By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
Hox (homeotic selector) gene complexes are thought to act as metagenes. A whole Hox complex can potentially function to pattern a body axis. One Hox gene cannot pattern an axis. This coordinated function depends on collinearity, where the Hox genes' properties coordinate with their 3′ to 5′ order in a chromosomal complex (Mainguy et al., 2007; Duboule, 2007). The best known collinear property is spatial collinearity, where the Hox genes' spatial expression patterns are ordered collinearly along an axis but there are other collinear properties. An important one is posterior prevalence, identified by E. B. Lewis in Drosophila (Lewis, 1978, 1995).
Lewis, who discovered Hox collinearity, proposed that BX-C (the Bithorax Complex: a Hox complex containing the posterior half of the Drosophila Hox cluster) contains genes that, in the course of evolution, have arisen by tandem duplications. The BX-C genes specify the identities of the posterior Drosophila thoracic and abdominal segments from T3 to A8 (+A9). Lewis originally proposed that one gene function is required for each segment but it has since emerged that BX-C contains only three genes that are expressed in a spatially collinear sequence, and that each specifies multiple segments. Lewis's findings fit the idea that posterior BX-C Hox genes are dominant to anterior ones: the property now known as posterior prevalence. It has generally been assumed, by Drosophilists, that this property acts postexpression, to enable dominance among coexpressed Hox genes. In fact, it acts at multiple levels and regulates transcription and mRNA stability of Hox genes, as well as regulating their function postranscriptionally and posttranslationally (Hafen et al., 1984b; Struhl and White, 1985; Beachy et al., 1988; Miller et al., 2001; Plaza et al., 2008; Yekta et al., 2004; Woltering and Durston, 2008). It is thus potentially capable of regulating the spatial collinearity of Drosophila Hox genes, but it is not, in fact, the primary regulator. Until recently, it was thought that Drosophila shows global posterior prevalence but this is incorrect. (section 2). The only animals that show global posterior prevalence so far are the vertebrates. We suspect this is because posterior prevalence mediates vertebrate temporal collinearity (section 4).
DROSOPHILA HOX PREVALENCE: AN ANCESTRAL STRATEGY?
Gehring et al. (2009) recently challenged the idea that Drosophila manifests global posterior prevalence. Posterior Drosophila Hox genes are posteriorly prevalent (Lewis, 1995) but Gehring and associates (2009) and the Kaufman group (see references in Table 1) showed that anterior Drosophila Hox genes are anteriorly prevalent so that loss-of-function mutations lead towards Antennapedia and gain-of-function mutations lead away from it (Table 1 and references therein; Fig. 1). What is the explanation of this situation? Gehring et al. (2009) argues that the Drosophila Hox cluster developed anterior and posterior prevalence because thoracic segment T2 and Antennapedia represent an ancestral ground state and its associated ur-Hox gene, respectively. Moreover, Gehring et al. (2009) argue that Hox collinearity evolved by evolution of anterior and posterior Hox genes from this ancestral gene by tandem duplication and sequential modification. There have been other interpretations: for example, Hox loss-of-function mutations lead to a Hox-less ground state, causing formation of antennae (Struhl, 1981; Percival-Smith et al., 1997); that the ur-Hox gene is Labial, not Antennapedia (Tihanyi et al., 2010), or that Drosophila actually does show global posterior prevalence (Duboule and Morata, 1994). Gehring et al.'s (2009) model fits the facts well.
There is anterior and posterior prevalence, centred around the Ur-Hox ground state represented by T2 and Antennapedia. The anterior genes Labial, Proboscipedia, Deformed, and Sex Combs Reduced are dominant to Antennapedia and so are the posterior genes Utrabithorax, Abdominal A, and Abdominal B. This situation presumably represents the ancestral state for metazoan prevalence.
What is the evolutionary logic to Gehring and coworkers' (2009) argument and why would Drosophila not need posterior prevalence? The reason ur-Hox relates to prevalence in Drosophila and other animals is unknown. Perhaps because each newly evolved Hox gene needs to be prevalent to be able to exert its new function. It can be assumed that ur-Hox-based prevalence is the ancestral condition. The global posterior prevalence that is uniquely detected in vertebrates (below) appears to be specific to them and is thus new in evolution. It supersedes ur-Hox-based prevalence. Posterior prevalence is presumably required for a specific feature of vertebrate development. We think this feature is the vertebrate form of temporal collinearity. See below.
Drosophila Hox collinearity is disorganised and disintegrating and not fully functional. For example: the Drosophila Hox complex is split into two parts, the Antennapedia and Bithorax complexes, and atypical Hox genes that are irrelevant for A-P axial (patterning is inserted into the Antennapedia complex; see Duboule, 2007; Durston et al., 2011). It, therefore, comes as no surprise that the spatial collinearity manifested by Drosophila's Hox cluster is primarily imposed by external upstream regulators, the spatially expressed gap genes and segmentation genes among others (Kehle et al., 1998; Mito et al., 2006; Rusch and Kaufman, 2000; Jack et al., 1988; White and Lehmann, 1986; Harding and Levine, 1988; Akam, 1987; Ingham and Martinez-Arias, 1986). Hox interactions also play a regulatory role here but they are not the primary spatial cues. We conclude that a Drosophila Hox cluster is not a fully functional metagene. It has no internal mechanism to fully mediate spatial collinearity and no posterior prevalence. It retains what is possibly a remnant of the ancestral form of prevalence. It is notable that many of the upstream regulators in Drosophila are also important in other insects, including Tribolium (which does have an intact Hox cluster) (Eckert et al., 2004; Maderspacher et al., 1998; Mito et al., 2006). Perhaps they have general importance. The fact that Tribolium, which has an intact Hox cluster, behaves like Drosophila regarding prevalence and Hox metagene function (see below), emphasizes that the connection between Hox cluster disintegration and Hox metagene function is not simply that the former prevents the latter (though it could). Rather, Hox complexes that fail to function as metagenes are free to disintegrate during evolution.
TRIBOLIUM AND CAENORHABDITIS
Do other invertebrates have posterior prevalence? Clearly, if there is an ancestral ground state/ur-Hox gene, this will be conserved in evolution and thus be available in all metazoa. The reason ur-Hox relates to prevalence in Drosophila is unknown. Perhaps this is because each newly evolved Hox gene needs to be prevalent in order to be able to exert its new function (see above) during evolution. It can be assumed that ur-Hox-based prevalence is the ancestral condition and that it may be generally important in metazoa. There are very few cases where there is data available to assess this. The cases where there is data are: insects other than Drosophila, notably the beetle Tribolium and nematode worms, especially Caenorhabditis. Both of these cases show a disorganised prevalence situation; that in Tribolium closely resembles that in Drosophila. The prevalence situation in Caenorhabditis is quite similar to that in Drosophila. See Tables 2 and 3 and Figure 1. It is notable that Tribolium has an intact Hox cluster and thus the disorganised prevalence in this insect does not depend on Hox cluster disintegration.
This is very like the situation in Drosophila, except that the ground state seems more likely to be represented by Scr. However, there can only be one metazoan ur-hox gene, so this situation needs to be resolved.
What is the spatial colinearity situation in other invertebrates? It is notable that many invertebrates have disorganised and/or disintegrating Hox complexes. The disintegration can be more advanced than in Drosophila. For example, Oikopleura has dispersed Hox complexes, where most of the Hox genes are separate (Seo et al., 2004, Monteiro and Ferrier, 2006). These organisms with disintegrating Hox complexes generally maintain spatial collinearity (in the sense that ancestrally anterior hox genes are expressed anteriorly and ancestrally progressively more posterior Hox genes progressively more posteriorly). In some cases, there is a scrambled Hox cluster (as in sea urchins). Here, ancestrally anterior and posterior Hox genes are mispositioned in a scrambled cluster. In the case of disorganised or dispersed clusters, spatial collinearity must depend on factors external to the Hox complex. We should suspect an upstream hierarchy equivalent to that in Drosophila. Other insects probably even have a very similar hierarchy. There is evidence for conservation of this hierarchy in other insects including the beetle Tribolium, which has an intact Hox cluster but nonetheless presumably uses this external hierarchy rather than acting as a metagene (Maderspacher et al., 1998; Eckert et al., 2004; Mito et al., 2006). It is thought that Hox collinearity evolved by tandem duplication of an ancestral ur-Hox gene and sequential modifications of the duplicates as above (Lewis, 1978, 1995; Gehring et al., 2009). If this is true, the clustered format of Hox genes will be ancestral, and dispersed Hox genes, as in dispersed and split clusters, will be derived forms (Monteiro and Ferrier, 2006).
POSTERIOR PREVALENCE IS UNIQUE TO VERTEBRATES: IS TIME OF THE ESSENCE?
Recent functional analysis shows that vertebrates so far uniquely manifest strict global posterior prevalence along the whole Hox-expressing part of the body axis. Each of the vertebrate Hox paralogue groups (pgs) except pg 12 and 13 has now been inactivated (Table 4, Fig. 1). In each case, this gives an anteriorising phenotype. In the case of Hox pg13, inactivation of individual genes has given anteriorising phenotypes. Nearly all vertebrate Hox gain-of-function phenotypes so far are posteriorising. Loss of functions for individual vertebrate Hox genes do not invariably follow this pattern; there are some exceptions (not surprising in the mass of data generated by 39 Hox genes). It is clear that loss of function for whole paralogue groups is the most significant observable for early developmental functions, where the different Hox complexes are clearly regulated in concert and Hox paralogues presumably act in concert. Hox genes are multifunctional and the few exceptions to posterior prevalence presumably reflect individual later functions of particular Hox genes. We note that Gehring et al. (2009) have proposed that “Experiments in the mouse also suggest that the ground state is a thoracic segment.” Our conclusions above argue that there is at any rate no similar connection between ur-Hox and prevalence in vertebrates as in Drosophila. It has also been suggested previously that vertebrates manifest posterior prevalence but the evidence was weak. This study was made before any of the vertebrate Hox paralogue groups had been inactivated, on the basis of homeosis caused by loss of function for individual Hox genes from only two of the Hox paralogue groups that show anterior prevalence in Drosophila. The same study proposed incorrectly that Drosophila manifests conserved posterior prevalence (Duboule and Morata, 1994). The global posterior prevalence in vertebrates is specific to them and is thus new in evolution. It supersedes ur-Hox-based prevalence. Why do vertebrates uniquely manifest posterior prevalence? We think that this is presumably required for a specific feature of vertebrate development. We think this feature is the vertebrate form of temporal collinearity (see text that begins with the final paragraph in the first column of the following page).
Shown is global posterior prevalence. We include all of the experiments showing loss of function for whole paralogue groups (designated in the table as PG, e.g., PG1 is paralogue group 1: labial) and loss of function for single Hox genes if this adds information.
Vertebrate Hox complex organisation is unique. These are the most compact Hox complexes known. Vertebrate Hox complexes are intact. The Hox genes are all transcribed in the same direction. Also, uniquely, these Hox complexes are duplicated. In most tetrapod vertebrates, there are 4 copies: A, B, C, D, via 2 genome duplications. There are thus up to 4 paralogues (copies) of each Hox gene number. In teleost fishes and a few other vertebrates, there are up to 8 copies, due to 3 genome duplications.
Vertebrate Hox genes have a second unusual collinearity feature. Hox temporal collinearity (see text below). Is this connected to posterior prevalence?
There is a key element that integrates the whole vertebrate anterior-posterior (A-P) pattern. This is time. The vertebrate A-P axis is a time axis. The head is made first and more and more posterior levels are made at later and later stages (Vasiliauskas and Stern, 2001). A central aspect of this timing is Hox temporal collinearity. The most 3′ anterior Hox genes are expressed first and progressively more 5′ posterior Hox genes are expressed progressively later. This temporally collinear sequence is used to generate the primary spatial A-P Hox pattern by time space translation and to integrate all of the successive derived patterns (Wacker et al., 2004; Durston et al., 2010). This is thus a different situation than in Drosophila, where the primary pattern guiding Hox spatial collinearity is generated externally, by the gap and segmentation genes. Posterior vertebrate Hox genes thus start expression later than anterior vertebrate Hox genes and will need posterior prevalence to exert their functions against a background of more anterior Hox gene expression.
Posterior prevalence may do more than this. It may be required for generating vertebrate temporal collinearity itself and, therefore, for generating the spatially collinear expression sequence of the vertebrate Hox genes. The primary vertebrate A-P axial pattern begins to be generated initially early in development, during gastrulation. No specific A-P regulators are known that operate before this stage. This patterning process has only been studied extensively in Amphibia (Slack and Tannahill, 1992). The involuting gastrula mesoderm acquires patterning information first and this mesoderm then copies information to an adjacent tissue layer, the overlying neurectoderm, which develops a clear A-P pattern by the end of gastrulation. This copying process is called vertical signalling (Mangold, 1933; Wacker at al., 2004; Durston et al., 2010). The mesoderm thus copies A-P information onto the neurectoderm. This vertical signalling process has been dealt with in other reviews (Durston et al., 2010, 2011). Initial Amphibian A-P patterning correlates with the initial gastrula stage expression of the Hox genes. In the Amphibian Xenopus, there is a temporally collinear sequence of Hox gene expression in the gastrula mesoderm during gastrulation and a spatially collinear pattern develops in the neurectoderm by the end of gastrulation (Wacker et al., 2004). There is evidence that the mesoderm's temporal collinearity is used to generate the primary Hox pattern of the embryo by time space translation. Each successive combination of Hox mRNA's expressed in the temporally collinear mesodermal Hox gene expression sequence involutes into the gastrula at a specific time and appears to be copied to generate an identical combination at the correct place in the neurectoderm's spatially collinear Hox pattern. It has been shown that this signalling process is A to P time dependent (Wacker et al., 2004; Durston et al., 2010). This timed patterning process presumably continues after gastrulation, because the gastrulation process continues in the chordaneural hinge in the Xenopus tailbud (Gont et al., 1993) and because Hox9 are the most posterior Hox genes expressed before the end of gastrulation The rest start later. The situation in other vertebrates is presumably comparable (Gaunt and Strachan, 1996; Alexandre et al., 1996; Deschamps et al., 1999).
The temporal collinearity of the vertebrate Hox complexes (Duboule, 2007) is thus used to generate the spatially collinear vertebrate axial Hox pattern, This is different from the situation in Drosophila where there is a requirement for the intervention of an external upstream patterning mechanism. The temporally collinear Hox sequence in the vertebrate gastrula mesoderm integrates and synchronises the expression of Hox genes from all the different Hox complexes (Durston et al., 2011). Temporal collinearity, therefore, clearly involves trans-interactions and, since we are dealing with multicellular tissues, cell interactions. We suspect that Hox genes are involved in these trans and intercellular interactions because they are available in the gastrula at exactly the right times and places. It is also known that the Hox genes undergo cell-autonomous and non-cell-autonomous activating and repressive interactions in the gastrula and in early development. These include posterior prevalence (Hooiveld et al., 1999; Woltering and Durston, 2008; Durston et al., 2011; Iimura and Pourquié, 2006) and anterior to posterior activation (Hooiveld et al., 1999; McNulty et al., 2005; In Der Rieden et al., 2010; Durston et al., 2011). We suspect that these interactions are central in generating vertebrate temporal collinearity. Posterior prevalence downregulates anterior Hox genes once more posterior Hox genes are expressed. Anterior-to-posterior activation means that anterior Hox genes activate expression of more posterior ones (Hooiveld et al., 1999). It has also been shown that loss of function for anterior Hox genes can downregulate more posterior ones. (McNulty et al., 2005). This may be the primary reason that vertebrates have correlated posterior prevalence and temporal collinearity, while Drosophila does not (Durston et al., 2011). Vertebrate Hox complexes are clearly fully functional metagenes. The issue of temporal collinearity is complex. Partial temporal collinearity also occurs in a few invertebrates. Our discussion here specifically addresses temporal collinearity in vertebrates only; we can not comment on these other cases.
We have examined the role of prevalence in Hox collinearity and the functioning of Hox complexes as metagenes.
1E.B. Lewis's pioneering work revealed Hox functional collinearity and posterior prevalence. Lewis already realised that Hox collinearity must have an evolutionary explanation.
2W. Gehring realised that Drosophila actually has ur-Hox-based prevalence. It does not have global posterior prevalence. It appears that Drosophila, an organism with an incompletely functional Hox metagene, may have retained traces of an ancestral prevalence situation.
3Not many other invertebrates have enough functional data to enable an examination of prevalence. Tribolium and Caenorhabditis do. These both have disorganised prevalence. Tribolium prevalence strikingly resembles that in Drosophila.
4Vertebrates uniquely have global posterior prevalence. They also have a time-dependent A-P axis that is set up via Hox temporal collinearity. We suspect that posterior prevalence partly mediates vertebrate Hox temporal collinearity. This leads to spatial collinearity and axial patterning via time space translation. Vertebrates have fully functional Hox metagenes.
Antennapedia complex: The anterior half of the Drosophila Hox complex. This is a separate complex, isolated from the Bithorax complex. It contains the Hox genes: Labial (most anterior), Proboscipedia, Deformed, Sex Combs Reduced, Antennapedia.
Bithorax complex: The posterior half of the Drosophila Hox complex. This is isolated from the anterior half (Antennapedia complex) and contains 3 genes. Ultrabithorax (most anterior), Abdominal A, Abdominal B (most posterior).
Collinearity:Hox genes are contained in chromosal complexes of up to 14 closely related genes. These complexes generally display collinearity, where the Hox genes' properties coordinate with their 3′ to 5′ order in the chromosomal complex (Mainguy et al., 2007; Duboule, 2007). The best known collinear property is spatial collinearity, where the Hox genes' spatial expression patterns are ordered collinearly along an axis but there are other collinear properties. Important ones are: posterior prevalence, identified by E. B. Lewis in Drosophila (Lewis, 1978, 1995) and temporal collinearity.
Hox genes: Homeotic selector genes are a subfamily of genes concerned with specifying particular regions of the anterior-posterior (A-P) body axis (among other functions). Hox genes encode a subfamily of transcription factors, each containing its own variant of a conserved DNA-binding domain, the homeodomain.
Paralogue groups: Vertebrates have uniquely undergone at least 2, and in the case of teleost fish and some other vertebrates 3, genome duplications since the evolution of this subphylum. This has led to formation of multiple copies (at least 4: A, B, C, D, sometimes 8) of the Hox complex. These different copies are generally equivalent to each other and each contains a copy of most of the Hox genes. They are called paralogous Hox complexes. The copies of each particular Hox gene in the different complexes generally have similar or identical functions and are called paralogues, or, collectively, a paralogue group.
Posterior Prevalence: A collinearity property discovered by E.B. Lewis. Posterior Hox genes dominate anterior ones collinearly.
Temporal Collinearity: The time of first expression of Hox genes in development or during a developmental event is collinear.
Time Space Translation: Each successive combination of Hox mRNA's expressed in the vertebrate temporally collinear mesodermal Hox gene expression sequence involutes into the gastrula at a specific time and appears to be copied to generate an identical combination at the correct place in the neurectoderm's spatially collinear Hox pattern. It has been shown that this signalling process is A to P time dependent.
Ur-Hox Gene: The ancestral Hox gene from which all others evolved by tandem duplication and sequential modification or by genome duplications in vertebrates.
Vertical Signalling: The signalling process whereby gastrula mesoderm copies positional information onto neurectoderm during gastrulation.