A Better Understanding of What Makes Some Proteins “Fat”

TEXT

The posttranslational modification of proteins by the covalent attachment of lipids has been a relatively niche area of research, predominantly focused on bacterial pathogens. This may be due to the experimental challenges involved in the purification of these proteins and the membrane proteins that are involved in performing the modification. Lipoproteins make up a considerable proportion of secreted proteins in some bacteria and have been shown to perform a variety of functions, including cell division, cellular infrastructure, protein localization, antibiotic resistance, nutrient adsorption, and signal transduction, to name but a few. The first step in this posttranslational modification is the addition of a diacylglyceryl moiety to a conserved C-terminal cysteine residue which is catalyzed by the enzyme prolipoprotein diacylglyceryl transferase (Lgt), further characterization of which is presented in this issue of the Journal of Bacteriology by Pailler et al (6).

The biogenesis of lipoproteins, or the “fattening” of proteins, is a three-stage process involving the initial cysteine modification. This is then followed by the removal of a signal peptide by the lipoprotein-specific signal peptidase (LspA), leaving a new amino terminus which is further modified by the enzyme apolipoprotein N-acyltransferase (Lnt). This process was first outlined 30 years ago by Tokunaga and coworkers as a four-step pathway, with the initial transfer of a glyceryl group to the cysteine followed by the action of one or more 0-acyltransferases acylating the sn-3 and sn-2 hydroxyls of the glyceryl moiety of glycerylcysteine to form the diacylglyceryl-modified prolipoprotein (11). This model was subsequently modified following the demonstration of direct transfer of a diacylglyceryl moiety from phosphatidylglycerol to the conserved amino cysteine residue (9) and is now still largely held as the accepted pathway for lipoprotein maturation. This pathway is of major interest as there is strong evidence to suggest that the three enzymes are essential for viability in Gram-negative bacteria (1, 2, 12), with the signal peptidase being a target for the antibiotic globomycin (4).

Pailler et al. (6) use classical and modern molecular biological techniques in tandem to provide information on the membrane topology of the enzyme prolipoprotein diacylglyceryl transferase, Lgt. Their study is important in reminding us of the limitations of in silico methods of protein analysis and the importance of experimentation for confirmation. While the importance of Lgt in diderm bacteria is without question, given its essential nature in cell viability, the importance of its role in some monoderms has recently been given some renewed attention with the observation that Lgt is required for the full virulence of Bacillus anthracis (5). The approaches taken by Pailler et al. (6) address their initial conflicting findings on the membrane topology of Lgt. Using a variety of in silico methods applied to Lgt from Escherichia coli, it could not be determined whether this protein had five or seven transmembrane domains. Their use of alkaline phosphatase and beta lactamase fusions in conjunction with a substituted cysteine accessibility method (SCAM) resolved this ambiguity, demonstrating that Lgt, from E. coli at least, does indeed contain seven transmembrane domains. This finding is perhaps, in part, the reason that Lgt has yet to be structurally characterized since its discovery 30 years ago.

Although the experiments conducted by Pailler et al. are convincing with respect to confirming an integral membrane localization, the authors acknowledge that they cannot explain, from their studies, a previous report demonstrating a cytoplasmic localization of Lgt (10). One possibility is that there is another enzyme that exhibits Lgt activity and that this is the membrane-associated/cytosolic Lgt observed by Selvan et al. However, as these authors employed an assay that exploited a recombinant strain overexpressing Lgt, this seems unlikely. Nevertheless, the identification of lgt paralogues in several Gram-positive species does not preclude this possibility. The roles of these lgt paralogues remain unclear, although it has been proposed that they may be dedicated to the processing of specific lipoproteins in a manner analogous to the processing of specific wall-anchored proteins by substrate-specific sortase enzymes (3). Given that the conserved signal peptide appears to be essential for Lgt activity, the presence of signal peptides of both hydrophobic and hydrophilic nature may support the idea of Lgts with different peptide specificities and, thus, different locales.

The presence of paralogues of Lgt has only been postulated due to the observation of genes encoding proteins that fall into the prolipoprotein diacylglyceryl transferase family PF01790 (http://pfam.sanger.ac.uk/family/PF01790). Lgt proteins are also characterized by the presence of a prolipoprotein diacylglyceryl transferase signature, PS01311 (http://prosite.expasy.org/PDOC01015). From the topology studies of Pailler et al. in this issue (6), this signature is split between the periplasmic space and the inner membrane, residues 142 to 154. Their study expands upon studying the membrane topology of Lgt in E. coli by investigating the essentiality of certain key conserved residues. From their bioinformatic analysis, five residues were identified that are conserved between all available Lgt sequences across Firmicutes, Proteobacteria, and Actinobacteria. Probably not surprisingly, two of these, R143 and G154, were located in the prolipoprotein diacylglyceryl transferase signature, although mutation of these residues to alanine failed to complement a conditional allele in their studies. Essential amino acids of Lgt have been investigated previously, with H103 and Y235 being implicated as critical for Lgt activity (8). However, as Pailler et al. (6) neatly demonstrate, the failure of a mutant allele to complement may not be solely attributable to the mutation per se, as the mutation may inhibit the expression of the protein. In the studies that implicated an H103Q mutation as essential for Lgt activity, there were no data presented to demonstrate that this allele was actually expressed, although Pailler et al. do comment from their studies that there is not always a correlation between levels of protein expression and functionality (6). It is interesting to note that an earlier study that used diethylpyrocarbonate (DEPC) chemical modification to characterize Lgt activity indicated that a histidine or a tyrosine was essential for enzyme function. Pailler et al. demonstrate that a Y26A substitution failed to complement a conditional allele and that this residue is one of the five in Lgt that is conserved across all bacteria included in their analysis. Previous studies have shown that a Y26F mutation had no effect on Lgt activity, but this may be due to the nature of this mutation, which retains the aromatic ring structure which may be key for the activity of Lgt (8).

So why the importance of characterizing this bacterial enzyme and understanding the mechanism of the addition of lipid moieties to microbial proteins, or alternatively, what makes them fat? Lipoproteins, by virtue of their physical characteristics, serve as molecules that can perform a vast array of functions at the aqueous environment-membrane interface, with the fatty lipid portion serving as a membrane anchor allowing the functional component of the protein to perform its function in the aqueous environment. Such a property has led to interest in this modification for such purposes as nanomaterials, enzyme-linked immunosorbent assay (ELISA), protein targeting, and biosensors. Lipopeptides from bacteria have also been shown to have antifungal behavior, which could have major implications for human health and food security issues (7). However, a major area of interest in protein acylation by bacteria is that of prophylaxis. Over the last 2 decades, a large number of molecules have been characterized that have the ability to recognize motifs of microbes and that subsequently trigger an inflammatory response. The Toll-like receptors (TLRs) constitute a family of these proteins forming part of the innate defense response and also functioning in activation of the acquired immune response. Ligand binding by TLRs is mediated by leucine-rich repeats (LRRs) which recognize microbial “patterns.” One of these patterns is the acyl chains of bacterial lipoproteins. TLR2 is key in the recognition of bacterial lipoproteins, heterodimerizing with either TLR6 or TLR1, depending on whether the protein is diacylated or triacylated, respectively (14). Therefore, characterizing how these proteins are modified and understanding the enzymes behind this may lead to novel targets for microbial intervention and/or the generation of novel vaccines. To highlight the importance of bacterial lipoproteins in prophylaxis and, consequently, the importance of understanding their biogenesis, the example of the abundant peptidoglycan-associated lipoprotein (PAL) from Legionella pneumophila can be cited (13). This lipoprotein has been shown to stimulate both the humoral and cellular immune response depending on its mode of administration, promoting its role as a potential future vaccine candidate. The findings reported by Pailler et al. in this issue go some way to characterizing the enzyme, Lgt, that is responsible for the first step in this bacterial posttranslational modification. Nevertheless, many questions, let alone the structure of this enzyme, remain unsolved. One future avenue of interest will be to answer the question, how do the three enzymes, Lgt, LspA, and Lnt, act holistically? Given the difficulty of characterizing just one, this will be a question not for the fainthearted.