Selective molecular transport through the protein shell of a bacterial microcompartment organelle

Bacterial microcompartments are widespread prokaryotic organelles that have important and diverse roles ranging from carbon fixation to enteric pathogenesis. Current models for microcompartment function propose that their outer protein shell is selectively permeable to small molecules, but whether a protein shell can mediate selective permeability and how this occurs are unresolved questions. Here, biochemical and physiological studies of structure-guided mutants are used to show that the hexameric PduA shell protein of the 1,2-propanediol utilization (Pdu) microcompartment forms a selectively permeable pore tailored for the influx of 1,2-propanediol (the substrate of the Pdu microcompartment) while restricting the efflux of propionaldehyde, a toxic intermediate of 1,2-propanediol catabolism. Crystal structures of various PduA mutants provide a foundation for interpreting the observed biochemical and phenotypic data in terms of molecular diffusion across the shell. Overall, these studies provide a basis for understanding a class of selectively permeable channels formed by nonmembrane proteins.The complex behavior of biological systems depends fundamentally on the controlled movement of molecules between cellular compartments. Such processes occur in a wide range of biological contexts through the movement of ions and small molecules across lipid bilayers via proteins—channels and pumps—embedded in the bilayer. Achievements in understanding molecular transport in transmembrane systems have contributed to scientific disciplines from cell biology and physiology to membrane biophysics (1, 2). Interestingly, there exists a second type of system for molecular transport through proteins that is fundamentally different and much less understood. Hundreds of species of bacteria produce large subcellular organelles known as microcompartments (MCPs), which consist of metabolic enzymes encapsulated within proteinaceous shells reminiscent of viral capsids (reviewed in ref. 3). For MCPs to function, substrates and products must move across their outer protein shell, which lacks any lipid-based membrane. In the last several years, 3D structures of the proteins that comprise MCP shells have revealed narrow pores through their centers that have been hypothesized to be the routes by which substrates enter (and products escape from) MCPs (4; reviewed in ref. 5). However, experimental evidence to support this key hypothesis and the molecular principles involved is lacking.The overarching function of MCPs is to optimize metabolic pathways that have toxic or volatile intermediates. MCPs are present across at least 11 different bacterial phyla, where they carry out diverse metabolic processes (6–12). The carboxysome MCP is used to enhance CO₂ fixation in nearly all bacteria that use the Calvin cycle, and it has been estimated that 25% of the carbon fixation on Earth occurs within this proteinaceous bacterial organelle (9). The 1,2-propanediol utilization (Pdu) and ethanolamine utilization (Eut) MCPs are used to optimize 1,2-propanediol (1,2-PD) and ethanolamine catabolism, respectively (13–15), and the degradation of these compounds is thought to promote enteric pathogenesis (12, 16, 17). Although the Pdu and Eut MCPs, the carboxysome, and other metabolically diverse MCPs of unknown function encapsulate distinct sets of enzymes, all have shells built from homologous proteins suggesting they operate by conserved functional principles. Most models of MCP function propose that the protein shell acts as a diffusion barrier that allows passage of substrates (and products) while limiting the escape of a toxic or volatile metabolic intermediate such as CO₂ or toxic aldehyde (9, 18), but selective permeability by MCP shells has not been established experimentally.The shells of MCPs are assembled primarily from a family of small proteins that have so-called bacterial microcompartment (BMC) domains (5). Many BMC domain proteins form flat, hexagonally shaped oligomers that tile into extended sheets that form the basis of the MCP shell (4, 19, 20) (Fig. 1). In most cases, MCP shells are composed of four to eight different types of functionally diversified BMC domain proteins, some of which have pores proposed to mediate the selective movement of metabolites across the shell (4, 7, 8, 18). For example, the PduA shell protein from the Pdu MCP has a small central pore (∼6 Å) that is lined with numerous hydrogen-bond donors and acceptors, leading to a suggested role in the preferential movement of 1,2-PD over the less polar propionaldehyde (a toxic intermediate) (21). In addition, a subgroup of BMC proteins have been crystallized in two distinct conformations where the central pore is either fully closed or opened widely (12–15 Å), suggesting that a gating mechanism might control the movement of larger molecules (such as enzymatic cofactors) across the MCP shell (22, 23). However, no physiological or biochemical studies demonstrating transport or selective movement specifically through any MCP pore have been reported. As a result, the idea that the MCP protein shell is capable of mediating selective diffusion has lacked a clear experimental basis. Furthermore, a recent alternative model for MCP function proposes that enzymes embedded in or tightly associated with the shell could move metabolites into MCPs by vectorial catalysis, in which case functional pores might not be required for metabolite movement (24).Open in a separate windowFig. 1.Structure and function of the propanediol utilization (Pdu) bacterial microcompartment. A few thousand shell proteins (mostly of the BMC family) encapsulate a series of enzymes for metabolizing 1,2-PD. The protein shell of the Pdu MCP has been hypothesized to be selectively permeable allowing substrates such as 1,2-PD to enter through small pores in the center of hexameric shell proteins while restricting the efflux of propionaldehyde, which is toxic to the cell. For clarity, the reaction scheme has been simplified by omitting some of the steps involved in coenzyme B₁₂ recycling.Here, we use the known structure of PduA—a canonical, hexameric BMC-type shell protein in the Pdu MCP—to design a series of mutant shell proteins having central pores with altered sizes and physicochemical properties. A combination of physiological studies on mutant bacteria, biochemical studies on isolated mutant MCPs, and crystal structure studies on the mutant shell proteins, show that the PduA pore serves as a key route for entry of the metabolic substrate (1,2-PD), and that the chemical properties of the PduA pore are tuned to limit the escape of the toxic propionaldehyde intermediate.