Final Report Summary - TRANSPORTER FUNCTION (Mass spectrometry of structural dynamics in secondary membrane transporters)
Introduction
Secondary active transporters are found in all kingdoms of life, reflecting their general importance for cellular function. Common to all proteins in this group is the ability to transport ions, drugs, or nutrients across the cell membrane using the existing sodium gradient between the extracellular and intracellular milieu as driving force. Due to their involvement in a range of cellular processes, secondary active transporters represent promising drug targets. However, little is known about their molecular structure, transport cycle, and interactions with candidate drugs.
The current project is focused on the use of MS-based strategies to study the conformational dynamics of secondary transporters. The first phase has expanded the MS methodology, which in the second phase was used to compare homologous secondary transporters and investigate whether a common structural framework for sodium-driven transport across the membrane can be established, with consequences for other membrane protein systems.
To achieve this, we employ a range of mass spectrometry (MS)-based tools in combination with computational modelling. Recently, studies from our lab have revealed that the combination of ion mobility MS, which informs about the overall architecture of a protein complex, with collisional unfolding, which informs about complex stability, can be used to uncover structural and functional roles of lipid interactions.
Results
We selected the sodium/proton exchangers NapA from Th. thermophilus and NhaA from E. coli as models for well-characterised secondary transporters. It has been proposed that NapA-like proteins in the sodium/proton antiporter family undergo large conformational shifts as part of their catalytic cycle, which is facilitated by their dimeric architecture. NhaA, which is currently used as model system for human sodium/proton exchange, is proposed to require only small conformational changes and is not critically dependent on dimerization. The pronounced differences lead us to believe that both classes of antiporters use membrane lipids in different ways to support their architecture and function. It is highly debated which of these two structural models applies to the human sodium/proton exchanger NHA2, a highly relevant drug target in hypertension.
To be able to perform comparative MS of these Na+/H+ transporters, we developed a common MS approach to analyse NapA, NhaA, and NHA2 under identical MS conditions. Using the same instrument settings and solution conditions, we then performed high-resolution MS analysis of all three proteins. The data clearly indicate that the proteins fall into two groups: NapA and NHA2 form stable dimers and retain virtually no lipids after purification, while NhaA dimers dissociate very easily and retain cardiolipin with high selectivity. The stark difference in dimer stabilities was then investigated using sequence analysis and homology modelling, which showed that NapA and NHA2, but not NhaA, share a 13-helix architecture with an additional helix that mediates extensive subunit contacts. NhaA, on the other hand, forms a 12-helix bundle with a very weak dimer interface that lacks the additional helix.
Using native MS, MS/MS, and ion mobility (IM-)MS, we established that the lipid-free NapA dimer tightly binds negatively charged phospholipids, while zwitterionic and uncharged lipids are bound weakly and provide stabilization against unfolding. The comparative MS analysis of NapA, NHA2, and NhaA suggests that lipid interactions in NapA are largely non-selective apart from head-group charge, while NhaA specifically requires cardiolipin for its function. We therefore conducted extensive MD simulations of both proteins in model bilayers and compared the resulting complexes to insights from ion mobility MS studies of both proteins with and without lipids. For this purpose, we developed a novel calibration protocol for travelling wave ion mobility measurements to accurately measure collision cross-sections (CCS) of low-charge membrane proteins. Briefly, the low charge of membrane proteins reduces their mobility in the gas phase, which is not accurately reflected by more highly charged soluble proteins commonly used for calibration. We found that calibration with larger proteins than the analyte can mitigate this effect, greatly improving the accuracy of CCS measurements. With the modified calibration protocol, we were able to connect MD simulations, theoretical CCS values, and IM-MS data.
Application of this approach to NapA revealed that non-selectively bound lipids preferentially interact with charged residues that are located at the interface between core and dimer domains, giving rise to stabilization in the gas phase and likely providing a form of “molecular grease” for domain movements during the catalytic cycle.
Next, we applied the same approach to the complex between NhaA and the selectively bound cardiolipin molecule. By MS, we observed that dimeric NhaA retains cardiolipin more easily than monomeric NhaA. In line with this, collision-induced dissociation of dimeric NhaA instantly releases cardiolipin. We then performed bilayer MD simulations of NhaA and found that cardiolipin binds to two basic clusters at the dimer interface, essentially bridging the dimer subunits and giving rise to dimer stabilization in solution and in the gas phase.
Based on the different selectivities for functional lipids displayed by two homologous secondary active transporters, we were able to formulate a sliding scale model for lipid selection in membrane proteins.
We reasoned that the observed stabilization of the NhaA dimer by selective lipid binding could represent a general phenomenon for proteins with weak dimer interfaces. Working closely with Dr Kallol Gupta, Joseph Donlan, and Dr Jonathan Hopper in our group, we compared lipid binding and dimer stabilities in two additional protein systems predicted to have weak dimer interfaces, namely the leucine transporter LeuT and the sugar transporter SemiSWEET. We found that all three proteins exhibit selective cardiolipin binding in native MS experiments and bilayer MD simulations, in all cases promoting a more stable dimer association. Our results show that membrane transporters are able to selectively bind membrane lipids to modulate their oligomeric stability. In summary, the present project has provided an exciting new view on the biological function of secondary transporters.
Secondary active transporters are found in all kingdoms of life, reflecting their general importance for cellular function. Common to all proteins in this group is the ability to transport ions, drugs, or nutrients across the cell membrane using the existing sodium gradient between the extracellular and intracellular milieu as driving force. Due to their involvement in a range of cellular processes, secondary active transporters represent promising drug targets. However, little is known about their molecular structure, transport cycle, and interactions with candidate drugs.
The current project is focused on the use of MS-based strategies to study the conformational dynamics of secondary transporters. The first phase has expanded the MS methodology, which in the second phase was used to compare homologous secondary transporters and investigate whether a common structural framework for sodium-driven transport across the membrane can be established, with consequences for other membrane protein systems.
To achieve this, we employ a range of mass spectrometry (MS)-based tools in combination with computational modelling. Recently, studies from our lab have revealed that the combination of ion mobility MS, which informs about the overall architecture of a protein complex, with collisional unfolding, which informs about complex stability, can be used to uncover structural and functional roles of lipid interactions.
Results
We selected the sodium/proton exchangers NapA from Th. thermophilus and NhaA from E. coli as models for well-characterised secondary transporters. It has been proposed that NapA-like proteins in the sodium/proton antiporter family undergo large conformational shifts as part of their catalytic cycle, which is facilitated by their dimeric architecture. NhaA, which is currently used as model system for human sodium/proton exchange, is proposed to require only small conformational changes and is not critically dependent on dimerization. The pronounced differences lead us to believe that both classes of antiporters use membrane lipids in different ways to support their architecture and function. It is highly debated which of these two structural models applies to the human sodium/proton exchanger NHA2, a highly relevant drug target in hypertension.
To be able to perform comparative MS of these Na+/H+ transporters, we developed a common MS approach to analyse NapA, NhaA, and NHA2 under identical MS conditions. Using the same instrument settings and solution conditions, we then performed high-resolution MS analysis of all three proteins. The data clearly indicate that the proteins fall into two groups: NapA and NHA2 form stable dimers and retain virtually no lipids after purification, while NhaA dimers dissociate very easily and retain cardiolipin with high selectivity. The stark difference in dimer stabilities was then investigated using sequence analysis and homology modelling, which showed that NapA and NHA2, but not NhaA, share a 13-helix architecture with an additional helix that mediates extensive subunit contacts. NhaA, on the other hand, forms a 12-helix bundle with a very weak dimer interface that lacks the additional helix.
Using native MS, MS/MS, and ion mobility (IM-)MS, we established that the lipid-free NapA dimer tightly binds negatively charged phospholipids, while zwitterionic and uncharged lipids are bound weakly and provide stabilization against unfolding. The comparative MS analysis of NapA, NHA2, and NhaA suggests that lipid interactions in NapA are largely non-selective apart from head-group charge, while NhaA specifically requires cardiolipin for its function. We therefore conducted extensive MD simulations of both proteins in model bilayers and compared the resulting complexes to insights from ion mobility MS studies of both proteins with and without lipids. For this purpose, we developed a novel calibration protocol for travelling wave ion mobility measurements to accurately measure collision cross-sections (CCS) of low-charge membrane proteins. Briefly, the low charge of membrane proteins reduces their mobility in the gas phase, which is not accurately reflected by more highly charged soluble proteins commonly used for calibration. We found that calibration with larger proteins than the analyte can mitigate this effect, greatly improving the accuracy of CCS measurements. With the modified calibration protocol, we were able to connect MD simulations, theoretical CCS values, and IM-MS data.
Application of this approach to NapA revealed that non-selectively bound lipids preferentially interact with charged residues that are located at the interface between core and dimer domains, giving rise to stabilization in the gas phase and likely providing a form of “molecular grease” for domain movements during the catalytic cycle.
Next, we applied the same approach to the complex between NhaA and the selectively bound cardiolipin molecule. By MS, we observed that dimeric NhaA retains cardiolipin more easily than monomeric NhaA. In line with this, collision-induced dissociation of dimeric NhaA instantly releases cardiolipin. We then performed bilayer MD simulations of NhaA and found that cardiolipin binds to two basic clusters at the dimer interface, essentially bridging the dimer subunits and giving rise to dimer stabilization in solution and in the gas phase.
Based on the different selectivities for functional lipids displayed by two homologous secondary active transporters, we were able to formulate a sliding scale model for lipid selection in membrane proteins.
We reasoned that the observed stabilization of the NhaA dimer by selective lipid binding could represent a general phenomenon for proteins with weak dimer interfaces. Working closely with Dr Kallol Gupta, Joseph Donlan, and Dr Jonathan Hopper in our group, we compared lipid binding and dimer stabilities in two additional protein systems predicted to have weak dimer interfaces, namely the leucine transporter LeuT and the sugar transporter SemiSWEET. We found that all three proteins exhibit selective cardiolipin binding in native MS experiments and bilayer MD simulations, in all cases promoting a more stable dimer association. Our results show that membrane transporters are able to selectively bind membrane lipids to modulate their oligomeric stability. In summary, the present project has provided an exciting new view on the biological function of secondary transporters.