Biological membranes are a defining feature of life and essential to compartmentalize biochemical processes. In eukaryotic cells, the delimiting membranes of organelles are composed of hundreds, often thousands of proteins and lipids. Despite their complex compositions and a constant exchange of membrane material, these organelles maintain characteristic properties, which determine organelle identity and warrant ‑ultimately‑ organelle function and cell survival. Biological membranes are responsive materials with unique properties: at once robust and rigid, yet also fluid, self-repairing, and selectively permeable. Upon dietary perturbation and environmental changes, they can readily adapt their composition whilst maintaining organelle-specific properties. The molecular rules and machines underlying membrane responsiveness, however, remain poorly understood.
- How does a cell maintain the characteristic properties and composition of their organelles?
- How do they membranes adapt during development, stress and ageing?
- What is the contribution of aberrant membrane properties to the pathogenesis of complex diseases such as type II diabetes or Alzheimer’s?
One reason that attempts to answer such questions have fallen short until now is that exploring these complex systems requires a multi-disciplinary approach, which is seldom available in a single lab and which takes years to establish.
Our approach and vision
Our lab has developed an experimental pipeline to dissect the structural dynamics and cellular functions of membrane property sensors in vivo and in vitro using interdisciplinary approaches including genetics, biochemistry, biophysics, and ‑in collaboration‑ theory and simulation. Our vision is deciphering the code, by which organellar membranes communicate their identity and execute cell fate decisions. We combine bottom-up and top-down approaches to tackle complex biological questions related to the interplay of protein folding and lipid metabolism in healthy and stressed cells. We seek to identify new membrane property sensors and to establish their mode-of-action. We have already learned how cells interrogate specific physicochemical properties of the endoplasmic reticulum (ER) membrane to control complex behavior with vast impact on cell physiology.
A systems-view on protein folding, lipid metabolism, and membrane biogenesis
As point of departure, we performed systematic genetic screens to identify critical connections between lipid metabolism, membrane traffic, and protein homeostasis (Surma et al. Mol. Cell 2013). This approach did not only establish the importance of crosstalk between protein turnover and lipid metabolism on the systems level, it also suggested that membrane biogenesis is controlled by just a few, specialized sensor proteins, which control large transcriptional programs. The observation that aberrant lipid compositions upregulate the production of ER chaperones provided a radically new view onto the role of biological membranes in controlling cell physiology.
Following up on these findings, we have established the molecular mechanisms of three sense-and-response systems, each controlling a key decision in membrane biology. The first system senses and controls the level of saturated and unsaturated lipids in cellular membranes (Surma et al. Mol. Cell 2013; Covino et al. Mol. Cell 2016; Ballweg et al. Nat. Comm 2020). The second system is the unfolded protein response (UPR) that senses lipid imbalances and unfolded proteins in the endoplasmic reticulum (ER) to balance the relative rates of protein and lipid production (Halbleib et al. Mol. Cell 2017). The third system senses a class of signaling lipids to orchestrate the cellular decision between membrane biogenesis and fat storage (Hofbauer et al. JCB 2018).
These sense-and-response systems use three remarkably distinct modes of interacting with the membrane. One senses at their surfaces, the second within the membrane, and the third senses across the membrane by squeezing it (Covino et al. Mol. Cell 2018; Ernst et al. Curr Opin. Cell Biol. 2018). But despite their distinct mechanisms and although they do not physically interact with each other, they are functionally connected via the ER membrane and form an integrated network (Covino et al. Mol Cell 2018).
I am certain, that these findings were only the beginning of a long scientific journey. Other organelles have different functions and properties from the ER, yet they face the same challenge: maintaining identity despite a constant exchange of membrane material with other organelles at steady state, during differentiation, stress, and upon dietary perturbation. We have shown that carefully designed genetic screens and bioinformatic approaches bear the potential to identify new sensor proteins surveilling membrane properties and they are readily adapted to other organelles such as the peroxisomes and mitochondria (Ernst et al. Curr. Opin. Cell Biol, 2018).
Molecular mechanisms of UPR activation by lipid bilayer stress
The unfolded protein response (UPR) is a large transcriptional program conserved from yeast to man. More than 5% of all genes are regulated by the UPR in S. cerevisiae. The UPR of metazoans is crucial for the differentiation and activity of secretory cells. Professionally secretory cells such as antibody-producing plasma cells or insulin producing β-cells are particularly dependent on the UPR to cope with high secretory demand. Originally identified as stress response to accumulating unfolded proteins in the lumen of the ER, it has also been speculated that aberrant lipid compositions can act as potent activators of the UPR. However, despite an obvious relevance to health and disease, the underlying mechanism remained unexplored. We addressed this major open question in the field and elucidated the underlying mechanism in a ground-breaking publication (Halbleib et al. Mol. Cell 2017), which was highlighted in Nat. Rev. Mol. Cell Biol. and Trends in Cell Biol. We could show that the highly unusual transmembrane domain of most conserved transducer of the UPR, namely IRE1, causes a local squeezing of the bilayer thereby rendering the core UPR transducer sensitive to aberrant lipid compositions and bilayer properties.
Towards crowding and folding sensors
All cells must balance the production of proteins and lipids to maintain membrane functions. Imbalances in protein folding and lipid metabolism cause ER stress associated with viral infections and a wide range of complex diseases including type II diabetes and neurodegeneration. The central homeostatic program of the ER is the unfolded protein response (UPR), which senses unfolded proteins in the ER to control protein synthesis, chaperone abundance, and lipid metabolism. Through these mechanisms, the UPR counteracts ER stress and controls decisions between cell survival, adaptation, and apoptosis. For years, the field focused on soluble proteins as triggers of the UPR, while the more abundant membrane proteins were overlooked. Our finding of UPR activation by membrane aberrancies therefore provided a radically new perspective. At the forefront of this field, we address fundamental questions in membrane biology in our MemDense project, which is funded by anERC Consolidator Grant.
Biological membranes are crowded with proteins, and their physicochemical properties are greatly affected by the protein-to-lipid ratio. The lateral diffusion of membrane proteins, for example, is >10-fold slower in crowded, biological membranes compared to model membranes. This suggests that free lipids, which act as a solvent for membrane proteins, can become limiting. In fact, a typical ER membrane protein is surrounded by ~40 lipids per bilayer leaflet, such that any further increase in protein density is likely to cause substantial problems. How is the crowding of membrane proteins in the ER (irrespective of their folding status) sensed and controlled? The unusual transmembrane architecture of the UPR stress sensors IRE1 and PERK render them sensitive to such membrane aberrancies and much more than unrelated proteins.
A wide array of diseases is associated with chronic ER stress, including type II diabetes and non-alcoholic fatty liver disease. Even though these diseases show signatures of a vastly perturbed lipid metabolism, a role of lipids in perpetuating ER stress has never been tested. We propose that chronic ER stress is the manifestation of a vicious circle: increased lipid saturation induces the UPR, which upregulates membrane lipid biosynthesis and causes the production of even more saturated lipids. If not properly regulated, this positive feedback loop will lead to dramatically increased levels of saturated lipids, the formation of non-fluid, gel phases in the ER membrane, membrane protein crowding in the remaining fluid regions of the ER, and ultimately cell death. Consistent with this idea, we could already provide definitive evidence that signals from the membrane can set off such positive feedback loop of deleterious UPR activation. Currently, we investigate, how precisely aberrant lipid compositions in the ER can switch the UPR from adaptive and beneficial to chronic and cyto-toxic. Our goal is to device specific, membrane-based therapies to counteract chronic ER stress.
The folding of membrane proteins and their assembly into macromolecular membrane complexes relies on the correct association of transmembrane helices (TMHs) in the hydrophobic core of the ER membrane. In analogy to misfolded soluble proteins, a misfolded membrane protein exposes hydrophilic residues to the lipid environment. We thus ask: How are misfolded membrane proteins selectively recognized by UPR effectors?
Our ultimate vision is to establish the origins of acute and chronic ER stress at the molecular and cellular level. By identifying and dissecting the contributions of membrane proteins and lipids to UPR activation, we expect nothing less than a paradigm shift to the current understanding of complex, metabolic diseases related to ER stress. Even if some of our ambitious goals may take longer than expected, we will provide unprecedented insight into the inner workings of the ER stress response and develop new tools and resources of sustained value for the scientific community.
Eukaryotic strategies to sense membrane lipid saturation
All organisms that cannot control their body temperature must adapt their membrane lipid composition to the ambient temperature, a phenomenon referred to as homeoviscous adaptation. Although the theory of homeoviscous adaptation is more than 40 years old, key questions regarding the machinery for sensing and responding to changes in membrane properties remain unresolved, especially for eukaryotes (Ernst et al. JMB 2016). The simple eukaryote S. cerevisiae encodes only a single, essential fatty acid desaturase named Ole1, whose activity is crucial to maintain membrane fluidity. We focused our research on the membrane-bound transcription factor Mga2. It uses its single, C-terminal transmembrane helix to sense the molecular packing of lipids in the membrane of the ER (Surma et al. Mol. Cell 2013; Covino et al. Mol. Cell 2016; Ballweg et al. Nat. Comm 2020).
Collaborating with experts in molecular dynamics (MD) simulations, we established a pipeline to study intra-membrane sensing processes in vivo, in vitro and in silico. We could show that Mga2 forms pre-assembled, membrane-bound homo-dimers. The transmembrane helixes in these dimers rotate against each other with their relative rotational orientations being particularly sensitive to the molecular packing of the lipid environment.
A loosely packed environment (i.e. enriched in lipids with unsaturated acyl chains) favors the OFF-state that hinders downstream transcription factor activation, while a tightly packed membrane environment stabilizes the alternative and productive ON-conformation. We identified in the transmembrane region of these proteins unusual sequence elements with physicochemical features that are crucial for intra-membrane sensing and that prompt us to study putative lipid packing sensors in other organelles and higher eukaryotes in the future.