Inhibition of Bacterial Gene Transcription with an RpoN-based Stapled Peptide

Summary

In response to environmental and other stresses, the σ⁵⁴ subunit of bacterial RNA polymerase (RNAP) controls expression of several genes which play a significant role in the virulence of both plant and animal pathogens. Recruitment of σ⁵⁴ to RNAP initiates promoter-specific transcription via the dsDNA denaturation mechanism of the cofactor. The RpoN box, a recognition helix found in the C-terminal region of σ⁵⁴, has been identified as the component necessary for major groove insertion at the -24 position of the promoter. We employed the hydrocarbon stapled peptide methodology to design and synthesize stapled σ⁵⁴ peptides capable of penetrating Gram-negative bacteria, binding the σ⁵⁴ promoter, and blocking the interaction between endogenous σ⁵⁴ and its target DNA sequence, thereby reducing transcription and activation of σ⁵⁴ response genes.

eTOC Blurb

The bacterial transcriptional regulator σ⁵⁴ is involved in the virulence of pathogenic bacteria. Stapled peptides blocking the interaction between σ⁵⁴ and its target DNA promoter inhibit activation of σ⁵⁴-dependent genes in bacteria. These compounds provide an avenue to target bacterial pathogens and circumventing therapeutic drug resistance.

Introduction

Transcription factors lie at the crux of cellular homeostasis in all living systems. The transcription of genes into messenger RNA ultimately leads to the expression of proteins responsible for the execution of cellular functions with a high degree of temporal and spatial specificity; thus, disruptions in gene transcription can have deleterious consequences in the regulation of biological processes necessary for cell survival. Alterations such as the overexpression or deletion of a transcription factor are responsible for numerous disease states, underscoring the widespread interest in them as therapeutic targets in human disease, most notably in the cancer field. Targeting a transcription factor is fraught with numerous challenges, not the least of which is the difficulty of designing small molecules capable of disrupting weak interactions that span wide surfaces.

Much like in eukaryotic systems, gene transcription in prokaryotes follows a process involving several macromolecules working in concert. The initiation of bacterial transcription requires complexation of the RNA polymerase (RNAP) core enzyme with a sigma (σ)factor. σ-factors bind to the catalytic core of RNAP to initiate gene transcription after recognition of a promoter sequence and melting of dsDNA, and are functionally divided into housekeeping σ-factors, which control most prokaryotic genes, and alternative σ-factors, which bind to specific promoter sequences in response to environmental and physiological stresses (Cannon et al., 1993). The alternative factor σ⁵⁴ directly regulates the transcription of mostly non-essential genes involved in nitrogen metabolism (Reitzer and Schneider, 2001). In several species of pathogenic bacteria such as Pseudomonas aeruginosa, Vibrio cholera, Borrelia burgdorferi, and Helicobacter pylori, the σ⁵⁴ regulon controls the production of virulence factors involved in toxin release, quorum sensing, flagellar biosynthesis, and biofilm formation (Kazmierczak et al., 2005).

The structural features of σ⁵⁴ have been characterized by NMR and provide a window into the intricacies of its cellular function (Doucleff et al., 2007). σ⁵⁴ contains leucine zipper motifs that enable it to melt DNA in addition to hydrophobic regions which make contacts with RNAP. The C-terminal DNA binding domain of σ⁵⁴, or RpoN box, binds to a specific promoter region leading to the formation of a complex activated by an ATP-dependent enhancer protein (Buck et al., 2000). σ⁵⁴ engages its target promoter, located between positions –12 and –24 base pairs (bp) upstream of the transcription start site, via a conserved sequence of residues folded into an α-helix. Multiple hydrogen bonds and van der Waals interactions are essential for specific binding of the RpoN box to the major groove of DNA (Figure 1A) and subsequent activation of gene transcription. Given its role in the pathogenicity of many bacterial species, we hypothesized that blocking the interaction between endogenous σ⁵⁴ and its promoter element would downregulate the expression of virulence genes.

Figure 1.

(A) NMR model of the C-terminal domain of σ⁵⁴ interacting with DNA, with the RpoN box highlighted in red, and the −24 and −12 promoter sequences highlighted in blue. Residues essential for binding which also make key contacts with DNA are highlighted in red on the RpoN sequence. (B) Helical wheel of the RpoN box with staple positions indicated. (C) Linear peptide sequence of Sσ⁵⁴. X = (S)-2-(4′-pentenyl)-alanine, Z = (R)-2-(7′-octenyl)-alanine, B = norleucine. All peptides contain β-alanine in the first position as a spacer (Supplementary Table 1). (D) Circular dichroism spectra of Sσ⁵⁴ peptides.

Owing to a lack of any appreciable binding pocket, it has been extremely difficult to develop small molecules that target DNA specifically. There have been successes in this arena, most notably by Dervan and co-workers, who have developed programmable polyamides that bind specific DNA sequences and modulate eukaryotic gene expression (Dervan et al., 2005). Previously, our group and others (Walensky and Bird, 2014) have been involved in the development of hydrocarbon stapled peptides for their use as α-helix mimetics to target numerous protein-protein interactions involved in human disease. In many cases, these have led to compounds with potential therapeutic use; however, they have never been used to block a protein-DNA interaction successfully. Taking advantage of the helical shape of the RpoN box, we designed hydrocarbon stapled peptides that target the interaction between σ⁵⁴ and DNA. Stapled σ⁵⁴ peptides permeate Gram-negative bacteria and inhibit the transcription of genes activated by nitrogen depletion. Our data suggest that stapled peptides targeting bacterial transcription factors have potential for use as therapeutics to target the virulence properties regulated by σ⁵⁴ in pathogenic bacteria.

Results

Design and Synthesis of Stapled σ⁵⁴ Peptides

Gene expression in bacteria is accomplished by RNA polymerase (RNAP) in concert with a σ factor. σ factors guide RNAP to specific promoters and assist in melting dsDNA to initiate gene transcription. Bacteria possess multiple σ factors, each controlling specific responses to different environmental stimuli. Unlike other σ factors, σ⁵⁴ requires the help of upstream ATP-dependent activator proteins to trigger transcription. A solution structure of σ⁵⁴ bound to its promoter shows that a helix-loop-helix segment on the C-terminal portion of σ⁵⁴ interacts with a region in the major groove of DNA located between −24 and −12 bp upstream of the gene transcription start site (Doucleff et al., 2007). Specifically, binding of σ⁵⁴ to DNA takes place via the RpoN box α-helix, which is conserved across multiple bacterial species (Doucleff et al., 2005) (Figure 1A). Mutational studies have shown that several residues around the entire span of the RpoN α-helix are involved in a complex hydrogen bond network in which DNA wraps around the helix and facilitates binding and activation of transcription (Wang and Gralla, 2001). A helical wheel analysis (Figure 1B) shows the amphipathic character of the RpoN box with mostly hydrophobic residues on one face of the helix and polar amino acids arranged on the opposite side. With these design considerations in mind, we synthesized a family of stapled σ⁵⁴ peptides (Sσ⁵⁴) in which we varied the size and location of the hydrocarbon crosslink and kept intact the residues responsible for binding to DNA (Figure 1C). We examined their secondary structure by circular dichroism and found that, while the same length wild type σ⁵⁴ peptide is unstructured, all Sσ⁵⁴ peptides adopt an α-helical secondary structure in solution (Figure 1D).

Sσ⁵⁴ Peptides Bind to the σ⁵⁴ Promoter Element

With compounds in hand, we investigated whether Sσ⁵⁴ peptides bind to a DNA oligo containing a target promoter sequence. We employed a fluorescence polarization binding assay to evaluate the ability of Sσ⁵⁴ peptides to engage a fluorescein-tagged 30 bp oligo containing the sequence for the glnA σ⁵⁴-dependent promoter sequence (glnAp2) (Smith and Bell, 2016) (Figure 2A). While the unstructured, wild type peptide did not bind appreciably to the target DNA sequence, all Sσ⁵⁴ peptides bound to DNA with similar affinities (K_d = 0.86–5.98 μM). Testing a scrambled peptide in the same assay revealed only a weak binding affinity (Figure S1A). Additionally, none of the peptides bound to a FITC-tagged scrambled 30-mer to any appreciable extent (Figure S1B). We corroborated these results by performing gel shift assays using the FITC-glnAp2 probe. Exposure of stapled peptide Sσ⁵⁴-2 to FITC-glnAp2 showed a significant reduction in the amount of free probe (Figure 2B, lanes 1 and 2). We likewise performed the experiment in the presence of the C-terminal fragment of σ⁵⁴ (residues 338-398) which contains the RpoN box, and we observed that σ⁵⁴(338-398) binds to the probe (lane 3), and Sσ⁵⁴-2 disrupts this binding interaction effectively (lane 4) in a dose-dependent fashion (Figure S1C). Performing the experiment using a scrambled DNA probe (Figure S1D) shows that none of the stapled peptides can interact with the probe, corroborating the fluorescence polarization results (Figure S1B).

Figure 2.

(A) Fluorescence polarization binding analysis of Sσ⁵⁴ peptides with FITC-glnAp2 30-mer. (B) Gel shift assay showing the association of Sσ⁵⁴-2 with FITC-glnAp2 and the disruption of σ⁵⁴(338-398) bound to FITC-glnAp2. (C) Polyacrylamide gel showing digestion of a 5′-6-FAM glnAp2 218-mer with DNase I. V = vehicle. (D) Densitometric quantification of remaining DNA relative to starting material (lane 3) from (B). Data are represented as mean ± SEM.

To further confirm the ability of Sσ⁵⁴ peptides to bind to DNA, we performed a DNase digestion assay by combining a 6-FAM-labeled dsDNA oligo containing Escherichia coli glnAp2 with peptide and exposing the mixture to DNase I. The reaction mixtures were separated by polyacrylamide gel electrophoresis and analyzed by fluorescence scanning (Figure 2C). Treatment with the wild type σ⁵⁴ peptide (lane 5) displayed a significant amount of DNA degradation as observed in vehicle-treated DNA (lane 4), confirming the fluorescence polarization results showing a lack of binding by this peptide. On the other hand, Sσ⁵⁴ peptides were far more effective at protecting DNA from enzymatic degradation. There were differences in the ability of each Sσ⁵⁴ peptide to shield DNA from cleavage, with Sσ⁵⁴-1 and -2 (lanes 6 and 7) being the most effective (Figure 2D). Notably, the binding affinities correlate with the ability of the peptides to protect DNA from degradation by an endonuclease. The differential binding of the peptides suggests that each of these compounds may result in varying degrees of activity in a biological setting.

Sσ⁵⁴ Peptides Enter Gram-Negative Bacteria by Diffusion

Armed with the knowledge that Sσ⁵⁴ peptides can bind to a target promoter, we investigated whether the compounds penetrated bacterial cells. While studies have been done on the ability of stapled peptides to enter human cells (Bird et al., 2016), there are no reports showing whether they can penetrate bacterial cells. Entry of hydrophobic compounds into Gram-negative bacteria is met with several physical barriers. In particular, the lipopolysaccharide (LPS) envelope surrounding the outer membrane of the cell is responsible for the inability of many organic compounds to penetrate Gram-negative bacteria (Denyer and Maillard, 2002). To evaluate whether Sσ⁵⁴ peptides enter bacterial cells, we treated BW25113 E. coli cells and PA01 P. aeruginosa cells with FITC-Sσ⁵⁴ peptides. Uptake was measured by flow cytometry (Figure 3A). In both bacterial species, all FITC-Sσ⁵⁴ peptides penetrated bacterial cells while the unstructured wild type FITC-σ⁵⁴ peptide did not. Bacterial cells also possess the ability to control transport of macromolecules through ATP-dependent transporters (Mousa and Bruner, 2016), so to rule out that internalization of Sσ⁵⁴ peptides is dependent on this system, we repeated the flow cytometry experiment with the addition of sodium azide as an ATPase inhibitor. Flow cytometry analysis showed no change in peptide permeability, suggesting that these compounds enter bacterial cells by diffusion. We confirmed the flow cytometry results by fluorescence confocal microscopy (Figure 3B). The images show that FITC-Sσ⁵⁴ peptides accumulate in the cytoplasm of bacterial cells while the unstructured wild type compound is not permeable. Taken together, these data show that stapled peptides are capable of penetrating Gram-negative bacteria.

Figure 3.

(A) Flow cytometry of BW25113 E. coli and PA01 P. aeruginosa treated with 4 μM FITC-Sσ⁵⁴ peptides. Data are represented as mean ± SD. (B) Fluorescence microscopy of BW25113 E. coli. FM 4-64: plasma membrane stain, FITC: Sσ⁵⁴. The scale bars represent 10 μm.

Stapled σ⁵⁴ Peptides Block the Upregulation of σ⁵⁴-dependent Genes

In E. coli, σ⁵⁴ is primarily responsible for triggering transcription in response to nitrogen starvation. We hypothesized that exposing E. coli cells to growth media deficient in nitrogen would initiate σ⁵⁴-mediated transcription of nitrogen metabolism genes, and subsequent treatment with Sσ⁵⁴ peptides would reverse or inhibit that transcriptional activation. To examine this, we cultured MG1655 E. coli cells in Gutnick minimal media (Brown, 2014) to induce nitrogen starvation (Figure S3). Cells were then treated with peptide, and RNA was later isolated and processed for RT-qPCR analysis (Figure 4A). We quantified RNA for the σ⁵⁴-dependent nitrogen starvation response genes glnA, yeaG and nac, which are involved in nitrogen fixation, metabolism, and assimilation, respectively (Reitzer and Schneider, 2001). We also examined pspA, a σ⁵⁴-dependent gene that responds to cell membrane disruptions and is not responsive to nitrogen levels, and we used the housekeeping gene cysG as a control. Exposure to nitrogen-deficient conditions activates transcription of glnA by a factor of 10, yeaG 100-fold, and nac close to 500-fold, while pspA was unaffected by nitrogen depletion. Treatment with the wild type σ⁵⁴ peptide showed no difference in mRNA levels when compared to the untreated control; however, exposure to Sσ⁵⁴ peptides, particularly Sσ⁵⁴-2 and Sσ⁵⁴-3, restricted the mRNA levels of the nitrogen response genes to the basal levels found in cells at the nitrogen-rich state (N+, Figure 4A). Some Sσ⁵⁴ peptides also increased the levels of pspA transcripts, which is not surprising given the involvement of this gene in responding to changes in the integrity of the bacterial envelope and membranes. Because nitrogen depletion with concomitant blockade of the transcription of nitrogen metabolism genes is likely to be toxic to bacterial cells, we verified that E. coli cells are alive at the dose used in the RT-qPCR experiment via cell viability (Figure S2B) demonstrating that the low levels of nitrogen metabolism gene transcripts detected in Sσ⁵⁴-treated cells are not due to cell death. We similarly evaluated whether the compounds were toxic to human cells, and the data show that, at these doses, Sσ⁵⁴ peptides are not toxic to human cells (Figure S2C) despite the ability of the peptides to penetrate human cells (Figure S3).

Figure 4.

(A) Relative normalized expression of σ⁵⁴-controlled genes after treatment with Sσ⁵⁴. Data are represented as mean ± SEM. N+ denotes nitrogen-rich media while N- indicates nitrogen-deficient media. (B) Glutamine synthetase transferase assay with A₅₄₀ values normalized to OD₆₀₀. Data are represented as mean ± SEM. (C) RNA-Seq heat map for transcripts that are differentially expressed in MG1655 E. coli in response to nitrogen starvation and treatment with peptides. RNA of E. coli grown under nitrogen-rich conditions (N+) was used as a control.

We also sought to confirm that inhibition of σ⁵⁴-dependent gene transcription led to a decrease in the nitrogen sensitivity response. We measured the intracellular activity of glutamine synthetase, the gene product of glnA, in a colorimetric assay (Rothstein et al., 1980), and found no appreciable enzymatic activity in cells treated with Sσ⁵⁴-2 and Sσ⁵⁴-3 (Figure 4B), which is consistent with our observations from the RT-qPCR experiment.

To understand more broadly the full-scale effects of stapled σ⁵⁴ peptides on the transcription of activated σ⁵⁴-dependent genes, we performed an RNA-Seq experiment to analyze the transcriptome of E. coli. MG1655 E. coli cells were grown in Gutnick media to induce nitrogen starvation, were treated with either DMSO, WT σ⁵⁴, Sσ⁵⁴-1, or Sσ⁵⁴-2, and RNA was isolated and processed for RNA-Seq. Comparison of the data set for Sσ⁵⁴-2 treatment versus DMSO control showed differential expression of 2,191 of the 4,273 genes detected, 1047 of which were positively regulated and 1140 were negatively regulated by a factor of 2 or greater. Gene set enrichment analysis (GSEA) (Subramanian et al., 2005) with all 155 E. coli gene sets with fewer than 500 genes per set from gene2go (Powell, 2014) showed that 6 gene sets were positively regulated with a false discovery rate of less 5%. These included carbohydrate and polysaccharide biosynthesis, amide biosynthesis and catabolism, and metal ion transport. 6 out of 155 gene sets were negatively regulated with a false discovery rate of less than 5%, including amino acid, organic acid, and organonitrogen catabolism. These results are consistent with the Sσ⁵⁴ having minimal impact on gene sets that describe a common function.

When GSEA was performed using gene sets for the RpoN (σ⁵⁴) and RpoH (σ³²) regulons, the RpoN regulon (Bonocora et al., 2015) was shown to be negatively regulated with a false discovery rate of less 0.1% (Figure S4). This result is consistent with Sσ⁵⁴-2 binding to the σ⁵⁴ promoter preventing transcription of genes under σ⁵⁴ control. In contrast, the RpoH regulon (Nonaka et al., 2006), which is regulated by the σ³² alternative sigma factor as part of the heat shock response, was neither positively nor negatively regulated by Sσ⁵⁴-2 (Figure S4). Similarly, GSEA of Sσ⁵⁴-1 versus DMSO showed the RpoN regulon was negatively regulated and the RpoH was not significantly impacted. Hierarchical clustering of the transcriptome data for genes under σ⁵⁴ (RpoN) and σ³² (RpoH) control (Figure 4C) clearly showed that genes from the RpoN regulon cluster together and were negatively regulated in the positive control (i.e., growth in nitrogen rich media) and Sσ⁵⁴-2 and Sσ⁵⁴-1 treatment, but not with WT σ⁵⁴ treatment. These data are consistent with and strongly support the selective, negative regulation of the RpoN regulon by Sσ⁵⁴.

Discussion

The emergence of antibiotic resistance along with the decreasing numbers of new antimicrobials available in the clinic underscores the need for creative ways to treat bacterial infections. We hypothesized that a direct approach targeting the transcriptional machinery of pathogenic bacteria might mitigate their virulence. We have previously synthesized stapled peptides that target transcription factors in human cells and restore the function of a tumor suppressor through the inhibition of a protein-protein interaction (Bernal et al., 2010). The present work takes this concept further by directly inhibiting the interaction between a transcriptional regulatory protein and its promoter. Targeting gene transcription with precision has been a longstanding goal in the search for potential therapeutics that minimize collateral damage. The complexity of eukaryotic genomes has made executing this strategy challenging (Edwards et al., 2016); thus, a study in a simpler system provides a springboard for the development of agents that control gene transcription with specificity. This strategy, however, is not without risks. σ⁵⁴ recognition sites are small and lend themselves to significant redundancy, increasing the likelihood of unexpected phenotypic responses. Here we present a case in which we target a σ⁵⁴-mediated cellular response to a specific environmental stress. Sσ⁵⁴ peptides successfully blocked the response without any appreciable side effects. Our work suggests that the inhibition of nitrogen sensitivity genes by treatment with Sσ⁵⁴ peptides is directly related to specific targeting of genes with active σ⁵⁴ promoters. It is likely that Sσ⁵⁴ peptides bind to other regions of the bacterial genome, raising the concerns about compound specificity; and if the compounds are nonselective, we would expect to see varied transcriptional responses. Nevertheless, our transcriptional profiling data strongly indicate that, even if the Sσ⁵⁴ peptides bound elsewhere, these other putative interactions are uneventful and do not affect the desired phenotypic outcome of the treatment. These results pave the way towards development of analogs to fine-tune the selective targeting of activated σ⁵⁴ promoters.

One of the bigger questions arising from this work was to determine whether stapled peptides could penetrate the multiple physical barriers erected by Gram-negative bacteria, the most notorious being the LPS envelope surrounding the outer membrane (Wiese et al., 1999). Hydrophilic compounds may enter Gram-negative bacteria through porins embedded in the outer membrane, but these channels limit the size of molecules that can go through to 600 Da or less, smaller than any of the compounds tested. Hydrophobic compounds can penetrate the LPS matrix, but its dense composition significantly slows their diffusion rates. Lipid A, a phosphorylated glycosamine, is an integral component of the LPS envelope, and its phosphate groups are linked to each other and held in place by bridging divalent cations. Some effective agents against Gram-negative bacteria are the polymyxins, polycationic antibiotics which function partly by destabilizing the envelope and promoting their own entry (Hancock and Chapple, 1999). The Sσ⁵⁴ peptides have charges of +3 and +4 at physiological pH, so it is likely that charge plays a role in their ability to penetrate E. coli and P. aeruginosa cells. Charge alone, however, is not sufficient to promote permeability, as the wild type σ⁵⁴ peptide does not enter cells despite also having a charge of +4. Sσ⁵⁴ peptides have a rigid α-helical shape with discrete hydrophobic and hydrophilic faces, and this likely allows them entry via disruption of the LPS matrix. This work demonstrates the potential of stapled peptides as antibiotics with the ability to access targets in the bacterial cytoplasm.

The search for agents to treat bacterial infections must keep up with the pace at which resistant pathogens are emerging. Antimicrobial peptides, in particular cationic host defense peptides, have shown promise as potential therapeutics (Overhage et al., 2008). Our work has shown that incorporating structural stability is essential in achieving bacterial cell permeability. Targeting the virulence properties of pathogenic bacteria opens a new therapeutic avenue that takes advantage of the hosts’ environment. By combining structural features that allow for permeability along with a defined genomic target controlling virulence, we are now able to target Gram-negative pathogens through a novel mechanism.

Significance

The emergence of bacterial drug resistance underscores the need for novel antibiotics. Gram-negative bacteria possess many physical barriers that make them impervious to most therapeutics. Additionally, virulence properties such as biofilm formation and toxin secretion make it significantly more difficult to target pathogens with synthetic agents. Disrupting virulence at the transcriptional level provides a new avenue to render bacteria vulnerable, but targeting protein-DNA interactions has been notoriously difficult. We show that using structurally defined hydrocarbon stapled peptides based on a DNA-interacting α-helix of σ⁵⁴ can block specific gene transcription in response to an environmental stress. These stapled peptides readily enter Gram-negative bacteria by diffusion and target σ⁵⁴ promoters in cells. This class of compounds will be developed for the treatment of infections by Gram-negative pathogens.

STAR Methods

Contact for Reagents and Resource Sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Federico Bernal (federico.bernal@nih.gov).

Experimental Model and Subject Details

All overnight bacterial cultures were either grown in LB broth (Fisher) or Gutnick media containing 10 mM NH₄Cl.

Method Details

Peptide Synthesis

Peptides were synthesized on a Tetras Peptide Synthesizer (Advanced ChemTech, Louisville, KY) using Fmoc-based solid phase peptide chemistry on Rink amide resin (30 μmol peptide per reaction). In brief, the synthesis protocol consists of removal of the Fmoc protective group with 25% piperidine in NMP, washing with NMP, and subsequent amino acid coupling using 10 equiv. amino acid (1 mL, 0.3 M), HCTU (0.99 mL, 0.3 M) and DIPEA (2.0 mL, 0.3 M) in NMP for 60 minutes before draining and washing. All peptide sequences include an N-terminal β-alanine spacing residue. Stapled peptides underwent a further ring-closing metathesis reaction on the resin-bound N-terminal Fmoc-protected peptide. After washing with NMP and 1,2-dichloroethane, the resin was exposed to three 6-hour cycles of bis(tricyclohexylphosphine)benzylidine ruthenium (IV) dichloride (Grubbs’ I catalyst, 3 mL, 1 mM) at room temperature in 1,2-dichloroethane. Peptides were primarily N-terminal acetylated for use in biochemical and cell-based assays. The acetylation reaction consisted of deprotection of the Fmoc group as outlined above, followed by reaction with neat acetic anhydride (1 mL) and DIPEA (2 mL, 0.3 M in NMP) for 1 hour. Other experiments made use of N-terminal fluoresceinated peptides. To this end, Fmoc-deprotected peptides were exposed to fluorescein isothiocyanate (2.8 mL, 25 mM) and DIPEA (0.2 mL, 0.3 M) in DMF for 12 hours. Final peptides were cleaved from the resin and fully deprotected by exposure to a solution containing 95% TFA, 2.5% water, and 2.5% triisopropyl silane for 2.5–3 hours. The cleaved peptides were precipitated into ice-cold 1:1 methyl-tert-butyl ether (MTBE), re-suspended in water and lyophilized as crude peptides overnight. The lyophilized peptides were purified by C-18 reverse phase HPLC on an Agilent 1200 HPLC system (Santa Clara, CA).

Circular Dichroism (CD) Spectroscopy

Acetylated compounds were dissolved in water to concentrations ranging from 60 to 75 μM. Final compound concentrations were determined by measuring sample absorbance at 205 nm using a NanoDrop2000 spectrophotometer (ThermoScientific, Wilmington, DE). Spectra were obtained on an Aviv Circular Dichroism Spectrometer, Model 420 (Aviv Biomedical, Inc, Lakewood, New Jersey) at 25°C. The spectra were collected using a 0.1 cm path length quartz cuvette (Hellma Analytics, Germany) with the following measurement parameters: wavelength, 240–190 nm; step resolution, 0.50 nm; averaging time, 5.0 sec per step. Spectra were processed using Aviv CDS Program software and converted to mean residue molar ellipticity using the cuvette path length (0.1 cm), the measured concentration, and the number of amino acids in the peptide (cross-linking amino acids and β-alanine cap were included as amino acids in this count).

Protein Purification

The C-terminal RpoN domain (CTD) of σ⁵⁴ (residues 338–398) from Klebsiella pneumoniae was cloned into the pET28b expression vector by PCR (Q5 Site-directed Mutagenesis Kit; New England Biolabs) and confirmed through Sanger sequencing. The plasmid was then transformed into E. coli Rosetta 2 (DE3) cells. The cells were grown at 37°C in 500 mL of LB medium and induced with 1 mM IPTG at OD₆₀₀: ~0.6. Four hours after induction cells were harvested via centrifugation and resuspended in 1X PBS supplemented with 1X Protease and Phosphatase Inhibitor Cocktail (ThermoScientific). Cells were lysed through sonication and the CTD was purified through Ni-affinity chromatography (Qiagen) and size exclusion chromatography (Superdex 75 10/300 GL; GE Healthcare).

Electrophoretic Mobility Shift Assays (EMSA)

FITC-glnAP2 DNA (5.625 nM) was initially incubated with our library of stapled peptides (20 μM) for 20 minutes at room temperature in the dark for 20 minutes, with DMSO as our negative control. The complexes were formed in 25 mM HEPES, 500 mM NaCl, 10% glycerol, pH 6.9. Next, the CTD protein (32 μM) was added to the mixture and further incubated at room temperature for 20 minutes. 12% polyacrylamide gels (29:1 acrylamide:bisacrylamide) were pre-run for 40 minutes at 4°C at 70 V. Samples were the n electrophoresed through the gels for approximately two hours at 10 V using 1X Tris-borate buffer. DNA was detected by fluorescence imaging on a Typhoon FLA 7000 (GE Healthcare).

Induction of Nitrogen Starvation

The genotypes and origins of all bacterial strains can be found in the key resources table. Nitrogen starvation was induced in one of two ways: MG1655 E. coli were grown in LB broth, then switched to M9 minimal media (1X M9 salts, MgSO₄, CaCl₂ and 20% (w/v) glycerol) containing either 40 mM (nitrogen rich) or 100 μM (nitrogen deficient) NH₄Cl. Conversely, overnight cultures were also grown in Gutnick minimal media (33.8 mM KH₂PO₄, 77.5 mM K₂HPO₄, 5.74 mM K₂SO₄, 0.41 mM MgSO₄) containing Ho-LE trace elements, 0.4% (w/v) glycerol and 10 mM NH₄Cl. The following day, cultures were prepared using Gutnick media with 3 mM NH₄Cl, a concentration that allows for complete depletion of nitrogen during 4 hours of growth. In Gutnick media, “nitrogen rich” is the point when OD₆₀₀ = 0.3, and “nitrogen deficient” being approximately 30 minutes after cultures have reached stationary phase.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Goat polyclonal anti-Glutamine Synthetase	Abcam	Cat# ab6585
Mouse monoclonal Anti-E coli RNA Sigma 54 (clone 6RN3)	BioLegend	Cat# 663303
Mouse monoclonal anti-GAPDH (clone GA1R)	Thermo Fisher	Cat# MA5-15738
Bacterial and Virus Strains
E. coli: MG1655	Coli Genetic Stock Center – Yale University	Cat# CGSC 6300
E. coli: BW25113	Coli Genetic Stock Center – Yale University	Cat# CGSC 7636
E. coli: Rosetta 2 (DE3)	Novagen	Cat# 71397
Chemicals, Peptides, and Recombinant Proteins
B-PER II	Thermo Fisher	Cat# 78260
2-methylimidazole	Alfa Aesar	Cat# A11949
2,4-dimethylimidazole	Sigma Aldrich	Cat# M50850
Adenosine diphosphate sodium salt	Sigma Aldrich	Cat# A2754
Hydroxylamine Hydrochloride	Sigma Aldrich	Cat# 159417
Sodium Tartrate	Sigma Aldrich	Cat# SX0795
Cobalt chloride hexahydrate	Sigma Aldrich	Cat# C2644
Sodium molybdate dihydrate	Sigma Aldrich	Cat# 221848
Copper chloride dehydrate	Sigma Aldrich	Cat# 459097
SYBR Select Master Mix	Applied Biosystems	Cat# 4472942
Critical Commercial Assays
CellTiter-Glo	Promega	Cat# G7571
BacTiter-Glo	Promega	Cat# G8231
Ammonia Quantification Assay	Sigma Aldrich	Cat# AA0100
RNeasy Mini Kit	Qiagen	Cat# 74106
Superscript III First strand cDNA synthesis system	Invitrogen	Cat# 18080051
Deposited Data
Deep sequencing data	GEO	Acc# GSE111317
Experimental Models: Cell Lines
Human: WS1 Fibroblast cells	ATCC	Cat# CRL-1502
Oligonucleotides
See Table S1 for DNA oligo and primer sequences	This study	N/A
Software and Algorithms
GraphPad Prism 7.01	GraphPad Software	https://www.graphpad.com/
SnapGene 3.0.3	GSL Biotech	www.snapgene.com
ImageJ	National Institutes of Health	https://imagej.net/NIH_Image
ChemDraw 15.1	PerkinElmer	http://www.cambridgesoft.com/software/overview.aspx

Flow Cytometry

BW25113 E. coli or PA01 P. aeruginosa cells were grown in LB media to an OD₆₀₀ of 0.4–0.6. 1.0 × 10⁶ cells were washed twice with PBS and treated with the FITC-labeled stapled-α-helical peptide (4 μM) and incubated at 37°C for 90 min. Samples were washed once with PBS and incubated at 4°C as pellets for 30 min. Trypan Blue (0.04% w/v) was added to all samples and removed and washed with PBS. The measurements were performed using a Beckman-Coulter Gallios Flow Cytometer in which a 488 nm laser was used to excite FITC and fluorescence was detected using the FL1 channel.

Confocal Fluorescence Microscopy

Bacterial Cells

MG1655 E. coli cells were grown to an OD₆₀₀ of 0.5–0.7 in LB media. The culture was split into separate 1 mL aliquots, centrifuged and resuspended in PBS. FITC-labeled compounds were added to a final concentration of 4 μM. Samples were incubated at 37°C in the dark for 90 minutes and then centrifuged to pellet the cells. The supernatant was discarded and washed twice with PBS. After centrifugation, the samples were incubated on ice as pellets in the dark for 30 minutes. FM 4-64 was added to a concentration of 2 μg mL⁻¹ and incubated for 5 min at 37°C in the dark. After cent rifugation and resuspension in PBS, 4 μL of each sample was pipetted onto the center of a square, 1% (w/v) agarose pad. Using a scalpel, the pad was placed directly atop a glass-bottom culture dish (35 mM, poly-lysine coated, MatTek Corporation). The cells were then visualized and images acquired on a Zeiss LSM780 confocal scanning microscope (Zeiss).

Human Cells

WS1 normal human fibroblasts (7.5 × 10⁵ cells) were seeded onto a white, clear 6-well plate in DMEM. Before seeding, a round glass coverslip was washed for 4 hours in 100% methanol, subsequently washed in ddH₂O, and placed inside each well. The cells were then treated with FITC-labeled peptides at a final concentration of 5 μM for 90 minutes at 37°C and 5% CO₂. The supernatant was discarded and the cells were fixed with formaldehyde solution (1.5% in 1X PBS) for 7 minutes. The formaldehyde solution was discarded and cells were subsequently washed for 30 min with 1X PBS. PBS was then aspirated, and the dsDNA was stained with 150 μL solution of DAPI (1:5000 in 1X PBS) (ThermoScientific) and incubated for 1 min. After aspirating the DAPI solution, a drop of antifade reagent in glycerol/PBS (ThermoScientific) was placed on a clean slide and the coverslip was gently placed cell down onto the slide. The edges were coated with nail polish and let dry for 30 min in the dark. The cells were then visualized and images acquired on a Leica TCS SP8 confocal laser scanning microscope (Leica Microsystems).

Fluorescence Polarization Binding Assays

FP assays were performed as described. Briefly, to determine the dissociation constants for peptide-DNA interactions, FITC-labeled oligonucleotides (5.625 nM) were incubated with different concentrations of stapled peptide (40 μM - 9.8 nM) at 25°C for 10 min in FPA buffer (50 mM Tris-HCl pH 7.5, 1 mM DTT, 0.1 mg mL⁻¹ BSA) in a black, opaque 96-well plate, and fluorescence polarization measured at equilibrium on a SpectraMax M5 microplate reader (Molecular Devices).

DNase I Digestion

A 216 bp oligonucleotide containing the glnA σ⁵⁴ promoter (glnAp2) was PCR amplified from pure, genomic DNA isolated from MG1655 E. coli. The resulting product was used as template in a subsequent PCR containing one 5′-6-FAM labeled primer and one unlabeled primer. The fluorescent PCR product was run on a 1% (w/v) agarose gel and gel purified. Reactions were performed as follows: 5 μL stapled peptide (C_final = 20 μM) and 35 μL 6-FAM DNA (250 ng) were incubated for 15 min at 37°C. 5 μL of DNase I (150 mU μL⁻¹) was added and reactions were incubated for 5 min at 37°C. Reactions were termina ted with 5 μL EDTA (500 mM) and vortexed immediately. The digestion products were purified via PCR Purification (Qiagen) according to the manufacturer’s protocol. The total, purified digestion products were run on a 12% polyacrylamide-TBE gel and imaged on a Typhoon Imager (GE Life Sciences). FIJI was used to quantitate remaining starting material relative to DMSO-treated samples.

Cell Viability Assays

Bacterial Cells

MG1655 E. coli cells were grown to an OD₆₀₀ of 1.0 and seeded onto a white, opaque 96-well plate in LB broth. The cells were incubated with compounds at the indicated doses for 1 h. Cell viability was assayed by addition of BacTiter-Glo^™ bioluminescence reagent according to the manufacturer’s protocol (Promega) and luminescence measured using a Spectramax M5 microplate reader (Molecular Devices). Data are normalized to vehicle-treated controls.

Human Cells

WS1 normal fibroblasts (1.5 × 10⁴ cells) were seeded onto a white, opaque 96-well plate. The cells were incubated with compounds at the indicated doses for 24 h. in Opti-MEM (Gibco). Cell viability was assayed by addition of CellTiter-Glo^™ bioluminescence reagent according to the manufacturer’s protocol (Promega) and luminescence measured using a Spectramax M5 microplate reader (Molecular Devices). Data are normalized to vehicle-treated controls.

Quantitative RT-PCR

MG1655 E. coli were grown in Gutnick minimal media to induce nitrogen starvation. At OD₆₀₀ 0.5, cells were treated with 10 μM stapled peptide. RNA was stabilized after 80 minutes after treatment using the RNAprotect Bacteria Reagent (Qiagen), and purified using the RNeasy Mini Kit and RNase-free DNase (Qiagen). Purified RNA was stored at −80°C in nuclease-free water. 0.5 μg of total RNA was used for cDNA synthesis, following manufacturer protocols (SuperScript III First-strand cDNA Synthesis System, Invitrogen). qPCR reactions were performed as follows: 1 μL cDNA, 5 μL SYBR-Select Master Mix (Applied Biosystems) and 250 nM forward and reverse primer (see Table S2 for sequences) were combined to a final volume of 10 μL. Plates were briefly centrifuged (1 min, 1,000 RPM, RT) before target amplification was performed on a C1000 Touch Real-Time PCR System (Bio-Rad) using the following conditions: 50°C (2 min), followed by 95°C (2 min), followed by 40 cycles of 95°C (15 s) and 60°C (1 min). A melting curve was obtained at the end of each experiment with the following conditions: 65°C (5 sec), followed by 60 × 5 sec cycles with increases of 0.5°C per cycle. All experiments were performed in triplicate and cysG was used as a reference gene. Threshold-cycle (C_t) values were automatically calculated for each replicate and used to determine the relative expression of the gene of interest relative to cysG for both treated and untreated samples by the 2^−ΔΔCt method (Livak and Schmittgen, 2001). p-values were calculated using one-way ANOVA analysis.

Quantification of Glutamine Synthetase Activity

MG1655 E. coli cells grown to an optical density (OD₆₀₀) of approximately 0.15 were suspended in 40 mL of M9 minimal media containing either 40 mM NH₄Cl (nitrogen-rich conditions) or 100 μM NH₄Cl (nitrogen-deficient conditions). Compounds were added to the cultures to a concentration of 10 μM and the cells were incubated in a shaker at 37°C for 4 h. The samples were then removed from the incubator and cetyltrimethylammonium bromide (CTAB) was added to a concentration of 1.5 mg mL⁻¹. Samples were incubated at 37°C for 2 min and subs equently centrifuged (3,500 RPM, 4°C) for 15 min. The supern atant was discarded, and the samples were washed with KCl (1% w/v) and centrifuged a second time. After discarding the supernatant, pellets were resuspended in 150 μL of KCl (1% w/v). An assay mixture (40 mM L-glutamine, 40 mM NH₂OH·HCl, 800 μM Na-ADP, 40 mM potassium arsenate, 100 mM mixed imidazole buffer (0.33 M each of imidazole, 2-methylimidazole, and 2,4-dimethylimidazole, pH 7.15), 600 μM MnCl₂, and 0.2 mg mL⁻¹ CTAB, pH 7.15) was added, and samples were incubated at 37°C for 30 minutes. After incubation with termination buffer (3.33% FeCl₃, 2% trichloroacetic acid, and 250 mM HCl) for 15 min, particulate matter was pelleted by centrifugation (3.500 RPM, RT) for 15 minutes. The absorbance at a wavelength of 540 nm was quantified using a Spectramax M5 plate-reader (Molecular Devices) to determine the amount of gamma glutamyl hydroxamate, indicative of active enzyme.

RNA-Seq

RNA for the RNA-Seq experiment was obtained exactly as described previously for the quantitative RT-PCR experiment. RNA libraries were prepared for sequencing using standard Illumina protocols. Sequenced reads were trimmed for adaptor sequence, masked for low-complexity or low-quality sequence, and then mapped to the E. coli K-12 genome using bowtie2 with default parameters. Reads were counted by feature count and reads-per-megabase of library size (CPM) were generated and normalized by the limmaVoom pipeline. All data are publicly available through GEO (accession # GSE111317).

Gene Set Enrichment Analysis

Gene set enrichment analysis was performed using GSEAPreranked version 6 module from GenePattern (Reich et al., 2006). The ranked list of genes based on fold change data from RNA-seq was used as the ranked list input for GSEAPreranked. A database containing 155 E. coli gene sets with fewer than 500 gene members each was obtained from gene2go (Powell, 2014). The gene set for rpoN regulon was generated by selecting all the genes with an intergenic sense rpoN consensus sequenced that were identified by ChIP-seq. The gene set for the rpoH regulon has been previously disclosed (Nonaka et al., 2006) and was used for this experiment. The rpoN and rpoH gene sets were combined to generate an alternative sigma factor gene set database. 1000 permutations of the gene set were used to assess the significance of the enrichment score for each gene set.

Hierarchical clustering analysis

Hierarchical clustering was performed using Heatmapper (Babicki et al., 2016) The distance measurement method used was Euclidean and clustering methods used was average linkage.

Quantification and Statistical Analysis

Unless otherwise specified, data are represented as the means of three, independent experiments (n=3) with error bars reflecting the standard error of the mean (SEM). p-values were determined with a 1-way ANOVA test compared to vehicle treatment (DMSO): (* = ≤ 0.05; ** = ≤ 0.005; *** = ≤ 0.0005; **** = ≤ 0.00005). All statistical analyses were performed with GraphPad Prism 7.01.

Data and Software Availability

Deep sequencing data have been deposited in the Gene Expression Omnibus (GEO) database under accession number GSE111317.

Highlights.

• Hydrocarbon crosslinks in σ⁵⁴ RpoN box-derived peptides confer increased helicity.

• Stapled σ⁵⁴ (Sσ⁵⁴) peptides bind to σ⁵⁴ promoters in vitro.

• Sσ⁵⁴ peptides enter Gram-negative cells by diffusion.

• Sσ⁵⁴ peptides block transcriptional activation of σ⁵⁴-dependent genes.