University of Groningen
Single-molecule fret study on structural dynamics of membrane proteins Aminian Jazi, Atieh
DOI:
10.33612/diss.135802718
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Aminian Jazi, A. (2020). Single-molecule fret study on structural dynamics of membrane proteins. University of Groningen. https://doi.org/10.33612/diss.135802718
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
81
Chapter 4:
smFRET study on the potassium-induced
conformational changes of the C-terminal
sensory domain in the betaine transporter
82
Molecular insight into the potassium-induced pre-activation in the
osmosensing betaine symporter BetP
Camilo Perez*, Atieh Aminian Jazi*, Rebecca M. Gärtner, Stanislav Maksimov, Reinhard Krämer, Thorben Cordes and Christine Ziegler (Manuscript in preparation)
Abstract
The trimeric betaine symporter BetP from Corynebacterium glutamicum senses hyperosmotic stress via a 45 amino acid long C-terminal domain. Biochemical data suggested a molecular switch model, in which the positively charged osmosensor is detached from the negatively charged membrane following a hyperosmotic upshift. To date it is still unknown which stimuli are sensed by BetP under hyperosmotic conditions. However, activation can be mimicked by an increase of K+ concentration representing a non-physiological stimulus. In all crystal structures of BetP, the osmosensor in one protomer forms an extended helix protrud ing into the cytoplasm towards the adjacent protomer via crystal contacts. In the remaining two protomers the C-terminal domains are not resolved in the structures assuming that they are gradually unfolded. Here, we report on a crystal structure of BetP de termined at 300 mM Rb+, a heavy atom K+ mimic. This structure shows that K+ binding stabilizes the C-terminal helical fold, strengthens the intratrimeric interactions of the osmosensor with loop 2, and interferes with lipid-protein interactions. A mutagenesis study confirms the coordination of three K+ by residues in the C-terminal domain and loop 2 and reveal their impact on regulation. However, the main chain conformation and the partial unfolding of the osmosensor was similar to that observed in structures without K+ suggesting that crystallization traps a pre-activated state of BetP. We further performed a comparative study of BetP using single -molecule FRET to characterize the conformational flexibility of the C-terminal osmosensor. A comparison of smFRET data in detergent and the membrane-mimicking amphipol A8-35, demonstrates that in detergent-solubilized BetP, the osmosensor is indeed not reactive to K+. Amphipol-reconstituted BetP enables a potassium-dependent conformational change. As potassium
83
weakens the interaction of the osmosensor with negatively charged membrane lipids it induces a pre-activated conformation, which seems to be the pre-requisite for the full activation by hyperosmotic stress.
84
Introduction
Hyperosmotic stress-regulated bacterial transporters (also referred to as osmoregulated transporters) can sense osmotic stress and react by a sudden increase in transport rate to rapidly accumulate compatible solutes in order to prevent dehydration of the cytoplasm [1]. They are able to perceive and translate physical stimuli into a protein conformational change, which directly affects the transport rate. The stimuli caused by hyperosmotic stress are diverse ranging from ionic strength, crowding, curvature stress or lateral pressure. In the case of BetP, hyperosmotic stress is sensed by changes of the physical state of the membrane most likely transduced by lipid-protein interactions to the transporter affecting betaine transport rate. By investigating the molecular mechanism of osmoregulated membrane transport in BetP we want to decipher the direct cellular consequences of hyperosmotic stress on the membrane itself, which would have a tremendous impact on our understanding of bacterial stress adaptation. The sodium-coupled betaine symporter from the soil bacterium Corynebacterium glutamicum, is a prime example for a lipid-dependent osmoregulated transporter. It is specific for betaine, which is accumulated in molar amounts against the concentration gradient into the cytoplasm under hyperosmotic stress, i.e., above 0.6 Osmol/kg. For this BetP exploits the inward directed sodium gradient across the membrane to energize betaine transport.
BetP consists of 12 transmembrane helices (TM) and adopts the so-called LeuT fold. The transporter core comprises two inverted structural five-helix-repeats (TM3-TM12), which are intertwined (Figure 4.01). Several crystal structures of BetP (pdb-entry 2WIT, 3PO3, 4DLL, 4DOJ, 4C7R, 4AIN) reveal an asymmetric trimer in which each protomer can independently adopt distinct conformational states of the alternating access cycle [2]. A substrate-bound closed state (pdb-entry 4AIN) reveals the tight coordination of the betaine molecule in a Tryptophan-box (W373, W374, and W377 in TM8) by cation-pi interactions of the tri-methylammonium group. In addition, the carboxyl of betaine is coordinated by a main chain carbonyl group of Met150 in the so-called glycine stretch in TM3. Both structural elements are also involved in sodium binding although by other residues and together with residues from TM5 (Na1) and TM10 (Na2). Latter are part of the coupling scaffold that comprises also TM6-7 and TM11-12. These helices are only indirectly involved in sodium coordination via water molecules, but essential in driving the
85
conformational changes from outward- to inward-facing conformation upon substrate binding, and vice versa upon substrate release.
Figure 4.01| Topology of the BetP monomer. BetP is formed by two inverted repeats, which causes the formation
of distinct structural topology TM3-4 (repeat 1) and TM8-9 (repeat 2) form the “bundle-domain”, which harbors the Tryptophan-prism (TM8) and the glycine stretch (TM3). Both elements are involved in sodium and betaine coordination. TM5 (repeat 1) and TM11-12 (repeat 2) form one coupling scaffolds, which is symmetry related to TM10 and TM6-7. While TM5 and 10 contribute residues for sodium ion coordination, the two hairpins TM6 -7 and TM11-12 drive conformational changes upon substrate binding and release.
The N-terminal and C-terminal domains are oriented into the cytoplasm. The positively charged C-terminal domain of BetP is formed by 45 amino acids (aa) and senses the yet unknown stress stimuli, which is linked to the physical state of the membrane. It consists of two structural clusters. The 1st cluster from Y550-R568 is separated from the 2nd cluster (E572-R595) by a short linker.
In all crystal structures of the BetP trimer, the C-terminal domains of two protomers point towards the adjacent protomer (C-terminal domain of chain A to chain C, and C-terminal domain of chain C to chain B) where they interact with an aspartate residue in the cytoplasmic loops 2 (Asp131, Figure 4.02). Furthermore, the C-terminal domain of chain C is only resolved until R568, in chain B even only to R558, suggesting that the C-terminal domain is not always entirely helically
86
folded or is rotating around its anchor point on the trimer. The C-terminal-loop2 interaction is not restricted to R558(A,C)D131(C,B), but can involve other charged residues, depending on the folding degree of the C-terminal helix and the conformation of loop 2. It can be assumed that there is a partner-switching interaction between these two elements, which depends on the activation state of BetP.
Figure 4.02| C-terminal domains and their interactions with loop 2 in the BetP trimer. Chain A (exemplary shown
for pdb-entry 4DOJ) is per definition the protomer with the longest resolved C-terminal domain (550-583; last 12 aa still missing), chain C can be resolved until R568, and chain B until R558. Inset blue: The C-terminal domain of Chain A is interacting with loop 2 of Chain C, in which only the 1st charged cluster of the C-terminal domain is resolved. The interaction R558A-D131C is always present, while R558C-D131B was observed in 2WIT and 4C7R suggesting that it is an interaction, which is not forced by crystal contacts.
Loop 2 comprises several charged residues, which interact with residues in loop4 and 8, in the cytoplasmic tip of TM12 and the linker between TM12 and the C-terminal domain. Most
87
interestingly, R568 is the last residue in the C-terminal helix, which can still interact with loop 2, here with I130. In addition, residues in loop 2 coordinate chloride ions and lipids. Loop 2 precedes TM3, which is the key helix in coupling as it harbors both coordinating residues for sodium and betaine. Taking into account the paramount importance of this flexible helix segment f or coupling and transport in BetP, it can be assumed that one important key to BetP’s regulatory mechanism is the way how the C-terminal domain of one protomer is linked to the glycine stretch of the adjacent TM3 via the interaction network of loop 2, loop 8, lipids and cytoplasmic ions (Figure 4.04 A).
Figure 4.03| Loop2 interaction network in the BetP trimer (pdb-entry 4C7R). Loop 2 is interacting with loop 4, loop
8, TM12 and the C-terminal domains.
In the most symmetric crystal structure 4C7R, in which all three protomers adopt an inward-facing state, several lipid densities were observed and assigned as POPG lipids. This structure reveals how the cytoplasmic segment of TM3 that is preceding the glycine stretch interacts via lipids and TM8 with the C-terminal domain. It has to be noted that also M150 in the glycine stretch is coordinated by a lipid fatty acyl chain and therefore is part in the regulatory interaction network. In addition, loop4 connects both C-terminal domains, e.g. the one from chain A and chain C affecting again now the interaction network with chain B. As the crystal contacts require only one C-terminal domain (per definition the C-terminal domain of Chain A), chain C is always trapped in a specific inward facing conformation. In Chain A and chain B, intra-trimeric
88
interactions of loop 2, 4, and 8 are more variable, which is the reason for the conformational asymmetry observed in different crystal structures depending on the amount of substrate or the presence of mutants in the glycine stretch that alter sodium/betaine affinities. The lipid-protein interactions have to be considered in the context of conformational changes from inward-facing (IF) state to outward-facing state (OF). The comparison of the three crystal structures: 4C7R ( A- C; IF), 4DOJ (A,B; OF occluded states) and 4LLH (A,B;OF) reveals the role of the cytoplasmic part of TM3 (TM3c) in the cycling from IF to OF (Figure 4.04 B). TM3c moves around the glycine stretch towards the center of BetP between bundle and coupling scaffold. In addition, loop 11 connecting the periplasmic parts of TM11 and TM12 (coupling scaffold for Na1) together with loop 6 connecting the cytoplasmatic parts of TM6 and TM7 (coupling scaffold for Na2) move away from the center.
Figure 4.04| Lipid-interaction network of TM3 and TM8 and conformational changes in BetP in the isomerization from inward-facing (IF) to outward-facing (OF) state. (A) TM3 is connected to conformational changes of the
C-terminal domain via lipids. In addition, the glycine stretch is undergoing lipid-protein interaction by M150. (B) Conformational changes in BetP involve mainly TM3 (red) and the two coupling scaffolds TM6-7 (blue) and TM11-12 (green).
89
In summary, the structural studies on BetP by X-ray crystallography suggest an intratrimeric regulatory network on the cytoplasmatic side of BetP, which involves lipids and engages mainly the 1st charged cluster of the C-terminal domains. Conformational changes of the C-terminal domain will affect all three TM3 helices in the BetP trimer, which is clearly indicated from the structural results. However, the structures do not point towards a link between the C -terminal domain and the coupling scaffolds, which also undergo significant main chain changes from IF to OF. Therefore, it can be assumed that the conformation and orientation of the C -terminal is susceptible to hyperosmotic stress stimuli and even slight changes in hyperosmotic stress stimuli will affect the C-terminal interactions with the loops 2, 4, and 8, respectively, as well as with lipids.
This is in very good agreement with functional studies on the regulation pattern of C -terminal truncated BetP mutant in cells and proteoliposomes revealed that especially the first cluster is essential for regulation. BetP can be truncated by 25 aa (BetPC25) and still maintaining full regulatory properties in E. coli cells and proteoliposomes [15]. Regulation is lost when more than these 25 aa are truncated or key residues are mutated, or mutations are inserted that disrupt the helical fold of the C-terminal domain. The biochemical properties of C-terminal domain has been studied extensively during the past decades and revealed that the osmodependent regulation of BetP is strongly dependent on the folding and orientation of the osmosensory domain [9,10]. Similar osmoregulatory function was seen in other transporters such OpuA, a type I ABC importer [9]. The specific requirement of potassium ions during activation and the characteristic helical C-terminal domain was, however, only observed in BetP [4,6]. Activation of BetP requires elevated cytoplasmic K+ concentrations above 220 mM. This concentration is maintained in C. glutamicum even in the absence of hyperosmotic stress and under mild hyperosmotic conditions, as this soil bacterium possesses a set of efficient potassium channels. However, activation can, at least partly, be achieved by increasing K+ concentrations in proteoliposome s system, but noteworthy K+ alone, i.e. in the absence of the membrane stimulus, is not able to fully activate BetP. Vice versa full activation cannot be reached by the membrane stimulus
90
alone. Therefore, the effect of K+ might be described as a non-physiological stimulus inducing similar mechanistic changes. However, to this end K+ up-regulation should not be confused with stress regulation, although it might result in very similar conformational changes at the C-terminal domains. Especially, charged residues in the beginning of the 2nd cluster E572, H73, and R574 turned out to be important for osmosensing in E. coli cells and in proteoliposomes, but less important in the C. glutamicum membrane, which consist entirely of negatively charge lipids. Mutations that disrupted the helical folding of the osmosensor, e.g., E572P and Y550P were insensitive to osmotic stress independent on the amount of negatively charged lipids. Negatively charged lipids such as phosphatidylglycerol and cardiolipin, which contribute together nearly 70% of the membrane lipid in C. glutamicum seem to stabilize an inactive conformation of BetP (Schiller et al. JBC 2006). Evidentially the non-physiological K+ stimulus is lost once the region around E572 is mutated. This is only visible in proteoliposomes measurements as they act as osmometers and cannot provide a stress signal via the membrane. The main effect of a hyperosmotic upshift is an instant shrinking of the proteoliposomes, which results in an increase in internal K+ concentration.
To understand the initial conformational changes in the C-terminal domain and their consequences on the cytoplasmic interaction network during K+ sensing, co-crystallization of BetP in the presence of 300 mM RbCl was performed by the group of Christine Ziegler. The current chapter is now separately described in the first part represents the basis and motivation of the smFRET investigation, which I conducted the structural biology with collaboration with Ziegler lab and FRET data that I elaborate and described in the results and discussion section in detail in current chapter of this thesis.
91
Results and Discussion
Structural and functional analysis of the C-terminal conformational changes
Activation of BetP can also be achieved by Rb+ and Cs+. Rb+ has a comparable ionic radius (r) to that of K+ (rRb+ = 166 pm; rK+ = 152 pm), so a similar coordination of Rb+ and K+ ions in BetP can be assumed. Co-crystallization of a crystallization mutant of BetP missing a part of the N-terminal domain BetP(ΔN29/E44E45E46/AAA) with Rb+ was performed by incubation of the purified protein with 300 mM RbCl prior to crystallization. Although 300 mM K+ (Rb+) does not correspond to full activation conditions for this crystallization mutant, which exhibits a shifted regulation profile in C. glutamicum cells and E. coli polar lipid proteoliposomes compared to the WT, crystallization conditions and radiation damage restricted the final concentration of Rb+ to a maximum of 300 mM.
A data set was collected at a wavelength of 0.8157 Å from co-crystallizations of BetP in Cymal-5 with 1 mM betaine, which yielded a structure to a resolution of 3.4 Å, in which all three protomers adopt an inward-facing open state. Anomalous scattering at the Rb+ edge allowed to unambiguously identify K+ binding sites. Three sites were identified close to the C-terminal domains of chain A and chain C, which will be discussed in detail in the next paragraph. However, K+ binding did not change the conformation or the interaction network of the C-terminal domain in chain A with loop 2 and the C-terminal domain of chain C (Figure 4.05). One explanation might be that in 3D crystals this protomer is trapped in a conformation that would correspond to the K+ activated state. However, there was a change in the lipid-protein interactions in loop 2 at K121. Most interestingly, this residue is tightly coordinated in 4C7R by Y553 and T124 in order to keep its orientation towards the lipids.
92
Figure 4.05| Comparison of Rb+-BetP with BetP from the pdb-entry 4C7R. Insets: K121 at the end of the
trimerization helix TM2 is changing its orientation due to missing interactions with Y553 in the C-terminal domain and T124 in loop 2.
In Rb+-BetP the Rb ions are located very close to K121, which is pushed away by electrostatic repulsions. The former interaction partners Y553 and T124 are now coordinating Rb+ instead of K121, which has now moved to a former lipid position (Figure 4.06). It is yet not clear from the structural comparison what the consequences of this conformational change in K121 might be. We can only assume that the lipid-lipid interaction network in the center of BetP will change and as observed in 2D crystals the protomers will adopt a different orientation with resp ect to the membrane.
Based on the crystal structures we suggest that K+ sensing involves binding of K+ to the 1st charged cluster of the C-terminal domains, which affects the conformations of T124 and Y553 and K121 via interaction with central lipid. As these lipids are crucial in locking the inward facing state, K+ binding might destabilize the IF state.
93
Figure 4.06| K+ coordination sites in the Rb+-BetP structure and conformational changes in K121 due to K+ binding.
K121 changes its orientation towards the center of the trimer and occupies lipid positions. K+ is coordinated mainly by E552, Y553 and Q557 (C-terminal domain), R210 (loop 4) and T124 (TM2), which were involved in the coordination K121 before.
Indeed, aside of K121, there are some consequence to the transporter core, which provide mechanistic insights into the regulatory aspect after K+ sensing. In Rb+-BetP, M150C is engaged in interaction with the two residues, Ala148 and L154 that flank the glycine stretch G149-M150-G151-I152-G153 (Figure 4.07). This interaction is replaced in 4C7R by a tight interaction with residues from the coupling scaffold, especially F464, which is part of the Na2 binding site and a POPG lipid. Apparently, the flexible region of the glycine stretch is now stabilized and disconnected from lipids in the presence of K+. A similar stabilization can also be observed in loop 6 at K300 and K489, which might be possible lipid interaction sites. Overall K+ seems to stabilize the interactions between TM3 and the coupling scaffold and in addition increase the interaction between the two coupling scaffolds.
94
Figure 4.07| Comparison of Rb+-BetP with BetP from the pdb-entry 4C7R. Overlay of Rb+-BetP (tan) with BetP from the pdb-entry 4C7R (grey). (A) Coordination of M150 in Rb+-BetP. (B) Coordination of M150 in BetP-4C7R. (C) stabilization of loop 6 in Rb+-BetP (tan) overlaid with BetP from the pdb-entry 4C7R (transparent).
In order to study the contribution of each site to the osmo-regulation activity profile we performed alanine replacement of the cytoplasmic side chains involved in the Rb+ interaction-sites (Figure 4.08). Alanine mutants of residues Thr124, Arg210, Glu552 and Tyr553 exhib it severely altered responses to osmotic stress. While Arg210A and Glu552A are no longer regulated, Thr124A and Tyr553A show a slight increase in activity at high osmolalities. The alanine mutant of Gln557 retains the ability to respond to osmotic stress although with a similar shift to higher osmolalities. In summary, almost all mutants of the residues participating in the cytoplasmic Rb+ interaction sites showed impaired osmo-sensing properties compared to wild-type BetP.
95
Figure 4.08| Functional characterization of Rb+ interaction-sites. Osmotic activation profile of BetP WT expressed
in E. coli MKH13 compared to those of single alanine mutants of residues involved in the formation of Rb+ interaction sites. All measurements were carried out in triplicate. Error bars indicate s.d.
The location of the K+ interaction sites suggests that K+ binding might modulate the interaction between the C-terminal domains and adjacent protomers in the trimer. Therefore, the C-terminal domain inter-protomeric interactions was investigated by measuring the formation of a crosslink between the C-terminal domain and loop 2 of adjacent protomers, as a function of osmotic stress (Figure 4.09). First, residues Arg565 and Ile130 were replaced with cysteine; this mutant was expressed in E. coli cells and the osmotic profile was measured to test whether its activity could still be osmo-regulated. The BetP-R565C/I130C mutant was also regulated by changes in osmotic stress, although the initial activity measured at 200 mOsmol/kg was ~2.5-fold higher than WT, and the overall increase of activity was only 1.5-fold, compared to ~4-fold for the fully regulated WT BetP.
The proximity of C-terminal domain and loop 2 in this mutant BetP-I130C/R565C at low osmolality and at the osmolality of the activation optimum was assessed by the addition of the homo-bifunctional cross-linker o-PDM. When comparing WT and the cross-linked BetP-I130C/R565C using SDS-PAGE and immunoblotting against the N-terminal StrepII tag, a significant amount of dimeric cross-linked BetP-I130C/R565C, with limited amounts of trimeric protein was
96
observed a low osmolality. At high osmolality the cross-linked protein exists in both dimeric and trimeric forms.
Figure 4.09| Close-up view of the residues selected for cysteine replacement and Osmotic activation profile of BetP
WT and the double cysteine mutant I130C/R565C expressed in E. coli MKH13 cells and western blot of BetP WT and I130C/R565C from membrane vesicles of E. coli MKH13 cells extracted after incubation in iso-osmotic conditions (blue) or hyper-osmotic conditions (green), in the presence of the homobifunctional cross-linker o-PDM (o-phenylenedimaleimide). T, trimer; D, dimer; M, monomer.
These functional and structural results indicate that the interaction between the C-terminal domain of one protomer and loop 2 of an adjacent protomer occurs even under low osmolality, e.g., there is a preference of two C-terminal domains to interact. However, at maximal activation, the presence of the cross-linked trimer indicates that all three protomers have become involved in intra-trimeric interactions involving the C-terminal domain and loop 2. The crosslinking data
97
suggest further that hyperosmotic stress
activation in the presence of activating K+ promotes greater interaction between adjacent protomers in the trimer through the C-terminal domain and loop 2 of BetP. We therefore assume K+ binding is a pre-requisite in activation to weaken interactions with lipids to liberate critical residues for the C-terminal-loop 2 interaction.However, the locked conformation of the C-terminal domain in crystal structures (Figure 4.010), which is adopted even without K+ raises the question on the role of the 2nd charged cluster (N571-R595) and its involvement in stress regulation especially in the presence of negatively charged lipids as in the membrane of C. glutamicum.
Figure 4.010| Crystal contacts involving the C-terminal domain of chain A and negatively charged residues of all
three protomers of the symmetry mate in the crystal. As a result, all crystal structures including the one obtained from co-crystallization show a similar asymmetric C-terminal arrangement, which domains from chain B and C partly unresolved missing the 2nd charge cluster entirely.
The systematic FRET study described in the second part of this chapter was intended to shed light into the C-terminal flexibility of the two cluster separately and in dependence of the K+ concentration. In addition, the use of two different membrane-mimics – detergent as in the crystal structures as well as amphipol – should shed light into the nature of the pre-activated state observed in all crystal structures. In fact, smFRET should allow to assign the activation state in these structures. To mimic native environment of membrane, we decided to use the amphipol
98
polymer A8-35, which from binding studies was even more close to the proteoliposomes system than POPG nanodiscs (Figure 4.011).
Figure 4.011| Comparison of Na+-coupled betaine binding to BetP reconstituted into nanodiscs, amphipols and proteoliposomes with Kd values of 0.98mM in proteoliposomes, 0.35mM in amphipol (inset: chemical formula) and 0.27mM in POPG nanodiscs.
Motivation for the FRET study of BetP in amphipol and detergent
In previous published work, we described a novel FRET methodology to investigate and characterize the structure of homotrimeric BetP using smFRET (see chapters 2/3 and ref. [14]). We demonstrated that optimized labelling in combination with customized data analysis methods are the crucial for smFRET studies of trimeric membrane transporters such as BetP [14]. In the current study described below, we established FRET measurements of BetP in detergent and amphipol. We identified the residues for fluorophore labelling from previous functional data and confirmed that these positions will not alter regulation or transport after labelling. We incorporated donor and acceptor fluorophores for smFRET studies at three different positions along the C-terminal domain to track the conformational state of the regulatory domain: (1) first
99
charged cluster (L551), (2) second charged cluster (E577 and R584). Especially the last residue was challenging as this position is only resolved in view structures namely in 2WIT and 4C7R, respectively (Figure 4.012).
Figure 4.012 I BetP trimer with predicted C-terminal orientations. (A) The three label positions are colored in
orange. The predicted C-terminal structures, which remain unresolved in all crystal structures are indicated as lines. The C-terminal domain of chain B always points away from the trimer center resulting in an asymmetric conformation of terminal domains. (B) The asymmetric orientations of the C-terminal domains (here predicted from the orientations observed in X-ray structures) lead to a variation in distances, which is more pronounced for the distal 2nd cluster label at position E577.
We investigated the effect of increasing concentrations of K+ in the absence and presence of substrates sodium and betaine on these three specific locations both in detergent and amphipol. In summary, our results show that only when BetP is reconstituted into the negatively charged amphipol polymer, which serves as a mimic for the negatively charged C. glutamicum membrane, the C-terminal domain of BetP undergoes potassium-dependent conformational changes. Most interestingly, these changes occur strongly in the second charged cluster, which is not resolved in crystal structures. Depending on the lipid environment, this 2nd charged cluster is playing
100
different roles in regulation. We assume that K+ activation results in a partial unfolding of this segment.
The FRET data reveal a striking difference between the C-terminal conformations of BetP in amphipol and in non-ionic detergent DDM, respectively. It seems that in DDM the last tip of the C-terminal domain (E577-R595) adopts a locked conformation independent on the presence of potassium, which corresponds to the up-regulated conformation of BetP in amphipols. These results confirm our hypothesis that the crystal structures reveal a K+-pre-activated state. We assume that when BetP is reconstituted in detergent, important down-regulating lipid interactions at this terminal segment are already missing, Surrounding environment of C-terminal domain in detergent micelles does not enable the protein to adopt a fully activated state and therefore cannot sense the pre-activation trigger of K+ , which in amphipol the protein undergoes the activation state first by K+ binding. This would correspond in the molecular switch model (Figure 4.013) to a detachment of the last tip of the C-terminal domain from the membrane or at least to a weaken interaction with the membrane in order to react to the membrane stimuli.
Figure 4.013 I Molecular switch model for BetP K+ activation.
The higher FRET efficiency of DDM BetP and high K+ amphipol BetP and the narrow FRET efficiency distribution suggests that in this pre-activated state the second charged cluster of the C-terminal domains approach each other or even move towards the own protomer. The first
101
charge cluster L551 remains more or less unaltered as expected from the tight interaction network between lipids and cytoplasmic loops. The position E577 reacts to K+ more pronounced in detergent, however, there are no titration measurements for this position in amphipol to confirm if this region is also affected by different membrane mimics as the functional data would suggest for the stretch E572-R574.
102
FRET experiments of BetP C-terminal mutants
We designed and incorporated cysteines at strategic positions of C-terminal helices (Figure 4.1). Since most residues in the C-terminal domain are critical for K+ sensing and activation, [9][15], we used published structural and biochemical data from previous site -directed mutagenesis analysis to select residues that are not directly involved in K+ sensing and activation. The selected mutations for smFRET had no critical effect on the activity of the protein and did not disturb the function and regulation of BetP [7].
Figure 4.1| Experimental approach for smFRET studies of BetP’s C-terminus. (A) Side view of the crystal structures
of BetP-Cysteine mutants C252T/R584C, C252T/E577 and C252T/L551C marking the thre e label positions (orange) in protomer A. PDB: 4C7R. (B) Top view of the crystal structure of the mutant C252T/R584C and related distances (top) and schematic showing FRET of homotrimeric BetP based on molecular switch activation model [23][24]. Schematic showing labeled diffusing protein in the confocal spot (bottom). (C) 2D ALEX histogram of Alexa555/647-labeled BetP R584C showing three distinct FRET populations.
103
To visualize the conformational states of C-terminal domain by single-molecule experiments, we selected and created cysteine mutants at three discrete distanced positions (Figure 4.1A). The cysteines mutants where labelled stochastically with donor (Alexa Fluor 555) and acceptor (Alexa Fluor 647) dyes as described before. We used confocal microscopy identifying single-molecule transits and burst analysis methods to extract FRET information of BetP (Figure 4.1B). As described previously [14] stochastic labelling of BetP homotrimer leads to three distinct smFRET species (Figure 4.1C): donor acceptor (high S), acceptor (intermediate S) and donor-acceptor-acceptor (low S). Count-rate histograms in Figure 4.2. reveal that we can select the relevant region with donor-acceptor molecules as shown for BetP-584C via choosing burst with intermediate S-values.
Figure 4.2I Photophysical propertiers of homotrimeric BetP C-terminal mutant R584C in amphipol. (A) Photon
count rate of single-molecule bursts from different subpopulations in the stoichiometry regions between 0.27 and 0.75. (B) One-dimensional E* histograms of the different species of the BetP C-terminal mutants R584C, showing three stoichiometry regions assigned as High, Intermediate and Low S.
To assess BetP C-terminal variants as a stable trimer in a detergent-free aqueous solution, we used a commercially-available amphipol, that was introduced previously for investigation of
104
several membrane protein structures by electron microscopy (cryo-EM)[25][26]. This approach using Amphipol A8-35 [27] was successfully applied to for spectroscopic studies of proteins [10]. Functional binding studies on amphipol reconstituted BetP confirmed that Amphipol A8-35 is a suitable membrane mimic for BetP. Moreover, betaine and sodium binding turned out to be K+ dependent in this environment. We determined and verified the accessible FRET-ruler range using a set of double-stranded DNA samples (dsDNA) carrying the donor Alexa555 and acceptor dye Cy5 (equivalent of Alexa647). A detailed description of the sample is provided in Plötz et al. [33]. We analyzed a subset of these dsDNA FRET samples with 23 bp, 13 bp and 8 bp donor-acceptor distance with results shown in Figure 4.3.
Figure 4.3| A/B) Two-dimensional E*/S* histogrammes of dsDNA samples with Alexa555/Cy5 donor-acceptor pair.
The data were analyzed with dual-colour burst search in order to hightlight the donor-acceptor-labelled fraction. The data was not corrected for background, spectral imperfection and quantum yield differences. Burst-search parameters were M = 15, T = 500 µs and L = 25[Nir et al., JPC B 2007]; the plot uses additional per-bin thresholds for all photons > 150, 61x61 bins for E*/S*. Experiments in A/B were done on the same day as protein measurements
105
on BetP-R584C in buffer as detailed in material and methods. B) Also, the influence of amphipol was tested and the 13 bp dsDNA data verifies that no changes on the fluorophore photophysics occur when amphipol is added to the solution, since E*-values are unaltered between buffer (B, left) and buffer with amphipol (B, right).
The data in figure 4.3 shows that the accessible distances of our FRET-assays ranges from of ~9 nm (23 bp, linear model: 23x0.34 nm = 7.82 nm; a more realistic cylindrical DNA model suggests a probe separation of 8.8 nm according to [32] with associated mean apparent FRET values of E* = 0.30, down to ~4 nm (8 bp, linear model: 8x0.34 bp = 2.72 nm; a more realistic cylindrical DNA model suggests a probe separation of 4.6 nm according [32] While all smFRET data in the manuscript from both BetP and dsDNA are not corrected for instrumental parameters, these experiments (Figure 4.3) define the accessible distance range clearly. This is supported by intermediate E*-values of 0.59 for 13 bp donor-acceptor separation (linear model: 13x0.34 nm = 4.42 nm; a more realistic a cylindrical DNA model suggests a probe separation of 5.7 nm according to [32]. Importantly, the DNA experiments not only define the dynamic range of the FRET assay for the selected fluorophores, but also reveal that one static distance in the DNA is smeared out in the histogramme (Figure 4.3) due to the single -molecule character of the experiments. We finally used this model system to test the influence of amphipol. Analysis of the 13 bp dsDNA data shows no changes on the fluorophore photophysics when amphiphol is added to the solution, since E* -values are unaltered between buffer and buffer with amphipol (Figure 4.3B).
As a next step we analyzed the different BetP mutants in terms of their FRET efficiency distribution to first obtain a qualitative picture of the width, which is tentatively related to number of states and the mean position of the distribution, which is related to separation of the labels in space; Figure 4.4.
106
Figure 4.4| FRET efficiencies of BetP labelled in the 1st and 2nd charged cluster in DDM and Amphipol.
One-dimensional E* histograms of the C-terminal BetP-Cysteine mutants in a non-activated state (no K+ present) measured in DDM and Amphipol. Apparent FRET values of C-terminal mutants increased with decreasing distance of C-terminus helices in both detergent and Amphipol. (A) One-dimensional E* histograms of BetP-Cysteine mutants C252T/R584C, C252T/E577 and C252T/L551C in DDM under non-activating state in absence of K+ (B) One-dimensional E* histograms of BetP-Cysteine mutants in Amphipol under non-activating state.
smFRET experiments at first sight reproduced the expected trend based on the crystal structure (Figure 4.1) for both detergent and amphipol. Based on Figure 4.1. we expected short distances for E551C (C- distance in the crystal 2.5 nm), intermediate distances of 5.6 nm (E577C) and 6.8 nm for E584C; this is qualitatively reflected in the apparent FRET efficiencies of BetP in Figure 4.4. In contrast to the qualitative agreement and correlation, however, only the L551C mutants shows a reasonable match between its C- distance in the crystal (2.5 nm) and the observed E*-value of the corresponding e.g., dsDNA (8 bp). Overall, we find much higher E* -values as expected from the crystal structure, e.g., for E577C and 584C with 5.6-6.8 nm fluorophore separation we expected E* values below 0.5 due to the separation of the dyes larger than the Förster radius.
107
Such an effect can be interpreted in two different ways: (i) The C-terminal helices are folded, yet they adopt a conformational state that brings the fluorophores closer together than suggested by the arrangement of chain A in the crystal, i.e., pointing towards the cytoplasm away from the membrane surface. (ii) Alternatively, the C-terminus could be (partially) unfolded or with no specific arrangement (as suggested by the crystal structure) causing fast averaging of FRET-efficiency values. (iii) The C-terminal modules dynamically moves in a folded state on the attachment point of the transmembrane part of the protein again resulting in averaging of FRET efficiency values. These alternatives might be distinguishable by use of time-resolved anisotropy experiments, which allows to track the rotational diffusion time of the attached fluorophores. For labels that residue in folded parts of BetP’s C-terminus, we would expect higher values and control experiments could be done with artificial unfolding of model proteins to provide control experiments for the rotational diffusions of folded/unfolded systems. Alternatively, HDX-mass spectrometry might be able to shed further light on to this important mechanistic aspect.
Interestingly, the apparent FRET efficiency distributions were also broader in amphipol and shifted towards lower E*-values for all positions (Figure 4.4.). This idea is supported by further analysis of the data. For all labelling positions there was a difference in proximity ratio, i.e., EPR values with crosstalk and background corrections, which was reduced by nearly 20% over all distances when BetP was reconstituted in amphipol (Figure 4.5).
108
Figure 4.5. (A) FRET Data on the level of apparent FRET of C-terminal variants stabilized in both Amphipol and DDM
increase with expected decrease in distances. FRET Data on the level of apparent FRET are corrected for background and spectral crosstalk values. DDM (red) and Amphipol (blue). Proximity ratio EPR of C-terminal mutants increased as expected with decreasing distance of C-terminus helices in both detergent and Amphipol environments.
This data shown in Figure 4.5 is interesting by itself and suggests that BetP’s C-terminal helix really adopts a different conformational state (or ensemble) in DDM. It has to be mentioned, however, that the observed changes between DDM and amphipol can be of different molecular nature (as discussed above).
Next, we investigated the effect of KCl on the FRET-efficiency distributions of BetP with labels at R584C in both amphipol and DDM, as this position showed the most pronounced difference between these two membrane mimics (Figure 4.4). For this mutant, we performed a series of smFRET experiments with increasing concentration of KCl (Figure 4.6). The results demonstrate that only in amphipol, increasing K+ alters the FRET-efficiency distributions from a broad to narrow form (Figure 4.6A). Interestingly, the narrow distribution with 500 mM KCl closely resembles those in DDM, which were not changed upon addition of KCl. We note that all experiments were done under conditions that allow to keep ionic strength in the solution constant. Low salt conditions were not possible to study since the protein was not stable under these conditions.
109
Figure 4.6 I One-dimensional E* histograms of BetP-R584C in amphipol (A), DDM (B) with increasing concentrations
of KCL.
The C-terminal mutant L551C in detergent did not show strong conformational changes in in response to increasing K+ concentration (Figure 4.7 A). This is in line with the little impact of K+ on R584C in DDM. Apparently, at L551 the C-terminal domain is very close to the linker to TM12 (S546-Y550) and therefore can perform very limited conformational changes even in the presence of K+ stimuli.
110
Figure 4.7 I One-dimensional E* histograms of the Alexa555/647-labeled BetP C-terminal Cysteine mutants in DDM (A) L551C and (B) E577, in non-activating and K-activating state and in the presence of saturated concentration of
ligand in detergent.
Interestingly, the E577C mutant revealed conformational changes in detergent in the presence of K+ as the FRET efficiency distribution is getting narrower under activating conditions (Figure 4.7 B). In comparison with our structural findings we can attribute this to a further stabilization of the C-terminal domain in this region and an increased interaction with loop 2 as observed in the cross-linking data of I130C-R565C.
Comparing the three positions, we conclude that the first part of the C-terminal domain (L551), which is close to the transporter core and interacting at D547 and Y550 with cytoplasmic loop is not strongly affected by K+ binding. The smFRET results presented are compatible with the idea that the middle part of the C-terminal domain at E577 is more flexible in detergent but becomes stabilized significantly by K+ binding, which would be in good agreement with the latest crystal structure, where K+ stabilize the C-terminal domain close to this position. The last tip of the C-terminal domain probed via R584C labelling remains unaffected by K+ in detergent but show significant changes in response to K+ in amphipols.
Since this position showed the most pronounced changes, we evaluated the effect of sodium, lithium and betaine in absence and presence of K+ for BetP-R584C (Figure 4.8). Most interestingly,
111
the addition of betaine yielded a broadening of the FRET efficiency distribution suggesting the C -terminal domain at this position is again more flexible. In fact, in crystal structure s at high betaine concentrations, the C-terminal domains appear less well resolved even in the presence of K+ as shown in the latest structures (see introduction part). This impact of betaine on the conformational flexibility is reversed at the 1st charged cluster at BetP-L551C in the presence of betaine in detergent (Figure 4.5). For R584C only K+ induces the conformational changes and no comparable effect is observed in the presence of sodium or substrate on the FRET efficiency E* histograms of variants (Figure 4.8).
Figure 4.8| SmFRET measurements of R584C in amphipol in presence and absence of substrate, Na+ and K+. (A)
One-dimensional E* histograms of R584C in non-activating condition in the presence of Na+ / Li+ alone 2) K+ alone,
112
The dynamics of the C-terminal domain in hyperosmotic osmotic stress sensing and regulation of transport activity was investigated already previously by EPR [28]. The movement of spin-labeled C-terminal domains in various lipid environments showed a restricted conformational motion that are strongly dependent on the surrounding lipid composition [29]. These EPR studies revealed a nearly 4-fold decrease in correlation times in the C-terminal mobility upon the reconstitution into liposomes compared to protein in detergent. Also, the relative amplitudes suggest a faster transition in detergent compare to various lipid environments. This is indeed in good agreement to the comparison of FRET efficiency in amphipol and detergent described in this chapter. If the C-terminal domain is hold in an ensemble of conformations due to an attachment to the membrane (here mimicked by the negative charges of the amphipol polymer), the transition to an activated state would be significantly slowed down.
Our results also support findings of the structural analysis of BetP. The reorientation of C -terminal domains was suggested already from the first crystal structures from the asymmetric arrangements of these domains. However, as the N-terminal domain in the crystallization mutant of BetP is missing and it was not clear only from crystal structures if this fact was causing the trapped conformation of the C-terminal domain. One possible scenario here would be that the helical fold of the C-terminal domain could not be maintained because of missing N-terminal interactions (Figure 4.9).
However, the N-terminal domain plays a modulatory role as truncations will not change the up-regulation just the shifts the activation optimum to higher osmolalities. When missing the activation maximus of the osmotic stress is shifted towards higher osmolarities. Based on our
smFRET results at position R584C, we suggest that during pre -activation by K+ the C-terminal
domain undergoes a partner-switch from lipids to N-terminal domain. Evidently when being detached from the membrane (down-regulated state) the distal end of the C-terminal domain can only hold its helical fold by interaction with a negatively charged partner. In amphipol this can be the carboxyl-groups of the polymer mimicking the negatively charged lipid head groups, in detergent this might be the N-terminal domain, which is missing in the crystallization
113
mutants. In crystals most interestingly the N-terminal structure consisting of three negatively charged clusters is mimicked by the periplasmic loops of the symmetry mates. Therefore, we assume that the conformation we observe in the crystals correspond to a pre -activated state
in which the strong lipid-interactions of the 2nd cluster are already replaced by N-terminal
interactions mimicked via the crystal contact.
In the smFRET study, BetP was not truncated N-terminally and we assume that once not stabilized by lipids the C-terminal domain in detergent is also stabilized by interactions with the N-terminal domain as previously reported by cross-linking studies. Amphipol-solubilize d variants indicate that surrounding environment with detergent micelles does not enable BetP to adopt a fully down-regulated state and therefore cannot sense the pre-activation trigger of
K+. The FRET data are in very good agreement with the molecular switch model. The crystal
structures revealed that the C-terminal interaction at R558-568 with negatively charged residues in loop 2 (D131, E132 and E135) is the key to regulation.
Figure 4.9 | Predicted N-terminal interactions with the last tip of the C-terminal domain.
Even a slight movement of this first helical folded segment of the C-terminal domain away from the trimer center would change the conformational flexibility of TM3 and definitely the opening degree and opening probability of the inward-facing state. This position is in direct
114
contact with lipids (PG and Cardiolipin) and affected by K+, which on one hand stabilize the
region and also the flexible linker between TM12 and the C-terminal domain, which bends like an ‘elbow’. This movement can influence the entire cytoplasmic network of loops that control via TM8, TM5 and TM3 both sodium and betaine binding sites. The C-terminal movement
suggested by the FRET data points towards the following activation mechanisms: Upon K+
binding the C-terminal domain gets detached mainly around the last charge cluster from the membrane surface. The missing interactions with lipids lead to a partial unfolding of the last
charge cluster, which is compensated by K+ binding to the first cluster eventually assisted by
N-terminal interactions. Both is accompanied by a slight orientational change of the C-terminal domain, however the interaction between loop 2 and the C-terminal domain is maintained. In
the presence of a hyperosmotic shock, this K+ stabilized assembly can immediately move in
order to unlock the transporter.
The smFRET data were paramount to finally assign this pre-activated state in crystal structures. Furthermore, they indicate that the conformational space of the C-terminal domain is large r than anticipated from the molecular switch model, which assumes just two distinct states. The fact of a mutual interaction of the C-terminal. Domains within the trimer and the asymmetry in BetP suggest that a part of regulation is the consecutive activation of individual protomers.
Materials and Methods:
Bacterial strains, plasmids and growth conditions
The site-directed mutagenesis was used with the cysteine-free BetP mutant C252T as PCR template for C-terminal variants. Protein expression and purification was performed as described previously [16][17]. For heterologous expression of Strep-BetP, DH5α™-T1 (Invitrogen) component cells were used for transformation of PASK-IBA5betp WT and additional mutants R584C, E577C and L551C. We used Luria-Bertani medium supplemented with Carbenicillin (50 μg/ml) to grow cells at at 37°C and 180rpm. The anhydro tetracycline (200 μg/L) was used for protein induction and cells were harvested after reaching stationary phase. Membrane were isolated and solubilized in 0.1 % N-Dodecyl β-dodecyl-maltoside (DDM) with buffer 50 mM Tris–
115
HCl (pH 7.5), 200 mM NaCl, 8.6% glycerol, and 0.1% DDM and then protein was purified using StrepTactin column (IBA GmbH with same buffer supplemented with 5mM destibiothin. As next step, the purified protein was incubated with 1mM DTT (Dithiothreitol) for 1hour and purified with an equilibrated size exclusion column (Superose 6 10/300 GL).
Labeling of BetP cysteine-containing mutants with thiol-specific reagent
BetP C-terminal cysteine-containing mutants were applied onto a size exclusion column (Superose 6 10/300 GL) equilibrated with 50 mM Tris, pH 7 and 150 mM NaCl, 0.1%DDM after STREP protein purification. SEC fractions were concentrated using Vivaspin (Sartorius, MW cut-off 50 kDa) and Amicon Ultra (Millipore, MW cut-cut-off 30 kDa) concentrator with centrifugation at 4° and then were stored at −20 °C in 200 μl aliquots of 1-3 mg/ml in 50 mM Tris, pH 7.5, 200 mM Nacl, 8,6% glycerol and 0.1% DDM. BetP variants were subjected for stochastic labeling with maleimide derivatives of donor and acceptor fluorophores with different ratio in order to obtain optimized labeling efficiency for single-molecule FRET experiments. In order to to fully reduce oxidized cysteines C-terminal derivatives treated with 10 mM DTT for 30 mins in a deoxygenated Buffer solution containing 50 mM Tris, pH 7.5, 200 mM NaCl, 8.6% glycerol and 0.1% DDM (Buffer A). The protein mixture was further diluted to a DTT concentration of 0.5 mM in buffe r A and then labeled with the appropriate dye Alexa 555-maleimide (donor) and Alexa647-maleimide (acceptor) in an optimized ratio of protein:donor:acceptor and then was loaded onto an equilibrated desalting column (ZEBA, 2ml) with the MCWO 7KDa to remove the DTT from the protein solution. Labeling was conducted after washing the protein with deoxygenated buffer 50 mM Tris, pH 7.5, 150 mM NaCl and 0.1%DDM (buffer B). The protein solution incubated with dye mixture for 5 hours at 4 °C (under mild agitation). After labeling reaction, unreacted dyes were washed with Buffer B and ZEBA desalting column as described in previous step. Next, labelled BetP was applied onto an equilibrated desalting size exclusion column (Superose 6 10/300 GL) with 50 mM Tris, pH 7 and 150mM NaCl, 0.1%DDM. The optimum labelled fraction was collected and directly followed by liposome reconstitution or an Amphipol incorporation.
116
Reconstitution of BetP variants into Amphipol A8-35
Amphipol (A8-35 from Anatrace) was added to labeled BetP C-terminal mutants in 50 mM Tris, pH 7.5, 200 mM NaCl, 0.1% DDM at 1:3 protein/APol weight ratios. Briefly, the protein/APol mixture incubated for 4 hours at 4°C (under mild agitation). As a next step, the polystyrene beads (Bio-Beads SM2) were washed with methanol, distilled water and reconstitution buffer without detergent and then were added to remove the detergent from protein/APol solution at a 1:10 detergent/bio-beads weight ratio and as the next step, solution was incubated for 5 hours at 4°C (under mild agitation). The bio-beads were removed by using an Eppendorf benchtop micro centrifuge at 6000 rpm speed at 4°C. The collected fractions were concentrated using Amicon Ultra (Millipore, MW cut-off 30 kDa) concentrator with centrifugation at 4°C and then subjected for removal of free Amphipols on the size exclusion column ( Superose 6 10/300 GL) equilibrated with reconstitution buffer without detergent in 50 mM Tris, pH 7.5 and 150mM NaCl.
Sample preparation for microscopy
All microscopy methods are described and detail in ref [14] and were adapted for use in this article. DNA reference samples were obtained as described before [19]. ALEX experiments were carried out at room temperature with 25-50 pM solution of protein and DNAs samples. For DNA sample we used imaging buffer phosphate-buffered saline (PBS) at pH 7.4, containing 2 mM Trolox. Protein samples were also analyzed at 25-50 pM in an imaging buffer containing 100 mM Tris pH 7.4, 200 mM NaCl, and 0,1% DDM. In typical single-molecule experiments, sample solutions were transferred to coverslips that were previously incubated with 1 mg/ml Bovine serum albumin (BSA) in 100mM Tris pH 7.4, 200 mM NaCl and 0.1% DDM for 5 minutes incubation for surface passivation.
Liposome reconstitution of BetP derivatives
Liposomes were prepared using an E. coli polar lipid extract (polar Lipid Extract, 20 mg/ml in chloroform, Avanti Polar Lipids, USA) as described before ( with reconstitution buffer 50mM Tris (pH:7.4), 200 mM NaCl and 1 mM mercaptoethanol (Bio-rad) and extruded through a polycarbonate membrane filter (pore size of 400 nm). Next, we performed freeze-thaw cycles of empty liposomes in order to homogenize the liposomes. The solubilisation of liposomes was
117
initiated with titration of 10 % (w/v) Triton-X-100 (Sigma-Aldrich) and the absorbance were measured at 540 nm. The labelled BetP derivatives were mixed in the reconstitution buffer and incubated with liposome for 1 hour at 4°C (under mild agitation). The polystyrene Bio-Beads SM-2 (Bio-Rad) were added in five steps to remove detergent from protein solution in buffer, for following last overnight incubation at 4°C. Next, proteoliposomes were washed twice with reconstitution buffer to remove the bio-beads from mixture by centrifugation at 70,000 rpm speed at 15°C for 15 min and then stored at -80°C. The protoliposome were subjected for three freeze-thaw cycles and then extruded through a pre-wet polycarbonate membrane filter (pore size of 100 nm) prior to measurements.
Single-molecule FRET and ALEX spectroscopy
We used a custom-built confocal microscope for μs-ALEX, which we described in detail previously)[20][21]. In brief, a water immersion objective with NA = 1.2 was used to generate a diffraction limited spot using two laser diodes (Cube, Coherent, Germany). The excitation intensity was typically set to 30-60 μW at 532 nm and 15-25 μW at 640 nm with an alternation period of 50 μs. Fluorescence emission was collected in epi-fluorescence mode, spatially filtered by a 50-μm pinhole, matching bandpass filters and registered by two avalanche photodiode detectors (-spad, Picoquant, Germany; and Count modules, Laser Components, Germany). In this mode, three photon streams were extracted from the data corresponding to donor-based donor emission F(DD), donor-based acceptor emission F(DA) and acceptor-based acceptor emission F(AA). S and apparent FRET efficiencies E* were calculated for each fluorescent burst during their diffusion time trough confocal spot above a certain threshold yielding a two-dimensional (2D) histogram. Uncorrected FRET efficiency E* were calculated according to E* = F(DA)/(F(DD)+F(DA)). Stoichiometry S was defined as the ratio between the overall green fluorescence intensity over the total green and red fluorescence intensity during the green excitation period and describes the ratio of donor-to-acceptor fluorophores in the sample S = F(DA)+F(DD)/(F(DD)+F(DA)+F(AA)).
118
Using published procedures[22] to identify fluorescent bursts corresponding to single molecules, we obtained bursts characterized by three parameters (M, T, and L). A fluorescent signal is considered a burst provided it meets the following criteria: a total of L photons, having M neighboring photons within a time interval of T microseconds. For data shown in Figure 4.1 an all-photon burst search with parameters M = 15, T = 500 µs and L = 50 was applied; for data shown in all other Figures, a dual colour burst search using parameters M = 15, T = 500 µs and L = 25 was applied; additional thresholding removed spurious changes in fluorescence intensity and selected for intense single-molecule bursts (all photons > 100/150 photons unless otherwise mentioned). Binning the detected bursts into a 2D E*/S histogram where sub-populations are separated according to their S-values. E*- and S-distributions were fitted using a Gaussian function, yielding the mean values µi of the distribution and an associated standard deviations
wi. Experimental values for E* and S were corrected for background and for spectral crosstalk according to published procedures [13].
119
REFERENCES CHAPTER 4:
[1] Farwick, M., R.M. Siewe, and R. Krämer, Glycine betaine uptake after hyperosmotic shift in Corynebacterium glutamicum. Journal of Bacteriology, 1995.
[2] K. Khafizov, C. Perez, C. Koshy, M. Quick, K. Fendler, C. Ziegler, and L. R. Forrest, “Investigation of the sodium-binding sites in the sodium-coupled betaine transporter BetP,” Proc. Natl. Acad. Sci., vol. 109, no. 44, p. E3035 LP-E3044, Oct. 2012.
[3] A. Vergara-Jaque, C. Fenollar-Ferrer, C. Mulligan, J. A. Mindell, and L. R. Forrest, “Family resemblances: A common fold for some dimeric ion-coupled secondary transporters,” J. Gen.
Physiol., vol. 146, no. 5, pp. 423–434, Nov. 2015.
[4] C. Perez, C. Koshy, Ö. Yildiz, and C. Ziegler, “Alternating-access mechanism in conformationally asymmetric trimers of the betaine transporter BetP,” Nature, vol. 490, p. 126, Sep. 2012.
[5] R. M. Gärtner, C. Perez, C. Koshy, and C. Ziegler, “Role of Bundle Helices in a Regulatory Crosstalk in the Trimeric Betaine Transporter BetP,” J. Mol. Biol., vol. 414, no. 3, pp. 327–336, 2011. [6] C. Perez, C. Koshy, S. Ressl, S. Nicklisch, R. Krämer, and C. Ziegler, “Substrate specificity and ion
coupling in the Na& betaine symporter BetP,” EMBO J., vol. 30, no. 7, p. 1221 LP-1229, Apr. 2011. [7] D. Schiller, R. Rübenhagen, R. Krämer, and S. Morbach, “The C-Terminal Domain of the Betaine
Carrier BetP of Corynebacterium glutamicum Is Directly Involved in Sensing K+ as an Osmotic Stimulus,” Biochemistry, vol. 43, no. 19, pp. 5583–5591, May 2004.
[8] R. Rübenhagen, S. Morbach, and R. Krämer, “The osmoreactive betaine carrier BetP from Corynebacterium glutamicum is a sensor for cytoplasmic K(+),” EMBO J., vol. 20, no. 19, pp. 5412–5420, Oct. 2001.
[9] V. Ott, J. Koch, K. Späte, S. Morbach, and R. Krämer, “Regulatory Properties and Interaction of the C- and N-Terminal Domains of BetP, an Osmoregulated Betaine Transporter from
Corynebacterium glutamicum,” Biochemistry, vol. 47, no. 46, pp. 12208–12218, Nov. 2008. [10] G. Güler, R. M. Gärtner, C. Ziegler, and W. Mäntele, “Lipid-Protein Interactions in the Regulated
120
Betaine Symporter BetP Probed by Infrared Spectroscopy,” J. Biol. Chem. , vol. 291, no. 9, pp. 4295–4307, Feb. 2016.
[11] van der M. B. Wieb, “Förster Theory,” FRET – Förster Resonance Energy Transfer. 04-Oct-2013. [12] S. Sindbert, S. Kalinin, H. Nguyen, A. Kienzler, L. Clima, W. Bannwarth, B. Appel, S. Müller, and C.
A. M. Seidel, “Accurate Distance Determination of Nucleic Acids via Förster Resonance Energy Transfer: Implications of Dye Linker Length and Rigidity,” J. Am. Chem. Soc., vol. 133, no. 8, pp. 2463–2480, Mar. 2011.
[13] J. Hohlbein, T. D. Craggs, and T. Cordes, “Alternating-laser excitation: single-molecule FRET and beyond,” Chem. Soc. Rev., vol. 43, no. 4, pp. 1156–1171, 2014.
[14] A. A. Jazi, E. Ploetz, M. Arizki, B. Dhandayuthapani, I. Waclawska, R. Krämer, C. Ziegler, and T. Cordes, “Caging and Photoactivation in Single-Molecule Förster Resonance Energy Transfer Experiments,” Biochemistry, vol. 56, no. 14, pp. 2031–2041, Apr. 2017.
[15] J. M. Wood, “Bacterial responses to osmotic challenges,” J. Gen. Physiol., Apr. 2015.
[16] R. Rübenhagen, H. Rönsch, H. Jung, R. Krämer, and S. Morbach, “Osmosensor and Osmoregulator Properties of the Betaine Carrier BetP from Corynebacterium glutamicum in Proteoliposomes ,” J.
Biol. Chem. , vol. 275, no. 2, pp. 735–741, Jan. 2000.
[17] S. N. Cohen, A. C. Y. Chang, and L. Hsu, “Nonchromosomal Antibiotic Resistance in Bacteria: Genetic Transformation of Escherichia coli by R-Factor DNA,” Proc. Natl. Acad. Sci. U. S. A., vol. 69, no. 8, pp. 2110–2114, Aug. 1972.
[18] P. Girard, J. Pécréaux, G. Lenoir, P. Falson, J.-L. Rigaud, and P. Bassereau, “A New Method for the Reconstitution of Membrane Proteins into Giant Unilamellar Vesicles,” Biophys. J., vol. 87, no. 1, pp. 419–429, Jul. 2004.
[19] E. Ploetz, E. Lerner, F. Husada, M. Roelfs, S. Chung, J. Hohlbein, S. Weiss, and T. Cordes, “Förster resonance energy transfer and protein-induced fluorescence enhancement as synergetic multi-scale molecular rulers,” Sci. Rep., vol. 6, p. 33257, Sep. 2016.
[20] van der V. J. H. M., P. Evelyn, H. Matthias, O. Jens, de V. J. Willem, R. Gerard, and C. Thorben, “Mechanism of Intramolecular Photostabilization in Self-Healing Cyanine Fluorophores,”
121
ChemPhysChem, vol. 14, no. 18, pp. 4084–4093, Dec. 2013.
[21] G. Gouridis, G. K. Schuurman-Wolters, E. Ploetz, F. Husada, R. Vietrov, M. de Boer, T. Cordes, and B. Poolman, “Conformational dynamics in substrate-binding domains influences transport in the ABC importer GlnPQ,” Nat. Struct. &Amp; Mol. Biol., vol. 22, p. 57, Dec. 2014.
[22] E. Sisamakis, A. Valeri, S. Kalinin, P. J. Rothwell, and C. A. M. Seidel, “Chapter 18 - Accurate Single-Molecule FRET Studies Using Multiparameter Fluorescence Detection,” in Single Single-Molecule Tools,
Part B:Super-Resolution, Particle Tracking, Multiparameter, and Force Based Methods, vol. 475,
N. G. B. T.-M. in E. Walter, Ed. Academic Press, 2010, pp. 455–514.
[23] F. Korkmaz, S. Ressl, C. Ziegler, and W. Mäntele, “K+-induced conformational changes in the trimeric betaine transporter BetP monitored by ATR-FTIR spectroscopy,” Biochim. Biophys. Acta -
Biomembr., vol. 1828, no. 4, pp. 1181–1191, 2013.
[24] S. Maximov, V. Ott, L. Belkoura, and R. Krämer, “Stimulus analysis of BetP activation under in vivo conditions,” Biochim. Biophys. Acta - Biomembr., vol. 1838, no. 5, pp. 1288–1295, 2014.
[25] C. Tribet, R. Audebert, and J.-L. Popot, “Amphipols: Polymers that keep membrane proteins soluble in aqueous solutions,” Proc. Natl. Acad. Sci., vol. 93, no. 26, p. 15047 LP-15050, Dec. 1996.
[26] V. Polovinkin, I. Gushchin, M. Sintsov, E. Round, T. Balandin, P. Chervakov, V. Shevchenko, P. Utrobin, A. Popov, V. Borshchevskiy, A. Mishin, A. Kuklin, D. Willbold, V. Chupin, J.-L. Popot, and V. Gordeliy, “Erratum to: High-Resolution Structure of a Membrane Protein Transferred from Amphipol to a Lipidic Mesophase,” J. Membr. Biol., vol. 250, no. 2, p. 237, 2017.
[27] M. Zoonens and J.-L. Popot, “Amphipols for Each Season,” J. Membr. Biol., vol. 247, no. 0, pp. 759–796, Oct. 2014.
[28] J. Botzenhardt, S. Morbach, and R. Krämer, “Activity regulation of the betaine transporter BetP of Corynebacterium glutamicum in response to osmotic compensation,” Biochim. Biophys. Acta -
Biomembr., vol. 1667, no. 2, pp. 229–240, 2004.
[29] S. C. T. Nicklisch, D. Wunnicke, I. V Borovykh, S. Morbach, J. P. Klare, H.-J. Steinhoff, and R. Krämer, “Conformational changes of the betaine transporter BetP from Corynebacterium glutamicum studied by pulse EPR spectroscopy,” Biochim. Biophys. Acta - Biomembr., vol. 1818,
122 no. 3, pp. 359–366, 2012.
[30] S. Elter, T. Raschle, S. Arens, A. Viegas, V. Gelev, M. Etzkorn, and G. Wagner, “The use of
amphipols for NMR structural characterization of 7-TM proteins,” J. Membr. Biol., vol. 247, no. 0, pp. 957–964, Oct. 2014.
[31] S. Ressl, A. C. Terwisscha van Scheltinga, C. Vonrhein, V. Ott, and C. Ziegler, “Molecular basis of transport and regulation in the Na+/betaine symporter BetP,” Nature, vol. 458, p. 47, Mar. 2009. [32] Reste, Ludovic & Hohlbein, Johannes & Gryte, Kristofer & Kapanidis, Achillefs. Characterization of
Dark Quencher Chromophores as Nonfluorescent Acceptors for Single-Molecule FRET. Biophysical journal. 102. 2658-68.2012.
[33] Ploetz, E., Lerner, E., Husada, F., Roelfs, M., Chung, S., Hohlbein, J., Weiss, S. & Cordes, “Forster resonance energy transfer and protein-induced fluorescence enhancement as synergetic multiscale molecular rulers “Scientific Reports. 6, 18 p., 33257,2016.
