Insulin Receptor-Related Receptor Regulates the Rate of Early Development in Xenopus laevis

The orphan insulin receptor-related receptor (IRR) encoded by insrr gene is the third member of the insulin receptor family, also including the insulin receptor (IR) and the insulin-like growth factor receptor (IGF-1R). IRR is the extracellular alkaline medium sensor. In mice, insrr is expressed only in small populations of cells in specific tissues, which contain extracorporeal liquids of extreme pH. In particular, IRR regulates the metabolic bicarbonate excess in the kidney. In contrast, the role of IRR during Xenopus laevis embryogenesis is unknown, although insrr is highly expressed in frog embryos. Here, we examined the insrr function during the Xenopus laevis early development by the morpholino-induced knockdown. We demonstrated that insrr downregulation leads to development retardation, which can be restored by the incubation of embryos in an alkaline medium. Using bulk RNA-seq of embryos at the middle neurula stage, we showed that insrr downregulation elicited a general shift of expression towards genes specifically expressed before and at the onset of gastrulation. At the same time, alkali treatment partially restored the expression of the neurula-specific genes. Thus, our results demonstrate the critical role of insrr in the regulation of the early development rate in Xenopus laevis.


Introduction
Receptor tyrosine kinases are among the most important members of the cell signaling system. They play a key role in the organism development and functioning, and participate in the regulation of the cell interactions, cell proliferation and differentiation, cell migration and metabolism, and control of the cell cycle. To activate most of the receptor tyrosine kinases, a ligand of protein or peptide nature is necessary, which is capable of binding to the receptor simultaneously in several sites, causing its dimerization and activation, resulting in autophosphorylation of the intracellular part of the receptor. An exception is the insulin receptor family, whose members are dimerized by disulfide bonds in an inactive state, and interaction of their extracellular part with the ligand leads to conformational changes followed by their activation [1]. This family includes three receptors: insulin receptor (IR), insulin-like growth factor receptor (IGF1R) and insulin receptor-related receptor (IRR). For the IRR, unlike its homologs, ligands of peptide-protein nature were not found [2]. All three receptors are highly homologous and, presumably, have similar mechanisms of functioning. Unexpectedly, we found that human IRR can be activated by alkaline extracellular medium [3,4]. Activation of IRR by alkali is specific, reversible, and dose-dependent [4]. IRR orthologs from frog and mouse also can be activated by alkaline medium. Activation of endogenous IRR by alkali in insulinoma MIN6 cells leads to IRS-1 phosphorylation and cytoskeleton remodeling in these cells [5].
Expression of the insrr in mice was detected in the stomach, kidneys, and pancreas in adult animals. All these organs are involved in excretion or secretion of fluids with an alkaline pH [2]. Mice with insrr knockout did not show gross phenotype under normal conditions [6]. However, the in vivo analysis of insrr knockout mice phenotype under experimentally induced alkalosis revealed the role of IRR as a regulator of bicarbonate excretion in the kidneys [4,7]. Also, insrr is expressed in the alpha and beta cells of the islets of Langerhans, in enterochromaffin-like cells of the stomach, and in certain types of neurons [6,[8][9][10][11][12][13][14]. The expression of insrr, as well as other members of the IR family (insr and igf1r), was found at the stage of unicellular mouse embryos and blastocysts [15]. Lately, IRR is found to be a potential therapeutic target for heart cardioprotection [16]. Our recent studies have revealed a possible role for the receptor tyrosine kinase IRR in the development of mouse preimplantation embryos. Using the Mouse Embryo Assay (MEA) we showed that blastocyst yield for insrr knockout animals was significantly lower than for wild-type mice: 6.7% and 43.8% of the total number of isolated zygotes, respectively [17].
Based on phylogenetic analysis of the IR family sequences, insrr is present in amphibians remaining conserved with the mammalian orthologs [18]. According to the transcriptome analysis, insrr ortholog in Xenopus laevis and Xenopus tropicalis frogs is highly expressed in embryogenesis before neurulation [19,20]. One gene from the IR family was found in flies and worms, and only two orthologs of IR and IGF1R were found in fish [18,21]. In the present work, we studied the physiological role of insrr in early development of Xenopus laevis using morpholino-mediated knockdown. As a result, we showed that downregulation of insrr leads to delay in development, accompanied by a shift in gene expression towards earlier expressing genes. At the same time, incubation of the embryos in the alkaline medium eliminated these effects.

Insrr Knockdown Elicits Retardation of Embryonic Development
To understand the role of insrr in Xenopus laevis development, we downregulated its expression by injecting morpholino antisense oligonucleotide (insrr MO1) into early embryos ( Figure 1). As a result, a significant retardation of development was observed in these embryos when comparing to their siblings injected with the control MO ( Figure 1A,B). The first signs of such retardation became visible only during neurulation, when retardation of the neural tube closure was seen in embryos injected with insrr MO1 compared to control siblings. Interestingly, most of the embryos with downregulated insrr and demonstrated developmental delay between the late neurula stage and tadpole stages 28-32, then gradually became equal in appearance with the control siblings ( Figure 1A).
Since IRR is known as a receptor for which activity is modulated by alkaline pH, we decided to test effects of alkaline medium on the development of embryos with downregulated insrr. To this end, we incubated embryos injected with insrr MO1 and with the control MO in solutions with pH 7.2 and 8.5. As a result, we revealed that whereas the development of the insrr MO1-injected embryos was significantly delayed compared to the control siblings at pH 7.2, no such difference was detected for insrr MO1-injected embryos incubated at pH 8.5 ( Figure 1C,D). In other words, alkaline pH medium "rescued" embryos with downregulated insrr from the retardation in development. It is noteworthy that in wild-type embryos grown in an alkaline medium (pH 8.5), no delay or acceleration of development was observed compared to control (pH 7.2).
we confirmed specificity of MO1 by demonstrating that the retardation effect caused by this MO could be rescued if insrr mRNA deprived of the MO1 target site was co-injected (Supplementary Figure S2). In addition, we proved efficiency of the MO suppressive effect upon translation of both insrr.L and insrr.S homeolog mRNAs, detecting N-fragments of proteins encoding by these mRNAs by Western blotting (see Material and Methods and Supplementary Figure S3).
. Importantly, the specificity of the development retardation effect elicited by the insrr downregulation was confirmed by the fact that insrr MO2, designed to another target site on insrr mRNA, elicited a similar effect to insrr MO1 (Supplementary Figure S1). Besides, we confirmed specificity of MO1 by demonstrating that the retardation effect caused by this MO could be rescued if insrr mRNA deprived of the MO1 target site was co-injected (Supplementary Figure S2). In addition, we proved efficiency of the MO suppressive effect upon translation of both insrr.L and insrr.S homeolog mRNAs, detecting N-fragments of proteins encoding by these mRNAs by Western blotting (see Material and Methods and Supplementary Figure S3).

Insrr Knockdown Causes a Shift of Gene Expression towards Genes Expressed in Earlier Stages of Development
As shown by PCA analysis, bulk RNA-seq data obtained for transcriptomes of all four groups of embryos clearly differed from each other (Figure 2A). In comparison with control group at pH 7.2, 1650 differentially expressed genes (DEG) were detected in transcriptome after insrr MO1 (designated as MO below) injection at pH 7.2 (padj < 0.05), including 807 upregulated and 843 downregulated genes. In case of alkali treatment of embryos injected with insrr MO, we found 1235 DEGs (padj < 0.05), including 635 upregulated and 600 downregulated genes ( Figure 2B). Also, comparison of transcriptomes of the control group at pH 8.5 and embryos injected with insrr MO1 at pH 8.5 revealed 1519 DEGs (padj < 0.05), including 770 upregulated and 749 downregulated genes ( Figure 2B, Supplement File S1). Interestingly, injection of this morpholino led to some increase of the insrr mRNA level, with log2(FoldChange) 0.647 at pH 7.2 and with log2(FoldChange) 0.513 at pH 8.5 (Supplement File S1). This effect is not surprising and was observed earlier after injections of morpholinos to mRNA of other genes. In particular, we detected a similar effect in the case of blocking by morpholino the translation of mRNA of the homeobox gene Xanf [22]. This effect likely indicates a negative feedback regulatory loop between the protein concentration and the level of transcription of its own gene.
To visualize the action of alkali we made graphic analysis of transcriptomes of the control and insrr MO-injected embryos at pH 7.2 depending on the degree of change in the expression of the genes. The density plot on a field of log2(FoldChange) vs. log10(baseMean) was plotted and genes that pass a threshold of padj < 0.05 in differential expression analysis between wild-type and insrr knockdown are colored red ( Figure 2C). Other genes are colored grey. Then, using multicolor indication, we labeled DEGs in the same plot, which were downregulated ( Figure 2D) and upregulated ( Figure 2E) in insrr MO-injected embryos under alkali treatment. We found that under alkaline treatment of insrr knockdown embryos, most downregulated genes had increase in their expression, and vice versa, upregulated genes had decreased expression.
Then, we annotated a list of DEGs in the insrr MO-injected embryos and control MO siblings at pH 7.2 (insrr MO 7.2/control MO 7.2 DEGs) (Supplementary Figure S4A). Upregulated DEGs were mostly enriched in microtubules and mitotic-cycle-regulation genes groups. Downregulated DEGs were found in RNA/DNA binding, ribosome, and metabolic regulation genes groups. Similar analysis was made for DEGs in insrr MOinjected embryos grown at pH 7.2 and pH 8.5 (insrr MO 8.5/insrr MO 7.2) (Supplementary Figure S4B). We found annotations only for upregulated DEGs in RNA binding, ribosome and gene expression groups (Supplementary Figure S4B). Among top genes with changed expression after insrr MO injection are actin and myosin coding genes, transcription factor pax6 and cycline ccna2 (Supplement File S1). levels much later: only after the beginning of neurulation (see a representative example of the temporal expression profiles of two DEGs from these groups in Figure 3A and a full table of profiles for all 30 selected DEGs in Supplementary Table S1). At the same time, when we had checked these 30 genes among insrr MO 8.5/insrr MO 7.2 DEGs, we found that under the influence of an alkaline medium, all of them had had a change in their expression levels to the opposite (Table 1). Interestingly, by analyzing temporal expression patterns of the15 most upregulated and most downregulated insrr MO 7.2/control MO 7.2 DEGs (Table 1) using the Xenbase online database, we revealed the following regularity consistent with the morphological retardation of development. Namely, in the wild-type embryos, these upregulated genes had maximum levels of their expression very early, either before or at the very beginning of gastrulation, and the expression of the downregulated genes reached its maximum levels much later: only after the beginning of neurulation (see a representative example of the temporal expression profiles of two DEGs from these groups in Figure 3A and a full table of profiles for all 30 selected DEGs in Supplementary Table S1). At the same time, when we had checked these 30 genes among insrr MO 8.5/insrr MO 7.2 DEGs, we found that under the influence of an alkaline medium, all of them had had a change in their expression levels to the opposite (Table 1).     For a more comprehensive analysis of the effect elicited by insrr knockdown and its alkali rescuing, we compared transcriptomes of Xenopus laevis embryos at stage 15 with downregulated insrr and their wild-type siblings grown under neutral (pH 7.2) and alkaline (pH 8.5) conditions as described in Materials and Methods.
To test if the revealed regularity could be also true in relation to other insrr MO 7.2/control MO 7.2 DEGs, we built the point diagram, plotted log2(Fold change insrr MO 7.2/control MO 7.2) versus log2(FoldChange stage10/stage15) (Supplement File S2) for each of 1461 DEGs ( Figure 3A), which were found in intersection of 1650 DEGs after insrr MO injection at pH 7.2 and 42,638 genes expressed during gastrulation [19] (Figure 3B,C). This diagram shows that most DEGs have log2(Fold change MO7.2/WT7.2) values comparable and similar as have log2(FoldChange stage10/stage15) values for the same genes. In other words, for most DEGs, the rule is true that if DEG has a higher expression level in embryos with downregulated insrr compared to the wild-type siblings, then one may expect that this gene should demonstrate a higher expression level at stage 10 compared to stage 15 during normal development. And vice versa, if DEG has a lower expression level in embryos with downregulated insrr, this indicates that this gene should demonstrate a lower expression level at stage 10 compared to stage 15. In sum, this confirms the overall validity of the regularity, suggesting a shift in gene expression towards earlier expressed genes in embryos with downregulated insrr.  Figure S5A and File S3) and compared their expression levels in conditions of insrr MO 7.2/control MO 7.2 and insrr MO 8.5/insrr MO 7.2 experiments. As one may see in the point diagram in Supplementary Figure S5B, direction of changes of expression levels of all these 331 DEGs after incubation of embryos in alkaline medium (pH 8.5) was truly the opposite to that observed in the neutral medium (pH 7.2). Importantly, the majority of these 331 DEGs sensitive to alkaline pH in insrr knockdown embryos are the genes that have the greatest change in expression in insrr knockdown embryos grown in the neutral medium (Table 1 and Supplementary Figure S6). Taking into account the known modulating effect of alkaline pH medium on IRR activity, one may conclude that this correlation once again confirms the specificity of changes in gene expression observed in embryos with downregulated insrr. Also, we annotated the list of the selected 331 DEGs and found them in ribosome, translation, and metabolic regulation genes groups (Supplementary Figure S5C).
In addition, we investigated differentially expressed genes in the wild-type Xenopus laevis embryos grown at pH 8.5 and pH 7.2 (control MO 8.5/control MO 7.2 DEGs). Alkali treatment of embryos resulted in 849 DEGs (padj < 0,05). Of those, 499 genes were upregulated, and 350 genes downregulated ( Figure 2B). Upregulated genes include factors involved in adherent junction, WNT signaling pathway (lef1, smad4), insulin signaling (insr, irs1), cell cycle (cdc27, cdk4, bub1), RNA metabolic process, regulation of gene expression and other processes (Supplement File S4). Genes with a reduced level of expression after alkali treatment were associated with spliceosome, embryo development, BMP signaling pathway and other processes (Supplement File S5).

Discussion
The physiological role of IRR in pH maintenance in the renal system and in regulation of some behavior traits was previously demonstrated in adult mice [2,23]. Although insrr knockout mice did not show defects in early embryonic development in vivo, blastocyst yield in knockout animals is significantly lower than in wild-type animals according to the results of in vitro analysis, suggesting a possible role for the receptor tyrosine kinase IRR in the development of mouse preimplantation embryos [17]. Acid-base balance regulation in frogs is different from mice or humans. For example, blood pH above 8.0 is usually observed depending on habitat and temperature [24,25], and due to this the role of IRR as an alkali sensor can be more important in frogs than in mice or humans. Also, frogs and toads can live and breed in a wide range of water pH, from acidic (pH about 4.0) to alkaline (pH about 10.0) [26], which indicates the presence of a mechanism for regulating the expression of various genes depending on the external pH.
We now studied the effects of temporary downregulation of insrr functioning in Xenopus laevis embryos. As a result, the following set of data, schematically shown in Figure 4, was obtained, indicating that IRR may regulate the rate of early development in frogs. Downregulation of insrr by anti-sense morpholinos resulted in a significant morphological retardation of embryo development. The specificity of this effect is confirmed by its reversibility after alkaline treatment of the insrr knockdown embryos, as well as by the rescue experiments, in which insrr MO was co-injected with insrr synthetic mRNA which lacked the MO target site.
ulation of IRR signaling. Furthermore, it can be assumed that although the protein ligand of IRR has not been found, this ligand may exist, and it regulates IRR signaling in Xenopus embryos. In this regard, early Xenopus embryos could be considered as a promising model for searching for IRR interacting proteins. Also, it should be noted that other membrane proteins such as alkaline pH-sensing channels (for example, TASK-2) [40] or alkaline pH activated receptors (Erbb2, c-Met) [41,42] may be involved in the action of extracellular alkaline pH medium on insrr knockdown in Xenopus laevis embryos.  In addition to the morphological retardation of development observed at the tadpole stages, we revealed in early neurula embryos with downregulated insrr a massive expression shift from later expressed genes towards earlier ones. Thus, in the first group were genes whose expression is increased during neurulation and which are responsible mainly for skeletal muscle and neural tube differentiation, including desmin, several genes involved in adrenergic signaling, foxg1, several genes of hox compex, mylfp, myl1, pax6, pitx2 and others. Among the second group were genes regulating overall patterning of the embryo before and at the very beginning of gastrulation: cer1, eomes, frzb, hhex, mixer, pou5f3, and many others. Importantly, as in the case of morphological retardation of development, expression of these genes could be temporarily rescued by incubation in alkaline pH medium.
Previous results revealed no developmental abnormalities of insrr genetic knockout mice. However, the developmental retardation observed in Xenopus embryos with downregulated insrr simply could be overlooked in mice due to the difficulty of observing intrauterine development, especially at early stages. In addition, there is growing evidence that in many cases genetic knockout in mice may not be phenotypic, due to compensatory mechanisms that bypass the knockout effects by activating homologous genes that can substitute the function of the knockout gene [27,28]. In particular, there is evidence that the overall rate of development also can be regulated by signaling through an insulin receptor [29,30]. Finally, the mechanisms of early development in mice and frogs could be different in aspects concerning IRR signaling.
The developmental retardation revealed in embryos with downregulated insrr was accompanied by a synchronous temporal shift in the expression of many genes that coordinately expressed in normal embryogenesis but are generally not mechanistically connected with each other. This indicates that signaling through IRR may regulate some chromatin remodeling, governing simultaneous switching between large gene ensembles. It can be assumed that such remodeling could most likely be associated with chromatin acetylation, which is known as a critical epigenetic factor responsible for altering the accessibility of promoters of genes at successive developmental stages [31][32][33]. Thus, histone hyperacetylation is essential to enhance accessibility for the transcription machinery of the promoters of genes normally expressed before the early gastrula stage, while promoters of genes expressed during neurulation and responsible for the muscle and neural crest differentiation do not need hyperacetylation [34].
The level of chromatin acetylation is determined by the balance between histone acetylases (HATs) and deacetylases (HDACs) activities [35]. As downregulation of insrr shifts the embryonic expression towards genes with maximum expression in the early stages, which are characterized by chromatin hyperacetylation, one may suppose that IRR signaling either inhibits HATs or stimulates HDACs activities. Indeed, in both of these cases, suppression of IRR signaling would lead to chromatin hyperacetylation and a shift in expression towards earlier expressed genes. However, the concentration of insrr transcripts is maximal at the beginning of embryogenesis and then decreases to a very low level by the onset of neurulation [19]. Such a temporal profile of insrr expression contradicts the assumption of its direct suppressive effect on HAT activity at developmental stages before gastrulation since these stages are characterized precisely by hyperacetylation of chromatin. Thus, it seems more likely that IRR signaling may not inhibit HATs but stimulate HDACs activity. In agreement with this, it was shown that treatment of embryos by butyrate, a noncompetitive inhibitor of HDACs [36], elicits retardation of development phenotypically similar to that we observed in the case of insrr downregulation by MO [32].
In a continuation of this, one may suppose that since IRR is the receptor tyrosine kinase, it could stimulate the activity of HDACs by regulating their phosphorylation. Indeed, it was shown, for example, that phosphorylation of HDAC1 and HDAC2 by protein kinase CK2 increases their activity by several times [37,38]. Furthermore, the activity of protein kinase CK2 itself is stimulated by phosphorylation of its two tyrosine residues [39]. In addition, other protein kinases capable of HDACs phosphorylation, for example, cAMPdependent protein kinase and protein kinase G in the case of HDAC1 [37], can be also considered as potential targets for IRR signaling in early embryonic cells. Obviously, all these speculations could be tempting topics for further experimental testing.
Interestingly, whereas we revealed clear retardation of development and rescuing effect of alkaline pH or synthetic insrr mRNA injected in embryos with downregulated insrr, we did not see visible changes in the developmental rate of the wild-type embryos treated by alkaline or acidic pH. Also, no acceleration of development compared to the wild-type embryos was detected in their siblings injected with synthetic insrr mRNA (data not shown). These results indicate that in the wild-type embryos neither pH nor the concentration of IRR is the limiting factor that determines overall signaling intensity through this receptor. Otherwise, it would be difficult to explain why only a strong decrease in the concentration of IRR in embryos injected with MO, but not pH or overexpression of insrr, allowed us to reveal the role of this receptor in regulating the rate of embryonic development (Figure 4). In turn, this may indicate that there are some other proteins in embryo, which interact with IRR and play a role of limiting factor in the regulation of IRR signaling. Furthermore, it can be assumed that although the protein ligand of IRR has not been found, this ligand may exist, and it regulates IRR signaling in Xenopus embryos. In this regard, early Xenopus embryos could be considered as a promising model for searching for IRR interacting proteins. Also, it should be noted that other membrane proteins such as alkaline pH-sensing channels (for example, TASK-2) [40] or alkaline pH activated receptors (Erbb2, c-Met) [41,42] may be involved in the action of extracellular alkaline pH medium on insrr knockdown in Xenopus laevis embryos. To test MO specificity, we performed the phenotype rescue experiments by co-injection of MO with insrr mRNA [3]. For mRNA synthesis, pCS2-insrr plasmid encoding the fulllength xenopus IRR was linearized with NotI restriction endonuclease, and the capped mRNA was synthesized using the SP6 mMESSAGE mMACHINE Transcription kit (Ambion). Then, we added synthetic insrr mRNA at final concentration 5 ng/µL into MO1 mixture for injection. Embryos injected with this mixture were compared with embryos injected with MO1 and control MO.

Embryo Manipulations
For RNA-seq, embryos injected with MO1 or with the control were incubated at pH 7.2 and 8.5 till embryonic stage 13 (early neurula) and five embryos of each type were collected in three replicates in RNAlater solution.
For injection experiments, MOs were diluted to the final concentration of 0.3 mM; mRNAs were diluted to the final concentration of 25 ng/µL. The mRNA or MO solutions were mixed with FLD (Fluorescein Lysinated Dextran, 40 kDa, 5 mg/mL, Invitrogen) and 4-5 nl of the mixture were injected into single blastomeres at two-cell stage. Injected embryos were grown until stage 13, at which myc-tagged IRR fragments in crude lysates of these embryos were analyzed by Western blotting as described in [43]. Coomassie stained gels were used as loading controls.

RNA Purification and RNA-seq
All RNA purifications and RNA-seq analysis were made by Genoanalytica (Moscow, Russia). Briefly, total RNA was extracted from the samples with Trizol reagent according to the manufacturer's protocol. Quality of RNA was checked with BioAnalyser and RNA 6000 Nano Kit (Agilent). Poly(A) RNA was purified using Dynabeads ® mRNA Purification Kit (Ambion). Illumina library was made from poly(A) NEBNext ® Ultra™ II RNA Library Prep Kit for Illumina (NEB) according to the manufacturer's protocol. Sequencing was performed on Illumina HiSeq1500 with 50 bp read length. At least 30 million reads were generated for each sample. Reads were mapped to the genome using Star Aligner and the fold changes of gene expression were calculated using DEseq2.0 software (Supplement File S1). Raw RNA-seq data were deposited to European Nucleotide Archive under the accession number PRJEB45623.

Gene Expression Analysis
For analysis of our RNA-seq data and previously published RNA-seq data [9], we used identical software. For reads mapping, STAR v2.7 software was used [44]. For reads counting we used HTSeq v0.11.2 software [45], and finally for differential transcriptional and principal component analysis (PCA), DESeq2 v3.9 software was used [46]. Enriched Gene Ontology terms were identified using DAVID bioinformatics resources with genes pre-ranked according to their fold change induced by insrr morpholino, pH treatment or both [47,48]. The temporal expression patterns for the selected genes were retrieved from the online resource xenbase.org.