Dimeric SecA is essential for protein translocation

doi:10.1073/pnas.0502774102

Dimeric SecA is essential for protein translocation

Department of Molecular Biology and Biochemistry, Wesleyan University, Middletown, CT 06459

Communicated by Linda L. Randall, University of Missouri, Columbia, MO, April 4, 2005 (received for review January 12, 2005)

Abstract

SecA facilitates bacterial protein translocation by its association with presecretory or membrane proteins and the SecYEG translocon channel. Once assembled, SecA ATPase undergoes cycles of membrane insertion and retraction at SecYEG that drive protein translocation in a stepwise fashion. SecA exists in equilibrium between a monomer and dimer, and association with its translocation ligands shifts this equilibrium dramatically. Here, we examined the proposal that protein translocation can occur by means of a SecA monomer. We produced a mutant SecA protein lacking residues 2–11, which was found to exist mostly as a monomer, and it was unable to complement a conditional-lethal secA mutant, was inactive for in vitro protein translocation, and was poorly active for translocation ATPase activity. Furthermore, we developed a technique termed membrane trapping, where wild-type SecA subunits became trapped within the membrane by overproduction of membrane-stuck mutant SecA proteins, and, in one case, a membrane-associated SecA heterodimer was demonstrated. Finally, we examined both endogenous and reconstituted membrane-bound SecA and found a significant level of SecA dimer in both cases, as assessed by chemical crosslinking. Collectively, our results strongly suggest that membrane-bound SecA dimer is critical for the protein translocation cycle, although these results cannot exclude participation of SecA monomer at some stage in the translocation process. Our findings have important implications regarding SecA motor function and translocon assembly and activation.

Footnotes

↵ * To whom correspondence should be addressed. E-mail: doliver{at}wesleyan.edu.
Author contributions: L.B.J. and D.O. designed research; L.B.J. and C.R.Z. performed research; L.B.J. and D.O. analyzed data; and D.O. wrote the paper.
Abbreviations: EDAC, 1-ethyl-3-(3-dimethyl-aminopropyl)carbodiimide; IMV, inverted membrane vesicle; IPTG, isopropyl-β-d-thiogalactopyranoside.
Note. While we were writing up this study, three additional papers appeared on the topic. Or et al. (36) published an additional paper indicating that SecA may function as a monomer based on analysis of a SecA mutant Δ11/N95 (missing residues 2–11 and 832–901). Although it is difficult to strictly compare our results, given that somewhat different SecA proteins were used and SecA function was often augmented through use of the PrlA4 (SecY) mutant protein, we worry whether the recA ⁺ background or the transient depletion conditions used by these investigators may have resulted in the presence of contaminating wild-type SecA. However, both of us found that the Δ11 truncation substantially monomerizes SecA, and that the truncate (either SecAΔ11 or Δ11/N95) has substantially diminished protein translocation activity (e.g., see figures 3C and 6A of ref. 36), although we clearly differ on the extent of the defect and its interpretation. Our results are more comparable with Randall et al. (37), who found that SecAΔ11 is largely monomeric and inactive for in vitro protein translocation of either maltose-binding protein or galactose-binding protein precursors. Karamanou et al. (38) reported that SecA9-861 (missing residues 2–8 and 862–901) is both dimeric and functional in vivo and in vitro. From our vantage point, these results suggest that the level of SecA dimer drops significantly in going from SecAΔ8 to SecAΔ11, in agreement with measurements by Randall et al. (37) for these two proteins, and they support our conclusion that membrane-integrated SecA dimer plays an important role in the activity of this protein.