Phospholipid transporter shifts into reverse

The multisubunit phospholipid transport system Mla has been under scrutiny to determine whether it functions as an exporter or an importer. Structural studies accompanied by the reconstitution of the entire Mla system into proteoliposomes now reveal that ATP binding and hydrolysis drive phospholipid import.

The maintenance of outer membrane lipid asymmetry (Mla) phospholipid transport system of Escherichia coli is an ATP-binding cassette (ABC) transporter that was initially proposed to extract phospholipids from the external leaflet of the outer membrane for delivery to the inner membrane (retrograde flow). This functional assignment was based on genetic screens showing that mutants of the Mla system aberrantly accumulate phospholipids at the cell surface. Subsequent biochemical studies of the inner membrane ABC transporter subunits revealed that the Mla system spontaneously moves phospholipids toward the outer membrane (anterograde flow), but the role of ATP binding and hydrolysis in this process was not apparent. The resulting debate as to whether the Mla system drives retrograde versus anterograde phospholipid transport has called for the reconstitution of the inner and outer membrane Mla components into proteoliposomes, as was achieved previously for the lipopolysaccharide (LPS) transport system1. In this issue of Nature Structural & Molecular Biology, Tang et al.2 report a high resolution cryo-EM structure of the Mla inner membrane ABC transporter complex. The authors also reconstitute the entire Mla system into inner and outer membrane proteoliposomes. They confirm the previously reported spontaneous anterograde flow of phospholipids but critically demonstrate that ATP binding and hydrolysis shifts the Mla system into reverse, consistent with most of the genetic observations that have been accounted for previously by retrograde phospholipid transport.

E. coli and its relatives encounter bile salts in the gut and must maintain outer membrane lipid asymmetry to resist these detergents. The inner membrane represents an ordinary phospholipid bilayer, which is impermeable to aqueous solutes, but it can be easily penetrated by hydrophobic antibiotics and disrupted by detergents. Negatively charged functional groups in the lipid A domain of LPS are bridged in the outer membrane by divalent cations to provide bacteria with a permeability barrier to hydrophobic antibiotics and detergents. Homeostatic mechanisms ensure that the phospholipid-to-LPS ratio is optimal to maintain outer membrane lipid asymmetry. Conventional energy sources like ATP are only available at the inner membrane, which drive the biosynthesis of LPS and phospholipids and power their transport between these two distinctly different membrane systems.

One genetic argument in favor of anterograde phospholipid transport is the grouping of the MlaF2E2D6B2 complex with LptB2FGC and related ABC transporters, which all drive anterograde transport. The highest resolution structures of MlaF2E2D6B2 were recently obtained by cryo-EM at 2.9 Å (ref. 3) and 3.05 Å (ref. 4), respectively, in the presence of a membrane scaffolding protein to produce lipid nanodiscs. These findings largely recapitulate the structure reported by Tang et al., which was solved at 3.3 Å resolution using a mixture of detergent and phospholipids. Different orientations of the bound phospholipids probably reflect the different lipid matrices employed, but the two structures solved in lipid nanodiscs display strikingly similar electron density for the bound phospholipids. A cavity in MlaF2E2D6B2 for binding phospholipids is observed within the transmembrane MlaE2 dimer in the outward open conformation, and it connects directly to a phospholipid cavity within the hexameric ring formed by MlaD6 (Fig. 1a). A narrow channel in the center of MlaD6 appears to receive phospholipids from MlaC for transfer to the external leaflet of the inner membrane via two similarly narrow lateral openings in the walls of the MlaE2 dimer. The single transmembrane helices that extend from each MCE domain of the MlaD6 ring were not apparent in an earlier 10 Å resolution structure5. These helices act like pendentives to connect the hexameric MlaD6 with the underlying dimeric MlaE2 membrane domain. In response to the binding and hydrolysis of ATP, the MlaF2 ABC subunits drive conformational changes in MlaE2 and MlaD6 that serve to extrude phospholipids into the membrane. A regulatory function for MlaB2, a STAS domain protein that stabilizes MlaF2, has recently been proposed6. New evidence suggests that MlaF2E2D6B2 may somehow flip-flop phospholipids to balance their distribution across the bilayer7.

Fig. 1: Cell envelope phospholipid homeostasis defects in E. coli.

a, Wild-type cells maintain outer membrane lipid asymmetry to provide resistance to large solutes and detergents. The Mla system initiates the removal of aberrantly surface-exposed phospholipids with the outer membrane lipoprotein MlaA, which is anchored to the OmpF or OmpC porins (PDB 5NUO) and provides a channel to move phospholipids toward MlaC (PDB 5UWA) in the periplasmic space. The MlaF2E2D6B2 ABC transporter (PDB 6XBD) allows spontaneous anterograde flow of phospholipids toward the outer membrane, but the binding and hydrolysis of ATP reverses the phospholipid flow back toward the inner membrane. A cavity for binding phospholipids is observed in the outward open conformation within the transmembrane MlaE dimer, which directly connects to a phospholipid cavity within the hexameric ring formed by MlaD. A narrow channel in the center of MlaD appears to receive phospholipids from MlaC for transfer to the external leaflet of the inner membrane via two lateral openings in the walls of the MlaE dimer. b, Treatment of wild-type cells with EDTA removes some LPS from the cell surface to promote its replacement with phospholipids. Mla mutants allow surface phospholipids to accumulate by preventing their removal. In both cases, the outer membranes resist large solutes and detergents but can trigger the PldA and/or PagP enzymes. LetB (PDB 6V0C) and PqiB proteins form stacked hexameric rings that can create an internal channel for phospholipids, making them candidates to facilitate phospholipid transport. c, Treatment of Mla mutant cells with EDTA removes more LPS from their cell surface, which allows the phospholipid bilayer domains to expand into separate phases with elevated phospholipid mobility and vulnerability to detergents. Additionally, phase boundary defects can create transient cracks to allow penetration of large solutes, and these are accentuated by fatty acyl chain length differences between the lipid A and phospholipid phases. Combining EDTA treatment with relatively impermeable lipid asymmetry mutants thereby enables their genetic selection by monitoring detergent-sensitivity. This antibiotic supersusceptible phenotype can also be directly produced in mutants that alter the flow of phospholipids (mlaA*) or LPS (lptD4213) or that perturb the regulation of LpxC.

Despite the tendency of MlaF2E2D6B2 to facilitate the spontaneous anterograde flow of phospholipids8,9, the MlaA channel is not designed to deliver phospholipids into the inner leaflet of the outer membrane; instead, it is meant to extract them from the external leaflet10. Indeed, the dominant negative mlaA* mutation reverses the protein’s normal function by allowing phospholipids to flow directly into the external leaflet11. MlaA is anchored to the OmpF or OmpC porins, which are permeable to solutes under 600 Da (Fig. 1a)12. The high affinity of MlaC for phospholipids ensures that phospholipids flow from the outer membrane external leaflet through MlaA and toward MlaC in the periplasmic space. MlaC also spontaneously accepts phospholipids supplied from inner-membrane MlaD.

Tang et al. have now taken a cue from the Kahne group, which has successfully reconstituted the inner and outer membrane components of the LPS transport apparatus into proteoliposomes and marked them with fluorescent probes to monitor the ATP-dependent flow of LPS, which moves, as expected, in the direction of the outer membrane13,14. By separately reconstituting the MlaA–OmpF and MlaF2E2D6B2 complexes into proteoliposomes and employing fluorescently labeled phospholipids in the presence of ATP and MlaC, Tang et al. demonstrate that phospholipids flow instead in the direction of the inner membrane. The question of how phospholipids flow toward the outer membrane now requires further explanation.

Like MlaD, the PqiB and LetB MCE-domain proteins form hexameric rings surrounding an interior cavity for phospholipids, but they are stacked upon each other to create an internal channel for phospholipid transport (Fig. 1b)5,15,16. It is unclear how these MCE-domain channels form an interface with the inner and outer membranes, whether they exhibit specificity for particular phospholipids, what direction the phospholipids flow, and whether they are coupled to any conventional sources of cellular energy. Although individual pqiB and letB mutations appear to be benign, in combination with mutations in mla genes they exhibit selectable defects in outer membrane lipid asymmetry17,18 and, along with yhdP11, are arguably the best candidates for controlling anterograde phospholipid flow.

The recent advances in understanding the structural biology of cell envelope phospholipid homeostasis have hinged on the development of selectable phenotypes to monitor lipid transport and biosynthesis. The antibiotic supersusceptible phenotype discussed by Vaara19 and Nikaido20 (Fig. 1c) can explain how cell envelope homeostasis disruption sensitizes cells to antibiotics and detergents. However, Malinverni and Silhavy noted in their initial report of the Mla system21 that our current understanding of outer membrane permeability defects needs to be re-evaluated because “...mutants of the Mla pathway are relatively impermeable to most compounds with the exception of SDS-EDTA, and they exhibit no additional defects in LPS or [outer membrane] protein levels; their only membrane defect is an increase in surface-exposed [phospholipids].” We know this because the calcium-dependent phospholipase PldA and the phospholipid palmitoyltransferase PagP remain latent in the outer membrane when phospholipids are absent from the external leaflet, but Mla mutants or EDTA treatment can trigger activity by delivering phospholipid substrates to the active sites of the enzymes on the cell surface.

As summarized in Fig. 1, treating wild-type cells with EDTA removes enough LPS from the cell surface to cause its replacement with phospholipids. It appears at first that, despite triggering the activity of PagP to mark their presence, the phospholipids in the external leaflet are relatively immobile because the membranes remain resistant to large solutes and detergents. Mla pathway mutants similarly trigger PagP and PldA while remaining resistant to large solutes and detergents (Fig. 1b). Treatment of Mla mutant cells with EDTA removes more LPS from their cell surface, which allows the phospholipid bilayer domains to expand into separate phases wherein the increased phospholipid mobility sensitizes cells to detergents. Additionally, phase boundary defects are accentuated by fatty acyl chain length differences between the lipid A and phospholipid phases, which can create transient cracks that permit the penetration of large solutes (Fig. 1c).

The mlaA* mutant accumulates surface phospholipids, but it also accelerates LPS production by triggering an increase in the fatty acyl CoA pool in response to the elevated activity of PldA22. When free fatty acids are produced by PldA in the outer membrane, they can act as second messengers to help coordinate the biosynthesis of LPS with phospholipids to maintain lipid asymmetry. LpxC controls LPS production and is stabilized by fatty acyl CoAs acting via an unknown mechanism on the PbgA–LapB–FtsH complex, which degrades LpxC when LPS accumulates to toxic levels in the external leaflet of the inner membrane23. The lptD4213 mutant reduces anterograde flow of LPS by impeding its delivery to the outer membrane, which also directly disturbs the lipid asymmetry but by LPS depletion. Viable alleles of pbgA and lpxC also result in LPS depletion, but by limiting its biosynthesis in the cytoplasm. By contrast, lapB and ftsH mutants result in toxic LPS accumulation, but they can be rescued by increasing phospholipid biosynthesis. These examples illustrate how the phospholipid/LPS ratio is maintained in conjunction with feedback-mediated regulation exerted at LpxC to adjust the LPS flow to match that of the phospholipids24,25. The Mla retrograde phospholipid transport system has an important role in maintaining the outer membrane lipid asymmetry so it can resist antibiotics and detergents, and it can finally be appreciated in terms of structure–function relationships that are consistent with both the genetics and biochemistry.


  1. 1.

    Lundstedt, E., Kahne, D. & Ruiz, N. Chem. Rev. (2020).

  2. 2.

    Tang, X. et al. Nat. Struct. Mol. Biol. (2020).

  3. 3.

    Chi, X. et al. Cell Res. (2020).

  4. 4.

    Coudray, N. et al. Elife 9, e62518 (2020).

    Article  Google Scholar 

  5. 5.

    Ekiert, D. C. et al. Cell 169, 273–285.e17 (2017).

    CAS  Article  Google Scholar 

  6. 6.

    Kolich, L. R. et al. Elife 9, e60030 (2020).

    CAS  Article  Google Scholar 

  7. 7.

    Hughes, G. W. et al. Preprint at bioRxiv (2020).

  8. 8.

    Hughes, G. W. et al. Nat. Microbiol. 4, 1692–1705 (2019).

    CAS  Article  Google Scholar 

  9. 9.

    Powers, M. J., Simpson, B. W. & Trent, M. S. Elife 9, e56571 (2020).

    CAS  Article  Google Scholar 

  10. 10.

    Abellón-Ruiz, J. et al. Nat. Microbiol. 2, 1616–1623 (2017).

    Article  Google Scholar 

  11. 11.

    Grimm, J. et al. Proc. Natl Acad. Sci. USA 117, 26907–26914 (2020).

    CAS  Article  Google Scholar 

  12. 12.

    Chong, Z.-S., Woo, W.-F. & Chng, S.-S. Mol. Microbiol. 98, 1133–1146 (2015).

    CAS  Article  Google Scholar 

  13. 13.

    Sherman, D. J. et al. Science 359, 798–801 (2018).

    CAS  Article  Google Scholar 

  14. 14.

    Xie, R., Taylor, R. J. & Kahne, D. J. Am. Chem. Soc. 140, 12691–12694 (2018).

    CAS  Article  Google Scholar 

  15. 15.

    Isom, G. L. et al. Cell 181, 653–664.e19 (2020).

    CAS  Article  Google Scholar 

  16. 16.

    Liu, C., Ma, J., Wang, J., Wang, H. & Zhang, L. J. Mol. Biol. 432, 1008–1019 (2020).

    CAS  Article  Google Scholar 

  17. 17.

    Nakayama, T. & Zhang-Akiyama, Q.-M. J. Bacteriol. 199, e00606–16 (2016).

    PubMed  PubMed Central  Google Scholar 

  18. 18.

    Isom, G. L. et al. Sci. Rep. 7, 8608 (2017).

    Article  Google Scholar 

  19. 19.

    Vaara, M. Antimicrob. Agents Chemother. 37, 2255–2260 (1993).

    CAS  Article  Google Scholar 

  20. 20.

    Nikaido, H. Chem. Biol. 12, 507–509 (2005).

    CAS  Article  Google Scholar 

  21. 21.

    Malinverni, J. C. & Silhavy, T. J. Proc. Natl Acad. Sci. USA 106, 8009–8014 (2009).

    CAS  Article  Google Scholar 

  22. 22.

    May, K. L. & Silhavy, T. J. Mbio 9, e00379–18 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Clairfeuille, T. et al. Nature 584, 479–483 (2020).

    CAS  Article  Google Scholar 

  24. 24.

    Cian, M. B., Giordano, N. P., Masilamani, R., Minor, K. E. & Dalebroux, Z. D. Infect. Immun. 88, e00758–19 (2019).

    Article  Google Scholar 

  25. 25.

    Guest, R. L., Rutherford, S. T. & Silhavy, T. J. Trends Microbiol. (2020).

Download references

Author information



Corresponding author

Correspondence to Russell E. Bishop.

Ethics declarations

Competing interests

The author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bishop, R.E. Phospholipid transporter shifts into reverse. Nat Struct Mol Biol 28, 8–10 (2021).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing