ABSTRACT Respiratory syncytial virus (RSV) is the leading cause of hospitalization of children under one year of age, but there are no approved vaccines or effective anti-viral treatments. RSV infects epithelial cells in the upper and lower respiratory tracts causing severe bronchiolitis in young children and adults. The use of primary human bronchial epithelial cultures (HBEC) has enabled us to identify its target cell type (ciliated) as well as better understand its pathogenesis in an ex vivo model. HBEC are grown at an air-liquid interface allowing differentia- tion of progenitor cells into pseudostratified, ciliated and mucus-producing cultures that mirror the human air- ways. We have found that RSV produced in HBEC (RSV/HBE) contains a much larger version of its attachment (G) glycoprotein (G/HBE; 170 kDa) than the “classical” 90 kDa form produced in immortalized cells. We have evidence consistent with this size difference being due to extra glycosylation of the G protein in HBEC compared to HEp-2 (immortalized) cells. We have also found that RSV/HBEC is 100-fold less infectious for HEp-2 cells compared to HBEC, based on qRT-PCR quantified viral genomes relative to the infectivity of GFP-expressing RSV. We will identify the structural reason for the larger G produced in HBEC and determine the importance of G glycosylation for infection of HBEC. The results may elucidate the role of G in vivo, improve RSV vaccine production and allow much more efficient isolation of RSV from clinical samples without selection for modified viruses able to grow in HEp-2 cells.