DispatchR545

Dispatches

Chromatin: Sub Out the Replacement

Nucleosomes with specific histone variants are incorporated into DNA at sitesof transcription and repair. The histone variant H3.3 has been linked totranscriptional regulation, but genetic tests in the fruit fly have yieldedsurprising results.

Oliver Bell1 and Dirk Schubeler2,*

In eukaryotic cells, DNA is embeddedin chromatin, consisting of the doublehelix wrapped around an octamer ofhistone proteins: H2A, H2B, H3 and H4.The majority of histones are assembledinto chromatin following replication inS phase to allow proper packaging ofnewly synthesized DNA. In addition,distinct variant histones are expressedoutside of S phase and depositedindependently of replication [1]. This isthought to provide a continuous sourceof histones, allowing replacement ofnucleosomes or compensating fortheir loss. Importantly, histones andtheir replacements are subject tomodifications, which can influencegene expression and chromatinarchitecture [2]. In the case ofcanonical histone H3, the closelyrelated variant H3.3 is highly enrichedfor several modifications associatedwith transcription [3,4], reflecting itsspecific incorporation at active genes[5–7]. Moreover, this replacementhistone is also highly abundant atregulatory regions [8]. Together thesestriking observations have led todifferent hypotheses on the functionof the histone variant H3.3 in generegulation. In this issue of CurrentBiology, Hoedl and Basler [9] usegenetics to directly address therole of active histone modificationsand histone variant H3.3. Verysurprisingly, in the absence of H3.3the authors find little effect onsomatic development and showthat only specific lysines are requiredfor the essential function of H3.3in the germline.

Chromatin is dynamically modified,which is thought to facilitate DNAaccessibility for protein binding [2].Indeed, nucleosomes are lost atactively transcribed genes andregained upon transcriptionalshut-down [10]. Deposition of varianthistone H3.3 has been observed at

promoters and throughout transcribedregions, and appears to partlycompensate for transcription-couplednucleosome displacement [5,7]. Inaddition, histones incorporated atactive genes are decorated by severalpost-translational modifications,including methylation at lysine 4 ofhistone H3 [7]. The relevance of thesemodifications and their link to thedeposition of H3.3 are importantquestions, which previously have notbeen addressed by genetic analysis.In metazoans, the genes encodingcanonical histones are organizedin tandem, multicopy clusters,which have hindered geneticstudies. Thus, research has mostlyfocused on the identification anddeletion of the enzymes that modifyhistones [11]. However, the possibilitythat several enzymes can modifythe same residue and the existenceof non-histone targets posesobvious limitations.

To directly address the role ofhistone modifications in Drosophiladevelopment, Hoedl and Basler [9]deleted the genes encoding histonevariant H3.3. Unlike canonical histones,H3.3 is expressed from two genecopies. As H3.3 deposition coincideswith the enrichment of several histonemodifications associated with geneactivity, its deletion may contributeto understanding the role of thesemodifications in transcriptionalregulation. In fact, H3.3 represents themajority of histone H3 methylated atlysine 4, which led to the hypothesis thatthere may be a link between variantdeposition and targeting of this mark [3].Surprisingly, Hoedl and Basler [9] findthat endogenous H3.3 is dispensablefor somatic development in flies. Thisfinding not only challenges some of thecurrent hypotheses for the function ofthis variant in gene regulation, it is alsosurprising given that deletion of onlya single H3.3 gene copy in mice leadsto severe developmental defects [12].

Notably, despite the deletion of H3.3,the levels of H3K4 methylation remainedunchanged, arguing that canonical H3can compensate at active genes for thelack of the variant, possibly by beingupregulated (Figure 1). However, similarexperiments performed in Tetrahymenasuggest that a lack of H3.3 at highlytranscribed genes is not compensatedby replication-independent assemblyof newly synthesized H3 [13].Alternatively, followingtranscription-dependent nucleosomaldisplacement, the resulting gaps mightbe masked by redistributing ‘old’nucleosomes from neighboringextragenic regions or by recycling freecanonical H3 histones (Figure 1).Nevertheless, one might speculatethat some gaps in non-dividing cellswill remain unfilled and, thus,potentially expose cryptic sitesfor spurious transcription initiationevents, as has been reportedin baker’s yeast [14,15].

Post-mitotic mammalian neuronsaccumulate H3.3 over time and largelyreplace the canonical isoform [16],which in turn would suggest late-onsetphenotypes also for Drosophila.Clearly, H3.3 is dispensable for somaticdevelopment, yet it will be interesting ifits lack results in enhanced sensitivityto genotoxic stress or phenotypesrelated to chromatin regulation, suchas heterochromatin-mediatedvariegation of transgene expression orpolycomb-group-mediated repressionof HOX genes.

H3.3 is also incorporated into thedecondensing male genome afterfertilization, once the sperm-specificprotamines are removed [17,18]. In linewith this function, female flies that lackH3.3 are infertile. Hoedl and Basler [9]further show that ectopic expressionof H3.2 in the embryos cannotcompensate for the loss of H3.3.Together with previous reports thatmutants in the machinery responsiblefor H3.3 deposition are also infertile[17,18], this argues that the processof protamine replacement is specificfor, and requires, maternal H3.3.This selectivity raises the questionwhether H3.3 plays a role in germline

Current Biology Vol 19 No 14R546

Histonerecycling

Nucleosomedisplacement

Nucleosomereplacement

Canonicalhistones

Varianthistones

A

B

C

Histone modifications

Current Biology

Figure 1. Compensation for loss of histone H3.3.

Under normal conditions, nucleosomes (blue) are displaced at sites of chromatin remodeling,transcription and repair. (A) Gaps are filled by nucleosomes containing variant isoforms (green)such as H3.3. (B) In the absence of H3.3, the remaining canonical nucleosomes could be redis-tributed to mask the gaps, nevertheless leading to an overall reduced nucleosomal density inthe genome. (C) Alternatively, canonical histones could be recycled by reinsertion.

development beyond the foremosttask of DNA packaging.

To directly ask if histonemodifications are required for germ cellfunction, Hoedl and Basler [9] mutatedH3.3 lysine 4 or lysine 9 to alanine(H3.3 K4A or K9A, respectively).Intriguingly, lysine 4 appears to beimportant for the particular role ofH3.3 in germ cell development asre-expression of H3.3 K4A did notrescue sterility, while flies with theK9A substitution were fertile. Itremains to be seen how andwhen mutations of lysine 4interfere with H3.3 function ingermline development or earlyembryogenesis.

Similar to the complete deletion ofH3.3, neither the K4A nor K9A mutantshad any apparent phenotype in somaticdevelopment. This is particularlyinteresting for the lysine 4 mutationsince reports from cultured Drosophilacells suggested that most of themodified lysine 4 occurs on H3.3 [3].Indeed, unlike in the absence of H3.3,expression of the H3.3 K4A mutantresulted in a strong reduction of bulklysine 4 methylation, indicating that themutated H3.3 is correctly incorporated.The resulting strong reduction of lysine4 methylation has no apparent effecton somatic transcription, which issurprising given the large number ofproteins interacting with methylated

K4 [11]. Nevertheless, the ultimatetest for the relevance of this markrequires mutation of both H3 isoforms.In the meantime, a more detailedanalysis of the distribution andmodification of the ectopicallyexpressed mutant histones will helpto better understand the lack of anyapparent phenotype.

The work by Hoedl and Basler [9]shows that genetic deletion is theultimate test of biological relevanceand can lead to surprisingobservations. This work also suggeststhat some of the current models forchromatin regulation by histonevariants should be revisited. In additionto challenging existing paradigms, thework further paves the way to approachhistone function genetically inmetazoans in order to increase ourunderstanding of genome regulationin a ‘chromatinized’ world.

References1. Henikoff, S., and Ahmad, K. (2005). Assembly

of variant histones into chromatin. Annu. Rev.Cell. Dev. Biol. 21, 133–153.

2. Felsenfeld, G., and Groudine, M. (2003).Controlling the double helix. Nature 421,448–453.

3. McKittrick, E., Gafken, P.R., Ahmad, K., andHenikoff, S. (2004). Histone H3.3 is enrichedin covalent modifications associated withactive chromatin. Proc. Natl. Acad. Sci. USA101, 1525–1530.

4. Waterborg, J.H. (1993). Dynamic methylationof alfalfa histone H3. J. Biol. Chem. 268,4918–4921.

5. Mito, Y., Henikoff, J.G., and Henikoff, S. (2005).Genome-scale profiling of histone H3.3replacement patterns. Nat. Genet. 37,1090–1097.

6. Schwartz, B.E., and Ahmad, K. (2005).Transcriptional activation triggers depositionand removal of the histone variant H3.3.Genes Dev. 19, 804–814.

7. Wirbelauer, C., Bell, O., and Schubeler, D.(2005). Variant histone H3.3 is deposited atsites of nucleosomal displacement throughouttranscribed genes while active histonemodifications show a promoter-proximal bias.Genes Dev. 19, 1761–1766.

8. Mito, Y., Henikoff, J.G., and Henikoff, S. (2007).Histone replacement marks the boundariesof cis-regulatory domains. Science 315,1408–1411.

9. Hodl, M., and Basler, K. (2009). Transcription inthe absence of histone H3.3. Curr. Biol. 19,1221–1226.

10. Lee, C.K., Shibata, Y., Rao, B., Strahl, B.D., andLieb, J.D. (2004). Evidence for nucleosomedepletion at active regulatory regionsgenome-wide. Nat. Genet. 36, 900–905.

11. Kouzarides, T. (2007). Chromatin modificationsand their function. Cell 128, 693–705.

12. Couldrey, C., Carlton, M.B., Nolan, P.M.,Colledge, W.H., and Evans, M.J. (1999). Aretroviral gene trap insertion into the histone3.3A gene causes partial neonatal lethality,stunted growth, neuromuscular deficits andmale sub-fertility in transgenic mice. Hum. Mol.Genet. 8, 2489–2495.

13. Cui, B., Liu, Y., and Gorovsky, M.A. (2006).Deposition and function of histone H3 variantsin Tetrahymena thermophila. Mol. Cell. Biol. 26,7719–7730.

14. Kaplan, C.D., Laprade, L., and Winston, F.(2003). Transcription elongation factors repress

DispatchR547

transcription initiation from cryptic sites.Science 301, 1096–1099.

15. Mason, P.B., and Struhl, K. (2003). The FACTcomplex travels with elongating RNApolymerase II and is important for the fidelity oftranscriptional initiation in vivo. Mol. Cell Biol.23, 8323–8333.

16. Pina, B., and Suau, P. (1987). Changes inhistones H2A and H3 variant composition indifferentiating and mature rat brain corticalneurons. Dev. Biol. 123, 51–58.

Bee Pheromones: Sof Manipulation?

Recent studies have provided a new perthe honey bee queen and her colony. Tha pheromone which pharmacologically m

Geraldine A. Wright

A honey bee colony is a complexeusocial society of largely sterile,female workers which are ‘ruled’ bya single reproductive monarch. Orderwithin this altruistic sorority is largelymaintained by the queen’s emissionof a chemical signal called queenmandibular pheromone (QMP) [1].When sensed or imbibed by herprogeny, QMP becomes the gluewhich binds the colony together asa self-organizing unit: it stimulatesthem to form a retinue around her [2],to rear brood, to forage for food, andto build comb [3]. It also establishesthe queen as the reproductive ruler,as it suppresses the rearing of newqueens and the development ofworker ovaries [4]. In the absenceof the queen’s chemical rule,reproductive anarchy ensues [5].

In an elegant series of studies,Mercer and colleagues have recentlyuncovered another function of QMP:it prevents young workers in thequeen’s retinue from forming aversiveolfactory memories [6]. What is more,they showed that this change inbehaviour is mediated by a singlecomponent of QMP, homovanillylalcohol (HVA). Of the five compoundsthat make up QMP, HVA is the leastconcentrated — it is w200-fold lessconcentrated than QMP’s majorcomponent, 9-ODA — yet it is essentialfor eliciting retinue behaviour [2]and may also be involved in thesuppression of ovary development inworkers [7]. The Mercer lab also foundthat exposure to HVA affects dopamine

17. Konev, A.Y., Tribus, M., Park, S.Y.,Podhraski, V., Lim, C.Y., Emelyanov, A.V.,Vershilova, E., Pirrotta, V., Kadonaga, J.T.,Lusser, A., et al. (2007). CHD1 motor protein isrequired for deposition of histone variant H3.3into chromatin in vivo. Science 317, 1087–1090.

18. Loppin, B., Bonnefoy, E., Anselme, C.,Laurencon, A., Karr, T.L., and Couble, P. (2005).The histone H3.3 chaperone HIRA is essentialfor chromatin assembly in the male pronucleus.Nature 437, 1386–1390.

ignal or Agent

spective on the relationship betweeney suggest that the queen produces

anipulates her workers.

receptor expression and reducesthe amount of dopamine in a youngbee’s brain [8].

Structurally, HVA looks likea dopamine molecule (Figure 1) andapparently it acts like one: in this issueof Current Biology, Beggs and Mercer[9] report that HVA is an agonist ofD2-like dopamine receptors. Theyselectively expressed the known beedopamine receptors — AmDOP1,

Figure 1. A honey bee queen (Apis mellifera c

Queens emit a pheromone containing themethoxyphenylethanol), which is structurally amitter dopamine. Photo courtesy of S.W. Cob

1Departments of Pathology andDevelopmental Biology, Stanford UniversityMedical School, Stanford CA 94305, USA.2Friedrich Miescher Institute for BiomedicalResearch, CH-4058 Basel, Switzerland.*E-mail: dirk.schubeler@fmi.ch

DOI: 10.1016/j.cub.2009.06.032

AmDOP2, and AmDOP3 — inmammalian cells in culture and testedligand specificity by measuringchanges in cAMP levels. AmDOP1 andAmDOP2 both responded to dopaminewith an increase in cAMP, which istypical of D1-like receptors, whileAmDOP3 responded as a D2-likereceptor, lowering cAMP. Interestingly,HVA also reduced cAMP in AmDOP3expressing cells.

By which mechanisms could HVAreduce aversive learning in bees? Itcould act as an antagonist ofdopaminergic receptors, binding withthe receptors and inactivating them,but Beggs and Mercer [9] found noevidence for this. On the other hand,as a D2-receptor agonist, perhapsHVA’s activation of AmDOP3 dampens

arnica) surrounded by her retinue.

molecule homovanillyl alcohol (4-hydroxy-3-nd pharmacologically similar to the neurotrans-ey.