Development Genes and Evolution

, Volume 213, Issue 5, pp 273–283

A genomewide survey of developmentally relevant genes in Ciona intestinalis

VII. Molecules involved in the regulation of cell polarity and actin dynamics

Authors

  • Yasunori Sasakura
    • Department of Zoology, Graduate School of ScienceKyoto University
  • Lixy Yamada
    • Department of Zoology, Graduate School of ScienceKyoto University
  • Naohito Takatori
    • Department of Zoology, Graduate School of ScienceKyoto University
  • Yutaka Satou
    • Department of Zoology, Graduate School of ScienceKyoto University
    • Department of Zoology, Graduate School of ScienceKyoto University
Original Article

DOI: 10.1007/s00427-003-0325-9

Cite this article as:
Sasakura, Y., Yamada, L., Takatori, N. et al. Dev Genes Evol (2003) 213: 273. doi:10.1007/s00427-003-0325-9

Abstract

In the present study, genes involved in the pathways that establish cell polarity and cascades regulating actin dynamics were identified in the completely sequenced genome of Ciona intestinalis, a basal chordate. It was revealed that the Ciona genome contains orthologous genes of each component of aPKC-Par and PCP pathways and WASP/WAVE/SCAR and ADF/cofilin cascades, with less redundancy than the vertebrate genomes, suggesting that the conserved pathways/cascades function in Ciona development. In addition, the present study found that the orthologous proteins of five gene groups (Tc10, WRCH, RhoD, PLC-L, and PSKH) are conserved in humans and Ciona but not in Drosophila melanogaster, suggesting a similarity in the gene composition of Ciona to that of vertebrates. Ciona intestinalis, therefore, may provide refined clues for the study of vertebrate development and evolution.

Keywords

Actin dynamicsBasal chordatesCell polarityCiona intestinalisGenomewide survey

Introduction

In the present study, genes with functions in the establishment of cell polarity, cell movement and the regulation of actin dynamics were investigated in the most thoroughly studied ascidian, Ciona intestinalis. While many reports have described the polarity, cell movements, and involvement of actin filaments in the processes, little is known about the molecules functioning in these events of ascidians (Conklin 1905a, 1905b; Zalokar 1974; Reverberi 1975; Sawada and Osanai 1981; Yoshida et al. 1997; Sasakura et al. 2000; Munro and Odell 2002a, 2002b). The aim of the present study was to describe how many genes, with or without redundancy, are encoded in the Ciona genome, and to reveal what molecules are used in these events during Ciona development.

Recent studies with model organisms have shown that conserved systems regulate cell polarity formation, cell movement and actin dynamics. As an example, both the A-P axis of fertilized eggs of protostomes (e.g. Caenorhabditis elegans and Drosophila melanogaster), and the apical-basal polarity of epithelial cells of mammals, are regulated by atypical protein kinase C (aPKC), Par1–6 and their homologous proteins (aPKC-Par pathway, shown in Fig. 1A; Rose and Kemphues 1998; Ohno 2001). Another example is the PCP pathway (Fig. 1B), in which orthologous proteins, which regulate the planar cell polarity (PCP) of D. melanogaster epithelial cell sheets, also function in the convergent extension during gastrulation in vertebrates (Kuhl et al. 2000, 2001; Aelst and Symons 2002; Myers et al. 2002; Wallingford et al. 2002a). In the PCP pathway, some Wnt/wingless homologues, such as Wnt-5a and -11, and their receptor Frizzled (Fz), activate two non-canonical pathways through different domains of Dishevelled (Dsh). One is the activation of the MAP kinase JNK (JNK pathway), and the other (Ca2+ pathway) is the activation of calcium/calmodulin protein kinase II (CamKII) and PKC by causing Ca2+ release through inositol 1,4,5-triphosphate (IP3) synthesized by phospholipase C (PLC; Kuhl et al. 2000). Moreover, some conserved proteins, Strabismus/Van Gogh (Stbm), Prickle (Pk), and Daam-1, function in the PCP pathway (Kuhl et al. 2000).
https://static-content.springer.com/image/art%3A10.1007%2Fs00427-003-0325-9/MediaObjects/s00427-003-0325-9flb1a-d.gif
Fig. 1A–D.

Four pathways/cascades investigated in this study. A aPKC-Par pathway. B PCP pathway. C WASP/WAVE/SCAR cascade. D ADF/cofilin cascade. A component which was not found in Ciona (Par-2) is in parentheses. Since the positions of Par-4 and Par-5 (14-3-3) in the aPKC-Par pathway were obscure, these molecules are not included in this figure. Ciona Par-4 orthologue and 14-3-3 proteins were found. Ciona Wnt, Fz, Dsh and JNK were identified and are to be described in other accompanying reports (Hino et al. 2003; Satou et al. 2003b)

Small G-proteins Rho/Rac/CDC42, which are key factors of actin regulation and have multiple roles in many cellular events, also function in the aPKC-Par and PCP pathways (Takai et al. 2001; Aelst and Symons 2002; Choi and Han 2002). However, the downstream cascade of Rho/Rac/CDC42 in the two pathways remains unclear. Rho/Rac/CDC42 activate many downstream cascades. Among these, we focused on two conserved cascades, WASP/WAVE/SCAR activation by Rac and CDC42 (Fig. 1C), and the regulation of actin depolymerizing factor ADF/cofilin by Rho and Rac (Fig. 1D; Erickson and Cerione 2001; Maciver and Hussey 2002). Both cascades regulate actin dynamics. In the ADF/cofilin cascade, many serine/threonine protein kinases (S/T-Ks), Lim-kinase (LIMK), Testicular-specific kinase (TESK), Rho-kinase (ROCK), p21-activated kinase (PAK), PKC and one phosphatase Slingshot (SSH) are involved. It is highly possible that the cascades function in the aPKC-Par and PCP pathways, because the regulation of actin dynamics is critical in both pathways.

These pathways and cascades are possibly used during Ciona development, and therefore it is important to reveal how many of the genes involved in these pathways are encoded in the Ciona intestinalis genome (Dehal et al. 2002). Our present study revealed that: (1) almost all of the orthologous genes are conserved in Ciona, with less redundancy than in vertebrates; and (2) the gene composition of the Ciona genome is closer to that of a vertebrate human than that of the protostome D. melanogaster.

Materials and methods

All of the methods used in this study are the same as those described by Satou et al. (2003a). The sequences used are represented by accession number, abbreviation of species (HS for Homo sapiens, HP for Hemicentrotus pulcherrimus, MM for Mus musculus, XL for Xenopus laevis, DR for Danio rerio, CS for Ciona savignyi, CE for Caenorhabditis elegans and DM for Drosophila melanogaster), and gene name. For example, human cPKC-β (accession number P05771) is represented as "P05771 HS cPKC- β". All sequences used in the present study are listed up in Table S1 (electronic supplementary material).

Results and discussion

aPKC-Par pathway

PKC

aPKC is part of the serine/threonine protein kinase (S/T-K) group, PKC (Moscat and Diaz-Meco 2000; Webb et al. 2000). PKC functions in many aspects of cellular events and developmental processes including the aPKC-Par pathway, the Wnt/Ca2+ pathway and ADF/cofilin regulation (Rose and Kemphues 1998; Kuhl et al. 2000; Ohno 2001; Maciver and Hussey 2002). PKC is divided into three groups: cPKC, nPKC and aPKC. PKC has conserved domains, C1 and C2, in addition to one S/T-K domain, and the composition of the C1/C2 domains is different among the groups. cPKC has a C1 and a C2 domain, nPKC has a C1 domain, and aPKC has a C2-like and a C1 domain from its N-terminal. In addition, another group of S/T-Ks, which has a kinase domain very similar to PKC, was also reported (PRKCL; Palmer et al. 1995). This group of proteins has HR1 domains and a C2 domain.

We found six candidates of PKC and PRKCL orthologues from the Ciona genome and cDNA/EST databases (Table 1). Among these, Ci-cPKC encoded a protein with a C1 and a C2 domain, Ci-nPKC-δ/θ and Ci-nPKC-μ had a C1 domain, and Ci-PRKCL had two HR1 domains and a C2 domain. From the domain composition, Ci-nPKC-δ/θ and Ci-nPKC-μ appeared to be orthologues of nPKC, and Ci-PRKCL an orthologue of PRKCL. Domains other than the S/T-K domain in the other candidate proteins were not found. This may have been because they are located near the end of the scaffold and the region coding the domains might have been lost during assembly. A phylogenetic tree drawn by the S/T-K domain is shown in Fig. 2. The tree indicated that Ci-cPKC is orthologous to cPKCs (bootstrap value 94%), Ci-nPKC-ε/η to human nPKC-ε, η and D. melanogaster DPKC98F (94%), Ci-nPKC-δ/θ to human nPKC-δ, θ and D. melanogaster PKC-δ (84%), Ci-aPKC to aPKCs (100%), Ci-PRKCL to human PRKCL1, 2 and D. melanogaster CG2049 (99%), and Ci-nPKC-μ to human nPKC-μ, nPKC-D2 and D. melanogaster CG7125 (100%). The result was in accordance with the orthology predicted by the domain composition. Although their domain structure is closely related, nPKC was divided into three distant subgroups, and Ciona had an orthologous gene in each subgroup. We concluded that Ciona has one cPKC, three nPKCs, one aPKC, and one PRKCL orthologue. The accuracy of the gene prediction and the expression of all of the Ciona PKC genes were confirmed by EST (Table 1).
Table 1.

Genes involved in aPKC-Par pathway, PCP pathway, WASP/WAVE/SCAR cascade, and ADF/cofilin cascade in Ciona intestinalis

Pathway/cascade

Gene name

The best gene model in the version 4 assembly

cDNA cluster

The best hit protein in the human proteome

Bootstrap support (%)

Best hit analysisa

Other supporting evidence

aPKC-Par

cPKC

grail.60.22.1

04699

P17252 PKC-α

94

 

nPKC-δ/θ

grail.53.49.1

03017

Q0565 nPKC-δ

84

domain composition

nPKC-ε/η

grail.9.1.1

31660

Q02156 nPKC-ε

94

 

aPKC

grail.136.32.1

05328

P41743 nPKC-iota

100

 

PRKCL

genewise.215.5.1

04819

Q16513 Protein kinase C-like 2

99

domain composition

nPKC-μ

grail.99.17.1

02012

O94806 nPKC-μ

100

domain composition

Par-1

grail.9.2.1

02316

Q96RG1 PAR-1A

100

domain composition

POPK-1

grail.41.79.1

00199

Q8TDC3 Putative serine/threonine protein kinase

99

 

Par-3

grail.159.20.1

16590

Q96RM6 Atypical PKC isotype-specific interacting protein longvariant b

35–78

Sequence similarity (Fig. S2)

Par-6

grail.8.56.1

10726

Q9BYG5 PAR-6 ß

100

domain composition

Par-4

grail.315.12.1

04784

Q9HBS3 Hypothetical protein

99

 

14–3-3-a

grail.603.8.1

00192

P35214 14–3-3 protein γ

-

Sequence similarity (Fig. S5)

14–3-3-b

grail.603.9.1

06461

Q04917 14–3-3 protein η

-

Sequence similarity (Fig. S5)

14–3-3-c

grail.5.2.1

12646

P35214 14–3-3 protein γ

-

Sequence similarity (Fig. S5)

14–3-3-d

grail.81.51.1

01417

P35214 14–3-3 protein γ

-

Sequence similarity (Fig. S5)

14–3-3 ε-1

grail.626.6.1

03329

P42655 14–3-3 protein ε

96

Sequence similarity (Fig. S5)

14–3-3 ε-2

grail.626.7.1

00726

P42655 14–3-3 protein ε

96

Sequence similarity (Fig. S5)

coronin

grail.520.1.1

01326

Q9BR76 Coronin 1B

100

Sequence similarity (Fig. S8)

Pod-1

grail.28.64.1

13395

P57737 70 kDa WD-repeat tumor rejection antigen homolog

100

domain composition

PCP

WRCH

grail.36.1.1

04923

Q9NPY5 WRCH-1

93

 

CDC42

grail.702.2.1

02045

P25763 CDC42 homolog

100

 

Rho/Rac/CDC42-related-a

grail.68.39.1

30344

P15154 p21-Rac1

98

 

Rho/Rac/CDC42-related-b

grail.68.45.1

03940

P15154 p21-Rac1

98

 

Rac-a

grail.14.22.1

01063

P15154 p21-Rac1

62

 

Rac-b

grail.14.18.1

02839

P15154 p21-Rac1

62

 

Rac-c

grail.14.19.1

14467

P15154 p21-Rac1

62

 

Rac-like-a

grail.14.20.1

NA

P15154 p21-Rac1

-

 

Rac-like-b

grail.14.21.1

NA

P15154 p21-Rac1

-

 

RhoA/B/C

grail.76.26.1

00588

P06749 RhoA

100

 

RhoD

grail.33.30.1

13518

P01121RhoB

40

 

Tc10

grail.1274.2.1

00329

P17081 TC10

56

 

PLC-β1/2/3

grail.18.81.1

32711

Q01970 PLC-β-3

61

 

PLC-β4

grail.1530.1.1

32529

Q15147 PLC-β-4

72

 

PLC-a

grail.769.4.1

10738

Q9BRC7 Hypothetical protein

-

 

PLC-b

grail.33.13.1

31523

Q9BRC7 Hypothetical protein

-

 

PLC-c

grail.4.88.1

16042

Q9BRC7 Hypothetical protein

-

 

PLC-L

grail.363.6.1

06743

Q9UPR0 KIAA1092

94

 

PLC-γ

grail.22.55.2

16654

P19174 PLC-γ-1

99

domain composition

CAMKI

grail.862.1.1

04737

Q9HD31 CamKI-like protein kinase

100

 

CAMKII

grail.58.7.1

00993

Q13557 CaMK-II δ subunit

100

 

PSKH

grail.118.1.1

12702

P11801 PSK-H1

100

 

CAMKIV

grail.6.60.1

34427

Q16566 CaMK IV

99

 

CAMK-like-a

grail.14.24.1

05688

O15075 DCAMKL1

100

 

CAMK-like-b

grail.10.105.1

10253

Q9C098 KIAA1765

100

 

CASK

grail.365.12.1

04851

O14936 CASK

100

 

Strabismus/Van Gogh

grail.6.19.1

02731

Q9ULK5 KIAA1215

-

Sequence similarity (Fig. S11)

Prickle

grail.1.77.1

05964

Q96MT3 FLJ31937

100

 

Triple-lim

grail.240.12.1

04107

Q9UGI8 Testin

-

outgroup

Daam-1

grail.1394.1.1

07435/31372

Q9Y4G0 KIAA0381

100

 

FHOS

grail.633.1.1

07069

Q9C0G8 KIAA1695

100

 

Formin/Cappuccino-a

genewise.1968.1.1

31842

Q9NZ56 Formin 2

98

 

Formin/Cappuccino-b

grail.20.36.1

NA

Q9NSV4 Diaphanous-related formin 3

98

 

Diaphanous

grail.326.6.1

16844

O60879 Diaphanous-related formin 2

100

 

Formin-like

genewise.76.8.1

02686

Q96PY5 KIAA1902

100

 

Formin/Diaphanous/ Daam-1-related-a

genewise.651.5.1

06823

Q9C0D6 KIAA1727

100

 

Formin/Diaphanous/ Daam-1-related-b

grail.67.28.1

32537

Q9BRM1 Hypothetical protein

100

 

WASP/WAVE/SCAR

WASP

grail.512.12.1

04333/12259

P42768 WASP

100

domain composition

Homer

grail.77.10.1

06270

Q9NSB8 Homer-2 protein

100

 

VASP/ENA

grail.49.1.1

15268

Q9UIC2 RNB6

99

 

WH1

grail.253.14.1

11957

Q9UIC2 RNB6

87

 

WAVE/SCAR

grail.1157.2.1

13213

Q9UPY6 WASP-family protein member 3

-

Sequence similarity (Fig. S15)

ARP3

grail.960.1.1

13270

Q9P1U1 ARP3β

100

 

ARP1

grail.96.24.1

00270

P42024 ARP1

99

 

ARP2

grail.1260.1.1

08111

O15142 ARP2

100

 

actin-like-a

grail.191.28.1

 

Q9NWY2 FLJ20537

98

 

actin-like-b

genewise.176.48.1

05771

P02571 Actin, cytoplasmic 2

-

 

p21-ARC

grail.362.7.1

00635

O15145 P21-ARC

-

Sequence similarity (Fig. S17)

p34-ARC

grail.490.3.1

02109

O15144 P34-ARC

-

Sequence similarity (Fig. S17)

p20-ARC

grail.1.155.1

01052

O15509 P20-ARC

-

Sequence similarity (Fig. S17)

p16-ARC

grail.21.37.1

03726

O15511 P16-ARC

-

Sequence similarity (Fig. S17)

p41-ARC

grail.988.4.1

01743

Q92747 Actin-related protein 2/3 complex subunit 1A

-

Sequence similarity (Fig. S17)

ADF/cofilin

ADF/Cofilin

grail.214.17.1

06006

Q9Y281 Cofilin 2

-

Sequence similarity (Fig. S18)

LIMK

genewise.648.29.1

00229

P53667 LIMK-1

98

domain composition

TESK

grail.220.30.1

34568

Q96S53 Testicular protein kinase 2

99

 

ROCK

grail.127.30.1

13543

O75116 KIAA0619

100

domain composition

SSH

grail.326.12.1

07760

Q8WYL2 HSSH-2

100

domain composition

PAK1/2/3

grail.81.56.1

02991

Q13153 PAK 1

100

 

PAK4/5/6

grail.135.11.1

09766

Q9ULS8 KIAA1142

100

 

a"↔" indicates a bi-directional best-hit relationship between a Ciona gene and a human protein, and "→" indicates a uni-directional best-hit relationship of a Ciona protein against a human protein

NA: not available

https://static-content.springer.com/image/art%3A10.1007%2Fs00427-003-0325-9/MediaObjects/s00427-003-0325-9flb2.gif
Fig. 2.

Phylogenetic tree of PKCs, generated by the neighbor-joining method based on the alignment of a serine/threonine kinase domain sequence. Ciona proteins are shown by large black dots. The number beside each branch indicates the percentage of times that a node was supported in 1000 bootstrap pseudoreplications. Protein names are explained in Materials and methods. Although this tree should be drawn as an unrooted tree, a rooted tree is shown here for simplicity. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position

Par-1

Par-1 encodes a well-conserved S/T-K with a KA1 domain at the C-terminal (Blot et al. 2002). C. elegans and D. melanogaster Par-1 is localized in the posterior half of the eggs, and is necessary for the establishment of A-P polarity and asymmetric cell division of the eggs (Rose and Kemphues 1998; Ohno 2001). Their orthologues in mammals are known as MAP/microtubule affinity-regulating kinases (MARKs). One candidate for the Ciona Par-1/MARK orthologue was found in the genome because it contained both a S/T-K and a KA1 domain (Table 1). The accuracy of the gene prediction and the gene expression were confirmed by EST (Table 1).

In addition to Par-1/MARK, there is another group of S/T-Ks that has a Par-1-like S/T-K domain but does not have a KA1 domain. HrPOPK-1 was identified as such a S/T-K in another ascidian, Halocynthia roretzi (Sasakura et al. 1998). Since HrPOPK-1 mRNA is localized to the posterior pole of fertilized eggs, HrPOPK-1 may have a role in the determination of the A-P axis of the egg. It was not known whether HrPOPK-1 orthologues are conserved among animals. We surveyed a candidate for Ciona POPK-1 (Table 1) and homologous kinases of humans (KIAA1811 and CAA07196), D. melanogaster (CG6114), and C. elegans (F15A2.6) in their completely sequenced genomes. Ciona POPK-1 had corresponding ESTs, confirming the accuracy of the gene prediction (Table 1).

A phylogenetic tree drawn by the S/T-K domain revealed that Par-1/MARK and the POPK-1 homologue formed different groups with bootstrap values of 100% and 99%, respectively (Fig. S1: electronic supplementary material). Ci-Par-1 is an orthologue of Par-1 and MARK. The tree also suggested that the last common ancestor of ascidians and vertebrates had a single MARK gene. On the other hand, Ciona Ci-POPK-1, human KIAA1811 and CAA07196, D. melanogaster CG6114, and C. elegans F15A2.6 are orthologous genes of HrPOPK-1 (99%). The conservation of HrPOPK-1 homologues among phyla is interesting, and their function during embryogenesis should be investigated.

Par-2

C. elegans Par-2 encodes a polypeptide with a RING finger and an ATP-binding domain (Levitan et al. 1994). An orthologous gene has not yet been reported in any other organism. Our extensive search of Par-2 did not find a homologous gene in the Ciona genomic and EST databases. Par-2 may be a nematode-specific gene, or became too diverged during evolution to identify an orthologue in any other animals.

Par-3 and Par-6

Both Par-3 and Par-6 proteins contain PDZ domains. Par-3 and Par-6 form a complex with aPKC, and the complex is necessary for the posterior localization of Par-1 (Rose and Kemphues 1998; Ohno 2001). Par-3 encodes a polypeptide with three PDZ domains. Par-3 homologues have been isolated in D. melanogaster (Bazooka) and vertebrates (ASIP). One candidate of the Ciona Par-3/ASIP homologue was found to have three PDZ domains (Table 1). The alignment of the amino acid sequence of Par-3/ASIP is shown in Fig. S2. Although the first PDZ domain of the Ciona homologue only showed a weak homology with the PDZ consensus sequence, the second and third PDZ domains showed a strong homology. Par-6 has a single PDZ and PB1 domain. The candidate for Ciona Par-6 had the same domain composition (Table 1). A phylogenetic tree was drawn by the amino acid sequence of each PDZ domain of Par-3/ASIP, and the PDZ domain of Par-6 (Fig. S3). Each PDZ domain of human Par-3, Par3-like, D. melanogaster Bazooka, and Ci-Par-3 formed a single group with bootstrap values of 35%, 78% and 68%, respectively. Although the bootstrap value of the first PDZ domain was low, the domain structure, similarity of amino acid sequence, and the high bootstrap values of the other two PDZ domains suggested that Ci-Par-3 is an orthologue of Par-3/ASIP. The PDZ domain of Par-6 constituted a single clade with a bootstrap value of 100%, indicating that Ci-Par-6 is an orthologue of Par-6. The expression of the two genes was confirmed by EST (Table 1).

Par-4

C. elegans Par-4 encodes a S/T-K, necessary for the proper orientation of the spindle (Rose and Kemphues 1998; Ohno 2001). Homologous genes have been identified in Xenopus (Xeek1) and mammals (LKB1). D. melanogaster has a Par-4 homologue (CG9374), the function of which is not known yet. One candidate for the Ciona Par-4 orthologue was retrieved (Table 1), and had corresponding ESTs, supporting the accuracy of the gene prediction and the expression of the gene. A phylogenetic tree drawn by the S/T-K domain of Par-4, with two outgroup kinases (human CamKKb and TSSK1), revealed that Ci-Par-4 is an orthologue of Par-4/LKB1 with a high bootstrap value of 99% (Fig. S4).

Par-5

C. elegans Par-5 encodes a 14-3-3 protein (Morton et al. 2002), which is a highly conserved protein family among eukaryotes (Ferl et al. 2002). It functions in many biological processes, such as response to stress, cell-cycle control and apoptosis. C. elegans Par-5 is necessary for the A-P polarity and the asymmetric cell division of the eggs (Rose and Kemphues 1998; Ohno 2001), suggesting yet another role of 14-3-3 in development.

C. elegans, D. melanogaster and humans have three, two and seven 14-3-3s, respectively. The length and structure of the 14-3-3 protein are well conserved; it has nine α-helices in the central region (core region). To identify the Ciona Par-5 orthologue, we tried to retrieve all 14-3-3 genes in the Ciona genome, and six candidates were found (Table 1). Ci-14-3-3ε-1 was the same as the previously isolated Ciona 14-3-3 gene (AB037679). As shown in Fig. S5, the amino acid sequences of the candidates were similar to the 14-3-3s of other animals. A phylogenetic tree was drawn by comparing the core region sequences (Fig. S6). The tree revealed that Ci-14-3-3ε-1 and Ci-14-3-3ε-2 are grouped with 14-3-3ε with a bootstrap value of 96%. The two genes were aligned in tandem in the Ciona genome. The Ciona protein clustered with C. elegans Par-5 was Ci-14-3-3-a; however, because the bootstrap value was low (55%), the orthologous relationship within the cluster was not so clearly interpreted. Similarly, there was no D. melanogaster and human 14-3-3 that showed a strong correlation with Par-5. The Ciona gene that showed the highest homology to Par-5 was Ci-14-3-3-a, and conversely, the C. elegans gene that showed the highest homology to Ci-14-3-3-a was 14-3-3-like2. Par-5 was the second highest homologous gene to Ci-14-3-3-a by Blast analysis. The D. melanogaster and human genes showing the closest homology to Ci-14-3-3-a were 14-3-3ζ and 14-3-3γ, respectively. It is important to reveal which 14-3-3 is the functional orthologue of Par-5. However, the orthology of the other Ciona 14-3-3s, Ci-14-3-3-b, Ci-14-3-3-c, and Ci-14-3-3-d, was obscure. The accuracy of the gene prediction and the expression of all the 14-3-3 genes of Ciona were confirmed by EST (Table 1).

Pod-1/coronin

Pod-1 was isolated as a causal gene of a Par-like mutant of C. elegans (Rappleye et al. 1999). Pod-1 encodes a polypeptide similar to coronin, a group of conserved proteins among animals. Coronin has two WD40 repeats, and because Pod-1 is composed of two coronin-like regions, Pod-1 has four WD40 repeats. C. elegans, D. melanogaster and humans have both coronin and Pod-1. The Ciona databases were used to find two candidates for coronin and Pod-1 orthologues (Table 1). Proteins encoded by Ci-coronin and Ci-Pod-1 had two and four WD40 repeats, respectively. A phylogenetic tree was drawn with the amino acid sequences of the entire region of coronin, and two coronin-like regions of Pod-1 (Fig. S7). Human, D. melanogaster and C.elegans coronin proteins and Ci-coronin formed a single clade with a bootstrap value of 89%. Ci-coronin was more distant from human coronins than the D. melanogaster and C. elegans counterparts. However, the high bootstrap score (89%) and the similarity of the amino acid sequence, shown in Fig. S8, suggested that Ci-coronin encodes a Ciona coronin orthologue. The N- and C-halves of Pod-1 homologues formed a single clade with bootstrap values of 59% and 78%, respectively (Fig. S7). Ci-Pod-1 was considered an orthologue of Pod-1 from the similarity of the domain structure and sequence homology. The accuracy of the gene prediction and the expression of Ciona coronin and Pod-1 were confirmed by EST (Table 1).

PCP pathway

Rho/Rac/CDC42

Rho/Rac/CDC42 is a group of conserved small G-proteins that function in many aspects of cellular events including actin regulation, vesicle transfer and gene expression (Takai et al. 2001; Aelst and Symons 2002). In mammals, at least 12 Rho type small G-proteins have been identified. In addition to these, human WRCH1 and WRCH2/ARHV were recently reported to be CDC42-like G-proteins that function in the Wnt-JNK pathway and downstream of Wnt-1 (Tao et al. 2001; Katoh 2002).

Our search yielded 12 Ciona genes encoding Rho-like G-proteins (Table 1). Among these, five genes encoding Rac-like proteins (Ci-Rac-a, Ci-Rac-b, Ci-Rac-c, Ci-Rac-like-a, and Ci-Rac-like-b) were aligned in tandem in the Ciona genome. Therefore, these five genes may have been created by the tandem duplication of a single Rac-like gene. The amino acid sequence encoded by Ci-Rac-like-a and Ci-Rac-like-b was shorter than other Rac proteins, and there were no corresponding ESTs. Therefore, the two genes are possibly pseudogenes and were excluded from further analysis. The accuracy of the gene prediction and the expression of the other Ciona Rho candidate genes were confirmed by EST (Table 1).

A phylogenetic tree drawn with almost the full-length of the sequences is shown in Fig. 3 (human Rab1A, H-Ras1, Ran and Arf1 were used as outgroups). There were at least seven subgroups of Rho/Rac/CDC42, and Ciona had orthologous genes in six subgroups. Ci-Rac-a, Ci-Rac-b, and Ci-Rac-c were likely orthologues of the Rac proteins (bootstrap value of 62%). However, the phylogenetic relationship among the Rac proteins of Ciona, humans and D. melanogaster was not evident. Ci-CDC42, Ci-Tc10, Ci-WRCH, Ci-RhoD, and Ci-RhoA/B/C were orthologues of CDC42 (100%), Tc10 (56%), WRCH/ARHV (93%), RhoD (40%) and RhoA/B/C (100%), respectively. There were two more ascidian candidate genes, Ci-Rho/Rac/CDC42-related-a and Ci-Rho/Rac/CDC42-related-b, but their orthologous relationships with known proteins were not clarified. Ciona orthologues of human Rho6/E/N, RhoG, RhoH and D. melanogasterMTL were not found in the Ciona genome. D. melanogaster did not have orthologous genes to Tc10, WRCH and RhoD, all of which are conserved in Ciona and humans.
https://static-content.springer.com/image/art%3A10.1007%2Fs00427-003-0325-9/MediaObjects/s00427-003-0325-9flb3.gif
Fig. 3.

Phylogenetic tree of Rho/Rac/CDC42 proteins, generated by the neighbor-joining method. Ciona proteins are shown by large black dots. The number beside each branch indicates the percentage of times that a node was supported in 1000 bootstrap pseudoreplications. Protein names are explained in Materials and methods. P11476 HS Rab1A, P01112 HS P21/H-Ras-1, P17080 HS Ran, and P32889 HS Arf1 were used as outgroup proteins. The scale bar indicates an evolutionary distance of 0.2 amino acid substitutions per position

PLC

Phosphoinositide-specific phospholipase C (PLC) creates second messenger molecules, IP3 and diacylglycerol, causes Ca2+ release, and activates kinases such as PKC and CamK (Williams 1999). It is suggested that PLC plays a role in the Wnt/Ca2+ pathway, possibly activated by Wnt/Fz (Kuhl et al. 2000).

PLC has two well-conserved domains, X and Y, both of which are a part of the catalytic domain of PLC. Humans have at least 14 proteins with both X and Y domains, and D. melanogaster has three proteins. Some PLCs were grouped as PLC-β, PLC-γ and PLC-δ by their domain structure. PLC-γ is distinct from other PLCs in that it contains two SH2 domains, one PH domain, and one SH3 domain between the X and Y domains.

We found seven Ciona genes encoding proteins with both X and Y domains (Table 1). The domain structure of the protein encoded by Ci-PLC-γ shared the same feature with known PLC-γ proteins. A phylogenetic tree was drawn with the X and Y domain sequences (Fig. S9). Two large groups, PLC-β and PLC-γ, were evident with a high bootstrap value of 100% and 99%, respectively. In contrast, the phylogenetic relationship of the other PLCs was less obvious. Ci-PLC-β4 and Ci-PLC-β1/2/3 were included in the PLC-β subfamily. In addition, there were two subgroups in the PLC-β subfamily, one (bootstrap value 72%) contained human PLC-β4, D. melanogaster NORPA, and Ci-PLC-β4, and the other subgroup (61%) contained human PLC-β1–3, Ci-PLC-β1/2/3, and D. melanogaster PLC21C. Ci-PLC-γ was included in the PLC-γ subfamily, as was revealed by its domain composition. Ci-PLC-L was grouped with human PLC-L and KIAA1092 (94%), whereas no D. melanogaster gene from this group could be found in its genome. The orthologous relationships among the remaining PLC proteins, including Ci-PLC-a, Ci-PLC-b, and Ci-PLC-c, were obscure. The accuracy of the gene prediction and the expression of all Ciona PLC genes were confirmed by EST (Table 1).

CamK

CamK is a group of S/T-Ks that is activated by the Ca2+-calmodulin complex (Soderling 1999). There are three major groups: CamKI, CamKII and CamKIV. CamKII is involved in the Wnt/Ca2+ pathway, and inhibits the canonical Wnt/β-catenin pathway (Kuhl et al. 2001). In addition, there are several groups of S/T-Ks with a kinase domain similar to CamKs, such as human PSKHs, CASK, D. melanogaster DCAMKL1, and Caki. We tried to identify all ascidian orthologues of these proteins.

Seven candidate genes were retrieved from the databases (Table 1), and a phylogenetic tree was drawn with the S/T-K domain sequences (Fig. S10). The tree revealed that Ci-CAMKI is an orthologue of CamKI (bootstrap value of 100%), Ci-PSKH of PSKH1/2 (100%), Ci-CAMKIV of CamKIV (99%), Ci-CASK of CASK/Caki (100%), and Ci-CAMKII of CamKII (100%). Ci-CAMK-like-a and Ci-CAMK-like-b were included in a group of CamK-like kinases (100%) whose functions have not yet been determined. Orthologues of PSKH and CamKIV were not found in the D. melanogaster genome. The accuracy of the gene prediction and the expression of all Ciona CamK genes were confirmed by EST (Table 1).

Strabismus/Van Gogh (Stbm)

Stbm is a novel type of transmembrane protein with four transmembrane domains (Darken et al. 2002). Homologous genes are known in D. melanogaster and in vertebrates, such as the C. elegans gene, B0410.2, that encodes a Stbm-like protein. In Ciona, we found a gene that encoded a protein with four transmembrane domains, and with a strong homology to Stbm (Table 1). The similarity of the amino acid sequences, shown in Fig. S11, suggests that Ci-Strabismus/Van Gogh is a Ciona orthologue of Stbm.

Prickle

Prickle (Pk) is a protein with three LIM domains. Pk proteins were identified in D. melanogaster, Xenopus, humans and Ciona, some of which exhibit functions in the PCP pathway (Tree et al. 2002; Wallingford et al. 2002b). Ciona intestinalis Pk, Ci-Pk1 and Ci-Pk2 have been isolated as downstream genes of Ciona Brachyury (Hotta et al. 2000). Ci-Pk1 and Ci-Pk2 were expected as the products of alternative splicing. Our search of the Ciona genome sequence revealed that Ci-Pk1 and Ci-Pk2 are encoded by a single gene, Ci-Prickle, and they are products of alternative splicing. There was another gene in the Ciona genome whose coding polypeptide contained three LIM domains (Ci-Triple-lim). However, the similarity between Ci-Triple-lim and the known Pks was very low. A phylogenetic tree revealed that Ci-Prickle is the only Pk gene in Ciona, shown by a high bootstrap value of 100% (Fig. S12). The Ci-Triple-lim branch was outside the Pk family of proteins.

Daam-1/Formin

Daam-1 is a recently identified Formin-homologue protein that functions in the convergent extension in Xenopus embryos (Habas et al. 2001; Myers et al. 2002). Daam-1 is involved in the activation of Rho by Wnt/Fz. Other Formin homologues also function in the Rho cascade and in actin regulation, such as the mammalian Formin homologue Diaphanous (Dia), which regulates actin polymerization when activated by Rho (Takai et al. 2001). Similarly, in D. melanogaster, Dia functions in the PCP pathway as a downstream factor of Rho (Aelst and Symons 2002). The D. melanogaster Formin-homologue cappuccino (cappu) has been isolated as a causal gene of a maternal mutant showing defects in the polarity of eggs (Emmons et al. 1995). Mammals have at least 14 Formin-homologous proteins, including Formin-like (FMNL), Daam-1, Dias, Formin homolog overexpressed in spleen (FHOS), and Formins (FMNs). D. melanogaster has at least six Formin-homologues.

Formin proteins contain two conserved domains, FH1 and FH2, from their N-terminal. FH1 is a proline-rich domain, and FH2 acts as an actin-binding domain. We found eight candidates of Ciona Formin-homologues (Table 1), and almost all of them, except Ci-Formin/Cappuccino-b, encoded a protein with a FH1 and a FH2 domain, suggesting that they are Formin-homologues. Ci-Formin/Cappuccino-b had a very short proline-rich domain and we could not judge whether this region is FH1 or not.

A phylogenetic tree of Formin-homologues constructed with the FH2 domain sequences is shown in Fig. S13. Formin-homologues were divided into at least five groups: Formin-like, Daam-1, Dia, FHOS, and Formin/Cappu. Ci-Formin-like was orthologous to human FMNL and its related proteins (bootstrap value of 100%), Ci-Daam-1 to Daam-1 (100%), Ci-Diaphanous to Dia (100%), and Ci-FHOS to FHOS (100%). Two Ciona genes, Ci-Formin/Cappuccino-a and Ci-Formin/Cappuccino-b, were orthologous genes of Formins and Cappuccino (98%). Ci-Formin/Cappuccino-a was closer to human Formins, whereas Ci-Formin/Cappuccino-b was a diversified gene in Ciona. Ci-Formin/Diaphanous/Daam-1-related-a and Ci-Formin/Diaphanous/Daam-1-related-b were clustered with the uncharacterized Formin-homologues human KIAA1727 and D. melanogaster CG18138 (100%).

An EST corresponding to Ci-Formin/Cappuccino-b did not exist, and thus it was not revealed whether this gene is active or not. Other Ciona Formin-homologues were correctly predicted and were expressed, as was revealed by the corresponding ESTs (Table 1).

WASP/WAVE/SCAR cascade

WASP

WASP has been identified as the causal gene of Wiskott-Aldrich syndrome (Erickson and Cerione 2001). WASP encodes a protein with two conserved domains, WH1 and WH2. Mammalian WASP and neural WASP (N-WASP) are activated by CDC42 to regulate F-actin formation (Rohatgi et al. 1999; Suetsugu et al. 2002). There are several proteins with a WH1 domain, such as Homer and ENA, although neither has a WH2 domain.

To identify a WASP orthologue, we retrieved all Ciona genes that encode proteins with WH1. Four candidate genes were found in the Ciona genome (Table 1). Among these, proteins encoded by Ci-WASP had one WH1 and WH2 domain, suggesting that Ci-WASP is a WASP orthologue. A phylogenetic tree drawn with WH1 domain sequences also suggested that Ci-WASP is a WASP orthologue, with a bootstrap value of 100% (Fig. S14). The other Ciona genes were orthologues of Homer (Ci-Homer), VASP/ENA (Ci-VASP/ENA), an unknown human gene FLJ33903, ENSP00000305579, and D. melanogaster CG10155 (Ci-WH1). The accuracy of the gene prediction and the expression of Ciona genes were confirmed by EST (Table 1).

WAVE/SCAR is a group of WASP-like proteins with a function in actin dynamics regulated by Rac (Miki et al. 1998). WAVE/SCAR has a WH2 domain near the C-terminus. The Ciona databases were searched and one candidate gene encoding a protein resembling WAVE/SCAR was found (Fig. S15). However, Ci-WAVE/SCAR did not have a WH2 domain, possibly because the Scaffold on which Ci-WAVE/SCAR is located is too short to cover the WH2-containing C-terminal half. In addition, there was also no EST for Ci-WAVE/SCAR with a WH2 domain. The Ciona gene that showed the closest similarity to WAVE/SCAR was Ci-WAVE/SCAR, and the human and D. melanogaster genes that showed the closest similarity to Ci-WAVE/SCAR were WAVE3 and SCAR, respectively. Therefore, Ci-WAVE/SCAR is the most likely candidate for the WAVE/SCAR orthologue.

ARP/ARPC

Actin-related protein complex (ARPC) is a conserved protein complex that functions in actin regulation with WASP (Takai et al. 2001; Suetsugu et al. 2002). Mammalian ARPC contains at least eight proteins, coronin, Arp2, Arp3, p16, p20, p21, p34, and p41 (Machesky et al. 1997). Some of these have homologues of yeast and C. elegans.

Arp2 and Arp3 are actin-like proteins; therefore, to identify Ciona Arp2/3 orthologues, we searched for all Arp2/3 and similar proteins. In the human genome, there are eight actin-like proteins, including Arp1–3, 3β and 11, and D. melanogaster has at least six actin-like proteins. In Ciona, we obtained four candidate genes encoding a protein similar to, but different from actin itself (Table 1). A phylogenetic tree was drawn with an almost full length of these proteins. Ciona muscle actin Ci-MA2 and cytoplasmic actin Ci-CA7 (Chiba et al. 2003) were used as outgroups (Fig. S16). This tree revealed that Ci-ARP1 is orthologous to human Arp1 and D. melanogaster Actz (bootstrap value of 99%), Ci-ARP2 to human Arp2 and D. melanogaster Arp14D (100%), and Ci-ARP3 to human Arp3, Arp3β, Arp11, FKSG72 and D. melanogaster Arp66B (100%). Ci-actin-like-a was clustered with the uncharacterized human protein FLJ20537 and D. melanogaster CG12235 (100%). A Ciona gene encoding a protein orthologous to human FLJ12934 or D. melanogaster CG7846 was not found. This was the only group in which the gene is conserved in humans and D. melanogaster but not in Ciona. An orthologue of Ci-actin-like-b was not found in other animals. This gene was located next to the tandem repeats of several actin genes in the Ciona genome. Therefore, it is possible that this gene is a duplicated, highly mutated Ciona actin gene. All of the Ciona actin-like genes were correctly predicted and were expressed, as was confirmed by EST (Table 1).

Next, the remaining proteins included in the ARPC were searched for, and p16, p20, p21, p34, and p41 homologous genes were found in the Ciona databases (Table 1). We concluded that they are orthologues of p16–p41, for two reasons. The first reason is the similarity of the sequences (Fig. S17), and the second reason is the fact that they all showed a strong similarity to the known p16–p41, and vice versa.

ADF/cofilin regulation cascade

ADF/cofilin

ADF/cofilin is a highly conserved actin regulator found throughout eukaryotes (Maciver and Hussey 2002). ADF/cofilin is regulated by its state of phosphorylation. When ADF/cofilin is not phosphorylated, it is active to depolymerize actin fibers. When ADF/cofilin is phosphorylated by S/T-Ks, such as LIMK and TESK, it is inactivated.

Humans have three ADF/cofilin proteins: ADF, muscle cofilin, and non-muscle cofilin, and D. melanogaster has one, twinstar. The Ciona gene that showed the closest similarity to ADF/cofilin was Ci-ADF/cofilin (Fig. S18). In addition, Blast analysis showed that human and D. melanogaster genes with the closest similarity to Ci-ADF/cofilin were cofilin2 and twinstar, respectively. These facts suggest that Ci-ADF/cofilin is a Ciona orthologue of ADF/cofilin. Ci-ADF/cofilin was too diversified to reveal its relationship with the three human ADF/cofilins (Fig. S19). In ADF/cofilin genes the first amino acid is encoded by an exon separated from other coding regions; however, this characteristic was not conserved in Ciona.

LIMK/TESK/ROCK

LIMK and TESK are S/T-Ks that directly phosphorylate ADF/cofilin (Maciver and Hussey 2002). LIMK also contains one LIM domain. TESK contains only a S/T-K domain, similar to that of LIMK. Humans have two LIMKs and two TESKs, and D. melanogaster has one LIMK and one TESK. One candidate for Ciona LIMK, that has a LIM and S/T-K domain, and one candidate for Ciona TESK, were found in the Ciona databases (Table 1).

Another S/T-K, ROCK, is a target of Rho, and an upstream regulator of LIMK. It contains a S/T-K domain, a coiled-coil structure, a HR1 domain, a PH domain, and a C1 domain from its N-terminus to its C-terminus. Mammals and D. melanogaster have two ROCKs and one ROCK (Drok), respectively. One candidate for the Ciona ROCK orthologue was found, sharing almost all domains, except HR1, with other ROCKs (Table 1).

A phylogenetic tree of these kinases with their S/T-K domains was constructed (Fig. S20). It was revealed that Ci-TESK is orthologous to TESK (bootstrap value of 99%), Ci-LIMK to LIMK (98%), and Ci-ROCK to ROCK (100%), as predicted by their domain compositions. These genes were accurately predicted and were expressed, as confirmed by EST (Table 1).

SSH

SSH was recently identified as the strongest candidate for ADF/cofilin phosphatase in D. melanogaster and humans (Niwa et al. 2002). In both species, SSH has two domains (domain A and B) and one catalytic domain. The phosphatase most similar to SSH is MAPK phosphatase (MAPKP), but it does not have domains A or B. One candidate for the Ciona ssh gene was found (Table 1). Ci-SSH encoded a polypeptide that had a domain A. Domain B of this Ciona product was less evident than the other SSHs. A phylogenetic tree was drawn with phosphatase catalytic domain sequences. Human MAPKP-1 and Ci-DUSP1/2/4/5 (Satou et al. 2003b) were used as outgroups. The tree indicated that Ci-SSH is a Ciona ssh orthologue with a high bootstrap value of 100% (Fig. S21). The accuracy of the gene prediction and the expression of Ci-SSH were confirmed by EST (Table 1).

PAK

p21-activated kinase (PAK) is a group of S/T-Ks that is activated by CDC42 and Rac. PAK is involved in the regulation of ADF/cofilin through the activation of LIMK, suggesting a role in actin dynamics (Jaffer and Chernoff 2002; Maciver and Hussey 2002). PAKs are divided into two groups, group 1 and group 2, by their domain structure. Group 1 contains a p21 binding domain (PBD) and an auto-inhibitory domain (AID), together with a S/T-K domain, whereas group 2 does not contain an AID. In mammals, six PAKs, PAK1–3 (group 1) and PAK4–6 (group 2), have been reported, and D. melanogaster has three PAKs, DPAK, STE20-like, and PAK3. From the Ciona databases two PAK candidates were found, both of which contained a PBD domain (Table 1). A phylogenetic tree was drawn with the S/T-K domain sequences (Fig. S22), and Ciona aPKC and human PKC-α were used as outgroups. The tree revealed that Ci-PAK1/2/3 is included in group 1, which includes human PAK1–3 and D. melanogaster DPAK with a bootstrap value of 100%. Ci-PAK4/5/6 is a member of group 2 (100%), which includes human PAK4–6 and D. melanogaster STE20-like. These results suggest that the last common ancestor of protostomes and deuterostomes had a group 1 and a group 2 PAK. It was predicted that two PAK genes from Ciona would be transcribed, and was confirmed by EST (Table 1).

Conclusions

Our present study revealed the Ciona intestinalis components of aPKC-Par and PCP pathways, and WASP/WAVE/SCAR and ADF/cofilin cascades, which indicated an ascidian genomic feature with less redundancy than vertebrates (Fig. 1). The conservation of each component suggests the conserved pathways/cascades also function in Ciona development. Further study is needed as to when and where these pathways/cascades are used in Ciona development. Ciona intestinalis would be an ideal model organism for such a study because of the compact nature of its genome. The present study also found that the orthologous proteins of five gene groups are conserved in humans and Ciona, but not in D. melanogaster. In contrast, only one gene group is conserved in humans and D. melanogaster, but not in Ciona, suggesting that Ciona is a better system for understanding vertebrate physiology. Ciona intestinalis, therefore, provides refined cues for the study of vertebrate development and evolution.

Acknowledgements

This research was supported by Grants-in-Aid for Scientific Research from MEXT, Japan to Y. Satou (13044001) and N.S. (12202001). Y. Sasakura was a Postdoctoral Fellow of JSPS with a research grant (14000967). We thank Kazuko Hirayama, Chikako Imaizumi, Asako Fujimoto, and Hisayoshi Ishikawa for their technical support.

Supplementary material

427_2013_325_MOESM1_ESM.zip (465 kb)
Supplementary material, approximately 476 KB.

Copyright information

© Springer-Verlag 2003