Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Curr Top Dev Biol. Author manuscript; available in PMC 2010 Jan 25.
Published in final edited form as:
PMCID: PMC2810641

Hox Specificity: Unique Roles for Cofactors and Collaborators


Hox proteins are well known for executing highly specific functions in vivo, but our understanding of the molecular mechanisms underlying gene regulation by these fascinating proteins has lagged behind. The premise of this review is that an understanding of gene regulation — by any transcription factor—requires the dissection of the cis-regulatory elements that they act upon. With this goal in mind, we review the concepts and ideas regarding gene regulation by Hox proteins and apply them to a curated list of directly regulated Hox cis-regulatory elements that have been validated in the literature. Our analysis of the Hox-binding sites within these elements suggests several emerging generalizations. We distinguish between Hox cofactors, proteins that bind DNA cooperatively with Hox proteins and thereby help with DNA-binding site selection, and Hox collaborators, proteins that bind in parallel to Hox-targeted cis-regulatory elements and dictate the sign and strength of gene regulation. Finally, we summarize insights that come from examining five X-ray crystal structures of Hox-cofactor-DNA complexes. Together, these analyses reveal an enormous amount of flexibility into how Hox proteins function to regulate gene expression, perhaps providing an explanation for why these factors have been central players in the evolution of morphological diversity in the animal kingdom.

1. An Introduction to the Problem

Hox proteins are homeodomain-containing transcription factors that have the capacity to carry out exquisitely precise functions in vivo that are critical for many aspects of animal morphogenesis. Most typically, each Hox gene is expressed in a subset of the anterior-posterior (AP) body axis, where it specifies cellular and tissue identities. Famous examples of the power that Hox genes have to sculpt animal morphogenesis include the antenna-to-leg transformation caused by the Antennapedia (Antp) mutation in Drosophila and several polydactyly syndromes in humans (Goodman, 2002; Lewis, 1978; Randazzo et al., 1991). These types of phenotypes have been a long-standing source of fascination for both biologists and lovers of science fiction.

An important and long-debated question for Hox biologists has been how these proteins achieve this apparently high degree of in vivo specificity. In this review, we summarize ideas and recent data bearing on the question of Hox specificity, with a special emphasis on what can be learned by analyzing native cis-regulatory elements that are directly bound and regulated by Hox proteins. Excellent reviews discussing the range of Hox target genes that have been identified using genome-wide and traditional approaches complement the emphasis of this chapter (Hueber and Lohmann, 2008; Pearson et al., 2005).

When Hox genes were first cloned and shown to encode homeodomain-containing proteins (Akam, 1989; Regulski et al., 1985), researchers initially speculated that Hox proteins would bind and regulate the correct subset of target genes according to the DNA recognition properties of their homeodomains. However, early work from a number of labs quickly established that homeodomains were not likely to be up to the task of dictating Hox-DNA-binding specificities on their own (Affolter et al., 1990; Desplan et al., 1988; Ekker et al., 1991, 1994; Hoey and Levine, 1988). Indeed, homeodomains, particularly the subset present in the Hox protein family, all bind to a very similar set of “AT”-rich DNA-binding sites, raising the fundamental question of how specificity is achieved. In addition to this classical problem of degenerate binding site recognition, experiments using chimeric Hox proteins—where bits of one Hox protein were replaced with homologous bits of another—highlight an additional complication. As expected, specific Hox functions required the homeodo-main. In some cases, however, specificity also required nonhomeodomain residues, in particular, those that lie immediately N- or C-terminal to the homeodomain (Chan et al., 1994; Dessain et al., 1992; Furukubo-Tokunaga et al., 1993; Gibson et al., 1990; Kuziora and McGinnis, 1989, 1990; Lin and McGinnis, 1992; Mann and Hogness, 1990; Zhao and Potter, 2001, 2002). How these nonhomeodomain residues may impact Hox specificity is just now coming into focus.

Because they are DNA-binding transcription factors, it is probably a safe bet that Hox proteins carry out the majority of their functions by binding to cis-regulatory elements (although alternative mechanisms have been proposed; Plaza et al., 2008). Because eukaryotic transcription is governed by cis-regulatory elements that typically integrate multiple inputs, each Hox-targeted element is likely to have binding sites for many transcription factors. Therefore, to understand how Hox proteins ultimately function to control target gene expression, it is necessary to consider two broad questions. First, how do Hox proteins recognize their DNA-binding sites and second, how do they interact with other transcriptional inputs that feed into the same cis-regulatory element? We suggest that it is helpful to break the problem of Hox specificity down into two conceptually separable steps (Fig. 3.1). In the first step, the question can be rephrased to ask: How do Hox proteins find the right DNA-binding sites in vivo? Many examples exist in the literature suggesting that Hox proteins solve this initial “DNA-binding specificity” step in multiple ways. As will be explored more fully below, one solution is by the use of cooperatively binding cofactors such as Extradenticle (Exd), Pbx, Homothorax (Hth), and Meis that increase Hox-DNA-binding specificities (previously reviewed by Mann and Affolter, 1998; Mann and Chan, 1996; Moens and Selleri, 2006). However, it is also increasingly clear that Hox proteins regulate many genes without the help of these cofactors. In the second step, the question is: Once bound, how do Hox proteins orchestrate a transcriptional response? As the same Hox protein can activate some target genes, and repress others, it is clear that this “activity regulation” step is also critical for how Hox proteins execute their in vivo functions. In fact, as will be described more fully below, there is now good evidence for both of the steps outlined in Fig. 3.1 playing critical roles in Hox specificity.

Figure 3.1
Two contributing steps in Hox specificity. In principle, Hox specificity can be broken down into two separate steps. The first step is DNA binding by Hox proteins, which can occur either with or without cooperatively binding cofactors. The second step ...

2. Too Many Binding Sites, Not Enough Specificity

Because all Hox proteins have a homeodomain, understanding how Hox proteins recognize their DNA-binding sites in vivo certainly depends, at least in part, on how this 60 amino acid domain recognizes DNA sequences. The basic DNA recognition principles for homeodomains were established from biochemical and structural studies (reviewed previously by Gehring et al., 1994). These studies show that all homeodomains fold into a bundle of three alpha-helices and an unstructured “N-terminal” arm. DNA contacts are formed primarily by residues 47, 50, 51, and 54 in the third alpha-helix (the so-called recognition helix) and by an arginine in position 5 of the N-terminal arm. While these studies provided a high resolution view of how homeodomains generally bind to DNA, they did not provide much insight into the problem of Hox specificity for three reasons. First, nearly all Hox homeodomains, even those with very disparate in vivo functions, have the same residues in the DNA-contacting residues visualized in these structures (Mann, 1995). Second, although nonhomeo-domain residues were known to play a role in Hox specificity from studies of chimeric Hox proteins (see above), these domains were not present in any of the initial structural studies. Third, the DNA sequences used in these early structural studies were not in vivo binding sites. Instead, these structures used high-affinity consensus sites that would not be expected to reveal insights into homeodomain specificity.

Another limitation in the early studies on homeodomain-DNA recognition was that only a small subset of homeodomain proteins were studied. Thanks to the advent of new, powerful methodologies, two large-scale studies have recently defined the individual DNA-binding specificities for nearly all homeodomains, including the subset present in the Hox proteins. One group used a bacterial one-hybrid approach (B1H) to analyze the DNA-binding preferences for all of the homeodomains encoded in the Drosophila melanogaster genome (Noyes et al., 2008) while the second group used an entirely in vitro platform, protein-binding microarrays (PBMs), to answer the same question for all mouse homeodomains (Berger et al., 2008). Although there are pros and cons for each approach (Affolter et al., 2008), these studies confirmed that Hox homeodomains like to bind “AT”-rich DNA sequences (Fig. 3.2). In particular, the so-called “Antennapedia (Antp) Group” of homeodomains, which includes all Hox homeodomains except for those of the Abdominal-B (Abd-B) class, like to bind the sequence TAAT[t/g][a/g]. There are 87,307 copies of the sequence TAATTA and 86,201 copies of the sequence TAATGA in the D. melanogaster genome, each more than five times the number of annotated protein-coding genes. Clearly, the presence of TAAT[t/g][a/g] cannot be sufficient information for Hox regulation. Moreover, as elegantly illustrated by the homeodomain-binding site survey studies (Berger et al., 2008; Noyes et al., 2008), TAAT[t/g][a/g] is readily bound by most Hox homeodomains, as well as many non-Hox homeodomains. Therefore, this and related binding sites cannot be sufficient to distinguish between Hox family members that carry out distinct functions in vivo.

Figure 3.2
Comparison of in vitro and in vivo Hox-binding site preferences. Shown are LOGO diagrams summarizing Hox-binding site preferences for most paralogs. The column on the left lists the LOGOs generated using the binding sites identified by the bacterial 1-hybrid ...

3. How Specific Do Hox Proteins Need to be?

Hox biologists can readily point to highly specific functions that are uniquely specified by individual Hox proteins. For example, in Drosophila, only the Hox gene Sex combs reduced (Scr) can orchestrate the development of a salivary gland, presumably by regulating a network of salivary gland-promoting genes (Bradley et al., 2001). The flip view that multiple Hox proteins probably share many targets is typically given less attention. We believe this discussion is highly relevant to how one thinks about Hox specificity, because it may be that only a subset of Hox targets for any particular Hox protein need be highly specific.

One example of a morphological process that is controlled by multiple Hox genes is appendage development in Drosophila. Leg development is limited to the thoracic segments due to repression of Distalless (Dll) by the abdominal Hox genes, Ultrabithorax (Ubx), abdominal-A (abd-A), and Abdominal-B (Abd-B) (Estrada and Sánchez-Herrero, 2001; Vachon et al., 1992). Moreover, Ubx and abd-A work, at least in part, through common binding sites in a Dll cis-regulatory element (Gebelein et al., 2002, 2004). Thus, Dll is a shared target for the abdominal Hox genes. Similarly, the antenna-specifying gene homothorax (hth) can be repressed by all of the trunk Hox proteins, suggesting that hth is also a shared target (Casares and Mann, 1998, 2001; Yao et al., 1999). A third example is the Drosophila head-promoting gene, optix, which is repressed by the trunk Hox genes and activated by the more anterior (head) Hox genes (Coiffier et al., 2008). For hth and optix, a limitation in this conclusion is that the Hox-binding sites in these genes (assuming the regulation is direct) have not yet been identified. Therefore, although these genes are clearly shared Hox targets by genetic criteria, it is possible that different Hox proteins use different binding sites within these genes to regulate their expression. Nevertheless, these and other examples (Greig and Akam, 1995; Hirth et al., 2001) support the idea that not all Hox functions need to be paralog-specific. If true, it follows that many bona fide Hox-binding sites may not need to discriminate between different Hox proteins.

How many Hox targets are shared and how many are paralog-specific? Although the field may be getting closer to a definitive answer to this question, by applying ChIP-chip and/or ChIP-seq methodologies to Hox proteins, the currently available data provide an interesting estimate. Using overexpression of Hoxc8 in mouse embryo fibroblasts, the expression levels of 34 genes were found to change by twofold or more (Lei et al., 2005). This relatively small number of potential Hoxc8 targets contrasts from the much larger numbers of regulated genes identified in whole embryo expression profiling experiments following the uniform (and ectopic) expression of individual Hox proteins during Drosophila embryogenesis (Hueber et al., 2007). An important strength of these experiments is that the global transcriptional response to multiple Hox proteins was analyzed in parallel, using the same experimental conditions. Remarkably, of the ~1500 genes (about 10% of all Drosophila genes) whose expression changed significantly in response to ectopic Deformed (Dfd), Scr, Antp, Ubx, Abd-A, or Abd-B expression, more than two-thirds (~69%) were regulated by only one of these six Hox proteins. About one-third (~30%) of all Hox-responsive target genes responded to multiple Hox proteins, while only ~1% responded to all six of these Hox proteins. There are, however, a few caveats to these experiments. For one, they measured responses to ectopic Hox expression, which may not always reflect accurate gene regulation in their native expression domains. Second, these experiments did not distinguish between tissue-specific responses and third, they could not unambiguously discriminate between direct and indirect effects. Nevertheless, the results from this study are remarkable because they suggest that many, and perhaps the majority of Hox target genes are paralog-specific. However, the results also support the view that a significant number of genes are targeted by multiple Hox proteins, again raising the possibility that not all Hox-binding sites need to discriminate between Hox proteins.

Developmental context is another issue that should be considered when thinking about Hox specificity. In Drosophila, for example, each embryonic segment is built using the same reiterated set of signaling pathways that provide them with a common coordinate system, known as positional information. Once this “ground plan” is established, the non-Hox regulatory inputs into a gene are largely the same from segment to segment. One way to think about the Hox genes is that they impose identity information on top of this developmental ground plan, thus providing each segment its unique characteristics. Because the other regulatory inputs are largely equivalent, a gene that is specifically expressed (or repressed) in a small subset of embryonic segments is likely to be a paralog-specific Hox target gene. For example, for salivary glands to form only in the first thoracic (T1) segment, the Hox gene Scr must activate the salivary gland program, including its directly activated target gene forkhead (fkh), in a paralog-specific manner; other Hox proteins cannot activate this target in other segments (Bradley et al., 2001). Similarly, while Dll is repressed by abdominal Hox proteins, the thoracic Hox proteins Scr and Antp must be permissive for Dll expression (Gebelein et al., 2002). Repression by the abdominal Hox proteins therefore must have specificity. However, the same is not true for tissues where the ground plan, and thus the other regulatory inputs, is different. For example, the Hox gene Ubx is expressed in all cells of the developing haltere (a balancing organ used during flight) of the fly, where it regulates the expression of many genes (Crickmore and Mann, 2007, 2008; Lewis, 1978; Weatherbee and Carroll, 1999). Other than the developing wing, where Hox genes are not expressed for most of development, no other tissue in the fly has the same ground plan as the haltere. Therefore, the cis-regulatory elements used by Ubx to regulate genes in the haltere may not need to be highly selective for Ubx: other Hox proteins never have the opportunity to regulate these genes in the haltere/wing tissue simply because they are not expressed there. Confirmation of this idea comes from the finding that other Hox proteins, when expressed in the wing, can result in haltere-like phenotypes and mimic Ubx-like regulation (Casares et al., 1996; R. S. Mann and M. Crickmore, unpublished observations). Similarly, although both Ubx and Abd-A have the potential to induce gonad development, the job is normally carried out by Abd-A simply because it is the only Hox protein that is expressed in the correct set of mesodermal progenitor cells (Greig and Akam, 1995). Finally, another remarkable example of functional redundancy among different Hox paralogs is that all of the Drosophila Hox proteins, except for Abd-B, have the ability to replace Labial in the specification of the tritocerebral neuromere in the fly's brain (Hirth et al., 2001). Like Ubx in the haltere, the ground plan for this portion of the brain may be sufficiently unique so that it does not require exquisite Hox specificity, accounting for why nearly all Hox paralogs can, at least to some degree, carry out the same functions as Labial.

In summary, when thinking about what Hox-bindings sites may look like based on these considerations, we suggest that it is useful to distinguish between three types of Hox target genes (Fig. 3.3) (1) those that must be highly specific for one Hox paralog (“paralog-specific”; e.g., Scr → fkh), (2) those that are shared by a subset of Hox proteins (“semi-paralog-specific”; e.g., Ubx, Abd-A, and Abd-B ⊣ Dll), and (3) those that are regulated by most or all Hox genes (“general”; e.g., all of the trunk Hox genes ⊣ optix). In addition, we argue that for some Hox functions that have the appearance a of paralog specificity (e.g., Ubx dictating haltere instead of wing fates), the Hox-binding sites, themselves, do not need to be paralog-specific as long as the developmental context is sufficient to specify a unique regulatory environment.

Figure 3.3
Three types of Hox target genes. “Paralog-specific” Hox target genes are those that are uniquely regulated by only a single Hox paralog, such as the activation of fkh by Scr. “Semi-paralog-specific” Hox target genes are ...

4. Hox Cofactors

Given that some Hox functions truly require a high degree of specificity, and that Hox homeodomains, themselves, are not sufficiently discriminating to account for this specificity, how is specificity achieved? One well-established way in which Hox proteins achieve specificity in vivo is to bind DNA cooperatively with other DNA-binding cofactors. To date, the best-characterized cofactors are all TALE (three amino acid loop extension) homeodomain proteins (Mann and Chan, 1996; Moens and Selleri, 2006). In Drosophila, the known TALE Hox cofactors are Extradenticle (Exd) and Homothorax (Hth). In the mouse, there are four Exd-related proteins (Pbx1, Pbx2, Pbx3, Pbx4) and five Hth-related proteins (Meis1, Meis2, Meis3, Prep1, and Prep2). In Caenorhabditis elegans, there are three genes encoding Exd-like proteins (ceh-20, ceh-40, ceh-60) and two that encode Hth-like proteins, unc-62 and psa-2, which encodes for a truncated form that has no homeodomain (see Mukherjee and Bürglin, 2007 for a thorough description of the TALE family genes). Here, we collectively refer to Exd/Pbx/Ceh-20 as PBC proteins. These proteins all have the ability, at least on some DNA sequences, to bind with Hox proteins in a highly cooperative manner. In addition, it is important to stress that where it has been analyzed, TALE family homeodomain proteins also carry out many Hox-independent functions in vivo (Bessa et al., 2008; Casares and Mann, 1998, 2001; Jiang et al., 2008; Laurent et al., 2007; Moens and Selleri, 2006). Because they have both Hox-dependent and Hox-independent functions the genetic analysis of TALE family genes needs to be interpreted with caution, since only a subset of the observed phenotypes is due to their role as Hox cofactors.

Protein interaction domains characterized in Hox, PBC, and Hth/Meis/Prep proteins have provided many insights into how these three factors assemble when bound to DNA. PBC proteins interact with Hth/Meis/Prep family members in a DNA-independent manner via highly conserved domains present in the N-terminal regions (PBC-A of PBC and HM of Hth/Meis) of these proteins (Mann and Affolter, 1998). In several cases, the nuclear localization and/or stability of these proteins has been shown to depend on this protein-protein interaction (Arata et al., 2006; Berthelsen et al., 1999; Haller et al., 2004; Huang et al., 2003; Mann and Abu-Shaar, 1996; Ryoo and Mann, 1999; Saleh et al., 2000; Stevens and Mann, 2007). In contrast, PBC-Hox interactions appear to be more complicated—and potentially more interesting. The traditional view, which has been supported by biochemical, in vivo, and structural studies, is that a motif common to most Hox proteins—YPWM—makes direct contacts with the TALE motif in PBC homeodomains, which creates a hydrophobic pocket that binds the W in YPWM (Chan and Mann, 1996; Chang et al., 1995; Joshi et al., 2007; LaRonde-LeBlanc and Wolberger, 2003; Lu and Kamps, 1996; Neuteboom et al., 1995; Passner et al., 1999; Phelan et al., 1995; Piper et al., 1999). For those Hox proteins that do not have an obvious YPWM motif, in particular the Abd-B paralogs, there is a conserved W residue that, at least for a subset of Abd-B paralogs, plays an important role in this protein-protein interaction (Shen et al., 1997). However, more recent studies make it clear that there is more to PBC-Hox-DNA complex formation than the YPWM-TALE interaction. On the one hand, two studies have shown that mutation of the YPWM motif in Ubx fail to abolish cooperative binding with PBC proteins in vitro and some Ubx functions in vivo (Galant et al., 2002; Merabet et al., 2003; Shen et al., 1997). On the other hand, a peptide immediately following the Ubx homeodomain, termed the UbdA motif because of its similarity between Ubx and Abd-A, is playing an important, and perhaps dominant, role in the interaction between these Hox proteins and Exd on some binding sites (Merabet et al., 2007). Interestingly, consistent with the idea that the UbdA motif is playing a role in PBC-Ubx (and likely Abd-A) interactions, it contributes to Ubx functional specificity in vivo (Chan and Mann, 1993; Gebelein et al., 2002). Although the UbdA motif is not found outside of arthropods, these findings suggest the more general idea that PBC proteins may have modes of interaction with other Hox proteins that are in addition to the classical YPWM-TALE interaction. The existence of multiple PBC interaction domains in a single Hox protein such as Ubx suggest that the way in which PBC-Hox complexes assemble onto cis-regulatory elements may vary from target to target, potentially expanding their ability to recruit additional transcription factors.

In addition to TALE family homeodomain proteins, the Drosophila homeodomain protein Engrailed (En) has also been shown to be a Hox cofactor (Gebelein et al., 2004). In this case, En bound cooperatively with both Ubx and Abd-A to a regulatory element from the Dll gene, and En input is required for Dll repression in the posterior compartments of the abdominal segments (Gebelein and Mann, 2007; Gebelein et al., 2004). Unlike the TALE cofactors, which can function with Hox proteins to both activate and repress target genes, it is likely that En-Hox dimers are more typically involved in gene repression due to En's ability to directly bind the corepressor Groucho (Jiménez et al., 1997).

A subset of Zn finger-containing transcription factors, most prominently Drosophila Teashirt (Tsh), has also been suggested to be Hox cofactors (Robertson et al., 2004; Taghli-Lamallem et al., 2007). Although a very appealing idea, these factors do not seem to exhibit the same robust cooperativity that is typically observed between TALE factors and Hox proteins. And, at least in the one target where Tsh-binding sites were identified (modulo), they are not adjacent to the Hox-binding sites (Taghli-Lamallem et al., 2007). Thus, at this time we prefer to classify these Zn finger factors as Hox “collaborators,” which provide additional, essential inputs into a subset of Hox-targeted cis-regulatory elements (discussed in more detail below) (Table 3.1).

Table 3.1
Direct Hox-DNA-binding sites

Given the high degree of specificity required for some Hox functions, and that there are dozens of Hox genes in vertebrates, it is perhaps surprising how few bona fide Hox cofactors have been identified. One answer to this paradox, discussed later in this chapter, comes from a recent atomic-level resolution view of how TALE cofactors bind to DNA with Hox proteins. Another possible answer, especially in vertebrates, is that multiple Pbx and Meis/Prep genes encode for proteins with different biochemical properties and thus expand the number of Hox cofactors. Perhaps analogously, although Drosophila has a single exd-like gene and a single hth-like gene, alternative splicing of Drosophila hth also adds to the repertoire of Hox cofactors present in the fly (Noro et al., 2006). Specifically, hth encodes both homeodomain-containing and homeodomain-less (HDless) isoforms. Not only do both of these isoforms contribute to hth functions, there is a clear division of labor for these isoforms. For example, the homeodomain-containing form of Hth is not required for a large number of embryonic functions, including many (but not all) Hox-dependent functions. In contrast, the homeodomain-containing form is essential for hth to specify antennal development, which is one of its Hox-independent functions. The fact that these two Hth isoforms exist suggests the possibility that they may also be used in different ways to achieve Hox specificity. Strikingly, in C. elegans, the gene psa-3 encodes an HDless Hth/Meis ortholog (Arata et al., 2006). Thus, C. elegans produces a very similar HDless isoform, but via gene duplication and truncation, instead of by alternative splicing as in Drosophila and vertebrates. The presence of HDless isoforms of Hth/Meis in C. elegans, Drosophila, and vertebrates reinforces the idea that it carries out critical functions that are distinct from those executed by homeodomain-containing isoforms.

Although exd appears to produce only a single isoform, some of the vertebrate Pbx genes produce multiple isoforms via alternative splicing (Milech et al., 2001; Monica et al., 1991; Wagner et al., 2001). Using the yeast two-hybrid assay, some evidence exists that a subset of isoforms have distinct abilities to interact with Meis 1, Meis 2a, and Prep1 (Milech et al., 2001). Such differences may also be important for Hox specificity, and may be reflected in the arrangement of binding sites for Hox and TALE proteins in Hox-targeted cis-regulatory elements.

5. What Do In Vivo Hox-Binding Sites Look Like?

An important approach to understand how Hox proteins regulate target gene expression, and to reveal potential generalizations, is to examine the cis-regulatory elements they directly bind to in vivo. Once a set of in vivo-validated Hox-targeted cis-regulatory elements are in hand, several questions can be asked. These include: How many also require input from known cofactors? How many Hox-binding sites are present in each element?, and What other regulatory inputs are there? To provide initial answers to these questions, we have surveyed the literature with the goal of cataloging the majority of the direct Hox-binding sites that have been examined to date, in both vertebrates and invertebrates. By “direct,” we included in this survey only Hox-binding sites that have been shown by a reporter gene assay (in cell culture (ex vivo)or in vivo) to be required for the activity of a cis-regulatory element (Table 3.1). Therefore, some recent genome-wide studies fell short of these stringent criteria for validation (Ebner et al., 2005; Hueber and Lohmann, 2008; Hueber et al., 2007; McCabe and Innis, 2005). Below, we discuss the results of this survey and their implications for Hox specificity.

We found 66 cis-regulatory elements for which there is strong experimental evidence for direct and essential Hox input (Table 3.1). Of these, 29 have been shown to use PBC cofactors (Exd/Pbx/Ceh-20). Two additional elements appear, by sequence, to have PBC-Hox composite sites, making a total of 31 elements with PBC-Hox sites in this data set (Table 3.1). For seven elements, there is experimental evidence that they do not use these cofactors. The remaining 30 targets have not been directly tested for PBC input, although two of these have been shown not to use Hth/Meis or Prep proteins (Table 3.1). Finally, among the 66 targets in this list, there are 11 examples in which other direct inputs have been shown to be required for the activity of the cis-regulatory element.

Although we need to cautiously interpret this relatively small number of elements, several interesting features emerge by analyzing these examples. First, it is noteworthy that for a large fraction of the elements (29 of the 36 elements where it was examined) Hox proteins bind their binding sites cooperatively with a PBC factor (Table 3.1). This is likely an overestimate of the frequency of PBC-Hox-binding sites, because sequence gazing of the 30 elements that were untested for PBC input suggests that many do not have an obvious PBC-binding site. Nevertheless, the abundance of PBC-Hox composite-binding sites in this list underscores the widespread contribution of these cofactors to Hox-binding in vivo. Strikingly, with only two exceptions (Dll DMX and fkh250con), PBC-Hox-binding sites are used for gene activation, not repression. In contrast, the Hox sites that do not have clear PBC input are used for both repression and activation. If this overall correlation continues to hold up, it suggests that PBC-Hox complexes are, in general, more likely to recruit transcriptional coactivators rather than corepressors.

Second, there is a trend for the anterior Hox proteins (paralogs 1-5) to use PBC cofactors more than the posterior Hox proteins (paralogs 6-13) (Table 3.1). Of the 30 elements targeted by a Hox 1-5 paralog, 20 have a required PBC-Hox-binding site. In contrast, of the 36 elements targeted by a Hox 6-13 paralog, only nine have been shown to have an essential PBC-Hox-binding site.

Third, elements that do not have PBC input are more likely to have multiple Hox-binding sites than elements that have PBC input (Table 3.1). For those elements that use Hox-PBC sites, the average number of Hox-binding sites is 1.2 (ranging from 1 to 3), whereas for those elements that do not appear to use Hox-PBC sites, the average number of sites is 2.8 (ranging from 1 to 30). Thus, from this data set, it appears that Hox sites without PBC input often function in groups. If this trend holds up, it may reflect a lower affinity for non-PBC Hox sites when compared to PBC-Hox sites. Perhaps multiple non-PBC Hox sites are therefore required in an additive manner to elicit a transcriptional response.

Although many of the targets listed in Table 3.1 have Hox-binding sites that do not appear to have direct PBC input, we avoid calling them Hox “monomer” sites because it is plausible that currently unidentified factors bind with Hox proteins (cooperatively or noncooperatively) to these sites. In fact, when true Hox “monomer”-binding sites—synthetic, but high-affinity-binding sites—were used to drive a reporter gene in Drosophila embryos, they did not produce expression patterns consistent with their ability to bind dozens of homeodomain proteins (Vincent et al., 1990). This experiment, together with the analysis of in vivo Hox targets listed in Table 3.1, suggest that Hox proteins never work as monomers.

The arrangement of binding sites also varies in interesting ways within this data set. Of the 40 PBC-Hox sites (distributed among 31 elements), 33 have the PBC half-site adjacent to the Hox half-site, and the majority of these (26) have the structure nnATnnATnn (where the first and second ATs form the core of the PBC and Hox half-sites, respectively). In one case (DMX-R from Dll) the PBC-Hox site has the structure nnATnnnATnn and there is one example (in the dpp midgut element) where the Hox site precedes the PBC site (Table 3.1). It is possible that these atypical arrangements help these sites be more selective for a subset of Hox proteins. It is also possible that the unique three-dimensional architectures of the protein complexes assembled by these atypical PBC-Hox-binding sites are important for recruiting additional, element-specific transcriptional effectors. Consistent with these ideas, both in vitro Hox-binding specificity and in vivo activity of DMX-R were reduced when the spacing between the two half-sites was changed from 3 to 2 bp (Gebelein et al., 2002). Of the 40 PBC-Hox sites, eight have been shown to have a nearby Hth/Meis or Prep-binding site. Although the low number of identified Hth/Meis/Prep-binding sites may in part be because they are not always looked for, it may also reflect the fact that there are isoforms of Hth and Meis that do not have a homeodomain and thus would not be expected to make DNA contacts.

Because both PBC-Hox and Hth/Meis/Prep-binding sites have a clear orientation, four possible arrangements of these two binding sites are possible. Interestingly, of these four, only one is not observed (TGACAG...PBC-Hox, where TGACAG represents one orientation of a Hth/Meis site) (Table 3.1). That three of the potential orientations have been observed suggests an inherent flexibility in how these complexes can bind DNA. Further, because the different orientations are expected to orchestrate the assembly of protein-DNA complexes that have unique three-dimensional architectures, these observations also suggest the possibility that they have unique biochemical properties, such as their ability to recruit additional transcriptional coactivators or corepressors.

An unusual arrangement of binding sites is found in an Abd-A-targeted element from the Drosophila rhomboid (rho) gene (Table 3.1) (Li-Kroeger et al., 2008). In this element, the order of binding sites is PBC-Hth-Hox. Robust cooperative DNA binding to this element was observed between a preformed Exd-Hth dimer and Abd-A. In principle, because both Exd and Hth are TALE homeodomain proteins, they both have the ability to bind the “YPWM” motif of Abd-A, raising the question of which, if either, TALE motif Abd-A is interacting with. Interestingly, mutation of the Hth site, but not the Exd site, dramatically reduced complex formation (Li-Kroeger et al., 2008), suggesting that the Hth-binding site and, perhaps, the Hth TALE motif are more critical for complex formation. Perhaps analogously, a direct Hox-Hth interaction was also proposed to exist in the repressor element from the Dll gene (Table 3.1) (Gebelein et al., 2004). The existence of these atypical arrangements suggests that there may be additional flexibility in how Hox, PBC, and Meis/Hth/Prep proteins assemble onto target DNAs. The dissection of the rho element in particular emphasizes that carrying out careful mutagenesis and follow-up in vivo studies will be critical for identifying additional novel architectures that are used by these factors in vivo.

As discussed above, two recent reports described the in vitro binding site preferences for nearly all mouse and fly homeodomains, including all Hox homeodomains (Berger et al., 2008; Noyes et al., 2008). How do these results compare with the in vivo Hox-binding sites listed in Table 3.1? To answer this question, we generated binding site logo diagrams using the B1H-derived binding sites for the Drosophila Hox homeodomains (Noyes et al., 2008) and the in vivo Hox-binding sites listed in Table 3.1 (Fig. 3.2). The B1H-derived logos are all based on at least 19 individual binding sites, while the number of individual in vivo binding sites for each Hox protein that went into this analysis ranged from 12 (for Scr/Hox5) to 57 (for Ubx/Hox7). Side-by-side analysis of these two sets of sequences reveals some noteworthy differences (Fig. 3.2). First, consistent with the high proportion of PBC-Hox sites in the Hox1/Labial targets, the in vivo consensus sequence readily identified a PBC half-site (TGAT). In addition, while the B1H selection tended to identify TAATTA for Hox1/Labial, GGATGG is commonly observed in the in vivo data set for these Hox proteins. Other, though less dramatic, differences are also observed between the B1H and in vivo data sets for nearly all of the Hox paralogs (Fig. 3.2). These comparisons reinforce the view that the in vivo environment, due to the presence of other cofactors, collaborators, or differences in DNA and/or chromatin structure, influences Hox-binding site preferences.

6. Insights into Hox Specificity from Structural Studies

Several monomeric homeodomain-DNA structures have been solved, and all reveal a very similar mode of DNA recognition by this DNA-binding domain (Gehring et al., 1994). Briefly, the third alpha-helix, also called the recognition helix, lies in the major groove of the DNA, where it makes several direct and water-mediated contacts with specific bases and the phosphate backbone. Ile47, Gln50, Asn51, and Met54, residues that are present in all Hox homeodomains, are primarily responsible for making these contacts. In addition, the so-called N-terminal arm, which precedes the first alpha-helix, is typically observed in the minor groove. Arg5, an N-terminal arm residue present in nearly all homeodomains, is the most commonly observed residue in the minor groove.

In addition to these monomeric homeodomain-DNA structures, we now have X-ray structures of five different PBC-Hox-DNA complexes (Fig. 3.4). The Hox homeodomains in these structures recognize the DNA using the same contacts that were observed in the monomeric structures, demonstrating that the presence of PBC does not grossly alter the way in which Hox homeodomains bind DNA. In all five of the PBC-Hox structures (with four different Hox proteins: Hoxb1, Scr, Ubx, and Hoxa9), the PBC and Hox homeodomains bind DNA in a head-to-tail orientation, with very similar overall arrangements. In all five structures the Hox YPWM motif binds the hydrophobic TALE pocket in the PBC home-odomain. Thus, these four Hox proteins have the capacity to bind DNA cooperatively with PBC proteins using very similar protein-protein and protein-DNA contacts. However, we note that the currently available structures provide an incomplete picture because, for some binding sites, other protein motifs, such as the UbdA motif of Ubx and Abd-A, play an important role in forming PBC-Hox-DNA complexes (Merabet et al., 2007). Currently, no structural information exists about these domains or how they contact PBC proteins.

Figure 3.4
Common and unique features of PBC-Hox-DNA complexes. (A) Consensus PBC-Hox-binding sites have a PBC half-site (typically TGAT or AGAT, blue) and a Hox half-site (typically NNATNN, red). Minor groove (Arg5) and major groove (Asn51) contacts observed in ...

PBC proteins not only bind cooperatively to DNA with Hox proteins, they also increase Hox-DNA-binding selectivity. This phenomenon is best illustrated with a few examples. In the absence of cofactors, the Hox1/Labial paralog shows a preference for binding the sequence TAATTA (Fig. 3.2) (Berger et al., 2008; Noyes et al., 2008). In the presence of a PBC protein, a PBC-Hox1/Lab heterodimer prefers to bind the sequence TGAT[t/g]GATgg, where [t/g]GATgg is the Hox-binding site (base pairs with brackets indicate multiple possibilities at the same position and lowercase letters indicate only partially preferred base pairs) (Fig. 3.2). In contrast, while Ubx also prefers to bind TAATTA as a monomer, it will readily bind TGATTTATTT as a PBC-Ubx heterodimer, where TTATTT is the Hox-binding site (Berger et al., 2008; Noyes et al., 2008) (Fig. 3.2). Thus, the presence of PBC changes the DNA-binding preferences of both Labial/Hox1 and Ubx, but toward different sequences.

These observations raise the question of how the same cofactor can generate two different outcomes for these two Hox paralogs. The likely answer is that the specificity information is in the Hox protein, but is only revealed in the presence of the cofactor. Two recent crystal structures of PBC-Hox-DNA complexes support this idea (Joshi et al., 2007). In one, an Exd-Scr heterodimer is bound to a paralog-specific PBC-Hox-binding site (fkh250) from the fkh gene, an in vivo Scr target gene. A second crystal structure shows the same two proteins bound to a consensus PBC-Hox sequence (fkh250con). Importantly, additional protein-DNA contacts were observed in the Exd-Scr-fkh250 complex, but not in the Exd-Scr-fkh250con complex. These contacts are derived from the N-terminal arm of the Scr homeodomain and, surprisingly, a nonhomeodomain residue in the linker region between Scr's YPWM motif and homeodomain. Both of these side chains are inserted into the minor groove of the fkh250-binding site (Figs. 3.4 and 3.5) (Joshi et al., 2007). Two additional observations are of interest. First, the DNA minor groove where these two side chains insert is unusually narrow, and significantly narrower than the analogous region of the fkh250con-binding site (Fig. 3.5). This suggests that subtle differences in DNA structure are likely to contribute to Hox-binding specificity. Second, the residues making these minor groove contacts, which are part of a normally unstructured region of Scr, require the YPWM-Exd interaction to be positioned in close proximity to the minor groove. Thus, the paralog-specific binding of Scr to its binding site in fkh depends on three contributing features (1) an unusual DNA structure; (2) paralog-specific residues in the Scr homeodomain and linker that insert into this DNA structure; and (3) the YPWM-Exd interaction, which positions the N-terminal arm and linker region so that this normally unstructured peptide can make these contacts.

Figure 3.5
Interactions between Hox proteins and the DNA minor groove. Shown are images from X-ray crystal structures of PBC-Hox-DNA complexes, focused only on the interaction between the minor groove (shown as the gray surfaces) and the amino acid side chains of ...

Although these structures provide insights into why Exd-Scr specifically binds fkh250, they raise the question of how general these findings are. Two additional observations suggest that the underlying principles revealed by these structures may, in fact, be a general feature of PBC-Hox-DNA interactions, at least for paralog-specific and semi-paralog-specific target sites. For one, Hox N-terminal arms and linker regions are evolutionarily conserved in a paralog-specific manner (Joshi et al., 2007; Mann, 1995; Morgan et al., 2000). The two minor groove inserting residues in Scr are conserved in all Hox5 paralogs, while other Hox paralogs have different N-terminal arm and linker sequences that are also equally conserved in a paralog-specific manner. Second, another PBC-Hox structure, that of Pbx-Hoxa9, also has significant N-terminal arm-minor groove contacts (LaRonde-LeBlanc and Wolberger, 2003) (Figs. 3.4 and 3.5). Although the binding site used in the Pbx-HoxA9 structure is not from an in vivo HoxA9 target, the N-terminal arm-minor groove contacts are also dependent on the YPWM-Pbx interaction, illustrating the potential generality of DNA minor groove recognition by PBC-Hox complexes. The Pbx-Hoxa9 structure also reveals significantly more contacts with the phosphate backbone of the DNA, perhaps accounting for its higher affinity compared to other PBC-Hox complexes (LaRonde-LeBlanc and Wolberger, 2003).

Along the same lines, it is also noteworthy that linker length—and, consequently, the distance between the YPWM motif and the N-terminal arm—varies significantly among Hox proteins. Not only are there huge differences in linker lengths (ranging from >50 in Labial to <5 in Hoxa9), linker length roughly correlates with Hox paralog: anterior (3′) Hox paralogs have a much greater tendency for long linkers than more posterior (5′) Hox paralogs. In addition, Hox linkers also vary for individual paralogs due to alternative splicing; the Ubx linker, for example, ranges from 8 to 51 depending on the Ubx isoform (Kornfeld et al., 1989; O'Connor et al., 1988). Consequently, the distance between the YPWM motif and the homeodomain varies and would therefore be expected to affect how the N-terminal arm and/or linker region interacts with the DNA. These intriguing observations contribute to the idea that DNA contacts made by linker and N-terminal arm residues may be generally critical for paralog-specific and cofactor-dependent DNA binding. These observations are also consistent with the alternative idea that linker residues interact with additional proteins, although such factors have not yet been identified (Merabet et al., 2003).

In summary, the common and unique features revealed in these five PBC-Hox-DNA structures, together with previously solved monomeric homeodomain-DNA structures, suggest that DNA recognition by Hox proteins uses two tiers of information that provide different degrees of specificity (Fig. 3.6). The first tier uses DNA-contacting residues that are common to all Hox proteins (Arg5, Ile47, Gln50, Asn51, and Met54) to promote Hox binding to “AT”-rich sequences, such as TAAT[gt][ga]. The second tier uses additional DNA-contacting side chains that come from the N-terminal arm and linker regions; these contacts are cofactor-dependent and paralog-specific (Fig. 3.6). Further, at least for the case of Exd-Scr bound to fkh250, these side chains recognize a DNA structure, rather than a specific DNA sequence. The fact that N-terminal arm and linker residues are conserved in a paralog-specific manner suggests that the two tiers of recognition may underlie the recognition of paralog-specific DNA sequences by other Hox-cofactor complexes.

Figure 3.6
Two tiers of Hox-DNA-binding specificity. Hox proteins bind DNA using two levels of protein-DNA contacts. DNA contacts made by Arg5 (in the N-terminal arm) and Ile47, Gln50, Asn51, Met54 (in the third helix) are used by all Hox proteins to bind “AT”-rich ...

7. Activity Regulation of Hox Proteins: The Role of Hox Collaborators

Although TALE family proteins clearly play an important role in DNA-binding site recognition, Hox proteins use these cofactors to both activate and repress target genes, raising the question of how gene activation versus repression is determined. Although there is currently only one example, one answer is that Hox proteins may use dedicated repressors, such as En, as Hox cofactors in gene repression (Gebelein et al., 2004). Another possibility, which will be no surprise to people used to thinking about cis-regulatory elements, is that additional factors bind to Hox-targeted elements and contribute to their activities. Given the increasing number of directly regulated Hox targets that have been characterized, several such accessory factors, which we refer to here as Hox collaborators, have been identified. We have classified a factor as a Hox collaborator if it provides a direct and essential input into a Hox-regulated element, but has not been definitively shown to bind DNA cooperatively with Hox proteins (Table 3.1). These factors include the Drosophila Forkhead domain protein Sloppy paired (Slp), which collaborates with Ubx and Abd-A to repress Dll in the Drosophila abdomen (Gebelein et al., 2004). Interestingly, recent results suggest that a vertebrate ortholog of Slp, FoxP1, collaborates with Hox proteins during the establishment of motor neuron identities in the mouse (see Chapter 1) (Dasen et al., 2008; Rousso et al., 2008). Thus, the collaboration of FoxP1/Slp (and perhaps other Forkhead domain factors) with Hox proteins appears to be evolutionarily conserved. Like En, Slp may be a dedicated transcriptional repressor due to its Groucho interaction domain, thus providing a mechanism to explain the regulatory sign of Hox-Slp-targeted genes. In addition to these cases, protein-protein interaction and genetic studies suggest that the range of potential Hox collaborators is extensive (Kataoka et al., 2001; Luo et al., 2004; Plaza et al., 2008; Prévôt et al.,2000).

Transcription factors that provide cells with identity information, like the Hox factors, have been generally referred to as Selector transcription factors, a term originally coined 40 years ago to describe common properties of the genes Ubx and en (García-Bellido, 1975; Mann and Carroll, 2002; Mann and Morata, 2000). More recently, selector proteins have been proposed to frequently, if not always, function together with effector transcription factors that are downstream of cell-cell signaling pathways (reviewed previously by Bondos and Tan, 2001; Curtiss et al., 2002; Mann and Affolter, 1998). Thus, not surprisingly, as more elements that are directly targeted by Hox proteins are dissected, signal effector transcription factors are being identified as Hox collaborators. In particular, vertebrate and Drosophila SMADs, effectors of the TGF-beta and Decapentaplegic (Dpp) pathways, have been identified as Hox collaborators in several cis-regulatory elements (Galant et al., 2002; Grienenberger et al., 2003; Shi et al., 1999, 2001; Walsh and Carroll, 2007) (Table 3.1). Although SMAD-Hox-DNA-binding cooperativity has not been described, there are several reports suggesting that SMADs and Hox proteins can directly interact with each other (Wang et al., 2006; Williams et al., 2005; Zhou et al., 2008). Such interactions may be critical for building an enhanceosome-like structure on Hox-targeted cis-regulatory elements. Although the number of examples shown to have direct inputs by signaling effectors is currently low, genetic analyses suggest that this phenomenon is likely to be a general feature of Hox-targeted cis-regulatory elements, and will probably extend to other signaling pathways, including Hedgehog (Hh), Wnts, and Notch (Arata et al., 2006; Crickmore and Mann, 2007, 2008; Hersh et al., 2007; Joulia et al., 2006; Marty et al., 2001; Merabet et al., 2005; Weatherbee and Carroll, 1999).

Like Hox cofactors, the presence of a particular Hox collaborator does not guarantee the sign of the transcriptional regulation. In the sal1.1 element, for example, Ubx collaborates with Mad/Medea to repress transcription, while in the XC midgut element from the wg gene, Mad/Medea collaborates with Abd-A to activate transcription (Grienenberger et al., 2003; Walsh and Carroll, 2007). This difference is not simply due to different Hox paralogs, because both Ubx and Abd-A can both directly repress and directly activate transcription (Table 3.1). Instead, these observations imply that additional, currently unknown, factors are being recruited to these elements to determine the sign of the transcriptional regulation.

Based on these direct examples, together with the larger number of genetically defined examples of Hox—signaling collaborations, we suggest that it may be a general feature of the multiprotein complexes that are built on Hox-targeted cis-regulatory elements. In fact, because Hox proteins work in so many different developmental contexts, it is likely that Hox collaborators will ultimately include a very large number of different types of transcription factors. Perhaps the ability of Hox proteins and PBC-Hox dimers to interact with a large number of different collaborators makes these proteins such ideal regulators of cell type and tissue identities.

Yet, despite this flexibility, it is critical to stress that the Hox factors play the central role in the function (and/or the assembly) of these multiprotein complexes, because without them, these complexes cannot function. Moreover, for paralog-specific functions, the activity and/or assembly of these complexes must depend on the correct Hox paralog and cofactors. Because of their central role, we would therefore like to coin the term “Hoxasome” to describe these multiprotein complexes, which include the Hox proteins, their cofactors, and their collaborators.

8. Insights into Hoxasome Function from cis-Regulatory Element Architecture

One straightforward view for how Hoxasomes function is that, once assembled, they recruit coactivators, corepressors, and/or chromatin remodeling complexes that ultimately carry out transcriptional regulation much like any other enhanceosome. Indeed, consistent with this view, there have been numerous reports describing direct interactions between Hox proteins and/or TALE cofactors with these more general components of the transcriptional machinery (Chariot et al., 1999; Prince et al., 2008; Saleh et al., 2000; Shen et al., 2004) and, in some cases, activation and repression domains have been mapped in Hox proteins (Rambaldi et al., 1994; Tour, 2005; Viganò et al., 1998; Zhao et al., 1996). Covalent modifications, such as phosphorylation, have also been shown to influence Hox activities in interesting ways (Berry and Gehring, 2000; Galant and Carroll, 2002; Jaffe et al., 1997; Ronshaugen et al., 2002; Taghli-Lamallem et al., 2008; reviewed elsewhere by Pearson et al., 2005).

In addition, to these mechanisms, there are some recent examples suggesting that Hoxasomes may regulate transcription—and be regulated themselves—in other, mechanistically distinct ways. One example concerns the way in which Drosophila Abd-A activates the expression of the gene rho in the peripheral nervous system (PNS). Abd-A-dependent expression of rho, which encodes a protease required for the processing of a ligand for the epidermal growth factor (EGF) receptor pathway, is necessary for abdominal-specific cell types (Brodu et al., 2002). Abd-A carries out this paralog-specific function by assembling a multiprotein complex that includes both Exd and Hth cofactors (Li-Kroeger et al., 2008). However, instead of the Abd-A Hoxasome directly activating rho transcription, it appears that it functions by competing with the binding of another transcription factor, Senseless, which is a rho repressor. Thus, the architecture of the rho cis-regulatory element is organized in a manner whereby the sequence-specific binding of an Abd-A Hoxasome permits rho expression by blocking binding of the Senseless repressor. Perhaps analogously, a zinc finger protein called ZFPIP has been shown to bind to Pbx1 and inhibit the binding of Hoxa9-Pbx1 complexes to a Hox-PBC consensus site (Laurent et al., 2007). Thus, competition in DNA binding, rather than a direct influence on transcription, may underlie other examples of gene regulation by Hox factors.

A second interesting example of the importance of cis-regulatory element architecture comes from the analysis of an element from the Drosophila bric-a-brac (bab) gene, which is a direct target of Abd-B (Williams et al., 2008). The activity of this element in the fifth and sixth abdominal segments of females—but not males—is critical for the dimorphic nature of abdominal pigmentation in male and female flies (Williams et al., 2008). As Abd-B expression is the same in male and female flies, the sex-specific activities of this element stem from the Hox collaborator, doublesex (dsx), which is a downstream effector in the sex-determination pathway (Christiansen et al., 2002). dsx encodes both male and female-specific isoforms. In males, the Dsx-M isoform collaborates with Abd-B to repress bab, while in females, the Dsx-F isoform collaborates with Abd-B to activate this bab element. In this element, there are two required Bab-binding sites, and more than 15 Abd-B-binding sites, suggesting the existence an unusually Hox-dense Hoxasome (Table 3.1). Moreover, the analysis of the same cis-regulatory element from other Drosophila species in which abdominal pigmentation pattern is the same in males and females suggests that the dimorphic activity of the D. melanogaster bab element is due to the relative orientation and specific spacing of the Dsx and Abd-B-binding sites. Mechanistically, it is currently unclear if these changes affect the stable assembly of this Hoxasome or, alternatively, its ability to recruit transcriptional coactivators.

Both of the studies highlighted above emphasize the value in characterizing bona fide Hox-targeted cis-regulatory elements at high resolution. Moreover, they also make it clear that a complete understanding of gene regulation by Hox proteins not only depends on understanding how these transcription factors bind DNA, but also how the bound factors, together with their cofactors and collaborators, assemble and regulate transcription. We suspect that the discovery of additional regulatory mechanisms will depend on similar fine-scale analysis of other Hox-targeted cis-regulatory elements.

9. Conclusions

In this review, we have summarized a wide range of mechanisms that Hox proteins employ to regulate their target genes. For one, Hox proteins often require cofactors to bind to their binding sites in paralog-specific and semi-paralog-specific target genes. Cofactors may not be as essential, however, for shared Hox functions or those executed by Hox proteins in a unique regulatory environment, such as the Drosophila haltere. Structural studies have suggested that TALE family cofactors not only increase the size of the binding site, they help to impose additional structure onto otherwise unstructured homeodomain and nonhomeodomain residues, allowing them to read additional features present in Hox-cofactor-binding sites. It will be interesting to see how generally applicable this model for Hox-DNA binding (and perhaps other homeodomain proteins) will be as more PBCHox-DNA complexes are characterized at high resolution. Finally, we have also seen that Hox-regulated cis-regulatory elements utilize a potentially large number of protein collaborators, such as effector transcription factors that are downstream of cell-cell signaling pathways. The assembly of these multiprotein-DNA complexes, which we have called Hoxasomes to emphasize the central importance of the Hox input, is essential for dictating the sign (repression or activation) of the transcriptional regulation.


We thank Matthew Slattery for comments on the manuscript and Barry Honig for discussions related to this review. This work was supported by an NIH RO1 grant awarded to R.S.M.


  • Affolter M, Percival-Smith A, Müller M, Leupin W, Gehring WJ. DNA binding properties of the purified Antennapedia homeodomain. Proc. Natl. Acad. Sci. USA. 1990;87:4093–4097. [PMC free article] [PubMed]
  • Affolter M, Slattery M, Mann RS. A lexicon for homeodomain-DNA recognition. Cell. 2008;133:1133–1135. [PubMed]
  • Akam M. Hox and HOM: homologous gene clusters in insects and vertebrates. Cell. 1989;57:347–349. [PubMed]
  • Appel B, Sakonju S. Cell-type-specific mechanisms of transcriptional repression by the homeotic gene products UBX and ABD-A in Drosophila embryos. EMBO J. 1993;12:1099–1109. [PMC free article] [PubMed]
  • Arata Y, Kouike H, Zhang Y, Herman MA, Okano H, Sawa H. Wnt signaling and a Hox protein cooperatively regulate psa-3/Meis to determine daughter cell fate after asymmetric cell division in C. elegans. Dev. Cell. 2006;11:105–115. [PubMed]
  • Bandyopadhyay S, Ashraf MZ, Daher P, Howe PH, DiCorleto PE. HOXA9 participates in the transcriptional activation of E-selectin in endothelial cells. Mol. Cell Biol. 2007;27:4207–4216. [PMC free article] [PubMed]
  • Berger MF, Badis G, Gehrke AR, Talukder S, Philippakis AA, Peña-Castillo L, Alleyne TM, Mnaimneh S, Botvinnik OB, Chan ET, Khalid F, Zhang W, et al. Variation in homeodomain DNA binding revealed by high-resolution analysis of sequence preferences. Cell. 2008;133:1266–1276. [PMC free article] [PubMed]
  • Berry M, Gehring W. Phosphorylation status of the SCR homeodomain determines its functional activity: essential role for protein phosphatase 2A,B’. EMBO J. 2000;19:2946–2957. [PMC free article] [PubMed]
  • Berthelsen J, Kilstrup-Nielsen C, Blasi F, Mavilio F, Zappavigna V. The subcellular localization of PBX1 and EXD proteins depends on nuclear import and export signals and is modulated by association with PREP1 and HTH. Genes. Dev. 1999;13:946–953. [PMC free article] [PubMed]
  • Bessa J, Tavares MJ, Santos J, Kikuta H, Laplante M, Becker TS, Gómez-Skarmeta JL, Casares F. meis1 regulates cyclin D1 and c-myc expression, and controls the proliferation of the multipotent cells in the early developing zebrafish eye. Development. 2008;135:799–803. [PubMed]
  • Bondos SE, Tan XX. Combinatorial transcriptional regulation: the interaction of transcription factors and cell signaling molecules with homeodomain proteins in Drosophila development. Crit. Rev. Eukaryot. Gene. Expr. 2001;11:145–171. [PubMed]
  • Bradley PL, Haberman AS, Andrew DJ. Organ formation in Drosophila: specification and morphogenesis of the salivary gland. Bioessays. 2001;23:901–911. [PubMed]
  • Brodu V, Elstob PR, Gould AP. abdominal A specifies one cell type in Drosophila by regulating one principal target gene. Development. 2002;129:2957–2963. [PubMed]
  • Bromleigh VC, Freedman LP. p21 is a transcriptional target of HOXA10 in differentiating myelomonocytic cells. Genes. Dev. 2000;14:2581–2586. [PMC free article] [PubMed]
  • Bruhl T, Urbich C, Aicher D, Acker-Palmer A, Zeiher AM, Dimmeler S. Homeobox A9 transcriptionally regulates the EphB4 receptor to modulate endothelial cell migration and tube formation. Circ. Res. 2004;94:743–751. [PubMed]
  • Capovilla M, Botas J. Functional dominance among Hox genes: repression dominates activation in the regulation of Dpp. Development. 1998;125:4949–4957. [PubMed]
  • Capovilla M, Brandt M, Botas J. Direct regulation of decapentaplegic by Ultrabithorax and its role in Drosophila midgut morphogenesis. Cell. 1994;76:461–475. [PubMed]
  • Capovilla M, Kambris Z, Botas J. Direct regulation of the muscle-identity gene apterous by a Hox protein in the somatic mesoderm. Development. 2001;128:1221–1230. [PubMed]
  • Caré A, Silvani A, Meccia E, Mattia G, Stoppacciaro A, Parmiani G, Peschle C, Colombo MP. HOXB7 constitutively activates basic fibroblast growth factor in melanomas. Mol. Cell Biol. 1996;16:4842–4851. [PMC free article] [PubMed]
  • Casares F, Calleja M, Sánchez-Herrero E. Functional similarity in appendage specification by the Ultrabithorax and abdominal-A Drosophila HOX genes. EMBO J. 1996;15:3934–3942. [PMC free article] [PubMed]
  • Casares F, Mann RS. Control of antennal versus leg development in Drosophila. Nature. 1998;392:723–726. [PubMed]
  • Casares F, Mann RS. The ground state of the ventral appendage in Drosophila. Science. 2001;293:1477–1480. [PubMed]
  • Ceccarelli E, McGrew MJ, Nguyen T, Grieshammer U, Horgan D, Hughes SH, Rosenthal N. An E box comprises a positional sensor for regional differences in skeletal muscle gene expression and methylation. Dev. Biol. 1999;213:217–229. [PubMed]
  • Chan SK, Jaffe L, Capovilla M, Botas J, Mann RS. The DNA binding specificity of Ultrabithorax is modulated by cooperative interactions with extradenticle, another homeoprotein. Cell. 1994;78:603–615. [PubMed]
  • Chan SK, Mann RS. The segment identity functions of Ultrabithorax are contained within its homeo domain and carboxy-terminal sequences. Genes Dev. 1993;7:796–811. [PubMed]
  • Chan SK, Mann RS. A structural model for a homeotic protein-extra-denticle-DNA complex accounts for the choice of HOX protein in the heterodimer. Proc. Natl. Acad. Sci. USA. 1996;93:5223–5228. [PMC free article] [PubMed]
  • Chan SK, Ryoo HD, Gould A, Krumlauf R, Mann RS. Switching the in vivo specificity of a minimal Hox-responsive element. Development. 1997;124:2007–2014. [PubMed]
  • Chang CP, Shen WF, Rozenfeld S, Lawrence HJ, Largman C, Cleary ML. Pbx proteins display hexapeptide-dependent cooperative DNA binding with a subset of Hox proteins. Genes Dev. 1995;9:663–674. [PubMed]
  • Chariot A, van Lint C, Chapelier M, Gielen J, Merville MP, Bours V. CBP and histone deacetylase inhibition enhance the transactivation potential of the HOXB7 homeodomain-containing protein. Oncogene. 1999;18:4007–4014. [PubMed]
  • Chen H, Rubin E, Zhang H, Chung S, Jie CC, Garrett E, Biswal S, Sukumar S. Identification of transcriptional targets of HOXA5. J. Biol. Chem. 2005;280:19373–19380. [PubMed]
  • Chen J, Ruley HE. An enhancer element in the EphA2 (Eck) gene sufficient for rhombomere-specific expression is activated by HOXA1 and HOXB1 homeobox proteins. J. Biol. Chem. 1998;273:24670–24675. [PubMed]
  • Christiansen AE, Keisman EL, Ahmad SM, Baker BS. Sex comes in from the cold: the integration of sex and pattern. Trends Genet. 2002;18:510–516. [PubMed]
  • Coiffier D, Charroux B, Kerridge S. Common functions of central and posterior Hox genes for the repression of head in the trunk of Drosophila. Development. 2008;135:291–300. [PubMed]
  • Crickmore MA, Mann RS. Hox control of morphogen mobility and organ development through regulation of glypican expression. Development. 2007;134:327–334. [PubMed]
  • Crickmore MA, Mann RS. The control of size in animals: insights from selector genes. Bioessays. 2008;30:843–853. [PMC free article] [PubMed]
  • Cui M, Han M. Cis regulatory requirements for vulval cell-specific expression of the Caenorhabditis elegans fibroblast growth factor gene egl-17. Dev. Biol. 2003;257:104–116. [PubMed]
  • Curtiss J, Halder G, Mlodzik M. Selector and signalling molecules cooperate in organ patterning. Nat. Cell Biol. 2002;4:E48–E51. [PubMed]
  • Daftary GS, Troy PJ, Bagot CN, Young SL, Taylor HS. Direct regulation of beta3-integrin subunit gene expression by HOXA10 in endometrial cells. Mol. Endocrinol. 2002;16:571–579. [PubMed]
  • Dasen JS, De Camilli A, Wang B, Tucker PW, Jessell TM. Hox repertoires for motor neuron diversity and connectivity gated by a single accessory factor, FoxP1. Cell. 2008;134:304–316. [PubMed]
  • Desplan C, Theis J, O'Farrell PH. The sequence specificity of homeodomain-DNA interaction. Cell. 1988;54:1081–1090. [PMC free article] [PubMed]
  • Dessain S, Gross CT, Kuziora MA, McGinnis W. Antp-type homeodomains have distinct DNA binding specificities that correlate with their different regulatory functions in embryos. EMBO J. 1992;11:991–1002. [PMC free article] [PubMed]
  • Di Rocco G, Gavalas A, Popperl H, Krumlauf R, Mavilio F, Zappavigna V. The recruitment of SOX/OCT complexes and the differential activity of HOXA1 and HOXB1 modulate the Hoxb1 auto-regulatory enhancer function. J. Biol. Chem. 2001;276:20506–20515. [PubMed]
  • Ebner A, Cabernard C, Affolter M, Merabet S. Recognition of distinct target sites by a unique Labial/Extradenticle/Homothorax complex. Development. 2005;132:1591–1600. [PubMed]
  • Edelman GM, Jones FS. Developmental control of N-CAM expression by Hox and Pax gene products. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 1995;349:305–312. [PubMed]
  • Ekker SC, Jackson DG, von Kessler DP, Sun BI, Young KE, Beachy PA. The degree of variation in DNA sequence recognition among four Drosophila homeotic proteins. EMBO J. 1994;13:3551–3560. [PMC free article] [PubMed]
  • Ekker SC, Young KE, von Kessler DP, Beachy PA. Optimal DNA sequence recognition by the Ultrabithorax homeodomain of Drosophila. EMBO J. 1991;10:1179–1186. [PMC free article] [PubMed]
  • Estrada B, Sánchez-Herrero E. The Hox gene Abdominal-B antagonizes appendage development in the genital disc of Drosophila. Development. 2001;128:331–339. [PubMed]
  • Ferretti E, Cambronero F, Tümpel S, Longobardi E, Wiedemann LM, Blasi F, Krumlauf R. Hoxb1 enhancer and control of rhombomere 4 expression: complex interplay between PREP1-PBX1-HOXB1 binding sites. Mol. Cell Biol. 2005;25:8541–8552. [PMC free article] [PubMed]
  • Ferretti E, Marshall H, Pöpperl H, Maconochie M, Krumlauf R, Blasi F. Segmental expression of Hoxb2 in r4 requires two separate sites that integrate cooperative interactions between Prep1, Pbx and Hox proteins. Development. 2000;127:155–166. [PubMed]
  • Furukubo-Tokunaga K, Flister S, Gehring WJ. Functional specificity of the Antennapedia homeodomain. Proc. Natl. Acad. Sci. USA. 1993;90:6360–6364. [PMC free article] [PubMed]
  • Galant R, Carroll SB. Evolution of a transcriptional repression domain in an insect Hox protein. Nature. 2002;415:910–913. [PubMed]
  • Galant R, Walsh CM, Carroll SB. Hox repression of a target gene: extradenticle-independent, additive action through multiple monomer binding sites. Development. 2002;129:3115–3126. [PubMed]
  • García-Bellido A. Genetic control of wing disc development in Drosophila. Ciba Found Symp. 1975;0:161–182. [PubMed]
  • Gebelein B, Culi J, Ryoo HD, Zhang W, Mann RS. Specificity of Distalless repression and limb primordia development by abdominal Hox proteins. Dev. Cell. 2002;3:487–498. [PubMed]
  • Gebelein B, Mann R. Compartmental modulation of abdominal Hox expression by engrailed and sloppy-paired patterns the fly ectoderm. Dev. Biol. 2007;308:593–605. [PMC free article] [PubMed]
  • Gebelein B, McKay DJ, Mann RS. Direct integration of Hox and segmentation gene inputs during Drosophila development. Nature. 2004;431:653–659. [PubMed]
  • Gehring WJ, Qian YQ, Billeter M, Furukubo-Tokunaga K, Schier AF, Resendez-Perez D, Affolter M, Otting G, Wüthrich K. Homeodomain-DNA recognition. Cell. 1994;78:211–223. [PubMed]
  • Gibson G, Schier A, LeMotte P, Gehring WJ. The specificities of Sex combs reduced and Antennapedia are defined by a distinct portion of each protein that includes the homeodomain. Cell. 1990;62:1087–1103. [PubMed]
  • Glenn ST, Jones CA, Pan L, Gross KW. In Vivo Analysis of Key Elements within the Renin Regulatory Region. Physiol. Genomics. 2008;35(3):243–253. [PMC free article] [PubMed]
  • Gong KQ, Yallowitz AR, Sun H, Dressler GR, Wellik DM. A Hox-Eya-Pax complex regulates early kidney developmental gene expression. Mol. Cell Biol. 2007;27:7661–7668. [PMC free article] [PubMed]
  • Goodman FR. Limb malformations and the human HOX genes. Am. J. Med. Genet. 2002;112:256–265. [PubMed]
  • Gould A, Morrison A, Sproat G, White RA, Krumlauf R. Positive cross-regulation and enhancer sharing: two mechanisms for specifying overlapping Hox expression patterns. Genes Dev. 1997;11:900–913. [PubMed]
  • Greig S, Akam M. The role of homeotic genes in the specification of the Drosophila gonad. Curr. Biol. 1995;5:1057–1062. [PubMed]
  • Grienenberger A, Merabet S, Manak J, Iltis I, Fabre A, Bérenger H, Scott MP, Pradel J, Graba Y. Tgfbeta signaling acts on a Hox response element to confer specificity and diversity to Hox protein function. Development. 2003;130:5445–5455. [PubMed]
  • Guazzi S, Lonigro R, Pintonello L, Boncinelli E, Di Lauro R, Mavilio F. The thyroid transcription factor-1 gene is a candidate target for regulation by Hox proteins. EMBO J. 1994;13:3339–3347. [PMC free article] [PubMed]
  • Haerry TE, Gehring WJ. A conserved cluster of homeodomain binding sites in the mouse Hoxa-4 intron functions in Drosophila embryos as an enhancer that is directly regulated by Ultrabithorax. Dev. Biol. 1997;186:1–15. [PubMed]
  • Haller K, Rambaldi I, Daniels E, Featherstone M. Subcellular localization of multiple PREP2 isoforms is regulated by actin, tubulin, and nuclear export. J. Biol. Chem. 2004;279:49384–49394. [PubMed]
  • Hassan MQ, Tare R, Lee SH, Mandeville M, Weiner B, Montecino M, van Wijnen AJ, Stein JL, Stein GS, Lian JB. HOXA10 controls osteoblastogenesis by directly activating bone regulatory and phenotypic genes. Mol. Cell Biol. 2007;27:3337–3352. [PMC free article] [PubMed]
  • Hersh BM, Carroll SB. Direct regulation of knot gene expression by Ultrabithorax and the evolution of cis-regulatory elements in Drosophila. Development. 2005;132:1567–1577. [PubMed]
  • Hersh BM, Nelson CE, Stoll SJ, Norton JE, Albert TJ, Carroll SB. The UBX-regulated network in the haltere imaginal disc of D. melanogaster. Dev. Biol. 2007;302:717–727. [PMC free article] [PubMed]
  • Hirth F, Loop T, Egger B, Miller DF, Kaufman TC, Reichert H. Functional equivalence of Hox gene products in the specification of the tritocerebrum during embryonic brain development of Drosophila. Development. 2001;128:4781–4788. [PubMed]
  • Hoey T, Levine M. Divergent homeo box proteins recognize similar DNA sequences in Drosophila. Nature. 1988;332:858–861. [PubMed]
  • Houghton L, Rosenthal N. Regulation of a muscle-specific transgene by persistent expression of Hox genes in postnatal murine limb muscle. Dev. Dyn. 1999;216:385–397. [PubMed]
  • Huang H, Paliouras M, Rambaldi I, Lasko P, Featherstone M. Nonmuscle myosin promotes cytoplasmic localization of PBX. Mol. Cell Biol. 2003;23:3636–3645. [PMC free article] [PubMed]
  • Hueber SD, Bezdan D, Henz SR, Blank M, Wu H, Lohmann I. Comparative analysis of Hox downstream genes in Drosophila. Development. 2007;134:381–392. [PubMed]
  • Hueber SD, Lohmann I. Shaping segments: Hox gene function in the genomic age. Bioessays. 2008;30:965–979. [PubMed]
  • Jaffe L, Ryoo HD, Mann RS. A role for phosphorylation by casein kinase II in modulating Antennapedia activity in Drosophila. Genes Dev. 1997;11:1327–1340. [PubMed]
  • Jave-Suárez LF, Schweizer J. The HOXC13-controlled expression of early hair keratin genes in the human hair follicle does not involve TALE proteins MEIS and PREP as cofactors. Arch. Dermatol. Res. 2006;297:372–376. [PubMed]
  • Jave-Suarez LF, Winter H, Langbein L, Rogers MA, Schweizer J. HOXC13 is involved in the regulation of human hair keratin gene expression. J. Biol. Chem. 2002;277:3718–3726. [PubMed]
  • Jeong S, Rokas A, Carroll SB. Regulation of body pigmentation by the Abdominal-B Hox protein and its gain and loss in Drosophila evolution. Cell. 2006;125:1387–1399. [PubMed]
  • Jiang Y, Shi H, Amin NM, Sultan I, Liu J. Mesodermal expression of the C. elegans HMX homolog mls-2 requires the PBC homolog CEH-20. Mech. Dev. 2008;125:451–461. [PMC free article] [PubMed]
  • Jiménez G, Paroush Z, Ish-Horowicz D. Groucho acts as a corepressor for a subset of negative regulators, including Hairy and Engrailed. Genes Dev. 1997;11:3072–3082. [PMC free article] [PubMed]
  • Jones FS, Holst BD, Minowa O, De Robertis EM, Edelman GM. Binding and transcriptional activation of the promoter for the neural cell adhesion molecule by HoxC6 (Hox-3.3). Proc. Natl. Acad. Sci. USA. 1993;90:6557–6561. [PMC free article] [PubMed]
  • Jones FS, Prediger EA, Bittner DA, De Robertis EM, Edelman GM. Cell adhesion molecules as targets for Hox genes: neural cell adhesion molecule promoter activity is modulated by cotransfection with Hox-2.5 and -2.4. Proc. Natl. Acad. Sci. USA. 1992;89:2086–2090. [PMC free article] [PubMed]
  • Joshi R, Passner J, Rohs R, Jain R, Sosinsky A, Crickmore M, Jacob V, Aggarwal A, Honig B, Mann R. Functional Specificity of a Hox Protein Mediated by the Recognition of Minor Groove Structure. Cell. 2007;131:530–543. [PMC free article] [PubMed]
  • Joulia L, Deutsch J, Bourbon HM, Cribbs DL. The specification of a highly derived arthropod appendage, the Drosophila labial palps, requires the joint action of selectors and signaling pathways. Dev. Genes Evol. 2006;216:431–442. [PubMed]
  • Kataoka K, Yoshitomo-Nakagawa K, Shioda S, Nishizawa M. A set of Hox proteins interact with the Maf oncoprotein to inhibit its DNA binding, transactivation, and transforming activities. J. Biol. Chem. 2001;276:819–826. [PubMed]
  • Knosp WM, Saneyoshi C, Shou S, Bächinger HP, Stadler HS. Elucidation, quantitative refinement, and in vivo utilization of the HOXA13 DNA binding site. J. Biol. Chem. 2007;282:6843–6853. [PubMed]
  • Koh K, Peyrot SM, Wood CG, Wagmaister JA, Maduro MF, Eisenmann DM, Rothman JH. Cell fates and fusion in the C. elegans vulval primordium are regulated by the EGL-18 and ELT-6 GATA factors -- apparent direct targets of the LIN-39 Hox protein. Development. 2002;129:5171–5180. [PubMed]
  • Kornfeld K, Saint RB, Beachy PA, Harte PJ, Peattie DA, Hogness DS. Structure and expression of a family of Ultrabithorax mRNAs generated by alternative splicing and polyadenylation in Drosophila. Genes Dev. 1989;3:243–258. [PubMed]
  • Kremser T, Hasenpusch-Theil K, Wagner E, Buttgereit D, Renkawitz-Pohl R. Expression of the beta3 tubulin gene (beta Tub60D) in the visceral mesoderm of Drosophila is dependent on a complex enhancer that binds Tinman and UBX. Mol. Gen. Genet. 1999;262:643–658. [PubMed]
  • Kuziora MA, McGinnis W. A homeodomain substitution changes the regulatory specificity of the deformed protein in Drosophila embryos. Cell. 1989;59:563–571. [PubMed]
  • Kuziora MA, McGinnis W. Altering the regulatory targets of the Deformed protein in Drosophila embryos by substituting the Abdominal-B homeodomain. Mech. Dev. 1990;33:83–93. [PubMed]
  • Lampe X, Picard JJ, Rezsohazy R. The Hoxa2 enhancer 2 contains a critical Hoxa2 responsive regulatory element. Biochem. Biophys. Res. Commun. 2004;316:898–902. [PubMed]
  • Lampe X, Samad OA, Guiguen A, Matis C, Remacle S, Picard JJ, Rijli FM, Rezsohazy R. An ultraconserved Hox-Pbx responsive element resides in the coding sequence of Hoxa2 and is active in rhombomere 4. Nucleic Acids Res. 2008;36:3214–3225. [PMC free article] [PubMed]
  • LaRonde-LeBlanc NA, Wolberger C. Structure of HoxA9 and Pbx1 bound to DNA: Hox hexapeptide and DNA recognition anterior to posterior. Genes Dev. 2003;17:2060–2072. [PMC free article] [PubMed]
  • Laurent A, Bihan R, Deschamps S, Guerrier D, Dupé V, Omilli F, Burel A, Pellerin I. Identification of a new type of PBX1 partner that contains zinc finger motifs and inhibits the binding of HOXA9-PBX1 to DNA. Mech. Dev. 2007;124:364–376. [PubMed]
  • Lei H, Wang H, Juan AH, Ruddle FH. The identification of Hoxc8 target genes. Proc. Natl. Acad. Sci. USA. 2005;102:2420–2424. [PMC free article] [PubMed]
  • Lewis EB. A gene complex controlling segmentation in Drosophila. Nature. 1978;276:565–570. [PubMed]
  • Li-Kroeger D, Witt LM, Grimes HL, Cook TA, Gebelein B. Hox and senseless antagonism functions as a molecular switch to regulate EGF secretion in the Drosophila PNS. Dev. Cell. 2008;15:298–308. [PMC free article] [PubMed]
  • Lin L, McGinnis W. Mapping functional specificity in the Dfd and Ubx homeo domains. Genes Dev. 1992;6:1071–1081. [PubMed]
  • Liu H, Strauss TJ, Potts MB, Cameron S. Direct regulation of egl-1 and of programmed cell death by the Hox protein MAB-5 and by CEH-20, a C. elegans homolog of Pbx1. Development. 2006;133:641–650. [PubMed]
  • Liu J, Fire A. Overlapping roles of two Hox genes and the exd ortholog ceh-20 in diversification of the C. elegans postembryonic mesoderm. Development. 2000;127:5179–5190. [PubMed]
  • Lohmann I, McGinnis N, Bodmer M, McGinnis W. The Drosophila Hox gene deformed sculpts head morphology via direct regulation of the apoptosis activator reaper. Cell. 2002;110:457–466. [PubMed]
  • Lu Q, Kamps MP. Structural determinants within Pbx1 that mediate cooperative DNA binding with pentapeptide-containing Hox proteins: proposal for a model of a Pbx1-Hox-DNA complex. Mol. Cell Biol. 1996;16:1632–1640. [PMC free article] [PubMed]
  • Luo L, Yang X, Takihara Y, Knoetgen H, Kessel M. The cell-cycle regulator geminin inhibits Hox function through direct and polycomb-mediated interactions. Nature. 2004;427:749–753. [PubMed]
  • Manak JR, Mathies LD, Scott MP. Regulation of a decapentaplegic midgut enhancer by homeotic proteins. Development. 1994;120:3605–3619. [PubMed]
  • Mann RS. The specificity of homeotic gene function. Bioessays. 1995;17:855–863. [PubMed]
  • Mann RS, Abu-Shaar M. Nuclear import of the homeodomain protein extradenticle in response to Wg and Dpp signalling. Nature. 1996;383:630–633. [PubMed]
  • Mann RS, Affolter M. Hox proteins meet more partners. Curr. Opin. Genet. Dev. 1998;8:423–429. [PubMed]
  • Mann RS, Carroll SB. Molecular mechanisms of selector gene function and evolution. Curr. Opin. Genet. Dev. 2002;12:592–600. [PubMed]
  • Mann RS, Chan SK. Extra specificity from extradenticle: the partnership between HOX and PBX/EXD homeodomain proteins. Trends Genet. 1996;12:258–262. [PubMed]
  • Mann RS, Hogness DS. Functional dissection of Ultrabithorax proteins in D. melanogaster. Cell. 1990;60:597–610. [PubMed]
  • Mann RS, Morata G. The developmental and molecular biology of genes that subdivide the body of Drosophila. Annu. Rev. Cell Dev. Biol. 2000;16:243–271. [PubMed]
  • Manzanares M, Bel-Vialar S, Ariza-McNaughton L, Ferretti E, Marshall H, Maconochie MM, Blasi F, Krumlauf R. Independent regulation of initiation and maintenance phases of Hoxa3 expression in the vertebrate hindbrain involve auto- and cross-regulatory mechanisms. Development. 2001;128:3595–3607. [PubMed]
  • Marty T, Vigano MA, Ribeiro C, Nussbaumer U, Grieder NC, Affolter M. A HOX complex, a repressor element and a 50 bp sequence confer regional specificity to a DPP-responsive enhancer. Development. 2001;128:2833–2845. [PubMed]
  • McCabe CD, Innis JW. A genomic approach to the identification and characterization of HOXA13 functional binding elements. Nucleic Acids Res. 2005;33:6782–6794. [PMC free article] [PubMed]
  • McCormick A, Coré N, Kerridge S, Scott MP. Homeotic response elements are tightly linked to tissue-specific elements in a transcriptional enhancer of the teashirt gene. Development. 1995;121:2799–2812. [PubMed]
  • Merabet S, Hombria JC, Hu N, Pradel J, Graba Y. Hox-controlled reorganisation of intrasegmental patterning cues underlies Drosophila posterior spiracle organogenesis. Development. 2005;132:3093–3102. [PubMed]
  • Merabet S, Kambris Z, Capovilla M, Bérenger H, Pradel J, Graba Y. The hexapeptide and linker regions of the AbdA Hox protein regulate its activating and repressive functions. Dev. Cell. 2003;4:761–768. [PubMed]
  • Merabet S, Saadaoui M, Sambrani N, Hudry B, Pradel J, Affolter M, Graba Y. A unique Extradenticle recruitment mode in the Drosophila Hox protein Ultra-bithorax. Proc. Natl. Acad. Sci. USA. 2007;104:16946–16951. [PMC free article] [PubMed]
  • Milech N, Kees UR, Watt PM. Novel alternative PBX3 isoforms in leukemia cells with distinct interaction specificities. Genes Chromosomes Cancer. 2001;32:275–780. [PubMed]
  • Moens CB, Selleri L. Hox cofactors in vertebrate development. Dev. Biol. 2006;291:193–206. [PubMed]
  • Monica K, Galili N, Nourse J, Saltman D, Cleary ML. PBX2 and PBX3, new homeobox genes with extensive homology to the human proto-oncogene PBX1. Mol. Cell Biol. 1991;11:6149–6157. [PMC free article] [PubMed]
  • Morgan R, In der Rieden P, Hooiveld MH, Durston AJ. Identifying HOX paralog groups by the PBX-binding region. Trends Genet. 2000;16:66–67. [PubMed]
  • Mukherjee K, Bürglin TR. Comprehensive analysis of animal TALE homeobox genes: new conserved motifs and cases of accelerated evolution. J. Mol. Evol. 2007;65:137–153. [PubMed]
  • Nakano T, Windrem M, Zappavigna V, Goldman SA. Identification of a conserved 125 base-pair Hb9 enhancer that specifies gene expression to spinal motor neurons. Dev. Biol. 2005;283:474–485. [PubMed]
  • Neuteboom ST, Peltenburg LT, van Dijk MA, Murre C. The hexapeptide LFPWMR in Hoxb-8 is required for cooperative DNA binding with Pbx1 and Pbx2 proteins. Proc. Natl. Acad. Sci. USA. 1995;92:9166–9170. [PMC free article] [PubMed]
  • Noro B, Culi J, McKay DJ, Zhang W, Mann RS. Distinct functions of homeodomain-containing and homeodomain-less isoforms encoded by homothorax. Genes Dev. 2006;20:1636–1650. [PMC free article] [PubMed]
  • Noyes MB, Christensen RG, Wakabayashi A, Stormo GD, Brodsky MH, Wolfe SA. Analysis of homeodomain specificities allows the family-wide prediction of preferred recognition sites. Cell. 2008;133:1277–1289. [PMC free article] [PubMed]
  • O'Connor MB, Binari R, Perkins LA, Bender W. Alternative RNA products from the Ultrabithorax domain of the bithorax complex. EMBO J. 1988;7:435–445. [PMC free article] [PubMed]
  • Pan L, Glenn ST, Jones CA, Gross KW. Activation of the rat renin promoter by HOXD10.PBX1b.PREP1, Ets-1, and the intracellular domain of notch. J. Biol. Chem. 2005;280:20860–20866. [PubMed]
  • Pan L, Xie Y, Black TA, Jones CA, Pruitt SC, Gross KW. An Abd-B class HOX.PBX recognition sequence is required for expression from the mouse Ren-1c gene. J. Biol. Chem. 2001;276:32489–32494. [PubMed]
  • Passner JM, Ryoo HD, Shen L, Mann RS, Aggarwal AK. Structure of a DNA-bound Ultrabithorax-Extradenticle homeodomain complex. Nature. 1999;397:714–719. [PubMed]
  • Pearson JC, Lemons D, McGinnis W. Modulating Hox gene functions during animal body patterning. Nat. Rev. Genet. 2005;6:893–904. [PubMed]
  • Pederson JA, LaFollette JW, Gross C, Veraksa A, McGinnis W, Mahaffey JW. Regulation by homeoproteins: a comparison of deformed-responsive elements. Genetics. 2000;156:677–686. [PMC free article] [PubMed]
  • Penkov D, Tanaka S, Di Rocco G, Berthelsen J, Blasi F, Ramirez F. Cooperative interactions between PBX, PREP, and HOX proteins modulate the activity of the alpha 2(V) collagen (COL5A2) promoter. J. Biol. Chem. 2000;275:16681–16689. [PubMed]
  • Phelan ML, Rambaldi I, Featherstone MS. Cooperative interactions between HOX and PBX proteins mediated by a conserved peptide motif. Mol. Cell Biol. 1995;15:3989–3997. [PMC free article] [PubMed]
  • Pinsonneault J, Florence B, Vaessin H, McGinnis W. A model for extradenticle function as a switch that changes HOX proteins from repressors to activators. EMBO J. 1997;16:2032–2042. [PMC free article] [PubMed]
  • Piper DE, Batchelor AH, Chang CP, Cleary ML, Wolberger C. Structure of a HoxB1-Pbx1 heterodimer bound to DNA: role of the hexapeptide and a fourth homeodomain helix in complex formation. Cell. 1999;96:587–597. [PubMed]
  • Plaza S, Prince F, Adachi Y, Punzo C, Cribbs DL, Gehring WJ. Cross-regulatory protein-protein interactions between Hox and Pax transcription factors. Proc. Natl. Acad. Sci. USA. 2008;105:13439–13444. [PMC free article] [PubMed]
  • Pöpperl H, Bienz M, Studer M, Chan SK, Aparicio S, Brenner S, Mann RS, Krumlauf R. Segmental expression of Hoxb-1 is controlled by a highly conserved autoregulatory loop dependent upon exd/pbx. Cell. 1995;81:1031–1042. [PubMed]
  • Potter CS, Peterson RL, Barth JL, Pruett ND, Jacobs DF, Kern MJ, Argraves WS, Sundberg JP, Awgulewitsch A. Evidence that the satin hair mutant gene Foxq1 is among multiple and functionally diverse regulatory targets for Hoxc13 during hair follicle differentiation. J. Biol. Chem. 2006;281:29245–29255. [PubMed]
  • Prévôt D, Voeltzel T, Birot AM, Morel AP, Rostan MC, Magaud JP, Corbo L. The leukemia-associated protein Btg1 and the p53-regulated protein Btg2 interact with the homeoprotein Hoxb9 and enhance its transcriptional activation. J. Biol. Chem. 2000;275:147–153. [PubMed]
  • Prince F, Katsuyama T, Oshima Y, Plaza S, Resendez-Perez D, Berry M, Kurata S, Gehring WJ. The YPWM motif links Antennapedia to the basal transcriptional machinery. Development. 2008;135:1669–1679. [PubMed]
  • Raman V, Martensen SA, Reisman D, Evron E, Odenwald WF, Jaffee E, Marks J, Sukumar S. Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature. 2000a;405:974–978. [PubMed]
  • Raman V, Tamori A, Vali M, Zeller K, Korz D, Sukumar S. HOXA5 regulates expression of the progesterone receptor. J. Biol. Chem. 2000b;275:26551–26555. [PubMed]
  • Rambaldi I, Kovàcs EN, Featherstone MS. A proline-rich transcriptional activation domain in murine HOXD-4 (HOX-4.2). Nucleic Acids Res. 1994;22:376–382. [PMC free article] [PubMed]
  • Randazzo FM, Cribbs DL, Kaufman TC. Rescue and regulation of proboscipedia: a homeotic gene of the Antennapedia Complex. Development. 1991;113:257–271. [PubMed]
  • Rao MV, Donoghue MJ, Merlie JP, Sanes JR. Distinct regulatory elements control muscle-specific, fiber-type-selective, and axially graded expression of a myosin light-chain gene in transgenic mice. Mol. Cell Biol. 1996;16:3909–3922. [PMC free article] [PubMed]
  • Regulski M, Harding K, Kostriken R, Karch F, Levine M, McGinnis W. Homeo box genes of the Antennapedia and bithorax complexes of Drosophila. Cell. 1985;43:71–80. [PubMed]
  • Robertson LK, Bowling DB, Mahaffey JP, Imiolczyk B, Mahaffey JW. An interactive network of zinc-finger proteins contributes to regionalization of the Drosophila embryo and establishes the domains of HOM-C protein function. Development. 2004;131:2781–2789. [PubMed]
  • Ronshaugen M, McGinnis N, McGinnis W. Hox protein mutation and macroevolution of the insect body plan. Nature. 2002;415:914–917. [PubMed]
  • Rousso DL, Gaber ZB, Wellik D, Morrisey EE, Novitch BG. Coordinated actions of the forkhead protein Foxp1 and Hox proteins in the columnar organization of spinal motor neurons. Neuron. 2008;59:226–240. [PMC free article] [PubMed]
  • Ryoo HD, Mann RS. The control of trunk Hox specificity and activity by Extradenticle. Genes Dev. 1999;13:1704–1716. [PMC free article] [PubMed]
  • Ryoo HD, Marty T, Casares F, Affolter M, Mann RS. Regulation of Hox target genes by a DNA bound Homothorax/Hox/Extradenticle complex. Development. 1999;126:5137–5148. [PubMed]
  • Saleh M, Rambaldi I, Yang XJ, Featherstone MS. Cell signaling switches HOX-PBX complexes from repressors to activators of transcription mediated by histone deacetylases and histone acetyltransferases. Mol. Cell Biol. 2000;20:8623–8633. [PMC free article] [PubMed]
  • Salsi V, Zappavigna V. Hoxd13 and Hoxa13 directly control the expression of the EphA7 Ephrin tyrosine kinase receptor in developing limbs. J. Biol. Chem. 2006;281:1992–1999. [PubMed]
  • Samad OA, Geisen MJ, Caronia G, Varlet I, Zappavigna V, Ericson J, Goridis C, Rijli FM. Integration of anteroposterior and dorsoventral regulation of Phox2b transcription in cranial motoneuron progenitors by homeodomain proteins. Development. 2004;131:4071–4083. [PubMed]
  • Serpente P, Tümpel S, Ghyselinck NB, Niederreither K, Wiedemann LM, Dollé P, Chambon P, Krumlauf R, Gould AP. Direct crossregulation between retinoic acid receptor {beta} and Hox genes during hindbrain segmentation. Development. 2005;132:503–513. [PubMed]
  • Shen W, Chrobak D, Krishnan K, Lawrence HJ, Largman C. HOXB6 protein is bound to CREB-binding protein and represses globin expression in a DNA binding-dependent, PBX interaction-independent process. J. Biol. Chem. 2004;279:39895–39904. [PubMed]
  • Shen WF, Rozenfeld S, Lawrence HJ, Largman C. The Abd-B-like Hox homeodomain proteins can be subdivided by the ability to form complexes with Pbx1a on a novel DNA target. J. Biol. Chem. 1997;272:8198–81206. [PubMed]
  • Shi X, Bai S, Li L, Cao X. Hoxa-9 represses transforming growth factor-beta-induced osteopontin gene transcription. J. Biol. Chem. 2001;276:850–855. [PubMed]
  • Shi X, Yang X, Chen D, Chang Z, Cao X. Smad1 interacts with homeobox DNA-binding proteins in bone morphogenetic protein signaling. J. Biol. Chem. 1999;274:13711–13717. [PubMed]
  • Stevens KE, Mann RS. A balance between two nuclear localization sequences and a nuclear export sequence governs extradenticle subcellular localization. Genetics. 2007;175:1625–1636. [PMC free article] [PubMed]
  • Streit A, Kohler R, Marty T, Belfiore M, Takacs-Vellai K, Vigano MA, Schnabel R, Affolter M, Müller F. Conserved regulation of the Caenorhabditis elegans labial/Hox1 gene ceh-13. Dev. Biol. 2002;242:96–108. [PubMed]
  • Stultz BG, Jackson DG, Mortin MA, Yang X, Beachy PA, Hursh DA. Transcriptional activation by extradenticle in the Drosophila visceral mesoderm. Dev. Biol. 2006;290:482–494. [PubMed]
  • Sun B, Hursh DA, Jackson D, Beachy PA. Ultrabithorax protein is necessary but not sufficient for full activation of decapentaplegic expression in the visceral mesoderm. EMBO J. 1995;14:520–535. [PMC free article] [PubMed]
  • Taghli-Lamallem O, Gallet A, Leroy F, Malapert P, Vola C, Kerridge S, Fasano L. Direct interaction between Teashirt and Sex combs reduced proteins, via Tsh's acidic domain, is essential for specifying the identity of the prothorax in Drosophila. Dev. Biol. 2007;307:142–151. [PubMed]
  • Taghli-Lamallem O, Hsia C, Ronshaugen M, McGinnis W. Context-dependent regulation of Hox protein functions by CK2 phosphorylation sites. Dev. Genes Evol. 2008;218:321–332. [PMC free article] [PubMed]
  • Tour E. Evolutionarily conserved domains required for activation and repression functions of the Drosophila Hox protein Ultrabithorax. Development. 2005;132:5271–5281. [PubMed]
  • Troy PJ, Daftary GS, Bagot CN, Taylor HS. Transcriptional repression of peri-implantation EMX2 expression in mammalian reproduction by HOXA10. Mol. Cell Biol. 2003;23:1–13. [PMC free article] [PubMed]
  • Tümpel S, Cambronero F, Ferretti E, Blasi F, Wiedemann LM, Krumlauf R. Expression of Hoxa2 in rhombomere 4 is regulated by a conserved cross-regulatory mechanism dependent upon Hoxb1. Dev. Biol. 2007;302:646–660. [PubMed]
  • Vachon G, Cohen B, Pfeifle C, McGuffin ME, Botas J, Cohen SM. Homeotic genes of the Bithorax complex repress limb development in the abdomen of the Drosophila embryo through the target gene Distal-less. Cell. 1992;71:437–450. [PubMed]
  • Viganò MA, Di Rocco G, Zappavigna V, Mavilio F. Definition of the transcriptional activation domains of three human HOX proteins depends on the DNA-binding context. Mol. Cell Biol. 1998;18:6201–6212. [PMC free article] [PubMed]
  • Vincent JP, Kassis JA, O'Farrell PH. A synthetic homeodomain binding site acts as a cell type specific, promoter specific enhancer in Drosophila embryos. EMBO J. 1990;9:2573–2578. [PMC free article] [PubMed]
  • Wagmaister JA, Miley GR, Morris CA, Gleason JE, Miller LM, Kornfeld K, Eisenmann DM. Identification of cis-regulatory elements from the C. elegans Hox gene lin-39 required for embryonic expression and for regulation by the transcription factors LIN-1, LIN-31 and LIN-39. Dev. Biol. 2006;297:550–565. [PubMed]
  • Wagner K, Mincheva A, Korn B, Lichter P, Popperl H. Pbx4, a new Pbx family member on mouse chromosome 8, is expressed during spermatogenesis. Mech. Dev. 2001;103:127–131. [PubMed]
  • Walsh CM, Carroll SB. Collaboration between Smads and a Hox protein in target gene repression. Development. 2007;134:3585–3592. [PubMed]
  • Wang N, Kim HG, Cotta CV, Wan M, Tang Y, Klug CA, Cao X. TGFbeta/BMP inhibits the bone marrow transformation capability of Hoxa9 by repressing its DNA-binding ability. EMBO J. 2006;25:1469–1480. [PMC free article] [PubMed]
  • Weatherbee SD, Carroll SB. Selector genes and limb identity in arthropods and vertebrates. Cell. 1999;97:283–286. [PubMed]
  • Williams TM, Selegue JE, Werner T, Gompel N, Kopp A, Carroll SB. The regulation and evolution of a genetic switch controlling sexually dimorphic traits in Drosophila. Cell. 2008;134:610–623. [PMC free article] [PubMed]
  • Williams TM, Williams ME, Heaton JH, Gelehrter TD, Innis JW. Group 13 HOX proteins interact with the MH2 domain of R-Smads and modulate Smad transcriptional activation functions independent of HOX DNA-binding capability. Nucleic Acids Res. 2005;33:4475–4484. [PMC free article] [PubMed]
  • Wu Y, Moser M, Bautch VL, Patterson C. HoxB5 is an upstream transcriptional switch for differentiation of the vascular endothelium from precursor cells. Mol. Cell Biol. 2003;23:5680–5691. [PMC free article] [PubMed]
  • Yao LC, Liaw GJ, Pai CY, Sun YH. A common mechanism for antenna-to-Leg transformation in Drosophila: suppression of homothorax transcription by four HOM-C genes. Dev. Biol. 1999;211:268–276. [PubMed]
  • Zeng C, Pinsonneault J, Gellon G, McGinnis N, McGinnis W. Deformed protein binding sites and cofactor binding sites are required for the function of a small segment-specific regulatory element in Drosophila embryos. EMBO J. 1994;13:2362–2377. [PMC free article] [PubMed]
  • Zhao JJ, Lazzarini RA, Pick L. Functional dissection of the mouse Hox-a5 gene. EMBO J. 1996;15:1313–1322. [PMC free article] [PubMed]
  • Zhao Y, Potter SS. Functional specificity of the Hoxa13 homeobox. Development. 2001;128:3197–3207. [PubMed]
  • Zhao Y, Potter SS. Functional comparison of the Hoxa 4, Hoxa 10, and Hoxa 11 homeoboxes. Dev. Biol. 2002;244:21–36. [PubMed]
  • Zhou B, Chen L, Wu X, Wang J, Yin Y, Zhu G. MH1 domain of SMAD4 binds N-terminal residues of the homeodomain of Hoxc9. Biochim. Biophys. Acta. 2008;1784:747–752. [PubMed]
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...