Figure 4. Initial stages of the Mfd loading pathway.

Color coding of RNA polymerase (RNAP) subunits and Mfd domains are shown in the keys on the left and right, respectively. (A) Apo-Mfd PDB 2EYQ (Deaconescu et al., 2006) combines with an elongation complex (EC) [PDB 6ALF (Kang et al., 2017) with upstream and downstream duplex DNA extended] to form a putative initial encounter complex [L0], which was modeled by superimposing apo-Mfd D4(RID) onto the Mfd_L1-D4(RID) and adjusting the trajectory of the upstream duplex DNA. (C) The [L0] → L1(atp) transition is shown. In this view, the downstream duplex DNA (and the direction of transcription) points away from the viewer. (Top) The Mfd-EC structures are shown as molecular surfaces with DNA shown in cartoon format. The boxed regions are magnified below. (Bottom) Mfd is shown as a transparent molecular surface surrounding a backbone ribbon. In the middle, the colored spheres denote the relative positions of the Mfd domain center-of-masses (com), with connecting lines denoting the motions from the L[0] → L1(atp) transition (the translations of the com's, as well as the relative rotation of the domains, are listed. The D4(RID) motion is negligible; also see Supplementary file 4). D.[L0]. E.L1(atp): The region of the upstream duplex DNA colored orange and denoted by the orange stripe was found to be required for Mfd function on an EC (Park et al., 2002). (F) View of the [L0] → L1(atp) → L2(adp) transition, highlighting the structural changes in the Mfd relay helix (RH) and hook helices (HH). In this view, the Mfd-EC complex is rotated ~180° about a horizontal axis, so the downstream duplex DNA (and direction of transcription) is toward the viewer. (Top) The RNAP is shown as a molecular surface, with nucleic acids shown in cartoon format. Mfd is shown with cylindrical helices. Color coding is as above but the RH is colored hot pink, and the HHs are colored dark green. The boxed region is magnified below. (Bottom) The complexes are shown in faded colors except for the RH and HHs. Also shown as a molecular surface are the residues of Mfd-D2 that interact with UvrA [determined from PDB 4DFC (Deaconescu et al., 2012)]. D.L[0]: The RH at the very N-terminus of TD1 extends for 30 residues and is surrounded by the HHs at the very C-terminus of TD2. The UvrA-interacting surface of Mfd-D2 is occluded by D7 (Deaconescu et al., 2006). E.L1(atp): The middle portion of the RH helix unfolds and the RH kinks about 112° around the second HH due to the translation/rotation of TD1 (denoted) and also TD2. The UvrA-interacting surface of Mfd-D2 is still occluded by D7. F.L2(adp): The transition from L1(atp) → L2(adp) involves a 259° rotation of TD1 around the backside of the DNA, as well as a 65 Å translation toward the RNAP (denoted). This is likely accomplished by ATP-hydrolysis-dependent walking of the Mfd translocation module and D7 along the DNA until it bumps into the RNAP. The corkscrewing translocation module unfolds the N-terminal half of the RH, wrapping it around the DNA as it goes. In this process D2 is separated from D7 but the UvrA-interacting surface of D2 is now occluded by the DNA. Also see Figure 2—video 1.

Figure 4—figure supplement 1. Putative structural pathway for Mfd activity and Mfd/DNA interactions.

(A) Putative ordered pathway of seven Mfd-elongation complex (EC) structures (Figure 2B–H). Each structure is shown in its position along with the number of particles that gave rise to that structural class (Figure 2—figure supplement 1). All the structures came from identically prepared samples, so the particle numbers are likely related to the stability of each complex. L1(atp) was placed first in the pathway because the Mfd component [Mfd(atp)_L1 most closely superimposes with apo-Mfd (Supplementary file 4)]. The structures can be grouped into two main groups, L1(atp) and L2(adp), which do not superimpose well with any of the other structures, and C1(ATP), C2(ATP), C3(adp), C4(ADP), and C5(ATP), which are all relatively similar to each other (Supplementary file 4, Figure 2—figure supplement 3A). We thus place L1(atp) and L2(adp) in a loading pathway (also see Figure 2—video 1), while C1–C5 represent the fully EC-engaged nucleotide hydrolysis cycle (NHC) for Mfd (see Figure 7—video 1). As described in the text, the complete loading pathway requires a minimum of 10 ATP hydrolysis events, while each cycle of the NHC requires one ATP hydrolysis. (B) (Top) The nucleic acid scaffold is shown (same as Figure 1A except the upstream single-stranded RNA is not shown for clarity). Above the sequences, the orange bar denotes the extent of upstream duplex DNA required for Mfd function (Park et al., 2002). The gray bars below denote the extent of Mfd/DNA interactions in the seven Mfd-EC structures (light gray bar, nt-strand interacts; dark gray bar, t-strand). The interactions of Mfd(atp)_L1 with the DNA explain the requirement for ~40 base pairs of upstream duplex DNA (Park and Roberts, 2006) and indicate that L1(atp) is an obligate intermediate in the pathway.