Trichothiodystrophy (TTD) is a rare autosomal recessive disorder characterized by brittle hair and also associated with various systemic symptoms. Approximately half of TTD patients exhibit photosensitivity, resulting from the defect in the nucleotide excision repair. Photosensitive TTD is due to mutations in three genes encoding XPB, XPD and p8/TTDA subunits of the DNA repair/transcription factor TFIIH. Mutations in these subunits disturb either the catalytic and/or the regulatory activity of the two XPB, XPD helicase/ATPases and consequently are defective in both, DNA repair and transcription. Moreover, mutations in any of these three TFIIH subunits also disturb the overall architecture of the TFIIH complex and its ability to transactivate certain nuclear receptor-responsive genes, explaining in part, some of the TTD phenotypes.
DNA can undergo various modifications, including strand break, base damage, helix distortion and strand cross-link from endogenous (reactive oxidative species) and exogenous sources (UV radiation). To maintain genome integrity and to avoid harmful effects of DNA damage, cells possess several DNA repair pathways that can be differentiated biochemically and genetically. If not repaired, the damaged DNA can lead to a range of human disorders that exhibit developmental defects, neurological abnormalities, photosensitivity, cancer and accelerate the aging process (1).
Nucleotide excision repair (NER) is one of the most important DNA repair system and is responsible for removing several kinds of DNA lesions, particularly those induced by anti-tumorigenic drugs and UV irradiation, such as cyclobutane pyrimidine dimmers (CPD) and pyrimidine (6-4) pyrimidone photoproduct, which distort the DNA helix. Defects in NER results in rare autosomal recessive diseases, known as xeroderma pigmentosum (XP), Cockayne syndrome (CS) and trichothiodystrophy (TTD) (2). Mutations in eleven genes have been associated with these diseases (3,4). The products of these genes participate in NER, and three of these eleven genes (XPB, XPD and p8/TTDA) are part of the basal transcription/repair factor TFIIH. Intensive work has been performed to understand why mutations in TFIIH and particularly in the XPD gene cause XP, CS and TTD (5). Here, we will consider the molecular functions of TFIIH and examine how a defect in one of them, may lead to the pathology of TTD.
TTD is a rare autosomal recessive disorder characterized by sulfur-deficient brittle hair and neuroectodermal symptoms (6). Clinical features of TTD are highly variable in expression, including Photosensitivity, Icthyosis, Brittle hair and nails, Intellectual impairment, Decreased fertility and Short stature (7,8). The acronyms, PIBIDS, IBIDS and BIDS represent the initials of these words. Light microscopy test of hair shaft reveals trichoschisis; there is a cleaving, breaking, irregular and flattening hair surface like a trichorrhexis nodosa. Polarizing microscopy shows the typical appearance of alternating light and dark bands, giving a ‘tiger tail’ pattern. There is no correlation between the extent of hair abnormalities and the severity of the rest of the clinical phenotype. However, amino acid analysis that quantifies sulfur (specifically cysteine), inversely correlates with the percentage of hairs showing one or more abnormalities and remains the definitive diagnostic test for TTD (7). In addition, the increased proportion of unstable disulfide conformers contributes to the reduced hair robustness (9). The sulfur-deficient brittle hair phenotype is not seen in either XP or CS patients (Table 1).
Clinical manifestations of more than 100 TTD patients from 20 countries all over the world, reported in the last 40 years, were statistically reviewed (10). The incidence for TTD was estimated at one per million in Western Europe. Only one TTD patient was found in Japan (11, personal communication). However, in Japan, the incidence for XP is 1 per 20 000–100 000 (12,13). The number of deaths at a young age of TTD patients is ∼20-fold higher compared with the US general population (10).
Cellular studies revealed that the photosensitive form of TTD is caused by the defective NER, also observed in XP and CS; however, non-photosensitive TTD display a normal NER capacity (14). Despite the fact that photosensitive TTD patients have a defect in the same gene as some XP patients characterized by photo-induced skin cancers, the TTD patients do not have an increased incidence of skin cancers (15) (Table 1). Neuroimaging examination revealed that the most common feature of TTD was hypomyelination, also found in CS patients. However, contrary to what occurs in CS, neurological abnormalities in TTD patients seemed to be caused by developmental defect (dysmyelination) rather than loss of myelin (demyelination). Progressive neurological degeneration was not reported in TTD patients (16).
Genetic complementation experiments (17) revealed that photosensitive TTD patients only result from mutations in XPD, XPB and p8/TTDA, three of the ten subunits of TFIIH. Moreover, in cells derived from TTD patients, the cellular concentration of TFIIH is significantly reduced (18,19). However, neither the extent of the DNA repair defect nor the degree of reduction in the level of TFIIH correlate with the severity of the clinical phenotype. A reduced (and even more significant reduction) level of TFIIH was also found in cells from some XPD patients.
Most of photosensitive TTD cases are mutated in the XPD gene, and the mutations are mostly located either at the R112 loop or at the C-terminal end of the protein (20–22) (Fig. 1). It should be mentioned that in a certain number of cases, different phenotypes were described for the same mutation, but the severity of clinical phenotype apparently correlates with the nature of the mutation located on the second allele, thus suggesting that both alleles might contribute to the TTD phenotype (23, unpublished data).
Epidemiological studies revealed that XPD, p.H201Y, p.D312N and p.K751Q polymorphisms, might correlate with either lung, colorectal, bladder or breast cancer but not with TTD-like symptoms. In vitro studies further demonstrated that recombinant TFIIH containing XPD polymorphism have normal DNA repair and basal transcription in vitro (24). This result indicates that for XPD, the single polymorphisms seem to be benign. It is also likely that combination of several polymorphisms, including that of XPD, might contribute to the ‘cancer’ phenotype.
Three mutations were found in XPB (XP mutation database: http://xpmutations.org); only the T119P mutation results in TTD. The tenth subunit of TFIIH, p8/TTDA, was recently identified as a responsible gene for the third group of photosensitive TTD-A (25). Patient cells with either p8/TTDA L21P, R56stop or T to C transition at start codon mutations are deficient in the p8/TTDA protein.
C7orf11 (TTDN1) was identified as the first disease gene for the remaining non-photosensitive form of TTD (26). Mutations were found in only six of the 44 unrelated non-photosensitive TTD patients (27). The TTDN1 patients are not defective in NER and the steady state level of TFIIH is normal. The severity of the clinical features does not correlate with the mutation map in TTDN1, suggesting that non-photosensitive TTD might be multi-factorial disease. Other factors besides TTDN1 mutations might influence the phenotype of this disorder. Indeed, it was shown that TTDN1 interacts with polo-like kinase1 through the phosphorylation by Cdk1 and plays a role in regulating mitosis and cytokinesis (28).
TFIIH IN NER
NER is a multi-step process, which proceeds via at least two alternative pathways. One is transcription-coupled DNA repair (TCR), which removes lesions only in the actively transcribed DNA strand, and the other is global genome repair (GGR), which removes lesions in any sequence of the genome (29). In eukaryotic cells, the process requires more than 30 proteins that sequentially target the damaged DNA: recognition of DNA damage, opening of the DNA around the lesions, single-strand incisions and excision of the lesion-containing DNA fragment, and gap filling DNA synthesis and ligation (30,31). Human TFIIH consists of a core of seven subunits: XPB, XPD, p62, p52, p44, p34 and p8/TTDA. This core forms a ring-shaped structure, that is associated to a module of cdk-activating kinase (CAK) composed of Cdk7, cyclin H and MAT1 (32). Once recruited to the site of DNA damage, either by XPC/HR23B (in GGR) or by the stalled RNA polymerase II (in TCR), TFIIH opens the DNA around the lesion and promotes the subsequent incision and excision, in both TCR and GGR, by recruiting the XPG and XPF endonucleases (33,34). To make the DNA ‘bubble’, the TFIIH core module drives its two helicase-proteins with ATP hydrolysis, whereas CAK is released from the core upon arrival of XPA and the other NER factors (35). Although the CAK kinase activity is critical in the basal transcription through the phosphorylation of C-terminal domain of the largest subunit of RNA polymerase II and many nuclear receptors (36,37), it is dispensable for the NER (Fig. 2A).
It has been predicted that DNA opening around the lesion could be driven through the helicase activity of XPB (38). A recent study has shown that the XPB helicase activity is not crucial in NER (39). It seems, however, that its ATPase activity rather than its helicase activity, in combination with the helicase activity of XPD, is needed to remove the DNA lesions. In addition to its seven helicase motifs, one of them being the ATP binding site, XPB possess two additional ‘ATPase’ motifs: the well-conserved R-E-D residue loop motif (at amino acids 472–474) and the positively charged thumb (ThM) region (at amino acids 514–537) (40) (Fig. 1). These motifs are specifically involved in the regulation of the DNA-dependant ATPase activity of XPB and help to stabilize the binding of TFIIH to the damaged DNA (41), to allow the recruitment of other NER factors. The XPB/T119P mutation might disturb the TFIIH architecture, even though located out of any helicase/ATPase motifs (Fig. 2B).
Another subunit of TFIIH, p52, interacts with XPB and stimulates its ATPase activity (42). Mutation in the Drosophila homolog of p52, Dmp52, destabilizes its interaction with XPB and results in a fly with UV sensitivity, cancer prone, brittle-bristle with cuticular-deformation and developmental defects. Some of these phenotypes displayed by this mutant fly were also observed in human TTD and XP (43) (Fig. 2C). Although it was found that mutation in p52 may abolish the physical and functional interaction between XPB and other factors in transcription or DNA repair, human with mutations in the p52 gene have never been identified.
In addition to XPB, TFIIH contains another DNA helicase subunit, XPD, which is dispensable for transcription initiation but plays a critical role in opening the DNA around lesions during NER (44–46). Recent structural studies of XPD revealed a four-domain structure consisting of two canonical helicase domains (HD1: helicase motif I, Ia, II and III, and HD2: helicase motif IV, V and VI), the iron-sulfur cluster binding (4FeS) and Arch domains (47–49) (Fig. 1). However, there is no relationship between the structural placement of the XPD mutations and the TTD phenotype. It is likely that the TTD mutations dispersed throughout all four domains might disturb the network of interactions between the various TFIIH subunits affecting the integrity of the complex (Fig. 2D).
Despite it is generally believed that the elevated incidence of cancer in patients with XP is a consequence of defective NER, one wonder why patients with photosensitive TTD are not cancer prone? Some answers might come from immunocytochemistry studies (50). It was observed that after UV irradiation, NER factors assemble at the damage site rapidly and are redistributed all over the nucleus, after completing the repair of most of the damaged DNA. XPC, the damage-recognition protein, localizes to damage sites rapidly in both XP and TTD cells (as well as in wild-type cells). However, in XP cells, XPC remains at damage site for more than 24 h, whereas in TTD cells XPC is redistributed 3 h after UV irradiation. Moreover, XP cells show a delay in recruiting TFIIH and this complex remains on the damage site over 24 h post-UV irradiation, whereas in TTD cells, TFIIH is hardly recruited to the DNA damage, (especially to CPD damage site) (51,52). These results partially support the hypothesis that although XP and TTD cells are both deficient in NER, the absence of NER complexes in TTD, at the damage site, may allow the trans-lesion synthesis or post-replication repair. In contrast, in XP cells, the persistence of NER factors at unrepaired sites may prevent any replication and/or transcription processes, resulting to genomic instability and an increase of cancer development.
As observed in the partnership XPB/p52, XPD interacts with p44, another subunit of TFIIH, and this association upregulates the helicase activity of XPD, but not its ATPase activity (46). Most of the XPD mutations are located in its C-terminal part and thus weaken its interaction with the N-terminal of p44, explaining the NER defect in these patients. Interestingly, p44 (as well as the p34 subunit) possesses a RING finger domain, typical of E3 ubiquitin ligase enzymes (53). Ssl1p, the budding yeast homolog of p44, exhibits an E3 ubiquitin ligase activity likely implicated in transcription (54). Although the role of p44 in regulating the helicase activity of XPD, and thus NER, is well established, no patients with mutations in the p44 gene have ever been identified (55).
Mutations in p8/TTDA, the tenth and smallest subunit of the TFIIH complex, only causes TTD (26,56). p8/TTDA stimulates the ATPase activity of XPB in vitro through a direct interaction with p52, which upregulates the activity of XPB (57). The intricate network of interactions between these three subunits was also demonstrated by genetic complementation experiments. Indeed, overexpression of p8/TTDA is able to re-establish normal levels of TFIIH not only in the TTD-A cells but also in TTD cells bearing mutations in XPD and in Drosophila cells bearing a mutation in Dmp52, the homolog of the human p52. These observations underline the role of p8 as a stabilizer of TFIIH (58). Two distinct kinetic pools of p8/TTDA exist: one is the form bound to TFIIH and alternatively p8/TTDA exist as a free module that shuttles between the cytoplasm and nucleus (59). Solution structure study revealed that this latter form might exist as a homodimer (60). After induction of DNA damage, the equilibrium between these two pools dramatically shifts towards a more stable form within the TFIIH complex. Structural studies revealed that the p8/TTDA L21P mutation possibly disrupt its conformation. Furthermore, the R56stop mutation weakens the interaction between p8/TTD-A and p52, and as a consequence, the stability of TFIIH (61) (Fig. 2E).
Despite all the work done to dissect the functions of each one of the TFIIH subunits, there is still a lot of work to be done (Table 2). Similarly, there are still many potential ways of modifying the function of this complex. Post-translational modifications of proteins are central to most aspects of cellular life and to investigate the modifications that TFIIH can undergo and how these modifications can modulate the activity of the complex could potentially aid us in the understanding of the molecular basis of the TTD phenotype. The ubiquitin proteasome degradation pathway seems to be implicated in the regulation of DNA repair (62). This is not so surprising given the fact that XPB was shown to interact with SUG1, a subunit of 26S proteasome (63). Moreover, ubiquitination of MAT1 regulates CAK activity (64). Other post-translational modification, such as the phosphorylation of XPB, might regulate NER, by preventing the further incision step by the XPF-ERCC1 endonuclease (65). Despite of many studies, unclear function and modification of every subunits of TFIIH still exist (Table 2).
|p8||Core stability/ATPase regulation|
|p8||Core stability/ATPase regulation|
Defects in DNA repair cause not only TTD but also many disorders, such as XP, CS, Bloom syndrome, Werner syndrome, Nijmegen breakage syndrome and Ataxia-telangiectasia among others. Moreover, defective DNA repair has been linked to neurological disorders (66). It must be noted that NER activity is strongly attenuated in terminally differentiated neurons (67). The fact that we observe neurological phenotypes (neurodegeneration) in a classical model of defective DNA repair could suggest that the accumulation of DNA damage causes neuronal death. However, UV penetrates deep into the dermis of the skin but never into the central nervous system (68). Moreover, it seems that the TTD patients suffer much more from developmental defect, e.g. dysmyelination, rather than from neurodegeneration (69), suggesting that the TTD phenotype arises from transcriptional deficiency.
Several studies show that NER factors are associated with transcription of protein coding genes. TFIIH bearing mutations in XPD, as found within TTD patients, but not XP patients, exhibits a deficiency of basal transcription in vitro (45). However, gene-expression profiling on microarrays of TTD and XP fibroblasts showed no significant differences (70). As the symptoms of TTD appear to be mainly caused by developmental defects, it is possible that significant differences in transcription occur in differentiating cells and specific tissues. A recent study in TTD mouse model highlights the importance of the co-activator function of TFIIH in the pathogenesis of TTD (71). TTD/Xpd (R722W) mutant mice, carrying a mutation found in a human TTD patient, showed the similar phenotype to those of TTD individuals, such as motor impairments associated with microcephaly and hypomyelination. It was suggested that such defects were due to a deregulation of thyroid hormone target genes in the central nervous system. Indeed, TFIIH containing the R722W mutation failed to properly stabilize the thyroid hormone receptor (TR) to its binding site, resulting in a defect in the TR transactivation (72). Other mutations within XPD weakens the anchoring of the CAK to the core of TFIIH and consequently its capacity to phosphorylate PPARs, RARa and/or VDR partners (37,72–75), which is required for efficient expression of the hormone-responsive genes.
It is well accepted that the onset of photosensitivity associates with the defects of NER. However, the rest of the pathophysiological symptoms of TTD patients are rather complicated. Currently, except for the TFIIH mutations and their link to deficient NER, and TTDN1 gene responsible for non-photosensitive TTD that regulates the mitosis and cytokinesis, no other mechanisms are linked to the TTD phenotype. Further studies of the cells from both photosensitive and non-photosensitive TTD may provide a clue to pathophysiology of TTD and possibly unveil a new function of the versatile TFIIH complex.
This study was supported by funds from an ERC Advanced-Scientist Grant (to J.M.E.), Agence Nationale de la Recherche (ANR-No 06-BLAN-0141-01 and ANR-No 08-MIEN-02203) and La Ligue contre la Cancer.
We are grateful to Renier Vélez-Cruz for critical reading of the manuscript. S.H. is a recipient of INSERM MD young investigator fellowship.
Conflict of Interest statement. None declared.