Click here to close
Hello! We notice that you are using Internet Explorer, which is not supported by Xenbase and may cause the site to display incorrectly.
We suggest using a current version of Chrome,
FireFox, or Safari.
Hayashi S
,
Kawaguchi A
,
Uchiyama I
,
Kawasumi-Kita A
,
Kobayashi T
,
Nishide H
,
Tsutsumi R
,
Tsuru K
,
Inoue T
,
Ogino H
,
Agata K
,
Tamura K
,
Yokoyama H
.
???displayArticle.abstract???
Many amphibians can regenerate limbs, even in adulthood. If a limb is amputated, the stump generates a blastema that makes a complete, new limb in a process similar to developmental morphogenesis. The blastema is thought to inherit its limb-patterning properties from cells in the stump, and it retains the information despite changes in morphology, gene expression, and differentiation states required by limb regeneration. We hypothesized that these cellular properties are maintained as epigenetic memory through histone modifications. To test this hypothesis, we analyzed genome-wide histone modifications in Xenopus limb bud regeneration. The trimethylation of histone H3 at lysine 4 (H3K4me3) is closely related to an open chromatin structure that allows transcription factors access to genes, whereas the trimethylation of histone H3 at lysine 27 (H3K27me3) is related to a closed chromatin state that blocks the access of transcription factors. We compared these two modification profiles by high-throughput sequencing of samples prepared from the intact limb bud and the regenerative blastema by chromatin immunoprecipitation. For many developmental genes, histone modifications at the transcription start site were the same in the limb bud and the blastema, were stable during regeneration, and corresponded well to limb properties. These results support our hypothesis that histone modifications function as a heritable cellular memory to maintain limb cell properties, despite dynamic changes in gene expression during limb bud regeneration in Xenopus.
Fig. 1. X. laevis and X. tropicalis exhibit comparable regenerative abilities. (A) Regeneration experiments. The leftlimb bud was amputated at a level corresponding to the knee or ankle in tadpoles at stages 52 or 53, respectively, and the digits on regenerated (left) hindlimbs were counted to establish regenerative ratios. (B) and (C). Graphs show regenerative ratios for (B) the X. laevis hindlimb after amputation at stages 52 or 53 (n=26, 16) and (C) the X. tropicalis hindlimb after amputation at stages 52 or 53 (n=12, 17.) Colors indicate the number of digits on the regenerated limb. (D) Ventral view of a completely regenerated hindlimb in a stage 58 X. tropicalis tadpole after amputation of the lefthindlimb at the putative knee level at stage 52. The morphology of the regenerated hindlimb (left) and that of the intact hindlimb (right) were similar. Scale bar: 5 mm.
Fig. 2. Trimethylation of lysine 27 on histone H3 (H3K27me3) is required for proper regeneration and development of limb buds. (A) Experimental schedule. To inhibit the trimethylation of H3K27, X. laevis tadpoles were reared in 15 μM DZNep (experimental) or DMSO (control) for 9 h without food and were subsequently fed for 15 h in dechlorinated tap water each day for 5 days. After 5 dpa, tadpoles were reared and fed as usual. (B) and (C) Regenerating limb buds at 5 dpa treated with DZNep (B) and DMSO (C). The green line indicates the amputation plane. Scale bar=500 μm. (D) and (E). Measured daily growth in lengths of the blastema (D) and stump (E). The distance from the base of the limb bud to the amputation plane and the distance from the amputation plane to the distal tip of the blastema were measured as the “stump†and the “blastema†lengths, respectively. δlength represents the difference in length between “X†dpa and “X−1†dpa. An asterisk indicates a statistically significant difference at p<0.05 (Welch’s t-test). Error bar indicates s.e.m (n=10 and 8 for DMSO and DZNep, respectively).
Fig. 3. Genome-wide histone modifications in X. tropicalis limb bud regeneration. (A) Left (diagram): Samples were collected from stage 53 tadpoles by amputating the limb bud at the putative ankle level (0 dpa), and the blastema was collected at 7 dpa. Right (diagram): Histone modifications and read-count range. H3K4me3 and H3K27me3 are tightly linked to the activation and inactivation, respectively, of gene expression. Reads were counted in gene loci from −1000 bp to +1000 bp around the transcription start site (0 bp). (B)−(E) Scatter plots comparing (B) H3K4me3 and (C) H3K27me3 in the X. tropicalis limb bud and blastema; histone-modification profiles were similar in (D) the limb bud and (E) blastema. Black dots show arbitrarily selected genes. (F) and (G) Distributions of the mapped read counts cumulated over all genes show high H3K4me3 and low H3K27me3 around the gene model transcription start sites in both the (F) limb bud and (G) blastema; these profiles indicate epigenetic regulation to an active state.
Fig. 4. Histone modification of morphogenetic genes in X. tropicalis hindlimb bud regeneration. (A) Read mapping shows epigenetic regulation of the hoxa cluster. K4 trimethylation peaked from hoxa5 to hoxa13, while K27 trimethylation peaked at hoxa1, indicating strong expression of 5′ hox genes. Modification profiles were similar in the stage 53 limb bud (0 dpa) and blastema amputated at the ankle level (7 dpa). The genome sequence is indicated in 140 kb. Read mapping range: 0−60. (B) and (C) For tbx5, which is specifically expressed in the forelimb, a high peak of H3K27me3 but not H3K4me3 indicated epigenetic repression of tbx5 in the hindlimb bud. In (C), a high peak of H3K4me3 but not H3K27me3 at tbx4, which is specifically expressed in the hindlimb, indicates chromatin regulation to an active state. Thus, ChIP-seq analysis reflected intrinsic hindlimb features. Limb bud (0 dpa) and blastema (7 dpa) histone-modification profiles are similar in each panel. Transcription start sites are indicated by a vertical dotted line. The genome sequence is shown in 50 kb. Read mapping range: 0−60.
Fig. 5. Histone modification of a limb-specific/FGF-dependent gene and its potential expression. (A) ChIP-sequence analysis of limb-specific gene in the X. tropicalis stage 53 hindlimb bud or blastema amputated at the ankle level at stage 53. Read mapping of prrx1, a blastema marker, showed a high peak for H3K4me3 but not H4K27me3. (B)–(D) A prrx1-GFP reporter assay in X. laevis stage 53 tadpoles. The hindlimb bud was amputated at a point corresponding to the ankle. The GFP reporter was expressed throughout the limb bud and blastema at 3 and 7 dpa. (E) In situ hybridization of prrx1 mRNA in X. laevis stage 52 tadpoles. The hindlimb bud was amputated at a point corresponding to the knee; prrx1 expression was sustained in the stump region at 5 dpa, although it had disappeared in the early blastema. (F) Read mapping of Fgf signaling downstream; dusp6 was highly enriched in H3K4me3. (G)–(I) A dusp6-GFP reporter was detected throughout the limb bud and blastema. (J) In situ hybridization of dusp6 mRNA was done as (E) dusp6 expression was detected in the mesenchyme underlying the epidermis at 5 dpa. The vertical dotted line indicates the transcription start site. The genome sequence is indicated in 50 kb. Read-mapping range: 0−60. The red line indicates the amputation plane. Scale bar: 500 µm in (B)–(D) and (G)–(I); 100 µm in (E) and (J). Up, posterior; right, distal.
Fig. 6. Histone modification of shh in X. tropicalis limb bud regeneration. (A) A gene model of the limb-specific shh enhancer MFCS1. The horizontal bar indicates the qPCR amplification region. (B) and (C). ChIP-qPCR analysis of the shh gene locus. Limb buds (stage 53), tadpole blastemas at 7 dpa (amputated at stage 53, ankle level), froglet forelimbs and froglet forelimb blastemas at 10 dpa were collected and lysed, and limb properties were compared with the shhlimb enhancer (MFCS1). Quantities were normalized to IgG. (B) No histone modifications were detected in the shhlimb enhancer. (C) H3K4 and H3K27 trimethylation was detected at shh transition start sites (TSS), with higher levels of H3K27 trimethylation in all samples. (D) and (E). In the gastrula stage, in contrast, neither (D) the limb enhancer region nor the (E) transcription start site was enriched in histone modifications. ChIP-qPCR was repeated three times with similar results; representative data are shown in (B)–(E). Error bars indicate the SEM of triplicates. (F) and (G). ChIP-sequence analysis for the shhlimb enhancer region and gene locus. The limb bud and blastema were collected as for ChIP-qPCR. (F) Read mapping at the shhlimb enhancer showed low trimethylation at K4 and K27 of H3. (G) As with the qPCR results, ChIP-seq read mapping showed that trimethylation was slight at K4 but higher at K27 at the shh gene locus. No mapping region is shown due to lack of sequence information in the genome. The vertical dotted line indicates the transcription start site. The genome sequence is indicated as 30 kb. Read-mapping range: 0−60.
Fig. 7. Model showing maintenance of cellular limb-lineage features during regeneration. Progenitor cells commit to a limb lineage and cell fate (upper panel). The limb bud expands and differentiates to form the limb as a motile organ. After amputation, limb cells re-express morphogenic genes to recapitulate limb development. Cellular memory in limb cells allows morphogenesis to start again. Expression patterns of gene products (mRNAs and proteins) change dynamically over time and by region during regeneration. Therefore, gene products cannot carry the memory for intrinsic cell features, such as commitment to a limb-cell lineage. Histone-modification patterns (H3K4me3 and H3K27me3) in the limb bud are maintained in the blastema during limb reconstruction. Studies in cell biology have shown that daughter cells inherit parental histone modifications. These modifications may serve as the cellular memory of limb lineage, enabling surviving cells to reconstruct a limb just like the original after an injury (lower panel).
