AmericanChemicalSociety.com

Our research program ultimately requires synthetic chemistry to study the biological functions of structurally unique natural products with significant biomedical relevance. We have a continued interest in the synthesis and biology of guanidine containing natural products, due to the importance of this motif in the molecular recognition of carboxylate, sulfate and phosphate groups in biological macromolecules (Curr. Bioact. Cmpds. 2009, 5, 39-78). We are interested in the Leucetta alkaloids of which naamidine A and kealinine C are representatives (Figure 1). These compounds display a diverse range of biological activities based on a relatively conserved structural core. In particular we are interested in the unique ability of naamidine A to modulate EGF dependent DNA synthesis. Ultimately this hyper stimulation leads to activation of caspase-dependent apoptosis. While allosteric kinase inhibitors are well characterized, the roles of allosteric agonists are relatively unknown. Synthetic access to these compounds will provide probes to further delineate the important activity of naamidine A. NA22598A₁ or the guadinomines also represent an important arena for target identification and mechanism of action studies.  The only difference between these compounds is the purported position of the carbamoyl group on the guanidine (either N1 or N2). Since both natural products are isolated from Streptomyces sp. It is our contention that these are the same compound and that erroneous HMBC data in the structure elucidation of NA22598A1 led to an incorrect structural interpretation. Initial spectroscopic investigations in our laboratory on model systems indicate that this is indeed the case. NA22598A1 was identified as a selective inhibitor of anchorage independent growth (AIG), suggesting a potential lead for small molecules that can selectively inhibit tumor metathesis. The guadinomines were identified as agents capable of inhibiting the Type 3 Secretion System (T3SS). If it is true that these are the same compounds, it would suggest a unique and unknown biochemical link between these two biological processes. We hypothesize that both of these processes may be modulated by the activity of matrixmetalloproteinases and that these natural products are thus inhibitors of MMPs. Further we suggest that the carbamoyl guanidine is a unique motif capable of divalent metal chelation and thus provides a hypothesis for the mechanism of inhibition of these Zn(II)–dependent metalloenzymes. If this is indeed the case, it would represent an important advance for the design of MMP inhibitors. Ageladine A is also an MMP inhibitor but its exact mode of inhibition has not been characterized with atomic resolution. Other marine metabolites with guanidines serve as synthetic inspiration such as saxitoxin, the crambescidins and the batzelladines.

Synthetic platform: Given the range of structural diversity in these natural products in respect to ring size, oxidation state and substitution we aimed to develop a unified synthetic platform to access these skeletons. In pursuance of our interests, a unified approach must:

     1)   provide access to predictable and controlled substitution patterns in short order,

    2) deliver these diverse hydrogen bond topologies, and

    3)   permit ring oxidation state adjustments within these cyclic or polyclic structures.

We felt that the addition of a guanidine N-H bond across a C-C p-system would represent a powerful tool for the preparation of these challenging heterocyles. The development of amine-alkene/alkyne hydroamination chemistry has provided an invaluable tool for the construction of nitrogen based heterocycles. The extension of this methodology to guanidines, however, remains largely underdeveloped. In order to manipulate the oxidation states of these heterocyclic cores we anticipated that this is best done by reduction, as oxidations of nitrogen rich heterocycles are usually problematic. This has led us to study the addition of guanidine N-H bonds across alkynes, thus providing access to cyclic structures at an elevated oxidation state (Figure 2). This disconnection also permits the rapid construction of diverse skeletal precursors via an imine-acetylide 3-component coupling.

To this end we examined the utility of an addition-hydroamination-isomerization sequence (Figure 3).  (Angew. Chem. Int. Ed. 2009, 48, 3116-3120). This strategy provided access to a variety of highly substituted 2-aminoimidazoles in just 3 steps and was generally high yielding. Further, the introduction of a removable R group permitted the preparation of any desired hydrogen bond donor-acceptor topology.

     We have also been able to adapt this reaction sequence to the addition of thiols, phenols and alcohols (Figure 5, J. Org. Chem.  2010, 75, 261-264). Consistent with what we had learned from the addition of amines to the cyanamides, this chemistry requires a fine balance of nucleophilicity and basicity to effect addition to the cyanamide without formation of the allenyl-cyanamide which readily decomposes. This manifold presents an interesting polarity reversal for the preparation of these heterocycles which are usually constructed by the addition of an intact thiourea or urea to an electrophile.

The major limitation of this methodology was that the addition of a nucleophile to the cyanamide was rate limiting and required forcing temperatures (>95 ^oC). Under these conditions only the stereo-electronically favored 5-exo-dig cyclization was observed followed by isomerization. Thus we set out to develop a more mild cyclization procedure that would allow us to A) selectively generate the 5-exo-dig or the 6-endo-dig product and B) maintain the fidelity of the alkene position so that it can be used for further functionalization. Under the simplistic assumption that a p-Lewis acid might trigger the hydroamination process, a series of metal salts were evaluated for their ability to affect the cyclization of an intact N,N-diboc-propargylguanidine (Figure 6, Angew. Chem. Int. Ed. 2011, 50, 684-687). In short we found that Ag(I) catalysts were optimum for the formation of the formal “5-exo-dig” product. We also discovered a unique role of dirhodium carboxylates to promote the selective formation of the formal “6-endo-dig” product. Of particular note is the fact that Rh(II) is highly selective for alkynes whose termini are not electronically differentiated (e.g. alkyl, see row 2). Rh(II) can also tolerate alkyl-halides which will be important for subsequent annulations etc. in the formation of more elaborate natural product skeletons.