(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Foxi2 and Sox3 are master transcription regulators that control ectoderm germ layer specification in Xenopus [1]

['Clark L. Hendrickson', 'Developmental', 'Cell Biology', 'University Of California', 'Irvine', 'California', 'United States Of America', 'Ira L. Blitz', 'Amina Hussein', 'Center For Complex Biological Systems']

Date: 2025-11 

Germ layer specification represents a critical transition where pluripotent cells acquire lineage-specific identities. We identify the maternal transcription factors Foxi2 and Sox3 to be pivotal master regulators of ectodermal germ layer specification in Xenopus. Ectopic co-expression of Foxi2 and Sox3 in prospective endodermal tissue induces the expression of ectodermal markers while suppressing mesendodermal markers. Transcriptomic analyses reveal that Foxi2 and Sox3 jointly and independently regulate hundreds of ectodermal target genes. During early cleavage stages, Foxi2 and Sox3 pre-bind to key cis-regulatory modules (CRMs), marking sites that later recruit Ep300 and facilitate H3K27ac deposition, thereby shaping the epigenetic landscape of the ectodermal genome. These CRMs are highly enriched within ectoderm-specific super-enhancers (SEs). Our findings highlight the pivotal role of ectodermal SE-associated CRMs in precise and robust ectodermal gene activation, establishing Foxi2 and Sox3 as central architects of ectodermal lineage specification.

Funding: This research was funded by National Institutes of Health (GM139617 and HD109696 to K.W.Y.C.) and National Science Foundation (1755214 to K.W.Y.C.). Additional support was provided by a GAANN Fellowship from the U.S. Department of Education (P200A220015 to C.L.H.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2025 Hendrickson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Here, we reveal that the ectodermal gene regulatory program is orchestrated by maternally expressed, and ectodermally enriched master regulatory TFs, Foxi2 and Sox3. These TFs prebind to the CRMs of the naïve embryonic genome prior to ZGA, preceding the appearance of major histone modifications and transcription. This prebinding event coordinates the subsequent ectodermal cell differentiation program. The CRMs bound by Foxi2 and Sox3 are associated with ectodermal super-enhancers (SEs), which ensure robust target gene expression. Our findings are consistent with the view that Foxi2 and Sox3 function as master regulators of the ectodermal germ layer.

Previous epigenetic studies have shown that the early embryonic genome lacks significant histone modifications before the blastula stage and therefore remains epigenetically naïve (reviewed in [ 15 ]). Histone modifications associated with active enhancers first appear during blastula stage, coinciding with major zygotic genome activation (ZGA) and the segregation of the pluripotent embryonic cells into specialized ectoderm, mesoderm, and endodermal lineages. We hypothesize that maternal, locally expressed master TFs bind to specific cis-regulatory modules (CRMs) to activate or repress the transcription of the first wave of zygotic target genes, thus orchestrating the gene regulatory programs required for cell lineage specification.

Our goal in this study is to determine the primary master regulators initiating ectodermal cell lineage differentiation at the highest positions in the programming of cell lineages. In vertebrates, germ layer specification is one of the earliest developmental decisions that pluripotent cells make. In amphibians, the three germ layers are organized along the animal–vegetal axis, where the ectoderm forms in the animal pole, the endoderm vegetally, and the mesoderm in the equator. The ectoderm comprises progenitor cell populations giving rise to the skin, nervous system, sensory organs, and numerous neural crest derivatives. Several experiments have demonstrated that Xenopus ectodermal cells of the blastula stage are pluripotent. Transplantation of blastula prospective ectodermal cells into the vegetal endoderm converts the ectodermal cells to endodermal fates [ 7 ]. In addition, explanted blastula ectodermal tissue (“animal caps”) can be converted to mesendodermal cells ex vivo (organoids), by soaking them in activin, a TGFβ-superfamily ligand [ 8 – 10 ]. With specific combinations of growth factors and/or small molecules, blastula ectodermal explants can be programmed to differentiate into a wide range of cell types, including neural tissue, muscle cells, blood, heart, and endoderm primordia [ 11 , 12 ]. Although this ectodermal pluripotency persists through blastula stages, it is autonomously lost by early gastrula [ 13 , 14 ]. Therefore, ectodermal germ layer and sub-lineage specification is an ideal system to study the dynamic processes controlling cellular potency and differentiation.

Additionally, discovery of these pluripotency TFs among many permutations highlights that not all TFs hold equal importance in development. TFs are considered master regulators when they play a more critical role in regulating specific differentiation pathways because they occupy a higher position within a gene regulatory program controlling downstream events in differentiation and maintaining pluripotency. One of the earliest TFs to support the concept of master regulators was Myod1, which was shown to drive fibroblasts into a muscle cell differentiation path [ 3 ]. Other examples include Spi1 (Pu.1), which specifies myeloid and B cell lineages in hematopoiesis [ 4 ], Gata1 in erythroid lineage specification [ 5 ], and Pdx1 for pancreas organogenesis [ 6 ]. These master regulatory TFs play a pivotal role in determining the identity and function of specific cell types. Identifying master regulatory TFs in developmental biology is fundamentally important for comprehending the mechanisms of cell lineage segregation and advancing regenerative medicine to facilitate direct cell lineage conversion.

A major breakthrough in biology was the first animal (Xenopus laevis) cloning experiment, which proved that differentiated cells retain the full genetic potential to develop into an entire organism [ 1 ]. This result demonstrated the concept of genomic equivalence, where cellular differentiation is governed by gene regulation rather than by irreversible modifications to the genome, such as differential loss of subsets of genes in cell lineages. These experiments further revealed that the egg cytoplasm contains necessary maternal regulators to reprogram the epigenetic memory of the differentiated cell nuclei to a totipotent state. Subsequently, it was shown in mammalian cells that networks of master regulator transcription factors (TFs), including most notably Klf4, Myc, Pou5f1, and Sox2, are critical for embryonic pluripotency [ 2 ]. This highlights the essential role of multiple TFs working in concert to orchestrate cellular identity, lineage specification, and reprogramming.

Results

Foxi2 and Sox3 prebind future ectodermal CRMs before Ep300 recruitment Previous work showed that maternal Foxi2 can bind to an upstream CRM of the foxi1 gene to activate its expression [25]. We wished to further examine the genome-wide regulatory role of Foxi2 during early development. Chromatin immunoprecipitation (ChIP)-seq analysis of Foxi2 was performed in mid- to late blastula (st. 8, 9) and early gastrula (st. 10.5) embryos and identified a total of 16,427 bound regions over the entire time course (Fig 2A). There are 13,158 high-confidence Foxi2 bound regions in the mid blastula, 6,668 in the late blastula, and 6,039 in the early gastrula (Fig 2A). Overall, Foxi2 binding is dynamic, displaying unique (class I, III, VI), shared (class II, IV, V), as well as persistently occupied (class VII) binding patterns across stages (Fig 2A and 2B). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Temporal dynamics and motif analysis of Foxi2 DNA binding and co-occupancy with Sox3 and Ep300 from blastula to early gastrula. (A) Venn diagram showing temporal dynamics of Foxi2 DNA binding, with ChIP-seq peaks categorized into Classes I–VII. (B) Analysis comparing Ep300 and H3K27ac signal from early gastrula ectoderm and endoderm dissections, within Foxi2 peak classification. (C) Proportion of Foxi2 class I–VII peaks overlapping Ep300 peaks from early gastrula ectoderm. (D) Temporal RNA expression profiles of genes within 20 kb of Foxi2/Ep300 co-bound peaks in classes I–VII. (E) Top predicted Fox motif and best-matched Sox and Pou motifs identified within Foxi2/Ep300 peaks. (F) Proportion of Fox, Sox, and Pou motifs present in Foxi2/Ep300 peaks. (G) 64-cell embryo ChIP-qPCR analysis of Foxi2, Sox3, and Ep300 binding at early gastrula Foxi2/Ep300 co-bound regions. Error bars represent standard deviation from biological duplicates, ChIP samples were normalized by their percent recovery compared to input (non-pulldown) samples. (H) Total Foxi2 (top) and Sox3 (bottom) peaks from early blastula to early gastrula stages, with corresponding Foxi2 and Sox3 ChIP-seq signal. (I) Early gastrula Sox3 peak overlap with Foxi2 class IV–VII “active” peaks. The data underlying this figure can be found in S2 Data and GSE288637. Xenopus illustrations © Natalya Zahn [27]. https://doi.org/10.1371/journal.pbio.3003476.g002 To characterize functionally relevant Foxi2 peaks, we examined the correlation of classified Foxi2 peaks with Ep300 and H3K27ac peaks in ectoderm and endoderm explants. Class IV, V, VI, and VII regions contain more H3K27ac deposition and Ep300 binding compared to classes I, II, and III (Figs 2B and S2B). In total, 3,050 Foxi2 peaks are Ep300 co-occupied, where class IV, V, VI, and VII Foxi2 peaks show the highest degree of overlap with both Ep300 and H3K27ac (Figs 2B, 2C and S2B). We bioinformatically assigned nearest zygotic genes to the Foxi2 peaks marked by Ep300, and examined whether these CRMs are linked to transcriptional activity of the genes. We find that class IV, V, VI, and VII genes become transcriptionally active during gastrulation, whereas class I, II, and III genes remain inactive (Fig 2D). These results suggest that Foxi2 peaks (1) are present at stage 10.5, either uniquely or overlapping with blastula stage binding, (2) are co-bound with ectodermal Ep300, (3) are decorated with ectodermal H3K27ac, and (4) are associated with genes expressed post ZGA, indicating that they represent active CRMs. TFs form complexes on CRMs and typically act in combination. To predict potential co-factors that may work together with Foxi2, we characterized TF motifs enriched within Foxi2-bound regions. As expected, de novo motif analysis reveals Fox motifs are the most centrally enriched (Fig 2E). Besides the Fox motif, the next most abundantly present and centrally enriched motifs are Sox and Pou. These Fox/Sox/Pou motifs co-occur in 33% of Foxi2 peaks bound by Ep300, potentially suggesting complex formation (Fig 2F). Interestingly, while 48% of Foxi2 peaks contain only Fox/Sox motifs, just 7% contain only Fox/Pou motifs (Fig 2E and 2F). Given Foxi2 and Sox3 high rate of motif co-occurrence, and their maternal expression, we assessed whether Foxi2 and Sox3 prebind functional ectoderm CRMs, before Ep300 recruitment and the onset of ZGA. Foxi2, Sox3, and Ep300 ChIP-qPCR analysis of 64-cell stage embryos (st. 6.5, ~3hpf) reveals that Foxi2 and Sox3, but not Ep300, occupy CRMs for dlx5, foxi1, hgf, krt7, krt70, tfap2a, and tp63 (Fig 2G), which are expressed in the ectoderm of the early gastrula. This demonstrates that Foxi2 and Sox3 prebind to these ectodermal CRMs during cleavage stages, thus prior to ZGA (st. 8.5, ~ 4–4.5hpf), and before the recruitment of Ep300. This finding demonstrates that these maternal TFs initially recognize the CRMs of future ectodermally expressed genes during the cleavage stages.

Independent and cooperative roles of Foxi2 and Sox3 in regulating ectodermal gene expression To investigate the molecular function of Foxi2 and Sox3, we generated Foxi2 and Sox3 knockdown phenotypes using translation-blocking morpholino antisense oligonucleotides (MOs) (S4A and S4B Fig). Embryos injected with Foxi2 or Sox3 MOs showed substantial depletion of Foxi2 and Sox3 proteins (Fig 4A). While only a minority of Sox3 morphants exhibited decreases in body length and inhibited blastopore closure, Foxi2 knockdown embryos showed robust defects in both dorsoventral patterning and body length, as well as reduced head size (Fig 4B). RNA-seq analysis of both Foxi2 and Sox3 morphants compared to wild-type embryos identified genes activated and repressed by Foxi2 and Sox3 (Fig 4C and 4D). We intersected these data with Foxi2 and Sox3 ChIP-seq datasets to identify direct target genes, defined as those showing at least a 2-fold change in expression in the morphants and having Foxi2 and/or Sox3 binding sites within 20 kb of the gene (Fig 4C and 4D). At the late blastula stage, Foxi2 and Sox3 directly activate 78 and 114 genes, and repress 68 and 18 genes, respectively (Fig 4C). A similar analysis in gastrula-stage embryos revealed that Foxi2 and Sox3 directly activate 72 and 164 genes and repress 54 and 136 genes, respectively (Fig 4D). Intersecting the lists of genes activated by Foxi2 and Sox3 identifies 34 genes co-bound and co-regulated by both factors. These genes include well-studied ectodermal TFs such as dlx3, dlx5, foxi1, hes5.1, tfap2a, and tp63 [30,31]. Reverse transcription (RT)-qPCR analysis of foxi2 and sox3 morphants confirms that these ectodermal TFs are jointly regulated by both factors (Fig 4E). Separately, 115 genes are regulated solely by Foxi2, while 243 genes are regulated solely by Sox3. Specific targets of Foxi2 include, amotl1, dlc, gata3, kit, klf5, and krt70, while direct targets of Sox3 include cdx4, foxb1, foxd4l1.2, hes5.10, neurog3, olig4, sox21, zic3, and zic4. These data suggest that Sox3 and Foxi2 regulate their specific target genes both jointly and independently, serving as master TFs that govern the early ectodermal program. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Independent and cooperative roles of Foxi2 and Sox3 in regulating ectodermal gene expression. (A) Western blot of Foxi2 and Sox3 morphant embryos at late blastula and early gastrula stages. (B) Phenotypic analysis of Foxi2, Sox3, and Foxi2/Sox3 morphants at the early tailbud stage. (C, D) RNA-seq analysis of late blastula embryos (top) and early gastrula embryos (bottom) showing activated and repressed genes within 20 kb of Foxi2-/Sox3-bound regions. Histograms show the total number of direct target genes affected in morphants. (E) Schematic diagram of loss-of-function experiment (left). RT-qPCR analysis of key ectodermally expressed genes (right) in early gastrula ectoderm of Foxi2, Sox3, and Foxi2/Sox3 morphants. eef1a1 expression was used for normalization. Error bars represent standard deviation from 3 biological replicates, asterisks indicate statistically significant differences (p < 0.05) determined by Student t test. The data underlying this figure can be found in S4 Data and GSE288636. Xenopus illustrations © Natalya Zahn [27]. https://doi.org/10.1371/journal.pbio.3003476.g004

Foxi2 and Sox3 directly co-regulate spatially distinct inner and outer layer cell states To gain a deeper understanding of gastrula ectoderm differentiation than can be provided by bulk RNA-seq analyses, we assessed nascent transcripts at the single-cell level by performing single-nucleus (sn) RNA-seq on stage 10.5 wild-type embryos. We then characterized the cell-type-specific expression of Foxi2 and Sox3 direct target genes within this dataset. Pre-processing using a 1,500 genes/cell cut-off yielded 13,711 high-quality early gastrula nuclei (Fig 5A). De novo clustering and marker gene expression analysis identified a total of 13 defined cell states, consisting of 5 ectodermal, 4 mesodermal, and 4 endodermal cell types (Figs 5A, S5A and S5B). The five ectodermal cell clusters map at the bottom of two major lobes in the Uniform Manifold Approximation Projection (UMAP) (Fig 5A) and express markers that show these represent inner neural plate, inner neural plate border, inner non-neural ectoderm, outer neural ectoderm, and outer non-neural ectoderm. Notably, inner ectoderm cell clusters are positioned in the left lobe, while outer ectoderm cell clusters segregate in the right lobe (Figs 5A, S5C, and S5E). Two small lobes in the upper right (gray) contain unannotated cells (UC) that may represent a novel cell state characterized by dscaml1 and tp73 expression (S5D Fig) among others, but the identity of these cells is currently unclear. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Foxi2 and Sox3 directly co-regulate genes in all ectodermal cell states. (A) Single nucleus RNA-seq UMAP (uniform manifold approximation projection) identifying 14 distinct cell types in the early gastrula. (B) Z-score expression analysis of Foxi2/Sox3 co-bound target genes shows distinct gene expression patterns marking outer and inner ectodermal cell types. The boxes represent genes expressed in both the inner and outer ectodermal layers, as well as genes specific to either the outer or inner ectodermal layer. (C) Schematic diagram depicting the regulatory roles of Foxi2 and Sox3, either independently or jointly, in activating the expression of target genes across various ectodermal cell types. The data underlying this figure can be found in S5 Data and GSE288638. https://doi.org/10.1371/journal.pbio.3003476.g005 After identifying 5 ectodermal cell clusters in stage 10.5 embryos, we mapped the expression patterns of Foxi2 and Sox3 direct target genes in the gastrula ectoderm to determine how Foxi2 and Sox3 contribute to the regional identity of this germ layer. There are 18 inner-layer and 14 outer-layer bound ectoderm target genes that are exclusively regulated by Foxi2, which include amotl1, klf5, krt70, lhx3, nkx3-1, and olfm4. Additionally, 53 inner and 49 outer layer direct ectoderm target genes are exclusively regulated by Sox3, including foxd4l1.2, neurog3, olig3, olig4, sal3, sox21, szl, zic3, and zic4, many of which are involved in neurogenesis. Pan-ectodermal targets (e.g., dlx3, hes5.1, klf5, lhx5, and tfap2a) are expressed across all five cell states. Some of these genes are specifically regulated by Foxi2 or Sox3 alone, while others require the combinatorial input of both Foxi2 and Sox3 (Fig 5C). The observations suggest that Foxi2 and Sox3 regulate genes in all five distinct ectoderm cell states. Interestingly, we identified target genes of Foxi2 and Sox3 that are specifically expressed in the inner (both non-neural and neural) ectodermal layer, such as foxi1, prdm14, szl, tp63, and zic4. Similarly, genes such as dlx5, hes4, and krt70 are expressed in the outer (both non-neural and neural) ectodermal cell states. Since Foxi2 and/or Sox3 regulate genes across both distinct ectodermal layers, we suggest that additional factors, like Prkci and Mark3, might play a role in differential segregation of these cell layers ([32,33], Tabler and colleagues [34]).

Foxi2 and Sox3 are master regulators shaping the ectodermal epigenetic landscape Given the potential synergy of Foxi2 and Sox3 within the ectoderm and their distinct roles in ectoderm target gene activation, could Foxi2 and Sox3 together serve as master regulators of ectodermal differentiation? To test this hypothesis, we injected Foxi2 and/or Sox3 mRNA into the vegetal pole and assessed the endoderm cell state in dissected vegetal tissue explants at gastrula stage 10.5 (Fig 6A). While overexpression of Foxi2 or Sox3 alone induced modest ectoderm gene expression, coinjection of both factors activated the ectodermal genes, dlx5, foxi1, krt70, lhx5, and tfap2a. Moreover, expression of key endodermal genes such as gata4, gsc, hhex, and otx2 was significantly reduced in endodermal explants co-expressing Foxi2 and Sox3 mRNA. These findings support Foxi2 and Sox3 working together as master regulators of ectodermal identity. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Foxi2 and Sox3 are master regulators shaping the ectodermal super enhancer landscape. (A) Schematic diagram of the experimental design (left). RT-qPCR analysis of key ectoderm and endoderm genes from early gastrula endoderm after ectopic expression of 1ng Foxi2, Sox3 or Foxi2/Sox3 mRNA. eef1a1 expression was used for normalization. Error bars represent standard deviation from 3 biological replicates, asterisks indicate statistically significant differences (p < 0.05) determined by Student t test. (B) Accumulation of H3K27ac, H3K4me1, and H3K27me3 histone modifications around Foxi2/Sox3/Ep300 (FSE) co-occupied, and other solely bound Ep300 peaks. (C) Deposition of Ep300, H3K27ac, H3K4me1 and H3K27me3 at FSE and other Ep300 peaks present in the early gastrula ectoderm and endoderm. (D) Ep300 signal intensity in dissected ectoderm at enhancers co-bound by Foxi2, Sox3, and Ep300 (FSEs). Box-and-whisker plots show the distribution of Ep300 signals, highlighting differences between WT and FS morphants. (E) Total ectoderm SEs associated with FSE co-occupied peaks and Ep300 solely occupied peaks. (F) Genome browser views of ectodermal SEs (shaded area), which are present (highly acetylated) in ectoderm but absent in endoderm. (G) Histogram comparing the total number of SEs associated with peaks containing Foxi2/Sox3/Ep300 (FSE), Ep300 with insignificant Foxi2/Sox3 binding (Ep300), and without Ep300 (Ep300(−)). FSE peaks are highly associated with ectodermal SEs. (H) Z-score expression analysis of genes located within 20 kb of ectoderm (top) and endoderm (bottom) SEs. snRNA-seq cell-type expression analysis reveals that genes associated with ectodermal SEs are preferentially expressed in ectodermal cells. (I) Average expression levels (top) of Foxi2- and/or Sox3-regulated ectodermal SE-associated genes compared to all ectodermal SE-associated genes. Coefficient of variation (COV) analysis (bottom) using snRNA-seq shows that Foxi2 and/or Sox3-regulated ectodermal SE-associated genes exhibit lower expression variability than the full set of ectodermal SE-associated genes. In contrast, ectodermal RE-associated-gene display lower average gene expression and higher COV. Asterisks indicate statistically significant differences (p < 0.05) determined by Wilcoxon Rank-Sum test. The data underlying this figure can be found in S6 Data, GSE288637, and GSE288638. Xenopus illustrations © Natalya Zahn [27]. https://doi.org/10.1371/journal.pbio.3003476.g006 In addition to the transcriptional role of Foxi2 and Sox3 in specifying the ectodermal cell state, we determined whether Foxi2 and Sox3 also shape the epigenetic landscape of ectodermal cells. Temporal analysis of H3K27ac and H3K4me1 signal deposition around 2,901 CRMs co-bound by Foxi2, Sox3, and Ep300 (FSE) (Fig 3E), reveals an increase from the early blastula to the early gastrula stage compared to independent Ep300 regions (Fig 6B). Interestingly, the repressive mark, H3K27me3 becomes gradually enriched around Ep300 peaks lacking Foxi2 and Sox3 binding (Fig 6B). These results suggest that Foxi2/Sox3 co-binding promotes H3K27ac and H3K4me1 enrichment at the expense of H3K27me3 deposition around ectodermal CRMs. To overcome the lack of germ layer resolution in whole-embryo histone marking studies, we also performed experiments using dissected ectodermal and endodermal tissues from the early gastrula stage. While no significant differences in Ep300 binding were observed between ectoderm and endoderm at both Foxi2-Sox3 co-bound (FSE) regions and Ep300 only peaks (lacking Foxi2/Sox3) (Fig 6C), we found a significant difference in histone modifications. In the ectoderm, FSE regions showed significant accumulation of H3K27ac and H3K4me1, while H3K27me3 deposition is notably absent (Fig 6C). In contrast, Ep300-only peaks showed modest H3K27ac and H3K4me1 accumulation but displayed strong H3K27me3 enrichment. These findings suggest that Foxi2 and Sox3 specifically regulate the epigenetic landscape of ectodermal genes within the ectodermal germ layer, but not in the endoderm. We sought to determine the role of Foxi2 and Sox3 in establishing the ectoderm-specific epigenetic landscape by measuring Ep300 genomic recruitment in the absence of both factors. To determine whether Foxi2 and Sox3 are required for tissue-specific Ep300 recruitment, we performed ChIP-seq on dissected ectoderm and endoderm from FS MO-injected and WT embryos (S6A Fig). Signal-based K-means clustering was performed to form FSE (I) and FSE (II) peak groups, which both display significant Ep300 deposition decreases in FS MO embryos compared to wild type (Fig 6D). In contrast, in FS MO embryos both non-FSE ectoderm Ep300, and vegetal Ep300-bound regions did not display significant decreases in Ep300 deposition (S6B Fig). Taken together, these data indicate that Foxi2 and Sox3 colocalization is required to recruit Ep300 to Foxi2 and Sox3-bound CRMs, comprising 49.7% of total ectodermal Ep300-bound regions in the early gastrula.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3003476

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/