The identities of the second through ninth abdominal segments of Drosophila are specified by two genes of the bithorax complex (BX-C), abdominal-A (abd-A) and Abdominal B (Abd-B). The correct deployment of these two genes requires an extensive region (the iab region) located between the two protein-coding transcription units. We show here that one iab mutation affects the pattern of expression of Abd-B. We also show that most or all of the DNA in this regulatory iab region is transcribed. In blastoderm stage embryos we can define three distinct domains within the iab DNA, each transcribed in a region that extends from a characteristic anterior limit to the posterior end of the segmented part of the embryo. The anterior limits of expression for the three regions are colinear with the sequence of the domains on the chromosome, and lie at about two-segment intervals. We suggest that these early transcription patterns reflect the initial activation of the BX-C.

The identities of some of the thoracic and all the abdominal segments of Drosophila are specified by the bithorax complex (BX-C) (Lewis, 1978). The BX-C contains three genes, Ultrabithorax (Ubx), abdominal-A (abd-A) and Abdominal-B (Abd-B), which in combination determine the characteristics of each segment within the BX-C domain (Sánchez-Herrero et al. 1985,a,b; Tiong et al. 1985). Mutations in any of these genes transform a series of parasegments (PS) (Martinez-Arias and Lawrence, 1985) which coincide with the anatomical region where the products of the genes are expressed (reviewed in Duncan, 1987).

The complex genetics of the BX-C can be simplified if we classify BX-C mutations into two broad groups: those that directly alter the products of the known protein-coding transcription units, and those that affect putative regulatory elements. For example, in the Ubx domain, Ubx mutations eliminate or alter the expression of the Ubx protein throughout the embryo (Beachy et al. 1985; Weinzierl et al. 1987), whereas the regulatory mutations abx and bxd, which transform just a subset of the parasegments affected in Ubx mutations (Hayes et al. 1984; Casanova et al. 1985), change the expression of Ubx in specific regions: abx in PS5 and bxd in PS6-13 (White and Wilcox, 1985; Cabrera et al. 1985; Beachy et al. 1985).

Within the abdominal region of the BX-C, the infra abdominal (iab) mutations each transform a subset of the segments (or parasegments) affected by abd-A or Abd-B mutations (Lewis, 1978, 1981; Kuhn et al. 1981; Sánchez-Herrero et al. 1985; Karch et al. 1985). With the exception of the iab-2 class, all the iab mutations map between the abd-A and Abd-B transcription units (Karch et al. 1985). All of these mutations fail to complement abd-A and Abd-B mutations (Karch et al. 1985), therefore implying that they act in cis to alter the expression of these genes (Karch et al. 1985; Casanova et al. 1987; Peifer et al. 1987; Tiong et al. 1987; Sánchez-Herrero et al. 1988). It is not known how the iab elements specify unique parasegmental patterns of expression for abd-A and Abd-B.

We show here by in situ hybridization to embryonic tissue sections that the iab region is transcribed. The patterns observed at blastoderm follow an anteroposterior order of expression and suggest an initial double-parasegment subdivision for the activation of the BX-C. We also demonstrate that an iab mutation changes the distribution of Abd-B products.

In situ hybridization

Whole lambda phages or gel-purified fragments were labelled with 35S by the oligo-random-priming method (Feinberg and Vogelstein, 1983) to specific activities of 4–19×108 disints-min-1μg-1. 1–1·5×105disintsmin-1μl-1 were used. In situ hybridization was as previously described (Sánchez-Herrero and Crosby, 1988). The Abd-B probe used is a 2-3 kb Abd-B cDNA (Hoey et al. 1986) that includes the homeobox. This probe hybridizes to sequences that are common to all classes of Abd-B transcripts (E.S. unpublished). The limits of expression of the different probes at stages later than blastoderm were ascertained by hybridizing alternate sections with Ubx, abd-A or Abd-B probes. Control experiments with wildtype lambda showed no specific signal at any stage of development.

Mutations

The iab-7MX2 mutation (previously called Abd-BMX2) is a chromosome break lying between +139-5 and 142 kb (Karch et al. 1985). It transforms abdominal segments(A) 5 to 7 (or PS 10 to 12) into segment A4 (or PS9) (Sánchez-Herrero et al. 1985; Duncan, 1987).

We have analyzed transcription of the entire abdominal region of the BX-C (+15 to +200 kb) (Karch et al. 1985) by hybridizing probes prepared from genomic DNA fragments to sections of Drosophila embryos. The rationale was to identify previously unknown transcription units that might be active in only a few cells or segments. To our surprise, we found that the majority of tested fragments showed a distinct and spatially restricted pattern of hybridization, although, in general, we did not find transcripts localized to specific abdominal segments.

Transcription patterns at the blastoderm stage

At blastoderm, only fragments to the left of the abd-A transcription unit (+15 to +30 kb) and those in the distal end of the Abd-B transcription unit (+192 to + 199 kb) showed no signal (see also Kuziora and McGinnis, 1988). At this stage, probes from +30 to +58 kb showed the expected pattern of expression corresponding to the abd-A gene (Rowe, 1987 and E. S., unpublished) and probes from +151 to +192 kb showed patterns corrresponding to those of the Abd-B gene (Harding and Levine, 1988; Sánchez-Herrero and Crosby, 1988; DeLorenzi et al. 1988; Kuziora and McGinnis, 1988). Probes from the iab region (+58 to + 150 kb) in between these genes showed one of three distinct patterns. Between +58 and +86 kb (domain I), all probes hybridized to a region of the blastoderm extending from about 38% to 10% egg length. From +99 to +125 kb (domain II), all probes hybridized to an overlapping but more restricted region, approximately from 28% to 10% egg length. Fragments from +125 to +150 kb (domain III) hybridized only to the region between around 20% and 10% egg length (Figs 1·3). By hybridizing alternate sections with iab and fushi tarazu (ftz) probes we confirmed that these three DNA domains are transcribed in regions that correspond approximately to the primordia for parasegments 9—15, 11–15 and 13–15, respectively (Martinez-Arias and Lawrence, 1985).

Fig. 1.

Map of the abdominal region of the BX-C according to Karch et al. 1985, showing the probes that detect specific transcription patterns at blastoderm (numbers above long probes correspond to lambda phages). Probes that detect each of the three different iab patterns are distinctively shaded. Probes that hybridize to abd-A or Abd-B transcripts are shown by black bars; those that detect no transcription at blastoderm are shown by white bars. The numbers below each iab region indicate anterior and posterior limits of expression in percentage of egg length (EL; 0% corresponds to the posterior pole).

Fig. 1.

Map of the abdominal region of the BX-C according to Karch et al. 1985, showing the probes that detect specific transcription patterns at blastoderm (numbers above long probes correspond to lambda phages). Probes that detect each of the three different iab patterns are distinctively shaded. Probes that hybridize to abd-A or Abd-B transcripts are shown by black bars; those that detect no transcription at blastoderm are shown by white bars. The numbers below each iab region indicate anterior and posterior limits of expression in percentage of egg length (EL; 0% corresponds to the posterior pole).

Fig. 2.

Measures in percentage of egg length (EL) of the different probes used. Each bar corresponds to a different probe and the bars arc shaded as in Fig. 1, according to the group to which the probe belongs. The first two bars in each group correspond to lamda phages (indicated by their numbers) and the rest of the probes are ordered according to their proximo-distal (left-right) order on the chromosome (see Fig. 1). The numbers to the right indicate the number of hcmi-blastoderms studied. The limits of transcription vary slightly along the dorso-ventral axis, but no attempt was made to control for this. This variation will account in part for the differences in measurements within each group.

Fig. 2.

Measures in percentage of egg length (EL) of the different probes used. Each bar corresponds to a different probe and the bars arc shaded as in Fig. 1, according to the group to which the probe belongs. The first two bars in each group correspond to lamda phages (indicated by their numbers) and the rest of the probes are ordered according to their proximo-distal (left-right) order on the chromosome (see Fig. 1). The numbers to the right indicate the number of hcmi-blastoderms studied. The limits of transcription vary slightly along the dorso-ventral axis, but no attempt was made to control for this. This variation will account in part for the differences in measurements within each group.

Fig. 3.

Patterns of expression at blastoderm of probes belonging to the three different iab regions. In this figure and in the following ones the anterior part of the embryo is to the left. (A) Hybridization observed with a 6-5 kb EcoRI fragment from +79-5 to +86kb, characteristic of domain I (+58 kb to +86kb). (B) Pattern observed with probes of domain II (+99 to + 125kb); the probe is a 8kb HindIII fragment from +113 to +121 kb. This section is alternate to the previous one. During and after gastrulation probes from part of region II, +99 to +113, detect stronger signal in the posterior cells that express this pattern. (C) Expression observed with a 6kb EcoRI fragment from +131·5 to+137·5 kb; this is characteristic of probes from +125 to +150kb (domain III).

Fig. 3.

Patterns of expression at blastoderm of probes belonging to the three different iab regions. In this figure and in the following ones the anterior part of the embryo is to the left. (A) Hybridization observed with a 6-5 kb EcoRI fragment from +79-5 to +86kb, characteristic of domain I (+58 kb to +86kb). (B) Pattern observed with probes of domain II (+99 to + 125kb); the probe is a 8kb HindIII fragment from +113 to +121 kb. This section is alternate to the previous one. During and after gastrulation probes from part of region II, +99 to +113, detect stronger signal in the posterior cells that express this pattern. (C) Expression observed with a 6kb EcoRI fragment from +131·5 to+137·5 kb; this is characteristic of probes from +125 to +150kb (domain III).

Probes from domains I and II detect transcripts from early cycle 14, just before the nuclei begin to elongate. The hybridization signals observed with these probes remain stronger than those obtained with comparable abd-A and Abd-B probes until late cycle 14, when the levels of abd-A and Abd-B transcripts increase substantially. Transcripts from domain III are first detected somewhat later, in mid-cycle 14; hybridization signals with domain III probes are never as strong as those observed with comparable Abd-B probes on the same embryos. Probes from the three domains detect signal first and predominantly in the nuclei, although later signal appears also over the cortical cytoplasm. Both the ectoderm and mesoderm shows the same spatially restricted expression of iab transcripts during gastrulation.

We obtained comparable signals with probes derived from whole lambda clones and with some probes derived from shorter subcloned fragments. In each case several non-overlapping fragments within each domain gave the same pattern of hybridization as the longer probes (Fig. 1). This makes it unlikely that the transcripts are detected by virtue of spurious hybridization to repeated sequences (No extensive repetitive sequences have been reported in the iab region, Karch et al. 1985). At least in one case (+113− +121 kb, within domain II), single-stranded probes from both strands detect transcripts with the same spatial distribution, implying the existence of at least two similarly regulated promoters.

Transcription after blastoderm

The iab region continues to be transcribed throughout embryogenesis, but the spatial patterns of expression change as development proceeds (Fig. 4). Probes spanning the whole iab region (from +58 to +150 kb) hybridize strongly to transcripts in PS13-15. When the germ band is extended both ectoderm and mesoderm show expression (Fig. 4C,E). Later, the signal is particularly strong in the ventral nerve cord, although the ectoderm is also labelled (Fig. 4D,F). We can attribute no function to these transcripts, as chromosome breaks in the iab region have no apparent effect on the development of PS13-15 (Karch et al. 1985). We do not consider them further. Probes in the iab-3 and iab-4 regions detect a more complex pattern superimposed on this basic one. Probes from +57 to +67kb show, in addition to strong expression in PS13–15, weaker signal in PS 8–12 (Fig. 4A). A probe extending from +79-5 to +86 kb also detects, at the same stages, transcripts anterior to PS13, extending to PS 8 or 9 (Fig. 4B). The anterior limits of expression of these two patterns seem to correlate with the most anterior parasegment affected by iab-3 and iab-4 mutations (Karch et al. 1985). The intervening DNA (from +67 to +79-5 kb) shows little or no hybridization in the region anterior to PS 13. We have made no attempt to characterize in detail the transcription units from which these transcripts arise.

Fig. 4.

Patterns of expression observed at stages of development later than blastoderm. Embryos oriented so that anterior is to the left and dorsal is up. (A) Signal observed in an embryo with the germ band extended probed with a 1·8 kb EcoRI fragment from +57 to +58·8 kb; signal is detected from PS 8 to 15, more strongly in PS13-15. (B) Expression detected with a 6·5 kb EcoRI fragment from +79·5 to +86 kb. Signal in the ventral cord extends from PS8 or 9 to PS15, with stronger expression in PS13–15. (C) Expression detected at the germ band extended stage by a 8 kb HindIII fragment from +113 to + 121 kb; signal is restricted to PS13–15. (D) Signal observed at late stages of embryogenesis with a 5 kb EcoRI fragment from +94-5 to +99·5 kb. The strong expression in the ventral cord is also limited to PS13–15. (E) A 6 kb EcoRI fragment from +131·5 to +137·5 kb also detects expression just in PS13–15 when the germ band is extended. (F) When the germ band is retracted the same fragment detects signal mainly in the ventral cord, also in PS 13–15.

Fig. 4.

Patterns of expression observed at stages of development later than blastoderm. Embryos oriented so that anterior is to the left and dorsal is up. (A) Signal observed in an embryo with the germ band extended probed with a 1·8 kb EcoRI fragment from +57 to +58·8 kb; signal is detected from PS 8 to 15, more strongly in PS13-15. (B) Expression detected with a 6·5 kb EcoRI fragment from +79·5 to +86 kb. Signal in the ventral cord extends from PS8 or 9 to PS15, with stronger expression in PS13–15. (C) Expression detected at the germ band extended stage by a 8 kb HindIII fragment from +113 to + 121 kb; signal is restricted to PS13–15. (D) Signal observed at late stages of embryogenesis with a 5 kb EcoRI fragment from +94-5 to +99·5 kb. The strong expression in the ventral cord is also limited to PS13–15. (E) A 6 kb EcoRI fragment from +131·5 to +137·5 kb also detects expression just in PS13–15 when the germ band is extended. (F) When the germ band is retracted the same fragment detects signal mainly in the ventral cord, also in PS 13–15.

The iab-7MX2mutation affects the expression of Abd-B products

To see if the iab region controls the expression of Abd-13 we have looked at the distribution of Abd-B transcripts in the mutation iab-7MX2. This mutation is an inversion with a breakpoint at about +140 that separates the iab-3-iab-7 region from the Abd-B transcription unit (Fig. 5A) (Karch et al. 1985), and transforms A5-A7 (or PS10–12) into A4 (or PS9; Fig. 5D) (Sánchez-Herrero et al. 1985; Duncan, 1987). We hybridized an Abd-B cDNA (Hoey et al. 1986) or an Abd-B genomic clone (the phage 8082, Karch et al. 1985) to embryos from the stock iab-7MX2/TM1. In wild-type embryos these two probes show the expression of all Abd-B transcripts. Most of the embryos from the iab-7MX2/TM1 stock showed the normal Abd-B pattern of expression (Fig. 5C), extending from PS10 to PS15 in late stages (Harding et al. 1985; Hoey et al. 1986; Harding and Levine, 1988; Sánchez-Herrero and Crosby, 1988; Kuziora and McGinnis, 1988). In about one quarter of the germ-band-retracted embryos (23/ 114) we observe that the expression of Abd-B is restricted to PS13-15 (Fig. 5E) in the epidermis and ventral cord. The parasegmental nature of this restricted expression is revealed in those sections that show segmental grooves. The normal extension of Abd-B expression into PS10–12 is therefore prevented by the separation of regulatory sequences that control the expression of Abd-B products. The control is over the transcripts corresponding to the m element of Abd-B (Casanova et al. 1986), since these are the only Abd-B transcripts expressed in PS10–12 (Sánchez-Herrero and Crosby, 1988; Kuziora and McGinnis, 1988). Alternate sections for three of the embryos with restricted Abd-B expression were hybridized with an abd-A probe. No change was observed with respect to the normal abd-A pattern.

Fig. 5.

Correspondence between spatial regulation of Abd-13 products and phenotypic transformation of an iab-7 mutation. (A) Map of the distal part of the iab region with the approximate location of the iab7MX2 mutation (Karch et al. 1985). The hatching indicates the DNA region that is separated from the promoter of the transcripts correponding to the m element of Abd-B (see Kuziora and McGinnis, 1988). (B) Abdomen of a wild-type male. Numbers to the right represent parasegments. Abdominal segments (A)l and 4 are indicated for reference. (C) Abd-B expression in an embryo from the iab-7MX2/TM1 stock that has completed germ band retraction. The signal in the ventral cord extends from PS10–15, as in wild-type embryos, and we think that the genotype of this embryo is iab-7MX2/TM1or TM1 /TM1. (D) Abdomen of a male of genotype iab-7MX2/ DfP9. PS10–12 are transformed into PS 9. We assume the adult transformation to be parasegmental, like the embryonic one (Duncan, 1987). (E) Abd-B expression in an embryo from the iab-7MX2/ TMI stock. The embryo is in a similar orientation and stage to that in C. The signal in the ventral cord is now restricted to PS13-15, and we believe the genotype is iab-7MX2/iab-7MX2. The cmbryos and E are from the same slide, so all experimental conditions are the same.

Fig. 5.

Correspondence between spatial regulation of Abd-13 products and phenotypic transformation of an iab-7 mutation. (A) Map of the distal part of the iab region with the approximate location of the iab7MX2 mutation (Karch et al. 1985). The hatching indicates the DNA region that is separated from the promoter of the transcripts correponding to the m element of Abd-B (see Kuziora and McGinnis, 1988). (B) Abdomen of a wild-type male. Numbers to the right represent parasegments. Abdominal segments (A)l and 4 are indicated for reference. (C) Abd-B expression in an embryo from the iab-7MX2/TM1 stock that has completed germ band retraction. The signal in the ventral cord extends from PS10–15, as in wild-type embryos, and we think that the genotype of this embryo is iab-7MX2/TM1or TM1 /TM1. (D) Abdomen of a male of genotype iab-7MX2/ DfP9. PS10–12 are transformed into PS 9. We assume the adult transformation to be parasegmental, like the embryonic one (Duncan, 1987). (E) Abd-B expression in an embryo from the iab-7MX2/ TMI stock. The embryo is in a similar orientation and stage to that in C. The signal in the ventral cord is now restricted to PS13-15, and we believe the genotype is iab-7MX2/iab-7MX2. The cmbryos and E are from the same slide, so all experimental conditions are the same.

20 blastoderm stage embryos from the same stock were hybridized with a probe corresponding to iab pattern II (the 8 kb Hmdlll fragment used in wild-type embryos, see Fig. 3). None of the embryos presented any apparent change in the pattern of expression compared to wild-type.

The genetics of the BX-C suggests that the DNA of the iab region controls the expression of abd-A and Abd-B products. Genetic tests indicate that iab mutations act essentially in cis. For example, the mutation iab-7MX2 when homozygous or in trans with a deficiency for Abd-B transforms PS10–12 to PS9. We have shown that in iab-7MX2 homozygous embryos Abd-B is not transcribed in PS 10-12. This observation supports the hypothesis that the iab elements act at the DNA level as distant regulatory elements to control abd-A and Abd-B (Karch et al. 1985; Casanova et al. 1987; Peifer et al. 1987; Tiong et al. 1987; Sánchez-Herrero et al. 1988).

Our results also reveal the unexpected finding that the iab region is itself transcribed and shows restricted patterns of expression. We discuss below the nature of these transcripts, and their significance with respect to the function of the BX-C.

Nature of iab transcripts

At present we do not know the number or structure of iab transcription units. In parallel experiments to those reported here, M. Crosby examined RNA from staged embryos by filter hybridization, using probes from domains II and III (+103—I-140 kb) and probes from the Abd-B transcription unit. Abd-B transcripts were readily detected (Sánchez-Herrero and Crosby, 1988) but, in the iab region, only weak smears of high molecular weight RNA were seen. No discrete transcripts were sufficiently abundant to be reproducibly detectable (M. Crosby, personal communication).

A possible clue as to the nature of these transcripts is provided by the similarities that they show with the bithoraxoid (bxd) transcripts of the Ultrabithorax domain (Hogness el al. 1985; Lipshitz et al. 1987). Both bxd (Akam et al. 1985; Irish et al. 1989) and iab transcripts are detected early in cycle 14, before transcripts from the adjacent protein-coding genes; both show a similar cellular distribution, first in the nuclei and later also in the cortical cytoplasm. The bxd transcripts derive from a long (25 kb) transcription unit, and are inefficiently processed into a complex family of variably spliced transcripts that appear not to code for proteins (Lipshitz et al. 1987). We suspect that the iab transcripts may have a similar organization.

The distribution of the iab transcripts does not appear to be consistent with their having a role as parasegment-specific products mediating iab functions. We would expect such products to be expressed at peak levels in those segments where they are required. Although the regions of expression of the early iab transcripts are in the same linear order as the effects of iab mutations, only very low levels of transcript (if any) are present in the most anterior segment affected by iab mutations within each domain. Thus mutations that affect PS10 and 11 (or A5 and A6) map within domain II, but domain II transcripts can be detected at high levels only from about PS11 backwards. A similar ‘offset’ is observed in the relationship between mutant effect and transcript distribution in the bxd region of Ubx (Akam et al. 1985; Irish et al. 1989).

The lack of a direct relationship between iab transcripts and function is supported by the observation that the accumulation of domain II transcripts appears to be normal in iab-7MX2 homozygous embryos. In this mutation domain II is intact (the chromosome break lies in domain III, which is not detectably transcribed in PS10–12). Thus the mutant phenotype in PS10–12 (and the lack of expression of Abd-B in these parasegments) is not likely to be caused by any disruption of domain II transcripts per se. Rather, it is likely to be due to the separation of PS10–12 regulatory sequences within domain II (and most of domain III) from the Abd-B transcription unit.

All of these arguments reinforce the conclusion drawn from genetic data that the iab region serves to regulate abd-A and Abd-B. In this context the iab transcripts might function in cis by some unprecedented mechanism. Alternatively, it may be the state of the iab DNA, or the act of transcription, that mediates iab function. The transcripts themselves might then be functionless by-products of this mechanism (see Lipshitz et al. 1987 for a discussion of this point with respect to bxd transcripts).

Transcription at blastoderm reveals early stages in the activation of the BX-C

Although early iab transcripts may have no function, they are significant because they reveal that, as early as mid-cycle 14, DNA of the BX-C has sensed and responded to a series of positional cues that subdivide the presumptive abdomen. Four different regions are defined by the transcription of 0, 1, 2 or 3 of the iab domains, in progressively more posterior cells of the blastoderm. This ordered activation of iab DNA occurs shortly before the abd-A and Abd-B genes are first transcribed (Rowe, 1988; Harding and Levine, 1988; Sánchez-Herrero and Crosby, 1988; DeLorenzi et al. 1988; Kuziora and McGinnis, 1988 and E.S., unpublished), and long before the differential expression of the abd-A and Abd-B genes establishes differences between the abdominal segments 3–7.

There is colinearity between the location of DNA domains in the chromosome and the anterior limit to which each is transcribed in the blastoderm: more proximal domains are transcribed in more anterior cells. This parallels the relationship between the position of each iab mutation on the DNA map and the segments that it affects (Lewis, 1978): mutations that map proximally affect anterior abdominal segments, while those mapping distally transform more posterior ones. We do not find a one-to-one correspondence between expression domains and phenotypic classes of mutations, but within our limits of resolution, each class of iab mutations (Duncan, 1987) falls entirely within a single transcriptional domain (Fig. 6): iab-3 and iab-4 mutations are included within domain I, iab-5 and iab-6 mutations are within domain II, and iab-7 mutations within domain III.

Fig. 6.

The relationship between transcriptional activity in blastoderm stage embryos and chromosomal position for each of the three iab domains. Above: The hatched bars at the top of the figure show the extent of the three blastoderm transcription domains relative to egg length (posterior pole = 0 %), and to the approximate position of parasegment primordia. Even-numbered parasegments are shaded on the scale bar. Below: The extent of chromosomal DNA showing each of these three patterns is depicted by corresponding shaded bars above the DNA map of the abdominal region of the BX-C. Shown below the map are the locations of abd-A and Abd-B transcription units (black bars), the spread of each class of iab mutation (square brackets) and the extent of the small deletions Mcp and Fab (paired round brackets).

Fig. 6.

The relationship between transcriptional activity in blastoderm stage embryos and chromosomal position for each of the three iab domains. Above: The hatched bars at the top of the figure show the extent of the three blastoderm transcription domains relative to egg length (posterior pole = 0 %), and to the approximate position of parasegment primordia. Even-numbered parasegments are shaded on the scale bar. Below: The extent of chromosomal DNA showing each of these three patterns is depicted by corresponding shaded bars above the DNA map of the abdominal region of the BX-C. Shown below the map are the locations of abd-A and Abd-B transcription units (black bars), the spread of each class of iab mutation (square brackets) and the extent of the small deletions Mcp and Fab (paired round brackets).

Peifer et al. (1987) proposed that each class of iab mutations identifies a parasegment-specific regulatory domain within the BX-C. They suggested that these were activated (‘opened’) in sequence along the chromosome in response to cues along the anteroposterior axis of the embryo. We believe that blastoderm transcription in the iab region may be a reflection of this activation event, as previously proposed for the bxd transcripts in the Ubx region (Lipshitz et al. 1987). If this is so, the initial domains of activation would appear to define units approximately two parasegments long, rather than individual parasegments, since the anterior limits of expression of the iab transcripts are separated by around 8–10% egg length.

The gap genes regulate the early activation of the BX-C (White and Lehman, 1986; Ingham et al. 1986; Harding and Levine, 1988; Irish et al. 1989) and are expressed at maximal levels just as these iab transcripts begin to accumulate (Knipple et al. 1985; Tautz et al. 1987). We would guess that they are directly involved in the initial activation of regulatory domains and therefore in the establishment of the iab patterns of expression. Mutations in gap genes alter the limits of expression of the iab transcripts (E.S., unpublished). After this initial activation by gap genes, ftz and perhaps other pair-rule genes may act to distinguish within these domains regions that are unique to single parasegments (Duncan, 1986; Ingham and Martinez-Arias, 1986).

The suggestion that the iab transcripts reflect the initial activation of the BX-C accounts for the colinearity of transcription boundaries with mutant effects. If the iab transcripts reflect a response to the gap genes but not to the pair-rule inputs there would be no reason to expect their limits of expression to be precisely parasegmental. However, these assumptions provide no obvious explanation for the apparent shift of about one parasegment between high levels of expression of iab transcripts from each domain, and the structures affected by mutations within the same domain. This discrepancy may disappear once the iab promoters are precisely located with respect to the DNA elements that sense position at blastoderm.

This model is similar to that previously proposed by Duncan (1986), in which the initial activation of the BX-C is made in double-parasegmental units (protosegments), subsequently divided by ftz. Both of these models imply two different processes required for the correct deployment of BX-C genes. Initially, gap and pair-rule genes define regions in the BX-C regulatory DNA that correspond to individual parasegments (the iab transcripts perhaps being a reflection of the initial activation). Thereafter, these regulatory regions with parasegmental ‘addresses’ regulate the expression of abd-A and Abd-B products in response to the evolving pattern of positional cues within each segment, so that segmental identity is achieved.

The location of two mutations, Miscadastral pigmentation (Mcp) and Frontabdominal (Fab), is of particular interest in relation to the delimitation of parasegmental domains in the iab DNA. These two mutations are short deletions (Karch et al. 1985; F. Karch, personal communication) yet they result in dominant phenotypes (Lewis, 1978; F. Karch, personal communication), Mcp activating the iab-5 regulatory domain in A4 and Fab activating the iab-7 regulatory domain in A6. It has been suggested (F. Karch personal communication) that they may act by removing the boundaries between two parasegmental regulatory domains. We note that they lie at or very close to the boundaries between transcription domains I and II (for Mcp) and II and III (for Fab) (Fig. 6).

We thank W. Bender and S. Sakonju for providing phages, plasmids and DNA, M. Crosby and F. Karch for communicating unpublished results, R. Weinzierl, A. Rowe and I. Dawson for help, M. Levine for the Abd-B cDNA, A. Martinez-Arias for the ftz probe, E. Reoyo for technical assistance and M. Crosby, G. Morata, S. Sakonju and J. Casanova for comments on the manuscript. E.S. is on leave from the CSIC, Spain. This work has been supported by EMBO and by the Medical Research Council.

Akam
,
M.
,
MartÍnez-arias
,
A.
,
Weinzierl
,
R.
&
Wilde
,
C. D.
(
1985
).
Function and expression of Ultrabithorax in the Drosophila embryo
.
Cold Spring Harbor Symp. quant. Biol
.
50
,
195
200
.
Beachy
,
P. A.
,
Helfand
,
S. L.
&
Hogness
,
D. S.
(
1985
).
Segmental distribution of bithorax complex proteins during Drosophila development
.
Nature, Land
.
313
,
545
551
.
Cabrera
,
C. V.
,
Botas
,
J.
&
Garcia-Bellido
,
A.
(
1985
).
Distribution of Ultrabithorax proteins in mutants of Drosophila bithorax complex and its trans-regulatory genes
.
Nature, Lond
.
318
,
569
571
.
Casanova
,
J.
,
Sánchez-herrero
,
E.
,
Busturia
,
A.
&
Morata
,
G.
(
1987
).
Double and triple mutant combinations in the bithorax complex of Drosophila
.
EMBO J
.
6
,
3103
3109
.
Casanova
,
J.
,
Sánchez-herrero
,
E.
&
Morata
,
G.
(
1985
).
Prothoracic transformation and functional structure of the Ultrabithorax gene of Drosophila
.
Cell
42
,
663
669
.
Casanova
,
J.
,
Sánchez-herrero
,
E.
&
Morata
,
G.
(
1986
).
Identification and characterization of a parasegment specific regulatory element of the Abdominal-B gene of Drosophila
.
Cell
47
,
627
636
.
Delorenzi
,
M.
,
Ali
,
N.
,
Saari
,
G.
,
Henry
,
C.
,
Wilcox
,
M.
&
Bienz
,
M.
(
1988
).
Evidence that the Abdominal-B r element function is conferred by a trans-regulatory homeoprotein
.
EMBO J
.
7
,
3223
3231
.
Duncan
,
I.
(
1986
).
Control of bithorax complex functions by the segmentation gene fushi tarazu of Drosophila melanogaster
.
Cell
47
,
297
309
.
Duncan
,
I.
(
1987
).
The bithorax complex
.
Ann Rev. Genet
.
21
,
285
-
319
(1987).
Feinberg
,
A. P.
&
Vogelstein
,
B.
(
1983
).
A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity
.
Anal. Bioch
.
132
,
6
13
.
Gyurkovics
,
H.
,
Gausz
,
J.
,
Kummer
,
J.
&
Karch
,
F.
(Submitted).
Harding
,
K.
&
Levine
,
M.
(
1988
).
Gap genes define the limits of Antennapedia and bithorax gene expression during early development in Drosophila
.
EMBO J
.
7
,
205
214
.
Harding
,
K.
,
Wedeen
,
C.
,
Mcginnis
,
W.
&
Levine
,
M.
(
1985
).
Spatially regulated expression of homeotic genes in Drosophila
.
Science
229
,
1236
1246
.
Hayes
,
P. H.
,
Sato
,
T.
&
Denell
,
R. E.
(
1984
).
Homoeosis in Drosophila: the Ultrabithorax larval syndrome
.
Proc. natn. Acad. Sci. U.S.A
.
81
,
545
549
.
Hoey
,
T.
,
Doyle
,
H. J.
,
Harding
,
K.
,
Wedeen
,
C.
&
Levine
,
M.
(
1986
).
Horneo box gene expression in anterior and posterior regions of the Drosophila embryo
.
Proc. natn. Acad. Sci. U.S.A
.
83
,
4809
4813
.
Hogness
,
D. S.
,
Lipshitz
,
H. D.
,
Beachy
,
P. A.
,
Peattie
,
D. A.
,
Saint
,
R. B.
,
Goldschmidt-Clermont
,
M.
,
Harte
,
P. J.
,
Gavis
,
E. R.
&
Helfand
,
S.
(
1985
).
Regulation and products of the Ubx domain of the bithorax complex
.
Cold Spring Harbor Symp. quant. Biol
.
50
,
181
194
.
Ingham
,
P. W.
,
Ish-Horowicz
,
D.
&
Howard
,
K. R.
(
1986
).
Correlative changes in homoeotic and segmentation gene expression in Kruppel mutant embryos of Drosophila
.
EMBO J
.
5
,
1659
1665
.
Ingham
,
P.
&
Martinez-Arias
,
A.
(
1986
).
The correct activation of Antennapedia and bithorax complex genes requires the fushi tarazu gene
.
Nature, Lond
.
324
,
592
597
.
Irish
,
V.
,
Martinez-Arias
,
A.
&
Akam
,
M. E.
(
1989
).
Spatial regulation of the Antennapedia and Ultrabithorax homeotic genes during Drosophila early development
.
EMBO J
.
8
,
1527
1537
.
Karch
,
F.
,
Weiffenbach
,
B.
,
Peifer
,
M.
,
Bender
,
W.
,
Duncan
,
L
,
Celniker
,
S.
,
Crosby
,
M.
&
Lewis
,
E. B.
(
1985
).
The abdominal region of the bithorax complex
.
Cell
43
,
81
96
.
Knipple
,
D. C.
,
Seifert
,
E.
,
Rosenberg
,
V. B.
,
Preiss
,
A.
&
Jâckle
,
H.
(
1985
).
Spatial and temporal patterns of Kruppel gene expression in early Drosophila embryos
.
Nature, Lond
.
317
,
40
44
.
Kuhn
,
D. T.
,
Woods
,
D. F.
&
Cook
,
J. L.
(
1981
).
Analysis of a new homeotic mutation (iab-2) within the bithorax complex in Drosophila melanogaster
.
Mol. gen. Genet
.
181
,
82
86
.
Kuziora
,
M. A.
&
Mcginnis
,
W.
(
1988
).
Different transcripts of the Drosophila Abd-B gene correlate with distinct genetic subfunctions
.
EMBO J
.
7
,
3233
3244
.
Lewis
,
E. B.
(
1978
).
A gene complex controlling segmentation in Drosophila
.
Nature, Lond
.
276
,
565
570
.
Lewis
,
E. B.
(
1981
).
Developmental genetics of the bithorax complex in Drosophila
. In
Developmental Biology using Purified Genes. ICN-UCLA Symp. Molec. cell. Biol
. (ed.
D.D.
Brown
,
C. F.
Fox
), pp.
189
208
.
New York
:
Academic Press
.
Lipshitz
,
H. D.
,
Peattie
,
D. A.
&
Hogness
,
D. S.
(
1987
).
Novel transcripts from the Ultrabithorax domain of the bithorax complex
.
Genes Dev
.
1
,
307
322
.
Martinez-Arias
,
A.
&
Lawrence
,
P. A.
(
1985
).
Parasegments and compartments in the Drosophila embryo
.
Nature, Lond
.
313
,
639
642
.
Peifer
,
M.
,
Karch
,
F.
&
Bender
,
W.
(
1987
).
The bithorax complex: control of segmental identity
.
Genes Dev
.
1
,
891
898
.
Rowe
,
A.
(
1987
).
Studies on the expression and regulation of the bithorax complex of Drosophila. D.Phil. Thesis, University of Cambridge
.
Sánchez-herrero
,
E.
,
Casanova
,
J.
,
Kerridge
,
S.
&
Morata
,
G.
(
1985
).
Anatomy and function of the bithorax complex of Drosophila. Cold Spring Harbor
Symp. quant. Biol
.
50
,
165
172
.
Sánchez-herrero
,
E.
,
Casanova
,
J.
&
Morata
,
G.
(
1988
).
Genetic structure of the bithorax complex
.
Bioessays
8
,
124
128
.
Sánchez-herrero
,
E.
&
Crosby
,
M.
(
1988
).
The Abdominal-B gene of Drosophila melanogaster: overlapping transcripts exhibit two different spatial distributions
.
EMBO J
.
7
,
2163
2173
.
Sánchez-herrero
,
E.
,
Vernós
,
I.
,
Marco
,
R.
&
Morata
,
G.
(
1985
).
Genetic organization of Drosophila bithorax complex
.
Nature, Lond
.
313
,
108
113
.
Tautz
,
D.
,
Lehmann
,
R.
,
Schnürch
,
H.
,
Schuh
,
R.
,
Seifert
,
E.
,
Kienlin
,
A.
,
Jones
,
K.
&
Jâckle
,
H.
(
1987
).
Finger protein of novel structure encoded by hunchback, a second member of the gap class of Drosophila segmentation genes
.
Nature, Lond
.
327
,
383
389
.
Tiong
,
S.
,
Bone
,
L. M.
&
Whittle
,
J. R. S.
(
1985
).
Recessive lethal mutations within the bithorax-complex in Drosophila
.
Mol. gen. Genet
.
200
,
335
342
.
Tiong
,
S. Y. K.
,
Whittle
,
J. R. S.
&
Gribbin
,
M. C.
(
1987
).
Chromosomal continuity in the abdominal region of the bithorax complex of Drosophila is not essential for its contribution to metameric identity
.
Development
101
,
135
142
.
Weinzierl
,
R.
,
Axton
,
J. M.
,
Ghysen
,
A.
&
Akam
,
M.
(
1987
).
Ultrabithorax mutations in constant and variable regions of the protein coding sequence
.
Genes. Dev
.
1
,
386
397
.
White
,
R. A. H.
&
Lehmann
,
R.
(
1986
).
A gap gene, hunchback, regulates the spatial expression of Ultrabithorax
.
Cell
47
,
311
321
.
White
,
R. A. H.
&
Wilcox
,
M.
(
1985
).
Regulation of the distribution of Ultrabithorax proteins in Drosophila
.
Nature, Lond
.
318
,
563
567
.