Centrin-2 (Cetn2) mediated regulation of FGF/FGFR gene expression in Xenopus

Centrins (Cetns) are highly conserved, widely expressed, and multifunctional Ca2+-binding eukaryotic signature proteins best known for their roles in ciliogenesis and as critical components of the global genome nucleotide excision repair system. Two distinct Cetn subtypes, Cetn2-like and Cetn3-like, have been recognized and implicated in a range of cellular processes. In the course of morpholino-based loss of function studies in Xenopus laevis, we have identified a previously unreported Cetn2-specific function, namely in fibroblast growth factor (FGF) mediated signaling, specifically through the regulation of FGF and FGF receptor RNA levels. Cetn2 was found associated with the RNA polymerase II binding sites of the Cetn2-regulated FGF8 and FGFR1a genes, but not at the promoter of a gene (BMP4) whose expression was altered indirectly in Cent2 morphant embryos. These observations point to a previously unexpected role of Cetn2 in the regulation of gene expression and embryonic development.

Scientific RepoRts | 5:10283 | DOi: 10.1038/srep10283 transport of RNAs and proteins 3 . It is clear that Cetn2 and Cetn3 have distinct functional roles in a number of systems.

Results and Discussion
The idiosyncratic aspects of X. laevis development can reveal gene functions hidden in other organisms 17 . We therefore set out to explore the roles of Cetns in early X. laevis development. Both X. laevis and X. tropicalis have multiple centrin genes, based on data accessed through Xenbase 18 . The gene/protein originally designated as Centrin (X. laevis: NCBI Reference Sequence: NP_001081398.1 and X. tropicalis: NP_001016387.1) or Centrin-1 (X. laevis: NCBI Reference Sequence: NP_001080127.1) display a Cetn2-like, rather than a Cetn1-like, genomic structure (see supplementary figure 1); we therefore refer to them as Cetn2 (see below). No Cetn1-like gene appears to be present in either X. laevis or X. tropicalis genomes. The Cetn3 and Cetn4 genes identified in X. laevis are similar in genomic structure to those found in mouse and human.
The latest version of the X. laevis genome (7.1 as searched through the Xenbase Blast function in March 2015) reveals two distinct Cetn2 (Cetn2a and Cetn2s) and Cetn3 (Cetn3l and Cetn3s) genes, and apparently a single Cetn4 gene. Our studies focus on the Cetn2a, Cetn3l, and Cetn4 genes. Cetn2a corresponds to the 172 amino acid polypeptide labeled cetn1 or centrin (see above); Cetn3l corresponds to the 167 amino acid polypeptide labeled Cetn3 (GenBank: AAI29791.1). We isolated full length cDNAs that correspond to Cetn2a, Cetn3l, and Cetn4. An analysis of gene expression during early Xenopus embryogenesis by Yanai et al. 19 (Fig. 1A) and our own RT-PCR data in X. laevis (Fig. 1B) indicate that Cetn2a, Cetn3l, and Cetn4 RNAs are supplied maternally and are present at high levels throughout early development; we have not directly examined the expression levels of the Cetn2s or Cetn3s genes.
We used two different antibodies to localized Cetn proteins in X. laevis. The first is a rabbit antibody (anti-XlCetn) 20 that recognizes Cetn2, Cetn3, and Cetn4 proteins based on its recognition of Cetn2a, Cetn3l, and Cetn4 polypeptides expressed from injected RNAs ( Fig. 1C and data not shown). The second antibody is a commercially available rabbit antibody (anti-HsCetn1) that reacts preferentially with X. laevis Cetn2a compared to Cetn3l (Fig. 1C). Because the anti-HsCetn1 antibody produced higher overall background labelling, we used the anti-XlCetn antibody for most staining studies. Both anti-Cetn antibodies stain the basal body region of epidermal ciliated cells (Fig. 1D-F). There is also discernible staining of the myotome; neuronal microtubules are not stained (Fig. 1G-J -anti-XlCetn staining shown). Basal body localization of all three Cetns was confirmed using C-terminally GFP tagged forms of Cetn2a, Cent3l, and Cetn4 expressed from injected RNAs (see below).
To down-regulate the levels of specific Cetn proteins in embryos we commissioned Gene-Tools LLC to design anti-sense translation blocking modified DNA oligonucleotides (morpholinos or MOs) specific for Cetn2a, Cetn3L, and Cetn4 RNAs (see supplemental figure. 2). The Cetn2s gene encodes a 201 amino acid long polypeptide that differs from Cetn2a primarily by the presence of a 29 amino acid insertion at its N-terminus. Similarly the Cetn3S gene encodes a 212 amino acid long polypeptide that differs from Cetn3l primarily by the presence of a 32 amino acid insertion at its N-terminus. Given the dramatic nature of the Cetn2 morphant phenotype (see below) we commissioned a second, completely non-overlapping anti-Cetn2 morpholino (Cetn2-MO2). The translation of the Cetn2s RNA is not expected to be altered by either of the Cetn2a morpholinos used in our studies; similarly expression of Cetn3s is not expected to be altered by the Cetn3l morpholino. Because of its dramatic nature, we concentrate the studies described here on the characterization of the Cetn2a morphant phenotype. Plasmids that encode versions of Cetn2a-GFP, Cetn3l-GFP, and Cetn4-GFP that contain sequences that match the Cetn2a-MO1, Cetn3l-MO, and Cetn4-MO morpholinos perfectly were created for specificity and rescue studies. The Cetn2-MO1 and Cetn2-MO2 morpholinos reduced the level of Cetn2 protein but not Cetn3 or Cetn4, while the Cetn3 and Cetn4 morpholinos specifically reduced the accumulation of their targets (Fig. 2).
In ectodermal explants, a single Cetn morpholino alone (10 ngs/embryo) did not dramatically or reproducibly disrupt the formation of cilia in multiciliated cells; cilia formation did appear to be disrupted when multiple Cetn morpholinos were used together (supplemental figure. 3 and data not shown). Cetn2 morphants exhibited embryonic phenotypes unlike those associated with a typical ciliopathy; they appeared similar to those associated with defects in FGF-mediated mesoderm formation 21 . This phenotype was distinct from that displayed by Cetn3 and Cetn4 morphants (supplemental figure. 3 and data not shown) as well as in Chibby (Cby) morphants (Cby is a basal body protein associated with the regulation of Wnt signaling) 22 .
To test whether Cetn2 morpholinos disrupted FGF signaling, we used ectodermal and mesodermal explants, as described previously 23 . In culture, such explants normally elongate and form notochordal tissue within the mesodermal domain (Fig. 3A). When the mesodermal domain was taken from a Cetn2 morphant embryo, elongation was inhibited and notochordal tissue failed to form (Fig. 2B). When the ectodermal region was derived from a Cetn2 morphant, morphological extension was somewhat suppressed but notochordal tissue formed (Fig. 3C). Similar results were obtained using mesoderm only explants. Again, control explants displayed an extended morphology (Fig. 3D) and formed notochordal tissue (Fig. 3I). Cetn2 morphant explants failed to elongate (Fig. 3E) or form notochord (Fig. 3J). The morphological extension and notochord phenotypes displayed by Cetn2 morphant explants were rescued by Cetn2 (Fig. 3G,K) but not by Cetn3 (Fig. 3L) RNA injection. The Cetn2 morpholino elongation defect was similar to that seen in explants derived from embryos injected with RNA encoding a dominant negative form of FGF Receptor 1 (dnFGFR1) (Fig. 3F). That the Cetn2 morphant phenotype involves effects on FGF signaling was further suggested by the fact that the phenotype could be rescued by the injection of FGF8 RNA (Fig. 3M).
Together with retinoic acid, Wnt, FGF, and BMP are three of the most prominent signaling pathways involved in early embryonic patterning, including mesoderm/notochord formation 24 . A preliminary RT-PCR analysis of the levels of Wnt8a, FGF8 and BMP4a RNAs in stage 11 control and Cetn2 morphant embryos revealed a dramatic decrease in FGF8 and an increase in BMP4a RNA levels, with no apparent change in Wnt8a RNA levels (Fig. 4A). This is a pattern of changing RNA levels quite distinct from that observed in ectodermal explants and whole embryos injected with morpholinos directed against either of two other cilia/basal body-associated proteins, Cby 22 and EFHC1 (Zhao et al., in preparation). We

Figure 2. A:
Embryos were injected into both cells at the two cell stage with RNAs encoding GFP (150 pgs per embryo) either alone or together with Cetn2MO1, Cetn2MO2, or Cetn3MO (10 ngs/side, 20 ngs total per embryo); at stage 11 or 25 the embryos were analyzed by SDS-PAGE and immunoblot using the anti-Human Cetn-1 antibody (which reacts preferentially with Cetn2 compared to Cetn3. There was a clear decrease in Cetn2 protein levels, persisting through stage 25. To confirm the specificities of the Cetn MOs both blastomeres of two cell embryos were injected with RNAs encoding GFP (200 pg/side) and RNAs encoding Cetn2a-GFP (B), Cetn3l-GFP (C), or Cetn4-GFP (D) RNAs with ("+ ") or without Cetn MO (10 ng/side). These Cetn RNAs contain the target sequence of the corresponding morpholino. In addition, uninjected ("UN") and embryos injected with GFP RNA alone were examined as controls for antibody specificity. Injected embryos were harvested at stage 11. Immunoblot analyses were carried out using an anti-rabbit GFP antibody. An apparent breakdown product of the Cetn2-GFP construct is indicated by the arrow in the Cetn2 MO panel. confirmed this result using quantitative reverse transcription PCR (qPCR) of Cetn2 morphant embryos prepared by injecting either the MO1 or MO2 Cetn2 morpholinos. These morpholinos decreased the levels of FGF2, FGF4, FGF8, FGFR1a, and FGFR1b RNAs but produced no significant change in FGFR2, FGFR3, or FGFR4a RNAs. The Cetn3 MO did not change the level of any FGF or FGFR RNA examined (Fig. 4B,C). In rescue studies, co-injection of Cetn2-GFP RNA could return the levels of FGF8 and FGFR1a to control levels, while Cetn3-GFP and Cetn4-GFP RNAs could not (Fig. 4D).
FGF and BMP have been found to regulate each other's expression in the early embryo 25 . To determine whether the effects of Cetn2 morpholinos were direct or indirect, we injected RNA encoding the BMP antagonist Noggin 26 into Cetn2 morphant embryos. Noggin RNA reduced the Cetn2 MO induced increase in the level of BMP4a RNA, but did not increase the level of FGF8 RNA (Fig. 4E). In contrast, the injection of FGF8 RNA reduced the Cetn2 MO induced increase in the level of BMP4a RNA, but failed to rescue the levels of FGFR1a RNA (Fig. 4F). This suggests that Cetn2 plays a direct role in the regulation of FGF8 and FGFR1a/b, while its effects on BMP4a RNA levels are indirect and the result of changes in FGF signaling.
Given Cetn2's established presence in the nucleus (see above) and the effects of the Cetn2 morpholinos on FGF8 and FGFR1a RNA levels, it seemed plausible that Cetn2 might directly influence the expression of these genes through interactions with chromatin. We initially examined the FGF8 gene in X. tropicalis using anti-XlCetn, anti-HsCetn1 (which preferentially reacts with Cetn2 over Cetn3 -see above), and anti-RNA polymerase II antibodies. We found that Cetn co-localized with polymerase at these loci (Fig. 5A). Further experiments in X. laevis used RNAs encoding GFP-tagged forms of X. laevis Cetn2a, Cetn3l, and Cetn4 (see Methods). In ectodermal explants derived from RNA injected fertilized eggs, myc-Cetn2-GFP (Fig. 5B,C) and Cetn4-GFP (SupFig. 4) polypeptides preferentially accumulated in ciliated cells and localized to the basal body region of cilia (Fig. 5D,E). myc-Cetn3-GFP was also found to accumulate preferentially in ciliated cells (Fig. 5F,G) and was localized to basal bodies (Fig. 5H,I), but its localization was not quite as cilia-specific as that observed for Cetn2. That said, in ectodermal  It is readily apparent that both myc-Cetn2-GFP and myc-Cetn3-GFP polypeptides accumulate in ciliated cells. In explants expressing both myc-Cetn3-GFP (J) and Cetn2-RFP (K; overlap in panel L, panel M is ATT staining), there was both extensive overlap in the localization of Cetn2 and Cetn3 polypeptides (arrow marked "2 + 3"), as well as sites where one or the other predominates (arrows marked either "2" or "3"). For ChIP studies in X. laevis, both blastomeres of 2-cell stage embryos were injected with RNAs encoding either GFP, myc-Cetn2-GFP, or myc-Cetn3-GFP; uninjected embryos were used as a control. Embryos were harvested at stage 11. GFP antibody was used to immunoprecipitate the injected embryos and Pol II antibody was used to immunoprecipitate the uninjected embryos. qPCR analysis was performed to check protein binding to the FGFR1a (N), FGF8 (O) and BMP4a (P) promoter regions.
Our work reveals a new and unexpected role for Cetn2 as a transcriptional regulator of a subset of FGF and FGFR genes. The fact that Cetn2 was found to co-localize with Pol II at some, but not all genes suggests that its promoter association is dependent upon the presence of other proteins. In addition to its nuclear role as part of the XPC DNA repair complex, Cetn2 has been reported to be an integral component of the Trex2 complex, which appears to interact with both nuclear pores and RNA polymerase and has been implicated in the nuclear export of mRNAs 27,28 . This suggests that the Cetn2 morphant phenotype could, in part, involve changes in Trex2 function or other, as yet unidentified interactions. To resolve this issue, we are currently in the process of identifying Cetn2-associated proteins in Xenopus and other systems.
The role of Cetn2 as a regulator of mesodermal differentiation, as revealed by notochord formation, in Xenopus early embryonic development may seem at odds with the reported phenotypes of morpholino treated Zebrafish embryos and Cetn2 null mice. In the Zebrafish Danio rerio, depletion of Cetn2 led to a ciliopathy-related cyst formation phenotype 11 . A syndromic ciliopathy, including dysosmia and hydrocephalus was reported in Cetn2 null mice 12 . We anticipate that the phenotypic differences between these three vertebrates can be attributed to differences in developmental mechanisms and Cetn2-containing complexes, the patterns of centrin gene expression during embryonic development, and perhaps some level of partial redundancy between the centrin genes, although neither Cetn3 or Cetn4 RNAs rescued the Cetn2 morphant gene expression phenotype. That said, a role of Cetn2 in the regulation of gene expression, and its integral role in the functions of XPC, Trex2, and perhaps other complexes, indicates the need for more subtle analyzes of the physiological roles of Cetn2 in particular and Cetns as a class of proteins.

Materials and Methods
Embryos, their manipulation and analysis. X. laevis embryos were staged, and explants were generated, following standard procedures 29 . Capped mRNAs were transcribed from linearized plasmid templates using mMessage mMachine kits (Ambion) following manufacturer's instructions. At the two-cell stage embryo injections were directed equatorially. As an injection tracer, we routinely included RNAs (150 pgs/embryo) encoding either β -galactosidase, green fluorescent protein (GFP) or GFP-CAAX, which is membrane-associated. In the case of GFP/GFP-CAAX RNA injection, embryos were examined at stage 10-11 by fluorescent microscopy to confirm the accuracy of injection. RNA isolation, cDNA synthesis, RT-PCR and qPCR analyses were carried out as described previously 22,23 . Real-time (quantitative) PCR was carried out using a Mastercycler Epgradient Realplex device (Eppendorf). PCR reactions were set up using DyNAmo SYBR Green qPCR kits (Finnzymes). Each sample was normalized to the expression level of ornithine decarboxylase (ODC). The cycling conditions used were: 95°C for 5 minutes; then 40 cycles of 95°C for 15 seconds, 56°C for 15 seconds, 60°C for 30 seconds. The Δ Δ CT method was used to calculate real-time PCR results. The primers used for RT-PCR analysis were Ornithine decarboxylase (ODC) [U 5′ -CAG CTA GCT GTG GTG TGG-3′ D 5′ -CAA CAT GGA AAC TCA CAC-3′ ]; Wnt8a  [U 5′ -TGA TGC CTT CAC TTC TGT GG-3′ D 5′ -TCC TGC AGC TTC TTC TCT  Morpholinos and plasmids. Cetn coding sequences were isolated from maternal RNA by RT-PCR and subcloned into pCS2 plasmids, following protocols used previously to isolate Cby coding sequences 22 ; where indicated such constructs had N-terminal myc and C-terminal GFP sequences 30 . We also generated plasmids that encode Cetn-GFP chimeras that either perfectly matched or were maximally mismatched to their respective morpholinos. Plasmids encoding either N-or C-terminally tagged Cetn2-RFP were obtained from Sergie Sokol (Mount Sinai School of Medicine) and John Wallingford (U. Texas, Austin). Plasmids encoding FGF8 and a dominant-negative form of FGFR1 were supplied by Enrique Amaya (U. Manchester). Morpholinos against X. laevis centrins were designed and synthesized by Gene Tools. These included two non-overlapping morpholinos against Cetn-2a [MO1 5′ CTTGTAGTTAGAAGCCATATCACAC 3′ and MO2 5′ TGCACACACCAACCTTCGACCTCGC 3′ ], a Cetn-3l morpholino [5′ CATCAGTCCTCACAGCCAGGCTCAT 3′ ], and a Cetn-4 morpholino