© 1999 Macmillan Magazines Ltd

letters to nature

NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com 691

Acknowledgements. We thank Z. Onsan for expert technical assistance, and E. Macpherson, B. Mickey,D. Moody, B. Pfingst and J. Schacht for comments on the manuscript. This work was supported by theNIH/HIHCD. The multi-channel recording probes were supplied by the Center for Neural Commu-nication Technology, which is supported by the NIH/NCRR.

Correspondence and requests for materials should be addressed to J.M. (e-mail: jmidd@umich.edu).

TheSILgene is requiredformouseembryonicaxial development andleft–right specificationShai Izraeli*, Linda A. Lowe†, Virginia L. Bertness*,Deborah J. Good‡, David W. Dorward§, Ilan R. Kirsch*& Michael R. Kuehn†

* Genetics Department, Medicine Branch, National Cancer Institute, NIH,Bethesda, Maryland 20889-5105, USA† Experimental Immunology Branch, National Cancer Institute, NIH, Bethesda,Maryland 20892-1360, USA‡ Department of Veterinary and Animal Sciences, University of Massachusetts,Amherst, Massachusetts 01003, USA§ Rocky Mountain Laboratories, National Institute of Allergy and InfectiousDiseases, NIH, Hamilton, Montana 59840, USA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The establishment of the main body axis and the determination ofleft–right asymmetry are fundamental aspects of vertebrateembryonic development. A link between these processes hasbeen revealed by the frequent finding of midline defects inhumans with left–right anomalies1. This association is also seenin a number of mutations in mouse2–4 and zebrafish1,5, and inexperimentally manipulated Xenopus embryos5. However, theseverity of laterality defects accompanying abnormal midlinedevelopment varies6, and the molecular basis for this variationis unknown. Here we show that mouse embryos lacking the early-response gene SIL have axial midline defects, a block in midlineSonic hedgehog (Shh) signalling and randomized cardiac looping.Comparison with Shh mutant embryos7, which have axial defectsbut normal cardiac looping, indicates that the consequences ofabnormal midline development for left–right patterning dependon the time of onset, duration and severity of disruption of thenormal asymmetric patterns of expression of nodal, lefty-2 andPitx2.

SIL is an early-response gene that was discovered at the site ofa leukaemia-associated chromosomal rearrangement8. It is ubiqui-tously expressed in proliferating cells9 and during early embryonicdevelopment (not shown). To analyse SIL function in vivo, wedeleted exons 3–5, including the translation start sites10, by homo-logous recombination in murine embryonic stem cells (not shown).Immunoblot analysis of protein extracts from SIL−/− cells confirmedthat this is a null allele (not shown). Mice derived from twoindependent SIL−/− embryonic stem-cell clones showed the samephenotype. Heterozygotes are normal, but mutant homozygotesdie in utero after embryonic day 10.5 (E10.5). Between E7.5 andE8.5, striking developmental anomalies appear in SIL−/− embryos.In addition to reduced size and limited developmental progresscompared to wild-type embryos (Fig. 1a, b), SIL mutants displayprominent midline neural-tube defects. These include delay orfailure of neural-tube closure and a holoprosencephalic defect,namely the lack of any midline separation at the anterior end ofthe cranial neural folds (Fig. 1c, d). In addition, left–right devel-opment is abnormal. In heterozygous and wild-type embryos theembryonic heart tube always loops to the right (Fig. 1e), whereas inSIL mutants the direction of heart looping is randomized (Fig. 1f, g;Table 1).

The transforming growth factor-b (TGF-b) family members

nodal, lefty-1 and lefty-2 and the transcription factor Pitx2 are allimplicated in the establishment of left–right-handed asymmetry inthe mouse11–18. Nodal, lefty-2 and Pitx2 are normally expressed onlyin the left lateral-plate mesoderm (LPM) before heart looping, withcontinued expression of Pitx2 on the left side of the looping hearttube (Fig. 2a–c). In contrast, SIL mutants showed bilaterallysymmetric expression of nodal and Pitx2 (Fig. 2d, f) at all stagesexamined (Table 1). For lefty-2, most SIL−/− embryos also showedbilaterally symmetric expression (Fig. 2e). However, a small numberof mutants expressed lefty-2 only on the right. These were all atvery early somite stages, when lefty-2 expression is first detectable(Table 1), indicating that abnormal lefty-2 expression on the rightside of SIL mutants may precede normal left-sided expression. Innormal embryos lefty-1 is expressed on the left side of the prospec-tive floorplate (PFP) at the ventral midline of the neural folds. InSIL mutants we found no midline lefty expression at any stageexamined (Fig. 2e). These results show that SIL activity is requiredto establish the normal pattern of left–right asymmetric geneexpression, both in the LPM and in the PFP, from the earlieststages of their expression.

The observed neural-tube defects, including the lack of lefty-1expression in the PFP, prompted an analysis of midline develop-mental pathways. Two key players in midline development are thesecreted signalling factor Shh and the winged helix transcriptionfactor HNF-3b (refs 7, 19, 20). Both are normally expressed in thenode and notochord at E7.5 to E8.5 (Fig. 3a, g for Shh; 3c, e, i, k forHNF-3b). HNF-3b is found also in the PFP (Fig. 3i, k), where itsexpression is dependent on Shh signalling from the adjacent node

Figure 1 Gross morphology of SIL−/− embryos. a, Lateral view of normal E9.5

embryo. b, E9.5 SIL mutant embryo, unturned with abnormal pericardial swelling

(arrow). c, Scanning electron microscopy (SEM) of normal E8.5 embryo head,

frontal view. Arrow, separation at anterior of neural folds. d, SEM of E8.5 SIL−/−

embryo, lacking normal midline separation at the anterior of the neural folds

(arrow). e, Ventral view of normal E9.5 embryo showing normal rightwards heart

looping (direction shown by red arrow). f, E9.5SIL−/− embryowith rightwardsheart

looping. g, E9.5 SIL−/− embryo with leftwards heart looping.

© 1999 Macmillan Magazines Ltd

letters to nature

692 NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com

and/or notochord7,21. In SIL mutants, both Shh and HNF-3b areexpressed in the notochord (Fig. 3h for Shh; 3j for HNF-3b),although by E8.5 the pattern of expression is discontinuous(Fig. 3b for Shh) because of gaps in the notochord (not shown).However, there was no HNF-3b expression in the PFP at E8.5(Fig. 3d, j) or at E7.5, when the neural folds and developingnotochord appear normal (Fig. 3f, l). This indicates that the lackof neural tube HNF-3b expression may be due to defective midlineShh signalling, rather than notochord or PFP degeneration. There-fore, we analysed expression of Patched and Gli1, which are alsoinduced by Shh signalling but are more widely expressed in theneural tube than HNF-3b (refs 22, 23). Expression of both Patchedand Gli1 was greatly reduced in SIL−/− embryos (Fig. 3m–p). As withHNF-3b, these molecular differences were seen before any grossnotochord or neural tube abnormalities (Fig. 3q, r). In contrast,Smoothened and Gli3, which are not induced by Shh signalling butare involved in the Shh pathway24,25, showed normal expression (notshown). The markedly reduced expression of multiple Shh targetgenes, despite Shh expression in the node and notochord, indicatesthat Shh signalling in the midline may be blocked in SIL−/− embryos.

A block in Shh signalling could explain several similarities

between the SIL−/− and mouse Shh knockout phenotypes, includingthe failure of floorplate development, notochord degeneration andholoprosencephaly. However, a pronounced difference is that Shh−/−

embryos have normal cardiac looping7. This finding has beenperplexing, because Shh has a role in left–right specification inthe chick26 and left–right development is highly conserved amongvertebrates. To determine whether defects in Shh signalling areinvolved in the abnormal left–right development of SIL−/− embryos,we examined heart looping in Shh mutants and found it to benormal (Table 1), confirming the original study7. Surprisingly, therewas bilateral expression of nodal, lefty-2 and Pitx2 in Shh−/−

embryos, although only in the anterior right LPM for lefty-2 andPitx2 (Fig. 2g–i). Not all Shh mutants had bilateral lefty-2 expres-sion. Those that did were among the older embryos examined(Table 1), indicating that lefty-2 turns on in the right LPM after left-sided expression has begun. Midline lefty-1 expression was notfound in any Shh mutant (Fig. 2h). Although these results showthat Shh is involved in regulating left–right gene expression in themouse, a block in Shh signalling alone cannot account for thecompletely symmetric bilateral pattern of nodal, lefty-2 and Pitx2expression found in SIL mutants.

In Xenopus embryos, removal of the ventral neural tube andnotochord before neural-tube closure leads to bilaterally symmetricnodal expression and randomized cardiac looping5,27. This indicatesthat more severe midline defects in SIL mutants than in Shh mutantsmight underlie the differences in left–right gene expression andrandomized heart looping. Examination of SIL mutant embryos byserial sectioning revealed not only abnormal histology but alsosignificantly fewer cells in the neural folds than normal (67 6 5 cellsper section in mutants, 107 6 5 in heterozygotes at mid-trunk level;P , 0:0001, t-test; Fig. 4a, b). Because SIL is expressed as animmediate early response to growth-factor stimulation9, reducedcell numbers might be due to reduced proliferation. However,bromodeoxyuridine (BrdU) labelling was similar in the neuralfolds of mutant and normal embryos (68:0 6 2:3% BrdU-positivecells in mutants, 63:5 6 2:7% in heterozygotes), as was the mitoticindex (1:79 6 0:3% in mutants, 1:76 6 0:2% in heterozygotes). Thelack of a proliferation defect led us to examine programmedcell death as a cause for decreased cell numbers. TUNEL analysisrevealed extensive apoptosis (Fig. 4c, d), almost exclusively in thedorsal midline, in the neural folds and somites. To explore furtherthe growth properties of SIL−/− cells, we injected SIL−/− and SIL+/−

embryonic stem cells subcutaneously into nude mice. The resultingteratomas were similar in size (1:3 6 0:4 g, n ¼ 7; 1:4 6 0:6 g,

Figure 2 Analysis of left–right asymmetric gene expression at E8.5. a, b, d, e, g

and h show dorsal views; c, f and i are ventral views. a, Normal nodal expression

in the left LPM. b, Normal lefty-1 expression in the left side of the PFP and lefty-2

expression in the left LPM. c, Normal Pitx2 expression in head mesenchyme,

heart and left LPM. d–f, In SIL−/− embryos nodal (d), lefty (e), and Pitx2 (f) are

bilaterally expressed. Note there is no lefty expression in the midline. g–i, In Shh

mutant embryos nodal (g), lefty (h) and Pitx2 (i) show a different pattern of bilateral

expression; again there is no midline lefty expression.

Table 1 Left–right development in SIL and Shh mutants

SILDirection of looping of embryonic heart tube

Right Left Ventral34 (46%) 32 (43%) 8 (11%)

Gene expression pattern in the LPMGene Somites Left Right Bilateral Absentlefty 3–4 2 3

5–6 1 107–8 3 2

nodal 3–5 56–8 6

Pitx2 6–7 78–10 6 2

.............................................................................................................................................................................

ShhDirection of looping of embryonic heart tube

Right Left Ventral29 (97%) 0 (0%) 1 (3%)

Gene expression pattern in the LPMGene Somites Left Right Bilateral Absentlefty 3–4 1 1

5–6 2 17–8 1 1

nodal 3–5 16–8 5 1

Pitx2 6–7 48–10 5

.............................................................................................................................................................................

© 1999 Macmillan Magazines Ltd

letters to nature

NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com 693

n ¼ 10, respectively) and histology, containing differentiated tissuesoriginating from all three germ layers (data not shown). Therefore,the lack of SIL function does not lead to a general growthdisadvantage, at least during the chaotic differentiation processesof teratoma formation. Rather, the localized apoptosis in SIL−/−

embryos indicates that SIL is involved in cell survival and growthwithin the context of the organized developmental program of themidline axial structures.

The early onset of lefty-2 expression in the right LPM of SILmutants before normal left-sided expression, and the expression ofnodal, lefty-2 and Pitx2 throughout the entire LPM of SIL mutants,contrast with the pattern seen in both Shh mutants and lefty-1mutant embryos14, in which right-sided expression is delayed andrestricted to the anterior LPM. These different patterns point to arole for SIL in additional, Shh- and lefty-independent pathwaysinvolved in the regulation of lefty-2 and Pitx2 expression in theright posterior LPM, from early somite stages. This is consistentwith a proposed model in which the midline actively represses theexpression of left–right asymmetric genes on the right27, ratherthan loss of a midline barrier14. Although lefty-1 mutants do showlaterality defects, most often left pulmonary isomerisms14, andembryos lacking Shh signalling function also have apparent leftisomerisms of the lung28,29, the direction of cardiac looping isnormal in lefty-1 and Shh mutants. We propose that the earlyonset and extended duration of nodal, lefty-2 and Pitx2 expressionthroughout the entire right LPM may be the key to why heartlooping is randomized in SIL mutants but not in Shh and lefty-1mutants. These observations may also explain why other mouse,human and zebrafish midline mutants show variable co-occurrenceand severity of left–right abnormalities. M

Figure 3 Analysis of Shh signalling. a–f and m–r show ventral views. Lines in a–f

indicate approximate level of transverse sections (10 mm) shown directly below in

g–l. Sections (g, h, k, l) were counterstained with nuclear fast red. a, g, Normal

E8.5 Shh expression in the notochord and gut endoderm. b, h, E8.5 SIL mutant

with discontinuous Shh expression in the notochord. c, i, Normal E8.5 HNF-3b

expression in the notochord, PFP and gut endoderm. e, k, Normal E7.5 HNF-3b

expression in the endoderm, notochordal plate and PFP. d, j, f, l, SIL−/− embryos

have reduced midline HNF-3b expression due to lack of PFP expression at E8.5

(d, j) and at E7.5 (f, l). Note the abnormal histology of the mutant neural folds (j).

m, q, Normal Patched expression at late E8.5 (m) and early E8 (q). n, r, Reduced

Patched expression in SIL−/− embryos at E8.5 (n) and early E8 (r). o, Late E8.5

normal Gli1 expression. p, Downregulation of Gli1 expression in a late E8.5 SIL−/−

embryo.

Figure 4 Analysis of cell proliferation and apoptosis at E8.5. a, b, Transverse

sections (10 mm; haematoxylin counterstained) from mid-trunk level of a normal

embryo (a) and a SIL mutant (b) analysed for cell proliferation by BrdU

incorporation (brown stain). c, d, Analysis of apoptosis by TUNEL assay done

in situ on sections adjacent to those analysed for BrdU uptake (c, normal embryo;

d, SIL mutant). Extensive apoptosis (green fluorescent nuclei) is seen in SIL−/−

axial and paraxial structures but not in gut or heart.

© 1999 Macmillan Magazines Ltd

letters to nature

694 NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Methods

Gene targeting, embryonic stemcells and generation ofmice. A 5 kilobaseEcoRI genomic fragment encompassing exons 3–5 of the mouse SIL gene,including putative translation start sites in exons 3 and 4 (ref. 10), was replacedby a PGK–NEO cassette. Gene targeting and generation of chimaeric mice weredone by standard methods30. Details of probes and primers for genotyping areavailable upon request.Generation of SIL−/− cells and teratomas. We generated embryonic stemcells homozygous for the SIL null mutation by culturing SIL+/− cells in thepresence of 4 mg ml−1 G418 for 4 days followed by maintenance in 1 mg ml−1

G418. Teratomas were generated by injecting 2.5 million embryonic stem cellsin 0.1 ml PBS subcutaneously into Swiss nude mice.Cell proliferation and apoptosis studies. E8 pregnant females were injectedintraperitoneally with BrdU (100 mg kg−1; Sigma). Embryos were dissected 1 hafter injection and fixed in Histochoice (Amresco) for 2–3 h at room temperature,mounted in 2% agarose in PBS to facilitate positioning, embedded in paraffin andcross-sectioned. BrdU was detected by monoclonal antibody (Becton Dickinson)following the manufacturer’s protocol. Apoptosis was detected using the In SituCell Death Detection Kit—Fluorescein (Boehringer Mannheim).In situ hybridization. Whole-mount in situ hybridization, plastic embeddingand sectioning were done as described11.

Received 5 February; accepted 5 May 1999.

1. Goldstein, A. M., Ticho, B. S. & Fishman, M. C. Patterning the heart’s left–right axis: from zebrafish toman. Dev. Genet. 22, 278–287 (1998).

2. Oh, S. P. & Li, E. The signaling pathway mediated by the type IIB activin receptor controls axialpatterning and lateral asymmetry in the mouse. Genes Dev. 11, 1812–1826 (1997).

3. King, T., Beddington, R. S. P. & Brown, N. A. The role of the brachyury gene in heart development andleft-right specification in the mouse. Mech. Dev. 79, 29–37 (1998).

4. Melloy, P. G. et al. No turning, a mouse mutation causing left–right and axial patterning defects. Dev.Biol. 193, 77–89 (1998).

5. Danos, M. C. & Yost, H. J. Role of notochord in specification of cardiac left–right orientation inzebrafish and Xenopus. Dev. Biol. 177, 96–103 (1996).

6. Chen, J. N. et al. Left-right pattern of cardiac BMP4 may drive asymmetry of the heart in zebrafish.Development 124, 4373–4382 (1997).

7. Chiang, C. et al. Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function.Nature 383, 407–413 (1996).

8. Aplan, P. D. et al. Disruption of the human SCL locus by ‘‘illegitimate’’ V-(D)-J recombinase activity.Science 250, 1426–1429 (1990).

9. Izraeli, S. et al. Expression of the SIL gene is correlated with growth induction and cellularproliferation. Cell Growth Differ. 8, 1171–1179 (1997).

10. Collazo-Garcia, N., Scherer, P. & Aplan, P. D. Cloning and characterization of a murine SIL gene.Genomics 30, 506–513 (1995).

11. Lowe, L. A. et al. Conserved left-right asymmetry of nodal expression and alterations in murine situsinversus. Nature 381, 158–161 (1996).

12. Collignon, J., Varlet, I. & Robertson, E. J. Relationship between asymmetric nodal expression and thedirection of embryonic turning. Nature 381, 155–158 (1996).

13. Heymer, J., Kuehn, M. & Ruther, U. The expression pattern of nodal and lefty in the mouse mutant Ftsuggests a function in the establishment of handedness. Mech. Dev. 66, 5–11 (1997).

14. Meno, C. et al. lefty-1 is required for left–right determination as a regulator of lefty-2 and nodal. Cell94, 287–297 (1998).

15. Piedra, M. E., Icardo, J. M., Albajar, M., Rodriguez-Rey, J. C. & Ros, M. A. Pitx2 participates in the latephase of the pathway controlling left–right asymmetry. Cell 94, 319–324 (1998).

16. Ryan, A. K. et al. Pitx2 determines left–right asymmetry of internal organs in vertebrates. Nature 394,545–551 (1998).

17. Yoshioka, H. et al. Pitx2, a bicoid-type homeobox gene, is involved in a lefty-signaling pathway indetermination of left-right asymmetry. Cell 94, 299–305 (1998).

18. Campione, M. et al. The homeobox gene Pitx2: mediator of asymmetric left–right signaling invertebrate heart and gut looping. Development 126, 1225–1234 (1999).

19. Ang, S. L. & Rossant, J. HNF-3 beta is essential for node and notochord formation in mousedevelopment. Cell 78, 561–574 (1994).

20. Weinstein, D. C. et al. The winged-helix transcription factor HNF-3 beta is required for notochorddevelopment in the mouse embryo. Cell 78, 575–588 (1994).

21. Roelink, H. et al. Floor plate and motor neuron induction by different concentrations of the amino-terminal cleavage product of sonic hedgehog autoproteolysis. Cell 81, 445–455 (1995).

22. Goodrich, L. V., Johnson, R. L., Milenkovic, L., McMahon, J. A. & Scott, M. P. Conservation of thehedgehog/patched signaling pathway from flies to mice: induction of a mouse patched gene byHedgehog. Genes Dev. 10, 301–312 (1996).

23. Hynes, M. et al. Control of cell pattern in the neural tube by the zinc finger transcription factor andoncogene Gli-1. Neuron 19, 15–26 (1997).

24. Akiyama, H. et al. Cloning of a mouse smoothened cDNA and expression patterns of hedgehogsignalling molecules during chondrogenesis and cartilage differentiation in clonal mouse EC cells,ATDC5. Biochem. Biophys. Res. Commun. 235, 142–147 (1997).

25. Ruiz i Altaba, A. Combinatorial Gli gene function in floor plate and neuronal inductions by Sonichedgehog. Development 125, 2203–2212 (1998).

26. Levin, M., Johnson, R. L., Stern, C. D., Kuehn, M. & Tabin, C. A molecular pathway determining left–right asymmetry in chick embryogenesis. Cell 82, 803–814 (1995).

27. Lohr, J. L., Danos, M. C. & Yost, H. J. Left–right asymmetry of a nodal-related gene is regulated bydorsoanterior midline structures during Xenopus development. Development 124, 1465–1472 (1997).

28. Litingtung, Y., Lei, L., Westphal, H. & Chiang, C. Sonic hedgehog is essential to foregut development.Nature Genet. 20, 58–61 (1998).

29. Motoyama, J. et al. Essential function of Gli2 and Gli3 in the formation of lung, trachea andoesophagus. Nature Genet. 20, 54–57 (1998).

30. Good, D. J. et al. Hypogonadism and obesity in mice with a targeted deletion of the NhIh2 gene.Nature Genet. 15, 397–401 (1997).

Acknowledgements. We thank H. Westphal and N. Takuma for Shh knockout mice; A.-M. Walsh and stafffor generation of chimaeras; J. Magrath and R. Mollina for animal care; G. Miller for pathology onteratomas; R. Dryfus for photography; P. Thomas for technical help; P. Aplan, H. Hamada, A. Joyner,C. Chiang, H. Hahn and M. Blum for providing probes; and M. Halpern, H. Arnheiter and B. Casey forcomments on the manuscript.

Correspondence should be addressed to M.R.K. (e-mail: mkuehn@box-m.nih.gov). Requests formaterials should be addressed to I.R.K. (e-mail: kirschi@exchange.nih.gov).

Regionsof variant histoneHis2AvD required forDrosophiladevelopmentMichael JohnClarkson*†, JulianR.E.Wells†‡, FrankGibson*,Robert Saint§ & David John Tremethick*

* The John Curtin School of Medical Research, the Australian National University,PO Box 334, Canberra, Australian Capital Territory 2601, Australia† Department of Biochemistry, and § Department of Genetics, University ofAdelaide, Adelaide, South Australia 5005, Australia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

One way in which a distinct chromosomal domain could beestablished to carry out a specialized function is by the localizedincorporation of specific histone variants into nucleosomes.H2AZ, one such variant of the histone protein H2A, is requiredfor the survival of Drosophila melanogaster1, Tetrahymenathermophila2 and mice (R. Faast et al., in preparation). Tosearch for the unique features of Drosophila H2AZ (His2AvD,also referred to as H2AvD) that are required for its essentialfunction, we have performed amino-acid swap experiments inwhich residues unique to Drosophila His2AvD were replaced withequivalently positioned Drosophila H2A.1 residues. MutatedHis2AvD genes encoding modified versions of this histone weretransformed into Drosophila and tested for their ability to rescuenull-mutant lethality. We show that the unique feature ofHis2AvD does not reside in its histone fold but in its carboxy-terminal domain. This C-terminal region maps to a short a-helixin H2A that is buried deep inside the nucleosome core.

Figure 1 shows the amino-acid sequence of Drosophila H2A.1(top) and the amino-acid differences found in the Drosophila H2AZprotein, His2AvD (bottom). A schematic representation of thesecondary structure of H2A is shown below the amino-acidsequence. In addition to the histone fold (the region encompassinga-helices 1 to 3), H2A contains two additional short helices thatflank the histone fold (helices a N and a C, respectively)3. Amino-acid differences between His2AvD and H2A span the wholesequence of His2AvD including the N and C termini (excludingthe extra-long C-terminal tail of His2AvD, His2AvD has a net gainof five serines and three threonines compared to H2A.1). Clearly,some or all of these amino-acid differences must specify the uniquefunction of His2AvD which cannot be provided by H2A. Therefore,to gain a molecular understanding of these special features thatdefine H2AZ function, we replaced the amino acids unique toHis2AvD, by changing the coding sequence, with H2A amino-acidresidues.

A deletion in the Drosophila His2AvD gene (His2AvD810) ishomozygous lethal1. However, the precise developmental step atwhich lethality occurs was unknown. Here, fly stocks containing theHis2AvD null allele, His2AvD810, were maintained against the TM6bbalancer chromosome which carries the dominant markers Tubbyand Humoral. Non-Tubby His2AvD810 homozygotes undergo aprotracted third instar and then die without entering pupation.

‡ Deceased