Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Aug 1.
Published in final edited form as: J Biomol Screen. 2012 May 31;17(7):946–956. doi: 10.1177/1087057112448216

High-Throughput Screening of a Diversity Collection Using Biodefense Category A and B Priority Pathogens

Esther W Barrow 1, Patricia A Clinkenbeard 1, Rebecca A Duncan-Decocq 1, Rachel F Perteet 1,2, Kimberly D Hill 1,3, Philip C Bourne 1, Michelle W Valderas 1,4, Christina R Bourne 1, Nicole L Clarkson 1, Kenneth D Clinkenbeard 1, William W Barrow 1
PMCID: PMC3700734  NIHMSID: NIHMS477948  PMID: 22653912

Abstract

One of the objectives of the National Institutes of Allergy and Infectious Diseases (NIAID) Biodefense Program is to identify or develop broad-spectrum antimicrobials for use against bioterrorism pathogens and emerging infectious agents. As a part of that program, our institution has screened the 10 000-compound MyriaScreen Diversity Collection of high-purity druglike compounds against three NIAID category A and one category B priority pathogens in an effort to identify potential compound classes for further drug development. The effective use of a Clinical and Laboratory Standards Institute–based high-throughput screening (HTS) 96-well–based format allowed for the identification of 49 compounds that had in vitro activity against all four pathogens with minimum inhibitory concentration values of ≤16 μg/mL. Adaptation of the HTS process was necessary to conduct the work in higher-level containment, in this case, biosafety level 3. Examination of chemical scaffolds shared by some of the 49 compounds and assessment of available chemical databases indicates that several may represent broad-spectrum antimicrobials whose activity is based on novel mechanisms of action.

Keywords: anti-infective drugs, automation, cell-based assays, compound repositories, high-content screening

Introduction

As summarized in a recent report, emerging infectious diseases (EID) represent an important and growing burden to global economies and public health; their emergence is likely to be driven by factors associated with social, environmental, and ecological factors.1 In their report, the authors reveal various interesting factors about global patterns of EIDs between 1940 and 2004.1 In general, EID events have risen significantly, with 54.3% of those events being caused by bacteria, and a large proportion of those are drug resistant.1 In addition, they observed that most EID events are caused by zoonotic pathogens, and 71.8% of those were caused by pathogens of wildlife origin.1

Biological terrorism is also a significant consideration with regard to global economies and public health, as demonstrated by the 2001 anthrax letters. Various studies have predicted that the economic impact of a biological attack with such agents as Bacillus anthracis, Francisella tularensis, and Brucella abortus could be in the billion dollar range, with human deaths approaching millions, depending on the amount of agent used and the population exposed to that agent.2 Antimicrobial drug resistance is a significant intrinsic health issue that continues to increase in importance as more emerging strains of bacteria threaten human health.3–5 The World Health Organization considers antimicrobial resistance as one of the three greatest threats to human health. As such, they declared antimicrobial resistance the focus of World Health Day 2011 (www.who.int/world-health-day/2011). Equally important is the limited pipeline of new antimicrobial candidates that is necessary to combat this global dilemma.6,7 Increasingly, pharmaceutical companies are showing less interest in antibacterial discovery and development; this makes society even more vulnerable to emerging infectious diseases.8 Because of these threats, the National Institutes of Allergy and Infectious Diseases (NIAID) is using the entire spectrum of biodefense resources to partner with other federal agencies and industry to develop a pipeline that “runs from bench to bedside” to combat these dread diseases.9

The purpose of this study was to address some of these major health issues by developing and using high-throughput screening (HTS) capabilities for screening compounds in upper-level biocontainment against biosafety level 3 (BSL-3) category A and B agents. Several of the methods had to be adapted because of certain constraints placed on laboratory personnel working in BSL-3 biocontainment facilities. In addition, electronic database management tools had to be developed to capture and assess resulting HTS data that were obtained under conditions requiring biosafety, biosecurity, and confidentiality constraints. These efforts were funded through NIAID contract HHSN266200400004I/N01-AI-40004 awarded to our institution from 2003 to 2011; the screening reported here took place from 2006 to 2010. It is our intent to assist NIAID in its research agenda of developing new and/or improved antimicrobials to combat emerging pathogens, potential biological threat agents, and drug-resistant bacterial strains. It is also our intent to share this information with the scientific community. The Center for Veterinary Health Sciences at Oklahoma State University is a veterinary school with more than half a century experience working with zoonotic diseases. This expertise will greatly improve our ability to address these issues, particularly because the four select agents that were evaluated in this study represent four important zoonotic pathogens, causing such diseases as anthrax, plague, tularemia, and brucellosis.

Materials and Methods

Bacterial Strains

The following bacterial strains were used in the HTS process. Bacillus anthracis Ames (NR-411), Yersinia pestis Colorado 92 (NR-641), and Francisella tularensis Schu S4 (NR-643) were obtained from the Biodefense and Emerging Infectious Research Resources Repository (BEI Resources, Manassas, VA), NIAID, National Institutes of Health (NIH). Brucella abortus 2308 was obtained from the U.S. Department of Agriculture National Animal Disease Center. Quality control strains Staphylococcus aureus 29213 and Escherichia coli 25922 were obtained from ATCC (Manassas, VA).

Quality Control for Bacterial Strains

Test strains were replaced annually with fresh isolates, and the old and new strains were evaluated phenotypically using the Biolog Gen II Identification System (Biolog, Hayward, CA) and genotypically by extraction of DNA and profiling using polymerase chain reaction (PCR) to identify a specific species as well as markers that differentiate individual strains or loci whose products are known to participate in virulence.10–14 DNA was extracted with the Promega Maxwell 16 automated genomic DNA extraction and purification kits and instrument (Promega, Madison, WI). Real-time PCR analysis was conducted with SYBR green using a Cepheid Smart Cycler II system (Cepheid, Sunnyvale, CA).

Microdilution Broth Screening Assay

A 96-well format broth microdilution assay,15 developed according to Clinical and Laboratory Standards Institute (CLSI) guidelines,16 was used to evaluate initial inhibition by each compound and subsequently to determine the minimal inhibitory concentration (MIC) for each agent that inhibited at the initial screening concentration of 16 μg/mL. Additional plates with appropriate quality control (QC) test strains and QC antimicrobials were included with each assay to verify appropriate CLSI MIC ranges.17 The assay uses AlamarBlue dye reduction (Invitrogen, Life Technologies, Grand Island, NY) to assess the MIC for all of the agents except Francisella tularensis, for which a noncolorimetric assay is required because of color interference with the media supplement. Drug plates were prepared with a Biomek 2000 Laboratory Automation Workstation (Beckman-Coulter, Brea, CA) to ensure reproducibility and accuracy. The Biomek 2000 used for drug plate preparation was housed in a Baker biosafety cabinet (BSC) to guarantee sterility. Inoculation of drug plates with bacteria was carried out in a BSL-3 facility using another Biomek 2000 housed in a Baker BioPROtect II BSC. Both of the Baker BSCs were specially designed to house Biomek 2000 units. Personnel conducting the screening wore approved personal protection equipment (PPE), which consisted of KleenGuard A60 protective coveralls with boots (Kimberly-Clark, Roswell, GA), double gloves, and a powered air purifying respirator with HEPA filters (3M, St. Paul, MN).

Initial screening of all 10 000 compounds against three of the four pathogens was conducted at a concentration of 16 μg/mL; each compound was assayed once in duplicate. Total final volume of drug and medium was 100 μL per well. Sterility, color, and growth control wells were included in each 96-well plate. After screening the first 480 compounds, it was noted that the hit rate for F. tularensis was disproportionate to the other pathogens (i.e., 45% versus 1%–5%). As a result, the initial screening concentration for F. tularensis was changed from 16 to 8 μg/mL. At that time, all 480 were repeated at the 8 μg/mL concentration, and those results are reported here. Minimal inhibitory concentration (MIC) determinations were then performed on all compounds showing inhibition at the initial screening concentration using a twofold dilution series ranging from 0.0625 to 16 μg/mL. MIC determinations were performed twice in duplicate (n = 4). Commercial antimicrobials for QC plates were obtained from Sigma-Aldrich (St. Louis, MO), Serologicals Proteins, Inc. (Kankakee, IL), and Mediatech (Manassas, VA).

Test compounds and commercial antimicrobials were prepared at the appropriate concentrations in cation-adjusted Mueller-Hinton broth (CAMHB; Becton Dickinson, Cockeysville, MD) and then aliquoted into 96-well plates for Bacillus anthracis and Yersinia pestis. AlamarBlue (Invitrogen; 10% vol/vol) was added to wells prior to inoculation. F. tularensis requires the addition of 2% Isovitalex (Becton, Dickinson and Company, Franklin Lake, NJ) to CAMHB. Because this causes interference with AlamarBlue (Invitrogen), the dye was not used in assays with this organism. As recommended by CLSI, Brucella broth was used with the Brucella abortus strain. Because AlamarBlue (Invitrogen) is stable for only 24-h incubation and B. abortus requires 48-h incubation, AlamarBlue (Invitrogen) was added 24 h after infection for that organism.

Inoculations with select agents were carried out in a BSL-3 facility. Appropriate wells were inoculated with ≈ 5 × 106/mL colony forming units (CFU), which resulted from a 1:20 dilution of an inoculum equivalent to 0.5 McFarland units (≈ 1–1.5 × 108 CFU). This was confirmed by plating an appropriate dilution onto solid medium. QC drugs and strains used for each organism were as follows: B. anthracis (penicillin PEN, doxycycline DX, ciprofloxacin CIP), Staphylococcus aureus ATCC 29213 and Escherichia coli ATCC 25922; F. tularensis (gentamicin GEN, CIP, DX), S. aureus ATCC 29213 and E. coli ATCC 25922; Y. pestis (GEN, DX, chloramphenicol CHL), E. coli ATCC 25922; B. abortus (CIP, tetracycline TET, rifampin RIF), S. aureus ATCC 29213 and E. coli ATCC 25922. Plates were read spectrophotometrically in an AD340a Beckman plate reader (Beckman-Coulter) programmed to subtract the absorbance at 600 nm from that at 570 nm; the MICs are reported as the lowest drug concentrations yielding a differential absorbance of zero or less.15,18 In the case of F. tularensis, absorbance was read at 600 nm, and the MIC was reported as the lowest drug concentration that inhibited visible growth.

MIC Evaluation with Commercial Antimicrobials

The four select agents were tested against 30 commercially available antimicrobials to obtain MIC values for each. As with the diversity collection, the MIC evaluations of each antimicrobial were performed twice in duplicate (n = 4). Likewise, each set of evaluations included a plate containing QC drugs and bacterial strains.

Diversity Collection

The MyriaScreen Diversity Collection was purchased from Sigma-Aldrich (St. Louis, MO). The collection was produced by collaboration between TimTec, Inc. (Newark, DE) and Sigma-Aldrich. Their compounds were filtered with TimTec’s proprietary software to obtain a pool of more than 300 000 TimTec and Sigma-Aldrich compounds on the basis of diversity. Additional filters were set to consider molecular weight (120–500), cLogP, H-acceptor, H-donors, and rotatable bonds (Sigma-Aldrich). The selection was refined by removing compounds that were overly represented or not well suited for medicinal chemistry follow-up (Sigma-Aldrich). The MyriaScreen “is composed of 10 000 high-purity screening compounds hand-picked to maximize chemical diversity while maintaining drug-likeness” (Sigma-Aldrich). The collection is set up to be interactive with HTS (e.g., shipped in 96-well plates containing individual tubes ready for manipulation with systems such as the Biomek 2000), and the compounds are dissolved in DMSO. For screening purposes, it was necessary to determine the effect of DMSO on the viability of each bacterial pathogen to ensure that inhibitory observations were due to the compound and not the DMSO. This consideration was one of the issues that went into the decision for the initial screening concentrations used. In preliminary experiments, we determined that for B. anthracis and B. abortus, the DMSO began to affect viability at 2% to 4% (vol/vol). For F. tularensis and Y. pestis, the DMSO began to affect viability at >1%. As a result, assays were run at a DMSO concentration of 0.8% for all organisms except F. tularensis, in which case 0.4% DMSO was used. Prior to preparation of drug-screening plates, each MyriaScreen compound was assigned an individual OSU number to maintain a blind study in which personnel were not aware of which compounds were being assayed at any one time. Drug plates were prepared for initial screening then stored at −80 °C until being transferred to the BSL-3 facility for inoculation.

Data Management

The assays that we have described in this article, both the colorimetric for B. anthracis, Y. pestis, and B. abortus and turbidimetric for F. tularensis, function by determining whether or not there is growth in a well. Any growth in a test compound (TC) or growth control (GC) well is considered a positive result; growth in a sterility control (SC) well indicates that there is potentially contamination, and the assay should be repeated. Growth is defined for the colorimetric assay as an A570–A600 value of >0. For the turbidimetric assay, a value greater than a cutoff value was considered growth. The cutoff value was determined as GCmean – ((GCmean – SCmean) × 70%), where GCmean = the mean absorbance (A600) value of the GC wells and SCmean = the mean absorbance (A600) value of SC wells. As described in the text, all plates contained positive and negative control wells to ensure the validity of results in test wells.

Data were copied directly from a plate reader into a Microsoft (MS; Redmond, WA) Excel spreadsheet. The spreadsheets were designed to automatically determine initial hits or calculate MIC values depending on the type of experiment. The spreadsheets would also check for skipped wells and assess the control wells on the plate. If there was a discrepancy in the calculated values, raw data were examined for anomalies. In cases in which questionable results remained, the assay was repeated. Resulting data were transferred to an MS Access database for compilation and reporting.

MDL ISIS/Base 2.55 SP4 (Symyx Technologies, Inc., Sunnyvale, CA) was used to manage and categorize the chemical frameworks. ChemBioDraw Ultra 12.02.1076 (Cambridge, MA) was used for assigning chemical scaffolding (frameworks) and obtaining chemical names and property information. SciFinder (CAS, American Chemical Society, Columbus, OH) was used to search for pertinent information regarding the compounds identified in the HTS process. All data were collected, stored, and maintained on a secure access hard drive accessible by authorized personnel from offices and biocontainment laboratories involved in the study.

Results

Adaption of HTS Procedures for BSL-3 Containment

Because of the safety and security nature of a number of biodefense and emerging infectious disease category A, B, and C priority pathogens, these agents can be studied only in BSL-3 containment facilities. As a result, automation of drug screening for these types of pathogens had to be developed in such a way as to overcome issues such as limited specialized space with appropriate physical barriers, higher maintenance and operation costs, regulatory issues, rules, permits necessary to conduct such work, and background checks for individuals working with or around select agents as well as personnel involved with critical equipment maintenance and IT services. Other issues include the need for special PPE, containment of equipment, annual required maintenance and training, security monitoring devices for limited-access laboratories, video surveillance, key card access, and security-capable computer systems and authorized IT personnel.

Precautions had to be implemented to ensure safety during inoculation and analysis of the assays performed in bio-containment. After 96-well plates had been inoculated with the pathogen, the sterile plastic lids were replaced, and the plates were placed in gas-permeable plastic bags with metal closures and incubated at appropriate temperatures. Following incubation, plates were transferred from the incubator to the BSC, where the lids were removed and a sterile seal was securely applied prior to spectrophotometric assessment in a plate reader.

The assays developed here have worked very well with issues involving BSL-3 containment operations. For learning purposes, personnel were initially trained on a Biomek 2000 housed in a BSC within a BSL-1 laboratory. Because no infectious agents are used, personnel can be trained on the instrument and learn robotic programming techniques without the need for PPE. That training can then be continued on the Biomek 2000 (Beckman) housed in the BSL-3 facility, where the inoculations and plate readings take place. This facilitates the training process by reducing the stress and difficulty of having to learn the procedure initially in a BSL-3 setting while suited up in PPE.

Real-Time PCR of Bacterial Strains Used in the Screening Process

As a part of the screening program, all of the bacterial strains used in the process were replaced with freshly propagated stocks annually. At that time, the DNA was extracted from both the old and new stocks and genetically profiled by real-time PCR by amplification of fragments of targets based on (1) specificity to a given bacterial species, (2) involvement in virulence, and (3) involvement in drug resistance as necessary. The amplified target fragments are listed in Table 1. An example of data resulting from the real-time PCR genotyping assay is shown in Figure 1 using B. anthracis cultures. These procedures were necessary to ensure genetic purity of each strain throughout the screening process. Use of the Promega Maxwell 16 (Promega) proved to be ideal for DNA extraction because it is small enough to be housed within a BSC and can be used to extract up to 16 samples at one time. All extracted DNA samples had to be streaked on appropriate medium and incubated for 48 h to ensure that no viable bacteria were present prior to the disinfected vials being transferred to a BSL-1 biocontainment facility for PCR evaluations.

Table 1.

Amplified Targets for Select Agents

Bacillus anthracis Ames capB2P Capsule B
anhC Unique identifying sequence designed in-house (putative anhydrolase)
LEFP Lethal factor
PAP Protective antigen
Brucella abortus 2308 omp2aC Outer membrane protein 2a
IS711C Insertion sequence, specific to Brucella species
alkBC Insertion sequence, specific to abortus subspecies
Francisella tularensis Schu S4 tul4C Tularemia reactive antigen
fopAC Outer membrane protein
Yersinia pestis CO92 plaP Plasminogen activator
caf1P F1 antigen
yopTP Type III secretion system protein
16S rRNAC Specific to Yersinia species
pgmC Locus containing select agent specific high pathogenicity island (HPI) as well as potential virulence factors
Open in a new tab
P

= plasmid-borne gene;

C

= chromosomal-borne gene.

Figure 1.

Figure 1

Open in a new tab

Representative real-time PCR genotyping assay melting curves from a fresh culture (A) used to replace a retired culture (B). DNA was extracted from the bacterial pathogen, in this case Bacillus anthracis Ames, and genetically profiled using real-time PCR by examining the following targets for amplification, capB2P (capsule B), anhC (unique identifying sequence designed in house; putative anhydrolase), LEFP (lethal factor), PAP (protective antigen). “C” and “P” indicate chromosomal or plasmid-borne genes. Amplified products were screened for their melting temperature to discern specificity of the amplification reaction for the desired sequence.

MIC Evaluations of Commercial Antimicrobials

The four select agents were screened against a panel of 30 commercially available antimicrobials that represent all the major classes of mechanism of action (MOA), including cell wall synthesis, protein synthesis (30S and 50S), protein synthesis (70S complex), RNA synthesis, DNA synthesis, folate synthesis, and cell membrane. The MIC values resulting from four assays are given in Table 2 for 16 of those compounds. The results for 14 other commercially available antimicrobials are included in Supplemental Table S2.

Table 2.

MIC Evaluations (μg/mL) of Commercial Drugs against Select Agents (n ≥ 4)

No Drug Structural Classes/MOA B.a. F.t. Y.p. Br.a.
1 Penicillin Penicillins/cell wall ≤0.12 >16 >16 16
2 Oxacillin Penicillins/cell wall 1–4 >16 >16 ≥16
3 Aztreonam Monolactam/cell wall >16 >16 >16 >16
4 Cephalothin Cephalosporins/cell wall 1–2 >16 >16 >16
5 Cefotaxime Cephalosporins/cell wall ≥16 >16 >16 16
6 Vancomycin Glycopeptide/cell wall 2 ≥8 >8 >8
7 Streptomycin Aminoglycosides/protein/30S 8 2 4 ≥8
8 Gentamicin Aminoglycosides/protein/30S 0.25–0.5 0.03–0.12 1 1–2
9 Doxycycline Tetracyclines/protein/30S 0.25 0.12–0.5 1–2 0.25
10 Tetracycline Tetracyclines/protein/30S 0.25 1–2 4 0.25–0.5
11 Azithromycin Macrolides/protein/50S 1 0.5–2 8 ≥8
12 Erythromycin Macrolides/protein/50S 0.5–1 1–2 >8 16
13 Chloramphenicol Chloramphenicol/protein/50S 8 4 2–4 1–2
14 Linezolide Oxazolidinone/protein/70S complex 4–8 8–16 >16 1–2
15 Rifampin Rifamycins/RNAase polymerase 0.06 0.5 4 0.5–2
16 Ciprofloxacin Fluoroquinolones/DNA gyrase 0.5–1 0.01–0.03 ≤0.12–0.25 0.25–1
Open in a new tab

B.a. = B. anthracis Ames; Y.p. = Y. pestis Co92; F.t. = F. tularensis SchuS4; Br.a. = B. abortus 2308.

Screening and MIC Evaluations of MyriaScreen against Select Agent Panel

Of the 10 000 compounds in the MyriaScreen, 2045 inhibited one or more of the four select agent bacterial pathogens at the initial screening concentration (Supplemental Data). The individual breakdown for those compounds that showed inhibition at ≤16 μg/mL were 547 (5.47%) for B. anthracis, 1714 (17.14% at ≤8 μg/mL) for F. tularensis, 125 (1.25%) for Y. pestis, and 406 (4.06%) for B. abortus. Forty-nine (0.49%) of the compounds inhibited all four of the select agent pathogens at ≤16 μg/mL (Table 3; Supplemental Table S3). Each is listed with its respective MyriaScreen compound ID and plate number (Sigma-Aldrich), along with chemical name and cLogP value, obtained with ChemBioDraw Ultra 12.0.2.1076 (CambridgeSoft, Cambridge, MA; Tables 3 and S3).

Table 3.

Representative MyriaScreen Compounds with MIC Values for All Four Select Agents

MIC (μg/mL)
ID Compound MW cLogP B.a. Y.p. F.t. Br.a.
L148016 6E06 2-(4-(5-nitrofuran-2-yl)-4H-1,2,4-triazol-3-yl)-1H-benzo[d]imidazole 296.24 1.837 0.25–0.5 >16 0.25 0.5
L148598 7E02 3-(3-(5-nitrofuran-2-yl)-1H-pyrazol-5-yl)pyridine 256.06 2.298 8 4 2–4 4–8
L161845 8C06 1-(2-cyclohexylethyl)-4-(3,4-dichlorobenzyl)-1H-tetrazol-5(4H)-imine 354.28 5.618 8 >16 8 16
L201014 15E04 3-(4-chlorophenylthio)-1-(furan-2-yl)propan-1-one 266.02 3.933 8–16 >16 4 4–8
L232238 19D03 1-(4-chlorophenyl)-3-(2,4-dimethoxyphenylamino)propan-1-one 319.1 4.005 >16 >16 2–4 4–8
L322776 24F05 4-bromo-2-((2,4,5-trifluorophenylamino)methyl) phenol 332.12 4.336 8 >16 2 8
R536628 70F06 4-chloro-N-(3,5-dichlorophenyl) benzenesulfonamide 336.62 5.114 4 >16 4 8
ST005324 10H11 N-(3-carbamoyl-4,5-dimethylthiophen-2-yl)furan-2-carboxamide 264.3 1.386 >16 8–16 >16 4
ST016021 37F11 N-(2,4-dimethylphenyl)-4-methyl-5-phenyl-2-(2,2,2-trifluoroacetamido)thiophene-3-carboxamide 432.11 5.546 0.5 4 1 2
Open in a new tab

Highlighted compounds R536628, ST005324, and ST016021 are three compounds discussed further in Tables 4 and 5. MyriaScreen compounds are listed as ID numbers; these include ID, plate, and well number. MW = molecular weight; B.a. = B. anthracis Ames; Y.p. = Y. pestis Co92; F.t. = F. tularensis SchuS4; Br.a. = B. abortus 2308.

The plate format for the initial screening and subsequent MIC evaluations was designed to accommodate as many compounds as possible, including growth control, solvent control, sterility control, and test compound color control (including test compound and medium only). For initial screening, the plate design accommodated 24 test compounds screened in duplicate, along with a test compound color control for each (Fig. 2A). Each evaluation was performed once in duplicate. In the case of the plate for MIC evaluations, each plate had four test compounds assayed in two sets of twofold dilutions ranging in concentration from 0.0625 to 16 μg/mL (Fig. 2B). Each MIC evaluation was performed twice in duplicate for a total of n = 4; each evaluation used a different inoculum preparation.

Figure 2.

Figure 2

Open in a new tab

(A) Plate format for initial screening at 16 μg/mL. (B) Plate format for minimum inhibitory concentration determination that will accommodate four test compounds. Dilutions are twofold, ranging in concentration from 0.0625 to 16 μg/mL. Appropriate control wells are shown, including ones for color control.

Data Mining for Potential MOA

MDL ISIS/Base 2.55 SP4 (Symyx Technologies) and ChemBioDraw Ultra 12.0.2.1076 (CambridgeSoft) software were used to analyze the accumulated data files for the compounds that inhibited all four select agent pathogens. In particular, common scaffolds (chemical frameworks) were examined based on each inhibitor. Further analyses revealed that some of the compounds that inhibited all four pathogens did so at varying levels of inhibition. As examples, two clusters will be discussed here to elucidate our methods.

A group of thiophen/carboxamide compounds was extrapolated from the database for further examination based on two compounds from the 49 that inhibited all four pathogens, compounds ST005324 (plate 10 well H1) and ST016021(plate 37 well F11). A substructure search of the MyriaScreen SD file with ISIS Base using the thiophen/carboxamide scaffold (Table 4) resulted in the detection of four other derivatives that shared this common scaffold and that had varying degrees of inhibition for some but not all of the four select agents (Table 4). In all, 132 compounds were identified in the database with the thiophen/carboxamide scaffold substructure search. Compounds ST005226 (plate 10 well D10), ST006747 (plate 16 well A11), and ST015518 (plate 37 well F08) inhibited three of the select agents, and compound ST026452 (plate 60 well E09) inhibited only two of the select agents (Table 4). This variation in inhibitory capacity suggests that the bacterial MOA for these compounds might be a specific target rather than a general mechanism and that they would be amenable to further chemical derivatization to improve inhibitory properties. A search with SciFinder (CAS Division of the American Chemical Society, Chemical Abstracts Service, Columbus, OH) revealed information that indicates this group of compounds represents a class of reverse transcriptase–associated ribonuclease H1 activity inhibitors.19,20

Table 4.

Thiophene/Carboxamide Inhibitors from MyriaScreen Library

MIC (μg/mL)
ID Chemical Structure Compound MW cLogP B.a. Y.p. F.t. Br.a.
graphic file with name nihms477948t1.jpg
ST005324 10H11 graphic file with name nihms477948t2.jpg N-(3-carbamoyl-4,5-dimethylthiophen-2-yl)furan-2-carboxamide 264.30 1.386 ≥16 8 – 16 ≥16 4–8
ST005226 10D10 graphic file with name nihms477948t3.jpg 5-bromo-N-(3-cyano-4,5-dimethylthiophen-2-yl)furan-2-carboxamide 325.18 2.876 NA ≥16 8 ≥16
ST00674716A11 graphic file with name nihms477948t4.jpg 4,5-dimethyl-2-(2-(3-nitro-1H-1,2,4-triazol-1-yl)acetamido) thiophene-3-carboxamide 324.32 −0.487 8–16 8 NA 8
ST01602137F11 graphic file with name nihms477948t5.jpg N-(2,4-dimethylphenyl)-4-methyl-5-phenyl-2-(2,2,2-trifluoroacetamido) thiophene-3-carboxamide 432.11 5.546 0.5 1 4 2
ST015518 37F08 graphic file with name nihms477948t6.jpg 2-amino-N-(4-methoxyphenyl)-4-methyl-5-phenylthiophene-3-carboxamide 338.11 4.156 NA 16 4 4 – 8
ST026452 60E09 graphic file with name nihms477948t7.jpg (E)-ethyl 4-((3-carbamoyl-4,5-dimethylthiophen-2-yl)amino)-4-oxobut-2-enoate 296.34 1.601 NA 16 NA 16
Open in a new tab

NA = not active at ≤16 μg/mL. The two highlighted in bold are from Table 3. MyriaScreen compounds are listed as ID number, which includes ID, plate, and well number. MW = molecular weight; B.a. = B. anthracis Ames; Y.p. = Y. pestis Co92; F.t. = F. tularensis SchuS4; Br.a. = B. abortus 2308.

Using this information, we have examined the potential for these compounds to inhibit this target by cloning and expressing the recombinant RNaseH1 proteins from each of the four bacterial strains tested. Initial studies using a previously reported assay21 with the two most potent thiophene/carboxamides (ST016021 and ST005324; Table 4) revealed marginal inhibition at 10 μM, which increased to almost complete inhibition at 100 μM (data not shown). These studies are ongoing.

The second cluster of inhibitors to be selected was a group of benzenesulfonamides represented by compound R536628 (plate 70 well F06). A substructure search using the benzenesulfonamide scaffold revealed a series of structurally similar inhibitors that showed varying degrees of inhibition to the four select agents (Table 5). In all, 726 compounds were identified in the MyriaScreen database that contained the benzenesulfonamide scaffold. Six of those were selected based on their structural similarity and ability to inhibit one or more of the select agents (Table 5). Although R536628 was able to inhibit all four select agents, five were not inhibitory for Y. pestis and showed only limited inhibition to B. abortus (Table 5). Again, this variation in inhibitory capabilities within this class of structurally similar compounds suggests that the MOA for these compounds might be a bacterial target amenable to drug development. A search using SciFinder resulted in bibliographic information regarding benzenesulfonamides showing activity against methicillin-resistant Staphylococcus aureus and vancomycin-resistant Entercoccus (VRE).22 The sulfonamide scaffold at first appearance suggests that they are acting similar to other sulfonamides that have inhibitory properties toward a key enzyme in the folate pathway, dihydropteroate synthase (DHPS).23 However, this particular collection lacks the amino group on the benzenesulfonyl moiety that is found on the sulfonamides that inhibit DHPS (e.g., sulfamethoxazole, sulfanilamide, and sulfadiazine). In an attempt to determine an MOA for these benzenesulfonamides, we examined their activity against recombinant DHPS from B. anthracis, an enzyme previously studied by our group.24,25 We were not able to show any inhibitory activity for this enzyme, despite inhibition of the organism, and thus concluded that DHPS inhibition is not likely the MOA for these compounds (data not shown). We have also assessed the phenotypic effect of one of the benzene sulfonamides using the phenotype microarray system (Biolog), which provided an inconclusive MOA and indicated it is not likely acting against DHPS.26

Table 5.

Benzenesulfonamide Inhibitors from MyriaScreen Library

MIC (μg/mL)
ID Chemical Structure Compound MW cLogP B. a. Y. p. F. t. Br.a.
graphic file with name nihms477948t8.jpg
R536628 70F06 graphic file with name nihms477948t9.jpg 4-chloro-N-(3,5-dichlorophenyl) benzenesulfonamide 336.62 5.114 4 ≥16 4 8
R577871 79F02 graphic file with name nihms477948t10.jpg N-(4-sec-butylphenyl)-4-methoxybenzenesulfonamide 319.43 4.814 16 NA 4 NA
R577898 79G02 graphic file with name nihms477948t11.jpg N-(3,4-dichlorophenyl)-4-methoxybenzenesulfonamide 332.21 4.267 16 NA 8 ≥16
R536601 70E06 graphic file with name nihms477948t12.jpg 4-chloro-N-(3-chloro-4-methylphenyl)benzenesulfonamide 316.21 4.876 8 NA 4 16
R536512 70H05 graphic file with name nihms477948t13.jpg 4-chloro-N-(3-chloro-4-fluorophenyl)benzenesulfonamide 320.17 4.544 16 NA 8 16
ST031828 82G08 graphic file with name nihms477948t14.jpg 4-bromo-N-(4-(sec-butyl)phenyl) benzenesulfonamide 368.29 5.70 4 NA 4 NA
Open in a new tab

NA = not active at ≤16 μg/mL. The one highlighted in bold is from Table 3. MyriaScreen compounds are listed as ID number, which includes ID, plate, and well number. MW = molecular weight; B.a. = B. anthracis Ames; Y.p. = Y. pestis Co92; F.t. = F. tularensis SchuS4; Br.a. = B. abortus 2308.

The entire screening results for all the MyriaScreen compounds are included as Supplemental Data.

Discussion

There are generally two major approaches used to screen diversity collections such as the MyriaScreen collection for lead antimicrobials. One involves the use of individual “targets” (e.g., enzymes), and the other involves the use of whole cells (e.g., bacteria). Toward the end of the 1980s through the 1990s, HTS efforts were generally focused on faster screening of small-molecule libraries using genomics and combinatorial chemistry.27 A large number of those projects used HTS that involved the use of individual targets such as receptors, kinases, and other enzymes.27 Limited success was reported for more than 60 different targets from about 34 companies using target-based HTS.28,29 For discussion purposes, we will examine the results from one of those companies.

After the first complete bacterial genome was reported in 1995, several companies, including GlaxoSmithKline, moved back into the area of antibacterials.30 In a review published in 2007, they reported their results regarding the evaluation of 300 genes and 70 HTS campaigns over a 7-year period from 1995 to 2001 using the company’s collection, which consisted of 260 000 to 530 000 compounds.30 Essentially, of 67 HTS campaigns involving numerous and varied bacterial targets and 3 HTS campaigns involving whole-cell assays using E. coli and S. aureus, only 16 HTS campaigns gave rise to hits, and only 5 of those resulted in leads.30 One of the problems observed in the process was that some of the hits obtained with target screening were later found to be nonspecifically toxic to both mammalian and bacterial cells, usually because of indiscriminate disruption of the cell membrane.30 Other leads lacked broad-spectrum activity (i.e., limited activity against a suitable range of Gram-positive and Gram-negative agents).30 One of their conclusions was that “whole-cell assays are favoured for finding a lead compound that has a modicum of antibacterial activity, but biochemical assays and genetic studies are vital to determine the MOA of these leads.”30

Our HTS was designed with whole-cell assays. The reasoning was in part based on results from HTS studies such as those reported by GlaxoSmithKline (i.e., initial hits could be obtained knowing that they were able to penetrate the whole cell and cause inhibition). Our drug-screening panel used one strain that is Gram-positive and three that are Gram-negative, so that our results could also demonstrate broad-spectrum activity. Further studies will be necessary to determine MOAs so that continued drug development will be possible. Our initial screening concentration of 16 μg/mL was based on two critical parameters. First, this concentration helped to eliminate the problems of DMSO toxicity observed at higher concentrations with some of the pathogens (discussed in the Materials and Methods section). Second, we wanted to use an initial concentration that was not too restrictive and allowed for identification of potential groups of inhibitors that showed a range of activity, including no activity. We felt that this would eventually allow us to identify compound series that might be more amenable to further drug development once the MOA was determined and molecular modeling and medicinal chemistry could come into play.

An examination of the 49 MyriaScreen compounds that inhibited all four of the select agent pathogens using SciFinder revealed some interesting compound classes such as the two discussed here, the thiophen/carboxamide and benzenesulfonamide groups. Although the precise MOA is not known at this time, we can speculate that the MOA for the thiophen/carboxamide inhibitors may be a bacterial RNase H, and the series could possibly be amenable to further drug development. In the case of the benzenesulfonamide inhibitors, it is unlikely that the MOA is DHPS, but the MOA may be another enzyme in the folate pathway. Further studies are necessary to make these determinations. Although compounds were identified that inhibited only one or more of the select agents (see Supplemental Data), we focused on those that inhibited all four because of NIAID’s interest in developing broad-spectrum inhibitors.

Of the 49 inhibitors identified, several are probably non-specific inhibitors (e.g., biocides) based on information derived from SciFinder. These would not be useful for further specific drug development. Similar searches using SciFinder for a number of the 49 inhibitors did not produce any information; therefore, we have no leads regarding their MOA at this time. Perhaps the use of more extensive library searching services could produce additional information for these compounds. However, the NIH contract budget did not allow for the licensing cost of those databases. Some of the inhibitors identified in our screening were found to represent known antimicrobials such as R462438 (plate 63, well F03; quinolones) and ST044514 (plate 12, well H11, quinolones) (Table S3), as well as ST024737 (plate 53, well B07, trimethoprim), ST024740 (plate 53, well D07, sulfamethoxazole), ST007538 (plate 18, well G10, sulfamethoxydiazine), and ST007537 (plate 18, well F10, sulfamethoxypyridazine) (Supplemental Data). This information was not known until after the screening results were completed and chemical names and SciFinder information were being assessed. The complete set of results for the HTS of the 10 000 compounds is presented as Supplemental Data. This is done so that other researchers can make use of this information by helping them to identify possible groupings of compounds that share common scaffolding. They may also be able to identify compounds similar to ones they have in their own collections and use this to make a decision on pursuing their own compounds for further drug development. It is advised that one should obtain a copy of the MyriaScreen SD file from Sigma-Aldrich to make full use of this information by using data management software such as MDL ISIS/Base. The link for the Sigma-Aldrich MyriaScreen SD file is http://www.sigmaaldrich.com/chemistry/drug-discovery/screening-compounds.html.

Supplementary Material

SUppData

Table S2. MIC evaluations (μg/mL) of additional Commercial Drugs against Select Agents. B.a. = B. anthracis Ames, Y.p. = Y. pestis Co92, F.t. = F. tularensis SchuS4, Br.a. = B. abortus 2308. (n 4).

Table S3. Additional MyriaScreen compounds with MIC values for compounds that inhibited all four Select Agents MyriaScreen compounds are listed as ID number which includes ID, plate and well number. MW = molecular weight, B.a. = B. anthracis Ames, Y.p. = Y. pestis Co92, F.t. = F. tularensis SchuS4, Br.a. = B. abortus 2308.

Acknowledgments

We would like to thank NIH/NIAID/DMID for the funding to conduct these services. We would like to thank our NIAID program officer, Kristin DeBord, PhD, and section chief, Judith Hewitt, PhD, in the Biodefense Research Resources Section, for their guidance and support during this project.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by NIH/NIAID/DMID contract HHSN2662004 00004I/N01-AI-40004.

Footnotes

Supplementary material for this article is available on the Journal of Biomolecular Screening Web site at http://jbx.sagepub.com/supplemental.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

  • 1.Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, Daszak P. Global Trends in Emerging Infectious Diseases. Nature. 2008;451:990–993. doi: 10.1038/nature06536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Kaufman AF, Meltzer MI, Schmid GP. The Economic Impact of a Bioterrorist Attack: Are Prevention and Postattack Intervention Programs Justifiable? Emerg Infect Dis. 1996;3(2):83–93. doi: 10.3201/eid0302.970201. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Fauci AS, Touchette NA, Folkers GK. Emerging Infectious Diseases: A 10-Year Perspective from the National Institute of Allergy and Infectious Diseases. Emerg Infect Dis. 2005;11:519–525. doi: 10.3201/eid1104.041167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Klee SR, Ozel M, Appel B, Boesch C, Ellerbrok H, Jacob D, Holland G, Leendertz FH, Pauli G, Grunow R, Nattermann H. Characterization of Bacillus anthracis-like Bacteria Isolated from Wild Great Apes from Cote d’Ivoire and Cameroon. J Bacteriol. 2006;188:5333–5344. doi: 10.1128/JB.00303-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Rambhia KJ, Gronvall GK. Antibiotic Resistance. Biosecur Bioterror. 2009;7:371–377. doi: 10.1089/bsp.2009.1105. [DOI] [PubMed] [Google Scholar]
  • 6.Boucher HW, Talbot GH, Bradley JS, Edwards JE, Gilbert D, Rice LB, Scheld M, Spellberg B, Bartlett J. Bad Bugs, No Drugs: No ESKAPE! An Update from the Infectious Diseases Society of America. Clin Infect Dis. 2009;48:1–12. doi: 10.1086/595011. [DOI] [PubMed] [Google Scholar]
  • 7.Talbot GH. What Is in the Pipeline for Gram-Negative Pathogens? Expert Rev Anti−infect Ther. 2008;6:39–49. doi: 10.1586/14787210.6.1.39. [DOI] [PubMed] [Google Scholar]
  • 8.Shlaes DM, Projan SJ, Edwards JE. Antibiotic Discovery: State of the State. ASM News. 2004;70:275–281. [Google Scholar]
  • 9.Peters NK, Dixon DM, Holland SM, Fauci AS. The Research Agenda of the National Institute of Allergy and Infectious Diseases for Antimicrobial Resistance. J Infect Dis. 2008;197:1087–1093. doi: 10.1086/533451. [DOI] [PubMed] [Google Scholar]
  • 10.Christensen DR, Hartman LJ, Loveless BM, Frye MS, Shipley MA, Bridge DL, Richards MJ, Kaplan RS, Garrison J, Baldwin CD, Kulesh DA, Norwood DA. Detection of Biological Threat Agents by Real-Time PCR: Comparison of Assay Performance on the R.A.P.I.D., the LightCycler, and the Smart Cycler Platforms. Clin Chem. 2006;52:141–145. doi: 10.1373/clinchem.2005.052522. [DOI] [PubMed] [Google Scholar]
  • 11.Bode E, Hurtle W, Norwood D. Real-Time PCR Assay for a Unique Chromosomal Sequence of Bacillus anthracis. J Clin Microbiol. 2004;42:5825–5831. doi: 10.1128/JCM.42.12.5825-5831.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Probert WS, Schrader KN, Khuong NY, Bystrom SL, Graves MH. Real-Time Multiplex PCR Assay for Detection of Brucella spp., B. abortus, and B. melitensis. J Clin Microbiol. 2004;42:1290–1293. doi: 10.1128/JCM.42.3.1290-1293.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Tomaso H, Reisinger EC, Al Dahouk S, Frangoulidis D, Rakin A, Landt O, Neubauer H. Rapid Detection of Yersinia pestis with Multiplex Real-Time PCR Assays Using Fluorescent Hybridization Probes. FEMS Immunol Med Microbiol. 2003;38:117–126. doi: 10.1016/S0928-8244(03)00184-6. [DOI] [PubMed] [Google Scholar]
  • 14.Versage JL, Severin DD, Chu MC, Petersen JM. Development of a Multitarget Real-Time TaqMan PCR Assay for Enhanced Detection of Francisella tularensis in Complex Specimens. J Clin Microbiol. 2003;41:5492–5499. doi: 10.1128/JCM.41.12.5492-5499.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Barrow EW, Valderas MW, Bourne PC, Barrow WW. Newly Developed Colorimetric Drug Screening Assay for Bacillus anthracis. Int J Antimicrob Agents. 2006;27:178–180. doi: 10.1016/j.ijantimicag.2005.12.001. [DOI] [PubMed] [Google Scholar]
  • 16.Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing; Nineteenth Informational Supplement. 2009;29:1–152. [Google Scholar]
  • 17.Clinical and Laboratory Standards Institute. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically. Approved Standard. 2009;29 [Google Scholar]
  • 18.Barrow EW, Dreier J, Reinelt S, Bourne PC, Barrow WW. In Vitro Efficacy of New Antifolates against Trim-ethoprim-Resistant Bacillus anthracis. Antimicrob Agents Chemother. 2007;51:4447–4452. doi: 10.1128/AAC.00628-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Chung S, Wendeler M, Rausch JW, Beilhartz G, Gotte M, O’Keefe BR, Bermingham A, Beutler JA, Liu S, Zhuang X, Le Grice SF. Structure-Activity Analysis of Vinylogous Urea Inhibitors of Human Immunodeficiency Virus-Encoded Ribonuclease H. Antimicrob Agents Chemother. 2010;54:3913–3921. doi: 10.1128/AAC.00434-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Wendeler M, Lee HF, Bermingham A, Miller JT, Cher-tov O, Bona MK, Baichoo NS, Ehteshami M, Beutler J, O’Keefe BR, Gotte M, Kvaratskhelia M, Le Grice S. Vinylogous Ureas as a Novel Class of Inhibitors of Reverse Transcriptase-Associated Ribonuclease H Activity. ACS Chem Biol. 2008;3:635–644. doi: 10.1021/cb8001039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Parniak MA, Min KL, Budihas SR, Le Grice SF, Beutler JA. A Fluorescence-Based High-Throughput Screening Assay for Inhibitors of Human Immunodeficiency Virus-1 Reverse Transcriptase-Associated Ribonuclease H Activity. Anal Biochem. 2003;322:33–39. doi: 10.1016/j.ab.2003.06.001. [DOI] [PubMed] [Google Scholar]
  • 22.Namba K, Zheng X, Motoshima K, Kobayashi H, Tai A, Takahashi E, Sasaki K, Okamoto K, Kakuta H. Design and Synthesis of Benzenesulfonanilides Active against Methicillin-Resistant Staphylococcus aureus and Vancomy-cin-Resistant Enterococcus. Bioorg Med Chem. 2008;16:6131–6144. doi: 10.1016/j.bmc.2008.04.040. [DOI] [PubMed] [Google Scholar]
  • 23.Brown GM. The Biosynthesis of Folic Acid. II. Inhibition by Sulfonamides. J Biol Chem. 1962;237:536–540. [PubMed] [Google Scholar]
  • 24.Valderas MW, Bourne PC, Barrow WW. Genetic Basis for Sulfonamide Resistance in Bacillus anthracis. Microbial Drug Resistance. 2007;13:12–21. doi: 10.1089/mdr.2006.9992. [DOI] [PubMed] [Google Scholar]
  • 25.Valderas MW, Andi B, Barrow WW, Cook PF. Examination of Intrinsic Sulfonamide Resistance in Bacillus anthracis: A Novel Assay for Dihydropteroate Synthase. Biochim Biophys Acta. 2008;1780:848–853. doi: 10.1016/j.bbagen.2008.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bourne CR, Wakeham N, Bunce RA, Nammalwar B, Berlin KD, Barrow WW. Classifying Compound Mechanism of Action for Linking Whole Cell Phenotypes to Molecular Targets. J Mol Recognit. 2012;25:216–223. doi: 10.1002/jmr.2174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Macarron R. Critical Review of the Role of HTS in Drug Discovery. Drug Discov Today. 2006;11:277–279. doi: 10.1016/j.drudis.2006.02.001. [DOI] [PubMed] [Google Scholar]
  • 28.Chan PF, Holmes DJ, Payne DJ. Finding the Gems Using Genomic Discovery: Antibacterial Drug Discovery Strategies—The Successes and the Challenges. Drug Discov Today. 2004;1:519–527. [Google Scholar]
  • 29.Livermore DM. Discovery Research: The Scientific Challenge of Finding New Antibiotics. J Antimicrob Chemother. 2011;66:1941–1944. doi: 10.1093/jac/dkr262. [DOI] [PubMed] [Google Scholar]
  • 30.Payne DJ, Gwynn MN, Holmes DJ, Pompliano DL. Drugs for Bad Bugs: Confronting the Challenges of Antibacterial Discovery. Nat Rev Drug Discov. 2007;6:29–40. doi: 10.1038/nrd2201. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SUppData

Table S2. MIC evaluations (μg/mL) of additional Commercial Drugs against Select Agents. B.a. = B. anthracis Ames, Y.p. = Y. pestis Co92, F.t. = F. tularensis SchuS4, Br.a. = B. abortus 2308. (n 4).

Table S3. Additional MyriaScreen compounds with MIC values for compounds that inhibited all four Select Agents MyriaScreen compounds are listed as ID number which includes ID, plate and well number. MW = molecular weight, B.a. = B. anthracis Ames, Y.p. = Y. pestis Co92, F.t. = F. tularensis SchuS4, Br.a. = B. abortus 2308.

NIHMS477948-supplement-SUppData.pdf (6.8MB, pdf)

RESOURCES