Figure 4. BofA TMS2 is near the SpoIVFB active site.

(A) Disulfide cross-linking of E44C at the cytTM-SpoIVFB active site to C46 in TMS2 of MBPΔ27BofA. pET Quartet plasmids were used to produce single-Cys E44C cytTM-SpoIVFB in combination with MBPΔ27BofA C46 (pSO91) or Cys-less MBPΔ27BofA C46S as a negative control (pSO110), and Cys-less variants of SpoIVFA and Pro-σ^K(1–127) in Escherichia coli. Samples collected after 2 hr of IPTG induction were treated for 60 min with Cu²⁺(phenanthroline)₃ (Cu+) to promote disulfide bond formation or with 2-phenanthroline (Cu–) as a negative control, then treated with TCA to precipitate proteins and resuspended in sample buffer with DTT (+) to reverse cross-links or without (–) to preserve cross-links, and finally subjected to immunoblot analysis with FLAG antibodies to visualize cytTM-SpoIVFB monomer, dimer, and complex with MBPΔ27BofA. (B) Disulfide cross-linking of V70C or P135C near the cytTM-SpoIVFB active site to C46 in TMS2 MBPΔ27BofA. pET Quartet plasmids were used to produce single-Cys V70C or P135C cytTM-SpoIVFB E44Q variants in combination with MBPΔ27BofA (pSO92 and pSO93) or Cys-less MBPΔ27BofA C46S as a negative control (pSO111 and pSO112), and Cys-less variants of SpoIVFA and Pro-σ^K(1–127) in E. coli. Samples collected after 2 hr of IPTG induction were treated and subjected to immunoblot analysis as in (A). A representative result from at least two biological replicates is shown in (A) and (B).

Figure 4—figure supplement 1. Models of SpoIVFB and BofA TMS2.

(A) Model of SpoIVFB. At Left, a side view of a SpoIVFB monomer. The model shows the six TMSs of the SpoIVFB membrane domain, the zinc ion (magenta) involved in catalysis, the interdomain linker, and the CBS domain. In the enlarged view of the active site cleft (Center), TMSs 1–6 and residues 44, 70, and 135 of SpoIVFB are labeled. At Right, a top view is shown. (B) Model of SpoIVFB with BofA TMS2. Labeling is as in (A) and BofA TMS2 (cyan) is modeled in the SpoIVFB active site cleft. The enlarged view of the active site cleft depicts experimentally observed disulfide cross-links (dashed lines) between BofA C46 and both E44C (in TMS2) and P135C (in a short loop in TMS4) of single-Cys cytTM-SpoIVFB variants (Figure 4).

Figure 4—figure supplement 2. Cys-less variants of SpoIVFA and MBPΔ27BofA inhibit cleavage of Pro-σ^K(1–127) by cytTM-SpoIVFB in Escherichia coli.

Pro-σ^K(1–127) and cytTM-SpoIVFB were produced from pYZ2 as a control (lane 1), or pET Quartet plasmids were used to produce Pro-σ^K(1–127), cytTM-SpoIVFB, and either SpoIVFA and GFPΔ27BofA from pSO40 as another control (lane 2), Cys-less SpoIVFA and MBPΔ27BofA from pSO90 (lane 3), or Cys-less SpoIVFA and Cys-less MBPΔ27BofA from pSO97 (lane 4). Samples collected after 2 hr of IPTG induction were subjected to immunoblot analysis with SpoIVFA, GFP, or penta-His antibodies as indicated. The single star (*) indicates cross-reacting proteins below SpoIVFA. The double (**) and triple (***) stars indicate breakdown species of GFPΔ27BofA and MBPΔ27BofA, respectively. A breakdown species below SpoIVFA (not indicated) is observed in some samples. The graph shows quantification of the cleavage ratio, as explained in the Figure 1B legend.

Figure 4—figure supplement 3. BofA TMS2 is in proximity to the active site of SpoIVFB.

Disulfide cross-linking of single-Cys E44C cytTM-SpoIVFB to MBPΔ27BofA C46. pET Quartet plasmids were used to produce single-Cys E44C cytTM-SpoIVFB (pSO91) or Cys-less cytTM-SpoIVFB E44Q as a negative control (pSO94) in combination with MBPΔ27BofA, and Cys-less variants of SpoIVFA and Pro-σ^K(1–127) in Escherichia coli. Samples collected after 2 hr of IPTG induction were treated as explained in the Figure 4 legend and subjected to immunoblot analysis with MBP antibodies to visualize MBPΔ27BofA monomer, dimer, and complex with cytTM-SpoIVFB. A representative result from at least two biological replicates is shown. The star indicates the complex in lane 2. The immunoblot images (raw and annotated) are in the Figure 4—source data 1 folder, with explanation on the annotated Figure 4A image.

Figure 4—figure supplement 4. Full-length BofA is in proximity to the active site of SpoIVFB.

Disulfide cross-linking of single-Cys cytTM-SpoIVFB variants to BofA C46. pET Quartet plasmids (pSO226–pSO231) were used to produce single-Cys E44C cytTM-SpoIVFB, or single-Cys V70C or P135C cytTM-SpoIVFB E44Q variants, in combination with BofA or BofA C46S, and Cys-less variants of SpoIVFA and Pro-σ^K(1–127) in Escherichia coli. Samples collected after 2 hr of IPTG induction were treated as explained in the Figure 4 legend and subjected to immunoblot analysis with FLAG antibodies to visualize cytTM-SpoIVFB monomer, dimer, and complex with full-length BofA. The star (*) indicates likely nonspecific cross-linking of single-Cys cytTM-SpoIVFB variants to E. coli proteins. A representative result from at least two biological replicates is shown.

Figure 4—figure supplement 5. BofA TMS2 has a preferred orientation in the active site cleft of SpoIVFB.

(A) Enlarged view of the model of SpoIVFB and BofA TMS2 shown in Figure 4—figure supplement 1B, which is based in part on disulfide cross-linked complexes shown in (C). At Left, the side view of SpoIVFB TMS2 and TMS6 (green) is shown with the zinc ion (magenta) involved in catalysis and BofA TMS2 (cyan). This view depicts experimentally observed cross-links (dashed lines) between BofA I56C and SpoIVFB A32C, and between BofA H57C and SpoIVFB Q181C. In the bottom view (Center), BofA TMS2 is shown with SpoIVFB TMS3. The dashed line indicates a cross-link between BofA A41C and SpoIVFB V86C that was observed. At Right, the top view of the model is shown. The dashed lines indicate observed cross-links between SpoIVFB residue M30C (located in the loop connecting TMS1 and TMS2) and BofA L62C and V63C (in a loop near the C-terminal end of TMS2). (B) Cleavage assays examining the effects of Cys substitutions for residues of interest in cytTM-SpoIVFB or GFPΔ27BofA. BofA TMS2 was modeled in the SpoIVFB active site cleft based on our initial cross-linking results (Figure 4). The initial model predicted proximity between residues at or near the ends of BofA TMS2 and residues of SpoIVFB, thus identifying residues of interest for cross-linking experiments. First, we examined the effects of Cys substitutions for the residues of interest using cleavage assays. pET Duet plasmids were used to produce Pro-σ^K(1–127) in combination with cytTM-SpoIVFB from pYZ2 as a control (lane 1) or with the indicated Cys-substituted cytTM-SpoIVFB from pSO141 or pSO256-pSO259 in E. coli (Left). pET Quartet plasmids were used to produce Pro-σ^K(1–127), cytTM-SpoIVFB, and SpoIVFA in combination with GFPΔ27BofA from pSO40 as a control (lane 7) or with the indicated Cys-substituted GFPΔ27BofA from pSO142, pSO143, or pSO260-pSO263 in Escherichia coli (Right). Samples collected after 2 hr of IPTG induction were subjected to immunoblot analysis, and the graph shows quantification of the cleavage ratio, as explained in the Figure 1B legend. Since cytTM-SpoIVFB with Cys substitutions for residues of interest cleaved Pro-σ^K(1–127) (lanes 2–6), we included the inactivating E44Q substitution in the single-Cys cytTM-SpoIVFB variants created for cross-linking. GFPΔ27BofA with Cys substitutions for residues of interest inhibited Pro-σ^K(1–127) cleavage by cytTM-SpoIVFB, although the G40C and H57C substitutions caused partial loss of inhibition, and the G40C substitution resulted in less accumulation of all four proteins (lanes 8–13). Ala substitutions at these positions had similar effects (Figure 2, lanes 2 and 15). (C) Disulfide cross-linking of single-Cys cytTM-SpoIVFB variants to single-Cys MBPΔ27BofA variants. pET Quartet plasmids (pSO93 as a positive control in lanes 1 and 2, pSO147, pSO148, and pSO186–pSO190) were used to produce single-Cys cytTM-SpoIVFB E44Q variants in combination with single-Cys MBPΔ27BofA variants or Cys-less MBPΔ27BofA from pSO144 as a negative control, and Cys-less variants of SpoIVFA and Pro-σ^K(1–127) in E. coli. Samples collected after 2 hr of IPTG induction were treated and subjected to immunoblot analysis as explained in the Figure 4A legend. A representative result from two biological replicates is shown. In agreement with the model shown in (A), I56C and H57C MBPΔ27BofA variants formed a cross-linked complex with A32C and Q181C cytTM-SpoIVFB variants, respectively (lanes 4 and 13). The I56C MBPΔ27BofA variant formed very little complex with the L33C cytTM-SpoIVFB variant (lane 10), suggesting that a preferred orientation of BofA TMS2 places I56C farther from L33C than from A32C in TMS2 of SpoIVFB. Similarly, comparison of complex formation by G40C and A41C MBPΔ27BofA variants with the V86C cytTM-SpoIVFB variant (lanes 7 and 16) suggested that a preferred orientation of BofA TMS2 places G40C farther than A41C from V86C near the C-terminal end of SpoIVFB TMS3. Likewise, L62C and V63C MBPΔ27BofA variants formed a cross-linked complex with the M30C cytTM-SpoIVFB variant (lanes 19 and 22), suggesting a loop near the C-terminal end of BofA TMS2 is in proximity to a loop between SpoIVFB TMS1 and TMS2.