A recurrent deletion in the ubiquitously expressed NEMO (IKK-γ) gene accounts for the vast majority of incontinentia pigmenti mutations

Human Molecular Genetics, Sep 2001

Incontinentia pigmenti (IP) is an X-linked dominant disorder characterized by abnormal skin pigmentation, retinal detachment, anodontia, alopecia, nail dystrophy and central nervous system defects. This disorder segregates as a male lethal disorder and causes skewed X-inactivation in female patients. IP is caused by mutations in a gene called NEMO, which encodes a regulatory component of the IκB kinase complex required to activate the NF-κB pathway. Here we report the identification of 277 mutations in 357 unrelated IP patients. An identical genomic deletion within NEMO accounted for 90% of the identified mutations. The remaining mutations were small duplications, substitutions and deletions. Nearly all NEMO mutations caused frameshift and premature protein truncation, which are predicted to eliminate NEMO function and cause cell lethality. Examination of families transmitting the recurrent deletion revealed that the rearrangement occurred in the paternal germline in most cases, indicating that it arises predominantly by intrachromosomal misalignment during meiosis. Expression analysis of human and mouse NEMO/Nemo showed that the gene becomes active early during embryogenesis and is expressed ubiquitously. These data confirm the involvement of NEMO in IP and will help elucidate the mechanism underlying the manifestation of this disorder and the in vivo function of NEMO. Based on these and other recent findings, we propose a model to explain the pathogenesis of this complex disorder.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://hmg.oxfordjournals.org/content/10/19/2171.full.pdf

A recurrent deletion in the ubiquitously expressed NEMO (IKK-γ) gene accounts for the vast majority of incontinentia pigmenti mutations

Swaroop Aradhya 2 Hayley Woffendin 6 Tracy Jakins 6 Tiziana Bardaro 4 5 Teresa Esposito 5 Asmae Smahi 3 Christine Shaw 2 Moise Levy 1 Arnold Munnich 3 Michele D'Urso 5 Richard A. Lewis 0 2 Sue Kenwrick 6 David L. Nelson 2 0 Department of Ophthalmology and the Cullen Eye Institute, Baylor College of Medicine , Houston, TX 77030, USA 1 Department of Dermatology 2 Department of Molecular and Human Genetics 3 Department of Genetics, Unit des Recherches sur les Handicaps Gntiques de l'Enfant INSERM-393 , Hopital Necker-Enfants Malades, 75015 Paris, France 4 BioGem, Naples, Italy 5 International Institute of Genetics and Biophysics , Area di Ricerca del CNR di Napoli, Naples, Italy 6 Wellcome Trust Centre for Molecular Mechanisms of Disease and University of Cambridge Department of Medicine, Addenbrooke's Hospital , Hills Road, Cambridge CB2 2XY, UK - Incontinentia pigmenti (IP) is an X-linked dominant disorder characterized by abnormal skin pigmentation, retinal detachment, anodontia, alopecia, nail dystrophy and central nervous system defects. This disorder segregates as a male lethal disorder and causes skewed X-inactivation in female patients. IP is caused by mutations in a gene called NEMO, which encodes a regulatory component of the IB kinase complex required to activate the NF-B pathway. Here we report the identification of 277 mutations in 357 unrelated IP patients. An identical genomic deletion within NEMO accounted for 90% of the identified mutations. The remaining mutations were small duplications, substitutions and deletions. Nearly all NEMO mutations caused frameshift and premature protein truncation, which are predicted to eliminate NEMO function and cause cell lethality. Examination of families transmitting the recurrent deletion revealed that the rearrangement occurred in the paternal germline in most cases, indicating that it arises predominantly by intrachromosomal misalignment during meiosis. Expression analysis of human and mouse NEMO/Nemo showed that the gene becomes active early during embryogenesis and is expressed ubiquitously. These data confirm the involvement of NEMO in IP and will help elucidate the mechanism underlying the manifestation of this disorder and the in vivo function of NEMO. Based on these and other recent findings, we propose a model to explain the pathogenesis of this complex disorder. Familial incontinentia pigmenti (IP; BlochSulzberger syndrome; MIM 308300) is a rare genodermatosis (14) that occurs in approximately 1 of 50 000 newborns. The most conspicuous sign of IP is a progressive skin pigmentation abnormality which begins with vesicular lesions containing apoptotic cells and infiltration of eosinophils. These lesions eventually heal and lead to hyperpigmentation due to incontinence of melanin from the superficial epidermis into the dermis. The clearance of melanin by macrophages finally leaves patients with linear or reticular hypopigmented patches along lines of X-inactivation. Although the skin phenotype can be quite dramatic, the most significant medical problems in IP are blindness due to retinal detachment and central nervous system defects, which cause mental retardation or seizures (1,3). A few minor signs include hair loss, conical or absent teeth and nail dystrophy. IP demonstrates complete penetrance, but its phenotypic expression is highly variable, even among related patients with the same mutation. Affected IP male conceptuses typically fail to survive past the second trimester and thus, this disorder has earned recognition as a classic male-lethal condition. Concordant with this observation, female individuals with IP mutations survive because of dizygosity for the X chromosome and selection against cells expressing the mutant X chromosome. Thus, female IP patients exhibit skewed X-inactivation (5,6), a feature that is often used to confirm diagnosis. We recently demonstrated that cells in IP patients lack NF-B function due to mutations of an upstream activator called NEMO (NF-B essential modulator) (7). In an initial screening of 50 patients, most had an identical deletion (hereafter termed NEMO 410) within the NEMO gene that eliminated exons 410 and consequently abolished protein function. The *To whom correspondence should be addressed. Tel: +1 713 798 4787; Fax: +1 713 798 5386; Email: The authors are members of the International IP Consortium The authors wish it to be known that, in their opinion, Swaroop Aradhya, Hayley Woffendin, Tiziana Bardaro and Asmae Smahi should be regarded as joint First Authors deletion alters sequence after nucleotide 399 (from ATG) in the NEMO mRNA and leads to a truncated protein containing the first 133 N-terminal amino acids. This recurrent rearrangement occurs between two identical, 878 bp MER67B repeats, the first of which is located in intron 3 and the second 4 kb telomeric to the gene. NEMO is a 23 kb gene composed of 10 exons (GenBank accession no. AJ271718) (7). The 48 kDa NEMO protein has two coiled-coil motifs and a leucine zipper which are required for dimerization and proteinprotein interactions, and a zinc finger at the C-terminus that appears to be necessary for posttranslational stability (810). It has also been shown that the C-terminus of NEMO is indispensable for function. NEMO is the regulatory component of IB kinase (IKK), a central activator of the NF-B transcriptional signaling pathway (11,12). In response to various cytokines, IKK phosphorylates the inhibitory IB molecules, which sequester NF-B in the cytoplasm. The removal of IB allows NF-B to translocate into the nucleus and activate transcription of various genes. Through this mechanism, NF-B regulates immune and inflammatory responses, and prevents apoptosis in response to TNF-. Therefore, NEMO mutations eliminate NF-B activity and cause potentially widespread disruption of downstream cellular responses, although the exact downstream effects are only now being elucidated. With respect to IP, loss-of-function mutations in NEMO create a susceptibility to cellular apoptosis in response to TNF- (7). This phenomenon explains the male lethality and skewing of X-inactivation in female patients. Our goal was to determine whether NEMO mutations could explain all cases of IP and to assess the mutation spectrum and the genotypephenotype correlations in IP patients. The analysis was performed in a cohort of 357 unrelated IP patients. We also examined the expression pattern of NEMO in an effort to relate its presence to the pathogenesis of IP. Based on this work and other recent findings, we propose a model to explain the basis for the complicated IP phenotype. A recurrent mutation in NEMO Analysis of the coding sections of NEMO (exons 210) in 357 unrelated patients revealed 277 mutations (Fig. 1 and Tables 1 and 2). Mutations in 50 of these patients have been reported previously (7). Among the 277 patients with mutations, 248 (90%) demonstrated the recurrent IP deletion (NEMO 410), which was detectable by Southern blotting of HindIII digests probed with exon 2, intron 3, or the whole NEMO cDNA (Fig. 2). The deletion was also detected in some patients with a diagnostic PCR assay described previously (7). Among the 277 mutations, 109 appeared de novo in isolated cases with no family history of IP, and the recurrent NEMO 410 deletion itself accounted for 101 of them (Table 1). The deletion was not detected in 50 unrelated normal women (100 X chromosomes) and when it was present in a family, it was only detected in affected members (data not shown). Although the Families with identified mutations Families without detectable mutations Recurrent deletion (NEMO 410) cases Small nucleotide mutations Recurrent (NEMO 410) deletion cases Small nucleotide mutations Total patients without detectable mutations X-inactivation investigated cases IP deletion cases surveyed for parental origin 27 248 (89.5% of mutations); 69.5% of patients tested recurrent deletion accounted for 90% of mutations, it was only found in 70% of the patients tested. Among 27 families with the recurrent IP deletion and in whom parental origin could be tested by linkage analysis, X-inactivation and/or Southern blot analysis, examination of parental DNA revealed that 19 (70%) of mutations originated from the paternal germline and eight (30%) from the maternal germline (Table 1). We considered the possibility that in cases of apparently maternal origin, in which a mother was diagnosed as clinically normal, the mother could actually have been misdiagnosed and have a deletion inherited from her father (i.e. the probands grandfather). Analysis of 30 such cases revealed the NEMO 410 deletion in only two clinically normal mothers. The origin of the deletion in these women could not be determined because their parents were unavailable. Small mutations in NEMO Analysis by CSGE or single strand conformation polymorphism (SSCP) analysis revealed that 29 out of 357 patients had smaller mutations, including microdeletions, substitutions and duplications (Fig. 1 and Table 2). Seven of these mutations have been described previously (7,13). Most small mutations C214 G220 T516 551588 G570 639645 C896 11631175 C1161 were predicted to cause frameshift and premature truncation of the resulting NEMO protein. Of the 29 mutations, 11 were nucleotide substitution mutations and only two of them changed the amino acid identity (M407V and E57K). Another two substitutions resulted in the elimination of splice donors at the 3 ends of exons 3 and 4, which almost certainly caused aberrant splicing. However, this could not be tested since mutant mRNAs were unavailable because of skewed X-inactivation in female individuals carrying these splice-site mutations. The fifth substitution mutation, described previously (7), changed the stop codon so that the open reading frame was extended by 27 novel amino acids. The remaining six substitution mutations changed a coding triplet codon to a nonsense codon. Only eight of 109 de novo cases were small mutations (Table 1). The NEMO protein has two coiled-coil motifs, a leucine zipper and a zinc finger (9,10). Of the 29 small mutations, four were located in a region bound by IKK- (14,15), six disrupted the human T-cell leukemia Tax protein binding region (16) and 12 were within the terminal region where cytokines bind (10) (Fig. 1). The exon 3 splice mutation was not within a known domain but is likely to have altered the NEMO protein sequence after the IKK- binding site. Lastly, with the exception of some exon 10 mutations, all of the remaining mutations were associated with a typical IP phenotype. The genotype phenotype correlations with respect to exon 10 mutations have been described previously (13). Polymorphisms in NEMO Four single nucleotide polymorphisms (SNPs) were discovered in exons 1a and 1c and introns 3 and 8 of NEMO (Table 3). All of these were found in unaffected members of Caucasian IP pedigrees. All SNPs were in either untranslated or intronic regions, emphasizing the importance of an undisrupted NEMO gene sequence. Two potential restriction fragment length polymorphisms were also discovered with a NEMO intron 3 probe. Two IP families (IP64 and XL206) demonstrated a 26 kb aberrant band on an EcoRI digest. This band was present in affected and unaffected members of the XL206 family, who carried the recurrent IP deletion (data not shown). Thus, the 26 kb fragment may represent a polymorphism and is likely to be quite rare, since it was not observed in more than 100 IP families and was found in only one of 48 normal female individuals. Three separate IP families (XL203, XL233 and XL306) exhibited a 3.1 kb aberrant band on an EcoRI blot, which segregated with disease in each family (data not shown). However, family XL203 had a P370fsX23 mutation in exon 9 (Table 2) and families XL233 and XL306 carried the recurrent IP deletion. Therefore, the 3.1 kb fragment could also represent a polymorphism. The mechanisms that give rise to these novel fragments were recently elucidated (17). X-inactivation status and screening of upstream exons CSGE, SSCP and Southern blot analyses did not detect mutations in 80 patients, of whom 38 were from families with multiple affected members and 42 were sporadic IP cases (Table 1). All of these patients were diagnosed with IP based on clinical data and, in familial cases, on linkage analysis. Skin pigmentation abnormalities were present in affected members, and some patients had other IP-associated signs. X-inactivation analysis, using the (CAG)n polymorphism at the HUMARA locus, NEMO intron 1a (77 bases distal to exon 1a) NEMO exon 1 (20 bases into exon; 5-UTR) NEMO intron 3 (34 bases distal to exon 3) NEMO intron 8 (6 bases proximal to exon 9) 3.1 kb EcoRI fragment (n = 120 X chromosomes) 26 kb EcoRI fragment (n = 216 X chromosomes) showed that 15 of a subset of 29 mutation-negative (CSGEnegative) patients showed completely skewed X-inactivation. Of the remaining CSGE-negative patients, six (all sporadic) were uninformative for the HUMARA (CAG)n polymorphic marker, and eight showed random X-inactivation. This suggested that many patients without detectable mutations are likely to be authentic IP cases (Table 1). The same subset of 29 patients was also examined for base alterations in the three non-coding exons upstream of exon 2. However, no base alterations were observed except for one SNP in exon 1c (Table 3). NEMO expression analysis Analysis of human tissue RNAs by northern blot showed ubiquitous expression of NEMO (Fig. 3). By visual comparison, the highest expression levels were noted in tissues of the central nervous system, liver and muscle. In addition, high levels were found in the placenta, bone marrow, lymph nodes, spinal cord and thyroid. Moderate expression was observed in the immune system, heart, kidney and prostate gland. The digestive system, lungs, blood, bladder and uterus exhibited low levels of NEMO transcript. Notably, a faint 3 kb band was also present in most tissues, possibly indicative of an alternative splice form. Only the small intestine, colon and bladder did not exhibit this band, while it was most evident in peripheral blood leukocytes, where it was of equal intensity compared with the 2 kb NEMO signal (Fig. 3). We have not investigated the significance of this potential splice form yet. To further evaluate the expression of NEMO, we examined mouse tissues and surveyed the public sequence databases. Tissues from four embryonic stages (days 7, 12, 14 and 18), newborn pups, and nine different tissues from 1-month-old mice (heart, brain, bladder, testicles, lung, liver, spleen, kidney and stomach) were examined. With equal amounts of RNA from the various embryonic and postnatal tissues, RTPCR indicated that Nemo expression began as early as embryonic day 7 and continued through the first postnatal month (data not shown). Relative quantities, by visual estimates, showed comparable levels of Nemo expression in all tissues. A survey of NCBIs LocusLink (ID# 8517) and Unigene (ID# 43505) database entries for NEMO corroborated the mouse RTPCR and human northern blot data in terms of presence in specific tissues. The databases also indicated expression in skin, foreskin, germ cells, pancreas, breast, and head and neck; we were not able to test these tissues. A BLAST search with the NEMO cDNA also identified ESTs from various tumor sources: T cells (T-cell leukemia), B cells (Burkitt lymphoma), neuroblastoma, melanoma, oligodendoglioma and human ovary tumors. We recently showed that IP is caused by mutations in NEMO and consequent disruption of NF-B activity. To corroborate these findings with a more extensive survey, we analyzed 357 unrelated IP patients of both familial and sporadic origin and found mutations in 277 (78%) patients. This high percentage strongly indicates that mutations in NEMO are the primary cause of IP. Of the 277 mutations, 90% were accounted for by an identical genomic deletion. However, the frequency of the common deletion is likely to be between 70 and 80%, since several patients have not yet demonstrated mutations. Moreover, many of these patients showed skewed X-inactivation, suggesting that they are likely to have NEMO mutations that functionally resemble those reported here, and that lead to the same conditionsevere reduction of NF-B activity. The remaining 10% of mutations among our patients included small nucleotide changes, including deletions, substitutions and duplications. Twenty of the 25 small mutations have not been reported previously. There is no specific clustering of mutations within specific regions of NEMO. Instead, the mutations are evenly scattered, indicating that the entire NEMO protein sequence is important and cannot be disrupted at any location without causing the same cellular lethality. Even the polymorphisms we detected were located in non-coding sections of the gene, in either introns or untranslated regions. Examples of coding sequence polymorphisms were not found among our patients (700 X chromosomes). However, some exon 10 mutations eliminate the zinc finger at the C-terminus of NEMO and cause mild, non-lethal phenotypes (13,1820). The E57K and M407V mutations are the only missense mutations in our cohort so far and these positions are conserved in the mouse as well. Notably, the E57K mutation near the N-terminus is associated with a slightly milder IP phenotype (M.DUrso, unpublished data), although there is no evidence yet that it is compatible with male survival. Typically, mutations in the C-terminal zinc-finger region are mild, allowing male individuals to survive (13,1820). However, at least one missense mutation close to the N-terminus has been reported, but it differs from E57K in that it leads to an atypical phenotype involving immune dysfunction (19). In contrast, the M407V mutation corresponds to a typical male-lethal phenotype; this is intriguing because it is located in the C-terminus, a region frequently associated with mild, variant phenotypes. The C-terminus is known to be indispensable for NEMO function, and there are some indications that the zinc finger is necessary for post-translational stability (9,10) (G.Courtois, personal communication). Thus, the M407V mutation might alter a very important region of the zinc finger and will thus be useful in elucidating the function of this domain. The discovery of a recurrent deletion in NEMO adds IP to a growing list of conditions termed genomic disorders which arise from large-scale genomic rearrangements that disrupt functional sequences (21,22). The high frequency of the NEMO 410 deletion has simplified diagnostic testing for IP. Interestingly, a section of NEMO has been duplicated along with the LAGE2 gene in a 35 kb sequence, and the NEMO 410 deletion happens to be one of four genomic rearrangements observed at this duplication locus (17). The NEMO 410 deletion mutation arises by rearrangement between two identical MER67B repeat sequences (7). Similarly, a recurrent inversion between identical repeats associated with the X-linked F8C gene accounts for most mutations that cause hemophilia A (23). This inversion occurs predominantly in the male germline, where the single X chromosome remains unpaired during meiosis (24). In contrast, our data show that the NEMO 410 deletion occurs at a relatively lower frequency (75%) in the paternal germline. Thus, while the hemophilia A inversion occurs predominantly by intrachromosomal exchange during male meiosis, the NEMO 410 deletion is most likely to originate by intrachromosomal misalignment in male individuals and, less frequently, from inter- or intrachromosomal exchange in female individuals. However, linkage and SNP haplotype analyses have indicated that the IP region in distal Xq28 exhibits a reduced rate of recombination (2528), possibly supporting recent evidence for extensive linkage disequilibrium in distal Xq28 (29). Therefore, if suppressed recombination is characteristic of the NEMO region, the recurrent NEMO 410 deletion may arise exclusively by intrachromosomal misalignment between the MER67B repeats, even in the maternal germline. The recurrent NEMO 410 deletion results in a truncated NEMO molecule containing the first 133 N-terminal amino acids plus several novel residues (7). Similarly, most small mutations prematurely disrupt the NEMO protein. Thus, genotypephenotype correlations cannot be made with NEMO mutations because the functional effects of the mutations are the samethey abolish NEMO activity and, consequently, NF-B activation. Although the truncated NEMO protein with an intact N-terminus may bind IKK-, as shown in vitro (10), the in vivo consequences are unclear. However, it is known that proteins lacking their C-terminus cannot activate NF-B (10), and this is the case with most NEMO mutations. Both the common NEMO 410 deletion and most small mutations cause a typical IP phenotype, including male lethality and skewed X-inactivation. We had expected that some mutations within NEMO would explain the phenotypic variation seen among female patients, but this variation now appears to be modulated by selection against the X chromosome, typical of several X-linked disorders, rather than by the functional consequences of NEMO mutations. However, specific exon 10 mutations in NEMO are an exception because they lead to variations in phenotype that include male survival and an ectodermal dysplasia (ED)-like manifestation (13,1820). These exon 10 mutations, unlike most NEMO mutations, are mild and do not abolish NF-B activity or cause skewed X-inactivation (13). Moreover, the phenotypic consequences of defective NF-B function are more exposed in patients with hypomorphic mutations than in patients with typical IP, in whom cell-lethal mutations cause prenatal male demise and skewed X-inactivation in female individuals. Together, our analyses identified mutations in 78% of the patients in our study. Although most of our patients exhibited a typical IP phenotype (except male patients with hypomorphic mutations) and were ascertained based on clinical presentation, X-inactivation, and/or linkage data, 22% of patients have not exhibited mutations. Since a significant number of these patients have shown skewed X-inactivation, they are likely represent authentic IP patients. Therefore, methods with greater sensitivity may be required to find their mutations. Alternatively, mutations may be possible in the promoter, introns, or other regions of NEMO that were not investigated. The NEMO promoter region was recently defined and will be studied in IP patients shortly (30). Complete gene deletions have not been excluded either, but Southern blot analysis of DNA from female patients who are negative for mutations by CSGE or SSCP has not revealed unusual fragments. Moreover, some of these individuals are heterozygous for SNPs within NEMO (Table 3) and thus cannot have a full gene deletion. The ubiquitous and early expression of NEMO is concordant with its vital requirement during embryonic and postnatal development. The absence of Nemo in hemizygous male mice causes lethality around day 12 (3133). Despite the universal expression of NEMO, the IP phenotype is restricted to specific tissues, particularly those of ectodermal origin. This could be attributed to the tissue- and time-specific interactions between NEMO and various upstream signaling molecules that execute certain developmental programs through NEMO and NF-B. However, these interactions and programs have yet to be defined. Despite the complexity of the IP phenotype, it has become clear that two factors contribute to the evolution of the IP phenotypeX-inactivation status in female individuals and the functions of NF-B (Fig. 4). Regardless of the type of mutation causing IP, X-inactivation is likely to modulate the severity of the disorder in female individuals and accounts for some of the variation found among them. As shown in this report, some female patients carry the common deletion but are clinically normal. In these individuals, selection against mutant cells may have commenced very early during their prenatal development. This has also been observed in mouse models of IP, in which the surviving Nemo+/ female mice show striking skewing of X-inactivation (31,32). Although X-inactivation might account for the phenotypic variation among female IP patients, a role for modifier genes cannot be excluded. In human male patients, in whom X-inactivation is not an issue, most NEMO mutations are lethal because they abolish NF-B activity, making cells susceptible to TNF--induced apoptosis (7). This has been demonstrated in Nemo-null male mice as well (31,32). The second causative factor in the manifestation of IP is NF-B function itselfto mediate the effects of signaling molecules impinging on NEMO. NF-B regulates immune reactions in response to cytokines (12,34,35). The growth and differentiation of keratinocytes also depends on NF-B (36 38). In addition, the recent discoveries of hypomorphic NEMO mutations in male IP patients suggest that IP is, in essence, a combination of several disordersprimary lymphedema (PL), familial expansile osteolysis (FEO), and EDsince the genes defective in these disorders function upstream of NEMO and require NF-B to implement their effects (Fig. 4A and C) (7,1820,3941). The proteins defective in these three disorders, namely VEGFR3, RANK and EDA/EDAR, are involved in osteoclastogenesis (42), vascular modeling (43,44) and hair follicle induction (45), respectively. Aside from those three disorders mentioned above, hyper-IgM syndrome is also related to IP because certain hypomorphic mutations in NEMO cause this immune dysfunction disorder (20). However, Xinactivation modifies the physical manifestation of the medical problems associated with FEO, PL, ED and hyper-IgM syndrome, thereby creating a distinct phenotype in heterozygous female IP patients. Several approaches will be used in the near future to address the pathogenesis of IP. Allelism between IP and other similar diseases, such as retinopathy of prematurity, should be investigated to gain insight into the IP phenotype. In addition, unexplained cases of skin pigmentation, hair loss, abnormal dentition or skeletal malformations could also be allelic conditions. Furthermore, investigating the regulation of the NEMO promoter by upstream factors will help address how IP manifests during development. Lastly, analyzing the consequences of NF-B malfunction on target gene expression will further illuminate the pathogenic mechanism underlying this complex disorder. MATERIALS AND METHODS All patients were enrolled with the approval of and according to guidelines established by internal institutional review boards for human subject research at each collaborating institution. Peripheral blood samples were obtained in ACD- or EDTAcontaining tubes, and genomic DNA was extracted by conventional salt precipitation protocols. PCR amplification was performed with 100 ng of template genomic DNA from patients for each exon with previously reported primers (7), and CSGE and SSCP were performed as described previously (4649). The primers were designed to amplify fragments of optimal size for CSGE and SSCP analyses, between 250 and 350 bp. If a band shift was noted in a familial IP case, all family members were tested to confirm that the band shift segregated with the disease and did not represent a polymorphism. When band shifts were found, longrange PCR was used to detect the mutation in the functional copy of NEMO, and not in a second similar X-linked locus, NEMO. The long-range PCR was performed with the EXPAND PCR kit (Roche) with primer SA15F (5-CTTGGCACATCACTTATCAG-3) and a respective reverse primer, as reported previously (7). The annealing temperature for all long-range PCR primer sets was 60C. To determine the base alteration, the long-range PCR products were sequenced using big-dye fluorescence technology either in-house or at SeqWright Inc. Southern blot analyses Five micrograms of patient genomic DNA was digested with HindIII overnight and electrophoresed on a 0.8% agarose gel for 20 h at 90 V. Following overnight transfer, the filters were prehybridized in a NaHP04 buffer for 4 h and hybridized overnight with a unique NEMO exon 2 probe that detects the functional NEMO copy and not another similar X-linked locus, NEMO. The unique 624 bp probe was amplified with primers SA22F (5-GCCGGTACCCTCCACTTCTCCTG-3) and SA11R (5-GAAACGCAGAGGTGCAGGTG-3). This probe detects the normal 12 kb and the mutant 8 kb HindIII bands. The blot was washed to a stringency of 2 SSC/0.1% SDS at 65C and autoradiographed for 12 days. X-inactivation analysis To determine which X chromosome was inactivated, we used an assay reported previously (6,13,50). Briefly, 250 ng of genomic DNA from patients was digested overnight with the methylation-sensitive restriction enzyme HpaII, or with a control enzyme, RsaI. Fifty nanograms of the digest was used in a PCR to amplify the (CAG)n polymorphic repeat at the HUMARA locus. The HpaII restriction site is within the amplicon and, if digested, prevents PCR amplification. Thus, only the methylated inactive X chromosome is detected, and the polymorphism determines which X chromosome is kept active or inactive. Northern blot and RTPCR analyses Commercially available multiple tissue northern blots (Clontech) were hybridized with an exon 3 probe and amplified with primers NEMO-3F and NEMO-3R (7) to detect the 2 kb NEMO transcript. The blots were hybridized according to the manufacturers suggestions. The blots were washed to a stringency of 0.1 SSC/0.1% SDS at 65C and autoradiographed for 3 days. For RTPCR analyses, total RNA was prepared according to the manufacturers suggestions using the Trizolbased kit (Gibco) from several tissues extracted into liquid nitrogen. Equal amounts of RNA were used to prepare cDNA with a commercial kit (Gibco). Subsequently, RTPCR was performed at 60C using two sets of mouse Nemo primers NEMO-2RTF (GGTGCAGCCCAGTGGTGGC, in exon 2) with NEMO-3RTR (CAGTCTCTCCACCAGCTTCCGG, in exon 3) to produce a 283 bp product, and NEMO-9RTF (ATGAGGAAGCGGCATGTAGAG, in exon 9) with NEMO10R (AGGGATGCTCCCTAAGAGCC, in exon 10) to produce a 451 bp fragment. Actin was detected as control with primers Actin-F (GTATGGAATCCTGTGGC) and Actin-R (CACCAGACAACACTGTGTTG), which produced a 128 bp RTPCR product. ACKNOWLEDGEMENTS The authors are members of the International IP Consortium, established with the help of the International IP Foundation (IIPF), New York. We would like to thank Susanne Emmerich, director of IIPF, for assistance in arranging collaborative meetings. We are grateful for the continuing participation and cooperation of the many families reported here and to their attending physicians. Kerry L. Wright was helpful in editing the manuscript. Technicians of the Baylor College of Medicine MRRC tissue culture core provided expert technical assistance with cell lines for all patients in the IP project. Alison Getz provided technical assistance with CSGE and sequencing, and Zhe Fang extracted mouse tissues for expression analysis. This work was supported in part by NIH grants 5 R01 HD35617 and 2 P30 HD24064 to D.L.N., MRC and Action Research grants in the UK to S.K., Telethon-Italy grant E0927 to M.D. and The Foundation Fighting Blindness and IIPF funding to R.A.L. R.A.L. is a Senior Scientific Investigator of Research to Prevent Blindness, New York. REFERENCES


This is a preview of a remote PDF: http://hmg.oxfordjournals.org/content/10/19/2171.full.pdf

Swaroop Aradhya, Hayley Woffendin, Tracy Jakins, Tiziana Bardaro, Teresa Esposito, Asmae Smahi, Christine Shaw, Moise Levy, Arnold Munnich, Michele D’Urso, Richard A. Lewis, Sue Kenwrick, David L. Nelson. A recurrent deletion in the ubiquitously expressed NEMO (IKK-γ) gene accounts for the vast majority of incontinentia pigmenti mutations, Human Molecular Genetics, 2001, 2171-2179, DOI: 10.1093/hmg/10.19.2171