Deficiency of Creb3l1 in Wnt1+ lineage led to shorter roots and thinner dentin

We firstly revisited the published scRNA-seq profile of the odontoblast lineage, and found regulon of CREB3L1 was specifically highly enriched at the terminal stage of odontoblast differentiation according to our previous single-cell RNA-seq results10 (Fig. 1a). RNAscope staining revealed that Creb3l1 mRNA was not expressed in preodontoblasts at embryonic day 15.5 (E15.5) (Fig. 1b), whilst its expression was evident in polarized odontoblasts at E18.5 (Fig. 1c), and it was continuously expressed in mature odontoblasts at postnatal day 2.5 (PN2.5) (Fig. 1d). To uncover its necessity in dentinogenesis, we generated Wnt1-Cre; Creb3l1f/f conditional knockout mice (cKO), which were used to specifically delete Creb3l1 in the neural crest lineage. All conditional knockout mice survived and displayed no statistically significant differences in either body size or weight compared to wildtype (WT) mice (Supplementary Fig. S1a, b). To assess the efficiency of knockout and examine tooth morphology in cKO mice, hematoxylin and eosin (HE) staining and immunohistochemistry (IHC) staining of CREB3L1 were conducted on molars at postnatal day 0.5 (PN 0.5) (Supplementary Fig. S2a, b). As expected, the dental papilla showed no expression of CREB3L1 (Supplementary Fig. S2a). However, there were no significant differences in tooth crown morphology between WT and cKO (Supplementary Fig. S2b).

Fig. 1

Creb3l1 mRNA exhibited high expression levels in polarized and mature odontoblasts. a Regulon of CREB3L1 in scRNA-seq profile for dental mesenchyme isolated from PN0 mouse lower molar.10 The CREB3L1 regulons were notably enriched in mature odontoblasts population. b–d RNAscope in situ hybridization was performed on the dental germ at embryonic day 15.5, 18.5 (E15.5, E18.5), and postnatal day 2.5 (PN2.5) developmental stages. The Creb3l1 mRNA was not expressed in dental mesenchymal cells but was specifically expressed in the epithelium and bone at E15.5 (b). Creb3l1 mRNA exhibited a high level of expression in the polarized odontoblasts and bone at E18.5 (c). Creb3l1 mRNA is localized in polarized and mature odontoblasts at PN2.5 (d). Scale bar = 100 μm

Micro-CT scans were conducted on mice at two separate developmental periods: postnatal 3 weeks (PN 3 W), the time of tooth root development, and postnatal 8 weeks (PN 8 W), the time of completed body development (Fig. 2a). The conditional deletion of Creb3l1 in the neural crest lineage demonstrated no effect on the number of teeth or crown morphology. The mandibular height below the root furcation of the first molar was analyzed (Supplementary Fig. S3a). It was found that this height was slightly decreased in cKO mice compared to WT mice at PN 3 W (Supplementary Fig. S3b), but significantly reduced at PN 8 W (Supplementary Fig. S3c). The mandibular alveolar bone (MAB) between the roots of the first molar (MAB1) and between the first and second molars (MAB2) were analyzed (Supplementary Figs. S4a, b, S5a, b). It was found that at PN 3 W, half of the cKO mice exhibited a significant decrease in bone mass. However, the other half did not exhibit any significant changes. Furthermore, there was no statistically significant difference in the bone volume (BV) to tissue volume (TV) ratio (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular separation (Tb.Sp) of cKO mice when compared with WT mice at PN 3 W (Supplementary Figs. S4c–f, S5c–f). A statistical analysis of all samples at PN 8 W revealed a statistically significant decrease in BV/TV in cKO mice in the mandible at both sites (Supplementary Figs. S4g, S5g). Additionally, a statistically significant reduction in trabecular number (Tb.N) was observed exclusively in the MAB1 in the PN 8 W cKO mice (Supplementary Fig. S4i). No statistically significant differences were observed for the remaining indicators (Supplementary Figs. S4h, j, S5h–j). Objectively, only a subset of the cKO mice exhibited a more pronounced reduction in bone mass when examined individually. The cementum of the first molar at PN 8 W exhibited a reduction in the amount of cementum in cKO mice compared to WT mice (Supplementary Fig. S6).

Fig. 2

Conditional deletion of Creb3l1 in neural crest lineage resulted in shorter root and thinner dentin. a Micro-CT scanning was performed to measure the mandibular first molar of PN 3 W and PN 8 W. Scale bar = 1 mm. b The measurement items, height of crown (violet line), crown dentin thickness (yellow line), root dentin thickness (red line), and height of root (green line) were included in the analysis. Scale bar = 1 mm. c HE staining was performed on wild-type (WT) and cKO mice at PN 3 W and PN 8 W. Scale bar = 100 μm. d The quantitative data of root-to-crown ratio from WT and cKO mice at PN 3 W. e The quantitative data of the root-to-crown ratio from PN 8 W mice. f The quantitative data of crown dentin thickness from WT and cKO mice at PN 3 W. g The quantitative data of crown dentin thickness from PN 8 W mice. h The quantitative data of root dentin thickness from WT and cKO mice at PN 3 W. i The quantitative data of root dentin thickness from PN 8 W mice. j The quantitative data of crown dentin density from PN 8 W mice. k The quantitative data of root dentin density from PN 8 W mice. n ≥ 5. ns not significant, P > 0.05; *P < 0.05; **P < 0.01; ****P < 0.000 1

The crown and root length, the thickness of crown and root dentin were analyzed to investigate potential abnormalities in tooth development20 (Fig. 2a, b). Shorter roots were observed in PN 3 W and PN 8 W cKO mice (Fig. 2a). The histological examination (Fig. 2c) and Micro-CT analysis of PN 3 W (Fig. 2d) and PN 8 W (Fig. 2e) mice provided further evidence to support this observation. Although there was a reduction in crown dentin thickness in cKO mice at PN 3 W in comparison to WT mice, this difference was not statistically significant (Fig. 2f). However, a significant difference was observed in the reduction in crown dentin thickness in cKO mice at PN 8 W (Fig. 2g). Both PN 3 W and PN 8 W cKO mice displayed thinner root dentin in contrast to WT mice (Fig. 2h, i).

Although there were macroscopic changes, it is unclear if any alterations in density and microstructure occurred in dentin. The Micro-CT analysis was employed to assess the density of dentin in both the crown and root of PN 8 W mice. The density of the crown and root exhibited no significant alterations (Fig. 2j, k). Scanning electron microscopy (SEM) was used to examine the dentinal tubules at the crown, cementoenamel junction (CEJ), and root. No significant changes in the dentinal tubule structure and dentinal tubules were observed (Fig. S7). The findings indicate that CREB3L1 is associated with total dentin deposition, but not with the density or structure of dentinal tubules.

Downregulation of Creb3l1 attenuated the odontoblastic differentiation capability of mDPCs

The RNAscope technique was employed to ascertain the mRNA expression of genes encoding proteins associated with odontoblast terminal differentiation in order to evaluate the potential role of CREB3L1 in regulating odontoblast differentiation at the transcriptional level in vivo. The expression of Dmp1 (Fig. 3a) and Dspp (Fig. 3b) in odontoblasts of cKO mice was observed to be diminished in comparison to WT mice in vivo, indicating that CREB3L1 may regulate the expression of Dmp1 and Dspp. To determine the mechanism underlying the shorter root and thinner dentin in cKO mice, the mDPC6T-Cas9 cell line,21,22 which maintained the differentiation capacity of primary cultured mouse dental papilla cells and constitutively expressed the CAS9 protein, was used to knock down Creb3l1 in vitro. The expression of CREB3L1 was initially verified during the differentiation induction of the mDPC6T-Cas9 cell line. The successful induction of differentiation was confirmed through the upregulation of differentiation-related proteins, including DMP1 and DSPP. Notably, the mRNA expression of Creb3l1 peaked on the fifth day after differentiation induction (Fig. 4a). The mRNA expressions of Dspp and Dmp1 increased during the differentiation induction (Fig. 4b, c) along with the increased protein levels (Fig. 4d, e). Since cleavage is required for CREB3L1 to function, the active form is the cleaved CREB3L1 fragment (Cleaved-CREB3L1). Western blot analysis detected two distinct bands, CREB3L1-FL (full-length CREB3L1) and Cleaved-CREB3L1 (Fig. 4d). Additionally, the highest expression of CREB3L1 was observed on the fifth day of differentiation induction as well (Fig. 4d, e). Furthermore, the role of CREB3L1 was explored by using CRISPR/CAS9 technology23 to knock out the Creb3l1 gene with a pair of single guide RNAs (sgRNAs) designated A8 (Fig. 4f, Supplementary Table S1). The P1 and P2 primers (Supplementary Table S1) were used to identify the chromosome that was successfully excised if a 553 bp fragment appeared (Fig. 4f). P3 and P4 primers (Supplementary Table S1) were used to identify the WT chromosome when 321 bp appeared (Fig. 4f). If both 553 bp and 321 bp were observed, the chromosome was considered to be incompletely cleaved and was regarded as a heterozygous cell line.

Fig. 3

Deficiency of CREB3L1 resulted in a significant decrease in the level of Dmp1 and Dspp in the first molar. a The RNAscope staining of Dmp1 in odontoblasts of the first molar of WT and cKO mice at PN 3 W. Scale bar = 100 μm. b The RNAscope staining of Dspp in odontoblasts of the first molar of WT and cKO mice at PN 3 W. Scale bar = 100 μm

Fig. 4

Downregulation of CREB3L1 inhibited the odontoblastic differentiation of mDPCs. a–c The mRNA expression level of Creb3l1, Dmp1, Dspp during differentiation induction. d The protein expression levels of the full length of CREB3L1 (CREB3L1-FL), Cleaved-CREB3L1, DMP1, and DSPP were detected during differentiation induction. e Quantification of the relative levels of protein expression in d. f Design of sgRNAs for Creb3l1 knockout. g Genotyping results for a monoclonal heterozygous cell line with knockdown of Creb3l1 (sgCreb3l1_A8-4). h–j The mRNA expression of Creb3l1, Dmp1, and Dspp decreased in sgCreb3l1_A8-4 cells after a 5-day induction of differentiation. k The expression of the CREB3L1-FL, Cleaved-CREB3L1, DMP1, and DSPP were decreased in sgCreb3l1_A8-4 during differentiation induction. l Quantification of the relative protein expression levels of k. m Alizarin red S staining was utilized to visualize calcium nodules in both the control (Ctrl) and sgCreb3l1_A8 groups after 14 days of differentiation induction. Scale bar = 1 cm, scale bar = 200 μm. n ≥ 3. ns, not significant, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.000 1

The monoclonal cell line was obtained via limiting dilution of the mDPC6T-Cas9 cell line treated with A8 sgRNAs. Two heterozygous monoclones, sgCreb3l1_A8-4 and sgCreb3l1_A8-8, were identified from the A8 sgRNA-treated cell pool (Supplementary Figs. S8a, Fig. 4g). The groups were divided into day 0 (D0), without differentiation induction, and day 5 (D5), with differentiation induction for 5 days. The mRNA levels of Creb3l1 (Fig. 4h), Dmp1 (Fig. 4i), and Dspp (Fig. 4j) were significantly decreased in sgCreb3l1_A8-4 cells on D5. These findings indicate that CREB3L1 was involved in the transcriptional regulation of Dmp1 and Dspp expression.

Finally, for the subsequent experiments, sgCreb3l1_A8-4 was selected for further studies. After 5 days of differentiation induction, the protein levels of CREB3L1, DMP1, and DSPP were found to be reduced in the sgCreb3l1_A8-4 group (Fig. 4k, l). A similar outcome was observed in the sgCreb3l1_A8-8 group (Supplementary Fig. S8b, c). The mineralized nodule formation ability of mDPC6T-Cas9 cells treated with A8 sgRNAs (sgCreb3l1_A8) after 14 days of differentiation induction (D14) was also impaired (Fig. 4m). These observations implied that CREB3L1 deficiency diminished the synthesis of dentin matrix proteins.

The deficiency of CREB3L1 altered chromatin accessibility and reduced gene expression associated with protein production and secretion

The established function of CREB3L1 is acting as a transcription factor.24 To validate this, immunofluorescence staining of N-terminal forms of CREB3L1 (CREB3L1-N) was conducted during the differentiation induction of mDPCs to examine the nuclear translocation of CREB3L1. The CREB3L1-N were predominantly localized in the nucleus after 12 and 24 h of differentiation induction (Fig. 5a). To elucidate the mechanism behind the impaired differentiation capacity of CREB3L1-deficient mDPCs, ATAC-seq and RNA-seq analyses were performed in Ctrl and CREB3L1-deficient mDPCs on D5. A significant change in chromatin accessibility was observed, with 888 regions exhibiting decreased accessibility and 510 regions exhibiting increased accessibility in the CREB3L1-deficient group (Fig. 5b, Supplementary Table S2). Gene Ontology (GO) analysis for the genes near the closed chromatin regions revealed a significant impact on biological processes including stem cell development, skeletal morphogenesis, regulation of cell-substrate adhesion, Notch signaling pathway, neural tube development, neural crest cell differentiation, mesenchymal development, epithelial-to-mesenchymal transition, and connective tissue development (Fig. 5c). The “mouse phenotype single KO” descriptions of these genes were associated with decreased neural crest cell number, decreased birth size, abnormal neural tube closure and abnormal craniofacial development (Fig. 5d). Downregulation of CREB3L1 resulted in decreased chromatin accessibility at the regulatory regions near the Dmp1 and Dspp genes (Fig. 5e, f).

Fig. 5

CREB3L1 deficiency led to substantial changes in chromatin accessibility and gene expression profile concerning protein production and secretion. a CREB3L1 acted as a transcriptional regulator by translocating to the nucleus during differentiation induction. Scale bar = 20 μm. b Heatmaps show the density of NFR summit-centered ATAC-seq signals in the Ctrl group and the sgCreb3l1_A8-4 group after 5 days of differentiation induction, respectively. The dot plot depicts the ATAC-seq outcomes, while the Gene Ontology (GO) enrichment analyses provided insight into the biological processes (c) and description of “Mouse Phenotype Single KO” (d) associated with the lost nucleosome-free regions (NFRs) in the sgCreb3l1_A8-4 group after 5 days of differentiation induction. e, f Visualization of ATAC-seq results for the Ctrl or sgCreb3l1_A8-4 cells after 5 days of differentiation induction from the UCSC Genome Browser. The blue box shows the closed NFRs in D5_sgCreb3l1_A8-4 near Dmp1 and Dspp. g Scatter plots show the significantly up-regulated or down-regulated genes identified by bulk RNA-seq analysis in D5_sgCreb3l1_A8-4. h The dot plot displays the GO enrichment for the downregulated genes in D5_sgCreb3l1_A8-4 from the bulk RNA-seq results. i The heatmap visualizes the differential expression of genes in specific GO terms identified by the bulk RNA-seq analysis

Apart from the chromatin accessibility changes, the difference in gene expression was also investigated. RNA-seq revealed that 1 007 genes were downregulated and 1 027 genes were upregulated in the CREB3L1-deficient mDPCs (P < 0.01) (Fig. 5g, Supplementary Table S3). Furthermore, GO analysis of the downregulated genes in CREB3L1-deficient mDPCs revealed their association with several cellular processes, including positive regulation of intracellular transport, such as Tmem30b;25,26,27,28 response to ER stress, such as Atf6b,29Creb3l1;13 positive regulation of protein phosphorylation, such as Bmp6,30Fgfr3,31,32Sox9;33,34,35 regulation of protein stability, such as Smad3,36Mdm2;37 regulation of collagen metabolic process, such as Notch1,38Itga2;39,40 extracellular matrix organization, such as Col11a1, an important component of collagen;41 intracellular protein transport, such as Notch1;42 growth hormone synthesis, secretion and action such as Atf6b,29Notch1;42 regulation of stem cell differentiation, such as Notch1,43Sox9;33,34,35 positive regulation of epithelial to mesenchymal transition, such as Notch1,43Smad344 (Fig. 5h, i). These findings are consistent with the reported role of CREB3L1 in ER stress, regulating transporter and extracellular matrix protein levels.45,46 These observations suggest that CREB3L1 altered chromatin accessibility and downregulated the expression of genes related to cell differentiation, protein biosynthesis, and protein secretion.

The CREB3L1-deficient mDPCs exhibited a weakened capacity for protein biosynthesis and secretion

Inspired by the top GO enriched term, we asked if loss of CREB3L1 leads to weakened capacity for protein biosynthesis and secretion. First, we chose to analyze marker protein for odontoblast terminal differentiation, DMP1 and DSPP. Consistent with the downregulation of RNA level for Dmp1 and Dspp in vitro and in vivo, the protein levels of DMP1 (Fig. 6a) and DSPP (Fig. 6b) in the odontoblasts of cKO mice were significantly lower than that of WT mice. Since the downregulation of DMP1 and DSPP protein may be due to the inhibition of transcription, we further investigate protein production in the knockout clone. The quantities of intracellular protein and supernatant secreted protein were compared between the Ctrl and sgCreb3l1_A8-4 groups. A significant reduction (30.74% in D0 and 41.97% in D5) in the quantity of intracellular proteins was observed in CREB3L1-deficient mDPCs (Fig. 6c). Notably, a more severe decrease (41.31% in D0 and 53.88% in D5) in the amount of cellular supernatant proteins was detected in CREB3L1-deficient mDPCs (Fig. 6d).

Fig. 6

Lack of CREB3L1 reduced the protein biosynthesis and export. a The DMP1 expression in odontoblasts of the first molar significantly decreased in PN 3 W cKO mice. Scale bar = 50 μm. b The DSPP expression in odontoblasts of the first molar significantly decreased in PN 3 W cKO mice. Scale bar = 50 μm. c The total amount of intracellular proteins and the percentage of decline to the Ctrl group. d The total quantity of extracellular supernatant proteins and the percentage of decline to the Ctrl group. e The protein degradation rates were measured after 24 h of exposure to 1 μg/mL cycloheximide in both the Ctrl and sgCreb3l1_A8-4 groups. n ≥ 3. ns not significant, P > 0.05; ***P < 0.001; ****P < 0.000 1

The reduction in total protein levels could be due to a reduced ability to produce protein or an increase in degradation. To determine whether the rate of protein degradation varied after inhibition of protein synthesis, cycloheximide (CHX)47 was applied to inhibit protein synthesis in the Ctrl and sgCreb3l1_A8-4 groups. Through a gradient concentration screen, it was determined that the application of 1 ug/mL CHX affected cellular protein synthesis with less cell death (Supplementary Fig. S9). Both Ctrl and sgCreb3l1_A8-4 groups were treated with this concentration of CHX for 24 h and the protein degradation rates before and after treatment were compared between the two groups. No significant difference was observed in the degradation rate between the two groups (Fig. 6e). These observations support the sequencing results that downregulation of CREB3L1 leads to a diminution of protein production and secretion.

CREB3L1 deficiency repressed TMEM30B expression in odontoblasts

In addition to its role as a transcription factor regulating the synthesis of mineralization-associated proteins, we are particularly interested in understanding how CREB3L1 regulates the protein secretion pathway. The BETA tool48 was utilized to identify the downregulated genes whose potential regulatory elements were also closed in ATAC-seq (Supplementary Tables S4, 5). To understand the role of TMEM30B, we initially confirmed its expression during the differentiation induction process. Its expression pattern was parallel to that of CREB3L1, peaking at D5 in mDPCs (Fig. 7a, b). Similarly, its expression pattern in dental papilla is similar to that of CREB3L1 in vivo (Supplementary Fig. S10). Additionally, in vivo histological analysis of the first molar from PN 3 W cKO mice revealed a significant reduction in TMEM30B expression levels (Fig. 7c).

Fig. 7

TMEM30B was identified to be regulated by CREB3L1 in odontoblasts. a The expression levels of TMEM30B protein significantly increased during the differentiation induction of mDPCs. b Quantification of the relative protein expression levels of a. c The expression of TMEM30B in odontoblasts of the first molar underwent a significant downregulation in PN 3 W cKO mice. d UCSC genome browser tracks demonstrate ATAC-seq peaks located at the Tmem30b locus. The blue box displays regions with closed chromatin in D5_sgCreb3l1_A8-4 near Tmem30b, while the black box exhibits the CREB family binding motif situated in the potential regulatory element region of Tmem30b. e Bar graph displays the results of the dual luciferase reporter assay. f The expression of the protein TMEM30B declined in the sgCreb3l1_A8-4 group after 5 days of differentiation induction. TMEM30B demonstrated predominant cytoplasmic expression. g Quantification of the relative protein expression levels of f. n ≥ 3. **P < 0.01; ***P < 0.001; ****P < 0.000 1

To determine if CREB3L1 can directly regulate Tmem30b transcription, a CREB family motif within Tmem30b located in the sgCreb3l1_A8-4 lost NFR was identified by JASPR website (Fig. 7d). Then, the corresponding region together with the regulatory element was cloned into the pGL3 promoter plasmid (T1) and the region without the motif (T1_mut). Dual-luciferase assays were conducted on the vector group, overexpressed Creb3l1 (oeCreb3l1)+T1 group, and oeCreb3l1 + T1_mut group. The results showed that overexpression of CREB3L1 could significantly upregulate the luciferase activity driven by the regulatory element of Tmem30b. However, the luciferase activity of the mutated Tmem30b regulatory element (oeCreb3l1 + T1_mut) group was downregulated compared to wildtype (oeCreb3l1 + T1) group. Thus, CREB3L1 is a positive regulator of this motif fragment (Fig. 7e).

The expression levels of TMEM30B protein were also detected during the 5-day differentiation induction of Ctrl and sgCreb3l1_A8-4 groups. As shown by the analysis of the isolated cytoplasm and nucleus, TMEM30B expression was exclusively observed in the cytoplasm and demonstrated a significant decrease in sgCreb3l1_A8-4 groups (Fig. 7f, g). Taken together, these observations suggest that TMEM30B is expressed in odontoblasts and regulated by CREB3L1.

TMEM30B deprivation impaired the protein synthesis and secretion capability of mDPCs

To examine the impact of TMEM30B deficiency on mDPC differentiation, Tmem30b expression was inhibited by siRNA (Fig. 8a). The suppression of Tmem30b resulted in a significant reduction in TMEM30B, DMP1, and DSPP protein levels during the differentiation process (Fig. 8b, c). Moreover, inhibiting Tmem30b greatly decreased the formation of mineralized nodules (Fig. 8d).

Fig. 8

TMEM30B deficiency impaired the odontoblastic differentiation capacity of mDPCs and reduced their intracellular and supernatant protein levels. a The mRNA expression level of Tmem30b was down-regulated in the siTmem30b group compared to the Ctrl group. b After 5 days of differentiation induction, the expression of TMEM30B, DMP1, and DSPP was reduced in mDPCs treated with siTmem30b when compared to the Ctrl group. c Quantification of the relative protein expression levels of b. d Alizarin red S staining was performed to observe the calcium nodules in both Ctrl and siTmem30b groups after 14 days of differentiation induction. Scale bar = 1 cm, scale bar = 100 μm. e Design of sgRNAs for Tmem30b knockout. f Genotyping results of the homozygous knockout cell line, sgTmem30b_60. g The total amount of intracellular proteins and the percentage of decline to the Ctrl group. h The total quantity of extracellular supernatant proteins and the percentage of decline to the Ctrl group. i The protein degradation rates were measured after 24 h of treatment with 1 μg/mL cycloheximide in both the Ctrl and sgTmem30b_60 groups. n = 3. ns, not significant, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.000 1

Targeted knockout of Tmem30b was performed using a pair of sgRNAs (sgRNA_a and sgRNA_b) (Fig. 8e) to evaluate the impact of Tmem30b deficiency on intracellular and supernatant proteins during differentiation induction. The P_a + P_b primers (Supplementary Table S6) were utilized to identify the wild-type (598 bp) and knockout (225 bp) chromosome (Fig. 8e). Subsequently, a homozygous knockout monoclone, sgTmem30b_60, was selected for further experiments (Fig.