Progress towards a stable cephalosporin-halogenated phenazine conjugate for antibacterial prodrug applications

Abstract

Resistant bacteria successfully evade the action of conventional antibiotic therapies during infection, often leading to significant illness and death. Our lab has discovered halogenated phenazine (HP) analogues which demonstrate potent antibacterial activities through a unique iron-starving mechanism. Herein, we describe synthetic efforts towards a stable cephalosporin-HP conjugate prodrug with the aim of translating HPs into useful clinical agents. Cephalosporin-antibiotic conjugates offer multiple advantages for antibacterial design, including the release of active agents through the targeting of intracellular cephalosporinase following selective ring-opening of the beta-lactam warhead. During these studies, carbonate-linked cephalosporin-HP conjugate 16 was synthesized; however, we were unable to successfully remove the ester group required for cephalosporinase processing. Cephalosporin-HP 16 was then utilized as a probe to investigate the stability of the carbonate linker in antibacterial assays and, as predicted, this compound proved to be inactive against Staphylococcus aureus (MIC > 100 μM). The lack of 16’s antibacterial activity can be attributed to the carbonate linker remaining intact throughout the MIC assay, thus not liberating the active HP moiety. These efforts have led to a more stable cephalosporin-HP conjugate joined through a carbonate linker compared to a highly unstable ether linked analogue we previously reported.

Pathogenic bacteria are able to effectively evade the action of conventional antibiotic therapies, leading to a significant number of resistant life-threatening infections.^1–4 In the United States alone, there are nearly 2.8 million cases of antibiotic-resistant infections that result in > 35,000 deaths each year.⁴ Bacteria utilize several well-defined resistance mechanisms that enable them to escape the activities of conventional antibiotics, which include: (1) changes in membrane chemistry to reduce antibiotic penetration, (2) removal of antibiotics through efflux pumps, (3) mutations of antibacterial targets, and (4) enzymatic inactivation of antibiotics.^2–10 In addition, very few new classes of antibiotics have entered the clinic over the past several decades further amplifying clinical challenges that result from resistance.³ Agents that operate through unique modes of action are desperately needed to address antibiotic resistance and select drug development strategies can enhance the likelihood of effectively treating multidrug resistant bacterial infections in the clinic.

Inspired by phenazine antibiotics, our group has discovered and explored an extensive series of halogenated phenazine (HP) small molecules that demonstrate excellent antibacterial activities against several antibiotic-resistant isolates of Staphylococcus aureus, S. epidermidis and Enterococcus faecium.^11–18 In addition to targeting free-floating planktonic bacteria, HP analogues are able to eradicate surface-attached biofilms formed by methicillin-resistant S. aureus (MRSA), methicillin-resistant S. epidermidis (MRSE), and vancomycin-resistant E. faecium (VRE). Using RNA-seq technology, we recently demonstrated that HP-14 (a potent analogue) induces rapid iron starvation in established MRSA BAA-1707 biofilms as its mode of action.¹⁷

Our active halogenated phenazine agents require the 1-hydro-xyphenazine moiety to bind, or complex, to iron (coordination of the hydroxyl group’s oxygen atom and adjacent nitrogen forms a stable 5-membered ring when bound to iron) providing an excellent platform to functionalize for the development of prodrug analogues regarding translational efforts (Fig. 1).^14,16–18 In previous studies, we have reported several carbonate-based prodrug versions of our most potent HP analogues (at the 1-hydroxyl position of the HP scaffold) and found these compounds to be stable and more soluble, yet very active in antibacterial assays.^15,16 In addition, we reported the initial synthesis of an ether-linked cephalosporin-HP conjugate; however, this molecule proved highly unstable and fragmented during NMR experiments (Fig. S1).¹⁶

Fig. 1.

Design of cephalosporin-halogenated phenazine conjugate prodrug 2.

Herein, we describe our efforts towards developing a more stable cephalosporin-halogenated phenazine conjugate prodrug using a synthetic approach that enabled the rapid assembly of a carbonate linker. The beta-lactam warhead of the cephalosporin scaffold is designed for intracellular bacterial targeting of cephalosporinase enzymes involved in cell wall synthesis (Fig. 2).^19–27 Following entry into a bacterial cell, a key serine residue of a cephalosporinase (or beta-lactamase) enzyme will undergo nucleophilic attack with the electrophilic beta-lactam motif of the cephalosporin scaffold leading to covalent modification of the target enzyme and subsequent fragmentation to release carbon dioxide and the halogenated phenazine. As such, the HP prodrug system is designed to stay intact by keeping the HP-linked to the cephalosporin scaffold until the beta-lactam moiety is first processed (see Fig. 3).

Fig. 2.

Cephalosporin-halogenated phenazine conjugate 2 and cephalosporinase-triggered release of the active halogenated phenazine moiety.

Fig. 3.

Progress to identify a stable carbonate linker regarding HP-cephalosporin conjugate prodrugs and ongoing work.

During these investigations, our initial synthesis goal was to convert 7-aminocephalosporanic acid (7-ACA, 3; commercially available) into target compound 7 (Scheme 1A). 7-ACA (3) underwent smooth esterification to 4 using thionyl chloride in methanol (90% yield)²⁸; however, we were unable to advance 4 to amide 5 despite several attempts. In a second route, 7-ACA was subjected to an amide-forming reaction upon treatment with thiophene-2-acetyl chloride to yield 6 in 85% yield.²⁹ The carboxylic acid in 6 was readily converted to its corresponding methyl ester 5 in 98% yield following treatment with (trimethylsilyl)diazomethane; however, base-promoted hydrolysis of the acetate group with potassium carbonate in methanol/water led to undesired lactone 8 as the sole product in 62% yield. Despite several attempts, we were unable to produce our initial target cephalosporin compound 7.

Scheme 1.

Synthetic efforts to modify the cephalosporin moiety. Reagents and conditions: (a) SOCl₂ MeOH, reflux, 5 h, 90%; (b) thiophene-2-acetic acid, EDCI, HOBt, NMM, CH3CN, rt; (c) thiophene-2-acetyl chloride, N,O-bis(trimethylsilyl)acetamide, CH₃CN, 85%; (d) TMSCHN₂, PhMe/MeOH (v/v = 4:1), 98%; (e) K₂CO₃, MeOH/H₂O (v/v = 5:1), 0 °C, 62%; (f) tert-Butyl acetate, BF₃•Et₂O, rt; (g) tert-Butyl acetate, BF₃•Et₂O, rt, 77%; (h) thiophene-2-acetyl chloride, pyridine, CH₂Cl₂, rt, 67%; (i) K₂CO₃, MeOH/H₂O (v/v = 5:1), 0 °C; (j) thiophene-2-acetyl chloride, N,O-bis(trimethylsilyl)acetamide, CH₃CN, 75%; (k) Diphenyldiazomethane, CH₂Cl₂, 0 °C, 5 h, 72%.

Since we were unable to prevent methyl ester 5 from undergoing lactone formation to 8 in the previous route, we designed a third route to utilize the more sterically encumbered tert-butyl ester to avoid the undesired lactonization reaction (Scheme 1A). 7-ACA (3) was treated with tert-butyl acetate and boron trifluoride to afford the corresponding tert-butyl ester 10 in 77% yield.³⁰ Following the installation of the desired tert-butyl ester group, 10 was then acylated with thiophene-2-acetyl chloride to form amide 9 in 67% yield following a literature protocol.³¹ The deacetylation of 9 has been reported with the use of an enzyme³²; however, we did not pursue this avenue. Instead, we attempted the deacylation of 9 with potassium carbonate in methanol/water; however, this reaction led to an inseparable mixture of 11 and 12. Although we were able to observe very small amounts of the desired compound 11 by NMR, the olefin migration product 12 was the major product. Despite this route overcoming the lactonization issue from the previous route, the olefin migration prevented this synthetic pathway from moving forward.

At this stage, we decided to pursue the synthesis of a suitable cephalosporin moiety from 13 in a fourth route (Scheme 1B). The amine of 13 was selectively acylated with thiophene-2-acetyl chloride in the presence of N,O-bis(trimethylsilyl)acetamide to afford amide 14 in 75% yield. Diphenyldiazomethane was then utilized to selectively functionalize the carboxylic acid of 14 to yield ester 15 in 72%.³³ We found compound 15 to be quite stable and we were encouraged to find that lactonization (to undesired compound 8) did not occur on this substrate.

With alcohol 15 in hand, we focused our attention on coupling 15 with HP 1 to yield carbonate-linked conjugate 16 (Scheme 2) upon treatment with triphosgene. Initially, we planned to condense 15 with triphosgene to yield an initial chloroformate intermediate, followed by treatment of HP 1 to yield the target carbonate 16. Unfortunately, this experimental approach proved unsuccessful as was initial treatment of 1 with triphosgene followed by the addition of alcohol 15. Interestingly, when a stoichiometric quantity of triphosgene dissolved in dichloromethane was added dropwise to a mixture of alcohol 15 and HP 1 in the presence of pyridine and triethylamine, the target carbonate 16 was observed in 72% yield on 270 mg scale.

Scheme 2.

Synthesis of carbonate-linked cephalosporin-HP conjugate 16 and efforts towards removal of the benzhydryl ester moiety. Reagents and conditions: (a) triphosgene, pyridine, triethylamine, CH₂Cl₂, 0 °C, 2 h, 72%.

With carbonate-linked cephalosporin-HP conjugate 16 in hand, we attempted to remove the benzhydryl ester moiety to target structure 2 using relatively standard trifluoroacetic acid (TFA) based protocols (Scheme 2); however, despite many efforts, we were unable to selectively cleave of the benzhydryl ester moiety as planned. Additional attempts were made to convert 16 to target structure 2 under hydro-genolysis conditions, but those efforts were also met with failure. The only product identified from these experiments was HP 1, which forms following the cleavage of the carbonate linker. In addition, we made several attempts to synthesize target structure 2 directly by reacting 1 with 14 in the presence of triphosgene; however, these attempts were also met with failure and we are currently working around these synthesis problems in an effort to access HP 2 and related analogues.

We were concerned about the overall stability of the carbonate linker in 16 since HP 1, the only observable product from the attempts to remove the benzhydryl ester, would result from carbonate fragmentation. To probe the stability of the carbonate-linker in 16, we evaluated this molecule in an MIC assay against MRSA to determine if this compound possesses any antibacterial activities. We hypothesized that 16 would not demonstrate antibacterial activity as the benzhydryl ester moiety on the cephalosporin scaffold is not a good substrate for a cephalosporinase to cleave the beta-lactam warhead that is required for subsequent carbonate fragmentation and release of active HP 1. When evaluated for antibacterial properties against MRSA BAA-1707, 16 proved inactive at the highest test concentration (MIC > 100 μM). As predicted, this “inactive” antibacterial result can be attributed to the carbonate linker of 16 remaining intact throughout the MIC assay and not liberating the active HP 1 moiety, which has an MIC ~ 1 μM against this MRSA isolate (see Supporting information for an image of this MIC assay).¹⁶ Despite the fragmentation of 16’s carbonate during chemical synthesis efforts, this linker proved to be highly robust in antibacterial assays.

In conclusion, we have synthesized a carbonate-linked HP-cephalosporin conjugate that demonstrates significant improvements in stability compared to an ether-linked conjugate analogue from our previous studies. Despite our progress, additional efforts are required to generate the carboxylic acid version of 16, which we hypothesize to be an ideal substrate for cephalosporinase-mediated processing and HP release and translational efforts. We believe that optimized HP-cephalosporin conjugate prodrugs could dramatically impact the treatment of antibiotic-resistant infections in humans and we look forward to continued advances in this area.