Crystallization and preliminary X-ray diffraction analysis of YejM from Salmonella typhimurium: an essential inner membrane protein involved in outer membrane directed cardiolipin transport

Salmonella typhimurium is responsible for over 35% of all foodborne illness related hospitalizations in the United States. This Gram-negative bacterium possesses an inner and an outer membrane (OM), the latter allowing its survival and replication within host tissues. During infection, OM is remodeled by transport of glycerophospholipids across the periplasm and into the OM. Increased levels of cardiolipin in the OM were observed upon PhoPQ activation and led to the discovery of YejM; an inner membrane protein essential for cell growth involved in cardiolipin binding and transport to the OM. Here we report how YejM was engineered to facilitate crystal growth and X-ray diffraction analysis. Successful structure determination of YejM will help us understand how they interact and how YejM facilitates cardiolipin transport to the OM. Ultimately, yejm, being an essential gene, may lead to new drug targets inhibiting the pathogenic properties of S. typhimurium.

Although being an interesting possibility, we decided that with the current body of work, our suggested interaction between YejL and YejM, is too speculative. For the purpose of transparency, we provide all additional body of work that we performed in response to the reviewers' critique in Supplementary File 1. We hope that this in-detail reply clarifies our decision and motivation to have made this revision of our manuscript. We have also revised Whitfield & Trent, 2014); however, the transport of glycerophospholipids across the periplasm and insertion into the inner-leaflet of the OM is not well understood.
Interestingly, increased levels of cardiolipin in S. typhimurium OM were observed upon PhoPQ regulator activation (Dalebroux et al., 2014). Recently the inner membrane protein YejM (also known as PbgA) was shown to bind cardiolipin and be involved in OM formation (Dalebroux et al., 2015). Furthermore, YejM is known to be an essential gene in E. coli and was shown to be involved in intrinsic multidrug resistance (De Lay & Cronan, 2008;Duo et al., 2008). YejM is comprised of 586 amino acids forming five predicted N-terminal transmembrane helices, followed by an arginine-rich periplasmic random coil linker region, and a C-terminal periplasmic domain ( Figure 1A), and it was shown that YejM associates as a tetramer in solution (Dalebroux et al., 2015). The arginine-rich linker region and periplasmic globular domain of YejM were demonstrated to bind cardiolipin and are required for OM remodeling and cell growth (Dalebroux et al., 2015). The molecular mechanism of the interplay between PhoPQ system and YejM, and how cardiolipin molecules are transported to the OM, need further structural and functional investigation.
Here we report crystal growth and present successful crystallization conditions for full-length YejM from Salmonella typhimurium using lipidic cubic phase crystallization and how the periplasmic domain of YejM was engineered to facilitate crystallization, and present preliminary X-ray diffraction analysis of the construct YejM241-586. Future successful structure determination of YejM will help us to understand cardiolipin transport to the OM and may lead to new drug targets inhibiting the pathogenic properties of S. typhimurium.

Results
Full-length YejM crystallized using lipidic cubic phase Full-length YejM was expressed and purified ( Figure 1B, first lane) as described in (Dalebroux et al., 2015), and concentrated to 15 mg/ml in the presence of 0.01% DDM. Micro crystals appeared in many conditions after two weeks. Figure 1C shows examples of LCP boli with micro crystals as judged by UV microscopy. Further work on full-length YejM will be aimed to optimize current crystal conditions and optimize detergents for LCP and regular vapor diffusion crystallization.
Specific construct development was needed to achieve YejM crystal growth We expressed and purified periplasmic construct YejM191-586 as described in (Dalebroux et al., 2015). The original construct YejM191-586 failed to crystallize and showed a degradation product after electrophoresis in the SDS polyacrylamide gel ( Figure 1B, second lane). To prevent degradation, reduce flexible protein parts and remove positively charged arginine clusters, and increase the chance for crystal growth, we deleted the linker region A191 to E240 in the YejM191-586 construct, resulting in YejM241-586 ( Figure 1B, third lane). Both constructs eluted as well-defined peaks when applied to size exclusion chromatography (SEC) ( Figure 1D). A clear shift in size between both constructs was observed in SEC when performed timely after affinity chromatography. When compared to the standards, both constructs eluted at later volumes than expected for their size ( Figure 1D). Initial crystals of YejM241-586 appeared after one week incubation at 18ºC under different conditions; e.g.  early hit crystals diffracted poorly, only up to 6Å and optimization and up-scaling of the crystallization setup from a 96 well format to a 24 well format did not improve the diffraction quality. Further screening using (Hampton PEGRx HT screen, Hampton Research) resulted in new crystal forms grown in condition C6 (0.1 M HEPES pH 7.5, 12% w/v polyethylene glycol 3,350) and C4 (0.1 M Citric acid pH 3.5, 25% w/v polyethylene glycol 3,350). Screening around these two conditions led to crystals in a condition consisting of (0.1 M citric acid pH 4, 18% w/v polyethylene glycol 3,350 ( Figure 1C). YejM241-586 crystals grown in condition C6 diffracted well, up to 1.6 Å ( Figure 1E). Data indexing and scaling with XDS (Kabsch, 2010) and further analysis with AIMLESS (Evans, 2006) resulted in a data set up to 1.8Å resolution and good overall statistics (Table 1).

Conclusions
Here we report successful crystallization using protein engineered specifically to enable crystal growth of YejM.
Our initial X-ray data analysis of YejM241-586 crystals suggests a dimer assembly of the periplasmic domain, therefore the membrane domain is very likely needed to form the YejM tetramer.
We also aim to solve the crystal structures of YejM alone and with bound cardiolipin. Ultimately these structures will help in understanding where and how cardiolipin binds to YejM, and whether YejM's architecture is that of a transporter or channel.

Competing interests
No competing interests were disclosed.

Grant information
The author(s) declared that no grants were involved in supporting this work. 1.

2.
3. . reported the structure of YejL and preliminary X-ray crystallographic et al characterization of the periplasmic domain (Residues 241-586) of YejM, an inner membrane protein potentially responsible for transport of glycerophospholipid cardiolipin. The authors also tested interaction between YejL and YejM using blue native gel electrophoresis and size exclusion chromatography. In general, it is an interesting study, yet the following issues need to be addressed before accepted for indexation:

Open Peer Review
It is not proper to claim that YejL is a ligand of YejM. "Ligand" is dedicated for "receptors" only. Also, there is no strong genetic evidence to show that YejL is involved in cardiolipin transport.
The authors claimed that YejM is a dimer based on its periplasmic domain, and probably, also because of YejL forming a dimer. This is not clear as the authors did not study the oligomerization state of the full-length YejM . Other than this, when running SEC on superpose 6, the authors did not include a standard marker for reference. It ishard to judge if the periplasmic domain of YejM forms a dimer in solution.
The author believed that the negative charge surfaces of YejL might bind with the arginine-rich loop of YejM. Why the fragment of YejM (241-586) that seems to not include the arginine-rich loop still binds YejL ? Please explain.
No competing interests were disclosed. Competing Interests:

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Author Response 28 Nov 2017 , Indiana University at Bloomington Susanne Ressl 1.) We apologize for the incorrect use of terminology and would rather term YejL as potential interaction partner of YejM.
2.) We carefully checked within the manuscript what section/sentence could have caused this reviewer's understanding that we claimed YejM to be a dimer. We found the following sentence in 1.
reviewer's understanding that we claimed YejM to be a dimer. We found the following sentence in the "Conclusions" section of our originally submitted manuscript: "Our initial X-ray data analysis of YejM241-586 crystals suggests a dimer assembly of the periplasmic domain, therefore the membrane domain is very likely needed to form the YejM tetramer". In this sentence, we suggested the existence of a periplasmic domain YejM241-586 dimer based on our analysis of the crystallographic unit. In this analysis the content of the asymmetric unit that builds the crystal resulted in a dimer of YejM 241-586, and therefore we proposed it to be the likely physiologically relevant state of the periplasmic domain alone. From the SEC-MALS studies in (Dalebroux 2015) that full-length YejM forms a tetramer when purified with the detergent et al. DDM. Therefore we reasoned in the above stated sentence that further oligomerization into a tetramer must be facilitated by the trans-membrane domain. Whether is exists as a dimer or a monomer either in the native membrane, in the presence of specific lipids or purified using a different detergent is a matter of current investigation by other laboratories (Miller and Dalebroux labs).
3.) Indeed, the arginine-rich linker region of YejM would be the most obvious binding region for YejL. However, we cannot exclude the possibility that YejL binds to solvent-accessible positively charged residues on the YejM241-586 surface.
of YejM. It is not predicted to contain a secretion signal and the PDB itself predicts it to be a soluble cytoplasmic protein. It is difficult to envision how it would interact with the periplasmic domain of YejM in a living cell. Cellular fractionation studies showing a periplasmic localization of YejL would need to accompany the data in Figure 2 in order for it to be meaningful. Genetic studies may also be useful-for example, does the YejL mutant show alterations in PhoPQ mediated survival or YejM tetramer formation?
The BNE gels are not convincing. Why does YejL migrate at such a large molecular weight? YejM migrates as a smear in both gels when loaded by itself. Adding YejL reduces the smear but it cannot be said that YejL alters the MW of YejM in either gel.

Minor issues:
Why not call YejM by its new name PbgA throughout? It would cause less confusion since the senior author was associated with the publication reporting the discovery last year. Figure 1A, the membrane domain in the figure should be labeled 0-190 not 191-240. Also the "j" in "YejL structure" is partially obscured.
I take issue with describing YejL as a 'ligand". To me and most biochemists a ligand is a small molecule that interacts with a target protein (receptor) causing some kind of conformational change and usually generation of a signal of some sort. What they are describing here is a protein-protein interaction and should be referred to as such, i.e. "YejL interacts with the periplasmic domain of YejM".
proteins that are Phospho-beta-glucosidases. To prevent confusion, we chose to use the official Uniprot database nomenclature of YejM (http://www.uniprot.org/uniprot/P40709). 2.) This figure is not included in the updated version of the manuscript. 3.) We agree that this was not the correct term to use and would rather term YejL as potential interaction partner of YejM.
We thank both reviewers for their helpful critique.
No competing interests.

Competing Interests:
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com