The past decade has witnessed the dramatic progress on site-specific incorporation of unnatural amino acids (Uaas) into proteins in live cells through the expansion of the genetic code.
1–4 Using engineered tRNA-synthetase pairs orthogonal to endogenous counterparts in the host cell, more than 70 Uaas have been incorporated into proteins in
Escherichia coli, yeast, or mammalian cells, which enable new opportunities to study biological problems with expanding non-proteinogenic chemistries. Fluorescent Uaas (), including L-dansylalanine (
1),
5–7 L-(7-hydroxycoumrin-4-yl)ethylglycine (
2)
8 and L-Anap (
3),
9 are of particular interest for molecular and cellular biology, which may complement and enchance the widely used fluorescent proteins (FPs).
10 These fluorescent Uaas can be incorporated at virtually any site of a protein, whereas the fusion of FPs is often limited to the N- or C-terminus or certain loop regions. In addition, fluorescent Uaas are much smaller than FPs, which helps to mitigate undesired perturbations to the structure and function of the target protein. Moreover, when developing fluorescent reporters, sensitivity to different environment cues such as pH and polarity can be chemically designed and synthesized
11 into the fluorescent Uaas, a flexibility not readily available for FPs.
Despite the potential advantages, effective application of fluorescent Uaas in live cells faces multiple challenges. Efficient synthesis of optically pure Uaas is a foremost demand. When racemic Uaas are fed to cells, the D-enantiomer competes for cellular uptake yet cannot be incorporated into proteins by the ribosome.
2 This reduces the net intracellular concentration of the bioactive L-enantiomer, leading to lower incorporation efficiency. For fluorescent Uaas in particular, the intracellular D-enantiomer further increases background fluorescence, which is a critical issue for cell imaging with Uaas. Synthetic routes to enantiomerically pure
1 and
2 have been developed,
5,12 but only a racemic synthesis is available for Uaa
3.
9 Anap is sensitive to polarity with changes in intensity and emission wavelength.
9 Uaa
3, when selectively incorporated into proteins, has the potential to fluorescently report protein conformational changes, interactions, modifications, and activities in live cells. Herein, we report a concise and efficient synthetic approach of L-Anap, which offers significant improvements over the current synthesis.
L-Anap has been incorporated into proteins in
Saccharomyces cerevisiae by Schultz and co-workers.
9 They developed a synthetic approach of racemic anap, which gives 8.1% of total yield over 8 steps (). The synthesis was based on the nucleophilic substitution of ethyl bromopyruvate oxime (
6) developed by Gilchrist and co-workers.
13,14 Presumably due to the low nucleophilicity of the Burcherer reaction product of
4 and
5, compound
7 was synthesized in only 26% yield over two steps from compound
4.
15 Another drawback of this synthesis resides in the nonstereoselective reduction of
7, which leads to the final racemic product.
To synthesize sufficient amounts of L-anap, we decided to develop an efficient and enantiospecific synthetic pathway toward L-anap (). Our synthetic strategy involves Fukuyama-Mitsunobu reaction
16–18 as the key step, which couples compound
10 with
N-trityl-L-serine methyl ester
1119 to give intermediate
12. This approach has two features: First, both the nucleophile and the electrophile are channeled to facilitate the intermolecular coupling by tuning the electronic effect of the protecting groups.
o-Nitrobenzenesulfonyl group makes
10 a good nucleophile under Mitsunobu condition and can be removed easily under mild conditions. Trityl group of compound
11 can conduct the reactive intermediate through the intermolecular coupling and circumvent
β-elimination as the side reaction. Second, the Mitsunobu reactions of
11 with different nucleophiles have been proved to be free of racemization,
19 therefore intermediate
12 are expected to be synthesized enantiospecifically.
We commenced with the synthesis of 6-acyl-2-naphthylamine (
13)
via two pathways (). In approach A, the selective Heck reaction
20–22 and Cu(I)-catalyzed C-N
23,24 bond formation were employed to introduce the acetyl and amino groups. 6- Bromo-2-naphthol (
14) was transformed to triflate
15, which was coupled with butyl vinyl ether through Cabri’s procedure,
20–22 followed by treatment with acid, to give intermediate
16. Under Buchwald’s condition,
25 compound
16 underwent the C-N bond formation to give intermediate
17 in 75% yield. After deprotection, compound
13 was obtained in 54% yield over 4 steps. Although this approach provided gram-scale synthesis of
13 with satisfactory yield, approach B, which took three steps less than A, was explored in parallel. Approach B features the direct conversion of hydroxyl group of compound
4 to amino group. Although this transformation
via Burcherer reaction
26 has been reported by Cho and his coworkers,
27 the procedure involves stirring the reaction mixture at 140 °C for 96 hours in a steel-bomb reactor, which is not viable in many biology laboratories. Inspired by recent advances in one-pot alkylation-Smiles rearrangment-hydrolysis
28 sequence, we applied Mizuno and Yamano’s procedure to the transformation of
4 to
13. Monitoring the reaction with TLC showed the complete conversion of the alkylation and Smiles rearrangement steps. However, the in situ hydrolysis of
19 under basic condition was incomplete and the final product
13 was difficult to separate from intermediate
19. Therefore, the crude product of
19 was separated and hydrolyzed under acidic condition, which provided compound
13 in 73% yield.
With compound
13 in hand, we set out to explore the Fukuyama-Mitsunobu reaction (). Compound
13 was treated with
o-nitrobenzenesulfonyl chloride and pyridine to give compound
10. In the presence of diisopropyl azodicarboxylate and triphenylphosphine, compounds
10 and
11 were coupled to give intermediate
12 in 88% yield over two steps. Sequential deprotection of trityl group and
o-nitrobenzensulfonyl group furnished L-anap methyl ester (
20). Deprotection of
20 under acidic condition
29,30 furnished L-anap as its hydrochloride salt. The enantiomeric purity of
3 was examined with the Mosher method,
31 and the dr value was above 40:1, which indicates that less than 3% racemization had occurred during deprotection (see
Supporting Information).
In conclusion, we developed an efficient synthetic route to enantiomerically pure L-anap, which gives 51% of total yield. Compared to the previous method, our synthetic approach not only afforded a bioactive L-enantiomer but also enabled a 12-fold increase in yield with fewer steps. All the reactions of our approach proceeded under mild conditions as well as in gram-scale. The general approach demonstrated here can also be applied to the synthesis of other 3-
N-aryl-2,3-diaminoproprionic acids, which serve as important building blocks for medicinal small molecules.
32–34 The application of L-anap to image protein modifications and activities in mammalian cells is undergoing and will be reported in due course.