Copy number variation of the bacteria-binding mucosal glycoprotein DMBT1: mutation rate, balancing selection and transcript length

The DMBT1 gene codes for the 340kDa DMBT1 glycoprotein, also known as SALSA or salivary agglutinin, which is predominantly expressed in saliva and at mucosal surface, and binds to a wide variety of pathogens through tandemly-arranged scavenger receptor cystine-rich (SRCR) domains. DMBT1 shows extensive multiallelic germline copy number variation (CNV) across all populations, with the tandemly-repeated CNV leading to different alleles with between 7 to 21 SRCR domains. The de novo mutation rate at the DMBT1 CNV1 and CNV2, using a segregation study from multigenerational CEPH pedigrees, was one of the fastest mutation rates that may affect the coding sequence, 1.4% and 3.3%, respectively. Population genetics analysis was used as an indirect method to estimate the recombination rate based on the historic patterns of genetic variations among DNA sequences of natural populations. The recombination rate showed a very high mutation rate at DMBT1 CNV2 under the standard neutral model, from all populations was 0.58% recombination per generation across the size of the average copy number of CNV2, similar values found also at the European and African populations, 0.76% and 0.32%, respectively. Genome scans using genetic variation data have identified regions of balancing selection, yet in almost all cases the genetic and functional basis is unknown. One such region is at the 5′ end of the DMBT1 gene. Here, using expression analysis, linkage in pedigrees, and long-range single transcript sequence, I show that the signal of balancing selection is best explained by variable length SRCR tandem repeats within the coding region of DMBT1. It is known that different length SRCR tandem repeats affect the binding affinity of salivary agglutinin to bacteria, including Helicobacter pylori and Streptococcus mutans. Taken together, our evidence suggests that balancing selection at DMBT1 is due to different bacterial binding affinities of encoded SRCR tandem repeat alleles.