2. The Structural and Physico-Chemical Effects of Halogen Atoms in Polypeptides
The introduction of halogen atoms provides many more benefits than other types of substitutions in drugs
[9]. In fact, it is estimated that up to one third of drugs undergoing clinical investigation is halogenated
[10]. This type of substitution can be used to control the degradability of pharmaceuticals as well as their lipophilicity, membrane permeabilization activity and catabolic stability
[11][12]. The ligand-receptor interaction can be modulated by the presence of halogen atoms because of their electron-withdrawing features and their steric-effects
[10][13]. This can interestingly find applications from very small molecules to large complexes such as antibodies. For example, halogenated Tyr improved 10 fold the binding affinity of the Fab fragment of an anti-EGF-receptor (059-152) for its antigen
[14].
In this role, the halogen atom can also act as an electron donor by allowing the formation of stabilizing interactions such as hydrogen bonds
[15][16]. Although it is hydrophobic in nature, the halogen introduces some interesting physico-chemical features not normally associated with this property, such as allowing for the possibility of F···H–N hydrogen bond formation. This is confirmed by through-space coupling in NMR studies
[17][18].
The introduction of fluorine into a peptide or a protein offers rich opportunities both when the fluorination site is on an amino acid side chain or on the backbone itself, because of its capacity to locally alter the electronic character of the peptide/protein
[19]. The fluorination also has an effect on the folding properties, changing the secondary structure propensity of peptides and protein segments containing aliphatic halogenated amino acids. In addition, fluorine has an impact on the proteolytic stability of peptides, in a manner that depends both on the nature of the fluorinated side chain but also on its effect on the immediate environment
[20]. Resistance towards proteolytic digestion due to the introduction of halogenated amino acids cannot however be generalized. It depends on the substitution position and number of introduced halogen atoms on the amino acid side chain. For example, a minimal presence of fluorinated, small side chain amino acids can improve remarkably proteolytic stability
[12]. Futhermore, bulky fluorinated residues, such as hexafluoroleucine or trifluoroisoleucine protect polypeptides from proteolytic enzymes. On the other side, fluorinated aromatic amino acids are comparable to their hydrocarbon counterparts in maintaining the proteolytic stability
[12]. These observations are rendered even more interesting by studies showing that the introduction of fluorine is relatively non-invasive with respect to structural stability. For example, Kovermann and co-workers investigated the thermodynamic stability and structural integrity of Cold shock protein B from Bacillus subtilis (BsCspB), containing fluorine-labelled phenylalanine or tryptophan residues. Time-resolved fluorescence kinetics was used to monitor chemical denaturation of fluorine-labelled BsCspB, while thermal denaturation was followed by high-resolution
1H and
19F NMR, X-ray
[21]. These experiments showed that the presence of F-Phe or F-Trp residue caused only a barely detectable change in thermodynamic stability in comparison to the wild type protein. Moreover, the X-ray structures of fluorinated CBsCspB variants were almost completely superposable with the structure of the wild type protein.
However, since the alignment of C-F bonds with adjacent C=O or C-N bonds is predictable, fluorinated amino acids might be used as a tool to study and control the secondary structure of polypeptides
[22]. In this context, 4-fluoroprolines are excellent building blocks to study the pharmacological and structural impact caused by their substitution, especially for the stability and kinetics of protein folding. In this context, the rational substitution of Pro or Hpy with a 4-fluorinated Pro analogue in peptides and proteins can enhance both the thermodynamic and hydrolytic stability, as well as the resistance to unfolding in organic solvents of the engineered protein
[23][24].
The pyrrolidine ring of proline introduces a conformational rigidity and constrains into polypeptide backbone, as two of the main chain atoms are constricted on the ring
[25][26][27][28]. The pyrrolidine ring adopts predominantly two different conformations. In the endo conformation (C
γ-endo), carbon 4 and α-carbonyl carbon are on the same side of the plane, while in the exo conformation (C
γ-exo), carbon 4 on the pyrrolidine ring points towards the opposite side of the carbonyl group of Pro (
Figure 2). The presence of fluorine on the C
γ atom, with its electron-withdrawing properties, can bias the proline ring pucker preferences and “pre-organize” the protein main chain, thus conferring additional stabilization to the whole protein
[29][30]. Thus, (2S, 4S)-4-fluoroproline ((4S)-FPro), prefers the C
γ-endo ring pucker and stabilizes the cis peptide-bond conformation, while (2S, 4R)-4-fluoroproline, ((4R)-FPro) stabilizes C
γ-exo ring pucker favoring the trans conformation
[28][31][32] (
Figure 2).
Figure 2. Fluorine effect on proline conformation. (
A) pyrrolidine ring puckers in proline, (4R)-FPro, and (4S)-FPro. Fluorine on the C
γ atom introduces a chiral center (indicated by *) with effect on the preference of endo/exo-pucker. (
B) cis/trans isomerization by Pro. Figure adapted from Ref.
[32].
In folded polypeptides, Xaa-Pro (Xaa = any amino acid) peptide bonds still display preferentially the trans conformation, although in this case both conformations are almost isoenergetic
[33]. The cis/trans isomerization of Xaa-Pro bonds is critical during the folding process in polypeptides. The refolding of globular proteins presents an intrinsically slow isomerization step when non-native trans prolyl-peptide bonds in the unfolded state undergo isomerization to reach their native cis state
[25]. Therefore, the cis/trans isomerization reaction is typically the rate-limiting step in the native protein folding or refolding processes. The substitution of 10 prolines (9 trans and 1 cis) in the Enhanced Green Fluorescent Protein (EGFP) with (4S)-FPro promoted a faster refolding
[25]. However, these results were attributed to several synergistic effects, arising mostly from the adoption of the endo pucker in 9 out of 10 Pro residues, rather than from cis/trans isomerization of prolines.
In 2011, Rubini and co-workers showed that the replacement of Pro 19, Pro 37, and Pro 38 (all displaying an exo pucker and all involved in trans peptide bonds) in human ubiquitin with (4R)-FPro led to the generation of a protein that was 4.71 kJ·mol
−1 more stable than the parent protein. Moreover, the introduction of (4R)-FPro did not affect the folding mechanism or biological activity
[26]. Recently, Rubini and co-workers studied the impact of 4- and 4,4-difluorinated proline analogues on thioredoxin refolding kinetics. The replacement of the conserved cis Pro76 in the E. coli thioredoxin variant Trx1P with (4S)-FPro
[27] led to a 9-fold acceleration of the rate-limiting folding step, i.e. cis/trans isomerization of Xaa-Pro bond. This step occurs in the context of a long-lived folding intermediate (I
trans) that displays an intact tertiary structure with a buried non-native trans Pro76
[34]. Interestingly, the Pro to (4R)-FPro
[27] and 4,4-difluoroproline (4,4-FPro)
[35] substitution at position 76 in Trx1P did not lead to an increase in the refolding rate (
Table 1). The high-resolution crystal structures of Trx1P and of its monofluorinated variants (
Figure 3) could not explain the difference in the refolding kinetics, as the three structures are almost indistinguishable
[36].
Figure 3. Overlay of chain A of Trx1P (cyan) on oxidized chain A of Trx1P-(4R)-FPro (light gray), on the right, and chain A of Trx1P-(4S)-FPro (brown), on the left. The cis-Pro76, cis-(4R)-FPro and cis-(4S)-FPro residues and the side chains of Cys32 and Cys35 forming the catalytic disulfide bond are shown in sticks, adapted from Ref.
[36].
Table 1. Kinetic refolding parameters of Trx1P, Trx1-(4R)-FPro, Trx1-(4S)-FPro, and Trx1-(4,4)-FPro.
|
k of Itrans→Ncis [s−1]a (In the Context of Tertiary Structure) |
Normalised Rate for k of Itrans→Ncis |
Ref. |
Trx1P |
9.3 ± 0.5 × 10−5 |
1 |
[35] |
Trx1-(4R)-FPro |
8.7 ± 2.5 × 10−5 |
0.94 |
[27] |
Trx1-(4S)-FPro |
8.0 ± 2.5 × 10−4 |
8.6 |
[27] |
Trx1-(4,4)-FPro |
7.9 ± 0.7 × 10−5 |
0.85 |
[35] |
A possible explanation for the folding behavior of these fluorinated protein variants could lie in the transition state of the refolding reaction, instead of in the 3D structure of the native state. The incorporation of (4
S)- and (4
R)-FPro at
cis Pro48 led to similar results for pseudo-wild type barstar protein
[37]. Similarly, to Trx1P, the
trans to
cis isomerization around Tyr47-Pro48 in pseudo wild type barstar during refolding, occurs in the context of a structured intermediate with Pro48 in
trans and it is the bottleneck of the refolding reaction. Incorporation of (4
S)-FPro at
cis Pro48 promoted a 3.5-fold acceleration of I
trans to N
cis (native state), whereas (4
R)-FPro at
cis Pro48 did not induced changes in the refolding kinetics, in comparison to the parent protein. However, incorporation of 4,4-FPro at the same position induced a 3.6-fold increase of the I
trans-to-N
cis folding rate.
Torbeev and Hilvert
[38] showed that the incorporation of 4,4-FPro at cis Pro32 into human β2-microglobulin eliminated the rate limiting step for the folding of this protein. The authors suggested that the strong electron-withdrawing effect of the two fluorine atoms on the Pro ring dramatically accelerated the
trans to
cis prolyl bond isomerization. In fact, in comparison to the 4-monofluorinated Pro isomers, 4,4-FPro shows the lowest energy barrier for
cis/trans isomerization; however, its application for the acceleration of folding of globular proteins seems to be context-dependent.
The introduction of (4
R)-FPro stabilized the conformation of proline-rich peptides in aqueous solution. The proline-rich sequences in proteins that adopt a polyproline type I (PP I) helical conformation are characterized by the
cis orientation for the Xaa-Pro bond, while PP II helical segments display prolyl amide
trans isomers. The effect of C
γ exo/endo isomerization of the Pro ring pucker influences the
cis/trans conformer ratio of the amide Xaa-Pro bond
[28][39][40][41] as recently demonstrated by NMR, X-ray and molecular dynamics (MD) studies
[42]. In collagen mimetics, the substitution of (4
S)-Hyp or (4
R)-FPro for the Pro residues in the sequence stabilizes the C
γ exo pucker and, thus, the PP II helix conformation. On the other hand, the substitution with (4
S)-FPro at the Yaa-Pro position caused a relative decrease in thermostability, highlighting the impact of the 4-position stereochemistry
[28]. The (4
R)-FPro enforces the
exo ring pucker, favoring the prolyl amide
trans isomer
[28][40][43]. With respect to the placement of F-substituted residues, Lin et al. found, using short model peptides, that C-terminal stereoelectronic effects might influence the stability of PP II conformation and the PP II/PP I conversion rate more than the N-terminal effects
[44]. All these investigations have demonstrated that the use of FPro is a rational tool to probe polypeptide stability, dynamics, and conformational and structural effects, although some complications were observed possibly due to the chemical and structural constraints imposed by these complex molecules.
The study of structure-activity relationships is now supported also by computational approaches. The application of molecular dynamics (MD) simulations on peptides/peptidomimetics has become an important tool. In particular, the successful prediction of the folding and dynamics of several proteins and peptides, using MD simulations, has significant applications in drug design studies
[45][46][47][48]. Their performance is linked to the accuracy of the empirical force fields used in the simulations
[49]. Among noncovalent interactions, that are recognized to be important, the halogen bond (or X-bond) is an intriguing tool for engineering protein–ligand interactions and for controlling the structures of proteins and nucleic acids. A structure−activity relationship study between halogenated compounds and a subtype of 5-hydroxytryptamine receptor (5-HT
7R) is a good example
[50]. The development of new computational modeling tools for X-bonds in biological molecules, has been reported based on a set of potential energy functions that describe the anisotropic electrostatic and shape properties of the halogens participating in them
[51][52]. Recently, the force field has been generalized by reducing the number of variables to just one for each halogen type and by estimating the electrostatic variable through a standard restricted electrostatic potential calculation of atomic charges. The simplified force field was validated against the AMBER force field, showing that Rappè et al.’s force field
[53] is more adaptable for incorporation into classical molecular mechanics/dynamics algorithms, including those commonly used to design inhibitors against therapeutic targets in medicinal chemistry and materials in biomolecular engineering. Halogenation, and in particular fluorination, can improve several features of proteins, such as thermal and proteolytic stability and/or their enzymatic activity. A study of the impact of fluorination on hydrophobicity was recently carried out using dynamics simulations, together with a new fixed-charge, atomistic force field, to quantify the changes in hydration free energy, ΔGHyd, for amino acids with alkyl side chains and with 1 to 6 -CH → -CF side chain substitutions
[54]. The results underline two main contributions traceable to alteration of side chain-water interactions and of the number of backbone-water hydrogen bonds. In conclusion, in recent years, many design tools have been developed to mechanically understand how fluorination plays a role in structure-activity relationships. This knowledge can be applied downstream in several ways, ranging from the effect on the hydrophobicity of (bio) polymers, to electrostatic properties.