Lack of evidence for conserved parasegmental grooves in arthropods

In the arthropod model species Drosophila melanogaster, a dipteran fly, segmentation of the anterior–posterior body axis is under control of a hierarchic gene cascade. Segmental boundaries that form morphological grooves are established posteriorly within the segmental expression domain of the segment-polarity gene (SPG) engrailed (en). More important for the development of the fly, however, are the parasegmental boundaries that are established at the interface of en expressing cells and anteriorly adjacent wingless (wg) expressing cells. In Drosophila, both segmental and transient parasegmental grooves form. The latter are positioned anterior to the expression of en. Although the function of the SPGs in establishing and maintaining segmental and parasegmental boundaries is highly conserved among arthropods, parasegmental grooves have only been reported for Drosophila, and a spider (Cupiennius salei). Here, we present new data on en expression, and re-evaluate published data, from four distantly related spiders, including Cupiennius, and a distantly related chelicerate, the harvestman Phalangium opilio. Gene expression analysis of en genes in these animals does not corroborate the presence of parasegmental grooves. Consequently, our data question the general presence of parasegmental grooves in arthropods.


Introduction
In the fruit fly Drosophila melanogaster, segmentation of the anterior-posterior (AP) body axis proceeds through the action of the so-called segmentation genes, i.e., gap, pairrule, and segment-polarity genes. At the end of this cascade, pair-rule genes (PRGs) such as even-skipped (eve) and fushi-tarazu (ftz) ensure that segment-polarity genes (SPGs) such as engrailed (en) and wingless (wg) are expressed in segmentally repeated stripes (reviewed in Sanson 2001). Expression of en and wg is initiated in adjacent regions of every segment, wg anteriorly abutting the domain of en.
After the initiation of this pattern by the PRGs, a positive auto-regulatory loop between en and wg expressing cells maintains their expression. The border between en and wg expressing cells is the parasegmental border, which acts as the primary organization center of the segment and functions as a clonal boundary (e.g., Vincent and O'Farrell 1992). The segmental boundaries that are represented by grooves form posterior to the expression of en (or in the most posterior en expressing cells). These are the morphological boundaries that are later seen in the larva and also in the adult fly. At least in Drosophila, morphological grooves also form at the parasegmental boundary, i.e., at the interface between en and wg expressing cells (anterior to en), but the function of these grooves is not known (e.g. Larsen et al. 2008). Interestingly, parasegmental grooves have also been reported in a distantly related arthropod, the spider Cupiennius salei (Damen 2002). A later study in the brine shrimp Artemia franciscana suggested that parasegmental groove may also (although less obviously) form in crustaceans (Prpic 2008). These data combined led to the opinion that parasegmental grooves indeed represent a conserved feature of arthropod development (e.g. Deutsch 2004;Mellenthin et al. 2006;Chipman 2010;Franke and Mayer 2014). Parasegments marked by the highly conserved expression of the SPGs in arthropods appear to be conserved among arthropods, but the occurrence of parasegmental grooves has indeed only been reported for the above-mentioned species. Some authors explicitly mention that they have never observed parasegmental grooves in their model organisms (i.e., Chipman et al. 2004;Brena et al. 2006). Published data on SPG expression in other spiders than Cupiennius also do not mention parasegmental grooves, although these studies indeed are not focusing on this topic (e.g., Turetzek and Prpic 2016;Pechmann 2020).
We therefore decided to carefully analyze the expression of the SPG en with respect to the presence or absence of parasegmental grooves in a number of spider species that cover most main branches of spiders, true spiders such as Parasteatoda tepidariorum (Entelegynae) and Pholcus phalangioides (Haplogynae) and the tarantula Acanthoscurria geniculata (Mygalomorphae). In neither of these species, we found grooves forming anterior to the expression of en showing that clearly visible parasegmental grooves as described by Damen (2002) for the spider Cupiennius may not form at all in these investigated species. We then re-investigated the expression of en in Cupiennius, a species that represents a different subgroup of Entelegynae (i.e., the RTA clade (e.g., Garrison et al. 2016)) to which Parasteatoda does not belong, and for which parasegmental grooves have indeed been reported (Damen 2002)). Even in this species, however, we were unable to identify parasegmental grooves. This suggests that the earlier report by Damen (2002) must have interpreted the data incorrectly.
We also found that all spiders and closely related groups of chelicerates that belong to Arachnopulmonata (e.g., spiders, scorpions, whip scorpions) all possess two paralogs (ohnologs) of en. The second paralog (en2) likely evolved new functions in this group of chelicerates after a whole genome duplication (WGD) that took place in the lineage leading to arachnopulmonate chelicerates (Schwager et al. 2017).

Methods
Sequence information of en genes have been identified in a sequenced genome (Parasteatoda (Schwager et al. 2017)) and sequenced embryonic transcriptomes (Cupiennius (Samadi et al. 2015), Pholcus (Janssen et al. 2015), Acanthoscurria (Pechmann 2020), Phalangium (Sharma et al. 2012), Marpissa muscosa (Harper et al. 2020), Charinus acosta (Harper et al. 2020), and Euphrynichus bacillifer (Harper et al. 2020). Potential orthologs were identified performing reciprocal tBLASTn searches against the single en gene of the onychophoran Euperipatoides kanangrensis (Eriksson et al. 2009). Phylogenetic analysis of en genes was done as described previously (Panara et al. 2019). The protein alignment and the nexus file are provided as Supplementary Files 1 and 2; gene identifiers of all genes used in this paper are summarized in the Supplementary File 3. Fragments of the genes were amplified using RT-PCR with gene-specific primers on cDNA that was reverse transcribed from total RNA isolated from embryos of different developmental stages; sequences of all used primers are summarized in Supplementary File 3. All fragments were cloned into the pCR-II Vector (Invitrogen) or the pJet1.2 CloneJet PCR cloning kit (Fisher Scientific) and sequenced on an ABI3730XL analyzer using Big Dye dye-terminators. In situ hybridization was performed as described in Janssen et al. (2018, supplement) and Panara et al. (2019) (Fas-tRed staining for confocal microscopy). Cell nuclei staining was done by incubation of the embryos in 1:10,000 Sybrgreen or 1 µg/ml 4-6-diamidino-2-phenylindole (DAPI) in phosphate-buffered saline with 0.1% Tween-20 (PBST) for 20-30 min. Stained embryos were investigated and photographed under a Leica dissection microscope that was equipped with a Leica DC100 digital camera. For confocal microscopy, we used a Leica SP8 inverted microscope. Embryos used for confocal microscopy were mounted in 1% low melting agarose. Confocal images were processed in Fiji/ImageJ (Schindelin et al. 2012). The image processing software Adobe Photoshop CC2018 was used for linear corrections of brightness and contrast.

Spiders and other arachnopulmonate chelicerates retained two engrailed paralogs after whole genome duplication (WGD)
In the spiders, Parasteatoda, Pholcus, and Acanthoscurria, so far only one en gene was isolated in previous studies (Pechmann et al. 2009;Turetzek and Prpic 2016;Pechmann 2020). However, in Cupiennius, two paralogs were described (engrailed-1 (en1) and engrailed-2 (en2) (Damen 2002)). We identified an en2 paralog in all investigated spiders, as well as in whip scorpions and a scorpion suggesting that all (or at least most) arachnopulmonate species may have retained two copies (ohnologs) of en after WGD (Schwager et al. 2017;Harper et al. 2020). Our phylogenetic analysis supports this as it shows monophyletic groups of en genes representing en1 and en2 respectively (Fig. 1). The most striking difference between the protein sequences of En1 (= En) and En2 is the lack of the C-terminal aspartic acid (D)-and glutamic acid (E)-rich domain that is usually present in En proteins, and a serine (S)-rich domain present in spider, whip-scorpion, mite, and harvestman En1 (En) proteins (Supplementary File 4).

Expression of engrailed genes reveals the absence of clearly-visible parasegmental grooves in spider development
In his influential paper, Damen (2002) described for the first time the expression of the SPG en (two paralogs) in a spider. Based on the conserved expression patterns of this gene, he concluded that the parasegmental boundaries that are set by the same genes in Drosophila and other arthropods are conserved in spiders (Damen 2002, and references therein). Additionally, Damen (2002) reported on the presence of transient morphologically visible parasegmental showing the phylogenetic relationship of chelicerates used in this study: 1, Arachnopulmonata; 2, spiders; 3, true spiders; 4, Entelegyne spiders boundaries (grooves) that allegedly form anterior to the segmental expression of en. Based on his finding, and the fact that parasegmental grooves exist in Drosophila (e.g., Martinez-Arias and Lawrence 1985; Larsen et al. 2008), he argued that parasegmental grooves are conserved and that they are thus representing an ancestral feature of arthropod development.
Although we fully agree with the idea of conserved parasegmental boundaries in spiders and indeed arthropods in general, we question the presence of transient parasegmental grooves in spiders. This is because expression of en genes in spiders other than Cupiennius has been studied, but neither of the corresponding papers described parasegmental grooves (e.g., Pechmann et al. 2009;Akiyama-Oda and Oda 2010;Turetzek and Prpic 2016;Pechmann 2020). This, however, may be due to the fact that these papers did not specifically address groove formation during segmentation. Additionally, numerous studies that provide comprehensive data on SPG expression (including en) in a large variety of different arthropod species did not report on the presence  Therefore, we first carefully analyzed the expression pattern of en1 in Parasteatoda (Entelegynae), Pholcus (Haplogynae), and Acanthoscurria (Mygalomorphae) with respect to the potential formation of parasegmental grooves that, if present, would form anterior to the segmental expression of en1. However, in neither of the investigated species, en1 is ever expressed posterior adjacent to any visible transverse morphological groove, but instead grooves are always associated with the most posterior segmental expression of en1 (Fig. 2), suggesting that these grooves indeed represent the "normal" segmental grooves that are present in all arthropods. Since we did not find any evidence for the formation of parasegmental grooves in any of these spiders, we then carefully re-investigated the expression of en1 in Cupiennius. To our surprise, we found that even in this species expression of en1 is clearly aligned with the formation of the segmental grooves (Fig. 3). Our data on the expression of Cupiennius en1 conflicts with the provided data by Damen (2002). However, together with our data on en1 expression in various spider species, we conclude that the grooves reported in Cupiennius represent segmental borders, and not parasegmental borders.
In order to get better insight into the morphology of the germ band, we also investigated its morphology before and around the onset of en1 expression in one spider, Parasteatoda, by means of confocal microscopy. We combined the  Fig. 1) with the nuclear dye DAPI (Fig. 4). Confocal microscopy offers a much higher resolution and allows to scan through all planes of a given embryo. We scanned embryos in slices of 1 µm and could not detect any sign of a groove anterior to en1 (FastRed in-situ hybridization signal; see Supplementary Fig. 1). Neither the position of nuclei (DAPI signal) nor the surface of the cells (visualized by the background signal of the FastRed stain) indicates the presence of grooves (Fig. 4).

Evolutionary significance
Engrailed and other SPGs are responsible for maintaining the conserved parasegmental borders in Drosophila and other arthropods (reviewed in Sanson 2001). Interestingly, the same genes appear to be responsible for segment border formation in annelids as shown for the nereid Platynereis dumerilii, thereby suggesting that the last common ancestor of protostomes was already segmented in at least some form and that the SPG-network was part of its segmentation process (Dray et al. 2010). In the annelid, however, segmental grooves (morphological segmental boundaries) form at the interface of en-and wg/Wnt1-expressing cells-where also the parasegmental boundaries of arthropods form. Conserved parasegmental grooves in arthropods would then be homologous to the segmental grooves of annelids. By showing that the parasegmental grooves are not a conserved ancestral feature of arthropod segmentation, our data weaken the idea that segmentation and groove formation in arthropods and annelids are conserved and thus question the idea of a segmental protostomian ancestor.

Expression of en2 reveals aspects of suband neo-functionalization after WGD
Arachnopulmonate chelicerates retained two copies of engrailed after WGD. Other chelicerates such as harvestmen and mites, however, ancestrally only possess one en gene (en1) (Fig. 1). In all chelicerates, en1 is expressed in the form of transverse segmental stripes suggesting that its function as a SPG is conserved (Figs. 2 and 3) (Telford and Thomas 1998;Damen 2002;Sharma et al. 2012;Turetzek and Prpic 2016;Sharma 2017;Pechmann 2020). Segmental grooves form in the posterior sector of this domain, as it is also the case in all other previously studied arthropods. Fig. 7 Expression of spider en2 in the head and the appendages. The anterior head region with the stomodaeum and the chelicerae are shown from ventral. The pedipalps and legs are shown from lateral (left panels) and dorsal (anterior to the right) (right panels). Note expression in the center of the pedipalps and legs. Abbreviations as in Fig. 5 Expression of the arachnopulmonate-specific second en paralog (en2), however, is different. Although it is expressed in the form of transverse segmental stripes, very similar to the expression of en1, the stripes of en2 are broader and extend further towards posterior spanning the segmental grooves. Consequently, en2 is expressed anterior (just like en1) and posterior to these grooves (Figs. 5 and 6). Additionally, spider en2 genes (but not en1) are expressed in the stomodaeum, and expression in the appendages is dorsal and more anterior than that of en1 (Figs. 5-7). Unlike en1 genes, however, orthologs of en2 are not expressed in the ventral nervous system, representing a case of sub-functionalization. Expression of the single harvestman en gene is very similar to that of spider en1: the segmental stripes are not split, there is no expression of en in the stomodaeum, expression in the ventral nervous system is present, and expression in the appendages is posterior (Fig. 8). The pattern of en2 is thus derived representing a case of neo-functionalization. And since this new pattern is highly conserved in at least spiders (expression data from other arachnopulmonate chelicerates are not available), it must have an important, likely new, role during development. Unfortunately, however, studies concerning the function of en2 in spiders are lacking. The remaining question is thus what (if any) morphological novelty may be correlated with the function of the second engrailed gene?
Author contribution RJ designed the study, performed all in situ hybridization experiments, and wrote the first draft of the manuscript. NT provided sequence information, cDNA and embryos of Pholcus. MP established the Acanthoscurria colony and provided sequence information, embryos, and cDNA. All authors discussed the data and worked on the final version of the manuscript.
Data availability All data of this study have been made public.

Declarations
Ethics approval Not applicable. Note that the domain of segmental en expression is not split (arrow and arrowhead) and that there is no expression in the stomodaeum. Abbreviations as in Fig. 5 Consent to participate Not applicable.

Conflict of interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.