University of Groningen Single-molecule fret study on structural dynamics of membrane proteins Aminian Jazi, Atieh

(1)

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.

(2)

Single-Molecule FRET Study on Structural Dynamics

of Membrane Proteins

Atieh Aminian 2020

(3)

2

Atieh Aminian PhD thesis

University of Groningen, The Netherlands

Zernike Institute for Advanced Materials PhD thesis series 2020-14

ISBN:978-90-830-7040-7

ISSN:1570-1530

The research described on this thesis was carried out in University of Groningen in the Zernike Institute of Advanced Materials of University of Groningen, the Netherlands . This work was financed by the Zernike Institute for Advanced Materials, the Centre for Synthetic Biology and ERC Starting Grant (ERC-STG 638536 – SM-IMPORT to T.C.). Atieh Aminian was supported by Ubbo Emmius funding (University of Groningen) and by the DFG in the frame of SFB699. E.P. was supported by a DFG fellowship (PL696/2-1).

Cover Design by: Atieh Aminian

(4)

3

Single-Molecule FRET Study on

Structural Dynamics of Membrane

Proteins

Proefschrift

ter verkrijging van de graad van doctor aan de Rijksuniversiteit Groningen

op gezag van de

rector magnificus prof. dr. C. Wijmenga en volgens besluit van het College voor Promoties.

De openbare verdediging zal plaatsvinden op

dinsdag 2 juni 2020 om 9.00 uur

door

Atieh Aminian Jazi

geboren op 04 September 1987 Tehran, Iran

(5)

4

Promotores

Prof. dr. T.M. Cordes Prof. dr. B. Poolman

Copromotor

Dr. G. Gkouridis

Beoordelingscommissie

Prof. dr. K. Duderstadt Prof. dr. W.H. Roos Prof. dr. H. Jung

(6)

5

Chapter 1:

(8)

7

Introduction & Background

Atieh Aminian Jazi, Christine Ziegler & Thorben Cordes(unpublished)

Integral membrane proteins perform a range of key functions that are crucial to cells such as material and information exchange with the environment, energy generation and transformation, and cell communication and signalling [4]. Membrane transport proteins, the biological focus of this thesis, mediate the solute transport across the membrane bilayer. Some have roles in the uptake of nutrients and other essential molecules [2], the generation and maintenance of ion gradients [3,12], the extrusion of waste metabolites and toxins [5,4], and the recapture of neurotransmitters and many other substances [2,6].

Given the importance of membrane proteins in so many biological processes, often occurring at the cell surface, it is no surprise that when these processes fail, diseases result [7]. Consequently, membrane proteins constitute the prime target for approximately 60% of all available pharmaceutical drugs [2]. An important step in designing novel therapeutic strategies is thus a detailed understanding of their molecular mechanisms. Structural and functional studies of transporters, aimed at understanding their molecular mechanisms, requires a combined understanding of (static) protein structure and structural dynamics something that I aspire here for active transporters. In this thesis, l will describe my efforts to establish novel methods to allow the single-molecule investigation of structural heterogeneity and dynamics of substrate-binding domains of ATP-binding cassette transporters [9,12] TRAP-transporters [11]and the secondary transporter BetP from Corynebacterium glutamicum [1,13].

The thesis contains the following chapters: Chapter 1 provides the scientific background on membrane transporters, the model protein BetP, all biophysical methodology and a synopsis of all scientific parts. In Chapter 2, I establish novel labelling schemes and advanced fluorop hores for membrane proteins. In Chapter 3, I overcome the hurdles to label BetP, as a trimeric protein with labels for single-pair FRET studies. In Chapter 4, I use the established methodologies to elucidate the activation mechanism of BetP, i.e. the conformational changes of the regulatory

(9)

C-8

terminal domain upon activation by the non-physiological stimulus K+_{. Chapter 5 summarizes and} discusses all results in light of the current state-of-the-art.

1.1 Background on membrane transporters

Molecules such as ions, peptides, small molecules, lipids and macromolecules and others are not only passively transported via channels or pores down their concentration gradient but can also be actively transported against chemical gradients. Membrane transporters are classified into channels or carriers (transporters) based on their biochemical functions, structure and location within the membrane. Transporter function by the alternating access model of transport in which, the substrate binding site can only and specifically be accessed from one side of the membrane at a time [14,13] (Figure 1.1). Many integral transmembrane proteins span the lipid-bilayer via stabilizing hydrophobic interactions [38]. Integral membrane transporters exhibit high specificity for their substrates. In contrast peripheral membrane protein are able to dissociate from the membrane to participate in a variety of cellular interactions such as lipid transport and cell signalling. Peripheral membrane proteins and their interactions with the lipid -bilayer are stabilized by both electrostatic or hydrogen bonds or other non-covalent interactions between the polar heads and polar regions of the protein [13,14].

Figure 1.1| Functional classification of membrane transport proteins: Cartoon sketch of different membrane proteins that passively or actively drive transport of substrates across membranes: (a) channels/pores that facilitate passive diffusion, (b) primary-active transporter where conformational changes in a transport protein

are driven by, e.g., ATP-hydrolysis and (c) secondary-active transporters that use an electrochemical gradient to co-transport a substrate. Substrate is shown as a black dot (a/b/c) and the co-substrate as a smaller red dot (c).

(10)

9

Active transport requires chemical energy stored in electrochemical gradients of ions (co-substrate) or molecules such as ATP to move the main substrate across the membrane. Primary-active transporter utilize the energy provided by ATP hydrolysis, for example in ABC transporters [10], to allow alternating access of a ligand binding site in the transmembrane domains. Secondary-active transporters mediate the uphill transport of the substrate by coupling it to the downhill flow of a co-substrate, often an ion (for example a proton or sodium-ion) in a thermodynamically unfavourable direction across the cell membrane, see Figure 1.1. The transporter investigated here is BetP [8,20] which is a member of the BCCT family of secondary transporters (betaine-choline-carnitine-transporters). These secondary transporters mediate the transport of one substrate, mostly a substrate with a trimethylammonium group against a concentration gradient using co-transport or counter-transport of another substrate or an ion. Members of the BCCT family are found in Gram-positive bacteria (BetL from Listeria monocytogenes [12,13]), Gram-negative bacteria (e.g. OpuD from Vibrio cholerae [43]) but also in archaea (e.g. glycine betaine transporter from Methanogenium cariaci [2,26]).

1.2 Scientific background on sodium-symporter BetP

BetP from Corynebacterium glutamicum is a well-characterized member of the BCCT family, Transporters belonging to the BCCTs family share common functional and structural features related to trimethylammonium substrate specificity [13], transported substrates contain a quaternary ammonium group. BetP’s function is transport of betaine via a sodium-symport mechanism. Crystal structures reveal that BetP is an asymmetric trimer [21]. Moreover, they show that a conserved tryptophan motif, the signature motif of the BCCT family, forms the binding site for the quaternary ammonium [15]. The transport core consists of two tightly nested structurally inverted repeats of five transmembrane helices allowing occlusion of substrate during transport [21].

Crystal structures demonstrated that BetP is an asymmetric trimer (Figure 1.2A), in which each of three protomers can independently adopt distinct conformations for transport. These were assigned as individual transport states in the alternating access cycle that allow uphill substrate

(11)

10

transport driven by the electrochemical Na+_{potential leading to accumulation of glycine betaine,} which is the exclusive substrate for BetP, to molar amounts in the cytosol under hyperosmotic conditions [13]. BetP requires both hyperosmotic stress as well as an elevated internal K+ concentration for full activation. However, as K+ _{is always above the Kd of BetP (220mM) in E.} coli phospholipid liposomes , the K+_{activation observed in proteoliposomes is considered as a} non-physiological stimulus [1,23]. Anyway, the detailed mechanism of stress sensing and activation is not yet described on atomistic level, but multiple functional and biochemical data have identified the C-terminal domain being key for sensing and regulation in BetP (Figure 1.2B). Thus, BetP has two distinct functions: sensing of osmotic stress and regulated transport of glycine betaine. It is assumed that the 45 amino-acid long C-terminal helix domain of BetP binds cytoplasmic K+_,_{which is a pre-requisite to activate BetP during hyperosmotic stress [21,25]. The} catalytic domain of BetP consists of 12 transmembrane helices (TM) and is divided into a transporter core of two inverted five-helix-repeats (TM3-TM12) and the two N-terminal helices TM1-2, which contribute to the trimer contacts. The symmetry between two repeats (TM3-TM7 and TM8-TM12) is a key to the alternating access mechanism in BetP [39]. Due to the variety of structural and biochemical studies, BetP serves as a paradigm for the mechanism osmoregulated transporters [28].

Figure 1.2| Structural and functional characteristics of BetP sodium-symporter from the BCCT family. Structure and FRET properties of the homotrimeric C252T/S516C BetP mutant. (A) Side and top views of the crystal structure of the mutant marking the three label positions and related distances. Protein Data Bank entry 4AIN. (B)

Normalized uptake rate of Cys-less BetP (wt) and BetP cysteine mutant C252T/S516C in E. coli cells depending on osmotic stress. The relative rate of uptake of 14C-labeled betaine for wild-type protein (green) is comparable to that of the mutant protein (red, C252T/S516C), which exhibits one-third of the total wild-type activity.

(12)

11

The 45 amino-acid long positively charged C-terminal domain (Figure 1.3, positively-charged residues in C-terminal domain are shown in blue) of BetP is suggested to sense directly hyperosmotic stress, however the stimulus that it is perceived by this domain is not known. As activation of BetP when reconstituted into proteoliposomes can be mimicked by increasing cytoplasmic K+_{concentration, the non-physiological K}+_{stimulus was used in the past to} investigate mechanistic features of activation, e.g. in spectroscopic measurements using AFM or PELDOR [23,28]. 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 surrounding biochemical environment of the helices [28]. The osmoregulatory function was seen in other transporters such as OpuA, a type I ABC importer [33]. Specific activation by potassium ions via a cytoplasmic C-terminal domain and instant regulation of transport activity was, however, only observed in BetP transporter [39].

Figure 1.3| BetP trimer in its surface representation. Positively-charged residues are color-coded in blue with possible label positions for fluorophores shown in orange. The blue colour positions indicate possible interaction within the C-terminals of adjacent protomers within BetP protein, PDB: 4C7R.

In order to study the activation mechanism of BetP in more detail, this thesis establishes smFRET experiments on BetP with the goal to study the conformational states of the C-terminal helix. For this purpose, we employed Förster-resonance energy transfer (FRET), a nanoscopic distance

(13)

12

ruler, in combination with single-molecule detection to investigate the conformational states and dynamics of BetP. smFRET provides a spatial resolution of sub-nanometers with a dynamic range of 2-10 nm with sub-second temporal resolution [40]. This combination is ideal to understand the structural changes required for osmoregulation in BetP and to characterize the flexible structural part within the C-terminal domain.

From a biological viewpoint I mainly investigate the effect of potassium-induced activation and conformational switching of the BetP C-terminal domain using smFRET in combination with new crystal structures which were co-crystallized with K+_{and Rb}+_{. For this, we designed various BetP} cysteine-mutants based on a cysteine-less BetP (C252T, TM5) containing engineered cysteines at the periplasmic and cytoplasmic positions in transmembrane helices and C-terminal domain of BetP. One example position of a potential periplasmic residue (S516C) is shown in Figure 1.2A.

Besides the achievement to label BetP for smFRET studies another major adv ance in my work was the use of amphipols (Figure 1.4) to mimic the natural membrane environment. Amphipols are amphipathic polymers that were successfully used before in order to investigate the structure of membrane proteins in structural biology and spectroscopic studies, e.g., biochemical characterization of transporters or for cryo-EM [22,37]. In this study, we used Amphipol polymer (A8-35) to stabilize the BetP variants for characterization and biophysical experiments of BetP in a detergent free solution [27,22].

(14)

13

Figure 1.4| Chemical Structure of Amphipol (A8-35), polyacrylate-based APol. Amphipathic polymer that allows solubilisation of membrane proteins in a detergent-free aqueous solution [41].

The hydrophilic backbone of these polymers (Figure 1.4) interacts with hydrophobic elemen ts within membrane proteins with high affinity and covers the transmembrane regions and keeps them water soluble in a protein-amphipol complex. Such a strategy is especially useful for smFRET experiments with diffusing molecules in the absence detergent [27]. In chapter 4 of this thesis, we used Amphipol (A8-35) for the first time in smFRET studies of C-terminal domain in Amphipol-stabilized variants of BetP.

1.3 Scientific background on biophysical methods and smFRET

Förster Resonance Energy Transfer (FRET) has become a powerful method to probe the structural dynamics and conformational heterogeneity of biomolecules [1]. FRET is a fluorescent -based photophysical mechanism where excitation energy is transferred from an excited donor fluorophore (D) to an acceptor fluorophore (A) in close-proximity via non-radiative dipole-dipole coupling; Figure 1.5A [18,30]. Due to its steep distance dependence FRET can be used as a spectroscopic ruler with a dynamic range of 3-8 nm (Figure 1.5B).

(15)

14

Figure 1.5| The concept of FRET. A) An electromagnetic transmitter-receiver is a macroscopic analog for B) the molecular dipole-dipole coulombic interaction between donor, D, and acceptor, A, fluorophores. The dependence of the efficiency of energy transfer from D to A on their distance provides a molecular ruler with a high dynamic range on the 2-8 nm scale. C) Simplified Jablonski diagram of donor D in the presence of a FRET acceptor A. After excitation (kex) to the excited competing pathways deplete the excited state S1: (i) kD, T which is the sum of radiative and non-radiative decay rates from S1 to S0. Radiative decay results in fluorescence emission; (ii) Förster-type energy transfer kFRET to an acceptor fluorophore A decaying with kA. Figure panel A/B and corresponding caption were reprinted with permission from Lerner & Cordes et al., Science 2018. Figure panel C and corresponding caption was reprinted with permission from Ploetz & Lerner et al., Scientific Reports 2016.

FRET can be described using a Jablonski diagram that explains the occurrence of FRET, i.e., excitation energy transfer as an additional pathway for donor excited state deactivation in a non-radiative fashion. The transfer of excitation energy occurs from the excited donor fluorophore to the nearby acceptor through a non-radiative dipole-dipole interactions (Figure 1.5C) [18,34]. Ideally, the FRET efficiency E depends only on the distance between both donor and acceptor fluorophores providing the ruler character as shown in Figure 1.5B. This holds only for the case that the relative orientation of the donor-emission and acceptor-absorption transition-dipole moment, spectral parameters (donor-emission spectrum, acceptor absorption spectrum) [36,40] are constant whenever conformational changes occur. In this case the FRET efficiency can be calculated via ratio metric measurements of fluorescence intensity ratios or fluorescence lifetimes (eqn. 1):

Eqn. 1.

In eqn. 1 FDA is the acceptor-fluorescence intensity after donor excitation and FDD the donor-fluorescence intensity after donor excitation corrected for spectral cross-talk and quantum yield differences. DA is the fluorescence lifetime of the donor-acceptor (DA) species, D the fluorescence lifetime of the respective donor-only (D-only) species. R is the distance between donor and acceptor, and R0 marks a dye-pair specific separation where the FRET transfers efficiency is equal to 50% (see Figure 1.5B), For accurate calculation of R-values, knowledge of the Förster radius R0 and determination of accurate FRET values E are required [18,29].

(16)

15

FRET is a powerful technique to monitor both inter- and intramolecular interactions that induce changes in donor-acceptor distances when being present on biomolecules. Such applications include protein-protein interactions, estimation of intracellular ion concentrations, kinetic studies for analysis of the enzymatic activities and protein conformational changes [18,29]. Figure 1.6 shows the main method used in this thesis.

Figure 1.6| Dynamic structural biology using smFRET. A) Scheme of data of microscopy setup for µs-ALEX measurements using a confocal microscope. The experimental setup is a combination of single -molecule fluorescence microscopy and spectroscopy which can be used to determine conformational states or dynamics in solution including conformational heterogeniety. Figure panel B) and caption were reprinted with permission from Lerner & Cordes et al., Science 2018.

For my work, single-molecule detection via confocal microscopy was combined with Förster-resonance energy transfer (FRET) via ratiometric detection of donor- and acceptor fluorescence [35,42]. The setup was able to excite both donor- and acceptor fluorophore to produce fluorescent bursts whenever labelled proteins would diffuse through the confocal excitation/detection spot (Figure 1.6A, top and Figure 1.6B). This information was used to extract information on the existing conformational states of BetP and other proteins in solution and possible conformational dynamics.

1.4 Thesis outline

The thesis contains the following chapters with novel scientific contributions: In Chapter 2, I establish novel labelling schemes and advanced fluorophores for membrane proteins. In Chapter

(17)

16

3, I overcome the hurdles to label BetP, as a trimeric protein with labels for single-pair FRET studies. In Chapter 4, I use the established methodologies to elucidate the activation mechanism of BetP, i.e., the conformational state of the regulatory C-terminal domain, relation conformational dynamics to static structure snapshots from X-ray crystallography.

Fluorescence labelling of membrane proteins and isolated soluble domains (chapter 2) :

Site-specific labelling of proteins and biomolecular complexes is a fundamental prerequisite for Förster-resonance energy transfer (FRET) experiments. Yet, especially for hydrophobic membrane proteins, only few approaches yielding high donor-acceptor fractions and favourable photophysical properties are established. In this chapter I provide novel methods for labelling (membrane) proteins with green and red organic fluorophores. Methods with resin-immobilized proteins (substrate-binding proteins) and free-diffusing membrane proteins (BetP/BasC) are compared and evaluated. Finally, I present smFRET characterizations of the performance of fluorophore-dye conjugates (see Figure 1.7) on substrate-binding domain 2 of ABC importer GlnPQ as a model system.

Figure 1.7| Design concept for photostabilizer–dye conjugates. UAAs are used to combine an organic fluorophore covalently with a photostabilizer on a biomolecular target or linker structure. This strategy was used to improve the properties of fluorescent dyes on isolated membrane transporter domains.

Caged FRET (chapter 3): Förster-resonance energy transfer (FRET), a nanoscopic distance ruler,

in combination with single-molecule detection has become a powerful tool to investigate the structural dynamics of biomolecular systems [35,8]. Here in the third chapter of this thesis, we introduce and elaborate a novel single-molecule approach to overcome the problems arise when using regular smFRET technique to characterize a complex biochemical species, such as

(18)

multi-17

labeled DNA and trimeric BetP. FRET labelling of a multi-subunit complex leads to the generation of heterogeneous mixture of FRET species that hamper the precise smFRET data acquisition (e.g., donor-acceptor-acceptor, Figure 1.2).

Figure 1.8| Working principle of Cage FRET methodology with corresponding time traces. Cartoon view of the excitation volume where diffusing species produce only green signals (top left, caged acceptor) and both green and red signals (top right, UV-activated acceptor) with corresponding smFRET photon streams of caged (left bottom) and after UV activation process (right bottom).

In this work, we establish caging of cyanine fluorophores and caged rhodamine dyes as photoactivatable fluorophore, i.e., chemical deactivation fluorescence, for single-molecule FRET experiments of freely diffusing molecules. The novel ‘caged FRET’ method was used to investigate the structure of the homotrimeric membrane transporter BetP, as an example of a multi-subunit protein, and also nucleic acids containing more than two fluorescent labels. The obtained results revealed that chemical caging and photoactivation (“uncaging”) by UV light allows temporal uncoupling of convoluted fluorescence signals from multiple donor or acceptor molecules. Therefore, the recovered fluoresce signals are use d to extract the desired FRET-related information [8]. Our study suggests that ‘caged FRET’ technique is not only suitable for smFRET analysis of complex biochemical systems, but it may also be used to study the intermolecular details of low-affinity binding interactions with diffusion-based smFRET. I will focus on Caged FRET methodology to elucidate the structure of membrane transporter BetP on chapter 3 of this thesis.

(19)

18

smFRET studies of BetP (chapter 4): This chapter was dedicated to provide new information on

the “molecular switch model” of BetP regulation, which describes how the C-terminal domain of BetP is changing conformation in order to detaches itself from the membrane in order to up -regulate transport. While the alternating access mechanism of betaine-sodium coupling and transport is very well described from a structural perspective, the conformational switching of the C-terminal domain during activation is not known on molecular level. This is attributed to the fact that crystal contacts force the C-terminal domain always in one distinct conformation, in which only one out of three terminal domains are resolved entirely. Here, we assessed potassium-induced conformational changes of the C-terminal domain by employing Förster-resonance energy transfer (FRET) in combination with single-molecule detection – an approach that allowed us to elucidate the conformational changes of the entire -helical C-terminal domain by using different labelling positions. For this, we first developed a direct approach to characterize the structural arrangement of BetP in different biochemical environments with homotrimeric cysteine mutants in the C-terminal helix (Figure 1.9A). We observed major differences in the conformational states in polymeric amphipols compared to detergent environment (Figure 1.9B). The observed differences and specifically the conformational space of the C-terminal domain identified in amphipol likely reflects a more native membrane-like state than detergent. Finally, we investigated the effect of potassium on the conformational states of the C-terminal domain. We demonstrate that in a negatively charged membrane -mimic conformational distribution of the C-terminal helices is narrowed down in response to increasing the concentration of K+_{(Figure 1.9B). We assume that crystallization contacts populate one of} these conformations. From a new Rb+ co-crystallized BetP structure and in the light of our new FRET data we could deduce that one purpose of K+_{binding is to strengthen the interaction of a} Cterminal domain with the cytoplasmatic loop of the adjacent protomer by weakening lipid protein interactions. However, again based on our FRET data (which shows protein is in a pre -activated state when it is in a non-ionic detergent) can now attribute the conformation observed in the X-ray crystal structures of BetP to a pre-activated state, in which these lipid interactions are not present as the protein is in a non-ionic detergent. Our studies also show that neither sodium nor lithium have a comparable effect on the conformational state of the protein. Our

(20)

19

findings confirm the role of potassium as a pre-requisite for full-activation to counteract the strong lipid-protein interaction in C. glutamicum, membranes consisting of negatively charged lipids.

Figure 1.9| smFRET studies of BetP (584C homotrimer) using ALEX spectroscopy. A) 2D-ALEX histogram showing three distinct labelling species in E*/S plot, where only the central DA population was selected for further analysis. B) 1D FRET efficiency plots of BetP in DDM (detergent) with no potassium versus BetP in amphipol (AMP) at different concentrations of potassium, showing that the activated state of the protein is formed both at high potassium concentrations in the native-like amphipol-state and in DDM without potassium.

(21)

20 REFERENCES CHAPTER 1:

[1] Reinhard Krämer, Susanne Morbach, “BetP of Corynebacterium glutamicum, a transporte with three different functions: betaine transport, osmosensing, and osmoregulation “, Biochimica et Biophysica Acta (BBA), 2004.

[2] Giacomini KM, Huang SM, et al., Membrane transporters in drug development. Transporter Consortium Nat Rev Drug Discov. 2010.

[3] Lodish H, Berk A, Zipursky SL, et al. Molecular Cell Biology. Section 15.2, Overview of Membrane Transport 4th edition. New York: W. H. Freeman; 2000.

[4] Vinothkumar KR, Henderson R. Structures of membrane proteins. Q Rev Biophys. 2010;

[5] Masato Otsuka, Takuya Matsumoto, Riyo Morimoto, Shigeo Arioka, Hiroshi Omote, Yoshinori Moriya, A human transporter protein that mediates the final excretion step for toxic organic cations, Proceedings of the National Academy of Sciences Dec 2005.

[6] Neurotransmitter transporters: molecular function of important drug targets, Trends in Pharmacological Sciences, Volume 27, Issue 7,2006.

[7] Sambamurti, Kumar; Suram, Anitha; Venugopal, Chitra; Prakasam, Annamalai; Zhou, Yan; Lahiri, Debomoy K.; Greig, Nigel H.Partial Failure of Membrane Protein Turnover May Cause Alzheimer's Disease: A New Hypothesis, Current Alzheimer Research, Volume 3, , pp. 81-90,2006.

[8] Jazi AA, Ploetz E, Arizki M, et al. Caging and Photoactivation in Single-Molecule Förster Resonance Energy Transfer Experiments. Biochemistry.;56(14):2031–2041.2017.

[9] Gouridis, G., Schuurman-Wolters, G., Ploetz, E. et al. Conformational dynamics in substrate-binding domains influences transport in the ABC importer GlnPQ. Nat Struct Mol Biol 22, 57–64,2015,

[10] Higgins, C.F. ABC transporters: from microorganisms to man. Annu. Rev. Cell Biol. 8, 67–113 .1992. [11] Christopher Mulligan, Marcus Fischer, Gavin H. Thomas, Tripartite ATP-independent periplasmic (TRAP) transporters in bacteria and archaea, FEMS Microbiology Reviews, Volume 35, Issue 1, January 2011.

(22)

21

[12] Biemans, Oldehinkel, E., Mahmood, N.A. & Poolman, B. A sensor for intracellular ionic strength. Proc. Natl. Acad. Sci. USA 103, 10624–10629 ,2006.

[13] Ziegler, C., Bremer, E. and Krämer, R. The BCCT family of carriers: from physiology to cryst al structure. Molecular Microbiology, 78: 13-34, 2010.

[14] Henderson RK, Fendler K, Poolman B. Coupling efficiency of secondary active transporters. Current Opinion in Biotechnology, 2019.

[15] Lomize APogozheva, I.D., Lomize, M.A. et al. The role of hydrophobic interactions in positioning of peripheral proteins in membranes. BMC Struct Biol 7, 44 ,2007.

[16] Oshy, C, Schweikhard, E.S., Gärtner, R.M., Perez, C., Yildiz, Ö. and Ziegler, C. Structural evidence for functional lipid interactions in the betaine transporter BetP. The EMBO Journal, 32: 3096-3105, 2013.

[17] 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, 1996.

[18] S. Sindbert et al., “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, 2011.

[19] D. Kapfhammer, E. Karatan, K. J. Pflughoeft, and P. I. Watnick, “Role for Glycine Betaine Transport in Osmoadaptation and Biofilm Formation within Microbial Communities,” Appl. Environ. Microbiol., vol. 71, no. 7, p. 3840 LP-3847, 2005.

[20] C. Perez, C. Koshy, S. Ressl, S. Nicklisch, R. Krämer, and C. Ziegler, “Substrate specificit y and ion coupling in the Na& betaine symporter BetP,” EMBO J., vol. 30, no. 7, p. 1221 LP-1229, 2011. [21] C., Faust, B., Mehdipour, A.R., Francesconi, K.A., Forrest, L.R., and Ziegler, C. Substrate-bound

outward-open state of the betaine transporter BetP provides insights into Na+ coupling. Nature communications. 5:4231, 2014.

(23)

22

[22] V.Polovinkin et al., “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.

[23] C., Khafizov, K., Forrest, L.R., Krämer, R., and Ziegler, C. The role of trimerization in the betaine transporter BetP. EMBO Rep. 12:804- 810. 2011.

[24] F.M. Goñi, “The basic structure and dynamics of cell membranes: An update of the Singer– Nicolson model,” Biochim. Biophys. Acta - Biomembr., vol. 1838, no. 6, pp. 1467–1476, 2014. [25] 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.

[26] M., Krämer, R., and Morbach, S. Characterization of compatible solute transporter multiplicity in Corynebacterium glutamicum. Applied microbiology and biotechnology. 76:701-708. 2007. [27] M. Zoonens, F. Giusti, F. Zito, and J.-L. Popot, “Dynamics of Membrane Protein/Amphipol

Association Studied by Förster Resonance Energy Transfer: Implications for in Vitro Studies of Amphipol-Stabilized Membrane Proteins,” Biochemistry, vol. 46, no. 36, pp. 10392–10404, 2007. [28] 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, 2008.

[29] 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, pp. 455–514, 2010.

[30] Truong, K., and Ikura, MThe use of FRET imaging microscopy to detect protein– protein interactions and protein conformational changes in vivo. Curr Opin Struct Biol. 11:573-578. 2001.

[31] C. Perez, B. Faust, A. R. Mehdipour, K. A. Francesconi, L. R. Forrest, and C. Ziegler, “Substrate-bound outward-open state of the betaine transporter BetP provides insights into Na+ coupling,” Nat. Communication, vol. 5, p. 4231, 2014.

(24)

23

[32] H. Peter, A. Burkovski, and R. Krämer, “Osmo-sensing by N- and C-terminal Extensions of the Glycine Betaine Uptake System BetP of Corynebacterium glutamicum,” J. Biol. Chem, vol. 273, no. 5, pp. 2567–2574, 1998.

[33] C. Horn, S. Jenewein, L. Sohn-Bösser, E. Bremer, and L. Schmitt, “Biochemical and Structural Analysis of the Bacillus subtilis ABC Transporter OpuA and Its Isolated Subunits,” J. Mol. Microbiol. Biotechnol., vol. 10, no. 2–4, pp. 76–91, 2005.

[34] B. Wallace and P. J. Atzberger, “Förster resonance energy transfer: Role of diffusion of fluorophore orientation and separation in observed shifts of FRET efficiency,” PLoS One, vol. 12, no. 5, p. e0177122, 2017.

[35] J. R. Widom, S. Dhakal, L. A. Heinicke, and N. G. Walter, “Single-molecule tools for enzymology, structural biology, systems biology and nanotechnology: an update,” Arch. Toxicol., vol. 88, no. 11, pp. 1965–1985, 2014.

[36] S. Bhuckory, O. Lefebvre, X. Qiu, K. D. Wegner, and N. Hildebrandt, “Evaluating Quantum Dot Performance in Homogeneous FRET Immunoassays for Prostate Specific Antigen,” Sensors (Basel)., vol. 16, no. 2, p. 197, 2016.

[37] A. C. Leney, L. M. McMorran, S. E. Radford, and A. E. Ashcroft, “Amphipathic Polymers Enable the Study of Functional Membrane Proteins in the Gas Phase,” Anal. Chem., vol. 84, no. 22, pp. 9841– 9847, 2012.

[38] J. U. Bowie, “Stabilizing membrane proteins,” Curr. Opin. Struct. Biol., vol. 11, no. 4, pp. 397–402, 2001.

[39] 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.

[40] 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.

[41] Hydrophobically driven attachments of synthetic polymers onto surfaces of biological interest: Lipid bilayers and globular proteins, Biochimie, Volume 80, Issues 5–6,1998.

(25)

24

[42] 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.

[43] C. Dagmar Kapfhammer, Ece Karatan, Kathryn J. Pflughoeft, Paula I. Watnick, Role for Glycine Betaine Transport in Vibrio cholerae Osmoadaptation and Biofilm Formation within Microbial Communities, Applied and Environmental Microbiology,2005.

(26)

25

Chapter 2:

Fluorescence labelling of membrane

transporters and isolated protein domains

(27)

26

2. Novel approaches for fluorescence labelling of membrane transporters and isolated protein domains

Abstract:

FRET has become a widely used method in combination with single-molecule detection to investigate protein conformational dynamics and heterogeneity. One of the main factors limiting mechanistic insights are poor photophysical properties of dyes and the ne ed for specific attachment strategies of fluorophore to the biomolecule [7]. Site-specific labelling is a crucial step in order to investigate biological systems with smFRET-based single-molecule methods, since it needs to map a relevant reaction coordinate of the system [21]. For proteins, cysteines are artificially introduced at non-conserved and solvent-exposed positions as anchor points for the fluorescent labels. While this strategy has various problems that deal with the molecular biology and preparation of the protein mutants (and their functionality), it is further complicated in multi-subunit proteins such as BetP since one cysteine (in one subunit) appears multiple times in the complex; thus multiple labelling sites are created. Designing an appropriate strategy for site-directed mutagenesis and further conjugation of fluorophores with optimized photophysical properties for FRET will allow to obtain more mechanistic insights. This chapter presents attempts to label homo- and heterotrimeric cysteine constructs of the membrane protein BetP for smFRET studies. Finally, I show labelling of ABC-transporter domains from the amino-acid importer GlnPQ as test systems for characterization of photostabilizer-dye conjugates. With this I established methods to mechanistically study the osmotic stress regulated sodium-coupled betaine symporter BetP. The long-term goals of our studies were monitoring catalytic movements of BetP during substrate transport under up-regulated conditions.

(28)

27

2.1 Site-specific labelling strategies for smFRET studies of BetP

Atieh Aminian Jazi, Reinhard Krämer, Christine Ziegler & Thorben Cordes(unpublished)

Cysteine labelling of homotrimic BetP: We first describe our studies on improving the labelling

of different BetP cysteine variants with high specificity and efficiency for studies of conformational dynamics in BetP. For this purpose, we first selected and characterized the residues of BetP that allow to establish reactive movements of the protein, then we identified non-conserved residues in cytoplasmic and periplasmic side for label attachment. Also, solvent accessibility of the residues within the protein sequence was determined by using the SWISS-MODE server on crystal structures.

Amino acids at positions, which neither affected functionality when mutated nor w ere considered to be involved in conformational changes were replaced by cysteine residues by site -directed mutagenesis. BetP variants were constructed on a cysteine -less BetP (C252T, TM5) as an initial template [26]. Then we demonstrated that the selected mutations had no critical effect on the transport function and activation profile of BetP. For labeling, disulfide bonds are formed in the interaction between sulfhydryl groups of the cysteine residue and commercially available fluorophore-maleimide conjugates. To obtain optimal results we investigated different fluorophore labeling techniques and biochemical experimental approaches to obtain BetP with higher degree of labelling. For this purpose, BetP was first solubilized in DDM detergent and purified via StrepTactin-affinity chromatography (Figure 2.1). We then conducted the FRET labeling (Figure 2.1C). The data shown in Figure 2.1 are the result of a long optimization process. For this we performed the set of experiments in combination with quantitative analysis for varying labeling conditions including variations of buffers, pH values, reducing agents and salts, different incubation times, temperature and distinct molar ratios of protein to fluorophores during labeling. In the first iteration of our final protocol we used 50mM Tris pH7.4 containing 0.05%-0.1% n-dode cyl β-D-maltoside (DDM), during purification of a Strep-Tactin® Macro-Prep column via the gravity flow method. To this a fluorophore mixture (of donor and acceptor-dye) were applied to immobilized BetP. The eluent fraction was subject to SEC-analysis and

(29)

28

revealed a high amount of non-specific dye-interactions with the column material. The amount of non-specific interaction increased also for hydrophobic fluorescent FRET probes such ATTO555 (Donor) and ATTO647N (Acceptor). Labelling with resin-immobilized BetP turned out to be impractical.

Figure 2.1| overexpression, protein purification and fluorophore labelling of BetP wild type and mutants.A) Western blot of BetP expression within E. coli cells, DH5αTM-T1R E. coli competent cell were transformed and were cultivated in LB media supplemented with 50 µg/ml carbenicillin. Protein production of BetP was induced through the addition of 200 µg/l Anhydrotetracycline hydrochloride (AHT) at an Optical Density (OD) of 600 of 0.5 and then cells were harvested for analysis after 2 hours. As marker Page Ruler Prestained Protein Ladder (Thermo Scientific Molecular Biology) was used, Lane 1: loaded cells before AHT cell induction. Lane 2: loaded cells after +2 hours AHT induction. B) BetP protein purification SDS-PAGE analysis. Solubilized membrane (10 mg/ml) were loaded onto an SDS-PAGE (12.5 %). Samples collected during purification and were examined by SDS PAGE and Western blot analysis and Each well was loaded with different fractions collected from Strep Tag purification. BetP wild type protein purification (left picture) and L516C_Transmembrane homotrimeric mutant SDS-PAGE (right picture). C)

(30)

Size-29

exclusion chromatography profile of labelled Betp_L516C mutant shows a single protein peak at 280 nm (blue)and absorptions at 550 nm and 660 nm(red) were monitored and corresponds to labelled protein with FRET donor fluorophore Alexa 555 and FRET Acceptor ALEXA 647N fluorophore. Protein was injected on a Superose 6 GE 10/300 column with a constant flow rate of 0.5ml/min and protein absorption was monitored at 280 nm. Green box shows selected labelled protein fraction for further smFRET analysis. D) Size-exclusion chromatography profile of labelled Betp_K489C mutant shows a single protein peak at 280 nm (blue)and absorptions at 550 nm and 660 nm(red) were monitored and corresponds to labelled protein with FRET donor fluorophore Alexa 555 and FRET Acceptor ALEXA 647N fluorophore. Green box shows selected labelled protein fraction for further smFRET analysis.

Subsequently, we tried to label the fresh purified protein and removed unbound labels via dialysis methods. For this, we performed series of dialysis experiments to remove the reducing agent dithiothreitol (DTT) used for stabilization of the protein and its cysteins followed by fluorophore incubation of the protein-to-dye ratio of Acceptor and donor 5:4 for ~4 nmol of protein in presence of 0.1mM, 0.5mM and 1M DTT in a deoxygenated reaction buffer. Subsequen tly, the unreacted dyes were washed with sequential dialysis steps to stop cross-reactions of DTT with the maleimide moiety. DTT prevents the interaction of thiol-specific reagents with the cysteine in further steps. The size-exclusion chromatography of labeled BetP protein revealed that the labeling efficiency was increased compared to convenient methods yet indicated protein aggregation. Furthermore, we replaced DTT by TCEP (tris(2-carboxyethyl) phosphine) and also protein immobilization on Strep tag column with desalting method as alternatives for smFRET labeling of BetP.

As a final result of our optimization process, we established a fast and reliable method to label BetP with a minimum incubation time using desalting methods in combination with an optimized fluorophore to protein ratio. This desalting technique aims at removing buffer salt, reducing agent and additional additives directly from purified protein, which was solubilized in detergent or amphipol, in a fast exchange for water. The key for success was to quickly desalt the freely-diffusing protein and to fully remove reducing agent, to allow a start of the sulfhydryl-reaction prior to cysteine-oxidation. Compared to the dialysis method, this approach has the advantage of a much faster reaction, i.e., within a few hours and higher stability of protein. The details of the protocol are provided in the material and methods parts of chapter 3/4.

(31)

30

After labelling was established for BetP cysteine-mutants, we monitored catalytic movements of BetP during substrate transport. This was done by using structural information on BetP to identify residues that move significantly during the isomerization from inward-facing (IF) to outward-facing (OF) state in the alternating access cycle. As one example we used the d ouble-cysteine mutant S140C/K489C to monitor catalytic activity of BetP via smFRET (Figure 2.2). Postion S140 is located at the cytoplasmic start of TM3, which carries the so-called glycine stretch of BetP. Structure in OF and IF showed that this part of TM3 is displaced by about 6 Å. K489 is located at the beginning of TM11, which is part of the coupling scaffold. This residue is discussed as a lipid interaction side and therefore anchored in the membrane. Structural comparison of IF and OF state suggests that this residue is not moving during alternating access. This mutant was already used in PELDOR measurements and exhibits a WT-type like activation profile. The top view of the BetP crystal structure (Figure 2.2A/B) reveals the challenges of the FRET approach in this particular BetP double-cysteine mutant. The most reduced activity was detected for BetP S140C/C252T/K489C, although normalized data indicated that regulation of the transport for this BetP mutant is still maintained (Figure 2.2D). Because the replacement of K489 in this mutant of BetP seemed to have an impact on the transport activity level, while BetP S140C/C252T alone exhibit only a slight reduced uptake rate compared to the BetP WT. Since the protein is expressed and purified as a homotrimer, the two cysteine residues appear in each subunit. Stochastic labelling with donor and acceptor fluorophore, here Alexa555 and Alexa647, results in FRET histograms that cannot be easily interpreted (Figure 2.2C).

(32)

31

Figure 2.2| Problems in smFRET investigations of multimeric proteins. Crystal structure of BetP in the double cysteine context S140C/C252T/K489C. PDB: 4AIN. (A) Side view showing the planar arrangement of BetP with its three monomers (one highlighted in blue). (B) Top view of the crystal structure showing the trimeric structure of BetP. Positions of cysteines and (selected) distances between them are indicated; note that many more distances in particular between neighbouring monomers are relevant and only some selected ones are displayed. (C) 2D ALEX histogram of Alexa555/647-labelled BetPS140C/K489C showing the convolution of FRET interactions and difficulties to use this data for structure analysis. (D) NormalizedUptake measurements of [14C]-labeled glycine betaine as substrate for two BetP-Cysteine mutants S140C/K489C and S516C substrates in E. coli MKH13 in dependence of osmotic stress as explained previously (Ott et al., 2008). uptake rate of two BetP-Cysteine mutants S140C/K489C and S516C in E. coli cells The relative uptake rate of [14C]-labeled betaine in E. coli cells for wild type protein (green circle) is comparable to the one of both mutant proteins (blue: S140C/C252T/K489C; red C252T/S516C), which exhibit 1/3 of total wild type activity.

We then used µs-ALEX (microsecond alternating-laser excitation), where fluorescently-labeled biomolecules diffuse through the excitation volume of a confocal microscope, for smFRET studies of BetP. During their diffusional transit the labelled protein produces fluorescent bursts in two distinct detection channels that are chosen to selectively monitor donor and acceptor emission. In ALEX, green excitation of the sample generates fluorescent signals that allow calculation of apparent FRET via photon streams DD excitation, donor emission) and DA (donor-excitation, acceptor-emission) as E* = DA/(DD+DA). The stoichiometry S is obtained from the ratio of total fluorescence in both channels during green excitation to the total fluorescence during green and red excitation S= (DD+DA)/(DD+DA+AA); AA, acceptor excitation, acceptor emission.

In ALEX brightness ratio S (Y-axis in Figure 2.2C) distinguishes molecular species by their labelling ratio of green to red fluorophores. Low S < 0.2 is indicative of acceptor-only labelled protein,

(33)

32

while high S > 0.8 corresponds to a donor-only species. Macromolecules containing both dyes are found at S values between these two boundaries (Figure 2.2C, 0.2 > S > 0.8). It is apparent that fluorophore interactions via FRET are highly complex since each protein has six possible labeled positions resulting in a mix of label compositions (heterogeneity of labelled species, Figure 2.2B). In such a scenario the relation of FRET efficiency and interprobe distance R is lost due to the ambiguous interaction of e.g., multiple donor with multiple acceptor fluorophores or signal loss via homo-FRET and energy dissipation. For homotrimeric variants of BetP we had to restrict ourselves to study single cysteine mutants, for which still problems with multiple distinctly labelled species arise. A solution to these problems is detailed in chapter 3.

Protein purification and fluorophore labelling of heterotrimeric BetP:

While problems related to convoluted smFRET distributions with homotrimeric proteins with one cysteine can be solved via photochemical tricks and data sorting (see chapter 3), double-cysteine mutants such as S140C/K489C required a different strategy. More recently, a heterotrimer approach was used as an interprotomeric crosstalk in regulation (Becker et al., 2014). Luckily, for BetP a specific purification strategy was developed by the Krämer lab and conducted in our lab (Figure 2.3 for Heterotrimeric BetP overexpression and purification) that resulted in a heterotrimeric protein complex, which we utilized here to label one protomer specifically for smFRET.

(34)

33

Figure 2.3| Heterotrimeric BetP overexpression, purification and fluorophore labelling of heterotrimer BetP smFRET of successfully purified and labelled heterotrimeric construct. A) Western blot analysis of Heterotrimeric BetP Sterp tag purification. Protein production of Heterotrimeric BetP were induced through the addition of 200 µg/l Anhydrotetracycline hydrochloride (AHT) at an Optical Density (OD) of 600 of 0.5 and then cells were harvested for analysis after 2 hours. As marker Page Ruler Prestained Protein Ladder (Thermo Scientific Molecular Biology) was used. B) Western blot analysis of Heterotrimeric BetP His tag protein purifications. C) Western blot analysis of Heterotrimeric BetP FLAG tag protein purification. D) Heterotrimeric BetP protein purification SDS PAGE (12.5 %). analysis. Solubilized membrane (10 mg/ml) were loaded onto an SDS PAGE Samples collected during purification and were examined by SDS-PAGE after Strep tag, His tag and FLAG tag purifications. E) protein structures of Heterotrimeric BetP (S140C/K489C) with label positions of donor and acceptor, PDB=4C7R. F) Corresponding two-dimensional histogram (2D) of Heterotrimeric BetP ( S140C/K489C) with Alexa555/Alexa647 FRET pair with values of S (labelling stoichiometry, dashed lines 0.4 < S < 0.6 donor-acceptor population DA) and Apparent FRET values of E*= 0.90 ±0.1 with excitation intensities of 30 kW/cm2_{at 532 nm and 20 kW/cm}2_{at 640 nm.}

(35)

34

Biochemical activity and functionality of labelled BetP: Another important point of

consideration relates to the biochemical function of the protein after fluorescent labelling, i.e., the labels may interfere with proper folding or function of the transporter. Thus, assays are needed that directly compare the protein activity and its degree of labeling as control experiments to conduct a relevant biophysical study. In Figure 2.2D, the normalized uptake rates of two BetP-Cysteine mutants S140C/K489C and S516C are shown in dependence of osmotic stress. We found a degree of FRET labelling of 76% for homotrimeric mutant BetP S140C-k489C and 40% for homotrimeric mutant L516C (See Figure 2.1 C, D). The relative uptake rate of [ 14C]--labeled glycine betaine for wild type proteins (green circle) was comparable to the one of both mutant proteins (See Figure 2.2 D, blue: S140C/C252T/K489C; red C252T/S516C). (Ultimately, the quality of the final FRET data is not only related to the functionality of the protein, but also to the degree of labeling and the percentage of molecules containing both the donor and the acceptor dye, since only those provide desired FRET information. Especially for smFRET studies, these two requirements, i.e., high labeling efficiency and retained biochemical functionality are challenging hurdles.

(36)

35

2.2 smFRET studies of substrate-binding proteins with photostabilizer-dye conjugates

Jasper HM van Der Velde, Jens Oelerich, Jingyi Huang, Jochem H Smit, Atieh Aminian Jazi, Silvia Galiani, Kirill Kolmakov, Giorgos Gouridis, Christian Eggeling, Andreas Herrmann, Gerard Roelfes, Thorben Cordes* (published in Nature Communications 2016)

A major limitation of smFRET experiments in general is related to currently available commercial fluorophores. The achievable count-rates and photobleaching properties of commercial fluorophore are restricting both spatial and temporal resolution of smFRET and other experiments. I will here characterize photostabilizer–dye conjugates as FRET donor-acceptor pairs in proteins.While synthetic organic fluorophores have been a major driving force for the recent success of fluorescence-based methods, they intrinsically suffer from transient excursions to dark states (blinking) and irreversible destruction (photobleaching) [8]. Both processes fundamentally limit their applicability and have, for a long time, hampered the development of advanced microscopy techniques with single -molecule sensitivity [9][10] or optical super-resolution <250 nm[11][12]. Lüttke and colleagues [13] introduced covalent binding of triplet-state quenchers and singlet-oxygen scavengers [9] to organic fluorophores as a strategy to reduce the above mentioned effects. Such photostabilizer–dye conjugates with intramolecular triplet-state quenching have ‘self-healing’[14] or ‘self-protecting’[15] properties, preventing photodamage without the use of solution additives. This non-invasive strategy has clear advantages compared with commonly used approaches, where micro- to millimolar concentrations of organic compounds are added to the biochemical buffer system.

We here establish the use of rhodamine- and carbopyronine dyes in conjugation to photostabilizers as FRET dyes. (S)-Nitrophenylalanine, NPA was used as a scaffold for

conjugation of a commercially available organic fluorophore (RhodamineB, ATTO647N and KK114), the photostabilizing p-nitrophenyl group [16][17] to a biomolecular target. Using maleimide chemistry, we facilitated the use of NPA-based photostabilizer–dye conjugates for direct labelling of proteins. For this purpose, two different synthesis strategies were

(37)

36

developed, accounting for the available quantity of the fluorophore. In the most straightforward case, that is, larger amounts of amine-reactive photostabilizer–dye conjugate are available, the NHS ester of the fluorophore (Figure 2.4, compound 17) can be coupled directly with 2-maleimidoethylamine (Figure 2.4 , compound 19), to yield a maleimide derivative of, for example, NPA–RhodamineB (Figure 2.4, compound 20).

Figure 2.4| Synthesis of reactive photostabilizer–dye conjugates of RhodamineB for direct labelling of primary amines and thiol residues. The strategy can be extended to other biochemical targets by a varying the linker of molecule. Adapted from van der Velde et al., Nature Communications 2016.

The second strategy is also feasible for small quantities of reactive fluorophore species (<10 mg) that could be due to high prices of commercially available precursors or complicate d synthesis. ATTO647N is a good example of a fluorophore that is often used in demanding fluorescence applications but is not readily available in large amounts. For these cases we optimized the synthesis, to yield a thiol-reactive derivative of ATTO647N containing the photostabilizer NPA (Figure 2.5, compound 25). As shown below both maleimide derivative s can covalently bind to recombinant proteins via solvent-exposed cysteine residues (Figure 2.4, compound 21 and Figure 2.5, compound 26).

To show the benefits of intramolecular photostabilization in fluorescence applications with proteins, we studied the substrate-binding domain 2 (SBD2) of the Lactococcus Lactis type 1 ABC transporter GlnPQ [18]. Using smFRET and alternating laser excitation (ALEX) spectroscopy method in combination with single molecule detection, the conformational

(38)

37

states of the protein were monitored (see Figure 2.6,a and b, open unliganded and closed liganded states of protein). The structural rearrangements of SBD2 on ligand binding cause a change of ∼0.9 nm regarding the distance between two selected amino acids in the protein [18]. As the FRET donor and acceptor fluorophore are attached via male imide chemistry at these positions in the protein (mutant of SBD2: T369C/S451C), the transfer efficiency E*reports on the conformational state of the protein, i.e., open or closed. As described previously [18], SBD2 was labelled stochastically using appropriate mixtures of donor and acceptor fluorophores.

Figure 2.5| Simplified synthesis of reactive photo stabilizer–dye conjugates where only small quantities of fluorophore are available. The resulting NPA–ATTO647N conjugate can be used for direct labelling of thiol residues, for example, in proteins (compound 25). The strategy can be extended to other biochemical targets by a variation of the linker molecule 19.

To understand the effects of intramolecular photostabilization in FRET-based assays, we used different fluorophore combinations: Cy3B or RhodamineB as donor fluorophores and red -emitting dyes such as ATTO647N and KK114 as the acceptor fluorophores. In experiments described further, either the donor (Cy3B or RhodamineB) or the acceptor (ATTO647N or KK114) was stabilized via covalent linkage to NPA. We used smFRET/ALEX, which allows to distinguish the desired proteins labelled with both donor and acceptor (do nor-acceptor) from those labelled with only donor (donor-only) or only acceptor (acceptor-only). Therefore, the

(39)

38

fluorescence emission of the donor-only (F(DD)), that of the acceptor when excited via FRET (F(DA)) and directly via red excitation light (F(AA)) is determined. In smFRET experiments, individual biomolecules are studied for short time periods of a few milliseconds while diffusing through the excitation volume of a confocal microscope. The challenge of such an experiment is to acquire intense fluorescent bursts during the short observation time under the required excitation intensity of 20-100 kW/cm2_.

Figure 2.6| (a) Crystal structures of the SBD2 (T369C/S451C) open (left panel, PDB: 4KR5) and closed state (right panel, PDB:4KQP, after binding of the ligand glutamine shown in red) with label positions of donor (D) and acceptor (A). (b) Corresponding one-dimensional histograms of E*-values for increasing amounts of ligand

(40)

39

glutamine where an increased E*=0.55±0.1 (dashed line, closed conformation) becomes visible (as opposed to the open conformation, E*=0.36±0.1, dashed line). Excitation intensities of 30 kW cm−2_{at 532 nm and}

20 kW cm−2_{at 640 nm; data evaluated with dual colour burst search and displayed with additional per -bin}

thresholds of F(DD)+F(DA)+F(AA)>75 and minimal number of counts per bin in ALEX histogram of 3. (c,d,e) 2D histograms of joint pair values of S (labelling stoichiometry) and E* (FRET efficiency, that is, interprobe distance) of Cy3B/ATTO647N in the presence (c) and absence (d) of 2 mM TX and Cy3B/NPA–ATTO647N (e) without ligand glutamine. Excitation intensities of 30 kW cm−2_{at 532 nm and 20 kW cm}−2_{at 640 nm; data evaluated with all}

photon burst search and displayed with additional per -bin thresholds of F(DD)+F(DA)+F(AA)>100 and minimal number of counts per bin in ALEX histogram of 3. (f,g,h) Histogram of fluorophore brightness values as determined from photon-counting histograms (PCHs) on single-molecule transits of labelled SBD2 diffusing through the observation volume, comparison of donor brightness F(DD) (f), acceptor brightness when excited via FRET, F(DA) (g) and acceptor brightness via direct red excitation F(AA) (h). Excitation intensities of 30 kW cm−2_{at 532 nm and 20 kW cm}−2_{at 640 nm. (i,j,k) Dependence of the mean count rate of F(DD), F(DA)}

and F(AA) of the different samples for increasing excitation intensity, respectively.

Results of such quality are, however, only available when using 2mM Trolox as a photostabilizer in FRET measurement solution, seen from comparison in Figure 2.6. Here we show data of apo-SBD2 protein (labelled with Cy3B/ATTO647N) in the presence and absence of Trolox compound in buffer solution. In agreement with Kong et al. [10], the high excitation intensities used in our experiments promote acceptor signal fluctuations, that is, blinking and/or bleaching. Cy3B and ATTO647N can hence be seen as a FRET couple where the acceptor photostability is limiting. This appears as a prominent bridge between the donor-only and donor-acceptor population (see Figure 2.6 d), altering both E*/S-values substantially. Closer inspection of the FRET histogram reveals that a significant portion of the molecules show these unwanted photophysical effects. Under these conditions, neither mean E* nor correction factors for accurate FRET determination are directly accessible. Besides the complete loss of information, the overall acquisition time has also increased in the absence of photostabilize r to obtain sufficient statistics from the relevant donor–acceptor species. It should be noted that such photophysical artefacts of the acceptor dye (Figure. 2.6d) are extremely problematic for data interpretation, as they suggest the existence of (non-biological) species in between the donor only and the actual FRET species (see 1D−E* in (Figure. 2.6d) that can only be fitted by the sum of two Gaussians with E*=0.19±0.07 and 0.43±0.1). Furthermore, it remains challenging to separate those two FRET species from D-only molecules when performing smFRET with green excitation in continuous-wave mode.

(41)

40

Strikingly, the bridge population can be removed by sole photostabilization of the acceptor fluorophore via scaffolding of ATTO647N to NPA. The ALEX data of apo-SBD2 with Cy3B/NPA– ATTO647N are shown in Figure 2.6e. Here the bridge between the donor–acceptor and donor-only population is fully removed without the addition of stabilizer to the solution (Figure 2.6d versus Figure 2.6e). The resulting mean E*/S values are changed compared to apo-SBD2 with Cy3B/ATTO647N (2mM Trolox in solution, Figure. 2.6c) accounting for decreased acceptor brightness; mean E* is now found at 0.34±0.09 and mean S at 0.67±0.08 (Figure. 2.6e). These differences are, however, expected considering the results from ATTO647N on DNA scaffold, where a decrease in the overall brightness is observed on conjugation to NPA to ATTO647N fluorophore. It should be mentioned that differences in the donor–acceptor population relative to donor and acceptor only comparing the samples Cy3B/ATTO647N and Cy3B/NPA–ATTO647N are not solely due to photophysics but also due to different labelling ratios of dyes and protein.

Next, we used Cy3B/NPA–ATTO647N to study the biomolecular function of SBD2 (see Figure. 2.6b). Upon addition of the ligand glutamine, the mean E* changes from 0.36±0.1 (open, interprobe distance of ∼4.9 nm) to 0.55±0.1 (closed, with a decreased interprobe distance of ∼4.0 nm). A concentration of 200 μM glutamine saturates the ligand binding and therefore results in a 100% population of the closed state of protein ( Figure. 2.6b). A ligand concentration of 1 μM, which is close to the Kd-value of the protein, consequently results in a mixture of open and closed conformational states of protein ( Figure. 2.6b)

As fluorophore brightness and the resulting photon budget ultimately determine the quality of the final histograms including the necessary measurement time, we quantitatively analysed Cy3B/ATTO647N and Cy3B/NPA–ATTO647N by means of photon-counting histograms. Data in Figure 2.5 f,g,h shows the three relevant photon streams used to determine E* and S -values of Cy3B/ATTO647N (donor–acceptor: 0.9>S>0.4; bridge: 0.9>S>0.7) and Cy3B/NPA–ATTO647N (donor–acceptor: 0.9>S>0.4). F(DD) shows that the strongest donor quenching and hence the most efficient FRET is found for the addition of TX to Cy3B/ATTO647N, whereas molecules in

(42)

41

the bridge (Cy3B/ATTO647N, no TX) show inefficient donor quenching due to a non-functional acceptor (Figure. 2.5 f). Cy3B/NPA–ATTO647N and healthy labelled molecules from the Cy3B/ATTO647N population with no addition of Trolox show a similar performance of the acceptor-based donor quenching. Then F(DA) correlates directly with quenching in F(DD) as seen in Figure 2.5g. Again, the best performance is observed from Cy3B/ATTO647N in the presence of Trolox, while molecules in the bridge show the lowest counts. A striking difference between the bridge and all other conditions is seen in F(AA), that is, direct excitation of the acceptor (Figure. 2.6h). The overall comparison suggest that NPA-based acceptor stabilization is sufficient to remove photophysical artefacts and hence make the smFRET of these pairs useful for biomolecular studies without addition of Trolox. Further optimization of the data quality could, however, be obtained by additional donor stabilization of, for example, Cy3B in this case. This interpretation is supported by excitation intensity-dependent count rates of Cy3B/ATTO647N and Cy3B/NPA–ATTO647N (Figure. 2.6 i,j,k). NPA-based acceptor stabilization improves the saturation characteristics in all three channels but the addition of Trolox to the solution (resulting in stabilization of both donor and acceptor fluorophores at the same time) remains superior in terms of achievable count rates.

To study the donor dependence in more detail, we repeated the described experiments using RhodamineB/ATTO647N and NPA–RhodamineB/ATTO647N. Here, the photostability of RhodamineB as donor is the limiting factor for smFRET data quality. For RhodamineB/ATTO647N we find prominent donor blinking in the absence of photostabilize r (see Figure 2.7). The bridge between the donor–acceptor and acceptor-only population can be removed by addition of Trolox in solution (Figure 2.7) or conjugation of RhodamineB to NPA (Figure 2.7). The overall magnitude of the observed effects is lower than for the case of acceptor bleaching/blinking shown in Figure 2.6. Photon-counting histograms and the intensity dependence show a similar trend as before, that is, correlation between F(DD) and F(DA) with the wish to increase donor photostability as much as possible. Our data makes clear that NPA–RhodamineB can be used at significantly higher excitation intensities than