The type IV pilus protein PilU functions as a PilT-dependent retraction ATPase

Type IV pili are dynamic cell surface appendages found throughout the bacteria. The ability of these structures to undergo repetitive cycles of extension and retraction underpins their crucial roles in adhesion, motility and natural competence for transformation. In the best-studied systems a dedicated retraction ATPase PilT powers pilus retraction. Curiously, a second presumed retraction ATPase PilU is often encoded immediately downstream of pilT. However, despite the presence of two potential retraction ATPases, pilT deletions lead to a total loss of pilus function, raising the question of why PilU fails to take over. Here, using the DNA-uptake pilus and mannose-sensitive haemagglutinin (MSHA) pilus of Vibrio cholerae as model systems, we show that inactivated PilT variants, defective for either ATP-binding or hydrolysis, have unexpected intermediate phenotypes that are PilU-dependent. In addition to demonstrating that PilU can function as a bona fide retraction ATPase, we go on to make the surprising discovery that PilU functions exclusively in a PilT-dependent manner and identify a naturally occurring pandemic V. cholerae PilT variant that renders PilU essential for pilus function. Finally, we show that Pseudomonas aeruginosa PilU also functions as a PilT-dependent retraction ATPase, providing evidence that the functional coupling between PilT and PilU could be a widespread mechanism for optimal pilus retraction. Author summary Bacteria interact with their surroundings using micrometre scale polymers called type IV pili. They allow bacteria to physically sense, attach and move on surfaces, and even to take up DNA. Consequently they represent important mechanisms of environmental survival and pathogenesis. The versatility of type IV pili is made possible by dedicated motors that power repeated cycles of extension and retraction. Curiously, although the ATPase PilT is well established as the retraction motor, many species have an additional PilT-like protein called PilU. However, how PilU functions has remained unclear, especially since when PilT is absent it is unable to take over its function. In this work we took a different approach. Instead of deleting pilT, we made inactivated variants and studied the functionality of two distinct types of pili used by the human pathogen Vibrio cholerae to survive in its natural aquatic environment. This allowed us to make the unexpected discovery that PilU is capable of acting as a retraction ATPase, but that it is not an independent motor and instead exerts its function via PilT. Our results suggest this functional coupling between PilT and PilU may be common in other bacteria.