The FlhA linker mediates flagellar protein export switching during flagellar assembly

The flagellar protein export apparatus switches substrate specificity from hook-type to filament-type upon hook assembly completion, thereby initiating filament assembly at the hook tip. The C-terminal cytoplasmic domain of FlhA (FlhAC) serves as a docking platform for flagellar chaperones in complex with their cognate filament-type substrates. Interactions of the flexible linker of FlhA (FlhAL) with its nearest FlhAC subunit in the FlhAC ring is required for the substrate specificity switching. To address how FlhAL brings the order to flagellar assembly, we analyzed the flhA(E351A/W354A/D356A) ΔflgM mutant and found that this triple mutation in FlhAL increased the secretion level of hook protein by 5-fold, thereby increasing hook length. The crystal structure of FlhAC(E351A/D356A) showed that FlhAL bound to the chaperone-binding site of its neighboring subunit. We propose that the interaction of FlhAL with the chaperon-binding site of FlhAC suppresses filament-type protein export and facilitates hook-type protein export during hook assembly.

fullly rescue switching. As stated, it sounds like they cause the switching defect in the first place.
8. line 130. Why conformational rearrangements specifically? This didn't seem clear to me. Also, are there any known instances wehre a defect in switching interferes with filament assembly but does not also give rise to longer hooks? I.e. doesn't the process of switching specificity, so far as we know, do both things-shut down early export and initiate late export? 9. line 138. the triple, like the single and double mutations, reduces binding of chaperones. but the Trp residue, in the present model, is involved in blocking chaperone binding, right? So its loss might be expected to permit binding? Unless what happens in the triple mutant is that the relationship between adjacent FlhAc domains is altered, and this relationship is important for chaperone binding; i.e., binding is near the interface between subunits. line 139 introduces the idea of binding to the dimple, but it's not entirely clear how this follows from what's gone before.
10. If it's a conformational equilibrium in the mostly folded protein, then wouldn't two bands on SDS-PAGE (where the protein is largely unfolded) be unexpected? The idea seems to be that the relevant regions of the protein remain folded, and capable of the hypothesized interaction, even in SDS-PAGE. An alternative possibility is that a part of the protein binds SDS differently (more) in the mutants; this might involve a local partially folded conformation, but might not (and so might be more likely).
11. line 162. But the triple mutant still shows some of the compact conformation, even though the side chain of Trp 354 is changed to a methyl group. So the Trp is not needed.
12. The crystal structure shows an interaction in which a hydrophobic residue introduced by mutation (Ala 351) makes a hydrophobic contact. The alanine replaces a residue that is polar in the wild-type protein. Doesn't this argue that the interaction is an artifact of the hydrophobicityincreasing mutation, together with the crowded conditions in a crystal that practically force interactions to occur? It's difficult to see, from the views shown in fig. 4, what would be the effect of having residues 351 and 356 still being acidic (wild type), but it seems likely that they would interfere. It can also be noted that the introduced Ala residues will increase the helical propensity of this segment very appreciably. So even if the alanine at 356 doesn't interact directly at the hypothesized binding site, it could help stabilize the helical conformation that is involved in the binding seen in the crystal.
13. Looking at the structure, it seems that the spacing of the two domains is fairly large, and it emerges that while the interaction involving the linker occurs there, the other, better established interactions between adjacent subunits actually do not occur-i.e. the relative placement of these subunits is dictated by the crystal. This heightens the suspicion that the interaction involving the linker is not a native one.
Summary: My overall view is that the mutant phenotypes don't point convincingly toward the specific idea proposed here, and the interaction seen in the crystal structure could be an artifact of crystallization and the non-native hydrophobicity introduced in key positions. I feel that the study needs more direct evidence in support of the proposed interaction, in the flagellum rather than in a crystal. This would ideally include evidence that the interaction is lost upon switching from early to late export substrates.
Reviewer #3 (Remarks to the Author): Inoue et al. present a paper titled "The FlhA linker mediates flagellar protein export switching during flagellar assembly" in which they present genetic, biochemical and structural data to support a role of the linker between the membrane embedded N-terminal and cytoplasmic Cterminal domains of FlhA in regulating the switch between secretion of hook-type to filament-type components.
Several FlhAL mutants impact flagellar filament assembly including E351A/D356A and W354A, which produce HBBs without filaments and hook length not controlled properly, and E351A/W354A/D356A which does not produce HBBs at all. The authors initially identified some genetic pseudorevertants that restore some level of motility to a flhA(E351A/W354A/D356A) mutant. These commonly increased cellular levels of other components involved in secretion and it was interpreted that this bypasses the reduced affinity between FlhAc(E351A/W354A/D356A) and FliJ (a published observation), which is present at higher intracellular concentration. They go on to characterize several mutants -the triple mutant, E351A/D356A and W354A -with gel-filtration, gel-mobility shifts, methoxypolyethylene modification and an X-ray structure. The gel-filtration results show a small shift in elution volume for the mutants which is interpreted to suggest the linker mutants result in a more compact monomeric structure due to binding of the mutated linker to the hydrophobic dimple in a cis manner. This binding site is also shared with flagellar chaperones so linker binding effectively inhibits chaperone binding. A more compact nature is also proposed based on mobility in SDS-PAGE gels with the mutants showing faster migration. Ultimately, the determined structure of E351A/D356A does show the linker of one monomer in the crystal asymmetric unit bound to a neighbouring "hydrophobic dimple" similar to the interaction observed for flagellar chaperones; however the relative disposition the two molecules in the presented crystal is different to that in the nomomeric ring. This is accounted for by the flexibility of the linker.
Overall, the manuscript is clearly written and presents an interesting structure capturing the FlhA linker bound to a neighbouring monomer, and model that offers a plausible explanation of the genetic data on hook length control. However, the accompanying biochemical and biophysical data does not convincingly support the conclusions drawn. The gel-filtration results and interpretation would benefit from further experiments to support the conclusions reached. For example, all samples should be run at the same concentration. FlhA homologues undergo transient selfassociation that can result in a concentration dependent change in elution profile with gelfiltration. Further, if the mutants affect self-association, which the linker is involved in, then this combination of different concentrations and mutants could account for the changes in elution position. Light scattering experiments (SEC-MALS and DLS) might provide a way to confirm the monomeric state and hydrodynamic properties of these constructs in solution. The different migration of the mutants in SDS-PAGE is surprising given these are denaturing gels. Running both boiled and unboiled samples might be further informative. Similarly the double band in the SDS-PAGE of the triple mutant (E351A/W354A/D356A), which doesn't appear to be present in the corresponding double Cys mutant used for the mPEG-maleimide modification, should be expected to disappear with boiling prior to SDS-PAGE if two different structural states. It is also somewhat confusing why the gel-filtration data is presented as suggesting an interaction of the linker with self, while the structure and model present an interaction with neighbouring molecule in the FlhA nonomer, this is briefly referenced in the very last sentence. The manuscript would additionally benefit from some expanded discussion on the respective interactions of the chaperones/FliJ with FlhA, which contribute to the temporal and specific regulation of substrate secretion, especially in light of the recent structure the complex between Chlmydia SctV and SctO.
Given the broad interest and novelty of the structure, the study merits publication in Communications Biology providing the outlined issues with the supporting biochemical data are addressed. Table 1: Data looks like it extended further than presented eg CC1/2 in highest resolution bin is 0.915 and I/sig 2.8? Figure 2: Isolation of pseudorevertants mPEG modification -seems much more than 5kDa, is this just a change in gel migration?

Additional points
Our responses are listed below. We highlighted all changes in red in the revised manuscript (Marked Up version).

To Reviewer #1
Inoue et al presented detailed analyses of the flexible linker of the major flagellar export apparatus protein FlhA (FlhAL) and its potential function in flagellar assembly. Given that FlhA is probably the most important protein in flagellar assembly and the FlhAL is the bridge between the cytoplasmic domain and the transmembrane domain and remains poorly understood, the studies presented in the manuscript will be valuable for the understanding of this important protein and the whole export apparatus. The major concerns are that the manuscript and the experimental results were poorly presented and written. In current format, it's very difficult to differentiate the results from the background information and the interpretation.
Re: Thank you so much for your comments. We agree that the background information for the experimental design may have been rather limited for readers to fully understand the purpose of this study. We added necessary background information in the Introduction and Results. We also tried to make the description of results clearly differentiated from the background information and interpretation. We hope the presentation of the results are sufficiently clearer in the revised manuscript. Additional experiments are needed if the authors want to make these claims.
Re: We carried out Bio-Layer Interferometry (BLI) measurements to understand why the FlhA linker with either ED or EWD mutation inhibits the binding of the FlgN chaperone to the flagellar chaperone-binding site of FlhAC (Please see new Figure 7 of the revised manuscript). We found that these mutations significantly affect the docking process of FlhAC to immobilized GST-FlgN. We also found that complete deletion of the FlhA linker region (FlhAL) increases the binding affinity of FlhAC for FlgN significantly compared to FlhAC-ED and FlhAC-EWD. These results suggest that FlhAL with these mutations inhibits the binding of FlgN to the chaperone-binding site of FlhAC.
Furthermore, we found that the binding affinity of the linker deletion mutant for FlgN was 30-fold lower than that of wild-type FlhAC, suggesting that FlhAL is required to keep Re: We modified it to make non-expert readers readable.

To Reviewer #2
This paper attempts to show that a linker segment of FlhA interacts with a hydrophobic region of an adjacent FlhA subunit (a region known to be involved in the binding of chaperones for late-cargo export), and by doing so interferes with export of late cargoes. It is hypothesized that this interaction plays a key role in regulating the switch between early and late export cargoes. The proposal is based on mutant phenotypes, and on a crystal structure of a mutant variant of the protein. The idea is interesting and might account for the results, but overall, I find the case to be not very persuasive.

Specific issues are detailed below.
Re: Thank you so much for all your comments and suggestions to improve our manuscript.

In the abstract, an important mutant is mentioned, and the mutational changes given.
But what is the phenotype? Is this a locked-early mutant, for example? We also modified our description to make our proposal clear as follows: "The interaction of FliJ with FlhAL is required for activation of the fT3SS, and FliH and Re: An interaction between FliK and FlhB triggers export switching of the fT3SS to shut down early export and initiate late export, thereby terminating hook assembly and initiating filament formation at the hook tip. So, many fliK mutants and certain flhB point mutants produce polyhooks without the filament attached. We have shown that hook lengths of the flhAED and flhAW mutants are 54.0 ± 22.3 nm (mean ± SD) and 52.9 ± 19.9 nm, respectively, where their SD values are larger than the wild-type one (51.0 ± 6.9 nm). Thus, their hook length is nearly the wild-type one although not controlled  Adv. 2018). These observations suggest that conformational rearrangements of FlhAC is required for the initiation of late export. To make it clearer, we modified our proposal as follows: "Because high-speed atomic force microscopy has shown that the flhAEWD mutation also inhibits highly cooperative FlhAC ring formation 10 , we propose that FlhAL regulates the conformational rearrangement of FlhAC, which is required for efficient termination of hook assembly and efficient initiation of filament formation at the hook tip."

If it's a conformational equilibrium in the mostly folded protein, then wouldn't two bands on SDS-PAGE (where the protein is largely unfolded) be unexpected? The idea seems to be that the relevant regions of the protein remain folded, and capable of the hypothesized interaction, even in SDS-PAGE. An alternative possibility is that a part
of the protein binds SDS differently (more) in the mutants; this might involve a local partially folded conformation, but might not (and so might be more likely).
Re: Although we really do not know why FlhAC-ED showed a slightly faster mobility in SDS-PAGE gels and why FlhAC-EWD showed two different bands on SDS-PAGE gels, with the slower mobility band corresponding to wild-type FlhAC and the faster one corresponding to FlhAC-ED, we assume that a strong hydrophobic interaction between FlhAL with the ED mutation and the hydrophobic dimple may allow this contact region to partially remain folded even in the SDS-PAGE and that the EWD triple mutation presumably weakens this strong hydrophobic interaction between FlhAL and the dimple, thereby causing two different bands on SDS-PAGE gels. We also agree with your alternative possibility.

line 162. But the triple mutant still shows some of the compact conformation, even though the side chain of Trp 354 is changed to a methyl group. So the Trp is not needed.
Re: Agreed and deleted this sentence. fig. 4, what would be the effect of having residues 351 and 356 still being acidic (wild type), but it seems likely that they would interfere. It can also be noted that the introduced Ala residues will increase the helical propensity of this segment very appreciably. So even if the alanine at 356 doesn't interact directly at the hypothesized binding site, it could help stabilize the helical conformation that is involved in the binding seen in the crystal.

The crystal structure shows an interaction in which a hydrophobic residue introduced by mutation (Ala 351) makes a hydrophobic contact. The alanine replaces a residue that is polar in the wild-type protein. Doesn't this argue that the interaction is an artifact of the hydrophobicity-increasing mutation, together with the crowded conditions in a crystal that practically force interactions to occur? It's difficult to see, from the views shown in
Re: We agree with this reviewer that the introduced Ala residues increase the helical propensity of residues 349-357of FlhAL and could help stabilize its helical conformation involved in the binding in the crystal. But the binding of a tryptophane residue to a hydrophobic pocket is strong enough to be typically used for specific binding of many protein partners. Also, when Ala-351 and Ala-356 of FlhAC-ED in the Rep. 2018), suggesting that this chaperone-binding site is also involved in the export of hook-type substrates. These genetic and biochemical data strongly suggest that the interaction between FlhAL and the chaperone-binding site of FlhAC coordinates the export of hook-type proteins with hook assembly in a highly organized and wellcontrolled manner. Therefore, we think that the interaction between FlhAL and the chaperone-binding site of FlhAC we observed in the crystal structure would reflect a functional state of the FlhAC ring during HBB assembly rather than an artifact of crystallization.

Summary: My overall view is that the mutant phenotypes don't point convincingly
toward the specific idea proposed here, and the interaction seen in the crystal structure could be an artifact of crystallization and the non-native hydrophobicity introduced in key positions. I feel that the study needs more direct evidence in support of the proposed interaction, in the flagellum rather than in a crystal. This would ideally include evidence that the interaction is lost upon switching from early to late export substrates.
Re: As mentioned just above, all our additional data strongly suggest that the interaction of FlhAL with the chaperone-binding site reflects the functional state of Re: Thank you so much for your comments. Our SEC-MALS is quite old and unfortunately did not work at all. So, we could not carry out SEC-MALS. Xing et al.
showed that FlhAC forms dimer in a protein concentration dependent manner and that FlhAL is required for its dimerization (Xing et al., Nat. Commun. 2018). They showed that FlhAC elutes as dimer at a protein concentration of 0.2 mM but as monomer at a protein concentration of 0.05 mM under a physiological salt concentration of 100 mM.
They also showed that FlhAC without FlhAL elutes as monomer even at a high protein concentration. Therefore, we carried out analytical size exclusion chromatography again at the same protein concentration of 10 μM in the presence of 150 mM NaCl. We also used BSA (66.4 kDa) as a marker. New SEC data clearly showed that they elute as monomer (Please see new Fig. 5a of the revised manuscript). Re: We carried out native PAGE and found that FlhAC-ED showed a slightly faster mobility even in native gels (Please see New Figure 5c of the revised manuscript).

The different migration of the mutants in SDS-
Because the F459C substitution is located in the chaperone-binding site of FlhAC, we assume that the F459C substitution may stabilize a compact conformation of FlhAC-EWD so that the upper band is gone.
It is also somewhat confusing why the gel-filtration data is presented as suggesting an interaction of the linker with self, while the structure and model present an interaction with neighbouring molecule in the FlhA nonomer, this is briefly referenced in the very last sentence.
Re: We modified the last sentence as follows to make it clearer: "Because FlhAC-ED monomer adopts a more compact conformation compared with the wild-type FlhAC monomer as judged by SEC (Fig. 5), FlhAL may bind to FlhAC in a cis manner as well. Therefore, it is also possible that FlhAL may block the docking of the flagellar chaperones to FlhAC by covering the binding site of the same FlhAC molecule." The manuscript would additionally benefit from some expanded discussion on the respective interactions of the chaperones/FliJ with FlhA, which contribute to the temporal and specific regulation of substrate secretion, especially in light of the recent structure the complex between Chlmydia SctV and SctO.
Re: Thank you so much for your comments. We analyzed the interaction of FlhAC with FliJ by BLI measurements and confirmed that FlhAL is required for the interaction with FliJ. We added new data to the revised manuscript to discuss the FliJ-FlhAC interaction as follows: "Effect of deletion of FlhAL on the interaction between FlhAC and FliJ. The crystal structure of a FliJ homologue, CdsO, in complex with CdsVC, which is a FlhAC homologue, has shown that CdsO binds to a large cleft between domains D4 of neighboring CdsVC subunits in the CdsVC ring structure but not to the linker region of CdsVC 34 . To confirm the importance of FlhAL in the interaction between FlhAC and FliJ, we analyzed the binding of FlhAC to immobilized GST-FliJ by Bio-Layer Interferometry (BLI) measurements 35 . The FliJ-FlhAC interaction showed a complex binding profile ( Fig. 3a) and did not fit the global one-state association-then-dissociation model.
Assuming that FlhAC binds to GST-FliJ to form a GST-FliJ/FlhAC complex, followed by a conformational change of this complex, the BLI data fitted well with a two-state reaction model and provided a KD value of about 1.3 μM (Fig. 3b). Unlike wild-type FlhAC, the association and dissociation processes of FlhAC with the flhAEWD mutation (FlhAC-EWD) or FlhAC lacking FlhAL (FlhAC-L) were observed at protein concentrations above 10 μM (Fig. 3a). Their association and dissociation processes were also different from those of wild-type FlhAC. The association profile of these mutant proteins was composed of two distinct (fast-on and slow-on) processes, and the dissociation profile was also composed of two distinct (fast-off and slow-off) processes. It has been shown that wild-type FlhAC forms dimer in a protein concentration dependent manner and that FlhAL is required for efficient dimerization of FlhAC 27 . So, their BLI data were fitted well with curves predicted by the Hill equation, with KD values of about 60 μM and 49 μM for the FliJ-FlhAC-EWD and FliJ-FlhAC-L interactions, respectively (Fig. 3b). Thus, both flhAEWD mutation and deletion of FlhAL significantly reduced the binding affinity of FlhAC for FliJ. Therefore, we conclude that FlhAL is required for stable interaction between FliJ and FlhAC." Given the broad interest and novelty of the structure, the study merits publication in Communications Biology providing the outlined issues with the supporting biochemical data are addressed.
Re: Thank you so much for all your supportive and constructive comments. Re: We analyzed our X-ray data at higher resolution than 2.8 Å, but CC(1/2) and I/σI deteriorated and so we determined the FlhAC-ED structure at 2.8 Å mPEG modification -seems much more than 5kDa, is this just a change in gel migration?