13 14 LPS composition determines OMV entry route into host cells |2 SUMMARY 15Outer membrane vesicles are microvesicles shed by Gram-negative bacteria and play important 16 roles in immune priming and disease pathogenesis. However, our current mechanistic 17 understanding of vesicle -host cell interactions is limited by a lack of methods to study the kinetics 18 of vesicle entry and cargo delivery to host cells in real-time. Here, we describe a highly sensitive 19 method to study the kinetics of vesicle entry into host cells in real-time using a genetically encoded 20 probe targeted to vesicles. We found that route of vesicular uptake, and thus entry kinetics and 21 efficiency of cargo release, are determined by the chemical composition of the bacterial 22 lipopolysaccharide. The presence of O-antigen facilitates receptor-independent entry, which 23 enhances both rate and efficiency of cargo uptake by host cells. Collectively, our findings highlight 24 the chemical composition of the bacterial cell wall as a major determinant of secretion-independent 25 delivery of virulence factors during Gram-negative infections. 26 27 28 Keywords: outer membrane vesicles, enterohemorrhagic Escherichia coli, FRET assay, 29 extracellular vesicles, endocytosis, vesicle trafficking; 30 31 LPS composition determines OMV entry route into host cells |3 LPS composition determines OMV entry route into host cells |4 Release of OMVs occurs during infection, and has advantages over other secretion systems. They 53 carry a broad range of cargo, from protein toxins such as VacA and shiga toxin, to hydrophobic 54 molecules such as Pseudomonas quinolone signal (PQS), the quorum sensing molecule of 55 Pseudomonas aeruginosa, and this cargo is protected from the potentially hostile extracellular 56 environment (Kulp and Kuehn, 2010; Berleman and Auer, 2013). OMV-mediated delivery of 57 virulence factors occurs without requiring close proximity between the bacterial cell and the host 58 cell (Bomberger et al, 2009). The small size of OMVs (20-200 nm) has made studying their 59 interactions with host cells in real time difficult. Previous work has often relied on OMVs labelled 60 with dyes such as fluorescein isothiocyanate (FITC) or dioctadecyloxacarbocyanine perchlorate 61 (DiO). While such dyes allow real time study of OMV entry and cargo delivery processes, the use 62 of membrane labelling of the vesicles may interfere with their physiological characteristics, and 63 alter the mechanism of OMV entry and cargo release (Bauman and Kuehn, 2009; Lulevich et al, 64 2009; Parker et al, 2010). Other approaches have used immunolabelling of OMV-associated 65 epitopes, such as hemolysin (HlyA) in enterohemorrhagic Escherichia coli (EHEC), but this 66 requires fixation of cells at pre-determined time points, and requires assumptions about OMV 67 cargo, which may ignore natural sub-populations of OMVs (Bielaszewska et al, 2013). Some 68 experiments have used host cell phenotypes as an indicator of OMV uptake, such as downstream 69 activation of NF-κB res...