How to take your sticky proteins out safely

Lipid-bearing proteins are essential players in the cell wall of gram-negative bacteria, where they contribute to the maintenance of the external membrane, the selective barrier against the outside world. Inevitably, getting lipoproteins from the cytoplasm to their final destination in the outer membrane is essential. In developing this transport mechanism, evolution faced the significant challenge of how to deal with the very hydrophobic lipid moiety that cannot be simply exposed to the aqueous periplasmic space environment during trafficking from the inner to the outer membrane. As is typical, the solution was provided by specialized proteins that have a dual role: They act as molecular chaperones that accompany the lipoprotein while at the same time shielding the hydrophobic lipid. In PNAS, a study has resolved the first structure of a chaperone lipoprotein complex, revealing a likely generic binding mechanism between the lipid anchor of the lipoprotein and the delivery chaperone LolA (1).

Escherichia coli, the workhorse of bacterial biology, possesses at least 90 species of lipoproteins that play key roles in a wide variety of physiological processes. These range from bacterial physiology (nutrient transport, signal transduction, and protein folding) to maintenance of the cellular structural integrity (envelope stability and contact with the peptidoglycan support mesh) and coupling it to cell division, but also to virulence and infection (surface adhesins, export of toxins, formation of biofilms, and immune system evasion and antibiotics resistance) (2). Like most proteins destined for export outside the cytoplasm, bacterial lipoproteins carry N-terminal signal sequences. These conserved extensions, termed type II in lipoproteins, play two main roles: First, they target the lipoprotein precursor from the cytoplasm, where they are synthesized, to the Sec translocase, the pump that will take the precursor to the external leaflet of the inner membrane. Second, they “guide” a posttranslocation maturation process that is reminiscent of introducing address tags. In a final step, they are removed and left behind. This process sequentially modifies a cysteine residue into a mature triacylated form that allows the soluble functional domain of the signal sequenceless mature lipoprotein to be anchored to the cell membrane (for a review, see ref. 2). The cysteine is located in the so-called lipobox, at the N terminus of the mature domain. Its conserved sequence features can dictate either retention at the inner membrane or escape to the outer membrane.

Mature lipoproteins destined to the outer membrane need to cross the periplasm, but, because of their extremely hydrophobic attachment, their release into this aqueous space does not occur spontaneously and requires a specific machinery. This lipoprotein outer-membrane localization (Lol) pathway comprises five proteins(LolABCDE): The tetrameric LolCDE complex is embedded in the inner membrane and is built of a heterodimer of the homologous proteins LolC and LolE that form a transmembrane channel, and a homodimer of the LolD ATPase, bound at the cytoplasmic face of the complex (Fig. 1). In an ATP-independent manner, LolC recruits the soluble carrier chaperone LolA at the periplasmic face by combining a Pad, three surface-exposed residues, and the interaction via a Hook, a conserved finger-like protrusion inserted in the LolA structure (3). During ATP hydrolysis by LolD, the trafficking lipoprotein sitting between LolC and LolE is then transferred from LolE to LolA bound to LolC (4). The loaded LolA disengages from the complex, to ferry the lipoprotein through the periplasm. At the outer membrane, LolA interacts with the anchored LolB receptor, itself a lipoprotein, where the freshly arrived lipoprotein passenger is spontaneously transferred from LolA to LolB (5). The latter then mediates the insertion of the passenger lipoprotein into the outer membrane.

Fig. 1.

Proposed model of lipoprotein trafficking based on the transport of the lipid anchor. By solving the structure of an LolA-triacylated peptide complex under different conditions, the authors (1) propose a model for lipoprotein trafficking in the cell wall based on the transfer of the lipid anchor. (A–E) After the extraction of the lipoprotein from the inner membrane (IM) by the LolCDE complex (A) upon ATP binding (red), the first acyl chain of the lipoprotein model (LP) interacts with the LolA cavity, disengaging the stabilizing Hook from LolC (B) and priming the wider opening of the LolA cavity (C). The second and third acyl chains can then engage the cavity in a coordinated packing (D), releasing the LolA shuttling chaperone to the periplasm. At the outer membrane (OM), LolA interacts with LolB, itself a lipoprotein, to relay the triacyl moieties of the passenger lipoprotein (E). (F–I) The R43L mutation in LolA, known to block the trafficking of lipoproteins, induces constitutive opening of its cavity prior to its substrate binding (F), resulting in an altered packing of the lipid anchor (G and H). LolA still reaches the outer membrane and LolB, but the delivery is not permitted (I). This altered packing is proposed as an original strategic target for the development of new antibiotics targeted at the cavity of LolA and perturbing the packing of the triacyl chains.

While structures of the soluble LolA and LolB were resolved decades ago (6), the complete molecular description of the Lol machinery was limited to homology modeling using similar ABC transporters until last year, when structures of the LolCDE complex in the absence or presence of a lipoprotein passenger were resolved by cryoelectron microscopy in detergent (7) or in nanodiscs (8), offering a molecular glimpse of the first transport step toward the outer membrane. This involves the extraction of the lipoprotein from the inner membrane. The LolCDE complex recognizes the lipoprotein passenger in the bilayer by specifically interacting mainly with the lipid anchor on the lipoprotein with the amide-linked acyl chain of the N-terminal peptide. The binding of ATP triggers a large conformational change of LolCDE extruding the lipoprotein trapped between the LolC and LolD in the membrane and relaying it from the periplasmic domain of LolE to LolA. However, the molecular details of this transfer to LolA have remained unclear, as has its capacity to carry and deliver over 90 evolutionary divergent lipoproteins.

To crack this mechanism, Kaplan et al. (1) managed to trap an LolA lipoprotein intermediate that was stable enough for structural investigation. However, while such a complex was easily isolated, it did not properly crystallize, possibly because of the flexible linker located at the interface between the triacyl group and the soluble domain of the lipoprotein. To overcome this issue, the authors engineered a modified model substrate (Pal, a peptidoglycan-associated lipoprotein) with a cleavage site for the TEV protease located between the triacyl group and the flexible linker of the protein, allowing for the soluble domain to be proteolytically removed after binding to LolA. As a result, a short triacylated fragment made of the 13 first residues of the globular domain (Pal₁₃) got trapped inside LolA. Since the TEV cleavage site did not hinder the trafficking of the lipoprotein passenger from the inner to the outer membrane in vivo, the authors copurified LolA with Pal₁₃ and crystallized the complex, overcoming the flexibility issue and resolving the first atomistic structure of the LolA substrate complex. The triacyl chains are packed in a cavity formed by the β-barrel of LolA stabilized by hydrophobic residues, while the ester linkage to the cysteinyl residue is stabilized by polar interactions. Residues essential for these interactions were confirmed by mutagenesis, corroborating their key roles for lipoprotein stabilization. The superposition of their LolA structures in the absence or presence of the triacyl chain peptide ligand demonstrated that the binding of the triacyl chains induced a conformational change in LolA, widening the hydrophobic cavity to accommodate the three triacyl chains. A step further, the alignment of their LolA-Pal₁₃ structure with the recent one of LolA-LolCDE (7) showed a steric clash between the first triacyl chain and the stabilizing Hook from LolC. Together, these observations allowed the authors to propose an updated molecular scheme for lipoprotein transfer from the LolCDE complex to LolA in which the lipoprotein extracted from the inner membrane by LolCDE during the ATP-driven step (Fig. 1A) is flipped to expose the triacyl chains to LolA bound to LolC by the Pad and Hook. The insertion of the first lipid chain disengages the Hook (Fig. 1B) and would induce the wide opening of the LolA cavity (Fig. 1C) to accommodate all three triacyl chains (Fig. 1D), resulting in the release of LolA to the periplasm (Fig. 1E), and eventually reach LolB.

The proposed mechanism is sensitive to the packing of the triacyl chains in the LolA cavity itself. Indeed, by solving the structure of an LolA mutant (R43L) that can bind lipoproteins but is defective to their transfer to LolB in the absence or presence of their Pal₁₃ model ligand, the authors (1) demonstrate that the mutation itself favored the premature “opening” of the LolA cavity while bound on LolC (Fig. 1E), resulting in an altered packing of the three triacyl chains (Fig. 1 G and H), compared to the wild-type LolA. This altered lipoprotein binding state interfered with some regions of the LolB binding domain, blocking the transfer of the lipoprotein passenger to its final partner, and consequently stopping the trafficking (Fig. 1I).

Kaplan et al. (1) propose a general mechanism for bacterial lipoprotein trafficking based on the trafficking of the triacyl chains themselves, rather than the proteins on which they are carried. Indeed, since the interaction between LolA apparently can occur with the triacyl chains Pal₁₃ alone, the authors propose this model as a general mechanism for lipoprotein trafficking, based on the binding and transfer of the triacyl chains. This model would easily explain away the apparent promiscuity of how more than 90 different lipoproteins can be transported by the Lol pathway.

The authors (1) leave some interesting questions open. First, their model suggests the flipping of the triacyl chains in the LolCDE complex from facing toward the inner membrane during extraction to facing the hydrophobic cavity of LolA at the periplasmic side of the complex during the transfer. It is not clear how this could happen when the triacyl chains are attached to the soluble part of the lipoprotein passenger. It suggests the intriguing possibility that the LolCDE complex might accommodate the rotation of the complete lipoprotein by actively helping the soluble domain to flip if required, or by just providing the protective environment to the lipid anchor while the mature domain attached to the lipid moiety by a very flexible linker remains free in the aqueous periplasmic space. Second, because of the high structural homology between LolA and LolB, the authors propose that the mechanism of binding of lipoproteins to LolB, and therefore their transfer between the two chaperones, is probably also based on the binding of the triacyl chains, rather than on the carrier protein moiety itself. This is also supported by previous work demonstrating unidirectional lipoprotein transfer from LolA to LolB driven solely by an increase in affinity (5), probably toward the triacyl chains. Also, if the transfer to LolB results in a similar binding mode to that of LolA, this opens the question of how LolB alone can invert the orientation of the triacyl chains from facing toward the periplasmic side to facing to the outer membrane to become inserted. Since an LolCDE-like complex is absent from the outer membrane, what drives the flipping and insertion of the lipoprotein? Finally, because different packing of the triacyl chains in the LolA cavity blocked trafficking, the authors (1) propose an original target for antibiotics design. Indeed, based on their model, small-molecule inhibitors could be screened to perturb the packing of the acyl chains in LolA, and block the trafficking of essential lipoproteins to the outer membrane, weakening the bacterial cell wall with lethal consequences to their viability.