Penicillin facilitates the entry of antisense constructs into Streptococcus mutans


  • Felicity S.A. McLeod,

    1. Department of Microbiology and Immunology, University of Otago, Dunedin, New Zealand
    Search for more papers by this author
  • Robin S. Simmonds

    Corresponding author
    1. Department of Microbiology and Immunology, University of Otago, Dunedin, New Zealand
    • Correspondence: Robin S. Simmonds, Department of Microbiology and Immunology, University of Otago, P.O. Box 56, 9054 Dunedin, New Zealand. Tel.: +64 3 479 7478: fax; +64 3 479 8540; e-mail:

    Search for more papers by this author


Antisense oligonucleotides (AS-ODN) target genes in a sequence-specific manner inhibit gene function and have potential use as antimicrobial agents. Cell barriers, such as peptidoglycan, cell surface proteins and lipopolysaccharide membranes, prevent delivery of AS-ODN into the bacterial cell, limiting their use as an effective treatment option. The β-lactam antibiotic penicillin was examined for its ability to deliver phosphorothioate oligodeoxyribonucleotides (PS-ODNs) and γ32 P-ODN into Streptococcus mutans OMZ175. Treatment of lag-phase S. mutans OMZ175 cells with penicillin and FBA (PS-ODN targeting the fructose-biphosphate aldolase gene), resulted in prolonged suppression of growth (> 24 h) and fba expression (656.9 ± 194.4-fold decrease at 5 h). Suppression of both cell growth and fba expression corresponded with a greater amount of γ32 P-ODN becoming cell associated, with a maximum γ32 P-ODN concentration per cell achieved 5 h after penicillin treatment (6.50 ± 1.39 × 108 molecules per CFU). This study confirms that for S. mutans OMZ175, the peptidoglycan layer acts as a major barrier preventing AS-ODN penetration and suggests that the use of agents such as penicillin that interfere with peptidoglycan integrity can significantly increase the uptake of PS-ODN by these cells.


Infection caused by antibiotic-resistant bacteria has become a major global healthcare issue, and there is a rapidly growing need for the development of new antimicrobials. Antibiotic development based on incremental change that produces new iterations of old chemical structures is unlikely to yield innovative and effective new antimicrobials. A transformational change is urgently required, and there is increasing need for antimicrobials with a novel method of action (Overbye & Barrett, 2005).

One technology currently under investigation is the use of antisense oligonucleotides (AS-ODN) to target genes in a sequence-specific manner and inhibit gene function. Antisense sequences can be designed to bind complementary RNA gene sequences through Watson–Crick pairing to block translational initiation or elongation (Bennett & Swayze, 2010). It is thought the presence of an AS-ODN molecule sterically blocks the ability of the ribosome to bind to the mRNA, preventing translation (Good, 2003). Degradation of the targeted RNA gene may involve the recruitment and activation of RNase-H (Rassmussen et al., 2007) or RNase-P (Shen et al., 2009). Numerous clinical trials using AS-ODN technology are being conducted for the treatment of diseases, such as cancer (Iversen et al., 2003), HIV/AIDS and other viral infections (Tripathi et al., 2005) and various autoimmune disorders (Ricotta & Frishman, 2012), with many showing promising therapeutic potential. By contrast, the use of AS-ODN's to treat bacterial infections is less advanced.

It might be expected that microbial cell surface structures, such as lipopolysaccharide membranes, lipoteichoic acids, teichoic acids and peptidoglycan, would present formidable barriers to entry of AS-ODN's to the bacterial cell cytoplasm, but little evidence has been reported regarding their effect. AS-ODN's conjugated to membrane-penetrating peptides show improved bacterial penetration (Tilley et al., 2006) and coupling of a positively charged peptide (KFFKFFKFFK) to a peptide nucleic acid (PNA) construct greatly increased penetration of the PNA into Escherichia coli (Good et al., 2001). These delivery peptides probably enhance cell penetration by localization of the PNA to the negatively charged lipopolysaccharide of the outer membrane. The coupling of AS-ODN's to membrane-penetrating peptides (RFFRFFRFFXB) has been shown to increase their permeability into E. coli (Mellbye et al., 2010), S. enterica (Mitev et al., 2009), Klebsiella pneumoniae (Kurupati et al., 2007) and Burkholderia multivorans (Greenberg et al., 2010). Encapsulation of AS-ODN in liposomes has also opened up new possibilities for cell delivery. Liposomes may fuse with bacterial cell walls improving AS-ODN penetration through the bacterial cell wall, as well as facilitating the penetration of AS-ODN through the cytoplasmic membrane (Meng et al., 2009). Injection of anionic liposome encased PS-ODN targeted towards mec-A, into the tail vein of mice infected with methicillin-resistant Staphylococcus aureus (MRSA), has been shown to restore the MRSA to oxacillin susceptibility and to rescue these animals from lethal sepsis (Meng et al., 2009). Zoocin A, a peptidoglycan hydrolase, can facilitate the uptake of PS-ODN by S. mutans OMZ175 (Dufour et al., 2011), suggesting that at least in Gram-positive bacteria, the peptidoglycan layer is a major barrier to PS-ODN entry. Penicillin is a β-lactam antibiotic known for its ability to inhibit the synthesis of the peptidoglycan layer of bacterial cells leading to a loss of structural integrity (Tomasz, 1979). It may also stimulate autolysis of target bacteria by the deregulation of genes encoding autolytic enzymes (Penyige et al., 2002). Therefore, the use of such cell wall targeting antibiotics presents a possible method of inducing bacterial cell permeabilization that has already been extensively studied and is already approved for in vivo use (Moellering & Weinberg, 1971; Chung et al., 2009).

In this study, treatment of S. mutans OMZ175 with penicillin was used to demonstrate that agents that compromise the peptidoglycan barrier can markedly affect the uptake of PS-ODN's into bacteria.

Materials and methods

Growth conditions and preparation of penicillin and zoocin A

Streptococcus mutans OMZ175 was grown at 37 °C in air + 5% CO2 in Todd Hewitt broth (THB, Difco). Columbia blood agar (Difco) supplemented with 5% whole human blood was used for routine culture. His-tagged recombinant zoocin A was produced from E. coli zooA and purified as described previously (Burne et al., 1999). Penicillin was prepared as a 200 mg mL−1 stock in reverse osmosis water and filter (0.22 μm, Millipore) sterilized.

PS-ODN preparation

The design of the phosphorothioate oligonucleotides (PS-ODN) used in this study has been described previously (Dufour et al., 2011). In brief, both were 18-nucleotide PS-ODNs, ATS was a randomly generated sequence with no extensive complementary sequence in the S. mutans UA159 genome, and FBA was designed to complement the ATG start codon of the fructose-biphosphate aldolase gene.

Growth inhibition assay

Antibacterial agents were serially diluted in THB in siliconized tubes to attain the desired concentrations and 10 μL volumes dispensed into the wells of a 96-well low cell binding microtitre plate (Nalgene NUNC International, Denmark). A 5% inoculum of an overnight culture of S. mutans OMZ175 was dispensed into the wells, and the total volume of each well made up to 200 μL with THB. The microtitre plate was then incubated in a plate reader (Multiskan Ascent Microtiterplate Reader, LabSystems, Finland) at 37 °C for 48 h with absorbance readings (595 nm) taken every hour. All tests were conducted in triplicate and controls included. Sigmoidal curves were fitted to each set of triplicate growth data (Microsoft Excel), and the equation for each curve used to calculate the time taken for that culture to reach an initial OD + 0.1 (lag phase). Differences between lag-phase values were analysed for statistical significance using the Tukey's multiple comparison test (Prism Software).

Corresponding viability assays were performed by preparing serial dilutions in peptone water. Five replicates of 10 μL of dilutions 10−3–10−8 were spotted onto CAB agar, and 100 μL of dilutions 10−0–10−2 were spread onto CAB agar. The dilutions, spot and spread plating of each culture were performed in triplicate.

Determination of sublethal penicillin concentration

The penicillin concentration (0.8 μg mL−1) selected as sub-lethal for lag-phase S. mutans OMZ175 cells significantly (P < 0.001) increased the lag phase without decreasing the OD of the culture at 24 h in comparison with the untreated control. The 24-h viable count of the lag-phase penicillin-treated cells (9.1 × 109 CFU mL−1) was not significantly different from that of the untreated cells (9.8 × 109 CFU mL−1).

Penicillin (0.05–100 μg mL−1) applied to exponential-phase (0.35 OD595 nm) S. mutans OMZ175 cells had no measurable effect on culture OD; however, it did significantly affect viable counts. Addition of 5 μg mL−1 penicillin to exponential-phase cells induced a significant decrease (P = 0.001) in cell viability, from 8.3 ×108 CFU mL−1 pretreatment, to 5.3 × 105 CFU mL−1 at 6 h post-treatment. By 20 h post-treatment, the cells had recovered to a level (8.1 × 108 CFU mL−1) not significantly different from the pretreatment level.

Growth inhibition of S. mutans OMZ175 in the presence of penicillin and PS-ODN

Lag and exponential-phase S. mutans OMZ175 cells were incubated with penicillin at final concentrations of 0.8 and 5 μg mL−1, respectively. FBA or ATS was added to a final concentration of 10 μM.

Determination of target gene sequence

Following treatment of lag-phase S. mutans OMZ175 with zoocin A + FBA, and penicillin + FBA, chromosomal DNA was extracted from 30 colonies and the sequence of fba for each determined as described previously (Dufour et al., 2011).

Determination of gene expression levels in S. mutans OMZ175

Eight millilitre volumes of lag and exponential-phase S. mutans OMZ175 cells were prepared and treated with penicillin and PS-ODN as described above for the growth inhibition experiments. Samples were removed at: the time of treatment, 30 min, 5 and 16 h post-treatment. The levels of mRNA transcript of fba, 16s RNA gene and gyrA in S. mutans OMZ175 were determined for each sample using a quantitative reverse transcriptase PCR (qRT-PCR) as described previously (Dufour et al., 2011).

Determination of PS-ODN bacterial cell association

ATS was radiolabelled at the 5-'end with 10U T4 polynucleotide kinase (New England Biolabs), 2 μL of [γ32 P] ATP (3000 Ci mmol−1, 10 mCi mL−1, 250 μCi) (Perkin Elmer), 3 μL of ligase buffer (New England Biolabs) and 6 μL ATS (final concentration 20 mM) in a total volume of 20 μL. Following incubation at 37 °C for 1 h, unincorporated [γ32 P] ATP was separated using a prepacked sephadex™ -μM membrane filters, three 1 μL volumes of the radiolabelled ATS were spotted onto separate membranes and each membrane placed in a 4 mL scintillation vial (Kartell). Two millilitres of scintillation fluid (Optiphase Hi-safe 2, Perkin Elmer) was added to each vial, and the vial counted using the γ32 P programme on a Quanta Smart scintillation counter by Perkin Elmer for Tri-Carb® liquid scintillation. The amount of radiolabel incorporated was determined using the method described by Roche for use with the Roche Mini Spin Columns.

Lag- and exponential-phase cultures of S. mutans OMZ175 were prepared as described previously, and 150 μL volumes added to 12 wells of a 96-well low cell-binding NUNC plate. To 6 wells, 40 μL of penicillin was added to attain a final concentration of 0.8 and 5 μg mL−1 for lag- and exponential-phase cells, respectively, and to the other six wells, 40 μL of MQ water was added. Each set of six wells was divided into two triplicates receiving either 10 μL γ32 P-ATS (20 μM final concentration) or 10 μL MQ water. Eleven sets were prepared, and the contents of one complete set assayed at 0, 0.5, 1.5, 2, 2.5, 3, 4, 5, 7, 10 and 12 h. Cells were collected by centrifugation (14 000 g, 10 min), washed three times with PBS, and resuspended in 200 μL of PBS before being pipetted on to separate 0.45-μM membrane filters held in a sealed Millipore Membrane Filter (1225 sampling manifold, Millipore). Suction was applied to draw all visible liquid through the membrane, and each membrane washed three times with 1 mL volumes of PBS. Once dry the membranes were transferred using sterile forceps to individual 4-mL scintillation vials. Scintillation fluid was then added to each vial, and each vial read as previously described.

The above radiolabel experiments were repeated using zoocin A instead of penicillin. The sublethal concentration of zoocin A (0.1 and 0.4 μg mL−1 for lag- and exponential-phase cells, respectively) had previously been determined (Dufour et al., 2011).

Results and discussion

Effect of penicillin and PS-ODN's on bacterial growth and expression of fba

Viability assays showed that following an initial reduction in viability, lag-phase S. mutans OMZ175 treated with 0.8 μg mL−1 penicillin alone rapidly recovered, while cultures treated with a combination of penicillin and FBA failed to show any recovery within 48 h (Fig. 1a). The addition of 0.8 μg mL−1 penicillin and 10 μM FBA to S. mutans OMZ175 cultures resulted in a lag phase that was significantly longer (P = 0.001) than that observed for the addition of penicillin alone (Table 1). The addition of 0.8 μg mL−1 penicillin and 10 μM ATS to S. mutans OMZ175 cultures resulted in lag phase times that were not significantly different from those obtained from cultures treated with penicillin alone. In the absence of penicillin, FBA (1–20 μM) had no significant effect on S. mutans OMZ175 growth (Table 1). It is unlikely that growth recovery occurred as the result of selection of a mutant phenotype, as PCR and sequencing of the target gene from 30 colonies of S. mutans OMZ175 obtained following penicillin + FBA treatment, failed to identify any mutation of fba.

Table 1. Growth of lag- or exponential-phase Streptococcus mutans OMZ175 cultures in the presence of penicillin and PS-ODN
Penicillin concentration (μg mL−1) plus PS-ODN (10 μM)Lag or recovery time (h) ± SDStatistical significancea (P)
  1. a

    All penicillin and PS-ODN alone values compared with the 0 + 0 control of the appropriate growth phase. All penicillin + PS-ODN values are compared with their corresponding penicillin only control.

  2. b

    As sigmoidal curves could not be fitted to the growth curves of penicillin-treated exponential-phase cells, statistical variance was calculated from viable counts at 24 h by use of the Student's t-test.

0 + 09.24 ± 0.52N/A
0 + ATS9.14 ± 0.14No (> 0.05)
0 + FBA8.88 ± 0.34No (> 0.05)
0.8 + 016.97 ± 0.35Yes (< 0.001)
0.8 + ATS17.22 ± 0.24No (> 0.05)
0.8 + FBA≥ 24Yes (< 0.001)
0 + 01.94 ± 0.31N/A
0 + ATS1.84 ± 0.38No (> 0.05)
0 + FBA1.81 ± 0.31No (> 0.05)
5 + 0≥ 24Yes (< 0.001)b
5 + ATS≤ 24bNo (> 0.05)b
5 + FBA≤ 24bNo (> 0.05)b
Figure 1.

Growth of Streptococcus mutans OMZ175 cultures treated with various inhibitory agents. Viability of lag-phase (a) or exponential-phase (b) S. mutans OMZ175 cultures treated with: (1), image_n/fml12286-gra-0001.pngTHB only; (2), image_n/fml12286-gra-0002.png10 μM ATS; (3), image_n/fml12286-gra-0003.png10 μM FBA; (4),image_n/fml12286-gra-0004.png 0.8 μg mL−1 (lag) or 5 μg mL−1 (exponential) penicillin; (5), image_n/fml12286-gra-0005.png0.8 μg mL−1 (lag) or 5 μg mL−1 (exponential) penicillin + 10 μM ATS; (6), image_n/fml12286-gra-0006.png 0.8 μg mL−1 (lag) or 5 μg mL−1 (exponential) penicillin + 10 μM FBA;. Error bars represent the SD of each triplicate. The time point at which the treatment was added is indicated by an arrow for each graph. (b) shows S. mutans at an initial lag-phase concentration, but treatment was not added until exponential phase (c. 108 Log10 CFU mL−1, OD of 0.35) was reached.

These experiments were replicated using exponential-phase S. mutans OMZ175 cells. While a penicillin concentration of 5 μg mL−1 significantly decreased the viability of exponential-phase S. mutans OMZ175 compared with the untreated control, the addition of penicillin + 10 μM FBA appeared to make no difference compared with penicillin alone (Fig. 1b). The addition of 5 μg mL−1 penicillin and 10 μM of either FBA or ATS to S. mutans OMZ175 in 0.05 M phosphate/potassium-buffered media did not significantly affect these results.

To demonstrate that the growth inhibition caused by penicillin and FBA was the result of inhibition of fba, qRT-PCR was used to measure relative gene expression levels. No treatments caused significant differences (P = 0.001) between transcript expression levels (math formula) for either control gene (16s RNA gene or gyrA). Compared with cultures treated with either penicillin or FBA alone, a significant decrease (P = 0.001) in the expression of fba was observed for lag-phase cultures treated with a combination of penicillin and FBA. Thirty minutes, 5 and 16 h after treatment addition, a 5.7 ± 0.2, 656.9 ± 192.4 and 640.6 ± 186.8-fold decrease in FBA mRNA expression, respectively, was observed. All other treatment types resulted in no significant differences in the expression level of fba. No significant differences in mRNA expression, for fba or either control gene, were found for exponential-phase S. mutans OMZ175, at any time point, regardless of treatment.

Association of PS-ODN with S. mutans cells

A radiolabel assay was used to demonstrate that the PS-ODN's entered S. mutans OMZ175 cells only in the presence of agents capable of compromising peptidoglycan integrity. For the treatment of S. mutans OMZ175 with γ32 P-ATS alone, all time points showed c. 104–105 molecules of ATS associated with each CFU. This remained true even as cell numbers increased, suggesting that free ATS molecules in the media associated with each new cell as it formed. Lag-phase S. mutans OMZ175 cultures treated with penicillin, and γ32 P-ATS showed a significant increase (P = 0.001) in the amount of γ32 P-ATS associated with each CFU (Fig. 2a). Increased uptake of γ32 P-ATS was shown to occur for the first 5 h of treatment (reaching 6.50 ± 1.39 × 108 γ32 P-ATS molecules per CFU), after which there was a rapid decrease in the amount of γ32 P-ATS present inside each cell. This rapid decrease in γ32 P-ATS corresponded with the time at which cultures treated with penicillin and ATS began to recover and grow. C. 2.5–3 log decrease in γ32 P-ATS per CFU, corresponded to c. 2.5–3 log increase in the number of viable cells per mL. Similarly, at the same time as both lag (Fig. 2b) or exponential (Fig. 2d)-phase S. mutans OMZ175 cultures treated with zoocin A + ATS began to recover and grow, a rapid loss in the amount of γ32 P-ATS per CFU was observed, with c. 3 log decrease in γ32 P-ATS per CFU corresponding to c. 2.5–3 log increase in the number of viable cells per mL for either growth phase. These results suggests that the observed decrease in γ32 P-ATS per CFU was due to dilution as the result of cellular replication, rather than any loss due to cellular leakage or diffusion out of the bacterial cells. It has been reported that after 14 days incubation with Mycobacterium tuberculosis, PS-ODN remained intact and chemically unchanged (Harth et al., 2002). Assuming the same holds true for S. mutans, it is unlikely that significant intracellular PS-ODN degradation occurred over the 24 h that these experiments were conducted.

Figure 2.

The amount of ATS associated with of Streptococcus mutans OMZ175 cultures treated with various inhibitory agents. Results for lag-phase cultures are given in (a and b), and for exponential-phase cultures in (c and d). Results for cultures treated with penicillin are given in (a and c), and for cultures treated with zoocin A in (b and d). The viable counts for each culture are given as: image_n/fml12286-gra-0005.png, ATS alone and image_n/fml12286-gra-0007.png, inhibitory agent + ATS. The number of ATS molecules per CFU are given as: image_n/fml12286-gra-0008.png, ATS alone and; image_n/fml12286-gra-0009.png, inhibitory agent + ATS. Error bars represent the SD of each triplicate.

Entry of PS-ODNs is determined by the physiological state of the cells and the mode of action of the permeabilizing agent

Both lag- and exponential-phase S. mutans OMZ175 cultures treated with zoocin A + γ32 P-ATS showed a rapid increase in the amount of ATS associated with each CFU (Fig. 2b and d). By contrast, while treatment of lag-phase cultures with penicillin and γ32 P-ATS also showed a rapid increase in the amount of ATS associated with each CFU (Fig. 2a), treatment of exponential-phase cells with penicillin and FBA did not affect the amount of ATS associated with each CFU (Fig. 2c). The rapidity with which PS-ODNs were able to enter zoocin A and penicillin treated cells was also different. Uptake of γ32 P-ATS by zoocin A–treated cells was detectable within minutes of addition of zoocin A to both lag- and exponential-phase cells (Fig. 2b and d), whereas the uptake of γ32 P-ATS by penicillin-treated cells was not detectable in exponential-phase cells (Fig. 2c) and detectable in lag-phase cells only after 90 min (Fig. 2a). The same pattern was observed in the degree of suppression of fba expression. It has previously been reported that a 37.354 ± 16.94-fold decrease in fba expression occurred 30 min after the zoocin A treatment of lag-phase cells (Dufour et al., 2011), while in this study, penicillin-treated cells showed only a 5.728 ± 0.23-fold decrease in fba expression at 30 min. This is probably a result of the differing modes of action of the two permeabilizing agents. Zoocin A, an endopeptidase, can lyse cells in any growth phase, and hydrolysis of peptidoglycan begins immediately upon addition of zoocin A to a culture (Simmonds et al., 1996). By contrast, cells just beginning to replicate in the lag phase are more susceptible to β-lactam antibiotics than are those in late log phase, and recently, it has been suggested that the assembly of a functioning divisome is a prerequisite for β-lactam-induced lysis of cells (Chung et al., 2009).

Concluding comments

Gram-positive bacteria like S. mutans have an array of surface exposed macromolecules such as teichoic acids that are covalently attached to the cell PG. These provide a negative charge to the cell surface (Ouhara et al., 2005), and it has been shown that alterations to the charge distribution in this layer can markedly change the susceptibility of S. mutans to cationic antimicrobial peptides (Mazda et al., 2012). It is possible that this negative charge could repel negatively charged PS-ODN and that the destruction of peptidoglycan by zoocin A and penicillin leads indirectly to a loss of this barrier. We conclude that a systematic search of compounds known to disrupt bacterial cell wall biosynthesis is likely to identify other agents capable of facilitating the entry of PS-ODNs into Gram-positive bacteria.


This work was undertaken with support from the Foundation for Research Science and Technology.