ARTICULO

 

MECHANISTIC VIEWS OF INTRAMOLECULAR HYDROXYCYCLOPROPANATION OF VINYL CARBOXYLIC ESTERS

 

 

José A. Bravo, * 'Patricia Moliinedo, J. Mauricio Peñarrieta, José L. Vila
Department of Chemistry, Universidad Mayor de San Andrés, P.O. Box 303, La Paz, Bolivia

 

 

---

Keywords: Organic Chemistry, Addition Reactions, Alheñes, Esters, Grignard Reagents, Mechanisms of Reactions, Natural Products, Organometallics, Kulinkovich, cyclopropanation, Jin Kun Cha.

ABSTRACT

The overwhelming presentation of plenty of synthetic steps in a verbally reduced or succinct form as appearing in any paper is here focused in an extensive and particularly graphical manner; just to extend the screen when the reader is boarding a published series of synthesis. The Intramolecular Hydroxycyclopropanation of ¿y-Vinyl Carboxylic Esters appeared to us to be a fascinating synthesis thematic and served us as an example to propose didactical and mechanistic views.

*Corresponding author: jabravo@umsa.bo

---

 

 

ANALYSIS AND MECHANISTIC PROPOSALS

As academics we are highly concerned with the didactical importance of covering the needs of debutant students in organic synthesis. This article presents an analytical and didactical approach to synthetic works by Jin Kun Cha and col. by means of reactions' theoretical mechanisms. We are continuing the previously published: "A Theoretical Mechanistic Approach to Diasteroselective Synthesis of cis-l,2-dialkenylcyclopropanols and Subsequent Oxy-Cope Rearrangement " (1) presenting now another synthesis by Cha and coworkers concerning Grignard reagents in the presence of transition metáis (2). Cha worked with organotitanium compounds to effect cyclopropane-mediated natural product synthesis. Prof. Cha group became interested in the Kulinkovich hydroxycyclopropanation (3). They published first the treatment of a carboxylic ester with an excess (3 equiv) of Grignard reagent in the presence of Ti(O/-Pr)₄ (1 equiv) affording cM-l,2-dialkylcyclopropan-l-ols (4). We already proposed a mechanistic approach (l) about such article. The second related article by Cha and col. (2) reports an intramolecular versión of the Kulinkovich hydroxycyclopropanation (4) and imposes the treatment of vinyl carboxylates with w-BuMgCl in the presence ofThe Kulinkovich hydroxycyclopropanation's scheme 1 (2) is reproduced here in Scheme 1 and it can be depicted through the next mechanistic proposal which has already been graphically and verbally described (l) and it is re-explained here.

To depict this Kulinkovich hydroxycyclopropanation scheme we start with our 2 equiv of Grignard reagent or A first interaction occurs between Grignard reagent and the titanium tetraisopropoxy derivative. Titanium is linked to the four alkoxy substituents. The reaction provokes the expulsion of two alkoxy residues to leave a species. The metal di-cationic charge is used by two carbanions
each from 1 equiv of the Grignard reagent, to obtain a dialkoxy-dialkyl titanium derivative 3. (Figure 1).

Titanium receives temporarily an extra electrón from an alkyl methylene to afford thespecies and the carbonium ion R'CH₂CH₂⁺. A reductive nucleophilic attack by the anionic titanium leads to the apparition of the Ti 3-membered ring intermedíate 4 (or a titanocyclopropanoid or titanacyclopropane (2)). (Figure 2).

The next step in this pathway consists of the condensation of intermedíate 4 with a carboxylic ester RCO₂R". The carbonyl group of this ester drives the reaction by interacting with the Titanacyclopropane in a transition state that implies a co-planarity of both species. A nucleophilic attack by the nucleophile carbonyl oxygen over titanium dispatches the two electrons of the bond Ti-CH(R')(CH₂) over that carbón. Henee the Titanacyclopropane splits to afford a carbanion and a cationic oxygen over the same molecule. A nucleophilic attack over the carbonyl carbón (with an exacerbated electrophilic character) leads to the formation of a titanium furanoid ring (Figure 3)

An intramolecular nucleophilic attack by the ethereal oxygen of thesubstituent group over the electrophilic titanium dispatches the two electrons of the Ti-CH₂- bond on intermedíate 5 to afford the corresponding splat species bearing a carbanion and a cationic oxygen, which in turn recovers neutrality by splitting the four-membered, di-oxygenated (oxirane-type) ring. The result is the dipolar (a carbanion and carbonium ion) species that neutralizes the charged extremes into a substituted cyclopropane or intermedíate 6 (Figure 4).

It is now time for two more equiv of Grignard reagent and for intermedíate 6 to interact to afford species 2, closing thus our looping cycle. This meeting takes place by means of cationic interchange, or the cationic moiety of Grignard MgX⁺ that replaces cationic Ti⁺(OiPr)₂(OR") through the temporary apparition of an anionic oxygen. The remaining Ti²⁺(OiPr)₂ species neutralizes the two anionic moieties liberated by the 2 equiv of Grignard reagent to form two new equiv of the species 3 to continué the cycle (Figure 5).

This synthetic pathway (Scheme 1) was largely applied by Cha and coworkers to produce a number of azulene sesquiterpene derivatives as published by these authors. Cyclisation (Scheme 1) goes on under stoichiometric or catalytic conditions involving the double alkylation of the tiatanacyclopropane intermedíate 4, itself formed by the reaction of and a Grignard reagent followed by elimination of the corresponding alkane (R'CH₂CH₃). The thesis that the putative intermediate 4 or the Titanacyclopropane intermediate could undergo a reversible exchange with an alkene has been thought. If a carboxylic ester is chained to the alkene moiety, then an intramolecular hydroxycyclopropanation is feasible. And this is the method currently under discussion. The numerous results afforded by Cha and coworkers include, structurally, two-fused cycles compounds all of them showing the cyclopropane feature, characteristic of the Kulinkovich synthesis. The second fused cycle comprises many models depending on the alkene used to be condensed to the Kulinkovich's intermediate substrate. The foliowing mechanisms correspond to enfries 1 to 15 in table 1, page 292, (2).

Entry 1.
It regards the first intramolecular hydroxycyclopropanation achieved by the original paper authors. Methyl 5-hexenoate with 3 or 5 equiv of Grignard's w-afforded the first intramolecular hydroxycyclopropanation product or cyclopropanol 8 (Scheme 2).

To explicitly understand this reaction we must follow the mechanism exposed above as a model; however some different explanations must be given in order to achieve the reported results (4). This implies a modification of the section comporting the intermolecular interaction between the Titanacyclopropane derivative 4 and the carboxylic ester for now we must explain the apparition of the intramolecular process product. Thus, once generated intermediate 4 as shown mechanistically above what is provoked now contrasting with the intermolecular versión is that before proceeding to a coplanar interaction between the Titanacyclopropane and the carboxylic ester, a splitting of 4 happens in order to genérate the 1-butene residue (Figure 6).

Titanium liberates one electrón splitting so the bonding with its neighboring methylene which in turn receives the two bonding electrons acquiring an anionic character. Indeed, titanium becomes a cationic nucleus. Neutrality returns to the Ti nucleus by means of receiving an electrón coming from the splitting of its bonding to the protoalkene. The alkene appears due to the driving forcé derived from the re-accommodation of the electronic excess over the carbanion into a tí bond. Titanium is now a neutralized nucleus; however there should be an electronic re-accommodation in different atomic orbitals generating so a dipolar situation giving titanium a nucleophilic character. This nucleophile will attack the terminal alkene of the ¿y-vinyl carboxylic ester. This fact collapses finally in the tricycle titanium derivative, some kind of eptitanium (like epoxy) or a Titanacyclopropane appearing after addition over the double bond of the ¿y-vinyl ester (Figure 7).

The intramolecular nucleophilic attack by the carbonyl oxygen in the methyl ester moiety occurs to produce a dipolar transition state. This nucleophilic attack exscinds heterolytically the bondand establishes a new link between oxygen an titanium. Stereochemistry in these mechanisms has a speculative character. A supraplanar attack over carbonyl electrophilic centeritself contained in a parallel plan to that containing the titanium derivative occurs placing the methoxy group upwards, adopting thus an axial position in the fusión bridge of the newest fused cycles compound. An isomerization towards a more comfortable conformer occurs changing all axial into equatorial bonds. The new conformer is particularly benefited by the absence of the previous through-space Van der Waals radii interaction of the axial methoxy and the axial isopropoxy substituents, all according to framework molecular models (Figure 8). All these steps conduct to the formation of the fused rings according to the reported reaction producís (2). The polar attraction manifested by oxygen and titanium establishes a bridge for a much tensed four-side cycle including a highly stressing double oxy function. This tensión incites the carbanion next to the stressed di-oxy ring to attack establishing a less tensed cycle a^v trois. This enterprise generates the stable enough 3/5 fused rings compound, which is analog to intermedíate 6 in the Scheme 1 of the present paper.

We return now to Scheme 1 to observe that intermedíate 6 will interact with the Grignard reagent that for entry 1 corresponds to «Bif^gO. This happens first by means of exscinding the Ti-O-Cvcles bond to genérate an alcoxy function which in turn is neutralized by the Lewis acid from Grignard r. namely "IVIgCl giving rise to the intermedíate to element 2 in the scheme. This Grignard adduct is finally destroyed in a protic médium to afford alcohol 8 (Figure 9).

What foliows next is the different interactions for catión ⁺Ti(OiPr)₂(OMe). This is first neutralized by the basic moiety of the Grignard reagent or the «Bu" carbanion (Figure 9). Next the basic ^-OMe is separated from titanium due to a better interaction with Lewis acidfrom a second equiv of Grignard r. The just formed Ti catión reacts with the basic moiety left by the second Grignard equiv to genérate intermediate 3 (Figure 10) in Scheme 1 thus giving continuity to the virtual reacting loop.

Entry 2.

Under the standard conditions (5 equiv of «BuMgCl, 0.5 equiv of ClTi(OiPr)₃, ether, room temperature) the authors achieved the synthesis of bicyclic 10 out of the homologue 9. The mechanism is the same as the previously presented. This demarche is graphically described as follows in Figure 11.

Entry 3.

According to Cha and coworkers (2) a precipitous decrease in yield was observed for the intramolecular hydroxycyclopropanation of 11 to afford bicyclo[5.1.0]octan-l-ol 12 in a 11% yield. Now we start our discussion from the apparition of the transition state (Figure 12).

As observed in the above graphics and after framework molecular models construction, the sp³ carbons of the fusión bridge become very deformed from the tetrahedral geometry. It means that the instability of compound 12 is increased regarding that of compound 10. This could be a reason to explain the lower yield for compound 12 with respect to the yield of 10 and 8.

Entries 4 and 5

The original paper's authors (2) mentioned that further extensión to methyl 8-nonenoate and methyl 4-pentenoate failed to produce the corresponding bicyclic cyclopropanols 14 and 16 (0% yield for both). See Figure 13.

Entries 6 and 7

According to Cha and coworkers, as a rule, the intramolecular hydroxycyclopropanations are not much influenced as a result of the presence of different substituents in the chain (entries 6-13, (2)). We start this reaction graphic description from the nucleophilic attack over the terminal vinyl of the methyl carboxylate 17. This compound is actually being substituted at positions a and y from carbonyl (Figure 14).

In order to better propose the corresponding theoretical mechanism, we will imagine stereochemistry for the two chiral centers under a speculative scope (Figure 15).

Entry 8 and entry 9

Contrasting to entries six and seven, the intramolecular hydroxycyclopropanation in entries 8 and 9 becomes influenced by the presence of substituents at the allylic position in the side chain. This can be translated as a dramatic diminution in yields (14% for product 22, entry 8, and 0% for the expected 24a, entry 9). This entry is graphically described as follow in Figures 16, 17 and 18.

At this point a competitive elimination of the HOTIPS group takes place to give the corresponding tricycle alkene. This fact is responsible for the low yield in the intramolecular product 22 (Figure 17). For entry 8:

For entry 9:
The expected intramolecular product 24a did not appear giving no yield. Instead product 24b (Figure 18) appeared as a result of a known intermolecular mechanism, it was yielded in a 61%. This result clearly shows that under the given conditions, and with respect to eventual substituent groups present at the allyl position of the alkene, both mechanisms, have a competitive character.

Entry 10
Mechanism for entry 10 is the same as that proposed for entries 6, 7 and 8. As a practice for the novel, it should be done apart using entry 8 as a guide. However the apparition of a mixture of epimers as resulting products encourages us to develop graphically one of such reactions (Figure 19).

This same mechanistic demarche gives rise to epimer 26b with the only fact of departing from at the /?-of-carbonyl inverted configuration position in the methyl ester compound.

Entry 11

Entry eleven presents new structural features and consequently new challenges in proposing a theoretical reaction mechanism. Now the side chain includes a substituted cyclohexane (Figure 20).

In order to obtain the isomer 28a (Figure 21), we must follow the mechanism below. We observe here that the stereochemical definition for one or the other isomer comes out just before establishing one or the other transition state. At that stage the free rotation about the sigma bond of the methyl ester group and the way for it to face the Lewis acid-base interaction between the carbonyl oxygen and the cationic titanium provokes a beta or alpha orientation for the tricycle. It is obvious that the most favored diasteroisomer in yielding was 28a due to an easier nucleophilic attack from the carbonyl oxygen to titanium in comparison to 28b (2:1).

The prodigious synthetic work by Cha and coworkers also permitted these authors to afford the trans-nng junction diasteroisomers 30a and 30b besides the cis-ring junction diasteroisomers 28a and 28b exposed in entry 11. The trans isomers resulted less favored in yielding due to the additional associated strain, all according to the original paper authors (2). No immediate references are given regarding the experimental conditions applicable in order to obtain cis or trans diateroisomers. As expected the mechanistic proposal regards a cióse similarity to the ones expressed in entry 11 (Figure 22).

At this point, the mechanism goes toward provoking an interaction between the ester moiety and titanium (Figure 23). To achieve such relationship an appropriate approaching of the ester function to titanium is necessary. Thus there must be a sigma bond free rotation of the tripod conformed by the cyclohexane and -OTBS as shown abo ve in Figure 22.

Once established the Ti-0 link, the formation of a 5 membered cycle comes out from a nucleophilic attack from the carbanion faced to the carbonyl carbón avid itself of electronic charge (Figure 22). The result is apparent in the three-fused-cycles non-ionic intermediate in Figure 24.

The proximity of an oxygen atom from a methoxy group to the titanium nucleus in such intermediate makes possible the continuation of the reaction by means of a nucleophilic attack over Ti. This splits the furanoid ring to give a highly instable 4-membered dioxy-titanium cycle and a methviene carbanion (Figure 25).

Being the 4-membered dioxy-titanium cycle besides an electronic deficient charged species, this reacts by splitting the 4-membered cycle by the attack of the nucleophile basic methylene. An important feature of the electronic movement at this stage of the intramolecular interactions is that the nucleophilic attack taking place between the already mentioned nucleophilic carbanionand the dioxygenated carbón at the just formed 4-membered dioxytitano cycle, implies as a result the apparition of the fused tricycle in an aorientation all through the inversión of the configuration of this chiral carbón(Figure 26).

The complementing mechanism to obtain the alcohol derivative form out fro the titanium intermediate is shown in Figure 27.

In order to propose a mechanistic explanation for product 30b we must observe at model A (Figures 23 and 28) that the two electrons are in the inner position, ready for a nucleophilic attack over the electrophilic carbón of carbonyl in the ester moiety actually linked to titanium through the carbonyl oxygen. To achieve the attended inverted (relative to the oí position of the tricycle in 30a) stereochemistry for the 3-membered cycle as shown in 30b (2), we lócate temporarily the two electrons in the outer position just as shown in Model B (Figure 28). Afterwards there must be an accommodation of the two electrons on that carbanion, currently at a sp² hybrid state, in order to be located (the 2 e") again in the inner position. So, model B must be turned for the 2 e" to regain the inner position being thus ready again for the nucleophilic attack over the carbonyl carbón. The result of turning atoms through bonds is shown below on model C (Figure 28).

This conformer (Model C, Figure 28) presents already the hydrogen of this carbanion (sp ) in an alpha position differing from the same cycle junction hydrogen in diasteroisomer 30a, where it appears in the beta position.

Let us proceed now to the nucleophilic attack by the electrons from the inner position into the carbonyl carbón. This provokes indeed the displacement of the n electrons of carbonyl to compénsate the cationic charge over deriving thus in an ether cyclic ether function (Figure 29).

The remaining operations regarás the replacement of the titanium moiety by the Grignard's catión and the final alcoholic water residue to afford 30b (Figure 30). This intramolecular nucleophilic attack driving to the apparition of theoriented-three-membered cycle characterizing compound 30b, happens with inversión of configuration at the chiral centre supporting the ⁺0Me and OTi residues.

Entry 13
Even though the failure to accomplish derivative 32 (0% yielding (2)) we propose with a didactical purpose how such reaction should happen from a strictly theoretical stand point (Figure 31).

The constructed model at this stage shows partially our molecule, particularly the conjugated C=C and C=O(OCH₃) systems. We do also identify the methine carbanion as well as the Ti⁺(OiPr)₂ catión. The next electronic movement comports the nucleophilic attack C=O—>Ti⁺ (Figure 31). This gives rise to an eight-membered cycle which according to the constructed model exhibits annular tensión because of failing to reach the previous and normally expected co-planarity of both n conjugated systems (Figure 32). Isn't this the cause for the reaction failure? However still theorizing we can suppose from there on another nucleophilic intramolecular attack from the methine carbanion to the electrophilic carbonyl carbón. As a result we've got a three fused cycles system 6-5-5 (Figure 31).

We can describe these steps inverting the order of the nucleophilic attacks episodes (Figure 33): we first propose an CH:~ attack on C=O and then, once free from the contracting non-planarity of the nconjugated systems (now existing only the C=C system), the second nucleophilic attack of C-O" to the Ti⁺(OiPr)₂ catión.

Anyway the first attack may never happen because of the currently absence of the Lewis acid catalytic action of Ti⁺(OiPr)₂, once linked to C=O as it was our first option. The problem arising from these mechanistic proposals is the trans disposition reached by the 1,2-fraws-H-OMe group at the ring junction after the methine attack on the carbonyl carbón of the ester moiety. This does not conduct to the expected stereochemistry at the ring junction of the final structure comporting the 5-and 3-membered fused rings. The good structure (2) 32 (Figure 31) presents a cis disposition for the just mentioned groups. It forces to another pre-disposition of the methine before attack on the C=O. For solving this we will only consider the case of a first nucleophilic attack on titanium by the carbonyl oxygen and then a second by the carbanion methine on the carbonyl carbón of the ester moiety. The transition state has already been described as we can see in Figure 34.

Nevertheless, the intramolecular carbanion attack on C=O as just described, proved to give rise to the 1,2-trans-K-OMe isomer. This TS shows the methine hydrogen in an outer position regarding the 8-membered ring. Thus it becomes necessary to employ the opposite position for that methine hydrogen before the carbanion attack, or the inner hydrogen position on the carbanion. For that purpose the 8-sided cycle must be re-accommodated throughout restricted sigma bonding rotation. The result can be drawn and constructed as follows (Figure 35):

This TS shows after rotation the inner hydrogen which means the electrón pair placed in an outer position. It signifies that there is no spatial disposition for the nucleophilic intramolecular action of the Lewis base (CH:~). This TS needs anyway the electrón pair in an inner or well disposed situation. In order to place the electrón par inside the 8-sided cycle, the sp³ carbón (CH:~) hybridize to sp² and then to sp³ state again resulting all in a configuration inversión of the chiral centre and thus the electrón pair placed inside the ring ready for a nucleophilic attack (Figure 36).

The next steps would involve the formation of the highly tensioned dioxotitano 4 membered ring and the formation of the three membered fused ring by the carbanion attack over the elctrophilic carbón supporting the MeO⁺ function. All this is followed by the known steps leading to the final alcohol form of the hypothetic product 32 (Figure 37).

Enfries 14 and 15
According to Cha and col. esters ofalken-1-ols, particularly 3-buten-ols, seem to be amenable to intramolecular hydroxycyclopropanations. Thus, Cha and col. also effected the reaction of benzoate 33 to give a 7:1 mixture of the cyclopropanols 34a and 34b. A further extensión to the homologue 35, lacked of good yield. These reaction entries can be mechanistically envisaged as follow (Figure 38).

For entry 15: The mechanistic approach for the 33-homologue 35 is the same as for 33 (see entry 14 mechanistic approach for details).

 

REFERENCES

1.     Bravo, J. Bol. J. ofChem. 2005, 23, 1-10. (http://www.bolivianchemistrviournal.org. 2005)        [ Links ]

2.     Lee, J.; Kang, C. H.; Kim, H.; Cha, J. K. J. Am. Chem. Soc. 1996,118, 291-292        [ Links ]

3.     Kulinkovich, O. G.; Sviridov, S. V; Vaasilevskii, D. A. ;. Pritytskaya, T. S. Zh. Org. Khim. 1989, 25, 2244.        [ Links ]

4.     Lee, J.; Kim, H.; Cha, J. K. J. Am. Chem. Soc, 1995,117, 9919        [ Links ]