SEARCH

SEARCH BY CITATION

Keywords:

  • segmentation;
  • somitogenesis;
  • evolution of development;
  • pair rule;
  • embryo;
  • Notch signaling

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

A fundamental characteristic of the arthropod body plan is its organization in metameric units along the anterior–posterior axis. The segmental organization is laid down during early embryogenesis. Our view on arthropod segmentation is still strongly influenced by the huge amount of data available from the fruit fly Drosophila melanogaster (the Drosophila paradigm). However, the simultaneous formation of the segments in Drosophila is a derived mode of segmentation. Successive terminal addition of segments from a posteriorly localized presegmental zone is the ancestral mode of arthropod segmentation. This review focuses on the evolutionary conservation and divergence of the genetic mechanisms of segmentation within arthropods. The more downstream levels of the segmentation gene network (e.g., segment polarity genes) appear to be more conserved than the more upstream levels (gap genes, Notch/Delta signaling). Surprisingly, the basally branched arthropod groups also show similarities to mechanisms used in vertebrate somitogenesis. Furthermore, it has become clear that the activation of pair rule gene orthologs is a key step in the segmentation of all arthropods. Important findings of conserved and diverged aspects of segmentation from the last few years now allow us to draw an evolutionary scenario on how the mechanisms of segmentation could have evolved and led to the present mechanisms seen in various insect groups including dipterans like Drosophila. Developmental Dynamics 236:1379–1391, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

A segmental organization of the anterior–posterior body axis is found in three major animal phyla: the arthropods, the annelids, and the vertebrates. This is only a small fraction of the 30 to 35 metazoan phyla that are recognized (Nielsen,2001). Yet, these groups are among the most successful and dominating. For example, arthropods are found in virtually every available habitat on earth. The repetitive building blocks allow flexibility with respect to adaptation to new tasks and may be of benefit for morphological and functional evolution. In current phylogenies, these three groups with segmented animals are not closely related; each of these three taxa belongs to a separate branch in the phylogenetic tree of the bilaterian animals (Fig. 1) (e.g., Aguinaldo et al.,1997, Adoutte et al.,1999, Telford2006). The vertebrates are chordates that belong to the deuterostomes, while the annelids belong to the lophotrochozoans, and the arthropods belong to the ecdysozoans (Aguinaldo et al.,1997). Each of these three phyla is more closely related to phyla with representatives that do not have a segmented body than they are to each other. Thus a key question is whether segmentation has evolved independently in the segmented phyla (arthropods, annelids, vertebrates), or is based on a common origin and dates back to their last common ancestor, the urbilaterian (Fig. 1). A common evolutionary origin for segmentation would suggest the use of common genetic programs for segmentation in the different phyla. However, before one can compare different phyla with each other, one first has to define what features of segmentation within each of these groups are ancestral features and which are derived. The present review focuses on segmentation within the arthropods and tries to define the conserved (and probably ancestral) aspects as well as the derived aspects of the mechanisms of arthropod segmen tation. This framework is then used to discuss the evolution of segmentation mechanisms and the origin of segmentation.

thumbnail image

Figure 1. Relationships of animal phyla. The diplobasts are in black, deuterostomes in green, lophotrochozoans in blue, and ecdysozoans in red. Segmented phyla are in blue boxes. Please note that not all phyla are included, but this does not affect the subdivision of the bilaterians into deuterostomes, lophotrochozoans, and ecdysozoans.

Download figure to PowerPoint

THE DROSOPHILA PARADIGM

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

The segmentation gene cascade that has been discovered in Drosophila still forms a paradigm for segmentation in arthropods. In Drosophila, a hierarchic gene cascade subsequentially subdivides the embryo into smaller units, eventually resulting in the repetitive pattern of the segments along the anterior–posterior axis (Nüsslein-Volhard and Wieschaus1980; St. Johnston and Nüsslein-Volhard1992; Ingham,1988; Pankratz and Jäckle,1993). There are several tiers in this gene cascade. Maternally provided gene products, like the bicoid and nanos gene products, activate a class of zygotically expressed genes, the gap genes. The first sign of a repetitive pattern then becomes visible after the activation of the next class of genes, the pair rule genes. They are expressed in a repetitive pattern of seven transverse stripes along the anterior–posterior axis. The final tier of genes that are turned on in the Drosophila segmentation gene cascade are the segment-polarity genes that are each expressed in 14 transverse segmental stripes. The segment-polarity genes delineate the boundaries of the parasegments, the embryonic segmental units, and define the anterior–posterior polarity of these parasegments. Segment-polarity genes are also involved in the generation of the intrasegmental patterning of the segments (Sanson,2001). Parallel to the segment-polarity genes, the Hox genes that specify the identity of the different segments are activated (Carroll,1995).

The huge amount of information on this hierarchic segmentation gene cascade in Drosophila still strongly influences our understanding of segmentation in arthropods. The arthropods consist of four extant groups: insects, crustaceans, myriapods (e.g., centipedes, millipedes), and chelicerates (e.g., spiders, mites, ticks, scorpions). The crustaceans and insects are sister groups, while the chelicerates are the most basally branched group (Fig. 2) (e.g., Friedrich and Tautz1995; Blaxter2001; Hwang et al.,2001; Giribet et al.,2001). The phylogenetic position of the myriapods is debated but most data currently suggest that the myriapods are a sister-group of the insect/crustacean clade and not of the chelicerates (e.g., Harzsch et al.,2004; Giribet et al.,2001).

thumbnail image

Figure 2. Relationships of major arthropod groups. Please note that the position of the myriapods is debated, and some authors place the myriapods as sister-group to the chelicerates. “Sh” and “Lo” indicate whether the representatives of these groups use short germ (Sh) or long germ (Lo) segmentation.

Download figure to PowerPoint

The way segments develop in Drosophila is rather untypical and derived. In contrast to most other arthropods, the segments in Drosophila basically all form at the same time at the syncytial blastoderm stage. In most arthropods—chelicerates, myriapods, crustaceans; and also most insects—however, the posterior segments are sequentially added from a posteriorly localized unsegmented celluarilized tissue, usually referred to as “growth zone.” The segmentation mode of Drosophila is called long germ development and segmentation via sequential addition of segments is called short germ segmentation. The latter represents the ancestral condition for arthropods (Tautz et al.,1994; Davis and Patel,2002).

The syncytial state of the early Drosophila embryo allows diffusion of products of segmentation genes that mostly code for transcription factors. This diffusion results in the formation of gradients, which are considered to be an important feature of Drosophila segmentation. But such diffusion of transcription factors is impracticable in a cellular environment; therefore, it is unlikely that a similar mechanism acts during segmentation in the cellularized growth zone of short germ arthropods. This raised two main questions: (1) To what extent is the Drosophila segmentation gene cascade an adaptation to this special condition? (2) What genetic mechanisms are used to pattern the segments in the growth zone of short germ arthropods? To answer such questions, we have to analyze the segmentation mechanisms in other arthropods to define which parts are conserved and which parts are derived and to what extent.

During the last decade, it became clear that despite these differences of segment formation, many of the Drosophila segmentation genes also play a role during segmentation in short germ insects and arthropods. The following sections discuss what is known about the different classes of genes and how they act in the different arthropod groups, and try to define which aspects might be ancestral aspects for arthropods and which are the derived aspects for particular groups.

CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

The role of the lowest level of the segmentation gene cascade, the segment-polarity gene network, appears to be the most conserved. The segment-polarity genes define the parasegmental boundaries and are required to maintain these boundaries. Parasegments are the fundamental genetic units in the arthropod embryo that are defined by functional compartment boundaries (Martinez-Arias and Lawrence,1985; Lawrence,1988; Patel,1994), but parasegments are out of phase compared to the later segments (Deutsch,2004). In the segmentation process of the embryo, the parasegments are the functional developmental entities and not the segments. This becomes obvious from the fact that key regulatory genes, like the Hox genes, obey functional boundaries that correspond to parasegment boundaries (Martinez-Arias and Lawrence,1985; Lawrence,1988; Damen,2002) and that the cells on either side of the boundary do not mix (Deutsch,2004). The initial organization in parasegments is observed in all arthropod groups, which means that the parasegment is an evolutionary conserved genetic entity among arthropods (Damen,2002; Hughes and Kaufman,2002b).

The boundaries of the parasegments are maintained by cross-regulatory interactions of the segment polarity genes that consist of both signaling molecules [e.g., wingless (wg), hedgehog (hh)] and transcription factors [e.g., engrailed (en), cubitus interruptus (ci), gooseberry (gsb)]. These genes are expressed in segmentally reiterated stripes (Martinez Arias et al.,1988). en and hh are expressed in the anterior compartment of the parasegmental unit, while wg is in the posterior compartment. hh and wg encode signaling molecules (HH and WG) that are transmitted to the other side of the boundary and in this way define and maintain mutually exclusive cell populations. This autoregulatory loop maintains the boundary in between these cell populations.

Functional data for the role of segment-polarity genes in defining the parasegment boundaries are only known from insects (DiNardo et al.,1988; Oppenheimer et al.,1999), but the expression patterns of segment-polarity genes in representatives of all four extant arthropod groups (insects, crustaceans, myriapods, chelicerates) are highly conserved, which strongly suggests their functions are conserved (Patel et al.,1989; Scholtz et al.,1994; Nagy and Carroll,1994; Nulsen and Nagy,1999; Damen,2002; Hughes and Kaufman,2002b; Janssen et al.,2004,2006; Simonnet et al.,2004). The segment polarity genes are expressed in segmental stripes in all arthropods. A stripe of cells expressing wg and ci is localized adjacently just anterior to a stripe of cells that express hh and en. Evidence that the boundary between these cell populations is a functional one comes from the expression of Hox genes that follows this boundary in various arthropods (Martinez-Arias and Lawrence,1985; Lawrence,1988; Damen,2002; Hughes and Kaufmann,2002c; Schwager et al.,2007). Furthermore, limb buds abut this boundary and, at least in the spider, grooves are associated with this boundary between wg and en expressing cells, identical to Drosophila (Damen,2002). In addition, in malacostracan crustaceans parasegments (and not segments) form as genealogical units that use the segment-polarity gene network to define their boundaries (Patel,1994; Scholtz et al.,1994; Scholtz,1995; Scholtz and Dohle,1996; Dohle et al.,2004).

The lowest level of the segmentation gene cascade appears to be highly conserved and organization in parasegments is thus a conserved feature of the arthropods (Damen,2002; Hughes and Kaufman,2002b; Deutsch,2004).

THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

The pair rule genes delimit the parasegments in Drosophila and define the domains that will express the segment polarity genes (Lawrence et al.,1987; DiNardo and O'Farrell,1987; Ingham,1988; DiNardo et al.,1988; Baker,1988). Pair rule genes form a crucial tier in the Drosophila segmentation gene cascade as this tier marks the transition from a non-periodic pattern to a periodic pattern. The pair rule genes thus are the first genes that are active in a repetitive pattern. The tiers upstream of the pair rule genes in the Drosophila segmentation gene cascade, i.e., the maternal and gap genes, are active in non-periodic patterns.

In other arthropods, pair rule gene orthologs also seem to be among the first genes that are activated in periodic patterns, although there is variation in the patterns in which they appear. Despite the variations, it is becoming clear that the activation of these genes forms a crucial step in the segmentation process of all arthropods.

But let us start first with Drosophila. In this long germ insect, the pair rule genes are expressed in repetitive patterns of seven transverse stripes along the anterior–posterior axis (Pankratz and Jäckle,1993). This initial repetition is in a double segmental periodicity. The double segmental periodicity of pair rule gene patterning in Drosophila is also obvious from their loss-of-function phenotypes that display cuticle deletions in alternating segments (Nüsslein-Volhard and Wieschaus,1980). Canonical pair rule patterning, therefore, is in such a two-segmental periodicity and this seems to be conserved at least among holometabolous insects (Davis and Patel,2003). In such short germ holometabolous insects like the beetle Tribolium castaneum, the initial expression patterns of pair rule gene orthologs are also in a double segmental periodicity (Sommer and Tautz,1993; Patel et al.,1994; Choe et al.,2006). In addition, gene inactivation leads to typical pair rule phenotypes for orthologs of the pair rule genes even skipped (eve), sloppy paired (slp), and paired (prd) in Tribolium (Schröder et al.,1999; Maderspacher et al.,1998; Choe and Brown,2007). In hemimetabolous insects that branch off more basally in the phylogenetic tree of the insects, some pair rule gene orthologs have been analyzed (e.g., the cricket Gryllus bimaculatus, the grasshopper Schistocerca americana, the milkweed bug Oncopeltus fasciatus). They are expressed in stripes in the growth zone, some of them segmental like eve in Oncopeltus, others double segmental like prd in Schistocerca, and again others partly segmental and partly double segmental like eve in Gryllus (Davis et al.,2001; Liu and Kaufman,2005a; Mito et al.,.2007). RNA interference (RNAi) experiments in Gryllus showed that eve acts as a canonical pair rule gene but only in the anterior segments (Mito et al.,2007). eve in Gryllus and Oncopeltus in addition seems to have a gap function (Liu and Kaufman,2005a; Mito et al.,2007), which will be discussed below. In insects, pair rule gene orthologs thus seem to act as in a double segmental periodicity and thus as canonical pair rule genes.

In other arthropod groups, there is expression data, but no functional data yet. In chelicerates (e.g., the spider Cupiennius salei, the spider mite Tetranychus urticae) and myriapods (e.g., the centipedes Strigamia maritima, Lithobius atkinsoni, and the millipede Glomeris marginata), pair rule gene orthologs are expressed in stripes in the growth zone before overt segmentation (Damen et al.,2000;2005; Schoppmeier and Damen,2005a; Dearden et al.,2002; Hughes and Kaufman,2002a,b; Chipman et al.,2004; Janssen,2005; Davis et al.,2005). Most data are available for the spider Cupiennius where orthologs of seven pair rule genes have been studied [eve, hairy (h), runt (run), odd-skipped (odd), odd-paired (opa), prd, and slp]. All these genes are expressed in a dynamic pattern in stripes that progress from posterior to anterior in the growth zone (Fig. 3). For one of these genes, the prd ortholog Cs-pby-3, it has been demonstrated that the stripes appear in a segmental periodicity (Schoppmeier and Damen,2005a), but also the periodicity of the other pair rule orthologs seems to be segmental in the spider rather than double segmental (Damen et al.,2005). Similarly, a prd ortholog is segmentally expressed in the growth zone of another chelicerate, the spider mite Tetranychus. But, in the anterior segments of the prosoma, this Tetranychus prd ortholog Tu-Pax3/7 is initially expressed in a double segmental pattern (Dearden et al.,2002). In myriapods, a similar dynamic expression is seen for pair rule gene orthologs in the unpatterned growth zone. Orthologs of the fushi tarazu (ftz) and eve genes in the centipede Lithobius atkinsoni (Hughes and Kaufman,2002a,b) and several pair rule orthologs in the millipede Glomeris marginata (Janssen,2005) are expressed in dynamically progressing stripes in the growth zone, in similar patterns to the spider. In these two myriapods, the patterns seem to be in a segmental periodicity. However, the odd ortholog odr1 is expressed in a clear double segmental pattern in the growth zone of the geophilomorphic centipede Strigamia maritima (Chipman et al.,2004). In crustaceans (e.g., the brine shrimp Artemia franciscana), the orthologs of two pair rule genes (eve, prd) are expressed in a dynamic posterior domain in the growth zone while transiently a stripe of expression is visible just anterior this domain that seems to be associated with the formation of the segments (Copf et al.,2003; Davis et al.,2005).

thumbnail image

Figure 3. Diversity of expression of pair rule gene orthologs in arthropods. Various patterns are observed for the activation of pair rule gene orthologs in different arthropods. In Drosophila, pair rule genes are expressed in a double segmental periodicity in seven transverse stripes in the embryo. In short germ insects like Tribolium, the initial periodicity is also double segmental, but the stripes appear sequentially after each other in the growth zone; the expression of most pair rule gene orthologs ceases once the segments have formed. In the spider Cupiennius, the stripes of pair rule gene ortholog expression progress dynamically from posterior to anterior in the growth zone; also in the spider, the expression of most pair rule gene orthologs discontinues once the segments have formed. In the crustacean Artemia, stripes of pair rule gene ortholog expression split off a posterior domain of expression. Pair rule gene orthologs are among the first genes in all arthropods to be activated in a periodic pattern.

Download figure to PowerPoint

The orthologs of pair rule genes, thus, are activated in stripes (in either a single-segmental or double-segmental periodicity) in all arthropods before overt morphological segmentation and before the segment-polarity genes are expressed. An open question still is whether canonical pair rule patterning in a two segmental periodicity is a derived mechanism used in (higher) insects or whether this is also used in other arthropods (Davis and Patel,2003; Damen,2004). Functional experiments using RNAi may solve this question.

There are a few other genes that are expressed at the same time as the pair rule orthologs in repetitive patterns, but this is only seen in restricted groups of arthropods. The Delta gene is expressed in stripes in the growth zone of spiders and myriapods (Stollewerk et al.,2003; Kadner and Stollewerk,2004; Janssen,2005) (discussed in more detail below), and the caudal (cad) gene is expressed in double segmental stripes in the centipede Strigamia (Chipman et al.,2004).

In all arthropods, the orthologs of the pair rule genes thus are among the first genes that are expressed in a repetitive pattern. Despite the variations seen for the pair rule gene orthologs among different arthropods (Fig. 3), it is clear that pair rule gene orthologs are involved in the transition from a non-segmental to a segmental pattern in all arthropods and that this is a conserved aspect of arthropod segmentation. However, the differences in the patterns of these genes in various arthropods also suggest that the segmentation mechanisms upstream of the pair rule gene orthologs are more diverged.

PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

Interactions among pair rule genes are important in Drosophila segmentation (Fujioka et al.,1996; Jaynes and Fujioka,2004; Schroeder et al.,2004). The comparative studies presented above have revealed that the expression patterns of pair rule orthologs in other insects and arthropods display a wide variety of expression patterns. But what are the genetic interactions of these genes in these species? The only data on pair rule gene interactions in short germ arthropods has recently been published by Choe et al. (2006). These authors have used RNAi to analyze the interactions among pair rule genes in the beetle Tribolium. They came up with the new concept of a “pair rule circuit” that describes the interactions of pair rule genes in the short germ insect Tribolium (Fig. 4). In this pair rule circuit, three pair rule genes regulate each other. eve is required to activate run, which, in turn, is required to activate odd. odd represses the expression of eve; in this way a stripe of eve is separated from a broad expression domain in the growth zone, resulting in a repetitive pattern of eve. Due to the inactivation of eve, run and odd expression also fades. These three genes also regulate their downstream targets prd and slp (Choe et al.,2006; Choe and Brown,2007). Together they define the parasegmental boundaries and the expression domains of the segment-polarity genes. This pair rule circuit thus defines segments sequentially in a double segmental periodicity in Tribolium. It is important to note that this mechanism is not a simple application of the Drosophila pair rule gene regulation, but is a new concept. The sequential activation of the genes of the pair rule circuit is a concept that can explain periodic gene expression in cellularized tissue like the growth zone of short germ arthropods. Despite the differences seen for the expression of pair rule gene orthologs (see previous section), these expression patterns are compatible with a circuit like the Tribolium pair rule circuit.

thumbnail image

Figure 4. A pair rule circuit for segmentation. The concept of a pair rule circuit has been proposed for the beetle Tribolium castaneum (Choe et al.,2006). Repetitive patterns of pair rule gene expression are generated via a regulatory circuit of the three pair rule genes eve, run, and odd. The circuit also regulates the downstream targets slp and prd. The pair rule genes define the stripes of segment-polarity gene expression and the boundaries of the parasegments. The periodicity of the gene interactions displays a genetic hierarchy in which a circuit of primary pair rule genes controls secondary pair rule genes, while segment polarity genes are further downstream.

Download figure to PowerPoint

There is another remarkable outcome of the study of Choe et al. (2006). As is the case in Drosophila, the pair rule genes act in two different functional levels. In Tribolium, eve, run, and odd regulate each other, as well as the downstream pair rule genes slp and prd. There thus exists a functional hierarchy among the primary pair rule genes (eve, run, and odd in Tribolium) and the secondary pair rule gene (slp and prd). Also in Drosophila, primary and secondary pair rule genes have been distinguished. The Drosophila primary pair rule genes are eve, run, and h. The expression of these genes in each of their seven transverse stripes is controlled directly by different combinations of maternal and gap gene products. This activation is accomplished by separate enhancers for the various stripes in the promoters of these primary pair rule genes (e.g., Akam,1989; Pankratz and Jäckle,1993; Klingler et al.,1996). In this way, the non-periodic pattern of the maternal genes and the gap genes is transformed into the periodic pattern of these pair rule genes. The metameric pattern of the other pair rule genes in Drosophila [odd, opa, prd, fushi tarazu (ftz), tenascin-m (ten-m), slp] depends at least in part on the primary pair rule genes (Jaynes and Fujioka,2004). Therefore, these genes are lower in the hierarchy and are called secondary pair rule genes. This is also obvious in the architecture of the promoter of the ftz gene, which does not contain separate stripe elements like the h, eve, and run promoter, but a single element responsible for all seven ftz stripes (Hiromi and Gehring,1987). run and eve are primary pair rule genes in both Drosophila and Tribolium; odd, however, is a primary pair rule gene in Tribolium but a secondary pair rule gene in Drosophila. With respect to this, it is notable that a bioinformatics approach demonstrated that odd behaves like a primary pair-rule gene in the global regulatory structure of the segmentation gene network of Drosophila (Schroeder et al.,2004). However, functionally it is a secondary pair rule gene. odd thus might have changed from a primary into a secondary pair rule gene in the lineage leading to Drosophila.

In the spider Cupiennius, a presumptive hierarchy among pair rule gene orthologs has also been detected (Damen et al.,2005) as these genes appear to be activated differently in the growth zone of the spider. As segmentation progresses in an anterior to posterior way in the growth zone, the more posterior parts of the growth zone undergo events that are earlier in the segmentation process than the more anterior parts of the growth zone. One thus can assume that genes that act more posterior in the growth zone are acting earlier in the segmentation hierarchy than genes that are only active in the anterior portion of the growth zone. The spider eve, run, h, and prd orthologs are expressed in the posterior portion of the growth zone, while stripes of slp, opa, and odd expression are never observed in the posterior portion of the growth zone, but only in the anterior portion of the growth zone (Damen et al.,2005). Comparison with Drosophila and Tribolium suggests that the hierarchy of these genes in the genetic network may be at least partially conserved.

DIVERGENT UPSTREAM MECHANISMS I: GAP GENES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

The Drosophila pair rule genes are regulated by the gap genes [e.g., hunchback (hb), Krüppel (Kr), giant (gt), knirps (kni)]. Gap genes are responsible for an early subdivision of the blastoderm stage embryo (Rivera-Pomar and Jäckle,1996). Each of the gap genes is expressed in one or two broad domains corresponding to several contiguous segments. The borders of these domains are set by cross-regulation between the gap genes (Jäckle et al.,1986; Hülskamp et al.,1990). Gap gene mutations result in the deletion of consecutive segments, producing a “gap” in the embryo (Nüsslein-Volhard and Wieschaus,1980). The syncytial state of the early Drosophila embryo is considered to be important for the action of the gap genes, which all are transcription factors, and permits them to act via short-range gradients.

In short germ insects, orthologs of the gap genes hb, Kr, and gt have been analyzed (Sommer and Tautz,1993; Wollf et al.,1995; Patel et al.,2001; Bucher and Klingler,2004; Cerny et al.,2005; Liu and Kaufman,2004a,b; Mito et al.,2005,2006). These gap gene orthologs show contiguous expression domains and also the order of the gap gene ortholog expression domains along the anterior–posterior axis is conserved, but the domains appear to be localized more anteriorly compared to Drosophila (e.g., Bucher and Klingler,2004). The posterior shift of gap gene domains in Drosophila might be an adaptation to long germ development. Although the expression patterns in short germ insects as well as the two Tribolium mutants with a gap phenotype that were found in a mutagenesis screen (Maderspacher et al.,1998) suggest that the function of gap gene orthologs might be largely conserved, RNAi against the gap gene orthologs point to a more complex picture. RNAi for these gap gene orthologs in Gryllus, Tribolium, and Oncopeltus results in both loss of segments and homeotic transformations. The homeotic effects are due to changes in the patterns of the Hox genes (Cerny et al.,2005; Mito et al.,2005,2006). Furthermore, the RNAi effect expands much more to the posterior than the expression domain of these genes. In contrast, in Drosophila the gaps of deleted segments more or less correspond to the expression domains. However, these longer-range functions might be a feature that is related to sequential segmentation in short germ arthropods. A mechanism of sequential activation of gap gene orthologs has been proposed for short germ insects (Liu and Kaufman,2005b). Such a sequential activation and thus sequential dependency of gap gene orthologs on each other could explain the RNAi effects that expand posteriorly beyond the expression domains of the genes.

Gap genes directly regulate the primary pair rule genes eve, h, and run in Drosophila. Although the gap genes are expressed early in the segmentation machinery of short germ insects, their influence on pair rule gene regulation in short germ insects is still obscure. Indeed, the expression of some pair rule genes is controlled by gap gene orthologs, but this is not necessarily in the domain where the particular gap gene ortholog is expressed. For instance in Tribolium, eve stripes 2 and 3 are within the Kr expression domain, but in the Kr RNAi embryos, eve stripes 2 and 3 are not affected; but there are defects visible for eve stripes that are posterior to the Kr domain (Cerny et al.,2005). The eve stripes 4 and 5 initially seem to form normally, but the splitting into secondary segmental stripes, especially for eve stripe 5, does not take place in Kr RNAi embryos. In addition, the more posterior eve stripes do not form at all (Cerny et al.,2005). Similar effects of Kr on eve expression are seen in Gryllus RNAi embryos (Mito et al.,2006). As the effects on eve are more posterior than where Kr is expressed, these effects must be indirect effects. Furthermore, the existence of stripe-specific enhancer elements in the promoter region of the hairy gene of Tribolium is an indication that gap genes may directly regulate this gene, as in Drosophila (Eckert et al.,2004). Gap gene orthologs thus seem to regulate pair rule genes in short germ insects, but they also act beyond their expression domains, and at least in the case of Kr this regulation seems to be indirect.

Recently, the new gap gene mille-pattes (mlpt) has been discovered in Tribolium that displays a segmental gap phenotype as well as homeotic transformations (Savard et al.,2006a). The Tribolium mlpt gene does not code for a transcription factor but for a polycistronic mRNA that encodes several peptides. Also the analysis of the eve gene uncovered a gap function for eve in both Oncopeltus and Gryllus (Liu and Kaufman,2005a; Mito et al.,2007). This data suggest some gap genes might have lost their gap function during the transition to Drosophila long germ segmentation.

An important result is the discovery of the prominent role of gap gene orthologs in regulating Hox genes. Hox genes are activated along the anterior posterior body axis and specify segmental identity and the control of the Hox gene expression domains might even be the ancestral role of the gap genes (see also Peel et al.,2005). One can imagine that a system of sequential anterior to posterior activation of gap genes, as proposed by Liu and Kaufman (2005b), defines the various expression boundaries of the Hox genes. Gap genes thus have a dual role in insect; they control the formation of segments and control the Hox genes that regulate the segmental identities.

Orthologs of gap genes have hardly been investigated in non-insect arthropods. hb and Kr seem not to be involved in segmentation in the centipede Strigamia (Chipman and Stollewerk,2005). However, this study focuses on the expression of these genes in the nervous system. It is presently not possible to decide whether gap genes also act in other arthropod groups. Therefore, studies in non-insect species for which functional studies like RNAi are possible would help to resolve the ancestral role of these genes in arthropods.

DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

The progressive nature of sequential addition of segments from a growth zone in short germ arthropods demands a periodic activation of genes like the pair rule genes. The discovery that the Notch/Delta signaling pathway is involved in spider segmentation (Stollewerk et al.,2003; Schoppmeier and Damen,2005b) is an important contribution to our understanding of the mechanisms of periodic gene activation in the growth zone of short germ arthropods. Notch/Delta signaling is a cell-to-cell signaling pathway that is found in all metazoans and that plays important roles in several processes that involve cell type specification and boundary formation (Artavanis-Tsakonas et al.,1999; Lai,2004). Notch is a transmembrane receptor that is cleaved once it has bound to its ligands Delta or Serrate, which are also transmembrane proteins, but localized on neighboring cells. The intracellular portion of Notch then can allocate to the nucleus and regulate gene expression. In the spider Cupiennius, the Notch/Delta signaling pathway is required for the formation of regularly shaped segments. The Notch gene is ubiquitously expressed in the growth zone, while the Delta-1 gene is expressed in stripes that progress dynamically into an anterior direction starting from the posterior end of the growth zone (Stollewerk et al.,2003). The dynamics of Delta expression are similar to those of the pair rule gene orthologs in the spider growth zone (Damen et al.,2000,2005). Gene silencing via RNAi of components of the Notch/Delta signaling pathway like Notch and Delta as well as the Presenilin (Psn) and Suppressor of Hairless (Su(H))genes, which are required for the transmission of the signal from the Notch receptor to the nucleus, result in severe segmentation phenotypes (Stollewerk et al.,2003, Schoppmeier and Damen,2005b). Interference of the Notch/Delta signaling pathway via RNAi also affects the expression of the hairy gene. The spider ortholog of the pair rule gene hairy is normally expressed in stripes in the growth zone (Damen et al.,2000) but after interference of the Notch/Delta signaling pathway, hairy is no longer expressed in periodic stripes but in a salt-and-pepper pattern (Stollewerk et al.,2003; Schoppmeier and Damen,2005b). The organization of hairy in stripes thus depends on Notch/Delta signaling.

Apart form the spider (a chelicerate), the Notch/Delta signaling pathway presumably also acts at least in myriapod segmentation. The Delta gene is expressed in dynamical stripes in the growth zone of the myriapods Lithobius forficatus (a centipede) and Glomeris marginata (a millipede) similar to that observed in the spider (Kadner and Stollewerk,2004; Janssen,2005). Unfortunately, no functional tests are available yet for myriapod species. Notch/Delta signaling is not involved in Drosophila segmentation, and presumably neither in the short germ insect Tribolium (Tautz,2004). Presently, it is not clear whether and to what extent Notch/Delta acts in segmentation of crustaceans and other short germ insects.

The Notch–Delta signaling pathway plays an important role in the dynamic periodic expression of genes in spiders and myriapods. Intringually, both the expression and phenotypes of Notch/Delta pathway components show strong similarities to the ones in vertebrates, where Notch/Delta signaling forms a core element of the segmentation clock that patterns the segmental units of the somites in the unpatterned presomitic mesoderm of vertebrate embryos (see below).

SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

To get insights into the evolutionary steps of arthropod segmentation, it is probably better not to use Drosophila as a starting point, but the last common ancestor of all arthropods, the urarthropod. Some important findings of the last few years now allow us to propose some of the mechanisms this urarthropod could have used for segmentation and to suggest an evolutionary scenario of how the mechanisms of segmentation might have changed and have led to the present situation seen in various insect groups including dipterans like Drosophila.

To do this, it is necessary to define what are the ancestral characters for segmentation, and hypothesize what mechanisms the urarthropod might have been using to make segments. First, the urarthropod was a short germ arthropod that developed segments via terminal addition. Second, the involvement of pair rule gene orthologs in all arthropods implies that pair rule gene orthologs were involved in segmentation of this urarthropod. It seems likely that these genes controlled the formation of parasegmental units as well as the downstream network of segment-polarity genes, which maintains the compartment boundaries of these parasegments. Third, the Notch/Delta signaling pathway presumably was involved in segmentation of the urarthropod. Fourth, at this stage it remains unclear whether and to what extent other genes, like for instance gap gene orthologs, played a role in segmentation of the urarthropod. The paucity of information that is available on gap gene orthologs in non-insect arthropods to date does not directly point to a prominent role in segmentation as is the case for gap genes in insects (Chipman and Stollewerk,2005). Taken together, there is good evidence to hypothesize a prominent role for pair rule gene orthologs as well as an involvement of Notch/Delta signaling in specifying the segmental units in the urarthropod.

The next question then is how these genes might have been acting. At this stage, two recent findings might help us. First, a circuit of pair rule gene orthologs similar to the pair rule circuit that has been discovered in the beetle Tribolium (Choe et al.,2006) may explain how pair rule gene orthologs could interact with each other in the cellular environment of the growth zone, and how such a circuit can generate repetitive patterns. Second, the use of a signaling pathway, like the Notch/Delta pathway, seems to be more suitable for patterning segments from celluarilized tissue and could control such a pair rule circuit in short germ arthropods. RNAi experiments in the spider, indeed, have shown a link between Notch/Delta signaling and the pair rule gene ortholog hairy (Stollewerk et al.,2003; Schoppmeier and Damen2005b). Interactions between Notch/Delta signaling and pair rule gene orthologs are further supported by the similarities seen in the dynamic expression patterns for pair rule gene orthologs and Delta in the growth zone of the spider and the millipede (Stollewerk et al.,2003; Damen et al.,2000,2005; Janssen,2005). Pair rule gene orthologs, therefore, might have been interacting with Notch/Delta signaling in the urarthropod as they still do with at least some of them in the spider, but, in addition, they might have controlled each other, as has been shown to be the case in insects (e.g., Drosophila and Tribolium) (Fuijoka et al.,1996; Jaynes and Fuijoka,2004; Choe et al.,2006). Presently, it is unclear whether a circuit of interacting pair rule gene orthologs may also be present in other short germ arthropods like chelicerates and myriapods. Expression patterns of these genes are compatible with a circuit (e.g., Chipman et al.,2004; Damen et al.,2005; Davis et al.,2005) but so far no functional data are available to confirm this.

We now can speculate that pair rule gene orthologs were interacting with each other in the ancestral mechanism of arthropod segmentation and that they were also interacting with the Notch/Delta signaling pathway (Fig. 5). This could show similarities to the present situation in chelicerates and myriapods (although in these groups only limited data are available on the genetic interactions). In the lineage leading to the insects, such a network or circuit of pair rule gene orthologs then might have become more and more independent of the Notch/Delta signaling, while on the other hand the influence of the gap genes might have been increased (Fig. 5) (see also Peel and Akam,2003; Peel,2004). Eventually, the influence of Notch/Delta signaling on pair rule genes was entirely reduced in the lineage to the insects, and the pair rule genes were only under the control of gap genes and other pair rule genes (pair rule circuit). Such a condition again might have facilitated the transition to syncytial long germ insects where the pair rule genes are activated and interacting under syncytial conditions and no longer act in a circuit. Notch/Delta signaling is a cell–cell signaling pathway, which is unlikely to act under such syncytial conditions. Thus, once the activation of pair rule gene orthologs no longer was controlled by this cell–cell signaling pathway but solely by transcription factors (gap genes, other pair rule genes), the transition to syncytial long germ segmentation as in Drosophila was possible. In such a scenario, syncytial long germ segmentation could easily have evolved several times independently, once the Notch/Delta cell–cell signaling pathway lost its influence on segmentation and the need for a cellularized state to allow this signaling no longer existed. Indeed, long germ segmentation is found in several groups of insects (dipterans, beetles, wasps) in different branches of the phylogenetic tree of the insects (Savard et al.,2006b); for instance, within the Coleoptera (beetles) there are both long germ and short germ beetles (e.g., Patel et al.,1994). This suggests that syncytial long germ segmentation as present in the various insect groups might have evolved several times independently from short germ segmentation. The use of a cascade of transcription factors seems to be more suitable for patterning segments in a syncytium, and could be the consequence of selection for faster development.

thumbnail image

Figure 5. Model for the evolutionary transition of pair rule gene regulation in segmentation. The urbilaterian presumable used both Notch/Delta signaling and a network of pair rule genes (maybe similar to the Tribolium pair rule circuit). In the lineage leading to the insects, the influence of Notch/Delta signaling was reduced, while presumably simultaneously the gap gene orthologs gained more influence on the control of the pair rule gene orthologs. Eventually in the syncytial long germ segmentation of Drosophila, pair rule genes are under the complete control of the gap genes.

Download figure to PowerPoint

COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

One of the long-standing questions with respect to segmentation is whether there is a common origin of segmentation in the phyla with segmented animals (Fig. 1).

But before we can make a more detailed comparison, we first have to look with more detail into the genetic mechanisms of vertebrate somitogenesis. The segmental pattern along the anterior–posterior axis in the vertebrate embryo is established during embryogenesis via the formation of the somites. The somites are mesodermal epithelial blocks of cells that are patterned in the presomitic mesoderm at the posterior end of the vertebrate embryo (Pourquié1999,2001,2003; Saga and Takeda,2001; Holley and Takeda,2002; Rida et al.,2004). The terminal addition of somites that bud off from the presomitic mesoderm is thus reminiscent of the way segments are added in short germ arthropods. This progressive sequential addition of somites has mainly been studied in three different vertebrate species: mouse, zebrafish, and chicken. The patterning of the somites involves oscillatory gene expression driven by a segmentation clock that acts in combination with fibroblast growth factor (FGF) and Wnt signaling (Palmeirim et al.,1997; Dubrulle et al.,2001; Aulehla et al.,2003; Dequéant et al.,2006). Notch/Delta signaling is a core component of the oscillator. Once Notch is activated, several downstream target genes are activated, such as her/hes genes, lunatic fringe, Delta, axin, and nkd1 (Jouve et al.,2000; Bessho et al.,2001; Forsberg et al.,1998; Aulehla et al.,2003; Ishikawa et al.,2004). Interactions between these genes result in oscillations of the gene expression. As a result, waves of gene expression progress in an anterior direction, which are obvious as stripes in the presomitic mesoderm. The repetitive patterns of gene expression thus are the result of oscillatory gene expression in vertebrates. Decreasing posterior to anterior gradients of FGF8 and WNT3a are involved in the conversion of the periodic oscillations into a stable repeated pattern of somite boundaries (Dubrulle et al.,2001; Aulehla et al.,2003).

A molecular oscillator based on Notch/Delta signaling thus forms a key component of vertebrate somitogenesis and this is a completely different genetic mechanism than the hierarchic segmentation gene cascade of Drosophila. Furthermore, orthologs of most of the gap genes, pair rule genes, and segment-polarity genes are present in vertebrates, but most of them are not involved in somitogenesis. The her/hes genes are the exception; these hairy-related genes are expressed in stripes in the presomitic mesoderm of vertebrates (e.g., Müller et al.,1996) but the expression is cyclic as part of the oscillator (e.g., Holley and Takeda,2002). Comparison of the genetic program used in vertebrate somitogenesis with the one used in Drosophila segmentation thus shows no similarities and would lead to the conclusion that there is no common genetic program for segmentation in vertebrates and arthropods, and thus argues against a common origin of segmentation for these groups.

However, the picture is completely different if one tries to define the ancestral genetic mechanism for segmentation in arthropods by incorporating data from other arthropod groups. A Notch/Delta based mechanism is an important element of spider segmentation (Stollewerk et al.,2003; Schoppmeier and Damen,2005b) and presumably also in myriapods (Kadner and Stollewerk,2005; Janssen,2005; Peel et al.,2005). Chelicerates and myriapods are basally branched arthropod groups suggesting that the involvement of Notch/Delta in segmentation likely is an ancestral character for the arthropods. Not only the expression patterns but also the segmentation phenotypes seen for various Notch/Delta pathway components in the spider show a high similarity to the ones in vertebrates. This involvement of the Notch/Delta signaling pathway in both spider and vertebrate segmentation is probably the best evidence for a common origin of segmentation. But is this enough to imply a common origin for the use of these mechanisms in segmentation?

The patterning of the segments/somites in both the growth zone of short germ arthropods and the presomitic mesoderm of vertebrates takes place in a cellular environment, in which cell-to-cell signaling pathways seem to form a predestinated way for patterning. As there are only a limited number of signaling pathways in metazoans, the independent incorporation of such a pathway in segmentation in spiders and vertebrates seems to be plausible. Indeed, parallel recruitment of signaling pathways seems to be rather widespread (Irvine and Rauskolb,2001); Notch/Delta signaling, for instance, is also used in segmentation of the legs, in neurogenesis, and in eye development in arthropods (Rauskolb and Irvine,1999). However, in these processes Notch/Delta signaling always acts somewhere at the end of the cascade of specification of boundaries and cell populations. The role of Notch/Delta signaling in spider segmentation and vertebrate somitogenesis is fundamentally different in this respect as here Notch/Delta signaling is far upstream in the cascade that lays down the segmental boundaries. This is a unique character. Therefore, these similarities are a strong argument for a common origin and indicate that formation of the segments in arthropods and vertebrates may have shared a genetic program in a common ancestor.

Apart from the Notch/Delta signaling pathway, there are other similarities. As already mentioned FGF and Wnt-signaling act in combination with Notch/Delta signaling in vertebrates (Dubrulle et al.,2001; Aulehla et al.,2003; Dequéant et al.,2006). There are to date no data on involvement of FGF signaling in arthropod segmentation, but the discovery that Wnt signaling may be a crucial partner of Notch/Delta signaling in the generation of periodic gene expression during mouse somitogenesis is interesting (Aulehla et al.,2003). It has been observed that wingless (one of the arthropod Wnt genes) is expressed at the most posterior end of the growth zone in various short germ arthropods like the beetle Tribolium, the crustacean Triops, the spider Cupiennius, and the centipede Lithobius (Nagy and Carroll,1994; Nulsen and Nagy,1999; Damen,2002; Hughes and Kaufman,2002b). This is very intriguing, but so far the role of this posterior wingless domain in short germ arthropods is unclear as clear functional evidence is lacking. RNAi knockdown of armadillo, a downstream component of Wnt-signaling cascade, indeed suggests involvement of Wnt-signaling in the cricket Gryllus (Miyawaki et al.,2004). However, armadillo is also involved in cell adhesion (Cox et al.,1996); one, therefore, has to be very careful with this conclusion, especially because RNAi for wg did not show any effect in the cricket (Miyawaki et al.,2004). Further research is required to solve the role of wg/Wnt in the growth zone of arthropods. Of equal interest is that also caudal (cad) is expressed in the posterior growth zone of short germ arthropods (Copf et al.,2004; Shinmyo et al.,2005; Chipman et al.,2004; Akiyama-Oda and Oda,2003; Damen, unpublished data). In the beetle Tribolium, the cricket Gryllus, and the brine shrimp Artemia cad is required in an early phase of segmentation for the formation of most body segments (Copf et al.,2004; Shinmyo et al.,2005). cad orthologs also are involved in the posterior addition of segments in annelids (de Rosa et al.,2005) and somites in vertebrates (e.g., van den Akker et al.,2002; Chawengsaksophak et al.,2004; Shimizu et al.,2005). Mutations in Cdx genes, the vertebrate orthologs of cad, cause posterior truncations and disturb axial patterning. The data from zebrafish indicate that the Cdx genes are downstream targets of the Wnt signaling during morphogenesis of the posterior body (Shimizu et al.,2005). The involvement of Cdx/cad genes and Wnt signaling in axis formation from a posterior growth zone, thus, may date back to the urbilaterian, the last common ancestor of vertebrates, annelids, and arthropods.

Thus, there are several arguments that suggest a common origin for segmentation. However, a common origin that dates back to the urbilaterian also would imply that parts of this genetic program for segmentation have been lost in a number of lineages. How likely is it to assume several occasions where segmentation has been lost? New findings show that it is apparently possible to lose segmentation. Recent evidence results in a change of the phylogenetic position of the Echiura, a small group of unsegmented animals, and considers the Echiura now as an ingroup of the annelids. The Echiura thus must have lost their obvious body segmentation that is seen in the other annelids (Hessling and Westheide,2002; Bleidorn et al.,2003). Also the relationships of the vertebrates, urochordates, and cephalochordates have recently been reinterpreted. In the new phylogeny, the unsegmented urochordates (tunicates) and not the segmented cephalochordates (lancelets) are the closest relatives of the segmented vertebrates (Delsuc et al.,2006; Gee,2006). If these phylogenetic relationships are correct, the urochordates must have lost segmentation. The examples of urochordates and echiurans show that it is apparently possible to lose segmentation, but these examples also show the importance of knowledge of the phylogenetic relationships for interpretation of similarities.

PERSPECTIVES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

Presently, we only understand a fraction of the evolutionary transitions in the arthropod segmentation network. Much still is based on the Drosophila paradigm. The functional work in short germ insects has demonstrated the importance of the use of newly developed functional methods, such as parental RNAi (Bucher et al.,2002) and transgenesis (Berghammer et al.,1999; Pavlopoulos et al.,2004; Pavlopoulos and Averof,2005), to dissect the underlying genetic network. Only a solid understanding of the genetic interactions of the segmentation network in the various arthropod groups will allow us to understand the evolution of segmentation. Expression patterns tell us much, but only functional tests will give us insights into the regulatory interactions in the genetic network underlying segmentation. A good example for this is the concept of the pair rule circuit (Choe et al.,2006). Only in such a way can evolutionary transitions be uncovered. The requirement for Notch/Delta signaling in spider segmentation (Stollewerk et al.,2003) or the discovery of the new gap gene mlpt in Tribolium (Savard et al.,2006a) show us that we not only should analyze candidate genes by looking at orthologs of Drosophila segmentation genes, but also perform unbiased approaches like EST screens to identify new factors and genetics interactions.

Future research has to solve important questions on the regulatory interactions in non-insect arthropods. It is of eminent importance to understand the way the Notch/Delta signaling pathway interacts with canonical Drosophila segmentation genes like the pair rule gene orthologs. Related to this is a requirement to understand the nature of the dynamic expression of pair rule gene orthologs and Delta in the growth zone of these animals. Do these genes oscillate like they do in the vertebrate presomitic mesoderm? It has been proposed, based on expression patterns, that these genes indeed oscillate, but unmistakable evidence is still missing. Furthermore, does FGF and Wnt signaling play a role in arthropod segmentation as they do in vertebrates? Expression patterns hint to an involvement of Wnt signaling, but what is its exact role and is this similar to the role in vertebrates? Another important question is to understand how and when gap gene orthologs became involved in segmentation? To that end, we have to analyze to what extent gap gene orthologs play a role in segmentation of non-insect arthropods.

Finally, to answer the question on the evolution of segmentation mechanisms, we also have to look beyond the arthropods. It will be important to analyze segmentation in onychophorans and tardigrades, which are the closest relatives of the arthropods and also have a segmented body plan (Fig. 1).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES

I am grateful to the people in my lab as well as the people of the Tribolium group in the lab of Diethard Tautz for inspiring discussions on the evolution of segmentation. I thank Evelyn Schwager and Alistair McGregor for a critical reading of the manuscript. The research in my lab is supported in part by grants from the Deutsche Forschungsgemeinschaft via Project A12 in SFB 572 of the Universität zu Köln and the European Union via the Marie Curie Research and Training Network ZOONET (MRTN-CT-2004-005624).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE DROSOPHILA PARADIGM
  5. CONSERVATION OF THE SEGMENT-POLARITY NETWORK AND THE CONSERVED ENTITY OF THE PARASEGMENT
  6. THE TRANSITION FROM A NON-PERIODIC TO A PERIODIC PATTERN: ORTHOLOGS OF PAIR RULE GENES
  7. PAIR RULE CIRCUIT, AND PRIMARY AND SECONDAY PAIR RULE GENES
  8. DIVERGENT UPSTREAM MECHANISMS I: GAP GENES
  9. DIVERGENT UPSTREAM MECHANISMS II: NOTCH/DELTA SIGNALING
  10. SCENARIO FOR EVOLUTIONARY TRANSITIONS IN ARTHROPOD SEGMENTATION
  11. COMMON ORIGIN OF SEGMENTATION?: SIMILARITIES WITH VERTEBRATE SOMITOGENESIS
  12. PERSPECTIVES
  13. Acknowledgements
  14. REFERENCES