ORIGINAL ARTICLE

Endocrine Care

A Recurrent Signal Mutation in the Releasing Hormone Receptor with Defective Translocation to the Cell Surface and Isolated Growth Hormone Deficiency

Michela Godi, Simona Mellone, Antonella Petri, Teresa Arrigo, Claudio Bardelli, Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 Lucia Corrado, Simonetta Bellone, Flavia Prodam, Patricia Momigliano-Richiardi, Gianni Bona, and Mara Giordano

Laboratory of Human Genetics, Department of Medical Sciences and Interdisciplinary Research Center of Autoimmune Diseases (M.Go., S.M., L.C., P.M.-R., M.Gi.), and Unit of Pediatrics, Department of Medical Sciences (A.P., S.B., F.P., G.B.), University of Eastern Piedmont, 28100 Novara, Italy; Department of Pediatrics (T.A.), University of Messina, 98100 Messina, Italy; and Laboratory of Pharmacology (C.B.), Department of Medical Sciences, University of Eastern Piedmont, 28100 Novara, Italy

Context: Mutations in the GHRH receptor (GHRHR) have been detected in the familial type-IB isolated GH deficiency (IGHD-IB) inherited as an autosomal recessive disorder and characterized by a low but detectable serum GH level and good response to substitutive GH therapy.

Objective: The aim of our study was the identification of mutations in sporadic patients with a IGHD-IB phenotype.

Subjects and Methods: The GHRHR gene was systematically screened by DHPLC in 134 IGHD pa- tients with no family history of the disorder or declared parental consanguinity.

Results: We identified a novel variation, Val10Gly, within the signal peptide at the heterozygous state in three patients and in one of 1084 controls (P ϭ 0.004), suggesting that it might contribute to IGHD. The functional analysis showed that the signal peptide is not cleaved from the mutant GHRHR, which in turn is not translocated to the cellular surface, demonstrating that 10Gly dras- tically affects the receptor correct processing. Because 10Gly was also present in normal-stature relatives of the patients as well as in a control, it is likely that it exerts its effects in the context of other genetic and environmental susceptibility factors.

Conclusion: At difference from previous papers reporting GHRHR mutations in familial cases with a clear recessive mode of inheritance, our study was conducted on a large sample of sporadic patients and allowed to discover a novel mechanism of the disease caused by a recurrent dominant mutation in the GHRHR signal peptide associated with incomplete penetrance. (J Clin Endocrinol Metab 94: 3939–3947, 2009)

he GHRH receptor (GHRHR) is a seven-transmem- tracellular cAMP and consequently the opening of a volt- T brane G -coupled receptor of 423 amino acids age-gated calcium channel. The influx of calcium deter- (1). After binding GHRH, it activates a stimulatory G mines release of GH stored in secretory granules. The protein by catalyzing GTP binding to its ␣-subunit. This increase of cAMP also stimulates protein kinase A to ac- stimulates the adenylate cyclase-mediated increase of in- tivate cAMP response element binding protein transcrip-

ISSN Print 0021-972X ISSN Online 1945-7197 Abbreviations: CHO, Chinese hamster ovary; ConA, concanavalin A; DHPLC, denaturing Printed in U.S.A. HPLC; ER, ; GHRHR, GHRH receptor; HA, hemagglutinin; IGHD, Copyright © 2009 by The Endocrine Society isolated GH deficiency; PMSF, phenylmethylsulfonyl fluoride; SDS, SD score; SRP, signal doi: 10.1210/jc.2009-0833 Received April 20, 2009. Accepted July 10, 2009. recognition protein. First Published Online July 21, 2009

J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 jcem.endojournals.org 3939 3940 Godi et al. A Novel Recurrent Mutation in the GHRHR J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 tion. cAMP response element binding protein stimulates moment of diagnosis, 84% of the patients were prepubertal, the transcription of the pituitary transcription factor PIT-1, mean age was 8.4 Ϯ 4.1 SD, and the mean bone age was Ϫ1.5 Ϯ which in turns activates transcription of many pituitary 1.0 SD. None of the patients had consanguineous parents. Del- eterious mutations in the GH1 gene were excluded in all the genes, including GHRHR and GH1 (2, 3). patients. GHRHR mutations were detected in about 10% of the Italian individuals of normal height (university and hospital autosomal recessive forms of isolated GH deficiency (IGHD) staff), not tested for GH levels, were used as controls. characterized by low but measurable serum GH levels and A written informed consent was obtained from the patients’ good response to exogenous GH therapy (IGHD-type IB): parents and from the normal-stature controls. seven missense mutations (4–8), two nonsense (9, 10), one impairing a PIT-1 binding site in the promoter (11), two Analysis of the GHRHR gene in the patients by microdeletions (5, 12), and five splicing mutations (10, 13– DHPLC Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 15). The missense mutations (His137Leu, Leu144His, GHRHR was amplified by PCR from genomic DNA (Sup- Ala176Val, Ala222Glu, Phe242Cys, and Lys329Glu) do not plemental Methods and Supplementary Table S1, published as alter the receptor surface expression but interfere with ligand supplemental data on The Endocrine Society’s Journals On- line web site at http://jcem.endojournals.org). Search for mu- binding, thus affecting GHRH signaling (16). Nonsense and tations was performed by denaturing HPLC (DHPLC) scanning splicing mutations are expected to lead to the synthesis of a on an HPLC instrument (Wave; Transgenomic, Glasgow, UK). severely truncated receptor or to the absence of the protein. To allow heteroduplex formation, PCR products were subjected With the exception of Leu144His, which is recurrent and has to 3 min 95 C denaturation followed by a gradual reannealing been found in different continents (6), each mutation is spe- from 95 C to 41 C in 30 min. The temperature required for cific for a particular geographical region and is likely to de- successful resolution of heteroduplex molecules was determined using the DHPLC Melt Program (http://insertion.stanford.edu/ rive from a common ancestor. Most of the reported patients melt.html) (Supplemental Table S1). All of the samples were run belonged to consanguineous pedigrees and were homozy- at the predicted temperatures (RTm) and at RTm ϩ 2 C as rec- gous. In a minority of the families, the affected members ommended (20). All detected heteroduplex fragments were sub- carried two distinct GHRHR mutations (4, 5). sequently sequenced using the Big-dye terminator reaction kit The aim of our study was the identification of GHRHR and an ABI Prism 3100 genetic analyzer (Applied Biosystems, causal mutations in a large cohort of unrelated patients Foster City, CA). (n ϭ 134) with no family history of IGHD, with no evi- dence of parental consanguinity, and with the IGHD-IB Genotyping phenotype. SNPs were genotyped by SnaPshot (Applied Biosystems) We identified a novel missense substitution in the signal following the manufacturer’s instructions. Eight fragments peptide at the heterozygous state in three patients with a containing nine SNPs were amplified in a single PCR using the QIAGEN Multiplex PCR Kit (QIAGEN Inc., Valencia, CA) drastic effect on GHRHR processing and translocation to with the same primers used for DHPLC analysis (see Supple- the cell surface. mental Methods and Supplemental Table S2).

Plasmid constructs and transfection Subjects and Methods A wild-type GHRHR cDNA cloned in pcDNA3 was kindly donated by Professor Kelly Mayo. Nucleotides encoding the Subjects FLAG epitope tag (DYKDDDDK) were inserted within the A total of 134 Italian patients (92 males and 42 females) with GHRHR N-terminal domain at residue 37 downstream of the sig- short stature consequent to IGHD were recruited from different nal peptide. The hemagglutinin (HA) epitope tag (YPYDVPDYA) centers. Their mean height SD score (SDS) was Ϫ2.4 Ϯ 0.9, and was inserted at the C terminal in a separate construct. Both plas- height velocity over 1 yr was less than Ϫ1.5 SDS. The patients mids were used as templates into which the T position ϩ29 was were considered sporadic because none of them had relatives substituted with a G (corresponding to the change 10ValϾGly) with a demonstrated GH deficit, although some had short-stat- using the Quik Change Site-Directed Mutagenesis Kit (Strat- ured parents. The mean target corrected height of the patients agene, La Jolla, CA). was Ϫ1.6 Ϯ 1.3. IGHD was defined as a peak GH response below Wild-type and mutant constructs were transiently expressed 10 ng/ml after two consecutive classical provocative tests (with in Chinese hamster ovary (CHO) cells not expressing endoge- arginine or clonidine or ) (17) or below 20 ng/ml after one nous GHRHR. double stimulus with GHRH ϩ arginine (18) and normal serum Cells were grown to 70–80% confluency in six-well plates in level of other pituitary hormones. The patients had a mean se- Ham’s F12 medium with 10% heat-inactivated fetal bovine se- cretion peak of 4.2 Ϯ 2.2 ng/ml after the single stimuli (n ϭ 113) rum, 100 U/ml penicillin, 100 ␮g/ml streptomycin, and 2 mM or 9.2 Ϯ 5.7 ng/ml after the double provocative test (n ϭ 21). L-glutamine (Sigma Aldrich, St. Louis, MO). Transfections were Patients with a known cause of acquired hypopituitarism were performed using 2 ␮g of each construct with Fugene 6 (Roche excluded. Skeletal maturation was estimated as bone age (radius, Diagnostics, Indianapolis, IN) following the manufacturer’s ulna, and short bone) according to Tanner et al. (19). At the instructions. J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 jcem.endojournals.org 3941

GHRHR mRNA detection Statistical analysis Total RNA was extracted 24 h after transfection using the Significance was evaluated from 2 ϫ 2 contingency tables by RNeasy kit (QIAGEN). cDNA was obtained by RT-PCR using ␹2 test (Yates correction) or, when required by the small number the RETROscript Kit (Ambion, Austin, TX), and a fragment of of expected cases, by two-tailed Fisher’s exact test. P values were 364 bp spanning exons 1–4 was amplified with primers 5Ј-GCT- not corrected for multiple comparisons. TACTGAGGCTGGTGGAG-3Ј (forward) and 5Ј-TAAGGTG- GAAAGGGCTCAGA-3Ј (reverse).

Results Western blotting Western blotting was performed with membrane extracts. Identification of a GHRHR variation associated to Twenty-four hours after transfection, cells were washed with IGHD Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 PBS supplemented with 1 mM phenylmethylsulfonyl fluoride A total of 134 sporadic IGHD Italian patients were (PMSF), scraped off the wells, and centrifuged at 14,000 ϫ g for 20 sec at 4 C. Cell pellets were resuspended in TE buffer (10 mM analyzed for the presence of causal mutations in the Tris, 1 mM EDTA), supplemented with 1 mM PMSF and protease GHRHR gene. The 809-bp sequence upstream the first inhibitors cocktail, and lysated by three freeze/thaw cycles. Ly- transcribed nucleotide, including the two Pit-1 binding sates were centrifuged at 14,000 ϫ g for 30 min at 4 C. Pellets sites, the 13 exons, and the intron-exon junctions, were were resuspended in Tris-sucrose buffer [10 mM Tris-HCl (pH amplified and analyzed by DHPLC, which detects the 8),2mM PMSF, protease inhibitors cocktail, 10% sucrose). About 10 ␮g of membrane extracts was separated on 10% presence of variations at the heterozygous state. Because SDS-PAGE, transferred to nitrocellulose filters (GE Health- the mutation Lys144His in exon 5 is recurrent, this am- care, Wauwatosa, WI), probed with a commercial anti-HA plicon was also mixed with a homozygous wild-type ref- antibody (Millipore, Billerica, MA), and diluted 1:1000 in erence sample before DHPLC analysis, under the hypoth- TBS-T 5% milk overnight at 4 C. were visualized esis that some patients could be homozygous, even in the using the enhanced chemiluminescence Western blotting de- absence of parental consanguinity. Although it is ex- tection reagents, according to instructions of the manufac- turer (Applied Biosystems). Chemiluminescence signals were tremely unlikely, we cannot exclude the presence of mu- analyzed under nonsaturating conditions with an image den- tations at the homozygous state in other exons. sitometer (Bio-Rad, Hercules, CA). Direct sequencing of the heteroduplex samples con- firmed the variations detected by DHPLC. Ј Immunofluorescence of epitope-tagged receptors Three variations (Table 1) were in the GHRHR 5 CHO cells were cultured on 12 mm2 round coverslips and at flanking region. They were all present in the public SNP 70% confluency were transiently transfected with plasmids con- database (http://www.ncbi.nlm.nih.gov/), and none fell in taining the GHRHR cDNA. a putative regulatory box. Forty-eight hours after transfection, cells were washed with One synonymous and five missense substitutions were PBS and fixed with 4% paraformaldehyde for 15 min. Immu- detected in the coding sequence (Table 1). The two mis- nofluorescence experiments were performed using both perme- sense in the signal peptide of the immature protein abilized (20 min with 0.2% Triton X-100 in PBS) and nonper- Ͼ meabilized cells. (Arg4Trp, Val10Gly) and the 489C T variation in exon Permeabilized cells were incubated overnight at 4 C with the 6 were not present in the SNP database. anti-FLAG M2 monoclonal antibody (1:200; Sigma Aldrich) or Sixteen variations were detected in the intronic se- the anti-HA polyclonal antibody (1:100, Santa Cruz Biotech- quences (Supplemental Table S3). Because none of them nology, Santa Cruz, CA) in PBS containing 0.1% Triton X-100 fell within a consensus splicing sequence, they were not and 4% fetal bovine serum. A secondary antibody conjugated with Cy3 (against the anti-FLAG; Jackson Immuno Research further considered. Laboratories, West Grove, PA), or with a fluorescein isothio- To determine the allele and genotype frequency of the cyanate (against the anti-HA; Sigma Aldrich) were added1hat promoter and exonic variations, all of the patients and 171 room temperature. For the colocalization experiments, a con- unrelated normal-stature matched controls were geno- jugated concanavalin A (ConA) antibody (Invitrogen, Chats- typed by a multiplex fluorescent primer extension assay worth, CA) was used as endoplasmic reticulum (ER) marker (SnaPshot; Applied Biosystems). None of the variations and added simultaneously to the above-described secondary antibodies. showed a statistically significant difference between pa- Nonpermeabilized cells were incubated under identical con- tients and controls (Table 1). ditions using primary and secondary antibodies diluted in PBS Because Arg4Trp and Val10Gly represented novel without Triton X-100. nonsynonymous changes, they were genotyped in a Confocal microscopy was performed using a 63X objective of larger number of normal-stature controls. An addi- the Leica confocal microscope TCS SP2 (Leica Microsystems AG, Wetzlar, Germany). Images were scanned using the same tional 170 individuals were analyzed for Arg4Trp (to- contrast settings for equivalent times and were representative of tal ϭ 341) and 913 others for Val10Gly (total ϭ 1084). at least three independent experiments. Arg4Trp was detected with comparable frequency in 3942 Godi et al. A Novel Recurrent Mutation in the GHRHR J Clin Endocrinol Metab, October 2009, 94(10):3939–3947

TABLE 1. Genotype frequencies of the GHRHR variations

Nucleotide Database changea entryb Position changea Patients Controls P 5Ј-723CϾT rs17159772 5Јflanking CC 62 (46.3) 83 (48.5) CT 58 (43.3) 69 (40.4) TT 14 (10.4) 19 (11.1) 5Ј-213CϾT rs2302019 5Јflanking CC 41 (30.6) 43 (25.1) CT 56 (41.8) 89 (52.0) TT 37 (27.6) 39 (22.9) 5Ј-187CϾT rs2302020 5Јflanking CC 65 (48.5) 85 (49.7) CT 54 (40.3) 67 (39.2) TT 15 (11.2) 19 (11.1) Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 10CϾT Exon 1 Arg4Trp CC 133 (99.3) 335 (98.3) CT 1 (0.7) 6 (1.7) TT 0 0 29TϾG Exon 1 Val10Gly TT 131 (97.8) 1083 (99.9) TG 3 (2.2) 1 (0.1) 0.004 GG 0 0 169GϾA rs4988496 Exon 3 Ala57Thr GG 120 (89.6) 150 (87.7) GA 14 (10.4) 19 (11.1) AA 0 2 (1.2) 363GϾT rs4988498 Exon 4 Glu121Asp GG 120 (89.6) 152 (88.9) GT 14 (10.4) 18 (10.5) TT 0 1 (0.6) 489CϾT Exon 6 CC 132 (98.5) 170 (99.4) CT 2 (1.5) 1 (0.6) TT 0 0 1265TϾC rs2228078 Exon 13 Met422Thr TT 132 (98.5) 169 (98.8) TC 2 (1.5) 2 (1.2) CC 0 0 Data for patients and controls are expressed as number (%). a The numbering refers to the GenBank entries NT_007819 for the gene and NM_000823 for the mRNA. Nucleotide numbering reflects cDNA numbering with Љ1Љ corresponding to the A of the ATG translation initiation codon in the reference sequence. b http://www.ncbi.nlm.nih.gov. patients and controls (Table 1). Conversely, the 10Gly In silico functional analysis of the Val10Gly allele was detected at the heterozygous state in three of variation 134 patients and only in one of 1084 normal-stature The Val10Gly substitution lies within the GHRHR sig- ϭ individuals (P 0.004), suggesting an involvement in nal peptide. Most eukaryotic signal comprise the disease. three functionally important regions: a positively charged The auxological, biochemical, and family data of the three N-terminal domain, a hydrophobic central core (h-re- patients carrying the 10Gly sequence are reported in Fig. 1. gion), and a polar C-terminal domain that ends at the The analysis of family members of patients 2 and 3 showed signal peptide cleavage site (21, 22). For the wild-type that 10Gly was inherited from one parent in both cases and 10Val sequence, the SignalP 3.0, a widely used software for was present in a sister of patient 3, thus excluding a de novo signal peptides (http://www.cbs.dtu.dk/services/SignalP/), mutation. Unfortunately, DNA of subject 1’s parents was predicted a signal peptide with a probability of 99%, with not available. Notably, both parents of patient 2 are short- statured individuals. Neither of them nor the other 10Gly a cleavage site between positions 22 and 23, and an h- carriers were investigated for serum GH levels. region between amino acids 9 and 18 (Fig. 2). All of the The allelic combination of promoter SNPs in the amino acids of the h-region had an S-score (Signal peptide three patients revealed that 10Gly lies within at least prediction) greater than 50%. Notably, for the 10Gly mu- two different haplotypes, indicating that this variation tant the S-score fell below the 50% cutoff at different po- is recurrent. sitions (Fig. 2), indicating that the 10Gly substitution al- Sequencing of the proximal promoter, the exons, and ters the h-region properties. Conversely, for the 4Trp exon-intron flanking regions of GHRHR did not identify sequence, which had a comparable frequency in cases and any other mutation in the three patients. controls (Table 1), the scores were very similar to those of J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 jcem.endojournals.org 3943

transiently cotransfected, both isoforms were equally coexpressed (lane 2), indi- cating that 10Gly does not exert a nega- tive effect on the wild-type GHRHR. The expression on the cell surface of the wild-type and mutant receptors was assessed by immunofluorescence. The 10Val and 10Gly plasmids, tagged with either the FLAG or HA epitopes, were used to transiently transfect CHO cells.

A cDNA encoding the receptor mutated Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 at position 144 (144His, kind gift of FIG. 1. Pedigrees of IGHD patients heterozygous for the Val10Gly mutation. The Val10Gly Prof. R. Salvatori) was used as a further genotype is indicated (V/V, homozygous for 10Val; V/G, heterozygous Val10/Gly10) for the control because it carries a pathogenic tested family members. Height and height SDS is indicated for the patients’ parents. mutation that does not interfere with Biochemical and auxological parameters at diagnosis are reported for the three patients. For the patients’ sibs, the auxological data at the patient’s retesting visit are reported. Black the expression on the cell surface (16). symbols, IGHD established by measurement of GH serum levels; gray symbols, short-statured The FLAG-tagged 10Gly-cDNA was individuals, not tested for GH serum levels. a, Height corrected for the parents’ target; nt, not not expressed on the cell surface of non- tested for the Val10Gly genotype; arg, arginine; clo, clonidine; ITT, insulin tolerance test; MRI, magnetic resonance imaging. permeabilized CHO cells, whereas both the wild-type and the 144His plasmids showed a uniformly distributed fluores- the wild-type 4Arg (data not shown), confirming the be- cence on the (Fig. 5). CHO cells cotrans- nign nature of this rare variation. fected with the same quantity of the wild-type and 10Gly The Val10Gly resides within the hydrophobic core of cDNA showed a reduced cell surface staining of the re- the signal peptide, which is critical for interaction with ceptor with respect to cells transfected with the wild-type the signal recognition protein (SRP) and for transloca- (Fig. 5D). Double immunofluorescent staining of perme- tion to the ER (23). This substitution replaces an amino abilized CHO cells for an ER marker and the tagged-GH- acid (Val) of elevated hydrophobicity (4.2 according to RHR showed that the mutant receptor was localized prev- the Kyte-Doolittle scale) (24) with an amino acid (Gly) alently in the perinuclear portion of the ER and was absent that occupies an intermediate position in the hydrophil- close to the cell membrane (Fig. 5, K and S), whereas the Ϫ ic-hydrophobic scale ( 0.4). The Kyte-Doolittle algo- wild-type showed the typical pattern of the ER extending rithm (http://gcat.davidson.edu/rakarnik/KD.html) predicts through the entire cell. Permeabilized cells cotransfected a reduction of the hydrophobicity of the GHRHR central with the mutant and the wild-type plasmids showed an core when Val is substituted by Gly (Fig. 3). Because the intermediate distribution pattern with a relatively higher hydrophobic core is crucial for translocation of proteins des- amount of the perinuclear localization (Fig. 5, L and T). tined for secretion across the membranes of the ER, we hy- pothesized that the mutant protein could interfere with this process. Discussion

In vitro functional analysis of the Val10Gly We detected a novel variation within the GHRHR gene in variation three of 134 IGHD patients, causing the amino acid sub- Comparable levels of GHRHR mRNA were detected stitution Val10Gly in the signal peptide. by RT-PCR in cells transfected with the wild-type and the Signal peptides are N-terminal sequences that mediate 10Gly cDNA (data not shown). the sorting of secreted or membrane bound proteins. The Western blotting of the membrane lysate of transfected first stage in the targeting of the nascent polypeptide is the cells performed on the wild-type showed a main band of interaction of the signal peptide with an RNA-multipro- 45 kD according to the expected size of the mature protein tein complex, the signal recognition particles (SRP). The product and a faint larger sized band (48 kD) probably SRP binds to a specific receptor, on the ER, that allows corresponding to the uncleaved preprotein (Fig. 4). The the interaction between the -nascent peptide mutant produced only the higher molecular weight band, complex and the , a complex of proteins indicating that the 10Gly variation interfered with the forming a membrane channel that transports the pro- cleavage of the 22-amino acid signal peptide. When equal teins across the lipid bilayer, whereas the signal peptide amounts of the wild-type and 10Gly GHRHR cDNA were is generally cleaved by a (23). 3944 Godi et al. A Novel Recurrent Mutation in the GHRHR J Clin Endocrinol Metab, October 2009, 94(10):3939–3947

FIG. 4. Western blotting of the GHRHR protein tagged with the HA epitope. CHO cells were transiently transfected with the 10Gly mutant (lane 1) or wild-type cDNA (lane 3) or cotransfected with both cDNAs (lane 2). Nontransfected cells served as negative control (lane 4). Naϩ/ Kϩ ATPase was used as reference (internal control). Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 of these regions are tolerant of aminoacidic substitutions as long as their hydrophobicity is retained (25). When comparing the homologous sequences of GHRHR in various mammalian species (Fig. 6), position 10 is occupied either by valine or isoleucine, two amino acids with very similar characteristics, but never by gly- cine. Gly differs from Val and Ile in several physicochem- ical properties thought to be important in the determina- tion of protein structure (26). In fact, it has the smallest side chain among nonpolar amino acids, and it occupies an FIG. 2. Comparison of the N-terminal sequence of the wild-type intermediate position on hydrophilic-hydrophobic scales (10Val) (A) and mutant (10Gly) (B) GHRHR performed with the SignalP (Ϫ0.4), whereas both Val and Ile have a larger molecular program (version 3.0). The software yields different parameters: 1) the mass and are included among the most hydrophobic S-score reports the signal peptide prediction for every amino acid position, with a high score indicating that the particular amino acid is amino acids (Val 4.20, Ile 4.50) (27). part of a functional signal peptide; 2) the C-score is the cleavage score The functional analysis of the Val10Gly mutation at every amino acid position; 3) the Y-score is a derivative of the was performed by transfecting the mutant and wild- C-score combined with the S-score, resulting in a better cleavage type GHRHR cDNA in CHO cells. The higher molecular site prediction. For the three scores, results greater than 50% (cutoff value) indicate proper signal peptide function and cleavage, weight of the GHRHR protein produced by the mutated whereas values below the cutoff indicate a compromised function. allele (Fig. 4) suggested that the mutation in the h-region The h-region is underlined. abolishes cleavage of the 22-bp signal peptide, although cleavage is more likely to be affected by mutations in the Signal peptides are between 20 and 30 residues in C-region. The change in the hydrophobicity of the core length, characterized by a central hydrophobic core of region might lead to poor recognition by the SRP and/or approximately 10–15 residues (h-region) that is impor- an altered positioning of the nascent peptide in the translocon tant for functioning as a translocation signal (21, 22). This that in turn could affect the interaction with the signal pep- domain is flanked by an N-terminal hydrophilic region tidase and consequently cleavage, in agreement with previ- and a neutral, but polar, C-terminal region (C-region). All ously reported signal peptide mutations (28–32). As a consequence of the incomplete maturation, the mu- tant GHRHR was absent at the cell surface because it was retained in the perinuclear zone within the ER, blocked at the first stages of intracellular trafficking (Fig. 5). This is in line with the results of other studies showing that mutations that impair cleavage of the signal peptide lead to the retention in the ER of proteins designated to be expressed on the membrane or secreted (33, 34). Cotransfection of the wild-type 10Val and mutant 10Gly plasmids resulted in an apparent minor amount of FIG. 3. Hydrophobicity profiles of the wild-type (10Val) and mutant (10Gly) signal peptide region of the GHRHR. The predicted signal the receptor on the plasma membrane. Thus, in our ex- peptide hydrophobicity was calculated considering the residue specific perimental conditions the presence of the mutant allele did values within a nine-residue window using the Kyte-Doolittle not exert an evident dominant-negative effect severely algorithm. Positive and negative values are defined as being hydrophobic or hydrophilic, respectively. Position 10 is boxed. The h- compromising the ability of the wild-type to translocate to region is underlined. the cell surface. We cannot exclude that a dominant-neg- J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 jcem.endojournals.org 3945

1)10Val2)144His 3)10Gly 4)10Val/ Gly A BCD

Intact cells Intact E F GH anti-FLAG

I JKL anti-FLAG+antiCon-A Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 M NOP anti-HA P FIG. 6. Multiple alignment of the GHRHR signal peptide sequences

Permeabilized cells Permeabilized from different species. The residue corresponding to the Val10Gly QRS T mutation is boxed. ant i-HA+ant iCon-A The previously reported mutations showed complete penetrance and caused IGHD only when present on both FIG. 5. Immunolocalization of epitope-tagged wild-type and mutant human GHRHRs transiently expressed in CHO cells. The four columns chromosomes. Instead, the 10Gly mutation was present report experiments conducted with plasmids bearing different GHRHR only on one allele in our IGHD patients. The most likely cDNAs: 1) 10Val; 2) 144His; 3) 10Gly; and 4) both 10Val and 10Gly. hypothesis is that the 10Gly sequence exerts its effects in The FLAG epitope was inserted at the N terminal, and the HA at the C terminal of the receptor (see Subjects and Methods). Cell surface the context of other unknown genetic and/or environmen- expression was determined on intact non-permeabilized cells, using tal factors. This is supported by the fact that 10Gly was the anti-M2 monoclonal antibody against the FLAG-epitope (A–D): the inherited from normal-stature parents at least in the case wild-type 10Val and the 144His plasmids, treated with anti-FLAG of patient 3. Moreover, it was present in a patient’s nor- antibodies (A and B), showed a uniform distribution of the fluorescence on the cell surface, whereas the 10Gly-cDNA showed no mal-stature sib (sister of patient 3; Fig. 1) and was detected fluorescence (C), similar to the negative control (cell transfected with in one male control individual with a height 0.56 SDS of Fugene alone; data not shown). CHO cells cotransfected with the wild- the mean. This incomplete penetrance might be explained type and 10Gly cDNA showed a reduced cell surface staining of the receptor with respect to cells transfected with the wild-type (D). by the co-occurrence in the three patients of a mutation Permeabilized cells were stained with both anti-FLAG (E–H) and anti- affecting another gene. Patient 2 might have inherited a HA (M–P). The same cells were simultaneously stained with the ConA- second unknown mutation from his 10Gly-negative Alexa Fluor633 antibody. Merged pictures show the anti-FLAG (I–L) or the anti-HA staining (Q–T) (red or green, respectively), together with mother who has a severe short stature, thus carrying two the staining of ER with anti-ConA (blue). The wild-type and the 144His mutations on two different genes (digenic inheritance) and control showed a staining with the typical reticular pattern of ER resulting in severe short stature and GH deficiency. In the extending through the entire cell and the cell membrane (I, J, and Q, three patients, we have excluded mutations in the coding R). In cells expressing the mutant receptor, the localization of the tagged GHRHR was prevalently in the perinuclear portion of the ER and proximal promoter sequences of one obvious candi- and was completely absent in the proximity of the cell membrane (G, date, the GHRH (data not shown), but there are many K, and O, S). Permeabilized CHO cells cotransfected with the mutant other candidate genes that regulate GH production and and the wild-type plasmids showed an intermediate distribution pattern with a relatively higher amount of the protein with a are worth considering. Notably, of five 10Val male car- perinuclear localization with respect to the wild-type (H, L, and P, T). riers (three IGHD, one relative not tested for GH, and one normal control), four had short stature, whereas the two ative effect is present on the somatotroph cells of the het- carrier females were unaffected. This difference is not sta- erozygous patients. Under the hypothesis that the 10Gly tistically significant due to the very low numbers, but it is sequence is associated in vivo,asin vitro, to a lower con- intriguing and is in line with the excess of males in IGHD centration of the membrane receptors, the heterozygous also evident in our sample (92 males and 42 females). The patients should be similar to the parents of homozygous presence of genetic factors on the X-chromosome contrib- patients carrying a null mutation. However, these muta- uting to IGHD might in part explain this sex distortion. tions are usually associated with an autosomal recessive In a recent report, de Graaff et al. (36) identified a mode of inheritance, and the patients’ parents were de- Pro18Leu variation in the GHRHR signal peptide in one scribed as normal, with the exception of a few cases in heterozygous IGHD patient out of 89. The authors did not which heterozygous subjects showed milder symptoms perform a functional or predictive analysis of this muta- (35, 12) and a possible dominant-negative effect of the tion; however, this is a nonconservative amino acid change frameshift mutation reported by Horikawa (12). that deserves further analysis. Variations in the signal pep- 3946 Godi et al. A Novel Recurrent Mutation in the GHRHR J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 tide may thus represent the cause of a novel form of IGHD 10. Salvatori R, Fan X, Veldhuis JD, Couch R 2002 Serum GH response manifesting in heterozygous patients and inherited with to pharmacological stimuli and physical exercise in two siblings with two new inactivating mutations in the GH-releasing hormone re- incomplete penetrance. The phenotype of these heterozy- ceptor gene. Eur J Endocrinol 147:591–596 gous patients might be associated with milder clinical fea- 11. Salvatori R, Fan X, Mullis PE, Haile A, Levine MA 2002 Decreased tures than the recessive form. For instance, of the three expression of the GHRH receptor gene due to a mutation in a Pit-1 patients heterozygous for 10Val, only one showed mag- binding site. Mol Endocrinol 16:450–458 12. Horikawa R 2002 Isolated GH deficiency due to inactivating mu- netic resonance imaging anomalies (Fig. 1). tation of GHRH receptor [Japanese]. Nippon Rinsho 60:297–305 At difference with the previous studies that detected 13. Salvatori R, Hayashida CY, Aguiar-Oliveira MH, Phillips 3rd JA, GHRHR mutations in pedigrees with a clear recessive Souza AH, Gondo RG, Toledo SP, Conceica˜o MM, Prince M, Maheshwari HG, Baumann G, Levine MA 1999 Familial Dwarfism transmission or declared consanguinity, our study was due to a novel mutation of the growth hormone-releasing hormone conducted on putatively sporadic IGHD patients. Thus, receptor gene. J Clin Endocrinol Metab 84:917–923 Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 the finding of a dominant mutation with incomplete pen- 14. Alba M, Hall CM, Whatmore AJ, Clayton PE, Price DA, Salvatori etrance is not in contrast with previous reports of a reces- R 2004 Variability in anterior pituitary size within members of a family with GH deficiency due to a new splice mutation in the sive inheritance of GHRHR mutations. GHRH receptor gene. Clin Endocrinol (Oxf) 60:470–475 15. Hilal L, Hajaji Y, Vie-Luton MP, Ajaltouni Z, Benazzouz B, Chana M, Chraïbi A, Kadiri A, Amselem S, Sobrier ML 2008 Unusual Acknowledgments phenotypic features in a patient with a novel splice mutation in the GHRHR gene. Mol Med 14:286–292 Address all correspondence and requests for reprints to: Mara 16. Alba M, Salvatori R 2005 Naturally-occurring missense mutations in the human growth hormone-releasing hormone receptor alter Giordano, Department of Medical Sciences, University of East- ligand binding. J Endocrinol 186:515–521 ern Piedmont, Via Solaroli 17, 28100 Novara, Italy. E-mail: 17. Growth Hormone Research Society 2000 Consensus guidelines for [email protected]. the diagnosis and treatment of growth hormone (GH) deficiency in This work was supported by Eastern Piedmont University childhood and adolescence: summary statement of the GH Research (Ricerca Locale 2007), Regione Piemonte (Ricerca Sanitaria Society. GH Research Society. J Clin Endocrinol Metab 85:3990– Finalizzata 2008bis), and Pfizer. 3993 Disclosure Summary: The authors have nothing to disclose. 18. Ghigo E, Bellone J, Aimaretti G, Bellone S, Loche S, Cappa M, Bartolotta E, Dammacco F, Camanni F 1996 Reliability of provoc- ative tests to assess growth hormone secretory status. Study in 472 normally growing children. J Clin Endocrinol Metab 81:3323–3327 References 19. Tanner JM, Whitehouse RH, Cameron N, Marshall WA, Healy WA, Goldstein H 1988 Assessment of skeletal maturity and predic- 1. Petersenn S, Schulte HM 2000 Structure and function of the growth- tion of adult height (TW2 method). San Diego: Academic Press hormone-releasing hormone receptor. Vitam Horm 59:35–69 20. Jones AC, Austin J, Hansen N, Hoogendoorn B, Oefner PJ, Cheadle 2. Mu¨ ller EE, Locatelli V, Cocchi D 1999 Neuroendocrine control of JP, O’Donovan MC 1999 Optimal temperature selection for muta- growth hormone secretion. Physiol Rev 79:511–607 tion detection by denaturing HPLC and comparison to single- 3. Mayo KE, Godfrey PA, Suhr ST, Kulik DJ, Rahal JO 1995 Growth stranded conformation polymorphism and heteroduplex analysis. hormone-releasing hormone: synthesis and signaling. Rec Prog Clin Chem 45:1133–1140 Horm Res 50:35–73 21. Martoglio B, Dobberstein B 1998 Signal sequences: more than just 4. Salvatori R, Fan X, Phillips 3rd JA, Espigares-Martin R, Martin De greasy peptides. Trends Cell Biol 8:410–405 Lara I, Freeman KL, Plotnick L, Al-Ashwal A, Levine MA 2001 22. Nothwehr SF, Gordon JI 1990 Structural features in the NH2-ter- Three new mutations in the gene for growth hormone (GH)-releas- ing hormone receptor in familial isolated GH deficiency type IB. minal region of a model eukaryotic signal peptide influence the site J Clin Endocrinol Metab 86:273–279 of its cleavage by signal peptidase. J Biol Chem 265:17202–17208 5. Salvatori R, Fan X, Phillips 3rd JA, Prince M, Levine MA 2001 23. Nagai K, Oubridge C, Kuglstatter A, Menichelli E, Isel C, Jovine L Isolated growth hormone (GH) deficiency due to compound het- 2003 Structure, function and evolution of the signal recognition erozygosity for two new mutations in the GH-releasing hormone particle. EMBO J 22:3479–3485 receptor gene. Clin Endocrinol (Oxf) 54:681–687 24. Kyte J, Doolittle RF 1982 A simple method for displaying the hy- 6. Salvatori R, Aguiar-Oliveira MH, Monte LV, Hedges L, Santos NL, dropathic character of a protein. J Mol Biol 157:105–132 Pereira RM, Phillips JA 2002 Detection of a recurring mutation in 25. Zanen G, Houben EN, Meima R, Tjalsma H, Jongbloed JD, Westers the human growth hormone- releasing hormone receptor gene. Clin H, Oudega B, Luirink J, van Dijl JM, Quax WJ 2005 Signal peptide Endocrinol (Oxf) 57:77–80 hydrophobicity is critical for early stages in protein export by Ba- 7. Carakushansky M, Whatmore AJ, Clayton PE, Shalet SM, Gleeson cillus subtilis. FEBS J 272:4617–4630 HK, Price DA, Levine MA, Salvatori R 2003 A new missense mu- 26. Taylor WR 1986 The classification of amino acid conservation. J tation in the growth hormone-releasing hormone receptor gene in Theor Biol 119:205–218 familial isolated GH deficiency. Eur J Endocrinol 148:25–30 27. Cornette JL, Cease KB, Margalit H, Spouge JL, Berzofsky JA, DeLisi 8. Haskin O, Lazar L, Jaber L, Salvatori R, Alba M, Kornreich L, C 1987 Hydrophobicity scales and computational techniques for Phillip M, Gat-Yablonski G 2006 A new mutation in the growth detecting amphipathic structures in proteins. J Mol Biol 195:659– hormone-releasing hormone receptor gene in two Israeli Arab fam- 685 ilies. J Endocrinol Invest 29:122–130 28. Arnold A, Horst SA, Gardella TJ, Baba H, Levine MA, Kronenberg 9. Wajnrajch MP, Gertner JM, Harbison MD, Chua Jr SC, Leibel RL HM 1990 Mutation of the signal peptide-encoding region of the 1996 Nonsense mutation in the human growth hormone-releasing preproparathyroid hormone gene in familial isolated hypoparathy- hormone receptor causes growth failure analogous to the little (lit) roidism. J Clin Invest 86:1084–1087 mouse. Nat Genet 12:88–90 29. Green PM, Mitchell VE, McGraw A, Goldman E, Giannelli F 1993 J Clin Endocrinol Metab, October 2009, 94(10):3939–3947 jcem.endojournals.org 3947

Haemophilia B caused by a missense mutation in the prepeptide 33. Beuret N, Rutishauser J, Bider MD, Spiess M 1999 Mechanism of sequence of factor IX. Hum Mutat 2:103–107 endoplasmic reticulum retention of mutant vasopressin precursor 30. Pidasheva S, Canaff L, Simonds WF, Marx SJ, Hendy GN 2005 caused by a signal peptide truncation associated with diabetes in- Impaired cotranslational processing of the calcium-sensing receptor sipidus. J Biol Chem 274:18965–18972 due to signal peptide missense mutations in familial hypocalciuric 34. Chan D, Ho MS, Cheah KS 2001 Aberrant signal peptide cleavage hypercalcemia. Hum Mol Genet 14:1679–1690 of X in Schmid metaphyseal chondrodysplasia. Implica- 31. Lanza F, De La Salle C, Baas MJ, Schwartz A, Boval B, Cazenave JP, tions for the molecular basis of the disease. J Biol Chem 276:7992– Caen JP 2002 A Leu7Pro mutation in the signal peptide of platelet 7997 glycoprotein (GP) IX in a case of Bernard-Soulier syndrome abol- 35. Maheshwari HG, Silverman BL, Dupuis J, Baumann G 1998 Phe- ishes surface expression of the GPIb-V-IX complex. Br J Haematol notype and genetic analysis of a syndrome caused by an inactivating 118:260–266 mutation in the growth hormone-releasing hormone receptor: 32. Kira´ly O, Boulling A, Witt H, Le Mare´chal C, Chen JM, Rosendahl Dwarfism of Sindh. J Clin Endocrinol Metab 83:4065–4074 J, Battaggia C, Wartmann T, Sahin-To´thM,Fe´rec C 2007 Signal 36. de Graaff LC, Argente J, Veenma DC, Herrebout MA, Friesema EC,

peptide variants that impair secretion of pancreatic secretory trypsin Uitterlinden AG, Drent ML, Campos-Barros A, Hokken-Koelega Downloaded from https://academic.oup.com/jcem/article/94/10/3939/2597261 by guest on 26 September 2021 inhibitor (SPINK1) cause autosomal dominant hereditary pancre- AC 2009 Genetic screening of a Dutch population with isolated atitis. Hum Mutat 28:469–476 growth hormone deficiency. Clin Endocrinol (Oxf) 70:742–750