High-Throughput Tiling of Essential mRNAs Increases Potency of Antisense Antibiotics

Abstract

Antimicrobial resistance is outpacing drug discovery, creating an urgent need for precision-based strategies to counteract resistant pathogens. Peptide nucleic acid (PNA)-based antisense molecules offer a promising approach by selectively inhibiting essential bacterial mRNAs, but their design rules for optimal efficacy remain incompletely understood. Here, a scalable high-throughput platform is developed for the nanomolar-scale one-shot synthesis of PNAs as carrier peptide conjugates (PPNAs). Parallel synthesis of up to 1,536 PPNAs composed of up to 21 PNA or peptide building blocks enabled systematic, base-by-base analysis of RNA hybridization, mRNA inhibition, and antimicrobial activity across nine essential genes in uropathogenic Escherichia coli. The accuracy and robustness of this high-throughput tiling platform are demonstrated through in-depth analysis of the acpP mRNA and identify potent antisense inhibitors of rpsH, ftsZ, and murA. This approach provides an efficient and scalable route to design and optimize PNA-based antimicrobials, facilitating empirical testing across diverse bacterial targets. By enabling large-scale exploration of the relevant mRNA sequence space, the sequence tiling platform accelerates the discovery of antisense-based antimicrobials, offering a scalable strategy to develop precision therapies against various pathogens and combat resistance.

Keywords: Antisense Antibiotics; Asobiotics; PNA; acpP; ftsZ; mRNA; murA; rpsH.

Figure 1

Nanomolar‐scale direct‐to‐biology screening approach. Schematic representation of the general workflow. 1). Starting from the selected (essential) target mRNA (pink), a library of PNA binders is created with a 1 base‐pair (bp) offset along the TIR, including the translational start codon “AUG”, which translates into “CAT” in the respective PNA antisense sequences. 2). PNAs were produced in parallel, in an automated cellulose‐based set up. 3). For testing the efficacy of bacterial growth inhibition, PNAs are directly synthesized as CPP‐conjugates, i.e., fused to (KFF)₃K, and subjected to broth microdilution assays for determination of their minimal inhibitory concentrations.

Figure 2

Coupling conditions for automated parallel CPP‐PNA (PPNA) conjugate synthesis. A). Schematic representation of PNA design and synthesis. PNA sequences are designed to hybridize along the target mRNA's TIR (here exemplarily shown for acpP) encompassing the translational start codon “AUG” which translates into “CAT” in the cognate PNA sequence. PNAs are then synthesized on a highly‐derivatized cellulose disc (n = 100 nmol) support following the outlined coupling, capping, and deprotection cycle. PPNAs are synthesized from C to N termini and cleaved in parallel in 96‐well format. B). Comparison of coupling chemistry in parallel PNA synthesis. In the prolonged parallel PNA synthesis, purities and yields are superior with the more stable and less reactive OxymaPure/DIC compared to PyOxima/DIPEA. C,D). Systematic analysis of the effect of truncation on yields. ©. Truncation of Cell Penetrating Peptide (CPP). (D). Truncation of Peptide Nucleic Acid (PNA). Sequence length of the conjugate does not negatively affect synthesis yield. Major side products are the capped/acetylated truncations.

Figure 3

MIC assays of our disc‐based PPNA (dPPNA) versus commercial PPNA (cPPNA) constructs. A). Schematic representation of the cellular mode of action of PPNAs in bacteria. The CPP mediates cell entry and dependent on the CPP used, here (KFF)₃K, it undergoes degradation in the extracellular or periplasmic space. In the cytoplasm, the PNA hybridizes with its target mRNA and based on the binding region leads to the inhibition of translation. When targeting essential bacterial mRNAs this ultimately causes bacterial death. B). Direct‐to‐biology validation against UPEC 536 used at 10^5 cfu mL⁻¹ in Mueller Hinton broth. Determination of the PPNAs MICs in a 384 well plate setup. Comparison of disc‐produced and HPLC‐purified conjugates titrated from 20–1.25  to 10–0.6 µm, respectively. A scrambled PNA control (“cactatctc”) was included. The lowest concentration of a PNA that inhibits growth, i.e., with a final OD₆₀₀ of ‘<0.05′, indicates the MIC. All PNA sequences are shown and their conjugation to the cell penetrating peptide (KFF)₃K is indicated. The constructs JVpna‐1763 and JVpna‐178 are scrambled sequence controls (scr ctrl) for each of the two PPNA sets. Water was added in an equal volume serving as growth control (ctrl). The final OD at 600 nm after 24 h is shown in a white‐to‐magenta color gradient. Grey cells indicate conditions that were not tested. MIC testing for dPPNA constructs was performed two to five times, and two times for all cPPNA constructs.

Figure 4

Disc‐based PNA synthesis platform enables simple identification of optimal PNA sequences with minimal length requirements (acpP). Direct‐to‐biology validation of 6–11mer PNA sequences complementary to the acpP target mRNA against UPEC 536 used at 10^5 cfu mL⁻¹ in Mueller Hinton broth. Magenta‐white heatmaps show the final optical density (OD₆₀₀) at 24 h post treatment together with each PNA's purine content (blue‐scale) and its predicted melting temperature (grey‐scale; https://www.pnabio.com/support/PNA_Tool.htm). The lowest concentration of a PPNA that inhibits growth, i.e., with a final OD₆₀₀ of “<0.05′, indicates the MIC. Scrambled PPNA controls (2nd last row per panel) were included for each set of PPNAs. MIC assays were performed at least two times and the average OD₆₀₀ is shown.

Figure 5

Sequence tiling of additional essential bacterial genes known to be druggable by conventional antibiotics or PPNAs. A). Systematic PNA sequence tiling analysis of drug‐related target genes or known PNA targets using dPPNAs. Sequences were designed to be complementary to the translation initiation region of the respective target mRNA within a window spanning nucleotides −14 to +11 each with a 2 bps offset. PPNAs were tested against UPEC 536 (10^5 cfu mL⁻¹) in Mueller Hinton broth and the final OD₆₀₀ after 24 h is shown as green‐white heatmap. MIC assays were performed at least two times and the average OD₆₀₀ is shown. B). Validation of the growth inhibitory activity of selected dPPNA hits targeting rpsH, murA, and ftsZ. Respective PPNAs were purchased in HPLC‐grade from Peps4LS GmbH (cPPNA). The gold standard acpP cPPNA was included as positive control, while a sequence unrelated cPPNA was added as scrambled control (scr). Water was added in an equal volume serving as growth control (ctrl). cPPNAs were titrated from 10 to 0.6 µm and tested against UPEC 536 (10^5 cfu mL⁻¹) in Mueller Hinton broth. The final OD at 600 nm after 24 h is shown in a pink‐to‐green color gradient. MIC testing was performed four times and the average OD is shown. The lowest concentration of a PPNA that inhibits growth, i.e., with a final OD600 of “<0.05”, indicates the MIC (summarized in Table S3, Supporting Information). C. UPEC 536 was treated with the indicated murA‐targeting dPPNAs at 5  and 10 µm for the time periods shown above. An acpP‐targeting cPPNA (‘ctcatactc”, JVpna‐177, 2.5 µm) and a respective scrambled cPPNA (“cactatctc”, JVpna‐178, 2.5 µm) were used as an internal positive and negative control, respectively. One representative example out of two spotting assays is shown.

Figure 6

Microarray PNA/RNA hybridization study. A). Schematic representation of the PNA/RNA hybridization read‐out. Microarray slides with immobilized acpP and rpsH PNA sequences are incubated with fluorescently 3′‐labeled template RNA probes and then imaged for binding strength quantification. Simultaneous incubation with differently labeled acpP‐ or rpsH‐specific RNA probes mirrors binding competition and specificity. B). Exemplary acpP‐specific hybridization outcome for a 10 nt acpP Cy5 3′‐labeled RNA probe (top), in presence of the respective Cy3 3′‐labeled rpsH probe (bottom) with noted bases overlap between probe and PNA sequence displayed on the array. C). acpP (top) and rpsH (bottom)‐specific hybridization outcome for 26 nt 3′labeled‐RNA probes. On the left, acpP‐based PNA sequence are displayed, meanwhile rpsH‐based PNA sequences are displayed on the right. Each microarray displays the library in duplicate. The experiments were performed three times independently and the average binding intensities are shown. C. Summary of MIC, purine content, melting temperature, and normalized binding intensities (BI) results for best performing sequences for each PNA length, based on MIC outcomes.