Click‐Mediated Labeling of Bacterial Membranes through Metabolic Modification of the Lipopolysaccharide Inner Core

Abstract

Metabolic glycan labeling has recently emerged as a very powerful method for studying cell-surface glycans, which has applications that range from imaging glycans in living multicellular organisms, such as zebrafish or mice, to the identification of metastasis-associated cell-surface sialoglycoproteins. This strategy relies on the cellular biosynthetic machinery assimilating a modified monosaccharide that contains a bioorthogonal chemical reporter. The metabolic incorporation of this reporter into glycans can be further visualized by chemical ligation with a label, such as a fluorescent probe. Somewhat surprisingly, previous studies have mainly focused on the labeling of vertebrate glycans by using derivatives of common monosaccharides, such as Nacetyl neuraminic acid (or its N-acetylmannosamine precursor), N-acetylglucosamine, N-acetylgalactosamine, and fucose. In spite of a much higher degree of diversity in their monosaccharide building blocks as well as an essential role in bacterium–host interactions and bacterial virulence, bacterial polysaccharides have been poorly explored in terms of in vivo structural modifications. Bacteria are divided into Grampositive and Gram-negative bacteria. Whereas Gram-positive bacteria are surrounded by a peptidoglycan cell wall, Gramnegative bacteria are covered by a dense layer of lipopolysaccharides that are embedded in their outer membrane. These lipopolysaccharides are involved in the structural integrity of the cell and are often considered as determinants of pathogenicity. Although lipopolysaccharides appear to be an interesting target for specific and well-defined glycan metabolic labeling in Gram-negative bacteria, attempts to achieve this goal have been limited to the introduction of modified l-fucose derivatives into a customized, genetically engineered strain of Escherichia coli. Although it is a very interesting proof of concept, this l-fucose-based approach has some limitations as l-fucose is not generally present within the lipopolysaccharides of all Gram-negative bacteria, but is found in the O-antigens of specific strains. Secondly, free lfucose is not an intermediate in the normal E. coli “de novo” pathway and, therefore, should not be directly activable into a nucleotide-sugar donor without the introduction of an alternative pathway, known as the “salvage pathway”, into the organism of interest by genetic engineering (metabolic pathway engineering). Furthermore, once activated in the form of a modified guanosine-5’-diphosphate–fucose (GDP– Fuc), the l-fucose analogue might be transformed into a correspondingly modified GDP–mannose (GDP–Man) by the reverse de novo pathway, and potentially further metabolized into various other compounds, a process which could result in the chemical reporter being spread through other pathways of sugar metabolism or beyond. As a result of all of these limitations, and as our goal was labeling the lipopolysaccharides of bacteria with no genetic modification, we investigated whether another sugar could be used as a target for the metabolic modification of glycans. From all of the potential targets, 3-deoxy-d-mannooctulosonic acid (KDO) appears to be a very attractive candidate. Indeed, KDO is a specific and essential component of the inner core of lipopolysaccharides, and has long been considered as being present in the lipopolysaccharides of almost all Gram-negative species (as well as higher plants and algae), in which at least one residue is directly connected to lipid A (Scheme 1a). Because of its vital importance, KDO has been considered as a determinant for the characterization of Gram-negative bacteria, and the KDO pathway as a potential target for the development of new antibacterial compounds. In the KDO pathway (Scheme 1b), arabinose5-phosphate (arabinose-5-P) is condensed with phosphoenolpyruvate (PEP) to give KDO-8-phosphate (KDO-8-P), which is then transformed into free KDO, and further activated to form the cytidine monophosphate (CMP)–KDO donor prior to lipopolysaccharide elaboration. For all of these reasons, we hypothesized that the KDO pathway, as a lipopolysaccharidespecific pathway, may be tolerant enough to incorporate a modified analogue of KDO, such as 8-azido-8-deoxy-KDO (1, Scheme 2), into the core of E. coli lipopolysaccharides, and potentially other Gram-negative bacteria. Given the presence of free KDO as an intermediate in the pathway, we postulated that if the cell penetration of this analogue of KDO was sufficient, it could then be directly activated, partially replace endogenous KDO in lipopolysaccharides, and be detected on the cell surface by azide–alkyne click chemistry (Figure S1 in the Supporting Information). Moreover, modification of the C8-position of KDOwith a bioorthogonal azido group should prevent reverse metabolism by KDO-8-P [*] Dr. A. Dumont, Dr. S. Dukan Aix Marseille Universit , Laboratoire de Chimie Bact rienne (UMR 7283), Institut de Microbiologie de la M diterran e (IMM), CNRS, 31 Chemin Joseph Aiguier 13402 Marseille (France) E-mail: sdukan@imm.cnrs.fr