Synthesis of Unsymmetrical Vicinal Diamines via Directed Hydroamination

Abstract

Vicinal diamines are a common motif found in biologically active molecules. The hydroamination of allyl amine derivatives is a powerful approach for the synthesis of substituted 1,2-diamines. Herein, the rhodium-catalyzed hydroamination of primary and secondary allylic amines using diverse amine nucleophiles, including primary, secondary, acyclic, and cyclic aliphatic amines to access a wide range of unsymmetrical vicinal diamines is presented. The utility of this methodology is further demonstrated through the rapid synthesis of several methylated bioactive molecules.

Graphical Abstract

Vicinal diamines are important building blocks in organic syntheses due to their ubiquity in medicines, ligands in transition-metal catalysis, and organocatalysts (Figure 1a).¹ These amines, often derived from ethylene or propylene diamine, can have important effects on the human body.² Interestingly, the minor change from a 1,2-ethylene diamine to a 1,2-propylene diamine has been correlated to a stark increase in drug potency. This is referred to as the “magic methyl effect”.³ Despite their importance, the preparation of these high-value-added materials and methyl analogues are still very challenging.⁴

Figure 1.

Bioactive molecules containing 1,2-diamines and the proposed hydroamination reaction.

Allyl amine-directed hydroamination is an attractive alternative route to 1,2-diamine synthesis due to the ease of coupling two readily accessible functional groups – amines and olefins – with 100% atom economy (Figure 1b).5^,6 This approach utilizes a Lewis basic amine moiety tethered to the olefin substrate, which serves as a directing group to achieve the desired reactivity with high chemo-and regioselectivities. However, the existing protocols uniformly require the use of secondary allylic amines or imines as directing groups and secondary amines or hydroxylamines as coupling partners.⁵ Herein, we present the general, efficient, stereoselective and atom-economical hydroamination of primary and secondary allylic amines with primary or secondary amine nucleophiles that allows direct access to a wide variety of 1,2-unsymmetrical diamines.

Previously, we reported the directed hydroamination of allyl imines and secondary amines to afford 1,2-diamines in a regio- and chemoselective fashion.^5a.,b Lewis basic directing group is essential to the reaction; it coordinates to the catalyst, increases the effective concentration of the olefin, thus facilitating reactivity, and promoting the formation of a metalacyclic intermediate and the formation of a single regioisomer. Upon aminometalation of the coordinated olefin, the resulting alkylrhodium(I) species is locked in a five-membered rhodacycle, the conformational rigidity of which prevents β-hydride elimination. Moreover, the dramatic difference in the rate of formation between 4- and 5-membered metalacyclic intermediates ensures a high degree of regioselectivity for the 5-membered metallacycle and, thus, the Markovnikov product. We hypothesized, therefore, that primary allylic amines may also participate in this type of directed hydroamination via a similar mechanism. To test this hypothesis, we examined the hydroamination reaction between 1 (1.0 equiv) and morpholine (1.2 equiv) under our previously reported reaction conditions:^5a [(DPEPhos)Rh(COD)]BF₄ (2.0 mol %), MeCN (4.0 M), 60 °C, 24 h under N₂. Promisingly, the desired hydroamination product 2 is formed in 45% yield and >20:1 dr along with a 13% of 2” (eq 1). The stereoselective generation of the trans-1,2-diamine is consistent with the formation of the less hindered trans-rhodacyclic intermediate I. Although morpholine is more nucleophilic than either 1 or 2, competitive homo- and cross hydroamination reactions with the primary amine of 1 or 2 serving as the nucleophile afforded 2’ and 2”, respectively. To circumvent these undesirable pathways, we envisaged that a lower reaction temperature may suppress the formation of these side products. Gratifyingly, the desired hydroamination product 2 was obtained in high yield (89%) at the room temperature using a moderate excess (2.0 equiv) of morpholine.⁵ Reducing to 1.2 equivalents affords 2 in synthetically useful 55% yield (>20:1 dr) (Table 1). Notably, only the trans diastereomer is observed under both conditions.⁷ Excitingly, when enantiomerically enriched 1 was employed the no epimerization was observed, suggesting that rhodium catalyst does not promote isomerization of the allylic position.

Table 1.

Hydroamination of primary allyl amines.^a

^aN-allyl amine (1.0 equiv), nucleophile (2.0 equiv), [(DPEphos)Rh(COD)]⁺BF₄^– (2.0 mol %), MeCN (4.0 M), room temperature;

^bnucleophile (1.2 equiv);

^cnucleophile (4.0 equiv);

^d[(DPEphos)Rh(COD)]⁺BF₄^– (4.0 mol %);

^e60 °C;

^f40 °C;

^gnucleophile (6.0 equiv);

^hnucleophile (5.0 equiv);

ⁱ[(DPEphos)Rh(COD)]⁺BF₄^– (5.0 mol %);

^j50 °C;

^kReductive aminations were performed with the corresponding aldehydes prior to isolation;

^lnucleophile (3.0 equiv);

^mMeCN (2.0 M), 48 h;

ⁿ[(DPEphos)Rh(COD)]⁺BF₄^– (7.5 mol %)’

^oMeCN (0.80 M), 72 h.

Next the scope of the primary amine directed hydroamination reaction was explored. Electronically diverse aryl and heteroaryl rings were well tolerated to afford 2–10 in good to excellent yields (70–89%) and excellent diastereoselectivities (>20:1). Interestingly, substrates bearing electron-rich substituents (5 and 6) and bromide (8) on para-position of the aromatic ring led to catalyst decomposition and low conversion was observed. Although high consumption of the allylic amine was observed at higher reaction temperatures, lower yields of the desired hydroamination products were obtained due to simultaneous increase in the rates of the aforementioned hydroamination side reactions. Instead, we found that increasing catalyst loading or equivalents of nucleophile was essential for achieving to higher yields of the desired products, 5 or 6/8, respectively. In addition to aryl-substituted allyl amines, a substrate bearing aliphatic substituent underwent the desired reaction smoothly, while a moderate reduction in diasteroselectivity was observed as phenethyl substituted 11 is formed in 73% yield and 13:1 dr. Impressively, an unsubstituted allyl amine can be employed to provide the desired hydroamination product 12 in 42% yield. The reaction requires a longer reaction time to afford a synthetically useful yield presumably due to the lack of Thorpe-Ingold effect, as after 48 hours 12 was formed in 72% isolated yield. Gladly, the reduced reactivity can be restored by increasing both the loading of the amine and temperature (12). Finally, we also observed a moderate 38% yield of 13 when 2-methylallyl amine is subjected to the reaction conditions.

Next, the scope of this methodology was explored with respect to the participating amine nucleophiles (Table 1, 14–28). Under the reaction conditions, pyrrolidine and piperidine both participate in the reaction to afford the products 14 and 15 good yields. As we observed with morpholine, the reaction is more efficient with 2.0 equivalents of nucleophiles, but a synthetically useful yield (52%) of 14 is obtained with only 1.2 equivalents of the secondary amine. Excitingly, a variety of acyclic secondary amines can be employed at 60 °C with higher loading of both the nucleophile and catalyst to generate 16 and 17 in 74% and 33% yield, respectively. However, more sterically encumbered acyclic secondary amines, such as diethyl amine, do not afford any of the desired hydroamination products likely due to their decreased nucleophilicity. We next investigated more complex and synthetically relevant nucleophiles. Impressively, 1-(2-hydroxyethy)piperazine can be used without protection of the free hydroxyl group. Furthermore, one-pot modification could afford 18 and 19 in good yields, suggesting that these hydroamination products can be easily modified without purification. The drug-like nucleophiles such as 1-pyrimidylpiperazine and norquetiapine successfully underwent the hydroamination reaction with reasonable stoichiometry (2.0–3.0 equiv) to provide synthetically useful yields (42–44%) of 20 and 21.

Despite their reduced nucleophilicity, primary amines are excellent coupling partners under reoptimized reaction conditions using higher loadings of the nucleophile and catalyst as well as increased reaction temperature. Aliphatic amines, including butyl amine, benzyl amine, iso-propyl amine, and cyclohexyl amine all afford the desired 1,2-diamine product in fair to good yields (22-25) Even sterically hindered primary amines were successfully utilized as nucleophiles, albeit giving lower yields, compared to those with n-butyl amine (24 and 26 vs 22). The allylic trisubstituted olefin of geranyl amine remained unaffected under reaction conditions (28), suggesting reactions are sensitive to steric bulk at nucleophile and olefin.

While the scope of this method is promising, the participation of secondary allylic amines, as an olefin partner, was of interest. Previously, we have demonstrated the rhodium-catalyzed asymmetric hydroamination of secondary allylic amines with 2,2’-bis(diphenylphosphino)biphenyl (BIPHEP) derivatives as ligands. However, the use of precious racemic BIPHEP ligand or two-steps protocol from allylic imine was required to obtain a racemic sample.^5b More importantly, sterically bulky substrates, such as cyclohexyl allyl amine (29), were not compatible under the reaction conditions. We contemplated, therefore, if a common bidentate phosphine ligand might be applicable for expanding a scope of directed hydroamination of secondary allylic amines. Thus, we investigated a variety of common bidentate ligands in the presence of sterically bulky 29 as a model substrate. The conditions optimized for primary amines, using DPEphos as the ligand, gave the desired 1,2-diamine albeit in about 3% yield; the major side product observed from these reaction conditions was α,β-unsaturated imine (30c) which could form via aldol condensation between an imine (30a) and an enamine (30b).⁷ Presumably, 30a and 30b form via a well-precedented Rh(I) catalyzed 1,3-hydrogen shift.⁸ Interestingly, a ligand screen revealed that the bite angle correlates with the chemoselectivity of these reactions: large bite angles favor 30c (Table 2, entries 1,2), mixtures of 30 and 30c are formed with intermediate bite angles (Table 2, entries 3–5), and, fortunately, small bite angles favor hydroamination product 30 (Table 2, entries 6,7). Under re-optimized reaction conditions, using DME is as the solvent, 30 is afforded in 78% in situ and 70% isolated yield (Table 2, entry 8).

Table 2.

Optimization of the hydroamination of secondary amine 29 with morpholine.^a

entry	ligand	βnb	solvent	Yield (30)c	Yield (30c)c
1	DPEphos	102 °	MeCN	< 5%	33%
2	Xantphos	111°	MeCN	< 5%	23%
3	dppb	98°	MeCN	25%	20%
4	dppf	96°	MeCN	39%	23%
5	dppp	91°	MeCN	38%	15%
6	dppe	85°	MeCN	21%	< 5%
7	dppbz	83°	MeCN	14%	< 5%
8	dppe	85°	DME	78% (70%d)	<5%
9	dppp	91°	DME	47%	10%

^aUnless otherwise noted, all reaction were conducted using 0.2 mmol of 29 in degassed solvent for 24 h;

^bnatural bite angle (β_n) are taken from ref.12;

^cIn situ yields determined by GC analysis of the crude reaction mixtures using 1-methylnaphthalene as an internal standard;

^disolated yield.

Next, the reaction scope was evaluated with a variety of secondary amine nucleophiles (Table 3). Secondary, cyclic amines, such as morpholine and pyrrolidine gave the desired 1,2-diamine in good yields (30–31). Drug-like nucleophiles, such as 1-pyrimidylpiperazine, affords the desired product 32 in a moderate yield (41%). Gratifyingly, other sterically hindered substates such as an adjacent cycloheptylring, were tolerated under these conditions as 33 is formed in 75% isolate yield. Unfortunately, more hindered substrates, such as 1-adamantylamine, did not afford the desired hydroamination product 34. However, this strategy allows direct access to sterically less crowded products when dppb is used as the ligand (35-41). It is worth noting that when enantiopure secondary allyl amine was employed, we observed the decrease in the stereochemical fidelity to 79% (35). Tolerance for potentially reactive functional groups such as electron rich arenes, aryl bromides, thiophenes, electron-rich alkenes was demonstrated (36-39). Moreover, the scope of nucleophiles that are amenable to less bulky substrate is likewise quite broad (36, 40, and 41). Unfortunately, despite the extensive exploration of reaction conditions, primary amine nucleophiles did not afford any of the desired hydroamination product with secondary allyl amines. This is likely due to the primary amine preferentially coordinating to the cationic Rh catalyst over the more hindered secondary amine, thus preventing the secondary amine from directing the hydroamination reaction.

Table 3.

Secondary Amine directing group scope^a

^aN-allyl amine (1.0 equiv), nucleophile (5.0 equiv), [Rh(COD)₂]⁺BF₄^– (5.0 mol %), dppe (5.0 mol %), DME (4.0 M), 60 °C, 24 h;

^bYields determined by ¹H NMR analysis of the crude reaction mixtures using 1-methylnaphthalene as an internal standard;

^cdppb;

^dWhen starting with enantiomerically pure allyl amine.

We next applied our methodology to the formal synthesis of biologically relevant molecules. The two step, formal synthesis of amnesia-reversal agent 43 was accomplished using our directed hydroamination reaction which allows access the key intermediate 42 in a single catalytic step.⁹ We were also interested in synthesizing GSK1018921, a potent GlyT1 inhibitor that has been shown to be effective at treating schizophrenia by increasing brain glycine level.^2e We were able to access desmethyl-GSK1018921 (45) from 44, accessed in 84% yield in a single catalytic step, which provides access to a new and rapid formal synthesis of 45.^2e Unfortunately, however, our directed hydroamination reaction failed to afford key intermediate 44’ for the synthesis of GSK1018921 (45’) presumably due to the steric clash between the methyl group and the phenyl ring in the requisite rhodacycle preventing its formation. It is worthwhile to emphasize that this method allows for the incorporation of a wide range of different nitrogen nucleophiles and thus would allow for rapid analogue synthesis.

We were interested in synthesis of methyl analogue of biologically relevant molecules because of its synthetic challenge and potential improvements in the activity or pharmacological properties due to the “magic methyl effect”.³ As we have shown in Scheme 1, the key intermediate 12 of methyl analogous of Moclobemide could be easily obtained with our methodology via coupling between allylamine and morpholine. Subsequent acylation of 12 afforded methyl analogue of Moclobemide (46) in a good 77% yield. This suggests that methyl analogs of Metoclopramide and Procainamide can be rapidly formed via similar strategy. Moreover, our strategy proved general for the synthesis of both regioisomers of biologically relevant 1,2-propylene diamines. For example, 47 and 48, methylated SQ109 derivatives, were synthesized from 26 and 28 in good yields.

Scheme 1.

Synthetic Application

In summary, a highly regio-, chemo-, and diastereoselective intermolecular, rhodium-catalyzed hydroamination reaction is reported. The olefin substrate scope has been significantly improved relative to imine/amine hydroaminations that were previously reported. Primary and secondary N-allyl amines function as excellent directing groups for this reaction. Gratifyingly, the nucleophile scope was substantially expanded to secondary, acyclic amines and primary amines, representing one of the limited hydroamination reactions that enables the use of such nucleophiles. Moreover, our directed hydroamination method can be perceived as a powerful strategy to access biologically relevant 1,2-unsymmetrical diamines. Finally, synthetic applications of this methodology have been demonstrated as a excellent approach to access biologically relevant molecules.