European Journal of Medical Genetics 58 (2015) 358e363
Contents lists available at ScienceDirect
European Journal of Medical Genetics journal homepage: http://www.elsevier.com/locate/ejmg
Clinical report
A new hereditary congenital facial palsy case supports arg5 in HOXDNA binding domain as possible hot spot for mutations Zehra Oya Uyguner*, Güven Toksoy, Umut Altunoglu, Hilal Ozgur, Seher Basaran, Hülya Kayserili Department of Medical Genetics, Istanbul Medical Faculty, Istanbul University, Istanbul, 34093 Turkey
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 January 2015 Accepted 18 May 2015 Available online 23 May 2015
Moebius syndrome (MBS) is a rare congenital disorder characterized by rhombencephalic mal development, mainly presenting with facial palsy with limited gaze abduction. Most cases are sporadic, possibly caused by a combination of environmental and genetic factors; however, no proven specific associations have been yet established. Hereditary congenital facial palsy (HCFP) is an autosomal dominant congenital dysinnervation syndrome, recognizable by the isolated dysfunction of the seventh cranial nerve. Mutant mice for Hoxb1 were reported to present with facial weakness, resembling MBS. Recently a homozygous mutation altering arg5 residue of HOXB1 homeodomain into cys5 was identified in two families with HCFP. We screened 95 sporadic patients diagnosed as MBS or HCFP for mutations in HOXB1. A novel homozygous alteration was identified in one HCFP case, affecting the same residue, resulting to his5. In silico protein analysis predicted stronger HOXB1-DNA binding properties for his5 than cys5 that resulted to milder phenotype. It should be noted that, inclusive of the previous report, only two mutations revealed in HOXB1 associated with HCFP involved the same amino acid arg5 in HOXB1 residing in HOXB1-DNA-PBX1 ternary complex. Ó 2015 Elsevier Masson SAS. All rights reserved.
Keywords: Hereditary congenital facial palsy Moebius syndrome HOXB1
1. Introduction A high percentage of MBS cases are sporadic and no specific genes are yet identified related to this condition. Two nested factors may delineate the fruitless efforts to identify the causative genes for MBS. First, linkage algorithms are generally powerless for sporadic occurrences due to genetic heterogeneity. Second, uncharacterized heredibility patterns may complicate the linkage analysis, since autosomal dominant with incomplete penetrance, autosomal recessive with homozygous or compound heterozygous mutations, or de novo heterozygous mutations, could all be under possibility. With improvements in diagnostic assessments, leading to the subdivision of patients according to their specific clinical features, may facilitate both the identification and the association of novel genes extracted from exome or whole genome sequencing data (MacKinnon et al., 2014). In 1996, it was reported that null mutant mice for Hoxb1, with disrupted development of the facial nerve, originating from the
* Corresponding author. Istanbul University, Istanbul Medical Faculty, Department of Medical Genetics, Capa, 34093, Istanbul, Turkey. E-mail address: [email protected] (Z.O. Uyguner). http://dx.doi.org/10.1016/j.ejmg.2015.05.003 1769-7212/Ó 2015 Elsevier Masson SAS. All rights reserved.
fourth rhombomere, had features resembling Moebius syndrome in humans (Goddard et al., 1996, Studer et al., 1996). Recently, a homozygous mutation in HOXB1, c.619C > T, altering arginine to cysteine at 207 (p.Arg207Cys), corresponding to arg5 > cys5, identified in two unrelated patients of GermanAmerican descent with bilateral facial weakness, feeding difficulties and hearing loss. Both affecteds share the same infrequent haplotype suggestive for a founder mutation implicated that the families can be distantly related (Webb et al., 2012). Thereupon, HOXB1 entitled to be the first gene associated with HCFP, addressed to HCFP3 (MIM#614744) at 17q21.32. HOXB1 is a homeodomain containing sequence-specific transcription factor of the HOX family, under the HOXL subgroup of ANTP-class (Holland et al., 2008). It has a characteristic helix-turnhelix DNA binding motif with three alpha helical regions (a1, a2, a3) where the specificity may be contemplated by heterodimerization with PBX1 (Piper et al., 1999). Yeast two-hybrid assays showed that Pbx homeodomain (TALE class) is necessary but not adequate for cooperation, which required conserved amino acids (Chang et al., 1995). Nuclear Magnetic Resonance (NMR) studies demonstrated that the conserved hexapeptide of Hoxb1 (TFDWMK) stabilizes binding of Pbx1 and Hoxb1 to DNA (Carolyn et al., 2001). The fifth amino acid of HOXB1, arginine (arg5),
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
demonstrated to confer DNA binding specificity to the first thymidine in the DNA recognition motif of 50 -1T-A-A-T-T/G-A/G-30 by hydrogen bonding and also by hydrophobic packing interactions (Wilson et al., 1996, Noyes et al., 2008). We report here a HCFP case with a novel homozygous mutation in the HOXB1 gene and present the results of in silico protein modeling for the prediction of protein function to investigate if compatible with phenotype. 2. Material and methods 2.1. Patients A total of 39 sporadic MBS/HCFP patients were enrolled for candidate gene screening. The investigation was approved by the Institutional Review Board of Istanbul Medical Faculty, Istanbul University. Written informed consent was obtained from patients and from their parents before sampling commenced. Additionally, DNA samples from 56 Moebius syndrome patients provided as Dutch cohort referred from Department of Human Genetics, Radboud University, Nijmegen Medical Centre, Netherlands, were also included. 2.2. Molecular analysis Genomic DNA was extracted from 2 ml of K3EDTA blood samples using DNA isolation kit (Mammalian Blood, Roche). Primers covering 50 and 30 of coding exons plus flanking introns designed for HOXB1 (NM_002144, NP_002135) gene and PCR for each region was performed in 50 mL reaction mixture, composed of 0.3 ml primer pairs, 1X reaction buffer, 2 mM MgCl2, 0.2 mM dNTP mix in 0.5 unite taq polymerase enzyme (Fermentas), in two step cycle (Tm of primers) program in thermal cycler (Bio Rad), starting with 10 min initial denaturation, followed by 30 s denaturation, 30 s annealing and 1 min/1 kb elongation periods. PCR products were purified by spin colon kit (High Pure PCR product purification kit, Roche) and sequenced (Macrogen, Seoul, Korea). Electropherograms were analyzed for sequence variants by using publicly available BLAST program (NCBI, USCS). DNA samples from 100 healthy control individuals were screened for c.620G > A alteration in HOXB1 gene by PCR and mutation specific restriction digestion.
359
2.3. Protein models Three-dimensional models for the full HOXB1 protein and mutants were created based on multiple-threading alignments by Local Meta-Threading-server (LOMET) by using in silico tool, I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER). HOXB1-HD domains for normal and mutants were created by a homology modeling tool, SWISS-MODEL (http://swissmodel.expasy.org). A HOXB1-DNAPBX1 model was downloaded from the Protein Data Bank in EuropeEBI website for the examination of P14653 (http://www.ebi.ac.uk/ pdbe-srv/view/entry/1b72/downloads.html) (Piper et al., 1999, Zhang, 2008, Roy et al., 2010). I-Tasser builds five predictions with individual confidence scores estimating the quality of the models based upon the capacity of the threading template alignments and the convergence parameters of the structure assembly simulations. Typically, the confidence score of 3D models in this system ranges between 5 and þ2, where the higher value signifies greater confidence. Among the five strongest 3D models created by the server, the highest confidence score recorded was 3.40 for arg5, 3.48 for cys5 and 3.58 for his5. This was the model selected and used for the analysis. Quality structure estimation of protein models built by the Swiss Model is stated by the Qualitative Model Energy Analysis 4 (QMEAN4) scoring system consisting of four statistical potential terms: C-beta interaction energies; all atom pair wise energy; solvation energy, and; torsion angle energy. Reliable values range between 0 and 1 (Arnold et al., 2006; Benkert et al., 2011). The created models recorded a QMEAN4 score of 0.544 for arg5, 0.5 for cys5 and 0.536 for his5.
3. Results 3.1. Clinical The patient, 23 months old female, was referred from pediatric neurology outpatient clinics due to facial paralysis with a preliminary diagnosis of Moebius syndrome. She was the only child of first-degree cousins, once removed, born at term via Caeserian section due to cephalo pelvic disproportion, with a weight of 3320 g (0.19SD). The child’s length and occipito frontal circumference (OCF) at birth were not recorded. Developmental milestones were achieved normally; the patient sat without support at seven
Fig. 1. (A) Pedigree of the family and facial appearance of the patient at the age of 23 months, 38 months and seven years of age, respectively (Note bilateral facial paralysis). (B) Partial electropherogram showing the mutation in HOXB1gene.
360
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
months of age, walked independently at 11 months, had first words at 12 months, and two-word sentences at 22 months. On physical examination at the age of 23 month, height was recorded as 81 cm (1.42 SD), weight 10.3 kg (1.04 SD) and OFC 48 cm (0.09SD). She had midface retrusion, left-sided ptosis with bilateral lagophtalmos, depressed nasal bridge, short nose with anteverted nares, tented vermilion of the upper lip, down turned corners of the mouth, high palate, and low-set ears (Fig. 1A). Left concomitant esotropia was noted, with mild inferior oblique muscle hyper function. There was no paralytic component of the
sixth cranial nerve. The patient was unable to smile or frown, and nasolabial folds were flattened. Physical and neurological examinations were otherwise unremarkable. These findings prompted for bilateral lower motor neuron palsy of the seventh cranial nerve. Auditory brainstem response test revealed normal result. Cranial MRI scans revealed minor dilatation of the temporal horns, and no significant structural anomalies or focal lesions. Karyotype was normal. At the age of seven, she was reported to have outstanding school performance and she had a two-years-old healthy brother.
Fig. 2. Genomic layout with homeodomain structure presented with GROMOS empirical force field energy for each amino acid of the HOXB1 models created for arg5, cys5 and his5. Organization of HOXB1 gene modified from its assembly at www.ensembl.org/Homo_sapiens presented at uppermost segment. Hexapeptide motif of HOXB1, found at N terminal hexapeptide, and three helical structures of its homeodomain presented with codon numbers shown at top and homeodomain peptide numbers shown below the simulated protein architecture. Peptide sequence of homeodomain with common conserved amino acids is separately colored blue for the HOXL subclass and yellow for the entire homeodomain class. The location of identified missense mutation his5 is presented as red dot on the architecture and also marked by arrow on the peptide sequence. Qmean Score 4 is shown at the upper right corner for each model. The X axis represents the amino acid sequence and the Y axis represents the energy for each amino acid in the structure. Negative energy values, shown in green, represent favorable whereas positive values, shown in red, indicate an unfavorable energy environment for a given amino acid. Hence, revert a movement of favorable energy for arg5 (w230) to unfavorable when is replaced by cysteine (w10) and more pronouncedly unfavorable when replaced by histidine (w3).
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
361
Fig. 3. Three-dimensional modeling of entire HOXB1 built for arg5, cys5 and his5, by I-Tasser. Confidence scores for each model are presented at the right lower corner. The model is visualized by PyMol, with rainbow painting starting from dark blue at ‘N terminal’ and ending red at ‘C terminal’. The N terminal hexapeptide-located codons between 179e184 or ‘24 to 16’ relative to homeodomain presented by molecular stick presentation and highlighted in red.
3.2. Molecular Sequencing of HOXB1 in total of 95 MBS/HCFP cases revealed homozygous c.620G > A, altering arginine to histidine at 207 (p.Arg207His) in one female patient. The parents of the patient and her healthy brother were heterozygous for this variation (Fig. 1B). Prediction analysis by silica algorithms, MutationTaster, Polyphen-2, and SIFT, revealed this alteration as the cause of the disease and its ascribing facial damage (Adzhubei et al., 2010; Schwarz et al., 2010; http://sift.bii.a-star.edu.sg). Furthermore, 200 alleles in control population did not carry this alteration. 3.3. Protein The empirical force field energy of arg5, cys5 and his5, respectfully, were compared by the computer program GROMOS (GROningen Molecular Simulation) via a Swiss-Model server upon web-
based integrated homology modeling study (Christen et al., 2005). The template was drawn from Drosophila melanogaster (1b72) and presented with w77% homology with the human ortholog, HOXB1. It was observed that the favorable energy environment of arginine reverted to unfavorable, when accommodated with cysteine and histidine (Fig. 2). I-Tasser modeling of each HOXB1 peptide containing one of the three, arg5, cys5 or his5, did not show critical differences between a1, a2 and a3 regions and in the secondary structure of the N-terminal hexapeptide conformation (Fig. 3). The number of predicted hydrogen bonding contacts of DNA molecule decreased from three to one when arginine was replaced by cysteine, and increased to four when replaced by histidine. Comparison of the mean values of the measurements between the oxygen of C2 of T11 with the hydrogens of the residues suggests that cys5 may have the weaker binding interaction than histidine (Fig. 4).
Fig. 4. HOXB1-PBX1-DNA ternary complex. Models of arg5 (A), cys5 (B), his5 (C) residues with proximity to thymidine 11 of DNA demonstrated by stick model for easy identification. The predicted hydrogen bond interactions between the residues and the oxygen of the second carbon of 11th thymidine of the DNA molecules are presented with closer shot of the region presented at D for arg5, E for cys5 and F for his5. The measured distances of 3.7 A to ε, 2.6 Å to h1 and 4.8 Å to h2 from the hydrogen atoms of arginine to oxygen (D); the distance of 8.3 Å from the hydrogen of sulfur of cycsteine to oxygen (E), and the distances of 7.3 Å to d1, 7.9 Å to ε1, 6.3 Å to d2 and 7.1 Å to ε2 from the hydrogen atoms of histidine to oxygen (F).
362
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
4. Discussion The prevalence of MBS and HCFP is currently unknown, but todate 39 sporadic individuals have been referred to our outpatient clinics since 1995 (Uzumcu et al., 2009, 2009). HOXB1 was the most relevant candidate gene, not only for its orthologue’s association with facial weakness in Hoxb1/ mouse, but also for its association with a human phenotype. HOXB1 gene screening revealed mutation altering arg5 to his5 in the coding peptide in a single case. This alteration was neither identified in our screened 100 healthy control individuals nor in 100 in-house groups of AGBRG (Advanced Genomics and Bioinformatic Research Group of National Research Institute of Electronics and Cryptology of Turkey) totaling up to 400 nationally matched alleles. Three-dimensional modeling studies comparing arg5 with published cys5, and here reported his5, predicted that the a helical structures of homeodomain of HOXB1 are sustained. Hydrogen bonding between the polar molecules and electronegative atoms, such as oxygen, are considered specific mediators of a proteinnucleic acid recognition mechanism (Sathyapriya et al., 2008). Side chains of residues form most of the crucial contacts through bidentate and complex hydrogen-bonding interactions with DNA base edges, hence conferring remarkable specificity (Luscombe and Thornton, 2002). It is known that arginine at position five of the homeodomain, which is conserved in 95% of known homeodomains, interacts with the core of 50 -TAAT-30 sequence. Comparison of the distances of the predicted hydrogen bonds between the C2 oxygen atom of thymidine (T11) and the side chains of arg5 to cys5 and to his5, suggests that his5 may present stronger bonds than cys5. Our measurement revealed 8.3 Å for the distance of hydrogen bonding between T11 and sulfhydryl group of cysteine; however, the distance has been reported to be 6 Å in published data by Webb et al. (2012). Ternary complex revealed that arg5 has three but shortest, cys5 has single and longest, his5 has four but longer connections, predicting that arg5 > his5 > cys5 binding interactions. Physicochemical specification of histidine is less basic (pH 7.59) than arginine (pH 10.76), but it shares common properties, including a positively charged polar basic format. As a result, the impact of his5 on the HOXB1 structure and its function may be milder than that of cys5, which correlates with the displayed
Table 1 Clinical findings of the four previously reported HCFP patients with homozygous mutation in HOXB1, compared to present case.
Origin Molecular findings NM_002144 NP_002135 Facial findings Masked facies Bilateral facial weakness Full oculomotor motility Esotrophia/esophoria Ptosis Smooth philtrum Midface retrusion Upturned nasal tip Low-set ears Micrognathia Other findings Feeding difficulty Sensorineural hearing loss Abnormal cochlear function Cochlear structural defect n.a. Not available.
Present case
Webb et al. (2012)
Turkish
German American
c.620G > A p.Arg207His
c.619C > T p.Arg207Cys
þ þ þ þ þ left e þ þ þ e
4/4 4/4 4/4 4/4 e 2/4 4/4 4/4 3/4 2/4
e e e n.a.
2/4 4/4 2/2 1/1
phenotypes (Table 1). Our patient had neither feeding difficulties nor sensorineural hearing loss as with previously reported cases. Heterozygous mutations involving arg5 in two other HOXL subclass homeobox genes are responsible for dominant disease phenotypes. Two patients with Currarino syndrome (MIM 142994) carrying heterozygous p.Arg245Gly and p.Arg245His in the HLXB9 (MNX1) gene, and one patient with Hand-Foot-Uterus (MIM 142959) syndrome with heterozygous p.Arg326Gly in HOXA13, were previously reported (Hagan et al., 2000, Parker et al., 2011). Out of 272 families with MBS or HCFP, 177 previously reported by Webb et al. (2012), along with 95 samples (39 from Turkey and 56 from Netherlands), only three cases were found to carry mutations in HOXB1, establishing the frequency of HOXB1 mutations to be 1.1% in patients afflicted with congenital facial paralysis, with or without the additional features of MBS. 5. Conclusion Arg5 of the homeodomain of HOXB1 is suggested to be a labile mutation site for HCFP. Stronger binding interactions of his5 than cys5 leading to a stronger HOXB1-DNA-PBX1 correlating with milder phenotype may propose that in silico protein modeling studies may provide qualified predictions for the consequences of the human mutations on the structure and function of the encoding genes and contribute to our understanding of the molecular pathology underlying the disease phenotype. Acknowledgments We would like to thank all of the participating patients and their families for their support in this ongoing research. We appreciate AGBRG (IGBAM, Bilgem, TUBITAK) for sharing with us their massively parallel sequencing results of 100 private individuals. We also wish to express our gratitude to Dr. Hans van Bokhoven for his kind contribution of Dutch cohort samples. This work received partial support from Istanbul University-BAP-10801, and full support from the Scientific and Technological Research Council of Turkey (TUBITAK, grant number 108S418 [CRANIRARE] and 112S398 [CRANIRARE-2]. The CRANIRARE consortium is supported by the European Research Area Network (E-RA net programs). All co-authors have read the manuscript agreed with its content and stated that there is no conflict of interest. References Adzhubei A, Steffen S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS, Sunyaev SR. A method and server for predicting damaging missense mutations. Nat. Methods 2010;7:248e9. Arnold K, Bordoli L, Kopp J, Schwede T. The SWISS-MODEL workspace: a web-based environment for protein structure homology modeling. Bioinformatics 2006;22:195e201. Benkert P, Biasini M, Schwede T. Toward the estimation of the absolute equality of individual protein structure models. Bioinform. 2011;27:343e50. Carolyn M, Slupsky CM, Sykes DB, Gay GL, Sykes BD. The HoxB1 hexapeptide is a prefolded domain: implications for the Pbx1/Hox interaction. Protein Sci. 2001;10:1244e53. Chang CP, Shen WF, Rozenfeld S, Lawrence HJ, Largman C, Cleary ML. Pbx proteins display hexapeptide-dependent cooperative DNA binding with a subset of Hox Proteins. Genes Dev. 1995;9:663e74. Christen M, Hünenberger PH, Bakowies D, Baron R, Bürgi R, et al. The GROMOS software for biomolecular simulation: GROMOS05. J. Comput. Chem. 2005;26: 1719e51. Goddard JM, Rossel M, Manley NR, Capecchi MR. Mice with targeted disruption of Hoxb-1 fail to form the motor nucleus of the VIIth nerve. Development 1996;122:3217e28. Hagan DM, Ross AJ, Strachan T, Lynch SA, Ruiz-Perez V, et al. Mutation analysis and embryonic expression of the HLXB9 currarino syndrome gene. Am. J. Hum. Genet. 2000;66:1504e15.
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363 Holland PWH, Booth HAF, Bruford EA. Classification and nomenclature of all human homeobox genes. BMC Biol. 2008;5:1e29. Luscombe MN, Thornton JM. ProteineDNA interactions: amino acid conservation and the effects of mutations on binding specificity. J. Mol. Biol. 2002;320:991e 1009. MacKinnon S, Oystreck DT, Andrews C, Chan WM, Hunter DG, Elizabeth C, Engle EC. Diagnostic distinctions and genetic analysis of patients diagnosed with moebius syndrome. Ophthalmology 2014;121:1461e8. Noyes MB, Christensen RG, Wakabayashi A, Stormo GD, Brodsky MH, Wolfe SA. Analysis of homeodomain specificities allows the family-wide prediction of preferred recognition sites. Cell 2008;133:1277e89. Parker L, Mangwani J, Wakeling E, Singh D. Hallux valgus interphalangeus and a novel mutation in HOXA13. Part of the broadening spectrum of hand-footgenital syndrome. Foot Ankle Surg. 2011;17:29e30. Piper DE, Batchelor AH, Chang CP, Cleary ML, Wolberger C. Structure of a HoxB1e Pbx1 heterodimer bound to DNA: role of the hexapeptide and a fourth homeodomain helix in complex formation. Cell 1999;96:587e97. Roy A, Kucukural A, Zhang Y. I-TASSER: a unified platform for automated protein structure and function prediction. Nat. Protoc. 2010;5:725e38. Sathyapriya R, Vijayabaskar R, Vishveshwara S. Insights into proteineDNA interactions through structure network analysis. PLoS Comput. Biol. 2008;4:1e15.
363
Schwarz JM, Rödelsperger C, Schuelke M, Seelow D. Mutation taster evaluates disease-causing potential of sequence alterations. Nat. Methods 2010;7:575e6. Studer M, Lumsden A, Ariza-McNaughton L, Bradley A, Krumlauf R. Altered segmental identity and abnormal migration of motor neurons in mice lacking Hoxb-1. Nature 1996;384(6610):630e4. Uzumcu A, Karaman B, Toksoy G, Uyguner ZO, Candan S, et al. Molecular genetic screening of MBS1 locus on chromosome 13 for micro deletions and exclusion of FGF9, GSH1, and CDX2 as causative genes in patients with moebius syndrome. Eur. J. Med. Genet. 2009a;52:315e20. Uzumcu A, Candan S, Toksoy G, Uyguner ZO, Karaman B, et al. Mutational screening of BASP1 and transcribed processed pseudogene TPJg-BASP1 in patients with moebius syndrome. J. Genet. Genomics 2009b;36:251e6. Webb BD, Shaaban S, Gaspar H, Cunha LF, Schubert CR, et al. HOXB1 founder mutation in humans recapitulates the phenotype of Hoxb1-/- mice. Am. J. Hum. Genet. 2012;91:171e9. Wilson DS, Sheng G, Jun S, Desplan C. Conservation and diversification in homeodomain-DNA interactions: a comparative genetic analysis. Proc. Acad. Sci. 1996;93:6886e91. Zhang Y. I-TASSER server for protein 3D structure prediction. BMC Bioinform. 2008;9:40e8.
Contents lists available at ScienceDirect
European Journal of Medical Genetics journal homepage: http://www.elsevier.com/locate/ejmg
Clinical report
A new hereditary congenital facial palsy case supports arg5 in HOXDNA binding domain as possible hot spot for mutations Zehra Oya Uyguner*, Güven Toksoy, Umut Altunoglu, Hilal Ozgur, Seher Basaran, Hülya Kayserili Department of Medical Genetics, Istanbul Medical Faculty, Istanbul University, Istanbul, 34093 Turkey
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 January 2015 Accepted 18 May 2015 Available online 23 May 2015
Moebius syndrome (MBS) is a rare congenital disorder characterized by rhombencephalic mal development, mainly presenting with facial palsy with limited gaze abduction. Most cases are sporadic, possibly caused by a combination of environmental and genetic factors; however, no proven specific associations have been yet established. Hereditary congenital facial palsy (HCFP) is an autosomal dominant congenital dysinnervation syndrome, recognizable by the isolated dysfunction of the seventh cranial nerve. Mutant mice for Hoxb1 were reported to present with facial weakness, resembling MBS. Recently a homozygous mutation altering arg5 residue of HOXB1 homeodomain into cys5 was identified in two families with HCFP. We screened 95 sporadic patients diagnosed as MBS or HCFP for mutations in HOXB1. A novel homozygous alteration was identified in one HCFP case, affecting the same residue, resulting to his5. In silico protein analysis predicted stronger HOXB1-DNA binding properties for his5 than cys5 that resulted to milder phenotype. It should be noted that, inclusive of the previous report, only two mutations revealed in HOXB1 associated with HCFP involved the same amino acid arg5 in HOXB1 residing in HOXB1-DNA-PBX1 ternary complex. Ó 2015 Elsevier Masson SAS. All rights reserved.
Keywords: Hereditary congenital facial palsy Moebius syndrome HOXB1
1. Introduction A high percentage of MBS cases are sporadic and no specific genes are yet identified related to this condition. Two nested factors may delineate the fruitless efforts to identify the causative genes for MBS. First, linkage algorithms are generally powerless for sporadic occurrences due to genetic heterogeneity. Second, uncharacterized heredibility patterns may complicate the linkage analysis, since autosomal dominant with incomplete penetrance, autosomal recessive with homozygous or compound heterozygous mutations, or de novo heterozygous mutations, could all be under possibility. With improvements in diagnostic assessments, leading to the subdivision of patients according to their specific clinical features, may facilitate both the identification and the association of novel genes extracted from exome or whole genome sequencing data (MacKinnon et al., 2014). In 1996, it was reported that null mutant mice for Hoxb1, with disrupted development of the facial nerve, originating from the
* Corresponding author. Istanbul University, Istanbul Medical Faculty, Department of Medical Genetics, Capa, 34093, Istanbul, Turkey. E-mail address: [email protected] (Z.O. Uyguner). http://dx.doi.org/10.1016/j.ejmg.2015.05.003 1769-7212/Ó 2015 Elsevier Masson SAS. All rights reserved.
fourth rhombomere, had features resembling Moebius syndrome in humans (Goddard et al., 1996, Studer et al., 1996). Recently, a homozygous mutation in HOXB1, c.619C > T, altering arginine to cysteine at 207 (p.Arg207Cys), corresponding to arg5 > cys5, identified in two unrelated patients of GermanAmerican descent with bilateral facial weakness, feeding difficulties and hearing loss. Both affecteds share the same infrequent haplotype suggestive for a founder mutation implicated that the families can be distantly related (Webb et al., 2012). Thereupon, HOXB1 entitled to be the first gene associated with HCFP, addressed to HCFP3 (MIM#614744) at 17q21.32. HOXB1 is a homeodomain containing sequence-specific transcription factor of the HOX family, under the HOXL subgroup of ANTP-class (Holland et al., 2008). It has a characteristic helix-turnhelix DNA binding motif with three alpha helical regions (a1, a2, a3) where the specificity may be contemplated by heterodimerization with PBX1 (Piper et al., 1999). Yeast two-hybrid assays showed that Pbx homeodomain (TALE class) is necessary but not adequate for cooperation, which required conserved amino acids (Chang et al., 1995). Nuclear Magnetic Resonance (NMR) studies demonstrated that the conserved hexapeptide of Hoxb1 (TFDWMK) stabilizes binding of Pbx1 and Hoxb1 to DNA (Carolyn et al., 2001). The fifth amino acid of HOXB1, arginine (arg5),
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
demonstrated to confer DNA binding specificity to the first thymidine in the DNA recognition motif of 50 -1T-A-A-T-T/G-A/G-30 by hydrogen bonding and also by hydrophobic packing interactions (Wilson et al., 1996, Noyes et al., 2008). We report here a HCFP case with a novel homozygous mutation in the HOXB1 gene and present the results of in silico protein modeling for the prediction of protein function to investigate if compatible with phenotype. 2. Material and methods 2.1. Patients A total of 39 sporadic MBS/HCFP patients were enrolled for candidate gene screening. The investigation was approved by the Institutional Review Board of Istanbul Medical Faculty, Istanbul University. Written informed consent was obtained from patients and from their parents before sampling commenced. Additionally, DNA samples from 56 Moebius syndrome patients provided as Dutch cohort referred from Department of Human Genetics, Radboud University, Nijmegen Medical Centre, Netherlands, were also included. 2.2. Molecular analysis Genomic DNA was extracted from 2 ml of K3EDTA blood samples using DNA isolation kit (Mammalian Blood, Roche). Primers covering 50 and 30 of coding exons plus flanking introns designed for HOXB1 (NM_002144, NP_002135) gene and PCR for each region was performed in 50 mL reaction mixture, composed of 0.3 ml primer pairs, 1X reaction buffer, 2 mM MgCl2, 0.2 mM dNTP mix in 0.5 unite taq polymerase enzyme (Fermentas), in two step cycle (Tm of primers) program in thermal cycler (Bio Rad), starting with 10 min initial denaturation, followed by 30 s denaturation, 30 s annealing and 1 min/1 kb elongation periods. PCR products were purified by spin colon kit (High Pure PCR product purification kit, Roche) and sequenced (Macrogen, Seoul, Korea). Electropherograms were analyzed for sequence variants by using publicly available BLAST program (NCBI, USCS). DNA samples from 100 healthy control individuals were screened for c.620G > A alteration in HOXB1 gene by PCR and mutation specific restriction digestion.
359
2.3. Protein models Three-dimensional models for the full HOXB1 protein and mutants were created based on multiple-threading alignments by Local Meta-Threading-server (LOMET) by using in silico tool, I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER). HOXB1-HD domains for normal and mutants were created by a homology modeling tool, SWISS-MODEL (http://swissmodel.expasy.org). A HOXB1-DNAPBX1 model was downloaded from the Protein Data Bank in EuropeEBI website for the examination of P14653 (http://www.ebi.ac.uk/ pdbe-srv/view/entry/1b72/downloads.html) (Piper et al., 1999, Zhang, 2008, Roy et al., 2010). I-Tasser builds five predictions with individual confidence scores estimating the quality of the models based upon the capacity of the threading template alignments and the convergence parameters of the structure assembly simulations. Typically, the confidence score of 3D models in this system ranges between 5 and þ2, where the higher value signifies greater confidence. Among the five strongest 3D models created by the server, the highest confidence score recorded was 3.40 for arg5, 3.48 for cys5 and 3.58 for his5. This was the model selected and used for the analysis. Quality structure estimation of protein models built by the Swiss Model is stated by the Qualitative Model Energy Analysis 4 (QMEAN4) scoring system consisting of four statistical potential terms: C-beta interaction energies; all atom pair wise energy; solvation energy, and; torsion angle energy. Reliable values range between 0 and 1 (Arnold et al., 2006; Benkert et al., 2011). The created models recorded a QMEAN4 score of 0.544 for arg5, 0.5 for cys5 and 0.536 for his5.
3. Results 3.1. Clinical The patient, 23 months old female, was referred from pediatric neurology outpatient clinics due to facial paralysis with a preliminary diagnosis of Moebius syndrome. She was the only child of first-degree cousins, once removed, born at term via Caeserian section due to cephalo pelvic disproportion, with a weight of 3320 g (0.19SD). The child’s length and occipito frontal circumference (OCF) at birth were not recorded. Developmental milestones were achieved normally; the patient sat without support at seven
Fig. 1. (A) Pedigree of the family and facial appearance of the patient at the age of 23 months, 38 months and seven years of age, respectively (Note bilateral facial paralysis). (B) Partial electropherogram showing the mutation in HOXB1gene.
360
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
months of age, walked independently at 11 months, had first words at 12 months, and two-word sentences at 22 months. On physical examination at the age of 23 month, height was recorded as 81 cm (1.42 SD), weight 10.3 kg (1.04 SD) and OFC 48 cm (0.09SD). She had midface retrusion, left-sided ptosis with bilateral lagophtalmos, depressed nasal bridge, short nose with anteverted nares, tented vermilion of the upper lip, down turned corners of the mouth, high palate, and low-set ears (Fig. 1A). Left concomitant esotropia was noted, with mild inferior oblique muscle hyper function. There was no paralytic component of the
sixth cranial nerve. The patient was unable to smile or frown, and nasolabial folds were flattened. Physical and neurological examinations were otherwise unremarkable. These findings prompted for bilateral lower motor neuron palsy of the seventh cranial nerve. Auditory brainstem response test revealed normal result. Cranial MRI scans revealed minor dilatation of the temporal horns, and no significant structural anomalies or focal lesions. Karyotype was normal. At the age of seven, she was reported to have outstanding school performance and she had a two-years-old healthy brother.
Fig. 2. Genomic layout with homeodomain structure presented with GROMOS empirical force field energy for each amino acid of the HOXB1 models created for arg5, cys5 and his5. Organization of HOXB1 gene modified from its assembly at www.ensembl.org/Homo_sapiens presented at uppermost segment. Hexapeptide motif of HOXB1, found at N terminal hexapeptide, and three helical structures of its homeodomain presented with codon numbers shown at top and homeodomain peptide numbers shown below the simulated protein architecture. Peptide sequence of homeodomain with common conserved amino acids is separately colored blue for the HOXL subclass and yellow for the entire homeodomain class. The location of identified missense mutation his5 is presented as red dot on the architecture and also marked by arrow on the peptide sequence. Qmean Score 4 is shown at the upper right corner for each model. The X axis represents the amino acid sequence and the Y axis represents the energy for each amino acid in the structure. Negative energy values, shown in green, represent favorable whereas positive values, shown in red, indicate an unfavorable energy environment for a given amino acid. Hence, revert a movement of favorable energy for arg5 (w230) to unfavorable when is replaced by cysteine (w10) and more pronouncedly unfavorable when replaced by histidine (w3).
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
361
Fig. 3. Three-dimensional modeling of entire HOXB1 built for arg5, cys5 and his5, by I-Tasser. Confidence scores for each model are presented at the right lower corner. The model is visualized by PyMol, with rainbow painting starting from dark blue at ‘N terminal’ and ending red at ‘C terminal’. The N terminal hexapeptide-located codons between 179e184 or ‘24 to 16’ relative to homeodomain presented by molecular stick presentation and highlighted in red.
3.2. Molecular Sequencing of HOXB1 in total of 95 MBS/HCFP cases revealed homozygous c.620G > A, altering arginine to histidine at 207 (p.Arg207His) in one female patient. The parents of the patient and her healthy brother were heterozygous for this variation (Fig. 1B). Prediction analysis by silica algorithms, MutationTaster, Polyphen-2, and SIFT, revealed this alteration as the cause of the disease and its ascribing facial damage (Adzhubei et al., 2010; Schwarz et al., 2010; http://sift.bii.a-star.edu.sg). Furthermore, 200 alleles in control population did not carry this alteration. 3.3. Protein The empirical force field energy of arg5, cys5 and his5, respectfully, were compared by the computer program GROMOS (GROningen Molecular Simulation) via a Swiss-Model server upon web-
based integrated homology modeling study (Christen et al., 2005). The template was drawn from Drosophila melanogaster (1b72) and presented with w77% homology with the human ortholog, HOXB1. It was observed that the favorable energy environment of arginine reverted to unfavorable, when accommodated with cysteine and histidine (Fig. 2). I-Tasser modeling of each HOXB1 peptide containing one of the three, arg5, cys5 or his5, did not show critical differences between a1, a2 and a3 regions and in the secondary structure of the N-terminal hexapeptide conformation (Fig. 3). The number of predicted hydrogen bonding contacts of DNA molecule decreased from three to one when arginine was replaced by cysteine, and increased to four when replaced by histidine. Comparison of the mean values of the measurements between the oxygen of C2 of T11 with the hydrogens of the residues suggests that cys5 may have the weaker binding interaction than histidine (Fig. 4).
Fig. 4. HOXB1-PBX1-DNA ternary complex. Models of arg5 (A), cys5 (B), his5 (C) residues with proximity to thymidine 11 of DNA demonstrated by stick model for easy identification. The predicted hydrogen bond interactions between the residues and the oxygen of the second carbon of 11th thymidine of the DNA molecules are presented with closer shot of the region presented at D for arg5, E for cys5 and F for his5. The measured distances of 3.7 A to ε, 2.6 Å to h1 and 4.8 Å to h2 from the hydrogen atoms of arginine to oxygen (D); the distance of 8.3 Å from the hydrogen of sulfur of cycsteine to oxygen (E), and the distances of 7.3 Å to d1, 7.9 Å to ε1, 6.3 Å to d2 and 7.1 Å to ε2 from the hydrogen atoms of histidine to oxygen (F).
362
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363
4. Discussion The prevalence of MBS and HCFP is currently unknown, but todate 39 sporadic individuals have been referred to our outpatient clinics since 1995 (Uzumcu et al., 2009, 2009). HOXB1 was the most relevant candidate gene, not only for its orthologue’s association with facial weakness in Hoxb1/ mouse, but also for its association with a human phenotype. HOXB1 gene screening revealed mutation altering arg5 to his5 in the coding peptide in a single case. This alteration was neither identified in our screened 100 healthy control individuals nor in 100 in-house groups of AGBRG (Advanced Genomics and Bioinformatic Research Group of National Research Institute of Electronics and Cryptology of Turkey) totaling up to 400 nationally matched alleles. Three-dimensional modeling studies comparing arg5 with published cys5, and here reported his5, predicted that the a helical structures of homeodomain of HOXB1 are sustained. Hydrogen bonding between the polar molecules and electronegative atoms, such as oxygen, are considered specific mediators of a proteinnucleic acid recognition mechanism (Sathyapriya et al., 2008). Side chains of residues form most of the crucial contacts through bidentate and complex hydrogen-bonding interactions with DNA base edges, hence conferring remarkable specificity (Luscombe and Thornton, 2002). It is known that arginine at position five of the homeodomain, which is conserved in 95% of known homeodomains, interacts with the core of 50 -TAAT-30 sequence. Comparison of the distances of the predicted hydrogen bonds between the C2 oxygen atom of thymidine (T11) and the side chains of arg5 to cys5 and to his5, suggests that his5 may present stronger bonds than cys5. Our measurement revealed 8.3 Å for the distance of hydrogen bonding between T11 and sulfhydryl group of cysteine; however, the distance has been reported to be 6 Å in published data by Webb et al. (2012). Ternary complex revealed that arg5 has three but shortest, cys5 has single and longest, his5 has four but longer connections, predicting that arg5 > his5 > cys5 binding interactions. Physicochemical specification of histidine is less basic (pH 7.59) than arginine (pH 10.76), but it shares common properties, including a positively charged polar basic format. As a result, the impact of his5 on the HOXB1 structure and its function may be milder than that of cys5, which correlates with the displayed
Table 1 Clinical findings of the four previously reported HCFP patients with homozygous mutation in HOXB1, compared to present case.
Origin Molecular findings NM_002144 NP_002135 Facial findings Masked facies Bilateral facial weakness Full oculomotor motility Esotrophia/esophoria Ptosis Smooth philtrum Midface retrusion Upturned nasal tip Low-set ears Micrognathia Other findings Feeding difficulty Sensorineural hearing loss Abnormal cochlear function Cochlear structural defect n.a. Not available.
Present case
Webb et al. (2012)
Turkish
German American
c.620G > A p.Arg207His
c.619C > T p.Arg207Cys
þ þ þ þ þ left e þ þ þ e
4/4 4/4 4/4 4/4 e 2/4 4/4 4/4 3/4 2/4
e e e n.a.
2/4 4/4 2/2 1/1
phenotypes (Table 1). Our patient had neither feeding difficulties nor sensorineural hearing loss as with previously reported cases. Heterozygous mutations involving arg5 in two other HOXL subclass homeobox genes are responsible for dominant disease phenotypes. Two patients with Currarino syndrome (MIM 142994) carrying heterozygous p.Arg245Gly and p.Arg245His in the HLXB9 (MNX1) gene, and one patient with Hand-Foot-Uterus (MIM 142959) syndrome with heterozygous p.Arg326Gly in HOXA13, were previously reported (Hagan et al., 2000, Parker et al., 2011). Out of 272 families with MBS or HCFP, 177 previously reported by Webb et al. (2012), along with 95 samples (39 from Turkey and 56 from Netherlands), only three cases were found to carry mutations in HOXB1, establishing the frequency of HOXB1 mutations to be 1.1% in patients afflicted with congenital facial paralysis, with or without the additional features of MBS. 5. Conclusion Arg5 of the homeodomain of HOXB1 is suggested to be a labile mutation site for HCFP. Stronger binding interactions of his5 than cys5 leading to a stronger HOXB1-DNA-PBX1 correlating with milder phenotype may propose that in silico protein modeling studies may provide qualified predictions for the consequences of the human mutations on the structure and function of the encoding genes and contribute to our understanding of the molecular pathology underlying the disease phenotype. Acknowledgments We would like to thank all of the participating patients and their families for their support in this ongoing research. We appreciate AGBRG (IGBAM, Bilgem, TUBITAK) for sharing with us their massively parallel sequencing results of 100 private individuals. We also wish to express our gratitude to Dr. Hans van Bokhoven for his kind contribution of Dutch cohort samples. This work received partial support from Istanbul University-BAP-10801, and full support from the Scientific and Technological Research Council of Turkey (TUBITAK, grant number 108S418 [CRANIRARE] and 112S398 [CRANIRARE-2]. The CRANIRARE consortium is supported by the European Research Area Network (E-RA net programs). All co-authors have read the manuscript agreed with its content and stated that there is no conflict of interest. References Adzhubei A, Steffen S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS, Sunyaev SR. A method and server for predicting damaging missense mutations. Nat. Methods 2010;7:248e9. Arnold K, Bordoli L, Kopp J, Schwede T. The SWISS-MODEL workspace: a web-based environment for protein structure homology modeling. Bioinformatics 2006;22:195e201. Benkert P, Biasini M, Schwede T. Toward the estimation of the absolute equality of individual protein structure models. Bioinform. 2011;27:343e50. Carolyn M, Slupsky CM, Sykes DB, Gay GL, Sykes BD. The HoxB1 hexapeptide is a prefolded domain: implications for the Pbx1/Hox interaction. Protein Sci. 2001;10:1244e53. Chang CP, Shen WF, Rozenfeld S, Lawrence HJ, Largman C, Cleary ML. Pbx proteins display hexapeptide-dependent cooperative DNA binding with a subset of Hox Proteins. Genes Dev. 1995;9:663e74. Christen M, Hünenberger PH, Bakowies D, Baron R, Bürgi R, et al. The GROMOS software for biomolecular simulation: GROMOS05. J. Comput. Chem. 2005;26: 1719e51. Goddard JM, Rossel M, Manley NR, Capecchi MR. Mice with targeted disruption of Hoxb-1 fail to form the motor nucleus of the VIIth nerve. Development 1996;122:3217e28. Hagan DM, Ross AJ, Strachan T, Lynch SA, Ruiz-Perez V, et al. Mutation analysis and embryonic expression of the HLXB9 currarino syndrome gene. Am. J. Hum. Genet. 2000;66:1504e15.
Z.O. Uyguner et al. / European Journal of Medical Genetics 58 (2015) 358e363 Holland PWH, Booth HAF, Bruford EA. Classification and nomenclature of all human homeobox genes. BMC Biol. 2008;5:1e29. Luscombe MN, Thornton JM. ProteineDNA interactions: amino acid conservation and the effects of mutations on binding specificity. J. Mol. Biol. 2002;320:991e 1009. MacKinnon S, Oystreck DT, Andrews C, Chan WM, Hunter DG, Elizabeth C, Engle EC. Diagnostic distinctions and genetic analysis of patients diagnosed with moebius syndrome. Ophthalmology 2014;121:1461e8. Noyes MB, Christensen RG, Wakabayashi A, Stormo GD, Brodsky MH, Wolfe SA. Analysis of homeodomain specificities allows the family-wide prediction of preferred recognition sites. Cell 2008;133:1277e89. Parker L, Mangwani J, Wakeling E, Singh D. Hallux valgus interphalangeus and a novel mutation in HOXA13. Part of the broadening spectrum of hand-footgenital syndrome. Foot Ankle Surg. 2011;17:29e30. Piper DE, Batchelor AH, Chang CP, Cleary ML, Wolberger C. Structure of a HoxB1e Pbx1 heterodimer bound to DNA: role of the hexapeptide and a fourth homeodomain helix in complex formation. Cell 1999;96:587e97. Roy A, Kucukural A, Zhang Y. I-TASSER: a unified platform for automated protein structure and function prediction. Nat. Protoc. 2010;5:725e38. Sathyapriya R, Vijayabaskar R, Vishveshwara S. Insights into proteineDNA interactions through structure network analysis. PLoS Comput. Biol. 2008;4:1e15.
363
Schwarz JM, Rödelsperger C, Schuelke M, Seelow D. Mutation taster evaluates disease-causing potential of sequence alterations. Nat. Methods 2010;7:575e6. Studer M, Lumsden A, Ariza-McNaughton L, Bradley A, Krumlauf R. Altered segmental identity and abnormal migration of motor neurons in mice lacking Hoxb-1. Nature 1996;384(6610):630e4. Uzumcu A, Karaman B, Toksoy G, Uyguner ZO, Candan S, et al. Molecular genetic screening of MBS1 locus on chromosome 13 for micro deletions and exclusion of FGF9, GSH1, and CDX2 as causative genes in patients with moebius syndrome. Eur. J. Med. Genet. 2009a;52:315e20. Uzumcu A, Candan S, Toksoy G, Uyguner ZO, Karaman B, et al. Mutational screening of BASP1 and transcribed processed pseudogene TPJg-BASP1 in patients with moebius syndrome. J. Genet. Genomics 2009b;36:251e6. Webb BD, Shaaban S, Gaspar H, Cunha LF, Schubert CR, et al. HOXB1 founder mutation in humans recapitulates the phenotype of Hoxb1-/- mice. Am. J. Hum. Genet. 2012;91:171e9. Wilson DS, Sheng G, Jun S, Desplan C. Conservation and diversification in homeodomain-DNA interactions: a comparative genetic analysis. Proc. Acad. Sci. 1996;93:6886e91. Zhang Y. I-TASSER server for protein 3D structure prediction. BMC Bioinform. 2008;9:40e8.
