Leonid, Thanks. As yet, there is no experimental proof that type III secretory proteins actually thread-pass through the rather narrow channel of injectisome. The number and types of pathogenicity proteins that have to be sequentially orchestrated through this narrow channel appears to be a kinetically losing game for the microbe. Also, fusion proteins necessary for needle tip to penetrate the host cell cytoplasmic membrane are unlikely to get mounted on the tip. On the other hand, the translocation process can be easily and visibly explained by outer membrane vesicles - which contain type III fusion proteins on the membrane and other pathogenicity proteins too in matrix and membrane. You may see full proposition at my researchgate page in article "eukaryotic cell intoxication by gram negative pathogens: ........"