Synthesis and SAR of 3,5-diamino-piperidine derivatives: Novel antibacterial translation inhibitors as aminoglycoside mimetics

Abstract

Aminoglycoside antibiotics target an internal RNA loop within the bacterial ribosomal decoding site. Here, we described the synthesis and SAR of novel 3,5-diamino-piperidine derivatives as aminoglycoside mimetics, and show they act as inhibitors of bacterial translation and growth.

Bacterial resistance to antibiotics is on the rise and represents a global medical threat. In hospitals in the United States, approximately two million patients per year are infected.¹ The majority of these nosocomial pathogens are resistant to at least one antibiotic and result in about 90,000 deaths per year; a number that has increased 7-fold over the last decade. The recent and rapid spread of community acquired methicillin resistant Staphylococcus aureus further highlights the threat of resistance development and illustrates the need for new antibiotics that work by novel mechanisms.² Given the broad genetic and physiological diversity of bacterial pathogens and the need for empiric therapies that cover a broad panel of organisms, it is not surprising that discovery of new antibiotics has advanced slowly. Central to antibiotic discovery is identifying broadly validated targets. One such proven target is the bacterial ribosome, which is the target for a significant number of clinically important antibiotics that bind at the ribosomal RNA (rRNA).³ Here, we expand on the description of a novel series of antibacterial compounds that target rRNA and blocks bacterial translation and growth.⁴

Three-dimensional structures of different aminoglycosides bound to the decoding site, or A-site, within the 16S rRNA have been determined by X-ray crystallography.⁵ Importantly, these studies have shown that 2-deoxystreptamine (2-DOS), a conserved core scaffold among aminoglycosides, binds in a similar manner regardless of the 4,5- or 4,6-disubstitutions found in the neomycin or gentamicin families, respectively (Fig. 1).⁶ The cis-1,3-configured amino groups of 2-DOS are predominantly involved in base recognition by forming conserved hydrogen bonds with A1493, G1494 and U1495 of the 16S rRNA. These interactions anchor the aminoglycoside scaffold within the A-site internal loop and displace residues A1492 and A1493 from the RNA interior. These two adenine residues act as a molecular switch that is involved in securing the fidelity of translation by interacting with the first two bases of the mRNA-tRNA codon-anticodon hybrid. Thus, aminoglycoside binding impairs the ribosome’s ability to discriminate against near-cognate tRNA-mRNA pairings, which ultimately causes cell death.

Figure 1.

The aminoglycoside gentamicin and the DAP scaffold that mimics 2-DOS of the aminoglycoside.

While 2-DOS is recognized as a key pharmacophore for binding rRNA, its synthesis and modification is challenging due to the five contiguous stereogenic centers.⁷ Therefore, we had developed synthetic mimetics of 2-DOS, one of which is cis-3,5-diamino-piperidine (DAP) (1) (Fig. 1).⁴ The DAP ring retains the characteristic cis-1,3-diamine configuration of 2-DOS that is important for RNA recognition. In contrast to aminoglycosides, the reduced chemical complexity renders the DAP series and derivatives amenable to rapid elaboration by parallel synthesis. Here, we focus on the synthesis and structure-activity relationships of DAP and other synthetic mimetics of 2-DOS.

Scheme 1 outlines the synthesis of the Boc-protected DAP (5). In this specific synthesis, the two NO₂ groups of 2-chloro-3,5-dinitropyridine (2) was hydrogenated under H₂ atmosphere at room temperature in the presence of 10% Pd/C to generate 3,5-diaminopyridine 3. The two amino groups were then protected as a di-Boc derivative 4 which was subsequently hydrogenated at 2200 psi using 5% Rh/C in the presence of acidic acid at 110 °C to yield the piperidine derivative 5.^8,9 Since aminoglycoside are relatively large molecules that engage in at least 15 contacts with rRNA,⁵ a single DAP scaffold would not allow high affinity binding to rRNA and, hence, additional functionality was added. Guided by modeling studies and the availability of a straightforward synthetic route, DAP was linked to a triazine core which resulted in molecules (termed DAPT) that had a high affinity for A-site containing RNA oligonucleotides.^4,10 The DAP scaffold and analogs (R¹ and R², termed headpiece) thereof were linked at the 2 and 6 positions of triazine (6) (Fig. 2). The so-called “tailpiece” substituents (R³) were attached at the triazine 4-position via a N-linkage. Representatives of R³ used in this study are shown in Figure 2.^4,10

Scheme 1.

(a) H₂, 10% Pd/C, rt. (b) Boc₂O, NaHCO₃, H₂O/MeOH/THF. (c) H₂, 2200 psi, 5% Rh/C, CH₃COOH, 110 °C.⁸,⁹

Figure 2.

DAPT (6) and tailpiece (R³) structures.

Table 1 highlights the SAR of headpieces where R¹ and R² are identical resulting in a symmetric configuration. The symmetrical arrangement of two DAPs in compounds 11 and 12 resulted in potent translation inhibitors. Importantly, in the symmetric configuration, the acyclic headpiece (13) and a dimethylated DAP (14) retained activity (Table 1). Both of these headpieces allow the amino groups to be located in identical spatial orientation found both in DAP and 2-DOS and, hence, may account for similar hydrogen bonding with target rRNA. Although two amino groups in the five-membered ring headpiece of compound 15 have a slightly different spatial orientation than in DAP, 15 was nevertheless active. In contrast, when one of the amino groups in the headpiece is removed, or the spatial orientation is significantly altered, activity is abolished (compounds 16–18). These results support our hypothesis that DAP can indeed substitute for the 2-DOS pharmacophore. Moreover, these findings show that the signature cis-3,5-diamino fragment of DAP, and by analogy the cis-1,3-diamino fragment of 2-DOS, can tolerate a variety of structure modifications as exemplified in compounds 11–15.

Table 1.

Symmetric headpiece substitutions and protein synthesis inhibitory activity.

Compounda	R1 and R2	R3	IC50b
11		7	10
12		8	7
13		8	7
14		7	5
15		8	8
16		8	100
17		7	>1000
18		7	800
Kan			0.4
Tet			2.8

^aKan, kanamycin; Tet, tetracycline.

^bCoupled in vitro transcription-translation assay with Escherichia coli extracts (μM).⁴

Table 2 summarizes headpiece SAR with asymmetric configurations of R¹ and DAP (R²). For most headpiece configurations there are as many as three R³ substitutions on the triazine core (6), which allowed detailed insight into headpiece (R¹) SAR. Interestingly, a mono-DAP compound (19) (R¹ = H) has similar activity as di-DAP compounds (12, 20 and 21). Consistent with the results shown in Table 1, active headpieces in symmetric configuration (13 and 14) also resulted in active inhibitors when they were asymmetrically combined with DAP (22–24, 25 and 29–31) (Table 2). Conversely, a headpiece with symmetric configuration that resulted in poor activity (Table 1; 16) was found to reduce activity when asymmetrically combined with DAP (38–40). In general, the replacement of an amino group with a hydroxy group on five- (32–34) or six-membered ring headpiece (38–40) slightly reduced activity compared to the di-amino substitutions (29–31 and 12, 20–21, respectively). Similarly, other mono-hydroxy substitutions exhibit moderate to weak activity (35–37, 41–43 and 44–46). An azetidin-3-ylamine substitution was tolerated (52–53), while a pyrrolidine-3,4-diamine substitutions (50–51) resulted in weak translation inhibitors. Di-hydroxy substitutions (47–49) were found to yield particularly weak inhibitors and appeared to destabilize the binding of the asymmetric DAP scaffold to its target. An acetamide substitution on the piperidine abolished activity (54–55). In summary, these results show that a symmetric substitution of two DAP moieties on the triazine core represents an optimal configuration for inhibitors of bacteria translation and growth.^4,10 However, this work also shows that there is room to alter the headpiece configuration while retaining antibacterial activity, which may be used to optimize other pharmaceutical properties such as bioavailability.

Table 2.

Asymmetric (R² is DAP) headpiece substitutions.

Compda	R1	R3	IC50	Ecb	Sac
19	H	8	16	4	nd
12		8	10	1	2
20		9	7	2	16
21		10	3	4	8
22		8	10	4	8
23		9	18	16	32
24		10	1	16	8
25		8	8	2	8
26		8	8	2	4
27		9	11	8	16
28		10	1	16	8
29		8	10	4	2
30		9	12	8	32
31		10	13	32	2
32		8	19	4	8
33		9	13	16	32
34		10	15	16	8
35		8	10	8	16
36		9	19	16	32
37		10	21	16	8
38		8	17	2	8
39		9	40	8	64
40		10	17	16	16
41		8	310	16	16
42		9	40	64	32
43		10	44	16	8
44		8	21	4	16
45		9	31	32	32
46		10	27	32	16
47		8	72	16	32
48		9	77	32	64
49		10	43	32	32
50		9	58	16	16
51		10	36	32	16
52		9	14	8	nd
53		10	11	16	16
54		9	>1000	nd	nd
55		10	>1000	nd	nd
Cm			1.5	4	8

nd, not determined.

^aCm, chloramphenicol.

^bE. coli ATCC 25922 minimum inhibitory concentration (μg/mL)⁴

^cS. aureus ATCC 25923 (MIC).

Summary and Perspective

This paper focused on headpiece optimization on our novel DAPT translation inhibitors. The active DAP pharmacophore was derived from structural information of the molecular recognition between the conserved 2-DOS scaffold of aminoglycosides and the A-site RNA.^5,6 This work expands on our efforts to rationally design 2-DOS mimetics, including the aminoazepane and novel acyclic deoxystreptamine scaffolds.¹⁰ After the initial design of DAPT, the series then evolved through an iterative process of parallel synthesis and high throughput testing that led to promising lead compounds. Rapid medicinal chemistry elaboration can now be used to potentially generate a clinical candidate that may improve on the pharmaceutical properties of aminoglycosides, including persisting resistance, poor bioavailability and undesirable toxicity.

In a future communication we will describe SAR around protein binding and the effect of serum on antibacterial potency. In the context of this study, a particularly intriguing observation was found with compound 22 in that the E. coli MIC improved in the presence of serum protein. This synergistic activity with serum was titratable and reached a peak reduction in MIC of 16-fold. Separately, particular core and tailpiece substitutions were found to reduce the serum MIC shift. A striking feature of the DAPT series is their potent antimicrobial activity against Pseudomonas aeruginosa, a pathogen for which there is a unmet clinical need to develop new antibiotics.² In one study, 53 clinical isolates of P. aeruginosa were tested against a DAPT compound (compound 1a in reference ⁴) and commercial antibiotics. The MIC₉₀ of the DAPT compound was 8 μg/ml, while the seven reference antibiotics that included amikacin, imipenem, ciprofloxacin, ceftazidime and tazobactam/piperacillin, all had higher MIC₉₀ values. Based on these and other findings we are particularly interested in developing the DAPT compounds as novel anti-P. aeruginosa agents.