription is highly regulated during early embryogenesis and isimportant to establish the distinction between somatic and germlinefates (Mello et al., 1992, 1996; Seydoux and Strome, 1999). Transcrip-tion transits rapidly from a generally inactive state in all lineages to anactive state specifically in the somatic lineage (Edgar et al., 1994;Seydoux and Dunn, 1997; Seydoux and Fire, 1994). In contrast, thegermline lineage remains poor in transcriptional activity.

An important determinant of somatic versus germline lineage isPIE-1 (Mello et al., 1992). PIE-1 has been shown to segregate in the Plineage and to disappear when P4 undergoes the final symmetricdivision (Mello et al., 1996). PIE-1 is an RNA binding protein that canrepress transcription in human cell culture (Zhang et al., 2003).Deletion of pie-1 causes a dramatic effect on the embryonic lineage;the P2 lineage loses its germline program and acquires a somaticprogram, causing the production of extra intestinal cells (Mello et al.,1996; Seydoux and Strome, 1999). Therefore, PIE-1 is essential to

suppress the somatic program, possibly by general repression oftranscription.

There is strongevidence that chromatinorganisation is important forgermline development. Work on MES-2, MES-3, and MES-6, thePolycomb Group-like complex in Caenorhabditis elegans, has shownthat these prevent degeneration of the germline and sterility throughdeposition of H3K27me2/3 marks (Bender et al., 2004; Capowski et al.,1991; Shin andMello, 2003). It was also shown that defects in erasure ofthe H3K4me2 marks in the Z2 and Z3 cells compromises germlineimmortality (Katz et al., 2009; Schaner et al., 2003; Schaner and Kelly,2006). Furthermore, inactivation of members of the NuRD (nucleosomeremodelling deacetylase) complex, a chromatin remodelling complexinvolved in repression of transcription, causes the ectopic expression ofgermline markers in the soma (Unhavaithaya et al., 2002). Therefore,chromatin organisation is critical to retain the germline program andalso to prevent its expression in the somatic lineage.

Di- or trimethylation marks on H3K4 and H3K27 correlate withactivation of transcription and with repression of transcription,respectively. Consequently, H3K4me2/3 and H3K27me2/3 are mainlyfound at different loci (Cao et al., 2002; Ringrose and Paro, 2004;Ruthenburg et al., 2007). Despite this anti-correlation, a limitednumber of loci have been found co-occupied by H3K4/K27me3marks.These loci tend to be transcriptionally inactive and to encode fordevelopmentally regulated transcription factors ready to be activatedat the appropriate time (Azuara et al., 2006; Bernstein et al., 2006).

Methylation marks are deposited and removed by enzymes. Anumber of these enzymes are part of the MLL3/SET1 complex inhumans, andwewill refer herein to its C. elegans counterpart as theMLL

228 S. Wang et al. / Developmental Biology 355 (2011) 227–238

complex. The MLL complex has a methyltransferase activity targetingH3K4 and a demethylase activity targeting H3K27 (Agger et al., 2007;Fisher et al., 2010; Issaeva et al., 2007; Lan et al., 2007; Lee et al., 2007).The complex can therefore depositmethylmarks onH3K4 (activation oftranscription) and remove methyl marks from H3K27 (relievingrepression of transcription) (Kouzarides, 2007). The MLL complexrequires core components to be effective: WDR-5, ASH-2, and RBBP-5(Ruthenburg et al., 2007). TheMLL complex can also contain at least twomethyltransferases (SET-16 or SET-2) (Fisher et al., 2010; Simonet et al.,2007). All these components are conserved in humans.

Here, we investigated the levels of di- and trimethylation at bothH3K4 and H3K27 sites during early embryogenesis. Since these markscorrelate with either activation or repression of transcription, weanticipated that the germline blastomeres would be poor inH3K4me2/3 and rich in H3K27me2/3 relatively to the somaticblastomeres. Interestingly, only the H3K4me3 mark shows levels ofdeposition that differ between the different lineages. Indeed, thegermline blastomeres and importantly some of the somatic blasto-meres remain poor in H3K4me3.We also show usingmutants that theMLL complex is active at an early stage in both the somatic andgermline blastomeres. Interestingly, both PIE-1 and RBR-2 (ademethylase targeting H3K4) can affect H3K4me3 levels duringearly embryogenesis, but only PIE-1 can prevent H3K4me3 depositionin the germline blastomeres. Taken together, our data show that thedeposition of the H3K4me3 mark is unique and dynamic compared toother epigenetic marks investigated here or reported by others.

Results

H3K4me3 in the early embryo and on chromosome X

The C. elegans early embryo is a good model to study how thesomatic lineage differs from the germline lineage, because theselineages can be tracked in vivo. During early embryogenesis, thelineage that will generate the gametes is called the P lineage (P0, P1,P2, P3, and P4). P1-P4 originate from four successive asymmetric celldivisions that also give rise to four somatic blastomeres (or foundercells) (Sulston et al., 1983). P4 will later divide symmetrically toproduce the primordial germ cells (PGCs), Z2 and Z3 (Fig. 1A and B).At this stage, the germline programme is distinct from the somaticprogramme (Schaner and Kelly, 2006; Seydoux and Strome, 1999).

It was previously shown that transcription in two- and four-cellstage embryos is poor (Edgar et al., 1994; Seydoux and Fire, 1994).Therefore, we investigated whether H3K27me2/me3, the repression-associated marks, and H3K4me2/me3, the activation-associatedmarks, are detectable by immunofluorescence at these stages. Wefound that H3K4me2/me3, and H3K27me2/me3 are detectable marksand appear evenly distributed between the different cells of theembryos at both two- and four-cell stages (Fig. 1C, data not shown,and (Schaner et al., 2003)). Therefore, the methylation marksanalysed herein are all present during early embryogenesis, andeach mark is distributed at comparable levels between cells.

TheH3K4me2,H3K27me2, andH3K27me3markshavebeen studiedduring post-embryonic germline development. During germline devel-opment the X chromosome is transcriptionally inactive (Reinke et al.,2000). Accordingly, the X chromosome is devoid of the H3K4me2mark(Kelly et al., 2002; Reuben and Lin, 2002), and enriched for both theH3K27me2/me3 marks (Bender et al., 2004). We predicted that

Fig. 1. The H3K4me3mark during early embryonic development. (A) Depiction of the early emlineages. The AB lineage is entirely formed of somatic cells. The P lineage produces the germand D (green)), and the primordial germ cells (Z2 and Z3 (blue)). The latter will generate alstages of the embryo. In green are the somatic blastomeres, and in green and blue are the gerof the embryos at two- and four-cell stages. Co-immunostaining for H3K4me3, H3K27me3 anarea at pachytene stage is depleted of the H3K4me3 mark. E) Competition assays usingH3K4me3 antibody is specific. Two-cell stage embryos are presented, but similar data were

H3K4me3 should adopt a similar localisation as H3K4me2, and indeedwe found a chromosomal area devoid of H3K4me3 (Fig. 1D). Thispattern of deposition suggests that the antibody against H3K4me3recognised the correct epitope.Nevertheless,weadditionally performeda peptide competition assay, and show that a peptide trimethylated atK4 can abrogate theH3K4me3 signal, but a peptide trimethylated atK27has no effect (Fig. 1E). Taken together, the pattern of deposition, thecompetition assays, and other data provided below indicate that thisantibody is specific to the H3K4me3 mark.

Cell lineage specific deposition of H3K4me3 at the eight-cell stage

Transcription has been shown to be active at the eight-cell stage(Edgar et al., 1994). Interestingly, it is at this stage that we observed astriking difference in H3K4me3 deposition. In contrast, the othermethylation marks (H3K4me2, H3K27me2, and H3K27me3) remainedunchanged. The cells poor inH3K4me3 are the germline blastomere (P3)and three somatic blastomeres: MS, E, and C (Fig. 2A). The four somaticblastomeres enriched for H3K4me3 are all AB descendants (Fig. 2A). Weensured that these four cells are descendants of the AB lineage byperforming co-immunofluorescence with an antibody against GLP-1.GLP-1 is a Notch homolog found at the membrane and within thecytoplasm. Importantly, it is expressed only in the AB lineage at the earlystages of embryogenesis (Evans et al., 1994). At four-cell stage only ABaandABpexpressGLP-1, confirming thatGLP-1 is specific to theAB lineage(Fig. 2B). We next analysed the eight-cell stage and found that only fourcells are positive for GLP-1 expression, and that these cells show highlevels of H3K4me3 deposition (Fig. 2C). This indicates that H3K4me3deposition is high in the AB lineage. We also performed a co-stainingusing an antibody against PGL-1, which confirms that levels of H3K4me3are low in the P3 cell (Supplementary Fig. 1). We next quantified theabundance of H3K4me3 relative to H3K27me2 using confocal microsco-py (Fig. 2D).We analysed 80 cells from ten embryos, and found a 3.3 foldenrichment in H3K4me3 deposition in the AB descendants comparedwith the P1 descendants (Fig. 2E). We have also produced 3D movies tohelp visualise these data (Supplementary Fig. 2). Accordingly, similarresults were obtained from another commercially available antibodyagainst H3K4me3 (Supplementary Fig. 3). Taken together, these datalead us to conclude that theH3K4me3mark is regulated according to thecell lineage. Indeed, the somatic blastomeres derived from P1, as well asP3, are resistant to accumulation of H3K4me3.

Since at eight-cell stagewe have observed four H3K4me3-depeletedcells, we expected that following an additional cell division, eight cells(seven somatic cells and P4) would remain depleted of H3K4me3, butthis is not the case. Indeed, only three somatic cells and P4 remain poorin H3K4me3. This implies that half of the descendants of the fourH3K4me3-depeleted cells are capable of acquiring the H3K4me3 mark(Fig. 3A). Accordingly, high levels of H3K4me3 are detected in allsomatic cells at thepost-gastrulation stage (Fig. 3B). As expected, Z2 andZ3, the PGCs, are depleted of the H3K4me3 mark (Fig. 3B) and of theH3K4me2 mark (Katz et al., 2009; Schaner et al., 2003). Therefore, thedeposition of the H3K4me3 mark is a very dynamic process.

The MLL complex is responsible for acquisition of H3K4me3 during earlyembryogenesis

We next addressed whether the deposition of H3K4me3 duringearly embryogenesis requires the MLL complex. We used mutants of

bryo lineage. The one cell embryo first divides asymmetrically to generate the AB and Pline blastomeres (P1, P2, P3, and P4 (green and blue)), the somatic blastomeres (EMS, Cl the gamete of the adult hermaphrodite. (B) Depiction of the two-, four-, and eight-cellmline blastomeres. (C) The H3K4me3 and H3K27me2 marks are detected in all the cellsd DNA by DAPI are shown as well as the corresponding DIC picture. (D) A chromosomicpeptides as indicated followed by immunostaining provide further evidence that thealso obtained at later stages of development and also in the germlines.

somatic blastomeres

P0

P1

P2

P3

P4

Z2 - Z3

AB

EMS

C

D

germline blastomeres

primordial germ cells

:one cell embryo

two-

cell

stag

e fo

ur-c

ell s

tage

A

C

P1

AB

P2

EMS

ABa

ABp

AB P1

P2 EMS

ABaABp

P3 E

C

MSABal ABpl

ABar ABpr

Two-cell stage

Four-cell stage

Eight-cell stage

B

D

E

DAPI

DAPI

H3K4me3

H3K4me3

H3K27me2

H3K27me2

DIC

DIC

DAPI H3K4me3

DAPI H3K4me3

DAPI H3K4me3

two-

cell

stag

e

P1 AB

P1

AB

P1 AB

DAPI

H3K4me3

H3K27me2

229S. Wang et al. / Developmental Biology 355 (2011) 227–238

60

50

40

30

20

10

0

A

B C

ED

Fig. 2. H3K4me3 deposition is enriched in the AB lineage. (A) Eight-cell stage embryo showing that H3K4me3 levels are high in the AB lineage, but H3K27me2 levels are similarregardless of the lineage. The AB lineage is indicated by a dotted line. (B) Co-immunostainings against PGL-1 and H3K4me3 at four-cell stage show that PGL-1 is specific to the ABlineage. (C) Co-immunostainings against PGL-1 and H3K4me3 at eight-cell stage show that H3K4me3 and PGL-1 are detected in the same cells, the AB descendants. (D) A confocalphotograph (stacking of multiple slices) shows that H3K4me3 (red) and H3K27me2 (green) only co-localise in half the cells. (E) Quantification using confocal images of co-immunostainings against H3K4me3, H3K27me2 and DAPI. Data are from 80 cells (ten embryos) and four separate experiments. Error bars represent standard error of the mean(SEM).

230 S. Wang et al. / Developmental Biology 355 (2011) 227–238

A

B

Fig. 3. H3K4me3 deposition is resumed in the somatic lineage. (A) An embryo at 16-cell stage showing that the H3K4me3 mark in the germline blastomere and three somaticblastomeres is not detectable or very faint (arrows). (B) A post-gastrulation embryo showing that the H3K4me3 mark is undetectable in primordial germline cells, Z2 and Z3, whichexpress PGL-1 (arrowheads).

231S. Wang et al. / Developmental Biology 355 (2011) 227–238

the core components (wdr-5 and rbbp-5) as well as a set-2 mutant.SET-2 is one of the two methyltransferases targeting H3K4 that havebeen shown to be part of the complex (Fisher et al., 2010; Simonetet al., 2007). Initially, we performed co-immunofluorescence at two-,and four-cell stages in the wdr-5 mutant. This mutant has alreadybeen shown to be impaired in deposition of H3K4me3 (Fisher et al.,2010; Simonet et al., 2007), but not at early stages of embryonicdevelopment. We started our analysis at the two-cell stage, and couldstill detect H3K4me3 deposition, though the signal appeared weakerthan in wild type worms (Fig. 4A). However, at four-cell stage, wewere unable to detect the H3K4me3 mark (Fig. 4B). Noteworthy, wefound no defect in levels of deposition of H3K27me2 (Figs. 4A and B)or of H3K27me3 (data not shown).

Additionally, we analysed the eight-cell stage and found that theH3K4me3 mark is undetectable in wdr-5, rbbp-5, and set-2 mutantscompared toN2 (Fig. 5A).Weperformed the samesystematic analysis atpost-gastrulation stage. We obtained the same result with wdr-5 and

rbbp-5 mutants, but surprisingly the set-2 mutant showed a detectablealbeit low level of H3K4me3 (Fig. 5B). To ensure that these observationsbased on immunofluorescence are not due to masking of specificepitopes,weperformedwesternblot analysis usingextracts fromwdr-5,rbbp-5, and ash-2 mutants and compared levels of H3K4me3 to wildtype embryoextracts.We found that in all themutant cases,H3K4me3 isundetectable, which is consistent with our immunofluorescence data(Fig. 5C).We conclude that theMLL complex is required for appropriatedeposition of H3K4me3 during embryogenesis.

H3K4me3 deposition remains low in germline blastomeres from rbr-2mutants

Our observations are consistent with a dynamic and lineagespecific regulation of H3K4me3 deposition during embryogenesis, inwhich theMLL complexmust be playing an important role in cells richin H3K4me3. However, poor deposition of H3K4me3 could be

A B

C D

Fig. 4. WDR-5 is required to generate the H3K4me3 mark. (A) Two-cell stage N2 embryo in which the H3K4me3 and H3K27me2 marks can be detected. (B) Two-cell stage wdr-5embryo in which the H3K4me3 mark appears diminished, and in which the H3K27me2 marks is not affected. (C) Four-cell stage N2 embryo in which the H3K4me3 and H3K27me2marks can be detected. (D) Four-cell stage wdr-5 embryo showing that the H3K4me3 mark is now undetectable, but H3K27me2 deposition is unaffected.

232 S. Wang et al. / Developmental Biology 355 (2011) 227–238

explained by both reduced MLL activity and increased demethylaseactivity. Two types of demethylases targeting H3K4 have beendescribed in C. elegans: LSD1-like and RBR-2 (Christensen et al.,

2007; Cloos et al., 2008; Greer et al., 2010; Katz et al., 2009). The LSD1ortholog, SPR-5, can demethylate H3K4me2/me1 and has been shownto be active in the PGCs (Katz et al., 2009). On the other hand, RBR-2

A

B

C

Fig. 5. Mutants of theMLL complex show reduced levels of H3K4me3. (A) The coreMLL complex, rbbp-5, andwdr-5, as well as themethyltransferase set-2, are required for deposition ofH3K4me3 at the eight-cell stage of embryogenesis. (B) The core MLL complex, rbbp-5, andwdr-5, as well as the methyltransferase set-2, are required for deposition of H3K4me3 at postgastrulation. (C) Western blot analysis against H3K4me3 using different antibodies shows that embryonic extracts from wdr-5, rbbp-5, and ash-2 have reduced levels of H3K4me3.

233S. Wang et al. / Developmental Biology 355 (2011) 227–238

234 S. Wang et al. / Developmental Biology 355 (2011) 227–238

can act on H3K4me3/me2, which makes RBR-2 a good candidate,since it can target H3K4me3 (Christensen et al., 2007). To testwhether demethylation by RBR-2 contributes to the H3K4me3pattern of deposition, we used an rbr-2 mutant, and analysed itseffect at multiple stages, including the eight-cell stage. At this stage,we found that the rbr-2 embryos deposit H3K4me3 in a patternsimilar to wild type animals (Fig. 6A, eight-cell stage). Importantly, wecould not detect H3K4me3 deposition in the P4 or later in Z2 and Z3,the PGCs (Fig. 6A, P4 and Z2-Z3). However, when we compared theH3K4me3 levels between the rbr-2 mutant and wild type, weobserved an enrichment of H3K4me3 in the somatic cells (Fig. 6B).We quantified this observation using western blot analysis, anddetected about four times the level of H3K4me3 in rbr-2 embryoscompared with wild type, with no effect on H3K27me3 deposition(Fig. 6C). Therefore, RBR-2 is not involved in preventing H3K4me3

A

Fig. 6. H3K4me3 deposition remains low in germline blastomeres in rbr-2mutants. (A) H3K4me3between rbr-2 and N2 embryos suggests an elevated deposition of H3K4me3 in rbr-2 embryos. (Cdeposition of H3K4me3 in rbr-2 embryos. K4_r indicates blotting using the rabbit anti-H3K4me3 (arabbit anti-H3.

deposition in the P lineage, but could play a role in regulating somaticdeposition of H3K4me3.

Absence of PIE-1 activates deposition of H3K4me3 in the P lineage

PIE-1 has been shown to be critical to maintain the germline fatethrough repression of transcription in the P lineage (Mello et al., 1992;Mello et al., 1996; Seydoux and Strome, 1999; Zhang et al., 2003).Herein, we have shown that deposition of H3K4me3, an activationmark, is low in the P lineage. Hence, we tested whether the absence ofPIE-1, a repressor of transcription, could promote the deposition ofH3K4me3 in the P lineage, and the somatic descendants of P1. Using apie-1 mutant and performing similar analysis as presented above, wedid not observe an effect on the deposition of H3K4me3 at two- andfour-cell stages (Figs. 7AB). However, we found that at eight-cell

B

C

deposition is not affected in germline blastomeres of rbr-2mutant embryos. (B) Comparison) Comparison between rbr-2 and N2 embryos using western blot analysis shows an elevatedb8580), H3K4_m, themouse anti-H3K4me3(ab1012), K27 a rabbit anti-H3K27me3, andH3 a

A B

DC

Fig. 7. H3K4me3 deposition is activated in germline blastomeres in pie-1 mutants. (A and B) Immunostainings at two- and four-cell stages with pie-1 embryos showing thatH3K4me3 deposition appears normal. (C) An eight-cell stage pie-1 embryo showing that the H3K4me3 mark can be detected in all cells, including P3. (D) A later stage of a pie-1embryo showing that H3K4me3 is present in P4.

235S. Wang et al. / Developmental Biology 355 (2011) 227–238

stage, the embryos depleted of PIE-1 acquire the H3K4me3mark in P3and in the somatic cells in which H3K4me3 is normally poor (Fig. 7C).Strikingly, even at later stages of development, we observed

deposition of H3K4me3 in P4 (Fig. 7D). Therefore PIE-1 is critical toprevent accumulation of H3K4me3 in germline blastomeres and in thesomatic descendants of P1.

236 S. Wang et al. / Developmental Biology 355 (2011) 227–238

H3K4me3 deposition is robust in the germline

The effect that the wdr-5 mutant has on levels of H3K4me3 inembryo is striking (Fig. 5). But surprisingly, previous studies haveshown that L4 or adults express levels of H3K4me3 that can easily bedetected by Westerns blot analysis (Fisher et al., 2010; Simonet et al.,2007). One explanation (other than the sensitivity of themethods) forthese results is that the germ cells of the wdr-5 animals maintainrobust deposition of H3K4me3. Hence, we tested this possibility byperforming immunostaining in the gonads of wdr-5 mutants. Wefound that in the wild type gonads, the H3K4me3 pattern ofdeposition is dynamic. It is at a lower level in the mitotic zone, butat a higher level in the other zones (Fig. 8A). In contrast, thedifferential distribution of the H3K27me2 modification is similarbetween the different zones, in particular if the DAPI signal is takeninto consideration (Fig. 8A). Importantly, in the wdr-5 gonads, theH3K4me3 mark is detectable, but its deposition pattern has changed(Fig. 8B). It appears that the zone normally low in H3K4me3 levels isextended in wdr-5 mutants (Fig. 8B). This may explain the source ofH3K4me3 detected by western blot analysis in wdr-5 animals (Fisheret al., 2010; Simonet et al., 2007), and also suggests that H3K4me3deposition in the germline is more robust than in the embryo.

A

Fig. 8. Germline deposition of H3K4me3 in N2 and a wdr-5mutant. (A) Wild type germlinesH3K27me2mark is evenly distributed relative to DAPI staining. (B) Thewdr-5mutant can stil

Importantly, our data from embryos and the germline imply thatembryonic viability and fertility could be affected in mutants of thecore MLL complex. Hence, we assessed these functions in the wdr-5,rbbp-5 and ash-2 mutants. We found that embryonic viability andfertility in all mutants are compromised at both 20 °C and 25 °C. Butsurprisingly, these functions are affected differentially. For example,the wdr-5 mutant produces 45% embryonic lethality at 25 °Ccompared with 16% for rbbp-5 (pb0.005) and 0% for ash-2(pb0.00005) (Table 1). On the other hand, brood size is more affectedin ash-2 (6 progeny, pb0.0005) or rbbp-5 (19 progeny, pb0.001) thanin wdr-5 (54 progeny) (Table 1). Perhaps this is an indication that theMLL core complex can associate with different partners according tothe site of expression. Despite this extra layer of complexity, our datasupport the view that proper embryonic development and fertilityrequires deposition of methyl groups on H3K4.

Discussion

Transcriptional activity during early embryogenesis is poor in thegermline blastomeres, but processive in the somatic blastomeres(Edgar et al., 1994; Schaner and Kelly, 2006; Schaner et al., 2003;Seydoux and Dunn, 1997; Seydoux and Fire, 1994; Seydoux and

B

showing that the deposition of the H3K4me3 mark is low level in the mitotic zone. Thel deposit H3K4me3 in the germline, but the zone in which the levels are low is extended.

Table 1Embryonic viability and fertility analysis of mutants of the MLL core complex. P values:*pb0.01, **pb0.005, ***pb0.0001, ****pb0.000001.

Genotypes, temp % emb. leth. aver. brood size

N2, 20 °C 0 204N2, 25 °C 0.5 149wdr-5, 20 °C 4** 111**wdr-5, 25 °C 45*** 54***rbbp-5, 20 °C 2 44***rbbp-5, 25 °C 16 19****ash-2, 20 °C 5* 38****ash-2, 25 °C 0 6****

237S. Wang et al. / Developmental Biology 355 (2011) 227–238

Strome, 1999). Previous studies have investigated the deposition ofmethylation and acetylation marks during early embryogenesis andfound that H3K4me2 is actively removed from the PGCs later duringembryogenesis (N26-cell stage). Prior to this stage, the levels ofH3K4me2 (and other marks) remain equivalent in all cells of the earlyembryo (Katz et al., 2009; Schaner and Kelly, 2006; Schaner et al.,2003). Similarly, we found that H3K4me2 and H3K27me2/me3 levelsremain equivalent prior to the generation of PGCs. However, weuncovered that the H3K4me3 mark has a unique distribution duringembryogenesis. Indeed, some of the somatic blastomeres and thegermline blastomeres fail to accumulate the H3K4me3 marks.Interestingly, this characteristic is regulated by PIE-1. Therefore, it islikely that deposition of the H3K4me3 mark is important for thedevelopment of the germline blastomeres.

Prior to this study there was no evidence that the chromatinarchitecture of germline blastomeres could be different from somaticblastomeres. Indeed, most data indicated that, at least at a superficiallevel, both types of blastomeres were equivalent in deposition ofnumerous metylation and acetylation marks thereby implying thattheir chromatin architecture was similar. Here, we provide newinsights into a possible role that a specific mark (H3K4me3) may playto ensure that the chromatin architecture adopted in the germlineblastomeres is in accordancewith its transcriptional status. Consistentwith this role, we found that in absence of PIE-1, deposition ofH3K4me3 is restored in germline blastomeres, and that thedifferential deposition between somatic blastomeres at the eight-cell stage is abolished. Therefore, these data imply interplay betweenthe general transcriptional repressor PIE-1 and the chromatinarchitecture of the germline blastomeres.

The observation that specific somatic blastomeres fail to accumu-late the H3K4me3 mark was unexpected. Interestingly, the somaticblastomeres “resistant” to the deposition of the third methylationmark on H3K4 are derived from the germline blastomere P1. Hence,we would like to raise the possibility that the P1 germline blastomereproduces an activity that prevents the MLL complex to deposit thethird methylation mark on H3K4 or/and that it stimulates its removalby a still undiscovered demethylase. However, this putative activityfrom P1 may not affect directly the MLL complex or the demethylase,but could affect the deposition or removal of other marks on thechromatin leading to the effect characterised herein. Our data appearto rule out RBR-2 as a potential demethylase, since it is not required tomaintain the H3K4me3 deposition pattern at the eight-cell stage(Fig. 6). Finally, considering that the germline blastomere is a type ofstem cell, our findingsmay provide new insights into how totipotency(and perhaps also pluripotency) can be maintained by controlling thedeposition of the third methyl mark on H3K4.

Materials and methods

Strains and general maintenance

Strains were maintained as previously described (Brenner, 1974).The strains used in this study were: wdr-5(ok1417), set-2 (ok952),

rbbp-5(tm3463), ash-2(tm1905), rbr-2(tm1231), pie-1(zu154) unc-25(e156)/qC1 dpy-19(e1259) glp-1(q339) and Bristol N2 as wild type.

Co-immunofluorescence

Immunofluorescence by freeze crack was performed on polylysinetreated slides at 0.3%. Slides treated by application of 75 μl ofpolylysine 0.3%, dried 10 min at 70 °C, quickly rinsed in distilledwater and excess liquid wiped off. The slides were 3×14 mm printedwells from Fisher Scientific LTD UK. About 30 mothers were placed ina well with a drop of M9 buffer, to wash off bacteria. These motherswere transferred into a 5-6 μl drop of M9 onto the polylysine treatedwell using an eyelash. A cover slip (22 mm×50 mm) was applied at aright angle and the slide placed at −80 °C for at least 20 min. Thecover slip was then promptly removed and embryosmethanol fixed at−20 °C for 10 min, washed 5 min in PBS, and then followed by twowashes in PBS-tween 0.2%. Primary antibodywas incubated overnightat 4 °C in a humid chamber. Washes performed as described above.Secondary antibody was incubated 2 h at 37 °C. Washes performed asdescribed above. Mowiol was applied to preserve fluorescence.Primary antibodies used: anti-H3K4me3 (Abcam, ab1012 andab8580), anti-H3K4me2 (Abcam, ab32356), anti-H3K27me3 (Milli-pore/Upstate, 07-449), anti-H3K27me2 (Millipore/Upstate, 07–452),and anti-H3 (Abcam, 1791). The secondary antibodies used are fromJackson ImmunoResearch: DyLight 594 AffiniPure Goat Anti-MouseIgG (H+L) (cat.115-515-146) and DyLight 488 AffiniPure Goat Anti-Rabbit IgG (H+L) (cat.111-485-144). Our controls performedwithout primary antibody show that these secondary antibodies donot produce significant background (data not shown).

Image processing

Confocal images were analysed using imageJ software. Individualslices were selected for every cell to discriminate between over-lapping cells. We used eight-cell stage embryos in which ABdescendants are distinctive from P1 descendants, i.e. the AB cellsenter mitosis faster than the P1 linage. We expressed the results as aratio of the H3K4me3 signal over the H3K27me2 signal for each cells.

Peptide competition assays

Peptides were preincubated in three fold excess with the anti-H3K4me3 (ab1012) at room temperature for 30 min prior toimmunostainings. Immunostainings were performed as describedabove. The peptides used were produced by Abcam (ab1432,H3K4me3) and (ab1782, H3K27me3).

Western blot analysis

Western blot analysis was performed as described in (Fisher et al.,2010) with a fewmodifications. Briefly, embryo protein extracts wereprepared by bleaching young mothers followed by washing theembryos four times with M9 buffer. A minimum volume of 20 μlpelleted embryos was collected for each protein sample, boiled inLaemmli buffer containing 100 mM DTT and sonicated.

Embryonic and fertility analysis

To calculate brood size and embryonic lethality ten clones perstrain were transferred onto fresh plates every day until the wormshad finished laying. The number of embryos and progeny werecounted each day. Embryos are considered dead if they did not hatchafter more than 24 h and appeared grossly abnormal.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.ydbio.2011.04.010.

238 S. Wang et al. / Developmental Biology 355 (2011) 227–238

Acknowledgments

SarahWoolner and Karel Dorey for confocal microscopy and adviseon quantification using imageJ. The anti-GLP-1 and anti-PGL-1 anti-bodies are kind gifts from Judith Kimble (Evans et al., 1994) and SusanStrome (Kawasaki et al., 1998), respectively. Some nematode strainsused in this work were provided by the Caenorhabditis Genetics Center,which is funded by the NIH National Center for Research Resources(NCRR). The rbbp-5 deletion strain was generated by the JapaneseConsortium (National BioResource Project for C. elegans). MRC forfunding GBP (Career Development Award G0600127).

References

Agger, K., Cloos, P.A., Christensen, J., Pasini, D., Rose, S., Rappsilber, J., Issaeva, I., Canaani,E., Salcini, A.E., Helin, K., 2007. UTX and JMJD3 are histone H3K27 demethylasesinvolved in HOX gene regulation and development. Nature 449, 731–734.

Azuara, V., Perry, P., Sauer, S., Spivakov, M., Jorgensen, H.F., John, R.M., Gouti, M.,Casanova, M., Warnes, G., Merkenschlager, M., Fisher, A.G., 2006. Chromatinsignatures of pluripotent cell lines. Nat. Cell Biol. 8, 532–538.

Bender, L.B., Cao, R., Zhang, Y., Strome, S., 2004. The MES-2/MES-3/MES-6 complex andregulation of histone H3 methylation in C. elegans. Curr. Biol. 14, 1639–1643.

Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry, B., Meissner,A., Wernig, M., Plath, K., Jaenisch, R., Wagschal, A., Feil, R., Schreiber, S.L., Lander, E.S.,2006. A bivalent chromatin structure marks key developmental genes in embryonicstem cells. Cell 125, 315–326.

Brenner, S., 1974. The genetics of Caenorhabditis elegans. Genetics 77, 71–94.Cao, R.,Wang, L., Wang, H., Xia, L., Erdjument-Bromage, H., Tempst, P., Jones, R.S., Zhang,

Y., 2002. Role of histone H3 lysine 27 methylation in Polycomb-group silencing.Science 298, 1039–1043.

Capowski, E.E., Martin, P., Garvin, C., Strome, S., 1991. Identification of grandchildlessloci whose products are required for normal germ-line development in thenematode Caenorhabditis elegans. Genetics 129, 1061–1072.

Christensen, J., Agger, K., Cloos, P.A., Pasini, D., Rose, S., Sennels, L., Rappsilber, J., Hansen,K.H., Salcini, A.E., Helin, K., 2007. RBP2 belongs to a family of demethylases, specificfor tri-and dimethylated lysine 4 on histone 3. Cell 128, 1063–1076.

Cloos, P.A., Christensen, J., Agger, K., Helin, K., 2008. Erasing the methyl mark: histonedemethylases at the center of cellular differentiation and disease. Genes Dev. 22,1115–1140.

Edgar, L.G., Wolf, N., Wood, W.B., 1994. Early transcription in Caenorhabditis elegansembryos. Development 120, 443–451.

Evans, T.C., Crittenden, S.L., Kodoyianni, V., Kimble, J., 1994. Translational control ofmaternal glp-1 mRNA establishes an asymmetry in the C. elegans embryo. Cell 77,183–194.

Fisher, K., Southall, S.M., Wilson, J.R., Poulin, G.B., 2010. Methylation and demethylationactivities of a C. elegans MLL-like complex attenuate RAS signalling. Dev. Biol. 341,142–153.

Greer, E.L., Maures, T.J., Hauswirth, A.G., Green, E.M., Leeman, D.S., Maro, G.S., Han, S.,Banko, M.R., Gozani, O., Brunet, A., 2010. Members of the H3K4 trimethylationcomplex regulate lifespan in a germline-dependent manner in C. elegans. Nature466, 383–387.

Issaeva, I., Zonis, Y., Rozovskaia, T., Orlovsky, K., Croce, C.M., Nakamura, T., Mazo, A.,Eisenbach, L., Canaani, E., 2007. Knockdown of ALR (MLL2) reveals ALR target genesand leads to alterations in cell adhesion and growth. Mol. Cell. Biol. 27, 1889–1903.

Katz, D.J., Edwards, T.M., Reinke, V., Kelly, W.G., 2009. A C. elegans LSD1 demethylasecontributes to germline immortality by reprogramming epigenetic memory. Cell137, 308–320.

Kawasaki, I., Shim, Y.H., Kirchner, J., Kaminker, J., Wood, W.B., Strome, S., 1998. PGL-1, apredicted RNA-binding component of germ granules, is essential for fertility inC. elegans. Cell 94, 635–645.

Kelly, W.G., Schaner, C.E., Dernburg, A.F., Lee, M.H., Kim, S.K., Villeneuve, A.M., Reinke,V., 2002. X-chromosome silencing in the germline of C. elegans. Development 129,479–492.

Kouzarides, T., 2007. Chromatin modifications and their function. Cell 128, 693–705.Lan, F., Bayliss, P.E., Rinn, J.L., Whetstine, J.R., Wang, J.K., Chen, S., Iwase, S., Alpatov, R.,

Issaeva, I., Canaani, E., Roberts, T.M., Chang, H.Y., Shi, Y., 2007. A histone H3 lysine27 demethylase regulates animal posterior development. Nature 449, 689–694.

Lee, M.G., Villa, R., Trojer, P., Norman, J., Yan, K.P., Reinberg, D., Di Croce, L., Shiekhattar,R., 2007. Demethylation of H3K27 regulates polycomb recruitment and H2Aubiquitination. Science 318, 447–450.

Mello, C.C., Draper, B.W., Krause, M., Weintraub, H., Priess, J.R., 1992. The pie-1 andmex-1 genes and maternal control of blastomere identity in early C. elegansembryos. Cell 70, 163–176.

Mello, C.C., Schubert, C., Draper, B., Zhang, W., Lobel, R., Priess, J.R., 1996. The PIE-1protein and germline specification in C. elegans embryos. Nature 382, 710–712.

Reinke, V., Smith, H.E., Nance, J., Wang, J., Van Doren, C., Begley, R., Jones, S.J., Davis, E.B.,Scherer, S., Ward, S., Kim, S.K., 2000. A global profile of germline gene expression inC. elegans. Mol. Cell 6, 605–616.

Reuben, M., Lin, R., 2002. Germline X chromosomes exhibit contrasting patterns ofhistone H3 methylation in Caenorhabditis elegans. Dev. Biol. 245, 71–82.

Ringrose, L., Paro, R., 2004. Epigenetic regulation of cellular memory by the Polycomband Trithorax group proteins. Annu. Rev. Genet. 38, 413–443.

Ruthenburg, A.J., Allis, C.D., Wysocka, J., 2007. Methylation of lysine 4 on histone H3:intricacy of writing and reading a single epigenetic mark. Mol. Cell 25, 15–30.

Schaner, C.E., Kelly, W.G., 2006. Germline chromatin. WormBook 1–14.Schaner, C.E., Deshpande, G., Schedl, P.D., Kelly, W.G., 2003. A conserved chromatin

architecture marks and maintains the restricted germ cell lineage in worms andflies. Dev. Cell 5, 747–757.

Seydoux, G., Dunn, M.A., 1997. Transcriptionally repressed germ cells lack asubpopulation of phosphorylated RNA polymerase II in early embryos ofCaenorhabditis elegans and Drosophila melanogaster. Development 124, 2191–2201.

Seydoux, G., Fire, A., 1994. Soma-germline asymmetry in the distributions of embryonicRNAs in Caenorhabditis elegans. Development 120, 2823–2834.

Seydoux, G., Strome, S., 1999. Launching the germline in Caenorhabditis elegans:regulation of gene expression in early germ cells. Development 126, 3275–3283.

Shin, T.H., Mello, C.C., 2003. Chromatin regulation during C. elegans germlinedevelopment. Curr. Opin. Genet. Dev. 13, 455–462.

Simonet, T., Dulermo, R., Schott, S., Palladino, F., 2007. Antagonistic functions of SET-2/SET1 and HPL/HP1 proteins in C. elegans development. Dev. Biol. 312, 367–383.

Sulston, J.E., Schierenberg, E., White, J.G., Thomson, J.N., 1983. The embryonic celllineage of the nematode Caenorhabditis elegans. Dev. Biol. 100, 64–119.

Unhavaithaya, Y., Shin, T.H., Miliaras, N., Lee, J., Oyama, T., Mello, C.C., 2002. MEP-1 and ahomolog of the NURD complex component Mi-2 act together tomaintain germline-soma distinctions in C. elegans. Cell 111, 991–1002.

Zhang, F., Barboric, M., Blackwell, T.K., Peterlin, B.M., 2003. A model of repression: CTDanalogs and PIE-1 inhibit transcriptional elongation by P-TEFb. Genes Dev. 17,748–758.