Figure 1.
Schematic representation of genomic organization of the SEDL gene, illustrating the 44 identified mutations.
The human SEDL gene consists of six exons that span approximately 22 Kb of genomic DNA and encode a 140 amino acid protein. The 420bp coding region (open boxes) is organised into 4 exons (exon 3 to exon 6) and 3 introns (indicated by a line, not to scale). Non-coding exons (filled boxes) consist of exons 1 and 2, the 5′ portion of exon 3 and the 3′ portion of exon 6. The sizes of the exons, and the translation Start (ATG) and Stop (TGA) codons in exon 3 and 6, respectively, are indicated. The locations of the 44 mutations are indicated, and these consist of: 7 nonsense, 4 missense, 10 splice site, 1 insertion, 15 intraexonic deletions, and 7 deletions that encompassed introns and exons. The four missense mutations (Asp47Tyr, Ser73Leu, Phe83Ser and Val130Asp) and one nonsense mutation (Gln131Stop) were selected for further functional studies (Figures 3, 4, 6 and 7).
Figure 2.
Three-dimensional model of SEDLIN showing locations of residues involved in the mutations studied.
The model is based on a published model of mouse Sedlin [2]. (A) A ribbon model of Sedlin showing the alpha helices in red, beta strands in blue and the residues involved in the missense mutations (Asp47Tyr, Ser73Leu, Phe83Ser, and Val130Asp) in yellow. The bold line indicates the 10 C-terminal amino acids that would be deleted by the nonsense SEDLIN mutation (Gln131Stop) shown in yellow. The dashed circle indicates the hydrophobic core of Sedlin. The Ser119 and Ser124 residues that were predicted to be phosphorylation sites (NetPhos2.0, MotifScan and ELM databases) are indicated in green. (B) An ∼45° rotated view of the ribbon model of Sedlin showing the SEDT-associated mutated residues Asp47 and Ser73 (yellow), in SEDT, which together with the residues Tyr60, Thr63, His67 and Gln91 (light blue) aid in forming the hydrophobic groove (indicated by dotted bar line). The alpha helices are shown in red, beta strands in blue, and the linker region in grey.
Figure 3.
Subcellular co-localization of wild-type and mutant SEDLINs, MBP1, PITX1 and SF1.
(A) COS7 cells were transfected with wild-type or mutant (Asp47Tyr, Ser73Leu, Phe83Ser, Val130Asp and Gln131Stop) cMyc-SEDLIN constructs and visualised by immunofluorescence. The wild-type cMyc-SEDLIN (green) and mutant forms (data not shown) were found to localize to the nucleus and cytoplasm. DAPI, which stains nuclei (shown as red), colocalized with SEDLIN (yellow in merged image). (B) Wild-type or mutant HA-SEDLINs (red) were co-transfected with cMyc-MBP1, cMyc-PITX1 or cMyc-SF1 constructs and visualized by immunofluorescence (green). Wild-type SEDLIN and mutant SEDLINs (data not shown) co-localized (yellow) to the nucleus with the three transcription factors and to punctate structures within the cytoplasm with MBP1. (C) Western blot analysis of subcellular fractions (N - nuclear, C–cytoplasmic) from COS7 cells transiently co-transfected with wild-type (WT) or one of the 5 mutant HA-SEDLIN constructs, and cMyc-MBP1, cMyc-PITX1 or cMyc-SF1 constructs. Use of anti-HA antibody detected the expected 18 kDa wild-type and mutant SEDLIN proteins, and anti-cMyc antibody detected the expected 36 kDa, 36 kDa and 54 kDa MBP1, PITX1 and SF1 proteins, respectively, which were seen in the nuclear and cytoplasmic fractions, thereby confirming the immunofluorescence results. Western blots with anti-α-Tubulin and anti-Lamin A/C antibodies confirmed that the nuclear and cytoplasmic fractions were free from detectable amounts of cytoplasmic and nuclear fractions, respectively. Untransfected (UT) cells-not transfected with HA-SEDLIN. Wild-type and mutant SEDLINs were found in the nuclear and cytoplasmic fractions, and the SEDLIN mutants did not lead to an altered subcellular localization of MBP1, PITX1 and SF1. Scale bars, 10 µm.
Figure 4.
Interactions between SEDLIN, and MBP1, PITX1 and SF1.
Co-immunoprecipitation studies using COS7 cells demonstrated interactions between wild-type SEDLIN, mutant SEDLINs (Asp47Tyr, Ser73Leu, Phe83Ser, Val130Asp and Gln131Stop, only data from Ser73Leu SEDLIN shown) and MBP1, PITX1 and SF1. (A) cMyc-MBP1 co-transfected with wild-type or mutant HA-SEDLIN; (B) cMyc-PITX1 co-transfected with wild-type or mutant HA-SEDLIN; (C) cMyc-SF1 co-transfected with wild-type or mutant HA-SEDLIN; and (D) empty cMyc vector co-transfected with HA-SEDLIN or empty HA vector co-transfected with cMyc-MBP1, cMyc-PITX1 or cMyc-SF1 constructs. Lysates were incubated with either anti-cMyc polyclonal antibody (M) or anti-HA polyclonal antibody (H), or without any antibody as a negative control (−) and immunoprecipitated with Protein G-Sepharose beads. Protein complexes were eluted and resolved on SDS-PAGE followed by Western blot analysis using an antibody to the cMyc epitope for MBP1, PITX1 and SF1, and to the HA epitope for SEDLIN. Five percent of the lysate (input (I)) that was used for the immunoprecipitation was electrophoresed in parallel with the immunoprecipitated lysates. Wild-type SEDLIN co-immunoprecipitated MBP1, PITX1 and SF1 and the SEDLIN mutations (Asp47Tyr, Ser73Leu, Phe83Ser, Val130Asp and Gln131Stop) did not disrupt these interactions (representative data for mutant Ser73Leu SEDLIN shown).
Figure 5.
Schematic model for SEDLIN homodimerization and SEDL expression in common cell lines.
(A) Homodimers that have formed between the transfected mutant cMyc-SEDLIN and the endogenously expressed wild-type SEDLIN could mask the loss of interaction between mutant SEDLINs and MBP1, PITX1 and SF1 in transfected COS7 cells. Thus, although the mutant SEDLIN may not directly interact with MBP1, PITX1 or SF1, the endogenously expressed SEDLIN will interact with these transcription factors, and hence the resultant homodimers will overall be seen to interact with the transcription factors. This proposed model provides an explanation for the observed results in the transfected cells. However, it is important to note that this situation of homodimers consisting of a wild-type and mutant SEDLIN would not normally occur in males affected with SEDT, as they are hemizygous and their cells would normally express the mutant SEDLIN; however this situation would occur in SEDT heterozygous carrier females because the SEDL gene escapes X-chromosome inactivation [13] and hence their cells would express both the wild-type and mutant SEDLINs. (B) RT-PCR analysis was used to detect the endogenous expression of SEDL in COS7, COS1, HEK293 and HK2 kidney cells, as a reliable SEDLIN antibody is not available. Detection of Calmodulin expression was used as an internal control for RNA quality and concentration; (+) with RTase, (−) without RTase.
Figure 6.
(A) Western blot analysis, using a monoclonal anti-cMyc antibody, of cell lysates obtained from COS7 cells transfected with cMyc-SEDLIN or cMyc vector alone, and resolved by continuous non-denaturing PAGE; lanes 1-3, cytoplasmic fractions; lanes 4–6, nuclear fractions; lane 1 and 4, transfected with cMyc-SEDLIN wild-type (WT) construct; lanes 2 and 5, transfected with cMyc vector alone; and lanes 3 and 6, untransfected (UT) cells. cMyc-SEDLIN, an 18 kDa protein (Figures 3 and 4) appeared as a 36 kDa protein on non-denaturing PAGE of cytoplasmic and nuclear fractions obtained from COS7 cells transfected with the cMyc-SEDLIN wild-type (WT) construct, consistent with homodimerization of SEDLIN protein. (B) Co-immunoprecipitation of wild-type (WT) cMyc-SEDLIN with wild-type or mutant HA-SEDLINs (data for mutant Ser73Leu, shown). Anti-cMyc antibody co-immunoprecipitated wild-type HA-SEDLIN in the presence of cMyc-SEDLIN (lane M) and anti-HA antibody co-immunoprecipitated cMyc-SEDLIN in the presence of HA-SEDLIN (lane H); in the absence of the anti-cMyc or anti-HA antibodies SEDLIN was not immunoprecipitated (lane -). Five percent of the lysate (input (I)) that was used for the immunoprecipitation was electrophoresed in parallel with immunoprecipitated lysates. Similar results were observed for the other SEDLIN mutants (Asp47Tyr, Phe83Ser, Val130Asp and Gln131Stop) (data not shown).
Figure 7.
Interactions between SEDLIN, MBP1, PITX1 and SF1 using the yeast two-hybrid assay.
Yeast cells, which do not have endogenous expression of SEDLIN, were used to investigate the interactions of wild-type or mutant SEDLIN (Asp47Tyr, Ser73Leu, Phe83Ser, Val130Asp and Gln131Stop) with wild-type SEDLIN, MBP1, PITX1 or SF1. The yeast reporter strain AH109 was used, and p53 and the SV40 large T antigen, which are known to interact [27], were used as a positive control. The yeasts were transformed with the vectors containing: (A) wild-type SEDLIN in pGADT7-AD (AD-WT) and either wild-type or mutant SEDLINs in pGBKT7-BD (BD-WT, BD-Asp47Tyr, BD-Ser73Leu, BD-Phe83Ser, BD-Val130Asp or BD-Gln131Stop). (B) wild-type and mutant AD-SEDLINs or BD-SEDLIN and each of the transcription factors BD-MBP1, AD-PITX1, AD-SF1. Yeast growth was monitored for 48 hrs after spotting and incubation at 30°C using either double drop out, DDO (Leu-Trp-), media as a control or quaternary drop out, QDO (Leu-Trp-Ade-His-), media in which the growth is dependent on the physical interaction between BD-SEDLIN and AD-transcription factors, or AD-SEDLIN and the BD-transcription factor. The wild-type SEDLIN interacts with wild-type SEDLIN and all of the mutant SEDLIN proteins, consistent with the proposed model for the formation of homodimers. However, the MBP1, PITX1 and SF1 fusion proteins, which interact with the wild-type SEDLIN, interacted only with the mutant Asp47Tyr SEDLIN, but not with the mutant Ser73Leu, Phe83Ser, Val130Asp and Gln131Stop SEDLINs.