Nasal embryonic LHRH factor (NELF) mutations in patients with normosmic hypogonadotropic hypogonadism and Kallmann syndrome  Ning Xu, Ph.D., Hyung-Goo Kim, Ph.D., Balasubramanian Bhagavath, M.D., Sung-Gyu Cho, M.S., Jae Ho Lee, Ph.D., Kyungsoo Ha, M.S., Irene Meliciani, M.S., Wolfgang Wenzel, Ph.D., Robert H. Podolsky, Ph.D., Lynn P. Chorich, M.S., Kathryn A. Stackhouse, B.S., Anna M.H. Grove, B.S., Lawrence N. Odom, M.D., Metin Ozata, M.D., David P. Bick, M.D., Richard J. Sherins, M.D., Soo-Hyun Kim, Ph.D., Richard S. Cameron, Ph.D., Lawrence C. Layman, M.D.  Fertility and Sterility  Volume 95, Issue 5, Pages 1613-1620.e7 (April 2011) DOI: 10.1016/j.fertnstert.2011.01.010 Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Figure 1 Location and conservation of NELF, KAL1, and TACR3 mutations. (A) Exon and intron structure of the 58-kb NELF gene (NM_015537) with locations of human missense and intronic mutations identified in sporadic IHH and KS patients. Important exons shown as blue rectangles to scale are numbered along with their size in base pairs (bp). The locations of start and stop codons are to scale, but the sizes of introns are not to scale. (B, C) Graphic views of KAL1-encoded anosmin-1 (NP_000207) and TACR3 (NP_001050) proteins with known domains. (B) Shown is the single amino acid cysteine in-frame KAL1 deletion in the whey-acidic-protein domain (WAP)—although the annotation (c.487_489delTGT) was incorrect in our previous report (17). (C) Also shown is the TACR3 truncation mutation p.Trp275X within a cytoplasmic domain between the fifth and sixth transmembrane domains. The relative size between anosmin-1 and TACR3 proteins and the size and locations of depicted domains are to scale. S = signal peptide; Cys-rich = cysteine-rich, FN-III = fibronectin III; H = histidine; T1–T7 = transmembrane 1–7. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Figure 2 Functional protein analysis of human NELF mutations. Western blot analysis is shown using the anti-NELF antibody (27), which recognizes the ∼63-kd NELF protein in lymphoblastoid cell lines extracts. β-Actin was used as internal loading control (∼42 kd). Lane 1 represents a normal male control subject, and lanes 2 and 3 represent individuals with NELF mutations. Additional control subjects and other mutations did not alter protein expression and are not shown. The Western blots were performed three times, and the results were consistently the same. Protein expression for patient C68 was not altered (data not shown). Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Figure 3 NELF splice mutant. Subcloning and sequencing analysis of reverse-transcription polymerase chain reaction products of NELF exons 8–12 from patient and control lymphoblastoid RNA confirms that the human NELF c.1160–13C>T mutation causes aberrant exon 10 skipping. Splicing patterns are compared between a normal control and the mutant by cloned cDNA sequencing (the wild-type or mutant band is a cloned fragment). An expected NELF product of 280 bp including exons 9 and 10 is observed in the control (lane 2), whereas an abnormal product of 246 bp skipping exon 10 (34 bp) is observed in a KS patient with c.1160–13C>T (lane 3). Exon 11 nucleotides and out-of-frame novel amino acid sequence are depicted in blue. A 123-bp DNA marker is shown in the first lane. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

(A) Ala253 in NELF (NP_056352) is evolutionarily fully conserved in seven available orthologues. Thus the change of this hydrophobic residue to a hydrophilic Thr is likely to be deleterious. This has also been shown by sorting intolerant from tolerant (SIFT), which compares the amino acid change to known orthologues and calculates a probability based on homology of whether it will be tolerated (likely benign) or not (likely a mutation). (B) The previously reported Thr478 NELF (NP_056352) missense mutation (see text) shows only partial conservation in four orthologues among seven available species. (C) Cys163 in the whey-acidic-protein (WAP) domain in anosmin-1 (NP_000207) is also evolutionarily fully conserved in all 12 available orthologues, suggesting that its deletion in patient C7 would be detrimental. All eight cysteine residues in the WAP domain in these 12 orthologues are invariant, and here only five cysteine residues are shown as invariant. (D) 4-Disulfide bonds of the WAP domain of KAL1-encoded anosmin-1 protein are shown with blue lines connecting two Cys molecules. The deleted Cys163 marked by an asterisk is forming a disulphide bridge with Cys134. Positions of residues affected by NELF missense mutation and Cys163 KAL1 deletion in KS patient C7 are marked by asterisks and red letters in all available orthologues. Amino acid abbreviations in green represent the invariant residues that match the consensus exactly, and abbreviations in blue show partial matching. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Quality assessment of the degree of homology in the 70–amino acid N-terminal region of NELF. Red regions indicate regions of high homology, and yellow and green regions indicate low homology. Dashes indicate insertions in the NELF sequence not present in the template. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Model for the extended N-terminal region of NELF model from amino acid 209 to 528. Although the homology in the 70–amino acid N-terminal region demonstrated in Supplemental Fig. 2 permits the construction of a detailed model, the resolution of the C-terminal region of the fragment is likely to be lower as a consequence of lower homology. However, the overall model is sufficient to demonstrate that the C-terminal region of the structure has no intramolecular contacts with the N-terminal region where the mutation Ala253Thr (highlighted in magenta) occurs, so that the effects and exposure of the mutation can be investigated on the basis of the model for the 70–amino acid model shown. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions

Protein modeling of p. Ala253Thr is shown Protein modeling of p.Ala253Thr is shown. Using multiple sequence alignment (Expresso), a protein model for the N-terminus (Supplemental Methods) was constructed based on homology to an acetylated Rsc4 tandem bromodomain histone chimera (PDB code 2R10). This site is identified by SPPIDER as a potential binding site of the protein. SPPIDER is a method enabling the recognition of protein-protein interaction sites, which is important in understanding functionally relevant amino acids and protein function. This method integrates enhanced relative solvent accessibility predictions with high-resolution structural data to discriminate between interacting and noninteracting sites on the surface of proteins. (Left) Model of N-terminal region of NELF spanning amino acids 209–331 obtained by alignment to 2R10. Blue regions in the model correspond to high-confidence regions based on the quality alignment to the template, and red regions correspond to low-confidence regions (Supplemental Methods). The site of the mutation, which is located at the outside of the protein within a high-confidence region, is indicated in a magenta sphere. (Right) Analysis using Interprosurf (http://curie.utmb.edu) also indicates an extended interaction region near p.Ala253. The observed mutation (shown in magenta) occurs in the central region of an extended protein-protein interaction region highlighted in green on the model of the 209–528–amino acid N-terminus of NELF. Interprosurf is a web server that can predict interacting residues on the surface of a monomeric protein, taking into account the 3-dimensional structures of protein subunits. This prediction method is based on the solvent-accessible surface area of residues in isolated protein subunits and includes methods to measure the relative importance of different amino acid residues in the protein-protein interface. Fertility and Sterility 2011 95, 1613-1620.e7DOI: (10.1016/j.fertnstert.2011.01.010) Copyright © 2011 American Society for Reproductive Medicine Terms and Conditions