13. The homeobox is a common coding motif in homeotic genes

29.12 Complex loci are extremely large and involved in regulation

Key terms defined in this section

Homeotic genes are defined by mutations that convert one body part into another; for example, an insect leg may replace an antenna.

Segment polarity genes control the pattern within each segment, including the polarity. Homeotic genes impose the program that determines the unique differentiation of each segment. Homeotic mutations cause cells of one compartment to develop the phenotype of a different cell compartment. Some homeotic mutants "transform" part of a segment or an entire segment into another type of segment. Most homeotic genes are expressed in a spatially restricted manner that corresponds to parasegments.

Homeotic genes interact in complicated interlocking patterns. Many homeotic genes code for transcription factors that act upon other homeotic genes as well as upon other target loci. As a result, a mutation in one homeotic gene influences the expression of other homeotic genes. The consequence is that the final appearance of a mutant depends not only on the loss of one homeotic gene function, but also on how other homeotic genes change their spatial patterns in response to the loss.

Homeotic genes act during embryogenesis. Their expression depends on the prior expression of the segmentation genes; we might regard the homeotic genes as integrating the pattern of signals established by the segmentation genes. Homeotic mutations cause one segment of the abdomen to develop as another, legs to develop in place of antennae, or wings to develop in place of eyes. Note that homeotic genes do not create patterns de novo; they modify cell fates that are determined by genes such as the segment polarity genes, by switching the set of genes that functions in a particular place. Indeed, the segment polarity genes are active at about the same time as the peak of expression of the homeotic genes.

The genetic properties of some homeotic mutations are unusual and led to the identification of complex loci. A conventional gene Xeven an interrupted one Xis identified at the level of the genetic map by a cluster of noncomplementing mutations. In the case of a large gene, the mutations might map into individual clusters corresponding to the exons. A hallmark of complex loci is that, in addition to rather well-spaced groups of mutations, extending over a relatively large map distance, there are complex patterns of complementation, in which some pairwise combinations complement but others do not. The individual mutations may have different and complex morphological effects on the phenotype. These relationships are caused by the existence of an array of regulatory elements extending far beyond the transcription units they control, which may be extremely large. Many of the bizarre results that are obtained in complementation assays turn out to result from mutations in promoters or enhancers that affect expression in one cell type but not another. We now recognize that complex loci do not have any novel features of genetic organization, apart from the fact that they have many regulatory elements that control expression in different parts of the embryo (for review see Scott, 1987).

Two of the complex loci are involved in regulating development of the adult insect body. The ANT-C and BX-C complex loci together provide a continuum of functions that specify the identities of all of the segmented units of the fly. Each of these complexes contains several homeotic genes. The two separate complexes may have evolved from a split in a single ancestral complex, as suggested by the evolution of the corresponding genes in other species. In the beetle Tribolium, the ANT-C and BX-C complexes are found together at a single chromosomal location. The individual genes may have been derived from duplications and mutations of an original ancestral gene. And in mammals, there are arrays of related genes whose individual members are related sequentially to the genes of the ANT-C and BX-C complexes (as we describe in the next section).

The homeotic genes clustered at the ANT-C and BX-C complexes show a relationship between genetic order and the position in which they are expressed in the body of the fly. Proceeding from left to right, each homeotic gene in the complex acts upon a more posterior region of the fly. The basic principle is that formation of a compartment requires the gene product(s) expressed in the previous compartment, plus a new function coded by the next gene along the cluster. So loss-of-function mutations usually cause one compartment to have the phenotype of the corresponding compartment on its anterior side. The individual genes code for a set of transcription factors that have related DNA-binding domains (see next section).

The identities of the most anterior parts of the fly (parasegments 1 V4) are specified by ANT-C, which contains several homeotic genes, including labial (lab), proboscipedia (pb), Deformed (Dfd), Sex combs reduced (Scr), and Antennapedia (Antp). The homeotic genes lie in a cluster over a region of ~350 kb, but several other genes are interspersed; most of these genes are regulators that function at different stages of development.

Figure 29.32 The homeotic genes of the ANT-C complex confer identity on the most anterior segments of the fly. The genes vary in size, and are interspersed with other genes. The antp gene is very large and has alternative forms of expression.

Figure 29.32 correlates the organization of ANT-C with its effects upon body parts. Adjacent genes are expressed in successively more posterior parts of the embryo, ranging from the leftmost gene labial (the most anterior acting, which affects the head) to the rightmost gene Antp (the most posterior acting, which affects segments T2 VT3).

The Antp gene gave its name to the complex, and among the mutations in it are alleles that change antennae into second legs, or second and third legs into first legs. Antp usually functions in the thorax; it is needed both to promote formation of segments T2 VT3 and to suppress formation of head structures. Loss of function therefore causes T2 VT3 to resemble the more anterior structure of T1; gain of function, for example, by over-expression in the head, causes the anterior region to develop structures of the thorax. (The molecular action of Antp is to prevent the action of genes hth and exd that promote formation of antennal structures. Hth causes exd to be imported into the nucleus, where it switches on the genes that make the antenna.)

The figure summarizes the organization of the gene. It has 8 exons, separated by very large introns, and altogether spanning ~103 kb. The single open reading frame begins only in exon 5, and apparently gives rise to a protein of 43 kD. The discrepancy between the length of the locus and the size of the protein means that only 1% of its DNA codes for protein.

Transcription starts at either of two promoters, located ~70 kb apart! One promoter is located upstream of exon 1, the other upstream of exon 3. Use of the first promoter is associated with omission of exon 3. The transcripts generated from either promoter end either within or after exon 8. All the transcripts appear to code for the same protein. No difference has yet been identified in the use of the two promoters, which suggests that their significance could lie in the different structures of the nontranslated leaders of the mRNAs (Scott et al., 1983).

The other genes of the ANT-C complex are expressed in the head and first thoracic segment. In the most anterior compartments, lab, pb, Dfd, have unique patterns of expression, so that deletions of segmental regions can result from loss-of-function mutations. An exception to the left-right/anterior-posterior order of action is that loss of Scr allows the overlapping Antp to function, that is, the direction of transformation is opposite from usual.

Figure 29.33 A four-winged fly is produced by a triple mutation in abx, bx, and pbx at the BX-C complex. Photograph kindly provided by Ed Lewis.

The classic complex homeotic locus is BX-C, the bithorax complex, characterized by several groups of homeotic mutations that affect development of the thorax, causing major morphological changes in the abdomen. When the whole complex is deleted, the insect dies late in embryonic development. Within the complex, however, are mutations that are viable, but which change the phenotype of certain segments. An extreme case of homeotic transformation is shown in Figure 29.33, in which a triple mutation converts T3A (which carries the halteres [truncated wings]) into the tissue type of the T2 (which carries the wings). This creates a fly with four wings instead of the usual two (Wickelgren, 2000).

Figure 29.34 The bithorax (BX-C) locus has 3 coding units. A series of regulatory mutations affects successive segments of the fly. The sites of the regulatory mutations show the regions within which deletions, insertions, and translocations confer a given phenotype.

The genetic map of BX-C is correlated with the body structures that it controls in the fly in Figure 29.34. The body structures extend from T2 to A8. The BX-C complex is therefore concerned with the development of the major part of the body of the fly. Like ANT-C, a crucial feature of this complex is also that mutations affecting particular segments lie in the same order on the genetic map as the corresponding segments in the body of the fly. Proceeding from left to right along the genetic map, mutations affect segments in the fly that become successively more posterior (Martin et al., 1995).

A difference between ANT-C and BX-C is that ANT-C functions largely or exclusively via its protein-coding loci, but BX-C displays a complex pattern of cis-acting interactions in addition to the effects of mutations in protein-coding regions. The BX-C occupies 315 kb, of which only 1.4% codes for protein. The individual mutations fall into two classes:

• Three transcription units (Ubx, abdA, AbdB) produce mRNAs that code for proteins. The transcription units are large (>75 kb for Ubx, and >20 kb each for abdA and AbdB). Each contains several large introns. (The bxd and iab4 regions produce RNAs that do not code for proteins; again, the transcription units are large, and the RNA products are spliced. Their functions are unknown.)
  
• There are cis-acting mutations at intervals throughout the entire cluster. They control expression of the transcription units. Cis-acting mutations of any particular type may occur in a large region. The locations shown on the map are only approximate, and the boundaries within which mutations of each type may occur are not well defined.

As a historical note, the complex was originally defined in terms of two "domains." Mutations in the Ultrabithorax domain were characterized first; they have the thoracic segments T2P VT3P and the abdominal compartment A1A as their targets (this corresponds to parasegments 5 V6). These mutations lie either in the Ubx transcription unit or in the cis-acting sites that control it. The mutations within the ultrabithorax domain are named for their phenotypes. The bx and bxd types are identified by a series of mutations, in each case dispersed over ~10 kb. The abx and pbx mutations are caused by deletions, which vary from 1 V10 kb.

Mutations in the Infraabdominal domain were found later; they have the abdominal segments A1P-A8P (parasegments 7-14) as targets.) These mutations lie either in the AbdA,B transcription units, or in the cis-acting sites that control them. Within the infraabdominal domain, cis-acting mutations are named systematically as iab2-9. These mutations affect individual compartments, or sometimes adjacent sets of compartments, as shown at the top of the figure (Karch et al., 1985).

Proceeding from left to right along the cluster, transcripts are found in increasingly posterior parts of the embryo, as shown at the bottom of the figure. The patterns overlap. Ubx has an anterior boundary of expression in compartment T2P (parasegment 5), abdA is expressed from compartment A1P (parasegment 7), and AbdB is expressed in compartments posterior to compartment A4P (parasegment 10).

Transcription of Ubx has been studied in the most detail. The Ubx transcription unit is ~75 kb, and has alternative splicing patterns that give rise to several short RNAs. A transient 4.7 kb RNA appears first, and then is replaced by RNAs of 3.2 and 4.3 kb. A feature common to both the latter two RNAs is their inclusion of sequences from both ends of the primary transcript. Of course, there may be other RNAs that have not yet been identified.

We do not yet have a good idea of whether there are significant differences in the coding functions of these RNAs. The first and last exons are quite lengthy, but the interior exons are rather small. Small exons from within the long transcription unit may enter mRNA products by means of alternative splicing patterns. So far, however, we do not know of any functional differences in the Ubx proteins produced by the various modes of expression.

Ubx proteins are found in the compartments that correspond to the sites of transcription, that is in T2P VA1 and at lower levels in A2 VA8. So the Ubx unit codes for a set of related proteins that are concentrated in the compartments affected by mutations in the Ultrabithorax domain. Ubx proteins are located in the nucleus, and they fall into the general type of transcriptional regulators whose DNA-binding region consists of a homeodomain.

We can understand the general function of the BX-C complex by considering the effects of loss-of-function mutations. If the entire complex is deleted, the larva cannot develop the individual types of segments. In terms of parasegments (which are probably the affected units), all the parasegments differentiate in the same way as parasegment 4; the embryo has 10 repetitions of the repeating structure T1P/T2A all along its length, in place of the usual compartments between parasegments 5 and 14. In effect, the absence of BX-C functions allows Antp to be expressed throughout the abdomen, so that all the segments take on the characteristic of a segment determined by Antp; BX-C functions are needed to add more posterior-type information.

Each of the transcription units affects successive segments, according to its pattern of expression. So if Ubx alone is present, the larva has parasegment 4 (T1P/T2A), parasegment 5 (T2P/T3A), and then 8 copies of parasegment 6 (T3P/A1A). This suggests that the expression of Ubx is needed for the compartments anterior to A1A. Ubx is also expressed in the more posterior segments, but in the wild type, abdA and AbdB are also present. If they are removed, the expression of Ubx alone in all the posterior segments has the same effect that it usually has in parasegment 6 (T3P/A1A).

The addition of abdA to Ubx adds the wild-type pattern to parasegments 7, 8, and 9. In other words, Ubx plus abdA can specify up to compartments A3P/A4A, and in the absence of AbdB, this continues to be the default pattern for all the more posterior compartments. The addition of AbdB is needed to specify parasegments 10 V14.

The general model for the function of the ANT-C and BX-C complexes is to suppose that additional functions are added to define successive segments proceeding in the posterior direction. It functions by reliance on a combinatorial pattern in which the addition of successive gene products confers new specificities. This explains the rule that a loss-of-function mutation in one of the genes of the ANT-C/BX-C complexes generally allows the gene on the more anterior side of the mutated gene to determine phenotype, that is, loss-of-function results in homeotic transformation of posterior regions into more anterior phenotypes (Beachy et al., 1985).

Expression of Ubx in a more anterior segment than usual should have the opposite effect to a loss-of-function; the segment develops a more anterior phenotype. When this is tested by arranging for Ubx to be expressed in the head, the anterior segments are converted to the phenotype of parasegment 6. So lack of expression of Ubx causes a homeotic transformation in which posterior segments acquire more anterior phenotypes; and over-expression of Ubx causes a homeotic transformation in which anterior segments acquire more posterior phenotypes.

This type of relationship is true generally for the cluster as a whole, and explains the properties of cis-acting mutations as well as those in the transcription units. These regulatory mutations cause loss of the protein in part of its domain of expression or cause additional expression in new domains. So they may have either loss-of-function or gain-of-function phenotypes (or sometimes both). The most common is loss-of-function in an individual compartment. For example, bx specifically controls expression of Ubx in compartment T3A; a bx mutation loses expression of Ubx in that compartment, which is therefore transformed to the more anterior type of T2A. This example is typical of the general rule for individual cis-acting mutations in the complex; each converts a target compartment so that it develops as though it were located at the corresponding position in the previous segment. The order of the cis-acting sites of mutation on the chromosomes reflects the order of the compartments in which they function. So the expression of Ubx in parasegments 4, 5, 6 is controlled sequentially by abx (affects parasegment 5), bx (affects T3A), etc.

The presence of only 3 genes within the BX-C complex poses two major questions. First, how do the combinations of 3 proteins specify the identity of 10 parasegments? One possibility is that there are quantitative differences in the various regions, allowing for the same sort of varying responses in target genes that we described previously for the combinatorial functioning of the segmentation genes. Second, how do the proteins function in different tissue types? The pattern of expression described above refers generally to the epidermis; the development of other tissues is controlled in a way that is parallel, but not identical. For example, although Ubx is expressed in all posterior segments up to A8 in the epidermis, in mesoderm, it is repressed posterior of segment A7. The posterior boundary reflects repression by abdA, since in abdA mutants, Ubx expression extends posterior in the mesoderm.

Why are loci involved in regulating development of the adult insect from the embryonic larva different from genes coding for the everyday proteins of the organism? Is their enormous length necessary to generate the alternative products? Could it be connected with some timing mechanism, determined by how long it takes to transcribe the unit? At a typical rate of transcription, it would take ~100 minutes to transcribe Antp, which is a significant proportion of the 22 hour duration of D. melanogaster embryogenesis.

Proceeding from anterior to posterior along the embryo, we encounter the changing patterns of expression of the genes of the ANT-C and BC-C loci. What controls their transcription? As in the case of the segmentation loci, the homeotic loci are controlled partially by the genes that were expressed at the previous stage of development, and partially by interactions among themselves. For example, the expression of Ubx is changed by mutations in bicoid, hunchback, or Kruppel. The anterior boundary of expression respects the parasegment border defined by ftz and eve. The general principle is that all of these regulatory genes function by controlling transcription, either by activating it or by repressing it, and that the gene products may exert specific effects by both qualitative and quantitative combinations.

Reviews

Scott, M. P. (1987). Complex loci of Drosophila. Ann. Rev. Biochem 56, 195-227.

Research

Beachy, P. A., Helfand, S. L., and Hogness, D. S. (1985). Segmental distribution of bithorax complex proteins during Drosophila development. Nature 313, 545-551.

Karch, F., Weiffenbach, B., Peifer, M., Bender, W., Duncan, I., Celniker, S., Crosby, M., and Lewis, E. B. (1985). The abdominal region of the bithorax complex. Cell 43, 81-96.

Martin, C. H. et al. (1995). Complete sequence of the bithorax complex of Drosophila. Proc. Nat. Acad. Sci. USA 92, 8398-8402.

Scott, M. P. et al. (1983). The molecular organization of the Antennapedia locus of Drosophila. Cell 35, 763-766.

Wickelgren, I. (2000). Hypertension. Mutation points to salt recycling pathway . Science 289, 23-26.