Table of Contents

Functional group manipulation continues to play a critical role in the field of synthetic organic chemistry (1). In any total synthesis, achieving a desired end target can only be brought about through careful planning and execution of numerous chemical reactions. The sole purpose of many of these reactions is to selectively alter a functional group, in order to make it amenable toward further use or simply to achieve the final desired moiety. Therefore, a tremendous effort is continually put forth to develop more efficient strategies and reagents, which can be used to alter functional groups (1).

For example, reducing tertiary amides to the corresponding aldehydes is generally difficult, requiring the use of extremely powerful reagents that may limit the use of tertiary amides in total synthesis (1). This is unfortunate, because many reactions can be carried out with the use of the amide functionality. Ortho-lithiation, which uses an amide as the ortho-directing group, works ideally with N,N-diethylamides. The limitation of this reaction, however, is the residual presence of the amide group when the reaction is complete. To remove this amide, stringent conditions must be used, which are often too harsh for the remaining functionalities in any newly constructed complex molecule. Any new procedure that would remove the amide under mild conditions would have a potential positive impact in the field of organic synthesis.

The Schwartz reagent

To reduce tertiary amides, our research group focused on bis(cyclopentadienyl)zirconium chloride hydride (Cp₂ZrHCl) (2). This reagent was initially developed by Jeffrey Schwartz and has been named after him (3).

The Schwartz reagent has seen continued development and increased use for functional group transformation (4–6). It was initially developed for the process of hydrozirconation, in which the addition of Cp₂ZrHCl to alkenes and alkynes leads to alkylzirconium and alkenylzirconium products, which can then undergo useful coupling or alkyl transfer reactions (for an example, see Figure 1). The reaction is most cleanly achieved by heating olefins or acetylenes with Cp₂ZrHCl. The reaction proceeds to give the hydrozirconated product in generally high yields (4–6).

The utility of organozirconium compounds was further illustrated with the development of palladium- and nickel- catalyzed cross-coupling procedures, which allowed further carbon–carbon bond formation reactions to proceed effectively (7). In addition, the relatively low electronegativity of zirconium permits the replacement of the zirconium metal with other metals such as aluminum, zinc, and tin via transmetallation, which further expands its chemistry. Furthermore, Zr(II) is an excellent reducing reagent (8).

More recently, Ganem and co-workers used the reducing potential of Cp₂ZrHCl effectively for the deoxygenation of -unsaturated organic substrates (9) (Figure 2).

Figure 2. Examples of the reducing potential of Cp2ZrHCI. -Ketoesters are reduced to ,-unsaturated esters and amides to N-substituted imines. TMS = Trimethylsilyl.

In this application, -ketoesters are reduced to the corresponding ,-unsaturated esters, and secondary amides are transformed to N-substituted imines. The latter process is important for functional group manipulation because it is difficult to control the reduction of amides to imines.

Reduction to aldehydes

Our research group found serendipitously that tertiary amides can be reduced to aldehydes with Cp₂ZrHC1. Our original impetus stemmed from attempts to generate an aromatic palladocycle, which could then react with an electrophile such as an acid halide (Figure 3, top). This would then regiospecifically provide an aryl ketone. Although we knew from the literature that this reaction proceeds by using groups such as amines, Schiff bases, and azo compounds, which can direct a five-membered palladium complex within an aromatic system (10, 11), we wanted to examine the potential use of N-methoxy-N-methylamides (see box, “Weinreb amides”). By using this amide, we could take advantage of the lone pair of electrons on oxygen to form a six-membered palladium complex within the aromatic system (Figure 3, bottom).

This reaction did not work, so we tried a transmetallation reaction of zirconium (using the Schwartz reagent) on the palladium complex. We initially thought that this might generate a more reactive species, with zirconium on the corresponding aromatic system, leading to an enhanced ability to accept an electrophile (Figure 3, bottom).

This reaction did not work either, and, in retrospect, the Schwartz reagent was not a wise choice for transmetallation because of the low electronegativity of zirconium. However, we noted a surprising result from the reaction: The sole product from this attempted reaction appeared to be the corresponding aromatic aldehyde. Therefore, what actually occurred was that the initial addition of lithium tetrachloropalladate to the reaction mixture acted simply as a spectator, whereas the subsequent treatment with Cp₂ZrHCl reduced the corresponding Weinreb amide to an aldehyde.

Following this surprising lead, we found that a variety of Weinreb amides, as well as other tertiary amides, were successfully reduced to the corresponding aldehydes using only Cp₂ZrHCl. Furthermore, the reduction proceeded relatively rapidly (~15–30 min) and cleanly, giving only the aldehyde and the zirconium byproduct, which was easily removed.

Although the original hypothesis was faulty and may not have been the best reaction to try, the surprising result was extremely interesting and fortunate. We therefore decided to pursue this finding and examine potential applications.

Related processes

Although a host of methods exist in the literature on the reduction of amides to aldehydes, many of them require harsh reaction conditions and are not very selective for reduction of amides. Reduction methods using the Schwartz reagent centered on the previously mentioned hydrozirconation processes; only a few papers describe the reduction of other functionalities, mainly in the context of functional group compatibility in the hydrozirconation processes.

Figure 4. Buchwald method for reducing tertiary amines to aldehydes (20).

Two relevant reactions are described in the literature, however, that detail the reduction of tertiary and secondary amides. In the first, Buchwald and co-workers describe the reduction of a variety of tertiary amides with titanium tetraisopropoxide–diphenylsilyl dihydride, followed by an aqueous workup to afford the corresponding aldehyde (Figure 4) (20). The reaction typically is carried out at room temperature, and the yields range from 50 to 90%. The proposed mechanism proceeds through a titanium-hydride–like complex, which is obtained by the reaction of Ph₂SiH₂and Ti(O-i-Pr)₄. Although this reaction is very mild and can tolerate numerous functional groups, such as terminal C=C and CC bonds, as well as nitriles and epoxides, the requirement of an aliphatic -hydrogen limits the reaction. Furthermore, epimerization of an optically active -stereogenic center would be expected to occur in chiral substrates.

In the second process, described by Ganem and colleagues, a variety of secondary amides can be reduced to the corresponding imines using the Schwartz reagent, as briefly mentioned above (9). In this procedure, the secondary amide is treated initially with potassium hydride, followed by treatment with the Schwartz reagent. Nonaqueous workup and diatomaceous earth filtration afforded the corresponding N-substituted imines in moderate to high yields (25–86%). Upon further mechanistic examination, the authors found that the reaction required 2 equiv of Cp₂ZrHCl to form the imine; and therefore, they suggested that the reaction probably involved binuclear complexes similar to those seen in carbonylation reactions (Figure 5).

Figure 5. Ganem method for reducing secondary amines to imines (9).

Ganem’s group later found that diisobutylaluminum hydride could be used to form the initial enolate, followed by the final reduction using Cp₂ZrHCl. This made it unnecessary to use 2 equiv of the Schwartz reagent, which can be costly in terms of reagent price and waste disposal. Although the reaction proceeds to reduce secondary amides to imines, little information was given about the reduction potential for other amide substrates.

Building on these studies, our findings have highlighted an additional use of the Schwartz reagent in the novel reduction of tertiary amides to aldehydes. The benefits are seen in the selectivity of the reagent for tertiary amides, even over groups such as esters, as well as the facile conversion under mild conditions, which gives the corresponding aldehyde in high yields.

Scope and selectivity

Our next goal was to examine further the utility of the reagent and determine the possible mechanism by which the reduction was proceeding. To do this, it was important to consider functional group compatibility. We knew after our initial studies that the reagent performed a mild reduction, but to use the reagent in any total synthesis, it must also be selective for the desired functionality. The reagent will reduce other functionalities such as carbonyl groups (ketones, esters, and aldehydes), cyano groups, and, of course, olefins, for which it is best known. Our goal was to examine the propensity to selectively reduce one functionality in the presence of another on the basis of the kinetic reactivity of the reagent.

We examined a variety of substrates that contained a tertiary amide and another functionality of interest to determine their compatibility. In every reaction, we found that the amide was selectively reduced to the corresponding aldehyde while keeping the other functionality intact. We were surprised that aldehydes were produced, because aldehydes are typically reduced by the Schwartz reagent to give the corresponding alcohol. At this point, the mechanism became a more critical question; we tried to discover how an aldehyde could be formed without additional reduction to the alcohol.

Establishing the mechanism

Our initial observations caused us to carry out probing mechanistic studies.

• Only slightly greater than 1 equiv of the zirconium reagent was required to elicit the transformation.
• The aldehyde product was not further reduced to the corresponding alcohol, even in the presence of excess reagent. This reduction would be expected if the aldehyde were placed alone in the presence of the reagent.
• The tertiary amides were reduced selectively. From this information alone, it seemed unlikely that a direct hydride transfer with subsequent elimination of the –NR₂ group was occurring. If this had been the case, some alcohol product would be expected, especially when >1 equiv was used.

Figure 6. Proposed mechanism for converting tertiary amides to aldehydes by Cp2ZrHCl via an iminium ion intermediate.

At this point, we postulated a plausible mechanism (Figure 6), in which the Schwartz reagent is added into the amide with subsequent hydride transfer. This zirconated intermediate can then proceed through an iminium ion intermediate, followed by reaction with water to give the desired aldehyde product. This mechanism seemed to correlate with the previously demonstrated mechanism for reducing secondary amides to imines (9). Furthermore, analogously to the Weinreb amides, a stable chelated intermediate is formed which is not subsequently reduced until it is exposed to water. This intermediate may account for the lack of overreduction to the corresponding alcohol.

We attempted to trap an intermediate to show the plausibility of this route. An additional mild reducing reagent was used, which is known to reduce imines and iminium species while not effecting the amide. The reagent of choice was tetrabutylammonium borohydride. The tertiary amide was treated with Cp₂ZrHCl followed immediately by adding the borohydride. The resulting product was a 1:1 mixture of the amine and alcohol.

These initial findings indicated that the iminium ion, or a related intermediate, must have formed in order to obtain the amine (control experiments showed that tertiary amides are not reduced by borohydride). Furthermore, we postulated that the alcohol could be formed if a trace of water were present; water would convert the iminium ion to an aldehyde, and this would be reduced by the borohydride to the corresponding alcohol.

To examine the mechanistic hypothesis further, we performed isotope experiments to determine the origin of each new atom. In the first study, we used the commercially available Cp₂ZrDCl. This experiment gave the deuterated aldehyde, showing the hydride transfer from the Schwartz reagent was occurring at some point. In the next isotope experiment, we used H₂¹⁸O to determine the source of the oxygen. We rationalized that if the proposed mechanism was occurring, and the aldehyde was being formed by water, the ¹⁸O label would be incorporated. By treating the tertiary amide with Cp₂ZrHCl and then subsequently quenching the reaction with 1 equiv of H₂¹⁸O, we found that the aldehydic carbonyl was labeled with ¹⁸O. This information supported the proposed mechanism shown in Figure 6.

Advantages of Cp₂ZrHCl

The reduction of tertiary amides to aldehydes via Cp₂ZrHCl has several distinct advantages. The reaction requires very short reaction times (~15 min) and provides high yields of the aldehydes with good chemoselectivity. Furthermore, the reaction does not require extensive workup procedures, nor does it require scrupulously dry conditions when the aldehyde is the desired product. Additionally, because the substrate dependence is minimal, this method will be useful for a wide variety of compounds. This is important because an ideal system will make it possible to manipulate the functional group of choice while leaving other groups intact. By generating a new method for carrying out functional group manipulations, we have provided synthetic chemists with another tool to tackle more complex problem.

References

1. Larock, R. C. Comprehensive Organic Transformations: A Guide to Functional Group Preparation; VCH: New York, 1989.
2. White, J. M.; Georg, G. I.; Tunoori, A. R. J. Am. Chem. Soc. 2000, in press.
3. Schwartz, J.; Labinger, J. A. Angew. Chem., Int. Ed. 1976, 15, 333–340.
4. Wipf, P.; Jahn, H. Tetrahedron 1996, 52, 12853–12910.
5. Negishi, E.; Takahashi, T. Aldrichim. Acta 1985, 18, 31–47.
6. Annby, U.; Karlsson, S.; Gronowitz, S.; Hallber, A.; Alvhall, J.; Svenson, R. Acta Chem. Scand. 1993, 47, 425–433.
7. Negishi, E.; Van Horn, D. E. J. Am. Chem. Soc. 1977, 99, 3168–3170.
8. Hudlicky, M. Reductions in Organic Chemistry, 2nd ed.; American Chemical Society: Washington, DC, 1996; pp 228–238.
9. Ganem, B.; Li, J.; Schedler, D.J.A. J. Org. Chem. 1996, 61, 4115–4119.
10. Holton, R. A.; Davis, R. G. J. Am. Chem. Soc. 1977, 99, 4175–4177.
11. Holton, R. A.; Natalie, K. J. Tetrahedron Lett. 1981, 22, 267–270.
12. Weinreb, S. M.; Nahm, S. Tetrahedron Lett. 1981, 22, 3815–3818.
13. Sibi, M. P. Org. Prep. Proceed. Int. 1993, 25, 15–40.
14. Mentzel, M.; Hoffmann, H.M.R. J. Prakt. Chem. 1997, 339, 517–524.
15. Maugras, I.; Ponchet, J.; Jouin, P. Tetrahedron 1990, 46, 2807–2816.
16. Braun, M.; Waldmüller, D. Synthesis 1989, 856–858.
17. Oppolzer, W.; Cunningham, A. F. Tetrahedron Lett. 1986, 27, 5467– 5470.
18. Irako, N.; Hamada,Y.; Shioiri, T. Tetrahedron 1992, 48, 7251–7264.
19. White, J. M.; Georg, G. I.; Tunoori, A. R. A one-flask synthesis of Weinreb amides from chiral and achiral carboxylic acids using the Deoxo-Fluor fluorinating reagent; Abstracts of Papers, 27th National Medicinal Chemistry Symposium, Kansas City, MO, June 13–17, 2000; American Chemical Society: Washington, DC, 2000; p B40.
20. Buchwald, S. L.; Bower, S.; Kreutzer, K. Angew. Chem., Int. Ed. 1996, 35, 1515–1516.

Ashok Rao Tunoori is a senior research scientist in the medicinal chemistry division of Coelacanth Corp. (ashok_tunoori@coelacorp.com). He earned his M.Sc. degree in organic chemistry from Kakatiya University, Warangal, India, and his Ph.D. from the National Chemical Laboratory in Pune, India. He was an Alexander Humboldt postdoctoral fellow for two years at the Martin Luther University, Halle-Wittenberg, Germany. He was a member of Gunda Georg’s group at the University of Kansas from 1996 to 1999. He has published more than 25 research papers and serves as a referee for scientific journals, including the Journal of the American Chemical Society and Chemical Communications.

Gunda I. Georg is a Kansas University Distinguished Professor in the Department of Medicinal Chemistry at the University of Kansas (Lawrence, Kansas 66045; 785-864-4498; georg@ukans.edu). She received her B.S. degree in pharmacy and her Ph.D. in medicinal chemistry from the University of Marburg, Germany. She has nearly 100 publications in the area of organic medicinal chemistry with a focus on the synthesis and structure–activity studies of anticancer natural products. She has served on committees for the ACS Medicinal Chemistry Division and Women Chemists Committee, and on scientific advisory boards for the National Institutes of Health. She is a cofounder of the startup company ProQuest, which specializes in drug delivery. In 1996, she was elected to the rank of AAAS Fellow.