Rhomboid Proteolysis is a Rate-Governed Reaction, Yet is Dispensable for E. coli Colonization of the Mouse Colon

Proteolysis, the controlled dissection of proteins into two parts, is essential to all life. Intramembrane proteolysis, whereby specialized enzymes cut membrane-resident segments, extends this potent chemical transformation to the unique membrane milieu. Intense research on rhomboid proteases, one of three widely conserved intramembrane protease families, has greatly advanced our knowledge of their biology in eukaryotes and their impact on human health and disease. Nonetheless, we have scant information about their roles in bacteria, which constitute the overwhelming diversity on Earth and burden humanity with myriad diseases. In this thesis, I tackle this deficit using the complementary approaches of genetics and biochemistry. In chapter II, I describe my work using reverse genetics and substrate identification to uncover the role of the Escherichia coli rhomboid protease, GlpG, with special focus on the lateral substrate gate. I have found that the lateral gate likely endows GlpG with allosteric regulation, and that GlpG cleaves two E. coli proteins, BasS and Rtn. Surprisingly however, GlpG is dispensable for normal growth in pure culture, response to diverse stresses, and colonization of and persistence within the mouse colon. In chapter III, I sought to generate new leads for bacterial rhomboid protease biology by investigating their biochemistry. I developed a steady-state kinetic assay and complementary thermodynamic methods to examine the kinetics of rhomboid protease cleavage in their natural membrane environment. Unexpectedly, nine diverse bacterial rhomboid proteases show identical and weak affinities (KM ~ 0.14 mole%, 135 μM; Kd ~ 0.08 mole%, 191 μM) with four substrate variants. In contrast, the turnover constant ranged 10,000-fold; yet, all rhomboid proteases displayed slow kinetics with the fastest (kcat ~0.1 s-1) being 1,000-fold slower than trypsin (kcat ~ 100 s-1). This kinetic profile exposes an unanticipated similarity to the DNA repair enzymes DNA glycosylases. I conclude with a new model for rhomboid proteolysis as a kinetically driven process, with the rich precedence of DNA glycosylases to guide future research