The folding mechanisms of helical membrane proteins remain largely uncharted. Here we examine two of the primary events during membrane protein folding: insertion/topology acquisition into the lipid bilayer and the assembly of alpha helices into their native state. Towards this goal we characterize the kinetics of bacteriorhodopsin folding and employ φ-value analysis to explore the folding transition state. We were able to obtain reliable φ-values for 16 mutants of bacteriorhodopsin well distributed throughout the protein. Every φ-value was less than 0.4, indicating the transition state is not uniquely structured. We suggest that the transition state is a loosely organized ensemble of conformations. Before the final folding of the alpha helices, the transmembrane segments must be properly aligned in the lipid bilayer. The topology of helical membrane proteins is generally defined during insertion of the transmembrane helices. It is now clear, however, that topology can remain malleable in some cases. We show that the entire EmrE protein can indeed invert topology after the full protein is translated and inserted. Wholesale topology flipping is remarkable given the physical constraints of the membrane and exposes new avenues to envision folding pathways, both productive and detrimental.

Membrane proteins are a neglected, but important class of proteins throughout the biological world. They carry out critical roles in the cell due to their unique location, such as transport across a membrane, transduction of exterior signals, and interaction between discrete aqueous regions. Despite the importance of these proteins, understanding of how they fold has lagged far behind that of soluble proteins. One of the primary challenges to studying membrane protein folding is developing methods that interrogate folding in the native environment of the lipid bilayer. Our lab has developed a method for measuring membrane protein stability under native conditions using a secondary protein that preferentially binds the unfolded state, obviating the need for harsh denaturants. Employing this method with a multimeric polytopic membrane protein, we measured an extremely slow unfolding rate, demonstrating that α-helical membrane proteins can have high kinetic stability under non-denaturing conditions. Efforts were made to expand the steric trap method for single-molecule fluorescence measurements in lipid vesicles, but were ultimately stymied by the inability to preserve the trapped complexes for measurement. Our lab has also applied single-molecule techniques to membrane protein folding. We were able to map the energy landscape of a membrane protein in a lipid bilayer using forced unfolding driven by magnetic tweezers. Further advancements to this technique simplified the attachment chemistry to ready the protein for tweezing. These techniques can be applied to a wide array of membrane proteins in a broad spectrum of membrane environments.

Cystic kidney diseases including polycystic kidney disease (PKD) and nephronophthisis (NPHP) are the most common genetic disorders leading to end-stage renal failure in humans. Animal models and human cases of PKD and NPHP have implicated the sterile alpha motif (SAM) domain containing proteins bicaudal C homolog 1 (BICC1) and ankyrin repeat and SAM-domain containing protein 6 (ANKS6) as being involved in these conditions and important for renal development. SAM domains are known protein-protein interaction domains that are capable of binding each other to form polymers and heterodimers. Using a negGFP native gel assay, we have identified the SAM domain of the previously uncharacterized protein ankyrin repeat and SAM-domain containing protein 3 (ANKS3) as a direct binding partner of the BICC1 and ANKS6 SAM domains. We found the ANKS3 SAM domain to polymerize with moderate affinity and determined the ANKS6 SAM domain can bind to a single end of this polymer. Crystal structures of the ANKS3 SAM domain polymer and the ANKS3 SAM-ANKS6 SAM heterodimer are presented to reveal typical ML-EH SAM domain interaction interfaces with a pronounced charge complementarity. A crystal structure of the BICC1 SAM domain, which forms a helical polymer with moderate affinity, is also presented. The SAM domains of BICC1 and ANKS3 exhibit high sequence and structural homology, allowing these SAM domains to bind each other using both of their conserved ML and EH interaction surfaces. The measured binding affinities of the two possible interfaces between the BICC1 and ANKS3 SAM domains are nearly identical, suggesting these SAM domains may associate to form the first observed alternating SAM domain co-polymer. The BICC1 and ANKS6 SAM domains are also shown to bind each other as a heterodimer with strong affinity. The R823W and I817N point mutations in the SAM domain of ANKS6 are responsible for cystic kidney disease in the PKD/Mhm(cy/+) rat and a newly identified mouse model, respectively. We show that these point mutations dramatically destabilize the ANKS6 SAM domain and are responsible for a loss of interaction with the SAM domains of ANKS3 and BICC1, suggesting these lost interactions are part of a possible disease mechanism. The network of interactions between the SAM domains of BICC1, ANKS3, and ANKS6 may allow regulation of polymer formation and the creation of diverse cellular scaffolds that are suggested to be important for renal development. This work provides a structural and biochemical foundation for the continued investigation of how the interactions between the BICC1, ANKS3, and ANKS6 SAM domains mediate cellular development.

To produce more diversified and more valuable commodities from a biobased economy, I have sought to upgrade ethanol into a wider range of products through cell-free platform technologies. I have built two styles of ethanol valorization system, one through acetaldehyde condensation, another through acetyl-CoA. Also, I have developed methods to power biochemical systems, one that converts electrical power into reducing equivalents and another that can regenerate ATP while co-producing acetone. I developed a 1,3-BDO production system that reached a maximum 0.16 g/L/h productivity, and, with replenishment of feedstock and enzymes, achieved a 1,3-BDO titer of 7.7 g/L. I also developed an isoprenol production system able to reach a maximum 1.0 � 0.05 mM/h productivity and 12.5 � 0.8 mM maximum titer. My ATP generator reached titers of 27 � 6 mM ATP and 59 � 15 mM acetone with maximum ATP synthesis rate of 2.8 � 0.6 mM/hour and acetone of 7.8 � 0.8 mM/hour. These results demonstrate a proof-of-concept blueprint for more diversified bioeconomy for commodity chemical productions.

Over the past century, fossil fuels have been an abundant and cheap source for petroleum-derived chemicals and fuels. The production and burning of these carbon-derived fuels have led to adverse global environmental changes such as increase in air pollution, climate change and fluctuations in sea level. To replace dwindling petroleum resources and to curb emissions of CO2, it is critical to develop alternative and renewable resources for energy and fuels. Genetically engineered microorganisms that can directly produce medium and long chain hydrocarbons have been one of the most promising potential routes to renewable biofuel synthesis.

Microbial fatty acids are an attractive source of precursors for a variety of renewable biofuels such as alkanes, alcohols, and biofuels. Enormous progress has been in engineering microbes to divert endogenous fatty acid synthesis and overproduce free fatty acids. However, there is an inherent problem of product toxicity that greatly reduces cell viability, increases cell lysis and product titers and there are previous reports suggesting membrane damage as the main mechanism of free fatty acid toxicity. In this work, we metabolically engineered Escherichia coli (E. coli) bacteria to overproduce medium chain free fatty acids and identified membrane stress as the leading factor in product toxicity. We found that membrane lipid composition can be altered by the direct incorporation of endogenously produced medium-chain fatty acids into lipids via the Aas pathway. The deletion of the aas gene and sequestering exported fatty acids reduces medium-chain fatty acid toxicity, partially restores normal lipid composition, and dramatically improves medium-chain fatty acid yields.

In the second part of this thesis, we genetically engineered E. coli to produce fatty acid methyl esters (FAMEs) by direct in vivo methylation of free fatty acid in the strains discussed above. Insect Drosophila melanogaster Juvenile Hormone Acid O-Methyltransferase (DmJHAMT) was identified as a candidate to methylate a variety of endogenous medium chain fatty acids in E. coli. By introducing DmJHAMT in E. coli engineered to produce medium chain fatty, we obtain medium chain FAMEs at titers of 0.56 g/L, more than two orders of magnitude higher than titers previously achieved. This one-step conversion process was optimized by expression of rat Mat1A gene that increased the S-adenosyl-L-methionine cofactor pool and providing a physical sink to extract FAMEs from culture.

The work presented here shows the viable method of producing microbial biodiesel by metabolical engineering of E. coli. Primary physiological stress associated with production of free fatty acid precursors was identified and higher titers of fatty acid production were reported. These free fatty acids were then converted to biodiesel in one step by expressing insect enzyme in E. coli. Although further work is needed for viable bacterial production of biodiesel, the simplicity of the pathway allows easier optimization and possibility to transport this pathway into photosynthetic microorganisms in the future.

The intricate functions of membrane proteins would not be possible without bends or breaks that are remarkably common in transmembrane helices. The frequent distortions are nevertheless surprising because backbone hydrogen bonds should be strong in an apolar environment, potentially rigidifying helices. It is therefore mysterious how distortions can be generated by evolutions. Through my studies on bacteriorhodopsin and Ca2+-ATPase, I found that helix distortions are facilitated by shifting hydrogen bonding partners. My results explained how evolution has been able to liberally exploit transmembrane helix bending by for optimizing membrane protein structure, function and dynamics.

It has recently been found that the stability of bacteriorhodopsin assessed through unfolding did not achieve equilibrium. Here I made a new equilibrium test by measuring the transition of folded bacteriorhodopsin to unfolded bacterioopsin. My result suggests that the energetic effects of most mutations that have been studied before may be focused on the folded state.

To further investigate how hydrogen bonds may shape transmembrane helices, I sought a non-perturbing method to measure strengths of both backbone and side-chain hydrogen bond strengths, because backbone hydrogen bonds cannot be probed by mutation and mutagenesis to remove side-chains may not mimic the breaking of side-chain hydrogen bonds. I therefore decided to employ the equilibrium hydrogen / deuterium fractionation factors (phi-value) for hydrogen bonds. I used model compounds to study the relationship between phi-value and the hydrogen bond free energy. By applying this relationship, I found that hydrogen bonding stabilizes enzyme intermediate state by up to 4 kcal/mol more than the resting state and marginally favor soluble-protein unfolding.

To measure hydrogen bond strengths in a membrane protein, I focused on the isolated voltage-sensor domain of the voltage-dependent potassium-selective channel. By utilizing its 2D NMR spectrum, I obtained phi-values and thus free energies for 70 % of the backbone hydrogen bonds in this membrane protein. I found that it has similar backbone hydrogen bond strength to water-soluble proteins on average. Moreover, I found that the flexible (dynamic) point of a transmembrane helix in this protein can be predicted from the backbone hydrogen bond strengths in that helix.

Over the last century the petro chemical industry has provided an abundant and cheap source of hydrocarbons that have impacted and transformed many facets of our lives. Petrochemicals not only provide fuel that revolutionized transportation industry but also provided cheap petrochemical feedstock molecules that form a basis of many textiles, plastics, adhesives, detergents, and lubricants that are indispensable in modern life. Our dependence on the petrochemical industry and the sheer quantity of petrochemical based products that are consumed have led to a number of adverse and unintended effects on our environment. Recent attention to unintended effects of the petrochemical industry such as climate change, air pollution, and landfill overflow have sparked renewed interest in green chemistry and research in renewables that could potentially decrease our dependence on the petrochemical industry.

Over the last two decades, with the advent of simple and cheap genetic engineering, a new field called metabolic engineering has emerged. A large focus of this field is centered around re-engineering the metabolism of simple organisms to produce bulk and specialty chemicals. We genetically engineered a lithoautotrophic organism, Ralstonia eutropha, to produce drop-in ready biofuels (isobutanol and 3-methyl-1-butanol) in an electro-bio reactor using CO2 as the sole carbon source and electricity as the sole energy input. This method integrates electrochemical formate production and biological CO2 fixation and higher alcohol synthesis. The liquid fuels generated are a relatively stable way to store energy and possess energy densities about 100 times higher than current-day batteries.

A different approach to metabolic engineering coined synthetic biochemistry seeks to reconstitute complex metabolic pathways outside of cells. The Synthetic Biochemistry approach eliminates the myriad unwanted side reactions that occur in cells and product toxicity, enabling near 100% theoretical yields. Unfortunately, eliminating the cellular environment also eliminates key regulatory networks involved in maintaining appropriate carbon flux and cofactor balance. Synthetic biochemistry seeks to replace these complex regulatory networks in vitro by creating self-balancing pathways that automatically regulate the production and consumption of cofactors while maintaining constant carbon flux through the pathway. To this end, we developed the concept of a molecular purge valve. Generally, the purge valve is made up of a NADPH dependent oxidase, a NADH dependent oxidase and a NADH specific oxidase (noxE). We first demonstrated this proof of concept by engineering the G. stearothermophilus pyruvate dehydrogenase complex to flip its cofactor specificity from NADH to NADPH and using it in the purge valve to transform pyruvate to either polyhydroxybutyrate bioplastic or isoprene at near quantitative yields. The concept of using a purge valve to maintain carbon flux in an inherently imbalanced pathway was expanded through the production of polyhydroxybutyrate from glucose using a synthetic pathway called the PBG cycle. The PBG cycle is made up of enzymes from the pentose phosphate pathway, the bifido shunt , and standard glycolysis (EMP). This pathway utilizes the purge valve concept at the two reductase enzymes of the pentose phosphate pathway and produced the bioplastic polyhydroxybutyrate from glucose. The PBG pathway ran semi-continuously for 55 hours and produced 86% yield from 110 mM glucose. The pathway had a maximum productivity of 0.7g/L PHB and maintained 50% of its productivity over the course of the 55 hours at room temperature. The PBG cycle serves as a demonstration that synthetic biochemistry could be an alternative to in vivo metabolic engineering for the production of bulk and specialty chemicals and could potentially be one way to replace some of the petrochemicals we use today with renewable bio-based chemicals.

Identifying polymer-forming SAM domains

Sterile Alpha Motif (SAM) domains are common protein modules in eukaryotic cells. It has not been possible to assign functions to uncharacterized SAM domains because they have been found to participate in diverse functions ranging from protein-protein interactions to RNA binding. Here we computationally identify likely members of the subclass of SAM domains that form polymers. Sequences were virtually threaded onto known polymer structures and then evaluated for compatibility with the polymer. We find that known SAM polymers score better than the vast majority of known non-polymers: 100% (7 of 7) of known polymers and only 8% of known non- polymers (1 of 12) score above a defined threshold value. Of 2901 SAM family members, we find 694 that score above the threshold and are likely polymers, including SAM domains from the proteins Lethal Malignant Brain Tumor, Bicaudal-C, Liprin-beta, Adenylate Cyclase and Atherin. In polymerization experiments, all of these predictions (except Adenylate Cyclase) were confirmed. As a result, the original SAM database was updated and additional predictions were obtained.

TMKink: A method to predict transmembrane helix kinks

A hallmark of membrane protein structure is the large number of distorted transmembrane helices. Because of the prevalence of bends, it is important to not only understand how they are generated but also to learn how to predict their occurrence. Here, we find that there are local sequence preferences in kinked helices, most notably a higher abundance of proline, which can be exploited to identify bends from sequence information. A neural network predictor identifies over two-thirds of all bends (sensitivity 0.70) with high reliability (specificity 0.89). It is likely that more structural data will allow for better helix distortion predictors with increased coverage in the future. The kink predictor, TMKink, is available at http://tmkinkpredictor.mbi.ucla.edu/.

Structural differences between mesophilic and thermophilic membrane proteins

Protein thermostability remains a focal point of interest for protein scientists. The differences in thermostability between mesophilic and thermophilic soluble proteins have been extensively studied. No differences in packing values have been found in soluble proteins. Membrane protein packing is different from soluble protein packing; thermophilic adaptation may be different as a result. Surprisingly, burial and packing values appear to be shared between mesophiles and thermophiles in both soluble and membrane proteins. We created a non-redundant database of unpaired and paired structures for the study of thermophile-mesophile structural differences in membrane proteins. We found little or no differences in burial or packing values in both the soluble and transmembrane regions of membrane proteins.