Study of monogenic forms of obesity has demonstrated the pivotal role of the central leptin–melanocortin pathway in controlling energy balance, appetite and body weight1. The majority of loss-of-function mutations (mostly recessive or co-dominant) have been identified in genes that are directly involved in leptin–melanocortin signaling. These genes, however, only explain obesity in <5% of cases, predominantly from outbred populations2. We previously showed that, in a consanguineous population in Pakistan, recessive mutations in known obesity-related genes explain ~30% of cases with severe obesity3,4,5. These data suggested that new monogenic forms of obesity could also be identified in this population. Here we identify and functionally characterize homozygous mutations in the ADCY3 gene encoding adenylate cyclase 3 in children with severe obesity from consanguineous Pakistani families, as well as compound heterozygous mutations in a severely obese child of European-American descent. These findings highlight ADCY3 as an important mediator of energy homeostasis and an attractive pharmacological target in the treatment of obesity.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors thank the patients and their families for participation in the study. We are grateful to A. W. Rathore for clinical facilitation, Z. Gilani for his advice regarding ADCY3 molecular modeling, M. Boissel for help with statistical analyses, and Q. Ain and I. Qureshi for their technical assistance. This study was supported by the Fédération de Recherche 3508 Labex EGID (European Genomics Institute for Diabetes; ANR-10-LABX-46) (to P.F.), the ANR Equipex 2010 session (ANR-10-EQPX-07-01; ‘LIGAN-PM’) (to P.F.), the European Community (FEDER) and the Region Hauts-de-France (to P.F.). The research leading to this study was also supported by funding from the European Research Council GEPIDIAB 294785 (P.F.) and the Pakistan Academy of Sciences (to M. Arslan).
Integrated supplementary information
Supplementary Figure 1 Sequencing scans of ADCY3 in the four probands (P1–P4) at the site of mutation.
mutADCY3-I is a 1-base deletion that lies in the last exon (coding exon 21) of the ADCY3 gene resulting in a truncated protein by introducing a stop codon after two aberrant amino acids; mutADCY3-II is a substitution at a splice acceptor site in intron 15 that abolishes this site as predicted by the Human Splicing Finder (Supplementary Note) resulting in skipping of exon 16. As a consequence, the protein carries 1,091 amino acids rather than the normal number of 1,144 amino acids; mutADCY3-III is a missense mutation (Asn>Ile) that lies in exon 1 of ADCY3; mutADCY3-IV is a compound heterozygous mutation, mutADCY3-IVa is a frameshift mutation and mutADCY3-IVb is an in-frame 3-base deletion. mutADCY3-IVa results in a change from Gly to Ala at amino acid 423 leading to 18 aberrant amino acids before introducing a stop codon. mutADCY3-IVb results in deletion of the Phe at amino acid position 1118.
Comparative sequence analysis showing conservation of the mutated ADCY3 regions at the protein level. Top, protein alignment for the homozygous mutations from consanguineous families. Bottom, protein alignment for the compound heterozygous mutation. Nucleotides in intronic regions are shaded gray.
Supplementary Figure 3 ADCY3 expression in BHK cells subsequent to transient transfection with pcDNA3 alone (CON) or human ADCY3 constructs (wt, mut I–IVb).
a, Representative blots with anti-ADCY3 antibody (top) and anti-HSP70 antibody (bottom) as loading control. Molecular weight standards (in kDa) are given to the right. b, Quantification of blots normalized for each experiment to values obtained for ADCY3 wt. n = 7–9. Means are represented as horizontal bars, s.d. as boxes, s.e.m. as whiskers and points represent individual data. c, Coomassie Blue staining of the PVDF membrane shown in a.
Supplementary Figure 4 Immunofluorescence images of BHK cells transiently transfected with plasmids encoding Myc-tagged ADCY3 wild-type or mutants (I, II, III, IVb).
ADCY3 expression was visualized by either anti-Myc or anti-ADCY3 antibody followed by Cy3-coupled secondary antibody. Nuclei were stained with Hoechst 33342.
CHO cells were transfected for 44 h with CreLuc reporter and either empty vector (Ctrl) or plasmids encoding Myc-tagged hADCY3wt (Wt) or hADCY3 mutants (Mut-I, Mut-II, Mut-III and Mut-IVb). Thereafter, culture medium was replaced and cells were cultured with a stimulatory medium containing 2% FCS and 1 μM forskolin plus 100 μM IBMX for 4 h. Luciferase activity was measured and normalized to β-galactosidase activity. The data are represented as the fold increase in luciferase activity in the stimulated condition over unstimulated cells in five independent experiments performed in triplicate and are expressed as the mean ± s.e.m. of three independent experiments performed in triplicate; an asterisk indicates significant differences: Mut-I vs. WT P = 0.017, Mut-II vs. WT P = 0.0025, Mut-III vs. WT P = 0.12, Mut-IV vs. WT P = 0.0041 (ANOVA and Bonferroni post hoc test).
a, cAMP generation in BHK cells stimulated by forskolin (1 μM, in the presence of 100 μM IBMX). An asterisk indicates significant differences: CON (n = 18) vs. WT (n = 18) P < 0.001, CON (n = 18) vs. mutant V (n = 8) P < 0.001; no significant difference between WT and mutant V (P = 0.054). Means are represented as horizontal bars, s.d. as boxes, s.e.m. as whiskers and points represent individual data. b, Western blot analysis for ADCY3 and tubulin (TUB) from transiently transfected BHK cells with Myc-tagged hADCY3wt (WT) or plasmid encoding ADCY3 carrying the variant rs11676272 (V). Molecular weight standards are given in kDa. Expression levels of mutant V were 70 + 41.4% of WT levels (n = 7–12). c, CHO cells were transfected for 44 h with CreLuc reporter and either empty vector (CON) or plasmid encoding either Myc-tagged hADCY3wt (WT) or ADCY3 carrying the variant rs11676272 (V). Thereafter, culture medium was replaced and cells were cultured with either a medium containing 2% FCS (unstimulated) or a medium containing 2% FCS and 1 μM forskolin plus 100 μM IBMX (stimulated) for 4 h. Luciferase activity was measured and normalized to β-galactosidase activity. The data are expressed as the mean ± s.e.m. of four independent experiments performed in triplicate. An asterisk indicates significant differences: WT vs. CON P = 0.032, CON vs. V P = 0.003, WT vs. V P = 0.66.
Supplementary Figure 7 Relative gene expression from cDNA of homozygous and heterozygous carriers of mutADCY3-1.
RNA was extracted from whole blood, and results were analyzed with the ΔΔCd method. Values are mean ± s.e.m. (n = 3 replicate measurements). No significant change was found in expression of the mutants relative to the wild-type.
a, Model-template sequence alignment. b, z score QMEAN, representing the model inside the non-redundant set of PDB structures regarding perfection. c, 3D structure of the C1 catalytic domain of wtADCY3. d, Superimposition of the C1 domain of Canis lupus (dog) ADCY5 (in pink) on the C1 domain of human wtADCY3 (in blue). More notably, the Mg-ATP key interacting residues in both structures are also conserved.
a, Model-template sequence alignment. b, z score QMEAN, representing the model inside the non-redundant set of PDB structures regarding perfection. c, 3D structure of the C2 catalytic domain of wtADCY3. d, Superimposition of the C2 domain of Rattus norvegicus (rat) ADCY2 (in pink) on the C2 domain of human wtADCY3 (in blue). Here the Mg-ATP key interacting residues in both structures are also conserved.