Transition Metal Complexes with Amino Acids, Peptides and Carbohydrates in Catalytic Asymmetric Synthesis: A Short Review

Abstract

The present review is devoted to the application of transition metal complexes with such ligands as amino acids, peptides and carbohydrates in catalysis. The literature published over the past 20 years is surveyed. Among the distinctive features of these ligands are their versatility, optical activity, stability and availability. Furthermore, depending on the specific synthetic task to be solved, these ligands open up almost infinite opportunity for modification. Largely thanks to their multifaceted reactivity, transition metal complexes with amino acids, peptides and carbohydrates can catalyze most of the known chemical reactions affording optically pure compounds. In this review, the emphasis is placed upon C(sp³)–H activation, cross-coupling and hydrogenation (including traditional hydrogenation in the presence of hydrogen gas and hydrogenation with hydrogen transfer) reactions. The choice is not accidental, since these reactions on the one hand display the catalytic versatility of the above complexes, and on the other hand, they are widely employed in industry.

1. Introduction

Nowadays, the chemistry of organometallic compounds is a diversified field of science that combines synthetic methodology, the investigation of the structure and reactivity of metal compounds with various ligands and the search for ways to practically apply them. Almost any molecule capable of coordinating with the transition metal atom(s) can be used as a ligand, provided that the target complex is retained as an independent unit even in solution, although it may undergo a partial dissociation. The dawn of this field of chemistry dates back to the end of the eighteenth century, but only the development of physical–chemical research methods in the beginning of the 1950s, which allowed us to study the structure of organometallic compounds in detail, gave impetus to rapid progress in this branch of knowledge. The practical application of EPR, NMR, IR and UV spectroscopy, as well as various types of polarimetry that allow us to conduct polarization-optical measurements provided deeper insights into the nature of chemical, optical, photo, biological, etc., activities of organometallic compounds. In addition, it is worth mentioning the virtually invaluable contribution of the Schlenk methodology to the development of this field of chemistry [1,2,3], since many organometallic compounds can exist only in an atmosphere of inert gases. Such a feature, for example, is inherent in a number of precursors to catalytically active centers [4,5]. The investigations of such structures are of paramount priority both for the theory of fundamental catalysis and the industrial application of metal complex catalysts.

Over the last decade, particular emphasis has been placed on metal complexes based on optically pure ligands [6,7,8,9]. On the one hand, such complexes exhibit intriguing biological and photoactive properties [10,11,12,13,14]. On the other hand, they are promising optically active catalysts and/or precursors of catalytically active sites [10,12,15,16,17,18,19,20,21,22]. Usually, a transition metal complex containing a chiral ligand catalyzes the transformation of a pro-chiral substrate to one enantiomer as the major product. The problem here is that designing molecules that correspond to the parameters of optically pure ligands is not always a trivial task. The synthesis of such molecules is often sophisticated and multi-stage. Consequently, the price of the desired product becomes rather high. Some model optically pure ligands were reported [23,24,25,26,27,28].

Therefore, some papers describe the application of natural optically pure, biologically active compounds, namely, amino acids, peptides or carbohydrates as ligands. Their complexes are stable in the absence of a high-purity inert gas atmosphere even under physiological conditions. They also possess other properties of organometallic complexes (optical, reactive, catalytic, photo, etc., activities) that attract the attention of researchers in the field of modern chemistry, physics, biology and medicine [29,30,31]. In addition, since such biologically important molecules are employed not only by chemists, but also by biologists and physicians, and are isolated from natural objects, they are more accessible as model “green” compounds than synthetic optically active congeners. One should also bear in mind the synergism of the properties of both ligands and metal-containing structures that constitute the target transition metal complexes [32].

It should also be stressed that α-amino acids and peptides, due to their multi-functionality, are ranked among the most versatile ligands. They can be bonded with atom(s) of transition metals in two ways. The first involves the coordination of α-amino acid or peptide ligands to the corresponding organometallic fragment of the initial transition metal complex via donor atoms of a functional group (for example, amino-, carboxylate or sulfanyl). In such a fashion, N,O-chelate-like complexes are formed, for example. The second way comprises the binding of amino acid moiety with a metal atom through one or more metal–carbon bonds. α-Amino acids containing the coordinating side chains (e.g., cysteine and histedine) can act as tridentate ligands.

At the same time, it was reported [33] that carbohydrates can be referred to as so-called “Privileged ligands” [24,34,35]. This inference is based on the fact that carbohydrate molecules can be diversely modified. In other words, the steric and electronic factors can be systematically varied via the introduction of various functional groups into the composition of the target product, depending on the specific synthetic task. Most often, the modification is employed in the case of monosaccharide molecules, since firstly, the processes of bond formation between transition metals and sugars are considered reversible in solution [36,37], and secondly, many sugars are able to be transformed in the coordination sphere of transition metals [38,39,40]. The modification of a ligand molecule is understood, first of all, as the introduction of achiral hydrophobic or coordinating fragments to furnish the hydrophobic and/or metal complexing derivatives. However, this strategy can be significantly expanded across the transformations delivering surfactants and food additives [41]. This methodology enables us to overcome the “overfunctionalization” of ligand molecules and also substantially diversify the target complexes. The latter, being chiral substances, are by definition the most important catalysts and auxiliaries for design of diastereomerically pure drugs, taking into account the availability and potential implementation of many industrially relevant processes [38,39,40].

A series of solid research works is devoted to the synthesis and characterization of transition metal complexes with amino acids, peptides or carbohydrates [37,42,43,44,45,46]. Commonly, such complexes are synthesized using the approaches of classical chemistry of non-organometallic compounds. It is notable that the choice of a suitable solvent is determined both by the nature of the initial transition metal complex and free functional groups contained in the amino acid or peptide molecule, and by the peculiarities of the target product. In some cases, however, the physical–chemical characteristics of the chosen system require and/or facilitate the alternative synthetic strategies. This is true, in particular, for the complexes in which the organometallic fragment is attached to the amino acid side chain via s- or p-coordination. The most important contribution to this area of research was probably made by Prof. W. Beck, who published more than 150 papers under the title “Metal Complexes of Biologically Important Ligands” (for example, [47,48,49,50]). Moreover, in 2018, Molecules published a Special Issue, “Metal Complexes of Biological Ligands”, which was edited by Prof. Beck [51]. Comprehensive reviews dedicated to the application of the above complexes remain limited in number. For these reasons, the present concise review of recent scientific literature (published already in the 21st century) will focus on transition metal complexes with amino acids, peptides or carbohydrates as applied to challenges to modern catalysis.

It is hoped that this review will help many researchers engaged in the field of organic catalytic synthesis to unveil the potential hidden in ligand molecules such as α-amino acids, peptides or carbohydrates. It should be noted that a single review, even if it covers the literature published over the last 20 years, cannot exhaustively analyze all the material accumulated in the area under consideration. The choice of papers included in this review is firstly a result of the potential opportunities that the proposed catalysts and/or catalytic systems open up for researchers. Secondly, this is the nature of the transition metal. Thus, the review deals with complexes designed on the basis of both traditionally popular metals in asymmetric catalysis, e.g., Ir, Ru, Rh and Os, and less known metals, such as Cu, Co, Mn and Yb. Of course, the palladium-derived structures are also considered [52]. The choice of such catalytic systems is nonrandom, since on the one hand, it highlights the general trends in the development of modern catalysis, and on the other hand, it evidences that expensive catalysts can be replaced by cheaper analogues. The latter is especially important in terms of potential industrial applications.

In addition, the review covers several different types of transformations which, in my opinion, can significantly contribute to the development of general organic synthesis. For example, in the long run, catalytically promoted enantioselective reactions of C(sp³)–H bond formation under mild conditions will allow the use of minimally functionalized bulk chemicals as the building blocks for the synthesis of complex molecules, thereby shortening synthetic chains, reducing undesired or by-products, and increasing the sustainability of chemical synthesis. Cross-coupling reactions are not only among the most popular reactions in modern organic chemistry due to the variety of the synthesized compounds and the ease of implementation; they are also excellent models for the study of catalytically active centers and the mechanisms of the catalytic act [53,54]. The so-called “transfer hydrogenation reaction” can become an indispensable tool for saturation of the multiple bonds. Notably, realization of such processes does not require the expensive infrastructure for the handling of hydrogen (cylinders, reducers, gas lines, special reaction vessels, etc.) [55]. This significantly simplifies the work of the researcher, especially in the case of “asymmetric transfer hydrogenation”, and also reduces the significant costs of strict safety regulations. The “asymmetric Michael addition reaction” and the “azide–alkyne cycloaddition reaction” are also briefly mentioned in the review to demonstrate that such complex processes can be realized under rather mild conditions.

That is why the data presented in the review will be classified not by the nature of the ligand molecules (that is, α-amino acids, peptides or carbohydrates), but by the catalytic reactions in which these ligands are employed. In addition, in order to achieve better results, in many cases molecules of ligands of one nature such as α-amino acids were modified by molecules of a different nature, such as sugars.

2. Enantioselective Reactions of C(sp³)–H Bond Activation

First, let us consider the processes of enantioselective functionalization of prochiral C–H bonds, since in the long run they can pave the way for the development of various expedient approaches to the synthesis of chiral compounds. Despite the efforts of many research teams, the efficiency of enantioselective reactions of C(sp³)–H bond activation is still insufficient for their wide application in asymmetric synthesis. Currently, the development of efficient chiral metal catalysts that can differentiate prochiral C–H bonds located on the same carbon methylene center through a metal insert remains a challenge. Given the ease of preparation from commercially available and naturally occurring amino acids, as well as their ability to coordinate with transition metals, amino acid derivatives were investigated in the desymmetrization processes of various C–H bonds.

Yu et al. [56] synthesized a Pd(II)/amino acid (monoprotected a-amino acids) complex (1) capable of catalyzing the asymmetric activation of prochiral C(sp²)–H and C(sp³)–H bonds to generate chiral products containing new C–C bonds with excellent enantioselectivity (Scheme 1). This is the first example of enantioselective functionalization of the C–H bond using an amino acid as a chiral ligand. Notably, the most common pyridine was employed as an orienting substituent to direct the asymmetric palladation of the C(sp²)–H bond. It was shown that two coordination centers of an amino acid ligand (carboxylate and mono-protected amine) were crucial for enantiocontrol. The steric repulsion between the bulky chain of N-Boc-isoleucine and the large aryl group of the substrate ensured the excellent enantioselectivity. Moderate yields of the products and enantioselectivity of the reaction were mentioned.

Another example of enantioselective activation of a methylene C(sp³)–H bond accelerated by the Pd(II) complex in the presence of bidentated N-acetyl-protected minomethyl chiral oxazoline (APAO) ligands (2) was documented by Yu et al. [57]:

Data taken from ref. [57].

It was found that new chiral acetyl-protected aminoethylquinoline ligands provided asymmetric introduction of palladium into prochiral C–H bonds on one methylene-carbon center. These palladium complexes were employed to implement the catalytic enantioselective functionalization of β-methylene C–H bonds in aliphatic amides. Furthermore, it was demonstrated [58] that C(sp³)‒H arylation of a wide range of aldehydes and ketones at β or γ positions can be carried out using amino acids as the transient directing groups in the presence of a Pd(OAc)₂ catalyst (3). The enantioselectivity of the C–H bond activation reactions was achieved with a chiral amino acid as a transient directing group:

Data taken from ref. [58].

A series of model aldehydes and ketones with various substituents tolerated the arylation reactions. The yields of the target products and enantiomeric ratios (er) were indicated. In addition, in 2017, Yu et al. [59] reported the first palladium(II)-catalyzed (4) enantioselective arylation of a methylene C(sp³)–H bond via the combination of a weakly coordinating monodentate directing group with a bidentate acetyl protected aminoethylquinolone ligand. A plausible scheme of the catalytically active center and the mechanism of enantioselective arylation were proposed. Also, it was assumed that the described enantioselective reactions of the β-C–H bond activation could provide a versatile platform for the creation of a-chiral centers in asymmetric synthesis:

Data taken from ref. [59].

After the publication of paper [59], much of the researchers’ attention was given to the transition metal-catalyzed enantioselective methylene C(sp³)–H bonds catalyzed by transition metals, especially to the acyclic methylene C(sp³)–H) bond. Amino acids as alternative chiral ligands were also successfully used for the desymmetrization of methylene C(sp³)–H and functionalization of cyclopropyl [60,61] and cyclobutyl [62]. However, it became clear that the methods described above were inefficient for the introduction of palladium (II) into the acyclic methylene C(sp³)–H bond [57]. Recently, Antilla and Kuninobu [63] reported a rare example of amino acid-controlled enantioselective arylation of acyclic methylene in the presence of (5). It was suggested that steric repulsion between the substrate methyl group and N-Boc-2-pentylproline pentyl moiety was the key to increasing the enantioselectivity of this reaction:

Data taken from ref. [63].

The first one-pot oxidative and catalytic enantioselective alkylation of the C(sp³)–H bonds adjacent to the nitrogen atom (Scheme 2), namely the cross-coupling of N-substituted glycine esters with α-substituted β-ketoesters, was described in [64]. The reaction was catalyzed using the system chiral BOX/Cu(II) complex (6) to produce various α-alkyl α-amino acids. The plausible mechanism of the model reaction was proposed. This new strategy was assumed to provide facile, straightforward and environmentally benign access to a variety of optically active α-alkyl α-amino acids and C1-alkylated tetrahydroisoquinoline derivatives.

He et al. [65] described the catalytic enantioselective amidation of biphenyl sulfoxides C(sp²)–H bond in the presence of Ir(III) complex (7) (Scheme 3), consisting of t-butyl cyclopentadienyl ligand and a modified chiral proline. An array of dibenzyl sulfoxides and dioxazolones tolerated the reaction to deliver a variety of densely functionalized sulfoxides bearing synthetically attractive amide groups in good yields and enantioselectivity. Moreover, the derivatization of the amidated sulfoxide gave diverse chiral sulfoxide scaffolds, which can be potentially employed in asymmetric catalysis as chiral bidentate and tridentate ligands:

Summarizing all the above, it may be assumed that the coordination of the chiral nitrogen atom to the metal center is probably crucial for enantiocontrol. However, such an assumption should be supported by the characteristics of transition complexes and the results of kinetic studies. An approach to so-called chiral recognition at the stage of C–H activation was proposed [66]. The approach can open up fresh opportunities for the development of enantioselective reactions of C–H activation with a wide coverage of substrates. Although the catalytic systems described in [66] are organocatalytic, it is logical to suggest that this approach can be extended to a certain extent across organometallic catalytic systems.

3. Cross-Coupling Reactions

Cross-coupling processes catalyzed using transition metal complexes represent a subset of general coupling reactions leading to the formation of carbon–carbon or carbon–heteroatom bonds in different organic molecules [67,68,69]. The awarding of the Nobel Prize to Heck, Negishi and Suzuki in 2010 highlighted the fundamental and applied importance of this type of transformation [70]. At the moment, it can be said that the cross-coupling reaction is one of the most dynamically developing directions of modern organic catalytic synthesis.

Ma et al. [71,72,73,74,75,76,77,78,79,80,81] showed that simple amino acids can serve as ligands for Cu(I) complex-catalyzed (8) mild cross-coupling reactions between aryl halides and various nucleophiles, e.g., amides, carbamates, amines, amino acids, N-heterocycles, sulfinic acids, alkynes and azides, for example:

Data taken from ref. [81].

It should be noted that the results obtained by Ma et al.’s team were, in many respects, superior to those published by Baskin and Wang [82] with (9):

Data taken from ref. [82].

In addition, the application of Cu(I) complexes with amino acids as ligands in coupling reactions of aryl iodides and bromides with aliphatic and aromatic thiols (under mild conditions) were also investigated by Guo et al. [83,84]. The yields of the cross-coupling products with (10) were 90–98% for aryl iodides and 72–85% for aryl bromides:

Data taken from ref. [84].

In continuation of the above research, it was demonstrated [85] that the system Cu(I)/amino acid was a powerful catalyst for N-arylation of sulfonamides with aryl bromides and iodides. The best results were achieved in the presence of 5–20 mol % of CuI as a catalyst, 20 mol % of N-methylglycine (for aryl iodides) (11) or N,N-dimethylglycine (for aryl bromides) (12) as ligands, and K₃PO₄ as a base:

Data taken from ref. [85].

It was stated that the developed cross-coupling ensured significantly higher yields of the products, and at the same time its implementation was much cheaper and more environmentally friendly compared to previous approaches.

Blum et al. [86] established that the Heck and Suzuki coupling reactions can be successfully catalyzed using palladium derivatives (13) of proline, tyrosine, or alanine, both in homogeneous and sol–gel entrapped versions:

Data taken from ref. [86].

It was found that in the systems under consideration, Pd(0) particles were not formed even when the cross-coupling was carried out in boiling mesitylene (161 °C) or over a long period of time (48 h), i.e., there was no heterogenization of the reaction mixtures. Accordingly, the reactions with palladium derivatives of amino acids proceeded at a higher rate than in the presence of Pd(OAc)₂ in silica sol–gel matrices (reference system), which was easily heterogenized under the reaction conditions.

Merola et al. [87] developed a method for the oxidative cross-coupling of phenylboronic acid with olefins catalyzed by nine palladium(II) bis-amino acid chelates with aliphatic ring structures, namely, proline, 4-fluoroproline, 4-hydroxyproline, 2-benzylproline, azetidine-2-carboxylic acid and pipecolinic acid (14):

Data taken from ref. [87].

The enantioselectivity of the studied model reactions was low. In the best case, with cis-bis(prolinato)palladium(II), enantiomeric excess was 24%, enantioselectivity being increased at lower temperatures. Despite the low enantioselectivity, these systems, in opinion of the authors, were suitable enough for the coupling of a variety of both electron-rich and electron-deficient phenylboronic acids with activated and non-activated olefins.

Kederien and Tatibouët et al. [88] showed that methyl 3-amino-1-benzothiophene-2-carboxylate was good precursor for the implementation of modern Ullmann-type cross-coupling reactions. An efficient method for N-arylation of the functionalized aminobenzo[b]thiophene with various (hetero)aryl iodides under mild conditions was elaborated. This reaction was catalyzed by CuI, using L-proline as a N,O donor ligand and Cs₂CO₃ (15) as a base in dioxane:

Data taken from ref. [88].

The yields of the target products ranged from moderate to good. The structures of all synthesized compounds were confirmed using the NMR and HRMS techniques. Notably, L-proline is a relatively inexpensive and readily available promoter, and it can be easily removed from crude reaction mixtures by simply washing with water.

Zhonggao and Xue et al. [89] reported the synthesis of sugar-based (Glu-IMS) imidazolium salts combining glucosidic and imidazolium head groups with diverse substituents. The PdCl₂/Glu-IMS pair (mol:mol = 1:2) turned out to be an efficient catalyst for the Heck–Mizoroki and Suzuki–Miyaura reactions in water. In the presence of such substrates as aryl iodides, bromides and activated aryl chlorides, ~4.0 nm palladium nanoparticles stabilized with Glu-IMS (16) were formed in the reaction system, which can act as real carriers of catalytically active sites:

Data taken from ref. [89].

Four air-stable, easy-to-handle pre-catalysts of the Glu-NHCs-Pd(II)-PEPPSI type were obtained on the basis of glucopyranoside-functionalized N-heterocyclic carbenes (Glu-NHC) [90]. These compounds were synthesized from the corresponding Glu-functionalized imidazolium bromide salts by direct coupling with Pd(OAc)₂ in pyridine.

Data taken from ref. [90].

It is worthwhile to note that a pure β-anomeric PEPPSI complex (with 1,2,3-trimethylbenzene) was obtained by carefully controlling the volume and rigidity of the substituent. It was the first example of a system that controls the absolute configuration of an anomeric carbon through a bridging group using the steric effects of substituents. The catalytic properties of these Glu-NHCs-Pd(II)-PEPPSI complexes (17) in the Suzuki reaction of aryl bromides and activated aryl chlorides (22 examples) in water were also examined. Thus, under the optimized conditions, a number of functional materials with a fluorene core and various aryl substituents were synthesized in excellent yields.

New glycosyl triazole ligands for the C–C coupling of terminal alkynes were documented in [91]. The copper complexes prepared from these ligands in situ (18) proved to be very efficient for the Sonogashira and Glaser couplings in a temperature-dependent competitive manner to give products in high yields:

Data taken from ref. [91].

The latter were achieved at 130 °C for a variety of substrates, including aliphatic and aromatic terminal alkynes, and aromatic halides with different substituents including 9-bromoscapine. In contrast, at room temperature, very low loading of a catalytic system consisting of a copper complex and a glycohybrid triazole ligand can provide excellent yields in Glaser reactions, including homocoupling and heterocoupling of various aliphatic and aromatic alkynes:

Data taken from ref. [91].

4. Asymmetric Transfer Hydrogenation (ATH)

The hydrogen transfer hydrogenation is an attractive alternative to conventional hydrogenation methodology. This method does not require the sophisticated experimental protocols, since in this case a molecule of an organic substance such as a solvent (called a liquid hydrogen carrier) acts as the hydrogen donor, which is transformed into another stable compound during the hydrogen donation. Ideally, this compound does not react with either the initial reactant or the target product. The hydrogen transfer hydrogenation typically occurs at a moderate temperature and pressure in the presence of organic or organometallic catalysts (often chiral), which allows the efficient asymmetric synthesis to be implemented. Over the last decades, due to its higher efficiency, atom economy and sustainable performance, the hydrogen transfer hydrogenation has received much attention both in academia and industry as an alternative to the traditional processes using high pressure hydrogen [22,92,93,94,95,96,97].

Therefore, Merola et al. [98] described amino acid half-sandwich iridium complexes, namely [(η⁵-Cp*^R)Ir(aa)Cl] (19), which appeared to be active in hydrogen transfer hydrogenation.

Data taken from ref. [98,99].

It was shown that the nature of R in Cp*^R dramatically affects both selectivity and the rate of the reaction. In addition, when water was replaced with organic solvent, strong effects of solvents were observed.

Later, the same research team [99] synthesized new complexes Ir(NHC)₂(L)(H)(X) (20), which were active in ATH of ketones. The maximum enantioselectivity was achieved in the presence of the Ir(IMe)₂(L-Pro)(H)(I) Ir(IMe)₂(L-Pro)(H)(I) catalyst, which reduced some acetophenone derivatives with ee up to 95%. The characteristics of these complexes were comparable to those of other well-known ATH catalysts and were superior to the catalysts containing NHC ligands. It was noted that this work outlined the design of highly enantioselective noble metal-based catalysts incorporating inexpensive, “off-the-shelf” chiral amino acids in combination with readily available N-heterocyclic carbenes that provided an alternative to expensive or synthetically complex chiral ligands.

Sun et al. [100] obtained iridium complexes based on chiral aminobenzimidazoles (derivatives of L-valine, L-phenylglycine and L-proline) (21), which were active in the ATH catalysis of quinolines in water or biphasic systems.

Data taken from ref. [100].

These catalysts do not require a protective inert atmosphere, while the initial iridium complexes are water-soluble, which greatly facilitates the experimental work. These ATH catalytic systems can operate with a catalyst loading of 0.001 mol. % (S/C = 100,000, TON = up to 33,000) under mild reaction conditions. The turnover frequency (TOF) value can reach 90,000 h⁻¹. Diverse quinoline and N-heteroaryl compounds were converted into the target products in high yields and up to 99% enantiomeric excess (ee).

Adolfson et al. studied [101,102,103] ruthenium and rhodium catalysts incorporating ligands based on pseudodipeptides (the combination of tert-butoxycarbonyl(N-Boc)-protected a-amino acids and chiral vicinal amino alcohols). The new ligands represent N-Boc-protected α-aminoamides with functional groups of α-amino alcohols. These highly modular ligands were merged with [{RuCl₂(p-cymene)}₂] (22) and the target metal complexes were examined as catalysts for the enantioselective reduction of acetophenone under transfer hydrogenation conditions using 2-propanol as a hydrogen donor. Excellent enantioselectivity (up to 98% ee) was found for 1-phenylethanol:

Data taken from ref. [103].

Although most ligands contained two stereocenters, it was shown that the absolute configuration of the formed alcohol was determined using the configuration of the amino acid counterpart of the ligand. L-amino acids-derived ligands gave the products of the S-configuration products, while d-amino acids-based catalysts allowed for the alcohols of the R-configuration. The combination of N-Boc-1-alanine and (R)-phenylglycinol (Boc-1-Ab) or its enantiomer (N-Boc-d-alanine and (S)-phenylglycinol, Boc-d-Aa) proved to be the most effective for the reduction process. The enantioselectivity of the hydrogen transfer hydrogenation was up to 96% ee.

It was demonstrated that the addition of LiCl to (23) had a positive effect both in terms of enantioselectivity and reaction rate [104]. Thus, it was suggested that the reaction may involve the transfer of hydride ion and lithium ion to the substrate and the realization of the six-membered transition state shown in Scheme 4a.

These pseudodipeptide ligands (24) were checked in ruthenium-catalyzed ATH of propargyl ketones [105] (Scheme 4b). In this case, the addition of LiCl had no effect on the catalytic process. The same was observed in tandem alkylation/ATH of acetophenones with primary alcohols [105], tandem isomerization/ATG of allyl alcohols [106] and ATH of heteroaromatic ketones [107]. Based on these results, a formal synthesis of the antidepressants (R)-fluoxetine and (S)-duloxetine in high yields (up to 94%) and excellent enantioselectivities (>99% ee) of the resulting 1,3-amino alcohols was carried out.

Diéguez, Adolfsson et al. [108] found that the introduction of a furanoside amino sugar moiety into the backbone of N-Boc-protected a-amino acids ligands (25) was a very efficient route for the transfer of chiral information to products (ee ranged from 98% to >99% during the reduction of a range of ketones) (see Scheme 5a). Unlike the above pseudodipeptides, enantioselectivity was exclusively controlled by the sugar moiety, which allows inexpensive achiral or racemic derivatives of α-amino acids to be employed. It was mentioned that due to the modular construction of the carbohydrate-based ligands, the structural diversity is readily achieved, so in the long run activity and enantioselectivity can be maximized with respect to other substrates as required.

This approach was extended to other ligands designed from pseudodipeptides or thioamides and carbohydrates containing no -CHOH fragments, which were also tested in ATH of aryl- and alkyl ketones (ee up to >99%) catalyzed by the corresponding ruthenium and rhodium complexes (26):

Data taken from ref. [111].

These ligands were easily synthesized from inexpensive D-glucose and D-xylose. The effect of the substituents in the ligand (positions of amino acid/thioamide groups either at C-5 or C-3 atoms of the furanoside fragment, etc.) on the catalytic activity was evaluated [109]. In addition, half-sandwich ruthenium(II), rhodium(III) and iridium(III) complexes of four different methyl 2,3-diamino-4,6-O-benzylidene-2,3-dideoxy-α-D-hexopyranosides and methyl 2-amino-3-tosylamido-α-D-glucopyranoside were described. These complexes were used as precursors of acetophenone ATH. The highest ee (68%) was obtained with a ruthenium catalyst [110].

Adolfsson et al. synthesized a number of thioamide ligands based on amino acids functionalized with 1,2,3-triazole [111,112]. The ligands of this series, when combined with dimeric [RhCp*Cl₂]₂ (27), proved to be good catalysts for ATH of aryl-,alkyl ketones in the presence of sodium isopropylate and lithium chloride in 2-propanol (Scheme 6). The conversion of the corresponding alcohols varied from good to high at 93% enantioselectivity.

It was stated [113] that the ligands depicted in Scheme 6 can be named the model ligands of the first generation. The second generation of ligands includes thioamides obtained via the reaction of D-glucose derivatives and various α-amino acids (Scheme 6). Rhodium catalysts based on the second generation of ligands promote ATH of aryl-, alkyl-, and heteroaromatic ketones with excellent enantioselectivity [113].

The third generation of ligands consists of amides and thioamides obtained from α-amino acids and D-mannose [114] (28). It was highlighted that the modular nature of the ligands enabled the easy and systematic variation of several parameters of the ligand so that activity and enantioselectivity could be maximized for each substrate as required (Scheme 7). By careful selection of the ligand components, excellent enantioselectivity (typically ranging from 95% to >99%) was achieved for a wide range of ketones, namely aryl-, alkyl-, alkyl-, heteroaryl and unsaturated ketones, as well as in tandem isomerization/ATH reactions (rhodium and ruthenium catalysts, up to 99% ee.) and in tandem alkylation/ATG of acetophenones and acetylpyridine (ruthenium catalyst, up to 92% ee).

Apart from α-amino acids-derived ruthenium, rhodium and iridium complexes, which catalyze ATH reactions, catalytic systems based on L-α-amino carboxylate osmium complexes [115], [(η⁶-p-MeC₆H₄iPr)Os(Aa)-Cl] (29), were reported. The elimination of chloride afforded the corresponding cationic trimmers, [{(η⁶-p-MeC₆H₄iPr)Os(Aa)}₃]³⁺, which were isolated as tetrafluoroborate salts. Synthetic procedures and characterization of the obtained products, including the determination of their absolute configuration, were described in detail. It was shown that the metal and the nitrogen atom (for aminocarboxylates with secondary or tertiary amino groups) acted as the stereogenic centers. Osmium was found to exist in both configurations, and the configuration of nitrogen was induced by the configuration of the initial L-α-amino acid carbon. Both neutral mononuclear compounds and cationic trimers catalyzed ATH of 2-propanol to ketones with ee up to 82%. The achieved enantiodifferentiation was explained by the Noyori bifunctional mechanism (Scheme 8).

The synthesis and characterization of carbohydrate–NHC Ru complexes was documented [116]. It was revealed that deprotection of the acetylated carbohydrate unit on the ruthenium complex can be accomplished in situ under alkaline conditions in protic solvents without affecting the Ru-triazolilidene bond. Thus, the first unprotected carbohydrate–NHC system with Ru (30) was achieved. This method provides access to carbene complexes containing hydroxyl functional groups. Carbohydrate functionality has a strong effect on the catalytic activity in the transfer hydrogenation of ketones. Glucose, directly bound to triazolilidene, significantly enhances the TOF of the catalytic system compared to more remote carbohydrate functionalization, though it reduces activity compared to non-functionalized carbene complexes. This prominent role of the carbohydrate substituent enables further development of catalysts toward other transformations involving hydrogen transfer, such as oxidation processes or hydrogen borrowing processes, as well as by modulating the carbohydrate moiety using other pyranoses or furanoses to exploit their useful properties in catalysis.

Scheme 8. Monomeric (a) and trimeric (b) arene-osmium catalyst for ATH of ketones Scheme was taken from ref. [115,117].

Scheme 8. Monomeric (a) and trimeric (b) arene-osmium catalyst for ATH of ketones Scheme was taken from ref. [115,117].

All the above ATH catalytic systems contain noble metals, the reserves of which on Earth are limited and are, therefore, very expensive. On the one hand, high values of enantioselectivity and yields can justify the high price of such catalytic systems, but on the other hand, a cheaper alternative is desirable.

In this line, Sun et al. [118] synthesized a series of Mn(I) complexes with chiral bidentate benzimidazole ligands (31) from readily available amino acids. These chiral Mn(I) catalysts demonstrated high activity and enantioselectivity in the ATH of a broad range of ketone substrates:

Data taken from ref. [119].

A tentative mechanism for the reaction was proposed on the basis of experimental and DFT studies. In addition, DFT calculations rendered a plausible model of enantiocontrol in ketone hydrogenation, in which the π–π stacking interaction between the catalyst and the substrate plays an important role. The data presented in [118] confirmed that Sun et al. developed an alternative approach to the synthesis of chiral aromatic alcohols.

Furthermore, catalytically active transition metal complexes are known to be synthesized from sugars, which participate in the classical hydrogenation of unsaturated bonds. The classical model of catalytic hydrogenation involves the reactions that are carried out in the presence of molecular hydrogen and transition metal complexes, which act as precursors of catalytic activity. For example, work was reported [119] on the hydrogenation of various functionalized and non-functionalized olefins (46 examples) in the presence of M-complexes (M = Ir, Rh) synthesized from a large family of phosphite-thioether/selenoether ligands (32)–(33). This family of ligands combines the advantages of carbohydrates, biaryl phosphite and thioether/selenoether moieties. New ligands are not only easy to handle (solid and air stable) but are also easily produced from cheap carbohydrates (L-tartaric acid and D-mannitol) and are readily modified using well-established methods of carbohydrate chemistry. Such modular architectures modularity were critical to obtain the most efficient catalyst (up to 99% ee enantioselectivity) for each type of olefin through the careful choice of ligand parameters. Previously, Prof. Diéguez et al. synthesized pyrrolidine-based phosphine/phosphite-O/S ligands from inexpensive carbohydrates (D-mannose, D-ribose and D-arabinose) and Ir-complexes based on these ligands (34) [120], which showed high enantioselectivity (ee up to 99%) in the hydrogenation of model tri- and disubstituted substrates.

Heck and Böge [121] disclosed the synthesis of novel Noyori-type trans-dichloro(bis-phosphane) ruthenium(II) complexes with four different methyl-2,3-diamino-4,6-O-benzylidene-2,3-dideoxy-α-D-hexopyranosides (35) and their application as precursors of catalysts in the asymmetric hydrogenation of acetophenone. It was found that these complexes demonstrated a remarkable effect of additional chiral diamino ligands on chiral induction. The data obtained are similar to those previously obtained by Noyori et al. for the first-generation catalyst precursors [122].

5. Other Types of Reactions

The asymmetric Michael addition reaction is recognized as one of the most important synthetic methods for the formation of C–C bonds. The Michael reaction represents the addition of a stabilized carbon nucleophile to an α,β-saturated compound bearing an electron-withdrawing group (e.g., nitro or carbonyl) [123]. The metal-catalyzed asymmetric Michael addition is a kind of classical Michael addition, but it affords the reaction products in a desired stereochemistry due to the so-called catalytic control [124]. Over the past two decades, the Michael addition has been widely employed in various areas of modern synthesis, including total synthesis [125]. In some works, water or mixtures of water and an organic solvent were used as the reaction medium. However, two-phase Lewis acid aqueous catalysis is still in its infancy. Meanwhile, such catalysis has huge potential for the development of more economical and environmentally friendly processes. Of particular promise is a methodology involving modification of the ligand for efficient and extremely facile transformation of the catalyst. For example, it was shown [126] that natural α-amino acids accelerated Yb(OTf)₃ (36)—catalyzed Michael addition (Scheme 9). The ligand acceleration factor of 138 was achieved for alanine. The highest enantioselectivity of the reactions was reached when natural α-amino acids were used as ligands for Lewis acids in water. The solubilizing and stabilizing properties of α-amino acids enabled researchers to “heterogenize” the metal catalyst in the aqueous phase and to reuse it many times without a noticeable loss of activity.

Two years later [127], new kinetic data on Yb(OTf)₃/α-amino acid-catalyzed Michael reactions in water were published. In addition, the effect of metal/ligand ratio, pH, temperature and structure of α-amino acids on the yields of the products and the selectivity of the process was evaluated. It was established that the reaction conditions requiring only 5 mol% Lewis acid ensured enantiomeric excess of up to 79% and were applicable to a broader range of donors and acceptors than previously [126]. In other words, this catalytic system may have the potential for large-scale applications because it demonstrates not only large ligand accelerations but also good solubility and stability in water. Moreover, the catalyst can be recycled many times without noticeable loss of activity.

Tiwari et al. [128] described the azide–alkyne cycloaddition reactions in the presence of copper complexes formed from novel pyridyl glycosyl triazole ligands (37). It was shown that such catalytic systems were effectively employed for Ullman coupling followed by additive cyclization to assemble a number of new quinazolinone fragments of pharmacologically relevant compounds in good to excellent yields. A series of model biologically important 2-amino-3-substituted and 3-substituted quinazolinones (14 novel compounds) were synthesized via a one-pot tandem reaction:

Data taken from ref. [129].

A tentative mechanism of assembly was proposed. It should be noted that Sharpless has referred to this cycloaddition as “the cream of the crop” of click chemistry [129] and “the premier example of a click reaction” [130]. Therefore, the further development of the azide–alkyne cycloaddition in the presence of new transition metal complexes, including those formed in situ, is certainly an urgent challenge to overcome.

6. Conclusions

As is evident from the abovementioned, the works covered in this review were performed only by a few research teams over a long period of time. In most cases, the potential carriers of catalytic activity are formed in situ, and the main emphasis is placed on the issues of catalytic activity and enantioselectivity.

Table 1. Description of the catalytically active systems discussed in the review.

Serial Number	Description of the Complex	Graphical Form	Literature
(1)	Pd(OAc)2 and amino acid (L)   formation in situ	Pd(OAc)2–L, L =	[56]
(2)	Pd(OAc)2 (10 mol%) and L (10 mol%) (L = APAO)   formation in situ	Pd(OAc)2–L, L =	[57]
(3)	Pd(OAc)2 (10 mol%) and L (10 mol%) (L = glycine)   formation in situ	Pd(OAc)2–L, L =	[58]
(4)	Pd(OAc)2 (10 mol%) and L (20 mol%) (L = glycine)   formation in situ	Pd(OAc)2–L, L =	[59]
(5)	Pd(OPiv)2 (5 mol%) and L (17.5 mol%) (L = N-Boc-2-pentyl proline)   formation in situ	Pd(OPiv)2–L, L =	[63]
(6)	Cu(OTf)2 (10 mol%) and L (12 mol%)   formation in situ	Cu(OTf)2 –L, L =	[64]
(7)	[Cp*IrIIICl2]2 (5 mol%), [Ag] (20 mol%) and L	[Cp*IrIIICl2]2–L, L =	[65]
(8)	CuI (0.5 mmol), L (1 mmol) and base (10 mmol)   formation in situ	CuI–L, L =	[81]
(9)	(CuOTf)2·PhH (5 mol%) and L (10 mol%) (1,2-Dimethylethylene-diamine)   formation in situ	(CuOTf)2·PhH–L, L =	[82]
(10)	CuI (5 mol%), L (20 mol%) and KOH (12.5 mmol)   formation in situ	CuI–L, L =	[84]
(11)	CuI (5–20 mol%), L (20 mol%) and K3PO4   formation in situ	CuI–L, L = (for aryl iodides)	[85]
(12)	CuI (5–20 mol%), L (20 mol%) and K3PO4   formation in situ	CuI–L, L = (for aryl bromides)	[85]
(13)	Pd(OAc)2 and amino acid = proline, tyrosine, or alanine (L)	Palladium colloids	[86]
(14)	Pd(acac)2 and amino acid (L) Palladium(II) Amino Acid Complexes were pre-synthesized	amino acid = L-proline/N-methyl-L-proline/4-hydroxy-L-proline/trans-4-fluoro-L-proline/2-benzylproline hydrochloride/L-azetidine-2-carboxylic acid/L-pipecolinic acid/D-proline/D-pipecolinic acid	[87]
(15)	CuI (0.05 mmol), L (0.048 mmol) and Cs2CO3 (0.965 mmol)   formation in situ	CuI–L, L =	[88]
(16)	PdCl2 and Glu-IMS, Glu-IMS = Glucoside-based imidazolium salts (combining glu- coside and imidazolium head groups)	~4.0 nm palladium nanoparticles stabilized by Glu-IMS	[89]
(17)	Pd(OAc)2 (1.0 mmol), Glu-NHCs precursor (1.0 mmol), KBr (2.0 mmol) and dry pyridine (3.0 mL) were pre-synthesized	Glu-NHCs-Pd(II)-PEPPSI complex R = methyl, phenyl, 1,2,3-trimethylbenzene and other	[90]
(18)	CuI (2–20 mol%), L (5–20 mol%) and K2CO3 (1–2 equiv.)   formation in situ	CuI–L, L = Glycosyl triazoles	[91]
(19)	[(η5-Cp*R)IrCl2]2, amino acid (L) and NaHCO3   were pre-synthesized	[(η5-Cp*R)IrCl2]2, Cp*R = tetramethyl(phenyl)cyclopentadienyl (Cp*Ph), tetramethyl(benzyl)cyclopentadienyl (Cp*Bn), tetramethyl(2-propyl)cyclopentadienyl (Cp*iPr), or tetra- methyl(cyclohexyl)cyclopentadienyl (Cp*Cy) L = L-alanine, L-phenylglycine, L-phenylalanine, L-proline, L-piperidine-2-carboxylic acid, L-phenylalanine, L-azetidine-2-carboxylic acid, L-piperidine-2-carboxylic acid, glycine, N,N-dimethyl-glycine	[98]
(20)	[Ir(COD)(NHC)2]X and amino acid (L)   were pre-synthesized	X = Cl, I NHC = N-heterocyclic carbene, L = glycine, L-alanine, L-valine, L-phenylglycine, L-azetidine-2-carboxylic acid, L-proline, D-proline, cis-4-fluoro-L-prolin, L-pipecolic acid	[99]
(21)	[Cp*IrCl2]2 (0.3 mmol) and L (0.3 mmol)   formation in situ	[Cp*IrCl2]2–L, L = or , where L1: R1 = t-Bu, R2 = CH3, R3 = H L2: R1 = t-Bu, R2 = Bn, R3 = H L3: R1 = t-Bu, R2 = CH3, R3 = Ph L4: R1 = i-Pr, R2 = CH3, R3 = H L5: R1 = Ph, R2 = CH3, R3 = H	[100]
(22)	[Ru(p-cymene)Cl2]2 (0.25 mol%) and L (0.55 mol%)   formation in situ	[Ru(p-cymene)Cl2]2–L, L =	[103]
(23)	[Ru(p-cymene)Cl2]2 (1 mol%) and L (2.2 mol%)   formation in situ	[Ru(p-cymene)Cl2]2–L, L = Ligand (1): R = R″ = Me, R′ = H Ligand (2): R = R″ = Me, R′ = H	[104]
(24)	[Ru(p-cymene)Cl2]2 (0.5 mol%) and L (1.1 mol%)   formation in situ	[Ru(p-cymene)Cl2]2–L, L =	[105]
(25)	[Ru(p-cymene)Cl2]2 (0.25 mol%) and L (0.55 mol%)   formation in situ	[Ru(p-cymene)Cl2]2–L, L = or	[105]
(26)	[RhCl2Cp*]2 (0.5 mol%) and L (1.1 mol%)   formation in situ	[RhCl2Cp*]2–L, L =	[111]
(27)	[RhCl2Cp*]2 or [IrCl2Cp*]2 (0.5 mol%), L (1.0 mol%) and SDS (10 mol%)   formation in situ	[MCl2Cp*]2–L, L = see Scheme 7, M = Rh or Ir	[114]
(28)	[Ru(p-cymene)Cl2]2 (0.25 mol%) and L (0.55 mol%)   formation in situ	[Ru(p-cymene)Cl2]2–L, L = ,	[115]
(29)	[{(η6-p-CH3C6H4iPr)OsCl}2(μ-Cl)2] (0.63 mmol), L (1.265 mmol) and aqueous KHCO3 (12.6 mL, 0.10 mol)   were pre-synthesized	[{(η6-p-CH3C6H4iPr)OsCl}2(μ-Cl)2], L = R1 = CH3, R2 = H, R3 = H R1 = Bn, R2 = H, R3 = H R1 = Bn, R2 = H, R3 = CH3 R1 = (CH2)2, R2 = (CH2)2, R3 = H R1 = (CH2)3, R2 = (CH2)3, R3 = H R1 = (CH2)4, R2 = (CH2)4, R3 = H R1 = Bn, R2 = CH3, R3 = CH3 R1 = (CH2)3, R2 = (CH2)3, R3 = CH3	[116]
(30)	[Ru(p-cym)Cl2]2 (0.16 mmol) and L (0.43 mmol)   were pre-synthesized	L = N-heterocyclic carbene hybrids The method of ligand (L) synthesis is described in the article [117]	[117]
(31)	Mn(CO)5Br (1 equiv.), L (0.3 mmol) and toluene, stirring at 110 °C for 4 h   were pre-synthesized	The method of ligand (L) synthesis is described in the article [119]	[119]
(32)	[Ir(μ-Cl)(cod)]2 (0.0185 mmol), L (0.037 mmol) and Na[{3,5-(CF3)2C6H3}4B] (0.041 mmol)   were pre-synthesized	[Ir(cod)(L)][{3,5-(CF3)2C6H3}4B], L = R1 = Ph, CH3, t-Bu and other	[120]
(33)	[Rh(cod)2]BF4 (0.05 mmol) and L (0.05 mmol)   were pre-synthesized	[Rh(cod)(L)]BF4, L =	[120]
(34)	[Ir(cod)2][{3,5-(CF3)2C6H3}4B] (0.01 mmol) and L (0.01 mmol)   formation in situ	[Ir(cod)2][{3,5-(CF3)2C6H3}4B]–L, L =	[121]
(35)	[(C6H6)RuCl2]2 (0.096 mmol), dppf (0.19 mmol) and L (0.19 mmol)   were pre-synthesized	[RuCl2(dppf)(L)], L =	[122]
(36)	Yb(OTf)3 (0.1 mmol), L (0.12 mmol) and NaOH (0.12 mmol)   formation in situ	Yb(OTf)3–L, L =	[127]
(37)	CuI (5 mol%), L (5 mol%) and CsCO3 (1.2 equiv.)	CuI–L, L =	[129]

presents the data on the composition of catalytically active systems, which are numbered in the order of their appearance in the text. The relevant references are also provided. Occasionally, attention is paid to the composition and structure of the catalytically active center (CAC) and tentative reaction mechanisms. It can be assumed that such peculiarities of the research in this field are due to the fact that the authors of most of the papers are well-known experts in the field of organic synthesis and their goals are traditionally aimed at achieving the maximum efficiency of the desired transformation of the initial substrate [131]. Meanwhile, the analysis of the CAC nature would allow the prediction of enantioselective regenerated compositions with given catalytic characteristics (activity, productivity and (enantio)selectivity). The development of this research direction is impossible without information about the interaction of the components of the catalytic system with each other, as well as with substrate(s) molecules. Given that the molecules of amino acids, peptides or carbohydrates can be easily modified to afford conformationally rigid molecular structures, it can be assumed that these structures will serve as multipoint-linked, optically active scaffolds for catalytically active centers, including nanosized ones, which will also permit the design enantioselective regenerated systems.

In addition, the number of works devoted to the role of non-precious metal complexes derived from amino acids, peptides or carbohydrates in catalytic organic synthesis should be expectedly increased in the near future. First of all, this is explained by the need to decrease the cost of technology and the intended application of precious metals in those branches of science where they cannot be replaced.

Probably, the combined efforts of researchers engaged in areas such as organic synthesis (namely, the modification of chiral molecules), the synthesis of transition metal complexes with optically pure ligands, the analysis and physical–chemical characterization of these complexes, as well as catalysis (the study of the nature of CAC(s) and the mechanism(s) of its (their) action in various chemical processes) would allow more rational implementation of the research process and interpretation of the data obtained.