On a new factor in the mechanism of bacterial infection