Congenital heart disease (CHD) is recognized as a pleiotropic and a complex genetic disease with a genetic etiology established only in <20% cases. We performed whole exome sequencing in 200 individuals with CHD including 17 affected individuals from 7 pedigrees, and 183 unrelated probands. Of 216 novel/rare protein-altering variants that segregated amongst affected members of 7 pedigrees, 58 were novel/rare variants in 57 genes expressed during early cardiac development. Gene ontology clustering revealed functional clustering among 9 structural genes involved in extracellular matrix (ECM) and cytoskeletal function (p=1.34× 10^-03). In the overall primary cohort of 200 CHD cases, novel/rare protein-altering variants (inherited or de novo) in these 9 genes were detected in 26 (13%) CHD cases. These included 10 patients with OBSCN variants and 4 with LAMA3 variants, two closely interacting genes. Importantly, 5 patients had co-segregating variants in OBSCN and/or LAMA3. Targeted sequencing for OBSCN and LAMA3 in a replication cohort of 241 CHD cases and 91 controls without CHD revealed a compound heterozygosity for rare protein-altering variants in OBSCN and LAMA3 in 11 patients compared to no compound heterozygosity in controls (odds ratio 9.21; p=0. 0.039). Combined expression knockdown of both genes together caused CHD using subclinical doses of morpholinos suggesting that obscn and lama3 disruption may interact to cause CHD. Our findings suggest that CHD may be caused by accumulation of compound mutations in interacting genes and supports a multi-genic contribution towards sporadic and inherited CHD.