Here we describe the cloning of the human Achaete Scute Homologue 2 (HASH2) gene, officially designated ASCL2 (Achaete Scute complex like 2), a homologue of the DrosophilaAchaete and Scute genes. In mouse, this gene is imprinted and maps to chromosome 7. We mapped the human homologue close to IGF2 and H19 at 11p15.5, the human region syntenic with mouse chromosome 7, indicating that this imprinted region is highly conserved in mouse and man. HASH2 is expressed in the extravillus trophoblasts of the developing placenta only. The lack of HASH2 expression in non-malignant hydatidiform (androgenetic) moles indicates that HASH2 is also imprinted in man.
The Mash genes (mammalian Achaete-Scute homologues) are conserved mammalian cognates of the Drosophila Achaete-Scute complex (1 ). In rodents, two Mash genes have been identified so far: Mash1 is expressed in neuronal progenitors, while Mash2 is expressed in spongiotrophoblast cells and their precursors (2 ,3 ). Both Mash1 and -2 are members of the basic helix-loop-helix (bHLH) gene family, which includes MYC and MYOD (4 ). Mash1 and -2 function as lineage-specific transcription factors essential for development of the neurectoderm and trophectoderm, respectively.
Mice deficient for Mash2 die at 10 days post coitum (d.p.c.) due to placental failure as a consequence of deficient spongiotrophoblast formation (3 ). In contrast, chimaeric mice with heterozygous extraembryonic tissues and homozygous embryonic tissues are viable and normal, indicating that this gene, although essential for development of the placenta, is not important for development of the embryo proper.
In the mouse, the Mash2 gene is subject to genomic imprinting with only the maternal allele being active (5 ). Mice carrying a homozygously mutated Mash2 gene die at 10 d.p.c. as do heterozygous mice who inherited the mutant allele from their mother. Heterozygous mice who inherited the mutant allele from their father are viable. This phenomenon of paternal imprinting of a gene being critical for development of the placenta is interesting as nuclear transplantation studies in mice have shown that imprinting is biased towards embryonic versus extraembryonic tissues with preferential expression of the maternal and paternal alleles in the embryo and placenta, respectively (6 ). In this respect, Mash2 behaves as an exception to the common imprinting rule.
Mash2 maps to distal chromosome 7 and is closely linked to three other imprinted genes, Ins2, Igf2 and H19. The conserved syntenic human chromosome region is 11p15.5 and we have taken advantage of this fact to isolate the human homologue of the Mash2 gene, designated HASH2 (human Achaete Scute homologue 2). We show that HASH2 maps proximal to, and in close proximity of, IGF2 and the gene order in this region is HASH2-INS-IGF2-H19-tel. In addition, imprinting disorders such as the Beckwith-Wiedemann syndrome (BWS), hemihypertrophy, and various childhood tumors associated with these disorders are mapped to this region.
Expression studies showed that HASH2 is expressed in extravillus trophoblast cells only. The gene order and similar expression patterns in extraembryonic tissues of mouse and man suggest conservation of this imprinted region in mouse and man.
A 138 bp DNA probe was generated by PCR using mouse Mash2 cDNA as template. This PCR product contained the basic helix-loop-helix region of the mouse Mash2 gene and was used to screen a human chromosome 11 specific cosmid library (7 ). Four positive cosmids (cSRL35E9, cSRL128B6, cSRL130H5 and cSRL147H5) were found, which were overlapping as determined by restriction enzyme fragment patterns (data not shown). The 138 bp mouse fragment hybridized to a 210 bp PstI and a 6.0 kb EcoRI genomic DNA fragment in all four cosmids (Fig. 1 ). Sequence analysis of the 210 bp PstI fragment, isolated from cSRL35E9, showed it contained the conserved bHLH region. In order to extend the coding sequence we partially sequenced the 6.0 kb EcoRI fragment using primers derived from the bHLH region and performed 5' RACE as detailed in the Materials and Methods. Three independent reactions were carried out and in each case the RACE clones are identical to the genomic sequence. Sequence comparison to the mouse Mash2 gene suggest that the mRNA transcript from the human gene contains sequences homologous to the mouse intron 1 (Fig. 2 C). We cannot eliminate the possibility that the RACE kit contained genomic DNA but three independent isolates of the same RACE clone from non-amplified starting material argues that this is unlikely. Mouse Mash2 is encoded by two exons, separated by a 251 bp intron. Exon 1 encodes the first three amino acids (MEA), the fourth amino acid (H) is formed by the splice junction and amino acids 5-253 are encoded by exon 2. The HASH2 sequence does not seem to be spliced.
It was our expectation that HASH2 would be located in human chromosome region 11p15.5, within the known conserved syntenic region with mouse chromosome 7. We therefore determined the localization of the HASH2 cosmids by fluorescent in situ hybridization (FISH) of human metaphase spreads. As predicted, HASH2 hybridized to the short arm of chromosome 11 (11p15.5) (Fig. 3 A). In addition, the gene was mapped between two balanced translocation breakpoints at 11p15.5 (11 ), placing HASH2 distal to the Beckwith-Wiedemann syndrome chromosome region 1 (BWSCR1) and proximal to HRAS1 (Fig. 3 B and C).
Murine Mash2 is expressed exclusively in the spongiotrophoblast in postimplantation stages and is paternally imprinted. To establish if HASH2 expression exhibited similar properties, RNA in situ hybridization was performed on early human extravillus trophoblast cells (the human equivalent of rodent spongiotrophoblast) of normal material and complete molar pregnancies, which are exclusively of androgenetic origin. We would expect HASH2 expression to be identifiable in normal trophoblast but not in other embryonic tissues. Complete molar pregnancies are entirely derived from the paternal germ line and we reasoned that if HASH2 is paternally imprinted, it may not be expressed in such cells.
In early normal human placentae, HASH2 expression was found in extravillus trophoblast cells only (Fig. 5 A). In contrast, in normal placentae, chorionic villus trophoblast cells, placental stromal cells, maternal placental bed cells and human embryos of Carnegie stages 19-21, HASH2 is undetectable. For comparison, control hybridizations performed on identically processed (extra)embryonic mouse tissues (12.5 d.p.c.) with a 132 bp Mash2 mouse probe showed expression in spongiotrophoblast cells only (Fig. 5 B). We conclude from these analyses that HASH2 expression is broadly similar to Mash2 expression in the developing embryo.
In this study, we report the isolation of the human homologue of Mash2, an imprinted gene coding for a lineage-specific transcription factor essential for development of the placenta. HASH2 maps to 11p15.5, the human syntenic region of mouse chromosome 7. The gene order HASH2-INS2-IGF2-H19-tel suggests conservation of this imprinted region in mouse and man. Recently, an additional gene on mouse chromosome 7 and human 11p15.5, p57kip2, was shown to be imprinted in both mouse and human (16 -19 ). In human this gene maps close to, and proximal to, IGF2 and the Beckwith-Wiedemann syndrome chromosomal region 1 (BWSCR1; 20 ), which extends the imprinted region further centromeric. This region thus includes the breakpoints of patients with Beckwith-Wiedemann syndrome, a syndrome associated with genomic imprinting.
In the embryo, expression of HASH2 is limited to the extravillus trophoblast, which is a similar exertion profile to the murine Mash2 gene. In mice, only the maternal Mash2 allele is active in spongiotrophoblast cells (the rodent equivalent of human extravillus trophoblast) at 8.5 d.p.c. The lack of expression of HASH2 in complete molar pregnancies, with two sets of paternal genes, is a strong indication, although an indirect one, for paternal imprinting of the HASH2 gene. Uniparental tissues have been shown to be useful for the study of imprinted genes. Both human and mouse uniparental tissues maintain the IGF2, IGF2r and MEST expression patterns, and the ZNF127 methylation pattern, which correlate with their parental origin (21 -24 ). However, in some cases expression of the paternally imprinted H19 gene was observed in hydatidiform moles (13 ,21 ,25 ). This can be explained by the finding that H19 is biallelically expressed in a subset of cells (extravallus trophoblasts) in the placenta (13 ,25 -27 ). Our experiments show that HASH2 is imprinted in non-invasive moles and that in malignant moles subsets of cells express HASH2, which might be due to loss of imprinting in these cells. Additional evidence for HASH2 imprinting could be provided by allele specific expression studies, but for these assays RNA polymorphisms are required, which are not yet available for HASH2.
If HASH2 is expressed only from the maternal allele this would mean that uniparental paternal disomy (UPD) of 11p15 is not viable. UPD is often seen in BWS patients and patients with hemihypertrophy. In most, if not in all, cases this UPD is not complete, but present in mosaics (28 ). An explanation for this finding might be that cases of complete paternal disomy of 11p15 in extraembryonic tissues is not viable, because of lack of HASH2 expression. In cases of complete UPD in BWS the disomy might be restricted to the embryo proper and not be present in the extraembryonic tissues. No data are available on this subject, therefore it would be interesting to determine whether normal cells are still present in placentae of patients with complete chromosome 11 UPD.
In early human pregnancy, extravillus trophoblasts actively and selectively migrate into the maternal tissues leading to selective adaptations of the maternal spiral arteries (29 ). Abnormal extravillus trophoblast differentiation is thought to be causative in pre-eclampsia, a pregnancy disorder with high morbidity and mortality which affects 7-10% of all pregnancies (29 ). Premature differentiation of extravillus trophoblast cells as reflected by defective integrin switching (no upregulation of [alpha]1 integrin) leads to incomplete spiral artery adaptation in pre-eclampsia and sets the stage for a cascade of maternal and fetal (de)compensation leading ultimately to hypertension, edema and proteinuria around weeks 26-28. In complete molar pregnancies, which are androgenetic in origin and characterized by the presence of hydropic placental tissues with highly proliferative trophoblast cells in the absence of embryonic tissues, a severe early form of pre-eclampsia occurs (30 ).
It will be interesting to see whether HASH2 expression or regulation is aberrant in the above-mentioned clinical disorders or in early intrauterine fetal death in general.
Plasmid 33.3 containing a nearly full-length cDNA insert of the mouse Mash2 gene was amplified in the presence of Mash2 primers (5' cgc gag cgc aac cgc gta 3' and 5' cag cgc acg aat gta ctc tac cgc gga 3') followed by cloning into pBluescript SKII. This yielded clone 531-14 containing a 138 bp insert homologous to the bHLH region of the Mash2 gene. This fragment was excised from the plasmid and used to screen a human chromosome 11 specific cosmid library (7 ) arrayed on filters, using the methods detailed in Ivens et al. (31 ).
RACE reactions were performed using Clontech's placenta marathon ready cDNA kit, according to the manufacturer's conditions. The gene specific primers used are R6: AAG TTC ACC AGC TTC ACG C for the first PCR and R5: AGG TGC AGG CAG AGG AAC or R4: GAC GGG GAA AAC TGT GG for the nested PCR.
Cosmids were labeled by nick-translation. Hybridization was performed on metaphase chromosomes, as described previously (32 ). In brief, slides were, after denaturation of target DNA, incubated overnight at 37oC with 10 [mu]l of hybridization mixture (50% formamide, 50 mM sodium phosphate, 10% dextran sulfate in 2* SSC) containing 50 ng of biotin- or digoxigenin-labeled denatured cosmid probe and 2 ng of centromere-11 specific probe under a sealed coverslip (24 * 24 mm2). Slides were washed in 50% formamide in 2* SSC followed by 2* SSC, both at 42oC. Biotin-labeled hybrids were detected by Fitc-conjugated avidin followed by amplification using goat-anti avidin and a second layer of Fitc-conjugated avidin. Digoxigenin-labeled hybrids were detected by subsequent incubation with mouse anti-digoxigenin, digoxigenin-conjugated sheep anti-mouse, and Tritc-conjugated mouse anti-digoxigenin. Chromosomes were counterstained with propidium iodide or DAPI.
Cell lines: probes were mapped to cell lines B12 and B8, carrying the balanced translocations t(11;12)(p15.5;q24.11) and t(11;16)(p15.5;q12), respectively (11 ).
Sequence reactions were performed using the ABI PRISMTM Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer) and following manufacturer's recommended conditions. Reactions were run on a ABI 377 automatic sequencer (Perkin Elmer).
Normal human lymphocytes were embedded in agarose, Proteinase K treated and digested overnight with SalI, NotI or NruI. Gels were run for 24 h with pulse time 150 s and 24 h with pulse time 90 s. Size-separated DNA was transferred to HybondN+ membranes. Hybridization with a 32P labeled 2.5 kb TaqI fragment derived from cosmid cSRL130H5 was carried out overnight at 42oC in hybridization mixture consisting of 40% formamide, 0.1% SDS, 5* SSC, 1* Denhardt's, 20 mM sodium phosphate and 0.24 mg/ml denatured herring sperm DNA. Filters were washed in 2* SSC, 0.1% SDS at 60oC.
Formalin-fixed, paraffin-embedded human extraembryonic tissues (first trimester) containing extravillus trophoblast of normal (n = 3) and androgenetic origin (n = 6) were obtained and processed as described previously (33 ). Molar pregnancies included tissues from complete hydatidiform moles without (n = 3) and with malignant potential (n = 3) (invasive mole = choriocarcinoma villosum). In addition, normal human embryos of Carnegie stages 19-21 (n = 3) and normal mouse embryos (12.5 d.p.c.) were obtained. Digoxigenin-labeled RNA probes were generated by promoter-mediated in vitro transcription using T3 and T7 RNA polymerase as described (33 ) from BamHI- or HindIII-linearized HASH2 plasmids containing a 211 bp PstI fragment of the HASH2 gene. For controls, HLA-G riboprobes were generated identically, except that T7 and SP6 RNA polymerase were used. Prior to use, size, yield and integrity of riboprobes were checked by chemiluminescent analysis of size-separated immobilized RNA (33 ). For non-radioactive in situ hybridization, formalin-fixed sections (4 [mu]m) on APES-coated slides were heated for 60 min at 60oC, dewaxed for 60 min in xylene at 50oC and rehydrated. Following blocking in methanol/hydrogen peroxide (0.3%) for 10 min and digestion with Proteinase K (Merck) for 30 min at 37oC in 0.1 M Tris, pH 8.0, 50 mM EDTA, slides were acetylated, washed in 2* SSC and dehydrated. Dried sections were covered with hybridization mix (25 [mu]l) containing 50% formamide, 50 mM Tris, pH 7.5, 5 mM EDTA, 0.5% NaPi, 0.05% SDS, denatured salmon testis DNA (Sigma) at 1 mg/ml, 0.05* SSC (HASH2) or 0.5* SSC (HLA-G), 0.2% PVP-360, 5% dextran sulfate, and riboprobes at saturating levels (2-3 [mu]g/ml.kb). Following sealing under acid-cleaned coverslips, slides were heated for 5 min at 80oC and hybridized overnight at 50oC. Following high stringency washes at 60oC in 0.1* SSC, and RNase treatment using RNase One (Promega), hybrids were detected as described (34 ) using Fab fragments of sheep anti-digoxigenin antibodies labeled with peroxidase (Boehringer Mannheim) followed by visualization using silver enhancement of DAB/Ni complexes. Negative controls with sense probes showed absence of signals.
This work was supported by the Netherlands Organization of Scientific Research (NWO grant 504-111), the European Community (EC grant 89(4)0521) and the Praeventionfund (grant 28-2372). Work in PFRL's laboratory was supported by an MRC Human Genome Mapping Project PhD studentship to MH and by grants from the MRC and Cancer Research Campaign.
*To whom correspondence should be addressed. Tel: +31 20 566 5168; Fax: +31 20 691 8626; Email: m.a.mannens@amc.uva.nl
Human Molecular Genetics
Pages
Introduction
Results
Cloning of the HASH2 gene
Localization of HASH2
Expression pattern of HASH2
Discussion
Materials And Methods
Cosmid library screening
Rapid amplification of cDNA ends (RACE)
Fluorescent in situ hybridization
Sequencing
Pulsed field gel electrophoresis
RNA in situ hybridization
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Mon May 12 18:10:04 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996