We report here a consanguineous Turkish family with three affected individuals having homozygous 8 bp truncation mutation, p.Gly559Aspfs*16. The mutation we found was previously reported in two other Turkish families, indicating founder effect. The phenotype of the affected individuals in our family is very similar to the other two families; except the uveitis in the index patient. The diagnosis of uveitis and possible ocular tuberculosis in the index patient is noteworthy. Nine months of anti-tuberculosis agents were given to the index patient and, for the first 2 months, corticosteroids were also used.
Fröjmark et al. were the first to link mutations in FZD6 gene to autosomal recessive nail dysplasia [3]. They identified two different mutations in two large consanguineous Pakistani families. Affected individuals were homozygous for the missense mutation p.Arg511Cys for one family and homozygous for the nonsense mutation p.Glu584Ter in the other family. Later, several reports of other patients from Pakistan, Iran, and Turkey were reported. Naz et al. reported two more Pakistani families where affected individuals were also homozygous for mutation p.Glu584Ter, indicating a common ancestor [4]. Raza et al. also reported another Pakistani family with a homozygous p.Gly422Asp; c.1265G > A mutation [5]. At the same year, two other families with new mutations were reported; in one family affected individuals were homozygous for missense mutation p.Arg509Ter and in the second family affected individuals were compound heterozygous for mutations p.Arg96Cys/p.Glu438Lys [34]. Moreover, in 2016, homozygosity for an 8 bp deletion, p.Gly559Aspfs*16; c.1676_1683delGAACCAGC was detected in two Turkish families [8]. In 2017, a homozygous 1 bp deletion variant, c.1859delC (p.Ser620Cysfs*75) was seen in an Iranian family [7]. To date, seven different mutations have been reported in eleven families, including two missense, two nonsense, two frameshifts, and one compound heterozygous [3,4,5,6,7,8]. Five out of seven mutations are clustered in the C-terminus, which suggests that the C-terminal region could be a mutation hotspot.
Through mutagenesis studies, it has been revealed that several residues in the intracellular loops and the C-terminus of FZDs play critical roles for signaling. Specifically, the mutation of the highly conserved internal KTxxxW motif between 498th and 503rd positions in the C-terminus, or single amino acid exchanges in the first (R340A) or the third (L524A) intracellular loops of, another protein from human FZD family, FZD5 completely abolished FZD signaling. The same mutations completely ablated the binding of the phosphoprotein DVL and its membrane recruitment by FZD [18] which is a central player in FZD-induced signal transduction and functionally necessary for all FZD signaling pathways [35, 36]. The PDZ domain of DVL directly binds the KTxxxW motif of FZD [37].
Also, in general, agonists binding to GPCRs were shown to induce changes in C-tail conformation that is necessary for activating heterotrimeric G protein [38]. Prolonged agonist stimulation catalyzes the phosphorylation of the C-tail, promoting arrestin binding, desensitization, and GPCR internalization [39]. Studies utilizing the peptides encoding the C-tail of FZD also suggest alpha-helicity in C-terminus of the Frizzleds is known as a need for efficient protein-protein interaction with DVL and other downstream signaling elements. Moreover, shortening the C-tail beyond C507 of, another protein from human FZD family, FZD5 impaired regular DVL recruitment and the ability of Wnt activator to activate Lef/ Tcf-dependent transcription [40, 41].
In terms of experimental studies, the existing knowledge about FZD6 is still limited. Fröjmark et al. expressed native and mutant (p.Glu584X and p.Arg511Cys) variants of FZD6 fused to green fluorescent protein (GFP) in HEK293T cells. While the missense mutation has no or little effect on total FZD6 levels, no expression was detected from the nonsense one [3]. The biological function of FZD6 protein was also studied by Cui et al., using a FZD6 knock-out mice, and their findings pointed out a regulatory role for FZD6-mediated Wnt signaling in the differentiation process of claw/nail formation [42]. Moreover, FZD6 was shown to be critical for the morphogenesis of hair follicles in Drosophila and mice. Fzd6−/− mice are viable and fertile; but among more than 100 FZD6 knock-out mice examined, all have abnormal macroscopic hair whorls [43]. Besides, FZD3−/− and FZD6−/− double-mutant mice die within minutes of birth and have a misoriented pattern of inner-ear sensory hair cells, this points out the role for FZD6 in planar-cell polarity. Also, FZD6 genes are expressed in all sensory hair cells and in many non-sensory epithelial cells in the inner ear [44]. Moreover, Fröjmark et al. reinvestigated the FZD6 knock-out mouse model and about 50% of male knock-out mice, but none of the female mice had absent or abnormal claws compared to wild-type mice. To link the expression of FZD6 to early nail development they also checked FZD6 expression in mouse embryos at several embryonic days and revealed that at E16.5 there was an expression of FZD6 in the epidermis of the digital tip in the region corresponding to the developing nail bed and ventral part of the digit [3]. In addition to that, Naz et al. reported a strong expression level of FZD6 in the ventral nail matrix and some FZD6 staining in the nail bed [6].
So far, ten types of NDNC have been reported in the literature, 6 of which are inherited in an autosomal dominant mode of inheritance. The genes associated with human hereditary nail disorders are listed as HPGD, RSPO4, PLCD1, COL7A1 and FZD6 [2]. HPGD gene is found to be associated with isolated congenital nail clubbing (OMIM 119900) and responsible for the metabolism of prostaglandins. Following irritation or injury, arachidonic acid (AA) is released and oxygenated by calcium-dependent enzyme systems leading to the formation of prostaglandins. Specifically prostaglandin E2 is readily detectable in equine acute inflammatory exudates. Moreover, both the influx of extracellular calcium and mobilization of intracellular calcium are very critical for the process of prostaglandin formation [45]. Another gene is RSPO4, linked to nail disorder, nonsyndromic congenital (NCDC4; OMIM 206800); a secreted protein with a known role in embryonic development and homeostatic self-renewal in adult tissues; besides its role in Wnt signaling which has both anti-inflammatory and pro-inflammatory functions. PLCD1 is linked to NDNC3 (OMIM 151600); a member of the phospholipase C family that regulates homeostasis of the immune system in skin. The lack of PLCD1 protein induces skin inflammation; since the skin of PLCD1-ko mice displays typical inflammatory phenotypes, including increased dermal cellularity, leukocyte infiltration and expression of pro-inflammatory cytokines. In addition, exogenously expressed PLCD1 attenuates LPS-induced expression of IL-1b [46]. Another gene related with nail disorders is COL7A1, which the alpha chain of type VII collagen that is associated with NDNC8 (OMIM 607523). Mutations in COL7A1 induce lifelong severe skin and mucosal blistering followed by scarring, caused by loss of adhesion between the epidermis and the dermis. COL7A1-ko mice also display blisters and erosion at sites of trauma, subepidermal blistering, and high postnatal lethality. Finally, FZD6 function as a negative regulator of the canonical Wnt/beta-catenin signaling. It was observed that FZD6 signalling activates beta-catenin in a study of patients affected by nail dysplasia. This study reported that Wnt3a signalling causes beta-catenin accumulation in healthy, but not FZD6-mutant fibroblasts, indicating a canonical role of FZD6 in this context [3]. Moreover, Kilander et al. showed via fluorescence recovery after photo-bleaching (FRAP) that recombinant WNT-1, − 2, 3A, − 4, −5A, −7A, -9B, and -10B affect FZD6 surface mobility and thus directly act on FZD6 [47]. The loss of interaction partners we proposed due to our truncation mutation could mainly be WNT family proteins. WNT pathway and innate immunity are also shown interrelated. There is an interaction among the WNT signaling network, inflammatory cytokines, and innate immune signaling pathways [48]. Individual WNT proteins were shown to have pro- or anti-inflammatory functions. WNT ligands and WNT/β-catenin signaling was found to positively regulate LPS-induced pro-inflammatory cytokines. The WNT signaling pathway plays a major role in regulating tolerance versus immunity, particularly in DCs, and more [49]. Therefore, it is not unexpected that immune-related problems are seen in NDNC patients. The common intersection point of all known NDNC genes is their association with the immune system, specifically innate immunity [50, 51]. The diagnosis of uveitis and possible ocular tuberculosis in the index patient is noteworthy as the innate immune system is a rapidly deployed, first response in host defense and its dysregulation leads to autoinflammation. Ocular follow-up of our patient will probably help differentiate between an auto-inflammatory granulomatous process and ocular tuberculosis.
Due to the lack of crystal structure of FZD6, computer-based analyzes of FZD6 are also very limited. The first attempt made by Mohammadi-asl, et al. as predicting the formation of multiple helical secondary structures in the distal cytoplasmic region of the p.Ser620Cysfs*75 mutant protein which does not exist in the native protein via I- TASSER [52]. Moreover, they used NtePhos 3.1 server and revealed pathogenic consequence of the mutation by disturbing the cytoplasmic domain structure and signaling through loss of phosphorylation residues [7]. They concluded that the nonsense mutation causes the elimination of the distal end of the topological domain (amino acids 495–706). This domain mediates Wnt/beta-catenin signaling by relocalization and phosphorylation of disheveled proteins. Even though, KTXXXW is present, loss of phosphorylation residues and formation of unusual helical secondary structures can result in lack of proper response to WNT-3A and WNT-5A activation consistent with previous studies [3, 4, 7].
For the same reason, we performed the homology modeling of native and mutant forms of FZD6 protein with I-TASSER. To gain more insight about the impacts of mutation on the structure of the protein, we performed 20 ns MD simulations and concluded that FZD6 mutant displayed higher RMSD pattern compared to native. This result suggests us that the introduction of stop codon to C-terminus, associated with the translation of new 15 amino acids upon the frame-shift, results in increased tendency for unfolding than native structure with higher backbone motion at 310 K. This result is also supported with RMSF pattern that Leu253-Cys282, Ala329-Phe380 and His549-Ser571 regions (mutant numbering) are more flexible in FZD6 mutant compared to native one. Specifically, His549-Ser571 region (mutant numbering) displays almost ~ 8-fold more flexibility compared to native. Hence, the loss of C-terminus, even being in partial, would disrupt the intramolecular interactions and we could end up with unstable protein. It is also crucial to notice that the conservation of KTxxxW motif in C-terminus; previously suggested as essential for FZD signaling in, another protein from human FZD family, FZD5 protein, is not enough for the FZD6 mutant. This fact suggests that the problem in our case can be the loss of structural integrity in addition to the loss of signaling region.
Along 20 ns MD trajectory, we also consider the impacts of mutations in terms of intramolecular interactions such as salt-bridge formations. Upon this particular mutation, almost 30% of salt-bridge interactions are lost. Even explaining protein stability is a complex issue, there is a well-known fact in literature that the intra-molecular interactions, such as salt bridge formations, are crucial elements for the stability of proteins and their positions on 3D structure of protein contribute to its stability in different manners, e.g. the salt-bridge formations on protein surface contribute to protein stability less than 1 kcal/mole [53] while those buried and positioned on hydrophobic core contribute more than 4 kcal/mole [54]. The loss of these seven salt bridge formations, established with β–sheet structures, considered as a part of the seven transmembrane-spanning receptor (Fig. 1), would adversely affect the protein stability and result in its non-functionality. Except for Glu697-Lys552 interactions, these particular salt bridge interactions are strong enough to contribute to the stability of the protein in a positive manner. As a well-known fact, entropy is a crucial element of thermodynamics of macromolecules to create a favorable environment for protein or substrate binding, happened in the signaling pathway. Upon the alterations in entropy of protein with negative manner caused by the loss of these salt bridge interactions, the expected interaction(s) of protein would be either lacked or disrupted in FZD6 mutant and this non-functionality happens.