A novel bicyclic 2,4-diaminopyrimidine inhibitor of Streptococcus suis dihydrofolate reductase

Bacteria strains and growth conditions

S. suis strains P1/7 and HE06 are serotype 2 isolates obtained from a pig with meningitis and an infected human, respectively (Jacobs, Van den Berg & Loeffen, 1996; Holden et al., 2009; Maneerat et al., 2013). Serotype 2 strains were chosen because of their association with pathogenicity (Lun et al., 2007; Wertheim et al., 2009a; Maneerat et al., 2013). S. suis was grown on brain heart infusion (BHI) plates (Beckton Dickinson, Franklin Lakes, NJ, USA) at 37 °C with 5% CO₂. Escherichia coli PA414 lacking folA and thyA genes, encoding dihydrofolate reductase (DHFR) and thymidylate synthase (TS), respectively (Ahrweiler & Frieden, 1988) was used as a surrogate for expressing S. suis DHFR. The deletion of thyA is required for survival of DHFR deficient mutant (Ahrweiler & Frieden, 1988). E. coli PA414 transformed with various expression plasmids was grown in Luria Bertani (LB) broth supplemented as necessary with 50 µg/mL kanamycin, 100 µg/mL ampicillin, 10 µg/mL chloramphenicol, 0.2% (w/v) arabinose, and/or 50 µg/mL thymidine.

Bacterial growth inhibition assay

The Clinical Laboratory Standard Institute (CLSI) broth microdilution method was used to test antibacterial activity of Pathogen Box compounds and selected DHFR inhibitors (The Clinical and Laboratory Standards Institute, 2018). Briefly, 5 × 10⁴ colony forming units (CFUs) of S. suis or E. coli PA414 were incubated with compounds at 10 µM (or other concentrations as indicated) in cation-adjusted Mueller Hinton broth (Beckton Dickinson, Franklin Lakes, NJ, USA) in 96-well plates at 37 °C with or without 5% CO₂, respectively. Growth was monitored by measuring optical density at 600 nm (OD₆₀₀) using a SpectraMax M5 spectrophotometer (Molecular Devices, San Jose, CA, USA). Percent growth of each compound-treated bacteria strain was calculated using the following formula: ${\displaystyle \left(\text{OD of compound treated bacteria}\times 100\right)∕\left(\text{OD of DMSO treated bacteria}\right).}$

Construction of expression plasmids

The dhfr gene encoding S. suis dihydrofolate reductase (SsDHFR) was amplified from clarified lysate of S. suis P1/7 and S. suis HE06 using high-fidelity Phusion DNA polymerase (New England Biolabs, Ipswich, MA, USA) and primers 5′ctggtgccgcgcggcagccatATGACTAAAAAGATTGTTGC3′ and 5′tcgggctttgttagcagccggatc CTAACCATCTCTTCTTTCATAG3′ (lower case denotes sequence homologous to pET15b while upper case denotes sequence homologous to the dhfr gene). The PCR products were cloned into NdeI and BamHI digested pET15b plasmid using a Gibson Assembly kit (New England Biolabs, Ipswich, MA, USA), according to manufacturer’s protocol. The resulting plasmids were subjected to Sanger sequencing (1st BASE, Singapore). Nucleotide sequences of the dhfr gene from S. suis P1/7 and S. suis HE06 were deposited in NCBI Genbank with accession number MH388486 and MH388487, respectively.

Complementation assay of E. coli surrogate

E. coli PA414 lacking functional DHFR and TS enzymes (Ahrweiler & Frieden, 1988) was used as a surrogate host to study the function of exogenously expressed SsDHFR. E. coli PA414 was co-transformed with (1) pBAD33 or pBAD33-EcTS and (2) pET15b or pET15b-SsDHFR. To test whether SsDHFR can complement for the loss of E. coli DHFR in the E. coli surrogate, overnight cultures of E. coli PA414 carrying various combination of plasmids were pelleted, washed, cell density adjusted to OD₆₀₀ of 4, serially-diluted, and spotted onto LB plates supplemented with appropriate antibiotics and either 50 µg/mL thymidine or 0.2% (w/v) arabinose. Plates were incubated at 37 °C overnight. Growth on agar plate was observed.

Overexpression of SsDHFR in E. coli BL21(DE3)

Cultures of E. coli BL21(DE3) carrying pET15b or pET15b-SsDHFR were grown with shaking at 37 °C in LB supplemented with 100 µg/mL ampicillin. When OD₆₀₀ reached 1, IPTG was added to 40 µM, temperature shifted to 16 °C, and incubated with shaking overnight. Bacterial pellets were collected by centrifugation, resuspended in lysis buffer (20 mM potassium phosphate buffer, 0.1 mM EDTA, 10 mM DTT, 50 mM KCl, 20% glycerol), and sonicated. Protein concentration of clarified bacteria lysate was determined by Bradford assay (Biorad, Hercules, CA, USA) according to the manufacturer’s protocol.

In vitro DHFR activity assay

DHFR uses NADPH as a cofactor for the conversion of dihydrofolate (DHF) to tetrahydrofolate (THF). Since the amount of NADPH consumed is directly proportional to the amount of THF product, monitoring NADPH consumed using absorbance at 340 nm (A₃₄₀; ε₃₄₀ of 12,300 M⁻¹ cm⁻¹) is indicative of DHFR enzymatic activity. The DHFR activity assay was performed as previously described (Tirakarn et al., 2012; Songsungthong et al., 2019). Briefly, using kinetics mode of a UV-Vis spectrophotometer (Agilent Technologies, Santa Clara, CA, USA), A₃₄₀ value of the reaction mixture (50 mM TES, pH 7.0, 75 mM β-mercaptoethanol, 1 mM EDTA, 1 mg/mL BSA, 0.1 mM NADPH, 0.1 mM dihydrofolate) was set to blank (A₃₄₀ = 0). Inhibitors were added to the reaction mixture as necessary. 1 µg of total protein from lysate of E. coli BL21(DE3) containing pET15b or overexpressing SsDHFR was added to reaction buffer to initiate the enzymatic reaction. The amount of SsDHFR containing lysate added was predetermined to give linear reaction kinetics. NADPH reduction was monitored by measuring A₃₄₀ reduction for 100 s. A₃₄₀ reduction per minute values were recorded.

Percent SsDHFR activity was calculated using the following formula:

(A₃₄₀ reduction per minute of compound-treated lysate × 100)/(A₃₄₀ reduction per minute of DMSO-treated lysate of BL21 expressing SsDHFR).

Statistical analysis

Data are shown as mean ± standard error of the mean (SEM). At least three independent experiments were performed. Statistical analysis using one-way ANOVA with Tukey’s post test was performed. Statistical significance is noted on the graph by asterisks. * denotes P < 0.05. ** denotes P < 0.01. ** denotes P < 0.001. *** denotes P < 0.0001. ns denotes not statistically significant.

Identifying compounds with S. suis growth inhibitory activity from Pathogen Box

We screened the Pathogen Box library for compounds active against S. suis P1/7 and S. suis HE06 isolated from an infected pig and human, respectively (Jacobs, Van den Berg & Loeffen, 1996; Holden et al., 2009; Maneerat et al., 2013). Thirty compounds with high inhibitory activity (with average percent inhibition of 90–100%) at 10 µM were identified (Table S1). Seven of the hits including rifampicin, levofloxacin, and linezolid are reference compounds known to be broadly effective against Gram-positive bacteria (highlighted in blue in Table S1). Twenty three other hits are compounds with structures distinct from commercially available antibiotics (highlighted in green in Table S1). The mechanisms of action of most non-reference hits are unknown. MMV675968, one of the hits, has been shown to inhibit the growth of protozoan parasites and Gram-negative bacteria (Lau et al., 2001; Nelson & Rosowsky, 2001; Popov et al., 2006; Songsungthong et al., 2019) via inhibition of dihydrofolate reductase (DHFR), an enzyme in the thymidylate cycle.

MMV675968 is an inhibitor of S. suis DHFR

MMV675968 inhibits DHFR of protozoan parasites and of Gram-negative bacteria (Lau et al., 2001; Nelson & Rosowsky, 2001; Popov et al., 2006; Songsungthong et al., 2019). We therefore hypothesized that MMV675968 inhibited S. suis growth through inhibition of S. suis DHFR (SsDHFR). An E. coli surrogate assay and an in vitro DHFR activity assay were used to test this hypothesis. For the E. coli surrogate assay, E. coli PA414 deficient in E. coli DHFR (EcDHFR) and E. coli thymidylate synthase (EcTS) of the thymidylate cycle making the strain a thymidine auxotroph (Ahrweiler & Frieden, 1988), was used as a host for exogenously expressing SsDHFR. E. coli PA414 transformant controls carrying two empty vectors (pET15b and pBAD33), or expressing EcTS alone did not grow in the absence of thymidine supplement as expected (Fig. 1A). E. coli PA414 expressing EcTS together with SsDHFR was able to grow without thymidine supplement (Fig. 1A), indicating that SsDHFR was functional and complemented for the loss of EcDHFR. Growth of E. coli PA414 in the absence of thymidine was therefore dependent on the function of exogenously expressed SsDHFR.

Trimethoprim (TMP), a bacterial DHFR inhibitor (Baker et al., 1981), inhibited the growth of E. coli PA414 expressing SsDHFR to 19% of untreated (Fig. 1B), confirming that trimethoprim inhibits SsDHFR. MMV675968 was more effective than TMP as seen by its ability to inhibit the growth of E. coli PA414 expressing SsDHFR to 7% (Fig. 1B). E. coli PA414 growth inhibition by MMV675968 was reversed in the presence of thymidine supplement (Fig. 1C), suggesting that the thymidine-related pathway is the principal target of MMV675968 in E. coli surrogate consistent with DHFR being a target.

Next, we tested whether MMV675968 could inhibit SsDHFR activity directly by an in vitro DHFR activity assay. DHFR activity present in lysate of E. coli BL21(DE3) carrying an empty pET15b plasmid accounted for only 2% of DHFR activity of lysate of E. coli BL21(DE3) overexpressing SsDHFR (Fig. 1D), indicating that most of the DHFR activity observed was from overexpressed SsDHFR rather than endogenous EcDHFR. In the presence of 1 µM MMV675968, DHFR activity decreased to 2% of that of untreated lysate (Fig. 1D), indicating that MMV675968 inhibits SsDHFR directly. Results from both E. coli surrogate assay and DHFR activity assay thus indicate that MMV675968 inhibits SsDHFR.

We also tested whether thymidine-related pathway is the main target of MMV675968 in S. suis by testing whether growth inhibition was rescued by thymidine supplement. Growth of both S. suis P1/7 and S. suis HE06 were inhibited in the presence of 1 µM MMV675968 (Fig. 1E). Growth inhibition of both S. suis strains was reversed with thymidine supplement (Fig. 1E), indicating that the main target of MMV675968 within S. suis is a thymidine-related pathway.

Bicyclic 2,4-diaminopyrimidines with long and flexible side chain associate with higher inhibitory activity against SsDHFR and S. suis

To investigate which chemical scaffolds are better at inhibiting SsDHFR and S. suis growth, eight DHFR inhibitors of various chemical structures (Figs. 2A–2C), which include commercially available compounds (pemetrexed, trimethoprim, pyrimethamine, cycloguanil, and methotrexate), compounds from Medicines for Malaria Venture (MMV)’s Malaria Box (MMV667486 and MMV667487), and a compound from Pathogen Box (MMV675968), were tested for SsDHFR inhibitory activity and S. suis growth inhibition. Pemetrexed, MMV667486, and MMV667487 did not show significant SsDHFR or S. suis growth inhibitory activity (Figs. 2A, 2D and 2E), indicating that these structures do not inhibit SsDHFR and consequently are unable to inhibit S. suis growth. Trimethoprim, pyrimethamine, and cycloguanil were moderately effective at inhibiting SsDHFR and S. suis growth (defined as being effective at 10 µM but not at 1 µM) (Figs. 2B, 2D–2F). Methotrexate and MMV675968 inhibited SsDHFR more effectively that trimethoprim, and completely inhibited S. suis growth, even at 1 µM (Figs. 2C–2F). Methotrexate and MMV675968, both of which are bicyclic 2,4-diaminopyrimidines with long and flexible side chains, are highly effective inhibitors of both SsDHFR and S. suis growth. Both S. suis P1/7 and S. suis HE06 strains shared a similar sensitivity pattern towards various DHFR inhibitors (Figs. 2E and 2F) consistent with the identical DHFR sequences between the two strains (Figs. S1A–S1B).