Dimeric Cinchona alkaloids

Nature is full of dimeric alkaloids of various types from many plant families, some of them with interesting biological properties. However, dimeric Cinchona alkaloids were not isolated from any species but were products of designed partial chemical synthesis. Although the Cinchona bark is amongst the sources of oldest efficient medicines, the synthetic dimers found most use in the field of asymmetric synthesis. Prominent examples include the Sharpless dihydroxylation and aminohydroxylation ligands, and dimeric phase transfer catalysts. In this article the syntheses of Cinchona alkaloid dimers and oligomers are reviewed, and their structure and applications are outlined. Various synthetic routes exploit reactivity of the alkaloids at the central 9-hydroxyl group, quinuclidine, and quinoline rings, as well as 3-vinyl group. This availability of reactive sites, in combination with a plethora of linker molecules, contributes to the diversity of the products obtained.


Introduction
The term alkaloid is used for many vastly different nitrogen heterocycles of mostly plant origin. Alkaloids are classified according to the heterocycle and the taxonomy of the species they were isolated from. The natural diversity of the alkaloids is further extended by the presence of numerous dimeric alkaloids (Fig. 1). Some dimers appear as byproducts by coupling of a small portion of the monomers (e.g., salutadimerine), while others are the final products of biosynthesis (e.g., cephalostatin). Alkaloid dimers can exhibit biological activities unrelated to that of the corresponding monomer [1], such as in vitro anti-HIV and antimalarial properties of Michelleamine A [2], or fungicidal activity of Bismurrafoline B [3]. Natural alkaloid dimers of different symmetry and heterodimers were isolated (Fig. 1). Additionally, dimers of alkaloids can be synthesized in the laboratory giving rise to a virtually unlimited number of combinations [1].
The bark of various Cinchona species contains four major alkaloids, namely quinine (QN), quinidine (QD), cinchonidine (CD), and cinchonine (CN) (Fig. 2). These quinoline alkaloids are isolated on an industrial scale in multi-ton amounts. Their structures contain a central hydroxyl group as well as quinoline and quinuclidine rings. The individual alkaloids differ in the configuration at two crucial stereogenic centers (C-8 and C-9, Fig. 2). Quinine has been used for nearly four centuries to treat malaria. Although currently it is largely replaced by newer medicines, such as chloroquine (1947) and artemisinin (1970's), its therapeutic use is limited to drug-resistant strains. On the other hand, quinidine is often used to treat certain arrhythmias. Cinchona alkaloids are also employed in enantioselective synthesis (catalysts, ligands) and separation processes (resolving agents, solid phases, assays) [4]. To date, no dimeric alkaloid in this family has been isolated from a natural source. Nevertheless, many synthetic dimers were made exploiting a few reactive sites in the Cinchona alkaloids (Fig. 2).
These synthetic dimers were examined for their biological activities and applicability in asymmetric reactions. For the purposes of medicinal chemistry, the multiplication of the pharmacophore in the dimers could improve binding or cause crosslinking of the biological receptors. The transition from a monomeric to dimeric alkaloid molecule results in accumu- For the purpose of this review, alkaloid derivatives are labeled with the corresponding alkaloid (QN, QD... cf. Fig.  2), 10,11-dihydroalkaloid (DHQN, DHQD…), or 9-epialkaloid descriptor (eQN, eQD, etc.) followed by a consecutive compound number. It has to be emphasized that some derivatization reactions were reported only for a single alkaloid, while others were exercised on a set of Cinchona alkaloids.

Dimers connected at the central C-9 position
The central 9-OH group offers an attractive site for modification (i.e., etherification and esterification reactions). Alternatively, the hydroxyl group can be replaced with a few other groups (e.g., NH 2 ) and subsequently used for dimerization.
In dimeric Cinchona aryl ethers (Fig. 3), the linkers are either electron deficient heterocycles or anthraquinones. All of these compounds were obtained through aromatic nucleophilic substitution. Thus, the alkaloid units are mostly in the para position with the exception for few meta derivatives, but  no ortho-diethers are known. The lack of such products arises from the reactivity of halo-aryls in the nucleophilic substitution, rather than steric interactions, since heavily substituted pyrimidine derivatives were obtained with relative ease.

Phthalazine derivatives
Up to 1992 there was an incremental progress in osmiumcatalyzed asymmetric dihydroxylation reactions, and promising results were obtained with monomeric Cinchona alkaloid derivatives in the role of ligands. Then, the discovery of dimeric phthalazine ether ligands (PHAL, 3) marked an enormous leap for asymmetric synthesis [27,28]. The respective dimers 3 were obtained from reactions of alkaloids with 1,4-dichlorophthalazine (2). The process required basic conditions and azeotropic removal of water with toluene [29]. In an alternative synthesis, the alkaloids were first deprotonated with NaH in DMF and subsequently treated with dichloride 2 [30]. This change in protocol often provided better preparative yields. Although 2 is commercially available, it can be efficiently obtained from phthalhydrazide (1), PCl 5 , and a catalytic amount of DMF (Fig. 5) [29].
Also, a stepwise protocol for the synthesis of unsymmetrical dimers was devised. Equimolar amounts of 1,4dichloro-phthalazine (2) and dihydroalkaloid gave reactive chloroderivative 4 that was used in a subsequent step to 9O-arylate another alkaloid. The resulting quinidine-dihydroquinidine and quinine-dihydroquinine heterodimers 3 had a single vinyl group that was used to anchor the molecule to polymer supports using the radical addition of thiols ( Fig. 6) [31][32][33].
A single reactivity averts crosslinking, and in the cases presented by the authors, also prevents significant distortion of geometry in the parent structure. The same approach was followed to obtain chiral stationary phase additive, by reaction of QD/DHQD-3 with octadecyl mercaptan [33]. The symmetrical phthalazine dimers 3 were also subject to many subsequent derivatization attempts. These include primarily immobilization, for example, direct copolymerization of quinine-based dimer with methacrylates [34], or copolymerization of more reactive alkaloid-derived acrylate DHQN-10 ( Fig. 7) with styrene/divinylbenzene in suspension [35].
Apart from simple alkaloids, also their elaborate derivatives were dimerized with phthalazine [36]. Didehydroquinidine (QD-11, vide infra) was coupled in a Sonogashira reaction with various aryl halides to yield alkaloids with extended carbon scaffold QD-12a-b. Also, 11-iodinated didehydroalkaloid QD-12c was prepared by addition of iodine to the triple bond of didehydro-alkaloid followed by elimination of HI. These three compounds (QD-12a-c) were used to obtain respective phthalazine dimers QD-13a-c (Fig. 8). The yields of the dimerization step were similar to that of unmodified quinidine [37].
Under osmium-catalyzed asymmetric dihydroxylation conditions, the two native vinyl groups in Cinchona dimers QN-3, QD-3 are transformed to the corresponding tetraols. These products catalyze the AD reaction, although arguably [27] their effectiveness is inferior to DHQN-3 and DHQD-3. Nevertheless, the polar character of these compounds was advantageous for reactions carried in special solvents, including ionic liquids, polyethylene glycol (PEG), and water. Thus, even more polar compounds were obtained by mono N 1-quaternization with benzyl [30] or allyl bromide. The N -allyl ammonium salt QD-14 was then directly used in the AD reaction. It was transformed in situ to water soluble ammonium salt DHQD-15 having six hydroxyl groups, which facilitated recycling of the ligand through aqueous extraction ( Fig. 9) [38]. Also, exhaustive quinuclidine Nalkylation was performed on phthalazine dimer DHQD-3. The obtained dimeric quaternary salts were not suitable for AD reactions, but were considered for phase transfer catalysis (PTC) [39].
Dimers of a different architecture were also obtained in a reaction of polyethylene glycols (PEG) with monomeric dihydroquinine ether DHQN-4 (Fig. 10). These immobilized soluble ligands DHQN-16a-c were still successful in aminohydroxylation reactions and could be recycled; however, significant catalyst loading was required [40].
Also, linkers with extended fused ring systems were applied. The synthesis of DHQD-32 by Corey involves two additional modifications. The alkaloid was first modified at the 6 -position of the quinoline ring (vide infra). Prior to coupling, dihydrocupreidine (DHQD-27), a derivative of quinidine with free 6 -hydroxy group, was 6 O-alkylated with a series of secondary alkyl bromides [44]. The reactive linker  molecule 30 was obtained analogously to phthalazine, starting from naphthalenedicarboxylic acid hydrazide and a mixture of PCl 5 and POCl 3 [45]. The coupling of quinidine derivative DHQD-31 with 1,4-dichlorobenzo[g]phthalazine afforded the dimer DHQD-31 in good yield. Subsequent partial N 1-quaternization with methyl iodide concluded the synthesis of this highly diversified structure (Fig. 12). DHQD-32 provided a highly regioselective and enantioselective course of AD reaction of terminal isopropylidene groups in selected terpenoids [44].

Pyridazine, pyrazine, and pyrimidine derivatives
Shortly after introduction of phthalazine-type ligands by Sharpless, Cinchona dimers with two single-ring heteroaromatic linkers were applied by Corey's group [46]. These included pyridazine DHQD-36 (PYDZ) and pyrazine spacers DHQD-41. The coupling was accomplished by refluxing the respective 3,6-or 2,5-dichloro heterocycles 35 and 40 with dihydroquinidine in toluene in the presence of a base and the azeotropic removal of water. 3,6-Dichloropyridazine (35) not only is commercially available, but can also be obtained in a short and efficient synthesis [47]. On the other hand 2,5dichloropyrazine (40) is more challenging to obtain (Fig. 13) [48,49].
Similarly to PHAL-dimers, a few modifications to the original PYDZ structure 36 were made. Double tethered derivatives of 36 are presented in the last section of this article. Dimer 36 was also partially quaternized with 9anthracenylmethyl group [50]. A pyridazine linker substituted with a short alkyl chain flanked with a terminal alkyne group was also obtained. The functionalized reactive dichloroheterocycle 44 was obtained via the sequential Diels-Alder and retro-Diels-Alder reactions of dichlorotetrazine (42) and 1,7-octadiyne (43) in one pot. After the coupling of 44 with dihydroquinidine the terminal alkyne group of the dimer DHDQ-45 remained reactive in the copper-catalyzed Huisgen 1,3-dipolar "click" cycloaddition (CuAAC) [51]. Thus 1,2,3-triazoles were obtained with various azides including small molecules [52] and polymers (Fig. 14) [51].
Pyrimidine-based dimers 51 (PYR) constitute another important group of ligands, particularly useful in AD of  The synthesis again relied on refluxing the dichloroheterocycle 50 with the alkaloid in the presence of a base in toluene and the azeotropic removal of water (Fig. 15). The reactive dichloride 50 was obtained in a two-step procedure starting from the condensation of adequately substituted diethyl malonate 47 and amidine 48. Subsequent reaction with either POCl 3 or PCl 5 gave the required reactive intermediate 50. An important feature of the pyrimidine scaffold is that 2-and 5-substituted derivatives are often easily accessible. Sharpless obtained dimers with pyrimidine linkers substituted at position 2 and 5 with combinations of phenyl and tert-butyl groups [53]. However, groups with greater steric demands at position 2 impede the formation of the dimer, and only monomeric alkaloid derivatives could be obtained from 2-CEt 3 -5-tBu-substituted pyrimidine [54]. In later reports, more differently 2-and 5-substituted and unsubstituted pyrimidine dimers were mentioned [55]. The diversity of the products was further enhanced in a synthesis of several 2-aryl substituted dimers. The commercially available 4,6-dichloro-2-methylthio-5-phenylpyrimidine (52) reacted with a series of arylboronic acids in a Suzuki-type reaction. The obtained intermediates with quinine provided the dimers QN-51a-d in very good yields ( Fig. 15) [56]. Also, an analogue of DHQN-51 substituted at the 2-position of the pyridine with 3,4,5-trimethoxyphenyl group was specifically designed for AD step in a synthesis of a natural product [57]. It was established that for applications in AD the presence of 2-tert-butyl is detrimental, while substitutions at 5-position are more tolerated [53]. Nevertheless, such tuning of the catalyst structure with bulky groups improved its performance in an asymmetric Feist-Bénary reaction [55,56].
Also, a related spacer with 1,3,4-triazine core was exploited. The synthesis was based on the reaction of inexpensive cyanuric chloride (53) with aniline to replace one of the reactive chlorides. Then, dichloride 55 was reacted with the prepared in situ quinine sodium salt in THF to provide the respective dimer QN-56 in nearly quantitative yield. Although the authors used only 4-bromoaniline (54), they proposed that a diverse array of products could be obtained using different aniline or amine derivatives. Interestingly, an excess of quinine sodium salt with cyanuric chloride gave C3-symmetric trimeric derivative QN-57 (Fig. 16).  However, only dimer 56 showed promise in AD reactions [58].

Anthraquinone derivatives
Apart from the heterocyclic spacers, also the anthraquinone unit was extensively studied. 1,4-Difluroanthraquinone (60) was obtained in the Friedel-Crafts reaction of phthalic anhydride with p-difluorobenzene. For the coupling, dihydroalkaloid was converted in situ into a lithium salt with butyllithium, and then a reaction with difluorocompound 60 yielded the anthraquinone dimers (AQN, DHQN-61, DHQD-61) in very good yield (Fig. 17) [59]. These ligands are superior in AD of alkenes with aliphatic substituents.
Similarly to the phthalazine ligands, also a stepwise synthesis was devised. Consecutive reactions of 60 were carried out with alkaloid sodium salts in DMF. This approach, although lower in yield, allowed for the synthesis of heterodimers namely, alkaloid-dihydroalkaloid pair QN/ DHQN-61 suitable for immobilization ( Fig. 18) [60].

9-Alkyl ethers
Dimeric alkaloid alkyl ethers constitute a much less studied group of compounds. Their synthesis is, however, straightforward and involves the Williamson etherification of an alkali metal alkaloid salt and the respective alkyl dihalide. For xylylene-linked dimers, all three isomers: ortho [65], meta [66], and para [67,68] were reported or mentioned in the literature. The shortest known link was made  with E-butene, still in very good yield [68]. A trimeric ether QD-77 was also obtained in an analogous reaction with 1,3,5-tris(bromomethyl)benzene (76) (Fig. 20) [69]. Some of the dimeric ethers were converted to polymeric quaternary ammonium salts (vide infra) with a series of bis(bromomethyl)arenes and served as effective PTC catalysts [68].
Alkyl ethers are much less suitable ligands for the asymmetric dihydroxylation reaction in comparison to the corresponding aryl ethers. For example, a dimer with p-xylylene linker (DHQD-74) gave merely 22 %ee in AD reactions where aryl ethers provided excellent enantioselectivity [67]. Nevertheless they showed promise in organocatalytic applications such as asymmetric aminooxygenation of oxindoles (meta-xylylene linker) [66].

9-Ester-linked dimers
The dimeric esters of alkaloids have the longest history of the presented groups of compounds. The carbonic acid diester is known from the patent literature dating to the end of the XIX century (German patent No. DE105666, 1898). In a later published work, the carbonate CD-79 was obtained in a reaction of excess of cinchonidine with a carefully controlled amount (0.25 equiv) of phosgene (Fig. 21). When the amount of phosgene was increased, an unreactive byproduct, identified then as alkaloid chloroformate, was formed. On the other hand, the carbonate CD-79 was reported to decompose in water [70].
Esters of dicarboxylic acids are generally more stable. They were most often obtained in a reaction of dicarboxylic  acid chlorides and the alkaloids. Usually the transformation was carried in the presence of a base such as triethylamine [71] and sometimes with catalytic amounts of 4dimethylaminopyridine (DMAP). The yields, when reported, were above 70 %. The dimers were also obtained in a one-pot procedure, where the starting diacids were first transformed to the corresponding chlorides with thionyl or oxalyl chloride and subsequently coupled with the alkaloids. Alternatively the dicarboxylic acids were activated with a carbodiimide (e.g., EDC); however, this milder method often resulted in poor yields [72]. Following one of these general methods, a relatively large array of dimeric esters was synthesized (Figs. 22,23). Links were formed from simple aliphatic diacids with 3-10 carbon atoms [71,72], as well as those with ether (German patent No. DE237450, 1909) and disulfide bonds [72,73].
A few of the esters were tested in AD reaction. Some, like the ferrocene-linked dimer DHQN-91 provided only moderate enantioselectivity (<61 %ee). In the group of simple diesters, hexadionate DHQN-81b turned out an effective catalyst (92 %ee) in contrast to malonate. The most attractive catalytic properties were found for the terephthalate ligand (DHQN-86, >98 %ee), which in AD of certain substrates outperformed the classic PHAL-type ligand DHQN-3 [71,78]. On the other hand, isophthalic ester, while still effective in AD, provided worse enantioselectivity than terephthalates in all the cases studied [74]. In a later study, improved results in AD and AA reactions were achieved for isophthalates and a series of analogous pyridine-linked dimers 89 [79].
Diversity in the linker structure was also introduced by subsequent modifications of an already dimeric molecule. In the reaction of fumaroyl dichloride with cinchonine and cinchonidine, the respective dimeric esters CN-96 andCD-  [80]. The reactivity of the activated double bond was further exploited in a Diels-Alder reaction with cyclopentadiene and isoprene (Fig. 26). Consequently, a set of cyclohexene-and bicyclic dicarboxylic acid esters 97 and 98 were obtained. These were, however, not viewed as target compounds, instead Cinchona alkaloids were used as chiral auxiliaries for the Diels-Alder reaction, and the esters were cleaved afterward. The transformations using dimers provided much improved enantioselectivity compared to that of monomeric esters also used in the study (94-99 % vs. 6-93 %ee) [80]. Interestingly, an inverted sequence of Diels-Alder and acylation reactions was also explored. The enantiomeric diacids, adducts of anthracene and fumaric acid, were converted to acid dichlorides and reacted with dihydrocinchonidine. The products were assayed in asymmetric dihydroxylation of stilbene giving 52-85 %ee. Better results were obtained for ester of 11R,12R configuration DHCN-94b than the 11S,12S diasteromer DHCN-94a (Fig. 24) [81].
Another strategy for the synthesis of dimeric esters of Cinchona alkaloids was to first obtain monomeric esters with linker precursors that could be coupled later. In one such approach esters with a terminal alkyne group 99a-d and azido-esters 100a-d of various chain length were formed by carbodiimide coupling. Using these components a combinatorial library of dimers 101aa-dd was obtained in a copper-catalyzed 1,3-dipolar 'click' cycloaddition ( Fig. 27). In products 101 the linker units are non-symmetric. Since the esterification step is performed separately, well-defined heterodimers could also be obtained. The products were studied for inhibition of P-glycoprotein, and optimum linker length was established at 6 methylene groups at each side of the triazole unit (as in QN-101cc) [82].
Few examples of dimers with more elaborate and functional linkers relevant to supramolecular and coordination chemistry were obtained. One such scaffold incorporated a chiral crown ether. Ether 102 with two hydroxymethyl groups was modified with phthalic anhydride, and the resulting diacid was converted to acid chloride 104. Subsequent reaction of 104 with cinchonine gave dimeric ester CN-105 in 78 % yield (Fig. 28). Although the authors saw potential in the product for phase transfer catalysis, they did not develop the idea further [83].
In a study on the derivatization of annulenes, a one-pot procedure to introduce one or two alkoxycarbonyl groups to nickel dibenzotetraaza [14]annulene complex (106) was developed. The sequence of reactions involved Friedel-Crafts acylation with oxalyl chloride, followed by decarbonylation and alkoxydehydrohalogenation [84]. The chirality of the product was assured with nonracemic alcohols, including quinine. The respective dimeric ester QN-107 was formed in 20 % yield (Fig. 29). This result is similar to that obtained for other explored alcohols [85].
A series of dimeric carbamates was obtained according to two general protocols [73,86]. In the first one, the alkaloid was treated with diisocyanate derived from the corresponding diamine. This process was highly efficient furnishing the dimers in 64-87 % yield; however, it was only attempted for the commercially available diisocyanates. Alternatively, a two-step procedure was used: First the alkaloid was converted to an active carbonic ester QN-111 in a reaction with nitrophenyl chloroformate (110). Then, reaction of an excess   (Fig. 30). In both approaches the lowest yields were noted for 1,2-diamine derivatives [86].
Some of the Cinchona carbamates were subsequently immobilized on silica (through addition of thiols to the vinyl group) and used as chiral solid phases for anion exchange chromatography of amino acids. The use of dimers, compared to monomeric carbamates, led to longer retention times of the analytes, however, offered no improvement in the enantioselectivity of the separation. Out of the dimeric modifiers of silica gel, the best separation of enantiomers was achieved using the 1,3-adamantyl-linked QN-114 [86].
The esters and carbamates of Cinchona alkaloids were also examined as antimalarial agents. They were tested in vivo against drug-resistant Plasmodium falciparum and for the inhibition of PfCRT CQR -a multidrug resistance transporter protein. Dimers linked with (CH 2 ) 8 ester QN-80e and (CH 2 ) 6 carbamate QN-112d (both with the same number of bonds separating two quinine units) turned out to be the most effective [73]. The 9-hydroxyl group in the Cinchona alkaloids can be replaced with an amino group with inversion of configuration at C-9 [87]. The alkaloid undergoes a Mitsunobu reaction with an azide source (e.g., HN3, diphenylphosphoryl azide) to give 9-epi-azido-alkaloid 116. This azide is subsequently reduced with triphenylphosphine to amine 117 (Staudinger reduction). This sequence was often followed as a one-pot procedure. Alternatively, the azido-alkaloid 116 can be obtained in an S N 2 reaction from alkaloid methanesulfonate and NaN 3 in DMF. The primary amino group of the 9-epi-9-deoxy-9-aminoalkaloid 117 was then subjected to reactions with various activated linker molecules producing a series of derivatives including amides, ureas, imines, etc. Reaction of oxalyl and isophthalic acid dichlorides with 9-aminocinchonine eCN-117 gave the respective dimers eCN-118 and eCN-119 [88]. Dimeric amide eCN-120 was obtained in a reaction of aminoalkaloid and sebacic acid applying HATU, a standard peptide coupling reagent (Fig. 31) [73]. The oxalyldiamide eCN-118 was tested as catalyst in the diethylzinc addition to benzaldehyde providing only moderate enantioselectivity, though the ee was improved compared to monomeric amide analogs [88].
Without an additive, the conformer populations were equal. However, the ratio varied in response to carboxylic and dicarboxylic acids [90]. The prevalent conformation and their equilibria were studied by circular dichroism (CD) [89] and 1 H NMR spectroscopy. In addition, an analytical system composed of eQD-124 and bromophenol blue allowed for visual identification of α-hydroxycarboxylic acids, as well as their spectrophotometric estimation. This indicator displacement technique allowed for the determination of tartaric acid in wine as well as some differentiation of enantiomers [91].
Amides of squaric acid (1,2-dihydroxy-cyclobuten-3,4dione) with Cinchona alkaloid units were recently shown to be effective hydrogen bond donors in organocatalysis [98]. The synthesis of dimeric squaramides was straightforward. It entailed mixing 9-aminoalkaloid 117 and squaric acid dimethyl ester (135a) in methanol for 24h, while the products 136 precipitated in nearly quantitative yields (Fig. 37) [99]. The dimers were exploited in dynamic kinetic resolution of azalactones. Unlike the monomeric Cinchona squaramides, the dimers do not form self-associates [99]. Nevertheless, these two classes of compounds have different application scope, and often unsymmetrically substituted monomeric squaramides were superior catalysts [100].      eCN-139 and eCN-140 (Fig. 38). First, 9-aminocinchonine eCN-117 was treated with an equivalent amount of diethyl squarate (135b), and the intermediate monoester-monoamide eCN-137 was treated with 0.5 equivalents of m-xylylenediamine to give a C2 symmetric analog of monomeric squaramides. Analogous C3-symmetric trimeric compounds were obtained in 81-87 % yield, when instead of diamine, the reactive eCN-137 intermediate was treated with 0.32 equivalents of selected triamines [101]. The trimers eCN-139 and eCN-140 were effective catalysts in the asymmetric Michael addition of 1,3-dicarbonyl compounds to nitrostyrenes [101] and Friedel-Crafts alkylation reactions [102].    with lithium or butyllithium in THF (Fig. 40). The same isomer of the product (9R) was obtained from both 9R-and 9Shalides: QN-145 and eQN-145, respectively. The most likely explanation for this process was the transient formation of a 9-radical (either direct, or through oxidation of carbanion), followed by radical recombination. The dimers QN-147 and DHQN-147 are sterically hindered, thus anti and syn atropisomers (arising from rotation of the quinoline ring, Fig. 41) were separated. The rotational barrier was estimated at ca. 23 kcal/mol in solution [106]. The reaction of organomagnesium reagents with 9-haloquinine [107] was shown to result in a stereoconvergent Würtz-type coupling (the similarity of this  process with the direct dimerization to QN-147 is only superficial, since the products had different configuration at C-9) [108]. The extension of this approach with divalent Grignard reagent 148 gave the respective dimer eQN-149 in moderate yield (Fig. 40) [106].
Also, a set of dimeric 9-disulfides was obtained in a sequence of Mitsunobu, reduction, and oxidation reactions. Seven different alkaloids including CD, QN, QD, DHQN, DHQD, as well as 9-epi-quinine (eQN), and 9-epi-quinidine (eQD) reacted giving the corresponding dimers 152 in 24-45 % yield with inversion of configuration. It is noteworthy that the oxidative dimerization step was the least demanding. The authors also regarded dimerization as a means for transient protection of the thiol group in 151, which could be cleanly regenerated with LiAlH 4 (Fig. 42) [109].

N1-Qarternary ammonium salts
Quaternary ammonium salts of Cinchona alkaloids were applied in the 1980's in the asymmetric synthesis under PTC conditions providing decent level of enantioselection. C 2 -symmetric quaternary ammonium salts derived from binaphthalene emerged as even more effective catalysts in 1999 [110]. Shortly thereafter, highly efficient C2-symmetric dimeric Cinchona alkaloid quaternary ammonium salts were developed [111]. Park and Jew obtained dimers CD-168a, CD-153a, CD-169a in a direct reaction of excess cinchonidine with ortho, meta, and para-xylylene dibromides, respectively, in a mixture of solvents (DMF/EtOH/CHCl 3 , 6:5:2 v/v). The obtained dimers were then alkylated at the 9hydroxyl group with allyl bromide. The preparative yields were in the range of 90-94 % over two steps (Fig. 43) [111].
The dimeric quaternary ammonium salts 153, 155, and 169 were also converted to ionic polymers (e.g., CD-184a) with adequate disulfonates 183a-i. The obtained materials were insoluble in water and most organic solvents (Fig. 46). Nonetheless, these polymers were effective in asymmetric transformations and were easily recovered from the reaction mixture [118,119].
Internal quaternary ammonium salts (i.e., betaines) where both cationic and anionic centers are present within a sin-gle molecule are also known. They differ in acid-base properties with the compounds described previously, in which there is no covalent bond between the oppositely charged species. The Gong group obtained a set of zwitterionic dimers 187 with a binaphthophenolate linker starting from all major Cinchona alkaloids and two axial enantiomers of BINOL [120]. The reactive MOM-protected 3,3 -bis(bromomethyl)-BINOL derivative 185 was obtained in 5 steps from commercially available BINOL in 68 % yield [121]. The reaction of enantiomeric (aR)-dibromide 185 with the alkaloids afforded the respective dimeric ammonium salts 186 in 64-74 % yields. The coupling was slightly more efficient for the alkaloids of quinine configuration (i.e., 8S,9R). Then, the protecting MOM groups were removed producing betaines 187 in 68-75 % yield (Fig. 47). Diastereomeric dimer (aS)-QD-187 was obtained from quinidine and (aS)-BINOL derivative (aS)-185 in overall 70 % yield. The authors also obtained species QD-188 with a net-positive charge, using monomethyl BINOL ether (Fig. 48). The obtained dimeric betaines 187-188 were tested as organocatalysts in Mannich reaction of azalactones and aliphatic imines. Dimeric betaine (aR)-QD-187 was particularly successful providing 96-98 %ee [120].
Few dimeric Cinchona alkaloid quaternary ammonium salts incorporate functional linkers relevant to supramolecu-    was subsequently used for double N 1-quaternization of both cinchonine (94 %) and cinchonidine (87 %). The obtained dimeric products 190a were subsequently 9O-allylated to give 190b in high yield (Fig. 49). In the original paper there are, however, some discrepancies between the reported spectral data and structures of 189-190 [122].
Polyethylene glycol (PEG2000) was also used in the role of a linker [125]. For this purpose, PEG was converted to a reactive intermediate 194 by introduction of terminal   [126,127]. A subsequent reaction with the alkaloids (CD, QN, and CN) in refluxing chloroform for 4 days afforded the respective dimers 195 (Fig. 51) [125].
Quaternary ammonium salts of Cinchona alkaloids were most often employed in asymmetric phase transfer catalysis (PTC), which is useful, for example, for the synthesis of nonracemic amino acids [110,132]. Phenylalanine derivatives can be obtained through enantioselective benzylation of glycine benzophenone imine esters under PTC conditions. This reaction serves as a benchmark for various catalysts (Table 1) [133].
The first generation of Park's catalyst CD-153b, with meta-xylylene linker applied in the PTC benzylation of glycine imine outperformed the monomeric Cinchona cat-alysts both in terms of enantioselectivity and reactivity. On the other hand, application of the isomeric dimer with ortho-xylylene linker gave poor ees. The allylation of the 9-hydroxyl in the catalysts often significantly improved the enantioselectivity in the PTC transformation [111]. Study of analogues of CD-153 revealed that 2-fluorine atom in the xylylene linker as well as 10,11-hydrogenated alkaloid unit further improved the enantioselectivity [112]. A highly hindered trimer CD-196b provided high enantioselectivity (94 % ee at −20 • C) at a cost of reactivity [128] The second generation Park's catalyst DHCD-170b with 2,7-naphthyl link was one of the most efficient and enantioselective catalysts for the benzylation of glycine imine under PTC conditions providing 97 %ee at 0 • C and 1 %mol loading [114]. This catalyst is commercially available and can be acquired from major reagent suppliers (Fig. 54).
Functional linkers in the dimeric quaternary ammonium salts, in some cases improved their application scope or facilitated the recycling of the catalyst. The macrocyclic dimer CD-190b was claimed to be more suitable for PTC reactions carried in low base concentration [122]. Incorporation of PEG within the linker facilitated recovery of the PTC epoxidation catalyst CD-195 [125]. Unfortunately, in many of the studied transformations the replacement of cinchonidine with pseudoenantiomeric cinchonine units resulted in more or less noticeably lower enantioselectivity and conversion. Another major concern in asymmetric catalysis is that the success of a particular catalyst structure in one reaction (e.g., PTC alkylation of glycine imine) does not necessarily translate to other asymmetric processes. For example, epoxidation of enones required a catalyst with free 9-hydroxyl group in the alkaloid unit, and the best results were obtained for quinine analogue QN-156a. Moreover, this epoxidation failed to proceed enantioselectively when using analogous monomeric catalysts [113]. Cyanation of aldehydes gave good enantioselectivities with CD-155a [134] and so did the Mannich reaction of azalactones catalyzed by QD-187 (up to 99 %ee for adequately modified reactants) (Fig. 55) [187].

3-Vinyl group
The vinyl group of Cinchona alkaloids is an attractive site of derivatization and was often used to couple the alkaloid (and even their dimers) to solid support. This was often achieved by a 'click' thiol-ene radical addition. The corresponding reaction of dithiols (butane-1,4-dithiol, and 2-mercaptoethyl ether) and cinchonidine led to dimeric products CD-200ab in fair yields [135]. The dimers were subsequently polymerized by tethering at the quinuclidine nitrogen atom (vide supra) producing an array of polymers CD-201 (Fig. 56). These were assayed in the asymmetric benzylation of glycine imine under PTC conditions providing 71-88 %ee.
Similar polymers, tethered alternatively by quinuclidine nitrogen and vinyl groups, were developed using the Mizoroki-Heck reaction at the alkaloid vinyl groups. First, the dimer CD-202 was obtained in a palladium-catalyzed reaction of cinchonidine and 4,4 -diiodobiphenyl and subsequently was polymerized by quaternization with 4,4bis(chloromethyl)biphenyl. Polymers CD-203 were effective in the benzylation of N -diphenylmethylene glycine tert-butyl ester. Surprisingly, when using an inverted reaction sequence, i.e., polymerization of the dimeric quaternary ammonium salt CD-179a in a Heck reaction, the obtained polymeric material CD-203 had superior catalytic qualities (Fig. 57) [136].  Cinchonidine dimer CD-205 tethered at the vinyl groups with silicon linker was also synthesized. First the 9-hydroxyl group of cinchonidine was transiently protected with TMS ether. Then, hydrosilylation of the ether CD-204 with bivalent silane (1,2-bis(dimethylsilyl)ethane) was performed applying Karstedt's catalyst. The coupling gave dimer DHCD-205 in moderate yield, similarly to reactions of CD-204 with other bulky mono-silanes. Finally, the TMS group was removed giving DHCD-206 (Fig. 58). The dimer CD-206 was applied for modification of Pt/Al 2 O 3 catalyst surface for asymmetric hydrogenation of ethyl pyruvate and phenylpropanedione. Unfortunately the enantioselectivity was lower than for unmodified cinchonidine (62 vs. 84 %ee) [137].
The terminal vinyl group is also reactive in alkene metathesis reactions. This approach was used to obtain a series of dimeric phosphite ligands derived from BINOL and Cinchona alkaloids. The phosphite esters QN-208, QD-208, and CN-208 were obtained in a one-pot reaction from BINOL (of both aR and aS configurations) and the corresponding alkaloids. The metathesis reaction of 208 with Hoveyda-Grubbs catalyst produced dimers 209 which were isolated in fair yields (Fig. 59). The authors also obtained libraries containing heterodimers by combining different stereoisomers of the alkaloid phosphites 208 in the metathesis reaction. Dimers 209 along with crude metathesis mixtures were subsequently used in asymmetric iridium-catalyzed asymmetric hydrogenation of α,β-unsaturated carboxylic acids. Out of the studied compounds, the dimer (aR)-QN-209 provided up to 92 %ee and outperformed the initial monomer ( ee 8-32 %) [138].
The vinyl group in the alkaloids can also be converted to a terminal alkyne by bromine addition followed by two HBr elimination reactions [139,140]. The alkynes QN-11 and QD-11 (didehydroalkaloids) were then directly coupled in a Sonogashira-type oxidation reaction. With iodine as an oxidant, the yields were good (71-72 %) and further improved when 9O-acetylated didehydroalkaloids 210 were used (86-95 %; Fig. 60) [36].
Direct dimerization of 210 using typical Glaser coupling (copper(I) salt/O 2 ) gave merely 15 % yield [36]. In contrast, a similar homocoupling of didehydroalkaloid quaternary ammonium salt QN-212 with copper(II) salt (Elington reaction) proceeded smoothly giving dimer QN-213 in 89 % yield. Also, Sonogashira coupling of the alkaloid alkyne QN-212 with aryl diiodides gave dimers 215a-b with phenylene and biphenylene linkers, respectively. Unfortunately, the authors found that the dimerization cannot be performed as a one-pot reaction, instead a stepwise protocol had to be followed using excess of reagents at each coupling step: First QN-212 was coupled with 1,4-diiodobenzene or 4,4diiodobiphenyl giving iodoaryl derivatives QN-214a and QN-214b, respectively. The subsequent coupling with QN-212 required significant catalyst loading only to proceed in low yields (Fig. 61). The products QN-213 and QN-215a-b were tested in the asymmetric aldol reaction under PTC conditions. Optimum performance was reported for the phenylene-linked dimer QN-215a, although only moderate enantioselectivity was achieved [141].
Nitration of the quinoline ring of dihydrocinchonine and dihydrocinchonidine occurs preferentially at the 8 position. Subsequent reduction of the nitro group with hydrazine on palladium catalyst yielded primary aromatic amine DHCD-222. The reactions of 8 -aminoalkaloid with several dicarboxylic acid chlorides yielded the corresponding dimers DHCD-225a-f in good to excellent yields. Similar reactiv- Both 6 -ethers DHQN-217a-f and 8 -amides DHCD-225a-f were assayed for inhibition of various strains of Plasmodium falciparum. In vitro tests showed that dimers containing an 8-methylene unit linker were most effective. The amide DHCD-225d (IC 50 0.02-0.05 μM) was more active than ether DHQN-217c (IC 50 0.08-0.26 μM). Unfortunately, in vivo study of DHCD-225d revealed lower activity compared to chloroquine and pronounced toxicity [144].
The 6 -methoxyquinoline ring of Cinchona alkaloids was also partially hydrogenated to give tetrahydroquinoline. The reaction for both QN and QD led to mixtures of 4 -epimers which were separated by 2-3 recrystallizations of mandelic acid salts (8-15 % yield for pure stereoisomers of 228). The secondary amines 228 were then coupled with glutaconic aldehyde enolate to give polymetine dyes 229 (Fig.  65). The dyes exhibited interesting chiral optical properties: markedly high specific rotation and maximum absorbance at 506-511 nm. The rotatory power was mostly dependent on the configuration at the 4 center, as exemplified by the [α] D (c0.01, EtOH) values of −1545 and +1135 for (4 R)-and (4 S)-DHQN-229, respectively [147].
The quinoline ring of the alkaloids is also susceptible to Grignard reagents addition at the 2 and 4 positions. In nonetheral solvents the Grignard adds at the 4 -position, and consecutively the deprotonated 9-hydroxyl group adds at the   Consequently, dimers QN-230 and CD-231 were obtained in 11-12 % yield in a reaction of divalent 1,4-phenylene and 4,4 -biphenylene Grignard reagents, respectively (Fig. 66). The formation of monomeric products was avoided by using an excess of the less reactive organomagnesium reagents [148].

Double-bridged dimers
There are only few reported cases where the Cinchona alkaloid units are connected simultaneously at two distinct sites using two independent linkers. These include linear polymers 201, 203 as well as copolymerized compound 10. However, multiple connections between just two alkaloid units would result in a cyclic product of restricted conformation. In studies conducted to explain the stereoselectivity of the asymmetric dihydroxylation, two such structures DHQD-235 and DHQN-237 were described by the groups of Corey and Lohray, respectively. Both used a pyridazine linker attached with 9-ether bond and additionally tethered the dimer by either the vinyl group (235) [149] or the quinoline ring (237) [150]. The 9,11-tethered structure was prepared in a sequence of reactions starting from a Brown hydroboration of quinidine at the vinyl group, followed by protection of the primary 11-hydroxyl group with the triisopropylsilyl (TIPS) group. The protected alcohol DHQD-232 was then coupled using dichloropyridazine (35), and the silyl ether was cleaved to form the dimeric 11-diol DHQD-234. In the last step, esterification of the dimer DHQD-234 with adipoyl chloride gave the bridged dimer DHQD-235, as confirmed by X-ray, in a total of 9.7 % yield over 7 steps (Fig. 67) [149,151].
Asymmetric dihydroxylation reaction using 9,11-double tethered dimer DHQD-235 provided enantioselectivity similar to AD with classic PYDZ-ligand (DHQD-36), and even outperformed it for some olefins. However, the additional 6 -link in DHQN-237 resulted in much lower level of enantioselectivity (32-72 %ee vs. 88-99 %ee). Rowan and Sanders studied dynamic self-organization processes of modified Cinchona alkaloids possessing both a reactive ester and hydroxyl functionalities. Instead of careful stepwise tethering of the alkaloid molecules, they built a dynamic library of cyclic oligomeric alkaloids. They hydroborinated the 3-vinyl group, and the resulting 11-alcohol 239 was oxidized to the corresponding acid using the Jones reagent and then esterified to give 241 [152]. In an alternative approach, the 11-alcohol 239 was converted to a halide and etherified with methyl 4-hydroxybenzoate to give 245 [153]. The transestrification of 241 and 245 was carried with catalytic potassium methoxide and 18-crown-6. Under these conditions dynamic libraries of cyclic dimers, trimers, tetramers and acyclic products were formed (Fig. 69). The composition of these mixtures depended on both the linker and the alkaloid configuration. For quinine and cinchonidinederived species, the cyclic trimer was the predominant (QN-247b, CD-247b) or nearly the exclusive product (QN-246b, CD-246b) that could be isolated upon crystallization [154]. In the case of quinidine, cyclic dimers were major products both with (QD-247a) and without a p-hydroxybenzoic spacer (QD-246a). In mixtures containing multiple types  Experiments toward the total synthesis of the Cinchona alkaloids led to the formation of an unexpected dimer. One of the steps in Jankowski's synthesis was the aldol condensation of 3-quinuclidinone (248) and quinoline-4-carbaldehyde (249). When this reaction was performed under high pressure conditions, in addition to the desired product 250, a dimeric hemiacetal 251 was formed. The dimer 251 did not dissolve in any of the usual solvents, but heating a suspension in methanol resulted in a cleavage of the hemiacetal bonds to form hydroxyketone 250 (Fig. 70) [155]. The structure of 251 was suggested based on MALDI-MS and IR data, and the relative configuration was not determined.

Closing remarks
There are many ways of combining multiple Cinchona alkaloid units into dimeric and oligomeric species, few avoid the use of linkers, and even highly hindered dimers can be formed. Most of the dimeric products were a result of planned stepwise syntheses of target molecules; nevertheless, methods of combinatorial chemistry were applied, including formation of dynamic libraries. The most exploited points of chemical diversity involve the central 9 position (Sharplesstype ligands and organocatalysts) and the quinuclidine N-1 nitrogen atom (phase transfer catalysts).
When compared to a monomeric analogue, the dimer has twice the number of reactive polar or electrically charged   ., 3, 51,  61, 170).
The dimeric Cinchona alkaloids are often more basic than their monomeric analogs and were more retained during their chromatographical isolation. Apparently, the same phenomenon seems to cause longer retention times on the solid phases modified with dimeric alkaloids.
In medicinal chemistry, the interaction of biomolecules with a tethered dimer may induce functional changes that could contribute to novel pharmacological properties. Few of the studies indicate that biological activity (cytotoxicity, transporter inhibition and antiplasmodial activity) can be found among the Cinchona dimers.