Abstract
The increasing utilization of fluorine in drug design parallels advances in understanding the physicochemical attributes of this element and an enhanced appreciation of how these unique properties can be exploited to address the numerous challenges encountered in pharmaceutical candidate optimization. Judicious placement of fluorine in a candidate compound can markedly affect potency, increase metabolic stability and enhance membrane permeability. The powerful electron-withdrawing nature of fluorine serves to modulate the pK a of proximal functionality, particularly basic amines, and can be an important tool for controlling physical properties. Fluorine also exerts a conformation bias that is significant and can be utilized strategically. The 18F isotope is of particular importance in positron emitting tomography (PET), an imaging technique of increasing importance for assessing drug-target engagement in both preclinical and clinical settings. This chapter provides a synopsis of some of the prominent and emerging applications of fluorine in the design and optimization of biologically active molecules.
Access provided by Autonomous University of Puebla. Download chapter PDF
Similar content being viewed by others
Keywords
- Gauche interaction
- Fluorine and basicity
- Fluorine and conformation
- Fluorine and hydrogen-bonding
- Fluorine and membrane permeability
- Fluorine and metabolic stability
- Positron emitting tomography imaging
1 Introduction
Fluorine is a remarkable and unique atom capable of imparting a myriad of advantageous properties to small molecules, particularly in the context of medicinal chemistry. Over the last five decades the knowledge and understanding of this almost “magical” substituent has been enhanced substantially. Ever improving technologies for the installation of fluorine into novel small molecules facilitate attempts to utilize fluorine to address numerous challenges encountered in drug design: conformational control; modulating pK a; improving metabolic stability, solubility, permeability and potency; or tracking the distribution of a compound in vivo using positron emission tomography (PET) imaging. Recently, fluorine has found utility in fragment-based drug design by taking advantage of the 19F NMR signal to both identify fragments that bind to targeted sites and to probe the presence of fluorophilic environments in proteins [1–3]. The applications of fluorine in medicinal chemistry continue to increase and its strategic deployment is considerably less empirical than it once was. However, there remains much to be understood regarding this privileged atom and its ability to aid in the fine tuning of pharmacodynamic (PD) and pharmacokinetic (PK) properties.
Fluorine is the most abundant halogen in the earth’s crust (and ranks 13th in abundance of all elements), but this has not translated into a proportional prevalence among naturally occurring organic compounds. There are only a handful of natural products known to contain fluorine, attributed to the fact that fluorine is found naturally in insoluble forms that are not readily available [4].
For example, seawater concentrations of fluoride have been measured at 1.3 ppm compared to ubiquitous chloride concentrations which are 1,900 ppm. Another factor contributing to the dearth of fluorine in organic molecules is the extremely high heat of hydration of fluoride [5]. Fluoride is heavily hydrated in water rendering it weakly nucleophilic and thus limiting its role in direct displacement chemistry. The high heat of hydration of fluoride relative to other halides is principally responsible for the inordinately high energy needed to generate F+ from aqueous F−[6]. This key difference from other halides precludes incorporation of fluorine into natural products via the haloperoxidase reaction that is a common pathway for the incorporation of chlorine, bromine and iodine [7, 8].
Despite the low concentrations of fluorine available on land and in the sea, some marine and terrestrial organisms can accumulate significant quantities of inorganic fluorine. However, the presence of organo-bound fluorine has only been identified in a handful of tropical and sub-tropical terrestrial plants and in only two microorganisms. A great deal of effort has gone into the elucidation of the biosynthesis of fluorine-containing natural products [8–11]. The most prominent example is the synthesis of highly toxic fluoroacetic acid by the bacterium Streptomyces cattleya from S-adenosyl-L-methionine and fluoride mediated by the enzyme 5′-fluoro-5′-deoxy adenosine synthetase [9–11]. However, to date there is no evidence to support the de novo synthesis of the C-F bond in the animal or insect kingdoms [4]. Indeed, caterpillars of the moth Sindus albimaculatus acquire trifluoroacetate from their host plant, Dichapetalum cymosum, in concentrations that render them toxic to predators [12]. It is interesting that despite the scarcity of fluorine-containing compounds found in the natural world, fluorine plays a prominent and ever increasing role in contemporary drug design.
One of the earliest synthetic fluorinated drugs is 5-fluorouracil, an antimetabolite synthesized for the first time in 1957 (Fig. 1) [13]. Since that discovery, the deployment of fluorine in small molecule drug candidates has grown, as have the technologies to install this atom. Today, 20–25 % of marketed drugs contain fluorine, including blockbusters such as fluoxetine (Prozac®, depression), celecoxib (Celebrex®, arthritis), efavirenz (Sustiva®, human immunodeficiency virus (HIV)), sitagliptin (Januvia®, diabetes), atorvastatin (Lipitor®, dyslipidemia), rosuvastatin (Crestor®, dyslipidemia), cinacalcet (Sensipar®, hyperparathyroidism) and the nucleosides emtricitabine (Emtriva®, HIV). Other prominent fluorinated drugs include sofosbuvir (Solvadi®), capecitabine (Xeloda®, breast and colorectal cancer), the HIV integrase inhibitors raltegravir (Isentress®, HIV), elvitegravir (HIV) and dolutergavir (Tivicay®), ciprofloxacin, the prototypical fluoroquinolone antibiotic, linezolid (Zyvox®, antibiotic) and paroxetine (Paxil®, antidepressant). The objective of this chapter is to provide an overview of the impactful nature of fluorine on drug discovery as it is currently understood.
2 Fluorine and Conformation in Drug Design
The high electronegativity of fluorine is responsible for the strong polarization of the C-F bond, manifested as the highest measured dipole in organic molecules at 1.85 D for CH3F. This phenomenon can strongly influence the interaction of C-F with a range of proximal functionality [14–22].
Depending on the nature of the functionality, these effects can be of sufficient magnitude to significantly affect conformational preferences, and can be exploited in the design of both drugs and organocatalysts [21, 22]. The basis for these intramolecular effects include dipole-dipole interactions, attractive electrostatic effects, repulsive effects between F and another electronegative atom, p-orbital repulsion, and a hyperconjugative effect in which the C-H bond of an adjacent atom interacts with the low lying σ* orbital of the C-F bond [21–23]. An understanding of the interactive relationship between F and a range of organic functional groups has been developed using density functional theory (DFT) calculations that provide some instruction on potential utility and applications [23]. Table 1 captures the predicted energy of an individual interaction and offers a possible explanation for the preferred conformation based on the underlying physical chemistry [23]. Whilst several of these specific structural elements are of limited interest in drug design and the inherent energy of some interactions are too low to be of significance in the absence of additional reinforcing effects, the underlying physical organic chemistry principles offer considerable insight into the role that fluorine plays in influencing conformation. The most prominent functionality for which the energy of interaction with fluorine is such that the effect might be gainfully exploited in drug design are the amine, alcohol and amide moieties [21, 22].
2.1 Fluorine-Fluorine Interactions
The conformation of fluoroalkane derivatives is of considerable contemporary interest. The influence of fluorine on the conformation of vicinally substituted compounds is based on a stabilizing gauche interaction that relies upon hyperconjugation between a C-H bond and the low lying σ* orbital of the C-F bond. The energy of this interaction is estimated to be 0.8 kcal mol−1, modest but nevertheless sufficient to exert an influence on conformation [21, 24] (Fig. 2a). In addition, there is a destabilizing repulsion between fluorine atoms in a 1,3-relationship that is of higher energy and which minimizes unfavorable dipole-dipole interactions [21, 24, 25]. The preferred conformation of 1,3-difluoropropane shown in Fig. 2b is estimated to be 3.3 kcal/mol−1 more stable than the alternative [21, 24, 25].
An interesting example of the effect of the preference for a gauche arrangement of vicinal fluorine atoms on physical properties is provided by the two diastereomers of 9,10-difluorostearic acid, 1 (erythro) and 2 (threo), which were carefully synthesized in a fashion that ensured stereochemical integrity (Fig. 3) [24, 26]. These molecules differ only by the stereogenicity of the fluorine atom at C-10 yet their melting points differ by 20 °C, with the erythro isomer 1 melting at 67–69 °C while for threo isomer 2, the melting point is 86–88 °C. The threo isomer 2 was anticipated to be the more stable isomer, with the fully extended form favored by the conformational preference exerted by the gauche interaction between the two F atoms, a conformation similar to that adopted by the native fatty acid. Langmuir isotherm measurements confirmed the similarity between 2 and stearic acid whilst the erythro isomer 1 displayed significant disorder, consistent with the lower melting point [24, 26]. In the Newman projections about the C9-C10 bond of 1, the stabilizing effect of the gauche-disposed F atoms is offset by unfavorable interactions between the two carbon chains such that none of the conformations are preferred although all are accessible.
Derivatives in which the vicinal difluoromethylene motif is extended to include 3, 4, 5 and 6 fluorine atoms have been examined to determine conformational preferences [24]. The most sophisticated study focused on the hexafluoromethylene motif, carefully assembled in the diastereomers 3 and 4, that extended earlier studies on tetra- and pentafluoro alkanes [24, 27]. The conformation of these compounds, both in the solid state and in solution, indicates that in order to avoid unfavorable 1,3-dipolar interactions between F atoms, 3 adopts a helical topography that simultaneously allows the vicinal fluorine atoms to adopt the preferred gauche arrangement. The same effect favors an extended anti conformation for 4 that presents no unfavorable 1,3-dipolar interactions and facilitates 3 out 5 possible gauche arrangements of the vicinal fluorine atoms [24, 27].
2.2 Fluorine-Amide Interactions
The conformation of α-fluoro-substituted amides is dominated by dipole-dipole interactions between the C-F and C=O bonds, with additional support from a productive electrostatic interaction between the fluorine atom and the amide NH if the molecule is configured appropriately [20, 28–33]. The synthesis and analysis of the enantiomers of α-fluoro capsaicin (6) as vanilloid receptor agonists is illustrative of the potential to take advantage of this kind of effect in drug design. Ab initio calculations of the rotational energy profile of 2-fluoro-N-methylpropanamide indicated that the trans conformer was the most stable with a 6 kcal mol−1 advantage over the gauche conformer which was 2 kcal mol−1 more stable than the cis conformer (Fig. 4) [28, 29, 33]. The stability of the trans conformation is attributed to a favorable alignment of the C-F and C=O bond dipoles supported by a productive electrostatic interaction between the F and amide H atoms. In the case of 6, the (R) and (S)-enantiomers would be expected to adopt conformations that stereo-differentiated the side chain projection vectors, potentially providing insight into the bound conformation of the natural 5 [33].
However, both the (R) and (S)-enantiomers of 6 performed similarly as TRPVI receptor agonists, leading to the conclusion that the conformation captured by 5 in Fig. 4 in which the alkyl side chains project in vectors parallel to each other and in the plane defined by the amide moiety represents that of the natural product rather than that in which the alkyl chains project orthogonally to the plane of the amide, as depicted in (R)-6 and (S)-6 in Fig. 4 [33].
A study examining the effect of fluorination on the conformational preferences of β-peptide derivatives installed the fluorine atom α- to the amide carbonyl, a topology that simultaneously placed the F β- to the amine moiety. This structural arrangement allowed an analysis of the roles of the dipole-dipole effect between F and C=O and the gauche interaction between vicinally-deployed F and amide moieties [34]. X-ray crystallographic structures of the synthesized compounds revealed that the C-F and C=O bonds aligned in an anti-periplanar conformation to satisfy the dipole interactions whilst the F and amide moieties adopted the anticipated gauche arrangement [34, 35]. In the representative compound 7, the torsion angle associated with the fluoropropionamide moiety in the solid state was −172.46° whilst that for the other α-fluoroamide was −179.7°. For the internal F-CH2-CH2-NHCO element, a torsion angle of −73.7° satisfied a gauche arrangement; however, in this particular example the F atom of the terminal NH-CH2-CH2-F moiety was disordered, interpreted as an effect of packing forces overcoming a gauche preference calculated to be 1.8 kcal mol−1 [34, 35]. In the ester 8, the torsion angle between the F atom and the ester C=O moiety was 154.5°, reflecting a reduced preference that is based on the weaker dipole associated with an ester that manifests as a calculated stabilization energy for the anti-periplanar form of 4.5 kcal mol−1, which compares to 7.5 kcal mol−1 for an α-fluoro amide and 2.2 kcal mol−1 for an α-fluoro ketone [20, 28, 29, 35, 36].
The difluoro succinamide esters 9 (threo) and 11 (erythro) and the corresponding acids 10 (threo) and 12 (erythro) were prepared and the solid state structures determined which revealed conformational differences that were reflected in the preservation of a vicinal F-F gauche interaction in all 4 molecules [31, 32]. In the case of 11, one of the fluorine-carbonyl pairs adopted a higher energy conformation that sacrificed the dipole-preferred anti-periplanar arrangement whilst in 9, 10, and 12 all arrangements between F and the C=O moieties were in, or very close, to the preferred anti-periplanar topography [31, 32]. NMR studies that analyzed 3 J HF and 3 J HH coupling constants revealed that these conformations were also prevalent in solution, underscoring the influence of the F-gauche effect.
The preference for a gauche relationship between fluorine and the N-CO moiety observed in 7 and 8 is of broader significance as a conformational tool. In the solid state, the conformation of 2-fluoro-N-(2-fluoroethyl)-propionamide (13) reflects a gauche relationship between the fluorine and NH-CO moieties, torsion angle = −69.9°, whilst the F atom α- to the amide carbonyl is aligned in a fashion consistent with the preferred dipole interactions in which the two are anti-periplanar, torsion angle = 2° [29]. A similar phenomenon is observed with 14 [35].
A particularly striking example of the gauche effect in β-fluoroamides is provided by the two 4-F proline diastereomers 15 and 16. Here the fluorine atom exerts a significant influence on the conformational mobility of this unique amino acid, providing insight into the stabilizing influences of 4-hydroxyproline (17, Hyp), prevalent in the natural protein collagen [22]. Collagen is comprised of repeat units of Xaa-Yaa-Gly in which Xaa and Yaa are frequently proline or 4-(R)-Hyp. The polypeptide associates into a tightly-wound triple helix fibril, the thermal stability of which is enhanced by the presence of the hydroxyl substituent [37–42]. Although this phenomenon was initially attributed to the H-bonding properties of the hydroxyl moiety, a triple helix assembled from (Pro-4-(R)-F-Pro-Gly)10 exhibits a CD spectrum identical to (Pro-Hyp-Gly)10 or (Pro-Pro-Gly)10 with all spectral data consistent with triple helix formation [37]. These observations support the conclusion that the stabilizing effect of Hyp (17) is based on an inductive effect rather than H-bonding. Following this observation, additional studies have evaluated the consequence of substituting the Xaa and Yaa positions with 15 and 16 rather than proline and Hyp (17), with the results leading to a deeper understanding of collagen conformation [38–43]. Proline has an inherent preference for the Cγ-endo conformation, calculated as 0.41 kcal mol−1, which favors this conformer by 2:1 at room temperature (Fig. 5) [44]. The introduction of a 4-(S)-F substituent (16) augments this preference, increasing the stability to a difference of 0.61 kcal mol−1, stabilized by the gauche effect between the F and amide moieties that facilitates a conformation in which the F atom adopts an axial disposition (Fig. 5) [43, 44]. This effect is manifested in the single crystal X-ray of Boc-4-(S)-fluoroproline [43]. However, in 4-(R)-fluoroproline (15) the Cγ-exo conformation is favored by 0.85 kcal mol−1, stabilized by the gauche effect between the F and amide moieties that also disposes the F atom axially [44]. This phenomenon is mimicked by Hyp (17) where the Cγ-exo conformer is calculated to be 0.48 kcal mol−1 more stable than the Cγ-endo conformer [45]. These insights provide an explanation for the structural similarity of (Pro-4-(R)-F-Pro-Gly)10, (Pro-Hyp-Gly)10 and (Pro-Pro-Gly)10. Additionally, in 15–17 the E trans isomer of the peptide bond is calculated to be more stable than the E cis isomer regardless of the ring pucker, with the energy differences providing a stability hierarchy of 4-(R)-Pro > Pro > 4-(S)-Pro (Fig. 5) [38, 44]. The preference for an E trans geometry has been attributed to this conformation allowing a favorable n-π* interaction between the O atom of the C=O moiety bound to the ring N atom and the C=O carbon of the exocyclic amide, as depicted in Fig. 5 [38, 46–48].
Molecular modeling studies have suggested that when proline is in the Xaa position of a (Pro-Pro-Gly)10 helical strand, it adopts a Cγ-endo pucker but when in the Yaa position a Cγ-exo conformation is preferred, consistent with the observations noted above that (Pro-4-(R)-F-Pro-Gly)10, (Pro-Hyp-Gly)10 and (Pro-Pro-Gly)10 form stable helical complexes [39, 49]. When a substituted proline is installed at the Xaa position, the conformational preference for the Cγ-endo conformation would suggest that 4-(S)-F-Pro would function to form a collagen-like helix structure but the 4-(R)-F-Pro diastereomer would not based on their conformational bias [39, 43]. This has been observed experimentally with both (4-(S)-F-Pro-Pro-Gly)10 and 4-(S)-F-Pro-Pro-Gly)7 forming stable triple helices while (4-(R)-F-Pro-Pro-Gly)10 and (4-(R)-F-Pro-Pro-Gly)7 do not [39, 43]. Similarly, 4,4-difluoroproline (18), which offers no conformational bias, does not substitute effectively for Hyp (17) to form stable triple helices in collagen-related peptides [42].
Incorporation of fluorine at the 3-postion of proline also introduces a conformational bias, a phenomenon initially explored with the diastereomeric esters 19 and 20 [50]. These molecules crystallized with the trans configuration between the amide and ester moieties, as depicted in 19 and 20, but the conformation of the proline rings differed, each stabilized by a gauche interaction between the F and amide moieties that deployed the halogen in an axial orientation. In 19, the pyrrolidine ring adopted the Cγ-exo conformation, which contrasts with 20 where the Cγ-endo arrangement prevailed [50]. The stereoelectronic effects of the 3-F-prolines 19 and 20 are complementary to those of the 4-F isomers 15 and 16. Thus, an anti orientation of the fluorine atom with respect to the ester moiety results in the Cγ-endo conformation for 20 but a Cγ-exo for 15 with the opposite effect for the syn epimers [50]. The Cγ-endo conformation of 20 was found to be present in solution as assessed by the coupling constant between the α- and β-protons in the 1H NMR spectrum. This coupling constant was ≤1 Hz and the calculated thermodynamic parameters suggested that this would be populated to the extent of 97 %. In contrast, the Cγ-exo conformation of 19 was predicted to amount to 69 % of the population, attributed to repulsion between the F atom and adjacent ester moiety as a consequence of unfavorable steric and electronic interactions [41, 50]. The conformational preferences of 19 and 20 were preserved when these proline derivatives were incorporated into short peptide sequences [41, 50, 51].
2.3 Fluorine-Amine Interactions
Protonated β-amino fluoroalkanes show a strong preference to adopt a gauche conformation based on a favorable electrostatic interaction between the ammonium moiety and the electronegative F atom, the energetics of which are estimated to be 5.8 kcal/mol [20, 23, 52]. In the unprotonated form, 2-fluoroethylamine experiences only a weakly stabilizing intramolecular interaction between the NH2 and F that has been viewed as a bridging H-bond estimated at ~1 kcal mol−1 in favor of the gauche conformation, although the precise nature of fluorine-hydrogen interactions is controversial (vide infra).
In cyclic amines, such as the protonated form of 3-fluoro-N-methyl-piperidine (21), these interactions provide significant conformational bias [53–55]. In 21, the ring F atom strongly prefers an axial disposition despite experiencing steric compression, with the calculated conformer populations provided in Fig. 6 [55]. As the global minimum, conformer A dominates to the extent of 95–96 %, stabilized by an electrostatic interaction between the F and NH+ moieties, whilst conformer D contributes 4–5 % at equilibrium with the unfavorable diaxial interactions compensated by a productive electrostatic effect.
Trans-4-fluoro-L-proline (15) and its cis- isomer 16 both adopt a defined conformation in aqueous solution in which the favored conformations project the F atom in an axial orientation that is gauche with respect to the amino moiety [56].
The two fluorinated enantiomers 23 and 24 of the neurotransmitter γ-aminobutyric acid (GABA, 22) represent an interesting example where the preferred gauche disposition between a protonated amine and a fluorine atom was used to glean insight into aspects of biologically-relevant conformations (Fig. 7) [57, 58]. The introduction of fluorine to GABA (22) reduces the basicity and increases the acidity of the natural amino acid in a fashion that preserves the zwitterionic nature at neutral pH. In solution, an extended conformation was observed for all molecules based on NMR analysis, stabilized in the case of 23 and 24 by the gauche interaction between the F and NH3 + moieties [57]. The range of conformations for each enantiomer are captured in Fig. 7 which also notes those that would be disfavored based on the absence of a gauche effect [57]. Both 23 and 24 were found to activate the cloned human GABAA receptor with similar potency but were markedly less potent than GABA (22) itself [57, 58].
This observation suggested that the bound form of the neurotransmitter was the extended form accessible to both 23 and 24 in a fashion that preserves the gauche relationship between F and NH3 +, represented by conformation B in Fig. 7 [57, 58]. However, GABA aminotransferase caused elimination of HF from the (R)-isomer 24 with 20-fold higher efficiency than for the (S)-isomer 23. This suggested that the conformation recognized by the enzyme placed the NH3 + and CH2CO2 − moieties in a gauche arrangement that would be destabilized, consistent with conformation C being preferred [57, 58] (Fig. 7). This conformation was also deduced to be that preferably recognized by the GABAC receptor [59]. Both 23 and 24 were found to act as agonists rather than antagonists at ρ1 and ρ2 GABAC receptors, with the (R)-isomer 24 the more potent but still ten-fold weaker than GABA (22), whilst the (S)-isomer 23 was 20-fold weaker than the natural ligand, attributed to weaker electrostatic interactions as a consequence of the reduced basicity. These observations suggested that GABA (22) adopted a folded binding orientation represented by conformation C at the GABAC receptor [59].
An analogous strategy was adopted to probe the bound conformation of N-methyl-D-aspartate (26), an agonist of the neurotransmitter glutamic acid (25), at recombinant GluN2A and GluN2B receptors expressed in Xenopus laevis oocytes (Fig. 8) [60]. (2S,3S)-3F-NMDA (27) was prepared in an enantiospecific fashion whilst (2S,3R)-3F-NMDA (28) was obtained along with the (2R,3S)- stereoisomer after separation of diastereomeric α-methylbenzylamine precursors. Analysis of 1H- and 19F NMR spectra indicated distinct chemical shifts and coupling constants for the individual compounds that were consistent with (2S,3S)-3F-NMDA (27) in solution adopting conformation A, shown in Fig. 8, whilst (2S,3R)-3F-NMDA (28) avoided conformer C and was concluded based on DFT calculations to prefer conformer B, the conformer observed in the single crystal X-ray structure. In the biological assays, (2S,3S)-3F-NMDA (27) evoked currents in oocytes expressing either GluN2A or GluN2B receptors, although the responses were less than that observed with NMDA (26), whilst (2S,3R)-3F-NMDA (28) and its enantiomer were essentially inactive [60]. These results suggested that only (2S,3S)-3F-NMDA (27) was able to adopt a conformation similar to that of bound NMDA (26), designated as A in Fig. 8. This result is consistent with X-ray crystallographic data of NMDA (26) bound to the GluN2D receptor which is highly homologous to the GluN2A and GluN2B receptors.
The effect of a fluorine atom β- to a protonated amine on conformational preference and its potential impact on biological recognition has further been documented in 8-membered rings and in fluorinated pyrrolidine elements incorporated to replace the peripheral pyrrolidines of G-quadruplex binding ligands based on an acridine scaffold [61, 62].
2.4 Fluorine-Hydroxyl and Fluorine-Alkoxy Interactions
The gauche conformation of β-fluoroethanol derivatives has been estimated to be favored by ~2 kcal mol−1 when the F and OH moieties can engage in an intramolecular H-bonding-like interaction [52, 63, 64]. In the absence of this interaction, the preference for the gauche relationship is essentially absent, with an energy estimated at only 0.1–0.2 kcal mol−1. However, protonation of the alcohol significantly enhances the stability of the gauche conformer, estimated at ~7 kcal mol−1 and attributed to a combination of both a stereoelectronic effect and an intramolecular HO-F interaction [52, 63, 64]. Several drug design studies have attempted to exploit the effects of deploying F and OH moieties in a vicinal arrangement that either take advantage of the gauche conformational preference or the effect of F on the H-bonding potential of the hydroxyl. An early example explored the effect of replacing the C-9 α-H atom of hydrocortisone acetate (29) with F to afford 30, which demonstrated ten-fold greater glucocorticoid activity in restoring liver glycogen deposition in the adrenalectomized rat and four- to nine-fold higher activity at increasing the survival time of adrenalectomized dogs with a concomitant reduction in sodium retention [65, 66]. Fludrocortisone (32) is an anti-inflammatory agent that improved on the therapeutic index of the progenitor hydrocortisone (31) and was the first marketed fluorinated pharmaceutical product [67]. The origin of the improved biological properties of fluorinated steroids remains enigmatic and may be a function of the combination of several effects that includes the increased acidity of the 11β-hydroxyl moiety, a distortion of the A ring due to non-bonded interactions between the F atom and proximal axial substituents and reduced metabolism in vivo [14, 68–70].
The judicious introduction of a fluorine atom into the HIV-1 protease inhibitor indinavir (33) and epi-indinavir (34) provided an opportunity to assess the effect of F-OH interactions on biological properties related to conformational disposition [71]. The four fluorinated derivatives 35–38 were prepared with the anticipation of complex effects due to the F atom which was expected to affect the acidity of the OH and influence the population of conformers in addition to affecting steric interactions and solvation. The Ki data for the inhibition of HIV-1 protease are summarized in Table 2 and reveal interesting profiles.
The syn, syn fluorinated derivative 35 fully retained the activity of the parent drug 33 while the anti,anti isomer 37 was an order of magnitude weaker. For the epi-indinavir derivatives 36 and 38, the syn,anti analogue 38 improved potency by eight-fold over 34 whilst the anti,syn diastereomer exhibited a 36-fold weaker Ki [71]. An analysis of the 1H-1H and 1H-19F coupling constants indicated that the syn,syn (35) and syn,anti (38) diastereomers adopted the fully extended conformation that characterizes indinavir (33) when bound to HIV-1 protease, attributed to the preference of F and OH to adopt a gauche relationship. In contrast, the other 2 diastereomers 36 and 37 populated additional conformers to a significant extent, providing an explanation for the observed reductions in potency. The lipophilicity of each of the diastereomers was determined from the measured 1-octanol/water partition coefficient and all were marginally increased compared to indinavir (33), with the exception of the anti,anti analogue 37.
Fluorination also impinges on the conformation of ethers, most prominently in anisoles [15, 72–74]. Arylmethyl ethers without ortho substituents adopt a conformation in which the methoxy moiety is close to coplanar with the aryl ring, a consequence of rehybridization of the substituent that maximizes electronic overlap of the oxygen atom with the π system and overcomes the inherent allylic 1,3-strain, with bond angles summarized in Fig. 9 [75–79]. In contrast, the trifluoromethoxy substituent is typically orthogonal to the plane of the benzene ring, favored by 0.5 kcal mol−1, inverting the calculated 3 kcal mol−1 preference for planarity observed with anisole [15, 80]. Difluoralkoxyaryl moieties behave similarly, indicating that these structural elements should not be considered as simple isosteres of alkoxy substituents but as moieties capable of exploring a broader range of conformational space [15, 74].
The arylethyl ether 39 is an inhibitor of cholesterol ester transfer protein, IC50 = 1.6 μM, but the fluorinated homologue 40 is eight-fold more potent, IC50 = 0.2 μM, attributed to a combination of steric effects and drug-protein interactions that are more favorable in 40 [73].
Ab initio calculations indicated that while the ethoxy moiety preferred a conformation coplanar with the aryl ring, the fluorinated ether projected in a vector that was more perpendicular. Interestingly, a 2-furyl substituent provided the optimal OCF2CHF2 mimetic, IC50 = 0.48 μM, attributed to the preferred conformation of the furan ring also being orthogonal to the plane of the phenyl moiety [73].
3 Modulation of pK a with Fluorine
Fluorine is the most electronegative element (Pauling electronegativity of 3.98) [84]. Due to the significant electron-withdrawing influence that fluorine exerts inductively through σ-bonds, fluorine substituents typically lower the pK a of neighboring functionality (although exceptions exist [81]): protic groups become more acidic as measured by a decrease in the pK a of the functional group itself, and basic groups become less basic as measured by a decrease in the pK a of the conjugate acid. The ionization state of basic and acidic functionality not only affects potency and the absorption, distribution, metabolism, and excretion (ADME) properties of a molecule, but may also play a role in reducing the potential for a compound to exhibit toxicity [82]. Specifically, molecules incorporating basic nitrogen atoms are known to be more promiscuous and may show increased potential for cardiovascular toxicity via promoting interference with the hERG potassium ion channel and may increase the occurrence of phospholipidosis [83–86]. The significant influence of fluorine on the pK a of adjacent functionality, in combination with other desirable properties including a small steric volume (similar in volume to oxygen) and high metabolic stability of the C-F bond, provides a powerful tool by which to modulate pK a during drug candidate optimization [14].
3.1 pKa Trends with Fluorine Substitution
An understanding of the correlation between fluorine substitution and pK a enables a rational approach to drug candidate optimization. For acyclic aliphatic amines, the observed pK a often represents the contribution of a conformational average of substituents and, as a consequence, the decrease in pK a due to fluorine substitution is generally additive [87]. As shown in Fig. 10, the shift in pK a of aliphatic amines due to fluorine substitution can be accurately predicted based on this additive property. The significance of the effect decreases with increased σ-transmission distance: each fluorine substituent at the β-carbon reduces the pK a by 1.7 units, at the γ-carbon by 0.7, and at the δ-carbon by 0.3. As a point of reference for the calculation, the pK a of alkyl amines can be generally approximated as primary amine = 10.7; secondary amine = 11.1; and tertiary amine = 10.9 [87, 88]. The chain length of the alkyl groups does not generally shift the pK a of the amine; however, N-methyl substitution is an exception (Fig. 11) [87].
For unstrained cyclic amines in saturated systems the change in pK a upon substitution can be approximated by summing the contribution of both σ-transmission pathways [87, 88]. Figure 12 demonstrates the application of this approach towards a series of substituted piperidines. For example, in 3-fluoropiperidine there are two σ-transmission pathways from the fluorine atom to the nitrogen atom: “path 1”=F-C-C-C-C-NH and “path 2”=F-C-C-NH. The pK a of 3-fluoropiperidine can therefore be predicted using the values of the linear system provided in Fig. 10: pK a = 11.1 (piperidine) – 0.3 (“path 1”) – 1.7 (“path 2”) = 9.1. This predicted value is in close agreement with the experimental pK a of 9.3.
That the observed pK a of 3-fluoropiperidine is slightly less acidic than predicted may be explained by stabilization of the ammonium species via dipole minimization, as depicted in Fig. 13a [88]. In fact, NMR studies with 3-fluoropiperidine have revealed that the 3-fluoro substituent inverts from a completely equatorial orientation to a completely axial orientation upon protonation of the amine [53]. Similarly, the fluoro substituent of cis-3-fluoro-4-benzylpiperidine adopts a completely axial orientation in the protonated species leading to stabilization and a pK a of 9.3 (Fig. 13b). Upon protonation of trans-3-fluoro-4-benzylpiperidine, the 3-fluoro substituent occupies both the axial and equatorial orientation equally, leading to reduced stabilization and a pK a of 8.8 [88].
The stabilization of ammonium species via syn-β-fluoro substituents appears to be a general phenomenon. The ΔpK a shift due to β-fluoro substitution can be refined to accommodate for this influence: ΔpK a = −1.4 for a full 1,3-syn CF-+NH interaction and ΔpK a = −2.0 in the complete absence of a 1,3-syn CF-+NH interaction [88]. This refinement enables the accurate prediction of pK a values for primary, secondary and tertiary amines with β-fluoro substituents (Fig. 14). Additionally, the syn-β-fluoro effect may, in part, account for the difficulty in predicting the pK a of imperfectly staggered heterocycles such as fluoro-substituted pyrrolidines [88].
Fluorine substitution of methanesulfonamides also leads to an additive decrease in the pK a of the NH proton. As shown in Table 3, the decrease in pK a is approximately 1.47 units per fluorine substituent and is independent of the R-group bound to nitrogen [90]. The linearity observed in this trend is postulated to be due to the small steric changes associated with replacing hydrogen by fluorine and the fact that fluorine substitution is somewhat remote (two atoms away from the acidic site) [90]. In the case of fluoroacetic acids, the decrease in pK a due to fluorine substitution is not additive: the pK a of acetic acid = 4.76; fluoroacetic acid = 2.59; difluoroacetic acid = 1.33; and trifluoroacetic acid = 0.50 [91].
Although fluorine exerts a strong electron-withdrawing influence through σ-bonds, fluorine is electron-donating through resonance (Fig. 15a). Due to the excellent 2p orbital overlap between fluorine and carbon, this influence can be quite strong and the significance of this effect is not matched by the other halogens (Fig. 15b) [14, 92]. Due to these opposing forces, the effect of fluorine on pK a in aromatic systems is strongly dependent on the through-resonance relationship of the fluorine substituent and the functional group of interest. Fluorine acts as an electron-withdrawing group in positions ortho- or meta- to functionality. Fluorine in a para- relationship to a functional group behaves as a neutral substituent because the through-resonance electron-donating character of fluorine essentially cancels out its inductive electron-withdrawing influence [92]. As shown in Table 4, this trend is observed with benzoic acids and phenols, archetypical systems of cross-conjugated and directly-bound functionality, respectively. Similarly, the changes in pK a of aniline (pK a = 4.62) upon fluorine substitution is: 2-fluoroaniline = −1.40, 3-fluoroaniline = −1.04 and 4-fluoroaniline = +0.03 [17]. Basic heterocycles such as pyridine also follow this trend (Fig. 16). Although the aqueous pK a of 4-fluoropyridine has not been reported, gas phase measurements indicate that 4-fluoro substitution leads to a small decrease in the pK a in the gas phase and a predicted aqueous pK a similar to that of pyridine [93]. It should be noted that the CF3 group is electron-withdrawing both inductively and through-resonance, and thus always reduces the pK a of adjacent functionality [93, 94].
4 The Effect of Fluorine Atoms on Drug Potency
Appropriate modulation of pK a, conformational preference, hydrophobic interaction, hydrogen bonding, lipophilicity, solubility or a combination of these properties exerted by the deployment of fluorine atoms can improve potency/binding affinity and sometimes even selectivity of drug molecules. This phenomenon is manifested in many examples taken from a wide range of areas of medicinal chemistry.
In keeping with the theme from the previous section, the influence of fluorine on the pK a of basic functionality was systematically studied in the context of optimizing the pharmaceutical properties of a thrombin inhibitor [95]. All permutations of single and double fluorine substitution of the amidinium-bearing arene were synthesized and the pK a of the amino group of the tricyclic skeleton (pK a1) and the pK a of the amidinium residue (pK a2) were experimentally determined (Fig. 17). The change in pK a upon introduction of fluorine followed the expected trends, with the exception of double fluorine substitution meta- to the amidinium group, compound 45. In this case, the pK a unexpectedly increased to 4.21 from 4.10 measured for the mono-fluoro analog 44 and insight into this phenomenon was sought through DFT calculations (RB3LYP, 6-31G**). For the mono-fluorinated arene, the exo-conformer, which minimizes allylic-1,3-like strain between the nitrogen lone pair and fluorine, is calculated to be 3.4 kcal mol−1 more stable than the endo-conformer (Fig. 18a). In the difluorinated arene, unfavorable allylic strain is unavoidable, as illustrated in Fig. 18b. Upon protonation, both the mono- and di-fluorinated arenes adopt conformations in which there is a favorable C-F…H-N+ interaction. However, because this favorable interaction is present in both analogs, the increased basicity of the difluoro analog is best explained by the differences observed in the unprotonated state, i.e., the electrostatic interaction between the C-F dipole and the nitrogen lone pair [91]. In all cases the decrease in pK a corresponded to a decrease in binding affinity (Ki). Although the change in binding affinity is likely complex in origin, linear free energy relationships correlating pK a and binding free enthalpies (−ΔG) calculated for inhibition of thrombin and trypsin demonstrated a relationship between affinity and pK a. Specifically, this analysis found that inhibition of thrombin is 2.5 times more sensitive to the pK a of the amidine functionality than is inhibition of trypsin, providing a possible means towards increasing thrombin selectivity [95].
Peroxisome proliferator-activated receptors α, δ and γ (PPARα, PPARδ and PPARγ) are implicated in regulating transcription of genes responsible for lipid and carbohydrate metabolism [96]. A series of fluorine-substituted analogs of the PPARδ modulator GW-501516 (48) was synthesized and studied based on the hypothesis that fluorine substitution at the α-carbon (49) would reduce the pK a of the adjacent carboxylic acid moiety and lead to stronger binding affinity (Fig. 19) [97]. In practice, the fluoro-substituted analog 49 was a less potent activator of oleic acid oxidation than the progenitor GW-501516 (48). However, when the affinity of each compound towards the PPAR subtypes was assayed, it was found that the fluoro-analog demonstrated a modestly positive effect on PPARα activation, essentially behaving as a modest dual PPARα/δ agonist. Computational modeling predicted that both enantiomers of the fluoro-analog bound similarly, and that the dual agonism is due to interactions remote from the fluorine atom.
Judicious fluorine substitution can provide a means of improving the solubility of drug molecules. The solubility of several classes of pharmaceutical compounds can be improved by disrupting molecular planarity and, in turn, potentially reduce crystal packing density [98]. Although this argument explains, in part, the observed improvements in solubility, fluorine substitution leading to increased solubility is explained “not only by disrupting molecular planarity but also via other mechanism(s)” [98]. In these cases, the trends observed in Fig. 20 can be understood as a manifestation of the inductively withdrawing/resonance-donating electronic effects of fluorine that underlie its influence on pK a. For example, fluorine substitution ortho and meta to a carboxylic acids leads to an increase in acidity which can consequently increase solubility. Fluorine is a strong through-resonance donor (vide supra) which can impart partial π-bond character to biphenyl systems, leading to a decrease in dihedral angle, but this effect is reduced if there is a strong resonance-accepting group such as a carboxylic acid ortho or para to a fluorine atom.
This through-resonance effect of a fluorine atom is also manifested in the properties observed with the series of fluorine-substituted aryl hydrocarbon receptor (AhR) agonists depicted in Fig. 21 [98]. The decrease in dihedral angle is most dramatic for the mono-fluoro analog 51 where solubility is increased. Both of these changes are likely due to an increase in polarization resulting from strong through-resonance communication between the fluorine substituent and the ketone functionality. In the case of the difluoro derivative 52, both the dihedral angle and solubility are similar to the mono-methyl substituted compound 53, suggesting that disrupting planarity may be the underlying influence. Fortuitously, fluorine substitution also led to an increase in potency as well as solubility.
An example of fluorine substitution in the context of the cysteine protease cathepsin K highlights not only the advantageous impact of fluorine on potency but also the isosteric relationship between a trifluoroethylamine moiety and an amide [99]. The replacement of amide bonds with isosteric functionality has been extensively explored within the field of medicinal chemistry and there are many well-established isosteres of the amide moiety. However, there are only a few isosteres that conserve the hydrogen-bond donating properties of the amide. Reduction of an amide to the corresponding amine not only removes a hydrogen-bond acceptor but also introduces a basic amine, typically providing poor mimetic properties. Addressing this deficiency, the strategic introduction of a trifluoromethyl moiety led to reduced basicity whilst preserving the near 120° backbone angle and conferring isopolarity with the carbonyl moiety, as illustrated by 55 in Fig. 22 [99, 100]. In addition to providing an acceptable isostere for the bis-amide 56, trifluoroethylamine 55 also exhibited improved potency and selectivity compared to 56. The trifluoroethylamine derivative 55 was the most potent compound identified, demonstrating an IC50 of less than 0.005 nM toward inhibition of cathepsin K. Modeling and SAR were used to rationalize the stark difference in potency between 55 and 56 [99]. Specifically, in 55 the sp2 nature of the amide leads to an interaction with the adjacent aromatic ring which enforces a suboptimal geometry for both the aromatic ring and for the hydrogen-bonding interaction to the protease. In the case of 55, the newly introduced sp3 center allows for a perpendicular arrangement of the aromatic ring thus stabilizing it in what is presumed to be the active conformation while simultaneously allowing for a more optimal hydrogen bond interaction with a residue in the active site. In addition to improved potency and selectivity with 55, the metabolic stability was also increased significantly: 55 demonstrated no susceptibility toward hydrolysis, a significant liability observed with dipeptide compounds.
The SF5 substituent has not been extensively exploited in drug design despite that this group has been referred to as a “super trifluoromethyl group” due to the fact that these two functionalities share many features (although this designation is probably imprecise) [101–104]. The shared features between SF5 and CF3 include high electronegativity, good thermal and chemical stability and high lipophilicity, with the SF5 group exhibiting each of these features to a greater extent than the CF3 group. Both of these substituents are xenobiotic and have remarkable stability under physiological conditions, lending to their usefulness in drug discovery. While there is some consensus on CF3 being similar in size (van der Waals volume) to the ethyl group, comparative comments on the relative size of an SF5 moiety are more speculative [16]. However, it has been suggested that the volume of the SF5 group is slightly less than that of a tert-butyl group but larger than the CF3 group [102]. In addition to differences in size, the geometries of these two groups are quite different. The SF5 group displays a pyramid of electron density whereas the CF3 group is associated with a cone shape of electron density [103]. The subtle differences in properties between these two groups provides opportunity for deployment of the SF5 group in a fashion that allows optimization of many of the PK and PD properties discussed in this chapter, including potency. An interesting example where an SF5 moiety provided an advantageous impact on potency is in the context of cannabinoid CB1 receptor ligands [104]. In two different pyrazole series illustrated by the prototype meta-CF3-aniline 57 and its SF5 analogue 58 (Fig. 23), the SF5 compounds consistently exhibited a lower K i for the CB1 receptor than the CF3-analogs. The increased potency was attributed to a potential “better fit” of the SF5 group in a pocket of the CB1 receptor relative to the CF3 group, indicating that in this setting the two groups are not biologically equivalent.
Another example where SF5 and CF3 substitution has been examined in a matched pairs-type of analysis was disclosed in the context of a P. falciparum dihydroorotate dehydrogenase (PfDHODH) inhibitor program [105]. The C-4 SF5 analogue 59 (Table 5) provided a two- to three-fold potency improvement over the CF3 analogue 60 as an inhibitor of PfDHODH. It was suggested that the enhanced potency may be due to the increased hydrophobicity of the SF5 substituent. In addition to increased potency, the SF5 analogue 59 exhibited good metabolic stability, promising PK parameters in vivo and no apparent liabilities. The 3,5-di-F-analog 61 also provided an interesting illustration of the effect of fluorine substitution. This compound compared well with 60 in many respects but showed significantly improved efficacy in vivo. These data illustrate again that the SF5 group can be a suitable substitute for CF3 with advantage in an optimization effort.
5 Effects of Fluorine on Metabolic Stability
Metabolic stability is a perpetual concern in drug discovery and there are many factors that govern the metabolism of a given compound. Perturbation of properties such as lipophilicity, polarity, sterics, and conformation can often reduce metabolic liabilities. When modifying a compound, changes must be carefully balanced such that the solution to one problem does not introduce another. One strategy to solving a metabolic stability issue is to identify the metabolic “soft spot”, i.e. the site of oxidation, and then block that site with a substituent resistant to metabolism. One way to accomplish this is to replace a labile/susceptible carbon-bound hydrogen with a fluorine atom. This type of substitution infrequently impairs drug-target interactions and is a strategy that has been employed successfully in many contexts [106–110].
One notable example is the discovery of ezetimibe (63) (Fig. 24), a marketed β-lactam derivative that lowers systemic cholesterol levels by blocking intestinal absorption of cholesterol [108, 109]. A key experiment led to the identification of active metabolites of the prototype compound 62 after dosing to rats [108], which in turn led to the discovery that metabolites were more efficiently localized to the putative site of action of the drug than 62 itself. Additional studies elucidated the structure of several metabolites, some of which offered advantageous localization to the intestine while others did not [111]. Armed with this knowledge, a second generation compound was designed in which sites of detrimental metabolism were blocked via strategic incorporation of substituents. The result was ezetimibe (63), a compound which contains hydroxyl functionality crucial to activity and selectivity, as well as fluorine atoms in two key locations which serve to blocked detrimental metabolic pathways without adversely affecting cholesterol absorption inhibitory activity significantly.
The metabolism of heterocycles which leads to reduced potency, poor PK properties or bioactivation into chemically reactive species is a particular concern in drug design, and strategies to mitigate these pathways have been reviewed and summarized [112]. As demonstrated by the discovery of ezetimibe (63), if the site of metabolism is known or even presumed, a useful and common strategy is to simply block that position with a fluorine atom. In a thorough effort towards improving metabolic stability in a series of piperidine-based 11β-hydroxysteroid dehydrogenase type I (11β-HSD1) inhibitors, it was discovered that introducing fluorine or polar substituents led to an increase in metabolic stability [113]. It can be reasoned that polar groups lowered the lipophilicity and log D, likely contributing to the increase in metabolic stability. Moreover, in a modification that presumably blocked a metabolic soft spot present in 64 (Fig. 25), installation of a para-fluoro substituent (65) had a positive effect on metabolic stability in mouse liver microsomes (MLM); a MLM T1/2 of >30 min was demonstrated which compared favorably with the unsubstituted piperidine analogue 64, which exhibited a T1/2 of just 6 min.
As suggested above, high lipophilicity can have an adverse impact on metabolic stability. In an effort to improve the metabolic stability of the lead histamine 3 receptor (H3R) inverse agonist 66 (Fig. 26), ring contraction of the azepine ring coupled with the introduction of fluorine substitution was explored [114]. While there was some confidence that this strategy would lead to increased metabolic stability, there was concern that the structural changes would adversely affect solubility. In the event, the smaller ring size associated with piperidines 67 and 68 resulted in more stable molecules with rat liver microsome (RLM) clearance values of 17 and 9 μL/min/mg, respectively, as compared to an RLM clearance of 55 μL/min/mg for 66. However, the structural changes were accompanied by some adverse affects, with 68 inhibiting the hERG channel by 73 % at a concentration of 10 μM. Attention was thus shifted toward the morpholine analogue 69 which had reasonable metabolic stability and did not share the hERG liability.
While these examples demonstrate the beneficial effect of blocking metabolism by taking advantage of the inherent strength of a carbon-fluorine bond, in extreme cases this property may be disadvantageous. An example of such a situation was observed in the discovery of SC-58635 (70), a cyclooxygenase 2 (COX 2) inhibitor known generically as celecoxib (Fig. 27). The early lead molecule 71 possessed a 4-fluorophenyl moiety that rendered the molecule extremely metabolically stable in vivo, represented by the greater than 220 h half life measured in rats [115]. Interestingly, a similar high level of metabolic stability was observed with the chloro analog 72 which had a measured half life of 117 h in rats. In an effort to bring this value into a more acceptable range, the introduction of functionality with greater susceptibility to metabolism was examined. As such, the methoxy analog 73 and the methyl derivative 70 were prepared with both compounds exhibiting a shorter half life in the rat: 5.6 and 3.5 h, respectively.
There are several interesting cases where the introduction of a fluorine atom does not prevent metabolic oxidation at the site of deployment [15]. A specific example is observed in phenyl rings that incorporate a nitrogen substituent para to the fluorine atom. In this arrangement, P450-catalyzed oxidation can facilitate fluorine rearrangement to the adjacent carbon atom resulting in a hydroxyl moiety being installed on the carbon atom vacated by the fluorine [116–118].
6 The Effect of Fluorine Substitution on Membrane Permeability
Absorption of an orally administered drug is governed primarily by one of two processes: passive transport, or active transport which requires energy that is typically supplied by ATP. The majority of drugs permeate membranes passively, the ease of which depends on the properties of both the membrane and the diffusing molecule. Membranes are a highly organized, anisotropic system with sufficient fluidity to allow translational and rotational movements of the constituent lipid and protein molecules; the packing density of the constituent elements influences binding to and permeation of drugs [119–121]. Lipophilicity and the cross sectional area of a drug are the two most influential properties governing membrane binding [121]. While the ability of fluorine to influence permeability may diminish with the size of a molecule, it is well suited for use in modulating lipophilicity in the context of small molecule drug discovery.
Lipophilicity is often expressed as log P, the logarithm of the partition coefficient of a compound between octanol and water. A distribution coefficient (log D) is also used to quantify lipophilicity in the event that charge states need to be taken into consideration. A log D value is the logarithm of the coefficient of the distribution of a molecule between water and octanol at a particular pH, typically 7.4 because of physiological relevance. Based on an analysis of the physical properties of orally bioavailable drugs, the optimal log P for an orally administered drug is between 1 and 5 [122].
It is a misconception to assume that fluorination of a molecule always increases lipophilicity. Monofluorination or trifluorination of saturated alkyl groups often decreases lipophilicity due to the relatively polar character of the monfluoro- and trifluoro-methyl alkanes which possess highly polar C-F and C-CF3 bonds. The same is true for difluorination when fluorine is introduced at the terminal carbon atom of an alkane (Table 6) [123]. In contrast to the data shown in Table 6, lipophilicity is typically increased upon aromatic fluorination, perfluorination, and fluorination adjacent to atoms with π-bonds, with the notable exception of some α-fluorinated carbonyl compounds [123]. Presumably, in these cases the lipophilicity is increased due to the excellent overlap between the F-C 2p orbitals which results in significant resonance electron-donation from fluorine to carbon that offsets the inductive electron-withdrawing influence of fluorine.
The situation becomes far less predictable when a heteroatom is proximal to the site of fluorine substitution. As indicated in Table 6, fluorination has little effect on lipophilicity when the site of fluorination is less than 3C-C bonds away, at least in the case in the case of terminal trifluoromethylated alcohols [123].
The impact of fluorine substitution on lipophilicity (as measured by log P) has been studied in the context of leukotriene (LT) receptor antagonists [124]. Leukotrienes are signalling molecules responsible for triggering constriction of smooth muscle, and the overproduction of leukotrienes is implicated in contributing to asthma in humans. The development of LT receptor antagonists is one approach towards moderating the bronchoconstrictive properties of leukotrienes. For a series of indole carboxamide-based LT antagonists represented by 74 (Fig. 28), an increase in lipophilicity via the careful installation of a trifluoromethyl group led to an increase in efficacy and oral availability. Extending the length of the alkyl chain of the indole amide led to increased lipophilicity but was also associated with a loss in affinity for the LT receptor. Fluorine substitution increased lipophilicity and, in comparison to non-fluorinated analogs, increased in vivo potency ten-fold. In a head-to-head comparison, the fluorinated analogs consistently exhibited greater lipophilicity than the non-fluorinated analog as measured by log P (Fig. 28).
The influence of fluorine substitution on permeability, as measured across a Caco-2 bilayer, has been studied in a series of factor Xa (fXa) inhibitors [125, 126]. It was observed during the optimization of the series of pyrazole-based fXa inhibitors 75–78 shown in Fig. 29 that the substitution of the aromatic ring with fluorine afforded a greater than two-fold increase in Caco-2 permeability (75 vs. 76). Similarly, the exhaustive introduction of fluoro to the methyl substituent bound to the aromatic ring (CH3 to CF3) resulted in a similar improvement in Caco-2 permeability (77 vs 75). Further, when both the aromatic ring and the methyl substituent were substituted with fluorine (78) the Caco-2 measured permeability increased to 4.86 × 10−6 cm/s, a >four-fold improvement over that non-fluoro analog 75. Furthermore, the increased permeability of 78 was complemented by the best in vivo potency within the series and 78 also exhibited a high selectivity ratio for fXa versus thrombin (1,000-fold) and fXa versus trypsin (300-fold).
The advantageous impact of fluorine on permeability has been explored in the discovery of a family of imidazolo[1,2-b]pyridazine-based IKKβ inhibitors where it was assumed that the central benzamide played a key role in pharmacokinetic properties [127]. As shown in Fig. 30, modifications to the benzamide region (C-3) of the imidazo[1,2-b]pyridazine had a dramatic affect on permeability: the measured PAMPA values for 79 and 80 were 0.84 × 10−6 cm/s and 19 × 10−6 cm/s, respectively [128]. Although 79 and 80 differ by only a single F-for-H substitution, the observed improvement in permeability is presumed to arise from a combination of factors. These factors include the enhancement of hydrophobicity and the formation of a weak interaction between the fluorine atom and the NH moiety, possibly a dipolar interaction. While the inhibitory activity and log D of 79 and 80 remained essentially identical, the installation of a second fluorine on the pyrrolidine moiety (81) resulted in an increase in both the measured log D and PAMPA values. In the case of 81, the added lipophilicity coupled with the small difference in size between H and F also afforded a two-fold increase in potency compared to 79 and 80.
6.1 Inter- and Intra-molecular H-Bonding of Fluorine
Numerous X-ray crystallographic structures of fluorine-containing compounds demonstrate close inter- and intra-molecular C-F⋯H contacts and the strong hydrogen-bonding capacity of hydrofluoric acid has long been recognized [30, 129–132]. The intriguing possibility that optimization of C-F⋯H contacts might lead to an increase in the potency of pharmaceutical agents has, in part, contributed to a significant interest in determining the capacity of fluorine to act as a hydrogen-bond acceptor. Even though the nature of this interaction remains a topic of considerable debate, it is recognized that the C-F⋯H interaction is much weaker than that of O⋯H hydrogen-bonding, and that interpretation of possible C-F⋯H interactions requires care and attention to multiple factors [20, 129, 133–138].
Due to the strong electronegativity of fluorine, the three lone pairs of electrons are tightly held, resulting in fluorine being a poorly polarizable atom [16, 20]. However, the substantial ionic nature of the C-F bond gives rise to a large dipole moment (μ). For example, in fluoromethane the dipole moment is 1.85 D while in difluoromethane it is 1.97 D [20]. In systems such as 3-fluoropiperidinium, dipole minimization rather than C-F⋯H bonding has been suggested to best explain the observed conformational bias [55]. The C-F⋯H interaction is variously described as a hydrogen-bonding or a dipolar interaction. The distinction is that a hydrogen-bonding interaction has a covalent component which is anisotropic, while a dipolar interaction is entirely electrostatic and not dependent on directionality. A recent analysis of C-F⋯H contacts documented in the Cambridge Structural Database (CSD) supports the absence of an anisotropic component in the C-F⋯H interaction. Specifically, in a survey of 100 compounds with C-F⋯H bond distances between 1.85 Å and 2.35 Å, the C-F⋯H bond angle was found to vary significantly between 95° and 165°, while demonstrating no dependence on F⋯H distance [30]. The absence of a directional dependence of C-F⋯H interactions is in agreement with an earlier analysis of the Protein Data Bank (PDB) [129]. The C-F⋯H interaction is therefore best described as a dipolar interaction.
The presence of inter- and intra-molecular C-F⋯H contacts is mostly supported by crystallographic data. Often a F⋯H distance shorter than the sum of the F and H van der Waals radii (~2.65 Å) is used as evidence for a C-F⋯H interaction [129]. A review of the PDB in 2004 using this distance criterion as well as directional constraints found that “…there is an 18% chance that a generic PDB entry with a co-crystallized fluorine-containing ligand presents a hydrogen-bond occurring at the fluorine” and that “10% of the overall amount of fluorine atoms from the PDB are involved in hydrogen bonding contacts” [129]. The mean F⋯H distances calculated for aliphatic, aromatic, and geminal fluorines (CF2H, CF3) are 2.113 Å, 2.698 Å, and 2.350 Å, respectively. When standard deviation values are considered, the differences between the three types of C-F bond are not statistically significant. Instead, a calculated mean F⋯H distance of 2.313 Å is more appropriate [129]. As a comparison, almost all H⋯O and H⋯N contacts observed crystallographically are less than 2.2 Å in length [138]. It is therefore suggested that a bond distance much shorter than 2.65 Å is often needed for a productive F⋯H interaction, and that this distance criterion is not a reliable indicator of F⋯H interactions [137].
Many studies have attempted to provide non-crystallographic evidence for a C-F⋯H interaction. For example, the magnitude of a 1hJF,H(O) NMR coupling constant has been used as evidence for a C-F⋯H interaction in 2′-fluoroflavonol (Fig. 31a) [139]. However, this analysis targets a minor conformer of 2′-fluoroflavonol and relies upon several assumptions regarding electronics. The same authors report that the 1hJF,H(O) coupling constants found for 2-fluorophenol and 4-bromo-2-fluorophenol arise from overlap of the electronic clouds surrounding both coupling nuclei (owing to their spatial proximity) rather than from transmission through a hydrogen bond [140].
In another study, the nature of intramolecular C-F⋯H-C bonding was interrogated using O-alkyl 9-dehydro-9-trifluoromethyl-9-epiquinidines as model systems (Fig. 31b) [134]. These compounds demonstrate many phenomena consistent with C-F⋯H-C bonding: X-ray crystal structures contain close F⋯H contacts, as low as 2.19 Å when R=Me; 19F NMR spectra contain three distinct resonances with first order 2 J F-F coupling indicating that the rotation of the CF3 group is greatly restricted; and noticeable deshielding is observed in 1H NMR for the proton most proximal to the CF3 group. A combination of crystallographic and NMR observations with Erying analysis and extensive computational analysis were used to probe the nature of this C-F⋯H-C interaction [134]. Despite the strong appearance of a C-F⋯H-C bonding interaction, it was concluded that steric crowding is responsible for the short F⋯H distances and restricted rotation; even the strongest C-F⋯H-C bonding interactions in the system contribute minimally to the ground state geometry or hinder rotation [134].
In contrast to C-F⋯H-C bonding, the C-F⋯H-N+R3 interaction is considered one of the strongest C-F⋯H bonding interactions. As a tool to further study the C-F⋯H-N+R3 interaction, 8-fluoro-N,N-dimethylnapthalen-1-amine, the fluoro analog of 1,8-bis(dimethylamino)naphthalene (“Proton Sponge”), was prepared and studied (Fig. 31c) [133]. The basicity of this compound is slightly greater (<1 pK a unit) than that of the parent compound N,N-dimethylnaphthalen-1-amine. Crystal structures of the protonated molecule with either a TfO− or B(C6F5)4 − counterion indicate an intramolecular C-F⋯H-N+R3 bonding interaction, with F⋯H distances of 2.131 Å and 2.027 Å, respectively. In both cases, hydrogen-bonding is observed between the proton and the counterion with significance equal to that of the C-F⋯H-N+R3 contact. Unlike Proton Sponge, the H-N+R3 proton is observed crystallographically to lie outside of the naphthalene plane, with deviations from planarity of 39.5° and 29.5° for TfO− and B(C6F5)4 − counter ions, respectively. Observation in 19F NMR of a large J F,H coupling constant is cited as further support of the C-F⋯H-N+R3 bonding interaction. Interestingly, the J F,H coupling observed with a B(C6F5)4 − counterion is suppressed if TfO− or Cl− are instead used as the counterion, or if the solvent for the B(C6F5)4 − counterion is switched from CH2Cl2 to acetonitrile. Taken together, the evidence points towards the existence of a weak C-F⋯H-N+R3 interaction which, in terms of significance, is of approximately equal in magnitude to that of the TfO−⋯H interaction.
The large number of C-F⋯H contacts observed in crystallographic databases may mislead into an overly optimistic assessment of the importance of this interaction. Care must be taken in attributing the interaction to C-F⋯H bonding rather than to other conformational or intramolecular influences. Where the strongest isolated C-F⋯H interaction, C-F⋯H-N+R3, is in fact relatively weak, less significant C-F⋯H interactions become negligible. However, even weak interactions can play a supporting role in drug-target interactions and, as the prevalence of C-F⋯H contacts observed crystallographically suggests, the significance of the C-F⋯H interaction may be amplified in specific environments such as within the cavity of a protein. That the potency of a compound can be increased via fluorine substitution is clear, but the contributing factors are likely complex in origin.
7 Fluorine in Positron Emitting Tomography (PET) Imaging
PET imaging has emerged as a useful in vivo imaging technique that is non-invasive in nature and is useful in both a pre-clinical and clinical setting. Positrons are antimatter to electrons and encounters between the two particles lead to annihilation and the release of energy in the form of two photons of light. These photons are of high energy (511 keV), travel in opposing directions, and are readily detected simultaneously by PET cameras surrounding the subject [141, 142]. Of particular utility, the coupling of a PET camera with a computed tomography (CT) scanner allows both anatomy and metabolism data to be combined into a unique image [141, 142]. Molecules containing PET atoms are proving to be useful for probing biochemical aspects of disease in vivo and drug target engagement [141–146]. The utility of the 18F isotope in PET imaging relies upon its 110 min decay half life, the longest of all of the short-lived positron-emitting radionuclides with applicability in small molecule imaging (Table 7), and the facility with which drug molecules will accommodate fluorine as a substitute for a hydrogen atom whilst preserving fundamental biological properties [141–146]. However, this aspect of biochemistry must be determined by careful experimentation, particularly in the context of allosteric modulators where structure-activity relationships can be subtle and small structural changes can invert pharmacological properties [147–150]. In addition, the energy of the positron emitted by 18F is such that it travels only short distances before the annihilation event, effectively localizing the signal within a specific tissue. The utility of 18F in PET imaging is further amplified by the frequency with which fluorine atoms are being incorporated into contemporary drug candidates and marketed drugs as medicinal chemists seek to take advantage of the unique and sometimes enigmatic properties that can be conferred by this atom [15–19, 67]. Recent estimates suggest that 20 % of the drugs in the current pharmacopeia and 9 of the 30 best selling small molecule drugs (vida supra) in 2009 contain fluorine, providing an opportunity to introduce the 18F isotope in a fashion that would be expected to exert no demonstrable effect on the biological profile of a prototype molecule [17, 67].
Against this backdrop, 18F derivatives have emerged as important tools in biomedical imaging, with broad application essentially limited only by the methodology to introduce this isotope into molecules in a convenient fashion at the very last stages of compound synthesis and under mild conditions that preserve molecular integrity [151–158]. Production of 18F requires a cyclotron in which, most commonly, 18O-enriched water is bombarded with accelerated protons to produce [18F] fluoride ions ([18F]F−) which are trapped on an ion exchange resin for the purpose of separation and released by using a carbonate or bicarbonate solution containing the appropriate metal counter ion. However, fluoride ion is poorly nucleophilic but quite basic, a problem solved by elution with a solution of 2,2,2-cryptand and K2CO3 in CH3CN which, after drying, provides a complex of [(crypt-222)K]+18F− in which chelation of the potassium ion increases fluoride reactivity [141]. The highly electrophilic [18F]-fluorine gas ([18F]F2) can also be prepared in a cyclotron by bombarding natural neon with high energy deuterons or, preferably, 18O2 with protons to provide alternative reactivity pathways with which to introduce 18F.
2-[18F]-Fluoro-2-deoxy-D-glucose ([18F]-FDG, 82) has found widespread application for assessing the metabolic status of heart, lungs and brain and, particularly, for imaging tumor cells, which accumulate the molecule based on their high metabolic demands (Fig. 32) [159]. [18F]-FDG (82) is recognized as glucose and taken up by cells via the glucose transporter and phosphorylated by hexokinase to [18F]-FDG-6-phosphate, which cannot be metabolized further due to the presence of the 2-fluorine substituent [142, 143]. This maximizes the value of [18F]-FDG (82) as a PET agent by localizing the molecule and preventing its metabolic conversion into a myriad of degradation products. Moreover, decomposition of [18F]-FDG (82) ultimately affords 2-18O glucose, a heavy atom analogue that is innocuously metabolized by normal pathways. 18[F]-FDG (82) is used to diagnose a range of cancers including lung and intestinal cancer, lymphomas and melanomas and metastatic tumors but is not without its limitations as a diagnostic tool since its utility is dependent on tumor location and homogeneity [142].
The radiolabeling of ligands for specific receptors, transporters or enzymes with 18F provides reagents of particular value for the development and assessment of central nervous system (CNS) function in vivo in a non-invasive fashion that has both preclinical and clinical utility [142]. In addition to providing insights into brain metabolism and neurotransmission, 18F-labeled ligands can be used to provide evidence of brain distribution and target engagement by an exploratory drug candidate for which a homologue incorporating a positron emitting element has been identified. However, the development of PET ligands to label specific receptors in the brain is not always straightforward and success depends on several factors including the intrinsic affinity of a ligand for the protein target of interest, specific radioactivity of the probe molecule, susceptibility to metabolic degradation that leads to loss of the radiolabel, blood-brain barrier permeability of the compound which can be a function of P-glycoprotein-mediated efflux, and the extent of non-specific binding to tissues which can lead to poor signal contrast [143, 144]. The latter is often a function of molecules that are highly lipophilic and which are attracted with facility to fatty acid elements in cell membranes, providing high background signals that are manifested as noise. For optimal brain penetration, a log P value of between 1.5 and 3 is generally considered to be optimal although log D measurements that take into account the charge at the physiological pH of 7.4 are considered to be the more relevant index of lipophilicity [143]. However, there are exceptions to these guidelines and the development of 18F-labeled PET ligands remains an experimental science.
Many useful PET ligands that demonstrate specificity for target receptors, transporters or enzymes have been developed, and these ligands have proven useful as probes for imaging receptor occupancy, monitoring changes in target density, and measuring in vivo drug distribution and kinetics after local administration, particularly to the CNS and the lung (Fig. 33) [144, 160–163]. A prominent example is [18F]-fluoro-L-DOPA (83), the first agent developed to image the dopamine system. 83 is used as a metabolic tracer to characterize neuroendocrine tumors which take up [18F]-FDG (82) only poorly and to monitor dopamine distribution and metabolism in the brain where it is converted to [18F]-fluorodopamine (84) and additional metabolites useful in the study of Parkinson’s disease [144, 164–167]. [18F]-2β-carbomethoxy-3β-(4-chlorophenyl)-8-(2-fluoroethyl)-nortropane ([18F]-FECNT) (85) is an analogue of cocaine used to assess the density of presynaptic dopamine transporters whilst [18F]-fallypride (86) is a useful ligand for labeling the dopamine D2 receptor [144, 168–170].
WAY-100635 (87) was originally described as a potent and selective 5HT1A receptor antagonist, K i = 2.2 nM, prompting the development of labeled forms used to illuminate the pharmacology of these receptors in vivo [144, 171]. A 11C label was readily introduced at the carbonyl carbon for PET imaging since the absence of fluorine in the molecule prevented simple replacement, necessitating the development and profiling of the analogues 88 and 89 in order to take advantage of the properties of 18F [172]. However, more recent studies have revealed that 87 binds to dopamine D4.2 receptors with a K i = 16.4 nM where it acts as a full agonist, providing caution on the use of 87 and its analogues in imaging [171].
18F-Altanserine (90) is used to label 5HT2A receptors while MA-1 (91) offers enhanced receptor selectivity and lipophilicity [144, 173, 174]. 2-[18F]-F-A85380 (92) and 6-[18F]-F-A85380 (93) have been developed as useful PET ligands for labeling the nicotinic acetylcholine receptor but take several hours to reach steady state concentration in the brain and exhibit low binding potentials (Bmax/Kd), prompting the development of 94 as a α2β4 antagonist, Ki = 240 pM, with improved brain kinetics [144, 175].
The 18F-labelled styrylpyridine derivative florbetapir (95) was approved by the FDA on April 6th, 2012 for use as a PET ligand that labels Aβ amyloid plaques in the brain (Fig. 34) [176, 177]. 18F-Florbetapir (95) derives from a series of stilbene derivatives that labeled amyloid plaques but were too lipophilic to be practically useful since they afforded high levels of non-specific binding in healthy brain [178–187]. The replacement of one of the phenyl rings with a pyridine heterocycle coupled with the introduction of a short polyethylene glycol moiety terminating with fluorine substitution provided 95, a molecule with improved physical properties that bound to amyloid Aβ aggregates from post-mortem Alzheimer’s disease brain tissue with a Kd of 3.72 nM and a Bmax of 8,811 fmol/mg protein [185]. Licensing of 18F-florbetapir (95) was based on the results of a Phase 3 clinical trial conducted in 29 patients dying of Alzheimer’s disease in which the amyloid Aβ plaque levels estimated by the tracer were compared with post-mortem samples, an analysis that revealed excellent concordance [187]. Three additional amyloid plaque labeling agents in clinical trials are 18F-florbetaben (96), a close analogue of 18F-florbetapir (95), 18F-flutemetamol (97) and 18F-AZD4694 (98) [188–196].
The introduction of 18F into potential PET ligands requires synthetic methodology that allows late stage installation of the radiolabel under conditions that are compatible with inherent and potentially labile functionality [141–145, 197]. The most common method of introducing 18F into an alkyl chain relies upon 18F fluoride displacement of a leaving group, typically a sulfonate, with cryptands used to sequester the counter ion and enhance the reactivity of the weakly nucleophilic anion [141–145, 197]. However, preparative procedures for a number of prosthetic moieties that incorporate alkyl fluorides have been developed that offer some advantage should this element be a part of the candidate molecule [144]. The application of microwave heating techniques has markedly facilitated some of these procedures, improving both yields and speed of synthesis. The 18F-fluoroethyl moiety is quite prominent in 18F PET ligands, with the label readily introduced by treating the corresponding tosylate or trifluoromethyl sulfonate derivative with 18F-fluoride [141–145, 168, 169, 197]. 18F-Labeled 2,2,2-trifluoroethyl 4-methylbenzenesulfonate (18F-99) is a prosthetic group that offers a useful approach to prepare potential PET ligands and is accessible from the unlabeled material which is sequentially subjected to nBuLi-mediated elimination of HF to afford the divinyl fluoride 100 which adds 18F quickly and efficiently under mild and carefully optimized conditions (Fig. 35) [198]. 18F-99 was prepared with good specific radioactivity and readily alkylated 4-cyanophenol, diphenylmethanethiol, 4-chlorobenzoic acid and dibenzylamine rapidly under mild conditions mediated by Cs2CO3 as the base in DMF as solvent [198].
Aryl fluorides are a much more common structural motif, ubiquitously explored in lead optimization campaigns where a phenyl ring is part of the pharmacophore. Whilst these moieties offer the potential to prepare an 18F-labeled analogue of a highly optimized ligand, the introduction of fluorine to these rings late in a synthetic scheme can present a considerable challenge and both nucleophilic and electrophilic processes have been developed [141–145, 197].
Aryl or heteroaryl ring fluorination can be accomplished using 18F− to displace a range of leaving groups, including F, Cl, Br, and I in addition to NO2, NMe3 + and ArI+, a reaction facilitated by the presence of an electron withdrawing moiety on the ring [151–158, 197, 199]. Electrophilic aromatic fluorination processes are more difficult and have typically relied upon the use of 18F2 gas, which is highly reactive and must be used under controlled conditions, with regiochemistry influenced by using trialkyl tin or mercury substituents to direct incorporation [197]. A promising new approach to electrophilic fluorination that exhibits greater compatibility with the kind of functionality found in drug molecules takes advantage of palladium catalysis with the palladium fluoride complex 102 prepared efficiently from 101 by exposure to fluoride for 5 min (Fig. 36) [200]. The palladium fluoride complex 102 functions as an electrophilic fluorinating agent and reacts with Pd-aryl complexes to afford aryl fluorides in good yield when heated in acetone at 85 °C for 10 min [200]. In a representative example, Pd complex 105 afforded the aryl fluoride 106 in 72 % yield under these conditions. Practical considerations relating this chemistry to 18F labeling applications include the stability of the Pd complexes 101 and 102, with the former stable at room temperature and to brief air exposure, whilst the palladium fluoride complex 102 is stable at 100 °C for 24 h and in 10 % aqueous CH3CN for 3 h at 23 °C. However, stoichiometric quantities of the catalyst 102 are required and care must be taken to purify the organic products to remove residual Pd. Complex 101 was shown to react with solutions of 18F−prepared under conventional conditions to afford 18F-labelled 102 which provided 18F-labelled 106 in an average 18 % radiochemical yield in two steps from 18F− [200].
A particularly interesting fluorination process that relies upon fluoride ion as the source and is radical-based in nature takes advantage of the manganese porphyrin catalyst Mn(TMP)Cl (107) to promote an oxidative fluorination that is applicable to aliphatic C-H moieties, potentially offering wide substrate versatility (Fig. 37) [201]. The experimental protocol exposes substrate to three equivalents of AgF, 6–8 equivalents of iodosylbenzene, 0.3 equivalents of tetrabutylammonium fluoride trihydrate and 6–8 mole % of catalyst 107 in mixture of CH3CN and CH2Cl2 at 50 °C under an inert atmosphere. The active catalyst in this process is believed to be trans-difluoroMn IV(TMP) (108) which transfers a fluorine atom to a carbon radical generated by a MnV oxo species produced by iodosylbenzene oxidation, as depicted in the catalytic cycle shown in Fig. 37. The procedure provides products in modest yields, typically just below 50 %, and with generally poor stereoselectivity at sites that are electronically unactivated. Nevertheless, the procedure is compatible with a wide range of functional groups and provides access to compounds that would otherwise require significant synthetic manipulation to access by a de novo approach.
8 Conclusion
Fluorine is a unique atom that has found widespread application in drug design to address problems associated with controlling conformation, enhancing potency, interfering with metabolism, increasing membrane permeability and modulating the pK a of proximal functionality. Although the physical chemistry underlying many of these effects is reasonably well understood, there remains much to learn about the influence of fluorine on drug properties and how fluorine can be exploited to full advantage. With continuing advances in the development of mild and efficient methods to introduce fluorine into organic molecules, it can be anticipated that applications of this remarkable element in drug discovery and development will grow in both sophistication and understanding. The promise of new and deeper insights into the properties and utility of fluorine should provide additional creative stimulus for improved synthetic methods, broadening and enhancing productive applications.
References
Vulpetti A, Hommel U, Landrum G, Lewis R, Dalvit C (2009) Design and NMR-based screening of LEF, a library of chemical fragments with different local environments of fluorine. J Am Chem Soc 131:12949–12959
Jordan JB, Poppe L, Xia X, Cheng AC, Sun Y, Michelsen K, Eastwood H, Schnier PD, Nixey T, Zhong W (2012) Fragment based drug discovery: practical implementation based on 19F NMR spectroscopy. J Med Chem 55:678–687
Vulpetti A, Dalvit C (2012) Fluorine local environments: from screening to drug design. Drug Discov Today 17:890–897
Harper DB, O’Hagan D, Murphy CD (2003) The handbook of environmental chemistry, vol 3 Part P. Springer, Berlin/Heidelberg, pp 141–169
Bowen HJM (1966) Trace elements in biochemistry. Academic, London
Sharpe AG (1967) In: Gutmann V (ed) Halogen chemistry, vol 1. Academic, New York, p 12
Neidleman SL, Geigert J (1986) Biohalogenation: principles, basic roles and applications. Ellis Horwood Ltd, Chichester
Vaillancourt F, Yeh E, Vosburg DA, Garneau-Tsodikova S, Walsh CT (2006) Nature’s inventory of halogen catalysts: oxidative strategies predominate. Chem Rev 106:3364–3378
O’Hagan D (2006) Recent developments on the fluorinase of streptomyces cattleya. J Fluor Chem 127:1479–1483
Zhu X, Robinson DA, McEwan AR, O’Hagan D, Naismith JH (2007) Mechanism of enzymatic fluorination in Streptomyces cattleya. J Am Chem Soc 139:14597–14604
Wagner C, El Omari M, König GM (2009) Biohalogenation: nature’s way to synthesize halogenated metabolites. J Nat Prod 72:540–553
Meyer M, O’Hagan D (1992) Rare fluorinated natural products. Chem Br 28:785
Heidelberger C, Chaudhuri NK, Danneberg P, Mooren D, Griesbach L, Duschinsky R, Schnitzer RJ, Pleven E, Scheiner J (1957) Fluorinated pyrimidines, a new class of tumour-inhibitory compounds. Nature 179:663–666
Bégué J-P, Bonnet-Delpon D (2008) Bioorganic and medicinal chemistry of fluorine. Wiley, Hoboken
Böhm H-J, Banner D, Bendels S, Kansy M, Kuhn B, Müller K, Obst-Sander U, Stahl M (2004) Fluorine in medicinal chemistry. ChemBioChem 5:637–643
Müller K, Faeh C, Diederich F (2007) Fluorine in pharmaceuticals: looking beyond intuition. Science 317:1881–1886
Hagmann WK (2008) The many roles for fluorine in medicinal chemistry. J Med Chem 51:4359–4369
Purser S, Moore PR, Swallow S, Gouverneur V (2008) Fluorine in medicinal chemistry. Chem Soc Rev 37:320–330
Shah P, Westwell AD (2007) The role of fluorine in medicinal chemistry. J Enzym Inhib Med Chem 22:527–540
O’Hagan D (2008) Understanding organofluorine chemistry: an introduction to the C-F bond. Chem Soc Rev 37:308–319
Hunter L (2010) The C-F bond as a conformational tool in organic and biological chemistry. Belistein J Org Chem 6. doi:103762/bjoc.6.38
Zimmer LE, Sparr C, Gilmour R (2011) Fluorine conformational effects in organocatalysis: an emerging strategy for molecular design. Angew Chem Int Ed 50:11860–11871
Buissonneaud DY, van Mourik T, O’Hagan D (2010) A DFT study on the origin of the fluorine gauche effect in substituted fluoroethanes. Tetrahedron 66:2196–2202
O’Hagan D (2012) Organofluorine chemistry: synthesis and conformation of vicinal fluoromethylene motifs. J Org Chem 77:3689–3699
Wu D, Tian A, Sun H (1998) Conformational properties of 1,3-difluoropropane. J Chem Phys A 102:9901–9905
Tavasli M, O’Hagan D, Pearson C, Petty MC (2002) The fluorine gauche effect. Langmuir isotherms report the relative conformational stability of (±)-erythro- and (±)-threo-9,10-difluorostearic acids. Chem Commun 1226–1227
Hunter L, Kirsch P, Slawin AMZ, O’Hagan D (2009) Synthesis and structure of stereoisomeric multivicinal hexafluoroalkanes. Angew Chem Int Ed 48:5457–5460
Banks JW, Batsanov AS, Howard JAK, O’Hagan D, Rzepa H, Martin-Santamaria S (1999) The preferred conformation of α-fluoroamides. J Chem Soc Perkin 2:2409–2411
Briggs CRS, O’Hagan D, Howard JAK, Yulfi DS (2003) The C-F bond as a tool in the conformational control of amides. J Fluor Chem 119:9–13
Dalvit D, Vulpetti A (2012) Intermolecular and intramolecular hydrogen bonds involving fluorine atoms: implications for recognition, selectivity, and chemical properties. ChemMedChem 7:262–272
Schϋler M, O’Hagan D, Slawin AMZ (2005) The vicinal F-C-C-F moiety as a tool for influencing peptide conformation. Chem Commun 4324–4326
O’Hagan D, Rzepa HS, Schϋler M, Slawin AMZ (2006) The vicinal difluoro motif: the synthesis and conformation of erythro- and threo-diastereomers of 1,2-difluorodiphenylethanes, 2,3-difluorosuccinic acids and their derivatives. Beilstein J Org Chem. doi:10.1186/1860-5397-2-19
Winkler M, Moraux T, Khairy HA, Scott RH, Slawin AMZ, O’Hagan D (2009) Synthesis and vanilloid receptor (TRPV1) activity of the enantiomers of α-fluorinated capsaicin. ChemBioChem 10:823–828
Peddie V, Butcher RJ, Robinson WT, Wilce MCJ, Traore DAK, Abell AD (2012) Synthesis and conformation of fluorinated β-peptidic compounds. Chem Eur J 18:6655–6662
O’Hagan D, Bilton C, Howard JAK, Knight L, Tozer DJ (2001) The preferred conformation of N-β-fluoroethylamides: observation of the fluorine amide gauche effect. J Chem Soc Perkin Trans 2:605–607
O’Hagan D, Rzepa HS (1997) Some influences of fluorine in bioorganic chemistry. Chem Commun 645–652
Holmgren SK, Taylor KM, Bretscher LE, Raines RT (1998) Code for collagen’s stability deciphered. Nature 392:666–667
Bretscher LE, Jenkins CL, Taylor KM, DeRider ML, Raines RT (2001) Conformational stability of collagen relies on a stereoelectronic effect. J Am Chem Soc 123:777–778
Hodges JA, Raines RT (2003) Stereoelectronic effects on collagen stability: the dichotomy of 4-fluoroproline diastereomers. J Am Chem Soc 125:9262–9263
Doi M, Nishi Y, Uchiyama S, Nisiuchi Y, Nakazawa T, Ohkubo T, Kobayasji Y (2003) Characterization of collagen model peptides containing 4-fluoroproline: (4(S)-fluoroproline-Pro-Gly)10 forms a triple helix, but (4(R)-fluoroproline-Pro-Gly)10 does not. J Am Chem Soc 125:9922–9923
Hodges JA, Raines RT (2005) Stereoelectronic and steric effects in the collagen triple helix: toward a code for strand association. J Am Chem Soc 127:15923–15932
Shoulders MD, Kamer KJ, Raines RT (2009) Origin of the stability conferred upon collagen by fluorination. Bioorg Med Chem Lett 19:3859–3862
Doi M, Nishi Y, Kiritoshi N, Iwata T, Nago M, Nakano H, Uchiyama S, Nakazawa T, Wakamiya T, Kbayashi Y (2002) Simple and efficient syntheses of Boc- and Fmoc-protected 4(R)- and 4(S)-fluoroproline solely from 4(R)-hydroxyproline. Tetrahedron 58:8453–8459
DeRider ML, Wilkins SJ, Waddell MJ, Bretscher LE, Weinhold F, Raines RT, Markley JL (2002) Collagen stability: insights from NMR spectroscopic and hybrid density functional computational investigations of the effect of electronegative substituents on prolyl ring conformations. J Am Chem Soc 124:2497–2505
Park S, Radmer RJ, Klein TE, Pande VS (2005) A new set of molecular mechanics parameters for hydroxyproline and its use in molecular dynamics simulations of collagen-like peptides. J Comput Chem 26:1612–1616
Hodges JA, Raines RT (2006) Energetics of an n → π* interaction that impacts protein structure. Org Lett 8:4695–4697
Hinderaker MP, Raines RT (2003) An electronic effect on protein structure. Protein Sci 12:1188–1194
Raines RT (2006) 2005 Emil Thomas Kaiser Award. Protein Sci 15:1219–1225
Improta R, Mele F, Crescenzi O, Benzi C, Barone V (2002) Understanding the role of stereoelectronic effects in determining collagen stability: a quantum mechanical/molecular mechanical study of (proline-proline-glycine)n polypeptides. J Am Chem Soc 124:7857–7865
Kim W, Hardcastle KL, Conticello VP (2006) Fluoroproline flip-flop: regiochemical reversal of a stereoelectronic effect on peptide and protein structures. Angew Chem Int Ed 45:8141–8814
Kitamoto T, Ozawa T, Abe M, Marubayashi S, Yamazaki T (2008) Incorporation of fluoroprolines to proctolin: study on the effect of a fluorine atom towards peptidic conformation. J Fluor Chem 129:286–293
Briggs CRS, Allen MJ, O’Hagan D, Tozer DJ, Slawin AMZ, Goeta AE, Howard JAK (2004) The observation of a large gauche preference when 2-fluoroethylamine and 2-fluoroethanol become protonated. Org Biomol Chem 2:732–740
Lankin DC, Chandrakumar NS, Rao SN, Spangler DP, Snyder JP (1993) Protonated 3-fluoropiperidines: an unusual fluoro directing effect and a test for quantitative theories of solvation. J Am Chem Soc 115:3356–3357
Snyder JP, Chandrakumar NS, Sato H, Lankin DC (2000) The unexpected diaxial orientation of cis-3,5-difluoropiperidine in water: a potent CF – NH charge-dipole effect. J Am Chem Soc 122:544–545
Sun A, Lankin DC, Hardcastle K, Snyder JP (2005) 3-Fluoropiperidines and N-methyl-3-fluoropiperidinium salts: the persistence of axial fluorine. Chem Eur J 11:1579–1591
Gerig JT, McLeod RS (1973) Conformation of cis- and trans-4-fluoro-L-proline in aqueous solution. J Am Chem Soc 95:5725–5729
Deniau G, Slawin AMZ, Lebl T, Chorki F, Issberner JP, van Mourik T, Heygate JM, Lambert JJ, Etherington L-A, Sillar KT, O’Hagan D (2007) Synthesis, conformation and biological evaluation of the enantiomers of 3-fluoro-γ-aminobutyric acid ((R)- and (S)-3F-GABA): an analogue of the neurotransmitter GABA. ChemBioChem 8:2265–2274
Clift MD, Ji H, Deniau GP, O’Hagan D, Silverman RB (2007) Enantiomers of 4-amino-3-fluorobutanoic acid as substrates for γ-aminobutyric acid aminotransferase: conformational probes for GABA binding. Biochemistry 46:13819–13828
Yamamoto I, Deniau GP, Gavande N, Chebib M, Johnston GAR, O’Hagan D (2011) Agonist responses of (R)- and (S)-3-fluoro-γ-aminobutyric acids suggest an enantiomeric fold for GABA binding to GABAC receptors. Chem Commun 47:7956–7958
Chia PW, Livesey MR, Slawin AMZ, van Mourik T, Wyllie DJA, O’Hagan D (2012) 3-Fluoro-N-methyl-D-asaprtic acid (3F-NMDA) stereoisomers as conformational probes for exploring agonist binding at NMDA receptors. Chem Eur J 18:8813–8819
Gooseman NEJ, O’Hagan D, Slawin AMZ, Teale AM, Tozer DJ, Young RJ (2006) The intramolecular β-fluorine…ammonium interaction in 4- and 8-membered rings. Chem Commun 3190–3192
Campbell NH, Smith DL, Reszka AP, Neidle S, O’hagan D (2011) Fluorine in medicinal chemistry: β-fluorination of peripheral pyrrolidine attached to acridine ligands affects their interactions with G-quadruplex DNA. Org Biomol Chem 9:1328–1331
Abraham RJ, Chambers EJ, Thomas AW (1994) Conformational analysis. Part 22. An NMR and theoretical investigation of the gauche effect in fluoroethanols. J Chem Soc Perkin Trans 2:949–955
Abraham RJ, Smith TAD, Thomas AW (1996) Conformational analysis. Part 28. OH-F hydrogen bonding and the conformation of trans-2-fluorocyclohexanol. J Chem Soc Perkin Trans 2:1949–1955
Fried J, Sabo EF (1954) 9α-Fluoro derivatives of cortisone and hydrocortisone. J Am Chem Soc 76:1455–1456
Fried J (1955) Biological effects of 9 α-fluorohydrocortisone and related halogenated steroids in animals. Ann N Y Acad Sci 61:573–581
O’Hagan D (2010) Fluorine in health care: organofluorine-containing blockbuster drugs. J Fluor Chem 131:1071–1081
Mock W, Zhang JZ (1990) Concerning the relative acidities of simple alcohols. Tetrahedron Lett 31:5687–5688
Weeks CM, Duax WL, Wolff ME (1973) A comparison of the structure of six corticosteroids. J Am Chem Soc 95:2865–2868
Wolff ME, Baxter JD, Kollman PA, Lee DL, Kuntz ID, Bloom E, Matulich DT, Morris J (1978) Nature of steroid-glucocorticoid receptor interactions: thermodynamic analysis of the binding reaction. Biochemistry 17:3201–3208
Myers AG, Barbay JK, Zhong B (2001) Asymmetric synthesis of chiral organofluorine compounds: use of nonracemic fluoroiodoacetic acid as a practical electrophile and its application to the synthesis of monofluoro hydroxyethylene dipeptide isosteres within a novel series of HIV protease inhibitors. J Am Chem Soc 123:7207–7219
Leroux F, Jeschke P, Schlosser M (2005) α-Fluorinated ethers, thioethers, and amines: anomerically biased species. Chem Rev 105:827–856
Massa MA, Spangler DP, Durley RC, Hickory BS, Connolly DT, Witherbee BJ, Smith ME, Sikorski JA (2001) Novel heteroaryl replacements of aromatic 3-tetrafluoroethoxy substituents in trifluoro-3-(tertiaryamino)-2-propanols as potent inhibitors of cholesterol ester transfer protein. Bioorg Med Chem Lett 11:1625–1628
Horne DB, Bartberger MD, Kaller MR, Monenschein H, Zhong W, Hitchcock SA (2009) Synthesis and conformational analysis of α, α-difluoroalkyl heteroaryl ethers. Tetrahedron Lett 50:5452–5455
Hehre WJ, Radom L, Pople JA (1972) Molecular orbital theory of the electronic structure of organic compounds. XII. Conformations, stabilities, and charge distributions in monosubstituted benzenes. J Am Chem Soc 94:1496–1504
Anderson GM III, Kollman PA, Domelsmith LN, Houk KN (1979) Methoxy group nonplanarity in o-dimethoxybenzenes. Simple predictive models for conformations and rotational barriers in alkoxyaromatics. J Am Chem Soc 101:2344–2352
Hummel W, Huml K, Bürgi H-B (1988) Conformational flexibility of the methoxyphenyl group studies by statistical analysis of crystal structure data. Helv Chim Acta 71:1291–1302
Brameld KA, Kuhn B, Reuter DC, Stahl M (2008) Small molecule conformational preferences derived from crystal structure data. A medicinal chemistry focused analysis. J Chem Inf Model 48:1–24
Johnson F (1968) Allylic strain in six-membered rings. Chem Rev 68:375–413
Klocker J, Karpfen A, Wolschann P (2003) On the structure and torsional potential of trifluoromethoxybenzene: an ab initio and density functional study. Chem Phys Lett 367:566–575
Graton J, Wang Z, Brossard A-M, Monteiro DG, Questel J-YL, Linclau B (2012) An unexpected and significantly lower hydrogen-bond-donating capacity of fluorohydrins compared to nonfluorinated alcohols. Angew Chem Int Ed 51:6176–6180
Meanwell N (2011) Improving drug candidates by design: a focus on physicochemical properties as a means of improving compound disposition and safety. Chem Res Toxicol 24:1420–1456
Peters JE, Schnider P, Mattei P, Kansy M (2009) Pharmacological promiscuity: dependence on compound properties and target specificity in a set of recent Roche compounds. ChemMedChem 4:680–686
Azzaoui K, Hamon J, Faller B, Whitebread S, Jacoby E, Bender A, Jenkins JL, Urban L (2007) Modeling promiscuity based on in vitro safety pharmacology profiling data. ChemMedChem 2:874–880
Jamieson C, Moir EM, Rankovic Z, Wishart G (2006) Medicinal chemistry of hERG optimizations: highlights and hang-ups. J Med Chem 49:5029–5046
Fischer H, Kansy M, Bur D (2000) CAFCA: a novel tool for the calculation of amphiphilic properties of charged drug molecules. Chimia 54:640–645
Clark J, Perrin DD (1964) Prediction of the strengths of organic bases. Q Rev Chem Soc 18:295–320
Morgenthaler M, Schweizer E, Hoffmann-Röder A, Benini F, Martin RE, Jaeschke G, Wagner B, Fischer H, Bendels S, Zimmerli D, Scheider J, Diederich F, Kansy M, Müller K (2007) Predicting and tuning physicochemical properties in lead optimization: amine basicities. ChemMedChem 2:1100–1115
Bruice PY (1984) Relative reactivity of amines and oxyanions toward proton abstraction from nitroethane. Electrostatic effects and the reactivity-selectivity principle. J Am Chem Soc 106:5959–5964
Trepka RD, Harrington JK, Belisle JW (1974) Acidities and partition coefficients of fluoromethanesulfonamides. J Org Chem 39:1094–1098
Dean JA (1999) Lange’s handbook of chemistry, 15th edn. McGraw-Hill, New York
Hansch C, Leo A, Taft RW (1991) A survey of Hammett substituent constants and resonance field parameters. Chem Rev 91:165c–195c
Taagepera M, Summerhays KD, Hehre WJ, Topsom RD, Pross A, Radom L, Taft RW (1981) Analysis of the acidities of 3- and 4-substituted pyridinium and anilinium ions. J Org Chem 46:891–903
Schlosser M (1998) Parametrization of substituents: effects of fluorine and other heteroatoms on OH, NH, and CH acidities. Angew Chem Int Ed 110:1496–1513
Olsen J, Seiler P, Wagner B, Fischer H, Tschopp T, Obst-Sander U, Banner DW, Kansy M, Muller K, Diederich F (2004) A fluorine scan of the phenylamidinium needle of tricyclic thrombin inhibitors: effects of fluorine substitution on pKa and binding affinity and evidence for intermolecular C-F CN interactions. Org Biomol Chem 2:1339–1352
Michalik L, Auwerx J, Berger JP, Chatterjee VK, Glass CG, Gonzalez FJ, Grimaldi PA, Kadowaki T, Lazar MA, O’Rahilly S, Palmer CAN, Plutzky J, Reddy JK, Spiegelman BM, Staels B, Wahli W (2006) International Union of Pharmacology. LXI. Peroxisome proliferator-activated receptors. Pharmacol Rev 58:726–741
Ciocoiu CC, Ravna AW, Sylte I, Rustan AC, Hansen TV (2011) Synthesis, molecular modeling studies and biological evaluation of fluorine substituted analogs of GW 501516. Bioorg Med Chem 19:6982–6988
Ishikawa M, Hashimoto Y (2011) Improvement in aqueous solubility in small molecule drug discovery programs by disruption of molecular planarity and symmetry. J Med Chem 54:1539–1554
Black WC, Bayly CI, Davis DE, Desmarais S, Falgueyret J-P, Léger S, Li CS, Massé F, McKay DJ, Palmer JT, Percival MD, Robichaud J, Tsou N, Zamboni R (2005) Trifluoroethylamines as amide isosteres in inhibitors of cathepsin K. Bioorg Med Chem Lett 15:4741–4744
Herrero S, Suárez-Gea ML, Suárez-Gea ML, González-Muñiz R, García-López MT, Herranz R, Ballaz S, Barber A, Fortuño A, Del Río J (1997) Pseudopeptide CCK-4 analogues incorporating the Ψ[CH(CN)NH peptide bond surrogate. Bioorg Med Chem Lett 855–860
Kirsch P (2004) Modern fluoroorganic chemistry, synthesis, reactivity applications. Wiley-VCH, Weinheim
Wipf P, Mo T, Geib SJ, Caridha D, Dow GS, Gerena L, Roncal N, Milner EE (2009) Synthesis and biological evaluation of the first pentafluorosulfanyl analogs of mefloquine. Org Biomol Chem 7:4163–4165
Welch JT, Lim DS (2007) The synthesis and biological activity of pentafluorosulfanyl analogs of fluoxetine, fenfluramine, and norfenfluramine. Bioorg Med Chem 15:6659–6666
Altomonte S, Zanda M (2012) Synthetic chemistry and biological activity of pentafluorosulphanyl (SF5) organic molecules. J Fluor Chem 143:57–93
Gujjar R, El Mazouni F, White KL, White J, Creason S, Shackleford DM, Deng X, Charman WN, Bathurst I, Burrows J, Floyd DM, Matthews D, Buckner FS, Charman SA, Phillips MA, Rathod PK (2011) Lead optimization of aryl and aralkyl amine-based triazolopyrimidine inhibitors of plasmodium falciparum dihydroorotate dehydrogenase with antimalarial activity in mice. J Med Chem 54:3935–3949
Smith DH, Waterbeemd H, Walker DK (2001) Pharmacokinetics and metabolism in drug design, methods and principles in medicinal chemistry, vol 13. Wiley-VCH, Weinheim
Barnette WE (1984) The synthesis and biology of fluorinated prostacyclins. CRC Crit Rev Biochem 15:201–235
Heek M, France CF, Compton DS, McLeod RL, Yumibe NP, Alton KB, Sybertz EJ, Davies HR Jr (1997) In vivo metabolism-based discovery of a potent cholesterol absorption inhibitor, SCH58235, in the rat and rhesus monkey through the identification of the active metabolites of SCH48461. J Pharmacol Exp Ther 283:157–163
Clader JW (2004) The discovery of ezetimibe: a view from outside the receptor. J Med Chem 47:1–9
Penning TD, Talley JJ, Bertenshaw SR, Carter JS, Collins PW, Docter S, Graneto MJ, Lee LF, Malecha JW, Miyashiro JM, Rogers RS, Rogier DJ, Yu SS, Anderson GD, Burton EG, Cogburn EG, Gregory SA, Koboldt CM, Perkins WE, Seibert K, Veenhuizen AW, Zhang AW, Isaakson PC (1997) Synthesis and biological evaluation of the 1,5-diarylpyrazole class of cyclooxygenase-2 inhibitors: identification of 4-[5-(4-methylphenyl)-3- (trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (SC-58635, celecoxib). J Med Chem 40:1347–1365
Rosenblum SB, Huynh T, Afonso A, Davis HR Jr, Yumibe N, Clader JW, Burnett DA (1998) Discovery of 1-(4-fluorophenyl)-(3R)-[3-(4-fluorophenyl)-(3S)-hydroxypropyl]-(4S)-(4-hydroxyphenyl)-2-azetidinone (SCH 58235): a designed, potent, orally active inhibitor of cholesterol absorption. J Med Chem 41:973–980
St Jean Jr DJ, Fotsch C (2012) Mitigating heterocycle metabolism in drug discovery. J Med Chem 55:6002–6020
Wan Z-K, Chenail E, Xiang J, Li H-Q, Ipek M, Bard J, Svenson K, Mansour TS, Xu X, Suri V, Hahm S, Xing Y, Johnson CE, Li X, Qadri A, Panza D, Perreault M, Tobin JF, Saiah E (2009) Efficacious 11β-hydroxysteroid dehydrogenase type I inhibitors in the diet-induced obesity mouse model. J Med Chem 52:5449–5461
Pierson PD, Fettes A, Freichel C, Gatti-McArthur S, Hertel C, Huwyler J, Mohr P, Nakagawa T, Nettekoven M, Plancher JM, Raab S, Richter H, Roche O, Rodríguez Sarmiento RM, Schmitt M, Schuler F, Takahashi T, Taylor S, Ullmer C, Wiegand R (2009) 5-Hydroxyindole-2-carboxylic acid amides: novel histamine-3 receptor inverse agonists for the treatment of obesity. J Med Chem 52:3855–3868
Seibert K, Zhang Y, Leahy K, Hauser S, Masferrer J, Perkins W, Lee L, Isakson P (1994) Pharmacological and biochemical demonstration of the role of cyclooxygenase 2 in inflammation and pain. Proc Natl Acad Sci U S A 91:12013–12017
Park BK, Kitteringham NR, O’Neill PM (2001) Metabolism of fluorine-containing drugs. Annu Rev Pharmacol Toxicol 41:443–470
Koerts J, Soffers AE, Vervoort J, De Jager A, Rietjens IM (1998) Occurrence of the NIH shift upon the cytochrome P450-catalyzed in vivo and in vitro aromatic ring hydroxylation of fluorobenzenes. Chem Res Toxicol 11:503–512
Dear GJ, Ismail IM, Mutch PJ, Plumb RS, Davies LS (2000) Urinary metabolites of a novel quinoxaline non-nucleoside reverse transcriptase inhibitor in rabbit, mouse and human: identification of fluorine NIH shift metabolites using NMR and tandem MS. Xenobiotica 30:407–426
Boguslavsky V, Bebecchi M, Morris AJ, Jhon DY, Rhee SG, McLaughlin S (1994) Effect of monolayer surface pressure on the activities of phosphoinositide-specific phospholipase C-β.1, -γ.1, and δ1. Biochemistry 33:3032–3037
Hanakam F, Gerisch G, Lotz S, Alt T, Seelig A (1996) Binding of hisactophilin I and II to lipid membranes is controlled by a pH-dependent myristoyl−histidine switch. Biochemistry 35:11036–11044
Fischer H, Gottschlich R, Seelig A (1998) Blood-brain barrier permeation: molecular parameters governing passive diffusion. J Membr Biol 165:201–211
Lennarnas H, Abrahamsson B (2005) The use of biopharmaceutical classification of drugs in drug discovery and development: current status and future extensions. J Pharm Pharmacol 57:273–285
Smart B (2001) Fluorine substituent effects (on bioactivity). J Fluor Chem 109:3–11
Jacobs RT, Bernstein PR, Cronk LA, Vacek EP, Newcomb LF, Aharony D, Buckner CK, Kusner EJ (1994) Synthesis, structure-activity relationships, and pharmacological evaluation of a series of fluorinated 3-benzyl-5-indolecarboxamides: identification of 4-[[5-[((2R)-2-methyl-4,4,4-trifluorobutyl)carbamoyl]-1-methylindol-3-yl]methyl]-3methoxy-N-[(2-methylphenyl)sulfonyl]benzamide, a potent, orally active antagonist of leukotrienes D4 and E4. J Med Chem 37:1282–1297
The Caco-2 cell line, derived from a human colorectal carcinoma, has become an established in vitro model for the prediction of drug absorption across the human intestine. When cultured on semi-permeable membranes, Caco-2 cells differentiate into a highly functionalized epithelial barrier with remarkable morphological and biochemical similarity to the small intestinal columnar epithelium. The membrane transport properties of novel compounds can thereby be assessed using these differentiated cell monolayers. The apparent permeability coefficients (Papp) obtained from Caco-2 cell transport studies have been shown to correlate to human intestinal absorption. Irvine JD, Takahashi L, Lockhart K, Cheong J, Tolan JW, Selick HE, Grove JR, (1999) MDCK (Madin–Darby canine kidney) cells: a tool for membrane permeability screening. J Pharm Sci 88:23–33
Pinto DJ, Orwat MJ, Wang S, Fevig JM, Quan ML, Amparo E, Cacciola J, Rossi KA, Alexander RS, Smallwood AM, Luettgen JM, Liang L, Aungst BJ, Wright MR, Knabb RM, Wong PC, Wexler RR, Lam PY (2001) Discovery of 1-[3-(aminomethyl)phenyl]-N-[3-fluoro-2′-(methylsulfonyl)-[1,1′-biphenyl]-4-yl]-3-(trifluoromethyl)-1H-pyrazole-5-carboxamide (DPC423), a highly potent, selective, and orally bioavailable inhibitor of blood coagulation factor Xa. J Med Chem 44:566–578
Shimizu H, Yamasaki T, Yoneda Y, Muro F, Hamada T, Yasukochi T, Tanaka S, Toki T, Yokoyama M, Morishita K, Iimura S (2011) Discovery of imidazo[1,2-b]pyridazines as IKKβ inhibitors. Part 3: Exploration of effective compounds in arthritis models. Bioorg Med Chem Lett 21:4550–4555 and refs therein
PAMPA: Parallel Artificial Membrane Permeability Assay is a method which determines the permeability of substances from a donor compartment, through a lipid-infused artificial membrane into an acceptor compartment. PAMPA models regularly exhibit a high degree of correlation with permeation across a variety of barriers including Caco-2 cultures, the GI tract, blood-brain barriers and skin
Carosati E, Sciabola S, Cruciani G (2004) Hydrogen-bonding interactions of covalently bonded fluorine atoms: from crystallographic data to a new angular function in the GRID force field. J Med Chem 47:5114–5125
Pfleiderer T, Waldner I, Bertagnolli H, Tödheide K, Fischer HE (2000) The structure of liquid and supercritical deuterium fluoride from neutron scattering and using high-pressure techniques. J Chem Phys 113:3690–3696
Barbarich TJ, Rithner CD, Miller SM, Anderson OP, Strauss SH (1999) Significant inter- and intra-molecular O-H⋯F-C hydrogen-bonding. J Am Chem Soc 121:4280–4281
Murray-Rust P, Stallings WC, Monti CT, Preston RK, Glusker JP (1983) Intermolecular interactions of the C-F Bond: the crystallographic environment of fluorinated carboxylic acids and related structures. J Am Chem Soc 105:3206–3214
Scerba MT, Leavitt CM, Diener ME, DeBlase AF, Guasco TL, Siegler MA, Bair N, Johnson MA, Lectka T (2011) NH + −F hydrogen-bonding in a fluorinated “Proton Sponge” derivative: integration of solution, solid-state, gas-phase, and computational studies. J Org Chem 76:7975–7984
Prakash GKS, Wang F, Rahm M, Shen J, Ni C, Haiges R, Olah GA (2011) On the nature of C-H⋯F-C interactions in hindered CF3-C(sp3) bond rotations. Angew Chem Int Ed 50:11761–11764
Chopra D (2012) Is organic fluorine really “not” polarizable? Cryst Growth Des 12:541–546
Dunitz JD (2004) Organic fluorine: odd man out. ChemBioChem 5:614–621
Cormanich RA, Freitas MP, Tormena CF, Rittner R (2012) The F⋯HO intramolecular hydrogen bond forming five-membered rings hardly appear in monocyclic organofluorine compounds. RSC Adv 2:4169–4174
Dunitz JD, Taylor R (1997) Organic fluorine hardly ever accepts hydrogen bonds. Chem Eur J 3:89–98
Fonseca TAO, Ramalho TC, Freitas MP (2012) F⋯HO intramolecular hydrogen bond as the main transmission mechanism for 1hJF, H(O) coupling constant in 2′-fluoroflavonol. Magn Reson Chem 50:551–556
Cormanich RA, Moreira MA, Freitas MP, Ramalho TC, Anconi CPA, Rittner R, Contreras RH, Tormena CF (2011) 1hJF, H coupling in 2-fluorophenol revisited: is intramolecular hydrogen bond responsible for this long-range coupling? Magn Reson Chem 49:763–767
Nolting DN, Nickels ML, Huo N, Pham W (2012) Molecular imaging probe development: a chemistry perspective. Am J Nucl Med Mol Imaging 2:273–306
Le Bars D (2006) Fluorine-18 and medical imaging: radiopharmaceuticals for positron emission tomography. J Fluor Chem 127:1488–1493
Miller PW, Long NJ, Vilar R, Gee AD (2008) Synthesis of 11C, 18F, 15O and 13N radiolabels for positron emission tomography. Angew Chem Int Ed 47:8998–9033
Ametamey SM, Honer M, Schubiger PA (2008) Molecular imaging with PET. Chem Rev 108:1501–1516
Pauwels EKJ, van der Klaauw AA, Corporaal T, Stokkel MPM (2002) Fluorine-18-radiolabeled pharmaceuticals for imaging with positron emission tomography, excluding [18F]-fluorodeoxyglucose. Drugs Future 27:655–667
Daniels S, Farah M, Tohid S, Velanguparackel W, Westwel AD (2010) The role and future potential of fluorinated biomarkers in positron emission tomography. Expert Opin Drug Discov 5:291–304
Sharma S, Rodriguez AL, Conn PJ, Lindsley CW (2008) Synthesis and SAR of a mGluR5 allosteric partial antagonist lead: unexpected modulation of pharmacology with slight structural modifications to a 5-(phenylethynyl)pyrimidine scaffold. Bioorg Med Chem Lett 18:4098–4101
Sams AG, Mikkelsen AK, Brodbeck RMG, Pu X, Ritzén A (2011) Efficacy switching SAR of mGluR5 allosteric modulators: highly potent positive and negative modulators from one chemotype. Bioorg Med Chem Lett 21:3407–3410
Packiarajan M, Mazza Ferreira CG, Hong S-P, White AD, Chandrasena G, Pu X, Brodbeck RM, Robichaud AJ (2012) N-Aryl pyrrolidinonyl oxadiazoles as potent mGluR5 positive allosteric modulators. Bioorg Med Chem Lett 22:5658–5662
Packiarajan M, Mazza Ferreira CG, Hong S-P, White AD, Chandrasena G, Pu X, Brodbeck RM, Robichaud AJ (2012) Azetidinyl oxadiazoles as potent mGluR5 positive allosteric modulators. Bioorg Med Chem Lett 22:6469–6474
Schirrmacher R, Wängler C, Schirrmacher E (2007) Recent developments and trends in 18F-radiochemistry: syntheses and applications. Mini Rev Org Chem 4:317–329
Cai L, Lu S, Pike VW (2008) Chemistry with [18F]-fluoride ion. Eur J Org Chem 2853–2873
Littich R, Scott PJH (2012) Novel strategies for fluorine-18 radiochemistry. Angew Chem Int Ed 51:1106–1109
Hollingworth C, Gouverneur V (2012) Transition metal catalysis and nucleophilic fluorination. Chem Commun 48:2929–2942
Prakash GKS, Chacko S (2008) Novel nucleophilic and electrophilic fluoroalkylation methods. Curr Opin Drug Discov Dev 11:793–802
Shimizu M, Hiyama T (2004) Modern synthetic methods for fluorine-substituted target molecules. Angew Chem Int Ed 44:214–231
Kirk KL (2008) Fluorination in medicinal chemistry: methods, strategies, and recent developments. Org Proc Res Dev 12:305–321
Furuya T, Kuttruff CA, Ritter T (2008) Carbon-fluorine bond formation. Curr Opin Drug Discov Dev 11:803–819
Pauwels EKJ (2001) 18F-Labeled fluorodeoxyglucose for PET imaging: the working mechanism and its clinical implication. Drugs Future 26:659–668
Gatley SJ, Volkow ND, Fowler JS, Ding Y-S, Logan J, Wang G-J, Gifford AN (2003) Positron emission tomography and its use to image the occupancy of drug binding sites. Drug Dev Res 59:194–207
Berridge MS, Heald DL, Lee Z (2003) Imaging studies of biodistribution and kinetics in drug development. Drug Dev Res 59:208–226
Hicks JW, Van Brocklin HF, Wilson AA, Houle S, Vasdev N (2010) Radiolabeled small molecule protein kinase inhibitors for imaging with PET or SPECT. Molecules 15:8260–8278
Badgaiyan RD (2011) Neurotransmitter imaging: basic concepts and future perspectives. Curr Med Imaging Rev 7:98–103
Jager PL, Chirakal R, Marriott CJ, Brouwers AH, Koopmans KP, Gulenchyn KY (2008) 6-L-18F-fluorodihydroxyphenylalanine PET in neuroendocrine tumors: basic aspects and emerging clinical applications. J Nucl Med 49:573–586
Minn H, Kauhanen S, Seppanen M, Nuutila P (2009) 18F-FDOPA: a multiple-target molecule. J Nucl Med 50:1915–1918
Garnett ES, Firnau G, Nahmias C (1983) Dopamine visualized in the basal ganglia of living man. Nature 305:137–138
DeJesus OT (2003) Positron-labeled DOPA analogs to image dopamine terminals. Drug Dev Res 59:249–260
Goodman MM, Kilts CD, Keil R, Shi B, Martarello L, Xing D, Votaw J, Ely TD, Lambert P, Owens MJ, Camp VM, Malveaux E, Hoffman JM (2000) 18F-Labeled FECNT: a selective radioligand for PET imaging of brain dopamine transporters. Nucl Med Biol 27:1–12
Davis MR, Votaw JR, Bremner JD, Byas-Smith MG, Faber TL, Voll RJ, Hoffman JM, Grafton ST, Kilts CD, Goodman MM (2003) Initial human PET imaging studies with the dopamine transporter ligand 18F-FECNT. J Nucl Med 44:855–861
Elsinga PH, Hatano K, Ishiwata K (2006) PET tracers for imaging of the dopaminergic system. Curr Med Chem 13:2139–2153
Chemel BR, Roth BL, Armbruster B, Watts VJ, Nichols DE (2006) WAY-100635 is a potent dopamine D4 receptor agonist. Psychopharmacol 188:244–251
Lang L, Jagoda E, Schmall B, Vuong B-K, Adams HR, Nelson DL, Carson RE, Eckelman WC (1999) Development of fluorine-18-labeled 5-HT1A antagonists. J Med Chem 42:1576–1586
Lemaire C, Cantineau R, Marcel Guillaume M, Plenevaux A, Christiaens L (1991) Fluorine-18-altanserin: a radioligand for the study of serotonin receptors with PET: radiolabeling and in vivo biologic behavior in rats. J Nucl Med 32:2266–2272
Herth MM, Kramer V, Piel M, Palner M, Riss PJ, Knudsen GM, Roesch F (2009) Synthesis and in vitro affinities of various MDL 100907 derivatives as potential 18F-radioligands for 5-HT2A receptor imaging with PET. Bioorg Med Chem 17:2989–3002
Gao Y, Kuwabara H, Spivak CE, Xiao Y, Kellar K, Ravert HT, Kumar A, Alexander M, Hilton J, Wong DF, Dannals RF, Horti AG (2008) Discovery of (−)-7-methyl-2-exo-[3′-(6-[18F]fluoropyridin-2-yl)-5′-pyridinyl]-7-azabicyclo[2.2.1]heptane, a radiolabeled antagonist for cerebral nicotinic acetylcholine receptor (α4β2-nAChR) with optimal positron emission tomography imaging properties. J Med Chem 51:4751–4764
Yang L, Rieves D, Ganley C (2012) Brain amyloid imaging – FDA approval of florbetapir F18 injection. New Engl J Med 367:885–887
Garber K (2012) First FDA-approved beta-amyloid diagnostic hits the market. Nat Biotechnol 30:575
Kung HF, Choi SR, Qu W, Zhang W, Skovronsky D (2010) 18F Stilbenes and styrylpyridines for PET imaging of Aβ plaques in Alzheimer’s disease: a miniperspective. J Med Chem 53:933–941
Kung HF, Lee C-W, Zhuang ZP, Kung MP, Hou C, Plossl K (2001) Novel stilbenes as probes for amyloid plaques. J Am Chem Soc 123:12740–12741
Wang W, Oya S, Kung MP, Hou C, Maier DL, Kung HF (2005) F-18 stilbenes as PET imaging agents for detecting β-amyloid plaques in the brain. J Med Chem 48:5980–5988
Qu W, Kung MP, Hou C, Benedum TE, Kung HF (2007) Novel styrylpyridines as probes for SPECT imaging of amyloid plaques. J Med Chem 50:2157–2165
Zhang W, Oyas S, Kung MP, Hou C, Maier DL, Kung HF (2005) F-18F PEG stilbenes as PET imaging agents targeting Aβ aggregates in the brain. Nucl Med Biol 32:799–809
Zhang W, Kung MP, Oya S, Hou C, Kung HF (2007) 18F-labeled styrylpyridines as PET agents for amyloid plaque imaging. Nucl Med Biol 34:89–97
Stephenson KA, Chandra R, Zhuang Z-P, Hou C, Oya S, Kung M-P, Kung HF (2007) Fluoro-pegylated (FPEG): imaging agents targeting Aβ aggregates. Bioconjug Chem 18:238–246
Choi SR, Golding G, Zhuang Z, Zhang W, Lim N, Hefti F, Benedum TE, Kilbourn MR, Skovronsky D, Kung HF (2009) Preclinical properties of 18F-AV-45: a PET agent for Aβ plaques in the brain. J Nucl Med 50:1887–1894
Wong DF, Rosenberg PB, Zhou Y, Kumar A, Raymont V, Ravert HT, Dannals RF, Nandi A, Brašić JR, Ye W, Hilton J, Lyketsos C, Kung HF, Joshi AD, Skovronsky DM, Pontecorvo MJ (2010) In vivo imaging of amyloid deposition in Alzheimer’s disease using the radioligand 18F-AV-45 (florbetapir F 18). J Nucl Med 51:913–920
Clark CM, Schneider JA, Bedell BJ, Beach TC, Bilker WB, Mintun MA, Pontecorvo MJ, Hefti F, Carpenter AP, Flitter ML, Krautkramer MJ, Kung HF, Coleman RE, Doraiswamy PM, Fleisher AS, Sabbagh MN, Sadowsky CH, Reiman PEM, Zehntner SP, Skovronsky DM (2011) Use of florbetapir-PET for imaging β-amyloid pathology. J Am Med Assoc 305:275–283
Fodero-Tavoletti MT, Brockschnieder D, Villemagne VL, Martin L, Connor AR, Thiele A, Berndt M, McLean CA, Krause S, Rowe CC, Masters CL, Dinkelborg L, Dyrks T, Cappai R (2012) In vitro characterization of [18F]-florbetaben, an Aβ imaging radiotracer. Nucl Med Biol 39:1042–1048
Ashford JW, Rosen A, Adamson M, Bayley P, Sabri O, Furst A, Black SE, Weiner M, Barthel H, Sabri O (2011) Florbetaben to trace amyloid-β in the Alzheimer brain by means of PET. J Alzheimer’s Dis 26(Suppl 3):117–121
Barthel H, Gertz H-J, Dresel S, Peters O, Bartenstein P, Buerger K, Hiemeyer F, Wittemer-Rump SM, Seibyl J, Reininger C, Sabri O (2011) Cerebral amyloid-β PET with florbetaben (18F) in patients with Alzheimer’s disease and healthy controls: a multicentre phase 2 diagnostic study. Lancet Neurol 10:424–435
Villemagne VL, Ong K, Mulligan RS, Holl G, Pejoska S, Jones G, O’Keefe G, Ackerman U, Tochon-Danguy H, Chan JG, Reininger CB, Fels L, Putz B, Rohde B, Masters C, Rowe CC L (2011) Amyloid imaging with 18F-florbetaben in Alzheimer disease and other dementias. J Nucl Med 52:1210–1217
Nelissen N, van Laere K, Thurfjell L, Owenius R, Vandenbulcke M, Koole M, Bormans G, Brooks DJ, Vandenberghe R (2009) Phase 1 study of the Pittsburgh compound B derivative 18F-flutemetamol in healthy volunteers and patients with probable Alzheimer disease. J Nucl Med 50:1251–1259
Wong DF, Moghekar Abhay R, Rigamonti D, Brasic JR, Rousset O, Willis W, Buckley C, Smith A, Gok B, Sherwin P, Grachev ID (2013) An in vivo evaluation of cerebral cortical amyloid with 18F-flutemetamol using positron emission tomography compared with parietal biopsy samples in living normal pressure hydrocephalus patients. Mol Imaging Biol 15: 230–237
Vandenberghe R, Van Laere K, Ivanoiu A, Salmon E, Bastin C, Triau E, Hasselbalch S, Law I, Andersen A, Korner A, Minthon L, Garraux G, Nelissen N, Bormans G, Buckley C, Owenius R, Thurfjell L, Farrar G, Brooks DJ (2010) 18F-flutemetamol amyloid imaging in Alzheimer disease and mild cognitive impairment: a phase 2 trial. Ann Neurol 68:319–329
Cselenyi Z, Joenhagen ME, Forsberg A, Halldin C, Julin P, Schou M, Johnstroem P, Varnaes K, Svensson S, Farde L (2012) Clinical validation of 18F-AZD4694, an amyloid-β-specific PET radioligand. J Nucl Med 53:415–424
Jureus A, Swahn B-M, Sandell J, Sandell J, Jeppsson F, Johnson AE, Johnstroem P, Neelissen JAM, Sunnemark D, Farde L, Svensson SPS (2010) Characterization of AZD4694, a novel fluorinated Aβ plaque neuroimaging PET radioligand. J Neurochem 114:784–794
Li Z, Conti PS (2010) Radiopharmaceutical chemistry for positron emission tomography. Adv Drug Deliv Rev 62:1031–1051
Riss PJ, Ferrari V, Brichard L, Burke P, Smith R, Aigbirhio FI (2012) Direct, nucleophilic radiosynthesis of [18F]-trifluoroalkyl tosylates: improved labelling procedures. Org Biomol Chem 10:6980–6986
Moon BS, Kil HS, Park JH, Kim JS, Park J, Chi DY, Lee BC, Kim SE (2011) Facile aromatic radiofluorination of [18F]-flumazenil from diaryliodonium slats with evaluation of their stability and selectivity. Org Biomol Chem 9:8346–8355
Lee E, Kamlet AS, Powers DC, Neumann CN, Boursalian GB, Furuya T, Choi DC, Hooker JM, Ritter T (2011) A fluoride-derived electrophilic late-stage fluorination reagent for PET imaging. Science 334:639–642
Liu W, Huang X, Cheng M-J, Nielsen RJ, Goddard WA III, Groves JT (2012) Oxidative aliphatic C-H fluorination with fluoride ion catalyzed by a manganese porphyrin. Science 337:1322–1325
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Meanwell, N.A., Eastman, K.J., Gillis, E.P. (2014). Tactical Applications of Fluorine in Drug Design and Development. In: Nenajdenko, V. (eds) Fluorine in Heterocyclic Chemistry Volume 1. Springer, Cham. https://doi.org/10.1007/978-3-319-04346-3_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-04346-3_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-04345-6
Online ISBN: 978-3-319-04346-3
eBook Packages: Chemistry and Materials ScienceChemistry and Material Science (R0)