Interakce iontů s proteiny Ion-Protein Interactions

Pˇr´ırodovˇedeck´a fakulta Univerzity Karlovy v Praze Katedra fyzik´aln´ı a makromolekul´arn´ı chemie Charles University in Prague, Faculty of Science Department of Physical and Macromolecular Chemistry Modelov´an´ı chemick´ ych vlastnost´ı nano- a biostruktur Modelling of Chemical Properties of Nano- and Biostructures Autorefer´at disertaˇcn´ı pr´ace Summary of the Ph.D. Thesis

Interakce iont˚ u s proteiny Ion-Protein Interactions Mgr. et Mgr. Jan Heyda

´ ˇ v.v.i. Ustav organick´e chemie a biochemie, AV CR Centrum biomolekul a komplexn´ıch molekulov´ ych syst´em˚ u ˇ Skolitel/Advisor: Prof. Pavel Jungwirth, DSc.

Praha, 2011

CONTENTS

2

Contents 1 Abstract in English 1.1 Introduction into complex systems . . . . . . . . 1.2 Aims of the study . . . . . . . . . . . . . . . . . . 1.3 Computational methods . . . . . . . . . . . . . . 1.3.1 Simulation protocol . . . . . . . . . . . . . 1.3.2 Methods of analysis . . . . . . . . . . . . . 1.4 Results and discussion . . . . . . . . . . . . . . . 1.4.1 Case study – Model of the peptide bond . 1.4.2 Amino acid proxies . . . . . . . . . . . . . 1.4.3 Single amino acids and short oligopeptides 1.4.4 Peptides and proteins . . . . . . . . . . . . 1.5 Summary . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

5 5 6 6 6 7 8 9 10 10 11 11

2 Autorefer´ at v ˇ cesk´ em jazyce ´ 2.1 Uvod do komplexn´ıch syst´em˚ u . . . . . . . . . . . . . . 2.2 C´ıle pr´ace . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 V´ ypoˇcetn´ı metody . . . . . . . . . . . . . . . . . . . . 2.3.1 Simulaˇcn´ı protokol . . . . . . . . . . . . . . . . 2.3.2 Druhy anal´ yzy . . . . . . . . . . . . . . . . . . 2.4 V´ ysledky a diskuse . . . . . . . . . . . . . . . . . . . . 2.4.1 Pˇr´ıklad – studie modelu peptidov´e vazby . . . . 2.4.2 Modely funkˇcn´ıch skupin . . . . . . . . . . . . . 2.4.3 Jednotliv´e aminokyseliny a kr´atk´e oligopeptidy 2.4.4 Peptidy a proteiny . . . . . . . . . . . . . . . . 2.5 Z´avˇer . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

13 13 14 14 14 15 15 16 18 18 19 19

. . . . . . . . . . .

. . . . . . . . . . .

Bibliography

21

Curriculum vitae

24

Selected Publications/Seznam publikac´ı

25

Title: Ion-Protein Interactions Author: Mgr. et Mgr. Jan Heyda Department: Physical and Macromoleculer Chemistry Advisor: Prof. Pavel Jungwirth, DSc., IOCB AS CR, v.v.i. Advisor’s e-mail address: [email protected] Abstract: Conventional molecular dynamics simulations in combination with advanced methods of analyses were used to improve the understanding of the interaction between ions and proteins in salt solutions. Thus systems of diverse complexity and size were investigated, starting with simple (and molecular) salt solutions with small fragments that mimic the various functional groups of amino acids such as N-methylacetamide representing the peptide bond or alkylated ammonium cations. Continuing with individual positively charged amino acids (arginine, histidine, lysine) a strong binding interaction with small fluoride anion that is significantly weakened for larger halides (Cl− , Br− , I− ) was described. This observation was extended by detecting the strong sensitivity of fluoride to charge distribution on ammonium, lysine side chain, and the N-terminal of glycine while sensitivity of iodide was found to be low. Later it was shown that the attractive side chain-side chain interactions are significant for short positively charged peptide fragments in polyarginine and dihistidine, while they are not present at all in case of polylysine. Considering the qualitative difference in the origin of ion-specific interactions, electrophoretic mobility measurements (for mono- and tetra- amino acids) were employed in tandem with MD simulations. The ion-specific arginine-sulphate and arginine-guanidinium interactions were proved, both pronounced as the specific decrease or increase in electrophoretic mobility in contrast to observations for lysine, chloride anion, and sodium cation. Cation-specific interaction was found, both experimentally and computationally, to be responsible for specific affecting of the enzymatic activity of HIV-1 protease and LinB enzyme from dehalogenase family. In both cases the general salting out effect was experimentally observed (pronounced as the increase of the enzymatic activity). Finally, the denaturant-specific unfolding pathway of TrpCage minipeptide was identified by comparing the denaturation process in urea and guanidinium chloride solution. In all the studies the aim was to shed more light on complex behaviour caused by ion-specific effects such as ordering in Hofmeister series, speeding up the enzymatic activity, preferential interactions with functional groups or the osmolyte specific denaturation pathways. Keywords: molecular dynamics, proteins, denaturation, salts, osmolytes.

N´azev pr´ace: Interakce iont˚ u s proteiny Autor: Mgr. et Mgr. Jan Heyda Katedra: Fyzik´aln´ı a makromolekul´arn´ı chemie ´ ˇ v.v.i. Vedouc´ı doktorsk´e pr´ace: Prof. Pavel Jungwirth, DSc., UOCHB AV CR, E-mail vedouc´ıho: [email protected] Abstrakt: V pˇredkl´adan´e pr´aci byly pouˇzity metody molekulov´e dynamiky v kombinaci s pokroˇcil´ ymi technikami anal´ yzy k z´ısk´an´ı detailn´ıch informac´ı a pro hlubˇs´ı pochopen´ı interakc´ı mezi ionty a proteiny v roztoc´ıch. Proto byly zkoum´any syst´emy o r˚ uzn´em stupni komplexity, poˇc´ınaje roztoky molekul´arn´ıch sol´ı s drobn´ ymi fragmenty, podobaj´ıc´ımi se funkˇcn´ım skupin´am aminokyselin, jako napˇr. N-methylacetamid reprezentuj´ıc´ı peptidovou vazbu nebo alkylovan´e amonn´e kationty. D´ale se pˇredmˇetem naˇseho studia staly jednotliv´e kladnˇe nabit´e aminokyseliny, u nichˇz byla pops´ana siln´a interakce s mal´ ym fluoridov´ ym aniontem, jeˇz je − − − vˇsak pro vˇetˇs´ı halogenidy (Cl , Br , I ) v´ yraznˇe zeslabena. Toto pozorov´an´ı bylo prohloubeno objevem vysok´e citlivosti fluoridu na rozloˇzen´ı n´aboje na amonn´e skupinˇe, boˇcn´ım ˇretˇezci lysinu a N-konci glycinu, zat´ımco jodid zde vykazoval pouze velice n´ızkou citlivost. N´aslednˇe bylo prok´az´ano, ˇze u kr´atk´ ych kladnˇe nabit´ ych peptidov´ ych u ´sek˚ u v polyargininu a dihistidinu jsou preferovan´e pˇritaˇzliv´e interakce mezi boˇcn´ımi ˇretˇezci, naopak v pˇr´ıpadˇe polylysinu pˇr´ıtomny nejsou. Na z´akladˇe existence kvalitativn´ıho rozd´ılu v p˚ uvodu iontovˇe specifick´ ych interakc´ı bylo spoleˇcnˇe s MD simulacemi provedeno mˇeˇren´ı elektroforetick´ ych pohyblivost´ı (pro mono- a tetra aminokyseliny). T´ımto zp˚ usobem byly odhaleny iontovˇe specifick´e interakce mezi arigininem a sulf´atem, a mezi argininem a guanidn´ ym kationtem – oba efekty se projevily jako charakteristick´e zv´ yˇsen´ı ˇci sn´ıˇzen´ı elektroforetick´e pohyblivosti ve srovn´an´ı s lysinem, chloridov´ ym aniontem a sod´ıkov´ ym kationtem. D´ale bylo experiment´alnˇe i v´ ypoˇcetnˇe zjiˇstˇeno, ˇze enzymatick´a aktivita HIV1 prote´azy a enzymu LinB z rodiny dehalogen´az souvis´ı s kationtovˇe specifickou interakc´ı, kter´a vysvˇetluje nˇekter´e zmˇeny v aktivitˇe. V obou pˇr´ıpadech byl experiment´alnˇe pozorov´an i efekt vysolov´an´ı (v podobˇe zv´ yˇsen´ı enzymatick´e aktivity). V neposledn´ı ˇradˇe byly srovn´any denaturaˇcn´ı procesy prob´ıhaj´ıc´ı v roztoc´ıch moˇcoviny a chloridu guanidn´eho a identifikov´any dva r˚ uzn´e zp˚ usoby rozbalen´ı minipeptidu TrpCage. Ve vˇsech zm´ınˇen´ ych studi´ıch bylo snahou v´ıce osvˇetlit komplexn´ı chov´an´ı v´ yˇse uveden´ ych syst´em˚ u, a to zejm´ena objasnit iontovˇe charakteristick´e jevy, jako napˇr. poˇrad´ı v Hofmeisterovˇe ˇradˇe iont˚ u, urychlen´ı enzymatick´e aktivity, favorizovan´e interakce nebo pr˚ ubˇeh denaturace z´avisl´ y na dan´em osmolytu. Kl´ıˇ cov´ a slova: molekulov´a dynamika, proteiny, denaturace, soli, osmolyty.

1 Abstract in English

1 1.1

Abstract in English Introduction into complex systems

Biomolecules are always in contact with solvents and solvents are rarely ionfree. Proteins in salt solutions are examples of complex systems which are arguably the most popular research targets in contemporary computational science [1–4]. The number and quality of publications rise in hand with increasing and persuading evidence of which effects are robust and which are not. In simple systems – such as binary or ternary solutions of electrolytes/non-electrolytes – it is becoming increasingly difficult to provide so needed novelty for high-quality publications. However, recent achievements appeared even in the field of neat water. The water is ‘alive’ and challenging as documented by examples, such as Laage [5], who studied water reorientation and hydrogen-bond cleavage, Sedlmeier [6], who critically compared water models with structure factors from neutron scattering experiment, or Dzubiella [7], who introduced a new generation of implicit solvents. In any case once the complex species, either polymer, peptide or even a pair of large spheres, is introduced into the simple solution, the number of study opportunities quickly increases [2,8–10]. Not only the general presence and evidence of an effect, but rather its absolute strength is important. It is typically the detailed balance between competing interactions (i.e. electrostatics vs. solvation) that makes this research relevant for biological systems. Proteins are very individual objects, therefore there are not many generally valid rules. However, the building blocks, twenty amino acids connected by the peptide bonds, are always present. In the crudest approximation we can assign all the relevancy to the functional groups, shown in Figure 1, and neglect everything else. This leads to a study of charged groups (carboxyl, ammonium, imidazolium, guanidinium), polar groups (alcohols, thiols, carboxamide), and also the peptide bond. It is not self evident if this is the correct approach. A breakthrough appeared in 2007, when Auton, Holthauzen and Bolen [11] redid the analysis of the old experimental data from the 1970s. Their study brought a completely new insight into interpretation of protein denaturation and reestab-

Figure 1. The functional groups that are supposed to be crucial for local interaction of salts and solvent with proteins and peptides. A charged carboxyl group of aspartate, glutamate and C-terminal (A), ammonium moiety of lysine or Nterminal (B), guanidinium moiety of arginine (C), imidazolium moiety of histidine (D), alcohols and thiols (E), carboxamide group of asparagine and glutamine (E), and the peptide bond (H).

5

1.2

Aims of the study

lished the validity of the so called transfer model [12], which is essentially nothing but the reductionist approach of cutting into pieces. This brought a lot of enthusiasm between theoreticians, as the simplified models are not stand-alone projects any more.

1.2

Aims of the study

We would like to deal with the systems in their full complexity, but there is usually no other way to proceed further than to use approximations. Therefore the reductionist approach is often used with advantage and a system is divided into pieces using the best accessible knowledge, or by a methodology that has shown to be recently the most successful and promising. The usual progress towards the understanding of ion-protein interactions would start with ion-ion, ion-amino acid proxy, ion-amino acid, and ion-peptide complexes so as to naturally reach the ion-protein system. Even though the first two steps capture most of the essence, the practical aspects (keeping the bio-track) force us to work in parallel on all fields. We started with ion-protein studies [13, 14] from the very beginning, having in mind that these systems with their wide range of interactions are a challenging task. The effects of sodium and potassium cations were investigated, and following the protocol of Vrbka et al. [2], we quantified the sodium preference when compared to potassium. However, the simulation data provided us with much more. The complexity encountered was another goal of these studies. In this way we faced the upcoming obstacles and learned which effects need to be carefully treated. While the ion-protein projects took long time, putting in front of us new obstacles and challenges, both of the methodological and fundamental origin, it supplied us continuously by new phenomena and effects that were itself of an interest and many smaller projects consequently arose.

1.3 1.3.1

Computational methods Simulation protocol

I focused on extensive direct classical molecular dynamics (MD) calculations, since my projects were quite suited for this technique. Also, in some of my studies the time-evolution was an important aspect [15], or even a key factor if the effect was dynamic [14]. The system for the ‘data-harvesting’ part of the simulation can be usually prepared in the following way. The salt and water molecules are mixed in amounts so as to reach the desired concentration and so that the simple energy minimization avoids the potential problems of close contacts. We are employing the periodic boundary conditions (PBC): an infinite number of periodic images of the original system are placed in all three dimensions, giving rise to the effectively infinite system (this is the common, but vital trick). System is heated and pressurized to the target temperature and density. One can now insert (expose) any object into the solution, let the system evolve (propagate) in time, visiting relevant configurations and collect the data for the

6

1.3

Computational methods

7

purpose of the analysis. Some of the analyzing techniques are described in the following section. Note that the atoms inside the system move in time (the time step is usually 1 fs) according to Newtonian equations of motion, interacting with each other via the so called force-field (the effective empirical potential). 1.3.2

Methods of analysis

For basic analyses we used the internal module PTRAJ [16]. An important example is the so called radial distribution function (RDF), which expresses the probability of finding two bodies in a given distance. It is the central object in the statistical thermodynamics, directly related to thermodynamics. To understand the ion-specific effects in different contexts we needed methods to be descriptive enough to capture both quantitative and qualitative pictures. These more advanced tools were coded as Python scripts. Two of them, descriptive by their nature – the spatial distribution function and the proximal distribution function are briefly described below. • The spatial distribution function provides a description of environment around the central particle with three dimensional resolution. Thus for a general (non-spherical) object, it is significantly more descriptive than the conventional one dimensional radial distribution function. An example of distribution of sodium cation and chloride and iodide anions around Nmethylacetamide is given in the left side of Figure 3. • The proximal distribution function is one dimensional, but unlike RDF (implicitly assuming sphericity) it takes into account the real shape of the central molecule, therefore is suited for descriptions in complex systems. Division of N-methylacetamide in polar peptide bond and ‘hydrophobic methyl groups’ is depicted in the right side of Figure 3. Finally, for the biomolecule in a salt solution the basic question is: “Is the salt attracted, or depleated from the vicinity of the biomolecule?” These situations are schematically drawn in Figure 2. For the quantification, the so called preferrential binding parameter, Γ, is introduced (Equation 1). This single number is very powerful as it is directly related to the change of thermodynamic properties of a biomolecule in a given solution.  Γ=

vicinity Nsalt

bulk Nsalt N vicinity − bulk Nwater water

 (1)

local bulk where the Nsalt , Nsalt are the number of salt molecules in the vicinity and far local bulk from the biomolecule, and the Nwater , Nwater represent the same for water. The formula simply displays if the salt is more attracted (Γ > 0) or excluded (Γ < 0) from the vicinity of a biomolecule compared to its concentration in solution.

1.4

Results and discussion

Figure 2. Competition of water (blue spheres) and cosolvent (yellow ellipses) molecules for the vicinity of the protein (a large gray sphere). The preferential binding of cosolvent (excess of cosolvent, Γ > 0 – the left figure), and preferential hydration (depletion of cosolvent, Γ < 0 – the right figure).

1.4

Results and discussion

This thesis consists of 13 articles in which we have investigated ion-specific effects in solutions. We focus on interactions of ions with peptide and protein surfaces, nevertheless, the objects of our interest range from solutions of molecular salts (i.e., ammonium chloride, or guanidinium sulphate), salts acting on small organic fragments, or amino acids, to the effect on surfaces of peptides and proteins. All these projects have shed more light on the following issues: • Understanding of ion-specific effects, and consequent ordering in Hofmeister series [17, 18] for cations, anions, and osmolytes. • Interaction of ions with organic molecules and biomolecules. • Rationalization of the reductionistic approach and understanding of its limitations. To these points, my contribution, following the complexity of the system, can be divided into three parts: 1. Interaction of ions with model systems. Examples are small molecules that mimic the peptide bond or amino acid side chain. 2. Ion-amino acid interaction, with dominant, but not exclusive, focus on positively charged amino acids. 3. Ion-peptide and ion-protein interaction. This part shows the diversity that is a genuine property of complex systems. It seems to be convenient to illustrate here on the molecule of N-methylacetamide the concepts that were used in our work. Other systems are more difficult to introduce within the limited abstract format. The reader is referred to our publications for more details.

8

1.4

Results and discussion

Figure 3. Spatial distribution (left) of iodide (top, in violet) and chloride (bottom, in gold) anion and sodium cation (green) around N-methylacetamide complemented by partial contributions (of carbonyl (red), amide (blue), and methyl groups (black)) of the proximal distribution function (right).

1.4.1

Case study – Model of the peptide bond

In our case, the ultimate goal is to understand the interaction between ions and proteins. The composition obviously varies, but the peptide bond is a unique feature that is always present. For that reason we investigated how the sodium/potassium halides solutions affect N-methylacetamide (NMA) – i.e., the proxy of the peptide bond. The simple and fairly rigid molecule is shown in the left part of Figure 3. It is the peptide bond to some extent protected on both ends by methyl groups. This model system is worth a deeper exploration mainly due to the recent results of Bolen and Rose [19, 20]. It was shown that the peptide bond is the key player for protein stabilization and denaturation. That is why the deeper knowledge about NMA properties in salt solutions is essential. Owing to the fact that NMA contains a polar part (the peptide bond) and also hydrophobic pieces (methyl groups) we conducted also polarizable simulations [15], which treat more accurately the weak, but important, interaction of anions with hydrophobes. Simulations were analyzed not only in terms of spatial, proximal and standard radial distribution functions, but also the distribution of residence times was estimated together with mean times of interaction that were calculated based on the first order kinetic assumptions. The spatial distribution functions in Figure 3 clearly show that the polar groups are the most prominent sites for both cations (carbonyl oxygen) and anions (amide hydrogen). However, since the accessible volume is rather limited, the corresponding proximal distribution functions carry this preference only for

9

1.4

Results and discussion

cations (height of the peak is roughly 4 for sodium), while being around unity for anions. The second very important observation is the smeared, but in total strong and robust, preference of softer anions for the methyl groups (peaks ranging from ∼1.5 for chloride to ∼3 for iodide). This fact leads to an exclusion of fluoride, neutral behaviour of chloride, and to attraction of bromide and iodide to NMA. The results of other projects are briefly summarized in the next sections. All the projects are framed by the concept of ion-specific interactions. 1.4.2

Amino acid proxies

Starting with the simpler systems, we were able to interpret and propose a mechanism by which ions affect the peptide bond, finding the strong affinity of cations, but only modest of anions. This effect has direct implications for proteins [15]. To probe the Collins concept of matching water affinities [21] for ammonium cation we employed the combined MC-MD approach [22] and compared with experimental data available (through the activity coefficients of ammonium salts). We observed significantly stronger pairing of ammonium with small halides implying that the ammonium should be considered the small (so called hard) cation, in contrast to the concept of water affinities. We touched the same topic using the neutron scattering experiments linked with MD [23] calculations, where we found that the ammonium and ammonium moieties in lysine and zwitterionic glycine are not equally attractive for fluoride anion. On the contrary, iodide binds similarly in all cases. 1.4.3

Single amino acids and short oligopeptides

The positively charged amino acids (monomers) are perfect targets for halide anions. We characterized the organization of small (F− ) and large (I− ) anions in the vicinity of positivey charged sidechains (arginine, lysine, histidine) [24]. We found not only that smaller anions bind in all cases stronger, but also that the binding strength is sidechain-dependent (arginine histidine lysine). The positively charged amino acids themselves already as dipeptides exhibit a surprising feature. Similarly to a pair of sodium cations one expect the repulsion of two sidechains of the like charge. However, we revealed and quantitatively characterized the origin of like-charge attraction (parallel stacking) in case of two arginines and two histidines [25–27]. This effect does not exist for two lysines. One always asks for a more direct proof of any qualitative effect. To that point, motivated by mounting evidence of specific behaviour of guanidinium (cationcation stacking, strong pairing with sulphate, etc.), we succesfully combined the MD framework with measurements of electrophoretic mobilities of tetra-argine and tetra-lysine [28, 29]. While MD provided the initial motivation and atomistic insight, capillary electrophoresis provided hard experimental data, placing the proposed effects on solid ground. In tandem with surface sensitive experimental techniques (second harmonic generation (SHG) spectroscopy) we explored the pH dependent surface activity of the β-amyloid 1-16 fragment, charge of which varies dramatically from +6 at pH = 3, through +2 at pH = 7 to -6 at pH = 11. The results captured the dominant

10

1.5

Summary

role of total charge and sketched the surface behaviour of positively and negatively charged groups [30]. 1.4.4

Peptides and proteins

β-amyloid 1-16 fragment was a 16 amino acid long peptide. However ,it does not exhibit features of larger proteins, such as stable secondary structure elements or three dimensional fold. The 20 amino acid long TrpCage minipeptide was designed and synthetized in 2002 so as to mimic the large proteins [31]. While other research groups concentrated on the thermodynamical description of peptide unfolding [32], we focused on the dynamics of peptide denaturation. Two widely used denaturants (urea and guanidinium chloride) were investigated when acting on the TrpCage minipeptide [14]. Two different (denaturantspecific) pathways of unfolding were observed, however, with a similar result – i.e., the ensemble of denatured states. The denaturation processes were experimentally characterized by three independent techniques (circular dichroism, differential scanning calorimetry and nuclear magnetic resonance). The different denaturating effect of neighbouring cations in Hofmeister series [18] (tetrapropyl ammonium and guanidinium) was rationalized based on the pair interactions. In this way we took into acount both the ion-specific interactions in salt solution and the ion-specific interaction with the protein surface [33]. The powerful denaturing action of guanidinium can be weakened by an addition of sulphate, which is related to strong cation-anion pairing. In contrast, tetrapropyl ammonium is insensitive to anions, but being unable to create hydrogen bonds it may be a denaturant for one and stabilizer for another protein. The predictions we made in 2009 came true in 2011, when they were experimentally verified [34]. Even though Hofmeister measured the salting out effect of egg white protein [17], the Hofmeister series miraculously appear across fields. Revealing of the effect of salts on enzymatic activity is still far from being reached, nevertheless, we have tried to tackle this problem. The Hofmeister series for cations was tested in two studies employing very relevant enzymes. In the first case, the enzymatic activity assay (within MichaelisMenten kinetics) of HIV-1 protease was performed in sodium chloride and potassium chloride, observing the activity consistently higher by 20 % in the latter case. MD calculations found generally a twice higher affinity of sodium cation to the enzyme surface, and on top of that, particular spatially resolved maxima around the active site [13]. The second study, which is still under progress, targets the catalytic activity of the LinB dehalogenase, in a large set of salt solutions at various concentrations. From the computational side, we aimed at alkali chloride salts where we found to a large extent direct Hofmeister ordering [18].

1.5

Summary

To summarize, in this thesis I demonstrated the wide range of applications for MD simulations for biologically relevant systems at various level of complexity. The all-atom description allowed to obtain information about ion-specific effects

11

1.5

Summary

in all studied contexts. Due to the fact that proteins are always exposed to solvents and salty solutions this work has potential applications to many different fields, such as biophysics, biochemistry, biology or biotechnology.

12

2 Autorefer´ at v ˇ cesk´ em jazyce

2 2.1

Autorefer´ at v ˇ cesk´ em jazyce ´ Uvod do komplexn´ıch syst´ em˚ u

Biomolekuly lze jen tˇeˇzko nal´ezt mimo vodn´e prostˇred´ı a to jen m´alokdy neobsahuje ˇz´adn´e ionty. B´ılkovina ve slan´e vodˇe je pˇr´ıkladem tzv. komplexn´ıho syst´emu, kter´ y b´ yv´a asi nejˇcastˇejˇs´ım pˇredmˇetem v´ yzkumu ve v´ ypoˇcetn´ı komuˇ nitˇe [1–4]. Cetnost a kvalita publikac´ı jde ruku v ruce s pozn´an´ım, kter´e jevy se uk´azaly b´ yt v´ yznamn´e a kter´e nikoliv. St´av´a se tak ˇc´ım d´al obt´ıˇznˇejˇs´ı pˇrij´ıt s pr˚ ulomov´ ym objevem, jenˇz je jedn´ım z nezbytn´ ych pˇredpoklad˚ u pro pˇrijet´ı do vysoce impaktovan´ ych ˇcasopis˚ u v oboru roztok˚ u jednoduch´ ych sol´ı a neelektrolyt˚ u. Pˇresto se i v posledn´ı dobˇe m˚ uˇzeme setkat se z´asadn´ımi poznatky i pro tak jednoduch´ y syst´em, jak´ ym je ˇcist´a voda. Voda je tak st´ale vdˇeˇcn´e t´ema vyz´ yvaj´ıc´ı k dalˇs´ımu b´ad´an´ı. Napˇr. Laage [5] se zab´ yval t´ım, jak pˇri rotac´ıch molekul vody doch´az´ı k vzniku a z´aniku vod´ıkov´e vazby, Sedlmeier [6] srovnal v´ ysledky poˇc´ıtaˇcov´ ych simulac´ı ˇcist´e vody s experimenty neutronov´eho rozptylu, a koneˇcnˇe Dzubiella [7] uvedl nov´ y model pro v´ ypoˇcet implicitn´ı solvatace. Jakmile je sledovan´ y syst´em, aˇt jiˇz se jedn´a o polymer, peptid nebo pouze hydrofobn´ı kouli, um´ıstˇen do roztoku, moˇznosti nov´ ych podnˇet˚ u ke studiu v´ yraznˇe narostou [2, 8–10]. V posledn´ı dobˇe vˇsak jiˇz nejde pouze o to, zda efekt existuje, ale zejm´ena o s´ılu tohoto efektu. Vˇetˇsinou je to pr´avˇe detailn´ı rovnov´aha mezi soupeˇr´ıc´ımi interakcemi, kter´a je kl´ıˇcov´a v bio-syst´emech (napˇr. elektrostatick´a interakce versus hydrataˇcn´ı energie). B´ılkoviny jsou velmi rozmanit´e molekuly a z toho d˚ uvodu je pro nˇe teˇzk´e nal´ezt obecnˇe platn´a pravidla. Nicm´enˇe z´akladn´ı kameny – dvacet aminokyselin, kter´e jsou spojov´any peptidovou vazbou – jsou vˇzdy pˇr´ıtomn´e. V nejhrubˇs´ım pˇribl´ıˇzen´ı pak m˚ uˇzeme veˇsker´ y v´ yznam pˇripsat jejich funkˇcn´ım skupin´am uveden´ ym na obr´azku 4, protoˇze to je to jedin´e, ˇc´ım se liˇs´ı. Vˇse ostatn´ı zanedb´ame. To vede ke studiu nabit´ ych skupin (karboxylov´a skupina, amonn´ y kation, imidazoliov´ y kation, guanidn´ y kation), pol´arn´ıch skupin (alkoholy, thioly, karboxamidy) a tak´e peptidov´e vazby. Bohuˇzel nebylo zat´ım dostateˇcnˇe prok´az´ano, zda je takov´ y postup opodstatnˇen´ y a spr´avn´ y.

Obr´ azek 4. Funkˇcn´ı skupiny aminokyselin, kter´ ym je pˇrisuzov´an z´asadn´ı vliv na lok´aln´ı interakci s ionty a solventy. Nabit´a karboxylov´a skupina aspart´atu, glut´amatu a C-konce (A), amonn´a skupina lysinu, resp. N-konce (B), guanidn´a skupina z postrann´ıho ˇretˇezce argininu (C), imidazoliov´a skupina histidinu (D), alkoholy a thioly (E), karboxamidov´a skupina asparaginu a glutaminu (E), a peptidov´a vazba (H).

13

2.2

C´ıle pr´ ace

Z´asadn´ı zvrat nastal v roce 2007, kdy Auton, Holthauzen a Bolen [11] kompletn´ım zp˚ usobem zrevidovali experiment´aln´ı a teoretick´a data ze 70. let. Jejich pr´ace vnesla zcela nov´e svˇetlo do popisu denaturace protein˚ u, ˇc´ımˇz znovu objevili a prok´azali platnost tzv. transferov´eho modelu [12], kter´ y nen´ı nic jin´eho neˇz proces rozdˇelen´ı proteinu na jednotliv´e ˇc´asti.

2.2

C´ıle pr´ ace

Bylo by velice uˇziteˇcn´e, kdyby bylo re´aln´e studovat syst´emy a jejich vz´ajemn´e interakce v jejich pˇrirozen´em prostˇred´ı, ale d´ıky jejich sloˇzitosti je to zat´ım mimo naˇse moˇznosti a mus´ıme se proto uch´ ylit k aproximac´ım. Nejbˇeˇznˇejˇs´ı je tzv. metoda redukce, kdy se p˚ uvodn´ı syst´em rozsek´a na kousky a ty se studuj´ı oddˇelenˇe v r´amci nejlepˇs´ı (a nejslibnˇejˇs´ı) dostupn´e teorie. Z tohoto pohledu se jako nejv´ yhodnˇejˇs´ı jev´ı zaˇc´ıt simulacemi interakc´ı iont˚ u mezi sebou, d´ale iont˚ u a mal´ ych organick´ ych molekul, a pot´e iont˚ u a aminokyselin, ˇc´ımˇz se koneˇcnˇe dostaneme ke k´ yˇzen´ ym protein˚ um. I pˇresto, ˇze d˚ uleˇzit´e jsou zejm´ena prvn´ı dva kroky, je z praktick´ ych d˚ uvod˚ u potˇreba postupovat souˇcasnˇe na r˚ uzn´ ych u ´rovn´ıch. V naˇsem pˇr´ıpadˇe jsme nav´azali na pr´aci Vrbka et al. [2] a zab´ yvali se vlivem chloridu sodn´eho a draseln´eho na HIV-1 prote´azu. Z v´ ypoˇcetn´ıho hlediska v´ ysledky potvrdily preferenci sodn´ ych kationt˚ u nad draseln´ ymi na povrchu HIV-1 prote´azy [13]. Podstatn´e bylo, ˇze jsme se setk´avali s d´ılˇc´ımi jevy, kter´e je tˇreba pochopit, abychom mohli plnˇe porozumˇet problematice protein˚ u v cel´e jej´ı ˇs´ıˇri. Projekty zkoumaj´ıc´ı proteiny byly dlouhodob´eho charakteru, neboˇt pˇred n´as opakovanˇe kladly nov´e pˇrek´aˇzky a v´ yzvy (aˇt fundament´aln´ı ˇci technick´e) [13, 14]. T´ım vˇsak s sebou pˇrin´aˇsely menˇs´ı projekty, kter´e byly zaj´ımav´e per-se.

2.3 2.3.1

V´ ypoˇ cetn´ı metody Simulaˇ cn´ı protokol

Klasick´a molekulov´a dynamika se uk´azala jako ide´aln´ı volba pro studium ˇ naˇsich syst´em˚ u. Casto bylo podstatn´e pouze z´ıskat sadu struktur, avˇsak v nˇekter´ ych pˇr´ıpadech byl d˚ uleˇzit´ y ˇcasov´ y v´ yvoj [14, 15]. K tom doch´azelo tehdy, kdyˇz byla studov´ana pˇr´ımo dynamika dan´eho dˇeje. Na zaˇc´atku kaˇzd´eho projektu stoj´ı pˇr´ıprava syst´emu. Pokus´ım se nast´ınit d˚ uleˇzit´e kroky. Nejprve se molekuly vody a soli n´ahodnˇe sm´ıs´ı v pomˇeru, kter´ y odpov´ıd´a c´ılov´e koncentraci. N´aslednˇe provedeme kr´atkou minimalizaci energie syst´emu, jeˇz n´as zbav´ı neˇz´adouc´ıch artefakt˚ u (napˇr. n´ahodn´eho pˇrekryvu dvou atom˚ u apod.). Prakticky vˇzdy pouˇz´ıv´ame v simulaci periodick´e okrajov´e podm´ınky, tj. rozmnoˇz´ıme n´aˇs syst´em ve vˇsech rozmˇerech, ˇc´ımˇz dostaneme nekoneˇcn´ y syst´em. Jde o d˚ uleˇzit´ y a bˇeˇznˇe pouˇz´ıvan´ y trik, jak ze syst´emu o cca 1000 atomech udˇelat obrovsk´ y syst´em, kde jiˇz beze zbytku plat´ı z´akony statistiky. Na z´avˇer syst´em zahˇrejeme a stlaˇc´ıme, aby mˇel poˇzadovanou hustotu a teplotu. Pokud chceme studovat chov´an´ı molekuly vystaven´e u ´ˇcink˚ um slan´eho roztoku, vloˇz´ıme ji do pˇripraven´eho syst´emu a nech´ame ho vyv´ıjet v ˇcase, ˇc´ımˇz prozkoum´ame zejm´ena relevantn´ı (ale do jist´e m´ıry i jin´e m´alo pravdˇepodobn´e)

14

2.4

V´ ysledky a diskuse

15

stavy syst´emu. Ty vytvoˇr´ı statistick´ y soubor, kter´ y m˚ uˇzeme n´aslednˇe analyzovat, napˇr´ıklad metodami popsan´ ymi v n´asleduj´ıc´ı kapitole. Je tˇreba jeˇstˇe podotknout, ˇze ˇcasov´ y krok pouˇz´ıvan´ y v simulac´ıch je velmi kr´atk´ y – typicky 1 fs – a veˇsker´ y pohyb je ˇr´ızen Newtonov´ ymi rovnicemi. Jednotliv´e ˇc´astice na sebe p˚ usob´ı tzv. silov´ ym polem, coˇz je pˇredem stanoven´ y efektivn´ı potenci´al (sada interakˇcn´ıch parametr˚ u). 2.3.2

Druhy anal´ yzy

Pro rutinn´ı anal´ yzy jsme pouˇz´ıvali volnˇe dostupn´ y modul PTRAJ [16]. D˚ uleˇzit´ ym pˇr´ıkladem je tzv. radi´aln´ı distribuˇcn´ı funkce (RDF), kter´a vyjadˇruje pravdˇepodobnost nalezen´ı dvou objekt˚ u (napˇr´ıklad kationtu a aniontu) v dan´e vzd´alenosti. Jej´ı z´asadn´ı v´ yznam pramen´ı ze skuteˇcnosti, ˇze se jedn´a o u ´stˇredn´ı veliˇcinu ve statistick´e termodynamice, pˇr´ımo souvisej´ıc´ı s termodynamikou syst´emu. Abychom porozumˇeli iontovˇe specifick´ ym interakc´ım, potˇrebujeme m´ıt k dispozici typy anal´ yzy, kter´e jsou dostateˇcnˇe popisn´e a z´aroveˇ n i kvantitativn´ı. Tyto pokroˇcilejˇs´ı, m´enˇe obvykl´e metody byly implementov´any v programovac´ım jazyku Python. Dvˇe z nich, mapa pravdˇepodobnosti a tzv. proxim´aln´ı distribuˇcn´ı funkce, jsou pops´any n´ıˇze. • Mapa pravdˇepodobnosti poskytuje prostorov´e rozloˇzen´ı jednoho typu ˇc´astic kolem druh´eho um´ıstˇen´eho ve stˇredu. Pro nesf´erick´e molekuly tak nese v´ yraznˇe v´ıce informace neˇz klasicky pouˇz´ıvan´a RDF. Pˇr´ıklad jej´ıho pouˇzit´ı je v lev´e ˇc´asti obr´azku 6. • Proxim´aln´ı distribuˇcn´ı funkce je v´ yznamovˇe podobn´a RDF, ale na rozd´ıl od n´ı bere do u ´vahy i tvar molekuly (proxim´aln´ı, tj. nejbl´ıˇze um´ıstˇen´ y), kolem kter´e je distribuˇcn´ı funkce poˇc´ıt´ana. Pˇr´ıklad jej´ıho pouˇzit´ı je zn´azornˇen v prav´e ˇc´asti obr´azku 6. ´ redn´ı ot´azkou, kterou si klademe pˇri studiu biomolekul v roztoc´ıch je, zda Ustˇ je v jej´ı bl´ızkosti sloˇzen´ı roztoku odliˇsn´e ˇcist´eho roztoku. Na obr´azku 5 jsou pro ilustraci zobrazeny tyto dvˇe krajn´ı moˇznosti. Pro kvantifikaci tohoto efektu je zaveden tzv. preferenˇcn´ı vazebn´ y parametr Γ (rovnice 2),  Γ=

vicinity Nsalt

bulk Nsalt N vicinity − bulk Nwater water

 (2)

local bulk kde Nsall , Nsalt vyjadˇruje poˇcet molekul soli (kosolventu) v bl´ızkosti a daleko local bulk od biomolekuly, a Nwater , Nwater m´a stejn´ y v´ yznam, avˇsak pro vodu (obecnˇeji solvent). Hodnota parametru m´a pˇr´ımoˇcar´ y v´ yznam; zda se s˚ ul vyskytuje v´ıce (Γ > 0), nebo m´enˇe (Γ < 0) v bl´ızkosti biomolekuly, neˇz odpov´ıd´a situaci daleko v roztoku.

2.4

V´ ysledky a diskuse

Tato doktorsk´a pr´ace sest´av´a z 13 ˇcl´ank˚ u v impaktovan´ ych ˇcasopisech, v nichˇz jsme studovali iontovˇe specifick´e efekty v nejr˚ uznˇejˇs´ıch prostˇred´ıch. Zaˇcali jsme

2.4

V´ ysledky a diskuse

Obr´ azek 5. Kompetice molekul solventu (napˇr. vody; ˇcerven´e krouˇzky) a kosolventu (napˇr. soli; ˇzlut´e elipsy) o okol´ı proteinu (velk´ y ˇsed´ y kruh). Nadbytek kosolventu (Γ > 0 – vlevo) a nadbytek molekul solventu (vylouˇcen´ı kosolventu, Γ < 0 – vpravo).

roztoky molekul´arn´ıch sol´ı (jako napˇr´ıklad chloridem amonn´ ym nebo guanidn´ ym), pokraˇcovali simulacemi aminokyselin vystaven´ ych slan´ ym roztok˚ um, abychom se nakonec zamˇeˇrili na povrchy peptid˚ u a protein˚ u. Vˇsechny projekty mˇely pˇrispˇet k hlubˇs´ımu porozumˇen´ı n´asleduj´ıc´ıch bod˚ u: • Iontovˇe specifick´e efekty a jejich projev v Hofmeisterov´ ych ˇrad´ach [17, 18] kationt˚ u, aniont˚ u a osmolyt˚ u. • Interakce iont˚ u a organick´ ych molekul a biomolekul. • Racionalizace metody redukce a stanoven´ı hranic jej´ı platnosti. M˚ uj pˇr´ıspˇevek lze rozdˇelit do n´asleduj´ıc´ıch tˇr´ı kapitol podle jejich rostouc´ı sloˇzitosti: 1. Interakce iont˚ u v modelov´ ych syst´emech. Pˇr´ıkladem budiˇz molekula napodobuj´ıc´ı peptidovou vazbu nebo postrann´ı ˇretezec aminokyseliny. 2. Interakce iont˚ u s aminokyselinami, zejm´ena pak kladnˇe nabit´ ymi. 3. Interakce iont˚ u a b´ılkovin. Tato ˇc´ast ukazuje velkou rozmanitost, kter´a je vlastn´ı pr´avˇe komplexn´ım syst´em˚ um. Na molekule N-metylacetamidu je vhodn´e uk´azat, jak´e postupy jsme obvykle pouˇz´ıvali pˇri naˇs´ı pr´aci. Ostatn´ı projekty jsou svou povahou komplikovanˇejˇs´ı na struˇcn´e shrnut´ı, proto odkazuji laskav´eho ˇcten´aˇre na naˇse publikace, kde m˚ uˇze nal´ezt vˇsechny detaily. 2.4.1

Pˇ r´ıklad – studie modelu peptidov´ e vazby

Naˇs´ım c´ılem je porozumˇet interakc´ı mezi ionty a proteiny. Zamysl´ıme-li se nad t´ım, co je to protein, zjist´ıme, ˇze zat´ımco sloˇzen´ım (tj. poˇcty r˚ uzn´ ych aminokyselin v sekvenci) se jednotliv´e proteiny mezi sebou liˇs´ı, peptidov´a vazba je unik´atn´ı prvek, kter´ y je vˇzdy pˇr´ıtomn´ y (pr´avˇe jednou za kaˇzdou aminokyselinu). Proto jsme studovali, jak sodn´e a draseln´e soli halogenid˚ u budou ovlivˇ novat N-metylacetamid (NMA), molekulu maj´ıc´ı charakter peptidov´e vazby. Tato pomˇernˇe

16

2.4

V´ ysledky a diskuse

Obr´ azek 6. Mapa pravdˇepodobnosti (vlevo) jodidov´ ych (nahoˇre, ve fialov´e barvˇe) a chloridov´ ych aniont˚ u (dole, zlatou barvou) a sodn´ ych kationt˚ u (zelenou barvou) kolem molekuly N-metylacetamidu vyj´adˇrena komplement´arnˇe tˇremi pˇr´ıspˇevky proxim´aln´ı distribuˇcn´ı funkce (vpravo; karbonylov´e skupiny (v ˇcerven´e barvˇe), amidov´e skupiny (v modr´e barvˇe) a metylov´e skupiny (v ˇcern´e barvˇe)).

rigidn´ı molekula s jednoduchou strukturou, zobrazena v lev´e ˇc´asti obr´azku 6, je v podstatˇe peptidovou vazbou metylovanou na obou stran´ach. Syst´em si zaslouˇz´ı bliˇzˇs´ı pozornost tak´e proto, ˇze ned´avn´e v´ ysledky Bolena a Rose [19, 20] silnˇe naznaˇcuj´ı, ˇze peptidov´a vazba je kl´ıˇcov´ ym elementem pro stabilitu a denaturaci. I proto je pozn´an´ı vlastnost´ı NMA v roztoc´ıch sol´ı tak d˚ uleˇzit´a. Jelikoˇz NMA nese kromˇe pol´arn´ı peptidov´e vazby i hydrof´obn´ı metylov´e skupiny, bylo tˇreba do naˇsich v´ ypoˇct˚ u zahrnout polarizovatelnost vody a aniont˚ u. Jenom tak bylo moˇzn´e spr´avnˇe vyhodnotit s´ılu interakce aniont˚ u s hydrof´obn´ımi skupinami. Data z´ıskan´a ze simulac´ı byla zpracov´ana pomoc´ı map pravdˇepodobnosti, proxim´aln´ıch distribuˇcn´ıch funkc´ı a klasick´ ych RDF (tj. anal´ yza, kde ˇcas nen´ı d˚ uleˇzit´ y). D´ale jsme charakterizovali d´elku tˇechto kontakt˚ u za pˇredpokladu, ˇze ionty se k peptidov´e vazbˇe v´aˇzou kinetikou prvn´ıho ˇr´adu (tj. explicitn´ı ˇcas je zde d˚ uleˇzit´ y). Mapy pravdˇepodobnosti na obr´azku 6 zcela jasnˇe ukazuj´ı, ˇze nejpravdˇepodobnˇejˇs´ı pozice v okol´ı molekuly NMA jsou pro kationty v bl´ızkosti karbonylov´eho kysl´ıku (kter´ y nese parci´aln´ı z´aporn´ y n´aboj) a pro anionty u vod´ıku (parci´aln´ı kladn´ y n´aboj) amidov´e skupiny. Druhou nem´enˇe d˚ uleˇzitou skuteˇcnost´ı je, ˇze prostor, kter´ y pˇr´ısluˇs´ı tˇemto energeticky v´ yhodn´ ym lokac´ım, je prostorovˇe vymezen´ y, zejm´ena pak pro anionty. Proto pˇri pˇrechodu k proxim´aln´ı distribuˇcn´ı funkci vid´ıme (vpravo), ˇze relativnˇe vysok´ y p´ık m´a pouze sodn´ y kation (cca o v´ yˇsce 4), nikoliv tak anionty (kolem 1). Tˇret´ım faktem je, ˇze anionty (na rozd´ıl od kationt˚ u)

17

2.4

V´ ysledky a diskuse

vykazuj´ı afinitu k metylov´ ym skupin´am, kde kˇrivka dosahuje hodnoty cca 1.5 pro chloridov´ y a v´ıce neˇz 3 pro jodidov´ y aniont. V´ ysledky ostatn´ıch projekt˚ u jsou shrnuty v dalˇs´ı kapitole. M˚ uˇzeme v nich naj´ıt jednoho spoleˇcn´eho jmenovatele – iontovˇe specifick´e interakce. 2.4.2

Modely funkˇ cn´ıch skupin

V pˇredchoz´ı kapitole jsme zaˇcali jednoduchou molekulou, na n´ıˇz jsme dok´azali pˇredpovˇedˇet a interpretovat s´ılu a mechanismus interakce, kter´ ym budou ionty p˚ usobit na peptidovou vazbu. Tato pozorov´an´ı maj´ı pˇr´ım´e d˚ usledky pro vlastnosti b´ılkovin ve slan´ ych roztoc´ıch [15]. Abychom ovˇeˇrili koncept ’matching water affinities’ pˇredpovˇezen´ y Collinsem [21], detailnˇe jsme kombinac´ı molekulov´e dynamiky a metody Monte Carlo prozkoumali vazebn´e preference amonn´eho kationtu, kdyˇz se nach´az´ı v roztoku s halogenidov´ ymi anionty [22]. V´ ysledky jasnˇe uk´azaly, ˇze amonn´ y kation preferuje mal´e anionty pˇred vˇetˇs´ımi, a je tud´ıˇz tzv. mal´ ym (nebo t´eˇz tvrd´ ym) kationtem. Takov´ y v´ ysledek je v rozporu s konceptem ’matching water affinities,’ ale je plnˇe v souladu s namˇeˇren´ ymi aktivitn´ımi koeficienty. Toto t´ema jsme studovali tak´e v kombinaci neutronov´eho rozptylu a MD simulac´ı [23]. V´ ysledky pregnantnˇe ukazuj´ı, ˇze amonn´ y kation a amonn´e skupiny lysinu a glycinu nejsou k fluoridov´emu aniontu pˇritahov´any se stejnou intenzitou. Naopak jodid se v´aˇze podobnˇe ve vˇsech pˇr´ıpadech. 2.4.3

Jednotliv´ e aminokyseliny a kr´ atk´ e oligopeptidy

Kladnˇe nabit´e aminokyseliny (v podobˇe monomer˚ u) pˇredstavuj´ı pro halogenidov´e anionty ide´aln´ı terˇc [24]. Charakterizovali jsme uspoˇr´ad´an´ı mal´ ych fluoridov´ ych a velk´ ych jodidov´ ych aniont˚ u v okol´ı kladnˇe nabit´ ych boˇcn´ıch ˇretˇezc˚ u (argininu, lysinu, histidinu). Menˇs´ı anionty se v´aˇzou silnˇeji ve vˇsech pˇr´ıpadech a nav´ıc se uk´azalo, ˇze vazebn´a s´ıla je z´avisl´a na typu boˇcn´ıho ˇretˇezce. Kladnˇe nabit´e aminokyseliny samy o sobˇe vykazuj´ı zaj´ımavou vlastnost. Stejnˇe jako je samozˇrejm´e, ˇze se odpuzuj´ı dva sodn´e kationty, je stejnˇe tak pˇrekvapiv´e, ˇze dva shodnˇe nabit´e postrann´ı ˇretˇezce argininu, nebo histidinu se pˇritahuj´ı a tvoˇr´ı paraleln´ı ’stack’. [25–27] Naproti tomu postrann´ı ˇretˇezce lysinu se chovaj´ı ’ˇr´adnˇe’ a odpuzuj´ı se. Pro kaˇzd´ y kvalitativn´ı jev, kter´ y je sv´ ym zp˚ usobem neobvykl´ y, je vhodn´e naj´ıt experiment, jenˇz jej dok´aˇze i kvantifikovat. V naˇsem pˇr´ıpadˇe se uk´azalo efektivn´ı pouˇzit´ı mˇeˇren´ı elektroforetick´e mobility pro peptidy tetraarginin a tetralysin [28, 29] v pˇr´ıtomnosti sol´ı obsahuj´ıc´ı jedno- a dvoumocn´e anionty. Zat´ımco MD pˇrispˇela k motivaci v´ yzkumu a poskytla vhled na atom´arn´ı u ´rovni, kapil´arn´ı elektrofor´eza namˇeˇrila potˇrebn´a experiment´aln´ı data a pˇremˇenila efekty z hypotetick´ ych na re´aln´e. Spolu s povrchovˇe citlivou spektroskopickou metodou (generov´an´ı druh´e harmonick´e frekvence (SHG)) jsme prozkoumali z´avislost povrchov´e aktivity β-amyloid 1-16 fragmentu na pH. β-amyloid 1-16 fragment m´a mnoho titrovateln´ ych skupin, a proto se jeho n´aboj v kysel´em (+6), bazick´em (-6) a neutr´aln´ım (+2) pH velmi liˇs´ı. Molekulov´a dynamika spr´avnˇe zachytila dominantn´ı roli celkov´eho

18

2.5

Z´ avˇ er

n´aboje a nast´ınila moˇzn´e d˚ uvody pro pozorovan´e rozd´ıly v kysel´em a bazick´em pH [30]. 2.4.4

Peptidy a proteiny

β-amyloid 1-16 fragment m´a 16 aminokyselin, pˇresto nevykazuje chov´an´ı vˇetˇs´ıch protein˚ u, napˇr. nenajdeme u nˇej stabiln´ı sekund´arn´ı strukturn´ı prvky nebo balen´ı do dominantn´ı terci´aln´ı struktury. Oproti tomu v roce 2002 umˇele navrˇzen´ y 20 aminokyselinov´ y TrpCage minipetid vˇsechny tyto rysy nese [31]. Zat´ımco ostatn´ı vˇedeck´e t´ ymy se zamˇeˇrily na popis termodynamick´ ych vlastnost´ı jeho sbalen´ı a rozbalen´ı [32], my jsme si dali za c´ıl charakterizovat proces denaturace. ´ cinek dvou nejbˇeˇznˇejˇs´ıch denaturant˚ Uˇ u (moˇcoviny a chloridu guanidn´eho) byl zkoum´an na minipeptidu TrpCage [14]. Pˇrekvapivˇe jsme zjistili, ˇze pr˚ ubˇeh denaturace je z´avisl´ y na denaturantu, nikoliv vˇsak v´ ysledn´ y efekt, kter´ y je stejn´ y– TrpCage minipeptid v rozbalen´em, denaturovan´em stavu. Denaturace minipeptidu byla nez´avisle promˇeˇrena tˇremi separ´atn´ımi technikami (cirkul´arn´ım dichroismem, kalorimetricky a pomoc´ı nukle´arn´ı magnetick´e rezonance). V n´asleduj´ıc´ı studii jsme se zamˇeˇrili na sousedy v Hofmeisterovˇe ˇradˇe kationt˚ u [18] – tetrapropylamonium a guanidinium. K vysvˇetlen´ı efekt˚ u jsme pouˇzili vˇsechny dosud zn´am´e efektivn´ı p´arov´e interakce, kter´e vykazuj´ı specificitu jak mezi samotn´ ymi ionty v roztoku, tak i mezi ionty a povrchem proteinu [33]. Jedinˇe v d˚ usledku siln´eho kation-aniontov´eho p´arov´an´ı m˚ uˇze b´ yt vysvˇetleno, proˇc se z velmi siln´eho denaturantu, kter´ ym je chlorid guanidn´ y, stane pr˚ umˇern´ y denaturant s´ıranu guanidn´eho. Naopak tetrapropyl amonium nen´ı v˚ ubec citliv´e na druh aniontu, s n´ımˇz se nach´az´ı v roztoku, z´aroveˇ n ale vzhledem k absenci moˇznosti tvorby vod´ıkov´ ych vazeb m˚ uˇze b´ yt pro jeden typ protein˚ u velmi siln´ ym denaturantem a pro druh´e slabˇe stabilizuj´ıc´ım osmolytem. Koneˇcnˇe naˇse pˇredpovˇedi z roku 2009 se vyplnily po proveden´ı experiment˚ u v roce 2011 [34]. Pˇrestoˇze Hofmeister v roce 1888 seˇradil soli podle jejich schopnosti vysolovat vajeˇcn´ y b´ılek [17] z roztoku, objevuj´ı se Hofmeisterovy ˇrady jako z´azrakem napˇr´ıˇc r˚ uzn´ ymi obory. Porozumˇet tomu, proˇc nˇekter´e soli zvyˇsuj´ı enzymatickou aktivitu, je bˇeh na dlouhou traˇt, ale i tak se mus´ı podstoupit. Hofmeisterova ˇrada pro kationty byla testov´ana ve dvou studi´ıch zab´ yvaj´ıc´ıch se vybran´ ymi enzymy. V prvn´ı z nich byl proveden rozbor enzymatick´e aktivity (s pouˇzit´ım kinetiky Michaelis-Mentenov´e) HIV-1 prote´azy v chloridu sodn´em a draseln´em, kter´ y v pˇr´ıpadˇe KCl roztoku vedl k objevu o 20% vyˇsˇs´ı aktivity neˇz u NaCl. MD simulace poskytovaly obecnˇe dvojn´asobnou afinitu sodn´eho kationtu k povrchu enzymu a nadto ukazovaly na prostorovˇe rozliˇsen´a maxima v okol´ı aktivn´ıho m´ısta. Druh´a studie, jeˇz je zat´ım ve f´azi pˇr´ıprav, m´a za c´ıl pozn´an´ı charakteru katalytick´e aktivity LinB dehalogen´azy pro sadu roztok˚ u sol´ı o r˚ uzn´ ych koncentrac´ıch. Z v´ ypoˇcetn´ıho hlediska jsme se zamˇeˇrili na chloridy alkalick´ ych kov˚ u, u kter´ ych jsme do znaˇcn´e m´ıry pozorovali Hofmeisterovo uspoˇr´ad´an´ı [18].

2.5

Z´ avˇ er

Z´avˇerem bych uvedl, ˇze v pˇredkl´adan´e dizertaˇcn´ı pr´aci bylo prezentov´ano ˇsirok´e spektrum aplikac´ı molekulovˇe dynamick´ ych simulac´ı, a to zejm´ena pro

19

2.5

Z´ avˇ er

biologicky relevantn´ı syst´emy s r˚ uzn´ ym stupnˇem komplexity. Popis na u ´rovni jednotliv´ ych atom˚ u a molekul n´am umoˇznil z´ıskat poznatky o iontovˇe specifick´ ych efektech ve vˇsech studovan´ ych kontextech. D´ıky skuteˇcnosti, ˇze proteiny jsou ve sv´em pˇrirozen´em prostˇred´ı vˇzdy vystaveny p˚ usoben´ı solvent˚ u a roztok˚ u sol´ı, m´a tato pr´ace potenci´aln´ı vyuˇzit´ı v mnoha vˇedn´ıch oborech, jako napˇr. v biofyzice, biochemii, biologii nebo biotechnologii.

20

REFERENCES

References [1] G. Hummer, S. Garde, A. E. Garcia, A. Pohorille, and L. R. Pratt, “An information theory model of hydrophobic interactions”, Proceedings of the National Academy of Sciences of the United States of America 93(17), pp. 8951–8955 (1996). [2] L. Vrbka, J. Vondr´ aˇsek, B. Jagoda-Cwiklik, R. V´acha, and P. Jungwirth, “Quantification and rationalization of the higher affinity of sodium over potassium to protein surfaces”, Proceedings of the National Academy of Sciences of the United States of America 103(42), pp. 15440–15444 (2006). [3] G. D. Rose, P. J. Fleming, J. R. Banavar, and A. Maritan, “A backbone-based theory of protein folding”, Proceedings of the National Academy of Sciences of the United States of America 103(45), pp. 16623–16633 (2006). [4] D. Horinek, A. Serr, M. Geisler, T. Pirzer, U. Slotta, S. Q. Lud, J. A. Garrido, T. Scheibel, T. Hugel, and R. R. Netz, “Peptide adsorption on a hydrophobic surface results from an interplay of solvation, surface, and intrapeptide forces”, Proceedings of the National Academy of Sciences of the United States of America 105(8), pp. 2842–2847 (2008). [5] D. Laage and J. T. Hynes, “A molecular jump mechanism of water reorientation”, Science 311(5762), pp. 832–835 (2006). [6] F. Sedlmeier, D. Horinek, and R. R. Netz, “Spatial correlations of density and structural fluctuations in liquid water: A comparative simulation study”, Journal of the American Chemical Society 133(5), pp. 1391–1398 (2011). [7] J. Dzubiella, J. M. J. Swanson, and J. A. McCammon, “Coupling hydrophobicity, dispersion, and electrostatics in continuum solvent models”, Physical Review Letters 96(8) (2006). [8] G. Hummer, S. Garde, A. E. Garcia, M. E. Paulaitis, and L. R. Pratt, “The pressure dependence of hydrophobic interactions is consistent with the observed pressure denaturation of proteins”, Proceedings of the National Academy of Sciences of the United States of America 95(4), pp. 1552–1555 (1998). [9] D. Chandler, “Interfaces and the driving force of hydrophobic assembly”, Nature 437, pp. 640–647 (2005). [10] D. Horinek, A. Herz, L. Vrbka, F. Sedlmeier, S. I. Mamatkulov, and R. R. Netz, “Specific ion adsorption at the air/water interface: The role of hydrophobic solvation”, Chemical Physics Letters 479(4-6), pp. 173–183 (2009). [11] M. Auton, L. M. F. Holthauzen, and D. W. Bolen, “Anatomy of energetic changes accompanying urea-induced protein denaturation”, Proceedings of the National Academy of Sciences of the United States of America 104, pp. 15317–15322 (2007). [12] Y. Nozaki and C. Tanford, “Solubility of amino acids and related compounds in aqueous urea solutions”, Journal of Biological Chemistry 238(12), pp. 4074– (1963). [13] J. Heyda, J. Pokorn´ a, L. Vrbka, R. V´acha, B. Jagoda-Cwiklik, J. Konvalinka, P. Jungwirth, and J. Vondr´ aˇsek, “Ion specific effects of sodium and potassium on the catalytic activity of HIV-1 protease”, Physical Chemistry Chemical Physics 11(35), pp. 7599–7604 (2009). [14] J. Heyda, M. Koˇz´ıˇsek, L. Bedn´arov´a, G. Thompson, J. Konvalinka, J. Vondr´aˇsek, and P. Jungwirth, “Urea and guanidinium induced denaturation of a trp-cage miniprotein”, The Journal of Physical Chemistry B (2011), doi: 10.1021/jp200790h. [15] J. Heyda, J. C. Vincent, D. J. Tobias, J. Dzubiella, and P. Jungwirth, “Ion specificity at the peptide bond: Molecular dynamics simulations of n-methylacetamide in aqueous salt solutions”, Journal of Physical Chemistry B 114(2), pp. 1213–1220 (2010).

21

REFERENCES

[16] T.E. Cheatham III C.L. Simmerling J. Wang R.E. Duke R. Luo R.C. Walker W. Zhang K.M. Merz B. Roberts B. Wang S. Hayik A. Roitberg G. Seabra I. Kolossvai K.F. Wong F. Paesani J. Van´ıˇcek J. Liu X. Wu S.R. Brozell T. Steinbrecher H. Gohlke Q. Cai X. Ye J. Wang M.-J. Hsieh G. Cui D.R. Roe D.H. Mathews M.G. Seetin C. Sagui V. Babin T. Luchko S. Gusarov A. Kovalenko D.A. Case, T.A. Darden and P.A. Kollman, “Amber 11” (2010). [17] F. Hofmeister, “Zur lehre von der wirkung der salze - zweite mittheilung”, Archiv fur Experimentelle Pathologie und Pharmakologie 24(4-5), pp. 247–260 (1888). [18] W. Kunz, J. Henle, and B. W. Ninham, “’zur lehre von der wirkung der salze’ (about the science of the effect of salts): Franz hofmeister’s historical papers”, Current Opinion in Colloid & Interface Science 9(1-2), pp. 19–37 (2004). [19] Y. F. Liu and D. W. Bolen, “The peptide backbone plays a dominant role in protein stabilization by naturally-occurring osmolytes”, Biochemistry 34(39), pp. 12884–12891 (1995). [20] T. O. Street, D. W. Bolen, and G. D. Rose, “A molecular mechanism for osmolyte-induced protein stability”, Proceedings of the National Academy of Sciences of the United States of America 103(38), pp. 13997–14002 (2006). [21] Kim D. Collins, “Ions from the hofmeister series and osmolytes: effects on proteins in solution and in the crystallization process”, Methods 34(3), pp. 300–311 (2004). [22] J. Heyda, M. Lund, M. Onˇc´ ak, P. Slav´ıˇcek, and P. Jungwirth, “Reversal of hofmeister ordering for pairing of NH+ 4 vs alkylated ammonium cations with halide anions in water”, Journal of Physical Chemistry B 114(33), pp. 10843–10852 (2010). [23] P. E. Mason, J. Heyda, H. E. Fischer, and P. Jungwirth, “Specific interactions of ammonium functionalities in amino acids with aqueous fluoride and iodide”, Journal of Physical Chemistry B 114(43), pp. 13853–13860 (2010). [24] J. Heyda, T. Hrob´ arik, and P. Jungwirth, “Ion-specific interactions between halides and basic amino acids in water”, Journal of Physical Chemistry A 113(10), pp. 1969–1975 (2009). [25] J. Vondr´ aˇsek, P. E. Mason, J. Heyda, K. D. Collins, and P. Jungwirth, “The molecular origin of like-charge arginine-arginine pairing in water”, Journal of Physical Chemistry B 113(27), pp. 9041–9045 (2009). [26] J. Heyda, P. E. Mason, and P. Jungwirth, “Attractive interactions between side chains of histidine-histidine and histidine-arginine-based cationic dipeptides in water”, Journal of Physical Chemistry B 114(26), pp. 8744–8749 (2010). [27] Mario Vazdar, Jiˇr´ı Vymˇetal, Jan Heyda, Jiˇr´ı Vondr´aˇsek, and Pavel Jungwirth, “Like-charge guanidinium pairing from molecular dynamics and ab initio calculations”, The Journal of Physical Chemistry A (2011), doi: 10.1021/jp203519p. [28] E. Wernersson, J. Heyda, A. Kub´ıˇckov´a, T. Kˇr´ıˇzek, P. Coufal, and P. Jungwirth, “Effect of association with sulfate on the electrophoretic mobility of polyarginine and polylysine”, Journal of Physical Chemistry B 114(36), pp. 11934–11941 (2010). [29] Anna Kub´ıˇckov´ a, Tom´ aˇs Kˇr´ıˇzek, Pavel Coufal, Erik Wernersson, Jan Heyda, and Pavel Jungwirth, “Guanidinium cations pair with positively charged arginine side chains in water”, The Journal of Physical Chemistry Letters 2(12), pp. 1387–1389 (2011). [30] A. E. Miller, P. B. Petersen, C. W. Hollars, R. J. Saykally, J. Heyda, and P. Jungwirth, “Behavior of beta-amyloid 1-16 at the air-water interface at varying ph by nonlinear spectroscopy and molecular dynamics simulations”, Journal of Physical Chemistry A 115(23), pp. 5873–5880 (2011). [31] J. W. Neidigh, R. M. Fesinmeyer, and N. H. Andersen, “Designing a 20-residue protein”, Nature Structural Biology 9(6), pp. 425–430 (2002).

22

REFERENCES

[32] D. R. Canchi and A. E. Garcia, “Backbone and side-chain contributions in protein denaturation by urea”, Biophysical Journal 100(6), pp. 1526–1533 (2011). [33] P. E. Mason, C. E. Dempsey, L. Vrbka, J. Heyda, J. W. Brady, and P. Jungwirth, “Specificity of ion-protein interactions: Complementary and competitive effects of tetrapropylammonium, guanidinium, sulfate, and chloride ions”, Journal of Physical Chemistry B 113(10), pp. 3227–3234 (2009). [34] Christopher E. Dempsey, Philip E. Mason, and Pavel Jungwirth, “Complex ion effects on polypeptide conformational stability: Chloride and sulfate salts of guanidinium and tetrapropylammonium”, Journal of the American Chemical Society 133(19), pp. 7300– 7303 (2011).

23

Curriculum vitae Name: Jan Heyda ˇ y Krumlov Born: 18th June 1983 in Cesk´ Email: [email protected] Address: Lipovsk´a 436, Praha 5 – Zliˇc´ın, 155 21 Education: • 2007Faculty of Science, Charles University in Prague, study program: Physical Chemistry; Modelling of Chemical Properties of Nano- and Biostructures PhD project: Ion-protein interactions Supervisor: Prof. Mgr. Pavel Jungwirth, DSc. • 2007Member of the International Max Planck Research School, Dresden “Dynamical Processes in Atoms, Molecules and Solids” • 2002-2008 Faculty of Mathematics and Physics, Charles University in Prague, study program: Mathematics; Mathematical Modeling in Physics and Engineering Diploma Thesis: Distribution of ions at surfaces of hydrated proteins Supervisor: Doc. Mgr. Pavel Jungwirth, CSc. • 2002-2007 Faculty of Science, Charles University in Prague, study program: Physical Chemistry; group: Electromigration and Separation Methods Diploma Thesis: Multidimensional Simulation of Electromigration Supervisor: Prof. RNDr. Bohuslav Gaˇs, CSc. Work Experience: • Since October 2006 Academy of Sciences of the Czech Republic, Institute of Organic Chemistry and Biochemistry • Since January 2006 – June 2008 Work in program COMSOL Multiphysics with respect to numerical solution of partial-differential equations • March 2007 – June 2008 Basic Research Project with Agilent Technologies

Selected Publications/Seznam publikac´ı 1. J. Heyda, T. Hrob´ arik, and P. Jungwirth, “Ion-specific interactions between halides and basic amino acids in water”, Journal of Physical Chemistry A 113(10), pp. 1969–1975 (2009). 2. P. E. Mason, C. E. Dempsey, L. Vrbka, J. Heyda, J. W. Brady, and P. Jungwirth, “Specificity of ion-protein interactions: Complementary and competitive effects of tetrapropylammonium, guanidinium, sulfate, and chloride ions”, Journal of Physical Chemistry B 113(10), pp. 3227–3234 (2009). 3. J. Vondr´ aˇsek, P. E. Mason, J. Heyda, K. D. Collins, and P. Jungwirth, “The molecular origin of like-charge arginine-arginine pairing in water”, Journal of Physical Chemistry B 113(27), pp. 9041–9045 (2009). 4. J. Heyda, J. Pokorn´ a, L. Vrbka, R. V´acha, B. Jagoda-Cwiklik, J. Konvalinka, P. Jungwirth, and J. Vondr´ aˇsek, “Ion specific effects of sodium and potassium on the catalytic activity of HIV-1 protease”, Physical Chemistry Chemical Physics 11(35), pp. 7599–7604 (2009). 5. J. Heyda, J. C. Vincent, D. J. Tobias, J. Dzubiella, and P. Jungwirth, “Ion specificity at the peptide bond: Molecular dynamics simulations of N-methylacetamide in aqueous salt solutions”, Journal of Physical Chemistry B 114(2), pp. 1213–1220 (2010). 6. J. Heyda, P. E. Mason, and P. Jungwirth, “Attractive interactions between side chains of histidine-histidine and histidine-arginine-based cationic dipeptides in water”, Journal of Physical Chemistry B 114(26), pp. 8744–8749 (2010). 7. J. Heyda, M. Lund, M. Onˇc´ ak, P. Slav´ıˇcek, and P. Jungwirth, “Reversal of hofmeister ordering for pairing of NH+ 4 vs alkylated ammonium cations with halide anions in water”, Journal of Physical Chemistry B 114(33), pp. 10843–10852 (2010). 8. E. Wernersson, J. Heyda, A. Kub´ıˇckov´a, T. Kˇr´ıˇzek, P. Coufal, and P. Jungwirth, “Effect of association with sulfate on the electrophoretic mobility of polyarginine and polylysine”, Journal of Physical Chemistry B 114(36), pp. 11934–11941 (2010). 9. P. E. Mason, J. Heyda, H. E. Fischer, and P. Jungwirth, “Specific interactions of ammonium functionalities in amino acids with aqueous fluoride and iodide”, Journal of Physical Chemistry B 114(43), pp. 13853–13860 (2010). 10. Anna Kub´ıˇckov´ a, Tom´ aˇs Kˇr´ıˇzek, Pavel Coufal, Erik Wernersson, Jan Heyda, and Pavel Jungwirth, “Guanidinium cations pair with positively charged arginine side chains in water”, The Journal of Physical Chemistry Letters 2(12), pp. 1387–1389 (2011). 11. A. E. Miller, P. B. Petersen, C. W. Hollars, R. J. Saykally, J. Heyda, and P. Jungwirth, “Behavior of beta-amyloid 1-16 at the air-water interface at varying pH by nonlinear spectroscopy and molecular dynamics simulations”, Journal of Physical Chemistry A 115(23), pp. 5873–5880 (2011). 12. J. Heyda, M. Koˇz´ıˇsek, L. Bedn´arov´a, G. Thompson, J. Konvalinka, J. Vondr´aˇsek, and P. Jungwirth, “Urea and guanidinium induced denaturation of a trp-cage miniprotein”, The Journal of Physical Chemistry B (2011), doi: 10.1021/jp200790h. 13. Mario Vazdar, Jiˇr´ı Vymˇetal, Jan Heyda, Jiˇr´ı Vondr´aˇsek, and Pavel Jungwirth, “Likecharge guanidinium pairing from molecular dynamics and ab initio calculations”, The Journal of Physical Chemistry A (2011), doi: 10.1021/jp203519p.